Q# 19-05: Hydrochlorothiazide and risk of non-melanoma skin cancer
In the adult patient population, is hydrochlorothiazide therapy causally associated with the risk of non melanoma skin cancer (i.e. squamous and basal cell carcinomas)?
If so, is there a cumulative dose-response relationship?
Non-melanoma skin cancer (NMSC) represents the most commonly diagnosed malignancy in Canada1 . The two most frequent histologic variants of NMSC are cutaneous squamous cell (SCC) and basal cell carcinomas (BCC). One in 8 (12.5%) Canadian will develop BCC in their lifetime and 1 in 20 (5%) will develop SCC2 . Ultraviolet exposure, photosensitizing drug use, and immunodeficiency are important risk factors for NMSC, with light skinned individuals most at risk1 .
Hydrochlorothiazide containing products are widely prescribed in Canada for a variety of indications, most commonly hypertension and edematous conditions. Hydrochlorothiazide is known to increase the sensitivity of the skin to sunlight and ultraviolet radiation3 • The current product monographs of hydrochlorothiazide containing products are not labelled for the risk of skin cancer or NMSC; photosensitivity, however, is labelled as an adverse reaction. In the US, labelling is similar for hydrochlorothiazide containing products.
Prompted by recent pharmacoepidemiologic studies, Health Canada recently completed a signal assessment for hydrochlorothiazide and the risk of non-melanoma skin cancer . The signal assessment leveraged EMA’s systematic review of the relevant evidence and updated the searches to identify subsequently published studies. An up-to-date systematic review of the literature and meta-analyses were performed and the certainty of evidence was rated with the GRADE approach. Total 5 pharmacoepidemiologic studies (from Europe and the US) including thousands of patients were directly relevant to the topic. Very low certainty evidence (for causal inference) suggested that the risk of NMSC increases with increasing use of hydrochlorothiazide (compared with its non-use); after several years of use, the risk could be four times higher for SCC and 1.25 folds higher for BCC. The certainty of evidence was downgraded for important methodological limitations identified in the included evidence. For example, sun exposure data and smoking status was not available or substantially missing; underlying disease, its duration and follow-up time were not matched; unclear disposition of exposure classification during the lag period in some studies; unclear adjustment for death as a competing risk for non melanoma skin cancer in other studies, etc.
Because of the existing uncertainty about an important patient-oriented outcome, Health Canada seeks researchers’ input and feasibility regarding the conduct of an observational study based on existing Canadian/foreign administrative digital healthcare data that could answer the aforementioned research questions with a higher degree of certainty by minimizing important biases identified in the existing literature on the topic in particular, and otherwise known to impact administrative database studies in general. Ideally, the observational study protocol would be designed emulating a hypothetical randomized trial of adult patients starting and adhering to hydrochlorothiazide therapy in particular and photosensitizing thiazide class of drugs in general versus alternative patient management strategies (e.g . use of non-photosensitizing, non-thiazide class of alternative pharmacological therapy; non-pharmacological therapies; or both, etc.). Subgroups of interest and important effect modifiers would be pre-specified.
CAN-AIM proposes to use data from three cohorts of the Canadian Partnership for Tomorrow Project (CPTP), all linked with administrative data collected by provincial Ministries of Health. The following cohorts are involved:
- The CARTaGENE cohort
- The BC Generations Project (BCGP)
- The Ontario Health Study (OHS)
The PMAP is finalized and approved for the current project. We are currently working on the analytical protocol and initial sample selection from the first CPTP cohort, the Ontario Health Study (OHS). We have also granted access to more recent years of data from the BC Generation Project (up to December 2019) and we are in process to apply for the same update with Cartagene cohort