NZHTA REPORT August 2004 Volume 7 Number 6
Evidence Based Review of Weight Loss Medicines:
A report commissioned by the New Zealand Accident Compensation Corporation (ACC)
Pamela Smartt New Zealand Health Technology Assessment
Department of Public Health and General Practice Christchurch School of Medicine Christchurch, NZ.
Division of Health Sciences, University of Otago
NEW ZEALAND HEALTH TECHNOLOGY ASSESSMENT (NZHTA) Department of Public Health and General Practice Christchurch School of Medicine, Christchurch, New Zealand
Evidence Based Review of Weight Loss Medicines:
A report commissioned by the New Zealand Accident Compensation Corporation (ACC)
Pamela Smartt
NZHTA REPORT August 2004 Volume 7 Number 6
This report should be referenced as follows: Smartt, P. Evidence based review of weight loss medicines: a report commissioned by the New Zealand Accident Compensation Corporation (ACC). NZHTA Report 2004 ; 7(6) . 2004 New Zealand Health Technology Assessment (NZHTA) ISBN 1 877235 70 9 ISSN 1174 5142
i
Acknowledgements
This report was prepared by Dr Pamela Smartt (Principal Investigator) who conducted the critical appraisals and prepared the project report. Dr Ray Kirk (NZHTA Director until February 2005) also provided comment on various drafts and coordinated the overall project. Dr Robert Weir (NZHTA Acting Director from February 2005) coordinated the project from February 2005. Ms Margaret Paterson (NZHTA Information Specialist) developed and undertook the search strategy and coordinated retrieval of documents. Mrs Ally Reid (NZHTA Administrative Secretary) provided document formatting.
The Canterbury Medical Library assisted with the retrieval of articles.
NZHTA is a Research Unit of the University of Otago funded under contract to the Ministry of Health.
Disclaimer
New Zealand Health Technology Assessment (NZHTA) takes great care to ensure the information supplied within the project timeframe is accurate, but neither NZHTA, the University of Otago, nor the contributors involved can accept responsibility for any errors or omissions. The reader should always consult the original database from which each abstract is derived along with the original articles before making decisions based on a document or abstract. All responsibility for action based on any information in this report rests with the reader. NZHTA and the University of Otago accept no liability for any loss of whatever kind, or damage, arising from reliance in whole or part, by any person, corporate or natural, on the contents of this report. This document is not intended as personal health advice. People seeking individual medical advice are referred to their physician. The views expressed in this report are those of NZHTA and do not necessarily represent those of the University of Otago or the New Zealand Ministry of Health.
Copyright
Copyright © to Accident Compensation Corporation 2005. All rights reserved. No part of this report may be reproduced or distributed by any person without prior written permission and/or licence from the Accident Compensation Corporation. http://www.acc.co.nz/
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC ii
Contact details
New Zealand Health Technology Assessment (NZHTA) Department of Public Health and General Practice Christchurch School of Medicine and Health Sciences PO Box 4345 Christchurch New Zealand Tel: +64 3 364 3696 Fax: +64 3 364 3697 Email: [email protected] Web Site: http://nzhta.chmeds.ac.nz/
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC iii Contents
Contents ...... iii Executive Summary ...... vii Recommendations ...... viii Background...... 1 Definition and measurement of obesity...... 1 The health consequences of obesity...... 1 Factors influencing body weight...... 2 Clinically significant weight reduction...... 3 Obesity and Injury...... 3 Obesity trends in NZ...... 3 The economic cost of obesity...... 4 Benefits associated with weight loss...... 4 Review questions ...... 7 Pharmacological treatments for obesity ...... 9 Pharmacological interventions registered in New Zealand...... 9 Phentermine (Duromine TM , Umine Timedcaps)...... 9 Diethylpropion hydrochloride (Tenuate Dospan)...... 10 Sibutramine hydrochloride (Reductil ®, Meridia ®) ...... 11 Orlistat (Xenical ®)...... 12 Meal replacement plans...... 14 Methodology ...... 15 Systematic review...... 15 Search strategy and information sources ...... 16 Study selection criteria...... 17 Methods of the review...... 17 Levels of evidence...... 18 Results...... 19 Meal replacement plans: weight loss and co morbidity risk reduction...... 19 Phentermine hydrochloride: weight loss and co morbidity risk reduction...... 20 Diethylpropion: weight loss and co morbidity risk reduction...... 22 Orlistat: weight loss and co morbidity risk reduction...... 23 Sibutramine: weight loss and co morbidity risk reduction...... 28 Comparative drug studies...... 34 Combined drug studies...... 39 Safety/Side effects ...... 41 Safety and side effects of sibutramine therapy...... 41 Safety and side effects of orlistat therapy ...... 42 Safety and side effects of phentermine therapy ...... 43
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC iv
Safety and side effects of diethylpropion therapy...... 43 Summary of potential harms of pharmacotherapy intervention ...... 44 Safety and side effects of meal replacement programs...... 44 Practice recommendations and guidelines ...... 47 National Institute of Health, USA...... 47 The National Institute for Clinical Excellence (NICE), UK ...... 47 High profile clinical trials...... 49 Completed studies...... 49 Childhood obesity ...... 49 Ongoing trials...... 50 Other anti-obesity therapies ...... 51 Horizon Scan ...... 53 Economic considerations ...... 55 The direct cost of medication...... 55 Economic analysis ...... 57 Economic evaluations of orlistat...... 57 Economic evaluations of sibutramine ...... 58 Economic evaluations of phentermine...... 63 Economic evaluations of diethylpropion...... 63 Discussion...... 65 Criteria for the evaluation of medication for the treatment of obesity ...... 65 Expected weight loss and duration of weight loss...... 66 Potential barriers to use ...... 66 Evidence Summary and Conclusions...... 67 Clinical effectiveness and circumstances of use...... 67 Safety and side effects...... 72 Ethnic groups...... 73 Economic considerations...... 73 Evidence Tables...... 75 Appendices...... 105 References...... 113
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC v
Tables
Table 1. Classification of body weight...... 1 Table 2. Obesity related co morbidities and risk...... 2 Table 3. Inclusion and exclusion criteria for review studies...... 17 Table 4. Scottish Intercollegiate Guidelines Network Evidence levels ...... 18 Table 5. Weight loss results at three months and one year for an unstratified and stratified study population...... 19 Table 6a. Phentermine effectiveness studies ...... 21 Table 6b. Phentermine vs. placebo post treatment outcomes, Haddock et al., 2002...... 21 Table 7. Diethylpropion effectiveness studies, Glazer et al., 2001...... 22 Table 8. Diethylpropion vs. placebo post treatment outcomes, Haddock et al., 2002...... 23 Table 9. Summary of publications reporting the effectiveness of orlistat for weight loss in overweight and obese participants...... 24 Table 10. Meta analysis of RCTs for Orlistat, long term studies ≥ 1 year (NICE HTA, ...... Avenell, May 2004)...... 25 Table 11. Orlistat 120 mg/d weight loss in otherwise healthy obese populations ...... 26 Table 12. Orlistat 30 120 mg/d weight loss in patients with hypercholesterolemia, results after one year...... 27 Table 13. Weight loss with orlistat: results of 1 year clinical trials, European Medicine Evaluation Agency...... 27 Table 14. Orlistat 120 mg/d in diabetic patients or patients with glucose intolerance...... 27 Table 15. Eligible studies reporting the effectiveness of sibutramine ...... 28 Table 16. Meta analysis of RCTs for Sibutramine, long term studies ≥ 1 year (NICE HTA, Avenell May 2004)...... 30 Table 17. Studies reporting the effectiveness of sibutramine in otherwise healthy overweight individuals...... 32 Table 18. Studies reporting the effectiveness of sibutramine in individuals with type 2 diabetes...... 32 Table 19. Sibutramine dose ranging studies. Bray 1999 reported in Leung et al., 2003...... 34 Table 20. Eligible studies comparing two or more review drugs ...... 34 Table 21. Summary of the comparative effectiveness of orlistat, sibutramine and low fat diets that may include meal replacements...... 35 Table 22. Summary of short term weight loss trials of sibutramine and orlistat not otherwise reported in systematic reviews and HTAs...... 35 Table 23. A comparison of the effectiveness of sibutramine, orlistat, phentermine and diethylpropion (Glazer et al., 2001)...... 37 Table 24. A comparative study of the effectiveness of sibutramine, orlistat, phentermine and diethylpropion (Haddock et al., 2002)...... 37 Table 25. Weight loss drugs trade off between benefit and harm...... 38 Table 26. The effect of sibutramine on blood pressure and heart rate; results from 20 studies (Nisoli and Carruba, 2003)...... 41 Table 27. NZ Drug pricing according to MIMS, 2004 ...... 55 Table 28. The estimated cost of a course of anti obesity treatment ...... 55 Table 29. Studies reporting on the cost or cost effectiveness of sibutramine, orlistat, phentermine and diethylpropion...... 57 Table 30. Sensitivity analysis around two base estimates of the cost effectiveness of sibutramine when different utilities are assigned to weight loss...... 59 Table 31. Economic evaluations of orlistat, sibutramine and phentermine...... 61
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC vi
Evidence Table 1. Meal replacement, Heymsfield, 2003 ...... 76 Evidence Table 2. Meal replacement, Allison et al., 2003...... 77 Evidence Table 3. Orlistat, Torgerson et al., 2004...... 78 Evidence Table 4. Orlistat, Rissanen et al., 2001 ...... 79 Evidence Table 5. Orlistat, Muls et al., 2001 ...... 80 Evidence Table 6. Orlistat, Foxcroft et al., 2000 ...... 81 Evidence Table 7. Orlistat, Derosa et al., 2003...... 82 Evidence Table 8. Orlistat, Hanefield et al., 2002...... 83 Evidence Table 9. Orlistat, Halpern et al., 2003...... 84 Evidence Table 10. Orlistat, Leung et al., 2003...... 85 Evidence Table 11. Orlistat, Heymsfield et al., 2000...... 86 Evidence Table 12. Orlistat, O’Meara et al., 2001...... 87 Evidence Table 13. Orlistat, Padwal et al., 2004 ...... 89 Evidence Table 14. Sibutramine, O’Meara et al., 2002...... 90 Evidence Table 15. Sibutramine,Tambascia et al., 2003 ...... 91 Evidence Table 16. Sibutramine, Padwal et al., 2004 ...... 92 Evidence Table 17. Sibutramine, Leung et al., 2003 ...... 93 Evidence Table 18. Sibutramine, Nisoli and Carruba, 2003...... 94 Evidence Table 19. Sibutramine, Berkowtitz et al., 2003...... 95 Evidence Table 20. Sibutramine, Hazenberg et al., 2000...... 96 Evidence Table 21. Sibutramine, Gokcel et al., 2001 ...... 97 Evidence Table 22. Sibutramine, Kim et al., 2003...... 98 Evidence Table 23. Sibutramine, Hauner et al., 2003...... 99 Evidence Table 24. Sibutramine, McNulty et al., 2003 ...... 100 Evidence Table 25. Sibutramine, Wadden et al., 2000 ...... 101 Evidence Table 26. Comparative drug study, Poston et al., 2001 ...... 102 Evidence Table 27. Comparative drug study, Haddock et al., 2002...... 103
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC vii Executive Summary
Obesity is a chronic, relapsing disease which increases the risk of a number of other serious diseases including diabetes, heart disease, hypertension, stroke and some cancers. Even a modest weight loss in obese individuals may be associated with significant health benefits arising from improvements in blood pressure, blood glucose and cholesterol levels.
Anti obesity drug treatment is perceived as an adjunct to other weight loss strategies in an overall program of supervised weight reduction, weight loss maintenance or management. Co interventions are typically a calorie controlled diet, behavioural therapy and/or physical exercise.
Weight loss drugs currently available in New Zealand include phentermine, diethylpropion, orlistat and sibutramine. Meal replacement products are readily available and popular non pharmacological weight loss adjuncts.
The evidence suggests that phentermine, diethylpropion, orlistat, sibutramine and specified meal replacement plans/products are all effective at achieving a moderate weight loss in obese individuals. On average, an additional weight loss of approximately 4.0 kilograms may be achieved when these products are used as part of an appropriate weight loss program. Clinically important weight loss is considered to be a loss of ≥ 5% of initial body weight; a substantial proportion of patients were reported to have achieved this goal with the aid of weight loss drugs.
In comparative studies, the newer weight loss drugs orlistat (Xenical ®) and sibutramine (Reductil ®) are reported to be safer, more acceptable and more effective weight loss agents than the earlier amphetamine related drugs, phentermine and diethylpropion. They can be prescribed for longer periods than the older drugs and both orlistat and sibutramine are currently being tested in adolescents. Sibutramine use has been cautioned in this age group outside a clinical trial while orlistat has recently been approved by the Food and Drug Administration (FDA) for use in adolescents.
Sibutramine may result in a higher weight loss than orlistat but with side effects that may limit its use in patients with hypertension or cardiovascular disease. For these patients, orlistat may provide a useful alternative although the accompanying gastrointestinal (GI) side effects may not be acceptable to all patients. Meal replacement plans may be useful for patients who cannot tolerate or do not respond to either of these medications.
Sibutramine is reported to be cost effective under most scenarios, with a lower estimated cost per quality adjusted life year (QALY) gained in patient groups with significant obesity related co morbidities such as diabetes. The cost per QALY gained with orlistat is reported to be high and for healthy obese patients orlistat may not be cost effective. However, for patients with obesity related co morbidities such as hypertension and/or hypercholesterolemia, orlistat may be considered to be good value for money.
It is not clear if phentermine, diethylpropion or meal replacement plans are cost effective treatments for obesity as no economic analyses of these therapies were identified, however, the direct cost of a course of phentermine or diethylpropion is much lower than the cost of orlistat or sibutramine.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC viii Recommendations
The weight loss drugs sibutramine and orlistat are effective adjuncts to an appropriate clinical program of weight loss for obese ACC claimants. Short term treatment of up to six months should be considered for maximum weight loss. Longer term treatment – i.e., up to two years should be considered if weight loss is to be maintained.
Patients with uncontrolled hypertension or cardiovascular disease may not be suitable candidates for sibutramine therapy; in these patients orlistat should be considered.
Meal replacement products should be considered for patients who are unable to tolerate these medications or who have not had an appropriate response. There is insufficient evidence to enable any specific product recommendations to be made.
Each of these interventions should be part of a medically monitored comprehensive regimen of weight reduction based on a calorie controlled diet, behaviour modification and where possible, physical exercise and lifestyle modifications.
Individuals with a body mass index (BMI) ≥ 30kg/m 2 are generally considered to be obese and should be considered for weight loss therapy. Claimants who have a BMI of ≥ 27kg/m 2 and obesity related co morbidities including diabetes, hypertension and sleep apnoea may also benefit from medical assistance with weight loss and should be considered.
Consideration should also be given to the differing recommendations for obesity thresholds in non Caucasian populations. The World Health Organisation (WHO) has defined a lower obesity threshold (BMI ≥ 25kg/m2) for use in Asian populations and a higher threshold (BMI ≥ 32kg/m 2) for Polynesians.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC
1 Background
Definition and measurement of obesity
Obesity is a chronic, relapsing disease in which there is an abnormal accumulation of body fat. Definitions and measurement of obesity vary but most researchers currently use BMI 1 to classify individuals in terms of weight; individuals with a BMI of ≥ 30kg/m 2 are generally considered to be obese, see Table 1.
Table 1. Classification of body weight
% Above BMI (kg/m 2) Weight Class Ideal weight Caucasian Asian* PI*
Normal 0% 18.5-24.9 18.5-22.9 18.5-22.9 Overweight 20% 25.0-29.9 23.0-24.9 26.0-31.9 Obese a Class I (mild) 20-40% 30.0-34.9 25-29.9 ≥32.0 Class II (moderate) 400-100% 35.0-39.9 ≥30.0 Class III (severe- morbid) >100% ≥40.0 - - *Figures taken from WHO, The Asia-Pacific Perspective: Redefining Obesity and its treatment. PI= Pacific Islander. a person with a BMI over 30 is considered to be obese and this is an arbitrary value obtained from epidemiological studies showing that a relationship between mortality and BMI follows a J-shaped curve and that mortality increases by 50-100% at BMI value above 30kg/m.
Generally BMI correlates well with body fat, however, BMI is not a foolproof guide to morbid fat accumulation, as it does not take into account body frame size, proportion of lean mass, age, gender or ethnic differences. Misclassification can occur when a BMI =30kg/m 2 threshold is used to define obesity in non Caucasian populations. Polynesians tend to have a lower fat percentage than Caucasians for any given BMI and the threshold for obesity in Polynesians is BMI ≥ 32kg/m 2. Asian populations, however, have more fat and co morbidities for any given BMI than Caucasians and a lower obesity threshold of BMI=26kg/m 2 has been suggested (Proietto and Baur 2004).
Other useful, but less commonly used, obesity indices include waist to hip ratio and waist circumference. These measures may provide additional useful information regarding risk factors associated with weight gain. In Caucasians, a WHR>1.02 for men and a WHR >0.88 for women is used to identify obesity and abdominal fat accumulation (NHANES III); this may be a better predictor of weight related cardiovascular risk than BMI.
Genetic disorders such as Alstrom’s syndrome and endocrine disorders such as hypothyroidism and Cushing’s disease must be ruled out before diagnosing obesity (Leung et al. 2003).
The health consequences of obesity
The 1997 New Zealand National Nutrition Survey (Wilson et al. 2001) estimated that the health risks associated with obesity were equivalent to a number of high risk, chronic diseases including type 2 diabetes, heart disease, hypertension, stroke, gallstones and some cancers.
1 BMI= weight in kilograms divided by height in metres squared (kg/m 2).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 2
The distribution of body fat is an important determinant of the type of associated health risk; abdominal or visceral fat (android obesity) is associated with increased risk of cardiovascular mortality, hypertension and non insulin dependent diabetes.
Increased health risks for conditions associated with obesity vary from a substantial to moderately raised risk for diabetes, heart disease, hypertension and a number of other metabolic disturbances (Brown et al. 2000; Carey et al. 1997; Chan et al. 1994; Manson et al. 1990; Rimm et al. 1995) to a slightly raised risk for some cancers and lower back pain (Bergstrom et al. 2001a; Bergstrom et al. 2001b; Calle et al. 2003; Oliveria et al. 1999). There is also an increased risk of adverse events from anaesthesia in obese patients, see Table 2.
Table 2. Obesity related co-morbidities and risk
Substantial Risk Increase (RR>3) Moderate Risk Increase (RR=2-3) Slight Risk Increase (RR= 1-2)
Type II diabetes Coronary heart disease Breast, endometrial, colon cancers Gall bladder Hypertension Reproductive hormone abnormalities Dyslipidaemia Osteoarthritis of the knee and hips Polycystic ovary syndrome Metabolic syndrome Hyperuricemia Impaired fertility Sleep apnoea Gout Low back pain Breathlessness Increased anaesthetic risk Foetal defects (maternal obesity) RR= relative risk is the probability of an event in the treatment group divided by the probability of the event in the control group.
Psychological problems such as clinical depression are also associated with obesity. These problems may impact on quality of life and result in job discrimination and other employment difficulties.
Factors influencing body weight
A number of factors influence body weight and fat levels making some individuals more susceptible to obesity. Known contributors are:
• ethnicity • gender • age • hormonal state • genetic make up.
People from southern Asia are at a greater risk of developing central obesity and their cardiovascular risk increases rapidly at lower levels of obesity relative to western standards. For this ethnic group increased cardiovascular risk starts at a BMI of 21 23kg/m 2. . The WHO 2 has proposed new obesity guidelines for Asian populations with an obesity threshold defined as a BMI ≥25kg/m 2. Individuals with certain learning disabilities such as Down’s syndrome are also at higher risk of obesity (NHS Centre for Reviews and Dissemination 1997).
2 WHO 2000, Redefining obesity and its treatment.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 3
Clinically significant weight reduction
Most obese individuals would need to shed 30% of their body weight to reach their “ideal weight”. Such goals are generally unrealistic. However, much smaller reductions in weight can be accompanied by substantial health gains. A clinically significant weight reduction can be achieved with a body weight loss of 5 10% within six months. Many obese people experience difficulty in achieving this amount of weight loss in the timeframe through diet and exercise alone (Derosa et al. 2003). These people may benefit from the pharmacological interventions reviewed here.
Obesity and Injury
Weight loss may be viewed as a modality to improve general health as well as an aid to treatment and recuperation from accident or injury. A number of obesity associated health problems may be present at the time of accident or injury, or acquired afterwards, as a result of mobility and lifestyle changes. Such problems may impede recovery and rehabilitation and impede or limit the effectiveness of treatment.
For individuals recovering from accidents and injury, obesity related problems may significantly interfere with treatment and recovery. Obese claimants may be refused surgical treatment because of increased risks associated with anaesthesia. Existing obesity or obesity arising from lack of mobility, activity and exercise after an accident or injury may interfere with a claimant’s return to independence or full employment. It may also substantially increase the claimant’s recuperation period. Personal care requirements may also increase, particularly in relation to attendant care and additional health interventions may be required for co morbid conditions, which may be initiated or exacerbated as a consequence of obesity.
Obesity trends in NZ
A substantial proportion of the New Zealand population is obese. In 1997, the National Nutrition Survey reported that more than half of New Zealand adults were overweight with 17% classified as obese; it is estimated that by 2011 this will have risen to 29% (Wilson et al. 2001). Obesity rates in Maori and Pacific Islanders are higher than for the general New Zealand population with 27% of Maori men, 26% of Pacific men, 28 % of Maori women and 47% of Pacific woman reported as obese by the 1997 survey. Obesity in New Zealand children is also increasing and is of some concern.
Given these statistics it is highly likely that a significant number of ACC claimants will be obese or at risk of obesity at the time of accident or injury. Since some ethnic groups are more at risk of obesity than others, a correspondingly greater proportion of obese claimants, or claimants with recovery related obesity risk, may be expected in these susceptible groups. It is estimated that 75% of Pacific peoples in New Zealand are overweight; an immobilising accident or injury could put these claimants at risk of becoming obese while recovering from an accident or injury.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 4
The economic cost of obesity
The high prevalence of obesity in New Zealand and its associated co morbidities suggests that obesity is likely to command a significant portion of the total health care costs incurred by any government agency.
Three component costs have been identified:
• the direct treatment cost to the individual and the service provider
• the opportunity costs to the individual and society arising from premature death or attributable morbidity
• the indirect costs to the individual and society of lost work production due to absenteeism from work and premature death.
An overall assessment of the economic cost of obesity by the WHO suggested that it accounts for between 2 7% of total health care costs in developed countries. As a significant proportion of ACC claimants are likely to be obese or overweight at the time of injury or accident, additional costs arising from obesity related complications may be expected. The additional cost may be expected to arise from the:
• direct cost of obesity treatment
• indirect costs associated with delays in returning the claimant to full employment as a result of obesity related co morbidities or treatment and rehabilitation problems arising from morbid obesity.
Benefits associated with weight loss
Even a modest weight loss of 5 10% has been associated with significant health benefits in the obese patients. These benefits arise from improvements in obesity related co morbidities including blood pressure, blood glucose and cholesterol levels. It has been estimated that for every one kilogram reduction in body weight there is a:
• 0.05mmol/l decrease in total cholesterol
• 0.02mmol/l decrease in low density lipoprotein cholesterol (LDL cholesterol)
• 0.015mmol/l decrease in triglycerides
• 0.007mmol/l increase in high density lipoprotein cholesterol (HDL cholesterol).
Weight loss and dietary fat modification appear to have independent and additive effects on the reduction in serum lipids: the net favourable effect of weight loss seems to be greater than that of dietary fat modification as weight loss per se is only responsible for about 60% of the fall in LDL cholesterol and 70% of the fall in triglycerides (Muls et al. 2001). Every 1% reduction in glycated haemaglobin (HbA 1c ) decreases cardiac complications from 9 40% depending upon the population and diabetes type (Gokcel et al. 2001).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 5
Higher weight losses have greater benefits (Jung 1997); a 10kg weight loss may be expected to incur:
• 10mmHg decrease in systolic blood pressure (SBP)
• 20mmHg decrease in diastolic blood pressure (DBP)
• 91% reduction in angina symptoms
• 33% raise in exercise tolerance
• 10% fall in total cholesterol
• 15% fall in LDL cholesterol
• 30% fall in triglycerides
• 8% increase in HDL cholesterol
• >50% reduction in the risk of developing diabetes
• 30 50% fall in fasting blood glucose
• 15% fall in HbA 1c .
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 6
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 7 Review questions
The following questions guided the current review:
• What are the most appropriate measures used to classify obesity and how is obesity defined across different ethnic and age groups – is the BMI of >35, as set out in ACC's 2001 guidance, still an appropriate definition of obesity for all claimants?
• What is the relative effectiveness of phentermine, diethylpropion, orlistat, sibutramine and meal replacement plans for claimants requiring treatment to aid weight loss and what degree of weight loss can be expected and maintained with each product?
• How long do the products take to achieve a clinically important weight loss and what are appropriate time limits for treatment with the various product regimes?
• What are the circumstances in which the various products should be used and what are the potential barriers to implementation?
• How do the specified products compare in terms of safety, contraindications and patient acceptability?
• What is the relative cost effectiveness of specified products for claimants requiring treatment to aid weight loss?
• Do the effectiveness and cost effectiveness of the various products vary according to patients’ ethnicity, and if so, how?
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 8
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 9 Pharmacological treatments for obesity
Pharmacological treatments for obesity are limited. Most have the potential to produce undesirable side effects, additionally, amphetamines administered for the purposes of weight loss may have euphoric actions and carry the potential for abuse (Silverstone 1992). Some previously available products have recently been withdrawn from the market in a number of countries, for example:
• phenylpropanolamine, an appetite suppressant, has been associated with increased haemorrhagic stroke in women and withdrawn in the USA
• fenfluramine and dexfenfluramine – two centrally acting appetite suppressants acting predominantly by releasing serotonin, were also withdrawn from the market recently because of an association with pulmonary hypertension and coronary valve damage (Gardin et al. 2000).
Pharmacological interventions registered in New Zealand
There are currently four generic drugs prescribed for the treatment of obesity in New Zealand:
• phentermine
• diethylpropion
• orlistat
• sibutramine.
Other non pharmacological interventions include meal replacement products and plans and bariatric surgery. Surgery for obesity is not considered in this review.
Phentermine (Duromine TM , Umine Timedcaps 3)
Phentermine has been available since the 1960s with a reported worldwide exposure of more than 50 million prescriptions (Glazer 2001). It is available in New Zealand as a resin under the brand name Duromine (15 or 30mg) or in a sustained release hydrochloride form under the brand name of Umine Timedcaps (30mg).
Phentermine is an appetite suppressant chemically related to amphetamine. It has its main effect on the dopaminergic and noradrenergic nervous systems and its actions include central nervous system (CNS) stimulation and blood pressure elevation. The resin complex (Duromine) is insoluble until it reacts with the GI fluids; phentermine is then released from the resin throughout the GI tract over a period of 10 14 hours. Phentermine HCL sustained release capsules (Umine) has peak concentrations at 2.4 hours after ingestion. There is almost complete absorption of the drug but eventually 70 80% of the oral dose is excreted unchanged in the urine with the remainder metabolised by the liver. The half life of phentermine is about 25 hours.
3 Also marketed under the product names Ionamine, Fastin, Adipex in the USA.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 10
Phentermine use in children under 12 years, elderly patients and pregnant or breastfeeding women is not advised as safety and efficacy in these populations has not been established.
Dosage and administration: the starting dose is usually 30mg/d (milligrams per day), however, in light framed patients or those who have suffered side effects a dose of 15mg/d may be prescribed. Failure to achieve a weight reduction of 5% within 12 weeks of starting phentermine therapy is an indication for discontinuation of treatment. In order to reduce the risk of dependence, the maximum continuous period of treatment should not exceed 4 8 weeks.
Circumstances of use: phentermine may be used as a short term adjunct in a medically monitored comprehensive regimen of weight reduction based on exercise, calorie controlled diet and behaviour modification in patients with a BMI ≥30kg/m 2 who have not had an appropriate clinical response to an appropriate weight reduction program alone. Patients with obesity related co morbidities such as sleep apnoea, insulin resistant diabetes mellitus, pre diabetes mellitus or impaired glucose tolerance (IGT) or high cardiovascular risk status, and have a BMI of less than 30kg/m 2 may also be considered for treatment with phentermine.
Mode of action: phentermine is an amphetamine which operates by stimulating the release of noradrenaline and dopamine; it also inhibits monoamine oxidase (Lean 2001). Its central catecholamine mechanisms cause appetite suppression for 12 14 hours (Thearle and Aronne 2003).
Indications: phentermine 15mg and 30mg are indicated as a short term adjunct in a medically monitored regime of weight reduction in obese patients that includes a calorie controlled diet, exercise and behaviour modification. It may also be appropriate to use phentermine in overweight patients with an increased risk of morbidity from other medical conditions including sleep apnoea, insulin resistant diabetes or IGT and cardiovascular disease.
Contraindications: patients with pulmonary artery hypertension, heart valve abnormalities or heart murmur, moderate to severe hypertension, cerebrovascular or cardiac disease, hypersensitivity to sympathomimetic drugs, hyperthyroidism, agitated states or psychiatric illness, depression, major eating disorder, glaucoma and patients with a history of substance abuse or dependence (Medsafe 2003).
Warning and precautions: serious cardiac valvular disease and primary pulmonary hypertension (PPH) has been reported in patients who have taken phentermine and fenfluramine or dexfenfluramine combinations (“phen fen”) for weight loss. There have been no reported cases of valvular disease and only very rare cases of PPH in patients taking phentermine alone. There is also a theoretical risk of cardiac valvular disease if phentermine is combined with selective serotonin reuptake inhibitors (SSRIs) such as fluoxetine and paroxetine, ergot derived drugs and clomipramine. Phentermine should not be administered with monoamine oxidase inhibitors and should be used with caution in patients with mild hypertension and patients under treatment with hypertensive agents as it may cause some loss of blood pressure control. The ability to perform activities requiring mental alertness such as driving and operating machinery may be impaired while taking phentermine (Medsafe 2003).
Diethylpropion hydrochloride (Tenuate Dospan)
Diethylpropion is an appetite suppressant or anorexiant. It is rapidly absorbed after oral administration with 75 100% of the dose excreted in the urine as diethylpropion hydrochloride or its metabolites. The half life of diethylpropion in the plasma is about two hours and the excretory half life of the drug and its metabolites is approximately 10 hours.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 11
The use of diethylpropion in children younger than 18 years and pregnant or breastfeeding women is not advised as safety and efficacy in these populations has not been established. Secondary organic causes of obesity should be excluded by diagnosis.
Dosage and administration: the recommended dose of sustained release diethylpropion is 75mg once daily, one hour before meals. It is intended for short term intermittent use only. Courses may be given over a period of up to 12 weeks with intervening period of at least one month without treatment. To limit unwanted exposure, treatment should be continued only if there is satisfactory weight loss within the first four weeks of treatment. Insulin requirements may need to be altered.
Circumstances of use: diethylpropion may be used as a short term adjunct in a medically monitored comprehensive regimen of weight reduction based on exercise, calorie restriction and behaviour modification in obese patients with a BMI ≥ 30kg/m 2 who have not had an adequate response to an appropriate weight loss program of diet and/or exercise alone. Patients with a BMI of less than 30kg/m 2 with co morbidities including sleep apnoea, insulin resistant diabetes mellitus, pre diabetes mellitus or IGT or high cardiovascular risk status, may require medical assistance with weight loss. Such patients may also be considered for treatment with diethylpropion.
Mode of action: diethylpropion is a sympathomimetic amine, it acts as a CNS stimulant, increasing the heart rate, raising blood pressure and decreasing the appetite.
Indications: diethylpropion is indicated as a short term adjunct in a medically monitored comprehensive weight loss program based on calorie restricted diet, exercise and behaviour modification in obese patients who have not achieved an adequate weight reduction using diet and exercise alone. Overweight patients with sleep apnoea, insulin resistant diabetes, IGT or with a high risk of cardiovascular disease may also be considered for treatment.
Contraindications: severe hypertension or pulmonary artery hypertension, advanced arteriosclerosis, hyperthyroidism, known hypersensitivity to sympathomimetic amines, glaucoma, agitated states and patients with a history of drug abuse. Failure to achieve a weight reduction of 5% in a period of 12 weeks is an indication for discontinuation of treatment. Use in conjunction with other anorectic agents is contraindicated (Medsafe 2004b).
Warning and precautions: diethylpropion should not be given during or within 14 days of monoamine oxidase inhibitors. It is not recommended for patients who have taken any anorectic agents within the prior year or for patients with a heart murmur or valvular heart disease. The use of anorexiants has been associated with PPH and an increased risk of PPH with repeated courses of therapy cannot be excluded. Epileptic patients should be carefully monitored. Diethylpropion should be used with caution in patients undergoing general anaesthesia (Medsafe 2003).
Sibutramine hydrochloride (Reductil ®, Meridia ®)
Sibutramine is an orally administered drug that promotes satiety after eating and stimulates energy expenditure (Lean 2001; Nisoli and Carruba 2003). It is the first product of its type to be used for the management of obesity (Medsafe, sibutramine data sheet). The drug is rapidly absorbed following ingestion with maximal plasma concentrations at one hour and an elimination half life of about one hour. Most (77%) of the drug is absorbed from the GI tract and up to 85% of the oral dose is excreted in the urine and faeces.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 12
Sibutramine and its metabolites do not enhance dopamine release and have not been shown to have a potential for abuse (Schuh et al. 2000).
The long term effects of sibutramine on the mortality and morbidity associated with obesity have not been established. Its use in obese children under 18 years, patients over 65 years and pregnant or breastfeeding women is not advised as safety and efficacy in these populations has not been established.
Dosage and administration: the recommended starting dose is 10mg once daily with or without food; in clinical trials sibutramine was given in the morning. If there is less than 2kg weight loss after four weeks and the 10mg dose is well tolerated the daily dose may be increased to 15mg. Patients usually achieve maximum weight loss – i.e., 5 10% of initial body weight, after six months of treatment. The European Union label states that patients responding well may be treated for up to a year, in the USA treatment may be maintained for up to two years (Astrup and Toubro 2001). The frequency of adverse events may be reduced if sibutramine is administered intermittently – i.e., alternating with placebo or drug free periods (Wirth and Krause 2001).
Circumstances of use: sibutramine may be used for the management of obesity including weight loss and weight maintenance in patients with a BMI ≥ 30kg/m 2 or a BMI of ≥ 27kg/m 2 in patients who have diabetes, dyslipidaemia or hypertension. It is intended for use when patients have not adequately responded to appropriate weight reducing therapy such as hypocaloric diet and/or exercise alone – e.g., patients having difficulty in achieving or maintaining a 5% weight loss within three months.
Mode of action: sibutramine is a centrally acting monoamine reuptake inhibitor which blocks the pre synaptic reuptake of serotonin and noradrenaline thereby potentiating the anorexic effect of these two neurotransmitters in the CNS.
Indications: sibutramine 10mg and 15mg is indicated as adjunctive treatment within a weight management program for patients between 18 65 years with a BMI of 30kg/m 2 (i.e., obese) or patients with a BMI of 27kg/m 2 (i.e., overweight) who also have type 2 diabetes or dyslipidemia. Patients must have failed to lose weight on diet and exercise in the past.
Contraindications: patients with severe hepatic or renal dysfunction, hypersensitivity to sibutramine, organic obesity or history of major eating disorders, patients with psychiatric illness, drug or alcohol abuse, inadequately controlled hypertension or history of cardiovascular disease, narrow angle glaucoma, hyperthyroidism, benign prostatic hyperplasia with urinary retention and phaeochromocytoma (Medsafe 2004a).
Warning and precautions: blood pressure and pulse rate should be monitored while patients are taking sibutramine. The ability to drive a vehicle or operate hazardous machinery may be impaired when taking sibutramine (Medsafe 2004a).
Orlistat (Xenical ®4)
Orlistat is a novel anti obesity agent that inhibits the digestion of fat. Undigested trigylcerides are excreted in the faeces 24 to 48 hours after administration resulting in a calorific deficit that has a positive effect on weight control.
4 Xenical®, Hoffman Roche.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 13
Orlistat is minimally absorbed into the systemic circulation with approximately 97% of the administered dose excreted in the faeces, 83% is unchanged orlistat (Medsafe 2004c). The time to complete excretion is three to five days.
The long term effects of orlistat on the mortality and morbidity associated with obesity have not been established . Its use in children, adolescents below the age of 18 years, patients over 65 years and pregnant or breastfeeding women is not advised as safety and efficacy in these populations has not been established.
Dosage, administration: the recommended dose of orlistat is 120mg three times a day (t.i.d.) with each main meal. Higher doses have not been shown to provide additional benefit. The greatest weight loss occurs within the first six months of treatment. In the USA, treatment may be maintained for up to two years (Astrup and Toubro 2001).
Circumstances of use: orlistat may be used for weight loss, weight maintenance and prevention of weight regain in adults with a BMI ≥ 30kg/m 2. It should be used in conjunction with a low fat calorie controlled diet that is nutritionally balanced, rich in fruit and vegetables and contains approximately 30% of calories from fat distributed over three meals.
Mode of action: orlistat is a synthetic derivative of lipstatin a naturally occurring lipase inhibitor produced by Streptomyces toxytricini . It is a potent and specific irreversible inhibitor of pancreatic and gastric lipases that acts by bonding at the active site of lipases in the lumen of the GI tract. The inactivated enzyme is unable to hydrolyse ingested triglycerides into free fatty acids and monoglycerides for absorption. At the recommended dose and when diet contains approximately 30% of calorific intake from lipids, about 20g (180kcal) of fat is excreted in the daily stool. Orlistat may also influence dietary choices as high fat meals can lead to more severe GI adverse events.
Indications: orlistat is indicated as adjunctive treatment within a low fat calorie controlled program for obese patients between 18 65 years with a BMI of 30kg/m 2. Patients must have demonstrated weight loss of 2.5kg one month before its use.
Contraindications: patients with chronic malabsorption syndrome, cholestasis, known hypersensitivity to orlistat or any of its components (Medsafe 2004c).
Warning and precautions: because of a possibility of a decreased absorption of fat soluble vitamins A, D, E, K and beta carotene, the use of multivitamin supplements may be considered while patients are taking orlistat. High fat meals are likely to increase the possibility of GI events and patients should adhere to dietary fat intake guidelines. A reduction of plasma levels of cyclosporine 5 is possible after orlistat administration; the effect of amiodarone may also be reduced. Coagulation parameters of patients on concomitant oral anticoagulants should be monitored. There is no information relating to effects on ability to drive and used machines (Medsafe, orlistat data sheet). The additive GI effects of concomitant use of orlistat and olestra, a non absorbable dietary fat substitute commonly available in snack foods such as potato chips, has been reported (Heck et al. 2002).
5 Cyclosporine absorption may be reduced by approximately one third (Zhi et al., 2003).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 14
Meal replacement plans
Low calorie diets (LCDs) – i.e., diets in the range of 1200 1600kcal/day, are the cornerstone of modern weight control efforts. Meal replacement strategies and products are used by millions of consumers worldwide, however, there are no established definitions of meal replacement or partial meal replacement (PMR) plans, (Heymsfield et al. 2003).
The term ‘meal replacement’ is generally used in the scientific literature to cover beverages, pre packaged shelf stable and frozen entities and meal or snack bars. Most of these products are fortified with vitamins and minerals designed to be consumed in place of one or more regular meals. They may be an adjunct in a LCD. Eating, behaviour modification and physical exercise advice usually accompany commercial meal replacement programs.
Meal replacement nutritional supplements may be a useful weight loss therapy in obese individuals who:
• do not want to use weight loss drugs
• cannot tolerate the side effects of weight loss drugs
• cannot maintain eating habit change
• do not respond to weight loss drugs.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 15 Methodology
The main methodological tools for this review include a comprehensive literature search for published studies that fulfil pre determined criteria, a systematic review of eligible studies and a synthesis of the evidence guided by specific review questions (see page 7).
Systematic review
A systematic review is a carefully defined process that involves systematically locating, appraising and synthesising evidence from scientific studies in order to obtain a reliable overview. The literature search is comprehensive involving as many relevant sources as possible being accessed; the study selection, data extraction and data pooling are performed according to pre set criteria. The adherence to scientific principles sets the systematic review apart from traditional literature reviews making it less biased and more objective (O'Meara et al. 1998).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 16
Search strategy and information sources
Search strategy
A comprehensive literature search of major relevant bibliographic and review databases was undertaken together with additional searches for clinical trials, guidelines, government, and other evidence based materials. A range of subject headings and keyword searches were used to search indexed databases (see Appendix II). The search was not restricted by date, but was restricted to English language articles. In addition, a methodology filter was used to identify systematic reviews, randomised controlled trials (RCTs), and economic studies. This comprehensive strategy identified 777 potentially relevant journal references. After screening the abstracts of these references, 211 full text articles from peer reviewed journals were retrieved for analysis. All relevant health technology assessment (HTA) resources were also examined and five international HTA reports of pharmacological treatment of obesity were retrieved.
Principal sources of information
The following databases were searched using the search strategy outlined in Appendix II. The searches were carried out during June and July 2004.
Bibliographic databases
Medline
Cinahl
Embase
Web of Science
Current Contents
Toxnet
Cochrane Controlled Trials Register
Review databases
Cochrane Database of Systematic Reviews
Database of Abstracts of Reviews of Effects
Health Technology Assessment Database
NHS Economic Evaluation Database
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 17
Study selection criteria
The FDA guidelines for the clinical evaluation of weight control drugs and the selection criteria set out in the recent HTAs of obesity treatment guided the determination of inclusion and exclusion criteria for this review. All studies of pharmacological agents included in the review were required to be randomised, placebo or active control double blind studies with weight loss as the primary outcome and obese participants as the primary study population. Additionally, studies had to be written in English and published between 1996 2004.
Table 3. Inclusion and exclusion criteria for review studies
Inclusion Criteria Study design Randomised placebo or active control, double-blind studies or HTAs or systematic reviews or meta-analyses of randomised, placebo-controlled, double-blind trials or RCTs, double blind, comparative drug trials, OR for meal replacement, a RCT against standard meal or diet plan. Primary study purpose Weight loss Study size Unrestricted Study duration/ length of follow-up Any duration of treatment. Follow–up of at least 12 months for long-term benefits, any follow- up length for short-term benefits. Participants Individuals with a BMI of at least 30kg/m 2, no age restriction. Interventions Licensed drugs, sibutramine, orlistat, phentermine, diethylpropion or meal replacement plans/products with or without co-interventions such as diet counselling, deficit diets, dietary advice sheets, education and encouragement to exercise if appropriate. Outcome measures/endpoints Primary endpoints - actual weight loss, relative weight loss (% of body weight or % excess over ideal weight or change in BMI), changes in central obesity. Secondary endpoints - measurement of obesity related risk factors – e.g., lipids, blood pressure and glucose tolerance. Health economics All economic and cost evaluations Language English language or English translation available Study dates/publication date Jan 1996-July 2004 Publication type Studies published in peer-reviewed journals or manufacturer’s unpublished material if relevant and available. Exclusion criteria Pregnant or breastfeeding mothers Abstracts and case reports Animal studies Non-clinical studies Preliminary exclusion of overweight +co-morbidities Quasi randomised, open label and cross-over trials Studies recruiting patients with eating disorders such as anorexia or bulimia nervosa Single RCTs with a crossover design* * Excluded because of (i) the possibility of carry over effects and (ii) because they are excluded from most meta-analyses because of their inadequacy in terms of estimating effect size (Kim et al. 2003).
Methods of the review
Studies were not limited in duration but in the analysis of results studies were considered in two groups to reflect the primary needs of ACC claimants:
• short term studies – i.e., treatment duration < 6 months
• long term studies – i.e., treatment durations ≥ 6 months.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 18
Further stratification by co morbidity was carried out where feasible and obese patients with:
• type 2 diabetes
• hypercholesterolemia
• hypertension were analysed separately to assess differences in need, weight loss goals and expectations that may be required by claimants with obesity related co morbidities.
Levels of evidence
The evidence was graded using the Scottish Intercollegiate Guidelines Network (SIGN) instrument, see Table 4.
Table 4. Scottish Intercollegiate Guidelines Network Evidence levels
Level Type of evidence/study Ia Evidence obtained from meta-analysis of randomised controlled trials Ib Evidence obtained from at least one randomised controlled trial IIa Evidence obtained from at least one well designed controlled study without randomisation IIb Evidence obtained from at least one other type of well-designed quasi-experimental study III Evidence obtained from well-designed non-experimental descriptive studies, such as comparative studies, correlation studies and case control studies. IV Evidence obtained from expert committee reports or opinions and/or clinical experience of respected authorities.
However, only evidence from RCTs, systematic reviews, and meta analysis of RCTs or HTAs including RCTs was considered in this review, limiting the evidence levels to Ia or Ib.
The main purpose of randomisation is to minimise bias, however, the quality of randomised studies can vary. Where the normal quality criteria are not met a number of biases may occur, these can be minimised in a number of ways:
• selection bias full explanation of the method of randomisation, including concealment
• attrition bias full reporting of the number and type of withdrawals and drop outs
• intention to treat (ITT) analysis preserves the baseline comparability between treatment groups achieved by randomisation
• detection bias blinding of outcome assessor and blinding of participants.
The robustness and validity of the eligible studies was further determined from the degree of bias minimisation based on the four criteria above. The most useful and reliable study was considered to be a well designed and properly conducted randomised, placebo controlled, double blind clinical trial or a meta analysis of such studies. Where blinding is not possible because of the nature of the intervention – e.g., meal replacements, a well designed and properly conducted RCT or meta analysis was considered to be the most useful type of study for the purpose of this review.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 19 Results
Meal replacement plans: weight loss and co-morbidity risk reduction
Meal replacement eating plans have not been critically evaluated for safety and efficacy until very recently. The first meta analysis evaluating RCTs of PMR plans and products was published in 2003 (Heymsfield et al. 2003). In this study a search of Medline, Embase and the Cochrane Clinical Trials Register between 1960 to January 2001 for clinical trials with meal replacements or meal plans, identified 276 potential publications of which 30 were meal replacement as defined in the protocol. Only six studies finally met the review criteria, 24 were excluded due to one or more of the following:
• lack of a control arm
• less than 12 weeks duration
• not meal intervention as defined in protocol
• inclusion of subjects <18 years
• inclusion of subjects BMI <25kg/m 2
• use of the anorexic drug diethylpropion.
For the purpose of the Heymsfield meta analysis, a PMR plan was defined as a program that:
“ includes one or more meals replaced by a commercially available, calorie reduced product(s) that are fortified with vitamins and minerals and at least one daily meal consisting of regular foods. As a LCD, the plan’s calorie content should be >800<=1,600kcal/day.”
Primary data were available from the six eligible RCTs. These data were used in a meta analysis and pooling analysis, see Table 5.
Table 5. Weight loss results at 3 months and one year for an unstratified and stratified study population ΛΛΛ (PMR- Significance of ΛΛΛ (PMR- Significance of RCD) PMR-RCD RCD) PMR-RCD N= kg(s.e.) difference N= kg(s.e.) difference 3 months p= 12 months p= Analysis of all participants - unstratified Meta-analysis Random effects 403 2.60(0.96) 0.006 219 2.43(1.65) 0.142 Meta analysis Fixed effects 403 3.01(0.33) <0.001 219 3.39(0. 72) <0.001 Pooled analysis of completers 403 2.54(0.37) <0.001 219 2.63(0.88) 0.003 Pooled analysis LOCF 485 2.39(0.35) <0.001 485 2.86(0.46) <0.001 Analysis - stratified by diabetic status Non-diabetic completers 305 2.79(0.37) <0.001 193 3.17(0.99) 0.002 Non-diabetic LOCF 367 2.67(0.35) <0.001 367 3.56(0.50) <0.001 Diabetic completers 98 2.46(1.84) 0.185 26 2.76(2.00) 0.183 Diabetic LOCF 118 2.62(1.89) 0.167 118 1.52(1.89) 0.424 Λ (PMR-RCD) = weight loss difference between the treatment (PMR) and the control (RCD) group measured from baseline body weight to body weight at the time of assessment. PMR = partial meal replacement, RCD = restricted calorie diet.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 20
For detailed results see Meal Replacement in Evidence Table 1, Heymsfield, 2003.
All methods of analysis indicated a significantly greater weight loss for individuals following the PMR plan than those following the conventional reduced calorie diet. On average, the PMR group lost 2.54 kilograms more than the control group after three months of treatment and 2.43 kilograms more after 12 months. The proportion of subjects losing ≥ 5% of initial body weight was 33% for the RCD group and 72% for the PMR group (p<0.001); at 12 months the proportions were 33% and 74% (p<0.001).
After stratification according to diabetic status, the non diabetic completers lost on average 2.79 kilograms more than the control group after three months (p<0.001) and 3.17 kilograms more after 12 months (p=0.002). The much smaller diabetic completers lost on average 2.46 kilograms more than the control at three months and 2.76 kilograms more than the control at 12 months. Neither difference reached significance, however, the number of diabetic cases was small and the standard error for the group large.
All of the risk factors for obesity related co morbidities showed improvement from their baseline values in both the treatment and control groups (p<0.001), however, there was no significant additional effect of PMR on risk factor improvement with the exception of plasma insulin (p<0.001). The authors concluded that the magnitude of the weight loss that could be achieved with PMR at 12 months was within the range observed for pharmacological agents and the range known to lower obesity related risk factors. They attributed the success of PMR to a general preference by participants for structured weight loss plans which promoted improved behavioural compliance, increased nutritional knowledge, more regular meals and less snacking. No reported adverse events were attributed to either the PMR plan or the control.
A comprehensive literature search for the current review carried out in July 2004 identified a further product and randomised trial that was eligible for inclusion. The randomised trial comprised a study of 100 overweight or obese participants (Allison et al. 2003) in which half were randomly assigned to a soy meal replacement program and half to a 1200kcal exchange diet program. Both groups received dietary counselling and a pamphlet describing good weight loss practices. After 12 weeks, the treatment group lost significantly more weight from the baseline than the control group 7.00kg vs. 2.90kg (p<0.001, ITT analysis). Fat mass was also significantly lower in the treated group 4.3 vs 1.4 (p=0.003). Obesity related risk factors showed reductions in both groups but the magnitude of the effect was variable with large standard errors; LDL cholesterol reductions were significantly greater at all time periods in the treatment group. Treatment was reported to be well tolerated with no serious side effects. For further details of results, see Meal Replacement in Evidence Table 2 (Allison et al. 2003).
Phentermine hydrochloride: weight loss and co-morbidity risk reduction
The efficacy and safety of phentermine has been assessed primarily in short term studies of up to 12 weeks duration. The longest study was a randomised, double blind, placebo controlled study of 100 subjects undergoing phentermine therapy for 36 weeks (Munro et al. 1968) and reported in a recent systematic review by Glazer (2001). No large scale, long term studies of phentermine for weight loss have been performed (Thearle and Aronne 2003).
The systematic review of pharmacotherapy of obesity reported by Glazer in 2001 included a comparative analysis of all randomised, placebo controlled, double blind trials of nine months or more duration.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 21
Trials limited to diabetic patients were excluded from this comparative study. Four phentermine studies were reported (Munro 1979; Munro et al. 1968; Steel et al. 1973; Truant et al. 1972; Weintraub et al. 1984), see Table 6a.
Table 6a. Phentermine effectiveness studies
Study Year Dose Duration Weight P value † Loss (kg) Munro 1986 30mg 36 weeks 12.2 <0.001 4/4 36 weeks 13.0 <0.001 Weintraub 1984 20 weeks 11.3 <0.01 Truant 1972 16 weeks 8.8 <0.01 Steel 1973 36 weeks 12.0 - 4/4 =4 weeks of phentermine followed by 4 weeks not on phentermine therapy. † compared to placebo
Data from these early studies suggested that phentermine was well tolerated. Adverse events were reported as “minor” with between 3 8% of patients affected. Stimulant effects such as agitation and insomnia were noted. Overall, phentermine was reported as efficacious, with a significantly greater weight loss than the placebo group reported in each study. Weight loss relative to baseline body weight was not reported and the results of the studies were not pooled. When the percentage weight loss in excess of placebo was compared across all of the reported drugs, phentermine appeared to have a weight loss advantage (8.1% of baseline weight lost) over sibutramine (5.0% of baseline weight lost), orlistat (3.4% of baseline weight lost) and diethylpropion ( 1.5% of baseline weight – i.e., weight gain). However, the high between group variation in completion rates and other study variables make these comparisons of doubtful value.
In 2002, Haddock et al., in a meta analysis of four decades of published randomised trials of pharmacotherapy for obesity, identified nine studies of phentermine published between 1969 1992. The resulting meta analysis of six eligible randomised studies, comprising 386 participants with a mean follow up of 13.2 weeks, reported a significantly greater weight loss for phentermine than the placebo (effect size <0.60). There was a mean weight loss of 2.8kg for the placebo group and 6.3kg for the group treated with phentermine. The overall difference in weight loss averaged 3.6 kilograms over the study period, see Table 6b. Most of the patients (>85%) in the reported trials were female.
Table 6b. Phentermine vs. placebo post treatment outcomes, Haddock et al., 2002
Drug Number post test Mean Weight loss (kg) Duration D-P Kg average dose Studies Subjects Placebo Phentermine Placebo Phentermine (Wks) (range) (dose range) (range) (range) (range) (range) Phentermine 27.7mg/d 6 386 13.2 (2.24) 29.4 (12-74) 32(15-76) 2.8 (1.5-5.2) 6.3(3.6-8.8) 3.6(0.6-6.0) (15-30mg/d)
When the treatment effect size for phentermine was compared to that obtained for other drugs in the study, the effect size for phentermine was second only to sibutramine and higher than either orlistat or diethylpropion. However, all confidence intervals overlapped and there were significant design differences between the studies.
There was no overall effect of treatment length for drugs evaluated in this study, suggesting an early treatment impact for most of the drugs. However, phentermine correlation results for this variable were large and it was suggested that treatment length may influence phentermine’s effect size.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 22
Follow up outcome analysis, carried out after the formal study completion and treatment had stopped, reported a placebo subtracted weight losses of 2.43kg and 2.37kg for phentermine and sibutramine respectively with unweighted effect sizes of 0.81 and 1.05 respectively. These results were based on a very small number of studies. Overall, the size effect was modest (less than 0.80) and the placebo subtracted weight loss was also modest <4kg.
Phentermine combinations In 1996, phentermine which was introduced into clinical practice more than 25 years ago, was administered with the newly introduced fenfluramine in the combination known as “Phen Fen”. Following reports of cardio valvopathy associated with this combination fenfluramide was withdrawn from the market the following year (Rothman and Baumann 2000; Rothman et al. 2000). There was no evidence that phentermine was responsible for these serious adverse events (Hensrud et al. 2003). Studies that included this drug combination have not been included in this review.
In 1999, Bradley et al. examined the use of phentermine and bupropion in combination for weight loss. This randomised, double blind, placebo controlled trial recruited 44 obese (>30kg/m 2) patients and ran for six months. Patients were randomly assigned in a double blind fashion to receive either phentermine 30mg + placebo or phentermine 30mg + bupropion SR (sustained release), 150mg twice daily. Both groups followed a 1200 calorie diet which included liquid meal replacement products. After six months, both groups of patients had lost more than 12% of their initial body weight. There were no serious complications or adverse effects noted over the period. Although there was no weight loss advantage in adding bupropion to phentermine, the Beck Depression Index scores improved in the groups taking bupropion. The study investigators concluded that bupropion might be a useful adjunct to phentermine in obese patients subject to mood swings and/or sub clinical depression.
Diethylpropion: weight loss and co-morbidity risk reduction
Diethylpropion is available as Tenuate 25mg t.i.d. and Tenuate Dospan 75mg extended release taken once a day. The longest eligible study was carried out by Silverstone in 1968 and reported in a systematic review of long term pharmacotherapy of obesity by Glazer (2001). Two shorter trials were also reported in this comparative review, see Table 7.
Table 7. Diethylpropion effectiveness studies, Glazer et al., 2001 Follow- Weight loss kg D-P Study Year up P value period Drug Placebo Kg Silverstone 1968 6 month 7.0 8.7 -1.7 NR 1 year 8.9 10.5 -1.5 NR Deramos 1964 6 months 7.8 1.9 5.9 >0.05 McKay 1973 6 months 11.7 2.5 9.2 <0.01
Data were reported in these trials for a total of 30 patients treated with diethylpropion for six months and five patients treated for 12 months. Results varied considerably for these small studies with one study reporting inferior weight loss for diethylpropion compared to the placebo, one reporting higher, but non significant, weight loss for diethylpropion compared to the placebo and the other reporting significantly higher weight loss for the diethylpropion compared to the placebo (McKay 1973). Diethylpropion was reported to have few stimulant related adverse events but weight loss beyond six months was not observed with these patients.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 23
Relatively rapid tolerance to the drug’s anorectic effects was reported. In a comparison of the efficacy of diethylpropion with phentermine, sibutramine and orlistat, diethylpropion was the only drug to record a weight gain. However, the between study heterogeneity was high and the diethylpropion study comprised only five subjects making the validity of the comparison doubtful and its usefulness limited.
In 2002, Haddock et al., in an analysis of four decades of published randomised trials of pharmacotherapy for obesity, critically reviewed 13 studies of diethylpropion published between 1965 1983 (Haddock et al. 2002), see Comparative Studies in Evidence Table 27. Nine of these studies were employed in a meta analysis, see Table 8 below.
The majority of patients were female (88%) and most studies used some form of life style management program. Some of the studies included in the analysis had no weight loss or had weight gain against the placebo as evidenced by the negative values in the drug placebo value range. Overall, however, diethylpropion produced a greater weight loss than the placebo with an effect size of approximately 0.6. The 95% confidence intervals for diethylpropion overlapped with sibutramine, which was the drug that produced the largest effect (i.e., > 0.80), suggesting that the effect size difference between the drugs was not significantly different.
Table 8. Diethylpropion vs. placebo post treatment outcomes, Haddock et al., 2002
Drug Duration Number post test Mean Weight loss (kg) D-P Studies (dose) (wks) Placebo DP Placebo DP (kg)
+3.00 Diethylpropion 9 17.6(6-52) 21.2(5-32) 18(4-29) 3.5 (-04-10.5) 6.5(1.9-13.1) (-1.6-11.5) DP=diethylpropion, D-P=drug minus placebo
When compared with 12 weeks of phentermine treatment (n=50), diethylpropion (n=49) patients lost 6.3kg against 8.3kg for phentermine (ns) with an effect size of 0.574. An analysis incorporating all drugs suggested that increasing length of treatment did not lead to more weight loss and overall, weight loss attributed to pharmacological intervention for obesity was considered to be modest – i.e., <4.0kg. No data were reported on the effect of diethylpropion of obesity related co morbidities in this study.
Orlistat: weight loss and co-morbidity risk reduction
A comprehensive literature search for randomised, placebo controlled, double blind trials or systematic reviews, meta analyses or HTAs of such studies published between January 1996 July 2004 identified 10 systematic reviews (Arterburn and Hitchcock 2001; Avenell et al. 2004; Foxcroft and Milne 2000; Glazer 2001; Haddock et al. 2002; Hensrud et al. 2003; Heymsfield et al. 2000; Leung et al. 2003; O'Meara et al. 2001; Padwal et al. 2004) and seven additional trials not included in the systematic reviews (Derosa et al. 2003; Halpern et al. 2003; Hanefeld and Sachse 2002; Krempf et al. 2003; Muls et al. 2001; Rissanen et al. 2001; Torgerson et al. 2004), see Table 9 (overleaf). These studies are further summarised in Evidence Tables 3 13, 26 and 27.
There was a good deal of overlap between the systematic reviews and HTAs in terms of the studies reviewed. All limited their reviews to randomised, placebo controlled, double blind trials of obese or overweight and obese individuals. The trials comprised mixed risk populations including healthy obese, diabetic and hypertensive patients. Two of the reviews permitted studies of any duration to be included in the review (Hensrud et al. 2003; O'Meara et al. 2001).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 24
Table 9. Summary of publications reporting the effectiveness of orlistat for weight loss in overweight and obese participants Study Size Study population Studies* Patients Study type Code Duration Health technology assessment, systematic reviews, Cochrane reviews, meta-analyses Avenell 2004 Obese Mixed 8 3885 HTA LT At least 1 year Padwal 2004 Obese Mixed 11 6021 CR LT At least 1 year Hensraud 2003 Obese Mixed 10 NR HTA MT Any Arterburn 2003 Obese mixed 6 1836 SR MT Any Leung 2003 Obese mixed 7 1830 SR MT More than 6 months Glazer 2001 Obese mixed 5 742 SR LT At least 9 months treatment Haddock 2001 Obese mixed 6 NR MA ST 48 weeks O’Meara 2001 Obese Mixed 11 5124 SR MT Any Foxcroft 2000 Obese mixed 3 551 SR LT More than 1 year Heymsfield 2000 Obese mixed 3 675 SR/MA LT 2 year follow-up Randomised, placebo-controlled, double-blind clinical trials Torgeson 2004 Obese diabetic 1 3305 RPCDBT LT 4 years Halpern 2003 Obese1 diabetic 1 365 RPCDBT ST 24 weeks Krempf 2003 Health obese 1 696 RPCDBT LT 10 months Dersosa 2003 Obese hypercholesterolemic 1 99 RPCDBT LT 1 year Rissanen 2001 Healthy obese 1 55 RBCDBT LT 1 year Hanefield 2002 Obese diabetic 1 492 RPCDBT ST 48 weeks Muls 2001 Obese hypercholesterolemic 1 441 RPCDBT ST 24 weeks + 24 weeks * Published studies, LT=long-tern, ST=short-term, MT=mixed term, RPCDBT=randomised placebo-controlled, double-blind trial, CR=Cochrane review, SR=systematic review, MA=meta-analysis, HTA=health technology assessment .
Most of the reviews were interested in the long term effects of orlistat in the treatment of obesity and therefore restricted their studies by duration of treatment or follow up. Most required participants to be treated for at least one year.
The patient populations varied considerably between studies, as did the reporting of results. Most studies reported absolute weight loss or % change from baseline rather than the more clinically useful weighted mean difference between treatment and control and the proportion of patients with ≥ 5% and ≥ 10% weight loss. Later studies intended to report clinically useful endpoints more often than earlier studies. Study results were reported for three patients groups:
• otherwise healthy overweight/obese patients
• patients with type 2 diabetes/glucose intolerance
• patients with hypercholesterolemia.
One study was a Cochrane Review (Padwal et al. 2004) and one was a National Institute for Clinical Effectiveness (NICE) HTA (O'Meara et al. 2001). The three most recent reviews (Avenell et al. 2004; O'Meara et al. 2001; Padwal et al. 2004) assessed the evidence presented in a large number of randomised, placebo controlled, double blind trials with total populations of 3,885, 5,124 and 6,021 respectively.
Overall effectiveness of orlistat in obese patients (mixed risk) The most recent meta analysis of nine randomised placebo controlled, double blind studies of the effect of orlistat for the treatment of obesity was conducted by Avenell et al., (2004) for the NHS R&D HTAprogram in the UK. The results of the meta analysis of these studies are summarised in Table 10 (overleaf).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 25
Table 10. Meta-analysis of RCTs for Orlistat, long-term studies ≥ 1 year (NICE HTA, Avenell, May 2004)
Total Weighted Mean Difference between drug and placebo, over all studies, fixed effects (95%CI) Fasting Study/ Weight TC LDL-C HDL-C TG HBA Ic plasma DBP SBP Year (kg) mmol/l mmol/l mmol/l mmol/l % glucose mmHg mmHg mmol/l
Orlistat, 360mg/d + diet vs. placebo + diet, 8 RCTs at 12 months
-3. 01 -0.34 -0.29 -0.03¥ -0.03 -0.17 -0.24 -1.64 -2.02 Broom, 2001 (95%CI, -3.48 to–2.54) (95%CI, -0.41 to–0.27) (95%CI, -0.34 to–0.24) (95%CI, -0.05 to–0.01) (95%CI, -0.04 to–0.10) (95%CI, -0.24 to–0.10) (95%CI, -0.34 to–0.14) (95%CI, -2.20 to–1.09) (95%CI, -2.87 to– Π Π Π Π Π Π Π Π Davidson, 1999 2 = 15.08 (p=0.04) 2 = 4.27 (p=0.64) 2 = 3.53 (p=0.74) 2 = 7.07 (p=0.22) 2 = 24.11 (p=0.0002) 2 = 8.23 (p=0.02) 2 = 13.02 (p=0.02) 2 = 19.11 (p=0.004) 1.17) Orlistat n=2166 Orlistat n=1509 Orlistat n=1059 Orlistat n=1250 Orlistat n=1399 Orlistat n=1399 Orlistat n=1399 Orlistat n=2056 Π2 = 10.09(p=0.12) Finer, 2000 Placebo n=1719 Placebo n=1496 Placebo n=1496 Placebo n=1233 Placebo n=1388 Placebo n=1388 Placebo n=1388 Placebo n=1611 Orlistat n=2056 Hauptman, 2000 Studies=8 Studies=7 Studies=7 Studies=6 Studies=6 Studies=6 Studies=6 Studies=7 Placebo n=1611 Studies=7 Hollander, 1998 Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Lindgarde, 2000 z=12.53(p<0.00001) z=9.67p(<0.00001) z=11.28(p<0.00001) z=2.43(p<0.01) z=0.77(p=0.44) z=4.67(p<0.00001) z=4.77(<0.00001) z=5.80(<0.00001) Test for overall effect Rossner, 2000 z=4.65 (<0.00001) Sjostrom, 1998
Orlistat, 360mg/d + diet vs. placebo + diet, 3 RCTs at 24 months
Hauptman, 2000 -3.26 -0.21 -0.22 -0.03¥ -0.04 -0.15 -1.20 -1.42 Rossner, 2000 (95%CI, -4.15 to–2.37) (95%CI, -0.34 to–0.09) (95%CI, -0.31 to–0.13) (95%CI, -0.07 to–0.0) (95%CI, -0.07 to–0.005) (95%CI, -0.24 to–0.07) (95%CI, -2.28 to–0.11) (95%CI, -3.08 to– Π Π Π Π Π Π Π Davidson, 1999 2 = 0.50 (p=0.48) 2 = 1.04 (p=0.60) 2 = 5.92 (p=0.05) 2 = 0.31 (p=0.86) 2 = 10.14 (p=0.006) 2 = 4.15 (p=0.13) 2 = 9.39 (p=0.002) 0.24) Orlistat n=451 Orlistat n=557 Orlistat n=555 Orlistat n=557 Orlistat n=557 Orlistat n=557 Orlistat n=451 Π2 = 0.22(p=0.64) Placebo n=448 Placebo n=537 Placebo n=536 Placebo n=537 Placebo n=537 NR Placebo n=537 Placebo n=448 Orlistat n=451 Studies=2 Studies=3 Studies=3 Studies=3 Studies=3 Studies=3 Studies=2 Placebo n=448 Studies=2 Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect z=7.18(p<0.00001) z=3.26p(<0.001) z=4.90(p<0.00001) z=1.84(p<0.07) z=0.68(p=0.55) z=3.50(p=0.0005) z=2.16(p=<=0.03) Test for overall effect z=1.68 (<0.09) ¥ =decrease in HDL in treatment group unfavourable, this result favours the control group.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 26
Over a period of 12 months, obese patients treated with orlistat lost on average 3.01 kilograms more than the placebo group. The difference between the two groups increased slightly at 24 months with orlistat treated patients losing on average 3.26 kilograms more than the placebo group; at both time points the difference was significant.
At 12 months, there was also significant and favourable differences reported for the orlistat treated group in terms of total cholesterol, LDL cholesterol, blood pressure, HbA 1c and fasting plasma glucose. A non significant decrease in triglycerides was reported in favour of the treatment group and an unfavourable decrease in HDL cholesterol. This pattern was repeated in the assessments undertaken at 24 months with the exception that SBPreduction was less marked and did not reach significance. HbA 1c was not reported at 24 months.
A chi squared test for between study heterogeneity revealed significant heterogeneity in weight loss, triglyceride, HbA 1c , fasting plasma glucose and DBP at the 12 month assessment. This heterogeneity persisted between a much smaller number of studies for triglycerides and DBP at 24 months. The high between study heterogeneity is likely to have arisen as a consequence of the inclusion of some studies with diabetic and hypercholesterolemic participants as well as studies comprising only healthy obese participants. Where possible, these sub groups have been reported separately in the current review.
Clinical-effectiveness of orlistat in otherwise healthy populations Three studies (two RCTs and a meta analysis of three RCTs) reported differences in weight loss between orlistat and the placebo in healthy obese populations. All of the studies reported the difference in mean weight loss or % weight lost from baseline between the treatment and placebo groups. In addition, two studies (Krempf et al. 2003; Torgerson 2004) reported the proportion of patients with clinically important weight loss – i.e., weight loss ≥ 5% and ≥ 10% of their original weight. The reported results are shown in Table 11.
Table 11. Orlistat 360mg/d weight loss in otherwise healthy obese populations Study Year No of No of Mean difference Mean Difference % with % with studies subjects in % change WL (kg) WL ≥5% WL ≥10% Krempf (RCT) 2003 1 95 -2.6* -2.9* NR NR -3.5** -3.6** 58 NR Torgerson (RCT) 2004 1 3305 NR -4.4 a 73 a 41 a NR NR 53 b 26 b Heymsfield (MA) 2000 3 675 -2.9* -2.5* NR NR * at one year, intention to treat (ITT) ** 18 months ITT, a difference at one year non-diabetic b difference at four years WL =weight loss
Reporting was variable across the three studies and no study reported all of the relevant weight loss parameters. Participants treated with orlistat lost a larger proportion of their original weight (2.6 2.9% at 12 months, 3.5% at 18 months) than those treated with the placebo, they also lost a greater amount of weight overall. Only two of the studies reported the proportion of patients who lost ≥ 5% of their original weight (Krempf et al. 2003; Torgerson et al. 2004). At one year, 58 73% of treated patients had lost 5% or more of their baseline weight this was a significantly higher proportion than the placebo treated group.
Clinical effectiveness of orlistat in patients with hypercholesterolemia Two RCTs, not reported in the systematic reviews, examined the effectiveness of orlistat in reducing lipid and blood pressure parameters in patients with hypercholesterolemia, see Table 12 (overleaf).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 27
The approach to the reporting of the results and the number of parameters reported differed in the two studies with one reporting mean placebo minus drug differences, the other reporting differences for each group from the baseline value. However, both studies indicated a significant improvement in LDL cholesterol, and total cholesterol. There was also evidence of a significant improvement in systolic and diastolic blood pressure. There was no evidence for improvement in the trigylceride and HDL cholesterol profiles.
Table 12. Orlistat 360mg/d weight loss in patients with hypercholesterolemia, results after one year No of Weight loss LDL-C HDL-C Total-C TG SBP DBP Study Year Studies Group mg/dl mg/dl mg/dl mg/dl mmHg mmHg (subjects) Muls LMS difference from placebo on % difference from baseline 2001 1 (411) P-O RCT -3.03kg* -10.00 -7.56 -8.37 -0.63 NR NR Derosa Mean % change from baseline for each group RCT 2003 1(23) Placebo -7.6kg** -10.8 +2.4 -12.1 -14.8 -3.0% -2.4% 1(25) Orlistat -8.6kg** -19.0 +2.3 -15.0 26.5 -4.6% -4.7% Bold = significant difference from placebo *difference Placebo-Orlistat ** from baseline, LMS = least mean square
Patients with type 2 diabetes or impaired glucose intolerance A European Medicine Evaluation Agency (EMEA) report of orlistat, comprised a meta analysis of seven randomised, double blind, placebo controlled trials of one to two years involving 4,188 diabetic and non diabetic patients, see Table 13. A higher proportion of diabetic and non diabetic patients receiving orlistat treatment experienced clinically significant weight loss (≥ 5% or ≥ 10% of baseline weight) than the placebo group. This difference was significant in all comparisons except for a weight loss of ≥ 10% for diabetic patients receiving orlistat vs. placebo. Overall, diabetic patients consistently lost less weight than their non diabetic counterparts regardless of the treatment they received.
Table 13. Weight loss with orlistat: results of 1 year clinical trials, European Medicine Evaluation Agency.
Non-diabetic Type2 diabetes Orlistat* + diet Placebo+ diet Orlistat* + diet Placebo+ diet Weight loss P value P value N=1561 N=1119 N=163 N=159 ≥5% weight loss 45.3% 23.4% <0.001 30.2% 13.2% <0.001 ≥10% weight loss 20.2% 8.3% <0.001 9.3% 4.4% ns * 360mg/d
Weight loss in diabetic and non diabetic participants was also reported in a systematic review by O'Meara et al. (2001). In a pooled analysis, non diabetic participants who had taken orlistat 360mg/d had a mean weight loss difference (placebo minus drug) of 2.9kg compared to 1.8kg for the diabetic group; both groups lost significantly more weight than the placebo group.
A meta analysis of studies examining glucose tolerance and changes in status from randomisation reported the weight loss achieved with orlistat 360mg/d (Heymsfield et al. 2000), see Table 14.
Table 14. Orlistat 360mg/d in diabetic patients or patients with glucose intolerance Glucose % Studies WL - Change % Study Year intolerant Improvement (no subjects) from baseline (kg) Progressing Participants IGT-NGT
Heymsfield Placebo group 3.79 (±4.1) 2000 2 (675) 7.6 49 Meta-analysis Orlistat group 6.27 (±0.41) 3.0 72 WL=weight loss, IGT=impaired glucose tolerance, NGT=normal glucose tolerance, kg=kilogram
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 28
Both placebo and orlistat treated groups exhibited a significant change in weight from their baseline body weight. The treatment group, however, had a much higher total weight change, a higher proportion of patients improving their glucose tolerance and only half as many patients with IGT at baseline progressing to diabetic status.
Two short term RCTs not otherwise reported in the systematic reviews, examined the effects of orlistat in obese diabetics (Halpern et al. 2003; Hanefeld and Sachse 2002), see Evidence Tables 8 and 9. In a 24 week study of 365 obese, non insulin dependent diabetic patients (Halpern et al. 2003), the difference in weight loss between the orlistat and placebo group was 1.6kg (least squares mean difference, p=0.0003, ITT analysis) with 30% of orlistat and 17% of placebo patients losing ≥ 5% of their initial body weight (p=0.003). There were also significant differences (least square, mean difference, orlistat minus placebo) in total cholesterol (p=0.0001) and LDL cholesterol (p=0.0002). HDL cholesterol levels improved in the placebo group but not in the orlistat group (p=0.038). There was significant improvement in glycaemic control in the orlistat group with significant decreases in HbA 1c (p=0.04), fasting plasma glucose (p=0.036) and post prandial glucose (p=0.05).
In a 48 week study of 492 overweight patients with type 2 diabetes treated with sulphonylurea therapy, Hanefeld et al., (2002) reported a weight loss of 4.6kg in the placebo group and 6.3kg in the orlistat group (p=0.07). Significantly more patients achieved a weight loss of ≥ 5% with orlistat than placebo (51.3% vs. 31.6% p=0.0001). There were also significantly higher improvements in HbA 1c (p=0.001), fasting glucose (p=0.004), post prandial glucose (p=0.003) and LDL cholesterol (p<0.05) in the patients treated with orlistat. There were no significant between group differences in trigylcerides or a reduction in blood pressure. HDL cholesterol did not change in the orlistat group, the placebo group showed an improvement over its baseline level (p=0.02).
Sibutramine: weight loss and co-morbidity risk reduction
A comprehensive literature search for randomised, placebo controlled, double blind trials or systematic reviews, meta analyses or HTAs of such studies, published between January 1996 July 2004 identified 10 systematic reviews (Avenell et al. 2004; Glazer 2001; Haddock et al. 2002; Hensrud 2004; Kim et al. 2003; Leung et al. 2003; McTigue et al. 2003; Nisoli and Carruba 2003; O'Meara et al. 2002; Padwal et al. 2004) and seven additional randomised trials not included in the systematic reviews (Berkowitz et al. 2003; Gokcel et al. 2001; Hauner et al. 2003; Hazenberg 2000; McNulty et al. 2003; Tambascia et al. 2003; Wadden et al. 2000), see Table 15 below. These studies are summarised in Evidence Tables 14 27.
Table 15. Eligible studies reporting the effectiveness of sibutramine Study Size Study population Studies* Patients Study type Code Duration Health technology assessment, systematic reviews, Cochrane reviews, meta-analyses Avenell 2004 Obese Mixed 5 1455 HTA LT At least 1 year Padwal 2004 Healthy obese 3 929 CR LT At least 1 year Hensraud 2003 Obese Mixed 11 >1924 HTA MT Any Arterburn 2003 Obese mixed 6a 1975 SR MT Any Leung 2003 Obese mixed 10 3150 SR MT More than 6 months Glazer 2001 Obese mixed 1* 160 SR LT At least 9 months treatment Haddock 2001 Obese mixed 4 NR MA ST 8-26 weeks, mean=15.5 O’Meara 2002 Obese Mixed 11 2037 HTA MT Any McTigue 2003 Obese Mixed 7 2815 MA MT 6-12 months Nisoli 2003 Obese Mixed, safety 19 4039 SR MT 8-52 weeks** Kim 2003 Obese mixed 21(31 a) 4528 MA MT 8-52 *in period and reported elsewhere, ** 1 study =5.5 hrs, a 21 published studies with 31 separately reported groups Randomised, placebo-controlled, double-blind clinical trials McNulty 2003 Obese diabetic 1 195 RPCDBT LT 12 months Gokcel 2001 Obese diabetic 1 60 RPCDBT ST 6 months Hazenberg 2000 Obese hypertensive 1 127 RPCDBT ST 12 weeks Berkowitz 2003 Healthy Obese adolescents 1 82 RPCDBT LT 12 months
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 29
Table 15. Eligible studies reporting the effectiveness of sibutramine (continued) Study Size Study population Studies* Patients Study type Code Duration Randomised, placebo-controlled, double-blind clinical trials Tambascia 2003 Healthy obese 1 40 RPCDBT ST 24 weeks Hauner 2004 Healthy obese 1 389 RPCDBT LT 54 weeks Sibutramine and orlistat Wadden 2000 Obese mixed 1 34 RPCDBT ST 16 weeks ST=short-term, LT=long-term, MT=mixed term, HTA = health technology assessment, CR = Cochrane review, SR = systematic review, MA = meta-analysis, RPCDBT = Randomised placebo controlled, double blind trial.
The patient populations varied considerably between trials as did the study size and duration. Study populations included healthy overweight and obese adults, mixed male and female participants, female only participants, healthy overweight adolescents, obese patients with poorly and well regulated type 2 diabetes, and obese hypertensive patients, patients with dyslipidemia. One study was carried out solely in a primary care setting.
Studies size varied between 34 389 for RCTs and 160 4,528 for systematic reviews, meta analyses and HTAs. Reporting parameters also varied, some studies reported weight loss in kilograms and % weight loss from baseline weight for placebo and sibutramine, others reported the difference in weight change between the two. The reporting of clinically important weight loss – i.e., weight loss of ≥ 5% was variable.
One systematic review focussed attention on adverse events and side effects reported in 19 RCTs of sibutramine (Nisoli and Carruba 2003), two were NICE Health Technology Reviews (O'Meara et al. 2002) and one was a Cochrane Review (Padwal et al. 2004). The 2003 review of safety and side effects is reported separately in the section on safety and side effects of sibutramine therapy (see page 41). The most recent high quality HTA (Avenell et al. 2004) reported results obtained in a primary care setting.
There was considerable duplication of studies and reporting overlap between the systematic reviews, HTAs and MAs; this overlap and the heterogeneity of the reported parameters made it difficult to summarise and compare these studies.
The most recent HTA (Avenell et al. 2004) reported on five long term studies including the Sibutramine Trial in Obesity Reduction and Maintenance (STORM), see Table 16 (overleaf). This was one of the most consistently reported studies with weighted mean differences between sibutramine reported for weight loss and risk factors. It was also a multi therapy study which allowed direct comparison to be made between sibutramine and a number of other weight loss strategies including orlistat, see Table 17 and Table 21.
Over a period of 12 months, obese patients treated with sibutramine lost on average 4.12 kilograms more than the placebo group (p<0.00001) . There was also a significant difference in favour of sibutramine treatment reported for HDL cholesterol (p=0.0004) and triglyceride levels (p=0.004) together with non significant differences favouring sibutramine for LDL cholesterol, HbA 1c and fasting glucose levels. There were overall differences in diastolic and SBP levels that favoured the placebo group; the difference was significant for DBP (p=0.0005). In both instances, blood pressure levels on average rose during the treatment period. There was no significant between study heterogeneity at 12 months.
Results were reported for the STORM study at 18 months. Significant differences in favour of treatment were reported for weight loss (3.40kg, p<0.00001), triglycerides (p=0.02) and HDL cholesterol (p=0.002). Differences in total cholesterol, LDL cholesterol, HbA 1c and fasting glucose levels also favoured treatment but the differences did not reach statistical significance. Blood pressure at 18 months was not reported.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 30
Table 16. Meta-analysis of RCTs for Sibutramine, long-term studies ≥ 1 year (NICE HTA, Avenell May 2004)
Total Weighted Mean Difference between drug and placebo, over all studies, fixed effects (95%CI) Fasting Study/ Weight TC LDL-C HDL-C TG HbA 1c plasma DBP SBP Year (kg) mmol/l mmol/l mmol/l mmol/l % glucose mmHg mmHg mmol/l Sibutramine + diet vs placebo + diet, 4 RCTs at 12 months, weight reduction FAVOURS Sibutramine Placebo Sibutramine Sibutramine Sibutramine Sibutramine Sibutramine Placebo Placebo -4.12 (95%CI, -4.97 to– 0.01 -0.08 0.10 -0.16 -0.07† -0.05 2.04‡ 1.16‡ Apfelbaum, 1999 3.26) (95%CI, -0.15 to 0.18) (95%CI, -0.23 to–0.07) (95%CI, 0.04 to 0.15) (95%CI, -0.26 to–0.05) (95%CI, -0.25 to 0.11) (95%CI, -0.26 to 0.16) (95%CI, 0.89 to3.20) (95%CI, 0.60 to2.39) Π Π Π Π Π Π Π Π Π McMahon, 2000 2 = 5.96 (p=0.11) 2 = 0.03 (p=0.98) 2 = 1.36 (p=0.24) 2 = 0.27 (p=0.60) 2 = 1.10 (p=0.78) 2 = NA 2 = 0.03 (p=0.98) 2 = 2.99 (p=0.22) 2 = 0.12 (p=0.94) Sibutramine n=530 Sibutramine n=378 Sibutramine n=212 Sibutramine n=214 Sibutramine n=459 Sibutramine n=265 Sibutramine n=378 Sibutramine n=445 Sibutramine n=445 Smith,2001a Placebo n=461 Placebo n=287 Placebo n=136 Placebo n=137 Placebo n=365 Placebo n=77 Placebo n=287 Placebo n=375 Placebo n=375 Smith, 2001b Studies=4 Studies=3 Studies=2 Studies=2 Studies=4 Studies=1 Studies=3 Studies=3 Studies=3
Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect z=9.44(p<0.00001) z=0.16(p=0.87) z=1.09 (p=0.28) z=3.57 (p=0.0004) z=2.87 (p=0.004) z=0.76 (p=0.45) z=0.47 (p=0.64) z=3.47(p=0.0005) z=1.29(p=0.20) Sibutramine + diet vs. placebo + diet, 1 RCTs at 12 months, weight maintenance, hypertensive patients -3.90 0.04 0.08 0.02 -0.18 -0.08 3.30 1.20 (95%CI, -5.75 to–2.05) (95%CI, -0.29 to0.37) (95%CI, -0.01to0.17) (95%CI,-0.21to0.25) (95%CI,-0.47to0.11) NR (95%CI, -0.49 to 0.33) (95%CI, 0.91-5.69) (95%CI, -2.45 to 4.85) STORM, 2000 Studies=1 Studies=1 Studies=1 Studies=1 Studies=1 Studies=1 Studies=1 Studies=1 † Weight maintenance STORM 2000, ‡ increased blood pressure in treatment group unfavourable.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 31
Table 16. Meta-analysis of RCTs for Sibutramine, long-term studies ≥ 1 year (NICE HTA, Avenell May 2004) (continued)
Total Weighted Mean Difference between drug and placebo, over all studies, fixed effects (95%CI) Fasting Study/ Weight TC LDL-C HDL-C TC HbA Ic plasma DBP SBP Year (kg) mmol/l mmol/l mmol/l mmol/l % glucose mmHg mmHg mmol/l Sibutramine + diet vs. placebo + diet, 1 RCT at 18 months
-3.40 -0.19 -0.16 -0.13¥ -0.33 -0.16 -0.12 (95%CI, -4.45 to–2.35) (95%CI,-0.49 to–0.11) (95%CI,-0.37 to–0.05) (95%CI,-0.05 to–0.21) (95%CI,-0.60 to–0.06) (95%CI,-0.36 to–0.04) (95%CI,-0.50 to–0.26) STORM, 2000 Π2 = NA Π2 = NA Π2 = NA Π2 = NA Π2 = NA Π2 = NA Π2 = NA Weight Sibutramine n=350 Sibutramine n=222 Sibutramine n=222 Sibutramine n=222 Sibutramine n=222 Sibutramine n=222 Sibutramine n=222 Placebo n=114 Placebo n=62 Placebo n=62 Placebo n=62 Placebo n=62 Placebo n=62 Placebo n=62 NR NR Maintenance Studies=1 Studies=1 Studies=1 Studies=1 Studies=1 Studies=1 Studies=1
Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect Test for overall effect z=6.33(p<0.00001) z=1.22(p=0.22) z=1.51 (p=0.13) z=3.12 (p=0.002) z=2.39 (p=0.02) z=1.57 (p=0.12) z=0.62(p=0.54) † Weight maintenance STORM 2000 . ¥ decrease in HDL is unfavourable, this result favours the control group .
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 32
Sibutramine therapy for otherwise healthy overweight individuals Three studies enabled an assessment of the effect of sibutramine in participants without significant obesity related co morbidities, see Table 17. The mean difference in the percentage weight loss between sibutramine and placebo treated patients was between 3% and 5% more in sibutramine treated patients (p=<0.05), a similar pattern was apparent for absolute weight loss.
Two of the three studies reported the proportion of patients with ≥ 5 % and ≥ 10% weight loss with difference between the treatment and placebo groups ranging from 21 34% for patients losing ≥ 5% and 15 21% losing ≥ 10% of their initial body weight. In both instances, the difference was significant. These results differ little from the overall picture obtained from the meta analysis of trials with both low and high risk participants, see Table 17.
Table 17. Studies reporting the effectiveness of sibutramine in otherwise healthy overweight individuals Mean Mean % % No of Dose difference difference Difference Difference Study Year studies Duration Mg/d in % change Weight loss kg WL ≥5% WL ≥10% (subjects) (s.d.) (s.d.) (s.d.) (s.d.) Padwal (SR) 2004 3(929) ≥1 years 15-20 4.6 (3.8-5.4) a 4.3 (3.6-4.9) a 34 (28-40) 15 (4-27) 10 -6.1 b -5.6 b NR NR Tambascia (RCT) 2003 1(40) 24 weeks - +1.1 c +0.9 c NR NR Hauner (RCT) 2003 1(389) 54 weeks 15 3.4 3.0 (-1.4t o-4.6) a 21.2 c 21.8 a a difference from placebo, b difference from baseline in sibutramine group, c difference from placebo. RCT=randomised controlled trial. WL=weight loss. Bold = a significant difference, s.d. = standard deviation.
Sibutramine therapy for patients with type 2 diabetes Two RCTs reported the effects of sibutramine treatment on weight loss, metabolic control and blood pressure in patients with type 2 diabetes, see Table 18.
Table 18. Studies reporting the effectiveness of sibutramine in individuals with type 2 diabetes
Mean difference placebo-drug, bold = significant difference Dose Study Year ≥5% ≥10% LDL-C HDL-C Total-C TG SBP DBP Mg/d N= Pulse rate WL WL mg/dl mg/dl mg/dl mg/dl mmHg mmHg McNulty a 2003 15 195 46 14 0.0 0.1 0.0 -0.2 4.6 2.8 5.9 (RCT) 20 65 27 0.1 0.1 0.1 -0.3 -1.3 0.0 5.8 Gokcel 2001 10 60 b NR NR 7.6 0.96 20.4 46.4 ns SR SR a study patients were receiving metformin b females with poorly regulated diabetes. NR=not reported, SR=significant reduction no data reported.
Significantly more patients in the sibutramine group lost ≥ 5% and ≥ 10% of their initial body weight than the control group. Triglycerides were significantly reduced in the treatment group in one study (McNulty et al. 2003), differences between the groups in terms of cholesterol levels were not significant. Blood pressure and pulse rate increased significantly in the treatment group at a dose level of 15mg/d; in a few individuals blood pressure increases were particularly marked. The authors noted that blood pressure increase tended to be offset by weight loss; only 37% of patients who achieved a ≥ 10% weight loss showed a rise of ≥ 5mmHg in SBP.
A sibutramine dose response effect was observed in one study (McNulty et al. 2003) with doses between 15 20mg/d eliciting increases in the proportion of participants losing 5% and 10% of their body weight.
In an earlier study, Gokcel et al. (2001) reported a 8.7kg greater weight loss for sibutramine after six months of treatment (drug placebo, p<0.0001). All risk parameters, except SBP, significantly improved with treatment.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 33
A number of other eligible studies (Leung et al. 2003) reported stratification and sub group analyses that included diabetic patients; all reported significant weight loss differences between sibutramine and placebo treated participants.
Sibutramine therapy for patients with hypertension The effectiveness of sibutramine in a patient population of mild to moderately obese hypertensive patients was examined in one RCT (Hazenberg 2000). After 12 weeks of treatment with 10mg sibutramine, the treatment groups had lost 2.4% more of their baseline body weight than the placebo group (p=0.002). In addition, 44% of treated patients had lost more than 5% of their baseline body weight compared to 17% in the placebo group (p=0.01). Blood pressure reductions of 4 5mmHg were observed in both groups. Sibutramine patients showed a slightly lower mean improvement than the placebo group; differences between the groups were not significant.
A meta analysis of 21 randomised, placebo controlled, double blind trials to determine the effect of >5mg/d sibutramine on weight loss and blood pressure was carried out by Kim et al., in 2003. The patient population was extremely heterogeneous and co morbidities of diabetes, hypertension, hyperlipidemia were allowed. Sibutramine had a significant effect on weight loss with an effect size 6 of 1.0 (range, 1.17 to 0.84) but increased blood pressure significantly. The net increase in blood pressure attributable to sibutramine was 1.6mmHg and 1.8mmHg for SBP and DBP respectively. Larger increases in blood pressure were observed in heavier and/or younger participants with an effect size of 0.16 (0.08 0.24) and 0.26 (0.18 0.33) respectively. The small increases in blood pressure observed were not considered to be important in normotensive patients or patients with well controlled hypertension being treated in a clinical setting. However, the increase was considered to be clinically important in patients who had borderline or high blood pressure.
Most other studies of sibutramine actually excluded patients with hypertension or cardiovascular disease, others only allowed well controlled hypertension (Berkowitz et al. 2003; Gokcel et al. 2001; Hauner et al. 2003; McNulty et al. 2003; Padwal et al. 2004).
The use of sibutramine in adolescents A recent dose ranging, randomised trial of sibutramine and behaviour therapy in obese adolescents aged 13 17 years reported an average weight loss of 7.8kg (s.d.=6.3kg) and 8.5% (s.d.=6.8 %) in an ITT analysis at six months (Berkowitz et al. 2003). The sibutramine treated children lost significantly more weight than the placebo treated group. Medication was reduced in 23 cases and discontinued in 10 cases to manage increases in blood pressure, pulse rate and other symptoms. The investigators concluded that medications for weight loss in children should only be employed in the context of a clinical trial until more extensive safety and efficacy data were available, see Evidence Table 19.
Sibutramine dose ranging studies The dosage of sibutramine used in most of the studies included in this review varied between 5 30mg/d. In some cases a sub group analysis was carried out for different dose levels, this was however uncommon. One dose ranging study (Bray et al. 1999), was identified in a systematic review of 10 RCTs by Leung et al., (2003).
6 Effect size = the standardised difference of changes in weight loss from baseline between the treatment and control group. An effect size of >0.80 in this context may be considered to be moderate, and effect size >1.0 may be considered to be large, however, interpretations of effect size importance may vary.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 34
Doses of sibutramine ranging from 1 30mg/d in 1047 obese patients produced a significant dose related weight loss at six months, see Table 19.
Table 19. Sibutramine dose ranging studies. Bray 1999 reported in Leung et al., 2003 % losing at least 5 or 10 Discontinuation SBP DBP Pulse Dose % of base weight due to Treatment group % Weight loss Change Change Change mg/d Adverse events mmHg mmHg bpm 5% 10% % Placebo - 1.2 19.5 0.0 8 -0.8 1.7 0.6 1 2.7 25.3 10.5 11 0.3 1.2 0.3 5 3.9 37.4 12.1 5 2.1 2.5 3.3 Sibutramine 10 6.1 59.6 17.2 9 2.8 4.2 6.0 24 weeks 15 7.4 67.3 34.7 11 2.7 3.4 6.1 N=1047 20 8.8 71.9 38.5 13 4.0 5.0 7.0 30 9.4 77.2 45.5 18 3.3 4.1 5.3 Bold = significant difference from placebo
In this study, weight loss at four weeks was found to be predicative of weight loss at six months and the proportion of patients withdrawing from the trial because of adverse events increased with increasing sibutramine dose. Blood pressure and heart rate increased with increasing doses of sibutramine up to 20mg/d. This trend was not followed in the 30mg/d group, however, it is not clear if susceptible patients had withdrawn at this stage or if there was an ITT analysis. Treatment related adverse events that resulted in discontinuation were all more common in the 30mg/d group and included hypertension, palpitations, tachycardia, insomnia and dyspepsia.
Comparative drug studies
Seven systematic reviews of pharmacotherapy for obesity between 2001 and 2004 compared the effectiveness of two or more of the drugs being examined in the current review (Arterburn et al. 2004; Avenell et al. 2004; Glazer 2001; Haddock et al. 2002; Hensrud et al. 2003; McTigue et al. 2003; Padwal et al. 2004), see Table 20.
Table 20. Eligible studies comparing two or more review drugs Study Phentermine Diethylpropion Orlistat Sibutramine Other weight loss therapies reported Avenell, 2004 Metformin (National Institute Selective serotonin reuptake inhibitors (SSRIs) For Clinical Acarbose Effectiveness) X X VLCDs LCDs Protein sparing modified fast Diet, exercise, behaviour therapy Hensraud, 2003 Ephedrine + Caffeine (ICSI Technology x Assessment Committee) Arterburn, 2003 Mazindol McTique, 2004 Metformin (US Preventative Fluoxetine Services Task Force) Haddock, 2002 Dexamphetemine (USA) Benzocaine Benzphentermine Dexfenfluramine Fenfluramine Fluoxetine Mazindol Methamphetamine Phendimetrazine Phenylpropanolamine Sertaline Glazer, 2001 Mazindol (USA) Fluoxetine (SSRI) Sertaline (SSRI) Fenfluramine Dexfenfluramine Padwal, 2004 None (Cochrane Review, X X Canada )
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 35
Avenell et al., May 2004. Systematic review of the long-term effects and economic consequences of treatments for obesity and implications for health improvement. This extensive and well reported review included a meta analysis of all randomised, placebo controlled, double blind, efficacy trials of obesity therapy with a follow up of at least one year. Pharmacological and non pharmacological interventions were examined. The objective of the review was to identify therapies that achieved weight reduction, risk factor modification or improved clinical outcomes. A summary of the effects of sibutramine and orlistat resulting from a meta analysis of eligible trials is shown in Table 21. Phentermine and diethylpropion were not included in this review, as they are no longer recommended for anti obesity therapy in the UK. The reported results for low fat diets (LFDs) are also included as they can include meal replacement programs.
Table 21. Summary of the comparative effectiveness of orlistat, sibutramine and low fat diets that may include meal replacements Weight change 95% Therapy Assessed at from baseline Confidence Change in risk factors (kg) Interval Sibutramine + diet 18 months -3.40 -4.45 to –2.35 Beneficial except for DBP Orlistat + diet 24 months -3.26 -4.15 to –2.37 Beneficial Low Fat diets* 12 months -5.31 -5.86 to –4.77 Beneficial * Can include meal replacement programs
Weight loss from baseline was greatest with LFD followed by sibutramine and orlistat. However, treatment duration varied between 12 24 months making it difficult to make a valid comparison of performance. A beneficial change in risk factors was reported for all three therapies with the exception of sibutramine and DBP. A weight loss of 10 kilograms was reported to be associated with a fall in total cholesterol of 0.25mmol/l and a fall in DBP of 3.6mmHg; a 10% weight loss was associated with a fall in SBP of 6.1mmHg.
Hensud, 2003. Pharmacological approaches to weight loss in adults. Institute for Clinical Systems Improvement, Technology Assessment Report, February 2003. Most of the long term trials reported in this HTA have been reported elsewhere in the current report. However, the inclusion of some short term trials of orlistat and sibutramine not reported elsewhere and a comparative summary of changes associated with a wide range of obesity related risk factors merits its inclusion here. The reported weight change from baseline and the proportions of patients losing clinically significant amounts of weight with the two drugs are shown in Table 22.
Table 22. Summary of short-term weight loss trials of sibutramine and orlistat not otherwise reported in systematic reviews and HTAs. % of participants % of participants Weight change with a loss of with a loss of Therapy Assessed at from baseline ≥5% of initial ≥10% of initial (kg) body weight body weight Sibutramine + diet 12-52 weeks -2.4-to -16.6 27-88% 6-76% Orlistat + diet 12 months -3.9 to –10.3 33-69% 10-39% 24 months -5.0 to –7.4 NR NR
Overall, 5 25% of participants failed to complete the pre randomisation run in phase of the studies and 9 54% of those randomised failed to complete the study (orlistat, sibutramine and placebo groups). The greatest benefits were reported for patients with co morbidities such as
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 36
diabetes, these participants experienced less weight loss but important physiological benefits based on Class A and M levels of evidence 7.
The type of weight loss programs that led to high proportions of patients achieving clinically important levels of weight reduction (>50% of patients losing ≥ 5% of their initial body weight and >30% of patients losing ≥ 10% of their initial body weight) were examined for predictors of this outcome. High performing programs were characterised by a number of features notably:
• a placebo run in period that identified individuals that were willing and able to make changes in their nutritional and activity patterns
• regular monitoring in a clinical setting
• exclusion of patients with serious disease, major depression or substance abuse
• information on behaviour modification
• changing nutritional choices
• increasing physical activity.
Padwal et al., August 2003. Long-term pharmacotherapy for obesity and overweight. Cochrane Review. Eight anti obesity agents were reviewed in this study but only orlistat and sibutramine studies met the review inclusion criteria. All of the studies included in this review were also reviewed elsewhere but not necessarily in the same review and not reported in the same way.
On average, orlistat treated patients lost 2.7kg (95% CI: 2.3 to 3.1kg) or 2.9% (95% CI: 2.3 to 3.4%) and sibutramine patients 4.3kg (95% CI: 3.6 to 4.9kg) or 4.6% (95% CI: 3.8 to 5.4%) more weight than the placebo group. The number of patients who achieved at least 10% weight loss was 12% (95% CI: 8 to 16%) higher with orlistat than placebo and 15 % (95% CI: 4 to 27%) higher with sibutramine than placebo. Orlistat and sibutramine were reported as being “modestly effective in promoting weight loss” but with “interpretation limited by high attrition rates”.
Glazer, 2001. Long-term pharmacotherapy of obesity, 2000. This systematic review included a comparative analysis of all of the drugs examined in the current review – i.e., phentermine, diethylpropion, orlistat and sibutramine. In addition, it evaluated some early studies of phentermine and diethylpropion not included in other reviews. The studies identified for review were considered to be too few and too heterogeneous to warrant a meta analysis of their findings and an average weight loss was reported for each of the drugs across the relevant studies. Trials lasted between 36 to 52 weeks. Three of the four drugs reviewed in the current report, sibutramine, orlistat and phentermine, were considered to have shown weight loss efficacy in long term trials, see Table 23 (overleaf).
7 Class A = randomised controlled trial, Class M = meta analysis or systematic review.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 37
Table 23. A comparison of the effectiveness of sibutramine, orlistat, phentermine and diethylpropion (Glazer et al., 2001) Weight loss Weight loss attributable attributable Therapy to drug s to drug a (kg) (%) Sibutramine 4.3 5.0 Orlistat 3.4 3.4 Phentermine 7.9 8.1 Diethylpropion -1.5 -1.5 a weight loss in excess of placebo
Patients receiving phentermine and sibutramine achieved a clinically important weight loss – i.e., weight loss of ≥ 5%, orlistat patients achieved a more modest weight loss while patients treated with diethylpropion on average gained weight. The magnitude of the weight loss was supported by Level I evidence 8 for orlistat (2 studies) and sibutramine (1 study) and by Level I/II/III evidence 9 for phentermine. With the exception of one short term study, the level of evidence reported for phentermine was below the standard required for the present report. The authors concluded that while the weight loss achieved with the aid of medication was modest, it was associated with important health and psychological benefits.
Haddock et al., 2002. Pharmacotherapy for obesity: a quantitative analysis of four decades of published randomised clinical trials This review considered 15 FDA approved medicines used for weight loss in the USA that included phentermine, diethylpropion, orlistat and sibutramine. The results of the meta analysis of studies reported between 1960 and 1999 are shown in Table 24.
Table 24. A comparative study of the effectiveness of sibutramine, orlistat, phentermine and diethylpropion (Haddock et al., 2002) Number at Weight loss Dosage Weeks of Number at Mean post test attributable Therapy Per day any treatment post-test effect placebo to drug a (mg) (range) drug (range) size c (range) kg (range) Sibutramine 14.0 (10.0-20.0) 14.5(8-26) 27.3(15-52) 26.8(15-49) 3.5(2.4-5.1) 1.0 Orlistat 302.9 (190-360) 47.5(16-76) 236.9(46.7-657) 164.5(46-340) 2.08(0.30-4.2) 0.50 Phentermine 27.5 (15-30) 13.2 (2-24) 32(15-76) 29.4 (12-74) 3.6(0.6-6.0) 0.62 Diethylpropion 75 (75-75) 17.6(6-52) 21.2(5-32) 18(4-29) 3.0(-1.6 b -11.5) 0.60 a Drug-placebo b negative numbers for kilo changes indicate a weight gain. c reported in a figure and without consideration of study design differences, values in table approximate only. 0=no treatment effect.
Sibutramine produced the largest mean effect size in the meta analysis with a significantly better weight loss than the other drugs. However, the 95% confidence intervals for the sibutramine estimate overlapped with diethylpropion and phentermine suggesting that there were no statistically significant differences between the effect sizes of these three drugs. Post treatment follow up studies were reported for phentermine and sibutramine, patients treated with both drugs sustained a placebo subtracted weight loss of 2.43kg and 2.37kg and large effect sizes of 0.810 and 1.05 10 respectively.
8 Level I evidence = randomised, double blind, placebo controlled studies.
9 Level II/III evidence = double blind, placebo controlled study/blind comparison.
10 Effect sizes >0.80 produced by meta analysis are considered to be in the “large” effects range.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 38
None of the 15 drugs included in the meta analysis exceeded a placebo subtracted weight loss of 4.0kg and none of the drugs were considered by the authors to have a demonstrated clear supremacy as an anti obesity medication.
Arterburn, 2004. Obesity: What are the effects of drug treatment in adults? An overview of RCTs and systematic reviews of RCTs, for nine drug treatments for obesity including phentermine, diethylpropion, orlistat and sibutramine was recently published by the British Medical Journal (Arterburn et al. 2004). All of the RCTs and systematic reviews assessed in the overview have been reported elsewhere and summarised in this review in the appropriate evidence tables. However, the study reported a useful summary of the evidence for effectiveness, harm and benefit for the study drugs and a number of key messages:
• no studies used the primary outcomes of functional morbidity or mortality and therefore the long term effect of obesity drugs was not known
• in many studies there was insufficient evidence of effect because of high drop out rates
• that overall diethylpropion, orlistat, phentermine and sibutramine had a favourable trade off between benefits and harms, see Table 25.
Table 25. Weight loss drugs trade off between benefit and harm Favourable trade off between benefits Unknown effectiveness Likely to be ineffective or harmful and harms Diethylpropion Sibutramine + orlistat Dexfenfluramine Orlistat Fenfluramine Phentermine Fenfluramine + phentermine Sibutramine Phenylpropanolamine Fluoxetine Mazindol Drugs included in the current review
McTique et al., 2003. Summary of the evidence for the US Preventive Services Task Force (USPSTF) on obesity An overview of the health outcomes from systematic reviews and RCTs of obesity drugs, was carried out by the US Preventive Services Task Force on obesity. Studies with at least one year follow up were identified by Medline and Cochrane Library searches for the period 1994 2003. Phentermine, diethylpropion, sibutramine and orlistat were amongst the drugs compared.
All of the reported systematic reviews (Arterburn and Hitchcock 2001) and RCTs (18 additional RCTs meeting the eligibility criteria) which included phentermine, diethylpropion, orlistat or sibutramine have been reported in separate studies elsewhere in this review and only the comparative efficacy observations are reported here:
• sibutramine and orlisat had high response rates, similar efficacy and promoted modest but significantly higher weight loss (3 5kg) than their controls
• in a single multi drug trial, participants treated with sibutramine lost significantly more weight than those treated with orlistat (13.4kg vs. 8kg)
• prolonged treatment with sibutramine and orlistat helped to sustain weight loss for up to two years
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 39
• phentermine had short term efficacy – i.e., promoted modest weight loss, but was not approved for long term use
• efficacy of diethylpropion was more limited and variable than the other reported drugs
• modest weight loss improved clinical outcomes
• a high number of patients can achieve a clinically significant (5% 10%) weight loss
• discontinuation of treatment may lead to rapid weight regain.
Combined drug studies
A single randomised trial examining the effectiveness of combined therapy with two weight loss drugs with different modes of action was identified (Wadden et al. 2000).
Wadden et al., 2000. Does adding orlistat to sibutramine induce further weight loss? The effect of adding orlistat to weight loss therapy with sibutramine was examined in a trial of 34 obese, female volunteers. The participants had already lost weight on sibutramine therapy and were continuing to take the drug at a dosage of 10 15mg/d. At the end of a 16 week combined treatment period, body weight remained essentially unchanged in both the sibutramine + orlistat and the sibutramine + placebo group.
In a sub group analysis, participants who had lost > 10% of their initial body weight in the previous period of sibutramine treatment gained weight in the extension study – regardless of whether they received orlistat or placebo. Patients who lost < 10 % of their initial body weight in the previous period of sibutramine treatment lost further amounts of weight when orlistat was added to their treatment. These participants lost more weight than those taking sibutramine alone; however, the difference did not reach a significance level of <0.05.
The authors suggested that most obese participants have a limit of 10 15% weight loss with further loss thwarted by a number of regulatory physiological processes including compensatory mechanisms that decrease energy expenditure.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 40
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 41 Safety/Side effects
Safety and side effects of sibutramine therapy
Sibutramine is currently being monitored by the New Zealand Intensive Medication Monitoring Program (IMMP) 11 .
A systematic review of the benefits and risks associated with the use of sibutramine in the management of obesity was carried out by Nisoli and Carruba (2003), see Table 26.
Table 26. The effect of sibutramine on blood pressure and heart rate; results from 20 studies (Nisoli and Carruba, 2003) No Patients No ∆∆∆SBP ∆∆∆DBP ∆∆∆Heart rate No of drop- Dose Study reporting Trial Subjects (mmHg) (mmHg) (beats/min) outs mg/d Length AEs P S P S P S P S P S P S Obese subjects Apfelbaum 10 52wk 78 82 NR NR -1.9 +1.5 +1.0 +8.0 § 63 72 5a 2a Fanghanel 10 24wk 54 55 -4.0 +1.7 -1.5 +0.3 -2.2 +0.1 23 31 10 15 Hanotin 10 12wk 114* 112 +0.6 +0.9 -0.1 +0.4 -0.9 +3.6 90 84 19 10 Cuellar 15 24wk 34 35 NR NR NR NR NR NR 16 23 25 13 Walsh 15 12 wk 9 10 NA NA NA NA NR NR NR NR NR NR Hanotin 5,10,15 12wk 59 177 NR NR NR NR -1.6 +3.2 42 126 12 29 Weintraub 5,20 8wk 20 40 NR NR NR NR NR NR NR NR 1 4 James 10,20 72wk 115 352 -2.4 +1.9 -.05 +3.4 +0.2 +4.6 5 48 58 148 McNulty 10,20 12wk 64 130 -0.2 +1.5 +0.5 +1.9 -0.8 +5.0 NR NR 18 32 Seagle 15,30 8wk 15 29 NR NR NR NR NR NR NR NR NR NR Hanson 30 5.5h 11 11 +5.0 +8.6 +1.4 -4.3 +2.8 +11.24 NA NA NA NA Patients with type 2 diabetes Knoll 10,20 52wk 163 332 -0.5 +0.7 -0.9 +0.8 +0.1 +2.7 NR NR 83 146 Knoll 15 NR 122 114 NR NR NR NR NR NR NR NR 12 a 9a Knoll 15,20 NR 64 131 NR NR NR NR NR NR NR NR 5A 14 A Bray 1.5-30 24wk 148 899 -0.8 +2.5 +1.7 +2.9 +0.6 +4.6 12 101 61 303 Fujioka 5,20 24wk 86 89 +2.4 +3.9 +1.4 +2.6 +0.7 +6.6 68 70 25 29 Gokcel 10 24wk 30 30 NR NR NR NR NR NR NR NR 5 1 Serrano- 15 24wk 65 69 -1.1 +0.5 NR NR NR +2.4 34 42 8 16 Rios Finer 15 12wk 44 47 NR NR NR NR +0.2 +7.5 45 42 4 4 Sibutramine use in children Berkowitz 5,10,15 24wk 39 43 +4.0 +0.1 +0.6 +1.8 2.0 +5.4 NR NR 5 3 a drop-out for adverse events, *comparison with dexfenfluramine § difference significant only at 6 months. P = Placebo, S = sibutramine
In general, the methodological quality of the reported trials was considered to be good. Double blinding, selection criteria and group comparability at baseline was reported for all trials. However, none of the trials included methods to determine if blinding had been successful and relatively few trials reported the use of a priori power calculation to estimate the required sample size.
All trials followed a similar protocol comprising a 1 3 week pre randomisation run in period, followed by a randomised assignment to sibutramine or placebo during a treatment period of between 8 52 weeks. In all trials adjunctive therapy comprising diet, exercise and behavioural modification advice was included in varying intensities.
11 IMMP is part of the NZ Pharmacological Vigilance Centre (formerly CARM), Department of Preventive and Social Medicine, Dunedin School of Medicine, University of Otago.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 42
Adverse events
The most commonly reported adverse events of sibutramine were headache, constipation and nausea. Certain adverse events associated with the nervous system including dizziness, dry mouth and insomnia were reported by >5% of patients.
The two most clinically important side effects reported for sibutramine were increased blood pressure and tachycardia; these were generally seen in the first eight weeks of treatment, see Table 26. Heart rate increases between 4 11 beats per min were reported for sibutramine against slightly decreased and increased rates ( 2 to +3 beats per min) for the placebo. In a number of instances the difference was significant, however, overall heart rate did not change dramatically during the course of treatment. A similar trend could be discerned in sub groups with type 2 diabetes with increased heart rates of 3 7 beats per minute for diabetic patients treated with sibutramine and +0.1 to +0.7 beats per minute for the placebo treated diabetic group.
Blood pressure changes
The effect of sibutramine on blood pressure is the consequence of a reducing effect due to its action on body weight and a stimulating effect based on noradrenaline reuptake inhibition. Therapeutic doses of sibutramine have been associated with increased blood pressure in a dose dependent manner (Weintraub et al. 1991), see Table 19.
SBP was higher in all studies in sibutramine patients (median +1.7mmHg, range +0.50 to +8.6) than those who had the placebo (median 0.5mmHg, range 4.00 to +5.00mmHg). Diastolic blood pressure followed the same pattern (sibutramine median =+1.5mmHg, range 4.3 to +3.4, placebo median =0.05mm Hg, range 1.9 to +1.7).
In a trial of obese hypertensive patients (Hazenberg 2000), the mean reduction in supine DBP was numerically but not statistically greater in the placebo group compared with the sibutramine group (5.7mmHg vs. 4.0mmHg; p=0.21). Similar reductions were reported in SBP.
In March 2002, the Italian Ministry of Health approached the EMEA for a reassessment of the benefit/risk ratio of sibutramine after 51 adverse events and two deaths were reported in a period of one year. Reporting incidence by number of fatal cases was calculated at 2.40 2.86 fatal events per 100,000 treatment years, which was lower than the estimated rate for the best available control population. EMEA reaffirmed a positive benefit risk profile for sibutramine based on these data.
Safety and side effects of orlistat therapy
No systematic review of the safety of orlistat and its side effects was identified. However, all of the studies reviewed reported on safety aspects of therapy and the adverse events associated with orlistat treatment.
The main adverse events reported by patients taking orlistat were GI in keeping with the local action of the drug. They included:
• fatty/oily stools
• liquid stools
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 43
• faecal urgency
• faecal incontinence
• flatulence
• lower serum levels of fat soluble vitamins A, D, E and K 1.
These adverse effects were generally mild to moderate and most events occurred early in the course of treatment. Although orlistat treated patients generally had significantly more adverse events than the control group, few serious adverse events were reported.
In the overview of systematic reviews and RCTs by McTigue (2003), it was noted that GI side effects were reported in 22 95% of orlistat users and that in addition to reduced vitamin absorption there was a reduced absorption of contraceptive pills (Arteburn, 2001). Reductions in plasma levels of warfarin and cyclosporine have been reported when co administered with orlistat. There is insufficient evidence on weight regain and long term safety for orlistat (Arterburn and Hitchcock 2001).
Safety and side effects of phentermine therapy
No systematic reviews of the safety and side effects of phentermine were identified. However, all efficacy studies reviewed reported on safety aspects of therapy and the adverse events associated with phentermine treatment. Most of the studies published between 1996 2004 were of short duration with small numbers of patients. There is little or no data on the effects of phentermine on obesity related co morbidities reported in these studies. No serious adverse events were reported in earlier studies of phentermine reported by McTigue et al. (2003).
In 1996, phentermine was administered with the newly introduced fenfluramine in the combination known as “Phen Fen”. Following reports of cardio valvopathy in patients treated with this combination fenfluramide was withdrawn from the market (Hensrud 2000; Rothman and Baumann 2000).
There was no evidence that phentermine was responsible for the valvopathy (Hensrud et al. 2003) and patients treated with phentermine alone had no reported adverse reaction of this type. Nonetheless, the Medicines Control Agency (MCA) warned that a link between phentermine and heart and lung problems could not be ruled out (Medicines Control Agency Committee on Safety in Medicines, 2004). Phentermine remains available for use in Europe, however, it is not recommended for use by the Royal College of Physicians (Wilding 2004).
Safety and side effects of diethylpropion therapy
No systematic reviews of the safety of diethylpropion and its side effects were identified. However, all of the studies reviewed reported on safety aspects of therapy and any adverse events associated with diethylpropion treatment. Data on the safety of diethylpropion is limited as published studies are generally of short duration (<3 months) in keeping with its conditions of use. Reported side effects include palpitations, tachycardia, elevated blood pressure and arrhythmia together with various adverse reactions associated with CNS stimulation – e.g., nervousness, restlessness, dizziness, euphoria and insomnia. As with amphetamine and other amphetamine derivatives, diethylpropion has a potential to induce dependence.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 44
Case reports of pulmonary hypertension have been reported and the European Medicines Control Agency warned that a link between diethylpropion and heart and lung problems could not be ruled out (Medicines Control Agency Committee on Safety in Medicines, 2004). Although diethylpropion still has a licence for use as an anti obesity treatment in the UK, its use is no longer recommended for this purpose by the MCA or the Royal College of Physicians (Wilding 2004).
Summary of potential harms of pharmacotherapy intervention
Overall, pharmacotherapy for obesity when carried out as indicated, is reasonably safe. Clinically significant adverse events were reported for sibutramine and orlistat in studies of one to two years; however, the longer term effects of these drugs is not known. Orlistat has predominantly GI side effects while sibutramine may cause clinically significant rises in blood pressure in some patient groups. Both of these drugs are reported to have clinically significant benefits in the treatment of obesity and its related co morbidities and a positive trade off between benefits and harm.
Phentermine and diethylpropion are only indicated for short term therapy. Their adverse events profile were somewhat sketchy in the studies reviewed, possibly because of the limits set on the publications dates for study inclusion for this review 12 . Phentermine in combination with fenfluramide (“Phen Fen”) was reported to be responsible for a number of cases of cardio valvopathy and although no cases have been reported in patients taking phentermine alone, it is no longer recommended for use by the Royal College of Physicians in the UK. Diethylpropion has been linked in a number of case reports with pulmonary hypertension and the European MCA have noted that a link between the two cannot be ruled out. Diethylpropion is no longer recommended for the treatment of obesity by the MCA or the UK Royal College of Physicians.
Safety and side effects of meal replacement programs
Meal replacement programs have not been critically evaluated until recently and there are few reports of their safety. In the two studies evaluated in the current review (Allison et al. 2003; Heymsfield et al. 2003), there were few reported adverse events and their impact was limited.
A lack of adverse events reporting in any study patients, including PMR treated diabetics was noted by Heysfield et al., (2003) in a systematic review of six studies examining meal replacement strategies. On the other hand, an unblinded RCT of a novel soy based meal replacement formula for weight loss in 100 obese volunteers (Allison et al. 2003) reported a large number of adverse events 13 including:
• decreased appetite/anorexia
• constipation
• diarrhoea
12 Eligibility criteria limited studies to the period 1996 2004; most studies reporting on these drugs were published much earlier.
13 Assessed by the monitoring of side effects system (MOSES).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 45
• drooling/salivation
• gas/indigestion
• abnormal taste
• lethargy
• excessive sleep
• enuresis/nocturesis
• weight gain
• weight loss.
For five of the events (constipation, gas or indigestion, abnormal or metallic taste, lethargy and weight loss), there was a significantly higher incidence in the meal replacement group. Generally however, treatment was reported to be well tolerated and free of serious side effects by the study authors.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 46
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 47 Practice recommendations and guidelines
National Institute of Health, USA
General guidelines for incorporating weight loss medicines into the overall treatment plan for overweight and obese individuals were published by the National Institute of Health in 2000. It was recommended that:
1. Therapy should be initiated with lifestyle changes (diet and behaviour modification and exercise) and that if these failed to promote a 10% weight loss (or at least 0.5kg/wk) over six months that pharmacotherapy should be considered.
2. Only individuals with a BMI ≥ 30kg/m 2 and no obesity related risk factors, or ≥ 27kg/m 2 if the patient also has hypertension, dyslipidemia, coronary artery disease, type 2 diabetes or sleep apnoea, should be eligible for treatment.
3. If the patient does not lose at least 2kg in the first month of therapy, the likelihood of response to that medication is low and an adjusted dose, discontinuation of the drug or changing the medication is advised.
4. If significant weight loss occurs with the medication or the initial weight loss is maintained, medication may be continued as long as it remains effective and the side effects tolerable.
In addition, it was suggested that drugs that were limited by the FDA to a maximum period of administration of 12 weeks would be best administered in the setting of a clinical trial.
The National Institute for Clinical Excellence, UK
NICE published best practice guidelines for the use of orlistat and sibutramine in 2001. 14
Orlistat
• Orlistat should only be prescribed for obese patients who fulfil the licensing criteria and have lost at least 2.5kg 15 by dietary control and increased physical activity in the month prior to treatment.
• When treatment is offered trained practice nurses and dietitians should be available to offer advice, support and counselling on diet, physical exercise and behavioural strategies.
• Continuation of orlistat therapy beyond three months should be supported by evidence of at least 5% body weight loss since the start of treatment.
14 http://www.nice.org.uk.
15 The FDA guidelines do not require prior weight loss using diet and exercise alone and continuation of treatment beyond 3 and 6 months does not require weight loss of 5% and 10% respectively (Ballinger et al 2002).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 48
• Continuation of orlistat therapy beyond six months should be supported by evidence of at least 10% weight loss from the start of treatment.
• Treatment should not usually be continued beyond 12 months and never beyond 24 months 16 .
Sibutramine
• Sibutramine should be prescribed only as part of an overall treatment plan for the management of obesity for people aged 18 65 years.
• When people are prescribed sibutramine, they should also be offered advice, support and counselling on diet, exercise and behaviour changes.
• It should be prescribed for the management of obesity in people who have a BMI of 30kg/m 2 or more, or who have a BMI of 27kg/m 2 or more together with significant disease – e.g., type 2 diabetes, high cholesterol.
• Sibutramine should not be prescribed unless the person taking it is monitored for side effects.
• Treatment should only be continued for more than four weeks if patients have lost 2kg in weight and should only continue treatment beyond three months if they have lost at least 5% (5kg for each 100kg) of their body weight from the start of drug treatment.
• Sibutramine should be stopped if patients do not lose weight as described.
• Blood pressure should be checked regularly. Increases in blood pressure may be a reason to stop treatment.
• Sibutramine is not recommended for patients who already have high blood pressure – i.e., 145/90 or above.
• Treatment is not recommended beyond 12 months. 17
• There is no evidence to show that prescribing sibutramine with other drugs used to treat obesity is beneficial.
16 The XENDOS trial of orlistat has recently reported on its safety and efficacy at four years.
17 This guidance is to be reviewed in September 2004.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 49 High profile clinical trials
Completed studies
Although there are a number of recently completed trials relating to the management of adult obesity (for a summary see Avenell et al., 2004, Appendix 9), most have little relevance to the ACC claimant population and/or do not involve any of the medications reviewed in the current report. Two high profile clinical trials of the newer weight loss drugs are notable exceptions. Their clinical outcomes have been reported in earlier sections of this review, they are examined individually here to describe the particular circumstances and goals of each trial and to highlight their contribution to this field of study.
STORM: STORM has been widely reported (Astrup and Toubro 2001; Hansen et al. 2001; James et al. 2000; James 2001) and its results incorporated in the recent systematic review by Avenell et al., (2004). This was a randomised double blind, placebo controlled, multi centre European study of sibutramine for the induction and maintenance of weight loss in obese subjects. A total of 605 male and female obese patients were treated for 24 months with sibutramine 10mg/d and a 600kcal deficit diet. If weight regain occurred sibutramine was increased to 20mg/d. At the time of the study, the European Union licence only permitted continuous treatment with sibutramine for up to one year, the STORM study led the US FDA to extend its clearance of sibutramine therapy to two years.
XENDOS: this prospective trial built upon a number of earlier diabetes prevention studies in overweight subjects with IGT (e.g., the STOP NIDDM, DPP, DPS and TRIPOD studies). Xendos was a randomised, double blind, placebo controlled, prospective, four year trial designed to determine the long term diabetes preventing and weight reducing effect of orlistat in combination with lifestyle changes in obese patients. Three thousand, three hundred and five obese patients were randomly assigned to orlistat 120mg t.i.d. or placebo. All patients had an energy reduced diet (reduced by up to 800kcal/day) with 30% fat intake. Patients were instructed to walk 1km a day above their normal level of physical activity.
At four years the orlistat treated patients had lost 5.8kg compared with 3.0kg in the placebo groups (p=<0.001). In IGT patients, the values were 5.7kg and 3.0kg respectively. The cumulative incidence of diabetes over the period of the study was orlistat 6.2 vs. placebo 9.0 (p=0.003) – i.e., 37% reduction of risk of developing diabetes. The orlistat group also experienced significantly greater four year improvements in SBP and DBP, total LDL and HDL cholesterol and the LDL/HDL ratio as well as waist circumference, fibrinogen and plasminogen activator inhibitor 1 levels. It is not known if the effect would be greater if orlistat was combined with metformin and/or acarbose (Torgerson 2004; Torgerson et al. 2004).
Childhood obesity
A number of studies of the use of Xenical in children and adolescents are soon to be published – i.e., in the latter months of 2004.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 50
Ongoing trials
Although there are a large number of recently completed and new trials relating to the management of obesity (for a summary see Avenell et al., 2004, Appendix 9), most have little relevance to the ACC claimant population and/or do not involve any of the medications reviewed in the current study. Two ongoing studies are, however, worthy of mention as follows:
SCOUT: a large cardiac outcomes study of sibutramine has been planned in response to a request by the European Committee for Proprietary Medicinal Products (CPMP) for a post approval commitment (Nisoli and Carruba 2003). The Sibutramine Cardiovascular Outcome Trial (SCOUT) is a placebo controlled, double blind, global multi centre study with the objective of comparing the effect of sibutramine and standard care for weight management to that of placebo on the incidence of a composite cardiovascular outcome. It is hoped that the trials will recruit 9,000 subjects ≥ 55 years with a BMI ≥ 27kg/m 2 and < 45kg/m 2 or BMI ≥ 25kg/m 2 and <27kg/m with a waist circumference of ≥ 102cm in males or ≥ 88cm in females, at risk of a cardiovascular event.
The NICHD 18 trial of Xenical in children and adolescents with obesity related diseases – this is a phase II clinical trial designed to determine the safety, tolerability and efficacy of orlistat in African American and Caucasian children aged 12 17 years with one or more obesity related co morbidities (hypertension, hyperlipidemia, hepatic steatosis, insulin resistance, IGT, type II diabetes or sleep apnoea syndrome). The trial is expected to recruit 370 children 19 .
18 National Institute of Child Health and Human Development. http://clinicalstudies.info.nih.gov/cgi/detail.cgi?A_1998 CH 0111.html.
19 As this report was being written, the FDA announced that it had approved the use of Xenical in children.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 51 Other anti-obesity therapies
A number of additional anti obesity therapies not considered in the current review have recently been reviewed by NICE (Avenell et al. 2004) and by the US Preventative Services Task Force (McTigue et al. 2003). The full list of therapies considered in each of these reviews is listed in Table 20.
Therapies that may be of particular interest to the ACC include:
• Acarbose
• Metformin
• Very low calorie diets (VLCD)
• LCDs other than those contained meal replacement products
• Mazindol (appetite suppressant)
• Fluoxetine (anti depressant, SSRI).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 52
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 53 Horizon Scan
There are a number of potential future therapies for obesity:
• Leptin: is a hormone released from adipocytes which signals energy availability to the brain. Its deficiency leads to severe obesity and most obese individuals have been found to have elevated leptin levels. However, like most diabetics most obese people are leptin resistant (Brower 2002). There have been a number of phase 2 trials of this agent with disappointing results to date (Thearle and Aronne 2003).
• Axokine: is recombinant human ciliary neurotrophic factor which has been observed to cause weight loss and, in rodents, does not lead to rebound weight gain upon cessation. Early phase trials have indicated that weight loss in excess of placebo is obtained and that weight loss is maintained after treatment has ceased (Brower 2002, Ettinger 2003).
• Topiramate: is an anti epilepsy drug that has been evaluated for weight loss (Bray et al. 2003, McElroy et al. 2003). A recent dose ranging study has suggested that a clinically significant weight loss and improvement in obesity associated risk factors may be obtained with the therapy but longer and larger studies are required to fully assess its efficacy and safety (Thearle and Aronne 2003).
• Bupropion: is a peptide analogue of the fragment of human growth hormone that causes lipolysis.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 54
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 55 Economic considerations
The direct cost of medication
The cost of treating a single patient for a month at the recommended dose for each of the drugs considered in the current review is shown in Table 27.
Table 27. NZ Drug pricing according to MIMS, 2004† Manufacturer Cost‡ Patient Cost per Recommended Proprietary Price per patient Drug Pack Charge Tablet Daily dose d Drug name NZ$ per month NZ$ NZ$ (per tablet) NZ$ Phentermine Duromine a 15mg [30] 21.52 37.12 1.24 15mg/d 37.82 C5controlled drug 30mg [30] 24.58 42.40 1.41 30mg/d 43.01 Umine 30mg[100] 35.90 61.93 0.62 30mg/d 18.91 Diethylpropion Tenuate Dospan b 75mg[100] 94.88 163.67 1.64 75mg/d 50.02 Sibutramine Reductil 10mg[30] 103.35 178.28 5.94 10mg/d 181.17 15mg[30] 119.10 205.45 6.83 15/mg/d 208.32 Orlistat Xenical 120mg[84] 112.00 193.20 2.30 360mg/d 210.45
† Propharma wholesalers may quote lower prices for these drugs. a = 3M New Zealand Ltd. b = Douglas Pharmaceutical Ltd. d based on the manufacturer’s price. ‡ calculated using patient charges.
The estimated direct drugs costs of a course of treatment are shown in Table 28.
Table 28. The estimated cost of a course of anti-obesity treatment Maximum permitted Minimum period Proprietary Daily Drug Prescribed period Maximum weight loss Drug name Dose Time Period Cost Time Period Cost Phentermine Duromine a 15mg 3 months $113.46 3 months† $113.46 C5controlled drug 30mg 3 months $129.03 3 months† $129.03 Umine 30mg 3 months $56.73 3 months† $56.73 Diethylpropion Tenuate Dospan b 75mg 3 months $150.06 3 months† $150.06 Sibutramine Reductil 10mg 1 yr $2,174.04 6 months $1,087.02 2 yr $4,348.08 15mg 1 yr $2,499.84 6 months $1,249.92 2 yr $4,999.68 Orlistat Xenical 360mg 1 yr $2,525.40 6 months $1,262.70 2 yr $5,050.80 † not reported but assumed to be maximum continuous prescription period allowed.
The cost of monthly general practitioner (GP) and or nurse consultations, dietary modifications, ancillary monitoring of obesity related co morbidities, laboratory tests and dietitians services have not been estimated.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 56
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 57 Economic analysis
Obesity is a major risk factor for a number of different health conditions and the economic implications of obesity are therefore significant. Cost of illness studies in the USA suggest that obesity accounts for 5.5 7.0% of national healthcare expenditure (Hensrud et al. 2003).
A comprehensive search of the literature for the present review identified five systematic reviews examining the economics of treating obesity with one or more of the review drugs. Upon inspection these reviews all drew upon data from a relatively small pool of studies 20 (Foxcroft and Milne 2000; Lamotte et al. 2002), see Table 29.
Table 29. Studies reporting on the cost or cost effectiveness of sibutramine, orlistat, phentermine and diethylpropion Systematic review/HTA Drug Reported data from Economic analysis O’Meara 2001 O’Meara 2002 Sibutramine Foxcroft 1999 Avenell 2004 Cost-effectiveness Orlistat Foxcroft 2000 Lamotte 2002 BASF Pharma/Knoll 2000 Foxcroft 2000 (rapid review) Hensraud 2003 Orlistat Cost-effectiveness Lamotte 2002 (HE model) O’Meara 2002 Sibutramine BASF Pharma/Knoll 2000 Cost-effectiveness O’Meara 2001 Orlistat Foxcroft 2000 (rapid review) Cost-effectiveness Phentermine Glazer 2001 Sibutramine Pharmacy chain average retail price Comparative cost Orlistat HE=health economics
A single new economic analysis of sibutramine (Warren et al. 2004) was identified in the comprehensive search carried out for this review.
Economic evaluations of orlistat
Two systematic reviews (Foxcroft and Milne 2000; O'Meara et al. 2001) used effectiveness from three RCTs (Davidson 1999; Hollander et al. 1998; Sjostrom et al. 1998) reporting on interventions with 120mg of orlistat three times a day in combination with a hypocaloric diet and compared to placebo plus diet to evaluate the cost effectiveness of orlistat. The follow up period was one to two years and outcomes were focussed on patients who lost ≥ 5% of their initial body weight.
The perspective adopted for the economic analyses was that of the health service provider and the cost estimates included:
• the initial consultation
• laboratory tests
20 BASF Pharma Knoll (2000). Unpublished study of economic analysis of Sibutramine reported in O'Meara (2002).
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 58
• outpatient consultations (4 per year)
• drug costs.
The incremental cost per QALY gained was estimated to be £45,888 (range £19,425 £55,391). It was noted that the trial data were not consistent with the EMEA prescription indication for orlistat 21 and that the reported costs may differ from those obtained in clinical practice. It was not clear if the estimated cost of £45,888 was considered by the authors of the report to be cost effective; however, an economic analysis conducted in 2004 (Warren 2004) suggested a cost effectiveness threshold of approximately £20,000 for another anti obesity drug. Using this threshold, orlistat would not be a cost effective option in this group of patients.
A study by Lamotte et al., (2002) reported in European Euros ( €) and funded by Roche Pharmaceuticals focussed on obese patients with type 2 diabetes. Using an economic modelling approach, the incremental cost effectiveness of orlistat for two years in patients with diabetes but free of clinical events and without hypertension or hypercholesterolemia was estimated to be €19,986 22 . This was just considered to be cost effective against a threshold of approximately €20,000 per QALY but lacking robustness. The cost per life year saved decreased in patients with complications such as hypercholesterolemia and AHT, with the cost per life year saved reducing to €3,462 in patients with both hypercholesterolemia and AHT (see Table 30). Sensitivity analysis with inputs assuming that all weight is regained in 2.5 years rather than 5 years, increased the costs from €19,986 to €26,527 for event free patients and from €3,462 to €4,565 for patients with hypercholesterolemia. Treating only patients with hypertension and or hypercholesterolemia was considered to be good value for money, these results were robust under sensitivity analysis.
A study by Glazer (2001) reported only drug costs and direct costs per kg lost or 1% of base weight lost. However, these parameters were also reported for sibutramine and phentermine allowing a snapshot of comparative costs at a single point of time in the USA. The drug costs per kg weight loss were US$433, US$323 and US$91 for orlistat, sibutramine and phentermine respectively. The cost per 1% weight loss was US$433, US$268, US$89 for orlistat, sibutramine and phentermine respectively. These results suggest that orlistat and sibutramine will cost more than phentermine to reduce weight by a similar amount or by a similar proportion of initial body weight. The relative benefits and harms of therapy were not considered.
Economic evaluations of sibutramine
HTAs of sibutramine in 2002 and 2004 (Avenell et al. 2004; O'Meara et al. 2002) did not identify any published economic evaluations of sibutramine. A more recent comprehensive search for the present review (June/July 2004) identified one published economic analysis (Warren et al. 2004) and a simple comparative analysis of direct costs of a number of drugs including sibutramine by Glazer (2001). An unpublished cost utility analysis was submitted to NICE by the manufacturer BASF Pharma/Knoll (2000) and critiqued by O’Meara et al., (2002), see Table 29.
21 European Medicine’s agency prescription indication = loss of >2.5kg by diet in four weeks previous treatment and a loss of ≥ 5% body weight after 12 weeks of treatment.
22 Year 2000 Euros.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 59
A recent economic analysis by Warren et al., (2004) estimated the incremental cost utility of 12 months of sibutramine therapy at a dose of 10mg/d combined with diet and lifestyle advice compared to diet and lifestyle advice alone in “healthy obese” individuals.
The clinical pathways modelled in this study were consistent with the European licensing for sibutramine 23 . The overall incremental cost per QALY was estimated at £4,780 based on the results of the sibutramine adiposity therapy (SAT) trial (central Estimate A, Table 29). The marginal gain for the cohort of 1,000 patients in this report was 48.15QALYs (0.048QALYs per patient). Sensitivity analyses showed that this value was sensitive to a number of parameters – particularly the utility associated with weight loss and the frequency of monitoring during and after treatment, see Table 30.
Table 30. Sensitivity analysis around two base estimates of the cost-effectiveness of sibutramine when different utilities are assigned to weight loss UK Cost per QALY UK Cost per QALY Central Estimate A Central Estimate B Scenario‡ SAT trial SAMSA analysis† £4,780 £10,530 WL benefits only £6,341 £19,125 WL and CHD benefits only £5,403 £12,952 WL and diabetes benefits only £5,567 £14,664 Utility per kg WL = lower CI £14,072 £16,682 Utility per kg WL= upper CI £3,001 £5,965 Worst case £20,602 £34,260 Best case £2,950 £5,809 Compliance = 30% £8,605 £34,905 Compliance =50% £7,077 £21,482 Compliance =75% £5,738 £14,271 Adverse events not included £4,769 £10,505 Placebo=no monitoring costs £9,034 £19,899 Estimate A is based on the utility gain per kg of weight lost reported in the SAT trial. Estimate B is based on the utility gain per kg of weight lost reported in the Samsa analysis. † used in the NICE report ‡ Estimates for a cohort of 1,000 patients receiving 12 months treatment with sibutramine.
The effect of differing assumptions and utilities on the estimated cost of sibutramine is also evident in the cost per QALY reported using the different utilities assumed in an economic analysis reported by Samsa et al., (2001), see Table 29. Sensitivity analyses on the central estimates of the cost per QALY for sibutramine vs. placebo based on the utilities reported in the two trials showed large variations in the estimated cost per QALY.
The basic model used in this analysis assumes coronary heart disease (CHD) and diabetes benefits in addition to weight loss and the effect of removing these effects from the model is shown for CHD and diabetes. They both appear to have effects of similar magnitude on the cost per QALY and when the effect of both is removed the cost rises from£4,780 to £6,341 for the estimates based on the SAT utilities and from £10,530 to £19,125 for estimates based on the Samsa utilities.
23 The European licensing for sibutramine only allows continued treatment for “responders” – i.e., patients who lose 2kg after one month and 5% of their initial weight after three months.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 60
Overall, cost was most sensitive to the utility 24 attached to each kg of weight lost, however, even at the lower confidence interval of this variable, sibutramine was considered by the authors to be cost effective by both analyses. Removing the effect of adverse events from the model makes very little difference to either estimate of the cost per QALY.
Only two obesity related co morbidities, CHD and diabetes, were included in the analysis; however, the authors noted that if all co morbidities were included in the model the cost per QALY of sibutramine would decrease.
It may be concluded from these analyses that the cost effectiveness of sibutramine is dependent not only on the amount of weight lost but also on the effect the weight loss has on obesity related co morbidities. Further calculations were carried out to identify the cost of treating sibutramine induced side effects that would make treatment no longer a cost effective option. For Estimate A, it was calculated that side effects that cost £900 per sibutramine patient would be required. Given the fact that sibutramine is well tolerated and that the most common side effects reported are constipation and dry mouth, it was considered unlikely that such a high cost would be incurred in the treatment of side effects.
A “comparative cost efficacy” of sibutramine, orlistat and phentermine reported by Glazer et al., (2001) estimated that the cost per kilogram of weight lost using sibutramine (US$323) was similar to that of orlistat (US$433) and higher than phentermine (US$91). The cost of each 1% weight loss with sibutramine was approximately 40% lower than a comparable weight loss with orlistat ($US268 vs. US$433) and again both of the newer drugs had higher costs than phentermine (US$89). An even weight loss over time was assumed in all calculations.
The BASF Pharma/Knoll model, Table 29, incorporated four weight loss scenarios:
• the effect of weight loss alone
• the effect of weight loss on cardiovascular risk reduction
• the effect of weight loss on diabetes risk reduction
• the combined effects of all three factors.
All weight lost was assumed to be regained after five years. There was some concern that the original estimates made by the company were based on unjustifiably high utility values resulting in an over estimate of the likely gain in QALY per kg of weight lost.
24 Utilities are calculated from a relationship between weight loss and quality of life.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 61
Table 31. Economic evaluations of orlistat, sibutramine and phentermine Type of study Study Costs and assumptions Economic analysis Type of analysis ORLISTAT Foxcroft Rapid review Cost estimates Base case analysis 2000 Cost-effectiveness d Patient consultation + lab tests a£118 Cost utility per QALY gained = £45,881 Monthly cost of orlistat 120mg tds =£45 (range £19,452 - £55,391) GP consultation cost=£16 per visit Annual treatment costs=£7344 (1998) Estimated short run gain over a 2-year treatment period =0.016QALYs per year Assumptions Drop-outs treated on average for 3 months Drop-out rate 1yr=27%, 2yr 23% 4 outpatient appointments per year Glazer Systematic review of Costs No analysis 2001 effectiveness which Cost of drug per month =US$119 included a comparison of Cost of drug per year=US$ 1,428 drug cost Cost per 1% weight lost=US$ 433 Cost per 1kg weight loss=US$433 Assumptions WMD Drug-placebo %=3.4, kg=3.4 ORLISTAT Lamotte Primary research report Drug cost Cost per life year gained (LYG) 2002 Per patients per year = €881 1. Event free diabetic= €19,986 Cost effectiveness Incremental costs Drug-placebo 2. Hypercholesterolemia = €7,407 Markov model 1.Event free diabetes = €1,608 3. AHT= €7,388
2.Hypercholesterolemia= €1,514 4. Hypercholesterolemia +AHT= €3.462 3.AHT= €1,678 4. Hypercholesterolemia+ AHT= = €1,641 Cost-effectiveness Life years gained Depending upon the risk profile (Lamotte) 1.Event free diabetes = 0.08 1. Event free diabetes = €1,816 2.Hypercholesterolemia= 0.204 2. Hypercholesterolemia= €1,835 3.AHT= 0.227 3. AHT= = €1,918 4. Hypercholesterolemia+ AHT= = 0.474 4. Hypercholesterolemia+ AHT= = €1,955 Assumptions Effectiveness data from Clark 1998 (1.=9.462, 2.=9.401,3.=8.93,4.=8.74) public health perspective Metformin base treatment = 119 euro per patient per year. 4.2% of patients could stop taking oral medication for diabetes and 10% could reduce their medication by approx 25%. Maetzel Primary research paper Cost estimates (2001 values: US$) Incremental cost-effectiveness ratio (CER) 2003 ATG +Orlistat 120mg t.i.d. for 1 year =$US 19,987 Markov state transition ATG =$US 18,865 $US 8,327 per event free LYG model of obese MI =$US 19,226 individuals with type 2 Stroke = $US 32,458 Life expectancy increased by 0.13 years diabetes Microvascular disease =$US 1,248 over an 11 year time frame by orlistat Congestive heart failure= $US 4,842 Sensitivity analysis for the base case Cateract =$US 2,163 Showed that 95% of CER fell under the Main Assumptions threshold of cost effectiveness where the Orlistat treatment 120mg t.i.d. for one year threshold was set at approximately $US All patients received standard type 2 diabetes treatment – i.e., 20,000 per event-free LYG. sulphonylureas, metformin or insulin) Conclusion No diabetes related complication at start of orlistat treatment Orlistat added to diabetes treatment is cost Simulation continued for 11 years effective and that the CER is likely to be Male patients aged 52 years conservative as no account was taken of Reduction in HbA 1c values directly translate to a reduction of lipid parameter reductions or BP changes. diabetes related complications CER for females was likely to be higher but Perspective= US healthcare still less than $US20,000 Note: shorter period of orlistat have less benefit SIBUTRAMINE Glazer Systematic review of Costs No analysis 2000 effectiveness which Cost per month =$116 included a comparison of Cost per year=$1,392 drug cost Cost per 1% weight lost=$268 Cost per 1kg weight loss=$323 Assumptions WMD Drug-placebo %=5.0, kg=4.3
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 62
Table 31. Economic evaluations of orlistat, sibutramine and phentermine (continued) Type of study Study Costs and assumptions Economic analysis Type of analysis SIBUTRAMINE O’Meara Systematic review BASF Pharma /Knoll Cost per QALY through: O’Meara recalculated the cost per QALY as 2002 Modelling c 1.Weight loss alone=£14,700 1.Weight loss alone=£19,000 cost utility analysis 2.CHD risk reduction = £32,000 2.CHD risk reduction = £42,000 3.Diabetes risk reduction= £58,260 3.Diabetes risk reduction= £77,000 BASF Pharma /Knoll Combined cost per QALY gained from the three Combined cost per QALY gained from the UK British £ influences=£7,860 three influences = £10,500 (£5,700 - Main Assumptions £35,200) 1000 BMI>30kg/m free of co-morbities at the start of the modelling period. All weight is regained in 5 years Warren Primary research Costs Incremental cost per QALY 2004 publication Sibutramine 10mg per month=£35, 15mg £39.09 Combined WL, CHD , diabetes GP visit £13 per month for 1 year SAM estimate =£4,780 Economic modelling Nurse practitioner £7.50 per month for 4 years SAMSAT estimate =£10,530 based Marginal costs of 1000 patients receiving sibutramine over WL only placebo I= £305,314 (WL), £298,328 (WL + CHD), £288,776 SAM estimate =£6,341 (WL+ diabetes) SAMSAT estimate =£19,125 Cost saving from avoided CHD events WL and CHD Non-fatal x1 =£2,577, fatal x1 =£2,160, ongoing yearly cost of SAM estimate =£5,403 non-fatal =£6,19. SAMSAT estimate =£12,952 Cost saving from reduced diabetes incidence WL and diabetes only Annual per capita cost of treating diabetes=£780 SAM estimate =£5,567 SAMSAT estimate =£14,664 Main Assumptions Both arms diet and lifestyle advice Participants =“healthy obese’ WL, CHD and diabetes benefits are assumed 25 SAT marginal QALY gain over 1000 patients =48.15 SAMSAT marginal QALY gain over 1000 patients =26.76 SAMSAT analysis used a smaller utility than SAM analysis Cost effectiveness threshold = £20,000 Most common AEs constipation and dry mouth PHENTERMINE Glazer Costs No analysis 2000 Cost per month =$60 Cost per year=$720 Cost per 1% weight lost=$89 Cost per 1kg weight loss=$91 Assumptions WMD Drug-placebo %=8.1, kg=7.9 a (NHS, Trust 1998) b Based of trials of efficacy by Sjostrom 1998, Davidson 1999, Hollander 1998. c Smith 1994, James 1999. d Sjostrom 1998, Davidson 1999, e Diethylpropion was not considered to be effective in the only long-term trial considered and was not included in the comparison Aes = adverse events. ATG = adherence to diabetes guidelines for the duration of the model. AHT = arterial hypertension. BP = blood pressure. CER = cost effectiveness ration. CHD = coronary heart disease. LYG = life year gained. MI = myocardial infarction. QALY = quality adjusted life year. SAT= Sibutramine Adiposity Therapy Trial, SAMSAT= the Samsa review trials. WL = weight loss. WMD = weighted mean difference.
O’Meara et al., (2002) recalculated the cost effectiveness estimates. The recalculated estimates were 20 25% higher than the original values, see Table 30, and the combined effects cost per QALY for sibutramine reported as £10,500 with a best case to worst case scenario range between £5,700 £35,200 per QALY. The cost per QALY gained through CHD reduction alone was £42,000, though diabetes incidence reduction alone was£77,000 and from weight loss alone was £19,000. The cost per QALY was again very dependent on the starting utilities and a sensitivity analysis carried out with lower utility gains per kilogram (0.00048 for sibutramine and 0.00058 for placebo), resulted in a cost per QALY of £50,400. Assuming a hypothetical cost effectiveness threshold of £20,000 (Warren 2004), careful evaluation of the likely utility gain per kilogram of weight lost is required for treatment to be cost effectiveness.
25 Two estimates were made with one based on the SAT utilities and another on the SAMSAT utilities, sensitivity analyses were computed around both estimates.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 63
Economic evaluations of phentermine
No published economic analyses of phentermine were identified for the period (1996 2004). One study reported drug cost and cost per unit of weight lost comparisons between phentermine, orlistat and sibutramine (Glazer 2001). The estimated annual cost of phentermine was $720, and, on the assumption of a weighted mean difference between phentermine and placebo weight loss of 8.1% and 7.9kg, a cost per 1% weight loss of $89 and 1kg weight loss of $91. The drugs cost and direct cost 1% weight loss for phentermine was much lower than for orlistat or sibutramine.
Economic evaluations of diethylpropion
No published economic analyses or drug cost comparisons were identified over the period (1996 2004) for diethylpropion.
Economic evaluations of meal replacement plans
No published economic analyses or cost comparisons were identified over the period 1996 2004 for meal replacement products or plans.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 64
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 65 Discussion
Criteria for the evaluation of medication for the treatment of obesity
Obesity is the result of long term mismatches in energy balance where daily intake exceeds daily energy expenditure. It has been suggested (Halpern and Mancini 2003) that a useful medication for obesity treatment should:
1. be effective for body weight reduction and result in improvement of overweight dependent conditions
2. provide evidence that side effects are tolerable and/or transitory
3. not be addictive
4. provide proof of long term efficacy and safety
5. have a known mechanism of action
6. be reasonably affordable.
None of the weight loss strategies included in the current review fulfil all six requirements. However, all of the pharmacological therapies examined were effective to a greater or lesser degree for body weight reduction and the amelioration of obesity associated co morbidities and had known mechanisms of action (criteria 1 and 5).
All of the drugs except phentermine and diethylpropion were reported to be unlikely to cause addiction (criterion 3); caution is advised in the administration of phentermine which is an amphetamine and diethylpropion which is an amphetamine derivative. However, all of the drugs except orlistat were reported to have a CNS stimulatory effect and were contraindicated for patients with a history of drug or substance abuse.
All of the pharmacological interventions reported side effects that for the majority of participants appeared to be tolerable and/or transient (criteria 2). All weight loss studies reported high drop out rates and many reported withdrawals due to drug side effects; however, clinically significant side effects or serious adverse events were rare.
More recent weight loss medications such as orlistat and sibutramine were subject to more intensive reporting of safety and side effects than earlier drugs such as phentermine and diethylpropion. Sibutramine was recently subjected to a reassessment of its safety profile in Europe and a positive risk benefit for the drug affirmed. Reported side effects for orlistat were generally mild to moderate although a number of serious adverse events possibly related to therapy were reported. Phentermine and diethylpropion are no longer recommended for use for the treatment of obesity by the Royal College of Physicians in the UK.
There was very limited evidence upon which to judge the affordability of treatment (criterion 6). The cost per patient per month was higher for the newer drugs – i.e., orlistat and sibutramine than for phentermine and diethylpropion.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 66
There were conflicting estimates of the cost effectiveness of orlistat. There is some evidence to suggest that orlistat treatment in patients with a number of co morbidities is likely to be better value for money than treating otherwise healthy obese patients. Orlistat treatment of patients with type 2 diabetes who were free of clinical events and without hypertension or cholesterolemia was considered to be on the margin of cost effectiveness in one study while in another study, the estimated incremental cost per QALY of orlistat treatment appeared to lie beyond the UK cost effectiveness threshold.
In contrast the cost effectiveness of sibutramine, as estimated in two separative evaluations with differing assumptions and utilities, appeared to be within the UK threshold. However, an assumption of lower utility gains per kilo of weight lost for sibutramine resulted in costs that exceeded the threshold for UK cost effectiveness. If the same cost effectiveness threshold is assumed for New Zealand and ACC claimants, the circumstances of use for orlistat needs careful assessment if the treatment is to be cost effective. Sibutramine treatment is likely to be cost effective for a much wider range of patients, although monitoring for side effects may impact on this assumption.
Expected weight loss and duration of weight loss
Most of the studies reviewed agree that weight loss achieved with pharmacotherapy is modest. For most purposes the mean difference between placebo and each of the drugs in the current review was under 4kg and maintenance of weight loss without continued treatment problematic. However, experience from clinical trials indicates that most weight loss occurs within the first six months of treatment with weight loss at four weeks reported to be predictive of weight loss at six months for sibutramine. Discontinuation of treatment of all of the drugs reviewed is advised if reasonable weight loss does not occur within 12 weeks of starting treatment.
Potential barriers to use
Weight loss targets set by obese people are frequently unrealistic. In most clinical trials, weight loss ranges from 500g – 1kg per week for about six months before stabilising spontaneously. Highly restrictive diets (< 800kcal/d) have not proven to be more effective in the long term than more modest dietary interventions such as meal replacement. Reduction of 500 1000kcal per day is recommended. The potential barriers to use for phentermine, diethylpropion, orlistat and sibutramine have been reported in detail in an earlier sectionof the review (see drug descriptions pages 9 13). A summary is presented in Appendix I.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 67 Evidence Summary and Conclusions
More than half of the New Zealand adult population is overweight or obese. The problem is increasing in children and adults and is more prevalent in lower socio economic groups and amongst Maori and Pacific peoples. Given the prevalence of obesity and overweight amongst New Zealanders, a significant proportion of ACC clients are likely to be obese at the time of accident or injury or overweight and in danger of becoming obese as a result of lifestyle changes that may result from injury or accident. There are a number potential consequences for ACC and its clients.
Obesity is associated with a number of co morbidities including heart disease, diabetes, stroke, high blood pressure and certain cancers. There is also an increased risk of adverse events from anaesthesia in these clients and psychological problems such as clinical depression that may result in job discrimination and other employment difficulties.
Personal care requirements may increase, particularly in relation to attendant care and additional health interventions may be required for co morbid conditions initiated or exacerbated as a consequence of obesity. These problems may impact on quality of life and significantly interfere with treatment, recovery and rehabilitation.
Overall, obese clients are likely to require additional treatment and/or health related resources compared to clients who are not obese. The potential cost implications of the additional health care and resources that may be associated with untreated obesity are likely to be significant.
Clinical effectiveness and circumstances of use
Weight loss
For all of the drugs examined, most weight loss was achieved within the first six months of treatment and weight loss was not generally maintained when medication was withdrawn. Pharmacologically aided weight loss was generally modest but clinically important because of concomitant reductions in risk levels for obesity associated co morbidities. For long term maintenance of weight loss, the evidence to date suggests that obesity requires sustained therapy in much the same manner as other chronic diseases.
All of the drugs examined were used as adjuncts to diet and or behaviour therapy, physical exercise and lifestyle changes. In most studies, the weight loss drugs were added to a calorie controlled diet or a similar diet plan. There was great heterogeneity between the individual studies in terms of these accompanying non drug interventions.
All of the appraised studies were evidence level 1a or 1b according to the SIGN criteria.
Phentermine hydrochloride 15mg and 30mg is indicated as a short term adjunct in a medically monitored comprehensive regimen of weight reduction based on exercise, calorie controlled diet and behaviour modification in patients with a BMI ≥30kg/m 2 who have not had a clinical response to an appropriate weight reduction program alone. Patients with obesity related co morbidities such as sleep apnoea, insulin resistant diabetes mellitus, pre diabetes mellitus or IGT or high cardiovascular risk status, and have a BMI of less than 30kg/m 2 may also be considered for treatment.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 68
Phentermine may be prescribed for up to 12 weeks when used as an adjunct in a medically monitored regimen of weight reduction. Failure to achieve a weight reduction of 5% within 12 weeks is an indication for discontinuation. Treatment may be continued beyond 12 weeks provided the patient is monitored for weight loss and medical conditions, and as long as weight loss is maintained.
The effectiveness of phentermine has been examined in a number of studies but most significantly in a comparative meta analysis of four decades of published randomised trials of anti obesity treatments. Patients who were prescribed phentermine lost on average 6.3 kilograms with a placebo subtracted average weight loss of 3.6 kilograms (mean follow up 13.2 weeks). Weight loss was maintained for an unspecified period 26 after the formal study completion at a placebo subtracted average of 2.43 kilograms. When the effect size for phentermine was compared to other anti obesity drugs, it was second to sibutramine and higher than either orlistat or diethylpropion. However, the confidence intervals for all four drugs overlapped suggesting that there was no significant difference in average weight loss between the medications.
Diethylpropion hydrochloride 75mg is an appetite suppressant with a recommended daily dose of 75mg taken once daily, mid morning. It is indicated for short term, intermittent use in a medically monitored regimen of weight reduction based on exercise, calorie restriction and behaviour modification in obese patients with a BMI ≥ 30kg/m 2 who have not had an adequate response to an appropriate weight loss program of diet and/or exercise alone. Patients with a BMI of less than 30kg/m 2 with co morbidities including sleep apnoea, insulin resistant diabetes mellitus, pre diabetes mellitus or IGT or high cardiovascular risk status, may also require medical assistance with weight loss. These patients may also be considered for treatment with diethylpropion. It is not recommended for children under 18 years.
Diethylpropion may be prescribed for courses of up to 12 weeks, with intervening periods of one month without treatment. Failure to achieve a weight reduction of 5% within a period of 12 weeks is an indication for discontinuation of treatment.
The effectiveness of diethylpropion was reported in a number of small studies and a single meta analysis of four decades of randomised trials of weight loss medication. The results varied considerably with both positive and negative placebo subtracted weight loss – i.e., superior and inferior weight loss when compared to the placebo group. In the most rigorous study, comprising a meta analysis of nine randomised trials of diethylpropion, effectiveness varied with some studies reporting a weight gain in comparison with the placebo group. Diethylpropion produced a 3.0 kilogram greater weight loss than the placebo with an effect size of 0.57. In a comparison of the effect sizes for diethylpropion, phentermine, orlistat and sibutramine, diethylpropion had a lower reported effectiveness than sibutramine and phentermine but a higher reported effectiveness than orlistat. The weight loss achieved by all of the drugs was similar and the confidence intervals for effect size overlapped.
Orlistat 120mg is indicated as a long term treatment for obesity, weight loss maintenance and prevention of weight regain in adults with a BMI ≥30kg/m 2. It should be used in conjunction with a low fat calorie controlled diet that is nutritionally balanced and contains approximately 30% of calories from fat distributed over three meals. The diet should be rich in fruit and vegetables. The recommended daily dose is one 120mg capsule three times a day at mealtimes.
26 Mean follow up assessment for all drugs studies = 17.3 weeks, median=6 weeks after the formal study completion.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 69
The efficacy of orlistat in children 27 and adolescents below the age of 18 years has not been established. There are currently no safety and efficacy data beyond two years and the duration of treatment should not extend beyond this period.
As one of two relatively new anti obesity medications with a novel mode of action, orlistat has been subjected to close scrutiny. A large number of relevant studies were identified including 10 systematic reviews and seven additional eligible RCTs. One of the systematic reviews was a Cochrane Review and two were performed as part of an overall HTA commissioned by the NHS R&D HTA Program on behalf of NICE in the UK (Avenell et al. 2004; O'Meara et al. 2001). Some studies examined the effect of treatment with orlistat over a period of up to two years.
Patient populations varied considerably between the studies. For the purposes of this review the results have been summarised for four patient populations (a) all patients, mixed co morbidity risk (b) healthy obese patients (c) diabetic or glucose intolerant patients and (d) hypercholesterolemic patients.
For all patients (mixed risk groups), the placebo subtracted weight loss was 3.01 kilograms after 12 months treatment and 3.26 kilograms after 24 months of treatment. In studies reporting otherwise healthy patients only the placebo subtracted weight loss was similar with a 2.5 2.9 kilogram loss after 12 months and 3.6 kilogram loss at 18 months. Clinically important weight loss – i.e., a weight loss of ≥ 5% was reported at one year for 58 78% of healthy obese patients.
For studies reporting on weight loss in diabetic or glucose intolerant patients, a clinically significant weight loss was reported in 30% of orlistat + diet treated patients against 13% of patients treated with placebo + diet. This was however, somewhat lower than that reported for the non diabetic group where 45% of patients achieved a clinically important weight loss. In a pooled analysis of diabetic and non diabetic patients, one HTA reported a placebo subtracted weight loss of 1.8 kilograms and 2.5 kilograms for diabetic and non diabetic patients respectively.
A meta analysis of diabetic only patients reported a greater proportion (72%) of patients improving their glucose tolerance with orlistat treatment than the placebo group (49%) and less than half as many patients (3.0% vs. 7.6%) with IGT progressed to diabetic status after treatment with orlistat.
Patients with hypercholesterolemia who had been treated with orlistat were reported to have lost up to 3kg more on average than those treated with placebo with significant improvements reported in LDL cholesterol and total cholesterol; there was also evidence of a significant reduction in SBP and DBP.
Sibutramine hydrochloride 10mg and 15mg may be used as a long term treatment and management of obesity including weight loss and weight maintenance in patients with an BMI ≥ 30kg/m 2 or ≥ 27kg/m 2 in patients who have diabetes, dyslipidaemia or hypertension. It is intended for use when patients have not adequately responded to appropriate weight reducing therapy such as hypocaloric diet and/or exercise alone – e.g., patients having difficulty in achieving or maintaining ≥ 5% weight loss within three months. The recommended starting dose is 10mg once daily with or without food; in clinical trials sibutramine was given in the morning. If there is less than 2kg weight loss after four weeks, and the 10mg dose is well
27 As this report was being completed, the FDA announced that Xenical ® had been approved for use in children.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 70 tolerated, the daily dose may be increased to 15mg. Sibutramine is not currently intended for general use in obese children under 18 years or in elderly patients over 65 years.
Patients who have demonstrated an adequate response – i.e., weight loss of ≥ 2kg within four weeks at a defined dose; usually achieve a maximum weight loss (5% 10% of initial body weight) after six months of continuous treatment with sibutramine. The current data sheet supports treatment for up to one year, however, the evidence of long term safety and efficacy reported in the STORM study led to the US FDA to extend its clearance of sibutramine therapy to two years. Treatment must be discontinued in patients who have not had an adequate response – i.e., patients whose weight loss stabilises below 5% of initial body weight or whose weight loss within three months of starting treatment is less than 5% of their initial body weight. Treatment should not be continued in patients who regain 3kg or more after previous weight loss with sibutramine.
As one of two relatively new anti obesity medications, sibutramine has been subjected to close scrutiny and a large number of relevant studies were identified including 10 systematic reviews and seven additional eligible RCTs. Patient populations varied considerably between the trials, as did the reporting of results. Study size varied considerably with RCTs varying between 34 389 participants and systematic reviews, HTAs and meta analyses between 160 4,528 participants. There was considerable duplication and overlap between the studies and considerable heterogeneity in the reporting of results.
The most recent HTA reported a significant placebo subtracted average weight loss of 4.12 kilograms after 12 months of treatment. The STORM trial reported a placebo subtracted average weight loss of 3.4 kilograms after 18 months of treatment. For patients without significant obesity related co morbidities a clinically significant weight loss – i.e., ≥ 5% of initial body weight was reported to be 21 34% higher for the sibutramine treated group than the placebo treated group, the placebo subtracted weight loss was between 3.0 4.3 kilograms.
For patients with type 2 diabetes, one study reported a clinically important weight loss difference between the treated groups and the placebo group of 45% at a dose of 15mg/d and 65% at a dose of 20mg/d. In another study, diabetic patients were reported to have lost 8.7 kilograms more weight than the placebo group (p<0.0001) after six months of treatment. In all other studies where stratification and sub group analyses of diabetic patients were reported, a significant difference between the sibutramine treated diabetic group and the placebo treated diabetic groups was reported.
For patients with mild to moderate hypertension, 12 weeks of treatment with 10mg/d of sibutramine resulted in 44% of patients achieving a clinically important weight loss compared with 17% in the placebo group (p=0.01). In a meta analysis of 21 RCTs, sibutramine showed a large effect size ( 1.0, range –1.17 to 0.84) but increased blood pressure significantly. Most other studies of sibutramine actually excluded patients with hypertension or cardiovascular disease, others only allowed participants with well controlled hypertension.
The effect of sibutramine varies with dose. The doses of sibutramine used in most studies in this review varied between 5 30mg/d. In some studies, a sub group analysis was carried out for different dose levels, this was however uncommon. One study specifically designed to examine the effect of dose on treatment response was identified. It included a systematic review of 10 RCTs of 1,047 obese patients treated with doses of sibutramine ranging from 1 30mg/d. A significant dose response effect was reported after six months of treatment with the proportion of patients losing ≥ 5% of their initial rate varying from 25% at a dose of 1mg/d to 77% at a dose of 30mg/d. The proportion of placebo treated participants reporting a weight loss of ≥ 5% of initial body weight was 20%. The proportion of patients who withdrew because of adverse events, increased with increasing sibutramine dose levels. Treatment related adverse
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 71 events were more common in the 30mg/d group. Weight loss at four weeks was found to be predictive of weight loss at six months.
A recent dose ranging trial of sibutramine in obese adolescents between the ages of 13 17 years was also identified. A placebo subtracted average weight loss of 4.6 kilograms (95 % CI, 2.0 7.4, p=0.001) was reported in an ITT analysis after six months of treatment. Medication was reduced or discontinued in a significant number of sibutramine treated patients due to increases in blood pressure, pulse rate and other symptoms. Although sibutramine treatment was successful, the authors cautioned against the use of sibutramine in adolescents outside clinical trials until further efficacy and safety data were available.
Meal replacement plans
Meal replacement strategies are often included in LCDs (diets in the range of 1200 1600kcal/day), which are the cornerstone of modern weight control efforts. However, there are no established definitions of meal replacements plans. In the scientific literature, “meal replacement” is a term generally used to cover beverages, pre packed, shelf stable and frozen entities and meal or snack bars. Most of the products are fortified with vitamins and minerals and designed to be consumed in place of one or more regular meals.
Searches for specific meal replacement products such as Complan and Ensure did not identify any published scientific literature relating to these products between 1996 2004. A wider search for meal replacement plans revealed a similar paucity of information.
Meal replacement eating plans have only very recently been critically evaluated for safety and efficacy. The first meta analysis evaluating RCTs of PMR plans was published in 2003. In this study of 276 potentially relevant publications, 30 were classified as meal replacement studies proper and only six of these studies met the review criteria. A meta analysis of these studies estimated that the PMR group lost an average of 2.54 kilograms more than the control group after three months of treatment and 2.43 kilograms more than the control group after 12 months of treatment. The proportion of subjects losing ≥ 5% of initial body weight after three months was 33% for the control group and 72% for the PMR group (p<0.001).
After stratification according to diabetic status, non diabetic completers lost on average 2.79 kilograms more than the control group after three months (p<0.001) and 3.17 kilograms more after 12 months (p=0.002). The diabetic group completers lost on average 2.46 kilograms more than the control at three months and 2.76 kilograms more than the control at 12 months; neither differences reached significance, however, the number of diabetic cases was small and the standard error for the group large.
A single additional product and RCT was identified by the comprehensive search strategy constructed for the current review. One hundred overweight or obese participants were randomly assigned to a novel soy meal replacement program plus dietary counselling or 1200kcal exchange diet program plus dietary counselling. After 12 weeks, the treatment group lost significantly more weight than the control group (7.00kg vs. 2.90kg, p<0.001, ITT analysis). Fat mass was also significantly lower in the treated group 4.3 vs. 1.4 (p=0.003).
Obesity related risk factors
No data were reported on the effect of phentermine, diethylpropion or meal replacement therapy on obesity related co morbidities and their associated risk factors.
Orlistat induced weight loss was accompanied by significant and favourable changes in a number of CHD risk factors including total cholesterol, LDL cholesterol, blood pressure and
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 72 glucose metabolism. When diabetic and glucose intolerant patients were treated with orlistat, a high proportion of diabetics improved their glucose tolerance and there was a reduction in the proportion of patients with IGT progressing to diabetic status.
Sibutramine treated patients also experienced significant and favourable changes in obesity related risk factors. Significant improvements were reported for HDL cholesterol and triglyceride levels and non significant improvements for LDL Cholesterol, HbA 1c and fasting glucose levels. However, blood pressure levels on average rose during treatment with sibutramine and in some instances there was a significant difference in favour of the control group. In patients with type 2 diabetes, blood pressure increases were reported to be offset by weight loss while in hypertensive patients blood pressure reductions were reported with sibutramine patients showing slightly lower mean improvements than the placebo group. Most studies of sibutramine excluded patients with hypertension or cardiovascular disease or only allowed well controlled hypertension and further data may be required to determine more fully the effect of this drug on patients with raised blood pressure. In dose ranging studies, blood pressure and heart rate increased with increasing dose levels of sibutramine.
Safety and side effects
Overall, the therapies examined in this review were reported to be well tolerated and mostly without significant side effects or serious adverse events. However, safety and side effects have only been reported for highly selected patient populations in closely monitored clinical trials and the results obtained may not apply to typical clinical populations or the ACC claimant population.
Early studies of phentermine suggested that it was well tolerated with only minor adverse events reported for 3 8% of patients. Stimulant effects such as agitation and insomnia were reported. Concerns have been raised about the safety of phentermine and the drug is no longer recommended by the Royal College of Physicians in the UK.
Diethylpropion was reported to have few stimulant adverse events; rapid tolerance to the drugs anorectic effects was reported. Concerns have been raised about the safety of diethylpropion and the drug is no longer recommended by the Royal College of Physicians in the UK.
Orlistat is a locally acting drug and the main adverse events reported were gastrointestinal. Commonly reported side effects were fatty, oily or liquid stools, faecal urgency or incontinence and flatulence, reported in 22 95% of orlistat users. Most GI adverse events occurred early in the course of treatment and may be moderated by dietary adjustments, nevertheless, some of the GI side effects may be unacceptable to some groups of patients. Other undesirable effects attributed to orlistat were lower serum levels of fat soluble vitamins and interference with the effectiveness of warfarin and cyclosporine treatment.
Sibutramine was generally well tolerated by the study populations at the recommended starting dose of 10mg/d. The most commonly reported side effects were headache, constipation and nausea. Adverse events associated with the CNS including dizziness, dry mouth and insomnia were reported by more than 5% of patients.
The two most clinically significant side effects were increased blood pressure and tachycardia. These symptoms were generally seen in the first eight weeks of treatment and although in some instances the increases reached statistical significance they were not generally considered to be clinically important in otherwise healthy obese patients. Sibutramine is contraindicated in patients with uncontrolled hypertension or cardiovascular disease.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 73
After a number of adverse events and two deaths the benefit/risk ratio of sibutramine was reassessed by the EMEA in 2002; a positive benefit risk profile for the drug was reaffirmed by the agency after further investigation.
As significant safety and efficacy data for orlistat and sibutramine does not extend beyond two years there is no evidence of their long term safety.
There are few reports on the safety and side effects of meal replacement plans. In the two studies evaluated in the current review, one (Heymsfield et al. 2003) reported a lack of adverse events in any study patients, including treated diabetics, while the other (Allison et al. 2003) reported a large number of adverse events. In the latter study, which was an unblinded RCT of a novel soy based meal replacement formula for weight loss, constipation, gas or indigestion, abnormal or metallic taste, lethargy and weight loss were significantly higher in the meal replacement group. Generally however, treatment was reported to be well tolerated and free of serious side effects by the study authors.
Ethnic groups
Some ethnic groups are more at risk of obesity than others and BMI thresholds for obesity are known to vary with ethnicity, as are the co morbidity risks. However, no studies of the efficacy of phentermine, diethylpropion, orlistat, sibutramine or meal replacement in different ethnic groups were identified. Most of the participants in the reported RCTs were white Caucasian females and the results obtained and the conclusions reached may not be applicable to other ethnic groups. A higher obesity threshold is advised for Polynesians and a lower threshold for Asians.
Economic considerations
Orlistat and sibutramine monthly drug costs are similar at $210.45 and $181.17 $208.32 (depending on dosage) respectively, the cost of phentermine and diethylpropion medication are much lower at $18.91 and $37.82 $43.01 respectively. Additional direct costs including GP visits and laboratory tests are likely to be the same for each of the drugs and are unlikely to exceed the monthly drug costs.
There were no published economic analyses of phentermine or diethylpropion, however, in one comparative study the estimated cost per kilogram of weight lost for phentermine, orlistat and sibutramine was US$91, US$433 and US$323 respectively. This is roughly in line with the differential daily NZ direct cost of the drugs.
Economic evaluations of orlistat from two systematic reviews suggested that the overall cost of orlistat therapy is high with the incremental cost per QALY gained estimated to be £45,888 (range £19,425 £55,391) in one study and approximately £31,978 in another. Overall, treating healthy obese patients was much less cost effective than treating patients with obesity related co morbidities. Treating only patients with obesity related hypertension and/or hypercholesterolemia was considered to be good value for money and the results of the analysis robust under sensitivity analysis.
Only one published economic analysis of sibutramine was identified. Two central estimates of cost per QALY gained were reported arising from differing assumptions and utilities associated with weight loss.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 74
The costs per QALY arising from the two calculations were £4,780 and £10,530. However, the basic economic model used in both analyses assumed benefits additional to weight loss arising from associated improved risk profiles in patients with co morbidities. The costs per QALY of weight loss alone were £6,341 and £19,125 and worst case scenarios estimated at £20,602 and £34,260 respectively in the two models. Sibutramine was considered to be cost effective even at the highest cost per QALY obtained in this analysis and it was noted that if all co morbidities were included in the model, the cost per QALY would decrease.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 75 Evidence Tables
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 76
Evidence Table 1. Meal replacement, Heymsfield, 2003
Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (Mean change from baseline) Safety Comments and year Studies Exclusion Wt-loss dif § s.e. p value Duration Heymsfield, Study design Participants N=487 Exposure (PMR) Results at 3 months Baseline comparison Within the context of a clinical 2003 Meta-analysis of six Male= 25% PMR – i.e., one or Meta-analysis fixed effects 3.01 0.33 <0.001 PMR and RCD differed trial, PMR appears to promote RCTs Female: =75%, more meals replaced Meta-analysis random effects 2.60 0.96 0.006 significantly in one factor – significantly greater weight loss Mean age 46.1 years by a commercially Pooled analysis of completers 2.54 0.37 <0.001 baseline glucose. than a corresponding RCD plan. Objective Mean BMI at baseline available, calorie- Pooled analysis LOCF 2.39 0.35 <0.001 Drop-out rate The overall magnitude of the To assess the =31 reduced product(s) Pooled analysis MI 2.23 0.35 <0.001 3 months (ns) weight loss in a pooled analysis at safety and Diabetic N=119 (24%) that are fortified with >=5% weight loss RCD=34%, <0.001 Exposure = 16% 1 year was in the range reported effectiveness of Non-diabetic vitamins and minerals PMR=72% Control=19% in some pharmacological weight meal replacement N=368(76%) and at least one daily Non-diabetic 1 year (p<0.001) control studies (Glaser 2000) and therapy meal consisting of Pooled analysis of completers 2.79 0.37 <0.001 Exposure=47% is at the level known to lower Inclusion regular foods, acting Pooled analysis LOCF 2.67 0.35 <0.001 Control=64% disease risk. Studies N=6 Randomised trial as a LCD and Pooled analysis MI 2.47 0.33 <0.001 None of the drop-out reported Ditchuneit et al., comparing PMR to a providing Diabetic program or product-related Four studies enrolled subjects 1999** traditional LCD >800<=16000kcal/day Pooled analysis of completers 2.46 1.84 ns adverse events. with no co-morbidities, the other 2 Fletcher-Mors et al., Study duration > Pooled analysis LOCF 2.62 1.89 ns Heterogeneity were designed for patients with 2000** 3months Comparison (RCD) Pooled analysis MI 1.67 1.73 ns Estimates of weight loss and type 2 diabetes. Rothacker 2001 Subjects >18 years Traditional LCD plan Results at 1 year (RCD N=4, PMR effectiveness of both groups Yip 2001 (LCD) with the same N=5) from the individual studies All six studies resulted in Hensrud 2001 Exclusion caloric composition as Meta-analysis fixed effects 3.39 0.72 <0.001 were heterogeneous significant weight loss from Ashley 2001 Studies with self the exposure. Meta-analysis random effects 2.43 1.65 0.142 (p=<0.001). baseline in both groups. Ahrens 2000 reported data. Pooled analysis of completers 2.63 0.88 0.003 Quality Pooled analysis LOCF 2.86 0.46 <0.001 The quality of all studies as PMR groups lost significantly Duration Pooled analysis MI 1.62 0.98 0.119 assessed by the Jadad criteria more weight than the control 3-51 months, median 12 >=5% weight loss RCD=33% <0.001 = moderate. group. months PMR=74% Blinding was not possible due Non-diabetic to the nature of the Note: % weight loss at 3 months Location Pooled analysis of completers 3.17 0.99 0.002 interventions. RCD=4%, PMR=7%. % weight USA (n=5) Pooled analysis LOCF 3.56 0.50 <0.001 Blinding of the outcome loss at 1 year RCD . 3-7%, Germany (n=1) Pooled analysis MI 2.92 0.66 <0.001 assessor was not described. PMR .7-8% Diabetic Randomisations description Note : 2 meals a day Pooled analysis of completers 2.76 2.00 ns was included in all protocols The first systematic evaluation of were replaced in the Pooled analysis LOCF 1.52 1.89 ns and all studies provided details RCTs evaluating PMR plans for exposure during the Pooled analysis MI 2.86 2.53 ns of dropouts. weight management suggests that weight loss phase, one they can safely and effectively meal a day during the Risk factors ⊥⊥⊥ 3 moWL p= 6 moWL p= PMR-RCD p= produce significant weight loss. weight maintenance Glucose (mg/dl) 0.028 0.009 ns There was no support for phase. SBP (mmHg) <0.001 <0.001 ns publication bias. DBP (mmHg) ns ns ns Insulin ( ΦU/ml) ns ns <0.001 Cholesterol (mmol/l) ns <0.001 ns HDL (mmol/l) ns ns ns LDL (mmol/l) 0.039 0.001 ns TG (mmol/l) 0.014 0.011 ns
*One trial >=30 years ** same study populations ns= no significant difference between groups § PMR = partial meal replacement, RCD= reduced calorie diet. Note: All risk factor values for both groups were significantly different from baseline values at 3 and 6 months, 3moWL= 3 month weight loss, 6moWL= 6 month weight loss. MI=multiple impatation, LOCF=last observation carried forward.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 77
Evidence Table 2. Meal replacement, Allison et al., 2003 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) (Mean change from baseline) Safety Comments and year Study Location(s) Exclusion Treatment Control ∈∈∈T-C Duration Recruitment (s.d.) (s.d.) P= Allison et al., Study Design Participants Exposure N=50, 12 week assessment (ITT)* Primary analysis ITT. Data collection on side 2003 RCT N=100 (120 screened) M=10, F=40 Weight change (kg) -7.0(4.6) -2.9(3.3) 0.001 Completers only analysis also undertaken, effects was collected BMI 28-41kg/m 2 Scan diet meal Total body fat mass (%) -1.5(3.6) -0.2(4.8) ns completers = 37/50 (74%) in each group. according to a Objective Age 35-65 replacement formula, Fat mass (kg) -4.3(4.0) -1.4(4.8) 0.003 standardised protocol. To assess the efficacy Male =20 5 Scan diet shakes + Waist circumference (cm) -6.0(4.2) -2.9(3.7) 0.003 Baseline comparison = ns and safety of a low Female =80 4 exchanges of fruit, Total cholesterol (mg/dl) -22.5(30.2) -6.8(24.7) 0.013 25% of the participants calorie soy-based 4 exchanges of LDL cholesterol (mg/dl) -21.2(23.5) -7.1(19.1) 0.009 Between group differences did not complete. meals replacement Inclusion vegetables and one HDL cholesterol (mg/dl) -1.5(7.8) -0.5(10.6) ns The protein, carbohydrate, fat proportions Food records were not program for the Overweight or obese fat exchange per day DBP (mmHg) -1.3(11.0) 0.9(6.6) ns between groups differed kept and compliance treatment of obesity. subjects instructions for use + SBP (mmHg) -4.8 (28.4) -1.5(12.3) ns Exposed = 33%: 52%: 15% was not checked in the Medically fit for safe a single session Control = 23%: 56%: 21% control group. Study location(s) weight loss dietary counseling + ** All assessments at 4 and 8 weeks USA a pamphlet were ns except for: Missing data - handled via multiple The study was not Exclusion describing good Fat mass at 8 weeks p=0.011,total imputation rather than LOCF powered to pick up rare Duration Weight loss of >5kg in weight loss practices. cholesterol at 4 weeks p=0.002, 8 adverse events. 12 weeks the past 3 months weeks p=0.0001,LDL cholesterol at 4 Eligibility - 115/120 screened patients were Use of weight loss drugs Comparison N=50 and 8 weeks p=<0.0001 eligible Conclusion in the last 6 weeks. M=10, F=40 Treatment Control Randomisation - 15/115 were not Soy protein meal Scores above the 90 th 1200kcal exchange Adverse events (MOSES) a mean mean p- value randomised. 6= BMI out of range, replacement LCD percentile on the Brief diet system. As assessed at 12 weeks 4=BSI>90%, 1= taking anti-obesity reduces body weight; Symptom inventory A single session Appetite: decreased/anorexia medication, 1= non-compliant, 3= medical total and LDL Disease not believed to dietary counselling Constipation 0.89 0.46 ns conditions cholesterol significantly be at least partially the and a pamphlet Diarrhoea 0.51 0.24 0.04 more than a LCD result of obesity and describing good Drooling/salivation 0.27 0.11 ns Withdrawals N=5 without a meal treatable by weight weight loss practices. Gas/indigestion 0.03 0.03 ns Diarrhoea =3 (1 with nausea) replacement product. reduction. Medical or Taste: abnormal/metallic Gastric reflux=1 1.59 0.51 0.0009 psychological Lethargy/no movement Other=1 contraindications Sleep: excessive 0.30 0.00 0.0223 Known hypersensitivity to Urinary: enuresis/nocturesis 0.19 0.00 0.0426 any of the ingredients of Weight gain 0.11 0.08 ns the formula. Weight loss 0.16 0.11 ns 0.08 0.14 ns Recruitment Note: 13 missing values for treatment 1.30 0.62 0.0208 Details not given. group and 13 missing values for control group at week 12. DBP=diastolic blood pressure, SBP=systolic blood pressure, ITT=intention to treat analysis, MOSES= monitoring of side effects scale (MOSES; Kalachnick, 1985) higher values are indicative of greater symptom severity.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 78
Evidence Table 3. Orlistat, Torgerson et al., 2004 Study Study Design Participants Exposure/ Outcome Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (Mean change from baseline) Safety Comments and year Study Location(s) Exclusion P O p=** Duration Torgerson et al., Study Design Participants Exposure Weight change at 1 yr (kg) 6.2 10.6 <0.001 Compliance was determined Compared with lifestyle changes 2004 Randomised, double N=3305 (placebo=1637, Orlistat 120mg t.i.d. + Weight change at 4 yr (kg) 3.0 5.8 <0.001 by counting the number of alone orlistat + lifestyle changes blind, placebo orlistat =1640) reduced calorie diet Completers only 1 yr (kg) 7.5 11.4 <0.001 capsules returned. resulted in a greater reduction in controlled multicentre Male=45% (800kcal/day deficit). Completers only 4 yr (kg) 4.1 6.9 <0.001 weight and a greater reduction in study Female=55% >=5% weight loss 1 yr (%) 45.1 72.8 <0.001 Withdrawals the incidence of type 2 diabetes Mean age 43.7 years Comparison >=5% weight loss 4 yr (%) 37.3 52.8 <0.001 Refusal of treatment, over 4 years. Objective (placebo), 43.0 years (orlistat). Placebo + reduced >=10% weight loss 1 yr (%) 20.8 41.0 <0.001 placebo=20% To determine the effect Normal glucose tolerance calorie diet >=10% weight loss 4 yr (%0 15.6 26.2 <0.001 orlistat=14%. Withdrawals The beneficial effect of orlistat in of adding orlistat to =79% (800kcal/day deficit). Cardiovascular risk factors (yr 1) due to insufficient therapeutic preventing diabetes was primarily lifestyle changes on IGT = 21% DBP (mmHg) -2.6 -3.6 <0.01 response, placebo=19%, due to a beneficial effect in IGT body weight and the SBP (mmHg) -5.2 -7.3 <0.01 orlistat=8%. patients. incidence of type 2 Inclusion Total cholesterol (%) -1.3 -8.8 <0.01 diabetes. Age 30-60 years LDL cholesterol (%) -1.6 -11.4 <0.01 Completers and non- The long-term safety or orlistat BMI >=30kg/m 2 HDL cholesterol (%) 8.5 3.4 <0.01 completers were not was demonstrated. Study location(s) Non-diabetic glucose LDL-HDL ratio -0.3 -0.5 <0.01 significantly different. Sweden tolerance or IGT (fasting Triglycerides (%) -6.3 -6.2 <0.05 1 During 4 years of treatment venous plasma glucose Waist circumference (cm) -7.0 -9.6 <0.01 Adherence orlistat + lifestyle changes Duration <7.8mmo;/l and 2 hr plasma Fasting whole blood* glucose (mmol/l) 0.2 0.1 <0.01 2 Placebo=92.8 significantly decreased the Four years glucose 7.8-11.1mmol/l) Fasting serum insulin (mmol/l) -17.0 -26.5 <0.01 Orlistat= 93.3% progression to type 2 diabetes Fibrinogen (µmol/l) 0.1 0.2 ns compared with placebo + lifestyle Exclusion Plasminogen activator inhibitor-1 (U/ml) -3.0 -7.1 <0.01 Safety change (log-rank p=0.0032). Diabetes Cardiovascular risk factors (yr 4) Orlistat was well tolerated There was a 37.3% decrease in Ongoing/ active DBP (mmHg) -1.9 -2.6 <0.01 and the overall incidence of the risk of developing diabetes cardiovascular disease SBP (mmHg) -3.4 -4.9 <0.01 adverse events was the with orlistat compared to placebo. Ongoing/active GI disease Total cholesterol (%) -2.3 -7.9 <0.01 same in both groups except LDL cholesterol (%) -5.1 -12.8 <0.01 for GI events. Most GI events The difference in weight loss was HDL cholesterol (%) 9.1 6.5 <0.01 3 were mild to moderate and the same whether assessed by LDL-HDL ratio -0.4 -0.6 <0.01 occurred during the early LOCF or BLCF analysis. Triglycerides (%) 2.9 2.4 ns phase of treatment. Waist circumference (cm) -4.4 -6.4 <0.01 Over the four-year period Only 52% of the treatment group Fasting whole blood* glucose (mmol/l) 0.2 0.1 <0.01 4 both groups had at least one and 34% of the placebo group Fasting serum insulin (pmol/l) -20.6 -32.0 <0.01 serious adverse event (13% were retained to the 4 th year. Fibrinogen (µmol/l) -0.5 -0.4 <0.01 4 placebo, 15% orlistat). Plasminogen activator inhibitor-1 (U/ml) 0.1 -3.0 <0.01 Overall 4% of placebo and 8% of orlistat patients withdrew because of adverse events primarily GI events. There were significant decreases in fat-soluble vitamins (A, D, E and K 1) after 4 years in the orlistat group. *venous blood, LOCF = last observation carried forward, BLCF =baseline observation carried forward, ITT=intention to treat. ** Between treatment p values for analyses, for cardiovascular risk factors analyses by LOCF ITT, BLCF ITT and observed data except where indicated. 1 =LOCF and observed, 2 BLCF p<0.05 3 LOCF and BLCF =NS, 4 BCLF =NS.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 79
Evidence Table 4. Orlistat, Rissanen et al., 2001 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) (pooled Mean change Comments and year Study Location(s) Exclusion from baseline at 1 year) Duration Change P (±sem) Rissanen et al., Study design Participants Exposure Weight (kg) -10.0(1.1) <0.001 Weight loss was There were beneficial changes in 2001 Randomised, 55 healthy obese women Orlistat 120mg t.i.d. BMI (kg/m 2) -3.8(0.4) <0.001 not the primary fibrinolysis and coagulation persisted placebo controlled, (51 completers) + hypoenergetic diet Body fat (kg) -6.9(1.0) <0.001 outcome, drug for one year if weight loss was double blind trial. Mean age =44 ±0.7 Comparison Lean body mass (kg) -3.1(0.5) <0.001 and placebo maintained. years Placebo + DBP (mmHg) -7.1(1.3) <0.001 results pooled. Objective Mean BMI= 36±0.5kg/m 2 hypoenergetic diet SBP (mmHg) -7.3 (1.9) <0.001 To study the effect of Smokers =37 Hypocaloric diet S- total cholesterol (mmol/l) -0.1(0.1) ns weight loss on blood Non-smokers=14 Designed to cause a S-LDL-cholesterol (mmol/l) -0.2(0.1) ns coagulability and weight loss of 0., 25- S-HDL-cholesterol (mmol/l) -0.0(0.0) ns fibrinolysis in health Inclusion 0.5kg/wk with 30% S-Triglycerides (mmol/l) -0.1(0.1) ns obese women. See Sjostrom fat, 50% S-insulin (mU/l) -3.4(0.8) <0.001 carbohydrate, and Location 20% protein. Placebo vs. orlistat weight loss at 12 Finland A maximum of months (7.2kg vs. 13.0kg) was not 300mg/day significant and the results of orlistat and Duration cholesterol was placebo were pooled. 12 months distributed in 3 main meals and a low fat snack. Alcohol consumption limited to 150g per week. Further reduction of 300kcal/d after 6 months.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 80
Evidence Table 5. Orlistat, Muls et al., 2001 Study Study Design Participants Exposure/ Outcome Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) (Mean change from baseline at Safety Comments and year Study Location(s) Exclusion 24 weeks± s.d., LOCF) Duration P O p=** Muls et al., 2001 Study Design Participants Exposure Weight loss (kg) -1.88±4.46 -4.66±3.77 <0.001 Patients were stratified by Conclusion Randomised, placebo N=441 recruited Orlistat 120mg t.i.d. Total cholesterol (mmol/l) 0.14±0.85 0.42±0.75 <0.001 study centre and weight loss Orlistat as an adjunct to controlled, double blind, N=294 randomised + diet (-600kcal/day) with <=30% LDL cholesterol (mmo;/l) -0.09±0.80 -0.53±0.65 <0.001 category during run-in dietary intervention promotes multicenter trial. N=255 completers calories from fat HDL cholesterol (mmo;/l) 0.17±0.24 0.07±0.22 <0.001 (<=2.00 vs. >2.0kg). clinically significant weight loss Male 19%, Female 81% Triglycerides (mmol/l) 0.14±9.55 0.08±0.72 0.896 and LDL-C beyond the effect Objective Mean age 48(P) and Comparison LDL-C/HDL-C ratio -0.46±0.70 -0.54±0.67 0.294 A 2.6% difference between of weight loss in overweight or To assess the effect of 50(O) years Placebo. + diet (-600kcal/day) with Lp(a) (mg/l) -6.50±26 3.00±25 0.225 placebo and treatment group obese patients with orlistat 120mg t.i.d. <=30% calories from fat. was judged to be clinically concomitant versus placebo on Inclusion relevant. A sample size hypercholesterolemia. weight loss and serum BMI 27-40kg/m 2 Data analysis calculation determined that a lipids in obese Age 18-70 years Analysis was based primarily on sample size of 100 in each Run-in weight loss and weight hypercholesterolemic Fasting LDL-C 4.1 to the ITT population for whom arm would be required for loss during the first 4 weeks patients. 6.7mmol/l and follow-up was available. Last there to be an 80% were predictors of longer-term TG<4.5mmol/l observation carried forward probability of showing a weight loss at 24 weeks. Study location(s) (<400mg/dl) data was included. 2.6% difference (effect size Belgium of 0.4) at the 5% significance Exclusion Adverse events level. It was estimated that Duration Serious disease, diabetes GI events 30% of patients would not be 24 weeks followed by an or uncontrolled Overall AEs 38 64 fully evaluable and that there optional 24-week open hypertension 67 80 0.016 would be a 20% withdrawal label orlistat extension Women of childbearing Liquid stools (%) rate in the placebo lead-in phase. potential without Increased defecation (%) 8 23 period. The total number adequate contraception Loose stools (%) 5 22 required to be randomised Previous bariatric surgery Decreased defecation (%) 3 16 was estimated to be 284. Use of appetite Bronchitis (%) 12 3 suppressants or lipid – 6 11 Note: lowering agents Limited reporting of weight Evidence of alcohol or loss parameters (BMI, %% substance abuse. 10% not reported). LDL-C = low density lipoprotein cholesterol, P=placebo, O=orlistat, **=difference from placebo. LOCF=last observation carried forward.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 81
Evidence Table 6. Orlistat, Foxcroft et al., 2000 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) Comments and year Study Reference Exclusion Participants Foxcroft et al., Study design Inclusion Exposure Outcomes for weight loss reported in 10kg weight reduction over 2 Only the “most plausible” results On average orlistat results in obese 2000 Rapid systematic review Systematic reviews or RCTs of orlistat Orlistat other systematic reviews. years as an adjunct to diet were used for further analysis. people losing an additional 3-4% of of RCTs. in the treatment of obesity Quality of life (NNT=12 95% CI 7-37) their initial body weight over diet Efficacy or cost effectiveness Estimated QALYs gained in a year EMEA prescription indications for alone over a 2-year period. Objective Human subjects only. of orlistat treatment (based on 100 1.60QALY per year of orlistat orlistat do not coincide with the To clarify the potential Only phase III clinical trials included. persons starting treatment with treatment published trial information. The estimated cost utility of orlistat benefits, disbenefits and responders those who lost and was £46,000 per QALY gained. costs of orlistat in the Exclusion maintained >10% of their initial body treatment of obesity. Phase II clinical trials weight over 2 years). There was no evidence that the short- Not stated but only 3/109 studies or Cost data term weight loss would have a longer- Studies included papers were included. Price of a patient consultation (NHS £118.00 term impact on morbidity and Sjostrom et al., 1998 1998) mortality. Davidson et al., 1999 Monthly cost of orlistat 120 tds. £45.00 Hollander et al., 1998 Average number of consultations per 4 year. GP consultation cost per visit. £16.00 Drop-out year 1 27% Drop-out year 2 23% Average cost per year for 100 £ 73,436 patients treated for 2 years. The cost /QALY of orlistat treatment £45,000 (13,541-131,918) per (extreme value sensitivity analysis) QALY gained The cost/QALY in primary care setting (extreme value sensitivity £26,635 (9,779 –66,505) analysis)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 82
Evidence Table 7. Orlistat, Derosa et al., 2003 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) (% change from baseline) Comments and year Study Duration Exclusion P O F OF Participants N=23 N=25 N=24 N=24 Derosa et al., Study design Participants Exposure/Comparison At 6 months Both groups comparable at baseline Significant decreases in 2003 Randomised, double N=99 Orlistat 120mg t.i.d.+ BMI (kg/m 2) NS NS NS p<0.05 Randomisation envelopes used weight, BMI and waist blind placebo-controlled Obese patients with controlled energy diet WCR (cm) NS NS NS p<0.05 Compliance assessed through tablet circumference was hypercholesterolemia BWL (kg) NS NS NS p<0.05 counting found only in the Objective M=48 Placebo + controlled At 1 year orlistat+fluvastin group To assess obese F=51 energy diet BMI (kg/m 2) NS P<0.02 P<0.05 p<0.01 96 (97%) completers, 3 subjects at 6 months and in all patients with Mean age =51 years WCR (cm) P<0.05 P<0.0 P<0.05 p<0.01 dropped out due to adverse reactions groups apart from BMI hypercholesterolemia (SD=9) Fluvastatin 80mg/d + BWL (kg) P<0.05 P<0.02 P<0.05 p<0.01 (orlistat), in the placebo group at who were prescribed a controlled energy diet At 6 months 1 year. The combined standardised diet, Inclusion SBP (mm Hg) NS NS NS NS Vitamin D levels were not measured. treatment groups comparing the action of BMI> 30kg/m 2 Orlistat 120mg t.i.d. + DBP (mm Hg) NS NS NS NS showed the greatest orlistat, fluvastatin, Age>40 years Fluvastatin 80mg/d + TC NS NS P<0.05 p<0.02 reductions. orlistat with fluvastatin Total cholesterol controlled energy diet LDL-C NS NS P<0.05 p<0.02 and placebo on >=240mg/dl HDL-C NS NS NS p<0.02 No significant changes anthropometric SBP>140mg Hg All participants undertook TGs NS NS NS p<0.05 in BP in any group at 6 measurements, blood DBP<90mm Hg >=30mins aerobic activity At 1 year months, but there was a pressure and lipid 4 days a week SBP (mm Hg) NS P<0.05 P<0.05 p<0.01 significant improvement profile. Exclusion DBP (mm Hg) NS P<0.05 P<0.05 p<0.01 in the treatment groups Smokers Diet: 1500kcal, 54% TC P<0.05 P<0.05 P<0.02 p<0.05 singly and combined at Duration Abnormal thyroid carbohydrate, 24% LDL-C P<0.05 P<0.05 P<0.02 p<0.05 1 year. The greatest One year function protein, 22% fats, 108mg HDL-C NS NS P<0.02 p<0.05 reductions in BP were in Patients taking cholesterol, 35g fiber TGs NS P<0.02 P<0.05 p<0.02 the combined treatment diuretics or beta during placebo lead-in groups. blockers period. Food diaries were Safety kept. Diet instruction every Orlistat and fluvastatin were Lipid profiles were 3 months. well tolerated. 3% of patients significantly better at 6 experienced side effects. months for all There were no serious parameters in the adverse events. combined treatment group. All groups showed significant improvements at 1 year for total cholesterol and low-density lipoprotein cholesterol. Again the combined treatment group had the greatest improvements for each parameter at 6 months and 1 year.
BMI = body mass index, WCR = waist circumference reduction, BWL= body weight loss, P= placebo group, O=orlistat 120mg t.i.d. group, F=fluvastatin 80mg/d group, OF= orlistat 120 t.i.d.+fluvastatin 80mg/d group. TC=total cholesterol, LDL-C=low density lipoprotein cholesterol, HDL-C+high-density lipoprotein cholesterol, TGs = triglycerides . Bold = raised values, normal=decreased values.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 83
Evidence Table 8. Orlistat, Hanefield et al., 2002
Study/Year Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions Location Objective Inclusion Comparison (including adverse events) (Mean change from baseline) Comments Duration Exclusion P O p=† Hanefield et Study Design Participants Exposure Weight loss after 1 yr of treatment Analysis Study conclusions al., 2002 Randomised, N= 492 enrolled Orlistat 120mg t.i.d.+ a Body weight (kg) 3.4 ±5.3 5.3 ±5.1 0.006 Efficacy assessed on an Orlistat + diet resulted placebo controlled, N=383 randomised 600 calorie reduced diet Body weight (%) 3.6 ±5.7 5.4 ±5.0 0.006 ITT basis, ITT analysis: in significant weight Location double blind clinical N=369 ITT population (30% fat, 50% Weight loss >=5% 31.6% 51.3% 0.0001 orlistat=189, loss, improved Germany trial. Male 181, Female= 188 carbohydrate, 20% Diabetic metabolic parameters after 48 weeks of placebo=180 glycaemic control and (Primary Average Age 56-57 protein, max 300mg/d treatment** LOCF model was used cardiovascular risk care and Run in period Average BMI= 34kg/m 2 cholesterol) HbA 1c -0.6 ±1.6 -0.9 ±1.1 0.47 for missing data factor profile in outpatient 4 week placebo + Fasting glucose (mmol/l) -0.9 ±2.8 -1.7 ±2.1 0.08 =5) overweight patients clinics) diet run in period. Inclusion Comparison Post prandial glucose(mmol/l) -0.7 ±4.5 -1.4 ±3.5 0.45 External validity with type 2 diabetes. Age 18-70 years Placebo t.i.d. + a 600 Triglycerides(mmol/l) -0.2 ±3.8 -0.1 ±1.4 0.12 Notes 2 Objective BMI>=28kg/M calorie reduced diet Diabetic metabolic parameters after 1yr* 1.Study conducted Comments To assess the long- HbA 1c =6.5-11% (30% fat, 50% HbA 1c 0.0003 according to GCP Effect was calculated -0.4 ±1.5% -0.9+1.3% term effect of Diagnosis of type 2 diabetes treated carbohydrate, 20% Fasting glucose (mmol/l) 0.004 guidelines from start of study – -0.7 ±3.2 -1.6 ±2.5 orlistat on body with sulphonylureas for at least 2 protein, max 300mg/d Post prandial glucose (mmol/l) 0.003 2.Compliance assessed i.e., including run in ± -1.8 ±3.8 weight, glycaemic months before screening or were cholesterol) Lipid profile at 48 weeks -0.5 4.6 by capsule count period and from
control and diagnosed with diabetes but not yet Total cholesterol 0.07 3.Trained monitor randomisation. ± ± cardiovascular risk treated with antidiabetic medication. LDL-cholesterol (mmol/L) 0.0 1.4 -0.2 0.9 0.12 checked protocol ± ± factors in HDL-cholesterol (mmol/L) +0.01 1.2 -0.09 0.7 0.02 adherence The run-in period was ± overweight patients Exclusion Lipid profile at 1yr +0.07 ±0.2 0.0 0.2 4.Study audit performed included in the with type 2 Diabetic patients treated with Total cholesterol <0.01 throughout the study reported conclusions, diabetes. antidiabetic drugs other than LDL-cholesterol (mmol/L) +1.8 ±22.0% -2.3 ±16.3% <0.05 5.Study sponsored by ideally the results sulphonylureas HDL-cholesterol (mmol/L) +5.1 ±34.3% -2.0 ±26.7% <0.01 Hoffmann-La-Roche during the Treatment Patients treated with medications Other CHD risk actors +6.4 ±24.5% +0.6 ±20.0% randomisation period Duration known to affect body weight, serum Triglycerides(mmol/l) ns should have been 48 weeks lipids or vitamins. SBP mmHg NR NR ns reported Patients with proliferative DBP mmHg -4.96 -4.98 ns retinopathy, papilloedema, Anthropomorphic indices -4.78 -4.80 Results were reported uncontrolled hypertension Waist circumference (cm) <0.01 from the start of (DBP>120mmHg), hypo or -3.0 ±5.6 -5.5 ±5.3 randomisation (48 hyperthyroidism, secondary type 1 weeks) in the main diabetes, cardiac insufficiency, or table but from the start presence or history of cancer or any of study (1 year) in the significant appetite, renal, hepatic, summary and GI, psychiatric, immunological or Reasons for discontinuation conclusion. A number metabolic disorders. Female Pre randomisation drop-out =109 Post randomisation drop-out =119 of significant patients were excluded if they were Violation of selection criteria=59 Adverse events =28 (orlistat =16, differences at 1 year pregnant, lactating or for Other protocol violation=16 placebo=12) were ns from the point childbearing potential and not taking Refused treatment =10 Withdrawal of consent=35 (orlistat =13, of randomisation adequate contraceptive measures. Failure to return=10 placebo=22) Did not co-operate=8 Lost to FU =24 (orlistat=13, placebo-11) Adverse event=3 Protocol violation=14 (orlistat=9, placebo Other=3 † = Change vs placebo, ITT = intention to treat analysis, LOCF= last observation carried forward, GCP=good clinical practice.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 84
Evidence Table 9. Orlistat, Halpern et al., 2003 Study/Year Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions Location Objective Inclusion Comparison (including adverse events) Comments P O P= Duration Exclusion Halpern et al., Study Design Participants Exposure Weight loss 2003 Randomised, double N=365 enrolled Orlistat 120mg ITT pop % WL 3.0± 1.3% 4.7 ±0.5% NR External validity Study blind, placebo N=343 randomised t.i.d. + Orlistat LMS diff P-O (kg) NR 1.6kg 0.0003 Highly selected patients conclusions controlled multi-centre, N=280 completed the study hypocaloric diet + Completers pop %WL -3.1±1.0% -4.9 ±0.7% NR many co-morbidities Orlistat is a safe Location clinical trial Orlistat N= 164 behavioural Completers pop WL (kg) 2.58±1.4kg 4.24±0.23 0.0003 excluded. adjunct to dietary Latin America: Run in period Placebo N=174 counselling >=5% WL 17% 30% 0.003 intervention and Brazil, Columbia, 2-week open, placebo Male=104 >=10% WL 3% 7% 0.072 High drop-out rate conventional Costa Rica period prior to Age=51 years Comparison Diabetic metabolic parameters (DMP), 24 Drop-out causes not antidiabetic Mexico randomisation BMI=35 Placebo t.i.d. + weeks specified pharmacotherapy Caucasian = 60% hypocaloric diet + Mean decrease in glucose levels 0.01±0.30 1.0±0.34 NR in the global Objective behavioural Fasting plasma glucose LSM diff P-O 0.22 ∀0.14 0.85 0.036 treatment of obese To determine if non- Inclusion counselling ITT pop decrease HbA 1c % - 0.61 ∀0.15 0.06 diabetic patients. It insulin dependent Age 18-70 years Completers pop decrease HbA 1c % 0.13 ∀0.14% -0.64 ∀0.16 0.04 promotes a diabetic patients lose BMI>27kg/m 2 Blood pressure at 24 wks values NR clinically more weight when Dx of NIDDM (WHO) SBP values NR values NR ns significant weight treated for 24 weeks HbA 1c 6.0-11.0% DBP values NR ns loss and improved with orlistat in Negative pregnancy test Lipid profile at 24 weeks glycaemic control conjunction with a adequate contraception in women of Total cholesterol (mmol/l) -0.03 ∀0.11 -0.40 ∀0.08 0.0001 and lipid profile. hypocaloric diet + child bearing potential LDL-Cholesterol (mmol/l) +0.03 ∀0.09 -0.31 ∀0.08 0.0002 behavioural HDL-cholesterol (mmol/l) +0.02 ∀0.02 -0.02 ∀0.02 0.038 counselling than Exclusion Fasting TGs (mmol/l) -0.10 ∀0.12 -0.18 ∀0.09 0.217 placebo t.i.d + History of recent MI, uncontrolled Post prandial TGs (mmol/l) 0.379 -0.11 ∀0.13 -0.03 ∀0.13 Comments hypocaloric diet + hypertension, significant cardiac, Anthropomorphic indices Weight loss began
behavioural renal, hepatic, GI, respiratory, High mean waist circumference cm 0.032 to slow at between -3.5 ∀5.4 -5.3 ∀5.3 counselling. neurological, psychiatric or endocrine Adverse events 16-20 weeks
disease, GI bariatric surgery, post Overall adverse events(%) 60.3 74 Duration surgical adhesion, history of Serious adverse events 4 5 24 weeks carcinoma, bulimia, laxative abuse, Patients withdrawn from study because of GI 5(2) 8(7) clinical symptoms of fat-soluble adverse events
vitamin deficiencies, substance abuse, Most common GI events
appetitive suppressants, insulin, Overall 65(37.0%) 110(65%) retinoids, systemic steroids other than Fatty/oily stool 18(10.4%) 77(45.6%) HRT, acarbose (wash-out =2 months), Increased defecation 19(11.0%) 68(40.2%) pregnant lactating and child-bearing Liquid stools 8(4.6%) 17(10.1%) potential women not taking Oily evacuation 0(0.0%) 17(10.1%) contraceptive measures. Flatulance with discharge 3(1.7%) 13(7.7%) Abdominal pain 6(3.5%) 13(7.7%) ITT= intention to treat, LMS=least mean squares, WL=weight loss.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 85
Evidence Table 10. Orlistat, Leung et al., 2003 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) Comments and year Study Reference Exclusion Leung et al., Study type Inclusion Exposure Weight reduction with orlistat associated with a significant Withdrawal rate not Orlistat and 2003 Systematic review Studies between 1966- Orlistat 30-120mg improvement in control of cardiovascular risk factors. P<0.05 reported sibutramine February 2002 t.i.d. demonstrated a Objective RCTs of > 6 months Comparison Patient with hypercholesterolemia favourable efficacy To describe the duration reporting the Placebo Total cholesterol orlistat vs. placebo 11.9% vs. 4.0%, p<0.001 and safety profile in pharmacological efficacy of LDL cholesterol, orlistat vs. placebo 17.6% vs. 7.6%, p<0.001 RCTs. management of obesity orlistat Plasma insulin at the end of 2 years, orlistat vs. placebo (one study) 66.5 vs. 86.3pmol/L, p<0.04127mm concentrating on orlistat Systolic blood pressure after one year, orlistat vs. placebo Hg vs. 129mmHg -=0.001980mm Current evidence Diastolic blood pressure after one year, orlistat vs. placebo. Hg vs. 82mmHg, p=0.002 supports their use as Included studies adjuncts to lifestyle Orlistat Patients with type 2 diabetes modifications in the Davidson, 1999 Weight loss Orlistat + hypocaloric diet vs. placebo 6.2 vs. 4.3kg p<0.001 treatment of obesity Finer, 2000 Improved glycaemic control orlistat vs. placebo p<0.001 Hauptman, 2000 Patients with IGT Hill, 1999 Normalisation of glucose levels, orlistat vs. placebo. 71.6% vs. 49.1%p=0,04 Hollander, 1998 Reduction on rate of progression to diabetes, orlistat vs. placebo 3.0% vs. 7.6%p=0.04 Rossner, 2000 Adverse events Sjostrom, 1998 Mild – moderate GI effects, orlistat vs. placebo 8-27% (yr 1) Low plasma levels of fat-soluble vitamins and beta carotene in the 6-22%(yr2) orlistat group Interactions reported with cyclosporine and warfarin p<0.05
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 86
Evidence Table 11. Orlistat, Heymsfield et al., 2000 Study Study Design Participants Exposure/ Outcomes Effect Size & Validity/Applicability % of studies Conclusions authors Objective Inclusion Comparison (including adverse events) Precision Comments and year Study References Exclusion Concomitant treatment Participants Heymsfield et al., Study design Inclusion Exposure Weight loss change from initial weight Data from the reported 2000 Meta-analysis of three Meta-analysis Orlistat, 120mg t.i.d. Orlistat 6.27 ∀0.41kg ITT analysis 100% studies suggest that randomised, double-blind- Randomised, double blind, Comparison Placebo 3.79 ∀4.1kg improvements in glucose placebo-controlled, placebo controlled trial Placebo. t.i.d. p<0.001 Similar at baseline tolerance depend upon multicentre clinical trials. At least 1 follow-up assessment of Concomitant Glucose tolerance relatively small changes in glucose tolerance mildly low energy diet for 1 (% of patients progressing to diabetic body weight and that Objective Follow-up of at least 2 full years year status ) modest weight loss may To test the hypothesis Individual trials Orlistat 3.0% reduce the risk of that orlistat combined with Age >18 years Placebo 7.6% developing diabetes in dietary intervention BMI 30-34 p-value 0.04 obese subjects. improves glucose Absence of weight loss >4kg in tolerance status and last 3 months Change from IGT to NGT Modest pharmacologically prevents worsening (IGT at baseline improved to NGT at facilitated weight loss diabetes status more Exclusion end of treatment) produces important effectively than placebo. Meta analysis orlistat metabolic benefits. Individual trials placebo 71.6% Included publications Subjects who had stopped p-value 49.1% Sjostrom ,1998 smoking in the last 6 months 0.04 Hauptman , 2000 Subjects with significant cardiac, Improvement in insulin area under renal, hepatic, GI, psychiatric or curve for subjects with normal glucose endocrine disorders, tolerance at baseline Subjects with drug-treated type 2 Orlistat vs. placebo diabetes, history or presence of p=0.03 substance abuse, excessive intake of alcohol or concomitantly used Fasting glucose levels normal medications that alter appetite or orlistat vs. placebo lipid levels. p=0.02 Participants Fasting glucose levels impaired N=675 orlistat vs. placebo Obese (BMI 30-43) p=0.01 39 centers in Europe and USA between 1992-1995.
EGO =Exposure Group Occurrence, CGO = Control Group Occurrence, RR = Relative Risk, RD = Risk Difference, NNT = number needed to treat, N/S = Not Stated, ITT= intention to treat
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 87
Evidence Table 12. Orlistat, O’Meara et al., 2001 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability % of Comments authors Objective Inclusion Comparison (including adverse events) Weighted mean difference † studies Conclusions and year Study Reference Exclusion
O’Meara et al., Study type Inclusion Orlistat Primary outcomes Randomised method 9% In general the 2001 RCTs Weight change described (n=11) methodological quality of Objective 12 weeks, 2 trials (50-60mg t.i.d) -1.24kg (-2.65 to 0.16), p=0.08 Concealed randomisation 100% the trials was moderate To systematically assess Any duration of 24 weeks, 2 trials -10.75kg orlistat, -7.34 placebo Similar at baseline, 100% to good. the evidence for the therapy* 1- and 2- year endpoints (120mg t.i.d) All -2.9kg (-3.61 to – 2.19) p=<0.00001 Patients blinded 100% clinical effectiveness and trials (n=4) Carers blinded Unclear The optimum regimen cost-effectiveness of Any length of From start of run-in period (n=2) -2.4kg(-3.62 to –1.47), p<0.00001 Blind outcome assessment 9%, 91% was 120mg t.i.d. orlistat in the follow-up* From double blind phase (n=2) -3.35kg(-4.44 to –2.27), p<0.00001 Intention to treat analysis 82% management of obesity At 1 year (n=2, 120mg t.i.d.) -2.38kg (-3.45 to –1.31) ,p<0.00001 Patient adherence assessed 91% Most trials were single Obese or At 2 years (n=2, 120mg t.i.d.) -3.19kg(-4.25 to – 2.12), p=0.00001 Withdrawals reported blind, placebo with a run- Studies included overweight Change in % of body weight at 2 years (n=2, -3.23kg (-4.77 to –1.69), p=0.00001 Selection criteria 100% in period prior to double (N=11) patients or patients 120mg t.i.d.) 100% blind trial. Davidson et al., 1999 who have % weight loss relative to initial weight at 24 Drent & van der Veen, previously been weeks There is a lack of 1993 obese or 30mg orlistat t.i.d. 8.5% information on the Drent et al., 1995 overweight wishing 60mg orlistat t.i.d. 8.8% effectiveness and safety Finer et al., 2000 to maintain weight 120mg orlistat t.i.d. 9.8% of orlistat in older people. Hauptman et al., 2000 loss. 240mg orlistat t.i.d. 9.3% Hill et al., 1999 Placebo 6.6% For orlistat trials that Hollander et al., 1998 Only studies % of patients losing >10% of body weight at matched the prescribing Micic et al., 1999 published in 24 weeks (n=2), guidelines most reported Rossner et al., 2000 English, French, 30mg orlistat t.i.d. 28% statistically significant Sjostrom et al., 1998 Dutch or German 60mg orlistat t.i.d. 28% results in favour of Van Gaal et al., 1998 were considered 120mg orlistat t.i.d. 37% orlistat for weight loss in for inclusion. 240mg orlistat t.i.d. 38% participants with and Placebo 19% without diabetes. Exclusion Reduction in waist circumference at 24 Trials recruiting weeks (n=2) people with eating 30mg orlistat t.i.d. 3.5cm disorders such as 60mg orlistat t.i.d. 5.1cm anorexia nervosa, 120mg orlistat t.i.d. 5.9cm and bulimia 240mg orlistat t.i.d. 6.3cm nervosa. Placebo 6.0cm RR of failure to achieve at least 5% loss of RR=0.72 (0.63-0.82), p<0.00001 initial weight at 1 year (orlistat, 120mg t.i.d, n=4) RR of failure to achieve at least 10% loss of RR=0.85 (0.80-0.91), p<0.00001 initial weight at 1 year (orlistat, 120mg t.i.d, n=5) RR of failure to achieve at least 10% loss of RR=0.86(0.79-0.93), p=0.0001 initial weight at 2 year (orlistat, 120mg t.i.d, n=3)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 88
Evidence Table 12. Orlistat, O’Meara et al., 2001 (continued) Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability % of Conclusions authors Objective Inclusion Comparison (including adverse events) Weighted mean NNT studies Comments and year Study Reference Exclusion difference † Participants O’Meara et al., Adverse events More frequent in 2001 (Cont) Gastrointestinal adverse orlistat group events: fatty stools, increased defecation, oily spotting. Lower serum levels of fat- soluble vitamins.
Economic evaluation Incremental cost-utility of orlistat treatment per QALY GBP 45,881 gained for WL of 3-4% during (19,452-55,391) first year of treatment.
** manufacturer’s trials had to have duration of at least 1 year.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 89
Evidence Table 13. Orlistat, Padwal et al., 2004 Study Study Design/Type Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability % of Conclusions authors Objective Inclusion Comparison (including adverse events) studies Comments and year Study References Exclusion
Padwal et al., Cochrane Systematic Inclusion Drug; Orlistat Primary outcomes Randomised/allocation 18% Baseline weight was 2004 Review BMI ≥ 30kg/m 2 Dose; 120mg t.i.d average weight loss in kg 2.7(2.3-3.1) concealment method calculated from the start of the study length ≥ 1 yr Co-interventions average weight loss as % change from baseline 2.9(2.3-3.4) described run in period rather than the Objective randomised Low fat hypocaloric diet (8 % with weight loss >= 5% 21(19-24) point of randomisation – To assess/compare the double-blind studies) % with weight loss >=10% 12(8-16) Similar at baseline 55% inflating the absolute change effects and safety of placebo or active Dietary counselling total population 8(5-10) in weight. single or combination control Exercise counselling high risk group 17(14-21) ITT last observation anti-obesity drug therapy Food intake diaries low risk group 0.7-3.4cm carried forward 0%* High attrition rates were the in clinical trials of at Six trials required a Educational literature or Comparative BMI reduction major methodological least one year compliance rate of videos Comparative waist circumference reduction for Double blind limitation. >75% in the run-in Orlistat (5 trials, not pooled, p<0.05 in 4 of 5 100% 11 trials phase studies) Blind outcome Is the degree of weight loss 6,021 participants Av BMI = 33.4kg/m 2 Secondary outcomes assessment achieved of clinical benefit? Exclusion (10 trials) NR It may be, particularly for high- Obesity of endocrine Av weight = 96kg Total cholesterol reduction greater in orlistat 0.33 mmol/L (0.28-0.38) Average drop-out rate risk groups but more definitive Included studies origin, group 33% data are needed. Weight Hollander et al., 1998 diabetes mellitus, Av age = 47 years Low-density lipoprotein reduction greater in 0.27mmol/L (0.22-0.31) (14-52) reduction of 5-10% of body Sjostrom et al., 1998 treatment with orlistat group (8 trials) weight is associated with Davidson et al., 1999 medication that may 80% female HDL reduction higher in orlistat group (7 trials) 0.02mmol/L (0.01-0.04) improvements in blood Finer et al., 2000 alter body weight, Triglyceride levels reduction greater in orlistat pressure, lipid and glucose Hauptman et al., 2000 uncontrolled 75 % Caucasian group 0.05mmol/L (0.07gain–0.17 parameters but data on the Lindgarde et al., 2000 hypertension loss) effect of mortality and Rossner et al., 2000 High risk populations SBP net decrease (9 trials) cardiovascular events are Bakris et al., 2002 Six trials limited enrollment SBP net increase (2 trials) 1.8mmHg (0.9-2.6) lacking. Broom et al., 2002 to higher risk patients, 3 DBP net decrease (8 trials) 0.4mm Hg (NS) All studies showed a positive Kelley et al., 2002 with type 2 diabetes on 1.6mm Hg (0.7-2.4) treatment effect. However Miles et al., 2002 stable doses of oral Fasting plasma glucose reductions (9 trials, publication bias cannot be hypoglycemic agents or significantly different to placebo in 5 trials) ruled out. insulin, 3 with at least one HbA 1c reduction (4 trials, high risk patients) 0.1-1.3 mmol/L The patient population studies additional CVD risk factor 0.2% (0.2-0.3) represent a highly selected (hypertension, Adverse events group of relatively healthy dyslipidemia, type 2 Increased in orlistat group: obese patients who were able diabetes IGT). fatty/oily stool and oily spotting to comply with study diet and faecal urgency (9 trials) 16-40% medication requirement. Most Faecal incontinence increase of the patients were female 6%(5-8%) Caucasians. Extrapolation to patients with different demographic profile must be made with caution.
Note: Levels of fat-soluble vitamins A, D, E and beta-carotene were reported as lowered in the orlistat group.
* all trials hampered by a high attrition rate of 14-52%, average 33% EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 90
Evidence Table 14. Sibutramine, O’Meara et al., 2002 Study Study Design/Type Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability % of Conclusions authors Objective Inclusion Comparison (including adverse events) Weighted mean difference, RR studies Comments and year Study References Exclusion Participants O’Meara et Study type Inclusion Sibutramine 5- Primary outcomes Randomised method described 19% Many of the trials al., 2002 NICE HDL RCTs 30mg/d Weight change (n=16) demonstrated a 8 weeks (3 trials, 10-20mg/d) -3.4kg (4.22 to –2.58) p<0.00001 Concealed randomisation 100% statistically significant Objective Any duration of therapy 24 weeks (4 trials, 10mg/d) -4.0 to 9.1kg Similar at baseline, manufacturers difference in favour of To systematically assess 1 year -4.1 to –4.8kg trials (n=5), other trials (N=11) 0% sibutramine in terms of the evidence for the Any length of follow-up 100% (a) absolute weight clinical effectiveness and RR of failure to achieve a weight RR=0.48 (0.39-0.60), p<0.00001 Patients blinded loss (b) % of patients cost-effectiveness of Obese or overweight patients or loss of at least 5% of initial body Carers blinded 100% achieving at least 5% sibutramine in the patients who have previously been weight at 24 weeks – 2 trials Blind outcome assessment unclear or 10% loss of initial management of obesity obese or overweight wishing to maintain (10mg/d). Note for 15mg/d dose unclear body weight. weight loss. significant heterogeneity was Intention to treat analysis Studies included observed. Patient adherence assessed 69% The differences were (N=11) Only studies published in English, Withdrawals reported 50% small and may not be Apfelbaum et al., 1999 French, Dutch or German were Fat content Selection criteria 94% clinically important. Bray et al., 1999 considered for inclusion. Fat-free mass loss at eight weeks -1.83kg(-2.48 to –1.19)p<0.00001 100% Cuellar et al., 2000 (2 trials, sibutramine) Weight loss seems to Fanghanel et al., 2000 Exclusion Fat change at eight weeks (2 -1.77kg (-2.58-0.96) p<0.00002 be dose related with Finer et al., 2000 Trials recruiting people with eating trials, sibutramine) statistically significant Fujioka et al., 2000 disorders such as anorexia nervosa, results reported for 10- Hanotin et al., 1998 and bulimia nervosa. Adverse events 20mg doses of Hansen et al., 1999 Pulse rate, blood pressure and Small but significant increase with sibutramine. Seagle et al., 1998 heart rate sibutramine, not considered to be Walsh et al., 1999 clinically significant. Weintraub et al., 1991 Between group Headaches, sleep difficulty, Reported in one small trial. difference in BP, lipids Studies from the irritability levels and glycaemic manufacturer (N=5)* control were Wirth et al., 1998 Manufacturer’s cost utility inconsistent. Smith et al., 1994 analysis (BMI<30kg/m 2, free of James et al., 1999 complications and co-morbidities, Rissan et al., 1998 no weight loss sustained at 56 Williams ey al., 1999 years, costs discounted at 6%, benefits at 1.5%))
QoL coefficient for sibutramine 0.00185/kg lost * manufacturer’s trials had to have a duration of Cost per QALY gained through £42,000 at least 1 year CHD reduction Cost per QALY gained through £77,000 diabetes incidence
Estimated cost per QALY £10,500 (£5,7000-£35,000) Best-worse case scenario EGO =Exposure Group Occurrence, CGO = Control Group Occurrence, RR = Relative Risk, RD = Risk Difference, NNT = number needed to treat, N/S = Not Stated NSD no significant difference, NS = not significant ‡individually significant data not pooled.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 91
Evidence Table 15. Sibutramine,Tambascia et al., 2003 Study/Year Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicabiilty Conclusions Location Objective Inclusion Comparison (including adverse events) Comments P S Duration Exclusion Tambascia et Study Design Participants Exposure Weight loss at 24 weeks Analysis Study conclusions al., 2003 Randomised, N=48 recruited Sibutramine 10mg/d + Weight loss (kg) +0.9 -5.6 Compliance tests – pill The authors concluded that Placebo-controlled, double N=40 randomised dietary advice from obese Weight loss (%) 1.2 -6.1 counting sibutramine had demonstrated Location blind clinical trial N=31 completers clinic Fat mass (%) +0.4 -1.6 Completers analysis- efficacy in reducing weight in Brazil Healthy obese, female Comparison Diabetic metabolic parameters at 24 analysis per protocol non-diabetic females and Run in period volunteers Placebo + diet advice from weeks Only difference from ameliorated insulin sensitivity None reported Age =41.1 ±9.9 years obese clinic Glucose (mg/dl) 0.0 -2.0 baseline computed for parameters and other metabolic Age-range =19-58 years Trigylcerides -8.0 -40.0 each treatment parameters. Objective BMI=34.3 ±2.9kg/m 2 Insulin ( :U/ml) +4.6 -9.3 separately. To assess the effect of HOMA-IR +0.8 -2.2 Proportion of participants sibutramine assisted weight Inclusion HOMA-≥ +51 -134 >=5% and 10% weight Comments reduction on insulin sensitivity BMI=30-40kg/m 2 Kitt +0.19 +1.06 loss not reported Weak analysis and metabolic parameters in Less than 20 per group obese, normal glucose Exclusion Blood pressure at 24 wks Reasons for No formal dietary intervention or tolerant individuals over a Psychiatric instability, history of SBP mmHg +4.0 -5.0 discontinuation behaviour modification program period of 24 weeks. anti-obesity surgery, recent DBP mmHg +1.0 -5.0 Adverse event =1 Sibutramine not administered as change in exercise pattern, Heart rate (bpm) +2.0 +5.0 Pregnancy=1 per licence. Duration smoking cessation within 6 Lipid profile at 24 weeks Protocol 24 weeks months, hypertension, history of Total cholesterol -8.0 -8.0 deviation=3(1S+2P) CHD, cardiac arrhythmia, HDL-cholesterol -1.0 +2.0 Lack of efficacy=4 endocrine disorders including LDL-cholesterol -5.0 -2.0 (1S+3P) type 2 diabetes and glucose intolerance, history of drug or Anthropomorphic indices External validity alcohol abuse, use of anti- Waist circumference cm +1.0 -4.0 100% female depressant agents, monoamine Adverse events very large number of oxidase inhibitors and use of Sibutramine = anorexia, dry mouth and exclusion criteria drugs that affect insulin nausea. sensitivity (corticoids, diuretics, Serious adverse events insulin sensitisers, oestrogen Cardiac arrhythmia =1 (sibutramine) replacement. Females were required to have a negative serum pregnancy test, to be post menopausal, surgically sterilised or using a medically accepted contraceptive method.
HOMA-IR, HOMA-≥ and K itt are insulin sensitivity tests, BMP = beats per minutes, P=placebo, S=sibutramine. Bold figures = significant difference.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 92
Evidence Table 16. Sibutramine, Padwal et al., 2004 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability % of Conclusions authors Objective Inclusion Comparison (including adverse events) Weighted mean studies Comments and year Study Reference Exclusion difference † Participants Padwal et al., Cochrane Inclusion Sibutramine 15-20mg daily Primary outcomes Randomised method described 33 Heterogeneity in 2004 Systematic Review BMI ≥≥≥ 30kg/m 2 average weight loss in kg 4.3(3.6-4.9) Similar at baseline 67 primary outcomes study length ≥1 yr Co-interventions = average weight loss % change from baseline 4.6(3.8-5.4) Intention to treat last observation 100 considered to be of To assess/compare randomised 3 trials (2 USA, 1UK) 929 % of patients with weight loss of least 5% 34(28-40) carried forward negligible clinical the effects and safety double blind participants % of patients with weight loss of at least 10% 15(4-27) Double blind 100 relevance. Stratifying of single or placebo or active control Comparative BMI reduction 1.5kg/m2 (1.2-1.8) Blind outcome assessment NR into high and low risk combination anti- Av BMI 33.4kg/m 2 Waist circumference reduction 4-5cm Compliance NR study populations obesity drug therapy Randomisation restricted to Drop-outs<20% Yes (43%) eliminated observed in clinical trials of at patients who could follow Av weight 96kg Secondary outcomes heterogeneity. least one year dietary advice or had 75% Improved blood glucose levels (2 studies) 0.08 and 0.9mmol/L (NS) compliance during the run-in Av age 47 years sibutramine treated patients 0.18-0.23 mmol/L McMahon et al 2000 phase. Triglycerides levels lower in sibutramine arm (p<0.05) Smith et al 2001 80% female Total cholesterol NSD McMahon et al 2002 Exclusion Low-density lipoprotein NSD Obesity of endocrine origin, 75 % Caucasian HDL higher in sibutramine studies (n=2) 0.08-0.09mmol/L diabetes mellitus, treatment (p=<0.05) with medication that may 2 trials limited enrollment to Adverse Events alter body weight, obese hypertensive patients (Sibutramine) uncontrolled hypertension with well controlled BP, in the Systolic blood pressure net increase remaining trial only *% had a 1.9mmHg (0.2-3.6)‡.Diastolic blood ‡ Note: one trial excluded treated cardiovascular condition pressure 1-4mm Hg because 36% of randomised Pulse rate increase 4-6 beats per min patients were not included in (p=<0.05) the analysis. 7-20% Duplicate populations Other AEs more common in sibutramine arm Insomnia Nausea Dry mouth Constipation
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 93
Evidence Table 17. Sibutramine, Leung et al., 2003
Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability % of studies Conclusions authors Objective Inclusion Comparison (including adverse events) Comments and year Study Reference Exclusion
Leung et al., Study type Inclusion Exposure Dose related weight loss at 6 months Overall withdrawal Conclusion 2003 Systematic review Studies between Sibutramine 1- Sibutramine rates Orlistat and sibutramine 1966-February 30mg t.i.d. 1mg 2.5% Sibutramine 9% demonstrated a favourable efficacy Objective 2002 Comparison 5mg 3.9% Placebo 7% and safety profile in RCTs. Current To describe the RCTs of > 6 Placebo 10mg 6.1% p<0.05 Discontinuation due to evidence supports their use as pharmacologic months duration 30mg 9.4% hypertension adjuncts to lifestyle modifications in management of reporting the Placebo 1.2% Sibutramine 0.4% the treatment of obesity obesity efficacy of Placebo 0.4% concentrating on Sibutramine Meta analysis of 4 long-term trials Discontinuation due to Note: In March 2002 the Italian sibutramine Waist circumference reduced (sibutramine) p<0.001 tachycardia Health Ministry suspended sales of Waist hip ratio reduced (sibutramine) p<0.02 Sibutramine 0.4% sibutramine for re-evaluation of its Weight loss greater in treatment group than Placebo 0.1% safety profile after 2 deaths and 50 Included studies placebo group for adverse events. After investigation Sibutramine Type 2 diabetes mellitus p<0.001 the EAEMP (June 2002) reported that Apfelbaum, 1999 Dyslipidemia p<0.05 No quality assessment the risk-benefit profile of sibutramine Cueller, 2000 Hypertension p<0.05 was reported remained positive. Dujovne, 2001 Greater reductions with sibutramine vs. placebo Fanghanel, 2000 Fasting HbA 1c and glucose levels, serum triglycerides p<0.05 Fanghanel , 2001 high density lipoprotein, cholesterol, urate, C-peptide Fujioka, 2000 Insulin p<0.001 James, 2000 Adverse events McMahon, 2000 Dry mouth, anorexia, insomnia, constipation, headache p<0.001 Smith, 2001 Blood pressure, sibutramine vs. placebo 1-3mmHg (mean increase) Wirth, 2001 Heart rate, sibutramine vs. placebo 4-5 beats/min (mean increase)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 94
Evidence Table 18. Sibutramine, Nisoli and Carruba, 2003
Study Study Design Participants Exposure/ Outcomes Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) Comments and year Included Studies Exclusion Adverse Event S P Participants Nisoli and Study design Inclusion Exposure Common adverse events All trials reported participant The authors Carruba, 2003 Systematic Review of all All clinical trials of sibutramine for Sibutramine associated with sibutramine selection criteria and comparability concluded that the trials of sibutramine weight management in overweight or Co-intervention Headache of groups at baseline. assessment-risk reporting unwanted side obese subjects reporting unwanted Low calorie diet Constipation profile of effects of sibutramine side effects of sibutramine Nausea sibutramine was Adjunctive therapy in the Nervous system (>5%) Relatively few trials described the positive although Objective form of diet, exercised Dizziness use of a priori power calculation to the product must A benefit-risk assessment and behavior Dry mouth estimate required sample size. be kept under of sibutramine in the modification advice in Insomnia review and BP and management of obesity varying intensities. Requiring regular None of the trials included heart rate monitoring methods to determine the success monitored Included studies Increased blood pressure of blinding of patients, care givers especially in obese Hansen et al., 1998 Increased heart rate or outcomes assessors. hypertensive Walsh et al., 1999 patients. James et al., 2000 STORM trial Reporting of numbers of Apfelbaum et al., 1999 Data relating to blood lipids withdrawals per group with Knoll 2000 and general metabolism reasons was variable. Knoll2000 showed substantial Knoll 2000 improvements in the A few trials included an McNulty et al., 2003 sibutramine group. assessment of patient’s adherence Bray et al., 1999 with the trial regimen; usually Fanghanel et al., 2000 Adverse events either counting returned capsules Cueller et al., 2000 Insomnia 8% vs. 3% or self-reporting (both are Fujioka et al., 2000 Nausea 7% vs. 1% potentially unreliable). Gokcel et al., 2001 Inc blood Serrano-Rios et al., 2002 pressure 8% vs. 3% Hanotin et al., 1998 Lassitude 7% vs. 11% Finer et al., 2000 Back pain 7% vs. 9% Hanotin et al., 1998 Weintaub et al., 1991 Withdrawals 14%vs5% Seagle et al., 1998 results form adverse events Sibutramine 0.8% withdrawals as a result of BP
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 95
Evidence Table 19. Sibutramine, Berkowtitz et al., 2003
Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability authors Objective Inclusion Comparison (including adverse events) Placebo Sibutramine p Conclusions and year Length of Exclusion (s.d.) (s.d.) value study Comments
Berkowtitz et Study design Participants Phase I, placebo Phase I Phase I During Phase I al., 2003 Randomised, N=146 adolescents controlled, 6months Mean weight loss 3.2kg (6.1) 7.8kg (6.3) 0.001 Completed 6 month 19/43(44%) participants double blind, evaluated at clinic, 82 weight loss trial % change in BMI 4.0%(5.4) 8.5% (6.8) 0.001 assessment had dose reductions or placebo randomised, 64 excluded Sibutramine + behavior Waist circumference -2.8cm(5.6) -8.2(6.9) <0.001 Placebo 34/39 discontinued treatment controlled trial (psychiatric conditions therapy, N=43 or placebo BMI reduction>=5% 14(36) 27(63) 0.02 Sibutramine (87%) with sibutramine. BP and (24), not interested (21), +behavior therapy, n=39 BMI reduction>=10% 6(15) 17(40) 0.02 40/43 HR results reflect a unable to attend group Medication protocol BMI reduction>=15% 1(3) 8(19) 0.02 Taken off study (93%) deliberate attempt to limit meetings (12), medical Week one : all participants Systolic BP (3 mo) 110.7mmHg 114.8mmHg 0.02† because of marked and increases in BP or HR. Length of conditions (2), other (7)). received placebo. Week two : Diastolic BP (3mo) 55.7mmHg 58.4mmHg 0.23† sustained BP increase 5/43 (12%) study Mean age =14.1 yrs either 5mg/d sibutramine or Systolic BP (6mo) 110.3mmHg 112.9mmHg 0.06† with sibutramine Comment Intention to treat (SD=1.2) placebo. Week three : Diastolic (6mo) 55.3mmHg 58.6mmHg 0.11† Sibutramine and analysis at 6 Mean weight=103.6kg 10mg/d or placebo. Week Pulse rate (3mo) 81.6 beats/min 85.6beats/min <0.001† ITT analysis behaviour therapy months (SD=15.4) seven : 15mg/d or placebo Pulse rate (6mo) 81.2beats/min 84.8beats/min 0.007† resulted in significantly Mean BMI=37.8(SD=3.8) Dose reduction: in subjects Hunger reduction greater 0.002 Study powered to greater weight loss than Mean waist with an increase in SBP or 19 dropouts detect 4% difference in placebo and behaviour circumference=110.8cm DBP of >=10mmHg or an Systolic BP‡ + 10mmHg <0.001 baseline BMI between therapy. (SD=10.0) increase of pulse rate of Diastolic BP‡ +8.6mmHg <0.001 treatment groups. Sex= 55F: 27M >=15%, dose reduced in Pulse rate +14.3/min <0.001 Behavioural and White race=45 5mg until desired rates No significant pharmacological Black race=34 achieved. Sibutramine Serum chemistry No between group difference between treatments appeared to discontinued if BP increased differences groups at baseline have additive effects that Inclusion 20mm Hg or more. maximised weight loss. adolescents Behavioural protocol No significant 13-17 years Weekly group sessions for difference in Sibutramine must be BMI =32-44. subject and separate groups withdrawals between carefully monitored to sessions for parents. Groups groups control increases in BP Exclusion/ were led by dietitians, and pulse rate. contraindications psychologists or Phase II, maintenance Cardiovascular disease psychiatrists. trials 7-12 months, all Diabetes (type 1 or 2) Diet subjects receive Major psychiatric disorders 1200-1500kcal/d 30% fat sibutramine Use of a weight loss 15% protein, 55% Not reported here, medication or weight loss carbohydrate. endpoints not relevant. of 5kg or more in the Exercise previous 6 months Eventual goal of walking or Use of medications similar aerobic activity promoting weight gain 120mins per week. (e.g., oral steroids) Other Use of medicines Daily eating and activity logs contraindicated with use of sibutramine Cigarette smoking † represents the difference between treatment conditions between baseline and month 3, baseline and month 6. ‡ Increase from baseline to time of dropout
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 96
Evidence Table 20. Sibutramine, Hazenberg et al., 2000
Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) (Mean change from baseline) % of studies Comments and year Exclusion P S p=
Hazenberg Study design Exposure Results at 12 weeks Completed study Conclusion et al., 2000 Randomised, double Sibutramine 10mg, N=62 Body weight (kg) -2.2 -4.4 0.002 Sibutramine n=50 (81%) blind, parallel-group, Participants Comparison Body weight % reduction +2.3 +4.7 <0.001 Placebo n=56(86%) Sibutramine 10mg/d placebo controlled N=127 Placebo N=65 BMI reduction (kg/m 2) -0.8 -1.6 <0.01 is effective in Age – median 48 years (19-65) More than 5% baseline body Withdrawals reducing body weight Objective M=38 Co-interventions weight loss (%) 17 44 <0.01 Four patients withdrew in the run-in phase. in obese patients. To assess weight loss, F=39 Dietary advice and diet Waist Hip Ratio change +0.01 -0.01 <0.06 14 patients erroneously given active safety and tolerability Hypertensive n=113 sheet Supine DBP (mm Hg) -5.2 -3.7 0.14 medication during the run-in phase were Weight reduction has associated with Supine SBP (mm Hg) -4.2 -4.1 0.97 withdrawn. a beneficial effect on sibutramine in mild to Heart rate (beats/min) -3.1 +1.8 <0.001 BP in obese moderately obese 4 patients withdrew (2 sibutramine, 2 placebo) hypertensive hypertensive patients. Inclusion because of adverse events patients. 18-65 years 7 patients withdrawn in double blind phase BMI in the range 27-40kg/m 2 Adverse events N=4 sibutramine (AEs =2 protocol violation =2) Hypertension stabilised for at least Sibutramine = 42% n=3 placebo (AEs=2, protocol violation =1). 4 weeks and medication not Placebo =43% changed in the period. Last observation carried forward analysis. Dry mouth 2 8 Exclusion Constipation 4 6 More than 3kg loss in previous 3 Nervousness 5 1 months Previous sibutramine More than borderline depression or taking any antidepressant. Obesity of endocrine origin Diabetes requiring insulin Any past or present significant illness Risk of pregnancy
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 97
Evidence Table 21. Sibutramine, Gokcel et al., 2001 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) (Mean change from baseline) % of studies Comments and year Exclusion P S P=
Gokcel et Study design Participants Exposure n=29 Six months Withdrawal - one subject withdrew The significant improvement al., 2001 Randomised, double blind, N=60, obese females Sibutramine 10mg Weight loss (kg) 0.91 ±0.53 -9.61 ±1.37 <0.0001 from the treatment group because in glycemic control in the placebo controlled. t.i.d. BMI change (kg/m 2) 0.36 ±0.21 -3.92 ±0.54 <0.0001 of hypertension. sibutramine group was Inclusion Diet restriction Waist circumference (cm) 0.92 ±0.49 -8.04 ±4.43 <0.0001 Exclusion - five subjects from attributed to weight loss. Objective HbA 1c >8% Hypoglycemic Fasting blood glucose (mg/dl) 15.76 ±3.89 124.88 ±8.58 placebo group suffered low 2 Evaluation of the efficacy of BMI >30kg/m drugs Postprandial blood glucose (mg/dl) 32.88 ±6.13 102.24 ±51.99 <0.0001 treatment efficacy and were There was also a significant : sibutramine in combination On maximum dosage of Comparison n=25 Insulin level ( U/ml) 0.68 ±0.43 5.66 ±0.97 <0.0001 excluded and switched to insulin. positive change in the lipid with hypoglycemic drugs in metformin (2,550mg/d) <0.0001 profile of the sibutramine Placebo HOMA IR 0.31 ±0.28 7.09 ±0.81 obese females with poorly and sulfonylureas Diet restriction HbA 1c (%) <0.0001 Comparable patient population at patients. 0.53 ±0.01 2.73 ±0.01 regulated type 2 diabetes. (gliclazide, 320mg/d or Hypoglycaemic Total cholesterol (mg/dl) <0.0001 baseline 7.68 ±5.04 28.08 ±4.93 glipizide, 20mg/d). drugs HDL cholesterol (mg/dl) <0.008 Conclusion LDL cholesterol (mg/dl) 0.01 ±1.10 0.97 ±1.58 >0.05 The addition of sibutramine to Exclusion VLDL cholesterol (mg/dl) 13.32 ±3.94 20.92 ±4.66 >0.05 oral hypoglycaemic therapy Type 1 diabetes Triglycerides (mg/dl) 0.01 ±2.58 8.68 ±2.58 <0.02 leads to significant weight loss Obesity of endocrine 0.36 ±7.26 46.76 ±13.09 <0.08 and improved metabolic origin parameters. Previous weight loss SBP – not reported drugs or intensive weight DBP & HR significant reduction s Sibutramine is well tolerated loss program reported but no data given and safe in this patient Uncontrolled population. hypertension, Side effects (sibutramine) SBP>160mm Hg or DBP> Dry mouth, n=11 100mm Hg. Constipation, n=16 Glaucoma Hypertension, n=1 (patients withdrew Pregnant women from study) Patients taking beta- blockers, antidepressants, monamine oxidase inhibitors or any drug affecting appetite or weight.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 98
Evidence Table 22. Sibutramine, Kim et al., 2003
Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) Standardised difference of Comments and year Studies Included Exclusion changes from baseline between treatment and control§ (95% CI) Kim et al., Study design Inclusion Exposure Weight change All studies were carried out in Western Sibutramine showed a 2003 Meta-analysis of Human studies Sibutramine >5mg/d Overall size effect -1.00 (-1.17 to -0.84) countries, which may limit applicability. large effect on weight but randomised double blind Double blind randomised, placebo Comparator Inclusion criteria may have led to under or increased blood pressure placebo controlled controlled trials of sibutramine Placebo over-estimation of the effect. significantly, particularly trials. English SBP in heavier and/or younger Objective Full text available Overall size effect 0.16 (0.08 to 0.24) Because only published studies were participants. The size of A comprehensive meta- Obese or overweight participants Size effect range -0.38 to 0.45 employed in the analysis there is a possibility the increase in blood analysis of RCTs on the Dosage of sibutramine > 5mg Net increased SBP attributable 1.6mm Hg of publication bias, however Rosenthal’s fail- pressure was small effect of sibutramine on Any duration to sibutramine safe N values were approximately N= 100 for however, it was weight loss and blood Co-morbidities of diabetes, SBP and N= 341 for DBP which makes considered to be clinically pressure. hypertension or hyperlipidemia DBP publication bias unlikely. important in patients who allowed. Overall size effect 0.26 (0.18 to 0.33) had borderline or high Studies included, N=21 Size effect range -0.11 to 0.73 blood pressure. Weintraub et al., 1991 Exclusion Net increased DBP attributable 1.8mm Hg Hanotin et al., 1998 Studies of patients with eating to sibutramine A greater increase in Seagle et al., 1998 disorders SBP was found in Apfebaum et al., 1999 Crossover studies subjects <44 years of age Bray et al., 1999 Studies that did not provide and in the larger trials Hansen et al., 1999 information for estimating a size (n>=120). Cuellar et al., 2000 effect of weight change. Fanghanel et al., 2000 Studies using drugs for weight Participants with a higher Finer et al., 2000 maintenance weight (>=92kg) Fujioka et al., 2000 Studies with multiple treatment experienced a higher Hazenberg et al., 2000 groups that did not compare each DBP and SBP. McMahon et al., 2000 group with the control group Dujovne et al., 2001 independently. The overall effect size Gokcel et al., 2001 was significantly larger Smith et al., 2001 when the sibutramine Wirth et al., 2001 dose was >15mg/d. Faria et al., 2002 McMahon et al., 2002 Serrano-Rios et al., 2002 Sramek et al., 2002 Zannad et al., 2002 DBP=diastolic blood pressure, SBP=systolic blood pressure
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 99
Evidence Table 23. Sibutramine, Hauner et al., 2003 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Comparison (including adverse events) Change in weight from baseline to 54 weeks Comments and year Inclusion P-S Placebo Sibutramine Exclusion p= Hauner et Study design Participants Exposure N=174 Weight loss (kg) 5.1(-6.1 to –4.1) 8.1(-9.2 to –6.9) <0.01 The populations were Adjunct treatment of obese al., 2003 Randomised, double N=389, ITT population Sibutramine 15mg/day Weight loss (%) 4.9 (4.0-5.8) 8.3 (7.1-9.6) ns similar at baseline except subjects with 15mg blind, placebo controlled N=348 given as an adjunct to a 5% weight reduction (%) 41.4 62.6 <0.001 for a significantly smaller sibutramine daily in clinical trial Consecutive, GP patients standard non- 10% weight reduction (%) 19.0 40.8 <0.001 number of males in the combination with a in the area of Cologne- pharmacological WHR -0.019 -0.024 ns sibutramine group. comprehensive non- Objective Dusseldorf, Germany. treatment. (-0.027 to –0.012) (-0.033 to –0.014) ns Analysis based on ITT and pharmacological weight To study, in a primary Male=89 (23%), Comparison N=174 Waist circumference (cm) -6.0(-7.1 to –5.0) -8.5cm (-9.7 to –7.2) ns at least one follow-up visits reduction program produces health care setting, the Female=259(67%) Placebo given as an SBP (mm Hg) -1.5 (-3.9 to 0.09) -2.9 (-5.1 to –0.7) ns with weight measurement. additional; weight loss and effect of a standardised Mean age =42.7 ± 11.7 adjunct to a standard non- DBP (mm Hg) -1.3(-2.8 to 0.1) -0.3(-1.8 to 1.3) <0.05 Last observation carried can be readily implemented non-pharmacological years pharmacological Heart rate (beats/min) -0.9 (-2.8 to 0.8) 1.9(0.1 to 3.8) ns forward (LOCF) was used in a primary care setting. program and 15mg treatment. Total cholesterol (mg/dl) -1.4(-7.0 to 4.2) -2.5 (-7.6 to 2.6) ns for the final analysis. sibutramine or placebo Inclusion Concomitant non- LDL- cholesterol (mg/dl) -8.7 (-15.0 to –2.3) -6.1 (-10.7 to –1.5) ns There were a significantly Treatment with sibutramine on long-term weight Age 18-65 years pharmacological therapy HDL - cholesterol (mg/dl) 3.9 (2.2 to 5.5) 5.6 (4.1 to 7.1) ns (p<0.05) smaller number of proved to be safe and reduction. BMI 30-40kg/m 2 Four educational sessions Triglycerides (mg/dl) 17.1 (-15.3 to 49.6) -9.9(-21.1 to 1.3) males in the sibutramine particularly effective in those Stable weight ( ±2kg) on food choice, physical group. obese subjects who were Trial duration during the preceding 3 activity, motivation and Adverse events less successful with the non- 54 weeks months behavioral modification 310/348 (85.6%) patients of the Withdrawals pharmacological program. Motivation and willingness and diet counselling. ITT population experienced 83.5% 87.8% Sibutramine =66 to reduce weight. Subjects were also adverse events including: Placebo=75 encouraged to attend 16 Back pain Main reasons for drop-out Exclusion further sessions. Bronchitis ns poor compliance and Serum creatinine >2mg/dl A dietitian made individual Sinusitis ns discontent with weight Hypercholesterolemia dietary counselling and Gastritis ns loss, sibutramine N=38, (total cholesterol energy requirement Ill defined experience ns placebo n=43. >320mg/dl) calculations. Subjects Headache ns Adverse event Uncontrolled hypertension were encouraged to record Infection ns withdrawals, sibutramine (>160/95mmHg, mean of their dietary intake for self- Pharyngitis ns n=22, placebo n=15. 3 independent control. Accidental injury ns measurements) Subjects were encouraged Enteritis ns Type 2 diabetes mellitus to increase their physical Dry mouth ns Coronary heart disease activity such that excess Surgery 0.001 Clinically significant energy expenditure by Flu syndrome ns dysrhythmia additional physical activity Fungal dermatitis ns Psychiatric diseases should exceed 1000kcal Upper respiratory tract infection 0.022 Child bearing potential per week. Gastrointestinal disorder ns without adequate Eczema ns contraception Hypertension ns Use of medication that Constipation ns may alter appetite. 0.002
Outcome figures in bold are statistically significant
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 100
Evidence Table 24. Sibutramine, Mcnulty et al., 2003 Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicabiilty Conclusions authors Objective Inclusion Comparison (including adverse events) (mean difference sibutramine – Comments and year Exclusion placebo ± SE or medians (range) S15mg/d 95%CI S20mg/d 95% CI McNulty et al., Study design Participants Exposure Weight change (kg) -5.1±0.9 -7.0 to –3.3 -7.8±1.0 -9.7 to-5.9 Relatively young patient Sibutramine was an 2003 Randomised, N=195 Sibutramine 15mg/day BMI (kg/m 2) -1.9-±0.3 -2.6 tp-1.2 -2.9±0.3 -3.6 to-2.2 population effective anti-obesity placebo controlled, Male=45%, Female=55% and 20mg/day and Waist circumference (cm) -4.9±0.9 -6.7 to-3.0 -6.8±1.0 -8.7 to-4.9 Strict inclusion criteria may agent in a substantial double blind, Mean Age=48-51years, standard diet advised by Hip -3.0±1.0 -5.0 to-1.0 -6.1± 1.0 -8.2 to-4.1 limit applicability. proportion of patients multicentre, and range 27-69years a dietitian. Waist hip ratio -1.4±0.9 -3.1 to 0.3 -0.9±0.9 -2.6 to 0.8 with type 2 diabetes. multi national trial. Duration=1 year Comparison >=5% weight loss (%) 46 65 Hypertensive =36% with Placebo and standard >=10% weight loss (%) 14 27 Average weight loss at Objective 29% taking anti- diet advised by a one year was 5.5kg To evaluate the hypertensive and 17% lipid dietitian. Metabolic changes with 15mg sibutramine effects of lowering drugs. Fasting glucose (mmol/l) -0.2 -1.0 to 0.7 -0.2 -1.0 to 0.5 and 8.0kg with 20mg – sibutramine (15 and HbA 1c (%) -0.34±0.33 -0.99 to 0.31 -0.10±0.34 -0.76 to 0.57 placebo patients lost no 20mg/day) on Inclusion Fasting insulin (pmol/l) -6.0 -24.0 to 12.0 -18.0 -30.0 to 0.0 weight. In fact 19% weight metabolic Type 2 diabetes > 6 mo Cholesterol (mmol/l) 0.0 -0.2 to 0.3 0.1 -0.1 to 0.3 gained more than 2kg control and blood BMI>=27kg/m 2 HDL 0.1 0.0 to 0.2 0.1 0.0 to 0.1 over 1 year. pressure in Metformin treatment 3 LDL 0.0 -0.2 to 0.3 0.1 -0.1 to 0.3 metformin obese months - 2 years Total cholesterol to HDL -0.5 -0.9 to-0.2 -0.2 -0.5 to 0.1 Weight loss of >= 10% subjects with type 2 Age 25-70 years cholesterol conferred metabolic diabetes. Fasting serum glucose 7- Triglycerides (mmol/l) -0.2 -0.6 to 0.1 -0.3 -0.6 to 0.0 and cardiovascular 15mmol/l benefits. No patients in Recruitment Cardiovascular effects the placebo group lost Eligible patients Exclusion SBP (mmHg) 4.6±2.2 0.3 to 8.8 -1.3 ±2.2 -5.6 to 3.1 >=10-% body weight. from case notes Ischemic heart disease, DBP (mmHg) 2.8±1.2 0.4 to 5.3 0.0±1.3 -2.5 to 2.4 heart failure or stroke. Pulse rate (beats per min) 5.9 ±1.7 2.5 to 9.4 5.8±1.8 2.3 to 9.3 Poorly controlled obese Seated pulse rate type 2 diabetic patients >100BPM, may gain the most from DBP>95mmHg Adverse events weight loss and anti- Fasting serum Sibutramine obesity medication. triglycerides>5.6mmol/l Dry mouth Lipid changes were Fasting serum cholesterol Constipation modest in this study >7.8mmol/l Insomnia and lower than other Serum creatinine studies but may be >120µmol/l explained by the fact Serum liver enzymes of that study participants bilirubin levels x2 upper limit were receiving of normal. metformin which has Weight change >3kg in favorable moderating preceding 3 months effect on lipids. Pregnant or childbearing women or childbearing Sibutramine tended to potential not taking increase blood adequate precautions again pressure in more pregnancy. patients than placebo and a few individuals showed marked rises. Figures in bold = significant, S=sibutramine.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 101
Evidence Table 25. Sibutramine, Wadden et al., 2000
Study Study Design Participants Exposure/ Outcomes Effect Size & Precision Validity/Applicability Conclusions authors Objective Inclusion Comparison (Mean change from baseline) Safety Comments and year Study Location(s) Exclusion S+P S+O P=** Recruitment Wadden et al., Study Design Participants Exposure N=17 Weight change in 16 week continuation Small study Conclusion 2000 Randomised, double 34 volunteers Sibutramine (10- trial –Last observation carried forward Short duration Body weight remained essentially blind, placebo Mean age 44.1 years 15mg/day) and orlistat + N=17 Diet modification /lifestyle therapy not unchanged in the two groups over controlled clinical trial. (±10.4) prescribed diet + a daily 1 month -0.7±1.2 -0.7±1.8 ns standard. the 16-week study period. Female =100% multivitamin 2 months -0.5±1.4 -0.7±2.9 ns Diets may have been very low fat. Objective BMI=33.9±4.9kg/m 2 3 months +0.1±1.8 -0.4±4.2 ns However, those that lost >=10% of To assess if adding 4 months +0.5±2.1 +0.1±4.1 ns Withdrawals their initial weight in the previous 1 orlistat to sibutramine Inclusion Comparison n=17 Weight change in 16 week continuation Sibutramine + orlistat, n=3 (18%) year gained weight during the induces further weight Females who had Sibutramine (10-15mg/d) trial –End point analysis Sibutramine, n=5 (29%) difference not second study regardless of loss or maintains completed 1 year and placebo + prescribed 1 month (N=16 S+P, N=14 S+O) -0.7±1.3 -0.9±1.9 ns significant. medication received. Those that weight lost in patients treatment with diet + a daily multivitamin. 2 months (N=10 S+P, N= 15 S+O) -0.3±1.6 -0.7±23.1 ns gained <10% in the initial study lost who have lost weight sibutramine (10- 3 months (N=12 S+P, N=15 S+O) +0.2±1.9 -0.6±4.4 ns Completers (EPA) and last observation a further amount of weight in the taking sibutramine 15mg/d) 4 months (N=10 S+P, N=14 S+O) +0.8±2.0 +0.3±4.2 ns carried forward (LOCF) analyses both second study with those taking alone and continue Prescribed diet reached the same conclusions. sibutramine + orlistat losing more taking sibutramine. 1200-1600kcal/day Adverse events (% of patients) Not clear if ITT analysis was carried weight than those taking representing a deficit of Soft stool out but a chi-squared test on the drop- sibutramine alone (ns). 9.1 50 0.04* Study location(s) 600-500kcal/day. Diet to Increased bowel movement out showed no significant difference. USA comprise 20% calories Faecal urgency 9.1 50 0.04* Further sub-group analyses of from protein, 50% Oily evacuation 9.1 42.9 ns Dosage patients thought most likely to Duration carbohydrate <=30% fat. Oily spotting 0 42.9 0.02* The medication dosage varied for both benefit from combination therapy 16 weeks after 1 year Healthy eating education. Flatus with discharge 9.1 28.6 ns treatments, did not increase their weight loss. of sibutramine Fat intake limited to Fatty oily stool 0 28.6 ns Sibutramine + placebo 60g/d. Liquid stool 0 28.6 ns 10mg/d sibutramine N=7 The authors suggest that most Exercise/activity Stomach pain/upset stomach 9.1 14.3 ns 15mg/d sibutramine N=-10 obese patients have a limit of 10- Monthly activity goals with Faecal incontinence 9.1 14.3 ns Sibutramine + orlistat 15% weight loss. Further weight eventual goal of 5 Decreased bowel movement 0 7.1 ns 10mg/d sibutramine N=6 loss appears to be thwarted by a sessions of 30-40mins Pellets/hard stool 0 7.1 ns 15mg/d sibutramine N=-11 toxic environment that discourages per week. 18.2 7.1 ns Orlistat 120mg t.i.d. physical activity, encourages The adverse events are reports for the Dose reduction of sibutramine had consumption of high fat diet and Overall – modest lifestyle last week of the trial. been made in the previous 1 year compensatory biological intervention. study because of side effects notably mechanism that decrease energy Combining the treatments did not appear insomnia, increased blood pressure expenditure. to result in any unexpected side effects. and pulse. Blinding 84.6% of patients had guessed what treatment they were on.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 102
Evidence Table 26. Comparative drug study, Poston et al., 2001
Study Study Design Participants Exposure/ Outcomes % of studies Validity/Applicability Conclusions authors Objective Inclusion Comparison (including adverse events) N=108 Comments and year Study Reference Exclusion
Poston et al., Study design/type Inclusion Included drugs Broad classification of components Most studies were conducted with Lifestyle treatments, with the exception of 2001 Meta-analysis of Studies published on or Amphetamine Behavioural modification studies 27.8% middle-aged overweight to moderately diets, have not been widely used in randomised, placebo before December 1999 Benzocaine Psychotherapy or cognitive behavioural 0.9% obese women in their 40s. Their randomised, placebo-controlled obesity drug controlled double FDA approved obesity Benzphetamine Exercise 17.6% outcomes may not be generally trials. blind trials of FDA medication Dexfenfluramine Diet 82.4% relevant to the larger population of approved anti-obesity Randomised studies where Diethylpropion Treatment manual for lifestyle intervention 3.7% obesity patients who seek treatment, Obesity-pharmacotherapy trials do not use agents. randomisation was not Fenfluramine Formal training by lifestyle interventionists 49.1% particularly those with more severe life-style treatments with the frequency broken Fluoxetine Specific classification of components obesity. expected based on the official positions of Objective Studies that included a Mazindol Very low calorific, pre-packaged food or Eating most professional organisations concerned Evaluation of life- direct comparison with Methamphetamine management 4.6% with the comprehensive management of style treatments used another anti-obesity drug or Orlistat Low-calorie diet 25.0% obesity. in published obesity placebo controlled. Phendimetrazine Balanced deficit diet 40.7% drug studies and their Data were for only human Phentermine Aerobics, Callisthenics or Lifestyle exercise 0.9% contribution to weight studies Phenylpropanalamine Weight lifting 0% losses associated English version of the study Sertaline Walking 2.8% with pharmacological available Sibutramine Self-monitoring 23.1% interventions. Data in published reports in Stimulus control or Contingency management 3.7% peer-reviewed journals Participants %outcome due to/ lifestyle component Studies included Relevant sub-groups 108 independent RCTs Trials with no lifestyle component 28.3% N=108 independent identifiable Trials with 1-3 lifestyle components 46-54% RCTs published in Studies with sufficient % weight loss attributable to specific 103 articles outcome data to compute a strategies in patients receiving drug 1960s n=10 size effect based on weight treatment 1970s n=46 loss Diet interventions vs. no diet interventions 49.5% vs. 26.5% 1980sn=13 1990s n=39 Exclusion Experimental obesity agents Nutritional supplements Weight maintenance studies
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 103
Evidence Table 27. Comparative drug study, Haddock et al., 2002 Study Study Design Inclusion Exposure/ Outcomes Wks Dose Drug minus Effect size Validity/Applicability Conclusions authors Objective Exclusion Comparison (including adverse events) Rx Mg/d placebo (kg) Comments and year Included Medications
Haddock et Study Design Inclusion Exposure Diethylpropion (9 studies) 18(6-51) 75 3.0(-1.6-11.5) <0.80 Standardised mean A large number of al., 2002 Meta-analysis of RCTs FDA approved anti- Orlistat Orlistat (6 studies) 48(16-76) 303(190-360) 2.08(0.30-4.2) <0.80 difference (d) based on studies did not obesity drugs Sibutramine Phentermine (6 studies) 13(2-24) 28 (15-30) 3.6(0.6-6.0) 0.81 change scores present codable Objective Prescription or OTC or Phentermine Sibutramine (4 studies) 15(8-26) 14(10-20) 3.5(2.4-5.1) 1.05 (reduction in weight) data. A comprehensive meta-analysis off label Diethylpropion used as a measure of of RCTs of medications for Studies published in effect size, where: Only sibutramine obesity. peer reviewed journals Comparison d1 = (Xt i -Xc I)/ S i had an effect size Medications Published before Placebo Xt i = mean weight loss exceeding 0.90. Amphetamine December 1999 treatment group (dexamphetamine) English version available Xc i = mean WL control The absolute Benzocaine Human studies group placebo-subtracted Benzphetamine Placebo or active Si = pooled SD of weight loss Dexfenfluramine† comparator change for both associated with anti- Diethylpropion* Randomised study Study weighting – obesity drug use in Fenfluramine† Outcome data available inverse function of the MA did not Fluoxetine sample variance exceed 4.0kg (i.e., Mazindol Exclusion modest). Methamphetamine Experimental obesity Female preponderance Sibutramine, which Orlistat* drugs was the drug with Phendimetrazine Dietary supplements the largest effect Phentermine (HCL and resin)* size, had Phenylpropanolamine (PPA) N=108 clinical trials overlapping CIs with Sertraline included phentermine. Sibutramine* Treatment length did * current review medications not influence effect size. † withdrawn 1997, NS = not significant – i.e., p=<0.05 * ns if Bonferroni’s correction for multiple tests is used. OTC = over the counter
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 104
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 105 Appendices
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 106
Appendix I. Summary of potential barriers to use of drugs included in the review
Potential barriers to use† Medication Contraindications Interactions Additional Cautions Phentermine Agitated patients # Hypotensive drugs ! May impair ability to engage in potentially HCL Alcoholics # Thyroid hormones hazardous activities – e.g., operating Advanced arteriosclerosis # MAO inhibitors machinery, driving a vehicle. 30mg Symptomatic cardiovascular disease # Chlorpromazine ! Potential for abuse (amphetamine family of Umine Patients with a history of drug abuse # Fenfluramine drugs) and dependence. Timedcaps Glaucoma # SSRIs ! Abrupt cessation after prolonged high doses Hypertension # Tricyclic antidepressants may result in extreme fatigue, depression and !5mg & 30mg Hyperthyroidism # Alcohol sleep pattern changes. Duromine TM Pregnant females ! # Anaesthetics Insulin requirement in diabetics may change Valvular heart disease and strict monitoring of blood glucose is Children and adolescents required. Diethylpropion Agitated patients # MAOI inhibitors ! Altered anti-diabetic drug requirements HCL Advanced arteriosclerosis # Other anorectic agents Patients with a history of drug abuse # General anaesthetics 75mg Glaucoma Tenutate Hypertension Dospan Hyperthyroidism Hypersensitivity or idiosyncrasy to the sympthomimetic amine Patients with pulmonary artery hypertension Severe hypertension Sibutramine Hypersensitivity to sibutramine or its ingredients # CNS active drugs ! Blood pressure and pulse rate should be HCL Organic causes of obesity # MAOI’s monitored in all patients as sibutramine has History of major eating disorders # Migraine therapy caused clinically relevant increases in blood 10mg & 15mg Psychiatric illness # Concomitant use of SSRI’s pressure in some patients. Reductil® Giulles de la Tourette’s syndrome # Agents that may raise blood pressure or MAOI’s in the past 2 weeks heart rate – e.g., decongestants, cough cold and ! Caution is advised in patients taking Other centrally acting drugs for the treatment of allergy medications and anti-inflammatory agents concomitant medicines known to affect psychiatric, weight reduction # Drugs that affect cytochrome P450metabolism haemostasis or platelet function. Tryptophan for sleep disorders # Co-administrations of ketoconazole or History of coronary artery disease, congestive ! erythromycin.Cimetidine Seizures have been reported in <0.1% of heart failure, tachycardia, peripheral arterial # Alcohol excess patients and sibutramine should be given with occlusive disease, arrhythmia or cerebro- caution to patients with a history of seizures. vascular disease. Inadequately controlled hypertension ! Sibutramine should be given with caution to Hyperthyroidism patients who have a history of motor or verbal Severe liver or renal impairment ticks. Benign prostatic hyperplasia/urinary retention Phaeochromocytoma Narrow angle Glaucoma Drug, medication or alcohol abuse Pregnancy or breastfeeding women Patients under 18 and over 65 years ! When given with anticoagulants INR values Orlistat Chronic malabsorption syndrome # Warfarin and other anti-coagulants should be monitored. 120mg Cholestasis ! Use of a multivitamin could be considered. Known hypersensitivity to orlistat or any of its # Decreased absorption of fat-soluble vitamins D, ! Patients should be advised to use dietary Xenical® components E and beta-carotene when co-administered with guidelines and fat intake should be orlistat. distributed over three main meals. ! Improved metabolic control in type 2 If a multivitamin is recommended it should be diabetes may allow dose reduction of oral taken at least two hours after the administration of hypo-glycaemic medication. orlistat or at bedtime. ! Monitor cyclosporine levels more frequently when co-administered with orlistat A reduction in cyclosporine plasma levels has ! Coagulation parameters should be been observed with orlistat. monitored in patients treated with oral anticoagulants. Reduction in the exposure to amiodarone and des- ! There may be reduced therapeutic effects ethylamiodarone has been recorded, but the of patients on amiodarone clinical effects are unclear. † For further details and information see http://www.medsafe.govt.nz/profs/Datasheet/u/uminetimedcaps.htm http://www.medsafe.govt.nz/profs/Datasheet/d/durominecap.htm http://www.medsafe.govt.nz/profs/Datasheet/t/tenuate%20dospantab.htm http://www.medsafe.govt.nz/profs/Datasheet/r/Reductilcap.htm http://www.medsafe.govt.nz/profs/Datasheet/x/Xenicalcap.htm
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 107
Appendix II. Search strategies
Medline strategy
1 exp Obesity/ (56799) 2 (obese or obesity).tw. (56362) 3 weight loss/ (8738) 4 appetite depressants/ (2734) 5 appetite/de (1343) 6 anti obesity agents/ (628) 7 or/1 6 (82997) 8 exp Phentermine/ (819) 9 duromine.tw. (3) 10 umine.tw. (0) 11 Diethylpropion/ (224) 12 tenuate dospan.tw. (9) 13 sibutramine hydrochloride.tw. (27) 14 Cyclobutanes/ (845) 15 reductil.tw. (8) 16 orlistat.ti. (184) 17 lactones/ (7467) 18 xenical.tw. (30) 19 complan.tw. (9) 20 Food, Formulated/ (4095) 21 or/8 20 (13334) 22 7 and 21 (1214) 23 randomized controlled trials/ (33133) 24 randomized controlled trial.pt. (191180) 25 random allocation/ (51130) 26 double blind method/ (78837) 27 single blind method/ (8255) 28 clinical trial.pt. (386534) 29 exp clinical trials/ (156308) 30 (clinic$ adj trial$).tw. (79327) 31 ((singl$ or doubl$ or treb$ or tripl$) adj (blind$ or mask$)).tw. (75413) 32 placebos/ (23136) 33 placebo$.tw. (84981) 34 randomly allocated.tw. (7750) 35 (allocated adj2 random).tw. (607) 36 meta analysis/ (5457) 37 (metaanaly$ or meta analy$).tw. (11523) 38 meta analysis.pt. (9354) 39 exp review, literature/ (2051) 40 (systematic adj (review$ or overview$)).tw. (5737) 41 or/23 40 (561105) 42 comment.pt. (253767) 43 letter.pt. (509370) 44 editorial.pt. (163392) 45 animal/ (3653674) 46 human/ (8547910) 47 45 not (45 and 46) (2808546)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 108
48 or/42 44,47 (3486417) 49 22 and 41 (408) 50 49 not 48 (390) 51 limit 50 to english language (348)
Medline (extra economics search)
1 economics/ (23811) 2 "costs and cost analysis"/ (32452) 3 cost allocation/ (1667) 4 cost benefit analysis/ (33314) 5 cost control/ (15851) 6 cost savings/ (4816) 7 cost of illness/ (6121) 8 cost sharing/ (949) 9 health care costs/ (11660) 10 direct service costs/ (677) 11 drug costs/ (6035) 12 hospital costs/ (4347) 13 health expenditures/ (8059) 14 economics, pharmaceutical/ (1367) 15 ((low or high) adj cost$).mp. (10695) 16 (health?care adj cost$).mp. (864) 17 (fiscal or funding or financial or finance).tw. (31542) 18 (cost adj (estimate$ or variable$)).mp. (810) 19 (unit adj cost$).mp. (578) 20 (economic$ or pharmacoeconomic$ or price$ or pricing).tw. (68154) 21 or/1 20 (210993) 22 exp Obesity/ (56799) 23 (obese or obesity).tw. (56362) 24 weight loss/ (8738) 25 appetite depressants/ (2734) 26 appetite/de (1343) 27 anti obesity agents/ (628) 28 or/22 27 (82997) 29 exp Phentermine/ (819) 30 duromine.tw. (3) 31 umine.tw. (0) 32 Diethylpropion/ (224) 33 tenuate dospan.tw. (9) 34 sibutramine hydrochloride.tw. (27) 35 Cyclobutanes/ (845) 36 reductil.tw. (8) 37 orlistat.ti. (184) 38 lactones/ (7467) 39 xenical.tw. (30) 40 complan.tw. (9) 41 Food, Formulated/ (4095) 42 or/29 41 (13334) 43 28 and 42 (1214) 44 comment.pt. (253767) 45 letter.pt. (509370)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 109
46 editorial.pt. (163392) 47 animal/ (3653674) 48 human/ (8547910) 49 47 not (47 and 48) (2808546) 50 or/44 46,49 (3486417) 51 21 and 28 and 42 (28) 52 51 not 50 (24) 53 limit 52 to english language (21) Embase
1 Diabetic Obesity/ (437) 2 obesity/ or morbid obesity/ (40399) 3 weight reduction/ (18379) 4 (obese or obesity).tw. (36765) 5 or/1 4 (62130) 6 Phentermine/ (751) 7 duromine.tn,tw. (11) 8 umine.tn,tw. (0) 9 Amfepramone/ (375) 10 tenuate dospan.tn,tw. (22) 11 Sibutramine/ (882) 12 reductil.tn,tw. (110) 13 Tetrahydrolipstatin/ (1055) 14 xenical.tn,tw. (323) 15 complan.tn,tw. (1) 16 complan/ (1) 17 reductil/ (882) 18 tenuate dospan/ (375) 19 duromine/ (103) 20 umine/ (0) 21 ensure/ (48) 22 (meal adj3 replace$).tw. (73) 23 or/6 22 (2308) 24 5 and 23 (1436) 25 exp meta analysis/ (18397) 26 (metaanaly$ or meta analy$).tw. (10151) 27 (systematic$ adj (review$ or overview$)).mp. (5300) 28 randomized controlled trials/ (85380) 29 clinical trial/ (296906) 30 randomization/ (10898) 31 single blind procedure/ (4761) 32 double blind procedure/ (47475) 33 crossover procedure/ (14990) 34 placebo/ (44490) 35 randomi?ed controlled trial$.tw. (13528) 36 (clinic$ adj trial$).tw. (63823) 37 ((singl$ or doubl$ or tripl$ or trebl$) adj (blind$ or mask$)).tw. (55628) 38 placebo$.tw. (65892) 39 (random$ adj allocat$).tw. (6611) 40 prospective study/ (38007) 41 or/25 40 (421878)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 110
42 case study/ (1663) 43 case report.tw. (66042) 44 abstract report/ or editorial/ or letter/ (409542) 45 or/42 44 (476020) 46 animal/ (7039) 47 exp animal experiment/ (644137) 48 46 or 47 (648030) 49 human/ (3829304) 50 48 not (48 and 49) (602008) 51 24 and 41 (588) 52 51 not 45 (554) 53 52 not 50 (553) 54 limit 53 to english language (497) 55 drug safety/ (67314) 56 drug contraindication/ (8022) 57 patient satisfaction/ (18988) 58 or/55 57 (92266) 59 58 and 24 and 45 (26) 60 54 or 59 (497) Cinahl
1 exp obesity/ or obesity, morbid/ (5024) 2 Weight Reduction Programs/ (295) 3 weight gain/ or "altered nutrition, more than body requirements (nanda)"/ (1092) 4 Weight Loss/ (1966) 5 (obese or obesity).tw. (3954) 6 or/1 5 (8392) 7 Appetite Depressants/ (216) 8 (phentermine or duromine or umine).tw. (20) 9 diethylpropion.tw. (0) 10 tenuate dospan.tw. (0) 11 Sibutramine/ (35) 12 reductil.tw. (0) 13 orlistat.tw. (54) 14 xenical.tw. (8) 15 complan.tw. (0) 16 food, formulated/ (609) 17 or/7 16 (916) 18 6 and 17 (227) 19 Meta Analysis/ (4451) 20 (meta analy$ or metaanaly$).tw. (2303) 21 (systematic$ adj (review$ or overview$)).mp. (5227) 22 "literature review"/ or "systematic review"/ (3178) 23 exp Clinical Trials/ (26696) 24 clinical trial.pt. (10842) 25 (clinic$ adj trial$).tw. (6139) 26 randomi?ed controlled trial$.tw. (5128) 27 ((singl$ or doubl$ or tripl$ or trebl$) adj (blind$ or mask$)).tw. (3569) 28 Placebos/ (2379) 29 placebo$.tw. (5466) 30 (random$ adj allocat$).tw. (774)
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 111
31 or/19 30 (36028) 32 18 and 31 (64) 33 limit 32 to english (64) Searches from other sources
In databases and all other sources without controlled vocabulary combinations of the index terms and additional keywords from the above strategies were used in the search.
Timing of the searches The searches were carried out during June and July 2004.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 112
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 113 References
Allison, D. B., Gadbury, G., Schwartz, L. G., Murugesan, R., Kraker, J. L., Heshka, S., Fontaine, K. R., et al. (2003). A novel soy based meal replacement formula for weight loss among obese individuals: a randomized controlled clinical trial. European Journal of Clinical Nutrition , 57, 514 522. Arterburn, D., & Hitchcock, N. P. (2001). Obesity. BMJ , 322, 1406 1409. Arterburn, D. E., Crane, P. K., & Veenstra, D. L. (2004). The efficacy and safety of sibutramine for weight loss: a systematic review. Archives of Internal Medicine , 164, 994 1003. Astrup, A., & Toubro, S. (2001). When, for whom and how to use sibutramine? International Journal of Obesity , 25, S2 S7. Avenell, A., Broom, I., Brown, T. J., Poobalan, A., Aucott, L., Stearns, S. C., Smith, W. C. S., et al. (2004). Systematic review of the long term effects and economic consequences of treatments for obesity and implications for health improvement. Health Technology Assessment , 8. Bergstrom, A., Hsieh, C. C., Lindblad, P., Lu, C. M., Cook, N. R., & Wolk, A. (2001a). Obesity and renal cell cancer: a quantitative review. British Journal of Cancer , 85, 984 990. Bergstrom, A., Pisani, P., Tenet, V., Wolk, A., & Adami, H. O. (2001b). Overweight as an avoidable cause of cancer in Europe. International Journal of Cancer , 91, 421 430. Berkowitz, R. I., Wadden, R. A., Tershakovec, A. M., & Cronquist, J. L. (2003). Behavior therapy and sibutramine for the treatment of adolescent obesity: a randomized controlled trial. JAMA , 289, 1805 1812. Bradley, P. S., Maddox, R. R., & North, W. K. (1999). Bupropion sr with phentermine for weight reduction, 51st Institute on Psychiatric Services: 29th October-2nd November 1999, New Orleans . Bray, G. A., Blackburn, G. L., Ferguson, J. M., Greenway, F. L., Jain, A. K., Mendel, C. M., Mendels, J., et al. (1999). Sibutramine produces dose related weight loss. Obesity Research , 7, 189 198. Bray, G. A., Hollander, P., Klein, S., Kushner, R., Levy, B., Fitchet, M., & Perry, B. H. (2003). A 6 month randomized, placebo controlled, dose ranging trial of topiramate for weight loss in obesity. Obesity Research , 11, 722 733. Brower, V. (2002). Fighting fat. New drugs against obesity in the pipeline. EMBO Reports , 3, 601 603. Brown, C. D., Higgins, M., Donato, K. A., Rohde, F. C., Garrison, R., Obarzanek, E., Ernst, N. D., et al. (2000). Body mass index and the prevalence of hypertension and dyslipidemia. Obesity Research , 8, 605 619. Calle, E. E., Rodriguez, C., Walker Thurmond, K., & Thun, M. J. (2003). Overweight, obesity, and mortality from cancer in a prospectively studied cohort of U.S. adults. New England Journal of Medicine , 348, 1625 1638.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 114
Carey, V. J., Walters, E. E., Colditz, G. A., Solomon, C. G., Willett, W. C., Rosner, B. A., Speizer, F. E., et al. (1997). Body fat distribution and risk of non insulin dependent diabetes mellitus in women. The Nurses' Health Study. American Journal of Epidemiology , 145, 614 619. Chan, J. M., Rimm, E. B., Colditz, G. A., Stampfer, M. J., & Willett, W. C. (1994). Obesity, fat distribution, and weight gain as risk factors for clinical diabetes in men. Diabetes Care , 17, 961 969. Davidson, M. H. (1999). Weight control and risk factor reduction in obese subjects treated for 2 years with Orlistat: a randomized controlled trial. JAMA , 281, 1174 1174. Derosa, G., Mugellini, A., Ciccarelli, L., & Fogari, R. (2003). Randomized, double blind, placebo controlled comparison of the action of orlistat, fluvastatin, or both on anthropometric measurements, blood pressure, and lipid profile in obese patients with hypercholesterolemia prescribed a standardized diet. Clinical Therapeutics , 25, 1107 1122. Ettinger, M.P., Littlejohn, T.W., Schwartz, S.L., Weiss, S.R., McIlwain, H.H., Heymsfield, S.B. (2003). Recombinant variant of ciliary neurotrophic factor for weight loss in obese adults: a randomized, dose ranging study. JAMA , 289, 1826 1832. Foxcroft, D. R., & Milne, R. (2000). Orlistat for the treatment of obesity: rapid review and cost effectiveness model. Obesity Reviews , 1, 121 126. Gardin, J. M., Schumacher, D., Constantine, G., Davis, K. D., Leung, C., & Reid, C. L. (2000). Valvular abnormalities and cardiovascular status following exposure to dexfenfluramine or phentermine/fenfluramine. Journal of the American Medical Association , 283, 1703 1709. Glazer, G. (2001). Long term pharmacotherapy of obesity 2000: a review of efficacy and safety. Archives of Internal Medicine , 161, 1814 1824. Gokcel, A., Karakose, H., Ertorer, E. M., Tanaci, N., Tutuncu, N. B., & Guvener, N. (2001). Effects of sibutramine in obese female subjects with type 2 diabetes and poor blood glucose control. Diabetes Care , 24, 1957 1960. Haddock, C. K., Poston, W. S. C., Dill, P. L., Foreyt, J. P., & Ericsson, M. (2002). Pharmacotherapy for obesity: a quantitative analysis of four decades of published randomized clinical trials. International Journal of Obesity , 26, 262 273. Halpern, A., & Mancini, M. C. (2003). Treatment of obesity: an update on anti obesity medications. Obesity Reviews , 4, 25 42. Halpern, A., Mancini, M. C., Suplicy, H., Zanella, M. T., Repetto, G., Gross, J., Jadzinsky, M., et al. (2003). Latin American trial or orlistat for weight loss and improvement in glycaemic profile in obese diabetic patients. Diabetes, Obesity & Metabolism , 5, 180 188. Hanefeld, M., & Sachse, G. (2002). The effects of orlistat on body weight and glycaemic control in overweight patients with type 2 diabetes: a randomized, placebo controlled trial. Diabetes, Obesity & Metabolism , 4, 415 423. Hansen, D., Astrup, A., Toubro, S., Finer, N., Kopelman, P., Hilsted, J., Rossner, S., et al. (2001). Predictors of weight loss and maintenance during 2 years of treatment by sibutramine in obesity. Results from the European multi centre STORM trial. Sibutramine Trial of Obesity
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 115
Reduction and Maintenance. International Journal of Obesity & Related Metabolic Disorders: Journal of the International Association for the Study of Obesity , 25, 496 501. Hauner, H., Meier, M., Jockel, K. H., Frey, U. H., & Siffert, W. (2003). Prediction of successful weight reduction under sibutramine therapy through genotyping of the G protein beta3 subunit gene (GNB3) C825T polymorphism. Pharmacogenetics , 13, 453 459. Hazenberg, B. P. (2000). Randomized, double blind, placebo controlled, multicenter study of sibutramine in obese hypertensive patients. Cardiology , 94, 152 158. Heck, A. M., Calis, K. A., McDuffie, J. R., Carobene, S. E., & Yanovski, J. A. (2002). Additive gastrointestinal effects with concomitant use of olestra and orlistat. Annals of Pharmacotherapy , 36, 1003 1005. Hensrud, D. D. (2000). Pharmacotherapy for obesity. Medical Clinics of North America , 84, 463 476. Hensrud, D. D. (2004). Diet and obesity. Current Opinion in Gastroenterology , 20, 119 124. Hensrud, D. D., Billington, C., Blum, H. T., & Gonzales Campoy, J. M. (2003). Pharmacological approaches to weight loss . Bloomington, M.N.: Institute for Clinical Systems Improvement (ICSI). Heymsfield, S. B., Segal, K. R., Hauptman, J., Lucas, C. P., Boldrin, M. N., Rissanen, A., Wilding, J. P., et al. (2000). Effects of weight loss with orlistat on glucose tolerance and progression to type 2 diabetes in obese adults. Archives of Internal Medicine , 160, 1321 1326. Heymsfield, S. B., Van Mierlo, C. A. J., Van Der Knaap, H. C. M., Heo, M., & Frier, H. I. (2003). Weight management using a meal replacement strategy: meta and pooling analysis from six studies. International Journal of Obesity , 27, 537 549. Hollander, P. A., Elbein, S. C., Hirsch, I. B., Kelley, D., McGill, J., Taylor, T., Weiss, S. R., et al. (1998). Role of orlistat in the treatment of obese patients with type 2 diabetes. A 1 year randomized double blind study. Diabetes Care , 21, 1288 1294. James, W. P., Astrup, A., Finer, N., Hilsted, J., Kopelman, P., Rossner, S., Saris, W. H., et al. (2000). Effect of sibutramine on weight maintenance after weight loss: a randomised trial. STORM Study Group. Sibutramine Trial of Obesity Reduction and Maintenance. Lancet , 356, 2119 2125. James, W. P. T. (2001). A new drug to maintain weight loss. Hospital Practice , 36, 22 23. Jung, R.T. (1997). Obesity as a disease. British Medical Bulletin, 53, 307 321. Kim, S. H., Lee, Y. M., Jee, S. H., & Nam, C. M. (2003). Effect of sibutramine on weight loss and blood pressure: a meta analysis of controlled trials. Obesity Research , 11, 1116 1123. Krempf, M., Louvet, J. P., Allanic, H., Miloradovich, T., Joubert, J. M., & Attali, J. R. (2003). Weight reduction and long term maintenance after 18 months treatment with orlistat for obesity. International Journal of Obesity & Related Metabolic Disorders: Journal of the International Association for the Study of Obesity , 27, 591 597.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 116
Lamotte, M., Annemans, L., Lefever, A., Nechelput, M., & Masure, J. (2002). A health economic model to assess the long term effects and cost effectiveness of orlistat in obese type 2 diabetic patients. Diabetes Care , 25, 303 308. Lean, M. E. J. (2001). How does sibutramine work? International Journal of Obesity , 25, S8 S11. Leung, W. Y., Neil Thomas, G., Chan, J. C., & Tomlinson, B. (2003). Weight management and current options in pharmacotherapy: orlistat and sibutramine. Clinical Therapeutics , 25, 58 80. Manson, J. E., Colditz, G. A., Stampfer, M. J., Willett, W. C., Rosner, B., Monson, R. R., Speizer, F. E., et al. (1990). A prospective study of obesity and risk of coronary heart disease in women. New England Journal of Medicine , 322, 882 889. McElroy, S.L., Arnold, L.M., Shapira, N.A., Keck, P.E., Rosenthal, N.R., Karim, M.R. (2003). Topiramate in the treatment of binge eating disorder associated with obesity: a randomized, placebo controlled trial. American Journal of Psychiatry, 160, 255 261. McKay, R. H. (1973). Long term use of diethylpropion in obesity. Current Medical Research & Opinion , 1, 489 493. McNulty, S. J., Ur, E., & Williams, G. (2003). A randomized trial of sibutramine in the management of obese type 2 diabetic patients treated with metformin. Diabetes Care , 26, 125 131. McTigue, K. M., Harris, R., Hemphill, B., Lux, L., Sutton, S., Bunton, A. J., & Lohr, K. N. (2003). Screening and interventions for obesity in adults: summary of the evidence for the U.S. Preventive Services Task Force. Annals of Internal Medicine , 139. Medicines Control Agency Committtee on Safety in Medicines (2004). Important safety message: European withdrawal of anorectic agents/appetite suppressants: new legal developments, no new safety issues: Licences for phentermine and amfepramone being withdrawn May 2001. Available from: http://www.mca.gov.uk/ourwork/monitorsafequalmed/safetymessages/anorectic.htm Medsafe (2003). Duromine™: phentermine capsules 15mg and 30mg: data sheet . Wellington: NZ, Ministry of Health. Medsafe (2004a). REDUCTIL®: sibutramine hydrochloride: data sheet . Wellington: NZ, Ministry of Health. Medsafe (2004b). Tenuate dospan: diethylpropion hydrochloride 75 mg tablet: data sheet . Wellington: NZ, Ministry of Health. Medsafe (2004c). Xenical®: orlistat 120mg capsules: data sheet . Wellington: NZ, Ministry of Health. Muls, E., Kolanowski, J., Scheen, A., & Van Gaal, L. (2001). The effects of orlistat on weight and on serum lipids in obese patients with hypercholesterolemia: a randomized, double blind, placebo controlled, multicentre study. International Journal of Obesity , 25, 1713 1721. Munro, J. F. (1979). Clinical aspects of the treatment of obesity by drugs: a review. International Journal of Obesity , 3, 171 180.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 117
Munro, J. F., MacCusih, A. C., & Wilson, E. M. (1968). Comparison of continuous and intermittent anorectic therapy in obesity. BMJ , 1, 352 354. NHS Centre for Reviews and Dissemination (1997). The prevention and treatment of obesity. Effective Health Care , 3. Nisoli, E., & Carruba, M. O. (2003). A benefit risk assessment of sibutramine in the management of obesity. Drug Safety , 26, 1027 1048. Oliveria, S. A., Felson, D. T., Cirillo, P. A., Reed, J. I., & Walker, A. M. (1999). Body weight, body mass index, and incident symptomatic osteoarthritis of the hand, hip, and knee. Epidemiology , 10, 161 166. O'Meara, S., Glenny, A. M., Sheldon, T., Melville, A., & Wilson, C. (1998). Systematic review of the effectiveness of interventions used in the management of obesity. Journal of Human Nutrition & Dietetics , 11, 203 206. O'Meara, S., Riemsma, R., Shirran, L., Mather, L., & ter Riet, G. (2001). A rapid and systematic review of the clinical effectiveness and cost effectiveness of orlistat in the management of obesity. Health Technology Assessment , 5, 1 81. O'Meara, S., Riemsma, R., Shirran, L., Mather, L., & ter Riet, G. (2002). The clinical effectiveness and cost effectiveness of sibutramine in the management of obesity: a technology assessment. Health Technology Assessment , 6, 1 97. Padwal, R., Li, S. K., & Lau, D. C. W. (2004). Long term pharmacotherapy for obesity and overweight. Cochrane Database of Systematic Reviews . Poston, W. S. C., Haddock, C. K., Dill, P. L., Thayer, B., & Foreyt, J. P. (2001). Lifestyle treatments in randomized clinical trials of pharmacotherapies for obesity. Obesity Research , 9, 552 563. Proietto, J., & Baur, L. A. (2004). Management of obesity. Medical Journal of Australia , 180, 474 480. Rimm, E. B., Stampfer, M. J., Giovannucci, E., Ascherio, A., Spiegelman, D., Colditz, G. A., & Willett, W. C. (1995). Body size and fat distribution as predictors of coronary heart disease among middle aged and older US men. American Journal of Epidemiology , 141, 1117 1127. Rissanen, P., Vahtera, E., Krusius, T., Uusitupa, M., & Rissanen, A. (2001). Weight change and blood coagulability and fibrinolysis in healthy obese women. International Journal of Obesity & Related Metabolic Disorders: Journal of the International Association for the Study of Obesity , 25, 212 218. Rothman, R. B., & Baumann, M. H. (2000). Neurochemical mechanisms of phentermine and fenfluramine: therapeutic and adverse effects. Drug Development Research , 51, 52 65. Rothman, R. B., Redmon, J. B., Raatz, S. K., Kwong, C. A., Swanson, J. E., & Bantle, J. P. (2000). Chronic treatment with phentermine combined with fenfluramine lowers plasma serotonin. American Journal of Cardiology , 85, 913 915. Schuh, L. M., Schuster, C. R., Hopper, J. A., & Mendel, C. M. (2000). Abuse liability assessment of sibutramine, a novel weight control agent. Psychopharmacology , 147, 339 346.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC 118
Silverstone, T. (1992). Appetite suppressants: a review. Drugs , 43, 820 836. Sjostrom, L., Rissanen, A., Andersen, T., Boldrin, M., Golay, A., Koppeschaar, H. P. F., & Krempf, M. (1998). Randomised placebo controlled trial of orlistat for weight loss and prevention of weight regain in obese patients. Lancet , 352, 167 172. Steel, J. M., Munro, J. F., & Duncan, L. J. (1973). A comparative trial of different regimens of fenfluramine and phentermine in obesity. Practitioner , 211, 232 236. Tambascia, M. A., Geloneze, B., Repetto, E. M., Geloneze, S. R., Picolo, M., & Magro, D. O. (2003). Sibutramine enhances insulin sensitivity ameliorating metabolic parameters in a double blind, randomized, placebo controlled trial. Diabetes, Obesity & Metabolism , 5, 338 344. Thearle, M., & Aronne, L. J. (2003). Obesity and pharmacologic therapy. Endocrinology & Metabolism Clinics of North America , 32, 1005 1024. Torgerson, J. S. (2004). Preventing diabetes in the obese: the XENDOS study and its context. British Journal of Diabetes & Vascular Disease , 4, 22 27. Torgerson, J. S., Hauptman, J., Boldrin, M. N., & Sjostrom, L. (2004). XENical in the prevention of diabetes in obese subjects (XENDOS) study: a randomized study of orlistat as an adjunct to lifestyle changes for the prevention of type 2 diabetes in obese patients. Diabetes Care , 27, 155 161. Truant, A. P., Olon, L. P., & Cobb, S. (1972). Phentermine resin as an adjunct in medical weight reduction: a controlled, randomized, double blind prospective study. Current Therapeutic Research, Clinical & Experimental , 14, 726 738. Wadden, T. A., Berkowitz, R. I., Womble, L. G., Sarwer, D. B., Arnold, M. E., & Steinberg, C. M. (2000). Effects of sibutramine plus orlistat in obese women following 1 year of treatment by sibutramine alone: a placebo controlled trial. Obesity Research , 8, 431 437. Warren, E., Brennan, A., & Akehurst, R. (2004). Cost effectiveness of sibutramine in the treatment of obesity. Medical Decision Making , 24, 9 19. Weintraub, M., Hasday, J. D., Mushlin, A. I., & Lockwood, D. H. (1984). A double blind clinical trial in weight control. Use of fenfluramine and phentermine alone and in combination. Archives of Internal Medicine , 144, 1143 1148. Weintraub, M., Rubio, A., Golik, A., Byrne, L., & Scheinbaum, M. L. (1991). Sibutramine in weight control: a dose ranging, efficacy study. Clinical Pharmacology & Therapeutics , 50, 330 337. Wilding, J. (2004). Clinical evaluation of anti obesity drugs. Current Drug Targets , 5, 325 332. Wilson, B. D., Wilson, N. C., & Russell, D. G. (2001). Obesity and body fat distribution in the New Zealand population. New Zealand Medical Journal , 114, 127 130. Wirth, A., & Krause, J. (2001). Long term weight loss with sibutramine: randomized controlled trial. JAMA , 286, 1331 1339.
EVIDENCE BASED REVIEW OF WEIGHT LOSS MEDICINES: A REPORT COMMISSIONED BY THE NEW ZEALAND ACC