We retrospectively analyzed the efficacy and safety of vedolizumab combined with basiliximab for SR-GI-aGVHD at four hospitals in China (Nanfang Hospital, Guangdong Second Provincial General Hospital, the Second Affiliated Hospital of Guangzhou Medical University, and the Seventh Affiliated Hospital of Sun Yat-Sen University). Inclusion criteria were any patient 18 years old who had undergone allo-HSCT using any donor or graft source, with any conditioning regimen between March 2019 and March 2022.

The diagnostic criteria for aGVHD are based on the literature criteria established by the Mount Sinai Acute GVHD International consortium group. SR aGVHD was defined as disease progression after 3 days of therapy onset with ≥2 mg kg⁻¹ day⁻¹ of systemic glucocorticoid or equivalent, or lack of improvement after 7 days of treatment initiation, or failure during methylprednisolone taper. This study was approved by the institutional review board of each participating hospital, and all patients (or their guardians) provided signed written informed consent before inclusion. The study was conducted in accordance with the Declaration of Helsinki.

Patients received methylprednisolone as the first-line treatment at a dose of 1 mg/kg. If the patients were diagnosed with SR aGVHD, combination therapy was started with basiliximab (20 mg per dose on days 1, 3, and 8 and repeated weekly until aGVHD was improved to grade < II) and vedolizumab. Patients were administered vedolizumab at the same dosage and schedule and approved for the therapy of IBD: an initial dose of 300 mg IV on days 1, 15, 43, and 71. Whether to continue vedolizumab was based on the response evaluated on day 28. In addition to cyclosporine A (CSA) and mycophenolate mofetil (MMF), no other immunosuppressants were administered during the vedolizumab–basiliximab treatment. CSA and MMF were initiated on day −9.

Patients with complete response (CR) and no response (NR) discontinued vedolizumab and basiliximab, but the CSA and MMF were continued, while patients with partial response (PR) continued to receive vedolizumab until aGVHD showed CR or vedolizumab was received for four courses. Supportive care followed department guidelines for each patient. Patients without prior fungal infection were administered fluconazole as prophylaxis against fungal infection, while patients with previous fungal infections received treatment with previously effective antifungal drugs. Steroids were gradually tapered by 30% every 5 days and stopped within 4 weeks after the second dose of basiliximab. Other immunosuppressants, such as MMF, methotrexate (MTX), ruxolitinib and mammalian target of rapamycin (mTOR) inhibitor, mesenchymal stromal cells (MSCs), and fecal microbiota transplantation (FMT), were allowed in NR patients.

All patients were diagnosed as LGI-aGVHD by pathological biopsy. aGVHD was staged and graded according to the consensus criteria, and the assessment of response was evaluated by standard definitions. CR was defined as complete remission and improvement of aGVHD signs and symptoms. PR was defined as an improvement of one stage in lower GI aGVHD without any additional systemic treatment, while other organs showed no progression. NR was defined as the absence of remission in lower GI aGVHD without progression in any other organ. Progression was defined as the deterioration in one or more organs by one or more stages and without improvement in any involved GVHD organ. Patients who died before day 28 were considered non-responders, and those who had worsened underlying disease before day 28 were considered non-evaluable.

Safety analysis was conducted by reviewing adverse events (AEs) and malignancy recurrence in all patients through medical records. AEs were recorded between the first dose of the combination and the last follow-up, including infusion reactions, infection, hepatotoxicity, or progressive multifocal leukoencephalopathy (PML). Follow-up care was conducted through physical examination and laboratory assessments, such as routine blood testing, biochemical tests, bone marrow assessment, and pathogen detection [cytomegalovirus (CMV) DNA, Epstein–Barr virus (EBV) DNA, etc.].

Assessments of AEs were performed daily for the first week, weekly from week 1 to week 8, monthly from the second month to the third month, and every 3 months thereafter to collect data on progression, survival, chronic GVHD (cGVHD), and safety outcomes including malignancy relapse and infections.

The primary endpoint was the overall response at day 28 (time from combination treatment to day 28), which was defined as the proportion of patients who had a CR or PR as compared with baseline organ staging. The key secondary endpoint was the duration of overall response at day 56, which was defined as the proportion of patients who had a CR or PR at day 28 and maintained response until day 56. Other secondary endpoints included failure-free survival (FFS) (time from combination to GVHD relapse or malignancy recurrence, non-relapse related death, or the addition of new systemic therapy for aGVHD), overall survival (OS), the onset of cGVHD [according to the National Institutes of Health (NIH) criteria], and non-relapse mortality (NRM).

The data cutoff was October 30, 2022. Complete response rate (CRR), overall response rate (ORR), and two-sided 95% confidence intervals (CIs) were calculated using the Wilson method. Fisher’s exact test was used for categorical variables and the Mann–Whitney U tests for continuous variables. Kaplan–Meier (KM) methodology was used to estimate FFS and OS, R was used for plotting the KM curve, and the hazard ratios (HRs) were also calculated, together with its 95% CI, using a stratified Cox model. All statistical tests were two-tailed with a significance level of 0.05. SPSS version 20.0 and R version 3.3.0 were used for data analysis.

Between March 2019 and March 2022, a total of 28 patients from four hospitals were included in this study. There were nine (32.1%) women and 19 (67.9%) men, with a median age of 36 years (range, 18 to 58 years). Acute leukemia was the most common indication for allo-HSCT. Sixteen (57.1%) patients received transplants from an HLA-haploidentical donor (HID) and 12 (42.8%) patients from an HLA-matched sibling donor (MSD). CsA was used in combination with both MTX and MMF as aGVHD prophylaxis for all patients; patients with HID were administered antithymocyte globulin (ATG). Prior to starting basiliximab and vedolizumab, all patients had grade III (n = 13, 46.4%) to IV (n = 15, 53.6%) LGI-aGVHD, only one patient had aGVHD of the LGI tract alone, 24 (85.7%) involved liver aGVHD, and 14 (50.0%) involved skin aGVHD. Diagnosis of all patients was confirmed by colonoscopy biopsy. The median exposure to glucocorticoid for all patients before combination treatment was 6 days (range, 3–15). All patients received the combination of basiliximab and vedolizumab as second-line therapy after failing to respond to glucocorticoid. The median days to start of combination treatment after GI aGVHD diagnosis was 7 days (range, 4–16). The median dose of basiliximab given was 3 (range, 1–5). The median dose of vedolizumab given was 2 (range, 1–3). The median follow-up was 13 months for all patients. The full details of the baseline are given in Table 1 .

The primary endpoint, overall response (OR) at day 28, was observed in 21 patients, producing an ORR of 75.0% (95% CI: 54.8%–88.6%). Among the 21 responders, 18 achieved CR (64.3%, 95% CI: 44.1%–80.7%), and three achieved PR (10.7%, 95% CI: 2.8%–29.4%); six (21.4%, 95% CI: 9.0%–41.5%) had no response, two (7.1%, 95% CI: 1.2%–25.0%) of whom died on day 28. The ORR at any time was 78.6% (95% CI, 58.5%–91.0%; CR, 71.4%), which included one patient who had a CR but died of infection before day 28 ( Figure 1 ).

At day 56, a durable OR was observed in 18 of 28 patients (64.3%, 95% CI: 44.1%–80.7%) and a CR in 17 of 28 patients (60.7%, 95% CI: 40.7%–77.9%). At day 100, 17 of these patients experienced a durable response and were free of LGI-GVHD. Among evaluable patients (n = 28), the median time to PR after starting basiliximab and vedolizumab was 8 days (3–24 days), and the median time to CR was 24 days (8–34 days) ( Figure 2 ).

Subgroup analysis of patients’ characteristics showed that day 28 response and durable OR (DOR) were associated with the onset time of treatment after GI-aGVHD was diagnosed. Patients (n = 17) who started treatment of basiliximab and vedolizumab within 7 days after the diagnosis of GI-aGVHD had a statistically significant difference in day 28 OR of 94.1% vs. 45.5% for patients (n = 11) who started after 7 days (p = 0.007). Durable OR at day 56 was 88.2% vs. 27.3% among patients who started treatment within 7 days and later initiation (p = 0.003), respectively ( Figure 3 ). It is worth mentioning that there was no significant difference in baseline characteristics between the two groups.

The 100-day, 6-month, and 12-month OS rates for the entire cohort of patients were 60.7% (95% CI: 45.1%–81.8%), 60.7% (95% CI: 45.1%–81.8%), and 47.6% (95% CI: 31.4%–72.1%), respectively. The median OS was 332 days (95% CI: 80–not evaluable) ( Figure 4A ). The median failure-free survival was 276 days (95% CI: 50–not evaluable) ( Figure 4B ). For patients in CR or PR (n = 21) at day 28, OS was greater for day 28 responders compared with non-responders (p < 0.0001) ( Figure 4C ). Administered time was also significantly associated with OS. OS was also greater for patients who started treatment of basiliximab and vedolizumab within 7 days compared with later initiation (p = 0.00013) ( Figure 4D ). There was no significant difference in ORR or survival between centers. By 12 months after basiliximab and vedolizumab initiation, two patients died of underlying disease relapse. Twelve patients (34.6%) died of causes other than malignancy relapse, including five (15.4%, 95% CI: 6.8%–37.6%) patients who died of serious infection, four (14.3%, 95% CI: 4.7%–33.6%) who died of progression of aGVHD, and three (10.7%, 95% CI: 2.8%–29.4%) who died of thrombotic microangiopathy (TMA). The 12-month NRM was 42.9% (95% CI: 24.1.0%–60.3%) ( Figure 4E ). Meanwhile, malignancy relapse was observed in two patients (12%, 95% CI: 1.9%–32.8%) at 1-year follow-up ( Figure 4F ).

Basiliximab and vedolizumab did not cause any infusion-related toxicity, hepatotoxicity, or PML in our study. There were no patients in our study who discontinued due to adverse events other than GVHD non-response. Fifty infections were recorded in 20 patients, of which 27 were bacterial, 18 were viral, and five were fungal ( Table 2 ). Nine (32.1%) patients developed CMV reactivation complicated with bacterial infections (25.0%, CMV infection; 7.1%, CMV viremia). EBV occurred in five patients (17.9%), and there were no posttransplant lymphoproliferative disorders in our case. Only three patients (10.7%) developed pseudomembranous colitis. Pulmonary infection (21 of 50) was the most common in nine patients. Six (21.4%) patients had coexisting intestinal infections (10 of 50 infections).

Among the 17 patients who survived for more than 100 days, cGVHD occurred in seven patients. The 1-year cumulative incidence of cGVHD was 35.3%, and 11.8% of patients developed moderate-to-severe cGVHD.