Background: Patients with heart failure (HF) are at higher risk of developing cancer. A recent observational study suggested statins could reduce cancer diagnosis in patients with HF.
Objectives: To evaluate the effect of statin treatment on the risk of cancer in patients with HF.
Methods: We conducted a target trial emulation using observational data from IQVIA Medical Research Database-UK (IMRD-UK, formerly known as THIN database) from 2000 to 2019. We included patients with a first HF diagnosis during the study period, and excluded patients who had a history of cancer, liver failure, HIV infection, and prior use of statin within 180 days of HF diagnosis. We emulated two hypothetical pragmatic trials: (1) initiating versus not initiating a statin within one year after HF diagnosis, (2) continuing a statin for ≤3 years, >3- to 6 years, and >6 years after initiation. The study outcome was a cancer diagnosis. In both emulated trials, we used a clone-censor-weight approach. All patients were replicated and each set of clones was assigned to each treatment strategy. The clones were artificially censored if they deviated from the assigned treatment. Finally, inverse probability of censoring weightings (IPCWs) were estimated following Robin’s G-methods to account for selection bias from artificial censoring. Weighted pooled logistic regression models were used to estimate the absolute risk differences, risk ratios (RR), and hazard ratios (HRs) of statin treatment on cancer risks. 95% confidence intervals (CIs) were estimated using non-parametric bootstrapping with 200 full resamples.
Results: 75,252 patients with HF were included in the first emulated trial. Compared to no statin treatment, statin treatment did not reduce the cancer risk in patients with HF (RR, 1.05; 95%CI, 0.94-1.15). 15,525 patients who initiated statin within 12 months of the HF diagnosis were included in the second emulated trial. Compared to treatment ≤3 years, statin treatment with longer durations did not reduce the cancer risk (3-6 years: RR, 0.94; 95%CI, 0.70-1.33. >6 years: RR, 0.97; 95%CI, 0.79-1.26). No significant risk difference was observed on any site-specific cancer diagnoses.
Conclusions: When using a target trial emulation framework to avoid immortal-time bias and selection bias in observational study design, we found statin treatment did not reduce the cancer risk in patients with HF.
