A multicentre, randomised, double-blind, placebo-controlled trial of efavirenz, an antiretroviral licensed for HIV since 1998, has been registered for Creutzfeldt-Jakob disease (CJD). According to the registration posted on ClinicalTrials.gov on 19 March 2026 (NCT07482085), the sponsor, Xuanwu Hospital in Beijing, plans to enrol 246 participants across 21 centres nationwide, randomised 1:1 to efavirenz or matching placebo. The efavirenz arm is listed as 200 mg once daily at bedtime for the first week, then 400 mg once daily, continued until death or study completion. The primary endpoint is median survival time from randomisation to death; secondary endpoints are the rate of functional decline and tolerability, with monthly telephone follow-up and a data monitoring committee. Eligibility is listed as age 18 to 80 with WHO criteria for probable sporadic CJD or genetically confirmed hereditary CJD, a Medical Research Council Prion Disease Rating Scale score above 10, and an available caregiver. As of the last registry update, on 15 March 2026, the status was not yet recruiting, with an estimated start in March 2026 and estimated completion in December 2028.
The reason for testing an old HIV drug is preclinical, and it is animal and cell work. In a study published in JCI Insight on 19 June 2025, a University of Calgary group led by Sabine Gilch inoculated tg650 mice, which overexpress human prion protein, with human sporadic CJD prions, then gave low-dose efavirenz in drinking water (1.68 mg/L) starting either 30 or 130 days after inoculation. Survival was significantly longer in both treated groups (10 mice each) than in the untreated infected group (9 mice) on log-rank testing. At the early clinical stage, treated mice had less accumulation of the misfolded prion protein PrPSc, lower brain cholesterol and lipid droplet content, and higher levels of the brain enzyme CYP46A1 and its product 24S-hydroxycholesterol. Overexpressing CYP46A1 in prion-infected neuronal cells also lowered PrPSc, which the authors read as evidence that the antiprion effect tracks with enzyme activation. Those are mice and cells, not people.
The authors of that paper state the central caution themselves: drugs that work against mouse-adapted scrapie prions have repeatedly failed in human prion disease. The most direct precedent is doxycycline, which looked promising in models and in observational series and was then tested in a phase 2, randomised, double-blind, placebo-controlled trial in Italy and France. In 121 patients reported in The Lancet Neurology in 2014, doxycycline was well tolerated but did not affect survival (hazard ratio 1.1, 95% confidence interval 0.8 to 1.7, p=0.50), and the trial was stopped for futility at the first interim analysis. Sporadic CJD is also unforgiving of delay: median survival from symptom onset is about five months, and a Beijing-led trial enrolling patients who already have symptoms is starting treatment far later in the disease than the mice were treated, one of which began before any clinical signs appeared.
There is a second unresolved question, about dose. The human evidence that efavirenz engages CYP46A1 at all comes from a pilot randomised study of five people with early Alzheimer's disease, published in Alzheimer's Research and Therapy in December 2022, in which four received 50 mg or 200 mg daily and one received placebo for 20 weeks, with plasma 24S-hydroxycholesterol as the target engagement marker. The registered CJD trial uses 400 mg daily, above the doses tested in that study and below the 600 mg standard HIV dose, and the registry lists exclusions including unstable serious illness, significant QTc prolongation, interacting medicines that cannot be adjusted, and pregnancy or breastfeeding. No efficacy or safety results in people with CJD have been published for efavirenz.