Estimated effect of reduced asymptomatic testing on missed Chlamydia trachomatis detections among women and heterosexual men in the Netherlands: a retrospective observational study

Summary

Background

Widespread testing for asymptomatic chlamydia is under debate. In the Netherlands, from 2025, Chlamydia trachomatis (CT) testing at Centers for Sexual Health (CSH) is indicated only for individuals with symptoms or partner notification. We aimed to quantify the potential impact of the new guideline on CT tests and detections for women and heterosexual men.

Methods

We included consultations of women and heterosexual men (≥16 years) tested for CT at the CSH Amsterdam, the Netherlands, 2015–2023. We assessed consultations where no test would have been performed under the new guideline (neither symptoms nor notified), including CT diagnosed during those consultations (=undetected CT). Determinants of undetected CT were assessed using multivariable logistic regression with cluster-robust standard errors. Analyses were stratified by age (<25/≥25 years) and gender (women/men).

Findings

In total 241,005 consultations among 124,527 individuals were analysed. Under the new guideline, CT testing would not have been performed in 167,476 (69.5%, 95% confidence interval [CI]; 69.3–69.7%) consultations; mostly among women <25 years (n = 100,756/167,476, 60.2%). In total 17,771/32,360 (54.9%, 95% CI; 54.4–55.5%) CT cases would have remained undetected and untreated: 12,092 among women <25 years, 1498 among women ≥25 years, 3428 among men <25 years, and 753 among men ≥25 years. Determinants of undetected CT were younger age, no/primary/secondary education, and recent bacterial STI.

Interpretation

The new guideline limiting CT testing might lead to a more than twofold reduction in diagnoses and antibiotic prescriptions in women and heterosexual men. Surveillance of STIs and associated complications is essential to monitor unintended consequences of the guideline change.

Funding

None.

Research in context.

Evidence before this study

We searched PubMed for studies on the impact of reduced testing for asymptomatic Chlamydia trachomatis (CT) on the number of tests and diagnoses up to 29 April 2025. We used the following search terms: “Chlamydia trachomatis” AND (“reduced testing” OR “discontinued testing” OR “non-screening”). No studies were found that reported any results regarding the number of CT tests and diagnoses. However, one randomized controlled trial (RCT) in Belgium reported on the incidence of CT and Neisseria gonorrhoeae in a sample of 1014 men who have sex with men and transgender women. In this study, all participants were screened at three anatomical sites (urethra, anus, and pharynx) every 3 months, but test results were not provided to the non-screening arm, if asymptomatic. The incidence of CT detection after 12 months was higher in the non-screening arm (incidence rate ratio 1.435 (95% confidence interval 1.098–1.875)), but this difference disappeared when two consecutive asymptomatic detections were counted as one infection.

Added value of this study

This study provides insight into the potential impact of a recently implemented guideline change in Centers for Sexual Health in the Netherlands, which restricts CT testing to women and heterosexual men with symptoms or partner notification, on the number of CT tests and diagnoses. We analysed routinely collected data covering approximately 241,000 consultations over a period of nine years. By calculating how many CT tests were performed and diagnoses were made among individuals who had neither symptoms nor a partner notification, we could estimate how many CT tests and diagnoses would not have been made if the revised guideline had been in place.

Implications of all the available evidence

We estimated that the new guideline limiting asymptomatic CT testing among women and heterosexual men might result in approximately 70% fewer CT tests and 55% fewer CT diagnoses. The majority (68%) of CT detections in those who would not have been eligible for testing under the new guideline were among women <25 years. These findings imply that antibiotic prescriptions for CT are likely to be reduced by more than half. Although existing evidence suggests that widespread asymptomatic CT testing does not lower chlamydia prevalence, it remains crucial to monitor symptomatic CT, related complications, and NG detections to evaluate any potential unintended outcomes following the guideline change.

Introduction

Chlamydia is the most commonly reported bacterial sexually transmitted infection (STI) in the world.¹

Widespread asymptomatic Chlamydia trachomatis (CT) testing strategies targeting persons at risk for chlamydia aim to reduce transmission, shorten the duration of infectiousness, and minimize long-term complications through early diagnosis and treatment. However, there is an ongoing debate on the harms and benefits of widespread asymptomatic testing.2, 3, 4 Reasons to reconsider asymptomatic CT testing include: (1) increasing evidence from epidemiological studies suggests that widespread asymptomatic testing does not reduce the population prevalence of chlamydia5, 6, 7; (2) the added value of widespread asymptomatic testing for the prevention of long-term complications attributable to CT infection (e.g., pelvic inflammatory disease [PID] and ectopic pregnancy) remains uncertain^8,9; (3) insights into the impact of high rates of antibiotic consumption due to widespread CT testing and detection on the microbiome, which may contribute to antimicrobial resistance.^3,10

The shift in the balance between harms and benefits has prompted adjustments to CT testing guidelines in several countries. For example, in the United Kingdom, the National Chlamydia Screening Programme (NCSP) initially recommended screening for all sexually active people below the age of 25 years. However, by 2021, screening in settings such as general practitioner (GP) practices was limited to women aged 15–24 years only.¹¹ In Belgium, the recommendation to test asymptomatic men who have sex with men (MSM) using HIV pre-exposure prophylaxis for CT and Neisseria gonorrhoeae (NG) has been discontinued since 2024.¹² This change was based on results from a randomized controlled trial, which showed that screening did not reduce NG incidence and only slightly reduced CT incidence among MSM.¹³ In the Netherlands, as of 2025, the CT testing guideline at Centers for Sexual Health (CSH) of Public Health Services shifted from testing all individuals regardless of symptoms to only offering a CT test to individuals with symptoms and/or those who have received a partner notification for chlamydia.¹⁴ CT testing conducted in GP practices in the Netherlands is guided by separate clinical guidelines and therefore not subject to this change. The 2025 guideline change did not affect testing practices for NG, for which asymptomatic testing and treatment remains part of routine care.

The implications of the Dutch guideline change on the number of tests no longer performed and CT cases no longer detected remain unclear. Understanding these effects is crucial to assess potential public health consequences, such as reduced antibiotic use due to less treatment of asymptomatic CT and potential unintended effects on other sexually transmitted infections (STIs). Therefore, this study aimed to quantify the potential impact of the new guideline on tests and diagnoses for women and heterosexual men attending the CSH Amsterdam, the Netherlands. More specifically, we assessed the number and proportion of CT tests that would no longer have been performed under the revised guideline, and the number and proportion of CT detections in those who would not have been eligible for testing (=undetected CT). Additionally, we examined determinants of a positive CT test among those who would no longer have been eligible for testing, to identify groups for whom intended and unintended consequences of the guideline change may be most relevant.

Methods

Study setting

We retrospectively analysed routinely collected data from the CSH of the Public Health Service Amsterdam, the Netherlands, which performs about 35% of the total number of CSH consultations in the Netherlands. The CSH provides free of charge, anonymous, and low-threshold STI testing, treatment and counselling to individuals seeking sexual health care. To get access to testing at the CSH, women and heterosexual men should meet at least one of the triage criteria: (1) being under the age of 25; (2) reporting STI-related symptoms; (3) being notified of STI exposure by a partner; (4) born in a country defined as an STI/HIV-endemic area, namely countries in Asia, Africa, the Dutch Caribbean islands, middle and South America; (5) reporting a partner originating from an STI/HIV-endemic area or, for women only, an MSM partner; (6) reporting sex work; (7) or being a victim of sexual violence.¹⁵ Routine practice entailed testing each woman for CT and NG at the urogenital site. Guidelines recommended additional anorectal testing when anal sex or any symptoms were reported. For women without symptoms, anorectal testing was not specifically recommended and was performed less frequently. Additional oral testing was performed when oral sex or oral symptoms were reported. For heterosexual men, routine practice involved testing for CT and NG exclusively at the urogenital site. Combined CT and NG testing was done using a highly sensitive nucleic acid amplification test (NAAT) which is the Aptima Combo2 test from Hologic.¹⁶

Data collection

At each consultation, clients self-reported sociodemographic (e.g., age, educational level, and region of origin) and sexual behaviour data (e.g., number of sexual partners, condom use with a casual partner, and sex work) including partner notifications and previous bacterial STI diagnosis (chlamydia/gonorrhoea/syphilis), which were collected through standardized questions. For men, ‘heterosexual’ was defined as self-reported sexual contact only with partners with a vagina in the preceding six months. Educational level was categorized as either up to secondary (general/vocational) education (including no education; elementary school; pre-vocational secondary education; secondary vocational education; senior general secondary education; and pre-university education), or university or university of applied sciences education. Region of origin was dichotomized into originating from an STI/HIV-endemic area (yes/no), based on whether the individual or at least one parent was born in a country classified as an STI/HIV-endemic area. All sexual behaviour variables referred to the preceding six months, except for condom use with a casual partner (prior to 2018 referring to last sexual encounter; from 2018 onwards referring to preceding six months) and previous bacterial STI diagnosis (referring to preceding 12 months). Both women and men were asked whether they had partners originating from an STI/HIV-endemic area. Women were additionally asked whether they had male partners who had sex with men. These responses were combined into a single variable: having a partner originating from an STI/HIV-endemic area or an MSM partner. Symptoms were self-reported by the client. Each individual was assigned a unique identifier, making it possible to account for multiple CSH visits throughout the study period.

In- and exclusion criteria for current analysis

All consultations between 1 January 2015 and 31 December 2023 of cisgender women and heterosexual cisgender men aged ≥16 years during which a CT test was performed were included. The start date of the study period was selected because 2015 marked a shift in sexual health care funding at CSHs, prioritizing individuals from higher-risk groups and subsequently altering client characteristics.¹⁷ Exclusion criteria for the current analysis were: (1) deviation from routine testing practice; (2) missing data on symptoms or partner STI notification; (3) consultations where only conjunctival or ulcer samples were tested; and (4) men tested anorectally.

Statistical analysis

All analyses were stratified by gender (i.e. cisgender women and cisgender men), given the gender differences in chlamydia prevalence.¹⁸ Study population characteristics were described using data at the time of each individual's first consultation; clinical data were presented per consultation. All analyses described hereafter were performed at consultation level and further stratified by age (<25 and ≥ 25 years), because of the age-related differences in triage criteria.¹⁵ We defined eligibility for CT testing under the new guideline as (1) reporting of STI-related symptoms during the consultation and/or (2) having received a partner notification for chlamydia. Conversely, non-eligibility was defined as (1) no symptoms reported and (2) no partner notification received. A CT detection was defined as a positive CT NAAT in at least one anatomical location (urogenital/anorectal/oral). Because we analysed the data at consultation level, individuals could contribute consultations to multiple analytic groups (e.g., being <25 years old in one consultation, and ≥25 in another).

We assessed the proportion of consultations where a CT test would not have been performed, as the number of consultations of those who would not have been eligible for testing, divided by the total number of consultations. We additionally calculated the number and proportion of CT detections among consultations from individuals who, under the revised guideline, would not have been eligible for CT testing, and compared these to consultations from individuals who would have remained eligible. 95% confidence intervals (95%CIs) for percentages were estimated using the binomial distribution. The average annual number of undetected CT cases (i.e. CT detections among individuals who would not have been eligible for CT testing under the new guideline) was calculated by dividing the total number of undetected CT cases by the number of study years. Additionally, given that NG testing guidelines remain unchanged, but reduced CT testing eligibility may lead to fewer individuals attending the CSH for STI testing, the guideline change may indirectly affect asymptomatic NG detection. To explore this possible unintended consequence, we assessed the number of NG diagnoses among individuals who would no longer have been eligible for CT testing.

To identify individuals at higher risk of undetected CT, we assessed determinants of a CT detection among those who would no longer have been eligible for testing across the four strata (gender and age). We used logistic regression with robust standard errors clustered at the individual level to account for within-person correlation. We addressed missing values in two ways: 1) if a variable had more than 5% missing values, a separate ‘missing’ category was created to preserve these records in the analysis (main analysis), and 2) we used multiple imputation of missing values by chained equations (MICE) creating 20 imputed datasets, including relevant covariates and the outcome variable (sensitivity analysis 1, see Supplementary Methods 1 for an in-depth explanation). Variables from univariable analysis were included in the multivariable model if the Akaike Information Criterion (AIC) of their univariable model was lower than that of the null model (i.e. a model including only the outcome variable). We assessed whether there was multicollinearity between included variables using the variance inflation factor (VIF), and dropped variables with a value of >10. Backward stepwise selection using the AIC was performed to determine which variables improved the model significantly. As we aimed to compare associations among people <25 and ≥ 25 years of the same gender, we included the same set of variables in both final models of a gender. Adjusted odds ratios (aORs), 95% CIs and p-values of the final models were reported.

We ran two additional sensitivity analyses: (1) excluding the years 2015–2018 because of lower data completeness in this period, resulting from changes in data collection systems and variable definitions, which were updated in 2019 and led to fewer missing values thereafter; this meant that ‘missing’ categories were not needed anymore for the variables anal sex, partners originating from an STI/HIV-endemic area or (for women only) MSM partners, and self-reported bacterial STI, and (2) excluding the time period with COVID-19 restrictions (13 March 2020–24 February 2022) as CSHs prioritized STI care for those with an increased risk of acquiring an STI (including individuals with symptoms). All analyses were performed in Stata 17.0 (Stata Corporation, College Station, TX, USA).

Ethics approval

No ethical approval was necessary as we used routinely collected, deidentified surveillance data.¹⁹

Role of the funding source

This study was conducted without dedicated external funding; all resources were provided by the authors’ host institution.

Results

There were 249,960 consultations among women and heterosexual men at the CSH Amsterdam between 1 January 2015 and 31 December 2023. We excluded 4042 consultations where no CT test was performed, 3487 test-of-cure consultations, and 1426 consultations for other reasons (e.g., men tested anorectally, age <16 years, or missing data on symptoms and partner notification) (Supplementary Figure S1). A total of 241,005 consultations of 124,527 individuals were included in the analysis, of which 168,201 (69.8%) consultations were among 81,824 women and 72,804 (30.2%) consultations among 42,703 men. The median age at first consultation was 22 years (interquartile range [IQR] = 20–24) for women and 24 years (IQR = 21–28) for men, respectively (Table 1). The majority of women (n = 57,684, 70.5%) and men (n = 25,906, 60.7%) had started or completed university or university of applied sciences education. Both women and men reported a median of 3 (IQR = 2–4 for women and IQR = 2–5 for men) sexual partners in the preceding six months.

Table 1.

Socio-demographic characteristics and sexual behaviour reported at first consultation of women and heterosexual men tested for CT at the Center for Sexual Health Amsterdam, the Netherlands, January 2015–December 2023.

	Women N = 81,824 n (%)a	Men N = 42,703 n (%)a
Age (years)
Median (IQR)	22 (20–24)	24 (21–28)
16–19	12,355 (15.1)	4005 (9.4)
20–24	50,306 (61.5)	22,180 (51.9)
25–34	15,280 (18.7)	12,312 (28.8)
≥35	3882 (4.7)	4205 (9.8)
Missing	1 (0.0)	1 (0.0)
Educational levelb
University or university of applied sciences	57,684 (70.5)	25,906 (60.7)
Up to secondary (general/vocational) education	23,508 (28.7)	16,139 (37.8)
Missing	632 (0.8)	658 (1.5)
Originating from STI/HIV-endemic areac
No	60,782 (74.3)	26,831 (62.8)
Yes	20,964 (25.6)	15,825 (37.1)
Missing	78 (0.1)	47 (0.1)
Gender of sexual partnersd
Same/both gender(s)	5439 (6.6)	NA
Opposite gender only	76,383 (93.4)	42,703 (100)
Missing	1 (0.0)	NA
Number of sexual partnersd
Median (IQR)	3 (2–4)	3 (2–5)
0–1	19,315 (23.6)	7439 (17.4)
2–4	44,968 (55.0)	20,766 (48.6)
≥5	16,895 (20.6)	14,194 (33.2)
Missing	646 (0.8)	304 (0.7)
Condom use with casual partnersd, h,e
No casual partner(s)	20,455 (25.0)	10,267 (24.0)
With condom	6485 (7.9)	4611 (10.8)
Without condom	50,585 (61.8)	25,934 (60.7)
Missing	4299 (5.3)	1891 (4.4)
Anal sexd, h,e, f
No	67,054 (81.9)	29,756 (69.7)
Yes	14,033 (17.2)	4867 (11.4)
Missing	737 (0.9)	8080 (18.9)
Partners originating from an STI/HIV-endemic area or (for women only) MSM partnersd, h,f, g
No	54,628 (66.8)	24,674 (57.8)
Yes	20,834 (25.5)	13,344 (31.2)
Missing	6362 (7.8)	4685 (11.0)
Sex workd, h
No	78,069 (95.4)	42,560 (99.7)
Yes	3706 (4.5)	109 (0.3)
Missing	49 (0.1)	34 (0.1)
Client of sex workerd, h
No	NA	39,026 (91.4)
Yes	NA	2793 (6.5)
Missing	NA	884 (2.1)
Self-reported bacterial STI in preceding yearg, h
No	68,539 (83.8)	36,722 (86.0)
Yes	6060 (7.4)	1961 (4.6)
Missing	7225 (8.8)	4020 (9.4)
HIV status
Negative	81,794 (100)	42,683 (100)
Positive	30 (0.0)	20 (0.0)

Abbreviations: CT, Chlamydia trachomatis; IQR, interquartile range; NA, not applicable; MSM, men who have sex with men; STI, sexually transmitted infection.

^aUnless indicated otherwise.

^bUp to secondary (general/vocational) education includes no education; elementary school; pre-vocational secondary education; secondary vocational education; senior general secondary education; pre-university education.

^cSTI/HIV-endemic areas include Asia, Africa, the Dutch Caribbean islands, middle and South America.

^dIn the preceding six months.

^eData prior to 2018 referred to the last sexual encounter, while data from 2018 onward referred to the preceding six months.

^fFor women: receptive anal sex; for men: insertive anal sex.

^gFor women: partner originating from an STI/HIV endemic area or a male partner who had sex with men; for men: partner originating from an STI/HIV endemic area.

^hBacterial STIs include chlamydia, gonorrhoea and syphilis.

The median number of consultations per individual was 1 (IQR = 1–2) for both women and men (Table 2). Symptoms were reported in 14.1% (n = 23,678/168,201) of consultations among women and 23.6% (n = 17,180/72,804) of consultations among men, and partner notifications were reported in 13.0% (n = 21,937/168,201) of consultations among women and 25.2% (n = 18,345/72,804) of consultations among men.

Table 2.

Characteristics of consultations with women and heterosexual men tested for CT at the Center for Sexual Health Amsterdam, the Netherlands, January 2015–December 2023.

	Women n (%)a	Men n (%)a
Total number of consultations	168,201	72,804
Number of consultations per individualb
Median (IQR)	1 (1–2)	1 (1–2)
1	49,381 (60.4)	28,542 (66.8)
2	14,648 (17.9)	7536 (17.7)
3	6656 (8.1)	3122 (7.3)
≥4	11,139 (13.6)	3503 (8.2)
Days since previous consultation in study period,c median (IQR)	193 (105–371)	254 (130–498)
Partner notification reported
No	146,263 (87.0)	54,459 (74.8)
Yes	21,937 (13.0)	18,345 (25.2)
Missing	1 (0.0)	0 (0.0)
Symptoms reported
No	144,521 (85.9)	55,624 (76.4)
Yes	23,678 (14.1)	17,180 (23.6)
Missing	2 (0.0)	0 (0.0)
Type of symptomsd
Burning sensation	4604 (2.7)	9220 (12.7)
Discharge	12,835 (7.6)	4490 (6.2)
Bleeding	2504 (1.5)	NA
Ulcers	1472 (0.9)	1188 (1.6)
Abdominal pain	3636 (2.2)	NA
Other	8610 (5.1)	6444 (8.9)
CT detection
No	147,032 (87.4)	61,613 (84.6)
Yes	21,169 (12.6)	11,191 (15.4)
NG detection
No	165,330 (98.3)	71,443 (98.1)
Yes	2871 (1.7)	1361 (1.9)
Concurrent CT and NG detection
No	167,265 (99.4)	72,277 (99.3)
Yes	936 (0.6)	527 (0.7)
PID diagnosis
No	168,026 (99.9)	NA
Yes	175 (0.1)	NA
Epididymitis diagnosis
No	NA	72,770 (100)
Yes	NA	34 (0.0)
Eligibility for testing under new guideline
Not eligible	126,554 (75.2%)	40,922 (56.2%)
Eligible	41,647 (24.8%)	31,882 (43.8%)

Abbreviations: CT, Chlamydia trachomatis; IQR, interquartile range; NA, not applicable; NG, Neisseria gonorrhoeae; PID, pelvic inflammatory disease.

^aUnless indicated otherwise.

^bOnly one value per individual (N = 81,824 women and N = 42,703 men).

^cOnly applicable to individuals with ≥2 consultations within the study period, corresponding to n = 86,377 consultations with women and n = 30,101 consultations with men.

^dIndividuals could report more than one symptom.

Testing eligibility

Under the new CT testing guideline, CT testing would not have been done in 167,476 (69.5% [95% CI 69.3%–69.7%] of n = 241,005) consultations: of which 100,756 consultations were among women <25 years (60.2%), 25,798 consultations among women ≥25 years (15.4%) 28,624 consultations among men <25 years (17.1%), and 12,297 (7.3%) consultations among men ≥25 years (Table 3; Fig. 1). This corresponds to an average of 18,608 CT tests per year that would not have been done.

Table 3.

Number of consultations and Chlamydia trachomatis (CT) detections for women and heterosexual men who would have been eligible and not eligible for CT testing under the new guideline, Center for Sexual Health Amsterdam, the Netherlands, January 2015–December 2023.

	Eligibility for testing under the new CT testing guideline	Consultations, n (%)	No CT detected, n (%)	CT detected, n (%)	CT positivity, %
Total	Eligible	73,529 (30.5)	58,940 (28.2)	14,589 (45.1)	19.8
	Not eligible	167,476 (69.5)	149,705 (71.8)	17,771 (54.9)	10.6
Women <25 y	Eligible	23,537 (18.9)	17,994 (16.9)	5543 (31.4)	23.6
	Not eligible	100,756 (81.1)	88,664 (83.1)	12,092 (68.6)	12.0
Women ≥25 y	Eligible	18,109 (41.2)	16,073 (39.8)	2036 (57.6)	11.2
	Not eligible	25,798 (58.8)	24,300 (60.2)	1498 (42.4)	5.8
Men <25 y	Eligible	14,692 (33.9)	10,481 (29.4)	4211 (55.1)	28.7
	Not eligible	28,624 (66.1)	25,196 (70.6)	3428 (44.9)	12.0
Men ≥25 y	Eligible	17,190 (58.3)	14,391 (55.5)	2799 (78.8)	16.3
	Not eligible	12,297 (41.7)	11,544 (44.5)	753 (21.2)	6.1

Abbreviation: CT, Chlamydia trachomatis.

Fig. 1.

(a) Number of consultations with Chlamydia trachomatis (CT) testing in people who would have been eligible and not eligible for testing under the new CT testing guideline; (b) Number of CT detections in people who would have been eligible and not eligible for testing under the new CT testing guideline. Both are stratified by gender (women and heterosexual men) and age group, at the Center for Sexual Health Amsterdam, the Netherlands, January 2015–December 2023. Abbreviation: CT, Chlamydia trachomatis.

Undetected CT

A total of 17,771 CT detections were found in consultations among individuals who would not have been eligible for testing under the new guideline (=undetected CT), corresponding to 54.9% (95% CI 54.4%–55.5%) of the total number of detections (n = 32,360). The highest number of undetected CT cases was found among women <25 years (n = 12,092/17,771, 68.0%), resulting in an average of 1343 undetected cases per year in this group (Supplementary Table S1). The average number of undetected CT cases per year was 166 for women ≥25 years, 380 for men <25 years, and 83 for men ≥25 years. The majority of CT detections in women not eligible for testing were single urogenital infections (<25 years: 68.1%; and ≥25 years: 49.9%) (Supplementary Table S2). In contrast, the majority of detections in women eligible for testing were concurrent urogenital and anorectal infections (<25 years: 57.5%; and ≥25 years: 54.1%).

When comparing NG detections between groups that would have been eligible and not eligible for CT testing, more than half of NG detections among women <25 years were found in the not eligible group (n = 1207/2072, 58.2%) (Supplementary Table S3). Among women ≥25 years, 48.3% (n = 386/799) of NG detections occurred in those who would no longer have been eligible for CT testing. For men, the proportion of NG detections was lower across groups that would not have been eligible for CT testing (<25 years: n = 267/600, 30.8%; ≥25 years: n = 51/494, 10.3%) compared to groups that would have been eligible. Within the groups that would no longer have been eligible for CT testing, the proportion of individuals with an NG detection was higher among those who also tested positive for CT compared to people who tested negative for CT (women <25 years, 3.3% vs. 0.9%; women ≥25 years, 4.3% vs. 1.3%; men <25 years, 3.2% vs. 0.6%; men ≥25 years, 3.6% vs. 0.2%; all p < 0.0001).

Determinants of undetected CT

The results of the univariable analysis of factors associated with undetected CT (i.e. a CT detection among those who would no longer have been eligible for testing) are shown in Supplementary Table S4 (women) and Supplementary Table S5 (heterosexual men). Although associations were mainly weak to moderate, multivariable models across all gender and age strata showed that several factors were associated with a higher risk of undetected CT (Fig. 2; Supplementary Table S6). Among these factors was younger age (<25 years: 16–19 vs. 20–24 years, ≥25 years: 25–34 vs. ≥35 years; women <25 years, aOR = 1.58 [95% CI 1.50–1.67]; women ≥25 years, 1.46 [1.24–1.73]; men <25 years, 1.44 [1.30–1.61]; men ≥25 years, 1.67 [1.33–2.07]). We also observed an association for a higher number of recent sexual partners. For 2–4 vs. 0–1 partners, the aORs were: women <25 years, 1.41 [95% CI 1.33–1.49]; women ≥25 years, 1.23 [1.04–1.46]; men <25 years, 1.52 [1.33–1.73]; men ≥25 years, 1.32 [1.60–2.14]. For ≥5 vs. 0–1 partners, the aORs were: women <25 years, 1.81 [95% CI 1.68–1.95]; women ≥25 years, 1.25 [1.01–1.54]; men <25 years, 1.85 [1.60–2.14]; men ≥25 years, 1.53 [1.15–2.03]. A self-reported bacterial STI in the preceding year was also associated with a higher risk of undetected CT: women <25 years, aOR = 1.32 [95% CI 1.26–1.39]; women ≥25 years, 1.56 [1.32–1.84]; men <25 years, 1.51 [1.36–1.68]; men ≥25 years, 2.15 [1.70–2.72].

Fig. 2.

Determinants of Chlamydia trachomatis (CT) detection in women and heterosexual men who would no longer have been eligible for testing under the new CT testing guideline, visiting the Center for Sexual Health Amsterdam, the Netherlands, January 2015–December 2023. Results were based on multivariable logistic regression with robust standard errors clustered at the individual level. As we aimed to compare people <25 and ≥ 25 years of the same gender, we included the same set of variables in the final models within the same gender. Models were adjusted for all other variables in the plot. ∗In the preceding six months. Abbreviations: aOR, adjusted odds ratio; mo, months; STI, sexually transmitted infection; y, years.

Among all women, recent anal sex was moderately associated with a higher risk of undetected CT (<25 years, aOR = 1.44 [95% CI 1.38–1.51]; ≥25 years, 1.36 [1.20–1.54]). Factors associated with an increased risk of undetected CT only in women <25 years were: up to secondary (general/vocational) educational level (vs. university or university of applied sciences; aOR = 1.32 [95% CI 1.26–1.39]) and having exclusively partners from the opposite gender (vs. partners from the same gender or both genders; aOR = 1.57 [95% CI 1.41–1.75]). Similarly, originating from an STI/HIV-endemic area was associated with undetected CT in women <25 years, although the risk was only slightly higher compared to the reference category (not originating from an STI/HIV-endemic area; aOR = 1.08 [95% CI 1.02–1.14]). Factors associated with a lower risk of undetected CT were using condoms with casual partners (vs. no casual partners; <25 years, aOR = 0.81 [95% CI 0.72–0.91]; ≥25 years, 0.71 [0.57–0.87]) and sex work (<25 years, aOR = 0.59 [95% CI 0.50–0.70]; ≥25 years, 0.79 [0.65–0.95]).

Among all men, up to secondary (general/vocational) education was moderately associated with an increased risk of undetected CT (<25 years: aOR = 1.28 [95% CI 1.18–1.39]; ≥25 years, 1.35 [1.15–1.60]). Originating from an STI/HIV-endemic area was associated with a slightly higher undetected CT risk only in men <25 years (aOR = 1.15 [95% CI 1.06–1.26]). In men ≥25 years, not using condoms with casual partners was associated with an increased undetected CT risk (aOR = 1.29 [95% CI 1.04–1.60]). Sexual behaviour variables associated with a modestly lower risk of undetected CT in all men were: having partners originating from an STI/HIV-endemic area (<25 years: aOR = 0.78 [95% CI 0.71–0.86]; ≥25 years, 0.81 [0.68–0.96]) and being client of a sex worker (<25 years: aOR = 0.55 [95% CI 0.43–0.69]; ≥25 years, 0.58 [0.43–0.79]). In men <25 years, factors associated with a slightly decreased undetected CT risk were using condoms with casual partners (aOR = 0.74 [95% CI 0.61–0.89]) and recent anal sex (aOR = 0.73 [95% CI 0.62–0.86]).

Sensitivity analyses

The sensitivity analysis where missing data were imputed using MICE yielded comparable results to our main analysis (Supplementary Table S7). In the sensitivity analysis excluding the years 2015–2018, the majority of associations remained consistent (Supplementary Tables S8 and S9). However, the previously observed weak associations between region of origin and undetected CT across all gender and age strata were no longer present, and among women, condom use with a casual partner was no longer associated with undetected CT. The sensitivity analysis excluding the time period with COVID-19 restrictions in the Netherlands also showed similar results in terms of the proportion of CT detections, positivity and determinants in groups not eligible for testing (Supplementary Tables S10 and S11). However, for women <25 years, the weak association between region of origin and undetected CT was no longer present. Similarly, for women ≥25 years, educational level was no longer associated with undetected CT.

Discussion

In this retrospective analysis of data from the CSH Amsterdam (2015–2023), we determined that in almost 70% of consultations with women and heterosexual men, no CT test would have been performed if the revised CT testing guideline had been in place. As a consequence, the number of CT detections would have been reduced by more than half. The highest number of undetected CT cases would have been in women below the age of 25 years. We observed that younger age, up to secondary (general/vocational) education (either started or completed), and a bacterial STI in the preceding year were associated with an increased risk of undetected CT, although the strength of these associations was modest.

To our knowledge, this study is the first to estimate the potential impact of reduced testing for asymptomatic CT in women and heterosexual men on the number of CT tests and diagnoses. Furthermore, we used data from the largest CSH in the Netherlands over several years, resulting in a high number of consultations. However, several limitations need to be addressed. First, we applied a counterfactual scenario to the data, and could not account for any potential changes in client testing or reporting behaviours that may follow the guideline change. For instance, individuals without symptoms might turn to their GP for CT testing, or they might be more likely to report symptoms under the new guidelines to increase their chances of receiving CT testing at the CSH. Future studies evaluating the real-world impact of the guideline change should examine whether there are changes in the proportion of consultations in which symptoms are reported before and after implementation. Moreover, reduced testing will result in fewer diagnoses which in turn leads to fewer people that are notified by a sexual partner. As a result, the size of the group not eligible for CT testing might be larger than estimated here. Second, since our analysis was limited to data from a single CSH, and client characteristics, number of CT tests performed, and positivity differ across settings,²⁰ our findings might not be directly generalizable to all clients attending STI screening. However, a substantial proportion of consultations at CSHs is among women under the age of 25 years, suggesting that in other regions young women are mostly represented in the group experiencing reduced testing. Last, there have been changes in definitions and reporting of a few variables over the years, which resulted in a substantial number of missing values in some variables. Nevertheless, a sensitivity analysis where missing values were imputed, as well as a sensitivity analysis restricted to years with few missing values, did not affect our results.

We looked at determinants of CT detection in the group that would no longer have been eligible for CT testing under the new guideline. Comparing our findings with the existing literature is not straightforward, as other studies typically include people with symptoms or a partner notification, whereas these factors were not present in our analysis of determinants of CT detection. However, our identified determinants were largely consistent with those reported in studies on general determinants of CT detection among women and heterosexual men in higher-income countries,^7,18,21, 22, 23, 24 although the strength of these associations varied both across previous studies and in comparison to our findings. For example, a Dutch national population-based survey reported that a low/medium educational level was a risk factor for CT among women aged 18–24 years, although a stronger association was observed (high education vs. low/medium education level: aOR = 0.25, 95% CI 0.08–0.69) compared to our data (up to secondary (general/vocational) education vs. university or university of applied sciences: aOR = 1.32, 95% CI 1.26–1.39).¹⁸ Another population-based study in Denmark reported younger age, a lower educational level, a higher number of sexual partners and not using condoms to be associated with an increased risk for incident CT.²² Younger age and a higher number of sex partners were reported as risk factors for CT among attendants of an STI clinic in Sweden.²³ This suggests that commonly reported CT determinants are also related to undetected CT. Conversely, in our study, insertive anal sex was associated with a lower risk of having undetected CT for heterosexual men below the age of 25 years. This contrasts with findings from a population-based survey and a survey among STI clinic visitors showing that heterosexual men who report insertive anal sex are more likely to engage in other sexual behaviours associated with a higher STI risk.^25,26 A possible explanation for this discrepancy in findings is that heterosexual men more often attend the CSH due to symptoms or a partner notification, diluting the role of insertive anal sex as a specific risk factor found in other studies.

We observed that most CT detections in women who would not have been eligible for testing were single urogenital infections, whereas most CT detections in women who would have been eligible for testing were concurrent urogenital and anorectal infections. While biological factors may contribute to this difference (e.g., symptomatic women might have a higher CT load than asymptomatic women,²⁷ and a higher vaginal CT load has previously been shown to increase the risk of incident rectal CT),²⁸ the disparity is more likely explained by the difference in anorectal testing guidelines based on symptom presentation. Guidelines recommended testing women with symptoms at both urogenital and anorectal sites, whereas no such recommendations existed for asymptomatic women. In line with these guidelines, 98.9% of symptomatic women in our data were tested anorectally, compared to only 41.5% of asymptomatic women.

The expected substantial reduction in CT detections reported in our study will likely result in a substantial decrease in azithromycin and doxycycline consumption. Considering the growing concern about antimicrobial resistance and the impact of antibiotic treatment on the healthy microbiome,^2,3 this reduction likely has positive consequences on the individual and population level. Furthermore, a chlamydia diagnosis may lead to feelings of stigma and anxiety regarding possible future infertility,²⁹ which could lead in some cases to psychological stress. Reducing unnecessary diagnoses may therefore not only help limit antimicrobial use, but also lessen the psychological burden on individuals, ultimately supporting a chlamydia control approach more focused on disease management. In this context, it is important to consider that many asymptomatic CT infections may have limited pathogenic potential and could represent transient colonization rather than progressive infection.³⁰ On the other hand, chlamydia screening has been found to be largely acceptable to women in various settings,³¹ and reduced testing might induce negative psychological impact, as individuals may miss out on the reassurance that a negative test result can provide.³²

Although the revised Dutch guideline specifically means a reduction in testing for CT and not for other STIs, reduced CT testing could have unintended side-effects on healthcare-seeking behaviour. For instance, women and heterosexual men might perceive testing for STIs in general as less important due to the guideline change. This could lead individuals to seek less testing for STIs in general, potentially leading to delayed diagnoses and treatment for a broader range of STIs. STI consultations also provide an opportunity for counselling and testing for blood-borne viruses, such as HIV. The impact on HIV testing will likely be limited, because HIV testing is not routinely offered to young women and heterosexual men as incidence is low.³³ However, this is more concerning in the case of gonorrhoea, for which prevalence among young women and heterosexual men is rising in the Netherlands^15,34 and other European countries.³⁵ Our findings indicate that NG and CT detections frequently co-occurred among people who would no longer have been eligible for CT testing. If this group reduces STI testing in general, this might impact NG detection. Since gonorrhoea cases can be asymptomatic, delayed testing for asymptomatic gonorrhoea may accelerate transmission and/or lead to complications and/or hinder timely identification of newly introduced or acquired resistant strains.³⁶ Together with the rise of gonorrhoea in the Netherlands, these findings underscore the need for communication strategies addressing the guideline change, emphasizing the importance of continuing testing for other asymptomatic STIs. It also emphasizes the need to strengthen primary prevention efforts. This is particularly urgent given the declining condom use among younger people in the Netherlands,¹⁵ despite primary prevention being a key strategy in STI management. Renewed campaigns promoting condom use and innovations in sexual health care, such as digital vending machines dispensing condoms at little or no cost,³⁷ could reinforce long-term strategies for reducing the burden of STIs in the population.

We found that factors moderately associated with a higher undetected CT risk under the new guideline were younger age and up to secondary (general/vocational) education. It is known that these factors are related to a higher risk of STIs^7,15 as well as adverse sexual health outcomes in general.³⁸ Yet, there are challenges in providing sexual health services to younger populations and those with up to secondary-level education.¹⁵ If those individuals would seek less STI testing in general as an unintended consequence of the guideline change, this might, to some extent, lead to delayed diagnosis and treatment of STIs other than chlamydia, potentially further exacerbating health disparities across different populations. This may also apply to individuals from socioeconomically disadvantaged backgrounds, although our dataset did not include information on such characteristics. Therefore, communication about the guideline change and STI prevention strategies should specifically target these already underserved populations. In addition, a prospective cohort study among Dutch women suggested that the risk of PID among asymptomatic women with CT was comparable regardless of whether they were treated,⁹ underscoring the continued importance of primary prevention in the context of chlamydia.

Importantly, while previous studies from high-income settings suggest that widespread asymptomatic CT testing does not reduce chlamydia prevalence,5, 6, 7 limiting asymptomatic CT testing requires ongoing monitoring of symptomatic CT detections, chlamydia-associated complications, and NG detections to detect any potential unintended consequences of the guideline change. This is especially relevant given that repeat diagnoses have been associated with an increased risk of complications,³⁹ although it has also been hypothesized that early detection and treatment through widespread testing may interrupt the development of a partially protective immune response.⁴⁰ Since our study focused on quantifying the potential reduction in chlamydia diagnoses rather than the potential clinical impact, modelling studies are crucial to explore how reduced asymptomatic testing may influence transmission dynamics and the development of sequelae at the population level. These studies would provide essential information for a comprehensive understanding of the long-term implications of the revised guidelines.

In conclusion, restricting CT testing to women and heterosexual men who have symptoms or were notified for chlamydia by a partner only might lead to substantial reductions in the number of CT tests, CT detections, and related treatment, thereby substantially reducing antimicrobial consumption. As many European countries are reconsidering their approach to CT testing,^11,12 insights from this study may contribute to a broader understanding of potential public health implications of restricting testing. Ongoing surveillance of symptomatic CT, long-term complications, and NG detections will be essential to ensure that the population-level benefits of decreased antimicrobial use for CT are realized while monitoring for any unintended consequences.

Contributors

JCMH, LW, and VWJ designed the study. LW and VWJ accessed and verified the data. JCMH, LW, and VWJ analysed the data. LW drafted the manuscript. All authors contributed to data interpretation, and reviewed and approved the final version of the manuscript. JCMH had final responsibility for the decision to submit for publication.

Data sharing statement

Data are available upon reasonable request. On request to the corresponding author (lwerner@ggd.amsterdam.nl), the following data will be available after publication: de-identified participant data. Data will be shared after approval of an analysis proposal by all authors (LW, VWJ, SMB, MFSvdL, EH, HdV, JCMH).

Declaration of interests

The institution of MFSvdL receives study funding from GSK for an investigator-initiated study.