Back to 2011 Program

Introduction: Gangrenous cholecystitis (GC) represents an advanced variant of acute gallbladder disease which carries a significant risk for increased morbidity and mortality. There are no diagnostic criteria for GC, making it a challenging diagnosis to make pre-operatively. We set out to determine which factors were predictive for GC in order to better identify which patients might benefit from early surgical management.
Methods:The medical records of 200 patients (76 with gangrenous and 124 with acute, non-gangrenous cholecystitis) treated at The Mount Sinai Medical Center from March 2003 to September 2009 were retrospectively reviewed. Specifically, presenting symptoms, physical examination, laboratory values, radiographic findings (ultrasonography and computed tomography), perioperative data and pathological findings were recorded. Univariate analysis was carried out with a two-tailed chi-square test for each of the categorical variables, and a two-sample t-test with the Welch correction for the continuous variables. Multivariate analysis was then performed using a binary logistical linear regression model and the model of best fit was determined. P-values of 0.05 or less were considered to indicate statistical significance. SPSS for Windows (Version 18.0.2, 2010 Chicago: SPSS Inc.) was used for all analysis.
Results:On univariate analysis, multiple comorbidities were associated with a significantly increased risk of GC, including age over 50 years (p=0.002), diabetes mellitus (p=0.002), coronary artery disease (CAD) (p<0.001), hypertension (p<0.001), hyperlipidemia (p=0.002) and steroid use (p=0.008). In addition, gangrenous pathology presented more frequently with fever >38C (p=0.010), elevated heart rate >100 (p=0.025), WBC >13,000 (p=0.002), and neutrophil shift >75% (p<0.001). Multivariate analysis confirmed that CAD, alcohol use, steroid use, nausea, leukocytosis, neutrophil shift and bilirubin levels displayed the highest predictive capacity for GC. Using a logistical regression model of best fit, we demonstrated that a scoring system utilizing the above factors could diagnose GC with a sensitivity of 81%, specificity of 77%, PPV 70% and NPV of 90%. GC was associated with a significantly higher rate of post-operative complications when compared to acute cholecystitis (OR=1.99, p=0.05).
Discussion: GC represents a complex clinical problem which is frequently difficult to diagnose preoperatively and often results in poor clinical outcomes. This study shows that a host of risk factors are correlated with GC and, therefore, may potentially be used for diagnostic purposes. The high sensitivity of our prediction model shows that relatively few clinical variables may be used to objectively stratify patients at risk for GC and, thus, determine which individuals may benefit from prompt surgical intervention.