In this prospective cohort study, which was conducted under the framework of an ongoing cohort study, 1228 prediabetic subjects were followed from 2003 to 2019. Changes in the BMI, HC, WC, WHR, and WHtR were evaluated over time using LMM. Three latent states were identified based on the patterns of changes in the mean values of anthropometric indices. The latent states were characterised according to the tendency for affecting by diabetes in future (low/ moderate/ high) and had latent state sizes of (29%/ 45%/ 26%).
For the best of our knowledge, this is the first attempt to classify prediabetic subjects into homogeneous states based on the changes in mean values of BMI, HC, WC, WHR, and WHtR over time using an advanced statistical model. However, there are previous studies that examined the general population, as well as some specific populations, by applying simple statistical methods (16, 19).
Previous studies have focused on the association between BMI, WC, HC, WHR, and WHtR with the risk of developing diabetes in the future separately. For instance, Sayeed et al. showed a significant association between WHtR and risk diabetes progression (11). In a meta-analysis based on the individual data of the Asian cohorts, Qiao et al. showed that in all studies included in this review, either BMI, WC, and WHR predicted or was associated with T2DM, separately (13).
It is believed that obesity indices abnormality are strong risk factor for T2DM (16, 20, 32). In the present study, the subjects with a high tendency for diabetes progression had obesity indices abnormalities. We observed that the mean of obesity indices was proportionally associated with low, moderate, and high tendency for diabetes progression. This finding is in line with the results of previous studies that have emphasized on the association of obesity disorders with the risk of diabetes progression (18, 19).
Previous studies have established that BMI is a predictor of T2DM in prediabetic subjects (18, 33, 34). In the pre sent study, the subjects with a high tendency for diabetes progression had a higher BMI mean value. Wei et al. obtained similar results to our study in terms of the association of BMI with future diabetes risk (34). In another study, Shakeri et al, found relationship between anthropometric indices and diabetes. They reported that the odds ratio of affecting by diabetes can elevate with increasing BMI (33). Furthermore, Haghighatdoost et al.'s study suggested that BMI is strongly associated with diabetes incidence and WC was moderately related to diabetes incident (20). Denmark et al.'s study indicated that who were overweight were more likely to develop diabetes (1.1% per unit of increase in the BMI). The study also revealed that risk of diabetes increases with weight gain and obesity (35).
WC and WHR are associated with progression of T2DM (8, 36). In the present study, the subjects with a high tendency of diabetes progression had a higher mean WC and WHR than other subjects. Klein et al. showed WC is a simple measure that can be used to identify subjects at increased risk of T2DM (36). In line with the present study, Sargeant et al. demonstrated WHR significantly associated with diabetes progression in future (8).
Earlier studies showed that WHtR is a predictor of T2DM in prediabetic subjects (32, 37). In the present study, the subjects with a high tendency for diabetes progression had a higher mean value of WHtR. In line with the present study, Tulloch-Reid et al, showed a strong association between BMI and WHtR abnormality with T2DM (37). In a longitudinal study, Hadaegh et al showed WHtR yielded the highest ability for future development of diabetes between other anthropometric measures (32).
In the present study, the subjects with a high tendency of diabetes progression had a higher mean value of obesity indices. This finding is in line with the results of Nayak et al, study that has emphasized the association WC, HC and BMI with the future risk of diabetes in subjects with PD (38).
Although numerous researchers have examined the risk factors of diabetes, most have ignored the complex, unstable, and variable conditions of prediabetic people in terms of obesity indices. Each year, 5–10% of people with PD will progress to diabetes, with the same proportion converting back to normoglycemia (39). In the current study, we followed and evaluated the trajectories of five important obesity indices simultaneously over a long-time horizon and identified people with a low, moderate, and high probability to progress diabetes or to remain in the same condition in the future. The results showed that the probability of transitioning from having a moderate tendency to develop diabetes to having a low tendency to develop diabetes was lower than the probability of transitioning in the opposite direction. Also, the probability of staying in the same state was higher than that of transitioning to a different state.