Congenital cytomegalovirus in a high HIV prevalent setting, South Africa

Abstract

Introduction: Prevalence of congenital cytomegalovirus (cCMV) is higher in infants born to women with non-primary than primary cytomegalovirus (CMV) infection and can result in debilitating neurological sequelae.

Objectives: To describe the epidemiology of cCMV in an urban, black South African population in a high human immunodeficiency virus (HIV) prevalence setting in the antiretroviral therapy (ART) era. The prevalence of cCMV in HIV-exposed and HIV unexposed neonates, the occurrence of symptomatic cCMV at birth and neurological sequelae within 12 months of age were determined. Postnatal CMV (pCMV) acquisition in relation to maternal CMV shedding was assessed and vaccination induced immune responses between cCMV, pCMV and CMV-uninfected infants explored. In stillbirths and neonatal deaths aged below 21 days, the prevalence of cCMV and attribution of CMV to foetal and neonatal demise was explored.

Materials and methods: Neonates born at Chris Hani Baragwanath Academic Hospital in Soweto, between May –December 2015 were screened and confirmed for cCMV by testing saliva and urine samples by polymerase chain reaction assay (PCR) within 21 days of birth. Confirmed cCMV cases were matched by gender, gestation and HIV-exposure to cCMV uninfected neonatal controls in a 1:2 ratio. Controls tested CMV negative on two independent saliva samples within 21 days of birth. Cases and controls were followed up in a matched cohort study until 12 months of age for neurological sequelae (sensorineural hearing loss and neurodevelopment). Controls had saliva tested at two, six and 12 months of age to identify pCMV acquisition. Mothers of cases and controls had a once off breast milk and vaginal fluid tested by PCR for CMV shedding within 21 days. Whole blood from all infants at seven viii

months of age was tested for antibodies to the primary series of childhood vaccines, combined Diphtheria-Tetanus-acellular Pertussis-Inactivated Polio-Haemophilus Influenzae type b- Hepatitis B (DTaP-IPV-Hib-HBV), by an in-house Luminex multiplex immunoassay. In neonates that died within 21 days of birth and stillbirths identified through a mortality surveillance, minimally invasive tissue sampling tested for CMV by PCR, and with review of antemortem clinical data, cause of death ascertained.

Results: Overall cCMV prevalence was 2.5% (95% Confidence Interval [CI] 1.9–3.2). Prevalence of cCMV was higher in children born to women living with HIV (i.e. HIV exposed, 5.2%, 95% CI 3.8–6.9) than HIV-unexposed (1.4%, 95% CI 0.9–2.0) neonates. Among all HIV-exposed neonates, those with cCMV were 20-fold (odds ratio [OR] 20.1, 95% CI 6.09–66.46) more likely to be infected with HIV in-utero (5/27, 18.5%) than cCMV uninfected infants (8/716, 1.1%). Symptomatic illness occurred in 7% of cCMV cases. Rate of neurological sequelae within 12 months of age was, however, similar in infants with cCMV (6%) compared to controls without cCMV infection (4%, OR 4.0, 95% CI 0.36-44.11), albeit with wide intervals of statistical uncertainty. Of the cCMV-uninfected, 79.7% acquired CMV postnatally by 12 months of age, which was unrelated to maternal CMV shedding within 21 days postpartum. The immune responses to the primary series DTaP-IPV-Hib-HBV vaccine was similar in cCMV, pCMV and CMV-uninfected infants at seven months of age. CMV was identified in 11.8% of stillbirths and neonatal deaths (<21 days age), however, cCMV was attributed as the cause of death in 3.9% of stillbirths and none of the neonates.

Conclusion: cCMV rates was higher in neonates born to women living with HIV than in HIV-unexposed neonates, even in the era of ART. Vertical HIV transmission was more frequent in neonates with cCMV, suggesting cCMV co-infection may account for a large fraction of the residual maternal to foetal HIV transmission in populations with efficient ART ix

programs. Neurological sequelae within one year did not differ between cCMV infected and uninfected children. CMV infection was higher in neonates dying within 21 days of birth and stillbirths than in live births; although clinical review of the cases did not attribute CMV to be in the casual pathway of neonatal deaths and in only 3.9% of stillbirths. Follow-up of cCMV infected infants into childhood is required to determine long term neurological sequelae and whether a universal or targeted cCMV screening programme in newborns would be of value in South African populations.