Introduction

Bariatric surgery is now universally recognised to be an effective treatment strategy for selected obese patients with type 2 diabetes mellitus (T2DM). In comparison, literature on the effect of bariatric surgery on type 1 diabetes, mellitus (T1DM) is somewhat limited. Most of the studies describing any experience with bariatric surgery in T1DM patients suffer with very small sample sizes and there are only three studies comparing the effect of bariatric surgery in T1DM patients with that in T2DM patients [1–3]. This is likely to become an even more important topic in the future with rising prevalence of obesity in patients suffering with T1DM. Approximately 12.6 % of youths with T1DM are obese [K] and up to half of the patients are either overweight or obese [4]. The obesity epidemic and intensive insulin treatment have been held responsible for it [4]. Moreover, it seems T1DM may occur at an earlier age in obese individuals with genetic predisposition and increases the risk of complications [5, 6]. Though typically type 1 DM presents in children and young adults, it can present at any age. There is another variant of type 1 diabetes mellitus called latent autoimmune diabetes in adults (LADA), which typically presents after the age of 30 [1, 7]. Currently, there is no systematic review in scientific literature on this topic. This review systematically examines published English language scientific literature on the effect of bariatric surgery on T1DM, in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines.

Methods

An online search of PubMed, Medline, Embase, and Google Scholar was independently carried out by two researchers using keywords like “bariatric surgery”, “metabolic surgery”, “gastric bypass” “Roux en Y Gastric Bypass (RYGB)”, “Mini Gastric Bypass”, “Single Anastomosis Gastric Bypass”, “Sleeve Gastrectomy (SG)”, “Gastric Banding”, “Biliopancreatic Diversion (BPD)”, “Duodenal Switch”, and “Type 1 Diabetes Mellitus” to identify all articles on the role of bariatric and metabolic surgery in patients with T1DM. Articles were also identified from references of relevant articles. Last of these searches was carried out on 25 June 2015.

A total of 22 articles were identified on the subject of bariatric surgery in type 1 diabetes mellitus. However, a number of them were excluded from cumulative analysis for various reasons. Some articles [8, 9] were excluded, as they did not describe any experience. Czupryniak et al. have published two papers [10, 11] on this topic. However, their latter paper [11] also includes the two cases described in the former [10]; the former study [10] was hence excluded. Similarly, Middlebeek et al. published two articles [12, 13]. However, most patients overlapped between the two studies. So, we included their latter study [13] in cumulative analysis, which also had one more patient than the former [12]. Animal studies and articles [14, 15] were also excluded from the systematic review. There was another study [16] where abstract had mentioned experience with T1DM patients but in reality the patients had T2DM and the description in abstract must have been in error. Finally, 15 articles were included in our cumulative analysis.

Three of these articles compared the results of bariatric surgery between T1DM and T2DM patients. Data from these articles were used to determine cumulative comparative experience in these two groups. We did not use any statistical comparison of data between type 1 and type 2 diabetes groups, as there will be significant risks of error in absence of raw data for any of these studies and their heterogeneous nature. Figure 1 gives a PRISMA flow chart for article selection.

Fig. 1
figure 1

PRISMA flow chart for article selection

Full size image

Results

This review identified a total of 15 studies describing experience with various bariatric surgical procedures in patients suffering with type 1 diabetes. Table 1 [1, 2–3, 13, 23, 17–26] lists qualitative characteristics of these studies.

Table 1 Experience with bariatric surgery in T1DM patients
Full size table

Cumulative Total Experience

A total of 15 studies describe bariatric surgery in 89 obese patients with T1DM over a period from August 2000 and May 2013. Out of these, 86.5 % (n = 58/67) were females. The mean age was 40.7 years (n = 67). The mean weight and BMI was 124 kg (n = 24) and 42.6 kg/m2, respectively. Roux-en-Y gastric bypass accounted for 70 % of the bariatric procedures (n = 62/89) in these patients. Sleeve gastrectomy (14.6 %, n = 13/89), biliopancreatic diversion (13.5 %, n = 12/89), and gastric banding (2.2 %, n = 2/89) accounted for the remainder. There was no mortality reported in any of these patients. Complications were reported in 9 patients but we cannot be certain of the denominator as most of these studies focussed on weight loss and diabetic outcomes. Weight loss and comorbidity resolution were satisfactory. It is not possible to provide cumulative numbers for weight loss and resolution in hypertension or dyslipidaemia because of significant variation in reporting methods in different studies. Total insulin requirement improved in almost all studies and requirement in terms of units/kilogram/day improved in 8 out of 15. However, two studies [3, 17] did not show any improvement in insulin requirement expressed per unit weight. Glycaemic control remained unaltered in most patients after surgery, improved in some, and worsened in a few.

Cumulative Comparative Experience

Three studies in this review compared the experience with bariatric surgery in T1DM and T2DM patients. Table 2 presents cumulative comparative experience. It is obvious from Table 2 that patients with T1DM have similar weight loss and comorbidity resolution but do not experience the significantly improved glycaemic control seen with T2DM patients.

Table 2 Comparison of cumulative quantitative data from studies comparing bariatric surgery in type 1 and type 2 diabetes mellitus patients
Full size table

Discussion

T1DM accounts for about 5–10 % of all cases of diabetes, and its incidence seems to be increasing worldwide [18–21]. It results from an autoimmune destruction of insulin-producing beta cells [22]. A large proportion of patients with T1DM suffer with metabolic syndrome and hypertension, and cardiovascular disease is now the leading cause of death in these patients [23, 24]. It is hence important that treatment strategies aimed at T1DM also address metabolic syndrome and cardiovascular disease risk factors. Bariatric surgery is recognised to be effective for both of these conditions [25, 26].

At the same time, T1DM is a different disease compared to T2DM. Whereas one could reasonably expect weight reduction and altered hormonal milieu, seen with bariatric surgery, to result in significant improvements in insulin resistance, patients with pure T1DM stand nothing to gain from these peripheral changes as the disease process mainly revolves around the pancreas. This has traditionally been the case with T1DM as patients were either normal or underweight. With the advent of obesity epidemic, the landscape of T1DM is changing and now approximately half of these patients are either overweight or obese. Bariatric surgery is recognised as one of the most effective long-term solutions for obesity.

Given that bariatric surgery can bring about real improvements in the lives of obese patients suffering with T1DM, it becomes important to understand the results with bariatric surgery in these patients. Over the last decade, a number of authors have reported on their experience with this group of patients. However, lack of a systematic review has made interpretation of data from small studies difficult. This prompted us to carry out the current review.

This review shows that total insulin requirement decreases in the majority of T1DM patients after bariatric surgery. Most studies also reported a significant reduction in insulin requirement expressed as units/kilogram but a few did not [3, 17]. These results are significant as the main problem in T1DM patients is impaired secretory capacity of beta cells and differences in pathogenesis compared to the commoner T2DM are widely recognised. Improved understanding of mechanisms underlying reduction in insulin requirement could open up newer areas for research into pathogenesis and management of diabetes mellitus (Reviewer 2, Comment 2, and Comment 4). Some of the potential factors responsible for this could be reduction in insulin resistance, improvement in function of remaining beta cell, preservation of beta cell mass, and increased hepatic insulin sensitivity. It is probable that just like patients with T2DM [27], there are multiple underlying mechanisms at play including weight loss [1], altered hormonal milieu, reduction in calorie intake, and reduction in pancreatic and liver triacylglycerol. Preservation of beta cell mass and function may be more important in those with LADA, which was seen in three patients (n = 3/89, 3.3 %) in this review [1, 28, 29]. Moreover, it is possible that some of these patients have “Double Diabetes” on the basis of recent findings that that some T1DM patients actually suffer with both T2DM and T1DM [30, 31]. These individual have both insulin resistance (due to obesity) and evidence of autoimmunity against beta cells. It is suggested that obesity may “accelerate” beta cell destruction in such genetically susceptible individuals.

Dirksen et al. [32] noticed significant improvement in insulin requirement within a week of surgery, which was associated with marked increases in postprandial glucagon-like peptide 1 (GLP-1) and peptide YY (PYY) secretions. Same authors also noticed a brisk rise in glucose reaching maximal concentration 1-h postprandial indicating faster glucose absorption. Authors suggested that treatment with a faster acting insulin analogue before meals may be better than regular human insulin to control postprandial hyperglycaemia. Authors wondered if GLP-1 response could even lead to regeneration of some beta cells in the long term. Though there is some basis [33–35] for such a theory, convincing evidence is lacking. Blanco et al. [3] found in their comparative study of the effect of RYGB on T1DM and T2DM patients that despite statistically “not different” effect on GLP-1 and glucagon response between the two groups, glycaemic control in T1DM group was significantly poorer than that in T2DM patients. At the same time, studies have shown that T1DM patients treated with GLP-1 analogues achieve better glycaemic control at lesser insulin dosages [36, 37]. These finding could indirectly support a role for hormonal factors.

Hypoglycaemic episodes have been seen after bariatric surgery by a number of authors in this review [2, 17]. Altered glucose kinetics and difficulty in insulin dosing are probable contributory factors [17]. Some authors believe that sleeve gastrectomy may be a more attractive procedure for patients with T1DM because of more predictable carbohydrate absorption [38]. However, it is worth remembering that hypoglycaemic episodes are also seen with intensive insulin management of T1DM [4].

Weight loss seen in these patients seems to be on expected lines. Even though these patients continue to need some insulin, it does not seem to have a significant impact on weight loss. This was further evident from the three studies [1–3] that compared results in these patients with those seen in T2DM patients. The same is true of comorbidity resolution. Most of the studies in this review have shown satisfactory resolution/improvement of comorbidities.

There are voices [14] suggesting that we might even consider undertaking “clinical studies that address the safety, tolerability, efficacy, and durability of such surgical procedures in non-obese patients with T1DM.” Others [22] feel bariatric surgery should be considered early in the course of T1DM to decrease pancreatic beta cell damage. This may have some appeal as intensive insulin treatment typically results in weight gain with consequent worsening of Insulin resistance, dyslipidaemia, and hypertension. Given the fact that a large number of T1DM patients are now obese, it has been suggested that strategies for weight control, including bariatric surgery, should form part of management strategy for these patients much like those with T2DM or even general population [4]. However, these patients will need close, perhaps lifelong, monitoring by an endocrinologist postoperatively because of the continued need for insulin treatment and the need for close titration of insulin regime to decrease the frequency of severe hypoglycaemic episodes.

Much has happened in the field of surgical management of diabetes since Pories first reported on the effect of bariatric surgery in patients with T2DM [39]. Progress has been somewhat slower in T1DM patients. Many authors argue that it is more common in bariatric surgery cohorts than has been appreciated thus far accounting for some of the failure of remissions [9]. In this review, we found many patients [38] were diagnosed with T1DM postoperatively as they were suffering with undiagnosed LADA at the time of surgery. LADA accounts for 2–12 % of cases of diabetes [40] and 3.3 % (n − 3) patients in this review. Prevalence of LADA in adults presenting with non-insulin-dependent diabetes is approximately 10 % [41].

It is currently unclear if this should prompt more routine investigations for evidence of autoimmunity and c-reactive peptide levels in patients considering bariatric surgery. Though the additional cost implications may not be justified for patients who would benefit from bariatric surgery as such on the grounds of their weight or comorbidities, screening for LADA and Double Diabetes may be appropriate for type 2 diabetics seeking pure metabolic surgery. This will help us target metabolic surgery for T2DM and improve its accuracy and chances of success, some of the biggest challenges facing metabolic surgeons. For all other insulin-dependent patients, it may make more sense to simply monitor diabetic status more closely in the earlier postoperative period with a lower threshold for investigations in patients with difficulties.

This review only includes obese patients with T1DM, who would as such qualify for bariatric surgery on the grounds of their weight and associated comorbidities. Our systematic review sheds light on what these patients can expect in terms of weight loss, comorbidity resolution, and diabetic control. As we can see, this review shows satisfactory weight loss and comorbidity resolution in obese type 1 diabetics undergoing bariatric surgery. In studies comparing outcomes in these patients with those seen with type 2 diabetics, weight loss and comorbidity resolution were similar in two groups. Glycaemic control however did not improve to the extent seen with type 2 diabetics. Bariatric or metabolic surgery is increasingly being recognised for even patients with BMI < 35 kg/m2 and patients with T2DM [42–44] (Reviewer 2, Comment 4). This review is not qualified to comment on the role of surgery for non-obese type 1 diabetics but one suspects it will offer limited advantages.

A number of surgical procedures have been used by surgeons in this group of patients but RYGB accounts for the majority of the procedure. This review cannot make any recommendation on the ideal bariatric procedure for obese patients with T1DM. There are other obvious weaknesses of this review that need to be recognised. There is no level 1 or 2 data available in scientific literature on this topic comprising the strength of the data in this review. Most of the studies describe retrospective experience with a small cohort of patients. Time is now ripe for a well-designed randomised controlled trial assessing the efficacy of adding bariatric surgery to intensive insulin management for selected obese patients with T1DM.

Conclusion

Obese T1DM patients can expect significant weight loss, comorbidity resolution, and reduction in insulin doses with bariatric surgery. Surgery does not however result in improved glycaemic control in a significant proportion of patients.