Clinical question — Does donepezil treatment benefit community resident patients with mild to moderate Alzheimer-type dementia?
The evidence. Long-term donepezil treatment in 565 patients with Alzheimer's disease (AD2000): randomised double-blind trial. Lancet 2004; 363: 2105-2115.1
Background. Dementia is a major healthcare issue. Currently it is estimated that 650 000 individuals have dementia in the UK at an estimated annual cost of £6.1 billion.
The advent of the cholinesterase inhibitors, such as donepezil, was heralded as a major advance in the management of patients with dementia. However, many of the trials have been criticised in relation to the unrepresentative nature of the participants, and the dubious relevance of the outcome measures.2 Particular concern has been expressed about the clinical and health service relevance of the various cognitive measures of patient response and the short lengths of patient follow-up.
GPs and primary care trusts require evidence to judge how cost-effective treatments, such as donepezil, are among unselected community dwelling patients with Alzheimer-type dementia.
Study design and intervention. AD2000 was a double-blind, randomised trial. It initially involved 565 patients referred to memory clinics within the West Midlands. To be included patients had to have a DSM IV diagnosis of Alzheimer-type dementia, be living in the community and have a regular carer. Moreover the treating/recruiting doctor had to be ‘substantially uncertain that the individual would obtain a worthwhile benefit from donepezil taking into account the available evidence and clinical circumstances’. The main trial was preceded by a randomised run-in treatment period of 12 weeks after which 486 patients were re-randomised to either 5 mg or 10 mg donepezil per day or to placebo.
From the GPs' perspective a key feature of the AD2000 trial is the clinical relevance of the study population to the types of patients with Alzheimer-type dementia likely to be encountered in day-to-day general practice. In particular, over half of the study sample had a co-morbidity (previous myocardial infarction, other cardiovascular disease, stroke, hypertension, or requirement for aspirin).
Outcomes and analysis. In addition to assessing the overall impact on institutionalisation or progression of disability, the authors examined behavioural and psychological symptoms (using the MMSE [Mini-mental state examination] and the BADLS [Bristol activities of daily living scale]), formal care costs, unpaid caregiver time and carer psychopathology using the General Health Questionnaire (GHQ).
It is important to appreciate that the original stated objective in AD2000 was to recruit 3000 patients. However, only 486 patients entered the study after the run-in period and a mere 20 patients reached the end of the third study year. In these circumstances great care must be taken to assure that the researchers have not succumbed to a beta error, whereby a true effect is missed due to inadequate numbers of participants. This problem is particularly pertinent in relation to the outcomes, such as institutionalisation, that are determined by an interaction of an amalgam of different influences, for example, biological, social, environmental and political.
Results. Cognition averaged 0.8 MMSE points better and functionality 1.0 BADLS points better with donepezil over the first 2 years. The authors also stated that there were no differences between the two treatment arms in terms of institutionalisation (42% compared with 44% at 3 years) or progression of disability (58% compared with 59% at 3 years). Overall healthcare costs were increased among the patients on donepezil largely due to the added burden of hospital overnight stays. However, some specific costs were reduced, for example, those relating to social workers, domestic helpers and ‘unpaid’ caregivers. Carers' psychological morbidity scores were 0.3 GHQ points lower with donepezil compared to placebo, albeit with broad confidence intervals due to the small sample size (95% CI = −0.3 to 0.9).
Commentary. AD2000 attempted to assemble a study population representative of the types of patients encountered in routine practice. Unfortunately, although the patients recruited undoubtedly had an enhanced co-morbidity burden than in many such trials, there is some evidence of imbalance between the two treatment groups. Individuals on donepezil had slightly more co-morbidities (149 compared with 138), were slightly older (163 patients aged 70–79 years compared with 155) and were more likely to be male (118 compared with 113). Some concerns must also be expressed about the impacts of doctors' judgements (for example, the requirement for substantial uncertainty that ‘the individual would obtain a worthwhile benefit from donepezil taking into account the available evidence and clinical circumstances’) and the run-in period on the characteristics of the study population. There is a particular concern that all these factors may combine together to bias the study results.
In interpreting the results of this trial it is necessary to be aware that the trial was underpowered (especially for the global and multifaceted outcomes) and probably biased in favour of the placebo. Thus, I would argue that there is a need to be wary of the authors' suggestion that there was no difference in institutionalisation, rate of disability or overall healthcare costs. However, it is worth noting that, even in this study, donepezil improved cognition and functionality and reduced the costs due to social workers, domestic helpers and ‘unpaid’ caregivers. Furthermore, as a GP, I was particularly interested to note that the carers' psychological morbidity scores were lower with donepezil compared with placebo.
The bottom line. In community resident patients with mild to moderate Alzheimer-type dementia, donepezil treatment improved cognition and functionality, reduced some community care costs and impacted on the psychological morbidity of carers.
- © British Journal of General Practice, 2004.