全文

Objective

The objective of the study was to evaluate the efficacy of intravaginal application of 5% 5-fluorouracil (5-FU) for the treatment of cervical intraepithelial neoplasia (CIN) 2 in women.

 

Study Design

Women aged 18-29 years with CIN 2 were recruited for this randomized controlled trial of observation vs treatment with intravaginal 5-FU. Women in the observation group returned in 6 months for a Papanicolaou smear, colposcopy, and a human papillomavirus (HPV) deoxyribonucleic acid test. Women in the 5-FU group were treated with intravaginal 5-FU once every 2 weeks for a total of 16 weeks and were similarly evaluated at 6 months. All women who had a baseline visit were included in the intention-to-treat analysis. Values of P < .05 were considered statistically significant.

 

Results

Between August 2010 and June 2013, 60 women were randomized and had a baseline visit for intervention (n = 31) vs observation (n = 29). Of women who had cervical biopsy results at 6 months, regression of disease was demonstrated in 93% of women in the 5-FU group (26 of 28) and 56% of women in the observation group (15 of 27). Under the intention-to-treat analysis, a relative risk for cervical disease regression of 1.62 (95% confidence interval [CI], 1.10–2.56) was found between the 5-FU and observation arms (P = .01). When the cervical biopsy, Papanicolaou smear, and HPV results were combined for the 6 month follow-up visit, 50% of the 5-FU group (14 of 28) had a documented normal biopsy, normal Papanicolaou smear, and negative HPV test compared with 22% in the observation group (6 of 27) (relative risk, 2.25; 95% confidence interval, 1.05–5.09; P < .05). There were no moderate or severe side effects in the intervention group.

 

Conclusion

Topical 5-FU appears to be an effective medical therapy for CIN 2 in young women. 5-FU is readily available and may be considered as an off-label treatment option for young women with CIN 2 who are interested in the treatment of disease but want to avoid excisional procedures.

 

Key words

cervical intraepithelial neoplasia; 5-fluorouracil; medical therapy
 

Historically, most women with cervical intraepithelial neoplasia (CIN) 2 and 3 underwent excisional therapy or ablation of the cervical transformation zone.1, 2 and 3 However, 5-26% of women have disease recurrence, even with negative surgical margins.4 Additionally, excisional treatment procedures have been associated with increased risk of premature delivery5, 6, 7 and 8 as well as anxiety, pain, bleeding, and health care expenditure.8, 9, 10 and 11 Because nearly half of CIN 2 lesions regress in young women, current guidelines endorse close observation of young women as a preferable management strategy for CIN 2 and acceptable management strategy for CIN 3.1, 2, 12, 13 and 14 Although expectant management is appealing, approximately one-third of CIN 2 cases will persist and the remaining may progress to CIN 3 on follow-up.12, 13 and 14 There are no medical therapies recommended to promote the clearance of human papillomavirus (HPV) or cervical dysplasia.

 

Topical 5-fluorouracil (5-FU) is used for the treatment of skin cancers and lesions caused by HPV, including genital warts, vulvar intraepithelial neoplasia and vaginal intraepithelial neoplasia.15, 16 and 17 The 5-FU treatment of genital disease is an off-label use of the medication because it has not been approved by the Food and Drug Administration or recommended by the American College of Obstetricians and Gynecologists for this use.

 

Initial treatment regimens were associated with severe side effects such as pain and chronic ulceration.18 These side effects were likely a dose-related response because standard treatments require multiple daily applications; studies limiting topical 5-FU to less frequent application or diluted doses have reported favorable side effect profiles.19, 20 and 21

 

The objective of this study was to assess the efficacy, safety, and acceptability of intravaginal 5-FU as a primary treatment for CIN 2 in young women.

 

Materials and Methods

This was a prospective nonblinded, randomized trial of intravaginal 5% 5-FU vs standard-of-care observation in young women with CIN 2 (no placebo). The primary outcome was regression of disease 6 months after the diagnosis of CIN 2. Secondary outcomes included high-risk HPV status at 6 months, 12 month pathological findings, and safety and acceptability data.

 

This study was approved by the University of North Carolina (UNC) Institutional Review Board. All participants underwent written informed consent procedures.

 

Women (aged 18-29 years) presenting to the UNC's Women's Hospital Clinics with satisfactory colposcopic examinations, a biopsy-confirmed diagnosis of CIN 2, and in whom follow-up observation with cytology and colposcopy every 6 months was planned were approached for study enrollment. Women who were non-English speaking, human immunodeficiency virus (HIV) infected, immunosuppressed, pregnant, planning pregnancy, or breast-feeding during the study time period or were unwilling to use condoms and another form of birth control during the treatment time period were excluded from the study.

 

Dual contraception (condoms plus 1 of the following: oral, intravaginal, injectable, implantable, or intrauterine conception) was required for the 5-FU group because of its potential teratogenic effects demonstrated in intravenous administration.22 Women were counseled regarding teratogenic risk during the consent process.

 

Order of randomization was generated based on a simple randomization table with a 1:1 allocation ratio, and assignments were placed into sequentially numbered opaque envelopes. After verbal interest was reported by a potential participant on the phone, women were randomized to an observation or treatment group (5-FU) by study staff. Participants in the observation group were given the option to have their written consent forms mailed and completion of a background survey by phone. They were scheduled for appointments in 6 months from biopsy date and received standard phone and written appointment reminders from the health care facility in addition to reminders from study staff. Reminders consisted of phone calls, texting, or e-mailing, depending on the preference of the participant.

 

The 5-FU group presented to the study site for written consent procedures and received written and verbal instructions for insertion of 2 g of 5-FU via vaginal applicators every 2 weeks for a total of 8 doses. This dosing schedule was based on its reported safety, tolerability, and efficacy in a prior trial of intravaginal 5-FU for the prevention of recurrence of CIN after excisional procedure in HIV-infected women.21 However, because our study's patient population was not potentially immunocompromised, we chose a shorter course of 16 weeks as per other treatment studies for HPV-related diseases.7, 8, 23 and 24

 

The 5-FU participants were instructed to insert 2 g of topical 5% 5-FU cream (Efudex; Valeant Pharmaceuticals International, Quebec, Canada) at night with a vaginal applicator, which could be twisted onto the study tube for the removal of 5-FU cream. After the medication was inserted into the vagina proximal to the cervix, participants placed a tampon per vagina overnight to keep the cream at the cervix. Participants were instructed to remove the tampon in the morning, shower, and frequently hand wash and change panty liners over the next 2 days to avoid irritation from the cream.

 

Women were supplied home pregnancy tests for use prior to each application of cream. Condoms or abstinence from sexual activity 48 hours after use of the cream was also recommended to diminish any potential irritation to the participant's sexual partner. All supplies were provided to the participants. Use of study drug was delayed if a participant was having her menstrual cycle.

 

Participants in the 5-FU group returned for a safety and acceptability visit at the study site after 8-16 weeks of use of intravaginal 5-FU to complete a survey and pelvic examination. A single unblinded coinvestigator assessed adherence and genital symptoms and performed a colposcopic examination of vaginal and cervical tissue as per the National Institutes of Health Division of AIDS Table for Grading the Severity of Adult and Pediatric Adverse Events: Female Genital Grading Table for Use in Microbicide Studies (Table 3 for a listing of criterion.).25

 

The interview included an acceptability questionnaire in which participants were surveyed regarding emotions felt while using the cream and side effects to their sexual partner; logistical concerns surrounding use of the cream were scored on a 5 point Likert scale. Prior to the 6 month visit, all participants received hospital-based and study-based reminders similar to the observation group described above.

 

At the 6 month visit, colposcopically guided cervical biopsies were performed at the site previously biopsied and diagnosed with CIN 2 on original histology 6 months prior for both intervention and control arms. In all participants, additional biopsies were obtained if there was clinical concern for other areas of dysplasia. We completed similar procedures at the 12 month visit. If the participant had normal results at the 6 month visit and had a normal colposcopy, no biopsies were taken at the 12 month visit unless abnormal colposcopic findings were noted. Women who had treatment procedures (because of a CIN diagnosis or personal preference because of persistent CIN 2) during the study were withdrawn from the study. The colposcopist, a board-certified obstetrician-gynecologist and the director of the colposcopy clinic, was not blinded to the study group of the participants.

 

HPV testing was performed at the 6 and 12 month visits using the Digene Hybrid Capture 2 high risk HPV DNA test at the Department of Pathology, Yale University (New Haven, CT). A single pathologist coinvestigator, board-certified in cytopathology with a focused subspecialty practice in gynecological surgical and cytopathology, was blinded to our randomization scheme and reviewed all cytology and pathologic specimens for entry criteria and for follow-up cytology and histology. Additional p16 immunohistochemistry staining was conducted if clinically indicated.

 

Sample size and statistical analysis

Sample size calculations were based on the primary outcome of the regression of disease. Based on published literature, we assumed that 50% of CIN 2 lesions would regress with observation alone,13 and 14 and a clinically significant difference was estimated at 30%. The total sample size required for the study was 89 women (45 in each arm, α = 0.05, β = 0.80 and 2-sided Student t test).

 

The differences between the treatment and observation groups were assessed for all categorical baseline variables and outcomes using 2-sided Fisher exact tests of association. Continuous baseline variables were assessed using exact Wilcoxon rank sum tests. Histological biopsy results at the 6 month visit were separated into a dichotomous variable indicating regression of disease vs persistence or progression of disease.

 

The association between the regression of disease (regression or persistence/progression) and group status (5-FU and observation) was determined using a 2-sided Fisher exact test of association. Additionally, relative risk (RR) with exact confidence interval (CI) was calculated to aid in understanding the direction of the association. The analyses were completed using intention-to-treat (ITT) methodology and by imputing histological biopsy results for the women with missing 6 month measurements. Imputed values were found using the last observation carried forward method (ie, women without results were designated to still have their baseline result [CIN 2, CIN 3]). Furthermore, a sensitivity analysis was conducted by removing the women with missing 6 month measurements and repeating the analysis to ensure the imputation method was conservative. Values of P < .05 were considered statistically significant. Other variables assessed were histological results at 12 months after diagnosis, cervical cytology, HPV presence, and genotype. All analyses were completed using SAS 9.3 (SAS Institute, Cary, NC) or StatXact version 9 (Cytel, Cambridge, MA).

 

Results

Between Aug. 1, 2010, and June 1, 2012, 93 women who met inclusion criteria were approached for enrollment (Figure). Twenty-six women declined and another 7 were subsequently excluded because of protocol violations prior to the baseline visit. Sixty women were randomized and underwent a baseline enrollment survey. Baseline characteristics of both groups are described in Table 1.

 

        
        

 

        
       
            

 

There were no statistically significant differences between the 5-FU and observation groups. The median age of participants was 24 and 23 years in the 5-FU and observation groups, respectively (P = .32). There were no differences between the groups in regard to the following CIN- or HPV-related parameters: tobacco use, first abnormal Papanicolaou smear, a prior history of CIN 2 or 3, or a prior history of cryotherapy or loop electrosurgical excision procedure. Two women in the 5-FU group and 1 woman in the observation group also were noted to have findings consistent with both CIN 2 and 3. All histological, cytological, and HPV results at each measured time point are described in Table 2.

 

        
        

 

         
        

 

Six month outcomes

Of the 60 women who underwent randomization and baseline procedures, 56 women, 28 women in each group, were analyzed at the 6 month visit. Cervical biopsy results are available on 55 of the 60 women (Figure). One woman in the observation group was not biopsied because she was pregnant at the 6 month visit; therefore, her 6 month cervical biopsy result was considered missing and was imputed using the last observation carried forward method, which is the same method used for the other 4 women missing 6 month visits. A range of 1–4 quadrant biopsies (mean, 2.7) was assessed for diagnosis in each participant.

 

With the ITT analysis, regression of disease was demonstrated in 84% of women in the 5-FU group (26 of 31) and 52% of women in the observation group (15 of 29). An RR of 1.62 (95% CI, 1.10–2.56) was observed, revealing that significantly more women in the 5-FU group had regression of disease (P = .01). Under the sensitivity analysis, regression of disease was demonstrated in 93% of women in the 5-FU group (26 of 28); 56% of women in the 5-FU group (15 of 27) were HPV negative compared with 26% of women in the observation group (7 of 27), a RR of 2.14 (95% CI, 1.07–5.37; P = .05).

 

Given the possibility we could miss an abnormality by basing our results solely on colposcopically directed biopsies, we repeated the ITT analysis, adjusting for the women with discrepant Papanicolaou and biopsy results (ie, women with regression of disease on biopsy results) but also with either a high-grade squamous epithelial lesion (HSIL) or atypical squamous cells of undetermined significance-favor high grade (ASCH) Papanicolaou result. Two women from the 5-FU group with normal biopsies and an ASCH and HSIL Papanicolaou result and 1 woman in the observation group with a normal biopsy and ASCH Papanicolaou were attributed to the persistent disease group.

 

Under ITT analysis, regression of disease was demonstrated in 77% of the 5-FU group (24 of 31) and 48% of the observation group (14 of 29). An RR of 1.60 (95% CI, 1.06–2.61) was found between the 5-FU and observation arms, respectively (P = .03). Under the sensitivity analysis, these 3 women were additionally removed from the analysis, and regression of disease was demonstrated in 92% of women in the 5-FU group (24 of 26) and 54% of women in the observation group (14 of 26), an RR of 1.71 (95% CI, 1.20–2.71, P < .01).

 

Lastly, we combined the biopsy, Papanicolaou smear, and HPV data into 1 variable of normal, normal, and negative results, respectively. When cervical biopsy, Papanicolaou smear and HPV results were combined for the 6 month follow-up visit, 50% of the 5-FU group (14 of 28) had a documented normal biopsy, normal Papanicolaou smear, and negative HPV test compared with 22% in the observation group (6 of 27) (RR, 2.25; 95% CI, 1.05–5.09; P < .05). Atypical squamous cells of uncertain significance Papanicolaou results with negative HPV testing were considered in the normal Papanicolaou smear category as per current clinical standards. 1

 

Twelve month outcomes

As a secondary outcome, we assessed the 36 women (20 in the 5-FU group and 16 in the observation group) who returned for follow-up procedures at 12 months. Five of the women who completed the 6 month visit were no longer enrolled because they had loop electrosurgical excision procedure procedures at the 6 month time point, and 15 of the women who completed the 6 month visit had withdrawn or were lost to follow-up. In the observation group, 55% of the women (16 of 29) had 12 month visits. Of these women, 31.3% (5 of 16) had normal biopsies, 12.5% (2 of 16) had CIN 1, 18.8% (3 of 16) had CIN 2, 12.5% (2 of 16) had CIN 3, and 25.0% (4 of 16) were not biopsied because of normal colposcopic findings. The 4 women who were not biopsied also had negative Papanicolaou smears and HPV results.

 

In the 5-FU group, data were collected at the 12 month visit for 65% of the women (20 of 31). Of these women, 50.0% (10 of 20) had normal biopsies, no women had CIN 1, 5.0% (1 of 20) had CIN 2, no women had CIN 3, and 45.0% (9 of 20) were not biopsied because of normal colposcopic findings. Of the women with normal colposcopic findings, 6 had normal Papanicolaou and negative HPV results, 1 had a normal Papanicolaou and positive HPV result, and 2 had atypical squamous cells of uncertain significance Papanicolaou smears and positive HPV results.

 

5-FU safety and acceptability

Ninety-four percent of 5-FU participants (29 of 31) completed an acceptability questionnaire and colposcopic examination after using a median of 4 doses (range, 4–8 doses) of the study drug. Findings on colposcopic examination during the use of the cream demonstrated that most women had grossly normal examinations. Abnormal examination findings were classified as minimal based the adverse events grading scale described in Materials and Methods 7 ( Table 3). Overall, 48% of women reported a side effect such as pain, bleeding, discharge, burning/itching/irritation, urinary concerns, or other concerns. The majority of women reported discharge and burning/itching/irritation as their chief concern.

 

Participants were asked whether any of the reported symptoms (Table 3) caused interference with any usual activities. The specific question was as follows: do any of the symptoms you described cause you interference with your usual activities? In other words, do they interfere with you working, shopping, cooking, going out, or doing things that you enjoy? No participant reported interference. Participants were then asked, “On a scale of 1-10, describe how much interference you have?” (1 is minimal, 10 is severe). The range of reported interference was 1–4. Of note, 2 women experienced vulvar irritation, which resolved within 2 weeks of stopping the cream. One of these women had a single ulceration and another had erythema on her vulva. Given the rare occurrence of such symptoms, investigators withdrew both women from further use of the cream. However, both women reported that the symptoms did not interfere with daily activities. A third woman discontinued the use of the cream prior to completion because of anxiety.

 

At the 6 month follow-up, participants in the 5-FU group self-reported a mean of 7.5 (range, 2.0–8.0) applications of the study drug; 86% of women (24 of 28) reported using all 8 doses. No women became pregnant while using the 5-FU cream.

 

Women were asked additional questions regarding their experience using intravaginal 5-FU. Despite 48% of patients reporting at least 1 side effect, 83% of participants (24 of 29) reported feeling overall satisfied with the use of the 5-FU cream. Ninety-seven percent women (28 of 29) believed the cream was safe, and all were confident they had used the cream correctly. Ten percent women (3 of 29) agreed that they were concerned that they “may hurt themselves while using the cream.” No women reported any adverse events experienced by their sexual partner. When asked regarding emotions felt surrounding use of the cream, women reported feeling the following: anxious (14%), afraid (3%), embarrassed (0%), empowered (10%), overwhelmed (0%), or intimidated (0%). Sixty-two percent reported no emotion at all.

 

Comment

In this randomized trial of clinical outcomes after the use of intravaginal 5-FU, we report that there was a significantly increased likelihood of histopathological regression of disease. Regression of CIN 2 was demonstrated in 93% of women in the 5-FU group (26 of 28) and 56% of women in the observation group (15 of 27). By the 6 month visit, women in the 5-FU group were twice as likely to be HPV negative than women in the control group. When all outcomes were combined, women in the 5-FU group were twice as likely to have negative cervical biopsies, Papanicolaou smears, and HPV tests at 6 months. Lastly, we demonstrated that topical 5-FU dosed every 2 weeks was well tolerated by healthy women. The efficacy and acceptability of 5-FU in this dosing format has not previously been studied in this population and explores the possibility of a patient-controlled management option for CIN.

 

This study provides additional evidence to support the pursuit of a feasible topical treatment of CIN. 5-FU is an off-label therapeutic option for the treatment of vaginal intraepithelial neoplasia and has previously been shown to have an impact on the recurrence of CIN.21 and 16 In a trial of intravaginal 5-FU use for the treatment of cervical dysplasia in HIV-infected women after excisional procedures for CIN 2-3, treatment with 5-FU was significantly associated with prolonged time to CIN development (P = .04). 21 Participants in the observation arm of this trial were more likely to have CIN 2-3 recurrence (P = .01). 21 Toxicity has always been a concern with 5-FU therapy; however, this study, from which we based our clinical trial design, reported no high-grade toxicities after 18 months of follow-up. Approximately one third of participants reported local reproductive tract side effects. 21

 

Studies of intravaginal cidofovir and imiquimod have shown efficacy in clinical regression of disease and clearance of HPV in women with CIN 2 and 3.27 and 26 In the imiquimod study, 73% of women underwent histological regression of CIN 2 and 3 compared with 39% of women using placebo (P = .009). 26 In the cidofovir study, histological clearance was demonstrated in 61% compared with 20% in the treatment and placebo groups, respectively (P < .01). 27 Although comparable efficacy to our study, both therapies required application multiple times per week (3 times per week for cidofovir, 1-3 times for imiquimod) and also reported local and systemic side effects. Those studies carried the advantage of placebo-control; however, in all studies, including our own, the pathologist was blinded to treatment results.

 

We purposely did not choose a placebo-control because we did not want to introduce any intervention that would interfere with the natural regression of disease (eg, increased condom use, placebo altering the vaginal milieu), which is the current standard of care. Our sample size is comparable with these other studies (59 and 48 participants in the imiquimod and cidofovir studies, respectively). Although larger sample sizes will be required for further investigation and validation of our results, we demonstrated regression of disease in women in the 5-FU group. None of these therapies, including 5-FU, are Food and Drug Administration approved or recommended by American College of Obstetricians and Gynecologists for the management of cervical dysplasia.

 

One limitation of our study is that biopsy outcomes were, for the most part, not based on conization biopsies or 4 quadrant biopsies with p16 immunohistochemical staining but clinically oriented, colposcopically directed biopsies (mean, 2.7). We do not have data on the size of the lesion at baseline. Although we had a single pathologist evaluate all specimens, this method introduces the concern for bias on the part of the investigator conducting biopsies, which may have led to erroneous diagnoses if nondiseased areas were biopsied. We attempted to address this issue by adjusting for women with discrepant Papanicolaou smears (HSIL and ASCH Papanicolaou smear result associated with CIN 1 or normal biopsies) and still found a resultant risk difference that was statistically significant in favor of the 5-FU group using 2 different analyses.

 

Additionally, we compiled the biopsy, Papanicolaou smear, and HPV data, and found that women in the 5-FU group were twice as likely to have a normal biopsy, normal Papanicolaou smear, and negative HPV test than the women in the observation group. Although some of the rigor of clinical trial design was lost in retrieving data from a clinical setting, this limitation was also an advantage of this study in that we were able to conduct a trial on a hard-to-reach young population who may not usually participate in clinical trials. More than half of the participants did not have private or government-sponsored health insurance. We had only an 8% dropout rate at the 6 month follow up.

 

In this prospective randomized trial, we demonstrated that intravaginal 5-FU was effective in promoting the regression of CIN 2 and was safe and acceptable to healthy women aged 18-29 years of age. However, it is important to note that despite the apparent efficacy of topical 5-FU in this study, we must caution that counseling for the prevention of pregnancy is needed when using this medication in women of child-bearing age because of its potential for teratogenic effects in its intravenous form.22 Additionally, because of the potential for mucosal disruption, we do not know whether this therapy could put a woman at higher risk of sexually transmitted infections. Nonetheless, these findings provide young women with a potential medical option for intervention after diagnosis of CIN 2 rather than having to choose between doing nothing (observation) or surgical treatment. The availability of such an option may reduce both patient and provider anxiety over a CIN 2 diagnosis because these fears may propel both parties toward more conservative surgical interventions with the potential for consequences in future pregnancies.

 

Future research aimed at understanding whether topical 5-FU decreases the time toward the inevitable regression of CIN 2, its effects on HPV infection itself, its use in CIN 3 or older women, or whether it is a comparable therapy with an excision or ablation procedure is needed. However, this study and prior research using this drug regimen demonstrate that topical 5-FU is a readily available, well-tolerated, patient-controlled medical treatment for cervical dysplasia in young women. Although excision and ablation are well-established, standard-of-care procedures, the time has come for our field to consider and study medical management options as alternatives for women who decline or cannot access surgical treatments.

 

Acknowledgments

We would like to thank nursing and support staff of the UNC Women's Hospital outpatient clinics for their assistance with completion of this study. We would also like to thank Jennifer Cayless (UNC Department of Pathology) and the UNC Center for Women's Health Research for study coordination.