Vergani, Patrizia MD; Locatelli, Anna MD; Ghidini, Alessandro MD; Andreani, Marianna MD; Sala, Francesca MD; Pezzullo, John C. PhD
OBJECTIVE: To study the causes of the observed association between presence of leiomyomata in pregnancy and increase in risk of cesarean delivery.
METHODS: We accessed the obstetric database of women who underwent second-trimester ultrasound screening between January 1996 and December 2004 and who delivered at more than 22 weeks of gestation. Ultrasonographic characteristics of the leiomyomata, including number, maximum diameter, change in size during pregnancy, and location were recorded. Pregnancy outcome was compared between the cohort with large uterine leiomyomata (5 cm or more) (n=251) and those without leiomyomata (n=24,546) using χ2, Fisher exact test, Student t test, and linear and logistic regression analysis, with an odds ratio (OR) with 95% confidence interval (CI) not inclusive of the unity considered significant.
RESULTS: Women with large leiomyomata had higher rates of cesarean delivery before onset of labor (OR 3.1, 95% CI 2.2–4.2) but not during labor (OR 1.0, 95% CI 0.6–1.6). The association between leiomyomata and cesarean delivery before labor remained significant after controlling for maternal age, nulliparity, fetal malpresentation, and placenta previa (OR 2.1, 95% CI 1.4–3.1). The only ultrasonographic leiomyoma characteristic independently associated with cesarean delivery before labor after controlling for confounders was maximum diameter of the largest leiomyoma (OR 1.3, 95% CI 1.1–1.4, P=.001).
CONCLUSION: Uterine leiomyomata 5 cm or larger are independently associated with cesarean delivery performed before labor, and the risk increases with the size of the leiomyoma.
LEVEL OF EVIDENCE: II
Leiomyomata in pregnancy are a relatively common occurrence, with a prevalence of 2–4%.1,2 Several series have noted an association between leiomyomata and risk of cesarean delivery. Although the earlier reports did not control for potential confounders, two more recent series using logistic regression analysis have confirmed an independent association between presence of leiomyomata and risk of cesarean delivery.1,3,4 Coronado et al,5 using International Classification of Diseases, 9th Revision, Clinical Modification, at discharge for identification of women with leiomyomata, found an adjusted odds ratio (OR) of 6.4 (95% confidence interval [CI] 5.5–7.5) for risk of cesarean delivery. Qidwai et al6 has shown that ultrasonographic detection of leiomyomata of any size is associated with a significant increase, in risk of both total cesarean delivery (OR 1.64, 95% CI 1.28–2.11) and cesarean delivery in labor (OR 1.57, 95% CI 1.16–2.13). However, important confounders (such as placenta previa) were not taken into account in either series. Moreover, it is unclear whether the presence of leiomyomata is associated with an increase in risk of cesarean delivery before labor (eg, because of increased risk of placenta previa or fetal malpresentation) or risk of cesarean delivery in labor (eg, due to dystocia or fetal distress).
Studies of the effect of ultrasonographic characteristics of the leiomyomata on the risk of cesarean delivery have shown that larger leiomyomata have a greater impact on risk of cesarean delivery than smaller tumors. For example, we have previously reported that leiomyomata of 5 cm or more in average diameter are independently associated with higher risk of cesarean delivery after accounting for known confounders (OR 2.6, 95% CI 1.2–5.5).3
Therefore, in the present study we have examined the effects of leiomyomata of 5 cm or more on risk of cesarean delivery in a large cohort of women identified during ultrasonographic screening in the second trimester. The aims of the study were the following: 1) to assess whether presence of leiomyomata is independently associated with increased risk of cesarean delivery before labor, in labor, or both; and 2) to evaluate which ultrasonographic characteristics of the leiomyomata can predict the risk of cesarean delivery.
MATERIALS AND METHODS
We conducted a retrospective cohort study based on a database of all women with singleton pregnancies and ultrasound diagnosis of uterine leiomyomata in pregnancy from January 1996 to December 2004 and who delivered at more than 22 weeks of gestation. This gestational age was chosen because it corresponds to the weeks of gestation when the routine sonographic screening is completed. All women enrolled in the prenatal clinic at our department undergo an ultrasound examination between 16 and 22 weeks of gestation for dating purposes and fetal anatomy survey. The examinations are performed and analyzed by physicians with specific training in obstetric sonography.
When uterine leiomyomata were visualized, anteroposterior, transverse, and longitudinal diameters were measured in each leiomyoma, and the average diameter of the tumor at the initial ultrasound examination was used to indicate the initial size. In the presence of multiple leiomyomata, the dimensions of the largest myoma were considered. Included in the present study were only those gestations with an average leiomyoma size at first ultrasound examination of 5 cm or more. Women were followed with monthly ultrasound scans until delivery, and the pertinent information was recorded in a dedicated database. The following leiomyoma characteristics as reported at the last ultrasound examination before delivery were used for the analysis: 1) location in relationship to the uterus (at or below the lower uterine segment versus the uterine corpus) and to the presenting fetal part; 2) location in relation to the placental insertion (retroplacental or not); 3) number of leiomyomata (single versus multiple); 4) maximum diameter of the largest leiomyoma; 5) cumulative myomatous diameter (calculated as the sum of the maximum diameters of all leiomyomata); and 6) change in size of leiomyomata between first and last scan (increased more than 10%, stable, or decreased more than 10%). The clinical management during labor for women with leiomyomata did not differ from that of the general obstetric population.
The ultrasound database was linked to the perinatal database, and the following pregnancy complications and outcomes were extracted: premature rupture of membranes, defined as amniorrhexis before the onset of labor; preterm delivery, defined as birth at less than 37 weeks; placental abruption; placenta previa, defined ultrasonographically when the lower edge of the placenta was 2 cm or less from the internal cervical os; and malpresentation at delivery, including any fetal presentation other than cephalic.
Cesarean delivery was divided into those performed before onset of labor and those performed during labor, including those for nonreassuring fetal testing and labor dystocia. Labor dystocia was diagnosed when cervical dilatation was arrested for 2 hours or progressed below the fifth percentile of the Friedman curve for more than 2 hours despite appropriate augmentation of labor with amniotomy or oxytocin or both. Blood loss at delivery was quantified with the use of graduated collection bags at vaginal delivery, and of suction bottles and weighed surgical pads at cesarean delivery. Postpartum hemorrhage was defined as blood loss of 1,000 mL or greater. The study was approved by the institutional review board of the University of Milano-Bicocca.
Statistical analysis used χ2, Fisher exact test, and Student t test. Multiple logistic regression analyses were used to assess the independent associations of variables found to be significantly associated with leiomyomata at univariable analysis. A two-tailed P<.05 or an OR with 95% CI not inclusive of the unity was considered significant.
During the study period, 608 consecutive pregnant women had sonographically identified leiomyomata out of a total of 25,154 women registered at our prenatal clinic (3%). After exclusion of those with gestational age at delivery of less than 22 weeks (n=5), lost to follow-up (n=10), and with average leiomyoma size at first ultrasound examination of less than 5 cm (n=342), 251 women fulfilled the study criteria and had leiomyomata of 5 cm or larger. Compared with women delivered during the study period without myomas at ultrasonographic screening (n=24,546), women with large leiomyomata were older (34.5±4.2 versus 30.8±4.4, P<.001), more frequently nulliparous (70% [n=176] versus 59% [n=14,451], OR 1.6, 95% CI 1.2–2.2), and had higher rates of placenta previa and fetal malpresentation. Race was white in 99% of women.
At the last ultrasound examination before delivery 40 of 251 women had multiple leiomyomata (16%), and 43 (17%) had tumors of 10 cm or greater in average diameter. In reference to the uterine location, 6% were retroplacental, 23% were located within the uterine isthmus, and 3% were located between the presenting part and the cervix.
Table 1 shows the pregnancy complications and mode of delivery in reference to the presence of large leiomyomata. Women with large tumors had higher rates of cesarean delivery before labor but not during labor (13 of 251 or 5.2% versus 1,303 of 24,546 or 5.3%, P=.93). Our study had a power of 94% for the detection of a significant difference between the two groups (α=0.05). As seen in Table 2, neonatal outcome was not significantly affected by the presence of large leiomyomata.
Multiple logistic regression analyses demonstrated that the presence of large myomas was independently associated with occurrence of placenta previa, fetal malpresentation, and cesarean delivery before labor (Table 3). In an attempt to understand whether a specific indication emerged as more frequent, we examined the indications for the 27 cases of cesarean delivery performed before the onset of labor that were not due to malpresentation or placenta previa in women with leiomyomata. The indication was nonreassuring fetal heart rate tracing in six cases, previous uterine scar in six, maternal obstetric complications in four, tumors located below the presenting fetal part in four, fetal growth restriction and oligohydramnios in three, third-trimester bleeding in two, maternal request in one, and maternal chorioamnionitis in one.
To explore which ultrasonographic leiomyoma characteristics could predict the risk of elective cesarean delivery, further analysis was limited to the 251 women with tumors of 5 cm or larger. Univariable analysis demonstrated that women undergoing cesarean delivery before labor were older, had lower gestational age at delivery, a more frequent history of previous cesarean delivery, or a scarred uterus, and they had fetuses in nonvertex presentation and of lower weight than the remaining women (Table 4). Moreover, those undergoing cesarean delivery before labor had significantly larger leiomyomata (Table 5). The small number of cases with increase or decrease in size of tumors did not allow a meaningful stratified analysis for change in size of the leiomyomata. Logistic regression analysis demonstrated that, after controlling for maternal age, gestational age at delivery, presence of a scarred uterus and nonvertex fetal presentation, maximum leiomyoma diameter was the only characteristic independently associated with odds of cesarean delivery before labor (OR 1.3, 95% CI 1.1–1.4, P=.001).
In women with leiomyomata, an increased rate of cesarean delivery is the obstetric complication most frequently observed.1,4. Consistent evidence has been gathered that malpresentation,5–7 placenta previa,5–7 and location of the tumors in the lower uterine segment below the presenting fetal part7 are the most common indications for cesarean delivery in women with leiomyomata. However, presence of leiomyomata remains significantly associated with increased risk for cesarean delivery even after correcting for the above indications. To explore this conundrum we have analyzed a cohort of women with leiomyomata of 5 cm or more in size because in a previous study we have shown that such a population is at greatest risk for cesarean delivery.3 Moreover, the cohort was identified prenatally with ultrasonography to avoid selection biases.
We confirm that leiomyomata of 5 cm or more in average diameter are associated with increased risk of cesarean delivery before, but not during, labor. This observation is interesting because the series which previously reported an independent association between leiomyomata and cesarean delivery originated from institutions with total cesarean delivery rates as high as 49%6 and 58%5 among women with uterine tumors, whereas the respective rate in our population was 26%. This suggests that, even where a policy of minimizing unnecessary cesarean delivery is implemented, as in our institution, there remains an excess rate of cesarean delivery associated with the presence of leiomyomata.
We did not find that large leiomyomata are associated with an increased risk of cesarean delivery in labor, and the lack of significance cannot be attributed to a lack of statistical power. In particular, we did not find a significant increase in risk of cesarean delivery for labor dystocia among women with large myomas. This finding supports the observation of Quidai et al6 who found that presence of leiomyomata is not associated with significant differences in duration of first or second stage of labor, as it would be expected in a dystocic labor, even after controlling for the size of the myomas. In summary, women with leiomyomata can be reassured that, if they are eligible for vaginal delivery and they initiate labor, they can expect similar rates of vaginal delivery as in the general obstetric population.
Are women with leiomyomata at increased risk for cesarean delivery before labor after exclusion of those with fetal malpresentation, placenta previa, and location of the tumors in the lower uterine segment below the presenting fetal part? In our cohort of women with prospective ultrasonographic identification of leiomyomata, large tumors are still associated with increased risk of cesarean delivery before the onset of labor, even after taking into account the above-mentioned indications. This observation originates from an institution like ours in which presence of large leiomyomata per se is not an indication for cesarean delivery. This suggests that sizable uterine tumors per se play a role in the decision to proceed with an operative abdominal delivery. In support of this hypothesis comes our observation that the maximum leiomyoma diameter is the only sonographic characteristic independently associated with risk of cesarean delivery before labor. We have found that the risk of cesarean delivery before labor increases by 26% for every 1-cm increase in maximum leiomyoma diameter. Similarly, independent investigators also found a trend for increased risk of planned cesarean delivery in women with leiomyomata of 10 cm or greater.6 When we attempted to explore the indications for cesarean delivery before the onset of labor in women without placenta previa or malpresentation, we were unable to identify a dominant indication. This suggests that perhaps the presence of a large uterine tumor, although it was not per se an indication for cesarean delivery before labor, possibly lowered the threshold to proceed with the abdominal route of delivery when other obstetric complications or factors were present, which, in isolation, would not have constituted an indication for elective cesarean delivery. In light of the above considerations and the good prognosis of labor in women with leiomyomata of any size, obstetricians should probably be wary of proceeding with an elective cesarean delivery in the absence of specific obstetric indications.
1. Katz VL, Dotters DJ, Droegemueller W. Complications of uterine leiomyomas in pregnancy. Obstet Gynecol 1989;73:593–6.
2. Exacoustos C, Rosati P. Ultrasound diagnosis of uterine myomas and complications in pregnancy. Obstet Gynecol 1993;82:97–101.
3. Vergani P, Ghidini A, Strobelt N, Roncaglia N, Locatelli A, Lapinski RH, et al. Do uterine leiomyomas influence pregnancy outcome? Am J Perinatol 1994;11:356–8.
4. Rice JP, Kay HH, Mahony BS. The clinical significance of uterine leiomyomas in pregnancy. Am J Obstet Gynecol 1989;160:1212–6.
5. Coronado GD, Marshall LM, Schwartz SM. Complication in pregnancy, labor, and delivery with uterine leiomyomas: a population-based study. Obstet Gynecol 2000;95:764-9.
6. Qidwai GI, Caughey AB, Jacoby AF. Obstetric outcomes in women with sonographically identified uterine leiomyomata. Obstet Gynecol 2006;107:376–82.
7. Sheiner E, Bashiri A, Levy A, Hershkovitz R, Katz M, Mazor M. Obstetric characteristics and perinatal outcome of pregnancies with uterine leiomyomas. J Reprod Med 2004;49:182–6.