Post-bariatric hypoglycaemia: an under-recognized complication of weight loss surgery - a case series

Abstract

Post-bariatric hypoglycaemia (PBH) has emerged as a significant, under-recognized complication following Roux-en-Y gastric bypass (RYGB) and other weight loss surgeries. The purpose of this study was to characterize the timing, clinical presentation, diagnosis and management outcomes of PBH in a diverse case series from a multi-speciality centre in Southern India owing to a predominantly high-carbohydrate diet and distinct bariatric phenotype. This was a retrospective observational case series analysis from a multidisciplinary bariatric clinic at a tertiary teaching hospital in Southern India from May 2017 to April 2024. Time to onset of PBH, nadir plasma glucose levels, management approach (diet, medication and surgery), complication rates and patient follow-up duration were evaluated. Nine adults (mean age: 41.5 years; both sexes) were diagnosed and managed for PBH following RYGB. PBH onset ranged from 1 month to 5 years following RYGB (mean: 17 months). Nadir glucose levels ranged from 16 to 67 mg/dL. Hyperinsulinaemia and nesidioblastosis were diagnosed in select cases. Patients received individualized medical nutrition therapy, medication (voglibose, octreotide and diazoxide) and, in severe/refractory cases, surgical procedures such as reversal of RYGB or subtotal pancreatectomy, as was required in two patients. Frequent and severe PBH episodes were more common in patients with higher initial BMI and persistent carbohydrate-heavy diet intake. PBH after bariatric surgery presents with variable onset, severity and duration, emphasizing the need for long-term multidisciplinary follow-up and individualized management, especially for high-risk populations. Early recognition and tailored intervention can improve clinical outcomes and patients’ quality of life.

Introduction

The prevalence of obesity is on an increasing trend all over the world (1). The National Family Health Survey 5 (NFHS 5) data from India showed the incidence of overweight and obesity to be 44.01% in males and 41.16% in females (2). The ICMR-INDIAB study showed a higher rate of abdominal obesity (39.5%) compared with generalized obesity (28.6%), along with a high frequency of metabolic complications, including hypertension (35.5%), dyslipidaemia (81.2%) and diabetes (11.2%).

Because of the increased metabolic risk, the threshold for surgical intervention for obesity in the Indian population is >37.5 kg/m² or >30 kg/m² with obesity-related comorbidities, such as diabetes and hypertension (3). The common procedures include Roux-en-Y gastric bypass (RYGB), sleeve gastrectomy, adjustable gastric banding, biliopancreatic diversion and one-anastomosis gastric bypass, of which sleeve gastrectomy and RYGB remain the most common procedures worldwide (4). While the results for weight loss (mean weight loss: 25.7%) (5) and resolution of metabolic complications, such as hypertension (46%), diabetes (58%) and dyslipidaemia (46%), are impressive (6), a significant proportion has faced complications, including nutritional deficiencies, surgical complications, anastomotic stricture, cholelithiasis, bone diseases such as osteoporosis or osteopenia, kidney stones and requirement of repeat surgery.

Post-bariatric hypoglycaemia (PBH) is a concerning and under-recognized issue (7). It usually manifests over a year after surgery, with symptoms occurring 1–3 h after eating. PBH significantly impairs the quality of life (QoL) (8) and is more commonly associated with RYGB compared with laparoscopic sleeve gastrectomy (LSG). There is no single agreed-upon criterion or blood glucose level to define PBH, but most commonly, it is characterized by neuroglycopaenic symptoms with postprandial plasma glucose less than 54 mg%, usually occurring ≥6 months after bariatric surgery, without fasting hypoglycaemia (9). The prevalence of PBH varied from 0.1 to 0.4% in various studies (10); however, symptoms were reported in up to 30% of patients (11) and in 54.4% of patients when data were collected with continuous glucose monitoring (12).

The mechanisms behind PBH involve multiple factors, including rapid delivery of nutrients to the lower intestine, which results in a post-meal glucose spike triggering excessive glucagon-like peptide-1 (GLP-1) secretion from gut cells, leading to an overly robust insulin output, eventually leading to hypoglycaemia (10). Treatment options for PBH are currently restricted, focusing mainly on diet adjustments and repurposed medications, such as acarbose, diazoxide, somatostatin mimics (such as octreotide and pasireotide) and calcium blockers. There are limited data on effectiveness and risks, and patient acceptance is limited and inconsistent.

There are no published data on PBH from India, where the bariatric phenotype includes higher VAT area, a diet with a larger proportion of carbohydrates, lower BMI cut-off for surgery and poor follow-up, in addition to the consequences of surgery (11). People across the landscape of India consume a diet rich in simple carbohydrates, which makes them more prone to hypoglycaemia following bariatric surgery, as simple carbohydrates are a strong insulinogenic signal (13). This study aims to explore the clinical presentation, onset time, and management approaches of PBH among patients who have undergone bariatric surgery, focusing on identifying effective treatment strategies.

Materials and methods

This was a retrospective observational study conducted at a multidisciplinary outpatient bariatric clinic in a quaternary care multi-speciality teaching hospital centre in Southern India from May 2017 to April 2024. The study protocol was approved by the Institutional Review Board (IRB Min No. 101046, dated 22 June 2016).

All patients presenting with documented hypoglycaemia following the RYGB surgery were included. Patients with symptoms suggestive of dumping syndrome (bloating, nausea and diarrhoea with/without vasomotor symptoms like flushing) were excluded. Evaluation of hypoglycaemia involved mixed-meal challenge test (MMCT). After an overnight fast during which only water was permitted, baseline plasma glucose, serum insulin and serum C-peptide were drawn at 06:00. Patients then consumed 9.5 scoops of a standard oral nutritional supplement (Ensure, approximately 82 g) reconstituted in 400 mL of water over 30 min, and paired plasma glucose, insulin and C-peptide were sampled hourly for 5 hours. Hyperinsulinaemic hypoglycaemia was confirmed with symptomatic hypoglycaemia (plasma glucose <55 mg/dL) and simultaneously increased serum insulin (>3.0 μIU/mL) and serum C-peptide (>0.6 ng/mL) (9). Plasma glucose was measured using hexokinase method; serum insulin was measured by the chemiluminescence method (Siemens Immulite assay CV 4.8%) and C-peptide was measured with chemiluminescent assay (Siemens Atellica, Germany CV 4%). Two patients required nuclear imaging with an Exendin Ga-68 DOTATATE scan to confirm nesidioblastosis.

Data regarding demographics, clinic characteristics, duration of symptoms, past comorbidities, metabolic profile and management were recorded. Patients were classified as early PBH as those with symptoms <6 months (after exclusion of dumping syndrome) and late PBH as those with symptoms occurring after 6 months.

Escalation to a surgical intervention was considered by a bariatric multidisciplinary team on a case-by-case basis rather than by a fixed protocol. The criteria applied were as follows:

• Failure of medical therapy: all patients had first received medical nutrition therapy followed by sequential pharmacotherapy with an α-glucosidase inhibitor, then diazoxide and then octreotide LAR, at maximally tolerated doses.

• Severity of hypoglycaemia: recurrent neuroglycopaenia – seizures, loss of consciousness or episodes requiring third-party assistance – with a documented plasma glucose nadir below 20 mg/dL.

• Biochemical confirmation: endogenous hyperinsulinaemic hypoglycaemia confirmed on MMCT, with insulinoma, exogenous insulin or sulphonylurea exposure and adrenal insufficiency excluded.

• Functional imaging pattern: 68Ga-DOTA-exendin-4 PET/CT showing diffuse pancreatic tracer uptake without a discrete lesion, supporting diffuse islet-cell hyperplasia rather than a resectable adenoma, was taken to favour a pancreatic rather than a purely anatomical target.

• Operative risk and comorbidity: cardiorespiratory reserve, prior thromboembolism, active or recent malignancy and nutritional status were weighed against operative risk.

• Counselling on weight regain and patient preference: all patients were counselled that reversal or conversion carries a substantial probability of weight regain, and the final decision was taken jointly with the patient.

Of the nine index RYGB procedures, four were performed at our institution (patients 1, 6, 7 and 9) and five elsewhere (patients 2, 3, 4, 5 and 8), the earliest in 2005. Operative notes were retrievable for the four in-house procedures, but recorded neither gastrojejunal anastomotic (stoma) calibre nor alimentary, biliopancreatic and common-channel limb lengths, since routine documentation of these variables was not departmental practice during the period concerned. The data were analysed using SPSS Statistics (version 21; SPSS, Inc., USA) and validated for normal distribution using the Shapiro–Wilk test. All non-parametric variables were analysed using the Kruskal–Wallis test, and all parametric variables were analysed using the Student t-test. All categorical variables were analysed using Fisher’s exact test rather than the Chi-square test. ANOVA was used for comparisons of continuous variables across more than two groups, as applicable. P < 0.05 was taken as significant.

Results

The characteristics of the study cohort are described in Table 1. We identified nine patients with post-bariatric hypoglycaemia (PBH) with a median age of 41.5 years (range: 28–54 years) and a slight female predominance (5/9). Most patients had cardiometabolic comorbidities, including hypertension (6/9), obstructive sleep apnoea (3/9) and diabetes mellitus (2/9). Pre-operative BMI was 42.15 kg/m² (range: 34.2–50.1 kg/m²), and all patients experienced substantial weight loss following bariatric surgery, with percentage weight loss typically exceeding 30% and reaching up to 60% in some cases. A subset subsequently regained a considerable amount of weight (≥10 kg in 4/9), whereas the remainder had minimal or no documented regain. Figure 1 shows the weight trajectory of the patients.

Table 1.

Characteristics of study population.

S. No.	Age and gender	Comorbidities	Initial weight (kg)	Weight lost (kg)	% Initial weight lost	Regained weight (kg)	Time to PBH in months	Nadir blood glucose (mg/dL)	Treatment	Follow-up (years)
1	40F	HTN, OSAS	118	60	50.85%	0	5	18	Diet, voglibose	4
2	54M	HTN	150	80	53.33%	12	4	16	Diet, voglibose, octreotide	19
3	28F	HTN, DM	108	34	31.48%	11	1	16	Diet, voglibose, diazoxide, octreotide, subtotal pancreatectomy	2
4	53F	DM, HTN, CA breast	112	41	36.61%	0	12	47	Prednisone, diazoxide, MF octreotide	4
5	53M	Asthma, OSAS, HTN	108.5	43	39.63%	12	10	33	Diet, voglibose, octreotide	7
6	28F	ASTH, HTN, OSAS	130	78	60%	0	12	54	Diet, acarbose, octreotide	7
7	44M	Chronic venous insufficiency	135	65	48.15%	0	30	67	Diet, voglibose	6
8	29M	Arnold–Chiari, cauda equina	140	50	35.71%	28	12	-	Diet, voglibose	6
9	35F	NIL	100.2	22	21.96%	0	60	64	Diet, voglibose	7

OSAS, obstructive sleep apnea syndrome; HTN, hypertension; DM, diabetes mellitus; CA, carcinoma; ASTH, asthma.

Figure 1.

Weight changes after RYGB during follow-up.

Time from surgery to onset of PBH was highly variable, ranging from as early as 1 month to as late as 60 months, with several patients presenting between 10 and 12 months. Nadir glucose levels ranged from 16 to 67 mg%, indicating a range of severity in hypoglycaemia. Hypoglycaemia was often profound, with nadir plasma glucose concentrations as low as 16–18 mg/dL in multiple cases. Eight out of nine patients had postprandial hypoglycaemia; one patient had both fasting and postprandial hypoglycaemia. All patients received medical nutrition therapy (MNT) as part of their treatment plan. For patients whose symptoms were not relieved with MNT alone, various medications were used, including octreotide LAR, voglibose and diazoxide. Two patients underwent pancreatic surgery (subtotal or partial pancreatectomy) for refractory symptoms. Follow-up duration ranged from 2 to 19 years, indicating the chronic nature and long-term management needs of PBH in this cohort.

Patient 1 lost 50.85% of his weight in 24 months and developed nadir blood glucose of 18 mg/dL, which was managed with MNT and voglibose. Patient 2, despite gaining back 20 kg of his initial 80-kg weight loss, continues to require octreotide and voglibose to manage PBH 19 years post-surgery. Patient 3 had hypoglycaemic seizures as presentation of PBH, which occurred 1 month post-surgery. Patient 4 developed pulmonary embolism and subsequently required revision of leak following surgery. While antibodies were negative, she required addition of both prednisolone and mycophenolate mofetil to manage PBH. Her previous comorbidities precluded pancreatectomy for management of nesidioblastosis. Patient 5 developed PBH 10 months post-surgery and required treatment with octreotide. He developed new-onset biliary sludge after octreotide, necessitating regular follow-up (Fig. 2). Patient 6 lost 60% of her initial weight post-surgery and required octreotide and acarbose for the management of PBH; however, bradycardia necessitated cessation of octreotide. Patient 7 had delayed onset of PBH occurring more than three years after initial surgery. Patient 8 regained 28 kg of his 50-kg weight loss after surgery after developing PBH at 12 months. He, however, was able to manage with MNT and was able to lose weight again resulting in PBH remission. Patient 9 also had delayed onset of PBH, which was managed with MNT and voglibose.

Figure 2.

Biliary sludge in patient 5 after starting octreotide.

Table 2 shows the comparison of early PBH (<6 months) vs late PBH (>6 months). Percentage weight loss is higher in the early PBH group vs the late PBH group (50.8% (41.2–52.1) vs 38.1% (35.9–46)), but it is not significant. Nadir blood glucose is significantly lower in the early PBH group (16.7 ± 1.2 vs 53.0 ± 13.7 mg/dL, P = 0.004). Two people out of the whole group had diabetes, both of whom developed early PBH. The correlation between nadir blood glucose and time to PBH is shown in Fig. 3. Regression demonstrates a strong positive association between nadir glucose and time to PBH (r = 0.78, P = 0.02), indicating that patients with lower nadir glucose values tend to develop PBH earlier after surgery. This pattern is consistent in both sexes, although the study is underpowered for sex-specific analyses.

Table 2.

Comparison of PBH (<6 months) vs late PBH (>6 months).

Variable	Early PBH (n = 3)	Late PBH (n = 6)	P value
Median % weight loss	50.8 (41.2, 52.1)	38.1 (35.9, 46.0)	0.601
Nadir glucose (mg/dL)	16.7 ± 1.2	53.0 ± 13.7	0.004
Age (years)	40.7 ± 13.0	40.3 ± 11.3	0.972
Sex (F/M)	2/1	3/3	1
Diabetes	2/3 (66%)	1/5 (20%)	0.453

Figure 3.

Relationship of severity of hypoglycaemia and time to PBH.

Discussion

Our study comprehensively presents the diagnosis and management strategies of PBH in patients who underwent RYGB in a multidisciplinary hospital in Southern India. While symptoms remain common post-surgery, its documented prevalence ranges from 0.1 to 0.4% in various studies (10). The prevalence in people undergoing RYGB with pre-existing diabetes has been reported to be 16.6% at 1 year post-surgery (14).

The median time of onset of PBH in our population was 17 months (range: 1–60 months). This is much earlier compared with the Swedish cohort, which reported a median onset of symptoms at 2.7 years (range: 1–14.8 years) post-surgery (15). Similarly, Lee et al. (11) reported a mean duration of 3.2 ± 1.2 years before the onset of hypoglycaemic symptoms.

In our study, two patients (patients 3 and 4) had severe hypoglycaemia, which was not managed on MNT, voglibose and diazoxide. Exendin Ga-68 DOTATATE scan showed diffuse hyperplasia of whole pancreas without any single lesion. Both patients had a previous history of diabetes, and the onset of symptoms was at 1 and 12 months, respectively, post-surgery after an average weight loss of 34.1%. In our series, this has occurred earlier and at a lower weight loss compared with Service et al. (16), who reported that the symptoms occurred after an average of 2.5 years and a median weight loss of 44.2%.

In our study, people with diabetes had early PBH and more severe presentations (e.g. nesidioblastosis requiring surgery). This is in contrast to Lee et al. (17), who reported neuroglycopaenic symptoms to be more common in non-diabetic individuals, but not shown to be significant on multivariate analysis. The Bariatric Outcome Longitudinal Database (BOLD), analysing a total of 275,514 cases, found that the incidence of PBH after RYGB reduced from 0.1 to 0.02% if patients with diabetes were excluded (18). A meta-analysis showed higher fasting glucose to be significantly associated with PBH (OR 3.16, P = 0.008); however, age and gender had no significant impact (19).

Our study has several strengths. This is the first case series of PBH from South Asia, highlighting the unique features and response of the population to RYGB surgery over a long follow-up. All patients had biochemical evidence of PBH in the presence of hyperinsulinaemia for diagnosis. The study acknowledges limitations, including a small sample size and potential biases due to the retrospective nature of the data collection, which precludes any definitive conclusion regarding the association between nadir glucose and time to PBH. The study did not include comparison with cases who underwent RYGB without PBH as this was a retrospective case series. The variations in surgical techniques and follow-up care may have affected the findings. While the MNT advice was given to all patients, adherence could not be assessed due to the retrospective nature of the study.

Conclusion

This study underscores the significant challenges in managing PBH following Roux-en-Y gastric bypass surgery. In this case series, PBH tended to occur earlier than that reported in some Western cohorts, particularly in patients with a prior history of diabetes, and appeared to be related to the amount of weight loss. This study highlights the need for personalized care, long-term follow-up and larger multi-centre studies to confirm these patterns. Public health strategies should emphasize education on potential long-term complications of bariatric surgery and the importance of adherence to dietary modifications and medical management. Future research should focus on larger, multi-centre studies to assess the regional differences in occurrence and severity of PBH and to explore new therapeutic non-surgical approaches to manage the refractory cases.

Declaration of interest

The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the work reported.

Funding

This work did not receive any specific grant from any funding agency in the public, commercial or not-for-profit sector.

Data availability

Data will be made available upon reasonable request to the corresponding author.

Ethics

This study was approved by the Institutional Review Board of Christian Medical College Vellore, India (IRB no. 10146, dated 22 June 2016). The data were collected retrospectively, but all care has been taken to ensure that the patient information and identification data are not accessible, except to primary investigators. The manuscript was read and approved by all authors, and each author believes that the manuscript represents honest work.

Artificial intelligence

No artificial intelligence was used to prepare the manuscript.