Response to McGeer et al.: Alzheimer’s Disease Can Be Spared by Nonsteroidal Anti-Inflammatory Drugs

23 April 2018

In their recent publication, McGeer and colleagues [1] recommend a simple diagnostic test at age 55 to identify people who, as they elsewhere describe, are “fated to develop Alzheimer’s” [2], further suggesting that they could benefit from a daily dose of a generic anti-inflammatory drug for prevention. However, their paper presents no real data (positive and negative predictive values) on how the test would perform if used clinically. We argue that the evidence provided for their diagnostic test in their recent [1] and previous [3] articles is insufficient to support their recommendations on widespread diagnosis and prevention.

The idea that anti-inflammatory treatment may slow or prevent Alzheimer’s disease has been around for a long time, and despite promising observational study data, no randomized clinical trial data has borne out the promise [4]. NSAIDs, especially with long term use, have significant side effects, including increased cardiovascular risk [5] (which itself a risk factor for Alzheimer’s disease [6]), liver and kidney damage, and gastrointestinal bleeding [7]. It is important to conclude accurately based on the results, as inappropriate conclusions can often lead to confusion and misrepresentation of science. As NSAIDs are available over-the-counter and no diagnostic tests are readily available to the public, this warrants due caution in reporting findings.

It appears as though the objective of this publication is to recommend widespread use of the diagnostic saliva test developed, and patented [8], by their company. Yet, they provide no information on the validity, sensitivity, or specificity of the test in their publications or patent filings describing the test [1,3,8]. In a previous publication, the authors describe their saliva test as “diagnostic,” saying:

“the high salivary Aβ42 levels confirm that the risk in these cases is definite rather than possible. Overall, the data demonstrate that AD can be diagnosed as well as predicted on the basis of salivary Aβ42 levels. Values in the range of 20 pg/ml indicate no risk, while values over 40 pg/ml indicate AD or pre-AD” [3]

This assertion was made on the basis of a study sampling saliva concentrations of Aβ42 from 10 cases and 27 controls [3]. Most of the controls members were markedly younger than the cases, and there was no follow-up to see whether they ended up developing AD [3]. As such, it is highly presumptive to claim that there is “no risk” in those showing low salivary levels of Aβ42, and “definite risk” in those with high levels. Especially since the cases presented in McGeer’s 2018 study found 6 controls with high levels of Aβ42, but that did not present with AD [1]. Again, the authors provide no mention of follow-up for confirmation of whether or not the subjects developed AD. Moreover, the authors claim that normal controls “at risk for AD secrete [saliva Aβ42] levels comparable to AD cases” [1], but they fail to provide any explanation as to how they concluded these patients were at high risk. Of further importance is that 13 of the 25 non-AD controls were below 55 years old (the typical age of onset) and without following these controls to determine whether they develop Alzheimer’s, the assertion that they “are not at risk for AD” [1] is highly speculative. Both of these studies [1,3] have considerable limitations: the directionality thwarts any certainty as to whether biomarkers definitively precede clinical presentation; the small sample size limits quality of evidence; and, no other information was provided on the patients, how they were recruited, or how AD cases were diagnosed.

A crucial question in Alzheimer’s research is whether the focus should be on preventing pathological changes in the brain, or preventing cognitive impairment [6]. Clinical significance, argues Solomon et al., should be the priority, considering the uncertainties in the literature on the association between biomarkers and clinical presentation [6]. The relationship is poorly understood, but Solomon et al. argue that, based on clinicopathological studies, not all those exhibiting pathological brain changes will develop overt cognitive impairment [6].

With the potential risks of long-term daily NSAID use, and the psychological suffering associated with such a diagnosis, any diagnostic test for widespread use must have clearly defined predictive values. Similarly, treatments would need to have careful weighting of evidence of associated benefits and harms. Until this information can be provided by higher quality prospective studies, these recommendations—for widespread application of this test and subsequent preventive therapy with over-the-counter drugs—are exceedingly premature, and potentially dangerous.

Lance G. Shaver, MPH(c)1*, Susan Molchan, MD, MA2, Ahmed Khawer, MPH(c)1

1Division of Community Health and Humanities, Faculty of Medicine, Memorial University of Newfoundland, St. John’s, NL, Canada
2Department of Psychiatry, Uniformed Services University of the Health Sciences, Bethesda, MD, USA

*Correspondence to: Lance G. Shaver, MPH, Candidate and Graduate Research Assistant in the Division of Community Health and Humanities, Faculty of Medicine, Memorial University of Newfoundland, St. John’s, NL A1B 3V6, Canada. Tel: +1 250 589 9035; E-mail:

[1] McGeer PL, Guo JP, Lee M, Kennedy K, Mcgeer EG (2018) Alzheimer’s disease can be spared by nonsteroidal anti-inflammatory drugs. J Alzheimers Dis 62, 1219–1222.
[2] IOS Press (2018) Neuroscientists say daily ibuprofen can prevent Alzheimer’s disease. ScienceDaily,, Posted 26 March 2018, Accessed 30 March 2018.
[3] Lee M, Guo J-P, Kennedy K, McGeer EG, McGeer PL (2017) A method for diagnosing Alzheimer’s disease based on salivary amyloid-β protein 42 levels. J Alzheimers Dis 55, 1175-1182.
[4] Wang J, Tan L, Wang H, Tan C, Meng X, Wang C, Tang S, Yu J (2015) Anti-inflammatory drugs and risk of Alzheimer’s disease: an updated systematic review and meta-analysis. J Alzheimers Dis 44, 385–396.
[5] Varga Z, Sabzwari S, Vargova V (2017) Cardiovascular risk of nonsteroidal anti-inflammatory drugs: an under-recognized public health issue. Cureus 9, e1144.
[6] Solomon A, Mangialasche F, Richard E, Andrieu S, Bennett DA, Breteler M, Fratiglioni L, Hooshmand B, Khachaturian AS, Schneider LS, Skoog I, Kivipelto M (2014) Advances in the prevention of Alzheimer’s disease and dementia. J Intern Med 275, 229–250.
[7] Moore N, Pollack C, Butkerait P (2015) Adverse drug reactions and drug-drug interactions with over-the-counter NSAIDs. Ther Clin Risk Manag 11, 1061–1075.
[8] Aurin Biotech Inc, McGeer PL, Moonhee L, Guo J-P, Kennedy K, McGeer EG (2018) Salivary Abeta42 levels as prognostic indicators for Alzheimer’s disease. Patent No. WO2018018138.


Submitted by Patrick McGeer on

Shaver and colleagues assert that our recommendations for prevention of Alzheimer disease (AD) by widespread use of ibuprofen are exceedingly premature and potentially dangerous. We consider our recommendations not to be exceedingly premature, and we further consider that there is no basis for Shaver et al.’s claim that widespread use of ibuprofen is potentially dangerous. Shaver et al. attempt to justify their assertion by drawing attention to such long term side effects of high concentrations of NSAID use such as increased cardiovascular risk, liver damage, and gastrointestinal bleeding. They do not appreciate that these rare occurences stem from long term administration of ibuprofen in the dose range of 2400-3200 mg/day. Such high doses are recommended for therapy in some severe chronic inflammatory disorders. They are far outside of the common usage of less than 600 mg/day of ibuprofen for mild pain relief. It is also 12-16 fold higher than the 200 mg/day shown to be effective in returning saliva levels to normal.

NSAIDs are non-prescription agents because of their well established safety in the overwhelming majority of circumstances. Restricting their use because of the extreme situations described by Shaver et al. has properly been dismissed by regulators world wide.

Since ibuprofen and other NSAIDS are non-prescription drugs, those fated to develop AD can follow the advice of Shaver et al. and become demented, or they can follow our advice to prevent this from happening. The truth of a prediction can only be validated by its outcome over time. In the case of an age related disorder such as AD, this can take decades.

Shaver et al. acknowledge that the concept that anti-inflammatory treatment may slow or prevent AD has been around for a long time. They express regret that no randomized clinical trial has born out the promise of this concept. They apparently do not realize that it would be completely impractical to carry out such a trial since it would take decades to complete. Moreover, it would condemn the non-ibuprofen control group to the usual risk of AD.

A control trial should be deemed unnecessary since the outcome can be predicted from epidemiological studies. More than 20 of these have already been carried out. They include case control, randomized, and population based studies. The first eighteen were reported by McGeer et al. in 1996 [1] and subsequent ones by Wang et al. in 2015. The overwhelming preponderance of these conclude that a high percentage of arthritics, or others suffering from other chronic inflammatory disorders, are spared from AD by consuming NSAIDs. A caveat has been that NSAID consumption must be started well before the clinical onset of AD. Failures in studies to show a sparing effect have been characterized by this fundamental criterion not being met. The necessity of this criterion has been established by biomarker studies which show that AD onset occurs at least a decade before clinical signs become detectable.

Shaver et al. object to our reporting any of the data we have accumulated on the grounds that there was no follow-up to see whether cases with high Abeta42 ended up developing AD. Such a follow-up will only be possible several decades from today. They also object to controls not matching precisely the ages of AD cases. Controls ranging in age from 12-92 years were deliberately chosen to show that Abeta42 does not vary with age. The genes responsible for generating this protein do not change their rate of production over time. Those at risk for AD fall into the category of those who have a continuously high rate of Abeta42 production. We have shown that such cases have an Abeta42 level more than double that of cases deemed not to be at risk [2].

Individuals for this study were not recruited. They were volunteers, which explains their random nature. To summarize, individuals at risk for AD can follow the advice of Shaver et al. and become demented. Alternatively, they can follow our advice of consuming a minimal daily dose of ibuprofen. Decades from now they can determine whether our prediction that they will never develop AD will come true in their lifetime.

[1] McGeer PL, Schulzer M, McGeer EG (1996) Arthritis and antiinflammatory agents as negative risk factors for Alzheimer disease: A review of seventeen epidemiological studies. Neurology 47, 425-432.
[2] McGeer PL, Guo JP, Lee M, Kennedy K, McGeer EG (2018) Alzheimer’s disease can be spared by Nonsteroidal anti-inflammatory drugs. J Alzheimers Dis 62, 1219-1222.