Background Discharge β-blocker prescription after myocardial infarction (MI) is recommended for all eligible patients. Numerous β-blocker choices are presently available with variable glycometabolic effects, which could be an important consideration in patients with diabetes mellitus (DM). Whether patients with DM preferentially receive β-blockers with favorable metabolic effects after MI and if this choice is associated with better glycemic control postdischarge is unknown. Methods Among patients from 24 US hospitals enrolled in an MI registry (2005-2008), we investigated the frequency of "DM-friendly" β-blocker prescription at discharge by DM status. β-Blockers were classified as DM-friendly (eg, carvedilol and labetalol) or non-DM-friendly (eg, metoprolol and atenolol), based on their effects on glycemic control. Hierarchical, multivariable logistic regression examined the association of DM with DM-friendly β-blocker use. Among DM patients, we examined the association of DM-friendly β-blockers with worsened glycemic control at 6 months after MI. Results Of 4,031 MI patients, 1,382 (34%) had DM. β-Blockers were prescribed at discharge in 93% of patients. Diabetes mellitus-friendly β-blocker use was low regardless of DM status, although patients with DM were more likely to be discharged on a DM-friendly β-blocker compared with patients without DM (13.5% vs 10.3%, P =.003), an association that remained after multivariable adjustment (odds ratio 1.41, 95% CI 1.13-1.77). There was a trend toward a lower risk of worsened glucose control at 6 months in DM patients prescribed DM-friendly versus non-DM-friendly β-blockers (Relative Risk 0.80, 95% CI 0.60-1.08). Conclusion Most DM patients were prescribed non-DM-friendly β-blockers - a practice that was associated with a trend toward worse glycemic control postdischarge. Although in need of further confirmation in larger studies, our findings highlight an opportunity to improve current practices of β-blockers use in patients with DM.