Sir,

We agree with Castanon and her colleagues (Castanon et al, 2013a) that the more than two-fold increase in cervical cancers registered in women aged 25–29 years in England in the last decade cannot entirely be explained by the cessation of screening women aged 20–24 years, which was first recommended in 2003. Nevertheless, we cannot believe that policy has not had some effect on the increase. Incidence of invasive cervical cancer per 100 000 women aged 25–29 years was higher in 2011 than the previous highest level in that age group: 19.3 compared with 14.8 in 1986 (Office for National Statistics).

Since 1992, registrations in England as a whole of invasive carcinoma of the uterine cervix in women aged 25–29 years have consistently represented 3% of total registrations of invasive and in situ cancer combined (cervical intraepithelial neoplasia grade 3, CIN3, is registered as carcinoma in situ), and the two diagnoses have increased in parallel, including during the so-called ‘Jade Goody effect’ in 2009, which is consistent with most of these cancers being screen-detected (Figure 1). The number of increased registrations of CIN3 since 2004 in women aged 25–29 years (the peak age group for CIN3 since the late 1980s) is greater than the simultaneous decrease in women aged 20–24 years, suggesting an increased risk in women born between about 1977 and 1983 (marked ‘+’ in Figure 1), which was before the effect of the new policy would have taken effect. This could be related to the year on year increase in the percentage of women being sexually active before the age of 16 years: 29.2% according to the latest Natsal survey (Mercer et al, 2013). It seems to us feasible that failing to treat several thousand women with CIN3, along with similar numbers with CIN2 (about a third of which progress to CIN3), may have contributed to the increase in invasive cancers seen in women aged 25–29 years.

Figure 1
figure 1

Registrations of in situ (CIN3) and invasive carcinoma of the uterine cervix in England: women aged 20–29 years in England 1992–2011 (Office for National Statistics data). '+', Women born 1977–1981, 1978–1982 and 1979–1983.

Full size image

Castanon et al (2013a) cite an article of ours (Herbert et al, 2008) that provided preliminary results of a 9-year audit of 133 cervical cancers that was published in 2010 along with concurrent cases of CIN2+ diagnosed at Guy’s and St Thomas’ (Herbert et al, 2010). In that audit we defined screen-detected cancers as cases diagnosed in asymptomatic women investigated for abnormal cytology and found that 15 (83.3%) of 18 cancers in women aged 20–29 years were screen-detected IA or IB1 cancers. The treatment of IA cancer is not the same as that of CIN3 as suggested by Castanon et al (2013a), and the effect of a ‘cancer diagnosis’ on a woman as young as 25–29 years may be devastating. During the period of our audit, only 3 of 41 IA cancers had a single large loop excision of the transformation zone (LLETZ), compared with 85 of 100 cases of CIN3 (Table 1). The most frequent treatment of IA cancer was LLETZ followed by knife cone biopsy, because many of these cancers arise in widespread CIN3 that may be difficult to excise completely on a LLETZ; 5 had trachelectomy and 15 had hysterectomies. Most women with CIN3 had a single LLETZ; those who had further treatment tended to be slightly older. LLETZ is less likely to cause premature rupture of membranes than repeated or larger excisional biopsies (Castanon et al, 2013b).

Table 1 Treatment of stage IA cervical carcinoma compared with CIN3
Full size table

Disallowing screening for women aged 20–24 years, whatever their clinical history of sexual activity, is an experiment that is unfortunately taking place during a period of time when there are birth cohorts at increased risk and screening coverage is falling in younger women.7 In our opinion, the view that screening women under age 25 years causes ‘more harm than good’ is letting down a generation of women who are above the ages of those who will benefit from vaccination in the future.