ARTICLE IN BRIEF
- ✓ New data suggest that timing, dose, and demography may prove critical in preventing dementia or Parkinson disease with nonsteroidal anti-inflammatory drugs.
Two new studies may help unravel the puzzle of conflicting data for and against the neuroprotective role of nonsteroidal anti-inflammatory drugs (NSAIDs) in dementia and Parkinson disease. Together, they suggest that timing, dose, and demography may prove critical in preventing both conditions with NSAIDs.
The dementia study found the strongest evidence to date that only people with the e4 variant of the apolipoprotein E (APOE) gene benefit from taking NSAIDs. The finding, reported in the Jan. 1 Neurology, could partly explain why randomized trials — none of which have controlled for the APOE genotype — discerned a benefit for NSAIDs.
“We have had this conundrum: observational studies have shown that NSAIDs reduce the risk of dementia, yet prevention trials show no effect or a worsening effect,” said the study's senior author, Peter P. Zandi, PhD, assistant professor in the department of mental health at the Johns Hopkins Bloomberg School of Public Health. “This paper starts to gather evidence on what might be going on.”
UNTANGLING THE DEMENTIA CONUNDRUM
As recently as last May, headlines in the lay press proclaimed that commonly used pain relievers do not prevent Alzheimer disease (AD), based on early results of the Alzheimer's Disease Anti-Inflammatory Prevention Trial (ADAPT). The results were reported in the May 22, 2007 Neurology. The randomized trial comparing naproxen and celecoxib against placebo had been cut short on Dec.17, 2004, days after another trial of celecoxib for colorectal cancer had reported an increased risk of fatal and non-fatal cardiovascular events.
But having begun enrollment of the participants in 2001, ADAPT investigators continued masked long-term follow-up. By early 2007 they had sufficient data to detect a trend toward increased risk — not of cardiovascular events, but of AD — among those who had taken the drugs before the trial's suspension, with a hazard ratio compared to placebo of 1.99 (CI 0.80 to 4.97; p=0.14) for celecoxib and 2.35 (CI 0.95 to 5.77; p = 0.06) for naproxen.
Even so, the lead investigator of the ADAPT study group emphasized in an interview with Neurology Today that those findings were only for short-term results. “Our paper was purposely titled ‘early results,’” said Constantine G. Lyketsos, MD, professor of psychiatry at the Johns Hopkins University School of Medicine and chair of psychiatry at the Johns Hopkins Bayview Medical Center. “We seem to have answered the question, ‘What if you're a few years from dementia and you go on NSAIDs, is that helpful?’ We've found it actually might be harmful. But what if you're five years away? The matter isn't closed. I don't think it's hugely likely, but it's possible that as we follow our cohort, we might find that one or both of the drugs will start showing a preventative benefit.”
The author of the Jan. 1 paper agreed that timing might be everything when it comes to the effect of NSAIDs on dementia risk. “When you take the NSAID might have a dramatic impact,” Dr. Zandi said. “If taken earlier, that's when the drugs would seem to exert their maximal benefit.”
Another reason for ADAPT's negative findings thus far, some have speculated, is that neither celecoxib nor naproxen modulates gamma secretase, an enzyme implicated in the production of beta amyloid. “People have speculated that maybe the ADAPT trial looked at the wrong NSAIDs,” Dr. Zandi said. “Maybe if ADAPT had looked at ibuprofen, they would have found a benefit.”
Dr. Zandi tested that hypothesis in the prospective observational study he led, the Cardiovascular Health Cognition Study, which included 3,229 people, aged 65 or older who were free of dementia at baseline. His group found no difference in risk reduction by NSAID type.
Besides waiting for longer-term results from ADAPT, Dr. Zandi suggested the data might be re-analyzed to see if a benefit might be distinguished among those NSAID users with the APOE e4 allele.
“They have the ability to do that,” Dr. Zandi said of the ADAPT researchers. “It's something I'd like to encourage them to do.”
In the meanwhile, neither he nor Dr. Lyketsos said they would recommend that neurologists have their at-risk patients tested for the APOE genotype and prescribed an NSAID if they're found to have the e4 allele.
“I wouldn't do it now based on this one study,” Dr. Zandi said. “If we can continue to build a case that it's specific among patients with the e4 allele, and we show it in a randomized prevention trial, we could take it to clinicians. We are not at that point yet.”
PARKINSON STUDY BUILDS A CASE
And what about the relationship of NSAIDs to Parkinson disease? The Parkinson study, reported in the Nov. 6, 2007 Neurology, detected a strong benefit for people who used non-aspirin NSAIDs regularly for two or more years, with an odds ratio of 0.44 (95 percent CI, 0.26 to 0.74) compared to non-users. Aspirin use benefited only women; regular users for at least two years had an odds ratio of 0.51 (95 percent CI, 0.26 to 1.02) compared to non-users.
Several prior studies reported an increased risk of PD among NSAID users, but they were hampered by being too small or assessing only the period immediately preceding diagnosis, said the senior study of the most recent study in Neurology, Beate Ritz, MD, PhD, professor of epidemiology and neurology in the School of Public Health at the University of California-Los Angeles.
“You need to take the NSAIDs in midlife, to slow down the process of brain cell death or prevent it,” she said. “People taking them shortly before diagnosis are doing it for the aches and pains that are often the first signs of the disease.” The only reason Dr. Ritz's team decided to ask participants in their case-control study about NSAIDs was because they suspected that use of the drugs immediately prior to diagnosis might indicate that patients were self-medicating for aches and pains associated with the disease's earliest stage.
“When we started the study in 2001, nobody thought NSAIDs or aspirin had a role in preventing Parkinson disease,” she said. “But we thought, maybe Parkinson patients have aches and pains that are predictive, and we knew that many patients early on take pain medications.”
Then came the 2003 Archives of Neurology paper analyzing data on 44,057 men in the Health Professionals Follow-up Study and 98,845 women in the Nurses' Health Study, showing a 45 percent reduced risk of developing PD among regular users of non-aspirin NSAIDs, and a non-significant trend toward benefit in those who took two or more tablets of aspirin per day. A second large cohort study published by the same group two years later in the Annals of Neurology found similar benefit, but only in ibuprofen users.
Dr. Ritz's team then analyzed the data on NSAID- and aspirin-use among the 293 incident, idiopathic PD cases and 286 age-, race-, and gender-matched controls. There was a protective effect for those who took two or more pills per week for at least one month, with the strongest effect seen in those who reported two or more years of use.
The lack of benefit for aspirin among men may prove to be an effect of dose rather than gender, Dr. Ritz said. “Men who regularly use aspirin usually take a baby aspirin for cardiovascular risk reduction,” she said. “Women may have taken it for aches and pains, at 500 mg or more.”
The lead author of the two epidemiologic papers that first identified a link between NSAIDs and PD praised the study's design. “However, it is important to point out that the evidence on this topic is not entirely consistent,” said Honglei Chen, MD, PhD, an investigator at the Aging and Neuroepidemiology Group at the National Institute of Environmental Health Sciences in Research Triangle Park, NC. “Three other case-control studies largely reported no associations. While there are a few design differences between the ‘positive’ studies and the ‘negative’ studies, more well-designed large epidemiological studies are needed to figure out this important relationship.”
Although neither Dr. Ritz nor Dr. Chen said they believe the findings thus far warrant clinical recommendations to patients, the Parkinson specialist who wrote the editorial accompanying Dr. Chen's 2003 paper said she is already recommending NSAIDs to patients at high risk of developing PD, such as those diagnosed with REM behavior disorder.
“We try to educate our patients with all the knowledge we have about a disease,” said Mya C. Schiess, MD, professor and vice chair of neurology at the University of Texas Medical Center in Houston and director of its Movement Disorders Clinic. “REM sleep behavior disorder (RBD) precedes symptomatic PD by years to decades.”
While Dr. Schiess called for more trials of those diagnosed with RBD randomized to placebo or NSAIDs, she said that in the meanwhile, “she would endorse prudent use of NSAIDs in a high-risk population. You'd want to be careful if you're dealing with someone who has renal function compromised by hypertension or diabetes. But for the most part, if the individual has shown tolerance before taking NSAIDs, it's really not that hurtful a recommendation.”
But it might be too late for patients already diagnosed with PD, Dr. Schiess said. “You need to catch it before the disease process really damages the substantia nigra and the other areas of the brain associated with motor impairment.”