Letter to the Editor: On epidemiology of fractures and variation with age and ethnicity.

We recently demonstrated differences in fracture incidence by ethnicity in the UK population using the Clinical Practice Research Datalink (CPRD) [1]. In their letter, Harper et al. [2] draw attention to the differential age structure of the UK population by ethnicity, and suggest that the lower incidence of fractures in Asian adults over 50 years might reflect the lower proportion of very elderly individuals within this demographic compared with the white population. The relatively small number of fractures within black and South Asian ethnicities meant that presentation of fracture incidence stratified by age band was not feasible. We have thus used a Poisson model to calculate incidence rate ratios (with white ethnicity as the referent), adjusting for age and calendar year to account for any contribution of age distribution and secular change within ethnicity strata. These data are summarised in Table 1 below, and clearly demonstrate that the differences in fracture incidence by ethnicity are congruent with the raw incidence data presented in our paper, with rates highest in whites, intermediate in South Asians and lowest in blacks. In conclusion, whilst there are ethnic