This review compared outcomes for exercise programmes in cognitively impaired and intact older adults and concluded that strength and endurance outcomes were similar for cognitively impaired and intact participants, which suggested that the former should not be excluded from exercise programmes. The authors' conclusions reflected data presented, but should be interpreted with caution due to uncertainties surrounding the methodologies used.

To compare endurance and strength outcomes in cognitively impaired (Mini Mental State Score (MMSE) below 24) and cognitively intact (MMSE 25 or above) older adults who participated in similar exercise programmes.

Abstracts International, Ageline, CINAHL, The Cochrane Library, Dissertation Online, ERIC, MEDLINE, PEDro, PsycINFO, PsycLIT, PubMed and Sport Discuss were searched. Search words were indicated. A manual search and cross-referencing from review and primary research articles were performed. Only English-language articles published in a peer-reviewed journal between January 1970 and December 2006 were included.

Studies were eligible if they were randomised controlled trials (RCTs) including more than five participants. (Of the included trials, two were non-randomised controlled trials.)

Eligible trials were of study participants older than 65 years who reported a baseline Mini Mental State (MMSE) score of less than 25 for cognitively impaired participants and 25 or above for cognitively intact participants. Alternatively, participants could have been diagnosed by a doctor as having some degree of cognitive impairment or have a pre-existing diagnosis of dementia. Mean age in included trials was 81 years (standard deviation (SD)=5, range 68 to 91). Mean MMSE score was 16 (SD=6, range 6 to 23) in trials of cognitively impaired participants and 28 (SD=1, range 26 to 29) in trials of cognitively intact participants. There were more women than men in the trials.

Trials had to compare an exercise programme or some other form of rehabilitative exercise, physical activity, fitness or recreational therapy with a non-intervention comparison group. Trials had to report at least one health-related physical fitness outcome (cardiovascular, endurance, strength, flexibility, body mass index) and include means and standard deviations, t-test or F-test and numbers of participants. Endurance or strength training or a combination was used. Mean exercise frequency was three sessions per week (SD=1, range 2 to 6), mean intensity was 50 minutes per session (SD=16, range 20 to 90) and mean duration was 16 weeks (SD=8, range 2 to 40).

Studies that did not report age, sample size or exercise mode were not considered in the analysis.

The authors stated neither how the papers were selected for the review nor how many reviewers performed the selection.

Methodological quality of included studies was assessed using the UTMB/TLC Interventions Trial Quality Form. Items assessed were listed. Studies were defined as very high quality if the score was 22 or above, high quality for scores between 19 and 21, medium quality for scores of 16 to 18 and low quality for scores of 15 and less.

Two reviewers independently assessed trial quality. Scoring differences were discussed until consensus was reached.

Individual effect sizes (ES) were calculated by dividing the difference between experimental and control groups by the estimated pooled standard deviation of the baseline outcome measure in treatment and control groups (Hedge’s g). Effect sizes were presented with 95% confidence levels.

The authors stated neither how data were extracted for the review nor how many reviewers performed data extraction. Items used for data extraction were listed.

Pooled data were presented as summary effect size (SES) based on the weighted mean of the unbiased individual efect size values, calculated using a fixed-effects method. Positive numbers reflected improvements in performance and negative numbers deterioration. Effect sizes of less than 0.49 were interpreted as small, 0.49 to 0.79 were interpreted as medium, and 0.80 or larger as large. Subgroup analysis was conducted separately by exercise type (strength or endurance).

Forty-one trials (n=2,921) were included, of which 21 were exercise trials in cognitively impaired participants (n=1,411) and 20 were exercise trials in cognitively intact participants (n=1,510). Mean trial quality was 18 (SD=2, range 13 to 23).

Large to moderate effect sizes were found for strength and endurance outcomes for cognitively impaired groups (SES 0.51, 95% CI 0.42 to 0.60; 21 trials, 31 comparisons) and cognitively intact groups (SES 0.49, 95% CI 0.40 to 0.58; 20 trials, 33 comparisons). Effect sizes for strength-only trials were moderate for cognitively impaired (13 trials) and large for cognitively intact participants (16 trials). Effect sizes for endurance-only trials were moderate for both cognitively impaired (18 trials) and cognitively intact participants (17 trials).

No statistically significant difference in effect size was found between studies in cognitively impaired and those in cognitively intact participants for strength and endurance combined (p=0.263), strength only (p=0.132) and endurance only (p=0.078). There seemed to be a larger variance in effect sizes of the trials in cognitively intact individuals, possibly due to a broader range of types of strength training and workloads used.

The results suggested that cognitively impaired older adults who participated in exercise programmes had similar strength and endurance training outcomes as age- and gender-matched cognitively intact older participants; therefore, cognitively impaired individuals should not be excluded from exercise rehabilitation programmes.

This systematic review addressed a clearly stated research question with appropriate definition of inclusion criteria. The literature search included a range of relevant databases and additional searches. Search terms were listed. But the authors included only published studies, which introduced potential for publication bias (which was not assessed in the review). Inclusion was restricted to English-language articles only, so relevant non-English articles may have been neglected. Appropriate methods were used to reduce reviewer bias and error in the validity assessment; it was unclear whether similar methods were used for study selection and data extraction. Study flow was described and sufficient details of the included studies were given. It was difficult to determine the appropriateness of pooling the results as heterogeneity between studies was not assessed. The authors did acknowledge that there may have differences in the type and intensity of training received by cognitively impaired and cognitively intact individuals, but this was not explored using subgroup analysis. Also, the analysis was based on indirect comparisons between trials in cognitively impaired and cognitively intact participants, but there were no data to compare cognitively impaired and cognitively intact participants directly in the same trial. There was no information about whether exercise rehabilitation programmes should be designed for cognitively impaired and cognitively intact groups separately or if mixed groups might be possible. It would have been interesting to see whether cognitively impaired individuals benefited from the exercise programmes with respect to other outcomes (such as quality of life); such outcomes were not considered. The authors’ conclusions follow from data presented, but should be interpreted with caution given uncertainties that surround the appropriateness of methodologies used and potential for missed studies.

Practice: The authors suggested that geriatric healthcare policy and practice should be reconsidered and cognitively impaired older adults should be provided with physical rehabilitation as a standard of care and opportunities to participate in community-based exercise programmes.

Research: The authors stated that more research was needed not only of direct comparisons of exercise programmes in cognitively impaired and intact individuals but also comparisons of participants with different degrees of cognitive impairment. More research was needed to define predictors of maximum benefit and determine the cost of implementing exercise programmes for older cognitively impaired adults.

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.