Not an Onion Study: Underpowered Analysis Of Poor Quality Data Finds Antipsychotics Actually Aren’t Linked To Increased Risk Of Death In The Elderly

The studies that led to warnings from health regulators against prescribing antipsychotics to elderly patients with dementia were biased, and there is actually no significant increase in risk of death linked to the drugs, according to a study in the Journal of the American Medical Directors Association. To arrive at these findings, University of Groningen researchers explained that they focused their analysis on only small, short-term clinical trials with data of generally “poor” quality.

According to the Dutch authors, long-term observational studies of large populations, which have consistently detected significant increases in mortality in elderly people taking antipsychotics, “might” have been subject to biases and errors. “One source of bias might be that none of the studies took severity of illness into account,” they wrote. “This is a potentially strong confounder because haloperidol and chlorpromazine are often used to treat the symptoms of delirium in terminally ill patients. These 2 drugs accounted for more than half of the conventional antipsychotics used.”

For their own study, the researchers identified 17 randomized, placebo-controlled clinical trials of antipsychotics in the elderly which met their criteria, involving 2,387 participants.

In these 17 studies, thirty-two total deaths occurred — 17 deaths in the treatment groups, and 15 in the placebo groups, yielding a mortality rate of 1.4% for the treatment groups and 1.3% for the placebo groups. “The pooled risk difference of 0.1% was not statistically significant,” the researchers wrote. They then concluded that, “The results of this meta-analysis refute the observational association between conventional antipsychotics in general, or haloperidol in particular. They do not confirm an increased mortality risk for conventional antipsychotics in elderly patients with dementia either. Except for one delirium-prevention trial, the trials that we used were available in 2008. Hence, our findings question the scientific support for the warning against conventional antipsychotics in patients with dementia that was based on the observational findings.”

A discussion of the study on Psychiatric Times was headlined, “Should We Rethink Warning Against Conventional Antipsychotics in Elderly?” A number of important factors, though, were not mentioned in the study’s abstract or Times article.

In their study the Dutch researchers acknowledged that observational studies on antipsychotics and the elderly have found that the risk of death increased between 4.2% and 7.3% during the first six months of antipsychotic use. A table in their article showed that the longest-duration clinical trial that the Dutch team examined was 4 months, and the average duration was only 6.3 weeks.

The authors also noted in their discussion section that in general the quality of the studies they examined “seemed poor” and that 15 of the 17 trials “scored unclear or high risk of bias on 4 or more of the 8 quality items that we scored.”

Eleven of the 17 trials that they reviewed did not actually even list the deaths that had occurred during the trials. The Dutch researchers managed to obtain supplementary information about the deaths that had occurred from the authors of three of those trials, and for the other eight trials the Dutch researchers wrote that the number of deaths was simply “assumed to be zero.”

Finally, due to the small size and short duration of these clinical trials, the authors acknowledged that the power of their analysis “was not sufficient to detect” a 1% increased mortality risk — which, they wrote, was the increased risk detected in an earlier, 2005 meta-analysis of clinical trials published in JAMA that originally led to the warning from health authorities against prescribing antipsychotics to elderly people with dementia. After conceding that their study was not sufficiently powered to even detect that increased risk if it were there, the authors then suggested that it was in any case “questionable” whether an increased risk of death of 1% was “clinically relevant.” They commented that, “Patients with dementia already have a poor prognosis, especially when psychosis or aggression co-occurs. The opportunity of diminishing these symptoms with an antipsychotic and thereby increasing quality of life may outweigh the small increase in risk of dying.”

Hulshof, Tessa A., Sytse U. Zuidema, Raymond W. J. G. Ostelo, and Hendrika J. Luijendijk. “The Mortality Risk of Conventional Antipsychotics in Elderly Patients: A Systematic Review and Meta-Analysis of Randomized Placebo-Controlled Trials.” Journal of the American Medical Directors Association 0, no. 0. Accessed June 14, 2015. doi:10.1016/j.jamda.2015.03.015. (Abstract)

Should We Rethink Warning Against Conventional Antipsychotics in Elderly? (Psychiatric Times, June 4, 2015)