Hard to Resist: Evaluating the Contribution of Insulin Resistance to Bone Density and Skeletal Fragility

Epidemiologic studies in diverse populations have shown that diabetes, both type 1 (T1D) and type 2 (T2D), is associated with an increased risk of fracture (1, 2). The magnitude of the increased fracture risk is much greater for T1D vs T2D (3). Paradoxically, patients with T2D are at increased risk of fracture despite having bone density that is normal or even above average (4). This same paradox is present in T1D, albeit to a lesser extent, where the deficits in bone density are mild and insufficient to fully explain the susceptibility to fracture (5). Although the fracture risk associated with diabetes has now been well characterized, the pathophysiologic origins of skeletal fragility in these disorders are poorly understood and likely multifactorial (6, 7). This knowledge gap is apparent in clinical practice, where there is a lack of evidence to effectively guide clinicians in the monitoring and management of bone health in people with diabetes. The current guidelines are nonspecific and limited to recommendations to increase weight-bearing activity in childhood and to optimize vitamin D status (8). Although these steps are rational and unlikely to be harmful, they are almost certainly inadequate as they fail to address the specific effects of diabetes on the skeleton.

The respective similarities (increased fracture risk) and differences (reduced vs elevated bone density) between the skeletal phenotypes of T1D and T2D are intriguing and have helped fuel a new area of research focused on investigating the links between energy metabolism, body composition, and the skeleton. The output of this work has led some to propose the existence of a bone-pancreas feedback loop whereby the skeleton responds to insulin and in-turn regulates insulin secretion and sensitivity through the action of undercarboxylated osteocalcin (9). Preclinical data suggest that insulin exerts an anabolic effect on bone and is a critical regulator of skeletal development and structural integrity (10, 11). It has therefore been hypothesized by some that hyperinsulinemia is the major contributor to the high bone mass phenotype of T2D (12). Other data suggest that the skeleton develops insulin resistance (IR) similar to other body tissues (13), implying that the skeletal cellular environment is actually one of deficient insulin signaling and that other factors such as increased body mass must be responsible for the high bone density findings. The presence and/or degree of IR present in the diabetic skeleton remains an important unsolved question.

In this issue of The Journal of Clinical Endocrinology & Metabolism, Napoli et al. (14) report the results of a study investigating the associations between IR, areal bone mineral density (aBMD), and fracture risk in older adults from the Health, Aging, and Body Composition Study. A unique aspect of this study was the decision to investigate the associations between IR and skeletal outcomes in individuals who were not affected by diabetes. This approach is an attempt to address a major barrier to interpreting mechanistic studies in humans with diabetes—namely that it is difficult to separate the potentially independent effects of hyperglycemia, insulin exposure, and insulin signaling on the skeleton. A limitation to this approach is that it requires relationships between IR and skeletal outcomes to be investigated in a population that lacks an abnormal skeletal phenotype and therefore may not be generalizable to people with diabetes.

A primary finding of this study was that greater Homeostatic Model Assessment for Insulin Resistance (HOMA-IR) appeared to associate with greater aBMD on univariable analysis, but that these associations did not persist after accounting for the effects of body mass index (BMI). These findings can be interpreted as support for the hypothesis that greater bone loading due to higher body mass is a more important contributor than hyperinsulinemia to the increased bone density seen in T2D, although the strength and significance of the associations between BMI and aBMD were not reported. At the other end of the lifespan, two recent studies reported that HOMA-IR was significantly inversely associated with bone mass in healthy adolescent males, independent of body composition (15, 16). Taken together with the findings of the current study, these results raise the possibility that the effects of IR may differ on the growing vs aging skeleton and warrant further study.

The reported relationships between IR, BMI, and incident fracture risk in the current study are less clear. Greater HOMA-IR appeared to be protective against incident fracture on univariable analysis. When BMI and aBMD were then added to subsequent models, the direction of the association flipped such that greater HOMA-IR appeared to be a risk factor for fracture. The statistical significance of these findings was not consistent across quartiles of HOMA-IR, however. The absence of a statistically significant trend for fracture risk across HOMA-IR quartiles and the finding that fracture risk was significantly increased in the third, but not the fourth quartile of HOMA-IR are troubling, and suggest the possibility of an unmeasured confounder. This was discussed by the authors, who suggested vitamin D as one such possible confounder. The presence of advance glycation end products (AGEs) is another variable that should be considered. Urinary pentosidine (an AGE) has been previously shown by the authors to associate with fracture risk in Health, Aging, and Body Composition Study participants with diabetes (17) and by others to improve fracture risk prediction in older adults not affected by diabetes (18); relationships between IR and AGEs are not well described. The authors additionally surmise that the lack of a linear relationship between HOMA-IR and fracture suggests the possibility of a threshold effect—a valid hypothesis. Rather than simply comparing fracture risk above and below the median IR in a study population, future studies should be specifically designed to identify and test the threshold for IR that confers increased fracture risk.

In summary, Napoli et al. (14) have provided an important, but incremental contribution to our understanding of the relationship between IR and bone mass. Their findings in a population of older adults without diabetes suggest that positive relationships between IR and aBMD are explained by the greater BMI in these individuals. It is intriguing to speculate that the fracture data support a role for IR in skeletal fragility; however, the results were not sufficiently robust to be convincing. Ultimately, the development of new in vivo methodologies to assess downstream markers of insulin signaling in skeletal cells may be required to definitively ascertain the contribution of IR to the skeletal complications of diabetes.

Acknowledgments

Financial Support: Supported by the National Institute of Diabetes and Digestive and Kidney Diseases (K23 DK114477).

Disclosure Summary: The author has nothing to disclose.

Abbreviations:

Abbreviations:
 
• aBMD

  areal bone mineral density

 
• AGE

  advance glycation end product

 
• BMI

  body mass index

 
• HOMA-IR

  Homeostatic Model Assessment for Insulin Resistance

 
• IR

  insulin resistance

 
• T1D

  type 1 diabetes

 
• T2D

  type 2 diabetes

References and Notes