ANTI–INFLAMMATORY DRUGS MAY PREVENT ALZHEIMER'S DISEASE—BUT NOT BY REDUCING INFLAMMATION

Although use of aspirin and nonsteroidal anti-inflammatory drugs (NSAIDs) is associated with a decreased risk of Alzheimer's disease, the low doses taken by most patients suggests that the medications exert their preventive effects through a mechanism other than reducing inflammation, according to a report published in the November Archives of Neurology.

Previous studies have found evidence that anti-inflammatory drugs are not only associated with lower rates of Alzheimer's disease but may also improve cognitive performance and slow cognitive decline in patients with the disease. However, the studies have largely ignored whether low doses of these medications have similar benefits. In the new study, researchers looked not only at the impact of anti-inflammatory drug dose on dementia risk but also whether more than 40 other medications or drug classes are associated with the disease. The investigators reviewed current medication usage, including both prescription and over-the- counter drugs, among 536 community residents age 75 or older (mean, 81) who had participated in the Sydney Older Persons Study. Cases of dementia were identified through detailed clinical interviews, which revealed that 78 patients had probable or possible Alzheimer's disease (but not vascular dementia), 45 had both Alzheimer's disease and vascular dementia, 40 had dementias other than Alzheimer's disease or vascular dementia, and 373 were dementia free.

Of the 19 medications or drug classes used by at least 50 subjects in the total sample, 15—including antacids, benzodiazepines, ß-blockers, laxatives, and vitamins—were used at similar rates by both controls and persons with Alzheimer's disease. In contrast, use of aspirin was considerably higher among controls (28%) than among subjects with Alzheimer's disease (16%); controls were also significantly more likely to be using NSAIDs (23% vs 8%). Surprisingly, two other medication classes, angiotensin-converting enzyme inhibitors and "other antirheumatics," were also used less frequently by subjects with Alzheimer's disease. Use of an analgesic without anti-inflammatory properties, acetaminophen, was identical between the two groups.

The fact that most other medications were used at similar rates in all four patient groups suggests that the differences observed with anti-inflammatory drugs were not due to underreporting or to problems in taking medication in general, the researchers noted. In addition, use of aspirin and NSAIDs did not differ between controls and subjects with vascular or other types of dementia, suggesting a specific effect with Alzheimer's disease.

The study's major finding, however, was that the dose of anti-inflammatory drug used was unrelated to the apparent protective effects; indeed, most of the subjects were taking rather low doses of these medications, the researchers noted. Among the 139 aspirin users, none were taking a high dose and more than 80% were taking 175 mg/d or less. Similarly, only 16 of 111 NSAID users (14%) were taking a high dose (naproxen dose equivalent > 1000 mg/d) and more were taking low doses (< 500 mg/d) than medium doses. "This suggests that effective protection [against Alzheimer's disease] for elderly persons may be gained at these low doses," the researchers noted. Because such doses are unlikely to suppress primary inflammatory processes in the brain, they continued, it is likely that other properties of these drugs are responsible for any protective effects. One possible mechanism involves cyclooxygenase, an enzyme that is inhibited by anti-inflammatory medications. Cyclooxygenase inhibition reduces platelet aggregation and in turn may reduce levels of circulating ß-amyloid produced by platelets. In addition, because cyclooxygenase converts oxygen to superoxide radicals, inhibiting the enzyme may limit oxidative damage in endothelial cells.

NOTHING STANDARD ABOUT MEASURES OF ANTIEPILEPTIC DRUGS

Inappropriate or ineffective tests to measure quality of life and behavioral measures in randomized controlled trials of antiepileptic drugs in patients with epilepsy are the norm rather than the exception, according to a study in the November Epilepsia.The authors found a consistent failure to apply behavioral or quality-of-life outcome measures in a systematic way—making any comparisons across the antiepileptic drugs difficult—and determined that in many of the trials they reviewed, not enough consideration had been given to the choice of measures used or their method of application.

The study, conducted at The Walton Centre, Department of Neurological Science, Liverpool, UK, was based on a review of a database of 44 randomized, controlled antiepileptic drug trials dating from 1966 to 1999.

Fifty-four different measures were used to assess the quality-of-life and behavioral outcomes of antiepileptic drugs in the 44 trials. The Profile of Mood States, the Minnesota Multiphasic Personality Inventory, and the Washington Psychosocial Seizure Inventory were used the most frequently, it was found.

Of the studies with active controls, four involved add-on therapy and 13 were monotherapy. And of these 17 reports, only four described the method of randomization and only three stated that an intent-to-treat analysis was conducted. Of the 25 placebo-controlled trials, 20 were add-on studies and five were monotherapy. Again, the method of randomization was described in only five studies and intent-to-treat was discussed in only seven.

The researchers also determined that only one of the 51 studies documented a priori hypotheses about behavioral outcomes, only one documented a rationale for selection of measures, only one provided evidence of reliability and validity in an epilepsy population, none performed a power calculation for behavioral change, and only two of the 51 provided a clinical interpretation of the results. Most studies were extremely poor in detailing this important information, the researchers reported.

Significant differences between two treatments were detected in only 16 of the 44 trials. Application of the Minnesota Multiphasic Personality Inventory showed a difference in three of the five trials, although the results of the Washington Psychosocial Seizure Inventory were not significant in any of the four trials in which it was used. A number of other measures that were able to demonstrate a significant difference were not applied to any later trials. Trials comparing an active drug with placebo were typically those reporting a significant change in behavioral outcomes.

The researchers urge that the selection of tests for use in clinical trials be made on the following grounds: the aims of the study, the rationale for selecting a particular measure, the psychometric properties of the scale, its previous application to epilepsy, and its sensitivity to change. They further recommend an international consensus on the use of behavioral and quality-of-life measures in clinical research and practice, and the establishment of recommendations for minimum standards for the use of quality-of-life and behavioral outcomes in clinical trials. They also stressed the need for uniformity in the selection of quality-of-life and behavioral outcomes—with evidence of reliability, validity, and sensitivity to change. And, of course, reporting of the outcomes of quality-of-life and behavioral assessments in randomized, controlled trials should be standardized.

Baker GA, Hesdon B, Marson AG. Quality-of-life and behavioral outcome measures in randomized controlled trials of antiepileptic drugs: a systematic review of methodology and reporting standards. Epilepsia.2000;41:1357-1363.

A NEW EPILEPSY SYNDROME?

While less common than benign childhood epilepsy with centrotemporal spikes, Panayiotopoulos-type benign childhood occipital epilepsy is a distinct syndrome that is well defined and recognizable by clinical and EEG features, according to a study in the October 24 Neurology.

Panayiotopoulos-type benign childhood occipital epilepsy is currently being considered as a classification of epilepsy by the International League Against Epilepsy. Although it is remarkably similar to rolandic epilepsy, there are markedly clear delineations in clinical (a clustering of unusual ictal manifestations) and EEG (occipital functional spikes) features. Ictal vomiting and deviation of the eyes, impaired consciousness, and progression to convulsions characterize the seizures, which are infrequent, often solitary, and often nocturnal. The peak age at onset is 5 years, said the authors. The prognosis is benign; the seizures usually remit within one to two years. (While one third of patients have one seizure and one half have between two and five seizures, between 5% and 10% may have more than 20 seizures. Some seizures occur intermittently for as long as five years, noted the authors.) According to the authors, the syndrome may be misdiagnosed as encephalitis, other serious brain insults, or migraine.

The seven-year prospective study was conducted by Roberto Caraballo, MD, and colleagues at the Hospital Nacional de Pediatría "Juan P. Garrahan" in Buenos Aires. The pediatric neurology unit serves a population of 1,200,000 children. While 145 children met the criteria for rolandic epilepsy, 66 children (38 boys) met the criteria for Panayiotopoulos-type benign childhood occipital epilepsy. Thus, the prevalence of Panayiotopoulos-type benign childhood occipital epilepsy was nearly half that of rolandic epilepsy. However, although the known prevalence of rolandic epilepsy is 15%, the prevalence of Panayiotopoulos-type benign childhood occipital epilepsy is difficult to estimate, as it has not yet been included in prevalence studies, said Dr. Caraballo and colleagues.

It is unclear how common this syndrome is, said Anne T. Berg, PhD, of Northern Illinois University, DeKalb, and C.P. Panayiotopoulos, MD, of St. Thomas' Hospital, London, UK, in an accompanying editorial. They agreed with Dr. Caraballo and colleagues that it is difficult to estimate the prevalence of the syndrome from the findings of the Argentine study. Although that study would suggest a prevalence of 5% or 6% in children with epilepsy, "referral center populations often over-represent difficult cases and under-represent easily diagnosed, well-recognized benign forms of a disorder." They added that rolandic epilepsy "occurs in roughly 10% of childhood onset epilepsy but represented only 0.5% of the reported ~13,000 children seen at the authors' center during their study period." More accurate estimates of prevalence will be facilitated by the pending inclusion of Panayiotopoulos-type benign childhood occipital epilepsy in the International League Against Epilepsy classification, they said. "With acceptance into the ILAE classification and with rich clinical descriptions such as that provided by Caraballo, this form of benign focal epilepsy of childhood may become more readily recognized," they concluded.

Berg AT, Panayiotopoulos CP. Diversity in epilepsy and a newly recognized benign childhood syndrome. Neurology.2000;55:1073-1074.

Caraballo R, Cersosimo R, Medina C, Fejerman N. Panayiotopoulos-type benign childhood occipital epilepsy: a prospective study. Neurology.2000;55:1096-1100.