Obesity and Skeletal Fragility

Abstract

Skeletal fracture has recently emerged as a complication of obesity. Given the normal or better than normal bone mineral density (BMD), the skeletal fragility of these patients appears to be a problem of bone quality rather than quantity. Type 2 diabetes mellitus (T2DM), the incidence of which increases with increasing body mass index, is also associated with an increased risk for fractures despite a normal or high BMD. With the additional bone pathology from diabetes itself, patients with both obesity and T2DM could have a worse skeletal profile. Clinically, however, there are no available methods for identifying those who are at higher risk for fractures or preventing fractures in this subgroup of patients. Weight loss, which is the cornerstone in the management of obesity (with or without T2DM), is also associated with an increased risk of bone loss. This review of the literature will focus on the skeletal manifestations associated with obesity, its interrelationship with the bone defects associated with T2DM, and the available approach to the bone health of patients suffering from obesity.

Traditionally, obesity is considered osteoprotective because of the positive association between body mass index (BMI) and bone mineral density (BMD) (ie, individuals with higher BMI have a higher BMD) (1). In recent decades, however, the concept that obese individuals are protected from fractures has been challenged by epidemiologic studies showing an increase in the risk of fractures, especially nonspine fractures, in obese individuals despite a normal or above normal BMD (2-7), an observation that has been coined by some in the scientific community as the “obesity paradox” phenomenon (8, 9). In this review, we synthesize the published data on the topic of obesity and skeletal fragility by a literature search of the relevant manuscripts in PUBMED, EMBASE, and Google Scholar until February 2023.

Bone Metabolism in Obesity

The higher BMD in obese patients is postulated to be due to skeletal adaptation to accommodate the higher load from excess weight (10-12), and from hyperestrogenemia secondary to the elevated aromatase enzyme activity in the abundant volume of adipose tissue (13, 14). However, this relationship is complex because, despite the positive correlation between BMI and BMD, a negative correlation between adiposity and BMD has been reported by several investigators (15-17). Although mechanical and hormonal factors generally favor bone anabolism (11, 12), the increased adipose tissue volume in obesity is associated with the production of a host of substances that may be harmful to bone. Furthermore, because vitamin D is sequestered in fat tissues, a low 25-hydroxyvitamin D level, which is common in the obese population, results in reduced calcium absorption from the gut (18, 19). This may lead to increased levels of PTH, which could have a negative impact on the skeletal health of obese subjects (20, 21) and may predispose them to bone loss (22, 23). On the other hand, aside from studies in overweight and obese children suggesting an increase in fracture risk among those with low vitamin D (24), fracture studies in obese adults with low vitamin D are lacking. The Longitudinal Aging Study Amsterdam, a study of community-dwelling older men and women with an average BMI in the overweight range, showed that a 25-hydroxyvitamin D level of ≤12 ng/mL was associated with an increased fracture risk among those who are aged 65 to 75 years but not among those who are aged >75 years (25). Although these studies support low vitamin D as a risk factor for fracture, further studies in obese older adults are needed.

Leptin and adiponectin are cytokines produced in the adipocytes, and obesity in the presence or absence of type 2 diabetes mellitus (T2DM) is associated with high levels of leptin and reduced levels of adiponectin (26). Because their effect on bone metabolism is variable, their overall impact on the skeleton remains unclear. For instance, leptin administered centrally resulted in decreased bone formation through stimulation of the sympathetic nervous system in both leptin-deficient and leptin-sufficient mice (27, 28). However, leptin administered peripherally by subcutaneous injection of recombinant leptin to children with congenital leptin deficiency led to increased bone mass and a reduction in body weight and fat (29, 30). Adiponectin has also been reported to have dual effects on bone. Although it has been shown to directly stimulate osteoblastogenesis, it has also been shown to promote osteoclastogenesis by stimulating the osteoclast receptor activator nuclear factor kappa-B ligand pathway (31). Others report that adiponectin can increase bone mass by suppressing osteoclastogenesis and at the same time activating osteoblastogenesis (32, 33). Overall, the majority of studies in humans showed a negative correlation between adiponectin and BMD (34-36). More importantly, results from the Osteoporotic Fractures in Men Study showed that increasing serum adiponectin is associated with an increase in the risk for fractures (37).

Adiposity in Obesity

Aguirre et al previously reported that in a population of obese older adults, BMI was positively correlated with BMD, whereas the opposite was true for body fat, which negatively correlated with BMD (15). In this study, fat mass remains an independent predictor of BMD in the different skeletal sites examined in both genders. To what extent does body fat influence BMD in relation to other compartments is demonstrated by the results of a meta-analysis of 44 studies investigating the effect of body composition on BMD. Data from a total of 20 226 subjects (15 260 women and 4966 men) showed that the correlations between lean mass with the spine, femoral neck, and total body BMD were much higher than that of fat mass in the entire population (38). Subgroup analysis also showed the same greater influence of lean mass over fat mass on BMD of the spine, femoral neck, and total body in men and premenopausal women. However, among postmenopausal women, the influence of lean and fat mass on the spine and femoral neck BMD was comparable. Moreover, the influence of body composition on fracture risk was examined in 49 050 women and 3600 men living in Manitoba, Canada (39). In this study, increasing lean mass was associated with increasing BMD but had no effect on the femoral strength index. Meanwhile, increasing fat mass had no influence on BMD but had a negative effect on femoral strength index. Nevertheless, higher fat mass was not an independent predictor of 5-year fracture risk, whereas the calculated Fracture Risk Assessment (FRAX) score remained the best predictor.

Although obese subjects generally have a greater amount of total adipose tissue volume than leaner individuals, it is now widely recognized that the metabolic dysfunction associated with obesity is more from the distribution of body fat rather than the total amount of adipose tissue (40). Truncal obesity is the accumulation of intra-abdominal or visceral adipose tissue (VAT) in and around vital organs such as the liver and pancreas. This, in conjunction with fat deposition in the muscle and the adipocyte size are major contributors to the development of metabolic dysfunction, insulin resistance, and subsequently T2DM in obese individuals (41).

In a cohort of 4865 participants (2642 females), aged 45 to 70 years, with normal to obese BMI, VAT mass negatively correlated with BMD at the total hip, femoral neck, and total body. In addition, men in the highest quartile of VAT had significantly lower BMD at all 3 sites than those in the lower quartiles (42). Among women, VAT mass also negatively correlated with BMD at the femoral neck, lumbar spine, and total body, with total body BMD significantly lower for those in the highest quartile of VAT compared with the lower quartiles (42). Other studies likewise confirmed the negative association between VAT with areal BMD measured by dual-energy X-ray absorptiometry (DXA) and with volumetric BMD measured by computed tomography (CT) of the spine cortical and/or trabecular bone (43-46). In a study by Vigevano et al, VAT volume and percentage of body fat were found to be negatively correlated with failure load and stiffness at the tibia, both of which were surrogate markers of bone strength measured by high-resolution peripheral quantitative CT (47). However, these findings were seen only in obese men who also have T2DM. Increasing VAT is associated with higher levels of proinflammatory cytokines such as C-reactive protein (CRP), IL-6, IL-1, and TNF-α, all of which are known to have unfavorable effects on the bone (15, 48-52). In the Study of Women's Health Across the Nation, fracture risk increased with increasing CRP, but only above the threshold of 3 mg/L (53). Although no relationship was found between CRP and BMD, composite femoral neck strength indices negatively correlated with CRP levels and may in part contribute to the increased skeletal fragility in obese subjects.

That the location of body fat matters has been highlighted in a study of 100 young women between the ages of 15 and 25 years, which found that subcutaneous fat had a positive predictive value on femoral bone structure and strength whereas the opposite was true for visceral fat. This suggests opposing effects of these fat depots on the bone (54). In another study of 228 Chinese men between 38 and 89 years old, total body and regional fat depots including subcutaneous and visceral fat were positively corelated with BMD, but this correlation disappeared with adjustments for covariates (55). Ectopic fat deposition in the muscles, which increases with aging and obesity, is also an important determinant of bone outcomes. The abundant inter/intramuscular fat has been reported to be associated with frailty, falls, osteoporosis, and fractures (56-58). Another important fat depot is the bone marrow adipose tissue, which increases with aging, menopause, osteoporosis, and obesity (59, 60). With obesity, changes in the hematopoietic and mesenchymal stem cell populations in the bone marrow occur with a shift in lineage differentiation from osteogenic to adipogenic, resulting in increased bone marrow fat content (61, 62). Likewise, with obesity, a change in the bone marrow microenvironment ensues that favors the production of pro-resorptive factors such as receptor activator nuclear factor kappa-B ligand, granulocyte colony-stimulating factor/monocyte colony-stimulating factor, tartrate-resistant acid phosphatase, including pro-inflammatory TNF-α with inhibition of the production of enhancers of osteoblast number/function such as osterix, runt-related transcription factor 2, bone morphogenetic protein 2, and Wnt10b (63). Altogether, these factors contribute to the development of skeletal fragility experienced by obese individuals.

To summarize, although total body fat may be a significant contributor to bone health, the location of increased fat mass seems to be a more important determinant. Available data suggest that fat accumulation in the viscera and ectopic sites such as the muscles have a more detrimental effect on the skeleton of obese patients. This negative influence is hypothesized to be secondary to the enhanced secretory capacity for proinflammatory cytokines by the adipocytes located in these fat depots. In addition, obesity is associated with increased bone marrow fat because of a shift in lineage differentiation away from the osteogenic to the adipogenic pathway, leading to enhanced production of pro-resorptive factors. Altogether, these events may result in a compromised skeleton.

Obesity and Falls

Obese individuals are at a higher risk for falls, and, in fact, for multiple falls (64-66) relative to nonobese subjects. Several factors have been proposed as underlying causes for the increased fall risk in these individuals. Increased fatty infiltration of the muscles in obese people may contribute to the weakness of the extremities (67, 68), leading to poor physical function. Sarcopenia, which is defined as low skeletal mass and function, is associated with increased risk of falls and even fractures (69-72). Sarcopenic obesity, which is a clinical and functional condition characterized by the coexistence of obesity and sarcopenia (69), is commonly associated with frailty, most especially in the elderly (73). This, in conjunction with postural instability (74, 75), may result in poor balance and increased risk for falls (76, 77). In a study of 9924 participants in the Women's Health Initiative (WHI) trial, those with sarcopenic obesity had a higher risk for falls during the 7-year observation period, higher among those aged 50 to 64 years than those aged 65 years and above, and highest among Hispanics (76). Moreover, the fall risk was highest among those with sarcopenic obesity compared with those who only had obesity or sarcopenia (76). Among participants of The Concord Health and Ageing in Men Project, those who had sarcopenic obesity had an increase in fall rates over 2 years and fracture risk over 6 years relative to obese subjects without sarcopenia (77).

Obese people also have reduced reaction time, and this coupled with failure to support the excess body weight during falls may result in fractures (78). More importantly, obesity is associated with several conditions such as diabetes, sleep apnea, hypertension, cardiovascular diseases, and chronic pulmonary diseases (79, 80). Notably, diabetes is associated with peripheral neuropathy and orthostatic hypotension that may predispose to falls. Osteoarthritis of weight-bearing joints such as the hip and knee commonly affects overweight and obese people (81), which further impairs activity and increases fall risk. Lastly, because of reduced agility and diminished protective mechanisms, obese subjects may be more likely to fall backward or sideways instead of forward (78), potentially accounting for their propensity for lower extremity and humeral fractures. Meanwhile, the increased soft-tissue padding around the hip may explain the protection against hip and pelvic fractures in obese individuals. Realizing these limitations, obese subjects may electively reduce physical activity, further perpetuating the cycle of weight gain and inactivity.

Obesity With and Without T2DM

Obesity Without T2DM

Bone mineral density, microarchitecture, biomechanical properties

In addition to higher BMD, obesity is associated with better bone microarchitecture. In a study by Evans et al, areal BMD by DXA, volumetric BMD by CT of the spine, and high-resolution peripheral quantitative CT of the radius and tibia in younger and older obese adults of both genders were higher compared with their normal weight counterparts (82). Furthermore, obese individuals had better bone microarchitecture and significantly higher bone strength (ie, failure load) than normal-weight participants, with the difference more pronounced in the older age group. By contrast, in a cohort of premenopausal or early perimenopausal women (N = 1924), baseline composite indices of bone strength were found to be significantly decreased with increasing BMI. After 9 years of follow-up, obesity was associated with increased fracture hazard, suggesting that, although BMD increases with greater skeletal loading, the magnitude may not be sufficient to compensate for the higher impact forces (from excess body weight) during a fall (83). However, the obesity–fracture association was attenuated with an adjustment for hip circumference, a surrogate for soft-tissue padding around the hip. On the other hand, data from the Study of Osteoporotic Fractures of 1377 obese women aged 65 years or older showed that the BMD T-scores at the total hip, femoral neck, and spine were significantly lower among those who experienced nonvertebral fractures compared with those who never had a fracture (84). These findings suggest that in the obese population, the subgroup who sustains a fracture may have “relative osteopenia” to account for the skeletal fragility.

Studies on the direct measurement of bone composition or biomechanical testing in obese subjects are lacking. However, microindentation, which is based on the principle that a bone with poor quality is more easily penetrated by a probe (needle <50 μm in size) than a bone with higher quality, has been gaining popularity as a method to assess bone quality (85). Impact microindentation, which is performed on the anterior tibia, measures the distance that the test probe indents into the bone relative to a reference that is located on the bone surface and is quantified as the bone material strength index (BMSi). It provides a direct measurement of tissue-level material properties of cortical bone in vivo (86-88). A higher BMSi is consistent with better bone strength, whereas the converse is true for lower BMSi values.

A study of 567 men aged 33 to 96 years in the Geelong Osteoporosis study reported a negative correlation between body weight and BMSi (89). In another study among older women, BMSi was found to inversely correlate with BMI, whole body fat mass, and subcutaneous fat at the tibia (90). Nonetheless, these studies were not designed to address differences in these parameters between obese and nonobese participants.

Fractures

Whereas fractures from T2DM can affect any site (91, 92), obesity-associated fractures consistently involve the lower extremities and to a certain extent the humerus (4). The Global Longitudinal Study of Osteoporosis in Women, which is a multinational 2-year observational study, enrolled 57 556 women, 23.8% of whom were obese, 74.4% were nonobese, and 1.9% were underweight (4). Prior fractures occurred in 16.6% (N = 2274) of obese and 17.3% (N = 7401) of nonobese women. Obese women were more likely to have had previous ankle or lower leg fractures and less likely to have experienced wrist, hip, rib, or pelvis fractures than nonobese women. Of the 46 443 who completed both 1- and 2-year follow-up surveys, 23.4% (n = 10 441) were obese, 74.9% (n = 33 349) were nonobese, and 1.7% (n = 744) were underweight. Incident fractures occurred in 6.1% (N = 633) of obese and 6.5% (N = 2170) of nonobese women. Of these, the ankle fracture rate was higher, whereas the wrist fracture rate was lower among obese compared with nonobese women. In this study, although there was no difference in prevalent and incident fractures between obese and nonobese patients, the fracture pattern appeared different. Obese women with fractures also had a higher incidence of falls over the 2-year period of observation and more comorbidities relative to nonobese women (4). In the WHI Observational Cohort, lower extremity fractures were significantly higher in overweight subjects with an upward trend for the different grades of obesity, whereas hip fractures were lower for overweight and obese subjects relative to normal weight subjects (5).

Others reported an increase in humeral fractures in obese women. In a cohort of 832 775 postmenopausal women of Spanish ancestry, compared with normal/underweight women, overweight and obese women had a lower risk of hip and pelvic fractures but a significantly higher risk of proximal humerus fracture (6). This protection from a hip fracture has been confirmed by a meta-analysis of 15 studies demonstrating that obese men and women had a reduced risk for hip fractures (93). Likewise, another study showed mostly ankle and humeral fractures in obese men and women. Although the humeral fractures mostly involved the proximal humerus in males, it mostly involved the humeral shaft in females (94). However, in a study of 5995 U.S. men aged ≥65 years who participated in the Osteoporotic Fractures in Men Study, although the authors reported that the hazard ratio (HR) for nonspine fracture increased with increasing weight classification (ie, overweight, obese 1, and obese 2) (7), they also observed that those with grade 2 obesity were 5 times more likely to experience a hip fracture than normal-weight men. Nevertheless, the majority of studies indicate that obesity-associated fractures mostly affect the lower extremities and the humerus to a certain extent, leading to a conclusion by some that fragility fractures among obese subjects are site specific (6).

Despite some authors reporting no increase in vertebral fractures in patients with obesity, a few investigators have reported contradictory findings. In a study of older women with a mean age of 60.8 ± 5.8 years and a mean BMI of 32.7 ± 4.4 kg/m² participating in a weight loss program, the authors reported that 44% of these women had vertebral deformities at baseline (95). However, these abnormalities were identified only from DXA images used to measure spine BMD; no radiograph confirmation was performed. Regardless, a systematic review and metanalysis of 23 published studies also reported a significant positive association between abdominal obesity and risk of vertebral fractures (96). In this meta-analysis, vertebral fracture determination varies from 1 publication to another through self-reports, review of medical records, and some by actual radiographs. Furthermore, obesity classification was not based on BMI but instead on either waist circumference, waist-to-hip ratio, abdominal fat mass, or VAT volume. Although the results of these 2 studies seem to contradict results from most studies discussed previously, the methods between studies vary (ie, using BMI vs. abdominal measures to define obesity) and fracture adjudication by self-report, medical record reviews, or radiographs. There is also a possibility that some of the obese subjects enrolled in these studies could have concurrent T2DM, which is associated with fractures in any skeletal site, and may account for this increase in vertebral fractures.

Obesity With T2DM

Bone mineral density, microarchitecture, biomechanical properties

As far as bone health is concerned, a similarity exists between obesity and T2DM. Both are associated with an increased risk of fractures despite normal to better than normal BMD (3, 8, 2, 97-99). This observation is not surprising because obesity is a risk factor for insulin resistance and subsequently the development of T2DM (100). Given that the prevalence of T2DM increases with increasing BMI (101), it may be difficult to separate the effect of obesity from the effect of diabetes on bone. In a recent publication, our group attempted to differentiate the effect of obesity alone from the effect of concurrent T2DM on the bone (47). Using components of the metabolic syndrome, 112 obese men were categorized into different phenotypes as proposed by prior investigators into metabolically healthy obese and metabolically unhealthy obese (MUHO) with or without T2DM (41). We observed that those who were MUHO with T2DM had significantly lower osteocalcin (a bone formation marker) and C-terminal telopeptide of type 1 collagen (a bone resorption marker) compared with both men who were metabolically healthy obese and MUHO without T2DM. Furthermore, parameters of bone strength (ie, bone stiffness and failure load) were significantly lower among those with T2DM compared with those without T2DM (47). These findings suggest that among obese men, diabetes could be the driver of the skeletal abnormalities observed. In this study, no normal-weight subjects were included for comparison. However, another group reported a reduction in bone turnover markers in normoglycemic young (between age 17 and 25 years) obese compared with nonobese subjects (102), suggesting that obesity by itself may be associated with an alteration in bone metabolism.

Given the relationship between obesity and T2DM, some of the underlying pathophysiology of skeletal dysfunction associated with obesity can also be found in patients with T2DM. But for T2DM, there are additional insults related to hyperglycemia that include but are not limited to increased accumulation of advanced glycation end-products (AGEs) in the bone, suppressed bone turnover, the side effects of glucose-lowering medications, the alteration in bone material tissue property, poor bone strength, the impact of complications from chronic kidney disease in some, and the increased propensity to falls resulting from neuropathy and retinopathy, all of which contribute to increased fracture risk (103-105). A study by Furst et al of 16 postmenopausal women with T2DM compared with 19 matched controls showed that BMSi was significantly reduced in subjects with T2DM, and this also inversely correlated with the duration of T2DM (106). Additionally, skin autofluorescence, an indirect measure of AGEs, also negatively correlated with BMSi and procollagen type 1 amino-terminal propeptide (an index of bone formation) only in patients with T2DM. Taken together, these findings suggested that increased AGEs contributed to the deterioration in bone material properties observed in patients with T2DM. Moreover, a study by Nilsson et al showed that compared with patients without diabetes, bone microarchitecture was better and bone strength by finite element analysis (failure load and stiffness) was higher in patients with T2DM. However, BMSi was much lower in patients with T2DM compared with those without (107). Another study also found similar BMD and bone microarchitecture among those with and without T2DM, but radial cortical porosity tended to be higher in individuals with diabetes (104). Furthermore, in agreement with the previous study, those with diabetes had reduced BMSi compared with controls, implying that skeletal fragility in patients with T2DM is perhaps a problem of bone material properties or bone composition rather than structure.

Fractures

Although obesity is associated with increased nonvertebral fractures with sparing of the hip, pelvis, and wrist, diabetes-associated fractures can involve any site (91, 92, 108). As far as hip fracture is concerned, many studies have found an increased risk among patients with T2DM (91, 92, 108). A meta-analysis of 8 studies showed an age-adjusted relative risk for hip fracture of 1.38 (95% CI, 1.25-1.53) in subjects with diabetes compared with those without (109). In the WHI observational cohort, the relative rate of hip fracture was 1.41 (95% CI, 1.17-1.70) and 1.29 for any fracture among those with T2DM compared with those without (91). Similar to patients with obesity, patients with T2DM have normal or above normal BMD. A study by Schwartz et al, showed that for a given T-score, the risk for nonvertebral and hip fractures among subjects with T2DM was much higher compared with those without T2DM (98). Because obesity is reported to be associated with protection from a hip fracture, a group of investigators in Taiwan examined whether obesity influenced the risk for hip fractures among patients with T2DM in a group of 22 048 men and women over the age of 40 years (110). They found that, compared with those without diabetes, an increased risk for hip fracture was only found among those with lower BMI (18.5 to <24 kg/m²) and not among those who have a higher BMI, including those with BMI of 30 kg/m² and above. There was no influence of gender. Another group from Singapore reported the result of a prospective hip fracture study of a cohort of 63 257 men and women, aged 45 to 74 years, followed for 12 years (111). They found that the risk of hip fracture was almost doubled among those with diabetes. Stratifying the subjects further according to BMI showed that for every BMI category (from underweight to obese), there is a significantly higher risk for hip fracture among participants with diabetes compared with those without diabetes. However, in the subgroup with diabetes, there was no significant difference in risk between the BMI categories, implying the lack of influence from body weight. Similarly, no influence from gender was found. A retrospective study by Adami et al of 59 950 women of Italian ancestry with an average age of 65.1 ± 11.0 years, of which 5091 had obesity but without diabetes, 3114 had diabetes but without obesity and 741 who had both obesity and diabetes, tried to separate the effects of obesity alone from diabetes alone or both on prevalent fractures. Relative to normal weight women, they found an odds ratio for vertebral or hip fracture of 1.2 (95% CI, 1.1-1.3) and for nonvertebral, nonhip fracture of 1.1 (95% CI, 1.1-1.2) among obese women without diabetes (108). The corresponding figures for those with both obesity and diabetes were higher, 1.5 (95% CI, 1.3-1.8) and 1.8 (95% CI, 1.5-2.1) for vertebral or hip fracture and nonvertebral nonhip fracture, respectively (108). The odds ratio for nonvertebral nonhip fracture in nonobese patients with diabetes compared with normal-weight individuals was 1.4 (95% CI, 1.2-1.5). However, the ORs for having had at least 1 vertebral or hip fracture were much higher in nonobese subjects with diabetes (odds ratio, 1.9; 95% CI, 1.7-2.1) when compared with those with both obesity and diabetes (OR 1.5, 95% CI 1.3–1.8). These somewhat variable findings could be due to differences in the population under investigation (ie, ethnic composition, gender, and age). Although analyses were adjusted for age and gender, the effect of ethnic variations between studies could not be controlled for. Moreover, there are differences in the BMI-based definition for overweight and obese in Asians compared with other racial groups, and this likely affects fracture adjudication across the different BMI categories.

Among patients with diabetes, the degree of glycemic control appears to influence fracture risk, as suggested by several studies demonstrating an association between glycemic control, measured by glycated hemoglobin (HbA1c), and risk for fractures. This is concordant with findings from previous studies demonstrating a significantly reduced bone turnover and bone strength among subjects with T2DM with a single or mean HbA1C over the preceding 12 months of >7%, respectively (112, 113). Although some reported a linear relationship between HbA1c and fractures (ie, high HbA1c was associated with increased risk for fractures), others did not (114-122). It is possible that the discrepancies in findings between studies may reflect differences in the HbA1c cutoffs used in assessing glycemic control (single vs. multiple HbA1c measurements). On the contrary, others reported an increase in fractures among patients with T2DM with tight glucose control. This was mostly observed among those who were on treatment for T2DM or those who were insulin users (121, 122). These findings were postulated to be secondary to an increase in hypoglycemic events and consequently falls. Regardless, there seems to be a preponderance of evidence supporting the notion that glucose control could be an important determinant in the incidence of adverse bone complications from diabetes.

Management of Bone Health in Obesity, T2DM and During Weight Loss

Thus, although obesity itself may have harmful effects on the skeleton, the concurrent presence of T2DM may make bones even more fragile. However, bone assessments in people with obesity and T2DM, whether alone or in combination, are quite limited. The only readily available measure for skeletal integrity in the clinical setting is the use of DXA for BMD testing, which may give falsely reassuring results. Therefore, skeletal fragility in these patients may be underestimated and bypass treatment or other preventive measures before the first fracture. Although the availability of trabecular bone score (TBS) may provide further information on skeletal health (123, 124), not every DXA machine has the software that performs this analysis on the spine because this test entails additional software and costs to the institution. More importantly, TBS is not considered valid in subjects with BMI greater than 37 kg/m² because of the effects of soft tissues on the measurement (125). TBS could be affected by the increased noise in the image brought about by the excessive volume of soft tissues in patients with obesity. A study using spine phantom scanned with an overlying soft tissue–equivalent material showed that at least 1 cm of soft-tissue thickness is associated with a significant decrease in TBS (126). To mitigate this technological imprecision associated with obesity, a software that corrects the TBS measurements with the soft-tissue thickness was developed (127). Also, given that the sites of fractures in obese patients are typically not the skeletal regions measured by conventional DXA imaging, the measured BMD may not reflect the actual level of skeletal fragility in these patients. Although adjustments in the FRAX score in patients with diabetes have been suggested (128), an updated version of FRAX will reportedly be available and will include diabetes as a risk factor. Despite the incorporation of weight and height in the FRAX calculator, the current model is designed to predict hip and major osteoporotic fractures but not peripheral skeletal fractures, which are more common in obese individuals.

To date, there are no clinical guidelines on the approach to the skeletal health of obese subjects. Among those who also have concurrent T2DM, one can follow the algorithm suggested by the Bone Working Group of the International Osteoporosis Foundation (129) for the management of bone health in patients with diabetes. Nevertheless, the current practice in these patients is primarily focused on the improvement of metabolic abnormalities with very little attention paid to bone health aside from routine vitamin D supplementation in those with vitamin D deficiency. Lifestyle intervention, with the goal of weight loss, is the cornerstone in the management of obese patients. The benefits of weight loss from the metabolic standpoint are undeniable (130). However, concerns about weight loss–associated bone loss were raised in previous observational and intervention studies that showed that bone loss at the hip and an increased risk of hip fractures occurred regardless of whether the weight loss was intentional or unintentional (131-133). In a randomized, controlled study by Shah et al, a 10% weight loss over 1 year was accompanied by a significant increase in markers of bone turnover and a decrease in bone density at the total hip among those on diet alone (133). The increase in bone turnover markers and bone loss at the total hip was attenuated but not prevented by the addition of aerobic plus resistance exercise to diet among those randomized to diet plus exercise. By contrast, there was no change in the markers of bone turnover and bone density among those randomized to control and exercise-alone groups in which, by study design, no weight loss occurred (134). All the participants in this study received calcium and vitamin D supplementation. To determine which type of exercise would be more beneficial in improving physical function and minimizing loss of bone and lean mass, the same group randomized a group of 160 frail, obese older adults to supervised weight loss by diet plus aerobic exercise, resistance exercise, or both vs. control for 6 months. Bone loss and loss of lean mass were greatest in the weight loss plus aerobic exercise group, the least in the weight loss plus resistance exercise, and intermediate among those randomized to the weight loss plus a combination of aerobic and resistance exercise (135) (Fig. 1). Improvement in physical function was, however, the best in the combined aerobic and resistance exercise group. Another study that randomized overweight to obese postmenopausal women with either weight loss or weight loss plus aerobic exercise for 6 months failed to find a difference in bone loss between the 2 groups (136). Taken together, these results suggest that patients undergoing voluntary weight loss should be counseled in performing exercises with a preference for resistance over aerobic exercise to prevent or at least minimize bone loss.

Figure 1.

Mean percentage changes from baseline in bone mineral density at the (A) total hip, (B) femoral neck, (C) trochanter, (D) intertrochanter, (E) lumbar spine, and (F) one-third radius during the interventions. *P < 0.05 for the comparison of the value from the control group; †P < 0.05 for the comparison of the value from the aerobic group. I bars indicate standard errors. (Reproduced from Armamento-Villareal R, Aguirre L, Waters DL, Napoli N, Qualls C, and Villareal DT. Effect of aerobic or resistance exercise, or both, on bone mineral density and bone metabolism in obese older adults while dieting: a randomized controlled trial. J Bone Miner Res. 2020;35:430–439. https://doi.org/10.1002/jbmr.3905).

A meta-analysis of 41 studies on weight loss–induced bone loss from lifestyle intervention demonstrated overall bone loss on the total hip but not on the spine (137). In addition, the authors also found an overall increase in markers of bone turnover, mostly in the first few months of weight loss (137). Although small changes in BMD may accompany weight fluctuations and may produce measurement artifacts between scans related to alteration in the amount of soft tissue around the bone, the increase in bone turnover reported in interventional studies (133, 138) and the increase in fractures reported in observational studies (131) suggest that bone metabolism is certainly affected by weight loss. One must keep in mind that the effect of weight loss on the skeleton can be rapid. In the Global Longitudinal Study of Osteoporosis in Women study, subjects who self-reported >10 lb of unintentional weight loss at baseline were noted to have a significantly increased risk of incident fractures of the clavicle, wrist, spine, rib, hip, and pelvis for up to 5 years after weight loss (139). The increase in fracture risk was found as early as 1 year after weight loss. Although this finding suggests a very rapid bone loss and decline in bone strength following weight loss, one must also consider that these women had unintentional weight loss, which may suggest concomitant illness that may have contributed to the skeletal deterioration. Follow-up of participants in the WHI Observational Study and Clinical Trials also showed increased fractures in women who experienced weight loss over a 3-year period of observation. However, the fracture pattern appears to differ between those with an unintentional vs. intentional weight loss. Relative to women with stable weight, those who had unintentional weight loss of 5% or more had a 33% higher incidence rate of hip fracture (1.33; 95% CI, 1.19-1.47) and vertebral fracture (1.16; 95% CI, 1.06-1.26). By comparison, those with intentional weight loss of 5% or more had an increased incidence rate of lower limb fracture (1.11; 95% CI, 1.05-1.17) but a decreased incidence of hip fracture (0.85; 95% CI, 0.76-0.95) (140). On the other hand, in patients who also have T2DM, it remains uncertain if there is a weight loss–associated increase in bone turnover, and, if so, would it be harmful or even potentially beneficial in this group of subjects who have suppressed bone turnover at baseline (47, 113, 141).

Armamento-Villareal et al reported that bone loss from weight loss could be in part mediated by an increase in sclerostin, an inhibitor of bone formation (142). Among subjects undergoing weight loss from lifestyle intervention, an increase in sclerostin level was observed among subjects randomized to weight loss by diet alone, which was not observed for those randomized to diet plus exercise who experienced an equivalent amount of weight loss. Sclerostin is increased in states of unloading and likely mediates bone loss from weight loss to a certain extent (143, 144). In this study, though the increase in sclerostin was prevented by adding exercise to diet, bone loss at the total hip was only attenuated but not completely prevented suggesting that mechanisms other than unloading could also be involved (142).

The bigger concern regarding weight loss-associated bone loss is the observation that bone loss continues even after weight loss has stopped or even after some weight regain has occurred (136, 145, 146). In a study among postmenopausal women who experienced significant bone loss at the lumbar spine after losing an average of 3.9 ± 3.5 kg over a 6-month period of lifestyle intervention, subsequent follow-up 12 months later with an average weight regain of 2.9 ± 3.9 kg showed no regain of lost bone (145). In addition, the bone turnover markers that increased during the weight loss period remained elevated despite the weight regain. A similar observation was found in a group of elderly men and women who experienced significant bone loss at the total hip after undergoing weight loss from a 1-year lifestyle intervention. Follow-up 18 months after the end of the intervention period showed that the participants regained ∼3% of weight but without bone regain (147). Another study showed that bone loss may even continue despite weight regain (136).

If bone loss correlates with weight loss, then bariatric surgery may be expected to have the most significant bone loss given the amount of weight loss. The malabsorption of nutrients and minerals along with mechanical unloading both contribute to bone loss and increased fractures observed in these patients (148-152). Recently, among surgical procedures, sleeve gastrectomy (SG) has outpaced Roux-en-Y gastric bypass (RYGB) surgery as the most common bariatric surgery performed (153). A meta-analysis of 33 studies showed that the BMI loss at 1 year is greater among those who underwent RYGB compared with those who had SG (154). On the other hand, compared with those who underwent SG, those who had RYGB may experience greater nutrient deficiencies, including vitamin D deficiency leading to secondary hyperparathyroidism, greater increases in bone turnover, and relatively more bone loss than those who had SG (155-159). Comparing bone turnover markers in a group of 95 patients randomized to medical weight loss, RYGB, and SG, Crawford et al reported significant increases in bone turnover markers at 5 years in the surgical groups, but the increase was greater in the RYGB group (157). In a study of 21 morbidly obese subjects with 11 undergoing RYGB and 10 SG, follow-up at 12 months showed increases in bone turnover markers in both groups but the increase in P1NP was greater for those who had RYGB than SG (156). More importantly, although the decline in BMD was experienced by both groups, the decline at the total hip and femoral neck was significantly greater among those who had RYGB surgery.

In terms of fractures, a meta-analysis of 6 studies showed that the risk for any type of fracture was increased by 29% in those who had bariatric surgery compared with nonsurgical controls, and most significantly in nonvertebral fractures, particularly upper limb fractures (160). Compared with controls, the risk for fracture was higher in the first 2 years after bariatric surgery. In a single-institution observational study of 3439 patients who underwent bariatric surgery between 1965 and 2015 (79% RYB and 11% SG), a significantly higher risk for fracture was observed in the bariatric surgery group compared with a matched nonsurgical group over a mean follow-up of 7.6 ± 5.6 years after surgery, with the risk much higher for RYGB than SG (161). A recent report showed an increase in major osteoporotic fractures only in patients with RYGB but not in those with SG (162). Comparing malabsorptive versus restrictive procedures showed a fracture incidence rate of 6.6% in the RYGB group and 4.6% for those who had adjustable gastric banding for a hazard ratio (HR) of 1.73 (95% CI, 1.45-2.08) after 3.5 years follow-up (150). Greater increases in HRs were noted at the hip, wrist, and pelvis for RYGB compared with adjustable gastric banding.

Calcium and vitamin D supplementation, considered as part of the standard of care for postbariatric surgery patients, were not enough to prevent bone loss in these subjects. Furthermore, preoperative zoledronic acid infusion was also unable to prevent the increase in bone turnover and bone loss on the total hip after RYGB at the end of the 24-week study period in a small group of subjects compared with a historical control group that did not receive the medication (163). On the other hand, a pilot study on the use of risedronate following SG showed that risedronate reduced bone loss at the hip and prevented bone loss at the spine compared with placebo after 12 months of therapy (164). Because sclerostin partly mediates bone loss in states of mechanical unloading, which could be significant in postbariatric surgery patients, the use of Romosozumab (an inhibitor of sclerostin) may need to be explored in this context. However, because the drug is only approved for 1 year, other long-term strategies for bone loss prevention outside of calcium and vitamin D, have to be considered. A regular exercise regimen, especially resistance training, should be encouraged in anyone who has had bariatric surgery. In a randomized controlled study of 70 women who had RYGB, a 6-month exercise training program consisting of a combination of resistance and aerobic exercise was only able to mitigate but not prevent the increase in bone turnover and bone loss at the total hip and femoral neck (165). Another study by Diniz-Sousa et al on 84 postbariatric subjects with half randomized to a multicomponent exercise program and the other half randomized to the usual postbariatric surgery care, showed that there was bone loss on the spine, total hip, and femoral neck at the end of 12 months in both groups (166). However, bone loss at the lumbar spine was attenuated in the exercise group. Though there was an increase in BMD from baseline at the wrist in the exercise group, bone loss at the total hip and femoral neck was not prevented by exercise. However, when analysis was limited to those whose exercise attendance was >50%, femoral neck BMD in the exercise group was much higher compared with the control group (166). Because exercise is relatively safe and can be maintained for an indefinite period without additional concerns, it should be encouraged as part of bone health management in anyone undergoing weight loss.

In conclusion, although obese individuals have normal or above-normal BMD, they are at an increased risk for fractures, suggesting poor bone quality. This skeletal complication appears to be aggravated by coexisting T2DM; diabetes alone, in the presence or absence of obesity, is associated with increased skeletal fragility and possibly worse outcomes. On the other hand, weight loss through lifestyle intervention or by surgical means can also result in bone loss and increased fractures, suggesting that even the standard of care in the management of obese patients could potentially further aggravate the existing problem of poor bone quality. Regardless, the cardiometabolic, physical function, and quality-of-life benefits (130, 134, 135, 167) from weight loss in overweight or obese individuals are overwhelming and greatly outweigh the negative effects on the skeleton. However, at this time, ways to prevent skeletal complications from weight loss are not particularly effective, and more research on this important topic is essential. Given the rapid increase in the number of individuals who are obese, the approach to the management of these patients should also include methods of improving bone health, which to date, appear to be of only secondary importance.

Abbreviations

AGE

advanced glycation end-product

BMD

bone mineral density

BMI

body mass index

BMSi

bone material strength index

CRP

C-reactive protein

CT

computed tomography

DXA

dual-energy X-ray absorptiometry

FRAX

Fracture Risk Assessment

HbA1c

glycated hemoglobin

HR

hazard ratio

MUHO

metabolically unhealthy obese

RYGB

Roux-en-Y gastric bypass

SG

sleeve gastrectomy

T2DM

type 2 diabetes mellitus

TBS

trabecular bone score

VAT

visceral adipose tissue

WHI

Women's Health Initiative

Funding

NIH R01 HD093047 and US Department of Veterans Affairs 101CX001665.

Disclosures

The authors have nothing to disclose. The contents of this manuscript do not represent the views of the US Department of Veterans Affairs or the United States Government.

Data Availability

Data sharing is not applicable to this article as no datasets were generated or analyzed during the current study.