To produce an evidence-based review to support recommendations from the US Preventive Services Task Force (USPSTF) concerning dementia syndrome screening in primary care settings.
We searched MEDLINE, PsycINFO, EMBASE, and the Cochrane Collaboration library from January 1994 to January 2001, with all searches limited to English language studies.
We developed an analytic framework comprising 9 key questions on dementia screening and treatment to be answered by systematic review. Next, we developed inclusion and exclusion criteria for each question. For questions of prevalence and accuracy of screening, we required cross-sectional or cohort studies in a primary care population with an acceptable reference standard test. For questions of treatment, we included randomized controlled studies (RCTs) of subjects with mild to moderate dementia. We included studies of 6 potential outcome domains: (a) cognitive, physical, and social function; (b) health care utilization rates; (c) behavioral symptoms of dementia; (d) caregiver stress; (e) accidents and injuries; and (f) health-related quality of life.
Two reviewers extracted data from included studies of fair to good quality for the preparation of evidence tables. We rated the quality of all selected studies using USPSTF methodology for study appraisal.
: Does screening for dementia in primary care settings affect any of the selected outcomes? We were unable to locate any RCTs or systematic reviews that addressed this question. : What is the prevalence of undiagnosed dementia in primary care patients? Two studies in North American populations showed that 1.8% and 5.7% of persons older than age 65 have undiagnosed dementia; 2 studies in non-US populations reported prevalence rates of undiagnosed dementia of 3.2% and 12%. : Does a reliable and valid screening test exist to detect dementia in primary care patients? Good evidence shows that the Folstein Mini-Mental State Examination (MMSE) has a sensitivity of 71% to 92% and specificity of 56% to 96% in primary care populations. : Do pharmacological interventions improve any of the selected outcomes? The efficacy of pharmacological intervention varies with the etiology of dementia. We found no evidence of benefit from anti-inflammatory drugs, estrogen, nimodipine, or aspirin in the treatment of dementia. We found no RCTs of treatments for vitamin B deficiency, thyroid disease, neurosyphilis, normal pressure hydrocephalus, or sleep apnea. Observational data show that no more than 1.5% of all cases of mild to moderate dementia are fully reversible. Multiple well-conducted RCTs show that for Alzheimer's disease, cholinesterase inhibitors improve cognitive and global function and delay functional decline by 3 to 5 months. One study shows that vitamin E and selegiline postpone functional loss by 7 months. Another study shows that gingko biloba produces a delay of approximately 3 months in cognitive decline. Some studies show that typical and atypical neuroleptics reduce agitated behaviors in patients with varied stages of dementia. One RCT found that clomipramine reduces depressive symptoms in early dementia. Another RCT found that sertraline reduced depressive symptoms in AD. : Do nonpharmacologic interventions improve any of the selected outcomes? Only limited evidence supports the use of nonpharmacologic behavioral interventions in advanced dementia, but this type of treatment has not been studied in early dementia. : Do caregiver interventions improve caregiver or patient outcomes? Five fair quality RCTs of intensive caregiver interventions found no direct benefit for either the patient or the caregiver. Two of these studies show a delay in nursing home placement of 11 to 19 months. : What are the adverse effects of dementia screening? No study meeting our inclusion criteria addressed this question. : What are the costs and cost-effectiveness of dementia screening? No study meeting our inclusion criteria addressed this question. : What are the side effects of dementia therapy? In RCTs of dementia therapy, dropout rates because of adverse effects ranged from 0% for antidepressant therapy to 27% from gastrointestinal side effects of high-dose rivastigmine.
The prevalence and burden of the dementia syndrome are high after age 65. The majority of patients with early dementia are undiagnosed in primary care practices. A brief interview screen can detect dementia with reasonable accuracy. Pharmacologic and nonpharmacologic treatments show benefit on outcomes in mild to moderate Alzheimer's disease, but it is not clear how many subjects in these studies were detected by screening. Evidence for benefit of treatment for other etiologies of dementia syndrome is more limited than that for Alzheimer's disease.
