Archives of Surgery

doi: 10.1001/archsurg.142.5.415pmid: N/A

Archives of Surgery expresses appreciation to the individuals who served as peer reviewers in 2006. A Ali F. AbuRahma; David B. Adams; Mark B. Adams; Syed Ahmad; Gretchen M. Ahrendt; Nita Ahuja; Jeff Allen; Stanley W. Ashley B Christopher C. Baker; Charles M. Balch; Jeffrey L. Ballard; Adrian Barbul; Philip S. Barie; Erik S. Barquist; Stephen T. Bartlett; J. Augusto Bastidas; Robert W. Beart; James Becker; Stephen W. Behrman; John Benfield; David H. Berger; Ramon Berguer; Thomas R. Biehl; Walter Biffl; Timothy R. Billiar; H. Scott Bjerke; James Black; Kirby Bland; Richard J. Bold; Carl E. Bredenberg; John J. Brems; Bruce Brenner; L. D. Britt; Rebecca Britt; Malcolm Brock; F. Charles Brunicardi; Eileen Bulger; Richard Burney; Ronald Busuttil; Karyn L. Butler; Jo Buyske; Aman U. Buzdar C Kurtis Campbell; David F. Canal; John Carrucci; Sally E. Carty; Kathleen Casey; Peter Cataldo; David C. Chang; William Chapman; William G. Cheadle; Herbert Chen; Kenneth J. Cherry; Michael Choti; James M. Church; Robert Cilley; William G. Cioffi; Orlo H. Clark; Christine S. Cocanour; Nicholas P. Coe; Thomas Cogbill; Herbert E. Cohn; John A. Coller; Paul Colombani; Steven D. Colquhoun; John Conte; Edward E. Cornwell; Myriam J. Curet; Steven A. Curley D Alan Dackiw; Michael D'Angelica; R. Clement Darling; Kenneth Davis; S. Scott Davis; Merril Dayton; Steven A. De Jong; Giacomo A. De Laria; Michael E. de Vera; E. P. Dellinger; Clifford W. Deveney; Karen E. Deveney; Beth Ditkoff; Todd Dorman; Quan-Yang Duh; James Dunn; Julie A. Dunn E David Wayne Easter; A. Brent Eastman; David Efron; E. Christopher Ellison; Kimberly S. Ephgrave; Carlos O. Esquivel; David M. Euhus F Timothy C. Fabian; Peter J. Fabri; S. T. Fan; Michael B. Farnell; Carlos Fernandez-del Castillo; George Ferzli; Anne Fischer; James W. Fleshman; Roger S. Foster; Amy L. Friedman; Eric Robert Frykberg G Vivian Gahtan; Susan Galandiuk; Richard L. Gamelli; Susan L. Gearhart; Hugh Gelabert; David A. Geller; John J. Gleysteen; Leonard Go; Brian Goh; Philip Gordon; Jessica Gosnell; Ian Gough; Jon C. Gould; Clive S. Grant; A. G. Greeenburg; David G. Greenhalgh; Baiba Grube; Nelson Gurll H Marie Hanna; Alden H. Harken; John W. Harmon; Hobart W. Harris; Michael J. Hart; Andrea Hayes-Jordan; James C. Hebert; David M. Heimbach; Charles P. Heise; Richard F. Heitmiller; Jennifer Heller; David Herndon; Jonathan R. Hiatt; O. Joe Hines; James Holcroft; Eddie L. Hoover; Thomas J. Howard; David B. Hoyt; Glenn C. Hunter; John G. Hunter I Karl A. Illig; David K. Imagawa J Danny O. Jacobs; Lisa Jacobs; Raymond J. Joehl; Abraham John; Lynt Johnson; Gregory Jurkovich K Gordon L. Kauffman; Howard S. Kaufman; M. Margaret Kemeny; M. R. Khan; Andrew Klein; M. Margaret Knudson; Clifford Ko; Debra G. Koivunen; Robert A. Kozol; Irving L. Kron L Julie Lange; Keith D. Lillemoe; Alex Little; David Livingston; Edward H. Livingston; Jayme Locke; Donald Low; Ann Lowry; Fred Luchette; Kirk A. Ludwig M John Maa; Charles D. Mabry; Alexandra MacLean; Jana MacLeod; Thomas Magnuson; Ronald V. Maier; Martin Makary; Mark A. Malangoni; Michael Marohn; Jeffrey B. Matthews; Patricia May; David McFadden; Mary H. McGrath; Christopher R. McHenry; W. Scott Melvin; Brent Miedema; Ravi Moonka; Francis D. Moore; A. R. Moossa; Leon Morgenstern; Jon Morris; David S. Mulder; Michel Murr; Thomas A. Mustoe N David M. Nagorney; Lena M. Napolitano; Heidi Nelson; Richard Nelson; Leigh A. Neumayer; Daniel Neuzil; Jeffrey A. Norton; Patricia Joy Numann O Grant O'Keefe; J. Patrick O'Leary; Kim M. Olthoff; James O'Neill; Hector Orozco; Mary F. Otterson P Adrian Park; Marco A. Patti; Timothy Pawlik; Bruce A. Perler; Gerald W. Peskin; Jeffrey H. Peters; Henry A. Pitt; Timothy H. Pohlman; Raphael Pollock; Susan Pories; Russell Glen Postier; Richard A. Prinz R Henry Randall; Howard A. Reber; Robert Reed; Robert V. Rege; Melanie Richards; J. David Richardson; Malcolm K. Robinson; Steven Rothenberg; Grace S. Rozycki S Juan Sarmiento; Andrew W. Saxe; William P. Schecter; David Schoetz; Richard D. Schulick; Michael A. Schweitzer; Carol Scott-Conner; Timothy Sielaff; Howard Silberman; Mark L. Silen; Rache Simmons; R. Stephen Smith; Joseph Solomkin; David A. Spain; Bruce Stabile; Steven C. Stain; Michael J. Stamos; William E. Strodel; Harvey Sugerman; Marc Sussman; Lee L. Swanstrom; Stephen G. Swisher T John L. Tarpley; Gordon L. Telford; Jon S. Thompson; James Tomlinson; Guido Torzilli; Robert J. Touloukian; Manuel Trias; Theodore Tsangaris; Anthony Tufaro; Patricia L. Turner; William Turnipseed U Robert Udelsman; John A. Ulatowski V R. James Valentine; Hernan Vargas; George Velmahos W Brad W. Warner; Andrew L. Warshaw; Sharon Weber; John Wei; John Weigelt; Giles F. Whalen; Samuel Eric Wilson; David Winchester; Jordan Winter; Leslie Wise; Bruce M. Wolfe; Christopher Wolfgang; Phil Wolfson; Jan H. Wong; Sherry M. Wren Y Stephen Yang; J. Nilas Young; David Yuh Z Herbert J. Zeh; Michael E. Zenilman; Emmanuel Zervos; Wei Zhou

Kozol, Robert A.;Voytovich, Anthony

doi: 10.1001/archsurg.142.5.417pmid: 17515482

The traditional quartet of vital signs (pulse, blood pressure, temperature, and respiration) has always played a prominent role in the diagnostic armamentarium and remains, appropriately, at the head of the written or verbally presented physical examination. The earliest manifestations of infection, sepsis, occult bleeding, and cardiopulmonary dysfunction often become apparent when looking at the vital signs. In the case of a patient with a high spinal cord injury and autonomic dysreflexia, profound hypertension and bradycardia may be the only clue to a medical or surgical crisis that otherwise remains silent because of interrupted pain pathways. Despite their value, these trusted measurements have always had their pitfalls and limitations. Blood pressure readings can be misleading because of measurement variation and error inherent in improper cuff width, auscultatory gap (the silent phases between systolic and diastolic determinations), and even the auditory acuity of the listener, not to mention the profound moment-to-moment variation in readings, which may be normal with exercise or anxiety. Students are taught not to reveal that they are counting respirations, for as patients become self-conscious, the rate will surely be affected. A carefully counted pulse can become chaotic the minute the observer walks away. Many will remember watching an oral temperature being taken for several minutes to avoid a spurious reading due to a hot or cold drink, an interruption to speak, or even a malingering patient rubbing the thermometer on the bed sheet. We have seen dramatic improvement in validity and reliability with the appearance of technologic advances in the form of telemetry, continuous ambulatory blood pressure monitoring, and tympanic membrane temperatures, all providing solid moment-to-moment data. Now, the emergence of the fifth vital sign presents new challenges. In the late 1990s, the measurement of pain was introduced as the fifth vital sign.1 We have no device or instrument with which to measure pain. We cannot count it or measure it with a timepiece. Instead, we rely on the patient's own interpretation of the magnitude of his or her pain. This is indicated on a 0-to-10 scale or on a visual analogue scale. The fifth vital sign remains totally subjective. A sign on physical examination is an observation. Symptoms are sensations reported by the patient. Thus, pain is a symptom and not a sign. For adults, reporting pain using the traditional 0-to-10 scale is limited for a number of reasons. It is a Likert scale and, as such, is an ordinal measurement tool. In other words, we cannot assume that the difference between a 2 and a 5 is the same as that between a 6 and a 9. It is even less likely that one patient's 5 is similar to another’s. The limitations of Likert scales have peen pointed out by others.2,3 Thus, although subjective pain scores may be helpful to approximate a trend in a given patient, they lack the precision and reliability (case to case or day to day) to exclusively direct analgesic dosing. Health care providers (a diverse group with different levels of training and experience) react to this subjective response with the administration of analgesics, frequently opioids. It is not surprising that this imprecise process frequently results in either underdosing with poor pain control or overdosing with the potential for dangerous complications. Since the use of the fifth vital sign, there has been a ground swell of effort to encourage health care providers to attend to patients' pain and to strive for pain scores approaching zero.4,5 Health care organizations have worked hard to develop pain management guidelines.6,7 A search for outcomes studies before and after the institution of the fifth vital sign yielded little useful data. Most studies looked at patient satisfaction but not at adverse events.8-11 The best data are from Vila et al,12 who compared patient satisfaction and respiratory events before and after the establishments of the fifth vital sign and pain management standards and found that patients were more satisfied with pain control during the use of the new guidelines. At the same time, they found more than double the rate of opioid-related respiratory events. Whether the establishment of the fifth vital sign or of new pain management efforts have increased opioid-induced adverse events on a national level remains unknown. In 2003, we published a prospective evaluation of sedation levels in postoperative patients receiving opioids.13 We were surprised to see the number of postoperative patients on patient-controlled analgesia or nurse-controlled dosing reaching sedation levels equal to those seen during conscious sedation. Procedures performed under conscious sedation now require special and frequent monitoring beyond that afforded to postoperative patients on the surgical floor. We followed that study with an evaluation of opioid-related respiratory events during the postoperative period.14 We noted clustering of such events during the first 24 hours postoperatively. We also noted an increased risk of such events in elderly patients. Lucas et al15 evaluated case surveys from trauma centers since the implementation of pain management standards. They determined that overmedication with opioids was responsible for 32 of the 867 deaths in the study and that similar overmedication contributed to deaths in another 14 of the patients. We do know that opioids impair judgment, yet we rely on our drugged and sedated patients to measure their fifth vital sign. Every surgeon has seen patients who could barely open their eyes respond “10” when awakened for the measurement of their pain level by a nurse. The administration of substantial doses of opioids to such a patient may lead to respiratory depression. Clearly, we need to refine our measurement and interpretation of the fifth vital sign. We must also optimize the sparing of opioids (eg, nonsteroidal anti-inflammatory drugs, cognitive intervention, acupuncture, and epidural catheters). The development of a more valid and precise measure of pain intensity is confounded by a variety of emotional and physiologic modifiers in both patient and observer. While the evaluation and measurement of pain is imperfect in the usual adult patient, the situation is more difficult in very old or very young patients. With regard to children, it may take until the age of 3 or 4 years to be able to describe pain with any degree of accuracy regarding location, quality, intensity, or duration.16 Furthermore, a child's frame of reference is based on prior experience, so considerable variation from case to case can be expected. There are a number of pediatric self-reporting tools, including the Oucher scale, that allow subjects to match their pain with photographs of children's faces (available in a variety of races) depicting increasing pain intensity.17 The Varni-Thompson Questionnaire has been developed to be more helpful in the case of chronic or persistent pain as, for example, in juvenile arthritis.18 The Child Facial Coding System has been described by Gilbert et al.19 Using videotapes of children admitted to a postanesthesia care unit, certain elements of facial expression were noted, including open lips, lowered brows, eyes squeezed shut, and raised cheeks. Facial action summary scores were positively correlated with visual analog ratings suggesting that this Child Facial Coding System measure has convergent validity. Elderly patients have more difficulty with self-reporting pain scales than younger adult patients.20 This is often further confounded by the presence of dementia. More than 50% of elderly patients in nursing homes have some degree of dementia. In addition, the type of dementia or neurocognitive disorder has an influence on pain and pain reporting.21 More than a dozen pain assessment tools have been developed for the elderly.22 Clearly, there are additional challenges as a result of cognitive and sensory impairment, language barriers, and cross-cultural gaps, any or all of which may be magnified when dealing with the elderly. There is a growing understanding of underlying physiologic mechanisms explaining interindividual differences in response to pain, including genetic polymorphism directing the metabolism of catecholamines, which alter activity in μ-opiate receptors.23 Additionally, recent studies suggest that nociceptive pathways interact with sympathetic, homeostatic, and emotional centers that—like cold, hunger, and thirst—drive adaptive behaviors.24 Thus, there seems to be genetic, psychological, and physiological determinants shaping the human response to pain. Finally, in our effort to establish valid and precise measures of pain and suffering, we encounter variations in the observer. The hierarchical structure of many academic medical centers and departments, and the lack of long-term relationships can undermine empathic accuracy. A colleague recently recounted an experience in an underdeveloped country in which he was astounded by the apparent indifference to the sufferings of patients lying in a primitive open ward. An overworked and exhausted nurse admitted: “These people are not real to us, Doctor!” In a study by Jackson et al,25 functional magnetic resonance imaging was performed on healthy nonphysicians who were shown photographs of individuals in painful situations (eg, fingers slammed in a door) as well as control images of nonpainful situations. When shown the painful images, functional magnetic resonance imaging revealed changes in the insular cortex (a key hub in the interoceptive system), the thalamus, and cingulate gyrus—all regions involved in pain processing. Furthermore, the authors suggest that variations in the subjects' assessment of the pain were partially a function of cerebral communications between the perception of another's suffering and experiencing that trauma themselves. Allcock26 has reviewed the many factors affecting nurses' assessment of patient's pain. In conclusion, it is important to appreciate the limitations in the use of pain as the fifth vital sign. As mentioned, pain is a symptom and not a sign. The current 1-dimensional measure of pain severity is flawed and cannot reliably guide analgesic therapy. Our traditional vocabulary in case presentation betrays our shaky confidence in patient self-reporting, as we state that the “patient claims” or the “patient denies,” while we assert that the radiograph “shows” or “proves.”27 The challenge is upon us to develop more precise and objective measures of pain. Until something better comes along, we had better dust off the time-honored bedside skill of listening, observing, and interpreting, because at this point in time, technology and existing tools have little to offer. Correspondence: Dr Kozol, University of Connecticut, Department of Surgery, 263 Farmington Ave, Mail Code 3955, Farmington, CT 06030 ([email protected]). Author Contributions:Study concept and design: Kozol and Voytovich. Drafting of the manuscript: Kozol and Voytovich. Critical revision of the manuscript for important intellectual content: Kozol and Voytovich. Administrative, technical, and material support: Kozol and Voytovich. Study supervision: Kozol and Voytovich. Financial Disclosure: None reported. References 1. McCaffery MPasero CL Pain ratings: the fifth vital sign. Am J Nurs 1997;9715- 16PubMedGoogle Scholar 2. Pell G Use and misuse of Likert scales. Med Educ 2005;39970- 971PubMedGoogle ScholarCrossref 3. Jamieson S Likert scales: how to (ab)use them. Med Educ 2004;381217- 1218PubMedGoogle ScholarCrossref 4. Apfelbaum JLChen CMehta SSGan TJ Postoperative pain experience: results from a national survey suggest postoperative pain continues to be under managed. Anesth Analg 2003;97534- 540PubMedGoogle ScholarCrossref 5. Filos KSLehmann KA Current concepts and practice in postoperative pain management: need for a change? Eur Surg Res 1999;3197- 107PubMedGoogle ScholarCrossref 6. American Society of Anesthesiologists Task Force on Acute Pain Management, Practice guidelines for acute pain management in the perioperative setting: an updated report by the American Society of Anesthesiologists Task Force on Acute Pain Management. Anesthesiology 2004;1001573- 1581PubMedGoogle ScholarCrossref 7. Gordon DBDahl JLMiaskowski C et al. American pain society recommendations for improving the quality of acute and cancer pain management: American Pain Society Quality of Care Task Force. Arch Intern Med 2005;1651574- 1580PubMedGoogle ScholarCrossref 8. Seers KCrichton NCarroll DRichards SSaunders T Evidence-based postoperative pain management in nursing: is a randomized-controlled trial the most appropriate design? J Nurs Manag 2004;12183- 193PubMedGoogle ScholarCrossref 9. Sauaia AMin SJLeber CErbacher KAbrams FFink R Postoperative pain management in elderly patients: correlation between adherence to treatment guidelines and patient satisfaction. J Am Geriatr Soc 2005;53274- 282PubMedGoogle ScholarCrossref 10. Stomberg MWWickstrom KJoelsson HSjostrom BHaljamae H Postoperative pain management on surgical wards: do quality assurance strategies result in long-term effects on staff member attitudes and clinical outcomes? Pain Manag Nurs 2003;411- 22PubMedGoogle ScholarCrossref 11. Meissner WUllrich KZwacka S Benchmarking as a tool of continuous quality improvement in postoperative pain management. Eur J Anaesthesiol 2006;23142- 148PubMedGoogle ScholarCrossref 12. Vila H JrSmith RAAugustyniak MJ et al. The efficacy and safety of pain management before and after implementation of hospital-wide pain management standards: is patient safety compromised by treatment based solely on numerical pain ratings? Anesth Analg 2005;101474- 480PubMedGoogle ScholarCrossref 13. Taylor SVoytovich AEKozol RA Has the pendulum swung too far in postoperative pain control? Am J Surg 2003;186472- 475PubMedGoogle ScholarCrossref 14. Taylor SKirton OCStaff IKozol RA Postoperative day one: a high risk period for respiratory events. Am J Surg 2005;190752- 756PubMedGoogle ScholarCrossref 15. Lucas CEVlahos ALedgerwood AM Kindness kills: the negative impact of the Visual Analog Scale. J Am Coll Surg In pressGoogle Scholar 16. Belter RMcIntosh JFinch ASaylor C Preschoolers' ability to differentiate levels of pain: relative efficacy of three self-report measures. J Clin Child Psychol 1988;17329- 335Google ScholarCrossref 17. Oucher! http://www.oucher.org/. Accessed May 2006 18. Varni JWThompson KLHanson V The Varni/Thompson Pediatric Pain Questionnaire, I: chronic musculoskeletal pain in juvenile rheumatoid arthritis. Pain 1987;2827- 38PubMedGoogle ScholarCrossref 19. Gilbert CALilley CMCraig KD et al. Postoperative pain expression in preschool children: validation of the child facial coding system. Clin J Pain 1999;15192- 200PubMedGoogle ScholarCrossref 20. Gagliese LMelzack R Chronic pain in elderly people. Pain 1997;703- 14PubMedGoogle ScholarCrossref 21. Scherder EOosterman JSwaab D et al. Recent developments in pain in dementia. BMJ 2005;330461- 464PubMedGoogle ScholarCrossref 22. Zwakhalen SMHamers JAbu-Saad HHBerger M Pain in elderly people with severe dementia: a systematic review of behavioural pain assessment tools. BMC Geriatr 2006;63PubMedGoogle ScholarCrossref 23. Zubieta JKHeitzeg MSmith Y COMT val158met genotype affects mu-opioid neurotransmitter responses to a pain stressor. Science 2003;2991240- 1243PubMedGoogle ScholarCrossref 24. Craig AD A new view of pain as a homeostatic emotion. Trends Neurosci 2003;26303- 307PubMedGoogle ScholarCrossref 25. Jackson PLMeltzoff ANDecety J How do we perceive the pain of others? a window into the neural processes involved in empathy. Neuroimage 2005;24771- 779PubMedGoogle ScholarCrossref 26. Allcock N Factors affecting the assessment of postoperative pain: a literature review. J Adv Nurs 1996;241144- 1151PubMedGoogle ScholarCrossref 27. Donnelly WJ The language of medical case histories. Ann Intern Med 1997;1271045- 1048PubMedGoogle ScholarCrossref

Livingston, Edward H.

doi: 10.1001/archsurg.142.5.419pmid: N/A

Emphasis on pain as the fifth vital sign has become the latest fad in medicine receiving undue attention among regulators and health care commentators. Kozol and Voytovich strike at the heart of this issue with a clear description regarding the differences between pain and other physiologic processes. As Kozol and Voytovich point out, pain is not a vital sign despite the popular new jargon; they elegantly describe the vital signs and how they are measured. Early in medical school, we are taught that signs are physiological processes that can be observed and measured. Symptoms are patients' subjective description of their disease state. Pain is a symptom and not a sign. As such, measurement tools will always fall short in precision and accuracy for pain assessment. Promoting pain as a measurable sign leads to misuse of pain measurement tools, overadministration of pain medication, and a rise in complications attributable to pain management. We live in a society that is unwilling to accept risk and expects solutions for all problems that arise. Laypeople attribute uncontrolled pain to inadequacy of the health care system. Consequently, activists demand that physicians provide complete pain relief for those with chronic problems. Litigation has worsened the situation because inadequate pain management has become the basis of several successful lawsuits against physicians. Consider Bergman v Eden Medical Center (case No. H205732-1, Superior Court of California, Alameda County, 2001): an 85-year-old man with known compression fractures and a poorly defined pulmonary process was admitted to the hospital with pain. While in the emergency department, he nearly had a respiratory arrest following morphine administration. The admitting physician treated his pain with various combinations of fentanyl citrate patches, meperidine hydrochloride, and acetaminophen and hydrocodone bitartrate, being cautious because of the respiratory arrest. While in the hospital, the nurses documented his pain as ranging from 7 to 10 on the 1-to-10 Likert scale. The physician testified that he was reluctant to use too much pain medication because of the patient's advanced age, pulmonary disease, and prior morphine intolerance. Once the patient was home, the family sought more pain medications from another physician who provided morphine, and the patient died a day later. The physician who cared for this patient in the hospital lost a $1.5 million judgment for elderly abuse following the courtroom testimony of a physician expert witness who described his care as “appalling” and “egregious.” Documenting pain as a vital sign resulted in a medical record that, to a lay jury, appeared to be replete with objective evidence for an untreated malady: pain. They failed to grasp the physician's dilemma of balancing pain relief with respiratory depression. Although the physician's concerns were valid (the patient died shortly after morphine administration), the jury focused on the untreated pain. Considerable harm can occur if we allow medical records to contain seemingly objective documentation of information, such as pain scores, which are inherently subjective. What should we do? We must insist on proper use of terminology. The concept of pain as a fifth vital sign must be abandoned. Society has made it clear that pain management must remain an important aspect of medical care. The physician community should insist that pain assessment remains in the proper context: pain is a subjective reporting of a patient's sensation and cannot be objectively measured or scaled. The sensation of pain is highly variable from patient to patient and cannot be precisely quantitated. Pain assessment must include an evaluation of structural problems causing pain, underlying psychological factors that affect the patient's ability to manage their pain, and the likelihood that medications will be effective. Treatment of pain with opiates risks creating addictive behaviors. As a physician community, we must abandon the notion that pain can be objectively quantified. We cannot allow laypeople to believe that there is a risk-free solution to pain management and that pain can always be controlled. Correspondence: Dr Livingston, University of Texas Southwestern Medical Center, 5323 Harry Hines Blvd, Room E7-126, Dallas, TX 75390-9156 ([email protected]). Financial Disclosure: None reported.

Thirlby, Richard C.

doi: 10.1001/archsurg.142.5.423pmid: 17515483

Abstract It is an honor to stand before yoeu with the opportunity to deliver the Presidential Address for the annual meeting of the Western Surgical Association (WSA). I have gained more from the WSA than I could ever return. To serve as the secretary for 5 years and the president this year represents the peak of my professional life. It is an honor that I will always cherish, and for which I will be forever grateful. My first task today is to thank several people who are responsible for my position at the podium. Anyone in the audience who knows me well realizes that I will be unable to accomplish that goal in the manner it deserves without completely dissolving into tears. Regrettably, by necessity, I must limit my remarks simply to saying “Thank you” to my family—Pat, Margie, and David. I would also like to introduce and thank the person who has run my surgical practice during my tenure at Virginia Mason, Sandhya Mishra. Finally, I would like to dedicate this talk to my father, Dr Richard Leeson Thirlby. It goes without saying that I have been contemplating the topic and content of my presidential address for 364 days. I reread many of the previous addresses of the presidents of the WSA. It became evident immediately that I could not match the eloquence or insight of my predecessors. I was left to abide by the sage credo that guides much of what I do: “Play to your strengths.” I am a very positive person. I believe much of my success results from an ability to ignore frustration and to act on the positives; hence, my topic. All too frequently, I hear tirades of surgical colleagues that could provide the content for another talk entitled “Top 10 Reasons Why I Hate My Job.” I will leave that talk for another individual. My objective today is to remind you or convince you that you practice the greatest profession on earth. View LargeDownload Richard C. Thirlby, MD Another guiding principle of any good address is to “know your audience.” From this principle emerges a potential problem with today's address. My audience today is dominated by several hundred successful, experienced general surgeons. Without sounding pretentious, my hope is that this address will be published, and so one day be read by young individuals contemplating a career in medicine and, eventually, surgery. This assumption affected my choice of topics you are about to hear. What could I say to convince a high school student, a college student, or a medical student that surgery is a great career? I suspect that some elements that you think unimportant may be very important to a college undergraduate. Thus, a potential problem with this talk is the inevitable “constructive feedback” I will receive from my colleagues about my Top 10 choices. They will point out things I forgot! Without doubt, I have omitted things and/or incorrectly ordered or ranked my reasons. I encourage you to let me know the errors of my ways. If I foster an enthusiastic discussion between you and your colleagues, challenging my thoughts with better ideas, I will have accomplished my goal today. So without further ado, with apologies to David Letterman, here we go. The “Top 10 Reasons Why General Surgery Is a Great Career.” 10. training is fun (you'll never forget it) and training never stops Surgical residency might be considered by some to be a deterrent to making general surgery a great career. I disagree. Is it arduous? Yes. Is it fun? Absolutely. The excitement starts on the first day of internship. Stephen Evans put it nicely in a recently published editorial: “No other profession comes close in terms of duplicating the level of intensity, fear, or anxiety that comes with the first day of internship.”1 This intensity continues for the duration of the training. It provides young people with great life experiences and memories. Let me illustrate the latter claim with an observation. If 2 or more surgeons who trained together are placed in a room together for more than 5 minutes, the conversation inevitably turns to recollection of experiences of residency. As any spouse of a surgeon can attest, stories of residency dominate conversation during alumni receptions at the American College of Surgeons Clinical Congress. Although the events may have occurred decades ago, they are recalled vividly, as if they occurred just yesterday. When one spends 80 to 100 hours per week with people in high-stress circumstances, personalities emerge. The experiences are remarkable and life-changing, and the friendships are enduring. They cannot be matched by any other career. The training never stops. Surgical or medical training does not stop with the end of residency. I can think of few, if any, careers in which the necessity of “continuing education” is so relevant. The judicial system has not changed appreciably since 1776; medicine and surgery change weekly. I asked our medical librarian to tally the English-speaking journals relating to general surgery. Her count totaled 75. Assuming the majority of these journals publish 12 issues annually, there are about 900 issues per year, probably in excess of 9000 manuscripts per year. What other career provides so much opportunity for mental advancement? There will be thousands of articles published this year that will be available to stimulate us to improve. Surgery is not unique among medical careers in this regard; however, I cannot think of another career that provides such stimulation, advancement, and change for an entire career. 9. job security Job security is a good thing. For decades, there has been debate regarding the supply and demand of physicians and surgeons. I have neither the time nor the desire to bore you with the details of the 1910 Flexner report, nor the multiple iterations of the Graduate Medical Education National Advisory Committee (or GMENAC) reports. Suffice it to say, these folks were not good at predicting things. This committee predicted that in 1990 there would be about 35 000 general surgeons in a country that needed about 23 000 surgeons, a surplus of about 12 000 surgeons.2 What happened? In 1990, there were about 29 000 surgeons in a country that needed about 28 000 surgeons.2 If I do the math correctly, that is about a 12-fold error. If any of us missed our targets by 12-fold on a consistent basis, we would be out of a job. The complexity of the variables involved exceeds the ability of most to make valid predictions. The current consensus is that there will not be enough general surgeons for the foreseeable future. There are several reasons. First, America is aging and general surgeons treat aged patients. At present, there are about 35 million Americans over the age of 65 years.3 By 2020, there will be about 53 million persons over the age of 65 years. The aged population, therefore, will increase by about 50%. This aging population will result in huge growth in demand for surgical services. It has been projected that the US population will increase by about 18% in the next 20 years.4 However, because of the aging population, it has been projected that the workload of general surgeons will increase by about 30%.3,4 How many general surgeons will there be? A recent publication by Stitzenberg and Sheldon5 suggests that progressive specialization within general surgery likely is going to decrease the numbers of general surgeons being produced and add to the complexity of workforce planning. Although about 1000 persons graduate from residencies in the United States annually, about 750 are currently entering fellowship training. Thus, the number of persons graduating and entering traditional general surgery practice in 2004 is only about 250 per year. Furthermore, surgeons are retiring earlier.6 The bottom line is that in the next few years, there will be a significant shortage of general surgeons, roughly 5000 or a 15% deficit. The Association of American Medical Colleges (AAMC) recently published a consensus statement on the physician workforce.7 They conclude that there will be a severe shortage of physicians in the next few decades. They suggest, therefore, that entry-level positions be increased in both medical schools and residency programs. They calculate that medical school enrollment should be increased by 30%.7 Similarly, the positions in graduate medical education programs, or residencies, should be increased by the same fraction. In a moment of remarkable insight, they go on further to recommend that “the AAMC should provide students, physicians, programs and hospitals with the best available and timely data on physician workforce needs in order to support informed decisions.”7 That is, the AAMC will give students information suggesting that we need general surgeons. However, I have a high level of cynicism toward those who try to predict the future. If given the opportunity to counsel young people, I would rather choose to state facts of the present and/or relate empirical observations to illustrate the point. As a program director in general surgery for 14 years, I have never had any trouble getting one of our graduates a good job; most have many great job offers. As a practicing surgeon, I can think of only 1 or 2 examples of good surgeons who left a community or a practice because of politics or lack of income or success of their practices. The bottom line is that a good general surgeon will be successful and have total job security for the next several decades. Name another desirable profession that can make that claim. You have total job security, and job security is a good thing. 8. the pay is not bad I live in Seattle, Wash. Not a week passes that does not provide me the opportunity to meet a “Microsoft millionaire.” A Microsoft millionaire is usually between the ages of 40 and 50 years, is worth millions, and has not worked for about 5 years. They do serve a purpose. They have the time to coach our children's Little League teams, which, regrettably, busy general surgeons do not. Yet, it does not seem fair. I work harder, am better educated and more skillfully trained, and deal with significantly more stress than they do. The surgical locker rooms of America are filled with surgeons complaining about inequities of financial compensation. I could relate to you countless examples of corporate greed, of insurance executives collecting obscene compensation packages, but I will not do that. Rather, let us look at the facts. The facts are that the average general surgeon in the United States has an annual salary of $280 000 per year.8 Pay increases progressively for the first 17 years of practice, peaking at about $300 000 per year. This pay scale puts him or her in the top 99.9% of income of employed Americans. The median household income in the United States is $44 000 per year.9 The average annual salaries of lawyers, computer/information systems, engineers, financial managers, financial advisors, and architects are $110 000, $102 000, $66 000 to $97 000, $96 000, $82 000, and $68 000, respectively.10 The Bureau of Labor Statistics calculates the average hourly wage for all professions in the United States.11 They provide data for 427 possible jobs in 2002. The highest hourly wage in the US is airline pilots, $95 per hour. The catch here is that pilots only work an average of 23 hours per week. The lowest salary, the 427th in the list, is that of waiters and waitresses, $4 per hour. The second highest hourly wage in the United States is that of physicians, $61 per hour. Physicians average 41 hours per week.11 The data are not broken down into physician specialty groups, but general surgeons are among the highest-compensated physicians. Lawyers and financial managers are ranked 14th and 34th in hourly wage, respectively. Are there inequities when it comes to compensation of general surgeons? Of course there are. If you are a general surgeon, 17 years into your career you likely have annual earnings ($300 000) that are substantially less than those of the radiologists or radiation oncologists in your institution ($450 000)—certainly not a fact that is based on common sense or justice. Even the anesthesiologist ($340 000) drinking coffee in the lounge while you are performing a Whipple resection is making more than you are. That is not fair or appropriate. However, the fact remains that general surgeons are paid very, very well. Our salaries afford us a lifestyle that is better than 99.9% of people in Western civilization. We should be embarrassed if we complain in public about our pay. 7. your mother will be proud of you The prestige of physicians and surgeons in the eyes of the American public has declined over the last 4 to 5 decades. The reasons for this are complex, and time does not permit me to discuss them at length. The condensed version or explanation of this decline in prestige is that we have “what we deserve.” My father practiced urology in northern Michigan for 40 years. One of his pet peeves was the phenomenon of physicians driving expensive cars with personalized license plates advertising the fact that they are physicians. I recently saw a Porsche with a license plate that said “skin doctor.” How can we, as a profession, complain that we have lost the respect of the American public when we do things like that? Thankfully, however, the fact remains that our patients, and the residents of the communities we serve, respect us a great deal. Surgery was ranked as the most prestigious occupation in a recent national survey, outranking the positions of college presidents, astronauts, big-city mayors, lawyers, and all other physicians. Recent data compiled by the National Opinion Research Center concluded that surgeons are the most respected professionals in the United States.12 We are highly respected for our technical skills, and the public values our ability to deal with very high on-the-job stress and appreciates our work ethic. Surgeons are respected in their communities. At work, surgeons are treated with respect and admiration by all levels of coworkers, from nurses to the hospital maintenance workers. Surgeons are revered by their patients. Your family is proud of you; even your mother-in-law is proud of you. Most important, however, your mother is proud of you. 6. surgeons have panache: the surgical personality and the culture of surgery As described before, surgeons are admired by the public. While you may argue that your life bears little resemblance to that portrayed by surgeons in Hollywood and modern television, there must be some justification for the dozens of movies or television shows depicting surgeons in action. The panache of the surgical personality is glorified by Hollywood. Surgeons on the movie screen are portrayed by people like Elliot Gould or Donald Sutherland, on television by the beautiful young men and women of Grey's Anatomy. Internists are portrayed by people like Robert Young in Marcus Welby, MD. The surgical personality is part of the culture of surgery. The recently published book, The Scalpel's Edge was written by a psychologist who had the opportunity to observe a team of academic surgeons.13 The author, Pearl Katz, very accurately describes the surgical personality and the culture of surgery. Like the writers of MASH, she observed that surgeons have a demeanor of confidence and apparent arrogance. She then attempts to explain why that might be the case. She suggests that: The kind of work that surgeons do influences their demeanor and behavior with others. What other profession makes decisions and takes risks that literally control patients' lives or deaths on a daily basis? Who but a surgeon routinely and boldly cuts into the most intimate depths of people's live bodies, penetrates their innermost body cavities? . . . Surgeons' detachment from their patients may be understood as necessary protections from these routine sights, smells, acts, and dramatic confrontations with mortality. Their demeanor of confidence and apparent arrogance may be partially explained by the almost superhuman requirements of their work.13(pviii) Katz observed that the surgical personality is characterized by a preference for acting over not acting. It has been part of the culture of surgery, since it was distinguished from medicine in the second century. “Surgeons' reluctance to admit doubt and uncertainty or error was likely to have permitted them to be sufficiently bold to carry out extremely difficult and risky procedures.”13(pviii) How is it that surgeons can deal with the stress of performing life-threatening operations on a daily basis? Katz's hypothesis is interesting. She was fascinated by the rituals of the glove and gown process, of the draping procedure that shielded the face and personality of the patient from the surgical team. She believes that: The rituals that are enacted in the operating room also shield surgeons from the emotions of other patients and from their own emotions. . . . In addition to separating the head of the patient from the rest of his body, operating room rituals facilitate a mind-set which focuses [on the technical aspects of the case, not the patient] . . . they facilitate surgeons' existing penchant for action and to avoid responding to distressing emotions.13(p110) The dominant themes of surgical culture—action, heroism, certainty, and optimism—are not compatible with identification with, or communication with, helpless, hopeless patients. . . . The process of repressing normal feelings of empathy is necessary during surgery. . . . Otherwise, they [the surgeons] would be forced to experience intolerable feelings of disgust, horror, fear, or excitement, and their own confrontations with death.13(p123) For decades, the culture of surgery was dominated by men. Is the traditional surgical personality by necessity masculine? The answer is no. Will the surgeons of the future have “panache”? The answer is yes. Katz accurately observed that “Surgical culture perpetuates itself by the recruitment and training of residents through a long apprenticeship process. Residents are often attracted to surgery by the image of the surgeon, with its prestige, power, and heroic mystique that is shared by most people in North American culture.”13 There is nothing related to gender in this attraction to surgery and the surgical personality. I would agree with Katz, who believes that women in surgery will make our profession better. The new surgical heroes may rather be those who can admit doubt and uncertainty when it exists, and can admit limits and be comfortable with palliation. They may communicate sensitively with patients. . . . The new surgical heroes may not be those men who make decisions and take risks for their patients, but rather those men and women who make decisions and take risks with their patients.13(p210) Women in surgery will enhance the prestige of our profession, while maintaining the culture and panache of surgery and, more important, improving the product. The culture of surgery is further illustrated by the WSA. Name another career that has resulted in organizations like the WSA, that drives busy professionals to spend thousands of dollars to spend time, learn, and socialize with colleagues from around the country. There is a bond, cemented by a profession that is greater than any other, that connects us all. In summary, if you choose surgery for a career, you will join a culture like no other. 5. you will have “heroes”; you will be a hero The Halstedian teaching model, the apprenticeship model, provides mentors. Mentors become role models and they become heroes. While many careers potentially foster a “hero-worship” relationship between the pupil and the teacher, I would contend that the relationship between surgeons and their mentors transcends that of all other careers. Claude Organ incorporated a feature called “Surgical Reminiscences” in the Archives of Surgery that illustrated the uniqueness of this bond. Each article describes what I call “hero-worship” between student and mentor. Western Surgical Association member Byron McGregor talked about his relationship with fellow WSA member and WSA president Chester McVay, who quoted the 12th-century chancellor Bernard in his classic anatomy text, saying that “we are like dwarfs seated on the shoulders of giants. If we see more and further than they, it not due to our own clear eyes or tall bodies, but because we are raised on high by their bigness.”14(p1) Byron astutely observed that McVay transformed himself from a humble dwarf into one of the very giants he so admired, “thereby making room up there for the rest of us . . . and, the view is fine!”14(p1) Doctor Organ wrote about his first interaction with Lester Dragstedt. Claude was a junior resident doing basic science research in a gastrointestinal physiology laboratory; the legendary Dr Dragstedt was a visiting professor. During the visit, a young Dr Organ had the opportunity to discuss his research with Dragstedt during a one-on-one meeting in his laboratory. Claude then transported Dragstedt to his hotel. As they departed, Dragstedt presented Claude with his card and offered to communicate with him about the research project. It was a moment, Claude observed, “in my surgical training and surgical experience I shall never forget.” He observed that “others might possess the card of legendary sports figures, but I had the card of Lester L.R. Dragstedt, MD, FACS, Professor of Surgery and Physiology.”15(p1) My mentors are heroes. I could list dozens of heroes . . . many are sitting in this room. I need to acknowledge two, both members of the WSA: Bill Fry and Robert McClelland or “Dr Mac.” It was my privilege to sit on Bill Fry's shoulders for 10 years in Dallas. He is the most talented surgeon I know. Dr Mac, the originator and editor of Selected Readings, has done more for surgical education than any surgeon ever. Dr Mac's knowledge of the surgical literature dwarfs that of any other individual on earth. A Dr Mac–trained surgeon knows that good surgical care can only be delivered by a surgeon who gets better every day. The best way to accomplish that goal is to read the surgical literature every day. I have photos of Drs Fry and McClelland in my office. I look at them daily for inspiration. To quote Dr McGregor, the view from their shoulders is “just fine.” Every surgeon in this room is somebody's hero. Whether you are in private practice in a small community hospital or a department chair at a major university, there are young surgeons, young medical students, or young people thinking of a career in surgery who had the privilege to work with you in a hospital and think of you often. You are somebody's hero. Of all the letters I have received from graduates of our program thanking me for being part of their training, perhaps the one that affected me most profoundly ended like this. “How does a student thank a mentor? . . . How do we thank you for those gifts? We emulate you.” What a great career. We have a profession and a culture that fosters heroes, we sit on the shoulders of giants, and we too can impact others' lives profoundly. 4. there's spirituality if you want it What do I mean by this? I can illustrate what I do not mean with a joke you have all heard. Although there are many variations on the joke, the short version goes something like this: “What is the difference between surgeons and God? God knows he is not a surgeon.” My circulating nurse of 17 years, Thea Nortness, told me that joke. I do not recall exactly the circumstances that prompted it, but I have encouraged her to retell the joke any time she feels it is appropriate. We would all probably benefit from having a hospital employee tell us this joke on a frequent basis. What I do mean is that, as surgeons, we are dealing with people on a daily basis who are using their spirituality to navigate through very stressful and/or life-threatening circumstances. A recent poll shows that 59% of the US population considers religion extremely, or very, important in daily life. However, being a surgical patient is not “daily life.”16 Nearly 100% of our patients use their spirituality during their surgical care.17 How is that relevant to the career of surgery? If spirituality is important in your life, if spiritual interactions give you strength, give you peace, surgery will make your life better every day. Surveys of hospital inpatients report that 77% of patients believed their physicians should consider their spiritual needs, 48% wanted their physicians to pray with them, but 68% said no physician had ever inquired about their spiritual or religious needs. . . . 94% thought it appropriate for physicians to inquire about their spiritual beliefs if they became gravely ill.16-20 How should these data influence my surgical care? I am not sure. I do know that the bond that can form between surgeon and patient is profound. The act of making a skin incision may be a simple technical exercise, but as the tissues open, the patient and the surgeon are bonded. From the patient's perspective, undergoing general anesthesia, being put to sleep while lying prone, is cause for thought. Is that spirituality? It usually is. All of us acknowledge the value of the recent emphasis in medical care on “patient-centered care” and the paramount importance of health-related quality of life. As John and Maggie Tarpley21 observe, if the majority of our patients view spirituality to be important in their ability to deal with illness, and “if the patient-centered approach to medicine is the gold standard, then spiritual aspects of each person must be considered.” Spiritual well-being is a major component of health-related quality of life.22 However, despite increasing evidence that patients would like their physicians to do so, spiritual issues are rarely addressed by 21st-century Western physicians.21 Clinicians who ignore the spiritual concerns of patients are, in effect, asking many patients to alienate themselves from beliefs that deeply define them. Astute clinicians pick up clinical clues from patients.23 The Hindu amulet, a copy of the Qur’an, rosary beads, a Bible, or Shabbat candles on the nightstand next to the bed may be as much communication to the surgeon as they are spiritual aids to the patient. They are signs of what the patient holds most dear. All that may be needed is a simple, open-ended question to engage the patient on a spiritual level. Showing respect for such defining features of a patient's life may constitute a healing act and can be integral to the care of the “whole patient.”23 How many careers can document that 90% of their customers use prayer or spirituality to help them through their relationship with the provider of the service? I can only think of a couple: airline pilots and casino dealers. In the case of commercial airlines, there are few atheists in the passenger section. However, the pilot has no personal or emotional contact or spiritual bond with the client. In the case of casinos, the dealer is the devil. The problem for me is that none of my training taught me to address the spiritual needs of my patients. How should physicians examine and engage religion in the lives of their patients? As a program director in general surgery for the last 14 years, I have supervised the curriculum of about 2100 didactic conferences. Ninety percent of my patients want me to address spirituality, yet I have never put the topic “on the table” for my residents. I have no great expertise or insight into this dilemma; I just know it is important. Jerome Groopman recently addressed this dilemma in an editorial entitled “God at the Bedside.”24 One of his dying patients told him she was frightened and said “Doctor, I want you to pray for me.” Groopman, however, had no training for dealing with this request. He eloquently describes the uncertainty of the boundary between professional and personal relationships. Finally, he drew on the customary practice of his teachers and answered a question with a question. He asked his patient: “What is the prayer you want?” “Pray for God to give my doctors wisdom,” the patient said. To that, he silently echoed, “Amen.” I will summarize with a statement that I will boldly call “Thirlby's dictum”: There are no atheists lying on operating room tables. Virtually all of your patients who are about to undergo major surgical procedures are having conversations with a being that is bigger than you or me. If you wish to do so, you can acknowledge this. You can do it with words, most patients welcome it, or you can do it with a smile, a gentle touch, and a few seconds of eye-to-eye contact that communicates clearly the spirituality of the moment. Few careers involve daily situations in which this intense, yet peaceful, exchange can occur. Name another career that evokes such emotions from the recipient and provider of services. 3. you will change patients' lives On the advice of a colleague a few years ago, I started saving thank you letters I receive from patients. They are in a file in my office. All of you have received these. I suggest you save them as well. If you read them again, it will convince you that you are in a great career. It will keep you going. You truly do change patients' lives. Here is a sampling from my file: “Thank you for successful efforts. It's great to have my life back.” “He was sick for a long time, but thanks to all of you, he has been given a chance to live a normal life again.” “Thank you all in helping rid me of a condition which gave me no peace. . . . Love you all for who [you are] and what you’ve done.” “Thanks so very much . . . you have freed me from an outer shell of anguish. God bless you.” “I intended to write this note a while ago to thank you for the wonderful care all of you provided during my recent surgical experience. You made such an impact on my life and I feel as if I have been given a second chance after having gone through this ‘miracle surgery.’ I am so thankful you have dedicated your lives to help improve the lives of others. I will never forget the incredible impressions all of you have made on my life and in my heart.” “I am doing wonderful! I can't even thank you enough. My appetite is crazy . . . too good! My system is finally feeling and doing its own natural things once again.” “Thank you so very much for all your kindness and care. I feel better than I have in years.” “I don't know how to begin to express my thanks and the thanks of my family.” “It is truly life changing. It's awesome.” “After living with UC [ulcerative colitis] for 25 years, for the first time in my memory, I know what ‘health’ feels like.” From another patient with UC, “I count my blessings every day to have had you as my surgeon. I think of my scars as beauty marks that remind me of how much I have to be thankful for. I wear them proudly.” And another patient with UC. “You have totally changed my life. . . . I had an absolute blast at the ninth-grade dance. My friends . . . said I looked gorgeous. And I owe it all to you . . . I will be eternally grateful, you erased the disease that had ravaged my body for 3 years.” “I am so thankful you have dedicated your lives to help improve the lives of others.” Many letters will draw on the importance of spirituality that I discussed above. “I know it was God who placed you in that ER room that Monday.” “God has given you the best gift of all . . . the ability to save lives. Thank you for everything.” “We have faith in you and a deep faith in God and He’ll watch over both of you.” “I thank God every day. . . . May you continue to be blessed with His healing gift.” “I don't know if you are a spiritual person, but I believe you are an angel.” While reading these may seem a bit self-promoting, that is certainly not my intent. All of you have received letters nearly identical to this; all surgeons have. I read these to remind all of you that you are truly changing patients' lives. Save the letters and reread them when you are down on your luck. It will remind you that you have a great career. 2. patients will change your life A patient, J.S., developed testicular cancer when he was a young man. His treatment included high-dose, external-beam radiotherapy to the pelvis. He presented to the Virginia Mason Medical Center a few years ago in his early 40s with a destructive, high-grade sarcoma of the left pubic ramus. The only effective treatment was radical excision. Our orthopedic oncologist, our vascular surgeon, and I performed a radical en bloc excision of the pubis and the hip joint. I reflected structures such as the colon and ureter out of harm's way; the femoral vessels were divided and reconstructed. The pubis and part of the acetabulum were resected en bloc. My orthopedic colleague assured me that ambulation would be possible despite the loss of the essential bony structures of the left hip, but that it would “hurt a little bit.” During J.S.'s preoperative evaluation, he was always remarkably upbeat. He never expressed any blame on the medical community for this “radiation-induced” tumor. He never seemed the least bit scared, bitter, or sad. He never intimated that he had been dealt an unfair circumstance; he only asked what we, and he, needed to do to get on with treatment and healing. He did not change after his operation. His pain management used the typical patient-controlled, narcotic analgesic techniques. On his second postoperative evening, his care team decided that he needed to mobilize to a bedside chair the next day. I entered J.S.'s room that morning on rounds. He was in the middle of his chair transfer. A nurse was in attendance. Every vein in J.S.'s forehead was popping out, sweat was beading on his brow, he was clearly in great pain, but was determined to transfer to the chair as he had been told to do. As I sized up the situation with my residents, I told J.S. to “hold on,” we could lift him to the chair. J.S. stopped his agonizing transfer, somehow managed to smile at me, and said “Don't worry Doc, I don't need any help.” That moment is etched in my brain. I shall never forget it. With tears in my eyes, I thanked him for changing my life. How many times have you treated a patient with a disabling or lethal disease? Despite pain, despite the prospect of mortality or severe disability, despite potential financial ruin, our patients routinely smile; they routinely exhibit a strength of spirit, mind, and body that transcends normal human behavior. Many times we see the anger. Many times we see the fear. Many times we see difficulty dealing with pain. However, when it is all said and done, what we are privileged to see on a daily basis is a strength of spirit that trivializes our daily worries. How can we, in good conscience, complain about anything in our daily life on the same day that a patient with a lethal cancer who is in extraordinary pain says “he doesn't need any help”? When our patients thank us for all our work on their behalf, our response should be “No, J.S., thank you. You did all the work. I did the easy part. Thanks for the privilege of getting to know you.” 1. i love to cut Rolling the right colon in a thin young patient with ulcerative colitis, putting a suture in the stomach, putting a stitch in Cooper's ligament: I could go on and on. A tough operation is completed with intuitive and clear anatomic dissection, minimal blood loss, a perfect anastomosis. You know the patient is going to do well. My vascular surgical colleagues tell me it is completing the perfect carotid endarterectomy in the perfect plane with a perfect feather. The patch looks like it was placed with a precise sewing machine. At the completion of a complex operation, you take your gloves off, take a cleansing breath, thank the operating room staff (and, in ingenuous fashion, the anesthesiologist), and find the family to relay the good news: you tell the family that “all is well, your loved one is going to be just fine.” What satisfaction! What other career provides both the rush of operating and the satisfaction of guiding a smooth perioperative course? I love to cut. Many a resident has heard me say, while rolling the right colon during an ileoanal pouch, “I can't believe somebody is paying me to do this.” With the exception of professional baseball, I cannot think of another job where the pay is so good to do something so fun. The cynics in the audience might respond, “Yes, but what about all of the baggage outside of the operating room?” My response to that is that the prize makes the other stuff worth it. We have a career that affords us the opportunity to be paid to do something that is truly fun. It is why we are all in surgery. Cherish it. In conclusion, members and guests of the WSA, you are all very fortunate indeed (Table). You were born with the mental and physical attributes necessary to excel as high school, college, and medical students. You were blessed to be provided with the opportunity to utilize your talents. At some time in your education, you were injected with the addicting drug called “surgery.” You had and/or developed panache. You have become a very successful surgeon, which affords you job security, financial comfort, great respect in your community, and daily personal interactions that are profound and fulfilling. Best of all, when you go to work, you get to cut. Don't ever complain; you have the greatest job in the world. Table. View LargeDownload Top 10 Reasons Why General Surgery Is a Great Career Back to top Article Information Correspondence: Richard C. Thirlby, MD, Section of General, Thoracic, and Vascular Surgery, Virginia Mason Medical Center, 1100 Ninth Ave, PO Box 900 (C6-SUR), Seattle, WA 98111-0900 ([email protected]). Accepted for Publication: December 5, 2006. Financial Disclosure: None reported. Previous Presentation: This paper was presented as the Presidential Address at the meeting of the Western Surgical Association; November 13, 2006; Los Cabos, Mexico; and is published after peer review and revision. Acknowledgment: The author gratefully acknowledges Cynthia Kirtland, chaplain at the Virginia Mason Medical Center; Susan Long, MLS, at the Virginia Mason Medical Center Medical Library; and Elisabeth Davis, MA, of the American College of Surgeons for invaluable support with literature search and review. References 1. Evans SRT A year like no other [editorial]. Arch Surg 2005;140592Google ScholarCrossref 2. Powell ACMcAneny DHirsch EF Trends in general surgery workforce data. Am J Surg 2004;1881- 8PubMedGoogle ScholarCrossref 3. Etzioni DALiu JHMaggard MAKo CY The aging population and its impact on the surgery workforce. Ann Surg 2003;238170- 177PubMedGoogle Scholar 4. Liu JHEtzioni DAO’Connell JBMaggard MAKo CY The increasing workload of general surgery. Arch Surg 2004;139423- 428PubMedGoogle ScholarCrossref 5. Stitzenberg KBSheldon GF Progressive specialization within general surgery: adding to the complexity of workforce planning. J Am Coll Surg 2005;201925- 932PubMedGoogle ScholarCrossref 6. Jonasson OKwakwa F Retirement age and the work force in general surgery. Ann Surg 1996;224574- 582PubMedGoogle ScholarCrossref 7. AAMC Statement on the Physician Workforce, June 2006. Association of American Medical Colleges Web site. http://www.aamc.org/workforce/workforceposition.pdf. Accessed October 2, 2006Google Scholar 8. US Department of Labor, Bureau of Labor Statistics, Occupational Outlook Handbook: Physicians and Surgeons. 2006;http://www.bls.gov/oco/ocos074.htm. Accessed August 2, 2006 9. Fronczek P American Community Survey Reports: Income Earnings, and Poverty From the 2004 American Community Survey. Washington, DC Housing and Household Economic Statistics (HHES) Division, US Census Bureau2005;1- 20 10. US Department of Labor, Bureau of Labor Statistics, Occupational employment and wages: 29-1067 surgeons. May2005;http://stats.bls.gov/oes/current/oes291067.htm. Accessed August 2, 2006 11. Buckley JE Research summary: rankings of full-time occupations by earnings, 2000. Mon Labor Rev 2002;12546- 57Google Scholar 12. National Opinion Research Center, General Social Survey. Chicago, Ill National Opinion Research Center1991; 13. Katz P The Scalpel's Edge: The Culture of Surgeons. Needham Heights, Mass Allyn & Bacon1999; 14. McGregor DB Chester Bidwell McVay, MD: the practicing anatomist. Arch Surg 2002;137226PubMedGoogle ScholarCrossref 15. Organ CH Jr Lester is coming [Surgical Reminiscence]. Arch Surg 2001;1361087PubMedGoogle ScholarCrossref 16. Maugans TAWadland WC Religion and family medicine: a survey of physicians and patients. J Fam Pract 1991;32210- 213PubMedGoogle Scholar 17. Hinshaw DB Spiritual issues in surgical palliative care. Surg Clin North Am 2005;85257- 272PubMedGoogle ScholarCrossref 18. Koenig HG Religion, spirituality, and medicine: research findings and implications for clinical practice. South Med J 2004;971194- 1200PubMedGoogle ScholarCrossref 19. King DEBushwick B Beliefs and attitudes of hospital inpatients about faith healing and prayer. J Fam Pract 1994;39349- 352PubMedGoogle Scholar 20. Ehman JWOtt BBShort THCiampa RCHansen-Flaschen J Do patients want physicians to inquire about their spiritual or religious beliefs if they become gravely ill? Arch Intern Med 1999;1591803- 1806PubMedGoogle ScholarCrossref 21. Tarpley JLTarpley MJ Spirituality in surgical practice. J Am Coll Surg 2002;194642- 647PubMedGoogle ScholarCrossref 22. Sulmasy DP Spiritual issues in the care of dying patients: “ . . . it's okay between me and God.” JAMA 2006;2961385- 1392PubMedGoogle ScholarCrossref 23. Puchalski C Spirituality in health: the role of spirituality in critical care. Crit Care Clin 2004;20487- 504, xPubMedGoogle ScholarCrossref 24. Groopman J God at the bedside. N Engl J Med 2004;3501176- 1178PubMedGoogle ScholarCrossref

Martin, Robert C. G.;Loehle, Jennifer;Scoggins, Charles R.;McMasters, Kelly M.

doi: 10.1001/archsurg.142.5.431pmid: 17515484

Abstract Background Hepatitis and cirrhosis are common etiologic factors in hepatocellular carcinoma (HCC) in the United States. However, noncirrhotic, nonfibrotic HCC has been recognized more frequently in Kentucky. The aim of this study was to evaluate the epidemiologic features of this variant of HCC. Hypothesis Kentucky hepatoma, defined as a noncirrhotic, nonfibrotic, hepatitis-negative HCC, occurs in an older population with more favorable preoperative factors when compared with other patients with HCC. Design A prospective review of our 1002 hepatopancreaticobiliary patients, the Kentucky Cancer Registry, and the Surveillance, Epidemiology, and End Results database. Setting An academic referral center. Patients All patients with HCC treated from January 1, 1999, through September 30, 2005, were reviewed for clinicopathologic factors, recurrence, and outcome. Results In a review of the region's 703 patients with HCC, we have seen a 4-fold increase in age-specific HCC diagnosis, with the most rapid increase seen in the 60- to 69-year-old age group. In our institution's 103 patients with HCC, 62 (60.2%) were without hepatitis or cirrhosis. These noncirrhotic, hepatitis-free patients were found to be significantly older (70 vs 55 years; P = .001), to be more often female (40.3% vs 24.4%; P = .01), to have a larger tumor size (6.5 vs 3.9 cm; P = .004), to have fewer liver lesions (1 vs 3; P = .22), and to more frequently undergo surgical therapy (75.6% vs 53.8%; P = .01) than the patients with cirrhosis or hepatitis (n = 41). Conclusions A larger percentage of the patients with HCC seen in our region are significantly different from those in other reports throughout the United States in preoperative clinical and pathologic presentation. The reason for this change is as yet unknown, but the incidence continues to rise annually. Hepatocellular carcinoma (HCC) is one of the top 5 most common cancers in the world and remains one of the top 3 most common causes of cancer mortality, with an incidence reported in excess of 560 000 cases.1 Most HCC cases (>80%) occur in either sub-Saharan Africa or East Asia. Aggressive attempts at prevention and public awareness have demonstrated encouraging trends in these high-incidence areas, with decreases in incidence having been reported in China, Hong Kong, and Singapore.2 In contrast and more alarmingly, HCC incidence in a number of low–incidence rate areas, including Australia, the United States, Canada, and the United Kingdom, has increased significantly and has not been easily explained by changes in immigration alone. It is widely know that chronic hepatitis B and C virus infection remains the most dominant risk factor in HCC incidence (Figure 1). Other known hepatocellular carcinoma risk factors include alcoholic cirrhosis, hemochromatosis, and the consumption of aflatoxin B1–contaminated foodstuffs. The universal belief has always been that cirrhosis precedes the development of HCC (Figure 1), even in the United States, where continued reports state that approximately 95% of HCC arises in the background of cirrhosis.3 Another challenge has been to determine the cause of HCC in patients without chronic hepatitis infection or without chronic cirrhosis but who still develop primary HCC (Figure 1). This etiologic question may be one of the reasons why there has been a 2-fold increase in the age-adjusted incidence rate of HCC in the United States between 1985 and 2002.4 This incidence has caused an increase of 1.3 per 100 000 population during 1978 to 1980 to 3.3 per 100 000 population during 1999 to 2001.4 Interestingly, the largest increase during this time has occurred in white individuals (including Hispanic persons), whereas the lowest increase has been within the Asian population. More recent reports have demonstrated that as much as 40% of the HCC diagnosed in the United States is of unknown cause (ie, not by alcohol, hepatitis, or cirrhosis).5 The origin of HCC in patients without chronic hepatitis infection or chronic cirrhosis from other causes remains unclear. To date, because of the overwhelming acceptance of hepatitis or cirrhosis as the cause of HCC, little attention has been paid to the increasing incidence of patients who present with nonhepatitic, noncirrhotic HCC. Because we have witnessed this etiologic variant of HCC so commonly in our practice during the past decade, we have termed it the Kentucky hepatoma. Thus, the aim of our study was to evaluate the incidence of this variant HCC. Methods A prospective review of our center's 1000 hepatopancreaticobiliary patients, the Kentucky Cancer Registry, and the Surveillance, Epidemiology, and End Results (SEER) database was conducted for both incident rates and overall survival in patients with HCC. The Kentucky Cancer Registry was searched from January 1, 1999, to September 30, 2005, to look at age-adjusted rates for HCC in patients 20 years and older.6 In addition, incidence rates per 100 000 persons were also searched by county to evaluate epidemiologic trends in certain areas of the state.6 The Kentucky Cancer Registry does not delineate HCC based on hepatitis, cirrhosis, or other etiologic factors. The evaluation of the Kentucky Cancer Registry was to evaluate trends in incidence rates and trends among the 65 counties in a predominantly rural state. De-identified patient-level data were obtained from the population-based SEER database maintained by the National Cancer Institute. Evaluation of the database included the 17 SEER areas: San Francisco, Calif; Connecticut; Detroit, Mich; Hawaii; Iowa; New Mexico; Seattle, Wash; Utah; Atlanta, Ga; San Jose–Monterey, Calif; Los Angeles, Calif; Alaska; rural Georgia; California (excluding the cities already mentioned); Kentucky; Louisiana; and New Jersey. The population captured by the SEER database is representative of the US population.7,8 To ensure the accuracy of the data, medical record abstractors undergo extensive training as described in the National Cancer Institute SEER publications. Malignancies are encoded based on the International Classification of Diseases, Ninth Revision (ICD-9)9 for oncology. The SEER database does not distinguish between patients with HCC who are diagnosed as having hepatitis B or C and those with underlying cirrhosis. The reason for the use of the SEER database is to compare the 5-year survival of patients in our region with that of others. All patients with HCC treated in the Department of Surgery at the University of Louisville from January 1, 1999, through September 30, 2005, were identified from the Division of Surgical Oncology's prospective hepatopancreaticobiliary database. At the University of Louisville, the care of HCC is planned under the auspices of a multidisciplinary hepatopancreaticobiliary disease management team that includes Society of Surgical Oncology–accredited and trained surgical oncologists with a specialty in hepatopancreaticobiliary malignancies, medical oncologists, gastroenterologists, interventional radiologists, transplant surgeons, and radiation oncologists. One hundred three patients with HCC were considered for treatment in this period. Data for these patients were then extracted from the database for review. Data examined included extensive demographics (age, sex, medical history, surgical history, social history, tobacco history, alcohol history, medication use, family history, and any other possible etiologic factors), pathologic findings of the liver lesion, pathologic or nonneoplastic liver parenchyma, hepatitis serologic test results, α-fetaprotein level, and outcome. Follow-up was obtained by the treating physician and is up to date from the end of the study. Nomenclature for the extent of resection was defined by Goldsmith and Woodburne.10 Evaluation of liver function was by the Childs-Pugh classification.11,12 Clinical staging of disease was performed by using the 2002 American Joint Committee on Cancer staging criteria and the Okuda staging criteria.13 Additional criteria in the form of extent of liver involvement were also used based on review of all cross-sectional imaging for each patient by 2 independent radiologists for extent of disease. The diagnosis of hepatitis was defined as the presence of hepatitis C antibody or the presence of hepatitis B surface antigen. Alcoholic cirrhosis was defined as excessive alcohol use in combination with either radiologic demonstration or pathologic demonstration of hepatic cirrhosis. A patient with Kentucky hepatoma was defined as an individual who was without hepatitis, had no history of excessive alcohol use, and had normal nonneoplastic hepatic parenchyma for patients who underwent resection or ablation, or normal hepatic parenchyma on computed tomography or magnetic resonance imaging for patients treated without resection or ablation. The χ2, t, and Mann-Whitney U tests for nominal, continuous, and ordinal variables, respectively, were used to evaluate the association of independent variables with surgical complications. Proportional hazards analysis was performed using all variables found to be significant by univariate analysis. Relative risks with 95% confidence intervals were calculated as a measure of association. Multiple logistic regression analysis was used to determine overall outcome and differences. P<.05 was considered statistically significant. Statistical analysis was performed using JMP software (SAS Institute Inc, Cary, NC). Results A review of the Kentucky Cancer Registry has demonstrated a near 4-fold increase in age-specific HCC rates in Kentucky, with the most rapid rise seen in the 60- to 69-year-old age group (Figure 2). In addition, a consistently increasing incidence of age-specific HCC has been seen in Kentucky in patients 70 years or older compared with other age groups. The incidence of HCC in Kentucky is evenly distributed among the north central counties, the eastern counties, and the western counties (Figure 3), without any 1 area as the dominant hotbed of incidence. In a review of our 103 patients with HCC at the University of Louisville, 62 (60.2%) were found to be without hepatitis or cirrhosis compared with 41 patients (39.8%) who had either underlying hepatitis infection or other etiologically induced cirrhosis. In a comparison of these 2 groups, patients who were without hepatitis or cirrhosis were found to be significantly older (70 vs 55 years; P = .001) (Table 1). Furthermore, in the patients without hepatitis or cirrhosis, a significantly greater proportion of women (40.3% vs 24.4%; P = .01) and a trend toward an increase in diabetes mellitus (37.1% vs 19.5%; P = .05) were seen when compared with patients with cirrhosis. Patients without underlying hepatitis or cirrhosis were less often smokers (40.3% vs 70.7%; P = .002) and had significantly lower α-fetoprotein levels at initial diagnosis (median level, 16 vs 320 ng/mL; P = .02). Patients who were found to be free of hepatitis and cirrhosis had a significantly greater tumor size (6.5 vs 3.9 cm; P = .02) and fewer tumors (median, 1) when compared with the patients with hepatitis or cirrhosis (median, 3). This translated into a trend toward a greater percentage of the patients with Kentucky hepatoma (57.6%) having less than 25% liver parenchymal involvement when compared with the cirrhotic patients (48.4%), which was not statistically significant (P = .58). These preoperative factors led to a greater use of surgical therapy in the patients with Kentucky hepatoma. The patients without hepatitis or cirrhosis underwent major liver resection more frequently (34.4% vs 2.0%; P<.001), with a similar fraction of patients who underwent minor resection (5.0% vs 4.7%) and a greater number of liver radiofrequency ablation procedures (27.3% vs 12.1%). Conversely, the hepatitis and cirrhosis population had a greater incidence of transplantation (39.3% vs 10.9%) when compared with the patients with Kentucky hepatoma. A similar T stage was seen in both groups, with a nonsignificant trend toward greater T1 stage in the patients with Kentucky hepatoma (50.0% vs 35.7%). This similar T stage also translated into similar Okuda stage and Childs-Pugh status (Table 1). In patients with Kentucky hepatoma, a trend was seen toward a greater incidence of vascular invasion, but this finding was not statistically significant (P = .06). In a review of the SEER database in regard to incidence, there has been a consistent age-adjusted incidence of more than 20 per 100 000 population for those 65 years and older for both men and women of all races. The 65 years and older aggregate incidence is 26.5 per 100 000 population. This increase has led to the third largest increase in the trend of SEER incidence rates by primary cancer site from 1994 to 2003, 1.6 per 100 000 population, with only thyroid (5.5) and melanoma (3.1) being greater. When we then compared our institutional data with the 17 SEER supporting data sites, there was a greater incidence of localized stage at diagnosis in the SEER database (33%) when compared with the University of Louisville patients (9.4%). A similar distribution was found of regional (25% vs 26%) and distant (21% vs 38%) disease in the SEER data areas compared with our institutional data. Interestingly, the 5-year survival rate by stage was significantly improved in the University of Louisville patients for localized disease (53.3% vs 21.9%; P = .001) and regional disease (26.3% vs 7.2%; P = .001), with similar survival for distant disease (5.4% vs 4.1%; P = .90) (Table 2). Comment The incidence of HCC has risen markedly in the United States during the past 2 decades, with an average yearly age-adjusted incidence rate of HCC (verified by histologic or cytologic analysis) from 1.3 per 100 000 population during 1978 to 1980 to 3.3 per 100 000 population during 1999 to 2001.4 This increase has been widely hypothesized to be related to the dormancy or latency of hepatitis B and C in patients infected in the 1960s and 1970s who are now developing underlying cirrhosis and HCC.4 The incidence of nonhepatitic, noncirrhotic HCC in the United States has not been evaluated. This report represents the first evaluation at a tertiary referral center of the incidence of this variant HCC (ie, nonhepatitic and noncirrhotic), which we have labeled Kentucky hepatoma. It is possible that a national increase in the incidence of HCC reflects similar increases in HCC among patients without hepatitis or cirrhosis in other regions.5 One reason for the increased incidence of HCC may be related to changes in the diagnostic criteria used in the SEER database. In the 1970s and early 1980s, the SEER database recorded only histologically confirmed cases of HCC. Although this is highly specific, it may underestimate the true number of patients with HCC. The average yearly age-adjusted incidence rates of all HCC captured by the SEER database irrespective of the method of diagnosis has increased approximately 30%. Because the diagnosis of HCC can now occur without histologic confirmation, based on the presence of underlying cirrhosis, a nodular mass seen on computed tomography or magnetic resonance imaging, and an elevated α-fetoprotein level, the corresponding increased incidence in the SEER database may be partially related to these changes in diagnostic criteria. Similarly, with the rapid rise in hepatic resections now being performed in all age groups, patients who were initially diagnosed as having and being treated for cancer of unknown primary are now being appropriately diagnosed as having HCC. Perhaps the incidence of HCC is not actually increasing, but the incidence of carcinoma of unknown primary is decreasing as a result of our ability to appropriately diagnose these patients as having either HCC or intrahepatic cholangiocarcinoma. The willingness to be surgically aggressive in patients who are initially diagnosed as having unknown primary cancers has led to more precise and accurate diagnosis and staging in patients with these primary malignancies. This may partially explain the relatively large proportion of patients at the University of Louisville diagnosed as having HCC who do not have underlying hepatitis and are free of any type of hepatic parenchymal abnormalities. Other potential causes for the increasing incidence of this variant HCC may be the ever increasing age of the US population. Elderly people (defined as those ≥65 years) will account for more than 61% of all new cancer cases and 70% of all annual cancer deaths.6 It has recently been estimated that the elderly patient population has 11 times the cancer risk of people younger than 65 years. It also has been estimated that in 2030, approximately 20% of the US population will be older than 65 years.7 These changes alone (age of the liver) coupled with environmental exposure and potential genetic effects may be other potential causes for this change in incidence rates. This reasoning is further solidified by the significant differences in the clinical and pathologic features of the patients with Kentucky hepatoma when compared with the common cirrhotic patient with HCC. Those with Kentucky hepatoma are significantly older, are more commonly female, are less often smokers, have a lower α-fetoprotein level, have a greater incidence of a normal α-fetoprotein level, have larger tumors, and have a smaller number of tumors (most commonly a single liver tumor). These criteria have led to an ability to be more surgically aggressive in these patients because of their underlying normal hepatic parenchyma and minimal comorbidities. This variant histologic presentation responds well to aggressive surgical resection. The variant HCC or Kentucky hepatoma has also shown an alarming rate of increase in the older patient (60-69 years of age) and remains high in the 70- to 79-year age group (Figure 1). In an attempt to evaluate specific regions of Kentucky that would elucidate this increase, a similar increase in HCC has been seen among all 5 major regions in Kentucky (Figure 2). Compared with other studies that have presented the incidence of this variant HCC, our study demonstrates the single largest incidence rate. A recent report from Reichman et al14 demonstrated an incidence of 26%, with a worse overall survival when compared with the cirrhotic population. A similar high incidence (31%) was reported by Fong et al,15 without a difference in overall survival. Both of these reports were from the United States, in New Jersey and New York, respectively, which has a higher incidence of Asian immigrants and thus a probable higher incidence of hepatitis. A similar low (<15%) incidence of this variant HCC is seen in Europe and Asia.16-18 Consistently, however, all of these studies, as well as ours, did not demonstrate a significant survival effect in patients with HCC. The underlying hepatic parenchyma did not affect the biologic features of this disease in our study or in previous studies evaluated. The fact that 60.2% of the patients evaluated at the Department of Surgery of the University of Louisville are without evidence of hepatitis or cirrhosis is markedly different from the experience of other centers. This finding certainly may reflect referral bias to our center and the fact that our surrounding area does not include a large population of patients with endemic hepatitis B and C. Because only 2 centers in Kentucky perform major hepatectomy, we reason that the referral of patients with HCC (with and without hepatitis or cirrhosis) is probably evenly distributed. Admittedly, our study does not capture all of the patients diagnosed with this condition in our region but only those referred for surgical consultation. However, because our center uses a multidisciplinary team approach to the evaluation and treatment of these patients and because the surgeons in our department treat all of the patients undergoing nonoperative liver-directed therapies (eg, embolization, chemoembolization, radioactive microsphere therapy, and percutaneous radiofrequency ablation), we believe that the results presented herein are probably an accurate reflection of the patients treated for HCC in our area. In addition, in other areas where multiple centers provide liver-directed therapy, it may be more common to refer patients with cirrhosis to larger institutions that have transplantation as an option and to treat older patients without cirrhosis in the smaller centers. The increased incidence of HCC in noncirrhotic, nonhepatitic patients with HCC may be common to other regions as well but may not be as well recognized in centers where there is a greater population at risk for hepatitis B and C. Whether our high proportion of patients with HCC who are without hepatitis or cirrhosis represents differences in diagnosis, surgical treatment, or as-yet-undiscovered etiologic factors remains to be elucidated. Other environmental exposures and genetic factors are being evaluated to delineate this epidemiologic variant. Correspondence: Robert C. G. Martin II, MD, University of Louisville School of Medicine, 315 E Broadway, Room 313, Louisville, KY 40202 ([email protected]). Accepted for Publication: December 13, 2006. Author Contributions:Study concept and design: Martin. Acquisition of data: Martin, Loehle, Scoggins, and McMasters. Analysis and interpretation of data: Martin. Drafting of the manuscript: Martin. Critical revision of the manuscript for important intellectual content: Martin, Loehle, Scoggins, and McMasters. Statistical analysis: Martin. Administrative, technical, and material support: Martin, Loehle, and McMasters. Study supervision: Scoggins. Financial Disclosure: None reported. Previous Presentation: This study was presented at the 114th Annual Meeting of the Western Surgical Association; November 14, 2006; Los Cabos, Mexico. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript. References 1. Greenlee RTMurray TBolden S et al. Cancer statistics, 2000. CA Cancer J Clin 2000;507- 33PubMedGoogle ScholarCrossref 2. McGlynn KATsao LHsing AW et al. International trends and patterns of primary liver cancer. Int J Cancer 2001;94290- 296PubMedGoogle ScholarCrossref 3. Botha JFLangnas AN Liver transplantation for hepatocellular carcinoma: an update. J Natl Compr Canc Netw 2006;4762- 767PubMedGoogle Scholar 4. El-Serag HBDavila JAPetersen NJ et al. The continuing increase in the incidence of hepatocellular carcinoma in the United States: an update. Ann Intern Med 2003;139817- 823PubMedGoogle ScholarCrossref 5. Davila JAMorgan ROShaib Y et al. Hepatitis C infection and the increasing incidence of hepatocellular carcinoma: a population-based study. Gastroenterology 2004;1271372- 1380PubMedGoogle ScholarCrossref 6. Kentucky Cancer Registry. http://www.kcr.uky.edu/. Accessed October 15, 2006 7. Surveillance, Epidemiology, and End Results. Bethesda, Md National Cancer Institute, Cancer Surveillance Research Program, Cancer Statistics Branch2006; 8. Liu JHChen PWAsch SM et al. Surgery for hepatocellular carcinoma: does it improve survival? Ann Surg Oncol 2004;11298- 303PubMedGoogle ScholarCrossref 9. World Health Organization, International Classification of Diseases, Ninth Revision (ICD-9). Geneva, Switzerland World Health Organization1977; 10. Goldsmith NAWoodburne RT The surgical anatomy pertaining to liver resection. Surg Gynecol Obstet 1957;105310- 318PubMedGoogle Scholar 11. Wantz GEPayne MA Experience with portacaval shunt for portal hypertension. N Engl J Med 1961;265721- 728PubMedGoogle ScholarCrossref 12. Pugh RNMurray-Lyon IMDawson JL et al. Transection of the oesophagus for bleeding oesophageal varices. Br J Surg 1973;60646- 649PubMedGoogle ScholarCrossref 13. Okuda KOhtsuki TObata H et al. Natural history of hepatocellular carcinoma and prognosis in relation to treatment: study of 850 patients. Cancer 1985;56918- 928PubMedGoogle ScholarCrossref 14. Reichman TWBahramipour PBarone A et al. Hepatitis status, Child-Pugh classification, and serum AFP levels predict survival in patients treated with transarterial embolization for unresectable hepatocellular carcinoma. J Gastrointest Surg 2005;9638- 645PubMedGoogle ScholarCrossref 15. Fong YSun RLJarnagin W et al. An analysis of 412 cases of hepatocellular carcinoma at a Western center. Ann Surg 1999;229790- 799PubMedGoogle ScholarCrossref 16. Figueras JRamos EIbanez L et al. Surgical treatment of hepatocellular carcinoma in cirrhotic and noncirrhotic patients. Transplant Proc 1999;312455- 2456PubMedGoogle ScholarCrossref 17. Torzilli GMakuuchi MInoue K et al. No-mortality liver resection for hepatocellular carcinoma in cirrhotic and noncirrhotic patients: is there a way? a prospective analysis of our approach. Arch Surg 1999;134984- 992PubMedGoogle ScholarCrossref 18. Tsukioka GKakizaki SSohara N et al. Hepatocellular carcinoma in extremely elderly patients: an analysis of clinical characteristics, prognosis and patient survival. World J Gastroenterol 2006;1248- 53PubMedGoogle Scholar Steven C. Stain, MD, Albany, NY: The authors from Louisville make the argument that HCC in Kentucky somehow differs from the rest of the country because of differences in clinical and pathologic presentation. This conclusion comes from analysis of data from 3 sources: a database maintained by the Division of Surgical Oncology, the Kentucky Cancer Registry, and the SEER database. There were 103 patients with HCC in the Louisville database and 703 patients in the Kentucky Cancer Registry. The authors contend that compared with a nationwide SEER 17 sample, the patients seen at the University of Louisville were more likely to have localized disease and that those with localized disease had significantly longer survival than patients in the SEER database. The authors have coined the term Kentucky hepatoma to describe a variant HCC that occurs in patients without hepatitis or cirrhosis. The Kentucky Cancer Registry became a SEER site in 2001, and the increased incidence of HCC identified in this registry appears to mirror national trends. Although it has been suggested that the national increase is likely due to hepatitis B and C infections secondary to transfusions received more than 30 years ago, today on the SEER Web site, I compared the rising incidence of liver and intrahepatic bile duct cancer in Kentucky with the nationwide sample, and increases seemed quite similar. To the authors, have you done statistical analyses to determine if the increasing incidence in Kentucky is greater than the trend observed in the nationwide SEER data? In the Louisville database, 60% of the patients did not have hepatitis or cirrhosis. I am unclear what the histological requirements for Kentucky hepatoma are. In 1956, Hugh Edmondson, MD, from the University of Southern California first described fibrolamellar HCC as a distinct, well-differentiated cancer with specific histologic characteristics that occurred in normal livers and most often in younger patients. Are there specific histologic features of Kentucky hepatoma that pathologists can identify, or are the authors referring to any HCC in the absence of cirrhosis or hepatitis serology? It would be useful to have some more information about the data available in the Louisville database. Did all patients with HCC in this database have serologic testing for hepatitis B or C? How did you determine the absence of cirrhosis? Was biopsy of normal liver required for patients who did not have resection? The authors compare the stage distribution and survival rate of patients in the Louisville database with the SEER data. While I am confident that the Louisville database included all patients seen by surgeons, did this database include all patients with HCC seen by all physicians within the region or even within the University of Louisville? I have concerns about comparing a surgical database with a population-based sample such as the SEER data set for both stage distribution and survival analysis. Surgeons might be much more likely to evaluate the patients who referring physicians believed were candidates for therapy. The authors point out that the Kentucky Cancer Registry does not contain data regarding hepatitis serology or cirrhosis histology. Do the authors believe that the increase in patients with HCC in Kentucky is primarily due to the increased prevalence of hepatitis that is seen nationally, or does Kentucky have an increase in the number of noncirrhotic hepatocellular carcinomas in the absence of hepatitis? Could this also be simply a change in referral patterns to the University of Louisville? I do not know if any of these are true, but I do not believe that this question can be answered by analysis of a single institutional database. Dr Martin: First, do we believe that our increasing trend in hepatocellular cancers is higher than the SEER database? The limitations of my comment are that I do not know why. The reason for this limitation is the difficulty with the SEER database and how much of the incidence is influenced by cholangiocarcinoma. One of the limitations with the SEER database, and hopefully one that can be easily corrected in the future, is that it includes in the data both HCC and cholangiocarcinoma as primary liver cancers. This limits the ability of the SEER database to evaluate trends in both of these diseases. The histologic characteristics of the HCC of these patients were all confirmed pathologically. None of these patients had the fibrolamellar variant, which is seen in the younger population and which is different from our data, since a majority of our patients with HCC were older. All patients did undergo serologic testing both for hepatitis as well as for α-fetoprotein levels. The hepatic parenchyma for the nonsurgically treated patients were evaluated by percutaneous biopsy to establish the diagnosis or were defined based on their lack of serologic, social, and radiologic characteristics for cirrhosis. We have established an inclusive referral pattern based on our multidisciplinary conferences, and since our division has established all of the new liver-directed therapies, this allows us to be confident that this is a true denominator. In addition, since there is really only 1 other academic center, and that is Lexington, which draws more toward the east of the state, we have been fortunate to be able to see a lot of these patients at the initial management. So I think our denominator is fairly robust. I think we probably do lose some patients south to Nashville as well as north to Cincinnati, but I would also argue that both the cirrhotic and noncirrhotic patients are probably being referred to those other centers because they may actually be closer geographically. The increasing incidence of HCC: yes, I think there is obviously the latency period of hepatitis B and C, but I do believe that this rising incidence of this variant HCC is real. The main focus of this manuscript is hopefully awareness, greater acceptance of it, and, more important, greater accuracy in treating these patients, since, as I have said before, systemic chemotherapy is really not beneficial for them. John J. Brems, MD, Burr Ridge, Ill: This is a phenomenon that I think others are seeing, too. One of my thoughts is, since you are seeing it in older patients and since this group seems to have a higher proportion of female patients, as opposed to your cirrhotic patients, do you think these could be hepatic adenomas that are undergoing malignant transformation? Dr Martin: Without a doubt, I think it is a possibility. The other is simply a new genetic distribution because of the aging population. Are we seeing that type of dedifferentiation or loss of that consistent regeneration as patients are getting older? It is a possibility. Obviously it remains to be seen. William C. Chapman, MD, St Louis, Mo: We live only a few hours west of you, and I will tell you that we do not see this phenomenon. Eighty percent of our patients who have HCC have associated liver disease. I think that is more in keeping, as you point out, with what has been published. I would be very skeptical, based on your single-center experience, to conclude that 60% of your patients with HCC are unassociated with liver disease. This point was mentioned earlier by Dr Stain: I would be pretty skeptical on the basis of 1 report. I think your point of saying to consider that your patient may be noncirrhotic and may be able to undergo standard resection is an important one, but I think you need a lot more data before you can jump to the conclusion that the epidemiology is changing. We are seeing a big increase in HCC, as are others. In our experience it is in large part related to increased incidence of nonalcoholic steatohepatitis and hepatitis C. A question for you: Who are you considering for liver transplantation? You mentioned that in your group you had transplant surgeons and, I guess, hepatologists involved. How do you structure that? You also have a pretty high recurrence rate. Are you considering transplantation in any of your noncirrhotic patients? You pointed out that 11%, I think, had had transplantations. Could you comment a little on how transplantation is factored in in your center? Dr Martin: I understand your skepticism, but we also mentioned that there must be a referral bias to a transplant center as well known as yours (ie, only cirrhotic patients). Obviously, the rate of 60% is probably higher than seen in other centers, but I think obviously the national trends have shown that the incidence rate is far higher than the 5%, or even the 10%, that continues to be quoted. I think the main issue is to keep it in mind, keep it aware. We have had some fairly stringent evaluation of our hepatic parenchyma, so we are very confident of our diagnosis. In regard to whom we perform transplantations on, transplantation, again, falls under the Milan criteria. I think there are few data to suggest that transplantation should be the first treatment option in a noncirrhotic patient with resectable disease. Some of our patients do undergo, as you reported in the Journal of American College of Surgeons, pretransplant therapy to better assess the biology of disease and to evaluate if transplantation would be indicated. The majority of these patients, though, do present with lesions greater than 5 cm as presented in the data, so they obviously fall outside the established Milan criteria. Stephen G. Jolley, MD, Anchorage, Alaska: In your nonhepatitic, noncirrhotic patients with hepatoma, are any of them related by blood? Dr Martin: A great question. I cut out a slide for the interest of time. Family history was not a factor in our incidence, and we take a thorough family history as part of the database just to begin to sort out those trends. Financial Disclosure: Dr Chapman is a founder of Pathfinder Therapeutics, Inc.

Bilimoria, Karl Y.;Cambic, Angela;Hansen, Nora M.;Bethke, Kevin P.

doi: 10.1001/archsurg.142.5.441pmid: 17515485

Abstract Hypothesis Women with newly diagnosed breast cancers may harbor additional ipsilateral or contralateral breast malignancies that are undetected by mammography and ultrasonography. Magnetic resonance imaging (MRI) has demonstrated excellent sensitivity in the detection of breast cancers. However, the impact of routine MRI on the surgical management of new, biopsy-proven breast cancers remains unclear. Design Retrospective analysis of a prospective database. Setting An academic, tertiary care center in a large metropolitan area. Patients A total of 155 women with breast cancer newly diagnosed by mammography, ultrasonography, and needle biopsy underwent preoperative bilateral breast MRI in a single-institution, single-surgeon setting during 1 year. Main Outcome Measures Change in surgical management based on breast MRI findings. Results The MRI demonstrated 124 additional suspicious lesions in 73 patients. Post-MRI follow-up mammograms or ultrasonograms were required in 65 patients, and 41 patients underwent additional image-guided biopsies. There was a change in surgical management as MRI discovered additional, otherwise undetected malignancies in 36 patients based on radiographic-pathologic correlation. Lumpectomy was converted to mastectomy in 10 patients (8 beneficial), wider excision was performed in 21 patients (10 beneficial), and 5 patients (2 beneficial) underwent contralateral surgery. Larger tumor size was an independent predictor of a beneficial change in surgical management (odds ratio, 1.66; 95% confidence interval, 1.04-2.66). Conclusions Breast MRI results in a beneficial change in surgical management in 9.7% of newly diagnosed breast cancers. The detection of additional, otherwise undetected ipsilateral and contralateral breast malignancies with MRI suggests that breast MRI may have a role in the evaluation of new breast cancers. Women with newly diagnosed breast cancer are at risk of harboring an occult, synchronous ipsilateral or contralateral breast cancer that is undetected by mammography or ultrasonography. The rate of multifocality and multicentricity varies widely from 11% to 57%.1-4 If these additional foci can be identified preoperatively, the planned surgical management can be altered. Unfortunately, mammograms and ultrasonograms are not sensitive enough to detect some of these synchronous lesions. Recently, breast magnetic resonance imaging (MRI) has been investigated as a screening modality and found to have a sensitivity of 93% to 100% in detecting breast cancers.5 Studies6,7 have shown that MRI in women with newly diagnosed breast cancers identifies additional, otherwise undetected synchronous tumor foci in 27% to 37% of patients. Thus, MRI has the potential to detect synchronous cancers, multifocality and multicentricity of the primary neoplasm, tumor extent, and early lesions that would otherwise develop into future cancers.6,7 Local recurrence rates after breast conservation therapy (BCT) range from 4.3% to 10%.8-10 These unrecognized lesions lead to worse local control of the cancer, lead to additional procedures and operations, and may result in worse outcomes.11-15 The purpose of this study was to determine the effect of routine breast MRI on the surgical management of newly diagnosed breast cancers. First, our objective was to determine how often MRI detected an additional lesion that was otherwise undetected by mammography and ultrasonography. Second, we sought to examine the change in preoperative evaluation as a result of the MRI findings. Third, we aimed to determine whether the change in surgical management was beneficial based on radiographic-pathologic correlation. Finally, we attempted to identify predictors of a beneficial change in surgical management based on preoperative breast MRI. Methods A standardized protocol was implemented by a single surgeon (K.P.B.) in the management of all new, biopsy-proven breast cancers starting in April 2005. The study includes patients identified from a prospective database from April 1, 2005, to April 1, 2006. Approval for this study was obtained from the Northwestern University Feinberg School of Medicine's institutional review board. The study included women 34 to 75 years of age with a new primary breast cancer. Women were excluded if MRI was deemed unnecessary: predominantly fatty breast tissue in cases where mammography was thought to be sufficient, women older than 75 years, insurance refusal of MRI precertification, claustrophobia, pregnancy or planned bilateral mastectomy (MRI would not have changed surgical management). We also excluded women presenting with a history of breast cancer and patients who had MRI performed at an outside institution. Imaging Patients who presented with a possible breast cancer underwent exhaustive mammography and ultrasonography. Biopsies were performed on suspicious lesions, typically with radiographic guidance. If the biopsy specimens were positive for malignancy, the patient was referred to a single surgeon (K.P.B.). A full office evaluation was performed, and a preliminary decision was made about the intended surgical treatment of the patient. After consultation with the surgeon, bilateral breast MRI was performed at 1.5 T, including bilateral dynamic scanning with axial acquisition. Kinetic analyses were facilitated by an MRI computer-aided detection system. If MRI detected an additional lesion that was not identified on prior imaging, the patient returned for a second-look ultrasonogram (for MRI-detected masses) or mammogram (for MRI-detected calcifications). All patients who underwent a second-look mammogram had an ultrasonogram as well. If the lesion was seen and still appeared suspicious, biopsy with image guidance was performed. If the lesion was not seen on follow-up ultrasonography or mammography, the patient underwent either an MRI-guided biopsy if the lesion was suspicious on initial MRI or 6-month follow-up MRI if the lesion was less concerning on initial MRI in the opinion of the attending breast radiologist. Change in surgical management If the pathologic findings of the MRI-detected lesion biopsy specimen were malignant or suspicious, the patient was reassessed by the same surgeon to determine if the MRI-identified malignancy altered the previously determined surgical management. The change in management was divided into 3 categories. First, MRI could result in a change from lumpectomy to mastectomy when the new lesion resulted in multicentric disease or the lesion appeared to be much more extensive on MRI so that more than a quadrantectomy would be required. Second, MRI could force a wider excision when an adjacent lesion or more extensive primary lesion was detected but lumpectomy was still possible. Finally, a change was also made when MRI discovered an otherwise undetected lesion in the opposite breast that resulted in contralateral surgery. Controversial changes in management based on breast MRI were discussed at the institutional multidisciplinary breast conference. Radiographic-pathologic correlation After surgery, all radiographic and pathologic results were reviewed to evaluate concordance by a single reviewer (K.Y.B.). In patients with a change in surgical treatment, the size of the lesion on mammography, ultrasonography, MRI, and final pathology reports was compared to determine whether the change was beneficial or unnecessary. The pathologists had no knowledge of the radiographic findings. Beneficial changes were those in which the pathologic results correlated with the MRI finding but not with those on mammography or ultrasonography. Moreover, the discordance had to be greater than 2.0 cm, a distance beyond what typical margins would encompass. Unnecessary changes were those in which MRI predicted an additional lesion or more extensive lesion that was missed on mammograms and ultrasonograms but on pathologic evaluation, the malignancy was concordant with the original mammography or ultrasonography findings rather than the subsequent MRI measurement. Statistical analysis Statistical analysis was performed using SPSS statistical software, version 14.0 (SPSS Inc, Chicago, Ill). Categorical variables were evaluated with χ2 analysis, and continuous variables were examined with independent-sample t tests. A binary logistic regression model was developed to identify factors that were associated with a beneficial change in management. The statistical significance level was set at P = .05. Results During the 1-year study period, 242 women with breast cancer were treated, and 190 women underwent bilateral breast MRI. Of the 190 who underwent MRI, 35 were excluded from the analysis because of undergoing MRI evaluation for recurrent breast cancer (n = 14), because of surgery for atypical ductal hyperplasia (ADH) that was subsequently found to be malignant (n = 2), because they were referred from an outside hospital for evaluation of positive margins (n = 8), or because they received neoadjuvant chemotherapy before MRI (n = 11). Of the remaining 155 women, the mean age was 53 years (range, 34-75 years). A total of 90 women (58.1%) presented with mammographic findings, and 64 (41.3%) presented with a palpable mass. Mastectomy was performed for 31 patients (20.0%), and lumpectomy was performed for 124 patients (80.0%). Thirty-three patients (21.3%) had ductal carcinoma in situ on final pathologic evaluation, 100 (64.5%) had invasive ductal carcinoma, 18 (11.6%) had invasive mammary carcinoma (a combination of ductal and lobular), and 4 (2.6%) had invasive lobular carcinoma. Forty patients (32.8%) with invasive cancer had node-positive disease. A total of 155 women with a newly diagnosed breast cancer underwent breast MRI in concordance with the study protocol. A total of 124 suspicious lesions were detected on MRI in 73 patients (Table 1). Sixty-five of the 73 patients required further imaging (ultrasonography or mammography) in an attempt to further evaluate the newly discovered MRI lesion. Of the 8 patients who did not undergo further imaging, 2 patients proceeded straight to wider excision and the other 6 had 6-month follow-up MRI recommended. In 24 of the 65 patients, the MRI finding appeared benign with follow-up ultrasonography and mammography. The remaining 41 patients in whom the MRI-detected lesion was still concerning after follow-up imaging underwent image-guided biopsy (ultrasonography guided, stereotactic, or MRI guided). Nine of those patients had malignancies apparent on biopsy specimens, 4 were found to have ADH or atypical lobular hyperplasia, and the remaining 28 had benign lesions. The false-positive rate for biopsy of an MRI-detected lesion was 78.0% (32/41). Breast MRI altered the surgical management of patients with newly diagnosed breast cancer in 36 (23.2%) of 155 patients (Table 2). Ten patients who were initially candidates for BCT were upgraded, based on MRI, to a mastectomy. Of the 10, 3 patients had an additional focus of cancer detected on MRI that resulted in multicentric disease, and all of these were biopsy proven. The other 7 patients were borderline candidates for BCT, but the MRI demonstrated that the primary lesion was more than 2.0 cm larger than previously thought by mammography and ultrasonography. Because these patients were borderline candidates for BCT even before the MRI, this larger size as seen on MRI resulted in these women being upgraded to a mastectomy as well. On the basis of MRI findings, 21 women required a wider excision but were still able to undergo a lumpectomy. In 13 patients there was a separate MRI-detected lesion larger than 2.0 cm from the primary site, which still allowed for a lumpectomy but mandated a wider excision. In 8 women, MRI showed that the primary lesion was larger than its appearance on mammography and ultrasonography, and as such, a wider excision was performed. Of the 21 women who required a wider excision, 19 women had wire-bracketed localization performed to ensure excision of the entire suspicious area, and 2 women had a wider excision at the time of surgery without bracketing assistance. All patients in the study received bilateral breast MRI, and 7 patients had a suspicious lesion discovered in the contralateral breast. Of these 7 patients, 2 had a biopsy-proven malignancy, and 2 patients had a biopsy specimen that demonstrated ADH. The remaining 3 patients had a suspicious lesion detected on MRI in the contralateral breast and instead chose not to undergo a needle biopsy of the lesion but to have a prophylactic mastectomy. A radiographic-pathologic correlation was performed to assess whether the change in surgical management based on MRI was beneficial owing to better concordance between MRI and surgical pathologic findings than between mammography or ultrasonography and surgical pathologic findings. Of the 36 women who had a change in surgical management based on MRI findings, 15 were found to have a beneficial change when MRI findings were confirmed on the final pathologic report. Eight of the 10 women who had an initially planned lumpectomy converted to mastectomy based on MRI were converted appropriately. Five of the 21 women who had a wider excision had a beneficial change because the MRI correlated with pathologic findings and the lesion was more than 2 cm larger than the mammogram or ultrasonogram had predicted. In the 7 women with contralateral MRI-detected lesions, the final pathologic report confirmed the 2 biopsy-proven malignancies. A malignancy was not found on the final pathologic report for the 2 patients who underwent contralateral lumpectomy for ADH or the 3 patients who elected to undergo a mastectomy rather that preoperative image-guided biopsy to determine the histologic features of the MRI abnormality. We analyzed patient demographics, breast cancer risk factors, radiographic data, pathologic features, and staging. On univariate analysis, the only significant factors were that patients with larger tumors on pathologic reports (T2 and T3) had a change in surgical management more often than those with smaller tumors (T0 and T1) (35.7% vs 18.9%; P = .03), and patients with advanced-stage (IIB, IIIA, IIIB, or IIIC) compared with early-stage disease (0, I, or IIA) had a beneficial change in management more often (22.2% vs 7.0%; P = .02). Multivariate analysis using logistic regression could not identify any significant factors that led to a change in surgical management; however, multivariate analysis showed that patients with larger tumors were more likely to have a beneficial change in surgical management (odds ratio, 1.66; 95% confidence interval, 1.04-2.66). The effect of our increasing experience with breast MRI during the 1-year course of the study was analyzed. Patients were divided into halves based on the date of initial breast MRI. Breast MRI led to a change in surgical management more frequently during the second half of the study (18.2% [16/88] vs 29.9% [20/67]; P = .08). The ability of breast MRI to prompt a beneficial change in surgical management improved as well (5.7% [5/88] vs 14.9% [10/67]; P = .05). No statistically significant change was found in the rate of recommendations for 6-month follow-up MRI (12.5% [11/88] vs 17.9% [12/67]; P = .49). Comment This study evaluated the impact of breast MRI on the surgical management of newly diagnosed breast cancers. Patients were treated by a strict protocol in which a single surgeon (K.P.B.) at a single institution assessed all patients before breast MRI. The timing of the MRI in the patient's workup was standardized. The patient was then reevaluated by the surgeon to determine whether a change in surgical plan was merited. A total of 23.2% of women had a change in management, and that alteration was found to be beneficial in 41.7% of those patients on radiographic-pathologic correlation. Thus, 9.7% of women had a beneficial change in surgical management based on preoperative bilateral breast MRI. Therefore, 10 women must undergo a breast MRI for 1 to have a beneficial change in management. Prior studies have attempted to retrospectively evaluate the effect of breast MRI on clinical management. Tillman et al16 found that the clinical management of 20% of patients changed with breast MRI; however, this change included the need for additional biopsies and changes in surgical management. Tillman and colleagues also attempted to quantify the benefit of the effect of MRI and found that 8% of those with a change in clinical management had a “strongly favorable effect” using a definition comparable to our definition of a beneficial change in surgical management. Bedrosian et al17 categorized the type of change in surgical management based on breast MRI and found that 4% underwent a wider excision, 5% underwent a separate lumpectomy, and 16.5% were upgraded from lumpectomy to mastectomy. However, these studies had multiple inconsistencies, including the failure to perform ultrasonography routinely, the timing of MRI with respect to surgery, the failure to perform bilateral MRI routinely, the instability of MRI technology owing to lengthy study periods, and the ultimate indication for breast MRI. Numerous reports6,18,19 have also described the identification of previously undetected lesions in the contralateral breast using MRI in 3.8% to 5.0% of patients. Our rate of detection of a contralateral malignancy was 1.3%, which may be attributable to the routine use of mammography and ultrasonography before MRI in our study and at our institution. Our analysis differs from previous studies in that mammography and ultrasonography were performed exhaustively before the use of breast MRI. Breast ultrasonography has gained acceptance in the management of patients with breast cancer and has been shown to adequately assess tumor size and extent.20 Compared with prior studies, we had a lower rate of identifying previously undetected lesions with MRI. Therefore, the use of ultrasonography before MRI may detect lesions around the primary tumor, thus lowering the apparent yield of subsequent breast MRI. Unfortunately, breast MRI also has a considerable false-positive rate and results in significant inconvenience and expense in the way of additional biopsies and imaging. In addition to patient anxiety about this additional workup, the costs of biopsies and additional MRIs are significant. Our study had an overall false-positive rate of 79.5% (58/73) in which MRI-detected lesions were ultimately benign. The biopsy false-positive rate was 78.0% (32/41). This is similar to the acceptable false-positive rate for stereotactic core biopsy based on screening mammography.21 Furthermore, 2 women were upgraded from lumpectomy to mastectomy who did not have a beneficial change in surgical management. Both of these women were borderline candidates for BCT, and they declined a biopsy of the area identified by MRI and chose to proceed with a mastectomy. On the final surgical pathologic report, the tumor size correlated with the size on mammography; thus, the MRI findings exaggerated the extent of the tumor, resulting in these women undergoing an unnecessary mastectomy. The cost of MRI technology has been a significant limitation and deterrent in the routine use of breast MRI. Reluctance on the part of insurance companies to cover the cost of the study has also restricted its use; however, the Centers for Medicare and Medicaid Services have broadened their list of indications for breast MRI to include “determination of the extent of disease in patients with known malignancy, prior to treatment (to assure confinement to 1 segment of the breast).”22 Furthermore, the reimbursement of breast MRI has been declining during the past 5 years. In 2006, the combined Medicare professional and hospital reimbursement for bilateral breast MRI (Current Procedural Terminology code 76094) was $1313.23 Cost-effectiveness must also consider follow-up imaging and potential biopsies based on the MRI findings; however, if the specificity of breast MRI continues to improve as it did between the first and second half of our study, these costs should continue to decrease. On the basis of our data, 10 women must undergo breast MRI to result in a benefit to 1 patient. Studies of prophylactic mastectomy in high-risk women have demonstrated a number needed to treat of 6 to avoid 1 case of cancer.24-26 Women with a newly diagnosed breast cancer must be considered high risk, and a number needed to treat of 10 is reasonable in our opinion. Some argue that that MRI-detected lesions are not clinically relevant.27,28 Women undergoing lumpectomy should receive postoperative external beam radiation, resulting in irradiation of any small ipsilateral lesions that may be missed by mammography and ultrasonography. Thus, MRI may simply be detecting disease that is already being addressed by radiation and could not be detected before the use of MRI. In addition, women with breast cancer are aggressively screened with physical examinations and mammography to identify a recurrent or new primary cancer. As a result, a lesion in the contralateral breast that is otherwise undetected by the initial mammography and ultrasonography would be detected early because of these aggressive screening practices. Opponents of routine MRI use often dismiss the significance of MRI-detected 1- to 2-mm lesions as too small, which would theoretically be addressed with postoperative radiotherapy. However, if we believe that it is important to clear lumpectomy margins of microscopic disease to minimize the risk of local recurrence, it would follow that small foci detected on MRI also warrant identification and excision. Use of MRI results in a beneficial change in surgical management in 9.7% of patients. This percentage is likely to increase with time as MRI technology progresses, the radiologists' experience improves, and the cost of MRI decreases. Additional malignancies are uncovered in 1 patient for every 10 who undergo MRI. These data suggest that breast MRI may have a role in the staging evaluation of newly diagnosed breast cancers. Back to top Article Information Correspondence: Kevin P. Bethke, MD, Northwestern Surgical Associates, 676 N St Clair, Suite 1525A, Chicago, IL 60611 ([email protected]). Accepted for Publication: December 13, 2006. Author Contributions: Drs Bilimoria and Bethke had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Bilimoria and Bethke. Acquisition of data: Bilimoria and Cambic. Analysis and interpretation of data: Bilimoria and Bethke. Drafting of manuscript: Bilimoria, Hansen, and Bethke. Critical revision of the manuscript: Bilimoria, Cambic, Hansen, and Bethke. Statistical analysis: Bilimoria. Study supervision: Bethke. Financial Disclosure: None reported. Funding/Support: Dr Bilimoria is supported by a grant from the Goldberg Family Charitable Trust. Previous Presentation: This study was presented at the 114th Annual Meeting of the Western Surgical Association; November 13, 2006; Los Cabos, Mexico. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript. Acknowledgment: We acknowledge Jacqueline Jeruss, MD, PhD, for her comments on the manuscript. We also appreciate the assistance of Holly Little, BS, and the Lynn Sage Comprehensive Breast Center Database staff for their help with data collection. References 1. Anastassiades OIakovou EStavridou NGogas JKarameris A Multicentricity in breast cancer: a study of 366 cases. Am J Clin Pathol 1993;99238- 243PubMedGoogle Scholar 2. Holland RVeling SHMravunac MHendriks JH Histologic multifocality of Tis, T1-2 breast carcinomas: implications for clinical trials of breast-conserving surgery. Cancer 1985;56979- 990PubMedGoogle ScholarCrossref 3. Schwartz GFPatchefsky ASFeig SAShaber GSSchwartz AB Clinically occult breast cancer: multicentricity and implications for treatment. Ann Surg 1980;1918- 12PubMedGoogle ScholarCrossref 4. Schwartz GFPatchesfsky ASFeig SAShaber GSSchwartz AB Multicentricity of non-palpable breast cancer. Cancer 1980;452913- 2916PubMedGoogle ScholarCrossref 5. Bluemke DAGatsonis CAChen MH et al. Magnetic resonance imaging of the breast prior to biopsy. JAMA 2004;2922735- 2742PubMedGoogle ScholarCrossref 6. Liberman LMorris EAKim CM et al. MR imaging findings in the contralateral breast of women with recently diagnosed breast cancer. AJR Am J Roentgenol 2003;180333- 341PubMedGoogle ScholarCrossref 7. Esserman LWolverton DHylton N Magnetic resonance imaging for primary breast cancer management: current role and new applications. Endocr Relat Cancer 2002;9141- 153PubMedGoogle ScholarCrossref 8. Bartelink HHoriot JCPoortmans P et al. Recurrence rates after treatment of breast cancer with standard radiotherapy with or without additional radiation. N Engl J Med 2001;3451378- 1387PubMedGoogle ScholarCrossref 9. Fisher BRedmond C Lumpectomy for breast cancer: an update of the NSABP experience: National Surgical Adjuvant Breast and Bowel Project. J Natl Cancer Inst Monogr 1992; ((11)) 7- 13PubMedGoogle Scholar 10. Veronesi UCascinelli NMariani L et al. Twenty-year follow-up of a randomized study comparing breast-conserving surgery with radical mastectomy for early breast cancer. N Engl J Med 2002;3471227- 1232PubMedGoogle ScholarCrossref 11. Fischer UZachariae OBaum Fvon Heyden DFunke MLiersch T The influence of preoperative MRI of the breasts on recurrence rate in patients with breast cancer. Eur Radiol 2004;141725- 1731PubMedGoogle Scholar 12. Harris JR Radiation therapy for invasive breast cancer: not just for local control. J Clin Oncol 2005;231607- 1608PubMedGoogle ScholarCrossref 13. Vinh-Hung VVerschraegen C Breast-conserving surgery with or without radiotherapy: pooled-analysis for risks of ipsilateral breast tumor recurrence and mortality. J Natl Cancer Inst 2004;96115- 121PubMedGoogle ScholarCrossref 14. Whelan TJJulian JWright JJadad ARLevine ML Does locoregional radiation therapy improve survival in breast cancer? a meta-analysis. J Clin Oncol 2000;181220- 1229PubMedGoogle Scholar 15. Early Breast Cancer Trialists' Collaborative Group, Effects of radiotherapy and of differences in the extent of surgery for early breast cancer on local recurrence and 15-year survival: an overview of the randomised trials. Lancet 2005;3662087- 2106PubMedGoogle ScholarCrossref 16. Tillman GFOrel SGSchnall MDSchultz DJTan JESolin LJ Effect of breast magnetic resonance imaging on the clinical management of women with early-stage breast carcinoma. J Clin Oncol 2002;203413- 3423PubMedGoogle ScholarCrossref 17. Bedrosian IMick ROrel SG et al. Changes in the surgical management of patients with breast carcinoma based on preoperative magnetic resonance imaging. Cancer 2003;98468- 473PubMedGoogle ScholarCrossref 18. Lee SGOrel SGWoo IJ et al. MR imaging screening of the contralateral breast in patients with newly diagnosed breast cancer: preliminary results. Radiology 2003;226773- 778PubMedGoogle ScholarCrossref 19. Lehman CDBlume JDThickman D et al. Added cancer yield of MRI in screening the contralateral breast of women recently diagnosed with breast cancer: results from the International Breast Magnetic Resonance Consortium (IBMC) trial. J Surg Oncol 2005;929- 16PubMedGoogle ScholarCrossref 20. Yang WTLam WWCheung HSuen MKing WWMetreweli C Sonographic, magnetic resonance imaging, and mammographic assessments of preoperative size of breast cancer. J Ultrasound Med 1997;16791- 797PubMedGoogle Scholar 21. Bassett LWinchester DPCaplan RB et al. Stereotactic core-needle biopsy of the breast: a report of the Joint Task Force of the American College of Radiology, American College of Surgeons, and College of American Pathologists. CA Cancer J Clin 1997;47171- 190PubMedGoogle ScholarCrossref 22. Breast ImagingMammography/Breast Echography (Sonography/Breast MRI Ductography). New York, NY Centers for Medicare and Medicaid Services2006; 23. Medicare Physician Fee Schedule for Illinois, Locality 16. 2006 http://www.wpsic.com/medicare/provider/2006_price-fees.shtml. Accessed August 12 24. Cook RJSackett DL The number needed to treat: a clinically useful measure of treatment effect. BMJ 1995;310452- 454PubMedGoogle ScholarCrossref 25. Hamm RMLawler FScheid D Prophylactic mastectomy in women with a high risk of breast cancer. N Engl J Med 1999;3401837- 1839PubMedGoogle ScholarCrossref 26. Hartmann LCSchaid DJWoods JE et al. Efficacy of bilateral prophylactic mastectomy in women with a family history of breast cancer. N Engl J Med 1999;34077- 84PubMedGoogle ScholarCrossref 27. Morrow M Magnetic resonance imaging in breast cancer: one step forward, two steps back? JAMA 2004;2922779- 2780PubMedGoogle ScholarCrossref 28. Morrow M Magnetic resonance imaging in the preoperative evaluation of breast cancer: primum non nocere. J Am Coll Surg 2004;198240- 241PubMedGoogle ScholarCrossref Baiba J. Grube, MD, New Haven, Conn: More than 30 years ago, the National Surgical Adjuvant Breast Project (NSABP) initiated the design of the B-06 randomized trial to determine if lumpectomy with or without radiotherapy was as effective as total mastectomy for the treatment of invasive breast cancer. With more than 20 years of follow-up, lumpectomy followed by radiotherapy for women with breast cancer continues to be appropriate therapy, provided that the margins of resection are tumor free and the cosmetic outcome is satisfactory. Likewise, the NSABP B-17 trial randomized women with noninvasive carcinoma to lumpectomy with or without radiotherapy and demonstrated a higher local recurrence rate without radiotherapy, but no difference in survival. These trials were initiated in an atmosphere of conflicting knowledge about the biology of breast cancer. Isolated nonrandomized studies showed that there were no significant differences in survival among the women treated with less invasive surgery, if followed by radiotherapy, when compared with more radical procedures. There was, however, knowledge that occult foci of breast cancer cells could be identified in quadrants away from the index lesion in autopsy and mastectomy specimens in 21% to 63% of cases. Despite this knowledge of more extensive multifocal and multicentric disease, local recurrence after breast conservation and radiotherapy was 14% in the NSABP B-06 trial, done in an era of less effective chemotherapy given only for node-positive disease. These data suggest that radiotherapy treats some occult disease and that perhaps some occult disease never manifests itself. Ipsilateral recurrence rates have further improved as the technology of standard low-cost breast imaging has improved, as our attention to and ability to assess margins have become more detailed, as chemotherapy regimens have achieved greater response rates, and as more sophisticated radiotherapy treatment planning protocols have been instituted. I am going to discuss 3 aspects of the study and ask 3 questions. 1. Breast conservation treatment has been recommended as the preferred treatment by the National Cancer Institute Consensus Conference since the early 1990s in appropriate candidates and according to patient preference. This recommendation was based on the results of randomized studies that enrolled women whose breast cancer had been assessed by standard mammographic imaging and clinical criteria. My first question is, Knowing the results of the NSABP B-06 trial and other similar randomized trials, that breast conserving surgery followed by radiotherapy results in similar survival and low in-breast recurrence and that MRI can detect occult malignant lesions in 10% to 54% of reported studies, what discussion did the authors undertake with patients about their choice to undergo MRI and how the findings on the MRI might affect local therapy in the absence of proven clinical trials? Should the findings of an MRI study now be considered a contraindication to breast conservation without randomized prospective data on the impact on in-breast recurrence? Is this not the same argument that was used in the 1970s by the naysayers to breast conservation that used the serial subgross pathologic data to defend the role of mastectomy for breast cancer and declare that studies of breast conservation were inappropriate? 2. In the manuscript, the authors proposed to test the hypothesis that the routine use of MRI can detect occult lesions that can alter surgical management and that this change in treatment is beneficial. The authors have provided an algorithm for the workup of a new breast cancer. In the center of the algorithm stands MRI evaluation, which dominates the decision tree. In fact, the authors provide thorough, well-analyzed data obtained from a single team of physicians with state-of-the-art imaging and multidisciplinary discussion on the immediate impact of MRI in their patient population. The term the authors use to describe the outcome is beneficial or unnecessary. The authors' definition of beneficial changes is when the pathology results correlate with MRI findings. As clinicians, the term beneficial should have a broader meaning than correlation of breast pathology with MRIs. It is easy to forget that MRI workup of the ipsilateral breast is unlikely to improve survival. The purpose of the MRI evaluation is to change treatment for the few patients who are destined to have in-breast recurrence and require mastectomy at a later date. More than half of the women in this study underwent a change in surgical management that could be described as “unnecessary.” The MRI data lead to 2 unnecessary ipsilateral mastectomies and 3 unnecessary contralateral prophylactic mastectomies based on MRI findings. This is a high price to pay for preventing a mastectomy at a later time for perhaps 5% of the breast conservation patients treated with current therapies. My second question is, Have the “beneficial” and “problematic” results of the authors' experience with the use of MRI on all patients diagnosed as having breast cancer altered their use, interpretation, and patient involvement in the decision-making process to even order an MRI been modified? Have the authors instituted quality-of-life questionnaires for women going through these repeat diagnostic studies and for those who have had an undesired change in treatment? 3. Securing clear surgical margins affects local recurrence, but there is not a uniform agreement on what constitutes a negative margin. Planned surgical resection is based on clinical and image-detected findings. The authors suggest that MRI can define extent of disease and delineate the area of resection. My third question: How often did the authors have to go back to reexcise positive margins in the patients evaluated with MRI compared with their previous experience without MRI? Magnetic resonance imaging is a highly sensitive technique with problems in specificity, which may be forcing women into mastectomies and extra biopsies unnecessarily. It can be heard in some surgical circles that all the great surgical questions have been answered in breast cancer. I challenge the authors to continue to be in the forefront of this area of surgical research and design a trial with colleagues in radiation oncology, pathology, and breast imaging that can provide hard data on the role of MRI in selection of the few patients who are not candidates for breast conservation and perhaps define a role for MRI in selecting patients who are good candidates for only partial breast irradiation or no breast radiotherapy at all. I would also compliment the program committee for selecting this topic and identifying this subject as a critical issue for surgeons to understand, whether to incorporate the use of MRI into practice without randomized prospective data or to challenge the push by our colleagues in breast imaging to change 30 years of breast cancer treatment founded on the principles of clinical trial data. Dr Bethke: The first question was, Should MRI findings be a contraindication to BCT? If the MRI showed additional suspicious findings in our study patients, and a subsequent biopsy confirmed more extensive malignancy than that seen on traditional breast imaging, we proceeded to a mastectomy if a standard lumpectomy or quadrantectomy was no longer possible. As breast surgeons we spend a lot of time and effort clearing margins of microscopic disease, and occasionally we will perform multiple reexcisions on a patient who is very keen on breast conservation. If we feel that it is important to clear the breast of microscopic disease, then it would follow that it is also important to clear the breast of macroscopic disease. The initial NSABP studies had local recurrence rates of 12% to 15%. We now have local recurrence rates of 5% or less. Presumably our aggressive approach to clearing the breast of residual disease plays a role in these improved results. More recently, some would argue that this improved local control translates into a modest improved survival. Most everyone would agree that our goal is to surgically remove as much of the cancer from the breast as possible prior to radiation, using safe, efficient, and cost-effective methods. Magnetic resonance imaging is not yet at the level of usefulness we desire but will continue to evolve and improve. The second question was, Have our indications for preoperative breast MRI changed based on the results of this study? Yes. During the study period, nearly everyone with a newly diagnosed breast cancer received a preoperative breast MRI. Our current indications are to perform MRI on patients younger than 40 years or on those whom our radiologists indicate have dense breasts. Patients with multifocal breast cancer involving an entire quadrant are evaluated with MRI to rule out multicentric disease, which would then require a mastectomy. We will also obtain an MRI on women with invasive lobular cancer and patients presenting with occult breast cancer manifested by nodal metastases in the absence of mammographic or sonographic abnormalities. We have not instituted a quality-of-life survey for these patients, but I’m sure it would show a fair amount of anxiety related to MRI's relatively low specificity and the need for additional testing. The last question was, What was our margin reexcision rate? Our overall reexcision rate for the entire group of 155 patients was 15.0%. For those patients in whom MRI changed management, the reexcision rate was 28%, and in those in whom it did not change management, 13%. The higher rate of reexcision in those whom MRI changed management probably reflects the more extensive nature of the malignancy and the fact that it was not seen on an ultrasonogram or mammogram, making localization difficult. We do not have an accurate estimate for the reexcision rate before our MRI study, but anecdotally MRI did not appear to lower our overall rate. Alden H. Harken, MD, Oakland, Calif: I think we tend to look at screening tests by virtue of their sensitivity and specificity. As we do so, I can understand how you can detect more breast cancer, or any kind of cancer, if you add more tests, so I look at the basic biology of breast disease and then the mechanism by which each additional testing technique asks the question. X-rays are actually testing for radiodensity, MRI testing examines some kind of polar molecule like water (or how the hydrogen atom “precesses”), positron emission tomographic scanning assesses how cells use glucose, and maybe ultrasonography explores how sonar dense the cells are. Is there something fundamental about breast cancer, the biology of breast cancer, that would make it uniquely visible by any one of these tests, or are you just adding more tests to increase sensitivity at the expense of specificity and cost? Dr Bethke: Our current breast imaging techniques, mammography, ultrasonography, and MRI, are not specific for breast cancer because they detect density differences (mammogram and ultrasonography) or changes in vasculature (MRI). Thus, they have difficulty differentiating many benign from malignant changes. We did this study because we were frustrated by the lack of guidelines for the use of breast MRI. Like you, we were concerned about its low specificity. We had hoped to develop clearer indications for MRI in the evaluation of newly diagnosed breast cancers; however, the only statistically significant predictor of a beneficial change in management was pathologic size, which isn't clinically useful because it is determined after excision. We are hopeful that breast MRI specificity will continue to improve as equipment and software improve. We were encouraged by the significant improvement in beneficial change noted between the first and second half of the study. Jose M. Velasco, MD, Skokie, Ill: I believe there are 2 points worth being highlighted: one was, who should have an MRI; and two, once an MRI is obtained, what should we do with the results? I would like to say that we should evaluate the role of MRI in the management of patients with breast disease, instead of just evaluating the role of breast MRI in the surgical management of these patients. In other words, do you think MRI could have a role in identifying patients who may have T2 or T3 lesions? Those patients could benefit from neoadjuvant therapy and thus be eligible for breast-preserving operations. Therefore, one could avoid mastectomy. Did you look at that? Along those lines, is there a controversy of the role, if any, of MRI for evaluating lymph node status? Is there any known correlation? Did you identify sentinel nodes on those patients? Finally, what criteria did you use to discern and identify the group of patients who needed a biopsy? Were they only those patients in whom the MRI, ultrasonogram, and mammogram identified positive lesions, or did you perform biopsies on all of them using MRI-guided biopsy? Dr Bethke: You asked if there might be a role for using MRI to direct breast cancer chemotherapy. The 11 patients who received a breast MRI and underwent neoadjuvant chemotherapy were excluded from our study because of the small number of patients and inconsistencies in their management. Magnetic resonance imaging has the potential to be helpful in following chemotherapy response, but at this time the poor specificity and high cost prohibit its routine use for this purpose. Alternatively, I believe that it may play a future role in optimizing patient selection for partial breast radiation by ruling out multicentricity. We did not use MRI to evaluate lymph nodes in our study, and there is no literature to support its use for this purpose. Sentinel lymph node biopsy was used to evaluate nodes. If a lesion was considered suspicious on MRI, a follow-up focused ultrasonogram was used to further evaluate the lesion. If it was visible on the ultrasonogram, an ultrasonography-guided core biopsy was performed. If it was considered suspicious and not seen on the ultrasonogram, an MRI-guided core biopsy was performed. Financial Disclosure: None reported.

Fatima, Javairiah;Baron, Todd H.;Topazian, Mark D.;Houghton, Scott G.;Iqbal, Corey W.;Ott, Beverly J.;Farley, David R.;Farnell, Michael B.;Sarr, Michael G.

doi: 10.1001/archsurg.142.5.448pmid: 17515486

Abstract Objective To review our experience with management of pancreaticobiliary and duodenal (PB/D) perforations after periampullary endoscopic interventions. Although pancreaticobiliary and duodenal perforations after periampullary endoscopic procedures are rare, their management has not been well described. Patients Individuals who experienced pancreaticobiliary and duodenal perforations. Main Outcome Measures Comorbidities, interventions performed, mechanism/site of perforation, management, and hospital morbidity/mortality. Results Seventy-five perforations (0.6%) occurred in 12 427 procedures; 20 perforations (27%) occurred during biliary stricture dilatation, 18 (24%) during diagnostic endoscopic retrograde cholangiopancreatography, and 15 (20%) during management of choledocholithiasis. Perforations were caused by guidewire insertion in 24 patients (32%), sphincterotomy in 11 (15%), passage of the endoscope in 8 (11%), or stent migration in 7 (9%) and were identified during the index procedure in 45 patients (60%). Delayed presentations included pain in 33 patients (44%), leukocytosis in 26 (35%), and/or fever in 13 (17%) and were diagnosed using computed tomography in 19 patients (25%) and abdominal radiography in 10 (13%); 9 cases (12%) were diagnosed more than 24 hours after the procedure. Indications for operative treatment were gaping duodenal perforations and perforations in patients with surgically altered anatomy. Indications for nonoperative management included contained bile duct perforations and focal duodenal perforations. Management was nonoperative in 53 patients (71%) and operative in 22 (29%). Patients with duodenal perforations, higher American Society of Anesthesia status (P<.01 each), and older age (mean ± SEM, 65 ± 4 vs 55 ± 2 years; P = .02) were more likely to require operative management. Hospital stay (mean ± SEM, 16 ± 4 vs 4 ± 1 days; P<.05) and mortality (13% vs 4%; P<.05) were greater in operative patients (P<.05 each). Conclusions Most (70%) pancreaticobiliary and duodenal perforations secondary to periampullary endoscopic interventions can be managed nonoperatively. Most biliary perforations can be managed nonoperatively; a requirement for operative treatment increases the mortality rate. Endoscopic retrograde cholangiopancreatography (ERCP) has evolved from a diagnostic tool to a primarily therapeutic intervention in pancreaticobiliary disorders. This evolution has led to an increase in the volume and complexity of these procedures, and, consequently, complications have increased in number and spectrum. Theseinvasive therapeutic procedures carry the risk of perforation of the bile duct, pancreatic duct, and/or duodenum.1-7 Duodenal perforations may be retroperitoneal (typically in the periampullary region and due to sphincterotomy) or intraperitoneal (typically in the lateral wall and due to endoscope passage). Although perforation as a result of ERCP is rare, the potentially serious nature of this complication, with a reported mortality of up to 25%, mandates a better understanding of the factors that predispose to this complication and the most appropriate management for perforations occurring after the various periampullary interventions.8-11Multiple studies6,9,12,13 have reported an incidence of ERCP-related perforation of 0.3% to 1.0% and have examined the risk factors that predisposed to perforation. Some groups1,9 have attempted to stratify the features of perforation based on clinical and radiographic findings to help guide the course of management. At our tertiary referral center, a high volume of diagnostic and therapeutic endoscopic periampullary procedures are performed. Hence, the aim of this study was to determine the incidence of pancreaticobiliary and duodenal perforations after periampullary interventions, the risk factors predisposing to perforations, the operative and nonoperative management options, and their respective clinical outcomes. Methods After approval from the Mayo Clinic institutional review board we queried our endoscopic database to identify all patients who underwent an endoscopic periampullary procedure that resulted in perforation of the duodenum, pancreatic duct, or bile duct between January 1, 1994, and December 31, 2004. Perforations resulting from transgastric manipulations were excluded from this study. Pertinent data were collected using the medical records of these patients regarding demographics, comorbidities, and history of a previous endoscopic or periampullary procedure. Indications for the procedure and the presence or absence of risk factors such as a periampullary diverticulum, precut sphincterotomy, and previous gastric, duodenal, or jejunal operations were assessed. The complexity of the periampullary endoscopic procedure, graded from 1 to 3 based on the degree of difficulty of therapeutic interventions, was evaluated.14 note was made of the site and mechanism of perforation, time of identification (during the procedure or <6, 6-24, or >24 hours after the procedure), diagnostic modality used, and other concomitant complications. Management approaches (endoscopic, operative, or nonoperative) and clinical outcomes (morbidity and mortality) were determined. Data were organized and are reported as mean ± SEM or the median and range depending on the distribution of data. Statistical analysis included tests of association with operative vs nonoperative management using Wilcoxon and Kruskal-Wallis (rank sum) tests for continuous or ordinal variables. χ2 and Fisher exact tests were used for nominal variables. Limited multivariate logistic regression modeling was then performed to adjust for confounding variables. A P<.05 was considered statistically significant. Results During the 11-year period, 12 427 endoscopic periampullary procedures, including ERCP and endoscopic retrograde cholangiography, along with therapeutic endoscopic interventions, were performed at Mayo Clinic, Rochester, Minn. Seventy-five (0.6%) of these procedures resulted in pancreaticobiliary and duodenal perforations in 23 males and 52 females. The median age of patients was 56 years (range, 14-91 years). The incidence of perforation was 0.8% (n = 70) during therapeutic procedures and 0.1% (n = 5) during diagnostic procedures. Comorbidities Almost half of these 75 patients with perforation had been treated previously for a chronic gastrointestinal disease, including primary sclerosing cholangitis (n = 21), ulcerative colitis (n = 16), and chronic pancreatitis (n = 4). Six patients had pancreatic cancer, and 2 had cholangiocarcinoma. A history of upper gastrointestinal operations was common; 25 patients had undergone a cholecystectomy, and 50 had an additional abdominal operation in the past. Whereas 64 patients (85%) had normal gastroduodenal anatomic continuity, 11 (15%) had surgically altered anatomy: 6 had Billroth II–type anatomy (loop gastrojejunostomy), 3 had Roux-en-Y gastrojejunal anatomy, and 2 had a previous gastric resection. Thirty (40%) of the 75 patients had a history of previous ERCP or other periampullary endoscopic procedures. Twelve of these 30 had a postinterventional complication at the time of the previous endoscopic periampullary procedure. Many of the patients who developed a perforation had substantial previous medical comorbidities. Thirteen patients were diabetic, 29 had clinically important cardiovascular disease, and 7 had underlying pulmonary disease. At the time of the procedure, 13 patients (17%) were identified as American Society of Anesthesia (ASA) class 1, 24 (32%) as class 2, 36 (48%) as class 3, and 2 (3%) as class 4. Indications for endoscopic intervention The most common indications for performing a periampullary procedure were dilatation of a biliary stricture in 20 patients (27%), diagnostic cholangiography in 18 (24%), and management of known or suspected choledocholithiasis in 15 (20%). Other less common indications included obstructive jaundice (n = 9; 12%), sphincter of Oddi dysfunction (n = 5; 7%), cholangitis (n = 5; 7%), and various pancreatic abnormalities (pancreatic ductal strictures, neoplasm, and ductal leak). The interventional procedures performed most commonly were sphincterotomy (biliary or pancreatic, n = 43), placement of a bile duct or a pancreatic duct stent (n = 22), and balloon dilatation of a bile duct stricture(s) (n = 17). Other less commonly performed procedures included transampullary choledocholithotomy (n = 4), cholangioscopy (n = 2), and brushing or biopsy of a bile duct lesion (n = 1). An exclusively diagnostic procedure with no additional intervention resulted in perforation in 5 patients. During the interventional procedure, 14 patients (19%) were found to have a periampullary diverticulum, 7 (9%) of whom required a precut sphincterotomy to allow the endoscopic intervention to be possible. Mechanism, site, and identification of perforation Guidewire manipulation was the presumed cause of perforation in 24 patients (32%), and 11 (15%) were sphincterotomy-related perforations, 8 (11%) were related to difficulties in passage of the endoscope, 8 (11%) occurred during cannulation, 7 (9%) occurred in the process of stent insertion, and 5 (7%) occurred during stricture dilatation (Table 1). The exact mechanism of perforation was unknown in 11 patients (15%). The duodenum and the bile duct were the most common sites of perforations (Table 1). Of the 75 perforations, 45 were identified immediately: 26 (35%) by direct endoscopic visualization and 19 (25%) by fluoroscopic identification of extravasation of contrast. Eight patients (11%) were suspected of having a perforation during the procedure, and it was confirmed using computed tomography (CT). Five other patients (7%) had the perforation diagnosed within 6 hours of the procedure, and in 8 (11%), the perforation was recognized within 6 to 24 hours. A delayed diagnosis (>24 hours) occurred in 9 patients (12%). Patients who were identified as harboring a perforation during the procedure presented later with symptoms and signs of abdominal pain (44%), leukocytosis (35%), fever (17%), hypotension (3%), and chest pain (1%). The modalities used to identify and confirm the suspicion of a perforation in these delayed presentations included CT (n = 19, 25%), abdominal radiographs showing free intraperitoneal or retroperitoneal air (n = 10, 13%), and magnetic resonance imaging (n = 1, 1%). Grade (complexity) of ercp Mean procedural complexity by grade was 1.5 ± 0.1. The technical difficulty of the periampullary interventional procedure performed was analyzed for the operatively and nonoperatively managed patients.14-16 Increasing grade (ie, increasing technical difficulty) of the interventional procedure was associated with the incidence of perforation requiring operative management (P = .02). Management The pre-intervention plan was executed endoscopically with success in 45 patients (60%); however, 30 procedures (40%) were abandoned prematurely because of the perforation. After recognition of the perforation, overall, 53 patients (71%) were managed nonoperatively, and 22 (29%) were managed operatively. The mean hospital stay for the entire group was 8 ± 1 days, with 5 deaths (7%). Endoscopic management of the perforation was attempted in 27 (36%) of these 75 patients, including 14 with duodenal perforations, 12 with biliary perforations, and 1 with pancreatic duct perforation. Endoscopic treatments for perforation included biliary stents in 10 patients, a nasobiliary drain or nasogastric tube in 5 patients each, endoscopic placement of a clip(s) in an attempt to “close” the site of perforation in 8 patients, a nasoduodenal tube in 3 patients, and a pancreatic duct stent in the 1 patient with a pancreatic duct perforation. Five of these 27 patients failed endoscopic management and were transferred to the operating room immediately. Twenty-two (81%) of the 27 patients were admitted for observation, 18 of whom were managed successfully using a nonoperative approach; 4 required operative treatment eventually. Nonoperative group There were 53 patients (71%) in this group, 15 with duodenal perforations (11 intraperitoneal and 4 retroperitoneal perforations) and 32 with bile duct perforations. Most perforations in this group resulted from endoscopic instrumentation during the procedure. Nearly half were guidewire-related perforations. A combination of antibiotic drugs and intravenous hydration and, when possible, gastric or bile duct decompression was used to manage all of these patients; 8 patients had a nasobiliary or nasoduodenal drain, 9 had a bile duct stent, and 1 had a duodenal stent. An endoscopic clip device (Hemoclip; Olympus, Melville, NY) was placed in 6 of these 53 patients. The mean hospital stay was 4 ± 1 days. Two patients with bile duct perforation died: 1 on postprocedure day 1 of a tension pneumothorax and the other on day 37 of multiorgan failure. In 6 of the 53 patients, the site of perforation was unknown. The most common presentation was abdominal pain (n = 5); 2 patients presented with fever. Computed tomography was used to diagnose the perforation in 5 patients, and an abdominal radiograph revealed free air in 1. All 6 patients had a biliary sphincterotomy performed during the procedure. The median hospital stay for this group of patients was 5 days (range, 2-13 days). Operative group Of the 22 patients (29%) in this group, 19 had duodenal perforations (9 intraperitoneal and 7 retroperitoneal perforations), 2 had perforations in the bile duct, and 1 had a perforation in the pancreatic duct; 2 patients with Billroth II anatomy had jejunal (intraperitoneal) perforations. One patient had 2 perforations: a focal duodenal perforation and a bile duct perforation. Fifteen patients underwent immediate operative intervention on identification of the perforation during the index procedure, 7 of whom had periampullary duodenal diverticula. Seven patients were managed nonoperatively for more than 6 hours before being taken to the operating room, resulting in 4 patients requiring intensive care unit admission with prolonged hospital stay and 1 death (Table 2). Seven of the duodenal perforations resulted from passage of the endoscope and were associated with large perforations that could not be managed endoscopically. Twelve patients underwent a primary duodenal repair, including 3 with an omental patch used to buttress the closure. Exploratory laparotomy with debridement and drain placement was performed in 7 patients. Three patients underwent a Roux-en-Y choledochojejunostomy, 2 with unresectable pancreatic cancer and 1 with a previous choledochocholedochostomy performed for biliary reconstruction as a part of orthotopic liver transplantation. The patient with pancreatic duct perforation underwent exploratory laparotomy on day 1 and had drainage of multiple intraperitoneal and retroperitoneal fluid collections. He developed severe necrotizing pancreatitis and cardiorespiratory complications that prolonged his hospital stay to 75 days. The mean hospital stay for patients who underwent operative management was 16 ± 4 days. Three patients (all with duodenal perforations) died on postoperative days 10, 12, and 15 with an ASA status of 3 and had required intensive care unit care. Morbidity and mortality The overall mortality due to perforations resulting from periampullary procedures was 7% (n = 5). Hospital mortality was greater in the operative group (13% vs 4%; P<.05). Four deaths were secondary to sepsis and multiorgan failure 10 to 35 days later. One patient, managed nonoperatively, died the evening of the endoscopic procedure from a tension pneumothorax. Complications after operative treatment included intra-abdominal abscess, hemorrhage, pancreatitis, and wound infection (n = 1 each). Patients in the operative group were older than those in the nonoperative group (64 ± 4 vs 55 ± 2 years; P = .02). The median ASA status was also greater (3 vs 2; P = .003), with 18 patients (82%) in the operative group having an ASA status of at least 3, whereas only 22 nonoperative patients (42%) had an ASA status of 3 or greater. Univariate analysis was performed to identify the association of variables such as age, sex, ASA status, grade of procedure, site of perforation, and duration of hospital stay between operative and nonoperative management. This analysis revealed that the mean age of patients requiring operative management was greater than that of those managed nonoperatively (65 ± 4 vs 55 ± 2 years; P = .02). An association was found between higher ASA status (≥3) and the need for an operative treatment (P = .003). Patients with duodenal perforations were more likely to require an operation than patients with bile duct perforations (P<.001). Increasing technical difficulty (grade) of the procedure was associated with requiring operative management (P = .02). The hospital stay was greater for patients requiring an operation vs those who could be managed nonoperatively (16 ± 4 vs 4 ± 1 days; P<.001). Sex difference and history of previous ERCP did not seem to predispose patients to require operative treatment; however, on applying the multivariate logistic regression model we found that the associations of ASA status and site of perforation with patients requiring operative management remained significant (P = .01 and .003, respectively). Comment Few studies have reported the incidence of perforation and its management, with no major series in the past decade.17 The retrospective review of 12 427 endoscopic periampullary procedures identified 75 perforations. The present major findings include an increase in the use of ERCP during the past decade, especially in terms of interventional procedures. This study identifies age, ASA status, site of perforation, and early recognition as critical factors in predicting patient outcome. Between 1994 and 2004, we performed 12 427 endoscopic periampullary procedures. The total number of procedures performed in 2004 was nearly twice the number in 1994 (Figure 1). Although the yearly rate of perforation (range, 0.4%-0.7%) seems to have remained relatively unchanged during this period, the absolute numbers of pancreaticobiliary and duodenal perforations show a rising trend. Seventy-three percent (n = 9109) of the periampullary endoscopies were therapeutic procedures, and 27% (n = 3318) were performed with only a diagnostic intent. During the decade, the number of therapeutic procedures performed each year showed an increasing trend, with 53% of endoscopic periampullary procedures being therapeutic in 1994 and 89% in 2005, whereas the trend for diagnostic procedures declined, probably due to the advent of better, noninvasive imaging modalities, such as endoscopic ultrasonography and magnetic resonance cholangiopancreatography. Seventy of the perforations in this period occurred as a result of therapeutic procedures, translating to an incidence of perforation in the therapeutic group of 0.8%, which was 8 times greater than the incidence of perforation in the diagnostic group (0.1%). As expected, therapeutic procedures carry a much higher risk of perforation due to the invasive nature of the endoscopic procedure. Patients who underwent operative management were older than those managed nonoperatively (65 ± 4 vs 55 ± 2 years; P = .02). Patients requiring operative treatment after a perforation were generally more ill, with multiple comorbidities, as indicated by the higher ASA status (median = 3), and required a longer hospitalization than patients requiring nonoperative management (12 ± 4 vs 4 ± 1 days); however, after adjusting for confounding variables, multivariate regression analysis demonstrated that greater ASA status (P = .001), higher grade (technical difficulty) of the endoscopic procedure (P = .06), and perforation in the duodenum (P = .003) were risk factors for the perforations requiring operative therapy. Associated medical illnesses and older age seem to be predictors of poor operative outcome.1,2 In the present study, approximately 20% of the patients with endoscopic perforations (14 of 75 patients) had a periampullary duodenal diverticulum. Seven of the 22 patients requiring operative intervention had a periampullary diverticulum, in which the perforation was noted in 5. Because the extensive database used does not detail the presence of a periampullary diverticulum, we cannot compare the presence or absence of a diverticulum and the association of a perforation in the entire experience. But, because the prevalence of periampullary diverticula is much less than 20% in the general population, these findings suggest that interventional periampullary endoscopic procedures in patients with a peri-Vaterian duodenal diverticulum are undertaken with a greater risk of perforation than when a diverticulum is absent. Early diagnosis and prompt treatment during the endoscopic procedure are vital for a better outcome.1,11,18-20 Suspicion and diagnosis of a procedure-related perforation can be facilitated greatly by clinical findings and especially radiographic imaging with contrast studies, CT, and even magnetic resonance imaging.11,21 Most important, suspicion of perforation should be high in patients with pain or fever after any periampullary endoscopic procedure. This concern is especially true when the procedure is therapeutic or technically difficult; the patient has local anatomic variations, such as periampullary diverticula22; strictures are dilated; there are changes in the usual anatomy related to previous operative interventions (eg, gastrectomy); or intubation is difficult. If there is any suspicion of perforation, prompt diagnosis and institution of systemic support with broad-spectrum antibiotics and intravenous resuscitation is mandatory; when perforation is confirmed, external biliary, duodenal, and/or pancreatic drainage may be indicated. Some patients could be managed by diversion or decompression of contents from the duodenum by placing a nasoduodenal tube alone or with internal pancreaticobiliary drainage. The management of perforations after therapeutic periampullary endoscopic procedures has remained a major controversy. The present data indicate that most ductal perforations caused by guidewire passage (21 of 24) or during attempted cannulation (6 of 8) cause focal tears or perforations, which can be managed nonoperatively using sphincterotomy, nasobiliary drains, or endoluminal stents.23-25 Successful nonoperative management of sphincterotomy-related retroperitoneal perforations is also possible, despite extensive retroperitoneal air noted on CT, provided the patients remain nontoxic.26,27 In contrast, if patients develop abdominal pain or fever or appear toxic clinically, surgical consultation should be obtained, and operative exploration for effective repair or drainage should be considered, especially in elderly or otherwise chronically ill patients less able to withstand the physiologic stress. We offer clinical management algorithms for patients noted to have a perforation either during or after completion of the procedure (Figure 2). Duodenal perforations represent a different category of injury.20,21 Many of these are not truly periampullary but rather occur on the lateral duodenal wall, are rather large free tears into the peritoneal cavity, and seem to occur secondary to trauma from difficult passage of the endoscope. These types of perforations are often recognized by the endoscopist, are difficult or impossible (currently) to repair endoscopically, and require emergency operative intervention. Early operative intervention usually allows a primary repair, similar in principle to closure of duodenal perforations secondary to duodenal ulcers. Delayed recognition or repair after failure of an attempt at nonoperative management can be devastating, requiring drainage alone without repair of the actual perforation. Moreover, several of the prolonged intensive care unit admissions and deaths in the present series came from delayed intervention, with local sepsis leading to multiorgan failure. Seven patients in the operative group (6 with duodenal perforations and 1 with perforation at the junction of the pancreatic duct with the ampulla) had a delay in receiving operative treatment; 4 of these resulted in a prolonged intensive care unit course and the death of 1 patient (Table 2). Therefore, intraperitoneal duodenal perforations require an aggressive approach; simple nasogastric/nasoduodenal tube drainage seems unwise except for very limited, focal perforations (Figure 2). A delay in diagnosis of duodenal perforations can lead to severe morbidity. The ability of endoscopic repair with intraluminally applied clips is poorly defined and seems unwise in large, nonfocal tears using the current endoscopic technology. Recent studies28 using other modalities, however, show promise for allowing endoscopic closure. Therefore, with the expanding spectrum of interventional ERCP procedures, the absolute number of resulting perforations has increased. Although it still is a rare entity, pancreaticobiliary and duodenal perforations can prove to be fatal, especially when the perforation occurs in the duodenum; operative treatment carries greater morbidity and mortality rates than perforations that can be contained nonoperatively. Correspondence: Michael G. Sarr, MD, Department of Surgery, Mayo Clinic College of Medicine, 200 First St SW (AL 2-435), Rochester, MN 55905 ([email protected]). Accepted for Publication: January 3, 2007. Author Contributions:Study concept and design: Fatima, Baron, Topazian, Houghton, Farnell, and Sarr. Acquisition of data: Fatima, Ott, and Sarr. Analysis and interpretation of data: Fatima, Houghton, Iqbal, Farley, Farnell, and Sarr. Drafting of the manuscript: Fatima, Baron, Houghton, Farley, and Sarr. Critical revision of the manuscript for important intellectual content: Fatima, Baron, Topazian, Houghton, Iqbal, Ott, Farnell, and Sarr. Statistical analysis: Fatima, Iqbal, and Sarr. Administrative, technical, and material support: Fatima, Ott, and Sarr. Study supervision: Baron, Topazian, Houghton, Farley, Farnell, and Sarr. Financial Disclosure: None reported. Previous Presentation: This paper was presented at the 114th annual scientific session of the Western Surgical Association; November 15, 2006; Los Cabos, Mexico; and is published after peer review and revision. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript. References 1. Preetha MChung YFChan WH et al. Surgical management of endoscopic retrograde cholangiopancreatography-related perforations. ANZ J Surg 2003;731011- 1014PubMedGoogle ScholarCrossref 2. Masci EToti GMariani A et al. Complications of diagnostic and therapeutic ERCP: a prospective multicenter study. Am J Gastroenterol 2001;96417- 423PubMedGoogle ScholarCrossref 3. Mallery JSBaron THDominitz JA et al. Complications of ERCP. Gastrointest Endosc 2003;57633- 638PubMedGoogle ScholarCrossref 4. Kuhlman JEFishman EKMilligan FDSiegelman SS Complications of endoscopic retrograde sphincterotomy: computed tomographic evaluation. Gastrointest Radiol 1989;14127- 132PubMedGoogle ScholarCrossref 5. Huguet JMSempere JBort I et al. Complications of endoscopic retrograde cholangiopancreatography in patients aged more than 90 years old [in Spanish]. Gastroenterol Hepatol 2005;28263- 266PubMedGoogle ScholarCrossref 6. Christensen MMatzen PSchulze SRosenberg J Complications of ERCP: a prospective study. Gastrointest Endosc 2004;60721- 731PubMedGoogle ScholarCrossref 7. Suissa AYassin KLavy A et al. Outcome and early complications of ERCP: a prospective single center study. Hepatogastroenterology 2005;52352- 355PubMedGoogle Scholar 8. Booth FVDoerr RJKhalafi RSLuchette FAFlint LM Jr Surgical management of complications of endoscopic sphincterotomy with precut papillotomy. Am J Surg 1990;159132- 135PubMedGoogle ScholarCrossref 9. Stapfer MSelby RRStain SC et al. Management of duodenal perforation after endoscopic retrograde cholangiopancreatography and sphincterotomy. Ann Surg 2000;232191- 198PubMedGoogle ScholarCrossref 10. Cotton PBLehman GVennes J et al. Endoscopic sphincterotomy complications and their management: an attempt at consensus. Gastrointest Endosc 1991;37383- 393PubMedGoogle ScholarCrossref 11. Bell RCVan Stiegmann GGoff JReveille MNorton LPearlman NW Decision for surgical management of perforation following endoscopic sphincterotomy. Am Surg 1991;57237- 240PubMedGoogle Scholar 12. Enns REloubeidi MAMergener K et al. ERCP-related perforations: risk factors and management. Endoscopy 2002;34293- 298PubMedGoogle ScholarCrossref 13. Pungpapong SKongkam PRerknimitr RKullavanijaya P Experience on endoscopic retrograde cholangiopancreatography at tertiary referral center in Thailand: risks and complications. J Med Assoc Thai 2005;88238- 246PubMedGoogle Scholar 14. Schutz SMAbbott RM Grading ERCPs by degree of difficulty: a new concept to produce more meaningful outcome data. Gastrointest Endosc 2000;51535- 539PubMedGoogle ScholarCrossref 15. Johanson JFCooper GEisen GM et al. Quality assessment of ERCP. Endoscopic retrograde cholangiopancreatography. Gastrointest Endosc 2002;56165- 169PubMedGoogle ScholarCrossref 16. Baron THPetersen BTMergener K et al. Quality indicators for endoscopic retrograde cholangiopancreatography. Am J Gastroenterol 2006;101892- 897PubMedGoogle ScholarCrossref 17. Scarlett PYFalk GL The management of perforation of the duodenum following endoscopic sphincterotomy: a proposal for selective therapy. Aust N Z J Surg 1994;64843- 846PubMedGoogle ScholarCrossref 18. Snyder WH IIIWeigelt JAWatkins WLBietz DS The surgical management of duodenal trauma: precepts based on a review of 247 cases. Arch Surg 1980;115422- 429PubMedGoogle ScholarCrossref 19. Lucas CELedgerwood AM Factors influencing outcome after blunt duodenal injury. J Trauma 1975;15839- 846PubMedGoogle ScholarCrossref 20. Howard TJTan TLehman GA et al. Classification and management of perforations complicating endoscopic sphincterotomy. Surgery 1999;126658- 663PubMedGoogle ScholarCrossref 21. Sarr MGFishman EKMilligan FDSiegelman SSCameron JL Pancreatitis or duodenal perforation after peri-Vaterian therapeutic endoscopic procedures: diagnosis, differentiation, and management. Surgery 1986;100461- 466PubMedGoogle Scholar 22. Vaira DDowsett JFHatfield AR et al. Is duodenal diverticulum a risk factor for sphincterotomy? Gut 1989;30939- 942PubMedGoogle ScholarCrossref 23. Sandha GSBourke MJHaber GBKortan PP Endoscopic therapy for bile leak based on a new classification: results in 207 patients. Gastrointest Endosc 2004;60567- 574PubMedGoogle ScholarCrossref 24. Davids PHRauws EATytgat GNHuibregtse K Postoperative bile leakage: endoscopic management. Gut 1992;331118- 1122PubMedGoogle ScholarCrossref 25. Ryan MEGeenen JELehman GA et al. Endoscopic intervention for biliary leaks after laparoscopic cholecystectomy: a multicenter review. Gastrointest Endosc 1998;47261- 266PubMedGoogle ScholarCrossref 26. Nam JSYi SY Massive pneumoperitoneum and pneumomediastinum with subcutaneous emphysema after endoscopic sphincterotomy. Clin Gastroenterol Hepatol 2004;2xxiiPubMedGoogle ScholarCrossref 27. Baron THGostout CJHerman L Hemoclip repair of a sphincterotomy-induced duodenal perforation. Gastrointest Endosc 2000;52566- 568PubMedGoogle ScholarCrossref 28. Mutignani MIacopinin FDokas S et al. Successful endoscopic closure of a lateral duodenal perforation at ERCP with fibrin glue. Gastrointest Endosc 2006;63725- 727PubMedGoogle ScholarCrossref Timothy Sielaff, MD, Minneapolis, Minn: This is an important data set for all of us who are involved in the management of patients with biliary tract and periampullary diseases and is important also as a description of the utilization of ERCP and the management of its complications. It is a single-institution experience, and that affords an opportunity to develop as uniform as possible a management approach to this very rare complication. The results, however, are broadly applicable to all of us who work in an institution where ERCPs are performed and for all of us who are called to help manage these complications. I just have 3 relatively simple questions. Number 1, there is a striking trend away from diagnostic ERCP, and that is very useful for all of our practices during this period. What are the indications for diagnostic ERCP in the year 2006 in Rochester and what should they be generalized to in the nation as a whole? Number 2, what are the predictors of perforation? We talked about those patients who had perforations, what their complicating factors were. But what is the denominator? What are the risk factors in the entire data set? And are there patients for whom the risk factors are so high that maybe endoscopic intervention should be avoided and other percutaneous interventions should be considered? Last, could you stratify the risks associated with these patients? If we get a call from the endoscopy suite about a perforation, could you stratify these patients into those who are very low risk and we need not be terribly concerned about, which patients require very close observation, and what tests need to be done to help monitor those patients and confirm that nonoperative therapy is feasible? And probably most important from the manuscript, we see that there were patients who died as a result of these complications. Which patients do we take straight to the operating room to try to salvage? Dr Farnell: The first question was, what are the indications for diagnostic ERCP in the year 2006? In our opinion, ERCP remains an important diagnostic tool. There are patients in whom there may be a suggestion of pancreatic malignancy on abdominal CT, magnetic resonance imaging, MRCP [magnetic resonance cholangiopancreatography], or endoscopic ultrasonography in whom the diagnosis remains unclear. A complementary or confirmatory ERCP in such patients may provide enough diagnostic information to achieve a clinical diagnosis and to plan operative therapy. In patients who have pancreatitis of unknown etiology, an anomaly of the entry of the bile duct into the pancreatic duct may be present. While MRCP may be useful in such patients, in our hands an ERCP may be critically important in such patients with pancreaticobiliary malunion or a choledochal cyst. In patients with disconnected duct syndrome or those in whom operative intervention is being considered for chronic pancreatitis, our preference would be to perform ERCP for operative planning. In addition, sphincter of Oddi manometry can be useful in a selected subset of patients with equivocal symptoms of biliary pain. Our preference in many patients with intraductal papillary mucinous neoplasm is to get an ERCP because it gives me a road map for operative planning, particularly in those with multifocal side branch disease. While endoscopic ultrasonography is the preference of my gastroenterology colleagues, my own personal preference is to have an actual picture provided by ERCP for operative planning. Dr Sielaff also asked about the predictors of perforation. Unfortunately, because of the design of the study, we were unable to address that. Dr Fatima is very industrious but she would have had to analyze over 12 000 charts in order to determine those factors predictive of perforation. We did analyze those factors that correlated with the need for operative intervention in those patients sustaining perforation, and those data were presented by Dr Fatima. Dr Sielaff also asked about a management algorithm. While the care of each patient should be individualized, there are some generalizations that can be made. It is important to talk to your endoscopist and ask whether he or she noticed the complication at the time of the procedure or not. Was the perforation guidewire related? Was it a perforation of the bile duct? If it was a guidewire perforation of the bile duct, it is highly likely that it can be managed nonoperatively. A sphincterotomy perforation is a complication that can go either way. In such a patient, CT to determine if there is objective evidence of perforation and careful clinical monitoring of the patient to determine if toxicity is present would help to dictate the need for operative intervention. Last, those patients who have altered anatomy, such as prior gastrectomy and Billroth II gastroenterostomy, those patients with peri-Vaterian diverticula, and those in whom scope passage resulted in a gaping perforation visible to the endoscopist or associated with gross extravasation on contrast examination represent a group of patients that will probably need to be operated on straight away. In this subset of patients, we would not wait for them to deteriorate clinically before making a recommendation to intervene operatively. Richard C. Thirlby, MD, Seattle, Wash: I have 1 comment and a question. I think I would quibble just a little bit with 1 of your conclusions, which I might paraphrase as “the majority of duodenal perforation should be operated on.” At the Mayo Clinic, the majority of these were operated on; however, there is not a good nonoperated-on cohort to compare to the operated-on cohort. Please expand a little more on how the surgeon decides who gets operated on and who doesn’t. You said a CT scan is essential. Would you require extravasation of contrast? Do you know how many patients were attempted to be managed medically who failed and subsequently went to operation? That might be helpful information. Dr Farnell: Dr Thirlby is making the point that judgment may be appropriate in patients who have sphincterotomy perforation. The patients that we have recommended be operated on immediately are those in whom passage of the endoscope results in an intraperitoneal gaping perforation. For patients with biliary sphincterotomy-induced perforations, we would agree with you that the approach to these patients should be individualized. If there is evidence of a perforation after endoscopic sphincterotomy, both the clinical status of the patient and imaging are important. If the patient has a negative abdominal examination, does not have signs of systemic toxicity, and has minimal or no extravasation from the site of the sphincterotomy on abdominal CT with oral contrast, antibiotic therapy, fluid resuscitation, and careful observation would be appropriate. In contrast, if there is gross extravasation into the peritoneal cavity or retroperitoneum and if there is evidence of toxicity, operative intervention is our recommendation. Our endoscopists do have some experience using hemoclips to manage small endoscopic perforations; however, their use was so selective that I am unable to make generalizations about their use based on our data. Of 11 patients experiencing perforation as a result of sphincterotomy, 7 were managed nonoperatively and 4 were managed operatively. Roger G. Keith, MD, Saskatoon, Saskatchewan: As an ERCPist, I think one of the problems that we encounter, and I think it is debated and you have mentioned it, is the presence of peri-Vaterian diverticula. There are numbers of interventional endoscopists who don't consider that a contraindication to sphincterotomy, and there are those who do. I wondered if you had in your review any evidence of patients who you had to operate on who had been perforated through the diverticulum? The second question is, more and more the interventional endoscopist is deploying metal stents for distal strictures, and perforation with a metallic stent, I think, creates a problem that one cannot avoid operating on. Perhaps you could comment on that. Dr Farnell: Indeed, there were patients who underwent sphincterotomy in spite of the presence of a peri-Vaterian diverticulum. Our experienced interventional endoscopists are willing to take on such patients for both diagnostic and therapeutic procedures selectively. In the majority of these patients, the procedures are performed safely. There were 14 patients in whom perforation occurred in the presence of a peri-Vaterian diverticulum. Seven were managed nonoperatively and 7 operatively. Of the 7 who were managed operatively, the perforation was at the site of the peri-Vaterian diverticulum in 5. There were 7 patients in whom stent insertion resulted in perforation. Five were managed nonoperatively and 2 operatively. In none of the 7 patients were the stents of the expanding metal variety. Financial Disclosure: None reported.

Chagpar, Anees B.;Scoggins, Charles R.;Martin, Robert C. G.;Carlson, David J.;Laidley, Alison L.;El-Eid, Souzan E.;McGlothin, Terre Q.;McMasters, Kelly M.

doi: 10.1001/archsurg.142.5.456pmid: 17515487

Abstract Hypothesis It has recently been proposed that only 3 sentinel lymph nodes (SLNs) are required for an adequate SLN biopsy. Others have advocated removing all nodes that are blue, hot, at the end of a blue lymphatic channel, or palpably suspicious or that have radioactive counts of 10% or greater of the most radioactive SLN. Our objective was to determine the false-negative rate (FNR) associated with limiting SLN biopsy to 3 nodes. Design Multicenter prospective study. Setting Both academic and private practice. Patients A total of 4131 patients underwent SLN biopsy followed by completion axillary node dissection. Main Outcome Measure The FNR associated with 3-node SLN biopsy. Results Of the 4131 patients in this study, an SLN was identified in 3882 (94.0%). The median number of SLNs identified was 2; more than 3 SLNs were removed in 738 patients (17.9%). Of the patients in whom a SLN was identified, 1358 (35.0%) were node positive. The overall FNR in this study was 7.7%. In 89.7% of node-positive patients, a positive SLN was found in the first 3 SLNs removed. If SLN biopsy had been limited to the first 3 nodes, the FNR would be 10.3% (P = .005 compared with removing >3 SLNs). The FNR increased with the strategy of limiting SLN biopsy to fewer SLNs (P<.001). Conclusion Removing only 3 SLNs cannot be recommended, because it is associated with a substantially increased FNR. Sentinel lymph node (SLN) biopsy is a well-accepted, minimally invasive technique that has been shown to accurately stage the axilla in patients with breast cancer. In this technique, a radioactive colloid and/or a blue dye is used to identify the first draining lymph node(s) in the axilla, which are subsequently removed. It has been recommended that all “hot” (or radioactive) nodes, all blue nodes, all nodes at the end of a blue lymphatic channel, all nodes with radioactive counts greater than 10% of the hottest node, and all palpably suspicious nodes be removed as sentinel nodes, the status of which dictates the need for further axillary dissection.1,2 On average, 2 to 3 SLNs are removed. However, on occasion, multiple SLNs can be identified. Although it is known that the false-negative rate (FNR) of SLN biopsy is greater in patients in whom only 1 SLN is removed,3 controversy exists regarding how many SLNs are sufficient for accurate staging of the axilla.4-7 Recently, it has been proposed that the SLN biopsy procedure should be terminated after 3 SLNs are identified.8 Adoption of such a policy, it is argued, will reduce the cost of the procedure by lessening operative time and decreasing the costs of pathologic analysis.8 If no benefit is to be gained by removing more than 3 SLNs, this policy certainly would be advantageous. The objective of this study, therefore, was to determine the FNR associated with limiting SLN biopsy to 3 SLNs. We hypothesized that removal of only 3 SLNs would be associated with a higher FNR than removal of all SLNs identified. Methods The University of Louisville Breast Sentinel Lymph Node Study is a multi-institutional prospective study in which patients with clinical stage I or II breast cancer underwent SLN biopsy followed by completion axillary lymph node dissection. Three hundred thirty-six surgeons from Canada and the United States participated in this study. The study was approved by the institutional review board at each site, and the patients signed informed consent forms before their participation. The technique of SLN biopsy used was left to the discretion of each surgeon. Sentinel node biopsy was not limited to a certain number of nodes in this study. To evaluate the FNR that would be expected if the procedure was terminated after finding 3 SLNs, the order in which SLNs were identified and the pathologic findings of each SLN were recorded. The FNR was defined as the percentage of node-positive patients in whom the results of the SLN biopsy were negative. Comparisons of the FNR between limiting SLN biopsy to 3 nodes or not were performed using the Fisher exact test, and the overall influence of the number of SLNs on the FNR was evaluated using the Mann-Whitney U test. All statistical analyses were performed using SPSS statistical software, version 13.0 (SPSS Inc, Chicago, Ill), with significance set at P = .05. Results From May 7, 1998, to August 2, 2004, 4131 patients were enrolled in this study. The median patient age was 60 years (range, 27-100 years), and the median tumor size was 1.5 cm (range, 0.1-11.0 cm). A sentinel node was identified in 3882 patients (94.0%). A median of 2 SLNs were removed (range, 1-18), with more than 3 nodes removed in 738 patients (17.9%). Of the patients in whom more than 3 SLNs were removed, the median number of SLNs removed was 5. Of the patients in whom a SLN was identified, 1358 (35.0%) were node positive on final pathologic analysis. These node-positive patients formed the cohort of interest for this study. The clinicopathologic features of this cohort are given in Table 1. The distribution of node-positive cases according to the number of SLNs removed is given in Table 2. A median of 13 nodes were removed after completion axillary dissection (range, 3-45), with a median of 2 positive nodes on final pathologic analysis (range, 1-28). Overall, 105 node-positive patients had a negative SLN biopsy result, yielding a FNR of 7.7%. The frequency of having the first sign of metastasis in the nth sentinel node and the cumulative true-positive rate for a given number of SLNs removed are given in Table 3. All of the SLN metastases were identified when 11 SLNs were removed. In this study, surgeons did not terminate the SLN biopsy procedure after any given number of SLNs were removed. However, the FNR that would have been realized if the number of SLNs removed was limited is given in Table 4. Overall, the FNR decreased with the increasing number of SLNs removed (P<.001). In 89.7% of node-positive patients, a positive SLN biopsy result was found in the first 3 SLNs removed. If SLN biopsy had been limited to the first 3 nodes, a positive SLN biopsy result would have been missed in 140 patients, yielding a FNR of 10.3% (P = .005 compared with removing >3 SLNs). In these patients, more than 1 positive node was found in 49 patients (35.0%). In addition, the SLN metastasis was found using hematoxylin-eosin staining in 50 (35.7%) of these cases. Comment Sentinel node biopsy has become a cornerstone of breast cancer management and has been shown to accurately stage the axilla in patients with breast cancer. Although the median number of SLNs identified is 2, more than 3 SLNs are found in 17.9% of cases. The significance of these latter SLNs has been questioned in recent studies.4-8 Sabel et al8 recently argued that SLN biopsy procedures should be terminated after finding 3 SLNs. In their study of 729 patients who underwent an SLN biopsy for breast cancer, the median number of SLNs removed was 2.5 (range, 1-9).8 More than 3 SLNs were removed in 40.7% of patients. Of the 133 node-positive patients in their study, metastatic disease was identified within the first 3 SLNs removed.8 Similarly, Schrenk et al6 found in a study of 263 patients with a mean of 1.8 SLNs removed (range, 1-5) that SLN metastases were found within the first 3 SLNs removed in all of the 105 node-positive patients. In addition, Low and Littlejohn,4 in a study of 113 patients with a mean of 1.9 SLNs removed (range, 0-6), found that all of the 33 node-positive patients were identified with the first 3 SLNs. In our larger study of 4131 patients, we found that 10.3% of the 1358 node-positive patients would have lymph node metastases that would have been missed if the procedure was terminated at 3 SLNs vs the overall FNR of 7.7% when no limit is placed on the number of SLNs removed. Therefore, 2.6% of the SLN-positive patients had their first sign of metastasis in their fourth or higher SLN. These data are similar to the Memorial Sloan-Kettering experience, which found in a study of 1561 patients that 2% of the 449 sentinel node–positive patients had their first sign of metastasis at their fourth or greater SLN.5 They found that 100% of the SLN metastases could only be found when 13 SLNs were removed.5 In addition, in a study of 720 patients, Woznick et al7 found 3% of the 172 SLN-positive patients had their first sign of metastasis in their fourth or higher SLN. In their study, however, only 6 nodes were required to identify all of the SLN metastases.7 Although the 2% to 3% rate of finding SLN metastases in higher-order SLNs may seem trivial, it must be understood that this is an incremental rate. In other words, limiting SLN biopsy to a particular threshold would increase the intrinsic FNR of SLN biopsy by 2% to 3%. In this study, in which many surgeons had little prior experience with SLN biopsy, the FNR was 7.7%. Other studies9-20 have found FNRs ranging from 0% to 29%, with an average of 8.4%. Given the current American Society of Clinical Oncology's recommendation that the FNR of SLN biopsy should be less than 5%,20 it seems impractical to limit SLN biopsy to a given number of SLNs, thereby increasing the FNR. However, as previously reported, the FNR declines to below 5% with surgeon experience of greater than 20 SLN cases.21 The rationale for suggestions to limit the number of SLNs has been to contain costs.8 However, because most patients will have 3 or fewer nodes identified, cost savings will be limited to few patients. Some, however, have suggested that costs can be contained by performing “focused” pathologic analyses only on the first 2 to 3 nodes and submitting the remaining SLNs for routine examination.22,23 In a study of 662 SLN biopsy procedures, Dabbs and Johnson22 found that in all of the patients in whom the first sign of metastasis was in the fourth or higher SLN, the micrometastasis was found by immunohistochemical analysis alone. Zervos et al,23 in their study of 509 SLN biopsy procedures, similarly found that metastatic disease detected by histologic criteria was always found within the first 2 SLNs. Interestingly, in our study, we found that more than 35% of patients whose first sign of metastasis was found in their fourth or higher SLN had disease detected by hematoxylin-eosin staining, suggesting that in at least a third of these patients the nodal metastases have clear prognostic significance. There may be concern that as many as 11 SLNs were removed in this study, which is nearly equivalent to the number of nodes removed at axillary dissection. However, the median number of SLNs removed was 2 in this study, and only 7.3% of patients had more than 5 SLNs removed. Clearly, removal of 11 SLNs is an outlier and an infrequent event possibly related to surgeon inexperience with the technique. Review of the data in Table 3 suggests that the incremental benefit of removing more than 5 SLNs is small. In conclusion, most patients will have 3 or fewer SLNs identified. However, if more than 3 SLNs are identified, these SLNs should be removed because there is a significantly higher FNR associated with limiting SLN biopsy procedures to 3 SLNs. It is not clear that the cost savings associated with restricting SLN biopsy procedures to 3 SLNs is worth the incremental FNR associated with this approach. Back to top Article Information Correspondence: Anees B. Chagpar, MD, MSc, Department of Surgery, University of Louisville, 315 E Broadway, Suite 312, Louisville, KY ([email protected]). Accepted for Publication: January 3, 2006. Author Contributions:Study concept and design: Chagpar, Scoggins, and McMasters. Acquisition of data: Chagpar, Martin, Carlson, Laidley, El-Eid, McGlothin, and McMasters. Analysis and interpretation of data: Chagpar. Drafting of the manuscript: Chagpar. Critical revision of the manuscript for important intellectual content: Chagpar, Scoggins, Martin, Carlson, Laidley, El-Eid, McGlothin, and McMasters. Statistical analysis: Chagpar. Administrative, technical, and material support: El-Eid, McGlothin, and McMasters. Study supervision: Scoggins and McMasters. Financial Disclosure: None reported. Previous Presentation: Presented at the 114th Annual Meeting of the Western Surgical Association; November 13, 2006; Los Cabos, Mexico. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript. Acknowledgment: We thank the University of Louisville, James Graham Brown Cancer Center, and Norton Healthcare, Center of Advanced Surgical Technology, for their support. References 1. Cox CEPendas SCox JM et al. Guidelines for sentinel node biopsy and lymphatic mapping of patients with breast cancer. Ann Surg 1998;227645- 651PubMedGoogle ScholarCrossref 2. Martin RCEdwards MJWong SL et al. Practical guidelines for optimal gamma probe detection of sentinel lymph nodes in breast cancer: results of a multi-institutional study. Surgery 2000;128139- 144PubMedGoogle ScholarCrossref 3. Wong SLEdwards MJChao C et al. Sentinel lymph node biopsy for breast cancer: impact of the number of sentinel nodes removed on the false-negative rate. J Am Coll Surg 2001;192684- 689PubMedGoogle ScholarCrossref 4. Low KSLittlejohn DR Optimal number of sentinel nodes after intradermal injection isotope and blue dye. ANZ J Surg 2006;76472- 475PubMedGoogle ScholarCrossref 5. McCarter MDYeung HFey JBorgen PICody HS III The breast cancer patient with multiple sentinel nodes: when to stop? J Am Coll Surg 2001;192692- 697PubMedGoogle ScholarCrossref 6. Schrenk PRehberger WShamiyeh AWayand W Sentinel node biopsy for breast cancer: does the number of sentinel nodes removed have an impact on the accuracy of finding a positive node? J Surg Oncol 2002;80130- 136PubMedGoogle ScholarCrossref 7. Woznick AFranco MBendick PBenitez PR Sentinel lymph node dissection for breast cancer: how many nodes are enough and which technique is optimal? Am J Surg 2006;191330- 333PubMedGoogle ScholarCrossref 8. Sabel MSKleer CGDiehl KMCimmino VMChang AENewman LA How many sentinel nodes should be removed in breast cancer? [abstract]. Ann Surg Oncol 2006;1327Google Scholar 9. Knauer MKonstantiniuk PHaid A et al. Multicentric breast cancer: a new indication for sentinel node biopsy: a multi-institutional validation study. J Clin Oncol 2006;243374- 3380PubMedGoogle ScholarCrossref 10. Goyal ANewcombe RGChhabra AMansel RE Factors affecting failed localisation and false-negative rates of sentinel node biopsy in breast cancer: results of the ALMANAC validation phase. Breast Cancer Res Treat 2006;99203- 208PubMedGoogle ScholarCrossref 11. Gui GPJoubert DJReichert R et al. Continued axillary sampling is unnecessary and provides no further information to sentinel node biopsy in staging breast cancer. Eur J Surg Oncol 2005;31707- 714PubMedGoogle ScholarCrossref 12. Kuehn TVogl FDHelms G et al. Sentinel-node biopsy for axillary staging in breast cancer: results from a large prospective German multi-institutional trial. Eur J Surg Oncol 2004;30252- 259PubMedGoogle ScholarCrossref 13. Nano MTKollias JFarshid GGill PGBochner M Clinical impact of false-negative sentinel node biopsy in primary breast cancer. Br J Surg 2002;891430- 1434PubMedGoogle ScholarCrossref 14. Bergkvist LFrisell JLiljegren GCelebioglu FDamm SThorn M Multicentre study of detection and false-negative rates in sentinel node biopsy for breast cancer. Br J Surg 2001;881644- 1648PubMedGoogle ScholarCrossref 15. Smillie THayashi ARusnak CDunlop WDonald Jvan der Westhuizen N Evaluation of feasibility and accuracy of sentinel node biopsy in early breast cancer. Am J Surg 2001;181427- 430PubMedGoogle ScholarCrossref 16. Molland JGDias MMGillett DJ Sentinel node biopsy in breast cancer: results of 103 cases. Aust N Z J Surg 2000;7098- 102PubMedGoogle ScholarCrossref 17. Hill ADTran KNAkhurst T et al. Lessons learned from 500 cases of lymphatic mapping for breast cancer. Ann Surg 1999;229528- 535PubMedGoogle ScholarCrossref 18. Snider HDowlatshahi KFan MBridger WMRayudu GOleske D Sentinel node biopsy in the staging of breast cancer. Am J Surg 1998;176305- 310PubMedGoogle ScholarCrossref 19. Harlow SPKrag DNJulian TB et al. Prerandomization surgical training for the National Surgical Adjuvant Breast and Bowel Project (NSABP) B-32 trial: a randomized phase III clinical trial to compare sentinel node resection to conventional axillary dissection in clinically node-negative breast cancer. Ann Surg 2005;24148- 54PubMedGoogle Scholar 20. Lyman GHGiuliano AESomerfield MR et al. American Society of Clinical Oncology guideline recommendations for sentinel lymph node biopsy in early-stage breast cancer. J Clin Oncol 2005;237703- 7720PubMedGoogle ScholarCrossref 21. Hutchinson JRChagpar ABScoggins CR et al. Surgeon and community factors affecting breast cancer sentinel lymph node biopsy. Am J Surg 2005;190903- 906PubMedGoogle ScholarCrossref 22. Dabbs DJJohnson R The optimal number of sentinel lymph nodes for focused pathologic examination. Breast J 2004;10186- 189PubMedGoogle ScholarCrossref 23. Zervos EEBadgwell BDAbdessalam SF et al. Selective analysis of the sentinel node in breast cancer. Am J Surg 2001;182372- 376PubMedGoogle ScholarCrossref Baiba J. Grube, MD, New Haven, Conn: In 1991, Dr Morton's group at the John Wayne Cancer Institute presented the scientific studies that confirmed the concept of sentinel node in an animal model. The feasibility of lymphatic mapping for breast disease was then first tested by Giuliano and colleagues in 1991 at the John Wayne Cancer Institute. The proof of principle was demonstrated by complete histopathologic analysis of all axillary nodes by Giuliano and colleagues in 1997. These results have been substantiated by others and have lead to the acceptance of intraoperative lymphatic mapping as an accurate and less invasive option for staging of the axilla in women with early breast cancer. The sentinel node hypothesis states that the sentinel node is the first draining lymph node from the primary tumor and that the sentinel node is the most likely node to harbor metastases if present. It is a functional, biological definition, not an operational, technical definition. The optimal number of nodes removed is really not the issue. The issue is the identification and removal of the true sentinel node. The presence of second-echelon nodes may occur because of injection technique, timing to dissection, location of tumor, surgeon experience, and patient characteristics. Identification of a second-echelon node may occur upon entry into the axillary space at the level of a secondary blue or hot node distal to the true sentinel node. Therefore, it is not unreasonable to discover that a more proximal node or deeper node is the true sentinel node. The criteria for removal of all blue nodes, nodes with blue lymphatic channels, and all radioactive nodes with a count of 10% of the hottest node as well as suspicious palpable nodes have resulted in axillary recurrence rates after sentinel node biopsy alone in the range of 0.4%. Many different techniques have been described to identify a node stained with vital dye, containing radioactivity or a combination of indicators. In the cumulative data from large series, recently reported by the ASCO [American Society of Clinical Oncology] group in 2005, with defined methodology, patient characterization, and experienced breast surgeons performing lymphatic mapping, the sentinel node identification rate is 96%, the false-negative rate is 7%, and the mean number of sentinel nodes removed is 1.92. The removal of multiple nodes, up to 11, as described in this study represents the equivalent of an axillary dissection and really defeats one of the goals of sentinel lymphadenectomy (ie, less morbidity with a minimally invasive technique), but this must be contrasted against a false-negative staging that could impair the adjuvant treatment decisions for a given patient. I am going to discuss 3 aspects of the study and ask 3 questions. The American Society of Breast Surgeons has proposed guidelines for surgeons learning the technique of sentinel node biopsy. These include the performance of 20 cases with a backup axillary lymph node dissection or the performance of sentinel node biopsy in the mentored situation in 20 cases. The identification rate should be at least 85% with a false-negative rate of 5%. The 300 surgeons who participated in the Louisville Sentinel Node Registry performed 4131 sentinel node procedures with an identification rate of 94%. Most surgeons had minimal experience prior to entry into the study. In previous studies from the Louisville Sentinel Node Registry, the predominant cluster of false-negative cases occurred within the first 10 cases in a surgeon's experience. My first question is: Did the number of sentinel nodes removed change as a function of the number of sentinel node procedures performed by each surgeon who initially required more than 4 nodes removed to accurately stage the axilla? The first major multi-institutional trial evaluating lymphatic mapping in breast disease presented by Krag and colleagues in 1998 demonstrated a wide variability in surgeons’ abilities, even in the hands of experienced breast surgeons, to accurately identify a true sentinel node and found that some individuals may never be able to learn the technique. Can the authors determine if the outliers in their study (ie, those who needed to remove more than 4 sentinel nodes to stage the axilla accurately) were the surgeons with lower rates of identifying any sentinel node? The goal of axillary staging with either axillary dissection or lymphatic mapping is to accurately stage the axilla for local control and adjuvant systemic treatment decisions and to minimize morbidity. In this cohort of surgeons, the false-negative rate even with multiple sentinel nodes removed was 7.7%, well below that recommended by the American Society of Breast Surgeons. The recent data from the NSABP B-32 [National Surgical Adjuvant Breast and Bowel Project B-32] sentinel node trial is also a reflection of community and academic surgeons using the sentinel node technique. The false-negative rate in this trial was 9.7%, but the number of nodes removed was not reported. Cox and colleagues have demonstrated that identification is higher by surgeons who perform more than 6 sentinel lymphadenectomies per month. My third question: What is the surgical volume for breast cases for each of the surgeons who removed more than 4 nodes to finally remove the true sentinel node and was there a difference in success rates for those who performed fewer cases than those who performed more cases? I would like to congratulate the authors on presenting the cumulative experience of a large number of surgeons representing academic practices, breast-only practices, and those with typical general surgery practices with limited volume of breast disease, using a variety of sentinel lymph node mapping techniques. Dr Chagpar: To begin with, it is clear that the false-negative rate associated with sentinel lymph node biopsy does improve with surgical experience. We and others have previously reported that having at least 20 cases does improve the false-negative rate. However, this is independent of the number of sentinel lymph nodes removed. Relative to the second question regarding surgeons who needed to remove more than 4 sentinel nodes having a lower rate of identifying any sentinel node, I need to be clear that this study was restricted to the patients in whom a sentinel node was identified. So those surgeons who could not identify any sentinel node clearly would not be in this cohort. Finally, as to the question of surgical volume and its impact on sentinel lymph node biopsy false-negative rates, this has also previously been reported. We did not look at surgical volume per se in determining whether surgical volume impacted whether surgeons removed more than 4 nodes or not. Anton J. Bilchik, MD, Santa Monica, Calif: It appears that the 18% of patients who had up to 5 sentinel lymph nodes removed had a mini-axillary lymph node dissection. Would you suggest to the surgeons here that perhaps if more than 1 to 2 sentinel nodes are found that the surgeon proceed with an axillary lymph node dissection? Dr Chagpar: I think that we need to be clear as to whether you mean 1 to 2 sentinel lymph nodes can't be found or whether those 1 to 2 sentinel lymph nodes are negative. Clearly, if no sentinel lymph node can be identified, then standard management would indicate that you would go ahead and do an axillary dissection. However, the question here was, if you find 1 to 2 sentinel lymph nodes and yet there is still another blue lymph node there, do you go ahead and remove it? Or do you say, “I have removed my 2 sentinel lymph nodes, I am done,” as would be suggested by Dr Sabel and his colleagues who presented at the Society of Surgical Oncology? I think that our data suggest that it behooves you to go ahead and remove all of the sentinel lymph nodes that you can identify, that is to say, any hot node, any blue node, any node at the end of a blue lymphatic channel, any node with counts greater than 10% of the hottest node, or any palpably suspicious node, as you will increase your false-negative rate by failure to remove those sentinel nodes. Theodore X. O'Connell, MD, Los Angeles, Calif: Begging the question of surgeon experience etcetera as brought up in the first discussion, the other question is whether the true-positive rate of approximately 90% with 3 lymph nodes recovered and a 92% true-positive rate with greater then 5 recovered, a difference of only 2%, is clinically significant. More important, what is the clinical impact of missing that 2%? I don't think this study or other studies have shown that there is a significant impact on survival, on local recurrence, or on the use of adjuvant therapy, especially now when the vast majority of patients get adjuvant therapy even with negative nodes. So what is the benefit to the patient of taking out more nodes vis-à-vis the negative impact as far as operating time and morbidity is concerned? Dr Chagpar: An excellent question. I think that to begin with we have to realize that this study did not address survival or local recurrence. However, I think that what our study did do is it demonstrated that in those patients in whom the fourth sentinel lymph node would be positive or the fifth sentinel lymph node would be positive when the first 3 were negative, remember that there was at least 1 positive node and more than 1 positive node in 35% of those cases. So one has to ask the question that if there is going to be more than 1 positive node left in the axilla, it then behooves us to remove that. It may in fact impact adjutant therapy, although I agree with you that we tend to give chemotherapy to nearly everyone these days. But remember that postmastectomy radiation, for example, is dictated by the number of positive nodes, so patients who have 4 or more positive nodes will have postmastectomy radiation therapy, and the absolute number of nodes that are positive now affects staging per the new AJCC [American Joint Committee on Cancer] staging system. Financial Disclosure: None reported.

Brasel, Karen J.;Lim, Hyun J.;Nirula, Ram;Weigelt, John A.

doi: 10.1001/archsurg.142.5.461pmid: 17515488

Abstract Hypothesis Although demographic and clinical information are known to affect hospital length of stay (LOS), we hypothesized that LOS after traumatic injury would be significantly influenced by nonclinical factors. Design Retrospective database analysis. Patients Trauma patients treated at hospitals participating in data submission to the National Trauma Data Bank. Methods The National Trauma Data Bank was queried for all patients older than 18 years with an LOS longer than 48 hours and complete demographic information including age, sex, and race/ethnicity; nonclinical factors including payment type (commercial, Medicaid, Medicare, uninsured, and other) and discharge destination (home, rehabilitation facility, nursing home, and other); and clinical information (body region injured, Injury Severity Score, and Revised Trauma Score). Statistical analysis was performed using generalized linear modeling adjusted for multiple comparisons. Main Outcome Measures Length of stay greater than the mean. Results We obtained 313 144 medical records. Mean LOS was 9.6 days. Discharge destination had the greatest effect on LOS. Mean LOS for patients with Medicaid (11.3 days) was significantly longer than for patients with commercial insurance and uninsured patients (each 9.3 days) and patients with Medicare (8.8 days). Length of stay was longer for patients discharged to a nursing home (14.2 days) or rehabilitation facility (11.5 days) compared with those discharged to any other facility (9.6 days). In multivariate analysis, factors significantly associated with extended LOS included age, sex, race/ethnicity, insurance status, discharge destination, and Revised Trauma Score. Conclusions Nonclinical factors significantly influence LOS. If LOS is used as a quality measure for injured patients, adjustment for these factors is necessary. Length of stay (LOS) has been suggested as a meaningful outcome measure that is a potential target for quality improvement activities.1-5 The American College of Surgeons Committee on Trauma uses LOS as an example outcome measure for a performance improvement program.6 Although the influence of many clinical factors on LOS is both intuitive and supported by data, studies from medical and elective surgical patients confirm the additional importance of nonclinical factors.7-12 One of these factors is insurance or payer status. We hypothesized that nonclinical factors, specifically payer status, would similarly affect LOS in injured patients. Methods The 2004 National Trauma Data Bank (NTDB) data set was used. Records were excluded from the analysis if they met any of the following criteria: patient younger than 18 years; LOS missing, less than 48 hours, or longer than 365 days; or LOS less than the number of days in the intensive care unit. Records were also excluded if any of the primary variables of interest were missing. Subgroup analysis investigated the effect of missing variables; risk estimates did not change significantly and, therefore, data imputation for missing variables was not performed. Descriptive statistics were used to summarize the data. Initial descriptive analyses were performed to determine whether to use parametric or nonparametric statistics. Analysis of variance was use for continuous variable and a χ2 test was used for categorical variables. The Tukey multiple pairwise comparison test was used to control the type I error rate.13,14 Before developing the regression model, the following variables were recategorized for further analysis. The variable of race/ethnicity was divided into the following 4 groups: white, black, Hispanic, and other (Asian/Pacific Islander, Native American, and Native Alaskan). The type of payment was categorized into the following 5 groups: commercial, Medicaid, Medicare, uninsured, and other. Discharge destination was grouped into the following 5 categories: home (home, home with home health assistance, jail, or psychiatric facility), rehabilitation facility, nursing home (skilled nursing facility or nursing home), another hospital, and other (unknown discharge destination or patient unable to complete treatment). Body region injured was categorized into the following 6 groups: head, face and neck, thorax, abdomen, spine, and upper or lower extremity. An extended stay was defined as an LOS greater than the mean LOS. For multivariate analysis, LOS was converted from a continuous variable to a dichotomous variable, LOS less than or equal to the mean LOS and LOS greater than the mean LOS. Because the analysis of variance showed differences in LOS by group, generalized linear models were created to examine the association of demographic, clinical, and nonclinical factors, as well as the interaction term of race/ethnicity × payer status, with an LOS greater than the mean. Nonclinical factors were defined as demographic and socioeconomic factors, including age, sex, race/ethnicity, payer status, and discharge status. Clinical factors were defined as body region injured, Injury Severity Score, and Revised Trauma Score. To identify potential predictors for the final model, the variables were first examined individually using univariate analysis. Variables were identified as significant using a .05 α level, and only these were included in a stepwise method to determine a final model. Results of the analysis of variance are given in Table 1; results of the generalized linear model analysis are given in Table 2. Intensive care unit LOS and need for admission to the intensive care unit were initially included in the models but were ultimately excluded because intensive care unit LOS is essentially a subset of total hospital LOS. In the final regression model, we examined not only the significance of the risk coefficients but also the strength of association. For each effect from the models, the odds ratio and 95% confidence interval were calculated. P<.05 was considered statistically significant.13,15 Analyses were carried out using SAS statistical software (version 9; SAS Institute, Cary, NC).14 Results The analysis included 313 144 medical records. Mean ± SD age of patients was 48.2 ± 21.1 years; 62.3% were men. Seventy-four percent of the patients were white, 17% were black, 6% were Hispanic, 1% were Asian/Pacific Islander, and 0.5% were Native American or Native Alaskan. Forty percent of the patients had commercial insurance, 11% had Medicaid, 25% had Medicare, 16% were uninsured, and 9% had other insurance. Mean ± SD Injury Severity Score was 12.3 ± 9.3. Overall survival was 97%. Discharge destination was home for 63% of patients, a rehabilitation facility for 16%, a nursing home or other long-term care facility for 13%, another hospital for 4%, and other for 4%. Mean ± SD LOS was 9.6 ± 12.8 days. Patients with Medicaid insurance had the longest stay (mean ± SD), 11.3 ± 15.8 days, while those with Medicare insurance had the shortest stay, 8.8 ± 11.3 days (Table 3). However, there was a significant interaction effect between race/ethnicity × payer status on LOS (Table 2). In the multivariate generalized linear models, 8 factors were associated with discharge destination (Table 2): sex, race/ethnicity, age, payment type, body region injured, discharge destination, Injury Severity Score, emergency department Revised Trauma Score, and the interaction term race/ethnicity × payer status. Discharge destination had the greatest association with extended LOS. Compared with patients discharged to home, the odds ratio for an extended LOS for patients discharged to another hospital, rehabilitation facility, or nursing home was 2.23, 3.74, and 3.83, respectively. Payer status had the next greatest effect; Medicare patients were significantly associated with an LOS less than the mean compared with patients with commercial insurance (odds ratio, 0.77; 95% confidence interval, 0.73-0.82). Medicaid patients and those designated as self-payers were associated with an extended LOS (Table 2). Patients with head, face or neck, and thorax injuries had stays shorter then the mean, and those with abdominal and spine injuries had stays longer then the mean, However, the magnitude of these associations, as well as the associations of sex, age, Injury Severity Score, and Revised Trauma Score, was relatively small compared with discharge destination and payer type status. Comment Quality medical care has become the primary focus in US health care since the 2000 report from the Institute of Medicine.16 Organizationally, performance improvement is the cornerstone on which advances and increases in quality are based. Building on the foundations established by Shewhart17 and Deming,18 Donabedian19 enumerated the 3 essential components necessary to measure the quality of health care—structure, process, and outcome. Structure is simply the infrastructure of the health care system, which includes the individuals, the equipment, and the physical plant. Process involves the various steps involved in the delivery of health care. Examples include the use of perioperative β-blockers, the timeliness of perioperative antibiotic administration, and the presence of an attending physician at a trauma resuscitation. Outcomes are, simply put, results. Some are relatively easy to measure, including mortality, perioperative myocardial infarction, and surgical site infection. Others, including pain, function, and quality of life, are more difficult to measure. Many current quality improvement efforts focus on measuring structure and process because these are usually easier to measure than outcomes.20,21 The relationship of structure and process to outcomes can be direct or indirect. However, an assumption that cuts through most performance improvement or quality programs is that attention to structure and process will result in better outcomes. This relationship is best illustrated with process measures. Compliance with evidence-based guidelines is a process measure. For example, adherence to a guideline for management of patients with splenic injuries improved outcome, measured as splenic salvage rate.22 Process measures are also used in the Institute for Healthcare Improvement 100 000 Lives initiative23 in an effort to reduce catheter-related bloodstream infection, ventilator-associated pneumonia, surgical site infection, in-hospital cardiopulmonary arrest, death after myocardial infarction, and adverse drug events. Early results suggest that compliance with the identified process measures improves the targeted outcomes.23,24 Other studies have not confirmed that standardized process implementation will have significant effect on outcomes.25 While process measures may correlate directly with desirable outcomes, it is not clear that all process or outcome measures will monitor or measure what is intended to be measured.24-26 Length of stay is a commonly used outcome measure. University HealthSystem Consortium (UHC), National Surgical Quality Improvement Program (NSQIP), and the NTDB all provide reports to participating hospitals providing LOS performance data indexed to some benchmark. This comparison between hospitals is used as one measure of the quality of care provided. It is suggested that if LOS decreases, care has become more efficient and more effective, because patients with extended LOS often consume substantial hospital resources. The relationship of LOS, effectiveness, and efficiency may not always be direct. Our data from the NTDB involving hospitalized injured patients suggest that nonclinical factors significantly affect LOS. Based on these data, it is tempting to suggest that LOS should not be used as a quality indicator. Rather, we suggest these data bring into question whether LOS should be considered a process measure or an outcome measure. Length of stay should not be used as a comparison between trauma centers or as a benchmark outcome measure unless it is appropriately adjusted. A simple comparison using injury severity will be inadequate. Although injury severity was significantly associated with extended LOS, other factors had much stronger associations. Clearly, age, physiologic status, and body region injured are also important. Adjustment for clinical factors, or risk adjustment, is also inadequate. Clinical factors alone may capture only 27% of the variation in extended LOS.1 Most risk-adjustment models do not include adjustment for the important nonclinical factors that affect LOS. As a process measure, LOS would be used differently. In hospitalized injured patients, LOS in an acute care hospital is clearly affected by the need for rehabilitation or long-term care. The current data demonstrate that discharge destination had the strongest association with extended LOS, with odds ratios ranging from 1.15 to 3.83 for discharge destinations other than home. While acute care is provided without concern for insurance status in most cases, rehabilitation or placement in a long-term acute care facility (LTACF) is done only after an assessment of payment has been made. Placement of patients without adequate and appropriate insurance requires substantial personnel time and effort and may often depend on a very few beds in LTACFs in any given community. In some communities, none of these beds may be available. These realities of rehabilitation and skilled subacute care needs suggest that using LOS as a process measure may be helpful to a trauma system of care. Length of stay could be part of assessing rehabilitation and long-term care similarly to assessing acute care. Transfer agreements for rehabilitation and LTACF care could be incorporated into a trauma center's overall care plan. Length of stay could become a monitor for how well a community and a trauma system is identifying and managing available LTACF beds in the community. As with other aspects of a well-run trauma system, this approach should be inclusive and enable postacute care for all patients including the uninsured or underinsured. Continued extended LOS might indicate an unmet LTACF bed need, which would need to be addressed at the system level. Although much of the influence of discharge destination on extended LOS may be because of insurance status, payment type has an independent influence in the multivariate model second only to discharge status. Some of this is moderated by race/ethnicity, inasmuch as interaction terms between groups were significant. The influence of race/ethnicity and payment type on other outcome measures of resource use has been noted by others.27-29 Liu et al27 found that Medicaid patients, blacks, Hispanics, and Asians were less likely to receive complex surgical care at high-volume hospitals. Other reasons for an extended LOS may relate to its surrogate for socioeconomic status, including nutritional status, social support networks, and incidence of posttraumatic stress disorder. Using LOS as a process measure rather than a benchmark outcome measure enables an individual trauma center to investigate and address all of these potential reasons. This study has limitations. The NTDB is a nonsystematic sample of hospitalized injured patients using data submitted voluntarily by participating hospitals, and population-based estimates or conclusions are impossible. However, the large number of hospitals from different regions of the United States that do contribute allow many possible regional influences to be overcome. Different hospitals may have different inclusion criteria for patient inclusion in a trauma registry. Because we included all injuries in patients surviving more than 48 hours, this difference in inclusion criteria should have a minimal effect. Comorbid conditions, clearly shown to affect LOS in hospitalized injured patients, were not considered in the model because of the large amount of missing data and quality of this field in the version used. However, the large cross-sectional sample suggests that the results reflect general associations with extended LOS. Definitions of excessive, increased, or extended LOS are varied, and there is no acceptable gold standard.1,30 Length of stay greater than the mean or median has been used by others31,32 and is appropriate in this large database containing patients with many different injury diagnoses. Attribution of inappropriate hospital days requires either prospective determination of clinical need or comparison of carefully selected similar patients with similar diagnoses, in addition to similar clinical and nonclinical factors that affect LOS. Discharge destination and insurance status are major factors in increased LOS in the injured patient. This confirms associations found by others, primarily in small, single-institution studies.28,29,33 Once this relationship is established, it is difficult to look at unadjusted LOS as an outcome measure evaluating quality of hospital-based medical care. Approximately one fourth to one third of the LOS may be associated with nonmedical factors.28 However, many organizations using LOS as a quality benchmark do not adjust for these important nonclinical factors.3,6,34 As an outcome measure, LOS must be adjusted for both clinical and nonclinical factors. While unadjusted LOS may not be a valid outcome measure, it may have value as a process measure for many trauma programs. Correspondence: Karen J. Brasel, MD, MPH, Department of Surgery, Medical College of Wisconsin, 9200 W Wisconsin Ave, Milwaukee, WI 53226 ([email protected]). Accepted for Publication: December 31, 2006. Author Contributions:Study concept and design: Brasel, Lim, Nirula, and Weigelt. Acquisition of data: Brasel. Analysis and interpretation of data: Brasel, Lim, and Nirula. Drafting of the manuscript: Brasel, Lim, and Nirula. Critical revision of the manuscript for important intellectual content: Brasel, Lim, Nirula, and Weigelt. Statistical analysis: Lim and Nirula. Administrative, technical, and material support: Brasel and Weigelt. Study supervision: Brasel and Weigelt. Financial Disclosure: None reported. Previous Presentation: This study was presented at the 114th Scientific Session of the Western Surgical Association, November 15, 2006; Los Cabos, Mexico; and is published after peer review and revision. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript. References 1. O’Keefe GEJurkovich GJMaier RV Defining excess resource utilization and identifying associated factors for trauma victims. J Trauma 1999;46473- 478PubMedGoogle ScholarCrossref 2. The best medical evidence for the best care management. Milliman Care Guidelines Web site. http://www.mnr.com. Accessed February 14, 2007Google Scholar 3. Wisconsin Collaborative for Healthcare Quality Web site. http://www.wchq.org. Accessed February 14, 2007 4. Englert JDavis KMKoch KE Using clinical practice analysis to improve care. Jt Comm J Qual Improv 2001;27291- 301PubMedGoogle Scholar 5. Guru VAnderson GMFremes SE et al. Canadian CABG Surgery Quality Indicator Consensus Panel, The identification and development of Canadian coronary artery bypass graft surgery quality indicators. J Thorac Cardiovasc Surg 2005;1301257PubMedGoogle ScholarCrossref 6. American College of Surgeons Committee on Trauma, Resources for Optimal Care of the Injured Patient. Chicago, Ill American College of Surgeons1999; 7. Khaliq AABroyles RWRoberton M The use of hospital care: do insurance status, prospective payment, and the unit of payments make a difference? J Health Hum Serv Adm 2003;25471- 496PubMedGoogle Scholar 8. Schnitzler MALambert DLMundy LMWoodward RS Variations in healthcare measures by insurance status for patients receiving ventilator support. Clin Perform Qual Health Care 1998;617- 22PubMedGoogle Scholar 9. MacKenzie EJMorris JA JrEdelstein SL Effect of pre-existing disease on length of hospital stay in trauma patients. J Trauma 1989;29757- 765PubMedGoogle ScholarCrossref 10. McAleese POdling-Smee W The effect of complications on length of stay. Ann Surg 1994;220740- 744PubMedGoogle ScholarCrossref 11. Case CJohantgen MSteiner C Outpatient mastectomy: clinical, payer, and geographic influences. Health Serv Res 2001;36869- 884PubMedGoogle Scholar 12. Kagan SHChalian AAGoldberg AN et al. Impact of age on clinical care pathway length of stay after complex head and neck resection. Head Neck 2002;24545- 548PubMedGoogle ScholarCrossref 13. Fahrmeir LTutz G Multivariate Statistical Modeling Based on Generalized Linear Models. New York, NY Springer Publishing Co2001; 14. SAS/STAT User's Guide, Version 9.1 Cary, NC SAS Institute Inc2002; 15. Nelder JWedderburn RW Generalized linear models. J R Stat Soc Ser A 1972;135370- 384Google ScholarCrossref 16. Institute of Medicine, To Err Is Human: Building a Safer Health System. Washington, DC National Academy Press2003;132- 154 17. Shewhart WA The application of statistics as an aid in maintaining quality of a manufactured product. J Am Stat Assoc 1925;20546- 548Google ScholarCrossref 18. Deming WE Quality, Productivity and Competitive Position. Cambridge MIT Center for Advanced Engineering Study1982; 19. Donabedian A Quality assessment and assurance: unity of purpose, diversity of means. Inquiry 1988;25173- 192PubMedGoogle Scholar 20. Crombie IKDavies HT Beyond health outcomes: the advantages of measuring process. J Eval Clin Pract 1998;431- 38PubMedGoogle ScholarCrossref 21. Brook RHMcGlynn EACleary PD Quality of health care, part 2: measuring quality of care. N Engl J Med 1996;335966- 970PubMedGoogle ScholarCrossref 22. Brasel KJWeigelt JAChristians KKSomberg LB The value of process measures in evaluating an evidence-based guideline. Surgery 2003;134605- 612PubMedGoogle ScholarCrossref 23. A resource from the Institute for Healthcare Improvement. Institute for Healthcare Improvement Web site. http://www.ihi.org. Accessed February 14, 2007Google Scholar 24. Spertus JARadford MJEvery NR et al. Acute Myocardial Infarction Working Group of the American Heart Association/American College of Cardiology First Scientific Forum on Quality of Care and Outcomes Research in Cardiovascular Disease and Stroke, Challenges and opportunities in quantifying the quality of care for acute myocardial infarction. Circulation 2003;1071681- 1691PubMedGoogle ScholarCrossref 25. Bradley EHHerrin JElbel B et al. Hospital quality for acute myocardial infarction: correlation among process measures and relationship with short-term mortality. JAMA 2006;29672- 78PubMedGoogle ScholarCrossref 26. Rocco MVFrankenfield DLHopson SDMcClellan WM Relationship between clinical performance measures and outcomes among patients receiving long-term hemodialysis. Ann Intern Med 2006;145512- 519PubMedGoogle ScholarCrossref 27. Liu JHZingmond DSMcGory ML et al. Disparities in the utilization of high-volume hospitals for complex surgery. JAMA 2006;2961973- 1980PubMedGoogle ScholarCrossref 28. Brasel KJRasmussen JCauley CWeigelt JA Reasons for delayed discharge of trauma patients. J Surg Res 2002;107223- 226PubMedGoogle ScholarCrossref 29. Chan LDoctor JTemkin N et al. Discharge disposition from acute care after traumatic brain injury: the effect of insurance type. Arch Phys Med Rehabil 2001;821151- 1154PubMedGoogle ScholarCrossref 30. Schoetz DJBockler MRosenblatt M et al. “Ideal” length of stay after colectomy: whose ideal? Dis Colon Rectum 1997;40806- 810PubMedGoogle ScholarCrossref 31. Schwartz RJJacobs LMYaezel D Impact of pre-trauma center care on length of stay and hospital charges. J Trauma 1989;291611- 1615PubMedGoogle ScholarCrossref 32. Holloway SSarosi GKim L et al. Health-related quality of life and postoperative length of stay for patients with colorectal cancer. J Surg Res 2002;108273- 278PubMedGoogle ScholarCrossref 33. Thomas SNMcGwin G JrRue LW III The financial impact of delayed discharge at a level I trauma center. J Trauma 2005;58121- 125PubMedGoogle ScholarCrossref 34. Collins TCDaley JHenderson WHKhuri SF Risk factors for prolonged length of stay after major elective surgery. Ann Surg 1999;230251- 259PubMedGoogle ScholarCrossref R. Stephen Smith, MD, Wichita, Kan: Since the report “To Err Is Human” was published by the Institute of Medicine in 2000, there has been an appropriate emphasis on quality of care and performance improvement. Business and industry have recognized the importance of this topic through the Leapfrog Initiative. Commercial entities such as HealthGrades have attempted to assess the quality of care provided by both institutions and physicians and have made their conclusions available to the public. We have all heard recently, and frequently, the mantra of “pay for performance,” which will greatly affect our practices regardless of our geographic locations and our practice setting. Dr Brasel and her colleagues have brought forward an important issue: How is quality measured? Can a single statistic serve as a valid indicator of quality and efficiency? The American College of Surgeons Committee on Trauma suggests the use of LOS as an example outcome measure in a performance improvement program. The authors have thoughtfully and convincingly demonstrated that a seemingly simple and straightforward statistic is influenced by myriad complex factors, many of which are nonclinical and beyond the control or influence of surgeons or hospitals. Some of these factors include discharge destination, the presence or absence of family support, payer status, and the availability of rehabilitation or long-term care facilities. While extended LOS certainly demonstrates holes in our health care system, is this an appropriate measure of quality of care in a trauma center? Based on this work, the answer appears to be no. Dr Brasel and her colleagues could have entitled this paper “Lies, Damn Lies, and Statistics.” I have several questions. Many facilities that submit data to the NTDB do not include financial information such as payer type. Are you certain that your data are broadly representative of patients with injury? For example, the percentage of uninsured patients, 16% in your study, seems small for a trauma population. Were there state or regional differences in either LOS or the nonclinical factors that seemed to influence LOS? Your categories of discharge destination were somewhat broad. For example, discharge to home included the subcategories of home, home with health care assistance, psychiatric facility, and jail. While my teenage children may consider home the equivalent of incarceration, I am not sure this is true of trauma patients. Did you assess LOS for the various subcategories of discharge destination? Did mechanism of injury, for example, blunt vs penetrating trauma or an assault vs a motor vehicle crash, have a significant effect on LOS? It is my impression that victims of violent acts are much more difficult to place and, thus, have a prolonged LOS even though their injuries are essentially equivalent to those of others. Finally, based on your analysis, is LOS a meaningless parameter of quality measurement? Should organizations that purport to measure quality of care abandon LOS as an indicator of effective and efficient care? Dr Brasel: First, with respect to financial data, we did not use either costs or charges in this analysis, although that would be interesting, and all records that we included did have payer source. With respect to the uninsured, one of the problems with the NTDB is that it is not population-based, so these results, although I believe they would be relatively similar given the population-based sample, can only be generalized to the NTDB records. We did not assess regional differences. That would be interesting to do. I grouped discharge destination primarily based on my experience in that patients whom I discharge to jail or psychiatric facilities essentially must be able to take care of themselves, somewhat similar to patients whom I discharge to home. I do not think putting them in a separate group would be particularly helpful because then they would be such a small minority, less than 5% in both groups. We included mechanism of injury, blunt vs penetrating trauma, in early analyses. It did not turn out to be independently significant, although, as you might imagine, patients who are injured through blunt mechanisms are significantly different with respect to some of those nonclinical factors from patients who are injured via penetrating mechanisms. However, it becomes overwhelmed by those other nonclinical factors. I do not think we are going to be able to throw out LOS. It is so easy to measure and is such a part of the administrative culture. What I would like to see is LOS used intelligently and to work with administrators rather than having them spend inordinate amounts of personnel time and statistical time fudging the statistics to make LOS look good. Length of stay should be used as a process measure rather than an outcome measure to truly improve the quality of care that we give. David B. Hoyt, MD, Orange, Calif: Embedded in your data is this problem with access to rehabilitation. My question is: How can we take these data and turn them around and create a national forum to really increase access to rehabilitation, which is one of the frontiers of trauma systems that is totally undeveloped. Dr Brasel: I am not sure I have a short or a simple answer to that question, but I do think it is important. And I think, going back to one of Dr Smith's questions, looking at this on a regional basis may help, or looking at a smaller subset where we might have days in rehabilitation, and use LOS as a system aspect rather than just in acute hospital care. Scott R. Petersen, MD, Phoenix, Ariz: Hospital administrators are constantly looking at LOS and, subsequently, direct variable costs. Your study demonstrates that LOS is really process-related and not a quality outcome. I think it is an important concept. As a medical director of a hospital trauma service line, I am informed on a quarterly basis about our direct variable costs and where the service is in relation to our goal. Ninety percent of direct variable cost is directly related to LOS. Your article has highlighted that LOS and indirectly controllable costs are not necessarily a provider-only outcome but a process of care. I have 2 questions: Have you looked at the direct variable costs in your facility? What are you tracking now in your own facility insofar as LOS? Dr Brasel: We looked at this about 4 years ago and found that about 25% to 30% of our extended LOS on a prospective basis was completely nonmedical. And that was not even because of some of these nonclinical factors. The patients were assessed as ready to go and the physicians were ready to write the discharge orders, but the consultant had not put a final opinion on the medical record, there was some nursing issue to be done, the discharge supplies were not ready, or there was no bed available in a rehabilitation facility or a nursing facility. I would like to say we have reduced that. We have case managers who work on that actively. We have not been very successful. Mark Talamonti, MD, Chicago, Ill: Length of stay is also used as a measure of quality in surgical oncology, and is very complex in that setting. Other papers that have examined LOS in surgical oncology have tried to relate it to high-volume centers and to complexity of cases. In your database, did you look at high-volume trauma centers, different levels of trauma centers, and how that influenced LOS? Dr Brasel: We did not look at level of trauma center. There are a lot of nondesignated centers in the NTDB or centers that have chosen not to report a designation. Gregory J. Jurkovich, MD, Seattle, Wash: Two quick questions: First, do you know what the UHC uses as risk adjustment? In your first slide, you showed that the UHC did have risk-adjusted LOS. What is the UHC using as risk adjustments in a trauma population for LOS? Second, have you considered interaction between the variables used in your regression analyses? Your multivariate logistic regression shows a number of these variables as being individually important, but I cannot imagine that many of these are not related to each other. Have you looked at interaction variables for those key nonmedical social sorts of issues that relate to LOS? Dr Brasel: All of the factors that the UHC adjusts for are clinical, unlike the Hospital Efficiency Index, which includes some nonclinical adjustment. The UHC adjusts for age and comorbidity. They do not adjust for injury severity or anything specific to trauma populations. We did look at some interaction terms. The interaction of race/ethnicity and insurance is significant. We did not overwhelm the model with interaction terms, but that interaction term was significant. Donald E. Low, MD, Seattle: Many of the comparisons in which we see LOS used as the most relevant factor have to do with the evolution from open to minimally invasive and even interventional types of procedures. Your study population was a diverse trauma population. It would be equally interesting to have the same objective comparisons in patients who have a single pathologic condition. However, do you believe that any assessment of LOS is relevant if readmission rates are not included? Dr Brasel: I would say, probably not. If you are just discharging people from the hospital after a first admission and they return and are readmitted for another 4, 5, or 6 days, that probably does not influence resource consumption and may be an indicator of poorer quality of care rather than improved quality of care. You might suggest that readmission rate, as a clinical or nonclinical factor, is influential, but it goes to the point that you need to know what goes into the data and intelligently use the data when you talk with your administrators, and look at your own practice, your hospital's practice, and your system's practice. Financial Disclosure: None reported.

Gali, Bhargavi;Findlay, James Y.;Plevak, David J.;Rosen, Charles B.;Dierkhising, Ross;Nagorney, David M.

doi: 10.1001/archsurg.142.5.467pmid: 17515489

Abstract Hypothesis Perioperative events of patients undergoing living donor (LD) right hepatectomy are similar to those of patients undergoing right hepatectomy for disease (DZ). Design Institutional review board–approved retrospective case-control study. Setting Eight hundred–bed tertiary care referral center. Patients and Methods We matched 40 patients who had LD with 40 patients who had DZ. Perioperative events (anesthesia, surgical events, transfusion, hemodynamic events, complications, and length of hospital stay) were compared using the signed rank test and exact McNemar test where appropriate. Main Outcome Measures Intraoperative time, transfusion requirements, postoperative complications, and hospital length of stay. Results There was a significant difference in surgical time between the LD and DZ groups (median, 4.1 vs 3.3 hours; P = .001). There was also a significant difference in anesthesia time between the LD and DZ groups (median, 5.6 vs 4.2 hours; P<.001). The level of autologous transfusion was higher in the LD group (median, 1.3 vs 0 U in the DZ group; P<.001), and that of packed red blood cell transfusion was lower in the LD group (mean, 0 vs 0.5 U; P = .008). There was no other significant intraoperative difference. Postoperative hemoglobin levels were significantly higher in the LD group (median, 12.6 vs 11.8 g/dL; P = .03). Comparison of the number of complications in the immediate postoperative period revealed no other significant differences. Conclusions The LD procedure took longer to perform because of the time required for hilar dissection. The difference in intraoperative transfusions is attributable to use of cell salvage and retransfusion of salvaged blood for all donors; this was not routine for DZ procedures. Perioperative outcomes were similar in all other respects. The LD procedure has similar outcomes to those of the DZ procedure. Living donor (LD) liver transplantation has emerged as an excellent option for adult patients awaiting liver transplantation who are unlikely to receive a deceased donor liver. Since 1988, 2085 adult LD liver transplantations have been performed in the United States.1 Donors are otherwise healthy individuals undergoing right hepatectomy, an operation with the potential to result in major complications and even death. Thus, there are serious concerns about the donors' safety during and after LD liver procedures. We began adult-to-adult LD liver transplantation at our institution in 2000. At that time, our paramount concern was donor safety. We wanted to do everything possible to ensure that all living donors received the best possible care. Our institution has a large experience with liver resection and a sizable hepatobiliary surgery practice. To take advantage of this experience, we decided that each and every potential living liver donor would be seen by our most experienced hepatobiliary surgeon and our most experienced liver transplant surgeon, and that the LD operation would be performed by a team including surgeons and operating room and anesthesia personnel with experience in both hepatobiliary and liver transplantation procedures. Our goals were to perform the LD hepatectomy safely, to match or to exceed results with liver resection for disease (DZ), and to avoid early mistakes commonly attributed to the “learning curve.” We reviewed our 5-year experience with right hepatectomy for LD liver transplantation and compared results with those of a control group of patients who underwent DZ-related right hepatectomy. Our primary aim was to determine whether the liver LD procedure is at least as safe as DZ liver resection. Our secondary aim was to determine whether our model of care—use of a combined liver transplant and hepatobiliary surgery team—enabled us to achieve our goals to maximize donor safety and avoid the learning curve. Methods This study was performed with approval of the institutional review board of the Mayo Clinic. Data were abstracted from patient medical records and transplantation and anesthesia databases maintained on all patients undergoing transplantation. We retrospectively compared a group of patients who underwent LD with a control group of patients who underwent DZ liver resection. The LD group included the first 40 donors who underwent LD right hepatectomy at the Mayo Clinic from June 2000 through January 2005. The DZ group was compiled by matching the LD group on age (within 10 years) and sex from patients who underwent right hepatectomy for benign and malignant disease during the same period. The LD operations were performed by the same team (a hepatobiliary surgeon [D.M.N.] and a transplant surgeon [C.B.R.]). The DZ operations were performed by the same hepatobiliary surgeon (D.M.N.). Ld group All donors in the LD group underwent ultrasonography with Doppler examination and volumetric computed tomography. Selected patients underwent angiography and liver biopsy. Criteria for the LD group included (1) liver resection limited to 70% of the parenchyma, (2) anticipated right liver graft size exceeding 0.8% of recipient weight, and (3) vascular anatomy favorable for resection and transplantation. Anesthetic management for all donors included neuraxial opioids to assist with postoperative analgesia before induction of general anesthesia. General anesthesia was maintained using volatile agents and fentanyl citrate. Intermediate-acting muscle relaxants were used. All patients had a central venous catheter placed, monitoring of central venous pressure throughout the procedure, and use of a low central venous pressure technique. Direct arterial pressure monitoring was used in all cases. Cell salvage was used in all of the LD procedures. The LD operation was performed through a bilateral subcostal incision with an upper extension in the midline to the xiphoid. After abdominal exploration and cholecystectomy, a cholangiogram was obtained through injection of the cystic duct. The right hepatic artery was dissected free from the bifurcation of the proper hepatic artery to the undersurface of the liver. If replaced, the replaced right hepatic artery was dissected free from the pancreas to the undersurface of the liver. The right portal vein was dissected free at the bifurcation of the portal vein. The right hepatic duct was not divided until midway through the parenchymal division. The right side of the liver and the retrohepatic vena cava were mobilized with division of the right adrenal vein. All small inferior right hepatic and caudate venous tributaries to the vena cava were divided, encircling the right hepatic vein and any additional veins larger than 0.5 cm in diameter. The falciform ligament was divided, but the other attachments to the left liver were left intact. The right hepatic artery and right portal vein were temporarily occluded, and the line of vascular demarcation was marked on the surface of the liver. The vessels were unclamped and left open throughout the parenchymal division (no in-flow occlusion). The parenchyma was divided using 1 of 2 commercially available dissectors (CUSA Excel [Tyco Healthcare, Mansfield, Mass] or Helix Hydro-Jet [Erbe, Tübingen, Germany]), ligating or oversewing bile ducts and larger vessels. The middle hepatic vein was identified early during the division and retained with the remnant left side of the liver. The right bile duct was divided during division of the parenchyma, probing the right duct, bifurcation, left duct, and common hepatic duct to ensure avoidance of injury to the remnant biliary system. The right hepatic duct stump was oversewn or ligated. All donors were treated with heparin after complete transection of the liver. The right side of the liver was removed by ligating and dividing the right hepatic artery, stapling and dividing the right portal vein, and dividing the right hepatic vein. The donor liver was exsanguinated in the operative field to enable collection of the blood with the cell saver. Meticulous hemostasis was achieved without the use of topical hemostatic agents, and the resection interface was carefully checked for bile leakage. The falciform ligament was reconstructed, vascular patency to the remnant liver was ensured, and a drain was placed in the right liver fossa before closure. All donors in the LD group were observed in the recovery room before transfer to the intensive care unit for an overnight stay. Dz group All patients in the DZ group received general anesthesia that included volatile agents and intermediate-acting muscle relaxants. Analgesia techniques were at the discretion of the attending anesthesiologist. Most patients received long-acting narcotic agents; however, intrathecal narcotics or epidural analgesia were used in some cases. All patients had central venous pressure monitoring, and a low central venous pressure technique was used; in addition, all patients had direct arterial pressure monitoring. Cell salvage was not used in most cases. The technique for right hepatectomy (polysegmentectomy, segments 5-8) was similar for patients with benign and malignant disease in the matched control group. After a midline or bilateral subcostal incision without an apical vertical extension, the liver was mobilized and a cholecystectomy was performed without cholangiography. The right hepatic artery and right portal veins were encircled. The right hepatic artery was divided between ligatures, and the portal vein was divided with an endovascular stapler. Small hepatic veins (<5 mm) were divided, and the right hepatic vein was transected with an endovascular stapler before parenchymal transection. Parenchymal transection was performed along the line of vascular demarcation using the CUSA Excel device. Biliary and vascular structures were suture ligated on the remnant liver and clipped on the specimen liver. The middle hepatic vein was retained within the remnant liver unless invaded by malignancy. In-flow vascular occlusion was used selectively for interface bleeding during parenchymal transection. The right hepatic duct was divided during parenchymal division. Hemostasis was secured with cautery or suture ligatures. Topical hemostatic agents were not used. The remnant liver was rotated into the right subdiaphragmatic space so that the interface abutted the retroperitoneum. The falciform ligament was not reconstructed. Only patients with concerns about hemodynamic events or hemostasis were sent to the intensive care unit. Data collection and analysis Intraoperative records were reviewed for anesthesia time, surgical time, blood product transfusion, and significant pulmonary or hemodynamic events. Perioperative and postoperative information was collected, including blood product transfusion, cardiac and pulmonary complications, biliary complications, infections, and length of hospital stay. Readmission to the hospital and mortality at the time of data collection were also recorded. Complications were documented on the basis of predefined criteria. Statistical comparisons were made using the signed rank test for continuous outcomes and the exact McNemar test for categorical end points. P<.05 was considered statistically significant. Results The demographic data for the LD and DZ groups are shown in Table 1. One of the 40 patients in the LD group requested noninclusion of the demographic information during research activities before donation, so that person was excluded from the analysis along with the matched patient in the DZ hepatectomy group. Comparison of intraoperative variables (Table 2) revealed a significant difference in surgical time, with a median of 4.1 hours for the LD group vs 3.3 hours for the DZ group (P = .001). There was also a significant difference in anesthesia time, with a median of 5.6 hours for the LD group and 4.2 hours for the DZ group (P<.001). Intraoperative transfusion of autologous blood was significantly different, with medians of 1.3 U for the LD group vs 0 U for the DZ group (P<.001). Transfusion of packed red blood cells intraoperatively was significantly different, with means of 0 U for the LD group and 0.5 U for the DZ group (P = .008). Estimated blood loss data were available on 17 of 39 matched pairs and were not significantly different between the LD and DZ groups, with medians of 900 and 425 mL, respectively (P = .29). There were no significant differences in any other intraoperative variables, including transfusion of coagulation factors (fresh frozen plasma, P>.90) or platelets (P>.90), albumin infusion (P = .48), and cardiac or pulmonary events (P>.90). Postoperative hemoglobin levels were significantly higher in the LD group (median, 12.6 vs 11.8 g/dL; P = .03). Postoperative transfusion of coagulation products was similar in both groups (P = .63). The postoperative infection rate was not statistically significant (P = .38). Surgical complications are listed in Table 3. There were more biliary complications in the LD group (6 vs 1). Although the difference did not reach statistical significance (P = .12), we were concerned about these leaks and changed our operative approach after encountering them early in our experience. We switched from oversewing the right bile duct stumps to simple ligation (as is done during the DZ procedure) and have avoided leaks since that time. There were no differences in the number of postoperative pulmonary, cardiac, vascular, or infectious complications (P = .19). The maximum number of complications for a patient in either group was 2, with a median of 0. Postoperative complications by the modified Clavien classification2 are shown in Table 3. The distributions of the type and total number of complications were similar between the groups. No cardiac events (ie, arrhythmia, myocardial infarction, or cardiac arrest) were noted in either group. One patient in the LD group required reintubation postoperatively for hypopnea, and 1 patient from each group required thoracentesis for a pleural effusion. No clinically apparent problems with pulmonary atelectasis or pneumonia were noted in either group. One patient in the DZ group had an ileus and another had pancreatitis. One patient in the LD group required ultrasound-guided drainage of a hematoma. There were no episodes of sepsis, acute renal failure, deep venous thrombosis, nerve palsy, or cerebrovascular accidents noted in either group. Lengths of stay for both groups were similar, with a mean stay of 7.4 days for the LD group and 7.2 days for the DZ group (P = .21). There was no significant difference in readmission rates for perioperative complications between the 2 groups, with 4 patients in the DZ group and 5 in the LD group readmitted within 1 month of surgery (P>.90). Four patients in the DZ group died after recovery from their operations. All 4 deaths occurred at least 6 months after the operations, and all 4 patients had significant metastatic disease. There were no deaths in the LD group. Comment We found that overall morbidity, as measured by the modified Clavien classification,2 was similar between the groups of patients undergoing right hepatectomy for living liver donation and disease, with complications arising in 13 (33%) of the 39 patients in the LD group and 7 (18%) of the 39 patients in the DZ group. (One patient in each group had 2 complications.) We also found that the type of modified Clavien classification outcomes was similar. Although published complication rates for LD vary widely, a systematic review found reported rates ranged from 0% to 67%.3 More recently, Shah and colleagues4 reported a 37% complication rate, with 6.9% grade IIIa and 12.9% grade IIIb complications according to the modified Clavien classification and no grade IV complications.2,5 Bak and colleagues6 noted complications in 17% of donors, including 2.4% grade I, 4.9% grade IIIa, and 9.8% grade IIIb complications according to the modified Clavien classification (see the complication grade key in Table 3). Most of these reports assessing morbidity and mortality after liver donation are compiled without comparison with a control group. A recent review of LD outcomes noted that morbidity ranged from 0% to 100%, with a median of 16%.7 The authors of that review commented that underreporting may have occurred in some of these studies of LD, and most of the data were from case series. Nevertheless, our LD results are comparable with the overall results reported in the literature. The right side of the liver is the most frequent site of graft donated by living donors in the West and at our institution. Consequently, patients undergoing the DZ procedure constitute the patient group most comparable to living liver donors. A previous report from our group found complications in 16.2% of patients after hepatic resection,8 and our current results compare favorably with that earlier report. We conducted our review using case-matched patients who underwent the DZ procedure. We recognize and acknowledge the limitations of our control group. Clearly, morbidity may differ significantly because of differences in operative techniques, residual functioning liver volume after hepatectomy, and patient characteristics related to the concurrent diseases.9 Despite reports that patient age has not significantly affected morbidity (32.2% for those <70 years and 39.1% for those ≥70 years),10 we did attempt to match the control group by age as a surrogate marker for comorbidity, given the differences in the overall LD and DZ populations.9 We would have also preferred to use only patients with benign disease in the control group (especially because we use the cell saver for all of our patients undergoing LD operations but avoid its use for patients with malignant disease). Unfortunately, it would have been impractical to limit our control group to benign disease owing to the relative infrequency of benign vs malignant disease requiring right hepatectomy. Although the control group operations were performed by the same hepatobiliary surgeon involved who shared responsibility for the LD operations, there were major differences in the operative approaches between the 2 groups. These differences included routine use of neuroaxial analgesia and the cell saver, extensive dissection of the right hepatic vessels before parenchymal transection, and division of the right bile duct close to the common hepatic duct bifurcation for the LD operation. Likewise, division of the right hepatic vessels (ie, artery, portal vein, and hepatic vein) before parenchymal transection and selective use of in-flow occlusion were used in the DZ group but not in the LD group operations. These differences had the potential to affect our results. However, these differences would only be important had we found major differences in morbidity between the 2 groups. Thus, our control group is not perfect, but it is the best group available with whom to compare our LD results. There was a significant difference in both the operative and anesthesia times between the LD and DZ groups. We attribute these differences to the additional time required for vascular dissection and administration of neuroaxial analgesia before induction in the LD group. Most importantly, however, is that the longer operative and anesthesia times did not appear to affect the postoperative morbidity or the other intraoperative factors, such as the likelihood of significant hemodynamic events or transfusion requirements. The higher autologous transfusion rate in the LD group was due to the routine use of intraoperative blood salvage, which is recommended for LD surgery and is routine for every such case at our institution.11 Cell saver was not routinely performed in the DZ group, because most of the patients were undergoing hepatectomy for malignant disease. Use of cell salvage may also explain the comparative increase in intraoperative packed red blood cell transfusion in the DZ group and account for the higher hemoglobin level postoperatively in the LD group. Despite these differences, there was no significant difference in postoperative transfusion of red blood cells between the 2 groups. Although the difference did not reach statistical significance, the percentage of patients with biliary complications was higher in the LD group. Biliary complications are the most frequent cause for LD morbidity, arising in 0% to 38.6% (mean, 6.2%) of living donors.7 We had biliary complication rates of 15% (6 occurrences) in the LD group compared with 3% (1 occurrence) in the DZ group. Despite our goal to avoid the learning curve, we had more biliary complications in the LD group. We had decided to oversew the bile duct stumps in the LD group in contrast to our practice of simple ligation in the DZ group. Reverting back to our standard approach of simple ligation avoided further problems. Despite our better results with bile duct stump ligation, we continue to be concerned about duct identification and division during the LD operation. Protection of the LD patient's ductal system is vital, as any injury is potentially devastating to the donor's future health. Infection is the second most commonly reported complication in LD subjects, with rates of 0% to 28.6% (median, 5.8%).7 Our infection rate was 10% (4 occurrences) in the LD group and 3% (1 occurrence) in the DZ group, which did not reach statistical significance. Major respiratory complications were reported in 9.8% of donors by Dondero et al,12 including pulmonary embolism, emphysema, and bacterial pneumonia. Minor respiratory complications were noted in 7.1%. In comparison, we encountered 5% major and minor complications in our LD group (1 postoperative reintubation and 1 pleural effusion requiring thoracentesis). Fortunately, there were no deaths in our LD group and no postoperative deaths in the DZ group. All of the deaths in the DZ group were due to metastatic disease and occurred more than 6 months after the operative procedure. Length of stay was not statistically different between the groups, with a median of 7.0 days for the LD group and 6.0 days for the DZ group. These results compare favorably with those reported by others.13 One of our goals when we began to perform LD liver transplantation was to avoid complications commonly attributed to the learning curve. We took advantage of our experience in both hepatobiliary and transplantation surgery to avoid these early problems. This study attests to our avoidance of the learning curve. Indeed, there were few differences in the perioperative and postoperative courses and minimal differences in morbidity for patients undergoing right hepatectomy for living donation vs disease. Our results demonstrate that LD right hepatectomy can be accomplished safely and with morbidity limited to that associated with DZ right hepatectomy by an experienced hepatobiliary surgical team. Correspondence: Bhargavi Gali, MD, Mayo Clinic College of Medicine, 200 First St SW, Rochester, MN 55905 ([email protected]). Accepted for Publication: January 7, 2007. Author Contributions:Study concept and design: Gali, Findlay, Plevak, and Rosen. Acquisition of data: Gali. Analysis and interpretation of data: Gali, Findlay, Plevak, Rosen, and Dierkhising. Drafting of the manuscript: Gali, Findlay, Plevak, Rosen, and Nagorney. Critical revision of the manuscript for important intellectual content: Gali, Findlay, Plevak, Rosen, Dierkhising, and Nagorney. Statistical analysis: Plevak and Dierkhising. Obtained funding: Gali. Administrative, technical, and material support: Rosen. Study supervision: Findlay, Plevak, and Nagorney. Financial Disclosure: None reported. Funding/Support: This study was supported solely by institutional and/or departmental sources. Previous Presentation: This paper was presented at the 114th Annual Scientific Session of the Western Surgical Association; November 14, 2006; Los Cabos, Mexico; and is published after peer review and revision. The discussions that follow this article are based on the originally submitted manuscript and not the revised manuscript. This article was corrected on 5/21/2007, prior to publication of the correction in print. References 1. Organ Procurement and Transplantation Network (OPTN), Living donor transplants in the US by state. http://www.optn.org. Accessed October 4, 2006 2. Dindo DDemartines NClavien PA Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 2004;240205- 213PubMedGoogle ScholarCrossref 3. Beavers KLSandler RSShrestha R Donor morbidity associated with right lobectomy for living donor liver transplantation to adult recipients: a systematic review. Liver Transpl 2002;8110- 117PubMedGoogle ScholarCrossref 4. Shah SAGrant DRGreig PD et al. Analysis and outcomes of right lobe hepatectomy in 101 consecutive living donors. Am J Transplant 2005;52764- 2769PubMedGoogle ScholarCrossref 5. Clavien PASanabria JRStrasberg SM Proposed classification of complications of surgery with examples of utility in cholecystectomy. Surgery 1992;111518- 526PubMedGoogle Scholar 6. Bak TWachs MTrotter J et al. Adult-to-adult living donor liver transplantation using right-lobe grafts: results and lessons learned from a single-center experience. Liver Transpl 2001;7680- 686PubMedGoogle ScholarCrossref 7. Middleton PFDuffield MLynch SV et al. Living donor liver transplantation—adult donor outcomes: a systematic review. Liver Transpl 2006;1224- 30PubMedGoogle ScholarCrossref 8. Tsao JILoftus JPNagorney DMAdson MAIlstrup DM Trends in morbidity and mortality of hepatic resection for malignancy: a matched comparative analysis. Ann Surg 1994;220199- 205PubMedGoogle ScholarCrossref 9. Schroeder RAMarroquin CEBute BPKhuri SHenderson WGKuo PC Predictive indices of morbidity and mortality after liver resection. Ann Surg 2006;243373- 379PubMedGoogle ScholarCrossref 10. Cescon MGrazi GLDel Gaudio M et al. Outcome of right hepatectomies in patients older than 70 years. Arch Surg 2003;138547- 552PubMedGoogle ScholarCrossref 11. Lutz JTValentin-Gamazo CGorlinger KMalago MPeters J Blood-transfusion requirements and blood salvage in donors undergoing right hepatectomy for living related liver transplantation. Anesth Analg 2003;96351- 355PubMedGoogle Scholar 12. Dondero FTaille CMal H et al. Respiratory complications: a major concern after right hepatectomy in living liver donors. Transplantation 2006;81181- 186PubMedGoogle ScholarCrossref 13. Shackleton CRVierling JMNissen N et al. Morbidity in live liver donors: standards-based adverse event reporting further refined. Arch Surg 2005;140888- 896PubMedGoogle ScholarCrossref John J. Brems, MD, Burr Ridge, Ill: Drs Gali, Nagorney, and their coauthors compared the outcome and complications between right hepatectomies for disease and transplantation. When LD liver transplantation emerged as a therapeutic option for adults, it was imperative that this would be a safe operation for the donor. Therefore, I congratulate the Mayo Clinic on the wisdom of having their most experienced hepatobiliary surgeon, who does the majority of hepatectomies for disease, perform the LD operations in conjunction with their most experienced liver transplant surgeon. As the authorsmention in their manuscript, there can be no learning curve for this operation. When you compare these 2 groups, you are really comparing apples to oranges. With right hepatectomies for disease, the primary goal is to resect the tumor with negative margins. Therefore, the blood supply to the right lobe of the liver is divided early before the parenchymal dissection. In right hepatectomies for liver transplantation, the goal is to prevent ischemia to the right lobe and, consequently, the blood supply to the right lobe is preserved until after the parenchymal dissection. Therefore, these are 2 fundamentally different operations with 2 very different outcome goals. Thus, it is difficult to compare them. I have a few questions for the authors. The authors state that after a right hepatectomy for disease, the remnant liver is rotated into the right subdiaphragmatic space to abut the retroperitoneum. This is not done in their donor transplant hepatectomies. Many hepatobiliary surgeons don't do this either because of concerns that it may cause hepatic venous outflow obstruction. I was wondering why they do it in the DZ group and not the transplant group. Also, with the LD operation, you are performing 2 simultaneous operations. I assume many of their LD livers were transplanted into recipients with hepatocellular or cholangiocarcinoma. Therefore, from a logistical standpoint, how did they perform these operations? Did they start with the recipient operation to ensure that there wasn't extrahepatic disease, or did they begin both the donor and recipient operations simultaneously? Did they ever have an instance where they began the donor operation and had to abort it because of problems with the recipient operation? If so, what did they do with the orphan donor liver? Last, I noted that they had 4 wound infections in the donor group and only 1 in the DZ group. I would have expected it to be the other way around, since many patients in the DZ group had probably received chemotherapy and were less healthy than patients in the donor group. Do they have any thoughts on why this may have occurred or happened? In closing, I would like to congratulate this group on their outstanding results. They have demonstrated how you can have a successful marriage between surgical oncologists and liver transplant surgeons. And, in the end, the patient is the one who benefits the most from this. Dr Nagorney: Thank you, Dr Brems. You are very kind with your comments. First, I admit that there were clear differences in patient groups and operative techniques. You wondered about rotating the liver into the right liver fossa with a potential concern for hepatic venous outflow obstruction. We didn't let the liver rotate into the subdiaphragmatic space for living liver donors to ensure that the normal anatomic position of the remaining middle and left hepatic veins was maintained to reduce any remote chance of hepatic venous outflow obstruction. In my own clinical practice for disease, most patients are at risk for disease progression. Intervention through the right subcostal space or flank directly, especially for percutaneous ablation, is easier if the liver is rotated. Thus, I let the liver rotate to the right. If there is no engorgement, I leave it there. If there is any engorgement or venous bleeding from the liver, I will ultrasound the hepatic veins. If there is hepatic venous outflow obstruction, I return the liver remnant to its original position. You asked a very important question about living liver donors. We utilize living liver donors for patients with malignancy, predominantly sclerosing cholangitis and associated bile duct cancer. Each recipient who is a candidate for living liver donor is prestaged several days before the LD liver operation. If metastatic malignancy is confirmed at the prestaging operation, that finding precludes retrieving a right liver, aborting a donor operation, and potentially discarding that organ. However, after staging, the 2 operations do proceed concurrently. Thus, such a case may still arise. We can abort the LD liver operation anytime before removal of the graft. Technically, even if we split the liver completely, liver donor function will be preserved. If the bile ducts haven't been divided, aborting the donor operation isn't a major technical issue. Your final question related to wound infections. I don't have a good explanation for our findings. Clearly, many recipients had chemotherapy, though most had been off chemotherapy for at least several weeks. We do ensure that the nadir of their white blood cell count has occurred and is returning to normal levels. Perhaps that fact may account for the lower than expected incidence of infection in that group. Theodore X. O’Connell, MD, Los Angeles, Calif: I have 2 questions for you. First, your basic hypothesis is that LD hepatectomy should be as safe as DZ hepatectomy. Shouldn't it be safer, since the patients having it for disease are having it for their own benefit, and obviously for life-threatening problems, while the people having it as a donor are simply doing it out of largesse and not for their own benefit? I would think that it should be safer for these individuals. Second, even though the complication rate in the live liver donor is 36% and 21% in the patients with hepatectomy for disease, you say there is no statistical difference, although there is almost a doubling in complication rate. Even though it is not statistically significant, it seems to be clinically significant, and I think you have a type II error in that you really don't have enough patients in either arm to prove that this is statistically significant. Dr Nagorney: I agree. An LD liver operation should be safer than that for one of disease. All I can say is that we try to make it as safe as we can. The operation itself precludes similar controls; thus, we compared it with patients undergoing a similar operation with disease. You asked an important statistical question about the clinically greater incidence of overall infections with the donor operation than those with disease. According to our statistician, we would have needed many more discordant pairs to show a statistical difference. Clinically we remained concerned about any complications in our living donors. I think we are reducing the frequency of complications from our initial experience, but I don't think we will ever abrogate it. Timothy Sielaff, MD, Minneapolis, Minn: Thank you very much for bringing this really interesting discussion to the meeting. A comment and a question. First the comment. About 3 years ago, we did essentially an identical analysis of our living donors and right liver resections and found actually that the length of stay was longer for living donors than in the resection for diseased patients, 7 days vs about 5 days. And that was almost entirely ascribable to increased pain medication requirements of ileus in the donor population. And my question is, could you comment about your routine use of drains and how often they are effective in avoiding problems or identifying problems? Dr Nagorney: Our drainage policy is changing. For the LD operations, we are still going to place drains to recognize bile leaks as soon as possible. We believe that we can act upon leaks earlier and avert subsequent complications like an infected biloma. In our patients with disease, my policy toward drainage has changed. There are numerous randomized trials that show that drains are unnecessary. Even though personally I have drained patients for years and found them effective, the need is really minimal. So, over the study period, I have minimized the use of drains in major liver resections. Financial Disclosure: None reported.