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.