Desquamative inflammatory vaginitis is an uncommon but severe form of purulent vaginitis that occurs predominantly in perimenopausal women, causing discharge and severe dyspareunia.1–3 Of unknown etiology and frequently undiagnosed, the syndrome is characterized by vaginal inflammation with a vaginal rash and purulent discharge. Microscopy reveals a marked increase in inflammatory cells, predominantly polymorphonuclear leukocytes and parabasal epithelial cells, together with abnormal vaginal flora.1–4 The term desquamative inflammatory vaginitis first was introduced in 1965 when Gray and Barnes described six women with a “reddened” vagina and “numerous pus cells … with oval and round parabasal cell.”3 Three years later, Gardner published the first case series including eight patients with similar clinical presentation in which no response to antimicrobial treatment was seen.2 After a gap of three decades, Sobel described his experience with 51 women with desquamative inflammatory vaginitis including successful treatment with topical 2% clindamycin, suggesting a bacterial etiology for this entity.1 Although 29% of patients initially responding relapsed when therapy was discontinued, results with topical clindamycin therapy were encouraging with short-term follow-up of 30–60 days only.1
During the past 15 years, we have increased our clinical experience managing women with desquamative inflammatory vaginitis including the use of alternative therapy. In this report, we describe patient outcomes and treatment at follow-up in a case series of 98 women diagnosed with this syndrome.
MATERIALS AND METHODS
We conducted a chart review of patients evaluated by a single clinician (J.D.S.) in a referral university-based vaginitis clinic diagnosed with desquamative inflammatory vaginitis in the years 1996–2007. Institutional Review Board approval from Wayne State University was obtained. The case definition of desquamative inflammatory vaginitis required all the following criteria: 1) symptomatic women with at least one of the following complaints: discharge, dyspareunia, pruritus, burning, or irritation; 2) vaginal inflammation, typically a spotted vaginal rash, focal or linear erosions; 3) vaginal pH higher than 4.5; and 4) saline microscopy showing increase in parabasal cells and inflammatory cells, defined as ratio of inflammatory cells to epithelial cells more than 1:1. Trichomoniasis was excluded in all cases, and cutaneous, oral, or vulvovaginal lesions suggestive of lichen planus excluded the cases being defined as desquamative inflammatory vaginitis. All patients reported were seen by a single clinician (J.D.S.), minimizing interobserver variability.
Demographic characteristics were abstracted from the medical record. A comprehensive medical history including sexual behavior, sexually transmitted diseases, and menopausal status was obtained at the intake visit in all women seeking care at the vaginitis clinic. Initial induction vaginal therapy consisted of either 4 to 5 g of 2% clindamycin or 10% hydrocortisone cream daily for 4 to 6 weeks. Symptoms, vulvovaginal signs, pH, wet mount, yeast culture results, and treatment were recorded for each visit. After discontinuation of therapy for at least 4 weeks, patients were defined as in remission at any follow-up visit if symptoms and signs were eliminated and vaginal inflammatory and parabasal cells were eradicated. We categorized as “controlled” those asymptomatic or minimally symptomatic, microscopy-normal patients in whom lower-dose maintenance topical therapy was continued because of concerns about potential for relapse. These included patients who relapsed after treatment cessation and patients in whom therapy had never been discontinued. Patients with residual persistent symptoms and clinical findings were categorized as partially controlled. All partially controlled patients continued on intensified treatment that included switch of 2% clindamycin and 10% hydrocortisone cream, use of 15% hydrocortisone cream, and a combined formulation including hydrocortisone and clindamycin. Patient outcomes and treatment at the 1-, 2-, and 4-year follow-up end points from initial visit were analyzed as the outcomes of interest in cross-sectional analyses. For longitudinal analyses, we used a follow-up period of 30 months total. Cure was defined as complete resolution of all signs, symptoms, and laboratory evidence of desquamative inflammatory vaginitis for at least 3 months off therapy.
Data were analyzed using SPSS 18.0 for Windows. For our case series, we estimated the proportion of women with particular demographic and clinical characteristics and the results of treatment as measured at the 2- and 4-year follow-up end points. Although recognizing the limitations of a case series in comparison with a randomized controlled trial, we nevertheless compared outcomes of treatment for women by characteristics of interest using the χ2 test. We used the Fisher exact test in all cases given the small numbers in our sample and the respective cells. The exact test was also used for proportions and their corresponding confidence intervals (CIs), a more conservative estimation.
We explored outcomes using survival analysis methods (in the SAS statistical program) that take account of the longitudinal nature of our cohort. We censored data at 30 months because the majority of women were monitored at least that long.5 The outcome was defined as cure or complete control, as defined above. We examined Kaplan-Meier survival plots to compare the probability of achieving either cure or complete control for fixed factors (eg, postmenopausal status).
During the 12-year period 1996–2007, 3,000 new patients were evaluated in the vaginitis clinic. Of the 180 charts initially classified as desquamative inflammatory vaginitis in a clinical database, 171 (95%) were available for review; however, 41 did not meet the strict case definition resulting in 130 women diagnosed with desquamative inflammatory vaginitis. Excluding charts not available for review, the estimated prevalence of desquamative inflammatory vaginitis in the vaginitis clinic was 4.3% (130 of 2,991; 95% CI 3.6–5.1%). Thirty-two desquamative inflammatory vaginitis patients were excluded from further analysis, including five with lichen sclerosus, six with suspected desquamative inflammatory vaginitis in whom erosive lichen planus could not be excluded, six who did not initially meet full case definition but fulfilled it subsequently, and 15 who did not have baseline data needed for analysis. Thus, 98 patients meeting strict case definition were analyzed.
Table 1 summarizes demographic characteristics and medical history at baseline. Ninety-seven of the 98 patients were white. Forty-eight women (49.5%, 95% CI 38.7–59.3%) were menopausal, of whom the majority (77%, 95% CI 62.7–88.9%) were already receiving systemic hormone therapy at their baseline visit.
All patients were symptomatic; discharge, usually purulent, was the primary complaint (88.9%, 95% CI 80.8–94.3%), with dyspareunia in 86.1% (95% CI 76.9–92.6%) of the 86 sexually active patients. Vaginal inflammation was evident initially in all patients including a typically spotted, often ecchymotic vaginal rash in 69 women (70.4%, 95% CI 60.3–79.2%), and the remaining 19 women experienced diffuse or focal vaginal erythema or linear erosions. A purulent discharge was apparent in 68 patients (69.5%, 95% CI 59.3–78.3%) on physical examination. Seventy-one patients had some vestibular abnormality at the initial visit including focal or diffuse erythema with variable thinning (n=63) or linear erosions and ecchymotic rash (72%, n=8). By case definition, all patients had an elevated vaginal pH and an increase in inflammatory and parabasal cells. Nearly all of the women (n=95, 96.9%, 95%, CI 91.3–99.4%) had abnormal bacterial flora, defined as nondominant Lactobacillus morphotype on saline microscopy. None of the women in the case series had a positive whiff test or clue cells present.
Initial treatment with topical 2% clindamycin daily was prescribed for 53 patients (54.1%, 95% CI 43.7–64.2%) and 45 women (45.9%; 95% CI 35.8–56.3%) were prescribed 10% topical hydrocortisone 3–5 g intravaginally once daily (n=39) or a cortisone acetate suppository 25 mg twice daily (n=6). At the first follow-up visit (median 3 weeks, range 1–19 weeks), 84 patients (85.7%, 95% CI 77.2–92%) were dramatically improved or asymptomatic, with the remaining 14 partially controlled. In these 14 women, changes were made in treatment, varying dosage frequency, medication type, or a combination of the three. There were no complete failures, with all women experiencing some improvement. Among the 84 women whose clinical manifestations resolved, treatment was discontinued in 53 (63.1%, 95% CI 51.9–73.4%) after a median of 8 weeks. Seventeen of these 53 women (32.1%, 95% CI 19.9–46.3%) relapsed within 6 weeks, and 23 (43.4%, 95% CI 29.8–57.7%) relapsed within 26 weeks off therapy. Treatment was repeated for all of these women. Relapse subsequent to 26 weeks was uncommon.
At 1 year (median 13.5, range 10.5–17.9 months), cure had been achieved in 25 women (26%); more than half of the women (n=57, 58%) were controlled but had continued on or required maintenance therapy. In 24 of 57 controlled patients (42%), treatment cessation was attempted but recurrence of symptomatology or microscopy abnormalities led to reinstitution of therapy. Fifteen women (16%), although improved, were only partially controlled at 1-year follow-up, with frequent attempts to intensify local anti-inflammatory vaginal and vestibular therapy (0.05% clobetasol, 0.01% tacrolimus, or 0.5 mg/g estriol cream).
Forty-six of the original 98 women (47%, 95% CI 36.8–56.3%) were still being monitored at 2 years after intake (mean 25.1, range 19–30 months). The cumulative cure rate at the 2-year end point was approximately 34.7% (n=34 women; 95% CI 25.4–45%) under the very conservative assumption that those women no longer monitored at 2 years all had relapsed. Under the opposite assumption that all women no longer monitored remained cured, the cumulative cure rate would be 87.8% (86 of 98, 95% CI 79.6–93.5%). Twenty-five women (25.8%) were still being monitored after the 4-year end point. Seven women were cured (having discontinued all treatment) and 12 were controlled on low-dose maintenance regimens. The remaining 8 of the 25 women monitored beyond 4 years were still only partially controlled (9.4%, 95% CI12.1–28.6%).
There were no statistically significant differences in response to treatment by sociodemographic characteristics in either the cross-sectional analyses at varying time points or the Kaplan-Meier survival curves (not shown). Table 2 presents the results of analyses seeking to identify the treatment factors associated with prognosis. A favorable early response to therapy (mean 9.5 weeks) was associated with favorable clinical prognosis in the subsequent 20 weeks (P=.01). Because of inconsistent criteria used to select anti-inflammatory therapy, comparative treatment efficacy evaluation was not feasible.
A subgroup of patients emerged over time with incomplete response to vaginal therapy with persistent symptoms, predominantly dyspareunia. An almost constant finding in this more recalcitrant group was the presence of persistent signs of vulvar vestibular syndrome, distinguishable from erosive lichen planus. This subgroup with vestibular inflammation was more likely to have a poor long-term prognosis (P=.03; Table 2) with more than twice the proportion categorized as partially controlled in this subgroup compared with other patients in our cohort (26.9% compared with 12.9%). Related to this subgroup risk, continuation or repeated therapy targeting the vestibule at a third follow-up visit (mean 19.3 weeks) was also associated with poorer long-term prognosis (P=.007, Table 2). None of the women who required therapy to the vestibule at the third visit were categorized as cured long term and the proportion categorized as only partially controlled was nearly three times higher compared with other women (33.3% compared with 12.7%).
The foremost adverse outcome was vulvovaginal candidiasis. Four women had a positive vaginal yeast culture when evaluated at baseline. Seventeen women with a history of vulvovaginal candidiasis received prophylactic fluconazole treatment (150 mg/wk) because both clindamycin and steroids are known to increase the risk for vulvovaginal candidiasis. Of the remaining 77 women, 31 (40.3%; 95% CI 29.2–52.1%) developed positive vaginal yeast cultures of whom 19 (61.3%, 95% CI 42.2–78.2%) were symptomatic and received antimycotic therapy.
Desquamative inflammatory vaginitis is a chronic inflammatory process in that women were symptomatic for a median of 15.5 months before diagnosis and the majority required therapy for more than 1 year. Moreover, once acute treatment was stopped in women experiencing clinical and laboratory resolution, 32% relapsed within 6 weeks.
Initial clinical improvement in response to therapy is almost universal to the extent that lack of improvement justifies reconsideration of diagnosis. However, only one fourth was cured by short-term therapy. Recognizing desquamative inflammatory vaginitis as chronic and not an acute process that frequently requires long-term therapy is fundamental. Nevertheless, although women frequently required therapy when seen at 2- and 4-year follow-up, most women were asymptomatic, indicating that the condition could be managed effectively.
Treatment for desquamative inflammatory vaginitis advocated in the early 1990s was that of daily topical clindamycin, and although duration of therapy was not determined, it was apparent that response was rapid but that relapse was common.1 At the time the beneficial effect of clindamycin was attributed to its antibacterial effect directed at an undetermined anaerobic pathogen.1,6 Lack of response of desquamative inflammatory vaginitis to metronidazole suggested a Gram-positive coccal species. With frequent relapses of the syndrome, alternative measures were required. Low-dose topical corticosteroids had been used for desquamative inflammatory vaginitis by Gardner with modest benefit and high relapse.2 Accordingly, faced with relapsing, occasionally refractory desquamative inflammatory vaginitis, we introduced a higher-dose steroid in the form of 10% vaginal hydrocortisone cream with dramatic salutary effect. Given the efficacy of 10% hydrocortisone in women failing clindamycin, bias has inevitably influenced drug selection priority. In the present study it was not possible to compare short- and long-term efficacy of clindamycin and hydrocortisone. Moreover, the favorable response to corticosteroids indicates a likely noninfectious etiology of this chronic syndrome. Of interest, the anti-inflammatory effects of clindamycin have been recognized, as have the anti-inflammatory properties of macrolides with inhibition of synthesis of proinflammatory cytokines IL-1, IL-6, IL-8, and TNF-α.4,6–10 Clindamycin is a lincosamide, which is a class of antibiotics capable of binding to the 23s segment of the 50s subunit of bacterial ribosomal RNA and inhibits protein synthesis.11 It has been shown that macrolides have overlapping binding sites with clindamycin to the 23s segment, suggesting a similar immunomodulatory mechanism for clindamycin, macrolides, and cortisone.11,12 Originally desquamative inflammatory vaginitis, as the name suggests, was considered to exclusively involve the vagina.1–3 Vestibular involvement if present was considered to be secondary to the purulent discharge.1 Consequently in the present study, patients presenting initially with vestibular findings rarely received specific topical treatment for the vestibule. With experience it became apparent that controlling vaginal inflammation and eliminating discharge were insufficient to heal the vestibule, and anti-inflammatory treatment for the vestibule was also necessary. The need to treat vestibular disease at the third follow-up visit was significantly associated with a poorer long-term prognosis.
The etiology of desquamative inflammatory vaginitis is unknown. However, the unique white predominance of patients in a clinic where 50% of individuals are African American, and favorable response to anti-inflammatory agents suggest that patients with desquamative inflammatory vaginitis have a predisposing genetic factor, allowing the immune system to react against elements in the vaginal mucosa resulting in inflammatory vaginitis. The trigger responsible for initiating the response of the immune system is not yet identified. It was hypothesized that lack of vaginal estrogen had a major role in initiating the cascade considering that a significant number of patients were diagnosed with estrogen deficiency, including postpartum nursing and perimenopausal and postmenopausal states.1 Regrettably, no measurements of systemic estrogen levels are available.
This uncontrolled review has several limitations including the loss of 48% of women to follow-up at 2 years. However, our conclusions at 2- and 4-year follow-up are minimal, and given the referral basis of this unique clinic, we believe attrition reflects cure rather than failure and surrender by these desperate women. Desquamative inflammatory vaginitis is a clinical syndrome without a definitive laboratory test. Nevertheless with a strict case definition, specificity is still possible. Other causes of purulent vaginitis were unlikely to have been included. Of patients with erosive lichen planus, there is a small subgroup of less than 10% that presents with vaginal disease only.13 These patients could clinically overlap with desquamative inflammatory vaginitis patients to an extent that some clinicians believe that desquamative inflammatory vaginitis is a variant form of erosive lichen planus.14,15 None of the patients in the current study had oral or cutaneous disease; furthermore, the clinical features of vaginitis were not typical of erosive lichen planus. A multicenter prospective randomized study is needed, including systemic estrogen measurements and comparison between clindamycin and steroid treatment.
In conclusion, we report the course and clinical response to treatment in women with desquamative inflammatory vaginitis. This study provides several new insights. Desquamative inflammatory vaginitis is an infrequent but not rare chronic inflammatory process that frequently involves the vestibule in addition to the vagina and responds well to long-term anti-inflammatory therapy. Physicians should be familiar with the clinical presentation and consider the diagnosis of desquamative inflammatory vaginitis in women presenting with purulent vaginitis, particularly in white women.