Background: Bariatric surgery is popularly used to treat or prevent morbidity in severely obese patients. Severe complications are rare, but their early detection has a significant impact on clinical outcomes. We aimed to determine whether blood tests in the first few postoperative days are reliable predictors for complications.
Methods: We retrospectively analyzed 1073 patients who underwent laparoscopic bariatric surgery between 2009 and 2018 at our center. Clinical outcome was correlated with postoperative serum C-reactive protein (CRP), white blood cell count, and vital signs, analyzed using a receiver operating characteristic (ROC) curve. A total of 570 procedures between 2009 and 2015 were used to calculate the best cutoff values (calculation group), which were validated with 330 different patients operated upon between 2016 and 2018 (validation group).
Results: Twenty-four patients (4.2%) developed anastomotic or staple-line leakages in the calculation group. The ROC curve showed a good reliability for CRP levels on day 2 (area under the ROC curve=0.86); the highest Youden index existed for a cutoff of 119 mg/L. White blood cell count and heart rate were poor predictors. Even though several characteristics differed in the validation cohort, test quality of the cutoff was high (sensitivity, 71.4%; specificity, 94.9%; positive predictive value, 23.8%; negative predictive value, 99.3%). The prediction was excellent especially for leakages appearing on days 2 to 9 (sensitivity 100.0%, negative predictive value 100%). Leakages from day 10 were rare and prediction poor (sensitivity 0%).
Conclusions: A CRP level on day 2 <120 mg/L is a good predictor of a postoperative course without leakage, even though the predictive value goes down for late-appearing events. An earlier CRP measurement added no predictive benefit. The cutoff value was validated in an internal cohort and could be applied to different populations.