Frobert and colleagues [1] found no abnormalities in patients with angina-like chest pain and normal coronary angiograms and concluded that "the rationale for routine esophageal investigations in this patient group is questionable." This observation flies in the face of evidence from throughout the world that an esophageal source of pain in these patients is common. Studies documenting a high prevalence of reflux (40% to 60%) in patients with unexplained angina-like pain have been reported over the past decade from four countries [2-5].

What might explain this new, apparently aberrant, observation? Selection bias may have occurred among these patients, who were from a group of cardiology patients from which patients with reflux were selected out; notably, these patients were from Scandinavia, where omeprazole is widely used. The data analysis may also be an explanation. Although Frobert and colleagues reported no differences in the mean time of acid exposure ("the reflux index"), they did not indicate how many patients in their patient groups actually had abnormal reflux. Patients were also discussed only as groups, possibly masking an association of reflux with chest pain in specific persons, regardless of total reflux time. This is suggested by the observation, portrayed in Frobert and colleagues' Figure 2, that there was excellent correlation between the number of pain episodes grouped in 2-hour periods and the number of corresponding reflux episodes.

We are concerned that these biases and methods of data analysis led Frobert and colleagues to "find what they sought," a phenomenon that is likely to apply to both cardiologists and gastroenterologists but in opposite ways. The truth probably lies somewhere toward the middle. We suggest that one not extrapolate from a single study to the conclusion that this report is the truth and all others are incorrect. Abandoning esophageal investigation in patients with unexplained chest pain would be a mistake.