TO THE EDITOR: I appreciate this informative article on irritable bowel syndrome (IBS). However, I do not agree with the SORT recommendation, “The absence of abdominal pain can be used to rule out IBS.” Given the data presented in Table 1 of the article and the cited systematic review,¹ the clinical evidence does not support this assertion.

Based on data in Table 1, the positive likelihood ratio of abdominal pain in diagnosing IBS is about 1.3, and the negative likelihood ratio is about 0.31. A negative likelihood ratio in this range can assist in ruling out a disorder, but does not definitively rule it out.² For abdominal pain to rule out a disorder, the patient's pretest probability must be considered. For instance, the disorder could not be ruled out in a patient with a pretest probability of 50 percent, which would result in a posttest probability of 24 percent. However, it would be useful to rule out the disease in a patient with a pretest probability of 10 percent, which would result in a posttest probability of 3 percent. Successfully ruling out the diagnosis of IBS in many patients would likely take more than the presence or absence of any one symptom, such as abdominal pain.

IN REPLY: In our article, we summarized information from a 2008 systematic review of the accuracy of individual symptoms to diagnose IBS.¹ For abdominal pain, the sensitivity is 90 percent and the specificity is 32 percent (positive predictive value = 9 percent; negative predictive value = 97 percent). The negative likelihood ratio is 0.016 to 0.034 (when weighted for a prevalence of 5 to 10 percent). Using the estimated prevalence of IBS in north America of 5 to 10 percent,² the posttest probability for IBS in a patient with abdominal pain is 6.5 to 12.8 percent, and the posttest probability for IBS in a patient without abdominal pain is 1.6 to 3.4 percent.

Because individual symptoms (e.g., abdominal pain, diarrhea, constipation) lack sufficient sensitivity and specificity to accurately diagnose IBS, clinical diagnostic criteria were developed. Of these, the Manning criteria are the most extensively studied (sensitivity, 63 to 90 percent; specificity, 70 to 93 percent).³, ⁴ The Rome I criteria in 1990 developed a consensus definition and criteria (sensitivity, 65 to 85 percent; specificity, 70 to 100 percent).³, ⁴ These were revised in 1999 with the Rome II criteria (sensitivity, 64 to 89 percent; specificity, 66 to 73 percent), and in 2006 with the Rome III criteria (sensitivity, 81 percent; specificity, 60 percent).³, ⁴ There have been eight validation trials for the Manning criteria, four for the Rome I criteria, three for the Rome II criteria, and none for the Rome III criteria.⁴ All of these criteria include abdominal pain, which is required for the diagnosis of IBS using the Rome II and III criteria.

IBS is a complex disorder with nonspecific symptoms defined as abdominal discomfort or pain associated with altered bowel habits for at least three days per month in the previous three months, with the absence of organic disease. An accurate diagnosis of IBS is important to minimize risks and reduce unnecessary medical procedures and tests while reducing cost. Without specific biomarkers or genetic tests, positive clinical diagnostic criteria in the absence of red flags, with positive history and physical examination findings, are the best way to diagnose IBS.