Different impact of sleeve gastrectomy and Roux-en-Y gastric bypass on thyroid function and levothyroxine requirements

Abstract

Purpose

Bariatric surgery might impact the thyroid function test and, in hypothyroid patients, the dose of levothyroxine (LT4), but data are not univocal. We evaluated changes in thyroid function during the first year following bariatric surgery in patients without pre-existing thyroid disease, as well as adjustments in LT4 dosage in those with thyroid disorders, comparing the effects of Roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (SG).

Methods

A retrospective observational study including 69 patients on LT4 therapy (Tr) and 85 patients with normal thyroid function (Nt) at pre-surgical work-up and at regular follow-up at 45 days, 3–6 months and 1-year visits after surgery.

Results

The mean body weight reduction for the entire cohort was in 31.4 ± 0.7% 1 year after surgery. In the two Nt groups, TSH levels remained stable throughout the observation period, without differences for type of surgery and without relationship with body weight. After bariatric surgery, the patients treated with LT4, belonging to both Tr-RYGB and Tr-SG, needed to increase the dosage per kg body weight to achieve stable TSH. At 1 year, the increase of LT4/kg/die was higher in the Tr-RYGB group than Tr-SG group (0.44 vs 0.30 mcg/kg/die, p = 0.032).

Conclusion

In patients with severe obesity and normal thyroid function, TSH levels are not related with body weight change. Patients treated with LT4, after bariatric surgery need to in-crease the dosage per kg of body weight of the LT4 especially after surgery with malabsorptive component (RYGB) compared to restrictive surgery (SG).

Introduction

Obesity is a clinical condition that is increasing throughout the world, so much so that it has taken on the characteristics of a global epidemic (globesity) [1].

The relationship between thyroid function and adipose tissue has yet to be understood, with numerous adipokines and other hormonal mediators acting at various levels of the hypothalamic-pituitary-thyroid-adipose tissue axis, influencing their respective functions bidirectionally [2]. Previous studies evaluating thyroid function among human obesity have reported conflicting results: of these, several have shown slightly increased values of thyrotropin (TSH) [3] and higher level of free triiodothyronine (fT3); others, normal thyroid profile [4–6] or increased TSH and lower, but still in the normal range, fT3 and free thyroxine (fT4) levels [7]. All these hormonal changes seem to normalize with weight loss [8]. Instead, the prevalence of hypothyroidism in patients with severe obesity range from 12 to 25% [3, 9].

In recent decades, the use of bariatric surgery has increased, especially due to the demonstrated effectiveness on weight loss itself and on its associated medical problems [10]. The increased number of obesity surgeries has led to the exploration of the link between weight loss induced by bariatric surgery and changes in circulating levels of thyroid hormone [11–17].

The data collected so far in the literature are not univocal: bariatric surgery is associated with a significant decrease in TSH and a non-significant change of fT4 in the first years after surgery [18–20]. However, results of TSH and fT4 in hypothyroid patients following bariatric surgery are sparse [15, 21, 22]. Few studies compared the effects of different types of bariatric operations on thyroid function and use of levothyroxine (LT4) [15, 21, 23]. It is therefore not clear whether there is an influence on the effectiveness of LT4 therapy by bariatric surgery interventions and whether there is a difference between the effects of restrictive interventions and interventions with a malabsorptive component.

The aim of the study was to evaluate the impact of two types of bariatric surgery tech-niques, Roux-en Y Gastric Bypass (RYGB), a restrictive operation with malabsorptive component, and Sleeve Gastrectomy (SG), a purely restrictive surgery, on thyroid hormones levels in patients with normal thyroid function and in patients with thyroid diseases. Furthermore, the study explored the effect of these two types of surgery, SG and RYGB, on LT4 dosage and its relationship with the body weight change.

Materials and methods

This is a retrospective study of patients submitted to bariatric surgery in a single tertiary referral center. Pre-surgery data was collected from medical records of hospitalization and follow-up data was prospectively collected and retrieved from outpatient medical records at the visit of 45 days (1st control), 3–6 months (2nd control) and 1 year after surgery (3rd control). The following parameters were considered: 1) anthropometric data (age, sex, height, body weight, Body Mass Index -BMI-); 2) LT4 dosage expressed as µg/week, µg/day and µg/Kg/day; 3) plasma indices of thyroid function (TSH, fT3 and fT4). The patients considered were divided into 4 groups based on the type of intervention (RYGB or SG) and whether there was a pre-existing indication (Nt, normal thyroid function or Tr, LT4 treated) to take hormone therapy with LT4 (tablet formulation), regardless of the underlying thyroid disease. The 4 groups (Tr-RYGB, Tr-SG, Nt-RYGB, Nt-SG) were compared for sex, age and BMI. The Investigator sought for written informed consent to the data analysis from all subjects. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the local Ethical Committee (ID 22077, Prot.BarLevo/Cam01/2018). The manuscript has been prepared according to STROBE guidelines. Serum free thyroxine (fT4, normal range 0.7–1.7 ng/dL), free triiodothyronine (fT3, normal range 2.7–5.7 pg/mL), and serum TSH (normal range, 0.4–4 μUI/mL) were determined in all samples by electrochemiluminescence immunoassay (Roche Corporation, Indianapolis, IN, USA).

Descriptive data are reported as means ± standard error. Parametric (ANOVA and ANOVA for repeated measure) and non-parametric tests (Wilcoxon and Mann–Whitney) were used for the comparison between and within groups, and Pearson correlation for bivariate correlations. A p-value of less than 0.05 was regarded as significant. Data management and analysis were performed with SPSS 27.0 statistical software (^®IBM) and StatView5.

Results

From a total cohort of 236 initially screened patients, we included 154 patients, 28 (18%) males and 126 (82%) females in our study for whom we could retrieve data for all three follow-up visits. The mean age of the included patients was 51 ± 1.2, the patients had a mean preoperative weight of 123 ± 1.85 kg and BMI of 45 ± 0.6 kg/m2. Of these, 91(59%) underwent to RYGB and 63 (41%) to SG. Sixty-nine (44%) were already on LT4 therapy before the surgery, (43 RYGB (Tr-RYGB) and 26 SG (Tr-SG); 85 patients had no history of thyroid dysfunction and LT4 treatment (48 RYGB (Nt- RYGB) and 37 SG (Nt-SG). There was no statistically significant difference among baseline characteristics between groups (Table 1).

Table 1.

Baseline characteristics of studied population. Data are mean ± SE

	Normal thyroid function			LT4 Treated
	Nt-RYGB	p*	Nt-SG	Tr-RYGB	p#	Tr-SG
n	48		37	43		26
Age	51 ± 3	ns	47 ± 2	51 ± 1	0.05	55 ± 2
Body weight (Kg)	123 ± 2	ns	129 ± 5	123 ± 2	ns	119 ± 4
BMI (Kg/m2)	44 ± 0.7	ns	45 ± 1	46 ± 0.9	ns	44 ± 2
TSH (µU/mL)	1.9 ± 0.9	ns	1.9 ± 0.15	2.1 ± 0.3	ns	2.7 ± 0.45
fT3 (pg/mL)	3.6 ± 0.09	ns	3.8 ± 0.13	3.5 ± 0.1	ns	3.7 ± 0.1
fT4 (ng/dL)	1.0 ± 0.2	ns	1.2 ± 0.3	1.2 ± 0.07	ns	1.13 ± 0.04

^*p were calculated using Mann Whitney U test

The group of patients treated with LT4 (Tr) and the one with normal thyroid function (Nt) showed a significant weight reduction (ANOVA, p = < 0.001) at the end of the observation period (1 year) equal to a mean 31.4 ± 0.7% of the initial body weight with a comparable trend over time without significant difference by type of surgical treatment, except, at 3–6 months and at 12 months, between Nt-RYGB and Tr-SG, and at 12 months, between Tr-RYGB and Tr-SG (after Bonferroni adjustment) (Fig. 1, Table 2).

Fig.1.

Percentage of total body weight loss (%TBWL) at 45 days, 3–6 months, 1 year after surgery in the four group of patients. Data are mean ± SEM

Table 2.

Percentage of total bodyweight loss (TBWL) at 45 days, 3–6 months and 12 months

	Total body weight loss (%)1
	45 days	3–6 months	12 months
Nt-RYGB	16.2 ± 1.9	31.47 ± 2.8#	33.1 ± 1.1°
Nt-SG	12.8 ± 0.7	25.97 ± 2.3	31.1 ± 1.3
Tr-RYGB	13.1 ± 0.9	24.91 ± 1.2	32.9 ± 1.4*
Tr-SG	13.1 ± 0.9	20.97 ± 1.1#	26.4 ± 1.5*°

¹Mean ± SEM; ^#p = 0.007 Nt-RYGB vs Tr-SG; °p = 0.003 Nt-RYGB vs Tr-SG; * p = 0.005 Tr-RYGB vs Tr-SG by one-way ANOVA with Bonferroni correction for multiple comparisons

Normal thyroid function group (Nt)

From the analysis of the data of the group of subjects with normal thyroid function, Nt-RYGB and Nt-SG, it emerges that 1 year after surgery the TSH levels are substantially stable in both intervention subsets (p = ns for both groups). Only a transient significant increase in TSH values, was observed at 2nd control (3–6 months) in the Nt-RYGB group, compared to baseline values (p = 0.029) (Fig. 2a, Table 3). FreeT3 levels showed a remarkable reduction compared to basal period, in all follow-ups, in the Nt-RYGB, (basal vs 45 days p = < 0.001; vs 3–6 months p = 0.009; vs 1 year p = 0.013), but not in Nt-SG (p = ns for all follow-ups). Furthermore, we observed a significant difference in fT3 levels between the two types of surgery (Nt-RYGB vs Nt-SG) only at 45 days (p = 0.043) (Fig. 2b, Table 3). FreeT4 levels were significantly higher at 1st follow-up only in Nt-RYGB compared to basal time (basal vs 45 days p = 0.024), and no significant difference between groups were detected 1 year after surgery (p = ns) (Fig. 2c, Table 3). Considering the body weight and hormone levels recorded before surgery and at each follow-up time we observed a direct and significant relationship between fT3 levels and body weight in both Nt-RYGB and Nt-SG groups (r = 0.311, p < 0.001; r = 0.337, p < 0.001, respectively). No relationship was observed between TSH and body weight in both groups of surgery (p = ns for both groups).

Fig.2.

Temporal pattern of hormones in Nt-RYGB and Nt-SG patients a TSH, b fT3, and c fT4. Data are mean ± SEM. * p < 0.05, p values were calculated vs. baseline by Wilcoxon signed rank. # p < 0.05, p values were calculated by Mann Whitney U test for difference between groups

Table 3.

Hormonal changes after bariatric surgery in “Nt” groups. Data are mean ± SE

	Nt-SG				Nt-RYGB
	Pre-surgery	45 days	3–6 months	1 year	Pre-surgery	45 days	3–6 months	1 year
TSH (µU/mL)	1.9 ± 0.20	1.9 ± 0.24	1.8 ± 0.20	1.8 ± 0.19	1.9 ± 0.17	2.3 ± 0.20	2.3 ± 0.17*	2.0 ± 0.17
fT3 (pg/mL)	3.8 ± 0.19	3.5 ± 0.21	3.4 ± 0.25	3.2 ± 0.20	3.6 ± 0.17	2.8 ± 0.18*#	3.1 ± 0.21*	3.1 ± 0.18*
fT4 (ng/dL)	1.2 ± 0.04	1.2 ± 0.05	1.2 ± 0.05	1.1 ± 0.05	1.0 ± 0.04#	1.1 ± 0.04*	1.0 ± 0.04	1.0 ± 0.04

^* p < 0.05 vs pre surgery by Wilcoxon; # p < 0.05 vs SG at the same time from surgery by Mann Whitney U test

LT4 treated group (Tr)

In the group of patients treated with LT4 and undergone SG (Tr-SG), TSH levels showed, compared to the pre-surgery value, a statistically significant reduction only at 3–6 months follow-up (basal vs 45 days p = ns; vs 3–6 months p = 0.023; vs 1 year p = ns). Between the two groups (Tr-RYGB vs Tr-SG), however, there were not statistically significant difference in TSH values at all controls (Fig. 3a, Table 4). A reduction of fT3 after surgery was observed in Tr-SG, without reaching statistically significance comparing basal value vs each follow-up time (basal vs 45 days; vs 3–6 months; vs 1 year, for all p = ns). On the contrary, the reduction in fT3 observed in the Tr-RYGB, was statically significant at all control compared to basal value (basal vs 45 days p = 0.001; vs 3–6 months p < 0.001; vs 1 year p < 0.001). Between the two groups, at 12 months, a statistically significant difference in terms of fT3 values was observed (p = 0.045) (Fig. 3b, Table 4). fT4 showed an increase with significant difference between groups at 45 days (p = 0.017) and at 3–6 months of follow-up (p = < 0.001). Compared to basal fT4 values, significant difference emerges for patients underwent SG (Tr-SG) at all follow-ups (basal vs 45 days p = 0.009; vs 3–6 months p = < 0.001; vs 1 year p = 0.022) (Fig. 3c, Table 4).

Fig.3.

Temporal pattern of a TSH, b fT3(b), and c in Tr-RYGB and Tr-SG patients. Data are mean ± SEM. * p < 0.05, p values were calculated vs. baseline by Wilcoxon signed rank. # p < 0.05, p values were calculated by Mann Whitney U test for difference between groups

Table 4.

Hormonal changes after bariatric surgery in “Tr” groups. Data are mean ± SE

	Tr-SG				Tr-RYGB
	Pre-surgery	45 days	3–6 months	1 year	Pre-surgery	45 days	3–6 months	1 year
TSH (µU/mL)	2.7 ± 0.46	2.5 ± 0.88	1.3 ± 0.49*	1.9 ± 0.34	2.1 ± 0.37	2.3 ± 0.71	1.6 ± 0.39	1.7 ± 0.27
fT3 (pg/mL)	3.7 ± 0.11	3.3 ± 0.25	3.4 ± 0.31	3.3 ± 0.31	3.5 ± 0.14	2.7 ± 0.18*	2.7 ± 0.14*	2.6 ± 0.16*#
fT4 (ng/dL)	1.13 ± 0.13	1.41 ± 1.01*	1.63 ± 0.73*	1.29 ± 0.56*	1.2 ± 0.10	2.7 ± 0.82#	2.1 ± 0.59#	1.9 ± 0.45
LT4 (µg/week)	722 ± 65	746 ± 69	711 ± 80	697 ± 83*	777 ± 45	789 ± 47	849 ± 64#	850 ± 66#

^* p < 0.05 vs pre surgery by Wilcoxon; # p < 0.05 vs SG at the same time from surgery by Mann Whitney U test

During the follow-up, no significant change was observed in the weekly total dosage of LT4 in Tr-RYGB (basal vs 45 days, 3–6 months, 1 year, p = ns for all), but there was a reduction (and statistically significant) at 1 year in the Tr-SG compared to basal value (p = 0.034). In the comparison between the two surgeries, there was a significant difference at the last two follow-ups (45 days, p = 0.055; 3–6 months, p = 0.041; 1 year p = 0.037), with higher weekly LT4 dosage in the Tr-RYGB group (3–6 months, mean 849 ± 45 µg/week vs 711 ± 36 µg/week; 1 year, mean 850 ± 47 µg/week vs 697 ± 34 mcg/week) (Fig. 4, Table 4). When the change LT4 dosage was analyzed daily per kg of body weight (ΔLT4 µg/Kg/day), a progressive increase emerged in both groups, Tr-RYGB and Tr-SG (p < 0.001) (Fig. 5a). One year after surgery, the increase of the daily LT4 dose per kg of body weight, needed to maintain a stable TSH value, was higher in the Tr-RYGB group than in Tr-SG group (53.7 vs. 37.6%, p = 0.015; 0.44 vs 0.30 mcg/kg/die, p = 0.032) (Fig. 5b).

Fig.4.

Temporal pattern of weekly LT4 dosage in Tr-RYGB and Tr-SG patients. Data are mean ± SEM. * p < 0.05, p values were calculated vs. baseline by Wilcoxon signed rank. # p < 0.05, p values were calculated by Mann Whitney U test for difference between groups

Fig. 5.

Change in LT4 dosage per kg of body weight. Delta (Δ) LT4 was calculated as a difference between each time of follow-up and pre-surgery value. a When the LT4 dosage was adjusted per kg of body weight, a progressive increase was observed in both groups. b After 1 year, in Tr-RYGB group the daily average increase of LT4/kg was higher than in Tr-SG group. * p < 0.05, p values were calculated vs. baseline by Wilcoxon signed rank. # p < 0.05, p value was calculated by Mann Whitney U test for difference between groups

Discussion

In this study we investigated, during the first year after surgery, (a) the longitudinal change in thyroid function in euthyroid patients, and (b) the progressive change in LT4 dose using a weight-based approach- in hypothyroid patients already under treatment before surgery, comparing, RYGB with SG for both outcomes.

The relationship between thyroid hormone alterations and weight change has been investigated in several studies with controversial results. There is evidence to support a significant reduction in TSH values after bariatric surgery, regardless of weight loss [24–27], determining in some cases the resolution of the subclinical hypothyroidism that often accompanies patients with severe obesity [9]. The reason for this decrease has been attributed to various mechanisms, including a reduction in the levels of the hormone leptin: in fact, leptin, produced by adipocytes, would be able to stimulate the secretion of TSH and its decrease, caused by the loss of fat mass, would cause the cessation of this stimulus [3, 11, 28]. Another hormone capable of stimulating TSH secretion and whose levels are reduced after bariatric surgery is ghrelin. In fact, ghrelin is produced by the cells of the gastric fundus, and it sees a significant reduction in its production especially in operations that involve the removal or bypass of this portion of the stomach, just as in the case of SG and RYGB [29]. In our study, on the contrary, we observed that, 1 year after bariatric surgery (SG or RYGB), TSH values do not change substantially in patients who were already euthyroid before surgery, except for transient variation during follow-up. This observation suggests that not in all patients with severe obesity TSH values are influenced by changes in weight and adipose tissue mass, especially in absence of underlying thyroid diseases, as emerged in previous investigations [4, 5]. In line with our study, other authors reported no change in TSH levels after bariatric surgery also obtained with several techniques in euthyroid patients [30–32]. A previous study also observed a reduction in 24-h TSH levels after biliopancreatic diversion, but unchanged fasting TSH levels compared to pre surgery [11]. These contrasting results might be due to the differences in subject populations, sample sizes, and surgical procedures. In the context of obesity, an enhanced peripheral conversion of fT4 to fT3 has been described and it has been interpreted as a compensatory mechanism to limit fat gain, improving energy expenditure through fT3-mediated mechanisms, such as thermogenesis [33]. Therefore, the significant correlation of fT3 and body weight detected in both Nt-RYGB and Nt-SG and the significant reduction of fT3 levels following the surgery, mirrored by a transient increase of TSH, could be explained by the reversion of the afore-mentioned phenomenon. Moreover, in our study, the fT3 reduction after surgery is particularly significant in the RYGB subgroup, where the malnutrition and malabsorption might play a more relevant role [34]. Considering our results as a whole, another possible explanation of the reduction in fT3 without a parallel increase of TSH could suggest an altered response of hypothalamus-hypophysis to the decline in fT3, perhaps due to site-specific changes in D2 activity, more specifically increased hypothalamic D2 activity during caloric restriction [35, 36]. The stability of TSH over the time was also observed, after surgery, in patients treated with LT4 with no significant difference in the total weekly dosage compared to the baseline. Nevertheless, considering the weight loss post-surgery, this translated in a progressive increase in the LT4 requirement per kg of body weight.

This last data could be explained by the subversion of the normal anatomy of the first gastrointestinal tract, which occurs after bariatric surgery and which causes different degrees of malabsorption affecting also the bioavailability of drugs. This is especially frequent after RYGB [37], as already suggested from pioneering studies on the subject [38, 39] and confirmed more recently [40, 41]. In this regard, a previous study showed a non-statistically significant increase in TSH values in hypothyroid patients submitted to various types of surgeries (RYGB and Biliopancreatic diversion (BPD)) which successfully required the switch- at the same dosage—of the LT4 tablet formulation to the liquid one [41]. In fact, the LT4 liquid formulation guarantees better bioavailability of the drug even in conditions of altered gastric acidity, including those secondary to bariatric surgery. Many of the studies have highlighted a significant reduction in the total dose of LT4 in patients who have undergone bariatric surgery1 [7, 41], as also confirmed synthetically by a recent meta-analysis [24] and this seems to be related to a decrease of the lean body mass [43]. Again, apparently, our results are in contrast with these data since they indicate instead a progressive and statistically significant increase of the LT4 dose both after SG and after RYGB. However, this discrepancy can be explained by the fact that in our study the dose of LT4 was evaluated “per kg” and not as “total amount” -per day or per week. According to our results, Trimboli et al. recently showed in a small cohort of patients with obesity underwent to bariatric surgery, that the total daily dose of LT4 did not change significantly but, when considered per weight, it increased significantly especially after RYGB [44]. Similarly, also other studies demonstrated a decrease in the total weekly dose of LT4, 6 and 12 months after surgery, that turned into an increment when adjusted for the body weight [45, 46]. In fact, when calculated on a “total dose” basis, we showed substantial stability of dosage compared to pre-surgery time in our cohort too (Fig. 4). Based on this evidence, we suggest quantifying the LT4 dose in patients with obesity undergoing bariatric surgery, according to body weight rather than as an absolute dose, as this may better reflect the true dose requirement and its changes over time. This is clinically relevant, especially in such subset of patients who are going to face a sure change in their body weight in the first 12–24 months after surgery, requiring close monitoring of thyroid function and therapy [43]. It should be underlined that the mal-absorption of LT4 after bariatric surgery, do not derive only from the altered gastric anatomy and acidic environment, necessary for the dissolution of the tablets, but it is also favored by the frequent concomitant intake of proton pump inhibitors (PPIs), which hinder the gastric pH, and of vitamin and mineral supplements (ie calcium and iron), that form insoluble complexes with L-T4, less absorbable by the intestinal mucosa [42, 47]. The main limitations of the study are represented by the retrospective nature of the analysis, even if the follow-up data were collected prospectively, and by the single center source of data with a limited sample size. On the other hand, most of the studies on the topic have generally involved a small number of patients, while our sample emerges as among the largest ones reported so far in the literature. Furthermore, the study considers a short follow-up period (12 months), which however seems to be suggestive to evaluate the changes in serum TSH levels and LT4 dose even if it does not consider direct measurement of the drug absorption. Finally, there were no data collected on the effective lean body mass lost.

Conclusions

In patients with severe obesity with normal thyroid function, TSH levels are not influenced by bariatric surgery. In hypothyroid patients receiving LT4 prior to bariatric surgery, the drug dose calculated per kilogram of body weight resulted increased following both SG and RYGB. This increase is more pronounced after RYGB, likely reflecting the malabsorptive component of the procedure. Taken together, these findings suggest the need for close follow-up of thyroid function in patients with thyroid disease undergoing bariatric surgery to ensure appropriate and timely adjustment. However, further prospective studies based on larger samples are necessary to confirm these results.

Author contributions

Conceptualization, S.C.; methodology, M.P., C.M., B.A. and R.B.; formal analysis, A.P., S.C. and G.R..; writing—original draft preparation, A.P.; writing—review and editing, S.C. and A.A. All authors have read and agreed to the published version of the manuscript.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data availability

The data presented in this study are available on request from the corresponding author.

Declarations

Ethics approval and consent to participate

All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. For this type of study formal consent is not required.

Competing interests

The authors declare no competing interests.