Defining pneumonia after esophagectomy for cancer: validation of the Uniform Pneumonia Score in a high volume center in North America

Defining pneumonia after esophagectomy for cancer: validation of the Uniform Pneumonia Score in a... Summary Surgery is a central component of multimodality therapy for esophageal and gastroesophageal junction cancer. Pneumonia is a common sequela of esophagectomy, leading to an increase in intensive care unit stay, hospital stay, readmission rates, and postoperative mortality. Developing strategies to reduce pneumonia after esophagectomy is hampered by the absence of a standardized methodology for defining pneumonia. This study aims to validate the Uniform Pneumonia Score (UPS) in a high volume center in the USA. The UPS was developed to define pneumonia after esophagectomy for cancer and is based on the assessment of temperature (°C), leukocyte count (×109/L), and pulmonary radiography. The UPS has been validated utilizing a prospective, Institutional Review Board approved database of esophageal cancer patients treated in a high volume esophagectomy center in the USA between 2010 and 2015. One hundred ninety-three consecutive patients were included and 21 (10.9%) were treated for pneumonia. The UPS was able to predict treatment for suspected pneumonia with a good sensitivity (85.7%, confidence interval (CI): 63.7%–96.7%), specificity (97.1%, CI: 93.4%–99.1%), positive predictive value (78.3%, CI: 59.9%–89.7%), and negative predictive value (98.2%, CI: 95.1%–99.4%). The diagnostic accuracy was 95.9%, CI: 92.0%–98.2%. The UPS demonstrated to be a reliable scoring system to define pneumonia after esophagectomy for cancer. Global application of this model will standardize the definition of pneumonia after esophagectomy. This will improve outcome reporting and comparisons of complications between individual institutions, clinical trials, and national audits. INTRODUCTION Esophagectomy remains an important component of multimodality treatment for regional esophageal or gastroesophageal junction (GEJ) cancer.1 However, the associated morbidity and mortality can be significant.2–5 Pulmonary complications, primarily pneumonia, are frequently observed and may increase mortality and prolong intensive care unit (ICU) and hospital stay.6–8 This highlights the need to develop new strategies to reduce pulmonary complications after esophagectomy. Unfortunately, research and quality improvements in pulmonary complications after esophagectomy are hampered by the lack of a standardized methodology for defining postoperative pneumonia. In the current literature, 16 different, nonvalidated definitions are used for pneumonia, leading to a wide variation of reported pneumonia incidences (2% to 39%).8,9 The variation in definitions has made it impossible to assess and compare pneumonia-related outcomes across individual institutions, clinical trials, and national audits. The lack of a validated and internationally accepted and used definition for pneumonia after esophagectomy led to the initiative by van der Sluis et al.6 to create an objective and easy applicable scoring system to define pneumonia. This scoring system, the Utrecht Pneumonia Score, was revised and both internally (University Medical Center Utrecht) and externally (Catharina Hospital Eindhoven) validated, resulting in the Uniform Pneumonia Score (UPS) (Table 1).10 This study aims to validate the UPS in a geographically distinct cohort of esophagectomy patients from a high volume center in the USA. Table 1 Uniform pneumonia score Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 A sum score of 2 points or higher, of which at least 1 point is assigned due to infiltrative findings on pulmonary radiography indicates treatment of pneumonia. View Large Table 1 Uniform pneumonia score Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 A sum score of 2 points or higher, of which at least 1 point is assigned due to infiltrative findings on pulmonary radiography indicates treatment of pneumonia. View Large MATERIALS AND METHODS Patients and data Data of all consecutive patients who underwent an esophagectomy for resectable esophageal or GEJ cancer (Siewert I and Siewert II) in a tertiary referral center (Virginia Mason Medical Center, Seattle, WA) between January 2010 and January 2016 were prospectively entered into an Institutional Review Board (IRB) approved database. Database entries included standard patient characteristics, and intraoperative and postoperative data. UPS specific variables (temperature [°C], leukocyte count [×109/L] and pulmonary radiography findings) were retrospectively collected from the computerized medical records, similarly to the previous validation study.10 For pulmonary radiography findings, both chest X-rays and CT scans were taken into account when available. In patients treated with antibiotics following the diagnosis of pneumonia, temperature, leucocyte count, and pulmonary radiography findings were collected on the day of treatment initiation. In patients not treated for pneumonia, temperature, leucocyte count, and pulmonary radiography findings were collected from the medical records on the fourth postoperative day to ensure sufficient time from ICU discharge. This approach is in consistence with the data gathering in the previously published development and validation study.6,10 We postulate, that with this timepoint, operation- or ventilator-associated abnormalities were minimized on routine pulmonary radiography. If variables were not available on day 4 postoperatively, data were collected on the closest consecutive day on which the variable of interest was available. After exclusion of patients with missing values for model variables, complete case analysis was performed. This study was conducted in compliance with the Health Insurance Portability and Accountability Act (HIPAA) and received ethical approval from the local medical ethical committee (IRB number 17–072). The need for informed consent was waived. Patients underwent either an open transhiatal, Ivor–Lewis, or left thoracoabdominal esophagectomy with a two-field lymphadenectomy. During induction of general anesthesia, all patients received prophylactic 3 gram Unasyn (2 gram ampicillin + 1 gram sulbactam), which was repeated every 2 hours during surgery. Neoadjuvant chemotherapy regimens at Virginia Mason Hospital & Medical Center routinely consist of five cycles of either the combination of carboplatin and taxol or cisplatin and 5FU delivered with radiotherapy with a typical dose of 50.4 gray delivered in 28 fractions. All patients were managed according to the latest version of the Virginia Mason Medical Center esophagectomy (enhanced recovery) clinical pathway.11 Outcomes The primary outcomes were sensitivity, specificity, accuracy, positive predictive- and negative predictive value of the UPS. Furthermore, calibration and discriminatory ability were assessed. In line with previous studies (and in the absence of a well-defined gold standard for the diagnosis of pneumonia after esophagectomy), the outcome measure ‘pneumonia’ was defined as the clinical decision to initiate antibiotic treatment for suspected pneumonia.6,10 Secondary outcomes included pulmonary complications other than pneumonia, anastomotic leakage, chylothorax, recurrent laryngeal nerve injury, cardiac complications, length of hospital stay, and length of ICU stay.9 Statistical analysis Data were analyzed using the SPSS for Windows, version 22.0 (IBM corp., Armonk, New York) and the R 3.1.2 open-source software (http://www.R-project.org). Analysis and reporting were performed in accordance with the TRIPOD statement.12 Patient and treatment-related characteristics as well as postoperative outcomes besides pneumonia were compared between patients who were treated for pneumonia and patients who were not, in order to provide insight in potential differences. Categorical data were compared using the chi-square test or Fisher's exact test, as appropriate. Continuous data were compared using the Mann-Whitney-U test or the Student's T-test for nonparametric or parametric variables, respectively. The model performance of the UPS was assessed for discriminatory ability and calibration. Discrimination is the ability to distinguish a patient with the outcome from a patient without the outcome, which was assessed using the concordance (C) statistic extracted from receiver operating characteristic curve analysis.13 The agreement between the predicted probability of pneumonia by the model and the observed probability is indicated by calibration and was determined by visual inspection of calibration plots. RESULTS Patient and treatment characteristics Figure 1 shows the process of patient selection. A total of 210 consecutive patients underwent esophagectomy at Virginia Mason Medical Center between January 2010 and January 2016. Sixteen patients were excluded because they underwent esophagectomy for benign disease and 1 patient was excluded due to the missing UPS specific data. Hence, 193 patients were included for analyses (Fig. 1). Mean age of the included patients was 66.1 (±5.1) years, 60.6% of patients were classified as American Society of Anaesthesiologist (ASA) score III, 36.3% of the patients had a pT3 tumor, and 69.9% underwent neoadjuvant chemoradiation (Table 2). Fig. 1 View largeDownload slide Flowchart. Fig. 1 View largeDownload slide Flowchart. Table 2 Patient characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: ASA score, American Society of Anaesthesiologist score; BMI, body mass index (kg/m2); CTx, chemotherapy; CRTx, chemoradiotion; ECOG, Eastern Cooperative Oncology Group; F, female; M, male; n, number. View Large Table 2 Patient characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: ASA score, American Society of Anaesthesiologist score; BMI, body mass index (kg/m2); CTx, chemotherapy; CRTx, chemoradiotion; ECOG, Eastern Cooperative Oncology Group; F, female; M, male; n, number. View Large Across the 193 patients in this study, 21 (10.9%) underwent treatment for pneumonia. Median time (range) from surgery to initiation of antibiotic treatment was 5 (2–12) days. Patients with a pre-existing comorbidity of Chronich Obstructive Pulmonary Disease or a history of smoking had a significantly higher incidence of pneumonia (Table 2). All but one patient received a gastric conduit reconstruction after esophagectomy. Approximately half of the patients (48.2%) underwent an open Ivor–Lewis procedure (intrathoracic anastomosis) and the other half (48.2%) underwent an open left thoracoabdominal operation with a cervical anastomosis. The remaining 3.6% underwent an open transhiatal esophagectomy. An R0 resection was achieved in 97.4% of patients and the median (range) lymph node yield was 21 (5–49) (Table 3). Table 3 Surgical characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: OR, operation room. View Large Table 3 Surgical characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: OR, operation room. View Large Significantly more patients who were clinically diagnosed with pneumonia and consequently underwent antibiotic treatment were diagnosed with a pneumothorax (23.8% vs. 9.3% (P = 0.044)), pleural effusion requiring treatment (28.6% vs. 10.6% (P = 0.018)), reintubation (28.6% vs. 1.7% (P < 0.001)), and recurrent laryngeal nerve injury (4.8% vs. 0.0% (P = 0.004)) when compared to the patients who were not treated for pneumonia. Patient who developed pneumonia demonstrated a significant longer hospital stay (12 (7–112) days vs. 7 (5–43) days (P < 0.001)) (Table 4). One (0.5%) patient died within 30 days after surgery and three (1.6%) patients died within 90 days after surgery. The patients who died during hospitalization both had multiorgan failure after a complicated course including aspiration pneumonia. Table 4 Clinical outcomes Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Table showing the baseline data. For continuous variables data shown represent median (range), all other data are presented as numbers (percentages). Abbreviations: ICU, Intensive Care Unit; n.a. = not applicable. View Large Table 4 Clinical outcomes Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Table showing the baseline data. For continuous variables data shown represent median (range), all other data are presented as numbers (percentages). Abbreviations: ICU, Intensive Care Unit; n.a. = not applicable. View Large External validation Pneumonia defined by the UPS demonstrated a sensitivity of 85.7% (confidence interval (CI): 63.7%–96.7%) and a positive predictive value of 78.3% (CI: 59.9%–89.7%). The specificity was 97.1% (CI: 93.4%–99.1%) and the negative predictive value 98.2% (CI: 95.1%–99.4%) (Table 5). The diagnostic accuracy was 95.9% (CI: 92.0%–98.2%). Furthermore, both the discriminatory ability (C-statistic of 0.97) (Fig. 2) and calibration (Fig. 3) were very satisfactory. In patients not classified as being treated for pneumonia, a low observed probability of pneumonia was demonstrated in the calibration plot. The calibration plot also showed a high probability of pneumonia treatment in all patients with a positive UPS. Of the UPS negative patients, none of the T0L0P0-, T0L1P0-, and T1L0P0 patients and 4 out of 35 T0L0P1 patients underwent antibiotic treatment. In the UPS positive group, the majority of the T0L1P1 and TXLXP2 patients received antibiotic treatment. Only 1 patient was scored T1L0P1 and did not receive antibiotic treatment (Table 2b). Fig. 2 View largeDownload slide Discrimination. Abbreviations: AUC, area under the curve; NPV, negative predictive value, PPV, positive predictive value, sens, sensitivity, spec, specificity; UPS, uniform pneumonia score. Fig. 2 View largeDownload slide Discrimination. Abbreviations: AUC, area under the curve; NPV, negative predictive value, PPV, positive predictive value, sens, sensitivity, spec, specificity; UPS, uniform pneumonia score. Fig. 3 View largeDownload slide Calibration plot. Abbreviations: UPS, uniform pneumonia score. Fig. 3 View largeDownload slide Calibration plot. Abbreviations: UPS, uniform pneumonia score. Table 5 Diagnostic performance Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Sensitivity (85.7%, CI: 63.7%–96.7%), positive predictive value (78.3%, 59.9%–89.7%), specificity (97.1%, CI: 93.4%–99.1%), negative predictive value (98.2%, CI: 95.1%–99.4%), accuracy (95.9%, CI: 92.0%–98.2%). View Large Table 5 Diagnostic performance Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Sensitivity (85.7%, CI: 63.7%–96.7%), positive predictive value (78.3%, 59.9%–89.7%), specificity (97.1%, CI: 93.4%–99.1%), negative predictive value (98.2%, CI: 95.1%–99.4%), accuracy (95.9%, CI: 92.0%–98.2%). View Large DISCUSSION In accordance with a previously published external validation analysis, this study demonstrated the UPS to be an accurate dichotomous scoring system to define pneumonia after esophagectomy for esophageal and GEJ cancer.6,10 In the previous validation study, the UPS demonstrated a sensitivity of 79.1% and a specificity of 96.6% in the University Medical Center Utrecht and a sensitivity of 82.9% and specificity of 95.0% in the Catharina Hospital Eindhoven, both high volume esophagectomy centers in the Netherlands.10 To support general applicability of any scoring system, external validation, achieved by evaluating the model performance in differing clinical settings, is essential.14 This study showed that the UPS can be a valuable tool for investigating pneumonia in a distinct patient population, using a different surgical approach and postoperative clinical pathway. Moreover, the UPS was found to be an easy applicable scoring system, since the measured variables (temperature, leucocyte count, and pulmonary radiography findings) were routinely available as reflected by the fact that only one patient was excluded due to missing variables. The results of this study support the implementation and application of the UPS in research on outcomes after esophagectomy for cancer. The need for a uniform definition of postesophagectomy pneumonia has been demonstrated by a meta-analysis, which showed that between 2004 and 2009 in a total of 56 studies reporting on pneumonia after esophagectomy only 18 defined pneumonia using 16 different definitions. This has resulted in a wide range of pneumonia incidence (2–39%) in previous reports of esophagectomy outcomes. The authors demonstrated that the reporting of morbidity and mortality after esophagectomy for esophageal cancer lacks standard definitions, making it difficult to compare results, and concluded that uniform definitions for complications following esophagectomy should be developed.8 This led to the development of a specific scoring model to define pneumonia after esophagectomy in 2014 (Utrecht Pneumonia Score) and the initiative of the Esophagectomy Complications Consensus Group (ECCG) to standardize reporting of complications following esophagectomy.6,9 The results of the Delphi survey, published as the ‘International Consensus on Standardization of Data Collection for Complications Associated with Esophagectomy,’ are considered to be the standard guideline for reporting complications after esophagectomy and is now also used in nationwide registries.9,15 Following the agreement to develop a comprehensive complication list, and in an effort to avoid confusion regarding already existing definitions, the participating centers decided to link many complications, including pneumonia, to standard definitions. Pneumonia was defined by an internationally accepted definition, published by the American Thoracic Society and the Center for Disease Control and Prevention.9,16 However, these guidelines focused on ventilator-associated pneumonia (VAP), defining pneumonia according to the clinical pulmonary infection score (CPIS). The CPIS has been developed in patients with VAP, without validation in patients with pneumonia.17–19 Moreover, it has been demonstrated that the CPIS is not generally applicable, since its validation in trauma patients failed.20 Hence, a general and widely accepted definition for pneumonia after esophagectomy was still lacking. To fill the gap, the Utrecht Pneumonia Score was revised, and an internal and external validation of the revised score led to the introduction of the UPS in 2016.10 The UPS consists out of three variables: leukocytes, temperature, and pulmonary radiography. Sputum culture is not included in the UPS since, as previously demonstrated, a positive sputum culture was not significant associated with the clinical diagnosis of pneumonia in multivariate analysis. Furthermore, in approximately a quarter of patients treated for pneumonia sputum culture results are not available, making it an insufficient variable for defining pneumonia in retrospective research.6 Since pneumonia is the most frequently observed complication after esophagectomy for cancer, it often serves as the primary outcome of comparative surgical studies.8 The UPS has already been utilized as a primary outcome measure in multiple retrospective and prospective studies.21–26 However, several landmark trials in which pneumonia was the primary outcome measure and on which current practice is based, like the TIME trial27 and the MIRO trial,2,28 have utilized different and nonvalidated definitions. The UPS has proven to be an accurate definition for pneumonia after esophagectomy in different centers within the Netherlands. It is a dichotomous scoring system, which can be applied easily in clinical practice. However, it only facilitates quantifying the diagnosis of pneumonia. Since it does not stratify on the basis of severity, grading of pneumonia severity may be scored using the Clavien–Dindo classification of surgical complications.29,30 In this study, the UPS was validated in a geographically distinct, external cohort in the USA.10 Esophagectomies at this North American center were performed open and peri- and postoperative care was guided, in contrast to the previous validation study, by an enhanced recovery after surgery (ERAS) protocol.11 The ERAS protocol aims to optimize perioperative care by minimizing surgical stress and complications and accelerate recovery.31 This may be an important reason why the incidence of pneumonia in this study was remarkably lower (11.9%) when compared to pneumonia rates in the previous validation study (36% in the University Medical Center Utrecht and 40% in the Catharina Hospital Eindhoven).10 Also, in the present cohort, the UPS was able to provide a reasonably accurate definition for pneumonia after esophagectomy, which shows its potential for broader application. Standardized definitions for postoperative complications will increase accuracy of future research and quality initiatives. The strength of this study is the evaluation of a prospective dataset without loss to follow-up, containing detailed information on postoperative complications from a high-volume esophageal cancer care center. Furthermore, the ease of score-variable collection is demonstrated, facilitating the wide implementation of the UPS in both retrospective and prospective studies. Since UPS-specific variables are collected routinely on the fourth postoperative day in most hospitals, also nationwide registries can make use of the UPS. Despite this, certain limitations apply to the current analysis. The gold standard for diagnosing pneumonia in this study was the clinical decision to initiate treatment for suspected pneumonia. This subjective definition has been used in previous studies and constitutes the only definition in which all factors are accounted for by the attending physician. Since there is no current gold standard for the definition of postesophagectomy pneumonia, this definition remains the best standard to validate the UPS, which is designed to fill this gap. Though the variables included in the UPS are of substantial importance in the clinical decision making process to initiate antibiotic treatment for pneumonia, many other factors contribute to the clinical diagnosis. However, not all factors can be included in an objective scoring system, which is applicable for research. This validation study demonstrates that the model variables used in the UPS have proven to be an adequate representation of all factors taken into account by the attending physician. In this study, model variables in patients who were not treated for pneumonia were collected on the fourth postoperative day, which is in accordance with the previous validation- and development study. This might have led to a small underestimation of the number of false positives (patients who had a positive UPS after postoperative day 4, but were not treated with antibiotics). However, from a methodological point of view, the same methodology as presented in the development study should be used to adequately validate a scoring system. Furthermore, the incidence of pneumonia was 10.9% in this study, hence relatively few events were observed. However, for external validation studies, formal sample size calculations for the number of events based on statistical power considerations are not well investigated.14 Within the context of these limitations, this study showed the UPS to be an accurate, objective, and easy to implement scoring system to define pneumonia after esophagectomy in esophageal cancer research. Therefore, we encourage the application of the UPS in future esophageal cancer research and nationwide registries reporting on outcomes after esophagectomy. Notes Specific author contributions: Conception and design: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum, Teus J. Weijs, Jelle P. Ruurda, Richard van Hillegersberg, Donald E. Low; Administrative support: Maarten F. J. Seesing, Andrea Wirsching; Provision of study materials or patients: Donald E. Low; Collection and assembly of data: Maarten F. J. Seesing, Andrea Wirsching; Data analysis and interpretation: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum; Manuscript writing: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum; Teus J. Weijs, Jelle P. Ruurda, Richard van Hillegersberg, Donald E. Low; Final approval of manuscript: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum; Teus J. Weijs, Jelle P. Ruurda, Richard van Hillegersberg, Donald E. Low. References 1 van Hagen P , Hulshof M C , van Lanschot J J et al. Preoperative chemoradiotherapy for esophageal or junctional cancer . N Engl J Med 2012 ; 366 : 2074 – 84 . Google Scholar CrossRef Search ADS PubMed 2 Mariette C , Robb W B . Open or minimally invasive resection for oesophageal cancer? Recent Results Cancer Res 2012 ; 196 : 155 – 67 . Google Scholar CrossRef Search ADS PubMed 3 Lerut T , Moons J , Coosemans W et al. Postoperative complications after transthoracic esophagectomy for cancer of the esophagus and gastroesophageal junction are correlated with early cancer recurrence: role of systematic grading of complications using the modified Clavien classification . Ann Surg 2009 ; 250 : 798 – 807 . Google Scholar CrossRef Search ADS PubMed 4 Goldberg R F , Bowers S P , Parker M et al. Technical and perioperative outcomes of minimally invasive esophagectomy in the prone position . Surg Endosc 2013 ; 27 : 553 – 7 . Google Scholar CrossRef Search ADS PubMed 5 Seesing M F J , Gisbertz S S , Goense L et al. A propensity score matched analysis of open versus minimally invasive transthoracic esophagectomy in the Netherlands . Ann Surg 2017; 266: 839 . 6 van der Sluis P C , Verhage R J , van der Horst S et al. A new clinical scoring system to define pneumonia following esophagectomy for cancer . Dig Surg 2014 ; 31 : 108 – 16 . Google Scholar CrossRef Search ADS PubMed 7 Markar S , Gronnier C , Duhamel A et al. Pattern of postoperative mortality after esophageal cancer resection according to center volume: results from a large European multicenter study . Ann Surg Oncol 2015 ; 22 : 2615 – 23 . Google Scholar CrossRef Search ADS PubMed 8 Blencowe N S , Strong S , McNair A G et al. Reporting of short-term clinical outcomes after esophagectomy: a systematic review . Ann Surg 2012 ; 255 : 658 – 66 . Google Scholar CrossRef Search ADS PubMed 9 Low D E , Alderson D , Cecconello I et al. International consensus on standardization of data collection for complications associated with esophagectomy: Esophagectomy Complications Consensus Group (ECCG) . Ann Surg 2015 ; 262 : 286 – 94 . Google Scholar CrossRef Search ADS PubMed 10 Weijs T J , Seesing M F , van Rossum P S et al. Internal and external validation of a multivariable model to define hospital-acquired pneumonia after esophagectomy . J Gastrointest Surg 2016 ; 20 : 680 – 7 . Google Scholar CrossRef Search ADS PubMed 11 Markar S R , Schmidt H , Kunz S et al. Evolution of standardized clinical pathways: refining multidisciplinary care and process to improve outcomes of the surgical treatment of esophageal cancer . J Gastrointest Surg 2014 ; 18 : 1238 – 46 . Google Scholar CrossRef Search ADS PubMed 12 Collins G S , Reitsma J B , Altman D G et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement . Ann Intern Med 2015 ; 162 : 55 – 63 . Google Scholar CrossRef Search ADS PubMed 13 Steyerberg E W , Vickers A J , Cook N R et al. Assessing the performance of prediction models: a framework for traditional and novel measures . Epidemiology 2010 ; 21 : 128 – 38 . Google Scholar CrossRef Search ADS PubMed 14 Han K , Song K , Choi B W . How to develop, validate, and compare clinical prediction models involving radiological parameters: study design and statistical methods . Korean J Radiol 2016 ; 17 : 339 – 50 . Google Scholar CrossRef Search ADS PubMed 15 Busweiler L A , Wijnhoven B P , van Berge Henegouwen M I et al. Early outcomes from the Dutch Upper Gastrointestinal Cancer Audit . Br J Surg 2016 ; 103 : 1855 – 63 . Google Scholar CrossRef Search ADS PubMed 16 American Thoracic Society, Centers for Disease Control and Prevention, Infectious Diseases Society of America . American Thoracic Society/Centers for Disease Control and Prevention/Infectious Diseases Society of America: controlling tuberculosis in the United States . Am J Respir Crit Care Med 2005 ; 172 : 1169 – 227 . CrossRef Search ADS PubMed 17 Rosbolt M B , Sterling E S , Fahy B G . The utility of the clinical pulmonary infection score . J Intensive Care Med 2009 ; 24 : 26 – 34 . Google Scholar CrossRef Search ADS PubMed 18 Pugin J , Auckenthaler R , Mili N et al. Diagnosis of ventilator-associated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic “blind” bronchoalveolar lavage fluid . Am Rev Respir Dis 1991 ; 143 : 1121 – 9 . Google Scholar CrossRef Search ADS PubMed 19 Singh N , Rogers P , Atwood C W et al. Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit. A proposed solution for indiscriminate antibiotic prescription . Am J Respir Crit Care Med 2000 ; 162 : 505 – 11 . Google Scholar CrossRef Search ADS PubMed 20 Pham T N , Cancio L C , Gibran N S et al. American Burn Association practice guidelines burn shock resuscitation . J Burn Care Res 2008 ; 29 : 257 – 66 . Google Scholar CrossRef Search ADS PubMed 21 Weijs T J , Berkelmans G H , Nieuwenhuijzen G A et al. Immediate postoperative oral nutrition following esophagectomy: a multicenter clinical trial . Ann Thorac Surg 2016 ; 102 : 1141 – 8 . Google Scholar CrossRef Search ADS PubMed 22 van der Sluis P C , Ruurda J P , van der Horst S et al. Robot-assisted minimally invasive thoracolaparoscopic esophagectomy versus open transthoracic esophagectomy for resectable esophageal cancer, a randomized controlled trial (ROBOT trial) . Trials 2012 ; 13 : 230–6215–13–230 . Google Scholar CrossRef Search ADS 23 Yamana I , Takeno S , Hashimoto T et al. Randomized controlled study to evaluate the efficacy of a preoperative respiratory rehabilitation program to prevent postoperative pulmonary complications after esophagectomy . Dig Surg 2015 ; 32 : 331 – 7 . Google Scholar CrossRef Search ADS PubMed 24 Valkenet K , Trappenburg J C , Gosselink R et al. Preoperative inspiratory muscle training to prevent postoperative pulmonary complications in patients undergoing esophageal resection (PREPARE study): study protocol for a randomized controlled trial . Trials 2014 ; 15 : 144–6215–15–144 . Google Scholar CrossRef Search ADS 25 Scarpa M , Cavallin F , Saadeh L M et al. Hybrid minimally invasive esophagectomy for cancer: impact on postoperative inflammatory and nutritional status . Dis Esophagus 2016 ; 29 : 1064 – 70 . Google Scholar CrossRef Search ADS PubMed 26 Berkelmans G H , Wilts B J , Kouwenhoven E A et al. Nutritional route in oesophageal resection trial II (NUTRIENT II): study protocol for a multicentre open-label randomised controlled trial . BMJ Open 2016 ; 6 : e011979–2016–011979 . Google Scholar CrossRef Search ADS 27 Biere S S , van Berge Henegouwen M I , Maas K W et al. Minimally invasive versus open oesophagectomy for patients with oesophageal cancer: a multicentre, open-label, randomised controlled trial . Lancet 2012 ; 379 : 1887 – 92 . Google Scholar CrossRef Search ADS PubMed 28 Briez N , Piessen G , Bonnetain F et al. Open versus laparoscopically-assisted oesophagectomy for cancer: a multicentre randomised controlled phase III trial—the MIRO trial . BMC Cancer 2011 ; 11 : 310–2407–11–310 . Google Scholar CrossRef Search ADS 29 Dindo D , Demartines N , Clavien P A . Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey . Ann Surg 2004 ; 240 : 205 – 13 . Google Scholar CrossRef Search ADS PubMed 30 Clavien P A , Strasberg S M . Severity grading of surgical complications . Ann Surg 2009 ; 250 : 197 – 8 . Google Scholar CrossRef Search ADS PubMed 31 Findlay J M , Gillies R S , Millo J et al. Enhanced recovery for esophagectomy: a systematic review and evidence-based guidelines . Ann Surg 2014 ; 259 : 413 – 31 . Google Scholar CrossRef Search ADS PubMed © The Authors 2018. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Diseases of the Esophagus Oxford University Press

Defining pneumonia after esophagectomy for cancer: validation of the Uniform Pneumonia Score in a high volume center in North America

Loading next page...
 
/lp/ou_press/defining-pneumonia-after-esophagectomy-for-cancer-validation-of-the-Lymt9fzj8b
Publisher
Oxford University Press
Copyright
© The Authors 2018. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus.
ISSN
1120-8694
eISSN
1442-2050
D.O.I.
10.1093/dote/doy002
Publisher site
See Article on Publisher Site

Abstract

Summary Surgery is a central component of multimodality therapy for esophageal and gastroesophageal junction cancer. Pneumonia is a common sequela of esophagectomy, leading to an increase in intensive care unit stay, hospital stay, readmission rates, and postoperative mortality. Developing strategies to reduce pneumonia after esophagectomy is hampered by the absence of a standardized methodology for defining pneumonia. This study aims to validate the Uniform Pneumonia Score (UPS) in a high volume center in the USA. The UPS was developed to define pneumonia after esophagectomy for cancer and is based on the assessment of temperature (°C), leukocyte count (×109/L), and pulmonary radiography. The UPS has been validated utilizing a prospective, Institutional Review Board approved database of esophageal cancer patients treated in a high volume esophagectomy center in the USA between 2010 and 2015. One hundred ninety-three consecutive patients were included and 21 (10.9%) were treated for pneumonia. The UPS was able to predict treatment for suspected pneumonia with a good sensitivity (85.7%, confidence interval (CI): 63.7%–96.7%), specificity (97.1%, CI: 93.4%–99.1%), positive predictive value (78.3%, CI: 59.9%–89.7%), and negative predictive value (98.2%, CI: 95.1%–99.4%). The diagnostic accuracy was 95.9%, CI: 92.0%–98.2%. The UPS demonstrated to be a reliable scoring system to define pneumonia after esophagectomy for cancer. Global application of this model will standardize the definition of pneumonia after esophagectomy. This will improve outcome reporting and comparisons of complications between individual institutions, clinical trials, and national audits. INTRODUCTION Esophagectomy remains an important component of multimodality treatment for regional esophageal or gastroesophageal junction (GEJ) cancer.1 However, the associated morbidity and mortality can be significant.2–5 Pulmonary complications, primarily pneumonia, are frequently observed and may increase mortality and prolong intensive care unit (ICU) and hospital stay.6–8 This highlights the need to develop new strategies to reduce pulmonary complications after esophagectomy. Unfortunately, research and quality improvements in pulmonary complications after esophagectomy are hampered by the lack of a standardized methodology for defining postoperative pneumonia. In the current literature, 16 different, nonvalidated definitions are used for pneumonia, leading to a wide variation of reported pneumonia incidences (2% to 39%).8,9 The variation in definitions has made it impossible to assess and compare pneumonia-related outcomes across individual institutions, clinical trials, and national audits. The lack of a validated and internationally accepted and used definition for pneumonia after esophagectomy led to the initiative by van der Sluis et al.6 to create an objective and easy applicable scoring system to define pneumonia. This scoring system, the Utrecht Pneumonia Score, was revised and both internally (University Medical Center Utrecht) and externally (Catharina Hospital Eindhoven) validated, resulting in the Uniform Pneumonia Score (UPS) (Table 1).10 This study aims to validate the UPS in a geographically distinct cohort of esophagectomy patients from a high volume center in the USA. Table 1 Uniform pneumonia score Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 A sum score of 2 points or higher, of which at least 1 point is assigned due to infiltrative findings on pulmonary radiography indicates treatment of pneumonia. View Large Table 1 Uniform pneumonia score Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 Uniform pneumonia score Diagnostic determinant Range Score Temperature [°C] ≥36.1 and ≤38.4 0 ≤36.0 and ≥38.5 1 Leukocyte count [×109/L] ≥4.0 and ≤11.0 0 <4.0 or >11.0 1 Pulmonary radiography No infiltrate 0 Diffused (or patchy) infiltrate 1 Well-circumscribed infiltrate 2 A sum score of 2 points or higher, of which at least 1 point is assigned due to infiltrative findings on pulmonary radiography indicates treatment of pneumonia. View Large MATERIALS AND METHODS Patients and data Data of all consecutive patients who underwent an esophagectomy for resectable esophageal or GEJ cancer (Siewert I and Siewert II) in a tertiary referral center (Virginia Mason Medical Center, Seattle, WA) between January 2010 and January 2016 were prospectively entered into an Institutional Review Board (IRB) approved database. Database entries included standard patient characteristics, and intraoperative and postoperative data. UPS specific variables (temperature [°C], leukocyte count [×109/L] and pulmonary radiography findings) were retrospectively collected from the computerized medical records, similarly to the previous validation study.10 For pulmonary radiography findings, both chest X-rays and CT scans were taken into account when available. In patients treated with antibiotics following the diagnosis of pneumonia, temperature, leucocyte count, and pulmonary radiography findings were collected on the day of treatment initiation. In patients not treated for pneumonia, temperature, leucocyte count, and pulmonary radiography findings were collected from the medical records on the fourth postoperative day to ensure sufficient time from ICU discharge. This approach is in consistence with the data gathering in the previously published development and validation study.6,10 We postulate, that with this timepoint, operation- or ventilator-associated abnormalities were minimized on routine pulmonary radiography. If variables were not available on day 4 postoperatively, data were collected on the closest consecutive day on which the variable of interest was available. After exclusion of patients with missing values for model variables, complete case analysis was performed. This study was conducted in compliance with the Health Insurance Portability and Accountability Act (HIPAA) and received ethical approval from the local medical ethical committee (IRB number 17–072). The need for informed consent was waived. Patients underwent either an open transhiatal, Ivor–Lewis, or left thoracoabdominal esophagectomy with a two-field lymphadenectomy. During induction of general anesthesia, all patients received prophylactic 3 gram Unasyn (2 gram ampicillin + 1 gram sulbactam), which was repeated every 2 hours during surgery. Neoadjuvant chemotherapy regimens at Virginia Mason Hospital & Medical Center routinely consist of five cycles of either the combination of carboplatin and taxol or cisplatin and 5FU delivered with radiotherapy with a typical dose of 50.4 gray delivered in 28 fractions. All patients were managed according to the latest version of the Virginia Mason Medical Center esophagectomy (enhanced recovery) clinical pathway.11 Outcomes The primary outcomes were sensitivity, specificity, accuracy, positive predictive- and negative predictive value of the UPS. Furthermore, calibration and discriminatory ability were assessed. In line with previous studies (and in the absence of a well-defined gold standard for the diagnosis of pneumonia after esophagectomy), the outcome measure ‘pneumonia’ was defined as the clinical decision to initiate antibiotic treatment for suspected pneumonia.6,10 Secondary outcomes included pulmonary complications other than pneumonia, anastomotic leakage, chylothorax, recurrent laryngeal nerve injury, cardiac complications, length of hospital stay, and length of ICU stay.9 Statistical analysis Data were analyzed using the SPSS for Windows, version 22.0 (IBM corp., Armonk, New York) and the R 3.1.2 open-source software (http://www.R-project.org). Analysis and reporting were performed in accordance with the TRIPOD statement.12 Patient and treatment-related characteristics as well as postoperative outcomes besides pneumonia were compared between patients who were treated for pneumonia and patients who were not, in order to provide insight in potential differences. Categorical data were compared using the chi-square test or Fisher's exact test, as appropriate. Continuous data were compared using the Mann-Whitney-U test or the Student's T-test for nonparametric or parametric variables, respectively. The model performance of the UPS was assessed for discriminatory ability and calibration. Discrimination is the ability to distinguish a patient with the outcome from a patient without the outcome, which was assessed using the concordance (C) statistic extracted from receiver operating characteristic curve analysis.13 The agreement between the predicted probability of pneumonia by the model and the observed probability is indicated by calibration and was determined by visual inspection of calibration plots. RESULTS Patient and treatment characteristics Figure 1 shows the process of patient selection. A total of 210 consecutive patients underwent esophagectomy at Virginia Mason Medical Center between January 2010 and January 2016. Sixteen patients were excluded because they underwent esophagectomy for benign disease and 1 patient was excluded due to the missing UPS specific data. Hence, 193 patients were included for analyses (Fig. 1). Mean age of the included patients was 66.1 (±5.1) years, 60.6% of patients were classified as American Society of Anaesthesiologist (ASA) score III, 36.3% of the patients had a pT3 tumor, and 69.9% underwent neoadjuvant chemoradiation (Table 2). Fig. 1 View largeDownload slide Flowchart. Fig. 1 View largeDownload slide Flowchart. Table 2 Patient characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: ASA score, American Society of Anaesthesiologist score; BMI, body mass index (kg/m2); CTx, chemotherapy; CRTx, chemoradiotion; ECOG, Eastern Cooperative Oncology Group; F, female; M, male; n, number. View Large Table 2 Patient characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Gender  Female 41 (21.4) 4 (19.0) 37 (21.5) 0.794  Male 152 (78.6) 17 (81.0) 135 (78.5) Age 66.1 ± 9.6 65 ± 11.2 66 ± 9.4 0.256 BMI 27.7 ± 5.1 27.1 ± 6.4 27.8 ± 5.0 0.338 ECOG score  O 94 (48.7) 9 (42.9) 85 (49.4) 0.559  I 84 (43.5) 10 (47.6) 74 (43.0)  II 14 (7.3) 2 (9.5) 12 (6.9)  III 1 (0.5) 0 (0) 1 (0.6) ASA score  II 76 (39.4) 7 (33.3) 69 (40.1) 0.549  III 117 (60.6) 14 (66.7) 103 (59.9) Alcohol user 174 (90.2) 19 (90.5) 155 (90.1) 0.980 Nicotine user  Current 30 (15.5) 9 (42.9) 21 (12.2) <0.001  Previous 121 (62.7) 12 (57.1) 109 (63.4)  Never 42 (21.8) 0 (0.0) 42 (24.4) Comorbidity  Cardial 44 (22.8) 5 (23.8) 39 (22.7) 0.907  Vascular 109 (56.5) 11 (52.4) 98 (57.6) 0.688  Diabetes 43 (22.8) 6 (28.6) 37 (21.5) 0.318  Asthma 17 (8.8) 1 (4.8) 16 (9.3) 0.488  COPD 20 (10.3) 5 (23.8) 15 (8.8) 0.032 Pathological T-stage  Complete pathological response 42 (21.8) 5 (23.8) 37 (21.5) 0.312  Tis 1 (0.5) 0 (0) 1 (0.6)  T1 43 (22.3) 6 (28.6) 37 (21.5)  T2 35 (18.1) 5 (23.8) 30 (17.4)  T3 70 (36.3) 5 (23.8) 65 (37.8)  T4 1 (0.5) 0 (0) 1 (0.6)  Tx 1 (0.5) 0 (0) 1 (0.6) Pathological N-stage  N0 106 (54.9) 14 (66.7) 92 (53.5) 0.172  N1 44 (22.8) 5 (23.8) 39 (22.7)  N2 29 (15.0) 2 (9.5) 27 (15.7)  N3 12 (6.2) 0(0) 12 (7.0)  Nx 2 (1.1) 0(0) 2 (1.1) Histology  Adenocarcinoma 168 (85.7) 19 (90.5) 146 (84.9) 0.746  Squamous cell carcinoma 26 (13.3) 2 (9.5) 24 (14.0)  Other 2 (1.0) 0 (0.0) 2 (1.2) Neoadjuvant therapy  None 53 (27.5) 7 (33.3) 46 (26.7) 0.624  CTx 5 (2.6) 1 (4.7) 4 (2.3)  CRTx 135 (69.9) 13 (62.0) 122 (71.0) Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: ASA score, American Society of Anaesthesiologist score; BMI, body mass index (kg/m2); CTx, chemotherapy; CRTx, chemoradiotion; ECOG, Eastern Cooperative Oncology Group; F, female; M, male; n, number. View Large Across the 193 patients in this study, 21 (10.9%) underwent treatment for pneumonia. Median time (range) from surgery to initiation of antibiotic treatment was 5 (2–12) days. Patients with a pre-existing comorbidity of Chronich Obstructive Pulmonary Disease or a history of smoking had a significantly higher incidence of pneumonia (Table 2). All but one patient received a gastric conduit reconstruction after esophagectomy. Approximately half of the patients (48.2%) underwent an open Ivor–Lewis procedure (intrathoracic anastomosis) and the other half (48.2%) underwent an open left thoracoabdominal operation with a cervical anastomosis. The remaining 3.6% underwent an open transhiatal esophagectomy. An R0 resection was achieved in 97.4% of patients and the median (range) lymph node yield was 21 (5–49) (Table 3). Table 3 Surgical characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: OR, operation room. View Large Table 3 Surgical characteristics Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Surgical approach Ivor–Lewis 93 (48.2) 9 (42.9) 84 (48.8) 0.410 Left thoracoabdominal 92 (48.2) 10 (47.6) 82 (47.6) Transhiatal 8 (3.6) 2 (9.5) 6 (3.5) Gastric conduit reconstruction 192 (99.5) 1 (4.8) 0 (0.0) 0.004 Jejumostomy 140 (72.5) 17 (81.0) 123 (71.5) 0.360 OR time (min) 401 (244–664) 446 (344–615) 401 (244–664) 0.110 Estimated blood loss (cc) 150 (50–700) 200 (75–500) 150 (50–700) 0.081 Radicality R0 188 (97.4) 21 (100.0) 167 (97.1) 0.429 R1 5 (2.6) 0 (0.0) 5 (2.9) Lymph node yield 21 (5–49) 22 (10–44) 21 (5–49) 0.446 Table showing the baseline data. For continuous variables data shown represent median (interquartile range), all other data are presented as numbers (percentages). Abbreviations: OR, operation room. View Large Significantly more patients who were clinically diagnosed with pneumonia and consequently underwent antibiotic treatment were diagnosed with a pneumothorax (23.8% vs. 9.3% (P = 0.044)), pleural effusion requiring treatment (28.6% vs. 10.6% (P = 0.018)), reintubation (28.6% vs. 1.7% (P < 0.001)), and recurrent laryngeal nerve injury (4.8% vs. 0.0% (P = 0.004)) when compared to the patients who were not treated for pneumonia. Patient who developed pneumonia demonstrated a significant longer hospital stay (12 (7–112) days vs. 7 (5–43) days (P < 0.001)) (Table 4). One (0.5%) patient died within 30 days after surgery and three (1.6%) patients died within 90 days after surgery. The patients who died during hospitalization both had multiorgan failure after a complicated course including aspiration pneumonia. Table 4 Clinical outcomes Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Table showing the baseline data. For continuous variables data shown represent median (range), all other data are presented as numbers (percentages). Abbreviations: ICU, Intensive Care Unit; n.a. = not applicable. View Large Table 4 Clinical outcomes Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Total Antibiotic treatment for pneumonia (n = 193) (%) Yes (n = 21) (%) No (n = 172) (%) P Postoperative complications total 108 (56.0) n.a. n.a. n.a. Pneumothorax 21 (10.9) 5 (23.8) 16 (9.3) 0.044 Pleural effusion requiring treatment 24 (12.4) 6 (28.6) 18 (10.5) 0.018 Reintubation 9 (4.7) 6 (28.6) 3 (1.7) <0.001 Anastomotic leakage 16 (8.3) 3 (14.3) 13 (7.6) 0.291 Chyle leakage 9 (4.7) 1 (4.8) 8 (4.7) 0.982 Recurrent laryngeal nerve injury 1 (0.5) 1 (4.8) 0 (0) 0.004 Atrial fibrillation 44 (22.8) 7 (33.3) 37 (21.5) 0.223 Hospital stay 7 (5–115) 12 (7–115) 7 (5–43) <0.001 ICU stay 1 (0–22) 1 (1–15) 1 (0–22) 0.234 Table showing the baseline data. For continuous variables data shown represent median (range), all other data are presented as numbers (percentages). Abbreviations: ICU, Intensive Care Unit; n.a. = not applicable. View Large External validation Pneumonia defined by the UPS demonstrated a sensitivity of 85.7% (confidence interval (CI): 63.7%–96.7%) and a positive predictive value of 78.3% (CI: 59.9%–89.7%). The specificity was 97.1% (CI: 93.4%–99.1%) and the negative predictive value 98.2% (CI: 95.1%–99.4%) (Table 5). The diagnostic accuracy was 95.9% (CI: 92.0%–98.2%). Furthermore, both the discriminatory ability (C-statistic of 0.97) (Fig. 2) and calibration (Fig. 3) were very satisfactory. In patients not classified as being treated for pneumonia, a low observed probability of pneumonia was demonstrated in the calibration plot. The calibration plot also showed a high probability of pneumonia treatment in all patients with a positive UPS. Of the UPS negative patients, none of the T0L0P0-, T0L1P0-, and T1L0P0 patients and 4 out of 35 T0L0P1 patients underwent antibiotic treatment. In the UPS positive group, the majority of the T0L1P1 and TXLXP2 patients received antibiotic treatment. Only 1 patient was scored T1L0P1 and did not receive antibiotic treatment (Table 2b). Fig. 2 View largeDownload slide Discrimination. Abbreviations: AUC, area under the curve; NPV, negative predictive value, PPV, positive predictive value, sens, sensitivity, spec, specificity; UPS, uniform pneumonia score. Fig. 2 View largeDownload slide Discrimination. Abbreviations: AUC, area under the curve; NPV, negative predictive value, PPV, positive predictive value, sens, sensitivity, spec, specificity; UPS, uniform pneumonia score. Fig. 3 View largeDownload slide Calibration plot. Abbreviations: UPS, uniform pneumonia score. Fig. 3 View largeDownload slide Calibration plot. Abbreviations: UPS, uniform pneumonia score. Table 5 Diagnostic performance Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Sensitivity (85.7%, CI: 63.7%–96.7%), positive predictive value (78.3%, 59.9%–89.7%), specificity (97.1%, CI: 93.4%–99.1%), negative predictive value (98.2%, CI: 95.1%–99.4%), accuracy (95.9%, CI: 92.0%–98.2%). View Large Table 5 Diagnostic performance Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Antibiotic treatment for pneumonia Yes (n = 21) No (n = 172) Uniform pneumonia score Positive 18 (85.7%) 5 (2.9%) Negative 3 (14.3%) 167 (97.1%) Sensitivity (85.7%, CI: 63.7%–96.7%), positive predictive value (78.3%, 59.9%–89.7%), specificity (97.1%, CI: 93.4%–99.1%), negative predictive value (98.2%, CI: 95.1%–99.4%), accuracy (95.9%, CI: 92.0%–98.2%). View Large DISCUSSION In accordance with a previously published external validation analysis, this study demonstrated the UPS to be an accurate dichotomous scoring system to define pneumonia after esophagectomy for esophageal and GEJ cancer.6,10 In the previous validation study, the UPS demonstrated a sensitivity of 79.1% and a specificity of 96.6% in the University Medical Center Utrecht and a sensitivity of 82.9% and specificity of 95.0% in the Catharina Hospital Eindhoven, both high volume esophagectomy centers in the Netherlands.10 To support general applicability of any scoring system, external validation, achieved by evaluating the model performance in differing clinical settings, is essential.14 This study showed that the UPS can be a valuable tool for investigating pneumonia in a distinct patient population, using a different surgical approach and postoperative clinical pathway. Moreover, the UPS was found to be an easy applicable scoring system, since the measured variables (temperature, leucocyte count, and pulmonary radiography findings) were routinely available as reflected by the fact that only one patient was excluded due to missing variables. The results of this study support the implementation and application of the UPS in research on outcomes after esophagectomy for cancer. The need for a uniform definition of postesophagectomy pneumonia has been demonstrated by a meta-analysis, which showed that between 2004 and 2009 in a total of 56 studies reporting on pneumonia after esophagectomy only 18 defined pneumonia using 16 different definitions. This has resulted in a wide range of pneumonia incidence (2–39%) in previous reports of esophagectomy outcomes. The authors demonstrated that the reporting of morbidity and mortality after esophagectomy for esophageal cancer lacks standard definitions, making it difficult to compare results, and concluded that uniform definitions for complications following esophagectomy should be developed.8 This led to the development of a specific scoring model to define pneumonia after esophagectomy in 2014 (Utrecht Pneumonia Score) and the initiative of the Esophagectomy Complications Consensus Group (ECCG) to standardize reporting of complications following esophagectomy.6,9 The results of the Delphi survey, published as the ‘International Consensus on Standardization of Data Collection for Complications Associated with Esophagectomy,’ are considered to be the standard guideline for reporting complications after esophagectomy and is now also used in nationwide registries.9,15 Following the agreement to develop a comprehensive complication list, and in an effort to avoid confusion regarding already existing definitions, the participating centers decided to link many complications, including pneumonia, to standard definitions. Pneumonia was defined by an internationally accepted definition, published by the American Thoracic Society and the Center for Disease Control and Prevention.9,16 However, these guidelines focused on ventilator-associated pneumonia (VAP), defining pneumonia according to the clinical pulmonary infection score (CPIS). The CPIS has been developed in patients with VAP, without validation in patients with pneumonia.17–19 Moreover, it has been demonstrated that the CPIS is not generally applicable, since its validation in trauma patients failed.20 Hence, a general and widely accepted definition for pneumonia after esophagectomy was still lacking. To fill the gap, the Utrecht Pneumonia Score was revised, and an internal and external validation of the revised score led to the introduction of the UPS in 2016.10 The UPS consists out of three variables: leukocytes, temperature, and pulmonary radiography. Sputum culture is not included in the UPS since, as previously demonstrated, a positive sputum culture was not significant associated with the clinical diagnosis of pneumonia in multivariate analysis. Furthermore, in approximately a quarter of patients treated for pneumonia sputum culture results are not available, making it an insufficient variable for defining pneumonia in retrospective research.6 Since pneumonia is the most frequently observed complication after esophagectomy for cancer, it often serves as the primary outcome of comparative surgical studies.8 The UPS has already been utilized as a primary outcome measure in multiple retrospective and prospective studies.21–26 However, several landmark trials in which pneumonia was the primary outcome measure and on which current practice is based, like the TIME trial27 and the MIRO trial,2,28 have utilized different and nonvalidated definitions. The UPS has proven to be an accurate definition for pneumonia after esophagectomy in different centers within the Netherlands. It is a dichotomous scoring system, which can be applied easily in clinical practice. However, it only facilitates quantifying the diagnosis of pneumonia. Since it does not stratify on the basis of severity, grading of pneumonia severity may be scored using the Clavien–Dindo classification of surgical complications.29,30 In this study, the UPS was validated in a geographically distinct, external cohort in the USA.10 Esophagectomies at this North American center were performed open and peri- and postoperative care was guided, in contrast to the previous validation study, by an enhanced recovery after surgery (ERAS) protocol.11 The ERAS protocol aims to optimize perioperative care by minimizing surgical stress and complications and accelerate recovery.31 This may be an important reason why the incidence of pneumonia in this study was remarkably lower (11.9%) when compared to pneumonia rates in the previous validation study (36% in the University Medical Center Utrecht and 40% in the Catharina Hospital Eindhoven).10 Also, in the present cohort, the UPS was able to provide a reasonably accurate definition for pneumonia after esophagectomy, which shows its potential for broader application. Standardized definitions for postoperative complications will increase accuracy of future research and quality initiatives. The strength of this study is the evaluation of a prospective dataset without loss to follow-up, containing detailed information on postoperative complications from a high-volume esophageal cancer care center. Furthermore, the ease of score-variable collection is demonstrated, facilitating the wide implementation of the UPS in both retrospective and prospective studies. Since UPS-specific variables are collected routinely on the fourth postoperative day in most hospitals, also nationwide registries can make use of the UPS. Despite this, certain limitations apply to the current analysis. The gold standard for diagnosing pneumonia in this study was the clinical decision to initiate treatment for suspected pneumonia. This subjective definition has been used in previous studies and constitutes the only definition in which all factors are accounted for by the attending physician. Since there is no current gold standard for the definition of postesophagectomy pneumonia, this definition remains the best standard to validate the UPS, which is designed to fill this gap. Though the variables included in the UPS are of substantial importance in the clinical decision making process to initiate antibiotic treatment for pneumonia, many other factors contribute to the clinical diagnosis. However, not all factors can be included in an objective scoring system, which is applicable for research. This validation study demonstrates that the model variables used in the UPS have proven to be an adequate representation of all factors taken into account by the attending physician. In this study, model variables in patients who were not treated for pneumonia were collected on the fourth postoperative day, which is in accordance with the previous validation- and development study. This might have led to a small underestimation of the number of false positives (patients who had a positive UPS after postoperative day 4, but were not treated with antibiotics). However, from a methodological point of view, the same methodology as presented in the development study should be used to adequately validate a scoring system. Furthermore, the incidence of pneumonia was 10.9% in this study, hence relatively few events were observed. However, for external validation studies, formal sample size calculations for the number of events based on statistical power considerations are not well investigated.14 Within the context of these limitations, this study showed the UPS to be an accurate, objective, and easy to implement scoring system to define pneumonia after esophagectomy in esophageal cancer research. Therefore, we encourage the application of the UPS in future esophageal cancer research and nationwide registries reporting on outcomes after esophagectomy. Notes Specific author contributions: Conception and design: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum, Teus J. Weijs, Jelle P. Ruurda, Richard van Hillegersberg, Donald E. Low; Administrative support: Maarten F. J. Seesing, Andrea Wirsching; Provision of study materials or patients: Donald E. Low; Collection and assembly of data: Maarten F. J. Seesing, Andrea Wirsching; Data analysis and interpretation: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum; Manuscript writing: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum; Teus J. Weijs, Jelle P. Ruurda, Richard van Hillegersberg, Donald E. Low; Final approval of manuscript: Maarten F. J. Seesing, Andrea Wirsching, Peter S. N. van Rossum; Teus J. Weijs, Jelle P. Ruurda, Richard van Hillegersberg, Donald E. Low. References 1 van Hagen P , Hulshof M C , van Lanschot J J et al. Preoperative chemoradiotherapy for esophageal or junctional cancer . N Engl J Med 2012 ; 366 : 2074 – 84 . Google Scholar CrossRef Search ADS PubMed 2 Mariette C , Robb W B . Open or minimally invasive resection for oesophageal cancer? Recent Results Cancer Res 2012 ; 196 : 155 – 67 . Google Scholar CrossRef Search ADS PubMed 3 Lerut T , Moons J , Coosemans W et al. Postoperative complications after transthoracic esophagectomy for cancer of the esophagus and gastroesophageal junction are correlated with early cancer recurrence: role of systematic grading of complications using the modified Clavien classification . Ann Surg 2009 ; 250 : 798 – 807 . Google Scholar CrossRef Search ADS PubMed 4 Goldberg R F , Bowers S P , Parker M et al. Technical and perioperative outcomes of minimally invasive esophagectomy in the prone position . Surg Endosc 2013 ; 27 : 553 – 7 . Google Scholar CrossRef Search ADS PubMed 5 Seesing M F J , Gisbertz S S , Goense L et al. A propensity score matched analysis of open versus minimally invasive transthoracic esophagectomy in the Netherlands . Ann Surg 2017; 266: 839 . 6 van der Sluis P C , Verhage R J , van der Horst S et al. A new clinical scoring system to define pneumonia following esophagectomy for cancer . Dig Surg 2014 ; 31 : 108 – 16 . Google Scholar CrossRef Search ADS PubMed 7 Markar S , Gronnier C , Duhamel A et al. Pattern of postoperative mortality after esophageal cancer resection according to center volume: results from a large European multicenter study . Ann Surg Oncol 2015 ; 22 : 2615 – 23 . Google Scholar CrossRef Search ADS PubMed 8 Blencowe N S , Strong S , McNair A G et al. Reporting of short-term clinical outcomes after esophagectomy: a systematic review . Ann Surg 2012 ; 255 : 658 – 66 . Google Scholar CrossRef Search ADS PubMed 9 Low D E , Alderson D , Cecconello I et al. International consensus on standardization of data collection for complications associated with esophagectomy: Esophagectomy Complications Consensus Group (ECCG) . Ann Surg 2015 ; 262 : 286 – 94 . Google Scholar CrossRef Search ADS PubMed 10 Weijs T J , Seesing M F , van Rossum P S et al. Internal and external validation of a multivariable model to define hospital-acquired pneumonia after esophagectomy . J Gastrointest Surg 2016 ; 20 : 680 – 7 . Google Scholar CrossRef Search ADS PubMed 11 Markar S R , Schmidt H , Kunz S et al. Evolution of standardized clinical pathways: refining multidisciplinary care and process to improve outcomes of the surgical treatment of esophageal cancer . J Gastrointest Surg 2014 ; 18 : 1238 – 46 . Google Scholar CrossRef Search ADS PubMed 12 Collins G S , Reitsma J B , Altman D G et al. Transparent reporting of a multivariable prediction model for individual prognosis or diagnosis (TRIPOD): the TRIPOD statement . Ann Intern Med 2015 ; 162 : 55 – 63 . Google Scholar CrossRef Search ADS PubMed 13 Steyerberg E W , Vickers A J , Cook N R et al. Assessing the performance of prediction models: a framework for traditional and novel measures . Epidemiology 2010 ; 21 : 128 – 38 . Google Scholar CrossRef Search ADS PubMed 14 Han K , Song K , Choi B W . How to develop, validate, and compare clinical prediction models involving radiological parameters: study design and statistical methods . Korean J Radiol 2016 ; 17 : 339 – 50 . Google Scholar CrossRef Search ADS PubMed 15 Busweiler L A , Wijnhoven B P , van Berge Henegouwen M I et al. Early outcomes from the Dutch Upper Gastrointestinal Cancer Audit . Br J Surg 2016 ; 103 : 1855 – 63 . Google Scholar CrossRef Search ADS PubMed 16 American Thoracic Society, Centers for Disease Control and Prevention, Infectious Diseases Society of America . American Thoracic Society/Centers for Disease Control and Prevention/Infectious Diseases Society of America: controlling tuberculosis in the United States . Am J Respir Crit Care Med 2005 ; 172 : 1169 – 227 . CrossRef Search ADS PubMed 17 Rosbolt M B , Sterling E S , Fahy B G . The utility of the clinical pulmonary infection score . J Intensive Care Med 2009 ; 24 : 26 – 34 . Google Scholar CrossRef Search ADS PubMed 18 Pugin J , Auckenthaler R , Mili N et al. Diagnosis of ventilator-associated pneumonia by bacteriologic analysis of bronchoscopic and nonbronchoscopic “blind” bronchoalveolar lavage fluid . Am Rev Respir Dis 1991 ; 143 : 1121 – 9 . Google Scholar CrossRef Search ADS PubMed 19 Singh N , Rogers P , Atwood C W et al. Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit. A proposed solution for indiscriminate antibiotic prescription . Am J Respir Crit Care Med 2000 ; 162 : 505 – 11 . Google Scholar CrossRef Search ADS PubMed 20 Pham T N , Cancio L C , Gibran N S et al. American Burn Association practice guidelines burn shock resuscitation . J Burn Care Res 2008 ; 29 : 257 – 66 . Google Scholar CrossRef Search ADS PubMed 21 Weijs T J , Berkelmans G H , Nieuwenhuijzen G A et al. Immediate postoperative oral nutrition following esophagectomy: a multicenter clinical trial . Ann Thorac Surg 2016 ; 102 : 1141 – 8 . Google Scholar CrossRef Search ADS PubMed 22 van der Sluis P C , Ruurda J P , van der Horst S et al. Robot-assisted minimally invasive thoracolaparoscopic esophagectomy versus open transthoracic esophagectomy for resectable esophageal cancer, a randomized controlled trial (ROBOT trial) . Trials 2012 ; 13 : 230–6215–13–230 . Google Scholar CrossRef Search ADS 23 Yamana I , Takeno S , Hashimoto T et al. Randomized controlled study to evaluate the efficacy of a preoperative respiratory rehabilitation program to prevent postoperative pulmonary complications after esophagectomy . Dig Surg 2015 ; 32 : 331 – 7 . Google Scholar CrossRef Search ADS PubMed 24 Valkenet K , Trappenburg J C , Gosselink R et al. Preoperative inspiratory muscle training to prevent postoperative pulmonary complications in patients undergoing esophageal resection (PREPARE study): study protocol for a randomized controlled trial . Trials 2014 ; 15 : 144–6215–15–144 . Google Scholar CrossRef Search ADS 25 Scarpa M , Cavallin F , Saadeh L M et al. Hybrid minimally invasive esophagectomy for cancer: impact on postoperative inflammatory and nutritional status . Dis Esophagus 2016 ; 29 : 1064 – 70 . Google Scholar CrossRef Search ADS PubMed 26 Berkelmans G H , Wilts B J , Kouwenhoven E A et al. Nutritional route in oesophageal resection trial II (NUTRIENT II): study protocol for a multicentre open-label randomised controlled trial . BMJ Open 2016 ; 6 : e011979–2016–011979 . Google Scholar CrossRef Search ADS 27 Biere S S , van Berge Henegouwen M I , Maas K W et al. Minimally invasive versus open oesophagectomy for patients with oesophageal cancer: a multicentre, open-label, randomised controlled trial . Lancet 2012 ; 379 : 1887 – 92 . Google Scholar CrossRef Search ADS PubMed 28 Briez N , Piessen G , Bonnetain F et al. Open versus laparoscopically-assisted oesophagectomy for cancer: a multicentre randomised controlled phase III trial—the MIRO trial . BMC Cancer 2011 ; 11 : 310–2407–11–310 . Google Scholar CrossRef Search ADS 29 Dindo D , Demartines N , Clavien P A . Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey . Ann Surg 2004 ; 240 : 205 – 13 . Google Scholar CrossRef Search ADS PubMed 30 Clavien P A , Strasberg S M . Severity grading of surgical complications . Ann Surg 2009 ; 250 : 197 – 8 . Google Scholar CrossRef Search ADS PubMed 31 Findlay J M , Gillies R S , Millo J et al. Enhanced recovery for esophagectomy: a systematic review and evidence-based guidelines . Ann Surg 2014 ; 259 : 413 – 31 . Google Scholar CrossRef Search ADS PubMed © The Authors 2018. Published by Oxford University Press on behalf of International Society for Diseases of the Esophagus. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Diseases of the EsophagusOxford University Press

Published: Apr 14, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off