Non–Sexually Related Acute Genital Ulcers in 13 Pubertal Girls: Title and subTitle BreakA Clinical and Microbiological Study FREE

The most common causes of acute genital ulcers (AGU) are herpes simplex virus (HSV) and aphthosis. Since the beginning of the 20th century, several cases of virgin patients with AGU that were not consistent with HSV infection or aphthosis were reported. In 1913, Lipschütz¹ proposed a classification of nonvenereal AGU in 3 categories. Two categories fall in the pattern of aphthosis, being idiopathic or secondary to Behçet disease or Crohn disease. The third type of AGU was described as sudden, gangrenous, self-limiting, with nonrelapsing ulcers, occurring mainly in non–sexually active girls and young women and associated with systemic signs evocative of infection. During the 20th century, nonherpetic AGU has been associated with several infectious diseases, including salmonellosis^{1 - 3} and infectious mononucleosis.⁴

However, the etiological diagnosis of AGU remains a difficult one, partly because of the large array of potential precipitating factors,⁵ many of which are probably underrecognized by clinicians. Moreover, when considering the literature on this issue since the seminal study by Lipschütz,¹ we suspected that in fact several distinct nosological entities may have been reported under the entry Lipschütz's ulcer, thus further hampering the understanding of this syndrome. We therefore suggest avoiding this appellation, and propose herein major and minor criteria to designate a peculiar clinical situation that requires prompt diagnosis and management, often in an emergency context: the first flare of AGU in an immunocompetent patient without any context of sexually transmitted disease (STD) or aphthosis. Our aim was to describe the clinical and microbiological features of the first flare of AGU in a series of these patients.

We designed a descriptive study of 13 consecutive immunocompetent patients with a first flare of non–sexually related AGU. Patients referred to the departments of dermatology of 3 university hospitals in Paris from January 1, 1996, through December 31, 2006, were prospectively included if they presented with 5 major and 1 of 2 minor criteria. Major criteria included (1) presenting with a first flare of AGU; (2) being younger than 20 years; (3) absence of sexual contact during the past 3 months; (4) absence of immunodeficiency; and (5) acute course of the genital ulcer (ie, abrupt beginning and healing without scarring within 6 weeks). Minor criteria were related to the symptomatology of the genital lesions, which may present in the following 2 possible patterns: (1) 1 or several deep, well-delimited, painful ulcers, with a necrotic and/or fibrinous center or (2) a bilateral “kissing pattern” (a mirrorlike vulvar distribution).

The exclusion criteria were (1) a history of genital aphthosis or STD; (2) clinical or microbiological evidence of genital herpes or another STD; and (3) immunodeficiency. Given its high incidence in the general population, a history of oral aphthosis was not an exclusion criterion. Because this was a descriptive, noninterventional study, the project did not require approval by an institutional review board.

We collected the following data: age, sex, sexual history, complete clinical examination, symptomatology of the genital ulcers, time to healing of the genital lesions, blood cell count, biochemistry (measurement of liver enzyme levels and kidney function), and results of a search for infectious agents including human immunodeficiency virus, cytomegalovirus (CMV), Epstein-Barr virus (EBV), syphilis, and toxoplasmosis. The presence of HSV was investigated by means of repeated serologic examinations and/or local swabbing (for culture and/or polymerase chain reaction). In some cases, other laboratory workups were implemented, including genital mucosal biopsy for histopathological examination (in 5 patients).

One year after the flare of AGU, a standardized form was sent to the patients to assess occurrence of relapses of AGU or other medical problems.

We conducted an extensive and systematic MEDLINE search using the key words acute and genital or vulvar and ulcers or ulcus vulvae acutum or Lipschütz. We further searched all references in the retrieved articles. Thus, we collected a total of 32 relevant articles on the topic.

In March 2000, a previously healthy 19-year-old woman presented with painful AGU, odynophagia for 10 days, and intense fatigue for 2 days. She denied previous sexual contact at repeated interviews. On physical examination, her temperature was 40°C and she had pharyngeal edema, lymphadenopathies, and vulvar ulcers (Figure, A, and Table 1). Her hymen was intact. At day 10, she developed a grayish white membranous tonsillitis, with enlarged and swollen cervical lymph nodes. Pharyngeal culture results were negative for group A streptococci. Cefpodoxime proxetil therapy was then started with substitution of josamycin at day 13. At day 25, she developed a rash of the 4 extremities, consisting of scattered erythematous small papules that resolved spontaneously within a few days. Fatigue persisted until day 37. At day 65, she presented with several painful aphthoid ulcerations of the tongue that healed within 15 days.

Serologic test results for CMV, human immunodeficiency viruses 1 and 2, parvovirus B19, hepatitis B and C viruses, syphilis (Treponema pallidum hemagglutination [TPHA] and VDRL tests), toxoplasmosis, and p24 antigenemia were negative at different sampling times.

At day 7 of the ulcers, findings were positive for IgM antibodies to the EBV capsid antigen (VCA) and negative for IgG antibodies. At day 15, findings for IgM and IgG anti-VCA antibodies were strongly positive, whereas those for anti-EBV nuclear antigen (EBNA) were negative, confirming primary infection with EBV. Other biological variables are summarized in Table 1.

The Tzanck test showed no cytopathogenic effect. Histopathological examination of a mucosal biopsy specimen showed a necrotic epithelium and a polymorphic dermal infiltrate that consisted of neutrophils and mononuclear cells expressing preferentially the CD8 antigen (Figure, B). Results of in situ hybridization revealed EBV transcripts (EBV-encoded small RNA [EBER] and BamHI H left frame 1 [BHLF-1 RNA]) (Figure, C), whereas immunohistochemistry revealed EBNA-2 and BamHI Z EBV replication activator (ZEBRA) proteins (Figure, D). Quantification of EBV was performed at sequential intervals in the serum samples and in the vulvar lesion.

At day 6, the serum load of EBV was in the range of the serum loads commonly found in patients with EBV primary infection (373 copies/mL), whereas the EBV load was higher in the mucosal swab (1737 copies/mL). The serum EBV load remained high at day 15 (336 copies/mL) and decreased to less than 3 copies/mL at day 24, without any subsequent flare of serum EBV load during the 4-month virological follow-up.

The CD8 cells specific for EBV were quantified at different times, using an interferon γ enzyme-linked immunospot assay (IFN-ELISPOT kit; Diaclone Research, Besançon, France). At day 21, a weak response against latent (EBNA-3A, EBNA-3C, and latent membrane protein 2 [LMP-2]) and lytic (EBV immediate-early gene product BMLF-1) antigens was detected. At day 65, the response against BMLF-1 displayed a 10-fold increase (257 spot-forming cells per 10⁶ peripheral blood mononuclear cells). At 6 months, the BMLF-1 response vanished and was replaced by a strong response against the LMP antigen (217 spot-forming cells per 10⁶ peripheral blood mononuclear cells).

Six months later, the patient remained free of symptoms, findings for anti-EBNA antibodies were positive at high titers, and blood lymphocyte levels were within the reference range. One year later, she had not experienced a relapse.

From 1996 to 2006, 13 patients (including the case described in the preceding section) were included in a clinical and virological study (Table 1). Unless otherwise indicated, data are expressed as mean (SD).

All of the patients were female. Mean age was 16.6 (2.8) years (median age, 18 years; range, 11-19 years). Eleven of 12 patients (92%) denied previous sexual contact. Patient 12 had had previous sexual intercourse, although not during the 3 months before the AGU.

Seven of the 13 patients (54%) had prodromic tonsillitis, which occurred with a mean delay of 8.6 (6.0) days before the AGU onset (median delay, 7 days; range, 3-18 days). A fever or flulike symptoms preceded AGU in 10 of 13 patients (77%), with a mean delay of 3.8 (3.1) days before the AGU onset (median delay, 3 days; range, 0-10 days).

The genital ulcers were bilateral in 10 of 13 patients (77%), with a peculiar kissing pattern in 9 of 10 (90%). Nine of 13 patients had a medical history of transient oral erosions, diagnosed as aphthae.

The AGU was associated with an EBV primary infection in 4 of 13 patients (31%). Biological mononucleosis syndrome (hyperbasophilic atypical lymphocytes) was present in 3 of the 4 patients with EBV primary infection (75%) vs 1 of the 7 patients without EBV primary infection (14%) (P = .09). No other infectious agents were detected in any of the 13 patients; all patients who underwent testing had negative serologic findings for HIV (12 patients) and syphilis (6). All patients who underwent testing had repeated negative test results or a formerly acquired immunity profile for HSV (10 patients), CMV (12), and toxoplasmosis (6) antibodies. The kissing pattern of the ulcerations was present in all 4 of the patients with EBV primary infection (100%) vs 5 of the 9 patients without EBV primary infection (56%).

A medical history of oral erosions diagnosed as aphthae was present in 1 of the 4 patients with EBV primary infection (25%) vs all 8 of those without EBV primary infection (100%) (P = .02).

Cytolytic hepatitis was present in 3 of the 4 patients with EBV primary infection (75%) vs 3 of the 7 without EBV primary infection (43%).

Results of the histopathological examination were not helpful for the diagnosis in any of the 5 patients who underwent mucosal biopsy and usually showed a nonspecific lymphocytic inflammatory dermal infiltrate.

Six patients were treated with valacyclovir hydrochloride (46%) and 4 with a β-lactamine (31%). The mean delay before healing was 16.8 (7.4) days (median delay, 15 days; range, 8-35 days), without a significant difference according to the final diagnosis (EBV primary infection vs other AGU).

Three patients (23%) experienced 1 relapse of AGU during the following year. None of them was diagnosed as having EBV primary infection during the first flare of AGU. One of them was later diagnosed as having Behçet disease.

The first flare of a nonherpetic AGU in an immunocompetent patient without an STD is a difficult etiological problem that is often diagnosed in excess as genital herpes or aphthosis or even sexual abuse in young patients. We presented herein a series of 13 consecutive cases of AGU in female patients younger than 20 years, and we showed that one-third of these AGUs were associated with EBV primary infection.

Unlike herpes and aphthosis, AGUs related to EBV primary infection are typically associated with the absence of further relapse of genital ulcers. This clinical syndrome should lead to an oriented microbiological workup in search of previously reported causative infectious agents, including EBV, CMV, salmonella, and toxoplasmosis.^{2 - 8} Our study confirms that the most frequently reported infectious agent associated with a first flare of nonherpetic AGU is EBV.

We retrieved 18 articles that reported 24 cases of a first flare of AGU associated with EBV primary infection,^{4 - 5 ,9 - 24} including 15 cases in girls without previous sexual contact (Table 2). All cases were confirmed by serologic test results consistent with EBV primary infection. In 4 cases, local workup identified EBV on the genital lesions.^{4 ,21 ,23}

In 2006, Huppert et al¹⁵ reported a series of 20 cases of presumed non–sexually transmitted AGU in female patients aged 10 to 19 years. Systemic signs such as fever, malaise, and headache were reported by 19 of 20 patients. Two of them had IgM antibodies to VCA (10%), and 2 had possible evidence of acute CMV infection. Summary rather than detailed individual data were provided; thus, peculiar findings in the 2 patients with associated EBV primary infection were not availabel. Other laboratory findings were nonspecific. The median duration of pain was 10 days, and 15 patients (75%) healed by 21 days. Huppert et al¹⁵ considered that most of these cases of AGU were possibly genuine genital aphthae. However, the inclusion criterion in their series was AGU with negative HSV test results and, therefore, that these cases might be different from those in our series, which may explain the lower rate of EBV-related AGU reported in the series by Huppert et al.¹⁵ For example, 3 of their patients had a history of AGU,¹⁵ whereas this feature was an exclusion criterion in our series.

In the highlighted case from our series of patients, we demonstrated the presence of EBV in mononuclear cells infiltrating the vulva. We showed the expression of both RNA (EBER and BHLF-1) and proteins (EBNA-2 and ZEBRA), suggesting that EBV may replicate in the mucosa, or at least that a part of the lytic cycle could occur. However, the proportion of EBV-positive cells was low, suggesting that the ulceration results more from an immune reaction associated with EBV than from a direct cytopathogenic effect of the virus itself. Moreover, in our series, as in others,¹⁵ most of the pubertal non–sexually active patients with AGU had no evidence of associated acute systemic infection. Therefore, it might be hypothesized that AGU in these patients result from a reactional nonspecific inflammatory process. This may explain why AGU have been described with a large spectrum of acute infection, such as CMV, influenza, salmonella, and toxoplasmosis.^{2 - 3 ,7 - 8 ,25 - 26}

In addition, some authors have suggested that these peculiar forms of AGU in non–sexually active adolescents are a clinical form of primary genital aphthosis.¹⁵

In 24 previously published cases of EBV primary infection associated with AGU,^{5 ,9 ,11 ,13 - 14 ,16 ,18 - 24} mean age was 15.1 years, which is similar to our findings, and 23 of 24 patients (96%) were female. In our series, 92% of the patients denied previous sexual contact; in the literature, 17 of 24 patients (71%) had never had sexual contact; 4 (17%) had had sexual intercourse^{4 - 5 ,17} ; and 3 (13%) had had only orogenital (2 patients)^{10 ,18} or digital-genital (1)²³ contact. On clinical examination, 8 of 22 previously published cases of AGU (36%) had a kissing pattern of ulcers, 13 (59%) had sore throat, 16 (73%) had lymphadenopathy, and all had systemic signs, the most common being fever (20 of 22 [91%]) and fatigue (14 of 22 [64%]). In our study, the mean duration of genital lesions was 17 days, which is comparable to the duration of 19 days reported in the literature. This healing delay is comparable to that of 12 to 16 days reported for genital herpes primary infection.^{27 - 29} In 9 of the 21 reported cases (43%), acyclovir was introduced before the correct diagnosis of EBV primary infection was made (5 of 13 [38%] in our series); this reflects that genital herpes is the most commonly suspected differential diagnosis. Indeed, genital herpes should always be ruled out first in a patient with AGU. However, the clinical presentation of genital herpes is different, with smaller and superficial erosions often coalescing in a polycyclic pattern, and this differential diagnosis can be confirmed by genital polymerase chain reaction or culture.

A much-debated question is how the EBV reaches the genital mucosa in AGU. Four modes of transfer of EBV to the genital mucosa may be discussed. First, potential transmission via genital-genital contact has been hypothesized,³⁰ given that EBV shedding has been identified in male semen³¹ and in the uterine cervix.³² Second, transmission via oral-genital contact is suggested by previous reports of long-term salivary shedding of EBV after infectious mononucleosis and in healthy adults.^{33 - 34} These 2 hypotheses seem unlikely because 92% of the patients in our series had had no previous sexual contact. Third, shedding of EBV in the urine has been shown during and after infectious mononucleosis³⁵ as well as in healthy adults.³⁶ Fourth, a hematogenous spread of EBV-infected lymphocytes^{14 ,16 ,23} or Langerhans cell precursors³⁷ has been reported. Furthermore, the clinical sequence of systemic signs (fever, asthenia, and prodromic tonsillitis) and genital signs is evocative of a hematogenous transfer of EBV to the genital mucosa. Overall, even if there is currently no solid evidence to definitely promote one of these hypotheses, the urinary and hematogenous routes are the most likely ones and may coexist. Once having reached the genital mucosa, the EBV may provoke a cytotoxic immune response, suggested by the CD8⁺ granzyme-B⁺ T-cell inflammatory infiltrate retrieved in some of our EBV-infected patients.

The heterophile antibody (monospot) test is widely used, but its sensitivity and specificity are weak. The monospot test should systematically be replaced by VCA-IgM and VCA-IgG tests, which are sensitive and specific and emerge early during the course of EBV primary infection.³⁸ Furthermore, in 3 of our 4 patients with EBV primary infection, findings of the local workup for EBV were negative, demonstrating the low sensitivity of this diagnostic method in this clinical setting. We conclude that the diagnosis of AGU related to EBV primary infection should be based essentially on VCA-IgM results. The histopathological examination of the mucosal biopsy specimen performed in 5 of our 13 patients showed only nonspecific inflammatory dermal infiltrates. We therefore consider this invasive diagnostic method as probably not relevant in most cases of AGU.

Primary infection with EBV is a largely underrecognized cause of AGU, leading patients to unnecessary investigations, treatment, and anxiety. Patients should be informed that this rare form of acute mononucleosis is self-limiting and seldom relapses. Supportive care primarily aimed at pain relief is indicated in all cases. Systemic corticosteroids have been administered by some authors for peculiarly painful or long-lasting AGU,⁹ but evidence-based data on the efficacy-safety balance are lacking. Despite showing in vitro activity, acyclovir demonstrated no clinical benefit in a meta-analysis of 339 patients with EBV primary infection.³⁹

Our series of consecutive patients is the first, to our knowledge, to permit an estimation of the rate of EBV primary infection among patients with AGU in the absence of STD, although the small size of our series does not allow for a definite estimation. The analysis of the present series and of the previous reports^{5 ,9 ,11 ,13 - 14 ,16 ,18 - 24} yields several messages for the clinician. (1) Acute mononucleosis should always be in a clinician's differential diagnosis of AGU, especially when there is no context of sexual transmission and no history of genital aphthosis. (2) Failure to consider causes other than genital aphthosis or herpes may expose patients and their family to unnecessary investigations, treatments, and stress. (3) The symptomatology of AGU is essentially nonspecific; therefore, a systematic initial workup is required and should be clinically oriented.⁵ (4) The physiopathology and etiology of nonherpetic AGU still constitute broadly unexplored research fields in which further prospective clinical and microbiological studies are needed. Further infectious causes of this little-known syndrome may be recognized in the near future.

Correspondence: Nicolas Dupin, MD, PhD, Service de Dermatologie, Pavillon Tarnier, CHU Cochin, 89 rue d’Assas, 75006 Paris, France (nicolas.dupin@cch.aphp.fr).

Accepted for Publication: May 16, 2008.

Author Contributions: All the authors had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: Farhi, Wendling, Pelisse, and Dupin. Acquisition of data: Wendling, Molinari, Raynal, Carcelain, Morand, Avril, Francès, Rozenberg, Pelisse, and Dupin. Analysis and interpretation of data: Farhi. Drafting of the manuscript: Farhi and Dupin. Critical revision of the manuscript for important intellectual content: Farhi, Wendling, Molinari, Raynal, Carcelain, Morand, Avril, Francès, Rozenberg, and Pelisse. Statistical analysis: Farhi. Study supervision: Farhi, Morand, and Dupin.

Financial Disclosure: None reported.

Funding/Support: This work was partly supported by l’Association de Recherche en Virologie et Dermatolgie (ARVD).

Additional Contributions: Agnès Carlotti, MD, and Philippe Gerhardt, MD, provided clinical contribution. Micheline Tulliez, MD, provided immunohistochemistry experiments.