Liraglutide 3.0 mg in the treatment of adults with obesity and prediabetes using real‐world UK data: A clinical evaluation of a multi‐ethnic population

UK guidelines recommend liraglutide 3.0 mg in adults treated within specialist weight management services with BMI ≥35 kg/m2, prediabetes and high cardiovascular disease risk. We aimed to clinically evaluate liraglutide 3.0 mg in specialist weight management services. We evaluated liraglutide 3.0 mg in weight management services at Guys and St Thomas' NHS Foundation Trust. Objective body weight (BW) was measured at baseline and 4 months, allowing classification as ‘responders’ (≥5% BW reduction) and ‘non‐responders’ (<5% BW reduction). One hundred and twenty‐one patients were evaluated. At 4 months, 76.0% attended follow‐up (82.6% responders, 17.4% non‐responders); BW (−8.6 kg, 95%CI:‐9.8, −7.4 kg), BMI (−3.2 kg/m2, 95%CI: −3.6, −2.8) and %‐BW (−6.6%, IQR: −8.8%, −5.2%) significantly reduced. In responders, HbA1c reduced by −5.0 mmol/mol (IQR: −7.0. −4.0 mmol/mol). In responders BW continued to reduce up to 12 months (4 m: −10.2 kg, p < .0001; 6 m: −15.6 kg, p < .0001; 9 m: −16.5 kg, p < .0001; 12 m: −16.7 kg, p < .01). Those of Black African and Caribbean ethnicity experienced less BW loss than those of white ethnicity (4.12 kg, p = .017) and had a greater attrition rate. In adults with obesity and prediabetes who are treated within specialist weight management services, liraglutide 3.0 mg reduces BW and HbA1c. Those of Black African and Caribbean ethnicity experienced less BW reduction and greater attrition at 4 months. Further evaluation of the ethnic differences in response to obesity pharmacotherapy is required.

• There is limited real-world data evaluating the efficacy of liraglutide 3.0 mg in those with obesity and prediabetes of various ethnicities.

What this study adds?
• In a publicly funded real-world evaluation of an obesity pharmacotherapy programme, liraglutide 3.0 mg in combination with lifestyle intervention significantly reduced body weight and improved glycaemic control in those with obesity and prediabetes.
• Those of Black African and Caribbean ethnicity experienced less body weight reduction and had a higher attrition rate than those of white ethnicity.This was not underpinned by differences in age, sex, BMI or socioeconomic status between ethnic groups.
• Further evaluation is required to determine whether ethnicity-specific differences in response to GLP-1 receptor agonists persist in a larger population.

| INTRODUCTION
Obesity prevalence is increasing globally. 13][4] Bariatric surgery is the most effective intervention for obesity; however, access is limited.Pharmacological interventions targeting GLP-1 have considerable scope to address this unmet clinical need.Clinical trial data delineates in patients with prediabetes and obesity that liraglutide 3.0 mg in combination with lifestyle intervention leads to significant body weight (BW) reduction. 36][7][8] In the SCALE Obesity and Prediabetes trial, liraglutide 3.0 mg reduced BW by À5.6 kg compared to placebo. 3 Meta-analysis shows that 65.3% of patients reduce weight by >5%. 9,10However, most participants in previous clinical trials are of white ethnicity; some research points towards those of Black African ethnicity experiencing less weight reduction. 11,12Clinical trials also do not commonly report participant socioeconomic status, and it is unclear whether efficacy translates to individuals of diverse socioeconomic backgrounds. 13For instance, lower socioeconomic status is associated with reduced BW loss 10 years following metabolic surgery. 14Overall, real-world clinical evaluation is required to determine how individuals of Black African and Caribbean ethnicity and from diverse socioeconomic statuses respond to obesity pharmacotherapy.
In the UK, the National Institute of Clinical Excellence (NICE) Technical Appraisal 664 recommends liraglutide 3.0 mg in adults with BMI ≥35 kg/m 2 who are with prediabetes, high cardiovascular disease risk and treated within specialist weight management services. 15ere are limited real-world data evaluating liraglutide 3.0 mg in this context.The Tier 3 South-East London Healthy Weight Programme (SELHWP) and Tier 4 Guys and St Thomas' NHS Foundation Trust Bariatric Service provide a unique perspective; these services treat a diverse population of various ethnicities and socioeconomic statuses. 16Therefore, our clinical evaluation aimed to determine the effectiveness of liraglutide 3.0 mg on weight reduction in adults with obesity, prediabetes, and high cardiovascular disease risk.We sought to further delineate changes in HbA1c, glycaemic status and how various demographic factors interacted with weight loss outcomes.

| METHODS
We analysed patients on an obesity pharmacotherapy pathway over 22 months (1 July 2021-1 May 2023) from South East London.This was a prospective analysis of all patients referred to the pathway.This project was approved by the Guys and St Thomas' NHS Foundation Trust audit committee (Project Number: 14214).

| Eligibility for pathway
We included adults with BMI ≥35 kg/m 2 , prediabetes and high CVD risk, who were adhering to a tier 3 or 4 weight management lifestyle intervention.Tier 3 services are specialist weight management clinics providing non-surgical intensive medical and lifestyle intervention within a multidisciplinary team (MDT).Tier 4 services provide surgical and non-surgical interventions supported by an MDT.Full details of eligibility criteria are in supplementary Table S1.

| Patient pathway
Supplementary Table S2 shows the obesity pharmacotherapy patient pathway.Eligibility for liraglutide 3.0 mg was based on NICE criteria.
Patients commenced liraglutide 3.0 mg once they had started a weight management lifestyle intervention via our services.The clinic involved an initial group onboarding session with the obesity physician.Liraglutide was titrated weekly from 0.6 mg to 3.0 mg in 0.6 mg increments as per its licence. 17When patients experienced side effects, a more gradual titration was implemented at the physician's discretion.For the initial 4-month treatment period, a dietitian telephone follow-up was booked at 2 months and a face-to-face follow-up with the obesity physician at 4 months.At 4 months follow-up with the obesity physician, objective BW was measured.Continuation was determined based on the 5% weight loss threshold: categorising into responders (≥5% BW loss) and non-responders (<5% BW loss).Non-responders were discharged from the obesity pharmacotherapy pathway; responders were continued for 2 years total treatment.Further follow-up for responders included weight management dietitian telephone appointments at 6 and 9 months, alongside obesity physician follow-up in person at 12 months.Liraglutide was stopped preoperatively in tier 4 patients who subsequently underwent bariatric surgery.

| Lifestyle intervention
Patients participated in a tier 3 or tier 4 weight management programme.Tier 4 weight management patients had been referred for bariatric surgery, endoscopic procedures or pre-optimisation dietitian input.Tier 3 weight management patients were reviewed in the South-East London Healthy Weight Programme (SELHWP).
The SELHWP features an MDT, including specialist dietitians, physiotherapists, psychologists, dietitian assistants and obesity physicians.
Patients were offered the choice between traditional diet and lifestyle advice (Balance) or a total meal replacement programme (FAST).
Patients had the choice between both programmes, but validated psychology and eating behaviour questionnaires were used to screen patients for suitability to participate in FAST.Those with disordered eating were excluded from the FAST programme.
Balance utilised a moderate and evidence-based calorie reduction at 500-600 kcal less than the calculated basal metabolic rate. 18These calculations were rounded to 1500 kcal or 1800 kcal/day; patients were provided with adaptable meal plans based on these calorie targets.These adaptable meal plans are underpinned by government 'eat well guide' portion and proportion guidance 19 and suggest moderately reduced starchy carbohydrate serving sizes.This meal plan specifically benefits those living with insulin resistance, PCOS, T2DM and those with reduced carbohydrate requirements, i.e., low physical activity levels. 20ST is a low-calorie diet programme that utilises nutritionally complete meal replacement products (Total Meal Replacement) for the first 12 weeks of intervention. 21,22This is followed by 6 weeks of moderated food re-introduction, known as partial meal replacement (PMR).Patients were advised to purchase products from Lighter Life, Exante or Tesco Ultra-Slim.
Balance and FAST programmes run for 12 months.They have front-loaded clinical contact, meaning sessions fall fortnightly in the first 6 months and move to monthly contact in the second 6 months.
Both programmes include 8 h of physiotherapy-led exercises, a smartphone application, and opt-in psychology support.

| Outcome assessment and stopping criteria
The primary outcome was objective change in BW at 4 months.Secondary outcomes included body weight change at other time points (2, 6, 9, 12 months), HbA1c and side effects.BW was objectively measured at baseline and 4 months using the same weighing scales for all patients and was gathered by self-report at 2, 6, 9 and 12 months.
HbA1c data was gathered at baseline and follow-up via primary care records.Sensitivity analysis was also performed to examine the impact of ethnicity, sex, lifestyle intervention group and socioeconomic status on outcomes.The Multiple Deprivation Index categorised deprivation; a score of 1 represents the most deprived decile, and 10 represents the least deprived decile.This score encompasses 7 domains: income, employment, education, health, crime, barriers to housing and services and the living environment. 23

| Discontinuation criteria
As per previously published literature, a 5% weight loss threshold was set to determine responders and non-responders to liraglutide following 3 months of treatment on full dose. 3,24This was set at 4 months, as the dose-escalation phase lasts 1 month.Patients were classed as responders if they achieved clinically meaningful weight reduction at 4 months (≥5%).Patients were classed as non-responders if they did not experience clinically meaningful weight reduction (<5%) at 4 months.Patients were also stopped if they experienced significant side effects which were not amendable to dose changes or if they were discharged from weight management services. 24

| Side effects
Side effect data was collected for all patients from physician and dietitian consultations documentation.Serious side effects were classified as per serious adverse event criteria, including any medical event resulting in hospitalisation, death, significant disability, congenital anomaly or birth defect.

| Statistics
All statistics were performed using Microsoft Excel and GraphPad Prism (GraphPad Prism Version 10.0.0,GraphPad Prism Software, Boston, USA).Statistical significance was defined at p-value <.05.
Baseline demographics were summarised as mean ± SD for parametric data and median with interquartile range (IQR) for non-parametric data.To assess differences in Body Weight, BMI, percentage BW change and HbA1c, pre and post-intervention data was analysed using appropriate tests.Body Weight and BMI followed a normal distribution so were analysed by parametric tests (paired t-test, unpaired t-test).HbA1c and % body weight change did not follow a normal distribution, so were analysed by non-parametric tests (Wilcoxon signed rank test, Mann Whitney U-test).One way Analysis of Variance (ANOVA) was performed as a sensitivity analysis to determine the effect of ethnicity, sex and lifestyle intervention group on 4-month weight loss outcomes.Missing Body Weight and HbA1c data were analysed by the last value carried forward. 25For HbA1c analysis, the non-responders included all patients who either discontinued liraglutide due to side effects, were discharged due to non-attendance or classed as non-responders at 4 months.2).
After adjustment for missing data using the last value carried forward method, responders significantly reduced weight at 2, 6, 9 and 12 months (Supplementary Table S3).Similarly, after adjustment for missing data using the baseline value carried forward method, body weight significantly reduced at (Table 2, Figure 2).BMI significantly reduced at all time points for responders (Table 2, Figures 2 and 3).
Compared to the total cohort, responders tended to be of white ethnicity (50%) and in the Balance group (65.8%) (Supplementary  S5).Those discontinuing liraglutide due to side effects were all females, tended to be younger (43 ± 13.5 years), mostly of white ethnicity (72.7%), and more commonly in the FAST (45.4%) or tier 4 (18.2%)intervention groups.
(Supplementary Table S5) Outcome groups had similar median multiple deprivation index and mean baseline BMI and BW (Supplementary Table S5).S6).missing data, the difference in HbA1c between groups remained significant ( p < .001)(Supplementary Table S7).All patients had prediabetes at baseline.For responders, at follow-up, 97.4% were with normoglycaemia, 2.6% were with prediabetes, and none were with T2D.For non-responders, at follow-up, 38.1% were with normoglycaemia, 47.6% were with prediabetes, and 14.3% were with T2D
(Supplementary Table S8.1).One serious side effect occurred during the programme, involving an episode of pseudo-bowel obstruction requiring hospitalisation.Other side effects included two cases of mild lower gastrointestinal bleeding, two cases of severe nausea and vomiting, one episode of possible cholecystitis and one case of gallbladderlike abdominal pain (Supplementary Table S8.2).One case of severe nausea and vomiting was on a background of severe heart failure; a slower titration of medication was trialled (0.6 mg every 2 weeks to max 3 mg); however, this did not improve the tolerability of liraglutide (Supplementary Table S8.2).One case of severe nausea and vomiting occurred during total meal replacement (FAST programme) (Supplementary Table S8.2)

| Sensitivity analysis
We conducted a sensitivity analysis to assess how ethnicity, sex, and lifestyle intervention group impacted BW outcomes at 4 months.An ANOVA comparing weight loss across ethnic groups was significant ( p = .006)(Supplementary Table S9.1).Sub-group analysis showed this was driven by those of white ethnicity reducing weight by À4.1 kg more than those of Black African and Caribbean ethnicity ( p = .017)(Supplementary Table S9.2).There was a near-significant trend towards those of white ethnicity experiencing greater BW reduction than those of Asian ethnicity ( p = .055)(Supplementary Table S9.2).Those of white ethnicity and Black African and Caribbean ethnicity had similar ages, baseline HbA1c, multiple deprivation index; however, there were more females and patients undergoing tier 4 intervention in the Black African and Caribbean group (Table 1).
Those of Black African and Caribbean ethnicity were proportionally more likely to be non-responders and not to attend follow-up (Table 1).Those of white ethnicity were more likely to be responders and to discontinue liraglutide due to side effects (Table 1).There were no significant between-group differences when analysing BW reduction by sex or lifestyle intervention group (Supplementary Tables S10   and S11).We report a clinical evaluation assessing liraglutide 3.0 mg in adults with obesity, prediabetes and high cardiovascular risk from a multiethnic population treated within weight management services in South East London.At 4 months follow-up, 82.6% were liraglutide responders (BW loss ≥5%); responders were generally of white ethnicity and in the Balance lifestyle intervention group.We show that overall body weight, %BW and BMI reduced significantly at 4 months, with responders experiencing further reductions up to 1 year.
Responders also significantly improved their HbA1c, with nearly all patients reversing from prediabetes to normoglycaemia.Those of Black African and Caribbean ethnicity experienced less weight reduction, were more likely to be non-responders and were less likely to persist with the programme.Combining total-meal replacement with liraglutide was associated with a higher non-responder rate and greater discontinuation due to side effects.Further evaluation is required with a larger sample size to determine whether ethnicityspecific differences with liraglutide persist at 1 year and beyond.
[7][8] Compared to other anti-obesity medications, liraglutide benefits glycaemic control, is useful in those with T2D or cardiovascular disease and has few contraindications. 26In 3-year data from the SCALE prediabetes and obesity trial, only 2% of participants treated with liraglutide developed T2D. 277][8]28 Meta-analysis delineates that >65% of patients achieve clinically meaningful BW reduction (≥5% BW loss). 9,10However, most participants included in the phase 3 evaluations and meta-analysis are of white ethnicity. 10Further mechanistic trials have investigated how liraglutide impacts weight reduction; this is via a reduction in energy intake and gut physiology changes.[31] When compared to previous data, those patients treated within our tier It is yet to report and is assessing 2 years of liraglutide 3.0 mg in patients with obesity and related complications. 381][42][43][44] Further evaluation of the long-term cardiometabolic health benefit associated with GLP-1 agonist therapy is warranted in patients with obesity and prediabetes.

| Ethnicity, demographic and clinical considerations
Our service evaluation shows that those of Black African and Caribbean ethnicity may experience less BW reduction and are less likely to persist with a state-funded obesity pharmacotherapy programme.This is important, as patients of Black African and Caribbean ethnicity have an increased risk of T2D compared to those of white ethnicity and equivalent BMI. 45Previous studies have assessed the efficacy of weight management interventions in diverse ethnicities.In the Look AHEAD study, which evaluated intensive lifestyle intervention, at 1 year, those of African American ethnicity experienced less BW reduction than those of white ethnicity.However, between-group differences did not persist at 4 and 8-year follow-up. 46In a post-hoc analysis of the SCALE programme, 1 year BW reduction was similar in Black and African-American ethnic groups when compared to those of white ethnicity; however, initial BW loss tended to be slower in those of Black and African-American ethnicity.Of note, these data represent African American and Black African individuals, whereas our cohort encompasses those of Black African and Caribbean ethnicity.Importantly, our population and previous data's populations are diasporic and do not demonstrate efficacy in an indigenous Black African and Caribbean community. 47Further work should assess, with a larger sample, the differential response to liraglutide by ethnic groups.Perhaps initial weight loss is less in these cohorts, but long-term BW loss is similar. 46Stopping criteria may require modification by ethnicity as if BW loss in those of Black African and Caribbean ethnicity is delayed, then a longer time may be required to determine medication responsiveness.
In line with previous data, we report that gastrointestinal side effects are common, with over one-third of patients experiencing nausea. 3,7,27For serious side effects, we report an episode of bowel pseudo-obstruction secondary to GLP-1 related constipation.Furthermore, those discontinuing due to side effects were more likely to be in the FAST (total-meal replacement) group; TMR and concomitant GLP-1 agonist therapy require further study.This could focus on whether the initiation of obesity pharmacotherapy should be during TMR, food reintroduction or in the weight maintenance phase.
Responder and non-responder criteria may need to be modified in this circumstance, as the GLP-1 agonist therapy may be more feasible for weight maintenance rather than further weight reduction.

| Strengths and limitations
Our clinical evaluation has several strengths.Firstly, when compared to other clinical evaluations, we included a well-characterised population who were all with BMI ≥35 kg/m 2 , prediabetes and high CVD risk.Secondly, we included a diverse population in terms of socioeconomic status and ethnicity, including a significant proportion of Black African and Caribbean ethnicity individuals.This is important as clinical trial data includes primarily those of white ethnicity. 48Third, our programme was entirely state-funded via the NHS, which may have increased the uptake of liraglutide compared to other healthcare systems.In fact, the median multiple deprivation index was 3, indicating our population was from relatively high levels of socioeconomic deprivation.
Our clinical evaluation has limitations.First, we did not include a control group because this is real-world data.6][7][8] Secondly, we used self-reported weight at 2, 6 and 9 months, which could bias findings.There may be a reporting bias for responders, whereby those losing more weight are most likely to selfreport or over-report to dietitians. 49,50However, after adjustment for missing data, clinically meaningful weight reduction was still achieved. 25Third, just under 50% of patients had HbA1c available at follow-up.Coupled with the relatively small sample size, caution should be used when interpreting the risk of incident type 2 diabetes.
Fourth, our clinical evaluation included a small number of those of Asian ethnicity; this is related to the local population's characteristics.
Further evaluation is required to compare weight loss outcomes in those of Asian vs white ethnicity.Fifth, there were a small number of self-reported weights at 1 year, with the exact therapeutic effect being relatively imprecise.Finally, given this sample was selected based on BMI there is a risk of regression to the mean biasing findings.However, compared to previous trials we had limited other exclusion criteria and present real-world data showing liraglutide's efficacy in those with prediabetes and obesity.

| CONCLUSIONS
Overall, our clinical evaluation shows that in a multi-ethnic population of adults with obesity and prediabetes that, liraglutide 3.0 mg, in com-

T A B L E 1
Baseline demographic and clinical characteristics.

2. 7 )
The mean difference in BW and BMI at time points compared to baseline was analysed using Paired t-test.The mean difference in BMI and BW change at 4 months between responders and nonresponders was analysed using an unpaired t-test.Change in % Body Weight at time points compared to baseline was analysed using the Wilcoxon-signed rank test.Change in % body weight at 4 months between responders and non-responders was analysed by the Mann-Whitney U test.Abbreviations: BMI, body mass index; IQR, interquartile range; Kg, kilograms; m, months; 95%CI, 95% confidence interval; %, percentage.*p < .05;**p < .01;***p < .001;****p < .0001.
Body Weight, HbA1c and BMI Change over 12 Months.A: Body Weight, B: HbA1c, C: BMI, D: % Body Weight Change, Graphs depicting Change in Body weight, HbA1c and BMI over 12 months for responders and non-responders.For Graphs A + C: Points and error bars represent mean and 95% confidence interval.For Graph B + D: points and error bars represent median and interquartile range.Mean Difference in BW and BMI at time points compared to baseline analysed using Paired t-test.Mean difference in BMI and BW change at 4 months between responders and non-responders analysed using unpaired T Test.Mean difference in HbA1c between groups was analysed using Mann Whitney U-test.Mean difference in % Body weight was analysed using Mann-Whitney U-test for 4 month difference between responders and nonresponders, for paired measurements this was analysed using Wilcoxon Signed-rank test.*: p < .05,**: p < .01,***: p < .001,****: p < .0001,BMI, Body Mass Index; Kg, kilograms; m, months; 95%CI, 95% confidence interval.

F I G U R E 3
Waterfall Plots of Body Weight Change Compared to Baseline.(A) 4 month weight change, (B) 2 month weight change, (C) 6 month weight change, (D) 9 month weight change, (E) 12 month weight change.

3 and 4
weight management services differ from trial participants, potentially being at the more severe end of the disease spectrum and of varying demographic factors.This underpins the importance of our realworld data in demonstrating efficacy in a diverse population.24Previous real-world studies have evaluated liraglutide 3.0 mg in weight management services.Data from multiple Canadian weight management clinics showed that %-BW reduced by 7.1% following 6 months of therapy.32In a real-world randomised controlled trial, patients treated with liraglutide 3.0 mg and lifestyle intervention experienced a mean BW reduction of À7.7% at 1 year.33Similar realworld data in populations from South Korea, USA, Belgium and Israel has demonstrated liraglutide's efficacy in those living with obesity.[34][35][36][37]However, compared to our clinical evaluation, outside of the UK, many patients are only covered by specific health insurance programmes and potentially self-fund.This may preclude those from high levels of socioeconomic deprivation from seeking treatment and affects the previous data's external validity when applied to the United Kingdom's (UK) state-funded health service.Alongside our data, The STRIVE study, which is a phase 4 trial of liraglutide within UK tier 3 weight management services, will provide further clarification.

$ 5 .
0 mmol/mol.The Diabetes Prevention Program provides clarity regarding the benefits associated with the reversal of prediabetes.39

F I G U R E 4
Outcome Data, Side Effect Profile and Glycaemic Status Change.(A) Outcome Status following 4 months of treatment for all patients, (B) Side Effect Profile, (C) Change in Glycaemic Status for Responders and Non-responders over follow-up.GI, gastrointestinal; DC, discharged; DNA, did not attend.
bination with lifestyle intervention, leads to clinically meaningful weight and HbA1c reduction.In those of Black African and Caribbean ethnicity, weight loss was less than those of white ethnicity; this was not underpinned by clinical or demographic factors.Those undergoing a total-meal replacement programme with liraglutide were more likely to stop liraglutide therapy.Further real-world evaluation of 3.0 mg liraglutide is required to determine whether (1) ethnicity-specific differences persist at 1 year and beyond, (2) different categorisation of responders and non-responders are required in those of Black African and Caribbean ethnicity, (3) the optimal timing of liraglutide initiation when used alongside total-meal replacement.
Service Evaluation Flow Chart. Figure depicting the number of patients at each stage of the service evaluation and the number of patients with available data.N, number; DNA, did not attend.

Table S5 )
. Non-responders were proportionally more likely to be males (37.5%), of Black African and Caribbean ethnicity or Asian eth-

Table S7
Body weight and BMI change across 12 months.