First round of screening, for women aged 24 to 59

Of 1,341,584 women, the first round included 300,677 aged 24-49 screened by HPV testing and 706,820 screened by cytology (Table 2; Supplementary Table S1 and Fig S2). HPV testing detected 5313 women with CIN3+, while cytology detected CIN3+ in 8232 women (17.67 v 11.65/1000 women screened; adjusted odds ratio 1.55, 95% confidence interval 1.50 to 1.61). Cervical cancer was detected in 259 women by HPV test and 441 women by cytology test (0.86 v 0.62/1000 screened; 1.38, 1.18 to 1.61).

Table 2

Test performance in women screened aged 24-59 in the first round of the UK HPV screening pilot; detection of CIN3+ and cervical cancer after a positive screening test by screening round, screening test and age group. Includes first cycles of the pilot in 2013-2016 and second cycles of the pilot by the end of 2019 for women with negative screening test results in the first cycle (cervical cancer information available until the end of 2018)

Eleven interval cancers were diagnosed in 837,960 woman-years after negative HPV tests at ages 24-49, and 62 in 2,135,315 woman-years after negative cytology (1.31 v 2.90/100,000 woman-years; adjusted relative risk 0.44, 95% confidence interval 0.23 to 0.84; table 3, figure 1). In the second round of screening, CIN3+ detection was lower in women screened negative for the HPV test in the first round (227 out of 188,318) than in women screened negative by cytology (1,177 out of 260,266; 1.21/ 1000 screened v 4.52/1000; adjusted odds ratio 0.26, 95% confidence interval 0.23 to 0.30; Table 2). The same was true for the detection of cervical cancer, one woman against 54 women (0.01/1000 screened v 0.21/1000 screened; 0.02, 0.00 to 0.17).

Table 3

Interval cervical cancer diagnosed between the first and second round of screening after a negative screening*, by screening test and age group. Includes women who tested negative in the first round of the English HPV screening pilot in 2013-2016, with information on cervical cancer available through the end of 2018

Fig. 1

Cumulative incidence of interval cervical cancer after a negative first-round screening test result, by age group or HPV test type. Interval cancers are those diagnosed between the first and second cycles of screening. Shaded areas indicate 95% confidence intervals. LBC = liquid-based cytology; HPV=human papillomavirus

A total of 45% (12,709) of women aged 24-49 years with positive and cytologically negative HPV screening specimens had a negative HPV screening test either at the early recall of 12 months (10,248, 81%) or at early recall of 24 months (2461, 19%; not tabulated). Among these 12,709 women, 5380 had been screened in the second round, with the detection of 29 CIN3+ lesions (5.39/1000 screened), none of which was cancerous. Compared with the 188,318 women who tested negative for HPV in the first round (Table 2), the risk in the second round was higher for women who tested negative for HPV at early recall (adjusted odds ratio 3, 27, 95% confidence interval 2.21 to 4.84) ​​. Of the 5,380 women with second-round data, 4,181 were screened at one of the four labs that used HPV genotyping testing. Here, CIN3+ was detected in the second round in 8.9/1000 women infected with HPV16/18, whereas this proportion was 4.5/1000 in women infected with 12 other high-risk genotypes. Although the risk per 1000 women is higher for HPV16/18 infections, approximately two-thirds of all CIN3+ lesions were detected in women infected with other genotypes (7 v 15, respectively).

At age 50-59, 79,040 women were screened with an HPV test and 175,973 with first-round pilot cytology (Table 2); CIN3+ was detected in 219 and 332 women, respectively (2.77 v 1.89/1000 screened; adjusted odds ratio 1.56, 95% confidence interval 1.31 to 1.85). With 31 and 52 cases detected in women screened by HPV testing and those screened by cytology, respectively (0.39 v 0.30/1000 screened), the adjusted odds ratio for the detection of cervical cancer in the first round was 1.41 (95% confidence interval from 0.90 to 2.21). Two women had interval cancer diagnosed during the 281,369 woman-years after a negative HPV test, as did 17 women out of 631,584 woman-years after a negative cytology (0.71 v 2.69/100,000 woman-years; relative risk 0.26, 95% confidence interval 0.06 to 1.15; Fig. 1).

In the second round, CIN3+ was detected in 14 of the 24,550 women screened by HPV test and in 20 of the 22,195 women screened by cytology (0.57 v 0.90/1000 dithered; odds ratio 0.63, 95% confidence interval 0.32 to 1.25). In the second round, one cancer was detected with HPV screening and two with cytology. Of 1570 women who were HPV positive at screening but HPV negative at early recall, 98 had been re-screened by the end of 2019, with no cases of CIN3+ reported (not tabulated).

With first-round HPV testing, detection of both CIN3+ (adjusted odds ratio 0.16, 95% confidence interval 0.14 to 0.18) and cervical cancer (0. 46, 0.32 to 0.67; based on data reporting lesion detection in Table 2) was significantly lower at ages 50-59 than at ages 24-49. Significantly fewer CIN3+ cases were also detected in the second round of screening five years after a negative HPV screen at age 50 to 59 than in the second round three years after a negative HPV screen at age 24 to 49 years (adjusted odds ratio 0.46, 95% confidence interval 0.27 to 0.79).

We compared 51,639 women screened with an HPV mRNA test (APTIMA) and 292,831 women screened with a first-round HPV DNA test (cobas or RealTime; Table 4). Detection was similar for both types of tests for CIN3+ (adjusted odds ratio 1.04, 95% confidence interval 0.96 to 1.12) and cancer (0.95, 0.67 to 1.34) in the first round, as well as for the detection of CIN3+ in the second round after negative tests in the first round (1.05, 0.73 to 1.50). One woman had interval cancer diagnosed after a negative APTIMA test versus 12 women after a negative cobas or RealTime test, resulting in a hazard ratio of 0.51 (95% confidence interval 0.07 to 3 ,91; table 5, fig 1). The adjusted odds ratio for second-round detection of CIN3+ was 1.38 (95% confidence interval 0.91 to 2.08) for APTIMA and 1.65 (1.22 to 2.23) for RealTime versus cobas. For interval cancer incidence, the hazard ratio was 0.35 (95% confidence interval 0.04 to 2.84) for APTIMA and 0.40 (0.12 to 1.33 ) for RealTime compared to cobas.

Table 4

Test performance in women screened aged 24-59 in the first round of the UK HPV screening pilot; detection of CIN3+ and cervical cancer after a positive screening test, by type of HPV detection target and by screening cycle. Includes first cycles of the pilot in 2013-2016 and second cycles of the pilot by the end of 2019 for women with negative screening test results in the first cycle (cervical cancer information available until the end of 2018)

Table 5

Interval cancer diagnosed between the first and second round of screening after a negative screen, by HPV detection target type. Includes women who tested negative in the first round of the English HPV screening pilot in 2013-2016, with information on cervical cancer available through the end of 2018