I read with interest the excellent editorial on drugs for obesity by Dr John Muir (Guidelines in Practice, March 1999). However, I disagree with some of his conclusions.

Obesity is associated with increased morbidity and mortality. In Britain, approximately 16% of the population are obese. The incidence of obesity has doubled in less than 20 years, suggesting an urgent need for effective treatment. Those who achieve 10% weight loss on diet have no need for drug intervention, but some people cannot lose weight unaided.

Dr Muir quotes orlistat as achieving an average weight loss of 7.6% over one year, with associated reduction in lipid and glucose risk factors. Quesenberry et al1 showed that a difference of weight of this amount will, on average, result in the heavier person spending approximately half a day more per year as a hospital inpatient.

Reducing hospital stays by half a day in 16% of the population would produce a vast potential cost saving – which Dr Muir has not brought into his equations – as well as other savings such as a 23% decrease in the need for oral hypoglycaemics in Type 2 diabetes.2

It seems unfair, in this instance, to subtract the so-called placebo response of 4.5% from the orlistat response rate. A placebo is, by definition, inactive. In the orlistat trials the placebo groups were far from inactive. The so-called placebo group were in monthly attendance at an obesity clinic. The true placebo group were those who simply carried on eating as before.

A real measure of the success of orlistat would be to add the projected 7.6% weight reduction to the estimated weight increase that statistics show would have been likely to occur had these patients not been on medication.

The studies show that 4.5% weight loss can be achieved by regularly attending a dietetic clinic. In primary care we are not funded to provide such clinics, and our area, like many parts of Britain, is covered by just one dietetic hospital outpatient clinic per week. It is not practical to swamp this by referring 16% of our patients to it on a monthly basis.

Presumably Dr Muir does provide these clinics, but he does not include their cost in his financial calculations. In addition, such attendances usually necessitate these patients taking a half day off work, with further cost to the community.

Twelve hospital clinic outpatients visits per year would cost my health authority more than a year's treatment with orlistat and would be little more than half as effective.

I feel that the facts quoted by Dr Muir do, in fact, support the use of orlistat as a cost-effective treatment for obesity.

Dr JR Stott, GP, Barry, South Wales

  1. Quesenberry CP, Caan B, Jacobson A. Arch Intern Med 1998; 158: 466-72.
  2. Lucas CP, Segal K, Hollander PA et al. Paper presented at the ICO, Paris 1998.

We read with interest Dr Muir's editorial (Guidelines in Practice, March 1999), and welcome the much-needed guidance provided by the Royal College of Physicians' Report on clinical management of the overweight and obese. However, we feel that the editorial was not entirely balanced.

Data from more than 7000 patients have been collected in clinical studies on orlistat (Xenical), and much of these data have been published. The RCP Report cites just three papers. However, these do refer to data from nearly 900 patients.

The Report recommends drug treatment only for selected patients and clearly states that 'not all obese patients respond to drug therapy'. The Xenical Summary of Product Characteristics states that: 'Treatment with orlistat should only be started if diet alone has previously produced a weight loss of at least 2.5kg over a period of four consecutive weeks. Treatment with orlistat should be discontinued after 12 weeks if patients have been unable to lose at least 5% of their body weight as measured at the start of drug therapy.' It is therefore the intention that only the responder population should receive Xenical.

When discussing the weight loss and cost implications, Dr Muir includes both responder and non-responder data from the study results in his calculations. Consequently, the cost per kilogram lost that he quotes has little relevance to clinical practice.

Regarding Dr Muir's comments on fat-soluble vitamin absorption, both of the published two-year studies have addressed this issue, and in both cases levels remained within the reference range in all treatment groups.

The gastrointestinal side-effects of the drug, according to Sjostrom et al,1 'happened early in orlistat treatment and were of short duration (<4 days)'. As the RCP Report says, these effects are 'predominantly related to its therapeutic action' and may help to enforce behavioural change.

The Report states that a 10% weight loss should be the primary goal of treatment. Nearly 40% (38.8%) of Xenical patients achieved >10% weight loss over the first year, and in the more recently published paper,2 to which Dr Muir refers, results were very similar, with 38.9% of orlistat-treated patients (responders and otherwise) achieving more than 10% weight loss. More than a third (34.1%) of subjects completing 2 years maintained >10% weight loss compared with 17.5% of subjects receiving placebo.

At the ECO Congress held in Paris last year, Sjostrom presented further analysis of his study.3 This suggests that 54% of the patients who, according to the licence, would be eligible to receive orlistat would achieve >10% weight loss in one year.

The Report also states that 'a weight loss of between 5% and 10% of the initial body weight reduces the health risks associated with obesity', and a significant proportion of patients receiving Xenical achieved and maintained >5% weight loss.

Sarah Ridout (Medical Director) and Dr Patricia Campbell (Medical Advisor), Roche Products Ltd

  1. Sjostrom et al. Lancet 1998; 352: 160-1.
  2. Davidson et al. JAMA 1999; 281: 235-42.
  3. Data on file 14119c presented at Paris ECO

I was very surprised to read, in the Guideline Development article (Guidelines in Practice, March 1999), that evidence obtained from meta-analysis is graded 1a.

I totally agree with Feinstein, quoted in Dr Haque's letter (March 1999, p. 41), that most meta-analyses are 'statistical alchemy', with similar weight being given to non-compatible trials of all levels, and then firm conclusions being reached at the end. Statistical validity is not improved by pooling incompatible data sets.

Dr Clive Brady, GP, Surrey

Guidelines need to be accepted by the whole primary health care team. Nurses are frequently given the role of collecting evidence, producing drafts etc.

Indeed, they are often the ones asking for practice guidelines to ensure the best/consistent treatment for patients, and to protect themselves.

What is needed is information about nurses and the law in relation to guidelines, as well as information for the whole primary health care team in the context of PCGs. Practice is not only about GPs.

Margaret Hanily, Primary Care Facilitator, Stockport

I perform audit in local surgeries and assist the health authority, and now PCGs, in developing local guidelines that are relevant to local needs.

I am currently developing practice policy on H. pylori eradication, and found the article on this topic (Jan/ Feb 1999) very useful. It has reinforced pertinent points.

Dr Sharon Hadley, GP Tutor/Course Organiser, Leigh-on-Sea

Guidelines in Practice, April 1999, Volume 2
© 1999 MGP Ltd
further information | subscribe