To review our experience with a screening programme that included four sequential cervical length (CL) measurements from 16 to 22 weeks of gestation.
To review our experience with a screening programme that included four sequential cervical length (CL) measurements from 16 to 22 weeks of gestation.
Historical cohort study.
Tertiary-care centre in a university hospital.
There were 312 singleton pregnancies in 321 women with a previous large loop excision of the transformation zone (LLETZ), and 62 pregnancies after a second-trimester miscarriage in a previous pregnancy.
The CL measurements were performed by transvaginal ultrasound at 16, 18, 20, and 22 completed weeks of gestation.
Early preterm delivery before 34 completed weeks of gestation.
Early preterm delivery was found in 7.4%. The CL at 16 completed weeks of gestation was smaller in the LLETZ group (36 mm, interquartile range 30–40 mm) compared with the control group (38 mm, interquartile range 32–42 mm; P = 0.040). For the analysis of risk factors for early preterm delivery after LLETZ, only cases with a complete data set were included (n = 145). In a multivariate analysis, two parameters remained significantly predictive, with CL at 16 completed weeks of gestation being the most significant measure (P < 0.001, OR 0.90, 95% CI 0.83–0.98), followed by conception using IVF treatment (P = 0.031, OR 0.64, 95% CI 1.54–34.80).
Even as early as 16 weeks of gestation, women with early preterm delivery reveal a significantly lower CL than those without. Dynamics in the CL do not add to this information.
Large loop excision of the transformational zone (LLETZ) is the most frequently used method of treatment for the ablation of cervical intraepithelial neoplasia.[1, 2] The advantages of this procedure include simplicity and cost-effectiveness; however, pregnancy-related morbidity associated with the procedure has also been reported. Two recent large studies proved that LLETZ significantly increases the risk for spontaneous preterm delivery by about two- to three-fold.[5, 6] Both studies were well-matched for confounding factors. These results are in accordance with previous studies that have been pooled in a meta-analysis, who demonstrated a relative risk of 1.7 for preterm delivery, and 2.7 for preterm prelabour rupture of membranes.
Although there is also evidence suggesting the opposite,[5, 6] an at least slightly increased risk for preterm delivery is currently assumed for pregnancies after LLETZ. This displays the need for an accurate predictive parameter identifying those at risk for preterm delivery after LLETZ. It has already been shown that spontaneous early delivery can effectively be predicted based on a combination of maternal characteristics and the ultrasonographic measurement of cervical length at 20–24 weeks of gestation.[8-11] The predictive value of cervical length measurement has recently been verified for women who have undergone LLETZ.
For several years, sequential cervical length screening from 16 to 22 weeks of gestation has been implemented at the Department of Fetomaternal Medicine of the Medical University of Vienna for women at risk for preterm delivery. The main study objective was to evaluate the gestational age with the most predictive value for preterm delivery in our large patient population that included 345 pregnancies after LLETZ. Moreover, we aimed to compare the kinetics of cervical length in LLETZ pregnancies with women after a second-trimester miscarriage.
At the Department of Fetomaternal Medicine of the Medical University of Vienna, Austria, a screening programme for pregnant women at risk for preterm birth has been established for many years. Sequential cervical length measurements are hypothesised to be a more appropriate measure for the risk of preterm birth, as they also allow for an assessment of the dynamics of cervical length.
Women at risk for preterm birth also include those who had undergone LLETZ. There were 370 pregnancies after LLETZ that both underwent regular follow-up examinations, beginning with first-trimester screening, and were delivered at the Department of Fetomaternal Medicine of the Medical University of Vienna, Austria, from January 2003 to July 2012. Women who had been treated with cervical cerclage or prophylactic progesterone, women who had had a multiple pregnancy, and/or women who had had a preterm delivery in a previous pregnancy were excluded from this study. As the study population consists of a historical cohort, and prophylactic progesterone treatment was used routinely in accordance with recent local guidelines only in the last months of the study period, the application of this exclusion criterion is thought to have only a minor influence. According to the local guideline on women at risk for preterm delivery, patients were offered cerclage if the cervical length was <25 mm at 16 or 18 completed weeks of gestation (prophylactic cerclage), or if the cervical length was <20 mm at 20 completed weeks of gestation (therapeutic cerclage). Within the study period, 17 women had received a cerclage after LLETZ, all of them being prophylactic. As a control group, we included 60 women who underwent cervical length screening because of one late miscarriage in a previous pregnancy, and who had not undergone LLETZ.
The main outcome measure was preterm delivery as a result of cervical insufficiency (defined as painless cervical dilatation leading to second-trimester birth), preterm labour, or preterm prelabour rupture of membranes (see below). Hence, cases were also excluded for preterm delivery for reasons other than those associated with a previous LLETZ: namely, pre-eclampsia/eclampsia, HELLP (haemolysis, elevated liver enzymes and low platelet count) syndrome, and imminent intrauterine danger to the fetus' life. Accordingly, in a historical cohort study, 312 consecutive pregnancies in 294 women with a previous LLETZ were included. This study was approved by the Institutional Review Board (IRB) of the Medical University of Vienna (IRB number: 1400/2012).
All women were seen for the first time between 11 and 14 weeks of gestation in the course of first-trimester screening. The screening programme for pregnancies at risk for preterm delivery included the following: cervical length measurement by transvaginal ultrasound at 16, 18, 20, and 22 completed weeks of gestation, and, if considered necessary, further examinations after that. All ultrasound examinations for placental invasion screening were performed by highly experienced operators. Thirty-two sonographers were involved. Because of the ‘historical cohort’ design, only a minority of sonographers had a certificate; however, all cervical length measurements were carried out according to the guidelines of the Fetal Medicine Foundation (available online at http://www.fetalmedicine.com/fmf/training-certification/certificates-of-competence/cervical-assessment). Each transvaginal scan was performed over a period of about 3 minutes. The shortest of at least three measurements was documented. Cervical length measurements are subject to internal quality assurance. All patients were treated individually.
Data analysis was performed by retrospective chart review. The outcome was confirmed by notes review. In our department, the PIA Fetal Database software (GE-Viewpoint, Wessling, Germany) is used as the basic perinatologic database. The main outcome measure was preterm delivery as a result of cervical insufficiency, preterm labour, or preterm prelabour rupture of membranes. In accordance with a recent large study, we chose 34 completed weeks of gestation (early preterm delivery) as a clinically relevant cut-off for preterm delivery. In addition to cervical length measurements, the following parameters were included: number of previous LLETZ procedures; the age of the mother at delivery; body mass index (BMI) at the initial visit; number of fetuses; parity; previous preterm birth; previous second-trimester miscarriage; urinary tract infection during pregnancy; and cigarette smoking.
Nominal variables are reported as numbers and frequencies, and continuous variables are reported as medians and interquartile ranges (IQRs). Statistical analysis was accomplished using contingency tables that enabled the calculation of overall accuracy, sensitivity, specificity, and positive and negative predictive values, and Pearson's chi-square test or Fisher's exact test, where appropriate. Paired Student's t–tests were applied to test for differences between repeated measures, where appropriate. A logistic regression model with Wald's tests was used to test the statistical significance of all coefficients. Odds ratios (ORs) are given, including the 95% confidence intervals (95% CIs). The optimal cut-off for cervical length was calculated automatically, based on the receiver-operating characteristic (ROC) curve, as the threshold value with the highest specificity and sensitivity. The discriminatory ability of each parameter investigated is described as the correlation of specificity and sensitivity, and was measured by the area under the ROC curve (AUC). Where appropriate, values are given with 95% CIs.
Statistical analysis was performed using the open-source statistical package, ® 2.13.0, the mass package, the faraway package, and the memisc package. Differences were considered to be statistically significant if P < 0.05.
The detailed characteristics about the pregnancies included are provided in Table 1. In LLETZ patients, preterm delivery before 37 completed weeks of gestation was found in 54 pregnancies (17.3%), and preterm delivery before 34 completed weeks of gestation (early preterm delivery) was found in 23 cases (7.4%). In the control group (women after a previous second-trimester miscarriage), the rates were 30.0% (18/60) and 13.3% (8/60), respectively.
|LLETZ group||Control group||P|
|Number of pregnancies (n)||312||60||–|
|Age (years)||33 (20–50)a||30 (16–43)a||0.010|
|Body mass index before pregnancy (kg/m2)||21.9 (18.0–30.1)a||22.3 (18–49)a||0.662|
|Parity (n)||0 (0–4)a||0 (0–4)a||0.581|
|Pregnancies after IVF treatment||13 (4.2)b||0 (0)b||0.140|
|Pregnancies in smokers||63 (20.2)b||16 (26.7)b||0.261|
|Pregnancies in women with two previous LLETZ conisations||16 (5.1)b||–||–|
For the analysis of the course of cervical length and risk factors for early preterm delivery, only cases with a complete data set were included. This resulted in a sample of 145 pregnancies (46.5%). Figure 1 shows subsequent cervical length measurements in pregnancies after LLETZ (Figure 1A) and in the control group, i.e. pregnancies after a late miscarriage in a previous pregnancy (Figure 1B). The cervical length at 16 completed weeks of gestation was smaller in the LLETZ group (36 mm, IQR 30–40) compared with the control group (38 mm, IQR 32–42; P = 0.040).
Figure 1A shows subsequent cervical length measurements in pregnancies after LLETZ with an early preterm delivery (group LLETZ-A), and in pregnancies with delivery after 34 completed weeks of gestation (group LLETZ-B). Notably, a smaller mean cervical length was found for group B even at 16 completed weeks of gestation. In both LLETZ groups, cervical length was found to have decreased thereafter. In paired Student's t-tests, the difference in cervical lengths between 16 and 22 weeks of gestation was highly significant in women with and without ‘early preterm delivery’ (P < 0.001). This was not the case in the control group (Figure 1B). At 16 completed weeks of gestation, cervical length was 39 mm (IQR 34–42 mm) in women with delivery after 34 completed weeks of gestation (control-B), compared with 35 mm (IQR 29 – 41) in women with early preterm delivery (control-A; P = 0.129). However, from 16 to 22 weeks of gestation cervical length declined only in control-A cases (P = 0.016), and not in control-B cases (P = 0.060).
We then focused on predictive factors for early preterm delivery (Table 2). In a univariate analysis, several parameters differed significantly between groups A and B, including all four cervical length measurements; however, after multivariate analysis, two parameters remained significantly predictive, with cervical length at 16 completed weeks of gestation being the most significant (P < 0.001, odds ratio 0.90), followed by conception using IVF treatment (P = 0.031, OR 0.64). Notably, cervical length measurements at 18, 20, and 22 completed weeks of gestation were no longer significantly predictive in the multivariate model, and neither were the differences between these measurements.
|Delivery <34 weeks (n = 17)||Delivery ≥34 weeks (n = 128)||Univariate analysis||Multivariate analysis|
|OR (95% CI)||P||OR (95% CI)||P|
|Age (years)b||34 (30–37)||32 (29–37)||1.08 (0.96,1.23)||0.214||–||–|
|Body mass index (kg/m2)b||22.7 (20.7–25.0)||21.0 (19.8–24.4)||1.02 (0.89,1.18)||0.752||–||–|
|Pregnancy after IVF treatmentc||3 (17.6)||6 (4.7)||7.31 (1.54,34.80)||0.023d||7.31 (1.54,34.80)||0.031|
|Urinary tract infection during pregnancyc||0||5 (3.9)||0 (0)||0.994||–||–|
|Parityb||0 (0–1)||1 (0–1)||0.81 (0.39,1.65)||0.557||–||–|
|Number of previous second-trimester miscarriagesc||0 (0–1)||0 (0–1)||1.50 (0.9,2.5)||0.112||–||–|
|Cigarette smokingc||5 (29.5)||27 (21.1)||2.13 (0.58,7.84)||0.255||–||–|
|Number of previous conisationsa||1 (1)||1 (1)||1.53 (0.19,12.01)||0.686||–||–|
|Cervical length in week 16 (mm)b||27 (17–32)||37 (31–42)||0.90 (0.83,0.98)||0.009d||0.90 (0.83,0.98)||<0.001d|
|Cervical length in week 18 (mm)b||26 (13.5–35)||36 (31–42)||0.91 (0.84,0.98)||0.013d||1.03 (0.93,1.15)||0.528|
|Cervical length in week 20 (mm)b||24 (16.5–28)||35 (29–41)||0.87 (0.79,0.96)||<0.001d||0.98 (0.87,1.10)||0.757|
|Cervical length in week 22 (mm)b||20 (15–25)||34 (28–40)||0.85 (0.77,0.94)||<0.001d||1.21 (0.83, 1.94)||0.250|
|Cervical length: ∆ week 18–16 (mm)b , e||−1.0 (−3, +2)||−2.0 (−5, +3)||0.99 (0.96,1.02)||0.575||–||–|
|Cervical length: ∆ week 20–16 (mm)b , e||−3 (–7, −1)||−1 (−1, +3)||0.98 (0.95,1.02)||0.312||–||–|
|Cervical length: ∆ week 22–16 (mm)b , e||−6 (−8, −4)||−2 (−8, +1)||0.97 (0.94,1.00)||0.044||0.97 (0.94,1.00)||0.052|
As a next step, we calculated the optimised cut-off values for cervical length at 16 completed weeks of gestation. The optimised cut-off was ≤ 30 mm, resulting in specifity, sensitivity, and positive and negative predictive values of 83.6% (95% CI 76.0–89.5%), 76.5% (95% CI 50.1–93.2%), 38.2% (95% CI 22.2–65.4%), and 96.4% (95% CI 91.0–99.0%), respectively. The area under the curve was 0.73, with an odds ratio of 16.6 (95% CI 4.4–67.6). In all four cases with a cervical length >30 mm and a preterm delivery, women reached at least 27 completed weeks of gestation.
Two recent large studies in women with singleton pregnancies demonstrated preterm delivery rates of 3.4 and 6.9% in women with a previous LLETZ.[12, 13] Nevertheless, there are controversial reports that challenge the role of the surgical procedure itself and the obstetric outcome. Although there seems to be a strong correlation between LLETZ and prelabour rupture of membranes, low birthweight, and preterm delivery, the association with excision width (cone size) remains unclear.[1, 14, 15] In accordance with previous reports, there was a 7% rate of early preterm deliveries after LLETZ in our cohort.
In addition, we identified one other risk factor for preterm delivery apart from previous LLETZ in our study cohort, namely pregnancies after IVF treatment. Notably, in contrast to previous studies,[12, 13] parity and cigarette smoking were not among these risk factors.
It has already been demonstrated that cervical length at 20–24 weeks of gestation was an accurate predictive parameter for preterm delivery in both an unselected patient population,[8-11] and in women who had had a previous LLETZ. Our study, which included pregnancies after LLETZ, adds information about sequential cervical length measurements. Notably, in the univariate analysis, cervical length was found to differ between women with and without an early preterm delivery at all times of measurement (Table 2). This allows for a quite early first appraisal of a women's prognosis; however, when focusing on the optimised cut-off value at 16 completed weeks of gestation (≤ 30 mm), the individual risk cannot be predicted with a highly reliable overall accuracy at this point in time (positive predictive value, 38.2%; negative predictive value, 96.4%). Notably, in a multivariate analysis of predictive factors for preterm delivery, cervical length measurement remained independently significant only at 16 completed weeks of gestation. It can, however, be concluded that at both 16 and 22 completed weeks of gestation, cervical length, when analysed independently of other parameters, can be used for the assessment of the individual risk for a preterm delivery before 34 completed weeks of gestation.
Notably, the kinetics of cervical length measurements did not add any predictive information in the multivariate analysis. This can be explained by the fact that an increased risk for preterm delivery could already be seen at the time of the first measurement, i.e. at 16 weeks of gestation, which in addition remained the only significant cervical length measurement after multivariate analysis, and that cervical length was found to decrease significantly in both women with and without early preterm delivery. When sequential cervical length measurements between LLETZ and control patients are compared, it becomes evident that LLETZ patients at risk for early preterm delivery already display smaller cervical lengths. This suggests that, in pregnancies after LLETZ, a majority of the risk for early preterm delivery is attributable to a primary mechanical impairment caused by the operation itself, whereas in other women prone to early preterm delivery, a dynamic, accelerated shortening of the cervix seems to occur, as can be seen in Figure 1B.
One might argue that the study was biased by the exclusion of all patients who underwent cerclage, as sequential cervical length measurements were performed for interventional and not observational reasons. However, all excluded cases had undergone a prophylactic cerclage because of very short cervical lengths at 16 or 18 completed weeks of gestation. Hence, the subsequent cervical length dynamics should not have been influenced. Moreover, several women had refused the intervention, although cervical length was below the threshold for cerclage according to the local guidelines. Thus, we consider the impact of this exclusion criterion to be minor. Naturally, one will have to consider the retrospective design as a study limitation when interpreting our results.
t is important that obstetricians ensure that repeated cervical length measurements do not lower the threshold for the use of cerclage. In general, we discourage the use of prophylactic cerclage in women after LLETZ because of the reported non-beneficial effects and possible adverse events of this procedure.[16, 17] In accordance with prior reports, we believe that the proper application of cervical length measurements may prevent unnecessary cerclage.[18, 19]
To the best of our knowledge, we are the first to have evaluated the value of periodic cervical length measurements for the prediction of early preterm delivery in pregnancies after LLETZ. Although the study must be interpreted within the limit of its retrospective design, the following conclusions can be drawn: (i) the cervical length decreases in women both with and without early preterm delivery from 16 to 22 weeks of gestation, and this decline does not provide reliable information for the prediction of early preterm delivery; (ii) even at 16 weeks of gestation, women with early preterm delivery reveal significantly lower cervical lengths than those without. The optimised cut-off value at this gestational age is <30 mm.
The findings should encourage obstetricians to measure cervical length as early as 16 completed weeks of gestation. Studies are needed that evaluate the value of a prophylactic cerclage in pregnancies after LLETZ with a cervical length <30 mm at 16 completed weeks of gestation.
All authors declare that they have no commercial interest, financial interest, and/or another relationship with manufacturers of pharmaceuticals, laboratory supplies, and/or medical devices, or with commercial providers of medically related services.
All authors substantially contributed to the conception and design of the article. SP, CN, and JO were responsible for the acquisition of data. RS performed the statistical analyses. SP, WE, and JO drafted the article and revised it for intellectual content.
All authors gave final approval of the version to be published.
The study was approved by the ethics committee of the Medical University of Vienna, Vienna, Austria (IRB number 1400/2012).
Although emphasised by Pils et al., the use of cervical sonography to predict the risk of preterm birth or to identify candidates for treatment with cerclage or progesterone is not the principal contribution of cervical imaging. Rather, the importance of cervical ultrasound is that it has provided the opportunity to observe the parturitional process at the cervix and decidual–membrane interface. These observations have overturned previous paradigms that emphasised myometrial activity, membrane rupture, and early cervical effacement as independent pathways to preterm birth. The association of a short cervix with an increased risk of preterm birth was initially interpreted in that paradigm as evidence of a continuum of competence, to which LLETZ was a surgical contributor, and for which cerclage, a surgical cure, was prescribed. After more than a decade of unsuccessful attempts to find a combination of cervical length threshold and reproductive history that could distinguish ‘cervical insufficiency’ from ‘preterm labour’, the results of two landmark studies produced a revelation: a medical therapy, progesterone, was shown to be effective for a short cervix (Fonseca et al., N Engl J Med 2007;357:462–469), whereas surgical treatment helped women with the shortest cervical length, of less than 15 mm, where a greater benefit would be expected when the cervix was 15–25 mm (Owen et al., Am J Obstet Gynecol 2009;201:375.e1–375.e8). These studies have made it clear that, with some exceptions, a short cervix is evidence of parturition in progress, a process that is more biochemical than biophysical, and that in the most severe cases is evident well before the now outdated 20-week delineation between ‘miscarriage’ and ‘preterm labor’.
Studies of paired measurements of cervical length, as reported by Pils et al., are welcome because they allow for the assessment of the timing of onset and the rate of progression of this process. Looking at cervical length over time rather than in a specific gestational age window produces different interpretations of the same data. For example, the 1996 National Institute of Child Health and Human Development (NICHD) Preterm Prediction Study concluded that a short cervix at 22–24 and at 26–28 weeks of gestation is a predictor of subsequent spontaneous preterm birth before 35 weeks of gestation (Iams et al., N Engl J Med 1996;334:567–573). A re-examination of paired cervical length measurements from that study at 22–24 and 26–28 weeks of gestation revealed that cervical change had already occurred before 22–24 weeks of gestation in women destined to deliver after 28 weeks of gestation, and that the rate of cervical change before spontaneous preterm birth was similar, regardless of the eventual clinical presentation as preterm labour or membrane rupture (Iams et al., Am J Obstet Gynecol 2011;205:130.e1–e6). The parturitional process was evident weeks before both presentations.
The LEETZ procedure inserts another variable into the interpretation of short cervix. Pils et al., have contributed more than they may have realised by including repeated cervical measurements in their analysis. The reader is invited to read the paper again keeping the ‘short cervix is parturition in progress’ model in mind.
I have no conflicts of interest - no commercial relationships, advisory panels, speakers bureaus, stocks/investments in medical products.
I receive contract funding to perform clinical research studies from the NICHD, Cincinnati Childrens Med Center, & CaseWestern/MetroHealth, and stipends & royalties from Elsevier as editor of Am J Obstet Gynecol & of Creasy/Resnik textbook
Ohio State University, Columbus, OH, USA