The short answer is: possibly, but the evidence is more complicated than a simple yes or no. Anti-inflammatory drugs — particularly a class known as NSAIDs, or nonsteroidal anti-inflammatory drugs — have shown real promise in reducing Alzheimer’s risk in large observational studies, including a March 2025 analysis of nearly 12,000 participants that found long-term NSAID users had a 12% lower risk of developing dementia. That is a meaningful finding. But when researchers have tried to replicate this in controlled trials — giving NSAIDs to at-risk individuals and tracking outcomes — the results have been disappointing, and in some cases the drugs appeared to do no good at all.
This disconnect between what population data suggests and what trials confirm is the central puzzle in this area of research. Someone who has taken ibuprofen for arthritis for the past decade might wonder whether they are inadvertently protecting their brain. Someone with a family history of Alzheimer’s might ask their doctor about starting an anti-inflammatory regimen. This article covers what the science actually shows, why the observational and trial data diverge, which populations may or may not benefit, the very real risks of long-term NSAID use, and what emerging research — including a December 2025 study pairing a cannabis extract with celecoxib — is beginning to reveal about the future of this approach.
Table of Contents
- What Does the Research Say About Anti-Inflammatory Drugs and Alzheimer’s Risk?
- Why Have Clinical Trials Failed to Confirm What Observational Studies Found?
- Which Anti-Inflammatory Drugs Have Been Studied, and Are They All the Same?
- What Are the Real Risks of Long-Term NSAID Use?
- The Timing Problem — Why When You Start May Matter More Than Whether You Start
- Emerging Research — Cannabis Extract Combined With Celecoxib
- Where the Research Is Heading
- Conclusion
- Frequently Asked Questions
What Does the Research Say About Anti-Inflammatory Drugs and Alzheimer’s Risk?
The hypothesis that inflammation drives Alzheimer’s disease is not new. For decades, researchers have noted that the brains of Alzheimer’s patients show signs of chronic neuroinflammation — activated microglia, elevated cytokines, and inflammatory markers clustering around amyloid plaques. The reasoning followed naturally: if inflammation is part of the problem, perhaps anti-inflammatory drugs could be part of the solution. The observational evidence has been, at times, striking. In a study published in the New England Journal of Medicine, subjects who took NSAIDs for a cumulative period of 24 or more months showed significantly reduced Alzheimer’s risk. Earlier population studies found that people who used ibuprofen for more than five years were over 40% less likely to develop Alzheimer’s disease.
A separate analysis found that consistent NSAID users over five or more years showed roughly a 25% lower risk compared to non-users. These are substantial reductions, and they fueled considerable optimism in the research community through the late 1990s and early 2000s. The most recent and comprehensive data comes from the Rotterdam Study, a large Dutch population cohort. A March 2025 paper in the Journal of the American Geriatrics Society analyzed 11,745 dementia-free participants with a mean age of 66.2 years. Long-term NSAID users showed a hazard ratio of 0.88 — translating to a 12% lower risk of developing dementia. Notably, the duration of use mattered more than the cumulative dose. How long someone had been taking NSAIDs was a stronger predictor of lower risk than how much they had taken in total, suggesting that sustained, consistent exposure is the relevant variable.
Why Have Clinical Trials Failed to Confirm What Observational Studies Found?
Despite the promising observational data, randomized controlled trials — the gold standard of medical evidence — have largely told a different story. The INTREPAD trial enrolled asymptomatic, high-risk individuals in their 60s and gave them daily naproxen for two years. The result was not just neutral; the primary outcome may have worsened in the treatment group. A trial using 800mg of ibuprofen found no overall cognitive benefit. Celecoxib, a selective COX-2 inhibitor, showed no impact on dementia risk in trials. The enthusiasm that had built up through the 1990s largely collapsed when the controlled trial results started coming in. The most commonly cited explanation for this discrepancy involves timing. The observational studies captured people who had been using NSAIDs for years — often beginning well before any cognitive symptoms appeared.
The trials, by contrast, enrolled people who were already identified as high-risk and were likely in the early stages of pathological change. The leading theory is that neuroinflammation may play a harmful role in the earliest, pre-symptomatic stages of Alzheimer’s development, but by the time someone is already on a clinical radar for cognitive risk, the inflammatory processes may have already set damage in motion that NSAIDs cannot reverse. This “timing hypothesis” remains unproven but is the most plausible reconciliation of the conflicting data. However, there is one important exception worth noting. The ibuprofen trial that otherwise showed no benefit did find a significant protective effect in a subgroup of participants carrying the ApoE4 gene variant — the most well-established genetic risk factor for late-onset Alzheimer’s. This raises the possibility that genetic background may substantially influence who benefits from anti-inflammatory treatment. A person without ApoE4 may simply not respond to NSAIDs the same way a carrier does. This has not been validated in a dedicated trial, but it points toward a more personalized approach to the question rather than a blanket recommendation.
Which Anti-Inflammatory Drugs Have Been Studied, and Are They All the Same?
Not all NSAIDs are equivalent in this context, and the research has not treated them as such. Ibuprofen and naproxen are the most widely studied, given their prevalence as over-the-counter medications. Indomethacin, a prescription NSAID, has shown more promising results in some laboratory and early clinical work than the more common alternatives. Celecoxib, a COX-2 selective inhibitor that was once thought to have advantages due to its more targeted mechanism, showed no impact on dementia risk in trials and has generally been a disappointment in this line of research. The distinction between COX-1 and COX-2 inhibitors is worth understanding here. NSAIDs generally work by blocking cyclooxygenase enzymes — COX-1 and COX-2 — which are involved in the production of prostaglandins, molecules that promote inflammation and pain.
Some researchers hypothesized that COX-2 selective drugs like celecoxib would be more effective at targeting the brain-specific inflammatory pathways implicated in Alzheimer’s, while also being gentler on the stomach than traditional NSAIDs. Neither of those advantages turned out to be decisive in practice. The more promising signals have generally come from non-selective NSAIDs taken over long periods, not from the more targeted agents. Indomethacin stands somewhat apart from the others. A small but notable early trial of indomethacin in mild-to-moderate Alzheimer’s patients showed it could slow cognitive decline compared to placebo — a result that has not been systematically followed up with larger trials, partly because of indomethacin’s side effect profile and partly because the research momentum shifted toward other drug classes. That early finding still occasionally surfaces in discussions about which specific NSAID, if any, might eventually prove useful in a clinical context.
What Are the Real Risks of Long-Term NSAID Use?
Any discussion of NSAIDs as a potential Alzheimer’s preventive strategy has to be grounded in a frank accounting of what long-term use actually does to the body. These are not benign supplements. The two most serious concerns with extended NSAID use are gastrointestinal bleeding and cardiovascular problems — both of which become substantially more significant with age, precisely the demographic that would be considering this approach for dementia prevention. Gastrointestinal bleeding occurs because NSAIDs inhibit COX-1, which plays a protective role in the stomach lining. Regular use erodes that protection over time, increasing the risk of ulcers and bleeding events. The risk is elevated further by combination with blood thinners, corticosteroids, or simply by age.
On the cardiovascular side, certain NSAIDs — particularly the COX-2 selective drugs — have been associated with increased risk of heart attack and stroke with prolonged use. Naproxen has a somewhat more favorable cardiovascular profile than ibuprofen at high doses, but no NSAID is without risk in this category for older adults. The 2025 Rotterdam Study authors were explicit about this tradeoff. Despite their finding of a 12% reduction in dementia risk among long-term NSAID users, they specifically recommended that no one begin long-term NSAID therapy for the purpose of dementia prevention without consulting a physician. That 12% benefit needs to be weighed against the individual’s baseline risk for gastrointestinal and cardiovascular complications — a calculation that looks very different for a healthy 60-year-old with no GI history than for someone already on anticoagulants with a history of peptic ulcer disease. The balance simply does not favor casual self-medication with ibuprofen in hopes of protecting the brain.
The Timing Problem — Why When You Start May Matter More Than Whether You Start
The timing hypothesis deserves its own examination because it fundamentally reshapes how to interpret both the positive and negative findings in this field. Alzheimer’s disease does not begin when symptoms appear. Pathological changes — amyloid accumulation, tau tangles, synaptic loss — begin years or even decades before the first memory complaint. By the time someone receives a clinical diagnosis, or even before that when they are identified as “high risk,” the biological cascade is already well underway. If neuroinflammation is most destructive during the early, silent phase of disease — when amyloid is just beginning to accumulate and the immune response is first being recruited — then intervening with an anti-inflammatory during that window might genuinely slow or redirect the process.
But that same intervention, applied years later when the pathology is established, might arrive too late to help and could even disrupt compensatory mechanisms the brain has already developed. This is the most credible explanation for why people who happened to take NSAIDs for arthritis in their 50s showed reduced Alzheimer’s risk in their 70s, while people given NSAIDs specifically because they were already considered at high risk in their 60s showed no benefit. The warning implied by this hypothesis is important: the observational data should not be read as license to start an NSAID regimen at 65 and expect protection. The protection seen in observational studies may have accrued to people who began NSAID use much earlier, for entirely unrelated reasons, and whose inflammatory pathways were modulated during a critical pre-symptomatic window that had already closed by the time the clinical trials enrolled their participants. Replicating that exposure intentionally — and safely — would require either very early biomarker-based identification of at-risk individuals or a fundamentally different drug strategy.
Emerging Research — Cannabis Extract Combined With Celecoxib
A study published in December 2025 in the journal Aging and Disease offered an unexpected direction for this line of research. Researchers at UT Health San Antonio found that combining a low-dose THC extract with celecoxib — the COX-2 selective NSAID that had previously shown little promise on its own — produced a shift in gene expression patterns that moved markers related to synaptic function, inflammation, and Alzheimer’s risk back toward healthier profiles. The combination, rather than either agent alone, appeared to be the active element.
This is very early-stage work, and the jump from gene expression profiles to clinical outcomes in humans involves many intervening steps that may not all succeed. But it adds a layer of nuance to the celecoxib story: a drug that appeared inert in isolation may have properties that become relevant in combination. It also points toward a broader shift in anti-inflammatory research for Alzheimer’s — away from single-agent, long-term suppression strategies and toward more targeted combinatorial approaches that address multiple pathways simultaneously.
Where the Research Is Heading
The field has not abandoned the inflammation hypothesis — if anything, it has become more sophisticated about it. The focus has shifted toward understanding which specific inflammatory pathways are harmful versus protective at different stages of the disease, toward genetic stratification that might identify who actually benefits from anti-inflammatory intervention, and toward novel drug targets beyond the standard COX-1/COX-2 inhibitor framework. Researchers are increasingly interested in the role of the NLRP3 inflammasome, microglial activation states, and the complement system as more specific targets within the neuroinflammatory cascade.
These pathways are not well addressed by conventional NSAIDs. Future clinical trials are likely to test more targeted anti-inflammatory agents in earlier, biomarker-defined populations — attempting to finally answer the timing question with the right tool in the right people at the right time. Whether that effort will eventually vindicate the original hypothesis remains genuinely uncertain, but the scientific rationale remains sound enough to sustain serious investigation.
Conclusion
The evidence on anti-inflammatory drugs and Alzheimer’s risk is real but unresolved. Large observational studies, including the 2025 Rotterdam Study, consistently show that long-term NSAID users develop dementia at lower rates than non-users. Historical population data found risk reductions of 25 to 40 percent in people using these drugs for five or more years. These are not trivial findings. But controlled trials — the only way to establish causation — have failed to confirm them, most likely because they enrolled participants too late in the disease process and because the specific drugs and populations tested may not have matched the conditions under which observational benefit occurred.
The ApoE4 subgroup signal and the emerging combinatorial research with celecoxib suggest that the story is not finished, but neither offers a clear clinical recommendation today. For anyone caring for a person with Alzheimer’s, or concerned about their own risk, the practical message from this research is measured. Do not start taking ibuprofen or naproxen daily on the assumption it will protect your brain — the risks of long-term use are serious, and the evidence does not support that decision without medical guidance. If you are already taking an NSAID regularly for another condition, that is something worth discussing with your physician in the context of your overall health picture. The inflammation hypothesis remains one of the more biologically compelling frameworks in Alzheimer’s research, and it may yet produce actionable preventive strategies — but that work is still in progress.
Frequently Asked Questions
Should I take ibuprofen regularly to prevent Alzheimer’s?
No. While some long-term population studies have found lower Alzheimer’s risk among regular NSAID users, controlled clinical trials have not confirmed a preventive benefit. Long-term NSAID use carries significant risks including gastrointestinal bleeding and cardiovascular problems. The 2025 Rotterdam Study authors explicitly advise consulting a physician before starting any such regimen.
Which anti-inflammatory drug shows the most promise for Alzheimer’s prevention?
The evidence is mixed across drugs. Ibuprofen has the most observational support. Indomethacin showed some early clinical promise but has a difficult side effect profile. Celecoxib has generally performed poorly in trials on its own, though December 2025 research suggests it may have value in combination with other agents. No NSAID is currently recommended for this purpose.
Why do population studies show benefit but clinical trials don’t?
The leading explanation is timing. People in observational studies often began NSAID use decades before any cognitive symptoms appeared, potentially interrupting the inflammatory process during a critical early window. Clinical trials enrolled people who were already identified as high-risk, likely too late in the disease process for anti-inflammatory intervention to redirect the outcome.
Does the ApoE4 gene affect whether NSAIDs help?
Possibly. One ibuprofen trial found no overall cognitive benefit but did identify a significant protective effect in participants carrying the ApoE4 variant — the strongest known genetic risk factor for late-onset Alzheimer’s. This has not been confirmed in a dedicated trial but suggests genetics may determine who, if anyone, responds to this approach.
What are the main risks of taking NSAIDs long-term?
Gastrointestinal bleeding is the most common serious risk, caused by erosion of the stomach lining’s protective layer. Cardiovascular risks, including elevated rates of heart attack and stroke, are also well documented with prolonged use, particularly in older adults. These risks are significant enough that no researcher currently recommends long-term NSAID use specifically for dementia prevention.
Is there new research combining cannabis and anti-inflammatory drugs for Alzheimer’s?
Yes. A December 2025 study from UT Health San Antonio found that combining a low-dose THC extract with celecoxib shifted gene expression patterns related to inflammation, synaptic function, and Alzheimer’s risk toward healthier profiles. This is preliminary research and has not yet translated into clinical recommendations, but it represents an emerging direction in the field.
