A recent article asking the question ‘will HPV vaccination prevent cervical cancer” by Claire Rees and co-authors from Queen Mary Hospital and Newcastle University in the UK raises important issues about the interpretation of trials, but should not be read as doubting whether we should use the HPV vaccine. The authors are correct that although HPV infection causes cancer of the cervix, it is too early to be able to say whether the HPV vaccine will prevent cervical cancer. This simply reflects the natural history of the disease when it can take up to 10 years from exposure to HPV to the development of the precancerous CIN3 (cervical intraepithelial neoplasia 3) and then a further 10 years for cancer of the cervix to develop. So, the vaccine has not been in use for sufficient time to know whether it will prevent cancer of the cervix.
The strategy which must be adopted in the early evaluation of the impact of the vaccine is similar to that used in trials of new treatments where it can take many years to see a survival advantage. That is to seek earlier endpoints that will act as surrogate markers for our primary distant endpoint. In the case of HPV vaccination, the endpoints have been measuring the incidence of the precancerous CIN, given that HPV infection can trigger this change.
Women who have been vaccinated should know the that reason for continuing to screen them relates to the uncertainty of the success of vaccination until more time has passed.
What are the reasons that our current trials may be inaccurate in predicting the efficacy of HPV vaccination against cancer of the cervix? For starters, the vaccines only protect against the common HPV oncogenic subtypes, and not all that are known to be oncogenic. That may not be vital if there is cross protection against other subtypes, but alternatively there may be the substitution of other oncogenic subtypes that are not covered by the vaccine.
Most infections with HPV are transient and are not associated with progression to cancer of the cervix. If in measuring the incidence of CIN this includes CIN1 which can come and go and where persistent infection is only associated with a 1% rate of progression to cancer of the cervix, rather than CIN3 which is more likely to progress to cancer, then there may be a false idea of the potential efficacy of the vaccine. Also, if the trial population does not match the population in general the trial result will not reflect the real-world outcomes. In the trials that were reviewed, the authors are concerned that the population studied is an older group of women than the general population.
Arguably, the major reason for expressing concerns over not yet knowing the efficacy of HPV vaccination stems from the great success of the Pap smear program of population screening. Now replaced by testing for HPV, this prevents cancer of the cervix by identifying and treating the potentially precancerous lesions and vaccinated women must still participate in this program, given that we do not know how the vaccine will perform over time. That is certainly the foremost public health message that should accompany the introduction of HPV vaccination. Women who have been vaccinated should know the that reason for continuing to screen them relates to the uncertainty of the success of vaccination until more time has passed. Also, the design of future trials should try to match the trial population with the general population and more critically select surrogate endpoints.
The second message should be that the introduction of the HPV vaccination should be accompanied by monitoring of the population to measure in the real world what impact the vaccine has over time on the development of cancer of the cervix.