. Donepezil in patients with severe Alzheimer's disease: double-blind, parallel-group, placebo-controlled study. Lancet. 2006 Apr 1;367(9516):1057-65. PubMed.


Please login to recommend the paper.


  1. This Swedish nursing home trial demonstrates the practical utility of donepezil in severely impaired Alzheimer patients and provides support for its use. It shows that more patients with severe illness actually improve their cognitive functioning and to a relatively marked extent over their baselines. By comparison, patients with mild to moderate AD generally only maintain their function with the drug over a 6-month trial. The overall group effect on the Severe Impairment Battery was a mean improvement of three or four points above baseline in the donepezil group and did not depend on the placebo group continuing to decline in order to gain statistical significance. More than half of the treated patients showed overall clinical improvement compared to about one-third of placebo patients on the CIBIC+, that is, 53-59 percent versus 36-38 percent.

    This trial demonstrates again the realistic limitations of these medications, that they can improve intellectual functioning, communication, and general clinical state, while not actually improving activities of daily living but attenuate their worsening.

    To a large extent, the trial contradicts the claims by manufacturers that these drugs must be used early in the course of illness and continued indefinitely, and undermines claims that the drugs alter the course of the illness. The fact that donepezil shows this robust an effect in severe dementia demonstrates the potential that many patients will show improvements at various times throughout their illness.

    Many physicians have been using cholinesterase inhibitors in patients with severe dementia even though the labeling is for mild to moderate AD. So the results support donepezil use here, provided there are no medical contraindications such as significant cardiovascular or pulmonary disease, syncope, gastrointestinal disease, seizures, etc.

    David Hogan, in an accompanying editorial, correctly questions aspects of the clinical significance of donepezil because this and nearly all other cholinesterase inhibitor trials were not designed to measure the number of individual patients who actually gain worthwhile benefits. (The way the CIBIC+ clinical global rating is used doesn’t quite do this). The study reports mainly average changes on test scales that don’t translate into the numbers of individuals who uniquely benefit from treatment. Yet the overall clinical impression from the CIBIC+ of 53-59 percent improving with donepezil versus 36-38 percent with placebo reported in the trial (a statistically “moderate” effect size implying that the number needed to treat for one patient to benefit could be about 5 or 6) would be quite clinically significant if those particular patients also could be shown to have sustained their improvements and to have improved their cognitive function. Without the use of appropriate individual patient outcomes, the clinical meaning of the statistical effects of cholinesterase inhibitors and memantine will continue to be elusive.

    There is no evidence, one way or another, that prescribing a cholinesterase inhibitor—or memantine for that matter—would either increase or decrease overall costs for nursing home patients. (No doubt we will soon see such claims after “economic” analyses are bootstrapped on to this trial). Regardless, and more importantly, there is nothing wrong with paying for a treatment that really improves patients’ function and quality of life. Moreover, people with severe dementia who live in nursing homes can and do have meaningful quality of life, and there is nothing wrong with prolonging their survival even if it costs more. There is no credible evidence, however, that these drugs prolong survival, just that they improve symptoms. (There is also no credible evidence that they delay nursing home placement in more mildly impaired patients).

    It is a bit of an overstatement for David Hogan to describe the effect in this study as “too little, too late,” as some patients can meaningfully benefit, probably even some who may have received cholinesterase inhibitors previously and had not benefited from them then. It could be that cholinesterase inhibitors have their best and most pronounced effects in people who are further along in their illness with more severe cholinergic deficits. This, however, might run counter to the interests of some who want the medications to be started early in the course and continued indefinitely. The political climate may not allow us to accept these trial results with as much enthusiasm as they deserve.

    View all comments by Lon S. Schneider

Make a Comment

To make a comment you must login or register.

This paper appears in the following:


  1. Donepezil—Relief for Patients with Severe AD?