The impact of pneumococcal 7- and 13- valent conjugate vaccines on the molecular epidemiology of invasive streptococcus pneumoniae in South Africa
No Thumbnail Available
Date
2018
Authors
Ndlangisa, Kedibone Maria
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Introduction
There are at least 98 known pneumococcal serotypes; however, the majority of disease is caused
by approximately 20 serotypes. The pneumococcal conjugate vaccine (PCV) significantly
decreased pneumococcal disease however following its introduction some countries reported
increases in non-PCV serotypes. Cases of capsular switching have been reported however, in
most cases increases were driven by pre-existing clones that were circulating prior to PCV
introduction. The 7-valent pneumococcal conjugate vaccine (PCV7) was introduced in South
Africa in 2009 and was replaced with the higher valency PCV13 in 2011. Pneumococcal disease
is usually caused by a single serotype and disease by more than one serotype is rarely reported.
Some pneumococci do not react with commercially available antisera used for serotyping and
are thus regarded as non-serotypeable (NT). NT pneumococci are commonly isolated during
carriage studies and very rarely cause invasive disease.
Materials and methods
Patient information and bacterial isolates were obtained through the national, laboratory-based
surveillance for invasive pneumococcal disease (IPD) in South Africa, coordinated by the
Group for Enteric, Respiratory and Meningeal Disease Surveillance in South Africa (GERMS
SA). Genotypes (sequence types) of isolates that caused IPD in 2005 through 2013 prior and
during routine use of PCV7 and PCV13 were compared. Sequence type diversity by serotype
was assessed using Simpson’s index of diversity. For dual serotype IPD (co-infection with two
serotypes) cases identified in 2005 through 2014, factors associated with dual serotype IPD
were assess and genomes of isolate pairs from co-infected individuals were sequenced. In two
cases of IPD where a non-serotypeable and a serotypeable isolate were co-detected during
routine serotyping, genomes of isolate pairs were compared.
Results
From 2005 through 2013, 36,267 IPD cases were reported to GERMS-SA of which 70%
(25,508/36,267) had viable isolates available for further characterization. The average incidence
of PCV serotypes declined 49% (95% confidence interval (CI), -44% to -54%) from
18.6/100,000 to 4.0/100,000 population during the pre-PCV (2005-2008) and PCV13 period
(2011-3013), among infants and young children (<5 years old). For this age group, while no
change occurred among all non-PCV serotypes combined, average 35B incidence increased
76% (95% CI, 26% to 94%) from 0.078/100,000 population in the pre-PCV period to
0.32/100,000 population in the PCV13 period. Among older children and adults (≥5 years old),
declines occurred in average incidences of PCV serotypes from 7.2/100,000 to 2.1/100,000
population and in non-PCV serotypes from 3.0/100,000 to 1.5/100,000 population between the
pre-PCV and PCV13 period. Similar genotypes circulated among isolates from infants and
young children and among older children and adults. For most serotypes, sequence type
diversity and predominant genotypes did not change between the pre-PCV and PCV13 period.
While globally disseminated clones were common among some serotypes (e.g., serotype 1
[clonal complex (CC) 217] and 14 [CC230]), some were represented mainly by clonal
complexes rarely reported elsewhere (e.g., serotype 3 [CC458, 60%] and 19A [CC2062, 83%]).
Increase in non-PCV serotypes appeared to be due to expansion of pre-existing clones. For
example, CC172, a genotype previously common among multiresistant serotypes 6A, 19A, 19F
and 23F, was represented by 13% (1/8) and 56% (9/16) of 35B isolates from infants and young
children during the pre-PCV and PCV13 period, respectively.
IPD co-infection with two serotypes was identified in 33 IPD patients in 2005-2014. For 30
(91%) of the 33 patients, one or both isolates was a PCV13 serotype. Dual serotype IPD was
associated with infants and young children ˂5 years of age [adjusted odds ratio (aOR), 4.7, 95%
CI, 1.8-11.7], underlying illness (other than HIV) (aOR, 2.8; CI, 1.1-6.6) and death (aOR, 2.5;
CI, 1.08-6.09). For each co-infecting pair, isolates were genotypically unrelated and their
genotypes were common among isolates of the same serotype in South Africa. Of 701 accessory
genes identified among dual serotype IPD isolates, four were common between isolate pairs.
A serotype 1 (isolate not available for further characterization) and 18C isolate were each co
detected with a non-serotypeable (NT) isolate in 2009 (case A) and 2010 (case B). Comparison
of a non-serotypeable case A isolate with a serotype 1 genome (isolate of the same genotype as
the NT case A isolate) revealed only the presence of four genes (rmlA, B, C and D) in the
capsular locus, all other capsular genes were absent. Nonetheless it had a ST associated with
serotype 1 (ST217 and ribosomal ST3462) and its core genome clustered with other ST217
isolates. The case B non-serotypeable isolate had all serotype 18C capsular genes except for
variation in two of the capsular genes (wchA and wze), compared to the 18C isolate. Both case
B isolates were ST9817 and their core genomes were identical.
Conclusions
In this study, genotypes circulating in South Africa before and during PCV use did not differ.
Genotypes rarely reported in other parts of the world but common among some of our serotypes
highlight the importance of these data. Genome comparison of 35B isolates belonging to CC172
and CC172 isolates expressing other capsular types may provide some information on the origin
and evolution of the 35B/CC172 clone, which appears to have increased during the PCV era in
South Africa. The association of dual serotypes with death warrants increased awareness of IPD
co-infection caused by two or more serotypes. Although IPD due to non-serotypeable
pneumococci is currently rare and there is no reported evidence of adaption from PCV serotypes
to non-serotypeable pneumococci thus far, the ability of pneumococci to alter capsule
production is a concern and therefore non-serotypeable pneumococci should be monitored.
Description
A thesis submitted to the Faculty of Health Sciences, University of the Witwatersrand,
Johannesburg, in fulfillment of the requirements for the degree of Doctor of Philosophy.
Johannesburg, November 2018