A cervical screening initiative for elderly women in Denmark

Abstract Introduction To assess outcome of a one‐time human papillomavirus (HPV)‐screening in 2017 of Danish women aged 70+. Material and methods Women born 1947 or before were personally invited to have a cell‐sample collected by their general practitioner. Screening‐ and follow‐up samples were analyzed in hospital laboratories in the five Danish regions and registered centrally. Follow‐up procedures varied slightly across regions. Cervical intraepithelial neoplasia 2 (CIN2) was recommended treatment threshold. Data were retrieved from the Danish Quality Database for Cervical Cancer Screening. We calculated CIN2+ and CIN3+ detection rates per 1000 screened women, and number of biopsies and conizations per detected CIN2+ case. We tabulated annual number of incident cervical cancer cases in Denmark for the years 2009–2020. Results In total, 359 763 women were invited of whom 108 585 (30% of invited) were screened; 4479 (4.1% of screened, and 4.3% of screened 70–74 years) tested HPV‐positive; of whom 2419 (54% of HPV‐positive) were recommended follow‐up with colposcopy, biopsy and cervical sampling, and 2060 with cell‐sample follow‐up. In total, 2888 women had histology; of whom 1237 cone specimen and 1651 biopsy only. Out of 1000 screened women 11 (95% confidence interval [CI]: 11–12) had conization. In total, 579 women had CIN2+; 209 CIN2, 314 CIN3, and 56 cancer. Out of 1000 screened women five (95% CI: 5–6) had CIN2+. Detection rate of CIN2+ was highest in regions where conization was used as part of first‐line follow‐up. In 2009–2016, number of incident cervical cancers in women aged 70+ in Denmark fluctuated around 64; in 2017 it reached 83 cases; and by 2021 the number had decreased to 50. Conclusions The prevalence of high‐risk HPV of 4.3% in women aged 70–74 is in agreement with data from Australia, and the detection of five CIN+2 cases per 1000 screened women is in agreement with data for 65–69 year old women in Norway. Data are thus starting to accumulate on primary HPV‐screening of elderly women. The screening resulted in a prevalence peak in incident cervical cancers, and it will therefore take some years before the cancer preventive effect of the screening can be evaluated.


| INTRODUC TI ON
In Denmark, as in several other high-income countries, the current age-specific incidence of cervical cancer is bipolar with a first peak around the age of 35 years and a second peak around the age of 80 years. The second peak has been hypothesized to be related to reactivation of latent human papillomavirus (HPV) as the immune system degenerates with increasing age and/or with new HPVinfections resulting from new sexual relationships in middle aged women. 1 Demography is a third possible explanation, as women presently in older age have not been screened systematically for precancerous cervical lesions while they were younger. The problem was considered by the Danish Health Authority.
Following the argument that demography was the most likely explanation for the second peak in incidence, it was decided to offer the currently elderly women a one-time screening-test. In 2017, all women born in 1947 or earlier were therefore personally invited to have a screening sample taken by their general practitioner. Samples were HPV-tested to maximize sensitivity.
Women responded positively to the screening initiative, as 30% of these 70+ aged women participated. In total, 4.1% of participating women tested high-risk HPV-positive, with the prevalence decreasing slightly with increasing age. 4 As planned, the histology outcome was reported after end of the recommended follow-up period. [5][6][7] However, it turned out that follow-up periods had been longer than anticipated. We therefore updated the follow-up with two more years of observations and report on both benefit and harm. As cervical intraepithelial neoplasia grade 2 (CIN2) was the threshold for treatment of CIN in postmenopausal women in Denmark, 8 we used detected cases of CIN2+ as the main measure for outcome of the intervention. Finally, we report on the impact so far on the national cervical cancer incidence.

| MATERIAL AND ME THODS
Organization of the one-time screening of elderly women in Denmark has been described in detail before. 4 In brief, during 2017 women were personally invited to have a cell-sample collected by their general practitioner, samples were analyzed region-wise at the hospital pathology laboratories, and all data were collected in the national Pathology Data Bank. For definition of target population, see  We registered all follow-up in terms of screens, biopsies and conizations over an approximate four-year period from start of The screening resulted in a prevalence peak in incident cervical cancers, and it will therefore take some years before the cancer preventive effect of the screening can be evaluated.

K E Y W O R D S
cervical cancer, cervical intraepithelial neoplasia, conization, human papillomavirus, screening

Key message
In one-time HPV-screening in elderly Danish women, 4.1% of women were HPV-positive, and 0.5% had CIN2+ detected, but more than twice as many had conization.

| RE SULTS
In this one-time screening, 359 763 elderly eligible women were invited and 108 585 (30%) participated, Figure 1. There were limited regional differences in participation with rates between 29% and 31% for the five regions, Table 2. HPV was detected in screening tests for 4479 women, 4.1%, ranging from 3.4% in Region South to 4.8% in the Central Region. Overall, 24% of the HPV-positive women were positive for HPV16/18, and 76% were positive for other HPVtypes only, and similar distributions were seen for the regions, except for the Capital Region where 20% of the samples were positive for HPV16/18 and 80% positive for other HPV only. These regional differences were probably explained by different assays with BD Onclarity© used in the Capital Region and Cobas4800© elsewhere, Table 1.

| DISCUSS ION
To reduce cervical cancer incidence in old age, a one-time, primary HPV-screening was offered to women aged 70+ in Denmark.
Women with detected CIN2+ were treated. One-third of the detection of one CIN2+ case, and close to 300 women for detection of one CIN3+ case. Screening outcome reflected regional differences in follow-up strategy. Radicality in follow-up was decisive for both benefit of screening in terms CIN2+ detection rate and for harm in terms of proportion of women undergoing conization. Although the purpose of screening is to detect precancerous lesions, our study demonstrated that also asymptomatic invasive cancers were detected at this prevalence screen.
The strength of this study was use of high-quality data from the national Pathology Data Bank ensuring complete registration of all HPV, cytological and histological diagnoses following the positive HPV-test. We did not correct for deaths and emigrations during the last two years of follow-up, which means that we slightly underestimated completeness of follow-up when we used number of screened women as denominator. Our regional data were not age-adjusted, as the regions in this age-range have similar age distributions. 9 We did not have data on individual cervical screening history, which could have been valuable as about half of cervical cancers are diagnosed in women with no or insufficient screening history. 10  Denmark.
In Denmark, the Parliament decided that the one-time offer of cervical screening to elderly women should have no upper age limit.
So, it caused some uproar from gynecologists when it became clear that about 1000 women aged 100 years or older would be invited for cervical screening. 16 The gynecologists argued that it can be very difficult to take a sample of representative cells from the cervix uteri in elderly women, 17  The argument in Denmark for a one-time screening of elderly women was that older generations of women had not been adequately screened while they were younger. 2 In England, the fact that elderly women had been screened with cytology, which is less sensitive than HPV-testing, has been an additional argument for a "catch-up" HPV-test of elderly women. 20 Arguments for a higher  The effect of cervical screening of elderly women on cervical cancer incidence depends on the women's willingness to participate.
In Australia, 30% of women aged 70-74 were hysterectomized, and 35% of remaining eligible women had a screening sample registered. 13 In Denmark, 15% of women aged 70+ were hysterectomized, and 30% of eligible women were screened; 44% in the age group 69-73 years. 4 Coverage is thus still a challenge when it comes to impact of screening on the national cancer rates. Within the routine screening program, self-sampling has proved to increase coverage in under-screened women. [27][28][29] It remains to be seen if this would be the case also for elderly women.
In the years before the screening of elderly women in Denmark, annually 64 women aged 70+ years were diagnosed with cervical cancer. In the screening, nine times as many women were diagnosed with CIN2+, and 19 times as many women underwent conization. It is therefore reasonable to expect that the one-time screening will lower cervical cancer incidence in elderly women. However, the screening resulted also in a prevalence peak of cervical cancers, and it will take some years before the full effect of the screening on cervical cancer incidence of elderly women can be evaluated. Elderly women and health care authorities will then have a clearer picture of benefit and harm. It was nevertheless promising that the number of cervical cancer cases had decreased to 54 in 2020 and remained at the same low level in 2021.

| CON CLUS ION
In 2017, Danish women aged 70+ received a one-time invitation for primary HPV-screening. The background was a relatively high incidence of cervical cancer in women above screening age. In total, 4.1% of screened women were HPV-positive. Based on HPV-type and cytology triage, about half of the women were recommended cell-sample follow-up, and the other half referral for CBC. In total, five per 1000 screened women had CIN2+ diagnosed, and three per 1000 CIN3+. Among the HPV-positive women 28% had a conisation performed, but less than half of these women had a diagnosis of CIN2+. The screening resulted in a prevalence peak in incident cervical cancers, and it will therefore take some years before the cancer preventive effect of the screening can be evaluated.

AUTH O R CO NTR I B UTI O N S
MS and EL conceptualized and designed the study. PVH undertook the statistical analysis. MS and EL drafted the first version of the manuscript. All authors commented on the table layout and manuscript, and all others accepted the final version for submission.

ACK N OWLED G M ENTS
We thank Danish Health Data Authority for providing updated cervical cancer incidence data.

FU N D I N G I N FO R M ATI O N
This study was financially supported by the Danish Cancer Society (R247-A14608).

CO N FLI C T O F I NTER E S T S TATEM ENT
BA reports non-financial support from Roche and Axlab outside the submitted work. RS is a consultant for Merck. EL reports nonfinancial support from Roche outside the submitted work. The remaining authors report no conflict of interest.