Hydrochlorothiazide and Risk of Skin Cancer

The study objective is to determine whether the use of hydrochlorothiazide is associated with an increased risk of skin cancer compared with the use of angiotensin-converting enzyme (ACE) inhibitors. More specifically, the investigators will assess whether hydrochlorothiazide is associated with the risk of non-melanoma and melanoma skin cancer. A common-protocol approach will be used to conduct retrospective cohort studies using administrative health data from seven Canadian provinces (Alberta, British Columbia, Manitoba, Nova Scotia, Ontario, Quebec, and Saskatchewan) and the United States (US) MarketScan database. Briefly, the Canadian databases include population-level data on physician billing, diagnoses and procedures from hospital discharge abstracts, and dispensations for prescription drugs. Prescription drug data are limited to those aged 65 years old and older in Alberta, Nova Scotia, and Ontario. The US MarketScan database includes individuals and their dependents covered by large US employer health insurance plans, and government and public organizations. A standardized mortality ratio weighted cohort analysis will be conducted. In each jurisdiction, the investigators will assemble a study cohort that includes all patients aged 40 years or older (or 66 years or older in Alberta, Nova Scotia, and Ontario) newly-treated with hydrochlorothiazide or an ACE inhibitor between April 1, 1995 and March 31, 2018 (or the latest date of data availability at each site). The date of study cohort entry will be defined by the dispensation date of the newly prescribed hydrochlorothiazide or ACE inhibitor. Patients will be followed starting 366 days after study cohort entry until an incident diagnosis of non-melanoma or melanoma, or censored upon switching to a study drug, death, end of coverage or end of the study period (March 31, 2018), whichever occurs first. Exposure will be defined as a prescription for hydrochlorothiazide or an ACE inhibitor on the date of cohort entry. The exposures will be lagged by one year in order to consider a minimum cancer latency period between treatment initiation and the diagnosis of skin cancer. Analyses will be conducted using a modified intention-to-treat approach. The outcomes of interest will be non-melanoma and melanoma skin cancer. Weighted Cox proportional hazards models with calendar year as a strata will be used to estimate site-specific adjusted hazard ratios (HR) and corresponding 95% confidence intervals (CI) for each outcome of interest among hydrochlorothiazide users compared to ACE inhibitor users. In order to reflect the Canadian context, the primary analysis will be restricted to Canadian data and the US MarketScan data will be used only in a secondary analysis. Secondary analyses will be conducted for each outcome of interest. These will include the following: 1) as-treated, 2) cumulative duration of use, 3) dose-response relation, 4) effect modification by age (≤65, 66-74, and ≥75 years), sex and immunosuppressive status, and 5) time since initiation of treatment (\<3, 3-5, and \>5 years). For non-melanoma, events will be further classified into basal cell carcinoma or squamous cell carcinoma sub-types in jurisdictions where data is available. Four sensitivity analyses will be performed to assess the robustness of study results and address some of the study limitations. Site-specific results from the Canadian sites will be combined by random-effects meta-analysis to provide an overall assessment of the risk of non-melanoma and melanoma skin cancer in hydrochlorothiazide users compared to ACE inhibitor users. Four additional analyses at the meta-analysis level will be conducted. These will include the following: 1) using fixed-effects model, 2) pooling results only from jurisdictions where non-melanoma sub-type differentiation is available, 3) pooling results from all sites including the US MarketScan data, and 4) pooling results by jurisdictions with and without cancer registry data.