Objective
We sought to evaluate whether testing for hereditary thrombophilia alone or in combination with second-trimester uterine artery Doppler (UAD) is useful in predicting recurrent complications in patients with previous preeclampsia, placental abruption, or stillbirth.
Study Design
Between 2001 and 2005, 110 consecutive women were included in the study and received 100 mg of aspirin daily. Adjustment was made for several maternal confounding factors using a logistic regression model.
Results
After multivariable logistic regression analyses, only abnormal UAD assessment was significantly associated with recurrent complications (odds ratio, 11.2; 95% confidence interval, 3.8–32.6) Combining the results of UAD and the presence of laboratory markers of thrombophilia failed to improve the accuracy of UAD to predict recurrent complications.
Conclusion
Hereditary thrombophilia testing is not useful in predicting recurrent complications in subsequent pregnancy.
Some inherited abnormalities of coagulation have been proved to be associated with an increased risk of venous thrombosis. In 1999, Kupferminc et al reported an increased frequency of these laboratory markers of genetic thrombophilia in women with pregnancies complicated by severe preeclampsia, abruption placentae, fetal growth retardation, or stillbirth. Since then, a large number of retrospective and case-controlled studies have been published and compiled in several reviews. However, the evidence to support a role for these biological markers in the pathogenesis of serious obstetric complications remains controversial with conflicting results from different studies.
The seventh American College of Chest Physicians (ACCP) conference on antithrombotic and thrombolytic therapy in 2004 and the consensus conference in France in 2003 have recommended screening for congenital thrombophilia in women with recurrent pregnancy loss and women with prior severe preeclampsia, abruptions, or unexplained intrauterine death. However, several questions remain unanswered after the publication of these recommendations. If the screening for congenital thrombophilia is positive, what should the clinician do with this information? We postulate that screening for laboratory markers of thrombophilia alone or in combination with second-trimester uterine artery Doppler (UAD) may be useful in predicting recurrent adverse obstetric outcomes in subsequent pregnancy. We conducted a prospective study to evaluate this hypothesis.
Materials and Methods
This study was prospective and conducted in 3 maternity hospitals in the Rouen, France, geographic area between 2001 and 2005. The study was approved by the Ethics Committee of Upper Normandy, and informed consent was obtained from each woman. Inclusion criteria were all pregnant patients with a documented history of severe preeclampsia, stillbirth, or abruption placentae reported during their immediate last pregnancy. Severe preeclampsia was defined in cases of severe hypertension (≥160 mm Hg systolic and/or ≥110 mm Hg diastolic blood pressure) and/or severe proteinuria (≥5 g/day), or in the presence of multiorgan involvement. Stillbirth was defined as fetal death >22 weeks’ gestation. Abruptio placentae was diagnosed on the basis of associated clinical criteria (vaginal bleeding, uterine tenderness, and fetal distress) with histologic analysis of the placenta to confirm the disease. Exclusion criteria were multiple pregnancies, chronic hypertension, chronic renal disease, diabetes preexisting to pregnancy, previous thromboembolic diseases, and immunological abnormalities associated with antiphospholipid syndrome. Exclusion criteria in cases of stillbirth were abnormal results of karyotyping, infectious diseases, and congenital anomalies detected at autopsy.
Eligible women were enrolled and received 100 mg of aspirin per day <16 weeks of gestation until 34 weeks of gestation as routinely prescribed for these patients at our institution. Patients were tested for antithrombin III, protein C, protein S, anticardiolipin antibodies, lupus anticoagulant, factor V Leiden mutation, prothrombin 20210A mutation, homozygous methylenetetrahydrofolate reductase (MTHFR) polymorphism, and homocysteinemia in cases of homozygous MTHFR polymorphism at Rouen University Hospital as described previously. Most patients had thrombophilia screening a few months after their first pregnancy complications, whereas others were tested early in the first trimester of subsequent gestation.
All patients had prenatal standardized care as routinely recommended in France. Women with thrombophilic disorders had graduated elastic compression stocking throughout the pregnancy and antithrombotic prophylaxis by low-molecular weight heparin (LMWH) during the postpartum period to reduce the risk of thromboembolism diseases. Patients with first-trimester miscarriage, hypertension or diabetes or isolated proteinuria diagnosed <20 weeks’ gestation, or multiple pregnancy were excluded from the study. Patients with antithrombotic prophylaxis prescribed any time during the pregnancy were also excluded from our study because that therapy might reduce adverse outcomes. Doppler assessment of the uterine arteries was performed between 22-28 weeks’ gestation as routine prenatal care for these patients at higher risk. An abnormal test result was represented by either a diastolic to systolic ratio (resistance index) >2 SD or the presence of a diastolic notch. We evaluated covariates such as gestational age at first delivery, interpregnancy interval, maternal age, body mass index, current smokers, UAD assessment, and hereditary thrombophilia testing. We categorized gestational age at delivery in the first pregnancy as <28, 29-34, and >34 weeks’ gestation. Body mass index was also categorized as normal (<25 kg/m 2 ), overweight (25-29.9 kg/m 2 ), and obese (≥30 kg/m 2 ).
Recurrent complications were considered in patients with preeclampsia defined by a blood pressure of at least 140/90 mm Hg with proteinuria >0.5 g per 24 hours, stillbirth defined as fetal death >22 weeks’ gestation, small for gestational age (SGA) defined as birthweight below the third centile depending on gestational age at birth and sex of the neonate according to the geographic pediatric criteria for the Upper Normandy population, and abruptio placentae diagnosed on the basis of clinical criteria and histologic analysis of the placenta.
Continuous data are reported as mean ± SD or as median (range). The Student t , Mann-Whitney, and χ 2 tests were used for statistical analysis as appropriate. The logistic regression model determined independent predictors of recurrent obstetric complications after the control for confounding variables. The variables presenting a particular clinical interest or P < .20 in univariate analysis were included in the model. Sensitivity, specificity, and positive and negative predictive values and likelihood ratios (LRs) were calculated with 95% confidence intervals (CIs) for diagnostic tests. Receiver operating characteristic (ROC) curves were constructed and the point farthest from the line of nondiscrimination (45-degree line) was chosen as the cut-off value. Areas under 2 ROC curves were statistically compared by calculating the standardized derivate value ( Z value) of the difference between the areas. A difference of P < .05 was considered statistically significant. The analysis was performed using software (Stata 9.0; Stata Corp, College Station, TX).
Results
During the study period 126 consecutive women were initially recruited but 16 of them were later excluded. Secondary exclusion criteria were patients who received antithrombotic prophylaxis (n = 6, 3 of them with thrombophilic disorders), patients with incomplete thrombophilic tests (n = 5), patients with antiphospholipid antibodies (n = 2), and various causes (n = 3, prematurely stopped aspirin treatment, termination of pregnancy for cystic fibrosis, and a neonate with cardiac defect). A total of 110 patients were eligible for the study. Previous vascular placental diseases were represented by severe preeclampsia (n = 74), stillbirth (n = 19), or abruptio placentae (n = 17). Mean gestational age at delivery for the pregnancy index was 32.1 weeks ± 4.5. Thrombophilic disorders were diagnosed in 29 (26.4%) patients: factor V Leiden mutation (n = 8, 1 patient with associated hyperhomocysteinemia and another with homozygous MTHFR mutation), prothrombin G20210A mutation (n = 7, 2 patients with associated homozygous MTHFR mutation), homozygous MTHFR mutation (n = 12, 7 patients with associated hyperhomocyteinemia), and protein S deficiency (n = 2). Demographic and clinical characteristics of subsequent pregnancy are reported in Table 1 with a total recurrent complication rate of 30.9% (n = 34). Recurrent complications were preeclampsia (n = 19), SGA (n = 10), and abruption placentae (n = 5). Variables associated with recurrent complications in univariate analyses are reported in Table 2 . After multivariable logistic regression analyses, only abnormal UAD assessment was significantly associated with recurrent complications (odds ratio, 11.2; 95% CI, 3.8–32.6; P = .001). Similar results were obtained in patients with only previous preeclampsia (data not shown). Abnormal uterine Doppler assessment was significantly associated with an earlier gestational age at delivery in previous pregnancy (30.2 weeks ± 0.6 vs 33.2 weeks ± 0.5; P = .001) and no association was found with interpregnancy interval, maternal age, body mass index, current smokers, and thrombophilic disorders. No significant association was found among thrombophilic disorders, gestational age at delivery in previous and subsequent pregnancy, and neonatal outcomes. Accuracy of UAD, thrombophilic disorders, and UAD or thrombophilic disorders in predicting recurrent complications are reported in Table 3 . UAD assessment was accurately classified in 76.3% of patients as regards recurrent complications whereas combining UAD assessment and the presence of thrombophilic disorders in a further diagnostic test was only obtained in 61.8% of patients. The area under the ROC curve as regards the probability of recurrent complications depending on the UAD assessment was significantly higher than that of the ROC curve depending on the combination of UAD assessment and presence of laboratory markers of thrombophilia (0.74; 95% CI, 0.64–0.83 vs 0.65; 95% CI, 0.55–0.74; P = .005).
