Central Resistance to Thyroid Hormones in Morbidly Obese Subjects Is Reversed after Bariatric Surgery-Induced Weight Loss

Endocrine abnormalities are common in obesity, including altered thyroid function. The altered thyroid function of obesity may be due to a mild acquired resistance to the thyroid hormone. The aim of this study was to investigate the effect of weight loss after bariatric surgery (BS) on resistance to thyroid hormones in patients with extreme obesity compared with a control group. We performed an observational study evaluating patients with extreme obesity who underwent BS. We included 106 patients (83 women) and 38 controls (24 women). The primary endpoint was the thyrotroph thyroxine resistance index (TT4RI) and thyroid stimulating hormone (TSH) index (TSHRI). The parameters were studied before and after surgery. TSHRI and TT4RI were higher in the obese patients than in the control group. TT4RI and TSHI decreased significantly over time after surgery, with this decrease being associated with the excessive body mass index (BMI) loss and C-reactive protein (CRP). In extreme obesity, BS promotes a significant decrease in the increased TT4RI and TSHI. This decrease of TT4RI and TSHI is progressive over time after BS and significantly associated with excess BMI lost and CRP. Extreme obesity is characterized by a mild reversible central resistance to thyroid hormones.


Introduction
Obesity has reached epidemic proportions. Since 1980, the prevalence of obesity has continuously increased in most countries. The disease burden related to high BMI has increased since 1990 [1]. The prevalence of obesity in the United States in 2013-2014 was 35% among men and 40.4% among code number: 2014/135), and written informed consent was obtained from all patients and controls. All of the studies were conducted in accordance with the Declaration of Helsinki. We included a total of 144 patients and controls (107 women) in our study, of which there were 106 patients (83 women) and 38 controls (24 women) selected from a pool of volunteers available to our unit. None of the controls had diabetes or other medical problems, nor were they taking any type of medication.

Assays and Other Methods
Serum TSH (mIU/L) was measured by a two-site chemiluminescent immunoassay (ADVIA Centaur, Siemens, Deerfield, IL, USA) with a sensitivity of 0.01 mIU/L and with intra-assay coefficients of variation of 2.5%, 2.4%, and 2.4% for low, medium, and high serum TSH levels, respectively; and with inter-assay coefficients of variation of 5.3%, 3.4%, and 2.1% for low, medium, and high TSH levels, respectively; as previously published [19]. Serum FT4 (ng/dL) was measured by a direct chemiluminescent immunoassay (ADVIA Centaur, Siemens) with a sensitivity of 0.1 ng/dL and with intra-assay coefficients of variation of 3.3%, 2.2%, and 2.5% for low, medium, and high plasma FT4 levels, respectively; and with inter-assay coefficients of variation of 2.5%, 4.0%, and 2.3% for low, medium, and high FT4 levels, respectively; as previously published [19]. Insulin (µU/mL) was determined with a chemiluminescent immunometric assay (Immulite 2000 Insulin, DPC, Los Angeles, CA, USA) and with an intra-assay CV of 5.5%, 3.3%, and 3.7% for low, medium, and high plasma insulin levels, respectively; and with an inter-assay CV of 7.3%, 4.1%, and 5.3% for low, medium, and high insulin levels, respectively. Serum GH (µg/L) and IGF-1 (ng/mL) were determined by a chemiluminescent assay (Immulite, EURO/DPC) as previously published [19]. Plasma glucose (mg/dL) was measured with an automatic glucose oxidase method (Roche Diagnostics, Mannheim, Germany). All samples from a given subject were analyzed in the same assay run.

Statistical Analysis
Descriptive analysis was used in order to determine the characteristics of both the patients and the control group. Continuous data are expressed as mean ± standard error (SE) and/or median and interquartile range (IR). Non-numerical variables are expressed as frequencies and percentages.
The control subjects and obese patients were compared using the Mann-Whitney test for quantitative parameters and the Chi-squared and Fisher's exact tests for qualitative parameters. The Wilcoxon signed-rank test was used to compare the preoperative and post-surgical values in the obese patients. Data Availability: All of the data analyzed during this study are included in this manuscript as well as in previously published articles cited in the references. If any information is required, please contact the corresponding author.
To assess the overall association of TSHRI and FT4RI with anthropometric, biochemical, and hormonal data during the post-surgery period, repeated measures correlation was calculated [28].
Generalized estimating equations (GEE) models, with the autoregressive correlation structure, were used to evaluate the trajectory of postoperative TSHRI and FT4RI, as well as to determine factors associated with their changes after BS. Bivariate GEE models were used in order to observe the effect of each variable on the change in TSHRI or FT4RI, taking time into account. Only those features significantly associated with TSHRI/FT4RI values in the bivariate analysis were finally included in a multivariate GEE model.
In the multivariate analysis, statistical analyses were performed with SPSS, version 24.0, and R, version 3.5.1., with the packages geepack and rmcorr added. p-values < 0.05 were considered as statistically significant.

Preoperative Characteristic of the Obese Patients and the Control Group
Preoperative characteristic of the obese patients and the control group are presented in Table 1. Among the 106 patients studied, 83 were women and the mean age was 47.5 ± 0.9 years ( Table 1). The surgical procedures performed were RYGB and SG. Among the 38 controls studied, 24 were women and the mean age was 45.0 ± 1.5 years ( Table 1). The two groups had similar sex and age as designed by the matching criteria.

Fasting Serum Levels
Fasting glucose, lipids, CRP, and hormones (mean ± SE, Median (interquartile ranges)) are presented in Table 2. TSHRI levels were higher in the obese group than in controls; 3.5 ± 0.1 vs. 2.5 ± 0.1 for the obese and control groups, respectively. TT4RI levels were higher in the obese group than in the controls; 66.0 ± 5.5 vs. 28.9 ± 2.2 for the obese and control groups, respectively. Fasting C-reactive protein levels were higher in the obese group than in the healthy controls; 0.9 ± 0.1 vs. 0.2 ± 0.1 for the obese and control groups, respectively.

Evolution over Time of the Clinical and Analytical Parameters
The mean EBMIL as a percentage 12 months after bariatric surgery was 71.6 ± 2.4%. The biochemical and hormonal parameters in obese patients before and 12 months after BS are presented in Table 3. TSHRI levels significantly decreased in the obese patients after BS weight loss; 3.5 ± 0.1 vs. 2.5 ± 0.1 for the obese patients before and 12 months after surgery, respectively. TT4RI levels significantly decreased in the obese patients after BS weight loss; 66.0 ± 5.5 vs. 30.9 ± 2.3 for the obese patients before and 12 months after surgery, respectively.   Figure 1 shows the TT4RI and TSHRI values (Median (IR)) in the control subjects and obese patients before and 12 months after surgery. TT4RI and TSHRI levels were higher in the obese group than in the healthy controls. TT4RI and TSHRI levels significantly decreased in the obese patients 12 months after surgery-induced weight loss.  Figure 2 shows the evolution over time of TT4RI and TSHRI (Median (IR)) before and after surgery (0, 1, 3, 6, and 12 months) in obese patients. The results show a decrease of TT4RI and TSHRI. The indices of central resistance to TH (TT4RI and TSHRI) significantly decreased in the obese patients at 3, 6, or 12 months after BS, when compared with presurgical values. TT4RI and TSHRI values were similar when compared with 3, 6, and 12 months after BS. Post-surgery changes in TSHRI and TT4RI values in relation to changes in selected anthropometric, biochemical, and hormonal data are presented in Figures 3 and 4. Changes in both indices were significantly negatively correlated with EBMIL (within-subjects repeated measures correlation rm = −0.64 and rm = −0.52, respectively) and significantly positively correlated with changes in CRP values (rm = 0.34 and rm = 0.48, respectively), but not correlated with changes in GH, IGF-1, fasting glucose, and HOMA-IR. Similar results were obtained from bivariate GEE models.  After adjusting for EBMIL and CRP in a multivariate model, both variables were significantly and independently associated with TSHRI and TT4RI values in the follow-up, with a higher excess weight loss associated with lower thyroid resistance indices, whilst higher C-reactive protein values were associated with higher values in both indices (Table 4).

Discussion
The main result of this study is that we have found that the indices of central resistance to TH are increased in patients with extreme obesity, and that weight loss after BS leads to a decrease in these indices through direct or indirect mechanisms. The decrease in the indices of central resistance to thyroid hormones is significantly and independently associated with excess BMI lost and CRP. The present study strongly suggests that obesity is a situation of mildly increased pituitary resistance to TH that is reversed with BS. As far as we are aware, this is the first time that indices of resistance to thyroid hormone have been evaluated in morbidly obese subjects before and after lifestyle intervention, pharmacotherapy, or surgery-induced weight loss.
Different circulating TH levels have been found in obesity [12], although the majority of the studies have only been carried out on TSH levels. In agreement with our results, most studies have found increased TSH levels in patients with morbid obesity. Rotondi et al. [13] reported increased TSH levels in morbidly obese patients. Reinehr et al. [14] have found elevated TSH levels in obese children. Increased FT4 in obesity has been found in some studies [14]. Several studies have reported different results regarding the variation of TSH after BS and the relation of TSH decrease with weight loss [15][16][17][18]. Most [15,17], but not all [18] studies have found a decrease in circulating TSH after the operation. In agreement with our results, Guan et al. [15] found that bariatric surgery was associated with a decrease in circulating TSH levels. Neves et al. [17] found that BS promotes a significant decrease of circulating TSH. In obese children, weight loss was associated with a decrease in TSH, thyroid volume, and structure, while FT4 remained unchanged. These data suggest that the alterations of thyroid function and structure in patients with obesity are common and reversible [29]. Elsewhere, Dall'Asta et al. [18] evaluated obese subjects with normal thyroid function before and after weight loss through BS and found that TSH levels did not change. In the meta-analysis of Guan et al., although bariatric surgery was associated with a significant decrease in TSH, FT4 was not significantly changed postoperatively [15]. In accordance with these data, a recent study has found that TSH levels decreased in parallel with decreased BMI after bariatric surgery. However, no significant change was observed in FT4 or FT3 levels [30]. The differences between studies may be due to the type of bariatric surgery performed, the characteristics of the control group and the studied patients, or to the statistical power of the studies. As far as we are aware the indices of central resistance to thyroid hormones have not been studied before and after bariatric surgery.
The mechanism and the clinical implications of TSH and FT4 elevation in obese patients is still not fully understood. Our results are in agreement with the findings of Laclaustra et al. that found evidence of the association between indices measuring resistance to TH and the prevalence of obesity and diabetes in the US population [27]. TH signaling is unique for each cell (tissue or organ), depending on circulating thyroid hormone levels and on the exclusive blend of membrane transporters, deiodinases, and thyroid receptors present in each cell [31]. The resistance to thyroid hormone indices measures central sensitivity and the degree of pituitary gland inhibition by thyroxine. Peripheral resistance could also be present because, despite having a higher thyroxine, the studied group has morbid obesity, a phenotype that is physiopathologically associated with hypothyroidism [27]. Moreover, there are several animal models of reduced TH signaling in the presence of obesity [31,32]. TH receptors are less expressed on adipocytes of obese vs. lean individuals [33]. In addition, the TH receptor β has been found to be inversely correlated with disease severity in liver biopsies from patients that underwent bariatric surgery with different stages of obesity-induced nonalcoholic steatohepatitis [34]. Moreover, reduced sensitivity to TH could be epigenetically modified [35]. The reduced TH receptor expression may induce a decrease in hormone action, thereby increasing plasma TSH and constituting a condition of peripheral thyroid hormone resistance [33]. In the present study, we found increased indices of central resistance to TH in obese subjects which decreased after weight loss in accordance with the hypothesis that the increase of TSH and FT4 may represent a compensatory activation of the thyroid axis [12].
The mechanisms underlying the reason why the indices of central resistance to TH decrease after bariatric surgery remain unknown. We have found that the decrease in the indices of central resistance to TH is significantly associated with excess BMI lost after BS. This decrease is not due to an intrinsic effect of BS. A decrease in TSH has been found in obese patients after non-surgical induced weight loss [36] and the decrease in circulating TSH has been found to be associated with excess body weight loss after BS [17,19]. The present results show a decrease in the indices of central resistance to TH that is statistically significant 3 months after BS, and there was no further significant decrement, maybe due in part to the faster excess BMI loss in the first months after BS. In agreement with these data, the reduced TH receptor is reversed by weight loss, improving the thyroid hormone sensitivity [33]. One of the plausible explanations is the decrease of leptin levels following bariatric surgery [37]. With a decreasing amount of fat, the decreasing leptin [37] could promote a decrease of circulating TSH levels. However, leptin treatment was not associated with changes in thyroid size or nodularity or function [38]. Another potential mechanism may be insulin resistance: patients with extreme insulin resistance have a high prevalence of thyroid nodules, and patients with a homozygous insulin receptor mutation have significantly enlarged thyroid glands [38]. In the present study we could not find a significant influence of the insulin resistance indices on the central resistance to TH indices. Another potential mechanism may be HbA1c [39], however in the present article we could not find a significant influence of the HbA1c value on the resistance to TH indices. There is an important relationship between the GH-IGF-1 axis and obesity [8][9][10]40], and circulating GH modulates the thyroid axis. Circulating GH levels increased serum FT3 and decreased serum FT4 in humans [41]. Treatment with GH in adults with GH deficiency causes variable changes in thyroid function, the most consistent effect being decreased circulating thyroxine levels [42]. GH replacement therapy in a large group of adult GH-deficient patients has been shown to induce a significant reduction in FT4 [43]. The decreased GH secretion of obesity could be a contributory factor to the increased circulating FT4 levels. We explored the correlation between the GH-IGF-1 axis and the indices of central resistance to thyroid hormones, and were unable to find any important correlation. By contrast, we found that the variation of the increased indices of inflammation such as CRP highly correlate with the thyroid hormone resistance index in obese subjects. In agreement with these data, laparoscopic sleeve gastrectomy promotes TSH reduction in patients with morbid obesity, which correlates with an improved inflammatory state after surgery [44]. In any event, our results show that the decrease of the indices of central resistance to TH are associated with body weight loss after BS and CRP, suggesting that the decrease in the indices of central resistance to TH is mainly weight-mediated and probably due in part to the improved inflammatory state after surgery. Our results also suggest that the modification of thyroid function could be a potential therapeutic aim in patients undergoing BS [45].
We must acknowledge a number of limitations of our study. First, the relatively small sample size of our control subjects or obese patients, which did not allow for the stratification of different subgroups in the analysis. Secondly, we did not consider certain variables that could influence the study, such as the concomitant use of other medications. Thirdly, the heterogeneous nature of our patient group, with different obesity-associated comorbidities, secondary hormonal changes, and treatments, which could interfere with the thyroid hormone axis. However, there are several strengths to our study. We included sex-and age-matched controls in order to reduce the chances of misclassifying individuals due to variability in these variables. We evaluated TT4RI and TSHRI at different time points after BS, as most studies evaluated the variation of TSH alone using only two moments (before and after surgery). Moreover, this is the first study to evaluate the indices of central resistance to TH before and after BS. Nevertheless, more studies are required in order to better understand the complex relation between thyroid dysfunction and obesity [46].

Conclusions
In conclusion, this study shows that in patients with extreme obesity, weight loss induced with BS through direct or indirect mechanisms brings about a decline in the increased indices of central resistance to thyroid hormones. This decrease in central resistance to thyroid hormones is progressive over time after BS, and significantly and independently associated with BMI loss and CRP. Extreme obesity is characterized by a mild reversible central resistance to thyroid hormones.

Conflicts of Interest:
The authors declare no conflict of interest.  TH  thyroid hormone  TSH  thyroid stimulating hormone  TSHRI  TSH index  T4  thyroxine  TT4RI thyrotroph thyroxine resistance index