|| Checking for direct PDF access through Ovid
Gallbladder cancer is a rare neoplasm with a poor prognosis. Early diagnosis and correct treatment strategy is important. The aim of this study was to identify predictors for incidental gallbladder cancer.Data from cholecystectomies registered in the nationwide Swedish Register for Gallstone Surgery between 2007 and 2014 were analyzed for incidental gallbladder cancer. Exclusion criteria were patients with a gallbladder not sent for histopathology, preoperative suspicion of polyps/gallbladder cancer, and indication for operation for other reasons than gallstone disease. Predictive factors for incidental gallbladder cancer were identified using multivariable logistic regression.A total of 86,154 procedures were registered in the Swedish Register for Gallstone Surgery. Of these, 36,355 patients were included in the analysis, and 215 of the included patients had incidental gallbladder cancer (0.59%). Mean age was 70 ± 11 years for index cases and 54 ± 16 years for the control group, and 80% of cases and 60% of controls were female. Predictors for incidental gallbladder cancer were older age (odds ratio = 1.08; P < .001), female sex (odds ratio = 3.58; P < .001), previous cholecystitis (odds ratio = 1.37; P = .045), and the combination of acute cholecystitis without jaundice (odds ratio = 1.39; P = .041) and jaundice without acute cholecystitis (odds ratio = 2.02; P = .009). A preoperative risk model including these factors gave an area under receiver operating characteristic curve of 0.82. By adding macroscopic evaluation of the gallbladder by the surgeon, the area under receiver operating characteristic curve increased to 0.87. Intraoperatively suspected gallbladder cancer was confirmed as cancer in 31% of the cases.Incidental gallbladder cancer is more likely to be diagnosed in older patients, women, and after previous cholecystitis. Jaundice and acute cholecystitis were also shown to be important risk factors. Intraoperative inspection of the gallbladder improved the risk model.