Surgery for Obesity and Related Diseases ] (2015) 00–00
Original article
Should we wait for metabolic complications before operating on obese patients? Gastric bypass outcomes in metabolically healthy obese individuals Elise Pelascini, M.D.a,b,c, Emmanuel Disse, M.D., Ph.D.b,c,d, Arnaud Pasquer, M.D.a,b,c, Gilles Poncet, M.D., Ph.D.a,b,c, Christian Gouillat, M.D., Ph.D.a,b,c, Maud Robert, M.D., Ph.D.a,b,c,* a Department of Digestive Surgery, Center of Bariatric Surgery, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France Centre Intégré et Spécialisé de l’Obésité de Lyon, Groupement Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite, France c Université Claude Bernard Lyon 1, Lyon, France d Department of Endocrinology, Diabetology and Nutrition, Groupement Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite, France b
Abstract
Background: A subgroup of obese patients without metabolic disorders has been identified and defined as metabolically healthy but morbidly obese (MHMO). Objectives: To compare Roux-en-Y gastric bypass (RYGB) outcomes between MHMO and metabolically unhealthy morbidly obese (MUMO) patients to assess whether the obesity phenotype could affect the results. Setting: A university-affiliated tertiary care center. Methods: One hundred nineteen consecutive patients underwent RYGB; 102 completed the 2-year follow-up and were divided into 2 groups (MHMO and MUMO) according to Wildman criteria, including blood pressure, triglycerides, high-density lipoprotein cholesterol (HDL-C), fasting blood sugar, C-reactive protein (CRP), and homeostasis model assessment of insulin resistance (HOMAIR). Weight loss and metabolic parameter changes were analyzed. Results: Twenty-one of 102 (20.6%) patients were identified as MHMO; they were mostly women (90.5%) and were significantly younger than MUMO patients (39.4 ⫾ 9.1 yr versus 47.2 ⫾ 10, P ¼ .001); 12.6% were lost to follow-up. MHMO phenotype was significantly associated with a greater percentage of excess body mass index loss (P ¼ .03), independent of gender, age, and redo procedures. All metabolic parameters were significantly improved 2 years after surgery in the MUMO group. HOMA-IR, CRP, and triglycerides were significantly lower 2 years after surgery in the MHMO group, whereas fasting blood sugar and HDL-C were unchanged. At 2 years of follow-up, 92.3% of the population was metabolically healthy. Conclusions: RYGB is an effective procedure to achieve weight loss and had a strong positive metabolic effect in both MHMO and MUMO phenotypes. RYGB led to an increase of the metabolically healthy status and may prevent or delay the onset of metabolic disorders. (Surg Obes Relat Dis 2015;]:00–00.) r 2015 American Society for Metabolic and Bariatric Surgery. All rights reserved.
Keywords:
Morbid obesity; Metabolically healthy; Metabolic disorders; Gastric bypass; Obesity phenotype; Weight loss
*
Correspondence: Maud Robert, M.D., Ph.D., Hôpital Edouard Herriot, Chirurgie Digestive, Pavillon D, 5 Place d’Arsonval, 69437 Lyon Cedex 03, France. E-mail: [email protected]
The increasing prevalence of obesity worldwide has major socioeconomic consequences. Obesity is often associated with metabolic disorders leading to type 2 diabetes mellitus (T2DM), higher cardiovascular risk, and higher
http://dx.doi.org/10.1016/j.soard.2015.04.024 1550-7289/r 2015 American Society for Metabolic and Bariatric Surgery. All rights reserved.
2
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
mortality rate [1,2]. Bariatric surgery is now considered the best option to achieve significant weight loss and has also been found to be superior as an optimized medical therapy to cure T2DM and other co-morbidities [3,4]. Therefore, indications for bariatric surgery are now evolving from weight loss considerations to metabolic objectives. Since the 1980s, a subgroup of obese patients without metabolic disorders has been identified and defined as metabolically healthy but morbidly obese (MHMO) patients [5–7]. Currently, there is still no standardized definition of metabolic health, resulting in a wide variation of metabolically healthy obesity prevalence, which is estimated to be between 6% and 40% [6,8]. MHMO characterized by less visceral fat would be at lower cardiovascular risk [8]. Thus, whereas indications for bariatric surgery tend to expand to nonobese T2DM patients with metabolic concerns, the best therapeutic strategy in MHMO patients, who could be at lower risk of morbidity and mortality, remains questionable. In the present study, we compared gastric bypass outcomes in MHMO versus metabolically unhealthy morbidly obese (MUMO) patients. The aim was to assess whether obesity phenotype influences the results of bariatric surgery and therefore if our surgical management can be based on stratification of obese individuals according to their metabolic health phenotype. Methods Definition of MHMO patients and population selection From April 2009 and May 2011, 119 patients underwent a Roux-en-Y gastric bypass (RYGB) in the Department of Digestive Surgery, Edouard Herriot Hospital, a universityaffiliated tertiary care center (Lyon, France). Fifteen patients were excluded from the study because of insufficient collected data (follow-up performed in another center) and were considered as lost to follow-up. Two others were excluded because of type 1 diabetes mellitus. Finally, 102 patients were included in the study. Patients' characteristics are presented in Table 1. The population was divided into 2 groups, MHMO and MUMO, according to Wildman’s definition [9]. Patients were considered as MHMO when having only 1 or no cardiometabolic abnormalities among the following 6: 1. Elevated blood pressure: systolic/diastolic blood pressure Z130/85 mm Hg or antihypertensive medication use; 2. Elevated triglycerides level: fasting triglycerides level Z1.70 mmol/L; 3. Decreased high-density lipoprotein cholesterol (HDL-C) level: o1.04 mmol/L in men or o1.30 mmol/L in women or lipid-lowering medication use; 4. Elevated glucose level: fasting blood sugar Z5.5 mmol/ L or antidiabetic medication use;
Table 1 Characteristics of the study population Population
n ¼ 102
Age (yr) Male gender (%) Height (m) Weight (kg) Waist Circumference (cm) BMI (kg/m²) T2DM (%) Insulin-treated T2DM (%) Arterial hypertension (%) Dyslipidemia (%) Redo surgeries (%)
45.6 (⫾ 10) 43% 1.66 (⫾ .09) 124.2 (⫾ 20) 131 (⫾ 14) 45 (⫾ 6.6) 51% (n ¼ 54) 21% (n ¼ 21) 41% (n ¼ 45) 46% (n ¼ 47) 27.4% (n ¼ 28)
BMI ¼ body mass index; T2DM ¼ type 2 diabetes. Data are presented as mean ⫾ standard deviation.
5. Insulin resistance: homeostasis model assessment of insulin resistance (HOMA-IR) 42.5 (i.e., 490th percentile); and 6. Systemic inflammation: C-reactive protein (CRP) level 45 mg/L (i.e., 490th percentile). Laparoscopic Roux-en-Y gastric bypass procedure A 5-port technique was used and consisted in a small gastric pouch (30 mL) by stapling the stomach using a linear stapler. The first jejunal loop was moved up into an antecolic position after an epiploic transection. An end-to-side gastrojejunal anastomosis was performed using a linear stapler. Closure of the anterior part of the anastomosis was done using a running suture. The alimentary limb was 150 cm long. A laterolateral jejunojejunal anastomosis was performed with a linear stapler. Closure of the mesenteric defect was systematic, using a nonabsorbable silk suture (2/0). Data analysis This is a retrospective study of prospectively collected data from our electronic database dedicated to bariatric surgery. Anthropometric and biological data were recorded preoperatively; at 2, 6, 12, 18, and 24 months; and every year after surgery. Weight loss was expressed in percentage of excess body mass index (BMI) loss (%EBMIL), absolute weight loss in kg (aWL), and BMI loss in kg/m2. The prevalence and evolution of metabolic co-morbidities in each group were assessed. Type 2 diabetes mellitus was defined using the American Diabetes Association criteria [10], and diabetes was considered in remission when patients stopped their antidiabetic treatments and HbA1c was o6% and/or fasting blood glucose levels were o100 mg/dL. Arterial hypertension was defined as systolic/ diastolic blood pressure Z130/85 mm Hg or use of antihypertensive drugs. Dyslipidemia was defined as lowdensity lipoprotein cholesterol (LDL-C) Z4.1 mmol/L or triglycerides Z1.7 mmol/L.
RYGB Outcomes in Metabolically Healthy Obesity / Surgery for Obesity and Related Diseases ] (2015) 00–00
Because redo surgery has been previously associated with poor weight loss results [11,12], and because of a high percentage of redo procedures in the present study population (MHMO: n = 12/21; MUMO: n = 16/81), weight loss results were analyzed considering previous history of bariatric surgery so as to limit any selection bias. Statistical analysis IBM SPSS Statistics Version 19 (IBM Corp., Armonk, NY, USA) was used for analyses. Baseline quantitative data were compared between groups by a Student t test. When the distribution of variables was nonnormal, the nonparametric test of Wilcoxon was used. Ordinal data were compared between groups by a χ2 test. Weight loss evolution (%EBMIL, BMI, BMI loss, and aWL) was analyzed by mixed ANOVA (3 factors) repeated measures; time was the within factor, and obesity phenotype (MUMO or MHMO) and redo surgery (yes/no) were the 2 between factors. Estimated marginal means of %EBMIL were used for the graphic representation of weight loss in MUMO and MHMO phenotypes. Evolution of metabolic parameters between groups over time was analyzed using ANOVA repeated measure. A multiple linear regression model was used to identify the predictive factors of weight loss. P o .05 was considered significant.
3
Table 2 Anthropometric and metabolic data in metabolically healthy but morbidly obese (MHMO) compared with metabolically unhealthy morbidly obese (MUMO) individuals
Age (yr) Male ratio (%) Height (m) Weight (kg) Waist circumference (cm) BMI (kg/m2) T2DM (%) HTN (%) Dyslipidemia (%) Redo surgeries (%) bands to RYGB Mason to RYGB CRP (mg/L) Triglyceride (mmol/L) HDL-C (mmol/L) Fasting blood sugar (mmol/L) Fasting insulin (mUI/L) HOMA-IR
MHMO
MUMO
n ¼ 21
n ¼ 81
39.4 (⫾ 9.1) 9.5% 1.6 (⫾ .07) 119 (⫾ 16) 125 (⫾ 16) 44 (⫾ 5) 5% 5% 14% 12 (57%) 6 6 6.7 (⫾ 5) 1.1 (⫾ .4) 1.3 (⫾ .2) 5.0 (⫾ 1.7) 8.6 (⫾ 4) 1.7 (⫾ .8)
47.2 (⫾ 10) 35.8% 1.6 (⫾ .09) 125 (⫾ 21) 132 (⫾ 14) 45.2 (⫾ 7) 65% 54% 54% 16 (20%) 12 4 10.6 (⫾ 10) 2 (⫾ 1.1) 1 (⫾ .4) 7.1 (⫾ 3.1) 18.2 (⫾ 22) 5.9 (⫾ 7)
P
.001* .04* .2 .2 .3 .4 o.0005* o.0005* .003* .002*
.1 .0009* .002* .002* .1 .03*
BMI ¼ body mass index; T2DM ¼ type 2 diabetes mellitus; HTN ¼ arterial hypertension; CRP ¼ C-reactive protein; HDL-C ¼ high-density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance. Data are presented as arithmetical mean ⫾ standard deviation. * Statistically significant.
Results Preoperative characteristics of MHMO versus MUMO groups Using Wildman criteria, 20.6% (n = 21) of the whole population was identified as MHMO. MHMO patients were mostly women (90.5%) and were significantly younger than MUMO patients (39.4 ⫾ 9.1 versus 47.2 ⫾ 10, P = .001). A previous history of bariatric surgery was more common in the MHMO group (57% versus 20%, P = .002). Most revisional procedures were for weight loss failure, after appropriate nutritional management. The others were for functional disorders like reflux or esophageal motility disorders. Characteristics of MHMO and MUMO groups are presented in Table 2. When 1 cardiometabolic abnormality was present in the MHMO group, low-grade systemic inflammation was the most common (38.9%). In MUMO individuals, altered glucose homeostasis was the most prevalent metabolic disorder (75.3% had elevated fasting blood sugar and 67.7% had insulin resistance), followed by altered lipid profile (68.3% had low HDL-C and 55.7% had hypertriglyceridemia). Follow-up and 30-day postoperative complications Mean follow-up after surgery was 31.6 (⫾9) months; 12.6% were lost to follow-up. In the whole population, the rate of minor complications was 4.9% and the rate of major
complications was 3.9%, according to Dindo-Clavien classification. In the primary RYGB group (n ¼ 74) we had 6 30-day postoperative complications (8%): 4 minor (1 pneumonia, 1 ovarian cyst hemorrhage, 2 abdominal sepsis treated and resolved with antibiotherapy) and 2 major (endoluminal hemorrhage of the gastrojejunal anastomosis). In the revisional gastric bypass group (n ¼ 28) we had 3 30day postoperative complications (10%): 1 minor (pneumonia) and 2 major (1 trocar hemorrhage, 1 stenosis of the jejunojejunal anastomosis). Weight loss outcomes Mean %EBMIL at 2 years in the whole population was 66.2% (⫾26). The mean weight was 88.3 kg (⫾18) and the mean BMI was 32.1 kg/m2 (⫾6). Using ANOVA repeated measures to assess the influence of time and obesity phenotype on %EBMIL, we observed a significant effect of time on %EBMIL (P o .001) and a greater %EBMIL in the MHMO group (P ¼ .06). Evolution of estimated marginal mean %EBMIL over time in MHMO compared with MUMO individuals is presented in Fig. 1; estimated marginal mean of %EBMIL at 24 months after RYGB was 71.6% versus 59.5%, respectively. Two years after surgery, BMI in MHMO and MUMO was, respectively, 30.9 and 32.3 kg/m-2; obesity phenotype had no significant effect on BMI over time (P ¼ .08). BMI loss at
4
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
Fig. 1. Graphic representation of percentage of excess body mass index (BMI) loss over time (estimated marginal means) in the whole population according to the obesity phenotype (metabolically healthy but morbidly obese versus metabolically unhealthy morbidly obese). Obesity phenotype effect over time: P ¼ .06 (ANOVA repeated measures). %EBMIL ¼ Excess BMI loss %; MHMO ¼ metabolically healthy but morbidly obese patients; MUMO ¼ metabolically unhealthy morbidly obese patients. MHMO: n ¼ 21, MUMO: n ¼ 81. There were none lost to follow-up at each time-point.
24 months was 13.2 and 13 kg/m-2 in MHMO and MUMO, respectively, and obesity phenotype did not affect BMI loss over time (P ¼ .2). Absolute weight loss at 24 months was 35.6 versus 35.9 kg in MHMO and MUMO respectively. Finally, using ANOVA repeated measures, obesity phenotype did not affect the absolute weight loss over time (P ¼ .3). Redo procedures were significantly associated with lower %EBMIL (P ¼ .001), lower BMI loss (P ¼ .01), and lower aWL (P ¼ .002) over time. Nevertheless, no significant effect of the interaction between obesity phenotype and redo procedures was observed no matter how weight loss results were expressed. Change in %EBMIL over time in MHMO and MUMO groups according to their previous history of bariatric surgery is presented in Fig. 2. In the subgroup of redo procedures, %EBMIL at 24 months was 59% in MHMO versus 49% in MUMO individuals (P ¼ .3). In the subgroup of primary RYGB, %EBMIL at 24 months was 85% (⫾19.6) in MHMO and 70% (⫾24.1) in MUMO (P ¼ .08). Evolution of metabolic parameters Two years after RYGB, 92.3% of the whole population exhibited r1 cardiometabolic disturbance and was considered "metabolically healthy" according to Wildman criteria. Only 7.7% of the whole population still had a metabolically unhealthy phenotype. Changes in fasting blood sugar,
HOMA-IR, triglycerides, HDL-C, and CRP over time are presented in Fig. 3. Using ANOVA repeated measures, we found that each metabolic parameter was significantly improved over time (all were decreased except HDL-C, which was increased). A significant effect of obesity phenotype on the evolution of postoperative glycemia (P ¼ .01) and triglycerides (P ¼ .003) was observed, which was significantly lower in MHMO over time. A trend toward lower HOMA-IR (P ¼ .06) and lower CRP (P ¼ 006) was observed in MHMO during the postoperative period. Changes in metabolic parameters from the preoperative period to the 2-year follow-up are presented in Table 3. All metabolic parameters were improved from baseline to 2 years after surgery in the MUMO group. HOMA, CRP, and triglycerides were significantly lower 2 years after surgery in the MHMO group, whereas glycemia and HDL-C were unchanged. Thus, despite significant decrease in glycemia after surgery in MUMO and no glycemic change in MHMO group, glycemia stayed higher in the MUMO versus the MHMO group (P ¼ .01). Predictive factors of weight loss at 2 years Performing multiple regression analysis and considering the 2-year %EBMIL as the dependent variable, we found that MHMO phenotype was significantly associated with a
RYGB Outcomes in Metabolically Healthy Obesity / Surgery for Obesity and Related Diseases ] (2015) 00–00
5
Fig. 2. Percentage of excess body mass index (BMI) loss over time after Roux-en-Y gastric bypass (RYGB) in metabolically healthy morbidly obese and metabolically unhealthy morbidly obese patients according to their previous history of bariatric surgery (primary RYGB versus revisional RYGB). Obesity phenotype effect: P ¼ .06. Previous bariatric surgery effect: P ¼ .001. Obesity phenotype x previous surgery interaction effect: P ¼ .6. %EBMIL ¼ excess BMI loss %; MHMO ¼ metabolically healthy but morbidly obese patients; MUMO ¼ metabolically unhealthy morbidly obese patients. MUMO primary RYGB: n ¼ 65, MHMO primary RYGB: n ¼ 9, MUMO revisional RYGB: n ¼ 16, MHMO revisional RYGBP: n ¼ 12. There were none lost to follow-up at each time point.
greater %EBMIL (P ¼ .03), independently of gender (P ¼ .8), age (P ¼ .4), and redo procedures, which were negatively correlated to %EBMIL (P ¼ .001). Discussion The MHMO phenotype is a quite new concept that was first described in the 1980s [13,14]. Whereas obesity is often associated with metabolic abnormalities leading to an increased risk of T2DM and cardiovascular disease, interest has grown for this paradoxical condition in which patients are still metabolically healthy despite excessive body fat. Indeed, understanding the determinants of the metabolically healthy status is a matter of great interest and could lead to change and improve the management strategies of obesity. Data in the literature regarding the prevalence and the determinants of the MHMO phenotype are heterogeneous and conflicting. First, there is still no standardized definition of metabolic health, making it difficult to compare the studies [6]. Second, variation among ethnicities, sample size, age groups, and study design can also account for the wide heterogeneity of the data. The prevalence of MHMO could be 6%–40%, depending on the studies and the criteria used [5,8]. Several definitions have been proposed to define MHMO and MUMO phenotypes, based on the cardiometabolic risk assessment. Whereas the first definitions were specifically based on the metabolic syndrome parameters, Wildman has proposed to add surrogate biomarkers of insulin resistance (i.
e., HOMA-IR) and inflammation (i.e., CRP). Thus, this last definition takes into account pathophysiologic traits of obesity that have been linked with T2DM risk and cardiovascular disease [8]. Therefore, we also used Wildman criteria to define MHMO individuals in our study [9]. It is of note that, using these criteria, we found a 20.6% rate of MHMO patients in our population, which is inferior to the 31.7% MHMO prevalence reported by Wildman et al. and by others [15,16]. This can be explained by the fact that we focused on a population with several co-morbidities and metabolic disorders, in whom gastric bypass is the procedure of choice because of its metabolic effect. We also found that our MHMO population was mostly female (90.5%) and significantly younger than MUMO patients (39.4 ⫾ 9.1 versus 47.2 ⫾ 10, P = .001), which is consistent with literature data [17,18]. The hypothesis that gynecoid fat distribution in women, who thus have less visceral fat than men, could be one of the determinants of MHMO status has been supported by many authors [19,20]. Indeed, visceral fat has been associated with insulin resistance and cardiovascular risk in several studies [21,22]. Nevertheless, some authors did not find any correlation between visceral fat assessed by waist circumference and MHMO phenotype, arguing that difference between MHMO and MUMO individuals mostly is due to the total amount of body fat accretion rather than body fat distribution [8]. In our study population, MHMO and MUMO had similar waist circumference (125 cm ⫾ 15.7 versus
6
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
Fig. 3. Postoperative evolution of (A) glycemia, (B) HOMA-IR, (C) triglycerides, (D) HDL cholesterol, (E) C-reactive protein, and (F) BMI at 6, 12, and 24 months in metabolically unhealthy morbidly obese (gray line) and metabolically healthy but morbidly obese (black line) phenotypes. Results are presented as arithmetical means and error bars represent standard error of the mean. Effect of obesity phenotype over time on glycemia (P ¼ .01), HOMA-IR (P ¼ .06); triglycerides (P ¼ .003), HDL-C (P ¼ .1), C-reactive protein (P ¼ .06). HOMA-IR ¼ homeostasis model assessment of insulin resistance; HDL ¼ highdensity lipoprotein; BMI ¼ body mass index. MHMO: n ¼ 21, MUMO: n ¼ 81. There were none lost to follow-up at each time point.
132 ⫾ 13.8, P ¼ .3). Beside whole body fat distribution and accretion, several histologic and functional features of adipose tissue in obesity have been associated with insulin resistance and metabolic disorders, including immune cells infiltration, proinflammatory cytokines secretion, fibrosis, adipocyte hypertrophy [23]. Thus, metabolic health status in obese patients is obviously not easy to assess with simple clinical and biological parameters such as BMI, waist circumference, glycemia, and lipid profile. In our study population, we found that low-grade inflammation was the most common metabolic disorder in the MHMO group (38.9% of patients with CRP 45 mg/L). This is consistent with the finding of Wildman et al. reporting that, despite no increase in cardiovascular risk in MHMO women, these individuals still displayed abnormal levels of inflammatory markers [24]. This could suggest that
MHMO individuals are at risk for insulin resistance and could evolve into a MUMO state. This hypothesis is supported by the study of Soriguer et al., who recently assessed the evolution of the metabolic health status at 6 and 11 years of follow-up in 1051 individuals representative of the general population [25]. They found that in MHMO patients the risk for becoming diabetic was lower than in MUMO patients, but this risk remained significant (odds ratio ¼ 3.13; 95% confidence interval ¼ 1.07–9.17; P ¼ .02). They also reported that in patients who lost weight during the study, the association between MHMO phenotype and T2DM incidence disappeared, even after adjusting for HOMA-IR. These results suggest that MHMO phenotype could be a changing and evolving state that should be considered over time and that MHMO individuals would gain metabolic benefit from weight loss.
RYGB Outcomes in Metabolically Healthy Obesity / Surgery for Obesity and Related Diseases ] (2015) 00–00 Table 3 Evolution of the metabolic parameters in metabolically healthy but morbidly obese (MHMO) and metabolically unhealthy morbidly obese (MUMO) patients 2 years after the gastric bypass Metabolic parameters
Obesity Preoperative 24 months after phenotype values surgery values
Fasting blood MHMO sugar (mmol/L) MUMO HOMA-IR MHMO MUMO Triglycerides MHMO (mmol/L) MUMO HDL-C (mmol/L) MHMO MUMO CRP (mg/L) MHMO MUMO
P
4.6 (⫾ .4)
4.5 (⫾ .5)
.1
7.2 (⫾ 3) 1.7 (⫾ .8) 5.5 1.1 (⫾ .5)
5.4 (⫾ 2) .7 (⫾ .3) 1.4 (⫾ 2.2) .8 (⫾ .2)
.0001* .003* o.001* .005*
2 (⫾ 1.1) 1.3 (⫾ .2) 1.1 (⫾ .3) 4.9 (⫾ 3.8) 10.1 (⫾ 8)
1.1 (⫾ .5) 1.4 (⫾ .2) 1.4 (⫾ .3) 1 (⫾ .7) 2.7 (⫾ 4.4)
o.0001* 1.5 o.0001* .03* o.0001*
HOMA-IR ¼ homeostasis model assessment of insulin resistance; HDLC: high-density lipoprotein cholesterol; CRP ¼ C-reactive protein. * Statistically significant.
In our study, we found a significant improvement in all metabolic parameters over the postoperative period and a significant favorable effect of MHMO status compared with MUMO phenotype on glycemia (P ¼ .01) and triglycerides (P ¼ .03). A favorable trend to improve HOMA-IR (P ¼ .06) and CRP (P ¼ .06) after surgery was also observed in the MHMO phenotype (ANOVA repeated measures). Finally, weight loss induced by RYGB had a beneficial effect on the metabolic profile of both MUMO and MHMO patients. Nevertheless, patients with preoperative MHMO phenotype still had a better biological profile of metabolic health 24 months after surgery than those with the preoperative MUMO phenotype, despite the strong decrease of MUMO phenotype prevalence 24 months after surgery. Thus, despite massive weight loss, MHMO and MUMO patients are still discernible after bariatric surgery. These results seem to indicate that MUMO phenotype is not totally reversible into a metabolically healthy condition. This is a further argument for considering that MHMO and MUMO phenotypes are different stages in the evolution of obesity disease. Few studies have assessed the outcomes of bariatric surgery in MHMO patients compared with MUMO individuals [18,26,27]. In a cohort of 190 patients who underwent gastric banding, Sesti et al. assessed weight loss and metabolic outcomes at 6 months of follow-up in the subgroups of MHMO and MUMO individuals. In this short-term study, the authors reported similar weight loss in both groups and a significant correlation between weight loss and improvement of insulin sensitivity [26]. Jimenez et al. studied the effects of bariatric surgery (RYGB and sleeve gastrectomies) on metabolic co-morbidities and anthropometric and inflammatory markers at 12 months of follow-up between 52 obese women presenting with a
7
HOMA-IR o2.94 and an age- and BMI-matched group of women with HOMA-IR 42.94 [27]. The authors concluded that bariatric surgery also results in improvement of the altered metabolic and inflammatory profiles in the insulin sensitive, morbidly obese group. More recently, Goday et al. conducted a prospective cohort study on 222 morbidly obese patients who underwent either a RYGB (n ¼ 135) or a sleeve gastrectomy (n ¼ 87) [18]. Forty-two (18.9%) patients fulfilled the criteria for MHMO phenotype and had similar weight loss to MUMO patients at 1 year of follow-up. In that study, MHMO patients exhibited a significant improvement of plasma glucose, HOMA-IR, total cholesterol, LDL-C, and HDL-C after surgery. In the present study, we noted a trend toward a favorable effect of MHMO phenotype on %EBMIL (ANOVA repeated measures, P ¼ .06) over time, and MHMO phenotype appeared to be an independent factor of greater weight loss (multiple regression, P ¼ .03). An interesting hypothesis could be that adipose tissue from MUMO is structurally different from adipose tissue from MHMO. Because of increased inflammation and fibrosis, reduction of adipose tissue from MUMO individuals could be limited and could affect weight loss negatively. Such a stimulating hypothesis needs to be tested in further clinical studies. This study has some limitations. The small sample size (n ¼ 102) is explained by the careful selection of RYGB with complete biological and anthropometric data during more than 24 months of follow-up. There was also a large prevalence of redo procedures in the MHMO group (57.1%), which could be a selection bias; first of all, it could lead to underestimating weight loss outcomes in this group compared with the MUMO patients. Indeed, we and others also had reported the negative effect of redo procedures on weight loss results [11,12]. When analyzing weight loss after excluding redo procedures, we found the 2 years %EBMIL tended to be greater in the MHMO group versus the MUMO group (85% ⫾ 19.6 versus 70% ⫾ 24.1 respectively, P ¼ .08). These results seem to confirm that bariatric surgery is as effective or more effective in MHMO than in MUMO individuals for achieving weight loss. Second, one could question if patients with a history of redo procedures are "real" MHMO individuals or if they became metabolically healthy as a result of the first bariatric procedure. However, as previously discussed, we believe that metabolic health status is a dynamic condition and that the aim of bariatric surgery is to avoid the onset of metabolic disorders whatever the previous history of obesity. Because of the study design, we could not assess the impact of the type of surgical procedure (malabsorptive versus restrictive) on metabolic outcomes. One might question the interest of performing an RYGB procedure, often considered a metabolic technique and which is at higher risk of complications, rather than a restrictive
8
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
procedure in metabolically healthy individuals. However, we believe that RYGB, by achieving good weight loss and durable results, is still an interesting option. Obviously, further trials are needed to answer this question and to confirm the scarce literature data. Conclusions RYGB appears to be an effective procedure to achieve weight loss in both metabolically healthy and metabolically unhealthy morbidly obese patients. RYGB produced a strong metabolic effect, leading to an increase of the metabolically healthy status in our study population from 20.6% to 92.3% at 2 years of follow-up, suggesting that metabolic health is a dynamic condition and not a definitive state. Even in metabolically healthy individuals, bariatric surgery has a favorable metabolic impact that could prevent or delay the onset of metabolic disorders. Further studies are needed to determine whether restrictive procedures are as effective as RYGB on weight loss and metabolic outcomes in MHMO individuals so as to define the best management strategies in this specific obesity phenotype. Disclosures The authors have no disclosures and no commercial interest in the subject of study. References [1] Sjostrom CD, Lystig T, Lindroos AK. Impact of weight change, secular trends and ageing on cardiovascular risk factors: 10-year experiences from the SOS study. Int J Obes (Lond) 2011;35 (11):1413–20. [2] Van Gaal LF, Mertens IL, De Block CE. Mechanisms linking obesity with cardiovascular disease. Nature 2006;444(7121):875–80. [3] Schauer PR, Bhatt DL, Kirwan JP, et al. Bariatric surgery versus intensive medical therapy for diabetes–3-year outcomes. N Engl J Med. 2014;370(21):2002–13. [4] Guo X, Liu X, Wang M, Wei F, Zhang Y. The effects of bariatric procedures versus medical therapy for obese patients with type 2 diabetes: meta-analysis of randomized controlled trials. Biomed Res Int 2013;2013:410609. [5] Ruderman NB, Schneider SH, Berchtold P. The "metabolicallyobese," normal-weight individual. Am J Clin Nutr 1981;34 (8):1617–21. [6] Phillips CM. Metabolically healthy obesity: definitions, determinants and clinical implications. Rev Endocr Metab Disord 2013;14 (3):219–27. [7] Sims EA. Are there persons who are obese, but metabolically healthy? Metabolism 2001;50(12):1499–504. [8] Martinez-Larrad MT, Corbaton Anchuelo A, Del Prado N, et al. Profile of individuals who are metabolically healthy obese using different definition criteria. A population-based analysis in the Spanish population. PLoS One 2014;9(9):e106641.
[9] Wildman RP, Muntner P, Reynolds K, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999-2004). Arch Intern Med 2008;168(15):1617–24. [10] American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2010;33(Suppl 1):S62–9. [11] Robert M, Pasquer A, Espalieu P, et al. Gastric bypass for obesity in the elderly: is it as appropriate as for young and middle-aged populations? Obes Surg 2014;24(10):1662–9. [12] Slegtenhorst BR, van der Harst E, Demirkiran A, et al. Effect of primary versus revisional Roux-en-Y gastric bypass: inferior weight loss of revisional surgery after gastric banding. Surg Obes Relat Dis 2013;9(2):253–8. [13] Sims EA, Berchtold P. Obesity and hypertension. Mechanisms and implications for management. JAMA 1982;247(1):49–52. [14] Andres R. Effect of obesity on total mortality. Int J Obes 1980;4 (4):381–6. [15] Lopez-Garcia E, Guallar-Castillon P, Leon-Munoz L, RodriguezArtalejo F. Prevalence and determinants of metabolically healthy obesity in Spain. Atherosclerosis 2013;231(1):152–7. [16] Iacobellis G, Ribaudo MC, Zappaterreno A, Iannucci CV, Leonetti F. Prevalence of uncomplicated obesity in an Italian obese population. Obes Res 2005;13(6):1116–22. [17] Phillips CM, Dillon C, Harrington JM, et al. Defining metabolically healthy obesity: role of dietary and lifestyle factors. PLoS One 2013;8 (10):e76188. [18] Goday A, Benaiges D, Parri A, et al. Can bariatric surgery improve cardiovascular risk factors in the metabolically healthy but morbidly obese patient? Surg Obes Relat Dis 2014;10(5):871–6. [19] Brochu M, Tchernof A, Dionne IJ, et al. What are the physical characteristics associated with a normal metabolic profile despite a high level of obesity in postmenopausal women? J Clin Endocrinol Metab 2001;86(3):1020–5. [20] Marin P, Andersson B, Ottosson M, et al. The morphology and metabolism of intraabdominal adipose tissue in men. Metabolism 1992;41(11):1242–8. [21] Dagenais GR, Yi Q, Mann JF, et al. Prognostic impact of body weight and abdominal obesity in women and men with cardiovascular disease. Am Heart J 2005;149(1):54–60. [22] Brochu M, Poehlman ET, Ades PA. Obesity, body fat distribution, and coronary artery disease. J Cardiopulm Rehabil 2000;20(2):96–108. [23] Rajala MW, Scherer PE. Minireview: the adipocyte—at the crossroads of energy homeostasis, inflammation, and atherosclerosis. Endocrinology 2003;144(9):3765–73. [24] Wildman RP, Kaplan R, Manson JE, et al. Body size phenotypes and inflammation in the Women's Health Initiative Observational Study. Obesity (Silver Spring) 2011;19(7):1482–91. [25] Soriguer F, Gutierrez-Repiso C, Rubio-Martin E, et al. Metabolically healthy but obese, a matter of time? Findings from the prospective Pizarra study. J Clin Endocrinol Metab 2013;98(6):2318–25. [26] Sesti G, Folli F, Perego L, Hribal ML, Pontiroli AE. Effects of weight loss in metabolically healthy obese subjects after laparoscopic adjustable gastric banding and hypocaloric diet. PLoS One 2011;6 (3):e17737. [27] Jimenez A, Perea V, Corcelles R, et al. Metabolic effects of bariatric surgery in insulin-sensitive morbidly obese subjects. Obes Surg 2013;23(4):494–500.
Original article
Should we wait for metabolic complications before operating on obese patients? Gastric bypass outcomes in metabolically healthy obese individuals Elise Pelascini, M.D.a,b,c, Emmanuel Disse, M.D., Ph.D.b,c,d, Arnaud Pasquer, M.D.a,b,c, Gilles Poncet, M.D., Ph.D.a,b,c, Christian Gouillat, M.D., Ph.D.a,b,c, Maud Robert, M.D., Ph.D.a,b,c,* a Department of Digestive Surgery, Center of Bariatric Surgery, Hôpital Edouard Herriot, Hospices Civils de Lyon, Lyon, France Centre Intégré et Spécialisé de l’Obésité de Lyon, Groupement Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite, France c Université Claude Bernard Lyon 1, Lyon, France d Department of Endocrinology, Diabetology and Nutrition, Groupement Hospitalier Lyon Sud, Hospices Civils de Lyon, Pierre Bénite, France b
Abstract
Background: A subgroup of obese patients without metabolic disorders has been identified and defined as metabolically healthy but morbidly obese (MHMO). Objectives: To compare Roux-en-Y gastric bypass (RYGB) outcomes between MHMO and metabolically unhealthy morbidly obese (MUMO) patients to assess whether the obesity phenotype could affect the results. Setting: A university-affiliated tertiary care center. Methods: One hundred nineteen consecutive patients underwent RYGB; 102 completed the 2-year follow-up and were divided into 2 groups (MHMO and MUMO) according to Wildman criteria, including blood pressure, triglycerides, high-density lipoprotein cholesterol (HDL-C), fasting blood sugar, C-reactive protein (CRP), and homeostasis model assessment of insulin resistance (HOMAIR). Weight loss and metabolic parameter changes were analyzed. Results: Twenty-one of 102 (20.6%) patients were identified as MHMO; they were mostly women (90.5%) and were significantly younger than MUMO patients (39.4 ⫾ 9.1 yr versus 47.2 ⫾ 10, P ¼ .001); 12.6% were lost to follow-up. MHMO phenotype was significantly associated with a greater percentage of excess body mass index loss (P ¼ .03), independent of gender, age, and redo procedures. All metabolic parameters were significantly improved 2 years after surgery in the MUMO group. HOMA-IR, CRP, and triglycerides were significantly lower 2 years after surgery in the MHMO group, whereas fasting blood sugar and HDL-C were unchanged. At 2 years of follow-up, 92.3% of the population was metabolically healthy. Conclusions: RYGB is an effective procedure to achieve weight loss and had a strong positive metabolic effect in both MHMO and MUMO phenotypes. RYGB led to an increase of the metabolically healthy status and may prevent or delay the onset of metabolic disorders. (Surg Obes Relat Dis 2015;]:00–00.) r 2015 American Society for Metabolic and Bariatric Surgery. All rights reserved.
Keywords:
Morbid obesity; Metabolically healthy; Metabolic disorders; Gastric bypass; Obesity phenotype; Weight loss
*
Correspondence: Maud Robert, M.D., Ph.D., Hôpital Edouard Herriot, Chirurgie Digestive, Pavillon D, 5 Place d’Arsonval, 69437 Lyon Cedex 03, France. E-mail: [email protected]
The increasing prevalence of obesity worldwide has major socioeconomic consequences. Obesity is often associated with metabolic disorders leading to type 2 diabetes mellitus (T2DM), higher cardiovascular risk, and higher
http://dx.doi.org/10.1016/j.soard.2015.04.024 1550-7289/r 2015 American Society for Metabolic and Bariatric Surgery. All rights reserved.
2
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
mortality rate [1,2]. Bariatric surgery is now considered the best option to achieve significant weight loss and has also been found to be superior as an optimized medical therapy to cure T2DM and other co-morbidities [3,4]. Therefore, indications for bariatric surgery are now evolving from weight loss considerations to metabolic objectives. Since the 1980s, a subgroup of obese patients without metabolic disorders has been identified and defined as metabolically healthy but morbidly obese (MHMO) patients [5–7]. Currently, there is still no standardized definition of metabolic health, resulting in a wide variation of metabolically healthy obesity prevalence, which is estimated to be between 6% and 40% [6,8]. MHMO characterized by less visceral fat would be at lower cardiovascular risk [8]. Thus, whereas indications for bariatric surgery tend to expand to nonobese T2DM patients with metabolic concerns, the best therapeutic strategy in MHMO patients, who could be at lower risk of morbidity and mortality, remains questionable. In the present study, we compared gastric bypass outcomes in MHMO versus metabolically unhealthy morbidly obese (MUMO) patients. The aim was to assess whether obesity phenotype influences the results of bariatric surgery and therefore if our surgical management can be based on stratification of obese individuals according to their metabolic health phenotype. Methods Definition of MHMO patients and population selection From April 2009 and May 2011, 119 patients underwent a Roux-en-Y gastric bypass (RYGB) in the Department of Digestive Surgery, Edouard Herriot Hospital, a universityaffiliated tertiary care center (Lyon, France). Fifteen patients were excluded from the study because of insufficient collected data (follow-up performed in another center) and were considered as lost to follow-up. Two others were excluded because of type 1 diabetes mellitus. Finally, 102 patients were included in the study. Patients' characteristics are presented in Table 1. The population was divided into 2 groups, MHMO and MUMO, according to Wildman’s definition [9]. Patients were considered as MHMO when having only 1 or no cardiometabolic abnormalities among the following 6: 1. Elevated blood pressure: systolic/diastolic blood pressure Z130/85 mm Hg or antihypertensive medication use; 2. Elevated triglycerides level: fasting triglycerides level Z1.70 mmol/L; 3. Decreased high-density lipoprotein cholesterol (HDL-C) level: o1.04 mmol/L in men or o1.30 mmol/L in women or lipid-lowering medication use; 4. Elevated glucose level: fasting blood sugar Z5.5 mmol/ L or antidiabetic medication use;
Table 1 Characteristics of the study population Population
n ¼ 102
Age (yr) Male gender (%) Height (m) Weight (kg) Waist Circumference (cm) BMI (kg/m²) T2DM (%) Insulin-treated T2DM (%) Arterial hypertension (%) Dyslipidemia (%) Redo surgeries (%)
45.6 (⫾ 10) 43% 1.66 (⫾ .09) 124.2 (⫾ 20) 131 (⫾ 14) 45 (⫾ 6.6) 51% (n ¼ 54) 21% (n ¼ 21) 41% (n ¼ 45) 46% (n ¼ 47) 27.4% (n ¼ 28)
BMI ¼ body mass index; T2DM ¼ type 2 diabetes. Data are presented as mean ⫾ standard deviation.
5. Insulin resistance: homeostasis model assessment of insulin resistance (HOMA-IR) 42.5 (i.e., 490th percentile); and 6. Systemic inflammation: C-reactive protein (CRP) level 45 mg/L (i.e., 490th percentile). Laparoscopic Roux-en-Y gastric bypass procedure A 5-port technique was used and consisted in a small gastric pouch (30 mL) by stapling the stomach using a linear stapler. The first jejunal loop was moved up into an antecolic position after an epiploic transection. An end-to-side gastrojejunal anastomosis was performed using a linear stapler. Closure of the anterior part of the anastomosis was done using a running suture. The alimentary limb was 150 cm long. A laterolateral jejunojejunal anastomosis was performed with a linear stapler. Closure of the mesenteric defect was systematic, using a nonabsorbable silk suture (2/0). Data analysis This is a retrospective study of prospectively collected data from our electronic database dedicated to bariatric surgery. Anthropometric and biological data were recorded preoperatively; at 2, 6, 12, 18, and 24 months; and every year after surgery. Weight loss was expressed in percentage of excess body mass index (BMI) loss (%EBMIL), absolute weight loss in kg (aWL), and BMI loss in kg/m2. The prevalence and evolution of metabolic co-morbidities in each group were assessed. Type 2 diabetes mellitus was defined using the American Diabetes Association criteria [10], and diabetes was considered in remission when patients stopped their antidiabetic treatments and HbA1c was o6% and/or fasting blood glucose levels were o100 mg/dL. Arterial hypertension was defined as systolic/ diastolic blood pressure Z130/85 mm Hg or use of antihypertensive drugs. Dyslipidemia was defined as lowdensity lipoprotein cholesterol (LDL-C) Z4.1 mmol/L or triglycerides Z1.7 mmol/L.
RYGB Outcomes in Metabolically Healthy Obesity / Surgery for Obesity and Related Diseases ] (2015) 00–00
Because redo surgery has been previously associated with poor weight loss results [11,12], and because of a high percentage of redo procedures in the present study population (MHMO: n = 12/21; MUMO: n = 16/81), weight loss results were analyzed considering previous history of bariatric surgery so as to limit any selection bias. Statistical analysis IBM SPSS Statistics Version 19 (IBM Corp., Armonk, NY, USA) was used for analyses. Baseline quantitative data were compared between groups by a Student t test. When the distribution of variables was nonnormal, the nonparametric test of Wilcoxon was used. Ordinal data were compared between groups by a χ2 test. Weight loss evolution (%EBMIL, BMI, BMI loss, and aWL) was analyzed by mixed ANOVA (3 factors) repeated measures; time was the within factor, and obesity phenotype (MUMO or MHMO) and redo surgery (yes/no) were the 2 between factors. Estimated marginal means of %EBMIL were used for the graphic representation of weight loss in MUMO and MHMO phenotypes. Evolution of metabolic parameters between groups over time was analyzed using ANOVA repeated measure. A multiple linear regression model was used to identify the predictive factors of weight loss. P o .05 was considered significant.
3
Table 2 Anthropometric and metabolic data in metabolically healthy but morbidly obese (MHMO) compared with metabolically unhealthy morbidly obese (MUMO) individuals
Age (yr) Male ratio (%) Height (m) Weight (kg) Waist circumference (cm) BMI (kg/m2) T2DM (%) HTN (%) Dyslipidemia (%) Redo surgeries (%) bands to RYGB Mason to RYGB CRP (mg/L) Triglyceride (mmol/L) HDL-C (mmol/L) Fasting blood sugar (mmol/L) Fasting insulin (mUI/L) HOMA-IR
MHMO
MUMO
n ¼ 21
n ¼ 81
39.4 (⫾ 9.1) 9.5% 1.6 (⫾ .07) 119 (⫾ 16) 125 (⫾ 16) 44 (⫾ 5) 5% 5% 14% 12 (57%) 6 6 6.7 (⫾ 5) 1.1 (⫾ .4) 1.3 (⫾ .2) 5.0 (⫾ 1.7) 8.6 (⫾ 4) 1.7 (⫾ .8)
47.2 (⫾ 10) 35.8% 1.6 (⫾ .09) 125 (⫾ 21) 132 (⫾ 14) 45.2 (⫾ 7) 65% 54% 54% 16 (20%) 12 4 10.6 (⫾ 10) 2 (⫾ 1.1) 1 (⫾ .4) 7.1 (⫾ 3.1) 18.2 (⫾ 22) 5.9 (⫾ 7)
P
.001* .04* .2 .2 .3 .4 o.0005* o.0005* .003* .002*
.1 .0009* .002* .002* .1 .03*
BMI ¼ body mass index; T2DM ¼ type 2 diabetes mellitus; HTN ¼ arterial hypertension; CRP ¼ C-reactive protein; HDL-C ¼ high-density lipoprotein cholesterol; HOMA-IR: homeostasis model assessment of insulin resistance. Data are presented as arithmetical mean ⫾ standard deviation. * Statistically significant.
Results Preoperative characteristics of MHMO versus MUMO groups Using Wildman criteria, 20.6% (n = 21) of the whole population was identified as MHMO. MHMO patients were mostly women (90.5%) and were significantly younger than MUMO patients (39.4 ⫾ 9.1 versus 47.2 ⫾ 10, P = .001). A previous history of bariatric surgery was more common in the MHMO group (57% versus 20%, P = .002). Most revisional procedures were for weight loss failure, after appropriate nutritional management. The others were for functional disorders like reflux or esophageal motility disorders. Characteristics of MHMO and MUMO groups are presented in Table 2. When 1 cardiometabolic abnormality was present in the MHMO group, low-grade systemic inflammation was the most common (38.9%). In MUMO individuals, altered glucose homeostasis was the most prevalent metabolic disorder (75.3% had elevated fasting blood sugar and 67.7% had insulin resistance), followed by altered lipid profile (68.3% had low HDL-C and 55.7% had hypertriglyceridemia). Follow-up and 30-day postoperative complications Mean follow-up after surgery was 31.6 (⫾9) months; 12.6% were lost to follow-up. In the whole population, the rate of minor complications was 4.9% and the rate of major
complications was 3.9%, according to Dindo-Clavien classification. In the primary RYGB group (n ¼ 74) we had 6 30-day postoperative complications (8%): 4 minor (1 pneumonia, 1 ovarian cyst hemorrhage, 2 abdominal sepsis treated and resolved with antibiotherapy) and 2 major (endoluminal hemorrhage of the gastrojejunal anastomosis). In the revisional gastric bypass group (n ¼ 28) we had 3 30day postoperative complications (10%): 1 minor (pneumonia) and 2 major (1 trocar hemorrhage, 1 stenosis of the jejunojejunal anastomosis). Weight loss outcomes Mean %EBMIL at 2 years in the whole population was 66.2% (⫾26). The mean weight was 88.3 kg (⫾18) and the mean BMI was 32.1 kg/m2 (⫾6). Using ANOVA repeated measures to assess the influence of time and obesity phenotype on %EBMIL, we observed a significant effect of time on %EBMIL (P o .001) and a greater %EBMIL in the MHMO group (P ¼ .06). Evolution of estimated marginal mean %EBMIL over time in MHMO compared with MUMO individuals is presented in Fig. 1; estimated marginal mean of %EBMIL at 24 months after RYGB was 71.6% versus 59.5%, respectively. Two years after surgery, BMI in MHMO and MUMO was, respectively, 30.9 and 32.3 kg/m-2; obesity phenotype had no significant effect on BMI over time (P ¼ .08). BMI loss at
4
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
Fig. 1. Graphic representation of percentage of excess body mass index (BMI) loss over time (estimated marginal means) in the whole population according to the obesity phenotype (metabolically healthy but morbidly obese versus metabolically unhealthy morbidly obese). Obesity phenotype effect over time: P ¼ .06 (ANOVA repeated measures). %EBMIL ¼ Excess BMI loss %; MHMO ¼ metabolically healthy but morbidly obese patients; MUMO ¼ metabolically unhealthy morbidly obese patients. MHMO: n ¼ 21, MUMO: n ¼ 81. There were none lost to follow-up at each time-point.
24 months was 13.2 and 13 kg/m-2 in MHMO and MUMO, respectively, and obesity phenotype did not affect BMI loss over time (P ¼ .2). Absolute weight loss at 24 months was 35.6 versus 35.9 kg in MHMO and MUMO respectively. Finally, using ANOVA repeated measures, obesity phenotype did not affect the absolute weight loss over time (P ¼ .3). Redo procedures were significantly associated with lower %EBMIL (P ¼ .001), lower BMI loss (P ¼ .01), and lower aWL (P ¼ .002) over time. Nevertheless, no significant effect of the interaction between obesity phenotype and redo procedures was observed no matter how weight loss results were expressed. Change in %EBMIL over time in MHMO and MUMO groups according to their previous history of bariatric surgery is presented in Fig. 2. In the subgroup of redo procedures, %EBMIL at 24 months was 59% in MHMO versus 49% in MUMO individuals (P ¼ .3). In the subgroup of primary RYGB, %EBMIL at 24 months was 85% (⫾19.6) in MHMO and 70% (⫾24.1) in MUMO (P ¼ .08). Evolution of metabolic parameters Two years after RYGB, 92.3% of the whole population exhibited r1 cardiometabolic disturbance and was considered "metabolically healthy" according to Wildman criteria. Only 7.7% of the whole population still had a metabolically unhealthy phenotype. Changes in fasting blood sugar,
HOMA-IR, triglycerides, HDL-C, and CRP over time are presented in Fig. 3. Using ANOVA repeated measures, we found that each metabolic parameter was significantly improved over time (all were decreased except HDL-C, which was increased). A significant effect of obesity phenotype on the evolution of postoperative glycemia (P ¼ .01) and triglycerides (P ¼ .003) was observed, which was significantly lower in MHMO over time. A trend toward lower HOMA-IR (P ¼ .06) and lower CRP (P ¼ 006) was observed in MHMO during the postoperative period. Changes in metabolic parameters from the preoperative period to the 2-year follow-up are presented in Table 3. All metabolic parameters were improved from baseline to 2 years after surgery in the MUMO group. HOMA, CRP, and triglycerides were significantly lower 2 years after surgery in the MHMO group, whereas glycemia and HDL-C were unchanged. Thus, despite significant decrease in glycemia after surgery in MUMO and no glycemic change in MHMO group, glycemia stayed higher in the MUMO versus the MHMO group (P ¼ .01). Predictive factors of weight loss at 2 years Performing multiple regression analysis and considering the 2-year %EBMIL as the dependent variable, we found that MHMO phenotype was significantly associated with a
RYGB Outcomes in Metabolically Healthy Obesity / Surgery for Obesity and Related Diseases ] (2015) 00–00
5
Fig. 2. Percentage of excess body mass index (BMI) loss over time after Roux-en-Y gastric bypass (RYGB) in metabolically healthy morbidly obese and metabolically unhealthy morbidly obese patients according to their previous history of bariatric surgery (primary RYGB versus revisional RYGB). Obesity phenotype effect: P ¼ .06. Previous bariatric surgery effect: P ¼ .001. Obesity phenotype x previous surgery interaction effect: P ¼ .6. %EBMIL ¼ excess BMI loss %; MHMO ¼ metabolically healthy but morbidly obese patients; MUMO ¼ metabolically unhealthy morbidly obese patients. MUMO primary RYGB: n ¼ 65, MHMO primary RYGB: n ¼ 9, MUMO revisional RYGB: n ¼ 16, MHMO revisional RYGBP: n ¼ 12. There were none lost to follow-up at each time point.
greater %EBMIL (P ¼ .03), independently of gender (P ¼ .8), age (P ¼ .4), and redo procedures, which were negatively correlated to %EBMIL (P ¼ .001). Discussion The MHMO phenotype is a quite new concept that was first described in the 1980s [13,14]. Whereas obesity is often associated with metabolic abnormalities leading to an increased risk of T2DM and cardiovascular disease, interest has grown for this paradoxical condition in which patients are still metabolically healthy despite excessive body fat. Indeed, understanding the determinants of the metabolically healthy status is a matter of great interest and could lead to change and improve the management strategies of obesity. Data in the literature regarding the prevalence and the determinants of the MHMO phenotype are heterogeneous and conflicting. First, there is still no standardized definition of metabolic health, making it difficult to compare the studies [6]. Second, variation among ethnicities, sample size, age groups, and study design can also account for the wide heterogeneity of the data. The prevalence of MHMO could be 6%–40%, depending on the studies and the criteria used [5,8]. Several definitions have been proposed to define MHMO and MUMO phenotypes, based on the cardiometabolic risk assessment. Whereas the first definitions were specifically based on the metabolic syndrome parameters, Wildman has proposed to add surrogate biomarkers of insulin resistance (i.
e., HOMA-IR) and inflammation (i.e., CRP). Thus, this last definition takes into account pathophysiologic traits of obesity that have been linked with T2DM risk and cardiovascular disease [8]. Therefore, we also used Wildman criteria to define MHMO individuals in our study [9]. It is of note that, using these criteria, we found a 20.6% rate of MHMO patients in our population, which is inferior to the 31.7% MHMO prevalence reported by Wildman et al. and by others [15,16]. This can be explained by the fact that we focused on a population with several co-morbidities and metabolic disorders, in whom gastric bypass is the procedure of choice because of its metabolic effect. We also found that our MHMO population was mostly female (90.5%) and significantly younger than MUMO patients (39.4 ⫾ 9.1 versus 47.2 ⫾ 10, P = .001), which is consistent with literature data [17,18]. The hypothesis that gynecoid fat distribution in women, who thus have less visceral fat than men, could be one of the determinants of MHMO status has been supported by many authors [19,20]. Indeed, visceral fat has been associated with insulin resistance and cardiovascular risk in several studies [21,22]. Nevertheless, some authors did not find any correlation between visceral fat assessed by waist circumference and MHMO phenotype, arguing that difference between MHMO and MUMO individuals mostly is due to the total amount of body fat accretion rather than body fat distribution [8]. In our study population, MHMO and MUMO had similar waist circumference (125 cm ⫾ 15.7 versus
6
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
Fig. 3. Postoperative evolution of (A) glycemia, (B) HOMA-IR, (C) triglycerides, (D) HDL cholesterol, (E) C-reactive protein, and (F) BMI at 6, 12, and 24 months in metabolically unhealthy morbidly obese (gray line) and metabolically healthy but morbidly obese (black line) phenotypes. Results are presented as arithmetical means and error bars represent standard error of the mean. Effect of obesity phenotype over time on glycemia (P ¼ .01), HOMA-IR (P ¼ .06); triglycerides (P ¼ .003), HDL-C (P ¼ .1), C-reactive protein (P ¼ .06). HOMA-IR ¼ homeostasis model assessment of insulin resistance; HDL ¼ highdensity lipoprotein; BMI ¼ body mass index. MHMO: n ¼ 21, MUMO: n ¼ 81. There were none lost to follow-up at each time point.
132 ⫾ 13.8, P ¼ .3). Beside whole body fat distribution and accretion, several histologic and functional features of adipose tissue in obesity have been associated with insulin resistance and metabolic disorders, including immune cells infiltration, proinflammatory cytokines secretion, fibrosis, adipocyte hypertrophy [23]. Thus, metabolic health status in obese patients is obviously not easy to assess with simple clinical and biological parameters such as BMI, waist circumference, glycemia, and lipid profile. In our study population, we found that low-grade inflammation was the most common metabolic disorder in the MHMO group (38.9% of patients with CRP 45 mg/L). This is consistent with the finding of Wildman et al. reporting that, despite no increase in cardiovascular risk in MHMO women, these individuals still displayed abnormal levels of inflammatory markers [24]. This could suggest that
MHMO individuals are at risk for insulin resistance and could evolve into a MUMO state. This hypothesis is supported by the study of Soriguer et al., who recently assessed the evolution of the metabolic health status at 6 and 11 years of follow-up in 1051 individuals representative of the general population [25]. They found that in MHMO patients the risk for becoming diabetic was lower than in MUMO patients, but this risk remained significant (odds ratio ¼ 3.13; 95% confidence interval ¼ 1.07–9.17; P ¼ .02). They also reported that in patients who lost weight during the study, the association between MHMO phenotype and T2DM incidence disappeared, even after adjusting for HOMA-IR. These results suggest that MHMO phenotype could be a changing and evolving state that should be considered over time and that MHMO individuals would gain metabolic benefit from weight loss.
RYGB Outcomes in Metabolically Healthy Obesity / Surgery for Obesity and Related Diseases ] (2015) 00–00 Table 3 Evolution of the metabolic parameters in metabolically healthy but morbidly obese (MHMO) and metabolically unhealthy morbidly obese (MUMO) patients 2 years after the gastric bypass Metabolic parameters
Obesity Preoperative 24 months after phenotype values surgery values
Fasting blood MHMO sugar (mmol/L) MUMO HOMA-IR MHMO MUMO Triglycerides MHMO (mmol/L) MUMO HDL-C (mmol/L) MHMO MUMO CRP (mg/L) MHMO MUMO
P
4.6 (⫾ .4)
4.5 (⫾ .5)
.1
7.2 (⫾ 3) 1.7 (⫾ .8) 5.5 1.1 (⫾ .5)
5.4 (⫾ 2) .7 (⫾ .3) 1.4 (⫾ 2.2) .8 (⫾ .2)
.0001* .003* o.001* .005*
2 (⫾ 1.1) 1.3 (⫾ .2) 1.1 (⫾ .3) 4.9 (⫾ 3.8) 10.1 (⫾ 8)
1.1 (⫾ .5) 1.4 (⫾ .2) 1.4 (⫾ .3) 1 (⫾ .7) 2.7 (⫾ 4.4)
o.0001* 1.5 o.0001* .03* o.0001*
HOMA-IR ¼ homeostasis model assessment of insulin resistance; HDLC: high-density lipoprotein cholesterol; CRP ¼ C-reactive protein. * Statistically significant.
In our study, we found a significant improvement in all metabolic parameters over the postoperative period and a significant favorable effect of MHMO status compared with MUMO phenotype on glycemia (P ¼ .01) and triglycerides (P ¼ .03). A favorable trend to improve HOMA-IR (P ¼ .06) and CRP (P ¼ .06) after surgery was also observed in the MHMO phenotype (ANOVA repeated measures). Finally, weight loss induced by RYGB had a beneficial effect on the metabolic profile of both MUMO and MHMO patients. Nevertheless, patients with preoperative MHMO phenotype still had a better biological profile of metabolic health 24 months after surgery than those with the preoperative MUMO phenotype, despite the strong decrease of MUMO phenotype prevalence 24 months after surgery. Thus, despite massive weight loss, MHMO and MUMO patients are still discernible after bariatric surgery. These results seem to indicate that MUMO phenotype is not totally reversible into a metabolically healthy condition. This is a further argument for considering that MHMO and MUMO phenotypes are different stages in the evolution of obesity disease. Few studies have assessed the outcomes of bariatric surgery in MHMO patients compared with MUMO individuals [18,26,27]. In a cohort of 190 patients who underwent gastric banding, Sesti et al. assessed weight loss and metabolic outcomes at 6 months of follow-up in the subgroups of MHMO and MUMO individuals. In this short-term study, the authors reported similar weight loss in both groups and a significant correlation between weight loss and improvement of insulin sensitivity [26]. Jimenez et al. studied the effects of bariatric surgery (RYGB and sleeve gastrectomies) on metabolic co-morbidities and anthropometric and inflammatory markers at 12 months of follow-up between 52 obese women presenting with a
7
HOMA-IR o2.94 and an age- and BMI-matched group of women with HOMA-IR 42.94 [27]. The authors concluded that bariatric surgery also results in improvement of the altered metabolic and inflammatory profiles in the insulin sensitive, morbidly obese group. More recently, Goday et al. conducted a prospective cohort study on 222 morbidly obese patients who underwent either a RYGB (n ¼ 135) or a sleeve gastrectomy (n ¼ 87) [18]. Forty-two (18.9%) patients fulfilled the criteria for MHMO phenotype and had similar weight loss to MUMO patients at 1 year of follow-up. In that study, MHMO patients exhibited a significant improvement of plasma glucose, HOMA-IR, total cholesterol, LDL-C, and HDL-C after surgery. In the present study, we noted a trend toward a favorable effect of MHMO phenotype on %EBMIL (ANOVA repeated measures, P ¼ .06) over time, and MHMO phenotype appeared to be an independent factor of greater weight loss (multiple regression, P ¼ .03). An interesting hypothesis could be that adipose tissue from MUMO is structurally different from adipose tissue from MHMO. Because of increased inflammation and fibrosis, reduction of adipose tissue from MUMO individuals could be limited and could affect weight loss negatively. Such a stimulating hypothesis needs to be tested in further clinical studies. This study has some limitations. The small sample size (n ¼ 102) is explained by the careful selection of RYGB with complete biological and anthropometric data during more than 24 months of follow-up. There was also a large prevalence of redo procedures in the MHMO group (57.1%), which could be a selection bias; first of all, it could lead to underestimating weight loss outcomes in this group compared with the MUMO patients. Indeed, we and others also had reported the negative effect of redo procedures on weight loss results [11,12]. When analyzing weight loss after excluding redo procedures, we found the 2 years %EBMIL tended to be greater in the MHMO group versus the MUMO group (85% ⫾ 19.6 versus 70% ⫾ 24.1 respectively, P ¼ .08). These results seem to confirm that bariatric surgery is as effective or more effective in MHMO than in MUMO individuals for achieving weight loss. Second, one could question if patients with a history of redo procedures are "real" MHMO individuals or if they became metabolically healthy as a result of the first bariatric procedure. However, as previously discussed, we believe that metabolic health status is a dynamic condition and that the aim of bariatric surgery is to avoid the onset of metabolic disorders whatever the previous history of obesity. Because of the study design, we could not assess the impact of the type of surgical procedure (malabsorptive versus restrictive) on metabolic outcomes. One might question the interest of performing an RYGB procedure, often considered a metabolic technique and which is at higher risk of complications, rather than a restrictive
8
E. Pelascini et al. / Surgery for Obesity and Related Diseases ] (2015) 00–00
procedure in metabolically healthy individuals. However, we believe that RYGB, by achieving good weight loss and durable results, is still an interesting option. Obviously, further trials are needed to answer this question and to confirm the scarce literature data. Conclusions RYGB appears to be an effective procedure to achieve weight loss in both metabolically healthy and metabolically unhealthy morbidly obese patients. RYGB produced a strong metabolic effect, leading to an increase of the metabolically healthy status in our study population from 20.6% to 92.3% at 2 years of follow-up, suggesting that metabolic health is a dynamic condition and not a definitive state. Even in metabolically healthy individuals, bariatric surgery has a favorable metabolic impact that could prevent or delay the onset of metabolic disorders. Further studies are needed to determine whether restrictive procedures are as effective as RYGB on weight loss and metabolic outcomes in MHMO individuals so as to define the best management strategies in this specific obesity phenotype. Disclosures The authors have no disclosures and no commercial interest in the subject of study. References [1] Sjostrom CD, Lystig T, Lindroos AK. Impact of weight change, secular trends and ageing on cardiovascular risk factors: 10-year experiences from the SOS study. Int J Obes (Lond) 2011;35 (11):1413–20. [2] Van Gaal LF, Mertens IL, De Block CE. Mechanisms linking obesity with cardiovascular disease. Nature 2006;444(7121):875–80. [3] Schauer PR, Bhatt DL, Kirwan JP, et al. Bariatric surgery versus intensive medical therapy for diabetes–3-year outcomes. N Engl J Med. 2014;370(21):2002–13. [4] Guo X, Liu X, Wang M, Wei F, Zhang Y. The effects of bariatric procedures versus medical therapy for obese patients with type 2 diabetes: meta-analysis of randomized controlled trials. Biomed Res Int 2013;2013:410609. [5] Ruderman NB, Schneider SH, Berchtold P. The "metabolicallyobese," normal-weight individual. Am J Clin Nutr 1981;34 (8):1617–21. [6] Phillips CM. Metabolically healthy obesity: definitions, determinants and clinical implications. Rev Endocr Metab Disord 2013;14 (3):219–27. [7] Sims EA. Are there persons who are obese, but metabolically healthy? Metabolism 2001;50(12):1499–504. [8] Martinez-Larrad MT, Corbaton Anchuelo A, Del Prado N, et al. Profile of individuals who are metabolically healthy obese using different definition criteria. A population-based analysis in the Spanish population. PLoS One 2014;9(9):e106641.
[9] Wildman RP, Muntner P, Reynolds K, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999-2004). Arch Intern Med 2008;168(15):1617–24. [10] American Diabetes Association. Diagnosis and classification of diabetes mellitus. Diabetes Care 2010;33(Suppl 1):S62–9. [11] Robert M, Pasquer A, Espalieu P, et al. Gastric bypass for obesity in the elderly: is it as appropriate as for young and middle-aged populations? Obes Surg 2014;24(10):1662–9. [12] Slegtenhorst BR, van der Harst E, Demirkiran A, et al. Effect of primary versus revisional Roux-en-Y gastric bypass: inferior weight loss of revisional surgery after gastric banding. Surg Obes Relat Dis 2013;9(2):253–8. [13] Sims EA, Berchtold P. Obesity and hypertension. Mechanisms and implications for management. JAMA 1982;247(1):49–52. [14] Andres R. Effect of obesity on total mortality. Int J Obes 1980;4 (4):381–6. [15] Lopez-Garcia E, Guallar-Castillon P, Leon-Munoz L, RodriguezArtalejo F. Prevalence and determinants of metabolically healthy obesity in Spain. Atherosclerosis 2013;231(1):152–7. [16] Iacobellis G, Ribaudo MC, Zappaterreno A, Iannucci CV, Leonetti F. Prevalence of uncomplicated obesity in an Italian obese population. Obes Res 2005;13(6):1116–22. [17] Phillips CM, Dillon C, Harrington JM, et al. Defining metabolically healthy obesity: role of dietary and lifestyle factors. PLoS One 2013;8 (10):e76188. [18] Goday A, Benaiges D, Parri A, et al. Can bariatric surgery improve cardiovascular risk factors in the metabolically healthy but morbidly obese patient? Surg Obes Relat Dis 2014;10(5):871–6. [19] Brochu M, Tchernof A, Dionne IJ, et al. What are the physical characteristics associated with a normal metabolic profile despite a high level of obesity in postmenopausal women? J Clin Endocrinol Metab 2001;86(3):1020–5. [20] Marin P, Andersson B, Ottosson M, et al. The morphology and metabolism of intraabdominal adipose tissue in men. Metabolism 1992;41(11):1242–8. [21] Dagenais GR, Yi Q, Mann JF, et al. Prognostic impact of body weight and abdominal obesity in women and men with cardiovascular disease. Am Heart J 2005;149(1):54–60. [22] Brochu M, Poehlman ET, Ades PA. Obesity, body fat distribution, and coronary artery disease. J Cardiopulm Rehabil 2000;20(2):96–108. [23] Rajala MW, Scherer PE. Minireview: the adipocyte—at the crossroads of energy homeostasis, inflammation, and atherosclerosis. Endocrinology 2003;144(9):3765–73. [24] Wildman RP, Kaplan R, Manson JE, et al. Body size phenotypes and inflammation in the Women's Health Initiative Observational Study. Obesity (Silver Spring) 2011;19(7):1482–91. [25] Soriguer F, Gutierrez-Repiso C, Rubio-Martin E, et al. Metabolically healthy but obese, a matter of time? Findings from the prospective Pizarra study. J Clin Endocrinol Metab 2013;98(6):2318–25. [26] Sesti G, Folli F, Perego L, Hribal ML, Pontiroli AE. Effects of weight loss in metabolically healthy obese subjects after laparoscopic adjustable gastric banding and hypocaloric diet. PLoS One 2011;6 (3):e17737. [27] Jimenez A, Perea V, Corcelles R, et al. Metabolic effects of bariatric surgery in insulin-sensitive morbidly obese subjects. Obes Surg 2013;23(4):494–500.
