Treatment of Alzheimer Disease and Prognosis of Dementia: Title and subTitle BreakTime to Translate Research to Results

In this issue of JAMA, 2 studies address the treatment and prognosis of patients with dementia. Quinn and colleagues¹ describe results from a well-conducted trial conducted in conjunction with the Alzheimer's Disease Cooperative Study (ADCS) network designed to determine if treatment with the omega-3 fatty acid docosahexaenoic acid (DHA) improved the symptoms and course of patients with mild to moderate Alzheimer disease (AD). The multisite trial enrolled 402 patients (60% were assigned to DHA and 40% to placebo). At the end of the 18-month trial, there were no treatment group differences on any of the primary or secondary outcomes despite elevations of plasma phospholipids and cerebrospinal fluid DHA (measured in a subgroup). Among the 102 participants with baseline and follow-up brain magnetic resonance imaging, there were also no effects of treatment on total brain or hippocampal volume changes. However, among patients without an apolipoprotein E (APOE) ε4 allele, those assigned to DHA had less decline compared with those receiving placebo on 1 primary outcome and 1 secondary outcome.

This trial adds to a growing literature that treatment with DHA does not improve symptoms of AD. Although several observational studies reported that diets rich in fish or supplements with omega-3 fatty acids were associated with reduced risk of developing AD, most randomized clinical trials for treatment of AD or mild cognitive impairment or in healthy elderly individuals have not found a beneficial effect.² The ADCS trial attempted to overcome some prior methodological issues by enrolling patients who consumed 200 mg/d or less of DHA and documenting elevation of phospholipids and DHA with treatment. The post hoc analysis indicating possible benefit in patients without an APOE ε4 allele is of great interest because the benefit was observed for the cognitive outcomes but not for the functional or behavioral outcomes. However, as the authors point out, the clinical significance of this finding is uncertain.

Access to hospice and other end-of-life care is critically important for patients with advanced dementia, although the use of these services has been limited. Part of the problem has been lack of accurate prognostic tools that could be used to identify patients eligible for hospice care. In their article in this issue of JAMA, Mitchell and colleagues³ describe efforts to prospectively validate a prognostic tool designed to identify those patients with advanced dementia with greater risk of mortality. The investigators used the Advanced Dementia Prognostic Tool (ADEPT) to estimate 6-month survival among more than 600 nursing home patients with advanced dementia and compared its performance with Medicare hospice criteria. Despite concerted efforts to gather data independently from the Minimum Data Set, to enroll only patients with severe impairment, and to use rigorous analytic methods, the ADEPT only modestly identified 6-month survival (using continuous scores, area under the receiver operating characteristic curve = 0.67). While the ADEPT was somewhat better than the Medicare hospice criteria and had excellent interrater reliability, the sensitivity and specificity using various cut points was far from clinically useful.

The fact that these 2 studies are basically “negative” in which both attempts at identifying a new treatment for AD and in trying to identify patients with dementia at highest risk of death were modest at best gives the field pause once again. Why, despite a remarkably productive 2 decades of neuroscientific advances in the underpinnings of AD and other forms of dementia, are there still only minimally effective treatments and only rough estimates for prognostication? Similar disappointing results in the prevention of AD were summarized in the recent report from the National Institutes of Health–sponsored State of the Science Conference on Preventing Alzheimer Disease and Cognitive Decline.⁴ The field seems to be at a crossroads. On the one hand, there have been tremendous scientific achievements especially in the genetic and molecular mechanisms of AD. In addition, the utility of biomarkers and diagnostic imaging for patients with dementia has blossomed, prompting new efforts to incorporate these into diagnostic guidelines.⁵ On the other hand, little improvement has occurred in the basic care of patients with AD and other dementias including treatment of cognitive and behavioral symptoms and ability to predict the progression and adverse outcomes of dementia.

It is likely that the individual course of dementia is too dependent on a complex interplay of biological, psychological, and cultural characteristics to lead to easy prognostication. For example, several studies suggest large racial and ethnic differences in rates of nursing home admissions and mortality rates for patients with dementia, but the reasons for these differences are not totally understood.⁶ Yet the need to be able to predict dementia prognosis is greater than ever, not only for identification of those at highest risk to develop dementia but also for identification of those most likely to be admitted to nursing homes and for those who might benefit from hospice or other palliative care programs.

Effective treatment strategies to prevent progression of AD will likely need to be initiated earlier in the course of AD in order to be more efficacious. Given the accumulating evidence that the pathophysiological process of AD begins years, if not decades, prior to the diagnosis of clinical dementia, treatment of mild to moderate AD may be “too late.” Viewing dementia more in a life course model with earlier interventions for those at risk might yield greater results. This will require tremendous shifts in trial design, outcomes, and funding mechanisms. At this point, there are few therapeutic strategies for early interventions other than an active lifestyle and reduction or prevention of cardiovascular disease⁷ ; nonetheless, efforts should be made to investigate whether early cardiovascular and lifestyle modifications can affect downstream cognitive function.

Because AD is such a devastating illness and current therapeutic choices are limited and only moderately effective, new treatment options are urgently needed. Recently, almost every trial for treatment of AD has led to disappointing results, including several with β-amyloid–modifying agents.⁸ If aging and AD is a complex adaptation to insults that begin decades before symptoms emerge, targeting one downstream mechanism may not be effective. In addition, while β-amyloid most likely has a central role in the cascade of neuronal degeneration, there are also important contributions from tau protein, synaptic dysfunction, and vascular changes.⁹ The next steps in AD treatment may need to incorporate a combined regimen similar to the treatment approaches used for other chronic diseases. Another possibility is to combine pharmacologic strategies with behavioral interventions, an approach that has not been investigated.

Those affected by AD and other forms of dementia will increase exponentially over the coming decades.¹⁰ Therefore, there is an even greater imperative to improve both the “T1” research in successfully moving from molecules to patients but also in “T2” research in moving from efficacy to effectiveness and public policy. The studies in this issue of JAMA highlight the high-quality investigation essential for progress, but also the continued frustration over lack of effective interventions and prognostication.