Within-class differences in cancer risk for sulfonylurea treatments in patients with type 2 diabetes (ZODIAC-55) – a study protocol

Patients diagnosed with type 2 diabetes (T2D) are at increased risk for developing cancer; especially the risk of obesity-related cancers [ 1– 5]. According to the most recent World Cancer Research Fund (WRCF) definitions obesity related cancers include oesophageal cancer, liver cancer, kidney cancer, stomach cardia cancer, colorectal cancer, advanced prostate cancer, post-menopausal breast cancer, gallbladder cancer, pancreatic cancer, ovarian cancer and endometrial cancer [ 6].

In addition to T2D and obesity, glucose-lowering agents used in the treatment of T2D have also been associated with cancer risk and some studies have reported that these relations can be drug specific. For example, the use of pioglitazone, not rosiglitazone, has been linked to the development of bladder cancer in some studies [ 7, 8], although the robustness of the evidence underlying this possible relationship remains unclear and was absent in recent reports [ 9, 10]. Also, insulin glargine has been linked to higher breast cancer risk in some studies [ 11], although - again - several studies reported no or even an inverse association [ 11– 13]. Metformin has more consistently been associated with a decreased cancer risk [ 14], however concerns have been raised that this association might have been influenced by several types of bias [ 15, 16].

The sulfonylureas (SUs) are one out of six classes of oral glucose-lowering agents advised by the EASD and ADA as a second step when the glycaemic treatment targets are not reached with metformin mono-therapy [ 17]. Sulfonylureas have been available for many years and are highly efficacious at low costs. In the Dutch primary care treatment guideline for T2D, gliclazide is the preferred SU, as opposed to both the ADA and EASD which do not recommend a specific SU [ 17]. Previous studies have shown that within the class of SUs differences exist with regard to hypoglycaemia risk [ 18], for example there have been no reports of severe hypoglycaemia events in gliclazide users [ 18]. In addition, within-class differences in risk of cardiovascular events and safety when prescribed to patients with renal failure have been reported [ 19, 20].

An association between the class of SUs and increased overall cancer risk has also been reported [ 21– 24]. Most previous studies are limited by methodological issues, for example many studies reported baseline SU use and did not account for duration of SU use [ 22– 24]. There is also evidence suggesting within-class SU differences in cancer risk, where gliclazide use has been associated with a lower cancer risk [ 21– 24]. There are several potential mechanistic explanations, one of which could be that gliclazide leads to a more selective glucose dependent insulin response and lower insulin levels. In what way and to what degree cancer risk is modified by different SUs in unclear and requires further investigation and confirmation.

Most evidence, however, is derived from small observational cohort studies and substantial knowledge gaps exist. This also holds true for the presumed favourable long-term cancer safety profile of gliclazide in particular. The relations between use of glucose lowering agents and cancer are complex and there is overlap in risk factors; for example, several glucose lowering agents have been associated with weight gain, which in itself has also been related to an increased cancer risk.

The primary aim of this study is to evaluate within-class SU differences in risk for obesity-related cancer (excluding non-melanoma skin cancers) accounting for weight changes during follow-up and drug exposure [ 6]. Secondary aims are to evaluate within-class SU differences concerning all cancers combined and the cancer risks of the five largest groups of site specific cancers (breast, colorectal, prostate, bladder and lung cancer) accounting for duration of drug use.

This study will be the first large observational cohort study investigating differences in cancer risk within the class of SUs. An estimated 8% of the global population is known with T2D [ 29] and this could translate into an increased risk of cancer for an substantial amount of people. The prevalence of T2DM is expected to rise evermore, at least for the next decades [ 30]. A minimal change in cancer risk could result in a substantial change in the relative and even absolute number of patients diagnosed with cancer. If this study confirms the presence of within-class SU cancer risk differences, it could help patients and physicians in making a shared decision for a specific SU. This could contribute to quality of life of the patients as well as contribute to increasing effective care and cost-effectiveness of healthcare. If no differences are present, the safety, efficacy, and cost of SU will remain the only criteria for selecting the best SU.