Between September 3, 2004, and August 29, 2008, we enrolled 4065 healthy boys and men from 71 sites in 18 countries in a randomized, placebo-controlled, double-blind study. A total of 3463 of the subjects were heterosexual (i.e., they reported that their sexual partners were exclusively female), and 602 had sex with male partners (i.e., they reported that they had engaged in insertive or receptive anal intercourse or oral sex with a male partner within the previous year). The median follow-up period, after administration of the first dose of vaccine or placebo, was 2.9 years.
For the heterosexual subjects, eligibility criteria were an age between 16 and 23 years and one to five female sexual partners during their lifetime; for the subjects who had sex with male partners, the criteria were an age between 16 and 26 years and one to five male or female partners during their lifetime. Subjects who had clinically detectable anogenital warts or genital lesions at screening that were suggestive of infection with non-HPV sexually transmitted diseases, or who had a history of such findings, were excluded.
STUDY DESIGN AND OVERSIGHT
Subjects were randomly assigned in a 1:1 ratio to receive quadrivalent HPV vaccine or placebo at day 1, month 2 (±3 weeks), and month 6 (±4 weeks). Vaccine or placebo was administered as a 0.5-ml injection in the deltoid muscle (with all three doses administered in the same arm). The quadrivalent HPV L1 vaccine (Gardasil or Silgard, Merck) contains HPV-6, 11, 16, and 18 viruslike particles conjugated to an amorphous aluminum hydroxyphosphate sulfate (AAHS) adjuvant; the vaccine and the visually indistinguishable AAHS-containing placebo have been described previously.12
Details on randomization and blinding are provided in the Supplementary Appendix
, available with the full text of this article at NEJM.org.
The trial was designed by the study sponsor, Merck, in collaboration with external investigators and an external data and safety monitoring board. The sponsor collated the data, monitored the conduct of the trial, performed statistical analyses, and coordinated the writing of the manuscript with all the authors. The academic authors were actively involved in the collection, analysis, and interpretation of the data; the initial drafting and revision of the manuscript for intellectual content; and the approval of the final manuscript. The first draft was written by the first author, with contributions from other authors. All authors vouch for the completeness and accuracy of the data presented. All authors had access to all the study data (with confidentiality agreements) and participated in the decision to submit the manuscript for publication.
The trial was conducted in accordance with the protocol (available at NEJM.org), which was approved by the institutional review boards at participating centers. Written informed consent was obtained from all subjects. At each study site, the trial was conducted in conformity with applicable country or local requirements regarding ethics committee review, informed consent, and other statutes or regulations regarding the protection of the rights and welfare of human subjects participating in biomedical research.
STUDY POPULATIONS FOR ANALYSES
Analyses were conducted in an intention-to-treat population consisting of subjects who received one or more doses of vaccine or placebo and returned for follow-up. These subjects, who might have been seropositive at enrollment or might have had positive results for the quadrivalent HPV vaccine types on polymerase-chain-reaction (PCR) assay, represented the general population of unvaccinated boys and men. A total of 175 subjects did not return for follow-up after receiving one dose of vaccine or placebo (82 in the vaccine group and 93 in the placebo group). Case counting in the intention-to-treat population commenced after day 1.
An efficacy analysis was also conducted in the per-protocol population — that is, subjects who were seronegative on day 1 and PCR-negative for both swab and biopsy specimens from day 1 through month 7 for the relevant vaccine HPV type (or types) and did not have any protocol violations (see Table S1 in the Supplementary Appendix
). Subjects in the per-protocol population received all 3 vaccinations within 1 year and had 1 or more follow-up visits after month 7. Case counting commenced at month 7.
Additional analyses were conducted in a population of subjects who were negative for HPV-6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59 DNA and were seronegative for HPV-6, 11, 16, and 18 at enrollment and who received at least one dose of vaccine or placebo. This population approximates a population of young men before sexual debut. Case counting in this population began after day 1.
STUDY OBJECTIVES AND MEASURES OF EFFICACY
The primary efficacy objective was to show that the quadrivalent HPV vaccine reduced the incidence of external genital lesions associated with HPV-6, 11, 16, or 18, as compared with placebo. The secondary efficacy objectives were to show that the vaccine reduced the incidence of persistent infection with these HPV types and the detection at any time of DNA associated with these viral types, as compared with placebo. We also analyzed the composite efficacy of the vaccine against the development of external genital lesions related to any HPV type (including HPV types that were identified with the use of a PCR assay [HPV-6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, and 59] and types that were not so identified).
Detailed anogenital examinations were performed on day 1 and at months 7, 12, 18, 24, 30, and 36. Biopsies were performed for external genital lesions judged by the investigator to be possibly, probably, or definitely related to HPV and for any lesion whose cause was not known. Repeat biopsy of recurrent lesions (occurring within 2 months after the previous lesion, at the same location and with the same appearance) was not performed in order to avoid overestimation of incident external genital lesions. All biopsy specimens were processed independently to prevent contamination of HPV DNA and were assessed in a blinded fashion, first for the purpose of clinical management by pathologists at the central laboratory (Diagnostic Cytology Laboratories) and then for end-point adjudication by a four-member panel of pathologists.8,9
Panel members disagreed on the interpretation of 0.3% of the biopsy specimens and resolved the problem by meeting to obtain a consensus. Clinical management was performed according to local standards of care. HPV testing of thin sections was performed at a central laboratory with the use of multiplex PCR assay.8,9
Specimens for HPV testing were collected separately from the penis, scrotum, and perineal and perianal regions with the use of a nail file and Dacron swab on day 1 and at months 7, 12, 18, 24, 30, and 36. In the group of subjects who had sex with male partners, intra-anal specimens were collected with a Dacron swab. All specimens were tested for HPV DNA to identify subjects who were infected before enrollment and those who acquired new HPV infections during the study. Each thin-section and swab specimen was evaluated with three different primer-pair sets per HPV type, which amplified a portion of three separate open reading frames.13
Less than 5% of the PCR-positive biopsy specimens from external genital lesions (11 of 268 positive specimens) were positive for only 1 of 3 genes, and less than 4% of all specimens from external genital lesions (151 of 4886) could not be amplified to at least 1 of 14 HPV types tested.
To allow for assessment of vaccine safety, subjects recorded (on vaccination report cards) oral temperature and any adverse events occurring at the injection site on days 1 through 5 after receiving each dose of vaccine or placebo. They also recorded systemic adverse events and all serious adverse events that occurred on days 1 through 15 after receiving each dose. All serious adverse events that investigators believed to be associated with the vaccine or the study procedure and all deaths were recorded during the entire study period.
The primary end point of the study was the presence or absence of external genital lesions associated with HPV-6, 11, 16, or 18, defined as condylomata acuminata (external genital warts); penile, perianal, or perineal intraepithelial neoplasia (PIN); or penile, perianal, or perineal cancer. All other external genital lesions were also recorded, whether or not they contained an HPV type identified on PCR assay (HPV-6, 11, 16, 18, 31, 33, 35, 39, 45, 51, 52, 56, 58, or 59). Primary end-point cases were confirmed by means of a consensual diagnosis of condyloma acuminatum, PIN grade 1 or grade 2–3, or penile, perianal, or perineal cancer by the pathology panel after examination of a biopsy specimen and by detection of HPV DNA in an adjacent section of the same tissue block with the use of a PCR assay.
Persistent infection was defined as detection of the same HPV type (6, 11, 16, or 18) in an anogenital swab or biopsy specimen collected on two or more consecutive visits, with an interval of at least 6 months (±1 month) between the visits. DNA detection was defined as detection of HPV 6, 11, 16, or 18 DNA in any clinical specimen on one or more visits (with visits made at any time during the study).
To evaluate the primary end point, a fixed-event design was used. Assuming that the true efficacy of the vaccine is 80%, 32 cases would provide at least 90% power to demonstrate an efficacy of more than 20% at a one-sided alpha level of 0.025, and 23 cases in the heterosexual group would provide more than 90% power to demonstrate that vaccine efficacy is more than 0% in that subgroup. Therefore, the primary analysis was conducted when at least 32 cases that satisfied the primary end point had been observed in the entire study population and when at least 23 cases had been diagnosed among the heterosexual subgroup. The statistical criterion for study success required the lower bound of the confidence interval for vaccine efficacy to exclude 20%. Vaccine efficacy was defined as 100% × (1 − [rv/rp]), with rv, the incidence rate among vaccine recipients, defined as Cv, the number of primary efficacy cases among vaccine recipients, divided by τv, the total person-years of follow-up among vaccine recipients. Similarly, rp, the incidence rate among placebo recipients, was defined as Cp, the number of primary efficacy cases among placebo recipients, divided by τp, the total person-years of follow-up among placebo recipients. The hypotheses related to vaccine efficacy were tested by constructing a two-sided exact confidence interval for vaccine efficacy under the assumption that the number of end-point cases among vaccine recipients followed a binomial distribution.
The confidence interval reported for the end point of persistent infection with HPV-6, 11, 16, or 18 in the per-protocol population (97.5%) differs from that reported in other analyses (95%) because of application of the Hochberg multiplicity adjustment. This adjustment was necessary because of the two secondary end points: persistent infection and DNA detection. The adjustment was made so that the combined alpha level of the hypothesis tests for these two end points would not exceed a one-sided alpha level of 0.025.