Key words bariatric surgery - diabetes - GI hormones - insulin sensitivity - insulin secretion
Introduction
The increasing epidemic of obesity and type 2 diabetes (T2DM) is long recognized as
a major healthcare problem. Bariatric surgery has emerged as the most successful therapeutic
option for morbid obesity, since it results in remarkable and sustained weight loss
and a dramatic improvement of glucose control in patients with T2DM [1 ]. The improvement/resolution of T2DM is associated with the extent of weight loss
and the type of surgery ranging from 55% after restrictive procedures to 95% after
malabsorptive interventions [2 ]. The beneficial effects on glucose metabolism occur early after surgery, before
any substantial weight loss, suggesting a role of weight loss-independent mechanisms,
possibly related to changes in gastrointestinal (GI) hormones in response to ingested
nutrients [3 ]
[4 ]
[5 ]. There is now ample evidence that both obesity and T2DM are associated with impaired
secretion/action of GI hormones, namely glucagon-like peptide 1 (GLP-1) and glucose-dependent
insulinotropic polypeptide (GIP) [6 ] and that the restoration of a more physiologic GI hormone profile brought about
by bariatric procedures may contribute to the improvement of glucose homeostasis.
Roux-en-Y gastric bypass (RYGB), one of the most widely performed bariatric techniques,
creates a gastrojejunal anastomosis so that ingested food moves to the distal small
intestine bypassing the duodenum and the proximal jejunum. For these characteristics,
it is considered a mixed-malabsorptive procedure. RYGB has proven its clinical efficacy
in obese patients with type 2 diabetes for more than 30 years, with a diabetes remission
of ~75% [1 ]. Vertical sleeve gastrectomy (VSG) is a more recent procedure that involves the
excision of most of the stomach while maintaining intestinal anatomy. It is considered
a merely restrictive procedure although it is associated with marked reduction of
plasma ghrelin concentrations due to the removal of the gastric fundus. VSG results
in remarkable weight loss and metabolic improvement, reaching a rate of T2DM remission
similar to that of RYGB, but without the potential complications inherent to intestinal
bypass procedures [7 ].
Previous studies have examined the effects of the 2 procedures with regard to diabetes
remission or the changes in GI hormones, but limited information is available on the
contribution of intervention-specific changes in GI hormonal pattern to the remission
of diabetes. To this end, in this prospective study we compared the changes in insulin
sensitivity, insulin secretion, and post-meal GI hormone levels in obese patients
with T2DM 1-year after VSG or RYGB, to evaluate the hormonal and metabolic mechanisms
involved in weight loss and remission of T2DM.
Materials and Methods
Selection and description of participants
The study group included 33 obese patients with T2DM (M/F: 14/19; mean age: 46±9 years,
BMI: 44±8 kg/m2 ), who were on a waiting list for bariatric surgery. Inclusion criteria included:
age 30–65 years, body mass index (BMI)>40 or≥35 kg/m2 with uncontrolled T2DM under medical treatment, no contraindications to VSG or RYGB.
The choice of surgical procedure was made by the patient together with the surgeon
after a full explanation of the risks and benefits of each procedure. In total, 14
subjects underwent RYGB and 19 subjects underwent VSG. All participants were examined
by a multidisciplinary and integrated medical team consisting of a diabetologist,
a bariatric surgeon, a psychiatrist, and a dietician. The clinical and metabolic evaluation
of participants were conducted at the Department of Clinical Medicine and Surgery
of Federico II University while surgical procedures were performed at the Department
of Surgery, San Giovanni Bosco Hospital, Naples.
Antidiabetic treatment was oral hypoglycemic agents (OAD) in 24 patients, combined
OAD plus bedtime insulin in 5 patients and diet alone in 4 patients. None was on multiple
insulin injection regimen. Fourteen patients (74%) in VSG and 9 (64%) in RYGB were
on antihypertensive drugs; 5 patients (36%) in the RYBG group and 3 patients (16%)
in the VSG group were on hypolipidemic treatment.
The protocol was approved by the local Ethics Committee; all patients were informed
of the risks and benefits of each procedure and provided written, informed consent
before the study.
Study design
Before and one year after the bariatric procedure, anthropometric, clinical and routine
laboratory parameters were collected together with data on medication use. On both
occasions, in the morning after a 12-h overnight fast, a standard glucose tolerance
test (OGTT, 75 g of glucose) was performed to evaluate insulin secretion and insulin
sensitivity. The day after, a mixed-meal test (MMT) was performed to evaluate GI hormonal
response. The week before the tests all patients consumed a standardized hypocaloric
diet (1 200 Kcal) containing 52% CHO, 18% protein, 30% fat. To avoid possible confounders,
OAD were withheld 2 days before the MMT, while long-acting insulin was discontinued
for 24 h.
Mixed meal test (MMT)
The liquid mixed meal (Resource® ENERGY Nestle Nutrition, 304 kcal in total), containing 41 g carbohydrate (glucose),
13 g protein, and 9 g fat, was consumed within a maximum of 15 min. Blood samples
were drawn through an indwelling cannula at times 0, 30, 60, 90, 120, and 180 min
for the measurement of glucose, insulin, active GLP-1 and total GIP concentrations
at 0, 60, 120, and 180 min for the measurement of total ghrelin. Blood samples were
collected in BD Vacutainer® EDTA Aprotinin Tube contained K3EDTA (1.6 mg per ml blood) and Aprotinin protease
inhibitor (50KIU per ml blood) and immediately centrifuged at + 4°C and 3 000 rpm.
Plasma samples were stored at − 80°C and rigorously kept at + 4°C during assay. The
collection of blood samples with EDTA/Aprotinin under cooled conditions was appropriate
to maintain GLP-1 and ghrelin stability as described by Di Marino et al. [8 ] and Tvarijonaviciute et al. [9 ].
Operative techniques
All operations were performed laparoscopically by the same surgery team, as previously
described [10 ]. There were no major intra-operative complications or conversions to laparotomy.
Analytical procedures
Plasma glucose concentration was measured by the glucose oxidase method. Plasma insulin
and C-peptide were determined by ELISA. Plasma lipids were measured by Roche Cobas
analyzer using a colorimetric assay. HbA1c was measured by high-performance liquid
chromatography [Diamat HPLC, Bio-Rad Laboratories (Canada) Ltd., Mississauga, Canada]
[11 ].
Active GLP-1 was assayed by a nonradioactive, highly specific sandwich ELISA method
(Merck-Millipore, Darmstadt, Germany) with 100% cross-reactivity with 7–36 amide and
7–37 glycine-extended, but no reactivity with 9–36 amide and 9–37 glycine-extended
GLP-1 isoforms, GLP-2 or glucagon. Total GIP was assayed by a nonradioactive highly
specific sandwich ELISA method with (Merck-Millipore, Darmstadt, Germany) 100% cross
reactivity to human GIP (1–42) and GIP (3–42). Human total ghrelin (both intact and
des-octanoyl forms) was assayed by a nonradioactive, highly specific sandwich ELISA
method (Merck-Millipore, Darmstadt, Germany) with 100% cross-reactivity with des-octanoyl
human ghrelin, 80% human ghrelin (active), and 70% canine ghrelin (active). The intra-
and interassay coefficient of variation for the GLP-1 assay was<5% and for GIP and
ghrelin assays was<10%.
Measurements
Weight loss was expressed as% change in BMI and as percent excess weight loss (%EWL)
calculated by the following formula: (preoperative weight – current weight)/(preoperative
weight – ideal weight)×100 [12 ]. Insulin sensitivity and insulin secretion indexes were derived from glucose, insulin,
and C-peptide values measured every 30 min for 3 h during the OGTT. Insulin sensitivity
was assessed as the oral glucose insulin sensitivity (OGIS), which has been validated
vs. the hyperinsulinemic euglycemic clamp demonstrating a good correlation between
the 2 measurements of insulin sensitivity [13 ]. Insulin secretion as total amount of the hormone released by the beta cells (ISR)
was calculated from C-peptide with the deconvolution method [14 ]. Beta-cell function, which reflects the release of the hormone normalized to the
glycemic stimulus, was assessed as “early” (IGI30 =ratio between incremental C-peptide concentration and incremental glucose concentration
at 30 min) and “total” insulinogenic index (IGItotal =AUCCpeptide /AUCGlucose ). The interplay between insulin sensitivity and secretion, that describes the beta-cell
adaptive response to changes of insulin sensitivity, was determined by the product
OGIS×AUCCpeptide (adaptation index, AI) [15 ].
Estimation of insulin secretion was based on plasma C-peptide concentrations (prehepatic)
to circumvent possible changes in insulin clearance after surgery, which may influence
peripheral insulin concentrations. The hormonal responses to the mixed meal were evaluated
as the incremental area under the curves (IAUC) for 3 h, calculated with the trapezoidal
rule. IAUC was obtained by subtracting the basal area from the total AUC. The response
of GLP-1 was also assessed as maximal increase (peak) during MMT. Partial T2DM remission
was defined as HbA1c<47.5 mmol/mol and fasting glucose<125 mg/dl in the absence of
antidiabetic medications. Complete remission was defined as a HbA1c<42.1 mmol/mol
and fasting glucose<100 mg/dl in the absence of antidiabetic medications
Statistical analysis
All statistical analyses were performed using the statistical software package, SPSS
version 13.0 (SPSS Inc., Chicago, IL, USA). Continuous variables were expressed as
mean±standard deviation. Group comparisons were performed by the χ2 test for categorical variables, the Wilcoxon test or the Mann-Whitney test. Time
course effect of glucose and hormones was analyzed by general linear model (GLM) for
repeated-measures with the aim of testing whether the changes in hormone levels between
pre- and post-surgery differed among types of surgery at each time point of MMT (treatment×visit×min).
GLM included weight loss as covariate. The association between changes after surgery
was assessed by Spearman’s correlation analysis. To analyze the role of factors associated
with remission of diabetes, we compared the features of remitters and nonremitters.
A logistic regression analysis was performed using remission as dependent variable
and the factors significantly associated with remission as covariate. A p-value<0.05
was considered statistically significant.
Results
Weight loss and metabolic control
[Table 1 ] provides the main clinical and metabolic characteristics of participants before
and one year after surgery. Age, BMI, duration of diabetes, glucose control, and lipid
profile at baseline were similar between RYGB and VSG groups. At one-year, weight
loss expressed as percent change in BMI was − 32±10% after RYGB and − 30±7% after
VSG (p=0.546). Likewise, excess weight loss (EWL%) was similar after the 2 interventions
(78±15 and 70±23%, p=0.146). Glycemic control improved similarly in the 2 groups with
a mean HbA1c reduction of 18–26 mmol/mol from baseline values. Fasting triglycerides
levels fell markedly after both procedures; plasma total and LDL-cholesterol decreased
after RYGB whereas they remained substantially unchanged after VSG ([Table 1 ]). The remission of diabetes (partial plus total) was achieved in 14 VSG patients
(74%) and in 12 RYGB patients (86%) (p=0.28). Antihypertensive medications were discontinued
in all except 3 patients of the RYGB group and in 2 patients of the VSG group. Four
patients of the RYGB group and one patient of the VSG group discontinued hypolipidemic
treatment.
Table 1 Clinical and metabolic characteristics of participants before and one year after
surgery.
RYGB (n=14)
VSG (n=19)
RYGB vs. VSG
Presurgery
One year
p-Value baseline vs. one year
Presurgery
One year
p-Value baseline vs. one year
p-Value at one year
Age (years)
49±7
–
–
44±10
–
–
–
Sex (M/F)
6/8
–
–
8/11
–
–
–
Weight (kg)
116±20
78±8
0.000
130±29
90±17
0.000
0.022
BMI (kg/m2 )
42±6
28±2
0.000
46±9
32±5
0.000
0.009
EWL (%)
–
78±15
–
–
70±23
–
0.259
ΔBMI (%)
–
−32±10
–
–
−30±7
–
0.546
T2DM duration (years)
5±5
–
–
4±4
–
–
–
Glucose (mg/dl)
166±63
76±8
0.000
140±46
87±18
0.006
0.184
HbA1c (mmol/mol)
65±1
39±0.1
0.003
60±0.8
38±0.1
0.004
0.759
Total cholesterol (mg/dl)
202±26
167±38
0.010
207±57
211±49
0.740
0.010
HDL-cholesterol (mg/dl)
44±8
51±10
0.009
44±10
59±16
0.004
0.089
LDL-cholesterol (mg/dl)
120±25
91±32
0.044
127±48
136±44
0.389
0.013
Triglycerides (mg/dl)
195±76
114±78
0.002
220±120
101±27
0.002
0.474
Therapy
Diet alone (%)
2 (14)
13 (93)
0.017
2 (11)
17 (90)
0.003
0.800
OAD users n (%)
9 (64)
1 (7)
0.026
15 (79)
2 (10)
0.007
0.887
OAD+insulin users n (%)
3 (21)
0 (0)
0.098
2 (10)
0 (0)
0.167
0.809
Antihypertensive drug users n (%)
9 (64)
3 (21)
0.142
14 (74)
2 (10)
0.009
0.392
Hypolipidemic drug users n (%)
5 (36)
1 (7)
0.135
3 (16)
2 (10)
0.674
0.387
Data are expressed as means (± SD); RYGB: Gastric bypass; VSG: Vertical sleeve gastrectomy;
BMI: Body mass index; EWL: Excess weight loss; T2DM: Type 2 diabetes mellitus; OAD:
Oral antidiabetic drugs; HDL: High-density lipoprotein; LDL: Low-density lipoprotein;
HbA1c: Hemoglobin A1c
Insulin secretion and insulin sensitivity (OGTT)
Total insulin secretion (ISR) did not change, while beta-cell function improved to
a similar extent one year after surgery (IGI30 =0.5±0.2 and 0.4±0.2 nmol/l/mg/dl and IGI180 =1 603±1 577 and 1 059±952 nmol/l/mg/dl for RYGB and VSG, respectively) ([Table 2 ]). Insulin sensitivity (OGIS) was similar in the 2 groups, preoperatively and markedly
improved after either procedures (p<0.001 for both). Adaptation index (AI) increased
to a similar extent after surgery with no difference between RYGB and VSG.
Table 2 Insulin sensitivity and insulin secretion, and glucose and hormonal response to MMT
before and one year after surgery.
RYGB (n=14)
VSG (n=19)
RYGB vs. VSG
Presurgery
One year
Δ
Presurgery
One year
Δ
pΔ
OGTT
IGI 30 min (nmol/l/mg/dl)
0.2±0.1
0.5±0.2 **
0.3±0.2
0.2±0.02
0.4±0.2*
0.2±0.2
0.406
IGI 180 min (nmol/l/mg/dl)
324±111
1 603±1 577*
1 279±439
281±190
1 059±952*
778±998
0.424
ISR (nmol/m2 )
65±21
58±18
−7±16
57±21
54±20
−3±21
0.681
OGIS (ml min−1 m−2 )
287±90
549±82 **
262±88
302±49
500±86 **
198±93
0.177
Adaptation index
0.5±0.2
0.7±0.3*
0.2±0.2
0.4±0.2
0.7±0.3*
0.3±0.3
0.390
MMT
Fasting glucose (mg/dl)
173±50
96±13 **
−76±50
151±46
104±18 **
−47±39
0.080
IAUC glucose (mg/dl·180′)
6 246±5 186
3 016±3 047
−3 230±5 710
8 681±6 678
3 048±3 490 **
−5 633±4 999
0.227
Fasting insulin (μU/ml)
30±18
11±4*
−19±17
28±16
17±6*
−11±17
0.203
IAUC insulin (μU/ml·180′)
4 313±2 759
5 118±2 076
805±2 808
7 019±4 343
5 250±2 866*
−1 769±2 959
0.017
Fasting GLP-1 (pmol/l)
4.7±4.5
4.5±3.7
−0.2±5.1
3.1±1.8
3.9±2.9
0.8±3.2
0.579
IAUC GLP-1 (pmol/l·180′)
27±130
2 119±1 200 **
2 092±1 266
85±142
235±278
150±319
0.000
GLP-1 30-min peak (pmol/l)
7±6
47±22 **
40±21
4±1
10±7*
6±7
0.000
Fasting GIP (pg/ml)
108±130
42±19*
−66±128
93±72
53±30*
−40±54
0.501
IAUC GIP (pg/ml·180′)
25 344±15 169
13 234±4 713 **
−12 110±15 733
34 132±11 527
21 871±5 619*
−12 261±9 057
0.977
Fasting ghrelin (pg/ml)
391±264
443±169
52±290
457±262
200±41
−257±250
0.013
Ghrelin nadir (pg/ml)
309±72
278±38
31±75
334±64
167±12*
167±61
0.035
Data are expressed as means (± SD); OGTT: Oral glucose tolerance test; MMT: Mixed
meal test; RYGB: Gastric bypass; VSG: Vertical sleeve gastrectomy; OGIS: Oral glucose
insulin sensitivity; ISR: Insulin secretion rate; IGI: Insulinogenic index; AI: Adaptation
index
*p<0.05, ** p≤0.001 vs. presurgery
Glucose and hormone profile (MMT)
IAUCGlucose decreased while IAUCInsulin increased after surgery with no difference between interventions ([Fig. 1 ]).
Fig. 1 Glucose, insulin, GLP-1, GIP, and ghrelin response to a mixed meal in RYGB and VSG
subjects before (continuous line) and one year (dotted line) after surgery: Data are
expressed as means (± SEM). GLM for repeated measures showed a significant meal effect
for GLP-1 (p<0.001). RYGB: Gastric bypass; VSG: Vertical sleeve gastrectomy; GLP-1:
Glucagone-like peptide-1; GIP: Glucose-dependent insulinotropic peptide.
Meal-stimulated GLP-1 concentrations were flat in all patients preoperatively. Following
RYGB, both GLP-1 peak and IAUC increased markedly (p=0.001), while after VSG, the
release of GLP-1, although increased compared to presurgery, was much lower than in
patients operated of RYGB (p=0.0001). Meal GIP response after surgery decreased by
50% (p=0.001 after RYGB and p=0.05 after VSG) with no difference between interventions.
Neither fasting nor nadir ghrelin during MMT changed after RYGB; in contrast, a marked
suppression in both variables occurred after VSG with a significant difference between
the 2 intervention (p=0.013 for fasting ghrelin and p=0.035 for nadir ghrelin concentrations)
([Fig. 1 ] and [Table 2 ]).
The increase in insulin sensitivity and beta-cell function was correlated with weight
loss (R=0.425, p=0.014 and R=0.461, p=0.035, respectively) ([Fig. 2 ]) while no association was found with GI hormone concentrations.
Fig. 2 Correletion between excess weight loss (EWL) and changes of insulin sensitivity (OGIS)
in SG (circle) and RYGB (square) subjects. (Color figure available online only).
Discussion and Conclusion
Discussion and Conclusion
In this study, we evaluated glucose homeostasis and the profile of GI hormones in
severely obese patients with T2DM before and one year following either RYGB or VSG
– 2 of the most frequently performed bariatric procedures – to gain insight into the
physiological mechanisms behind weight loss and remission of T2DM. The 2 interventions
resulted to be equally effective in terms of weight loss and improvement of glycemic
control, with a similar rate of T2DM remission at 1 year (76% after VSG and 86% after
RYGB). Actually, the 2 major determinants of glucose homeostasis, that is, beta-cell
function and insulin sensitivity, improved to a similar extent after either procedures.
Interestingly, total insulin secretion remained unchanged while beta-cell function
increased significantly after surgery, indicating an amelioration of the pancreatic
glucose sensitivity, since a similar secretion occurs with much lower blood glucose.
These results are in agreement with those of Keidar A et al. [16 ] and Nannipieri et al. [17 ] but differ from those of Kashyap et al. [18 ], Lee et al. [19 ] and Schauer et al. [20 ] who demonstrated that RYGB is more effective than VSG in terms of metabolic improvement.
Differences in the degree of weight loss achieved with the 2 procedures, study population,
length of follow-up and experimental methods to assess metabolic functions may contribute
to these variable results. A distinct GI hormonal pattern followed the 2 procedures.
After RYGB, we found a marked increase in meal-induced GLP-1 response, a significant
reduction in post-meal GIP and ghrelin response. Increased GLP-1 levels are well documented
after RYGB, due to the accelerated nutrient entry into the small intestine [3 ]
[4 ]. Data regarding GIP are more inconsistent, with some studies reporting an increased
postprandial GIP level early after RYGB [21 ] and others no change [22 ] or even an early decline in fasting GIP levels in diabetic but not in nondiabetic
patients [23 ]. This discrepancy may be due to differences in analytical methods to measure GIP
(total vs. active form), the characteristics of the population studied (diabetics
vs. nondiabetic), the length of the limbs of the Roux anastomosis and the duration
of the follow-up.
Following VSG, a marked suppression of both fasting and post-meal ghrelin levels occurred
as a consequence of gastric fundus removal; GLP-1 concentration increased although
to a much lower extent than RYGB while GIP levels decreased by 50%.
The finding that RYGB and VSG are equally effective on weight loss and metabolic improvement
in the face of a different pattern of GI hormone profile, lead us to hypothesize that
the changes in gut hormones are not the main determinant of the metabolic improvement,
at least several months after surgery. However, since our evaluations were performed
one year after surgery we cannot rule out that the changes in GI hormonal profile
may have contributed to diabetes remission early after surgery. This hypothesis is
in line with a recently published commentary, which underlined that the mechanisms
behind the remission of diabetes after VSG or RYGB may differ in relation to the time
at which they are studied. Early after surgery, the improvement of glycemic control
is due to increased hepatic insulin sensitivity and to the improved beta-cell function
consequent to the exaggerated postprandial GLP-1 secretion. Later on, with progressive
weight loss the improvement in peripheral insulin sensitivity becomes the prevalent
mechanism [24 ]. On the other hand, a number of mechanisms have been highlighted which may contribute
to the improvement of glucose tolerance after BS, including neural activation [25 ], modifications of intestinal microbiota [26 ], and changes in the expression of genes regulating glucose and fatty-acid metabolism
induced by low nutrient availability [27 ].
A significant reduction in fasting triglycerides occurred in our patients after either
procedure, as reported in previous studies [10 ]
[28 ]
[29 ], whereas total and LDL-cholesterol decreased after RYGB but not VSG. This finding
is in line with recent studies demonstrating that bariatric procedures differentially
affect cholesterol metabolism with malabsorptive procedures (biliointestinal bypass)
providing a much greater reduction than restrictive surgery independent of weight
loss and insulin resistance [30 ].
A weakness of this study might be the lack of patient randomization to the 2 types
of operations. However, since the 2 procedures differ in terms of unwanted effects
and frequency of monitoring during follow-up, the patient’s preference should not
be ignored. Noteworthy is the fact that the 2 groups were comparable for anthropometric
and biochemical measures, as well as for medication use, thus minimizing the possibility
of selection bias.
In conclusion, RYGB and VSG exert similar beneficial effects in terms of weight loss
and remission of T2DM in the face of remarkable differences in GI hormone profile.
These findings highlight the importance of weight loss and challenge a primary role
of incretins in mediating the metabolic improvement achieved in obese patients with
T2DM one year after VSG or RYGB.
