Prospective Study of Appetitive Sensations after Metabolic and Bariatric Surgery Compared with Low-Calorie Diet

Abstract

Introduction

Metabolic and bariatric surgery (MBS) causes greater weight loss (WL) compared with low-calorie diet (LCD) that may be due to changes in appetite and gut hormones. The objective of this study was to quantify appetitive sensations in individuals who underwent LCD or MBS at equivalent weight loss (T2) and at one year (T3).

Methods

Visual analog scales (VAS) and the food craving inventory were used to assess appetite. Glucagon-like peptid-1 (GLP-1) levels were quantified by ELISA.

Results

Participants available for both T1 and T2 were as follows: LCD (n = 15), surgery (n = 24). By T3, LCD (n = 12) and surgery (n = 15). At T2, percent total WL (%TWL) was similar between LCD vs surgery (14.9% vs 14.6%; p = 0.94). At T3, there was greater %TWL after surgery compared with LCD (30.2% vs 14.6%; p < 0.0001). At T2, there was a significant increase in postprandial fullness and decreases in hunger and prospective eating only after surgery. Cravings decreased in both groups at T2 but remained decreased at T3 only after surgery. There was almost a two-fold increase from T1 to T2 in postprandial GLP-1 after surgery (p < 0.0001) that correlated with the increase in fullness (r = 0.69; p = 0.038); no change was noted after LCD (p = 0.34).

Conclusions

After equivalent WL, MBS results in favorable changes in appetitive sensations and GLP-1 levels compared with LCD. Such changes may support the ability to achieve greater reduction in body weight after MBS.

Supplementary Information

The online version contains supplementary material available at 10.1007/s11695-025-08473-5.

Introduction

Metabolic and bariatric surgery (MBS) is currently the most effective treatment for obesity that results in greater and more sustained total weight loss (TWL) compared with nonsurgical approaches [1, 2]. Effectiveness of the most common bariatric procedures, Roux-en-Y Gastric Bypass (RYGB) and Sleeve Gastrectomy (SG), is due in part to restriction of nutrient flow and likely to beneficial changes in key regulators of eating behavior and appetite control [3–10]. MBS is involved in the deactivation of the central reward system and the activation of inhibitory control toward food exposure through alterations in gastrointestinal hormones [11, 12]. Modifications of the gut anatomy increase postprandial concentrations of so-called satiety hormones, glucagon-like peptide-1 (GLP-1) and peptide YY (PYY), that may favorably impact weight loss [7, 9, 10, 12, 13]. In contrast, diet-induced WL results in increased levels of the orexigenic (appetite stimulating) hormone ghrelin with little to no change in anorexigenic hormones creating an environment that favors weight regain [10, 14, 15].

Sensations of hunger, fullness, and desire to eat may affect the ability to lose and maintain reduced body weight although results on changes in appetitive sensations after diet-induced WL vary showing no change, increased hunger, and increased or decreased ratings of prospective food consumption [9, 10, 14, 16–19]. Studies comparing the impact of MBS or LCD also report variable results indicating no change in appetite in either group or decreased hunger and increased fullness after MBS compared with LCD [9, 10, 18]. Discrepancies between studies may be due in part to differences in study design, such as amount of weight loss, duration of weight loss and size of meal or composition, or characteristics of the study cohort. Therefore, the goal of this study, which was a secondary endpoint to a study that will analyze changes in the plasma proteome, was to quantify subjective appetitive sensations at matched WL and at one year follow-up to understand mechanisms for greater weight loss after MBS compared with LCD.

Methods

Ethical Considerations

Data for this study was collected as part of a longitudinal weight loss intervention trial that was conducted at an academic Medical Center. We certify that all applicable institution and government regulations concerning the ethical use of human volunteers were obeyed during this study. The study was approved by the Columbia University Institutional Review Board and written informed consent was obtained from all individual participants included in the study.

Participants

Adult participants (age 18–70 years) with obesity (BMI ≥ 35 kg/m²) were enrolled in the main study in which the primary endpoint (unpublished) is to quantify changes in the plasma proteome with different weight loss interventions. The data presented in this analysis is from secondary endpoints. Main exclusion criteria were the presence of type 2 diabetes, major cardiovascular, kidney, liver, or psychiatric disease, current eating disorder, and the use of medications that might affect body weight (i.e. psychiatric medications), or weight change of > 5% body weight within 3 months prior to enrollment. Three groups of participants were enrolled: (1) individuals willing to participate in LCD (n = 17); (2) patients who were scheduled to undergo SG (n = 13); and (3) patients scheduled to undergo RYGB (n = 14). The choice of surgical procedure was based on patient and surgeon preference and performed as described [7]. Lean individuals (n = 12; BMI 18–24.9 kg/m²) matched for age and sex were included as controls for the groups with obesity prior to intervention.

Study Design

After the initial visit (T1), patients in the LCD cohort had weekly visits with the Registered Dietitian (RD) for approximately 12 weeks (T2), biweekly (every two weeks) visits for approximately 12 weeks (T2), followed by 6 monthly visits for the remainder of the study until approximately 1 year (T3). For the first phase (T1-T2), participants consumed 5–6 shakes/day, totaling 800–960 calories per day (Robard Corporation, Mt Laurel Township, NJ, USA or Optifast, Novartis, Minneapolis, MN, USA). After the T2 visit, participants were given instructions on a partial meal replacement transition diet: 3 meal replacement products, 1 meal (455-560 kcal), 1 snack (160 kcal) for one month; 2 meal replacement products, 2 meals, 1 snack for 2 months. For the remaining time subjects were instructed to consume 1200–1500 calories/day. All meal replacements were provided. Participants had 6 clinical visits to draw venous blood for safety labs during the complete meal replacement diet.

Patients participating in the surgical cohort underwent testing prior to their WL surgery (T1). The surgical groups were monitored until participants had lost a similar percentage of weight expected with the LCD group (T2). Additional follow-up occurred at approximately 1 year (T3) for LCD and surgery groups. At all three visits, a validated visual analog scale (VAS) and food craving inventory (FCI) questionnaire were completed. Weight and height were also measured followed by consumption of a liquid test meal over a 15-minute period (Ensure^® Compact Nutrition Shake, 236 ml, 440 kcal, 60% carbohydrate, 16% protein, 24% fat). Venous blood was collected in the fasted state and at 15, 30, 60, 90 and 120 min after meal consumption, centrifuged at 4 °C and stored as serum or plasma at -80 °C.

Data Collection

Appetite Scores

Participants completed a validated visual analog scale (VAS) questionnaire in the fasted state and at 30, 60, and 120 min after consumption of the test meal at T1, T2, and T3 [20, 21]. The VAS consisted of 100-mm lines with words (“not at all” or “extremely”) anchored at each end describing extreme sensations of hunger, fullness, and prospective eating. Patients were asked to make a vertical mark across the line corresponding to their feelings. Quantification was performed by measuring the distance from the left end of the line to the mark. A validated Food Craving Inventory (FCI) was used to measure general and specific food cravings [22].

GLP-1 Assay

Total GLP-1 levels were measured at T1 and T2 from blood collected prior to and 15, 30, 60, 90 and 120 min after the mixed meal from a subset of subjects: LCD (n = 7); SG (n = 5); RYGB (n = 4). Plasma levels of GLP-1 (7–36) and (9–36) were measured by ELISA (Millipore, Billerica, MA, USA) with an assay sensitivity of 1.5pM.

Statistical Analysis

The power analysis was based on the primary endpoint of the study which was to compare changes in the plasma proteome between the different intervention groups. For the primary endpoint, differences in GLP-1 values obtained from our prior studies were used to determine sample size which was calculated to be 12 subjects in each group to achieve over 95% power to detect expected differences in GLP-1 values. For the secondary endpoints, power was calculated with the actual number of subjects studied. For unpaired t-tests comparing T1 to T2 changes between LCD and surgery with 15 and 24 subjects, respectively, there was 80% power to find a surgery-LCD difference to be significant (p = 0.05) if the difference exceeded one standard deviation. For comparisons of T1 to T3 changes, with 12 and 14 subjects in the two groups, significance could be detected if the difference exceeded 1.1 times the standard deviation. Because this was not a randomized trial, outcomes were analyzed as changes from baseline rather than as absolute values at follow-up time point to minimize any bias from baseline group differences. Except for lean controls, only subjects who had visits for at least T1 and T2 were included in the data analysis. Two participants in the LCD group were excluded from analysis due to lack of adherence to study protocol. Postprandial VAS data was excluded from 3 participants in the RYGB due to meal test not being performed (2) or incorrect completion of questionnaire (1). At T3, there was loss to follow-up in both groups with patients assumed to be missing at random (Fig. 1; LCD n = 3, surgery n = 9). Area under the curve (AUC) was computed using the trapezoidal rule from 0 to 120 min. All statistical analyses were carried out using R (Version 4.3) with the cufunctions package to facilitate statistical analyses in R [23]. The intervention groups at baseline were compared to lean controls by unpaired t-tests. The surgery and LCD groups were compared with respect to changes in appetitive sensations and food cravings over time using a repeated-measures analysis with two fixed effects (time and group) and a random effect (subject) within the time factor, with interaction between group and time to allow for differences in time effects between groups. A correlation analysis was used at each follow-up time-point to examine the association between %TWL and appetite. A multiple linear regression model was employed to examine the association between %TWL and various measures of appetite. The best model was chosen by the modified Akaike Information Criterion. As results were similar in the surgery groups, the data presentation was combined, with any differences in results between SG and RYGB indicated in the text. A p < 0.05 was considered statistically significant and data were reported as means and standard errors unless otherwise noted.

Fig. 1.

Consort diagram. MT = meal test; LTFU = lost to follow-up. Reasons for withdrawal (WD) included relocation to a different state, lack of time due to job responsibilities, unrelated medical conditions, and 1 death

Results

Baseline Characteristics and Weight Loss

Subject enrollment and follow-up are summarized in the consort diagram (Fig. 1). Baseline characteristics for lean controls and subjects with obesity (LCD + surgery) were similar for age and sex but differed for race and ethnicity (Table 1). Baseline characteristics for the LCD and surgery groups were similar for age, sex, race, weight, and BMI (Table 1). The number of days from intervention to T2 were 89.6 ± 2.6 and 62.5 ± 10 for the LCD and surgery groups, respectively (p = 0.15) and from intervention to T3 were 371 ± 5.2 and 421 ± 23 in the LCD and surgery groups, respectively (p = 0.007). The time intervals at T2 were used to align groups for matched weight loss rather than fixed number of days. At T2, there were no differences in %TWL between LCD and surgery group (Table 2; Fig. 2). At T3, there was significantly greater %TWL in the surgery group compared to LCD (p ≤ 0.001).

Table 1.

Baseline characteristics of Lean, LCD, and surgery participants

	Lean (n = 12)	LCD (n = 15)	Surgery (n = 24)	Lean vs. Obese (LCD + Surgery) P-value	LCD vs. Surgery P-value
Age (yr)
Mean	37.2 ± 3.2	44.5 ± 2.8	40.7 ± 2.6	0.21	0.34
Sex
Female– n. (%)	7 (58)	12 (80)	18 (75)	0.27	1.00
Male– n. (%)	5 (42)	3 (20)	6 (25)
Race
White – n. (%)	6 (50)	5 (33)	7.5 (31)	0.01	0.10
Black – n. (%)	1 (8)	10 (67)	16.5 (69)
Ethnicity
Hispanic– n. (%)	1 (8)	3 (20)	15 (62.5)	0.02	0.02
Non-Hispanic– n. (%)	11(92)	12 (80)	9 (37.5)
Weight/BMI
Weight (kg)	67.2 ± 2.5	129.0 ± 5.1	126 ± 3.8	0.0001	0.63
BMI (kg/m2)	22.6 ± 0.5	45.1 ± 1.1	44.2 ± 1.1	0.0001	0.60
Appetite Sensations
Hunger Fasting	41.4 ± 8.3	41.4 ± 7.2	49.4 ± 5.3	0.58	0.40
Hunger AUC	2690 ± 740	2700 ± 560	3770 ± 400	0.37	0.11
Fullness Fasting	29.3 ± 8.7	28.2 ± 5.0	19.1 ± 4.5	0.39	0.27
Fullness AUC	7070 ± 820	6560 ± 680	4960 ± 480	0.09	0.06
Prospective Eating Fasting	48.5 ± 6.4	54.1 ± 5.4	58.8 ± 4.6	0.24	0.56
Prospective Eating AUC	3870 ± 840	4150 ± 820	4530 ± 540	0.59	0.66
Sweet Cravings	16.2 ± 1.0	18.1 ± 1.6	18.5 ± 1.1	0.21	0.82
Carbohydrate Cravings	17.0 ± 1.2	17.3 ± 1.5	21.3 ± 0.9	0.13	0.02
High Fat Cravings	14.6 ± 1.0	14.0 ± 0.9	18.1 ± 1.1	0.23	0.01
Fast-Food Cravings	9.2 ± 0.8	9.7 ± 0.7	11.8 ± 0.6	0.07	0.02
Total Cravings	56.9 ± 3.0	59.1 ± 3.7	69.7 ± 2.8	0.07	0.02

Values are presented as mean ± SEM. P-values reflect group comparisons at baseline. Bold font indicates difference with p ≤ 0.05

Table 2.

Changes over time in visual analog scale scores after LCD and surgery

	LCD	Surgery	LCD P-value	Surgery P-value	LCD vs. Surgery P-value
Total Weight Loss (%)
T2-T1	14.9 ± 0.7	14.6 ± 2.0	-	-	0.94
T3-T1	14.6 ± 2.0	30.2 ± 1.5	-	-	0.0001
T3-T2	-0.7 ± 2.1	19.3 ± 1.8	-	-	0.0001
Hunger Fasting
T2-T1	-7.4 ± 8.4	-0.8 ± 7.7	0.38	0.79	0.61
T3-T1	-14.0 ± 7.7	-3.1 ± 10	0.04	0.49	0.26
T3-T2	-16.5 ± 8.5	2.0 ± 5.4	0.22	0.66	0.52
Hunger AUC
T2-T1	-549 ± 570	-1810 ± 610	0.36	0.001	0.11
T3-T1	-794 ± 530	-328 ± 580	0.24	0.17	0.98
T3-T2	-436 ± 590	843 ± 700	0.75	0.08	0.15
Fasting Fullness
T2-T1	5.2 ± 8.0	3.7 ± 5.5	0.50	0.62	0.84
T3-T1	-1.0 ± 9.3	9.1 ± 7.8	0.92	0.10	0.31
T3-T2	0.5 ± 9.5	14.1 ± 6.4	0.59	0.25	0.24
Fullness AUC
T2-T1	618 ± 590	2740 ± 720	0.37	0.0001	0.02
T3-T1	-361 ± 540	2390 ± 870	0.73	0.001	0.01
T3-T2	-505 ± 580	-218 ± 680	0.23	0.61	0.61
Fasting Prospective Eating
T2-T1	-10.6 ± 7.3	-27.2 ± 5.1	0.13	0.0001	0.07
T3-T1	-10.3 ± 4.6	-18.8 ± 8.5	0.08	0.002	0.49
T3-T2	-6.7 ± 8.9	11.2 ± 5.6	0.73	0.28	0.33
Prospective Eating AUC
T2-T1	-966 ± 540	-2960 ± 560	0.09	0.0001	0.02
T3-T1	-638 ± 480	-1030 ± 720	0.32	0.03	0.49
T3-T2	162 ± 750	1480 ± 360	0.58	0.001	0.12

Values presented are linear mixed model mean ± SEM. P-values represent paired comparisons of change within group between T2 and T1, T3 and T1, and T3 and T2 and between group comparisons of the changes. Bold entries highlight differences where p ≤ 0.05

Fig. 2.

Total weight loss percentage (%TWL) and change in BMI after LCD and surgery. Values presented are linear mixed model mean ± SEM for changes between T2 and T1, and between T3 and T1. ****p < 0.0001

Visual Analog Scale

There was no significant difference between lean controls and participants with obesity at T1 (Table 1). Appetitive changes in the intervention groups over time are reported in Table 2; Figs. 3, and 4. Figures presenting additional appetite-related outcomes, including change in values for fasting and AUC measures of hunger, fullness, and prospective eating and cravings, are included in the supplementary file to enable comparison across groups and timepoints (Figs. S1, S2). VAS ratings for hunger, fullness, and prospective eating were similar in both groups at baseline, although there was a trend towards decreased fullness during the test meal as measured by AUC, in the LCD versus surgery group (p = 0.06). There were no notable differences in changes of fasting hunger and fullness between the LCD and surgery group, however a greater increase in fasting fullness was observed after SG compared to RYGB at both T2 and T3 (p = 0.04 for both). There was a significant decrease between T1 and T2 visits in fasting prospective eating in the surgery group, but only SG showed a sustained decrease from T1 to T3 (p = 0.002). Although the decrease in postprandial hunger from T1 to T2 was quite marked in the surgery group, this change did not reach statistical significance compared with LCD group (p = 0.11). The most marked differences between groups were the changes in postprandial fullness. There were significant increases only in the surgery group at both T2 and T3. There was also a significant decrease in fasting and postprandial prospective eating sensation that occurred only in the surgery group and remained decreased at one year compared with baseline. None of the appetitive sensations at T3 correlated with %TWL (p > 0.20).

Fig. 3.

Visual analog scale results of fasting and AUC appetitive ratings after LCD and surgery. Fasting and AUC values presented are linear mixed model mean ± SEM at T1, T2, and T3. P-values reflect the comparison between groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Fig. 4.

Fasting and postprandial VAS scores. Subjective appetite response to test meal pre and post weight loss in participants undergoing LCD or surgery. Curves for hunger (A), fullness (B), and prospective eating (C) are shown at 0, 30, 60, and 120 min following consumption of the test meal

Food Craving Inventory

Baseline cravings are presented in Table 1 which shows that sweet cravings were similar between lean controls and the groups with obesity, and between LCD and surgery groups, although there was a trend towards greater fast-food cravings in the group with obesity (p = 0.07). Cravings for carbohydrates, high fats, and fast-food fats were significantly higher in the surgery group than LCD at baseline (Table 1). Changes in cravings are presented in Table 3; Fig. 5. With the exception of high fat cravings in the LCD group, there were decreases in all cravings for both groups between T1 and T2. The decreases in carbohydrate, high fat, and fast-food cravings were greater after surgery and were maintained at T3 in all subscales except sweet cravings. There were no significant correlations between %TWL and cravings at T3 although high fat cravings at T3 tended to correlate with %TWL (r = 0.527; p = 0.06).

Table 3.

Changes over time in food cravings after LCD and surgery

	LCD	Surgery	LCD P-value	Surgery P-value	LCD vs. Surgery P-value
Sweet Cravings
T2-T1	-5.3 ± 1.4	-7.8 ± 1.3	0.001	0.0001	0.19
T3-T1	-1.9 ± 1.6	-1.4 ± 1.6	0.43	0.25	0.88
T3-T2	4.0 ± 1.6	5.3 ± 0.9	0.01	0.0001	0.33
Carbohydrate Cravings
T2-T1	-3.7 ± 1.6	-7.3 ± 1.7	0.02	0.0001	0.07
T3-T1	-2.6 ± 1.4	-5.9 ± 1.7	0.52	0.0001	0.01
T3-T2	3.1 ± 1.9	0.3 ± 0.8	0.15	0.83	0.35
High Fat Cravings
T2-T1	-1.6 ± 0.9	-5.2 ± 1.4	0.20	0.0001	0.03
T3-T1	-2.0 ± 1.3	-1.1 ± 1.3	0.34	0.05	0.58
T3-T2	-0.2 ± 1.1	2.2 ± 1.1	0.86	0.03	0.19
Fast-Food Fat Cravings
T2-T1	-2.1 ± 0.9	-4.4 ± 0.8	0.01	0.0001	0.04
T3-T1	-1.7 ± 0.7	-2.2 ± 0.8	0.14	0.001	0.24
T3-T2	0.4 ± 1.0	1.3 ± 0.7	0.46	0.07	0.50
Total Cravings
T2-T1	-12.7 ± 3.9	-24.6 ± 4.5	0.003	0.0001	0.03
T3-T1	-8.2 ± 3.7	-10.6 ± 4.1	0.32	0.001	0.13
T3-T2	7.3 ± 4.8	9.0 ± 2.5	0.08	0.01	0.68

Values presented are linear mixed model mean ± SEM. P-values represent paired comparisons of change within group between T2 and T1, T3 and T1, and T3 and T2 and between group comparisons of the changes. Bold entries highlight differences where p ≤ 0.05

Fig. 5.

Food craving inventory results of fasting cravings after LCD and surgery. Values presented are linear mixed model mean ± SEM between T1, T2, and T3. P-values reflect the comparison between groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Plasma GLP-1 Levels

Fasting plasma levels of GLP-1were similar between LCD and surgery groups at baseline (p = 0.27) and decreased somewhat from T1 to T2 in the surgery group (Fig. 6). In contrast, there was almost a two-fold increase from T1 to T2 in postprandial levels of GLP-1 in the surgery groups (p < 0.0001), whereas no change was noted after LCD (p = 0.90; Fig. 6). Interestingly, the increase in GLP-1 AUC was positively correlated with the increase in fullness AUC in the surgery group (r = 0.69; p = 0.038) but not in the LCD group (p = 0.34). This correlation was not significant in either group for change in hunger or prospective eating AUC values.

Fig. 6.

Fasting and postprandial GLP-1 levels after LCD and surgery. Fasting and AUC values presented are linear mixed model mean ± SEM at T1 and T2. P-values reflect the comparison between groups. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001

Discussion

The main objective of this study was to compare the effects of LCD versus MBS on indices of appetite and food cravings. Significant differences in the effects of these two interventions on appetitive sensations were demonstrated even when %TWL was matched between the groups. Most strikingly, with matched %TWL, there was a significant increase in postprandial fullness and decrease in postprandial prospective food consumption – these changes did not occur after LCD. As expected, subjects in the surgery group continued to lose weight, and despite two-fold greater mean %TWL compared with the LCD group, surgery subjects continued to report significantly greater fullness and reduced prospective food consumption. Despite some previous findings stating that caloric restriction increases hunger [9, 14, 24–26], there was no increase in fasting or postprandial hunger in the LCD group which is a finding also reported by Aukan et al. [10]. In contrast, there was a marked decrease in postprandial hunger, specific to the surgery group, although this effect was not maintained over the long-term.

An overarching question is why MBS results in greater and more sustained weight loss compared to calorie restriction alone, which is almost universally associated with weight regain [27]. There was no increase in hunger in the LCD group in this study, and the decrease in hunger in the surgery group was not maintained, suggesting that the sensation of hunger is not a major contributor to the different outcomes. In contrast, marked increases were observed in postprandial levels of fullness only after surgery. We have previously shown that SG and RYGB are associated with significant postprandial increases in the gastrointestinal satiety hormones PYY and GLP-1 that are known to elicit sensations of fullness after continuous infusion administration [7, 28–30]. Results of a subset of participants in this study also show a substantial increase in postprandial GLP-1 levels, whereas no change was observed after equivalent weight loss after LCD. Increases in satiety hormones are also minimal or absent after diet or laparoscopic gastric banding, a less successful procedure for weight loss which results in restriction to nutrient flow, but does not alter the rate or pathway of flow and has largely been abandoned [6, 14]. Thus, the anatomical changes that occur with SG and RYGB, resulting in increased anorexigenic signaling, are consistent with the sensation of increased fullness and likely allow patients to experience satiation with smaller caloric loads. In addition, the decrease in prospective food consumption that occurred only after MBS, in both the fasting and postprandial states, and was sustained at 1 year, may also contribute to greater %TWL, as this variable correlates with the amount of food intake at a meal [10, 31]. The use of GLP-1 receptor agonists is now the most effective medical treatment for obesity. Interestingly, studies have shown that use of these medications increases fullness and decreases prospective eating and hunger [32, 33], changes that are mimicked in our results. Some studies have shown that postprandial GLP-1 levels are associated with greater weight loss after MBS [34]. Indeed, data from this study show that change in postprandial GLP-1 levels after surgery correlates positively with change in fullness, although it cannot be determined whether this correlation indicates association and/or causation.

In concurrence with previous findings, total cravings decreased after MBS [7, 35]. All craving subsets were decreased in the short and long-term, with the exception of sweet cravings, which returned towards baseline at one year. All craving subscales other than sweet were also decreased in the LCD group, however, the magnitude of the changes were less marked compared with the surgery group and were not sustained in the long-term. High fat cravings may be of particular interest after surgery as these tended to correlate with weight loss. Of note, most craving subsets were greater at baseline in the surgery compared with LCD group. This difference is likely due to the relatively small sample size and non-randomized study design. Another consideration, however, is that the FCI was validated in a mostly non-Hispanic population. The LCD group consisted of a greater percentage of Hispanic participants (80% vs. 37.5%; p = 0.02) which may have influenced the scores.

Major strengths of this study include the prospective design in which participants were studied after equivalent mean %TWL in the short term over a similar period of time and after approximately year. A major limitation is that participants were not randomly assigned to interventions which could have resulted in selection bias and inherent group differences, such as greater cravings at baseline in the surgery vs. LCD group and differences in ethnicity. Such limitations were addressed by considering baseline variables in the statistical analysis and quantifying change in values however the effect of these differences may not be fully accounted for by statistical means. The sample size is also relatively small, which may limit the generalizability of the findings. Other factors that contribute to WL and maintenance such as physical activity, emotional and behavioral characteristics, or socioeconomic status were not addressed. A solid meal stimulus or a different macronutrient composition, administered at different times of the day or at different rates, could have elicited different effects on appetite and additional validated instruments may have addressed other aspects of eating behavior such as restraint, disinhibition, and hedonic hunger. Although the interval to T2 was shorter in the surgery group, this difference did not reach statistical significance (p = 0.15) and is unlikely to have contributed to greater satiety at that timepoint, particularly given that the difference in fullness between groups persisted through T3. Finally, longer-term outcomes would be needed to address nadir weight loss and weight regain. Although baseline measures were not predictive of %TWL outcomes, these findings, brought about by both anatomical and physiological alterations from surgery, might explain the greater %TWL as compared to those who lose weight with caloric restriction alone [7, 36–38].

Conclusion

Results of this study demonstrate differences between the effects of MBS versus LCD on appetitive sensations and GLP-1 levels, providing insight into the potential mediators that result in greater weight loss after surgery. In particular, MBS results in greater postprandial fullness and decrease in prospective food consumption ratings, both of which may contribute to the ability to sustain lower caloric intake over the long-term.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

L.T. and J.K. wrote the main manuscript text, prepared the figures, and performed data analysis; R.S. provided statistical support; S.B. coordinated and performed all study visits; J.L. managed participants on the low calorie diet; T.R., A.K., M.B., J.K. provided clinical care; J.K. conceived the study and obtained research funding; all authors reviewed the manuscript.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Humans Ethics and Consent to Participate

All procedures performed in this study involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all individual participants included in the study.

Competing Interests

A.K. served as Consultant for The Network Effects. J.K. has served on the Advisory Boards of Found Health, Gila Therapeutics and Morphic Medical; has stock ownership in Found Health; has served on the Board of Directors of the American Board of Obesity Medicine; has served as Consultant for the Federal Trade Commission, Apotex, Equinox, Canaccord Genuity, Sago, Bristol Meyers Squibb; has received roylaties for UptoDate.