Infant deaths from respiratory syncytial virus in Lusaka, Zambia from the ZPRIME study: a 3-year, systematic, post-mortem surveillance project

Abstract

Background: Respiratory syncytial virus (RSV) is the leading cause of acute lower respiratory tract infections and a key driver of childhood mortality. Previous RSV burden of disease estimates used hospital-based surveillance data and modelled, rather than directly measured, community deaths. Given this uncertainty, we conducted a 3-year post-mortem prevalence study among young infants at a busy morgue in Lusaka, Zambia-the Zambia Pertussis RSV Infant Mortality Estimation (ZPRIME) study.

Methods: Infants were eligible for inclusion if they were aged between 4 days and less than 6 months and were enrolled within 48 h of death. Enrolment occurred mainly at the University Teaching Hospital of the University of Zambia Medical School (Lusaka, Zambia), the largest teaching hospital in Zambia. We extracted demographic and clinical data from medical charts and official death certificates, and we conducted verbal autopsies with the guardian or next of kin. RSV was identified using reverse transcriptase quantitative PCR and stratified by age, time of year, and setting (community vs facility deaths). By combining the PCR prevalence data with syndromic presentation, we estimated the proportion of all infant deaths that were due to RSV.

Findings: The ZPRIME study ran from Aug 31, 2017, to Aug 31, 2020, except for from April 1 to May 6, 2020, during which data were not collected due to restrictions on human research at this time (linked to COVID-19). We enrolled 2286 deceased infants, representing 79% of total infant deaths in Lusaka. RSV was detected in 162 (7%) of 2286 deceased infants. RSV was detected in 102 (9%) of 1176 community deaths, compared with 10 (4%) of 236 early facility deaths (<48 h from admission) and 36 (5%) of 737 late facility deaths (≥48 h from admission). RSV deaths were concentrated in infants younger than 3 months (116 [72%] of 162 infants), and were clustered in the first half of each year and in the poorest and most densely populated Lusaka townships. RSV caused at least 2·8% (95% CI 1·0-4·6) of all infant deaths and 4·7% (1·3-8·1) of community deaths.

Interpretation: RSV was a major seasonal cause of overall infant mortality, particularly among infants younger than 3 months of age. Because most RSV deaths occurred in the community and would have been missed through hospital-based surveillance, the global burden of fatal RSV has probably been underestimated.

Funding: Bill & Melinda Gates Foundation.

Figure 1

Study enrolment flow diagram Infants who died at sites other than UTH or the four satellite sites that could issue death certificates could contribute deaths outside our enrolment system. Most of our data were collected at UTH, yielding detailed long-form data. These data had sufficient detail to be used in all three study aims. By contrast, the limited short-form data, collected at satellite sites, contributed mainly to aim 1 and, to a partial extent, aim 2. The three aims were: aim 1 (proportion of RSV among all deaths) is a straightforward prevalence calculation, where the numerator is RSV-positive deaths, and the denominator is all infant deaths; aim 2 stratifies prevalence by community versus facility, and into early facility deaths versus late facility deaths; and aim 3 follows the approach of the Global Enteric Multicenter Study to estimate the population attributable fraction for deaths caused by RSV. UTH=University Teaching Hospital of the University of Zambia Medical School.

Figure 2

RSV deaths over time The histogram shows the numbers of infant deaths by month. The percentages at the top of each bar are the proportion of RSV-positive deaths out of total deaths in each month. RSV=respiratory syncytial virus.