Human papillomaviruses (HPV) are a group of viruses that infect the skin and mucous membranes. Some types of HPV are sexually transmitted and are common in young people. Most infections will be cleared by the immune system, but some people will experience persistent infection with certain HPV types that go on to cause abnormalities in infected cells. These changes are called 'precancerous' because they can develop into cancers of the cervix, vagina, vulva, anal canal, penis, and head and neck. Infection with other HPV types causes warts in the genital area or around the anus.
Vaccination aims to prevent future HPV infections. Three HPV vaccines are in use – a bivalent one (protects against two HPV types), a quadrivalent one (protects against four HPV types), and a nonavalent one (protects against nine HPV types). In women, three doses of the bivalent or the quadrivalent HPV vaccines protect against precancer of the cervix caused by the HPV types contained in the vaccine. Evidence about the nonavalent vaccine, about the effects of the quadrivalent vaccine in males, and about the effects of HPV vaccines in people with HIV infection, has not yet been reviewed thoroughly. Uptake of HPV vaccines remains low in many countries. Simpler vaccine schedules, or giving the vaccine to both girls and boys, could increase the number of people being vaccinated.
Trials of HPV vaccines are not always designed to collect data about precancer and cancer, for several reasons. Firstly, HPV vaccine is routinely given before girls become sexually active, and it is not ethical to take specimens from the cervix of girls who have not had sex. Secondly, HPV-related precancer and cancer are rare and do not develop until years after HPV infection has occurred. Thirdly, participants in a trial will be offered treatment if precancer develops, so progression to cervical cancer would be even rarer, even without vaccination. An international committee of experts states that, in some circumstances, antibody levels (i.e. showing a strong immune system response), can be used to demonstrate protection against cervical and anal cancer. The antibody levels following vaccination in a trial should not be lower than those found in other studies on adults in whom the vaccine has been shown to protect against severe HPV-related cervical or anal disease.
How effective or harmful are different HPV vaccine schedules (i.e. number and timing of doses) and different HPV vaccines in females and males?
These results are based on research evidence to 27 September 2018. We analysed 20 studies involving 31,940 people.
Studies comparing two doses of HPV vaccine to three doses, or comparing the time interval between doses, focus on immune system responses rather than infection or disease outcomes. Two doses of HPV vaccine result in similar immune system responses to three doses, and a longer interval (up to 12 months) between doses gives a stronger immune system response than a shorter interval. There is insufficient evidence to determine whether there was a difference between the vaccine schedules for serious adverse events and death.
In 16- to 26-year-old men, one study showed evidence of moderate certainty that a quadrivalent HPV vaccine provides better protection against external genital lesions and genital warts than a dummy treatment (control). In 16- to 26-year-old women, one study showed that the nonavalent and quadrivalent vaccines provide the same levels of protection against cervical, vaginal, and vulval precancer lesions and cancer (high-certainty evidence).
There was evidence that the quadrivalent vaccine resulted in more local adverse events (such as pain, swelling, and redness at the injection site) than a control treatment in males, and that the nonavalent vaccine resulted in more local adverse events than the quadrivalent vaccine in males and females. Evidence about serious adverse events and deaths from studies comparing different HPV vaccine types or dose schedules was of low or very low-certainty.
In people living with HIV, HPV vaccines result in reasonable levels of immune system response, but evidence about their effects on persistent HPV infection or HPV-related disease outcomes and harms is limited.
Certainty of the evidence
No major issues were identified with the methodological quality of the studies for the measurements of infection and disease outcomes, or for immune system responses. Our certainty in the evidence about serious harms and deaths across all the studies comparing different HPV vaccines and vaccine schedules is low, either because of their low frequency, or because the evidence is indirect, or both. Evidence graded as high certainty means that we were confident that further research is unlikely to change our findings. Moderate-certainty evidence means that there is a possibility that further research may have an important effect on our findings, whilst low-certainty evidence means that our confidence was limited and further research may have an important impact on our findings. Very low-certainty evidence means that we were uncertain about the result.
A two-dose schedule of HPV vaccines in young females results in immune system responses that are comparable with a three-dose schedule. In males, the quadrivalent HPV vaccine appears to be effective in the prevention of external genital lesions and genital warts. Quadrivalent and nonavalent HPV vaccines in young women result in similar levels of protection against cervical, vaginal, and vulval precancer lesions and cancer. Evidence about the efficacy and harms in people living with HIV is limited. Further long-term population-level studies are needed to continue monitoring safety of these vaccines, to determine for how long two doses of vaccine can provide protection against HPV-related disease, the effect against HPV-related cancer, and whether a two-dose immunisation schedule will increase vaccine coverage.
The immunogenicity of two-dose and three-dose HPV vaccine schedules, measured using antibody responses in young females, is comparable. The quadrivalent vaccine probably reduces external genital lesions and anogenital warts in males compared with control. The nonavalent and quadrivalent vaccines offer similar protection against a combined outcome of cervical, vaginal, and vulval precancer lesions or cancer. In people living with HIV, both the bivalent and quadrivalent HPV vaccines result in high antibody responses. For all comparisons of alternative HPV vaccine schedules, the certainty of the body of evidence about serious adverse events reported during the study periods was low or very low, either because the number of events was low, or the evidence was indirect, or both. Post-marketing surveillance is needed to continue monitoring harms that might be associated with HPV vaccines in the population, and this evidence will be incorporated in future updates of this review. Long-term observational studies are needed to determine the effectiveness of reduced-dose schedules against HPV-related cancer endpoints, and whether adopting these schedules improves vaccine coverage rates.
Uptake of human papillomavirus (HPV) vaccine remains low in many countries, although the bivalent and quadrivalent HPV vaccines given as a three-dose schedule are effective in the prevention of precancerous lesions of the cervix in women. Simpler immunisation schedules, such as those with fewer doses, might reduce barriers to vaccination, as may programmes that include males.
To evaluate the efficacy, immunogenicity, and harms of different dose schedules and different types of HPV vaccines in females and males.
We conducted electronic searches on 27 September 2018 in Ovid MEDLINE, the Cochrane Central Register of Controlled Trials (CENTRAL) (in the Cochrane Library), and Ovid Embase. We also searched the WHO International Clinical Trials Registry Platform, and ClinicalTrials.gov (both 27 September 2018), vaccine manufacturer websites, and checked reference lists from an index of HPV studies and other relevant systematic reviews.
We included randomised controlled trials (RCTs) with no language restriction. We considered studies if they enrolled HIV-negative males or females aged 9 to 26 years, or HIV-positive males or females of any age.
We used methods recommended by Cochrane. We use the term 'control' to refer to comparator products containing an adjuvant or active vaccine and 'placebo' to refer to products that contain no adjuvant or active vaccine. Most primary outcomes in this review were clinical outcomes. However, for comparisons comparing dose schedules, the included RCTs were designed to measure antibody responses (i.e. immunogenicity) as the primary outcome, rather than clinical outcomes, since it is unethical to collect cervical samples from girls under 16 years of age. We analysed immunogenicity outcomes (i.e. geometric mean titres) with ratios of means, clinical outcomes (e.g. cancer and intraepithelial neoplasia) with risk ratios or rate ratios and, for serious adverse events and deaths, we calculated odds ratios. We rated the certainty of evidence with GRADE.
We included 20 RCTs with 31,940 participants. The length of follow-up in the included studies ranged from seven months to five years.
Two doses versus three doses of HPV vaccine in 9- to 15-year-old females
Antibody responses after two-dose and three-dose HPV vaccine schedules were similar after up to five years of follow-up (4 RCTs, moderate- to high-certainty evidence). No RCTs collected clinical outcome data. Evidence about serious adverse events in studies comparing dose schedules was of very low-certainty owing to imprecision and indirectness (three doses 35/1159; two doses 36/1158; 4 RCTs). One death was reported in the three-dose group (1/898) and none in the two-dose group (0/899) (low-certainty evidence).
Interval between doses of HPV vaccine in 9- to 14-year-old females and males
Antibody responses were stronger with a longer interval (6 or 12 months) between the first two doses of HPV vaccine than a shorter interval (2 or 6 months) at up to three years of follow-up (4 RCTs, moderate- to high-certainty evidence). No RCTs collected data about clinical outcomes. Evidence about serious adverse events in studies comparing intervals was of very low-certainty, owing to imprecision and indirectness. No deaths were reported in any of the studies (0/1898, 3 RCTs, low-certainty evidence).
HPV vaccination of 10- to 26-year-old males
In one RCT there was moderate-certainty evidence that quadrivalent HPV vaccine, compared with control, reduced the incidence of external genital lesions (control 36 per 3081 person-years; quadrivalent 6 per 3173 person-years; rate ratio 0.16, 95% CI 0.07 to 0.38; 6254 person-years) and anogenital warts (control 28 per 2814 person-years; quadrivalent 3 per 2831 person-years; rate ratio 0.11, 95% CI 0.03 to 0.38; 5645 person-years). The quadrivalent vaccine resulted in more injection-site adverse events, such as pain or redness, than control (537 versus 601 per 1000; risk ratio (RR) 1.12, 95% CI 1.06 to 1.18, 3895 participants, high-certainty evidence). There was very low-certainty evidence from two RCTs about serious adverse events with quadrivalent vaccine (control 12/2588; quadrivalent 8/2574), and about deaths (control 11/2591; quadrivalent 3/2582), owing to imprecision and indirectness.
Nonavalent versus quadrivalent vaccine in 9- to 26-year-old females and males
Three RCTs were included; one in females aged 9- to 15-years (n = 600), one in females aged 16- to 26-years (n = 14,215), and one in males aged 16- to 26-years (n = 500). The RCT in 16- to 26-year-old females reported clinical outcomes. There was little to no difference in the incidence of the combined outcome of high-grade cervical epithelial neoplasia, adenocarcinoma in situ, or cervical cancer between the HPV vaccines (quadrivalent 325/6882, nonavalent 326/6871; OR 1.00, 95% CI 0.85 to 1.16; 13,753 participants; high-certainty evidence). The other two RCTs did not collect data about clinical outcomes. There were slightly more local adverse events with the nonavalent vaccine (905 per 1000) than the quadrivalent vaccine (846 per 1000) (RR 1.07, 95% CI 1.05 to 1.08; 3 RCTs, 15,863 participants; high-certainty evidence). Comparative evidence about serious adverse events in the three RCTs (nonavalent 243/8234, quadrivalent 192/7629; OR 0.60, 95% CI 0.14 to 2.61) was of low certainty, owing to imprecision and indirectness.
HPV vaccination for people living with HIV
Seven RCTs reported on HPV vaccines in people with HIV, with two small trials that collected data about clinical outcomes. Antibody responses were higher following vaccination with either bivalent or quadrivalent HPV vaccine than with control, and these responses could be demonstrated to have been maintained for up to 24 months in children living with HIV (low-certainty evidence). The evidence about clinical outcomes and harms for HPV vaccines in people with HIV is very uncertain (low- to very low-certainty evidence), owing to imprecision and indirectness.