Predictors of Neutropenia Secondary to Ganciclovir Exposure During Treatment of Congenital Cytomegalovirus Disease

Abstract Background Congenital cytomegalovirus (CMV) infection is the leading cause of non-genetic sensorineural hearing loss (SNHL) in childhood. Treatment of infants with symptomatic congenital CMV disease with intravenous (IV) ganciclovir or oral valganciclovir improves audiologic outcomes when started in the first month of age. Neutropenia is the most common adverse event of antiviral treatment, occurring in 19-63% of infants. No correlation between ganciclovir drug exposure and development of neutropenia has been identified. Methods We utilized pharmacokinetic (PK), pharmacodynamic (PD), and hematologic data from three international, NIH-funded studies of IV ganciclovir or oral valganciclovir conducted by the Collaborative Antiviral Study Group (CASG) from 2002-2018. Drug exposure was estimated using area-under-the-curve (AUC) values and was defined as high (AUC≥40 mg•hr/mL) or low (AUC<40 mg•hr/mL). Baseline absolute neutrophil count (ANC) was defined as high (ANC>1000 cells/mm3) or low (ANC≤1000 cells/mm3) and regularly monitored for development of neutropenia (ANC<800 cells/mm3). Ordinal logistic regression determined if AUC or baseline ANC was a predictor of neutropenia severity across the course of treatment. Time to neutropenia was defined as the first study day that a participant had an ANC<800. Median time to neutropenia was determined using nonparametric estimates for each group of subjects and Kaplan-Meier curves were created. Fisher’s Exact Test compared proportions. Results 139 subjects were included. Neither AUC nor baseline ANC were significant predictors of neutropenia severity using ordinal logistic regression (P=0.62, P=0.35, respectively). However, a greater proportion of subjects with a baseline low ANC developed neutropenia (75% vs 40%, P=0.004; Odds Ratio 1.9, 95% CI 0.48-7.67). Subjects with a baseline low ANC also developed neutropenia more quickly than those with a baseline high ANC (Figure 1, median of 1.5 vs 216 days, P<0.0001), as did subjects with high drug exposure compared with low drug exposure (median of 28 vs 215 days, P=0.08). When considering only subjects with low baseline ANC, median time to neutropenia development was similar in those with low versus high drug exposure (median of 1 vs 2 days, P=0.23). These findings are limited by 50% of the 16 subjects with low baseline ANC being neutropenic on day 1. Likewise, drug exposure made no difference in time to development of neutropenia in subjects with high baseline ANC (P=0.36). All 3 subjects with high drug exposure and low baseline ANC developed neutropenia. Conclusion Low ANC at baseline was correlated with the development of neutropenia in patients treated for symptomatic congenital CMV disease. High drug exposure hastened the development of neutropenia compared with low drug exposure. Further studies should investigate if different dosing regimens such as body weight versus body surface area affect development of neutropenia. Figure 1. Kaplan-Meier Curve of Neutropenia Development by Baseline ANC.