Obstetric outcomes after conservative treatment for cervical intraepithelial lesions and early invasive disease

Maria Kyrgiou; Antonios Athanasiou; Ilkka E J Kalliala; Maria Paraskevaidi; Anita Mitra; Pierre PL Martin‐Hirsch; Marc Arbyn; Phillip Bennett; Evangelos Paraskevaidis

doi:10.1002/14651858.CD012847

. 2017 Nov 2;2017(11):CD012847. doi: 10.1002/14651858.CD012847

Obstetric outcomes after conservative treatment for cervical intraepithelial lesions and early invasive disease

Maria Kyrgiou ^1,^✉, Antonios Athanasiou ², Ilkka E J Kalliala ³, Maria Paraskevaidi ⁴, Anita Mitra ³, Pierre PL Martin‐Hirsch ⁵, Marc Arbyn ⁶, Phillip Bennett ⁷, Evangelos Paraskevaidis ²

Editor: Cochrane Gynaecological, Neuro‐oncology and Orphan Cancer Group

PMCID: PMC6486192 PMID: 29095502

Abstract

Background

The mean age of women undergoing local treatment for pre‐invasive cervical disease (cervical intra‐epithelial neoplasia; CIN) or early cervical cancer (stage IA1) is around their 30s and similar to the age of women having their first child. Local cervical treatment has been correlated to adverse reproductive morbidity in a subsequent pregnancy, however, published studies and meta‐analyses have reached contradictory conclusions.

Objectives

To assess the effect of local cervical treatment for CIN and early cervical cancer on obstetric outcomes (after 24 weeks of gestation) and to correlate these to the cone depth and comparison group used.

Search methods

We searched the following databases: Cochrane Central Register of Controlled Trials (CENTRAL; the Cochrane Library, 2017, Issue 5), MEDLINE (up to June week 4, 2017) and Embase (up to week 26, 2017). In an attempt to identify articles missed by the search or unpublished data, we contacted experts in the field and we handsearched the references of the retrieved articles and conference proceedings.

Selection criteria

We included all studies reporting on obstetric outcomes (more than 24 weeks of gestation) in women with or without a previous local cervical treatment for any grade of CIN or early cervical cancer (stage IA1). Treatment included both excisional and ablative methods. We excluded studies that had no untreated reference population, reported outcomes in women who had undergone treatment during pregnancy or had a high‐risk treated or comparison group, or both

Data collection and analysis

We classified studies according to the type of treatment and the obstetric endpoint. Studies were classified according to method and obstetric endpoint. Pooled risk ratios (RR) and 95% confidence intervals (CIs) were calculated using a random‐effects model and inverse variance. Inter‐study heterogeneity was assessed with I² statistics. We assessed maternal outcomes that included preterm birth (PTB) (spontaneous and threatened), preterm premature rupture of the membranes (pPROM), chorioamnionitis, mode of delivery, length of labour, induction of delivery, oxytocin use, haemorrhage, analgesia, cervical cerclage and cervical stenosis. The neonatal outcomes included low birth weight (LBW), neonatal intensive care unit (NICU) admission, stillbirth, perinatal mortality and Apgar scores.

Main results

We included 69 studies (6,357,823 pregnancies: 65,098 pregnancies of treated and 6,292,725 pregnancies of untreated women). Many of the studies included only small numbers of women, were of heterogenous design and in their majority retrospective and therefore at high risk of bias. Many outcomes were assessed to be of low or very low quality (GRADE assessment) and therefore results should be interpreted with caution. Women who had treatment were at increased overall risk of preterm birth (PTB) (less than 37 weeks) (10.7% versus 5.4%, RR 1.75, 95% CI 1.57 to 1.96, 59 studies, 5,242,917 participants, very low quality), severe (less than 32 to 34 weeks) (3.5% versus 1.4%, RR 2.25, 95% CI 1.79 to 2.82), 24 studies, 3,793,874 participants, very low quality), and extreme prematurity (less than 28 to 30 weeks) (1.0% versus 0.3%, (RR 2.23, 95% CI 1.55 to 3.22, 8 studies, 3,910,629 participants, very low quality), as compared to women who had no treatment.

The risk of overall prematurity was higher for excisional (excision versus no treatment: 11.2% versus 5.5%, RR 1.87, 95% CI 1.64 to 2.12, 53 studies, 4,599,416 participants) than ablative (ablation versus no treatment: 7.7% versus 4.6%, RR 1.35, 95% CI 1.20 to 1.52, 14 studies, 602,370 participants) treatments and the effect was higher for more radical excisional techniques (less than 37 weeks: cold knife conisation (CKC) (RR 2.70, 95% CI 2.14 to 3.40, 12 studies, 39,102 participants), laser conisation (LC) (RR 2.11, 95% CI 1.26 to 3.54, 9 studies, 1509 participants), large loop excision of the transformation zone (LLETZ) (RR 1.58, 95% CI 1.37 to 1.81, 25 studies, 1,445,104 participants). Repeat treatment multiplied the risk of overall prematurity (repeat versus no treatment: 13.2% versus 4.1%, RR 3.78, 95% CI 2.65 to 5.39, 11 studies, 1,317,284 participants, very low quality). The risk of overall prematurity increased with increasing cone depth (less than 10 mm to 12 mm versus no treatment: 7.1% versus 3.4%, RR 1.54, 95% CI 1.09 to 2.18, 8 studies, 550,929 participants, very low quality; more than 10 mm to 12 mm versus no treatment: 9.8% versus 3.4%, RR 1.93, 95% CI 1.62 to 2.31, 8 studies, 552,711 participants, low quality; more than 15 mm to 17 mm versus no treatment: 10.1 versus 3.4%, RR 2.77, 95% CI 1.95 to 3.93, 4 studies, 544,986 participants, very low quality; 20 mm or more versus no treatment: 10.2% versus 3.4%, RR 4.91, 95% CI 2.06 to 11.68, 3 studies, 543,750 participants, very low quality). The comparison group affected the magnitude of effect that was higher for external, followed by internal comparators and ultimately women with disease, but no treatment. Untreated women with disease and the pre‐treatment pregnancies of the women who were treated subsequently had higher risk of overall prematurity than the general population (5.9% versus 5.6%, RR 1.24, 95% CI 1.14 to 1.34, 15 studies, 4,357,998 participants, very low quality).

pPROM (6.1% versus 3.4%, RR 2.36, 95% CI 1.76 to 3.17, 21 studies, 477,011 participants, very low quality), low birth weight (7.9% versus 3.7%, RR 1.81, 95% CI 1.58 to 2.07, 30 studies, 1,348,206 participants, very low quality), NICU admission rate (12.6% versus 8.9%, RR 1.45, 95% CI 1.16 to 1.81, 8 studies, 2557 participants, low quality) and perinatal mortality (0.9% versus 0.7%, RR 1.51, 95% CI 1.13 to 2.03, 23 studies, 1,659,433 participants, low quality) were also increased after treatment.

Authors' conclusions

Women with CIN have a higher baseline risk for prematurity. Excisional and ablative treatment appears to further increases that risk. The frequency and severity of adverse sequelae increases with increasing cone depth and is higher for excision than it is for ablation. However, the results should be interpreted with caution as they were based on low or very low quality (GRADE assessment) observational studies, most of which were retrospective.

Plain language summary

Obstetric outcomes after conservative treatment for cervical intraepithelial lesions

The issue   Cervical intra‐epithelial neoplasia (CIN) is a pre‐cancerous lesion of the cervix uteri (neck of the womb) caused by human papillomavirus (HPV), which may develop into cervical cancer, if not treated. Local treatment involves destroying or removing the abnormal area of the cervix, leaving most of the cervix, and the uterus in place maintaining the ability to become pregnant in the future, if desired. Certain types of local treatment may also be suitable for very early cervical cancer (stage IA1) if the tumour is very small and very unlikely to have spread beyond the cervix. There are many studies investigating whether the local treatment for CIN and early cervical cancer increases the risk of preterm birth (PTB) in subsequent pregnancies. However, there is no definite conclusion and this creates confusion for both the medical staff and women who may be recommended treatment, but also want to have children in the future.

The aim of the review   We aimed to assess whether the local conservative treatment techniques for cervical precancer (CIN) and early cervical cancer increased the risk of complications for mother and baby during pregnancy occurring after treatment, and especially whether treatment is associated with an increase in the risk of PTB. We also studied whether the risk of PTB increases with increasing amount of cervical tissue removed.

Selection criteria   We included all studies that investigated the effect of treatment of CIN and early cervical cancer on late pregnancy outcomes (beyond 24 weeks of gestation) in women who had been treated previously for CIN and early cervical cancer, as compared to women who had not been treated. We excluded studies that had no untreated comparison group, reported pregnancy outcomes in women who had undergone treatment during pregnancy, or had a high‐risk treated, comparison group or both.

What are the main findings?   We included 69 studies (6,357,823 pregnancies: 65,098 pregnancies of treated and 6,292,725 pregnancies of untreated women). Treatment was associated with an increased risk of PTB before 37 pregnancy weeks, as well as an increased risk of severe PTB (less than 32 to 34 pregnancy weeks), extreme PTB (less than 28 to 30 pregnancy weeks) and pPROM (premature preterm rupture of the membranes) as compared to untreated women. The risk of overall PTB was higher for women treated by excisional methods (where tissue is cut away) than by ablative treatments (where tissue is destroyed instead of being cut away). Multiple treatments, as well as increasing amounts of tissue removed at the time of treatment, were associated with an increased risk of overall PTB. However, women with CIN who were not treated also had a higher risk of overall PTB than the general population. Low birth weight (LBW) < 2500g), neonatal intensive care unit (NICU) admission and perinatal mortality rates were also found to be increased after treatment.

What is the quality of the evidence?   Due to the nature of the intervention and outcomes studied, we were only able to include observational studies, of which the majority were retrospective. These types of studies are of low quality with a high level of variability between the studies, therefore the level of evidence for most outcomes can only be considered to be of low or very low quality.

What are the conclusions?   Women with CIN have a higher baseline risk for PTB than the general population and the treatment for CIN probably increase this risk further. The risk for PTB is probably higher when excisional techniques are used than for ablative treatments. Also, the risk of PTB appears to increase with multiple treatments and increasing amounts of tissue removed. However, these results should be interpreted with caution due to the low and very low quality of the included studies.

Summary of findings

Background

Description of the condition

Cervical cancer remains the commonest gynaecological malignancy worldwide, and accounts for 7.5% of female cancer deaths. Over half a million new cases are diagnosed each year around the world, with the vast majority occurring in the developing world, where a woman's risk of cervical cancer by age 74 is almost double that in the developed world (1.6% versus 0.9%) (Ferlay 2015).The introduction of cervical screening programmes over the last 20 years has produced a profound decrease in the incidence and mortality from cervical cancer (Arbyn 2009; Quinn 1999). This is due to the treatment of pre‐invasive lesions, cervical intra‐epithelial neoplasia (CIN), detected by screening (IARC 2005).

CIN is an abnormality in the squamous cells of the cervix and, if left untreated, cervical cancer may develop. The condition is asymptomatic and interventions are usually performed only on women with higher grade CIN of grade (CIN 2 or 3). This is because cervical treatment has been correlated with adverse obstetric sequelae (Kyrgiou 2006), while many of the low‐grade lesions (also known as LSIL, low‐grade squamous intra‐epithelial lesions or CIN 1) resolve spontaneously in young individuals (NHS Cervical Screening Programme 2016).

The average age of a woman diagnosed and treated for CIN is between 25 and 30 years of age, although it may occur in women considerably younger (NHS Cervical Screening Programme 2016). As the pre‐cancerous lesions typically occur in young women of reproductive age, the impact of their treatment on the outcomes of subsequent pregnancies has been an area of active research for the past decade. Whilst it is paramount that effective treatment is undertaken, it is also important that this treatment has minimal adverse effects on future fertility and pregnancy outcomes for this young female population.

Description of the intervention

The conservative methods for treatment of CIN are classified into excisional and ablative. These techniques remove or destroy the transformation zone (TZ) containing the abnormal cells whilst preserving cervical function. Excisional methods include cold knife conisation (CKC), laser conisation (LC), needle excision of the transformation zone (NETZ), also known as straight wire excision of the transformation zone (SWETZ), large loop excision of the transformation zone (LLETZ) (Kitchener 1995; Prendiville 1989), also known as loop electrosurgical excisional procedure (LEEP) and Fischer cone biopsy excisor (FCBE). Ablative methods include laser ablation (LA), radical diathermy (RD), cold coagulation (CC) and cryotherapy (CT).

The mean age of women undergoing treatment for pre‐invasive cervical disease is around their 30s and similar to the age of women having their first child (Herbert 2000; Paraskevaidis 1992). Local cervical treatment has been correlated with an increased risk of preterm birth, perinatal morbidity and mortality in a subsequent pregnancy (Albrechtsen 2008; Arbyn 2008; Bruinsma 2011; Kyrgiou 2006; Kyrgiou 2014; Noehr 2009a). The underlying mechanism is unclear; hypotheses include immunomodulation relating to human papillomavirus (HPV) infection affecting parturition pathways, unexplained confounders in women with CIN and acquired ‘mechanical weakness’ secondary to loss of cervical tissue (Kyrgiou 2012).

In England alone in 2013 to 2014, 3.6 million women aged between 25 and 64 years attended for cervical screening and over 23,800 cervical procedures were carried out (CervicalCancerScreening 2015), the vast majority in an outpatient setting. In contrast in the USA, there are approximately 400,000 cases of pre‐invasive disease per year (Henk 2010). The regulations in colposcopy are more liberal than in the UK leading to wide variation in clinical practice. In Germany, treatment for CIN is still commonly performed with the cold knife under general analgesia (Petry 2008). The long‐term sequelae of treatment remains therefore an important international issue to both healthcare professionals and women, whatever the clinical setting.

How the intervention might work

The characteristics of the conservative methods of treatment are well‐described and established in the medical literature (Martin‐Hirsch 2013). LLETZ, LC and LA are usually performed under local anaesthesia in an outpatient setting, while CKC requires general anaesthesia and hospitalisation. Theoretically, the excisional techniques (CKC, LC, NETZ, LLETZ, FCBE) are superior over ablative techniques (LA, RD, CC, CT) as they allow a comprehensive histological evaluation of the excised tissue and the whole TZ with precise evaluation of excision margins. Ablative techniques destroy the TZ epithelium and preclude histological evaluation and, therefore, demand accurate pre‐treatment biopsy at a separate visit. LLETZ is the most favoured technique (Kitchener 1995), by combining all the advantages of the excisional techniques mentioned above together with a relatively shorter duration, low cost, good compliance, simplicity and easier learning curve for practitioners.

The best available evidence suggests that these methods (CKC, LC, LLETZ, LA) present similar low morbidity and are equally successful, in terms of eradicating CIN (Martin‐Hirsch 2013; Nuovo 2000) and in preventing invasive cervical cancer (Chew 1999; Paraskevaidis 1991; Soutter 1997). However, the existing data regarding future fertility and pregnancy outcomes are conflicting.

Why it is important to do this review

Observational studies have indicated that the treatment of CIN could have detrimental effects on fertility and pregnancy outcome; although the conclusions are usually equivocal, perhaps due to the weakness associated with small sample sizes used in the studies. We have not found any published randomised controlled trials (RCTs) comparing pregnancy outcome between treated and untreated women in the literature and because of the pre‐malignant nature of the condition treated, it is perhaps unlikely that one will be ever conducted. Thus, the only available level of evidence on this subject may have to be provided by systematic reviews and meta‐analyses of observational studies.

Media publicity has heightened public awareness that treatment for cervical precancer is associated with an increased reproductive morbidity. There has been a substantial increase in enquiries from patients and clinicians on the risks associated with different treatment techniques and cone depths (Founta 2010; Kyrgiou 2015a), and as to how this risk may be managed and prevented. With a rapidly evolving evidence base and lack of a robust synthesis of the published literature, these questions are becoming increasingly difficult to answer.

Since the first systematic review of the reproductive risk associated with treatment almost a decade ago (Kyrgiou 2006), more than 50 observational studies have been published confirming (Jakobsson 2007; Ortoft 2010) or disputing these associations (Castanon 2012; Reilly 2012); some of these reporting data from large population‐based datasets. Individual attempts to synthesise parts of this rapidly evolving evidence base in small systematic reviews and meta‐analyses reached contradictory conclusions (Arbyn 2008; Bruinsma 2011; Conner 2014; Crane 2003; Danhof 2015; Jin 2014; Kyrgiou 2006; Kyrgiou 2014;) and initiated debates and confusion within the scientific community (Arbyn 2008; Conner 2014; Crane 2003; Danhof 2015; Jin 2014). Whether these discrepancies were due to questionable quality of some of these primary and secondary studies or differences in the explored comparisons (Bruinsma 2011; Conner 2014; Danhof 2015; Jin 2014), the subject is open to a definitive comprehensive high‐quality synthesis of the existing evidence that will be highly informative to women, clinicians and policy makers (Arbyn 2008; Bruinsma 2011; Conner 2014; Danhof 2015; Jin 2014; Kyrgiou 2014). Because many large studies (Albrechtsen 2008; Bruinsma 2007; Castanon 2012; Heinonen 2013; Jakobsson 2007; Jakobsson 2009; Noehr 2009a; Ortoft 2010; Reilly 2012; Shanbhag 2009) have been published since Kyrgiou's first meta‐analysis (Kyrgiou 2006), we decided to update it in order to incorporate the latest studies and pay attention especially to the effect of the comparison group and the depth of the excised cone; two areas on which other meta‐analyses do not emphasise.

Objectives

Methods

Criteria for considering studies for this review

Types of studies

We included all studies reporting on late obstetric outcomes (beyond 24 weeks of gestation) in women with one or more previous local cervical treatments for CIN or early invasive disease (stage IA1), as compared to women without treatment. The interventions included any type of conservative treatment, either excisional or ablative (See:Types of interventions).

Studies were included irrespective of the type of untreated control group, which could have been drawn from one of the following sources: a) external group from general population that was mostly matched or adjusted for confounders; b) internal group with self‐matching of the pregnancies for the same women before and after treatment; c) internal group with the pre‐treatment and post‐treatment pregnancies of a given population; d) women attending colposcopy with or without CIN/biopsy but no treatment; e) women with high‐grade disease but no treatment (high‐grade squamous intra‐epithelial lesion (HSIL)). As the studies are non‐randomised, the choice of comparison group can impact on the magnitude of effect of the proposed comparisons. We know that women with CIN may have demographic and behavioural characteristics or even background immunological imbalances that place them at higher baseline risk of adverse reproductive outcomes. The different comparison groups have advantages and disadvantages and subgroup analyses for the different groups will allow better assessment of the true effects of treatment. More details are described in Assessment of risk of bias in included studies.

We excluded studies that did not include an untreated control group, compared different treatment techniques without an untreated control, reported on only fertility or early obstetric outcomes (before 24 weeks of gestation), reported only on obstetric outcomes beyond 24 weeks of gestation that are not listed below (see: Types of outcome measures), compared outcomes for treatments performed during pregnancy, or those that described outcomes in high‐risk women (i.e. women with history of miscarriage or women conceiving through assisted reproductive technology (ART))

Types of participants

We included women who had a pregnancy with or without a previous conservative treatment for CIN/early cervical cancer (stage IA1). We included women irrespective of the grade of the lesion for both squamous and glandular intra‐epithelial neoplasia. There was also no age restriction.

Types of interventions

Any comparison of interventions for treatment of CIN or stage IA1 cervical cancer by conservative methods of either:

excision (cold knife conisation (CKC); laser conisation (LC); needle excision of the transformation zone (NETZ), also known as straight wire excision of the transformation zone (SWETZ); large loop excision of the transformation zone (LLETZ), also known as loop electrosurgical excisional procedure (LEEP); Fischer cone biopsy excisor (FCBE));
ablation (laser ablation (LA); radical diathermy (RD); cold coagulation (CC); cryotherapy (CT)).

In studies that reported on the impact of several treatment techniques, we extracted data for each specific method, where possible. If the outcomes were not reported separately for each technique, we analysed the intervention under broader terms, i.e. excisional treatment not otherwise specified (NOS), ablative treatment NOS and treatment NOS.

Types of outcome measures

Primary outcomes

Maternal outcomes

- overall (less than 37 weeks) prematurity (both iatrogenic and spontaneous)
- severe (less than 32 to 34 weeks) prematurity
- extreme (less than 28 to 30 weeks) prematurity
- overall prematurity in singleton and multiple pregnancies
- overall prematurity in nulliparous and parous women
- overall prematurity for single and multiple cones
- overall prematurity for different cone depths and volumes
- overall prematurity for different comparison groups

Secondary outcomes

Maternal outcomes

- overall spontaneous (i.e. non‐iatrogenic) prematurity
- severe spontaneous prematurity
- extreme spontaneous prematurity
- threatened preterm birth
- premature rupture of the membranes
- chorioamnionitis
- mode of delivery (caesarean section, instrumental deliveries)
- length of labour (precipitous, prolonged)
- induction of labour or use of oxytocin
- haemorrhage (antepartum, postpartum)
- analgesia (epidural, pethidine, not otherwise specified)
- cervical stenosis
- cervical cerclage

Neonatal outcomes

- low birth weight (less than 2500 g, less than 2000 g, less than 1500 g, less than 1000 g)
- admission to neonatal intensive unit (NICU)
- perinatal mortality
- stillbirth
- Apgar score

In cases of heterogeneity in the cut‐offs used for cone depth and prematurity classification, these were grouped together when possible (i.e. 32 to 34 weeks to include both cut‐offs, 10 to 12 mm cone depth to include studies grouping at both these cut‐offs including or not the values equal to these numbers).

Search methods for identification of studies

The literature searches started from 1948 when the conservative methods of treatment for CIN were introduced into clinical practice and included references published up to June 2017.

Electronic searches

We searched the following electronic databases.

Cochrane Central Register of Controlled Trials (CENTRAL) (CENTRAL, 2017, Issue 5) (Appendix 1).
MEDLINE (1948 to June week 4, 2017) (Appendix 2).
Embase (1980 to 2017, week 26) (Appendix 3).

The searches started from inception to date in order to capture all studies published since the late 1970s. The treatment techniques used predominantly to manage the disease have changed over the years, although there are still clinical indications for the oldest techniques.

We used the 'related articles' feature in MEDLINE to retrieve additional references.

Searching other resources

We searched Metaregister, Physicians Data Query, www.controlled‐trials.com/rct, www.clinicaltrials.gov and www.cancer.gov/clinicaltrials for ongoing studies. We contacted the main investigators of any relevant ongoing trials for further information.

We searched conference proceedings and abstracts through ZETOC (http://zetoc.mimas.ac.uk)) and WorldCat Dissertations. We searched reports of conferences in the following sources.

Annual Meeting of the British Society of Colposcopy and Cervical Pathology.
Annual Meeting of the International Federation of Cervical Pathology and Colposcopy.
Annual Meeting of European Federation of Colposcopy.
Annual Meeding of the American Society of Colposcopy and Cervical Pathology.

In an attempt to identify any articles missed by the initial search or any unpublished data, we handsearched the references of the retrieved articles and meta‐analyses and the proceedings of relevant conferences. We contacted experts in the field, including directors of UK cancer and colposcopy registries, to identify further reports of studies.

We included both published and unpublished studies, if they met the inclusion criteria for the review.

Data collection and analysis

Selection of studies

We downloaded all titles and abstracts retrieved by electronic searches to the reference management database Endnote. We also added titles and abstracts retrieved from other sources to Endnote. We removed duplicates and two review authors (MK, AA) independently examined the remaining references. Titles and abstracts retrieved from other sources were also added to the EndNote database. We excluded those studies which clearly did not meet the inclusion criteria and we obtained copies of the full text of potentially relevant references. We assessed the eligibility of retrieved papers independently, compared the results and resolved disagreements by discussion. If necessary, we reached consensus with the involvement of a third review author (MA). We documented reasons for exclusion.

Data extraction and management

We classified the studies according to treatment modality (i.e. CKC, LC, LLETZ, LA etc) and in groups of excisional or ablative techniques.

From each study, we extracted data on the study design and setting, the study population, the interventions examined, the comparison group, the quality of the data and risk of bias and the outcomes assessed. We retrieved from each study and outcome, the number of events in treated and untreated women. If required, we contacted authors to obtain additional data if the numbers provided in the published report did not allow sufficient precision in the data extraction.

More specifically, we extracted the following data.

Author, year of publication, journal and language.
Country.
Setting where the study was conducted (hospital‐based versus population‐based).
Inclusion and exclusion criteria.
Study design, methodology, source of information.
Study population:

- total number enrolled and number included in each group;
- grade of CIN;
- cone size;
- single/multiple pregnancy;
- single/multiple treatment;
- nulliparous/parous women;
- control for confounding factors:
  - age;
  - smoking;
  - parity;
  - socio‐economic status;
  - race;
  - history of previous preterm birth (PTB);
  - others.

Intervention details:
- type of procedure used (excisional or ablative);
- specific type of procedure used (excisional: CKC, LC, NETZ, LLETZ, FCBE; ablative: LA, RD, CC, CT).
Comparison group:
- external untreated comparison group (general population);
- internal comparison group:
  - self‐matching: the treated group consisted of only parous women and the pregnancy after treatment was compared to the pregnancy before treatment;
  - pre‐treatment pregnancies: in some studies, the treated group consisted of both nulliparous and parous women and the comparison group consisted of the pregnancies of the parous women before treatment.
- untreated women with colposcopy with or without a biopsy who did not undergo treatment;
- untreated women with untreated high‐grade disease.
Risk of bias (Assessment of risk of bias in included studies).

Outcomes reported in each study.

- Primary outcomes:
  - overall (less than 37 weeks) prematurity (both iatrogenic and spontaneous);
  - severe (less than 32 to 34 weeks) prematurity;
  - extreme (less than 28 to 30 weeks) prematurity;
  - overall prematurity in singleton and multiple pregnancies;
  - overall prematurity in nulliparous and parous women;
  - overall prematurity for single and multiple cones;
  - overall prematurity for different cone depths and volumes;
  - overall prematurity for different comparison groups.
- Secondary outcomes:
  - maternal outcomes: overall (less than 37 weeks) spontaneous (i.e. non‐iatrogenic) prematurity, severe (less than 32 to 34 weeks) spontaneous prematurity, extreme (less than 28 weeks) spontaneous prematurity, threatened preterm birth, premature rupture of the membranes, chorioamnionitis, mode of delivery (caesarean section, instrumental deliveries), length of labour (precipitous, prolonged), induction of labour or use of oxytocin, haemorrhage (antepartum, postpartum), analgesia (epidural, pethidine, not otherwise specified), cervical stenosis, cervical cerclage;
  - neonatal outcomes: low birth weight (LBW) (less than 2500 g, less than 2000 g, less than 1500 g, less than 1000 g), perinatal mortality, stillbirth, Apgar score.
- For each reported outcome, we extracted information on:
  - the outcome definition;
  - number of participants allocated to each group;
  - for the dichotomous outcomes of interest: number of adverse pregnancy events in each group (treated and untreated), in order to estimate the risk ratio (RR), and missing participants.

Two review authors (MK, AA) abstracted data independently in a data abstraction form specially designed for the review. They resolved differences by discussion or by appeal to a third review author (EP) if necessary.

Assessment of risk of bias in included studies

To assess the risk of bias in included RCTs, we planned to use Cochrane's'Risk of bias' tool, comprising assessments of the following study characteristics: sequence generation; allocation concealment; blinding (of participants, healthcare providers and outcome assessors); incomplete outcome data; selective reporting of outcomes; other possible sources of bias (Higgins 2011).

As RCTs comparing women with CIN with non‐treated are not feasible or ethical due to the pre‐malignant nature of the condition, we anticipated that published evidence might rely only on observational cohort studies. As the comparison groups (treated for CIN with a particular procedure versus non‐treated) are non‐randomised, effects and effect sizes cannot be attributed with certainty to the treatment alone. The differences in the size of the treatment effect across studies may be partly explained by the choice of control population, because women with CIN may have demographic and behavioural characteristics or even background immunological imbalances that place them at higher baseline risk of adverse reproductive outcomes.

It should also be noted that all eligible comparison groups have advantages and limitations. A recent meta‐analysis showed that the use of historical external controls might produce inherent biases that could inflate the contribution of cervical treatment to adverse outcomes, even if the authors control for possible confounders (such as age, parity, smoking etc; Bruinsma 2011). The use of internal controls (pregnancies in the index woman before treatment) is an attractive alternative approach, but even this might be inadequate for confounders that are liable to change with time. Women with mild precancerous lesions that do not warrant excision treatment probably provide the best, although still imperfect, comparator. In contrast, those with high‐grade disease who neglect treatment advice aimed at preventing cancer may have high risk for confounders related to low socioeconomic class that may influence fertility or pregnancy outcomes.

For non‐randomised studies (NRS), we assessed the risk of bias in the following domains: Incomplete outcome Data (Attrition bias) was considered to be present if for more than 20% of the cohort the outcome data were missing or the method to collect outcome data was not systematic; Selective reporting (Reporting bias) was considered to be present if outcome data were not reported stratified according to all included study types; Performance and selection bias was assessed in whether the treatment assignment was reported appropriately, based on the representativeness of both the treatment and comparison groups, i.e. whether the treated cohort indeed represented the population at risk and was not subjected to possible selection bias, whether the comparison groups were drawn from the same source as the treated group and on their comparability, i.e. whether the authors used internal or self‐matched comparison group, or otherwise matched or adjusted for the possible confounders or effect modifiers; no other possible sources of bias were assessed.

We used the GRADE (Grading of Recommendations, Assessment, Development and Evaluation) (GRADE Working Group 2004) approach to assess the quality of evidence provided by the included studies. We used GRADEpro (GRADE profiler) software to generate 'Summary of findings' tables to include an assessment of the more clinically relevant outcomes .

Measures of treatment effect

We calculated risk ratio (RR) and 95% confidence intervals (95% CIs) for each adverse pregnancy outcome in the treated versus untreated women for dichotomous outcomes. We separated studies by general type of treatment (excisional or ablative) and we further grouped them by specific treatment procedure and by specific comparison group. We used a random‐effects model to pool RR (Dersimonian 1986). We used unadjusted data for the analyses.

Unit of analysis issues

In studies with multiple treatment groups, we proportionally divided the ‘shared’ comparison group into the number of treatment groups (i.e. based on the number of treated women for each technique), in order to avoid duplicate inclusion of some untreated women in the same forest plot. We treated comparisons between each treatment group and the split comparison group as independent comparisons.

When more than one comparison group was described in the included studies, the comparison groups were summed together if appropriate (i.e. external, any CIN or HSIL without treatment). If an external and internal self‐matching group was available, only data on the external group were included. In one study (Castanon 2012), with both internal controls, pre‐conisation population was used in preference to self‐matching.

Dealing with missing data

We contacted study authors about inclusion/exclusion criteria and eligibility of their study. We further contacted authors for additional data not presented in the original manuscript i.e. data according to different cone depths.

Assessment of heterogeneity

We assessed inter‐study heterogeneity with Cochran Q test (Cochran 1954), by visual inspection of forest plots, by estimation of the percentage heterogeneity between studies which cannot be ascribed to sampling variation (I² statistic) (Higgins 2003), and by a formal statistical test of the significance of the heterogeneity (Deeks 2001). If there was evidence of substantial heterogeneity, we investigated and reported the possible reasons for this.

Assessment of reporting biases

For the outcomes presented in the 'Summary of findings' tables that included more than 10 studies, we explored potential publication bias graphically by the funnel plot in the Cochrane Review Manager software (Review Manager 2104).

Data synthesis

We pooled the results of the studies in meta‐analyses. For dichotomous outcomes, we calculated RR and 95% CIs and these were then pooled. We analysed the data separately for each treatment modality, in groups of ablative and excisional techniques, and as a whole, irrespective of the type of method used. We further analysed the data according to the cone depth.

Several studies provided separate data for overall and spontaneous PTB (sPTB) (less than37 weeks of gestation). If only data on sPTB were provided, these were also included in the overall PTB analysis (Crane 2006; Noehr 2009a; Ortoft 2010; Poon 2012; Stout 2015). If studies provided data on nulliparous and parous women separately, these were presented together and were also included in the respective forest plots for overall PTB (i.e. Analysis 1.15; Analysis 1.16). This was also the case for single and multiple treatment and singleton and multiple pregnancies .

1.15 — Comparison 1 Maternal Outcomes‐PTB, Outcome 15 PTB (<37w)‐Nulliparous women.

1.16 — Comparison 1 Maternal Outcomes‐PTB, Outcome 16 PTB (<37w)‐Parous women.

We used random‐effects models with inverse variance weighting for all meta‐analyses (Dersimonian 1986).

If data were not of suitable quality for meta‐analysis, we reported the results as a narrative in the text of the review.

Subgroup analysis and investigation of heterogeneity

Given the non‐randomised nature of the included studies, we assessed whether the choice of comparison group impacts on the risk estimate for each outcome and over‐inflates the effect of treatment that could be partly attributed to other confounders. We therefore distinguished the different untreated comparison groups used across studies and performed analyses for the risk of PTB for each individual comparator (external; internal (self‐matching); internal (pre‐treatment pregnancies); colposcopy but no treatment; HSIL but no treatment). We performed separate analyses according to the comparison group for PTB (less than 37 weeks of gestation).

Furthemore, for the outcome of PTB (less than 37 weeks of gestation), we also performed analyses according to parity (nulliparous and parous women separately), number of treatments (women with single and multiple treatments separately), and number of fetuses (women with singleton and multiple pregnancies separately).

Sensitivity analysis

Finally, we performed meta‐regression analysis to assess the impact of some factors on the risk of PTB (less than37 weeks). These included the year of study (1979 to 1989, 1990 to 1999, 2000 to 2009, 2010 to 2015); type of treatment (excision or ablation); type of comparator (external, internal – pregnancies before treatment, internal – self‐matching, CIN but no treatment, HSIL but no treatment).

Results

Description of studies

The characteristics of the included and excluded studies and the outcomes examined are described in the Characteristics of included studies and in the Characteristics of excluded studies, respectively.

Results of the search

We retrieved 3219 citations from the literature search. Of those, 2849 were excluded based on the title or abstract; 370 were retrieved in full text for evaluation. We identified 69 studies that fulfilled the inclusion criteria and 289 were excluded (of these 289 studies, 242 were reviews/meta‐analyses, conference proceedings, letters or duplicates; the remaining 47 studies are listed in detail in Excluded studies and Characteristics of excluded studies); 12 studies are awaiting classification. No unpublished studies could be identified. The details, including reasons for exclusion, are present in the PRISMA flowchart (Moher 2009; Figure 1).

Included studies

Sixty‐nine studies fulfilled the inclusion criteria of this systematic review and were also included in the meta‐analysis. We identified no unpublished studies. The detailed characteristics of the included studies are shown in Characteristics of included studies. All studies were cohorts and retrospective, except for five prospective studies (Fischer 2010; Frega 2013; Guo 2013; Poon 2012; Simoens 2012). Fifteen studies were population‐based (the data were drawn from registers, or they included a large number of hospitals covering a large area) (Albrechtsen 2008; Andia 2011; El‐Bastawissi 1999; Frega 2013; Heinonen 2013; Jakobsson 2007; Jones 1979; Kristensen 1985; Kristensen 1993; Larsson 1982; Noehr 2009a; Noehr 2009b; Reilly 2012; Shanbhag 2009; Sjoborg 2007) and the remaining studies were hospital‐based (the data were drawn from hospital records). There were no RCTs.

Seven studies were identified in non‐English language and were not included (He 2007; Kalitsaris 1991; Kasum 1991; Lund 1986; Praest 1979; Spuhler 1995; Zornoza‐Garcia 2009) given the large number of included studies and the low quality of these small studies we considered that their inclusion would not alter the conclusions of the review. In future updates we will consider the inclusion of these reports (Characteristics of studies awaiting classification). Five studies published after April 2016, week 2 are awaiting classification in future updates of this review (Aleman 2016; Bjorge 2016; Brie 2016; Jancar 2016; Zebitay 2017).

Many of the included studies included cohorts treated with a variety of treatment modalities. Specifically, 14 studies examined the impact of cold knife conisation; (CKC) (Bruinsma 2007; Buller 1982; Crane 2006; Ehsanipoor 2014; Guo 2013; Jones 1979; Klaritsch 2006; Kuoppala 1986; Larsson 1982; Ludviksson 1982; Moinian 1982; Ortoft 2010; Sozen 2014; Weber 1979), 10 of laser conisation (LC) (Andersen 1999; Bekassy 1996; Forsmo 1996; Hagen 1993; Lima 2011; Raio 1997; Sadler 2004; Sagot 1995; Simoens 2012; Spitzer 1995), one of needle excision of the transformation zone (NETZ) (Ortoft 2010), 32 of large loop excision of the transformation zone (LLETZ) (Acharya 2005; Andia 2011; Blomfield 1993; Braet 1994; Bruinsma 2007; Crane 2006; Cruickshank 1995; Ehsanipoor 2014; Frega 2013; Frey 2013; Gunasekera 1992; Guo 2013; Haffenden 1993; Heinonen 2013; Himes 2007; Jakobsson 2009; Kitson 2014; Lima 2011; Martyn 2015; Noehr 2009a; Noehr 2009b; Ortoft 2010; Paraskevaidis 2002; Parikh 2008; Poon 2012; Sadler 2004; Samson 2005; Simoens 2012; Stout 2015; Tan 2004; Turlington 1996; Werner 2010), one of Fischer cone biopsy Excisor (FCBE) (Anwar 2016), eight of laser ablation (LA) (Anderson 1984; Bruinsma 2007; Forsmo 1996; Gunasekera 1992; Sadler 2004; Saunders 1986; Spitzer 1995; van Rooijen 1999), one of radical diathermy (RD) (Bruinsma 2007), two of cryotherapy (CT) (Crane 2006; Hemmingsson 1982), 15 of excision not otherwise specified (NOS) (Albrechtsen 2008; Armarnik 2011; Castanon 2012; El‐Bastawissi 1999; Fischer 2010; Jakobsson 2007; Martyn 2015; Miller 2015; Reilly 2012; Shanbhag 2009; Simoens 2012; Sjoborg 2007; van de Vijner 2010; Van Hentenryck 2012; Wuntakal 2013), five of ablation NOS (Ehsanipoor 2014; El‐Bastawissi 1999; Jakobsson 2007; Reilly 2012; Shanbhag 2009) and three of treatment NOS (Kirn 2015; Kristensen 1985; Kristensen 1993).

There were five types of comparison groups: external (general population) (Acharya 2005; Albrechtsen 2008; Andersen 1999; Anderson 1984; Andia 2011; Armarnik 2011; Bekassy 1996; Blomfield 1993; Braet 1994; Castanon 2012; Crane 2006; Cruickshank 1995; Ehsanipoor 2014; El‐Bastawissi 1999; Fischer 2010; Forsmo 1996; Frega 2013; Frey 2013; Gunasekera 1992; Haffenden 1993; Hagen 1993; Heinonen 2013; Jakobsson 2007; Jakobsson 2009; Jones 1979; Kirn 2015; Klaritsch 2006; Kristensen 1985; Kristensen 1993; Kuoppala 1986; Lima 2011; Ludviksson 1982; Miller 2015; Noehr 2009a; Noehr 2009b; Ortoft 2010; Paraskevaidis 2002; Parikh 2008; Poon 2012; Raio 1997; Reilly 2012; Samson 2005; Saunders 1986; Shanbhag 2009; Simoens 2012; Sjoborg 2007; Sozen 2014; Tan 2004; van de Vijner 2010; Van Hentenryck 2012; van Rooijen 1999; Weber 1979; Werner 2010); internal (pre‐treatment pregnancies) (Acharya 2005; Albrechtsen 2008; Andia 2011; Buller 1982; Castanon 2012; Cruickshank 1995; Hemmingsson 1982; Larsson 1982; Moinian 1982; Sagot 1995; Spitzer 1995; Stout 2015; Werner 2010; Wuntakal 2013); internal (self‐matching) (Anwar 2016; Bekassy 1996; Castanon 2012; Jakobsson 2009; Kristensen 1993; Ortoft 2010; Raio 1997; Sjoborg 2007); women who attended colposcopy with or without biopsy who did not undergo treatment (Bruinsma 2007; Castanon 2012; Frey 2013; Guo 2013; Himes 2007; Kitson 2014; Martyn 2015; Miller 2015; Noehr 2009a; Poon 2012; Sadler 2004; Stout 2015; Wuntakal 2013), and women with untreated high‐grade squamous intra‐epithelial lesion (HSIL) (El‐Bastawissi 1999; Ortoft 2010; Shanbhag 2009).

As many studies were old, dating back to the 1980s, we ensured that we avoided overlapping the same patients in different reports, particularly those from the Scandinavian countries. More specifically, seven studies were identified from Denmark. Weber 1979 was an old hospital‐based study (delivery during 1974 to 1975), with no overlapping with the other more recent studies. Kristensen 1985 identified treated women from the registry of a Danish county (1973 to1980), but only women delivering in a specific university hospital (up to 1982) were included. Kristensen 1993 was a population‐based study from the whole of Denmark including only parous women, with their first infant delivered in 1982, and their second during 1982 to 1987 (treatment during 1977 to 1987). There was no overlapping with Kristensen 1985. Andersen 1999 was a hospital‐based study including women treated during 1985 to 1989. There was a negligible degree of overlapping with Kristensen 1993 (for women with delivery in 1982, treatment during 1985 to 1987 and subsequent delivery until 1987). Noehr 2009a was a population‐based study from the whole of Denmark that was comprised of women treated during 1997 to 2005 with subsequent singleton pregnancy at the same time period. There was no overlapping with the previous studies. Noehr 2009b had the same design with Noehr 2009a and the only difference was that only women with twin pregnancies were included. Ortoft 2010 identified treated women from the Danish nationwide pathology database (1989 to 2007), but only women delivering at Aarhus University Hospital until 2007 were included. Approximately 8% of all Danish births take place in this hospital. Women delivering during 1997 to 2005 (nine years) were also included in Noehr 2009a, but women delivering during 1989 to 1996 and 2006 to 2007 (10 years) were not included in Noehr 2009a. Because there was no way to eliminate the overlapping, we decided to include both studies.

Four studies were identified from Finland. Some of the authors in Jakobsson 2007 (population‐based study from the whole of Finland), Jakobsson 2009 (hospital‐based study) and Heinonen 2013 (population‐based study from the whole of Finland) were common, and we carefully avoided duplication. More details about the outcomes extracted from each study are listed in Characteristics of included studies. Kuoppala 1986 was another hospital‐based study from Finland, but there was no overlapping with the aforementioned population‐based studies.

Five studies were identified from Norway. Albrechtsen 2008 was a population‐based study from the whole of Norway (excisional treatment during 1953 to 1979 or 1986 to 2003 and subsequent pregnancy during 1967 to 2003). All participants in Acharya 2005 and Sjoborg 2007 were also included in Albrechtsen 2008, thus we excluded Acharya 2005 and Sjoborg 2007 from the analyses, which also included Albrechtsen 2008. There was no overlapping between Acharya 2005 and Sjoborg 2007. Forsmo 1996 was a hospital‐based study which included women treated with LLETZ or LA during 1983 to 1988. There was overlapping with Albrechtsen 2008 for women treated with LLETZ during 1986 to 1988, but there was no way to eliminate this overlapping. In Hagen 1993, all women had received LLETZ during 1983to 1985 and there was no overlapping with Albrechtsen 2008.

Six studies were identified from Sweden. Five of these studies were hospital‐based (Bekassy 1996; Hemmingsson 1982; Ludviksson 1982; Moinian 1982; van Rooijen 1999). Larsson 1982 identified women from the South Swedish Regional Tumour Registry, but only women delivering in two hospitals were included. One of these hospitals was also included in Bekassy 1996, but the studies took place in different periods with no overlapping.

Although case‐control studies and studies assessing the impact of treatment performed during pregnancy were excluded, we included the study by Ortoft 2010, as only 18 women (2.5%) were treated during pregnancy, and the case‐control study by Castanon 2012, as additional data were provided by the authors.

There was no risk of overlapping for studies from other countries.

Excluded studies

The characteristics of the 47 excluded studies (not including reviews/meta‐analyses, conference proceedings, letters or duplicates) are shown in Characteristics of excluded studies. The PRISMA flowchart is shown in Figure 1 . We excluded studies without an untreated comparison group (Althuisius 2001; Berghella 2004; Berretta 2013; Bull‐Phelps 2007; Chevreau 2017; Conner 2013; Ferenczy 1995; Gordon 1991; Gronroos 1979; Khalid 2012; Kim 2016; Kindinger 2016; Kullander 1971; Leiman 1980; Liu 2014; Liverani 2016; Macvicar 1968; Mariya 2016; Masamoto 2008; Michelin 2009; Monaghan 1982; Nam 2010; Novikova 1994; Patrelli 2008; Radha Bai Prabhu 2010; Rafaeli‐Yehudai 2014; Sangkarat 2014; Shin 2010; Wakita 1990; Wongtiraporn 2014), studies with women treated during pregnancy (Mitsuhashi 2000; Rosen 1991; Seki 2010; Sljivancanin 2013), or late obstetric outcomes (beyond 24 weeks) that we did not study in this meta‐analysis (Ciavattini 2015; Gentry 2000; Kalliala 2012; Naleway 2015; Ricciotti 1995; Spracklen 2013), case‐control studies (Watson 2012), studies with a high‐risk treated and/or comparison group (i.e. previous history of mid trimester loss (Pils 2014), conceived through assisted reproductive technology (ART) (Ciavattini 2014; Pinborg 2015)), and studies assessing the impact of CIN on outcomes without information as to whether treatment was performed (Al‐Halal 2013; Smaldone 2010; Zuo 2011).

Risk of bias in included studies

The included studies were non‐randomised studies (NRS); they were prospective or retrospective cohorts and were therefore at high risk of underlying bias. The included studies varied with regard to design, the data source, the study and comparison populations, the reported outcomes, the length of follow‐up and the matching for possible confounders, as described above.

The summary of the authors' judgements about each 'Risk of bias' item is presented in Figure 2 and the detailed evaluation of 'Risk of bias' domains separately for each included study in Figure 3.

'Risk of bias' graph: review authors' judgements about each risk of bias item presented as percentages across all included studies.

'Risk of bias' summary: review authors' judgements about each risk of bias item for each included study.

A description of the quality of the evidence is provided based on the GRADE assessment on maternal outcomes (Table 1) and on fetal outcomes (Table 2).

Summary of findings for the main comparison. The effect of treatment for CIN on maternal outcomes.

The effect of treatment for CIN on maternal outcomes
Patient or population: women with known obstetric outcomes  Setting: hospitals/clinics  Intervention: treatment for CIN before pregnancy  Comparison: women with no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with [comparison]	Risk with [intervention]
PTB (< 37 w)	Study population		RR 1.75  (1.57 to 1.96)	5,242,917  (59 observational studies)	⊕⊝⊝⊝  VERY LOW ¹
	54 per 1000	95 per 1000  (85 to 106)
PTB (< 32 to 34 w)	Study population		RR 2.25  (1.79 to 2.82)	3,793,874  (24 observational studies)	⊕⊝⊝⊝  VERY LOW ²
	14 per 1000	32 per 1000  (26 to 40)
PTB (< 28 to 30 w)	Study population		RR 2.23  (1.55 to 3.22)	3,910,629  (8 observational studies)	⊕⊝⊝⊝  VERY LOW ³
	3 per 1000	7 per 1000  (5 to 11)
PTB (< 37 w) ‐ Repeat cones versus No Treatment	Study population		RR 3.78  (2.65 to 5.39)	1,317,284  (11 observational studies)	⊕⊝⊝⊝  VERY LOW ⁴
	41 per 1000	156 per 1000  (109 to 222)
pPROM (<3 7 w)	Study population		RR 2.36  (1.76 to 3.17)	477,011  (21 observational studies)	⊕⊝⊝⊝  VERY LOW ⁵
	34 per 1000	80 per 1000  (60 to 108)
PTB (< 37 w) ‐ Depth ≤ 10 mm to 12 mm versus No Treatment	Study population		RR 1.54  (1.09 to 2.18)	550,929  (8 observational studies)	⊕⊝⊝⊝  VERY LOW ⁶
	34 per 1000	53 per 1000  (37 to 75)
PTB (< 37 w) ‐ PTB (< 37 w) ‐ Depth ≥10 mm to 12 mm versus No Treatment	Study population		RR 1.93  (1.62 to 2.31)	552,711  (8 observational studies)	⊕⊕⊕⊝  LOW ⁷
	34 per 1000	66 per 1000  (55 to 79)
PTB (< 37w) ‐ PTB (<37w) ‐ Depth ≥15 to 17mm versus No Treatment	Study population		RR 2.77  (1.95 to 3.93)	544,986  (4 observational studies)	⊕⊕⊕⊕  VERY LOW ⁸
	34 per 1000	94 per 1000  (66 to 134)
PTB (< 37 w) ‐ PTB (< 37 w) ‐ Depth ≥ 20 mm versus No Treatment	Study population		RR 4.91  (2.06 to 11.68)	543,750  (3 observational studies)	⊕⊕⊕⊕  VERY LOW ⁹
	34 per 1000	167 per 1000  (70 to 397)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

The effect of treatment for CIN on neonatal outcomes
Patient or population: women with known obstetric outcomes  Setting: hospitals/clinics  Intervention: treatment for CIN before pregnancy  Comparison: women with no treatment
Outcomes	*Anticipated absolute effects^ (95% CI)**		Relative effect  (95% CI)	№ of participants  (studies)	Quality of the evidence  (GRADE)	Comments
	Risk with No Treatment	Risk with Treatment
LBW (< 2500 g) ‐ Treatment versus No Treatment	Study population		RR 1.81  (1.58 to 2.07)	1,348,206  (30 observational studies)	⊕⊝⊝⊝  VERY LOW ¹
	37 per 1000	66 per 1000  (58 to 76)
NICU Admission ‐ Treatment versus No Treatment	Study population		RR 1.45  (1.16 to 1.81)	2557  (8 observational studies)	⊕⊕⊝⊝  LOW ²
	89 per 1000	130 per 1000  (104 to 162)
Perinatal Mortality ‐ Treatment versus No Treatment	Study population		RR 1.51  (1.13 to 2.03)	1,659,433  (23 observational studies)	⊕⊕⊝⊝  LOW ³
	7 per 1000	11 per 1000  (8 to 14)
*The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).    CI: Confidence interval; RR: Risk ratio
GRADE Working Group grades of evidence   High quality: We are very confident that the true effect lies close to that of the estimate of the effect  Moderate quality: We are moderately confident in the effect estimate: The true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different  Low quality: Our confidence in the effect estimate is limited: The true effect may be substantially different from the estimate of the effect  Very low quality: We have very little confidence in the effect estimate: The true effect is likely to be substantially different from the estimate of effect

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 sPTB (<37w)	14	1.024731E6	Risk Ratio (IV, Random, 95% CI)	1.76 [1.47, 2.11]
1.1 CKC vs No Treatment	3	7320	Risk Ratio (IV, Random, 95% CI)	3.53 [2.05, 6.05]
1.2 LC vs No Treatment	2	222	Risk Ratio (IV, Random, 95% CI)	1.40 [0.51, 3.81]
1.3 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	5.83 [3.80, 8.95]
1.4 LLETZ vs No Treatment	11	773123	Risk Ratio (IV, Random, 95% CI)	1.60 [1.22, 2.08]
1.5 LA vs No Treatment	1	356	Risk Ratio (IV, Random, 95% CI)	0.95 [0.34, 2.68]
1.6 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	1.86 [0.08, 43.87]
1.7 Excisional Treatment NOS vs No Treatment	2	95985	Risk Ratio (IV, Random, 95% CI)	1.70 [1.17, 2.46]
1.8 Ablative Treatment NOS vs No Treatment	2	134720	Risk Ratio (IV, Random, 95% CI)	1.42 [1.20, 1.70]
1.9 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	1.30 [1.00, 1.69]
2 sPTB (<32‐34w)	7	655675	Risk Ratio (IV, Random, 95% CI)	2.63 [1.91, 3.62]
2.1 CKC vs No Treatment	2	6990	Risk Ratio (IV, Random, 95% CI)	4.38 [1.08, 17.65]
2.2 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	10.53 [4.33, 25.65]
2.3 LLETZ vs No Treatment	6	530985	Risk Ratio (IV, Random, 95% CI)	2.37 [1.82, 3.08]
2.4 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	1.86 [0.08, 43.87]
2.5 Excisional Treatment NOS vs No Treatment	1	264	Risk Ratio (IV, Random, 95% CI)	13.92 [0.73, 266.57]
2.6 Ablative Treatment NOS vs No Treatment	1	109979	Risk Ratio (IV, Random, 95% CI)	1.57 [0.97, 2.53]
3 sPTB (<28w)	2	626670	Risk Ratio (IV, Random, 95% CI)	3.18 [1.64, 6.16]
3.1 CKC vs No Treatment	1	6956	Risk Ratio (IV, Random, 95% CI)	5.41 [0.74, 39.84]
3.2 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	14.74 [4.50, 48.32]
3.3 LLETZ vs No Treatment	2	502336	Risk Ratio (IV, Random, 95% CI)	2.57 [1.96, 3.36]
3.4 Ablative Treatment NOS vs No Treatment	1	109979	Risk Ratio (IV, Random, 95% CI)	1.22 [0.54, 2.74]
4 pPROM (<37w)	21	477011	Risk Ratio (IV, Random, 95% CI)	2.36 [1.76, 3.17]
4.1 CKC vs No Treatment	4	36733	Risk Ratio (IV, Random, 95% CI)	4.11 [2.05, 8.25]
4.2 LC vs No Treatment	4	635	Risk Ratio (IV, Random, 95% CI)	1.89 [0.97, 3.66]
4.3 NETZ vs No Treatment	1	7279	Risk Ratio (IV, Random, 95% CI)	8.83 [5.39, 14.46]
4.4 LLETZ vs No Treatment	8	302974	Risk Ratio (IV, Random, 95% CI)	2.15 [1.48, 3.12]
4.5 LA vs No Treatment	2	548	Risk Ratio (IV, Random, 95% CI)	1.62 [0.74, 3.55]
4.6 CT vs No Treatment	1	180	Risk Ratio (IV, Random, 95% CI)	1.13 [0.21, 6.00]
4.7 Excisional Treatment NOS vs No Treatment	5	98372	Risk Ratio (IV, Random, 95% CI)	2.66 [1.13, 6.24]
4.8 Ablative Treatment NOS vs No Treatment	1	24742	Risk Ratio (IV, Random, 95% CI)	1.47 [1.01, 2.15]
4.9 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	1.44 [1.05, 1.97]
5 pPROM (<32w)	1	72788	Risk Ratio (IV, Random, 95% CI)	8.30 [2.03, 33.98]
5.1 CKC vs No Treatment	1	6842	Risk Ratio (IV, Random, 95% CI)	5.32 [0.72, 39.19]
5.2 NETZ vs No Treatment	1	7279	Risk Ratio (IV, Random, 95% CI)	25.38 [9.80, 65.74]
5.3 LLETZ vs No Treatment	1	58667	Risk Ratio (IV, Random, 95% CI)	3.74 [1.66, 8.41]
6 pPROM (<28w)	1	72788	Risk Ratio (IV, Random, 95% CI)	9.09 [1.04, 79.18]
6.1 CKC vs No Treatment	1	6842	Risk Ratio (IV, Random, 95% CI)	6.64 [0.38, 115.16]
6.2 NETZ vs No Treatment	1	7279	Risk Ratio (IV, Random, 95% CI)	43.51 [11.48, 164.86]
6.3 LLETZ vs No Treatment	1	58667	Risk Ratio (IV, Random, 95% CI)	1.81 [0.25, 13.08]
7 Threatened PTB	5	903	Risk Ratio (IV, Random, 95% CI)	2.44 [1.37, 4.33]
7.1 CKC vs No Treatment	1	126	Risk Ratio (IV, Random, 95% CI)	1.40 [0.45, 4.34]
7.2 LC vs No Treatment	1	112	Risk Ratio (IV, Random, 95% CI)	1.56 [0.53, 4.62]
7.3 LLETZ vs No Treatment	1	237	Risk Ratio (IV, Random, 95% CI)	4.0 [0.75, 21.37]
7.4 Excisional Treatment NOS vs No Treatment	2	428	Risk Ratio (IV, Random, 95% CI)	4.51 [1.68, 12.06]
8 Chorioamnionitis	4	29198	Risk Ratio (IV, Random, 95% CI)	3.43 [1.36, 8.64]
8.1 CKC vs No Treatment	1	28531	Risk Ratio (IV, Random, 95% CI)	2.39 [0.61, 9.43]
8.2 LC vs No Treatment	1	112	Risk Ratio (IV, Random, 95% CI)	3.33 [0.14, 80.11]
8.3 LLETZ vs No Treatment	1	237	Risk Ratio (IV, Random, 95% CI)	10.00 [1.19, 84.15]
8.4 Excisional Treatment NOS vs No Treatment	1	318	Risk Ratio (IV, Random, 95% CI)	3.0 [0.51, 17.68]
9 Caeserean Section	36	272360	Risk Ratio (IV, Random, 95% CI)	1.06 [0.98, 1.14]
9.1 CKC vs No Treatment	6	30462	Risk Ratio (IV, Random, 95% CI)	1.24 [0.91, 1.68]
9.2 LC vs No Treatment	5	1038	Risk Ratio (IV, Random, 95% CI)	1.38 [0.90, 2.11]
9.3 LLETZ vs No Treatment	14	5436	Risk Ratio (IV, Random, 95% CI)	1.04 [0.94, 1.15]
9.4 LA vs No Treatment	4	1258	Risk Ratio (IV, Random, 95% CI)	0.86 [0.61, 1.20]
9.5 CT vs No Treatment	2	238	Risk Ratio (IV, Random, 95% CI)	2.47 [1.02, 6.01]
9.6 Excisional Treatment NOS vs No Treatment	9	203532	Risk Ratio (IV, Random, 95% CI)	1.03 [0.89, 1.20]
9.7 Ablative Treatment NOS vs No Treatment	2	24848	Risk Ratio (IV, Random, 95% CI)	1.38 [0.42, 4.58]
9.8 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	1.13 [1.00, 1.27]
10 Instrumental Deliveries (ventouse/forceps)	16	9588	Risk Ratio (IV, Random, 95% CI)	0.97 [0.88, 1.08]
10.1 CKC vs No Treatment	2	454	Risk Ratio (IV, Random, 95% CI)	1.33 [0.66, 2.70]
10.2 LC vs No Treatment	2	668	Risk Ratio (IV, Random, 95% CI)	1.16 [0.65, 2.07]
10.3 LLETZ vs No Treatment	6	1418	Risk Ratio (IV, Random, 95% CI)	0.89 [0.68, 1.17]
10.4 LA vs No Treatment	3	550	Risk Ratio (IV, Random, 95% CI)	0.94 [0.62, 1.41]
10.5 Excisional Treatment NOS vs No Treatment	3	950	Risk Ratio (IV, Random, 95% CI)	0.71 [0.46, 1.10]
10.6 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	1.01 [0.89, 1.15]
11 Precipitous Labour (<2hours)	5	1059	Risk Ratio (IV, Random, 95% CI)	1.26 [0.80, 1.96]
11.1 CKC vs No Treatment	2	289	Risk Ratio (IV, Random, 95% CI)	1.24 [0.47, 3.27]
11.2 LLETZ vs No Treatment	4	770	Risk Ratio (IV, Random, 95% CI)	1.26 [0.76, 2.08]
12 Prolonged labour (>12hours)	7	1854	Risk Ratio (IV, Random, 95% CI)	1.25 [0.92, 1.69]
12.1 CKC vs No Treatment	2	325	Risk Ratio (IV, Random, 95% CI)	1.99 [0.89, 4.45]
12.2 LC vs No Treatment	1	500	Risk Ratio (IV, Random, 95% CI)	0.92 [0.41, 2.04]
12.3 LLETZ vs No Treatment	4	673	Risk Ratio (IV, Random, 95% CI)	0.96 [0.55, 1.70]
12.4 LA vs No Treatment	2	356	Risk Ratio (IV, Random, 95% CI)	1.41 [0.88, 2.26]
13 Induction of Labour	11	4668	Risk Ratio (IV, Random, 95% CI)	1.01 [0.89, 1.15]
13.1 CKC vs No Treatment	2	137	Risk Ratio (IV, Random, 95% CI)	1.11 [0.54, 2.29]
13.2 LLETZ vs No treatment	8	4056	Risk Ratio (IV, Random, 95% CI)	0.99 [0.82, 1.20]
13.3 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	0.61 [0.22, 1.66]
13.4 Excisional Treatment NOS vs No Treatment	2	417	Risk Ratio (IV, Random, 95% CI)	0.90 [0.64, 1.28]
14 Oxytocin Use	6	2006	Risk Ratio (IV, Random, 95% CI)	0.90 [0.64, 1.26]
14.1 CKC vs No Treatment	1	103	Risk Ratio (IV, Random, 95% CI)	0.98 [0.59, 1.63]
14.2 LLETZ vs No Treatment	4	1804	Risk Ratio (IV, Random, 95% CI)	0.76 [0.43, 1.34]
14.3 Excisional Treatment NOS vs No Treatment	1	99	Risk Ratio (IV, Random, 95% CI)	1.18 [0.67, 2.05]
15 Epidural Use	5	105488	Risk Ratio (IV, Random, 95% CI)	1.02 [0.68, 1.53]
15.1 LLETZ vs No Treatment	4	818	Risk Ratio (IV, Random, 95% CI)	0.86 [0.64, 1.16]
15.2 Excisional Treatment NOS vs No Treatment	1	104670	Risk Ratio (IV, Random, 95% CI)	1.79 [1.29, 2.50]
16 Pethidine Use	2	394	Risk Ratio (IV, Random, 95% CI)	0.94 [0.72, 1.24]
16.1 LLETZ vs No treatment	2	394	Risk Ratio (IV, Random, 95% CI)	0.94 [0.72, 1.24]
17 Analgesia Use NOS	1	103	Risk Ratio (IV, Random, 95% CI)	1.11 [0.62, 1.98]
17.1 CKC vs No Treatment	1	103	Risk Ratio (IV, Random, 95% CI)	1.11 [0.62, 1.98]
18 Cervical stenosis	2	680	Risk Ratio (IV, Random, 95% CI)	2.26 [0.24, 21.59]
18.1 LC vs No Treatment	1	500	Risk Ratio (IV, Random, 95% CI)	3.0 [0.12, 73.29]
18.2 CT vs No Treatment	1	180	Risk Ratio (IV, Random, 95% CI)	1.71 [0.07, 41.31]
19 Antepartum Haemorrhage	4	1245	Risk Ratio (IV, Random, 95% CI)	1.11 [0.40, 3.12]
19.1 CKC vs No Treatment	1	34	Risk Ratio (IV, Random, 95% CI)	1.24 [0.26, 5.83]
19.2 LC vs No Treatment	1	168	Risk Ratio (IV, Random, 95% CI)	17.84 [0.98, 325.68]
19.3 LLETZ vs No Treatment	2	277	Risk Ratio (IV, Random, 95% CI)	0.52 [0.16, 1.67]
19.4 LA vs No Treatment	1	708	Risk Ratio (IV, Random, 95% CI)	8.00 [0.90, 71.18]
19.5 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	0.41 [0.07, 2.25]
20 Postpartum Haemorrhage (>600ml)	1	149	Risk Ratio (IV, Random, 95% CI)	4.60 [1.38, 15.36]
20.1 CKC vs No Treatment	1	149	Risk Ratio (IV, Random, 95% CI)	4.60 [1.38, 15.36]
21 Massive Obstetric Haemorrhage (>1000ml)	1	149	Risk Ratio (IV, Random, 95% CI)	3.95 [0.45, 34.48]
21.1 CKC vs No Treatment	1	149	Risk Ratio (IV, Random, 95% CI)	3.95 [0.45, 34.48]
22 Cervical cerclage	8	141300	Risk Ratio (IV, Random, 95% CI)	14.29 [2.85, 71.65]
22.1 CKC vs No Treatment	3	30744	Risk Ratio (IV, Random, 95% CI)	31.42 [2.32, 426.22]
22.2 LC vs No Treatment	1	112	Risk Ratio (IV, Random, 95% CI)	6.68 [0.83, 53.69]
22.3 LLETZ vs No Treatment	1	56	Risk Ratio (IV, Random, 95% CI)	11.00 [0.64, 189.96]
22.4 Excisional Treatment NOS vs No Treatment	2	104840	Risk Ratio (IV, Random, 95% CI)	42.45 [28.99, 62.16]
22.5 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	2.16 [1.24, 3.76]

Outcome or subgroup title	No. of studies	No. of participants	Statistical method	Effect size
1 LBW (<2500g)	30	1.348206E6	Risk Ratio (IV, Random, 95% CI)	1.81 [1.58, 2.07]
1.1 CKC vs No Treatment	5	30304	Risk Ratio (IV, Random, 95% CI)	2.51 [1.78, 3.53]
1.2 LC vs No Treatment	4	786	Risk Ratio (IV, Random, 95% CI)	1.76 [0.72, 4.35]
1.3 LLETZ vs No Treatment	12	3357	Risk Ratio (IV, Random, 95% CI)	2.11 [1.51, 2.94]
1.4 LA vs No Treatment	4	1104	Risk Ratio (IV, Random, 95% CI)	1.07 [0.59, 1.92]
1.5 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	3.67 [0.47, 28.47]
1.6 Excisional Treatment NOS vs No Treatment	10	823648	Risk Ratio (IV, Random, 95% CI)	2.01 [1.62, 2.49]
1.7 Ablative Treatment NOS vs No Treatment	4	483402	Risk Ratio (IV, Random, 95% CI)	1.36 [1.19, 1.55]
1.8 Treatment NOS vs No Treatment	1	5547	Risk Ratio (IV, Random, 95% CI)	1.35 [1.14, 1.60]
2 LBW (<2000g)	3	74981	Risk Ratio (IV, Random, 95% CI)	2.49 [0.97, 6.36]
2.1 LC vs No Treatment	1	181	Risk Ratio (IV, Random, 95% CI)	4.46 [1.36, 14.59]
2.2 LA vs No Treatment	2	772	Risk Ratio (IV, Random, 95% CI)	0.95 [0.39, 2.29]
2.3 Excisional Treatment NOS vs No Treatment	1	74028	Risk Ratio (IV, Random, 95% CI)	4.60 [3.32, 6.37]
3 LBW (<1500g)	5	76836	Risk Ratio (IV, Random, 95% CI)	3.00 [1.54, 5.85]
3.1 LC vs No Treatment	1	181	Risk Ratio (IV, Random, 95% CI)	12.75 [1.53, 106.44]
3.2 LLETZ vs No Treatment	1	378	Risk Ratio (IV, Random, 95% CI)	7.0 [0.36, 134.59]
3.3 LA vs No Treatment	2	772	Risk Ratio (IV, Random, 95% CI)	0.68 [0.16, 2.80]
3.4 Excisional Treatment NOS vs No Treatment	2	75505	Risk Ratio (IV, Random, 95% CI)	3.34 [2.02, 5.54]
4 LBW (<1000g)	2	2185	Risk Ratio (IV, Random, 95% CI)	2.09 [0.06, 74.71]
4.1 LA vs No Treatment	1	708	Risk Ratio (IV, Random, 95% CI)	0.29 [0.01, 5.50]
4.2 Excisional Treatment NOS vs No Treatment	1	1477	Risk Ratio (IV, Random, 95% CI)	11.10 [1.44, 85.79]
5 NICU Admission	8	2557	Risk Ratio (IV, Random, 95% CI)	1.45 [1.16, 1.81]
5.1 CKC vs No Treatment	2	71	Risk Ratio (IV, Random, 95% CI)	1.40 [0.52, 3.75]
5.2 LLETZ vs No Treatment	5	1994	Risk Ratio (IV, Random, 95% CI)	1.42 [1.01, 1.99]
5.3 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	2.44 [0.29, 20.49]
5.4 Excisional Treatment NOS vs No Treatment	2	434	Risk Ratio (IV, Random, 95% CI)	1.76 [1.13, 2.75]
6 Perinatal Mortality	23	1.659433E6	Risk Ratio (IV, Random, 95% CI)	1.51 [1.13, 2.03]
6.1 CKC vs No Treatment	7	50588	Risk Ratio (IV, Random, 95% CI)	1.46 [0.83, 2.57]
6.2 LC vs No Treatment	3	906	Risk Ratio (IV, Random, 95% CI)	1.89 [0.26, 13.87]
6.3 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	9.99 [3.13, 31.92]
6.4 LLETZ vs No Treatment	7	302271	Risk Ratio (IV, Random, 95% CI)	1.53 [0.88, 2.67]
6.5 LA vs No Treatment	2	258	Risk Ratio (IV, Random, 95% CI)	3.0 [0.12, 72.74]
6.6 CT vs No Treatment	2	238	Risk Ratio (IV, Random, 95% CI)	0.19 [0.01, 4.59]
6.7 Excisional Treatment NOS vs No Treatment	5	820028	Risk Ratio (IV, Random, 95% CI)	1.85 [1.02, 3.36]
6.8 Ablative Treatment NOS vs No Treatment	2	472197	Risk Ratio (IV, Random, 95% CI)	0.69 [0.42, 1.13]
6.9 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	1.00 [0.63, 1.58]
7 Perinatal Mortality (<37w)	1	73992	Risk Ratio (IV, Random, 95% CI)	9.40 [2.01, 43.89]
7.1 CKC vs No Treatment	1	6956	Risk Ratio (IV, Random, 95% CI)	5.33 [0.31, 90.71]
7.2 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	30.96 [8.71, 110.13]
7.3 LLETZ vs No Treatment	1	59637	Risk Ratio (IV, Random, 95% CI)	3.92 [1.24, 12.38]
8 Perinatal Mortality (<32w)	1	73992	Risk Ratio (IV, Random, 95% CI)	12.81 [2.70, 60.87]
8.1 CKC vs No Treatment	1	6956	Risk Ratio (IV, Random, 95% CI)	6.75 [0.39, 117.10]
8.2 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	44.23 [11.67, 167.61]
8.3 LLETZ vs No Treatment	1	59637	Risk Ratio (IV, Random, 95% CI)	5.43 [1.71, 17.30]
9 Perinatal Mortality (<28w)	1	73992	Risk Ratio (IV, Random, 95% CI)	13.76 [2.37, 79.89]
9.1 CKC vs No Treatment	1	6956	Risk Ratio (IV, Random, 95% CI)	9.21 [0.51, 164.95]
9.2 NETZ vs No Treatment	1	7399	Risk Ratio (IV, Random, 95% CI)	51.61 [13.17, 202.29]
9.3 LLETZ vs No Treatment	1	59637	Risk Ratio (IV, Random, 95% CI)	4.49 [1.09, 18.45]
10 Stillbirth	12	249855	Risk Ratio (IV, Random, 95% CI)	0.98 [0.63, 1.52]
10.1 CKC vs No Treatment	3	935	Risk Ratio (IV, Random, 95% CI)	1.61 [0.48, 5.40]
10.2 LC vs No Treatment	2	725	Risk Ratio (IV, Random, 95% CI)	0.33 [0.03, 3.18]
10.3 LLETZ vs No Treatment	4	242473	Risk Ratio (IV, Random, 95% CI)	1.42 [0.62, 3.26]
10.4 LA vs No Treatment	1	64	Risk Ratio (IV, Random, 95% CI)	0.0 [0.0, 0.0]
10.5 Excisional Treatment NOS vs No Treatment	1	110	Risk Ratio (IV, Random, 95% CI)	0.0 [0.0, 0.0]
10.6 Treatment NOS vs No Treatment	1	5548	Risk Ratio (IV, Random, 95% CI)	0.77 [0.42, 1.40]
11 Apgar score (≤5)(1min)	1	225	Risk Ratio (IV, Random, 95% CI)	0.57 [0.12, 2.68]
11.1 LC vs No Treatment	1	225	Risk Ratio (IV, Random, 95% CI)	0.57 [0.12, 2.68]
12 Apgar score (<7)(1min)	1	152	Risk Ratio (IV, Random, 95% CI)	0.63 [0.07, 5.71]
12.1 LLETZ vs No Treatment	1	87	Risk Ratio (IV, Random, 95% CI)	0.16 [0.01, 3.30]
12.2 CKC vs No Treatment	1	65	Risk Ratio (IV, Random, 95% CI)	1.61 [0.15, 16.90]
13 Apgar score (<7)(5min)	2	297	Risk Ratio (IV, Random, 95% CI)	0.82 [0.19, 3.59]
13.1 CKC vs No Treatment	1	32	Risk Ratio (IV, Random, 95% CI)	0.0 [0.0, 0.0]
13.2 LLETZ vs No Treatment	1	120	Risk Ratio (IV, Random, 95% CI)	0.93 [0.16, 5.37]
13.3 CT vs No Treatment	1	58	Risk Ratio (IV, Random, 95% CI)	0.61 [0.04, 9.28]
13.4 Excisional Treatment NOS vs No Treatment	1	87	Risk Ratio (IV, Random, 95% CI)	0.0 [0.0, 0.0]

Methods	Retrospective cohort study Comparison group: A) External ‐ matching for age (+/‐ 3 years), parity, date of delivery, smoking (+/‐ 5 cigarettes per day) and previous obstetric history B) Internal (pre‐treatment pregnancies) Information source ‐ Hospital records of the University Hospital of Northern Norway
Participants	A) Treated group: 79 women < 45 years who had a LLETZ (December 1995 to December 2000) and subsequently delivered (> 20 weeks) at the University Hospital of Northern Norway. Inclusion criteria: only first pregnancies (> 20 weeks) following LLETZ Exclusion criteria: Women with ectopic pregnancies, miscarriages, TOPs. Untreated group: 158 matched women who were identified using routinely entered data from the birth register. B) Of the 79 women of the treated group, 45 were parous. The last pregnancy before LLETZ of these 45 women can serve as an internal comparison group.
Interventions	LLETZ
Outcomes	PTB (< 37 weeks); threatened PTL; chorioamnionitis; induction of labour; LBW (< 2500 g); perinatal mortality
Notes	Because all women included in this study have been also included in Albrechtsen 2008, we excluded it from the analyses in which Albrechtsen 2008 has been also included. A total of 428 women < 45 years had LLETZ performed during the study period and 89 of them had a pregnancy after the procedure. Ten women were excluded (three ectopic pregnancies, two TOPs and five miscarriages) from the study. Data from 79 women whose pregnancies progressed > 20 weeks and 158 matched controls were analysed. The histological diagnosis was normal in 3 (3.8%), CIN1 in 5 (6.3%), CIN2 in 18 (22.8%), and CIN3 in 53 (67.1%) of cases.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias)   All outcomes	Low risk	The information was obtained from hospital records
Selective reporting (reporting bias)	Low risk	No reporting bias is obvious
Other bias	Low risk	No other obvious source of bias
Relevant assignment described?	Low risk	Yes, treatment performed on clinical grounds
Representative intervention group?	Low risk	All eligible for the study women having LLETZ and subsequently delivering in a single university hospital between December 1995 to December 2000
Representative comparison group?	Low risk	The untreated group was drawn from the same source as the treated group
Comparability of treatment groups?	Low risk	Matching for date of delivery, age, parity, previous obstetric history and smoking habit

Methods	Castanon 2012 (this is the main study): Retrospective cohort study Comparison groups ‐ A) External (general population) B) Women with punch biopsy but no treatment C) Internal (pre‐treatment pregnancies) D) Internal (self‐matching: same women before and after treatment) Regression analysis for age, parity and study site for some comparison groups, but not for the ones that we used in our meta‐analysis (see "Notes") Information source ‐ Records of the 12 participating NHS hospitals (for the ascertainment of the exposure); hospital episode statistics of inpatient obstetric records for the whole of England (for the ascertainment of the outcomes) Castanon 2014: Case‐control study nested in a retrospective cohort study Matching for age, parity, study site and whether the birth occurred before or after the first colposcopy; regression analysis for age, parity, study site and index of multiple deprivation
Participants	Castanon 2012: Study period for the treatment/biopsy: January 1987 to December 2009 Study period for the delivery: April 1998 to April 2010 Location for the treatment/biopsy: one of the 12 participating NHS hospitals Location for the delivery: any NHS hospital Treated group ‐ 4776 deliveries of women who had undergone excisional treatment and subsequently had a pregnancy Untreated group ‐ A) 510,660 deliveries (general population of England) B) 7263 deliveries of women who had undergone punch biopsy but no treatment and subsequently had a pregnancy C) The deliveries (1173) of the treated group before their treatment D) For 372 women who had at least one delivery both before and after the delivery, internal matching (self‐matching) was also possible: the first delivery after treatment was compared with the last delivery before treatment Exclusion criteria (for all groups): Pregnancies with no gestational age recorded, with gestational age > 43 weeks, with gestational age < 20 weeks or with no year of birth recorded as well as multiple pregnancies. Additional exclusion criteria (for the women with treatment or punch biopsy): Women for whom the date of histology was unknown Castanon 2014: Study period for the treatment/biopsy: April 1988 to December 2011 Study period for the delivery: April 1998 to March 2011 Location for the treatment/biopsy: one of the 12 participating NHS hospitals Location for the delivery: any NHS hospital Cases ‐ 768 women with a preterm birth after excisional treatment or punch biopsy. Only the earliest occurring singleton preterm birth (with any parity) in each woman was included Controls ‐ 830 matched women with a term birth after excisional treatment or punch biopsy Inclusion criteria: births at 37 weeks' gestational age, women with incomplete colposcopy records, women for whom the only pathology sample reported was non‐cervical, women who were recorded as being sterilised while pregnant, women with a diagnosis of cervical cancer at any time, women whose pregnancy was at high risk (diabetes mellitus, hypertension, placenta praevia with haemorrhage, supervision of high risk pregnancy, mental disorders, and diseases of the nervous system complicating pregnancy, childbirth, and the puerperium)
Interventions	Excision NOS (CKC, LC, LLETZ, other)
Outcomes	PTB (< 37 weeks); PTB (< 37 weeks) (D < 10 mm); PTB (< 37 weeks) (D ≥ 10 mm); PTB (< 37 weeks) (singleton pregnancies); PTB (< 33 weeks)
Notes	Castanon 2012: In addition to the treated/comparison groups that we used in our meta‐analysis, the authors had a variant of those as well: only the first birth recorded in the dataset for each woman during the study period, after exclusion of antepartum stillbirths and stillbirths of indeterminate timing. For these groups there was regression analysis for age, parity and study site, but no regression analysis for the groups that we selected. However, we selected the latter, because the population is bigger and we had no reason to restrict to the first pregnancy of each woman during the study period. Castanon 2014: this was a case‐control study nested in this retrospective cohort study. The cases were 768 preterm births and the controls were 830 term births, all occurring after excisional treatment or punch biopsy. The main outcome of this study was the depth of the cone in the cases and controls stratified in the following categories: 1 mm to ‐9 mm, 10 mm to 14 mm, 15 mm to 19 mm, ≥ 20mm. The cases and the controls, as a case‐control study, were different from the cases and the controls of all the other studies. We have contacted the investigators of the study and also used the published data to extract the PTB (< 37 weeks) rate for women with excision of < 10 mm in depth and ≥10 mm in a treated group versus women who had punch biopsy but no treatment.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias)   All outcomes	High risk	Of the total number of pregnancies (n = 26897) of the women with a pregnancy before or after treatment/punch biopsy, 8050 pregnancies (29.9%) had an unknown gestational age
Selective reporting (reporting bias)	Low risk	No reporting bias is obvious
Other bias	Unclear risk	The authors do not have information as to whether the punch biopsy group may have had a history of ablative treatment or whether the treatment group had antenatal interventions when pregnant. There is no evidence of the efficacy of these interventions so it is unclear whether this is a source of bias.
Relevant assignment described?	Low risk	Yes, treatment performed on clinical grounds
Representative intervention group?	Low risk	Women having treatment at one of the 12 participating NHS hospitals (representation from the whole England) and then delivering at any NHS hospital
Representative comparison group?	Low risk	A/B) The untreated group was drawn from the same source as the treated group C/D) Internal controls
Comparability of treatment groups?	Low risk	A) General population of the whole England B) Women with punch biopsy (it is considered that possible confounding factors are not different between this group and the treated group and thus, this is one of the best comparison groups, in general) C/D) Internal controls No regression analysis for the comparison groups that we used (see "Notes" above)

Methods	Retrespective cohort study Comparison group: A) External ‐ matching for age, parity, husband's or partner's social class, height and daily cigarette consumption B) Internal (pre‐treatment pregnancies) Information source ‐ Aberdeen Maternity and Neonatal Databank, postal questionnaires
Participants	A) Treated group ‐ 149 women who had undergone LLETZ between 1989 and 1991. Only the first singleton pregnancies following treatment that progressed to 20 weeks of gestation were included. Multiple pregnancies were excluded. We also excluded 2 miscarriages, giving a total of 147 women. Untreated group ‐ 298 women without previous treatment (two controls for each case). We excluded 3 miscarriages, giving a total of 295 women B) The 147 deliveries of the treated group after LLETZ were compared with the 133 deliveries of the treated group before LLETZ
Interventions	LLETZ
Outcomes	PTB (< 37 weeks); PTB (< 37 weeks) (singleton pregnancies); PTB (< 28 weeks); CS; precipitous labour (< 2 hours); stillbirth
Notes	1000 women who had undergone LLETZ between 1989 and 1991 were identified via Aberdeen Maternity and Neonatal Databank. A postal questionnaire was sent to these women in 1993 and 653 replied. Of these, 149 had a singleton pregnancy after treatment and were included in the treated group. The control group was also pooled from Aberdeen Maternity and Neonatal Databank.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias)   All outcomes	High risk	Postal questionnaires were used for the selection of the treated group and many women did not reply (347/1000 = 34.7%)
Selective reporting (reporting bias)	Low risk	No reporting bias is obvious
Other bias	Low risk	For the treated group, questionnaires were used for the ascertainment of the outcome; there is a risk of recall bias and misclassification of the outcome
Relevant assignment described?	Low risk	Yes, treatment performed on clinical grounds
Representative intervention group?	High risk	Postal questionnaires were used for the selection of the treated group and many women did not reply (34.7%). The women who replied are more likely to have a higher educational level.
Representative comparison group?	Low risk	A) The untreated group was drawn from the same source as the treated group B) Internal comparison group (pre‐treatment pregnancies)
Comparability of treatment groups?	Low risk	A) Matching for age, parity, husband's or partner's social class, height and daily cigarette consumption B) Internal comparison group (pre‐treatment pregnancies)

Methods	Prospective cohort study Comparison group: External ‐ matching for age (+/‐ 5 years), race, the number of prior vaginal deliveries at ≥ 20 weeks and gestational age at the time of cervical sonography (+/‐ 2 weeks) Information source ‐ medical records of one of the southern New Jersey maternal–fetal medicine offices
Participants	Treated group ‐ 85 pregnant women presenting to one of the southern New Jersey maternal–fetal medicine offices (during 2001 to 2007) with a history of LLETZ (n = 68), CKC (n = 15), or both (n = 2) Unterated group ‐ 85 pregnant women referred from the referred obstetrical ultrasound population (during 2007 to 2008) without previous cervical surgery Exclusion criteria (for both groups): multiple gestations, a clinical history of cervical insufficiency (defined as a history of repeat midtrimester pregnancy loss associated with painless cervical dilatation), presence of a cerclage or planned cerclage, ruptured membranes, or a fetal aneuploidy or major anomaly recognized at the time of cervical sonography
Interventions	Excision NOS (CKC; LLETZ)
Outcomes	PTB (< 37 weeks); PTB (< 37 weeks) (singleton pregnancies); PTB (< 34 weeks); CS; cervical cerclage
Notes	No enrolled patients were excluded from analysis after cervical sonography had been performed. The researchers had difficulty finding a matched control for one of the study participants, a 40‐year‐old Caucasian with four previous vaginal deliveries. They finally identified a Filipino gravida who otherwise matched the study patient.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias)   All outcomes	Low risk	No enrolled patient was excluded from analysis after cervical sonography had been performed
Selective reporting (reporting bias)	Low risk	No reporting bias is obvious
Other bias	Low risk	No other source of bias is obvious
Relevant assignment described?	Low risk	Yes, treatment performed on clinical grounds
Representative intervention group?	Low risk	All eligible for the study women having a cervical sonography during 2001 to 2007
Representative comparison group?	Low risk	The untreated group was drawn from the same source as the treated group
Comparability of treatment groups?	Low risk	Matching for age (+/‐ 5 years), race, the number of prior vaginal deliveries at ≥ 20 weeks, gestational age at the time of cervical sonography (+/‐ 2 weeks)

Methods	Retrospectice cohort study Comparison group: Internal (pre‐treatment pregnancies) Information source ‐ Hospital records of the Hospital Mere‐Enfant, Nantes, France
Participants	Treated group ‐ 54 women (53 pregnancies) under 39 years of age who had undergone LC between 1 July 1982 and 30 June 1992 and had one or more pregnancies after their operation. Untreated group: 38 women of the treated group (59 pregnancies) who had also a pregnancy before treatment Inclusion criteria: Only pregnancies leading to the birth of a child were included.
Interventions	LC ‐ two different techniques: Before 1986, hand‐held laser (10/54) under GA with 2 stitches, cone‐shaped 1 cm to 2 cm deep, radius 1 cm to 1.5 cm, LA for haemostasis ‐ After 1986, micromanipulator (44/54), less radical, cylinder, 0.8 cm to 1.8 cm deep, radius 0.6 cm to 0.8 cm.
Outcomes	PTB (< 37 weeks); Threatened PTL; pPROM; CS; chorioamnionitis; cervical cerclage
Notes	Between 1 July 1982 and 30 June 1992, 222 women under 39 years of age underwent CO2 laser conisation for CIN. 27 had subsequent hysterectomy or tubal sterilisation, and 48 others could not be recontacted. Thus, 147 women were available for study. Of these 147 women, 54 had a total of 71 pregnancies after the operation. Of these 71 pregnancies, 53 led to the birth of a live child (two sets of twins). These 53 pregnancies were included in the treated group. Of the 54 women which made out the treated group, 38 had also a pregnancy before the operation. These 38 women had a total of 82 pregnancies before the operation. Of these 82 pregnancies, 59 (all monofetal) led to the birth of a live child. These 59 pregnancies were included in the control group.
*Risk of bias*
Bias	Authors' judgement	Support for judgement
Incomplete outcome data (attrition bias)   All outcomes	High risk	Of the 222 women who underwent LC between 1 July 1982 and 30 June 1992, 48 (21.6%) could not be recontacted.
Selective reporting (reporting bias)	Low risk	No reporting bias is obvious
Other bias	Low risk	No other obvious source of bias
Relevant assignment described?	Low risk	Yes, treatment performed on clinical grounds.
Representative intervention group?	Low risk	All women that had LC in a single university hospital in France between July 1982 to June 1992
Representative comparison group?	Low risk	Internal comparison group (pre‐treatment pregnancies)
Comparability of treatment groups?	Low risk	Internal comparison group (pre‐treatment pregnancies)

Study	Reason for exclusion
Al‐Halal 2013	Comparison of women with CIN (+/‐treatment) versus general population. We could not include this in any comparison as we had no information as to whether women with CIN had or not treatment and the increase in the PTB could be attributed to either.
Althuisius 2001	No untreated comparison group
Berghella 2004	No untreated comparison group
Berretta 2013	No untreated comparison group
Bull‐Phelps 2007	No untreated comparison group
Chevreau 2017	No untreated comparison group
Ciavattini 2014	Both the treated and untreated group consisted of a high‐risk population (women that conceived via assisted reproductive technology (ART))
Ciavattini 2015	This study reports the miscarriage rates in the treated and untreated group but none of the outcomes that we are studying
Conner 2013	No untreated comparison group
Ferenczy 1995	No untreated comparison group
Gentry 2000	This study reports the cervical length before and after LLETZ (self‐matching), but we do not study this outcome
Gordon 1991	No untreated comparison group
Gronroos 1979	No untreated comparison group
Kalliala 2012	This study reports the Incidence of any pregnancy, livebirths, miscarriages, extrauterine pregnancies, molar pregnancies, and  terminations of pregnancies (TOPs), but none of the outcomes that we are studying
Khalid 2012	No untreated comparison group
Kim 2016	No untreated comparison group
Kindinger 2016	No untreated comparison group
Kullander 1971	No untreated comparison group
Leiman 1980	No untreated comparison group
Liu 2014	No untreated comparison group
Liverani 2016	No untreated comparison group
Macvicar 1968	No untreated comparison group
Mariya 2016	No untreated comparison group
Masamoto 2008	No untreated comparison group
Michelin 2009	No untreated comparison group
Mitsuhashi 2000	1) Treatment during pregnancy 2) No untreated comparison group
Monaghan 1982	No untreated comparison group
Naleway 2015	This study reports the percentage of the women that fell pregnant in the treated and untreated group but not the pregnancy outcomes
Nam 2010	No untreated comparison group
Novikova 1994	No untreated comparison group
Patrelli 2008	No untreated comparison group
Pils 2014	The untreated comparison group was high risk: consisted only of women with a history of second‐trimester miscarriage
Pinborg 2015	Both treated and untreated groups consisted of a high‐risk population (women that conceived via assisted reproductive technology (ART))
Radha Bai Prabhu 2010	No untreated comparison group
Rafaeli‐Yehudai 2014	No untreated comparison group (treated women with cerclage versus treated women without cerclage)
Ricciotti 1995	This study reports the cervical length before and after LLETZ (self‐matching), but we are not studying this outcome
Rosen 1991	Treatment during pregnancy
Sangkarat 2014	No untreated comparison group
Seki 2010	1) Treatment during pregnancy 2) No untreated comparison group
Shin 2010	No untreated comparison group (treated women with cerclage versus treated women without cerclage)
Sljivancanin 2013	Treatment during pregnancy
Smaldone 2010	Comparison of women with CIN (+/‐treatment) versus general population. We could not include this in any comparison as we had no information as to whether women with CIN had or not treatment and the increase in the PTB could be attributed to either.
Spracklen 2013	Data only on fertility
Wakita 1990	No untreated comparison group
Watson 2012	Case‐control study
Wongtiraporn 2014	No untreated comparison group
Zuo 2011	Comparison of women with CIN (+/‐treatment) versus general population. We could not include this in any comparison as we had no information as to whether women with CIN had or not treatment and the increase in the PTB could be attributed to either.

Methods	Retrospective cohort study
Participants	Women who had undergone conisation at Antwerp University Hospital and subsequently delivered + control group
Interventions	Conisation
Outcomes	Rates of adverse obstetric outcomes (such as LBW and CS) after conisation and comparison to the control group
Notes	Identified after April 2016, week 2; it will be included in the next version of this meta‐analysis

Methods	Population‐based retrospective cohort study
Participants	Women from whole of Norway who had undergone treatment for CIN/early invasive cervical cancer during 1998 to 2014 and subsequently had singleton pregnancy + control group
Interventions	Treatment for CIN/early invasive cervical cancer
Outcomes	Rates of adverse obstetric outcomes (such as PTB) after treatment and comparison to the control group
Notes	Identified after April 2016, week 2; it will be included in the next version of this meta‐analysis

Methods	Retrospective case‐control study
Participants	Women who had undergone conisation in a university hospital between January 2002 to January 2012 and subsequently delivered + control group
Interventions	Conisation
Outcomes	Rates of adverse obstetric outcomes (such as PTB < 37 weeks, PTB < 32 weeks, PTB < 28 weeks and PROM) after treatment and comparison to the control group
Notes	Identified after April 2016, week 2; it will be included in the next version of this meta‐analysis

Methods	Retrospective study
Participants	Women who underwent conisation from 1999 to 2005 in Peking Union Medical College Hospital
Interventions	Conisation
Outcomes	Fertility and pregnancy
Notes	Article in Chinese ‐ pending translation for inclusion in the next update of this meta‐analysis

Methods	Retrospective
Participants	Women with previously made conisation of the uterine cervix
Interventions	Conisation
Outcomes	Term, deliveries, preterm deliveries and spontaneous abortions
Notes	Article in Yogoslavian ‐ pending translation for inclusion in the next update of this meta‐analysis

Methods	Population‐based multicentric trial
Participants	Turkish women with singleton pregnancy during 2007 to 2013 who had previously undergone CKC + control group
Interventions	CKC
Outcomes	Rates of PTB after CKC and comparison to the control group; data about removed volume and height of the removed cone and correlation to PTB
Notes	Identified after April 2016, week 2; it will be included in the next version of this meta‐analysis

PERMALINK

Obstetric outcomes after conservative treatment for cervical intraepithelial lesions and early invasive disease

Maria Kyrgiou

Antonios Athanasiou

Ilkka E J Kalliala

Maria Paraskevaidi

Anita Mitra

Pierre PL Martin‐Hirsch

Marc Arbyn

Phillip Bennett

Evangelos Paraskevaidis

Abstract

Background

Objectives

Search methods

Selection criteria

Data collection and analysis

Main results

Authors' conclusions

Plain language summary

Summary of findings

Background

Description of the condition

Description of the intervention

How the intervention might work

Why it is important to do this review

Objectives

Methods

Criteria for considering studies for this review

Types of studies

Types of participants

Types of interventions

Types of outcome measures

Primary outcomes

Secondary outcomes

Search methods for identification of studies

Electronic searches

Searching other resources

Data collection and analysis

Selection of studies

Data extraction and management

Assessment of risk of bias in included studies

Measures of treatment effect

Unit of analysis issues

Dealing with missing data

Assessment of heterogeneity

Assessment of reporting biases

Data synthesis

1.15. Analysis.

1.16. Analysis.

Subgroup analysis and investigation of heterogeneity

Sensitivity analysis

Results

Description of studies

Results of the search

1.

Included studies

Excluded studies

Risk of bias in included studies

2.

3.

Summary of findings for the main comparison. The effect of treatment for CIN on maternal outcomes.

Summary of findings 2. The effect of treatment for CIN on neonatal outcomes.

Allocation

Blinding

Incomplete outcome data

Selective reporting

Other potential sources of bias

Effects of interventions

MATERNAL OUTCOMES

Preterm birth

1.1. Analysis.

1.2. Analysis.

1.3. Analysis.

1.4. Analysis.

1.5. Analysis.

1.6. Analysis.

1.7. Analysis.

1.8. Analysis.

1.9. Analysis.