The Role of Donor CMV Serostatus on Outcome after T-Cell Replete Haplo-SCT and Post-Transplant Cyclophosphamide: A Cohort Analysis on 207 Consecutive Adult Patients

INTRODUCTION. In the setting of T-cell replete haploidentical transplantation, the question about the best stem cell donor is still open. Cytomegalovirus (CMV) serostatus represents one of the main factors affecting outcome, however its role has to be yet characterized after haploidentical stem cell transplantation with post-transplant high-dose cyclophosphamide (haplo-SCT with PT-Cy).

METHODS. Four cohorts were followed up, according to patient/donor CMV serostatus, on a total of 207 haplo-SCT with PT-Cy performed at two institutions (Institut Paoli-Calmettes, Marseille and Istituto Clinico Humanitas, Milano) from September 2009 to April 2015; cohort 1, patient/donor CMV -/- (n=36); cohort 2, -/+ (n=26), cohort 3, +/+ (n=110), cohort 4, +/- (n=35). CMV reactivation, OS and TRM were observed in the four cohorts and the adjusted effect of patient/donor CMV serostatus was evaluated through multivariate Cox regression on OS and TRM. 

RESULTS. Median follow-up from transplant was 660 days (range: 42-1826). The main characteristics (patient/donor gender, diagnosis, patient’s age, disease status before transplant, DRI, AB0 matching, source of stem cells, conditioning intensity, HCT-CI, year of transplant) were analyzed and did not significantly differ among the groups except for the source of stem cells that was unbalanced in cohort 1 (approx. ratio PBSC:BM of 3:1) vs. the cohorts 2,3,4 (ratio 1:1, p=0.04). The rate of CMV reactivation was 42% for the entire population and 0%, 45%, 52%, 51% in cohorts 1 to 4, respectively, with a longer time to first reactivation in the cohort 2 vs. 3 and 4: day+63 vs. +41 and +42. Two-year OS was 62%, 65%, 50%, 42% in cohorts 1 to 4, respectively (64% vs. 48% in cohorts 1+2 vs. 3+4, p=0.01); TRM was 9%, 17%, 24% and 31% in the same groups.  After adjustment, OS was not significantly impaired if a CMV seropositive donor was used for a CMV seronegative patient instead of a CMV seronegative donor (table 1); mortality risk was higher among CMV+ patients, without differences according to donor CMV serostatus. Results on TRM were similar (table 1). Of note, no significant differences of acute or chronic GvHD were observed among the 4 cohorts (p=0.86 and p=0.12, respectively).

CONCLUSIONS: for a CMV- patient, the use of a CMV+ donor did not appear to impair survival after haplo-SCT with PT-Cy. Outcome of CMV+ patients is worse than CMV- patients, without significant differences according to donor CMV serostatus. Further studies with larger series (i.e. collaborative and/or registry studies) are warranted to better clarify this issue.    

Table 1. Multivariate Cox regression on OS and TRM.

OS				TRM
Patient/donor CMV serostatus	HR	95% CI	P	HR	95% CI	P
-/-	1			1
-/+	0.98	0.37-2.54	0.96	1.41	0.28-7.02	0.68
+/+	2.28	1.16-4.47	0.02	3.64	1.09-12.20	0.04
+/-	2.18	1.01-4.69	0.05	3.80	1.02-14.09	0.05

Variables with p<0.05 in the OS model: source of stem cells, HCT-CI, DRI.

Variables with p<0.05 in the TRM model: none (trend for HCT-CI, p=0.09)