Why aren’t we seeing the effect of vaccination against HPV 16/18 on cervical cancer registration in England?

Photo by @NAO

A vaccination programme against the HPV virus was introduced in England in 2008 for girls aged 12-13 along with catch-up programme for those aged 14-18. Many people expected cervical cancer rates in women aged 20-24 to fall by 2014, when the vaccinated cohorts entered their twenties. However, we do not think that is reasonable.

As can be seen in Table 1 below, the first vaccinated women (those aged 17-18 in 2008) would have been invited for screening for the first time in September 2015. Less than 50% of these women received the vaccine.

The HPV vaccine works by preventing infection with HPV types 16 and 18. Therefore women already infected by the time they are vaccinated, have little or no benefit of the vaccine. It is estimated that in 2008 between 17% and 20% of women aged 16-18yrs in England were infected with the HPV 16/18 virus. Therefore, at most only 40% of those invited for screening in September 2015 (80% of the 50% vaccinated is 40%) will have benefited from the vaccine. The vaccine should prevent about 70% of cervical cancer (maybe slightly more in young women, particularly, as seems to be the case, if the vaccine has partial protection against other types of the virus). Thus, we might expect to see a fall in cervical cancer in 25 year old women in 2016 of about nearly 30% (70% of 40% is 28%).

It will not be until September 2018 that a substantial proportion of the population coming for screening will have been vaccinated before being exposed to the HPV virus.

Table 1. Summary of the school-based HPV vaccine programme in 2008
Vaccine age Birth cohort Coverage Year enter screening
12/13 01 Sept 95 – 31 Aug 97 86% 2020-2022
14/15 01 Sept 94 – 31 Aug 95 68.5% 2019
15/16 01 Sept 93 – 31 Aug 94 68.6% 2018
16/17 01 Sept 92 – 31 Aug 93 41.7% 2017
17/18 01 Sept 91 – 31 Aug 92 38.9% 2016
17/18 01 Sept 90 – 31 Aug 91 47.4% 2015

Interestingly, national cancer statistics for 2016 do show a (20%) reduction in cervical cancer diagnosis at age 20-24 (73 in 2015 and 58 in 2016). Even more surprisingly the number of cancers diagnosed at ages 25-29 also fell from 401 in 2014 to 341 in 2016. These statistics on their own are not sufficient to conclude that the decrease in cervical cancer is directly related to vaccination, particularly since the number of women attending for cervical screening is at a 20-year low and many cancers in women in their early 20s are picked up on screening.

The impact of vaccinating 17-18 year old women on cancer aged 20-24 is likely to be less because it typically takes at least 8 years from HPV infection to cancer, so a women with cancer at aged 22 is likely to have been infected with HPV before she was 17.

Unfortunately, we only have data for women aged 25-29. If the decrease in cancers in women aged 25-29 is all in women aged 25, then the fall would be rather more impressive and would suggest an early impact of HPV vaccination.

Assessing the effect of vaccination on cervical disease and on cervical cancer by looking at routinely reported cervical cancer statistics is going to be further complicated by the upcoming change in the cervical screening programme. At the end of 2019 testing for the HPV virus will replace cytology as the primary screening test in England. HPV testing is a more sensitive test which will mean that we will be detecting more disease at screening. In Australia, where vaccination began in 2007 and where they have achieved higher coverage among their catch-up cohort, a recent data modelling study showed that the introduction of primary HPV testing will lead to an increase in diagnosis of cervical cancers from 2016 to 2018 before rates begin to fall.  Hence, in England, just when we would expect the vaccination effect to be most apparent we may instead see an increase in the number of cancers diagnosed due to the introduction of primary HPV testing.

A similar phenomenon was observed in 2009 when the English reality TV star Jade Goody was diagnosed and subsequently died from cervical cancer. The screening programme expected an increase in women attending cervical screening, what they did not foresee was the huge impact it had on cervical cancer diagnosis. In England in 2008 there were 1808 cases of cervical cancer diagnosed in women aged 20-64, however in 2009 there were 2207; 399 more cancers than the year before! By 2010 diagnosis had decreased to 1846. This phenomenon was short lived and was most pronounced in women under age 40.

Therefore, in writing this I hope to encourage those of you who are watching cervical cancer trends for proof of vaccine efficacy to be patient. Almost certainly just when a decrease in the incidence of cervical cancer should be seen we will see a steep increase. Hence routinely reported cancer incidence statistics will not be an appropriate source to evaluate the effectiveness of vaccination unless cancer incidence is reported by year of birth or vaccination status.

You may be interested in the following related content: HPV vaccine in Scotland

The views expressed are those of the author. Posting of the blog does not signify that the Cancer Prevention Group endorse those views or opinions.

Subscribe to our mailing list to get updates of new posts…



  1. I find that the figures quoted by charities and the media about cervical cancer are blatantly misleading. I have widely read that about 3000 women are diagnosed each year with cervical cancer, but here you are saying there were only 1808 cases in 2008 and 2207 in 2010. A long way off the widely quoted 3000. We are also told that the screening programme saves 5000 lives per year, but other sources say the figure is less than 2000.
    If we are dealing with such small numbers for such a rare disease, why is the public presented with so much misleading information about this cancer?

    • You are right to point out how confusing it is to try and navigate the multiple different numbers of cervical cancers quoted. Careful attention needs to be paid to what people are reporting. In fact, all the numbers you quote above are “correct”. The often quoted 3,000 figure includes cervical cancers at all ages from England, Wales and Scotland (i.e. the UK). In England the number is closer to 2,500 (but varies year on year). In the blog I quote the number of women diagnosed between the age to 20 and 64 in 2009 (1,808) and in 2010 (2,207). In terms of lives saved, the number must be inferred rather than observed which makes things complicated. It is particularly difficult for cervical cancer, because there is lots of evidence that cervical cancer was becoming more common before cervical screening was introduced and that, in the absence of screening, it would have continued to increase. The figure quoted by the screening programme is based on a 2004 publication by Professor Julian Peto and colleagues which compared the death rate prior to screening to the death rate following the introduction of screening in England. The Peto paper assumes that the relatively high-rates of cervical cancer seen in young women in the 1980s and 90s would have led to extremely high rates when those women reached their 40s, 50s and 60s. The 2,000-figure quoted by others is based on a 2016 publication Dr Rebecca Landy and colleagues which uses incidence data (a non-conventional approach) to estimate the number of lives saved. The more conservative figure comes from estimating the effect of screening by comparing screening in women with and without cervical cancer, rather than by looking at trends over time. It is hard to know which one is more accurate, but in my opinion the newest estimation being more conservative probably reflects the current number of lives saved each year by the screening programme more accurately.

      To estimate how many lives are saved we have to estimate what would have happened had there been no screening. In 2016 there were 854 cervical cancer deaths in the UK corresponding to a rate of 2.6 per 100,000 women per year. In 1971 the rate was 10.5 per 100,000. That rate today would yield 3449 deaths – 2595 more than actually occurred – but we don’t know that all of the fall is due to screening (there have also been advances in treatment) and we don’t know that without screening rates would not have been even higher (we think they would have been because we know that the infection that causes cervical cancers has become more common).

      You may not think of 2000 lives per year as very many – but even that is 40,000 over the last 20 years and given the ages involved it is likely that most of those 40,000 women are still alive today. Potentially something like 1 in 300 women alive today aged 50+ would be dead had they not been for cervical screening. We think those are not small numbers.

  2. You said above: “Interestingly, national cancer statistics for 2016 do show a (20%) reduction in cervical cancer diagnosis at age 20-24 (73 in 2015 and 58 in 2016). Even more surprisingly the number of cancers diagnosed at ages 25-29 also fell from 401 in 2014 to 341 in 2016.”

    This is very misleading. In 2014 the 20-24 year old group was wholly populated by women who had been eligible for the vaccine and government estimates take up of the vaccine to be 49%. In this year the number of cervical cancers in this age group was 82 – the highest number of cervical cancers (and also the rate was the highest) since cervical cancer stats were available in 1999. This may well be a blip in the stats but it still looks very suspicious that this peak occurred in a cohort where nearly half of the women were vaccinated.

    Its also worth noting that in the 15-19 year old age group, the number of cervical cancers in the 8 years prior to the vaccine being introduced was 16, and the number in the 8 years after the vaccine was introduced was 16 – there appears to be no change whatsoever in the pattern of cervical cancer in this age group. I appreciate the numbers are very small, and quite possible may be a different and more aggressive type of cervical cancer like small cell cervical cancer, but it’s important for women not to become too blase about the protection offered by the HPV vaccination when there is no evidence yet of just how the vaccine will impact the number of cervical cancers. Assurances and headlines about cervical cancer being eradicated are very premature and based on estimates, assumptions and modelling, which may, or may not be accurate, as so many other factors are involved.

    Diane Harper, an investigator in the trials for both Cervarix and Gardasil has said publicly that if the vaccines don’t offer good protection for more than 15 years, it won’t make a difference to the numbers of cervical cancers, as it will merely postpone the cancers.

    I also recently had sight of a paper in which GSK stated their Cervarix vaccine only had VLPs for two strains of HPV (HPV16 & HPV18) because adding more strains caused antibodies to wane for HPV18. Although Cervarix only contains protection against two HPV strains, it has shown good cross protection for other high risk strains.

    In Diane Harpers 10 year review of the HPV vaccines, studies showed:

    Duration of antibody response is critical for clinical prevention of HPV infection. Cervarix has high anti-HPV16 and HPV18 antibody titers for at least 9.4 years [15]in longitudinal follow-up studies; Gardasil has plateaued anti-HPV16 titers well above natural infection titers for at least 9 years, but anti-HPV18 titers that are not different from natural infection titers as early as 24 months after vaccination [16].”

    “Nearly 20% of Gardasil9 recipients had a loss of detectable anti-HPV18 titers after 24 months. In Gardasil recipients, after 1.5 years over 10% of women had no detectable anti-HPV18 titers, after 3 years over 20% of women lost detectable titers, and after 5 years nearly 35% of women lost detectable titers [18]“

    Women are not told this, and I doubt very much the modelling done will have taken this into account, as the favourable studies seem to be cherry picked to boost the perception of both the efficacy and safety of HPV vaccines, as has been demonstrated by the recently announced BMJ investigation into the May Cochrane Review on HPV vaccines, where critics have accused the review of being ‘incomplete’ and ‘biased’.

    HPV vaccine safety: Cochrane launches urgent investigation into review after criticisms https://www.bmj.com/content/362/bmj.k3472.full

    • We appreciate that people are very eager to see an impact of HPV vaccination on cervical cancer rates and that some people are interpreting the lack of observable effect so far as an indication that the vaccines do not work. There is currently no data suggesting that the HPV vaccines do not work as intended. The evidence that they work in preventing HPV infection and pre-cancer caused by HPV is overwhelming. We anticipate that vaccination will first have a substantial impact on cervical cancer in 2019-20 and that sufficient high-quality data should be available in 2021.

      We have shown previously that the increase in cervical cancer in women aged 20-24 in 2014 was restricted to cases in women aged 24.5-24.9 and resulted from screen-detection of cancers six months earlier than in previous years when women were first invited for screening at 25.0 (as opposed to at 24.5 in 2014). Anyone aged 24.5 in 2014 would have been 18 or 19 years old in September 2008 when vaccination was first introduced in England. The majority were not vaccinated in 2008-09 and even among those that were, a substantial proportion would already have been infected with HPV16 or 18 by the time they were vaccinated. As we predicted in 2010, it is unlikely that cancers will be reduced within eight years of vaccination (because it typically takes at least 8 years for cancer to develop after HPV infection).

      Cervical cancer in 15-19 year old women is extremely rare (as you say about 2 cases per year in England) and may not be preventable by vaccination of adolescents. Some will be in teenagers infected with HPV at birth or in childhood, and some may be unusual types of cervical cancer not caused by HPV.

      We agree that talking about cervical cancer “eradication” created false expectations. Over the next 50 years cervical cancer will become a very rare disease in England and a major challenge is to make it very rare all over the world. But unlike smallpox or polio, it is extremely unlikely that there will be no cases of the disease.

      Although there is little evidence of the HPV vaccines providing protection for 15 years or more, there is no evidence that they don’t and considerable reason to expect that they will. Even if the vaccines only work for 10 years that would have a huge impact on cervical cancer. HPV infections are most common in 16-24 year olds. Preventing infection in that age group will have a huge knock-on effect in reducing the lifetime risk of HPV infection and hence of cervical cancer. Even postponing a cervical cancer from age 29 to age 39 is a good thing.

      Levels of HPV18 antibodies are no higher 24 months after vaccination than they are after natural infection. But natural infection also protects against re-infection and multiple studies show that HPV18 infections are extremely rare 5-7 years after vaccination. It has been suggested that HPV vaccines also induce protection against new infections via immune memory.
      There were no cases of HPV16/18 positive pre-cancer (CIN2+) over 12 years of follow-up of 2084 women vaccinated with Gardasil in Nordic Countries. Comparison with controls shows that the vaccine prevents at least 90% of such disease for at least 10 years.

      Any reading of the literature would conclude that the commercially available HPV vaccines are extremely effective at preventing persistent HPV infection and resultant neoplasia. A recent review formalised this view. It also concluded that the vaccines are safe. The same conclusions had been independently reached by the FDA, the MHRA and other medicines regulators. A group who previously unsuccessfully challenged the European Medicines Agency’s position on the HPV vaccine are now challenging this review. The challenge will be formally investigated, but it is noteworthy that in announcing the review, David Tovey, editor-in-chief of the Cochrane Library, said: “we also have no reason to believe that the main conclusions of the review relating to benefit and serious adverse effects are unsafe”.

1 Trackback / Pingback

  1. Further evidence of the success of the National HPV Immunisation programme in England – Cancer Prevention Group Blog

Leave a Reply

Your email address will not be published.