SGLT2 inhibitors and severe infection: no excess against DPP-4 inhibitors in real-world cohorts

Design

Systematic review and meta-analysis of real-world active-comparator cohorts

Population

Adults with type 2 diabetes on SGLT2 vs DPP-4 inhibitors; 18 cohorts, 11 pooled

Primary outcome

Incident pneumonia and sepsis

Effect

Pneumonia ratio 0.65 (95% CI 0.55 to 0.75), I² 89.8%; sepsis prediction interval 0.38 to 1.13

This systematic review pooled observational cohorts comparing SGLT2 inhibitors with DPP-4 inhibitors in type 2 diabetes, with pneumonia and sepsis as co-primary outcomes. Eighteen cohorts were included and 11 contributed numbers; all were at moderate risk of bias.

For pneumonia, the pooled ratio was 0.65 (95% CI 0.55 to 0.75), with very high heterogeneity (I² about 90%); removing one cohort with delayed follow-up gave 0.61 with no heterogeneity. For sepsis the pooled ratio was also 0.65, but the prediction interval crossed 1 (0.38 to 1.13). GRADE certainty was very low for both. Comparisons with GLP-1 receptor agonists showed no consistent pattern.

The useful reading is reassurance, not protection. Prescribers worried that SGLT2 inhibitors raise severe infection risk — beyond the known genital mycotic infections — have no support for that here. The authors themselves caution against treating the lower estimates as a preventive effect, because healthier patients may be the ones started on SGLT2 inhibitors.

• Do not withhold an SGLT2 inhibitor out of concern for pneumonia or sepsis risk.
• Continue to counsel about genital mycotic infections, the infection risk that is established.
• Pause SGLT2 inhibitors during acute severe infection, as for any acute illness.
• Do not cite these data as a reason to start an SGLT2 inhibitor to prevent infection.

Why it matters

Infection worry is one reason some clinicians hesitate over SGLT2 inhibitors in frail patients.

Don't overread it

Observational and very low certainty; it does not show SGLT2 inhibitors prevent infection.

The statistics, in plain English

A pooled ratio of 0.65 suggests about a third fewer events, but I² around 90% means the studies disagreed sharply, and a prediction interval crossing 1 for sepsis means a new study could plausibly find no difference or harm. Very low certainty means the true effect may be substantially different. Patients chosen for SGLT2 inhibitors may simply be healthier, which would make them look protected.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.