Elderly Women Above Screening Age Diagnosed with Cervical Cancer Have a Worse Prognosis

Discussion

In the present study, 26% of the women who were diagnosed with cervical cancer were above the age of 65 years. The recommended cervical cancer screening system in Sweden covers women aged up to 65 years, which is why our results indicate that more than one-quarter of cancer cases are not detected by the national screening program. In 2011, women in Sweden had an estimated life expectancy of 83.5 years, which means that for one-third of their lives they are not screened (7). The breast cancer screening program ends at 75 years of age. The major causes of death are cardiovascular diseases and malignancies (8). However, a large proportion of women over 65 years are healthy and have an active sexual life (9). They are susceptible for curative treatment for dysplasia and cervical cancer if the disease is discovered at an early stage, as shown by the results from another study (2). The recommended screening programs are ended at 60 to 65 years of age in many countries. The screening program accomplishes substantial risk reduction up to the last smear and some years beyond the last smear (10). However, the efficiency of cytological screening has been shown to be lower in women 50 years or older compared to women of younger age (11). The high-risk Human Papilloma Virus (hr-HPV) test has higher sensitivity (12) than a single cytological test at finding cervical dysplasia, especially in women above 50 years of age (13). A Dutch study of cost-effectiveness indicated that primary HPV screening would be preferred in women over the age of 30 years (14). The somewhat lower specificity of the hr-HPV test compared to liquid-based cytology may be compensated for by also analyzing the hr-HPV-positive samples with normal liquid-based cytology. Incorporating screening with the modern hr-HPV test and triage of hr-HPV-positive women with cytology provides a good balance between maximizing sensitivity and specificity by limiting the number of referrals for colposcopy and conisation (15), especially in post-menopausal women (13, 16, 17).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

The number of patients with screen-detected cervical cancer (×) and the number of patients with symptoms (black shading) in relation to the International Federation of Gynecology and Obstetrics stage. All screen-detected women with cervical cancer had disease stage I a-b except for one patient with stage II.

Since organized screening in Sweden began in the 1960s, the number of women with cervical cancer has decreased. The mean age of diagnosis for cervical cancer has increased to 55 years, and the proportion of elderly women aged 70 years or above has increased from 5.4% to 27.3%, supporting the results of this study (18). It is, therefore, important to identify the high-risk individuals above screening age, or to extend the screening program to older women. An Italian study shows results parallel to ours. Almost 40% of Italian women with cervical cancer had not had a Pap smear taken. Interestingly, 16% of the cases were above 64 years of age, and thus not included in the screening program (19). In a Dutch study including 401 women with cervical cancer, one-third of the women were outside the screening age, which is consistent with our findings (2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Kaplan–Meier estimates of survival probabilities for patients with cervical cancer diagnosis related to whether the woman had followed the screening program (n=60, deaths=8), not followed the screening program (n=62, deaths=16) or was beyond the screening age (>65 years old) with normal screening history (n=22, deaths=11) or beyond screening age (>65 years old) not followed the screening program (n=21, deaths=11). Ticks indicate censored patients. Overall survival using the Cox proportional hazards model showed that women beyond screening age who had not followed the screening program had a Hazard Ratio (HR) of 5.9 (95% Confidence Interval (CI)=2.4-14.6, p<0.001), women beyond screening age with normal screening history had an HR of 4.8 (95% CI=1.9-12.1, p=0.001), while the women of screening age who had not followed the recommended screening program had an HR of 2.0 (95% CI=0.9-4.7, p>0.05 compared to women who had followed the screening program. There was no difference in the overall survival between women beyond screening age if they had a normal screening history or not (p=0.6).

An adequate compliance in cervical screening programs is 85% according to European guidelines (20). In Sweden, the compliance rate in women 50-60 years of age was 84% but in Skåne was 74% in 2012 (21). In non-attending women, cervical cancer is detected at a more advanced stage, leading to worse prognosis (22), similar to the results in this study. Some commonly stated reasons for not attending cervical screening are “uncomfortable with vaginal examination”, “lack of time”, “feeling healthy” or “experience of unfriendly health workers” (23). An HPV self-test may be a suitable way to address these issues and increase compliance with cervical screening in both non-attendants and also in elderly women above the screening age. In Finland, the coverage of cervical cancer screening was increased to the desired level of 85% with a self-screening test if a third reminder was sent (24). Recently, we also showed that an HPV self-test increases compliance with the cervical screening program in long-term non-attendant women (23).

The present study confirms that screening-detected cervical cancer is diagnosed at an earlier FIGO stage than is symptom-detected cancer. Andrae et al. presented similar results showing a cure proportion of 92% in screening detected cancer (22). Among the symptomatic women, the cases that were discovered between screening intervals (interval cancer) had a higher cure rate compared to cancer in women who were not screened at all, or who were underexposed to screening. The more advanced stages of cancer were more common amongst the under-screened/unscreened persons, as in our study. In a Dutch study, 16% of screened women had advanced stages of cancer, while in the under-screened group, the figure was 48% (5). The cure rate is related to the FIGO stage, but the prognosis is better for screening-detected cancer, even after adjustment for stage (22). Further studies show that women above the age of 50 years are more likely to present with advanced stages compared to women under 50 years of age (25). Women not attending the recommended screening program and women beyond the screening age have a significantly more advanced FIGO stage when diagnosed with cervical cancer. We suggest that the cervical screening program should continue even after 65 years of age similarly to the breast cancer screening program up to 75 years of age. The introduction of cervical screening in elderly women must be assessed by the national quality register in Sweden.