Obesity Surgery

, 18:1035

A Comparison of a Personal Series of Biliopancreatic Diversion and Literature Data on Gastric Bypass Help to Explain the Mechanisms of Resolution of Type 2 Diabetes by the Two Operations


    • Department of SurgeryUniversity of Genoa Medical School—Azienda Ospedaliera, Universitaria San Martino
    • Dipartimento di ChirurgiaAzienda Ospedaliera Universitaria San Martino
  • Francesco Papadia
    • Department of SurgeryUniversity of Genoa Medical School—Azienda Ospedaliera, Universitaria San Martino
  • Giovanni Camerini
    • Department of SurgeryUniversity of Genoa Medical School—Azienda Ospedaliera, Universitaria San Martino
  • Giuseppe Marinari
    • Department of SurgeryUniversity of Genoa Medical School—Azienda Ospedaliera, Universitaria San Martino
  • Dario Civalleri
    • Department of SurgeryUniversity of Genoa Medical School—Azienda Ospedaliera, Universitaria San Martino
  • Adami Gian Franco
    • Department of SurgeryUniversity of Genoa Medical School—Azienda Ospedaliera, Universitaria San Martino

DOI: 10.1007/s11695-008-9531-x

Cite this article as:
Scopinaro, N., Papadia, F., Camerini, G. et al. OBES SURG (2008) 18: 1035. doi:10.1007/s11695-008-9531-x



Roux-en-Y gastric bypass (RYGBP) and biliopancreatic diversion (BPD) are highly beneficial operations for type 2 diabetes mellitus (T2DM) in obese patients, leading to complete T2DM resolution in 75–90 and 97–99% of cases, respectively. In both RYGBP and BPD, the foregut is excluded from the food stream and the distal small bowel receives the food stimulation, while following BPD fat intestinal absorption is also extremely limited. This study was carried out to identify clinical features that could give insight on the different mechanisms of action on diabetes resolution.


The files of 443 severely obese patients with T2DM undergoing BPD from May 1976 to May 2007 were examined, and the presence of T2DM (fasting serum glucose >125 mg/ml) at 1–2 months, at 1 year, at 10 years, and at ≥20 years following the operation was recorded.


The percentage of patients cured (fasting serum glucose reduced to ≤110 mg/dl, on free diet and with no therapy) was 74% at 1 month, 97% at 1 and 10 years, and 91% at ≥20 years, the 26% of uncured patients at 1 month being those with most severe preoperative T2DM.


As the early results after BPD resemble those reported after RYGBP, it can be hypothesized that the duodenal exclusion and the distal small bowel stimulation are the first mechanisms acting in BPD, immediately after the operation, that only subsequently the myocellular fat depletion, which cannot be immediate, takes over, and that the minimal fat absorption is the mechanism accounting for the long-term results of BPD.


Morbid obesityObesity surgeryBiliopancreatic diversionRoux-en-Y gastric bypassType 2 diabetes mellitus


The beneficial effect of biliopancreatic diversion (BPD) on type 2 diabetes mellitus (T2DM) is well known. Recently, we published 237 cases of preoperatively type 2 diabetic obese patients who showed at 10 years after BPD full resolution (normoglycemia, no medication, free diet) in 98% of cases [1]. The same figure was reported by Buchwald in his meta-analysis [2], referring to both BPD and BPD with duodenal switch. Marceau [3], using the latter operation, found that 96% of formerly diabetic patients no longer needed treatment after operation. Besides a nonspecific benefit from weight reduction, diabetes resolution is due to specific effects independent of weight changes, this being proved by the very early normalization of insulin action, when body weight is still high [4].

The other bariatric operation that shows a specific effect on T2DM is Roux-en-Y gastric bypass (RYGBP), which was reported to result in resolution of T2DM in 74–89% of cases [2, 510]. The two operations have a common anatomico-functional attribute, represented by foregut exclusion from the food stream and gut hormone release by the distal small bowel due to stimulation by food [7, 1113], while BPD has one more specific functional feature, consisting of the extremely limited fat absorption [14]. It can be hypothesized then that the different success rate of the two procedures is accounted for by this additional BPD specific action.

The foregut bypass and the ileal stimulation would cause diabetes resolution in about 75–90% of cases and amelioration in the others as a consequence of gut-hormonal changes, which are differently reported, even if great importance is attributed by all authors to the postoperative change of incretion and action of the glucose-dependent insulinotropic polypeptide (the former “gastric inhibitory polypeptide” or GIP) [15, 16]. Other authors give equal or greater importance to the postoperative increase of glucagon-like peptide 1 (GLP-1), which is denied by some, and ignored by other authors [1521]. Whatever the effect of foregut and hindgut hormonal changes, this anatomo-functional condition is shared by BPD, which has its additional fat-absorption limiting action. The latter would act by causing lipid depletion of the target cell, particularly the muscular one, with consequent return to glucose utilization as the energy source, and thus disappearance of insulin resistance [22, 23]. Some useful information on the importance and action timing of the two mechanisms mentioned above can be derived from data obtained by us in BPD patients where the limited fat absorption was not yet or no longer present.

Materials and Methods

The study was retrospectively carried out on the charts of the 3,020 obese patients undergoing BPD at the Department of Surgery of the University of Genoa School of Medicine from May 1976 to May 2007.

Among the whole cohort, 443 patients were considered as having T2DM (fasting serum glucose concentration >125 mg/dl) prior to the operation: there were 171 men and 272 women, with age from 14 to 68 years (mean 42), mean body weight (BW) 135 kg (ranging from 82 to 236) and mean body mass index (BMI) 49.9 kg/m2 (30.2 to 87.7). Postoperatively, T2DM was considered as resolved when fasting serum glucose (FSG) level was <110 mg/dl on free diet with no medication.

For each patient, the time in years between T2DM diagnosis and operation and the use of antidiabetic medication was recorded: at the time of BPD, 20 patients were receiving insulin, and 79 patients were on oral antidiabetic agents, while in the others, diabetes was diagnosed at hospitalization for bariatric surgery before operation.

The FSG mean values and the number of patients who were still diabetic among those available for follow-up were considered at the following times: (1) prior to the operation; (2) at 1–2 months after operation, when BW values were still in the morbidly obese range; (3) at 1 year after operation, when BW was close to that of stabilization and food intake was equal to or greater than preoperatively; (4) at 10 years after operation; (5) at ≥20 years after operation. Because of the very large size of the sample, data were analyzed with parametric statistics: the differences between FSG mean values were assessed by Student’s t test for paired data; multiple logistic regression was used to evaluate the relation between data, gender, age, current BW and duration of T2DM being the independent variables, and the normalization of FSG at 1–2 months following the operation the dependent variable. The preoperative use of antidiabetic agents was dummy-coded (0 = no therapy, 1 = oral antidiabetics, 2 = insulin), and data were entered in the regression model as independent variable.

The 12 patients with preoperative T2DM among the 77 who for any reason underwent elongation of the common limb were also considered.


At 1–2 months after BPD, in three quarters of the preoperatively T2DM patients, FSG fell within the normal limit, and a sharp and highly significant reduction of mean values was observed (Table 1). A further significant reduction of mean FSG was found at 1 year following the operation, and the data remained substantially unchanged 10 and ≥20 years thereafter. One year after BPD, FSG was normal in the near totality of cases and this T2DM resolution rate was unchanged at 10 years and essentially maintained in the very long term.








Prior to BPD

135.0 ± 26.0

49.9 ± 9.15

179 ± 60

443/443 (100%)

443/443 (100%)

1–2 months after BPD

116.3 ± 21.3

43.0 ± 8.4

110 ± 31*

383/443 (86%)

95/383 (25%)

1 year after BPD

89.0 ± 19.9

32.9 ± 7.2

85 ± 20

363/443 (82%)

10/363 (2.8%)

10 years after BPD

86.5 ± 18.8

32.3 ± 7.0

90 ± 26

293/407 (72%)

8/293 (2.7%)

≥20 years after BPD

84.5 ± 20.9

29.6 ± 6.8

93 ± 33

34/56 (61%)

3/34 (8.8%)

*p < 0.001 vs. prior to BPD

§p < 0.01 vs. 1–2 months after BPD

The multiple logistic regression model showed that the early normalization of FSG is predicted by the preoperative T2DM treatment, the early normalization being more probable in patients who preoperatively were not treated (chi-square 19.264, p < 0.001, confidence interval 1.9–3.357). On the contrary, age, gender, current BW and duration of T2DM were found unrelated to early diabetes resolution. However, if preoperative treatment is not entered in the multiple regression, T2DM duration becomes a predictive variable (chi-square 11.91, p < 0.006, confidence interval 1.05–1.120).

Because at longer term FSG fell within the normal range and remained unchanged in the nearly totality of subjects, the search for a predicting factor for diabetes resolution would obviously be meaningless.

Ten of the 12 patients with preoperative T2DM who underwent elongation of the common limb maintained full diabetes resolution. Two of them (17%) had relapse of diabetes 8 and 14 years after elongation, respectively.


As mentioned in the introduction, our hypothesis was that the beneficial effect of BPD on T2DM is explained both by the gut hormonal changes caused by the food stream being excluded from duodenum and allowed to reach the distal small bowel and the minimal fat absorption with consequent cellular depletion. The minimal fat absorption is witnessed by our study on intestinal absorption [14], while we could demonstrate the intramyocellular fat depletion directly by means of MR spectrophotometry [23]. The reason why we consider that the minimal fat absorption is a specific action of the BPD is the fact that the operation allows a mean maximum daily fat absorption of <40 g, a fat intake that is not obtainable by any other means. In fact, any chronic diet containing such a low quantity of fat could hardly supply enough protein for survival.

If the beneficial effects other than weight loss after BPD are considered at the minimum follow-up time after which no further substantial modifications were observed, this time is 1 month for hypercholesterolemia, and 1 year for T2DM [24]. This means that 1 year is required to reach the maximum effect on T2DM, i.e., a nearly 100% resolution rate (Table 1). The percentage of patients cured is 74 at 1 month, 97 at 1 and 10 years, and 91 at 20 years. In reality, only a minority of the patients referred to in Table 1 as uncured could be considered true diabetes relapse. In fact, many of these patients showed, especially at long term, a condition of mild hyperglycemia, ranging from 110 to 160 mg/dl, which was maintained on free diet and with no therapy, for many years (up to >10), without any tendency to evolve towards a frank diabetic status. This phenomenon, which was observed in a large percentage of the T2DM patients who underwent BPD with a BMI < 35, was already discussed in a previous article [25] and deserves further investigation.

The ∼25% of uncured patients at 1–2 months were those with the most severe preoperative diabetes. As a matter of fact, the short-term resolution, which can reasonably be assumed to be inversely related to preoperative diabetes severity, was related to the duration of diabetes itself and, even more, to the preoperative specific antidiabetic treatment. Indeed, the ∼75% of patients who were diagnosed at operation, even if from an epidemiological point of view they are considered part of the general T2DM population who are unaware of their disease, they were necessarily in an early stage of diabetes, and thus much less severely ill than the aware population. The early results of a pilot study that we are carrying out in T2DM patients with BMI < 35 kg/m2 submitted to BPD show at 1 month only 30% of full resolution of diabetes, thus resembling much more those of the aware than those of the unaware patients in this study population. This is not surprising, considering that the patients in the pilot study are recruited in an obviously aware and treated population.

As the early results closely resemble those reported after RYGBP [9, 11], what happens in our opinion is that the foregut exclusion and ileal food stimulation are the first mechanisms acting in BPD, immediately after the operation, and that only subsequently the myocellular fat depletion, which cannot be immediate, takes over. The question to be answered remains if this latter mechanism accounts for the entire effect or if the sum of the two actions is necessary to obtain the 100% result.

Moreover, again, ∼20% relapse was observed after elongation of the common limb, when fat absorption is necessarily increased, and this could also be interpreted as if, the reversal of insulin resistance due to minimal fat absorption having disappeared, the only antidiabetic mechanisms left were those of the foregut and hindgut hormonal changes, resulting in ∼80% resolution rate. However, the long time elapsed after elongation, and particularly the fact that hyperglycemia appeared many years after the completion of post-elongation weight regain, suggests as a more likely explanation a long-term exhaustion of beta-cell function that even if favored by the increased insulin resistance possibly consequent to the increased fat absorption would have occurred anyway.

The immediacy of the T2DM amelioration or resolution that follows both BPD and RYGBP strongly suggests that either a nervous or a gut hormonal mechanism takes place. The gut hormones possibly involved in this phenomenon most probably include GLP-1, whose incretion has been found increased after both BPD and RYGBP, and which has a beneficial effect on beta cell function [2630]. Regarding the long-term results, the myocellular fat depletion mechanism is apparently still fully active. The question of the relative importance of the two above mechanisms at long term is still open. However, the vast majority of reports dealing with cure of T2DM after RYGBP have a relatively short follow-up, while the scarce literature reporting long-term results refer to data which is ambiguous, inaccurate, or based on a very low follow-up rate. This could suggest that the minimal fat absorption is the only or the main mechanism accounting for the long-term results of BPD.

However, it cannot be excluded that the better long-term results of BPD are partly due to a prolonged and increased beneficial effect on beta-cell function from persistent stimulation of hindgut hormones, especially GLP-1. However, after RYGBP, the food stimulation of the distal small bowel, due to the increased transit speed allowing food to reach distal ileum, could easily reduce or even disappear following the intestinal adaptive phenomena. On the contrary, after BPD, the stomach empties directly into the ileum, which would provoke distal gut hormone stimulation stronger and longer lasting than that one caused by RYGBP. This should result in, besides an increased insulin production capacity by the existing beta cells, a true increase of beta-cell mass due to the trophic action of GLP-1.

Copyright information

© Springer Science + Business Media, LLC 2008