Article Text

Original article
Extra-rectal lymphogranuloma venereum in France: a clinical and molecular study
  1. Arnaud Desclaux1,
  2. Arabella Touati2,3,
  3. Didier Neau1,2,
  4. Cécile Laurier-Nadalié2,3,
  5. Cécile Bébéar2,3,
  6. Bertille de Barbeyrac2,3,
  7. Charles Cazanave1,2,3
  1. 1 CHU Bordeaux, Service des Maladies Infectieuses et Tropicales, F-33000 Bordeaux, France
  2. 2 Univ. Bordeaux, USC EA 3671, Infections humaines à mycoplasmes et chlamydiae, F-33000 Bordeaux, France
  3. 3 Institut National de la Recherche Agronomique, USC EA 3671, Infections humaines à mycoplasmes et chlamydiae, F-33000 Bordeaux, France
  1. Correspondence to Dr Arnaud Desclaux, CHU Bordeaux, Service des Maladies Infectieuses et Tropicale, F-33000 Bordeaux, France; arnaud.desclaux{at}


Objectives To describe a series of extrarectal lymphogranuloma venereum (LGV) cases diagnosed in France.

Methods Consecutive LGV cases confirmed at the French Reference Centre for chlamydiae with an extrarectal sample from January 2010 to December 2015 were included. The first part of the study consisted of a retrospective case note review and analysis. In a second part, the complete ompA gene sequence of our samples was determined.

Results There were 56 cases overall: 50 cases of genital LGV and six cases of pharyngeal LGV. Subjects were all men, median age 39 years, 27/53 were HIV-positive, 47/51 reported having sex with other men, 43/49 reported multiple sexual partners (a mean 25 in the last 6 months). Median time from symptom onset to diagnosis was 21 days. Subjects most commonly presented with inguinal adenopathy alone (19 of 50 genital cases) and adenopathy with genital ulcer (17 of 50). Three pharyngeal cases were symptomatic. Fever was reported in 11 cases. Inguinal abscess was reported in 22 of 42 cases presenting with lymphadenopathy. Co-infections were frequent: eight cases of syphilis, four non-LGV Chlamydia trachomatis infections, one case of gonorrhoea. Cure was always achieved with doxycycline therapy but prolonged treatment was necessary in eight cases with inguinal abscess. Genotyping according to ompA sequencing showed the co-circulation of genovars L2 (16 of 42 strains successfully typed) and L2b (24 of 42). There was no association between HIV status and disease severity or genovar distribution.

Conclusion In the span of 6 years, 56 extrarectal LGV cases were confirmed through genotyping in France. Extrarectal LGV seemed to share a common epidemiological background with rectal disease in terms of affected population and genovar distribution. HIV prevalence was lower than expected.

  • chlamydia
  • lymphogranuloma venereum
  • lymphadenopathy
  • genital ulcer
  • HIV

Statistics from


Lymphogranuloma venerum (LGV) is a STI due to Chlamydia trachomatis (CT) genovars L1 to L3. It can manifest either as a genital ulceration followed by inguinal lymphadenopathy (genital LGV) or as an acute proctitis (rectal LGV), generally in men who have sex with men (MSM) who report unprotected anal sex. LGV was considered a tropical illness until an epidemic was detected in Europe in 2003 among MSM. Since then circulation of the disease has been increasing in this population and presents as rectal LGV in a vast majority of cases. Genital LGV is rarely described and accounts for approximately 5% of cases according to surveillance data from the UK.1 Hypotheses have been put forth to explain the rarity of genital cases and understand modes of transmission but have not been confirmed so far: preferential rectal tropism of the epidemic L2b genovariant, sexual practices such as rectal douching or fisting, host-specific factors such as HIV status.2 There is a need to clarify the place of genital and pharyngeal LGV in the current epidemic in terms of clinical presentation, at-risk population, efficacy of treatment and molecular characteristics of infecting strains. To do so, we sought to analyse extrarectal LGV cases confirmed at the French Reference Centre for chlamydiae which carries out every CT typing prescribed nationally.


Study population and setting

In 2010 a national surveillance network has been set up in France to monitor clinical and epidemiological characteristics of CT proctitis.3 It relies on voluntary laboratories and physicians throughout the country to send all rectal CT-positive specimens along with demographic, clinical and behavioural data, information on concurrent STI, and HIV status. This surveillance network was approved by the French Personal Data Protection Agency. Extrarectal LGV cases are usually excluded from this surveillance. If such a diagnosis is suspected, CT-positive samples can be sent to the French Reference Centre for chlamydiae on medical request. The diagnosis is then confirmed by real-time PCR.

We retrospectively analysed every extrarectal LGV case diagnosed in our laboratory from the implementation of the surveillance network in January 2010 until December 2015. All cases with an extrarectal sample positive for CT genovar L were included. Subjects gave their written consent after the diagnosis of CT infection as part of the LGV surveillance network.

LGV diagnosis

CT infection was initially diagnosed at the referring laboratory using a commercially available nucleic acid amplification test. Primary samples or DNA extracts were sent to our institution for genotyping. Total nucleic acid was extracted when necessary using the NucleoSpin Tissue kit (Macherey-Nagel). LGV genotyping was performed using a TaqMan real-time PCR exploiting a 36-base pair deletion in the pmpH gene specific to LGV strains.4 LGV-positive samples were further tested with a TaqMan real-time PCR targeting a 9-base pair insertion unique for the L2b variant in the pmpH gene.5

Data collection

Physicians in charge of the subjects at the time of the illness were asked retrospectively between August 2015 and March 2016 to fill an anonymised and standardised questionnaire. The information was retrieved through the subject’s medical record at the local institution. The following data were collected: demographic information, clinical information such as nature and duration of symptoms, sexual behaviour, concurrent or past STI, HIV status, nature of the treatment received and outcome. Data were combined in a standard Excel spreadsheet format.

ompA sequencing

To confirm the genotype identity of CT as determined by real-time PCR, the entire ompA gene of all pharyngeal, inguinal and genital LGV-positive samples included in our study was sequenced. Briefly, frozen DNA extracts were thawed and a nested PCR using two sets of primers (NLO and NRO then PTCM3 and SERO2A6) finally amplifying a 1010 bp sequence was carried out. Sanger sequencing was then performed in both directions. DNA sequences were read and edited with the BioEdit software, compared with other ompA sequences available in the BLAST algorithm and aligned using MultAlin.7

Statistical analysis

Our sample was described using mean, median and extreme values. Simple univariate analyses were performed using Student’s t-test to compare quantitative variables, χ2test and Fisher’s exact test to compare qualitative variables, with statistical significance reported at p<0.05.


Distribution of LGV cases

Fifty-six extrarectal LGV cases were diagnosed during the study period according to our definition. A sustained increase in the diagnoses could be seen during the last 3 years (42 cases, table 1). Thirty-seven cases (66.1%) were diagnosed in the Paris area. Testing for CT infection was prescribed in an STI clinic (41%), an internal medicine ward (40%), a general practice setting (9%), a surgical ward (8%) and an unknown location (2%). Subjects were all men, their median age at diagnosis was 39 years. Most of them were from France or the European Union (94%, omitting the nine patients whose nationality was unknown). One subject had two cases of genital LGV that took place 2.5 years apart.

Table 1

Characteristics of the 56 cases of extrarectal LGV

CT-positive samples

Sixty-two LGV-positive samples were included in our study. Twenty-five (40.3%) were recovered from lymph node aspirates or biopsies, 17 (27.4%) from genital ulcer swabs, 14 (22.6%) from first-void urine or urethral swabs, 6 (9.7%) from oropharyngeal swabs. In five patients, LGV CT could be recovered from both urine and lymph node (two cases), both urine and genital ulcer (two), the three different sites (one case). Co-infection with non-LGV CT was found in four cases (two urine samples, one anal and one throat swabs). Three more urine samples were positive for CT but have not been typed. CT serology was realised in 23 cases including two without lymphadenopathy and was always positive.

Clinical presentation

Of those 56 extrarectal LGV cases, we first have to separate the six defined by the presence of LGV CT in an oropharyngeal swab. Asymptomatic pharyngeal infection was found in two cases: one involved recent sexual contact with an LGV-infected partner, the other a non-LGV CT-positive anal sample. Three displayed oral symptoms: pharyngitis (inflamed pharynx and a sore throat) in one, presence of an ulcerated lesion on the tongue associated with cervical lymphadenopathy in the second, purulent pharyngeal exudate in the third. No clinical data were available for the last case.

A remaining 50 genital LGV cases were diagnosed during the 6-year period. None of them displayed rectal symptoms. Subjects mostly presented with inguinal adenopathy either isolated (19/50, 38%) or associated with genital ulceration (17/50, 34%). Six cases (12%) showed a combination of inguinal adenopathy and urethritis, with or without genital lesion. Urethritis or a penile ulceration was the only clinical sign in eight cases (16%).

Fever was reported in 11 cases and was always associated with the presence of lymphadenopathy. Inguinal lymphadenopathy was predominantly unilateral (29/42, 69%). Lymph nodes were usually tender (37/42, 88%) and inflammatory (29/42, 69%). An inguinal abscess, the classical ’inguinal bubo', with or without formation of a fistula was present in half of those cases (22/42, 52%).

Duration of symptoms

Time from beginning of symptoms to diagnosis of genital LGV varied greatly among cases. Overall, the mean duration of symptoms was 29.3 days (median 21 days) with values ranging from 2 days to 180 days (information was unavailable in 6/50 episodes). A box plot representation of these data (figure 1) shows the duration of symptoms according to the anatomical site where LGV CT was found. Diagnoses made on an ulcer swab or a urine sample tended to occur early in the disease process (median 7.5 days and 8.5 days, respectively) compared with those established by a lymph node aspirate/biopsy (32.5 days). Similarly, subjects who presented with inguinal adenopathy (alone or with genital signs) had probably suffered symptoms for a longer period of time (median 30 days and 21 days, respectively) than those who presented with urethritis or a genital ulcer alone (median 6 days).

Figure 1

Duration of symptoms from beginning of symptoms to collection of the lymphogranuloma venerum (LGV)-positive sample. Box plot representation with ends of the whiskers showing extreme values. Time to diagnosis depending on sample site. Gu, genital ulcer; IA, inguinal adenopathy. * p=0.014.

LGV CT was detected from urine in 14 cases. Of those, only 4 subjects presented with urethritis, the other 10 presented with inguinal adenopathy, genital ulceration or both. In two cases the urine sample was collected 75 days and 180 days after inguinal swelling was noticed.

HIV status and STI co-infections

Twenty-seven subjects were HIV-positive at LGV diagnosis corresponding to 28 LGV cases (51.9%, status was not documented in two cases). Median time from first HIV-positive serology was 5.5 years (mean 8.8 years, extremes 0.4–29 years, information unavailable in two cases). All subjects were on a combined antiretroviral therapy at the time of LGV diagnosis (information unavailable in three cases) and all were virologically suppressed (HIV viral load <50 copies/mL in 19/28, no data for the remaining 9 cases). CD4 cell count at LGV diagnosis was available in 20 cases and ranged from 350/mm3 to 1076/mm3 (median 541/mm3). Characteristics of LGV cases did not differ according to HIV status except for earlier diagnosis in HIV-positive subjects (Student’s t-test, p=0.018) (table 1).

Concurrent STIs were frequently diagnosed: active or recently treated syphilis (positive venereal disease research laboratory (VDRL) and treponemal test) in eight cases (14.8%), non L-genovar CT infection in four (8.2%), gonorrhoea in one (2%). A history of syphilis (positive treponemal test and negative VDRL) was documented in 21 cases (38.9%), hepatitis B in 8 according to anti-HBc antibody titre (no current infection found). Hepatitis B vaccination and hepatitis A vaccination were documented in 29 (isolated positive anti-HBs antibody in 20, history of hepatitis B virus vaccination in 9) of 47 (61.7%) and 12 (positive antihepatitis A virus (anti-HAV) antibody in 4, history of HAV vaccination in 8) of 37 (32.4%) cases, respectively. One patient had chronic hepatitis C (positive viral load).

Sexual behaviour

Forty-eight cases involved MSM, four involved men having sex with women, information was unavailable in the remaining four. In 43 cases, subjects reported more than one sexual partner in the six preceding months (median 15, no available information in seven cases). There was no history of travel outside of France in 38 of 39 cases. The infecting partner could be allegedly identified in 16 of 41 cases (39%) and was only rarely a stable partner (6/35, 17.1%). Inconsistent condom use (non-existent or occasional) was reported in 28 of 34 documented cases.

Antibiotic treatment and outcome

Forty-seven cases were treated with doxycycline (200 mg per day): treatment duration was 21 days in 39 cases and a mean 56 days in 8. Subjects treated for more than 3 weeks had an inguinal abscess at diagnosis. A follow-up visit was reported in 40 cases. Only one subject had persisting symptoms at the last visit: he had to be admitted after 3 weeks of therapy for febrile abdominal pain. An inguinal abscess was surgically drained, laparoscopic exploration of the peritoneum found a purulent exudate in the pouch of Douglas and an inflamed appendix. Inguinal pus was positive for CT genovar L. He had no rectal sign or symptom.

Seven cases were treated with azithromycin 1g orally as a single dose (five episodes) or at days 1, 8 and 15 (two episodes). Clinical presentation was urethritis, genital ulceration or asymptomatic pharyngeal infection. Treatment failure was reported in one subject who had received single-dose azithromycin for isolated urethritis. Outcome was not specified in four cases.

Treatment was not documented in two cases. No late complications such as lower limb/scrotal oedema or urethral stricture were described.

Genotyping according to ompA sequencing

Out of 62 samples, 42 complete ompA sequences were obtained from 42 different cases. Eleven DNA extracts could not be amplified, six samples were lost and three were not used (successful sequencing already achieved from another sample). Initial typing at the Reference Centre using real-time PCR had identified 36 L2b-positive specimens, 6 had been categorised as L strains.

The six samples from 2010 and 2011 had an ompA genotype identical to the reference L2b strain L2b/UCH-1/proctitis (AM884177.1). Out of 36 L2b-positive specimens, only 13 were confirmed through ompA sequencing as identical to the reference L2b strain since 2012 (36.1%), whereas 16 ompA sequences (38.1%) were actually identical to the L2 reference strain L2/434/Bu (AM884176.1). Five sequences differed from the L2b reference strain by a single C→A non-synonymous nucleotide polymorphism at position 493 (His165Asn) or 517 (Leu173Ile) (one and four cases, respectively). These L2b variants have already been described8–10 (figure 2). Finally two specimens were identical to genovars E and G despite the repeated positivity of the L genovar and L2b variant-specific real-time PCR.

Figure 2

Repartition of the genotypes involved in 40 cases of pharyngeal and genital lymphogranuloma venerum (LGV) according to ompA sequencing. The two sequences found to belong to genovars E and G are not shown.

Disease severity was not linked to a particular genovar. Ten of 22 genital cases (45%) infected with an L2b strain presented with an inguinal bubo compared with 9 of 15 genital cases (60%) infected with an L2 strain (χ2 test, p=0.6). The infecting strain did not significantly differ between HIV-positive and HIV-negative subjects (L2b 50% vs 45%, p=1; L2 45% vs 40%, p= 1; L2b variants 10% vs 13.6%, p=1).


This report is the largest case series including extrarectal LGV since the beginning of the epidemic in 2003 and the first French case series since 1989.11 During the last 6 years, 56 cases were confirmed at the French Reference Centre of chlamydiae of which none had concurrent rectal infection. According to the data provided by the national LGV surveillance network, they represent 3.1% of LGV cases diagnosed during the same period throughout the country confirming the rarity of this entity. Typing of extrarectal CT strains in France is only carried out on clinicians' demand; therefore the incidence of extrarectal LGV is probably underestimated.

However small in magnitude, we noticed a surge in the number of cases diagnosed from 2013 onwards, following the trend seen with rectal LGV in France and the UK.12 Cases exclusively involved men, largely MSM, with a median age of 39 years, fitting the characteristics of previously described rectal LGV cohorts.1 The main epidemiological difference consisted in the lower prevalence of HIV infection in our patients (51% vs 75%), in accordance with previous studies from the Netherlands and the UK.13 14 Seroadaptive practices among MSM such as seropositioning could potentially explain this finding (insertive rather than receptive anal sex is associated with less risk of HIV acquisition).15

Clinical findings differed slightly from historical descriptions16 (ie, prior to 2003): features of primary stage disease (genital ulceration or urethritis) were much more common than expected (62% of genital cases), and overlap with secondary stage (inguinal syndrome) was frequent (50% of genital cases). Nevertheless, diagnosis took a mean 30 days from the onset of symptoms to be reached, leaving many opportunities for transmission events to occur in a population engaging in high-risk sexual behaviour. We showed that the urethra can be a reservoir for shedding of CT genovar L until a diagnosis of inguinal LGV is made, even though case-finding studies have so far failed to demonstrate the importance of urogenital transmission of the disease among MSM.17–19 Our six pharyngeal cases featured either sexual contact with a partner infected with LGV, oropharyngeal symptoms or anal symptoms linked to non-LGV CT proctitis. In these situations, CT typing could be valuable.

We found no difference between HIV-positive and HIV-negative patients in terms of disease severity and co-infections but the comparison was hampered by the small sample size and missing data. We nonetheless noticed that HIV-seropositive patients were diagnosed significantly earlier. This could be due to easier access to appropriate medical care (HIV was successfully treated in all but one patient) or increased STI awareness leading to better screening and testing practices.

This study provides new data on genetic characteristics of LGV strains localised to the pharyngeal and genital areas. It first shows that the infecting strain belongs primarily to genovar L2b, also responsible for the vast majority of rectal LGV cases since 2003.5 20–22 It suggests a change in the molecular epidemiology of extrarectal LGV over time with the emergence of L2 strains in 2012 and L2b variants (differing from the reference strain by a single nucleotide polymorphism) in 2013. This shift has been observed as well in the last 5 years of the rectal LGV epidemic.9 10 23 Whether L2b strain is associated with a more severe clinical phenotype in rectal LGV is a debated issue. Our results did not show any impact of the LGV genovar on the severity of genital disease contrary to what limited clinical data had suggested.24 Nevertheless this finding is biased by the sampling of a large majority of symptomatic patients; systematic surveillance of these variants in the at-risk population would be necessary to reach a definitive conclusion. Finally, ompA sequencing of our specimens revealed that the L2b real-time PCR had misclassified/misidentified 47% of strains (16/34) as L2b. This finding confirms an earlier report that the 9-base pair insertion in the pmpH gene is no longer specific to genovar L2b.23 This target should therefore not be used for LGV typing. We hypothesise that recombination events occurring at the omp1 site between L2 and L2b could explain the discrepancies between omp1 sequencing and the so-called L2b-specific PCR. Identification of one genovar E and genovar G CT despite initial typing as LGV is most likely due to double infection, resulting possibly in the preferential amplification on the non-LGV strain. Recombination events between invasive and non-invasive strains cannot be excluded as such recombinants have already been described.25

The weakness of the study lies in its retrospective design. It hampered sexual behaviour data collection and prevented case-finding in the sexual contacts of our patients, leaving questions on the modes of transmission unanswered. The strengths of our work reside in the large number of included cases compared with previously published series,12 13 26 27 the detailed information retrieved for each case, and in the genetic characterisation of 42 extrarectal LGV strains.

To conclude, 50 genital and 6 pharyngeal LGV cases were diagnosed in France since 2010. The disease affected MSM in their late 30s engaging in high-risk sexual behaviour. Half the patients were HIV-positive. Genital LGV diagnosis was reached a month after symptoms onset. Genital ulcer and urethritis were surprisingly frequent at diagnosis. Evolution of inguinal bubo was slow but cure was always achieved under doxycycline therapy. Genotyping of extrarectal strains showed the circulation of L2 and L2b genovars similarly to the rectal LGV epidemic. Host and pathogen factors leading to the underrepresentation of genital LGV are still to be identified.

Key messages

  • From 2010 to 2015, 56 extrarectal  lymphogranuloma venereum (LGV) cases were PCR-proven in France.

  • Extrarectal LGV exclusively affected men of whom the vast majority were men who have sex with men. Fifty per cent were infected with HIV.

  • The infecting genovars were L2 and L2b.

  • HIV status and Chlamydia trachomatis genotype were not predictive of disease severity and response to antibiotic therapy.


The authors thank all the clinicians and biologists who participated in the study: Dr Karen AIDAN, Dr Philippe ARSAC, Dr Véronique BACLET, Dr Charlotte BIRON, Dr Jean-Marc BOHBOT, Dr Philippe BONHOMME, Dr Dorothée CHOPIN, Dr Nesrine DAY, Dr Dominique DECRÉ, Pr Constance DELAUGERRE, Dr Jean DEROUINEAU, Dr Philippe DHOTTE, Pr Nicolas DUPIN, Dr François DURUPT, Dr Cécile FICKO, Dr Sophie FLORENCE, Dr Sébastien FOUÉRÉ, Pr Pierre-Edouard FOURNIER, Dr Daniel GOSSET, Dr Claude GOURDIN, Dr Guillaume GRAS, Dr Claire GUGLIELMINOTTI, Dr Philippe GUIGARD, Dr Thomas GUIMARD, Dr Gérard ISRAËL, Pr Michel JANIER, Dr Philippe KRAEMER, Dr Ludovic LASSEL, Dr Marc de LAVAISSIÈRE, Dr Marie-Gisèle LEBRETTE, Dr Maeva LEFEBVRE, Dr Jean-Yves LIOTIER, Dr Jean-Baptiste LOUISON, Dr Laura MONFORT, Dr Yann NGUYEN, Dr Diane PONSCARME, Dr Bernard PROUVOST-KELLER, Dr Christian REGNIER, Dr Elliott SCHRETER, Dr Martin SIGUIER, Dr Gabriela SPIRIDON, Dr Françoise-Julie TIMSIT, Dr Didier TROISVALLETS, Dr Heïdi WILLE, Dr Stephen WILSON.


View Abstract


  • Contributors AD, BdB and CC wrote the paper. AT performed the molecular analyses, and described and reviewed the paper. CL-N performed statistical analysis. DN and CB reviewed and corrected the paper.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.