Letters

Effectiveness of rivastigmine in Alzheimer's disease

Improvements in functional ability remain unestablished

Authors' reply

Patients' view on quality of life should be assessed

Guidelines do not ignore clinically relevant end points

Improvements in functional ability remain unestablished

EDITORTwo recent reports on rivastigmine in Alzheimer's disease ^{1 2} provide further proof that cholinesterase inhibitors produce modest improvements in cognitive testing and in clinical impression of change. The new claim is of improved functionality with rivastigmine, which, if true, would be an important advance in the management of Alzheimer's disease.

Unfortunately, however, these studies do not establish that functional ability is improved. Both studies rated functionality using the progressive deterioration scale, which was developed to assess quality of life not activities of daily living.³ It contains considerable duplication (for example, four questions on handling finances), and only two items relate peripherally to the basic activities of dressing and eating. It cannot be concluded, therefore, that improved scores equate to improved functionality.

Moreover, Rösler et al misrepresent the small improvement in progressive deterioration score seen with rivastigmine (2.8 on a 100 point scale) by citing in the discussion that one third of patients taking higher dose rivastigmine attained at least a 10% improvement in score without noting that 20% of placebo patients also improved to this extent. The benefit was actually only 13% (33% v 20%), which is reduced to 10% (29% v 19%) on more appropriate intention to treat analysis.

The intention to treat analyses are also potentially biased because of non-random drop outs: 77 (32%) of 243 higher dose rivastigmine patients did not have a 26 week assessment compared with 31 (13%) of the 239 placebo patients. Alzheimer's disease is progressive and so replacing missing data by carrying forward values obtained earlier in the trial underestimates natural deterioration. No improvements in progressive deterioration score were seen in the lower dose rivastigmine group, which had the same drop out rate as the placebo group.

Thus it remains unclear whether cholinesterase inhibitors produce sufficient benefit in Alzheimer's disease to justify their widespread use. Clearly, any delays in progress to severe dependency or institutionalisation would be worth while both clinically and economically. Improved functionality, fewer neuropsychiatric symptoms, and reduced burden and stress on carers would also be important. But none of these has been reliably established for rivastigmine or donepezil. Longer term placebo controlled trials addressing these outcomes are urgently required.⁴ One such study, the national AD2000 donepezil trial, has recently opened and already includes 150 patients. To resolve current uncertainties about the best use of cholinergic agents, widespread supportfrom clinicians and purchasersfor studies such as AD2000 should be encouraged.

Peter Bentham, consultant in old age psychiatry. 
Mental Health Services for the Older Adult, Queen Elizabeth Psychiatric Hospital, Birmingham B15 2QZ

Richard Gray, professor of medical statistics. 
Elizabeth Sellwood, AD2000 trial coordinator. 
University of Birmingham Clinical Trials Unit, Institute of Clinical Research, Medical School, Birmingham B15 2TH

James Raftery, professor of health economics. 
University of Birmingham Health Economics Facility, Park House, Birmingham B15 2RT on behalf of the AD2000 Trial Steering Committee

^a Competing interests: None declared.

Authors' reply

EDITORWe agree with Sellwood et al on the need for longer term trials (2 years) to assess economic and disease specific outcomes. However, such trials cannot be performed until efficacy is proved in six month studies and can be performed only once the drug is registered.

It should be noted that items in the progressive deterioration scale are based on a comprehensive evaluation of activities performed in day to day life by patients with Alzheimer's disease and were selected following input from carers¹: those who deal with patients on a day to day basis and best know their activities. Therefore the scale meaningfully reflects patients' functional ability. Although some items seem to cover similar activities, this was done deliberately as an internal cross check to ensure the validity of information provided by the carer.

Our data show clearly that the functional ability of patients treated with rivastigmine improves over six months. Also, significantly more patients treated with rivastigmine experience a highly clinically relevant improvement in activities of daily living (10% improvement on the progressive deterioration scale) compared with placebo. In comparison, clinical studies have shown that untreated patients worsen consistently. ^{2 3} Although the change reported in our study may not seem large, any stabilisation or reduction in loss of functional ability results in important clinical benefits in this progressive and debilitating disease. Furthermore, in a pooled analysis of phase III studies with rivastigmine, clinically and statistically significant improvement was noted for 22 of these items compared with placebo.⁴

We agree that an intention to treat analysis is not appropriate for a disease characterised by progressive worsening. Indeed, an observed case analysis also showed significant clinical benefits for rivastigmine (29% v 18% for placebo; P=0.012).

Finally, Sellwood et al will be pleased to note that a placebo controlled study with a duration of treatment and follow up of three years is under way to examine the delay to diagnosis of Alzheimer's disease with rivastigmine in an at-risk patient population. In addition to cognitive, behavioural, and pharmacoeconomic measures, this trial (which includes 900 subjects from 12 countries) will examine the effect of rivastigmine on reducing the rate of brain atrophy using quantitative magnetic resonance imaging.

Michael Rösler, Head. 
Sektion Gerontopsychiatric, Psychiatrische Universitätsklinik, D 97089 Würzburg, Germany On behalf of the B303 Exelon Study Group

^b Competing interests: Four of the authors of the paper are employed by Novartis.

Patients' view on quality of life should be assessed

EDITORIn commenting on the recent paper showing efficacy and safety of rivastigmine in patients with Alzheimer's disease,¹ Flicker refers to a modest improvement in carer rated quality of life.² There is no consensus, however, on how to assess quality of life in dementia and no quality of life instrument used in clinical trials to date has been satisfactory.³ The progressive deterioration scale, which was used in this trial, is regarded as a measure of functional ability (activities of daily living) and not quality of life. Patients with mild to moderate dementia are, however, able to describe and rate their quality of life, and their views on treatment should be taken into consideration.

In a study of patients starting on adjunctive antiepileptic drugs we found that side effects and adverse events were important indicators of quality of life. Indeed, some patients who became completely free of seizures opted to stop taking the adjunctive drugs because of unwanted side effects, notably weight gain.⁴ Without objective data we cannot know how patients with dementia would balance small improvements in cognition or activities of daily living against side effects such as nausea, vomiting, and diarrhoea. Similarly, we have few data to guide us on how changes in the various domains assessed in dementia trials relate to quality of life from the patients' perspective.

Assessment of quality of life in dementia is in its infancy and raises many technical and ethical issues. Several measures of patient self report have, nevertheless, been developed and are now becoming available. In future, clinical trials of antidementia drugs should incorporate measures of patient-reported quality of life alongside proxy measures.⁵

Caroline E Selai, research psychologist. 
c.selai{at}ion.ucl.ac.uk

Michael R Trimble, professor in behavioural neurology. 
Raymond Way Neuropsychiatry Research Group, National Hospital For Neurology and Neurosurgery, London WC1N 3BG

Martin N Rossor, professor of clinical neurology. 
Richard J Harvey, director of research. 
Dementia Research Group, National Hospital For Neurology and Neurosurgery, London WC1N 3BG

Guidelines do not ignore clinically relevant end points

EDITORIn his commentary on the European rivastigmine study¹ Bayer states that the absence of measures of neuropsychiatric outcome and the burden on carers is unfortunate but that the choice of these end points was governed by requirements of regulatory authorities rather than the aim of measuring the real impact of the illness on the lives of patients and their families. He says that the need for clinically relevant outcome measures should now be better appreciated.

However, the European guideline on medicinal products in the treatment of Alzheimer's disease recommends that improvement of symptoms should be assessed in the following three domains: cognition, as measured by objective tests (cognitive end point); activities of daily living (functional end point); and overall clinical response, as reflected by global assessment (global end point).² Efficacy variables should be specified for each of the three domains. Two primary variables should be stipulated, one evaluating the cognitive end point and the other the clinical relevance of the improvement in cognition. The protocol should specify this second primary variable and to which domain (global, or preferably functional) it is related. Moreover, the instruments which measure burden on carers and activities of daily living should have been validated for Alzheimer's disease. The study should be designed to show significant differences in at least two of the primary variables. If this is achieved, then the overall benefit (response) should be assessed in individual patients, and the effect of treatment should be illustrated in terms of the proportion of patients who achieve a meaningful benefit (responders).

In our view these recommendations show that the emphasis of regulatory authorities is on clinically relevant end points, and the Committee for Proprietary Medicinal Products guideline stresses the need to develop these. Therefore, the statement that the choice of the outcomes in clinical drug trials in dementia was governed by the requirements of regulatory authorities is not justified.

J G Storosum, psychiatrist. 
B J van Zwieten-Boot, pharmacologist. 
A J A Elferink, clinical epidemiologist. 
Medicine Evaluation Board of the Netherlands, Kalvermarkt 53, PO 16229, 2500 BE Den Haag, Netherlands