Detection of Multiple Double-Stranded DNA Viruses after Cord Blood Transplantation Is Frequent and Persistent

Background: DNA viral infections remain important causes of morbidity following cord blood transplantation (CBT). The cumulative incidence, risk factors, and kinetics of reactivation of multiple double-stranded (ds)DNA viruses after CBT are unknown. This lack of understanding limits development of strategies for broader prophylaxis.

Methods: Weekly plasma samples through 100 days post-CBT were retrospectively tested by quantitative PCR for HHV-6B, HHV-6A, BK, adenovirus (ADV), and EBV; twice-weekly tests for CMV were performed prospectively. Children and adults with ≥1 year of follow up and availability of >60% of samples while alive, with <14 days between samples, were included. We identified a cohort of 125 CBT recipients from 2007-2014 with a median of 13 samples per patient (range, 2-14). Cumulative incidence curves of any detection of ≥1, ≥2, ≥3, or ≥4 viruses within 100 days were created, risk factors were analyzed in Cox models, and reactivation dynamics were characterized. Patients with CMV reactivation were treated with antivirals; this was not accounted for in these data.

Results: Cohort characteristics are presented in Table 1: 95% of patients had ≥1 virus. Detection rates were: CMV, 58%; HHV-6B, 74%; HHV-6A, 0%; BK, 62%; ADV, 10%; and EBV, 3% (Table 2). Detection of multiple viruses at any time and concurrently was frequent (Fig. 1); ≥3 viruses were detected in 39% of the cohort at any time and in 14% concurrently. The proportion of patients with detection of any and multiple viruses peaked by wk 4 and persisted through wk 14. Grade 3-4 acute GVHD increased risk for detection of ≥2 (aHR, 3.1; p=0.01) and ≥3 (aHR, 3.1; p=0.08) viruses.

Median time to detection was similar for all viruses (3-4 wks) except EBV (7.6 wks; Table 2). Median time to peak viral load from first detection was longest for BK (7 wks) and shortest for HHV-6B (3 days). Despite preemptive therapy for CMV, median time to peak viral load was 3 weeks. Among patients with ≥5 samples tested, the median proportion of positive samples after first detection was <50% for each virus except BK (median positive samples, 100%; Fig. 2).

Median viral loads after reactivation ranged from 2.3 (CMV) to 3.4 (BK) log₁₀ copies/ml; CMV was mitigated by treatment. Max viral loads were significantly higher than first for all except EBV; mean differences ranged from 0.1 (EBV) to 1.6 (BK) log₁₀ copies/ml (Table 2, Fig. 3). Viral load did not markedly differ if viruses were detected alone or concurrent with other viruses (Fig. 3).

Conclusions: We demonstrate frequent and persistent detection of multiple dsDNA viruses through day 100 after CBT. BK demonstrated the greatest and most sustained expansion, whereas HHV-6B reached max levels soon after first detection. These findings provide the rationale to study the impact of multiple virus reactivations on organ disease, health care utilization, and mortality in larger cohorts. These data will be critical to design trials using novel, safer therapies (e.g. CMX001 [brincidofovir], multi-virus-specific T cells) for broad prevention of viral reactivation.