Subjects

Women were consecutively recruited from antenatal and ultrasound clinics (performing nuchal translucency scans) at Royal Prince Alfred Hospital, Sydney, Australia. At this time, nuchal translucency scans were offered to all women who consented. All women <14 weeks’ gestation (measured by ultrasound) with a singleton pregnancy were included in the study. Written informed consent was obtained from all participants. The women were prospectively followed up longitudinally throughout their pregnancy.

Women were categorized into 3 groups at time of delivery: group 1 had PE, group 2 had GHT, and group 3 were normotensive and acted as the control group. Women were considered to have PE or GHT as defined by the International Society for the Study of Hypertension in Pregnancy guideline of 2001. Briefly, PE was defined as blood pressure ≥140 mm Hg systolic and/or ≥90 mm Hg diastolic >20 weeks’ gestation on 2 occasions at least 4 hours apart. This was associated with evidence of fetal growth restriction and end-organ dysfunction, including proteinuria (≥2+ on dipstick or 300 mg/24 h) or renal insufficiency (serum creatinine >0.09 mmol/L), liver disease (raised serum transaminases or severe epigastric pain), neurological problems (including convulsions), and hematological disturbances (thrombocytopenia, disseminated intravascular coagulation, or hemolysis), all of which resolved within 3 months postpartum. GHT was defined as women with blood pressure ≥140 mm Hg systolic and/or ≥90 mm Hg diastolic >20 weeks’ gestation.

This study was approved by the South Western Sydney Human Ethics Committee.

Data collection

Baseline clinical data were collected for all women recruited into the study at the first antenatal visit. Baseline data included maternal age at time of delivery, height (measured in meters), weight (measured in kilograms), and body mass index (BMI) (calculated as weight/height squared). Detailed personal, family medical, and obstetric history (in particular chronic infections, hypertension, PE, miscarriages, termination) along with results of routine infectious serology (hepatitis B and C antibody testing, syphilis, and HIV), blood group, smoking history, medications taken prior to and during pregnancy, illicit drug usage, and previous delivery methods were documented.

Clinical data were collected throughout the pregnancy at time points 11-14, 16-20, 26-28, and 34-36 weeks’ gestation; term; and 6 weeks’ postnatal. Data collected included blood pressure (measured by midwife or obstetrician in antenatal clinic, sitting down after 5-minute rest) and results of antenatal investigations such serum PAPP-A, BHCG levels, glucose tolerance, and nuchal translucency. At the time of delivery the following data were collected: gestation at delivery, weight (grams at birth), infant length (centimeters), infant sex, infant Apgar scores at 1 and 5 minutes, requirement of any neonatal supportive care (eg, intensive care, or high-dependency monitoring), delivery method, and any complications that occurred during the delivery (eg, postpartum hemorrhage). In the postpartum period the data collected included: blood pressure, breast-feeding status, and postdelivery complications ( Table 1 ).

Sample collection and processing

Maternal blood was collected at 6 time points: 11-14, 16-20, 26-28, and 34-36 weeks’ gestation; term; and 6 weeks’ postnatal. Venous blood was collected into 2 tubes: a polyethylene terephthalate tube containing a clot activator (silica particles) for serum samples (sFlt-1, PlGF, sEng) and a polyethylene terephthalate tube containing 7.2 mg of potassium ethylenediaminetetraacetic acid (BD, Sydney, Australia) for plasma samples (AT1AA). The blood was centrifuged, within 4 hours of venipuncture, for 10 minutes at 3000 rpm (GS-6 Centrifuge, Beckman Coulter Inc, Brea, CA). The serum was then aliquoted (250 μL) and stored at –80°C until required.

Enzyme-linked immunosorbent assay

Concentrations of sFlt-1 PlGF, sEng, and AT1AA were measured using commercial enzyme-linked immunosorbent assay (ELISA) kits (sFlt-1, Bender Medsystems, Vienna, Austria; PlGF and sEng, R&D Systems, Minneapolis, MN; and AT1AA, One Lambda, Canoga Park, CA). All ELISAs were performed according to the manufacturer’s instructions, in duplicate. If the coefficients of variation (CV) was ≥10%, the results of the ELISA were discarded and the entire ELISA plate repeated. The interassay and intraassay CV were 5.8% and 9.9%, respectively, for sFlt-1; 2.3% and 7.7%, respectively, for PlGF; 3.8% and 10.8%, respectively, for sEng; and 3.29% and 11.41%, respectively, for AT1AA. As reported by the manufacturers the intraassay precision (CV) for 10 identical samples was 3.9% and the interassay precision (CV) for 20 identical samples was 5.1%. The lower limit of detection for the AT1AA assay was 2.5 U/mL. In the current study the AT1AA ELISA we utilized has undergone robust scientific validation by Dragun, who first used the kit in the renal transplant population, however, the kit has not been validated in pregnancy.

Statistical methods

Statistical analysis was carried out using software (SPSS, Version 23; IBM Corp, Armonk, NY). The distribution of the data was analyzed using the Kolmogorov-Smirnov test and, where applicable, the data were logarithmically transformed. The data are presented as mean ± SEM. In cases where the data could not be normalized, nonparametric tests such as Kruskal-Wallis or Mann Whitney U tests were applied. In this case the data are presented as median (interquartile range). Although the overall significance level was set at .05, the sharpened Bonferroni method was used to adjust the individual alpha level when multiple comparisons were performed. A sFlt-1 to PlGF ratio was calculated by dividing the sFlt-1 concentrations by the PlGF concentrations at various time points for each individual patient. The sample size was powered on a 50% difference in protein concentrations at 12 weeks specifically designed to look at mechanisms for predicting PE. The AT1AA assay was subsequently available and a nested case-control study was conducted based on sample availability.

An analysis of repeated measures was employed to assess the changes of the markers over time and to assess between-subject effects. Cross-sectional analysis at a single time point was performed by Kruskal-Wallis tests. Bivariate correlation analysis using Spearman tests were conducted to assess relationships between the markers and demographic risk factors. A univariate analysis was conducted to assess whether individual demographic or serum factors were predictive of disease. Diagnosis was grouped for this analysis to include all hypertensive disorders of pregnancy (GHT and PE).

Subjects

Women were consecutively recruited from antenatal and ultrasound clinics (performing nuchal translucency scans) at Royal Prince Alfred Hospital, Sydney, Australia. At this time, nuchal translucency scans were offered to all women who consented. All women <14 weeks’ gestation (measured by ultrasound) with a singleton pregnancy were included in the study. Written informed consent was obtained from all participants. The women were prospectively followed up longitudinally throughout their pregnancy.

Women were categorized into 3 groups at time of delivery: group 1 had PE, group 2 had GHT, and group 3 were normotensive and acted as the control group. Women were considered to have PE or GHT as defined by the International Society for the Study of Hypertension in Pregnancy guideline of 2001. Briefly, PE was defined as blood pressure ≥140 mm Hg systolic and/or ≥90 mm Hg diastolic >20 weeks’ gestation on 2 occasions at least 4 hours apart. This was associated with evidence of fetal growth restriction and end-organ dysfunction, including proteinuria (≥2+ on dipstick or 300 mg/24 h) or renal insufficiency (serum creatinine >0.09 mmol/L), liver disease (raised serum transaminases or severe epigastric pain), neurological problems (including convulsions), and hematological disturbances (thrombocytopenia, disseminated intravascular coagulation, or hemolysis), all of which resolved within 3 months postpartum. GHT was defined as women with blood pressure ≥140 mm Hg systolic and/or ≥90 mm Hg diastolic >20 weeks’ gestation.

This study was approved by the South Western Sydney Human Ethics Committee.

Data collection

Baseline clinical data were collected for all women recruited into the study at the first antenatal visit. Baseline data included maternal age at time of delivery, height (measured in meters), weight (measured in kilograms), and body mass index (BMI) (calculated as weight/height squared). Detailed personal, family medical, and obstetric history (in particular chronic infections, hypertension, PE, miscarriages, termination) along with results of routine infectious serology (hepatitis B and C antibody testing, syphilis, and HIV), blood group, smoking history, medications taken prior to and during pregnancy, illicit drug usage, and previous delivery methods were documented.

Clinical data were collected throughout the pregnancy at time points 11-14, 16-20, 26-28, and 34-36 weeks’ gestation; term; and 6 weeks’ postnatal. Data collected included blood pressure (measured by midwife or obstetrician in antenatal clinic, sitting down after 5-minute rest) and results of antenatal investigations such serum PAPP-A, BHCG levels, glucose tolerance, and nuchal translucency. At the time of delivery the following data were collected: gestation at delivery, weight (grams at birth), infant length (centimeters), infant sex, infant Apgar scores at 1 and 5 minutes, requirement of any neonatal supportive care (eg, intensive care, or high-dependency monitoring), delivery method, and any complications that occurred during the delivery (eg, postpartum hemorrhage). In the postpartum period the data collected included: blood pressure, breast-feeding status, and postdelivery complications ( Table 1 ).

Sample collection and processing

Maternal blood was collected at 6 time points: 11-14, 16-20, 26-28, and 34-36 weeks’ gestation; term; and 6 weeks’ postnatal. Venous blood was collected into 2 tubes: a polyethylene terephthalate tube containing a clot activator (silica particles) for serum samples (sFlt-1, PlGF, sEng) and a polyethylene terephthalate tube containing 7.2 mg of potassium ethylenediaminetetraacetic acid (BD, Sydney, Australia) for plasma samples (AT1AA). The blood was centrifuged, within 4 hours of venipuncture, for 10 minutes at 3000 rpm (GS-6 Centrifuge, Beckman Coulter Inc, Brea, CA). The serum was then aliquoted (250 μL) and stored at –80°C until required.

Enzyme-linked immunosorbent assay

Concentrations of sFlt-1 PlGF, sEng, and AT1AA were measured using commercial enzyme-linked immunosorbent assay (ELISA) kits (sFlt-1, Bender Medsystems, Vienna, Austria; PlGF and sEng, R&D Systems, Minneapolis, MN; and AT1AA, One Lambda, Canoga Park, CA). All ELISAs were performed according to the manufacturer’s instructions, in duplicate. If the coefficients of variation (CV) was ≥10%, the results of the ELISA were discarded and the entire ELISA plate repeated. The interassay and intraassay CV were 5.8% and 9.9%, respectively, for sFlt-1; 2.3% and 7.7%, respectively, for PlGF; 3.8% and 10.8%, respectively, for sEng; and 3.29% and 11.41%, respectively, for AT1AA. As reported by the manufacturers the intraassay precision (CV) for 10 identical samples was 3.9% and the interassay precision (CV) for 20 identical samples was 5.1%. The lower limit of detection for the AT1AA assay was 2.5 U/mL. In the current study the AT1AA ELISA we utilized has undergone robust scientific validation by Dragun, who first used the kit in the renal transplant population, however, the kit has not been validated in pregnancy.

Statistical methods

Statistical analysis was carried out using software (SPSS, Version 23; IBM Corp, Armonk, NY). The distribution of the data was analyzed using the Kolmogorov-Smirnov test and, where applicable, the data were logarithmically transformed. The data are presented as mean ± SEM. In cases where the data could not be normalized, nonparametric tests such as Kruskal-Wallis or Mann Whitney U tests were applied. In this case the data are presented as median (interquartile range). Although the overall significance level was set at .05, the sharpened Bonferroni method was used to adjust the individual alpha level when multiple comparisons were performed. A sFlt-1 to PlGF ratio was calculated by dividing the sFlt-1 concentrations by the PlGF concentrations at various time points for each individual patient. The sample size was powered on a 50% difference in protein concentrations at 12 weeks specifically designed to look at mechanisms for predicting PE. The AT1AA assay was subsequently available and a nested case-control study was conducted based on sample availability.

An analysis of repeated measures was employed to assess the changes of the markers over time and to assess between-subject effects. Cross-sectional analysis at a single time point was performed by Kruskal-Wallis tests. Bivariate correlation analysis using Spearman tests were conducted to assess relationships between the markers and demographic risk factors. A univariate analysis was conducted to assess whether individual demographic or serum factors were predictive of disease. Diagnosis was grouped for this analysis to include all hypertensive disorders of pregnancy (GHT and PE).