Systemic and Topical Antiviral Control of Cytomegalovirus Anterior Uveitis: Treatment Outcomes - Trials I and II

Cytomegalovirus (CMV) anterior uveitis is an increasingly recognized cause of recurrent and chronic ocular inflammation in immunocompetent individuals, accounting for up to 25% of anterior uveitis cases at specialized centers. The condition typically presents with unilateral hypertensive uveitis, coin-shaped or stellate keratic precipitates, and variable anterior chamber inflammation. Without adequate treatment, CMV anterior uveitis can lead to secondary glaucoma from trabecular meshwork damage and corneal endothelial decompensation, both of which may result in permanent vision loss. Diagnosis requires PCR detection of CMV DNA in aqueous humor via anterior chamber paracentesis.

Despite growing recognition, particularly in Asian and Asian-American populations, no randomized controlled trial has previously established an evidence-based treatment protocol. Observational data and small case series suggest that both oral valganciclovir and topical ganciclovir 2% ophthalmic solution can suppress CMV replication and control inflammation, and that some cases improve without antiviral therapy. However, the relative efficacy of these approaches has not been formally compared, and whether long-term antiviral suppression prevents disease recurrence remains undefined. Prior observational work has demonstrated a correlation between higher aqueous CMV DNA viral load and worse anterior segment outcomes, including greater corneal endothelial cell loss, providing biological rationale for targeting viral load as a primary endpoint in a treatment trial.

Study Design: STACCATO is a sequential, double-masked, placebo-controlled randomized clinical trial consisting of two linked sub-studies (Trial I and Trial II) conducted at six international sites: the Francis I. Proctor Foundation at UCSF (clinical and data coordinating center), UCLA, Chulalongkorn University (Bangkok, Thailand), Khon Kaen University (Khon Kaen, Thailand), Chiang Mai University (Chiang Mai, Thailand), and Chang Gung Memorial Hospital (Taoyuan, Taiwan). Both trials employ a three-arm parallel-group design with 1:1:1 allocation. Randomization uses site-stratified permuted blocks (block sizes 3 and 6) to ensure balance across sites. Allocation concealment is maintained through a centralized electronic randomization system. Both participants and treating ophthalmologists are masked to treatment assignment throughout each trial period. Active and matching placebo tablets and eye drops are identical in appearance, packaging, and administration schedule.

Trial I: Acute Treatment (7-Day Intervention) Trial I evaluates the short-term virologic and clinical effects of antiviral therapy in participants with PCR-confirmed active CMV anterior uveitis. Participants are randomized to one of three treatment arms: Arm 1: oral valganciclovir 900 mg (two 450 mg tablets) twice daily for 7 days, plus matching placebo eye drops six times daily; Arm 2: topical ganciclovir 2% ophthalmic solution six times daily for 7 days, plus matching placebo tablets twice daily; Arm 3: matching placebo tablets twice daily plus matching placebo eye drops six times daily. All participants receive topical prednisolone acetate 1% as standard of care for inflammation control; ocular hypotensive medications are added as clinically indicated for intraocular pressure management.

The primary endpoint of Trial I is log10-transformed aqueous humor CMV viral load at Day 7, measured by quantitative PCR (qPCR) at the central laboratory (Stanford University). Aqueous humor samples are obtained at baseline (Exam 0, pre-enrollment paracentesis) and at Day 7 (Exam 2) via anterior chamber paracentesis; approximately 100 µL are collected at each visit, of which approximately 50 µL is used for local directed qualitative PCR testing for eligibility determination (CMV, HSV, and VZV), and the remaining sample is stored at -80°C and shipped to the central laboratory for quantitative CMV PCR analysis. The pre-specified aqueous lower limit of quantitation (LLOQ) is 2,700 IU/mL (3.43 log10 IU/mL), based on the assay's plasma-validated linear range of 135-6,750,000 IU/mL (2.13-6.83 log10 IU/mL) with 50 µL samples diluted to 1 mL. Viral load values reported as not detected are imputed using maximum likelihood estimation (MLE) in the primary analysis, with sensitivity analyses employing alternative imputation strategies.

Secondary endpoints in Trial I include: (1) proportion of participants achieving clinical inactivity (≤0.5+ anterior chamber cell per SUN criteria) at Days 7 and 21; (2) intraocular pressure (IOP) in mmHg at Days 7 and 21; and (3) visual acuity (logMAR) at Days 7 and 21. A nested observational analysis evaluates the effect of pre-enrollment topical corticosteroid use on baseline aqueous CMV viral load.

Trial II: Long-Term Suppressive Therapy (12-Month Follow-up) Trial II evaluates whether long-term antiviral suppression reduces the rate of inflammatory recurrence in participants who have achieved clinical quiescence following PCR-confirmed CMV anterior uveitis. Participants may enter Trial II via two pathways: (1) participants who complete Trial I and subsequently achieve clinical inactivity, defined as ≤0.5+ anterior chamber cell sustained for at least 2 weeks with a minimum 2-week washout from antiviral therapy; or (2) patients with previously PCR-confirmed CMV anterior uveitis who are already clinically inactive and have not participated in Trial I.

Upon enrollment, participants are re-randomized (independently of their Trial I assignment) to one of three long-term suppressive regimens: Arm 1: oral valganciclovir at suppressive dosing for 12 months, plus matching placebo eye drops; Arm 2: topical ganciclovir 2% ophthalmic solution at maintenance dosing for 12 months, plus matching placebo tablets; Arm 3: matching placebo tablets plus matching placebo eye drops for 12 months.

The primary endpoint of Trial II is recurrence of anterior uveitis inflammation, defined as ≥1+ anterior chamber cell, over 12 months of follow-up. Participants are assessed for recurrence at scheduled visits (approximately months 1, 3, 5, 7, 9, 11, and 12 after randomization) and at unscheduled visits prompted by participant-reported symptoms. If recurrent inflammation is detected at any visit, an anterior chamber paracentesis is performed to assess for the presence of detectable CMV viral load in aqueous humor, and the participant exits the study. A secondary endpoint of Trial II is the prevalence of CMV mutations conferring antiviral resistance in aqueous humor at recurrence, assessed by viral genomic sequencing.

Genomic and Translational Aims: Aqueous humor samples from baseline (Trial I, Exam 0) and recurrence visits (Trial II) are subjected to RNA sequencing to characterize host transcriptional signatures and viral genomic features associated with CMV anterior uveitis. Specific hypotheses include: (a) participants who experience recurrent inflammation after antiviral suppression will have a higher prevalence of CMV resistance mutations compared to baseline samples from Trial I; (b) host transcriptional profiles will distinguish CMV-positive from CMV-negative anterior uveitis; and (c) unique transcriptional signatures will be associated with recurrent inflammation in Trial II participants. This aim is designed to identify biomarkers of treatment response and recurrence risk, and to elucidate mechanisms of viral persistence and resistance informing future therapeutic development.

Statistical Analysis Plan Sample Size: Trial I is powered to detect a 0.8 log10 IU/mL difference in post-treatment CMV viral load between oral valganciclovir and each comparator arm. With baseline viral load as a covariate (ANCOVA), 39 participants per arm (117 total) provides 80% power at a two-sided alpha of 0.05, accounting for approximately 10% loss to follow-up. For Trial II, assuming a 25% recurrence rate in the placebo group, 33 participants per arm (99 total) provides 80% power to detect a 36 percentage-point reduction in recurrence risk at a two-sided alpha of 0.05.

Primary Analyses: The primary efficacy analysis for Trial I compares log10-transformed post-treatment viral loads using ANCOVA, adjusting for baseline viral load, treatment arm, and research site. Multiple comparison adjustment uses the max-T method for the two primary comparisons (oral valganciclovir vs. topical ganciclovir 2%, and oral valganciclovir vs. placebo); topical ganciclovir 2% vs. placebo is a pre-specified secondary comparison. The primary analysis for Trial II uses log-binomial regression to evaluate the risk of inflammatory recurrence over 12 months, adjusted for baseline viral load, treatment arm, and research site.

Secondary and Sensitivity Analyses: Secondary clinical endpoints in both trials are analyzed using linear regression (continuous outcomes) or log-binomial regression (binary outcomes) with analogous covariate adjustment. Pre-specified sensitivity analyses for Trial I evaluate robustness to alternative approaches for undetectable viral load values, including substitution with 1 IU/mL, MLE imputation below the LLOQ, exclusion of participants with undetectable pre-treatment viral loads, and binarization by LLOQ or detectability thresholds. Subgroup analyses include stratification by sex and adjustment for glaucoma medication use as a proxy for baseline disease severity.

Interim Analyses and Stopping Rules: An independent Data Safety Monitoring Board (DSMB) will review accumulating safety and efficacy data at scheduled intervals throughout both trials. Formal interim efficacy analyses employ the O'Brien-Fleming alpha spending approach to preserve overall Type I error. The DSMB has authority to recommend early study modification or termination for safety or overwhelming efficacy.

Prior Work: Pilot STACCATO Trial The current study builds on the completed pilot STACCATO trial (NCT03586284), which enrolled 51 participants across three sites (UCSF, Chulalongkorn University, and Khon Kaen University) between January 2020 and March 2024. The pilot demonstrated feasibility of double-masked RCTs with repeated anterior chamber paracentesis in CMV anterior uveitis. Oral valganciclovir produced the largest absolute reduction in aqueous CMV viral load at Day 7 (0.44 log10 IU/mL), though between-arm differences did not reach statistical significance in this underpowered cohort (early termination due to COVID-19). Baseline and Day 7 viral loads were strongly correlated (Pearson's r=0.774). No serious procedural complications (endophthalmitis, traumatic cataract) were observed. The current study is powered to detect the virologic and clinical differences suggested by these pilot data.

Study Oversight and Data Management The study is overseen by an independent DSMB comprising a uveitis/ocular infectious disease specialist, a biostatistician with clinical trial expertise, and an infectious disease/antiviral pharmacology specialist, none affiliated with the participating institutions. Study data are captured in a validated electronic data capture system with built-in range checks and logic validation; source data verification is conducted at each site. Quantitative PCR analyses of all aqueous humor samples are performed at the central laboratory (Stanford University) to ensure assay comparability across sites and throughout the study period. The study is funded by the National Eye Institute-National Institutes of Health.