Roux-en-Y gastric bypass and Sleeve gastrectomy does not affect food preferences when assessed by an ad libitum buffet meal

Abstract

BackgroundRoux-en-Y gastric bypass (RYGB) and Sleeve gastrectomy (SG) lead to a reduction in energy intake. It is uncertain whether this reduction is simply due to eating smaller portions orif surgerychangesfood preferences towards less energy-dense food. Previous results rely on verbal reports, which may beprone to recall bias and underestimation of especially unhealthy foods.

MethodsUsing an ad libitum buffet meal targeting direct behavior,we investigated if RYGB and SG surgery leads to changes in food preferences. In addition, we assessed food preferences by a picture display test to explore differences between a method relying on verbal reportand a method assessing direct behavior.

ResultsForty-one subjects (BMI: 45.0±6.8kg/m2) completed a visit pre and 6 months postRYGB (n=31) and SG (n=10). Mean BMI decreased with 11.7±0.6kg/m2and total energy intake at the buffet meal with54% (4491±208kJ vs. 2083±208kJ, P<0.001), respectively.However, relative energy intake from the following food categories: high-fat, low-fat, sweet, savory,high-fat-savory, high-fat-sweet, low-fat-savory and low-fat-sweet, as well as energy density did not change following surgery (all P≥0.18). In contrast, the picture display test showed that food from the low-fat-savory group was chosen more often post-surgery (34±8% vs. 65±9%, P=0.02).

ConclusionThe reduction in energy intake after RYGB and SG surgery and the subsequent weight loss seems to be primarily related to a reduction in portion sizesand not by changes in food preferences towards less energy-dense foods. These results underline the necessity of investigating eating behavior by targeting direct behavior.

Key words: eating behavior, food preferences,ad libitumbuffet meal, roux-en-y gastric bypass, sleeve gastrectomy

Introduction

Bariatric surgery leads to a decrease inenergy intake(1-5), with reported reductions of about 40-50%sixmonths post-surgery(1,5). Energy intake gradually increases during the first two post-operative years(1,2); however, the reduction in energy intake does not seem to return to pre-surgery levels, andisstill significantly reduced 8-10 years after surgery(4,6).Whether this decrease in energy intake is simplydue to eating smaller portions of the same food items or a change in food preferences towards less energy dense foods(7) remains uncertain. Available studiessuggest that both Roux-en-Y gastric bypass (RYGB) and Sleeve gastrectomy(SG) surgery affect food preferences leading to reduced intake of high-fat and sweet foods(1,2,8-12).However, anessential limitation inprevious studiesis the use offood records, questionnaires and interviews, which are prone to recall bias and underestimation of food intake(13,14), especially of unhealthy foods(15).

Animal studies seem to replicate findings in humans, reporting a reduction in preferences for high-fat and sweet foods after RYGB(9,16-18) and SG surgery(19,20).This suggests that potential changes in food preferences could bedriven by mechanisms other than an effectofdietary counsellingorincreased motivation for weight loss(21).

Proposed mechanisms include changes in taste sensitivity for sweet taste, and decreased hedonic value for sweet and fatty foods, potentially driven by the alterations in taste sensitivity(7,22). These changes might cause an unconditioned shift in food selection towards less sugary and fatty foods (23). Another proposed mechanism is the experience of unpleasant post-ingestive responses such as nausea and abdominal pain after surgery, especially after intake of sugary and fatty foods.This unpleasant response can lead to avoidance of the triggering food through a learned process, and thereby modify eating behaviour(23).

Using an ad libitum buffet meal targeting direct behavior,we investigated if RYGB and SG surgery leads to changes in food preferences.We hypothesized that RYGB and SG surgery would lead to a relative reduction in intake of high-fat and sweet foods and a relative increase in intake of low-fat and savoury foods. In addition, we assessed food preferences by a picture display test, to explore potential differences between a method relying on verbal report and a method assessing direct behavior.Furthermore, we examined whether changes in food preferences were associated with experiences of unpleasantpost-ingestive responses.

Methods

This study was part of a larger study, the GO Bypass study.The aim of the GO Bypass study was to identify the multiple factors determining the variation of weight loss after RYGB and SG surgery using a novel interdisciplinary approach.

Study population

From March 2014 to July 2015patients scheduled for RYGB or SG surgery at Bariatric Clinic, Køge Hospital, Denmark,were recruited. Eligibility for surgery required: age ≥25y and BMI ≥ 50 kg/m2 or BMI ≥ 35 kg/m2combined with type 2 diabetes, obstructive sleep apnea, hypertension, polycystic ovary syndrome or osteoarthritis in the lower extremities (local criteria for bariatric surgery).Exclusion criteria included pregnancy as well as inability, physically or mentally, to comply with the procedures required by the study protocol.In total, 61 participants were recruited. Six participants dropped out prior to the first visit (rejected/declined surgery n=3, withdrawal of informed consent n=1, missing contact n=2) and 12 dropped out shortly after the first visit (rejected/declined surgeryn=6, surgery postponed indefinitely n=4, SG surgery n=2 [at n=22 we changed inclusion criteria’s to include SG surgery]), leaving a total sample of 43 participants. The study was approved by the Scientific Ethic Committees of the Capital Region of Denmark (J.no H-3-2013-138) and registered in the database (ID no NCT02070081). All study participants gave written informed consent.

Design

Food preferences were assessed approximately 3 months pre-surgery and 6 months post-surgery. The pre-surgery visit was held before the participants received nutritional counseling from a dietician, in order to obtain a mandatory 8% weight loss as preparation for surgery (local requirement). At each visit, the subjects arrived at Department of Nutrition, Exercise and Sports at 9 AM, whereanthropometric data were collected. In order to standardize energy intake prior to the buffet meal, a standardized liquid meal was served at 10 AM (Cambridge Weight Plan®, 954 kJ) and at 1.15 PM (Cambridge Weight Plan®, 1674 kJ). At 3.30 PM the picture display test was carried out, and at 4.30 PM the buffet meal was served.

The surgical procedures

RYGB and Sleeve gastrectomy were performed laparoscopically with the patient in a 10 to 20 degree anti-Trendelenburg position using 5 trocars. In gastric sleeve resection, using a 34-FR calibration tube, a vertical resection of the greater curvature is performed to reduce the size of the stomach to 80 to 100 ml.In RYGB, dividing the stomach with stapling instruments creates a small gastric pouch of approximately 30 ml. Sixty cm from the ligament of Treitz the jejunum is divided and the distal part of the small bowel is then anastomized to the pouch by placing the alimentary limb antecolic and antegastric. The oral part of the transected jejunum is anastomized to the alimentary limb at 120 cm distal to the pouch. This leaves an alimentary limb of 120 cm, a biliary limb of 60 cm and a common limb of the rest of the small bowel.

Outcomes

The buffet meal test:The following 20 food items were served at the buffet meal: pork rib roast, chicken fillet strips, fish cakes, nuggets, Danish omelet, French fries, creamy potato gratin, bread (rye bread, baguette and crispbread), ketchup, remoulade, mayonnaise, skyr with berries, cut raw vegetables, cut fruit, vanilla ice cream, chocolate sauce, cocoa meringues, biscuit cones with chocolate, sweet liquorice, and Danish pastries. The only liquid served at the meal was water. The participants were instructed to eat according to their preferences and for as long as they wanted. They ate unaccompanied, in order to diminish social desirability bias. The staging of the 20 food items wasidentical at each visit.

The 20 food items varied along two dimensions associated with food choices - fat (high or low) and taste (sweet or savory) (24), thereby making it possible to organize the food itemsinto separate categories (high-fat and low-fat or sweet and savory) or combined food categories (high-fat savory, low-fat savory, high-fat sweet and low-fat sweet) (supplementary table 1).Total energy intake, intake from each of the food categories and the duration of the meal were registered as g eaten from each food item and as time from the participants started the meal until they left the test room.The participants were unaware of this registration. As a distraction from considerations of healthy food choices, a blood sample was drawn after the meal, and the participants were informed that the aim was to investigate how “eating according to preferences”affected different hormones in the blood.

The picture display test: Participants were presented to standardized pictures of the following 20 food items: pork rib roast,turkey strips, nuggets, Danish omelet, French fries, crispbread, salty crackers,skyr with berries, carrots, cut fruit, cheese, smoked fillet (cold cuts),vanilla ice cream,cocoa meringues, sweet liquorice,Danish pastries, milk chocolate, pound cake, cookies and wine gum. Pictures were displayed in a randomized order and the participants were instructed to choose the food item they would most like to eat.

Previous experience of unpleasant post-ingestive responses: Assessed by a semi-structured qualitative interview 6 weeks and 6 months after surgery. Data from the interviews were quantified into a 4-point score (0-3), with 0 indicating no experience of discomfort related to eating after surgery, whereas 1, 2 and 3 indicate low, medium and high levels of discomfort, respectively.

Pre-mealhunger:Assessed by a visual analogue scale (100 mm) approximately 15 minutes prior to the buffet meal.

Anthropometric measures:Body weight and height were measured after an overnight fast.

Statistical analysis

Descriptive data summaries are presented as mean ± standard deviation (SD).

Pre-meal hunger, total energy intake, energy density, eating time, eating rate, and weight changes were analyzed using linear mixed models with visit as fixed effect.

Intake from the buffet meal was analyzed using a two-stage approach to accommodate for semi-continuous data (consisting of zero intake and positive intake;(25)). First a logistic mixed-effects model was fitted to the binary data indicating if intake was 0 or not. Second a linear mixed model was fitted to the square-root transformed positive intake values only. Both models included the food category-visit interaction as fixed effect. Estimates from these two analyses were then combined to obtain unconditional estimated mean intakes per category and visit and differences in mean intakes between combinations of category and visit. Binary outcomes from the picture display test were analyzed using logistic mixed-effects models (one model per category) with visit as fixed effect.

All models were adjusted for age and gender and included subject-specific random effects. Results are reported as mean ± standard error (intake as mean percentages). Associationsbetween changes in intake at the buffet meal test and previous experience of post-ingestive responseswere evaluated using simple linear regression adjusted for age and gender.P-values<0.05 were considered significant. Statistical analyses were carried out using R (26) and RStudio version 0.99.896 (

Results

Participants

Forty-three subjects participated in the study; however, one subject had missing data from the pre-surgery visit due to illness on the test dayand one subject was unable to attend the post-surgery visit. These two subjects were excluded from the analysis, leaving a sample of 41 subjects. Thirty-one subjects underwent RYGB surgery and 10 underwent SG surgery.

The characteristics of the participants are presented in Table 1. Mean age was 39.8±9.4years and mean BMI 45.0±6.8kg/m2. Ten subjects were diagnosed with type 2 diabetes prior to the surgery, of which seven were under medical treatment.

Weight loss

BMI decreased with 11.7±0.6kg/m2 (P<0.001)sixmonths post-surgery. RYGB and SGsurgery resulted in acomparable decrease in BMI (RYGB: 12.2±0.2kg/m2 vs. SG: 10.4±0.5kg/m2; P=0.15), whereas non-diabetics had a larger decrease compared to diabetics (non-diabetics: 12.6±0.2 kg/m2 vs. diabetics: 8.9±0.4 kg/m2; P=0.004).

Buffet meal test and pre-meal hunger

Pre-meal hunger did not differ pre- to post-surgery (P=0.31), yet total energy intake at the buffet meal decreased with 54% (4491±208kJ vs. 2083±208kJ, P<0.001) andeating rate decreased with 28% (293±15kJ/min vs. 211±15kJ/min, P<0.001). The energy density of the meal did not change pre- to post-surgery (Table 2).

Relative energy intakefrom thecombined food categories: high-fat savory, low-fat savory, high-fat sweet and low-fat sweet,and the separate food categories: high-fat vs. low-fat and sweet vs. savorydid not differ between the pre- and post-operative visits(Table 2). Likewise, no differences were observed in relative energy intake from each of the 20 food items served at the buffet(Supplementary Table 2).

Previous experiences of unpleasant post-ingestive responses were not associated with intake of high-fat or sweet foods, respectively (Table 3).

Relative energy intake from the combined and separate food categories did not differ between surgical groups (all P≥0.45) (supplementary table 3) or between diabetics and non- diabetics (all P≥0.47), justifying combining data in one pooled analysis.

Picture Display test

Pictures from the low-fat savory group were chosen more often post-surgery compared to pre-surgery in the picture display test(34±8% vs. 65±9%, P=0.02). Subjects’ choice of picture from the high-fat savory, high-fat sweet and low-fat sweet groups did not change after surgery(all P≥0.21) (Table 4).

Discussion

The main finding in this study was that the post-operative reduction in energy intake following RYGB and SG surgerywas not explained by changes in food preferences, since we found no changes in intake of food from pre-defined food categories and no changes inenergy densityfollowing bariatric surgery.This indicates that the post-operative reduction in energy intake and the subsequent weight loss is caused simply by eating smaller portions of the same food items. These data were obtained using an ad libitum buffet meal test,which has neverbefore been applied in a population of bariatric patients. Interestingly, when measuring food preferences by means of verbal report, we found an increasedpreference for low-fat savory foods following surgery, which is in line with post-operative dietaryguidelines. This underlinethe necessity of including methods targeting direct behaviorsand not exclusively rely on verbal report measures, when investigating complex behaviors.

Previous studies using verbal report to assess changes in food preferences report conflicting results both in regard to the effect on intake from specific food groups and macronutrient composition(7).While decreased intake of sugary and fatty foods has been reported in some studies(1,2,8,10,12,27)this is not consistently seen across all studies and at all time points after surgery(1,2,10,12). When changes in food preferences are based on macronutrient intake, intake of calories from fat and carbohydrateswas unchanged(1,4) or decreased(1,2,8)whereas intake of calories from protein was unchanged(4)or increased(1). Thus, in combination with differences in study designs, the methods for analyzing data seem to be a likely explanation for the inconsistent findings.However, the main limitationof these findings is that they rely on verbal reports. Inaccurate reporting,especially among obese people who are often confronted with the social stigma of obesity, is well described(28). Thus,previous results based on self-report must be interpreted with caution as they do notnecessarily reflect actual behavior, as supported byour findings.

Astrength of the buffet meal test is that it targets both consummatory and appetitive eating behaviors, whereas verbal report measures, such as the picture display test, have the limitation of only targeting appetitive eating behavior.At the buffet meal, the subjects are exposed to differentsensory cues,includingsmell and sight of food thattriggers appetitive behavior as well as taste and texture that triggersconsummatory behavior. Together these triggers direct food preferences and portion selection (29) and are mechanisms at play in everyday eating behavior. These triggers are neglected when food behavior is investigated by questionnaires and interviews.

One proposed mechanism for changes in food preferences after surgical treatment include a learned restrictive behavior linked to experiences of unpleasant post-ingestive responses after intake of especially high-fat and sugary foods(22,23).However, we were not able to find support for this hypothesis, since we found no association between post-operative experiences of negative responses related to eating and food preferences at the buffet meal. It has been suggested, that negative post-ingestive signalsin response to food intake lead toconditioned food avoidance elicitedthrough a learned process. The triggering foods can still be consumed, as long as the patients learn to stop consuming quantities that triggers a negative response(22). It can be argued that 6 months post-surgery might be too early for detecting changes in behaviors related to negative response from eating certain foods as patients are more compliant to the prescribed post-operative diet during the early post-operative period. Yet, nearly all of the included patients (98%) reported one or more experiences with unpleasant physiological responses to food after surgery,indicating that the patients in our study have deviated from the strict post-operative diet and exposed themselves to high-fat and sugary foods. However, despite this high frequency of patients reporting experience with unpleasant responses related to eating, we were unable to detect any differences in intake from the high-fat and sweet categories, and the majority of calories consumed post-surgery was still from thehigh-fat-savory category. Results from animal studies show that RYGB operated rats do not completely avoid high-fat and sugary foods, but still consume the majority of their calories from this category, indicating that high-fat and sugary foods are still liked(17); however, it seems that they learn at what threshold they can avoid an unpleasant response.It is likely that total energy intake post-surgery in our study was so low that intake of food from the high-fat and sweet categories was below a threshold for unpleasant responses.