Marked exercise‐induced T‐wave heterogeneity in symptomatic diabetic patients with nonflow‐limiting coronary artery stenosis

Background T‐wave heterogeneity ( TWH ) independently predicted cardiovascular mortality in Health Survey 2000 based on 12‐lead ECG s recorded at rest. We investigated whether TWH is elevated during exercise tolerance testing (ETT) in symptomatic diabetic patients with nonflow‐limiting coronary artery stenosis compared to control subjects without diabetes. Methods Cases were all patients ( n = 20) with analyzable ECG recordings during both rest and ETT who were enrolled in the Effects of Ranolazine on Coronary Flow Reserve ( CFR ) in Symptomatic Patients with Diabetes and Suspected or Known Coronary Artery Disease ( RAND ‐ CFR ) study ( NCT 01754259); median CFR was 1.44; 80% of cases had CFR <2. Control subjects ( n = 9) were nondiabetic patients who had functional flow reserve ( FFR ) >0.8, a range not associated with inducible ischemia. TWH was analyzed from precordial leads V 4 , V 5 , and V 6 by second central moment analysis, which assesses the interlead splay of T‐waves about a mean waveform. Results During exercise to similar rate‐pressure products ( p = .31), RAND ‐ CFR patients exhibited a 49% increase in TWH during exercise (rest: 49 ± 5 μV; exercise: 73 ± 8 μV, p = .003). By comparison, in control subjects, TWH was not significantly altered (rest: 52 ± 11 μV; ETT : 38 ± 5 μV, p = .19). ETT ‐induced ST ‐segment depression >1 mm ( p = .11) and T peak ‐T end ( p = .18) and QT c intervals ( p = .80) failed to differentiate cases from controls. Conclusions TWH is capable of detecting latent repolarization abnormalities, which are present during ETT in diabetic patients with nonflow‐limiting stenosis but not in control subjects. The technique developed in this study permits TWH analysis from archived ECG s and thereby enables mining of extensive databases for retrospective studies and hypothesis testing.