Study population

In this retrospective case-control study, blood samples were prospectively collected from women with singleton pregnancies diagnosed antenatally with IUGR following written informed consent, between November 2004 and August 2007 at BC Women’s Hospital in Vancouver, Canada. Ethics approval was granted by the University of British Columbia Children’s and Women’s Health Centre Research Ethics Boards. Other data pertaining to this cohort of women have previously been published. Eligible women were consecutively recruited from inpatient and outpatient services at BC Women’s Hospital. Women were excluded if they were in active labor at the time of eligibility or were within 48 hours of antenatal betamethasone administration.

The antenatal diagnosis of IUGR was defined as fetal AC less than the 10th percentile for gestational age identified by antenatal ultrasound. Maternal blood samples were collected at this time but were not included into the study until after delivery. Inclusion required either birth weight less than the fifth percentile for gestational age at delivery and sex or birthweight less than the 10th percentile with either uterine artery Doppler notching at 22 ⁺⁰ to 24 ⁺⁰ weeks’ gestation, absent/reversed umbilical artery end diastolic flow, or oligohydramnios (amniotic fluid index <50 mm) documented during pregnancy.

A total of 19 confirmed low-birthweight cases were recruited. Two were excluded from the primary analysis because of IUGR being attributed to fetal congenital anomalies confirmed at delivery (one Treachers-Collins syndrome, one Cornelia de Lange syndrome). One additional case was excluded because of IUGR of unknown origin. During this pregnancy, the women had an appendectomy at 10 weeks’ gestation, followed by cerclage for cervical incompetence at 12 weeks. Preterm spontaneous rupture of membranes occurred at 29 ⁺² weeks followed by the delivery of a live fetus 3 weeks later (birthweight less than the 10th percentile). Placenta pathology showed chorioamnionitis and funisitis but no evidence of abnormal placentation. No infection was noted.

The remaining 16 cases were subgrouped based on the presence or absence of placental IUGR. Placental IUGR (n = 9) was defined as IUGR with an abnormal placental pathology report. Abnormal placenta pathology included decidual necrosis, adherent thrombus, low placental weight, fibrin deposition, fetal surface vessel thrombosis, decreased fetal vascularization, decidual vasculopathy, calcification, infarcts, intervillous thrombus, and advanced/delayed villous maturation and villitis.

A constitutionally small fetus (n = 7) was defined as a fetus clinically diagnosed antenatally with IUGR and with a normal placental pathology report. Although not included in the primary analysis, the 3 syndromic IUGR cases all had normal placental pathology reports. Placental pathology status was determined by a perinatologist blinded to all aspects of the woman’s history, pregnancy course, and pregnancy outcomes.

A total of 79 non-smoking women served as normal pregnancy controls. These women had no documented concerns of hypertension, proteinuria, gestational diabetes or IUGR during their pregnancy. Normal pregnancy controls were matched for maternal age (±5 years), gestational age (±2 weeks), and parity (0, 1, ≥2).

Sample collection and PlGF analysis

Maternal venous blood was collected antenatally in EDTA tubes in the standard fashion. Plasma was obtained through centrifugation and samples were frozen at –80°C. Laboratory staff was blinded to the clinical diagnosis of all women. Samples were analyzed for PlGF using the automated Triage PlGF assay kit (Alere San Diego, San Diego, CA) according to the product insert. This immunoassay uses fluorescently labeed murine monoclonal antibodies against PlGF for PlGF quantification. Briefly, plasma is thawed to room temperature and mixed by inversion. A total of 250 μL of thawed plasma is pipetted into the sample port of a PlGF test cartridge. The cartridge is inserted into the Triage meter. The results are displayed on the meter in approximately 15 minutes in picograms per milliliter.

The cartridge contains chemistries for on-board positive and negative control systems. Control systems at the level of the cartridge and meter ensure that the quantitative PlGF result is valid. Calibration information is supplied by the manufacturer in the form of a lot-specific EPROM chip that is contained within each kit of devices. The measurable range of the assay is 12–3000 pg/mL. Concentrations less than 12 pg/mL are value assigned based on the calibration curve, but this value is displayed to the user as a qualitative result less than 12 pg/mL. A positive test was defined as a PlGF concentration below the fifth percentile for gestational age of normal controls, as described in the product insert and derived from Knudsen et al.

Samples were batch assayed to minimize any effect of interassay variability.

Statistics

Data were analyzed using Prism 4.0 (GraphPad, San Diego, CA). Descriptive data are expressed as median and interquartile range (IQR) for nonnormally distributed data. A χ ² or Fisher’s exact test was used for the comparison of categorical variables. A Kruskal-Wallis analysis of variance or Mann-Whitney U test was used for continuous variables.

The primary outcome of a positive test identifying placental IUGR was evaluated using a 2 × 3 contingency table (positive vs negative tests for each group) and a test of association (the Freeman-Halton test for the complete 2 × 3 table and the Fisher exact test of the relevant 2 × 2 subtable). Sensitivity, specificity, positive predictive values (PPVs) and negative predictive values (NPVs) for a positive PlGF test identifying placental IUGR from constitutionally small fetuses were calculated with 95% confidence intervals as a secondary analysis. We performed this analysis with and without the 3 syndromic IUGR cases excluded from the primary analysis.

PPV and NPV were calculated to characterize the 2 × 2 tables, despite the artificial prevalence in this case-control study. Also, qualitative PlGF results less than 12 pg/mL were set equal to 12 pg/mL for the purpose of statistical analysis. This approximation does not affect the reported test performance in terms of clinical sensitivity, or specificity, but may slightly underreport the significance of the difference between groups in the Kruskal-Wallis analysis.

Differences with P < .05 were judged to be statistically significant. The STARD (Standards for Reporting of Diagnostic Accuracy) Initiative guidelines were consulted throughout the design and analysis of this study.