March 2002 · Vol. 51, No. 3
Patient-Oriented Evidence that Matters
Does cerclage prevent preterm birth or decrease perinatal morbidity when performed on the identification of a short cervix by second-trimester ultrasound?
Camille Andy, MD
Moses Cone Health System Family Medicine Residency Program
BACKGROUND: Because strong evidence supports the association between cervical length and preterm delivery and perinatal morbidity, transvaginal ultrasound (TVUS) has been used to identify patients with premature cervical change who may benefit from therapeutic cerclage placement. Observational studies report conflicting results regarding the benefits of therapeutic cerclage. Few randomized trials regarding the efficacy of cerclage therapy have been reported.
POPULATION STUDIED: This study enrolled 113 women from an urban outpatient perinatal testing center that had cervical changes identified by TVUS during the second trimester of pregnancy. Specific changes were dilation of the internal os and either (1) prolapse of membranes of at least 25% of the total cervical length or (2) a distal cervical length of less than 2.5 cm. Subjects were excluded if they had membrane prolapse beyond the external os or any other contraindication to cerclage. All participants were similar with regard to age, risk factors for preterm labor, and history of preterm deliveries. However, patients in the no-cerclage group tended to have an increased rate of second-trimester deliveries (12.1 vs 27.3, P = .07).
STUDY DESIGN AND VALIDITY: This was an open randomized controlled trial. Participants with cervical changes were randomized to receive either McDonald cerclage (n = 55) or no cerclage (n = 58). Before randomization, all patients received amniocentesis, multiple urogenital cultures, indomethacin, and clindamycin. All subjects were treated identically (including serial ultrasonography and modified bed rest) after the intervention. Routine prenatal care continued and cerclage was removed at 36 weeks’ gestational age or for any of the following reasons: rupture of membranes, preterm labor refractory to tocolytic therapy, or other indication for delivery. Analysis was done by intention to treat. The authors developed a stepwise logistic regression model to analyze dependent and independent variables. It is uncertain whether allocation assignment to treatment group was concealed.
The study population is generalizable and large enough to have adequate power to support a negative result. Despite reporting good data that cerclage is not beneficial in preventing preterm deliveries and decreasing neonatal morbidity, the authors wanted to identify several risk factors that were associated with failure of the cerclage. The logistic regression model, although statistically sound, makes this study difficult to follow and the results do not add important clinical information. In addition, not all the significant data are clearly given, specifically the data regarding perinatal morbidity between the 2 groups.
OUTCOMES MEASURED: The primary outcomes measured were gestational age at delivery and neonatal morbidity, defined as none, minimal, severe, or death. The authors analyzed a number of other variables to determine any associations with the primary outcomes but did not address cost effectiveness or patient satisfaction.
RESULTS: There were no statistical differences in the primary outcomes of gestational age at delivery of less than 34 weeks’ gestation (34.9 vs 36.2, P = .8) or perinatal morbidity, reported only as perinatal death (12.7 vs 11.9, P = .9). The regression model analysis identified preterm labor, chorioamnionitis, and abruption as significant risk factors associated with the primary outcomes.
RECOMMENDATIONS FOR CLINICAL PRACTICE
In the current study, therapeutic cerclage did not benefit patients identified as having a short cervix by second-trimester ultrasound. These findings do not support the routine use of second-trimester ultrasound to screen for premature cervical changes. Risk factor analysis provided further data showing that preterm deliveries are a multifactorial process and that patients with preterm labor, infection, and abruption do not benefit from cerclage.
Another well-designed but smaller randomized controlled trial1 studied a subset population of pregnant women at high risk for cervical incompetence and found therapeutic cerclage to be beneficial. Comparison of these 2 studies suggests that a subset population might benefit from cerclage therapy. Unfortunately, with current medical diagnostics, this ideal patient is not easily identifiable. An ongoing Cochrane Review that addresses cerclage efficacy may help clarify these discrepancies.
- Althuisius SM,
van Geijn HP.
Final results of the Cervical Incompetence Prevention Randomized Cerclage Trial (CIPRACT): therapeutic cerclage with bedrest versus bedrest alone. Am J Obstet Gynecol 2001;185:1106–12.
Copyright© 1995-2008 John Wiley & Sons, Inc. All rights reserved.
The Journal of Family Practice ©2002 Quadrant HealthCom Inc.