Several studies now have found that oral sex is linked to HPV, even when many confounding variables (such as tobacco and alcohol, number of sexual partners, etc.) have been adjusted; specifically, by examining separate samples having different health status. This is probably a better method than holding constant health status in a multiple regression equation (see here).
The main idea is that the human papillomavirus (HPV) is commonly spread via oral sex. HPV increases the probability of getting an oral cancer. This is true whether women practice oral sex on men or men practice oral sex on women. I has also been found that open-mouthed kissing is associated with greater HPV prevalence, although the researchers may acknowledge that it is difficult to isolate the effects because these variables are all correlated.
(last update : November 2014)
A review of the relevant studies.
Here’s the result from D’Souza et al. (2007) study.
The analysis involves logistic regression, and the main focus is on the odds ratio (OR), which is a measure of relative risk. The OR is calculated as OR=(a/c)/(b/d) or equivalently OR=(a*d)/(b*c), where, a is the number of exposed cases, b the number of exposed non-cases, c the number of unexposed cases, d the number of unexposed non-cases. An OR of say, 0.25 is equivalent to say that the event occurs in one case out of four cases. A value greater (lower) than 1 is indicative of higher (lower) prevalence.
Judging by its values that the odds of having an oropharyngeal cancer increases drastically similarly for all patients and patients having oral HPV type-16 infection. As the number of vaginal- and oral-sex partners increases, so does the odds of having an oropharyngeal cancer. For example, compared to having 0-5 vaginal-sex partners, having 6-25 and 26+ partners are associated with ORs of 2.2/2.7 and 3.1/4.2, respectively. The risk is thus increased by two to threefold.
When the sample is restricted to patients with an HPV-16–positive tumor, the ORs are all greater. This means that HPV-16 status does not moderate this relationship.
The association between HPV-16 exposure and oropharyngeal cancer was also investigated among patients with varied use of tobacco and alcohol. The association was greatly increased among patients without a history of smoking or drinking who were seropositive for HPV-16 L1 (odds ratio, 44.8; 95% CI, 5.9 to 338.5) or had an oral HPV-16 infection (odds ratio, 43.7; 95% CI, 4.2 to 452.7). So, smoking and drinking can’t play a role.
To evaluate whether combined exposure to HPV and tobacco (or alcohol) may further increase ORs of oropharyngeal cancer, they calculate the Synergy index. If greater than 1, the index indicates the presence of interaction. But such was not the case (Table 4). Furthermore, among those patients being seropositive for HPV-16 L1 protein, tobacco and alcohol and their interaction were not associated with ORs greater than 1 (but instead, all lower than 1). Inversely, among patients being seronegative for HPV16 -L1 protein, the ORs are truly high. Similarly, the patients being negative (positive) for oral HPV-16 infection have (do not have) ORs associated with tobacco and alcohol that are larger than 1. Consequently, the authors believe that tobacco and alcohol were important risk factors for oropharyngeal cancer, but they may not have acted as cofactors in HPV-mediated carcinogenesis in the oropharynx.
In their multivariate analysis, oropharyngeal cancer was independently associated with HPV-16 L1 seropositivity (OR, 32.2), poor dentition (OR, 4.1), infrequent toothbrushing (OR, 6.9), history of squamous-cell carcinomas of the head and neck in a first-degree family member (OR, 5.4), and heavy tobacco use (OR, 2.5) after adjustment for age, sex, and alcohol use. These factors were collectively estimated to be responsible for 90% of cases of oropharyngeal cancers (the attributable risk; 95% CI, 72 to 96), with 55% of cases (95% CI, 45 to 63) attributable to HPV-16 exposure alone.
In the end, the role of HPV in the cancer is robust to health status. And the authors add :
Although HPV-16 alone accounts for more than 90% of cases of HPV-positive squamous-cell carcinomas of the head and neck, 8 a more accurate and probably higher proportion might be found by testing for other types of HPV (e.g., types 18, 31, 33, and 35), which are infrequently detected in oropharyngeal cancers. …
The public health implications of our findings are underscored by the annual increases in the incidence of tonsillar and base-of-tongue cancers in the United States since 1973. 36, 37 The widespread oral sexual practices among adolescents may be a contributing factor in this increase. 38
D’Souza et al. (2009) replicated this conclusion, using logistic regression, on two distinct populations : 332 control patients from an outpatient otolaryngology clinic and 210 college-aged men. A new finding was that open-mouthed kissing may lead to HPV.
Table 1 shows the result for the adult patients. In the univariate analysis, being homo- or bi-sexual is associated with an OR of 1.5. The lifetime number of oral sex partners has an OR of 3.19 and 7.29 (depending on the n. of sex partners), the non-use of oral barrier during oral sex has an OR of 1.92. The lifetime number of vaginal sex partners has an OR of 1.67 and 4.48 (depending on the n. of sex partners). Tobacco has an OR of 0.19 if former user but an OR of 3.05 if current user. Curiously, the intensity of current smoking is associated with an OR much lower than 1. Drinking has an OR of 1.77 if former user and an OR of 1.29 if current user, and the number of days of drinking has no effect on HPV infection. The use of marijuana has an OR of 2.0. In the multivariate analysis (covariates being age, sex, lifetime number of oral sexual partners and current tobacco use), the lifetime number of oral sex partners has an OR of 2.37 and 5.20, while the lifetime number of vaginal sex partners has an OR of 1.23 and 3.91. Tobacco has an OR of 0.21 if former user but an OR of 3.86 if current user.
Table 2 shows the result for the college-aged men. In the multivariate analysis, the adjustments are made for age and lifetime number of vaginal sex partners. Being homo- or bi-sexual is associated with an unadjusted OR of 4.3 and adjusted OR of 15.6. The lifetime (past year) number of oral sex partners has an unadjusted OR of 20.0 (10.8) and adjusted OR of 7.4 (7.9), the non-use of oral barrier during oral sex has an unadjusted OR of 4.4 and adjusted OR of 7.4. The lifetime (past year) number of vaginal sex partners has an unadjusted OR of 2.1 (3.8) and adjusted OR 0.7 (1.5). The lifetime (past year) number of open-mouthed kisses has an unadjusted OR of 8.4 (11.7) and adjusted OR of 9.5 (17.4); the adjustments were made for age and lifetime number of oral sex partners. Some coefficients are probably exaggerated, maybe due to the small number of the cases in some cells.
Another case-control study has been conducted by Pintos et al. (2008) in which DNA testing of HPV and serological assays are both used as markers of HPV infection. The interesting feature is that they looked at the association between HPV-related factors and tonsil-related cancers as well as overall cancers, when the other studies use overall cancers as their dependent variable.
Table 3 shows regressions unadjusted and adjusted for age, sex, schooling, race, religion, language, tobacco smoking, and alcohol drinking. HPV DNA was the marker of HPV infection. Regarding tonsil-related cancers, the unadjusted OR of positive HPV DNA detection (high-risk type) is 23.06 and the adjusted OR is 19.32. Regarding all oral cancers, the respective unadjusted OR and adjusted OR are 6.89 and 4.81. Regarding oral cancers that are not related to tonsil, the unadjusted and adjusted ORs are 2.67 and 2.14. HPV DNA positivity was strongly associated with cancers, especially tonsil-related cancers, but to a lower extent than for high-risk types.
Studies of serological response to HPV may better represent the temporal relationship between the virus and the tumor, since serological response is more a marker of past cumulative HPV exposure rather than current HPV infection. Such analysis was conducted and results reported in their Table 5, which shows the odds ratios of oral carcinoma according to HPV 16 seropositivity. Three models are of interest, the first model (m1) is the crude analysis, the third model (m3) is adjusted for SES, tobacco, alcohol, the last model (m4) is adjusted for SES, tobacco, alcohol and additionally for detection of oral HPV DNA. For the all oral cancers condition, HPV 16 seropositivity has an unadjusted OR of 6.62, adjusted OR of 7.48 and adjusted OR 6.45. For the tonsil-related cancers, the ORs are, respectively, 27.68, 182.27, 99.34. For not tonsil-related cancers, the ORs are, respectively, 2.8, 3.87, 3.93. The low number of cases may make the estimations rather erratic (e.g., when looking at ORs for tonsil-related cancers). Unlike HPV-16, the HPV-18 and HPV-31 seropositivity don’t have a strong relationship with tonsil-related cancers. The HPV 16 association with this cancer did not decrease after adjustment for markers of sexual activity, suggesting that genital infections are not likely to explain the serological response to HPV 16.
Table 6 shows the ORs for serological assay and HPV DNA detection, individually, and combined. We see that the OR for both markers was largely higher than either marker alone. This indicates once more that if researchers use both markers of HPV infection, the correlation will be even higher.
Concerning the issue of reverse causality, they write :
Another potential limitation, intrinsic to case-control studies in which exposure and outcome are assessed at the same point in time, is the uncertainty to confirm whether the exposure precedes the outcome. It is extremely unlikely that reverse causality bias could explain the magnitude of the association between HPV infection and cancers of the palatine and lingual tonsils. Since serological response is more a marker of past than recent infection, the positive association between seropositivity and these cancers – in addition to the biological evidence – further supports a causal link between HPV and these malignancies.
Another study (Pickard et al., 2012) confirmed the link between HPV and sexual practices in a much larger sample using covariate adjustments from multiple logistic regression.
Their Table 2 shows the ORs (adjusted for difference in lifetime number of vaginal sex partners) associated with HPV prevalence. Having smoked cigarettes, drank alcohol, or smoked marijuana, does not seem to be associated with greater odds of HPV. However, the frequency of marijuana use has a great effect on HPV prevalence. Having sexually transmitted infection (STI) and genital warts also have a large effect on HPV prevalence.
The number of open-mouth kissing partners (lifetime) shows a large effect on the unadjusted (OR=5.7) and adjusted coefficient (OR=4.0). The number of oral sex partners (lifetime) shows a large effect on the unadjusted (OR=5.3) and adjusted coefficient (OR=4.0). The number of vaginal sex partners (lifetime) has large effect (OR=3.8) on the unadjusted regression which has greatly diminished (OR=1.6) on the adjusted regression; the coefficient of vaginal sex partners here is adjusted for lifetime number of oral sex partners. They affirm these associations were similar if we adjust for the number of oral sex partners instead of vaginal sex partners. Unless there are some unknown indirect pathways between these variables, it should be concluded that sexual behaviors have strong connection to the transmission of HPV, that are not mediated by health status.
Studies tell us that oral cancer is often discovered when it is too late. The incidence of HPV-16 has increased dramatically over the last decades, although people with HPV do not necessarily develop oral cancer, and it affects men more often than it affects women. Mehanna et al. (2013) report that the overall HPV prevalence in oropharyngeal cancer (OPC) has increased from 40.5% before 2000 to 64.3% in 2000-2004 and to 72.2 in 2005-2009. Before 2000, the rates were 50.7% and 35.3% in North America and Europe, respectively, and, at 2005+, these rates are now 69.7 and 73.1. Thus, Europe has experienced a much steeper increase in HPV. In comparison, there was no increase (but instead a modest or strong decrease) in prevalence of non-OPC (Table 1).
Researchers tested tumor samples from 271 patients with certain types of throat cancer diagnosed from 1984 to 2004. The [HPV type 16] virus was found in only 16 percent of the samples from the 1980s — but in 72 percent of those collected after 2000.
Often oral cancer is only discovered when the cancer has metastasized to another location, most likely the lymph nodes of the neck. Prognosis at this stage of discovery is significantly worse than when it is caught in a localized intra oral area. Besides the metastasis, at these later stages, the primary tumor has had time to invade deep into local structures. Oral cancer is particularly dangerous because in its early stages it may not be noticed by the patient, as it can frequently prosper without producing pain or symptoms they might readily recognize, and because it has a high risk of producing second, primary tumors. This means that patients who survive a first encounter with the disease, have up to a 20 times higher risk of developing a second cancer. This heightened risk factor can last for 5 to 10 years after the first occurrence. There are several types of oral cancers, but around 90% are squamous cell carcinomas. It is estimated that approximately $3.2 billion is spent in the United States each year on treatment of head and neck cancers. (2010 numbers)
HPVs, also called human papillomaviruses, are a group of more than 150 related viruses. More than 40 of these viruses can be easily spread through direct skin-to-skin contact during vaginal, anal, and oral sex. Some types of HPV can cause genital warts. Other types can cause cancers of the penis, anus, or oropharynx (back of the throat, including base of the tongue and tonsils.) HPV types 6 and 11 cause 90 percent of all genital warts.
… Types 16 and 18 are responsible for about 70 percent of all cases of cervical cancer (1). HPV also causes anal cancer. 85 percent of all anal cases are caused by HPV-16. Close to half of vaginal, vulvar, and penile cancers are caused by HPV types 16 and 18 (2). Its only recently, HPV infections have been found to cause cancer of the oropharynx, which is the middle part of the throat including the soft palate, the base of the tongue, and the tonsils.
In the United States, more than half of the cancers diagnosed in the oropharynx are linked to HPV-16 (3). Oral HPV infection is more common among men than women, explaining why men are more prone than women to develop an HPV related head and neck cancer. …
About 1 percent of the population had an HPV 16 infection, with it being five times more common in men than women, correlating with the higher incidence of HPV-related cancer in men than women. … Patients infected with oral HPV type 16 have a 14 times greater risk of developing one of these cancers.
Over all, 10.1 percent of men were infected orally with HPV of one type or another, compared with 3.6 percent of women. The reason for the higher rate among men was unclear. The men in the study tended to have higher numbers of sexual partners than the women, but statistical analyses showed that this accounted for 16 percent of the difference in virus prevalence. One line of speculation was that hormonal differences in women might play a protective role, or that oral sex on women for some reason causes a greater likelihood of transmission.
Nearly one-third of college-aged women who have had just one sexual partner contracted human papillomavirus within one year of becoming sexually active, according to a study to be published this week in the Journal of Infectious Diseases, the Canadian Press/CBC News reports.
For the study, Rachel Winer of the University of Washington and colleagues followed 125 women ages 18 to 22 who had not had sex previously or who first had sex with one male partner in the three months prior to the beginning of the study. The women were asked to keep diaries of their sexual activities and to estimate how many sexual partners their partners had had. The women also underwent a gynecological exam every four months. Researchers stopped collecting data from women who reported beginning a sexual relationship with a second partner.
One-third of the women had become infected with HPV within one year of starting their first sexual relationship, and 50% of the women were infected with HPV three years later, despite the fact they’d still only had a single sexual partner, the study found. The study also found that the rate of HPV was higher among women who estimated that their partners had had at least two previous sexual partners. …
Winer said the study, which was funded by the National Institute of Allergy and Infectious Diseases, “shows that even just with one partner there’s a high risk of infection.” She added that HPV is “different” from other sexually transmitted infections “in that it’s just very common among everyone who’s having sex. So even just being exposed to one partner makes you susceptible to infection,” she said.
Like any other cancers, oral cancers are rare, especially HPV-related oral cancers, but the fact that the “rates of HPV-related throat cancer had risen 225 percent in the previous 16 years” is of great concern. My feeling is that a strong event must have changed sexual behaviors (sexual revolution ?) to explain this increase of HPV-16 prevalence over the last decades. Anyway, this is probably a sad news for white men, but not for asian men. Asians certainly do not think much, if any, about sex. But the contrary is true for white men who generally speak and think about sex virtually all the time : “the median number of young men’s thought about sex stood at almost 19 times per day”.
I found here a laughable comment : “Oh no, hope my girlfriend doesnt read about this”. I would not be surprised if men are much more skeptical than women about these studies but I hope that most men do not think that way. At this point, women should pass this information on.
- D’Souza, G., Kreimer, A. R., Viscidi, R., Pawlita, M., Fakhry, C., Koch, W. M., … & Gillison, M. L. (2007). Case–control study of human papillomavirus and oropharyngeal cancer. New England Journal of Medicine, 356(19), 1944-1956.
- D’Souza, G., Agrawal, Y., Halpern, J., Bodison, S., & Gillison, M. L. (2009). Oral sexual behaviors associated with prevalent oral human papillomavirus infection. Journal of Infectious Diseases, 199(9), 1263-1269.
- Mehanna, H., Beech, T., Nicholson, T., El‐Hariry, I., McConkey, C., Paleri, V., & Roberts, S. (2013). Prevalence of human papillomavirus in oropharyngeal and nonoropharyngeal head and neck cancer—systematic review and meta‐analysis of trends by time and region. Head & neck, 35(5), 747-755.
- Pickard, R. K., Xiao, W., Broutian, T. R., He, X., & Gillison, M. L. (2012). The prevalence and incidence of oral human papillomavirus infection among young men and women, aged 18–30 years. Sexually transmitted diseases, 39(7), 559-566.
- Pintos, J., Black, M. J., Sadeghi, N., Ghadirian, P., Zeitouni, A. G., Viscidi, R. P., … & Franco, E. L. (2008). Human papillomavirus infection and oral cancer: a case-control study in Montreal, Canada. Oral oncology, 44(3), 242-250.