Forget NSAIDs for Alzheimer's Prevention

BALTIMORE, April 25 -- Neither naproxen (Aleve) nor celecoxib (Celebrex) reduced the risk of developing Alzheimer's disease, at least in the short term, investigators from the suspended ADAPT trial reported.

According to results of the randomized study funded by the National Institute on Aging, there was even a suggestion that both naproxen and celecoxib were associated with a slightly increased risk of dementia, reported Constantine Lyketsos, M.D., M.H.S., of Johns Hopkins, and colleagues in the ADAPT trial.

The findings run counter to those of earlier observational studies suggesting that chronic use of NSAIDs could help to prevent Alzheimer's and other forms of dementia, the investigators reported in an online release from the May 22 issue of Neurology.

"Although our study was conducted to test the hypothesis that celecoxib or naproxen would reduce the incidence of Alzheimer's disease, these results indicate no such effect, at least within the first few years after treatment begins," said Dr. Lyketsos.

The ADAPT (Alzheimer's Disease Anti-inflammatory Prevention Trial) study was halted in December of 2004, after investigators in a different trial (the Adenoma Prevention with Celecoxib study) announced that the drug was associated with significantly increased cardiovascular risk.

Interest in the use of NSAIDs for Alzheimer's prevention was sparked by a study, published in Neurology in 1993, indicating that indomethacin in doses of 100 to 150 mg/day appeared to protect mild-to-moderately impaired Alzheimer's disease patients from the degree of cognitive decline exhibited by well-matched controls, Dr. Lyketsos and colleagues noted.

"However, several subsequent trials have shown no benefit from treatment of Alzheimer's disease with rofecoxib, naproxen, nimesulide, or diclofenac," they wrote. "A trial of rofecoxib to prevent progression to Alzheimer's disease in those with mild cognitive impairment (often a prodromal phase of Alzheimer's disease) was similarly disappointing."

The ADAPT study was designed to test the hypothesis that NSAIDs might be able to protect cognitively normal people against the depredations of Alzheimer's.

The study was a randomized, placebo-controlled, double-masked clinical trial at six U.S. dementia research clinics. The participants were volunteers 70 and older who were considered cognitively normal, with cognitive screening scores above designated cutoffs, and a family history of Alzheimer's disease.

Beginning in early 2001, the patients were randomized to celecoxib at 200 mg twice daily, naproxen sodium at 220 mg twice daily, or placebo. The main outcome measure was a diagnosis of Alzheimer's after randomization.

Patients with possible cognitive syndromes were identified with the Cognitive Assessment Battery, and returned for a dementia evaluation conducted by physicians, trained study nurses, and psychometrists, who conducted physical, neurologic, and mental status exams, and a detailed psychometric assessment battery.

The authors included in their analysis all events that occurred up to July 17, 2005, six months after the study was halted.

They found that "neither treatment was associated with reduction in the incidence of Alzheimer's disease, all cause-dementia, or the Alzheimer's disease prodromes. Instead, there was a suggestion of increase in the incidence of Alzheimer's disease with both treatments."

The hazard ratio for an on-treatment event for patients on celecoxib compared with placebo was 1.99 (95% confidence interval, 0.80 to 4.97, P=0.14). The hazard ratio vs. placebo for patients on naproxen was 2.35 (95% CI, 0.95 to 5.77, P=0.06).

Because of imperfect screening measures, seven patients with dementia, and 46 others with milder cognitive syndromes had been erroneously enrolled in the study, the authors acknowledged. When the patients with dementia or any cognitive syndrome were excluded from the analysis, the hazard ratios for both treatments vs. placebo increased. For celecoxib, the hazard ratio for Alzheimer's bumped up to 4.11 (95% CI, 1.30 to 13.0, P=0.02) and for naproxen went up to 3.57 (95% CI, 1.09 to 11.7, P=0.04).

"It is probably noteworthy that the effects of the two treatments seem broadly interchangeable," the authors wrote. "This fact would seem to reduce the probability that the observed hazard ratios are attributable purely to chance.

Furthermore, the findings trend in the same direction as the results of a recent trial of the selective Cox-2 inhibitor rofecoxib in patients with mild cognitive impairment, and also the results of a widely cited trial of naproxen and rofecoxib for treatment of Alzheimer dementia that showed trends toward negative effects of rofecoxib."

The investigators noted that they will need to continue masked follow-up of the patients for an indefinite period, but for the moment recommend against the use of either naproxen or celecoxib for primary prevention of Alzheimer's disease.

"Because the observational data suggest that NSAIDs may have protective effects for individuals with 'healthier' brains (i.e., for those whose onset of Alzheimer's disease would be some years in the future), continued observations in the ADAPT cohort could show mitigation -- or even reversal -- of the treatment effects that presently appear null or negative," they wrote.