Correspondence
Nonsteroidal Drugs and Alzheimer's Disease
N Engl J Med 2002; 346:1171-1173April 11, 2002
- Article
To the Editor:
The study by in 't Veld et al. (Nov. 22 issue)1 on the prevention of Alzheimer's disease with nonsteroidal antiinflammatory drugs (NSAIDs) includes no mortality data. It is entirely possible that patients who received long-term treatment with NSAIDs were dying sooner from the complications. One may then underestimate the risk of Alzheimer's disease, giving a false positive result. There are also no data on morbidity. The small group of patients receiving NSAIDs for more than 24 months was the only group with a statistically significant reduction in the risk of Alzheimer's disease. It may be inappropriate to base a positive conclusion on an analysis of only 233 patients.
Although there is some evidence of an effect of NSAIDs, the safety and cost effectiveness of such an intervention are questionable. Gastrointestinal toxicity of NSAIDs alone causes 100,000 to 400,000 hospitalizations every year in the United States, with an estimated yearly cost of $2 billion.2 Unless we develop a reliable way to predict Alzheimer's disease, it will be hard to justify a randomized, controlled, blinded trial with these potentially toxic medications.
2 ReferencesAmit Sood, M.D.
Ocean Beach Hospital, Ilwaco, WA 98624
amitricha@pol.net 1
in 't Veld BA ,Ruitenberg A ,Hofman A , et al. Nonsteroidal antiinflammatory drugs and the risk of Alzheimer's disease. N Engl J Med 2001;345:1515-1521
Full Text | Web of Science | Medline2
Smalley WE ,Griffin MR . The risks and costs of upper gastrointestinal disease attributable to NSAIDs. Gastroenterol Clin North Am 1996;25:373-396
CrossRef | Web of Science | Medline
To the Editor:
In the report by in 't Veld et al., the investigators used a sophisticated Cox model and found a significantly reduced risk of Alzheimer's disease in patients who had taken NSAIDs for a cumulative period of 24 months or more. We have serious reservations about the main outcome modeled by the Cox regression. The authors “calculated the person-time between January 1, 1991, and death, a diagnosis of dementia, or the end of the study period, whichever came first.” However, they provided no information concerning the time of censoring of competing events such as death and a diagnosis of dementia. We understand that the event of interest was dementia and that death was censored. But the statistical model may be misleading, because a patient who has died is no longer at risk for dementia. In addition, if death was due to drug toxicity, it could artificially increase the benefit attributed to NSAIDs. On the other hand, if death was not censored, death and dementia could be considered a composite outcome, but they probably do not share the same prognostic factors. Our suggestion is to use the multistates extension of the Cox model.1 We used it to explore long-term, treatment-related morbidity in patients with lymphoma and showed that the assumption of the uniformity of prognostic factors or treatment effect on different event rates such as morbidity or mortality was not valid.2
2 ReferencesNicholas Mounier, M.D.
Marc André, M.D.
Eric Lepage, M.D., Ph.D.
Centre Hospitalier Universitaire Henri Mondor, 94010 Creteil, France
nicolas.mounier@sls. ap-hop-paris. fr 1
Therneau TM, Grambsch PM. Modeling survival data: extending the Cox model. New York: Springer-Verlag, 2000:350.
2
Andre MPE ,Mounier N ,Leleu X , et al. Second cancers and late toxicities after treatment of aggressive non-Hodgkin lymphoma (NHL) with the ACVBP regimen: a GELA (Groupe d'Etude des Lymphomes de l'Adulte) cohort study on 2849 patients. Blood 2001;98:768a-768a abstract.
Web of Science
To the Editor:
In 't Veld et al. report an association between the use of NSAIDs and a reduced incidence of Alzheimer's disease. They suggest that NSAIDs may exert this effect by blockade of cyclooxygenase-1 and cyclooxygenase-2 or by activation of the peroxisome proliferator γ (PPARγ) nuclear transcription factor. However, more than 50 percent of the NSAIDs recorded in their study do not activate PPARγ (e.g., diclofenac), which underlines the importance of the suppression of cerebral prostaglandin production by NSAIDs as a possible mechanism to reduce the risk of Alzheimer's disease.
The authors state that the use of acetaminophen has not been assessed, because acetaminophen is available without a prescription. A previous study found acetaminophen to be protective against Alzheimer's disease.1 Whereas acetaminophen has only minimal effects on prostaglandin synthesis in monocytes, prostaglandin synthesis is markedly suppressed by acetaminophen in microglia.2 Therefore, reduction of cerebral prostaglandin production may have an even greater effect on Alzheimer's disease than the effect reported by the authors, since patients using acetaminophen may have been included in the group of subjects without prescriptions for NSAIDs.
2 ReferencesMichael Hüll, M.D.
Klaus Lieb, M.D.
Bernd L. Fiebich, Ph.D.
University of Freiburg Medical School, D-79104 Freiburg, Germany
michael_huell@psyallg.ukl. uni-freiburg. de 1
Breitner JC ,Welsh KA ,Helms MJ , et al. Delayed onset of Alzheimer's disease with nonsteroidal anti-inflammatory and histamine H2 blocking drugs. Neurobiol Aging 1995;16:523-530
CrossRef | Web of Science | Medline2
Fiebich BL ,Lieb K ,Hull M , et al. Effects of caffeine and paracetamol alone or in combination with acetylsalicylic acid on prostaglandin E(2) synthesis in rat microglial cells. Neuropharmacology 2000;39:2205-2213
CrossRef | Web of Science | Medline
To the Editor:
In 't Veld et al. found that those taking NSAIDs had a lower risk of Alzheimer's disease. The prevailing view has been that this effect occurs through inhibition of cyclooxygenase, reducing the inflammation associated with Alzheimer's disease.1 Weggen et al. recently found that certain NSAIDs lower the amount of amyloidogenic 42-residue β-amyloid (Aβ42) protein independently of their effects on cyclooxygenase.2 Although the effect was found at clinically relevant concentrations for ibuprofen, sulindac, and indomethacin, it was not found for naproxen. Given the large number of patients taking either ibuprofen or naproxen in the study by in 't Veld et al., it would be helpful to know whether the relative risk of Alzheimer's disease differed between these two groups.
2 ReferencesDaniel Press, M.D.
Beth Israel Deaconess Medical Center, Boston, MA 02215
dpress@caregroup.harvard. edu 1
Rogers J ,Webster S ,Lue LF , et al. Inflammation and Alzheimer's disease pathogenesis. Neurobiol Aging 1996;17:681-686
CrossRef | Web of Science | Medline2
Weggen S ,Eriksen JL ,Das P , et al. A subset of NSAIDs lower amyloidogenic Abeta42 independently of cyclooxygenase activity. Nature 2001;414:212-216
CrossRef | Web of Science | Medline
Author/Editor Response
The authors reply:
To the Editor: Both Sood and Mounier et al. point out that no mortality data were provided in our report on NSAIDs and the risk of Alzheimer's disease. They suggest that a protective effect of NSAIDs might be explained by increased mortality. We considered this possibility, but we found no association between the use of NSAIDs and mortality in the remaining 6595 elderly patients without dementia. Hence, increased mortality in our users of NSAIDs without dementia does not appear to explain our results. The risk of death from any cause was 1.07 (95 percent confidence interval, 0.82 to 1.40) in patients who had used NSAIDs for two years or more and 0.97 (95 percent confidence interval, 0.86 to 1.08) in those who had used NSAIDs for less than two years, as compared with patients who did not use NSAIDs. We disagree with Sood that a positive conclusion cannot be based on 233 patients, as long as the data are cautiously interpreted. We agree, however, that the risk of gastrointestinal toxicity is high and that the cost effectiveness of the use of NSAIDs as a therapeutic intervention would require careful study. The suggestion by Hüll et al., who studied the effect of acetaminophen on microglial effects in rats,1 is interesting. A problem is that acetaminophen is widely available over the counter in the Netherlands and misclassification of exposure may be substantial. Press refers to a study published shortly before ours that suggested that some NSAIDs (e.g., ibuprofen) may protect against Alzheimer's disease by lowering levels of amyloidogenic Aβ42 peptide, whereas other NSAIDs (e.g., naproxen) have no such effect.2 Although ibuprofen and naproxen were among the most commonly used NSAIDs in our study, the numbers were too low to justify a comparison between individual agents.
2 ReferencesBruno H.C. Stricker, M.D., Ph.D.
Albert Hofman, M.D., Ph.D.
Monique M.B. Breteler, M.D., Ph.D.
Erasmus Medical Center, 3000 DR Rotterdam, the Netherlands1
Fiebich BL ,Lieb K ,Hull M , et al. Effects of caffeine and paracetamol alone or in combination with acetylsalicylic acid on prostaglandin E(2) synthesis in rat microglial cells. Neuropharmacology 2000;39:2205-2213
CrossRef | Web of Science | Medline2
Weggen S ,Eriksen JL ,Das P , et al. A subset of NSAIDs lower amyloidogenic Abeta42 independently of cyclooxygenase activity. Nature 2001;414:212-216
CrossRef | Web of Science | Medline
