Elevated Absolute Lymphocyte Counts Following Acute Graft Vs. Host Disease Treatment Is Associated with Therapeutic Responses and Lower Non-Relapse Mortality Following Umbilical Cord Blood Transplant,

Abstract 4085

Acute graft vs. host diseases (aGVHD) is a life threatening complication of allogeneic hematopoietic cell transplantation (allo-HCT). In a large series of patients undergoing allo-HCT (n=863) we recently demonstrated that following the initiation of corticosteroids, a clinical response at day 28 was associated with an increased likelihood of overall treatment response and a reduction in treatment related mortality (TRM) (MacMillan, Blood, 2010). We have also shown that patients who have rapid lymphocyte recovery following umbilical cord blood transplantation (UCBT) have improved outcomes (Burke, BBMT, 2010). To date, no study has addressed the influence of absolute lymphocyte count (ALC) on aGVHD treatment responses. We hypothesized that high ALC might predict favorable responses to aGVHD therapy. To test this hypothesis we reviewed the ALC at the time of aGVHD diagnosis and weekly following the initiation of corticosteroid therapy (48 mg/m² × 14 days followed by a 10% taper over 8 weeks). ALCs were collected on patients who had clinical laboratory data available in the electronic medical record and who had available data on clinical response to corticosteroid therapy. There were 211 patients transplanted at our center from 2001–2007 who fit the above criteria. Median age was 42 years (range 0.2–69 yrs). Thirty nine patients (19%) were <18 years of age and 135 (64%) were >35 years old. The majority of patients underwent UCBT (n=134, 64%) and most others received a sibling PBSC transplant (n=71, 34%). Myeloablative conditioning was used in 55% of patients. GVHD prophylaxis was mainly with CSA/MMF (n=142, 67%) or MTX/CSA (n=42, 22%). Median time to aGVHD onset was 37 days (11–92). Patterns of aGVHD included skin only (n=91, 43%), gut only (n=36, 17%) or multi-organ involvement (n=82, 39%). At the time of GVHD diagnosis, ALC was not predictive of response to therapy. Likewise, there was no association with the ALC at D7 (after the start of steroid treatment) and therapeutic response. The D14 post-treatment ALC showed an association with clinical response at D28 (a time point previously associated with long-term responses and NRM). Patients with an ALC of 0–0.14, 0.15–0.34 and >0.35 at D+14 after corticosteroid therapy had a 40%, 54% and 68% chance of a clinical response at D28 (p>0.01). This translated into a reduction in non-relapse mortality for patients with higher ALC (0–0.14 vs. >0.15; 33% [19–47%] vs 23% [17–29%], p=0.04). Subgroup analysis showed that these observations were mainly driven by the myeloablatve UCB group. Treating ALC as a continuous variable and using a repeated measures approach, we observed that for every unit increase in ALC between the day of diagnosis and D14, there was a 2.26 higher increased likelihood of having a complete clinical response at D28 (p>0.001). Similar results were seen for NRM (OR=0.89 [0.81–9.8], p=0.01). In multivariate analysis, repeated measures of ALC (from D0-14) remained significant for both D28 clinical response (p=0.001) and NRM (p=0.05). Patients with steroid refractory aGVHD (n=17) were treated with ATG. Because steroid refractory aGVHD has poor outcomes and because ATG can negatively impact ALC, we removed these patients from the analysis to determine the impact of ALC. In this subgroup analysis, ALC at D14 after aGVHD treatment remained associated with D28 clinical response in multivariate analysis (OR=3.01, 95%CI[1.24–7.35], p=0.02). Using repeated analysis, the change in ALC from aGVHD treatment day D0 to 14 was associated with D28 treatment response (OR=4.42 [1.88–10.37], p<0.001). Collectively, these results show that absolute ALC at D14 and increases in ALC from the day of treatment to day 14were associated with D28 clinical response and NRM. ALC maybe a simple and cost effective method to monitor response to aGVHD therapy.