Revisiting the HPV vaccine

[danaeris]

Does the HPV vaccine really work?

I've been meaning to look into this since I had a conversation with a very sexy doctor friend (*winks*) about the pros and cons of Gardasil, the HPV vaccine.

The conversation called the efficacy of the vaccine into question, which left me wanting to learn more about how they tested the vaccine, both for safety and efficacy.

Here's what I found:
- in women ages 9-26, they found that the vaccine induced "high and persistent anti-HPV antibody titres"
- they followed more than 20,000 women who had been administered the vaccine for two years, and found "the vaccine was highly effective in preventing cervical dysplasia of any grade and external genital lesions related to HPV types 6, 11, 16 and 18 infection."
http://www.ingentaconnect.com/content/adis/dgs/2006/00000066/00000009/art00008

Sounds fair enough, except for the fact that the incubation period is MUCH longer than 2 years.

"The incubation period for HPV 16 to give rise to cervical cancer is quite long and rather variable. It has been estimated that among HPV 16-positive women, the median incubation period from first detected HPV infection to cervical carcinoma in situ was between 7 and 12 years."
http://www.stanford.edu/group/virus/papilloma/2004goglincarnevale/Papilloma/HPV16.htm

I was unable to find the incubation period for the other three strains... but how can they claim to be effective after only following these women for two years, when one of the strains they are claiming to prevent has an incubation period of 7 to 12 years... I couldn't say. It seems rather ridiculous. (I'm wondering if there are separate incubation periods, and they are talking about different ones here... even so, progression to SIL, which is the next stage towards cervical cancer, takes about 5.6 years on average, which barely puts them in the zone for the HPV-16 vaccine... IF that's what these numbers mean.)

And, like with all approved drugs, we know that the short term safety is good, but we don't know what the long term safety is, and won't know until it has been on the market for a good long time.

I'm not saying that the vaccine WON'T protect you. But I AM saying that you don't know for sure, and we won't for another decade or so.

Given the cost of the vaccine, I know I'm going to have to think carefully about whether or not to get it. In the long run, I imagine it will afford some resistance to HPV, which seems worthwhile for someone like me. But there is no guarantee.

Comments

[dagibbs.livejournal.com]

I think you are comparing apples and oranges. You are comparing the time from HPV -> actual cancer to what they are looking at: HPV vaccine -> symptoms of HPV (genital warts) and cervical dysplasia (the appearance of unusual cells in a pap smear). A quick look at cervical dysplasia, suggest that the time from first appearance of cervical dysplasia -> actual cancer is about 10 years. So, if they have 2 years -> CD + 10 years CD -> cancer, that does cover the 12 years (out end median) for HPV -> cancer in the 2nd value.

Now, assuming that the HPV causes cancer link is already "proven", they seem to have provided good evidence the vaccine reduces/prevents HPV (of the named types), and therefor should also reduce/prevent cancer.

[danaeris.livejournal.com]

*sigh* That's why I put in brackets that thing about incubation periods.

I will not feel that this is clear to me until I've actually done full reporting (i.e. talked to the actual company that made the vaccine, or to some other relevant expert).

The fact of the matter is that a doctor who I've known to actually know his shit told me that there's no evidence that this vaccine actually works. So don't be so sure that you know what's up with it until the actual reporting has been done.

[johngnassi.livejournal.com]

Ah, but even if the only way to get cervical cancer were to go through the steps of HPV -> warts -> dysplasia -> cancer over a period of a decade, and even if the vaccine stopped all HPV infections, it's still not clear if it's a good idea to use the vaccine. Because the outcome that you are looking for is not just avoidance of cancer, but living a long and healthy life.

What if the vaccine works but it causes problems? Then it may "work", but...

A classic example of looking at a surrogate outcome that was the wrong outcome to look at was the drug flecainide, used to treat premature ventricular contractions (PVCs). Sometimes people die suddenly, and relatively unexpectedly, from PVCs that worsen to a lethal rhythm (ventricular fibrillation, VF). People with frequent PVCs were known to die more frequently than people who didn't have them.

So the cardiologists, when I was med student, were putting lots of patients on flecainide, which made the heart less irritable and decreased the number of PVCs in patients who had a lot of them. They figured that they were doing a good thing. They hadn't done studies to make sure that people who had frequent PVCs and took the drug had better longer lives.

So, finally someone with enough skepticism did do the right study. One in which some people got the drug and some people got placebo. What did they find out? That the drug actually did work; it suppressed the PVCs. BUT people who took the drug DIED more frequently than the ones who didn't. What was going on?

Well, by the surrogate measure, the drug worked: it decreased the frequency of PVCs. And we knew that in unmedicated people, increased frequency of PVCs increased the death rate.

But many smart cardiologists hoped flecainide would be a good treatment for sudden cardiac death. They were so afraid of death that they emotionally felt that they had to do something. But the stupid part was that they didn't check to see if what they were doing was overall beneficial to the patient - they looked at just the surrogate marker (frequency of PVCS) instead of something that really counts - long and healthy life. In their lust to stop death and cure disease, many had the hubris to believe they knew enough about the pathophysiology of the problem that they could just tweak something here or there... and voila! - save lots of people from an early death.

Life isn't that simple. What they were actually doing was killing people even faster by putting them on the drug. What went wrong?

Well, they didn't admit the limits of their knowledge - that they had an idea that something might be good, but they didn't admit to themselves that the treatment was still experimental. So they pushed it on people who trusted them. And killed them with their arrogance.

Why did this happen? Well, in further studies it became clear that flecainide stabilized cell membranes - that is, it reduces the chance of changing from one rhythm to another, say from frequent PVCs to normal sinus rhythm (NSR, which is healthy). But what they didn't realize is that it also inhibited the change from PVCs to NSR. And they didn't understand the natural history of the disease, that short runs of PVCs changed back spontaneously to NSR. Unless there was flecainide in the mix - which kept the PVCs going and increased the chance of death. So, while the drug decreased the chance you would go into a bad rhythm which could degenerate into a lethal rhythm, it also decreased the chance that once you went into that rhythm that you could spontaneously get out of it. And that turned out to be the bigger effect.

In the case of the HPV vaccine, we're still in the surrogate marker of disease state of knowledge. We don't know if it will eventually prevent cervical cancer years down the road, or cause another bad problem... or both... or neither!

[johngnassi.livejournal.com]

So... what I've usually have been discussing with patients about HPV, and the one vaccine available for it, is the limits of knowledge about how it works, and how little is known about the long term effects. Some people, who have sex more frequently with different people, or for some reason believe that they are at increased risk of getting HPV, might want to take a chance on the vaccine improving their lives. Others might not. But at least they understand they are making an emotional decision, not one based on definitive science. In ten years we'll have more information and maybe then can say. But until then, all we will have is surrogate marker of disease and the imperfect conclusions derived from them.

Life isn't simple, is it? One thing I frequently remind myself of is, that when you don't really know what the better thing to do it, you can do either with complete complacency. If you get it, just realize no one knows if it will predispose you to a problem in a few years, before it helps keep you from getting a cancer. After patients get this, I'm fine with them making whatever choice keeps them away from the thing they're more afraid of.

Ideally, it would be ideal to maintain a registry of people who were willing to be randomized to placebo or vaccine, and follow them for 10 years to see which group did better overall. But I don't know of any such trial yet, so we're back to where we were with the flecainide :).

[danaeris.livejournal.com]

You clearly are saying that we don't know if it is SAFE. But do you ALSO question its efficacy?

[johngnassi.livejournal.com]

Yes. Even looking at just efficacy, we don't know if it will decrease cervical cancers. It has not been studied long enough to show that.

[johngnassi.livejournal.com]

What we do know is that almost all cervical cancers co-occur with HPV (but remember that correlation is not causality):

"Clinical question
Does HPV-negative cervical cancer ever occur?

Bottom line
In this group of more than 500 women with biopsy results showing cervical intraepithelial neoplasia (CIN) grade 3 and cervical cancer, human papillomavirus (HPV) DNA testing was positive in more than 99% of cases. Patients with negative HPV DNA testing are very unlikely to have cervical cancer. (LOE = 1b)

Reference
Bohmer G, van den Brule AJC, Brummer O, Meijer CJLM, Petry KU. No confirmed case of human papillomavirus DNA-negative cervical intraepithelial neoplasia grade 3 or invasive primary cancer of the uterine cervix among 511 patients. Am J Obstet Gynecol 2003; 189:118-20."

Add to that that the HPV vaccine does, over a 2 year period, reduce HPV infections in women who have had no more than 6 sex partners previously:

"Clinical question
Is a vaccine effective against human papillomavirus strains associated with cervical cancer?

Bottom line
A bivalent vaccine against human papillomavirus (HPV) types 16 and 18 is well-tolerated and effective in reducing HPV infection and HPV-associated cytologic abnormalities. What we need now is a larger, longer-termed, adequately powered study to look at the effect of this vaccine on the development of cervical cancer. (LOE = 1b)

Reference
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 randomised controlled trial. Lancet 2004; 364:1757-65.

Study design: Randomized controlled trial (double-blinded)

Allocation: Concealed

Setting: Population-based

Synopsis
This team of researchers randomly assigned healthy women aged 15 to 25 years with no more than 6 sexual partners and no history of condyloma or cervical cancer to receive a bivalent vaccine active against HPV serotypes 16 and 18 or placebo. They administered the vaccine or placebo at 0, 1, and 6 months. They evaluated the patients after 27 months to determine the presence of HPV infection or cytologic abnormalities. Using an intention-to-treat approach to these outcomes, the vaccine was 95% effective against persistent HPV infection and 93% effective against cytologic abnormalities associated with HPV. In the intention-to-treat analysis, the absolute reduction in new HPV infections was 6.4% (number needed to treat [NNT] = 16) and 3.5% for persistent infections (NNT= 29). Other than local injection site symptoms, there were no differences in side effects between the active and placebo vaccines."

If these results hold up over time (i.e. a 2% per year reduction in chronic HPV infection, although I don't know what even that means for the development of clinical cancers yet), they could be *very* significant.

[johngnassi.livejournal.com]

Another study just looking at an HPV16 vaccine showed excellent results in that narrow domain:

"Clinical question
Does human papillomavirus-16 vaccination prevent cervical intraepithelial neoplasia?

Bottom line
For young women who have not been previously infected with human papillomavirus-16 (HPV16), vaccination prevents HPV16-related cervical intraepithelial neoplasia (CIN). (LOE = 1b)

Reference
Mao C, Koutsky LA, Ault KA, et al. Efficacy of human papillomavirus-16 vaccine to prevent cervical intraepithelial neoplasia. Obstet Gynecol 2006;107:18-27.

Study design: Randomized controlled trial (double-blinded)

Allocation: Concealed

Setting: Population-based

Synopsis
HPV16 is responsible for approximately half of all cervical cancers. The vaccine studied here consists of 40 mcg purified capsid polypeptide specific to this virus. Nearly 2400 women, aged 16 years to 23 years, were randomized to 3 injections of vaccine or placebo, administered at enrollment, month 2, and month 6. All women underwent cervical testing at enrollment, which included sampling for cytology; tests for HPV16 DNA, chlamydia, and gonorrhea; and wet mount to detect bacterial vaginosis or trichomoniasis. Serum samples for HPV16 antibody titers were obtained at enrollment and serially afterward. Women with abnormal cytology were referred for colposcopy and all women underwent colposcopy at 48 months. The key analysis was of women who were seronegative and cervical HPV16-negative at study entry and who received at least one dose of vaccine. Among those women were 32 cases of HPV-related CIN identified during the study, all in the placebo group (0.0% vs 1.1%; 100% efficacy; 95% CI, 88-100). There were no significant differences in losses to follow-up between vaccine and placebo groups (70% completion of 4-year follow-up in each group)."

But for this stunning result they analyzed the subgroup of people who were already at a lower likelihood of developing HPV - those who were (-) at the start, which will include lucky people, people with lower rates of exposure, and those with some native resistance. By focussing on the people who already were less likely to get HPV (and just one subtype of HPV), the analysis would tend to make the treatment look better than it is. An even then, they only found a difference of 1% between placebo and control groups for a surrogate marker of cancer, and that after 4 years.

So, it may turn out that vaccination is great strategy for reducing cervical cancers, but the jury is really still out, and thus the treatment is still experimental.

[dagibbs.livejournal.com]

That makes sense.

So what we have is a treatment that some THINK is likely to help. But we don't know for sure. How much study do we need to do in order to decide? If it really does work, how long have we waited being careful?

So, this isn't conclusive -- just suggestive. Do we try it, and continue to monitor with long-term studies? Or just wait until we know enough... 10 or 20 years?

[johngnassi.livejournal.com]

Studies!! Enroll people in studies! That's how these things get figured out.
That and waiting maybe 7-10 years.

[velvetpage.livejournal.com]

That's the progression to actual cervical cancer. Abnormal pap smears appear much earlier, often within a year or two.

[danaeris.livejournal.com]

As posted above...
I will not feel that this is clear to me until I've actually done full reporting (i.e. talked to the actual company that made the vaccine, or to some other relevant expert).

The fact of the matter is that a doctor who I've known to actually know his shit told me that there's no evidence that this vaccine actually works. So don't be so sure that you know what's up with it until the actual reporting has been done.

[velvetpage.livejournal.com]

Fair enough. My daughters won't hit puberty for close to a decade. By then, some long-term information should be available.

[payoolay.livejournal.com]

They gave it to my daughter in grade 7 this year, along with some other mandatory vaccines they get at that age (there was a form to sign).

[echan.livejournal.com]

All the advertisements I've seen (here in Phoenix, Arizona, USA) have been very careful to say that it only prevents against HPV, which causes some cervical cancers. However, where I just said 'causes' seems to be somewhat contradicted by something you quoted: "highly effective in preventing [stuff] related to HPV types 6, 11, 16 and 18 infection." So, perhaps their claims are true, if you accept their claim of a direct causal relationship between HPV and cancer.

HPV vaccine in texas

[admiralthrawn.livejournal.com]

As an interesting follow-up to this discussion, it was announced friday that the state of Texas will begin requiring this vaccine for middle school and high school girls.

Interestingly, the reactions seem to be split between "This is horrible, they're encouraging middle school girls to have sex!" and "This is horrible, they're clearly being bribed by big pharma!". I've read several articles on it, and nobody but you has raised the issue of whether it is actually proven effective.

The lenghts of studies you cite also raise the question of how long the protective effect lasts; do we know yet whether you will need booster vaccines after N years to keep the useful level of protection? That question seems hard to answer with two-year studies, even if you accept that preventing HPV will prevent the cancers.