Lack of Association Between HF and Cancer

Study Questions:

Is heart failure (HF) associated with cancer incidence and cancer-specific mortality?


The study cohort comprised participants from the Physicians’ Health Studies I and II, two randomized controlled trials of aspirin and vitamin supplements conducted from 1982 to 1995 and from 1997 to 2011, respectively, which included annual health evaluations and determination of cancer and HF diagnoses. In the primary analysis, the study authors modeled HF as a time-varying exposure. Secondarily, they measured cumulative cancer starting at landmark ages (they performed sensitivity analyses at 65, 75, and 80 years of age). In these ways, they were able to mitigate confounding by age, which is the single strongest risk factor for incident cancer.


Among 28,341 Physicians’ Health Study participants, 1,420 developed HF. A total of 7,363 cancers developed during a median follow-up time of 19.9 years (25th-75th percentile, 11.0-26.8 years). In the 177 HF participants who subsequently developed cancer, a median time of 3.4 years (25th-75th percentile, 1.5-6.9 years) elapsed between HF diagnosis and cancer diagnosis. HF was not associated with cancer incidence in crude (hazard ratio [HR], 0.92; 95% confidence interval [CI], 0.80-1.08) or multivariable-adjusted analysis (HR, 1.05; 95% CI, 0.86-1.29). The risk for cancer, adjusted for covariates, was similar among key subgroups (age at baseline >60 years of age or ≥60 years of age, aspirin use, and smoking status) (p value for interaction >0.05 for all comparisons). HF did not increase the risk of cancer death on unadjusted analysis (HR, 0.89; 95% CI, 0.71 to –1.12; p = 0.32) or after adjusting for covariates in model 1 (HR, 1.16; 95% CI, 0.82-1.65; p = 0.40). No association was found between HF and site-specific cancer incidence (for example, prostate cancer, HR, 0.78; 95% CI, 0.59-1.05; p = 0.10), or cancer-specific mortality after multivariable adjustment. Results were similar when using the landmark method at all landmark ages. In their base case analysis at 70 years of age, 10.5% of HF and 9.3% of non-HF participants developed cancer during the initial 5 years of follow-up. The unadjusted model showed no association between HF and time to first cancer diagnosis (HR, 1.05; 95% CI, 0.79-1.40; p = 0.71). Multivariable adjustment did not alter the findings (HR, 1.04; 95% CI, 0.72-1.51; p = 0.84). HF did not increase the risk of cancer death on unadjusted (HR, 1.28; 95% CI, 0.78-2.09; p = 0.33) or multivariable-adjusted (HR, 1.44; 95% CI, 0.68-3.06; p = 0.34) analysis.


The study authors confirmed that HF is not associated with an increased risk of cancer among male physicians.


These findings are important because this study concludes that there was no association between HF or current guideline-mediated therapy for HF and incident cancer. However, as more and more novel therapies emerge, it is possible that surveillance for cancer in HF may need to be done.

Clinical Topics: Cardio-Oncology, Geriatric Cardiology, Heart Failure and Cardiomyopathies, Prevention, Acute Heart Failure, Smoking

Keywords: Aspirin, Cardiotoxicity, Geriatrics, Heart Failure, Incidence, Neoplasms, Prostatic Neoplasms, Risk Factors, Secondary Prevention, Smoking, Vitamins

< Back to Listings