Type 2 diabetes mellitus (T2DM) has been associated with an increased risk of fractures. Furthermore, SGLT2-I use (a drug used to treat T2DM) is also hypothesized to contribute to this increased risk of fractures in patients with diabetes. Therefore, the aim of this study is to analyse whether there is a relationship between SGLT2-I use and fracture risk. More specifically, we will study whether SGLT2-I use affects fracture risk in subjects suffering from T2DM when compared to DPP4-I use. DPP4-I was chosen as a comparator, because previous observational and randomized controlled studies have shown that this drug does not affect fracture risk in this study population. Furthermore, DPP4-I is likely to be prescribed to a similar population.
To perform this research, we will make use of the British CPRD, an ongoing primary care database of medical records from general practitioners, with coverage of over 11.3 million patients. From this database, we will select all first-time SGLT2-I users over 18 years old and find a comparable first-time DPP4-I user. In this way, two comparable groups of both SGLT2-I and DPP4-I users will be formed, in which we can analyse and compare the fracture risk.
SGLT2-I are a relatively new drug class to treat T2DM. Data on more long-term putative side effects have recently been brought forward, such as an increased risk of fractures (2). Although studies have been performed on this matter, results are conflicting and hampared by a short duration of follow-up. Therefore, our aim is to study the association between longer term use of SGLT2-Is as compared to use of DPP4-I (both as add-on to metformin) and the risk of any fracture and major osteoporotic fractures (MOFs).
We will select all patients aged 18 and older with at least one year of valid data collection and with a first ever prescription of a DPP4-I or a SGLT2-I between 1-1-2013 and 31-12-2020. Propensity score matching will be used to control for confounding. Cox regression models will be used to estimate the risk of any fracture with SGLT2-I as compared to DPP4-I use in CPRD GOLD, and risk of any fracture as well as MOFs in AURUM. As secondary analyses we will estimate the risk of MOFs and a fracture of the hip, radius/ulna, proximal humerus, and clinical symptomatic vertebral fracture individually. Furthermore, we will stratify the analyses by sex and age categories. Duration of use will be studied, for which we will divide follow-up time into intervals of 90 days and determine duration of use at the start of each interval. SGLT2-I use will then be stratified by categories of duration of use. Lastly, we will study the class effect of SGLT2-Is, comparing the different classes of SGLT2-I initiators (Canagliflozin, Dapagliflozin, Empagliflozin and Ertugliflozin) to DPP4-I initiators.
Primary outcome:
Major osteoporotic fractures (defined as a fracture of hip, humerus, radius/ulna or clinical symptomatic vertebrae) as well as any fracture in AURUM.
We expect the number of events for MOFs to be relatively low in CPRD GOLD, and therefore our primary outcome in CPRD GOLD will be:
Any fracture
Secondary outcomes:
MOFs and individual MOF sites
Patrick Souverein - Chief Investigator - Utrecht University
Patrick Souverein - Corresponding Applicant - Utrecht University
Coen Stehouwer - Collaborator - Maastricht University
Jakob Starup-Linde - Collaborator - Aarhus University Hospital
Johanna Driessen - Collaborator - Utrecht University
Joop van den Bergh - Collaborator - Maastricht University
Olaf Klungel - Collaborator - Utrecht University
Peter Vestergaard - Collaborator - Aalborg University Hospital
Rikke Viggers - Collaborator - Aalborg University Hospital
Veerle van Hulten - Collaborator - Utrecht University
Zheer Kejlberg Al-Mashhadi - Collaborator - Aarhus University Hospital