Topical cyclosporine vs tacrolimus for ocular GVHD prophylaxis after outpatient allogeneic hematopoietic stem cell transplantation: A Phase I–II trial

Abstract

Abstract Background Ocular graft-versus-host disease (oGVHD) is a common complication after allogeneic hematopoietic stem cell transplantation (alloHCT), where dry eye symptoms affect up to 90% of patients. Symptoms usually begin after the 3 month post-transplant and persist indefinitely. Current therapies target established disease, with limited evidence supporting prophylactic strategies. This study evaluated the safety, tolerability and early efficacy of topical tacrolimus (Tac) vs cyclosporine (Cys) for oGVHD prevention in the outpatient alloHCT setting. Methods Non-blinded phase I–II trial (NCT06348602) in patients >18 years undergoing outpatient peripheral blood alloHCT from matched sibling or haploidentical donors. Exclusion criteria: active ocular infection/intolerance to topical Cys or Tac. After engraftment (≥500 neutrophils/µL and ≥20,000 platelets/µL for 3 days), patients were randomized to Tac 0.03% ointment or Cys 0.1% solution twice daily in both eyes for 12 months. Artificial tears were allowed. Systemic GVHD prophylaxis included post-transplant cyclophosphamide, a calcineurin inhibitor, and mycophenolate mofetil.Ophthalmologic evaluations were performed every 3 months using NIH and ICCGVHD criteria. Tolerability was defined as relative (with adverse effects but no discontinuation) or absolute (treatment withdrawal). Adherence was classified as good (>85% of doses) or poor (<14 applications/week). Ocular quality of life (oQF) was assessed with NEI VFQ-25. The primary endpoint was safety/tolerability; secondary endpoints included oGVHD prevention, ocular symptoms, quality of life and relapse-free survival (GFRS).Results Twenty-four patients were enrolled (54.2% Cys, 45.8% Tac); median age 35 years; 70% male. Diagnoses included acute lymphoblastic leukemia (50%), acute myeloid leukemia (29.2%), aplastic anemia (12.5%), and myeloproliferative neoplasms (8.3%). Haploidentical donors (75%). Half received reduced-intensity conditioning. Median CD34+ 8.94×10⁶/kg. Median neutrophil and platelet recovery occurred at 13 and 12.9 days.Acute GVHD developed in 62.5%, with grade 3–4 in 20.8%, mainly skin (93%), GI tract (33%), and liver (20%). Chronic GVHD occurred in 37.5%, moderate-severe in 16.7%. Relapse occurred in 33.3% of patients with acute leukemia. Overall mortality was 20.8%. Median event-free survival was 66.4 weeks, and GFRS was 6.2 months. Ocular GVHD was confirmed in 2 patients (1 per group), and considered probable in 2 additional cases (both in the Cys group). All oGVHD cases were mild, median onset: 18.3 weeks. Dry eye symptoms (DES): 23.1% of Cys-treated had Schirmer <5 mm at baseline vs 0% in Tac. After engraftment, DES were more prevalent in Cys (61.5% vs 18.2%). At 12 months, Schirmer <5 mm was observed only in Cys (33.3% vs 0%). A significant difference between groups was at baseline (Fisher's test, p = 0.0287). While no significant differences in Schirmer 6–10 mm plus positive staining tests at any timepoint, including at 12 months (50% Cys vs 0% Tac, p > 0.2). Both treatments were well tolerated. Burning: 70% Cys vs 55% Tac. Pruritus: 15% Cys vs 45% Tac. Blurred vision: 30.7% Cys vs 36.4% Tac. Foreign body sensation: 23% Cys vs 45% Tac. Tearing: 7.7% Cys vs 27% Tac. Irritation: 23% Cys vs 9% Tac. Pain and warmth reported in Tac (18.2% each). No patients discontinued treatment due to intolerance. Adherence was: 81.8% Tac, 76.9% Cys. Poor adherence occurred more in Cys, associated with depression & disease relapse.oQF was better in Cys at baseline (p = 0.015), before starting treatment. After engraftment, most patients in both groups reported good oQF (≥70%). No significant differences were found between groups at any time. oQF remained stable over 12 months. Poor adherence did not significantly affect quality of life (Wilcoxon p = 0.93 for Cys; p = 0.50 for Tac). Conclusions Both topical Tac and Cys were safe, well tolerated, and associated with high adherence in outpatient alloHCT. Tac showed a trend toward fewer dry eye symptoms and less ocular toxicity, although most comparisons did not reach statistical significance. At six months, both groups maintained good oQF (50%; p = 0.84). The most consistent finding was the preservation of ocular quality of life regardless of prophylaxis. Poor adherence did not negatively affect outcomes. These findings support the feasibility of using topical immunomodulators for oGVHD prophylaxis in outpatient alloHCT. Longer follow-up is needed to assess sustained efficacy.