Participants and trial design

A randomized, multicenter, parallel, open-label clinical trial was undertaken between August 2010 and August 2012 in 8 tertiary teaching hospitals in Latin America with experience in the management of severe preeclampsia. The hospitals were selected among 21 others because of their ability to immediately respond to obstetric emergencies and the presence of neonatal specialized care 24 hours a day. These hospitals also had a track record of conducting clinical studies.

Pregnant women with singleton or twin pregnancy and severe hypertensive disorders at 28-33 weeks of gestation were invited to take part in the Expectant Management of Preeclampsia, or MEXPRE (Manejo Expectante de Preeclampsia) Latin study. Pregnancies less than 28 weeks were excluded because the neonatal mortality for those babies in the intensive care units of the hospitals involved is high and subject to limited resources.

Severe preeclampsia was diagnosed if there was presence of blood pressure of 140/90 mm Hg or greater on 2 occasions at least 4 hours apart and 0.3 g or greater of protein in a 24 hour urine specimen with 1 or more of the following additional criteria: blood pressure greater than 160 mm Hg systolic or greater than 110 mm Hg diastolic; proteinuria of at least 5 g per 24 hours; or symptoms suggesting significant end-organ involvement, such as headache, visual disturbances, epigastric pain, or tinnitus.

A diagnosis of severe gestational hypertension was made with a presence of blood pressure of 140/90 mm Hg or greater on 2 occasions at least 4 hours apart and less than 0.3 g of protein in a 24 hour urine specimen with 1 or more of the following additional criteria: blood pressure greater than 160 mm Hg systolic greater than 110 mm Hg diastolic; or symptoms suggesting significant end-organ involvement, such as headache, visual disturbances, epigastric pain, or tinnitus.

Chronic hypertension was defined as hypertension present before pregnancy or before the 20th week of gestation. Superimposed preeclampsia in women with chronic hypertension was defined as the development of new-onset proteinuria (excretion of greater than 300 mg of protein over 24 hours), with 1 or more of the following criteria: blood pressure greater than 160 mm Hg systolic or greater than 110 mm Hg diastolic; or symptoms suggesting significant end-organ involvement, such as headache, visual disturbances, epigastric pain, or tinnitus.

Exclusion criteria were eclampsia, HELLP syndrome (hemolysis, elevated liver enzymes, and low platelet count), preeclampsia with renal failure or pulmonary edema, active vaginal bleeding, ruptured membranes, placenta previa, diabetes mellitus or gestational diabetes, preexisting renal disease, or autoimmune disease. In addition, major fetal abnormalities, fetal growth restriction (fetal estimated weight below the 10th percentile for gestational age determined by ultrasound), oligohydramnios, and reverse umbilical artery Doppler flow. All patients had an ultrasound before enrollment to rule out fetal growth restriction.

Eclampsia was defined as generalized convulsions in pregnancy not caused by epilepsy or other causes and HELLP syndrome as platelet count of 150,000 or less and aspartate aminotransferase of 70 units/L or greater, alanine aminotransferase of 40 units/L or greater, and lactate dehydrogenase of 600 units/L or greater.

Pulmonary edema was defined as tachypnea greater than 40 per minute, gas diffusion deficit, and compatible chest X-ray and renal insufficiency as urine output less than 500 mL/d, serum creatinine greater than 1.2 mg/dL, and creatinine clearance less than 20 mL/min.

Disseminated intravascular coagulation was defined as the presence of 3 or more of the following criteria: low platelet count (<100,000 cells/mL), low fibrinogen concentration (<300 mg/dL), presence of D-dimers (≥40 mg/dL), or prolonged prothrombin time (≥14 seconds), and partial thromboplastin time (≥40 seconds). Oliguria was defined as the urine output less than less than 0.5 mL (kilograms per hour) or less than 100 mL per 4 hours. Gestational age was judged from the menstrual history and confirmed by measurement of fetal crown-rump length at first-trimester scan.

The diagnosis of small for gestational age (SGA) was defined as birthweight less than the 10th percentile for gestational age. Abruptio placentae was defined as clinical/pathology diagnosis of retroplacental hematoma at delivery.

Stillbirth or fetal death was defined as death in utero and neonatal death as death from birth to 28 days after birth. Perinatal death was defined as fetal plus neonatal deaths.

The ethics committees or institutional boards for all participating hospitals approved the protocol. Trial coordinators regularly undertook quality control of screening data handling and verification of adherence to the protocols at the different hospitals.

Randomization and masking

After written informed consent was obtained from women, they were randomly allocated to the prompt delivery group (PD) or expectant management group (EXM) in a 1:1 ratio.

All patients were extensively counseled regarding the maternal and perinatal risks of expectant or early delivery management. The principal investigator made the assignment centrally for each hospital by using a computer-generated code with variable block size of 4 and 6, with concealment of allocation by sealed envelopes. The trial coordinator or recruiters in each hospital did not have access to the randomization sequence. This study was open label because of the nature of the intervention.

Interventions

In the PD group, the patients received glucocorticoid therapy followed by delivery in 24-72 hours; in the EXM group, the women were treated expectantly (glucocorticoid therapy followed by delivery only for specific maternal/fetal indications or reaching 34 weeks of gestation).

Antepartum management included bed rest; to prevent and control seizure, all women initially received magnesium sulfate as a 4 g intravenous loading dose followed by 1 g intravenous per hour for 24-48 hours. In the PD group, magnesium sulfate was continued until 24 hours after delivery. Bolus doses of hydralazine, labetalol, or oral nifedipine were administered to control severe hypertension (≥160/110 mm Hg). Four doses of 6 mg of dexamethasone intramuscularly given 12 hours apart were administered or 2 doses of 12 mg of betamethasone intramuscularly given 24 hours apart. Maternal evaluation included frequent blood pressure measurements along with questioning for symptoms of worsening preeclampsia.

Oral antihypertensive medications were used only in some hospitals, in those women with severe hypertension to maintain systolic blood pressure (BP) below 160 mm Hg and diastolic BP less than 110 mm Hg, after use of intravenous bolus doses of hydralazine, labetalol, or oral nifedipine to control severe hypertension. The oral antihypertensive drugs used were α-methyl dopa, nifedipine, or hidralazine. The administration of oral antihypertensives after the acute treatment of severe hypertension was at the discretion of the treating physician and varied between hospitals.

Laboratory evaluation included serial measurement of liver function tests, complete blood cell count, coagulation profile, and renal function tests at admission and each day for 48 ± 72 hours in both groups; in the expectant group, the laboratory evaluation continued twice per week.

Fetal status was assessed with a non-stress test each day and ultrasound to rule out fetal growth and restriction, and assessment of amniotic fluid was performed at admission prior to randomization and then serially. Ultrasounds to evaluate fetal growth were performed every 1-2 weeks according to findings as long as the patient remained pregnant in the expectant group.

Specific indications for delivery in the expectant group were uncontrollable blood pressure, nonreassuring fetal heart rate tracing, abnormal fetal testing, fetal growth restriction (fetal weight below the 10th percentile for gestational age determined by ultrasound), abruption placentae, decline in renal function, the HELLP syndrome, persistent severe headache or visual changes, or epigastric pain and 34 weeks of gestation. Magnesium sulfate was restarted when delivery was indicated and continued for 24 hours after delivery.

The investigators and the neonatologist who followed up the infants prospectively collected data on the pregnant women and their newborns on a standardized form. Outcome data included maternal demographics and maternal and neonatal outcomes. Demographic data included maternal age, gestational age at diagnosis, days of expectant management, gestational age at delivery, the indication for delivery, and the mode of delivery. An attempt was made to blind the treatment allocation to the data abstracter, especially to the neonatologist.

Maternal morbidities included placental abruption, pulmonary edema, the HELLP syndrome, renal insufficiency, eclampsia, disseminated intravascular coagulation, and death.

Neonatal data included fetal or neonatal death (perinatal death), birthweight, Apgar scores, neonatal intensive care unit admissions and length of stay, composite neonatal morbidities (respiratory distress syndrome [RDS], intraventricular hemorrhage [IVH] grade III and IV, necrotizing enterocolitis, neonatal sepsis).

Statistical analysis

The primary outcome was perinatal mortality (fetal and neonatal death). Secondary outcomes were composite neonatal morbidities (RDS, IVH, necrotizing enterocolitis, neonatal sepsis), neonatal data (birthweight, SGA, Apgar scores, neonatal intensive care unit admissions and length of stay), and maternal morbidities (placental abruption, pulmonary edema, the HELLP syndrome, renal insufficiency, eclampsia, disseminated intravascular coagulation) and death.

The sample size was calculated, assuming a 15% rate of perinatal mortality (fetal death and neonatal death at ≤28 days) in the prompt delivery group, a 2-sided α error of 5%, and 80% power to detect a reduction of 30% in the rate of perinatal mortality with the expectant management. The high rate of perinatal mortality assumed in the study reflects the reality in Latin American countries. To detect that difference, 260 patients with 130 in each group needed to be recruited.

The baseline characteristics of the participants in each group were compared with a Student t test for continuous variables with normal distribution and a Mann-Whitney U test for discrete and ordinal variables or those with nonnormal distribution. Univariate comparisons of dichotomous data were done with a Fisher exact test. The P values for all hypotheses were 2 sided, and values of P < .05 were judged to be significant. Statistical analysis was performed with Epi Info software version 7 (Centers for Disease Control and Prevention, Atlanta, GA), under the intention-to-treat principle. The trial is registered (ClinicalTrials.gov no. NCT01164852 ).