Abstract
Aim
The existing predictive risk models for the surgical outcome of acute diffused peritonitis (ADP) need renovation by adding relevant variables such as ADP's definition or causative etiology to pursue outstanding data collection reflecting the real world. We aimed to revise the risk models predicting mortality and morbidities of ADP using the latest Japanese Nationwide Clinical Database (NCD) variable set.
Methods
Clinical dataset of ADP patients who underwent surgery, and registered in the NCD between 2016 and 2019, were used to develop a risk model for surgical outcomes. The primary outcome was perioperative mortality.
Results
After data cleanup, 45 379 surgical cases for ADP were derived for analysis. The perioperative and 30‐day mortality were 10.6% and 7.2%, respectively. The prediction models have been created for the mortality and 10 morbidities associated with the mortality. The top five relevant predictors for perioperative mortality were age >80, advanced cancer with multiple metastases, platelet count of <50 000/mL, serum albumin of <2.0 g/dL, and unknown ADP site. The C‐indices of perioperative and 30‐day mortality were 0.859 and 0.857, respectively. The predicted value calculated with the risk models for mortality was highly fitted with the actual probability from the lower to the higher risk groups.
Conclusions
Risk models for postoperative mortality and morbidities with good predictive performance and reliability were revised and validated using the recent real‐world clinical dataset. These models help to predict ADP surgical outcomes accurately and are available for clinical settings.
Keywords: acute diffuse peritonitis, national clinical database, prediction, risk model
Risk models for predicting postoperative surgical outcomes of acute diffused peritonitis (ADP) were created and validated using the recent real‐world clinical dataset of the Japanese Nationwide Clinical Database (NCD). Evaluation of predictive power and calibration of these risk models showed they were sufficiently reliable to be used in clinical settings.
1. INTRODUCTION
Globally, acute diffused peritonitis (ADP) is a typical medical and surgical emergency that remains a disease with high mortality despite improvements in surgical and intensive care management. 1 , 2 , 3 It is essential to promptly diagnose ADP and appropriate treatments because potentially fatal infectious complications such as septic shock and disseminated intravascular coagulation (DIC) would deteriorate if adequate intervention is delayed. The treatment for ADP involves a comprehensive approach that includes medical and surgical interventions. Prompt and aggressive management is crucial to prevent complications and improve outcomes. Surgical intervention is the mainstay therapy for addressing the underlying cause of ADP and drainage of intra‐abdominal infectious fluid 4 ; however, the postoperative mortality rates of ADP were reported to be between 8.4% and 34% from isolated studies. 5 , 6 , 7 , 8 Indeed, preoperative conditions of ADP patients can be associated with high mortality and severe complications. 9 , 10 Given that emergency surgery, which could be invasive for clinically ill patients, must be considered for ADP, precious prediction of surgical outcomes is extremely important for decision support between medical providers and the patients or their families.
Knowledge about risk factors for mortality of ADP has been accumulated as several scoring systems for evaluating the severity of ADP were reported, such as APACHE‐2, P‐POSSUM, or Mannheim peritonitis index. 9 , 10 , 11 , 12 Similarly, most scoring systems were developed from a dataset in an author's unit. With the increasing usage of electric health records (EHR) in the medical and healthcare field, it has been recognized that using real‐world data from the EHR is significant in developing machine learning‐based predictive risk models. 13 , 14 In Japan, in previous studies, postoperative mortality risk models for eight major gastroenterological surgeries, including ADP, were created using the national clinical database (NCD) and reported separately. 7 , 15 , 16 , 17 , 18 , 19 , 20 , 21 Regarding ADP, a surgical risk model for predicting mortality was initially developed using 8482 surgical cases registered in the NCD in 2011, 7 followed by risk models for predicting operative morbidities using 16 930 cases reported in 2011 and 2012. 22 The developed risk models were implemented as risk calculators on the NCD web system. The probabilities of mortality and morbidities in a patient were calculated and available for any medical provider (http://www.ncd.or.jp).
ADP is a clinical syndrome rather than a definitive diagnosis as the causative etiology varies widely: frequent etiologies were ruptured appendix, ischemic bowel disease, gastrointestinal tract perforation, and so on. 23 , 24 , 25 In addition, the severity of ADP is considered to vary depending on its cause and responsible site. In the initial database between 2011 and 2015, the classification of ADP (primary, secondary, or tertiary peritonitis) and clinical information about the detailed etiology and responsible organ (or site) was not contained in the NCD registry items, which became a limitation of the initially developed risk model. Since 2016, Such information has been added to the NCD items. Under a new NCD project, the current study was conducted to create new NCD risk models for predicting surgical outcomes of ADP using the current and larger database from 2016.
2. METHODS
2.1. Data source and patient selection
The NCD is Japan's largest nationwide surgical registry, established in 2010. The NCD project systematically collects and analyzes data and uses them to improve the quality of surgical procedures in Japan. 26 The accuracy of the NCD data was ensured through systematic audits for standard NCD input initiated by the Japanese Society of Gastroenterological Surgery (JSGS) database committee in 2016. 27 In the current study, perioperative data were retrieved from the NCD in patients who underwent surgery for ADP between January 2016 and December 2019. Preoperative and perioperative variables registered in the NCD system were used in the current study.
2.2. Clinical data registry regarding the definition of ADP
Acute diffused peritonitis can be classified based on various factors, including the underlying etiology or the organs/sites responsible within the abdomen. When classifying ADP, primary peritonitis occurs due to bacterial translocation, hematogenous spread, or iatrogenic contamination of the abdomen without a macroscopic defect in the gastrointestinal tract (e.g., idiopathic). On the other hand, secondary peritonitis arises from direct contamination of the peritoneum by spillage from the gastrointestinal or urogenital tracts or their associated solid organs (e.g., perforation, acute appendicitis, or ruptured suture). Tertiary peritonitis is persistent secondary peritonitis lasting 48 h after an attempted surgical source control. 2 , 28 , 29 A clear definition of ADP is crucial for developing a relevant risk model. The NCD registry from 2016 has included variables related to the classification of ADP (primary, secondary, or tertiary peritonitis), detailed etiology, and the responsible organ/site as data items (refer to Table S1).
2.3. Outcome measures
The primary outcome of this study was perioperative mortality of ADP, which was defined as 30‐day postoperative mortality and in‐hospital mortality. For the secondary outcomes, 11 postoperative complications strongly associated with perioperative death were selected according to the number of events and their association with mortality. The 11 postoperative complications included 30‐day mortality, organ space surgical site infection (SSI), pneumonia, unplanned intubation, mechanical ventilation, acute renal failure, impaired consciousness, respiratory arrest, sepsis, disseminated intravascular coagulation (DIC), and postoperative transfusion.
2.4. Statistical analysis
Descriptive statistics were produced for the study subjects' demographic, clinical, and laboratory characteristics. Categorical variables were expressed as numbers and percentages. A risk model was generated for each outcome using a backward stepwise logistic regression with the Akaike information criterion (AIC). According to previous studies and clinical knowledge, candidate covariates for inclusion in the risk models were selected from registry variables relevant to the outcomes. Avoiding quasi‐complete separation problems in a logistic regression model, when the number of events (or non‐events) for any one of the covariate levels was <5, the levels were merged, or the covariate was excluded from the candidate. If the correlation coefficient between two covariates was >0.9, the covariate considered more important was left in the candidate covariate list to avoid multicollinearity. The performance of the risk model was evaluated using the c‐index and a calibration plot. The optimism of the c‐index was corrected using Harrell's bias correction method. 30 Confidence intervals for the c‐index were calculated using the location‐shifted bootstrap method 31 since the sample size was sufficiently large. The number of bootstrap samples was set to 200.
Similarly, bootstrapping was used to correct optimism in the calibration plot. 30 The statistical significance level was set at 0.05; all tests were two‐tailed. All statistical analyses were performed using R software version 4.0 and later (R Foundation, Vienna, Austria, https://www.r‐project.org/). Backward stepwise variable selection method and calibration plots were performed using the rms R package version 6.2 and later (Frank E Harrell Jr., https://CRAN.R‐project.org/package=rms).
3. RESULTS
3.1. Patients characteristics and preoperative risk profiles and laboratory data of the study population
Between 2016 and 2019, the NCD registered 3 088 308 surgical cases in the gastroenterological surgery section. We collected clinical data on 45 379 surgical cases for ADP after excluding duplicate patients on the NCD registry (n = 1663), a patient with categorization mismatch (n = 1), and patients with missing data (n = 11 966) (Figure 1). The characteristics of the study cohort are summarized in Table 1. Notably, emergency surgery and ambulance transportation accounted for 93.4% and 44.3% of cases, respectively, which is comparable with the previous NCD reports. 7 , 22 In the classification of ADP, secondary peritonitis was dominant, accounting for 92.7%. Regarding etiology, bowel perforation was a leading cause, accounting for 68.5% of cases. “Upper gastrointestinal tract (stomach, duodenum, or small bowel)” and “large bowel” were the frequent organs causing ADP, accounting for 38.5% and 37.8% of cases, respectively (Table 1).
FIGURE 1.
Patient flow.
TABLE 1.
Profiles and laboratory data of the entire study population and outcome groups.
| Entire study population | Surgical outcomes | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variables | Perioperative mortality | 30‐day mortality | Organ space SSI | Pneumonia | Unplanned intubation | Respiratory failure | Acute renal failure | Impaired consciousness | Cardio‐respiratory arrest | Sepsis | DIC | Transfusion | |
| Number of patients (% of the entire population) | n = 45 379 | n = 4809 (10.6) | n = 3259 (7.2) | n = 3997 (8.8) | n = 2869 (6.3) | n = 1568 (3.5) | n = 5535 (12.2) | n = 2317 (5.1) | n = 1508 (3.3) | n = 743 (1.6) | n = 6647 (14.6) | n = 2295 (5.1) | n = 5490 (12.1) |
| Demographics | No. (%) | ||||||||||||
| Age, Under 60 years | 12 496 | 413 (3.3) | 263 (2.1) | 820 (6.6) | 250 (2.0) | 172 (1.4) | 672 (5.4) | 260 (2.1) | 180 (1.4) | 89 (0.7) | 902 (7.2) | 250 (2.0) | 779 (6.2) |
| 61–65 years | 3911 | 338 (8.6) | 206 (5.3) | 361 (9.2) | 137 (3.5) | 104 (2.7) | 357 (9.1) | 168 (4.3) | 75 (1.9) | 32 (0.8) | 452 (11.6) | 146 (3.7) | 442 (11.3) |
| 66–70 years | 6076 | 588 (9.7) | 360 (5.9) | 613 (10.1) | 342 (5.6) | 208 (3.4) | 728 (12.0) | 304 (5.0) | 177 (2.9) | 88 (1.4) | 892 (14.7) | 287 (4.7) | 751 (12.4) |
| 71–75 years | 5785 | 720 (12.4) | 476 (8.2) | 611 (10.6) | 442 (7.6) | 246 (4.3) | 812 (14.0) | 351 (6.1) | 215 (3.7) | 107 (1.8) | 995 (17.2) | 353 (6.1) | 826 (14.3) |
| 76–80 years | 6063 | 831 (13.7) | 573 (9.5) | 634 (10.5) | 519 (8.6) | 305 (5.0) | 991 (16.3) | 425 (7.0) | 253 (4.2) | 135 (2.2) | 1158 (19.1) | 427 (7.0) | 924 (15.2) |
| 81 and over years | 11 048 | 1919 (17.4) | 1381 (12.5) | 958 (8.7) | 1179 (10.7) | 533 (4.8) | 1975 (17.9) | 809 (7.3) | 608 (5.5) | 292 (2.6) | 2248 (20.3) | 832 (7.5) | 1768 (16.0) |
| Sex, male | 27 000 | 2763 (10.2) | 1829 (6.8) | 2396 (8.9) | 1840 (6.8) | 980 (3.6) | 3256 (12.1) | 1388 (5.1) | 827 (3.1) | 423 (1.6) | 3753 (13.9) | 1269 (4.7) | 3077 (11.4) |
| BMI: <26 kg/m2 | 39 900 | 4219 (10.6) | 2830 (7.1) | 3429 (8.6) | 2563 (6.4) | 1348 (3.4) | 4728 (11.8) | 1938 (4.9) | 1290 (3.2) | 629 (1.6) | 5770 (14.5) | 2012 (5.0) | 4811 (12.1) |
| ≥26, and <30 kg/m2 | 4114 | 435 (10.6) | 310 (7.5) | 424 (10.3) | 228 (5.5) | 155 (3.8) | 560 (13.6) | 267 (6.5) | 152 (3.7) | 91 (2.2) | 630 (15.3) | 212 (5.2) | 506 (12.3) |
| ≥30 kg/m2 | 1365 | 155 (11.4) | 119 (8.7) | 144 (10.5) | 78 (5.7) | 65 (4.8) | 247 (18.1) | 112 (8.2) | 66 (4.8) | 23 (1.7) | 247 (18.1) | 71 (5.2) | 173 (12.7) |
| General | |||||||||||||
| Ambulance transportation | 20 111 | 2350 (11.7) | 1735 (8.6) | 1510 (7.5) | 1465 (7.3) | 736 (3.7) | 3043 (15.1) | 1291 (6.4) | 878 (4.4) | 406 (2.0) | 3542 (17.6) | 1245 (6.2) | 2658 (13.2) |
| Emergent surgery | 42 397 | 4524 (10.7) | 3076 (7.3) | 3776 (8.9) | 2717 (6.4) | 1479 (3.5) | 5321 (12.6) | 2204 (5.2) | 1429 (3.4) | 718 (1.7) | 6370 (15.0) | 2208 (5.2) | 5196 (12.3) |
| ADL within 30 days, any assistance | 7143 | 1667 (23.3) | 1160 (16.2) | 898 (12.6) | 1072 (15.0) | 467 (6.5) | 1752 (24.5) | 643 (9.0) | 594 (8.3) | 257 (3.6) | 1885 (26.4) | 653 (9.1) | 1571 (22.0) |
| ADL just before surgery, any assistance | 11 942 | 2391 (20.0) | 1678 (14.1) | 1416 (11.9) | 1435 (12.0) | 736 (6.2) | 2694 (22.6) | 1013 (8.5) | 837 (7.0) | 371 (3.1) | 2792 (23.4) | 1098 (9.2) | 2390 (20.0) |
| ASA classifications 3, 4 and 5 | 21 736 | 3975 (18.3) | 2753 (12.7) | 2632 (12.1) | 2221 (10.2) | 1237 (5.7) | 4599 (21.2) | 1887 (8.7) | 1282 (5.9) | 628 (2.9) | 5147 (23.7) | 1868 (8.6) | 4239 (19.5) |
| Smoking habit within 1 year | 9491 | 661 (7.0) | 432 (4.6) | 755 (8.0) | 483 (5.1) | 253 (2.7) | 869 (9.2) | 388 (4.1) | 212 (2.2) | 113 (1.2) | 1071 (11.3) | 346 (3.6) | 909 (9.6) |
| Alcohol consumption habitat | 10 447 | 816 (7.8) | 534 (5.1) | 936 (9.0) | 574 (5.5) | 337 (3.2) | 1092 (10.5) | 487 (4.7) | 260 (2.5) | 140 (1.3) | 1271 (12.2) | 473 (4.5) | 1115 (10.7) |
| Diabetes | 7150 | 925 (12.9) | 626 (8.8) | 768 (10.7) | 571 (8.0) | 307 (4.3) | 1105 (15.5) | 494 (6.9) | 317 (4.4) | 146 (2.0) | 1282 (17.9) | 455 (6.4) | 1065 (14.9) |
| Esophageal varix within 6 months | 248 | 57 (23.0) | 44 (17.7) | 35 (14.1) | 25 (10.1) | 14 (5.6) | 60 (24.2) | 35 (14.1) | 20 (8.1) | 6 (2.4) | 69 (27.8) | 17 (6.9) | 75 (30.2) |
| Chronic steroid use | 1927 | 407 (21.1) | 282 (14.6) | 240 (12.5) | 244 (12.7) | 126 (6.5) | 423 (22.0) | 177 (9.2) | 114 (5.9) | 67 (3.5) | 505 (26.2) | 179 (9.3) | 427 (22.2) |
| Blood coagulation abnormality without anticoagulant | 1695 | 439 (25.9) | 313 (18.5) | 296 (17.5) | 281 (16.6) | 114 (6.7) | 531 (31.3) | 282 (16.6) | 187 (11.0) | 73 (4.3) | 619 (36.5) | 236 (13.9) | 478 (28.2) |
| Blood coagulation abnormality with anticoagulant | 2468 | 465 (18.8) | 338 (13.7) | 305 (12.4) | 305 (12.4) | 160 (6.5) | 595 (24.1) | 267 (10.8) | 165 (6.7) | 85 (3.4) | 638 (25.9) | 232 (9.4) | 545 (22.1) |
| Sepsis just before surgery | 7063 | 1851 (26.2) | 1370 (19.4) | 1261 (17.9) | 1105 (15.6) | 602 (8.5) | 2575 (36.5) | 1267 (17.9) | 853 (12.1) | 352 (5.0) | 3440 (48.7) | 1099 (15.6) | 2024 (28.7) |
| Cardiovascular factors | |||||||||||||
| Hypertension within 30 days | 16 508 | 2133 (12.9) | 1473 (8.9) | 1596 (9.7) | 1318 (8.0) | 723 (4.4) | 2529 (15.3) | 1115 (6.8) | 681 (4.1) | 340 (2.1) | 2950 (17.9) | 1098 (6.7) | 2401 (14.5) |
| Congestive heart failure within 30 days | 801 | 270 (33.7) | 198 (24.7) | 121 (15.1) | 161 (20.1) | 86 (10.7) | 299 (37.3) | 156 (19.5) | 117 (14.6) | 60 (7.5) | 289 (36.1) | 111 (13.9) | 245 (30.6) |
| Myocardial infarction within 6 months | 219 | 69 (31.5) | 49 (22.4) | 28 (12.8) | 32 (14.6) | 25 (11.4) | 63 (28.8) | 28 (128) | 21 (9.6) | 19 (8.7) | 73 (33.3) | 18 (8.2) | 56 (25.6) |
| Angina pectoris within 30 days | 527 | 98 (18.6) | 74 (14.0) | 44 (8.3) | 60 (11.4) | 38 (7.2) | 104 (19.7) | 47 (8.9) | 28 (5.3) | 26 (4.9) | 125 (23.7) | 39 (7.4) | 97 (18.4) |
| Previous PCI | 1139 | 225 (19.8) | 158 (13.9) | 127 (11.2) | 129 (11.3) | 80 (7.0) | 248 (21.8) | 99 (8.7) | 62 (5.4) | 43 (3.8) | 279 (24.5) | 105 (9.2) | 232 (20.4) |
| Previous cardiac surgery | 742 | 154 (20.8) | 105 (14.2) | 98 (13.2) | 79 (10.6) | 55 (7.4) | 207 (27.9) | 83 (11.2) | 47 (6.3) | 28 (3.8) | 187 (25.2) | 86 (11.6) | 175 (23.6) |
| Previous PVD with a symptom | 196 | 75 (38.3) | 49 (25.0) | 44 (22.4) | 32 (16.3) | 25 (12.8) | 77 (39.3) | 31 (15.8) | 26 (13.3) | 16 (8.2) | 68 (34.7) | 29 (14.8) | 53 (27.0) |
| Previous PVD surgery | 301 | 88 (29.2) | 57 (18.9) | 46 (15.3) | 45 (15.0) | 24 (8.0) | 87 (28.9) | 33 (11.0) | 29 (9.6) | 16 (5.3) | 91 (30.2) | 42 (14.0) | 79 (26.2) |
| Respiratory factors | |||||||||||||
| Ventilator dependent within 48 h | 1161 | 477 (41.1) | 345 (29.7) | 246 (21.2) | 255 (22.0) | 145 (12.5) | 726 (62.5) | 251 (21.6) | 244 (21.0) | 101 (8.7) | 532 (45.8) | 196 (16.9) | 476 (41.0) |
| COPD | 1390 | 276 (19.9) | 185 (13.3) | 190 (13.7) | 252 (18.1) | 118 (8.5) | 341 (24.5) | 145 (10.4) | 87 (6.3) | 47 (3.4) | 352 (25.3) | 126 (9.1) | 271 (19.5) |
| Current pneumonia | 1026 | 331 (32.3) | 233 (22.7) | 178 (17.3) | 455 (44.3) | 128 (12.5) | 456 (44.4) | 172 (16.8) | 127 (12.4) | 49 (4.8) | 393 (38.3) | 150 (14.6) | 323 (31.5) |
| Respiratory distress within 30 days | 2100 | 706 (33.6) | 494 (23.5) | 386 (18.4) | 467 (22.2) | 225 (10.7) | 856 (40.8) | 361 (17.2) | 290 (13.8) | 127 (6.0) | 783 (37.3) | 277 (13.2) | 650 (31.0) |
| Renal factors | |||||||||||||
| Acute renal failure within 24 h | 1384 | 495 (35.8) | 366 (26.4) | 291 (21.0) | 270 (19.5) | 129 (9.3) | 644 (46.5) | 623 (45.0) | 265 (19.1) | 103 (7.4) | 723 (52.2) | 261 (18.9) | 476 (34.4) |
| Dialysis within 14 days | 1632 | 540 (33.1) | 379 (23.2) | 258 (15.8) | 192 (11.8) | 119 (7.3) | 453 (27.8) | 127 (7.8) | 175 (10.7) | 84 (5.1) | 501 (30.7) | 197 (12.1) | 474 (29.0) |
| Cerebral nerve factors | |||||||||||||
| Previous cerebrovascular disease | 2469 | 382 (15.5) | 258 (10.4) | 277 (11.2) | 321 (13.0) | 145 (5.9) | 529 (21.4) | 219 (8.9) | 164 (6.6) | 60 (2.4) | 568 (23.0) | 218 (8.8) | 475 (19.2) |
| Oncological factors | |||||||||||||
| Cancer chemotherapy within 30 days | 1845 | 410 (22.2) | 227 (12.3) | 241 (13.1) | 134 (7.3) | 69 (3.7) | 233 (12.6) | 108 (5.9) | 65 (3.5) | 24 (1.3) | 356 (19.3) | 138 (7.5) | 358 (19.4) |
| Cancer chemotherapy within 90 days | 2487 | 531 (21.4) | 281 (11.3) | 347 (14.0) | 171 (6.9) | 84 (3.4) | 287 (11.5) | 125 (5.0) | 72 (2.9) | 27 (1.1) | 449 (18.1) | 154 (6.2) | 462 (18.6) |
| Cancer radiotherapy within 90 days | 275 | 62 (22.5) | 36 (13.1) | 35 (12.7) | 25 (9.1) | 5 (1.8) | 38 (13.8) | 15 (5.5) | 6 (2.2) | 2 (0.7) | 42 (15.3) | 19 (6.9) | 39 (14.2) |
| Cancer immunotherapy within 90 days | 146 | 23 (15.8) | 19 (13.0) | 12 (8.2) | 12 (8.2) | 8 (5.5) | 23 (15.8) | 7 (4.8) | 7 (4.8) | 6 (4.1) | 30 (20.5) | 13 (8.9) | 30 (20.5) |
| Advanced cancer with multiple metastases | 1721 | 557 (32.4) | 323 (18.8) | 240 (13.9) | 151 (8.8) | 72 (4.2) | 270 (15.7) | 144 (8.4) | 103 (6.0) | 35 (2.0) | 433 (25.2) | 131 (7.6) | 383 (22.3) |
| Benign tumor | 472 | 27 (5.7) | 20 (4.2) | 12 (2.5) | 17 (3.6) | 7 (1.5) | 33 (7.0) | 12 (2.5) | 8 (1.7) | 2 (0.4) | 36 (7.6) | 8 (1.7) | 31 (6.6) |
| Malignant tumor | 5029 | 973 (19.3) | 625 (12.4) | 509 (10.1) | 376 (7.5) | 181 (3.6) | 776 (15.4) | 355 (7.1) | 236 (4.7) | 104 (2.1) | 1114 (22.2) | 385 (7.7) | 901 (17.9) |
| Hematological factors | |||||||||||||
| Preoperative transfusion within 72 h | 1412 | 432 (30.6) | 301 (21.3) | 294 (20.8) | 217 (15.4) | 131 (9.3) | 523 (37.0) | 222 (15.7) | 187 (13.2) | 85 (6.0) | 491 (34.8) | 167 (11.8) | 615 (43.6) |
| Other factors | |||||||||||||
| Open wound | 610 | 133 (21.8) | 77 (12.6) | 145 (23.8) | 79 (13.0) | 46 (7.5) | 162 (26.6) | 52 (8.5) | 57 (9.3) | 15 (2.5) | 144 (23.6) | 50 (8.2) | 165 (27.0) |
| Preoperative laboratory values | |||||||||||||
| WBC < 3000/mL | 4458 | 839 (18.8) | 590 (13.2) | 520 (11.7) | 432 (9.7) | 272 (6.1) | 1031 (23.1) | 464 (10.4) | 300 (6.7) | 127 (2.8) | 1260 (28.3) | 522 (11.7) | 903 (20.3) |
| Hemoglobin <10 g/dL | 9215 | 1699 (18.4) | 1088 (11.8) | 1194 (13.0) | 867 (9.4) | 462 (5.0) | 1593 (17.3) | 633 (6.9) | 444 (4.8) | 208 (2.3) | 1813 (19.7) | 601 (6.5) | 2368 (25.7) |
| Hematocrit <30% | 8522 | 1600 (18.8) | 1023 (12.0) | 1096 (12.9) | 804 (9.4) | 445 (5.2) | 1517 (17.8) | 603 (7.1) | 415 (4.9) | 192 (2.3) | 1671 (19.6) | 546 (6.4) | 2251 (26.4) |
| Platelet count: ≥100 000, and <150 000/mL | 4245 | 738 (17.4) | 525 (12.4) | 456 (10.7) | 377 (8.9) | 231 (5.4) | 816 (19.2) | 344 (8.1) | 249 (5.9) | 126 (3.0) | 914 (21.5) | 401 (9.4) | 755 (17.8) |
| ≥50 000, and <100 000/mL | 1688 | 497 (29.4) | 350 (20.7) | 263 (15.6) | 229 (13.6) | 137 (8.1) | 506 (30.0) | 196 (11.6) | 160 (9.5) | 70 (4.1) | 523 (31.0) | 229 (13.6) | 511 (30.3) |
| <50 000/mL | 423 | 211 (49.9) | 160 (37.8) | 75 (17.7) | 75 (17.7) | 46 (10.9) | 165 (39.0) | 83 (19.6) | 59 (13.9) | 17 (4.0) | 187 (44.2) | 86 (20.3) | 177 (41.8) |
| Albumin: ≥2.0, and <3.0 g/dL | 14 646 | 2312 (15.8) | 1499 (10.2) | 1758 (12.0) | 1264 (8.6) | 736 (5.0) | 2474 (16.9) | 1007 (6.9) | 688 (4.7) | 324 (2.2) | 2801 (19.1) | 970 (6.6) | 2562 (17.5) |
| <2.0 g/dL | 4018 | 1093 (27.2) | 714 (17.8) | 724 (18.0) | 509 (12.7) | 282 (7.0) | 1059 (26.4) | 425 (10.6) | 301 (7.5) | 138 (3.4) | 1143 (28.4) | 398 (9.9) | 1138 (28.3) |
| Total bilirubin: ≥3.0 mg/dL | 1435 | 385 (26.8) | 268 (18.7) | 216 (15.1) | 142 (9.9) | 82 (5.7) | 359 (25.0) | 168 (11.7) | 120 (8.4) | 38 (2.6) | 353 (24.6) | 119 (8.3) | 329 (22.9) |
| AST >35 U/L | 9092 | 1726 (19.0) | 1240 (13.6) | 1082 (11.9) | 859 (9.4) | 504 (5.5) | 1894 (20.8) | 866 (9.5) | 599 (6.6) | 272 (3.0) | 1986 (21.8) | 755 (8.3) | 1711 (18.8) |
| ALT >35 U/L | 6884 | 1041 (15.1) | 748 (10.9) | 780 (11.3) | 528 (7.7) | 316 (4.6) | 1167 (17.0) | 539 (7.8) | 363 (5.3) | 170 (2.5) | 1224 (17.8) | 432 (6.3) | 1097 (15.9) |
| ALP >340 U/L | 8157 | 1440 (17.7) | 959 (11.8) | 896 (11.0) | 619 (7.6) | 359 (4.4) | 1265 (15.5) | 544 (6.7) | 370 (4.5) | 182 (2.2) | 1488 (18.2) | 560 (6.9) | 1348 (16.5) |
| BUN >20 mg/dL | 20 132 | 3500 (17.4) | 2453 (12.2) | 2260 (11.2) | 1942 (9.6) | 1077 (5.3) | 3915 (19.4) | 1804 (9.0) | 1144 (5.7) | 548 (2.7) | 4450 (22.1) | 1642 (8.2) | 3706 (18.4) |
| Serum creatinine ≥2.0 mg/dL | 5502 | 1411 (25.6) | 1046 (19.0) | 714 (13.0) | 610 (11.1) | 389 (7.1) | 1486 (27.0) | 844 (15.3) | 472 (8.6) | 251 (4.6) | 1633 (29.7) | 645 (11.7) | 1337 (24.3) |
| Serum sodium <130 mEq/L | 2122 | 391 (18.4) | 265 (12.5) | 258 (12.2) | 183 (8.6) | 117 (5.5) | 388 (18.3) | 177 (8.3) | 106 (5.0) | 65 (3.1) | 440 (20.7) | 174 (8.2) | 413 (19.5) |
| Serum sodium ≥147 mEq/L | 606 | 200 (33.0) | 154 (25.4) | 91 (15.0) | 109 (18.0) | 74 (12.2) | 243 (40.1) | 81 (13.4) | 62 (10.2) | 37 (6.1) | 190 (31.4) | 82 (13.5) | 185 (30.5) |
| CRP: ≥10, and <20 mg/dL | 10 007 | 1186 (11.9) | 762 (7.6) | 945 (9.4) | 694 (6.9) | 363 (3.6) | 1207 (12.1) | 482 (4.8) | 334 (3.3) | 154 (1.5) | 1425 (14.2) | 480 (4.8) | 1321 (13.2) |
| ≥20 mg/dL | 11 847 | 1404 (11.9) | 920 (7.8) | 1489 (12.6) | 886 (7.5) | 500 (4.2) | 1794 (15.1) | 813 (6.9) | 462 (3.9) | 215 (1.8) | 2137 (18.0) | 707 (6.0) | 1693 (14.3) |
| APTT ≥30 s | 32 990 | 3891 (11.8) | 2604 (7.9) | 3165 (9.6) | 2269 (6.9) | 1245 (3.8) | 4416 (13.4) | 1836 (5.6) | 1213 (3.7) | 582 (1.8) | 5206 (15.8) | 1738 (5.3) | 4355 (13.2) |
| INR of PT: >1.1, and <1.67 | 15 597 | 2207 (14.2) | 1451 (9.3) | 1719 (11.0) | 1267 (8.1) | 734 (4.7) | 2556 (16.4) | 1090 (7.0) | 707 (4.5) | 314 (2.0) | 2892 (18.5) | 1033 (6.6) | 2521 (16.2) |
| ≥1.67 | 1762 | 509 (28.9) | 396 (22.5) | 243 (13.8) | 218 (12.4) | 135 (7.7) | 573 (32.5) | 267 (15.2) | 175 (9.9) | 98 (5.6) | 566 (32.1) | 224 (12.7) | 512 (29.1) |
| Classification of ADP | |||||||||||||
| Primary peritonitis | 2566 | 349 (13.6) | 264 (10.3) | 151 (5.9) | 192 (7.5) | 89 (3.5) | 353 (13.8) | 147 (5.7) | 120 (4.7) | 66 (2.6) | 412 (16.1) | 107 (4.2) | 342 (13.3) |
| Secondary peritonitis | 42 057 | 4316 (10.3) | 2903 (6.9) | 3729 (8.9) | 2596 (6.2) | 1436 (3.4) | 5036 (12.0) | 2119 (5.0) | 1339 (3.2) | 657 (1.6) | 6071 (14.4) | 2132 (5.1) | 4997 (11.9) |
| Tertiary peritonitis | 756 | 144 (19.0) | 92 (12.2) | 117 (15.5) | 81 (10.7) | 43 (5.7) | 146 (19.3) | 51 (6.7) | 49 (6.5) | 20 (2.6) | 164 (21.7) | 56 7.4) | 151 (20.0) |
| Etiology of ADP | |||||||||||||
| Idiopathic | 1491 | 183 (12.3) | 135 (9.1) | 108 (7.2) | 93 (6.2) | 59 (4.0) | 187 (12.5) | 77 (5.2) | 62 (4.2) | 38 (2.5) | 215 (14.4) | 63 (4.2) | 188 (12.6) |
| Bowel perforation | 31 098 | 3721 (12.0) | 2548 (8.2) | 2843 (9.1) | 2091 (6.7) | 1172 (3.8) | 4194 (13.5) | 1782 (5.7) | 1146 (3.7) | 572 (1.8) | 5190 (16.7) | 1829 (5.9) | 4124 (13.3) |
| Bowel ischemia | 1154 | 316 (27.4) | 238 (20.6) | 141 (12.2) | 167 (14.5) | 82 (7.1) | 361 (31.3) | 149 (12.9) | 119 (10.3) | 53 (4.6) | 381 (33.0) | 149 (12.9) | 284 (24.6) |
| Bowel obstruction | 1494 | 205 (13.7) | 138 (9.2) | 117 (7.8) | 148 (9.9) | 70 (4.7) | 244 (16.3) | 97 (6.5) | 80 (5.4) | 31 (2.1) | 276 (18.5) | 104 (7.0) | 235 (15.7) |
| Inflammation of the biliary tract | 1032 | 91 (8.8) | 59 (5.7) | 77 (7.5) | 81 (7.8) | 37 (3.6) | 114 (11.0) | 52 (5.0) | 33 (3.2) | 17 (1.6) | 124 (12.0) | 35 (3.4) | 130 (12.6) |
| Pelvic peritonitis due to gynecological disorder | 314 | 15 (4.8) | 9 (2.9) | 16 (5.1) | 15 (4.8) | 5 (1.6) | 26 (8.3) | 12 (3.8) | 6 (1.9) | 1 (0.3) | 37 (11.8) | 9 (2.9) | 34 (10.8) |
| Bowel inflammation (appendicitis, diverticulitis, or IBD) | 6893 | 161 (2.3) | 111 (1.6) | 316 (4.6) | 185 (2.7) | 72 (1.0) | 248 (3.6) | 111 (1.6) | 50 (0.7) | 32 (0.5) | 342 (5.0) | 115 (1.7) | 209 (3.0) |
| Others | 4398 | 466 (10.6) | 270 (6.1) | 615 (14.0) | 300 (6.8) | 179 (4.1) | 548 (12.5) | 213 (4.8) | 155 (3.5) | 65 (1.5) | 570 (13.0) | 190 (4.3) | 631 (14.3) |
| Organ or site responsible for ADP | |||||||||||||
| Esophagus | 184 | 29 (15.8) | 18 (9.8) | 43 (23.4) | 31 (16.8) | 20 (10.9) | 42 (22.8) | 16 (8.7) | 9 (4.9) | 6 (3.3) | 32 (17.4) | 16 (8.7) | 34 (18.5) |
| Appendix | 7143 | 87 (1.2) | 61 (0.9) | 268 (3.8) | 138 (1.9) | 54 (0.8) | 167 (2.3) | 78 (1.1) | 30 (0.4) | 20 (0.3) | 242 (3.4) | 64 (0.9) | 129 (1.8) |
| Biliary tract | 1160 | 113 (9.7) | 67 (5.8) | 107 (9.2) | 98 (8.4) | 48 (4.1) | 136 (11.7) | 57 (4.9) | 38 (3.3) | 19 (1.6) | 151 (13.0) | 47 (4.1) | 168 (14.5) |
| Unknown | 932 | 153 (16.4) | 106 (11.4) | 80 (8.6) | 68 (7.3) | 45 (4.8) | 129 (13.8) | 58 (6.2) | 41 (4.4) | 23 (2.5) | 157 (16.8) | 46 (4.9) | 135 (14.5) |
| Stomach, duodenum, or small bowel | 17 460 | 2097 (12.0) | 1408 (8.1) | 1567 (9.0) | 1196 (6.8) | 704 (4.0) | 2126 (12.2) | 892 (5.1) | 584 (3.3) | 311 (1.8) | 2302 (13.2) | 735 (4.2) | 2306 (13.2) |
| Large bowel | 17 140 | 2294 (13.4) | 1603 (9.4) | 1810 (10.6) | 1277 (7.5) | 664 (3.9) | 2851 (16.6) | 1192 (7.0) | 797 (4.6) | 371 (2.2) | 3682 (21.5) | 1371 (8.0) | 2590 (15.1) |
| Others | 2096 | 214 (10.2) | 121 (5.8) | 228 (10.9) | 129 (6.2) | 74 (3.5) | 275 (13.1) | 102 (4.9) | 83 (4.0) | 26 (1.2) | 260 (12.4) | 80 (3.8) | 296 (14.1) |
Abbreviations: ADL, activity of daily living; ADP, acute diffused peritonitis; ALT, alanine aminotransferase; APTT, activated partial thromboplastin time; ASA, American Society of Anesthesia; AST, aspartate aminotransferase; BMI, body mass index; BUN, Blood urea nitrogen; COPD, chronic obstructive pulmonary disease; INR, international normalized ratio; PCI, percutaneous coronary intervention; PT, prothrombin time; PVD, peripheral arterial disease; WBC, white blood count.
3.2. Occurrence of mortality and morbidity in patients with ADP
The incidence of perioperative mortality and the 30‐day postoperative mortality, and the 10 morbidities are summarized in Table 2. Perioperative mortality and the 30‐day mortality rates for the 45 379 surgical cases were 10.6% and 7.2%, respectively. The association between morbidities of ADP and perioperative mortality or the 30‐day mortality is shown in Table 2. The 10 mortalities with a high number of events relatively strongly associated with perioperative mortality were selected as secondary outcomes: organ space SSI, pneumonia, unplanned intubation, respiratory failure, acute renal failure, impaired consciousness, cardio‐respiratory arrest, sepsis, and DIC. The “reoperation” was excluded from the secondary outcomes because the mortality rate was not so high as that of “unplanned intubation” or “cardio‐respiratory arrest”.
TABLE 2.
Association between morbidities and perioperative mortality or 30‐day mortality.
| Morbidities | No. of event | Perioperative mortality (n = 4809) | 30‐day mortality (n = 3259) | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| No. of death | Mortality rate (%) | OR | 95% CI | No. of death | Mortality rate (%) | OR | 95% CI | |||
| Selected as secondary outcomes | Sepsis | 6647 | 2503 | 37.7 | 9.54 | 8.94–10.18 | 1895 | 28.5 | 10.92 | 10.13–11.79 |
| On ventilator (>48 h) | 5535 | 2092 | 37.8 | 8.30 | 7.77–8.88 | 1490 | 26.9 | 7.93 | 7.35–8.56 | |
| Postoperative transfusion | 5490 | 1730 | 31.5 | 5.50 | 5.14–5.89 | 1166 | 21.2 | 4.87 | 4.50–5.27 | |
| Acute renal failure | 2317 | 1064 | 45.9 | 8.91 | 8.16–9.74 | 828 | 35.7 | 9.29 | 8.46–10.21 | |
| DIC | 2295 | 1061 | 46.2 | 9.02 | 8.26–9.86 | 787 | 34.3 | 8.57 | 7.79–9.43 | |
| Impaired consciousness (>24 h) | 1508 | 1021 | 67.7 | 22.18 | 19.81–24.84 | 846 | 56.1 | 21.96 | 19.68–24.50 | |
| Pneumonia | 2869 | 980 | 34.2 | 5.24 | 4.82–5.70 | 601 | 20.9 | 3.97 | 3.60–4.38 | |
| Organ space SSI | 3997 | 856 | 21.4 | 2.58 | 2.38–2.80 | 451 | 11.3 | 1.75 | 1.57–1.94 | |
| Unplanned intubation | 1568 | 694 | 44.3 | 7.66 | 6.90–8.51 | 467 | 29.8 | 6.23 | 5.56–6.99 | |
| Cardio‐respiratory arrest | 743 | 626 | 84.3 | 51.74 | 42.37–63.20 | 573 | 77.1 | 52.64 | 44.17–62.74 | |
| Not selected | Reoperation | 3395 | 725 | 21.4 | 2.52 | 2.31–2.75 | 348 | 10.3 | 1.53 | 1.36–1.72 |
| Wound disruption | 2213 | 572 | 25.8 | 3.20 | 2.90–3.54 | 301 | 13.6 | 2.14 | 1.88–2.43 | |
| Intraabdominal abscess | 3112 | 544 | 17.5 | 1.89 | 1.71–2.08 | 247 | 7.9 | 1.12 | 0.98–1.29 | |
| Failure of the sutures | 1811 | 458 | 25.3 | 3.05 | 2.73–3.41 | 248 | 13.7 | 2.14 | 1.86–2.46 | |
| Pulmonary atelectasis | 1691 | 427 | 25.3 | 3.03 | 2.70–3.40 | 246 | 14.5 | 2.3 | 2.00–2.64 | |
| Heart failure | 695 | 316 | 45.5 | 7.46 | 6.40–8.69 | 230 | 33.1 | 6.8 | 5.78–8.00 | |
Abbreviations: CI, confidence interval; DIC, disseminated intra‐vascular coagulation; OR, odds ratio; SSI, surgical site infection.
3.3. Risk models for predicting postoperative outcomes and their predictive performance
A risk model was created for each surgical outcome based on multivariate logistic analysis (Table 3 and Table S2). The risk model for perioperative mortality, primary outcome, retained 55 predictors, including variables of etiology and organ or site responsible for ADP (Table 3). The top five relevant predictors for perioperative mortality were selected according to an absolute value of beta coefficient value in the multivariate logistic model. The top five relevant predictors were “age ≥81 (OR = 3.83)”, “age;76–80 (OR = 2.82)”, “advanced cancer with multiple metastases (OR = 2.43)”, “platelet count <50,000/mL (OR = 2.67)”, and “serum albumin <2.0 g/dL (OR = 2.60)”.
TABLE 3.
Risk models for predicting mortality of acute diffused peritonitis.
| Variables | 30‐day mortality | Perioperative mortality | ||||||
|---|---|---|---|---|---|---|---|---|
| B. Coefficient | p value | OR | 95% CI | B. Coefficient | p value | OR | 95% CI | |
| Age <60 years | Reference | – | – | – | Reference | – | – | – |
| 61–65 years | 0.51 | <0.001 | 1.67 | 1.36–2.05 | 0.6 | <0.001 | 1.82 | 1.53–2.15 |
| 66–70 years | 0.53 | <0.001 | 1.71 | 1.42–2.04 | 0.64 | <0.001 | 1.9 | 1.64–2.21 |
| 71–75 years | 0.83 | <0.001 | 2.3 | 1.93–2.73 | 0.86 | <0.001 | 2.37 | 2.05–2.74 |
| 76–80 years | 1.02 | <0.001 | 2.77 | 2.34–3.28 | 1.04 | <0.001 | 2.82 | 2.45–3.25 |
| 81 and over the years | 1.31 | <0.001 | 3.71 | 3.17–4.35 | 1.34 | <0.001 | 3.83 | 3.35–4.38 |
| BMI: <26 kg/m2 | Reference | – | – | – | Reference | – | – | – |
| ≤26, and <30 kg/m2 | 0.17 | 0.016 | 1.19 | 1.03–1.37 | 0.15 | 0.018 | 1.16 | 1.03–1.32 |
| ≥30 kg/m2 | 0.42 | <0.001 | 1.53 | 1.22–1.92 | 0.33 | 0.002 | 1.39 | 1.13–1.71 |
| Brickman index: 0 | Reference | – | – | – | ||||
| 1–199 | −0.29 | 0.029 | 0.75 | 0.58–0.97 | ||||
| 200–399 | −0.01 | 0.923 | 0.99 | 0.82–1.20 | ||||
| 400 and over | 0.06 | 0.268 | 1.06 | 0.96–1.17 | ||||
| Sex: Female | Reference | – | – | – | ||||
| Male | 0.11 | 0.004 | 1.12 | 1.04–1.20 | ||||
| Diabetes: No | Reference | – | – | – | ||||
| Yes | −0.09 | 0.062 | 0.92 | 0.84–1.00 | ||||
| Alcohol consumption habitat: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.17 | 0.003 | 0.84 | 0.75–0.94 | −0.18 | <0.001 | 0.83 | 0.76–0.92 |
| Respiratory distress within 30 days: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.25 | <0.001 | 1.28 | 1.11–1.47 | 0.31 | <0.001 | 1.36 | 1.20–1.55 |
| ADL within 30 days, no assistance | Reference | – | – | – | Reference | – | – | – |
| Any assistance | 0.17 | 0.008 | 1.18 | 1.05–1.34 | 0.22 | <0.001 | 1.25 | 1.12–1.40 |
| ADL just before surgery, no assistance | Reference | – | – | – | Reference | – | – | – |
| Any assistance | 0.27 | <0.001 | 1.31 | 1.17–1.47 | 0.22 | <0.001 | 1.25 | 1.13–1.38 |
| Esophageal varix within 6 months: No | Reference | – | – | – | ||||
| Yes | 0.37 | 0.053 | 1.44 | 0.99–2.09 | ||||
| Hypertension within 30 days: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.08 | 0.064 | 0.92 | 0.85–1.00 | −0.07 | 0.059 | 0.93 | 0.87–1.00 |
| Previous cerebrovascular disease: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.22 | 0.005 | 0.81 | 0.69–0.94 | −0.16 | 0.022 | 0.85 | 0.75–0.98 |
| Chronic steroid use: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.27 | <0.001 | 1.3 | 1.12–1.52 | 0.31 | <0.001 | 1.36 | 1.19–1.56 |
| Blood coagulation abnormality: No | Reference | – | – | – | ||||
| Yes (without anticoagulant) | 0.12 | 0.11 | 1.12 | 0.97–1.29 | ||||
| Yes (with anticoagulant) | −0.07 | 0.269 | 0.93 | 0.81–1.06 | ||||
| Sepsis just before surgery: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.56 | <0.001 | 1.76 | 1.61–1.92 | 0.47 | <0.001 | 1.6 | 1.47–1.73 |
| ASA classification: 1 and 2 | Reference | – | – | – | Reference | – | – | – |
| 3, 4 and 5 | 0.76 | <0.001 | 2.14 | 1.92–2.38 | 0.7 | <0.001 | 2.02 | 1.85–2.20 |
| Ambulance transportation: No | Reference | – | – | – | ||||
| Yes | 0.11 | 0.01 | 1.11 | 1.03–1.21 | ||||
| Emergent surgery: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.13 | 0.14 | 0.88 | 0.74–1.04 | −0.14 | 0.055 | 0.87 | 0.75–1.00 |
| Cancer chemotherapy within 90 days: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.24 | 0.004 | 1.27 | 1.08–1.50 | 0.54 | <0.001 | 1.72 | 1.51–1.96 |
| Cancer immunotherapy within 90 days: No | Reference | – | – | – | ||||
| Yes | 0.42 | 0.128 | 1.52 | 0.89–2.60 | ||||
| Ventilator dependent within 48 h: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.45 | <0.001 | 1.57 | 1.33–1.85 | 0.54 | <0.001 | 1.72 | 1.47–2.00 |
| COPD: No | Reference | – | – | – | ||||
| Yes | 0.14 | 0.086 | 1.15 | 0.98–1.36 | ||||
| Current pneumonia: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.15 | 0.108 | 1.16 | 0.97–1.39 | 0.16 | 0.052 | 1.18 | 1.00–1.39 |
| Congestive heart failure within 30 days: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.29 | 0.003 | 1.34 | 1.10–1.63 | 0.31 | <0.001 | 1.37 | 1.14–1.64 |
| Myocardial infarction within 6 months: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.46 | 0.015 | 1.58 | 1.09–2.29 | 0.49 | 0.005 | 1.64 | 1.16–2.31 |
| Previous cardiac surgery: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.27 | 0.032 | 0.76 | 0.60–0.98 | −0.23 | 0.042 | 0.8 | 0.64–0.99 |
| Previous PVD with symptoms: No | Reference | – | – | – | ||||
| Yes | 0.53 | 0.003 | 1.69 | 1.19–2.41 | ||||
| Acute renal failure within 24 h: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.13 | 0.093 | 1.14 | 0.98–1.33 | 0.19 | 0.01 | 1.21 | 1.05–1.40 |
| Dialysis within 14 days: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.25 | 0.002 | 1.28 | 1.09–1.50 | 0.41 | <0.001 | 1.51 | 1.31–1.75 |
| Advanced cancer with multiple metastases: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.66 | <0.001 | 1.93 | 1.64–2.27 | 0.89 | <0.001 | 2.43 | 2.12–2.79 |
| No neoplastic disease | Reference | – | – | – | Reference | – | – | – |
| Benign tumor | −0.08 | 0.739 | 0.92 | 0.56–1.51 | −0.17 | 0.435 | 0.84 | 0.55–1.30 |
| Malignant tumor | 0.54 | <0.001 | 1.72 | 1.53–1.92 | 0.56 | <0.001 | 1.74 | 1.58–1.92 |
| WBC: ≥3000/mL | Reference | – | – | – | Reference | – | – | – |
| <3000/mL | 0.29 | <0.001 | 1.34 | 1.20–1.49 | 0.33 | <0.001 | 1.39 | 1.26–1.53 |
| Hemoglobin: ≥10 g/dL | Reference | – | – | – | ||||
| < 10 g/dL | 0.12 | 0.113 | 1.13 | 0.97–1.30 | ||||
| Hematocrit: ≥30% | Reference | – | – | – | Reference | – | – | – |
| <30% | −0.11 | 0.028 | 0.9 | 0.82–0.99 | −0.14 | 0.063 | 0.87 | 0.75–1.01 |
| Plate count: ≥150 000/mL | Reference | – | – | – | Reference | – | – | – |
| ≤100 000, and <150 000/mL | 0.29 | <0.001 | 1.33 | 1.19–1.49 | 0.26 | <0.001 | 1.29 | 1.17–1.43 |
| ≤50 000, and <100 000/mL | 0.44 | <0.001 | 1.56 | 1.34–1.81 | 0.47 | <0.001 | 1.59 | 1.40–1.82 |
| <50 000/mL | 0.98 | <0.001 | 2.67 | 2.10–3.40 | 0.98 | <0.001 | 2.67 | 2.11–3.37 |
| Albumin: ≥3.0 g/dL | Reference | – | – | – | Reference | – | – | – |
| ≤2.0, <3.0 g/dL | 0.38 | <0.001 | 1.47 | 1.33–1.62 | 0.58 | <0.001 | 1.78 | 1.63–1.94 |
| <2.0 g/dL | 0.71 | <0.001 | 2.04 | 1.79–2.33 | 0.96 | <0.001 | 2.6 | 2.32–2.92 |
| Total bilirubin: <3.0 mg/dL | Reference | – | – | – | Reference | – | – | – |
| ≥3.0 mg/dL | 0.53 | <0.001 | 1.71 | 1.43–2.03 | 0.63 | <0.001 | 1.88 | 1.60–2.20 |
| AST: ≤35 U/L | Reference | – | – | – | Reference | – | – | – |
| >35 U/L | 0.39 | <0.001 | 1.48 | 1.35–1.62 | 0.4 | <0.001 | 1.49 | 1.36–1.64 |
| ALT: =<35 U/L | Reference | – | – | – | ||||
| >35 U/L | −0.14 | 0.012 | 0.87 | 0.78–0.97 | ||||
| ALP: =<340 U/L | Reference | – | – | – | Reference | – | – | – |
| >340 U/L | 0.36 | <0.001 | 1.43 | 1.30–1.57 | 0.39 | <0.001 | 1.48 | 1.37–1.61 |
| BUN: =<20 mg/dL | Reference | – | – | – | Reference | – | – | – |
| > 20 mg/dL | 0.46 | <0.001 | 1.59 | 1.44–1.75 | 0.44 | <0.001 | 1.56 | 1.43–1.69 |
| Serum creatinine: < 2.0 mg/dL | Reference | – | – | – | Reference | – | – | – |
| ≥2.0 mg/dL | 0.67 | <0.001 | 1.96 | 1.76–2.18 | 0.57 | <0.001 | 1.77 | 1.61–1.96 |
| Serum sodium: <130 mEq/L | Reference | – | – | – | Reference | – | – | – |
| ≥130, and < 147 mEq/L | −0.36 | <0.001 | 0.7 | 0.60–0.81 | −0.35 | <0.001 | 0.7 | 0.61–0.80 |
| ≥147 mEq/L | 0.09 | 0.494 | 1.09 | 0.85–1.41 | 0.02 | 0.859 | 1.02 | 0.81–1.29 |
| CRP: <10 mg/dL | Reference | – | – | – | Reference | – | – | – |
| ≥10, and < 20 mg/dL | −0.34 | <0.001 | 0.71 | 0.64–0.79 | −0.28 | <0.001 | 0.76 | 0.69–0.83 |
| ≥20 mg/dL | −0.56 | <0.001 | 0.57 | 0.51–0.63 | −0.51 | <0.001 | 0.6 | 0.55–0.66 |
| INR of PT: ≤1.1 | Reference | – | – | – | Reference | – | – | – |
| 1.1<, <1.67 | 0.15 | 0.001 | 1.16 | 1.06–1.27 | 0.19 | <0.001 | 1.21 | 1.12–1.30 |
| ≥1.67 | 0.64 | <0.001 | 1.89 | 1.63–2.19 | 0.57 | <0.001 | 1.77 | 1.54–2.03 |
| Etiology of ADP | ||||||||
| Idiopathic: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.32 | 0.009 | 1.37 | 1.08–1.75 | 0.26 | 0.027 | 1.3 | 1.03–1.65 |
| Bowel perforation: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.3 | <0.001 | 1.34 | 1.19–1.52 | 0.36 | <0.001 | 1.44 | 1.23–1.69 |
| Bowel ischemia: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.47 | <0.001 | 1.6 | 1.32–1.95 | 0.53 | <0.001 | 1.69 | 1.39–2.06 |
| Bowel obstruction: No | Reference | – | – | – | ||||
| Yes | 0.19 | 0.048 | 1.21 | 1.00–1.47 | ||||
| Inflammation of the biliary tract: No | Reference | – | – | – | ||||
| Yes | −0.3 | 0.099 | 0.74 | 0.52–1.06 | ||||
| Appendicitis, diverticulitis, or IBD: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.24 | 0.046 | 0.79 | 0.62–1.00 | −0.21 | 0.06 | 0.81 | 0.66–1.01 |
| Others: No | Reference | – | – | – | ||||
| Yes | 0.14 | 0.145 | 1.15 | 0.95–1.38 | ||||
| Organ or site responsible for ADP | ||||||||
| Esophagus: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.45 | 0.117 | 1.56 | 0.89–2.73 | 0.61 | 0.014 | 1.84 | 1.13–2.98 |
| Appendix: No | Reference | – | – | – | Reference | – | – | – |
| Yes | −0.55 | 0.001 | 0.58 | 0.41–0.80 | −0.56 | <0.001 | 0.57 | 0.42–0.78 |
| Biliary tract: No | Reference | – | – | – | ||||
| Yes | 0.44 | 0.017 | 1.55 | 1.08–2.22 | ||||
| Others: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.19 | 0.159 | 1.21 | 0.93–1.58 | 0.42 | 0.002 | 1.52 | 1.17–1.99 |
| Unknown: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.65 | <0.001 | 1.92 | 1.39–2.63 | 0.75 | <0.001 | 2.11 | 1.55–2.88 |
| Stomach, duodenum, or small bowel: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.57 | <0.001 | 1.77 | 1.45–2.15 | 0.64 | <0.001 | 1.89 | 1.52–2.36 |
| Large bowel: No | Reference | – | – | – | Reference | – | – | – |
| Yes | 0.64 | <0.001 | 1.89 | 1.55–2.31 | 0.67 | <0.001 | 1.96 | 1.58–2.44 |
| (Intercept) | −5.61 | <0.001 | 0.00 | 0.00–0.01 | −5.40 | <0.001 | 0.00 | 0.00–0.01 |
Abbreviations: ADP, acute diffused peritonitis; CI, confidence interval; DIC, disseminated intra‐vascular coagulation; OR; odds ratio; SSI, surgical site infection.
The C‐index, which is the area under the ROC curve, and the 95% confidence interval (CI) were calculated with the bootstrap‐based optimism correction method to evaluate the discriminative performance of each model. 30 The C‐indices of perioperative mortality and 30‐day mortality were 0.859 (95% CI: 0.855–0.865) and 0.857 (95% CI: 0.853–0.864), respectively, which were comparable to those of the initially reported model. 7 Regarding postoperative morbidities, the C‐indices were >0.80 except for organ space SSI (C‐index:0.704), pneumonia (C‐index:0.799), and unplanned intubation (C‐index: 0.774) (Table 4). To summarize, these findings indicated the good discrimination performance of the risk model developed in the current study.
TABLE 4.
C‐index of the logistic regression models for predicting the outcome.
| Outcomes | C‐index | 95% CI |
|---|---|---|
| Mortality within 30 days | 0.857 | 0.853–0.864 |
| Perioperative mortality | 0.859 | 0.855–0.865 |
| Organ space SSI | 0.704 | 0.700–0.717 |
| Pneumonia | 0.799 | 0.794–0.810 |
| Unplanned intubation | 0.774 | 0.768–0.790 |
| On ventilator (>48 h) | 0.849 | 0.845–0.854 |
| Acute renal failure | 0.858 | 0.852–0.867 |
| Impaired consciousness (>24 h) | 0.856 | 0.849–0.867 |
| Cardio‐respiratory arrest | 0.817 | 0.810–0.834 |
| Postoperative transfusion | 0.81 | 0.806–0.817 |
| Sepsis | 0.837 | 0.832–0.842 |
| DIC | 0.823 | 0.819–0.834 |
Abbreviations: CI, confidential interval; DIC, disseminated intravascular coagulation; SSI, surgical site infection.
The calibration plots of the risk model for predicting each outcome are shown in Figure 2. The predicted value of all the surgical outcomes except unplanned intubation, cardio‐respiratory arrest, and DIC was highly fitted with the actual probability from the lower to the higher risk groups. The consistency of the calibration plots in the three models remained in the low‐risk group; however, they overpredicted the incidence in the higher‐risk group.
FIGURE 2.
Calibration plot of prediction models for mortalities and morbidities of acute diffused peritonitis. Observed event rates (95% CI) were plotted with predicted event rates for each prediction model in the whole dataset (n = 45 379).
4. DISCUSSION
In the present study, new risk models for predicting postoperative mortality and morbidities of ADP were revised with newly added variables, causative etiology, and responsible organ or site and validated using the recent NCD dataset. The discrimination performance of the developed risk models of perioperative mortalities and the 30‐day mortality was sufficiently high, with 0.857 (95% CI: 0.853–0.864) and 0.859 (95% CI: 0.855–0.865) of the C‐index, respectively, which were comparable to those of initial report. 7 The calibration assessment showed all the models except unplanned intubation, cardio‐respiratory arrest, and DIC were fitted well from the lower to the higher risk groups, suggesting reliable risk models in clinical use. Therefore, the newly developed risk models were considered valuable and reliable for predicting postoperative outcomes of ADP.
Given the wide variation in causes of ADP, it is crucial to define ADP when developing a risk model precisely. In this study, the latest dataset from the NCD was utilized, providing new relevant variables such as ADP classification, etiology, and causative organ. This addition was aimed at addressing the limitations of the previous study. To our knowledge, this database represents the largest national registry with comprehensive clinical variables for defining the ADP population. Notably, within the entire cohort, 92.7% had secondary peritonitis, which often requires prompt surgical intervention. Therefore, the newly developed risk models exhibit significant strength in predicting surgical outcomes for these patients.
In addition, a novelty of the current study is to reveal the impact of the causative etiology or causative organ of ADP for predicting surgical outcomes. These variables were selected as a predictor of the developed risk models, which was considered to complement the limitation of the previous study. The number of factors comprising the risk models increased; however, most were factors selected in the previous studies.
A recent report of gastroenterological surgery in Japan using the NCD between 2011 and 2019 showed that surgical cases of ADP have been increasing and that the 30‐ and 90‐day mortalities have been steadily decreasing in these years, while the postoperative morbidities classified as Clavien–Dindo grade 3 or higher remained high. 16 Considering the transition, it is meaningful to develop a new risk model based on the current database for accurate prediction in future use. In this study, the incidence rate of the 30‐ and perioperative mortality were lower than those of the initial reports (7.2% vs. 8.9%, and 10.6% vs. 13.9%, respectively). 7 , 22 Possible reasons for this may be (1) improvement of perioperative management for ADP patients and (2) a contribution of the NCD risk calculator developed with the initial study, which may lead to excluding the most seriously ill patients of ADP for surgery.
When comparing the discrimination performance of the newly developed risk models with the initial risk models, the c‐indices of the new mortality models were compared with those of previous models (Table S3). Regarding the 10 morbidities, the eight morbidities, except “pneumonia” and “acute renal failure,” were newly selected as secondary outcomes in this study. The c‐index of the two common outcomes was slightly lower than those of the previous study. 22 However, it is considered that this result can not conclude which is superior to the other because their discrimination performance was not compared directly using the same dataset. In addition, the background difference in mortality between the initial and the present dataset mentioned above may need to be considered.
Regarding calibration of the risk prediction models, the predicted value of each surgical outcome using the developed risk model was highly consistent with the actual probability except for unplanned intubation, cardio‐respiratory arrest, and DIC (Figure 2). In the three models, poor calibration was found in the high‐risk group, suggesting that the reliability of the predictions from the models was inferior. A possible cause of this may be unknown preoperative or intraoperative factors not included in explanatory variables for developing the new risk models in this study. Except for the three models, the risk models for predicting surgical outcomes of ADP were reliable enough for clinical use.
In the newly developed risk models, the number of independent predictors increased compared with the previous models, reaching 55 for perioperative mortality. Indeed, the sample size for this study was five times larger (n = 45 379 vs. n = 8482). For the perioperative mortality, the top five were “age >81 (OR = 3.83)”, “advanced cancer with multiple metastases (OR = 2.43)”, “platelet count <50 000/mL (OR = 2.67)”, “serum albumin <2.0 g/dL (OR = 2.60)”, and “unknown site responsible for ADP (OR = 2.11)”, among which the three variables were consistent with those of 30‐day mortality. This finding suggested that preoperative severity, especially in elderly patients, was highly relevant to mortality after surgery. Naturally, this also applies to the risk models of the 10 morbidities.
ADP is a clinical syndrome rather than a definitive diagnosis, as the causative etiology varies widely from one setting to another, and it seems to be correlated to mortality. 32 For example, it is recognized by clinicians that ADP due to large bowel perforation must be more serious than that of the upper intestinal tract, such as gastric or duodenum. 32 , 33 To our knowledge, few studies have quantified a relative risk measure of the causative etiology. Therefore, revealing the impact of the causative etiology for predicting outcomes of ADP could be a novel finding of the present study. This study's new risk models selected causative etiology or responsible organ as a covariate. Etiology (“bowel perforation”, or “bowel ischemia”) and causative site (“esophagus”, “unknown”, “upper gastrointestinal tract”, or “lower gastrointestinal tract”) are shown to be relevant to perioperative mortality (Table 3). Regarding the 10 morbidities, the variable of the causative site of ADP was selected in the top five relevant predictors (Table S2). As expected, “esophagus” or “lower gastrointestinal tract” had a greater negative impact on morbidities, whereas “appendix” was less associated with severe complications. “Esophagus” was strongly relevant to the three outcomes associated with respiratory failures, such as “pneumonia”, “unplanned intubation,” and “on ventilator >48 h”. “Lower gastrointestinal tract” was relevant to the incidence of sepsis, DIC, and impaired consciousness. The definitive preoperative diagnosis of a detailed etiology of ADP is challenging; however, recent advances in imaging modalities could evaluate the etiology or causative organ (site) of ADP in most cases.
This study has several limitations. First, the newly developed risk models need to be validated using a truly external dataset, even though internal validation was performed in this study. In addition, validation using a database outside Japan is mandatory for evaluating generalization. Existing literature shows that etiologies of peritonitis vary by geographic location and local environmental factors with genetic predisposition. 21 Second, the present models seem to be complicated regarding many independent variables, which may burden the working time for data input. However, as far as Japanese users, these variables are already registered in the NCD system, so prediction calculated by the models is available automatically at participating hospitals.
5. CONCLUSION
We constructed and validated the new prediction models of postoperative surgical outcomes of ADP using the recent NCD dataset. Evaluation of predictive power and calibration of these risk models showed they were sufficiently reliable for clinical settings. The developed risk models in this study were updated and revised with the latest real‐world data, which can lead to providing realistic predictions. Since treatment decisions for ADP are made under dynamic circumstances, accurate prediction of postoperative complications is useful for shared decision‐making processes between patients and/or their families and medical providers.
AUTHOR CONTRIBUTIONS
Study concepts: Kitagawa, Oshikiri, Kakeji, Marubashi, Miyata. Study design: Sato, Marubashi, Oshikiri, Kakeji. Data acquisition: Kitagawa, Oshikiri, Kakeji. Quality control of data and algorithms: Kitagawa, Oshikiri, Kakeji, Marubashi, Miyata, Tachimori. Data analysis and interpretation: Sato, Hirakawa, Tachimori, Marubashi Kakeji. Statistical analysis: Hirakawa, Tachimori, Miyata. Manuscript preparation: Sato, Hirakawa, Marubashi. Manuscript editing: Sato, Marubashi. Manuscript review: Marubashi. All authors approved the final version of this manuscript.
FUNDING INFORMATION
This work was supported by the Japanese Society of Gastroenterological Surgery (APJ‐2019‐02).
CONFLICT OF INTEREST STATEMENT
Author Yuko Kitagawa is Editor in Chief of Annals of Gastroenterological Surgery. Author Yoshihiko Kakeji is an editorial board member of the Annals of Gastroenterological Surgery. Shinya Hirakawa and Hisateru Tachimori belong to an endowed course funded by Takeda Pharmaceutical Company Limited. Shinya Hirakawa, Hisateru Tachimori, and Hiroaki Miyata belong to a department that accepts financial support from National Clinical Database, Johnson & Johnson K.K., Nipro Corporation, and Intuitive Surgical Sàrl.
ETHICS STATEMENTS
Approval of the research protocol: The study protocol was approved by the Japanese Society of Gastroenterological Surgery and the Ethics Committee of Kobe University (approval number 20190128).
Informed Consent: N/A.
Registry and the Registration No. of the study/trial: N/A.
Animal Studies: N/A.
Supporting information
Table S1.
Table S2.
Table S3.
ACKNOWLEDGMENTS
We thank all the data managers and hospitals participating in the NCD project for their continued efforts to register surgical and other medical data.
Sato N, Hirakawa S, Marubashi S, Tachimori H, Oshikiri T, Miyata H, et al. Predicting surgical outcomes of acute diffuse peritonitis: Updated risk models based on real‐world clinical data. Ann Gastroenterol Surg. 2024;8:711–727. 10.1002/ags3.12800
DATA AVAILABILITY STATEMENT
Research data are not shared. We used (1) data from the National Clinical Database (NCD) registry in Japan.
REFERENCES
- 1. Sartelli M, Catena F, Ansaloni L, Coccolini F, Corbella D, Moore EE, et al. Complicated intra‐abdominal infections worldwide: the definitive data of the CIAOW study. World J Emerg Surg. 2014;9:37. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2. Gauzit R, Pean Y, Barth X, Mistretta F, Lalaude O, Top Study T. Epidemiology, management, and prognosis of secondary non‐postoperative peritonitis: a French prospective observational multicenter study. Surg Infect. 2009;10(2):119–127. [DOI] [PubMed] [Google Scholar]
- 3. Scapellato S, Parrinello V, Sciuto GS, Castorina G, Buffone A, Cirino E. Valuation on prognostic factors about secondary acute peritonitis: review of 255 cases. Ann Ital Chir. 2004;75(2):241–245; discussion 6. [PubMed] [Google Scholar]
- 4. Kumar D, Garg I, Sarwar AH, Kumar L, Kumar V, Ramrakhia S, et al. Causes of acute peritonitis and its complication. Cureus. 2021;13(5):e15301. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Ballus J, Lopez‐Delgado JC, Sabater‐Riera J, Perez‐Fernandez XL, Betbese AJ, Roncal JA. Surgical site infection in critically ill patients with secondary and tertiary peritonitis: epidemiology, microbiology and influence in outcomes. BMC Infect Dis. 2015;15:304. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Jang JY, Lee SH, Shim H, Choi JY, Yong D, Lee JG. Epidemiology and microbiology of secondary peritonitis caused by viscus perforation: a single‐center retrospective study. Surg Infect. 2015;16(4):436–442. [DOI] [PubMed] [Google Scholar]
- 7. Nakagoe T, Miyata H, Gotoh M, Anazawa T, Baba H, Kimura W, et al. Surgical risk model for acute diffuse peritonitis based on a Japanese nationwide database: an initial report on the surgical and 30‐day mortality. Surg Today. 2015;45(10):1233–1243. [DOI] [PubMed] [Google Scholar]
- 8. Adesunkanmi AR, Oseni SA, Adejuyigbe O, Agbakwuru EA. Acute generalized peritonitis in African children: assessment of severity of illness using modified APACHE II score. ANZ J Surg. 2003;73(5):275–279. [DOI] [PubMed] [Google Scholar]
- 9. Malik AA, Wani KA, Dar LA, Wani MA, Wani RA, Parray FQ. Mannheim peritonitis index and APACHE II‐‐prediction of outcome in patients with peritonitis. Ulus Travma Acil Cerrahi Derg. 2010;16(1):27–32. [PubMed] [Google Scholar]
- 10. Yelamanchi R, Gupta N, Durga CK, Korpal M. Comparative study between P‐POSSUM and Apache II scores in predicting outcomes of perforation peritonitis: prospective observational cohort study. Int J Surg. 2020;83:3–7. [DOI] [PubMed] [Google Scholar]
- 11. Kumar S, Gupta A, Chaudhary S, Agrawal N. Validation of the use of POSSUM score in enteric perforation peritonitis – results of a prospective study. Pan Afr Med J. 2011;9:22. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Nachiappan M, Litake MM. Scoring Systems for Outcome Prediction of patients with perforation peritonitis. J Clin Diagn Res. 2016;10(3):PC01–PC05. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Coombs L, Orlando A, Wang X, Shaw P, Rich AS, Lakhtakia S, et al. A machine learning framework supporting prospective clinical decisions applied to risk prediction in oncology. NPJ Digit Med. 2022;5(1):117. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Rodriguez PJ, Veenstra DL, Heagerty PJ, Goss CH, Ramos KJ, Bansal A. A framework for using real‐world data and health outcomes modeling to evaluate machine learning‐based risk prediction models. Value Health. 2022;25(3):350–358. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Takeuchi H, Miyata H, Gotoh M, Kitagawa Y, Baba H, Kimura W, et al. A risk model for esophagectomy using data of 5354 patients included in a Japanese nationwide web‐based database. Ann Surg. 2014;260(2):259–266. [DOI] [PubMed] [Google Scholar]
- 16. Watanabe M, Miyata H, Gotoh M, Baba H, Kimura W, Tomita N, et al. Total gastrectomy risk model: data from 20 011 Japanese patients in a nationwide internet‐based database. Ann Surg. 2014;260(6):1034–1039. [DOI] [PubMed] [Google Scholar]
- 17. Kurita N, Miyata H, Gotoh M, Shimada M, Imura S, Kimura W, et al. Risk model for distal gastrectomy when treating gastric cancer on the basis of data from 33 917 Japanese patients collected using a Nationwide web‐based data entry system. Ann Surg. 2015;262(2):295–303. [DOI] [PubMed] [Google Scholar]
- 18. Kobayashi H, Miyata H, Gotoh M, Baba H, Kimura W, Kitagawa Y, et al. Risk model for right hemicolectomy based on 19 070 Japanese patients in the National Clinical Database. J Gastroenterol. 2014;49(6):1047–1055. [DOI] [PubMed] [Google Scholar]
- 19. Matsubara N, Miyata H, Gotoh M, Tomita N, Baba H, Kimura W, et al. Mortality after common rectal surgery in Japan: a study on low anterior resection from a newly established nationwide large‐scale clinical database. Dis Colon Rectum. 2014;57(9):1075–1081. [DOI] [PubMed] [Google Scholar]
- 20. Kenjo A, Miyata H, Gotoh M, Kitagawa Y, Shimada M, Baba H, et al. Risk stratification of 7732 hepatectomy cases in 2011 from the National Clinical Database for Japan. J Am Coll Surg. 2014;218(3):412–422. [DOI] [PubMed] [Google Scholar]
- 21. Kimura W, Miyata H, Gotoh M, Hirai I, Kenjo A, Kitagawa Y, et al. A pancreaticoduodenectomy risk model derived from 8575 cases from a national single‐race population (Japanese) using a web‐based data entry system: the 30‐day and in‐hospital mortality rates for pancreaticoduodenectomy. Ann Surg. 2014;259(4):773–780. [DOI] [PubMed] [Google Scholar]
- 22. Saze Z, Miyata H, Konno H, Gotoh M, Anazawa T, Tomotaki A, et al. Risk models of operative morbidities in 16 930 critically ill surgical patients based on a Japanese Nationwide database. Medicine (Baltimore). 2015;94(30):e1224.26222854 [Google Scholar]
- 23. Blot S, De Waele JJ. Critical issues in the clinical management of complicated intra‐abdominal infections. Drugs. 2005;65(12):1611–1620. [DOI] [PubMed] [Google Scholar]
- 24. Pieracci FM, Barie PS. Management of severe sepsis of abdominal origin. Scand J Surg. 2007;96(3):184–196. [DOI] [PubMed] [Google Scholar]
- 25. Sartelli M, Chichom‐Mefire A, Labricciosa FM, Hardcastle T, Abu‐Zidan FM, Adesunkanmi AK, et al. The management of intra‐abdominal infections from a global perspective: 2017 WSES guidelines for management of intra‐abdominal infections. World J Emerg Surg. 2017;12:29. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Marubashi S, Takahashi A, Kakeji Y, Hasegawa H, Ueno H, Eguchi S, et al. Surgical outcomes in gastroenterological surgery in Japan: report of the National Clinical Database 2011–2019. Ann Gastroenterol Surg. 2021;5(5):639–658. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kanaji S, Takahashi A, Miyata H, Marubashi S, Kakeji Y, Konno H, et al. Initial verification of data from a clinical database of gastroenterological surgery in Japan. Surg Today. 2019;49(4):328–333. [DOI] [PubMed] [Google Scholar]
- 28. Ordonez CA, Puyana JC. Management of peritonitis in the critically ill patient. Surg Clin North Am. 2006;86(6):1323–1349. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Holzheimer RG, Muhrer KH, L'Allemand N, Schmidt T, Henneking K. Intraabdominal infections: classification, mortality, scoring and pathophysiology. Infection. 1991;19(6):447–452. [DOI] [PubMed] [Google Scholar]
- 30. Frank E, Harrell J. Regression Modeling Strategies With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. London: Springer; 2015. [Google Scholar]
- 31. Noma H, Shinozaki T, Iba K, Teramukai S, Furukawa TA. Confidence intervals of prediction accuracy measures for multivariable prediction models based on the bootstrap‐based optimism correction methods. Stat Med. 2021;40(26):5691–5701. [DOI] [PubMed] [Google Scholar]
- 32. Chichom‐Mefire A, Fon TA, Ngowe‐Ngowe M. Which cause of diffuse peritonitis is the deadliest in the tropics? A retrospective analysis of 305 cases from the south‐west region of Cameroon. World J Emerg Surg. 2016;11:14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Agarwal N, Saha S, Srivastava A, Chumber S, Dhar A, Garg S. Peritonitis: 10 years' experience in a single surgical unit. Trop Gastroenterol. 2007;28(3):117–120. [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Table S1.
Table S2.
Table S3.
Data Availability Statement
Research data are not shared. We used (1) data from the National Clinical Database (NCD) registry in Japan.