The detection of CIN2 and CIN3 during each study
period, among women aged 35–60 years at enrolment, is
shown in table 3. There was no signifi cant diff erence in
the group eff ect between study phases (p for heterogeneity
>0·15 in all comparisons). During the fi rst screening
round, the detection of CIN2 or CIN3 was signifi cantly
higher in the HPV group versus the cytology group, with
a detection ratio of 2·03 (95% CI 1·60–2·57). Conversely,
in the second round, the detection of CIN2–3 was
signifi cantly lower in the HPV group versus the cytology
group, with a ratio of 0·51 (0·28–0·93). For the total
detection of disease during the fi rst two screening
rounds, the ratio between the HPV group and the cytology
group was 1·66 (1·34–2·06). Results were similar for
detection of CIN2 and CIN3, separately. Among women
35 years and older, age (considered a continuous variable
in logistic regression) was not a modifi er of the group
eff ect on detection of CIN2 and 3 in round one, or in the
two rounds combined. However, the modifi cation of
group eff ect by age was of borderline signifi cance
(beta=–0·17; p=0·058) for the detection of CIN3 in round
two, suggesting that the gain in protection obtained by
HPV testing could increase with increasing age.
Among women aged 25–34 years at recruitment, there
was a signifi cant diff erence between the two recruitment
phases in the relative detection of CIN3 by HPV testing
(table 4). For phase one, the detection ratio between
groups was close to 1, both during the fi rst and the second
trial rounds. In phase two, detection of CIN3 was much
higher in the HPV group than in the cytology group in
the fi rst screening round, and lower in the second round.
In phase two, the total detection ratio, for both screening
rounds, of CIN3 in the HPV group versus the cytology
group was 2·14 (95% CI 1·28–3·59). These eff ects were
observed both among women aged 25–29 and 30–34 years
at recruitment (table 5).
For ages 25–34 years, pooling both phases, the detection
ratio of CIN2 in the HPV versus cytology group was
4·54 (95% CI 3·00–6·88) in the fi rst round and
0·54 (0·23–1·27) in the second round, with no heterogeneity
between phases. For the total detection of CIN2
over the fi rst two screening rounds, the ratio between the
HPV and cytology group was 3·11 (2·20–4·39). This ratio
was 3·84 (2·26–6·52) among women aged 25–29 years at
recruitment and 2·61 (1·65–4·13) among women aged
30–34 years, with no evidence of heterogeneity between
these two subgroups (table 5).
For all ages combined, no evidence of heterogeneity
between centres was found for the eff ect of HPV
screening on the detection of high-grade CIN during
round one, round two, or both combined.