- Top of page
- Disclosure of interests
- Contribution to authorship
- Details of ethics approval
Please cite this paper as: Stock S, Norman J. Treatments for precursors of cervical cancer and preterm labour. BJOG 2012;119:647–649.
Women who have treatment for cervical cancer precursors are at an increased risk of preterm birth. It is generally thought that this risk relates to a decrease in the volume of the cervix, as in some studies the risk of preterm labour has been shown to increase with the volume of tissue removed,1,2 and the number of procedures,3 suggesting a ‘dose response’ to tissue removal. In England in 2009–2010 over 23 800 cervical excisional procedures were carried out, the majority being performed in women of childbearing age.4 Around 98% were performed by large loop excision of the transformation zone (LLETZ), which, although lower in risk than knife cone biopsy,2 is associated with poor pregnancy outcomes, including preterm birth, spontaneous preterm labour, preterm prelabour rupture of membranes, low birthweight and perinatal mortality.2,3,5–7 Overall, LLETZ is associated with a between two- and ten-fold increased relative risk of preterm delivery, which is similar to the risk of preterm labour in a woman who has had a previous spontaneous preterm delivery.
Three articles in this edition of the journal investigate the relationship between LLETZ excision and preterm delivery.8–10 Papoutsis et al.8 used 3D transvaginal ultrasound to investigate regeneration of the cervix following LLETZ, prospectively recruiting 73 women undergoing LLETZ, and measuring cervical volume both before and 6 months after the procedure. They related cervical regeneration to the volume of tissue removed, and found that the smaller the volume of tissue excised, the greater the regeneration at 6 months. The authors used an arbitrary definition of ‘incomplete regeneration’, based on a deficit in cervical volume of 25% at 6 months, compared with the pretreatment volume. Removal of more than 11% of pretreatment cervical volume by LLETZ was associated with ‘incomplete regeneration’. It is unclear whether this definition of incomplete regeneration has clinical significance in terms of future pregnancy outcome, but these findings support the hypothesis that increased tissue excision compromises cervical integrity.
Khalid et al.9 also provide evidence of an association between tissue excision and preterm birth. They performed a retrospective study of women with a singleton pregnancy following LLETZ, and examined the relationship between volume, length (distance from distal to proximal margin), and thickness (distance from stromal margin to surface) of the tissue removed and the gestational age at delivery. A total of 344 women were eligible for inclusion, and data was available for 321 of these. Greater specimen volume and thickness was associated with preterm birth (later than 24 weeks and less than 37 weeks of gestation). There was a three-fold increase in the risk of preterm delivery if the excision volume exceeded 6 cm3 (RR 3.00, 95% CI 1.45–5.92), and if the thickness was >12 mm (RR 3.05, 95% CI 1.31–7.08). No association was found between the time from LLETZ to pregnancy and preterm birth.
Poon et al.10 aimed to investigate whether the risk of spontaneous preterm labour associated with LLETZ is reflected in cervical length measured by transvaginal ultrasound at mid-gestation. They performed a secondary analysis of 26 867 singleton pregnancies, with cervical length measurement at 20–25 weeks of gestation. They found that women who had previous LLETZ (n = 473) had a 3.4% rate of spontaneous preterm labour before 34 weeks of gestation, compared with a 1.3% rate in women without LLETZ. The median cervical length was lower in the group with previous LLETZ, at 32 mm, compared with 34 mm in the group without LLETZ. They found that when cervical length was included in a logistic regression model, LLETZ was no longer independently associated with the likelihood of preterm labour, indicating that any risk of preterm birth associated with LLETZ is entirely attributable to a reduction in cervical length.
These three studies demonstrate some of the complexities of research in this area. Firstly, a lack of uniform definitions limits the comparability of study findings. For example, Papoustis et al.8 report cone ‘depth’, and measured ‘volume’ of the cone by displacement of fluid, using the Archimedes’ principal. Khalid et al.9 used an alternative approach, reporting ‘length’ and ‘thickness’ of the cone, and calculating the volume based on these measurements along with the perimeter. A further complication is the observation that the ratio of biopsy to cervical size may be more important than its absolute dimensions, as information about pre-treatment cervical volume is not often available. The careful description of obstetric outcomes is equally important. Cervical treatment is most likely to cause pregnancy complications relating to ‘incompetence’, i.e. spontaneous preterm labour or mid-trimester loss. To most accurately examine the effect of cervical treatment, spontaneous preterm labour should be differentiated from medically indicated preterm birth. Khalid et al.9 did not make these distinctions, so their findings may underestimate the effect of LLETZ on preterm labour. Furthermore, mid-trimester losses were not included, perhaps because late miscarriages are not always recorded in maternity data collection systems. This has contributed to a relative lack of evidence regarding the relationship between cervical treatment and mid-trimester pregnancy loss. It seems clear that establishing consistent definitions regarding pre-treatment findings, the volume of tissue removed during cervical treatment and obstetric outcomes would help in the interpretation of research findings.
Secondly, cervical cancer precursors and spontaneous preterm labour share common risk factors, such as smoking and social deprivation.11 These may confound findings of observational studies. Despite the growing body of evidence that tissue excision is associated with preterm labour, some observational studies have suggested an increased risk of preterm birth associated with CIN itself, regardless of treatment.11–14 This association has been attributed to the fact that some women are at high risk of both CIN and preterm birth, or that both conditions are mediated by a common process, such as infection. Prospective studies are required to fully assess the relationship between pre-invasive cervical disease, its treatment and obstetric outcomes. These should include appropriate comparison groups, such as women who attend colposcopy who have ablative treatments or no treatment. Such trials require a collaborative approach, with a commitment from funders for adequate follow-up in order to capture complications, which may occur many years after treatment. However, they are essential to determine optimal management strategies that minimize both the recurrence of pre-invasive cervical disease and poor obstetric and neonatal outcomes.
The prediction of preterm labour in women who have had excisional treatment needs to be improved if successful treatments are to be initiated. Although mid-trimester cervical length measurement is associated with preterm labour, in the study by Poon et al.10 the prediction rate of spontaneous labour earlier than 34 weeks of gestation was only 52.6% (95% CI 47.3–57.7) and 32.7% (95% CI 29.6–35.9) for preterm labour at less than 37 weeks of gestation, even when cervical length was combined with maternal characteristics and obstetric history. This suggests that a single mid-trimester cervical length measurement is of limited clinical value. Serial cervical length ultrasounds from earlier in gestation, coupled with other predictors such as fetal fibronectin, may be a better approach. However, in order to assess clinical effectiveness, studies must evaluate the predictive ability of these tests, in conjunction with interventions to prevent preterm labour. Treatments applied to the cervix, such as progesterone or cervical cerclage, show the greatest potential to improve outcomes in women at high risk of preterm labour, although there still remains little evidence that neonatal outcomes are improved with their use.
One barrier to the development of both predictive tests and effective treatments to prevent preterm labour is the current lack of understanding in how excisional treatments affect cervical function and biology. Excision of tissue may reduce the tensile strength of the cervix, allowing it to open with pressure from the increasing weight of intrauterine contents, or with mild myometrial contractions. Alternatively, the excision of functional epithelium and immune cells may disrupt cervical immunity, predisposing the cervix to infection or inflammation that can cause preterm labour. Normally the immune system of the cervix is well developed to protect it from the diversity of microbes and foreign antigens it is exposed to. The epithelia expresses toll-like receptors to detect potential pathogens, produces cytokines that mediate inflammation and produces antimicrobial peptides that have both direct antibacterial effects and act as signalling molecules to the immune cells.15 Endogenous macrophages, neutrophils, dendritic cells, and natural killer cells are also present in the cervical epithelium and stroma.16 There is increasing evidence to support the theory that dysregulation of these cervical innate immune components is involved in the pathogenesis of preterm labour. For example, in early pregnancy low cervicovaginal cytokine concentrations have been associated with subsequent chorioamnionitis,17 indicating that immune hyporesponsiveness may predispose the woman to ascending infection. Low cervicovaginal levels of antimicrobial peptides in pregnancy have been associated with bacterial vaginosis, which itself is associated with preterm labour.18 Lastly, women with idiopathic preterm labour have fewer cervical macrophages at the beginning of the second trimester of pregnancy than those whose pregnancy progressed beyond 35 weeks of gestation.19 These findings support the hypothesis that cervical innate immune function is important for pregnancy maintenance. It seems likely that excisional treatments for cervical cancer precursors may disrupt both the structural and functional integrity of the cervix, and its capacity to effectively retain pregnancy to term.
Given the number of women undergoing treatment for pre-invasive cervical cancer, and the increasing prevalence of preterm labour, improving obstetric outcomes for women requiring treatment for cervical cancer precursors is becoming a crucial issue in women’s health. The data reported to date suggest that minimizing unnecessary tissue removal during cervical treatment in women of reproductive age seems a sensible approach. Further investigation is needed to determine the optimal approach for treating cervical intraepithelial neoplasia (CIN) to effectively treat pre-invasive disease with minimal adverse effect on future reproductive health, including pregnancy outcome. This includes research to determine the effect of LLETZ on cervical function and regeneration, and how best to manage those women at highest risk.