Increasing trends of lymphogranuloma venereum among HIV-negative and asymptomatic men who have sex with men, the Netherlands, 2011 to 2017

Introduction Lymphogranuloma venereum (LGV), an invasive form of Chlamydia trachomatis infection, has been reported among (mainly HIV-positive) men who have sex with men (MSM) since 2003. In the Netherlands, LGV testing recommendations changed from selective to universal testing in 2015. Changes in tested populations could have led to incomparable LGV positivity rates over time. Aim We investigated LGV trends among MSM attending Centres for Sexual Health using surveillance data between 2011 and 2017. Methods LGV positivity was calculated among MSM tested for rectal Chlamydia infection and MSM tested specifically for LGV. With multivariable logistic regression analysis, the association between years and LGV was adjusted for testing indicators and determinants. Results We included 224,194 consultations. LGV increased from 86 in 2011 to 270 in 2017. Among LGV-positives, proportions of HIV-negative and asymptomatic MSM increased from 17.4% to 45.6% and from 31.4% to 49.3%, respectively, between 2011 and 2017. Among MSM tested for rectal chlamydia, LGV positivity increased from 0.12% to 0.33% among HIV-negatives and remained stable around 2.5% among HIV-positives. Among LGV-tested MSM, LGV positivity increased from 2.1% to 5.7% among HIV-negatives and from 15.1% to 22.1% among HIV-positives. Multivariable models showed increased odds ratios and significant positive associations between years and LGV. Conclusions Although increased testing and changes in LGV incidence are difficult to disentangle, we found increasing LGV trends, especially when corrected for confounding. LGV was increasingly attributed to HIV-negative and asymptomatic MSM, among whom testing was previously limited. This stresses the importance of universal testing and continuous surveillance.


Background
Lymphogranuloma venereum (LGV) is a sexually transmitted infection (STI) caused by invasive L1, L2 and L3 serovars of Chlamydia trachomatis (CT). Rectal LGV has been reported at a low level among men who have sex with men (MSM) in Europe since 2003, while urogenital infections are rarely seen [1][2][3]. Rectal LGV can cause proctitis or proctocolitis characterised by severe rectal pain, mucoid and/or haemorrhagic rectal discharge, tenesmus, diarrhoea or constipation and also systemic symptoms such as fever, malaise and weight loss. If left untreated, LGV proctitis can lead to chronic colorectal fistulas and strictures [4,5]. In addition, the ulcerative nature of LGV may increase the risk of STI and HIV transmission [6,7]. Differentiation of the LGVcausing L serovars from the non-LGV (D-K) serovars is important, as rectal LGV requires an extended doxycycline treatment regimen (21 days) compared with non-LGV rectal infection (7 days) [8,9].
To date, the exact LGV transmission dynamics are not well understood. A previous Dutch study found a rectal-urethral LGV ratio of 15:1 and therefore concluded that urethral LGV could not explain all rectal LGV transmissions [3]. Indications for tissue tropism have not been found. Urethral epithelium may be infected only briefly, long enough to contribute to ongoing transmission, but too short to become symptomatic and clinically detected [10]. This could explain why LGV is confined to networks of MSM engaged in high-risk sexual behaviours, including high rates of partner change. An oro-anal transmission route has been proposed to further explain the rectal-urethral LGV ratio [11].
Previous studies and the European Centre for Disease Prevention and Control underlined the importance of active LGV testing and the implementation of surveillance systems [4,7,12,13]. Still, surveillance programmes have not been implemented in many European countries owing to the limited availability of diagnostics [13]. In the Netherlands, LGV was added to the routine national STI surveillance system of Centres for Sexual Health (CSH) as part of the response to the outbreak in 2003. In addition, LGV testing recommendations were developed, which included rectal CT testing for only MSM who reported having had receptive anal sex. Among those with a positive rectal CT test result, LGV testing was recommended, in particular for HIV-positive MSM or MSM presenting with LGV-related symptoms.
Between 2006 and 2012, the number of LGV diagnoses fluctuated, but the incidence remained consistently higher than before the outbreak in 2003 most probably resulting from increased testing after 2003. However, it was estimated that selective LGV testing could have resulted in an underdiagnosis of 20% between 2006 and 2012, as LGV was also diagnosed among asymptomatic MSM and among HIV-negative MSM [12]. Therefore, the CSH testing recommendations changed in 2015 from selective to universal rectal chlamydia testing for all MSM and universal LGV testing in all rectal chlamydia-positive MSM [14].
Changes in the indications for LGV testing over time complicate the interpretation of LGV positivity trends (confounding by indication). A decreased positivity is to be expected when MSM at lower risk for LGV are increasingly tested, the incidence remains stable and LGV is truly uncommon in the population among whom testing was limited before. Because of selective testing procedures, positivity is usually calculated among only those positive for rectal CT who were further tested for LGV, thus excluding those negative for rectal CT [12]. In a universal testing situation however, the LGV positivity rate could also be calculated including rectal CT negatives into the denominator, which would provide a more representative estimate [15,16].
This study aimed to investigate rectal CT testing, LGV testing and LGV positivity using different denominators among MSM attending CSH between 2011 and 2017. In addition, we aimed to explore LGV trends with changes in the characteristics of the population tested for LGV taken into account.

Methods
This surveillance study used routinely collected data from all consultations performed at CSH in the Netherlands between 2011 and 2017 [17].

Data source and variables
The CSH offer low-threshold and free of charge STI/HIV care exclusively for predefined high-risk populations. MSM are eligible for STI/HIV testing independent of risk behaviour. A selection of information registered in the electronic patient records of the CSH, including demographic variables, sexual risk behaviour, STI/HIV history, STI testing and diagnosis, was sent to the CSH surveillance database for national surveillance purposes. For the current study, all MSM (men who had sex only with men and men who had sex with both men and women) consultations between 2011 and 2017 were included. Trends including earlier years have been published previously [12].

Laboratory testing and case definition
Testing for LGV applied a two-step procedure. First, samples were tested by a generic nucleic acid amplification test (NAAT) for CT. Second, positive samples were further tested by a second PCR to discriminate between LGV and non-LGV genotypes. Only laboratoryconfirmed LGV cases are reported to the surveillance system. Indeterminate LGV test results were considered as negative in the analyses.

Data analysis
A client identification number was not available over the full study period and therefore the unit of analysis is 'STI consultation'. Therefore, repeat testers could have been counted several times over the period analysed. We calculated the proportion of consulting MSM who were tested for rectal CT infection, the proportion of rectal CT-positive MSM among those tested and the proportion of MSM who were tested for LGV serovars among those with a rectal CT infection.
LGV positivity was calculated by dividing the number of LGV-positives by the number of MSM who were (i) tested for rectal CT-infection, including those positive but not tested for LGV as 'non-LGV' and (ii) tested for LGV, thus excluding rectal CT-negative MSM. Proportions and positivity were stratified by year of consultation and by HIV status. All trends were explored using the using the Cochran-Armitage test for trends.
We performed uni-and multivariate logistic regression to adjust for possible confounders that could influence the association between year (categorical: 2011-17) and LGV. Indicators for LGV testing and variables that were previously identified as LGV risk factors using the same Dutch CSH surveillance data between 2006 and 2012, were added to the multivariate model: age group (< 35; 35-39; 40-44; ≥ 45 years), STI symptoms (no; yes; unknown), HIV status (negative; positive: known or new infections), sexual contact (sex with men only; sex with both men and women), number of sexual partners in the past 6 months (0-1; 2-5; 6-50; ≥ 51) and condom use in last contact (yes; no; unknown) [12]. Records with missing data were excluded from analysis.

Ethical statement
Ethical committee approval was not sought since we used de-identified routinely collected surveillance data.

Rectal CT and LGV testing
The proportion of MSM who were tested for rectal CT infection increased from 77.6% to 96.3% between 2011 and 2017 (  LGV  Table S2).
Overall, LGV positivity was higher among HIV-positives than among HIV-negatives over the full study period, but the trends were comparable for both groups, with a rapid increase and decrease in 2012 and 2013, respectively ( Figure 2). LGV positivity was low and stable around 0.6% (p = 0.232) when calculated among all MSM tested for rectal Chlamydia infection. Among HIVpositives the trend was stable around 2.5% (p = 0.534) and among HIV-negatives, the trend increased slightly but significantly from 0.12% (

Discussion
The Dutch CSH surveillance data showed that rectal CT testing, LGV testing and the number of LGV diagnoses increased among MSM between 2011 and 2017, with an increasing proportion of LGV attributed to HIV-negative or asymptomatic MSM. Given the increasing rectal CT and LGV testing trends, we expected positivity to decrease. In line with this, we found a decreasing trend for rectal CT. However, LGV positivity remained stable when calculated among all MSM tested for rectal CT infection and increased when calculated among LGVtested MSM. These crude LGV positivity trends were affected by changes in the population tested over time, as we found stronger and significant positive associations between years and LGV after correction for LGV testing indicators and risk factors.
A strength of the current study is that CSH surveillance data have national coverage including all consultations at all CSH, resulting in a large number of observations. Unique is the collection of data on LGV since the first outbreak in 2003. Limitations of the current study include the lack of information on LGV-related symptoms, LGV risk factors such as unprotected receptive anal sex and LGV testing and diagnoses at healthcare settings other than CSH. In addition, we were not able to correct for repeat testing in the logistic regression analyses as a unique client identifier was not available for the full study period (available only July 2014-2017). However, the results of a generalised estimating equation model correcting for within-subject correlations, including the years 2015 to 2017 only, did not differ from the results of a logistic regression model (data not shown). Lastly, the interpretation of LGV positivity is challenging because changes in rectal CT positivity are also captured if LGV positivity is calculated among all MSM tested for rectal CT infection. Using LGV-tested MSM only as denominator, selection bias limits the interpretation of LGV positivity. Therefore, all trends and changes in the characteristics of the population tested should be considered.
Internationally, testing procedures vary from country to country and long-term surveillance data on LGV is lacking, which hampers comparison of trends between countries. The number of reported LGV cases continues to increase in European countries, which may be explained by increased testing [13]. In the United Kingdom, the reported numbers of LGV were higher than in the Netherlands and a similar overall upward trend with three rapid increases in 2005, 2010 and 2014 was observed [18]. In 2016, however, the UK reported a small decrease in the number of reported cases, which coincided with lower numbers of gonorrhoea and HIV infections among MSM. In the Netherlands, the increasing number of LGV infections coincided with increasing numbers and positivity rates of gonorrhoea and infectious syphilis [17]. A rising trend in the number of LGV infections was also observed in France, which was explained by a combination of improved surveillance, better access to testing and an increasing incidence possibly due to increases in condomless anal sex among MSM [19]. Outbreak investigations and surveillance data consistently showed that the majority of LGV cases occurred among HIV-positive MSM and that most of the cases presented with proctitis [2,4]. In contrast to previous reports showing that this profile of individuals with LGV has remained largely unchanged over the years, we found increasing proportions of HIVnegative and asymptomatic MSM among LGV-positives, up to almost 50% in 2017 [2,13,19,20].
The increasing proportions of HIV-negative MSM and asymptomatic cases among LGV-infected MSM could be explained by increased testing and previous underdiagnosis. Besides changes in the testing policy, the increasing proportion of HIV-negative MSM among LGV-positives may also be related to HIV biomedical interventions in two different ways: On the one hand, greater awareness of 'undetectable is untransmittable' for HIV and the availability of pre-exposure prophylaxis (PrEP) might have resulted in increased risky behaviour including sexual mixing between HIVnegative and HIV-positive MSM [21]. On the other hand, increasing STI trends could occur among HIV-negative MSM regardless of changes in behaviour, as it could also be explained by a decrease in HIV seroconversion among MSM using PrEP exhibiting high-risk behaviour [17,22]. With respect to asymptomatic LGV, our results showed an increase in the proportion of asymptomatic cases among HIV-positive MSM only, among whom LGV testing was already high in earlier years. Thus, the change in testing policy may not be the sole explanation. A hypothetical explanation could be a change in circulating LGV strains over time, as observed in France where the predominating strain changed from L2b to L2 between 2010 and 2015 [23]. Although not supported by French data, less severe clinical features or no symptoms have been observed among patients infected with the L2 strain compared with patients infected with the L2b strain in Spain [24]. In the Netherlands, LGV control is based on offering universal rectal CT testing and universal LGV testing to all MSM attending CSH, followed by treatment and (voluntary) partner notification for those with LGV. Although recommended by the Dutch guideline for general practitioners (GPs), rectal testing is rarely performed by GPs [25]. Therefore, insight into the occurrence of rectal CT and LGV among MSM attending other clinical settings in the Netherlands is needed, which would help to increase awareness of the current LGV epidemic among Dutch GPs. In general practice, HIV PrEP prescribing may provide an opportunity to increase rectal CT and LGV testing as part of the routine 3-monthly PrEP checkups in a subgroup of high-risk HIV-negative MSM. In addition, asymptomatic infections deserve more attention, as an increasing number of asymptomatic infections may lead to delayed or missed diagnosis and consequently to increased LGV transmission. Another challenge of asymptomatic cases to LGV control is a longer recall period (6 months) for partner notification compared with symptomatic LGV (4-6 weeks before onset of symptoms) according to the partner management guideline for CSH [26]. In the literature, the relevance of detecting asymptomatic LGV is debated as a shorter course of doxycycline (7 days, similar to rectal Chlamydia infection) may effectively treat these infections. However, strong evidence is not available yet.

Conclusion
It is challenging to disentangle the effects of increased testing, previous underdiagnosis and changes in LGV incidence over time. Current LGV control measures are mostly performed by CSH, which may not be enough to interrupt LGV transmission. To guide additional prevention efforts, insight into LGV occurrence among MSM attending clinical settings other than CSH and insight into circulating LGV strains are needed. Furthermore, it is important to take changes in the characteristics of the tested population into account in the interpretation of trends as we found stronger and significant increasing LGV trends when corrected for confounding. Notably, we found increasing proportions of LGV attributed to HIV-negative and asymptomatic MSM, among whom testing was previously limited. These findings stress the importance of universal rectal CT and LGV testing for MSM irrespective of behaviour, HIV status or symptoms and the importance of continuous surveillance to improve LGV control and to improve insight into the epidemic.