BMI trajectories tied to diabetes incidence

Longitudinal trajectories in body mass index (BMI) may be predictive of new-onset diabetes, a recent study has found.

In 4,519 adults (median age, 42.0 years; 53.9 percent) free of diabetes, researchers identified four linear BMI trajectories: low (n=884), moderate (n=1,511), moderate-high (n=1,510) and high (n=614). Trajectories were calculated using group-based trajectory modelling (GBTM), which assumes that differences in longitudinal patterns of change may be summarized by discrete sets of measurements. Thus, those with only one BMI measurement were excluded.

There were 168 new-onset cases of diabetes during the study duration, resulting in an incidence rate of 3.7 percent. This increased progressively and with direct proportionality to the BMI trajectories.

For instance, in the low-BMI trajectory group, 0.6 percent of the participants developed diabetes, while the same was true for 2.0 percent of those in the moderate-trajectory group. In the moderate-high and high groups, the incidence rates of diabetes were 5.1 percent and 9.1 percent, respectively.

Fully adjusted regression models confirmed these findings, showing that relative to those with low BMI trajectories, the high-trajectory group was at a greater risk of new-onset diabetes (relative risk, 3.24, 95 percent CI, 1.27–8.24).

Moreover, researchers found that even just four annual BMI tests were sufficient to reliably predict new-onset diabetes risk. Using this dataset, GBTM still yielded four linear BMI trajectories that correlated with diabetes incidence and risk. Similar results were observed when only the first three annual BMI tests were analysed.