Bariatric Surgery for Patients With Early-Onset vs Late-Onset Type 2 Diabetes

Importance  The prevalence of early-onset type 2 diabetes mellitus (T2DM), which responds poorly to medical treatment, is increasing. Bariatric surgery has been well recognized for its effectiveness in the remission of T2DM, but its effectiveness and durability in the remission of early-onset T2DM has not yet been explored.

Objective  To compare the short- and long-term outcomes of bariatric surgery with a specific focus on the rate of remission of T2DM in patients with early-onset (age <40 years) and late-onset (age ≥40 years) T2DM.

Design, Setting, and Participants  In this cohort study, 558 Taiwanese patients (339 with early-onset T2DM and 219 with late-onset T2DM) with a body mass index (calculated as weight in kilograms divided by height in meters squared) above 25 underwent bariatric surgery to ameliorate T2DM between January 1, 2007, and December 31, 2013. Patients were followed up for at least 1 year. Preoperative, perioperative, and postoperative clinical and laboratory data were prospectively collected and compared between the 2 groups.

Main Outcomes and Measures  Rate of remission of T2DM (hemoglobin A_1C <6.0% without antiglycemic medication) was the primary outcome measure.

Results  Of the 558 patients (345 women) in the study, mean (SD) ages were 33.5 (7.5) for those with early-onset T2DM and 50.6 (6.5) with late onset T2DM. Those with early-onset T2DM had higher mean (SD) preoperative BMI and hemoglobin A_1C values (39.4 [8.5] and 8.7% [3.8%] of total hemoglobin [to convert hemoglobin to a proportion of total hemoglobin, multiply by 0.01], respectively) than did patients with late-onset T2DM (36.7 [7.5] and 8.2% [1.6%], respectively). Distribution of surgical procedures and major complications were similar between the 2 groups. At 1 year, patients with early-onset T2DM achieved greater weight loss than those with late-onset T2DM, although the difference was not statistically significant. A higher rate of complete remission of T2DM was observed in patients with early-onset T2DM than in those with late-onset disease (193 [56.9%] vs 110 [50.2%]; P = .02). At 5 years, patients with early-onset T2DM still maintained a higher rate of weight loss (mean [SD], 30.4% [11.8%] vs 21.6% [11.7%]; P = .002) and higher rate of remission (47 of 72 [65.3%] vs 26 of 48 [54.2%]; P = .04) than did those with late-onset disease. Age at bariatric surgery, duration of T2DM, and preoperative C-peptide level were independent predictors of remission. The remission rate was directly related to extent of weight loss. Multivariate analysis confirmed the higher rate of remission in the group with early-onset T2DM.

Conclusions and Relevance  This article describes the largest long-term study examining bariatric surgery for patients with early-onset T2DM. Bariatric surgery may achieve better and more long-lasting glycemic control in select patients with early-onset T2DM than in those with late-onset T2DM.

Type 2 diabetes mellitus (T2DM) is becoming a global epidemic, with an ongoing decline in the age of onset.^1-3 The incidence of T2DM among young Americans increased from 2.9 per 10 000 in 2001 to 3.6 per 10 000 in 2009.³ The same trend was observed among young Asians.⁴ About 20% of T2DM in Asians was reported as early-onset T2DM (newly diagnosed before age 40 years). To our knowledge, the cutoff age for early-onset T2DM is yet to be standardized globally. To include women who are of reproductive age, the cutoff age for early-onset T2DM is usually set at 45 years in Western populations but at a younger age in Asians owing to the difference in demographics. The Joint Asian Diabetes Evaluation program, which enrolls patients from 9 Asian countries or regions, set a cutoff age for early-onset T2DM at 40 years.⁴ In contrast with late-onset T2DM, early disease onset imposes a longer exposure to the risk of progression to severe complications following chronic hyperglycemia.^5-8 Patients with early-onset T2DM had higher risks for cardiovascular and renal complications at any age.^5,7-11 Compared with age-matched members of the population, individuals with early-onset T2DM have 30 times the risk of stroke and 14 times the risk of myocardial infarction.⁸ The increase in cardiovascular risks was driven primarily by metabolic syndrome, which can accompany T2DM. This finding highlighted the importance of controlling T2DM in patients with early-onset disease.

However, treatment of early-onset T2DM is more difficult than treating late-onset T2DM; an intensive lifestyle intervention was not effective in the treatment of youth with T2DM.^12,13 The use of surgical procedures, such as bariatric surgery, to treat diseases related to metabolic syndrome has been explored in patients with poorly controlled T2DM.^14-16 This surgery is the application of bariatric surgical procedures for patients with a body mass index (BMI) (calculated as weight in kilograms divided by height in meters squared) of less than 35 who also have severe comorbidities. Several randomized trials have justified the use of bariatric surgery as a viable treatment in patients with a BMI of less than 30 and poorly controlled T2DM.^17-20 To our knowledge, the efficacy of bariatric surgery in the treatment of patients with T2DM with different ages of onset has not been examined. In a consecutively recruited, prospective Taiwanese cohort that received bariatric surgery to treat T2DM, we compared the baseline clinical characteristics and the efficacy of glycemic control between patients with early- and late-onset disease.

Box Section Ref ID

Prospective clinical trials (NCT00540462 and NCT01317979) of bariatric surgery in the treatment of patients with T2DM began January 1, 2007.^21,22 This study was approved by the Ming-Sheng General Hospital Institutional Review Board. All patients provided written informed consent for the use of their data. Patients were eligible for bariatric surgery if they met American Diabetes Association criteria for T2DM,²³ were aged 18 to 67 years, had a BMI greater than 25 with poor control of their T2DM with diet and medical therapy (hemoglobin A_1C [HbA_1C] >6.5% of total hemoglobin [to convert hemoglobin to a proportion of total hemoglobin, multiply by 0.01]), and received approval from the endocrinologists (including K.C.) that they would benefit from bariatric surgery. Exclusion criteria were the presence of major organ dysfunction with American Society of Anesthesiologists class greater than 3, pregnancy, and previous gastrointestinal surgery. In addition, patients were excluded if they had been diagnosed with type 1 diabetes mellitus, had anti–glutamic acid decarboxylase or islet cell autoantibodies, had a fasting C-peptide level of less than 1 ng/mL (to convert to nanomoles per liter, multiply by 0.331), or were unresponsive to a stimulus test. Patients with a fasting C-peptide level of less than 1 ng/mL or an increase in C-peptide level of less than 50% in response to a glucagon stimulation test were classified as unresponsive.

As of December 31, 2013, a total of 558 patients who had received bariatric surgery and completed at least 1 year of follow-up were included in this study. The patients who were first diagnosed with T2DM before the age of 40 years were defined as early-onset T2DM patients.⁴ Patients with late-onset T2DM were those who were first diagnosed with T2DM at or after age 40. We examined and compared the outcomes between the 2 groups (eFigure in the Supplement).

Anthropometric measures (BMI and blood pressure) and blood chemical data (fasting plasma glucose, HbA_1C, serum insulin, and lipids) were measured at baseline and at annual follow-ups. Insulin resistance was measured by the homeostasis model assessment (HOMA) index, calculated as plasma glucose (millimoles per liter) × insulin (international units per milliliter)/22.5 (to convert glucose to millimoles per liter, multiply by 0.0555; insulin to picomoles per liter, multiply by 6.945).²⁴ Weight loss percentage was calculated as weight loss (kilograms)/initial weight (kilograms) × 100.

All the procedures were performed by 3 senior surgeons (W.-J.L., K.-H.S., and C.-C.W.) with more than 10 years of experience in various kinds of bariatric surgical procedures, following a standardized operative technique. The surgical procedures comprised 2 types of gastric bypass procedures: laparoscopic Roux-en-Y gastric bypass and laparoscopic single anastomosis (mini) gastric bypass.²⁵ In brief, using a standard 5-port laparoscopic technique, Roux-en-Y gastric bypass was performed by the antecolic and antegastric route with an alimentary limb of different length according to the patient’s BMI: 100 cm for a BMI of 35 to 45, 150 cm for a BMI of 46 to 55, or 200 cm for a BMI of 56 or more. The gastric pouch was approximately 20 mL, and the gastrojejunostomy was created by stapler technique with a 1.2-cm-diameter anastomosis. Laparoscopic single anastomosis (mini) gastric bypass was performed by creating a long-sleeved gastric tube that was approximately 2.0 cm wide along the lesser curvature from just below the angle at the lesser curvature to the angle of His. A Billroth II–type loop gastroenterostomy was then created with the intestine at 150 to 250 cm distal to the ligament of Treitz. We used 150 cm for patients with a BMI of 39 or less, 200 cm for a BMI of 40 to 49, and 250 cm for a BMI of 50 or above. Nonbypass procedures included laparoscopic adjustable gastric banding and laparoscopic sleeve gastrectomy^26,27; the latter was performed by creating a sleeve gastrectomy over a 36F bougie and leaving a 4-cm antrum. The operation type was usually decided by the patient together with the multidisciplinary team based on the patient’s BMI and other medical conditions. Gastroesophageal bypass is usually recommended unless the patient did not want to be subjected to the higher risks with a surgical procedure and the potential long-term complications of bypass surgery, such as nutritional deficiencies.

Patient follow-up was scheduled at the first, third, sixth, and 12th month for the first year and then annually. Basic clinical parameters and laboratory test results for nutritional parameters were recorded at every visit. Treatment success was defined as complete remission (HbA_1C <6.0% without antidiabetic medication), partial remission (HbA_1C <6.5% without antidiabetic medication), and prolonged remission (complete remission for 5 years).²⁸

All statistical analyses were performed using SPSS, version 12.01 (SPSS Inc), with baseline comparisons made using χ² tests and 2-sample t tests. Continuous variables were expressed as mean (SD). The differences between the groups in pertinent characteristics were established with t tests for independent samples. P ≤ .05 (2-sided) was considered statistically significant. Logistic regression analysis was used to control the effect of baseline characteristic variables (Table 1) and follow-up outcome variables on remission after bariatric surgery. Categorical variables were converted to dummy variables before logistic regression analysis. The variables with P ≤ .05 were selected for multivariate analysis to determine the factors predicting or associated with remission of T2DM.

Of 558 patients (345 women), 339 (60.8%) had early-onset T2DM and 219 (39.2%) had late-onset T2DM. Mean (SD) ages were 33.5 (7.5) for those with early-onset T2DM and 50.6 (6.5) with late onset T2DM. The duration of T2DM in both groups was similar. According to preoperative data, compared with patients with late-onset T2DM, those with early-onset disease were heavier, had a larger waist circumference, and had higher fasting glucose levels, HbA_1C, C-peptide levels, and inflammatory markers, as well as dyslipidemia (Table 1). Use of insulin, HOMA insulin resistance index, and types of operative procedures were comparable between the 2 groups, but those with early-onset T2DM used less oral antidiabetic medication than those with late-onset disease. The early-onset group had a shorter operation time and longer hospital stay than the late-onset group (Table 2). Although the late-onset group had a significantly higher rate of minor morbidity than the early-onset group (31 [14.2%] vs 24 [7.1%]; P = .03), there were no significant differences in rates of major complications between the groups (3 [1.4%] vs 6 [1.8%]; P = .89). Anastomotic line or stapler line leakage occurred in 6 patients (5 [1.5%] with early-onset T2DM and 1 [0.5%] with late-onset disease); the leakage was diagnosed clinically and confirmed by computed tomography scan. All 6 patients recovered after reoperation. Two patients with major bleeding were treated with a blood transfusion and laparoscopy for hemostasis. No patients in this study died.

At 1 year, patients with early-onset T2DM achieved a larger reduction in BMI and waist circumference than did patients in the late-onset group (Table 3). Percentage of weight loss also was higher in the early-onset group, but it was not statistically significant. A significantly lower fasting glucose level and HbA_1C, greater reduction in HbA_1C and HOMA insulin resistance index, and better dyslipidemia control were observed in the early-onset group than in the late-onset group. Patients with early-onset T2DM also used less insulin and oral antidiabetic medication than did those with late-onset disease. The remission rate of patients with early-onset T2DM was significantly higher than in patients with late-onset disease (193 [56.9%] vs 110 [50.2%]; P = .02) (Table 3). There was no significant difference in the distribution of type of operation between the 2 groups, which can affect outcomes (P = .24) (Table 1) and follow-up (eFigure in the Supplement). Duration of T2DM (P < .001), age at operation (P < .001), and preoperative C-peptide level (P < .001) were independent predictors of remission. Direct association of remission with weight loss (P < .001) was observed in both groups. Higher rates of remission in the early-onset group persisted on multivariate analysis after adjusting for age at operation, duration of T2DM, use of antidiabetic medications, type of operative procedure, HbA_1c, C-peptide, highly sensitive C-reactive protein, and weight loss.

Five-year follow up data were available for 122 of 249 eligible patients (49.0%; 74 with early-onset T2DM and 48 with late-onset T2DM). Patients with early-onset T2DM had a greater percentage of weight loss and higher remission rate at 5 years than did those with late-onset disease (47 of 72 [65.3%] vs 26 of 48 [54.2%]; P = .04) (Table 4). Improvement in laboratory markers of control and remission of T2DM, as well as reduction in the use of insulin and antidiabetic drugs, supported the higher rates of remission in the early-onset group. Age at operation (P < .001) and duration of T2DM (P < .001) were independent predictors of remission. After adjusting for the above variables, multivariate analysis confirmed that the early-onset group had higher rates of remission. There were no reports of relapse of T2DM in our series, which could be owing to underreporting, as data from approximately one-fourth of patients were available for 5-year follow-up.

Duration of T2DM inversely affected the rate of remission at 1 year in both groups (P < .001 for both) (Table 5). The rate of remission at 1 year was highest among patients in both groups who underwent bariatric surgery less than 1 year after diagnosis (early-onset group, 105 of 115 [91.3%]; late-onset group, 38 of 50 [76.0%]). The rate of remission was lowest in patients in both groups who underwent bariatric surgery more than 10 years after diagnosis (early-onset group, 11 of 44 [25.0%]; late-onset group, 3 of 23 [13.0%]). A similar effect of disease duration on the rate of remission persisted at 5 years after surgery for the early-onset group (P = .003) and late-onset group (P = .14). The rate of remission was still higher at 5 years in the early-onset group than in the late-onset group on multivariate analysis after adjusting for duration of T2DM and other baseline characteristics. Multivariate analysis again confirmed that duration of T2DM (P < .001) was an independent predictor of remission at 5 years after bariatric surgery. Two-way analysis demonstrated that weight loss was inversely affected by duration of T2DM in those with early-onset disease at 1 year (P < .001) and 5 years (P = .002). A similar effect was observed in the late-onset group, but it was not significant. A direct association between the rate of remission and extent of weight loss was observed (P < .001).

The increasing prevalence of T2DM in youth and young adults is a significant societal health burden owing to the difficulty of treatment and the consequences of associated complications.^7,9,12 In this prospective cohort of patients who underwent bariatric surgery, preoperative data showed that patients with early-onset T2DM had poorer glycemic control and worse dyslipidemia than did patients with late-onset disease in spite of a similar duration of T2DM. These findings corroborated and highlighted the difficulties in managing early-onset T2DM.^5-7,9,12,13

The most important finding of this study is that bariatric surgery can provide better control of blood glucose and dyslipidemia in patients with early-onset T2DM than in those with late-onset disease. After surgical treatment, patients with early-onset T2DM had higher rates of complete and prolonged remission of T2DM than did patients with late-onset disease. To our knowledge, this is the largest longitudinal study to compare the long-term effect of bariatric surgery on lowering BMI in patients with early-onset T2DM vs those with late-onset disease. The overall 5-year 60% rate of prolonged, complete remission is compatible with previous reports.^29,30 Results from our study especially demonstrate the safety and efficacy of bariatric surgery in patients with early-onset T2DM.

The difficulty in treatment of early-onset T2DM can mainly be attributed to the physiological, mental, and social factors that may limit the efficacy of lifestyle management, such as diet and exercise.^12,13 On the other hand, dyslipidemia and chronic inflammation may cause the early manifestation of diabetic complications and premature death in patients with early-onset T2DM.⁵ A recent study demonstrated that, among patients with early-onset T2DM, obesity (BMI of 30-35) reduced healthy life-years by 11.8 years in men and 14.6 years in women and reduced life expectancy by 5.9 years in men and 5.6 years in women.³¹ Those with a BMI greater than 35 had even greater reductions in healthy life-years and life expectancy. Although bariatric surgery is recommended as the primary treatment for patients with T2DM and a BMI greater than 35, as it has an 80% rate of complete remission, mounting evidence has also suggested that bariatric surgery can be considered as a secondary treatment for mildly obese or overweight patients with T2DM.^18-21 The International Diabetes Federation has recommended that patients with T2DM should be conditionally eligible for bariatric surgery if they have a BMI above 30 (or 27.5 in Asian countries) and an HbA_1C level greater than 7.5% despite fully optimized conventional therapy, especially if their weight is increasing or other weight-responsive comorbidities are not achieving targets with conventional therapies.³² However, this treatment should be provided early after the diagnosis, as the success of bariatric surgery decreased markedly in patients whose duration of T2DM was more than 10 years.^33-35

Caution must be exercised in the selection of patients with early-onset T2DM who will benefit from bariatric surgery. It is important to differentiate between type 1 diabetes and T2DM in patients with early-onset disease. For preoperative testing, anthropometric measures and biochemical blood tests (fasting plasma glucose, HbA_1C, serum insulin, and lipids) may be helpful to differentiate type 1 diabetes from T2DM. In addition to these routine clinical evaluations and laboratory tests, measurement of autoimmune antibodies, such as anti–glutamic acid decarboxylase or anti–islet cell autoantibodies, should be performed for every patient with early-onset T2DM to rule out possible latent autoimmune adult diabetes. The HOMA index may confirm a diagnosis of T2DM. However, young patients with T2DM exhibit marked heterogeneity in terms of etiologic factors and presentation between different ethnic groups. The prevalence of childhood type 1 diabetes was 4.45 per 100 000 person-years in a white population but was only 2 per 100 000 person-years in an Asian population.^2,11,12 Conversely, the prevalence of childhood T2DM was 3 per 100 000 person-years in Asia.¹⁰ Incidence of T2DM in Asia was 4% to 6% in those aged 18 to 40 years, and patients with early-onset T2DM also experienced higher renovascular complications than did those with late-onset disease or those in an age-matched group with type 1 diabetes.¹¹

Another important consideration in selecting patients with early-onset T2DM for bariatric surgery is to select patients with preserved islet cell function (ie, normal C-peptide levels). In this study, we excluded patients with C-peptide levels less than 1.0 ng/mL. Higher rates of remission of T2DM are reported after bariatric surgery for patients with higher C-peptide levels, which was also supported by results of our study.^30,33-35

There are some limitations to this study. First, because it is an observational study, it is subject to selection bias. However, this study provides important evidence that warrants further studies on management of this particular group of patients. Second, the study population is Taiwanese patients of Asian ethnicity in a single institution. The results may not be generalizable for other ethnic groups and regions. Furthermore, even if the same principles may apply to other ethnicities, the selection criteria may vary with the characteristics of the cohort. Last, data on only 122 of 249 eligible patients were available for 5-year follow-up because, at the time of the study, it was less than 5 years since the remaining patients underwent surgery. A follow-up study at a later date with extended inclusion criteria using patients with available 5-year data instead of 1-year data will provide a larger amount of 5-year follow-up data from which to draw a conclusion with more strength regarding long-term remission.

Taiwanese patients with early-onset T2DM had poorer glycemic control and worse dyslipidemia than did patients with late-onset disease. However, patients with early-onset T2DM may benefit from bariatric surgery and, as a result, have a higher rate of remission of T2DM than patients with late-onset disease. Our results show that bariatric surgery is a viable option for the treatment of early-onset T2DM and that treatment early after diagnosis is preferred.

Accepted for Publication: March 21, 2016.

Corresponding Author: Wei-Jei Lee, MD, PhD, Department of Surgery, Min-Sheng General Hospital, No. 168, Chin Kuo Road, Tauoyan, Taiwan, ROC (wjlee_obessurg_tw@yahoo.com.tw).

Published Online: June 1, 2016. doi:10.1001/jamasurg.2016.1130.

Author Contributions: Dr W.-J. Lee had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: W.-J. Lee, Ser.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Aung, W.-J. Lee, S.C. Chen, Ser, Wu, J.-C. Chen.

Critical revision of the manuscript for important intellectual content: Aung, W.-J. Lee, Ser, Chong, Y.-C. Lee.

Statistical analysis: Ser, Y.-C. Lee.

Administrative, technical, or material support: W.-J. Lee, S.C. Chen, Ser, Wu, Chong, J.-C. Chen.

Study supervision: Ser, Chong.

Conflict of Interest Disclosures: None reported.