Cervical intraepithelial neoplasia

Introduction: precancerous changes of cervix in pregnancy

Most of the precancerous lesions of the cervix occur in women of childbearing age. The mean age of incidence of precancerous changes of cervix requiring treatment is 30 years (3).

Consequently, it is common to diagnose precancerous lesions during pregnancy. Many expectant mothers are anxious regarding the po­tential for fetal and pregnancy complications during pregnancy, and the news of a diagnosis of a precancerous condition heightens these anxieties.

Generally, clinicians and their patients need to decide between conservative management until after delivery or treatment during pregnancy, both of which can cause significant psychological mor­bidity. The overall risk of progression of CIN3 to invasive carcinoma is uncertain, but the limited observational data available suggests that it may be in the region of 30-50%. A retrospective observational study of 143 women in New Zealand with high-grade CIN managed conservatively showed a cumulative incidence of invasive cancer of the cervix or vaginal vault of 31.3% (95% confidence interval (CI) 22.7-42.3) at 30 years, and 50.3% (95% CI 37.3-64.9) in the subset of 92 such women who had persistent disease within 24 months (5). It is stated that the rate at which invasive cancer develops from CIN is usually slow, measured in years and perhaps decades (6). However, a meta-analysis of the published literature showed a small risk of progression of high-grade cervical cytology to invasive cancer over 24 months of 1.44% (95% CI 0-3.95%) (7).

Pregnancy and pathogenesis of CIN and cervical cancer

CIN arises from the transformation zone, where columnar epithe­lium transforms through a normal process called metaplasia into squamous epithelium. Dysplasia is the pathological process which distorts this process, under the impact of oncogenic ‘high-risk’ human papillomavirus (HPV) infection. Squamous metaplasia ac­celerates during puberty and most importantly during first preg­nancy. The impact of first pregnancy on the transformation zone and the increased vulnerability to dysplasia appears to be greater in very young women, leading to AN increased risk of high-grade CIN and invasive cancer. Pooled data worldwide suggest the relative risk for developing cervical cancer in later life women with the first full- term pregnancy at age less than 17 years compared with 25 years or older is 1.77 (95% CI 1.42-2.23) (9).

Cervical cytology in pregnancy

Indication

In countries such as the United Kingdom with national cervical cancer screening programmes with effective call/recall systems, rou­tine cervical cytology samples are generally avoided in pregnancy (10). However, in many countries pregnancy is thought to be an excellent opportunity to screen not only for precancerous cells in the cervix but also for various sexually transmitted infections.

Pitfalls

Performing a cytology sample in early pregnancy is safe, with no apparent risk to the pregnancy. However, cytological interpretation can be challenging and in some cases can cause confusion and a di­lemma both for patients and clinicians. The pathologist must be in­formed if the sample is taken from a pregnant woman. Since 1960, a number of researchers have investigated unique pregnancy-related changes on cervical cytology samples in an effort to reduce diag­nostic errors.

The cervix undergoes both glandular and stromal changes during pregnancy, similar to those occurring in the endometrium (13). The glandular epithelium frequently everts onto the ectocervix to form a glandular ectropion, caused by pregnancy-related hormonal effects, and subsequently the acidic vaginal milieu precipitates extensive squamous metaplasia. Immature metaplastic cells in a large quantity can be misinterpreted on cervical cytology as high-grade precan- cerous change.

Sloughed off decidual cells can be mistaken for high- or low-grade CIN as they acquire orangeophilic cytoplasm and pyknotic nuclei in the process of degeneration (14). Hyperplastic and hypertrophic endocervical glands with Arias-Stella type changes can mimic atyp­ical glandular cells of uncertain significance (15, 16). The recom­mended investigations in non-pregnant patients of endocervical curetting and endometrial biopsy for atypical glandular cells are ab­solutely contraindicated in the pregnant population (17). Therefore, this has the potential of giving rise to anxiety for the entire dur­ation of the pregnancy, before the changes can be effectively inves­tigated after the puerperium. Transrectal ultrasound and diagnostic conization during pregnancy have been considered in these circum­stances. However, the transrectal ultrasound has low sensitivity for small volume endocervical pathology in pregnancy, and conization is associated with significant pregnancy risks and cervical haem­orrhage risk, and therefore neither are considered routine clinical practice (18, 19).

Colposcopy in pregnancy

Indication

Under most circumstances, colposcopy in pregnancy is performed to exclude invasive malignancy in order to safely defer treatment of high-grade CIN until after the natural completion of the pregnancy. The indications for colposcopy are the same as in non-pregnant patients, except that the follow-up colposcopy after adequate exci­sional treatment of CIN lesions can be deferred until completion of the postpartum period (10). Performing colposcopy is safe in preg­nancy but the interpretation of findings is challenging because of varied pregnancy-induced changes of the cervix. Beyond 20 weeks’ gestation, it is advisable to use corrective measures for the colpos­copy couch to prevent gravid uterus-induced supine hypotension.

Pitfalls

Increased oestrogen levels in pregnancy increase the volume of the cervix, and cause pronounced eversion of the columnar epithelium leading to a glandular ectropion (20). When the upper limit of the transformation zone is hidden within the endocervical canal in early pregnancy, this frequently changes during the course of pregnancy leading to adequate colposcopy with full visualization of the trans­formation zone by the third trimester. However, during the second half of pregnancy the more relaxed vaginal walls and abundant cer­vical mucous can obscure visualization. Third-trimester colposcopy is more technically challenging and is best avoided where possible. The cervix appears vascular and oedematous, often with a blue discol­ouration of the squamous epithelium and florid glandular ectropion with contact bleeding, leading to inaccurate colposcopic suspicion of significant cervical pathology. It is easy to under- or overestimate the severity of the lesions during pregnancy. Proliferation and dilatation of cervical surface vessels progressively increase with pregnancy and lesions can appear higher grade colposcopically than revealed on histology. Decidualization of cervical stroma can mimic high-grade lesions with dense acetowhite plaques and spidery superficial blood vessels.

Management of CIN in pregnancy

In view of the difficulties and limitations of colposcopy in preg­nancy, it is prudent to repeat colposcopy examination for the initial assessment of cervical cytology abnormalities 12 weeks following delivery, even if high-grade CIN is not detected. CIN1 in early preg­nancy should be managed with initial conservative management. A significantly higher tendency of regression of CIN1 lesions in pregnancy compared with non-pregnant controls has been reported (69% vs 48.7%; P = 0.03) (21). The United Kingdom National Health Service Cervical Screening Programme recommends that if CIN2 or -3 is suspected, repeat colposcopy is recommended at the end of the second trimester, to exclude progression to invasive carcinoma. If the pregnancy has already advanced beyond the second trimester, colposcopy can be deferred until 3 months following delivery (22).

Excisional treatment is rarely performed by most colposcopists during pregnancy, balancing the apparent low risk of progression to invasive disease with the potential for disruption to the pregnancy and the risk of haemorrhage. There is little evidence to quantify the risk of excisional or ablative treatment precipitating miscarriage. The risk of miscarriage and severe bleeding is likely to be small in the first trimester and some colposcopists advocate treatment of high-grade CIN if it presents very early in pregnancy. However, due to the apparent very low-risk of progression to cancer during the course of the pregnancy, most advocate conservative management.

If biopsy is indicated, sharp biopsy forceps should be used to re­duce tearing damage to the vascular cervix. The pregnant cervix should not be biopsied without adequate recourse to means of haemostasis. Monsel’s paste and silver nitrate should available, but both are caustic and can cause considerable slough if used for pro­longed period on a pregnant cervix. Gelatine paste, oxidized re­generated collagen, microfibrillar collagen, or a thrombin-soaked gelatine sponge can be applied for haemostasis with or without va­ginal packing if significant bleeding is encountered (8).

Particularly during pregnancy, cervical punch biopsy can miss invasive disease. If there is suspicion of an invasive carcinoma, a diagnostic wedge or diathermy loop biopsy should be considered, potentially during examination under anaesthesia. In keeping with other surgical procedures during pregnancy, the cumulative ma­ternal and fetal risks may be lowest if the general anaesthesia pro­cedure is performed during the second trimester. If carcinoma is confirmed in the first trimester, termination of pregnancy may be considered, while carcinoma diagnosed during the second trimester may be managed by expediting delivery once adequate fetal maturity has been achieved (23). Cases have been reported performing rad­ical trachelectomy during pregnancy with preservation of the uterus (24) and utilizing chemotherapy during late pregnancy to manage the cancer while the pregnancy progresses (25). Investigation with magnetic resonance imaging can be performed during pregnancy, and depending on the stage, a Wertheim’s hysterectomy caesarean section can be planned for delivery in conjunction with the obstet­rics and neonatology teams (23).

Mode of delivery

It has been hypothesized that vaginal delivery will lead to extensive remodelling of the cervix, allowing a higher chance of spontaneous regression of precancerous lesions compared to caesarean section delivery (26). However, other authors suggest that mode of delivery is not a significant risk factor for either progression or regression of precancerous lesions (27).

There is concern regarding the vertical transmission of high-risk subtypes of HPV during vaginal delivery, posing a theoretical risk of conditions including recurrent respiratory papillomatosis. However, respiratory papillomatosis is most commonly caused by HPV subtypes 6 or 11, associated with benign genital warts, while high­grade CIN is most commonly associated with HPV 16 or 18 (28). It has been estimated that the risk of a child contracting recurrent re­spiratory papillomatosis from a mother who has active condylomata and delivers vaginally is approximately 1 in 400, and the maternal and fetal risks associated with caesarean section likely outweigh the small fetal risk of HPV transmission during vaginal delivery (29).

Therefore, the evidence suggests that the route of delivery of preg­nant women with precancerous lesions of the cervix should be based on standard obstetric parameters.

Puerperium

Observational studies suggest that there is a trend of natural regres­sion of severity of cervical precancerous changes through pregnancy and the puerperium (30-33). However, it is not clear if this relates to the extensive remodelling and inflammation that occurs in the cervix during pregnancy and the postpartum period, or because of the nat­ural history of HPV infection. Both colposcopic and cytological as­sessments may be difficult during the immediate postpartum period because of the relative hypo-oestrogenic state. It is therefore advis­able to defer assessment for at least 8-10 weeks following delivery, to allow involution of the cervix. In the United Kingdom, a 12-week interval is recommended (22). Cases of abnormal cervical cytology tests undergoing colposcopy during pregnancy should be reassessed with another colposcopy regardless of the findings during preg­nancy. Colposcopists should be aware that compared to antenatal observations, lesions may appear smaller and more centrally located in the cervix after the puerperium (17).

Effect of excisional treatment on subsequent pregnancy

On fertility

It has been suggested that destruction and removal of part of the mucous-producing glands of the endocervix can interfere with fertility and a healthy pregnancy (34). Additionally, authors have proposed that cicatrization of the cervix after excisional treatment leads to cervical stenosis which may compromise fertility (35). Studies to evaluate such risks are difficult to conduct as there are many variables and confounding factors including social, clinical, psychological, and lifestyle issues. A large Finnish study involving follow-up of over 250,000 women-years did not report any nega­tive effect on fertility (36). However, a United States-based study evaluating 152 treated women and 1172 untreated women, sug­gested an increased risk of subfertility with prolonged (>1 year) time to conceive (37). A systematic review and meta-analysis found no conclusive evidence to suggest that fertility is affected adversely by excisional treatment (38).

On pregnancy and preterm birth

First trimester

The rate of first-trimester miscarriage is not significantly different between women treated for CIN and untreated women. The rate of termination of pregnancy and ectopic pregnancy were found to be higher in treated women. The reason for this is unclear but may relate to lifestyle factors and sexual behaviours of the group of women studied, power of the studies analysed, or publication bias (38).

Second trimester

A meta-analysis identified a significantly higher rate of second- trimester miscarriage in women previously treated for CIN com­pared to the cohort without a history of CIN (relative risk 2.60; 95% CI 1.45-4.67). It is unclear if this is related to cervical weak­ness, ascending infection due to possible dysfunction of the mu­cous plug, or the method of excisional treatment (38). Analyses are complex, as the increased risk of adverse pregnancy outcomes such as second-trimester miscarriage may not be attributable solely to the treatment itself, but to common risk factors that predispose to both precancerous cervical conditions and these obstetric complica­tions. Data pertaining to valid control groups are difficult to capture. Additionally, treatment modalities are likely to have varying impacts on the cervix, and the impact on the cervix for any given modality may vary from operator to operator.

Third trimester

Meta-analyses published in 2006 and 2008 suggested adverse third- trimester obstetric outcomes in women who underwent excisional treatment of CIN (39, 40). Cold knife conization had the greatest increased risk of preterm delivery, low birthweight, and caesarean section. The most prevalent excisional treatment (large loop exci­sion of the transformation zone (LLETZ )) was also associated with a significant increase in premature delivery. However, extreme pre­term birth (<32 weeks) was associated with cold knife conization but not LLETZ. Perhaps most importantly, there was no difference seen in perinatal mortality between women previously treated for CIN and untreated women.

A meta-analysis in 2015 included 20,832 women who gave birth after treatment for CIN before pregnancy, 52 women who gave birth after treatment for CIN during pregnancy, 64,237 women with CIN who gave birth before treatment, and 8,902,865 women who gave birth without CIN. Compared to women with untreated CIN, women treated for CIN before or during pregnancy had a signifi­cantly higher risk of preterm birth (<37 weeks) (odds ratio (OR) 1.7; 95% CI 1.0-2.7). Women treated for CIN before pregnancy had no significant difference in the rate of premature birth to women with CIN where treatment was delayed until after pregnancy (OR 1.4; 95% CI 0.85-2.3). Women treated during pregnancy had a clearly increased risk for premature delivery (OR 6.5; 95% CI 1.1-37), and premature rupture of membranes (OR 1.8; 95% CI 1.4-2.2) (41). These data support the policy of deferral of treatment of high-grade CIN until the puerperium.

While these systematic reviews have established an association between excisional treatments for CIN and premature delivery, the cause remains uncertain. An association between diagnosis of high­grade CIN and other factors predisposing for premature delivery such as smoking is hypothesized. Additionally, anatomical changes following cervical treatment, cicatrization of cervix, and alteration of cervical mucous and flora may predispose to prematurity. It has also been proposed that the pathophysiology of CIN could be con­tributory (42). The dose-effect relationship of cervical tissue excised appears relevant, with greater cervical stromal excision increasing the subsequent cervical deficit and prematurity risk (43).

A meta-analysis in 2008 suggested that ablative treatments are not associated with increased perinatal mortality (40). LLETZ has superseded ablative treatment in Western countries because it allows confirmation of clearance of disease and full histological examination, but many studies show that both are equally effective (44). Analysis is ongoing to explore the safety and validity of ab­lative techniques with satisfactory colposcopy in women with re­productive wishes, to minimize the potential impact on future pregnancies (45). Patients who intend to have a further pregnancy should undergo documented and well-informed counselling re­garding potential risks and benefits before any excisional treatment is performed.