Cordocentesis, first introduced by Daffos et al,1 has been well accepted as a standard invasive procedure for fetal diagnosis. Although it is relatively safe, the procedure is associated with an increased risk of fetal loss ranging from 1.5–2.8%.2,3 Nevertheless, the risk of other serious complications potentially associated with cordocentesis, eg, preterm birth, intrauterine growth restriction, etc, has not been adequately investigated. Most of the previous studies regarding cordocentesis-related morbidity involved pregnancies of relatively late gestation and lacked controls.4–10 Based on our preliminary study on cordocentesis-related fetal loss,2 the rates of small for gestational age, preterm birth, and low birth weight were increased and have a tendency to increase among pregnancies undergoing second-trimester cordocentesis. This increase is however not statistically significant. Theoretically, cordocentesis may be associated with umbilical cord or placental injury, which leads to thrombosis or emboli. If such thromboembolic events occurred, they might interfere with fetal circulation or cause obstruction in some critical vessels resulting in small for gestational age, fetal loss, or chronic hypoxia secondary to inadequate perfusion. Furthermore, bleeding in the placenta in cases of penetration through the placenta during the procedure could gradually extend to form a large hematoma, chronic abruption, or thrombosis in the intervillous space resulting in placental infarction. Therefore, we conducted this study to ascertain the rates of small for gestational age, preterm birth, and low birth weight associated with cordocentesis in the second trimester.
MATERIALS AND METHODS
A cohort study was conducted at the tertiary care Maharaj Nakorn Chiang Mai Hospital, Chiang Mai University with ethical approval by the institutional review board (Research Ethics Committee 4, Faculty of Medicine, Chiang Mai University, Study Code: OBG-13-1433-EX/Research ID 1433). During the study period (from September 1995 to December 2013), singleton pregnant women attending our antenatal care clinic were enrolled in the study after written informed consent. The inclusion criteria were as follows: 1) gestational age of 16–22 weeks based on fetal ultrasonographic biometry in the first half of pregnancy; 2) no fetal anomaly on ultrasound screening before the cordocentesis; and 3) no known medical or obstetric complications at the time of recruitment.
The cordocenteses were performed by the authors using a freehand technique under transabdominal ultrasound guidance and using a convex 3.5-MHz transducer and Voluson E8 ultrasound machine. The details of the techniques were described elsewhere.2,11 The procedures were carried out in an outpatient setting with real-time ultrasound guidance to confirm the number and viability of the fetus, gestational age, location of the placenta, and site for puncture (cord insertion or free loop of cord). A 22- or 23-gauge spinal needle was used under local anesthesia without fetal paralysis. The puncture site, either near cord insertion or free-floating loop, was chosen based on the accessibility and quality of visualization regardless of location. Cordocenteses were performed in only one center (Maharaj Nakorn Chiang Mai Hospital) but antenatal care and delivery were managed by the doctors in the network hospitals and the recorded outcome data were collected by mail and telephone. The data were entered by one of three researchers. The control pregnancy was selected and matched for each study patient on the basis of a one-to-one ratio. The control met the following criteria: 1) a pregnant woman in a first visit to an antenatal clinic before 22 weeks of gestation; 2) undergoing ultrasound examination for anomaly screening at 16–22 weeks of gestation and no fetal structural anomaly; 3) maternal age within 2 years of the age of the case study; 4) the same parity (nulliparous, parous, or grand multiparous [greater than five]); 5) the same ethnic group; and 6) low-risk pregnancy without obvious medical and obstetric complications at the time of recruitment (16–22 weeks of gestation). The exclusion criteria for the matched pairs were: 1) serious fetal structural or chromosomal anomalies at the time of screening; 2) results of the cordocentesis suggestive of abnormalities, eg, fetal hemoglobinopathy (fetal carrier of a hemoglobinopathy were not excluded); 3) incomplete outcome data or loss to follow-up; and 4) previous cordocentesis, amniocentesis, or chorionic villous sampling in the current pregnancy.
All patients were recruited between 16 and 22 weeks of gestation and were reviewed for medical and obstetric history at the time of the procedures and were followed up until delivery (Fig. 1). The matched pairs would be excluded if either of the women was lost to follow-up or underwent repeated cordocentesis. All demographic data and details of ultrasound findings, techniques of cordocentesis, and pregnancy outcomes were prospectively recorded in a computerized database. Cordocentesis profiles were immediately recorded during the procedure such as duration of the procedure, placenta penetration, puncture site, umbilical cord bleeding and its duration, bradycardia and its duration, etc.
Definitions used in this research were as follows: 1) small for gestational age: birth weight of less than 10th percentile for each gestational week based on our population standard growth curve12; 2) preterm birth: delivery after 22 weeks and before 37 complete weeks of gestation; 3) early preterm birth: preterm birth before 34 complete weeks of gestation; 4) abortion: a pregnancy ending before 22 complete weeks of gestation; 5) low birth weight: birth weight of less than 2,500 g; 6) fetal loss: a pregnancy complicated by spontaneous abortion or fetal death, 7) midpregnancy: pregnancy at gestational age between 16 and 22 weeks; 8) early cordocentesis: the procedure performed during 16–17 weeks of gestation; 9) experienced operator: a performer who had experience in cordocentesis of at least 40 cases; 10) bleeding: bleeding from umbilical cord puncture, categorized into three groups: prolonged bleeding (more than 1 minute), transient bleeding, or no obvious bleeding; 11) fetal bradycardia: fetal heart rate of less than 100 beats per minute lasting more than 1 minute, 12) puncture site: location for needle puncture categorized into cord insertion (a distance from the placenta less than 5 cm) or free loops; 13) difficult procedure: procedures with duration time of more than 30 minutes; and 14) failed procedure: procedures resulting in inadequate fetal blood for examination. The primary outcomes of measures included the small for gestational age, preterm births, and low birth weight. The secondary outcomes included fetal loss rate and maternal obstetric complications. All analyses were performed using IBM SPSS 21.0. The sample size was determined on the assumption that a relative risk of the small for gestational age among pregnancies undergoing cordocentesis was 1.4 and the expected incidence was 4% in the control group of low-risk pregnancy. Thus, 4,633 cases per group was required to achieve significance at the 5% level with a power of 95%. The rates of pregnancy outcomes were compared using McNemar's χ2 test and paired t test as appropriate; the significance of factors related to cordocentesis influencing the outcomes was tested using logistic regression (forward stepwise method) analysis. A P value <.05 was considered statistically significant.
During the study period, 7,228 cordocenteses in the second trimester were performed by the authors. Of these, 216 cases were multifetal pregnancies and were initially excluded. Of the remaining 7,012 women, 624 were further excluded because of medical or obstetric complications. Of the 6,388 eligible pregnancies, 882 pregnancies had fetal abnormalities detected by fetal blood analysis or ultrasound scanning and were further excluded. A total number of 5,506 women were allocated to the study group and 5,506 women in a control group were identified for matching based on the criteria mentioned. Of these, 467 were further excluded because of loss of follow-up or incomplete data of the final outcomes. Thus, 5,039 matched pairs were available for analysis.
The baseline characteristics of the two groups were similar (Table 1). The most common indications for cordocentesis were fetal risk for severe thalassemia, 3,175 cases (63.0%), the risk for hemoglobin Bart's disease (23.0%), β-thalassemia major (17.0%), and β-thalassemia or hemoglobin E disease (23.0%) (Table 1). Most women who underwent cordocentesis were recruited by our established screening program for prenatal diagnosis of severe thalassemia and Down syndrome in our network of hospitals.13 Rapid karyotyping was the second most common indication (1,408 cases [27.9%]), mostly recruited from our Down screening project established in the northern part of Thailand.
The main pregnancy outcomes including small for gestational age, preterm birth or early preterm birth, low birth weight, and fetal loss rates were significantly higher in the study group (Table 2). Of note, the risk of small for gestational age among pregnancies undergoing second-trimester cordocentesis was 1.5 times higher than that in the control group (6.9% compared with 4.6%, P<.001, McNemar test). Likewise, the rates of preterm birth, early preterm birth, and overall fetal loss were significantly higher with relative risk of 1.7 (95% confidence interval [CI] 1.5–1.9), 1.9 (95% CI 1.4–2.6), and 1.9 (95% CI 1.4–2.7), respectively. Additionally, maternal outcomes such as preeclampsia, preterm rupture of membranes, placental abruption, and cesarean delivery rate were not significantly different between the two groups. Mean gestational age at delivery of the study group was slightly but significantly lower than that of the control group (38.1±2.2 compared with 38.3±1.7 weeks of gestation, mean difference −0.14, P value by paired t test <.001). Mean birth weight of the study group was slightly but significantly lower than that of the control group (2,971±515 compared with 3,051±468 g, mean difference −79 g, P value by paired t test <.001).
In subgroup analysis of risk factors of the poor fetal outcomes in the study group, univariate analysis showed that experience level of the operators, prolonged bleeding during performance, penetration of the placenta, the presence of fetal bradycardia, failure of the procedure, and difficulty of the procedure were significantly associated with small for gestational age as presented in Table 3. Logistic regression analysis showed that experience levels of the operators, prolonged bleeding during the procedures (range 61–410 seconds, mean 102.7±61.74 seconds, median 90 seconds), penetration through the placenta, and fetal bradycardia after the procedure were independent factors significantly associated with increased rates of the four poor outcomes (Table 4). We generated equations for predicting the probability of poor outcomes derived from the logistic regression analysis based on various risk factors during the performance of the cordocentesis (see the Appendix, available online at http://links.lww.com/AOG/A563).
Of note, the rates of fetal anomalies were not significantly different between the two groups. However, there were four fetuses with late-occurring anomalies including two fetuses with hydranencephaly and one fetus with late-occurring hydrocephalus in the study compared with one fetus with late-occurring hydrocephalus in the control group.
The main findings in this study is that rates of small for gestational age, preterm birth, low birth weight, and fetal loss are increased among pregnancies with cordocentesis in the second trimester. The cordocentesis-related fetal morbidity rates in this study are more likely to reflect the true situation than other published studies. This is because of the adequate sample size of properly matched patients from a single center and the same baseline characteristics between the two groups, especially no underlying maternal risk and fetal abnormalities in both groups that might have been at risk of a loss unrelated to the procedure. Additionally, cordocentesis in this study was uniquely performed at 16–22 weeks of gestation instead of 28–30 weeks of gestation as was done in most previous reports.4–10 Our results suggest that cordocentesis-related complications have a wide spectrum of severity, not just all or none between perfectly normal outcomes and fetal death. We observed that several cases of fetal loss occurred long after the procedure, several days or even several weeks, suggesting that sublethal complications secondary to cordocentesis may exist.
The difference in fetal loss rate between the study group (1.9%) and the control group (1%) was 1%, lower than in some studies2,3 but comparable to that reported by Weiner et al5 in which the mean gestational age at cordocentesis was 29 weeks. Of interest, four fetuses in the study groups had late-occurring cerebral anomalies including two fetuses with hydranencephaly and one with late-occurring hydrocephalus. Only one case was found in the control group. Although the events cannot reliably be linked to the procedure, it is possible that they represent a late consequence of thromboembolism secondary to cordocentesis, because fetal brain disruption after cordocentesis has been reported by Villo et al.14
The major factors significantly associated with adverse outcomes in this study include performance during the learning curve, penetration through the placenta, prolonged bleeding, bradycardia after the procedure, and an unsuccessful procedure at the first attempt. The equation of log odds derived from logistic regression analysis may be helpful in calculating the probability of adverse outcomes (Appendix, http://links.lww.com/AOG/A563). The strengths of this study include 1) large sample size; 2) high homogeneity of the women enrolled to the study in terms of gestational age, parity, maternal age, and socioeconomic status; 3) exclusion of the cases with known confounding factors for adverse outcomes; 4) cordocenteses performed in the same single center with the same techniques; and 5) control participants who were matched by same gestational age and before the study outcomes had occurred. Finally, some limitations of this study may include 1) the large number of women with loss to follow-up; 2) inability to discriminate between spontaneous and indicated preterm birth; and 3) no evaluation of neonatal morbidity and mortality. In conclusion, this cohort study compared the fetal outcomes between singleton pregnancies undergoing cordocentesis and control participants matched for age, race, parity, socioeconomic status, and gestational age at the time of recruitment. It was observed that second-trimester cordocentesis significantly increased rates of small for gestational age, preterm birth, low birth weight, and fetal loss.
1. Daffos F, Capella-Pavlovsky M, Forestier F. A new procedure for fetal blood sampling in utero: preliminary results of fifty-three cases. Am J Obstet Gynecol 1983;146:985–7.
2. Tongsong T, Wanapirak C, Kunavikatikul C, Sirirchotiyakul S, Piyamongkol W, Chanprapaph P. Fetal loss rate associated with cordocentesis at midgestation. Am J Obstet Gynecol 2001;184:719–23.
3. Ghidini A, Sepulveda W, Lockwood CJ, Romero R. Complications of fetal blood sampling. Am J Obstet Gynecol 1993;168:1339–44.
4. Weiner CP. Cordocentesis for diagnostic indications: two years' experience. Obstet Gynecol 1987;70:664–8.
5. Weiner CP, Okamura K. Diagnostic fetal blood sampling-technique related losses. Fetal Diagn Ther 1996;11:169–75.
6. Maxwell DJ, Johnson P, Hurley P, Neales K, Allan L, Knott P. Fetal blood sampling and pregnancy loss in relation to indication. Br J Obstet Gynaecol 1991;98:892–7.
7. Donner C, Simon P, Karioun A, Avni F, Rodesch F. Experience of a single team of operators in 891 diagnostic funipunctures. Obstet Gynecol 1994;84:827–31.
8. Donner C, Rypens F, Paquet V, Cohen E, Delneste D, van Regemorter N, et al.. Cordocentesis for rapid karyotype: 421 consecutive cases. Fetal Diagn Ther 1995;10:192–9.
9. Daffos F, Capella-Pavlovsky M, Forestier F. Fetal blood sampling during pregnancy with use of a needle guided by ultrasound: a study of 606 consecutive cases. Am J Obstet Gynecol 1985;153:655–60.
10. Boulot P, Deschamps F, Lefort G, Sarda P, Mares P, Hedon B, et al.. Pure fetal blood samples obtained by cordocentesis: technical aspects of 322 cases. Prenat Diagn 1990;10:93–100.
11. Tongsong T, Wanapirak C, Kunavikatikul C, Sirirchotiyakul S, Piyamongkol W, Chanprapaph P. Cordocentesis at 16–24 weeks of gestation: experience of 1,320 cases. Prenat Diagn 2000;20:224–8.
12. Tongsong T, Simaraks S, Sirivatanapa P, Sudasna J, Wanapirak C, Kunavikatikul C, et al.. Study of intrauterine growth from birthweight at Maharaj Nakhon Chiang Mai Hospital. J Med Assoc Thai 1993;76:482–6.
13. Wanapirak C, Tongsong T, Sirivatanapa P, Sa-nguansermsri T, Sekararithi R, Tuggapichitti A. Prenatal strategies for reducing severe thalassemia in pregnancy. Int J Gynaecol Obstet 1998;60:239–44.
14. Villó N, Beceiro J, Cebrero M, de Frias EG. Fetal brain disruption sequence in a newborn infant with a history of cordocentesis at 21 weeks gestation. Arch Dis Child Fetal Neonatal Ed 2001;84:F63–4.