HPV test alternative to smear

Thursday May 15 2008

“Hope over smear test alternative” is the headline from BBC News today. It reports that testing for the common sexually transmitted infection human papillomavirus (HPV) may be a better screening tool for cervical cancer than smear tests. In addition, “testing for HPV was so sensitive it only needed to be done every six years - compared to three years for smears”, the BBC adds.

The story is based on UK research which has looked at the proportion of people who are initially reassured by a negative smear test or by the new HPV test and who later go on to show early signs of cervical cancer and how long it takes these women to develop this abnormality. This second report of a large, local study provides reliable information that supports extending the use of the test; however, more research into how the test might be used, either separate from the smear test or in addition to it, will be required. Larger studies will also need to look at the number of cases of cancer itself that could potentially be prevented by different screening strategies.

It is worth pointing out to women who may be confused by the headline “hope over alternative” that the HPV test is carried out on a sample from the cervix obtained in the same way as a conventional smear test; the alternative is not, for example, a blood test that removes the need for a cervical examination.

Where did the story come from?

Dr Jack Cuzick and colleagues from Queen Mary’s School of Medicine and Hammersmith Hospital carried out this research which was funded by a Cancer Research UK programme grant. Dr Cuzick had received funding from the makers of the test. The study was published in the peer-reviewed International Journal of Cancer .

What kind of scientific study was this?

This was a retrospective cohort study that reported the long-term findings from the Hammersmith Study in women aged 35 or over. This study, of almost 3,000 women, reported on how accurate the usual cervical smear test (cytology) was compared with the newer DNA-based tests for the HPV virus. The researchers found that the newer test was more sensitive: it picked up more women with an abnormal result in the initial test who were subsequently shown to have abnormal cervical cells when examined more closely by colposcopy and biopsy. Colposcopy is a procedure where the cervix is closely examined in the clinic and often a piece of tissue, a biopsy, is taken to confirm the degree of abnormality found on the cervical smear.

Signs of early cervical cancer are pre-cancerous cells called cervical intraepithelial neoplasia (CIN), and these are divided into “grades” of severity, graded CIN1+, CIN 2+ and CIN3+. The researchers wanted to know if these new tests were specific; that is, how good they were at correctly excluding women who did not go on to show abnormal cells (graded more severe than CIN2+ or 3+) when a biopsy was carried out. Women went on to have colposcopy and biopsy if they had an abnormal initial result either on a conventional cervical smear test or on the newer HPV test. The biopsies were the standard by which the accuracy of both tests were judged.

Patients were recruited from April 1994 to September 1997. The researchers excluded women who had already been treated for CIN of any grade or who had had any cervical abnormality in the previous three years. Data was available for 2,516 women over a nine year period (included in this was the initial study period which had an average follow up of 6.4 years). Some women, 466 of those initially recruited, were removed from the surveillance system because they had died, had hysterectomies or moved out of the area. Others could not be traced for some reason or were followed up for less than one year. The DNA tests and the cervical smear results were all collected from the results in the first year.

For the women that had complete follow-up data, the researchers counted the number of abnormal biopsy results that occurred after the first year of the study and compared this with the results of the groups who at the initial test in the first year of the study had normal, borderline, mild, moderate or severe cellular abnormalities, with either a negative or positive HPV test. In this way they were able to calculate the sensitivities and specificities of the two tests and the chance that if a person tested positive at the initial test, they would turn out to have abnormal CIN on biopsy (the positive predictive value). The researchers also report the risk for people who had a negative initial test in the first year of the study of subsequently developing CIN2+ or more.

What were the results of the study?

Twenty new cases of CIN2+ or more were identified amongst the 2,516 women who had at least one further smear in the period after the first year of the study. Including any disease identified with the initial test in the first year of the study, the risk of developing CIN2+ or more at one, five and nine years after a normal cervical smear test was 0.33%, 0.83% and 2.20% respectively. The equivalent figures after a negative HPV test were 0.19%, 0.42% and 1.88%.

What interpretations did the researchers draw from these results?

The researchers conclude that “HPV testing offered excellent protection from CIN2+ or more for at least six years after a negative test, whereas the protection from cytology (the cervical smear test) began to wane after about three years.” They say that about six times as many CIN2+ or more lesions were found in the follow-up period in those initially HPV positive compared with those HPV negative, whereas there was little difference between normal and abnormal initial cervical smear tests.

What does the NHS Knowledge Service make of this study?

This topical study provides data that will inform policies on how best to screen for cervical cancer. There are several limitations to the study, which the researchers acknowledge, that mean it is too soon to conclude that the newer tests will replace the conventional smear test.

  • The study only included women aged 35 and above, so the results may not apply equally to younger women. While true, this limitation may not be relevant as primary HPV screening for women is only recommended above the age of 30 and this study covers most of this age group.
  • The DNA HPV testing was done retrospectively on samples collected from the initial testing round. The type of DNA testing changed over the period of the study, and 58 women who were initially negative to both the first type of DNA test and the smear test were later shown to be positive to HPV on the newer DNA test. They did not receive the initial colposcopy and so, in fact, may have had CIN2+ or more at baseline. The potential effect of these factors in reducing the accuracy of the outcome is unknown.

Larger studies which follow women for longer may show us whether these new types of testing reduce cancer incidence, rather than just reducing the rates of cervical intraepithelial neoplasia or pre-cancerous changes. The place of this test in the pathway of cervical cancer testing has yet to be determined.

Sir Muir Gray adds...

Good research; well reported; needs assessment in an ordinary service setting.

Analysis by Bazian
Edited by NHS Choices