Article Text
PDF
Abstract
Objectives To report sex and age-specific Chlamydia trachomatis (Ct) seroprevalence estimates in the general population of the Netherlands between 1996 and 2017 and identify risk factors associated with Ct seropositivity.
Methods Participants (n=5158, aged 15–59 years) were included from three independent nationwide population-based serosurveillance studies in 1996, 2007 and 2017. Participants completed a questionnaire on demographics and sexual behaviour. Serum antibodies were analysed using Medac Ct IgG ELISA test. Census weights were assigned to achieve seroprevalence estimates representative of the general Dutch population. Weighted seroprevalence estimates were stratified by gender, age and birth cohort. Trends and risk factors in men and women were identified using multivariable logistic regression.
Results Weighted overall Ct seroprevalence was 10.5% (95% CI: 9.2% to 12.0%) in women and 5.8% (95% CI: 4.7% to 7.0%) in men. Among women <25 years, there was a non-significant increase in seroprevalence from 5.9% (95% CI 3.7% to 9.2%) in 1996, to 7.6% (95% CI 5.1% to 11.1%) in 2007 and 8.8% (95% CI 5.5% to 13.9%) in 2017. Among women ≥25 years, the seroprevalence significantly decreased from 15.6% (95% CI: 12.2% to 19.7%) in 1996 to 9.5% (95% CI: 7.2% to 12.4%) in 2007 but did not further drop (11.2% (95% CI 8.1% to 15.3%) in 2017). In men, we did not observe trends between study rounds. In both men and women, having a non-Western migration background was a risk factor for seropositivity. In women, having had a prior sexually transmitted infection and ≥2 recent sex partners were risk factors for seropositivity as well.
Conclusions We have not found evidence for a decrease in population seroprevalence in those under 25 years old despite decades of intensified testing-and-treatment efforts in the Netherlands. This suggests further monitoring of Ct burden in the general population is needed. If serum banks are used for this, specifically individuals <25 years old and with diverse migration backgrounds should be included.
- population surveillance
- seroepidemiologic studies
- chlamydia infections
Data availability statement
Data are available upon reasonable request.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
Nationwide chlamydia seroprevalence surveys are a tool for understanding the impact of control activities in countries.
In the Netherlands, chlamydia reporting rates are rising, but it is unclear if this is due to increased transmission or better case finding.
WHAT THIS STUDY ADDS
There was no evidence for a decrease in population chlamydia seroprevalence in people under 25 years old, despite decades of targeted chlamydia testing-and-treatment efforts.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
In settings with targeted chlamydia testing, repeated population-based serology studies provide essential insight in chlamydia exposure in the general population.
Future seroprevalence studies for monitoring of chlamydia should specifically include individuals <25 years old and with diverse migration backgrounds.
Introduction
Widespread testing and treatment of Chlamydia trachomatis (Ct) in asymptomatic women and heterosexual men are under debate in various Western countries.1 2 The goals of Ct control can be twofold: first, to curb transmission of Ct infection in the population and second, to reduce reproductive tract sequelae. While there is evidence from trials that Ct control programmes can reduce the risk of pelvic inflammatory disease (PID) in women, there is a lack of evidence supporting a reduction in population prevalence.1–3
To estimate population prevalence, Ct serology can be used in addition to nucleic acid amplification tests (NAATs), as it estimates cumulative exposure over time. Recent studies of Ct serology using existing serum banks have provided valuable insights into the impact of Ct control activities.4–6
In the Netherlands, Ct is monitored by surveillance in all Sexual Health Centre (SHC) clinics, where free-of-charge sexually transmitted infection (STI) testing and treatment are provided for high-risk groups (eg, heterosexuals aged <25 years, those who report STI symptoms or are notified of STI exposure), and in a representative sample of general practitioners (GPs). Reported Ct infections have increased steeply between 2010 and 2021 by 51% in GPs and 77% in STI clinics, adding up to around 63.000 reported Ct infections in 2021.7 8 However, surveillance estimates cannot be used as indicators of population incidence and prevalence as Ct infections are often asymptomatic and diagnosis depends on client-initiated testing behaviour and client characteristics, which changed over time. Repeated nationwide surveys are needed to put these surveillance data into perspective.
Ct seroprevalence in the Netherlands was previously assessed using a national biobank for population-based seroprevalence studies from 1995/1996 and 2006/2007.9 We now have seroprevalence data from 2017 available. This allows us to study trends and gain more insight in Ct prevalence following the implementation of a uniform national testing policy in 2006.8 The aim of this study was to report sex and age-specific Ct seroprevalence, evaluate trends between 1996–2017 and risk factors for seropositivity.
Methods
Study design
Samples were used from large serum banks in the Netherlands, collected from three consecutive cross-sectional serological surveys conducted in 1995–1996 (Pienter 1, referred to as 1996), 2006–2007 (Pienter 2, ‘2007’) and 2016–2017 (Pienter 3, ‘2017’). The primary objective of the Pienter studies was to periodically monitor seroprevalence of national immunisation programme-targeted diseases. The designs of the Pienter studies were described in detail elsewhere.10–12 In summary, a representative national sample was drawn from the Dutch population. In randomly sampled municipalities (selected from geographical regions proportional to size), an age-stratified sample was randomly drawn from the register (age 0–79 years). Participants were invited to fill in a questionnaire and provide a blood sample.
We included participants aged 15–39 years as this group is most relevant for Ct transmission and bears the largest burden of sexual and reproductive ill-health due to Ct. A sufficient number of samples were included to detect a 3% change in overall Ct prevalence across concurrent Pienter rounds.9 This resulted in including all Pienter samples from 2007 with sufficient serum and from the samples collected in 1996, a random selection was made matched for age and sex to 2007.9 In 2017, the age group was expanded to 15–59 years to describe birth cohort effects. Samples from 2017 were randomly selected to achieve a similar age distribution as in 1996 and 2007.
Laboratory methods
The Medac Ct IgG ELISA (Wedel, Germany) test was used to test for Ct IgG antibodies. This test was the most used serology test in clinical practice in the Netherlands.13 14 The ELISA uses a synthetic peptide from the immuno-dominant region of the major outer membrane protein (MOMP) as antigen, providing minimal cross-reactivity with other Ct species and therefore high specificity (97%).14 15 Sensitivity of the Medac was found to be 71% as compared with the Microimmunofluorescence assay, and varied between 38% and 66% in comparison with NAAT, depending on time since infection.14 16 Levels of Ct IgG antibodies (arbitrary units (AU)/mL) were calculated per manufacturer’s instructions, and outcomes were classified as negative (IgG concentration <22 AU/mL), grey zone (IgG concentration 22–28 AU/mL) or positive (IgG concentration ≥28 AU/mL).
Questionnaire data
From all participants, information on age, sex, urbanisation, country of birth and country of birth of both parents was available from the population register. Migration background was based on country of birth of the participant and both parents and was categorised as Dutch if both parents and participant were born in the Netherlands. First and second-generation migrants were grouped together and further divided into Western non-Dutch (West European countries, Australia, North America, Indonesia and Japan); Morocco and Turkey; Suriname, Aruba and the former Dutch Antilles; and non-Western (other countries). Other demographic and sexual risk behaviour variables were self-reported and definitions can be found in online supplemental table 1. Education was divided into theoretical (higher professional level, university level) or practical (all other levels). No data on previous STI testing were available, only on diagnosis which was included in the analyses as ‘STI diagnosed ever’ and ‘Ct diagnosed ever’.
Supplemental material
Weighting of the data
We incorporated census weights to achieve estimates representative for the general Dutch population. The weights were obtained using census data for the Netherlands by sex, age and migration background, for each survey period. Weights were obtained by dividing the proportion of the population in the Netherlands in a specific group (eg, women aged 25–29 years with a Turkish or Moroccan migration background) by the proportion of survey participants in the same group.
Statistical analyses
We first determined differences between participants in study rounds by comparing demographic and behavioural variables using the χ² test. Prevalence of IgG antibodies against the Ct MOMP (hereafter ‘seroprevalence’) was calculated with corresponding 95% CIs. Antibody levels in Ct seropositives stratified by age group and sex were compared between study rounds using the Mann-Whitney U test. P values below 0.05 were considered significant.
Ct IgG seroprevalence
Numbers of Ct seropositives and seroprevalence were all estimated taking into account weights. Analyses were stratified by sex, age group, study round and birth cohort. Strata were compared using the χ² test. Trends between three or more groups were assessed using the χ² test for trends. P values below 0.05 were considered significant. Several comparisons were made among participants 15–39 years old, all stratified by sex and age group. Age was divided into: (1) binary variable corresponding to high-risk triage criterion in Dutch SHC clinics7 (<25 years and ≥25 years), and (2) in 5-year age bands (15–19, …, 35–39 years) to assess trends in more detail.
First, we compared overall seroprevalence between women and men by age group. Second, we compared seroprevalence between age groups by study round. Last, we compared seroprevalence between study rounds by age group, which we plotted against the yearly number of tests conducted in SHC based on national surveillance data.7 8
In addition, we compared seroprevalence between study rounds by cross-sectional birth cohorts for which participants 15–59 years were included. For example, for women born in 1977–1981, we compared seroprevalence between 1996, 2007 and 2017, when they were, respectively, 15–19, 25–29 and 35–39 years.
Determinants of Ct seropositivity
To identify determinants of Ct seropositivity, logistic regression analyses were performed stratified by sex. Due to low number of seropositives per study round, the data from all rounds were combined. We excluded variables which were highly correlated (condom use: regular partner vs casual partner and last sexual contact) or which had an inconsistent definition across study rounds (sexual practice). We included the following categorical sexual risk behaviour variables: STI diagnosed ever, age at first time sex, number of recent sexual partners, having a regular partner and condom use with regular partner (online supplemental table 1). If a variable had more than 5% missingness, a separate missing category was included in the analyses to avoid loss of information.
We first performed univariable logistic regression. We then performed multivariable logistic regression. The multivariable model was built via backward stepwise selection based on Akaike Information Criterion (AIC). Age and migration background could affect seropositivity differently across study periods, for example, because access to sexual healthcare has become more available to certain groups over time. To account for potential bias introduced by combining study rounds, interaction terms for age and migration background with study round were included. Interaction terms were checked one by one by comparing the AIC with and without. Variables with subgroups with less than 15 cases were excluded from the multivariable analyses. Weighted ORs and weighted adjusted ORs with 95% CIs were calculated.
Statistical analyses were performed using R V.4.2.0, and weights were incorporated in Ct seroprevalence estimates and logistic regression analyses using the R survey package.17
Results
Characteristics of the study population
We included a total of 5158 participants. For ages 15–39 years, we included 958 women and 620 men from 1996, 972 women and 625 men from 2007, and 569 women and 587 men from 2017. For ages 40–59 years, we included additional participants from 2017 (431 women and 413 men). Participants with Ct grey zone results were excluded from all analyses (n=32, 0.6%).
Characteristics are described for participants aged 15–39 years (table 1). The majority of participants had a Dutch migration background, but this decreased from 88% in 1996 to 83% and 82% in 2007 and 2017, respectively. Participants in 1996 were more often from low urbanisation municipalities and practically educated as compared with 2007 and 2017. The proportion of participants with a known history of STI increased over time from 2.7% in 1996 to 4.0% in 2007 and 7.3% in 2017. More participants reported first sexual contact <18 years over time from 31% in 1996 to 42% in 2017. Most participants (70–80%) had <2 recent sexual partners. Condom use increased between 2007 and 2017 with last sexual contact, with regular partner and casual partner (p<0.05).
Sociodemographic characteristics and sexual risk factors of the study population (15–39 years old) in 1996, 2007 and 2017
Ct IgG seroprevalence
In total, 347 chlamydia seropositives were identified. Among both women and men, median Ct antibody levels by age group did not differ significantly between study rounds (online supplemental figure 1).
Supplemental material
Overall, the weighted Ct seroprevalence was 10.5% (95% CI: 9.2% to 12.0%) in women and 5.8% (95% CI: 4.7% to 7.0%) in men (table 2).
Weighted numbers of Ct IgG positives and weighted Ct IgG seroprevalence in the general population in the Netherlands in 1996, 2007 and 2017
Between women and men
Overall seroprevalence was significantly lower in men compared with women ≥25 years (6.0%; 95% CI: 3.9% to 7.7% men vs 12.2%; 95% CI: 10.4% to 14.3% women). Among those <25 years, seroprevalence was not significantly different between men and women (5.5% men vs 7.5% women).
Between age groups by study round
In 1996, seroprevalence was significantly lower in women <25 years than in women ≥25 years (5.9%, 95% CI: 3.7 to 9.2 vs 15.6%, 95% CI: 12.2 to 19.7, p<0.001), but this difference diminished in 2007 and 2017 (table 2). In more detail, figure 1A showed seroprevalence increases with each 5-year age band in 1996. On the contrary, in 2017, this increase with age is only visible in women <30 years. Among men, within each study round, seroprevalence was not significantly different between those below and above 25 years (table 2 and figure 1C).
Chlamydia trachomatis (Ct) IgG seroprevalence by study round and by age strata for women (A) and men (C). Ct IgG seroprevalence by birth cohort for women (B) and men (D) for 1967–1971, 1972–1976 and 1977–1981, with corresponding 95% CI for seroprevalence. Flat lines in B and D indicate no data were collected for this birth cohort and age group. P value is given for χ² test between two strata or linear trend in proportions between three strata. P values are indicated as: *p<0.05, **p<0.01, ***p<0.001.
Between study rounds by age group
Among women ≥25 years, the seroprevalence significantly decreased from 15.6% (95% CI: 12.2% to 19.7%) in 1996 to 9.5% (95% CI: 7.2% to 12.4%) in 2007 (p<0.01), but did not further drop with 11.2% (95% CI 8.1% to 15.3%) in 2017 (p=0.42 for 2007 vs 2017). The decrease is especially visible in women 30–34 and 35–39 years (figure 1A). Among women <25 years, there was a non-significant increase in seroprevalence from 5.9% in 1996, to 7.6% in 2007 and 8.8% in 2017, during a period of intensified Ct testing and treatment in this group (online supplemental figure 2). This increase is especially visible in women 20–24 years (figure 1A). Among men ≥25 years, there is a non-significant increase in seroprevalence from 4.5% in 1996 to 6.9% in 2007 and 6.7% in 2017 (table 2), which is not observed in men <25 years.
Supplemental material
Between study rounds by birth cohort
A significant increase in seroprevalence by age was found within the female birth cohorts of 1977–1981 and 1988–1992 (p=0.01 and p=0.007, figure 1B) and within the male birth cohort of 1967–1971 (p=0.001, figure 1D). The largest significant increase within a birth cohort is found for women born in 1988–1992 from 4.7% (95% CI: 2.2% to 9.8%) at 15–19 years to 15.0% (95% CI: 8.9% to 24.2%) at 25–29 years (online supplemental table 3).
Determinants of Ct seropositivity
Having a non-Western migration background, ever prior STI and ≥2 recent sex partners were associated with Ct seropositivity in women (table 3). Furthermore, consistent with previous analyses, we found that the odds of Ct seropositivity among women aged ≥25 years differed between 1996 and 2007/2017. Among men having a non-Western migration background was associated with Ct seropositivity.
Determinants for Ct IgG seropositivity among women and men aged 15–39 years old in the Netherlands in 1996, 2007 and 2017
Discussion
Based on three nationwide population-based probability surveys in the Netherlands, we found an overall chlamydia seroprevalence of 10.5% (95% CI: 9.2% to 12.0%) among women 15–39 years over the time period from 1996 to 2017. Seroprevalence in women ≥25 years was highest in 1996, dropped in 2007, but plateaued by 2017. In women <25 years, seroprevalence has increased non-significantly but consistently between 1996 and 2017. Within birth cohorts, the strongest effect was observed in women born 1988–1992 where seropositivity tripled between those 15–19 years and 25–29 years old. For men, seroprevalence was half of that of women. We did not observe trends within and between study rounds.
This is the first Ct seroprevalence study in the Netherlands to assess trends during decades of increased Ct control efforts. There are several limitations to this study. First, response rates of Pienter decreased from 55% to 16% between 1996 and 2017,10–12 similar to falling rates in European health surveys.18 Although participation determinants were consistent across Pienter surveys, healthy responder bias could be present which may result in underestimating Ct seroprevalence.19 20 We (partly) corrected for this by weighting for demographic characteristics, but generalisability might be decreased in subgroups at a population level. Second, no gold standard exists for Ct serology, making any serology test suboptimal, particularly in men who often do not develop a detectable antibody response after infection.4 21 Due to sensitivity of the used serology test (38–66%), our estimates are likely an underestimation of prior Ct infections.14 15 Furthermore, it is known that individuals with repeated infections and recent infections are more likely to develop and maintain detectable antibodies.16 22 As a consequence, people with first or non-recent infections might be missed in our seropositives. Third, because of the relative low number of identified seropositives, few stratified analyses could be conducted and a certain degree of uncertainty was observed around our estimates. For women, we did observe consistent trends, but these were often not significant, likely due to low numbers. For men, the uncertainty is largest due to relatively lower sensitivity of serology to detect past infection in males.4 21
Our results on prevalence are as expected when compared with a Dutch study among participants of a prior Ct screening implementation trial, in which the same antibody test was used as in our study.22 Among women 25–39 years, a seropositivity of 15% (95% CI 14% to 16%) was found in the cohort study in 2016,22 whereas 11% (95% CI 8% to 15%) was found in our study. The difference is likely explained by the self-selection of high-risk participants into the Ct screening implementation trial. In an English study in 2010/2012, a seroprevalence of 25% was found in women using an in-house Pgp3 antibody test, while our study found 9–10% in 2007/2017.4 A comparison study among women with a prior Ct infection, in which the Pgp3 assay found 74% seropositives and the Medac assay found 45%, suggests the difference likely reflects the higher sensitivity of the Pgp3 test and not necessarily a higher prevalence of Ct infection in England compared with the Netherlands.4 16 Both the English study and our study observed an increase in seroprevalence with age in women, which levelled out around age 30 years.4
Identified risk factors for Ct seropositivity found in our study are largely consistent with those found in other population-based studies.4 22 23 We found an association between migration background and Ct seropositivity, with similar effect size as found for race/ethnicity in a study in the USA.23 Reasons for this association might be that immune response to Ct differs across ethnic groups.24 25 It could also be that migration background is a proxy of lower socioeconomic status. However, we did not find an association with level of education or urbanisation and Ct seropositivity.
For women, we identified trends across study rounds which could be explained by differences in Ct exposure.
Among women ≥25 years, seropositivity was highest in 1996 and significantly decreased in in 2007. This suggests that exposure to Ct was historically high before 1990 and subsequently decreased, possibly as a consequence of changes in sexual behaviour following extensive AIDS/HIV media coverage.26 Other explanations might include Ct control efforts, though the absence of a further decline from 2007 to 2017 contrasts with increased control efforts in the Netherlands in this period. In 2017, seroprevalence increased with age, as expected for cumulative exposure, but levelled out around age 30 years. This may suggest reduction of exposure or waning of Ct antibodies in women >30 years, relative to those <30 years.
Among women aged <25 years, we observed no significant changes in age-specific seroprevalence between 1996 and 2017, which is surprising given increased control efforts during that time. Multiple explanations could explain this. First, it could be that due to limited statistical power, a change could not be detected in this age group. However, if anything, we observe an increase in seroprevalence over time rather than a decrease in women <25 years. This is supported by the birth cohort results which suggest the strongest increase by age is in the most recent cohort of young women (1988–1992). Second, targeted testing and treatment of high-risk populations could have an effect in that group, but that effect cannot be observed in a sample of low-risk individuals. Finally, seroprevalence trends may not directly indicate changes in exposure, but rather relate to factors such as treatment, frequency or duration of infections influencing seroconversion or reversion,22 27 28 for which no data were collected in this study.
If test and treat of current Ct infection is scaled down in the foreseeable future, serology could be used to obtain reliable estimates of general population exposure.1 Based on this study, we would recommend samples to be collected more frequently to be able to assess trends (eg, every 5 years), focus on those under 25 years old (because interpretation is closer to incidence) and include sufficient numbers from different migration backgrounds. Furthermore, it should be considered that scaling back Ct control programmes may not necessarily impact transmission but could lead to an increase in sequelae. Therefore, monitoring of sequelae is crucial, and biomarkers to predict the progression to scarring sequelae, such as PID and infertility, are needed. Assays that use Ct antibody profiles as predictors for disease progression are currently under development and could be incorporated into Ct serosurveillance.21 29 30
We have not found evidence for a decrease in population seroprevalence in those under 25 years old despite decades of intensified testing-and-treatment efforts in the Netherlands. Repeated seroprevalence studies in the future would be useful to monitor Ct burden in the general population.
Data availability statement
Data are available upon reasonable request.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and ethical approval for the Pienter seroprevalence surveys was obtained by an Ethical Committee in the Netherlands: Pienter 1 (1996) by the Medical Ethical Committee of Netherlands Organisation for Applied Scientific Research; Pienter 2 (2007) by the Medical Ethics Testing Committee of the foundation of therapeutic evaluation of medicines in Almere (number: ISRCTN 20164309); Pienter 3 (2017) by the Medical Ethics Committee Noord-Holland (number: M015-022). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
First of all, we gratefully acknowledge the participants of the Pienter studies. Second, we thank the Pienter team from the National Institute of Public Health and Environment (RIVM), Bilthoven, the Netherlands, more specifically Fiona van der Klis, for making the Pienter data and samples available. Finally, we thank Eleanne van Ess and Jolande Land for their contribution to the research proposal and laboratory analyses.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
Handling editor Laith J Abu-Raddad
Contributors SAM, BMH and FvA selected the serum samples. SAM conducted the laboratory analyses. ZWA analysed and interpreted the data. JCMH designed the study and supervised the data analyses, interpretation and drafting of the manuscript. ZWA drafted the manuscript. All authors contributed comments on the manuscript and approved the final version. ZWA and JCMH are the guarantors of the study.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.