Abstract
Background
Organized Pap cytology screening for cervical cancer or its precursors has been practiced for decades in most Western countries. Having led to substantial reductions in cervical cancer mortality, it is currently considered one of the greatest success stories in public health history. Traditionally, screening guidelines have dictated that women are to be invited for their first Pap test shortly after becoming sexually active, but recently there has been some debate about when to initiate screening. In July 2009, Sasieni and colleagues published updated results from a large case–control study looking at the age-specific effectiveness of cytology-based screening in the UK [1].
In 2004, the NHS in England decided to increase the screening age from 20 to 25 years based, in part, on an earlier report from the same group of investigators [2]. They showed that the benefit associated with screening was substantially less in women aged 20–34 years than in older women and suggested that policy changes should be considered. Despite the evidence from this audit study, at a recent British Medical Association meeting, representatives voted in favor of lowering the screening age to 20 years [3]. This decision followed a high-profile campaign to reduce screening age and following the death of popular television star Jade Goody, who died at 27 years of age from the disease.
Human papillomavirus (HPV) infection is very common in young women after the onset of sexual activity and is usually associated with low-grade cervical intraepithelial neoplasia (CIN), following initial infection. While the vast majority of these dysplasias regress, it is not possible at present to determine which ones will progress to high-grade CIN and eventually to invasive cancer. Screening at this point in a woman's life will likely detect such cytological abnormalities and, depending on the clinician's perception of risk, may prompt unnecessary ablative treatment, which poses substantial risk of present or future adverse reproductive outcomes (e.g., miscarriage or premature delivery during subsequent pregnancies and low birth weight) [4].
Compounding the problem is the low sensitivity of a single Pap test in detecting pre-cancerous and cancerous lesions, at a little over 50% [5]. In order to increase a screening programme's sensitivity, this low sensitivity is typically compensated by frequent screening. Not uncommonly, screening guidelines recommend annual or biennial Pap tests to minimize the risk of false-negative results on successive smears. This in turn makes cytology screening a costly process that is very demanding on the public health system to ensure that it is done frequently enough while maintaining the standards of quality, high coverage and follow-up of all cases considered abnormal to permit an impact on cervical cancer incidence and mortality. In North America, guidelines tend to place greater emphasis on annual Pap tests beginning at 18 years of age, or 3 years after the onset of sexual activity, but not later than 21 years of age. It is only after three consecutively negative annual Pap tests that the frequency of screening can be decreased to 2- or 3-year intervals [6,7]. Owing to their efficient call–recall systems with invitations to screen sent to all eligible women, European countries with organized screening programmes, such as the UK, Sweden and Finland, begin screening at later ages (e.g., 25 years), but opportunistic screening at younger ages is also common.
In light of the differences between North American and European guidelines, it is not surprising that the issue of age at screening initiation has not yet been solved. Because of the low risk of cervical cancer before 30 years of age, randomized, controlled trials would be impractical (and unethical in jurisdictions that have policies with early age of onset) to study the problem. We are thus left with findings from observational studies to make decisions about screening young women. In addition to the study by Sasieni et al. [1], which clearly indicates that the benefits of beginning screening before 25 years of age are nonexistent, there have been three other studies published on this topic. Two of them, one from Italy [8] and the other from Australia [9], support Sasieni's original findings, but the third, which was carried out in Sweden, found no evidence that screening was less effective in young women [10].
Results
Sasieni et al. present results from a population-based case–control study examining the effectiveness of cervical cancer screening, with a particular focus on women screened between 20 and 24 years of age. Cases (n = 4012) included women with either invasive or microinvasive cervical cancer (identified using histopathology laboratory records) from different centers across the UK and diagnosed between January 1990 and April 2008. Controls (n = 7889) were individually matched to cases on age and place of residence and, when possible, two controls were identified per case. The investigators report odds ratios (OR) calculated by conditional logistic regression to estimate the association between cytology screening during a particular 3-year age band (e.g., 22–24 years) and incidence of cervical cancer in the subsequent 5-year band (e.g., 25–29 years). The authors carefully considered membership in the NHS to select cases and controls to avoid biases related to lack of screening opportunity. While one control was matched to the index case on the same attending clinician the other was matched on neighborhood but from a different clinician, to avoid a conservative bias that could occur by only matching on the same general practitioner (physicians tend to have their own preferences in terms of being liberal or conservative in following policies, which would reflect in the screening histories of all their patients).
Overall, the benefit from screening was shown to increase with age. At older ages the ORs were well below 1.0, but for younger women (<25 years) they were above 1.0. For example, for women screened between 22 and 24 years of age, the OR was 1.11 (95% CI: 0.83–1.50), between 32 and 34 years of age the OR was 0.55 (95% CI: 0.44–0.69), between 42 and 44 years of age the OR was 0.37 (95% CI: 0.29–0.48), and between 52 and 54 years of age the OR was 0.26 (95% CI: 0.19–0.36). To test the sensitivity of these results to stage and histological type of cancer, the authors performed analyses restricted to specific cancers. When restricting the analysis to either squamous cancers only, stage IB+ (any histology), squamous and stage IB+ or stage IIB+ (any histology) the authors showed that the overall pattern of increasing ORs with decreasing age stayed the same and that at young ages, the ORs never varied significantly from unity. The authors also showed that the ORs were generally lower for more advanced disease, supporting the notion that advanced cancers are less common in screened populations.
Significance
Consistent with what is already well established in the cervical cancer literature, investigators found that screening older women (aged 35 – 64 years) reduces cervical cancer risk by approximately 60 to 80%. Researchers also found that screening younger women (aged 20–24 years) was consistent with no reduction in risk. Taking into account the anxiety and pregnancy-related morbidity associated with screening and treatment, screening younger women could actually be harmful.
The study confirms the author's previous findings and supports the NHS's decision to increase the screening age to 25 years. Even though approximately half of the data were previously included in analysis for the original manuscript, when limiting the analysis to only new data, the results were qualitatively similar. This suggests that the impact of screening young women is the same despite recent improvements in the quality of screening throughout the UK [1]. Major strengths of this study include its size and the use of the NHS database to identify subjects and collect data. As data were collected on all randomly selected controls and information was prospectively recorded, there was no opportunity for selection or recall bias. However, since this was not a randomized trial there was still opportunity for confounding factors. In fact, it is likely that women who attend screening differ from women who do not on at least some factors. But as the authors point out, in order for the observed difference in the benefit from screening to be attributable to confounding, confounders would have to affect the results differently at different ages, which is unlikely.
Considering the good reputation of the UK's cervical cancer screening program, the results from this study are likely generalizable to other countries, including the USA, which at present has the most risk-averse policies. Can we assume that the findings from this study will have a major impact on policy decisions regarding cervical cancer screening? Unfortunately, this study did not examine the net balance of harms and benefits of screening at younger ages, which is required for sound policy decisions. It is likely that the decision in a particular country on when to initiate screening will depend on the preferences by professional societies and public health stakeholders. In addition, in countries where screening is currently offered to women aged 20–24 years and where policy decisions are heavily influenced by public opinion, lowering the screening age may be a politically risky decision.
Future perspective
Human papillomavirus vaccination is expected to have a major impact on how we screen for cervical cancer. As successive cohorts of vaccinated women reach screening age, lesion prevalence will be much lower and the positive predictive value of Pap cytology will be negatively affected. Pap screening will no longer remain an efficient first line of defense and the question of the effectiveness of screening women aged 20–24 years using this technology may become a moot point. Following vaccination, the sensitivity and specificity of Pap cytology are also expected to be negatively affected [11]. Smear reading is a highly repetitive process that is prone to human error and when most smears are negative, the signal (squamous abnormalities) to noise (inflammation and reactive atypias) ratio will be reduced and this may lead to errors in interpretation. Less conspicuous lesions may be missed and in an attempt to compensate there may be more overcalls of benign abnormalities.
To overcome these new challenges, policy makers will need to consider HPV DNA testing to replace Pap cytology as the primary screening test for cervical cancer [11]. This is because HPV testing is less prone to human error and is much more sensitive than cytology in detecting high-grade CIN. Despite the many issues associated with Pap cytology, it will likely remain important as a second-line test reserved for the more labor-efficient task of triaging HPV positive cases. In this situation, lesion prevalence will be high since the cytology workload will only include smears from women harboring HPV infection, and as a result, Pap cytology would be expected to perform with sufficient accuracy. Via linkage of vaccination and screening databases, an HPV followed by Pap testing strategy would also provide a low-cost strategy to monitor long-term vaccine efficacy. If HPV DNA testing is eventually adopted as a population screening tool, age to initiate screening will become an important concern. Furthermore, research needs to be conducted to address this question so that policy makers will be better informed.
Executive summary
In many developed countries, Pap cytology screening has been effective in reducing mortality from cervical cancer. It is uncertain whether women aged 20–24 years receive any benefit from screening.
Results of a large case-control study with cases identified between 1990 and 2008 from centers across the UK suggest that women aged 20–24 years receive no benefit from screening. Consistent with what is already well established in the literature, investigators found that screening older women (aged 35–64 years) greatly reduces cervical cancer risk. The results were not sensitive to stage or histologic type of cancer.
Screening younger women is costly and treating these women may be associated with adverse reproductive outcomes. The results support the NHS's decision to increase the screening age to 25 years in England.
In the post-vaccination era, Pap cytology will likely be replaced by human papillomavirus (HPV) DNA testing as the primary screening test for cervical cancer. Research should be carried out to determine the best age at which to initiate screening using HPV testing.
Footnotes
Eduardo Franco has served as occasional consultant to companies that produce human papillomavirus vaccines (GSK and Merck), cervical cytology tests (Cytyc and Ikonisys), and human papillomavirus assays (Roche, Gen-Probe and Qiagen). The authors have no other relevant affiliations or financial involvements with organizations or entities with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript.
No writing assistance was utilized in the production of this manuscript.