Obstetrics & Gynecology:
Ultrasonographic Cervical Length and Risk of Hemorrhage in Pregnancies With Placenta Previa
Stafford, Irene A. MD; Dashe, Jodi S. MD; Shivvers, Stephan A. MD; Alexander, James M. MD; McIntire, Donald D. PhD; Leveno, Kenneth J. MD
From the Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center, Dallas, Texas.
Presented orally at the Society of Maternal Fetal Medicine Annual Meeting, February 1–6, 2010, Chicago, IL.
Corresponding author: Jodi S. Dashe, MD, Department of Obstetrics and Gynecology, University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas, TX 75390-9032; e-mail: firstname.lastname@example.org.
Financial Disclosure The authors did not report any potential conflicts of interest.
OBJECTIVE: To estimate the relationship between cervical length and hemorrhage leading to preterm delivery in women with placenta previa.
METHODS: Between October 2007 and May 2009, transvaginal cervical-length measurements were obtained in all singleton pregnancies with placenta previa identified at or beyond 24 weeks of gestation. Only women who delivered liveborn or stillborn neonates at our hospital and had placenta previa confirmed at delivery were included. Cervical length of 30 mm or less was considered short. Clinicians were blinded to cervical-length measurements. Chi-square and logistic regression were used for analysis.
RESULTS: Of 89 identified women with placenta previa at initial ultrasonography, 68 had placenta previa at delivery, and 29 (43%) of these had a short cervix. Gestational age at cervical-length measurement was 32±4 weeks in women with a short cervix and 33±2 weeks in those with a longer cervix (P=.4). Women with previa and a short cervix were more likely to require delivery for hemorrhage, 79% compared with 28%, and to deliver preterm, 69% compared with 21% (both P<.001). Tocodynamometer evidence of regular uterine contractions was more common with a short cervix than with a longer cervix, 69% compared with 21% (P<.001). Conversely, 64% with a cervical length greater than 30 mm had no bleeding episodes and progressed to term.
CONCLUSION: In pregnancies with placenta previa, a third-trimester cervical length of 30 mm or less is associated with increased risk for hemorrhage, uterine activity, and preterm birth.
LEVEL OF EVIDENCE: II
Although it affects only one in 300–400 deliveries,1,2 placenta previa and associated hemorrhage can be a potentially devastating complication of pregnancy and continues to be a leading cause of maternal morbidity and mortality. The etiology of hemorrhage is formation of the lower uterine segment and dilation of the internal os, with tearing of placental attachments, compounded by the inability of the lower uterine segment myometrium to constrict the torn vessels.1 This can lead to massive and unpredictable hemorrhage and emergent preterm delivery. Indeed, neonatal mortality is increased threefold in pregnancies with placenta previa with the major cause being prematurity.3
There is a clear relationship between ultrasonographic cervical length and preterm birth, particularly among women with a prior preterm birth.4,5 The reported tenth percentile for cervical length in the late second and early third trimester has ranged from 26 mm to 31 mm in unselected populations.4,6 A cutoff of 30 mm has been associated with a relative risk for preterm birth of 3.8.7 However, data are limited regarding cervical length in the setting of placenta previa. Because the mechanism of hemorrhage in placenta previa likely includes cervical effacement, it is plausible that there is an association among ultrasonographic cervical shortening, bleeding, and preterm birth. Women diagnosed with placenta previa in the second trimester typically return for ultrasonography in the mid–third trimester, and cervical length could easily be assessed at that visit. Our objective was to estimate the relationship between third-trimester cervical length and subsequent bleeding and preterm birth in women with placenta previa. An additional objective was to estimate the relationship between cervical length and uterine contractions in these pregnancies to better understand whether women with bleeding were experiencing contractions leading to cervical effacement.
MATERIALS AND METHODS
Between October 1, 2007, and May 30, 2009, we performed transvaginal cervical-length measurements on all singleton gestations with placenta previa undergoing ultrasound evaluation at our hospital at or beyond 24 weeks of gestation. It is our policy that when placenta previa is identified in the second trimester, follow-up ultrasonographic evaluation is performed at approximately 32 weeks to evaluate whether the placenta remains over the cervix. For study purposes, the last cervical-length measurement was used for analysis. Placenta previa was termed complete if the placenta completely covered the internal os and was termed incomplete if the placental edge partially covered or reached the margin of the internal os.8,9 Cervical length assessment, including presence or absence of funneling, was performed according to a protocol used by the National Institutes of Child Health and Human Development, Maternal–Fetal Medicine Units Network.4 Considering that women with placenta previa are at increased risk for preterm birth, we decided to use a cervical length cutoff at the conservative end of the published tenth percentile range.4,6,7 For study purposes, a cervical length of 30 mm or less was termed short. Cervical-length measurements were performed for the study protocol only and were not reported to the attending clinician; thus, the clinicians were blinded to cervical-length measurements throughout the study period. At our institution, women with placenta previa are not electively delivered before term. They are delivered for serious hemorrhage, eg, active, ongoing bleeding that is clinically significant, or in the absence of hemorrhage, they are delivered when the fetal age reaches term. Our ultrasonography unit is accredited by the American Institute of Ultrasound in Medicine, and all ultrasonographers are certified by the American Registry of Diagnostic Medical Sonographers.
We included only women who delivered liveborn or stillborn neonates at our hospital and who had placenta previa at delivery. This diagnosis was based on clinical assessment and ultrasonography at the time of delivery.9 Excluded were women with multifetal gestation or history of uterine anomalies. Fetal growth restriction was diagnosed using birth weight percentiles derived for our population adjusted for maternal ethnicity and neonate sex.10 If women were admitted antepartum for vaginal bleeding, they were approached within 12 hours of admission to obtain verbal informed consent, after which they were specifically questioned about their perception of uterine contractions in the days before and on the day of hospitalization. Tocodynamometer tracings were reviewed for the hospitalization in which the patient was delivered using criteria of at least four contractions per hour to signify uterine activity as described by Iams and colleagues.11 Women were similarly questioned about their perception of contractions before delivery. Outcome information was ascertained by review of an obstetric database that contains selected pregnancy outcomes, including anomaly information, about all neonates delivered at our hospital. It is maintained for clinical operations, including quality assurance. Perinatal nurses assess the data for consistency and completeness and enter outcomes for all women and their neonates into this database.
Sample size estimates were based on the assumption that among women with placenta previa, the ratio of cervical length greater than 30 mm to that 30 mm or less would be 3:1. For the outcome of preterm birth resulting from hemorrhage, a sample size of 100 would be sufficient to identify an absolute increase of 30% between groups, from 10% with cervical length more than 30 mm to 40% with cervical length less than 30 mm, with 80% power. After enrollment of 68 women, we noted that the combined preterm birth rate in the two groups was higher than anticipated, 41% compared with the planned 20%, and that more women than anticipated had cervical length 30 mm or less, 43% compared with the planned 10%, providing a ratio of 3:2 rather than 3:1. This prompted an interim analysis for which a P value of .014 was used for statistical significance using a Lan-DeMets spending function and altering the terminal significance level to .045 if necessary. Although not an originally planned intervention, it was made in a manner blinded to the results of the effects under study. Statistical analyses included Student's t test, chi-square test, Wilcoxon rank-sum test, and logistic regression with P<.05 considered significant. The study was approved by the Institutional Review Board at the University of Texas Southwestern Medical Center.
There were 89 women with placenta previa identified during the study period who subsequently delivered at our hospital. Of these, 68 women (76%) had placenta previa at delivery. Twenty-nine women (43%) with placenta previa had cervical lengths 30 mm or less. Maternal demographic characteristics are presented in Table 1. There were no differences in maternal age, ethnicity, nulliparity, prior preterm birth, prior cesarean delivery, prior dilatation and curettage, or reported tobacco use between the cervical length 30 mm or less and more than 30-mm groups. Ten percent of women in each group had a history of preterm birth in a prior pregnancy.
Ultrasound data are shown in Table 2. There were no differences between groups in terms of gestational age at cervical length assessment, type of placenta previa, or placental location (anterior, posterior, or lateral). Women with a short cervix were significantly more likely to have funneling, 28% compared with 0% (P<.001). The interval from ultrasonography until delivery was approximately 10 days shorter in women with a cervical length 30 mm or less (18.8±13.5 days compared with 29.0±16.7 days, P=.009).
Prevalence of vaginal bleeding and uterine contractions is presented in Table 3. Overall, 32 women (47%) with placenta previa were hospitalized for vaginal bleeding, 75% of those with a short cervix and 28% with a longer cervix (P<.001). In 12 women, 10 of whom had a short cervix, bleeding requiring hospitalization developed before the 32-week ultrasonography. If women with vaginal bleeding before 32 weeks are excluded from analysis, the prevalence of hospitalization for vaginal bleeding remained greater in women with a short cervix, 55% compared with 22% with a longer cervix (P=.01). Eleven of the women hospitalized for bleeding (34%) reported symptomatic uterine contractions. Among those hospitalized with bleeding, symptomatic uterine contractions were reported more often in the setting of a short cervix both on the day of hospitalization (34% compared with 3%, P<.001) and on the day before hospitalization (17% compared with 0%, P<.007). Thus, vaginal bleeding was associated with both a short cervix and self-detection of uterine contractions. Symptoms of uterine activity were also assessed on the day the patient delivered. Women with a short cervix were significantly more likely to experience uterine contractions at delivery, 55% compared with 13%, and to report having had contractions on the day before delivery (both P<.003). Tocodynamometry tracings confirmed that significantly more women in the short cervix group were having regular contractions on the day of delivery, 69% compared with 21% (P<.001).
Delivery information is presented in Table 4. Women with placenta previa and a short cervix were more likely to require delivery for hemorrhage than those with a longer cervix, 79% compared with 28% (P<.001). The probability of delivery secondary to hemorrhage according to cervical length using logistic regression analysis is shown in Figure 1. Women with a short cervix were also more likely to require emergent cesarean delivery, 76% compared with 28% (P<.001), and to require cesarean hysterectomy for placental invasion, 31% compared with 8% (P=.02). In addition, women with a short cervix were more likely to require delivery preterm, 69% compared with 21% (P<.001). The relative risk of preterm delivery in women with a short cervix was 3.36 (95% confidence interval:1.73–6.53). Conversely, those with a cervix more than 30 mm were significantly more likely to be delivered at term with no antecedent bleeding or hospitalization (64% compared with 17%, P<.001). This analysis was repeated after excluding the 10% of women in each group with a history of preterm birth, and women with a cervical length of 30 mm or less still had an increased rate of preterm birth as compared with those with a longer cervix (68% compared with 25%, P<.001). Of the preterm deliveries, 24 (86%) were for hemorrhage, 95% of preterm births in the short cervix group, and 63% of preterm births in the group with a longer cervix (P=.026). Indications for delivery in those without hemorrhage were oligohydramnios in two pregnancies, severe preeclampsia in one, and one stillbirth.
Neonatal outcomes are shown in Table 5. There was one stillbirth in the group with cervical length greater than 30 mm. The patient had received prenatal care from the first trimester, had no medical complications, and was found to have an unexplained stillbirth at 34 weeks of gestation. For the remaining 67 liveborn neonates, there was no difference between groups in the prevalence of fetal growth restriction, either less than the third percentile or less than the tenth percentile. Similarly, there were no differences between groups in the number of neonates with Apgar scores less than 3 at 5 minutes or umbilical artery pH less than 7.1. Birth weight was lower in the group with a short cervix, 2,524±694 g compared with 2,921±566 g (P=.01), corresponding to the difference in gestational age at delivery between cohorts. The rate of neonate admission to the intensive care nursery, 21% in the group with a short cervix and 8% with a longer cervix, was not statistically significant and directly paralleled the number of neonates delivered before 34 weeks. There were no neonatal deaths.
There are three main findings from this prospective study of ultrasonographic cervical length in women with placenta previa. First, women with placenta previa and cervical length of 30 mm or less were three times more likely to deliver preterm than those with longer cervices. This relative risk is similar to that reported for women with a prior preterm birth and a cervical length of 30 mm in the late second trimester.4 Second and perhaps more importantly, women with previa and a short cervix were more likely to develop bleeding. Indeed, antepartum hospitalization for vaginal bleeding was more than twice as common among women with cervical length 30 mm or less as was the risk of delivery for hemorrhage. Third, although vaginal bleeding with placenta previa has traditionally been considered painless, we found that more than one third of women with a short cervix who presented with bleeding were experiencing perceptible contractions as compared with only 3% of women who had a longer cervix. Moreover, significantly more women with a short cervix reported having contractions at least 1 day before presenting with bleeding.
Strengths of this series include that all clinicians (and patients) were blinded to cervical-length measurements so that we were able to study the relationship of cervical length to various outcomes without concern that management might be biased by the ultrasonographic findings. Another strength is that because women were not intentionally delivered preterm, we were better able to assess the likelihood that complications would develop according to cervical length. A limitation of our study is that our study design did not allow us to determine the temporal relationship between cervical shortening and the development of complications. For example, it is not possible to know if contractions preceded the development of a short cervix and subsequent bleeding or whether women first developed cervical shortening followed by later contractions. This is similar to difficulties encountered in the diagnosis of cervical incompetence. A prospective study of serial cervical-length measurements in women with placenta previa would better address this. A prospective study might also better address a potential role for cervical length in the prediction of placental invasion.
The role of bleeding in inciting contractions and that of contractions in inciting further bleeding has long been problematic in the management of women with placenta previa. Once bleeding develops, hemorrhage often ensues and leads to preterm birth. Berghella and colleagues12 studied the frequency of uterine contractions in pregnant women with a prior preterm birth as a function of cervical length. They found that in women with a cervical length below 25 mm in the second trimester, there was a trend toward an increased risk for preterm birth with tocodynamometer evidence of at least four contractions per hour.12 In our series, women with a short cervix were significantly more likely to have at least four contractions per hour on admission for delivery. Even if an ultrasonographic short cervix is merely a marker for relatively subtle contractions in these pregnancies, its important relationship to hemorrhage and preterm birth would suggest that it may be a useful tool in predicting whether hemorrhage will force delivery for placenta previa before term. The results of this study showing that hemorrhage resulting from placenta previa occurs in the setting of perceptible uterine contractions plus cervical shortening also suggests that the pathophysiology of bleeding resulting from previa is a result of labor effects on the placenta implanted onto a changing cervix.
1. Obstetrical hemorrhage. In: Cunningham FG, Leveno KJ, Bloom SL, Hauth JC, Rouse DJ, Spong CY, editors. Williams obstetrics. 23rd ed. New York (NY): McGraw-Hill; 2009. p. 769–73.
2. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2003. National vital statistics reports, Vol 54, No 2. Hyattsville (MD): National Center for Health Statistics; 2005.
3. Salihu HM, Li Q, Rouse DJ, Alexander GR. Placenta previa: neonatal death after live births in the United States. Am J Obstet Gynecol 2003;188:1305–9.
4. Iams JD, Goldenberg RL, Meis PJ, Mercer BM, Moawad A, Das A, et al. The length of the cervix and the risk of spontaneous premature delivery. National Institute of Child Health and Human Development Maternal Fetal Medicine Unit Network. N Engl J Med 1996;334:567–72.
5. Owen J, Yost N, Berghella V, Thom E, Swain M, Dildy GA, et al. Mid-trimester endovaginal sonography in women at high risk for spontaneous preterm birth. JAMA 2001;286:1340–8.
6. Gramellini D, Fieni S, Molina E, Berretta R, Vadora E. Transvaginal sonographic cervical length changes during normal pregnancy. J Ultrasound Med 2002;21:227–32.
7. Hibbard JU, Tart M, Moawad AH. Cervical length at 16–22 weeks' gestation and risk for preterm delivery. Obstet Gynecol 2000;96:972–8.
8. Feldstein VA, Harris RD, Machin GA. Ultrasound evaluation of the placenta and umbilical cord. In: Callen PW, editor. Ultrasonography in obstetrics and gynecology. 5th ed. Philadelphia (PA): Saunders Elsevier; 2008. p. 731–5.
9. Dashe JS, McIntire DD, Ramus RM, Santos-Ramos R, Twickler DM. Persistence of placenta previa according to gestational age at ultrasound detection. Obstet Gynecol 2002;99:692–7.
10. McIntire DD, Bloom SL, Casey BM, Leveno KJ. Birth weight in relation to morbidity and mortality among newborn infants. N Engl J Med 1999;340:1234–8.
11. Iams JD, Newman RB, Thom EA, Goldenberg RL, Mueller-Heubach E, Moawad A, et al. Frequency of uterine contractions and the risk of spontaneous preterm delivery. N Engl J Med 2002;346:250–5.
12. Berghella V, Iams JD, Newman RB, MacPherson C, Goldenberg RL, Mueller-Heubach E, et al. Frequency of uterine contractions in asymptomatic pregnant women with or without a short cervix on transvaginal ultrasound scan. Am J Obstet Gynecol 2004;191:1253–6.
© 2010 The American College of Obstetricians and Gynecologists