Randomized Clinical Trial of Morphine in Acute Abdominal Pain

Study objective Administration of analgesia to patients with acute abdominal pain is controversial. We test the hypothesis that morphine given to emergency department (ED) patients with acute abdominal pain will reduce discomfort and improve clinically important diagnostic accuracy. Methods Pain was measured with a standard 0- to 100-mm visual analog scale. ED patients with acute abdominal pain were randomized in a double-blind fashion to 0.1 mg/kg intravenous morphine or placebo. The primary endpoint was the difference between the 2 study arms in clinically important diagnostic accuracy. Clinically important diagnostic accuracy was defined a priori by its complement, clinically important diagnostic error, using 2 independent, blinded investigators to identify any discordance between the provisional and final diagnoses that might adversely affect the patient’s health status. The provisional diagnosis was provided by an ED attending physician, who examined the patient only once, 15 minutes after administration of the study agent. The final diagnosis was obtained through follow-up at least 6 weeks after the index ED visit. Results We randomized 160 patients, of whom 153 patients were available for analysis, 78 patients in the morphine group and 75 patients in the placebo group. Baseline features were similar in both groups, including initial median visual analog scale scores of 98 mm and 99 mm. The median decrease in visual analog scale score at 15 minutes was 33 mm in the morphine group and 2 mm in the placebo group. There were 11 instances of diagnostic discordance in each group, for a clinically important diagnostic accuracy of 86% (67/78) in the morphine group and 85% (64/75) in the placebo group. The difference in clinically important diagnostic accuracy between the 2 groups was 1% (95% confidence interval [CI] −11% to 12%). Analysis by efficacy and intention to treat yielded similar results. κ for interobserver concordance in classification of clinically important diagnostic accuracy was 0.94 (95% CI 0.79 to 1.00). No patients required naloxone. Conclusion Although administration of intravenous morphine to adult ED patients with acute abdominal pain could lead to as much as a 12% difference in diagnostic accuracy, equally favoring opioid or placebo, our data are most consistent with the inference that morphine safely provides analgesia without impairing clinically important diagnostic accuracy.