Literature Search and Characteristics of Included Studies
The study selection process is presented in Figure 1. The literature search retrieved 616 references. In addition, we received 12 references from the pharmaceutical companies with marketing authorization for HPV vaccines in Norway. 46 publications, reporting on 13 different RCTs, were selected.
Selection process of randomized control trials.
Characteristics from included studies are presented in Table 1. Totally, the studies included nearly 40 000 participants. They were conducted in North America (USA and Canada), South America, Europe and Asia. Most clinical trials had a maximum follow-up period of 4 years, with two reporting results after a follow-up of 6, and 8 years. The participants were healthy and non-pregnant women aged 15 to 45 years of age. One of the studies included women aged 9 to 23 years, but the mean age was 17 years. The FUTURE (protocol 19) trial included women aged 24 to 45 (mean age 34 years). However, we included this study since one of our inclusion criterion was women aged 16 and older. Some studies included only participants with no history of HPV infection and negative HPV tests at entry into the study, but most studies recruited participants with fewer than four to six lifetime sex partners.
Vaccines used in the trials were the bivalent vaccine containing HPV 16 and 18 virus-like particles (VLP) from GlaxoSmithKline, or the monovalent vaccine containing HPV 16 VLP and the quadrivalent vaccine containing HPV 6, 11, 16 and 18, both from Sanofi Pasteur MSD. All trials used placebo as comparator except for two studies: one used hepatitis B vaccine in both the intervention and the control groups, and one compared the bivalent and the quadrivalent vaccines. All vaccines were given as three doses during a six months period (at day 1, months 2 and 6; or at months 0, 1 and 6).
Some of the included studies had unclear allocation concealment and unclear blinding. However, all studies were assessed as having low risk of bias.
While, one of our aims was to investigate the effect of a HPV catch-up vaccination on overall and cancer related mortality, and on the incidence of cervical cancer, no RCT examining these outcomes were identified.
Effect of HPV Vaccines on Outcomes Identified in Relevant Studies
Table 2 summarises the effects of HPV vaccine versus placebo or no vaccine and the quality of evidence for each outcome.
Overall Mortality. Overall mortality was seldom reported, and primarily only in the text. Overall mortality was reported in 7 RCTs. The authors reported that none of the recorded deaths were considered to be related to the intervention in the vaccine or control groups.
CIN2+. The intention-to-treat (ITT) analysis based on five studies showed a borderline statistically significant reduction in CIN2+ lesions associated with HPV vaccination with a pooled RR of 0.80 (95% CI: 0.62–1.02) for a follow-up period of 4 years (Figure 2). The quality of the evidence was judged moderate. Two of the included studies published results for longer follow-up periods, and reported RRs of 0.29 (0.11–0.78), and 0.64 (0.27–1.52) for follow-up of 6 and 8 years, respectively (results not shown).
Risk of cervical intraepithelial neoplasia grade 2 and higher lesions associated with HPV vaccination.
The reported RR using the per protocol population (PPP) showed a non statistically significant reduction for all CIN2+ lesions after a four year follow-up period (RR: 0.49; 95% CI: 0.21–1.14) (Figure 2). However, this finding was based on only one RCT, and the quality of the evidence was considered low.
When considering only CIN2+ HPV related lesions, we found statistically significant reductions in risk with HPV vaccination both for studies using ITT and PPP populations (Figure 3). The pooled RRs were 0.54 (0.44–0.67) for the ITT population, and 0.05 (0.01–0.16) for the PPP population, both for a 4 years follow-up period. The quality of the evidence was considered high for both estimates. Two studies reported data for 721 participants from the ITT population for 8 years follow up (Figure 3). The pooled RR was 0.29 (0.09–0.96). The quality of the evidence for this outcome was considered moderate.
Risk of HPV related cervical intraepithelial neoplasia grade 2 and higher lesions associated with HPV vaccination.
VIN2+, VaIN2+. We found a statistically significant reduction in risk of all VIN2+ or VaIN2+ lesions with HPV vaccination (RR = 0.49; 95% CI = 0.32–0.76) based on two RCTs reported in one publication (Figure 4). The quality of the evidence for this outcome was considered moderate. However, when considering only published estimates on HPV related VIN2+ or VaIN2+ (from four studies), the reduction in risk was not statistically significant (pooled RR = 0.72; 0.03–15.02) (Figure 4). The quality of the evidence for this outcome was low.
Risk of vulvar and vaginal intraepithelial neoplasia lesions associated with HPV vaccination. Legend: VIN2+: vulvar and vaginal intraepithelial neoplasia grade 2 and higher. VaIN2+: vaginal intraepithelial neoplasia grade 2 and higher
Condyloma Acuminata. HPV vaccination was associated with a reduction in risk of condyloma, both for all condyloma and for those related to HPV types included in HPV vaccines in the ITT population (Figure 5). The reported RR, based on two RCTs, was 0.38 (0.31- 0.47) for all condyloma, and the pooled RR was 0.28 (0.12–0.65) for HPV related condyloma. The quality of the evidence for these two outcomes was high.
Risk of condyloma associated associated with HPV vaccination in girls aged 16 and older.
Serious Adverse Events. We included 14 studies that reported estimates of the association between HPV vaccination and serious adverse events. The risk of having a serious adverse event was similar in both the vaccine and control groups (Figure 6). The pooled RR was 0.99 (0.91–1.08). The quality of the evidence for this outcome was moderate.
Risk of serious adverse events associated with HPV vaccination in girls aged 16 and older.
We identified one publication that compared the bivalent vaccine with the quadrivalent vaccine, and examined possible differences safety between these two vaccines. However, the quality of the evidence was judged low, and no statistically significant difference was found (RR = 1.05; 95% CI: 0.59–1.05) (results not shown).