EFFECTS OF BLOOD PRESSURE-LOWERING ON CANCER RISK: AN INDIVIDUAL PARTICIPANT DATA META-ANALYSIS OF 300,000 PARTICIPANTS

Objective: Evidence for the effects of pharmaceutical blood pressure (BP)-lowering on cancer risk is inconsistent and based on observational data. We therefore investigated the effect of BP-lowering on the risk of cancer in a large collaborative study. Design and method: We included randomised trials participating in the Blood Pressure-Lowering Treatment Trialists’ Collaboration. Placebo-controlled trials, drug class comparison trials and trials comparing more-vs-less intensive BP-lowering that provided individual-level participant data (IPD) on cancer events were pooled. We investigated the effects of BP reduction on cancer risk by conducting one-stage IPD meta-analyses using Cox proportional hazard models, stratified by trial and accounting for competing risks. We also investigated effects stratified by age, gender, body mass index (BMI), smoking and previous antihypertensive use at baseline. Results: This analysis included 300,098 participants (42% women) from 39 trials. At baseline, the mean age of participants was 66 (standard deviation [SD] = 9), mean BMI was 28 (SD = 5), 18% were current smokers and 75% were previously on BP-lowering medication. Over a median duration of 4 years, 16,748 participants were diagnosed with cancer, and 4347 cancer deaths were reported. The hazard ratio (HR) per 5mmHg reduction in systolic BP was 1.03 (95% confidence interval [CI] 0.99–1.07) for any cancer and 1.05 (95% CI 0.98–1.12) for cancer death. We found heterogeneity in the effects of BP-lowering across age, gender, BMI and smoking groups for any cancer and cancer death, and across groups defined by previous antihypertensive use for any cancer. However, there was no evidence that BP-lowering significantly increased the risk of developing cancer in specific patient subgroups. We found no evidence for trends in increasing or decreasing risk over time for either outcome (P for trend: any cancer = 0.98, cancer death = 0.99). Conclusions: This large-scale IPD meta-analysis found no evidence that BP-lowering had an important effect on cancer risk. Although we found that BP-lowering effects differed across several patient characteristics, there was no evidence for trend in cancer risk over time. We plan to further investigate the effects of BP-lowering on site-specific cancers (breast, colorectal, kidney, lung, prostate and skin) and present these results at the meeting.