Abstract

Background: Blood stream infections (BSI) are a major source of morbidity and mortality after allogeneic blood and marrow transplantation (BMT). In studies of risk for BSI various factors have been identified. The impact of cytomegalovirus (CMV) viremia, however, on risk has not been assessed. This is important since both CMV infection and ganciclovir (GCV), the mainstay of pre-emptive therapy, have myelosuppressive and immunosuppressive effects. We conducted a retrospective analysis to test the hypothesis that CMV viremia predisposes allogeneic BMT patients to BSI.

Methods: We retrospectively analyzed 278 allogeneic BMTs performed at Children's Healthcare of Atlanta between January 1st 2005 and December 31st 2014 that met eligibility criteria. The primary outcome was the first episode of BSI occurring between day +30 and +100 seen in engrafted patients following allogeneic BMT. We compared clinical characteristics between patients with and without BSI. This analysis was based on a time dependent competing risk model. Risk events including acute GVHD and CMV viremia were counted only when they preceded the BSI. We performed a multivariate analysis to estimate the effect of CMV viremia on risk for BSI in the post-engraftment period (days +30-100).

Results: The median age was 9 years (range 0-22 years). 44.6% of the patients were Caucasian, while 38.5% were African-American. 59.7% received an unrelated marrow, unrelated cord or the alternative donor transplant. 60.8% were transplanted for a hematologic malignancy. The cumulative incidence of BSI between day +30 and+100 was 21.9% (95% confidence interval (CI), 17.5-27.3%). The median day of BSI development was 54 (range 30-99). The leading cause of BSI was Staphylococcus epidermidis. Gram-positive cocci were responsible for 71.4% of all BSIs. In the multivariate analysis, three factors were significant: grade III/IV aGVHD (hazard ratio (HR)=3.490, 95% CI 1.530-7.980, P=0.003), CMV viremia (HR=3.410, 95% CI 1.410-8.250, P=0.007), and African-American ethnicity (HR=2.520, 95% CI 1.130-5.600, P=0.024). Form of insurance (Medicaid/no insurance vs. other), employed as a surrogate for income level, had no influence on risk. In the 57 patients with CMV viremia, the frequency of neutropenia (<500/mm3) after the development of CMV viremia until day +100 was 26.3% (15/57), significantly higher than that observed in patients without CMV viremia (5.4% (12/221), P<0.001).

Conclusion:

Our analysis of risk for post-engraftment BSI in children yielded three factors. One, aGVHD, is well established. The other two, CMV viremia and African-American ethnicity, have not been previously recognized. That CMV viremia would predispose to subsequent BSI is highly plausible. Our results suggest that the risk of CMV viremia may be mediated in part by neutropenia stemming from CMV or its treatment with ganciclovir. That African-American ethnicity has not previously been recognized as a risk factor may simply reflect the ethnic make up of previous studies. Our findings are consistent with those of studies performed in other settings, showing patients of African descent to be at increased risk for bacteremia and sepsis. An increased risk for BSI among patients of African descent could partly explain the concerning excess risk for transplant related mortality previously observed among these patients.

Disclosures

No relevant conflicts of interest to declare.

Author notes

*

Asterisk with author names denotes non-ASH members.