Should routine screening Papanicolaou smears be done for women older than 65 years?

Abstract

O F ALL MEDICAL DEcisions, those regarding clinicianinitiated tests and trea tments in asymptomatic individuals require the most thoughtful consideration: as clinicians, we are careful to “do no harm” to healthy people. Populationbased preventive interventions fall in this category and include decisions about screening women older than 65 years for cervical cancer. As women age it becomes increasingly important to consider the benefits and harms that can be expected. With advancing age, the balance between benefit and harm continually changes: greater benefits and fewer harms can be expected in younger, healthy, at-risk women who have never been screened; older, infirmed women with prior normal test results and short life expectancies can expect fewer benefits and greater harms. Assessing where individual women reside at any one time on the balance provides the objective foundation of informed decision making regarding screening. Armed with the best quantitative information, clinicians can then incorporate individual women’s preferences to craft a screening strategy that is rational, safe, acceptable, and effective. A framework for cancer screening decisions in older individuals has recently been proposed. The quantitative information that clinicians need to consider are a woman’s likelihood of developing and dying from cervical cancer in her remaining lifetime, the potential benefits that will accrue from screening, and the potential harms directly related to the screening intervention. Much of the information, though difficult to estimate, is critical for optimal decision making. An unscreened woman’s lifetime risk of cervical cancer is estimated at 3.5% and can be reduced to 0.8% with screening. Throughout an average woman’s lifetime, events occur that modify her risk of developing cervical cancer. The most striking is surgical removal of the cervix by total hysterectomy. Up to 37% of women in the United States have had their cervix removed by the age of 65 years. Women who have had total hysterectomies for nonneoplastic conditions are no longer at risk of cervical cancer and should not be screened for this malignancy. Clinicians should be aware that some women may have had a supracervical hysterectomy in which the cervix was retained. As long as a woman has a cervix she will be at some risk, but these risks become smaller as the number of prior Papanicolaou tests (Pap smears) with normal results increases. By the age of 65 years, a woman with a long history of normal test results has an extremely high likelihood of being among the women who will not develop cervical cancer. Cumulative lifetime risks of incident cancers are more important for risk counseling, but these have not yet been well described in this group. Continued screening, however, can be expected to provide additional benefit in terms of cervical cancer prevention; but this relatively small additional benefit needs to be placed in the context of life expectancy. Women 90 years and older, for example, have average life expectancies of less than 5 years and are unlikely to benefit from cancer screening of any type. Younger women with severe medical conditions also may be expected to achieve marginal benefits from screening. Given that most cervical cancers diagnosed in older women occur among unscreened or poorly screened women, healthy women older than 65 years who have never been screened will probably achieve the greatest benefit. In the future, tests with more powerful predictive value, such as those screening for the presence or absence of highrisk human papillomavirus (HPV), may be more accurate than cytologic evaluation in risk-stratifying older women, though this strategy requires further study. No medical interventions are without harm, and the benefits of screening need to be weighed against harms of screening. Harms are easier to accept if patients are symptomatic than if they are not, as patients are more willing to accept risk, and From the Department of Obstetrics, Gynecology and Reproductive Sciences and the Department of Epidemiology and Biostatistics, University of California, San Francisco. The author has no relevant financial interest in this article. George F. Sawaya, MD CONTROVERSIES IN INTERNAL MEDICINE