Archives of Gynecology and Obstetrics

, Volume 289, Issue 1, pp 85–99

Pregnancy outcome following loop electrosurgical excision procedure (LEEP) a systematic review and meta-analysis

Maternal-Fetal Medicine

DOI: 10.1007/s00404-013-2955-0

Cite this article as:
Jin, G., LanLan, Z., Li, C. et al. Arch Gynecol Obstet (2014) 289: 85. doi:10.1007/s00404-013-2955-0

Abstract

Objective

This study aimed at assessing the association of the relative risk (RR) of adverse pregnancy outcomes with previous treatment of loop electrosurgical excision procedure (LEEP) for the management of cervical intraepithelial neoplasia (CIN).

Methods

Data sources were from MEDLINE, EMBASE, and SCI citation tracking. Selection criteria: The eligible studies had data on pregnancy outcomes of women with or without previous treatment for CIN. Considered outcomes were severe preterm delivery (<34/32 weeks), extreme preterm delivery (<28 weeks), low birth weight (<2,500 g), stillbirth, preterm spontaneous rupture of membranes, perinatal mortality, and neonatal mortality and induction.

Results

36,954 cases and 1,794,174 controls in 4 prospective cohort and 22 retrospective studies were included in this meta-analysis. LEEP was associated with a higher risk of severe preterm delivery (<32 weeks, relative risk 1.98, 95 % CI [1.31, 2.98] 159/11,337 vs. 7,830/860,883), extreme preterm delivery (<28 weeks, RR, 2.33, 95 % CI [1.84, 2.94] 97/9,611 vs. 1,559/618,332), preterm premature rupture of the membranes (RR, 1.88, 95 % CI [1.54, 2.29] 126/2,837 vs. 7,899/313,094), and low birth weight (<2,500 g, RR, 2.48, 95 % CI [1.75, 3.51] 110/1,451 vs. 55/1,742). A cervical length of less than 3 cm was significantly increased in LEEP as compared with that of control group (RR, 4.88, 95 % CI [1.56, 15.25]), but increasing LEEP volume or depth was not associated with an increased rate of preterm birth <37 weeks. And LEEP was not associated with a significantly increased risk of perinatal mortality, cesarean section, stillbirth mortality, neonatal mortality, induction, and neonatal intensive care unit admission.

Conclusions

LEEP is associated with an increased risk of subsequent preterm delivery (<32/34, <28 weeks) and other serious pregnancy outcomes. But increasing LEEP volume or depth is not associated with an increased rate of preterm birth.

Keywords

Cervical intraepithelial neoplasia (CIN) Cesarean section (CS) Cervical volume Large loop excision of the transformation zone (LLETZ) Loop electrosurgical excision procedure (LEEP) Lower birth weight (LBW) Preterm delivery (PD) Perinatal mortality (PM) Pregnancy outcome Preterm spontaneous rupture of membranes (PPROM) Stillbirth mortality 

Introduction

More than 1 million women each year are diagnosed with grade 1 cervical intraepithelial neoplasia (CIN-1), and 500,000 would be found to have high-grade cervical cancer precursor lesions, referred to as CIN-2 and CIN-3, as indicated by laboratory surveys from the College of American Pathologists (CAP) [1]. Managing preinvasive cervical neoplasia has been a substantial public health burden because of the wide prevalence of infection of human papillomavirus, a virus from the papillomavirus family capable of infecting humans which can cause precancerous lesions and invasive cancer [2]. Women diagnosed with CIN are usually at their reproductive age, at a mean age of about 30 years, and even at a much younger age including adolescents [3, 4, 5]. Treatment for CIN must be a safe and effective procedure for evaluation and treatment of high-grade squamous intraepithelial lesions (HSIL).

Loop electrosurgical excisional procedure (LEEP), also known as large loop excision of the transformation zone (LLETZ) or LLETZ-conisation procedure, is a method of management for women with an abnormal cervical smear, which offers a deep excision of the transformation zone with minimal tissue damage. LEEP has also become the most popular method of treatments for CIN, because it is performed as a new diagnostic-treatment procedure for patients with cervical cytological diagnoses consistent with an epithelial abnormality [6]. Moreover, its clinical advantages include its ability to examine the margins of the extirpated transformation zone and to allow a comprehensive histological investigation of the extirpated tissue and entire transformation zone with the precise assessment of excision margins.

Although LEEP is by far the most popular procedure [7], it has its adverse effects on pregnancy outcomes. For instance, Crane in 2003 [8] reviewed five studies which compared women treated with LEEP with the untreated control subjects and concluded that women treated LEEP were more likely to have a preterm delivery in subsequent pregnancies (OR 1.8; 95 % CI [1.18,2.76]) and more likely to have low birth weight infants (<2,500 g; OR 1.60; 95 % CI 1.01–2.52) than the women untreated with LEEP were. A widely cited meta-analysis found the same evidence [9]. However, another meta-analysis published in 2008 [9] examined data from eight studies with 3,600 women treated with LEEP and reported no increase in the risk of preterm delivery at 32 or 34 weeks of gestation (RR 1.2; 95 % CI 0.5–2.9) and perinatal mortality rates (RR 1.17; 95 % CI 0.74–1.87) with huge heterogeneity. Despite the lack of a demonstrable increase in the risk of preterm delivery at any gestational age, the authors still cautioned that LEEP cannot be considered completely free of adverse pregnancy outcomes.

As the currently existing data for pregnancy outcomes are conflicting, we need a new more comprehensive and systematic review on more original studies and a meta-analysis with a focus on preterm delivery and more impact factors like excision depth/length that previous reviews have been unable to analyze. The objective of this meta-analysis was to assess the potential association of the relative risk (RR) of adverse pregnancy outcomes with previous treatment with LEEP for the management of CIN.

Methods

Types of studies, interventions, and inclusion and exclusion criteria

We included all the case-control studies that compared obstetric outcomes in pregnant women with or without LEEP treatment published in English between January 1980 and October 2012. The considered intervention was the treatments for CIN or cervical cancer with LEEP. The studies that used LEEP as the only treatment method were included in this mega analysis. In addition, we also included several studies which did not include control groups and the studies which were self-matched [10, 11, 12]. We excluded those studies in which more than one treatment methods were used but the outcomes for each treatment method were not reported separately. Furthermore, we also excluded those studies in which treatment was undertaken during the period of pregnancy.

Measures of pregnancy outcomes

The measures of pregnancy outcomes were related to either maternal or fetal outcomes or both. Maternal outcomes included preterm delivery (<37, <32/34, and <28 week), cesarean section rates, and preterm spontaneous rupture of membranes (PPROM). Fetal outcomes included low birth weight (<2,500, and <1,500 g), perinatal mortality, and neonatal intensive care unit (NICU) admission. The adverse obstetric or neonatal events were perinatal mortality, PPROM, and stillbirth mortality.

Retrieval of related eligible studies and data extraction

The eligible studies published between 1960 and November 2012 were retrieved with the four keywords including loop electrosurgical excision procedure (LEEP OR LETZ), loop excision, large loop excision of the transformation zone (LLETZ), and pregnancy. There was no language restriction. Two authors (Gong and Zhang) first verified inclusion and exclusion criteria independently and reached consensus. The results were then compared and a consensus was reached, with the involvement of a third investigator (Zhang) if necessary in case of discrepancies or disagreements. For every selected study, we retrieved the number of cases and controls in which every outcome was observed. Pooled relative risk (RR) and weighted mean difference (WMD) were used to report binary and continuous outcomes, respectively. We used fixed-effects models when there was no significant heterogeneity and/or random-effect models or when significant heterogeneity was present. RRs and 95 % CIs were calculated with the Revman 5.0 software. Heterogeneity among studies for every outcome was assessed by using a Cochrane Q test and the risk of publication bias was assessed by using funnel plots.

Result

Four (4) prospective cohort studies and 22 retrospective studies that fulfilled the inclusion criteria were included in this mega-analysis; three (3) of 26 studies were the multicenter cohort studies. A blind, sham-surgery controlled, and randomized trial cannot be done because patients cannot be randomly assigned to a sham procedure. Table 1 summarizes the characteristics of these studies. Table 2 shows the total effects of the meta-analysis for every outcome.
Table 1

Characteristics of the included studies

Source

Participants

Study design

Study period

Match/adjust

Inclusion/exclusion criteria

Pregnancy Outcomes

Assessment of study quality

Treated

Untreated

Blomfied et al. [30]a

40

80

Retrospective case control study

1989–1992

Matching for age, parity and ethnic group

UK

PD. PPROM, LBW

Low

Braet et al. [29]b

78

78

Retrospective cohort study

1988–1992

Matching for age, parity, smoking

UK

PD, PM

Low

Cruickshank et al. [14]

149

298

Retrospective matched cohort

1989–1991

Matching for age, parity, smoking, height, social class

Only first singleton pregnancies. Gestation ≥20 weeks

PD <28 weeks

Medium

Ricciotti et al. [12]b

29 (after)

29 (before)

Prospective observational cohort trial

1994–1995

UK

Patients with grade IIor grade IIICIN

UK

Medium

Gentry et al. [11]

20

20

Prospective study

1997–1998

UK

Patients being treated in the colposcopy, clinic the outpatient surgery center

UK

High

Paraskevaid et al. [44]

28

28

Case-control study

1990–1998

Matching for smoking, age, multiple pregnancies, history of previous preterm deliveries and parity

All women had cytology and colposcopy every 4 months during the first two post-treatment years and thereafter every 6 months indefinitely

PD, LBW, CS

Medium

Tan et al. [45]

119

119

Retrospective cohort study

1995–1998

Matching for age, parity

Only first pregnancies following treatment

Miscarriages, PD, gestation at delivery

Low

Sadler et al. [27]

278

256–426

Retrospective cohort study

1988–1999

Adjusted: age, ethnicity, smoking, history of preterm delivery, inter-hospital transfer

A singleton pregnancy to at least 20 completed weeks’gestation, and delivered or had post-partum care at the study hospital

PD, PPROM

High

Acharya et al. [26]

79

158

Retrospective matched cohort

1995–2000

Matching for age, parity, smoking, date of delivery, previous obstetric history

Age <45 years, gestation >20 weeks, only first pregnancies. Ectopic pregnancies excluded

PM

High

Samson et al. [16]

571

571

Retrospective matched cohort

1992–1999

Matching for age, parity, smoking, and date and place of delivery

Age <45 years, gestation >20 weeks, only first pregnancies

PM, PD <34 weeks

High

Crane et al. [37]

75

81

Prospective cohort

2001–2004

Adjustment for age, parity, smoking, third trimester bleeding by logistic regression

Only singleton pregnancies. Women with known risk factors for PD (previous PD, PD for maternal or fetal reasons) excluded

PM, PD <34 weeks

Medium

Bruinsma et al. [15]

69

2,294

Retrospective population based cohort

1982–2000

Obstetric antecedents, illicit drug use during pregnancy, major maternal medical condition, birth, being single, age, referral cytology were significant covariates in logistic regression model for PD. Parity and country of birth were insignificant

Gestation ≥20 weeks or > 400 g

PM, PD <32 weeks, <28 weeks, LBW <1,500, <1,000 g

High

Jakobsson et al. [34]

2,690

472,533

Retrospective population based cohort

1997–2004

Age, parity and smoking were adjusted for by logistic regression

Age 15–49 years. Only singleton pregnancies

PM, PD <32, <28 weeks

High

Jakobsson et al. [10]

625

NC

Retrospective cohort study

1997–2003

Adjusting for maternal age, parity, and maternal smoking

Patients have an outpatient treatment for CIN and a subsequent delivery

PD

High

Noehr et al. [46]

8,180

544,498

Retrospective matched cohort

1997–2005

Adjusted: age, smoking, previous pregnancy

Singleton deliveries with 21–45 completed weeks of gestation

PD

High

Werner et al. [28]

1,353

240,348

Retrospective cohort study

1992–2008

Age, race, and null parity

Singleton births, outpatient clinic settings

PD, LBW, PM

High

Ortoft et al. [24]

572

72,899

Population-based cohort study

1989–2007

Mother’s age, smoking status, and parity

Only singleton deliveries with valid gestational ages and birth weights

PD

High

Fischer et al. [17]

85

85

Prospective observational cohort trial

2001–2007

Matched: age, race, gestational age, prior vaginal deliveries

Pregnant women with a history of a cone biopsy or LEEP

Proportion of cervical lengths, CS, PD

Low

Paavonen et al. [47]a

20,011

430,975

Retrospective register-based study

1997–2009

Matched for Maternal age, parity, socio -economic status, or marital status, or history of previous preterm birth

Women who had LEEP and a subsequent delivery

PD

High

Lima et al. [22]

18

58

Retrospective case-control study

2000–2005

Matched for each case studied

Patients have subsequent delivery

LBW, PD

Low

Andia et al. [23]

378

189

Retrospective, case–control, multicenter study

1998–2007

Matched for age, parity

Age ranged from 20 to 43 years, and parity from 0 to 5

LBW, PD

Medium

Stout et al. [18]

616

1,839

Retrospective cohort study

UK

Adjusting for confounding variables

Women who underwent Pap screening and did or did not undergo LEEP

PD

High

Poon et al. [19]

473

25,772

Multicenter cohort study

1998–2006

Adjusting for racial origin, smoking, parity and previous PD

Women with major fetal abnormalities, painful regular uterine contractions, history of ruptured membranes or cervical cerclage in situ were excluded

PD

Medium

Macones et al. [20]

625

602

Multicenter cohort study

2000–2006

Age, ethnicity, BMI, smoking, and prior PD

Reproductive-aged women who underwent LEEP, PAP smear, or cervical punch biopsy without LEEP

PD

High

Khalid et al. [33]b

321

NC

Retrospective observational study

1999–2002

None

Case records and histology reports for eligible women

PD and miscarriage rate (<24 weeks of gestation)

Low

Simoens et al. [21]

97

194

Multicenter cohort study

2008–2010

Ethnicity, HIV status, education, age, smoking and parity

A history of previous CIN treatment

PD

Medium

PD preterm delivery, PM perinatal mortality, LBW low birth weight, CIN cervical intraepithelial neoplasia, CS cesarean section, PPROM preterm spontaneous rupture of membranes, NC no control group, UK unknown

aWe cannot get full text, abstract only

bTreated = after treatment, untreated = before treatment

Table 2

Total effect of the meta-analysis for every outcome and treatment method used

Outcome

Studies (participants)

Positive case (total cases)

Positive controls (total controls)

Pooled relative risk [95 % CI]

Heterogeneity I2 (P)

Preterm delivery <32/34 weeks

8 (872,220)

159 (11,337)

7,830 (860,883)

1.98 [1.31, 2.98]

65 % (0.001)

1.68 [1.16, 2.41]a

25 % (0.24)a

Preterm delivery <28 weeks

5 (627,993)

97 (9,611)

1,559 (618,382)

2.33 [1.84, 2.94]

0 % (<0.00001)

Stillbirth

2 (242,148)

6 (1,502)

849 (240,646)

0.98 [0.43, 2.22]

0 % (0.71)

PPROM

5 (315,931)

126 (2,837)

7,899 (313,094)

1.88 [1.54, 2.29]

79 % (< 0.0007)

2.91 [2.20, 3.86]a

0 % (0.75)a

Cervical length <3 cm

2 (326)

18 (160)

4 (166)

4.88 [1.56, 15.25]

2 % (0.31)

Introduction

3 (1,535)

177 (725)

170 (810)

1.12 [0.88, 1.43]

65 % (0.06)

1.17 [0.96, 1.42]a

3 % (0.31)a

Cesarean delivery

9 (3,007)

334 (1,435)

332 (1,572)

1.12 [0.97, 1.29]

60 % (0.010)

Low birth weight <2,500 g

8 (3,193)

110 (1,452)

55 (1,742)

2.48 [1.75, 3.51]

26 % (0.22)

Low birth weight <1,500 g

2 (2,929)

8 (447)

41 (2,482)

2.27 [0.53, 9.76]

42 % (0.19)

NICU admission

3 (1,354)

90 (674)

74 (680)

1.26 [0.91, 1.75]

0 % (0.72)

Perinatal mortality

9 (761,152)

51 (5,527)

6,749 (755,625)

1.04 [0.78, 1.39]

13 % (0.32)

Neonatal mortality

2 (242,853)

17 (1,934)

854 (240,919)

1.25 [0.68, 2.32]

0 % (0.56)

aSensitivity analysis

Preterm delivery, stillbirth, and PPROM

We identified 19 studies that fulfilled the eligibility criteria and provided data on preterm delivery [10, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30]. We add 11 studies besides the 8 studies of Kyrgiou in 2006 [31].

LEEP was significantly associated with PD <37 weeks (RR 1.84, 95 % CI [1.56, 2.18], 1,446/16,759 (8.6 %) vs. 115,935/2,508,445 (4.6 %)). Because of significant heterogeneity between procedures (P = 0.031), we have not observed an overall pooled RR. The risk of PD <32/34 weeks was significantly increased in women treated with LEEP [RR 1.98, 95 % CI (1.31, 2.98), 159/11,137 (1.4 %) vs. 7,830/860,883 (0.9 %), I2 = 65 % (Fig. 1a)]. Noehr et al. [13] performed a large scale population-based study. Omission of this study still yielded a pooled RR that was not significantly different from unity (RR 2.33,95 % CI [1.84, 2.94], I2 = 25 %). The risk of PD <28 weeks also was significantly increased in women treated with LEEP (RR 1.68, 95 % CI [1.16, 2.41], 97/9,611 (1.0 %) vs. 1,559/618,382 (0.3 %), I2 = 0 % (Fig. 1b).
Fig. 1

Forest plot showing association between LEEP and a preterm delivery <32/34 weeks and b preterm delivery <28 weeks

However, LEEP was not associated with an increased risk of stillbirth [RR 0.98, 95 % CI (0.43, 2.22) 6/1,502 (0.4 %) vs. 849/240,646 (0.4 %), I² = 0 %, (P = 0.97)] or PPROM (RR 2.91, 95 % CI (2.20, 3.86), 126/2,837 (4 %) vs. 7,899/313,094 (2.5 %) I² = 0 %, (P < 0.00001)] (Fig. 2a, b).
Fig. 2

Forest plot showing association between LEEP and a stillbirth, b PPROM

Cervical length and excision volume/depth

Figure 3 showed that there was a significant increase in the existence of cervical length less than 3 cm in the LEEP groups as compared with that of the control groups. It was shown in two studies that a cervical length that was <3.0 cm was found to be more frequently present in the LEEP groups than in the control groups (pooled RR 4.88, 95 % CI [1.56, 15.25], 18/160 (11.2 %) vs. 4/166 (2.4 %), I2 = 2 %). The results were not heterogeneous. Usually, two studies did not identify a raised risk. However, these two studies did show a significant association of the frequent presence of a cervical length <3.0 cm with LEEP treatment. Further analysis of the dimensions of the excised tissues was also undertaken for those studies that reported these data Fig. 4a showed that the mean cervical length was significantly shorter in the LEEP group (WMD: −0.62, 95 % CI [−0.78, 0.46], I2 = 0 %) than that in the control group. Moreover, Fig. 4b displayed the mean cervical lengths as measured by transvaginal ultrasonography before and after LEEP treatment (WMD: 0.33, 95 % CI [−0.28, 0.95], I2 = 56 %), respectively.
Fig. 3

Forest plot showing the presence of a cervical length <3 cm

Fig. 4

Cervical length in LEEP and control group, excision volume and depth in LEEP

Figure 4c, d showed in both studies [32, 33] (case groups were preterm, control groups were term) that increasing LEEP volume or depth caused not difference between term and preterm deliveries. A recent study by Macones et al. [32] reported no association between LEEP volume and preterm birth (OR 1.01, 95 % CI [0.95–1.08]) and no association between LEEP depth with preterm birth (OR 1.01, 95 % CI [0.67–1.54]). In contrast, Khalid et al. [33] showed that there was a threefold increase in the risk of PD when the excision volume of LEEP exceeded 6 cm3 (RR = 3.00; 95 % CI [1.45–5.92]), or when the thickness of the excised tissue was >12 mm (RR = 2.98; 95 % CI [1.27–7.01]). Besides, they concluded that the thickness and volume of LEEP specimens could predict the RR of pregnancy-related morbidity.

It has been shown [26, 33] that when the size of the diathermy loop used for the procedure was 25 mm, the risk of preterm birth (OR 4.0; 95 % CI [1.0–16.0]; P = 0.05) and low birth weight birth (OR 14.0; 95 % CI [1.7–114]; P = 0.01) caused by LEEP treatment was significantly higher than that of the control group. The results also indicated a higher risk of obstetrical problems associated with the use of a relatively larger loop (>25 mm), although the confidence intervals were very wide.

Perinatal mortality, induction, and neonatal mortality

Figure 5a, b, c shows the RR for perinatal mortality, induction, and neonatal mortality associated with LEEP. We add two studies in our comparison with Crane’s study conducted in 2003 [8]. We identified eight studies that fulfilled the eligibility criteria and provided data on perinatal mortality [15, 16, 24, 25, 26, 28, 30, 34] (RR 1.04; 95 % CI [0.78, 1.39], 51/5,527 (0.92 %) vs. 6,749/755,625 (0.89 %), I² = 13 %, P = 0.79), neonatal mortality (RR 1.25; 95 % CI [0.68, 2.32] 17/1,934 (0.9 %) vs. 854/240,919 (0.4 %), I² = 0 %, P = 0.47), induction (RR 1.17; 95 % CI [0.96, 1.42], 177/725 (24.4 %) vs. 170/810 (21.0 %) I² = 3 %, P = 0.11). The pooled results did not show any heterogeneity between procedures. Thus, LEEP is not considered to be significantly associated with perinatal mortality, induction, or neonatal mortality.
Fig. 5

Forest plot showing association between LLETZ and a perinatal mortality, b induction, c neonatal mortality

Low birth weight (<2,500/<1,500 g)

Eight studies [13, 14, 16, 21, 22, 23, 25, 26] showed birth weight (<2,500 g, RR 2.48, 95 % CI [1.75, 3.51], I² = 26 %, P < 0.00001). Only two studies reported low birth weight <1,500 g ([15, 23]). Andia et al. in 2011 showed a RR of 7.25, but Bruinsma et al. showed a RR of 0.81. This yielded a pooled RR of 2.27, 95 % CI [0.53, 9.76], I2 = 42 %, and P = 0. 27 (Fig. 6a, b). The results also indicated a higher risk of obstetrical problems associated with LEEP, although dispute existed in birth weight <1,500 g.
Fig. 6

Forest plot showing association between LEEP and a birth weight <2,500 g, b birth weight <1,500 g

Others outcomes

We found no significant difference with respect to cesarean section or NICU admission. Each individual comparison did not show inter-study heterogeneity. NICU admission showed a RR of 1.26; 95 % CI [0.91, 1.75], 90/674 vs. 74/680 (P = 0.17, I² = 0 %) and cesarean section showed a pooled RR of 1.12, 95 % CI [0.97, 1.29], 334/1,435 vs. 332/1,572, (P = 0.13, I2 = 60 %) (Fig. 7a, b).
Fig. 7

Forest plot showing association between LEEP and a NICU admission, b cesarean delivery

Publication bias

The funnel plots appeared to be symmetrical for PM (Fig. 8b) and LBW (Fig. 8c). However, the funnel plot suggested that there may be publication bias in the identified trials for PD <32 weeks (Fig. 8a).
Fig. 8

Funnel plots of risk ratios for overall PD, PM, LBW

Discussion

CIN is the potentially premalignant transformation and abnormal growth (dysplasia) of squamous cells on the surface of the cervix. Some patients with CIN have the risk of suffering from cervical cancer, usually cervical squamous cell carcinoma (SCC). Currently, LEEP is the most commonly used method for CIN treatment. However, LEEP treatment could cause a number of pregnancy outcomes. In this study, we conducted a mega-analysis to assess the association of RR of adverse pregnancy outcomes with previous treatment of LEEP for the management of CIN. With respect to pregnancy outcomes, our meta-analysis has established a new conclusion that LEEP was associated with an increased risk of preterm delivery <28 or <32/34 weeks. We also found that while depth and volume of LEFP were not associated with PD, there was a significant increase in the presence of the cervical length <3 cm o in LEEP group compared with that of the control group. Small sample size may fail to detect significant differences and the chance of a type 2 error can be reduced by the combining of these existing studies. In our study, for all pregnancy outcomes apart from PD, the comparisons showed no evidence of inter-study heterogeneity, confirming the consistent direction and magnitude of effect.

Several new original studies and reviews on pregnancy outcomes after LEEP treatment have recently been published. The availability of these new data increased the sample size and statistical power and thus, enabled us, for the first time, to address the rarer but more serious pregnancy outcomes such as severe preterm delivery (<32/34 weeks), and extreme preterm delivery (<28 weeks), stillbirth, induction, impact of cervical length on preterm. Our results of this mega-analysis on these new data were totally different from those reported in the widely cited meta-analysis conducted by Arbyn et al. [9]. We found that LEEP treatment has a considerable impact not only on the outcomes of pregnant mothers and their babies concerned but also on the health budgets for neonatal intensive care.

Kyrgiou et al. [31] reported that the RR of preterm birth based on eight reports related to LEEP was 1.70 [1.24–2.35]. However, the RRs were close to 1.0 in five of the eight studies. Thus, the preponderance of individual studies in this meta-analysis suggested that LEEP was not associated with preterm birth. In our study, we observed that PD <37 weeks showed huge heterogeneity. As previously discussed, the causes of preterm birth are complex and incompletely understood, and the increased rate of PD that has been reported by many investigators may be a result of some other, as yet unidentified, factors. An observational study of over 30,000 singleton pregnancies has demonstrated that the risk of PD was inversely related to the gestation at previous spontaneous delivery [35]. Racial differences in the rate of spontaneous preterm birth have also been observed, with the reported risk being higher in African and South Asian women than in Caucasian women [36]. At the same time, there are several explanations for PD. One potential explanation for any noted relationship is the association of both conditions with genital tract colonization and infection. However, Stout et al. [18] showed that active vaginal infection during pregnancy did not amplify risk for preterm delivery and LEEP did not independently increase the risk of PD.

Our mega analysis suggests that the effect of LEEP on preterm delivery is mediated or is reflected in shortening cervical length. Our analysis indicated that the cervical length <3 was more frequently seen in LEEP group and in the control group. The suggested mechanisms for the pathogenesis of preterm delivery in women with a history of LEEP because the shortening cervical and impairment of local immunological defense mechanisms after removal of the cervical glands [37, 38, 39]. And then, the fact that the removal or destruction of part of the cervix will affect its function, especially during subsequent pregnancies, is unsurprising. One would logically expect that the more cervical tissue that is removed, the greater effect on function there might be. However, our study did not show this. The results of our study, suggest that the proportion of the total cervical volume or endocervical canal removed might be more important than the actual depth of excision by LEEP.

At the same time, we also changed the pregnancy outcome compared with those in the 2003 and 2008 studies [8, 9], LBW increased 2.37 [1.70, 3.30], which was higher than that reported in the 2003 study (OR 1.60, 95 % CI [1.01, 2.52], P = 0.04). More interesting, only two studies showed LBW <1,500 g, 7/378 vs. 0/189, RR 7.52, 95 % CI [0.43, 130.97]. However,the meta-analysis in the 2003 and 2008 studies failed to find the increasing risk level.

It should be noted that our study did not have enough power to confirm the increasing tendency of the risk of perinatal mortality. Thus, we need more data on stillbirth and neonatal mortality, though current data did not show any increase.

Finally, any meta-analysis may be subject to publication bias. If more studies with striking results are published, new interventions might be mistakenly assumed to be effective at publication time of this meta-analysis.

Practical implications

Whether there is a critical threshold in the amount of tissue excised or destroyed that determines obstetric morbidity and the success of treatment in terms of recurrent CIN or cancer are key questions that remain to be answered. Having a clear understanding of this relation would be useful in guiding clinical decision-making.

Three recent studies have shown that women threated with LEEP are still at higher risk than the general population in developing subsequent invasive cervical cancer, even many years after treatment, which can help with the follow-up of women after treatment [40, 41, 42]. Some gynecologists warned that the less aggressive treatments might also increase these risks [43]. Practitioners should provide counseling to their patients regarding the potential for future obstetric complications and discuss the timing and appropriate prenatal considerations to maximize outcomes in future.

Conclusion

Evidence obtained from this study clearly indicates that LEEP is associated with a small but real increase in risk of PD <37 weeks and that the risk of PD <32/34 and <28 weeks is significantly increased in women treated with LEEP. Additionally, removing large amounts of cervical tissue by LEEP probably does not differ between term and preterm delivery. Moreover, the precise conditions of the safety of using LEEP have been insufficiently known and require further research. Nevertheless, this information should be considered when women are counseled before their consent to treatment and lends support to the philosophy of the doctors.

Conflict of interest

We declare that we have no conflict of interest.

Copyright information

© Springer-Verlag Berlin Heidelberg 2013

Authors and Affiliations

  1. 1.West China Second HospitalSichuan UniversityChengduPeople’s Republic of China
  2. 2.West China HospitalSichuan UniversityChengduPeople’s Republic of China

Personalised recommendations