==> picture [215 x 22] intentionally omitted <==

Single Technology Appraisal

Semaglutide for managing overweight and obesity [ID3850]

Committee Papers

==> picture [215 x 22] intentionally omitted <==

NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE

SINGLE TECHNOLOGY APPRAISAL

Semaglutide for managing overweight and obesity [ID3850]

Contents:

The following documents are made available to consultees and commentators: The final scope and final stakeholder list are available on the NICE website.

1. Company submission from Novo Nordisk

2. Clarification questions and company responses

3. Patient group, professional group and NHS organisation submissions from: a. Obesity UK

4. Evidence Review Group report prepared by Southampton Health Technology Assessments Centre (SHTAC)

5. Evidence Review Group report – factual accuracy check

6.

• Technical engagement response from company

7. Technical engagement responses and statements from experts:

• a. Beverley Burbridge, Support Group Leader – patient expert, nominated by Obesity UK

• b. Kenneth Clare, Director of Bariatric and Metabolic Surgery Services – patient expert, nominated by Obesity UK

• c. Carel le Roux, Professor of Metabolic Medicine – clinical expert, nominated by Novo Nordisk

• d. John Wilding, Professor of Medicine and Honorary Consultant Physician – clinical expert, nominated by Novo Nordisk

• e. Gary McVeigh, Clinical Advisor – clinical expert, nominated by NHS England and NHS Improvement

8. Technical engagement responses from consultees and commentators:

• a. Association for the Study of Obesity

• b. British Dietetic Association

• c. British Obesity and Metabolic Surgery Society

• d. Royal College of Physicians

• e. UK Obesity Organisation

9. Evidence Review Group critique of company response to technical engagement prepared by SHTAC

Any information supplied to NICE which has been marked as confidential, has been redacted. All personal information has also been redacted.

NATIONAL INSTITUTE FOR HEALTH AND CARE EXCELLENCE

Single technology appraisal

Semaglutide for managing overweight and obesity [ID3850]

Document B

Company evidence submission

20 July 2021

File name Version ID3850_Semaglutide 1.0 2.4mg NICE Document B_Final_redacted

Contains Date confidential information Yes 20 July 2021

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Instructions for companies

This is the template for submission of evidence to the National Institute for Health and Care Excellence (NICE) as part of the single technology appraisal (STA) process. Please note that the information requirements for submissions are summarised in this template; full details of the requirements for pharmaceuticals and devices are in the user guide.

This submission must not be longer than 150 pages, excluding appendices and the pages covered by this template. If it is too long it will not be accepted.

Companies making evidence submissions to NICE should also refer to the NICE guide to the methods of technology appraisal and the NICE guide to the processes of technology appraisal.

In this template any information that should be provided in an appendix is listed in a box.

Highlighting in the template (excluding the contents list)

Square brackets and grey highlighting are used in this template to indicate text that should be replaced with your own text or deleted. These are set up as form fields, so to replace the prompt text in [grey highlighting] with your own text, click anywhere within the highlighted text and type. Your text will overwrite the highlighted section.

To delete grey highlighted text, click anywhere within the text and press DELETE.

Grey highlighted text in the footer does not work as an automatic form field, but serves the same purpose – as prompt text to show where you need to fill in relevant details. Replace the text highlighted in [grey] in the header and footer with appropriate text. (To change the header and footer, double click over the header or footer text. Double click back in the main body text when you have finished.)

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Contents

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Tables and figures

Table 1: The decision problem ................................................................................. 12 Table 2: Technology being appraised ...................................................................... 16 Table 3: Clinical effectiveness evidence ................................................................... 31 Table 4: Summary of STEP 1 study methodology .................................................... 34 Table 5: Patient baseline demographics and disease characteristics, STEP 1 ........ 38 Table 6: Baseline comorbidities, STEP 1 ................................................................. 40 Table 7: Summary of statistical analyses in STEP 1 ................................................ 41 Table 8: Glycaemic status at baseline during STEP 1.............................................. 52 Table 9: Co-primary, confirmatory and selective supportive secondary endpoints for the BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity subgroup ................................................ 55 Table 10: Co-primary, confirmatory and selective supportive secondary endpoints for the BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high CV-risk subgroup 56 Table 11: Baseline characteristics of patients in the STEP 1 and SCALE 1839 trials: patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk 58 Table 12: Base-case results of the indirect comparisons of semaglutide 2.4 mg (STEP 1) versus liraglutide 3.0 mg (SCALE 1839) ................................................... 62 Table 13: Most common adverse events reported in ≥ 10% of patients ................... 65 Table 14: Summary of STEP trials and rationale for their inclusion in the submission 70 Table 15: Summary list of published cost-effectiveness studies .............................. 81 Table 16: Baseline characteristics for populations of interest ................................... 86 Table 17: Starting non-diabetic hyperglycaemia distribution for both populations of interest ..................................................................................................................... 90 Table 18: Definition of baseline cohort characteristics ............................................. 92 Table 19: Features of the economic analysis ........................................................... 96 Table 20: Definition of treatment effects on physiological parameters included in the model 100 Table 21: Change in physiological parameter values – STEP 1 – all patients ........ 102 Table 22: Percentage of non-diabetic hyperglycaemia patients with glycaemic status reversal – STEP 1 – all patients ............................................................................. 104 Table 23: Change in physiological parameter values – Liraglutide 3.0 mg – all patients ................................................................................................................... 105 Table 24: Adverse event rates ............................................................................... 108 Table 25: Incremental probability of adverse events per cycle for Semaglutide 2.4 mg versus diet and exercise ......................................................................................... 108 Table 26: Bariatric surgery criteria.......................................................................... 109 Table 27: Bariatric surgery efficacy and case fatality ............................................. 109 Table 28: Efficacy data sources in relation to stopping rules and proportion of responders ............................................................................................................. 113 Table 29: Definition of catch-up rate ...................................................................... 115 Table 30: Risk equations used for obesity-related complications ........................... 117 Table 31: Clinical inputs for calibration of composite cardiovascular endpoints ..... 118 Table 32: BMI-dependent risk functions for knee osteoarthritis .............................. 120 Table 33: Relative risks used to adjust general population mortality ...................... 121 Table 34: Short-term mortality from complications ................................................. 122 Table 35: Re-estimate association between BMI and utility ................................... 125 Table 36: Coefficients to adjust BMI dependent baseline utility as a function of age from Søltoft et al. 2009 and Burström et al. 2001 studies....................................... 125 Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Table 37: Disutility values associated with acute events ........................................ 126 Table 38: HRQoL decrements of acute events ...................................................... 127 Table 39: HRQoL decrements for bariatric surgery ................................................ 127 Table 40: Treatment-related adverse event disutilities ........................................... 128 Table 41: Summary of utility values for cost-effectiveness analysis ....................... 129 Table 42: Annual treatment costs (list price) .......................................................... 131 Table 43: Annual obesity related complication costs applied to health states in the model 132 Table 44: Costs of treatment-related adverse events ............................................. 134 Table 45: Bariatric surgery costs ............................................................................ 134 Table 46: Acute event costs ................................................................................... 135 Table 47: Comparison of cost inputs between current submission and TA664 ...... 136 Table 48: Summary of baseline cohort characteristics and AE probabilities in the economic model ..................................................................................................... 138 Table 49: Summary of efficacy inputs applied in the economic model ................... 139 Table 50: Summary of epidemiological inputs, utilities and costs applied in the economic model ..................................................................................................... 140 Table 51: Summary of assumption applied in the economic model ....................... 143 Table 52: Base-case results for semaglutide 2.4 mg versus diet and exercise ...... 147 Table 53: Base-case results (probabilistic) for semaglutide 2.4 mg versus diet and exercise .................................................................................................................. 148 Table 54: Deterministic sensitivity analysis for semaglutide 2.4 mg versus diet and exercise .................................................................................................................. 151 Table 55: Key scenario analyses ........................................................................... 153 Table 56: Summary scenario results for semaglutide 2.4 mg versus diet and exercise 155 Table 57: Base case results (deterministic) for semaglutide 2.4 mg versus liraglutide 3.0 mg .................................................................................................................... 158 Table 58: Base-case results (probabilistic) for semaglutide 2.4 mg versus liraglutide 3.0 mg .................................................................................................................... 158 Table 59: 915 analysis for semaglutide 2.4 mg versus liraglutide 3.0 mg .............. 160 Table 60: Summary scenario results for semaglutide 2.4 mg versus liraglutide 3.0 mg 162 Table 61: Summary of evidence regarding therapeutic weight loss for complications of obesity from Cefalu et al. 2015 ........................................................................... 164

Figure 1: Tiered model of obesity services ............................................................... 24 Figure 2: Number of items (orlistat) prescribed for the treatment of obesity ............. 26 Figure 3: STEP 1 study design schematic ................................................................ 33 Figure 4: Mean (%) change in body weight from baseline by week, observed data . 48 Figure 5: Proportion of patients achieving ≥ 5%, ≥ 10%, ≥ 15% and ≥ 20% weight loss at 68 weeks ....................................................................................................... 49 Figure 6: Mean (%) change in body mass index from baseline by week .................. 50 Figure 7: Mean change in HbA1c by week ............................................................... 51 Figure 8: Change in SF-36 physical functioning score by week ............................... 53 Figure 9: Change in IWQOL-Lite physical function score by week ........................... 54 Figure 10: Prevalence and duration of gastrointestinal events ................................. 67 Figure 11: Obesity model structure .......................................................................... 89

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Figure 12: Projected BMI change for semaglutide 2.4 mg and diet and exercise over time for patients with BMI ≥ 30 kg/m[2] with one or more obesity related comorbidities 103 Figure 14: Projected BMI change for semaglutide 2.4 mg and liraglutide 3.0 mg over time for patients with BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD 107 Figure 15: Annual probability of knee replacement surgery by level of BMI ........... 119 Figure 16: BMI related HRs from Bhaskaran et al. 2018 ........................................ 121 Figure 17: Cost-effectiveness plane for semaglutide 2.4 mg versus diet and exercise 149 Figure 18: Cost effectiveness acceptability curve plane for semaglutide 2.4 mg versus diet and exercise ......................................................................................... 150 Figure 19: Deterministic sensitivity analysis for semaglutide 2.4 mg versus diet and exercise .................................................................................................................. 152 Figure 20: Cost-effectiveness plane for semaglutide 2.4 mg versus liraglutide 3.0 mg 159 Figure 21: Deterministic sensitivity analysis for semaglutide 2.4 mg versus liraglutide 3.0 mg .................................................................................................................... 161

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

List of abbreviations

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

B.1. Decision problem, description of the technology and clinical care pathway

B.1.1. Decision problem

Table 1 presents the decision problem for the submission. The population defined in the final scope is consistent with the anticipated marketing authorisation from the Medicines and Healthcare products Regulatory Agency (MHRA) for semaglutide 2.4 mg (''''''''''''''''''''').

Semaglutide 2.4 mg is indicated as an adjunct to a reduced-calorie diet and increased physical activity for chronic weight management (weight loss and weight maintenance) in adult patients with an initial body mass index (BMI) of:

• ≥ 30 kg/m[2] (obesity), or

• ≥ 27 kg/m[2] to < 30 kg/m[2] (overweight) in the presence of at least one weightrelated comorbidity

Note, that as part of the MHRA licence, a stopping rule may be applied to ‘nonresponders’ of semaglutide 2.4 mg; however, discussions with the Medicines and Healthcare products Regulatory Agency (MHRA) are still ongoing. The rule states that if patients have not lost at least 5% of their initial body weight after 6 months on the 2.4 mg/day maintenance dose of semaglutide 2.4 mg, a decision is required on whether to continue treatment, taking into account the benefit-to-risk profile in the individual patient.[1]

This submission focuses on adult patients with a BMI of ≥ 30 mg/kg[2] in the presence of at least one weight-related comorbidity.

As per NICE quality standards, QS127, these patients are eligible for treatment within specialist weight management services (SWMS; defined as a weight management service led by a specialist multidisciplinary team e.g. Tier 3 or integrated Tier 3 and Tier 4 services where pharmacotherapy is provided [see Section B.1.3.4.1 for tier definitions]) in the UK.[2] This population is narrower than the technology’s anticipated marketing authorisation because patients with obesity who

have comorbidities are anticipated to benefit most from pharmacological treatment Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

within SWMS in National Health Service (NHS) England clinical practice. The comorbidities recorded at baseline in the key clinical trial for this submission (STEP 1) were considered by clinicians to be highly representative of the weight-related comorbidities typically seen in SWMS in UK clinical practice. This submission will also address the patient population eligible to receive liraglutide 3.0 mg (Saxenda[®] ), which is available within SWMS for patients with a BMI ≥ 35 mg/kg[2] with non-diabetic hyperglycaemia and high cardiovascular (CV) risk.

Obesity is a serious chronic disease associated with significant morbidity and mortality.[3] Obesity is also associated with a substantial number of weight-related comorbidities that have significant detrimental impact on patients and healthcare systems (see Section B.1.3.3).[4-15] Therefore, treating obesity is paramount to reducing the burden on patients and the healthcare system.

Indeed, studies have shown that as little as 5% weight loss in patients with obesity can lead to significant improvements in patient health.[16] Furthermore, weight loss of ≥ 10%, and in particular, weight loss of 15%, is associated with substantial benefits to patients with obesity who have comorbidities (see Section B.1.3.3.1). Weight loss of 10–15% can reduce in the burden of existing comorbidities, reduce the risk of developing further weight-related comorbidities and improve patient quality of life.[16-] 22 This view is supported by the clinical community, who have also advised that weight loss of 10–15% is considered highly significant and seldom achieved with standard care, suggesting that there are a substantial number of comorbidities that can be meaningfully alleviated with weight loss of this magnitude.[23]

Despite the benefits of treating obesity with pharmacotherapy, there are currently very few efficacious pharmacological treatment options for patients with obesity who are seeking SWMS. Orlistat was the mainstay of treatment but is rarely used today and is associated with undesirable side effects, insufficient weight loss and poor adherence (discussed further in Section B.1.3.4.2).[24-26] More recently, NICE approved liraglutide 3.0 mg for use within SWMS for patients with a BMI ≥ 35 mg/kg[2] with non-diabetic hyperglycaemia and high CV risk (TA664).[27] However, there remains a substantial unmet clinical need for patients with obesity and comorbidities who can access SWMS but do not meet the criteria for liraglutide 3.0 mg, meaning these patients may be deprived of benefitting from clinically meaningful weight loss. Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

This view is shared by the clinical community, who have expressed a desire to use pharmacotherapy to treat a broader population of patients in SWMS, rather than just those eligible for liraglutide 3.0 mg.[23]

The need for additional treatment options for obesity is also apparent when considering the increasing prevalence of the disease[28] , which will result in more patients requiring SWMS. If additional pharmacotherapy treatment options are not made available within SWMS, an increasing number of patients will be at risk of developing additional weight-related comorbidities or dying from obesity-related complications such as COVID-19.[22, 29, 30] Furthermore, there is a substantial burden on the healthcare system due to increasing numbers of obesity-related hospitalisations and comorbidities that require treatment[31, 32] ; this burden could be alleviated with the introduction of new, efficacious therapies.

Semaglutide 2.4 mg has the potential to address this unmet clinical need by providing a treatment option for those patients with obesity and comorbidities who are referred to a SWMS. The weight loss observed with semaglutide 2.4 mg (~15% weight loss observed across the entire STEP trial programme) is a significant advancement for obesity treatment with more than double the weight loss observed over existing pharmacotherapy options (patients typically achieve between 5–7% weight loss in SWMS[23] ), representing a substantial step change in treatment.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

Table 1: The decision problem

B.1.2. Description of the technology being appraised

Table 2 presents a description of semaglutide 2.4 mg. The draft summary of product characteristics (SmPC) is presented in Appendix C. The European Public Assessment Report (EPAR) is expected to be available in 2022. The publication date of the Public assessment report (PAR) of the MHRA is still to be confirmed.

Table 2: Technology being appraised

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

16 of 176

B.1.3. Health condition and position of the technology in the treatment pathway

B.1.3.1 Disease background

Obesity is a serious chronic disease and is considered one of the greatest long-term health challenges facing the United Kingdom (UK).[3, 30] According to the World Health Organization (WHO), obesity is defined as abnormal or excessive fat accumulation

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

17 of 176

that may impair health, and is characterised by a BMI of ≥ 30 kg/m[2] . Patients with obesity can be further classified into one of the following BMI categories:[33]

• Obesity Class I: BMI of ≥ 30 – < 35 kg/m[2]

• Obesity Class II: BMI of ≥ 35 – < 40 kg/m[2]

• Obesity Class III: BMI of ≥ 40 kg/m[2]

Obesity is a complex and multifactorial disease driven by imbalances in energy intake and expenditure, leading to weight gain.[34] A number of causative mechanisms for obesity have been observed, including appetite dysregulation and abnormal hormone signalling, as well as a range of genetic factors (e.g. genes associated with excess weight gain), psychological factors (e.g. depression and food addiction) and environmental factors (e.g. poor socioeconomic conditions, and certain diseases or medications).[34-38]

Regarding the hormonal mechanism of the disease, a gastrointestinal (GI)-derived hormone known as glucagon-like peptide 1 (GLP-1) integrates in the hypothalamus where it regulates feelings of appetite and satiety.[39] More specifically, GLP-1 is released in response to food intake and regulates blood glucose by enhancing insulin secretion and inhibiting glucagon secretion from the pancreas.[39, 40] Activation of GLP-1 receptors in the brain also regulates appetite by increasing satiety and reducing hunger, a process modulated by several other hormones and gut peptides, including leptin and ghrelin.[41-43] In addition, GLP-1 is known to reduce energy intake in humans.[41, 42, 44]

B.1.3.2 Epidemiology

As of 2019, the majority of adults in England were overweight or obese (68% of men and 60% of women); this included 27% of men and 29% of women who were obese.[45] These numbers reflect a sharp rise in the prevalence of obesity in England over recent decades, increasing from 13% of men and 16% of women in 1993.[45] The prevalence of obesity in England increases with age, from 13% of adults aged between 16 and 24 years, to 36% of those aged 65 to 74 years; although it was lower among adults aged ≥ 75 years (26%).[45] Among all adults, the prevalence of obesity was highest in the North East and West Midlands (34% of adults) and was

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

18 of 176

lowest in the South East (24%) and London (23%), suggesting a socioeconomic trend (further discussed in B.1.4).

By 2035, an estimated 37% of adult men and women in the UK will be obese (BMI ≥30)[28] , highlighting the importance of tackling obesity now.

B.1.3.2.1 Mortality and life-expectancy

Obesity is associated with significant morbidity and mortality. Public Health England estimates that obesity is responsible for more than 30,000 deaths each year.[32] In 2018/2019, 11,117 hospital admissions were directly attributable to obesity and it was a factor in 876,000 hospital admissions, an increase of 4% and 23% on the previous year, respectively.[31]

Obesity in adulthood is associated with a decrease in life expectancy of approximately 6–13 years.[46-48] A systematic review and meta-analysis of 97 studies including more than 2.88 million individuals and 270,000 deaths found patients with obesity had an 18% increase in all-cause mortality compared with individuals with a normal BMI.[49] Similarly, a population-based cohort study of 3.6 million adults in the UK using data from the Clinical Practice Research Datalink (CPRD) found that in patients with obesity aged ≥ 40 years, life expectancy was 4.2 years shorter in men and 3.5 years shorter in women compared with individuals of healthy weight (BMI: 18.5–24.9 kg/m[2] ).[50] These results are unsurprising given that patients who are overweight or obese are at increased risk of developing weight-related comorbidities than individuals with a normal BMI.[51]

B.1.3.3 Burden of disease

B.1.3.3.1 Weight-related complications and the impact of weight loss

Patients with a BMI of ≥ 30 kg/m[2] have a significant risk of developing weight-related comorbidities[4] , of which a large number are reported throughout the literature. Perhaps the most pertinent given recent events is the strong relationship between obesity and COVID-19.[29, 52] Patients with obesity are at an increased risk of hospitalisation, severe symptoms, advanced treatment requirements (e.g. a need for mechanical ventilation or admission to intensive care units) or dying from COVID19.[29, 30] Furthermore, the higher a patient’s BMI, the more likely they are to die from

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

19 of 176

the disease. Public Health England estimated that having a BMI of 35–40 mg/kg[2] could increase a person’s risk of dying from COVID-19 by 40%, while a BMI > 40 mg/kg[2] could increase the risk by 90%.[53]

Other commonly reported weight-related complications include: pre-diabetes, type 2 diabetes (T2D), cardiovascular disease (CVD, e.g. coronary heart disease, dyslipidaemia and hypertension), knee osteoarthritis, liver disease (e.g. NASH or NAFLD), reproductive system disorders (e.g. infertility), kidney disease, gout, asthma and obstructive sleep apnoea.[4-14] Note that this list is not comprehensive and that the number of different weight-related complications observed in UK clinical practice, and their burden on the healthcare system, is vast.

The benefits of weight loss in patients with obesity who have comorbidities are well documented. Indeed, as little as 5% weight loss can lead to significant improvements in glycaemic status, blood pressure, triglycerides, and high density lipoprotein (HDL) cholesterol.[16] Weight loss of 5−10% is associated with a reduction of intrahepatocellular lipids in NAFLD; reduction in triglycerides, increase in HDL cholesterol and reduction in non-HDL cholesterol; improvements in ovulation and regularisation of menstrual cycles; and prevention of T2D and various cancers.[54] Furthermore, the Look AHEAD study, which examined the impact of short-term weight loss on the incidence of cardiovascular disease in 4,384 patients with obesity with T2D, found that patients who achieved 10% weight loss in the first year after treatment had a 20% reduction in the risk of cardiovascular events.[17] Weight loss ≥ 10% has also been shown to improve symptomatology and increase physical function in people with osteoarthritis, gastroesophageal reflux disease, and obstructive sleep apnoea; reduce inflammation and fibrosis in NASH; and reduce the frequency of incontinence.[18]

Substantial health benefits have been observed with weight loss of 15% or greater. For instance, weight loss of ≥ 15% has also been shown to greatly reduce blood pressure in patients with hypertension, much more so than seen with 5% weight loss.[54] Similarly, although patients with dyslipidaemia (i.e. elevation of plasma cholesterol or triglycerides) and hyperglycaemia (elevated haemoglobin A1C [HbA1C]) benefit from as little as 3% weight loss, triglycerides and HbA1C are further

reduced with ≥ 15% weight loss.[54] Furthermore, weight loss > 15% has been shown to reduce the risk of heart failure and cardiovascular mortality.[20, 21]

Besides reducing the impact of current weight-related comorbidities, weight loss also reduces the chance of developing additional weight-related comorbidities. A UK study of 571,961 patients from the CPRD GOLD database found that a median weight loss of 13% was associated with risk reductions in developing T2D (41%), sleep apnoea (40%), hypertension (22%), dyslipidaemia (19%), asthma (18%), chronic kidney disease (15%), hip or knee osteoarthritis (13%) and heart failure (8%).[22]

B.1.3.3.2 Patient health-related quality of life and the impact of weight loss

People who are obese typically have poorer health-related quality of life (HRQoL) than individuals with normal weight.[55] Individuals living with being obese internalise feelings of being stigmatised and often feel shame or distress about their size and ‐ habits; this can contribute to low self esteem, impaired work and social life, and ‐ diminished overall psychological well being.[56]

Perhaps the most prominent impact obesity has on patient HRQoL is its effect on physical functioning.[57, 58] A cross-sectional study involving approximately 9,000 individuals in the UK reported statistically significant differences in the Short Form 36 Health Survey Questionnaire (SF-36) physical functioning subscale scores between overweight (BMI: 25.0−29.9 kg/m[2] ), moderately obese (BMI: 30.0−39.9 kg/m[2] ), and morbidly obese (BMI: ≥ 40.0 kg/m[2] ) patients, with scores worsening as BMI increased.[57] The authors noted that physical well-being deteriorated markedly with increasing weight.[57] Similarly, a systematic review based on 43,086 patients found that adults who were obese had significantly reduced physical component scores of the SF-36 questionnaire compared with individuals of healthy weight (weighted mean difference referent to normal weight: –2.54 points for Class I obesity; –3.91 for Class II obesity; –9.72 for Class III obesity; all p < 0.001).[59]

Improvements in physical functioning have been observed among patients achieving weight loss.[60, 61] More specifically, weight loss of 5−10% is associated with improvements in certain aspects of HRQoL (e.g. mental functioning), with the most notable improvements observed in physical functioning scores for both SF-36 and

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

21 of 176

the Impact of Weight on Quality of Life-Lite (IWQOL-Lite) domains.[57, 62] Pertaining to this, a study of overweight and obese individuals (mean BMI, 32.5 kg/m[2] ) who underwent a 6 month clinical weight loss programme showed increases in scores on the SF-36 physical functioning subscale.[60] This finding echoes results of an earlier study which found a statistically significant improvement in physical functioning subscale scores among obese individuals (n = 38) following a weight loss treatment programme.[61] Overall, these results demonstrate that weight loss leads to substantial improvements in patient HRQoL, particularly with regards to physical functioning.

B.1.3.3.3 Socioeconomic burden

Obesity incurs a significant financial burden through both increased total direct healthcare costs and indirect costs.[63, 64] In the UK, the overall cost of obesity to the wider society is estimated to be £27 billion and is estimated to increase to approximately £50 billion in 2050 if obesity rates continue to rise.[32] The total cost to the NHS specifically was estimated to be £6.1 billion in 2014/15 and is projected to reach £9.7 billion by 2050.[32]

A significant portion of the economic burden of obesity is driven by the associated comorbidities, which impose substantial medical costs from their treatment. In a UK study of 250,046 patients, healthcare costs were greater as the BMI category increased, reaching a maximum mean annual cost per person of £456.[15] However, after adjusting for BMI, the presence of a comorbidity was the single largest predictor of healthcare costs, with an additional £1,366 mean increase in annual patient costs if a comorbidity was present. The second greatest predictor was depression, at £1,044 per patient per year.[15]

Studies investigating the indirect costs of obesity in the UK are scarce, and the value is hard to quantify given the large number of weight-related complications.[65] As such, any estimate of indirect cost is likely to be an underestimate as calculations cannot include all diseases associated with obesity and because differing societal variables are used in different studies.[65] In 1998, the indirect costs of obesity were estimated to be greater than the direct costs (£2.1 billion versus £500 million, respectively).[63] Some examples of the indirect costs associated with obesity include absenteeism

22 of 176

from work, the cost of time spent in travel and waiting by the patient and by caregivers, as well as the time spent in actually receiving treatment.[63, 66]

Taken together, this abundance of literature supports the notion that substantial weight loss is associated with significant and far-reaching downstream benefits, both in terms of clinical outcomes for the patients and their quality of life, and ultimately the cost-savings associated with reducing the burden and incidence of weightrelated comorbidities.

B.1.3.4 Clinical care pathway and proposed positioning of the technology

Semaglutide 2.4 mg should be administered as an adjunct to a reduced-calorie diet and increased physical activity to patients with a BMI ≥ 30 mg/kg[2] and at least one weight-related comorbidity (listed in Section B.1.3.4.3) within a SWMS [defined below in Section B.1.3.4.1]). It should be noted that as part of NICE quality standards, QS127, these patients are eligible for SWMS in UK clinical practice (also further discussed in Section B.1.3.4.1).[2]

The following sections provide an overview of the UK clinical guidelines for managing obesity in the UK, the current treatment options and unmet need, and justification for proposed use of semaglutide 2.4 mg in SWMS.

B.1.3.4.1 UK clinical guidelines

Within the NHS in England, obesity management is currently delivered through a tiered system.[67] In 2014, the Department of Health Working Group report on the joined up clinical pathways for obesity described the four weight management tiers; these are depicted in Figure 1.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

23 of 176

Figure 1: Tiered model of obesity services

==> picture [452 x 185] intentionally omitted <==

Source: Welbourn et al. 2018.[68]

The different tiers of weight management services cover different activities; however, it should be noted that definitions of weight management tiers are intended as a guide and local definitions vary.[69, 70] Tier 1 is a universal intervention aimed at prevention and re-enforcement of healthy lifestyle principles[71] , and Tier 1 forms the basis of all future treatment. Tier 2 covers lifestyle interventions that aim to reduce a person's energy intake and help them to be more physically active by changing their lifestyle behaviour; interventions may also include pharmacotherapy in appropriate clinical circumstances.[72] Broadly, Tier 3 and Tier 4 comprise multidisciplinary team (MDT) weight assessment and management clinics (WAMCs) located in primary or secondary care that provide advice for patients[68] ; at a minimum, MDTs must include a clinician, specialist dietician, physician and clinical psychologist.[69, 73] In this

submission, these settings will be referred to as SWMS to reflect settings where pharmacotherapy can be offered under the guidance of a specialist MDT. The primary aim of Tier 3 is to achieve clinically meaningful weight loss (defined by NICE as weight loss of 5–10%)[74] in patients with weight-related comorbidities who are not considering bariatric surgical intervention (i.e. Tier 4).[71] Another important remit of Tier 3 is to prepare appropriately selected patients for bariatric surgery.[70, 71, 73]

NICE has issued guidance on the management of obesity (CG189), which states that the referral of patients to Tier 3 services should be considered if:[70]

• The underlying causes of living with overweight or obesity need to be assessed

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

24 of 176

• The person has complex disease states or needs that cannot be managed adequately in Tier 2 (for example, the additional support needs of people with learning disabilities)

• Conventional treatment has been unsuccessful

• Drug treatment is being considered for a person with a BMI > 50 kg/m[2]

• Specialist interventions (such as a very-low-calorie diet) may be needed; or

• Bariatric surgery is being considered

Following commissioning guidance from NHS England in 2016[75] , NICE is currently updating CG189 to align with published guidance from The Royal College of Surgeons of England and The British Obesity & Metabolic Surgery Society (BOMSS), which recommends referral to Tier 3 settings for patients with severe and complex obesity (i.e. patients with a BMI > 40 kg/m[2] or a BMI ≥ 35 kg/m[2] with comorbidities; these patients are currently recommended for Tier 4 as part of CG189.[76, 77] However, it should also be noted that as part of the accompanying quality standards for the CG189 guidelines (QS127), NICE states that adults with a BMI ≥ 30 kg/m[2] for whom Tier 2 interventions have been unsuccessful should discuss their choice of alternative interventions for weight management, including referral to a SWMS.[2]

B.1.3.4.2 Current treatment options and relevant comparators for semaglutide 2.4 mg

Interventions for managing obesity include lifestyle interventions (such as diet and exercise); pharmacotherapy, namely orlistat and liraglutide 3.0 mg; and bariatric surgery.

Lifestyle intervention in the form of diet and exercise counselling is essential in treating obesity and forms the basis of all treatment programmes. However, evidence suggests that lifestyle interventions alone are not enough to help patients lose weight and maintain weight loss. In the UK, it was reported that only 20% of individuals in the general population successfully lose 10% of their body weight and maintain that weight loss for 1 year.[78] Given that in clinical practice, lifestyle intervention is prescribed for all patients who are obese, lifestyle intervention without

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

25 of 176

pharmacotherapy is considered a relevant comparator to semaglutide 2.4 mg for the purpose of this submission.

Regarding pharmacotherapy, orlistat is currently recommended for managing obesity in adults who have a BMI ≥ 28 kg/m[2] with associated risk factors such as T2D, hypertension, or hypercholesterolaemia; or a BMI ≥ 30 kg/m[2] . However, as part of the NICE technology appraisal for naltrexone–bupropion (TA494), clinical experts and consultees reported that orlistat is not commonly prescribed in clinical practice due to efficacy and tolerability issues.[25] These sentiments were reflected in the appraisal of liraglutide 3.0 mg (TA664), where clinicians stated that many people decide not to have orlistat or stop taking it because of undesirable, and socially unacceptable side effects.[27] Orlistat also appears to be much less effective in general practice than in randomised clinical trials, leading to undesirable side effects, poor adherence and insufficient weight loss outcomes.[24, 25] Pertaining to this, data from NHSE highlight a long-term downward trend on the prescription of orlistat in the UK , as depicted in Figure 2.

Figure 2: Number of items (orlistat) prescribed for the treatment of obesity

==> picture [452 x 257] intentionally omitted <==

Notes: y-axis = number of items prescribed in thousands. Source : National Health Service England, 2021.[26]

For these reasons, orlistat is not considered a relevant comparator to semaglutide 2.4 mg for this submission. Conversely, liraglutide 3.0 mg is considered a relevant comparator to semaglutide 2.4 mg for this submission for those patients with a BMI ≥ 35 mg/kg[2] with non-diabetic hyperglycaemia and high CV risk. Liraglutide 3.0 mg is recommended by NICE for use in a restricted population of adult patients with obesity, only if:[27]

• They have a BMI ≥ 35 kg/m[2] and non-diabetic hyperglycaemia (defined as a HbA1c level of 42–47 mmol/mol or a fasting plasma glucose level of 5.5– 6.9 mmol/litre

• They have a high risk of cardiovascular disease based on risk factors such as hypertension and dyslipidaemia

• It is prescribed in secondary care by a specialist multidisciplinary Tier 3 weight management service

In the UK, bariatric surgery is only available for patients with a high BMI (a BMI ≥ 40 kg/m[2] , or between 35–40 kg/m[2] and other significant disease) and is rarely used in clinical practice (between 0.002–0.1% of eligible patients).[25, 71] Hence, bariatric surgery was not included in the scope and is not considered a relevant comparator to semaglutide 2.4 mg for the purpose of this submission.

B.1.3.4.3 The unmet clinical need and proposed use of semaglutide 2.4 mg

Currently, there are limited efficacious pharmacological treatment options for patients with obesity who are seeking SWMS. Orlistat was the mainstay of treatment but is rarely used today and is associated with undesirable side effects, insufficient weight loss and poor adherence (as discussed in Section B.1.3.4.2).[24-26] More recently, NICE approved liraglutide 3.0 mg for use within SWMS for patients with a BMI

≥ 35 mg/kg[2] with non-diabetic hyperglycaemia and high CV risk (TA664).[27] However, there remains a substantial unmet clinical need for patients with obesity and comorbidities who can access SWMS but fall outside of the criteria for liraglutide 3.0 mg treatment; these patients may be deprived of benefitting from clinically meaningful weight loss.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

27 of 176

Indeed, studies have shown that patients with a BMI below 35 mg/kg[2] (and even below 30 mg/kg[2] ) can benefit from weight loss.[71] Furthermore, clinicians explained that weight loss of 10–15% is considered highly significant and not something routinely possible with standard care, and that there are a significant number of weight-related comorbidities that can be meaningfully alleviated from a weight loss of this magnitude.[23] Patients with a BMI ≥ 35 mg/kg[2] may also benefit from additional pharmacotherapy options, and NICE acknowledges that weight loss > 10% may be needed for patients with a BMI ≥ 35 mg/kg[2] .[74] However, in SWMS currently, clinicians advised that patients typically achieve between 5–7% weight loss[23] , and the most recent data indicates that only between 20–25% are currently achieving ≥ 10% weight loss with the current pharmacotherapy options.[71]

The need for additional treatment options for obesity is also apparent when considering the increasing prevalence of the disease[28] , which will result in more patients requiring SWMS. Those patients with a BMI ≥ 30 kg/m[2] with comorbidities who do not meet the criteria for liraglutide 3.0 mg in SWMS will be left without an effective pharmacotherapy treatment option, which may deprive them of benefitting from clinically meaningful weight loss and cause their disease to worsen. This can put patients at risk of developing additional weight-related comorbidities or dying from obesity-related complications such as COVID-19.[22, 29, 30] Furthermore, there is a substantial burden on the healthcare system due to increasing numbers of obesityrelated hospitalisations and comorbidities that require treatment[31, 32] ; this burden could be alleviated with the introduction of new, efficacious pharmacotherapies. As such, there is a clear need for additional pharmacotherapy options for patients being treated in SWMS. This view is shared by the clinical community, who have expressed a desire to use pharmacotherapy to treat a broader population of patients in SWMS, rather than just those eligible for liraglutide 3.0 mg.[23]

Taken together, there is a clear need for a more widely accessible and effective pharmacotherapy option within SWMS for treating patients with obesity, particularly for those patients who have comorbidities that can be meaningfully alleviated with weight loss. As such, the proposed target population for semaglutide 2.4 mg is patients with a BMI ≥ 30 mg/kg[2] (obese) and at least one weight-related comorbidity.

These patients represent a population that has the potential to experience substantial clinical benefits from weight loss with semaglutide 2.4 mg.

B.1.4. Equality considerations

Several equality considerations are relevant to this submission:

• BMI variations between different ethnicities – NICE has issued public health guidance (PH46) in line with advice from WHO, which states that members of black, Asian (South Asian and Chinese) and other minority ethnic groups are at an increased risk of chronic health conditions at a lower BMI than the white population (BMI < 25 kg/m[2] ).[79] NICE recommends using lower BMI thresholds (23 kg/m[2] to indicate increased risk and 27.5 kg/m[2] to indicate high risk) for BMI to trigger action to prevent T2D among Asian populations (compared to 25 kg/m[2] and 30 kg/m[2] for the white population, respectively)[79]

• Access inequalities – There are often hurdles to overcome to gain access to treatment

• According to a report by the Royal College of Surgeons, 31% of Clinical Commissioning Groups have at least one mandatory policy on BMI level and weight management.[80] This means patients who require rapid weight loss to get another procedure (e.g. a hip/knee operation) may end up waiting for prolonged periods of time to access services and may or may not achieve the target weight loss to allow their procedure to be performed

• Socioeconomic inequalities – A higher prevalence of obesity has been found in people of lower socioeconomic status, as highlighted by data published by the House of Commons Library:

• In the most deprived areas in England, the prevalence of excess weight is 9% higher than the least deprived areas.[81] This difference was particularly pronounced for women, where 39% of women in the most deprived areas were obese, compared with 22% in the least deprived areas.[45] The most recent available data covers surveys from 2018/19, and shows that levels of excess weight are estimated to be highest in the West Midlands, the North East, and Yorkshire & the Humber[81]

• Among people with disabilities, excess weight is 10% higher than among those without disabilities[81]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

29 of 176

• Compared with all other ethnic groups, black people have the highest rates of excess weight, followed by white British people[81]

• Among people with no qualifications, rates of excess weight are 12% higher than among people with Level 4 qualifications or higher (i.e. a degree)[81]

• COVID-19 inequalities – People with obesity are at an increased risk of

hospitalisation, severe symptoms, advanced treatment requirements or dying from COVID-19.[29, 30]

If semaglutide 2.4 mg is recommended for use within SWMS, a greater number of patients with obesity would have access to effective pharmacotherapy, which could help to mitigate some of these inequalities.

B.2. Clinical effectiveness

Semaglutide 2.4 mg is currently being investigated in a series of clinical trials known as the STEP programme. The STEP programme is a comprehensive examination of semaglutide 2.4 mg for weight management in a variety of populations and treatment settings. In total, the programme consists of 17 individual studies: eight global Phase IIIa studies, two regional Phase IIIa studies, six Phase IIIb studies and one Phase IV study. An overview of the STEP programme is presented in Section B.2.11.

B.2.1. Identification and selection of relevant studies

A systematic literature review (SLR) was conducted to identify randomised controlled trial (RCT) evidence for this submission from the current treatment landscape for patients who are overweight or obese.[82] Full details of the process and methods used to identify and select the relevant clinical evidence are in Appendix D.

B.2.2. List of relevant clinical effectiveness evidence

The clinical SLR identified two potentially relevant RCTs from the STEP programme: STEP 1 and STEP 3. A full overview of the STEP clinical trial programme, including rationale for inclusion or exclusion of the trials from the submission, is provided in Section B.2.11.

STEP 3 was a Phase IIIa randomised, double-blind, multicentre, placebo-controlled trial of 611 adults who were obese (BMI ≥ 30 kg/m[2] ), or overweight (BMI ≥ 27 kg/m[2] ) Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

30 of 176

with at least one weight-related comorbidity and without diabetes or HbA1c ≥ 6.5%. During the study, semaglutide 2.4 mg was administered in conjunction with intensive behavioural therapy (IBT), which consisted of combined behavioural counselling, reduced-calorie diet, and increased physical activity. However, as discussed in TA664, IBT is not standard clinical practice in the UK. Therefore STEP 3 is not considered relevant for this submission.

The Phase III STEP 1 trial is the pivotal RCT providing evidence of the clinical benefits of semaglutide 2.4 mg for treating patients with a BMI ≥ 30 mg/kg[2] and at least one weight-related comorbidity. The trial is summarised in Table 3.

Table 3: Clinical effectiveness evidence

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

31 of 176

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

32 of 176

B.2.3. Summary of methodology of the relevant clinical effectiveness evidence

B.2.3.1 Study design

STEP 1 was a Phase III, randomised, double-blind, placebo-controlled study in 1,961 adults who were obese (BMI ≥ 30 kg/m[2] ), or overweight (BMI ≥ 27 kg/m[2] ) with at least one weight-related comorbidity, and without diabetes or HbA1c ≥ 6.5%.[83] The study was conducted across 129 sites in 16 countries, including 10 sites in the UK (of which nine where in England). Figure 3 presents a study design schematic for the STEP 1 study.[83, 84] The primary objective of the study was to compare the effect on body weight of semaglutide 2.4 mg once weekly versus placebo as an adjunct to lifestyle intervention in patients who were overweight or obese.[83]

Figure 3: STEP 1 study design schematic

==> picture [452 x 162] intentionally omitted <==

Source: Novo Nordisk (STEP 1 clinical study report), 2020.[85]

In the main phase, 1,961 patients who were overweight or obese were randomised 2:1 to receive either semaglutide 2.4 mg once weekly or placebo once weekly as an adjunct to lifestyle intervention (counselling, a reduced calorie diet [500 kcal/day deficit relative to estimated total energy expenditure at Week 0] together with 150 minutes/week of physical activity).[83] The study included an initial 16-week doseescalation period during which the dose of semaglutide was gradually increased to the maintenance dose of 2.4 mg once weekly. Treatment was continued for an additional 52 weeks until Week 68 (end of treatment).[85] The study also included a

further 52-week off treatment extension phase, during which a subset of patients who had completed the main phase on the maintenance dose of semaglutide 2.4 mg

33 of 176

once weekly or placebo discontinued both treatment and structured lifestyle intervention.

Table 4: Summary of STEP 1 study methodology

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

34 of 176

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

35 of 176

Other outcomes used in Change from baseline to Week 68 in: the economic • BMI model/specified in the • Incidence of type 2 diabetes scope • Waist circumference

• Glycaemic status

• Cardiovascular events

• Mortality

• Adverse effects of treatment

• Health-related quality of life

• Change in fasting lipid profile from baseline (specifically, HDL and total cholesterol) Pre-planned subgroups A subgroup of 140 patients with a BMI ≤ 40 mg/kg[2] underwent DEXA to assess body composition.

Key: AE, adverse event; BMI, body mass index; DEXA, dual-energy x-ray absorptiometry; DMC, data monitoring committee; EAC, event adjudication committee; eCRF, electronic case report form; HbA1C, haemoglobin A1C; IWQOL-Lite-CT, Impact of Weight on Quality of Life-Lite Clinical Trials Version; SF-36, Short Form-36. Notes:[a] , the full list of permitted and disallowed concomitant medication is available in the STEP 1 clinical study report. Source: Novo Nordisk (STEP 1 clinical study report), 2020[85] ; Wilding et al. 2021.[83]

B.2.3.2 Patient baseline demographics, disease characteristics and comorbidities

Table 5 presents the baseline demographics and disease characteristics for patients in the full analysis set (FAS, see Section B.2.4.1 for a definition) and the BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity subgroup from the STEP 1 study. Patient disposition data for STEP 1 is presented in Appendix D2, alongside a Consolidated Standards of Reporting Trials (CONSORT) diagram of patient flow.

Overall, across both the FAS and the BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity subgroup (hereby referred to as the target population), baseline demographics and disease characteristics were well balanced between the semaglutide 2.4 mg and placebo treatment arms.[83, 86]

Baseline demographics and disease characteristics were also very similar across the FAS and the target population. As expected given the exclusion of the lower BMI patients, some of the disease characteristics were slightly higher in the target population, such as lipid levels, waist circumference and blood pressure (Table 5).

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

36 of 176

Note that ethnicity and BMI category were not reported for the target population; however, in line with the other baseline characteristics, the proportion of patients in each group is expected to be similar to the FAS.

Table 6 presents the baseline comorbidities for patients in the FAS and the target population from the STEP 1 study. As with the baseline demographics and disease characteristics, the distribution was similar between the semaglutide 2.4 mg and placebo treatment arms in both the FAS and target population. Furthermore, the baseline comorbidities were similar between the FAS and the target population.

Clinicians considered the baseline demographics, disease characteristics and comorbidities observed in the STEP 1 study to reflect the UK obesity patient population, including those patients commonly referred to SWMS.[23]

37 of 176

Table 5: Patient baseline demographics and disease characteristics, STEP 1

Table 6: Baseline comorbidities, STEP 1

B.2.4. Statistical analysis and definition of study groups in the relevant clinical effectiveness evidence

Table 7 presents the hypothesis and associated statistical analysis methods adopted in the STEP 1 study.

Table 7: Summary of statistical analyses in STEP 1

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

41 of 176

B.2.4.1 Analyses sets and evaluations

• The FAS included all randomised patients according to the intention-to-treat principle.

• The efficacy evaluation for the STEP 1 study was based on the FAS

• The safety analysis set (SAS) included all randomised patients exposed to at least one dose of randomised treatment (semaglutide 2.4 mg or placebo)

• The safety evaluation for the study was based on the SAS

• The dual energy X-ray absorptiometry (DEXA) analysis set included a subgroup of patients with a BMI ≤ 40 mg/kg[2] at screening and a DEXA scan performed at baseline considered to be of acceptable quality by the imaging laboratory

• The body composition subgroup study was performed on the DEXA analysis set

B.2.4.2 Post-hoc subgroup analyses

The subgroups analysed post-hoc were as follows:

• Patients with a BMI ≥ 30 mg/kg[2] who have at least one weight-related comorbidity (i.e. the target population for this submission)

• Patients with a BMI ≥ 35 kg/m[2] who have non-diabetic hyperglycaemia and high CV risk (i.e. the TA664 population)

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

42 of 176

The efficacy evaluation conducted for the FAS was also conducted for these two subgroups. Note that no safety analyses were conducted for the subgroups.

B.2.4.3 Observation periods

Two observation periods were defined for the efficacy and safety evaluations[85] :

• The in-trial period – defined as the uninterrupted time interval from randomisation to last contact with trial site; used for:

• Efficacy: observed values

• Safety: death and events with potential long latency to diagnosis

• The on-treatment period – defined as the interval from first to last trial product administration plus 2 or 7 weeks of follow-up and excluding any period of temporary treatment interruption defined as > 2 or > 7 consecutive missed doses (corresponding to > 2 or > 7 weeks off treatment).

• The on-treatment period (+2 weeks) was used for:

  • Efficacy: observed values

  • Safety: ECG, laboratory assessments, physical examination and pulse

• The on-treatment period (+7 weeks) was used for:

  • Safety: adverse events (AEs) and adjudicated events

The in-trial and on-treatment periods define the patient years of observation (PYO) and patient years of exposure (PYE), respectively, as the total time duration in the periods.[85]

B.2.4.4 Trial estimands

An estimand is a detailed description of the treatment effect estimated to address a scientific question of interest; more than one estimand can be defined for the same endpoint.[89] The International Council for Harmonisation of Technical Requirements for Registration of Pharmaceuticals for Human Use (ICH) E9 (R1) draft addendum on estimands and sensitivity analysis in clinical trials sets out five estimand strategies for estimating the treatment effect (note that this is used by both the EMA and US Food and Drug Administration [FDA]).[90, 91]

According to the guidance, an estimand is defined by four inter-related attributes: the population of interest, the variable (endpoint) of interest, the way intercurrent events

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

43 of 176

are handled and the population level summary.[90, 91] Different estimands are employed to handle intercurrent events that occur during a trial. In the case of STEP 1, the intercurrent events were the initiation of other anti-obesity therapies (weight management drugs or bariatric surgery) and premature discontinuation of the trial product.

For STEP 1, two estimand strategies were applied to the efficacy analyses to address two different aspects of the treatment effect of semaglutide 2.4 mg:

• The treatment policy estimand (primary estimand used in the trial analyses) – estimated the treatment effect of semaglutide 2.4 mg relative to placebo for all randomised patients regardless of premature discontinuation of trial product or initiation of other anti-obesity therapies (weight management drugs or bariatric surgery)

• The treatment policy estimand estimates the population level treatment effect of semaglutide 2.4 mg regardless of treatment adherence and/or other anti-obesity therapies, and therefore the overall treatment impact of semaglutide 2.4 mg for the indicated patient population. Therefore, these results are presented throughout Sections B.2.6 and B.2.7 (note that these results were also used in regulatory approval)

• The hypothetical (trial product) estimand (secondary estimand used in the trial analyses) – estimated the treatment effect of semaglutide 2.4 mg relative to placebo for all randomised patients, assuming they remained on their randomised treatment for the entire planned duration of the trial and had not initiated other anti-obesity therapies

• The trial product estimand excludes the effects of any other anti-obesity therapies and any effects after first treatment discontinuation, and provides a clinically relevant estimate of the average treatment effect of semaglutide 2.4 mg. These results are used to inform the economic model (Section B.3.3.1.1), which captures the effects of alternative anti-obesity therapy use based on published literature; the results are presented in Appendix E2

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

44 of 176

B.2.4.5 Prediabetes definition used in the analyses

For the analyses presented in this submission, pre-diabetes was defined according the NICE preferred definition (as used in TA664), that is non-diabetic

hyperglycaemia, defined as a HbA1c level of 42–47 mmol/mol (6.0–6.4%) or a FPG level of 5.5–mmol/L.[27, 92] Note that this is different to the definition used in the primary analyses, which used the American Diabetes Association (ADA) definition: HbA1c 5.7−6.4% both inclusive; or FPG ≥ 5.6 mmol/L and ≤ 6.9 mmol/L; or 2-hour post-challenge (OGTT) PG ≥ 7.8 mmol/L and ≤ 11.0 mmol/L.[93]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

45 of 176

B.2.5. Quality assessment of the relevant clinical effectiveness evidence

A summary of the quality assessment for the STEP 1 trial is presented in Appendix D3.

STEP 1 was conducted in accordance with Good Clinical Practice (GCP) as defined by the International Council on Harmonisation (ICH), and in accordance with the ethical principles underlying the Declaration of Helsinki.[85] For each study site, the adequacy of the research facility to execute the protocol requirements was confirmed by the local trial management in the country, before study initiation. The study was conducted by qualified investigators, in accordance with a single protocol to promote consistency across sites and measures taken to minimise bias. The study was also monitored by the sponsor by means of central and off-site monitoring, on-site visits, telephone calls, and regular inspection of the electronic case report form (eCRF) with sufficient frequency to verify the study conduct. In addition, all patients provided written consent before study initiation.[85]

Several committees were involved, including the Novo Nordisk safety committee, which performed safety surveillance during the study.[85] An independent external event adjudication committee (EAC) performed ongoing blinded adjudication of selected event types, using definitions and guidelines pre-specified in the EAC Charter. All protocol amendments and patient-informed consent forms received approval by the Institutional Review Board/Independent Ethics Committee at each site, before study initiation.[85]

Baseline demographics and disease characteristics for patients in the STEP 1 study were generally well-balanced between treatment arms (see Section B.2.3.2), and according to clinical expert opinion, the overall population was representative of the general UK obesity patient population.[23]

As a consequence of the COVID-19 pandemic, it was decided as of 23 March 2020 to stop source data validation of remaining data, as monitors were not able to visit the sites and remote validation was not possible. However, all data were entered into the eCRF and checked for completeness, and data cleaning and casebook sign-off Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

46 of 176

were ensured. The decision was in alignment with ICH GCP and regulatory guidance regarding COVID-19.[85]

B.2.6. Clinical effectiveness results of the relevant trials

This section presents the efficacy results of the STEP 1 trial. The results of the intrial efficacy analysis for the FAS (n = 1,961), evaluated using the treatment policy estimand (see Section B.2.4.3 for definitions), are presented in the following sections. The results of the in-trial efficacy analysis for the FAS using the trial product estimand is presented in Appendix E2. The results of the in-trial efficacy analysis for the BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity subgroup (i.e. the target population) and the BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high CV risk subgroup using the treatment policy estimand is summarised in Section B.2.7; the corresponding results of the in-trial efficacy analysis using the trial product estimand are presented in Appendix E2.

The results of the FAS are the primary focus of the submission and are subsequently used in the model to demonstrate the effectiveness of semaglutide 2.4 mg. The results of the FAS, and not the target population (BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity; summarised in Section B.2.7), were selected as the primary data source given NICE’s preference not to use post-hoc trial data for economic modelling. However, as discussed in the subsequent sections, the results of the FAS and the BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity subgroup are consistent with ~75% of the FAS population having a BMI ≥ 30 and at least 1 weight-related comorbidity (presented in Appendix E2). As such, the FAS can be considered a reasonable proxy for the target population for this submission.

B.2.6.1 Co-primary endpoint: percentage weight change

In the semaglutide 2.4 mg treatment arm, weight loss was observed from the first post-randomisation assessment (4 weeks) onward, reaching a nadir at 60 weeks (Figure 4).[83] In the placebo arm, weight loss decreased less and a plateau was reached after approximately 20 weeks of treatment. The estimated mean weight change at 68 weeks (based on observed data) was −14.9% with semaglutide 2.4 mg, compared to −2.4% with placebo (estimated treatment difference [ETD]: −12%; 95% CI: −13.4, −11.5; p < 0.001).[83] Overall, approximately 95% of patients receiving

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

47 of 176

semaglutide 2.4 mg experienced weight loss, compared with 65% of patients receiving placebo.[85]

The mean change in body weight (based on observed data) from baseline to 68 weeks was −15.3 kg in the semaglutide 2.4 mg treatment arm compared with −2.6 kg in the placebo treatment arm (ETD: −12.7 kg; 95% CI: −13.7, −11.7).[83]

Figure 4: Mean (%) change in body weight from baseline by week, observed data

==> picture [452 x 202] intentionally omitted <==

Source : Wilding et al. 2021.[83]

B.2.6.2 Co-primary and secondary endpoints: categorical weight change

Patients in the semaglutide 2.4 mg treatment arm were more likely to lose ≥ 5% (i.e. co-primary endpoint), ≥ 10%, ≥ 15% and ≥ 20% (i.e. secondary endpoints) of their baseline body weight at Week 68 compared with those who received placebo (Figure 5); the difference was significant (p < 0.001) for the ≥ 5%, ≥ 10% and ≥ 15% thresholds (the 20% threshold was not part of the statistical testing hierarchy).[83] Note that at 68 weeks, categorical weight loss results were available for 1212 patients in the semaglutide 2.4 mg treatment arm and 577 patients in the placebo treatment arm (Figure 5).

48 of 176

Figure 5: Proportion of patients achieving ≥ 5%, ≥ 10%, ≥ 15% and ≥ 20%

weight loss at 68 weeks

==> picture [337 x 251] intentionally omitted <==

Source: Wilding et al. 2021.[83]

B.2.6.3 Secondary endpoint: change in BMI

At baseline, mean BMI was similar in the semaglutide 2.4 mg and placebo treatment arms (37.8 mg/kg[2] versus 38.0 mg/kg[2] , respectively). Figure 6 presents the mean change in BMI by week for patients in the semaglutide 2.4 mg and placebo treatment arms. After 68 weeks, semaglutide 2.4 mg was associated with greater reductions from baseline than placebo in average BMI (-5.54 versus -0.92, respectively; ETD [95%CI]: -4.61 [-4.96, -4.27]).[83]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

49 of 176

Figure 6: Mean (%) change in body mass index from baseline by week

==> picture [452 x 286] intentionally omitted <==

Source : Novo Nordisk (STEP 1 clinical study report), 2020.

B.2.6.4 Secondary endpoint: change in systolic blood pressure

Full details of the waist circumference analysis are presented in Appendix R1. Overall, treatment with semaglutide 2.4 mg resulted in greater reductions from baseline in systolic blood pressure (SBP) compared with placebo (−6.16 mmHg versus −1.06 mmHg; ETD: –5.1; 95% CI: –6.3, –3.9).[83]

B.2.6.5 Fasting lipid profile

Full details of the fasting lipid profile analysis are presented in Appendix R2.

Regarding total cholesterol levels, the mean (standard deviation [SD]) change from baseline to 68 weeks in the semaglutide 2.4 mg and placebo treatment arms was - 0.04 (0.14) and 0.00 (0.13), respectively.[94] For HDL levels, the mean (SD) change from baseline to 68 weeks in the semaglutide 2.4 mg and placebo treatment arms was 0.04 (0.14) and 0.01 (0.14), respectively.[94] Note that these results are on the log-scale.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

50 of 176

B.2.6.6 Change in HbA1c from baseline

Semaglutide 2.4 mg was associated with greater reductions from baseline in HbA1c than placebo at 68 weeks (mean reduction of −0.5% versus −0.2%, respectively; ETD: −0.29%; 95% CI: −0.32, −0.26), as presented in Figure 7.[85] The mean HbA1c decreased from baseline through to Week 52 in both treatment arms, but to a larger extent in the semaglutide 2.4 mg arm (Figure 7).[85]

Figure 7: Mean change in HbA1c by week

==> picture [415 x 280] intentionally omitted <==

Key: HbA1C, haemoglobin A1C. Source: Novo Nordisk (STEP 1 clinical study report), 2020.[85]

B.2.6.7 Secondary endpoint: change in waist circumference

Full details of the waist circumference analysis are presented in Appendix R3. Overall, semaglutide 2.4 mg was associated with greater reductions from baseline in waist circumference than placebo (mean reduction of –13.54 cm versus –4.13 cm, respectively; ETD: –9.42 cm; 95% CI: –10.30¸–8.53).[83]

B.2.6.8 Secondary endpoint: change in glycaemic status

Glycaemic status (normo-glycaemia, prediabetes [non-diabetic hyperglycaemia] and T2D [per NICE preferred definitions]) was assessed at baseline and 68 weeks. At

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

51 of 176

these time points, data were available for 1,306 patients in the semaglutide 2.4 mg arm and 655 patients in the placebo arm.[95] Table 8 presents the glycaemic status at baseline for patients in the semaglutide 2.4 mg and placebo treatment arms. Note that the proportion of patients in each category was similar between treatment arms.

Table 8: Glycaemic status at baseline during STEP 1

Of those patients who had non-diabetic hyperglycaemia at baseline, a greater proportion of patients treated with semaglutide 2.4 mg shifted to normo-glycaemic at 68 weeks than with placebo (79.8% versus 39.1%, respectively).[95] Similarly, of those patients who had either T2D or normo-glycaemia at baseline, a greater proportion of those treated with semaglutide 2.4 mg shifted to or remained normo-glycaemic at 68 weeks than with placebo (61.1% versus 23.5%, and 95.3% versus 83.4% respectively).[95]

B.2.6.9 Secondary endpoint: health-related quality of life

This section presents the results of the HRQoL analyses using the 36-Item Short Form Survey (SF-36) and the short form of Impact of Weight on Quality of Life-Lite for Clinical Trials (IWQOL-Lite-CT). The two measures are the most commonly used to assess the effect of weight loss in patient functioning and HRQoL.[57] The focus of this section is the impact of semaglutide 2.4 mg on patient physical functioning relative to placebo given the prominence of this factor on HRQoL in patients with obesity (as discussed in Section B.1.3.3.2)[57, 58] ; the overall results are presented in Appendix R4. Information about the instruments used in the analyses is also provided in Appendix R4.

B.2.6.9.1 SF-36

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

52 of 176

At baseline, mean SF-36 scores (all domains) were similar between the semaglutide 2.4 mg and placebo treatment arms.[85] After 68 weeks, semaglutide 2.4 mg was associated with improvements in patients’ physical and mental functioning, as measured using SF-36. SF-36 scores are norm-based scores (NBS). NBS are scores transformed to a scale where the 2009 US general population has a mean of 50 and a standard deviation of 10.

Figure 8 presents the mean change in SF-36 physical functioning score from baseline to 68 weeks. The estimated mean change in physical functioning score was significantly greater (p < 0.001) for semaglutide 2.4 mg compared with placebo (2.21 versus 0.41, respectively; ETD: 1.8; 95% CI: 1.2, 2.4).[83] Note also that a higher proportion of patients achieved a clinically meaningful within-patient change (i.e. an increase of at least 3.7 points) with semaglutide 2.4 mg compared with placebo (40% versus 27%; OR [95% CI]: 2.08 [1.60, 2.70]).[83]

Figure 8: Change in SF-36 physical functioning score by week

==> picture [390 x 276] intentionally omitted <==

Notes: SF-36 scores are norm-based scores (NBS). NBS are scores transformed to a scale where the 2009 US general population has a mean of 50 and a standard deviation of 10. Source: Wilding et al. 2021.[83]

B.2.6.9.2 IWQOL-Lite-CT scores

At baseline, mean IWQOL-Lite-CT scores (all domains) were similar between the semaglutide 2.4 mg and placebo treatment arms.[85]

Figure 9 presents the mean change in IWQOL-Lite-CT physical functioning score from baseline to 68 weeks. The estimated mean change in physical function score was significantly greater (p < 0.001) for semaglutide 2.4 mg compared with placebo (14.67 versus 5.25, respectively; ETD: 9.43; 95% CI: 7.5, 11.35).[83] Again, a higher proportion of patients achieved a clinically meaningful within-patient change (i.e. an increase of at least 14.6 points) with semaglutide 2.4 mg compared with placebo (51% versus 33%; OR [95% CI]: 2.72 [2.14, 3.47]).[83]

Figure 9: Change in IWQOL-Lite physical function score by week

==> picture [411 x 302] intentionally omitted <==

Key: IWQOL-Lite-CT, short form of Impact of Weight on Quality of Life-Lite for Clinical Trials. Source: Wilding et al. 2021.[83]

B.2.7. Subgroup analysis

This section presents the post-hoc subgroup analyses for the STEP 1 trial. The intrial efficacy results evaluated using the treatment policy estimand are presented for

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

54 of 176

the target population (i.e. patients with a BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity), which represented 75.0% of the FAS, and the BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high CV risk subgroup, which represented 21.5% of the FAS (presented in Sections B.2.7.1 and B.2.7.2, respectively). The corresponding results of the efficacy analysis for these subgroups using the trial product estimand are presented in Appendix E2. Note that the results of the pre-specified analysis for the DEXA subgroup are provided in Appendix E3.

Overall, semaglutide 2.4 mg performed consistently well across all subgroups analysed.

B.2.7.1 Patients with a BMI ≥ 30 mg/kg[2] with at least one comorbidity

Table 9 presents the results of the efficacy analyses for the target population relative to the efficacy results of the FAS analyses. Overall, the results of the efficacy analyses conducted for the target population were consistent with those of the FAS.[83, 96]

Table 9: Co-primary, confirmatory and selective supportive secondary endpoints for the BMI ≥ 30 mg/kg[2] plus ≥ 1 comorbidity subgroup

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

55 of 176

B.2.7.2 Patients with a BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high cardiovascular risk

Table 10 presents the results of the efficacy analyses for the BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high CV-risk subgroup relative to the efficacy results of the FAS analyses. Overall, the results of the efficacy analyses conducted for the BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high CV risk subgroup were similar to those of the FAS.[83, 96]

Table 10: Co-primary, confirmatory and selective supportive secondary

endpoints for the BMI ≥ 35 mg/kg[2] plus non-diabetic hyperglycaemia plus high CV-risk subgroup

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

56 of 176

B.2.8. Meta-analysis

All efficacy data supporting the use of semaglutide 2.4 mg for the treatment of adult patients with a BMI ≥ 30 kg/m[2] and at least one weight-related comorbidity were provided by the STEP 1 trial; therefore, a meta-analysis was not required.

B.2.9. Indirect and mixed treatment comparisons

The STEP 1 trial provides head-to-head data for semaglutide 2.4 mg as an adjunct to lifestyle intervention (counselling, a reduced calorie diet [500 kcal/day deficit relative to estimated total energy expenditure at Week 0] together with 150 minutes/week of physical activity) versus lifestyle intervention without pharmacotherapy. In the absence of STEP 8, a head to head trial of semaglutide 2.4mg vs liraglutide 3.0mg, which will report in Q4 2021, an indirect treatment comparison (ITC) was conducted between semaglutide 2.4 mg and liraglutide 3.0 mg using individual patient data (IPD).An SLR was conducted to identify relevant studies Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

57 of 176

for inclusion in the indirect comparisons of semaglutide 2.4 mg versus liraglutide 3.0 mg; details of the SLR and a summary of the included studies is presented in Appendix D. The ITCs included data from two trials: STEP 1, providing patient data for semaglutide 2.4 mg; and SCALE obesity and pre-diabetes (SCALE 1839), providing patient data for liraglutide 3.0 mg. A summary of the methodology and outcomes of the STEP 1 trial is presented in Sections B.2.3-B.2.10. A summary of the methodology and outcomes of the STEP 1 trial is presented in Appendix D.1.3.1.

The main patient population of interest consisted of sub-populations of patients from these trials, those with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk.[97] Table 11 presents the baseline characteristics of patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk in the STEP 1 and SCALE 1839 trials. Note that only those characteristics identified as potential effect modifiers are presented (the effect modifiers are discussed further in Section B.2.9.1.1). Overall, the baseline characteristics for patients with a BMI ≥ 35 kg/m[2] with nondiabetic hyperglycaemia and high CV risk were similar between the STEP 1 and SCALE 1839 trial.

Table 11: Baseline characteristics of patients in the STEP 1 and SCALE 1839 trials: patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

58 of 176

B.2.9.1 Methods

Full details of the statistical methods used in the ITCs are provided in the ITC report.[97] The ITCs were conducted using linear regression analyses for the continuous outcomes of change from baseline (listed above).[97] A logistic regression analysis was performed for the binary endpoint of reaching normo-glycemic status.[97] Methods were conducted in accordance with NICE Decision Support Unit (DSU) Technical Support Document (TSD) 18.[98]

B.2.9.1.1 Potential effect modifiers

For both semaglutide 2.4 mg and liraglutide 3.0 mg, exposure is inversely dependent on body weight.[97] In published pharmacokinetic analyses of semaglutide 2.4 mg and liraglutide 3.0 mg, gender was found to have an impact on exposure to liraglutide 3.0 mg, whereas the effect of gender on exposure to semaglutide 2.4 mg was minor. In addition, published evidence indicates that the treatment effect of liraglutide depends on baseline HbA1c in diabetic populations. Race, ethnicity and age were not found to have a clinically relevant effect on exposure for either semaglutide 2.4 mg or liraglutide 3.0 mg. Note that the factors affecting exposure to semaglutide 2.4 mg and liraglutide 3.0 mg were independent of the patient populations (i.e. obese versus diabetic).[97]

Age, dyslipidemia, hypertension and cardiovascular disease (CVD) were also included in the ITC as potential effect modifiers.[97] While these additional included effect modifiers were not thought to be effect modifiers for treatment with GLP-1 receptor agonists, they represent key comorbidity measures where a population adjustment can provide reassurance.[97]

B.2.9.1.2 Analyses conducted

Several analyses were conducted during the ITC. The unadjusted analysis was chosen as the base case because the results of the population adjustment had no impact on the outcomes of the ITC. A series of additional analyses were conducted to explore the impact of adjustments for potential effect modifiers (population adjustment 1 & 2 – see below), time-points, estimands (treatment policy versus trial product) and population (patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk versus patients with non-diabetic

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

59 of 176

hyperglycaemia).[97] The following analyses were conducted for the primary time point of interest (Week 52 for STEP 1 and Week 56 for SCALE 1839):[97]

• Unadjusted analysis – performed on the sub-population of patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk, and for the broader subpopulation of patients with non-diabetic hyperglycaemia

• Population adjustment 1 (results adjusted to SCALE 1839 population) – for all endpoints, gender and body weight were included as potential effect modifiers. For HbA1c and glycaemic status endpoints, baseline HbA1c was also included as a potential effect modifier

• Population adjustment 2 (results adjusted to SCALE 1839 population) – in addition to the potential effect modifiers in population adjustment 1, the following potential effect modifiers were also included:

• Age (in years)

• Dyslipidaemia (yes/no)

• Hypertension (yes/no)

• Cardiovascular disease (CVD) (yes/no)

• Analysis of patients with non-diabetic hyperglycaemia

Additional analyses were conducted at various time-points.[97] It was assumed that if no differences between the results of the unadjusted and adjusted analyses were observed at the primary time point of interest, then adjustment would also have no impact on the results of the analyses at different time points. As this was the case in all analyses (except for the analysis of normo-glycaemic status), only unadjusted analyses were conducted for ITCs considering the following alternative timepoints:[97]

• Week 56 in STEP 1 and Week 56 in SCALE 1839

• Week 68 in STEP 1 and Week 56 in SCALE 1839

• Week 28 in STEP 1 and Week 28 in SCALE 1839

Further scenarios considered the impact of estimands on the ITC and analyses were conducted to explore the impact of using a trial product estimand for all unadjusted analyses conducted (at the primary time point of interest and two alternative time points considered):[97]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

60 of 176

• Week 52 in STEP 1 and Week 56 in SCALE 1839 - primary time point considering the trial product estimand

• Week 68 in STEP 1 and Week 56 in SCALE 1839 considering trial product estimand

• Week 28 in STEP 1 and Week 28 in SCALE 1839 considering the trial product estimand

B.2.9.1.3 Outcomes investigated during the indirect comparisons

As a base case, outcomes in STEP 1 at week 52 were compared with outcomes in SCALE 1839 at week 56, corresponding to approximately 1 year after randomisation in each trial.[97] This comparison was considered to be conservative as week 52 in STEP 1 corresponded to 36 weeks of treatment with the maintenance dose of semaglutide 2.4 mg, whereas week 56 in SCALE 1839 corresponded to 52 weeks on the target dose of liraglutide 3.0 mg.[97] The endpoints compared in the ITCs of

semaglutide 2.4 mg and liraglutide 3.0 mg included change from baseline to Week 52 (STEP 1) or Week 56 (SCALE 1839) in:[97]

• Body weight (%)

• SBP (mmHg)

• HDL cholesterol (mg/dl)

• Total cholesterol (mg/dl)

• HbA1c (%)

• Waist circumference (cm)

Glycaemic status at the end of treatment was also compared:[97]

• T2D (Yes/No)

• Normo-glycaemic (FPG < 5.5 mmol/L and HbA1c < 6.0%; Yes/No)

Note that lipid data were not collected at week 52 or 56 in STEP 1.[97] Therefore, the lipid values corresponding to week 56 in STEP 1 were obtained using linear interpolation of the values collected at week 20 and week 68.[97]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

61 of 176

B.2.9.2 Results

The results of the unadjusted population analysis (base case) at the primary time point of interest (Week 52 for STEP 1 and Week 56 for SCALE 1839) using the treatment policy estimand are presented in Table 12. The results of the unadjusted population analysis at the primary time point of interest using the trial product estimand, and the results of the scenario analyses (population adjustment 1, population adjustment 2, and unadjusted non-diabetic hyperglycaemia population), are provided in Appendix D4.

The results of the ITC suggest that semaglutide 2.4 mg was associated with a statistically significant reduction in change from baseline in body weight, HbA1c and waist circumference compared with liraglutide 3.0 mg.[97] In addition, Semaglutide 2.4 mg was also associated with a numerically greater reduction in SBP compared with liraglutide 3.0 mg, although the results did not reach statistical significance.

Furthermore, the results of the ITC suggested that across both trials, treatment with semaglutide 2.4 mg and liraglutide 3.0 mg resulted in similar changes from baseline in HDL and total cholesterol. For the binary endpoint of reversion of T2D status to normo-glycaemic status, semaglutide 2.4 mg was associated with significantly higher odds of achieving normo-glycaemic status compared with liraglutide 3.0 mg. For all but one endpoint, the results were consistent across all analyses conducted. The only exception was the normo-glycaemia analysis, which was no longer statistically significant after adjusting for differences in trial populations.[97] This was driven by a slightly lower baseline HbA1c in SCALE 1839 (5.8%) versus STEP 1 (5.9%); the closer a population is to being normo-glycaemic (i.e. HbA1c < 5.7%), the lower the incremental glycaemic effect of adding a more potent GLP-1 receptor agonist.

Table 12: Base-case results of the indirect comparisons of semaglutide 2.4 mg (STEP 1) versus liraglutide 3.0 mg (SCALE 1839)

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

62 of 176

B.2.9.3 Uncertainties from the indirect comparisons

A general uncertainty in anchored indirect treatment comparisons is the potential for bias due to effect modifiers that are in imbalance between trial populations. Since STEP 1 and 1839 were conducted in very similar populations, the potential for imbalance of effect modifiers was low, and unadjusted comparisons were considered as the primary indirect treatment comparisons. However, these were supplemented with population-adjusted analyses, with effect modifiers elicited based on published evidence. As a sensitivity analysis, the population-adjusted analyses were repeated to include further adjustment variables into the model to verify that populationadjusted results were robust to the choice of putative effect modifiers.

Another general uncertainty is related to the handling of intercurrent events of treatment discontinuation and the use of rescue medication in the statistical analysis. To compensate for this, a trial product estimand strategy was applied as an alternative strategy to the primary treatment policy estimand strategy, the results of which were broadly consistent with the base case.

A specific uncertainty in the indirect treatment comparison of STEP 1 and 1839 is the different study duration. This was assessed by means of sensitivity analyses in regard to evaluation time points for the endpoints studied, which again were broadly consistent with the base case.

B.2.9.4 Conclusions from the indirect comparisons

In a population of patients with BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk, when considering the effect of approximately one year of treatment based on the treatment policy estimand, the results of the ITC suggested that semaglutide 2.4 mg was associated with a statistically significant reduction in body Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

63 of 176

weight, HbA1c and waist circumference compared with liraglutide 3.0 mg.[97] For the binary endpoint of reversion of T2D status to normo-glycaemic status, semaglutide 2.4 mg was associated with higher odds of achieving normo-glycaemic status compared with liraglutide 3.0 mg. However, the results of the normo-glycaemia analysis were no longer statistically significant after adjusting for differences in trial populations (see Appendix D4). This was driven by the minor population difference in HbA1c of 0.1%-points, because the glycaemic effect of adding a more potent GLP-1 receptor agonists diminishes in a population that is predominantly normo-glycemic.[97]

Across all continuous outcomes, the results of the ITCs which adjusted for potential treatment effect modifiers (population adjustment 1 & 2, Appendix D4) were consistent with the unadjusted analyses.[97] Similarly, the results of the analyses conducted in the pre-diabetic population, using alternative time points, and considering the trial product estimands were consistent with the unadjusted basecase analysis across all outcomes (Appendix D4).[97]

B.2.10. Adverse reactions

This section presents the results of the STEP 1 safety evaluation, which was based on the safety analysis set (see Section B.2.4.1 for a definition). The on-treatment period is used for most evaluations as it represents the period when patients were considered exposed to treatment. AEs with onset during the on-treatment period correspond to treatment-emergent AEs (TEAEs). For deaths and event types with potential long latency to diagnosis, the in-trial period is used (see Section B.2.4.3 for observation period definitions).

Given the 2:1 randomisation for semaglutide 2.4 mg versus placebo, the primary focus of the safety evaluation of AEs and adjudicated events is the proportion of patients with events and event rates to ensure a valid treatment comparison. Whenever the number of patients with events or number of events are looked at, and when evaluating treatment differences for infrequent/rare events, the 2:1 randomisation must be kept in mind.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

64 of 176

B.2.10.1 Summary of adverse events

A summary of adverse events that occurred during STEP 1 is provided in Appendix F1. Overall, a similar proportion of patients in the semaglutide 2.4 mg and placebo treatment arms reported AEs (89.7% and 86.4%, respectively).[83]

B.2.10.1.1 Most common adverse events

Table 13 presents the most common AEs reported in ≥ 10% of patients. In the semaglutide 2.4 mg treatment arm, the most common AEs were gastrointestinal disorders, particularly nausea (44.2% of patients), diarrhoea (31.5% of patients), vomiting (24.8% of patients), and constipation (23.4% of patients).[83] In the placebo treatment arm, the most common AEs were nasopharyngitis (20.3% of patients), nausea (17.4% of patients), diarrhoea (15.9% of patients), headache and upper respiratory tract infection (both experienced by 12.2% of patients).[83]

Table 13: Most common adverse events reported in ≥ 10% of patients

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

65 of 176

B.2.10.2 Adverse events of particular interest

Based on therapeutic experience with GLP-1 receptor agonists and in line with regulatory feedback and requirements, a number of safety focus areas were of special interest in the safety evaluation, as presented in Appendix F2.[83]

B.2.10.2.1 Gastrointestinal disorders

Gastrointestinal disorders (typically nausea, diarrhoea, vomiting, and constipation; Table 13) were the most frequently reported AEs and occurred in more patients receiving semaglutide 2.4 mg than those receiving placebo (74.2% versus. 47.9%; Table 13).

Figure 10 depicts the prevalence and duration of gastrointestinal events by severity. Overall, most gastrointestinal AEs were mild or moderate in severity, were transient, and resolved without permanent discontinuation of the regimen.[83]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

66 of 176

Figure 10: Prevalence and duration of gastrointestinal events

==> picture [685 x 285] intentionally omitted <==

Notes: Figure presents the proportion of patients reporting nausea (A), diarrhoea (B), vomiting (C), or constipation (D). Events classed as mild, moderate, or severe, over the course of the treatment period and the median duration of the event. Data are on treatment observation period data (during treatment with trial product [any dose of trial medication administered within the previous 49 days, i.e. any period of temporary treatment interruption with trial product was excluded]). Adverse events were classified by severity as mild (green), moderate (orange) or severe (red). Source: Wilding et al. 2021.[83]

B.2.10.3 Study withdrawal

Refer to Appendix F3 for a summary of study withdrawals that occurred during STEP 1.

B.2.10.4 Adverse events leading to discontinuation of study treatment, dose interruptions or dose adjustments

Refer to Appendix F4 for a summary of AEs leading to discontinuation of study treatment, dose interruptions or dose adjustments during STEP 1.

B.2.10.5 Use of rescue therapy Refer to Appendix F5 for a summary of rescue therapy used during STEP 1.

B.2.10.6 Safety summary

Overall, semaglutide 2.4 mg administered once weekly as an adjunct to a reducedcalorie diet and increased physical activity was well tolerated in patients who are overweight or obese, and the majority of reported AEs were mild or moderate in severity. Furthermore, the safety and tolerability profile of semaglutide 2.4 mg was consistent with previous studies of semaglutide as well as with that reported for the GLP-1 receptor agonist class in general; no new safety concerns were identified (further discussed in Section B.2.13).[83, 99, 100]

As is typical of the GLP-1 receptor agonist drug class[99] , transient, mild or moderate gastrointestinal disorders were the most frequently reported AEs with semaglutide 2.4 mg.[83] Nausea was the most common gastrointestinal event, occurring primarily during the dose-escalation period, a finding similar to that reported with liraglutide at a dose of 3.0 mg.[101] Gallbladder-related disorders, principally cholelithiasis, were also more common with semaglutide 2.4 than with placebo. Gallbladder-related disorders have been previously reported with GLP-1 receptor agonists and are consistent with the known effects of rapid weight loss.[83]

B.2.11. Ongoing studies

B.2.11.1 STEP programme

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

68 of 176

Table 14 presents an overview of the STEP clinical trial programme.

Table 14: Summary of STEP trials and rationale for their inclusion in the submission

B.2.12. Innovation

Over recent years, despite pharmacological advancements for managing patients with obesity, the clinical unmet need in a broad population remains.[23] Previously, patients with obesity who have comorbidities have been unable to receive effective pharmacological intervention within SWMS. Liraglutide 3.0 mg, recently approved by NICE, is restricted for use in a narrow subset of patients with obesity (patients with a BMI ≥ 35 kg/m[2] plus non-diabetic hyperglycaemia plus high CV-risk). However, semaglutide 2.4 mg provides a pharmacological treatment option within SWMS for a broader group of patients with obesity, that could benefit from significantly greater weight loss than previously achieved with other pharmacological interventions.

The weight loss observed with semaglutide 2.4 mg as part of the STEP 1 study (a mean of 15% over 68 weeks) is a significant advancement of more than double the weight loss reported with existing pharmacotherapy options for obesity, which typically provide weight loss of between 5–7%.[23] Furthermore, clinicians have highlighted that there are various weight-related comorbidities impacting patients with obesity that are best placed to benefit from weight loss of 10–15%.[23] As discussed in Section B.1.3.3.1, the benefits of 10–15% weight loss have been demonstrated across a range of weight-related comorbidities, including but not limited to T2D, cardiovascular disease, osteoarthritis, gastroesophageal reflux disease, obstructive sleep apnoea, NASH, hypertension, dyslipidaemia and hyperglycaemia.[17-21, 54] As such, semaglutide 2.4 mg provides a step change in treatment for obesity.

In addition to the direct impact semaglutide 2.4 mg can have in alleviating the current burden of comorbidities in patients with obesity, it can also have additional downstream benefits to patients and wider society by reducing the risk of patients developing future weight-related comorbidities.[22] Weight loss with semaglutide 2.4 mg may also give patients their independence back by allowing them to participate in daily activities, sports and hobbies, or by returning to work.[102] Use of this treatment may also reduce the burden of obesity on the healthcare system through alleviation and prevention of weight-related comorbidities, reducing weight-related hospitalisations and mortality.

B.2.13. Interpretation of clinical effectiveness and safety evidence

In the pivotal STEP 1 study, semaglutide 2.4 mg as an adjunct to lifestyle intervention (diet and exercise) demonstrated considerable benefit over lifestyle intervention alone (placebo arm). On average, patients in the FAS receiving semaglutide 2.4 mg experienced six-times more weight loss compared with patients receiving placebo (−14.9% versus −2.4%, respectively). Moreover, a significantly greater proportion of patients treated with semaglutide 2.4 mg experienced clinically meaningful weight loss of ≥ 5% (86.4% versus 31.5%; p < 0.001), ≥ 10% (69.1% versus 12.0%; p < 0.001) and ≥ 15% (50.5% versus 4.9%; p < 0.001) than with placebo.[83] Greater reductions in waist circumference, systolic blood pressure and HbA1C were also observed with semaglutide 2.4 mg than with placebo.[83] Furthermore, of the patients with non-diabetic hyperglycaemia (i.e. pre-diabetes) at baseline, more than double the proportion of those treated with semaglutide 2.4 mg became normo-glycaemic at 68 weeks compared with those treated with placebo.[95]

The results of the efficacy analyses in the target population (patients with a BMI ≥ 30 mg/kg[2] with at least one weight-related comorbidity) were consistent with those of the FAS, which is unsurprising given the similarities between the two patient populations. In addition, semaglutide 2.4 mg also demonstrated effectiveness in a more severe population of patients with obesity with a BMI ≥ 35 mg/kg[2] with nondiabetic hyperglycaemia and high CV risk (i.e. the liraglutide 3.0 mg eligible population).

Importantly, the clinical benefit obtained with semaglutide 2.4 mg was not achieved at the risk of patient wellbeing. Treatment with semaglutide 2.4 mg resulted in significant and clinically meaningful improvements in HRQoL in patients with obesity. After 68 weeks, semaglutide 2.4 mg was associated with significant improvements in patients’ physical functioning, as measured using IWQOL-Lite-CT and SF-36.[83] Regarding safety, semaglutide 2.4 mg was well tolerated and demonstrated a safety profile consistent with that previously reported for semaglutide, and with the GLP-1 receptor agonist class in general, and no new safety concerns were identified.[83, 99, ] 100 This finding mirrors that of the SUSTAIN 6 trial, a large Phase IIIa pre-approval Cardiovascular Outcomes Trial.[103] SUSTAIN 6 investigated the effects of subcutaneous semaglutide (0.5 mg and 1.0 mg) versus placebo on Major Adverse

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

76 of 176

Cardiovascular Events (MACE) in patients with T2D and high CV-risk. The trial was designed and conducted in accordance with the US 2008 FDA guidance for evaluating cardiovascular safety in new glucose-lowering therapies. Overall, there was a 26% reduction in the risk of MACE with semaglutide 0.5 mg and semaglutide 1.0 mg relative to placebo (note: these doses are different to that used in obesity [2.4 mg]). In line with STEP 1, the safety profile of semaglutide was overall consistent with that of the GLP-1 RA class, with gastrointestinal adverse events being the most frequently reported adverse drug reactions with semaglutide.[103] It should also be noted that during STEP 1, more patients discontinued treatment due to a lack of efficacy in the placebo arm than in the semaglutide 2.4 mg arm[85] , highlighting a desire from patients to be treated with pharmacological intervention.

The results of the indirect treatment comparisons show that in patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk, semaglutide 2.4 mg was significantly more effective in reducing body weight, HbA1c and waist circumference compared with liraglutide 3.0 mg.[97] The results also show that patients with non-diabetic hyperglycaemia would benefit greatly from the introduction of semaglutide 2.4 mg.[97] This sentiment is reflected by clinicians who, if approved, would prioritise treatment with semaglutide 2.4 mg ahead of treatment with liraglutide 3.0 mg.[23]

Overall, semaglutide 2.4 mg offers a safe and effective treatment option for patients with a BMI ≥ 30 mg/kg[2] with at least one weight-related comorbidity, offering the potential for weight loss previously unobtainable in SWMS with current pharmacotherapy treatment options.

B.2.13.1 Strengths and limitations of the evidence base

The STEP 1 study was a high-quality trial that enrolled a large and broad population of patients with obesity. STEP 1 included nine clinical trial sites in England and clinicians also considered the patient demographics, disease characteristics and comorbidities to be highly reflective of those patients managed in SWMS in UK clinical practice.[23] Of all the STEP trials, the patient population of STEP 1 was the most reflective of the target population, and therefore STEP 1 was considered the most appropriate primary evidence source for this submission, a view shared by clinicians.[23]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

77 of 176

The target population for this submission (patients with a BMI ≥ 30 mg/kg[2] with at least one weight-related comorbidity) was analysed post-hoc and was narrower than the FAS population of the STEP 1 trial. However, the target population represented 75.0% of the FAS and the effiacy results were consistent between the two patient populations. Similarly, although a post-hoc subgroup from STEP 1 was used for the ITC analysis (patients with a BMI ≥ 35 kg/m[2] with non-diabetic hyperglycaemia and high CV risk), this population was reflective of the liraglutide 3.0 mg-eligible population and was required for a comparison between semaglutide 2.4 mg and liraglutide 3.0 mg to be made.

B.3. Cost effectiveness

B.3.1. Published cost-effectiveness studies

An SLR of cost-effectiveness studies in obesity was conducted in October 2018 with a further update conducted in April 2021. Systematic searches for costeffectiveness/cost-utility, costs and healthcare resource studies were carried out simultaneously as one combined search to identify all relevant studies on adult patients with obesity. Appendix G provides details of the SLR. In summary, a total of seven published cost-effectiveness analyses which reported results from the UK NHS perspective were identified and reviewed (Table 15). None of the studies identified in the review precisely met the definition of the target population for the present submission, being patients with obesity with a BMI of ≥ 30 mg/kg2 in the presence of at least one weight-related comorbidity (see Section B.1.1 for the full decision problem), nor were any subgroup analyses in this population published. Further, orlistat and bariatric surgery, which were identified as treatments in a number of these studies, were not included as the comparators in this submission.

Of the reviewed studies, one was a piggy-back cost-effectiveness analysis conducted alongside a clinical trial[104] ; all other published cost-effectiveness analyses involved some degree of modelling to estimate treatment effect on costs and health outcomes. Cohort, state-transition modelling was used in three of the published costeffectiveness studies[105-107] , two studies used simple decision analyses based on a 1- year and 5-year time horizon,[108, 109] and the final study used a microsimulation state transition model (Monte Carlo simulation) with a 30-year time horizon with computer generated individuals.[110] With the exception of one study published in 2005,[109] and the piggy-back trial analysis[104] – both conducted on a 1-year time-frame analysis – all of the reviewed cost-effectiveness analyses modelled long-term consequences of T2D and CVD in patients with obesity. The association between obesity, onset of T2D and CV disease risk are thus well established in the health-economic literature herein reviewed.[111] One study modelled the association of BMI with colon cancer.[107] The methods used to extrapolate short term changes in BMI to onset of complications differed across studies. For example, Ara et al. 2007[108, 112] used the Framingham risk model to calculate incidence of CVD as a function of BMI. Later,

the same authors published a de novo analysis informed by a set of newly Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

79 of 176

developed risk models;[106] the models were developed on a random sample of adults with data from the General Practice Research Datalink (GPRD) (n = 100,000). Risk models were developed for: onset of type 2 diabetes, incidence of acute myocardial infarction (AMI), stroke and death from any cause as function of BMI, age, gender, smoking status, aspirin, insulin, statin and blood pressure treatment, for T2D and non-T2D cohorts. A natural disease progression model of the BMI trajectory was also estimated by the authors[106] based on the GPRD data.

Table 15: Summary list of published cost-effectiveness studies

B.3.2. Economic analysis

The cost-effectiveness model used for this submission was designed to evaluate the treatment of obesity and to reflect the disease’s natural history, expected prognostic pathways in the absence of intervention, and treatment effects, is adapted from the model used for the previous NICE appraisal for liraglutide 3.0 mg for managing overweight and obesity (TA664), and uses committee preferred assumptions from that appraisal.[27] The model has undergone numerous improvements and additions during which it was validated against real world data (Section B.3.10). It is a statetransition cohort model that uses risk equations for extrapolation and captures the benefit of weight loss on the key weight-related comorbidities. It is consistent with other models that have been used for obesity and diabetes modelling.[106, 107, 122]

Key results of the economic analysis ICERs with and without PAS are presented. Where PAS ICERs are presented, these are based on the proposed PAS price. Semaglutide 2.4mg vs. diet and exercise: • Patient population: BMI ≥ 30 kg/m[2] patients with one or more obesity related comorbidities o Deterministic ICER: ''''''''''''''''''''' per QALY (without PAS) o Probabilistic ICER'' '''''''''''''''''''' per QALY (without PAS) o Deterministic ICER: £14,827 per QALY (with PAS) o Probabilistic ICER: £14,733 per QALY (with PAS) Semaglutide 2.4mg vs. liraglutide 3.0mg • Patient population: BMI ≥ 35 kg/m[2] , prediabetes and high risk for CVD o Deterministic ICER: ''''''''''''''''''''' (without PAS) o Probabilistic ICER: '''''''''''''''''''''' (without PAS) o Deterministic ICER: Dominant (with PAS) o Probabilistic ICER: Dominant (with PAS)

Key: BMI, Body mass index; CVD, Cardiovascular disease; ICER, Incremental cost-effectiveness ratio; QALY, Quality-adjusted life year.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

85 of 176

B.3.2.1 Patient population

Two populations are considered, as discussed in more detail in Sections B.1.1 and B.2.7:

1. BMI ≥ 30 kg/m[2] patients with one or more obesity related comorbidities (base case)

• BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD

  • (subpopulation)

The subpopulation with BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD is considered in order to align with the recommended target population for the pharmacotherapy, liraglutide 3.0 mg which was recommended by NICE in TA664.[27]

The characteristics of the starting cohort in terms of baseline demographic parameters are defined at model entry. The model structure described in detail in Section B.3.2.2, is a cohort model and so average patient characteristics are used as model inputs. These were sourced from a post-hoc analysis of the corresponding subset population in the STEP 1 clinical trial. The baseline characteristics for both populations of interest are shown in Table 16.

Table 16: Baseline characteristics for populations of interest

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

86 of 176

B.3.2.2 Model structure

A Markov cohort model was developed in Microsoft Excel[®] , based on the model previously used in TA664 for liraglutide 3.0 mg in obesity.[27] The model is a statetransition cohort model that uses risk equations for extrapolation.

B.3.2.2.1 Model heath states

The model includes 11 health states, shown below in Figure 11, and is an updated and simplified version of the model used for TA664.[27] Obesity is associated with numerous possible comorbidities as evidenced by the WHO consultation on obesity, which noted the complications that respond to weight loss and have substantial consequences on healthcare resources and costs, patients’ quality of life, and/or life expectancy.[111]

Comorbidities included in the model were those with strong evidence of association with obesity (non-diabetic hyperglycaemia, T2D, OSA, acute coronary syndrome [ACS] and stroke), and osteoarthritis. Cancer health states had limited impact on the cost-effectiveness model used in TA664 for liraglutide 3.0 mg and so were not included.

Given that neither all weight related comorbidities from STEP 1 nor all weight-related comorbidities observed in real life are included, the model is expected to be conservative and therefore underestimating the benefit of treatment with semaglutide 2.4 mg.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

87 of 176

The majority of the cohort enters the model in the normal glucose tolerance (NGT) or non-diabetic hyperglycaemia health states. In addition, patients in an arm with pharmacotherapy may be ‘on’ or ‘off’ pharmacotherapy. Patients who have discontinued pharmacotherapy treatment still receive diet and exercise. This is tracked by tunnel states that capture time since pharmacotherapy was stopped (to model catch up of surrogate outcomes with control values). Patients in the CV event states are also divided between those with and without diabetes.

The analysed cohort is distributed across the starting health states based on the proportion with non-diabetic hyperglycaemia and proportion with NGT. For each patient population this is based on the observed prevalence in the STEP 1 clinical study.[90] The percentage of patients with history of CVD was '''''' and was used to inform the number of patients entering the model in a post-CVD health state.

''''''’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’’

’’’’’’’’’’’’’’’’’’’’’ .

The prevalence for obesity-related comorbidities for both populations of interest are shown in Table 17.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

88 of 176

Figure 11: Obesity model structure

==> picture [430 x 518] intentionally omitted <==

Key: ACS, Acute coronary syndrome; BMI, body mass index; HDL, high density lipoprotein; OSA, obstructive sleep apnoea; SBP, systolic blood pressure; T2D, type 2 diabetes.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

89 of 176

Table 17: Starting non-diabetic hyperglycaemia distribution for both populations of interest

To limit the number of health states in the model, the following assumptions, which still allow the main costs and benefits of treatment to be captured, were made:

• Non-diabetic hyperglycaemia patients move to T2D + post-ACS or T2D + poststroke states following an ACS or stroke event respectively. This assumption has previously been used in TA664 where clinical experts explained that people are more likely to be diagnosed with T2D after a cardiovascular event.[27] This assumption is tested in a scenario analysis

• T2D micro-vascular complications are not included as distinct health states. Rather, for a proportion of patients residing in the T2D health state, higher costs apply, reflective of possible micro-vascular complications. This is a conservative assumption as the reduction in micro-vascular complications within T2D is not captured directly from treatment

• Osteoarthritis is not accounted for as a separate health state as this would have tripled the number of health states considered. It is considered in the model in the form of knee replacement event rather than a separate health state. Nonetheless, it is expected that the underestimation from a cost perspective would be low, given that managing osteoarthritis does not involve high medical expenditure (e.g. management with analgesics)

• OSA is not accounted for as a separate health state as this would also have substantially increased the number of health states considered. The costs and quality of life decrements are accounted for by estimated prevalence of OSA each cycle from using a risk equation

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

90 of 176

• Once individuals develop a condition which bears substantial higher costs, quality of life and/or survival implications, the less ‘serious’ condition is superseded by the more ‘serious’ one. For example, if 50% of the cohort are in the temporary nondiabetic hyperglycaemia reversal health state and then experiences non-fatal myocardial infarction (MI) at a rate of 0.3% per year, in the first year, 0.15% of the cohort will then move into the post-ACS state and leave the temporary nondiabetic hyperglycaemia health state

• A small proportion of patients (1.8%) have T2D at baseline. This is due to some of the patients developing T2D between screening and Week 0. As these patients are not in the target population for the submission, they are accounted for in the model scaling up the NGT and prediabetes health states to 100%.

B.3.2.2.2 Transitions and transition probabilities

During every model cycle, the cohort moves between health states or may remain in the same health state. The likelihood of a transition occurring is given by a transition probability. Transition probabilities are derived based on the risk of developing each of the comorbidities associated with obesity included in the analysis (e.g. T2D) or the likelihood of an event occurring (e.g. non-fatal MI is followed by transition of the cohort to post-ACS health states).

In this model, the transition to T2D, ACS, stroke, OSA, osteoarthritis (knee replacement) or death are predicted based on short-term effects of interventions on surrogate outcomes – BMI, SBP, total cholesterol, HDL cholesterol and HbA1c which are typical endpoints in obesity clinical trials and extrapolated over a given time horizon. Effects on surrogate outcomes are translated into lifetime risks of obesity complications through risk-prediction equations or lookup risk tables obtained from published studies.

Demographic and cardio-metabolic risk factor parameters are defined as either static i.e. do not change over time (e.g. sex, height, smoking status, use of lipid-lowering medication) or as dynamic i.e. change over the time horizon of the model. For simplification, some variables which can be dynamic in real-life have been assumed static in this model (e.g. smoking status). Such variables are not believed to be

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

91 of 176

affected by the interventions considered, and therefore, this simplifying assumption is not expected to affect the outcomes of the cost-effectiveness analyses conducted.

Dynamic variables can change in the model because of

• Time and/or natural progression of the disease (e.g. age);

• Weight reduction interventions (e.g. SBP, cholesterol levels); or

• Both of the above (e.g. BMI, HbA1c in T2D)

Table 18 exhibits the model parameters characterising the starting cohort, and their assumed nature (static or dynamic). A brief description of their behaviour throughout the time horizon of the model is also provided.

Table 18: Definition of baseline cohort characteristics

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

92 of 176

In the model, the characteristics of the cohort differ between states. For example, if a therapy impacts BMI, patients in a treated health state will have a different BMI from untreated patients. Transitions depend on cohort characteristics such that if a therapy modifies cardiovascular risk factors, then the transition to a post-CVD event health state is affected by the change in risk of serious cardiovascular disease. State payoffs may also depend on events which do not influence the likelihood of transitioning to a different health state and are included as one-off costs and decrements in the estimated total quality of life.

To reflect the higher likelihood of some obesity co-morbidities occurring simultaneously and to consider the possible lifelong implications of such conditions despite the memory-less feature of Markov cohort models, combinations of comorbidities have been defined as separate health states. More specifically, if for example, a proportion of the cohort has T2D and goes on to experience an MI event, the respective cohort moves into a health state post-ACS + T2D in the next cycle. If a proportion of this cohort goes on to experience a stroke, this proportion will move

93 of 176

into a post-ACS + post-stroke + T2D health state. This approach allows the model to reflect the chronic nature of most obesity-related conditions.

A limitation of this model is that non-diabetic hyperglycaemia can only be included as a comorbidity at baseline and the cohort with NGT cannot develop non-diabetic hyperglycaemia over the time horizon of the analysis. An SLR was conducted to obtain risk equations for obesity related comorbidities.[123] There are risk equations on development of T2D from any non-diabetic health state, but not for development of non-diabetic hyperglycaemia from NGT. To overcome this limitation, the model allows the user to assign a higher risk of developing T2D to the proportion of the cohort that had non-diabetic hyperglycaemia at baseline and who reverted to NGT during treatment than for those with NGT at baseline. As such, the model includes a temporary non-diabetic hyperglycaemia reversal health state. This adjusts for the fact that reversal of non-diabetic hyperglycaemia to NGT is a temporary situation and after treatment, the cohort would return to have the same risk of T2D as observed in non-diabetic hyperglycaemia, which is higher compared to NGT patients. Therefore, transitions to the temporary non-diabetic hyperglycaemia reversal health state may occur only from non-diabetic hyperglycaemia. This transition is based on the percentage reduction in prevalence observed in STEP 1, applied at the end of the first cycle in the model (3 months from the beginning of the model).

Transition to a post-ACS state occurs following a non-fatal MI or non-fatal unstable angina. Transition to post-stroke occurs following a non-fatal stroke or transient ischemic attack (TIA). The cohort residing in post-ACS can experience further MI or unstable angina events in the next cycles but remains in the same post-ACS state or experiences a cerebrovascular event and transitions to a post-ACS + post-stroke health state. Similarly, the cohort residing in post-stroke can experience a further stroke or TIA event in the next cycles but remains in the same post-stroke state or experiences a cardiovascular event and transitions to a post-ACS + post-stroke health state. Transition to death can occur from any of the model health states either as a fatal event occurs or based on disease specific and general population mortality.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

94 of 176

As a general rule, the transitions allowed, and the respective transition probabilities, depend on the health state in which the cohort resides at the end of the previous cycle. That is:

• Higher probability of developing T2D applies to patients residing in a non-diabetic hyperglycaemia health state than for patients with NGT

• Higher rates of MI, unstable angina, stroke or TIA apply to patients residing in the T2D state than patients without T2D

• Higher rates of MI, unstable angina, stroke or TIA apply to patients residing in post-ACS or post-MI health states than patients who did not have a previous CV event

• Higher rates of MI, unstable angina, stroke or TIA apply to patients residing in post-ACS+T2D or post-MI+T2D health states, than for patients in post-ACS or post-MI health states with NGT or in the T2D health state without previous CV event

• Higher rates of mortality apply to T2D, post-ACS, post-stroke health states compared with health states with NGT or no complications

B.3.2.2.3 Time horizon

The base case uses a time horizon of 40 years to capture all costs and consequences of semaglutide 2.4 mg and the comparator. At the end of the model time horizon, a majority of patients are deceased with a difference of less than 0.1% of patients alive between treatment arms. Therefore, any subsequent differences beyond the modelled time horizon are expected to be minimal and discounted such that the model approximates lifetime costs and benefits of the intervention in obesity.

B.3.2.2.4 Cycle length

The cycle length of the model is the time interval elapsing from one transition to another. The cycle length was defined considering the condition analysed and the likely frequency of changes in patients’ health status.

For the current decision problem, a cycle length of 3 months was defined for the first year with annual cycles thereafter. Shorter cycles at the start of the model allow a more accurate representation of the treatment effects, dosing schedule,

incorporation of treatment stopping rules and changes in disease status after therapy Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

95 of 176

initiation. Although shorter cycles will always yield more precise estimates, the error becomes very small when the number of cycles increases with increasing time horizon. Thus, for computational efficiency, annual cycles were defined after the first year. Half-cycle correction is applied.

Key features of the economic analysis along with a comparison of the approach to TA664, are summarised in Table 19.

Table 19: Features of the economic analysis

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

96 of 176

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

97 of 176

B.3.2.3 Intervention technology and comparators

The model was developed to allow the comparison of the intervention, semaglutide 2.4 mg in combination with diet and exercise hereby referred to as ‘semaglutide 2.4 mg’, versus relevant comparators in each population of interest.

B.3.2.3.1 Semaglutide 2.4 mg (Intervention)

Semaglutide 2.4 mg is a once weekly pharmacotherapy which involves a titration phase and a maintenance phase. The titration phase starts at treatment initiation and lasts for 16 weeks during which time the dose is increased every 4 weeks. The applied dosage regime is:

• Week 1–4: 0.25 mg/week

• Week 5–8: 0.5 mg/week

• Week 9–12: 1.0 mg/week

• Week 13–16: 1.7 mg/week

• Week 17 onwards: 2.4 mg/week

Semaglutide 2.4 mg maintenance dose is 2.4 mg subcutaneous (s.c.) injection per week. At the time of writing, it is unclear whether the marketing authorisation for semaglutide 2.4 mg will include a stopping rule where patients not achieving a defined weight loss discontinue treatment. In the model it is assumed that that

treatment should be discontinued after 28 weeks if patients have not lost < 5% of their initial body weight, in anticipation of this expected stopping rule in the marketing authorisation (see Section B.1.1).

B.3.2.3.2 Comparators

Diet and exercise

For the population with BMI ≥ 30 kg/m[2] with one or more obesity related comorbidities, no pharmacotherapy is provided in SWMS and thus the relevant comparator to semaglutide 2.4 mg in combination with diet and exercise, is diet and exercise alone.

Diet and exercise refers to the standard management in obesity which includes a reduced calorie diet and increased physical activity.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

98 of 176

Liraglutide 3.0 mg

For the population with BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD, liraglutide 3.0 mg in combination with diet and exercise, hereby referred to as ‘liraglutide 3.0 mg’ is recommended by NICE in TA664. Therefore, this treatment is a relevant comparator to semaglutide 2.4 mg in combination with diet and exercise only in this specific subpopulation.

Liraglutide 3.0 mg (Saxenda, manufactured by Novo Nordisk) is a GLP-1 receptor agonist similar to semaglutide 2.4 mg. Liraglutide 3.0 mg is approved as an adjunct to a reduced-calorie diet and increased physical activity for weight management in adult patients (≥ 18 years) with an initial BMI of:

• ≥ 30 kg/m² (obese), or

• ≥27 kg/m² to <30 kg/m² in the presence of at least one weight-related comorbidity such as dysglycaemia, hypertension, dyslipidaemia, or OSA

Liraglutide 3.0 mg treatment involves a titration phase and a maintenance phase. The titration phase starts at treatment initiation and lasts for 4 weeks during which time the dose is increased each week. The applied dosage regime is:

• Week 1: 0.6mg/day

• Week 2: 1.2mg/day

• Week 3: 1.8mg/day

• Week 4: 2.4 mg/day

• Week 5 onwards: 3.0 mg/day

Liraglutide 3.0 mg maintenance dose is 3 mg s.c. injection per day. In line with the marketing authorisation treatment should be discontinued after 12 weeks on the 3.0 mg daily maintenance dose if patients have not lost at least 5% of their initial body weight (referred to as a stopping rule).[143]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

99 of 176

B.3.3. Clinical parameters and variables

B.3.3.1 Treatment effects

The clinical effectiveness of the interventions considered is introduced in the model through changes in BMI and cardio-metabolic risk factors, namely SBP, HDL cholesterol, total cholesterol and HbA1c (in diabetes). These intermediate endpoints are used in risk equations/risk tables to calculate transition probabilities in the model, guiding the progression of the cohort throughout the time horizon of the analysis.

In addition, a temporary reversal of non-diabetic hyperglycaemia in the model is allowed by treatment arm. Finally, adverse events for each treatment are also incorporated.

B.3.3.1.1 Effect on surrogate outcomes

The relative reduction in BMI is the main driver of clinical effectiveness. Changes in BMI influence the risk of all obesity complications in the model (except the incidence of secondary cardiovascular events). Changes in SBP, HDL cholesterol and total cholesterol influence the risk of T2D, CVD in primary prevention and the risk of CVD in secondary prevention. Changes in HbA1c influence the risk of CVD in primary prevention in the cohort with T2D, however, average HbA1c is assumed to not increase over time. Table 20 describes how the effect on surrogate outcomes, measured through changes in physiological parameters, is quantified in the model.

Table 20: Definition of treatment effects on physiological parameters included in the model

The initial effect of treatment on the physiological parameters is applied at model start and relative to the baseline values. It should be noted however that the incidence of complications takes over only from Cycle 2 of the model and as such,

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

100 of 176

the initial weight reduction will start influencing the incidence of complications at the beginning of Cycle 2 of the model (starting from Month 4).

Treatment effects for semaglutide 2.4 mg and diet and exercise for both subgroups were sourced from the population containing all patients of the STEP 1 clinical trial. This population efficacy was assumed to be representative of the subgroups of interest and since this population was defined a priori in the STEP 1 trial, it provided a statistically robust measure of the treatment effect. This assumption was validated in an advisory board with key opinion leaders (KOLs)[23] and the post hoc analysis is explored in a scenario analysis for validation.

The trial product estimand from STEP 1 was used to reflect the efficacy of patients who stay on treatment. This estimand, which estimated the treatment effect of semaglutide 2.4 mg relative to placebo for all randomised patients, assuming they remained on their randomised treatment for the entire planned duration of the trial and had not initiated other anti-obesity therapies, as discussed in more detail in Section B.2.4.4, was considered most appropriate for the modelling of efficacy in the economic model. This estimand in combination with the stopping rule for nonresponders provides a clear representation of patients who would receive treatment in a SWMS setting. As described in Section B.3.2.3.1 the stopping rule for semaglutide 2.4 mg is 28 weeks so the full analysis set efficacy is used for Week 20 and Week 28 STEP 1 data. After Cycle 2, early responder efficacy for STEP 1 Week 68 data is used for patients continuing treatment with semaglutide 2.4 mg. The efficacy inputs from the population containing all patients in STEP 1 for semaglutide 2.4 mg and diet and exercise, as used in the base case, are shown in Table 21. Week 28 data for weight change and SBP change and Week 20 data for total cholesterol and HDL cholesterol were used to inform model cycles up to 9 months. Week 68 data informed on treatment model cycles after 9 months.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

101 of 176

Table 21: Change in physiological parameter values – STEP 1 – all patients

Figure 12 illustrates the progression of BMI over time for semaglutide 2.4 mg and diet and exercise, including the additional effect of bariatric surgery, for patients with BMI ≥ 30 kg/m[2] with one or more obesity related comorbidities. Semaglutide 2.4 mg shows a greater reduction in BMI compared to diet and exercise. After age 57 the cohort no longer receives bariatric surgery and the BMI of the cohort increases at a greater rate compared to after the catch up period until age 68 where the cohort no Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

102 of 176

longer gains a natural increase in BMI (assumptions discussed in more detail in Sections B.3.3.2 and B.3.3.3 respectively).

Figure 12: Projected BMI change for semaglutide 2.4 mg and diet and exercise

over time for patients with BMI ≥ 30 kg/m[2] with one or more obesity related comorbidities

==> picture [448 x 255] intentionally omitted <==

Key : BMI, body mass index.

B.3.3.1.2 Reversal of non-diabetic hyperglycaemia to normal glucose tolerance

An immediate effect of interventions in terms of temporary reduction in non-diabetic hyperglycaemia prevalence is applied in the model such that patients temporarily have normal glucose levels.

Non-diabetic hyperglycaemia reversal is applied in the model through a proportion of the cohort with non-diabetic hyperglycaemia transitioning from the non-diabetic hyperglycaemia health state to the temporary non-diabetic hyperglycaemia reversal health state. Patients in the temporary non-diabetic hyperglycaemia reversal health state have normal glucose levels. Following obesity treatment discontinuation, patients return to their pre-treatment glycaemic status as defined in Section B.3.2.1 and therefore transition back to the non-diabetic hyperglycaemia health state.

103 of 176

As in the model used for TA664, non-diabetic hyperglycaemia reversal is applied at Month 3 of the analysis, since it is expected to occur immediately after treatment start, even though observation on glycaemic status from a given trial is available only at 1 year. The parameter is sourced from Week 52 glycaemic status results for the population containing all patients in STEP1 for semaglutide 2.4 mg and diet and exercise. Section B.3.3.1.3 provides detail for the liraglutide 3.0 mg estimate in the BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD population from SCALE 1839. Table 22 shows the percentage of non-diabetic hyperglycaemia patients with glycaemic status reversal for semaglutide 2.4 mg and diet and exercise.

Table 22: Percentage of non-diabetic hyperglycaemia patients with glycaemic

status reversal – STEP 1 – all patients

B.3.3.1.3 Indirect treatment comparison

In the absence of a completed head-to-head trial, an indirect treatment comparison of semaglutide 2.4 mg versus liraglutide 3.0 mg was conducted. A patient-level regression was undertaken including STEP 1 and SCALE 1839 trial data, as discussed in Section B.2.9.

The ITC was not able to produce adjusted estimates for efficacy in responders (further details are contained within Appendix D). The placebo arms of the two trials were very similar in terms of baseline characteristics but did produce slightly different results for change from baseline in BMI and other risk factors. The efficacy of liraglutide 3.0 mg was adjusted in the model to reflect this difference. Observed efficacy in SCALE 1839 for all patients was used for this adjustment.[144]

The adjustment made was to increase the estimated efficacy of liraglutide 3.0 mg on BMI, SBP, HDL and total cholesterol efficacy estimates by the size of difference

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

104 of 176

between the efficacy estimates in the placebo arms of STEP1 (all patients) and SCALE 1839 (all patients). The percentage of non-diabetic hyperglycaemia patients with glycaemic status reversal was estimated using the odds ratio between liraglutide 3.0 mg (all patients) and placebo (all patients) in SCALE 1839, applied to STEP1 odds for placebo (all patients). Resulting efficacy parameters applied for liraglutide 3.0 mg are presented in Table 23.

Figure 13 illustrates the progression of BMI over time for semaglutide 2.4 mg and liraglutide 3.0 mg, including the additional effect of bariatric surgery, for patients with BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD. Semaglutide 2.4 mg shows a greater reduction in BMI compared to liraglutide 3.0 mg.

Table 23: Change in physiological parameter values – Liraglutide 3.0 mg – all patients

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

105 of 176

Figure 13: Projected BMI change for semaglutide 2.4 mg and liraglutide 3.0 mg

over time for patients with BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk for CVD

==> picture [457 x 260] intentionally omitted <==

Key : BMI, body mass index; CVD, cardiovascular disease.

B.3.3.1.4 Adverse events associated with treatment

Treatment-related AEs are included in the model through a per cycle probability of

occurrence in a treatment arm. Adverse events have a disutility and cost that is

applied in the cycle that it occurs.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

106 of 176

For the semaglutide 2.4 mg and diet and exercise arms, AE rates were sourced from the population containing all patients of the STEP1 trial.[85] The liraglutide 3.0 mg arm AE rates were sourced from the population containing all patients of the SCALE 1839 trial.[145] The AEs included in the model were non-severe hypoglycaemic events as these are most relevant for clinical practice (no severe hypoglycaemic events were observed), and severe gastrointestinal AEs as these were most common. The rate of AEs for each treatment included in the model is shown in Table 24, the resulting incremental probabilities are presented in Table 25. The utilities and costs applied for severe and non-severe hypoglycaemic events and severe gastrointestinal events are described in Section B.3.4.4 and Section B.3.5.3 respectively.

Table 24: Adverse event rates

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

107 of 176

Table 25: Incremental probability of adverse events per cycle for Semaglutide

2.4 mg versus diet and exercise

B.3.3.2 Bariatric surgery following non-surgical weight management intervention

Bariatric surgery refers to the use of surgical procedures for body weight management. To date, there are three main types of bariatric surgery: gastric banding, gastric bypass and sleeve gastrectomy.[146] Details of how bariatric surgery is included in the model are provided in Appendix L1.

The base case inputs for bariatric surgery eligibility reflect UK data and are shown in Table 26. It should be noted that the minimum BMI level is just to define whether patients are eligible or not. The actual BMI at the time of surgery is higher in clinical practice. The impact of using a higher BMI level for eligibility for bariatric surgery is explored in a scenario analysis. The default model efficacy by type of bariatric surgery on BMI, SBP, total cholesterol, HDL cholesterol and HbA1c, and case fatality are shown in Table 27.

108 of 176

Table 26: Bariatric surgery criteria

Table 27: Bariatric surgery efficacy and case fatality

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

109 of 176

B.3.3.3 Natural increase in BMI

The STEP 1 trial provided data for the BMI trajectory while on treatment for 68 weeks and an assumption that patients maintain treatment benefit is made for an additional year based on the efficacy plateau seen in the BMI data (Figure 12 and Figure 13).

After the treatment period the path taken for BMI in the arm receiving diet and exercise alone follows natural history of progression, a 0.1447 kg/m[2] (males) and 0.1747 kg/m[2] (females) annual increase in BMI as estimated by Ara et al. from a 100,000 UK Clinical Practice Research Datalink (CPRD) sample of individuals with obesity.[106]

This natural weight increase is a common assumption in obesity models and is supported by a model developed by NICE[150] Support is also found in Heitmann and Garby[151] and the analysis on the UK Clinical Practice Research Datalink.[106] Weight gain reaches a plateau at the ages of 65-70 years on average.[150] As a base case input in the model, weight can increase up to an average of 68 years based on KOL opinion.

B.3.3.4 Treatment discontinuation

Discontinuation is applied in the model in three different ways:

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

110 of 176

• Non-responder early discontinuation or stopping rule: Close to treatment initiation, i.e. after the first 3-months cycle in the model

• Maximum treatment duration: After a determined duration of treatment (2 years)

• Per cycle discontinuation: During treatment period, patients can discontinue due to any reason such as adverse events

Discontinuation is applied in the semaglutide 2.4 mg and liraglutide 3.0 mg arms of the model only.

B.3.3.4.1 Non-responder discontinuation / Stopping rule

Obesity interventions usually have a regulatory-imposed rule in their label to be stopped early-on in case of lack of response typically defined as <5% weight loss from baseline. In the case of liraglutide 3.0 mg, the regulatory approval by the European Medicines Agency determined that treatment with liraglutide 3.0 mg should be discontinued after 12 weeks on the 3.0 mg/day dose if patients have not lost at least 5% of their initial body weight. This was applied as such in the model.

In the base case analysis, a stopping rule was also used for semaglutide 2.4 mg (in the assumption that semaglutide 2.4 mg marketing authorisation will include such a rule), whereby treatment should be discontinued after 12 weeks on the maintenance dose, i.e. after 28 weeks (16 weeks to reach maintenance dose, 12 weeks on maintenance dose) if patients have not lost <5% of their initial body weight.

A scenario explored no stopping rule for semaglutide 2.4 mg, however as it is a regulatory requirement for liraglutide 3.0 mg, a stopping rule was always enabled for liraglutide 3.0 mg. When enabled, the following will be applicable to the proportion of non-responders in the cohort:

• Liraglutide 3.0 mg has a 4-week titration period before starting the maintenance dose at 3.0 mg/day. Hence, data on the percentage of non-responders from any of the trials of liraglutide 3.0 mg were based on responder-status assessment at 16 weeks after treatment initiation (4-week titration period + 12-week maintenance dose)

• Semaglutide 2.4 mg responder status in STEP 1 was evaluated after 28 weeks (16-week titration period + 12 weeks of maintenance dose)

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

111 of 176

• Non-responders will discontinue liraglutide 3.0 mg and semaglutide 2.4 mg immediately (i.e. they do not get liraglutide 3.0 mg/semaglutide 2.4 mg efficacy) but the model will account for the initial time on treatment pharmacy costs. Early discontinuation assumes that the non-responder part of the cohort goes on to receive diet and exercise and achieve diet and exercise efficacy

For clarity, Table 28 describes how efficacy parameters are applied to each treatment and proportion of the cohort, responder/non-responder in case of early treatment discontinuation or no early discontinuation. For semaglutide 2.4 mg the stopping rule is assumed to be applied at 28 weeks, so the STEP 1 full analysis set efficacy (responders and non-responders) is used up to and including Cycle 2. After Cycle 2, early responder efficacy is used for patients continuing semaglutide 2.4 mg. The population containing all early responder patients of STEP 1 was assumed to be representative in terms of semaglutide 2.4 mg treatment efficacy for both subgroups. A scenario analysis explores a post hoc analysis of STEP 1 using subgroup data. The proportion of non-responders for liraglutide 3.0 mg is based on the SCALE 1839 trial and is not matched to STEP 1 patient characteristics. Matching was not possible because of the difference in definition of response between the two trials.

Table 28: Efficacy data sources in relation to stopping rules and proportion of responders

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

112 of 176

B.3.3.4.2 Maximum treatment duration

The base case analysis assumes a maximum treatment duration of 2 years as in SWMS, treatment is typically provided for a maximum of two years.[27] This is also in line with TA664.

B.3.3.4.3 Per cycle discontinuation

In the base case analysis, the probability of discontinuation per cycle was sourced from the Kaplan–Meier curve of time to discontinuation for early 5% responders in the population containing all patients of STEP 1 for semaglutide 2.4 mg and diet and exercise, and this population’s discontinuation was assumed to be representative across both subgroups. For liraglutide 3.0 mg, this was sourced from the SCALE 1839 trial, safety analysis set for patients with BMI ≥ 35 kg/m[2] , non-diabetic hyperglycaemia and high risk of CVD in line with TA664.[27] The assumptions around responder discontinuation during treatment are as follows:

• The proportion of the cohort that discontinues before the maximum treatment duration, catches-up to the weight, SBP, total cholesterol and HDL cholesterol level as if patients stayed with diet and exercise according to the catch-up rate. The glycaemic status in patients in the temporary non-diabetic hyperglycaemia reversal health state that discontinue, revert back to the non-diabetic hyperglycaemia health state as defined by the catch-up rate

• After catch-up, physiological parameters will be equal to that in the diet and exercise arm and from there on follow the same path, i.e. both arms have weight going up following natural weight increase.

The catch-up period thus starts sooner for the proportion of the cohort discontinuing before the end of the 2-year treatment duration and defines the rate with which the respective proportion of the cohort will reach the same levels as the diet and exercise arm.

Note that for the proportion of the cohort discontinuing after 6 months, the catch-up duration will be slightly longer than defined by users (e.g. if catch-up period is defined to run over 3 years, the catch-up period for this cohort will be 3.5 years). Similarly, for the cohort discontinuing after 9 months of treatment, catch-up duration

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

113 of 176

will also be slightly longer than defined by users (e.g. if catch-up period is defined to run over 3 years, the catch-up duration for this cohort will be 3 years and 3 months). The proportion of responders discontinuing after 3 months of treatment for any reason (i.e. at start of Cycle 2) will behave as non-responders (non-responder discontinuation is described in above).

B.3.3.5 Catch-up period after treatment for BMI and glycaemic status

The effect of treatment on BMI, surrogate outcomes (SBP, HDL and total cholesterol) and glycaemic status is not lost immediately after treatment is stopped, but it is initially retained and wanes off over time.

The catch up is implemented so that at the end of the catch-up period BMI and surrogate markers for patients who had received active therapy have returned to the same levels as patients who received diet and exercise alone. The glycaemic status in patients in the temporary non-diabetic hyperglycaemia reversal health state reverts back to the non-diabetic hyperglycaemia health state for all treatments in the model after the catch-up period.

The definition of catch-up rate is provided in Table 29. By default, the catch-up rate for both surrogate outcomes and glycaemic status is assumed to be a cumulative 33% each year until the treatment effect is lost by Year 3 following treatment stop. Treatment effect waning rates are in line with assumptions used in TA664[27] , and in line with previously published evidence.[106]

Table 29: Definition of catch-up rate

B.3.3.6 Post-treatment level of SBP, cholesterol and HbA1c

In the base case analyses, although SBP and cholesterol may also be associated with a natural progression, for reasons of simplicity – as indeed patients may receive Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

114 of 176

blood pressure and cholesterol lowering drugs – the model only accounts for their evolution related to treatment effects, and there is no natural increase over time. When treatment is discontinued, the cohort returns to the value at baseline which is then maintained over the entire time horizon of the model.

HbA1c in diabetic patients is observed to increase with diabetes duration. The modelled cohort of interest enter the model without T2D but may develop T2D over time. It becomes difficult to represent HbA1c progression because diabetes duration will be different for all those who develop diabetes over the years, and they will all be at a different HbA1c level. This is because diabetes duration will be different for all those who develop diabetes over the years, and they will all be at a different HbA1c level.

Thus, to simplify, when conducting analyses and the cohort do not all have T2D, an average HbA1c of 7.5% is applied to the entire proportion of the cohort with T2D (either since baseline, or developed over time), and this does not increase further with time.

B.3.3.7 Obesity related complications

The model was adapted from the model submitted to NICE as part of TA664 for liraglutide 3.0 mg. This model was constructed from an SLR conducted in 2017 on the links between obesity and complications. This SLR was further updated to validate and identify sources to inform transition probabilities.[123] As described in Section B.3.2.2, not all the benefits of treatment are captured in the model as not all the comorbidities defined in that section could be modelled. Therefore, the risk equations used to extrapolate clinical efficacy parameters in disease outcomes for which data exist are included in Table 30, which presents which risk equation is used for each complication. More details on the risk equations for onset of T2D and CV events are provided in Appendix L.

115 of 176

Table 30: Risk equations used for obesity-related complications

B.3.3.7.1 Cardiovascular events – individual risks

Individual risk models were available for the following outcomes: first ischaemic heart disease (IHD) considered angina herein, first MI and first stroke, recurrent MI and recurrent stroke. There was no model to predict recurrent angina, hence, the estimated risk for recurrent events (MI and stroke) was adjusted based on the proportions of MI and stroke of total CVD exhibited in Table 31.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

116 of 176

Table 31: Clinical inputs for calibration of composite cardiovascular endpoints

B.3.3.7.2 Obstructive sleep apnoea

The proportion of the cohort having OSA depends on the BMI level in the cycle. The prevalence of OSA by BMI level was sourced from the Sleep Heart Study.[128] This study found that the prevalence of OSA as defined according to Apnoea-Hypopnea Index (AHI) ≥15 is 13% at BMI levels between 24.4–28.0 kg/m[2] (irrespective of gender). The reported odds ratio corresponding to one standard deviation (SD) increment in BMI was 1.6 (1.45, 1.76). This study was preferred to other available studies because it was the largest in sample size (n=5,615) and it provided sufficient data to calculate a prevalence rate per unit BMI. It was also preferred to other studies as it investigated the prevalence of moderate-to-severe OSA (AHI ≥15), given that in the present health-economic analysis, OSA was assigned a hospital cost for continuous positive airway pressure treatment. The BMI-prevalence table used in the model is illustrated in Appendix L.

Throughout the time horizon, the proportion of the cohort having sleep apnoea may reside in any non-dead health state (i.e. OSA co-occurs with any obesity complication including non-diabetic hyperglycaemia at baseline). This was possible as OSA was assumed not to influence the progression to and from other health states or events.

117 of 176

B.3.3.7.3 Knee replacement

The annual incidence of knee replacement surgeries in the reference BMI group (2022.5 kg/m[2] ) for ages <65 years and ≥65 years was sourced from the study of Wendelboe et al. 2003, reporting figures of 0.053% and 0.12%, respectively.[129] The study provided granular data on the association between BMI and incidence of knee surgeries by 2.5 BMI-unit steps for observed BMI levels between 17.50 and 42.49 kg/m[2] and was hence preferred to other studies available in the literature. To derive a continuous function of the BMI-risk of knee replacement, and to extrapolate the association beyond the observed 42.49 kg/m[2] BMI in the study, a second-order polynomial trend was fitted to the calculated probabilities. Separate trend lines were fitted for males and females, and for patients aged 64 years or lower, and above 65 years. Figure 14 illustrates the incidence rate applied corresponding to an average age of the cohort of 48 years. The fitted polynomial trend functions are illustrated in Table 32.

Figure 14: Annual probability of knee replacement surgery by level of BMI

==> picture [435 x 215] intentionally omitted <==

Key: BMI, body mass index. Source: Wendelboe.[129]

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

118 of 176

Table 32: BMI-dependent risk functions for knee osteoarthritis

B.3.3.8 Mortality

Survival of the model cohort is estimated using a combination of disease specific and BMI adjusted mortality. Disease specific mortality is split into long term mortality, associated with the different comorbidity health states adjusted by BMI and, shortterm fatality associated with CVD events and knee replacement.

B.3.3.8.1 Long term mortality

General population mortality, defined as age and gender-specific all-cause mortality, was included in the model based on UK lifetables.[136] The general population mortality was adjusted by excluding the mortality of obesity related comorbidities accounted for elsewhere in the model such as diabetes, IHD, cerebrovascular diseases and arthrosis. The long-term mortality rate associated with obesity complications were obtained from mortality statistics by age, sex and underlying cause of death using ICD10 codes.[135]

The long-term mortality rate for each health state in the model was adjusted by the hazard ratio for BMI. This obtained the long-term disease specific mortality rate specific to the BMI. The association between all-cause mortality and BMI was obtained from Bhaskaran et al. 2018.[50] The digitised curve, used as a model input, is shown in Figure 15.

119 of 176

Figure 15: BMI related HRs from Bhaskaran et al. 2018

==> picture [413 x 255] intentionally omitted <==

Key: BMI, body mass index; HR, hazard ratio. Source: Bhaskaran.[50]

In addition to all-cause mortality, the model includes a higher risk of mortality associated with obesity complications which is applied to the cohort in each respective health state. The higher long-term mortality risk associated with diabetes and CVD was used to adjust general population mortality for the cohort in health states reflecting such complications.

General population mortality for the cohort who developed CVD was up-adjusted by a factor representing the long-term higher risk of death from having had an ACS or an MI (Table 33).

Table 33: Relative risks used to adjust general population mortality

120 of 176

B.3.3.8.2 Obesity complications short-term mortality

Short-term fatality from CVD events and knee replacement and bariatric surgery is considered in the model at the event occurrence (Table 34).

Case-fatality associated with angina, MI and stroke are considered in the model and used to account for a higher risk of death in the year when the event occurs. Casefatality associated with angina and MI was weighted and applied to the cohort developing ACS, and stroke related mortality was applied to the cohort developing a stroke event. Case-fatality was defined as death that occurs within 1 month of the event.

Table 34: Short-term mortality from complications

B.3.4. Measurement and valuation of health effects

B.3.4.1 Health-related quality-of-life data from clinical trials

The SF-36 and IWQOL-Lite-CT HRQoL measures were administered to patients in the STEP 1 trial at baseline, Week 8, Week 16, Week 20, Week 36, Week 52 and Week 68. For both SF-36 and IWQOL-Lite-CT mean scores (all domains) at baseline were similar between the semaglutide 2.4 mg and placebo treatment arms.

The mean change in SF-36 physical functioning score from baseline to 68 weeks was significantly greater (p < 0.001) for semaglutide 2.4 mg compared with placebo (2.21 versus 0.41, respectively; ETD: 1.8; 95% CI: 1.2, 2.4).[83] Figure 8 in Section B.2.6.9.1 shows that the ETDs in change from baseline to Week 68 for all individual

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

121 of 176

health domain scores and for the physical component summary (PCS) and mental component summary (MCS), scores of SF-36 were all in favour of semaglutide 2.4 mg.

The estimated mean change in IWQOL-Lite-CT physical function score was significantly greater (p < 0.001) for semaglutide 2.4 mg compared with placebo (14.67 versus 5.25, respectively; ETD: 9.43; 95% CI: 7.5, 11.35).[83] Figure 9in Section B.2.6.9.2 shows that the ETDs in change from baseline to Week 68 for the physical function, physical, psychological and total scores of the IWQOL-Lite-CT were all in favour of semaglutide 2.4 mg.

Neither SF-36 or IWQOL-Lite-CT HRQoL are representative of the NICE reference case, nor do they yield utilities that can be applied to an economic model. They were therefore not used for the present economic analysis.

B.3.4.2 Mapping

No mapping was performed for this cost-effectiveness model as the model uses published utility values throughout the whole lifetime of the analysis.

B.3.4.3 Health-related quality-of-life studies

Systematic searches for HRQoL studies were carried out to identify all relevant studies on adult patients with obesity. The SLR conducted for semaglutide in April 2021 is an update to the previous SLR that was conducted for the liraglutide NICE TA664 in October 2018. For the original SLR and SLR update combined, a total of 45 records were included for data extraction, corresponding to 26 unique HRQoL studies.

Findings from the SLR demonstrated the lack of comprehensive published utility data, and studies identified in the SLR were therefore not used in the base case analysis. HRQoL inputs used in the base case were instead based on a large UK population-based study that has demonstrated a robust association between BMI and utility (see Section B.3.4.5).[137] Detailed information on the HRQoL literature review and included studies is provided in Appendix H.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

122 of 176

B.3.4.4 Adverse reactions

As discussed in Section B.3.3.1.4, the adverse events included in the model are nonsevere hypoglycaemia and severe gastrointestinal AEs. Their impact on the cohort of patients is included through disutilities that are applied in the cycle that the event occurs.

B.3.4.5 Health-related quality-of-life data used in the cost-effectiveness analysis

In the model HRQoL is varied by age and BMI level, defined as baseline utility values without obesity complications. Baseline utility values are then adjusted for HRQoL decrements associated with obesity related complications. HRQoL decrements may be associated with the disease, e.g. diabetes or CVD, and therefore applied to baseline utility values to derive health state utilities at each cycle; or may be associated with an event, e.g. MI, and therefore applied in the model as one-off disutilities.

B.3.4.5.1 Baseline utility – per BMI level and age

The model allows the derivation of baseline utility as a function of BMI of the cohort with obesity and no comorbidities based on a polynomial model. The polynomial model reflects clinical opinion that the marginal rate of utility is most likely different from changing BMI from 26 to 25 compared with from 46 to 45. Furthermore, a polynomial model more accurately reflects that HRQoL may increase up to a BMI of around 25kg/m[2] .

The source used for deriving such a model is Søltoft et al.[137] . Appendix N describes why this was selected as the most appropriate source. The utility curves were reestimated using a polynomial of third degree for BMI ranges 15–35 kg/m², as in Søltoft et al.[137] Then, a logarithmic function was fitted to utilities in the BMI range 27– 35 kg/m² and extrapolated up to a BMI of 60. The coefficients of the two functions (polynomial and log) for men and women are exhibited in Table 35. The coefficients for the age-adjustment (two scenarios) are reported in Table 36. Details of the two studies are presented in Appendix N.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

123 of 176

Table 35: Re-estimate association between BMI and utility

Table 36: Coefficients to adjust BMI dependent baseline utility as a function of age from Søltoft et al. 2009 and Burström et al. 2001 studies

B.3.4.5.2 Acute decrements in utility

Non-fatal acute events considered in the model include ACS, osteoarthritis, stroke and TIA. Upon the occurrence of one of these events, a one-off disutility is applied in the cycle in which the event occurs. To account for HRQoL impact associated with severe musculoskeletal disorders, the model includes a disutility applied in the year a knee replacement surgery occurs. The disutility is applied once, to the event, and is multiplied with a factor of three to account for three years of living with a chronic, debilitating condition prior to surgery. This assumption was in line with modelling knee replacement in TA664.[27]

124 of 176

A summary of all events for which an acute disutility is included in the model is shown in Table 37. No disutility associated with treatments was included in the model.

Table 37: Disutility values associated with acute events

For fatal events, it is assumed that death occurs in the middle of a cycle. Thus, the total disutility for stroke, for example, for the proportion of the cohort experiencing death from the event, was calculated by multiplying the proportion of the cohort dying within 30 days from the event by half the negative total utility value of the cycle in which the proportion of the cohort died.

The total disutility at each cycle is calculated by taking the sum of non-fatal acute events and fatal acute events.

B.3.4.5.3 Health state utility values

In addition to acute decrements from baseline utility, long-term absolute HRQoL decrements associated with each obesity related complication are considered in the model (Table 38). Such HRQoL decrements are derived from the literature and are subtracted from age, gender and BMI dependent baseline utility values at each cycle to derive health state utility values. By doing so, health state utility values continue to be adjusted for age, gender and BMI level, in addition to HRQoL decrements associated with a given disease.

No HRQoL decrement is assigned for non-diabetic hyperglycaemia.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

125 of 176

Table 38: HRQoL decrements of acute events

When health states combine two or more obesity complications, the HRQoL decrement associated with each single complication is summed together and the total is then subtracted from the baseline utility. For example, in the health state ‘T2D + Post ASC’ the total HRQoL decrement subtracted from the baseline utility is equal to the sum of the HRQoL decrement for T2D and the HRQoL decrement for post ASC. Gough et al. 2009 concluded that the HRQoL decrements associated with T2D and obesity showed no significant interaction and thus could be assumed to be additive.[157] HRQoL decrements associated with each potential combination of obesity complications included in the model are not reported in the literature and so the current approach was therefore seen as the most appropriate solution to account for the HRQoL impact of obesity complications in a transparent way without underestimating the associated severe humanistic burden.

B.3.4.5.4 Disutilities related to bariatric surgery

A one-off disutility is applied to the proportion of the cohort receiving bariatric surgery at each cycle (Table 39). The disutility represents the decrement in quality of life associated with the surgical procedure and related complications.

Table 39: HRQoL decrements for bariatric surgery

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

126 of 176

B.3.4.5.5 Disutilities related to AEs

Table 40 illustrates modelled disutilities per type of treatment-related adverse event in the model.

Table 40: Treatment-related adverse event disutilities

B.3.4.5.6 Summary of utility values used in the model

Table 41 provides a summary of utility values used in the model as well as a comparison to values used in TA664.[27] Where values are dependent on BMI and age, the value for the mean of the baseline population is presented.

127 of 176

Table 41: Summary of utility values for cost-effectiveness analysis

B.3.5. Cost and healthcare resource use identification, measurement and valuation

Appendix I outlines the SLR to search for studies reporting cost and healthcare resource data for the treatment of patients with obesity in the UK. The SLR conducted for semaglutide in April 2021 is an update to the previous SLR that was conducted for the liraglutide NICE TA664 in October 2018.

Three studies identified in the combined searches were from the UK perspective.[159-] 161 Information extracted was not found to be relevant for the current economic analysis as studies did not focus on the patient population or treatments identified as relevant to the decision problem, and thus was not utilised (see Appendix G and I).

B.3.5.1 Intervention and comparators’ costs and resource use

B.3.5.1.1 Treatment costs

Treatment costs include the acquisition cost of pharmacological treatment and the cost of diet and exercise.

The cost of pharmacological treatment is included in the model based on the dose as per the list price and the proposed confidential discount price (PAS price). Costs for both the titration and maintenance doses are included. The total treatment cost per cycle is calculated by multiplying the cost per mg by the dose in mg required in a given cycle. The cost of needles is provided by Novo Nordisk. To account for the stopping rules used within the 3-month cycle length during the first year of the model, an adjustment is made to the pharmacy cost. For semaglutide 2.4 mg, the

adjustment removes 2 weeks of treatment acquisition costs after Cycle 2 (6 months). For liraglutide 3.0 mg, the adjustment adds 1 month of treatment acquisition costs to Cycle 1 (3 months) which is removed in Cycle 2.

The cost of the diet and exercise intervention is included in the obesity monitoring cost. Table 42 reports the annual cost of treatment and monitoring using list prices for semaglutide and liraglutide.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

129 of 176

B.3.5.1.2 Obesity monitoring

The cost of obesity monitoring is defined as the average cost of all routine visits, examinations, diet and exercise required for the management of an adult patient with obesity. The routine obesity monitoring cost is included per year; however, it is readjusted automatically to fit the 3-month cycles in the first year of the simulation.

The cost of obesity monitoring is applied to each comparator in the model since all pharmacological treatments are assumed to be administered in conjunction with diet and exercise and to require visits/examination follow-up. Table 42 reports the costs of obesity monitoring, diet and exercise.

Table 42: Annual treatment costs (list price)

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

130 of 176

B.3.5.2 Health-state unit costs and resource use

Annual cost of obesity related complications includes costs associated with monitoring and treating a given disease and is used in the model to define health state costs. For T2D, the cost of insulin treatment is considered separately and is based on the dose and type of insulin treatment available in UK practice. The costs of health states including multiple obesity complications are calculated by summing the costs associated with each condition. Obesity related complication costs are derived from UK published studies or real-world data. Table 43 displays annual costs for each health state in the model and the pharmacy costs related to T2D.

Table 43: Annual obesity related complication costs applied to health states in

the model

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

131 of 176

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

132 of 176

B.3.5.3 Adverse reaction unit costs and resource use

Table 44 illustrates the costs per type of treatment-related adverse event in the model.

Table 44: Costs of treatment-related adverse events

B.3.5.4 Miscellaneous unit costs and resource use

B.3.5.4.1 Bariatric surgery costs

As bariatric surgery is included in the model as a downstream event, it has a corresponding one-off cost. The cost is applied to the proportion of the cohort receiving the surgery at each cycle and includes preoperative management, procedure, postoperative follow-up and surgery related complications.[146]

The average procedure cost is calculated as the weighted average cost of the three types of procedures. The cost of leaks is used as a proxy for bariatric surgery related complications[147] and is included in the model as an average across all patients, i.e. already weighted by the incidence of complications. Table 45 reports bariatric surgery costs.

Table 45: Bariatric surgery costs

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

133 of 176

B.3.5.4.2 Acute event cost

The model includes the one-off costs of the obesity related acute events angina, MI, stroke, knee replacement and TIA. The costs of obesity related acute events represent the economic burden associated with managing the patient when the acute event occurs, including hospitalisation costs. In the case of knee replacement, this is also associated with pre-surgery visits/examinations and post-surgery followup. The cost of managing osteoarthritis before surgery (e.g. analgesics) was considered negligible and not accounted for in the model.

The costs of obesity related acute events are included in the model as one-off, and separately from health state costs. Table 46 reports acute event costs.

Table 46: Acute event costs

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

134 of 176

B.3.5.5 Comparison of costs inputs to TA664 costs inputs

Table 47 presents a comparison if cost inputs between the current submission and TA664. Where there is a difference between the costs, this is either because costs sources have been updated, or more appropriate cost sources were identified in a targeted literature review (TLR).

Table 47: Comparison of cost inputs between current submission and TA664

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

135 of 176

B.3.6. Summary of base-case analysis inputs and assumptions

B.3.6.1 Summary of base-case analysis inputs

A summary of base case model inputs is provided in Table 48 to Table 50. A default margin of error of 25% around the mean estimate was applied where standard errors of the mean were not available/not reported.

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

136 of 176

Table 48: Summary of baseline cohort characteristics and AE probabilities in the economic model

Company evidence submission for semaglutide 2.4 mg for managing overweight and obesity [ID3850]

137 of 176

Table 49: Summary of efficacy inputs applied in the economic model

Table 50: Summary of epidemiological inputs, utilities and costs applied in the economic model