Introduction

Obesity has increased in the last decades, assuming an epidemic character as it affects approximately 600 million people globally and about one-third of the North American population [1]. In this context, bariatric surgery has been increasingly used as a therapeutic option for obese individuals who do not obtain satisfactory results from conventional treatments. Currently, it is considered the most effective and long-lasting treatment for obesity [2, 3]. Among the available surgical techniques, Roux-en-Y gastric bypass (RYGB) has been one of the most used, because of its lower rates of morbidity and mortality and high excess weight loss percentage [4].

Despite the benefits related to weight loss, this procedure has negative repercussions on bone metabolism [5–8], which could be particularly concerning in women [6]. Although the pathophysiological mechanisms responsible for these changes are still not totally elucidated, a possibility would be that weight and body composition changes induced by surgery would promote changes in hormone metabolism and adipose tissue, leading to variations in the levels of important adipokines [9–12], which could modulate the expression of the receptor activator of nuclear factor kappa β ligand (RANKL) and osteoprotegerin (OPG) [13–17].

Moreover, the lower mechanical load stemming from rapid and intense weight loss after surgery may increase the expression of sclerostin [6, 18–20]. This protein, coded by the SOST gene and produced exclusively by osteocytes, inhibits bone formation by inhibiting the terminal differentiation of osteoblasts and promotes their apoptosis [21, 22]. Sclerostin also reduces bone formation by Wnt, binding to the receptor LRP5/6 in osteoblast cell membrane, blocking the signaling pathway Wnt/β-catenin, and inhibiting the transcription of osteogenic genes [23–26].

Thus, changes in the levels of adipokines, RANKL, OPG, and sclerostin could be involved in the bone remodeling changes that occur after bariatric surgery, with bone repercussions proportional to weight loss [19, 20, 27–29]. However, studies on the subject are still scarce, and the relationships that exist between these proteins and bone remodeling after surgery have not been totally clarified. Thus, the objective of this study was to assess the repercussions of RYGB-related weight loss and body composition, biochemical, and hormonal changes on serum bone remodeling markers.

Materials and Methods

This prospective and observational study assessed the biochemical, hormonal, and bone remodeling parameters of obese women before and after RYGB, and it was approved by the local Research Ethics Committee (protocol number 3710-2010) and the studies were conducted according to the principles of the 1964 Helsinki Declaration. Informed consent was obtained from all individual participants included in the study.

Patients

The sample consisted of menstruating women aged 18 to 50 years, with obesity grades II (gII; BMI 35–39.99) or III (gIII; BMI ≥40.0). The women underwent RYGB according the institution’s protocol between May 2009 and December 2011. After the women were informed about the objectives of the study and the data collection procedures, they read and signed an informed consent form.

The exclusion criteria were the presence of the following conditions diagnosed by protocol tests established by the American Consensus for Bariatric Surgery [2]: severe renal or hepatic diseases, collagen diseases, hypogonadism, hyperparathyroidism, hyperthyroidism, type 2 diabetes mellitus (DM) decompensated or hypercortisolism, or use of drugs that knowingly affect bone metabolism (corticosteroids, thiazides diuretics, anticonvulsant, and furosemide); discontinuation of nutritional or medical follow-up; and gestation.

Preoperatively, the patients were followed at an outpatient clinic for 12 months by dietitians, psychologists, and endocrinologists to better prepare them for surgery and minimize perioperative risks. During this time, the patients underwent tests to exclude endocrine causes of obesity and assess surgical risk. Of the 35 initial candidates, one patient died 1 week after surgery, and another, 6 months after surgery, both from obesity-related complications. Three patients were lost to follow-up. Thus, 30 patients were effectively studied.

Study Design

The patients eligible for surgery were called 3 days before the procedure for data collection (age, self-reported ethnicity) and baseline anthropometric and biochemical assessments (T0). These assessments were repeated 3 (T3), 12 (T12), and 24 months (T24) after surgery. In each visit, the patients received medical and nutritional advice and were prescribed a multivitamin/multimineral supplement (Materna®), protein supplement (30 g/day, Isofort®, for the first 30 days after surgery), calcium citrate (1200 to 1500 mg/day), cholecalciferol (3000 IU/day), and intramuscular vitamin B12 (5000 IU every 3 months, Citoneurin®) [30].

Anthropometric measurements (weight (in kilograms (kg)) and height (in meters (m)), body mass index (BMI), weight loss (%), and body composition (fat mass (FM, in kilograms and percentage) and the fat-free mass (FFM, in kilograms and percentage) determined by bioelectrical impedance analysis (Maltron BF model 906; Rayleigh, Essex, England))) were collected on all study occasions by the same researcher (MFGB).

Laboratory Analysis

The studied bone remodeling markers were collagen type 1 C-telopeptide (CTX) and bone-specific alkaline phosphatase (BSAP). The other biochemical and hormonal variables were serum calcium (Ca), parathyroid hormone (PTH), 25-hydroxy vitamin D (25-OH-VitD), leptin, adiponectin, RANKL, OPG, and sclerostin.

Blood was collected in tubes containing separator gel components, using a needle coupled with the Vacutainer® system, and the tubes were centrifuged at 3000 rpm by a cooled centrifuge for 10 min. The serum was frozen to −80 °C until determination of the Ca and PTH levels by the dry chemistry method and chemiluminescence, respectively. BSAP and CTX were determined by chemiluminescence (Beckman Coulter, Fullerton, CA); 25-OH-VitD was measured by high-performance liquid chromatography (HPLC) (DiaSorin kit, Stillwater, MN, USA). The serum was also used for the measurements of adiponectin, leptin, RANKL, OPG, and sclerostin, which were measured by enzyme-linked immunosorbent assay-ELISA (Biomedica Medizinprodukte GmbH & Co. KG, Vienna, Austria).

Statistical Analyses

Statistical analyses of the anthropometric and laboratory data were performed by the software SPSS® version 21.0 for Windows (Statistical Package for the Social Sciences) and Microsoft Office Excel version 2010. The study occasions were compared by the non-parametric Friedman test, followed by Dunn’s multiple comparison test. Spearman’s correlation coefficient verified the correlation between the various study variables. All analyses used a statistical significance level of 5%.

Results

Clinical Characteristics

The median age of the patients at baseline was 41.5 years, and 70% (n = 21) reported being white. The calculated 24-month excess weight lost (%EWL) for these patients was 71.9 ± 18.1% using the calculation method [62]. The median (min-max) BMI decreased from 48.3 (37.8; 62.2) kg/m2 before surgery to 39 (30.8; 50.9) kg/m2, 33.1 (23.1; 41.2) kg, and 31.8 (23; 40.4) kg/m2 at 3, 12 (p < 0.001), and 24 months of procedure, respectively. Before surgery, 90% of the patients had obesity gIII; at the end of the follow-up, 57% (17) of the patients continued to be obese, and only 10% (3) had a BMI compatible with their ideal weight (p < 0.01; Table 1). Surgery promoted progressive and significant loss of weight and its compartments, which were more expressive in the first year after surgery (p < 0.05; Table 1). Total weight decreased by a median (min-max) of 40.95 (21.0; 75.0) kg, of which 33.45 (16.3; 63.5) kg were FM and 7.9 (1.4; 17.1) kg were FFM. In percentages, fat mass decreased by 12 (5.1; 30.8) %, and FFM increased by 12.03 (5.1; 30.9) % (p < 0.05).

Table 1 Weight, body mass index, and body compartments of 30 patients submitted to Roux-en-Y gastric bypass (RYGB) at baseline and 3, 12, and 24 months after surgery
Full size table

Laboratory Data

The baseline serum levels of 25-OH-VitD were compatible with vitamin D insufficiency or deficiency in 90% of the patients. This percentage had decreased to 70% by T3 and T12 and to 66% by T24. Ca and PTH levels did not change between T0 and T24 (p > 0.05) (Table 2).

Table 2 Laboratory test results of 30 patients submitted to Roux-en-Y gastric bypass (RYGB) before and 3, 12, and 24 months after surgery
Full size table

Adiponectin increased significantly while leptin decreased between T0 and the other occasions. Unlike adiponectin, whose level changed between T3 and T12, leptin level did not change between the postoperative follow-ups (Table 2; Fig. 1).

Fig. 1
figure 1

Changes in serum adipokine, bone remodeling markers, and other bone markers in 30 patients submitted to Roux-en-Y gastric bypass (RYGB) before (T0) and 3 (T3), 12 (T12), and 24 (T24) months after surgery. a Adiponectin (T0 < T3, T12, T24; T3 < T24; p < 0.05). b Leptin (T0 > T3, T12, T24; p < 0.05). c Collagen type 1 C-telopeptide (CTX) (T0 < T3, T12; T3 > T24; p < 0.05). d Bone-specific alkaline phosphatase (BSAP) (T0 < T12, T24; T3 < T12, T24; p < 0.05). e Receptor activator of nuclear factor kappa β ligand (RANKL) (T0 < T3; T3 > T24; p < 0.05). f Osteoprotegerin (OPG) (T0 > T3, T12, T24; p < 0.05). g Serum sclerostin changes in 30 patients submitted to Roux-en-Y gastric bypass (RYGB) before and 3, 12, and 24 months after surgery (T0 < T3 ; T3 > T12, T24; p < 0.05)

Full size image

RANKL and sclerostin were higher at T3, but OPG was lower. Postoperatively, RANKL changed between T3 and T12, sclerostin changed between all three postoperative follow-ups, and OPG did not vary (Table 2; Fig. 1).

Serum CTX increased significantly between T0 and T3, remained high at T12, and decreased by T24 (p < 0.001) (Table 2; Fig. 1). BSAP was higher at T12 (p < 0.001) and remained high at T24 (Table 2; Fig. 1).

At T24, CTX correlated inversely with body weight (r = −0.451; p = 0.024), FM (r = −0.418; p = 0.038), and FFM (r = −0.417; p = 0.038) in kilograms. CTX also correlated negatively with serum leptin preoperatively (r = −0.40; p = 0.031), at T24 (r = −0.554; p = 0.004), and with OPG at T3 (r = −0.37; p = 0.048); finally, it correlated positively with adiponectin at T24 (r = 0.453; p = 0.023). RANKL was directly correlated with body weight at T3 (r = 0.366; p = 0.046). Sclerostin correlated inversely with body weight at T0 (r = −0.379; p = 0.039) and with FM at T3 (r = −0.391; p = 0.03). BSAP correlated negatively with 25-OH-VitD at T12 (r = −0.412; p = 0.033) and positively with PTH at T24 (r = 0.49; p = 0.024).

Discussion

The negative influence of weight loss and the hormonal and biochemical changes induced by bariatric surgery on bone health has been reported [7, 11, 20, 31–35], in addition to the association between excess body weight loss and bone mineral density (BMD) 1 to 2 years after RYGB [20, 34, 36].

Bone turnover increased in the study sample, evidence by higher CTX and BSAP after RYGB. Interestingly, CTX increased 3 months after the procedure, remained high at the 1-year follow-up, and decreased by the 2-year follow-up. These findings corroborate Yu et al. (2015) who found that this marker increased significantly and was associated with BMD 2 years after RYGB [7]. Elias et al. (2014) compared RYGB with vertical-banded gastroplasty, found that these changes did not occur in women submitted to gastroplasty, and suggested that bone loss may be more strongly associated with techniques that promote malabsorption [37]. The BSAP levels in the study sample had increased by T12 and remained high at T24, confirming Bruno et al. (2010), who observed this same behavior in this enzyme after the same follow-up period [38]. These findings suggest that bone formation occurs after bone resorption. Indeed, another study did not find changes in this marker shortly after RYGB [39].

The mechanisms involved in the relationship between bariatric surgery and bone changes are complex and not fully understood. A possibility would be the effect of adipokines on bone remodeling and markers. However, data on this effect are controversial. Some studies find an association with bone formation markers and report positive [13, 19, 40] or negative action, regardless of gender, age, and BMI [41, 42]. The adiponectin of the present sample was directly correlated with CTX at T24. Its levels increased progressively from T3 to T12, but had stabilized by T24. As a matter of fact, this adipokine is inversely related to the amount of adipose tissue and can be associated with postoperative bone changes [43]. Although scarce, studies have reported an increase in adiponectin and its association with higher bone resorption, evidenced by higher collagen type 1 cross-linked N-telopeptide (NTX), or lower total, spinal, and pelvic BMD in patients submitted to RYGB [29, 44, 45].

Leptin was inversely correlated with CTX at T0 and T24. Its level decreased significantly between T0 and T3, but stabilized after T3. Weight loss reduces leptin secretion, suggesting that this protein may have a paracrine or endocrine effect on bone remodeling [42, 46]. Leptin seems to behave contrary to adiponectin as it correlates directly with weight [29, 47, 48] and inversely with bone resorption markers, such as NTX [38, 42, 49].

In addition to adipokines, other proteins seem to be involved in the pathophysiological mechanism associated with bariatric surgery-induced bone changes. In bone remodeling, the balance between bone formation and resorption is sustained by the RANK/RANKL/OPG axis [50], which regulates osteoclastogenesis and is responsible for bone resorption or osteoclast apoptosis, stimulating bone formation [51]. The adipose tissue may unbalance the RANKL/RANK/OPG axis and consequently, change bone mass. The significant increase in RANKL in the study sample associated with OPG decrease after surgery, as well as the inverse relationship between CTX and OPG at T3 suggest higher osteoclastic activity and early bone resorption after RYGB. These findings confirm other studies that suggest a strong correlation between weight loss and body composition changes as bone resorption increases and BMD decreases [13, 52, 53]. Studies on RANKL and OPG behavior after RYGB in humans were not found.

Finally, complementing the hypothesis of the relationship between adipose tissue loss and higher bone resorption, researchers have proposed that skeletal load may be an efficient means of avoiding bone loss and increasing the formation of new bone [54]. Thus, load changes may induce a compensatory increase in bone remodeling [55] mediated by osteocytes and sclerostin and inhibit the Wnt/β-catenin pathway [23–25, 56]. When the body is not properly exercised or the skeletal load decreases very quickly due to the intense and rapid weight loss promoted by bariatric surgery, bone turnover increases, as indicated by bone resorption markers, which are sometimes associated with lower BMD and higher risk of fractures [6, 57]. The endogenous inhibitor of the Wnt pathway, sclerostin, is considered a key regulator of bone homeostasis, and it can be used as a new marker for bone resorption assessment [20, 58]. This protein increases in individuals who join a weight loss program after they lose at least 10% of their body weight [38, 59–61]. Nevertheless, its role after bariatric surgery is controversial. The increase in serum sclerostin provides important information about the continuous loss of BMD after surgery [19, 20]. Nonetheless, Grethen et al. (2012) did not find serum sclerostin changes in obese women submitted to RYGB and normal-weight women [19]. Sclerostin increased significantly in the study sample between T0 and T3, but had decreased by T12. This protein was negatively correlated with total body weight before surgery and with FM 3 months after RYGB, which could be partly explained by the mechanostat theory. Therefore, while preoperative excess weight would be associated with low sclerostin, protecting against bone resorption, the rapid initial weight loss, especially the loss of body fat induced by bariatric surgery, could increase the protein level in the first months after RYGB. Thus, lower sclerostin in the long-term follow-ups could be related, among others, to weight loss deceleration during this phase.

Some of the study limitations include its observational character and sample size. Despite these limitations, this study is clinically important because it established that FM and FFM loss after RYGB can be associated with changes in proteins intrinsically related to bone tissue metabolism, namely adiponectin, leptin, RANKL, OPG, and sclerostin, which can be involved in the mechanisms associated with higher bone remodeling after surgery.

Conclusion

Bariatric surgery remains the most effective treatment for severely obese patients, although the potential effects on bone health are only partially understood. There is a complex relationship between weight loss and biochemical, hormonal, and body composition changes induced by RYGB, with many peripheral and central mediators potentially involved in the regulation of bone homeostasis.

Changes of sclerostin, CTX, and adiponectin in this study provide further important information on bone metabolism in patients after bariatric surgery. Apparently, the weight reduction and the body compartments and changes of adipokines interfere on bone remodeling. The clinical implications for osteoporosis and fracture risk are still unclear and additional data are clearly needed to characterize the clinical impact of the observed bone loss, but, in order to prevent bone loss and fractures, the health care team should start monitoring the aspects of the daily life of patients who undergo RYGB early and thoroughly.