Anti-HSV med falls flat in Alzheimer’s disease trial

Valacyclovir appears to have little benefit in the treatment of individuals with early symptomatic Alzheimer’s disease (AD) and herpes simplex virus (HSV) seropositivity, according to the results of the phase II VALAD study.

After 78 weeks, valacyclovir-treated participants showed greater cognitive worsening than those who received placebo, with scores on the 11-item AD Assessment Scale (ADAS)-Cognitive Subscale increasing by 10.86 vs 6.92, respectively, from baseline (between-group difference, 3.93, 95 percent confidence interval [CI], 1.03–6.83; p=0.01), reported primary investigator Dr Davangere Devanand from Columbia University in New York City, New York, US, and colleagues in their paper. [JAMA 2026;335:511-522]

Daily functioning, assessed using the Alzheimer's Disease Cooperative Study-Activities of Daily Living (ADCS-ADL) Scale, was similar between the valacyclovir and placebo arms (ADCS-ADL Scale score, –13.78 vs –10.16; between-group difference, −3.62, 95 percent CI, −8.16 to 0.93).

“Valacyclovir did not significantly change the characteristic brain imaging indices of amyloid, tau, and neurodegeneration in AD. These findings weaken the case for HSV as a modifiable factor affecting AD neuropathology,” Devanand and colleagues said.

No reliable biomarker

The rationale for using an anti-HSV medication comes from evidence implicating HSV in the pathogenesis of AD. [Lancet Infect Dis 2021;21:1557-1567; Adv Exp Med Biol 2023;1423:279-280; Virology 2025;607:110523; Curr Neurol Neurosci Rep 2018;18:55-69]

In an accompanying editorial, Dr David Knopman from Mayo Clinic in Rochester, Minnesota, US, described VALAD as important, “because it refutes the idea that valacyclovir (a drug known to engage the HSV-1 target at the doses used) plays any therapeutic role in symptomatic cognitive impairment due to AD.” [JAMA 2026;335:493-494]

“What is sorely lacking,” Knopman noted, “is a bioassay for pathogenic reactivated virus in the brain.” He argued that the existence of a biomarker to enrich the study population with individuals harbouring reactivated HSV-1 species would have been ideal, “because the target of treatment was presumably a pathogenic form of reactivated HSV-1.”

“The trial conducted by Devanand [and colleagues] lacked the opportunity to enrich the study group with the target pathology and, therefore, was forced to treat all patients without regard for their burden of reactivated HSV-1,” Knopman pointed out.

Until a reliable biomarker emerges, “HSV-1 remains a spectral presence and a difficult target for therapeutics,” he said.

VALAD trial

VALAD was an 18-month, double-blind, randomized clinical trial. The population comprised 120 adults (mean age 71.4 years, 55 percent female) with a clinical diagnosis of probable AD or a clinical diagnosis of mild cognitive impairment with positive biomarkers for AD, a positive serum antibody test (IgG or IgM) for HSV-1 or HSV-2, and a Mini-Mental State Examination score of 18 to 28.

The participants were randomized to receive either valacyclovir at a target dose of 4 g/day (n=60) or matching placebo (n=60). Valacyclovir was given initially at 2 g/day and then increased to 3 g/day at week 2 and to 4 g/day at week 4 and the remainder of the study.

The primary outcome was 78-week change in the ADAS-Cognitive Subscale score (range 0–70, with higher scores indicating greater impairment). The secondary outcomes included changes in the following: ADCS-ADL Scale score, 18F-florbetapir amyloid positron emission tomography (PET) standardized uptake value ratio (SUVR) for six brain regions, and 18F-MK-6240 tau PET medial temporal SUVR for four brain regions.

As for safety, valacyclovir was well tolerated. The number of serious adverse events (AEs) and mild or moderate AEs were similar between the valacyclovir and placebo arms. The most common serious AE was COVID-19 (5 percent vs 3.3 percent), and the most common mild or moderate AE was elevated serum creatinine level (8.3 percent vs 3.3 percent). There were two deaths in each treatment arm that were deemed unrelated to the study treatments.