Overview and rationale: Through Swedish health care registers it is possible vaccine effectiveness. The overall aim is to develop tools for fast and reliable evaluation of vaccines focusing on effectiveness. In this study the investigators will assess the population impact on condyloma prescribed drugs after HPV-vaccination and to estimate the economic burden of hospitalizations due to condyloma. Genital warts (condyloma) are one of the most common sexually transmitted diseases in Sweden, estimated 20 000 identified cases per year, and are primarily treated with Podofyllotoxin and Immiquimod. One of the approved HPV vaccines also protects against condyloma. In clinical trials, this vaccine is safe and highly efficacious (90-100%) against persistent infection with HPV 6 and 11 and genital warts in women and men. However, these results are from clinical trials and do not answer the question weather the vaccine works when delivered, as it would be in the real world. By linkage between the Svevac, the patient register and the Drug prescription register the investigators can estimate the effectiveness of the HPV vaccine in preventing cases of condyloma in a defined population.

Method and execution: This is a retrospective collection of data from Swedish health care registries where the study population will consist of persons who are HPV-vaccinated as well as a non-HPV-vaccinated control group (identified through the population-based register, matched on sex, age and living area). The primary statistical analysis will be a calculation of prescriptions of condyloma treatments after vaccination compared to the non-vaccinated control group.

GARDASIL® POPULATION-BASED CONDYLOMA EFFECTIVENESS STUDY IN SWEDISH WOMEN BACKGROUND Human papilloma virus (HPV) has been established as a necessary but not sufficient cause of cervical cancer. Over 100 HPV types have been identified, of which 40 infect the genital tract. HPV16, followed by HPV18, are the most frequent in cervical cancer and contribute to 70% of all invasive cervical cancers. All together, around 15 high-risk (or oncogenic) HPV types have been identified. Cervical cancer is the second most common cancer among women globally, with an estimated half a million new cases and a quarter of a million deaths each year.

Two of the 40 genital HPV types, 6 and 11, are so called low risk types causing genital warts (condyloma). Genital warts are one of the most common sexually transmitted infections and the clinical burden has been increasing the past decades. The warts are benign but often cause mental or emotional stress in those experiencing them.

There are two available vaccines against HPV. Both of these vaccines have shown 95% efficacy against HPV16 and 18-related precancerous lesions in clinical trials but the follow-up time is so far limited to 5.5 years. One of the vaccines, Gardasil®, also offers protection against HPV 6 and 11. In clinical trials, this vaccine is safe and highly efficacious (90-100%) against persistent infection with HPV 6 and 11 and genital warts in women and men. Recently the first study on the effect of Gardasil® on national population level was published. The study showed a substantial decline of condyloma in Australian women after introduction of GARDASIL®. However, there were some limitations to this study. It was not based on entirely true population-based data, they could not adjust for residency status and they could not link cases to vaccination status.

The aim of this study is to investigate whether Gardasil® is effective against condyloma in vaccinated women compared to non-vaccinated women by using registry-based data. The study is also going to investigate whether Gardasil® has an effect on cases with previous history of condyloma, which is possible with our data. Further, we will try to estimate the economic burden of hospital visits due to genital warts and investigate whether Gardasil® has an affect on these hospital visits. It is estimated that out of 20 000 visits to health care, for condyloma, one third requires surgery due to recurrence, and of those one third will require surgery again. The study will take advantage of the unique Swedish population-based health care databases, which allows linking and follow-up at the individual level.

OBJECTIVES

• To assess the population impact on Podofyllotoxin/Aldara prescription and hospital visits on condyloma in Sweden after vaccination with GARDASIL®.
• To assess if vaccination with GARDASIL® has an effect on recurrence of genital warts.
• To estimate the economic burden of hospital visits due to genital warts.

MATERIALS AND METHODS Study design The study is a historically prospective analysis of population-based nationwide cohorts comprising participants with individual-level longitudinal information on HPV vaccination status and prescription of Podofyllotoxin/Aldara. Vaccination status will be obtained from national vaccination registers, and drug prescription register (Sweden) and information about Podofyllotoxin/Aldara from the drug prescription register. Information about hospital visits due to genital warts will be obtained from the national in- and out patient registry. The study is observational and as such vaccination will be according to local regulations, policies and practices.

Registers used HPV vaccination status will be obtained from the Swedish vaccination register (SVEVAC) and the Swedish drug prescription register. The Swedish vaccination register (SVEVAC) currently registers, through a user-friendly web-based large-scale system, all HPV vaccinations in Sweden nationally, since the start of HPV vaccination in 2006. SVEVAC holds information on personal ID-number, sex, age, date of vaccination, time of vaccination, type of vaccine, vaccine lot number, and health unit (county and municipality). Registration with personal ID-number is voluntary, but subject must actively state if they don't want to be registered. Written information to the subject and inquiry on consent for all aspects of the registration and register based follow-up will be administered in an efficient way (subject to approval from Ethics Committee). Present experience with HPV vaccination is that 95% of vaccinated subjects accept this registration. In Sweden opportunistic HPV vaccination has been available since May 2006 and is recommended to young adults. The vaccine has been subsidised for girls age 13-17 since May 2007 through the Swedish drug reimbursement system. A prescription from a medical doctor is then required. All drug prescriptions are since 2005 registered in a national drug prescription register. To compare all HPV vaccinated subjects the drug prescription register will be scrutinized to complete information on HPV vaccination. The drug register holds information on prescription date, number of doses, ID-number, sex, type of vaccination. Today approximately 100 000 Swedish females have been vaccinated with the HPV vaccine, as well as 800 males. The majority of vaccinations were with Gardasil® (98%).

From the drug prescription register we will obtain information on Podofyllotoxina and Aldara prescriptions, date of prescription and dose. In the majority of cases prescriptions are for self-treatment at home.

All in-patient health care in Sweden is registered with personal ID-number at a national level since 1987. The so-called Patient register is kept by the Swedish National Board of Health and Welfare (NBHW). From 2001, out-patient care at hospitals has been included in this register. Information on adverse events will be obtained from this register. The International Classification of Diseases 10th revision (ICD-10) has been used to code diagnoses since 1997.

Many specialist clinics are based at hospitals and using patient register it will be possible to catch visits to those clinics as well as day surgery.

Study population and study size This is a population-based study and will comprise girls/women vaccinated with GARDASIL® in Sweden between May 2005 and October 2010. The study will also include all women who were prescribed Podofylloxin/Aldara between January 2005 and December 2010 and also women who have a recorded hospital visit due to genital warts from 2005 to December 2009. The study cohort will include girls/women born between 1989 and 1996. In this way we will include the oldest and youngest girls/women who may benefit from subsidised vaccination within the study period.

As of October 2010, 98% of the vaccinations registered were Gardasil®. 108 670 women have been vaccinated with at least one dose, 92 296 with two doses and 66 150 with three doses. Approximately 87% of the vaccinated girls/women are between 13 to 17 years of age and 10% are between 18-25 years of age. In a report from the NBHW (2007) the estimated number of cases with condyloma is about 20 000 every year, taking both prescription of Podofyllotoxin/Aldara and hospital visits into account. Approximately half of those are women.

To link information on vaccination status, and health care data from population based registers data the Swedish personal ID number will be used.

Study outcome To study the effect of Gardasil on condyloma incidence we will compare incidence of genital warts in vaccinated versus non-vaccinated women. The majority of cases will be identified through the Drug prescription register and about 20% will be found in patient register. To find cases through the Drug prescription register ACT codes will be used for Podofyllotoxin and Aldara (see below) and ICD-10 codes for the In- and Out patient registers.

The described register will also be used to investigate whether vaccination with Gardasil has changed the burden of recurrent warts by identifying women who had genital warts before vaccination and after and compare with those not vaccinated. The burden of hospital visits will also be investigated by comparing vaccinated and non-vaccinated women.

ACT codes used in the Drug prescription register ACT-code used for Gardasil®: J07BM01 ACT-code for Podofyllotoxin: D06BB04 ACT-code used for Aldara: D06BB10 ICD-10 codes Hospital visit due to genital warts will be found in the in- and out patient register. The ICD-10 code used to indentify patients are: A63 and D26.

Statistical methods This is a cohort study and the data will be analyzed using statistical methods for survival analysis. In particular, to study the population impact on genital wart incidence in Sweden after vaccination with GARDASIL® we will model the association between exposure (vaccination status) and outcome (genital warts determined via either prescription of Podophyllotoxin/Aldara or hospital visit due to genital warts) using Cox and/or Poisson regression. Incidence rate ratios (IRR) and 95 % confidence intervals will be used to quantify the effect of the vaccine on the risk for genital warts. All girls will be followed up from May 2006 and attained age will be used as the underlying time scale in the model. Because full effectiveness of the vaccine is assumed to be achieved after 3 doses the exposure will be included in the statistical model as a time-varying exposure classified from no vaccination (unexposed), partially completed vaccination (1 to 2 doses) and fully vaccinated (exposed). The analysis will account for possible confounding factors such as sexual risk taking behaviors, using genital wart history as marker, socioeconomic status of the parents and type of prescribed treatment (Podophyllotoxin orAldara).

To assess risk of recurrence of condyloma, IRR adjusted for prior genital wart history, determined in the same way as for outcome, will be used.

To estimate temporal trends in the economic burden of hospital visits due to genital warts we will identify and report descriptive data on all recorded hospital visits due to genital warts in the Swedish Patient registries as a function of calendar year (2005-2009) as well as mode of treatment (Podophyllotoxin/Aldara /surgery).

Limitations We cannot exclude that we will not be able to catch all cases that seek medical care. Some cases that visit primary health care may be treated during the visit and this will not be recorded anywhere. However, most of these cases do receive a prescription for self-treatment.

Because the majority of vaccinees are between 13-17 years of age they may not have had genital warts before vaccination and therefore recurrence of warts may be difficult to catch in our study population.

Ethical approval The study will be conducted after approval of the Ethics Committee of Karolinska Institutet.

Inclusion Criteria:

• Individuals born 1989-1996

Exclusion Criteria:

• Individual vaccinated with Cervarix

Historically prospective study of population-based nationwide cohorts.