This review concluded that specific classes of medications could be withdrawn successfully without causing harm in people aged 65 years and older. Although the authors' conclusions appeared to reflect the evidence, these should be interpreted with caution given the poor quality of the included studies, the limited literature search and problems with the reporting of the review.

To assess the benefits and harms of withdrawal of medications in people aged 65 years and older.

PubMed was searched for articles published between 1966 and 2007. Search terms were reported. References of identified articles were manually searched.

Randomised controlled trials (RCTs) that compared the effects of withdrawal of certain classes of medications versus a control (continued medication) in adults with a mean age of 65 years or more were eligible for inclusion. Prospective observational studies were also eligible for inclusion. Eligible studies were required to be part of a combined multi-factorial intervention and report on at least one clinically relevant outcome after medication withdrawal. Studies that assessed general reductions in medications and trials that did not have a factorial design (which prohibited the effect of medication withdrawal to be established) were excluded.

Included studies were conducted in in-patient or out-patient settings, nursing homes, general practices, health centres or the community. Medication/class withdrawal included thiazide (diuretics), antihypertensives, psychotropics and miscellaneous (including digoxin, donepezil and nitrates). Where reported, between 20% and 100% of patients were successfully withdrawn from medication. Treatment durations ranged from three weeks to five years.

One reviewer screened the relevant studies for inclusion.

The authors did not state how the validity assessment was performed.

Two reviewers independently extracted data on clinical benefits and harms. Data for individual studies were not presented. It was unclear how discrepancies were resolved.

There was evidence of heterogeneity, so data were presented as a narrative synthesis and in tables, grouped by class of medication and by study type for antihypertensives and diuretics.

Thirty two studies (n=8,972) were included in the review: 19 RCTs and 13 prospective observational studies. There was a discrepancy in the number of patients reported in the text and tables (we used the number reported in the text). Sample sizes ranged from 15 to 6,291 patients; most studies included 300 or fewer participants. The authors reported that few studies were of high quality, but no further details were reported.

Diuretics (four RCTs): Between 87% and 100% of patients treated with thiazides were successfully withdrawn from medication over a period of six to 52 weeks compared to between 51% and 100% of controls. Serious adverse effects (such as heart failure) were reported to be rare.

Antihypertensives (10 observational studies): Between 20% and 85% of patients were successfully withdrawn over a four-week to five-year period. The main reason for recommencing antihypertensive therapy was a gradual increase in blood pressure, and sometimes heart failure.

Psychotropics (11 RCTs): The studies showed inconsistent results between patients who withdrew from psychotropics compared to those who did not; most studies reported no significant differences in withdrawal symptoms between the two groups.

Outcomes for miscellaneous medication withdrawal (seven studies) were inconsistent and were reported in the review.

There was some evidence to suggest that specific classes of medications could be successfully withdrawn in older people without causing harm.

The review question was clear and was supported by broad inclusion criteria. The search of the literature was limited to one electronic database and one other appropriate source. The authors acknowledged difficulties with identifying appropriate search terms to capture all the relevant studies. As no search for unpublished studies was undertaken, potentially relevant studies may have been missed. The authors reported that the overall quality of the included studies was not high. The authors did not report the process for undertaking the validity assessment, the review process for study selection was only carried out by one reviewer and it was unclear how discrepancies for data extraction were resolved. Thus, reviewer error and bias were not ruled out. Few details on study and patient characteristics were reported, therefore, it was unclear whether patients and treatment regimens were comparable at baseline. The authors acknowledged the presence of heterogeneity among studies and a narrative synthesis was appropriate. The authors also acknowledged certain limitations with the included studies (such as small trial numbers).

Although the authors' conclusions appeared to reflect the evidence, these should be interpreted with caution given the poor quality of the included studies, the limited literature search and the problems with the reporting of the review.

Practice: The authors stated that the results of the review could not be generalised beyond withdrawal of thiazides, antihypertensives and psychotropic medications, and were limited to the main outcome of successful withdrawal rate.

Research: The authors stated that future research should use easily identifiable and unique search terms to identify relevant studies. The authors also stated that there is a need for further research into the benefits and harms of adverse reactions in older people after medication withdrawal to guide recommendations on withdrawal.

Not stated.

Alexander GC, Sayla MA, Holmes HM, et al. Prioritizing and stopping prescription medicines. CMAJ 2006; 174: 1083-1084.

Holmes HM, Hayley DC, Alexander GC, et al. Reconsidering medication appropriateness for patients late in life. Arch Intern Med 2006; 166: 605-609.

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.