Skip Navigation


American Journal of Epidemiology Advance Access originally published online on January 18, 2006
American Journal of Epidemiology 2006 163(6):534-543; doi:10.1093/aje/kwj077
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
163/6/534    most recent
kwj077v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (29)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Burchell, A. N.
Right arrow Articles by Franco, E. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Burchell, A. N.
Right arrow Articles by Franco, E. L.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

American Journal of Epidemiology Copyright © 2006 by the Johns Hopkins Bloomberg School of Public Health All rights reserved; printed in U.S.A.

Original Contribution

Modeling the Sexual Transmissibility of Human Papillomavirus Infection using Stochastic Computer Simulation and Empirical Data from a Cohort Study of Young Women in Montreal, Canada

Ann N. Burchell1,2, Harriet Richardson1,3, Salaheddin M. Mahmud1,4, Helen Trottier1, Pierre P. Tellier5, James Hanley2, François Coutlée6 and Eduardo L. Franco1,2

1 Department of Oncology, Faculty of Medicine, McGill University, Montreal, Quebec, Canada
2 Department of Epidemiology and Biostatistics, Faculty of Medicine, McGill University, Montreal, Quebec, Canada
3 Division of Cancer Care and Epidemiology, Faculty of Medicine, Queen's University, Kingston, Ontario, Canada
4 Department of Community Health Sciences, Faculty of Medicine, University of Manitoba, Winnipeg, Manitoba, Canada
5 Department of Family Medicine, Faculty of Medicine, McGill University, Montreal, Quebec, Canada
6 Département de Microbiologie, Centre Hospitalier de l'Université de Montréal, Montreal, Quebec, Canada

Correspondence to Dr. Eduardo Franco, Division of Cancer Epidemiology, McGill University, 546 Pine Avenue West, Montreal, Quebec, Canada H2W 1S6 (e-mail: eduardo.franco{at}mcgill.ca).

Received for publication July 26, 2005. Accepted for publication October 26, 2005.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
The authors estimated plausible ranges of the probability of human papillomavirus (HPV) transmission per coital act among newly forming couples by using stochastic computer simulation. Comparative empirical data were obtained in 1996–2001 from a cohort study of female university students in Montreal, Canada. Female prevalence and frequency of sexual intercourse and condom use were set equal to those in the cohort. Simulations included 240 combinations of male prevalence, the relative risk for protected versus unprotected sex, and per-act transmission probabilities. Those that produced expected HPV incidence within the 95% confidence interval observed in the cohort were selected. The observed 6-month cumulative incidence following acquisition of a new partner was 17.0% (95% confidence interval: 11.4, 23.0). Expected incidences consistent with those from cohort findings occurred in 54/240 simulations. The range of per-act transmission probabilities was 5–100% (median, 40%). Male HPV prevalence was the same as or greater than that for women in all consistent simulations. Varying condom effectiveness did not produce better-fitting data. This simulation suggests that HPV transmissibility is several-fold higher than that for other viral sexually transmitted infections such as human immunodeficiency virus or herpes simplex virus 2. With high transmissibility, any potential protective effect of condoms would disappear over multiple intercourse acts, underlining the need for an effective HPV vaccine.

disease transmission; papillomavirus, human; sexually transmitted diseases; uterine cervical neoplasms

Abbreviations: HIV, human immunodeficiency virus; HPV, human papillomavirus; STI, sexually transmitted infection


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 
Human papillomavirus (HPV) is the most common sexually transmitted infection (STI). Cervical HPV infection is found in 5–40 percent of asymptomatic women of reproductive age (1Go), and as many as 75 percent of adults may eventually be infected in their lifetime (2Go). Risk rises with increasing number of sexual partners, younger age at sexual debut, and recent acquisition of new partners (3Go–7Go). The vast majority of these infections will be transient (3Go, 8Go–12Go). However, a substantial increase in risk of cervical neoplasia exists for women who develop persistent, long-term infections with oncogenic HPV types (3Go, 9Go, 13Go–15Go). It is now well established that HPV infection is the central, probably necessary cause of cervical cancer (16Go).

The acknowledgment that cervical cancer is caused by an STI has produced a change from a noninfectious to an infectious disease paradigm, with corresponding changes in prevention strategies. There is currently great enthusiasm concerning the possible application of HPV testing as an adjunct to Papanicolaou cytology screening for cervical cancer (17Go) and widespread interest in the development of HPV vaccines (18Go). However, assessments of the potential impacts of these proposed strategies are hampered by limited information on the sexual transmissibility of HPV. To date, most natural history models that predict the impact of HPV testing and vaccination strategies have been based on empirical data that have come exclusively from epidemiologic studies of women (19Go–23Go). A better understanding of the sexual transmission dynamics of HPV would lead to more informed decision making when different prevention strategies are compared through more valid mathematical prediction models.

In the absence of empirical data on HPV transmissibility, computer simulation may be a useful tool for estimation. The objective of this study was to simulate probabilities of HPV transmission per coital act in a hypothetical population to estimate plausible ranges for this parameter that would be coherent with observed rates of HPV incidence among young, sexually active women enrolled in a cohort study we previously conducted in Montreal, Canada (24Go).


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 References
 
Hypothetical populations of newly forming heterosexual couples were simulated. Acquiring a new partner has been shown to be a key determinant of HPV acquisition (7Go). Therefore, newly forming rather than long-standing couples were the object of analysis.

McGill-Concordia Cohort Study
The source of empirical data was the prospective McGill-Concordia Cohort Study of young female university students in Montreal, Canada. Women attending either the McGill or Concordia university health services clinics were recruited for a study of the natural history of HPV infection and cervical neoplasia. The study methodology is described in detail elsewhere (24Go). In brief, 621 female participants were followed for 24 months at 6-month intervals in 1996–2001. At each visit, a cervical specimen was collected and tested for 27 HPV types using L1 consensus primers MY09/MY11 and HMB01 and the line blot assay (Roche Molecular Systems, Basel, Switzerland) (25Go). Women also self-completed questionnaires, which collected information on sexual history and behavior since the last visit. In the overall cohort, baseline cervical HPV prevalence was 29 percent for any type, 22 percent for high-risk oncogenic types, 15 percent for low-risk types, and 7 percent for HPV-16 (24Go).

Of 2,058 follow-up study visits, there were 238 visits by 182 women in which a new sexual partner was reported since her last visit, and no other partners ("new partner visits"). Empirical estimates of simulation parameters and cumulative HPV incidence were based on data from these new partner visits. Each new partner visit was assigned two time points: time t, the visit at which a new sexual partner was reported; and time t – 1, the visit immediately preceding time t. The median duration of the interval between t – 1 and t was 6 months (range, 3–28 months). Cumulative incidence of any new type of HPV was calculated using the Kaplan-Meier method. To account for repeated event times, the 95 percent confidence interval was estimated using bootstrap sampling of the 182 women who reported at least one new partner visit (26Go).

Simulation approach
A stochastic Monte Carlo computer simulation produced hypothetical cohort data for a population of 10,000 newly forming heterosexual couples. The assumed values for fixed and variable parameters used in the simulations are summarized in table 1.


View this table:
[in this window]
[in a new window]
  		TABLE 1. Fixed and variable parameters used in simulations of male-to-female transmission of HPV*,{dagger}

 
The first step was to assign the initial type-specific HPV positivity for each hypothetical female. Doing so involved drawing a random variable from the standard uniform distribution, which was compared with the observed type-specific prevalence at time t – 1 (table 1). For example, HPV-16 prevalence was 4.37 percent. Then, if the drawn random variable was less than or equal to 0.0437, the hypothetical female was assigned to be HPV-16 positive at time t – 1 (i.e., female HPV-16 positivity ~ Bernoulli (0.0437)). Because HPV prevalence among male partners of McGill-Concordia cohort women was unknown, the male-to-female prevalence rate ratio at time t – 1 was varied from 0.5 to 2.0. With respect to HPV status, random mixing of males and females was assumed. That is, HPV positivity in one partner was considered independent of that in the other when the couple was initially formed.

Simulated data on the frequency of intercourse over a 6-month interval were then generated for each couple. Intercourse frequencies per month were set to be the same as in the cohort (table 1) using randomly drawn numbers from the gamma distribution, rounded to the nearest integer, that most closely matched the empirical distribution (shape = 1, scale = 10). Each couple was also randomly assigned condom use frequency. It was assumed that women interpreted "sometimes" as condom use 50 percent of the time and "regularly" as condom use 75 percent of the time.

Given the uncertainty regarding condom efficacy, the relative risk of HPV transmission for a single act of protected versus unprotected intercourse was varied from 0.1 to 1.0 (table 1). A lower bound of 0.1 was selected because it most closely approximates that for another viral STI, human immunodeficiency virus (HIV), for which considerable data on condom effectiveness have been accumulated (27Go).

To simulate the incidence of HPV in the female partner, per-act transmission probabilities were varied from 0.001 to 1.0 (table 1). A lower bound of 0.001 was selected because it is the estimate for HIV given conditions of low viral load and long-standing partnerships (28Go, 29Go). For each act of intercourse between HPV-discordant couples, a variable was randomly drawn from the standard uniform distribution and transmission events were assigned, taking into account condom use.

The simulation outputs a data set with the type-specific HPV status of simulated women for times t – 1 and t. Kaplan-Meier analysis was used to calculate the expected cumulative incidence of any new HPV type at 6 months, as it would have been observed in a hypothetical cohort.

For each of the 240 possible combinations of the male-to-female prevalence rate ratio, relative risk for condom use, and per-act transmission probability value, 100 simulations of 10,000 couples were run. Resulting cumulative incidences were averaged over the 100 simulations to provide the best estimate of what would be expected under those conditions. Expected incidences were then compared with the 95 percent confidence interval for the observed cumulative incidence. Simulated conditions that produced expected cumulative incidences within this range were considered compatible.


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 References
 
Cumulative incidence of HPV infection at 6 months for women reporting the 238 new partner visits in the McGill-Concordia Cohort Study was 17.0 percent (95 percent confidence interval: 11.4, 23.0). Of 240 simulations, 54 (22.5 percent) produced expected cumulative incidences that fell within the range of 11.4–23.0 percent.

All of the simulations that produced expected incidences consistent with the observed data assumed that men's prevalence was the same as women's, or greater (figure 1). The highest proportion (54 percent) of consistent simulations assumed that men's prevalence was 1.5 times that for women.


Figure 1
View larger version (13K):
[in this window]
[in a new window]
  		FIGURE 1. Frequency distribution of values of the male-to-female human papillomavirus (HPV) prevalence rate ratio in 54 simulated conditions that were consistent with the observed cumulative incidence of any new HPV type among women in the McGill-Concordia Cohort Study, Montreal, Canada, 1996–2001.

 
Figure 2 shows that values of consistent per-act transmission probabilities ranged from 0.05 to 1.00, with a median of 0.40. Per-act transmission probabilities of 0.001–0.025 were not consistent with the observed cohort data. The median per-act probability values were 0.625 when the male-to-female prevalence rate ratio was assumed to equal 1.0, 0.30 when the prevalence rate ratio was assumed to equal 1.5, and 0.10 when the prevalence rate ratio was assumed to equal 2.0.


Figure 2
View larger version (16K):
[in this window]
[in a new window]
  		FIGURE 2. Frequency distribution of human papillomavirus (HPV) transmission probabilities per coital act in 54 simulated conditions that were consistent with the observed cumulative incidence of any new HPV type among women in the McGill-Concordia Cohort Study, Montreal, Canada, 1996–2001.

 
The probability of transmission over a specific number of acts, n, can be estimated with the following equation: Probability (infection) = 1 – (1 – {lambda})n, where {lambda} is the per-act transmission probability (28Go). At the median value of the per-act transmission probability for all consistent simulations (0.40), a woman would have a 99.6 percent probability of becoming infected within 11 acts of intercourse.

Figure 3 shows that no single estimate of per-act effectiveness of condoms produced better-fitting data. Simulations with relative risks ranging from 0.1 to 1 gave expected cumulative incidences that fit the observed data.


Figure 3
View larger version (18K):
[in this window]
[in a new window]
  		FIGURE 3. Frequency distribution of values of the relative risk of the effectiveness of condoms for a protected versus an unprotected coital act in 54 simulated conditions that were consistent with the observed cumulative incidence of any new human papillomavirus (HPV) type among women in the McGill-Concordia Cohort Study, Montreal, Canada, 1996–2001.

 
Figures 4 and 5 show how the relation between the expected 6-month cumulative incidence of any new HPV type and the per-act transmission probability varies with the male-to-female prevalence rate ratio, under the assumption that condoms offer no protection (figure 4) and that they offer fourfold protection (figure 5). The observed 95 percent confidence interval for cumulative incidence in the McGill-Concordia Cohort Study is shown for comparison (dotted area). If one assumes that HPV prevalence is equivalent in men and women, then the per-act transmission probability most consistent with the observed data is greater than 0.20. However, if one assumes that HPV is more prevalent among women than among men, then the per-act transmission probability may be as low as 0.03.


Figure 4
View larger version (26K):
[in this window]
[in a new window]
  		FIGURE 4. Expected cumulative incidence of human papillomavirus (HPV) in a simulated cohort of 10,000 women, by transmission probability per coital act and male-to-female prevalence rate ratio ({diamond} = 0.50, {square} = 0.67, {blacksquare} = 1.00, {blacktriangleup} = 1.50, • = 2.00). The empirically observed 95 percent confidence interval (0.114, 0.230) for the incidence among women in the McGill-Concordia Cohort Study, Montreal, Canada, 1996–2001, is dotted. Results assume no protective effect of condoms.

 

Figure 5
View larger version (26K):
[in this window]
[in a new window]
  		FIGURE 5. Expected cumulative incidence of human papillomavirus (HPV) in a simulated cohort of 10,000 women, by transmission probability per coital act and male-to-female prevalence rate ratio ({diamond} = 0.50, {square} = 0.67, {blacksquare} = 1.00, {blacktriangleup} = 1.50, • = 2.00). The empirically observed 95 percent confidence interval (0.114, 0.230) for the incidence among women in the McGill-Concordia Cohort Study, Montreal, Canada, 1996–2001, is dotted. Results assume that condoms offer fourfold protection.

 
Similarly, figure 6 shows how the relation between the expected 6-month cumulative incidence of any new HPV type and the per-act transmission probability varies with the assumed protective effects of condoms if male HPV prevalence is 1.5 times that of females. All simulated relative risk values for condom effectiveness were compatible with the observed data, but higher condom effectiveness implies higher transmissibility. That is, if the relative risk is 1, then the plausible range for transmissibility is about 0.09–0.40; if the relative risk is 0.1, then this range shifts to 0.16–1.00.


Figure 6
View larger version (28K):
[in this window]
[in a new window]
  		FIGURE 6. Expected cumulative incidence of human papillomavirus (HPV) in a simulated cohort of 10,000 women, by transmission probability per coital act and relative risk of transmission of HPV infection ({diamondsuit} = 0.10, {blacksquare} = 0.25, {blacktriangleup} = 0.50, • = 1.00) for a protected versus an unprotected act. The empirically observed 95 percent confidence interval (0.114, 0.230) for the incidence among women in the McGill-Concordia Cohort Study, Montreal, Canada, 1996–2001, is dotted. Results assume that male prevalence is 1.5 times that among women.

 
The Monte Carlo standard error of the expected incidence was estimated for each of the 240 configurations of parameter values. Consider, for example, the instance in which the male-to-female prevalence rate ratio was 1.5, the relative risk for condom use was 0.25, and the per-act transmission probability value was 0.40. Over 100 replications of a simulation of 10,000 couples, the mean incidence rate was 0.19900. The standard error of the mean, or Monte Carlo standard error, was calculated by dividing the observed standard deviation of the 100 estimates by sqrt(100), and it equaled 0.00047. Over the 240 sets of parameter values, this Monte Carlo standard error ranged from 0.00002 to 0.0007, indicating considerable precision in the simulated incidence rates.

The 95 percent confidence interval for observed incidence was chosen for comparison since it is the conventional level of confidence for the dispersion of parameter values in most decision-making situations in public health. For comparison, other interval boundaries were also used. For instance, the 99 percent confidence interval for observed incidence was 10.6, 25.3; although more (66/240) simulated conditions produced expected incidence values that were consistent with this interval than with the 95 percent confidence interval, the parameter values in those 66 simulated conditions were identical. Comparison with the 50 percent confidence interval (15.0, 18.8) led to fewer (12/240) simulated conditions being consistent with the observed rate. Per-act transmissibility values were in the lower range (0.05–0.30), whereas the male-to-female prevalence ratio was in the higher end of the range (1.5–2.0). The values of the risk ratio for condom use were no different from those obtained using the 95 percent confidence interval.

Further analysis was carried out to determine the influence of specific assumptions. Results were similar when incidence density, rather than cumulative incidence, was used as the comparative outcome (data not shown). The simulations presented above assumed that regular condom use reported by women in the McGill-Concordia Cohort Study indicated use 75 percent of the time; results were similar when regular use was assumed to be 95 percent of the time (data not shown). Finally, random assignment of female HPV positivity at time t – 1, sexual frequency, and condom use frequency assumes that these factors are uncorrelated. To test this assumption, HPV incidence was simulated among the observed 238 new partner visits using reported data on female HPV positivity at time t – 1 and sexual and condom use frequency, and results were similar (data not shown).


   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
References
 
The modeled HPV per-act transmission probabilities that were consistent with observed cumulative incidence among young female university students ranged from a lower limit of 5 percent per act to an upper limit of 100 percent per act. At the median, 40 percent per act, the probability of male-to-female transmission would reach virtually 100 percent with only 11 acts of intercourse. Per-act transmissibility values of less than 5 percent were inconsistent with the observed data.

The results suggest that HPV prevalence among male partners of this university student population in Montreal was equal to or greater than that among women. Other research of HPV prevalence in both sexes of the same university student population has reported slightly less to equivalent prevalence among males compared with females (7Go, 30Go, 31Go). In sexually transmitted disease clinic populations, higher prevalence was observed among males compared with females in Denmark and Greenland (32Go). However, comparison of sex-specific prevalence within the same population assumes that sexual networks are confined to that population. This assumption may not be true if female students have partners outside the student population. Partnership studies would be needed to verify the true infection status of women's partners.

STI transmission dynamics involve three distinct components: 1) transmissibility from an infected to an uninfected partner upon exposure, 2) the likelihood of sexual exposures between infected and uninfected persons, and 3) the duration of the infection (2Go). The first, transmissibility, can be measured empirically only in studies of couples (33Go, 34Go). One such study, conducted by Oriel (35Go), examined the transmission of genital warts before HPV was identified as the causal agent. Participants were patients at a hospital's venereology department in London, England. Sexual partners of the index patient in the 9-month period before and after the appearance of warts were recorded for 97 patients. Sixty percent (53/88) of the sexual partners of the index patients subsequently developed warts, suggesting high transmissibility.

To our knowledge, there have been no published reports of the transmissibility of HPV itself based on data from couples, but it is thought to be high (36Go, 37Go). Unlike most STIs, HPV is not concentrated in "core groups"—small groups of highly sexually active individuals (2Go, 37Go). An epidemiologic pattern of high prevalence among moderately sexually active individuals may result from either a long duration of infectivity and/or high infectivity (37Go). There is evidence that the duration of HPV infection is short for women (3Go, 10Go, 24Go, 38Go), and the same may also be true for men (39Go). This evidence suggests that high transmissibility may explain the observed prevalence in most populations.

The estimated per-act transmission probabilities for HPV in this simulation study were high in comparison with other viral STI but were comparable to those presumed for bacterial STI. Studies of HIV-discordant couples indicate that the probability of HIV transmission is 1 per 1,000 acts of intercourse (28Go). This probability is believed to increase as much as 10-fold with high seminal viral load, which may occur during acute primary infection or when either partner is coinfected with other STIs (29Go, 40Go). Even in such circumstances, the range of plausible HPV per-act transmission probabilities indicates that HPV would still be considerably more infectious than HIV. Similarly, the probability of transmission of herpes simplex virus type 2 is estimated to be 1 per 1,000 acts among stable, long-standing couples (41Go). Transmission probabilities for other STIs are available; however, they are typically reported as the probability of transmission per partnership, not per coital act, and are considered an average across partnerships of varying duration. They range from 20 percent for Chlamydia and 50 percent for gonorrhea (42Go) to 60 percent for syphilis (43Go) and 80 percent for Haemophilus ducreyi, the infectious agent for genital ulcers (42Go). The higher rate of transmission of the latter two agents is related at least in part to the presence of genital ulcers that increase transmission of STI.

The present study used stochastic computer simulation to model HPV transmissibility. Deterministic models have also been developed for HPV, specifically to estimate the population impact of vaccination (36Go, 44Go). Hughes et al. (36Go) assumed a per-partner male-to-female transmission probability of 0.8 based on the epidemiology of HPV. Using Finnish HPV-16 seroprevalence data for calibration, Barnabas and Garnett (44Go) estimated a per-partner male-to-female transmission probability of 0.6 for that type. Both of these values are consistent with the range of per-act transmission probabilities deemed plausible in this simulation study.

The high per-act transmission probability estimated in this simulation study suggests that women exposed to an infected partner would acquire HPV within the first acts of intercourse. Consistent with high transmissibility, neither the frequency of sex nor the number of sex acts was associated with incident HPV infections among women in the McGill-Concordia cohort (data not shown).

This simulation study relied on the accuracy of the measured cumulative incidence of HPV in the Montreal cohort. In any given 6-month period in which women reported a single new partner, and no other partners, the cumulative incidence was 17.0 percent (95 percent confidence interval: 11.4, 23.0). This rate is consistent with that for women starting their first sexual relationship (45Go), where cumulative incidence of any type of HPV was 20 percent at 6 months following the first act of intercourse. A concern in any study of HPV among sexually experienced women is the possibility that "incident" infections may be reactivation of previously latent infections. In the McGill-Concordia Cohort Study, such misclassification would have been uncommon since women were young (aged 18–24 years). Furthermore, the 6-month cumulative incidence among women who reported no sexual activity was nearly five times less, at 3.8 percent, than among women who reported a new partner.

Epidemiologic investigations of HPV also have to contend with sampling variability due to anatomic site chosen for the specimen, collection method, sample processing, and assay error. The McGill-Concordia Cohort Study used accepted methods for cell sampling and HPV testing. Nevertheless, sampling and assay variability is an issue that our simulation work did not address. Such variability is likely to become compounded in studies involving both partners, especially given that the sampling methods for males are evolving. Results from modeling may therefore be complementary to empirical studies of transmission.

Assumptions must be made in any simulation exercise, and this study was no exception. The simulation of couples assumed random mixing of men and women, at least with respect to HPV status. Surveys of sexual behavior show that mixing may not be random; rather, it may tend to be moderately assortative, such that "like" mix with "like" (2Go, 46Go). High rates of HPV even among moderately sexually active populations (37Go) suggest that an assumption of random mixing with respect to HPV status may not be untenable. Nevertheless, if substantial assortative mixing was present, our simulation would have resulted in an underestimation of per-act transmissibility.

This simulation assumed that couples remained together and that no partnerships dissolved. This assumption, if violated, would have led to an underestimate of transmissibility, but this bias was minimized by the short time interval for simulation (6 months). Furthermore, per-act transmission probabilities were presumed constant. It is possible that the risk of STI transmission varies with the number of acts, and future efforts to study transmissibility should examine this issue (47Go). The random assignment of female HPV positivity at time t – 1, sexual frequency, and condom use frequency in the hypothetical couples presumes that these variables are uncorrelated. Such correlations were not influential when they were simulated, nor did analysis of the cohort itself reveal correlation among these variables. It was also assumed that women who reported "regular" condom use had in fact used condoms 75 percent of the time. Regular use was not assumed to indicate 100 percent use of condoms; even among those who always use them, partial condom use can occur (i.e., not applying the condom before insertion, removing the condom sometime during intercourse, and condom breakage or slippage). As many as 38 percent of young heterosexual condom users report delaying application of the condom at least occasionally (48Go–50Go). Nevertheless, when the simulations were repeated assuming that regular use indicated use 95 percent of the time, the results were nearly equivalent.

Whether or not condoms provide any level of protection against HPV transmission remains a subject of debate (51Go). In vitro studies demonstrate that latex condoms are impermeable to all known sexually transmitted pathogens (52Go), although they cannot protect the entire surface of the genital epithelium from infection. HPV research has found equivocal results (27Go, 51Go). A paradoxical effect is occasionally reported, such that condom use appears to increase risk of HPV infection (5Go, 51Go, 53Go). Methodological issues that have limited the evaluation of condom effectiveness include imprecise measurement and the inability to distinguish with whom participants use condoms or the infection status of that partner (51Go, 54Go).

A critical implication of high transmissibility found in this simulation is that condoms may not offer effective protection over multiple acts of intercourse, which could explain an absence of observed effects in many empirical studies. A protective effect of condoms, even if one exists, is virtually lost with high infectivity (55Go). Simulated conditions in this study showed that high per-act transmission probabilities result in substantial transmission, even with a 10-fold protective effect of condoms. Although condoms may offer protection in relatively brief encounters involving few acts of intercourse, they would be ineffective in partnerships where multiple sex acts occur in an ongoing relationship. For example, if the true per-act transmission probability is 40 percent, transmission occurs within 11 acts of intercourse. If condoms reduce risk of transmission by half to 20 percent, then transmission would occur within 24 acts, which is within about 10 weeks according to the intercourse frequency reported by women in the McGill-Concordia Cohort Study.

That the simulation was unable to provide an estimate of the effect of condoms leads to study design considerations for observational studies of transmission among couples. It would not be possible to obtain more specific estimates of condom effectiveness using the McGill-Concordia study design, which was similar to other longitudinal studies of HPV in young women (3Go, 4Go, 7Go, 38Go). To obtain an estimate of the relative risk of infection for a protected versus an unprotected act, and to finely distinguish it from the per-act transmission probability, one would need to compare couples who always used condoms correctly with couples who never used them. Alternatively, one could study couples who engaged in few acts of sexual intercourse.

Fortunately, HPV vaccines are a promising alternative to condoms. Preliminary evidence from proof-of-principle trials shows great promise for vaccines against HPV-16 alone (56Go), HPV-16 and -18 (57Go), and HPV-6, -11, -16, and -18 (58Go). The findings of this simulation study provide a strong rationale for maximizing coverage of an HPV vaccine upon licensure and for considering the benefits of extending vaccination to young men before they engage in sexual activity. A second implication is that high transmissibility will magnify the impacts of poor vaccine coverage, poor "take," or waning of immunity over time. Close monitoring of population coverage and vaccine effectiveness over time will be necessary. A first generation of validated natural history models has been used to assess the potential impact of changes in these parameters on long-term vaccine efficacy (21Go–23Go). However, these Markov models have been built exclusively on the basis of probabilistic assumptions consistent with findings from epidemiologic studies of the natural history of HPV and cervical neoplasia in women. The approach described here may provide the HPV transmissibility framework that could be incorporated into these models to enhance their ability to make projections of vaccine efficacy under a wider range of scenarios than has been possible with the first-generation models.


   ACKNOWLEDGMENTS
 
The Canadian Institutes of Health Research (CIHR) provided support for the McGill-Concordia Cohort Study (grants MT-13649 and MOP-53111 to E. L. F.). A. N. B. is supported through a Richard H. Tomlinson doctoral fellowship to McGill University; F. C. is the recipient of a National Scholar award from the Fonds de la Recherche en Santé du Québec; and E. L. F. is the recipient of a Distinguished Scientist award from the CIHR.

The authors are grateful to research nurses Gail Kelsall and Suzanne Dumais for specimen and data collection.

Conflict of interest: none declared.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 References
 

  1. Human papillomaviruses. IARC monographs on the evaluation of carcinogenic risks to humans. Vol 64. Lyon, France: International Agency for Research on Cancer, 1995.
  2. Aral SO, Holmes KK. Social and behavioral determinants of epidemiology of STDs: industrialized and developing countries. In: Holmes KK, Sparling PF, Mardh P, et al, eds. Sexually transmitted diseases. New York, NY: McGraw-Hill, 1999:39–76.
  3. Ho GY, Bierman R, Beardsley L, et al. Natural history of cervicovaginal papillomavirus infection in young women. N Engl J Med 1998;338:423–8.[Abstract/Free Full Text]
  4. Moscicki A, Hills N, Shiboski S, et al. Risks for incidence human papillomavirus infection and low-grade squamous intraepithelial lesion development in young females. JAMA 2001;285:2995–3002.[Abstract/Free Full Text]
  5. Richardson H, Franco EL, Pintos J, et al. Determinants of low-risk and high-risk cervical human papillomavirus infections in Montreal university students. Sex Transm Dis 2000;27:79–86.[Web of Science][Medline]
  6. Sellors J, Karwalajtys TL, Kaczorwski J, et al. Incidence, clearance, and predictors of human papillomavirus in women. Can Med Assoc J 2003;168:421–5.[Abstract/Free Full Text]
  7. Winer RL, Lee SK, Hughes JP, et al. Genital human papillomavirus infection: incidence and risk factors in a cohort of female university students. Am J Epidemiol 2003;157:218–26.[Abstract/Free Full Text]
  8. Hildesheim A, Schiffman MH, Gravitt PE, et al. Persistence of type-specific human papillomavirus infection among cytologically normal women. J Infect Dis 1994;169:235–40.[Web of Science][Medline]
  9. Moscicki AB, Shiboski S, Broering J, et al. The natural history of human papillomavirus infection as measured by repeated DNA testing in adolescent and young women. J Pediatr 1998;132:277–84.[CrossRef][Web of Science][Medline]
  10. Franco EL, Villa LL, Sobrinho JP, et al. Epidemiology of acquisition and clearance of cervical human papillomavirus infection in women from a high-risk area for cervical cancer. J Infect Dis 1999;180:1415–23.[CrossRef][Web of Science][Medline]
  11. Thomas KK, Hughes JP, Kuypers JM, et al. Concurrent and sequential acquisition of different genital human papillomavirus types. J Infect Dis 2000;182:1097–102.[CrossRef][Web of Science][Medline]
  12. Liaw KL, Hildesheim A, Burk RD, et al. A prospective study of human papillomavirus (HPV) type 16 DNA detection by polymerase chain reaction and its association with acquisition and persistence of other HPV types. J Infect Dis 2001;183:8–15.[CrossRef][Web of Science][Medline]
  13. Remmink AJ, Walboomers JM, Helmerhorst TJ, et al. The presence of persistent high-risk HPV genotypes in dysplastic cervical lesions is associated with progressive disease: natural history up to 36 months. Int J Cancer 1995;61:306–11.[Web of Science][Medline]
  14. Liaw KL, Glass AG, Manos MM, et al. Detection of human papillomavirus DNA in cytologically normal women and subsequent cervical squamous intraepithelial lesions. J Natl Cancer Inst 1999;91:954–60.[Abstract/Free Full Text]
  15. Ylitalo N, Josefsson A, Melbye M, et al. A prospective study showing long-term infection with human papillomavirus 16 before the development of cervical carcinoma in situ. Cancer Res 2000;60:6027–32.[Abstract/Free Full Text]
  16. Bosch FX, Lorincz A, Munoz N, et al. The causal relation between human papillomavirus and cervical cancer. J Clin Pathol 2002;55:244–65.[Abstract/Free Full Text]
  17. Franco EL. Primary screening of cervical cancer with human papillomavirus tests. J Natl Cancer Inst Monogr 2003;31:89–96.
  18. Franco EL, Harper DM. Vaccination against human papillomavirus infection: a new paradigm in cervical cancer control. Vaccine 2005;23:2388–94.[CrossRef][Web of Science][Medline]
  19. Fahs MC, Plichta SB, Mandelblatt JS. Cost-effective policies for cervical cancer screening. An international review. Pharmacoeconomics 1996;9:211–30.[Web of Science][Medline]
  20. Myers ER, McCrory DC, Nanda K, et al. Mathematical model for the natural history of human papillomavirus infection and cervical carcinogenesis. Am J Epidemiol 2000;151:1158–71.[Abstract/Free Full Text]
  21. Goldie SJ, Grima D, Kohli M, et al. A comprehensive natural history model of HPV infection and cervical cancer to estimate the clinical impact of a prophylactic HPV-16/18 vaccine. Int J Cancer 2003;106:896–904.[CrossRef][Web of Science][Medline]
  22. Goldie SJ, Kohli M, Grima D, et al. Projected clinical benefits and cost-effectiveness of a human papillomavirus 16/18 vaccine. J Natl Cancer Inst 2004;96:604–15.[Abstract/Free Full Text]
  23. Kulasingam SL, Myers ER. Potential health and economic impact of adding a human papillomavirus vaccine to screening programs. JAMA 2003;290:781–9.[Abstract/Free Full Text]
  24. Richardson H, Kelsall G, Tellier PP, et al. The natural history of type-specific human papillomavirus infections in female university students. Cancer Epidemiol Biomarkers Prev 2003;12:485–90.[Abstract/Free Full Text]
  25. Gravitt P, Peyton CL, Apple RJ, et al. Genotyping of 27 human papillomavirus types by using L1 consensus PCR products by single-hybridization, reverse line blot detection method. J Clin Microbiol 1998;36:3020–7.[Abstract/Free Full Text]
  26. Akritas MG. Bootstrapping the Kaplan-Meier estimator. J Am Stat Assoc 1986;81:1032–8.[CrossRef][Web of Science]
  27. Stone KM, Timyan J, Thomas EL. Barrier methods for the prevention of sexually transmitted diseases. In: Holmes KK, Sparling PF, Mardh P, et al, eds. Sexually transmitted diseases. New York, NY: McGraw-Hill, 1999:1307–21.
  28. Gray RH, Wawer MJ, Brookmeyer R, et al. Probability of HIV-1 transmission per coital act in monogamous, heterosexual, HIV-1-discordant couples in Rakai, Uganda. Lancet 2001;357:1149–53.[CrossRef][Web of Science][Medline]
  29. Wawer MJ, Gray RH, Sewankambo NK, et al. Rates of HIV-1 transmission per coital act by stage of HIV-1 infection, in Rakai, Uganda. J Infect Dis 2005;191:1403–9.[CrossRef][Web of Science][Medline]
  30. Shin HR, Franceschi S, Vaccarella S, et al. Prevalence and determinants of genital infection with papillomavirus, in female and male university students in Busan, South Korea. J Infect Dis 2004;190:468–76.[CrossRef][Web of Science][Medline]
  31. Weaver BA, Feng Q, Holmes KK, et al. Evaluation of genital sites and sampling techniques for detection of human papillomavirus DNA in men. J Infect Dis 2004;189:677–85.[CrossRef][Web of Science][Medline]
  32. Svare EI, Kjaer SK, Worm AM, et al. Risk factors for HPV infection in women from sexually transmitted disease clinics: comparison between two areas with different cervical cancer incidence. Int J Cancer 1998;75:1–8.[CrossRef][Web of Science][Medline]
  33. Aral SO, Peterman TA. A stratified approach to untangling the behavioral/biomedical outcomes conundrum. Sex Transm Dis 2002;29:530–2.[Web of Science][Medline]
  34. Fenton KA, Johnson AM, McManus S, et al. Measuring sexual behaviour: methodological challenges in survey research. Sex Transm Infect 2001;77:84–92.[Free Full Text]
  35. Oriel JD. Natural history of genital warts. Br J Vener Dis 1971;47:1–13.[Web of Science][Medline]
  36. Hughes JP, Garnett GP, Koutsky L. The theoretical population-level impact of a prophylactic human papillomavirus vaccine. Epidemiology 2002;13:631–9.[CrossRef][Web of Science][Medline]
  37. Koutsky LA, Kiviat NB. Genital human papillomavirus. In: Holmes KK, Sparling PF, Mardh P, et al, eds. Sexually transmitted diseases. New York, NY: McGraw-Hill, 1999:347–59.
  38. Woodman CB, Collins S, Winter H, et al. Natural history of cervical human papillomavirus infection in young women: a longitudinal cohort study. Lancet 2001;357:1831–6.[CrossRef][Web of Science][Medline]
  39. Van Doornum GJ, Prins M, Juffermans LHJ, et al. Regional distribution and incidence of human papillomavirus infections among heterosexual men and women with multiple sexual partners: a prospective study. Genitourin Med 1994;70:240–6.[Web of Science][Medline]
  40. Chakraborty H, Sen PK, Helms RW, et al. Viral burden in genital secretions determines male-to-female sexual transmission of HIV-1: a probabilistic empiric model. AIDS 2001;15:621–7.[CrossRef][Web of Science][Medline]
  41. Wald A, Langenberg AGM, Link K, et al. Effect of condoms on reducing the transmission of herpes simplex virus type 2 from men to women. JAMA 2001;285:3100–6.[Abstract/Free Full Text]
  42. Anderson RM. Transmission dynamics of sexually transmitted infections. In: Holmes KK, Sparling PF, Mardh P, et al, eds. Sexually transmitted diseases. New York, NY: McGraw-Hill, 1999:25–37.
  43. Garnett GP, Aral SO, Hoyle DV, et al. The natural history of syphilis: implications for the transmission dynamics and control of infection. Sex Transm Dis 1997;24:185–200.[Web of Science][Medline]
  44. Barnabas R, Garnett G. Assessing the effect of HPV 16 transmission and HPV 16 progression rates on the potential impact of HPV vaccines. (Abstract F-07). Presented at the 22nd International Papillomavirus Conference and Clinical Workshop 2005, Vancouver, British Columbia, Canada, April 30–May 6, 2005.
  45. Collins S, Mazoomzadeh S, Winter H, et al. High incidence of cervical human papillomavirus infection in women during their first sexual relationship. Br J Obstet Gynaecol 2002;109:96–8.
  46. Laumann EO, Gagnon JH, Michael RT, et al. The social organization of sexuality: sexual practices in the United States. Chicago, IL: University of Chicago Press, 1994.
  47. Røttingen J, Garnett G. The epidemiological and control implications of HIV transmission probabilities within partnerships. Sex Transm Dis 2002;29:818–27.[Web of Science][Medline]
  48. de Visser RO, Smith AM. When always isn't enough: implications of the late application of condoms for the validity and reliability of self-reported condom use. AIDS Care 2000;12:221–4.[CrossRef][Web of Science][Medline]
  49. Quirk A, Rhodes T, Stimson GV. "Unsafe protected sex": qualitative insights on measures of sexual risk. AIDS Care 1998;10:105–14.[CrossRef][Web of Science][Medline]
  50. Warner L, Clay-Warner J, Boles J, et al. Assessing condom use practices: implications for evaluating method and user effectiveness. Sex Transm Dis 1998;25:273–7.[Web of Science][Medline]
  51. Manhart LE, Koutsky LA. Do condoms prevent genital HPV infection, external genital warts, or cervical neoplasia? Sex Transm Dis 2002;29:726–35.
  52. Stratton P, Alexander NJ. Prevention of sexually transmitted infections: physical and chemical barrier methods. Infect Dis Clin North Am 1993;7:841–59.[Web of Science][Medline]
  53. Franco EL. Understanding the epidemiology of genital infection with oncogenic and nononcogenic human papillomaviruses: a promising lead for primary prevention of cervical cancer. Cancer Epidemiol Biomarkers Prev 1997;6:759–61.[Free Full Text]
  54. Warner L, Newman DR, Austin HA, et al. Condom effectiveness for reducing transmission of gonorrhea and chlamydia: the importance of assessing partner infection status. Am J Epidemiol 2004;159:242–51.[Abstract/Free Full Text]
  55. Mann JR, Stine CC, Vessey J. The role of disease-specific infectivity and number of disease exposures on long-term effectiveness of the latex condom. Sex Transm Dis 2002;29:344–9.[Web of Science][Medline]
  56. Koutsky LA, Ault KA, Wheeler CM, et al. A controlled trial of a human papillomavirus type 16 vaccine. N Engl J Med 2002;347:1645–51.[Abstract/Free Full Text]
  57. 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 randomized controlled trial. Lancet 2004;364:1757–65.[CrossRef][Web of Science][Medline]
  58. Villa LL, Costa RL, Petta CA, et al. Prophylactic quadrivalent human papillomavirus (types 6, 11, 16, and 18) L1 virus-like particle vaccine in young women: a randomised double-blind placebo-controlled multicentre phase II efficacy trial. Lancet Oncol 2005;6:271–8.[CrossRef][Web of Science][Medline]

Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
International Collaboration of Epidemiological Stu
Cervical Carcinoma and Sexual Behavior: Collaborative Reanalysis of Individual Data on 15,461 Women with Cervical Carcinoma and 29,164 Women without Cervical Carcinoma from 21 Epidemiological Studies
Cancer Epidemiol. Biomarkers Prev., April 1, 2009; 18(4): 1060 - 1069.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
B. Dondog, G. M. Clifford, S. Vaccarella, T. Waterboer, D. Unurjargal, D. Avirmed, S. Enkhtuya, F. Kommoss, N. Wentzensen, P. J.F. Snijders, et al.
Human Papillomavirus Infection in Ulaanbaatar, Mongolia: A Population-Based Study
Cancer Epidemiol. Biomarkers Prev., July 1, 2008; 17(7): 1731 - 1738.
[Abstract] [Full Text] [PDF]

Home page
 Cancer Epidemiol. Biomarkers Prev.Home page
A. R. Giuliano and D. Salmon
The Case for a Gender-Neutral (Universal) Human Papillomavirus Vaccination Policy in the United States: Point
Cancer Epidemiol. Biomarkers Prev., April 1, 2008; 17(4): 805 - 808.
[Full Text] [PDF]

Home page
 CMAJHome page
E. L. Franco MPH DrPH, A. de Pokomandy MD, A. R. Spence MSc, A. N. Burchell MSc, H. Trottier PhD, M.-H. Mayrand MD MSc, and S. Lau MD
Vaccination against human papillomavirus
Can. Med. Assoc. J., December 4, 2007; 177(12): 1524 - 1525.
[Full Text] [PDF]

Home page
 Epidemiol RevHome page
E. J. Dasbach, E. H. Elbasha, and R. P. Insinga
Mathematical Models for Predicting the Epidemiologic and Economic Impact of Vaccination against Human Papillomavirus Infection and Disease
Epidemiol. Rev., August 1, 2006; 28(1): 88 - 100.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
163/6/534    most recent
kwj077v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (29)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Burchell, A. N.
Right arrow Articles by Franco, E. L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Burchell, A. N.
Right arrow Articles by Franco, E. L.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?