Diabetes & silent MIs: Prevalent, prognostic or imprecise?

About one in six patients with newly diagnosed type 2 diabetes mellitus and no overt cardiovascular disease had evidence of silent MIs on electrocardiograms (ECGs), according to a study published March 5 in Circulation. The results don’t support routine use of screening ECGs in practice, though, wrote the author of an accompanying editorial.

Timothy M.E. Davis, MBBS, DPhil, of Fremantle Hospital in Fremantle, Australia, and colleagues reported on an analysis of the United Kingdom Prospective Diabetes Study (UKPDS), a trial that collected data on patients newly diagnosed with type 2 diabetes mellitus over a 30-year time frame. UKPDS had randomized patients to either an intensive glucose control or conventional treatment. Using UKPDS data on 5,102 patients, Davis et al set out to determine the prevalence of silent MIs, and the relationship between silent and future MIs and all-cause mortality.

The data included ECGs recorded every three years throughout the study. They defined a silent MI as definite or probable Q-wave MI (Minnesota codes 1.1 and 1.2) without a history of or typical symptoms of coronary heart disease. Excluding patients for missing ECG and clinical data and other factors, they identified 4,827 patients for the prevalence assessment.

The prevalence of silent MIs increased from 17.5 percent at baseline to 30.1 percent at 12 years, with the presence of Q waves appearing transient over time. For instance, 53 percent of patients with silent MI at baseline showed ECG evidence of silent MI at three years. Of those with complete baseline risk factor data (1,967 patients), patients with silent MIs were more likely to be older, female, sedentary, non-smokers, and be more likely to have higher blood pressure and take aspirin and lipid-lowering medications.

Davis et al determined that silent MI was significantly and independently associated with a 58 percent increased rate of fatal MI and a 31 percent increase rate of all-cause mortality. A separate analysis found that silent MI did not alter risk classification.

“Although SMI [silent MI] at diagnosis occurred more often in female patients, the SMI-associated increased rate of a subsequent fatal MI was independent of sex,” they wrote. “Some modifiable cardiovascular risk factors, including hypertension and microalbuminuria, were more prevalent among patients with SMI. Nevertheless, when SMI was added to UKPDS Risk Engine variables in multivariate analyses, the improvement in prediction of key cardiovascular outcomes was marginal.”

Based on their results, they suggested that Q waves on ECG’s of otherwise asymptomatic patients may hold prognostic significance and would warrant a review of modifiable risk factors in these patients.

In an accompanying editorial, Benjamin M. Scirica, MD, MPH, of Brigham and Women’s Hospital in Boston, admired the extended time frame used in the analysis but detailed several areas of concern. The prevalence findings were “surprisingly high” compared with other diabetes studies, prompting him to question whether the ECG criteria were “not specific enough to exclude non-infarct-related ECG abnormalities.”

The fact that the risk analysis offered only marginal improvement in predicting outcomes was in line with guidelines recommending against ECG screening for primary prevention, Scirica wrote. Nor did it appear that test results would give physicians reason to change therapy. Patients in UKPDS whose baseline ECG showed evidence of a Q wave and who were randomized to the more intensive glucose group appeared to gain no additional benefit from the regimen, he observed. The transience of Q waves, as shown in the three-year findings, suggest most transient Q waves were false positives, he added.

“For clinical care, the evidence and guidelines do not support the routine use of screening ECGs to detect new Q waves in patients with diabetes mellitus,” Scirica advised, while aggressive control of cardiovascular risk factors should remain a priority.