We read with interest the narrative Review by Khosla et al. (Khosla, S., Samakkarnthai, P., Monroe, D. G. & Farr, J. N. Update on the pathogenesis and treatment of skeletal fragility in type 2 diabetes mellitus. Nat. Rev. Endocrinol. 17, 685–697 (2021))1. The authors have clearly elucidated the aetiopathogenesis of diabetic bone disease; nevertheless, certain aspects of the disease deserve further thought, especially when it comes to diagnosis and management of bone fragility in type 2 diabetes mellitus (T2DM).

The fact that bone mineral density (BMD) is higher in people with T2DM than in age-matched and sex-matched healthy individuals is not universally applicable. Most of the data with regard to BMD in people with T2DM is based on the white population. Similar data from Asia, where the majority of the global population of people with T2DM reside, are scarce. As an example, available data from India suggests that BMD is highly variable in people with T2DM; in fact, BMD in people with T2DM was observed to be lower than in healthy individuals in some studies2,3.

It is widely accepted that a correction factor of 0.5 needs to be applied while considering the BMD T-scores in people with T2DM, mainly while calculating the ten-year fracture probabilities using the Fracture Risk Assessment Tool (FRAX). However, this evidence is based on a 2011 study that had included participants from only three observational studies, namely: Study of Osteoporotic Fractures, Osteoporotic Fractures in Men and Health, and Ageing and Body Composition Study. Here again, the participants were older community-dwelling adults residing in the USA who were white or Black4, which limits the generalizability of the evidence. It therefore seems inappropriate to use a correction factor of 0.5 while considering the T-scores in Asian Indian people with T2DM.

Furthermore, two additional corrections that have been proposed for calculating ten-year fracture probability scores in people with T2DM, apart from correcting the T-score, are to increase the patient’s age by ten years or to select the option ‘rheumatoid arthritis’ in the FRAX calculation tool5. Nevertheless, each of these three corrections, when applied, yield numerically different ten-year fracture probabilities for hip fracture and major osteoporotic fractures. Hitherto, there are no studies comparing one correction modality over the other and so it remains uncertain which modality should be preferred while calculating FRAX in T2DM.

With regard to the choice of anti-osteoporotic drugs in diabetic bone disease, the authors1 have rightly said that the available data are scant and limited either to observational studies or post hoc analysis of randomized controlled trials. Head-to-head clinical trials comparing anti-resorptive therapies with anabolic therapies are lacking, especially in the setting of low bone turnover observed in people with diabetes mellitus6. Besides, the effectiveness of various combination therapy regimens versus sequential therapy regimens have not been tested in people with diabetes mellitus.

Thus, the diagnosis and management of diabetic bone disease remain enigmatic. Extensive collaborative research involving participants across various ethnicities is required in this regard. Randomized controlled trials that compare the efficacy of different anti-osteoporotic agents with pre-specified endpoints primarily in people with diabetes mellitus need to be undertaken. One such trial is underway in India (CTRI/2022/02/039978).