Increased accuracy in diagnosing diverticulitis using predictive clinical factors.

Abstract

Background: The aim of this study was to identify clinical factors leading to increased diagnostic accuracy for acute colonic diverticulitis.

Methods: Patients with clinical suspicion of acute colonic diverticulitis verified with computed tomography (CT) from two hospitals in Sweden between 9 January 2017 and 31 October 2017 were prospectively included. Symptoms, comorbidities, and laboratory results were documented. Candidate variables were analyzed using logistic regression, and the final variable set that yielded the most accurate predictions was identified using least absolute shrinkage and selection operator regression and evaluated using the area under the receiver operating characteristic (ROC) curve.

Results: In total, 146 patients were included (73% women; median age 68 years; age range, 50-94 years). The clinical diagnostic accuracy was 70.5%. In the multiple logistic regression analysis, gender (female vs male odds ratio [OR]: 4.82; confidence interval [CI], 1.56-14.91), age (OR, 0.92; 95% CI, 0.87-0.98), pain on the lower left side of the abdomen (OR, 15.14; 95% CI, 2.65-86.58), and absence of vomiting (OR, 14.02; 95% CI, 2.90-67.88) were statistically significant and associated with the diagnosis of CT-verified diverticulitis. With seven predictors (age, gender, urinary symptoms, nausea, temperature, C-reactive protein, and pain left lower side), the area under the ROC curve was 0.82, and a formula was developed for calculating a risk score.

Conclusion: We present a scoring system using common clinical variables that can be applied to patients with clinical suspicion of colonic diverticulitis to increase the diagnostic accuracy. The developed scoring system is available for free of charge at https://phille-wagner.shinyapps.io/Diverticulitis_risk_model/.