Cancer or a history of cancer could plausibly be linked to risk of zoster (known commonly as shingles) because both diseases can be affected by the health of the immune system, and this may be further affected by cancer treatments. Recent US data suggest a substantial increase in zoster risk among people with a previous cancer diagnosis. A detailed investigation of this in CPRD would help establish how overall/site-specific cancers are related to zoster risk. We will identify people with a zoster diagnosis, and others with similar general characteristics but no zoster diagnosis, and we will look back in the GP medical records of both groups to identify previous cancers, and to examine whether previous cancers were more likely among those with a zoster diagnosis. We will also account for differences in other important characteristics between the groups. The analysis will help us to understand whether cancer or its treatment is linked to an increased future risk of zoster; the results will be of interest to patients with cancer or a history of cancer and their doctors. Since a zoster vaccine is now available, the results may also help to inform future vaccination policies for zoster.
The objective of the study is to investigate the association between having had a cancer diagnosis and future risk of herpes zoster (also known as shingles). The two diseases could plausibly be associated because both have associations with immune function, which may be further affected by cancer treatments. Recent US data suggested a substantial increase in zoster risk among people with a previous cancer diagnosis. Our aim is to investigate this further by looking at how "any cancer" and common site-specific cancers (breast, colorectal, lung, prostate) are associated with future zoster risk, adjusting for potential confounders. We will conduct a matched case-control study: all individuals with an incident zoster diagnosis will be identified and matched on age, sex and GP practice to people who are under follow-up in CPRD at the same time but have no history of zoster. Conditional logistic regression models will be used to examine associations between cancer and zoster, both unadjusted, and adjusted for a range of potential confounders. The temporal relationship between cancer and zoster will also be examined by categorising the time between cancer diagnosis and index date. The results will be of interest to people with a history of cancer, and may help to inform zoster vaccination policy.
Health Outcomes to be Measured:
All cancers and the following site-specific cancers: breast, colorectal, lung and prostate cancer.