Epidemiology of streptococcus pneumoniae post-pneumococcal conjugate vaccine introduction in South Africa
Date
2016
Authors
Von Mollendorf, Claire Emily
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
BACKGROUND
Streptococcus pneumoniae is a leading cause of severe invasive bacterial infections globally;
estimated to cause over 300,000 deaths in children <5 years in 2015. Pneumococcal
conjugate vaccine (PCV) introduction in South Africa has been associated with changes in
invasive pneumococcal disease (IPD) risk groups and emerging serotypes. Serotype 1
pneumococcal disease is highly invasive, fluctuates annually but tends to have lower
mortality and antibiotic-resistance than other serotypes. Paediatric antiretroviral treatment
(ART) and HIV prevention of mother-to-child transmission programme improvements in
South Africa have resulted in a growing number of HIV-exposed-uninfected (HEU) children
who have higher rates of infectious diseases than unexposed children. It is important to
identify risk groups changes with new interventions and define IPD burden pre- and post-
PCV introduction in developing countries.
OBJECTIVES
In South Africa we aimed to estimate severe pneumococcal disease burden in the pre- (2005-
2008) and post-PCV era (2013) amongst HIV-infected (HI) and HIV-uninfected (HU)
children <5 years of age; describe the epidemiology of serotype 1 IPD in all age groups from
2003 to 2013; describe the epidemiology of IPD in HEU children <1 year of age from 2009
to 2013 and the risk factors related to IPD in HI and HU children post-PCV introduction
(2010 to 2012). All analyses included PCV introduction impact.
METHODS
A model using national laboratory-based IPD surveillance data as the baseline was used to
determine the total burden of severe hospitalised pneumococcal disease and related mortality
in South Africa in children aged <5 years. Adjustments were made for specimen-taking
practices and care seeking differences. Vaccine probe studies were used to calculate nonbacteraemic
pneumococcal pneumonia case numbers. Observed case fatality ratios were
applied to estimated case numbers to determine pneumococcal death numbers.
All patients with laboratory-confirmed IPD were included in the serotype 1 analysis. We
calculated incidence rates, determined factors associated with serotype 1 disease and
conducted a space-time analysis using SaTScan with a Bernoulli model for comparison.
Maps to visualise serotype 1 clusters were generated using ArcGIS.
Surveillance data was used to compare IPD incidence and mortality in HEU, HIVunexposed-
uninfected (HUU) and HI infants. Factors associated with HIV status were
compared using a multinomial regression model and logistic regression for mortality factors.
A matched case-control study nested within the surveillance programme was used to
determine risk factors associated with IPD in HU and HI children aged <5 years. Data was
analysed using conditional logistic regression.
RESULTS
In the pre-vaccine era (2005-2008) in South Africa, roughly 196,100 (148,000-251,000) cases
of severe pneumococcal disease were estimated annually in children aged <5 years, an
incidence of 3799/100,000; the rate was reduced by 67% in 2013, likely due to PCV and
other interventions. In addition 8600 (7000-10220) pneumococcal-related annual deaths were
estimated pre-vaccine and 3600 in 2013, a rate difference of 99/100,000 child-years.
Over an 11-year period two clusters (2003-2004 and 2008-2012) of serotype 1 infection were
detected in all age groups with reductions in incidence noted in 2013. Among children aged
<5 years, those with serotype 1 IPD had shorter hospital stays, fewer penicillinnonsusceptible
cases (adjusted odds ratio (aOR) 0.02, 95% confidence interval (CI) 0.01–
0.05), lower HIV prevalence (aOR 0.19, 95% CI 0.12–0.31) and lower in-hospital death rates
(aOR 0.38, 95% CI 0.19–0.76) than children with non-serotype 1 IPD.
The incidence of IPD was greatest in HI infants (272-654/100,000), then HEU infants (33-
88/100,000) and HUU infants (18-28/100,000). Young HEU infants (37% [59/175]) were
more likely to die than HUU infants (32% [51/228]; adjusted relative risk ratio, 1.76, 95% CI
1.09–2.85]). On case-control analysis a number of factors were shown to be associated with
an increased risk of IPD in the post-PCV period. In HU children these factors included
underlying medical conditions (aOR = 1.99, 95% CI 1.22–3.22), attending day care (aOR =
1.58, 95% CI 1.01–2.47) or having been exposed to HIV perinatally (aOR = 1.62, 95% CI
1.10–2.37), while PCV vaccination reduced the odds of IPD (aOR = 0.67, 95% CI 0.46–
0.99). Predisposing factors in HI children included malnutrition (aOR = 2.68, 95% CI 1.40–
5.14) and recent tuberculosis (aOR = 5.12, 95% CI 1.69–15.50), while current ART reduced
the odds of IPD (aOR = 0.13, 95% CI 0.05–0.38).
CONCLUSION
Pneumococcal disease represents a major public health burden in young children in South
Africa. PCV and other HIV-associated interventions resulted in a significant reduction in
both invasive disease and non-bacteraemic pneumonia. Serotype 1 IPD has distinctive
clinical features with temporal decreases noted post-PCV13 introduction. With improvements
in interventions to prevent and treat HIV, a resultant growing HEU infant population has
been observed with an increased risk of IPD compared with HUU children. Risk factors
related to socio-economic conditions and intense exposure to infection continues to be
important causes of IPD in children. A full understanding of PCV impact on pneumococcal
disease burden is needed to support ongoing national policy decisions on PCV use.
Description
A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand,
Johannesburg, in fulfilment of the requirements for the degree of Doctor of Philosophy.
Johannesburg, 11 November 2016