In the RECORD trial, 5292 participants aged ≥ 70 years with a recent previous osteoporotic fracture were randomised from 1999 to 2002 within a blinded, factorial design to oral 800 IU (20 μg) daily vitamin D3, 1000 mg calcium (calcium carbonate), both, or placebo, and followed up for 24–62 months. The trial was based in 21 centres in England and Scotland. Ethical approval was obtained from the Multicentre Research Ethics Committee for Scotland and each centre's Local Research Ethics Committee. The participants gave written informed consent. Full details and main results of the trial are reported elsewhere.
The main outcome of the trial was new low-energy fracture. The effects of vitamin D3 (with or without calcium) on the development of diabetes or initiation of tablets or injections for diabetes were prespecified secondary outcomes. At baseline and by annual postal or telephone questionnaire, the participants were asked if they had diabetes, or took tablets or injections for diabetes. All data on diabetes and medications for diabetes were therefore collected by self-report.
We compared all participants who had been randomised to take vitamin D3 with all those who had not been randomised to D3 (i.e. intention to treat analysis), using multiple logistic regression. Odds ratios and 95% confidence intervals were presented adjusted for the trial minimisation factors of gender, age, type of enrolling fracture and time since fracture. A corresponding, secondary, per protocol analysis was undertaken, based upon participants still reporting tablet consumption on > 80% of days at 2 years post-randomisation. We also undertook post hoc analyses to test for an interaction effect between vitamin D3 and calcium under a per protocol basis.