Review

Phase III efficacy of Gardasil^®

CIN2 and CIN3 were accepted by regulatory agencies as intermediate disease end points for cervical cancer in phase III vaccine efficacy trials on the basis of the low incidence of cervical cancer in screened women, the long interval from HPV infection to cervical cancer, and the large body of screening data showing that detection and treatment of high-grade CIN grades 2 and 3 reduced cervical cancer incidence and mortality.³⁶ Selected highlights from public presentations, including the reporting of interim phase III data provided to the ACIP for Gardasil^®, and information from this vaccine's product label, have been described.^{29, 31}

The efficacy studies for Gardasil^® were conducted as two main substudies, Females United To Unilaterally Reduce Endo/Ectocervical Disease (FUTURE) I and FUTURE II. Both substudies demonstrated high efficacy.^{29, 37} To date, publicly available efficacy information for phase III studies is available only for Gardasil^®, and is based on approximately 1.5 years of follow-up after dose three of the vaccine. In the FUTURE II trial, for women (aged 15–26 years) who completed the vaccination regimen, did not violate the protocol, and had no virological evidence of infection with the respective HPV vaccine type at study entry through 1 month following the third vaccine dose (vaccine n = 5,301 versus placebo n = 5,258), vaccine efficacy was shown to be 100% (97.96% CI 76–100%) for preventing HPV-type-16-related or HPV-type-18-related CIN2 and CIN3 and adenocarcinoma in situ.^{29, 37}

In the FUTURE I trial, for women (aged 16–23 years) who completed the vaccination regimen, did not violate the protocol, and who had no virological evidence of infection with the respective HPV vaccine type at study entry through 1 month following the third vaccine dose (vaccine n = 2,261 versus placebo n = 2,279), vaccine efficacy was also reported as 100% (97.5% CI 88–100%) for preventing external genital warts or vulvar/vaginal intraepithelial neoplasia of any grade associated with HPV types 6, 11, 16 or 18. In 2,240 women who received the vaccine 100% of HPV type 6/11/16/18-related cervical lesions of any grade were prevented.²⁹

Combined data from phase II and III databases were used to conduct preliminary analysis of the impact of Gardasil^® on the overall clinical trial populations.²⁹ The reduction of HPV type 16/18-related CIN2 and CIN3 or adenocarcinoma in situ was estimated at 39% (95% CI 23–52%). The reduction in HPV type 6/11/16/18-related CIN or adenocarcinoma in situ was 46% (95% CI 35–56%), and HPV type 6/11/16/18-related genital warts was 69% (95% CI 58–77%). In addition, an interim analysis of combined phase II and III Gardasil^® studies (median follow-up 1.9 years) demonstrated a reduction in CIN2 and CIN3 of 12.2% (95% CI 3.2–25.3%) compared with placebo, when extended to include HPV types not targeted by the vaccine.²⁹ These data are based on a relatively short-term follow-up period and the analyses include vaccine recipients with prevalent HPV infections; longer follow-up will allow more-accurate measures of HPV vaccine impact and a determination of whether immunity is sustained and incident HPV infections prevented. No significant benefit was observed with Gardasil^® in women who were already infected with HPV vaccine types.^{29, 38} The efficacy of Gardasil^® is unknown for younger girls and for males, although trials are ongoing.²⁹ Mathematical modeling studies indicate that vaccination of males may not be cost-effective, particularly if high population penetration is achieved in females.^{39, 40}

Safety of Gardasil^® and potential risks

Few safety issues have arisen during any of the trials of Gardasil^® or Cervarix^®.²⁹ Injection site reactions were reported in 83% of Gardasil^® recipients and in 73.4% of placebo recipients. The most common injection-site experiences were erythema, pain, and swelling, with severe symptoms being more often reported by vaccine recipients than those on placebo. The most common systemic adverse experiences were fever, headache and nausea. There was no statistically significant difference in adverse experiences between the vaccine and placebo recipients. Trial data show short term equivalent safety for already-exposed women (i.e. those infected with HPV types recognized by the vaccine, or with evidence of type-specific antibodies) receiving the vaccine and those unexposed to the vaccine HPV types. Also, Gardasil^® booster doses given at approximately 4.5 years following completion of the initial three-dose vaccine series were shown to be safe and to elicit robust memory immune responses.³⁸

Women with positive pregnancy tests before planned vaccinations were excluded from receiving Gardasil^®.²⁹ Some women became pregnant, however, during the few weeks or months following the receipt of either vaccine or placebo. During Gardasil^® clinical trials, 2,266 women (1,115 receiving the vaccine and 1,151 receiving placebo) reported at least one pregnancy each. The proportions of pregnancies with an adverse outcome were comparable in subjects who received Gardasil^® and those who received placebo: 40 individuals in the Gardasil^® group, and 41 in the placebo (3.6% of all subjects who reported a pregnancy in either group) experienced a serious adverse experience during pregnancy. The most common events reported were conditions that can result in Cesarean section (e.g. failure of labor, malpresentation and cephalopelvic disproportion), premature onset of labor (e.g. threatened abortions and premature rupture of membranes), and pregnancy-related medical problems (e.g. pre-eclampsia and hyperemesis). Congenital anomaly during pregnancy occurred in 15 subjects who received Gardasil^® and 16 who received placebo. Further subanalyses were conducted to evaluate pregnancies with estimated onset within 30 days of administration of a Gardasil^® or placebo dose, and those with an estimated onset of more than 30 days after administration. In pregnancies with estimated onset within 30 days of vaccination, five cases of congenital anomaly were observed in the group that received Gardasil^®, compared with no cases of congenital anomaly in the group that received placebo. The congenital anomalies observed included pyloric stenosis, congenital megacolon, congenital hydronephrosis, hip dysplasia and clubfoot. Conversely, in pregnancies with an onset beyond 30 days following vaccination, 10 cases of congenital anomaly were observed in the group that received Gardasil^®, compared with 16 cases in the group that received placebo. Regardless of when pregnancy occurred in relation to vaccination, the types of anomalies observed were consistent with those generally observed in pregnancies in women aged 16–26 years.

Gardasil^® was assigned a pregnancy risk factor category B by the US FDA, but additional information regarding the risk in pregnant women is needed. Category B assignment can include drugs that have no demonstrated adverse reproductive or developmental effects in animals and for which there are no adequate and well-controlled data in humans. Category B can also include drugs that have demonstrated adverse reproductive or developmental effects in animals but did not demonstrate adverse effects in adequate and well-controlled trials in humans. A summary of possible pregnancy outcomes can be found within the Gardasil^® product label.²⁹ A post-marketing registry for monitoring of the marketed vaccine has been developed to further evaluate general safety and pregnancy outcomes. The product label encourages patients and health-care providers to contact the registry to report any exposures to Gardasil^® that occur during pregnancy.²⁹

Although not measured in HPV vaccine trials, there are other potential risks from HPV vaccination following introduction into the population. Reductions in safer sex practices and screening for cervical cancer due to potential misconceptions about HPV vaccine protection should be of concern. Monitoring such effects will require research and the creation of special surveillance programs capable of linking HPV vaccination with a number of HPV-related disease outcomes as well as with provider and patient practices. Similarly to other vaccines, widespread use of HPV vaccines may elicit rarely seen adverse reactions and unanticipated outcomes. Questions and concerns have been raised about whether other carcinogenic HPV types not contained in the vaccines will replace HPV types 16 and 18 in the ecological niche. One can only speculate regarding potential unknowns, but, in the public's interest, pharmacovigilence and surveillance activities will be required.

Durability of immunity

Little information is currently available on the duration of HPV-vaccine-induced immunity. Phase II studies have demonstrated diminishing vaccine-generated antibody titers, which plateau after 24 months and then remain relatively stable for 4–5 years.^{28, 33, 35} To date, no studies have been able to identify any immune correlate of HPV-vaccine-induced protection. Geometric mean antibody titers are similar among vaccinated individuals presenting as cases or non-cases; therefore, there is no way at present to determine whether specific levels or the presence or absence of measurable antibodies will correlate with loss of vaccine protection.

In the US, given the broad ACIP recommendation,^{30, 31} many women who are currently infected with HPV vaccine types will be vaccinated. Population-based surveillance to assess durability of vaccine immunity will detect HPV vaccine types in women who received an HPV vaccine when already infected with the virus. For the near future, the determination of post-licensure durability of HPV vaccine immunity will be difficult. Questions may also arise regarding the utility of HPV testing before vaccination—an issue that needs clarification. There are presently no validated and approved HPV typing tests, and, even if these tests become available, HPV testing and typing before vaccination is inappropriate because single HPV tests are inadequately informative and there is no good measure of past HPV exposure.³¹

Optimum age for maximizing vaccine efficacy

There is general agreement that it is important and most effective to vaccinate girls and women before the age at which sexual exposure to HPV is likely to occur. In the US, 26% of females are sexually active by the age of 15 years, 40% by age 16 years, and over 70% by age 18 years.⁴¹ Approximately half of sexually active women aged 19–21 years report having had four or more sexual partners in their lifetime,⁴² greater than 10% of sexually active ninth graders (13–15-year-olds) have had four or more sexual partners,⁴³ and HPV acquisition occurs rapidly after sexual initiation. In one report, 39% of college-aged women (typically 18 years) acquired HPV within 24 months of sexual debut.⁴⁴ Vaccinating before the onset of sexual intercourse would obviously provide the greatest benefit to the largest numbers of individuals.

Although it is expected that vaccine effectiveness will decline with increasing number of sexual partners and that this correlates with increasing age, further clinical trials are ongoing to study HPV vaccine efficacy in women within the age range 26–55 years.⁴⁵ It is important for states, health-care organizations, health-care providers and individuals to consider the likely diminishing benefits if populations of women who have a high probability of previous HPV exposures are vaccinated. It is also important to recognize that vaccine efficacy in clinical trials does not represent vaccine effectiveness when transferred to real world settings; nor does it incorporate cost–benefit considerations.

When considering the optimum age for HPV vaccination, the actual HPV exposure data from women enrolled in any HPV vaccine trial should be assessed along with the body of literature on HPV population prevalence and incidence. At the time women were enrolled in the Gardasil^® trials, 27% were estimated to have had previous exposure to one or more of the HPV vaccine types.²⁹ This estimate was made on the basis of the combined HPV DNA type-specific polymerase chain reaction results from genital specimens (current exposure) and serological testing (past or current exposure) of women at trial entry. These laboratory test measurements are associated with underestimation errors. Averages taken from published studies indicate that 40% of women with detectable cervical HPV DNA never develop measurable HPV-type-specific antibodies.^{46, 47} In addition, HPV DNA point prevalence estimated from single tests represents an underestimate of current HPV infection by a factor ranging from 1.4 to above 2.0.^{48, 49, 50} Given the measurement errors inherent in defining past and present HPV infection, one could estimate that approximately twice as many trial participants were exposed to HPV vaccine types when entering the trials than was indicated by research laboratory assays. A twofold increase in estimated HPV exposures would be consistent with previous studies and indicates that approximately half of all American women aged between 16 and 26 years with an average of only two sexual partners will have been exposed to one or more of the vaccine HPV types. As noted earlier, one survey performed in the US reported that, on average, sexually active women aged 19–21 years had had four or more sexual partners,⁴¹ which is twice the average number of sexual partners reported for Gardasil^® trial participants. The benefit and impact of the HPV vaccine for a general population of sexually active women is, therefore, presumably overestimated by current HPV vaccine trial measurements.

On the basis of a large number of epidemiological studies conducted throughout the world, the likelihood of already having been exposed to HPV types 6 or 16 will most certainly be high in sexually active women. The impact of HPV vaccines on disease reduction in the general population will be determined by the baseline prevalence of the HPV vaccine types in the population and the level of population exposure estimated by the average lifetime number of sexual partners in women receiving the vaccine. These data become relevant when considering the benefits and effectiveness of the vaccine and, ultimately, competing health-care resources.

Cost-effectiveness, and ethical and other considerations

Screening programs are available to sexually active women in many industrialized nations. As noted earlier, these programs already represent approximately $5 billion in US health-care expenditure. The estimated US costs of HPV vaccination under ACIP recommendations are as follows. To vaccinate 20% of the 4 million person cohort of American 11–12-year-old females, the three-dose series would require 2.4 million doses.³⁰ At the current US market price of $120 per dose, or $360 per three-dose series, this would equate to a vaccine expenditure of $288 million. At 80% penetration for 11–12-year-olds, the US vaccination costs would approach $1.15 billion. If 80% penetration were achieved in females from 11 to 18 years old, the vaccination costs alone for the US would approach $4.6 billion. Extension of HPV vaccination to women aged 26 years would, at any penetration level, double the costs required to vaccinate a similar number of females aged 18 years and younger. These cost estimates do not include administration and other provider costs or potential government or manufacturer vaccine discounting. The figures also reflect the costs resulting from recommended 'catch-up' vaccination in girls and women aged 13–26 years. The steady-state recurring costs to vaccinate 11–12-year-old girls would be restricted to approximately $600 million per year if vaccination of 13–26-year-old girls and women was not widely adopted.

On an individual basis, some older women—such as those who have not previously been sexually active—could benefit fully from HPV vaccination, but it must also be recognized that many of these women might be expected to pay for the vaccination. In any setting, women without the access to or resources for HPV vaccination are often those who could have benefited most. Health-care disparities are known to affect the rates of both vaccination and screening,^{5, 6, 7, 51} and thus these same at-risk individuals might also fail to either participate in or have access to both primary or secondary cervical cancer prevention programs. For cervical cancer incidence to be reduced, women will require both screening and vaccination, as the first-generation HPV vaccines do not provide protection against a number of carcinogenic HPVs. Previously existing disparities will also need to be addressed. In countries with good screening programs, HPV vaccines may have little impact on the incidence of invasive cervical cancer. The benefits of HPV vaccines in these countries could be obtained primarily through a reduction in morbidity and costs associated with diagnosing and treating genital precancers.

In terms of providing HPV vaccines to those in most need, it should be acknowledged that the greatest burden of cervical cancer is found in underserved, resource-poor populations living in developing countries in which screening may simply be unavailable or inadequately supported. In total, 80% of all incident cervical cancer and related mortality occurs in the developing world.² Bringing HPV vaccination to these women might be the only hope of reducing the global burden of cervical cancer.⁵¹ Such a scheme will require extreme discounting of existing HPV vaccines, the development of alternative low-cost HPV vaccines, or both. The second approach has received some support.⁵²

Reduction or elimination of cervical cancer risk in any population through vaccination will require many decades and most likely second-generation, and possibly third-generation, HPV vaccines that provide protection against a broader spectrum of carcinogenic HPV types. Heptavalent and octavalent HPV VLP vaccines and other potential non-VLP HPV vaccines under consideration, such as those based on the HPV L2 minor capsid protein cross-reactive epitopes,^{53, 54} might provide broader coverage against the majority of carcinogenic HPV types. Reduction of cervical cancer incidence will ultimately be determined by several factors, including the baseline prevalence of carcinogenic HPV, the level of vaccination coverage in the population, the number of carcinogenic HPV types included in the vaccines, the durability of vaccine protection, the adequacy of accompanying provider and patient education programs, whether recommended screening practices are maintained at high levels, and finally how well we address continuing health-care disparities. These issues bring us to consideration of cervical cancer screening programs and improvements in secondary prevention likely to be realized through HPV testing.

Primary screening

HPV testing as a primary screening tool has been examined in a number of settings, and the evidence supporting this application has been carefully analyzed and reviewed elsewhere.⁵⁶ A recent overview of European and North American studies on HPV testing in primary cervical cancer screening showed that the overall sensitivity of HPV testing for CIN grade 2 or above was substantially better than that of cytology (96.1% versus 53%), although the specificity was slightly reduced (90.7% versus 96.3%).⁵⁷ The sensitivity of HPV testing was high in women of all ages, whereas the sensitivity of cytology was significantly better in women over the age of 50 years. HPV test sensitivity was reproducible between different areas of Europe and North America, whereas cytology sensitivity was highly variable. The specificity of both HPV testing and cytology increased with age. Testing efficiency increases in older women because the prevalence of newly acquired HPV infection declines, as does that of CIN2. At the same time, the prevalences of cancer and its immediate precursor, CIN3, increases. The increase in HPV test sensitivity and its high reproducibility compared with cytology indicates that it could be viable as the primary screening test in older women, with triage to cytology only occurring in women who are HPV-positive. Several studies are ongoing to examine the length of protection provided by a negative HPV test and to determine the appropriate management strategies for HPV-positive, cytology-negative women. Regional differences have been observed for HPV prevalence in women without high-grade CIN; therefore, the most age-appropriate strategies for HPV DNA testing alone or as an adjunct to cytology will most likely be region-specific and will be related to available resources and implementation issues, as well as to HPV prevalence.

New primary screening approaches based on HPV testing alone might be more readily welcomed in settings where screening programs are either virtually absent or undersupported. Recently, efforts have been undertaken to develop low-cost HPV tests for resource-poor regions.⁵⁸ In resource-rich environments, despite a large body of supportive data for HPV testing, and numerous ongoing efforts, the long and successful history of Pap screening is likely to retard the adoption of routine HPV testing as a preferred primary screening test.

Atypical and low-grade cervical abnormalities

Low-grade and equivocal cervical cytology diagnoses represent the greatest number of abnormal diagnoses associated with cervical cancer screening programs.⁴ The largest proportion of all histologically confirmed high-grade abnormalities are detected among women with these cytologic diagnoses.⁵⁹ The potential cost-effective management of these diagnostic categories through HPV testing has, therefore, been a very active area of study. In the Atypical Squamous Cells of Undetermined Significance/Low-Grade Squamous Intraepithelial Lesions Triage Study (ALTS), a randomized multisite trial conducted by the US National Cancer Institute, a high rate of HPV positivity was found among women with low-grade squamous intraepithelial lesions, thereby undermining the prospect of any utility for HPV testing in distinguishing women at increased risk for true cervical cancer precursors.⁶⁰ In ALTS and in other studies,⁶¹ however, HPV testing has been shown to have utility in the management of women diagnosed with equivocal or atypical squamous cells of unknown significance (ASCUS).⁶² Approximately half of all women with equivocal cytology diagnoses are HPV-negative and do not require further costly colposcopic evaluations. Triage of ASCUS Pap smears can be performed either by the use of residual liquid cytology specimens or by co-collection of a second sample in HPV test transport media.⁶³ This 'reflex' HPV testing for atypical Pap tests has been shown to be cost-effective.⁶⁴

Human papillomavirus as a test of cure

Overall, standard treatments of cervical cancer precursors are very effective. After therapeutic excision or ablation of cervical precancers, the risk of recurrence is about 5–10%.^{65, 66} Generally, a combination of cytology and colposcopy has been used to monitor post-treatment outcomes. HPV testing at 4–6-month intervals following therapy is highly sensitive and specific and better than cytology alone for monitoring the risk of disease recurrence.^{66, 67} Women who tested positive after LEEP for carcinogenic HPV types, especially HPV type 16, had a high risk of subsequent CIN2 or above. HPV-based detection methods, alone or in combination with cytology, could be usefully incorporated into post-LEEP management strategies.⁶⁶