Safety of dipyridamole testing in 73,806 patients: the Multicenter Dipyridamole Safety Study

Dipyridamole imaging is widely used as an alternative to exercise testing to identify and risk stratify patients with coronary artery disease. Safety data on intravenous dipyridamole stress testing has been derived largely from individual institutional data.

Data were collected retrospectively by 85 coinvestigators from 73,806 patients who underwent intravenous dipyridamole stress imaging in 59 hospitals and 19 countries to determine the incidence of major adverse reactions during testing. The dose of dipyridamole infused was 0.56 mg/kg in 64,740 patients, 0.74 mg/kg in 6551 patients, and 0.84 mg/kg in 2515 patients. Combined major adverse events among the entire 73,806 patients included seven cardiac deaths (0.95 per 10,000), 13 nonfatal myocardial infarctions (1.76 per 10,000), six nonfatal sustained ventricular arrhythmias (0.81 per 10,000) (ventricular tachycardia in two and ventricular fibrillation in four), nine transient cerebral ischemic attacks (1.22 per 10,000), (with speech or motor deficit), one stroke, and nine severe bronchospasms (1.22 per 10,000) (one intubation and eight near intubations). In addition to the safety data, detailed demographic, peripheral hemodynamic, side effect, and concomitant drug data were examined in a subgroup of 3751 patients. End points from subsets of patients were compared with those of the group as a whole. Multivariate analysis revealed that dipyridamole-induced chest pain was more common in patients less than 70 years old (p = 0.0017), those with a history of coronary revascularization (p = 0.002), or patients taking aspirin (p = 0.0001). Minor noncardiac side effects were less frequent among the elderly (p = 0.0053) and more frequent in women (p = 0.0001) and patients taking maintenance aspirin (p = 0.0034). When a patient was judged on the basis of the adequacy of hemodynamic response to be a dipyridamole "nonresponder" (< 10 mm Hg drop in systolic blood pressure and 10 beats/min increase in heart rate), the only significant predictor was angiotensin-converting enzyme inhibitor intake (p = 0.0025). Inferoposterior hypoperfusion was significantly more frequent in patients with dipyridamole-induced hypotension: 57% (44/77) (p < 0.0001) of those who had hypotension and 89% (8/9) (p = 0.0076) who had severe symptomatic bradyarrhythmias displayed inferoposterior defects on thallium scanning. Caffeine levels were determined in 391 consecutive patients: levels greater than 5 mg/L were observed in only eight patients (2%), suggesting that methylxanthine levels sufficient to alter the hemodynamic response to dipyridamole resulting in suboptimal hyperemic stress are unlikely when patients take nothing by mouth after midnight.

The risk of serious dipyridamole-induced side effects is very low and is comparable to that reported for exercise testing in a similar patient population.