Could This Neuroleptic Ward Off RA?

Patients taking haloperidol were significantly less likely to develop new-onset rheumatoid arthritis (RA), relative to those receiving other antipsychotic medications, researchers found.

Data pooled from the VA health system and two systems that record commercial insurance claims indicated that individuals receiving haloperidol for either schizophrenia or Tourette syndrome had new RA diagnoses 31% less often than such patients taking competing drugs (adjusted HR 0.69, 95% CI 0.61-0.79, P=0.0001), according to Joseph Magagnoli, MS, of the University of South Carolina's College of Pharmacy in Columbia, and colleagues.

"These findings suggest a potential benefit of haloperidol in RA and provide a rationale for randomized controlled trials to provide causal insights," the group wrote in a brief manuscript published online at medRxiv, which has not yet been peer-reviewed.

The study was prompted by two earlier publications: one, appearing more than 40 years ago, reported a case series in which the drug appeared to benefit patients with established RA; the other, from 2021 and similar to the current study, examined Japanese insurance claims and found an association between use of haloperidol and incident RA. The latter's authors cited previous research indicating that antipsychotic drugs affect cytokine expression as the jumping-off point for their investigation.

Magagnoli and colleagues sought to confirm the association in U.S. patients. They drew on data from the VA (2001-2021), PearlDiver Mariner (2010-2021), and IBM MarketScan (2006-2020), looking for patients who had received antipsychotics for schizophrenia or Tourette syndrome. Collectively, the authors said, these databases cover most U.S. adults. To be included in the study, these patients had to have at least 6 months of follow-up and no record of RA at the time they were diagnosed with the neuropsychiatric conditions.

The researchers employed meta-analytic techniques in pooling the three systems' data (and hence they called their study a meta-analysis). They also calculated propensity scores for included patients to assemble cohorts with similar characteristics to minimize potential biases. No indication was given in the manuscript of how many patients were included, though given the 15- to 20-year time span and the substantial prevalence of antipsychotic-treated schizophrenia, it was likely in the hundreds of thousands, if not millions.

Kaplan-Meier curves showing rates of incident RA over time indicated that the reduction with haloperidol was already apparent within 5 years of first dosing, and was especially prominent in the VA data, with an adjusted hazard ratio of 0.52 (95% CI 0.27-0.98). But the other two databases also showed significant differences between haloperidol and comparator drugs.

Given the magnitude of benefit suggested in the study, Magagnoli and colleagues estimated that a prospective clinical trial would "have a >95% probability of identifying a protective effect."

Limitations to the study were the usual ones for this type of analysis: reliance on administrative data and lack of information on potentially significant biasing factors.

Comment