When to test women for human papillomavirus

Increased understanding of human papilloma-viruses (HPV) and cervical carcinogenesis has led to prevention strategies that are very promising. Two articles in this issue (pp 79, 83) find that DNA testing for HPV is a cost effective way to clarify the meaning of equivocal results of cervical cytology.¹,²

Cervical HPV infections are very common, particularly among young women, and are sexually transmitted.³ They usually clear within one or two years. However, persistent infections by 15 or so carcinogenic HPV genotypes cause most cases of cervical precancer, which is generally diagnosed years after the causal infection, and of invasive cancer, which is typically diagnosed at least two decades after infection.⁴

Reproducible methods for testing for HPV are already available that are more sensitive (but less specific) than cytology for detecting prevalent and incipient precancer and cancer.⁵ Testing for HPV might be used to improve the four sequential steps of programmes for preventing cervical cancer: population screening, triaging equivocally abnormal cytology, diagnosing women with abnormal results linked to deciding when to treat, and assuring cure after treatment. Until recently, the standard model for prevention was based on identifying morphological changes in cervical cells. This model comprises cytology for screening, repeated cytology to triage equivocal results, colposcopy and biopsy for histological diagnosis and treatment decisions, and cytology or colposcopy to assess cure. Other strategies based on HPV testing are gaining empirical support.

Cytology (Pap smear) screening has greatly reduced rates of cervical cancer in regions with well organised, population-wide programmes. Because cytology is limited by moderate sensitivity and poor reproducibility, the effectiveness of such programmes relies on many rounds of screening throughout adult-hood. Testing for HPV is a more sensitive and reproducible tool than cytology for cervical precancer and cancer and is theoretically a better primary screening test if applied with proper understanding of the clinical course of HPV infection.

For example, HPV screening should not begin until 10-15 years after the average age of sexual debut, past the ages of frequent acquisition and clearance of HPV infection. Efficiency of HPV testing increases with age because the prevalence of benign, recently acquired infections declines as the prevalence of precancer and treatable cancers rises. These trends together increase the predictive value of a positive HPV test but maintain the reassurance provided by a negative test, which in turn permits the lengthening of screening intervals.⁶

If HPV testing is used for primary screening the management of women found to be positive for carcinogenic types of the virus will depend on regional resources. To increase the specificity of screening for HPV, positive women can be triaged by using Pap tests⁷; if they are cytologically negative, they can be rescreened by HPV testing a year or two later to identify persistent infection. When assays for specific types of HPV become available, persistent infection with the most carcinogenic types, such as HPV16 and HPV18, would imply a particularly high risk of cancer. Where diagnostic resources are limited, women aged 35 and above might have HPV testing in the context of screen and treat strategies because, in many populations, a positive HPV test at older ages will probably represent persistent infection and an associated raised risk of precancer or treatable cancer.⁸

When the results of screening cytology are equivocal, “reflex” HPV testing is cost effective in deciding whether colposcopy is needed, as confirmed by the two articles in this issue.¹,² True precursors of cervical cancer (and the cancer) are caused by carcinogenic HPV; “look alike” cells are negative for carcinogenic HPV. In any setting HPV testing is useful only to clarify results for the cytological categories that harbour true uncertainty. Technical efficacy of triage is no longer questionable, but the cost effectiveness of triage by HPV testing compared with cytological or colposcopic methods will vary between populations and regions.

Recent improvements in screening for cervical cancer have not been matched by concomitant advances in colposcopic evaluation and diagnosis. Indeed, the sensitivity of biopsy directed by colposcopy to detect underlying precancer is only around 70%.⁹ Thus, despite its historical status as the diagnostic gold standard, this procedure is now a technically weak link in the prevention of cervical cancer. Cost effectiveness analyses should be adjusted to reflect these limitations.

If the reproducibility and sensitivity of HPV testing proves to be superior to the combination of cytology and colposcopy, are we willing to treat women surgically—and remove the entire zone of cancer susceptibility, the cervical squamocolumnar transformation zone—on the basis of virological risk status alone? In general, the specificity and positive predictive value of HPV testing are mediocre. But there are notable exceptions: for example, even in the absence of histologically confirmed precancer, women with abnormal cytology and HPV16 infection are at very high absolute risk of having a missed, small precancerous lesion (cervical intraepithelial neoplasia grade 3).¹⁰

After excisional or ablative treatments of the cervix for precancer, the absolute risk of recurrence is about 5-10%. Testing for HPV four to six months after treatment is highly sensitive and specific for the risk of recurrence and is better than cytology alone for monitoring cure.¹¹

Regardless of their promise, switching to prevention strategies based on HPV testing warrants caution. Only one test has had extensive clinical validation. Before widespread adoption, all new HPV tests will require robust, real life evidence of reliability and accuracy in detecting precancer and cancer; otherwise, testing errors could mislead clinical management.¹² Furthermore, even the low cost HPV tests now being developed will have to be used in targeted populations if they are to be cost effective. Excessive or misguided use will increase costs without adding benefit. Like most revolutionary technologies, HPV testing must be managed wisely to do good rather than harm.