JAMA Surgery

Schwenk, Wolfgang; Böhm, Bartolomäus; Witt, Christoph; Junghans, Tido; Gründel, Kerstin; Müller, Jochen M.

doi: 10.1001/archsurg.134.1.6pmid: 9927122

BackgroundLaparotomy causes a significant reduction of pulmonary function, and atelectasis and pneumonia occur after elective conventional colorectal resections.ObjectiveTo evaluate the hypothesis that pulmonary function is less restricted after laparoscopic than after conventional colorectal resection.DesignA randomized clinical trial.SettingThe surgical department of an academic medical center.PatientsSixty patients underwent laparoscopic (n=30) or conventional (n=30) resection of colorectal tumors. The 2 groups did not differ significantly in age, sex, localization or stage of tumor, or preoperative pulmonary function.Main Outcome MeasuresForced vital capacity, forced expiratory volume in 1 second, peak expiratory flow, midexpiratory phase of forced expiratory flow, and oxygen saturation of arterial blood.ResultsThe forced vital capacity (mean±SD values: conventional resection group, 1.73±0.60 L; laparoscopic surgery group, 2.59±1.11 L; P<.01) and the forced expiratory volume in 1 second (conventional resection group, 1.19±0.51 L/s; laparoscopic surgery group, 1.80±0.80 L/s; P<.01) were more profoundly suppressed in the patients having conventional resection than in those having laparoscopic surgery. Similar results were found for the peak expiratory flow (conventional resection group, 2.51±1.37 L/s; laparoscopic resection group, 3.60±2.22 L/s; P<.05) and the midexpiratory phase of forced expiratory flow (conventional resection group, 1.87±1.12 L/s; laparoscopic surgery group, 2.67±1.76 L/s; P<.05). The oxygen saturation of arterial blood, measured while the patients were breathing room air, was lower after conventional than after laparoscopic resections (P<.01). The recovery of the forced vital capacity and forced expiratory volume in 1 second to 80% of the preoperative value took longer in patients having conventional resection than in those having laparoscopic resection (P<.01). Pneumonia developed in 2 patients having conventional resection, but no pulmonary infection occurred in the laparoscopic resection group (P>.05).ConclusionsPulmonary function is better preserved after laparoscopic than after conventional colorectal resection. Pulmonary complications may be reduced after laparoscopic resections because of the better postoperative pulmonary function.THE SUPPRESSION of pulmonary function is a well-known sequela of abdominal surgery and was first described in 1933 by Beecher.Following upper abdominal incisions, forced vital capacity (FVC) and forced expiratory volume in 1 second (FEV1) are reduced by almost 60%because of a reflectory dysfunction of the diaphragm. Because the functional residual capacity is decreased postoperatively, small airways collapseand atelectasis occurs in most patients.Regardless of the anesthetic technique, pulmonary function does not recover to preoperative values within the first postoperative week after conventional abdominal surgery, and intensive physiotherapy does not prevent pulmonary dysfunction.Pneumonia is clinically apparent in more than 5% of all patients undergoing elective conventional colorectal resection and constitutes the most common general postoperative complication after conventional colorectal resection.Pulmonary function is better after simple laparoscopic procedures than after conventional surgery. Arterial oxygen saturation (SaO2) is less impaired and atelectasis occurs less often after laparoscopic than after conventional cholecystectomy.Although the laparoscopic approach to colorectal diseases has been reported by some authorsto be beneficial, it is still questionable whether laparoscopic colorectal resection actually results in better postoperative pulmonary function.A randomized study was conducted to determine whether postoperative pulmonary function is better following laparoscopic than conventional colorectal resections.PATIENTS AND METHODSHYPOTHESIS, END POINTS, AND SAMPLE SIZE CALCULATIONWe hypothesized that pulmonary function is less suppressed and recovers faster following laparoscopic than conventional colorectal resections. To test this hypothesis, we recorded postoperative changes in FVC, FEV1, peak expiratory flow (PEF), the midexpiratory phase of forced expiratory flow (FEF25%-75%), the relation of the FEV1to the FVC (FEV1/FVC ratio), and SaO2. The sample size needed to test the hypothesis was calculated according to the methods described by Altman.The FVC and FEV1were chosen as major criteria for calculating the sample size. It was assumed that FVC and FEV1values decrease by 50%±20% (±SD) following conventional colorectal resections.A difference of 15% in the postoperative FVC and FEV1values between the laparoscopic and conventional resection groups can be detected by a 2-tailed test with an α level of .05, a β level of .20 (power, 80%), and 30 patients in each group.STUDY POPULATIONAll patients with the diagnosis of a colorectal tumor who were scheduled for elective ascending colectomy, sigmoidectomy, proctosigmoidectomy, or abdominoperineal resection were included in the study. Patients who were scheduled for a sphincter-preserving anterior resection with total mesorectal excision of carcinoma of the middle or lower rectum (<12 cm from the anal verge) were excluded. Further exclusion criteria were intestinal obstruction, intra-abdominal abscess or sepsis, infiltration of tumor into adjacent organs or a tumor of more than 8 cm in diameter on computed tomographic scan, severe obesity (body mass index [or Quetelet index], calculated as weight in kilogramsdivided by the square of the height in meters, >32), operative risk greater than the American Society of Anesthesiologists' class III, and uncorrectable coagulopathy or thrombocytopenia.STUDY DESIGNThe study was approved by the local ethics committee. Concomitant cardiopulmonary diseases of all patients were recorded, and medical treatment was optimized before surgery. An informed consent was obtained from every patient. Mechanical bowel preparation and the perioperative administration of antibiotics were the same in the 2 groups. Anesthesia with endotracheal intubation was performed in a standardized manner by the same anesthesiological team using sufentanil, propofol, and atracurium besylate. Intraoperative sufentanil doses in micrograms per kilogram of body weight per minute were comparable in both groups.OPERATIVE TECHNIQUE AND INTRAOPERATIVE RANDOMIZATIONAll patients underwent a diagnostic laparoscopy. When the surgeon decided during the diagnostic laparoscopy that laparoscopic resection of the tumor was feasible, intraoperative randomization was accomplished, and a laparoscopic or conventional resection was carried out. If the surgeon decided that laparoscopic resection could not be performed, the patient was excluded from further evaluation and underwent a conventional resection. All laparoscopic procedures were performed by an experienced laparoscopic team using a standardized 5-trocar technique (infraumbilical and middle and lower abdomen bilaterally) that has been described in detail elsewhere.During proctosigmoidectomy and abdominoperineal resection, high ligation of the inferior mesenteric artery was accomplished with a linear endoscopic stapling device (EndoGIA 30; Autosuture Germany, Toenisvorst, Germany). During ascending colectomies, the ileocolic and right colic arteries were dissected with a stapler close to their origin from the superior mesenteric artery. In all resections with curative intent, a systematic regional lymphadenectomy was performed. After resection, the specimen was retrieved through a minilaparotomy of 3 to 4 cm in the left lower abdomen (proctosigmoidectomy) or the infraumbilical region (ascending colectomy). During ascending colectomy, a functional end-to-end ileotransversostomy was performed extracorporeally with a linear stapling device (GIA80; Autosuture Germany). For a proctosigmoidectomy, the anvil of a circular stapling device (Premium Plus CEEA; Autosuture Germany) was inserted into the descending colon extracorporeally. Then the minilaparotomy was closed, the pneumoperitoneum was reestablished, and the anastomosis between the descending colon and rectum was performed using the "double-stapling" technique. The anatomical extent of resection and the anastomotic technique were similar in the conventional surgery, but the resection was accomplished through a wide midline laparotomy.POSTOPERATIVE ANALGESIA, PULMONARY FUNCTION, AND SaO2All patients received patient-controlled analgesia with morphine sulfate until the morning of the fourth postoperative day and, from then on, tramadol hydrochloride orally until discharge. Pain was assessed using a visual analog scale during rest and during coughing. The patient-controlled anesthesia bolus was increased when the visual analog scale score at rest was higher than 50. All patients received supplemental oxygen (2-6 L/min) until the morning of the first postoperative day. Patients were discharged from the surgical intensive care unit to the regular nursing floor on the first postoperative day. Bedside spirometry (Renaissance Spirometer; Firma Puritan Bennett Hoyer, Gräfeling, Germany) was carried out with the patients lying in bed and the upper body elevated by 45°. Each test was repeated 3 times, and the best of the 3 results for FVC, FEV1, PEF, and FEF25%-75%were chosen for further analysis.The FEV1/FVC ratio (in percentage) was calculated from these values. Spirometry was performed preoperatively, 3 times per day from the first to the third day, twice a day from the fourth to the sixth day, and once a day from the seventh day until discharge. At the same time, SaO2was measured by pulse oximetry (Oxyshuttle+2; Critikon, Norderstedt, Germany) while the patients were breathing room air. Body plethysmography was performed preoperatively and on the fifth postoperative day in the Department of Pulmonology to validate the results of the spirometric tests.All patients received patient-controlled analgesia with morphine sulfate immediately after surgery until the eighth postoperative day. The doses were adjusted according to the patients' subjective pain perception, which was assessed every 8 hours by visual analog scale. All intraoperative and postoperative complications and mortality were recorded until 30 days after surgery. Pneumonia was diagnosed when patients had a temperature higher than 38°C, a productive cough, and radiological evidence of infection. To complete the measurements of the first postoperative week, patients were not discharged before the seventh postoperative day.DATA COLLECTION AND STATISTICAL ANALYSISNormally distributed continuous data are given as mean±SD and were compared between the groups using the Student ttest. If appropriate, the Wilcoxon rank sum test was performed. Categorical data were compared using the Fisher exact test. Correlations between continuous values were calculated using the Spearman rank test. A Pvalue of .05 was considered significant. Statistical analysis of all data was performed using commercial software (SAS for Windows; SAS Institute Inc, Cary, NC).RESULTSFrom April 19, 1995, to October 24, 1996, 60 patients were randomly assigned to laparoscopic (n=30) or conventional (n=30) resection of colorectal tumors. An 86-year-old patient in the laparoscopic resection group was switched to a laparoscopic-assisted sigmoidectomy because of tumor infiltration of the left ovary and was included in the laparoscopic group for an intent-to-treat analysis. After conventional sigmoidectomy, 1 patient required a second laparotomy because of hemorrhage from the greater omentum at the splenic flexure. The patient received mechanical ventilation for 24 hours after primary surgery.Patient demographics are shown in Table 1. There were no significant differences in these characteristics between the groups. Preoperatively, the FVC and FEV1values, the FEV1/FVC ratio, and the PEF, FEF25%-75%, and SaO2values were comparable between the groups (Table 2). None of the patients had pulmonary function values below 70% of expected for their age. Preoperative results of the spirometric tests were correlated to the results of body plethysmography (FVC, r=0.93, P<.001; and FEV1, r=0.89, P<.001).Table 1. Characteristics of Patients Undergoing Laparoscopic or Conventional Resection of Colorectal Tumors&ast;Patient GroupCharacteristicLaparoscopic Resection (n = 30)Conventional Resection (n = 30)P†Age, mean ± SD, y63.3 ± 12.264.8 ± 14.7.67BMI, mean ± SD, kg/m224.7 ± 2.924.7 ± 2.4>.99SexMale14 (47)16 (53).80Female16 (53)14 (47)Concomitant diseases‡Cardiac6 (20)10 (33).38Arterial hypertension7 (23)4 (13).51Diabetes mellitus2 (7)4 (13).67Pulmonary1 (3)3 (10).61Hepatic01 (3)>.99None20 (67)13 (43).12ASA classI14 (47)9 (30)II14 (47)19 (63).29III2 (7)2 (7)Type of resectionAscending colectomy4 (13)3 (10)Proctosigmoidectomy22 (73)24 (80).83Abdominoperineal resection4 (13)3 (10)Tumor stage§0 (adenoma)1 (3)3 (10)I9 (30)8 (27)II12 (40)5 (17).19III6 (20)8 (27)IV2 (7)6 (20)&ast;Values are number (percentage) unless otherwise indicated.†For age and BMI, the Student ttest was used; for all other variables, the Fisher exact test.‡Some patients had more than 1 concomitant disease.§From the Union Internationale Contre le Cancer.Table 2. Preoperative Values of Pulmonary Function Tests and Pulse Oximetry in Patients Undergoing Laparoscopic or Conventional Resection of Colorectal Tumors&ast;Patient GroupTestLaparoscopic Resection (n = 30)Conventional Resection (n = 30)P†FVC, L3.62 ± 1.263.38 ± 1.07.44FEV1, L/s2.57 ± 0.902.36 ± 0.95.39FEV/FVC ratio, %81.2 ± 4.678.9 ± 10.8.30PEF, L/min5.83 ± 2.645.57 ± 2.89.71FEF25%-75%, L/min4.11 ± 2.194.07 ± 2.58.94SaO2, %95.76 ± 1.2495.77 ± 1.52.98&ast;Data are given as mean ± SD. FVC indicates forced vital capacity; FEV1, forced expiratory volume in 1 second; PEF, peak expiratory flow; FEF25%-75%, midexpiratory phase of forced expiratory flow; and SaO2, arterial oxygen saturation.†Student ttest.The operating time was 219±64 minutes for the laparoscopic resection group and 146±41 minutes for the conventional resection group (P<.01). Postoperative morbidity and the number of deaths are shown in Table 3. The mean length of postoperative hospital stays was 10.1±3.0 days in the laparoscopic resection group and 11.6±2.0 days in the conventional resection group (P<.05). One patient with sigmoid carcinoma and diffuse liver metastases was discharged 9 days after conventional surgery but died 14 days later of hepatic failure.Table 3. Postoperative Morbidity and Death After Laparoscopic and Conventional Resection of Colorectal Tumors&ast;Patient GroupVariableLaparoscopic Resection (n = 30)Conventional Resection (n = 30)P†Major complications‡Hemorrhage on day of primary surgery01>.99Intraperitoneal abscess§01>.99Minor complications&par;Pneumonia02.49Perineal wound healing impairment01>.99Symptomatic hyperglycemia01>.99Central venous (CV) line infection01>.99Brachial plexus lesion (CV line)01>.99Urinary tract infection20.49Surgical morbidity03.20Total morbidity28.08DeathIn hospital00>.99<30 d01>.99&ast;Data are the number of patients.†Fisher exact test.‡Requiring a laparotomy.§On 17th postoperative day.&par;Requiring medical treatment.The postoperative suppression of pulmonary function (FVC, FEV1, PEF, and FEF25%-75%) was more severe after conventional than after laparoscopic resection (Table 4). The results of spirometric tests and body plethysmography in the postoperative period were also highly correlated (FVC, r=0.96, P<.001; and FEV1, r=0.94, P<.001). Improvement of pulmonary function had the same slope in both groups (Figure 1, Figure 2, and Figure 3). Recovery to 80% of the preoperative FVC value was achieved after 2.9±2.0 days in the laparoscopic resection group and after 5.2±2.6 days in the conventional resection group (P<.01) (Figure 1). Recovery of 80% of the preoperative FEV1value took 3.0±2.2 days in the laparoscopic resection group and 5.7±3.2 days in the conventional resection group (P<.01) (Figure 2). The postoperative values of the PEF reached 80% of the preoperative level within 3.9±3.3 days in the laparoscopic resection group and 5.7±3.7 days in the conventional resection group (Figure 3). The FEF25%-75%reached 80% of the preoperative value within 3.2±2.9 days after laparoscopic resection and 4.9±3.2 days after conventional resection (P<.05). There were no differences in the FEV1/FVC ratio between the groups.Table 4. Pulmonary Function at 2 PM on the First Day After Laparoscopic or Conventional Resection of Colorectal Tumors&ast;Patient GroupTestLaparoscopic Resection (n = 30)Conventional Resection (n = 30)P†FVC, L2.59 ± 1.111.73 ± 0.60<.01FEV1, L/s1.80 ± 0.801.19 ± 0.51<.01PEF, L/s3.60 ± 2.222.51 ± 1.37<.05FEV1/FVC ratio, %69.6 ± 11.067.7 ± 9.8.51FEF25%-75%, L/s2.67 ± 1.761.87 ± 1.12<.05SaO2, %93.8 ± 1.992.1 ± 3.3<.05&ast;Data are given as mean ± SD of actual values and percentage of preoperative values. The abbreviations are the same as explained in the first footnote of Table 2.†Student ttest.Figure 1. Postoperative changes in forced vital capacity (FVC) following laparoscopic or conventional resection of colorectal tumors. Pre indicates preoperative measurement; Op, day of surgery; asterisk, P=.01; dagger, P<.05; and double dagger, P=.06.Figure 2. Postoperative changes in forced expiratory volume in 1 second (FEV1) following laparoscopic or conventional resection of colorectal tumors. Pre indicates preoperative measurement; Op, day of surgery; asterisk, P<.01; and dagger, P<.05.Figure 3. Postoperative changes in peak expiratory flow (PEF) following laparoscopic or conventional resection of colorectal tumors. Pre indicates preoperative measurement; Op, day of surgery; asterisk, P<.05; dagger, P=.06; and double dagger, P<.01.Preoperative SaO2values were comparable between the groups (Table 2). From the morning of the first postoperative day, the SaO2value was lower after conventional resection than after laparoscopic surgery (Figure 4). Although the SaO2value remained almost unchanged in the laparoscopic resection group, it continuously decreased in the conventional resection group until 2 PM of the second day after surgery (Figure 4). The SaO2value at least once was lower than 90% in 22 patients (73%) having conventional resection but in only 14 patients (47%) having laparoscopic resection (P=.06).Figure 4. Postoperative changes in oxygen saturation of the arterial blood (SaO2), measured while patients are breathing room air, following laparoscopic or conventional resection of colorectal tumors. Asterisk indicates P<.05; dagger, P<.01.COMMENTPostoperative pneumonia occurs in more than 5% of all patients after elective conventional colorectal resection.The reason for this incidence of postoperative pneumonia is a prolonged impairment of pulmonary function induced by laparotomy.After a midline laparotomy, pulmonary function is depressed to about 50% of the preoperative value,and complete recovery of pulmonary function after abdominal surgery usually takes 7 or more days.Pulmonary function is more depressed after incisions in the upper abdomen than after lower abdominal laparotomy.Colorectal carcinoma, inflammatory bowel disease, or diverticular disease require the resection of larger bowel segments.Mobilization of the splenic or hepatic flexure is necessary in many cases, and exploration of the upper abdomen and the liver is mandatory in patients with malignant neoplasms. Therefore, the midline laparotomy is extended to the upper abdomen in most colorectal resections.The main reason for the suppression of pulmonary function after laparotomy is a decreased or even a paradoxical upward movement of the diaphragm during inspiration.Because of this diaphragmatic malfunction, a postoperative shift from abdominal (ie, diaphragmatic) to thoracic (ie, rib cage) breathing occurs, tidal volume is markedly decreased, and respiratory frequency is increased.The small airways of the lung (<1.0 mm in diameter) are not supported by cartilage and are influenced by transmitted intrapleural pressures. Normally, pleural pressures are less than atmospheric pressure, producing a positive transpulmonary pressure that distends these small airways.Because of the postoperative diaphragmatic malfunction, the intrapleural pressure rises and a negative transpulmonary pressure develops. This negative transpulmonary pressure causes the small airways to collapse.Collapse or narrowing of small airways will result in a reduction of ventilation to affected lung regions and produce a low ventilation-perfusion relationship. The lung volume at which a small airway begins to close is called the closing capacity of the lung. If the functional residual capacity (ie, volume of air remaining in the lungs at the end of a normal expiration) is decreased below the closing capacity of the lung, regions with a low ventilation-perfusion ratio will develop, which leads to impaired gas exchange. The failure of closed airways to reopen leads to a total collapse of the lung unit served by the airway, producing atelectasis and a reduced SaO2value.The influence of modern anesthesiological,analgesic,and physiotherapeutic techniqueson postoperative pulmonary function after conventional abdominal surgery has been evaluated to reduce the high incidence of pulmonary complications after gastrointestinal tract resections. Pain relief or intensive physiotherapy has improved the postoperative recovery of pulmonary function only marginally.A reduction of the functional residual capacity by 25%and the FVC or FEV1by about 50% appears to be inevitable because the degree of diaphragmatic dysfunction after upper abdominal laparotomy cannot be substantially influenced. This reduction of the pulmonary function will cause atelectasis in more than 50% of patients,and pneumonia will develop in up to 7% of patients after conventional colorectal resection, regardless of supportive therapy.LAPAROSCOPIC SURGERY, however, may reduce the degree of diaphragmatic dysfunction and the incidence of pulmonary complications.Residual pneumoperitoneum does not influence pulmonary function after diagnostic laparoscopy,but changes in diaphragmatic function have been found after laparoscopic cholecystectomy.After laparoscopic cholecystectomy, the FVC decreases to 54% to 79%, the FEV1to 54% to 80%, the PEF to 49% to 76%, and the FEF25%-75%to 68% to 88% of the preoperative value.Furthermore, the total lung capacity is suppressed to 92% and the maximum minute ventilation to 78% of the preoperative values.Several randomized trials have proved that FVC, FEV1, and FEF25%-75%values are suppressed by almost 50% after conventional cholecystectomy, whereas they are reduced by only 19% to 27% after laparoscopic gallbladder surgery.The postoperative SaO2decreased one third less in a laparoscopic resection group than in a conventional resection group,and there were significant differences in postoperative pulmonary function between the 2 groups, even when conventional cholecystectomy was performed by minilaparotomy.Furthermore, the SaO2, measured while patients were breathing room air, was significantly lower after conventional than after laparoscopic cholecystectomy.Few data have been available regarding postoperative pulmonary function after laparoscopic and conventional colectomy. Franklin et alcompared 84 laparoscopic and 110 conventional colorectal resections and found that "pulmonary toilet" was promoted after laparoscopic surgery. Senagore et aldescribed 5 cases of pneumonia (19%) after 26 laparoscopic colectomies and 11 cases of pneumonia (10%) after 110 conventional colorectal resections. All cases of pneumonia were diagnosed early in this series after surgery with a pneumoperitoneum of 15 mm Hg. Later, all resections were performed with an intra-abdominal pressure of 10 mm Hg, and no further patients with pulmonary infection were observed.In our department, laparoscopic colorectal resections are performed with a maximum intra-abdominal pressure of 12 mm Hg, and after more than 100 laparoscopic colorectal procedures, no cases of pneumonia have developed.Only 2 clinical trialsaddressing pulmonary function after laparoscopic and conventional colorectal resection have been published. Azagra et alperformed laparoscopic proctosigmoidectomies in 7 patients using trocar positions comparable to those used in our technique. Postoperative changes in FVC, FEV1, and PEF values after laparoscopic surgery were compared with those of laparoscopic-assisted (n=7) and conventional (n=7) proctosigmoidectomies. The authors did not find any significant difference in postoperative pulmonary function between the 2 groups. On the first postoperative day, the difference in PEF values between both groups was about 15%. This difference is almost identical to the difference in PEF values 24 hours after surgery in our trial (18.8%). To detect a 15% difference in PEF values between laparoscopic and conventional resections with a power of 80%, a sample size of 30 patients in each group would have been necessary.Therefore, it can be assumed that Azagra et alwere not able to detect significant differences in pulmonary function after laparoscopic, laparoscopic-assisted, and conventional proctosigmoidectomy because of the small sample size of their study.Stage et alrandomly assigned 29 patients from 3 different departments to laparoscopic (n=15) or conventional (n=14) colectomy and did not find any significant differences in pulmonary function between the groups. There are, however, some important differences between this study and our own trial. All patients investigated by these authors underwent perioperative thoracic epidural analgesia, which was not used in our trial. Almost 50% of their patients (n=14) underwent an ascending colectomy, whereas 46 of our patients (77%) underwent proctosigmoidectomy, which is the most common laparoscopic and conventional colorectal cancer resection. Although no exact data for mean preoperative pulmonary function are given by Stage et al, and no SD is shown in the figures given, the preoperative pulmonary function seemed to be much worse in their study (FVC, 2.3-2.5 L; and FEV1, about 1.8 L/s) than in our patients (preoperative FVC, 3.6±1.3 L and 3.4±1.0 L; and preoperative FEV1, 2.6±0.9 L/s and 2.4±1.0 L/s). Although the mean patient age was about 10 years lower in our study, other factors might be responsible for this considerable difference in preoperative pulmonary function between the groups, and these may explain the different postoperative findings.Our randomized study supports 3 theses reported by other authors: First, a wide midline laparotomy causes a postoperative depression of pulmonary function by 50% to 60%; second, complete recovery of pulmonary function takes more than 7 days; and third, significant decreases in SaO2values may occur several days after laparotomy, despite an uneventful course, suggesting the formation of atelectatic areas in the lung.In contrast to the conventional resection group, pulmonary function was reduced by only 35% after laparoscopic colorectal resection. Furthermore, it recovered within 3 days, and the postoperative SaO2was almost unchanged in the laparoscopic resection group. However, although a greater initial diminution of pulmonary function occurred in patients undergoing conventional colorectal resection compared with those having laparoscopic surgery, the daily amount of pulmonary recovery was comparable in both groups. The previously suggested pathogenesis may explain why the SaO2was more severely reduced from day 1 to day 4 after conventional colorectal resection in our study. Lindberg et alhave shown a significant correlation between pulmonary function, atelectatic areas, and the PaO2value after conventional colorectal resection. Therefore, the differences in SaO2values between the groups in our study (Figure 4) indicate a greater impairment of functional residual capacityand a higher incidence of atelectasis.Furthermore, the lower levels of SaO2may indicate an increased risk for postoperative complicationsafter conventional than after laparoscopic surgery.Clinical data from randomized trialsregarding pulmonary function after conventional and laparoscopic cholecystectomy support the assumption that the incidence of pneumonia may be reduced by laparoscopic surgery because radiological evidence of atelectasis occurred less often (40%) after laparoscopic than after conventional (90%) surgery, and the incidence of chest infection was higher after minilaparotomy (8%) than after laparoscopic cholecystectomies (1%). In our study, pneumonia developed in 2 patients (7%) after conventional colorectal resection, but not after laparoscopic surgery; this difference was not significant. Further randomized multicenter trials will need to evaluate whether the incidence of pneumonia is lower after laparoscopic colorectal resections than after conventional procedures.The data of our randomized trial demonstrate that pulmonary function is 30% to 35% less suppressed and recovers 40% to 45% faster after laparoscopic surgery than after conventional colorectal resection. No other change in surgical, anesthesiological, or physiotherapeutic technique has resulted in such an improvement of postoperative pulmonary function after colorectal resection. Because of these major advantages, the use of the laparoscopic approach may decrease the incidence of postoperative pulmonary complications after elective colorectal resections. Therefore, laparoscopic resection should be considered in every patient scheduled for an elective segmental resection of benign colorectal disease. Laparoscopic procedures, however, cannot be recommended for the curative treatment of malignant disease, unless randomized multicenter studies have determined the long-term results of the new technique.HKBeecherThe measured effect of laparotomy on the respiration.J Clin Invest.1933;12:639-650.CDWilliamsJBBrenowitzVentilatory patterns after vertical and tranverse upper abdominal incisions.Am J Surg.1975;130:725-728.RGLatimerMDickmannECDayVentilatory pattern and pulmonary complications after upper abdominal surgery determined by preoperative and postoperative computerized spirometry and blood gas analysis.Am J Surg.1971;122:622-632.DBCraigPostoperative recovery of pulmonary function.Anesth Analg.1981;60:46-52.GHedenstiernaMechanisms of postoperative pulmonary dysfunction.Acta Chir Scand Suppl.1989;550:152-158.PLindbergLGunnarssonLTokicsAtelectasis and lung function in the postoperative period.Acta Anaesthesiol Scand.1992;36:546-553.BEMarshallMQWyche JrHypoxemia during and after anesthesia.Anesthesiology.1972;37:178-209.HHendolinJLahtinenELansimiesTTuppurainenKPartanenThe effect of thoracic epidural analgesia on respiratory function after cholecystectomy.Acta Anaesthesiol Scand.1987;31:645-651.RHBartlettMLBrennanALGazzanigaELHansonStudies on the pathogenesis and prevention of postoperative pulmonary complications.Surg Gynecol Obstet.1973;137:925-933.EFChristensenPSchultzOVJensenPostoperative pulmonary complications and lung function in high-risk patients: a comparison of three physiotherapy regimens after upper abdominal surgery in general anesthesia.Acta Anaesthesiol Scand.1991;35:97-104.AJSenagoreMALuchtefeldJMMackeiganPWMazierOpen colectomy versus laparoscopic colectomy: are there differences?Am Surg.1993;59:549-553.OHansenWSchwenkHPHuckeWStockColorectal stapled anastomoses: experiences and results.Dis Colon Rectum.1996;39:30-36.BBöhmKNouchirvaniHPHuckeWStockMorbidity and mortality after elective resections of colorectal cancers [in German].Langenbecks Arch Chir.1991;376:93-101.PRSchauerJLunaAAGhiatasMEGlenJMWarrenKRSirinekPulmonary function after laparoscopic cholecystectomy.Surgery.1993;114:389-399.LGunnarssonLTokicsHGustavssonGHedenstiernaInfluence of age on atelectasis formation and gas exchange impairment during general anaesthesia.Br J Anaesth.1991;66:423-432.AJMcMahonJNBaxterGKenneyPJO'DwyerVentilatory and blood gas changes during laparoscopic and open cholecystectomy.Br J Surg.1993;80:1252-1254.MEFranklinDRosenthalRFNoremProspective evaluation of laparoscopic colon resection versus open colon resection for adenocarcinoma.Surg Endosc.1995;9:811-816.JSAzagraMGoergenEGilbartDJacobsPLejeuneECarlierAnterior resection: the total laparoscopic approach.In: Monson JRT, Darzi A, eds. Laparoscopic Colorectal Surgery.Oxford, England: Isis Medical Media; 1995:38-55.DGAltmanPractical Statistics for Medical Research.London, England: Chapman & Hall; 1991.JWMilsomBBöhmLaparoscopic Colorectal Surgery.New York, NY: Springer-Verlag NY Inc; 1996.American Thoracic SocietyStandardization of spirometry: 1987 update.Am Rev Respir Dis.1987;136:1285-1298.Union Internationale Contre le CancerTNM Classification of Malignant Tumours.Berlin, Germany: Springer Publishing Co; 1987.GHansenPADrablosRSteinertPulmonary complications, ventilation and blood gases after upper abdominal surgery.Acta Anaesthesiol Scand.1977;21:211-215.SRinnert-GongoraPITartterMultivariate analysis of recurrence after anterior resection for colorectal carcinoma.Am J Surg.1989;157:573-576.GTFordWAWhitelawTWRosenalPJCruseCAGuenterDiaphragm function after upper abdominal surgery in humans.Am Rev Respir Dis.1983;127:431-436.BDureuilJPCantineauJMDesmontsEffects of upper or lower abdominal surgery on diaphragmatic function.Br J Anaesthiol.1987;59:1230-1235.BDureuilNViiresJPCantineauMAubierJMDesmontsDiaphragmatic contractility after upper abdominal surgery.J Appl Physiol.1986;61:1775-1780.RWMWahbaPerioperative functional residual capacity.Can J Anaesth.1991;38:384-400.BMRademakerJRingersJAOdoomLTde WitCJKalkmanJOostingPulmonary function and stress response after laparoscopic cholecystectomy: comparison with subcostal incision and influence of thoracic epidural analgesia.Anesth Analg.1992;75:381-385.FBonnetCBleryMZatanOSimonetDBrageJGaudyEffect of epidural morphine on post-operative pulmonary dysfunction.Acta Anaesthesiol Scand.1984;28:147-151.GSimonneauAVivienRSarteneDiaphragm dysfunction induced by upper abdominal surgery: role of postoperative pain.Am Rev Respir Dis.1983;128:899-903.JSprungEYChengNNimphiusRDHubmayrJRRodarteJPKampineDiaphragm dysfunction and respiratory insufficiency after upper abdominal surgery.Plucne Bolesti.1991;43:5-12.DBenhamouGSimonneauTPoynardMGoldmanJCChaputPDurouxDiaphragm function is not impaired by pneumoperitoneum after laparoscopy.Arch Surg.1993;128:430-432.FEriceGSFoxYMSalibERomanoJLMeakinsSAMagderDiaphragmatic function before and after laparoscopic cholecystectomy.Anesthesiology.1993;79:966-975.IKanellosKZarogilidisEZiogasIDadoukisProspektiv-vergleichende Studie der Lungenfunktion nach laparoskopischer, Mini-Lap oder konventioneller Cholezystektomie.Minim Invasive Chir.1995;4:169-171.CKKumEEypaschAAljaziriHTroidlRandomized comparison of pulmonary function after the "French" and "American" techniques of laparoscopic cholecystectomy.Br J Surg.1996;83:938-941.RCFrazeeJWRobertsGCOkesonOpen versus laparoscopic cholecystectomy: a comparison of postoperative pulmonary function.Ann Surg.1991;213:651-654.JGStageSSchulzePMollerProspective randomized study of laparoscopic versus open colonic resection for adenocarcinoma.Br J Surg.1997;84:391-396.Corresponding author: Wolfgang Schwenk, MD, Universitätsklinik für Allgemein, Viszeral Gefäss und Thoraxchirurgie, Medizinische Fakultät der Humboldt–Universität zu Berlin, Charité, Schumannstr 20/21, 10117 Berlin, Germany (e-mail: [email protected]).

Shelton, Andrew A.; Madoff, Robert D.

doi: 10.1001/archsurg.134.1.13pmid: N/A

Keenan, Heather T.;Brundage, Susan I.;Thompson, Diane C.;Maier, Ronald V.;Rivara, Frederick P.

doi: 10.1001/archsurg.134.1.14pmid: 9927123

Abstract Background The relationship between facial fractures and traumatic brain injury is controversial. Some studies show an increased risk of brain injury with the presence of facial fractures while others claim that facial fractures protect against brain injury. Objective To examine the association between facial fractures and traumatic brain injuries. Design Case-control study. Setting Subjects were recruited from the emergency departments of 7 hospitals in the Seattle, Wash, area. Patients Three thousand eight hundred forty-nine injured bicyclists and 5 scene deaths were identified from March 1, 1992, to August 31, 1994, with complete data available on 3388 bicyclists. Interventions None. Results The study group was composed of 1602 cases with injuries to the head, face, or brain and 1540 control subjects. There were 203 bicyclists with traumatic brain injuries, of whom 62 had an identifiable intracranial injury and 141 suffered a concussion. A total of 81 patients sustained facial fractures. The odds ratio for the risk of intracranial injury associated with facial fractures after adjustment for significant confounders was 9.9 (95% confidence interval, 5.1-19.3). The effect was less strong but still present when all traumatic brain injuries including concussions were considered (odds ratio, 2; 95% confidence interval, 1.1-3.7). No association was found for concussion only. Conclusions This study demonstrates no evidence that facial fractures help prevent traumatic brain injury. Data indicate that facial fractures are markers for increased risk of brain injury. IT HAS BEEN proposed that the face protects the brain from injury the way an airbag protects the chest in a motor vehicle crash. Actual data on this are scant and conflicting. Lee et al1 reported that facial fractures are associated with a decreased risk of a traumatic brain injury, while Davidoff et al2 found facial fracture to be highly associated with traumatic brain injury. This question has important clinical implications. Multiple origins and potentially significant confounding variables make accurate assessment of the association between traumatic head injury and facial fractures difficult. In our study, we used a large database of individuals with injuries from bicycle crashes to examine the association between traumatic brain injury and facial fracture using a case-control design. Materials and methods Data collection Our current study data were collected as part of a case-control study examining head injuries caused by bicycle crashes. Methods of data collection are described in detail in a larger study that examined the effectiveness of bicycle safety helmets in preventing head injuries.3 Our case-control study was conducted at 7 major hospitals in the Seattle, Wash, area: Central and Eastside Hospitals of Group Health Cooperative of Puget Sound, a large health maintenance organization; University of Washington Medical Center; Harborview Medical Center, a regional level I trauma center; Overlake Hospital, a community hospital that cares for a large portion of trauma patients; and Children's Hospital and Medical Center of Seattle and the Mary Bridge Hospital and Medical Center of Tacoma, 2 children's hospitals. The medical examiner's offices of Pierce and King counties (Washington) were visited regularly to identify deaths at the scene that might have eluded the surveillance system. Subject identification Subjects were identified prospectively by regular surveillance (1 to 2 times a week) of the emergency department (ED) logs and records in each of the 7 study hospitals from March 1, 1992, to August 31, 1994. Automated hospital admission records were screened monthly by International Classification of Diseases, Ninth Revision, Clinical Modification coding (ICD-9-CM E codes 800-802, 805-807, 810-816, 818-823, 825-829) to identify admissions that were not recorded in ED records. Any individual injured on a bicycle was eligible for the study, including child passengers (<6 years) riding in a child carrier, (n=4), but excluding older bicycle passengers, pedestrians injured by a bicycle, and individuals assaulted while riding a bicycle. Medical record abstraction Trained abstractors reviewed the complete medical record of each patient, recording the following: mental status (nausea, headaches, amnesia, and length of unconsciousness), description of the incident, Glasgow Coma Scale scores, radiological procedures, helmet use, and a detailed description of all injuries. All radiology reports, surgical procedures, and discharge summaries were reviewed. Medical examiners' reports were abstracted for all deaths. A computer software program (TRICODE, TRIANALYTICS, Bel Air, Md) was used to convert text descriptions into ICD-9-CM codes and to calculate Abbreviated Injury Scale score and Injury Severity Score. Definitions Patients who sought care for a bicycle-related traumatic brain injury in the ED of a study hospital during the period from March 1, 1992, to August 31, 1994, were enrolled in this study. They were classified by injury into 2 groups: those with intracranial injuries and those with concussions. Intracranial injury included the following: cerebral lacerations, cerebral contusions, and subarachnoid, subdural, and extradural hemorrhages. Patients with loss of consciousness without signs of intracranial hemorrhage were defined as having a concussion. Loss of consciousness was defined by either a transient state of witnessed unresponsiveness or a patient report of temporary loss of awareness. A large group of individuals with traumatic brain injury were defined as having intracranial hemorrhage/contusions and a concussion. Control subjects were defined as bicyclists treated in the same EDs who reported hitting their helmet, head, or face or sustaining an injury to their head or face, without any traumatic brain injury. This control group was chosen to include only those bicyclists at risk of traumatic brain injury and/or facial fracture. Facial injury was defined as any injury to the jaw, lips, cheeks, nose, ears (external), eyes (external), forehead, or mouth (intraoral). Facial fracture was defined as injury to the maxilla, orbits, zygoma, or mandible. Questionnaires Detailed questionnaires were sent to all injured bicyclists treated in this period regarding the circumstances of the crash, including helmet use, speed, and motor vehicle involvement. Telephone follow-up occurred within 2 weeks of the ED visit. Response rate to the questionnaires was 88%; complete data were available for analysis on 3388 injured bicyclists. Statistical analysis The SAS statistical package was used for all analyses (SAS Institute Inc, Cary, NC). Descriptive information and crude odds ratios (ORs) were generated for all traumatic brain injuries. Unconditional logistic regression modeling was performed for traumatic brain injuries to estimate the OR and 95% confidence intervals (CIs) of traumatic brain injury while controlling for multiple confounders.4 To determine whether facial fracture was associated with severity of brain injury, analyses of all cases of traumatic brain injury, as well as the subcategories of intracranial injury and concussions were done. Results Seven hospital EDs treated 3849 injured bicyclists, 5 of whom died at the scene during the 212-year study. Two bicyclists were excluded because of incomplete data, leaving 3388 injured cyclists for analysis. Of these bicyclists, 1602 had head, face, or brain injury or recounted that they hit their helmet, head, or face. They constitute the study group. There were 203 bicyclists (14.4%) with traumatic brain injuries of whom 62 (3.9%) had intracranial injuries. Cases (intracranial injuries, n=62) and controls (all other face or head injured bicyclists, n=1540) did not differ significantly by age or sex; however, more controls (45.8%) than cases (24.2%) were wearing a helmet (Table 1). Characteristics of the crash (Table 2), including the bicyclists' self-reported speed and the type of surface (paved vs other) were not different between cases and controls; however, bicyclists whose speed was unknown were at a higher risk of intracranial injury (OR, 10.7; 95% CI, 3.3-33.3), most likely reflecting a recall bias for patients suffering intracranial injury. Collision with a motor vehicle occurred in 55% of cases, which significantly increased the odds of intracranial injury (OR, 6.6; 95% CI, 4.2-10.4). A total of 81 patients had facial fractures distributed between cases (29%) and controls (4.1%), including the following: 31 mandibular fractures or dislocations (28.3%); 6 maxillary fractures (7.4%); 29 nasal fractures (35.8%); and 15 orbital fractures (18.5%). Characteristics of cyclists' injuries are given in Table 3. Bicyclists with intracranial injuries were more likely to have concomitant neck injuries (OR, 3.1; 95% CI, 1.3-7.1). Injury Severity Scores ranged from 0 to 75; 24.2 % of the patients and 99.6% of controls had scores lower than 15. Patients with traumatic brain injuries had higher Injury Severity Scores, with 66.1% of the cases having scores of 16 to 40 compared with 0.4% of controls, and 9.7% of the cases having scores of 41 to 75 compared with no controls. The Glasgow Coma Scale scores were higher in the control group, with 98.9% of controls receiving scores from 12 to 15 compared with 56.5% of the cases. The OR for the effect of facial fracture on the risk of intracranial injury was adjusted for significant confounders including helmet use, motor vehicle involvement, and speed. Further adjustment for other variables including sex, age, and surface type had almost no effect on the OR. The adjusted OR for the risk of intracranial injury (n=62) associated with facial fracture was 9.9 (95% CI, 5.1-19.3). The effect was weakened, but still present, when all traumatic brain injuries (n=203), including concussions, were considered (OR, 2; 95% CI, 1.1-3.7); when only concussions (n=141) were considered, no association was found (OR, 0.6; 95% CI, 0.2-1.6). Comment This study demonstrates no evidence that facial fractures help prevent traumatic brain injury. In fact, the risk of intracranial injury in those bicyclists with facial injury was increased almost 10-fold, and the risk for all brain injuries including concussion was doubled. This does not imply that facial fractures cause traumatic brain injury, but suggests that blunt impact with enough force to break facial bones could also produce brain injury. Patients with intracranial injuries were more likely to have been involved in a crash involving a motor vehicle, to have had a higher Injury Severity Scale score, and to have coexisting neck injuries than bicyclists who had an injury to the head or face that suggested a more severe crash. Importantly, no protective effect of facial fracture was found, even in bicyclists with less severe brain injury such as concussion. In the group who had concussions, no association with facial fracture was observed. These results disagree with the previously reported decreased risk of traumatic brain injury in patients with facial fracture. Lee et al1 theorized that the facial bones act as a protective cushion for the brain to explain why injuries that crush the facial bones frequently caused no apparent brain damage. Our results are more in agreement with Davidoff et al2 who reported a 55% incidence of concomitant facial fracture and brain injury (defined as loss of consciousness or posttraumatic amnesia) and a 6% incidence of intracranial injury in a retrospective case series of 156 patients admitted to an ED with facial fractures from a variety of causes; however, the Davidoff study lacked a control group. There are some limitations to our study. The controls were limited to patients who fell and hit their head or face and sought care for a bicycle-related injury. Those patients who fell, but were not badly injured, and did not seek medical care were excluded from this study. This could underestimate the association of facial fracture with brain injury. Our study only examined 1 population, bicyclists who crashed, so we cannot comment on other mechanisms of injury that are reported as causes of facial fractures such as motor vehicle crashes or assaults.5,6 While our choice of patient population limits generalizability, it is a homogeneous data set from which to draw associations. Conclusions Our study shows no evidence that facial fractures are protective of traumatic brain injury. In fact, the risk of intracranial hemorrhage in bicyclists with a facial injury was increased by almost 10-fold, and the risk of any traumatic brain injury including a concussion was doubled. This suggests that the presence of enough force to cause facial fractures is likely to produce brain damage. Facial fractures are a marker for an increased risk of injury to the brain rather than protection against traumatic brain injury. Patients with facial fractures seen in the ED need to be evaluated for potentially serious brain injury. Reprints: Frederick P. Rivara, MD, MPH, Department of Pediatrics and Epidemiology, Harborview Injury Prevention and Research Center, Box 359960, 325 Ninth Ave, University of Washington, Seattle, WA 98104-2499 (e-mail: [email protected]). References 1. Lee KFWagner LKLee YESuh JHLee SR The impact-absorbing effects of facial fractures in closed-head injuries. J Neurosurg. 1987;66542- 547Google ScholarCrossref 2. Davidoff GJakubowski MThomas DAlpert M The spectrum of closed-head injuries in facial trauma victims: incidence and impact. Ann Emerg Med. 1988;176- 9Google ScholarCrossref 3. Thompson DCRivara FPThompson RS Effectiveness of bicycle safety helmets in preventing head injuries: a case-control study. JAMA. 1996;2761968- 1973Google ScholarCrossref 4. Breslow NEDay NE Statistical methods in cancer research. Breslow NEDay NEeds. The Analysis of Case-Control Studies. Vol 1 Lyon, France International Agency for Research on Cancer1980;International Agency for Research on Cancer Series, No. 32Google Scholar 5. Turvey TA Midfacial fractures: a retrospective analysis of 593 cases. J Oral Surg. 1977;35887- 891Google Scholar 6. Cannell HKing JBWich RD Head and facial injuries after low speed motor-cycle accidents. Br J Oral Surg. 1982;20183- 191Google ScholarCrossref

Viljakka, Mikko; Saali, Keijo; Koskinen, Matti; Karhumäki, Lauri; Kössi, Jyrki; Luostarinen, Markku; Teerenhovi, Ossi; Isolauri, Jouko

doi: 10.1001/archsurg.134.1.18pmid: 9927124

ObjectiveTo evaluate the influence of antireflux surgery on gastric emptying.DesignNonrandomized controlled trial 3 months before and after surgical intervention.SettingSecondary and tertiary referral center.Patients and Control SubjectsTwenty consecutive patients (7 women, 13 men), mean age 49.2 years, with symptomatic, objectively confirmed gastroesophageal reflux disease and 10 healthy control subjects (3 women, 7 men), mean age 37.3 years.InterventionLaparoscopic or open Nissen fundoplication (in 1 case Toupet 180° posterior hemifundoplication).Main Outcome MeasuresGastric emptying scintigraphy, using solid food, in control subjects and patients 3 months before and 3 months after the operation; time to halving of the maximal activity and the activity remaining at 60, 100, and 120 minutes.ResultsPreoperative symptoms included pyrosis in 19 of 20 patients and regurgitation in 18. Three months postoperatively, 19 patients were symptom-free. The mean time to halving of the maximal activity decreased from 113 to 78 minutes (P=.001). Delayed gastric emptying was found postoperatively in 3 patients, compared with preoperative values, using activity at 60, 100, 120 minutes and the mean time to halving of the maximal activity as the variables. Compared with control subjects, gastric emptying was slower in patients preoperatively and faster postoperatively, but the difference was not statistically significant.ConclusionGastric emptying is enhanced after antireflux surgery, along with cessation of symptoms and healing of esophagitis.THE PATHOGENESIS of gastroesophageal reflux disease is multifactorial.In surgical management, Nissen fundoplication is the favored method, with success rates of 91% to 95%.Multiple mechanisms of action have been suggested for Nissen fundoplication.Restoration of the normal lower esophageal sphincter pressure with reduced frequency of transient lower esophageal relaxation is documented,but its role in curing the esophagitis has been questioned.Esophageal motor activity was found to improve after fundoplication.The role of gastric emptying in the pathogenesis of gastroesophageal reflux disease is controversial. Some investigators reported delayed gastric emptying in about 40% of patients with refluxand esophagitis,whereas others found no correlation between reflux and delayed gastric emptying,and such delay was associated with normal rather than inflamed esophageal mucosa.Accelerated gastric emptying could be demonstrated after Nissen fundoplication,but not after the Hill operation.The failure of antireflux operations has been associated with delayed gastric emptying.The aim of the present study was to evaluate the influence of antireflux surgery on gastric emptying.PATIENTS, CONTROL SUBJECTS, AND METHODSPATIENTS, CLINICAL EXAMINATIONS, AND SUBJECTSTwenty patients (7 women, 13 men), mean age 49 years (range, 28-70 years), underwent surgery for the treatment of symptomatic gastroesophageal reflux disease at the Kanta-Häme Central Hospital, Hämeenlinna, Finland, between October 1, 1995, and August 31, 1996. The diagnosis was based preoperatively on the presence and duration of pyrosis, regurgitation, and dysphagia and the use of medication to treat reflux symptoms. Preoperative examinations included esophagogastroduodenoscopy, esophageal manometry, and 24-hour pH monitoring. Pathological reflux was found by the 24-hour pH recording (pH <4.00 for >4.2% of the recording time) in 18 of 20 patients (mean, 23.3% of the total monitoring time). The other 2 patients had erosive esophagitis and had experienced symptoms for many years. The clinical assessment was repeated 1 month and 3 months after the operation, and at 3 months, a control esophagogastroduodenoscopy was performed to observe the healing of esophagitis and the location and competence of the fundic wrap.Ten healthy volunteers (3 women, 7 men; mean age, 37 years; age range, 25-52 years) served as control subjects. The control subjects were using no medications.OPERATIONNissen fundoplication was performed on 16 patients by using a laparoscopic procedure and on 3 patients by using an open procedure (2 of these owing to previous abdominal surgery). A wrap 2 cm in length was constructed with a 50F Maloney bougie inside the esophagus. The short gastric vessels were not divided. One of the open operations was begun as a laparoscopic procedure but was converted to an open procedure owing to technical problems. For 1 patient with propulsive, low-amplitude peristalsis, the Toupet hemifundoplication was preferred.GASTRIC EMPTYING SCINTIGRAPHYThe patients fasted from midnight before the test. The test meal was taken with a glass of water. The test steak consisted of technetium Tc 99m colloid–labeled egg (20 MBq), 50 g of minced meat, and 1 to 2 spoons of bread crumbs cooked in a microwave oven. Immediately after the meal, imaging with a gamma camera was begun with the patient in a semisitting position. Two cobalt 57 (57Co) markers were attached on the skin. The dynamic stage of imaging lasted 30 minutes. Acquisition, at 1 frame per 20 seconds, was performed in an anteroposterior direction. Thereafter, static images for 5-minute periods were collected at 15- to 30-minute intervals, during which the patients were permitted to get up. The camera and seat positions were kept constant, and minor shifts were corrected with the aid of the 57Co marker. Imaging was continued until the activity of gastric contents had decreased to less than half of the maximal value, or for a maximum of 240 minutes. The 10 control subjects underwent the same gastric emptying scintigraphy.ANALYSIS OF GAMMA IMAGINGThe dynamic imaging and the 5-minute static images were linked in a file in temporal order. At this stage, if necessary, the static images were adjusted with the spots of the 57Co markers. Dynamic images were added as 5-minute periods to obtain dynamic stage activity values at 5, 10, 15, 20, 25, and 30 minutes. Half-time correction of technetium Tc 99m was performed. By using dynamic and static values, an activity curve was plotted, with background activity subtracted, expressing the activity as a percentage of the first time point. Activities (as percentages of the maximal activity) at 60, 100, and 120 minutes (A60, A100, and A120) were interpolated from the 2 nearest measurement points, and the time required for the activity to decrease to half of the maximal activity was defined (T1/2).STATISTICAL ANALYSISWe used the ttest for paired samples to compare preoperative and postoperative data and the ttest for unpaired samples to compare data for the patients and control subjects, with P<.05 considered significant. The mean differences with 95% confidence limits were calculated. Correlation coefficients were calculated for age and emptying data (T1/2 and A60). The study was approved by the ethics committees of Kanta-Häme Central Hospital and Tampere University Hospital, Tampere.RESULTSCLINICAL EXAMINATIONSAll patients had received medication for reflux symptoms; 14 had used proton pump inhibitors, and 8 had used histamine2blocking agents (some patients had used both types of medications). Cisapride, antacids, and sucralfate also were used. Preoperatively, 19 patients had pyrosis, 18 had regurgitation, and 5 had experienced dysphagia at some time. The respective numbers of patients experiencing these symptoms at 1 month after the operation were 0, 1, and 9, and at 3 months, they were 1, 1, and 2, respectively. The preoperative esophagogastroduodenoscopy grades for esophagitis (using Savary-Miller grading [grade 1, linear, nonconfluent erosions; grade 2, longitudinal, confluent, noncircumferential erosions; grade 3, longitudinal, confluent, circumferential erosions bleeding readily; and grade 4, ulcer]) were as follows: grade 0, 7 patients; grade 1, 7 patients; grade 2, 4 patients; grade 3, 1 patient; and grade 4, 1 patient. Two patients had Barrett esophagus, that was histologically confirmed (gastric-type epithelium >3 cm above the most cranial fold of gastric mucosa). Postoperative esophagogastroduodenoscopy showed no esophagitis (grade 0) in 18 patients and grade 1 in 1 patient; 1 patient refused to undergo esophagogastroduodenoscopy.GASTRIC EMPTYINGIn the preoperative imaging, the activity had fallen below 50% of the maximal before 100 minutes in 2 patients and before 120 minutes in 3 patients. In the postoperative imaging, the respective numbers of patients were 7 and 13. For these patients, the values at respective times were linearly extrapolated for statistical comparison. The means for all variables (Table 1) indicated enhanced gastric emptying 3 months after the operation. All changes were significant. For 3 patients, gastric emptying at 60, 100, or 120 minutes was slower or the T12was longer than in the preoperative examinations. Compared with control subjects, the preoperative gastric emptying was slower in patients (Table 1), but no statistical difference was noted between the patients for preoperative or postoperative examinations and the control subjects. The variation in the gastric emptying rate was wide for the patients and the control subjects (Figure 1and Figure 2). No significant correlation was found between age and T1/2 or A60in the preoperative and postoperative examination results for patients or between age and T1/2 or A60for the patients and the control subjects.Comparison of Preoperative vs Postoperative Gastric Emptying Data for Patients and Data for Patients vs Control Subjects&ast;PreoperativeMean Difference† (95% Confidence Interval)PostoperativeP†Control Subjects‡T12, min113.1 (40.5)−35 (−54 to −17)77.6 (26.6).00193.3 (40.8)A60, %83.1 (13.3)−20 (−31 to −9)63.2 (20.4).00176.9 (28.5)A100, %56.4 (16.7)−28 (−40 to −16)28.5 (24.4)<.00140.7 (31.6)A120,%42.9 (20.0)−26 (−38 to −15)16.5 (22.1)<.00129.3 (29.3)&ast;Unless otherwise indicated, data are given as mean (SD). T12indicates time until half of the radioactivity leaves the stomach; A60, A100, and A120, activity remaining after 60, 100, and 120 minutes, respectively.†Preoperative vs postoperative data.‡Comparisons of preoperative and postoperative data for patients vs control subjects were not significant.Figure 1. The time required for half of the gastric contents to leave the stomach in patients before and after the operation and in control subjects.Figure 2. The percentage of solid food retained at 60 minutes in patients before and after the operation and in control subjects.COMMENTThree months after antireflux surgery, gastric emptying was enhanced in most patients. By using T1/2 or activity at 60 minutes as the variables, gastric emptying was enhanced in 18 of 20 patients. Simultaneously, there was total abolition of symptoms and esophagitis in 18 patients (90%). Although these findings may not be mutually explanatory,they agree with the findings of previous studiesand could suggest a role for accelerated gastric emptying in the action mechanism of antireflux procedures.For 3 of the patients in the present study, gastric emptying was slower after fundoplication than before, but for 2 of them, the preoperative rate was faster than the average of the patients. Delayed gastric emptying has been associated with the failure of antireflux procedures to abolish symptoms of reflux,but these 3 patients were totally asymptomatic 3 months after the operation. These cases suggest that enhanced gastric emptying is not essential to the effectiveness of antireflux surgery. All patients with postoperative symptoms at 3 months had improved gastric emptying. Furthermore, no statistical difference in gastric emptying was found between preoperative and postoperative examination results for patients or between patients and control subjects.Although the preoperative gastric emptying was slower in patients, we could not demonstrate as clear a distinction between patients with gastroesophageal reflux disease and control subjects as was demonstrated in previous studies,because there was considerable overlap between and variation within the present study groups.The enhancement of gastric emptying could have resulted from the reduction in gastric volume related to the use of the fundus for plication. Similar enhancement was not found with the Angelchik prosthesis.Functional studies showed that the proximal stomach initially stores the solid component of food before its redistribution and emptying from the stomach,and removal of this reservoir may lead to accelerated emptying. It also has been suggested that delayed proximal rather than total gastric emptying could be important in the induction of transient lower esophageal sphincter relaxation.With simultaneous restoration of the lower esophageal sphincter pressure,enhanced gastric emptying can promote the progress of gastric contents into the intestine, thus inhibiting reflux into the esophagus.Esophageal transit times tend to normalize after Nissen fundoplication when the fundus is not mobilized,as in the present study. The improvement of esophageal transitand gastric emptying together may reflect improved upper gastrointestinal tract motility after antireflux surgery. Optimum esophageal motor function is desirable, since impairment is associated with esophagitis and reflux disease,although the significance of this finding has been questioned.After fundoplication, some patients complain of postprandial fullness,which may be related to delayed gastric emptying.Postoperatively, the increased gastric wall tension together with the reduced gastric volume also may account for the fullness.Heightened gastric tone can, in turn, enhance emptying. Flatulence also increases after fundoplication in some patients,but it is not accompanied by an increase of the intra-abdominal gas volume,which likewise may be explained by accelerated gastric emptying and possibly by normal, instead of abnormal, esophageal motility.To study gastric emptying, we used solid food, which seems to be preferable to liquids for detecting delayed emptying in patients with gastroesophageal reflux diseaseor esophagitis.Furthermore, solid food (with a glass of water) more closely mimics the postprandial state, when some patients experience excessive reflux.In postoperative tests, the percentage of retention of solid food at 100 minutes was similar to the findings in previous studies,as was the T1/2.The ages of the control subjects (mean, 37 years) and patients (mean, 49 years) differed, but there was no correlation between the age and gastric emptying, and gastric emptying was not influenced by age in previous studies.The symptoms of gastroesophageal reflux disease were relieved by antireflux surgery. At the 3-month postoperative follow-up, scintigraphy showed significantly enhanced gastric emptying. The preoperative gastric emptying for the patients was slower than for the control subjects, but in the small groups in the present study, no significant difference was found.POKatzPathogenesis and management of gastroesophageal reflux disease.J Clin Gastroenterol.1991;13(suppl 2):S6-S15.TRDeMeesterLBonavinaMAlbertucciNissen fundoplication for gastroesophageal reflux disease: evaluation of primary repair in 100 consecutive patients.Ann Surg.1986;204:9-20.SShiraziKSchulzeRSoperLong-term follow-up for treatment of complicated chronic reflux esophagitis.Arch Surg.1987;122:548-551.FHEllisRECrozierReflux control by fundoplication: a clinical and manometric assessment of the Nissen fundoplication.Ann Thorac Surg.1984;38:387-392.ACIrelandRHHollowayJToouliJDentMechanisms underlying the antireflux action of fundoplication.Gut.1993;34:303-308.JBancewiczMMughalMMarplesThe lower oesophageal sphincter after floppy Nissen fundoplication.Br J Surg.1987;74:162-164.RSFisherLSMalmudIFLobisWPMaierAntireflux surgery for symptomatic gastroesophageal reflux.Dig Dis.1978;23:152-160.RCGillKLBowesPDMurphyYJKingmaEsophageal motor abnormalities in gastroesophageal reflux and the effects of fundoplication.Gastroenterology.1986;91:364-369.LGrandeGLacimaERosDysphagia and esophageal motor dysfunction in gastroesophageal reflux are corrected by fundoplication.J Clin Gastroenterol.1991;13:11-16.RWMcCallumDMBerkowitzELernerGastric emptying in patients with gastroesophageal reflux.Gastroenterology.1981;80:285-291.GJMaddernBEChattertonPJCollinsMHorowitzDJCShearmanGGJamiesonSolid and liquid gastric emptying in patients with gastro-oesophageal reflux.Br J Surg.1985;72:344-347.AGLittleTRDeMeesterPTKirchnerGCO'SullivanDBSkinnerPathogenesis of esophagitis in patients with gastroesophageal reflux.Surgery.1980;80:101-107.SSShayDEggliCMcDonaldLFJohnsonGastric emptying of solid food in patients with gastroesophageal reflux.Gastroenterology.1987;92:459-465.AKeshavarzianDLBushnellSSontagEJYegelwelKSmidGastric emptying in patients with severe reflux esophagitis.Am J Gastroenterol.1991;86:738-742.WSchwizerRAHinderTRDeMeesterDoes delayed gastric emptying contribute to gastroesophageal reflux disease?Am J Surg.1989;157:74-80.GJMaddernGGJamiesonFundoplication enhances gastric emptying.Ann Surg.1985;201:296-299.GGJamiesonGJMaddernJCMyersGastric emptying after fundoplication with and without proximal gastric vagotomy.Arch Surg.1991;126:1414-1417.NVelascoLDHillRMGannanCEPope IIGastric emptying and gastroesophageal reflux.Am J Surg.1982;144:58-62.GJMaddernGGJamiesonBEChattertonIs there an association between failed antireflux procedures and delayed gastric emptying?Ann Surg.1985;202:162-165.GJMaddernJCMyersNMcIntoshFHGBridgewaterGGJamiesonThe effect of the Angelchik prosthesis on esophageal and gastric function.Arch Surg.1991;126:1418-1422.PJCollinsLAHoughtonNWReadRole of proximal and distal stomach in mixed solid and liquid meal emptying.Gut.1991;32:615-619.LLundellMAnvariPJCollinsJCMyersGGJamiesonThe association between gastric emptying, gastric distension and transient lower esophageal sphincter relaxations in patients with gastroesophageal reflux disease.Gastroenterology.1992;102:A478.MESLuostarinenMOKoskinenJOIsolauriEffect of fundal mobilisation in Nissen-Rossetti fundoplication on oesophageal transit and dysphagia.Eur J Surg.1996;162:37-42.PSinghAAdamopoulosRHTaylorDGColin-JonesOesophageal motor function before and after healing of esophagitis.Gut.1992;33:1590-1596.SLinMKeJXuPJKahrilasImpaired esophageal emptying in reflux disease.Am J Gastroenterol.1994;89:1003-1006.RTimmerRBreumelhofJHSMNadorpAJPMSmoutAmbulatory esophageal pressure and pH monitoring in patients with high-grade reflux esophagitis.Dig Dis Sci.1994;39:2084-2089.JBNegrePost-fundoplication symptoms: do they restrict the success of Nissen fundoplication?Ann Surg.1983;198:698-700.LRLundellJCMyersGGJamiesonDelayed gastric emptying and its relationship to symptoms of "gas-bloat" after antireflux surgery.Eur J Surg.1994;160:161-166.RHHollowayMHongoKBergerRWMcCallumGastric distension: a mechanism for postprandial gastroesophageal reflux.Gastroenterology.1985;89:779-784.MLuostarinenMKoskinenPReinikainenJKarvonenJIsolauriTwo antireflux operations: floppy versus standard Nissen fundoplication.Ann Med.1995;27:199-205.RMBremnerTRDeMeesterPFCrookesThe effect of symptoms and nonspecific motility abnormalities on outcomes of surgical therapy for gastroesophageal reflux disease.J Thorac Cardiovasc Surg.1994;107:1244-1250.TRDeMeesterLFJohnsonGJJosephMSToscanoAWHallDBSkinnerPatterns of gastroesophageal reflux in health and disease.Ann Surg.1976;184:459-469.JSVassilakisEXynosPKasapidisEChrysosAMantidesThe effect of floppy Nissen fundoplication on esophageal and gastric motility in gastroesophageal reflux.Surg Gynecol Obstet.1993;177:608-616.This study was supported by grants from the Finnish Foundation for Gastroenterological Research, Helsinki, and the Medical Research Fund of Tampere University Hospital, Tampere; and Medical Research Fund of Kanta-Häme Central Hospital, Hämeelinna.Reprints: Jouko Isolauri, MD, Turku University Central Hospital, Kiinamyllynkatu 4-8, 20520 Turku, Finland (e-mail: [email protected]).

Seale, A. Kent; Ledet, Jr, Walter P.

doi: 10.1001/archsurg.134.1.22pmid: 9927125

ObjectiveTo show that primary closure of the common bile duct following minicholecystectomy is safe and effective.DesignEighty-nine primary common bile duct closures done over 13 years.SettingGeneral community hospital.PatientsPatients with cholangiographic evidence of stones in the common bile duct.InterventionsMinicholecystectomy followed by a primary common bile duct closure or common bile duct closure over a T tube.Main Outcome MeasuresSuccessful clinical result of primary common bile duct closure following minicholecystectomy.ResultsEighty-nine primary closures of the common bile duct between 1983 and 1986, resulting in successful clinical outcomes.ConclusionPrimary closure of the common bile duct following minicholecystectomy is safe, effective, and inexpensive.OBSTRUCTION of the common bile duct as a result of the presence of stones has historically been corrected by opening the common bile duct and removing the stones. The duct is then closed around a T tube. Unfortunately, the use of T tubes carries a significant morbidity. The tube can be dislodged before a tract has developed and reoperation becomes necessary. A T tube must be left in place from 3 to 4 weeks before removal. There is a large amount of pain associated with the presence of the T tube over such a long period. In addition, a drain must also be positioned, and postoperative x-ray films are necessary before removal. Patients are reluctant to return to work with a T tube in place and can lose from 3 to 4 weeks of work.We believe that primary common bile duct closure using modern operative techniques can be as safe as primary common bile duct closure using a T tube. The patient returns home with 1 drain instead of 2 and the drain remains in place an average of 3 to 5 days. Early dislodgment of the drain does not always require reoperation and postdischarge x-ray films are not required. The patient has a smoother postoperative course, with less pain and an earlier return to work.Recently, common bile duct stones found during laparoscopic cholecystectomy have been handled by laparoscopic transcystic duct exploration or endoscopic sphincterotomy. Liberman and Phillipsreviewed a group of 76 patients, 59 of whom had a laparoscopic cholecystectomy and laparoscopic transcystic duct exploration and 17 of whom had laparoscopic cholecystectomy and endoscopic sphincterotomy. Their length of hospital stay of 6 to 7 days and their cost of $15,000 to $20,000 was considerably higher than a common bile duct exploration with primary closure.As early as 1917, Halsteaddescribed primary closure of the common bile duct that was drained using a tube through the cystic duct stump. Later closures were done using only a Penrose drain in the hepatorenal recess. Mayo,Kirschner,Mirrizzi,Edwards and Herrington,and Herrington et alhave written articles supportive of primary closure of the common bile duct. We believe that primary common bile duct closure can be both safe and effective.MATERIALS AND METHODSMethods for primary bile duct closure were developed in 1983. After a minicholecystectomy, a cholangiogram is done on the patients with laboratory and visual indications of an obstructed common bile duct. The ducts with stones are opened longitudinally in a convenient area. The relationship of the cystic duct to the incision in the common bile duct is not proven to be of consequence. The incision should not be extended within 1 cm of the bifurcation of the common bile duct. The stones in the common bile duct are removed either blind or using a choledochoscope, as well as Fogarty catheters, grasping forceps, irrigation, or stone baskets. The absence of stones in the common bile duct is verified using the choledochoscope. Using a headlight and ×2 magnification, the common bile duct is closed primarily using a 3-0 chromic running stitch (Figure 1). It is imperative that only small bites of the common bile duct are taken. A cystic duct cholangiogram is repeated after the bile duct has been primarily closed. A closed suction drain is always positioned.A, The initial stitch approximating only muscularis and mucosa is positioned at the proximal apex. B, The tail of the stitch is held and not tied. C, The suture proceeds distally to the apex, holding tension behind the running stitch. D, After reaching the apex, the suture line proceeds proximally approximating only the serosal layer. On reaching the proximal apex, the suture is tied.RESULTSOf the 1882 cholecystectomies performed between 1983 and 1996, 107 common bile duct explorations were performed: primary closure was performed in 89 and T-tube closure was performed in 18.Our patients for whom primary closure was performed ranged in age from 22 to 90 years (Table 1). The length of hospital stay for these patients ranged from 0 (day of surgery) to 7 days. Drains remained in place 3 to 10 days. We had no wound infections or drain tract infections. There were no associated intra-abdominal infections or abscesses and no phlebitis or pulmonary emboli. Postoperative bile leakage in all patients was controlled with the closed suction drainage; no reoperations were necessary. No patient returned with common bile duct strictures or retained stones and no patient who underwent primary common bile duct closure died postoperatively or within 90 days of discharge from the hospital.Characteristics of Patients Who Underwent Primary Common Duct ClosurePatient Age Group, yPrimary ClosureAverage Stay, dT Tube†20-3072.5130-4072.1140-5053.8250-6092.1060-70163.1470-80274.11080-90154.5090-10036.00Total893.618&ast;The duration of hospital stay is an average of the length of stay for each designated age group. In calculating the average, there were some patients who stayed as short as 0 days (day of surgery) and some who stayed as long as 7 days.†A comparison of the average stay of patients with a T tube is inappropriate as this procedure is done on sicker patients with multiple problems.CONCLUSIONSPrimary closure of the common bile duct offers improved patient care and should be considered a primary option. The average hospital costs for a minicholecystectomy, common bile duct exploration, and primary closure of the common bile duct was $5000 and the average length of hospital stay was 3.6 days. As opposed to T-tube drainage, the patient remained in the hospital for a shorter period and was not burdened over a 3- to 6-week interval by the T tube. Compared with laparoscopic transcystic exploration or endoscopic sphincterotomy, the average stay in the hospital was 2.6 less days, and the average cost was approximately $10,000 less.Closure of the common bile duct is a safe, effective, and less expensive means of handling stones in the common bile duct. It requires a shorter hospital stay and is notably less expensive. Many patients return to work within 5 to 10 days.MALibermanEHPhillipsCost-effective management of complicated cholecdocholithiasis: laparoscopic transcystic duct exploration or endoscopic sphincterotomy.J Am Coll Surg.1996;182:488-494.WSHalsteadSurgical Papers.Vol 2. Baltimore, Md: Johns Hopkins University Press. 1924:427-472.WIMayoAn address on the surgery of the hepactic and common bile ducts.Lancet.1923;1:1299-1302.MKirschnerOperations on the gall bladder and the bile passage.In: Radvin IS, trans. Vol 2. Abdomen and Rectum. In: Operative Surgery. Philadelphia, Pa: JB Lippincott; 1933:460-517.PLMirrizziPrimary suture of the common bile duct in choledocholithiasis.Arch Surg.1942;44:44-54.LWEdwardsJLHerrington JrClosure of the common bile duct following its exploration.Am Surg.1953;137:189.JLHerrington JrREDawsonWHEdwardsLWEdwardsFurther consideration in the evaluation of primary closure of the common bile duct following its exploration.Am Surg.1957;145:153-161.Corresponding author: A. Kent Seale, MD, Sulphur Surgical Clinic, 914 Cypress St, Sulphur, LA 70663.

Rutkow, Ira M.

doi: 10.1001/archsurg.134.1.105pmid: 9927143

THE SON OF A lawyer, Henry Hollingsworth Smith was born in Philadelphia on December 10, 1815. He attended the University of Pennsylvania, Philadelphia, where he also completed his medical studies (1837), and at the same time served as a private student of William Horner (1793-1853), professor of anatomy and dean of the school. Following graduation, Smith took a position as resident physician of Pennsylvania Hospital for 2 years. Believing his medical background to be incomplete, he spent 1840 and 1841 in Europe. Intending to make surgery his specialty, Smith returned to Philadelphia in the fall of 1841, where he began to teach private classes in minor surgery. The growth of Smith's practice was slow, such that he was able to devote much time to his work as a medical author. During that first year he translated from the French Jean Civiale's (1792-1867) Treatise on the Medical and Prophylactic Treatment of Stone and Gravel. In 1843, Smith prepared under the supervision of Homer an Anatomical Atlas, Illustrative of the Structure of the Human Body. Smith and Horner's relationship was strengthened no doubt by the former's marriage in October of that year to the latter's eldest daughter. Smith's first solo book length volume was also published in 1843, A Treatise on Minor Surgery. The following year Smith helped edit the fourth edition of his father-in-law's The United States Dissector, or, Lessons in Practical Anatomy. It was not unexpected that with his large literary output, the presentation of numerous lectures, and a little bit of nepotism, Smith was elected a surgeon to St Joseph's Hospital, Philadelphia, in 1849. This was soon followed by the Episcopal and Blockley Hospitals in 1854. For several years he was also an assistant lecturer on clinical surgery in his alma mater and, in 1855, was named professor of surgery. During the 1850s, Smith authored 2 well-known textbooks, A System of Operative Surgery (1852) and A Treatise on the Practice of Surgery (1856) (Figure 1). His A System of Operative Surgery was a massive compendium of various surgical operations and would prove particularly important to the photographic history of American medicine because it was the country's first surgical textbook to have illustrations based on daguerreotypes. In addition, Smith provided an extremely detailed introductory section describing the progress of American surgical history, a bibliographical index of American surgical writers with a list of their works by subject, and an extensive alphabetical list of American surgeons including the titles of their most important papers. View LargeDownload A limited number of Smith's A System of Operative Surgery were produced with "colored to life" plates. The shading helped highlight certain aspects of surgical technique, in this particular case "operations practiced at the lower portion of the neck." At the beginning of the Civil War, Smith was selected by the governor of Pennsylvania to reorganize the hospital department of the state and received the title of Surgeon-General of Pennsylvania. Returning to civilian life in late 1862, Smith found that the fifth edition of his Minor Surgery, the second edition of A System of Operative Surgery, and the first edition of his A Treatise on the Practice of Surgery had been completely sold out. Through an arrangement with J. B. Lippincott, it was decided to revise all 3 works and publish them together as a new 2-volume set, The Principles and Practice of Surgery Embracing Minor and Operative Surgery (1863). The work was an immediate success and is historically important because it represents the only systematic textbook of surgery to be authored by an American surgeon during the Civil War. Following his retirement from active academic life (1871), Smith continued a private surgical practice while participating in the politics of organized medicine. In 1878, he became president of the Philadelphia County Medical Society. Five years later, Smith was elected president of the Medical Society of the State of Pennsylvania. In 1887, he was named chairman of the Section of Military and Naval Surgery and Medicine of the Ninth International Medical Congress, conducted in Washington, DC. Smith died on April 11, 1890.

Cacciarelli, Thomas V.; Keeffe, Emmet B.; Moore, Dan H.; Burns, Washington; Busque, Stephan; Concepcion, Waldo; So, Samuel K. S.; Esquivel, Carlos O.

doi: 10.1001/archsurg.134.1.25pmid: 9927126

ObjectiveTo evaluate the effect of intraoperative transfusion of red blood cells (RBCs) on patient and graft survival.DesignA retrospective study.SettingA tertiary care referral center.PatientsBetween January 1, 1992, and December 31, 1994, medical records from 225 adult patients who underwent primary liver transplantations were analyzed.ResultsOverall patient survival was 90% at 1 year and 86% at 3 years, while graft survival was 89% at 1 year and 85% at 3 years. The following factors were associated with patient and graft survival: age, sex, medical condition at the time of transplantation, and intraoperative transfusion of RBCs. When these factors were subjected to a multivariate analysis, all were independently associated with survival. Fifty-four recipients (24%) underwent transplantation without intraoperative transfusion of RBCs, while 171 recipients (76%) received at least 1 U of RBCs intraoperatively. Recipients who did not receive transfusion of RBCs had higher patient and graft survival rates than patients who did receive RBCs. By multivariate analysis, transplantation without intraoperative transfusion of RBCs no longer remained statistically significant, and only sex and the patient's medical condition were independently associated with patient and graft survival. Patient and graft survival decreased if 5 or more U were transfused, but transfusion of 5 or more U was not independently associated with survival by multivariate analysis.ConclusionsIncreased transfusion requirement for RBCs was independently associated with patient and graft survival. While transplantation without transfusion of intraoperative RBCs was associated with superior patient and graft survival, these effects were overridden by patient sex and medical condition at the time of transplantation.IMPROVEMENTS IN patient selection, surgical techniques, postoperative management, and immunosuppression for orthotopic liver transplantation (OLT) have led to patient survival approaching 80% at 1 year.While attempts have been made to identify factors affecting patient and graft survival, no set of uniform predictive variables has been described.The effect of intraoperative blood loss and transfusions on survival after liver transplantation has been assessed by several different centers, with most data showing a correlation between blood use and postoperative morbidity and mortality rates.During a 3-year period, a substantial proportion of our adult patients underwent OLT without intraoperative transfusion of red blood cells (RBCs). These patients were compared with recipients who received at least 1 U of RBCs intraoperatively to determine the effect of not giving patients RBC transfusions on patient and graft survival rates after OLT.PATIENTS AND METHODSBetween January 1, 1992, and December 31, 1994, a total of 334 OLTs were performed at California Pacific Medical Center, San Francisco. Pediatric recipients (82 transplants) and retransplantations (27 transplants) were excluded from this study, leaving a total of 225 adult primary OLT patients for analysis. There were 105 female and 120 male patients, respectively. The mean age for the entire group was 49.5 years, 47.4 years for those patients who did not receive RBC transfusions, and 50.2 years for those who did receive RBC transfusions. Techniques for procuring and transplanting the donor liver have been described elsewhere.The mean (±SD) cold ischemia preservation times in the transfused and nontransfused groups were 11.6 ± 0.2 and 10.7 ± 0.4, respectively. The difference was not statistically significant. The 225 OLTs were performed by the following techniques: 169 piggyback (75%), 36 standard with venovenous bypass (16%), and 20 standard without venovenous bypass (9%). Recipients were divided into 2 groups based on intraoperative RBC transfusion requirement. The transfused group included 171 recipients (76%) who received at least 1 U of RBCs intraoperatively, while 54 recipients (24%) received no RBCs (either packed cells or salvaged blood) intraoperatively (nontransfused group). Cell saver was not used in patients requiring less than 30 U of blood. For the effect of RBC transfusion on patient and graft survival, the groups were arbitrarily divided into 3 groups: intraoperative transfusion of 0 (54 patients), 1 to 4 (64 patients), and 5 or more (107 patients) U of RBCs, respectively. The transfused group included 171 recipients (76%) who received at least 1 U of RBCs intraoperatively.STATISTICAL ANALYSISThe influence of the following variables on patient and graft survival after OLT were determined by univariate analysis: age, sex, primary diagnosis, medical condition at the time of transplantation, previous abdominal surgery, transjugular intrahepatic portosystemic shunts, preoperative prothrombin time, intraoperative RBC requirement, cold ischemia time, and operative time. The patient's condition at the time of transplantation was based on Child's classification and United Network for Organ Sharing (UNOS) status. Patient and graft survival were also calculated for the transfused and nontransfused groups. Variables that had a substantial effect on patient and graft survival by univariate analysis were entered into a proportional hazards regression analysis along with transfused and nontransfused recipients to determine whether transfusion of packed RBCs was independently associated with survival or whether confounding variables were present.Univariate analysis was performed by the Mann-Whitney nonparametric test. Patient and graft survival were determined using a log rank (Mantel-Cox) test and Kaplan-Meier cumulative survival plot. A proportional hazards model was used to control for any intergroup differences based on the aforementioned variables found to have a substantial effect on survival. Factors that were significant in univariate (Mann-Whitney) tests (P<.05) were considered in a multivariate proportional hazards regression model. Backward stepwise selection was used to find a group of factors that best influenced outcome. In the first step, all factors that had a significant effect on outcome as determined by univariate analysis were considered. At each step, the factor with the largest Pvalue for significance was removed until only those factors with P< .05 remained. These remaining factors were then considered to be those that were statistically significant in predicting outcome. Statistical software (S-PLUS, Version 3.2, StatSci, Division of MathSoft Inc, Seattle, Wash) was used to perform the logistic regression analysis.RESULTSIn the present study, the overall patient survival rates were 90% at 1 year and 86% at 3 years, and graft survival rates were 89% at 1 year and 85% at 3 years. Patient and graft survival for patients undergoing OLT with and without intraoperative transfusion of RBCs are shown in Figure 1. Transplant recipients who did not receive RBCs intraoperatively had significantly higher patient and graft survival compared with transfused recipients (P=.01 and P=.03, respectively). However, the following factors were also associated with patient survival: age (P=.04), sex (P=.005), and the patient's medical condition (P=.002). When these factors, along with whether a patient was transfused intraoperatively, were subjected to stepwise multivariate analysis, intraoperative transfusion of RBCs was not significant and only age (P=.03), sex (P=.004), and patient's medical condition (P=.003) remained significant (Table 1). The results of similar univariate and multivariate analyses for graft survival are given in Table 1. Men had poorer survival than women and survival was worse for critically ill patients (status 1 and 2a based on current UNOS classification).Figure 1. Patient survival (top) and graft survival (bottom) by intraoperative transfusion of red blood cells in 225 orthotopic liver transplantations (OLTs). The differences in patient survival (P=.01) and graft survival (P=.3) between recipients transfused and not transfused intraoperatively were statistically significant.Table 1. Variables Influencing Patient and Graft Survival After OLT in Recipients Receiving No RBCs vs 1 Unit or More of RBCs Intraoperatively&ast;VariableUnivariate AnalysisMultivariate AnalysisPatient survival0 vs ≥1 U RBCs.01.17Age.04.03Sex.005.004Patient's medical condition.002.003Graft survival0 vs ≥1 U RBCs.03.30Sex.008.005Patient's medical condition.002.002&ast;OLT indicates orthotopic liver transplantation; RBCs, red blood cells. Pvalues were calculated using proportional hazards regression analysis.When survival rates were examined, patient and graft survival decreased if 5 U or more were transfused (P=.006 and P=.01, respectively) (Figure 2). When subjected to multivariate analysis along with age, sex, and patient's medical condition, transfusion of 5 U or more was not independently associated with patient or graft survival (Table 2).Figure 2. Patient survival (top) and graft survival (bottom) by intraoperative transfusion of red blood cells (RBCs) in 225 orthotopic liver transplantation (OLT) recipients. The differences in patient survival (P=.006) and graft survival (P=.01) between recipients transfused 0 U and 5 or more U of RBCs intraoperatively was statistically significant.Table 2. Variables Influencing Patient and Graft Survival After OLT in Recipients Receiving Less Than 5 vs 5 Units or More of RBCs Intraoperatively&ast;VariableUnivariate AnalysisMultivariate AnalysisPatient survival<5 vs ≥ 5 U RBCs.004.12Age.04.02Sex.005.005Patient's medical condition.002.006Graft survival<5 vs ≥5 U RBCs.01.20Sex.008.006Patient's medical condition.002.01&ast;OLT indicates orthotopic liver transplantation; RBCs, red blood cells. Pvalues were calculated using proportional hazards regression analysis.Transfusion of packed RBCs, when examined as a continuous variable, was found to influence patient and graft survival by univariate analysis (P<.001). When subjected to multivariate analysis, blood transfusion as a continuous variable remained significant for both patient (P=.008) and graft (P=.03) survival, along with age, sex, and the patient's medical condition (Table 3). Figure 3shows relative survival as a function of amount of RBC transfusion and demonstrates decreasing survival with increasing transfusion requirement.Table 3. Variables Influencing Patient and Graft Survival After OLT, Analyzing Transfusion of RBCs as a Continuous Variable&ast;VariableUnivariate AnalysisMultivariate AnalysisPatient survivalRBC transfusion.0001.39Age.04.02Sex.005.003Patient's medical condition.002.003Graft survivalRBC transfusion.0001.61Sex.008.004Patient's medical condition.002.005&ast;OLT indicates orthotopic liver transplantation; RBCs, red blood cells. Pvalues were calculated using proportional hazards regression analysis.Figure 3. Relative probability of patient and graft survival based on transfusion requirements when analyzed as a continuous variable.When causes of patient and graft loss were examined, only 1 patient died (primary nonfunction) and 1 patient lost a graft (hepatic artery thrombosis) in the nontransfused group. Twenty-nine recipients in the transfused group died, and 45% (13 patients) died of septic complications, including generalized sepsis, endocarditis, aspergillosis, and bacterial pneumonia. The remaining causes of death were cancer (5 patients), cardiovascular disease (3 patients), trauma (3 patients), intraoperative cardiac arrest (2 patients), primary graft nonfunction (2 patients), natural causes (2 patients), and recurrent disease (1 patient). Causes of graft loss in the group included biliary strictures, hepatic artery thrombosis, and primary graft nonfunction, 1 patient each.COMMENTSeveral centers have examined factors that influence patient and graft survival after OLT, but no consensus exists as to which variables accurately predict outcome after OLT.According to these studies, factors that have increased the risk of death or graft loss for patients undergoing OLT vary and include UNOS status, Child's classification, fulminant hepatic failure, ABO incompatibility, compromised renal function, and infection before transplantation. In addition, some studies have reported a detrimental effect of intraoperative blood loss and massive transfusion requirements after liver transplantation.In a previous study, we conducted an analysis of several factors on the need for blood transfusion.These factors were recipient's age at the time of transplantation, sex, Child's classification, UNOS status, preoperative hematocrit, prothrombin time, partial thromboplastin time, platelet count, and fibrinogen level. In addition, other factors examined were primary liver disease, year of transplantation, history of abdominal surgery, transjugular intrahepatic portosystemic shunt placed prior to transplantation, use of venovenous bypass technique for transplantation, cold ischemia time, and operative time. The univariate analysis showed the following factors to be associated with OLT without blood transfusions: Child's classification, UNOS status, lack of previous right upper quadrant surgery, preoperative hematocrit, prothrombin time, activated partial thromboplastin time, piggyback technique, operative time, adult recipient status, and year of transplantation. A regression analysis showed that UNOS status (healthier patients required less blood), preoperative hematocrit, piggyback technique, operative time, and year of transplantation remained independently associated with OLT without transfusion of RBCs.Unfortunately, donor information was not available for the analysis.The present study examined the effect of OLT without transfusion of any RBCs on patient and graft outcome. A significant improvement was noted in both patient and graft survival when recipients underwent OLT without intraoperative transfusion of RBCs. However, other factors, rather than the presence or absence of intraoperative blood transfusions, appear to have a more substantial effect on outcome. Male sex and severity of the patient's medical condition (patients in an intensive care unit) had a negative influence on patient and graft survival. This finding is in agreement with UNOS scientific registry data, which have shown that sex and patient's medical condition affect both patient and graft survival.Why OLT without RBCs transfusion was not independently associated with superior patient and graft survival may be related to several factors. Previous studies demonstrating an adverse effect of transfusion requirements on outcome were conducted during an earlier era of OLT and examined the effects of massive intraoperative transfusion of blood and blood products.A more recent study also divided patients into a group of "bleeders" who received 10 U or more of RBCs.The present study compared patients receiving no RBCs with a group who received at least 1 U of RBC intraoperatively, and the study population did not contain a large proportion of bleeders. In fact, 50 patients (29%) in the transfused group received 3 U of packed RBCs or less intraoperatively (mean 5.7 U, median 4.0 U, range 1-67 U). The small number of RBCs required intraoperatively in the present study is probably due to technical improvements in performing OLT. Even when a group of recipients who received 5 U or more of RBCs was examined, an independent association with survival was not seen. Because many patients in the transfused group received few RBCs, the detrimental effect of massive transfusions in the few patients who required it were most likely masked by the remaining patients receiving only a few units of RBCs.When cutoff points were not employed and blood transfusion was examined as a continuous variable, it did have an independent association with survival. The concept that increased blood loss, and hence, a greater transfusion requirement during the transplantation, affects survival agrees with other data.Although blood loss is not the only factor that influences survival, it does point to the need for attention to detail during the procedure.Previous studies have reported a high mortality rate secondary to infectious complications in general,and particularly in patients who have received blood transfusions.Almost half the deaths in the present study involved septic complications, similar to the previous studies. In addition, all deaths due to sepsis occurred in the group that received transfusions, and the mean intraoperative RBC transfusion requirement for these patients was 13 U.The present study demonstrates that recipients who undergo OLT without transfusion of RBCs have superior patient and graft survival when compared with transfusion recipients, but other factors have a substantial effect on outcome after OLT, namely, age, male sex, and medical condition at the time of transplantation. When the effect of intraoperative blood transfusion on survival was examined as a continuous variable, an independent association with survival was noted, suggesting that blood loss should be kept to a minimum during the transplantation.Not Available1995 Annual Report of the US Scientific Registry of Transplant Recipients and the Organ Procurement and Transplantation Network-Transplant Data: 1988-1994.Bethesda, MD: Division of Organ Transplantation, Bureau of Health Resources Development, Health Resources and Services Administration, US Dept of Health and Human Services; 1995. No. 1076-8874.BWShaw JrRPWoodRDGordonSIwatsukiWPGillquistTEStarzlInfluence of selected patient variables and operative blood loss on six-month survival following liver transplantation.Semin Liver Dis.1985;5:385-393.JJBremsJRHiattJOColonnaVariables influencing the outcome following orthotopic liver transplantation.Arch Surg.1987;122:1109-1111.PBeligaRMMerionJGTurcottePreoperative risk factor assessment in liver transplantation.Surgery.1992;112:704-711.VCuervas-MonsIMillanJSGavalerTEStarzlDHVan ThielPrognostic value of preoperatively obtained clinical and laboratory data in predicting survival following orthotopic liver transplantation.Hepatology.1986;6:922-927.TLMotschmanHFTaswellMEBrecherJRakelaPMGrambschJJLarson-KellerTransplantation: their relationship to clinical and laboratory data.Mayo Clin Proc.1989;64:346-355.EMorLJenningsTAGonwaThe impact of operative bleeding on outcome in transplantation of the liver.Surg Gynecol Obstet.1993;176:219-227.DARouchJRThistlethwaiteLLichtorEffect of massive transfusion during liver transplantation on rejection and infection.Transplant Proc.1988;20:1135-1137.PNakazatoWConcepcionWBryTotal abdominal evisceration: an en-bloc technique for abdominal organ harvesting.Surgery.1992;111:37.TEStarzlSIwatsukiBWShaw JrTechnique of liver transplantation.In: Blumgart CH, ed. Surgery of the Liver and Biliary Tract.Edinburgh, Scotland: Churchill Livingstone Inc; 1987:1537.ATzakisSTodoTEStarzlOrthotopic liver transplantation with preservation of the inferior vena cava.Ann Surg.1989;210:649-652.TVCacciarelliEBKeeffeDHMoorePrimary liver transplantation without transfusion of red blood cells.Surgery.1996;120:698-705.VCuervas-MonsAJMartinezADekkerAdult liver transplantation: an analysis of the cause of death in 40 consecutive cases.Hepatology.1986;6:495-501.BWShaw JrRPWoodRJStrattaTJPillenANLangnasStratifying the causes of death in liver transplant recipients: an approach to improving survival.Arch Surg.1989;124:895-900.We thank Katie Allen for the skillful preparation of the manuscript.Corresponding author: Carlos O. Esquivel, MD, PhD, Professor of Surgery, Stanford University Medical Center, 750 Welch Rd, Suite 319, Palo Alto, CA 94304-1510.

Glasgow, Robert E.; Showstack, Jonathan A.; Katz, Patricia P.; Corvera, Carlos U.; Warren, Robert S.; Mulvihill, Sean J.

doi: 10.1001/archsurg.134.1.30pmid: 9927127

BackgroundVolume-outcome relations have been established for several complex therapies. However, few studies have examined volume-outcome relations for high-risk procedures in general surgery, such as hepatectomy for hepatocellular carcinoma (HCC).ObjectiveTo evaluate the relation between hospital volume and outcome for patients undergoing hepatectomy for HCC.DesignRetrospective cohort study.SettingAll acute-care hospitals in California.PatientsHospital discharge data were analyzed for each patient in California who underwent major hepatic resection for HCC from January 1, 1990, through December 31, 1994. Hospitals were grouped according to number of hepatectomies performed at each center during the 5-year study.Main Outcome MeasuresOutcome measures included operative mortality and length of hospital stay. Regression analyses were used to adjust for differences in patient mix.ResultsFive hundred seven patients underwent hepatectomy for HCC during the study. Hepatic resections were performed in 138 hospitals, with an overall in-hospital mortality rate of 14.8%. Three quarters of patients were treated at hospitals that average 3 or fewer hepatic resections for HCC per year. These low-volume providers represent 97.1% of all hospitals treating patients with HCC statewide. Significant reductions in risk-adjusted operative mortality rates (22.7%-9.4%; P=.002, multiple logistic regression) and risk-adjusted length of stay (14.3-11.3 days; P=.03, multiple linear regression) were observed as hospital volume increased.ConclusionsLow operative mortality and length of stay were associated with high-volume centers. These data support regionalization of high-risk procedures in general surgery, such as hepatectomy for HCC.TODAY'S CHANGING health care environment is being driven, in part, by external pressures on providers to deliver economical, high-quality care. For some medical therapies, quality of care varies little among providers, making cost a primary focus.For other treatments, however, quality of care is not uniform. Such is the case with coronary angioplasty, coronary surgery, and bone marrow and solid organ transplantation. For these complex therapies, a volume-outcome relationship exists where poor patient outcome, such as in-hospital mortality, is related to low provider volume and inexperience.These volume-outcome relations serve as the basis for the argument that high-risk procedures should be regionalized to centers of excellence.In the case of coronary angioplasty, coronary surgery, and transplantation, regionalization is beginning to occur as payers selectively contract with providers for these services. However, this is not the case with other complex therapies. In general surgical practice, standards for the minimum of experience necessary to perform highly complex and risky procedures, ie, major hepatic, pancreatic, or esophageal resection for neoplasia, do not exist. The number of these complex operations performed each year is insufficient for all surgeons and hospitals to have experience. Most of these operations are performed on an elective rather than emergent basis. Thus, if centers with superior patient outcomes could be identified, these procedures could be regionalized as a means of providing the most efficacious and cost-effective care.A key goal of any reorganization of health care delivery practices in the United States is to preserve or improve quality while reducing costs. Quality of a surgical procedure is measured by operative morbidity and mortality, outcome, effectiveness compared with alternate therapies, and patient satisfaction. It is an open question whether regionalization of high-risk procedures in general surgical practice is desirable or warranted from this standpoint. To help answer this question, we analyzed the relation between hospital volume and postoperative outcome in 1 high-risk, complex general surgical operation, major hepatic resection for hepatocellular carcinoma (HCC). We hypothesized that the risk, as measured by operative mortality, and the cost, as measured by length of hospital stay, are reduced when these patients are cared for in hospitals with greater experience with the procedure.MATERIALS AND METHODSDATA SOURCESWe retrospectively analyzed standardized patient discharge abstracts obtained from the California Office of Statewide Health Planning and Development (OSHPD), Sacramento. This database contains discharge data abstracts for every patient hospitalization from every acute-care facility in the state of California. Each abstract includes a variety of demographic, clinical, and hospitalization data that characterize a specific hospitalization. Each patient is assigned a principal diagnosis and procedure and up to 16 secondary diagnoses and procedures. The OSHPD database uses diagnostic and procedural codes derived from the International Classification of Diseases, Ninth Revision, Clinical Modification(4th ed) (ICD-9-CM), issued by the US Department of Health and Human Services.All discharge abstracts from January 1, 1990, through December 31, 1994, were included in the initial search of the OSHPD database. From these abstracts, all patients who underwent hepatic lobectomy (ICD-9-CMcode 50.3) or partial hepatectomy (ICD-9-CMcode 50.22) were examined. From this group, a subset of patients undergoing hepatic resection for HCC was selected (ICD-9-CMcode 155.0). Hospitals were characterized with regard to the number of acute and intensive care beds, discharges and patient hospital days per year, yearly overall surgical volume and number of hepatic resections for benign and malignant neoplasia, presence of a liver transplantation program and general surgery residency program, university affiliation, and capability for other complex surgery, as determined by the presence of cardiac surgery services. These data were derived, in part, from the Licensed Services and Utilization Profiles: Annual Report of Hospitalsfor January 11, 1991, through December 31, 1991.Frequency distributions for the individual patient characteristics within the data set and hospital characteristics listed above were computed.DATA ANALYSISPatients were grouped according to hospital identification number. Hospitals were then classified into quartile groups based on the number of hepatic resections performed in the study period. Crude operative mortality rate and length of hospital stay were calculated for each volume range. Operative mortality in this study was defined as patient death before hospital discharge. Because length of hospital stay is directly related to events within the postoperative course, patients with long hospital stays are most likely patients in whom significant perioperative complications develop. Thus, the percentage of patients with hospital stays longer than the 75th percentile (14 days) was calculated for each volume group. This measure served as a surrogate for postoperative complications, as other reliable objective measures of postoperative complications were not directly available within this database.To characterize a profile of hospitals within each volume group, the distribution of the various hospital characteristics was analyzed.Regression modeling was used to evaluate the independent associations of patient and hospital characteristics with the primary outcomes of interest (ie, operative mortality and length of hospital stay). The patient was the unit of analysis, with hospital volume group defined as a patient characteristic. This allowed for a volume group effect to be assessed while controlling for the characteristics of individual patients. Multiple logistic regression was used to model the dichotomous outcome, in-hospital mortality, and multiple linear regression was used to model length of hospital stay.The independent variables in these analyses included hospital volume, age group, sex, year of surgery, source of admission, type of resection (hepatic lobectomy or partial hepatectomy), presence of chronic liver disease, and presence of other preoperative comorbid illnesses. Age was entered into the regression equations as the following sets of dummy variables: 45 to 60 years, 60 to 75 years, and older than 75 years, with younger than 45 years as the reference group. Significant preoperative comorbid illnesses within a given patient abstract were grouped into 1 dichotomous variable to minimize potential colinearities among the various comorbidities. For example, patients with a history of congestive heart failure are likely to also have coronary artery disease. We believed the following comorbidities to have a significant influence on operative risk: coronary artery disease (ICD-9-CMcodes 412-414), chronic obstructive pulmonary disease (ICD-9-CMcodes 490-496), diabetes mellitus (ICD-9-CMcodes 250), congestive heart failure (ICD-9-CMcode 428), nutritional deficiencies (ICD-9-CMcodes 260-263), and preoperative intra-abdominal hemorrhage (ICD-9-CMcode 459). The presence of chronic liver disease, including cirrhosis (ICD-9-CMcode 571), was treated as a separate dichotomous variable, as it represents an independent factor associated with poor operative risk. The dependent variables for these analyses were operative mortality or death before discharge and length of hospital stay.Adjusted means for operative mortality rate and length of hospital stay were calculated from regression equations that included all of the independent variables. A complete description of the process of adjustment is provided by Cohen and Cohen.An adjusted mean is an estimate based on the hypothetical situation that all hospital volume groups had the same mean values on each of the independent variables that were entered into the equation. In other words, the adjusted mean represents the estimated operative mortality rate or length of stay if each of the volume groups treated patients with similar patient characteristics. To further evaluate this volume-outcome relation, an additional analysis was performed, where hospital volume was defined by the total number of hepatic resections for benign and malignant neoplasia, including metastatic disease.RESULTSDuring the study, 507 patients underwent major hepatic resection for HCC in the state of California. The number of resections performed each year was relatively constant, ranging from 117 in 1990 to 87 in 1993 (Table 1). A total of 138 hospitals reported to the database during the 5 years but, because a hepatic resection was not performed at all hospitals in each year, the yearly average number of hospitals in which a hepatic resection for HCC was performed was 53 (Table 1).Table 1. Distribution of Hospitals and Patients by Year of SurgeryYearNo. of Hospitals&ast;No. of Patients1990591171991579519925010119934587199456107Total138†507&ast;Indicates hospitals at which hepatic resection for hepatocellular carcinoma was performed.†Total number of hospitals reporting during the study period.Patient ages were distributed as seen in the following tabulation: Age Range, yNo. (%) of Patients<45126 (24.8)45-59125 (24.6)60-74203 (40.0)≥7553 (10.4)The median age was 60 to 64 years. Men outnumbered women by a ratio of nearly 3:2 (293 [57.8%] vs 214 [42.2%], respectively). The most common race of treated patients was white (263 [51.9%] of the study population), followed by Asian (135 [26.6%]), Hispanic (68 [13.4%]), other (21 [4.1%]), and African American (20 [3.9%]). Payer source is presented in the following tabulation: SourceNo. (%) of PatientsMedicare160 (31.6)Medi-Cal76 (15.0)Blue Cross/Blue Shield28 (5.5)Private insurance94 (18.5)Health maintenance or preferred health provider organization128 (25.2)Other21 (4.1)Source of admission and presence of comorbid medical illnesses were analyzed as indicators of severity of illness. Patients transferred from other acute-care hospitals and those admitted on an elective basis for surgery far outnumbered patients admitted from the emergency department (453 [89.3%] vs 54 [10.6%]). Preoperative comorbidities included diabetes mellitus (46 [9.1%]), coronary artery disease (37 [7.3%]), chronic obstructive pulmonary disease (36 [7.1%]), chronic nutritional deficiencies (17 [3.4%]), congestive heart failure (9 [1.8%]), and chronic renal insufficiency (6 [1.2%]). Eight patients, or 1.6% of the total, presented emergently with a diagnosis of intra-abdominal hemorrhage. Six (11.1%) of the 54 patients admitted through the emergency department had a diagnosis of intra-abdominal hemorrhage, compared with 2 (0.4%) of the 453 patients admitted on a routine basis or transferred from another acute-care facility. The median number of significant comorbidities per patient was 0, with a mean±SE of 0.70±0.04.Chronic liver disease, including cirrhosis, was present in 192 (37.9%) of patients overall. For patients with chronic liver disease, the crude operative mortality rate was 27.1% compared with 7.3% in patients without chronic liver disease (P<.001 by χ2analysis). Likewise, increasing number of comorbidities correlated with increased crude operative mortality. For patients with no comorbidities, the operative mortality was 5.0%. This rate increased significantly with increasing number of secondary diagnoses (P<.001 by linear regression analysis). For patients with 4 or more comorbidities, the operative mortality was 50.0%. Partial hepatectomy (ICD-9-CMcode 50.22) was performed in 299 patients compared with 208 patients who underwent hepatic lobectomy (ICD-9-CMcode 50.3).Each of the 4 volume groups included approximately 127 patients (25.0% of the total number of patients). One half of all patients were treated at centers where 6 or fewer resections were performed during the study (Table 2). These centers accounted for 88.4% of all reporting hospitals. In contrast, the highest-volume centers averaged 37 patients each during the 5 years and accounted for only 2.9% of the reporting hospitals statewide.Table 2. Hospital Volume Group AssignmentsHospital Volume Group, No. of Operations/5 yHospitals, No. (%)&ast;Patients, No. (%)1-290 (65.2)115 (22.7)3-632 (23.2)130 (25.6)7-1612 (8.7)116 (22.9)≥174 (2.9)146 (28.8)&ast;Indicates hospitals at which hepatic resections for hepatocellular carcinoma were performed.Hospital volume groups varied widely with regard to hospital characteristics (Table 3). The highest-volume providers were larger, with more acute and intensive care unit beds and more discharges and patient days per year. They also were more likely to be university hospitals with a general surgery residency training program. In addition to a higher overall surgical volume per year, the highest-volume providers were more likely to perform other complex operations (ie, coronary artery bypass) and to have a higher overall volume of hepatic resections for benign and malignant neoplasia. Three of the 4 highest-volume providers had a liver transplantation program, indicating an institutional interest in hepatology and hepatic surgery. In contrast, none of the lowest-volume providers performed liver transplantation.Table 3. Hospital Characteristics by Hospital Volume GroupsHospital Volume Groups, No. of Operations/5 yCharacteristicAll1-23-67-16≥17University hospital, %&ast;6.50.012.525.050.0Liver transplantation program, %&ast;8.00.012.533.375.0Residency training center, %&ast;18.77.037.533.375.0No. of acute-care beds, mean ± SEM&ast;302 ± 18243 ± 20337 ± 32493 ± 53734 ± 92No. of patient days/y, mean ± SEM&ast;63,877 ± 413047,548 ± 422474,753 ± 6884114,606 ± 11,242171,701 ± 19,472No. of discharges/y, mean ± SEM&ast;11,906 ± 7389498 ± 78912,643 ± 0.221,273 ± 209929,056 ± 3637No. of intensive care unit beds, mean ± SEM&ast;21.3 ± 1.714.4 ± 1.728.2 ± 2.939.8 ± 4.756.7 ± 8.1No. of operations/y, mean ± SEM&ast;7356 ± 4825808 ± 5328022 ± 86213,456 ± 140816,245 ± 2439No. of hepatectomies for neoplasia in 1990 through 1994, mean ± SEM&ast;†55.1 ± 1.36.3 ± 1.019.8 ± 0.952.4 ± 1.0139.8 ± 0.92Cardiac surgery center, %‡5445.662.583.3100&ast;P<.001 by analysis of variance, linear regression analysis.†Indicates January 1, 1990, through December 31, 1994.‡P<.01 by analysis of variance, linear regression analysis.The overall mortality rate for the study population was 14.8%. Crude operative mortality rates decreased with increasing hospital volume, from 24.4% in the lowest-volume centers to 6.2% in the highest-volume centers (Table 4). This inverse relationship between decreasing operative mortality rate and increasing hospital volume is summarized in Figure 1. This relationship was highly significant (P<.001 by logistic regression analysis). When the crude mortality rates for each volume group were adjusted to account for differences in patient characteristics, this highly significant volume-outcome relationship persisted (Table 4).Table 4. Crude and Adjusted Operative Mortality Rates by Hospital Volume GroupsHospital Volume Group, No. of Operations/5 yCrude Operative Mortality Rate, Mean %&ast;Risk-Adjusted Operative Mortality Rate, Mean %†‡1-224.422.73-616.213.37-1614.715.4≥176.2§9.4All14.8. . .&par;&ast;P<.001, using logistic regression analysis.†Indicates risk-adjusted using multiple logistic regression analysis for sex, age, year of operation, source of admission, type of resection, presence of chronic liver disease, and presence of significant comorbid illnesses.‡P<.05, hospitals with 17 or more operations vs those with 1 to 2, 3 to 6, and 7 to 16; hospitals with 1 to 2 operations vs those with 3 to 6, 7 to 16, and ≥17; and hospitals with ≥17 operations vs those with 1 to 2, using multiple logistic regression analysis.§P<.05, hospitals with ≥17 operations vs those with 1 to 2, 3 to 6, and 7 to 16, using χ2analysis.&par;Not applicable.Scatterplot diagram of hospital volume plotted against operative mortality rate.A similar volume-outcome relationship was observed when length of hospital stay was analyzed. The mean hospital stay was 12.9 days. The lowest-volume providers had a mean length of stay of 14.7 days compared with 10.8 days in the highest-volume providers (Table 5). When the length of hospital stay for each volume group was adjusted to account for differences in patient characteristics, this significant volume-outcome relationship persisted (Table 5). In addition, lowest-volume providers were more likely to have patients with hospital stays beyond the 75th percentile (Table 5). This implies that patients treated at lowest-volume providers were twice as likely to have complicated postoperative courses than patients treated at highest-volume providers. The mean (±SEM) length of stay for patients who died was longer than that for patients who survived (18.4±2.24 vs 12.0±0.42 days). When patients who died were excluded from the length-of-stay analysis, a significant volume outcome relationship persisted.Table 5. Crude and Adjusted Operative Length of Hospital Stay by Hospital Volume GroupHospital Volume Group, No. of Operations/5 yLength of Hospital Stay, Mean ± SEM,d&ast;Risk-Adjusted Length of Hospital Stay, d†Length of Stay >75th Percentile, Mean%‡§1-214.7 ± 1.014.330.43-613.7 ± 1.013.428.57-1612.9 ± 1.013.124.1≥1710.8 ± 0.911.3&par;15.8All12.9 ± 0.5. . .24.3&ast;P<.05, using linear regression analysis, hospitals with ≥17 operations vs those with 1 to 2, using Tukey-Kramer honestly significant difference.†Indicates risk-adjusted using multiple linear regression analysis, adjusting for sex, age, year of operation, source of admission, type of resection, presence of chronic liver disease, and presence of significant comorbid illnesses. Ellipses indicate not applicable.‡Indicates more than 14 days.§P<.05, using linear regression analysis, hospitals with 17 or more operations vs those with 1 to 2 and 3 to 6, using χ2analysis.&par;P<.05, hospitals with 17 or more operations vs those with 1 to 2, 3 to 6, and 7 to 16; and hospitals with 17 or more operations vs those with 1 to 2, using multiple linear regression analysis.When hospital volume was defined as total number of hepatic resections for benign and malignant neoplasia, these significant volume-outcome relations persisted. In 2004 patients, the overall operative mortality rate was 6.9%, ranging from 9.3% in the lowest-volume providers to 4.1% in the highest-volume providers.COMMENTIn an era of limited resources, increased attention is being paid in the United States to the most efficient delivery of health care services. The desired goal is to provide the highest quality of care while using the fewest resources. Volume-outcome analyses have shown that for certain complex surgical procedures, including heart transplantation and coronary artery bypass, the results of treatment are directly linked to experience. These observations have led to the suggestion that these complex, high-risk procedures should be regionalized to high-volume centers of excellence. To determine if similar volume-outcome relations exist in general surgical practice, we examined the relation between hospital volume and operative outcome in 1 complex, high-risk procedure, hepatic resection for HCC.Hepatic resection was selected for this analysis because it represents a technically challenging, resource-intensive procedure. Hepatic resection may be performed at any hospital, as there are no certificate-of-need requirements or other measures in place to regulate where such operations are performed. Similar to other complex, high-risk operations in general surgery, most hepatic resections are performed on an elective basis, making it possible for such cases to be regionalized to high-volume centers of excellence. Furthermore, hepatectomy for HCC is an infrequently performed procedure for which the average surgeon or hospital is not likely to have substantial experience. Hepatocellular carcinoma was selected as a model in this study, because it is a diagnosis for which major hepatic resection would be considered and represents a disease entity associated with significantly higher rates of operative morbidity and mortality compared with hepatic resection for other indications.Our main finding was a highly significant relationship between hospital volume and operative outcome for patients undergoing hepatic resection for HCC in California during the 5-year study. The sample size was large, encompassing 507 patients treated at 138 hospitals. The effect of hospital volume on operative mortality was substantial, with a 4-fold difference in mortality between the lowest- and highest-volume groups. In addition to reduced operative mortality rates, high-volume centers were found to have shorter average lengths of hospital stay. These findings suggest that the highest-volume hospitals not only treat patients undergoing hepatic resection for HCC with substantially lower operative mortality, but also accomplish this by using fewer resources.There are 3 possible explanations for these results. First, the data may have been flawed, and systematic reporting or coding errors may have favored the highest-volume providers. In previous studies of population database reliability, a 10% to 30% error rate in diagnosis and procedure coding has been reported.In our study, concordance between diagnosis and procedure codes was required for each patient as part of the inclusion criteria, thereby reducing the number of erroneously coded patient discharge abstracts from further analysis. An error rate of only 1% has been identified in the reporting of the end point of patient mortality when the OSHPD database has been reconciled to primary data from individual patient medical records.This error rate is not sufficiently large to explain the differences in operative mortality found between low- and high-volume hospitals in this study.Second, differences in patient characteristics could account for the observed variations in patient outcome. We controlled for these differences in 2 ways. First, the data were pooled by hospital volume such that the number of patients in each volume group was sufficiently large to offset small differences in patient mix at any one hospital. Second, we used multivariate statistical techniques to adjust mortality rates and length of hospital stay for differences in patient characteristics. This risk adjustment did not alter the primary finding of a highly significant association between hospital volume and patient outcome. Additional risks may have been present but not identifiable in the OSHPD database, and the distribution of patients with these unknown risk factors may have been skewed toward hospitals with low volume, but this seems unlikely.Finally, our findings probably represent true differences in outcome related to variations in patient care in hospitals of various volume groups. Supporting evidence suggests that this interpretation of the data is correct. First, the overall operative mortality rate of 14.8% and the rates for the lower-volume hospital groups in this study are comparable with previously reported operative mortality rates for this procedure in other large-population studies, including the nationwide Veterans Administration experience.Second, the operative mortality rates of the highest-volume centers in our study mirror the recently published results of other high-volume centers around the world.Therefore, the recently cited improvements in operative outcome in patients undergoing hepatectomy for HCC may apply only to high-volume centers and not more generally to the medical community. Finally, when the analysis in our study was extended to hepatic resection for all benign and malignant neoplasia, the volume-outcome relation persisted.Only a limited insight into potential differences in patient care leading to lower operative mortality in the high-volume centers could be gleaned from the available data. High-volume centers were larger hospitals with higher overall surgical volume and university and residency program affiliation. They were more likely to perform other complex operations, such as cardiac surgery. The highest-volume providers performed a greater overall number of hepatic resections per year, and 3 of 4 had active liver transplantation programs, compared with none of the lowest-volume providers. Furthermore, a greater percentage of patients treated at low-volume providers had extended hospital stays, suggesting complicated postoperative courses. Patients treated at high-volume centers were less likely to have postoperative complications, or these centers have developed the means to better manage these complications should they arise. Thus, the improved in-hospital mortality rates and shorter lengths of hospital stay for patients with HCC at high-volume centers likely represent not only increased volume or experience, but an institutional interest and expertise in the treatment of patients undergoing liver surgery.In today's health care environment, quality of care is measured not only in terms of patients' clinical outcome, but also cost-effectiveness. In California, most patients undergoing hepatic resection for HCC are treated at hospitals with limited experience. This study demonstrates a strong relationship between hospital volume and outcome in these patients. The standard of patient care for hepatic resection for HCC is not uniform across California. If the adjusted outcomes estimated for the highest-volume hospitals were applied statewide in 1990 through 1994, 44 additional patients would have survived their operation, with a savings of 1067 hospital days. Our results support the regionalization of high-risk general surgical procedures, such as hepatic resection for HCC, as a means of providing the most efficacious and cost-effective care.ELHannanJFO'DonnellHKilburn JrHRBernardAYaziciInvestigation of the relationship between volume and mortality for surgical procedures performed in New York State hospitals.&jama;.1989;262:503-510.GERosenthalDLHarperLMQuinnGSCooperSeverity-adjusted mortality and length of stay in teaching and nonteaching hospitals.&jama;.1997;278:485-490.HSLuftJPBunkerACEnthovenShould operations be regionalized?N Engl J Med.1979;301:1364-1369.JAShowstackKERosenfeldDWGarnickHSLuftRWSchaffarzickJFowlesAssociation of volume with outcome of coronary artery bypass graft surgery: scheduled vs nonscheduled operations.&jama;.1987;257:785-789.GLaffelABarnettSFinkelsteinMKayeThe relation between experience and outcome in heart transplantation.N Engl J Med.1992;327:1220-1225.ELHannanMRaczTJRyanCoronary angioplasty volume-outcome relationships for hospitals and cardiologists.&jama;.1997;279:892-898.JPBunkerHSLuftAEnthovenShould surgery be regionalized?Surg Clin North Am.1982;62:657-668.TGordonGBurleysonJTielschJCameronThe effects of regionalization on cost and outcome for one general high-risk surgical procedure.Ann Surg.1995;221:43-49.MJonesKBrouchDHallWAaronSt Anthony's Compact ICD-9-CM: Code Book for Physician Payment, 1995.Reston, Va: St Anthony's Publishing Inc; 1994:1-2.Office of Statewide Health Planning and DevelopmentLicensed Services and Utilization Profiles: Annual Report of Hospitals.Sacramento, Calif: Office of Statewide Health Planning and Development; 1991.JCohenPCohenApplied Multiple Regression/Correlation Analysis for the Behavioral Sciences.2nd ed. Hillsdale, NJ: Lawrence Eribaum Associates Inc; 1983.JITsaoJPLoftusDMNagorneyMAAdsonDMIlstrupTrends in morbidity and mortality of hepatic resection for malignancy.Ann Surg.1994;220:199-205.RDociLGennanriPBignamiMorbidity and mortality after hepatic resection of metastatic colorectal carcinoma.Br J Surg.1995;82:377-381.DENadigTPWadeRBFairchildKSVirgoFEJohnsonMajor hepatic resection: indications and results in a national hospital system from 1988 to 1992.Arch Surg.1997;132:115-119.AChenEMeuxGCoxReport of the Results From the OSHPD Reabstracting Project: An Evaluation of the Reliability of Selected Patient Discharge Data July Through December 1990.Sacramento, Calif: Office of Statewide Health Planning and Development; 1993.Not AvailableHow good are the data?Am J Kidney Dis.1992;46:675-683.SSLloydJPRissingPhysician and coding errors in patient records.&jama;.1985;254:1330-1336.TSegawaRTsuchiyaJFuruiKIsawaTTsunodaTKanematsuOperative results in 143 patients with hepatocellular carcinoma.World J Surg.1993;17:663-668.JVautheyDKlimstraDFranceschiFactors affecting long-term outcome after hepatic resection for hepatocellular carcinoma.Am J Surg.1995;169:28-35.HBismuthLChicheDCastaingSurgical treatment of hepatocellular carcinoma in noncirrhotic liver.World J Surg.1995;19:35-41.ECLaiS-TFanC-MLoK-MChuC-LLiuJWongHepatic resection for hepatocellular carcinoma.Ann Surg.1995;221:291-298.Corresponding author: Sean J. Mulvihill, MD, Department of Surgery, Room U-122, University of California–San Francisco Medical Center, San Francisco, CA 94143-0788 (e-mail: [email protected]).

Herfarth, Christian; Kienle, Peter

doi: 10.1001/archsurg.134.1.35pmid: N/A

Gibbs, James; Cull, William; Henderson, William; Daley, Jennifer; Hur, Kwan; Khuri, Shukri F.

doi: 10.1001/archsurg.134.1.36pmid: 9927128

ObjectiveTo improve the precision and reliability of estimates of the association between preoperative serum albumin concentration and surgical outcomes.DesignProspective observational study. Patients were followed up for 30 days postoperatively. Multiple logistic regression models were developed to evaluate serum albumin level as a predictor of operative mortality and morbidity in relation to 61 other preoperative patient risk variables.SettingForty-four tertiary care Veterans Affairs (VA) medical centers.PatientsA total of 54,215 major noncardiac surgery cases from the National VA Surgical Risk Study.Main Outcome MeasuresThirty-day operative mortality and morbidity.ResultsA decrease in serum albumin from concentrations greater than 46 g/L to less than 21 g/L was associated with an exponential increase in mortality rates from less than 1% to 29% and in morbidity rates from 10% to 65%. In the regression models, albumin level was the strongest predictor of mortality and morbidity for surgery as a whole and within several subspecialties selected for further analysis. Albumin level was a better predictor of some types of morbidity, particularly sepsis and major infections, than other types.ConclusionsSerum albumin concentration is a better predictor of surgical outcomes than many other preoperative patient characteristics. It is a relatively low-cost test that should be used more frequently as a prognostic tool to detect malnutrition and risk of adverse surgical outcomes, particularly in populations in whom comorbid conditions are relatively frequent.HYPOALBUMINEMIA has been shown to be associated with increased mortality and morbidity rates in both hospitalized patientsand samples of community-dwelling elderly persons.In surgery, an association between hypoalbuminemia and adverse outcome has been recognized for many years. In an early series of 26 patients, most operated on for diseases of the digestive tract, Jones and Eatonfound that postoperative edema was associated with low concentrations of serum albumin and serum protein, which they attributed to preoperative and postoperative undernutrition. More recently, Rich et alfound that patients undergoing cardiac surgery who had lower serum albumin levels showed a trend toward having higher postoperative mortality rates and had significantly higher rates of several complications than did patients with higher serum albumin levels, while controlling for other risk factors. Albumin also has been found to predict postoperative mortality and morbidity for patients undergoing elective surgeryand postoperative morbidity for those undergoing gastrointestinal tract surgery.Although it is well established that hypoalbuminemia, as a marker of malnutrition and disease, is associated with greater risk of adverse surgical outcome, previous studies have been based on relatively small samples and selected types of operations and have failed to adequately separate the predictive ability of albumin level from other risk factors. We used a large database, the National VA Surgical Risk Study,which included all major surgical procedures, to improve the precision and reliability of estimates of the association between albumin levels and mortality and morbidity. This database was sufficiently large and detailed to allow separate as well as summary analysis for 21 postoperative complications. The availability of comprehensive preoperative risk data permitted comparison of the predictive ability of serum albumin level with many other variables and measurement of the contribution of albumin level to prediction independent of the effects of other risk factors. Our results should better define the usefulness of preoperative albumin levels as a prognostic indicator.MATERIALS AND METHODSOVERVIEW OF THE NATIONAL VA SURGICAL RISK STUDYThe National VA Surgical Risk Study was initiated in 44 academically affiliated Veterans Affairs (VA) medical centers with the aim of developing and validating risk-adjusted models for the prediction of surgical outcome (30-day mortality and morbidity).Clinical nurse reviewers at each site prospectively collected 46 preoperative, 12 operative, and 24 postoperative variables for 87,078 major surgery cases between October 1, 1991, and December 31, 1993. Preoperative variables included health risk behaviors, comorbidities, and 14 laboratory test results. All surgical operations conducted under general, spinal, and epidural anesthesia were eligible for inclusion as the index operation except (1) operations on patients who had been enrolled in the study for an index operation within the previous 30 days, (2) operations with very low mortality rates (eg, ophthalmologic procedures), and (3) organ transplantation, because this is performed at a limited number of VA medical centers. The study protocol is described in detail elsewhere.ALBUMIN STUDY DATABASEThe analyses presented here were based on 54,215 (62%) of the 87,078 cases from the National VA Surgical Risk Study database for which preoperative serum albumin values were reported. The omitted cases, those not tested for albumin level (38%), were similar in age to those included in the analysis but were somewhat less likely to have comorbid conditions or severe disease. For example, 33% of the omitted cases vs 38% of the included cases had hypertension requiring medication, 12% vs 15% had diabetes mellitus requiring an oral agent or insulin, and 51% vs 64% were assigned an American Society of Anesthesiology (ASA) class of 3 or higher. (An ASA class of 1 indicates a healthy patient; 2, mild systemic disease but no functional limitations; 3, severe systemic disease with definite functional limitations; 4, severe systemic disease that is a constant threat to life; and 5, moribund patient unlikely to survive 24 hours with or without operation.) An analysis using all cases, in which missing albumin values were imputed from scores on the other preoperative variables, yielded very similar results to those reported here for cases tested for albumin level.INDEPENDENT VARIABLESPreoperative serum albumin values closest to the day of surgery were used in the analysis. Only those measurements that occurred within 30 days prior to surgery were considered valid. The number of days intervening between the last albumin measurement and surgery (mean=5.03 days, SD=6.15 days) was not found to affect the association between albumin level and mortality as indicated by a lack of statistical interaction between time of measurement and albumin level (χ2=1.47, P=.23). The predictive ability of serum albumin levels was evaluated against 61 other preoperative patient characteristics. These variables are listed in a previous publication.They include age, sex, tobacco use, alcohol use, substance abuse, functional status, weight loss, emergency operation, presence or absence of comorbidities covering all major organ systems, and 14 preoperative laboratory values.DEPENDENT VARIABLESOperative mortality was defined as death due to any cause occurring within 30 days of the operation. Operative morbidity included 21 predefined complications recorded in the 30 days after the operation. The summary measure used in this analysis was the presence or absence of 1 or more of the 21 complications. In other analysis of the National VA Surgical Risk Study database, this score was found to be strongly correlated with summary scores based on several weighting schemes, including the total number of morbidities reported (r=0.74), weights based on the regression relationship between postoperative length of stay and the 21 morbidities (r=0.75), weights based on the relative proportion of a morbidity occurring with other morbidities (r=0.96), and the mean ratings by the chiefs of surgery at the 44 participating VA medical centers on a scale of 1 to 5 as to the likelihood that the morbidity could result in death, permanent sequelae, prolonged length of stay, or patient dissatisfaction (r=0.97).STATISTICAL ANALYSISFirst, the association between preoperative serum albumin levels and mortality and morbidity was assessed univariately. We compared the predictive ability of albumin level with that of each of the 61 other preoperative variables. The predictive ability of each variable was determined by the cindexfrom logistic regressions with 1 independent variable. A cindex of 0.5 or less indicates prediction no better than chance, and values from 0.5 to 1.0 (perfect prediction) indicate improvement over chance.Second, multivariate logistic regression models were developed to assess the predictive ability of albumin level independent of the effects of other variables. The large number of candidate predictor variables was reduced in several steps. First, the variables were screened for significant univariate association (P<.001) with the dependent variables, which eliminated only 10 of the 62 candidate variables for mortality modeling and only 4 for morbidity modeling. Next, we eliminated variables that predicted the dependent variables with a cindex less than 0.57, a cutoff level that substantially reduced the number of variables but retained a diverse set of variables. We further reduced the number of remaining variables by developing separate stepwise regression models for the remaining comorbidity and laboratory variables, requiring P≤.0001 (Wald χ2) to enter the models and P≤.0005 (Wald χ2) to stay, and then eliminating variables that did not increase the cindex for the model by at least 0.01. These steps identified a diverse set of 7 strong predictors of mortality, including serum albumin level, that were forced in a final model to assess their relative predictive ability. Likewise, 6 variables were selected for a final morbidity model. These variables were tested for intercorrelation, and none of the correlations between pairs of predictor variables exceeded r=0.37.In addition to analysis of all operations in the database, we assessed the predictive ability of albumin levels within 3 subspecialties chosen to represent variation in mortality and morbidity rates as well as different types of surgery—general surgery and noncardiac thoracic surgery, which had the highest subspecialty 30-day mortality rates (6.4% and 6.0%, respectively) and higher than average morbidity rates (26.3% and 23.6%, respectively), and orthopedics, with a mortality rate of 2.5% and a morbidity rate of 13.7%. The model development process described for all operations was also used for the subspecialty analyses except that less stringent screening and model entry criteria were used owing to smaller sample sizes. Split-sample cross-validation of the all operations and subspecialty models showed little degradation of the cvalues, ranging from no decrease from learning to test samples in 3 of the models to a decrease of 0.03 in the cvalue for the orthopedic mortality model.RESULTSSAMPLE CHARACTERISTICSReflecting the composition of all VA hospital patients, 97.1% of the albumin study cases were male, and the mean age was 61.0 years (SD=13.0 years). Seventy-six percent were white, 18% were black, and 6% were of other racial backgrounds (eg, Hispanic, Asian American). The most frequent subspecialty was general surgery (28.3%), followed by orthopedics (18.0%), urology (14.8%), vascular surgery (11.7%), neurosurgery (8.4), thoracic surgery (7.3%), otolaryngology (5.9%), plastic surgery (3.7%), and other (2.0%). The 30-day mortality rate was 3.9%, and the 30-day morbidity rate (1 or more of the 21 complications) was 19.6%. The mean preoperative serum albumin level was 38 g/L (SD=6.5 g/L). Twenty-three percent of the cases had albumin values less than 35 g/L, which is generally considered the lower boundary of the normal range.UNIVARIATE ANALYSESTable 1shows the 10 strongest predictors of 30-day mortality and morbidity, ranked by the cindex, from the set of 62 preoperative variables. For both outcomes, albumin level was the best predictor, with ASA class and hematocrit ranking second and third, respectively. Albumin level alone correctly discriminated between survivors and nonsurvivors 78% of the time. In general, the cvalues are higher for mortality than morbidity, indicating that the variables predict mortality better than the occurrence of 1 or more complications.Table 1. Preoperative Predictors of 30-Day Mortality and Morbidity After Major Surgery, Ranked by cIndex Magnitude (Albumin Study Database)&ast;MortalityMorbidityVariable†cIndexVariablecIndexAlbumin, g/L0.78Albumin, g/L0.68ASA class0.77ASA class0.67Hematocrit, %0.72Hematocrit, %0.66Serum urea nitrogen, mmol/L0.69Functional status0.60Functional status0.67Age, y0.60Prothrombin time, s0.67Prothrombin time, s0.59Age, y0.64Serum urea nitrogen, mmol/L0.58Emergency operation0.64Alkaline phosphatase, U/L0.58WBC, >11,000/µL0.63Emergency operation0.57Alkaline phosphatase, U/L0.62WBC, >11,000/µL0.57&ast;ASA indicates American Society of Anesthesiology; WBC, white blood cell count.†All variables were examined using single-variable logistic regression models, and all variables were significant at P<.001.Figure 1and Figure 2display the relationship between serum albumin and mortality and morbidity for all operations and for the 3 selected subspecialties. For all operations, the mortality rate increases from less than 1% for albumin levels of 46 g/L or higher to 28% for albumin levels below 21 g/L. The increase in mortality seems to be exponential as albumin level decreases from a level of approximately 40 g/L. Figure 2shows that morbidity increases from approximately 10% to 65% as albumin values decline from 46 g/L to less than 21 g/L. The subspecialties show similar trends except that the mortality and morbidity rates are lower for orthopedics than the other subspecialties.Figure 1. Thirty-day mortality rate by preoperative serum albumin level for all operations and 3 subspecialties.Figure 2. Thirty-day morbidity rate by preoperative serum albumin level for all operations and 3 subspecialties.MULTIVARIATE ANALYSESTable 2presents the final models developed from the 62 preoperative variables for all operations and for 3 selected subspecialties. The models predict mortality very well for all operations, general surgery, and orthopedics, with cvalues ranging from 0.86 to 0.91. The cvalues for the morbidity models are not as high, ranging from 0.67 to 0.76. The predictive variables are listed in Table 2in order of the magnitude of their Wald χ2values, which indicates their relative importance in the models. Albumin level appears in all of the models and is the strongest predictor in both the mortality and morbidity models for all operations and in several of the subspecialty models. The odds ratios for albumin level in the all operations models indicate that a decrease of 10 g/L in albumin value was associated with more than a 2-fold increase in the odds of dying and almost a 2-fold increase in the odds of a complication. The ASA class was also a strong predictor, particularly in the mortality models. Both serum albumin levels and weight loss appear in the mortality model for thoracic surgery, indicating that albumin level detects risk not associated with weight loss of greater than 10% of body weight in the 6 months prior to surgery.Table 2. Multivariate Models of 30-Day Mortality and Morbidity After Major Surgery (Albumin Study Database)&ast;MortalityMorbidityVariableWald χ2†OR95% CIVariableWald χ2†OR95% CIAll Operations‡Albumin level, g/L541.00.440.41-0.48Albumin level, g/L813.90.580.56-0.60ASA class416.12.332.15-2.53ASA class745.81.711.65-1.78Emergency operation212.02.262.02-2.53Emergency operation422.12.001.87-2.14Serum urea nitrogen, mmol/L109.21.011.01-1.02Functional status351.21.481.42-1.55Age, y108.91.031.02-1.03Age, y115.11.011.01-1.01Dyspnea63.81.381.27-1.49Serum urea nitrogen, mmol/L82.21.011.01-1.01Functional status47.41.281.19-1.37General Surgery§Albumin level, g/L162.90.490.44-0.55Albumin level, g/L245.20.610.57-0.65ASA class140.12.181.91-2.49ASA class199.61.601.50-1.71Ascites86.14.473.23-6.17Emergency operation181.81.961.77-2.16Age, y45.71.031.02-1.03Functional status84.71.461.34-1.58Serum urea nitrogen, mmol/L44.11.011.01-1.02Age, y55.21.011.01-1.02Functional status36.61.411.26-1.57Transfusions (>4 U preoperatively)34.41.751.44-2.11Emergency operation34.41.681.41-2.01Serum urea nitrogen, mmol/L25.81.011.01-1.01Prothrombin time, s28.51.111.07-1.15Thoracic Surgery&par;Albumin level, g/L40.50.510.41-0.63Age, y29.71.021.02-1.03Serum urea nitrogen, mmol/L13.91.021.01-1.03Ventilator use (preoperative)29.64.662.64-8.19Weight loss12.51.861.31-2.65Albumin level, g/L25.70.700.61-0.81Dyspnea12.01.451.17-1.79ASA class15.61.351.16-1.58ASA class8.51.471.12-1.92Functional status12.01.391.15-1.69Age, y7.71.021.01-1.04Prothrombin time, s6.61.081.02-1.15Emergency operation6.91.821.15-2.86Dyspnea5.41.181.02-1.36Orthopedic Surgery¶ASA class70.42.952.28-3.18Age, y36.11.021.01-1.02Albumin level, g/L40.30.490.39-0.61Wound infection (preoperative)28.61.711.40-2.09Age, y27.71.041.02-1.05Functional status27.71.381.22-1.56Emergency operation26.12.381.70-3.34ASA class19.31.311.16-1.47Weight loss24.23.141.97-5.00Hematocrit, %17.00.970.96-0.99DNR orders17.12.471.60-3.83COPD16.01.431.19-1.70Functional status16.21.541.24-1.90Alkaline phosphatase, U/L15.71.001.00-1.00Serum urea nitrogen, mmol/L10.01.011.01-1.02Albumin level, g/L12.50.780.68-0.90Emergency operation11.81.421.16-1.74WBC, >11,000/µL11.11.321.12-1.56Impaired senses9.11.381.12-1.72Serum urea nitrogen, mmol/L5.91.011.00-1.01&ast;OR indicates odds ratio; CI, confidence interval; ASA, American Society of Anesthesiology; DNR, do not resuscitate; COPD, chronic obstructive pulmonary disease; and WBC, white blood cell count. Total sample size is reduced slightly from 54,215, and sample sizes differ between the mortality and morbidity models because of varying numbers of missing values associated with each set of predictor variables. The number of cases for the subspecialty models do not add to the totals for the all operations models because the all operations models also include cases from all other subspecialties in the database.†P≤.001 for all variables in the all operations models. P≤.05 for all variables in the subspecialty models.‡For all operations, mortality data, N = 53,055, with 2076 deaths and a cindex of 0.86. For morbidity data, N = 53,136, with 10,358 cases of complications and a cindex of 0.73.§For general surgery, mortality data, N = 13,566, with 864 deaths and a cindex of 0.87. For morbidity data, N = 14,782, with 3888 cases of complications and a cindex of 0.74.&par;For thoracic surgery, mortality data, N = 3740, with 228 deaths and a cindex of 0.74. For morbidity data, N = 3685, with 877 cases of complications and a cindex of 0.67.¶For orthopedic surgery, mortality data, N = 9467, with 235 deaths and a cindex of 0.91. For morbidity data, N = 8673, with 1141 cases of complications and a cindex of 0.76.COMPLICATION-SPECIFIC ANALYSISIn addition to analysis of morbidity using the summary measure of complications, we analyzed the association between preoperative serum albumin and each of the 21 predefined complications. Table 3gives the mortality rate, incidence, and regression results for each of the complications, listed in order of the magnitude of the cindex for the regression of the complication on serum albumin level. The complications with the higher cvalues tend to have higher mortality rates. They also are characterized by incidence rates that are markedly higher for cases with albumin levels below normal (<35 g/L) than for cases in the normal range (≥35 g/L).Table 3. Association Between Preoperative Serum Albumin Level and Selected Complications After Major Surgery (Albumin Study Database)&ast;Complication (Mortality Rate, %†)Incidence When Albumin Level <35 g/L, %Incidence When Albumin Level ≥35 g/L, %cIndex for Univariate Prediction‡Entry Step of Albumin Level in Stepwise Logistic RegressionSystemic sepsis (40.0)8.01.30.781Acute renal failure (47.8)2.10.40.762Coma (64.1)2.30.40.762Renal insufficiency (41.3)3.00.60.753Failure to wean from ventilation (30.9)11.22.30.753Bleeding/transfusions (27.3)7.11.60.722Cardiac arrest (77.8)4.41.10.722Pneumonia (22.1)10.62.90.711Urinary infection (8.1)8.92.60.692Pulmonary edema (27.1)5.91.90.693Reintubation (35.6)6.22.00.682Deep wound infection (10.2)5.92.00.661Wound dehiscence (11.3)2.00.80.643Prolonged ileus (14.3)3.91.90.61. . .Pulmonary embolism (44.0)0.60.30.601Myocardial infarction (42.2)1.10.60.59. . .Neurological deficits (29.9)1.70.80.59. . .Superficial wound infection (3.6)4.42.40.59. . .DVT/thrombophlebitis (12.0)0.90.50.58. . .CVA (27.6)0.90.50.58. . .Graft failure (9.2)1.00.70.5310&ast;DVT indicates deep vein thrombosis; CVA, cerebrovascular accident; and ellipses, not applicable.†Thirty-day mortality associated with occurrence of the complication, occurring by itself or in combination with other complications.‡All relationships were significant at P<.001.The last column of Table 3shows the entry step of albumin level in stepwise logistic regression using all 62 preoperative variables and an entry criterion of P<.05. In these multivariate analyses, albumin level was a relatively strong predictor of most of the complications, entering the models at the first, second, or third step. Albumin level was the first variable to enter for the major infection complications (systemic sepsis, pneumonia, deep wound infection). For several complications, albumin level was not a significant predictor in the regression analyses.The variation in the predictive ability of albumin level for different complications partially explains why albumin level was not as strong a predictor of operative morbidity for some types of surgery as it was for others in our analysis using a summary measure of morbidity (Table 2). For example, albumin level did not predict morbidity as well in orthopedic surgery as it did for all operations, and we found that most of the complications that were predicted best by albumin level for all operations in Table 3, such as systemic sepsis and failure to wean from ventilation, were proportionately underrepresented in orthopedic surgery, whereas several that were predicted less well by albumin level for all operations, such as pulmonary embolism and deep vein thrombosis/thrombophlebitis, were overrepresented in orthopedic surgery.Veterans who receive surgical care at VA hospitals are predominantly male and older, with higher rates of comorbid conditions than those undergoing surgery elsewhere. To demonstrate the applicability of the findings beyond VA patients, we performed separate analyses for a lower-risk segment of the sample and for women. All cases were identified in which the patient was younger than 70 years, had an ASA classification of 1 or 2 (healthy patient or mild disease), was independent in functional status, and did not report a weight loss of more than 10% within the last 6 months prior to surgery. There were 15,555 patients in our sample who met these criteria. The 30-day death rate for these low-risk patients was 0.32%. As shown in Figure 3, there was a negative association (c=0.72, P<.001) between serum albumin level and mortality even within this lower-risk subgroup. Likewise, for the women in the sample (n=1575), there was a strong association between serum albumin level and mortality (cindex=0.86, P<.001), with a graded increase in mortality rates as albumin values declined, similar to the finding for the total sample.Figure 3. Thirty-day mortality by preoperative serum albumin for low-risk cases (American Society of Anesthesiology class of 2 or lower, independent functional status, no weight loss of >10% in 6 months prior to surgery, and age younger than 70 years).COMMENTWe used a large national sample of VA patients to determine the association between preoperative serum albumin level and 30-day postoperative mortality and morbidity. We found large, graded increases in mortality and morbidity as albumin level declined from high to low levels. We found that for all major operations combined and for selected surgical subspecialties, serum albumin level was a strong predictor of mortality and morbidity independent of the effects of a large, diverse set of prospectively determined patient risk variables. The association between albumin concentration and operative mortality persisted in a subsample of patients who were classified as low risk on the basis of clinically recognizable risk factors. With regard to operative morbidity, we showed that albumin level was a better predictor of some types of complications than others.Although serum albumin level may also be affected by acute factors such as trauma and surgical stress, it is predictive of operative outcome because it is a marker of disease and malnutrition as well as possibly conferring a direct protective effect through several biological mechanisms.It is a better prognostic indicator than anthropomorphic markers of nutritional statusbecause of its ability to detect protein-energy malnutrition, which is not necessarily accompanied by lower body weight and may not be clinically recognizable, but is associated with significantly increased risk of morbidity and mortality.Protein-energy malnutrition results from increased protein and energy requirements associated with the stress of illness, injury, or infection.If the increased needs are not met from dietary or therapeutic sources, visceral protein stores are depleted, leading to abnormal function in organ systems, including gastrointestinal malabsorption, impaired immunologic response, and impaired production of albumin and other plasma proteins in the liver.However, albumin infusion usually is not an effective therapy because the albumin will degrade quickly and infusion does not address the underlying causes of adverse operative outcome.There was wide variation in the rate of preoperative serum albumin testing among the 44 participating hospitals, ranging from 20% to 98% of cases, with a median testing rate of 60%. Even if the analysis is restricted to cases with an ASA score of 3 or higher (severe systemic disease with functional limitation; 59% of all cases), the variation by hospital in the proportion tested remains large (from 27% to 97% of cases, with a median of 72%). In contrast, the ordering of blood counts and portions of the SMA-7 (eg, serum sodium and potassium) was consistently high (>90%) across hospitals.These findings suggest that preoperative serum albumin testing is underutilized by some surgeons. The cost of the test ($2-$4, based on cost data from 2 VA hospitals and 1 large community hospital) is low in relation to its prognostic value. In addition to its value as a prognostic tool, serum albumin level is a key element in accurate nutritional assessment,and it has been shown that nutritional therapy is more effective in reducing operative mortality and morbidity when the diagnosis of malnutrition is based on objective nutritional indices that include serum albumin rather than on subjective assessment.Particularly in patient populations with high rates of comorbidities, such as VA patients, it would seem that the test should be used more frequently as a prognostic tool and for detecting malnutrition.RPFergusonPO'ConnorBCrabtreeABatchelorJMitchellDCoppolaSerum albumin and prealbumin as predictors of clinical outcomes of hospitalized elderly nursing home residents.J Am Geriatr Soc.1993;41:545-549.FRHerrmannCSafranSELevkoffKLMinakerSerum albumin level on admission as a predictor of death, length of stay, and readmission.Arch Intern Med.1992;152:125-130.GFReinhardtJWMyscofskiDBWilkensPBDobrinJEManganRTStannardIncidence and mortality of hypoalbuminemic patients in hospitalized veterans.JPEN J Parenter Enter Nutr.1980;11:140-143.MWRichAJKellerKBSchechtmanWGMarshall JrNTKouchoukosIncreased complications and prolonged hospital stay in elderly cardiac surgical patients with low serum albumin.Am J Cardiol.1989;63:714-718.MCCortiJMGuralnikMESaliveJDSorkinSerum albumin level and physical disability as predictors of mortality in older persons.&jama;.1994;272:1036-1042.HKlonoff-CohenELBarrett-ConnorSLEdelsteinAlbumin levels as a predictor of mortality in the healthy elderly.J Clin Epidemiol.1992;45:207-212.CMJonesFBEatonPostoperative nutritional edema.Arch Surg.1933;27:159-177.JLMullenMHGertnerGPBuzbyGLGoodhartEFRosatoImplications of malnutrition in the surgical patient.Arch Surg.1979;114:121-125.GPBuzbyJLMullenDCMatthewsCLHobbsEFRosatoPrognostic nutritional index in gastrointestinal surgery.Am J Surg.1980;139:160-167.ASDetskyJPBakerKO'RourkePredicting nutrition-associated complications for patients undergoing gastrointestinal surgery.JPEN J Parenter Enter Nutr.1987;11:440-446.SFKhuriJDaleyWHendersonThe National VA Surgical Risk Study: risk adjustment for the comparative assessment of the quality of surgical care.J Am Coll Surg.1995;180:519-531.SFKhuriJDaleyWHendersonRisk adjustment of the postoperative mortality rate for the comparative assessment of the quality of surgical care.J Am Coll Surg.1997;185:315-327.JDaleySFKhuriWHendersonRisk adjustment of the postoperativemorbidity rate for the comparative assessment of the quality of surgical care.J Am Coll Surg.1997;185:328-340.JDaleyMGForbesGJYoungValidating risk-adjusted surgical outcomes: site visit assessment of process and structure.J Am Coll Surg.1997;185:341-351.JAHanleyBJMcNeilThe meaning and use of the area under a receiver operating characteristic (ROC) curve.Radiology.1982;143:29-36.PGoldwasserJFeldmanAssociation of serum albumin and mortality risk.J Clin Epidemiol.1997;50:693-703.JLMullenGPBuzbyMTWaldmanMHGertnerCLHobbsEFRosatoPrediction of operative morbidity and mortality by preoperative nutritional assessment.Surg Forum.1979;30:80-82.NAgarwalFAcevedoLSLeightonPredictive ability of various nutritional variables for mortality in elderly people.Am J Clin Nutr.1988;48:1173-1178.KNApelgrenJLRombeauPLTwomeyRAMillerComparison of nutritional indices and outcome in critically ill patients.Crit Care Med.1982;10:305-307.GLBlackburnKBHarveyNutritional assessment as a routine in clinical medicine.Postgrad Med.1982;71:46-63.GLBlackburnKBHarveyPrognostic strength of nutritional assessment.Progr Clin Biol Res.1981;77:689-697.DALipschitzProtein calorie malnutrition in the hospitalized elderly.Primary Care.1982;9:531-543.FNHomsyGLBlackburnModern parenteral and enteral nutrition in critical care.J Am Coll Nutr.1983;2:75-95.DALipschitzProtein-energy malnutrition.Hosp Pract.1988:87-99.SAMcClaveTEMitorajKAThielmeierRAGreenburgDifferentiating subtypes (hypoalbuminemic vs marasmic) of protein-calorie malnutrition: incidence and clinical significance in a university hospital setting.JPEN J Parenter Enter Nutr.1992;16:337-342.MARothschildMOratzSSSchreiberSerum albumin.Hepatology.1988;8:385-401.GPBuzbyWOWillifordOLPetersonA randomized clinical trial of total parenteral nutrition in malnourished surgical patients: the rationale and impact of previous clinical trials and pilot study on protocol design.Am J Clin Nutr.1988;47:357-365.JLMullenGPBuzbyDCMatthewsBFSmaleEFRosatoReduction of operative morbidity and mortality by combined preoperative and postoperative nutritional support.Ann Surg.1980;192:604-613.GPBuzbyGBlouinCLCollingPerioperative total parenteral nutrition in surgical patients.N Engl J Med.1991;325:525-532.This research was supported by a grant from the American Association for Clinical Chemistry Inc,Washington, DC, to the Institute for Health Services Research and Policy Studies at Northwestern University, Evanston, Ill; and by the Office of Patient Care Services and Cooperative Studies Program, US Department of Veterans Affairs Headquarters, Washington. Dr Daley is a Senior Research Associate in the Health Services Research and Development Program in the Department of Veterans Affairs.Reprints: James Gibbs, PhD, CSPCC (151K), Edward Hines, Jr, VA Hospital, PO Box 5000, Hines, IL 60141-5151 (e-mail: [email protected]).