Objective
To compare outcomes of physical examination-indicated cerclage in women with twin gestations to those with singleton gestations and to identify whether risk factors for extremely preterm birth (before 28 weeks) differ between these 2 groups.
Study Design
This is a single institution retrospective cohort study of women who underwent a physical examination-indicated cerclage between Jan. 1, 1980, and Aug. 15, 2012. Differences in characteristics and outcomes were compared between women with twin and singleton gestations. A multivariable analysis was performed to examine whether twin gestation was independently associated with delivery before 28 weeks and whether any effect modification was present between risk factors for preterm birth and the presence of a twin gestation.
Results
Of the 442 women who underwent a cerclage during the period of study, 104 (23.5%) had twins. Mean gestational age and digital cervical length at placement did not differ by plurality. Although twins were more likely to deliver at a slightly earlier median gestation than singletons (31.9 weeks; interquartile range, 24.9–35.1 vs 32.7 weeks; interquartile range, 24.6–38.3; P = .015), the frequency of delivery before 28 weeks did not differ between these 2 groups (33.7% vs 35.8%, P = .69). Greater cervical dilation and prolapsing membranes were identified as risk factors for birth <28 weeks in both groups; digital cervical length <2 cm appeared to be a risk factor particularly for women with twin gestations.
Conclusion
Women with a twin pregnancy who received a physical examination-indicated cerclage had similar risk factors for extreme preterm birth and may experience similar obstetric outcomes as women with singleton gestations.
See related editorial, page 5
Twin births have increased dramatically and now account for more than 3% of all live births. This marked increase has corresponded with an increase in the frequency of preterm delivery, as twins account for a disproportionate share of preterm deliveries and their associated perinatal morbidity. One factor leading to preterm delivery is cervical insufficiency, often considered to be present when there is painless cervical dilation in the midtrimester. This condition is estimated to account for 10-25% of all second-trimester pregnancy losses. The rate of cervical insufficiency in women with multiple gestations (5.0%) is significantly greater than among women with singleton pregnancies (0.05-1.8%).
There has been some evidence, both from observational studies as well as a randomized trial, that a cerclage placed in response to midtrimester cervical dilation (ie, a physical examination-indicated cerclage) can prolong singleton gestations. In addition, several authors have elucidated factors that are associated with success of physical examination-indicated cerclage in singleton pregnancies. However, current literature provides little guidance on the outcome of twin gestations after placement of a physical examination-indicated cerclage.
Given this lack of information, the ability to counsel a woman with a twin pregnancy who is thought to be a candidate for a physical examination-indicated cerclage is limited. Thus, in this study we have evaluated whether the characteristics and outcomes of women who receive a physical examination-indicated cerclage with a twin gestation differ from those of women with a singleton pregnancy.
Materials and Methods
This was a retrospective cohort study of all patients who underwent a physical examination-indicated cerclage placed at Northwestern Memorial Hospital between Jan. 1, 1980, and Aug. 15, 2012. Subjects were identified through a search of inpatient medical records for cervical cerclage placement ( International Classification of Diseases, 9th Revision , codes 67.5 and 67.59). All physical examination-indicated cerclages were placed on the inpatient labor and delivery unit during the study period. Women included in the study were those with singleton or twin gestations who presented between 16 0/7 and 23 6/7 weeks’ gestational age with painless cervical dilation diagnosed by digital examination and who subsequently delivered at the same institution. Women were excluded from the study if cerclage placement occurred after delivery of 1 fetus of a twin gestation, as a revision of a prior history-indicated cerclage, or in response to sonographic findings without a dilated cervical examination.
The medical records of all subjects were reviewed and data regarding cerclage placement and delivery were abstracted. Abstracted demographic data included maternal age, race, gravidity, parity, and gestational age at cerclage placement. Pertinent factors from a woman’s past obstetric and gynecologic history also were recorded including prior pregnancy outcomes and risk factors for cervical insufficiency (ie, history of diethylstilbestrol exposure, uterine anomalies, or history of cervical surgery including a conization or loop electrosurgical excisional procedure). Cervical dilation, effacement, and location of membranes on digital examination at time of presentation for cerclage placement were recorded.
Outcome data included gestational age at delivery (both continuous and dichotomized by delivery prior to 28 or 32 weeks), latency postcerclage, the occurrence of preterm premature rupture of membranes (PPROM) or chorioamnionitis, and neonatal intensive care unit (NICU) admission.
During the study period, physical examination-indicated cerclages were placed by a maternal-fetal medicine attending physician with the assistance of a resident and/or fellow using a McDonald or Shirodkar technique with 5 mm Mersilene suture. Before cerclage placement, clinical evaluation for evidence of preterm labor or intraamniotic infection was performed. Amniocentesis was not routinely used. Adjuvant techniques such as perioperative tocolytics or antibiotics, retrograde filling of the bladder, or use of an intracervical Foley for membrane prolapse reduction were used at the discretion of the treating physician and were recorded.
Comparisons were made between twins and singletons using Student t , Mann-Whitney U , Fisher exact, or χ 2 tests, as appropriate. A Kaplan-Meier analysis with a log-rank test was performed to compare gestational age at delivery as a continuous variable by plurality. Patient characteristics associated with an extremely preterm birth (ie, before 28 weeks) at a P < .05 level were identified in bivariable analysis. The presence of a twin gestation was forced into the model. Interaction terms were created between twin gestation and each of these associated risk factors. A multivariable logistic regression was then performed including all patient factors significantly ( P < .05) associated with preterm birth before 28 weeks, plurality, and the interaction terms to identify whether effect modification was present. Statistical analysis was performed using Stata version 11.1 (StataCorp LP, College Station, TX). This study was approved by the Institutional Review Board of Northwestern University with a waiver of informed consent.
Results
During the study period, 442 women underwent a physical examination-indicated cerclage: 104 (23.5%) twins and 338 (76.5%) singletons. Table 1 shows a comparison of baseline characteristics stratified by plurality. Women with twins were more likely to be older, nonHispanic white, less parous, and less likely to have experienced a prior preterm birth or midtrimester loss. Baseline preoperative and intraoperative characteristics are shown in Table 2 . Only 7 (1.6%) cerclages were performed using a Shirodkar. There were no differences in gestational age or cervical exam (ie, dilation, length, the presence of membrane herniation) at cerclage placement or the use of adjunctive interventions between the 2 groups. Notably, although this study spanned 32 years, outcomes did not change over the 2 halves of the study. Both median latency after cerclage (83 days; interquartile range [IQR], 24–115 vs 64 days; IQR, 22–108; P = .15) and median gestational age at delivery (32.7 weeks; IQR, 24.7–38.0; vs 31.3 weeks; IQR, 24.9–36.9; P = .22) were similar regardless of whether women received cerclage in the first or second half of the study period. Similarly preterm birth before 28 weeks did not differ by epoch (92 [34.2%] vs 70 [37.8%], P = .44).
| Characteristic | Twin gestation n = 104 | Singleton gestation n = 338 | P value |
|---|---|---|---|
| Age, y | 32.4 ± 5.5 | 30.1 ± 6.1 | < .01 |
| Race/ethnicity, n (%) | < .001 | ||
| NonHispanic white | 62 (60.9) | 92 (28.5) | |
| NonHispanic black | 19 (18.6) | 139 (43.0) | |
| Hispanic | 7 (6.9) | 62 (19.2) | |
| Other | 14 (13.7) | 30 (9.3) | |
| Nulliparous, no prior PTD | 58 (55.8) | 119 (35.2) | < .001 |
| Multiparous, no prior PTD | 22 (21.2) | 39 (11.5) | .013 |
| Mutiparous, prior PTD | 24 (23.1) | 180 (53.3) | < .001 |
| Prior midtrimester loss | 7 (6.8) | 98 (28.9) | < .001 |
| Cigarette use | 2 (1.9) | 18 (5.3) | .14 |
| Uterine anomaly | 4 (3.9) | 9 (2.7) | .53 |
| DES exposure | 1 (1.0) | 3 (0.9) | .94 |
| Prior LEEP/CKC | 17 (16.5) | 34 (10.0) | .08 |
| Variable | Twin gestation n = 104 | Singleton gestation n = 338 | P value |
|---|---|---|---|
| Gestational age at placement, n (%) | .21 | ||
| <18 wks | 6 (5.8) | 39 (11.5) | |
| 18-19 wks | 19 (18.4) | 55 (16.2) | |
| 20-21 wks | 37 (35.9) | 136 (40.1) | |
| ≥22 wks | 41 (39.8) | 107 (31.6) | |
| Dilation, cm | 1.5 (1–2.5) | 1.5 (1–2.5) | .86 |
| Digital cervical length <2 cm, n (%) | 53 (51.0) | 187 (55.3) | .45 |
| Membrane prolapse, n (%) | .23 | ||
| None | 56 (54.3) | 154 (45.4) | |
| Into cervical canal | 25 (24.3) | 100 (29.5) | |
| Beyond external os | 22 (21.4) | 85 (25.1) | |
| Adjunctive procedures, n (%) | |||
| Bladder retrofilled | 39 (37.9) | 129 (38.1) | .90 |
| Foley in cervix | 6 (5.8) | 28 (8.3) | .40 |
| Perioperative antibiotics | 56 (54.3) | 200 (59.0) | .34 |
| Perioperative tocolytics | 59 (57.3) | 187 (55.2) | .80 |
Gestational age at delivery differed between the 2 groups, with twins delivering at a median of 31.9 weeks (IQR, 24.9–35.1) compared with singletons at 32.7 weeks (IQR, 24.6–38.3) ( P = .015). Rates of delivery before 28 and 32 weeks did not differ. Thus, this difference was primarily related to differences in delivery at gestational ages later than 32 weeks ( Figure , P < .01). Rates of PPROM and chorioamnionitis did not differ between the 2 groups, however, latency between cerclage placement and PPROM was longer in twin gestations compared with singleton gestations. NICU admission was more common in twin neonates compared with their singleton counterparts ( Table 3 ). NICU admission frequency did not significantly change over the course of the study period ( P = .21).
