Effectiveness of HPV vaccines: Reporters ought to read their own stories

An LA Times report on a study of the HPV vaccine summarizes it by saying "Overall, the new results indicate that the vaccine is not living up to its initial prospects." But is that true? Here's what the reporter said mere paragraphs earlier about the findings:

Among women who had not previously been exposed to types 16 and 18, the vaccine reduced the risk of precancerous lesions caused by those two strains by 98%.

"The overall message, in my mind, is that among susceptible young women, the vaccine was highly effective in preventing HPV-16 or -18 precancerous cervical lesions," [study author] Koutsky said.

When the researchers included all the women enrolled in the study, the vaccine reduced the risk of lesions caused by types 16 and 18 by 44%. [The vaccine only targets two out of many strains of HPV, and other strains can also cause cancer. The two targeted strains are among the most widely found in cervical cancers.]

The number of women who were previously infected by the viruses was not large enough to account for this low rate of protection. [I don't know what this is supposed to mean, and I don't see anything in the paper that clarifies it. -TfK]

But when Koutsky and her colleagues considered lesions caused by all strains of the virus, the vaccine reduced the risk by only 17%.

Because researchers had previously believed that 50% of all serious precancerous lesions were caused by types 16 and 18, this rate of protection seemed inexplicably low. [Again, i don't see this in the paper. "Inexplicable" to whom? -TfK] …

Koutsky noted, however, that 17% was an improvement over the 13% figure that was presented to the FDA more than a year ago, which resulted in the administration's approval.

So when administered to uninfected women, it almost totally prevents lesions. Among infected women, it reduces lesions over all and especially those caused by the two strains of the virus that the vaccine is meant to block. As the text I bolded shows, this is statistically the same as what the clinical trials showed. How can matching the expectations exactly mean "not living up to its initial expectations"?

The part of the article I was most interested in was this:

Sawaya suggested in the editorial he co-wrote that the small size of the protection [the 17% figure above] could be because other strains of HPV are filling the gap when types 16 and 18 are eliminated.

Dr. Douglas Lowy of the National Cancer Institute, who originally developed Gardasil, noted that even if the other types proliferate, there might not be a significant increase in cancers because the other strains are less carcinogenic.

If the other strains are, in fact, proliferating, USC's Kast said, "then the addition of other HPV types to the vaccine would be a logical option." Merck and GlaxoSmithKline are working on that.

I study competition, after all, and competition among virus strains would be a cool little tidbit. But when I looked in the original paper, it explained:

Speculation that elimination of HPV-16 and HPV-18 will open a niche for other high-risk viruses is not supported by the literature. Young women are often infected with multiple high-risk HPV types, and the risk of new infection is greater for women who are infected with one or more HPV types than for uninfected women. In addition, levels of viral DNA in clinical specimens may be higher for one HPV type when there is coinfection with another type.

In other words, viral infections do not compete with one another, but facilitate additional infection (if anything). The idea that blocking one virus opens up opportunities for others is unlikely.

What seems more likely is that a woman would be infected with multiple strains simultaneously. As they point out, most infected people have multiple strains. So the woman's partner probably had more than just HPV-16 or HPV-18, there were probably other strains. The vaccine blocks those two strains, but not the others. If she is susceptible to HPV-caused cancer, those other strains do what HPV-16 or HPV-18 would have done otherwise. This may seem like a fine distinction, but I don't think it is.

Gardasil, after all, is not a cancer vaccine, it's a virus vaccine. It blocks viral infections. If blocking one infection allowed a different infection to replace it, you really wouldn't have done anything, and the vaccine would be worthless. But in this case, the vaccine works as expected against the target virus. We knew before that there were other cancer-causing strains of HPV, and that many people are infected with multiple strains. This result is not surprising in that sense.

What these results do show is that vaccination ought to be done early, well before the onset of sexual activity. A study in the same issue of the New England Journal of Medicine shows that oropharyngeal cancer is probably associated with HPV transmission due to oral sex, which means the vaccine should be administered not just before kids initiate vaginal intercourse, but before oral intercourse and other experimentation.

As an opinion piece in the same issue points out, "access to these vaccines has already become more a political than a public health question." Perhaps Matt Nisbet and Chris Mooney will address that side of the issue in this evening's presentation at the Stowers Institute.