Abstract
Background
Infection with human papillomavirus (HPV) 16 or HPV18 elicits an antibody response, but whether the elicited antibodies protect women against subsequent infection by a homologous HPV type compared with seronegative women is unknown.
Methods
Study participants were women aged 18–25 years at enrollment in the control group of the ongoing National Cancer Institute–sponsored, community-based, randomized HPV16/18 Costa Rica Vaccine Trial. At enrollment, 2813 participants were negative for cervical HPV16 DNA and 2950 for HPV18 DNA. Women were interviewed regarding sociodemographic data and medical and health history. Medical and pelvic examinations were conducted for all consenting sexually experienced women. Serum samples taken at enrollment were tested for total HPV16/18 antibodies with a polyclonal enzyme-linked immunosorbent assay, and cervical specimens were tested for type-specific HPV DNA over 4 years of follow-up. Using Poisson regression, we compared rate ratios of newly detected cervical HPV16 or HPV18 infection among homologous HPV-seropositive and HPV-seronegative women, adjusting for age, education, marital status, lifetime number of sexual partners, and smoking.
Results
There were 231 newly detected HPV16 infections during 5886 person-years among HPV16-seronegative women compared with 12 newly detected HPV16 infections during 581 person-years among HPV16-seropositive women with the highest HPV16 sero-levels. There were 136 newly detected HPV18 infections during 6352 person-years among HPV18-seronegative women compared with six new infections detected during 675 person-years among HPV18 seropositives with the highest sero-levels. After controlling for risk factors associated with newly detected HPV infection, having high HPV16 antibody titer at enrollment was associated with a reduced risk of subsequent HPV16 infection (women in the highest tertile of HPV16 antibody titers, adjusted rate ratio = 0.50, 95% confidence interval = 0.26 to 0.86 vs HPV16-seronegative women). Similarly, having high HPV18 antibody titer at enrollment was associated with a reduced risk of subsequent HPV18 infection (women in the highest tertile of HPV18 antibody titers, adjusted rate ratio = 0.36, 95% confidence interval = 0.14 to 0.76 vs HPV18-seronegative women).
Conclusion
In this study population, having high antibody levels against HPV16 and HPV18 following natural infection was associated with reduced risk of subsequent HPV16 and HPV18 infections.
CONTEXTS AND CAVEATS
Prior knowledge
Immunization with human papillomavirus (HPV) virus-like particles has been shown to be highly effective in preventing cervical HPV16/18 infections and their associated lesions. Natural infections with HPV also elicit an antibody response, but the degree of protection against subsequent HPV infections is unknown.
Study design
Serum samples were collected from women aged 18–25 years in the control group of the HPV16/18 Costa Rica Vaccine Trial, along with demographic and medical data. Serum samples taken at enrollment were tested for total HPV16/18 antibodies, and cervical specimens were tested for HPV DNA over 4 years of follow-up.
Contribution
High HPV16 and -18 antibody titers at enrollment were associated with a reduced risk of subsequent HPV16 and -18 infection for women in the highest tertiles of HPV16 and -18 antibody titers.
Implications
High antibody levels against HPV16 and HPV18 following natural infection offer some protection against subsequent infections.
Limitations
The effect of HPV antibody titers among women with incident persistent HPV infection could not be evaluated because the trial was still ongoing. The duration of protection after natural infection and whether the protective mechanism can be attributed to neutralizing antibodies or to some other immune protective mechanism are still unknown.
From the Editors
Passive transfer of immune sera or purified IgG induced by immunization with animal (canine and cottontail rabbit) papillomavirus virus-like particles (VLPs) have shown protection from subsequent canine and cottontail rabbit papillomavirus challenge (1,2). Immunization with the human papillomavirus (HPV) capsid (L1) VLPs generates high levels of polyclonal antibodies against conformational epitopes on the viral capsid (3), which have been shown to be highly effective in preventing cervical HPV16/18 infections and their associated lesions (4–9). Although the effector mechanism of protection in humans is not completely understood, it is believed that the subset of antibodies induced by vaccination capable of neutralizing virions is responsible for the high degree of protection observed with this vaccine (10). Antibodies against these and other relevant epitopes are probably induced following natural HPV infection, albeit at lower levels than those observed following vaccination. Moreover, natural history studies have shown that incidence and/or prevalence of newly detected HPV infections decline with age in many populations (11). Although part of this decline likely reflects reduced exposure to HPV secondary to fewer sexual partners at older ages, the cumulative effect of naturally acquired immunity may also be contributory. It is therefore reasonable to hypothesize that antibodies that are elicited by natural infection might protect the host against HPV reinfection with the same or related HPV type.
A few studies have examined this protection hypothesis (12–14). One study among young university students showed approximately 50% decreased risk of subsequent cervical infection with HPV16 and its genetically related types among women with persistent IgG and IgA antibodies to HPV16 (12). A second study conducted within a population-based natural history cohort in Guanacaste, Costa Rica, showed a 30% decreased risk of subsequent cervical infection with HPV16 among women who were seropositive for HPV16 VLP antibodies, although it was not statistically significant (13). A third study conducted among HIV-positive and HIV-negative women found no statistically significant evidence that presence of antibodies to HPV16 VLPs reduced rates of cervical HPV16 reinfection (14)
Our objective was to evaluate whether antibodies generated following natural infection with HPV16 and HPV18 are associated with reduced risk of detection of the same or related HPV types in the cervix. We used data from the control group of the ongoing, community-based, National Cancer Institute–sponsored Costa Rica HPV16/18 Vaccine Trial of 7466 young women. Serum samples taken at enrollment were tested for total (polyclonal) HPV16/18 antibodies by enzyme-linked immunosorbent assay (ELISA), and cervical specimens taken at enrollment and follow-up were tested for type-specific HPV DNA.
Methods
We analyzed data from the Costa Rica HPV16/18 Vaccine Trial (15), a publicly funded randomized trial of the efficacy of the HPV16/18 vaccine manufactured by GlaxoSmithKline for the prevention of HPV16/18 infection and related precancerous lesions (defined for the trial as cervical intraepithelial neoplasia 2 or 3 [CIN2+] or adenocarcinoma in situ). Enrollment took place between June 2004 and December 2005, and the 4-year follow-up is nearly complete. Participants are women in good general health between 18 and 25 years of age at enrollment from Guanacaste and nearby areas of Puntarenas, Costa Rica. Eligibility criteria included a negative history of chronic conditions requiring treatment, willingness to use birth control during the vaccination period, and stable residence in the study area. Approximately one-third of the women identified in a previous census fulfilled the inclusion criteria and participated in the study.
At enrollment, women provided written informed consent and underwent a urine pregnancy test. Eligible women were randomly assigned to receive either the study vaccine against HPV16/18 or the control vaccine against hepatitis A, administered at enrollment, 1-month, and 6-month visits. At enrollment, women were interviewed regarding demographics, sexual activity, contraceptive use, reproductive history, cigarette use, and family history of cancers. A detailed medical questionnaire was also administered, and medical and pelvic examinations were conducted for all consenting sexually experienced women. During the pelvic examination, cervical cells were collected and placed in liquid cytology medium (PreservCyt; Cytyc Corporation, Marlborough, MA) for liquid-based cytology (ThinPrep; Cytyc Corporation) and for cervical HPV detection. Aliquots destined for polymerase chain reaction (PCR) were stored in liquid nitrogen, whereas the remaining PreservCyt samples were kept at room temperature (∼20°C) until they were used to make liquid cytology slides and tested for HPV, Chlamydia trachomatis, and Neisseria gonorrhoeae. All testing was done with the investigators masked to the results of other tests or cytology results from the same individual.
After vaccination, women with negative screening results returned for follow-up annually. Women with low-grade squamous intraepithelial lesions or who were carcinogenic HPV positive with atypical squamous cells of undetermined significance were asked to return every 6 months until their cervical cytology tested as normal three consecutive times. At every follow-up visit, cervical samples were collected from sexually experienced (nonvirgin) women for HPV DNA testing. All study protocols were reviewed and approved by the National Cancer Institute and Costa Rican Institutional Review Boards.
HPV Assessment
Serum collected at enrollment was used to determine HPV16 and HPV18 serological status using a VLP-based direct ELISA, a standard measure of immunogenicity that measures polyclonal antibodies, performed at GlaxoSmithKline Biologicals in Rixensart, Belgium, as described previously (16,17). Briefly, testing was performed using ELISA microtiter plates coated separately with 2.7 μg/mL of either HPV16 or HPV18 VLPs, which were produced in a baculovirus expression system. After incubation and washing steps, plates were incubated further to block nonspecific antibodies. Next, they were washed and incubated with serum from participants, positive and negative quality controls, and standards, serially diluted starting at 1:100 in twofold increments. After the washing steps, a peroxidase-conjugated anti-human polyclonal antibody was added to react with the specific antibody. Excess conjugate was removed by washing, following which an enzyme substrate and chromogen were added and the color allowed to develop. Reactions were stopped, and optical density read at 450 and 620 nm. The optical density at 620 nm is read to measure any background signal, which is then subtracted from the optical density at 450 nm. ELISA titers were calculated from a reference curve determined from a reference pool of serum samples from human vaccines using a four-parameter logistic equation from SoftMax Pro (Molecular Devices, Sunnyvale, CA) (16,17). ELISA titers of VLP-specific polyclonal antibodies were calculated by averaging the values from all dilutions that fell within the working range of the reference curve relative to the optical density readings and expressed as ELISA units (EU)/mL. Antibody results were dichotomized using standard cutoff points calculated from antibody titer values 3 SDs above the geometric mean titers taken from a group of HPV-negative individuals (16). Cut points were optical density of at least 8 EU/mL for anti-HPV16 and at least 7 EU/mL for anti-HPV18 (16,17).
Total DNA was isolated from 200 μL of a 1 mL PreservCyt aliquot drawn before ThinPrep preparation by using a MagNA Pure LC instrument (Roche Diagnostics, Almere, the Netherlands) and a Total DNA isolation kit (Roche Diagnostics). DNA was eluted in 100 μL of water.
A 10-μL aliquot of extracted DNA was used for each short PCR fragment 10 (SPF10) PCR. The SPF10 PCR primer set was used to amplify at least 54 HPV genotypes, as described earlier (18,19). Briefly, this primer set amplifies a small fragment of 65 base pairs from the L1 region of HPV. Reverse primers contain a biotin label at the 5′ end, enabling capture of the reverse strand onto streptavidin-coated microtiter plates. The captured amplimers are denatured by alkaline treatment, and the captured strand is detected by a defined cocktail of digoxigenin-labeled probes that detect a broad spectrum of HPV genotypes. This detection method is designated the HPV DNA enzyme immunoassay (DEIA), which provides an optical density value. If the SPF10-DEIA yielded a borderline value (75%–100% of the cutoff value), the PCR was repeated and samples retested by DEIA.
The same SPF10 amplimers (from SPF10-DEIA-positive samples) were used to identify the HPV genotype by reverse hybridization on line probe assay (LiPA) containing probes for 25 HPV genotypes (SPF10 HPV LiPA25, version 1; Labo Bio-Medical Products, Rijswijk, the Netherlands). The LiPA25 detects HPV types 6, 11, 16, 18, 31, 33, 34, 35, 39, 40, 42, 43, 44, 45, 51, 52, 53, 54, 56, 58, 59, 66, 68/73, 70, and 74.
Because the Costa Rica HPV16/18 Vaccine Trial is focused on HPV16 and HPV18, to maximize sensitivity for these types, type-specific PCR primer sets were used to selectively amplify HPV16 and HPV18 from 2513 specimens that tested positive by the SPF10-DEIA PCR but that did not contain HPV16 or HPV18, as determined by LiPA25 (20). Amplimers from the type-specific PCRs were detected by DEIA, similar to the method used for SPF10 amplimer detection. Each run (from DNA isolation to type distinction) was accompanied by parallel positive and negative controls to monitor each step of the procedure, ensuring that DNA isolation, PCR, and post-PCR analyses were individually checked for positivity and negativity. Contamination rates were actively measured and recorded. Over the past 9 years (2001–2010), contamination incidence has been well below 1%.
Statistical Methods
Analyses for HPV16 and HPV18 infections were performed separately. Women in the control group of the trial who were HPV16 and/or HPV18 DNA negative at enrollment with at least 6 months of follow-up were included in the analysis. Participants who were referred to colposcopy at enrollment were excluded to avoid confounding by management/treatment.
The outcome of interest was cervical HPV16 or HPV18 DNA infection newly detected (referred to as “incident” infection) during follow-up. The main exposure of interest was enrollment HPV16 serostatus for the HPV16 analysis and enrollment HPV18 serostatus for the HPV18 analysis. We categorized the serotiters according to tertiles, which we had established a priori. Thus, we compared the cumulative incidence rate of incident HPV infection among seronegative participants vs seropositive participants in the lowest, medium, and highest serotiter groups.
Incidence rates and rate ratios of newly detected HPV infection were calculated using person-time methods for respective HPV-seropositive relative to HPV-seronegative women (separately for HPV16 and HPV18). Virgins contributed to the analysis at first report of sexual activity. Because we were not able to identify the exact timing of the incident infection, we assumed the midpoint between last negative detection and first positive detection as the time of infection. We used Poisson regression to evaluate the effect of various exposure variables measured at enrollment on the risk of subsequent cervical HPV DNA detection. Behavioral data (age [used continuously], education [categorized as <7, 7–9, >10 years, and university], marital status [categorized as married, single, divorced, or widowed], number of lifetime partners [categorized as 1, 2, 3, or >4], oral contraceptive use [categorized as never used, used in the past, and current user], smoking cigarettes [categorized as never smoked, smoked in the past, and currently smoking], cervical Chlamydia infection [determined from cervical samples and categorized as infected or not infected], history of pregnancy [never pregnant vs ever pregnant], gravidity [categorized as 0, 1, 2, 3, and 4 or more], and lastly years since first sex [determined by subtracting current age from age of first sex reported and categorized as within a year, 2, 3, 4, or 5 and more years]) collected at the enrollment visit were used as covariates. To determine the independent effect of serological status and other covariates of interest mentioned above on the risk of newly detected HPV infection, multivariable Poisson models were generated. The relative contributions of each exposure variable, adjusting for the simultaneous effects of covariates statistically significant at P < .05 for either the HPV16 or HPV18 analyses in the final models, were expressed as adjusted rate ratios (ARRs). To ensure that our results were not attributable to false-negative cervical HPV DNA test results at enrollment, we also performed multivariable models that restricted the analytic sample to those with two consecutive negative cervical HPV DNA tests for HPV16 or HPV18, as appropriate.
Results
For the HPV16 analysis, of the 3736 women in the control group of the HPV16/18 Costa Rica Vaccine Trial, we hierarchically excluded 265 (7.1%) women who were cervical HPV16 DNA positive at enrollment, 28 (0.8%) women who were referred to colposcopy, 68 (1.8%) women for whom enrollment HPV16 serology was missing, and 234 (6.3%) women with no follow-up visits (Figure 1, A). For the HPV18 analysis, we hierarchically excluded 93 (2.5%) women who were cervical HPV18 DNA positive at enrollment, 40 (1.1%) women who were referred to colposcopy, 83 (2.2%) women for whom enrollment HPV18 serology was missing, and 242 (6.5%) women with no follow-up visits (Figure 1, B). For both the HPV16 and HPV18 analyses, we excluded 327 (8.8%) virgins who did not initiate sexual activity during the follow-up and one 17 years old. Women excluded were younger and less likely to have higher numbers of lifetime partners compared with women in the analytic sample, mainly because of the exclusion of women who remained virgins.
Figure 1.
CONSORT diagram of the study population and analysis for human papillomavirus (HPV) 16 (A) and HPV18 (B). PY = person-years.
Of the 2813 women eligible for the HPV16 analysis, 699 (24.8%) were HPV16 seropositive at enrollment (Figure 1, A). Similarly, of the 2950 women eligible for the HPV18 analysis, 731 (24.8%) were HPV18 seropositive at enrollment (Figure 1, B). For both HPV16 and HPV18 analyses, the median age was 21 years (interquartile range [IQR], 19–23 years), median number of visits per person was 4, and median interval between study visits was 12 months (IQR, 10–12 months). Among the HPV16-seropositive women, median HPV16 titers were 28 EU/mL (IQR 14–71, range 8–3202); among HPV18-seropositive women, median HPV18 titers were 16 EU/mL (IQR 10–35, range 7–2541). However, the assays for the two HPV types are not identical and titers could not be directly compared.
A total of 291 women acquired a new HPV16 infection over 7843 person-years of observation, a crude incidence rate of 3.7 per 100 person-years (95% confidence interval [CI] = 3.3 to 4.2). A total of 178 women acquired a new HPV18 infection over 8407 person-years, a crude incidence rate of 2.1 per 100 person-years (95% CI = 1.8 to 2.4).
At enrollment, among HPV16-seropositive women, median IgG levels among those with incident HPV16 (n = 60) was 26 EU/mL (IQR 13–49) compared with 29 EU/mL among those with no infection (n = 639) (IQR 14–73, P = .5, Wilcoxon rank sum). Among HPV18-seropositive women, enrollment median IgG levels among those with incident HPV18 (n = 42) was 12 EU/mL (IQR 8–21) compared with 26 EU/mL among those with no infection (n = 689) (IQR 10–37, P = .05, Wilcoxon rank sum).
Overall, there were 231 HPV16 infections newly detected during 5886 person-years among HPV16-seronegative women (crude incidence rate of 3.92 per 100 person-years) compared with 12 HPV16 infections during 581 person-years incidentally detected (crude incidence rate of 2.06 per 100 person-years) among HPV16-seropositive women with the highest HPV16 serotiters. One hundred and thirty-six HPV18 infections were detected during 6352 person-years among HPV18-seronegative women (crude incidence rate of 2.14 per 100 person-years) compared with six HPV18 infections during 675 person-years (crude incidence rate of 0.89 per 100 person-years) among HPV18-seropositive women with the highest HPV18 serotiter levels. Incident HPV16 and HPV18 infections declined with increasing age but were higher among women with a greater number of lifetime partners. HPV16 and HPV18 infections were also increased among single, divorced, or widowed participants compared with married women and among past or current smokers compared with nonsmokers (Tables 1 and 2). In multivariable analyses, performed for HPV16 and HPV18 separately, several factors remained statistically significantly associated with incidentally detected HPV16 and HPV18 infection, including decreasing incidence with increasing age and increased incidence in single, divorced, or widowed participants and those with more lifetime partners.
Table 1.
Univariate and multivariable risk factors associated with incidently detected human papillomavirus 16 (HPV16) infection*
| Risk factor | PY | No. of new HPV16 infections | Incidence per 100/PY | Univariate, RR (95% CI) | Multivariable, RR (95% CI)† |
| Age at enrollment, y | |||||
| 18 | 1149 | 53 | 4.61 | 1 | 0.95 (0.89 to 1.00)‡ |
| 19 | 1189 | 47 | 3.95 | 0.86 (0.58 to 1.27) | |
| 20 | 988 | 46 | 4.66 | 1.01 (0.68 to 1.50) | |
| 21 | 864 | 28 | 3.24 | 0.70 (0.44 to 1.10) | |
| 22 | 1017 | 40 | 3.93 | 0.85 (0.56 to 1.28) | |
| 23 | 903 | 29 | 3.21 | 0.70 (0.44 to 1.09) | |
| 24 | 908 | 28 | 3.08 | 0.67 (0.42 to 1.05) | |
| 25 | 826 | 20 | 2.42 | 0.52 (0.31 to 0.86) | |
| Education, y | |||||
| <7 | 2367 | 71 | 3.00 | 1 | 1 |
| 7–9 | 1785 | 65 | 3.64 | 1.21 (0.87 to 1.70) | 1.08 (0.76 to 1.52) |
| ≥10 (technical) | 2498 | 99 | 3.96 | 1.32 (0.98 to 1.80) | 1.08 (0.78 to 1.50) |
| University | 1158 | 55 | 4.75 | 1.58 (1.11 to 2.25) | 1.29 (0.88 to 1.88) |
| Marital status | |||||
| Married | 3775 | 95 | 2.52 | 1 | 1 |
| Single | 3839 | 181 | 4.72 | 1.87 (1.47 to 2.41) | 1.64 (1.24 to 2.17) |
| Divorced/widowed | 221 | 14 | 6.34 | 2.52 (1.38 to 4.27) | 2.40 (1.31 to 4.11) |
| No. of lifetime partners | |||||
| 1 | 3879 | 122 | 3.15 | 1 | 1 |
| 2 | 1683 | 72 | 4.28 | 1.36 (1.01 to 1.82) | 1.42 (1.05 to 1.92) |
| 3 | 949 | 40 | 4.21 | 1.34 (0.93 to 1.90) | 1.36 (0.93 to 1.96) |
| ≥4 | 989 | 44 | 4.45 | 1.41 (0.99 to 1.98) | 1.44 (0.98 to 2.10) |
| Oral contraceptive use | |||||
| Never | 2593 | 111 | 4.28 | 1 | |
| Past | 1455 | 51 | 3.51 | 0.82 (0.58 to 1.13) | |
| Current (past month) | 3780 | 129 | 3.41 | 0.80 (0.62 to 1.03) | |
| Smoking | |||||
| Never | 6809 | 229 | 3.36 | 1 | 1 |
| Past | 457 | 26 | 5.69 | 1.69 (1.10 to 2.49) | 1.54 (0.99 to 2.30) |
| Current | 568 | 35 | 6.17 | 1.83 (1.26 to 2.58) | 1.42 (0.94 to 2.07) |
| Chlamydia infection | |||||
| No | 5712 | 208 | 3.64 | 1 | |
| Yes | 925 | 44 | 4.76 | 1.31 (0.93 to 1.79) | |
| Pregnancy | |||||
| Never | 3677 | 153 | 4.16 | 1 | |
| Ever | 4166 | 138 | 3.31 | 0.80 (0.63 to 1.00) | |
| Gravidity | |||||
| 0 | 3677 | 153 | 4.16 | 1 | |
| 1 | 2622 | 87 | 3.32 | 0.80 (0.61 to 1.04) | |
| 2 | 1109 | 36 | 3.25 | 0.78 (0.54 to 1.11) | |
| 3 | 324 | 13 | 4.01 | 0.96 (0.52 to 1.63) | |
| ≥4 | 111 | 2 | 1.80 | 0.43 (0.07 to 1.35) | |
| Years since first sex | |||||
| 0/1 | 754 | 37 | 4.91 | 1 | |
| 2 | 676 | 36 | 5.32 | 1.08 (0.68 to 1.72) | |
| 3 | 848 | 45 | 5.31 | 1.08 (0.70 to 1.68) | |
| 4 | 947 | 37 | 3.91 | 0.80 (0.50 to 1.26) | |
| ≥5 | 3423 | 96 | 2.80 | 0.57 (0.40 to 0.85) | |
Factors evaluated at enrollment. We used Poisson regression to compare enrollment risk factors for detection of a new HPV16 over the follow-up. CI = confidence interval; PY = person-years; RR = rate ratio.
Model adjusted for all the other variables.
Age entered as continuous variable in the multivariable model.
Table 2.
Univariate and multivariable risk factors associated with incidently detected human papillomavirus 18 (HPV18) infection*
| Risk factor | PY | No. of new HPV18 infections | Incidence per 100 PY | Univariate, RR (95% CI) | Multivariable, RR (95% CI)† |
| Age at enrollment, y | |||||
| 18 | 1271 | 32 | 2.52 | 1 | 0.95 (0.88 to 1.02)‡ |
| 19 | 1271 | 32 | 2.52 | 1.00 (0.61 to 1.64) | |
| 20 | 1082 | 24 | 2.22 | 0.88 (0.51 to 1.49) | |
| 21 | 906 | 23 | 2.54 | 1.01 (0.58 to 1.72) | |
| 22 | 1110 | 21 | 1.89 | 0.75 (0.43 to 1.29) | |
| 23 | 937 | 16 | 1.71 | 0.68 (0.36 to 1.22) | |
| 24 | 967 | 18 | 1.86 | 0.74 (0.41 to 1.30) | |
| 25 | 864 | 12 | 1.39 | 0.55 (0.27 to 1.04) | |
| Education, y | |||||
| <7 | 2540 | 55 | 2.17 | 1 | 1 |
| 7–9 | 1890 | 41 | 2.17 | 1.00 (0.67 to 1.50) | 0.88 (0.58 to 1.33) |
| ≥10 (technical) | 2701 | 64 | 2.37 | 1.09 (0.76 to 1.57) | 0.89 (0.61 to 1.31) |
| University | 1253 | 17 | 1.36 | 0.63 (0.35 to 1.06) | 0.53 (0.30 to 0.92) |
| Marital status | |||||
| Married | 3997 | 56 | 1.40 | 1 | 1 |
| Single | 4180 | 110 | 2.63 | 1.88 (1.37 to 2.61) | 1.95 (1.37 to 2.79) |
| Divorced/widowed | 223 | 12 | 5.37 | 3.83 (1.96 to 6.90) | 3.60 (1.83 to 6.54) |
| No. of lifetime partners | |||||
| 1 | 4049 | 65 | 1.61 | 1 | 1 |
| 2 | 1854 | 47 | 2.54 | 1.58 (1.08 to 2.29) | 1.59 (1.08 to 2.34) |
| 3 | 1041 | 28 | 2.69 | 1.68 (1.06 to 2.58) | 1.67 (1.04 to 2.62) |
| ≥4 | 1110 | 34 | 3.06 | 1.91 (1.25 to 2.87) | 1.83 (1.16 to 2.86) |
| Oral contraceptive use | |||||
| Never | 2751 | 66 | 2.40 | 1 | |
| Past | 1598 | 31 | 1.94 | 0.81 (0.52 to 1.23) | |
| Current (past month) | 4041 | 81 | 2.00 | 0.84 (0.60 to 1.16) | |
| Smoking | |||||
| Never | 7263 | 138 | 1.90 | 1 | 1 |
| Past | 518 | 21 | 4.06 | 2.13 (1.31 to 3.30) | 1.81 (1.10 to 2.84) |
| Current | 619 | 19 | 3.07 | 1.61 (0.97 to 2.54) | 1.14 (0.65 to 1.87) |
| Chlamydia infection | |||||
| No | 6128 | 129 | 2.11 | 1 | |
| Yes | 1050 | 32 | 3.05 | 1.45 (0.97 to 2.10) | |
| Pregnancy | |||||
| Never | 3889 | 91 | 2.34 | 1 | |
| Ever | 4519 | 87 | 1.93 | 0.82 (0.61 to 1.10) | |
| Gravidity | |||||
| 0 | 3889 | 91 | 2.34 | 1 | |
| 1 | 2849 | 53 | 1.86 | 0.79 (0.56 to 1.11) | |
| 2 | 1186 | 24 | 2.02 | 0.86 (0.54 to 1.33) | |
| 3 | 367 | 8 | 2.18 | 0.93 (0.42 to 1.80) | |
| ≥4 | 117 | 2 | 1.72 | 0.73 (0.12 to 2.32) | |
| Years since first sex | |||||
| 0/1 | 810 | 20 | 2.47 | 1 | |
| 2 | 744 | 27 | 3.63 | 1.47 (0.83 to 2.65) | |
| 3 | 950 | 18 | 1.89 | 0.77 (0.40 to 1.45) | |
| 4 | 1019 | 25 | 2.45 | 0.99 (0.55 to 1.81) | |
| ≥5 | 3670 | 71 | 1.93 | 0.78 (0.49 to 1.32) | |
Factors evaluated at enrollment. CI = confidence interval; PY = person-years; RR = rate ratio.
Model adjusted for all the other variables.
Age entered as continuous variable in the multivariable model.
In univariate models, newly detected HPV16 and HPV18 infections were statistically significantly lower in participants with the highest titers compared with seronegative women (Table 3). In multivariable models, after adjustment for the confounders identified in Tables 1 and 2, the decreased incidence among those with the highest HPV16 and HPV18 titers remained statistically significant (highest HPV16 tertile: ARR = 0.50, 95% CI = 0.26 to 0.86 compared with HPV16-seronegative women; highest HPV18 tertile: ARR = 0.36, 95% CI = 0.14 to 0.76 compared with HPV18-seronegative women).
Table 3.
Univariate and multivariable association of enrollment HPV16 and HPV18 antibody with incidently detected HPV16 and HPV18 infection*
| Antibody titers | PY | No. of new HPV16/18 infections | Incidence per 100 PY | Univariate, RR (95% CI) | Multivariable,RR (95% CI)† |
| Enrollment HPV16 titers | |||||
| Negative | 5886 | 231 | 3.92 | 1 | 1 |
| 0–33rd ptl (8–16 EU/mL) | 627 | 20 | 3.19 | 0.81 (0.50 to 1.25) | 0.79 (0.48 to 1.22) |
| 33–66th ptl (17–59 EU/mL) | 749 | 28 | 3.74 | 0.95 (0.63 to 1.38) | 0.83 (0.53 to 1.25) |
| ≥66th ptl (≥60 EU/mL) | 581 | 12 | 2.06 | 0.53 (0.28 to 0.90) | 0.50 (0.26 to 0.86) |
| Enrollment HPV18 titers | |||||
| Negative | 6352 | 136 | 2.14 | 1 | 1 |
| 0–33rd ptl (7–10 EU/mL) | 620 | 19 | 3.06 | 1.43 (0.86 to 2.25) | 1.36 (0.81 to 2.15) |
| 33–66th ptl (11–27 EU/mL) | 760 | 17 | 2.24 | 1.04 (0.61 to 1.68) | 1.01 (0.59 to 1.65) |
| ≥66th ptl (≥28 EU/mL) | 675 | 6 | 0.89 | 0.42 (0.16 to 0.86) | 0.36 (0.14 to 0.76) |
CI = confidence interval; EU = ELISA unit; ptl = percentile; PY = person years; RR = rate ratio.
Models adjusted for age (continuous), education, marital status, lifetime number of partners, and smoking status.
To ensure that our results were not attributable to false-negative cervical HPV DNA test results at enrollment, we performed the multivariable models that restricted the analytic sample to those with two consecutive negative cervical HPV DNA tests for HPV16 or HPV18, as appropriate. Independent predictors of new HPV infection were similar to those of the overall analysis, and the magnitude of the observed associations remained similar for the respective HPV titers (highest HPV16 tertiles: ARR = 0.40, 95% CI = 0.17 to 0.80 compared with HPV16-seronegative women; highest HPV18 tertiles: ARR = 0.17, 95% CI = 0.03 to 0.53 compared with HPV18-seronegative women) (data not shown).
We also evaluated the association between enrollment HPV16 titers and incident HPV31, a closely related HPV genotype for which some cross-protection has been reported following VLP vaccination (4,7,21–23). We did not observe any association between enrollment HPV16 titers and incident HPV31 infection (ARR = 0.96, 95% CI = 0.55 to 1.58 in the highest titer group, compared with HPV16-seronegative women). Similarly, we evaluated the association between enrollment HPV18 titers and incident HPV45 and, despite some cross-protection seen after vaccination (4,7,21–23), did not observe any association between enrollment HPV18 titers and subsequent HPV45 infection (ARR = 1.16, 95% CI = 0.64 to 1.94 in the highest titer group compared with HPV18-seronegative women).
Discussion
The major finding of this study was that unvaccinated women with higher antibody titers of HPV16 and HPV18 were at statistically significantly reduced risk of subsequent new HPV16 (50% reduction) and HPV18 (64% reduction) infection relative to seronegative women. The results we observed were independent of sexual behavior or other factors related to risk of cervical HPV DNA infection. Although the underlying mechanisms that explain the protection we observed cannot be defined with certainty based on the results from this study, two broad possibilities exist. First, women with higher levels of antibodies against HPV16 or HPV18 produce neutralizing antibodies that are capable of preventing new HPV infections from occurring. Alternatively, the antibody test used in our study might serve as a marker for other, associated, immunological responses (such as cell-mediated T-cell responses) that might be responsible for protection via rapid clearance of new infections. To the extent that neutralizing antibodies are the mechanism of protection we observed among naturally infected individuals, our findings suggest that the levels of antibodies required for protection are modest and at least an order of magnitude lower than plateau levels of antibodies seen among vaccinated women even 5 years after vaccination (23), suggesting, in turn, long-term durability of protection afforded by vaccination with the current VLP-based vaccines.
The fact that we observed partial protection may be a result of using the polyclonal ELISA assay, which measures total antibodies, not the specific neutralizing epitopes, although earlier studies among vaccinated individuals showed strong association between the total ELISA used in this study and the secreted alkaline phosphatase neutralization assay, which measures total neutralizing activity (24). The HPV serology assays currently used are heterogeneous, some measuring total antibodies, whereas others measure specific neutralizing antibodies. Whether assays that measure specific neutralizing epitopes better measure immunity needs to be assessed.
Our results are in accord with a study of 608 female university students, which found that sustained detection of IgG and IgA antibodies to HPV16 VLPs was associated with approximately 50% decreased risk of subsequent HPV16 infection (12). In another study conducted within a population-based natural history cohort in Costa Rica, a non-statistically significant 30% decreased risk of subsequent infection was initially observed among women who were seropositive for HPV16 VLP antibodies (13). A recent reanalysis of those data redefining exposure and outcome more stringently (M. Schiffman and N. Wentzensen, unpublished observation) showed evidence for a 50% decreased risk of reinfection among seropositive women, although this effect was not statistically significant because of the reduced sample size in the more stringent analysis. A third study of 1242 women (829 HIV positive and 413 HIV negative) (14) did not show statistically significant evidence that presence of antibodies to HPV16 VLPs reduced rates of reinfection among both HIV-negative and HIV-positive women. The lack of consistency observed in some of the previous studies might be partly explained by the fact that levels of antibodies generated following natural infection are many orders of magnitude lower than those generated through vaccination, resulting in levels of protection that are partial, at best, and therefore difficult to detect reliably. In addition, the previous studies compared women who were seropositive for the specific HPV antibody with women without the antibody and thus did not categorize antibodies among the seropositive women into finer categories, as we have done in this analysis. Indeed, when we examined our data dichotomizing the antibody levels as seropositive vs seronegative, we observed that HPV16-seropositive women had a rate ratio of 0.79 (95% CI = 0.59 to 1.03) compared with HPV16-seronegative women. Similarly, HPV18-seropositive women had a rate ratio of 0.94 (95% CI = 0.69 to 1.36) compared with HPV18-seronegative women. Thus, our analysis indicates that the protection conferred is among those with higher antibody levels. This lack of consistency among studies may be more pronounced because of the fact that the ELISA-based antibody detection assays used in the pre-vaccination era were more prone to misclassification (25) than currently available assays because they used VLPs produced in a less standardized manner and less pure than those currently available. Both studies that found a protective effect of antibody levels and subsequent infections did so by using a more stringent and specific definition of seropositivity; Ho et al. (12) defined seropositivity at two consecutive visits, and Schiffman’s (unpublished) work, based on assays from two different laboratories, considered as positive only those samples that were determined to be positive by both laboratories.
Results from the Papilloma Trial Against Cancer In Young Adults (4) also indicated that protection against lesions may be conferred following natural infection. Using data from the supplemental tables of that study (4), we calculated the relative risk of CIN2+ as 0.56 during average follow-up of 34.9 months among the participants in the control group of the study who were HPV16/18 DNA negative but seropositive compared with 1.13 among the control subjects who were HPV16/18 DNA negative and seronegative, a 50% reduced risk (4). The lower rates of CIN2+ among women with antibodies in this study may also be an indication of conferred protection against lesions following natural infection.
Ho et al. (12) also found that persistent IgG levels were associated with reduced risk for subsequent infection with HPV types that are genetically closely related to HPV16 (ie, HPV31, -33, -35, -52, and -58) (12). However, we did not find any association between enrollment HPV16 titers and incident HPV31 infection or between HPV18 titers and incident HPV45 infections. This lack of association between antibody levels and protection to related HPV types in our study may be related to low antibody levels in natural infection that do not provide sufficient cross-neutralization potential to be protective, to limited statistical power to detect associations, or to differences in the study populations.
Not surprisingly, our data indicated several other variables that were associated with incident HPV infection. Incident infections decreased with increasing age. Women with more lifetime partners were at increased risk of incident HPV infection, as were single, widowed or divorced women. These findings are in accord with other studies investigating factors associated with incident HPV infection (12,26).
Our population was the control group of the community-based randomized trial, which included women selected from a population census. To ensure that our findings were not explained by confounding sexual behaviors for HPV acquisition, we adjusted for sexual risk factors at enrollment in the models. We also examined sexual behaviors of women during follow-up and found no differences in number of recent sexual partners (before detection of the new infection) between seropositive and seronegative women (data not shown). Finally, it is reassuring to note that any residual confounding by sexual behavior is likely to have biased our results toward the null because seropositive women are more likely to have had more lifetime partners, and hence would be more likely to be exposed than seronegative women.
Our findings of a potential protective effect of higher antibody titers on incident HPV infection remained consistent in the analysis restricted to participants with two consecutive negative cervical HPV DNA results and reduced the possibility that our findings are related to false-negative HPV PCR tests for HPV16 or HPV18 at enrollment. A limitation of our analysis at this time is our inability to evaluate the effect of HPV antibody titers among women with incident persistent HPV infection (women positive for cervical HPV16/18 DNA at two consecutive visits) and CIN2+ because our trial is still ongoing. Future studies evaluating the role of antibodies to natural infection and subsequent newly detected cervical HPV DNA infection should consider incident persistent infection at two consecutive positive visits.
In summary, in this cohort of more than 2500 sexually active women, higher HPV16 and HPV18 antibody levels following natural infection were associated with a statistically significant 50% and 64%, respectively, reduced risk of subsequent new HPV16 and HPV18 infection compared with seronegative women. This is a very encouraging finding and an indication that a fraction of young women with higher antibody titers are protected from subsequent HPV infection. Whether this protection can be attributed to neutralizing antibodies or some other immune protective mechanism needs to be further elucidated. If protection is conferred through antibodies, however, it is encouraging to note that the titers required for protection are much lower than those observed in vaccinated women. It will be of interest to investigate the longitudinal patterns of antibody titers in the context of natural infection and duration of protection conferred by natural infection.
Funding
Intramural Research Program of the National Cancer Institute, National Institutes of Health. National Institutes of Health Office of Research on Women's Health.
Footnotes
The funding agency did not have any involvement in the design of the study; the collection, analysis, and interpretation of the data; the writing of the article; or the decision to submit the article for publication.
The Costa Rica HPV16/18 Vaccine Trial is a long-standing collaboration between investigators in Costa Rica and the National Cancer Institute (NCI). The trial is funded by intramural NCI and the NIH Office of Research on Women's Health and is conducted in agreement with the Ministry of Health of Costa Rica. Vaccine was provided for our trial by GLAXOSMITHKLINE Biologicals, under a clinical trials agreement with NCI. GLAXOSMITHKLINE also provided support for aspects of the trial associated with the regulatory submission needs of the company. NCI and Costa Rican investigators make final editorial decisions on this publication; GLAXOSMITHKLINE has the right to review/comment.
The affiliations of the members of the Costa Rica HPV16/18 Vaccine Trial group are as follows. At the Proyecto Epidemiológico Guanacaste, Fundación INCIENSA, San José, Costa Rica: Mario Alfaro (Cytologist), Manuel Barrantes (Field Supervisor), M. Concepcion Bratti (Coinvestigator), Fernando Cárdenas (General Field Supervisor), Bernal Cortés (Specimen and Repository Manager), Albert Espinoza (Head, coding, and data entry), Yenory Estrada (Pharmacist), Paula Gonzalez (Coinvestigator), Diego Guillén (Pathologist), Rolando Herrero (Coprincipal Investigator), Silvia E. Jimenez (Trial Coordinator), Jorge Morales (Colposcopist), Lidia Ana Morera (Head Study Nurse), Elmer Pérez (Field Supervisor), Carolina Porras (Coinvestigator), Ana Cecilia Rodriguez (Coinvestigator), and Maricela Villegas (Clinic Physician); at the University of Costa Rica, San José, Costa Rica: Enrique Freer (Director, HPV Diagnostics Laboratory), Jose Bonilla (Head, HPV Immunology Laboratory), Sandra Silva (Head Technician, HPV Diagnostics Laboratory), Ivannia Atmella (Immunology Technician), and Margarita Ramírez (Immunology Technician); at the National Cancer Institute, Bethesda, MD: Nora Macklin (Trial Coordinator), Allan Hildesheim (Coprincipal Investigator and NCI Co-project Officer), Douglas R. Lowy (HPV Virologist), Mark Schiffman (Medical Monitor and NCI Co-project Officer), John T. Schiller (HPV Virologist), Mark Sherman (Quality Control Pathologist), Diane Solomon (Medical Monitor and Quality Control Pathologist), and Sholom Wacholder (Statistician); at SAIC, NCI—Frederick, Frederick, MD: Ligia Pinto (Head, HPV Immunology Laboratory) and Alfonso Garcia-Pineres (Scientist, HPV Immunology Laboratory); at Womens and Infants’ Hospital, Providence, RI: Claire Eklund (Quality Control, cytology) and Martha Hutchinson (Quality Control, cytology); and DDL Diagnostics Laboratory, Voorburg, the Netherlands, Wim Quint (HPV DNA testing), and Leen-Jan van Doorn (HPV DNA testing), GLAXOSMITHKLINE Biologicals, Rixensart, Belgium Catherine Bougelet (HPV16/18 ELISA testing).
References
- 1.Breitburd F, Kirnbauer R, Hubbert NL, et al. Immunization with viruslike particles from cottontail rabbit papillomavirus (CRPV) can protect against experimental CRPV infection. J Virol. 1995;69(6):3959–3963. doi: 10.1128/jvi.69.6.3959-3963.1995. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Suzich JA, Ghim SJ, Palmer-Hill FJ, et al. Systemic immunization with papillomavirus L1 protein completely prevents the development of viral mucosal papillomas. Proc Natl Acad Sci U S A. 1995;92(25):11553–11557. doi: 10.1073/pnas.92.25.11553. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Schiller JT, Lowy DR. Papillomavirus-like particles and HPV vaccine development. Semin Cancer Biol. 1996;7(6):373–382. doi: 10.1006/scbi.1996.0046. [DOI] [PubMed] [Google Scholar]
- 4.Paavonen J, Naud P, Salmeron J, et al. Efficacy of human papillomavirus (HPV)-16/18 AS04-adjuvanted vaccine against cervical infection and precancer caused by oncogenic HPV types (PATRICIA): final analysis of a double-blind, randomised study in young women. Lancet. 2009;374(9686):301–314. doi: 10.1016/S0140-6736(09)61248-4. [DOI] [PubMed] [Google Scholar]
- 5.Harro CD, Pang YY, Roden RB, et al. Safety and immunogenicity trial in adult volunteers of a human papillomavirus 16 L1 virus-like particle vaccine. J Natl Cancer Inst. 2001;93(4):284–292. doi: 10.1093/jnci/93.4.284. [DOI] [PubMed] [Google Scholar]
- 6.Koutsky LA, Ault KA, Wheeler CM, et al. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med. 2002;347(21):1645–1651. doi: 10.1056/NEJMoa020586. [DOI] [PubMed] [Google Scholar]
- 7.Paavonen J, Jenkins D, Bosch FX, et al. Efficacy of a prophylactic adjuvanted bivalent L1 virus-like-particle vaccine against infection with human papillomavirus types 16 and 18 in young women: an interim analysis of a phase III double-blind, randomised controlled trial. Lancet. 2007;369(9580):2161–2170. doi: 10.1016/S0140-6736(07)60946-5. [DOI] [PubMed] [Google Scholar]
- 8.Garland SM, Hernandez-Avila M, Wheeler CM, et al. Quadrivalent vaccine against human papillomavirus to prevent anogenital diseases. N Engl J Med. 2007;356(19):1928–1943. doi: 10.1056/NEJMoa061760. [DOI] [PubMed] [Google Scholar]
- 9.Future 11 Study Group. Quadrivalent vaccine against human papillomavirus to prevent high-grade cervical lesions. N Engl J Med. 2007;356(19):1915–1927. doi: 10.1056/NEJMoa061741. [DOI] [PubMed] [Google Scholar]
- 10.Schiller JT, Lowy DR. Immunogenicity testing in human papillomavirus-like-particle vaccine trials. J Infect Dis. 2009;200(2):166–171. doi: 10.1086/599988. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Bosch FX, Burchell AN, Schiffman M, et al. Epidemiology and natural history of human papillomavirus infections and type-specific implications in cervical neoplasia. Vaccine. 2008;26(suppl 10):K1–K16. doi: 10.1016/j.vaccine.2008.05.064. [DOI] [PubMed] [Google Scholar]
- 12.Ho GY, Studentsov Y, Hall CB, et al. Risk factors for subsequent cervicovaginal human papillomavirus (HPV) infection and the protective role of antibodies to HPV-16 virus-like particles. J Infect Dis. 2002;186(6):737–742. doi: 10.1086/342972. [DOI] [PubMed] [Google Scholar]
- 13.Viscidi RP, Schiffman M, Hildesheim A, et al. Seroreactivity to human papillomavirus (HPV) types 16, 18, or 31 and risk of subsequent HPV infection: results from a population-based study in Costa Rica. Cancer Epidemiol Biomarkers Prev. 2004;13(2):324–327. doi: 10.1158/1055-9965.epi-03-0166. [DOI] [PubMed] [Google Scholar]
- 14.Viscidi RP, Snyder B, Cu-Uvin S, et al. Human papillomavirus capsid antibody response to natural infection and risk of subsequent HPV infection in HIV-positive and HIV-negative women. Cancer Epidemiol Biomarkers Prev. 2005;14(1):283–288. [PubMed] [Google Scholar]
- 15.Herrero R, Hildesheim A, Rodriguez AC, et al. Rationale and design of a community-based double-blind randomized clinical trial of an HPV 16 and 18 vaccine in Guanacaste, Costa Rica. Vaccine. 2008;26(37):4795–4808. doi: 10.1016/j.vaccine.2008.07.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Dessy FJ, Giannini SL, Bougelet CA, et al. Correlation between direct ELISA, single epitope-based inhibition ELISA and pseudovirion-based neutralization assay for measuring anti-HPV-16 and anti-HPV-18 antibody response after vaccination with the AS04-adjuvanted HPV-16/18 cervical cancer vaccine. Hum Vaccin. 2008;4(6):425–434. doi: 10.4161/hv.4.6.6912. [DOI] [PubMed] [Google Scholar]
- 17.Harper DM, Franco EL, Wheeler C, et al. Efficacy of a bivalent L1 virus-like particle vaccine in prevention of infection with human papillomavirus types 16 and 18 in young women: a randomised controlled trial. Lancet. 2004;364(9447):1757–1765. doi: 10.1016/S0140-6736(04)17398-4. [DOI] [PubMed] [Google Scholar]
- 18.Kleter B, van Doorn LJ, ter SJ, et al. Novel short-fragment PCR assay for highly sensitive broad-spectrum detection of anogenital human papillomaviruses. Am J Pathol. 1998;153(6):1731–1739. doi: 10.1016/S0002-9440(10)65688-X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kleter B, van Doorn LJ, Schrauwen L, et al. Development and clinical evaluation of a highly sensitive PCR-reverse hybridization line probe assay for detection and identification of anogenital human papillomavirus. J Clin Microbiol. 1999;37(8):2508–2517. doi: 10.1128/jcm.37.8.2508-2517.1999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.van Doorn LJ, Molijn A, Kleter B, Quint W, Colau B. Highly effective detection of human papillomavirus 16 and 18 DNA by a testing algorithm combining broad-spectrum and type-specific PCR. J Clin Microbiol. 2006;44(9):3292–3298. doi: 10.1128/JCM.00539-06. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Brown DR, Kjaer SK, Sigurdsson K, et al. The impact of quadrivalent human papillomavirus (HPV; types 6, 11, 16, and 18) L1 virus-like particle vaccine on infection and disease due to oncogenic nonvaccine HPV types in generally HPV-naive women aged 16-26 years. J Infect Dis. 2009;199(7):926–935. doi: 10.1086/597307. [DOI] [PubMed] [Google Scholar]
- 22.Wheeler CM, Kjaer SK, Sigurdsson K, et al. The impact of quadrivalent human papillomavirus (HPV; types 6, 11, 16, and 18) L1 virus-like particle vaccine on infection and disease due to oncogenic nonvaccine HPV types in sexually active women aged 16-26 years. J Infect Dis. 2009;199(7):936–944. doi: 10.1086/597309. [DOI] [PubMed] [Google Scholar]
- 23.Harper DM, Franco EL, Wheeler CM, et al. Sustained efficacy up to 4.5 years of a bivalent L1 virus-like particle vaccine against human papillomavirus types 16 and 18: follow-up from a randomised control trial. Lancet. 2006;367(9518):1247–1255. doi: 10.1016/S0140-6736(06)68439-0. [DOI] [PubMed] [Google Scholar]
- 24.Kemp TJ, Garcia-Pineres A, Falk RT, et al. Evaluation of systemic and mucosal anti-HPV16 and anti-HPV18 antibody responses from vaccinated women. Vaccine. 2008;26(29–30):3608–3616. doi: 10.1016/j.vaccine.2008.04.074. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Wang SS, Schiffman M, Herrero R, et al. Determinants of human papillomavirus 16 serological conversion and persistence in a population-based cohort of 10 000 women in Costa Rica. Br J Cancer. 2004;91(7):1269–1274. doi: 10.1038/sj.bjc.6602088. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Winer RL, Lee SK, Hughes JP, Adam DE, Kiviat NB, Koutsky LA. Genital human papillomavirus infection: incidence and risk factors in a cohort of female university students. Am J Epidemiol. 2003;157(3):218–226. doi: 10.1093/aje/kwf180. [DOI] [PubMed] [Google Scholar]