某PDE5抑制剂在阿尔茨海默病中的研究结果
Erectile Dysfunction Drug Shows Mixed Results in Alzheimer's
AR1001在阿尔茨海默病患者中总体未达主要终点,但在未使用胆碱酯酶抑制剂或美金刚的患者中显著减缓了认知和功能下降。次要终点趋势向好,但需进一步研究。
研究结果显示药物在特定亚组中有效,但总体未达主要终点,为后续研究提供了方向。
— AR1001 missed the primary endpoint in one population, but slowed decline in another
by Judy George, Deputy Managing Editor, MedPage Today
November 17, 2026 • 3 min read
Mirodenafil (known as AR1001), an oral phosphodiesterase-5 inhibitor approved in South Korea to treat erectile dysfunction, showed mixed results in people with early Alzheimer's disease in the phase III POLARIS-AD trial.
The treatment missed the primary endpoint in one prespecified group but showed positive results in another, reported Sharon Sha, MD, MS, of Stanford University in Palo Alto, California, at the Clinical Trials on Alzheimer's Disease conference in Boston.
In the overall study population, AR1001 showed a nonsignificant difference of -0.11 points versus placebo in Clinical Dementia Rating-Sum of Boxes (CDR-SB) scores at 52 weeks (P=0.4015). CDR-SB scores measure cognitive and functional impairment and can range from 0 to 18.
In a population of monotherapy patients -- those not on an acetylcholinesterase inhibitor or memantine at baseline -- AR1001 demonstrated a significant 52-week treatment difference of -0.44 points versus placebo (P=0.0266), representing a 31% slowing of decline.
Secondary endpoints of cognitive and functional measures trended in favor of AR1001 in both populations, with a nominally significant difference in Mini-Mental State Examination scores emerging in the monotherapy population.
The most frequent adverse events were diarrhea, falls, and dizziness. Adverse events led to discontinuation in 5.2% of patients in the AR1001 group and 1.7% in the placebo group.
Serious adverse events were similar in the AR1001 group (10.0%) and the placebo group (9.8%). Four deaths occurred in the AR1001 group (acute myeloid leukemia, cerebrovascular accident, suicide, and ruptured cerebral aneurysm) and three deaths occurred in the placebo group.
The CDR-SB data and safety profile are encouraging, Sha noted. "POLARIS-AD made its primary efficacy endpoint in the prespecified monotherapy population," she said. "The secondary endpoints had trends in favor of AR1001."
AR1001 is a brain-penetrant phosphodiesterase-5 inhibitor that enhances cyclic guanosine monophosphate (GMP) signaling. In preclinical Alzheimer's models, the treatment improved cognitive performance and reduced amyloid and tau burdens.
"AR1001 takes a novel approach, relaxing blood vessels to increase blood flow to the brain," noted Richard Oakley, PhD, of the Alzheimer's Society in England, who wasn't involved in the study.
"AR1001 is from the same family of medicines as Viagra [sildenafil], but it is being developed specifically for Alzheimer's disease. People should not take Viagra to try to treat Alzheimer's," Oakley said in a statement. "We look forward to seeing the full results from this trial and how this drug progresses."
In 2024, a study of 270,000 men in the U.K. linked erectile dysfunction drugs with a reduced risk of Alzheimer's disease. Those findings conflicted with data from the NIH DREAM initiative, but supported an earlier study of Medicare beneficiaries that suggested that sildenafil users were less likely to develop Alzheimer's than non-users.
The POLARIS-AD trial included 1,535 amyloid-positive Alzheimer's patients with mild cognitive impairment or mild dementia in the overall population; the monotherapy subgroup had 604 patients. The average patient age in the study was 73, and just over half were women. Participants were randomized to receive AR1001 30 mg or placebo once daily for 52 weeks.
The primary endpoint was the 52-week change in CDR-SB score in the overall and monotherapy populations. The primary efficacy objective required a significant treatment difference compared with placebo at week 52 in at least one population under the prespecified multiplicity procedure. The prespecified plan was suggested by AR1001's phase II findings, Sha pointed out.
Biomarker tests showed that AR1001 increased cyclic GMP levels in cerebrospinal fluid compared with placebo at week 52 in both the overall (P<0.0001) and monotherapy populations (P=0.0058), supporting target engagement, Sha said.
The presentation included no data about sex differences in treatment effects and Sha said the company has not yet run those analyses. "These results are hot off the press as of Friday," she noted, adding that "continuing analysis is happening, and the open-label extension is ongoing."
来源:MedPage Today · medpagetoday.com