Depression, anxiety, and dementia, as well as older age, female gender, lower education level, and decreased physical activity, have all been associated with memory loss reported by patients or family members (strength of recommendation [SOR]: B, cross-sectional studies). Memory complaints in patients with no cognitive impairment on short cognitive screening tests, such as the mini-mental status exam, may predict dementia (SOR: B, longitudinal studies). No consistent evidence supports pharmacologic treatment of reported memory loss that is not corroborated by objective findings (SOR: B, nonrandomized, poor-quality studies).
Is depression or polypharmacy at work?
Rajasree Nair, MD
Baylor College of Medicine, Houston, Tex
As the population ages, primary care physicians encounter a significant number of patients with memory loss and dementia. In clinical practice, patients with subjective memory complaints but normal cognitive testing present a diagnostic dilemma. Close attention to comorbid psychiatric conditions such as depression, anxiety, and substance use disorders, as well as polypharmacy, is essential.
While the US Preventive Services Task Force indicates that there is insufficient evidence to screen, it notes that recognizing cognitive impairment early not only facilitates diagnostic and treatment decisions, but also allows clinicians to anticipate problems the patient may have in understanding and adhering to recommended therapy. Even though evidence of early or minimal dementia may be difficult to detect, identifying it promptly enables physicians to counsel patients and caregivers on the course of disease progression, warning signs, medication adherence, finances, and safety.
Evidence Summary
Several cross-sectional studies indicate that patients with subjective memory loss are more likely to be older, female, less physically active, in poorer health, less educated, and more depressed or anxious than unaffected patients.1-4 These studies concentrate mostly on elderly people living in the community.
A study of 1883 patients with normal baseline short-cognitive test results found that those with subjective memory complaints had a higher incidence of dementia.5 At 5-year follow-up, 15% of patients with baseline subjective memory complaints had developed dementia compared to only 6% of those without such complaints (odds ratio=2.7; 95% confidence interval [CI], 1.45-4.98).
A prospective cohort study that followed 158 patients with no evidence of dementia showed a significant correlation between informant-reported memory problems and development of dementia at 5 years.6 Forty-five percent of patients with informant-reported memory problems developed dementia after 5 years compared with 25% of patients who had only self-reported memory problems (P =.02). This result suggests that subjective memory problems reported by observers (family or caregivers) may be more predictive of dementia than self-reported memory complaints.
Donepezil, ginkgo biloba may not help these patients
Most trials of interventions to preserve memory have not enrolled patients with subjective memory complaints. However, data from trials that enrolled either asymptomatic elderly patients or patients with mild cognitive impairment don’t support the use of donepezil, ginkgo biloba, NSAIDs, COX-2 inhibitors, vitamin E, vitamin B6, vitamin B12, statins, hormone replacement therapy, or omega-3 fatty acids to delay progression to dementia.7-16
Could mental exercise help?
One systematic review of 22 longitudinal cohort studies, which included more than 29,000 patients, evaluated complex patterns of mental activity in early, mid-, and late-life in relation to the incidence of dementia. Dementia was diagnosed at a significantly lower rate in patients with a higher level of cognitive exercise, such as memory-based leisure activities and social interactions, than those with less rigorous daily cognitive challenges (relative risk=0.54; 95% CI, 0.49-0.59).17
This raises the possibility that mental exercise has neuroprotective effects. No randomized trials exist to support this hypothesis, however.
