How often should we screen for cervical cancer?

Over the past 60 years, the mortality from cervical cancer — once one of the most common and lethal cancers in women in the United States — has decreased dramatically. Much of the reduction has been due to the widespread use of the Papanicolaou test, which has enabled clinicians to detect cervical intraepithelial neoplasia before it progresses to cervical cancer and to detect cervical cancer at an early stage. When cervical cancer is detected early, the five-year survival rate is more than 90 percent. 1 The Papanicolaou test is now the most widely used cancer-screening tool in the United States by far, with more than 80 percent of women undergoing screening in any two-year period and more than 90 percent having been screened at least once. By contrast, the rate of screening for colon cancer (by means of a fecal occult blood test) among women older than 50 years of age is only 34 percent within a two-year period. Despite the widespread acceptance of cervicalcancer screening, questions remain about optimal screening strategies. One key question is the optimal frequency of testing. Cost–benefit analyses have suggested that lifelong annual screening may not result in substantially better outcomes than less frequent screening and is much more costly. With this in mind, the American Cancer Society recently revised its guidelines for screening, 2 recommending intervals between screenings ranging from one to three years, depending on several factors, such as age, screening history, type of Papanicolaou smear, and history of immunosuppression. Other guidelines have also suggested screening less frequently than annually after three consecutive normal annual Papanicolaou tests and pelvic examinations. Yet there are not many data to support these recommendations. In this issue of the Journal, Sawaya et al. (pages 1501–1509) report the rates of cervical neoplasia in a relatively high-risk cohort of women 30 to 64 years of age who had had at least three previous negative Papanicolaou tests. They conclude that subsequent screening at three-year intervals would confer an excess risk of invasive cervical cancer of less than 3 in 100,000 — a risk that they note is roughly equal to the annual risk of breast cancer among men 45 to 64 years of age. This low risk might be lowered even further with improvements in the sensitivity of Papanicolaou testing achievable with liquid-based technology or human papillomavirus DNA testing and with adherence to new guidelines for the interpretation and appropriate follow-up of abnormal Papanicolaou tests. Sawaya et al. also provide an estimate of the excess number of tests that would be required with continued annual screening. They estimate that in a hypothetical cohort of 100,000 women who have already had three normal consecutive Papanicolaou tests, annual screening, as compared with screening once three years after the last negative test, would result in an additional 69,665 Papanicolaou tests and 3861 additional colposcopic examinations in women 30 to 44 years of age and an additional 209,324 Papanicolaou tests and 11,502 colposcopic examinations in women 45 to 59 years of age, while detecting only one additional case of cancer. These additional tests are both costly and emotionally upsetting to the patient, without materially affecting the rate of cervical cancer. Despite the guidelines indicating that the interval between screenings can be lengthened for many women, both physicians and patients have been reluctant to change it. This resistance probably stems from several factors, including the success and simplicity of annual screening, patients’ concern about cancer, and physicians’ concern about medicolegal issues. The data presented by Sawaya et al. may help to alleviate these worries. A broader concern is that if the recommended interval between screenings were increased to three years, patients might inadvertently be tested even less frequently. Ensuring that women are tested no less frequently than every three years is important, since 10 percent of cases of cervical cancer occur in women who have had a Papanicolaou test but who have not been screened within the previous five years. As patients move away from annual cervicalcancer screening, we must make sure that they do