Smoking is associated with incident heart failure (HF), yet limited data are available exploring the association between smoking status and long-term outcomes in HF with reduced vs. preserved ejection fraction (i.e., HFrEF vs. HFpEF).

We performed a retrospective analysis of HF patients undergoing coronary angiography from 1990–2010. Patients with coronary artery disease (CAD) and HF were stratified by EF (< 50% vs. ≥50%), smoking status (prior/current vs. never smoker), and level of smoking (light/moderate vs. heavy). Time-from-catheterization-to-event was examined using Cox proportional hazard modeling for all-cause mortality (ACM), ACM/myocardial infarction/stroke (MACE), and ACM/HF hospitalization with testing for interaction by HF-type (HFrEF vs. HFpEF).

Of 14,406 patients with CAD and HF, 85% (n = 12,326) had HFrEF and 15% (n = 2080) had HFpEF. At catheterization, 61% of HFrEF and 57% of HFpEF patients had a smoking history. After adjustment, there was a significant interaction between HF-type and the association between smoking status and MACE (interaction P = .009). Smoking history was associated with increased risk for MACE in patients with HFrEF (adjusted hazard ratio [HR] 1.18 [1.12–1.24]), but not HFpEF (HR 1.01 [0.90–1.12]). Active smokers had increased mortality following adjustment compared to former smokers regardless of HF-type (HFrEF HR 1.19 [1.06–1.32], HFpEF HR 1.30 [1.02–1.64], interaction P = .50). Heavy smokers trended towards increased risk of adverse outcomes versus light/moderate smokers; these findings were consistent across HF-type (interaction P > .12).

Smoking history was independently associated with worse outcomes in HFrEF but not HFpEF. Regardless of HF-type, current smokers had higher risk than former smokers.