Renal colic. Utility of the plain abdominal roentgenogram.

We conducted a retrospective cohort study of 85 consecutive symptomatic patients to evaluate the sensitivity, specificity, and predictive value of plain abdominal roentgenography (PAR) compared with clinical evaluation alone in diagnosis and treatment of renal colic. With a positive intravenous pyelogram and/or actual stone retrieval used as the gold standard, 72 patients had documented stones. The calculated sensitivity and specificity for PAR were 58% (95% confidence interval, 47% to 69%) and 69% (95% confidence interval, 44% to 94%), respectively. In this population with a stone prevalence of 85%, the positive predictive value of PAR was 91%. A clinical scoring system based on signs and symptoms fared as well as PAR (sensitivity, 73%; specificity, 46%; and positive predictive value, 88%). The strategy of clinical scoring followed by selective use of intravenous pyelography was more cost-effective than strategies using PAR. We conclude that the PAR has low sensitivity and specificity and improves the predictive value only marginally, and its routine use is not cost-effective.