The care of patients with type 2 diabetes mellitus (T2DM) should focus not only on glucose control but also on management of complications such as heart disease, dyslipidemia, obesity, and cancer. However, the prevalence of adult diabetes has been rising globally [1,2] and the economic and social burdens of diabetes threaten the world's health systems [3,4].

It has been demonstrated that metabolic surgery resolves T2DM in morbidly obese patients [5-7]. In contrast, whether metabolic surgery is effective in nonmorbidly obese patients still remains to be solved. Moreover, Asian people develop T2DM at a younger age with a lesser degree of obesity, and tend to have the complications and die earlier [8,9].

As life expectancy is increased, the number of patients with gastric cancer and T2DM is increased as well. It was noticed that gastric resection induce the resolution of T2DM in non-morbidly obese patients [10-12].

Recently, we studied the outcome of T2DM after gastrectomy and conventional reconstruction in non obese gastric cancer patients [13]. The study was a large-series retrospective study including about 400 patients and only 15.1% were resolved from T2DM after operation, which were not satisfactory. This prompted us to develop a effective reconstruction method for diabetic control after gastrectomy in gastric cancer patients with non obese T2DM. We grafted the bariatric type operation, long-limb Roux-en Y reconstruction into gastric cancer surgery. This study aimed to investigate the safety and efficacy of long limb Roux-en Y reconstruction after gastrectomy in non obese gastric cancer patients.

A total of 15 non obese T2DM patients with gastric cancer were enrolled. There were 10 males and 5 females and their mean age was 62.1 ± 8.7 years. The baseline demographic data are presented in Table 1. Four patients received total gastrectomy and 11 patients received subtotal gastrectomy. Eleven patients received laparoscopic operation.

The mean surgical time was 265.5 ± 84.8 minutes. There was no surgery-related mortality, but four patients experienced complications; one experienced anastomotic stricture that was recovered with conservative management, one experienced anastomosis site kinking managed with temporary stent insertion and one had myocardial infarction treated with percutaneous transluminal coronary angioplasty with stent insertion. There was one wound complication. There were no patients who were complaints of Roux stasis or A-loop syndrome. Final pathologic results revealed that nine patients were early gastric cancer and six patients were advanced. Four patients received chemotherapy. During the follow-up period, there was one patient who had poorly differentiated neuroendocrine carcinoma and died due to recurrence four months after surgery, but the blood glucose level was well controlled (HbA1c < 6.0%) without medication. In the remained patients, there was no evidence of recurrence in the regular check-up at 6 months after operation. Also, there was no stricture or jejunal ulcer in the follow-up endoscopic examination.

The patient who died of recurrence was excluded from the efficacy evaluation. Six months after surgery (n = 14), the mean BMI was significantly decreased to 21.7 ± 3.1 kg/m² from 25.2 ± 3.4 kg/m² (P < 0.05). The preoperative mean HbA1c was 7.7 ± 1.4% with antidiabetic medications and the mean HbA1c was also significantly down to 6.3 ± 0.8% (P < 0.05).

At the end of the study, the follow-up duration was 12.5 ± 5.5 months (6.0 to 21.7 months). HbA1c < 6% without any antidiabetic medications were observed in 11 patients (78.6%). A total of four patients who received total gastrectomy and seven patients (70%) who received subtotal gastrectomy were included. All of four patients who received chemotherapy discontinued anti-diabetic medication after operation. Although improvement of insulin sensitivity was unstable, insulin secretion was significantly improved (P < 0.05) (Table 2). The FPG levels of 2 patients (14.3%) were between 6.1 and 7.0 mmol/L and their HbA1c were well controlled (between 6% and 7%) with or without medications. Only one patient had HbA1c above 7%, but the FPG and 2-hour plasma glucose levels were significantly decreased after surgery. There were no patients who had anemia, and/or malnutrition after surgery. Even though the albumin levels were decreased at the end of study compared to the preoperative levels, they were still within normal range. Also, the levels of serum triglyceride and serum cholesterol decreased after operation, but remained in normal range.

Calorie restriction and/or weight loss are one of many effective ways to control T2DM in morbidly obese patients. However, calorie restriction and/or weight loss are not easy in non obese patients, and Asian people develop T2DM with a lesser degree of obesity.

The early recovery course and mismatched degree of T2DM improvement after gastric bypass compared with equivalent weight loss from other treatment modalities implicate weight-independent effects of bariatric surgery on glucose metabolism [6,19]. In addition, some reported that non-morbidly obese patient with T2DM can be recovered after bariatric surgery [20-25]. These results suggested that surgery might induce remission even in non obese T2DM patients.

We had studied the long-term results of gastrectomy and conventional reconstruction in non obese T2DM patients with gastric cancer [13]. It was a very large scale study and the complete remission was only 27.3% even though in total gastrectomy and Roux-en Y esophagojejunostomy. There were some reports that demonstrate the effect of subtotal gastrectomy and Roux-en Y gastrojejunostomy on T2DM in non obese patients, but the remission rate was still not high [12]. Before we started this study, we tried some modifications of Roux limb length after gastrectomy and were not able to control glucose homeostasis without medications in gastric cancer and T2DM patients. In this study, we tested the safety and efficacy of Roux-en Y reconstruction with a 100-cm biliopancreatic limb and a 100-cm Roux limb after gastrectomy in T2DM gastric cancer patients. The main difference with previous reports was the longer length of biliopancreatic and Roux limbs.

Among 15 patients enrolled in this study, 9 patients received laparoscopic surgery with two anastomotic stricture or kinking. We assumed that this complication was related with technical failure in the learning period.

Rearrangement of the intestine and/or modification of the stomach might induce functional obstruction of the small bowel and metabolic problems such as vitamin and trace element deficiencies [26]. Longer length of a Roux limb has been suggested to be effective in super obese patients without significant complications [27]. In this study, we found that there was no Roux stasis syndrome, anemia and hypoalbuminemia after surgery. Some patients complained of frequent diarrhea after eating fatty food, but normal diet was tolerable without malnutrition. In addition, most patients maintained their weight within normal range after surgery.

Mechanisms of antidiabetic effects after gastrointestinal surgery are not yet fully characterized. Except effects due to weight loss and calorie restriction, exclusion of the proximal small intestine from ingested nutrient, enhanced distal-intestinal nutrient delivery, impaired ghrelin secretion, modulations of intestinal nutrient sensing and regulation of insulin sensitivity, bile acid perturbations, changes in gut microbiota and other changes were proposed [19,28]. Our study showed that all of patients who received total gastectomy and/or chemotherapy stop the anti-diabetic medication with normal glucose level and these might result from weight loss or calorie restriction. However, seven among ten patients who received subtotal gastrectomy without chemotherapy demonstrated normal glucose control after operation. In this study, bypassed segments of the small bowel were longer than conventional Roux-en Y reconstruction after gastrectomy. Also, nutrient stimulation of lower intestinal hormones (e.g., GLP-1) might be enhanced in the patients of this study. It was beyond the scope of this study to elucidate the mechanism by which T2DM was resolved. However, insulin secretion was significantly improved after surgery. We were afraid that the β cell function in non obese T2DM patients was decreased and such decrease might be irreversible. Fortunately, animal experiments demonstrated that T2DM in lean T2DM rats can be improved by rearrangement of the gastrointestinal tract without significant weight loss [29,30]. As mentioned above, these results are consistent with some human trials [20,21,24,25]. Now, we suggest that even when weight is maintained within the normal range, T2DM is still reversible in non obese patients. In this study, the improvement of insulin sensitivity was not clear. A long term follow-up study with increased enrolled patients is warranted to elucidate the improvement of insulin sensitivity.

Taken together, we considered that long limb Roux-en Y reconstruction is feasible and has the potential to cure T2DM in gastric cancer patients. This was a pilot study and the duration of follow-up was short. To confirm the effectiveness of long limb Roux-en Y procedure, randomized controlled trial (RCT) is needed and our results will be helpful to design the RCT.