HPV Vaccine in Women With Cervical and Vulvar Disease
HPV Vaccine in Women With Cervical and Vulvar Disease
The primary objective of this retrospective analysis was to determine whether administration of quadrivalent HPV vaccine, compared with placebo, reduced the incidence of subsequent HPV related disease among women who had undergone surgery for cervical disease or who were diagnosed with vulvar or vaginal disease (genital warts, vulvar intraepithelial neoplasia, or vaginal intraepithelial neoplasia) after enrolment into the FUTURE I or FUTURE II clinical trials (Figure 1). Subsequent HPV related disease is defined as any disease that was detected at least 60 days after treatment or diagnosis. We measured the vaccine’s impact on end points that were associated with HPV types 6, 11, 16, and 18 specifically and on end points irrespective of causal HPV type.
(Enlarge Image)
Figure 1.
Study design for assessing effect of quadrivalent HPV vaccine on incidence of subsequent HPV related disease among women who had undergone surgery for cervical disease or who were diagnosed with vulvar or vaginal disease. Subsequent disease was measured from 60 days after surgery or diagnosis
Between December 2001 and May 2003, 17,622 women aged 15–26 years were enrolled in one of two randomised, double blind, placebo controlled trials (FUTURE I and FUTURE II). The studies were conducted in accordance with principles of Good Clinical Practice and were approved by the appropriate institutional review boards and regulatory agencies.
The study designs and the results of the primary hypotheses have been described elsewhere. Women were eligible to participate if they were not pregnant, did not report previous abnormal results on a cervical smear test, and had had a lifetime number of no more than four sex partners. Neither study included HPV testing (DNA or serology testing) or clinical examination before randomisation; thus, the trials allowed the enrolment of women with ongoing HPV infection or disease.
In both trials, participants were randomly assigned (1:1) to receive intramuscular injections of quadrivalent HPV vaccine (Gardasil or Silgard, Merck, Whitehouse Station, NJ) or visually indistinguishable placebo at day 1, month 2, and month 6. Comprehensive anogenital examinations and ThinPrep cervical cytology (Cytyc, Boxborough, MA, USA) were performed during scheduled visits, which occurred every 6–12 months after the vaccination phase. Cytology specimens were classified according to the Bethesda System 2001. Women reported any abnormality or suspected disease in between the scheduled study visits. All participants were required to use birth control during the vaccination phase (day 1 through to month 7), and all participants were evaluated for pregnancy before the administration of each dose of vaccine or placebo with a human chorionic gonadotropin assay sensitive to 25 international units. Counselling about safe sex and contraceptive use was part of the study protocol.
Protocol specified guidelines were used to triage women with smear test abnormalities to colposcopy. Colposcopists were trained to locate and biopsy all discrete abnormal areas on the cervix and lower genital tract. Loop electrosurgical excision was the preferred method for definitive therapy for cervical intraepithelial neoplasia grade II-III, adenocarcinoma in situ, and persistent cervical intraepithelial neoplasia grade I. However, other methods included cervix conisation, cryotherapy, and electrocauterisation. The frequencies of each technique recorded were loop electrosurgical excision (84.7%), cervical conisation (12.5%), cryotherapy (0.7%), and other (2.1%). As loop electrosurgical excision and cervix conisation accounted for 97.2% of the procedures performed, with no difference between the treatment groups (97.1% with placebo, 97.4% with vaccine) we use the term “cervical surgery” to describe any method used to treat cervical disease. Management of vulvar or vaginal disease followed the investigational site’s standards and practice and included surgical or topical therapy.
In this retrospective, intention to treat analysis, we identified all women who underwent definitive cervical therapy or who were diagnosed with vulvar or vaginal disease in the pooled FUTURE I and II studies after being randomised to and receiving at least one dose of HPV vaccine or placebo. The incidence rates for subsequent disease were calculated with case counting starting 60 days after cervical surgery or diagnosis of vulvar or vaginal disease.
As stated previously, women were referred for colposcopy and cervical surgery in accordance with a triage algorithm. In FUTURE I, the pre-specified primary end points included all HPV related anogenital disease, thus the study included stringent criteria for disease determination and follow-up, including more frequent examinations and screening and more aggressive colposcopy triage for suspected disease. In FUTURE II, the primary end point was cervical intraepithelial neoplasia grade II-III and adenocarcinoma in situ. Treatment of vulvar or vaginal disease in FUTURE II was based on local standards of care, and the investigators were not required to record the date or method of treatment. We therefore examined the average time from diagnosis to treatment of genital warts, vulvar intraepithelial neoplasia, and vaginal intraepithelial neoplasia within the FUTURE I study. That information was used to determine an appropriate time frame after which most women would have been treated, so that follow-up after that window would probably capture new, rather than residual, disease. In FUTURE I, the mean number of days between diagnoses and treatment of genital warts, vulvar intraepithelial neoplasia, and vaginal intraepithelial neoplasia was 28 days (interquartile range 0–42 days). As the trials were less than four years of total duration (mean of 3.6 years), and since 82% had treatment within 60 days of the pathology panel diagnosis, we choose this as the cut off to ensure there would be enough follow-up time to detect subsequent disease after treatment. We assumed a similar time window between treatment and diagnosis in FUTURE II, and used the same window (60 days after diagnosis) for both studies. For consistency, the same time window was applied for counting subsequent disease after cervical surgery. A sensitivity analysis was also performed using a 90 day time window, as 91% of subjects had treatment within 90 days.
All specimens from biopsies and excisional procedures were tested for 14 HPV types (6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59) using an assay based on the polymerase chain reaction. All tissue specimens underwent histopathological review by a pathology panel, who were masked to the women’s vaccination group and HPV status. A woman was considered to have developed an end point related to vaccine HPV types (HPV 6, 11, 16, or 18) or an end point related to 10 non-vaccine types (HPV 31, 33, 35, 39, 45, 51, 52, 56, 58, or 59) if the respective HPV DNA were detected in the same lesion that was diagnosed by the pathology panel as cervical intraepithelial neoplasia, vulvar intraepithelial neoplasia, vaginal intraepithelial neoplasia, or genital warts. For analyses of all disease, irrespective of causal HPV type, a woman was considered to have developed an end point if she had a lesion diagnosed by the pathology panel without consideration of HPV status.
In the analyses (presented in Table 2 , Table 3 , Table 4 ) a woman is counted only once for each end point (that is, once in each row), but a woman may have developed more than one end point during the trial (that is, a women may appear in more than one row). For example, a woman who is diagnosed with a cervical intraepithelial grade 3 lesion that is positive to HPV types 16, 31, and 45 would be counted once for (1) any disease; (2) any cervical disease; (3) any cervical intraepithelial grade II or worse; (4) any cervical intraepithelial grade III or worse; (5) any disease related to vaccine HPV types; (6) cervical intraepithelial grade I or worse related to vaccine HPV types; (7) cervical intraepithelial grade II or worse related to vaccine HPV types; and (8) cervical intraepithelial grade III or worse related to vaccine HPV types.
Women contributed to person years at risk from 60 days after the initial surgery or diagnosis until the day of ascertainment of the subsequent disease end point, or, for those without a subsequent disease end point, until the day of their last follow-up visit. Incidence, or rate of subsequent disease, is expressed as the number of women with an end point per 100 person years at risk. A point estimate of vaccine efficacy (that is, percentage reduction of the relative end point) and the 95% confidence interval were calculated on the basis of the observed case split between vaccine and placebo recipients, adjusted for the accrued person time in each group. The criterion for statistical significance (P<0.05) was equivalent to requiring that the lower bound of the confidence interval for vaccine efficacy exclude 0%. An exact conditional procedure was used to evaluate vaccine efficacy under the assumption that the numbers of cases in the vaccine and placebo groups were independent Poisson random variables.
Kaplan-Meier estimates (and 95% confidence intervals) of cumulative incidence rates for subsequent disease were plotted with case counting starting 60 days after cervical surgery or vulvar or vaginal disease diagnosis. The plots were not part of a formal survival analysis; rather they give a visual demonstration of the divergence of the incidence rates between the two vaccination arms over time. The 95% confidence interval for the Kaplan-Meier incidence at intervals of six months are not directly comparable with vaccine efficacy estimates. Vaccine efficacy was calculated from the exact conditional procedure and is included in the Kaplan-Meier plots as a reference.
Methods
Objectives
The primary objective of this retrospective analysis was to determine whether administration of quadrivalent HPV vaccine, compared with placebo, reduced the incidence of subsequent HPV related disease among women who had undergone surgery for cervical disease or who were diagnosed with vulvar or vaginal disease (genital warts, vulvar intraepithelial neoplasia, or vaginal intraepithelial neoplasia) after enrolment into the FUTURE I or FUTURE II clinical trials (Figure 1). Subsequent HPV related disease is defined as any disease that was detected at least 60 days after treatment or diagnosis. We measured the vaccine’s impact on end points that were associated with HPV types 6, 11, 16, and 18 specifically and on end points irrespective of causal HPV type.
(Enlarge Image)
Figure 1.
Study design for assessing effect of quadrivalent HPV vaccine on incidence of subsequent HPV related disease among women who had undergone surgery for cervical disease or who were diagnosed with vulvar or vaginal disease. Subsequent disease was measured from 60 days after surgery or diagnosis
Study Population
Between December 2001 and May 2003, 17,622 women aged 15–26 years were enrolled in one of two randomised, double blind, placebo controlled trials (FUTURE I and FUTURE II). The studies were conducted in accordance with principles of Good Clinical Practice and were approved by the appropriate institutional review boards and regulatory agencies.
The study designs and the results of the primary hypotheses have been described elsewhere. Women were eligible to participate if they were not pregnant, did not report previous abnormal results on a cervical smear test, and had had a lifetime number of no more than four sex partners. Neither study included HPV testing (DNA or serology testing) or clinical examination before randomisation; thus, the trials allowed the enrolment of women with ongoing HPV infection or disease.
Intervention
In both trials, participants were randomly assigned (1:1) to receive intramuscular injections of quadrivalent HPV vaccine (Gardasil or Silgard, Merck, Whitehouse Station, NJ) or visually indistinguishable placebo at day 1, month 2, and month 6. Comprehensive anogenital examinations and ThinPrep cervical cytology (Cytyc, Boxborough, MA, USA) were performed during scheduled visits, which occurred every 6–12 months after the vaccination phase. Cytology specimens were classified according to the Bethesda System 2001. Women reported any abnormality or suspected disease in between the scheduled study visits. All participants were required to use birth control during the vaccination phase (day 1 through to month 7), and all participants were evaluated for pregnancy before the administration of each dose of vaccine or placebo with a human chorionic gonadotropin assay sensitive to 25 international units. Counselling about safe sex and contraceptive use was part of the study protocol.
Protocol specified guidelines were used to triage women with smear test abnormalities to colposcopy. Colposcopists were trained to locate and biopsy all discrete abnormal areas on the cervix and lower genital tract. Loop electrosurgical excision was the preferred method for definitive therapy for cervical intraepithelial neoplasia grade II-III, adenocarcinoma in situ, and persistent cervical intraepithelial neoplasia grade I. However, other methods included cervix conisation, cryotherapy, and electrocauterisation. The frequencies of each technique recorded were loop electrosurgical excision (84.7%), cervical conisation (12.5%), cryotherapy (0.7%), and other (2.1%). As loop electrosurgical excision and cervix conisation accounted for 97.2% of the procedures performed, with no difference between the treatment groups (97.1% with placebo, 97.4% with vaccine) we use the term “cervical surgery” to describe any method used to treat cervical disease. Management of vulvar or vaginal disease followed the investigational site’s standards and practice and included surgical or topical therapy.
Population for Retrospective Subanalysis
In this retrospective, intention to treat analysis, we identified all women who underwent definitive cervical therapy or who were diagnosed with vulvar or vaginal disease in the pooled FUTURE I and II studies after being randomised to and receiving at least one dose of HPV vaccine or placebo. The incidence rates for subsequent disease were calculated with case counting starting 60 days after cervical surgery or diagnosis of vulvar or vaginal disease.
Rationale for 60 Day Window to Define Subsequent Disease
As stated previously, women were referred for colposcopy and cervical surgery in accordance with a triage algorithm. In FUTURE I, the pre-specified primary end points included all HPV related anogenital disease, thus the study included stringent criteria for disease determination and follow-up, including more frequent examinations and screening and more aggressive colposcopy triage for suspected disease. In FUTURE II, the primary end point was cervical intraepithelial neoplasia grade II-III and adenocarcinoma in situ. Treatment of vulvar or vaginal disease in FUTURE II was based on local standards of care, and the investigators were not required to record the date or method of treatment. We therefore examined the average time from diagnosis to treatment of genital warts, vulvar intraepithelial neoplasia, and vaginal intraepithelial neoplasia within the FUTURE I study. That information was used to determine an appropriate time frame after which most women would have been treated, so that follow-up after that window would probably capture new, rather than residual, disease. In FUTURE I, the mean number of days between diagnoses and treatment of genital warts, vulvar intraepithelial neoplasia, and vaginal intraepithelial neoplasia was 28 days (interquartile range 0–42 days). As the trials were less than four years of total duration (mean of 3.6 years), and since 82% had treatment within 60 days of the pathology panel diagnosis, we choose this as the cut off to ensure there would be enough follow-up time to detect subsequent disease after treatment. We assumed a similar time window between treatment and diagnosis in FUTURE II, and used the same window (60 days after diagnosis) for both studies. For consistency, the same time window was applied for counting subsequent disease after cervical surgery. A sensitivity analysis was also performed using a 90 day time window, as 91% of subjects had treatment within 90 days.
Efficacy End Points and Statistics
All specimens from biopsies and excisional procedures were tested for 14 HPV types (6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59) using an assay based on the polymerase chain reaction. All tissue specimens underwent histopathological review by a pathology panel, who were masked to the women’s vaccination group and HPV status. A woman was considered to have developed an end point related to vaccine HPV types (HPV 6, 11, 16, or 18) or an end point related to 10 non-vaccine types (HPV 31, 33, 35, 39, 45, 51, 52, 56, 58, or 59) if the respective HPV DNA were detected in the same lesion that was diagnosed by the pathology panel as cervical intraepithelial neoplasia, vulvar intraepithelial neoplasia, vaginal intraepithelial neoplasia, or genital warts. For analyses of all disease, irrespective of causal HPV type, a woman was considered to have developed an end point if she had a lesion diagnosed by the pathology panel without consideration of HPV status.
In the analyses (presented in Table 2 , Table 3 , Table 4 ) a woman is counted only once for each end point (that is, once in each row), but a woman may have developed more than one end point during the trial (that is, a women may appear in more than one row). For example, a woman who is diagnosed with a cervical intraepithelial grade 3 lesion that is positive to HPV types 16, 31, and 45 would be counted once for (1) any disease; (2) any cervical disease; (3) any cervical intraepithelial grade II or worse; (4) any cervical intraepithelial grade III or worse; (5) any disease related to vaccine HPV types; (6) cervical intraepithelial grade I or worse related to vaccine HPV types; (7) cervical intraepithelial grade II or worse related to vaccine HPV types; and (8) cervical intraepithelial grade III or worse related to vaccine HPV types.
Women contributed to person years at risk from 60 days after the initial surgery or diagnosis until the day of ascertainment of the subsequent disease end point, or, for those without a subsequent disease end point, until the day of their last follow-up visit. Incidence, or rate of subsequent disease, is expressed as the number of women with an end point per 100 person years at risk. A point estimate of vaccine efficacy (that is, percentage reduction of the relative end point) and the 95% confidence interval were calculated on the basis of the observed case split between vaccine and placebo recipients, adjusted for the accrued person time in each group. The criterion for statistical significance (P<0.05) was equivalent to requiring that the lower bound of the confidence interval for vaccine efficacy exclude 0%. An exact conditional procedure was used to evaluate vaccine efficacy under the assumption that the numbers of cases in the vaccine and placebo groups were independent Poisson random variables.
Kaplan-Meier estimates (and 95% confidence intervals) of cumulative incidence rates for subsequent disease were plotted with case counting starting 60 days after cervical surgery or vulvar or vaginal disease diagnosis. The plots were not part of a formal survival analysis; rather they give a visual demonstration of the divergence of the incidence rates between the two vaccination arms over time. The 95% confidence interval for the Kaplan-Meier incidence at intervals of six months are not directly comparable with vaccine efficacy estimates. Vaccine efficacy was calculated from the exact conditional procedure and is included in the Kaplan-Meier plots as a reference.
