SEARCH

SEARCH BY CITATION

Keywords:

  • conisation;
  • preterm delivery

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

Please cite this paper as:Ørtoft G, Henriksen T, Hansen E, Petersen L. After conisation of the cervix, the perinatal mortality as a result of preterm delivery increases in subsequent pregnancy. BJOG 2010;117:258–267.

Objective  To determine the effects of one or two conisations on preterm delivery and perinatal mortality in subsequent pregnancies.

Design  A population-based cohort study.

Setting  Aarhus University Hospital.

Population  Preterm delivery and mortality rates were evaluated in 721 deliveries after one conisation, and in 37 deliveries after two conisations, and were compared with 390 deliveries after dysplasia and 74 552 deliveries that were not preceded by conisation or dysplasia.

Methods  Cox regression was used to evaluate preterm delivery rates and perinatal mortality.

Main outcome measures  Birthweight, gestational age (prior to 28, 32, and 37 weeks of gestation, respectively) and perinatal mortality.

Results  The risk of preterm delivery was increased after one conisation [adjusted hazard ratios (95% CI): <37 weeks, 2.8 (2.3–3.5); <28 weeks, 4.9 (2.5–9.7)], and was further increased after two conisations [adjusted hazard ratios (95% CI): <37 weeks, 9.9 (6–17); <28 weeks, 9.8 (1.4–70)], compared with no conisation. One conisation was associated with an increased perinatal mortality [<28 weeks, 9.9 (4.0–25)]. All three methods of conisation [large loop excision of the transformation zone, electroknife and cold knife] increased the risk of preterm delivery.

Conclusions  A single conisation was associated with a 2.8-fold increased risk of perinatal death, most likely because of a 4.9-fold increase in extreme preterm delivery. Only 37 patients had two conisations, and the results showed a ten-fold increase in the risk of preterm delivery.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

Among the 5 million people in Denmark, approximately 400 women are diagnosed with cancer of the cervix, and approximately 150 of these women will die from their disease every year (data from screening of cervical cancer, National Board of Health, Denmark; http://www.sst.dk/English.aspx). Formal screening programs for cervical cancer were introduced in the late 1960s, and organised screening has been taking place in most but not all danish counties since the late 1980s (Pap smear every third year for age 23–50 years, and every 5 years for age 50–60 years). The number of new cervical cancer cases has decreased from almost 1000 in 1966 to <500 in 2003 (data from screening of cervical cancer, National Board of Health, Denmark; http://www.sst.dk/English.aspx). However, the incidence of cervical cancer in Denmark is high compared with other Scandinavian countries, with an incidence of 14 per 100 000 women per year, compared with nine and four in Sweden, and Finland, respectively.1 As a consequence of the screening program, around 4600 women will need a conisation to treat cervical intraepithelial neoplasia every year, so as to prevent an estimated 500 cases of cervical cancer. In the 1990s, conisation was regarded as a small and safe surgical procedure, and several small studies showed no increased risk of preterm delivery.2–6 However, in 1999 El-Bastawissi was the first to demonstrate an increased risk of preterm delivery after conisation (OR = 1.6),7 and since then several studies including a Danish study have confirmed these results.1,8–12 A single study has also demonstrated an increase in perinatal mortality after one conisation.13 In Denmark, the preterm delivery rate has increased over the past 10 years,14 and spontaneous preterm births in primiparous women with singleton spontaneous pregnancies have increased by 51% during this period, from 3.8% to 5.7%.14 Also, the age at which women give birth for the first time has increased over this period, probably making conisation before pregnancy and childbirth more common. Whereas other risk factors for preterm delivery, such as smoking in pregnancy, have decreased, conisation may be one of many factors contributing to the increased preterm delivery rate. Because preterm delivery is the single most important factor related to perinatal mortality and serious newborn morbidity, risk factors and the potential for preventive activities are important areas of research. Conclusions are difficult regarding the association between conisation and preterm delivery because of several methodological problems in previous studies (i.e. small sample size, case–control design and potential problems with a proper control group, poor control for confounding factors, no censoring).13 Thus, larger cohort studies with an emphasis on the pathophysiology of preterm delivery are needed. In this study, we aimed to secure valid details on gestational age, delivery, preterm inductions, and potential confounding factors. Furthermore, we have included some observations on the risk related to two conisations and also to dysplasia not treated with conisation. The current study was performed to evaluate the consequences of one or two conisations on preterm delivery and perinatal mortality in subsequent pregnancies, with special emphasis on the specific gestational age and the pathophysiology of preterm delivery.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

The pathology database

The pathology database is a Danish nationwide database containing the results of all cytology and histopathology performed in Denmark. The Department of Histopathology at the University Hospital of Aarhus has online access to the database, and we retrieved the pathology reports for all conisations and all dysplasia not treated with conisation, including: information on the degree of dysplasia, cervical procedures, dimensions of the biopsies, preconisation diagnosis, biopsy margins, and final histological diagnosis. If a prebiopsy diagnosis was missing, these data were collected from the patient’s hospital records.

The Aarhus birth cohort

The Aarhus birth cohort covers all pregnancies and deliveries at Aarhus University Hospital, Denmark, from 1989 onwards: approximately 8% of all Danish births. When a pregnant woman is referred from a general practitioner to the hospital for prenatal visits, she receives a letter with information about the cohort and a set of structured questions related to her current characteristics (height, weight, age), obstetric history (menstrual history, previous pregnancies, including outcome), medical history, and lifestyle factors (smoking, alcohol and coffee intake, drug abuse), and socio-economic factors (education, marital and employment status). The completed questionnaires are returned during the early second trimester. Only 86% of women in the database answered the questionnaires; non-responders were, however, not excluded. At the time of delivery, the midwife gave information on gestational age at birth, ultrasound examinations during pregnancy, and details about the course of delivery (mode and time of delivery, rupture of membranes, contractions, stimulation, and complications) and the newborn. A structured coding sheet was used to collect these data, and all information was validated by a research midwife with the use of an extensive coding manual and the women’s medical records. Data on the course of and complications during pregnancy and delivery were available for all women in the cohort.

Study population

All conus biopsies examined at the Institute of Histopathology, University Hospital of Aarhus, between 1989 and 2007 were identified from the Danish nationwide pathology database, and 5111 conus biopsies were identified: 4783 patients had one conisation, 161 patients had two conisations, and two patients had three conisations. Most conisation procedures were performed at the University Hospital of Aarhus, but a small number were performed at specialist clinics in Aarhus or at the County Hospital in Odder. To find the patients who had given birth after the conisation, each patient’s unique personal identification number was checked in the Aarhus Birth Cohort, and all singleton deliveries that had given birth after conisation at the Aarhus University Hospital until March 2007 were identified. Only singleton deliveries with valid gestational ages and birthweights were analysed. Gestational age was estimated from an ultrasound measure of: the highest fetal crown–rump length of 5–44 mm (if available); the smallest crown–rump length of 45–85 mm (if available); or the smallest fetal biparietal diameter of 22–50 mm (if available). For 80.3% of all deliveries, an ultrasound-based gestational age was known. If no ultrasound was available, the gestational age was based on the last menstrual period corrected for cycle length (6.6% of deliveries) or the last menstrual period (2.0% of the deliveries). In 2.2% of the deliveries the gestational age was based on the midwife’s best estimate from the combination of less reliable ultrasound measures and the date of the last menstrual period, and was given in days. In the remaining 6780 (9.1%) deliveries, the gestational age was based on the midwife’s best estimate from the combination of less reliable ultrasound measures and the date of the last menstrual period, and was given in weeks. Analysis of the preterm delivery rate with and without these less reliable gestational ages revealed a preterm deliveries rate in the no-conisation group of 4.1% if the less reliable data were included, and 3.7% if the less reliable data were excluded. We therefore decided to include these less reliable gestational ages to give as conservative an estimate as possible of the risk of preterm delivery after conisations. Preterm delivery was defined as delivery prior to 37 weeks of gestation, and was further categorised into deliveries prior to 32 and 28 weeks of gestation. Perinatal death was defined as stillbirth or death within the first week of life. Primary rupture of the membranes was considered present when the rupture of the membranes occurred before contractions. Birthweight was subcategorised as birthweights below 2500, 2000, or 1500 g.

Conus procedure

Conisations were carried out after cervical biopsies showed severe dysplasia or higher stage findings (i.e. cervical intraepithelial neoplasia stage 3, CIN 3), persisting moderate dysplasia (>18 months) in women under 35 years of age, and moderate dysplasia for women aged 35 years or more. Furthermore, conisations were carried out if Pap smear results showed planocellular or adenocarcinoma cells, persistent carcinoma in situ (CIS), or persistent severe dysplasia.

Of the 710 women with one conisation (excluding preterm induction of labour; Table 1) prior to pregnancy, the conus procedures were distributed as follows: 572 women had a loop electrosurgical excision procedure (large loop excision of the transformation zone, LLETZ), 71 women had an electrosurgical needle procedure (electroknife), and 67 had a cold-knife procedure. From 1989 to 1992, all conisations were performed with the cold-knife procedure, and from 1992 to 1995 the cold-knife procedure was only used for high conisations and in pregnant women. After 1995, the cold-knife procedure was omitted, and instead the electroknife was used when a high conus biopsy was warranted. The conus biopsies were submerged and transported in saline, cut open, fixed in formaldehyde in a stretch position, embedded in paraffin, sectioned, and stained for further histopathology examination. The height and width of the conus were measured by the pathologist, and the volume of the conus biopsies was calculated as the volume of a cone: volume = hr2/3), where h is the conus height and r is the conus radius. In our department, a so-called ‘cowboy hat’ conisation (conisation with an additional slice taken from the endocervical canal) is seldom used because we prefer to use the electroknife if a high conus biopsy is warranted. If the pathology report revealed a cowboy hat, the heights of the two conus slices were added for the height and volume calculations.

Table 1.   Total number of patients in the birth cohort and detailed information on patients censored for the Cox analysis of the preterm delivery rate shown in Table 2
 Total numberCensored patientsCensored because of
Preterm induced birthPreterm acute caesarean section before labourPreterm elective caesarean section
One conisation72111452
Two conisation371010
No conisation745521653776530347
Dysplasia3907241

Conus diagnosis

Because of the department’s conservative indications for conisations, the indications for the first conisation were severe dysplasia (CIN3) or higher stages in almost all patients (89%), and only 11% were performed because of moderate (CIN2) or mild dysplasia (CIN1). The final pathologic examination confirmed the indication, except in six women who had invasive squamous cell carcinoma in their final pathology report.

Statistics

The data were analysed by stata.15 Hazard ratios (HR), odds ratios (OR) and liniar regression coefficients (r) are given with 95% confidence intervals (CI). Gestational age and birthweight were presented as mean values with standard deviations (SD). Differences between mean gestational ages and birthweights were tested by a Mann–Whitney U test because the variances in the different groups were unequal. The adjusted HR was calculated by Cox regression analysis. The mother’s age, smoking status, and parity were entered in all multivariate models, whereas marital status (cohabiting versus single) and educational level (<10 years versus 10 years+) were assessed in these models, and remained in the final models if they changed the HR by at least 10%. To avoid excluding all women with missing values for potential confounding factors in the Cox regression, we evaluated smoking status as ‘yes’, ‘no’ or ‘not known’ as valid categories of the variable. For the Cox regression, the preterm inductions prior to delivery before contractions or preterm premature rupture of membranes (PPROM) were censored in the analysis at time of delivery. Linear regression was used to test the associations between gestational age and size of conus (conus height, circumference, and volume) and time from conus until pregnancy. For dichotomised outcomes, crude ORs with 95% CIs were calculated. For the patients who gave birth before as well as after the conisations, paired Student’s t tests were used for comparing differences in birthweights, whereas the Wilcoxon matched-pairs signed-rank test was used for testing differences between gestational ages because of uneven variances. The OR was tested by the McNemar matched-pairs test. Two-sided P values of <0.05 were considered to be statistically significant.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

Two exposure categories were identified: women with their first delivery after one conisation (n = 721) and women with their first delivery after two conisations (n = 37). The women could therefore be primiparous or multiparous, depending on which pregnancy came after their conisation. In women with one or two conisations prior to pregnancy, no deliveries were excluded, giving 721 deliveries after one conisation, and 37 deliveries after two conisations. If the women with one or two conisations had had further deliveries during the study period, these deliveries were excluded from the control group. In the control groups, 335 deliveries were excluded because of missing gestational age (132) and missing birthweight (203), leaving a total of 74 552 eligible deliveries without previous conisations and 390 deliveries with dysplasia not treated with conisation. Preterm inductions before delivery were identified, and the number of inductions is shown in Table 1. Both preterm induction of birth and preterm caesarean section before labour were censored for the Cox regression. Data that exclude censored patients are presented in Tables 2–5.

Table 2.   Gestational age (GA), birthweight (BW), perinatal mortality (peri. death), and PPROM of the first pregnancy after one or two conisations, compared with women with no conisation or women with previous dysplasia
 One conisationTwo conisationNo conisationDysplasiaHR/adjusted HR (95% CI for adjusted HR)
One conisation versusTwo conisation versus
  1. Results are given as means ± SD or % (n), where % = event of total number and n = number of event (not including censored patients as a result preterm induction or acute or elective caesarean section before labour). *P < 0.001 compared with no-conisation control group. **P < 0.01 compared with dysplasia control group.

n7103672 899383No conisationDysplasiaNo conisationDysplasia
GA275*,** ± 19266*,** ± 26279 ± 13280 ± 131.2/1.2 (1.1–1.3)1.2/1.3 (1.1–1.4)1.3/1.4 (1.0–2.0)1.4/1.4 (1.0–2.1)
GA <28 weeks1.3 (9)2.8 (1)0.3 (206)0 (0)4.5/4.9 (2.5–9.7)9.9/9.8 (1.4–70)
GA <32 weeks2.5 (18)8.3 (3)0.7 (491)1.0 (4)3.8/4.0 (2.5–6.4)2.5/2.4 (0.8–7.0)13/12 (4–38)8.3/10 (2–50)
GA <37 weeks11.1 (79)33.3 (12)4.1 (2995)3.9 (15)2.9/2.8 (2.3–3.5)3.0/3.0 (1.7–5.3)9.8/9.9 (6–17)10/11 (5–24)
BW3411* ± 6903088*,** ± 8293537 ± 5553506 ± 598    
BW <1500 g1.5 (11)2.8 (1)0.5 (376)0.5 (2)    
BW <2000 g4.5 (32)11.1 (4)1.0 (758)2.9 (11)    
BW <2500 g8.0 (57)19.4 (7)3.0 (2204)5.0 (19)    
Peri. death1.0 (7)0 (0)0.4 (312)0 (0)2.7/2.8 (1.3–5.9)
Peri. death <28 GA0.7 (5)0 (0)0.08 (57)0 (0)9.1/9.9 (4.0–25)
Peri. death <32 GA0.8 (6)0 (0)0.1 (71)0 (0)8.8/9.3 (4.1–22)
Peri. death <37 GA0.8 (6)0 (0)0.13 (98)0 (0)6.4/6.8 (3.0–15)
N7103671701377    
PPROM <28 GA0.6 (4)2.8 (1)0.1 (70)0 (0)5.9/6.8 (2.5–19)29/29 (4–207)
PPROM <32 GA1.7 (12)8.3 (3)0.3 (199)0.8 (3)6.3/6.8 (3.8–12)2.2/2.0 (0.6–7.1)31/30 (9–93)11/12 (2–60)
PPROM <37 GA8.0 (57)31 (11)2.3 (1598)2.6 (10)3.9/3.7 (2.8–4.8)3.2/3.2 (1.6–6.3)17/17 (9–31)14/14 (6–34)
Table 3.   Gestational age (GA), birthweight, perinatal mortality, and PPROM of the last children born before and first child born after one conisation
 After conisationBefore conisationMcNemar test for matched pairs Odds ratio (95% CI)
  1. Results are given as means ± SD or % (n). *P < 0.001.

n170170 
Gestational age at birth274* ± 16.5279 ± 12.2 
Birthweight3528 ± 6353517 ± 520 
GA <37 weeks10 (17)4.7 (8)2.8 (1.0–10)
Perinatal mortality00
PPROM <37 weeks6.5 (11)2.9 (5)2.5 (0.7–11)
Table 4.   Height, circumference, and volume of conisation for the different types of conisation, and gestational age and birthweight for children born of women who had had one conisation
 AllLLETZElectroknifeCold knife
  1. Means ± SDs. *P < 0.001 against the LLETZ group.

n710 572 7167
Conisation height (mm)16.4 ± 4.715.8 ± 4.219.0* ± 5.618.8* ± 5.9
Conisation circumference (mm)38.9 ± 9.237.8 ± 8.438.7 ± 9.148.1* ± 10.4
Conisation volume (mm3)709 ± 449634 ± 372828* ± 5121209* ± 606
Gestational age275 ± 19275 ± 18266* ± 26277 ± 19
Birthweight3411 ± 6903421 ± 6583252 ± 8323489 ± 772
Table 5.   Gestational age (GA), birthweight, perinatal mortality (peri. death), and PPROM of the first pregnancy after one LLETZ, electroknife or cold-knife conisation, compared with women with no conisation
 LLETZElectroknifeCold knifeNo conisationHR/adjusted HR (95% CI for adjusted HR)
LLETZ versus no conisationElectroknife versus no conisationCold knife versus no conisation
  1. Results are given as means ± SDs or % (n) where % = event of total number and = number of event (not including censored patients as a result of preterm induction or acute or elective caesarean section before labour). *P < 0.001 compared with the no-conisation control group. **P < 0.05 compared with the no-conisation control group.

n572716772 899   
GA275* ± 19266* ± 26277 ± 19279 ± 131.2/1.2 (1.1–1.3)1.8/1.9 (1.5–2.4)1.0/1.0 (0.8–1.4)
Birthweight3421* ± 6903252** ± 8323489 ± 7723537 ± 555   
GA <28 weeks0.9 (5)4.2 (3)1.5 (1)0.3 (206)3.1/3.3 (1.4–8.0)15.6/18.5 (5.9–58)5.5/7.0 (1.0–50)
GA <32 weeks1.9 (11)7.0 (5)3.0 (2)0.7 (491)2.9/3.0 (1.6–5.4)11.1/12.6 (5.2–30)4.6/5.4 (1.3–22)
GA <37 weeks9.6 (55)23.9 (17)10.4 (7)4.1 (2995)2.4/2.4 (1.8–3.1)6.8/7.0 (4.3–11.2)2.7/2.9 (1.4–6.1)
Peri. death0.7 (4)4.2 (3)0 (0)0.4 (312)1.9/1.9 (0.7–5.2)14.2/15.5 (5.0–48)
Peri. death <28 GA0.3 (2)4.2 (3)0 (0)0.08 (57)4.5/4.8 (1.2–20)55.9/65.0 (20–207)
Peri. death <32 GA0.5 (3)4.2 (3)0 (0)0.1 (71)5.4/5.7 (1.8–18)45.1/52.3 (16–166)
Peri. death <37 GA0.5 (3)4.2 (3)0 (0)0.13 (98)3.9/4.1 (1.3–13)33.3/37.9 (12–120)
n572716771 701   
PPROM <28 GA0.2 (1)4.2 (3)0 (0)0.1 (70)1.8/2.1 (0.3–15)46.1/54.1 (17–172)−/−
PPROM <32 GA1.0 (6)7.0 (5)1.5 (1)0.3 (199)3.9/4.2 (1.8–9.4)27.3/30.9 (12–75)5.6/6.3 (0.9–44.9)
PPROM <37 GA6.7 (38)19.7 (14)7.5 (5)2.3 (1598)3.2/3.0 (2.2–4.1)10.6/10.5 (6.2–17.7)3.6/3.9 (1.6–9.3)

No conus group

The no-conisation group consisted of all singleton deliveries in the study period in mothers without prior conisation or dysplasia. Of the 74 552 deliveries, 86.3% were vaginal deliveries and 13.7% were caesarean sections (6.3% planned and 7.4% emergency caesarean sections). Of the babies, 6.2% were born before 37 weeks of gestation (4.1% were spontaneous preterm deliveries, and the remaining 2.1% were born after preterm induced birth, preterm acute caesarean section, or preterm planned caesarean section as a result of intrauterine death or disease in the baby or mother), and 9% were born after 42 weeks of gestation. The overall perinatal mortality was 0.65%. The perinatal mortality after spontaneous deliveries was 0.42%.

Delivery after one conisation

Gestational age at delivery after one conisation was 275 days compared with 279 days after no conisation (< 0.0001) and 280 days in women with dysplasia (< 0.0001) (Table 2). The frequency of preterm delivery was 11.1% in women with a conisation prior to pregnancy, compared with 4.1% in women without and 3.9% in women with dysplasia. The risk of preterm delivery before 37, 32, and 28 weeks of gestation was increased with one conisation compared with none, and the risk increased with decreasing gestational age (Table 2). The perinatal mortality was 1.0% in children of women with one conisation, and 0.4% in children of women with no conisation [adjusted HR = 2.8 (1.3–5.9); P = 0.007]. The excess mortality was seen mainly in babies delivered prior to 28 weeks of gestation (Table 2). PPROM also occurred much more frequently in women with a conisation than in women without (< 0.0001 for PPROM before 37, 32, and 28 weeks of gestation) (Table 2). This difference was less pronounced when women with one conisation were compared with women with dysplasia only (P < 0.0001 and P = 0.18 for PPROM before 37 and 32 weeks of gestation, respectively). A total of 72% of the spontaneous preterm deliveries after one conisation started with PPROM, compared with 56% of the preterm deliveries in women without a conisation (< 0.01) and 67% in women with dysplasia (not significant). Similar to the findings related to preterm delivery, children of women with one conisation had a significantly decreased birthweight (3411 g) compared with children of women with no conisation (3537 g) (< 0.0001), excepting those in the dysplasia group (3506 g) (P = 0.09) (Table 2). Eight percent of children weighed <2500 g and 1.5% weighed <1500 g with one maternal conisation, compared with 3.0% and 0.5% with no conisation [OR 2.8 (2.1–3.7) and 3.0 (1.5–5.5)], and 5.0% and 0.5% with maternal dysplasia [OR 1.7 (0.9–3.0) and 3.0 (0.6–28)] (Table 2). All results remained essentially unchanged and statistically significant after adjustment for potential confounding factors. Only the mother’s age, parity, and smoking status remained in the final Cox regression.

A total of 170 women gave birth prior to a conisation and again some time after the operation (Table 3). This enabled us to compare pregnancy outcome before and after conisation in the same women, and we found that mean gestational ages at delivery were 279 and 274 days, respectively (Table 3). Of the 170 women, 10% experienced a preterm birth after conisation, whereas only 4.7% experienced preterm birth before conisation; this, however, was not quite statistically significant [OR 2.8 (1.0–10)]. Of the 25 preterm deliveries in this group, three women experienced preterm delivery before as well as after conisation, five women experienced preterm delivery before conisation, and 14 women experienced preterm delivery after conisation. Of the 25 preterm deliveries, four children were born before 32 weeks of gestation, one was born before conisation and three were born after conisation.

Delivery after two conisations

Gestational age at delivery after two conisations was 266 days, and was significantly lower than the 279 days after no conisation and 280 days after prior dysplasia (Table 2). The frequency of preterm delivery was 33% in women with two conisations prior to pregnancy. The risk of preterm delivery before 37, 32, and 28 weeks of gestation was significantly increased with two conisations compared with none (Table 2). None of the 36 children born after two conisations died perinatally. PPROM occurred much more frequently (92% of spontaneous preterm deliveries) in women with two conisations than in women without (Table 2). Similar to the findings related to preterm delivery, the number of children with a low birthweight was also increased. The numbers of children with body weights below 2500 and 1500 g were also raised in women with two conisations (two conisations, 19% and 2.8%, respectively; no conisation, 3.0% and 0.5%, respectively). All results remained essentially unchanged after adjustment for potential confounding factors. Only the mother’s age, parity, and smoking status remained in the final Cox regression.

Delivery after prior dysplasia not treated with conus

Three hundred and eighty-three women had current or earlier dysplasia (360 with moderate and 30 with severe dysplasia) not treated with conisation prior to pregnancy. The time from first diagnosis of dysplasia to delivery averaged 1339 days (66–5160). When this group was compared with the no-conisation group, no differences between the two groups were found, except for a significantly increased number of PPROM at <32 weeks of gestation [adjusted HR 1.92 (1.2–3.2)].

Effect of treatment procedure

The maximum heights of the conus biopsies were known for 646 of the 710 women with a spontaneous delivery after one conisation (Table 4). The mean height was 16.4 ± 4.7 mm (Table 4). The height of the conus biopsy was significantly associated with gestational age at delivery after one conisation (r = 0.03; P = 0.001). However, the maximal conus vaginal circumference and the conus volume were not associated with gestational age. Most of the 710 first cervical procedures were carried out by the LLETZ (n = 571). Gestational age was significantly lower in women who were operated on with the electroknife procedure (n = 71). This finding could be explained by the higher conus biopsies taken with this procedure. All data in patients having only one conisation were reanalysed to evaluate the effects of the different conisation procedures on preterm delivery (Table 5). All three procedures increased the risk of preterm delivery. The electroknife that was used for high conisations, however, revealed higher risk estimates for both preterm delivery and perinatal death than the two other conisation methods (LLETZ and cold knife).

In the mothers of the six children born prior to 32 weeks of gestation that died, the indications for the cervical procedures were severe or higher stages of dysplasia, and their cervical procedures had been performed with either LLETZ (four patients) or the electroknife (three patients).

Time from conisation to pregnancy

The time interval between the conisation and the first pregnancy after conisation varied between 99 and 4891 days (13 years), with a median of 1059 days (2.9 years). No association was found between the number of days from the conisation to delivery and the gestational age at birth [r = −0.001 (−0.003 to 0.0005)]. Eighteen women had a conisation during pregnancy. None of these patients experienced preterm birth, and the mean gestation was 279 ± 12.9 days, and the mean birthweight was 3539 ± 688 g, similar to that of women with no conisation.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

This study demonstrates a three-fold increase in the risk of preterm delivery after one conisation, and indicates a ten-fold increased risk of preterm delivery after two conisations. However, this latter group consisted of only 37 women. All three methods of conisation (LLETZ, electroknife and cold knife) increased the risk of preterm delivery. Larger studies from Denmark and Finland both found twice the risk, and a Norwegian study found more than three times the risk, of preterm delivery after a single conisation.9,12,13 The increased risk of preterm delivery after one or two conisations demonstrated in the present study is supported by the decrease in gestational age and birthweight found in these groups, although the absolute differences in gestational age and birthweight may have little clinical significance.

The strength of the present study was based on our ability to adjust for a variety of potential confounding factors (mother’s age, parity, smoking status, educational level, and marital status), our valid data on gestational age and detailed information on delivery (preterm inductions, rupture of membranes), and our detailed information on pathology, operative technique, and conus size. A further strength of the present study was the inclusion of a group of women with earlier dysplasia who had not been treated with conisation as a separate exposure group, as this group is thought to have characteristics that were more similar to women with a conisation.

A limitation of the present study was the small number of patients with two conisations and the small number of deliveries before 32 and 28 weeks of gestation in both conus groups. Because of the very small number of previous preterm deliveries and the high frequency of primiparous women, previous preterm delivery could not be included in the analyses as a confounding factor; however, we repeated all analyses after excluding women with previous preterm delivery. These analyses revealed essentially unchanged results. A slightly higher risk of preterm delivery and perinatal mortality were seen if we only included women of first parity, and essentially unchanged results were revealed if we restricted analyses to include only the first child of the women registered in the cohort. We also repeated the analyses after the exclusion of all pregnancies referred for special care because some of these could have been referred as a result of imminent extremely preterm delivery. This also revealed unchanged results. Information on in vitro fertilisation pregnancies was unfortunately not available in the present study, which may lead to unadjusted confounding.

The very high risk of preterm deliveries after two conisations has, to our knowledge, not been demonstrated previously. We also demonstrated an almost three-fold increase in the risk of perinatal mortality after one conisation, mainly because of the death of babies born prior to 28 weeks of gestation. Even though the number of events (seven perinatal deaths after one conisation) is low, the differences between the one conisation group and the no-conisation group are striking. In the Finnish study, the overall perinatal mortality was increased by 74% after one conisation.13 Arbyn et al.11 concluded in a recent meta-analysis that cold-knife conisations were associated with a serious adverse pregnancy outcome, whereas large loop excisions seem to have less severe adverse effects. In the present study we demonstrate that large loop excisions seem to have a similar serious adverse effect on pregnancy outcome compared with cold-knife conisations.

The pathophysiology behind the increase in preterm delivery after conisation is not well known. We found that conus height was associated with preterm delivery, in agreement with previous findings.8,16–18 However, we failed to find an association between conus circumference or conus volume and preterm delivery. Thus, the proportion of the cervical canal lost may be more important than total tissue mass lost as a result of the conisation. This was corroborated by the finding that two conisations increased the risk of preterm delivery even further. The cervical canal is thought to be important for the development of the cervical mucous plug during pregnancy. The cervical mucous plug contains high levels of immunoglobulin and phagocytic cells, and is thought to act as a gatekeeper to protect the fetomaternal unit against ascending microbes.19 The loss of the cervical canal’s ability to produce a sufficient mucous plug may increase the risk of infection and preterm delivery. The much higher occurrence of PPROM in the present study may also have been caused by this loss of a functional mucous plug. A mechanical weakening of the cervix because of the loss of tissue is another possible explanation.

The results of the conisation with the electroknife showed that this procedure results in significantly higher conus and higher risk estimates for preterm delivery than is seen after LLETZ conisation. As a higher conisation increases the risk of preterm delivery, this procedure should probably be used with care. However, if a high conisation is needed, for example if the transformation zone is expected to be situated high in the cervical canal, or if severe changes are expected to be located high in the cervical canal, electroknife surgery may still be indicated.

El-Bastawissi et al.7 have shown increased odds for preterm delivery after conisation for CIS. He also demonstrated a nonsignificant increase in preterm delivery after untreated CIS.7 Therefore, the presence of dysplasia, and perhaps also the severity of the dysplasia leading to the operation, might affect the preterm delivery rate per se. We included 383 patients with current or previous dysplasia, mostly moderate dysplasia and not treated with conisation, and found that these patients had no increased risk of preterm delivery compared with deliveries after no conisation. In fact, the gestational age at delivery was almost identical in women with dysplasia (280 days) and in women with no conus (279 days), indicating no effect of dysplasia per se on the preterm delivery rate. The authors do, however, accept that some degree of residual confounding could be caused by the lesser degree of dysplasia (CIN) in these patients when compared with patients with prior conisations.

In the present study the indications for conisation of women of fertile age have, for the entire period, been quite strict. Patients with severe dysplasia, CIS, or early carcinoma were offered a conisation; whereas fertile patients with moderate or light dysplasia were offered control every 4–6 months with Pap smear and colposcopy, which included biopsies from the transformation zone. A recent study in 13- to 22-year-olds showed that a conservative approach to light dysplasia will result in regression of dysplasia in 61% after 12 months and in 91% after 36 months of observation, supporting a conservative approach in order to prevent preterm deliveries resulting from conisation.20

If our results are true, the study indicates that one of approximately nine women after one conisation and one in three women after two conisations will experience a preterm delivery, compared with one in 24 women without a prior conisation. For extreme preterm deliveries, the numbers are 1:77 after one conisation and 1:36 after two conisations, compared with 1:333 of women with no conisation. Furthermore, approximately one in 143 will experience perinatal death related to extreme prematurity after one conisation, compared with one in 1250 women without a prior conisation. Severe adverse pregnancy outcome as a result of conisation has also been demonstrated in a recent meta-analysis performed by Arbyn et al.11 They also demonstrated an increased risk of preterm delivery (28/30 weeks of gestation), with a similar ‘number needed to treat for harm’ for cold-knife conisation (1:53), but with less severe estimates for large loop excision (1:250) compared with that of the present study.11

This increase in the risk of preterm delivery and perinatal mortality related to conisation prior to pregnancy may encourage the introduction of a nationwide vaccination program for human papillomavirus to minimise the need for prophylactic conisation. However, even with widespread vaccination programs, dysplasia will continue to be a problem, and studies on safety related to conservative approaches to cervical dysplasia are warranted. We suggest that all women should be informed about the risk of preterm delivery prior to conisation, and more studies are needed in these women during their pregnancies to determine their optimal management and treatment. If a conisation is necessary, the conus height should be restricted to the safest minimum.

Disclosure of interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

All authors declare that no organisation providing funds had any influence on the research. All authors declare that they have no conflicts of interest.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

GØ: project planning, data collection, data management, statistics, manuscript. TBH: project planning, data collection, statistics, manuscript. ESH: project planning, data collection, manuscript. LKP: project planning, manuscript.

Details of ethics approval

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

No approval from the ethics committee was warranted under Danish law. The study was approved by the Danish data protection board.

Funding

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

MD and HF Haderslevs Fund. University of Aarhus research initiative (Forsknings Initiativet Aarhus Universitets hospital), The FIGO fund, The Sanitorielæge Ellen Pedersens Fund , and The Classenske Fideicommis Surgical Fund.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References

We would like to acknowledge data manger Lise Viskum Hansen for her help and advice with data management, Morten Frydenberg for statistical advice, and Edwin Stanton Spencer for linguistic corrections.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Conclusions
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethics approval
  10. Funding
  11. Acknowledgements
  12. References
  • 1
    Arbyn M, Auier P, Ferlay J. Burden of cervical cancer in the 27 member states of the European Union: estimates for 2004. Ann Oncol 2007;18:14235.
  • 2
    Blomfield P, Buxton J, Dunn J, Luesley D. Pregnancy outcome after large loop excision of the cervical transformation. Am J Obstet Gynecol 1993;169:6205.
  • 3
    Cruickshank M, Flannelly G, Campbell D, Kitchener H. Fertility and pregnancy outcome following large loop excision of the cervical transformation zone. Br J Obstet Gynaecol 1995;102:46770.
  • 4
    Ferenczy A, Choukroun D, Falcone T, Franco E. The effect of cervical loop electrosurgical excision on subsequent pregnancy outcome. Am J Obstet Gynecol 1995;172:124650.
  • 5
    Keijser K, Kenemans P, van der Zanden P, Schijf C, Vooijs G, Rolland R. Diathermy loop excision in the management of cervical intraepithelial neoplasia: diagnosis and treatment in one procedure. Am J Obstet Gynecol 1992;166:12817.
  • 6
    Paraskevaidis E, Koliopoulos G, Lolis E, Papanikou E, Malamou-Mitsi V, Agnantis N. Delivery outcomes following loop electrosurgical excision procedure for microinvasive (FIGO stage IA1) cervical cancer. Gynecol Oncol 2002;86:103.
  • 7
    El-Bastawissi A, Becker T, Daling J. Effect of cervical carcinoma in situ and its management on pregnancy outcome. Obstet Gynecol 1999;93:20712.
  • 8
    Nohr B, Tabor A, Frederiksen K, Kjaer S. Loop electrosurgical excision of the cervix and the subsequent risk of preterm delivery. Acta Obstet Gynecol Scand 2007;86:596603.
  • 9
    Sjoborg K, Vistad I, Myhr S, Svenningsen R, Herzog C, Kloster-Jensen A, et al. Pregnancy outcome after cervical cone excision: a case–control study. Acta Obstet Gynecol Scand 2007;86:4238.
  • 10
    Samson S, Bentley J, Fahey T, McKay D, Gill G. The effect of loop electrosurgical excision procedure on future pregnancy outcome. Obstet Gynecol 2005;105:32532.
  • 11
    Arbyn M, Kyrgiou M, Simoens C, Raifu AO, Koliopoulos G, Martin-Hirsch P, et al. Perinatal mortality and other severe adverse pregnancy outcomes associated with treatment of cervical intraepithelial neoplasia: meta-analysis. BMJ 2008;337:a128410.1136.
  • 12
    Noehr B, Jensen A, Frederiksen K, Tabor A, Kjaer SK. Loop electrosurgical excision of the cervix and subsequent risk for spontaneous preterm delivery: a population-based study of singleton deliveries during a 9-year period. Am J Obstet Gynecol 2009;201:16.
  • 13
    Jakobsson M, Gissler M, Sainio S, Paavonen J, Tapper A. Preterm delivery after surgical treatment for cervical intraepithelial neoplasia. Obstet Gynecol 2007;109:30913.
  • 14
    Langhoff-Roos J, Kesmodel U, Jacobsson B, Rasmussen S, Vogel I. Spontaneous preterm delivery in primiparous women at low risk in Denmark: population based study. BMJ 2007;332:9379.
  • 15
    Stata Stastical Software, Release 9.0. Collage Station, USA, 2005.
  • 16
    Raio L, Ghezzi F, Di Naro E, Gomez R, Luscher K. Duration of pregnancy after carbon dioxide laser conization of the cervix: influence of cone height. Obstet Gynecol 1997;90:97882.
  • 17
    Sadler L, Saftlas A, Wang W, Exeter M, Whittaker J, McCowan L. Treatment for cervical intraepithelial neoplasia and risk of preterm delivery. JAMA 2004;291:21006.
  • 18
    Acharya G, Kjeldberg I, Hansen S, Sorheim N, Jacobsen B, Maltau J. Pregnancy outcome after loop electrosurgical excision procedure for the management of cervical intraepithelial neoplasia. Arch Gynecol Obstet 2005;272:10912.
  • 19
    Hein M, Petersen A, Helmig R, Uldbjerg N, Reinholdt J. Immunoglobulin levels and phagocytes in the cervical mucus plug at term of pregnancy. Acta Obstet Gynecol Scand 2005;84:73442.
  • 20
    Moscicki A, Shiboski S, Hills N, Powell K, Jay N, Hanson E, et al. Regression of low-grade squamous intra-epithelial lesions in young women. Lancet 2004;364:167883.