Published: 14/10/2004, Volume II4, No. 5927 Page 29
The issue Around 650,000 people have dementia in the UK at an estimated cost of£6.1bn per year. Cholinesterase inhibitors such as donepezil were heralded as a major advance in managing patients with dementia. But many trials have been criticised for the unrepresentative nature of the participants and the possible impact of commercial sponsorship.
The study AD2000 was an independently funded, double-blind, randomised trial (Lancet 2004; 363: 2105-2115, www. lancet. com). It initially involved 565 patients referred to memory clinics in the West Midlands. To be included, patients had to be diagnosed with Alzheimer-type dementia, live in the community and have a regular carer. Crucially, the treating/recruiting doctor had to be 'substantially uncertain that the individual would obtain a worthwhile benefit from donepezil'. The main trial was preceded by a run-in period of 12 weeks after which 486 patients were selected to enter the study proper, using 5mg/10mg of donepezil per day or a placebo.
The study population had high clinical relevance to patients with Alzheimer-type dementia encountered in day-to-day general practice.
In particular, over half had a comorbidity.
The study demonstrated that donepezil improved cognition and functionality, reduced some community care costs and impacted on the psychological morbidity of carers. But it failed to identify improvement in terms of institutionalisation, overall healthcare costs or the progression of the disability.
In practice Many primary care trusts remain concerned about the effectiveness of expensive drugs such as donepezil. The risk of trials such as AD2000 is that they are incorrectly interpreted, leading to inappropriate reductions in donepezil provision.
Although the original objective in AD2000 was to recruit 3,000 patients, only 486 were selected to enter the study proper and a mere 20 patients reached the end of the third study year. It is therefore possible that true effects are missed due to inadequate numbers of participants.
This is particularly pertinent to outcomes such as institutionalisation that are determined by a complex interaction of influences. The authors' suggestion that there were no differences in institutionalisation or overall healthcare costs should therefore be treated with caution.
AD2000 attempted to assemble a study population representative of patients encountered in routine practice. Unfortunately, although the patients had a higher comorbidity burden than in many such trials, there is evidence of imbalance between the two treatment groups.
Concerns must also be expressed about the selective effect of doctors' judgements in the run-in period on the study population. These may have biased the study results in favour of placebo.
The bottom line is that in an independent study, donepezil improved cognition and functionality and reduced costs of social workers, domestic helpers and 'unpaid' care givers. Furthermore, the carer's psychological morbidity scores were lower with donepezil.
Dr Nick Summerton is a GP, reader in primary care medicine and medical director of Yorkshire Wolds and Coast PCT.