Will HPV testing transform cervical screening?
HPV is a sensitive measure of cancer risk, but results must be interpreted with care, says Dr Anne Szarewski
Infection with certain 'higher-risk' types of human papillomavirus (HPV), in particular HPV 16 and HPV 18, is the most important risk factor for development of cervical cancer. But it is important to note that visible genital warts and HPV changes on a smear are not usually due to the higher-risk types. Biochemical tests, such as polymerase chain reaction (PCR) or hybrid capture, are necessary to distinguish between types of HPV, but such tests are not routinely available.
Infection with HPV appears to be extremely common in young people, but is usually transient. An analogy would be that many people catch colds, but only a small proportion go on to develop a serious chest infection. It appears the presence of HPV is more meaningful in older women (over 30), who have persistent infection. But the majority of HPV infections seems to be cleared within two years, probably mostly even earlier than that1.
Although HPV is the most important cause of cervical cancer it does not do so on its own. It appears the immune system determines how an individual responds to HPV infection and anything that reduces the immune response (such as immunosuppressive therapy) increases the risk.
The combined oral contraceptive pill may interact in some way with HPV to slightly increase the risk but it appears the HPV infection has to be present first. It is not clear how much of this might be related to differences in other social or sexual behaviour characteristics. Smoking has emerged as a risk factor in its own right, perhaps by affecting the local cervical immune response and making HPV persistence more likely. There is also a suggestion that early cervical abnormalities may regress with smoking cessation2.
The newest HPV tests are very sensitive – they detect more disease than smears – but the question of specificity (the number of false positives) is still an issue. However, a negative HPV test virtually rules out the chance of any abnormality being present.
In future, therefore, we might be able to reassure women who have a mild or borderline abnormal smear but a negative HPV test that they do not need to worry, while concentrating on those who test positive for the high-risk types. The liquid-based cytology (LBC) pilot included three centres using HPV testing as triage – results should be available by summer 2004.
Screening using HPV testing, either in combination with cytology or even alone, is also under investigation. A recent study using HPV testing in primary screening of women over 30 showed high sensitivity, but false positives were still a problem3.
It has been suggested that perhaps women with a positive HPV test but a negative smear could simply have another HPV test a year later, since in that study 45 per cent of them had a negative HPV test within that time. It also is possible that a longer interval between screens could be used if HPV testing was introduced.
Looking further ahead, HPV testing has potential as a self-sampling technique, making screening more acceptable to women who find gynaecological examinations embarrassing or unacceptable. A note of caution is necessary, however: we should not forget to consider the psychosocial consequences of HPV testing.
There has been very little research on this aspect, but recent studies suggest testing positive for HPV has an adverse psychosocial impact on women, with increased anxiety, distress and concern about sexual relationships.
Education will be extremely
important if HPV testing is to be more generally available.
Anne Szarewski is clinical consultant at the Margaret Pyke Centre, London, and a senior clinical research fellow in gynaecological oncology
at the Imperial Cancer Research Fund
The impact of screening
Although the number who develop cervical cancer is relatively small, the condition's unusual preinvasive stage (cervical intraepithelial neoplasia, or CIN) means it is almost completely preventable by screening. Around 200,000 women each year have an abnormal smear in the UK.
Since the programme of five-yearly screening began in 1988, coverage has greatly improved, from about 50 per cent in the mid-1980s to about 85 per cent overall in 1998. GPs are paid per smear taken at a rate dependent on their overall coverage, with increments at 50 per cent and 80 per cent.
Between 1987 and 1997 the mortality rate dropped by about 40 per cent; since 1990 the incidence rate has also fallen, from 16 women per 100,000 in 1986 to 9.6 per 100,000 in 2000. A recent paper estimated that cervical screening saved more than 8,000 lives between the introduction of the computerised call-recall system and 19974.
In October 2003 the NHS screening programme announced a change in screening frequency to three-yearly from age 25-49 and five-yearly from 50-64.
Cervical cancer affects approximately 3,200 women in this country each year, of which roughly 1,200 die.
It is very rare in women under 25, however. In 2002, just 26 cases of cervical cancer were registered for women aged between 15 and 24.
Indeed, it has been suggested that screening women under the age of 25 may do more harm than good. Young women frequently have borderline or mildly abnormal smears and yet, when referred for colposcopy, are not found to have any disease. Such false positive results lead to a lot of unnecessary investigation and anxiety.
Liquid based cytology (LBC) screening is to be introduced in England and Wales (Scotland is already in the process of implementing LBC). The main advantage of LBC is that it will greatly reduce the number of inadequate smears, from around 10 per cent to 3 per cent.
It also allows the possibility of doing other tests, such as HPV testing, from the same sample.
1 Schiffman M, Castle PE. Human Papillomavirus: epidemiology and public health. Arch Pathol Lab Med. 2003;127:930-4
2 Szarewski A. et al. The effect of smoking cessation on cervical lesion size.
Lancet 1996; 347: 941-3
3 Cuzick J. et al. Management of women who test positive for high-risk types of human papillomavirus: The HART study. Lancet 2003; 362:1871-6
4 Sasieni P, Adams J. Effect of screening on cervical cancer mortality in England and Wales: analysis of trends with an age period cohort model.
BMJ 1999; 318: 1244-5