Study design
The hypothesis for this paper was tested as a secondary hypothesis within a wider study investigating changes in energy intake following GBP, which is described in detail elsewhere (14, 15). Owing to the novel study protocol, the sample size was estimated from the patient population recruited for a randomised controlled trial (16) that detected significant differences in self-reported energy intake between Vertical Banded Gastrectomy (VBG (n = 7) and GBP (n = 9) participants at 6 years post-surgery. The SD associated with the change in dietary fat intake (% energy) from pre- to post-surgery and a 95% confidence interval was applied as follows:
$$n= {\left(\frac{confidence level x SD}{margin of error}\right)}^{2}$$
$$n= {\left(\frac{1.96 x 1.9}{1}\right)}^{2}$$
It was estimated that at least 16 participants are required for the present study based on a 14% attrition rate reported by another similar intake study (17). However, as the proposed study protocol was intensive for study participants, 32 GBP patients and a similar number of weight-stable comparators were recruited to account for a potentially higher attrition rate.
In brief, in this study, all participants were required to complete four fully residential study assessments at 1-month pre-surgery, 3, 12- and 24-months post-surgery at the Human Intervention Studies Unit (HISU) at Ulster University. HISU consists of en-suite bedrooms, a communal sitting room, a metabolic kitchen (closed access to participants) and communal dining room. Participants arrived at HISU at approximately 6pm on day 1 for an initial acclimation period where no measurements were performed. Following a standard meal of Spaghetti Bolognese for dinner (if requested), participants fasted from 10pm. Measurements started early on day 2 (approx. 7am) and lasted until bedtime (approx. 11pm) on day 2. Participants remained sedentary throughout but were free to engage in light activities such as reading, crafts and watching television.
Ethical approval
was granted by the West of Scotland Research Ethics Service (WoSRES) (REC 16/WS/0056, IRAS 200567). The study was registered as a clinical trial (NCT03113305) (clinicaltrials.gov) and was conducted according to the principles of the Helsinki declaration. The primary outcome of the work was changes in dietary energy intake, with the work presented in this study included as secondary outcomes. Prior to the start of the study written informed consent was obtained from all participants.
Patients were recruited from four sites in the United Kingdom (Phoenix Health NHS, Phoenix Health Private, London Imperial Weight Centre and North Bristol NHS Trust) and one site in the Republic of Ireland (Letterkenny University Hospital). Inclusion criterion were ≥ 18 years old with a scheduled GBP. Weight-stable (> 6 months) comparators were time-matched and recruited by posters, email circulations, radio, and social media platforms. The purpose of the comparator group was to account for external factors which could potentially impact GBP patients over the study period, as well as any change in behaviour in the residential unit over the four time points. Inclusion criterion were ≥ 18 years old with no plans to alter current body weight. Exclusion criteria for all participants were: presence of physical or psychological conditions affecting food intake; strict dietary restrictions, food allergies and pregnant or lactating women.
Total body weight, body mass index (BMI), FM, FFM, visceral adipose tissue (VAT) and subcutaneous adipose tissue (SAT) were assessed under standardised conditions using the total body GE Lunar iDXA scan (GE Healthcare, USA). Height was measured during the initial visit to the nearest 0.1cm using a wall-mounted stadiometer (Seca Ltd, Hamburg, Germany (%CV = 0.23%). If the participant’s body width exceeded the standard dimension of the DXA’s scanning area, they were positioned such that the right half of the body was fully within the scan field. Half scans have shown satisfactory validity (18). DXA measurements were conducted by trained researchers and verified by a qualified health care professional.
Percentage total weight loss (%TWL) was calculated as \(\%TWL=\frac{\left(weight prior to surgery - follow-up weight\right)}{weight prior to surgery} x 100 .\)Postoperative weight loss was also expressed as a percentage excess of weight loss (%EWL) following the formula:\(EWL=\frac{(\text{w}\text{e}\text{i}\text{g}\text{h}\text{t} \text{p}\text{r}\text{i}\text{o}\text{r} \text{t}\text{o} \text{s}\text{u}\text{r}\text{g}\text{e}\text{r}\text{y} - \text{f}\text{o}\text{l}\text{l}\text{o}\text{w}-\text{u}\text{p} \text{w}\text{e}\text{i}\text{g}\text{h}\text{t})}{(\text{w}\text{e}\text{i}\text{g}\text{h}\text{t} \text{p}\text{r}\text{i}\text{o}\text{r} \text{t}\text{o} \text{s}\text{u}\text{r}\text{g}\text{e}\text{r}\text{y} - \text{w}\text{e}\text{i}\text{g}\text{h}\text{t} \text{c}\text{o}\text{r}\text{r}\text{e}\text{s}\text{p}\text{o}\text{n}\text{d}\text{i}\text{n}\text{g} \text{t}\text{o} \text{B}\text{M}\text{I} = 25 \text{k}\text{g}/{\text{m}}^{2})} x 100\)
BMR was measured under standardised conditions following an overnight fast (from 10pm) using open-circuit portable indirect calorimetry (ECAL, Metabolic Health Solutions) by a trained researcher. Each participant was awakened at approximately 7am in the morning to empty their bladder and return to rest for at least 30 minutes in a quiet, darkened and thermoneutral room before the measurement was made. Distractions such as use of mobile phones were not permitted. Data were recorded for a minimum of 8-minutes and was terminated after readings had been stable for 45-seconds. The first 2-minutes of the measurement period were automatically discarded by the ECAL software, with any other anomalous recordings (e.g., coughing, removal of mouthpiece) also discarded as ‘false’ readings. BMR values were calculated using the Weir formula (19).
To determine the magnitude of metabolic adaptation following GBP, this study used the gold standard methodology (6, 7, 10, 11, 20, 21). The baseline BMR (dependent variable) for both patient and comparator groups was used to generate a linear regression model with multiple predictor variables (independent variables) that may affect BMR values - baseline FM, FFM, age, gender, medications, group (participants) and medical conditions. This model was used to predict the BMR (pBMR) at 3-,12- and 24-months post-surgery.
$$\text{p}\text{B}\text{M}\text{R} \left(\text{M}\text{J}/\text{d}\text{a}\text{y}\right)= 3.529-\left(1.509 \text{x} \text{P}\text{a}\text{r}\text{t}\text{i}\text{c}\text{i}\text{p}\text{a}\text{n}\text{t}\text{s}\right)+\left(0.511 \text{x} \text{G}\text{e}\text{n}\text{d}\text{e}\text{r}\right) - \left(0.001 \text{x} \text{A}\text{g}\text{e} \text{i}\text{n} \text{y}\text{e}\text{a}\text{r}\text{s}\right) + \left(0.022 \text{x} \text{F}\text{M} \text{i}\text{n} \text{k}\text{g}\right) + \left(0.088 \text{x} \text{F}\text{F}\text{M} \text{i}\text{n} \text{k}\text{g}\right) + \left(0.936 \text{x} \text{M}\text{e}\text{d}\text{i}\text{c}\text{a}\text{t}\text{i}\text{o}\text{n} \text{t}\text{h}\text{a}\text{t} \text{a}\text{f}\text{f}\text{e}\text{c}\text{t} \text{B}\text{M}\text{R}\right) - \left(0.513 \text{x} \text{D}\text{i}\text{s}\text{e}\text{a}\text{s}\text{e} \text{t}\text{h}\text{a}\text{t} \text{a}\text{f}\text{f}\text{e}\text{c}\text{t} \text{B}\text{M}\text{R}\right)$$
Participants (1 for Patient, 2 for Comparator), Gender (1 Female, 2 Male), Medications that affect BMR (1 Prescribed, 2 Not prescribed), Diseases that affect BMR (1 Present, 2 Absent).
Finally, the residual BMR (resBMR) is defined as the difference between the observed BMR (as measured by indirect calorimetry) from the predicted BMR based on the above linear regression equation.
$$BMR residual =(\text{m}\text{e}\text{a}\text{s}\text{u}\text{r}\text{e}\text{d} \text{B}\text{M}\text{R} - \text{p}\text{r}\text{e}\text{d}\text{i}\text{c}\text{t}\text{e}\text{d} \text{B}\text{M}\text{R})$$
And the presence of metabolic adaptation is defined as resBMR being significantly different from zero.
Statistical analysis
Statistical analyses were performed using IBM SPSS for windows (UK, version 26.0) and R (version 4.2). Baseline summary statistics are expressed as mean (SD) for continuous variables, or as numbers (percentage) for categorical variables. Results from linear mixed models were presented as least squares mean (SEM).
At each time-point, there were some random missing values due to missed appointments (Fig. 1) and, in a few cases, technical issues with measuring equipment. Given that it is reasonable to assume that such values were missing purely at random, mixed effects linear models were fitted for the main outcome measures of interests (Weight, BMI, FM, FFM, LBM, VAT, SAT, and BMR). In each of these linear mixed models, participant IDs were fitted as random effects, with participant group (patient or comparator), time and the interaction between group and time as fixed effects. For such linear mixed modelling, no imputation of missing values was conducted as this was unnecessary. From the fitted linear mixed models, the estimated means and standard errors of the outcome measure were then obtained for all group and time point combinations. Where applicable and deemed interesting, comparative analysis between different time points per group, or between the two groups per time point were conducted by testing the corresponding general linear hypothesis.
Mixed model analysis was applied to the residual BMR to determine the presence or absence of metabolic adaptation. Metabolic adaptation was considered to have occurred if BMR residual (magnitude of metabolic adaptation) was significantly different from zero (p ≤ 0.05).
Pearson correlation coefficients were used to study associations between changes in FM, FFM, %FFM/ weight and BMR in patients. P-values of ≤ 0.05 were considered as statistically significant.