OBJECTIVES--To investigate the association between exertional chest pain and gastroesophageal reflux in patients with normal coronary angiograms and in controls by measuring oesophageal pH during treadmill exercise tests and to compare the results with routine ambulatory monitoring. DESIGN--Case control study. SETTING--Tertiary referral cardiac unit. PATIENTS--50 consecutive patients with chest pain and completely normal coronary angiograms and 16 controls with coronary artery stenoses. MAIN OUTCOME MEASURES--Episodes of acid reflux and chest pain during treadmill exercise; a symptom index expressing the percentage of episodes of pain related to acid reflux during ambulatory monitoring. RESULTS--Four (8%) patients and two (12%) controls had reflux during treadmill exercise (NS). 32 (64%) and 16 (100%) reported chest pain, but only three (6%) and two (12%) had coincident reflux (NS). Reflux was as frequent before, during, and after treadmill exercise (five (8%) v six (9%) v two (3%)) in the 66 subjects; (NS). 19 (38%) patients and three (19%) controls had abnormal reflux on ambulatory monitoring (NS). Eight (16%) and three (19%) had a symptom index > 50%, but six and two of these reported pain without coincident reflux during treadmill exercise. CONCLUSION--There are many potential causes of chest pain in patients with angiographically normal coronary arteries. Although gastroesophageal reflux is commonly implicated and many patients have a high incidence of spontaneous reflux during ambulatory monitoring, it rarely occurs during exertion and the association with chest pain is poor.