Abstract
Cyclosporine A (CSA) is commonly used after bone marrow transplantation for graft versus host disease (GVHD) prophylaxis. There have been conflicting reports regarding dosing of CSA and its effects on relapse and GVHD. Higher doses of CSA after bone marrow transplantation have been associated with greater leukemia relapse. Because it was unknown whether umbilical cord blood transplantation (UCBT) has a similar risk, we studied the impact of CSA levels on UCBT outcomes in patients with acute leukemia.
We reviewed the records of 242 consecutive patients with acute leukemia who received UCBT and CSA for GVHD prophylaxis at our institution from 2004-2011. Eighty (33%) patients had acute lymphoblastic leukemia and 162 (67%) had acute myelogenous leukemia. Fifty-seven patients (24%) had a single UCBT; one hundred fifty-four patients (64%) underwent myeloablative conditioning. The median age at transplant was 30 (1-71) years and median follow-up 4.95 (0.56-10.29) years. Serum CSA levels were collected at least once a week around day -1 and continuing through ∼day 180. We hypothesized that early CSA levels would have the greatest impact on both graft vs. host disease (GVHD) and graft vs. leukemia (GVL) reactions. For the purpose of this study, we examined CSA levels during the first 6 weeks of administration and determined the median CSA level both early (day -3-20) and late (day 21-42). A median of 13 (range was 4-28) CSA levels were drawn during this time period. The median CSA level determined from day-3-20 was 273.5 and the median CSA from days 21-42 was 249.5. Based on the medians in each three week period, patients were divided into four different groups based CSA level and whether it was above or below the median (low-low; low-high; high-low and high-high).
In multivariate analysis for patients with ALL, disease free survival (DFS) at one year and three years was significantly improved in patients that maintained high levels of CSA from days 21-42 (p=0.02 for low-high and p=0.01 for high-high). Overall survival (OS) was improved at one year for patients with high CSA levels late (p=<0.01 for low-high and high-high) but at three years was improved in only patients that had maintained high CSA levels throughout the entire 6 weeks (p=0.03 for high-high). Patients with high levels of CSA from days 21-42 were also associated with less transplant related mortality (TRM) (p=0.04 for low-high and p=0.01 for high-high). There was no association of CSA levels and relapse or acute GVHD grades II-IV for ALL patients. In multivariate analysis of AML patients, we found that DFS at one year was significantly improved in patients that maintained high CSA levels at any time during the six weeks (p=0.01 for high-low, low-high and high-high) whereas at 3 years, patients that maintained high CSA levels early had a better chance of DFS (p=0.02 for high-low and high-high). OS at three years was also significantly greater in patients with high levels of CSA early (p=0.05 for high-low and high-high). TRM was decreased in patients that had high levels of CSA early (p=0.02 for high-low and high-high). Interestingly, in multivariate analysis, relapse was improved in patients that maintained initially low levels and then high levels of CSA (p=0.05). Acute GVHD grades II-IV was significantly improved in AML patients with high CSA levels for the entire 6 weeks (p=0.04 for high-high).
Taken together, these data suggest that high median CSA levels in the first six weeks is critical to ensure improved DFS and OS in patients with acute leukemia undergoing UCBT. Interestingly, in ALL patients, the improvement was seen in patients that maintained high CSA levels late, whereas in AML patients, it was seen in those that maintained high CSA levels early. This implies that the interaction between CSA dose and GVL may be different between myelogenous and lymphoid leukemia following UCBT.
Wagner:Novartis: Research Funding.
Author notes
Asterisk with author names denotes non-ASH members.
