Aus dem Institut für Medizinische Mikrobiologie
(Direktor Universitätsprofessor Dr. med. Mathias Hornef)
Spatial Epidemiology of Invasive Pneumococcal Disease Isolates
from Children under Six in Germany
Von der Medizinischen Fakultät der Rheinisch-Westfälischen Technischen Hochschule Aachen
zur Erlangung des akademischen Grades einer Doktorin der Theoretischen Medizin
genehmigte Dissertation
vorgelegt von
Stephanie Russo Perniciaro geb. Russo
aus
Champaign, Illinois (Vereinigte Staaten von Amerika)
Berichter: Herr Privatdozent
Dr. rer. nat. Mark van der Linden Frau Privatdozentin
Dr. med. Irene Burckhardt Tag der mündlichen Prüfung: .....28.05.2019............. Diese Dissertation ist auf den Internetseiten der Universitätsbibliothek online verfügbar.
NOTICE
Parts of this dissertation are pre-published in the following peer-reviewed journal
article:
Perniciaro S., Imöhl M., Fitzner C., van der Linden M. Regional variations in serotype
distribution and vaccination status in children under six years of age with invasive
pneumococcal disease in Germany. PLoS ONE 14(1):e0210278.
Dedication
To the boo.
Table of Contents
1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
1.1 Invasive Pneumococcal Disease . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Spatial Analysis in Epidemiology . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 The German National Reference Center for Streptococci . . . . . . 3
2. Objectives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
3. Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
3.1 Microbiological Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5
3.1.1The identification of Streptococcus pneumoniae . . . . . . .5
3.1.2 Serotyping Streptococcus pneumoniae . . . . . . . . . . . . . . 6
3.2 Statistical Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
3.2.1 IPD case definition and study inclusion criteria . . . . . . . 7
3.2.2 Geographic analysis groups . . . . . . . . . . . . . . . . . . . . . . .7
3.2.3 Vaccination status definitions . . . . . . . . . . . . . . . . . . . . . 8
3.2.4 Statistical analyses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .10
4.1 Invasive Pneumococcal Disease Rates . . . . . . . . . . . . . . . . . . . .10
4.2 Serotype Distribution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11
4.2.1 Vaccine serotypes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
4.2.2 Non-vaccine serotypes . . . . . . . . . . . . . . . . . . . . . . . . . . 13
5. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15
6. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
7. Literature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .19
8. Appendices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Figure 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29
Figure 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .30
Figure 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .31
Figure 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .32
Figure 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33
Figure 6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34
Figure 7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .35
Figure 8 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .36
Figure 9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .37
Table 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38
Table 2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .39
Table 3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .40
Table 4 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .42
Table 5 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .44
9. List of Publications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .48
10. Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49
Listing of Abbreviations
ATCC American Type Culture Collection
CI Confidence Interval
CLSI Clinical & Laboratory Standards Institute
CO2 Carbon Dioxide
GNRCS German National Reference Center for Streptococci
IPD Invasive Pneumococcal Disease
mL Milliliter, 1 x 10-3 liter
µL Microliter, 1 x 10-6 liter
mm Millimeter, 1 x 10-3 meter
µg Microgram, 1 x 10-6 gram
NVT Non-Vaccine Type
OR Odds Ratio
PCV7 Seven-valent Pneumococcal Conjugate Vaccine (serotypes 4, 6B, 9V, 14, 18C, 19F, 23F)
PCV10 Ten-valent Pneumococcal Conjugate Vaccine (serotypes 4, 6B, 9V, 14, 18C, 19F, 23F, 1, 5, 7F)
PCV13 Thirteen-valent Pneumococcal Conjugate Vaccine (serotypes 4, 6B, 9V, 14, 18C, 19F, 23F, 1, 5, 7F, 3, 6A, 19A)
PCV15 Fifteen-valent Pneumococcal Conjugate Vaccine (serotypes 4, 6B, 9V, 14, 18C, 19F, 23F, 1, 5, 7F, 3, 6A, 19A, 22F, 33F)
PCV20 Twenty-valent Pneumococcal Conjugate Vaccine (serotypes 4, 6B, 9V, 14, 18C, 19F, 23F, 1, 5, 7F, 3, 6A, 19A, 22F, 33F, 15B, 12F, 10A, 11A, 8)
SSI Statens Serum Institute
VE Vaccine Effectiveness
VT Vaccine Type
1
1. Introduction
1.1 Invasive Pneumococcal Disease in Germany Invasive pneumococcal
disease (IPD) is responsible for nearly half a million deaths per year in children under
five, and also represents 5% of all-cause child mortality (World Health Organization
2013). IPD is caused by the bacterium Streptococcus pneumoniae, or pneumococcus, a
Gram-positive diplococcus whose major virulence factor is a polysaccharide capsule,
which helps the bacteria evade phagocytosis. The polysaccharide capsule antigens
provide the basis for classifying pneumococci by serotype (Frankel et al. 1996). To date,
approximately 94 serotypes have been identified, with ongoing debate (Burton et al.
2016, Geno et al. 2017, Manna et al. 2018) about structural and genetic differences
accounting for the necessary approximation.
Pneumococci are common nasopharyngeal flora, as well as a frequent cause of
noninvasive pediatric infections, particularly otitis media (Somech et al. 2011). When
pneumococci are involved in invasive infections, a majority of cases are confined to three
clinical presentations: meningitis, bacteremic pneumonia, and sepsis. Mortality rates for
IPD depend on the presentation and population in which the disease is occurring, and so
reports vary from 10 to 50% (Cartwright 2002). In addition to being deadly, IPD can also
result in long-term disability, including paralysis, deafness, and amputation (Imöhl et al.
2015). Disease surveillance on pediatric IPD is ongoing throughout the world, with a
notable uptick following the development of pneumococcal conjugate vaccines (PCVs),
which are a common component of childhood immunization programs (World Health
Organization 2017).
Three PCV products have been licensed in Germany: PCV7, approved in July 2006,
followed by PCV10 and PCV13, both of which were approved in 2009, with the latter
replacing PCV7. PCV13 currently has the vast market share of infant pneumococcal
vaccination in Germany (Weiss et al. 2015). The choice of vaccine products, as well as
the decision to vaccinate, is made by the parents (in consultation with the pediatrician).
The Standing Committee on Vaccinations of the Robert Koch Institute issued a
recommendation for all infants to receive the pneumococcal conjugate vaccine in 2006 (a
recommendation for premature, chronically ill, or immunodeficient children was made in
2001) (Robert Koch Institute 2006, Robert Koch Institute 2001). The original vaccination
recommendations were for a 3+1 dosing schedule (third, and fourth, and fifth months of
life, with a booster at 11-14 months), but in August 2015, the committee announced a
2
change to a 2+1 dosing schedule (third and fifth months of life, with a booster at 11-14
months) (Robert Koch Institute 2015).
Following the introduction of the childhood vaccination program in Germany, a complete
shift in IPD serotype distribution has been described (van der Linden et al. 2016).
Vaccine-type serotypes in the country as a whole have decreased significantly, and some
herd protection effects, in which vaccine-type IPD decreases even in populations without
direct protection, like unvaccinated children and adults, have been reported as well (van
der Linden et al. 2015).
1.2 Spatial Analysis in Epidemiology Typical methods in spatial analysis focus
on distances (which consider the relationship between points on a map) and areas (which
focus on counts in sub-regions of the study area). In epidemiology, the points or counts
typically come from cases of a disease or health outcome of interest, referred to
generically as “events”. The pattern between these events can either be randomly
distributed or clustered (Gatrell et al. 1996). These clusters can then be described in
relation to other disease-relevant variables, such as vaccination uptake for vaccine-
preventable diseases (Thanh et al. 2016), local air pollution measurements for asthma
(Nnoli et al. 2018), or neighborhood safety measurements for incidents of gun violence
(Walker et al. 2014, Larsen et al. 2017).
Most epidemiological studies using spatial analysis have either tracked acute outbreaks
of highly-contagious infectious diseases, such as Zika virus, malaria, or H5N1 influenza
(Zambrana et al. 2018, Odhiambo et al. 2018, Souris et al. 2010), or long-term
environmental exposure data relating to chronic diseases, ranging from childhood cancer
to Type II diabetes mellitus to Parkinson’s disease (Oliveira Friestino et al. 2018, Zhang
et al. 2018, Shin et al. 2018). With improvements in computing power and data collection,
spatial analyses are increasingly being included as a component of non-epidemic, or
endemic, infectious disease surveillance (Lee et al. 2018), though few studies from spatial
analysis of IPD surveillance data exist (Gjini 2017, Amrine-Madsen et al. 2008).
For the surveillance of IPD in Germany, spatial analysis has the potential to identify
geographic areas with lower vaccine uptake or higher incidence of disease, which can
signal to public health agencies and health educators that more interventions would be
beneficial in these areas. In addition, applying spatial analysis to the German National
3
Reference Center for Streptococci (GNRCS) database allows for a finer focus on serotype
distribution and antibiotic resistance trends, which can help identify the origin of resistant
or widespread clones.
1.3 The German National Reference Center for Streptococci
The sample collection of the GNRCS is generated by voluntarily-participating hospital
and commercial microbiological laboratories, which send in bacterial isolates from cases
of IPD and a questionnaire containing limited, non-identifying clinical and demographic
information about the patient. The GNRCS performs microbiological and genetic testing
on the isolates (methods relevant to this study are described in detail in section 3.1) and
returns a detailed report to the laboratory of origin for each isolate. All data from IPD
isolates are recorded in a database housed at the GNRCS, which currently contains
information from over 83,000 streptococcal isolates from around the world.
The GNRCS has been collecting invasive pneumococcal isolates from children since
1997, and the surveillance program has been enhanced several times over the years, with
outreach from GNRCS personnel to the staff at microbiological laboratories. In North-
Rhine Westphalia, surveillance enhancement took place in 2001, and in Bavaria and
Saxony, surveillance enhancement took place in 2006, after which the GNRCS
surveillance area covered approximately 32 million residents of Germany. In 2010, the
remaining 13 federal states were also brought into enhanced surveillance, resulting in the
whole population of Germany (around 80 million people) being included in the enhanced
surveillance area (Rückinger et al. 2008, van der Linden et al. 2012).
Coinciding with the universal infant PCV recommendation, the Robert Koch Institute
also founded the online reporting system Pneumoweb in 2007, in which descriptions of
the invasive isolates are sent in by clinical and microbiological laboratories. The
laboratories can then print out this data and send it in with their isolates, in lieu of
completing the GNRCS questionnaire. In the most recent year of data collection, 1 July
2017 to 30 June 2018, the GNRCS received over 4700 samples, 3,281 of which were
from cases of IPD in Germany, of which, in turn, 175 were from children under 16.
4
2. Objectives
The expected output of this project is a statistical analysis comparing the estimated
incidence and serotype distribution of invasive pneumococcal disease occurring in
children under six years of age between population-normalized regions of Germany.
Of particular public health interest, this analysis will allow for regional comparisons of
the changes in the distribution of vaccine-type serotypes following the implementation of
the childhood immunization program in 2006, which will in turn provide insights into
vaccine impact and contribute to the ongoing, hotly-contested issue of pneumococcal
vaccine schedule recommendations.
The value of this analysis lies in the potential to identify differences in regional disease
rates, compositions, and contributing factors thereto, and to develop strategies to reduce
disease in problem areas.
5
3. Materials and Methods
3.1 Microbiological Methods
3.1.1 The identification of Streptococcus pneumoniae Streptococcus
pneumoniae strains, also called pneumococci, can be differentiated from other alpha-
hemolytic (viridans) streptococci by morphological and chemical methods. While some
pneumococci are slimy and shiny, others show a small depression at the center of their
circular colonies. Thus, the appearance of the colonies can vary, and so further
confirmation is often needed. The GNRCS protocols (Isenberg 1992) for the
identification of Streptococcus pneumoniae rely on two chemical test criteria: optochin
sensibility and bile solubility. Pneumococci are generally optochin-susceptible and bile-
soluble. In Germany, very few strains are optochin-resistant, but even these few strains
remain bile-soluble, so the combination of both techniques offers a stable, reliable
diagnosis.
To determine optochin-susceptibility, alpha-hemolytic streptococci are grown on sheep
blood agar plates and incubated for 24 hours with 5% CO2. From this plate, a sterile swab
is used to prepare a suspension in a 0.9% saline solution. This suspension is then
transferred to a separate sheep blood agar plate to ensure smooth and even inoculation.
An optochin disk (5µg, oxoid, DD0001B) is placed in the middle of the plate, and the
plate is incubated overnight at 37°C and 5% CO2. An inhibition zone (a generally circular
area of no bacterial growth surrounding the disk) of > 14 mm indicates a pneumococcus.
A complete lack of an inhibition zone indicates other viridans streptococci. An inhibition
zone of < 14 mm indicates questionable results and these samples also require bile-
solubility testing.
To determine bile solubility, three colonies from the original plating of alpha-hemolytic
streptococci are suspended in 1 mL 0.9% saline solution, which is distributed evenly
between two vials, one labeled “Test” and the other labeled “Control”. Four drops of 10%
bile (deoxycholic acid) solution are then added to the “Test” vial. Four drops of 0.9%
saline solution are added to the “Control” vial. Both vials are gently mixed by vortex
mixer and incubated for two hours at 37o C in the absence of CO2. The clarity of the
resultant solutions is used to indicate a positive or negative reaction. When the “Test” vial
is less cloudy than the “Control” vial, this is considered a positive reaction and indicates
that the sample is Streptococcus pneumoniae. When both the “Test” and “Control” vials
remain cloudy, this is a negative result and indicates that the sample is a different alpha-
6
hemolytic Streptococcus species. All reagents used in the bile-solubility test are checked
weekly with control strains Streptococcus pneumoniae ATCC 46919 as a positive control,
and Streptococcus salivarius ATCC 13419 as a negative control.
All alpha-hemolytic Streptococci that are both bile-soluble and optochin-susceptible are
recorded as Streptococcus pneumoniae. All samples that are not bile-soluble and are
optochin-resistant are considered to be other alpha-hemolytic Streptococci and are
separated for further identification tests. Any samples with questionable results (for
example, optochin-resistant and bile-soluble) are not considered a positive identification
of Streptococcus pneumoniae, and these samples are also separated for further
identification tests.
3.1.2 Serotyping Streptococcus pneumoniae Samples that have been
positively identified as Streptococcus pneumoniae are then further differentiated by their
type of capsular polysaccharide, which is their serotype. The GNRCS protocol relies on
the Neufeld test to determine serotype (Neufeld 1902, Lund and Henrichsen 1978). The
Neufeld test is based on the optical appearance of swelling under the microscope, caused
by a reaction of the polysaccharide capsule of Streptococcus pneumoniae when the strains
are exposed to specific antigens (antibodies derived from rabbits) present in sera.
From the original plating of a sample positively identified as Streptococcus pneumoniae,
an inoculation loop is passed two or three times over the plate, collecting several colonies,
and placed directly into an Eppendorf tube containing 300 µL 0.9% saline solution and
one drop of methylene blue dye, creating a bacterial suspension.
Then, 4 µL of the bacterial suspension is mixed with 4 µL of each applicable test
antiserum on a single glass microscope slide. The resulting mixture on the slide is then
covered with a cover slip the reactions are observed under the microscope.
The Neufeld test follows a fixed scheme for the identification of Streptococcus
pneumoniae. Each sample is tested by omni sera, pool antisera (sera A-I and P-T), type
sera/group sera (sera 1-48), and factor sera as necessary to complete the serotype
identification. Detailed descriptions of the sera reactions necessary to identify particular
serotypes are excerpted as Figure 1 from pages 27-29 of Streptococcus pneumoniae:
Textbook in Diagnosis, Serotyping, Virulence Factors and Enzyme-linked
7
Immunosorbent Assay (ELISA) for Measuring Pneumococcal Antibodies (Skovsted
2017).
As an example, a sample is first tested with omni serum, which comprises all known
serotypes, with a positive result. The sample is then tested against pool sera A, B, C, D,
E, F, G, H, I, P, Q, R, S, and T. Pool sera C and P are positive, so the sample is from
serogroup 7. The sample is then tested with factor sera 7b, 7c, 7e, and 7f. Factor serum
7e is positive, so the sample is serotype 7B.
Very rarely, mutation events cause some strains to have damaged capsules or no capsules
at all, and so no reactions occur during the Neufeld test. These strains are separated for
further identification methods. All reactions and results are documented on a laboratory
report form, which is included in each sample’s record when it is scanned and stored at
the GNRCS.
3.2 Statistical methods
3.2.1 IPD case definition and study inclusion criteria IPD cases were
limited to isolates in the GNRCS collection from invasive infections confirmed to be
Streptococcus pneumoniae, see Section 3.1.1. An invasive infection was defined as the
isolation of Streptococcus pneumoniae from a normally-sterile site, such as the blood or
cerebrospinal fluid (European Union Commission 2018). In this study, isolates from
children residing in Germany who were at least 90 days old but younger than six years of
age, and who were born after 1 July 2007 (one year after the onset of the general infant
vaccination recommendation) were included in the analysis to ensure that all children in
the post-vaccination period were within the age window to have been vaccinated in
infancy. A flowchart describing IPD cases included in the study analysis and the study
population breakdown can be found in Figure 2.
3.2.2 Geographic analysis groups The 16 federal states of Germany were
combined into four geographic analysis groups of between one and seven federal states,
based on the population of children under six residing in districts which had ever sent a
sample to the GNRCS during the study period, shown in Figure 3. Each geographic group
corresponds to approximately 700,000 children.
8
3.2.3 Vaccination status definitions Information on vaccination status was
either included on the GNRCS questionnaire that accompanies the isolate, or retrieved
through telephone follow-ups by the GNRCS staff. No identifying information was stored
at any time by the GNRCS. Vaccination status was separated as follows: “unknown”
(excluded from analysis), “unvaccinated” (child had not received any doses of vaccine at
time of infection), “vaccinated” (child had received at least one dose of vaccine at time
of infection), and “correctly-vaccinated” (child had received the age-appropriate number
of vaccine doses within 14 days of the prescribed timeframe). Children who received
mixed doses (i.e., one dose of PCV7 and one dose of PCV13) were excluded from
analysis. That is to say, children in the respective analysis groups (PCV7, PCV10, and
PCV13) were either vaccinated exclusively with the specified vaccine or wholly
unvaccinated.
The pneumococcal seasons July 2007 through June 2015 are covered by the same vaccine
schedule recommendation (three primary doses plus one booster dose). Later seasons
were excluded due to the change in vaccine schedule recommendation (to a 2+1
schedule). Vaccination age cohorts are defined as follows: at least one dose (≥90 days
old), post-primary series (between 150 and 449 days old), and post-booster (>449 days
old).
3.2.4 Statistical analyses Traditional spatial analysis methods, such as
cluster analysis and spatial regression, were not appropriate for this dataset for two
reasons. First, the sample sizes of IPD cases when separated by geographic district were
very low, which results in highly variable incidence estimates and unreliable models, see
Figure 4. Second, the GNRCS collection is generated from voluntary submissions by
microbiological laboratories, so there is no guarantee that all IPD cases are captured. In
fact, an audit of laboratories and hospitals in North Rhine-Westphalia showed that closer
to 50% of IPD cases are sent to the GNRCS (Reinert et al. 2005). Each geographic district,
therefore, does not have equal likelihood of submitting a sample to the GNRCS, even
when adjusted for population and age. Because of the poor fit with traditional spatial
analysis methods, the data were grouped by geographic analysis groups as described in
Section 3.2.2. Individual federal states provided stable incidence levels overall, but when
analyzing serotype distribution trends, sample sizes were again too small to ensure model
stability, see Figure 5.
9
Multivariate models using Firth’s bias-reduced logistic regression (Heinze 2006) were
designed with invasive infection from a specified serotype or specified serotypes fixed as
the outcome variable and vaccination status, age of child, year of infection (divided in
pneumococcal seasons from 1 July to 30 June), geographic group of residence, residence
in the former East German federal states (Brandenburg, Mecklenburg-Western
Pomerania, Saxony, Saxony-Anhalt, Thuringia) versus the former West German federal
states (Baden-Württemberg, Bavaria, Bremen, Hesse, Hamburg, Lower Saxony, North
Rhine-Westphalia, Rhineland-Palatinate, Schleswig-Holstein, Saarland) as predictor
variables. Samples from Berlin were excluded in East Germany versus West Germany
comparisons. Regional district-level demographic data from the 2011 census for median
household size, per capita income, percentage of adults without secondary education,
percentage of unemployment, and percentage of children under six years of age enrolled
in public daycare were added as additional predictor variables. Models were also run with
the vaccination status as the outcome variable and the demographic and geographic
variables as the predictor variables.
Univariate models were constructed first, for all vaccine type (VT) serotypes in groups,
and for single serotypes with at least 10 occurrences in the study period (Vittinghoff and
McCulloch 2006). Predictor variables with P ≤0.20 were selected for the multivariate
models. Multivariate models were constructed using forward stepwise logistic regression
and McFadden’s pseudo R2 to optimize model fit. Multiple logistic regression models
were run separately for each age cohort and each vaccine type. For multiple regression
models, 95% confidence intervals (95%CIs) of the odds ratio (OR) that did not cross one
were considered statistically significant.
Differences in the proportion of serotype distribution and differences in the proportion of
vaccinated versus unvaccinated cases across the geographic groups were measured with
Fisher’s exact test, with a Dunn-Šidák correction for multiple testing, yielding a
significance threshold of 0.002.
All analyses were performed with R (version 3.4.0, The R Foundation for Statistical
Computing, 2015). Map graphics were created with QGIS 2.18.10 (Quantum Geographic
Information System 2017, an Open Source Geospatial Foundation Project) using
shapefiles from the Global Administrative Areas database; bar graphs and tables were
made with Microsoft Excel 2016 and Microsoft Publisher 2016.
10
4. Results
For the study population, the GNRCS received samples from 162 different laboratories
over the study period. When separated by geographic group, case numbers ranged
between 3 (the 2007-2008 season in the northeastern states) and 33 (the 2014-2015 season
in the southern states) during the post-vaccination period. The median age of infection
was 404 days (13 months).
A clear vaccination status was established for 71% (591/832) of IPD cases in children
over 90 days and less than six years old born after the onset of vaccination. Of these
children, 137 (23.2%) were unvaccinated and 454 (76.8%) had received at least one dose
of any PCV. Among cases for which the vaccination status was known, unvaccinated
children ranged from 19.5% in North Rhine-Westphalia to 29.2% in the southern states;
vaccinated children ranged from 70.8% in the southern states to 80.5% in North Rhine-
Westphalia. Vaccination status in Germany as a whole, by geographic analysis group, and
by former political division can be seen in Table 1. Multivariate models describing
variables associated with vaccination status can be found in Table 2. For the PCV7 group,
being unvaccinated was associated with the year of infection, and being correctly
vaccinated was positively associated with the percentage of adults without a secondary
education. No demographic variables reached statistical significance in the PCV10 group,
and in the PCV13 group, increasing per capita income was associated with being
unvaccinated, and in the post-booster cohort, residence in the southern federal states was
associated with being unvaccinated.
A comparison of the demographic variables from the 2011 census included in the
multivariate logistic regression models (percentage of unemployment, median household
size, annual per capita income, percentage of adults with no secondary education, and
percentage of children enrolled in public daycare), separated by geographic analysis
group and by former political group, can be found in Figure 6.
4.1 Invasive Pneumococcal Disease Rates
As mentioned in section 3.2.4, when examining IPD cases reported to the GNRCS per
100,000 residents under six years of age, the residential district level proved too narrow
of a geographic focus to provide sufficient model stability. The 2013-2014 pneumococcal
season, used as an example for Figure 4 and Figure 5, showed a range from zero to seven
11
cases at the regional district level. However, of the 402 regional districts, 321 (79.9%)
reported zero cases and 57 (14.2% overall, and 70.4% of regional districts reporting any
cases during the 2013-2014 season) reported only one case.
The federal state level was more stable for an overall assessment of reporting rates on a
season-by-season basis. For the 2013-2014 season, each of the 16 federal states reported
at least one case, with a range of one to 29 cases. These case numbers, per 100,000
residents under six years old, yielded case reporting rates of between 1.51 per 100,000
(Berlin) and 5.91 per 100,000 (Brandenburg). However, when looking for differences in
serotype distribution, the case numbers at the federal state level again became too small
for meaningful analysis. With a single simple classification (PCV13 serotypes versus
non-vaccine serotypes), the reported case number in the 2013-2014 season ranged from
zero to eight, with eight (50%) federal states reporting zero PCV13-serotype IPD cases
and an additional four (25%) reporting only one case of PCV13-type IPD. This resulted
in a vaccine-type IPD case per 100,000 residents range of 0 to 2.03 per 100,000, with a
mean value of 0.59 per 100,000.
To have a more reliable estimate of reported IPD rates across Germany, geographic
analysis groups were formed as described in section 3.2.2. Here, cases reported in the
2013-2014 season ranged from 4 to 12, resulting in a range of 0.51 to 1.81 PCV13-type
IPD cases per 100,000 residents, with a mean rate of 0.91 cases per 100,000. The
Germany-wide rate was 0.88 vaccine-type IPD cases reported per 100,000, so the
geographic analysis groups provide a more similar mean rate than the individual federal
states. The sample sizes of grouped and individual serotypes across the post-vaccination
pneumococcal seasons were too low to examine at the regional district or federal state
level, so the regression models were applied only to the geographic analysis groups to
ensure reliable results.
The distribution of IPD samples from children under six years of age with a known
vaccination status across each geographic analysis group per 100,000 residents during the
post-vaccination seasons is shown in Figure 7.
4.2 Serotype Distribution
4.2.1 Vaccine serotypes The persistence of vaccine type serotypes following
the onset of the infant vaccination recommendation is shown in Figure 8. In the PCV7
12
group (n = 244), summarized in Table 3, the OR for invasive infection with a VT serotype
was 6.84 in unvaccinated children for children old enough to have received at least one
dose, while the OR for serotype 19F was 7.17. In the post-primary series cohort (n =110),
unvaccinated children had significantly higher odds of contracting serotype 19F IPD. In
the post-booster cohort (n = 88), odds of VT IPD were significantly higher in
unvaccinated children. In the PCV7 group, in addition to vaccination status, residence in
the northeastern federal states, residence in North Rhine-Westphalia, residence in the
southern federal states, residence in former East Germany, median household size, and
percentage of children enrolled in public daycare were statistically significant predictor
variables for at least one serotype.
The PCV10 group, summarized in Table 4, was smaller (n =149), with many individual
serotypes yielding sample sizes too small to ensure model stability. In children old enough
to have received one dose, VT IPD (OR = 4.52, adjusted for year of infection, median
household size, and residence in the southern federal states) and IPD from the three non-
PCV7 serotypes (OR = 13.35) had higher odds of occurring in unvaccinated children. In
the post-primary series cohort (n = 74), VT IPD was significantly higher (OR = 7.29,
adjusted for per capita income, year of infection, residence in former East Germany, and
percentage of children in public daycare) in vaccinated children. For the PCV10 group,
variables besides vaccination status which reached statistical significance in at least one
of the multivariate models included median household size, year of infection, percentage
of children enrolled in public daycare, and residence in North Rhine-Westphalia, though
many of these also had wide confidence intervals.
In the PCV13 group (n=374), summarized in Table 5, in children old enough to have
received at least one dose, odds of VT IPD were higher in unvaccinated children across
all VT groupings: VT IPD had an OR of 6.21 (adjusted for year of infection, age of child,
residence in the southern federal states, per capita income, and percentage of children
enrolled in public daycare); PCV10 serotypes had an OR of 7.85; PCV7 serotypes had
an OR of 7.65; PCV13non7 types had an OR of 3.03; PCV13non10 types had an OR of
2.13. Among single serotypes, unvaccinated children had higher odds of IPD caused by
serotypes 19F, 6A, 6B, and 7F (OR=3.67, OR= 8.83, OR= 16.05, and OR= 7.48,
respectively).
13
In the post-primary series cohort (n=177), unvaccinated children had significantly higher
odds of VT IPD, PCV7 type IPD, PCV10 type IPD, PCV13non7 IPD, PCV13non10 IPD,
as well as IPD from single serotypes 14 and 19A. Aside from vaccination status, the
following variables reached statistical significance in the multivariate models at least
once: year of infection, age of child, residence in the northeastern federal states,
percentage of unemployed adults, percentage of children enrolled in public daycare,
residence in North Rhine-Westphalia, per capita income, median household size,
residence in the central federal states, and residence in the southern federal states.
4.2.2 Non-vaccine serotypes In the PCV7 group, in children old enough to
have received one dose, residence in the former East Germany was significantly
associated with serotype 19A IPD and serotype 23B IPD, though this association
disappeared in the older age cohorts. In the post-primary series cohort (n =110) of the
PCV7 group correctly-vaccinated children had higher odds of contracting Serotype 12F
IPD.
The smaller PCV10 group still had a few significant associations. In children old enough
to have received one dose, IPD from serotypes 10A and 15B had higher odds (OR = 7.54
and OR = 11.87, respectively) of occurring in correctly-vaccinated children. In the post-
primary series cohort (n = 74), IPD from serotype 15B was significantly associated with
correctly-vaccinated children (OR = 10.10). In the post-booster cohort (n = 46), serotype
3 was significantly associated with correctly-vaccinated children, but the confidence
interval was very wide (OR 79.47, CI: 2.45-15531.60, adjusted for age of child and
percentage of children enrolled in public daycare).
In the PCV13 group (n=374), in children old enough to have received at least one dose
had lower odds of IPD from serotype 24F (OR= 0.36). In the same age cohort, correctly
vaccinated children were associated with infection from serotype 10A (OR=3.23). In the
post-primary series cohort (n=177), unvaccinated children continued to have significantly
lower odds of IPD from serotype 24F. Correctly-vaccinated children in this cohort had
higher odds of getting IPD from serotype 10A. In the post-booster cohort (n=147) of the
PCV13 group, correctly-vaccinated children had higher odds of infection from serotype
15C.
14
The proportion of non-vaccine serotypes significantly increased over the study period
across three of the four geographic groups and in Germany as a whole; however, no single
serotype increased significantly when comparing the 2007-2008 season to the 2014-15
season. Some regional differences can still be seen among non-vaccine serotypes.
Percentages of selected non-vaccine serotypes by geographic analysis group are shown
in Figure 9.
15
5. Discussion
This study shows only the data from children who had IPD, and so making demographic
characterizations about the population of German children as a whole is difficult, as we
do not know the vaccination status and adherence to the recommended vaccination
schedule of children who did not contract IPD. The voluntary nature of IPD reporting in
Germany is a continual hurdle for studies using the GNRCS collection. However, the data
here show a fairly consistent rate of representation across the federal states, taking
population into account, see Figure 5.
The variable most often significantly associated with IPD from particular serotypes in the
141 multivariate models generated in this study was vaccination status (23.4%), followed
by year of infection and demographic variables (each 14.1%), followed by age (9.9%),
with geographic variables showing a significant association the least often (in 8.5% of the
models). These results show that vaccination status is a strong indicator of VT IPD
infection: unvaccinated children had higher odds of VT infections as a whole, and in some
individual VT serotypes as well (serotype 19F in the PCV7 group and serotypes 14, 19A,
19F, 6A, 6B, 7F in the PCV13 group), while correctly-vaccinated children had higher
odds of non-VT IPD (serotype 12F in the PCV7 group, serotypes 10A, 15B and 3 in the
PCV10 group, and serotypes 10A and 15C in the PCV13 group). These results are
consistent with several PCV effectiveness studies in countries with PCV programs
(Balsells et al. 2017, Cho et al. 2017, Camilli et al. 2017) and with serotype distribution
studies that have been conducted in countries yet to implement these programs (Tam et
al. 2017, Singh et al. 2017, Zhang et al. 2017), which collectively emphasize the
importance of vaccination in preventing VT IPD.
Interestingly, no significant differences were found in IPD incidence estimates, serotype
distribution or vaccination status between former east and west federal states in this study,
which is a departure from previous results (Siedler et al. 2005). Extending this analysis
to the full population of Germany may help establish if this trend is only occurring in
young children, or if it can be applied across all age groups.
Some differences between geographic groups are evident, particularly in non-vaccine
serotypes, which may indicate that population-normalized regional analysis is useful for
determining newly ascendant serotypes and identifying pockets of persistent vaccine
serotypes. Upcoming serotypes identified here from the 2014-2015 season include 12F in
16
the central and northeastern states of Germany, which is echoed by studies in France
(Alari et al. 2016, Lepoutre et al. 2015), Brazil (Azevedo et al. 2016), and Israel (Ben-
Shimol et al. 2016). Serotype 38, common in the northeastern states of Germany, was
also seen in the UK (Brueggemann et al. 2003), Finland (Hanage et al. 2005), and
Hungary (Tóthpál et al. 2015). Serotype 15C was on the rise in North Rhine-Westphalia,
the northeastern states, and the southern states, which is corroborated by results from
Canada (Deng et al. 2015), Uruguay (Gabarrot et al. 2014), and South Korea (Cho et al.
2016). Serotype 10A was increasing in North Rhine-Westphalia, the central states, and
the northeastern states, as well as in non-invasive isolates in Japan (Kawaguchiya et al.
2016) and IPD in South America (Azevedo et al. 2016, Gabarrot et al. 2014, Hawkins et
al. 2017). All four of the preceding serotypes appear in a global review of upcoming
serotypes (Balsells et al. 2017).
Serotypes 7F, 38, 19A, 3, and 33F have been identified as having a high invasive capacity
(Yildirim et al. 2017). In the PCV13 group, these serotypes comprised 26.3% of IPD
cases (26.9% in unvaccinated children; 21.7% in correctly-vaccinated children) in
children under six during the most recent four seasons, which supports the importance of
the role of these serotypes in childhood IPD.
While few studies (Amrine-Madsen et al. 2008, Brueggemann et al. 2004) have included
the use of geographic analysis groups in IPD surveillance, the use of postal codes to
approximate socioeconomic status was recently validated (Link-Gelles et al. 2016) and
presents further potential for expanding upon spatial epidemiology studies of IPD in
Germany, particularly for adults (preliminary studies using postal codes in children under
six had sample sizes too low for model stability), which may soon be of greater interest
due to the possible implementation of a PCV recommendation for older adults.
Unvaccinated and incorrectly-vaccinated children represent nearly all (94%, 207/220) VT
cases of known vaccination status in the post-PCV era. While several studies have shown
both the 3+1 and 2+1 administration schedules to be highly effective at preventing VT
IPD in children (Weinberger et al. 2016, Deceuninck et al. 2015), previous work by the
GNRCS (van der Linden et al. 2016) has described troubling laxity in vaccine
administration in Germany. While receipt of any PCV dose has been shown to be better
than no dose at all (Moore et al. 2015), the increased effect of being unvaccinated shown
here for VT IPD in the post-primary series age cohorts of the PCV10 and PCV13 groups
17
(the PCV7 group having lost statistical significance in the post-primary series age cohort)
provides additional evidence (Andrews et al. 2014) that more PCV doses provide better
protection from VT serotypes.
While it is encouraging to see that rates of vaccination in children with IPD are fairly
consistent throughout Germany, with a range from 70.8% in the southern federal states
to 80.5% in North Rhine-Westphalia, the low rates of correctly-vaccinated children (from
13.5% in the southern federal states to 22.7% in the northeastern states) are problematic,
particularly in light of the recent change to the 2+1 schedule (Robert Koch Institute 2015).
Future analyses will examine the effect of a reduced schedule on childhood IPD.
Additionally, the third generation of PCVs is on the horizon (Alderson 2016), such as
PCV15, which includes all of the PCV13 serotypes plus 22F and 33F, and PCV20, which
includes all of the PCV13 serotypes plus 15B, 22F, 33F, 12F, 10A, 11A, and 8. The
anticipated coverage of these new products will continue to affect the serotype
distribution, so it will be of great interest to see if further regional differences in serotype
distribution develop as a result. Continued efforts to improve systemic surveillance IPD
in Germany are also underway, the impact of which will also need to be carefully
monitored.
18
6. Summary
The protective effect of the infant pneumococcal conjugate vaccine recommendation can
be seen in Germany as a whole and in smaller regional groups. Comparisons between
population-normalized geographic regions of Germany show different serotype
distributions after program implementation, particularly in non-vaccine serotypes. The
prior distinct differences in serotype distribution in children between the former East and
former West German federal states have vanished.
Children under six remain a vulnerable group, but the occurrence of VT IPD in children
correctly vaccinated (using a three-dose primary series plus one booster dose) with
PCV13 was low. However, less than one in five children in Germany with IPD were
correctly vaccinated with PCV13 according to the recommended schedule.
For all PCV products used in Germany (PCV7, PCV10, and PCV13), vaccination status
was the most common statistically significant predictor of infection with a particular
serotype: Unvaccinated children old enough to have received at least one dose of vaccine
in the PCV7 group, the PCV10 group, and the PCV13 group had significantly higher odds
of contracting VT IPD.
Pneumococcal conjugate vaccines have had a massive impact on IPD cases in German
children, playing a major role in both risk of infection and in serotype distribution.
Despite a largely uniform population in terms of overall size, rate of vaccination, and
several demographic characteristics, there are regional differences in serotype
distribution, especially in non-vaccine serotypes.
Continued surveillance and better schedule adherence are essential to definitively
establishing the most effective PCV administration schedule, which is particularly
important given the recent change to the vaccine administration schedule in Germany and
the anticipated arrival of the third-generation pneumococcal conjugate vaccine.
19
7. Literature
Alari A., Chaussade H., De Cellés M.D., Le Fouler L., Varon E., Opatowski L., Guillemot
D., Watier L. Impact of pneumococcal conjugate vaccines on pneumococcal
meningitis cases in France between 2001 and 2014: a time series analysis. BMC
Medicine (2016) 14:211 DOI 10.1186/s12916-016-0755-7.
Alderson M.R. Status of research and development of pediatric vaccines for
Streptococcus pneumoniae. Vaccine (2016) 34:2959-2961.
Amrine-Madsen H., Van Eldere J., Mera R.M., Miller L.A., Poupard J.A., Thomas E.S.,
Halsey W.S., Becker J.A., O’Hara F.P. Temporal and Spatial Distribution of Clonal
Complexes of Streptococcus pneumoniae Isolates Resistant to Multiple Classes of
Antibiotics in Belgium, 1997 to 2004. Antimicrobial Agents and Chemotherapy
(2008) 52:9: 3216-3220.
Andrews N.J., Waight P.A., Burbidge P., Pearce E., Roalfe L., Zancolli M., Slack M.,
Ladhani S.N., Miller E., Goldblatt D. Serotype-specific effectiveness and correlates
of protection for the 13-valent pneumococcal conjugate vaccine: a postlicensure
indirect cohort study. The Lancet Infectious Diseases (2014) 14:9:839-846.
Azevedo J., dos Anjos E.S., Cordeiro S.M., dos Santos M.S., Escobar E.C., Lobo P.R.,
Carvalho M.G., Reis M.G., Reis J.N., Campos L.C. Genetic profiles and
antimicrobial resistance of Streptococcus pneumoniae non-PCV10 serotype
isolates recovered from meningitis cases in Salvador, Brazil. Journal of Medical
Microbiology (2016) 65:1164-1170.
Balsells E., Guillot L., Nair H., Kyaw M.H. Serotype distribution of Streptococcus
pneumoniae causing invasive disease in children in the post-PCV era: A systematic
review and meta-analysis. Public Library of Science ONE (2017) 12:5:e0177113.
Ben-Shimol S., Greenberg D., Givon-Lavi N., Schlesinger Y., Miron D., Aviner S.,
Dagan R. on behalf of the Israel Bacteremia and Meningitis Active Surveillance
Group. Impact of PCV7/PCV13 introduction on invasive pneumococcal disease
(IPD) in young children: Comparison between meningitis and non-meningitis IPD.
Vaccine (2016) 34:4543-4550.
Brueggemann A.B., Peto T.E., Crook D.W., Butler J.C., Kristinsson K.G., Spratt B.G.
Temporal and geographic stability of the serogroup-specific invasive disease
potential of Streptococcus pneumoniae in children. Journal of Infectious Diseases
(2004) 190:7:1203-1211.
Brueggemann A.B, Griffiths D.T., Meats E., Peto T., Crook D.W., Spratt B.G. Clonal
Relationships between Invasive and Carriage Streptococcus pneumoniae and
20
Serotype- and Clone-Specific Differences in Invasive Disease Potential. Journal of
Infectious Disease (2003): 187:9:1424-1432.
Burton R.L., Geno K.A., Saad J.S., Nahm M.H. Pneumococcus with the “6E” cps Locus
Produces Serotype 6B Capsular Polysaccharide. Journal of Clinical Microbiology
(2016) 54:4: 967-971.
Camilli R., D’Ambrosio F., Del Grosso M., Pimentel de Araujo F., Caporali M.G., Del
Manso M., Gherardi G., D’Ancona F., Pantosti A., Pneumococcal Surveillance
Group. Impact of pneumococcal conjugate vaccine (PCV7 and PCV13) on
pneumococcal invasive diseases in Italian children and insight into evolution of
pneumococcal population structure. Vaccine (2017) 35:4587-4593.
Cartwright K. Pneumococcal disease in western Europe: burden of disease, antibiotic
resistance and management. European Journal of Pediatrics (2002) 161: 188-195.
Cho E.Y., Choi E.H., Kang J.H., Kim K.H., Kim D.S., Kim Y.J. Ahn Y.M., Eun B.W.,
Oh S.H., Cha S.H., Cho H.Y., Hong Y.J., Kim K.N., Kim N.H., Kim Y.K., Kim J.H.,
Lee H., Lee T., Kim H.M., Lee K.S., Kim C.S., Park S.E., Kim Y.M., Oh C.E., Ma
S.H., Jo D.S., Choi Y.Y., Lee J., Bae G.R., Park O., Park Y.J., Kim E.S., Lee H.J.
Early Changes in the Serotype Distribution of Invasive Pneumococcal Isolates from
Children after the Introduction of Extended-valent Pneumococcal Conjugate
Vaccines in Korea 2011-2013. Journal of Korean Medical Science (2016) 31:1082-
1088.
Cho Y.C., Chiu N.C., Lu C.Y., Huang D.T., Huang F.Y., Chang L.Y., Huang L.M., Chi
H. Redistribution of Streptococcus pneumoniae Serotypes after Nationwide 13-
valent Pneumococcal Conjugate Vaccine Program in Children in Northern Taiwan.
The Pediatric Infectious Disease Journal (2017): DOI
10.1097/INF.0000000000001664
Deceuninck G., De Serres G., Boulianne N., Lefebvre B., De Wals P. Effectiveness of
three pneumococcal conjugate vaccines to prevent invasive pneumococcal disease
in Quebec, Canada. Vaccine (2015) 33:23:2684-2689.
Deng X., Gitanjali A., Memari N., Mackenzie R., MacMullin G., Low D., Pillai D.R.,
Gubbay JB. Analysis of Invasive Pneumococcal Isolates from Children in Ontario,
Canada, 2007-2012. Pediatric Infectious Disease Journal (2015) 34:6:594-598.
European Union Commission. Decision 2018/945/EU on the communicable diseases and
related special health issues to be covered by epidemiological surveillance as well
as relevant case definitions. Official Journal of the European Union (2018) 170/1:
33-34.
21
Frankel R.E., Virata M., Hardalo C., Altice F.L., Friedland G. Invasive pneumococcal
disease: clinical features, serotypes, and antimicrobial resistance patterns in cases
involving patients with and without human immunodeficiency virus infection.
Clinical Infectious Diseases (1996) 23:3:577-584.
Gabarrot G.G., Vegal M.L., Giffoni G.P., Hernández S., Cardinal P. Félix V., Gabastou
J.M., Camou T., Uruguayan SIREVA II Effect of Pneumococcal Conjugate
Vaccination in Uruguay, a Middle-Income Country. Public Library of Science ONE
(2014) 9:11:e112337.
Gatrell A.C., Bailey T.C., Diggle P.J., Rowlingson B.S. Spatial point pattern anaylsis and
its application in geographical epidemiology. Transactions of the Institute of British
Geographers (1996) 21:1:256-274.
Geno K.A., Saad J.S., Nahm M.H. Discovery of novel pneumococcal serotype 35D: a
natural WciG-deficient variant of serotype 35B. Journal of Clinical Microbiology
(2017): DOI 10.1128/JCM.00054-17.
Gjini, E. Geographic variation in pneumococcal vaccine efficacy estimated from dynamic
modeling of epidemiological data post-PCV7. Nature Scientific Reports (2017) 7:
Article 3049.
Hanage W.P., Kaijalainen T.H., Syrjanen R.K., Auranen K., Leinonen M., Mäkelä P.H.,
Spratt B.G. Invasiveness of Serotypes and Clones of Streptococcus pneumoniae
among Children in Finland. Infection and Immunity (2005) 73:1:431-435.
Hawkins P., Mercado E., Chochua S., Castillo M.E., Reyes I., Chaparro E., Gladstone R.,
Bentley S.D., Breiman R.F., Metcalf B.J., Beall B., Ochoa T.J. McGee L. Key
features of invasive pneumococcal isolates recovered in Lima, Peru determined
through whole genome sequencing. International Journal of Medical Microbiology
(2017) 307:415-421.
Heinze G. A comparative investigation of methods for logistic regression with separated
or nearly separated data. Statistics in medicine (2006) 25:24:4216-4226.
Imöhl M., Möller J., Reinert R.R., Perniciaro S., van der Linden M., Aktas O.
Pneumococcal meningitis and vaccine effects in the era of conjugate vaccination:
results of 20 years of nationwide surveillance in Germany. BMC Infectious
Diseases (2016): DOI 10.1186/s12879-015-0787-1.
Isenberg H.D. Clinical Microbiology Procedures Handbook, Vol. I, (1992) American
Society for Microbiology.
Kawaguchiya M., Urushibara N., Aung M.S., Morimoto S., Ito M., Kudo K., Sumi, A.,
Kobayashi N. Emerging non-PCV13 serotypes of noninvasive Streptococcus
22
pneumoniae with macrolide resistance genes in northern Japan. New Microbes and
New Infections (2016) 9:66-72.
Larsen D.A., Lane S., Jennings-Bey T., Haygood-El A., Brundage K., Rubinstein R.A.
Spatio-temporal patterns of gun violence in Syracuse, New York 2009-2015. Public
Library of Science ONE (2017) 12:3:e0173001.
Lee E.C., Asher J.M., Goldlust S., Kraemer J.D., Lawson A.B., Bansal S. Mind the
Scales: Harnessing Spatial Big Data for Infectious Disease Surveillance and
Inference. The Journal of Infectious Diseases (2016) 214(S4): S409-S413.
Lepoutre A., Varon E., Georges S., Dorléans F., Janoir C., Gutmann D., Lévy-Bruhl D.,
the Microbiologists of the Epibac and the ORP Networks. Impact of the
pneumococcal conjugate vaccines on invasive pneumococcal disease in France,
2001-2012. Vaccine (2015) 35: 359-366.
Link-Gelles R., Westreich D., Aiello A.E., Shang N., Weber D.J., Holtzmann C.,
Scherzinger K., Reingold A., Schaffner W., Harrison L.H., Rosen J.B., Petit S.,
Farley M., Thomas A., Eason J., Wigen C., Barnes M., Thomas O., Zansky S., Beall
B., Whitney C.G., Moore M.R. Bias with respect to socioeconomic status: A closer
look at zip code matching in a pneumococcal vaccine effectiveness study. SSM-
Population Health (2016) 2: 587-594.
van der Linden M., Falkenhorst G., Perniciaro S., Fitzner C., Imöhl M. Effectiveness of
Pneumococcal Conjugate Vaccines (PCV7 and PCV13) against Invasive
Pneumococcal Disease among Children under Two Years of Age in Germany.
Public Library of Science ONE (2016) 11:8: e0161257.
van der Linden M., Falkenhorst G., Perniciaro S., Imöhl M. Effects of Infant
Pneumococcal Conjugate Vaccination on Serotype Distribution in Invasive
Pneumococcal Disease among Children and Adults in Germany. Public Library of
Science ONE (2015) 10:7:e0131494.
van der Linden M., Weiß S., Falkenhorst G., Siedler A., Imöhl M, von Kries R. Four
years of universal pneumococcal conjugate infant vaccination in Germany: Impact
on incidence of invasive pneumococcal disease and serotype distribution in
children. Vaccine (2012) 30:5880-5885.
Lund E., and J. Henrichsen. 1978. Labaratory Diagnosis, Serology and Epidemiology of
Streptococcus pneumoniae. Methods in Microbiology 12: 242-262.
Manna S., Ortika B.D., Dunne E.M., Holt K.E., Kama M., Russell F.M., Hinds J., Satzke
C. A novel genetic variant of Streptococcus pneumoniae serotype 11A discovered
in Fiji. Clinical Microbiology and Infection (2018) 24: 428.e1-428.37.
23
Moore M., Link-Gelles R., Schaffner W., Lynfield R., Lexau C., Bennett N.M., Petit S.,
Zansky S.M., Harrison L.H., Reingold A., Miller L., Scherzinger K., Thomas A.,
Farley M.M., Zell E.R., Taylor T.H., Pondo T., Rodgers L., McGee L., Beall B.,
Jorgensen J.H., Whitney C.G. Effect of use of 13-valent pneumococcal conjugate
vaccine in children on invasive pneumococcal disease in children and adults in the
USA: analysis of multisite, population-based surveillance. The Lancet Infectious
Diseases (2015) 15:3:301-309.
Neufeld F. Über die Agglutination der Pneumokokken und über die Theorie der
Agglutination. Zeitschrift für Hygiene und Infektionskrankheiten (1902) 34: 54-72.
Nnoli N.C., Linder S.H., Smith M.A., Gemeinhardt G.L., Zhang K. The combined effect
of ambient ozone exposure and toxic air releases on hospitalization for asthma
among children in Harris County, Texas. International Journal of Environmental
Health Research (2018) 28:4:358-378.
Odhiambo J.N. and Sartorius B. Spatio-temporal modelling assessing the burden of
malaria in affected low and middle-income countries: a scoping review. British
Medical Journal Open (2018) 8:e023071.
Oliveira Friestino J.K., Mendonça D., Oliveira P., Oliveira C.M., de Carvalho Moreira
Filho D. Childhood cancer: incidence and spatial patterns in the city of Campinas,
Brazil, 1996-2005. Salud Colectiva (2018) 14:1: 51-63.
Reinert R.R., Haupts S., van der Linden M., Heeg C., Cil M.Y., Al-Lahham A., Fedson
D.S. Invasive pneumococcal disease in adults in North-Rhine Westphalia,
Germany, 2001-2003. Clinical Microbiology and Infection (2005) 11:12:985-991.
Robert Koch Institute. Mitteilung der Ständigen Impfkommission am RKI:
Wissenschaftliche Begründung zur Änderung der Pneumokokken-Impfempfehlung
für Säuglinge. Epidemiologisches Bulletin (2015) 36: 377-389.
Robert Koch Institute. Mitteilung der Ständigen Impfkommission am Robert Koch-
Institut: Begründung der STIKO-Empfehlungen zur Impfung gegen
Pneumokokken und Meningokokken vom Juli 2006. Epidemiologisches Bulletin
(2006) 31: 255-259.
Robert Koch Institute. Mitteilung der Ständigen Impfkommission am Robert Koch-
Institut: Impfempfehlungen der Ständigen Impfkommission (STIKO) am Robert
Koch-Institut/ Stand: Juli 2001. Epidemiologisches Bulletin (2001) 28: 216-218.
Rückinger S., von Kries R., Reinert R.R., van der Linden M., Siedler A.,. Childhood
invasive pneumococcal disease in Germany between 1997 and 2003: Variability in
24
incidence and serotype distribution in absence of general pneumococcal conjugate
vaccination. Vaccine (2008) 26:3984-3986.
Shin S., Burnett R.T., Kwong J.C., Hystad P., van Donkelaar A., Brook J.R., Copes R.,
Tu K., Goldberg M.S., Villeneuve P.J., Martin R.V., Murray B.J., Wilton A.S.,
Kopp A., Chen H. Effects of ambient air pollution on incident Parkinson’s disease
in Ontario, 2001 to 2013: a population-based cohort study. International Journal of
Epidemiology (2018) 0:0: 1-11.
Siedler A., Reinert R.R., Toschke M., Al-Lahham A., von Kries R., ESPED. Regional
Differences in the Epidemiology of Invasive Pneumococcal Disease in Toddlers in
Germany. The Pediatric Infectious Disease Journal (2005) 24:12: 1114-1115.
Singh J., Sundaresan S., Manoharan A., Shet A. Serotype distribution and antimicrobial
susceptibility pattern in children <5 years with invasive pneumococcal disease in
India- A systematic review. Vaccine (2017) 35:4501-4509.
Skovsted I.C. Streptococcus pneumoniae: Textbook in Diagnosis, Serotyping, Virulence
Factors and Enzyme-linked Immunosorbent Assay (ELISA) for Measuring
Pneumococcal Antibodies (2017) SSI Diagnostica, Oxford BioSystems.
Somech I., Dagan R., Givon-Lavi N., Porat N., Raiz S., Leiberman A., Puterman M. Peled
N., Greenberg D., Leibovitz E. Distribution, dynamics and antibiotics resistance
patterns of Streptococcus pneumoniae serotypes causing acute otitis media in
children in southern Israel during the 10 year period before the introduction of the
7-valent pneumococcal conjugate vaccine. Vaccine (2011) 29:25: 4202-4209.
Souris M., Gonzalez J., Shanmugasundaram J., Corvest V., Kittayapong P. Retrospective
space-time analysis of H5N1 Avian Influenza emergence in Thailand. International
Journal of Health Geographics (2010) 9:3.
Tam P.I., Thielen B.K., Obara S.K., Brearley A.M., Kaizer A.M., Chu H., Janoff E.N.
Childhood pneumococcal disease in Africa- A systematic review and meta-analysis
of incidence, serotype distribution, and antimicrobial susceptibility. Vaccine (2017)
35:1817-1827.
Thanh D.P., Thompson C.N., Rabaa M.A., Sona S., Sopheary S., Kumar V., Moore C.,
Thieu N.T.V., Wijedoru L., Holt K.E., Wong V., Pickard D., Thwaites G.E., Day
N., Dougon G., Turner P., Parry C.M., Baker S. The Molecular and Spatial
Epidemiology of Typhoid Fever in Rural Cambodia. Public Library of Science
Neglected Tropical Diseases (2016) 10:6:e004785.
Tóthpál A., Kardos S., Laub K., Nagy K., Tirczka T., van der Linden M., Dobay O.
Radical serotype rearrangement of carried pneumococci in the first 3 years after
25
intensive vaccination started in Hungary. European Journal of Pediatrics (2015)
174:373-381.
Vittinghoff E. and McCulloch C.E. Relaxing the Rule of Ten Events per Variable in
Logistic and Cox Regression. American Journal of Epidemiology (2006): 165:6:
710-718.
Walker B.B, Schuumann N., Hameed S.M. A GIS-based spatiotemporal analysis of
violent trauma hotspots in Vancouver, Canada: identification, contextualization and
intervention. British Medical Journal Open (2014) 4:e003642.
Weinberger R., van der Linden M., Imöhl M., von Kries R. Vaccine effectiveness of
PCV13 in a 3+1 vaccination schedule. Vaccine (2016) 34:18:2062-2065.
Weiss S., Falkenhorst G., van der Linden M., Imöhl M., von Kries R. Impact of 10- and
13-valent pneumococcal conjugate vaccines on incidence of invasive
pneumococcal disease in children aged under 16 years in Germany, 2009 to 2012.
Eurosurveillance (2015) 20(10):pli=21057.
World Health Organization. Estimated Hib and pneumococcal deaths for children under
5 years of age, 2008: Child mortality estimates due to Hib and pneumococcal
infections. WHO Immunization, Vaccines and Biologicals (2013):
http_//www.who.int/immunization/monitoring_surveillance/burden/estimates/Pne
umo_hib/en/
World Health Organization. Pneumococcal vaccines WHO position paper. Weekly
Epidemiological Record (2012) 14:87: 129-144.
Yildirim I., Hanage W.P., Lipsitch M., Shea K.M., Stevenson A., Finkelstein J., Huang
S.S., Lee G.M., Kleinman K., Pelton S.I. Surveillance of pneumococcal
colonization and invasive pneumococcal disease reveals shift in prevalent carriage
serotypes in Massachusetts’ children to relatively low invasiveness. Vaccine (2017)
35:4002-4009.
Zambrana J.V., Bustos Carrillo F., Burger-Calderon R., Collado D., Sanchez N., Ojeda
S., Carey Monterrey J., Plazaola M., Lopez B., Arguello S., Elizondo D., Aviles
W., Coloma J., Kuan G., Balmaseda A., Gordon A., Harris E. Seroprevalence, risk
factor, and spatial analyses of Zika virus infection after the 2016 epidemic in
Managua, Nicaragua. Proceedings of the National Academy of Science (2018):
DOI 10.1073/pnas.1804672115
Zhang X., Chen X., Gong W. Type 2 diabetes mellitus and neighborhood deprivation
index: a spatial analysis in Zhejiang, China. Journal of Diabetes Investigation
(2018): DOI 10.1111/jdi.12899
26
Zhang X., Tianmei J., Shan W., Xue J., Tao Y., Geng Q., Ding Y., Zhao G., Zhang T.
Characteristics of pediatric invasive pneumococcal diseases and the pneumococcal
isolates in Suzhou, China before introduction of PCV13. Vaccine (2017) 35:4119-
4125.
27
8. Appendices
28
29
Figure 1. Key to Streptococcus pneumoniae Diagnostic Antisera Reactions. Excerpted from
pages 27-29 of The State Serum Institute Streptococcus pneumoniae Textbook in Diagnosis,
Serotyping, Virulence Factors and Enzyme-linked Immunosorbent Assay (ELISA) for Measuring
Pneumococcal Antibodies (2017) SSI Diagnostica, Oxford BioSystems.
30
Figure 2. Flowchart of IPD cases included in regression analysis. All cases of IPD in Germany
occurring in children between 90 days and 6 years of age born after the start of the study period,
1 July 1 2007, to 30 June 2015, which fit the study criteria. Cases were grouped for analysis into
a PCV7 group, a PCV10 group, and a PCV13 group, as shown above.
31
Figure 3. Population-based geographic analysis groups of children under 6 in Germany. The
population under 6 years of age based on data from the German Federal Statistical Office from
the 31 December 2015. Darker shades show districts which submitted at least one IPD isolate to
the GNRCS during the study period, 1 July 2007 to 30 June 2015.
32
Figure 4. IPD by regional district per 100,000 residents under the age of six. Cases of IPD reported to the GNRCS during the 2013-2014 season from children less
than six years of age are shown. At left, all cases per 100,000 residents under six years of age are depicted; at right, districts reporting only one case of IPD during the
2013-2014 season are highlighted.
33
Figure 5. IPD and VT IPD by federal state per 100,000 residents under the age of six. On the top left graphic, all reported IPD cases per 100,000 residents under
six years old are shown by federal state. In the top center and top right graphics, only PCV13-serotype IPD cases are shown. The bottom right shows PCV13-type case
reporting rates for the geographic analysis groups.
34
Figure 6. Distribution of demographic variables from the 2011 census. Variables are displayed by geographic analysis group (at left, in color) or by former political
group (at right, in grayscale).
35
Figure 7. IPD cases in children younger than 6 received by the GNRCS over the study period, 2007-2015. Cases of IPD per 100,000 residents under 6 years of
age are shown. Only cases with a known pneumococcal vaccination status are included.
36
Figure 8. Persistence of Vaccine Serotypes in IPD following implementation of PCV13. Vaccine-type cases of IPD in children under six with a known vaccination
status and who received either only PCV13 or no vaccine at all are shown. No IPD cases from serotypes 5, 23F, or 4 occurred during the PCV13 period (seasons 2009-
2010 through 2014-2015). One case of serotype 18C IPD occurred in the 2010-2011 season, and one case of serotype 9V IPD occurred in the 2011-2012 season, both
of which were in unvaccinated children.
37
Figure 9. Rising non-vaccine serotypes following PCV program implementation in
Germany. Percentage of selected non-vaccine serotypes causing IPD in children under six per
pneumococcal season, seen across the geographic analysis groups and Germany as a whole. While
the proportion of non-vaccine serotype IPD increased significantly in three of the four geographic
groups and across all of Germany (P=0.0388 in North Rhine-Westphalia, P=0.002 in the central
states, P=0.002 in the northeastern states, P=0.0006 in the southern states, P =1.30x10-9 in
Germany overall), no individual serotype reached significance when comparing the first and last
years of the study period.
38
Table 1. Vaccination Status of children with IPD in Germany, 2007-2015. Vaccination status and percentage for each PCV are shown for Germany as a whole, for
the geographic analysis groups, and for Former East and Former West Germany.
39
Table 2. Factors Influencing Pneumococcal Vaccination Status. Multivariate logistic
regression results describing associations of demographic and geographic variables with
pneumococcal vaccination status in children under 6 with IPD in Germany. No variables were
significantly associated in the PCV10 group.
40
Table 3. Multivariate logistic regression results for the PCV7 group. Results are divided into
three age cohorts: old enough to receive one dose of vaccine (≥90 days), old enough to receive
the full primary series (150-449 days old), and old enough to receive the booster dose (>449 days
old). Bold font indicates a statistically-significant correlation between the serotype and the listed
variable.
41
42
Table 4. Multivariate logistic regression results for the PCV10 group. Results are divided into
three age cohorts: old enough to receive one dose of vaccine (≥90 days), old enough to receive
the full primary series (150-449 days old), and old enough to receive the booster dose (>449 days
old). Bold font indicates a statistically-significant correlation between the serotype and the listed
variable.
43
44
Table 5. Multivariate logistic regression results for the PCV13 group. Results are divided into
three age cohorts: old enough to receive one dose of vaccine (≥90 days), old enough to receive
the full primary series (150-449 days old), and old enough to receive the booster dose (>449 days
old). Bold font indicates a statistically-significant correlation between the serotype and the listed
variable.
45
46
47
48
9. List of Publications
Perniciaro S., Imöhl M., van der Linden M. Invasive Pneumococcal Disease in Refugee
Children, Germany. Emerging Infectious Diseases (2018) 24:10 DOI:
10.3201/eid2410.180253
Perniciaro S., van der Linden M. Reassessing the 1 + 1 pneumococcal vaccine schedule.
The Lancet Infectious Diseases (2018) 18:4:381-382.
Imöhl M., Fitzner C., Perniciaro S., van der Linden M. Epidemiology and distribution of
10 superantigens among invasive Streptococcus pyogenes disease in Germany from
2009 to 2014. Public Library of Science ONE (2017) 12:7:e0180757.
van der Linden M., Falkenhorst G., Perniciaro S., Fitzner C., Imöhl M. Effectiveness of
Pneumococcal Conjugate Vaccines (PCV7 and PCV13) against Invasive
Pneumococcal Disease among Children under Two Years of Age in Germany. Public
Library of Science ONE (2016) 11:8:e0161257.
van der Linden M., Falkenhorst G., Perniciaro S., Imöhl M. Effects of Infant
Pneumococcal Conjugate Vaccination on Serotype Distribution in Invasive
Pneumococcal Disease among Children and Adults in Germany. Public Library of
Science ONE (2015) 10:7:e0131494.
van der Linden M., Perniciaro S., Imöhl M. Increase of serotypes 15A and 23B in IPD in
Germany in the PCV13 vaccination era. BioMedCentral Infectious Disease (2015)
15:207. DOI: 10.1186/s12879-015-0941-9.
Imöhl M., Möller J., Reinert R.R., Perniciaro S., van der Linden M., Aktas O.
Pneumococcal meningitis and vaccine effects in the era of conjugate vaccination:
results of 20 years of nationwide surveillance in Germany. BioMedCentral Infectious
Disease (2015) 15:61. DOI: 10.1186/s12879-015-0787-1.
van der Linden M., Winkel N., Küntzel S., Farkas A., Perniciaro S., Reinert R.R., Imöhl
M. Epidemiology of Streptococcus pneumoniae Serogroup 6 Isolates from IPD in
Children and Adults in Germany. Public Library of Science ONE (2013) 8:4:e60848.
49
10. Acknowledgements
My profound and sincere thanks are heartily given to the following individuals for their
expertise, support, patience, guidance, assistance, and inspiration:
Leon Perniciaro
Mark van der Linden
Matthias Imöhl
Mareike Jaschek, Julia Junker, and Natalja Levina
Christina Fitzner
Mathias Hornef
Maike Schleibach
The following organizations have also contributed substantially to the successful
completion of this dissertation:
Destatis, the German Federal Statistical Office
The R Foundation for Statistical Computing
The Open Source Geospatial Foundation
Robert Koch Institute
Pfizer Deutschland GmbH
Uni und Kind eV
50
Erklärung § 5 Abs. 1 zur Datenaufbewahrung Hiermit erkläre ich, dass die dieser Dissertation zu Grunde liegenden Originaldaten - in der Nationales Referenzzentrum für Streptokokken, Institut für
Medizinische Mikrobiologie des Universitätsklinikums Aachen
hinterlegt sind.
51
Eidesstattliche Erklärung gemäß § 5 Abs. (1) und § 11 Abs. (3) 12. der Promotionsordnung Hiermit erkläre ich, Stephanie Russo Perniciaro an Eides statt, dass ich folgende in der von mir selbstständig erstellten Dissertation „Spatial Epidemiology of Invasive Pneumococcal Disease Isolates from Children under Six in Germany“ dargestellten Ergebnisse erhoben habe1: Bei der Durchführung der Arbeit hatte ich folgende Hilfestellungen, die in der Danksagung angegeben sind
Doktorandin Stephanie Russo Perniciaro
Kooperationspartner Matthias Imöhl
Statistikerin Christina Fitzner
Assistentinnen Mareike Jascheck Julia Junker Natalja Levina
Doktorvater Mark van der Linden
Summe (%)
Studienüberwachung 100 100
Studiendesign/Konzeption 60 10 10 20 100
Untersuchung der IPD Proben
100 100
Datenauswertung 100 100
Statistische Auswertung 100 100
Bereitstellung von Materialien
100 100
Interpretation der Datenauswertung
60 10 20 10 100
______________________________________ Unterschrift der Doktorandin/des Doktoranden Als Betreuer der obigen Dissertation bestätige ich die Angaben von Stephanie Russo Perniciaro ______________________________________ Unterschrift des Doktorvaters
52
Stephanie Russo Perniciaro, geb. Russo
Geburtsdatum: 20 Oktober 1984 Geburtsort: Champaign, Illinois Geburtsland: Vereinigte Staaten von Amerika
Krefelderstraße 4, 52062 Aachen 017676684453 [email protected]
Beruflicher
Werdegang
Promotionsstudentin Universitätsklinikum Aachen
52074 Aachen
Doktorarbeit: Spatial epidemiology of invasive pneumococcal disease isolates from children under six in Germany
Wissenschaftliche Hilfskraft
Universitätsklinikum Aachen
52074 Aachen
mikrobiologische Laborarbeit, Nationales Referenzzentrum für Streptokokken, Institut für Medizinische Mikrobiologie
Clinical Research Coordinator
Tulane University School of Medicine New Orleans, LA
USA
klinische Forschung, Anästhesiologie Abteilung
Research Technician
University of Pittsburgh
Pittsburgh, PA USA
biologische Laborarbeit, Epidemiologie Abteilung
Public Health Intern Consumer Health Coalition
Pittsburgh, PA USA
klinische Übersetzungen und Veranstaltungsplanen
Teaching Assistant Creighton University
Omaha, NE USA
Aufsicht auf Laborkurse, Biologie Abteilung
Research Technician
Creighton University
Omaha, NE USA
Versuchstierhandlung, Biomedizin Abteilung
Academic Learning Advisor
Creighton University
Omaha, NE USA
Unterrichten und Nachhilfe: Spanisch, Englisch, sonstiges
Laboratory Technician
North Shore Sanitary District Gurnee, IL USA
Abwasserbehandlungslaborarbeit
Ausbildung
Bachelor of Science, Biology
Creighton University Omaha, NE USA
2007
Master of Public Health, Epidemiology
University of Pittsburgh Pittsburgh, PA USA
2010
Fähigkeiten Muttersprache: Englisch; weitere Sprachen: Deutsch (C1.2), Spanisch (C2)
Statistische Software: R, SAS, SPSS sowie Standard Software (Microsoft Office); grundniveau HTML, Python, QGIS
Standard mikrobiologische/wissenschaftliche Techniken
Referenzen Referenzen werden auf Wunsch zur Verfügung gestellt