OBJECTIVE

The aim of this study was to assess and compare the diagnostic value of elements of the disease history, physical examination, and routine laboratory tests in patients with suspected acute left-sided colonic diverticulitis (ALCD).

BACKGROUND

Misdiagnosis rates for diverticulitis vary in literature between 34% and 68% which needs improvement. Because of the frequent misdiagnosis, liberal use of imaging has been recommended. Before making a plea for routine imaging, the diagnostic accuracy of different variables of disease history, physical examination, and routine laboratory tests needs to be specified.

METHODS

All patients seen on the emergency department because of acute abdominal pain suspected of ALCD in whom an abdominal computed tomography was performed, between January 2002 and March 2006, were studied. Univariate logistic regression was used to study differences in patients' characteristics and symptoms, findings at physical examination and routine laboratory tests between patients with and without ALCD. Independent predictors to the risk of ALCD were identified using multivariate logistic regression and used to create a clinical scoring system.

RESULTS

Of 1290 patients with acute abdominal pain, 287 patients were eligible for analysis. Acute left-sided colonic diverticulitis was the final diagnosis in 124 patients (43%). Age, 1 or more previous episodes, localization of symptoms in the lower left abdomen, aggravation of pain on movement, the absence of vomiting, localization of abdominal tenderness in the lower left abdomen, and C-reactive protein 50 or more were found to be independent predictors of ALCD. A nomogram was constructed based on these independent predictors with a diagnostic accuracy of 86%.

CONCLUSIONS

This study showed that the clinical diagnosis of diverticulitis is difficult to make but can be improved using a clinical scoring system. In case of a high chance of ALCD based on the nomogram, additional imaging may not be needed.