Introduction

Roux-en-Y gastric bypass (RYGB) has been the most efficient method to help morbidity obesity patients to lose weight [1, 2]. The success of the surgery is due to the rapid weight loss that is a consequence of diet restriction, gastric reduction, duodenal exclusion, and fast delivery of nutrients to the ileum, which, combined, contribute to the metabolic improvement of the patient [3, 4]. This surgical method results in weight reduction along with improvement in the quality of life and most associated comorbidities, such as diabetes, systemic arterial hypertension, and cardiovascular dysfunction [5].

Most studies have shown that the metabolic improvement is due to the secretions of some incretins and these can be of particular relevance for the antidiabetic effect of the RYGB [6,7,8]. An important incretin that acts in insulin secretion is glucagon-like peptide-1 (7–36) amide (GLP-1), a peptide secreted by L-cells from the distal ileum after proglucagon cleavage. This peptide, in response to food intake, potentiates insulin glucose-stimulated secretion, inhibits the action of glucagon, and delays gastric emptying [9,10,11]. Studies have also shown that GLP-1 is responsible for increased insulin secretion as a result of food intake [12, 13].

Additionally, GLP-1 is a gut-derived hormone usually associated with the improvement of blood glucose after RYGB, since it has been observed that high levels of GLP-1 are secreted from patients after RYGB surgery [14, 15]. Several studies have shown this secretion of GLP-1 in short-term RYGB follow-ups [16,17,18], with most of them being 2–5 years long. Other studies that did 10-year follow-up of RYGB did not keep track of the same patients [19, 20].

Therefore, this article aims to show changes during 10-year follow-up of RYGB from the same patients during 4 different periods, focusing glucose metabolism, GLP-1 secretion, insulin secretion, and loss or regain of weight.

Materials and methods

Study design and participants

This study was done at the Metabolic Unit, Faculty of Medical Sciences, University of Campinas, through the follow-up evaluations of patients treated at the Bariatric Surgery Ambulatory beginning in August 2005 and concluded in July 2017.

The evaluations were as follows: T1—the first evaluation occurred when the patient arrived at the clinic for the first medical appointment and preparation for surgery, without previous treatment or nutritional monitoring; T2—the second occurred after the patient reached a 10% weight loss with diet, average of 5–7 months after the initial assessment; T3—9 months after the RYGB bariatric surgery; and the last evaluation T4—10 years after the RYGB bariatric surgery.

After 18 months of the RYGB surgery, the patient’s body weight was measured to calculate the percentage of excess weight loss (%EWL). The data were obtained through analysis of patient records, and the 18-month mark was chosen based on scientific reports that demonstrated that the higher amount of weight loss happens after 18 months of surgery [20, 21].

The patients were recruited at the Bariatric Surgery Ambulatory of the Faculty of Medical Sciences University of Campinas and underwent the 4 evaluations: T1, T2, T3, and T4. The data assessments were done within 3–5 months of the end of the data collection for each period.

At the Bariatric Surgery Ambulatory, all patients are screened to participate in the surgery groups. When a patient starts a group, he must follow a protocol to lose 10% of body weight so he can be ready for bariatric surgery. If a patient fails to do so, they will not undergo RYGB.

The study started in 2005, period T1, with 87 patients, but only 19 patients lost 10% body weight and went to the second evaluation (T2), before RYGB. These same 19 patients were also evaluated at period T3. This first part of the study, covering periods T1 to T3, was the research of a master’s degree thesis and was published as an article [14]. During these 3 periods, the patients had frequent visits to the Bariatric Surgery Ambulatory under the supervision of doctors, nutritionists, and psychologists.

After 10 years, period T4, these 19 patients of Bariatric Surgery Ambulatory were contacted but only 13 fit the inclusion criteria of the research. These 13 patients evaluated at T4 were not going often to the ambulatory, and they were having a regular lifestyle without any diet restrictions or use of weight loss medications.

During the T1, T2, and T3 periods, 11 morbidly obese patients with normal glucose tolerance (NGT group) and 8 patients who, besides obesity, had diabetes or glucose intolerance (abnormal glucose metabolism; AGM group) were evaluated and monitored. At T4, 9 individuals of the NGT group and 4 in the AGM group were assessed. In both groups, there were difficulties in the follow-up process, either due to patient health issues or due to lack of up-to-date information on patients. In the AGM group, one patient died of cancer and three could not retake the exams because they were using antidepressant medications that could tamper the evaluation results. Similarly, in the NGT group, three patients could not be re-evaluated since their contact information was not up to date and, therefore, the residential addresses and telephone numbers on file did not belong to these patients anymore.

The periods T1, T2, and T3 of the study had the protocol approved by the Research Ethics Committee of University of Campinas-UNICAMP, School of the Faculty of Medical Sciences (FCM-UNICAMP).

The period T4 of the study had the protocol approved by the Research Ethics Committee of University of Campinas-UNICAMP, School of the Faculty of Medical Sciences (FCM-UNICAMP) (number 31989314.3.0000.5404).

All volunteers included were previously informed about the study and signed an informed consent form at 2 different moments; the first informed consent was signed at T1 period and the second informed consent was signed at T4 period (10 years).

Patients who agreed to participate in the study were between 18 and 60 years old; BMI > 40 kg/m2 for obese and BMI ≥ 35 kg/m2 for diabetic obese. Exclusion criteria were as follows: individuals with chronic diseases such as heart disease, nephropathy, or thyroid disorder; patients on hormone replacement therapy, corticosteroid treatment, and renal or hepatic impairment. Patients who were using drugs that could interfere with metabolism such as antidepressants, corticosteroids, and diuretics were excluded.

Diabetic patients were classified according to the revised American Diabetes Association (ADA) criteria 2018 [22].

Assessments

At each of these periods, all patients that accepted to participate in the study underwent the Oral Glucose Tolerance Test (OGTT) Association (ADA) criteria 2018 [22]. The patients drank the 75 g of glucose; after surgery, they drank slowly than before, and they took almost 1 to 3 min to drink all the glucose.

During OGTT, blood samples were collected at 0, 30, 60, 90, and 120 min for quantification of plasma glucose and insulin, being considered as time 0 the final intake of 75 g of glucose. At times 0, 30, and 60 min of OGTT, samples were collected for later dosages of active GLP-1, and DPP-IV inhibitor was added to the blood samples.

All collected blood samples were centrifuged for serum, aliquoted, and stored in a freezer at − 80 °C for later dosing. All samples were dosed within a period of 3–5 months after the samples were collected.

Laboratory methods

Plasmatic glucose: Dosed in duplicate by the enzymatic method automated in the YSI 2300-Stat equipment; YSI, Yekkiw Springs, OH, USA.

Insulin: Quantified in duplicate by the enzyme-linked immunosorbent assay (ELISA) using Merck Millipore (Billerica, MA, USA) high-sensitivity human-specific kit (sensitivity: 1µU/mL).

Plasma active GLP-1: Measured by a fluorescence immunoassays method using high-sensitivity specific ELISA kit, sensitivity: 2 pM, Merck Millipore (Billerica, Ma, USA) [23,24,25].

Anthropometric measurements

Weight and height: we used a 100-g division mechanical scale for weight measurement (standardization: wearing light clothes and no shoes), and a portable stadiometer (Secca brand) with 1-mm division for height evaluation (participant without shoes and the five body points against a wall without skirting board).

Body composition assessment

Body composition was assessed by bioelectrical impedance using the Biodynamics Model 310 appliance after 4-h fasting [26]. All participants were instructed not to drink alcohol, coffee, coke, guarana, and tea (mate or black), and not to exercise the day before the test.

All participants were evaluated for anthropometric, body composition, and biochemical measurements.

Data analysis

The patients were classified according to weight loss sustenance that was calculated by the relationship between the %EWL (ideal weight based on the BMI of 25 kg/m2) at the last weight assessment and the maximum %EWL reached [1 − (last%EWL/maximum%EWL)] [20], which in fact represents the amount of weight regain of the maximum weight lost, expressed as percentages. Classifications: A: successful weight loss sustenance after the initial successful weight loss (maximum %EWL > 50%) and maintaining up to 50% of the weight lost—A1: excellent, with regain of 50% of the weight lost; A2: regular, with regain corresponding to 10–50% of the weight lost; B: unsatisfactory weight loss sustenance after the initial successful weight loss, with regain of > 50% of the weight lost. The maximum weight loss was obtained at 18 months after RYGB [20, 21]. Body weight at 18 months was obtained through analysis of patient records.

HOMA-IR was calculated as fasting glucose in mmol/L times the fasting insulin in µU/mL divided by 22.5 (fasting insulin vs. fasting glucose)/22.5 [27,28,29,30].

Statistical analysis

Descriptive statistics of numerical variables were calculated to describe the sample profile according to the variables under study, with mean values, standard deviation (SD), and standard error (SE).

To study the parameters over time, the generalized estimating equation (GEE) method was used. The estimates were calculated by maximum likelihood for treating losses over time. The data were transformed into ranks due to the lack of normal distribution. No covariate was analysed, only the comparison between the AGM X NGT groups [31,32,33,34,35,36,37].

For comparison of groups, the Mann–Whitney test was used. The significance level adopted for this study was 5% (p ˂0.05).

Software: The SAS System for Windows (Statistical Analysis System), version 9.4. SAS Institute Inc, 2002–2012, Cary, NC, USA [38].

All numeric data are given as mean ± SD, with only the graphics at Fig. 1 showing means and standard error bars in order to better illustrate the results.

Fig. 1
figure 1

OGTT curves. (A, B) Plasma glucose levels during OGTT for groups NGT and AGM respectively during periods T1, T2, T3, and T4; (C, D) Plasma insulin levels during OGTT for groups NGT and AGM respectively during periods T1, T2, T3, and T4; (E, F) Plasma insulin levels during OGTT for groups NGT and AGM respectively during periods T1, T2, T3, and T4; NGT, group normal-glucose tolerance; AGM, group abnormal-glucose metabolism. Numeric data are expressed as means ± SD and graphic data are expressed as means ± SE bars, black line and symbol (T1 period: first evaluation 11 patients NGT; 9 patients AGM), red line and symbol (T2 period: pre-surgery 11 patients NGT; 9 patients AGM), blue line and symbol (T3 period: 9 months after RYGB 11 patients NGT; 9 patients AGM), and pink line and symbol (T4 period: 10 years after RYGB 9 patients NGT; 4 patients AGM)

Full size image

Results

Groups were matched for age and gender; as shown in Table 1, NGT and AGM initially presented BMI of 46.31 ± 5.03 kg/m2 and 50.87 ± 10.31 kg/m2, respectively, and a high percentage of body fat of 44.29 ± 4.27% and 47.33 ± 5.07% for NGT and AGM, respectively. Both groups showed an increase in BMI after 10 years and remained obesity grade I with BMI for NGT 32.45 ± 4.99 kg/m2 and AGM 34.85 ± 4.46 kg/m2.

Table 1 Anthropometric characteristics
Full size table

The waist measurement comparing T3 and T4 periods showed an increase, 6.98% for the NGT group, and 15.16% for the AGM group. Regarding the body fat mass percentage between the T1 and T3 periods, the NGT group showed a reduction from 44.29 ± 4.27 to 28.26 ± 4.72, and a fat mass regain of 28.26 ± 4.72 at T3 and 37.17 ± 6.41 at T4. The AGM group showed a reduction in body fat mass percentage from 47.33 ± 5.07 at T1 to 32.74 ± 5.92 at T3, and a regain at T4 38.20 ± 4.94.

The mean weight in kilogrammes at T1 was NGT 122.84 kg ± 23.82 and AMG 114.86 kg ± 16.35. After 18 months of RYGB (moment of maximum loss), the mean was NGT 66.60 kg ± 12.49 and AMG 68.60 kg ± 16.36. The percentage of loss between these two moments was NGT 54.64% ± 7.23 and AMG 60.64% ± 11.54. And 10 years after RYGB, the mean weight was NGT 83.48 kg ± 17.055 and AMG 85.23 kg ± 13.75 and the percentage of weight regain between 18 months and 10 years after RYGB was NGT 26.37% ± 23.26 and AMG 26.34% ± 11.75.

Regarding the HOMA-IR and ΔHOMA-IR evaluations, Table 2 shows that NGT and AGM presented a significant reduction of the values over time. AUGC total showed a decrease from T1 to T3 periods for both groups, and at T4 period, the values return to be similar to T1. Fasting insulin (Table 2) in NGT and AMG have levels compatible with IR at T1, and at T4 period, fasting insulin levels became normality. The HbA1c at T4 for the NGT group was 5.5 ± 0.31% and the AMG group was 6.4 ± 1.29.

Table 2 OGTT data
Full size table

According to the plasma glucose curves of NGT (Fig. 1A) and AGM (Fig. 1B), patients presented an average glycaemic level of 117.8 ± 17.53 mg/dL and 202.07 ± 32.82 mg/dL respectively at 120 min for T1 period. When analysed separately, these groups showed a significant reduction in the mean plasma glucose levels after RYGB T4 period at 120 min after OGTT with NGT 55.49 ± 17.15 mg/dL and AMG 124.58 ± 60.94 mg/dL (p˂0.001), with glycaemic levels of the NGT and AGM groups statistically different at 120 min. Observing glucose curve of the AMG group (Fig. 1B), the glycaemia means at t30, t60, and t90 min are above 200 mg/dL. For the NGT group (Fig. 1A), it is noted that the mean 120-min glycaemia at T4 period is below 60 mg/dL.

For the primary analysis using GEE model when comparing NGT (Fig. 1A) with AGM (Fig. 1B), there was a significant change (p < 0.0001) between means of T1 vs. T2, T1 vs. T3, T2 vs. T3, T3 vs. T4, time intervals between 0 vs. all times (30, 60, 90, and 120 min), 30 min vs. 120 min, 60 min vs. 90 min and 120 min, and 90 min vs. 120 min.

Fig. 1C and D show the mean plasma insulin levels during OGTT for the NGT and AGM groups performed at T1, T2, T3, and T4. There was a change from T1 to T3 at the NGT and AGM groups, and a significant reduction from the first hour of OGTT, including fasting values. At T4 period, the NGT group had mean insulin levels increased at t30, t60 and t0, and t90 and t120 which were lower compared to T1. However, the AMG group had mean insulin levels at T4 lower than those at T1 and reduction at t90 and t120 min

Fig. 1E and F show the results of mean plasma levels of GLP-1 during the OGTT. The GLP-1 curves are statistically different (p = 0.001) for the NGT and AGM groups. From T3 to T4 periods, there is a significant increase (p < 0.0001) in GLP-1 secretion, at times 30 and 60 min. The AGM group at T4 had a higher increase in GLP-1; t30 of OGTT for AMG was 63.85 ± 37.98 pmol/L (Fig. 1F) when compared to NGT and t30 OGTT was 50.73 ± 24.82 pmol/L (Fig. 1E)

Discussion

Currently, RYGB is one of the most effective methods for treating morbid obesity, especially when the patient is unsuccessful after trying clinical, nutritional, and drug treatment [39,40,41,42].

Postoperatively, there are several long-term benefits for patients, such as control of systolic and diastolic pressures; normalization of serum levels of total cholesterol, triglycerides, uric acid and HDL cholesterol fraction, and decreasing cardiovascular risk; reduction in cases of asthma and obstructive sleep apnea, improvement in glycaemic levels and, consequently, carbohydrate metabolism, as well as decreased risk of type 2 diabetes; and psychological improvement, influencing self-esteem, social relationship, and reducing anxiety and depression [43, 44].

In our study, at T4 period, the BMI was NGT 32.45 ± 4.99 kg/m2 and AGM 34.85 ± 4.46 kg/m2, despite the weight regain that the patients had; according to Magro [21], there was no surgical failure (the failure is considered when the patient has BMI > 35 kg/m2). And even with weight and waist regain, these values remained lower in T4 when compared to T1; so, it is important to note that even with this weight recovery, NGT still have an improvement of metabolic levels.

Another factor evaluated was GLP-1 levels, this hormone has been studied after RYGB and it is related to the increase or even restoration of glucose-induced insulin secretion [19, 45]. Roux-en-Y gastric bypass (RYGB) is a bariatric surgery developed by Mason and Ito in 1967 for the management of obesity [46]. It is well known that GLP-1 mechanism and its potential increased levels in patients after RYGB [19, 45]. It has been hypothesized that the early delivery of nutrients to the distal gut may lead to improvement in diabetes after RYGB, the exact mechanism by which GLP-1 levels increase after RYGB remains unclear. In the hindgut hypothesis, secretions and nutrients reach the distal ileum and colon more quickly following gastric bypass, which may lead to more stimulation of L-cells and therefore a higher level of GLP-1 [47]. A recent [48] publication demonstrated that GLP-1 expressing cells are located in the stomach and contribute to the active levels of GLP-1 in the portal vein. These cells undergo changes in plasticity after bariatric surgeries in rats and obese individuals and can modulate gastric functions locally, contributing to the control of glucose homeostasis.

At T4 period, in the NGT group, GLP-1 levels increased at t30 and t60 min of OGTT and this fact is concomitant with the increase of insulin that occurs at t30–t60 min of OGTT. After this period, both levels decreased. However, the AGM group had at T4 expressive levels of GLP-1; they did not show corresponding insulin levels (Supplementary 1), probably this happens because this group patients were DM2 or were progressing to the clinical picture of DM2, so we must consider a possible beta cell degeneration process or beta cell loss (a probable pancreatic insufficiency) [49] so in our study there was no coincidence of the GLP-1 and insulin curves in the AMG after 10 years of RYGB; also, the AGM group at T4 had HbA1c 6.4 ± 1.29%. Another point we must consider is that we did not do the OGTT with an antagonist of GLP-1, like Jørgensen et al. [7] did to better elucidated the relation of GLP-1 and insulin levels.

The role of GLP-1 in T2DM remission after bariatric surgery is difficult to determinate in humans, because of other predictive factors such as duration of disease and degree of b-cell destruction before surgery; those may be more critical in determining whether b-cells can recover sufficiently to resolve T2DM [49]. These findings could be an explanation for maintaining normalization of blood glucose and insulin levels as well as part of the resolution of type 2 diabetes after the RYGB [50, 51]. In a recent study, Mingrone [52] was found that metabolic surgery resulted in continuous remission of DM2 within 10 years in 37.5% of patients with advanced type 2.

Improvement in glucose metabolism after weight loss occurs in either type 2 diabetes, carbohydrate-intolerant, or healthy oral glucose-tolerant patients, and in some cases, they have drug doses reduction [49, 53]. We observed a significant improvement in the blood glucose levels after OGTT specially at the NGT group. The AGM group presented levels higher than the NGT at T4. AMG also at T4 had mean glucose levels above T1, so these mean levels at T4 were still compatible with DM2 levels as ADA18 [22]. At the NGT group, even though had weight regain, the HOMA-IR levels are below value considered reference for IR diagnosis [54], demonstrating evidence of metabolic improvement for NGT after 10 years of RYGB. Moreover, in a previous study, this same group of morbidly obese individuals undergoing RYGB has demonstrated increased GLP-1 secretion at 9 months after RYGB and an improvement of insulin sensitivity, a fact that was reinforced by findings in other studies [55] and a study made by our group; postoperative RYGB patients were evaluated by a hyperinsulinemic-euglycemic clamp, and it was found that weight loss was associated with increased insulin sensitivity and decreased basal insulin secretion, with patients reaching values similar to the control group [56, 57].

It was also observed at the NGT group that even though there was weight regain at T4, HOMA-IR decreased throughout periods with a statistical difference between them; and at T4, the levels were lower compared to T1. The AGM group showed an important reduction in HOMA-IR when compared to the NGT group, but there was a decrease in the mean insulin levels also at T4 which may have decreased without leading an improvement in DM2; and at T4, the AMG group still has HbA1c 6.4%, according to ADA 18 [22], lower value for DM2 diagnosis.

We also observed that 55% of patients in the NGT group at T4 period had blood glucose lower than 60 mg/dL. Between these patients, the mean of blood glucose was 42 mg/dL at 120 min of OGTT, which was accompanied by symptoms such as nausea, vomiting, diarrhoea, sweating, and headache in some of them. This picture was promptly reverted with intravenous glucose infusion, which was made after the last collected blood sample at time 120 min of the OGTT. Also, at the 30 min of OGTT, GLP-1 secretion in T4 was 12.3 times higher than T1 in the NGT group. One study evaluated postoperative patients and observed very high levels of GLP-1 after RYGB, and such patients presented severe hypoglycaemia after RYGB with GLP-1 concentrations 10 to 20 times higher than usual [58, 59]. Moreover, at T4, the NGT group had glucose levels NGT = 55.49 ± 17.15 mg/dL, so a plausible explanation that hypoglycaemia presented by several patients undergoing this surgical procedure could be due to elevated GLP-1 levels.

It is important to note that several other authors have evaluated the consequences of the RYGB but most of them in the early postoperative period (2–5 years). PubMed survey [6, 60,61,62] and very few articles [19, 20] have shown GLP-1 after 10 years of RYGB, with our study exclusively assessing the same patients before and after 10 years of the RYGB.

This study has shown that after 10 years of RYGB, patients had regained weight, but the metabolic improvement remained specially at the NGT group; this fact can be attributed to the high levels of GLP-1 that these patients presented even after 10 years of RYGB. In order to better understand the benefits and consequences of this surgery through insulin secretion, glucose, and GLP-1 levels, further research is needed.

Study limitations

An important limitation of this study refers to the sample size, which is presented in a small number, allowed to consider the results found only in the population in question.

In addition, the fact that OGTT was not performed with an antagonist of GLP-1, like Jørgensen et al. did [7] to better understand long-term DM2 reversal, and also, we did not measure the C-peptide to assess insulin secretion.

However, with the absence of long-term follow-up articles using the same patients, the results of this research can be further complemented with future studies that will focus on the long-term consequences of a RYGB surgical procedure.

Conclusion

According to the results found in this study, it was noted by the evaluation of the same patients during 4 different times in the 10-year post-operatory RYGB period; even with weight regain, high levels of GLP-1 remained which can be associated with metabolic improvement specially at normal glucose-tolerant obese. Therefore, long-term follow-up after RYGB studies, using the same patients, is needed to better understand the mechanisms involved in the metabolic improvement.