The Bethesda system for interpretation of cervical cytology, introduced in 1988, was intended to standardize and simplify evaluation of an abnormal Papanicolaou smear. Although this seems to be the case for smears that read as squamous intraepithelial lesions (SIL), its utility in diagnosis of atypical squamous cells of undetermined significance (ASCUS) and description of inflammatory changes are uncertain. The proper management of women with ASCUS on a screening Papanicolaou smear remains controversial and results in substantial expenditure of health care dollars. Although colposcopic examination offers the most definitive follow-up evaluation of the abnormal Papanicolaou smear, the cost of using this test for all cytologic abnormalities is prohibitive. Although opinions vary widely, the most common guidelines for the management of women with ASCUS on Papanicolaou smear results include repeated cytologic examination in 3 to 6 months, with referral for colposcopy if results demonstrate persistent ASCUS or SIL.1,2
Clinical and laboratory data have linked overgrowth of vaginal anaerobic bacteria with cervical dysplasia, yet short-term alterations in the vaginal flora with local antibiotics has not been shown to reverse abnormal cytologic patterns.3–7 Several retrospective reviews presented in abstract form only have suggested that vaginal antianaerobic therapy in the absence of clinical bacterial vaginosis may increase the rate of regression to normal for women with a first-time finding of ASCUS on Papanicolaou smear. Although published or prospective data to support the use of empiric vaginal antimicrobials as first-line management of ASCUS are lacking, this has become a common clinical practice. In a randomized trial done before development of the Bethesda system, empiric treatment of atypical findings on Papanicolaou smear with triple sulfa cream was not found to be more efficacious than observation in clearing atypical changes on repeated cytologic examination.8 A similar study of vaginally applied metronidazole has not yet been reported. We therefore conducted a prospective, double-masked, randomized trial of vaginal metronidazole versus carrier-gel placebo in women with ASCUS on Papanicolaou smear. We sought to determine whether empiric vaginal metronidazole treatment of ASCUS on Papanicolaou smear increases the number of cases that regress to normal on repeated cytologic examination.
We assumed a rate regression to normal of 50% in untreated women. Power analysis showed that 45 women per study arm would allow the study to determine a 50% increase in normal results on cytologic examination in metronidazole-treated women, with a significance of α = .05 and a power of 80% in one-sided analysis. Because of the historical problem of compliance among women with abnormal Papanicolaou smears, we adjusted the study to allow for a 20% drop-out rate.9,10
Masked tubes of active vaginal metronidazole or carriergel placebo were packaged in boxes according to computer-generated random numbers from Curatek/3M Pharmaceuticals (Buffalo Grove, IL) before enrollment. Each box was labeled with a participant number and a hidden identification of the content of the tubes that could be accessed in case of problems, such as hypersensitivity reaction. The randomization code was maintained by Curatek/3M until collection of study data was completed.
Women were referred from the general gynecology clinics and the faculty practices of the Department of Obstetrics and Gynecology, University of Illinois at Chicago. This group accounts for approximately 30,000 clinic visits and more than 15,000 Papanicolaou smears per year. Sixty percent of the women seen are in the public assistance/Medicaid system, and another 12% are uninsured. On the basis of monthly statistics from the cytology laboratory, 9% to 11% of the Papanicolaou smears performed in this group are interpreted as ASCUS. Just over 50% of ASCUS results are seen in obstetric clinic smears and are not included in this study.
Table 1 shows inclusion and exclusion criteria. After the study was approved by the institutional review board, all primary care providers in the department were educated on the inclusion/exclusion criteria and were asked to refer potential participants. The authors interviewed all potential participants to determine eligibility and to obtain informed consent. After agreeing to participate, each candidate underwent standard speculum examination to collect a sample for potassium hydroxide testing and sodium chloride wet preparation. Women were excluded if this examination demonstrated active trichomonas vaginitis, fungal overgrowth, or asymptomatic bacterial vaginosis. Asymptomatic bacterial vaginosis was diagnosed by a positive result on potassium hydroxide “whiff” test and the presence of clue cells on sodium chloride wet preparation. Participants not excluded by this examination were assigned the next randomized tube in numeric order and were instructed on the use of the vaginal gel and applicator. Written and pictorial instructions were included with each tube of gel. Patients were given an appointment for 90 days after enrollment for repeated cytologic examination. Participants were contacted by phone when possible to evaluate gel use or address any problems or concerns.
At the follow-up visit, women were asked to describe how they used the gel and any problems they believed were related to the gel use. A speculum examination was performed to obtain samples for repeated cytologic examination. Papanicolaou smears were obtained by using a cytobrush from the endocervix and a wooden Ayers spatula from the exocervix/transformation zone. Both specimens were smeared onto one slide and fixed with spray fixative. Cytology requisitions did not include study participation in the history given to avoid bias in the interpretation. Although enrollment smears were evaluated by several different laboratories, all repeated smears were evaluated by the same cytology laboratory. Women with persistent ASCUS or another SIL diagnosis were referred for colposcopy. Women with normal results on repeated cytology were instructed to return to their primary care provider for cytologic examination every 6 months. After two normal smears, women were recommended to have Papanicolaou smear screening annually.
We used the Student t test to evaluate differences between continuous variables and the χ2 test to evaluate categorical variables; in each case, significance was set at P < .05. All statistical analyses were performed using BMDP statistical software (BMDP, Los Angeles, CA).11
During the 40-month study period, 197 women were referred for possible enrollment. Forty-eight women (24%) made appointments for initial evaluation but did not keep the appointment. Twenty-four (12%) women were eligible but declined enrollment after initial evaluation. These women were counseled to return to their primary care provider for repeated cytologic examination in 3 to 6 months. Nineteen women were evaluated and excluded for the reasons outlined in Table 2. The most common reason for exclusion was active vaginitis or cervicitis. Two women were excluded because they had already been treated with vaginal metronidazole although bacterial vaginosis was not documented.
One hundred and six tubes of gel were dispensed, 54 containing active metronidazole and 52 containing placebo. Age, gravidity, parity, and time between enrollment and repeated Papanicolaou smear did not differ between groups (Table 3). Enrollment Papanicolaou smears were subclassified as ASCUS not otherwise specified (n = 53), ASCUS favor reactive (n = 35), and ASCUS favor dysplasia (n = 18). The distribution of enrollment cytology subclasses was similar in each group (P = .8, χ2 test).
Eighty-four women (79%) returned for repeated cytologic examination (41 active gel recipients and 43 placebo recipients) at a mean of 6.5 months (range 3–36 months). Overall, 50 patients (60%) had normal cytology on repeated Papanicolaou smear and 34 (40%) had persistent abnormal findings. Twenty-two repeated Papanicolaou smears demonstrated persistent ASCUS (10 ASCUS not otherwise specified, 10 ASCUS favor reactive, and 2 ASCUS favor dysplasia), 11 demonstrated SIL (10 low-grade and 1 high-grade lesions), and 1 was unsatisfactory due to inflammation. The incidence of normal results on repeated cytologic examination did not significantly differ by subclassification of enrollment Papanicolaou smear: 56% for ASCUS not otherwise specified, 63% for ASCUS favor reactive, and 63% for ASCUS favor dysplasia (P = .8, χ2 test).
Regression to normal cytology was found in 22 (54%) of the 41 women who received active metronidazole compared with 28 of 43 placebo recipients (65%) (P = .28, χ2 test). The incidence of ASCUS (31% versus 25%) or SIL (15% versus 10%) on repeated cytologic examination was also similar between recipients of active drug and placebo recipients.
Nine of the 11 women with SIL on repeated Papanicolaou smear underwent colposcopy. The one woman with high-grade SIL on repeated cytologic examination was found to have cervical intraepithelial neoplasia (CIN) II on biopsy. Of the 10 women with low-grade SIL, 4 (40%) had CIN I, 2 (20%) had histologic changes consistent with human papillomavirus, and 2 (20%) had no pathologic abnormalities. Two women (20%) did not appear for colposcopy appointments.
Twenty-two women had ASCUS on repeated cytologic examination and were referred for colposcopy. Of these, 12 have had colposcopic evaluation with biopsy. Eleven of the 12 women had no histologic evidence of dysplasia, and 1 had an endocervical CIN II lesion on endocervical curettage that was not visible on colposcopy.
Since the introduction of the Bethesda System for Papanicolaou smear reporting, the optimal management of the ASCUS result has been controversial. Several studies have indicated an association between bacterial vaginosis and CIN. In 1975, Harington identified the presence of nitrosamines in vaginal discharge and postulated that these chemicals, which are the by-products of anaerobic vaginal floral, may be carcinogens in CIN and cervical cancer.3 This hypothesis was reinforced in 1984 by Pavic, who demonstrated the production of nitrosamines by the vaginal organisms found in bacterial vaginosis and trichomonas infection.4 Byrne et al noted that the most common cervical vaginal infection among women referred for colposcopy was bacterial vaginosis.5 Similarly, a series of 106 women with CIN and 79 controls evaluated by Guijon et al found a 10-fold increase in vaginal anaerobic bacteria in the CIN group. Bacterial vaginosis was diagnosed in 48% of women with CIN compared with 22% of the controls.6 In a retrospective review of 6150 Papanicolaou smears, Platz-Christensen et al found clue cells consistent with a diagnosis of bacterial vaginosis in 10% of cases. Overall, the relative risk of CIN was 8.0 in women with bacterial vaginosis. In women with bacterial vaginosis, the relative risk for CIN III or carcinoma in situ was 5.0 (95% confidence interval 4.1, 15.6).7
Although many data link clinically evident overgrowth of anaerobic bacteria and CIN, no published evidence indicates that short-term or subclinical alterations in the vaginal flora alter the rates of abnormal cytologic findings or CIN. It has been theorized that some cytologic changes that could result in a reading of ASCUS may be due to subclinical anaerobic overgrowth and that short-term treatment with antianaerobic agents before repeating cytologic examination may improve the rate of regression to normal, obviating the need for colposcopy. Two retrospective reviews presented at local district meetings of the American College of Obstetricians and Gynecologists suggested that vaginal anti-anaerobic therapy in the absence of clinical bacterial vaginosis may increase the rate of regression to normal by as much as 50% in women with a first-time finding of ASCUS on Papanicolaou smear. Although published or prospective data to substantiate the practice are lacking, empiric therapy with vaginal antimicrobial preparations is widely used as the first step in the management of ASCUS on Papanicolaou smear. In the only prospective trial to date, Reiter reported on 512 patients with atypical Papanicolaou smears who were randomized to receive vaginal triple sulfa cream or no therapy.8 Of the 494 women who completed the study, the rate of normal results on repeated cytologic examination did not differ between the treatment group (76.3%) and control group (74.0%). No similar data have been published on use of vaginal metronidazole as empiric treatment of ASCUS on Papanicolaou smear.
Our prospective data on the role of vaginal metronidazole indicate no benefit to its empiric use in the initial management of the patient with ASCUS on Papanicolaou smear. The rate of regression to normal was 8% higher in placebo recipient than in women who used active drug; this difference was not statistically significant. We designed the study to show improvement of 50% or more to be clinically significant at 80% power in women treated with active metronidazole gel. However, because data on only 43 placebo recipients and 41 active drug recipients were available, the power of the study to demonstrate the less than 10% difference between groups is only 50%. To reach a power of 80% with this narrow margin of difference, the study would have required 360 women per study arm, and we believe that the less than 10% difference would be of questionable clinical significance.
Experience with the new Bethesda classifications of atypical squamous cells, published in 2001, and increasing use of human papillomavirus testing as a secondary triage of findings of atypical squamous cells, are hoped to clarify the uncertainty surrounding the management of women with ASCUS on cytologic examination.12–14 At this time, we conclude that empiric use of antimicrobials, in particular vaginally applied metronidazole, before repeated cytologic examination does not improve the management of women with the diagnosis of ASCUS on Papanicolaou smear and that this practice should be discouraged.
1. Kurman RJ, Henson DE, Herbst AL, Noller KL, Schiffman MH. Interim guidelines for the management of abnormal cervical cytology. JAMA 1994;271:1866–70.
2. Kaufman RH. Atypical squamous cells of undetermined significance and low-grade squamous intraepithelial lesion: Diagnostic criteria and management. Am J Obstet Gynecol 1996;175:1120–8.
3. Harington JS. Epidemiology and aetiology of cancer of the uterine cervix. S Afr Med J 1975;49:443–5.
4. Pavic N. Is there a local production of nitrosamines by the vaginal microflora in anaerobic vaginosis/trichomoniasis? Med Hypotheses 1984;15:433–6.
5. Byrne MA, Turner MJ, Griffith M, Taylor-Robinson D, Soutter WP. Evidence that patients presenting with dys-karyotic cervical smears should be screened for genital-tract infections other than human papillomavirus infection. Eur J Obstet Gynecol Reprod Biol 1991;41:129–33.
6. Guijon F, Paraskevas M, Rand F, Heywood E, Brunham R, McNicol P. Vaginal microbial flora as a cofactor in the pathogenesis of uterine cervical intraepithelial neoplasia. Int J Gynecol Obstet 1992;37:185–91.
7. Platz-Christensen JJ, Sundstrom E, Larsson PG. Bacterial vaginosis and cervical intraepithelial neoplasia. Acta Obstet Gynecol Scand 1994;73:586–8.
8. Reiter RC. Management of initial atypical cervical cytology: A randomized, prospective study. Obstet Gynecol 1986;68:237–40.
9. Lachin JM. Introduction to sample size determination and power in clinical trials. Control Clin Trial 1981;2:93–113.
10. Hulley SB, Cummings SR. Designing clinical research. Baltimore: Williams & Wilkins; 1988.
11. Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. 2nd ed. Boston: PWS-Kent; 1988.
12. Solomon D, Schiffman M, Tarone R. Comparison of three management strategies for patients with atypical squamous cells of undetermined significance: Baseline result of a randomized trial. J Natl Cancer Inst 2001;93:293–9.
13. The Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesion Triage Study (ALTS) Group. Human papillomavirus testing for triage of women with cytologic evidence of low-grade squamous intraepithelial lesions: Baseline data from a randomized trial. J Natl Cancer Inst 2000;92:397–402.
14. Manos MM, Kinney WK, Hurley LB, Sherman ME, Shieh-Ngai J, Kurman RJ, et al. Identifying women with cervical neoplasia: Using human papillomavirus DNA testing for equivocal Papanicolaou results. JAMA 1999; 281:1605–10.