Research

www. AJOG.org

OBSTETRICS

First-trimester placental protein 13 screening for preeclampsia and intrauterine growth restriction Ilana Chafetz, MSc; Ido Kuhnreich, MSc; Marei Sammar, PhD; Yossi Tal, PhD; Yair Gibor, MD, PhD; Hamutal Meiri, PhD; Howard Cuckle, PhD; Myles Wolf, MD, MMSc OBJECTIVE: The purpose of this study was to evaluate first-trimester serum placental protein 13 (PP13) as a screening test for preeclampsia and intrauterine growth restriction (IUGR). STUDY DESIGN: We performed a prospective, nested case-control study in the Massachusetts General Hospital Obstetric Maternal Study. PP13 was measured by solid-phase sandwich enzyme-linked immunosorbent assay in serum samples that were collected at the first prenatal visit (9-12 weeks of gestation) from women who subsequently experienced preeclampsia (n ⫽ 47), IUGR (n ⫽ 42), or preterm delivery (n ⫽ 46). Women with uncomplicated term deliveries served as control subjects (n ⫽ 290) and were matched to cases by gestational age when serum was collected and for the duration of specimen storage.

(86.6 pg/mL; P ⬍ .001), and preterm delivery (84.9 pg/mL; P ⫽ .007). When PP13 was expressed as multiples of the gestational agespecific medians among the control subjects, the multiples of the medians were 0.2 for preeclampsia, 0.6 for IUGR, and 0.6 for preterm delivery (P ⬍ .001 for each disorder compared with control subjects). Receiver operating characteristic analysis yielded areas under the curve of 0.91, 0.65, and 0.60 for preeclampsia, IUGR, and preterm delivery, respectively. At a 90% specificity rate, the corresponding sensitivities were 79%, 33%, and 28%, respectively. CONCLUSION: The screening of maternal PP13 levels in the first trimester is a promising diagnostic tool for the prediction of preeclampsia with high sensitivity and specificity.

RESULTS: The median first-trimester PP13 level was 132.5 pg/mL in

the control subjects. Median PP13 levels were significantly lower among women who had preeclampsia (27.2 pg/mL; P ⬍ .001), IUGR

Key words: intrauterine growth restriction, placental protein 13, preeclampsia

Cite this article as: Chafetz I, Kuhnreich I, Sammar M, et al. First-trimester placental protein 13 screening for preeclampsia and intrauterine growth restriction. Am J Obstet Gynecol 2007;197:35.e1-35.e7.

A

bnormal placental development with placental insufficiency is associated with the development of preeclampsia, intrauterine growth restric-

From Diagnostic Technologies Ltd, Haifa, Israel (Ms Chafetz and Drs Sammar, Gibor, and Meiri); TechnoStat Ltd, Kfar Saba, Israel (Mr Kuhnreich and Dr Tal); the Department of Reproductive Epidemiology, University of Leeds School of Medicine, Leeds, England (Dr Cuckle); and the Renal Unit, Department of Medicine, Massachusetts General Hospital, Harvard Medical School, Boston, MA (Dr Wolf). Reprints: Myles Wolf, MD, MMSc, Renal Unit, Department of Medicine, Harvard Medical School, Bulfinch 127, Massachusetts General Hospital, 55 Fruit St, Boston, MA 02114; [email protected] Sponsored in part by National Institutes of Health grant RR017376 (M.W.) and grants 31851 and 42872 from Israel Chief Scientist (H.M.). 0002-9378/$32.00 © 2007 Mosby, Inc. All rights reserved. doi: 10.1016/j.ajog.2007.02.025

tion (IUGR), and preterm delivery, thereby contributing substantially to fetal, neonatal, and maternal morbidity.1,2 Preeclampsia affects 5%-6% of pregnancies and accounts for 18% of maternal deaths during pregnancy in the United States.3 IUGR, which complicates 3%-7% of moderately preterm (33-36 weeks of gestation) and 5%-14% of very preterm (20-32 weeks of gestation) newborn infants,4 is associated with increased perinatal morbidity and prolonged neonatal hospitalization, which incurs substantial cost.5,6 Women with a history of preeclampsia are at increased risk of the development of future cardiovascular disease.7,8 Although there is currently no evidence that early prediction of increased risk of preeclampsia and IUGR will lead to improved outcomes for these disorders, predictive markers that stratify a woman’s risk for these adverse pregnancy outcomes are needed to improve the design of future clinical trials. Several

promising candidates that have been identified include uterine artery ultrasonography and maternal serum/urinary levels of human chorionic gonadotropin, inhibin A, activin A, pregnancyassociated plasma protein A, sex hormone-binding globulin, placental growth factor, and soluble fms-like tyrosine kinase 1.9-18 Nonetheless, early pregnancy screening for preeclampsia and IUGR remains insufficient, and randomized controlled trials that used biomarkers to identify high-risk women have been disappointing perhaps because the sensitivity of most of these markers is high in the second trimester, long after the placental dysfunction that culminates in clinical disease is already established. Placental protein 13 (PP13) is a 32-kd dimer protein that is produced only in the placenta and is thought to be involved in normal implantation and maternal artery remodeling.19-22 Studies with real-time polymerase chain reac-

JULY 2007 American Journal of Obstetrics & Gynecology

35.e1

Research

Obstetrics

tion revealed a reduced level of PP13 messenger RNA expression in placentas that were obtained from pregnant women with preeclampsia who delivered preterm (26-34 weeks of gestation).22 In a recent prospective, nested-case control study, PP13 levels that were measured at 11-13 weeks of gestation were decreased significantly among 10 women who experienced early preeclampsia that necessitated delivery at ⬍34 weeks of gestation, compared with 423 matched control subjects.23 Our objective was to test the hypothesis that PP13 testing that is performed earlier in the first trimester will identify women who are at risk for diseases that are associated with placental dysfunction, specifically preeclampsia and IUGR.

M ETHODS Study design and population We performed a nested case-control study within the Massachusetts General Hospital Obstetric Maternal Study, which was a prospective cohort study that began in 1998 and completed enrollment in 2006.15,16,24 All women who received prenatal care through the Obstetrical Service at the Massachusetts General Hospital and its affiliated community health centers were eligible for inclusion. The Massachusetts General Hospital obstetrics service provides primary prenatal care to an ethnically and socioeconomically diverse population at its main campus in Boston and at several affiliated health centers in surrounding neighborhoods. More than 97% of women who received prenatal care through the network delivered at the main Massachusetts General Hospital campus. Serum samples were collected at the first prenatal visit from women who provided written informed consent. The samples were centrifuged, aliquoted, and stored at ⫺80°C for future analyses. Baseline demographic, medical, laboratory, and subsequent pregnancy characteristics were collected prospectively by practitioners throughout the period of prenatal care and were entered into an electronic medical record that served as the primary record for pregnancy and as the source of clinical 35.e2

www.AJOG.org data that were downloaded directly into the Massachusetts General Hospital Obstetric Maternal Study database. All affiliated clinical centers used the identical electronic medical record. Follow-up evaluation continued through the early postpartum period, which allowed pregnancy outcomes to be verified according to research definitions. All blood pressure measurements and urinalyses from the first prenatal visit through the postpartum period were captured by the database. For the current study, all women with a singleton gestation who delivered at ⬎26 weeks were eligible for inclusion. We excluded women with AIDS or hepatitis, cases of major fetal anomaly, and fetal death and women with placenta previa, placenta accrete, or placental abruption. Fifty consecutive cases of preeclampsia were identified according to the criteria of the International Society for the Study of Hypertension in Pregnancy25,26: hypertension (blood pressure, ⬎140/90 mm Hg) that first develops at ⬎20 weeks of gestation in a previously normotensive woman and proteinuria that exceeds 300 mg in a 24hour collection or ⱖ2⫹ by dipstick on a spot urinalysis. Fifty consecutive cases of IUGR, which was defined as birthweight at ⬍5th percentile for gestational age4 were also selected. Thirty-five of the 50 cases of preeclampsia had no evidence of IUGR (pure preeclampsia); the other cases manifested both preeclampsia and IUGR (mixed IUGR-preeclampsia). There were no differences in mean PP13 levels among women with pure preeclampsia vs women with mixed IUGRpreeclampsia (P ⫽ .13), whereas levels were significantly decreased among women with any preeclampsia compared with any IUGR (P ⬍ .01). Therefore, for the analyses, cases of mixed IUGR-preeclampsia were included in the preeclampsia group; the IUGR group consisted only of IUGR without preeclampsia. The incidence of preeclampsia increases with gestational age. Therefore, women with early preeclampsia that has not yet reached clinical attention and who deliver prematurely for other reasons may be misclassified as not having

American Journal of Obstetrics & Gynecology JULY 2007

had preeclampsia. To address this issue, we also selected 50 consecutive cases of preterm delivery, which was defined as infants who were born at ⬍37 weeks of gestation.27,28 Women who delivered prematurely without evidence of IUGR or preeclampsia were included; the study group included mostly women with premature rupture of membranes and idiopathic preterm delivery. Two control subjects were selected randomly per case (n ⫽ 300 subjects total) by frequency matching by gestational age at the time of blood sampling (at enrollment) to ensure a similar distribution of gestational age among cases and control subjects. In addition, control subjects were chosen from the same recruitment period as the cases (⫾1 month) to minimize potential confounding because of sample decay as a function of storage time. To be eligible as a control subject, women had to complete a pregnancy that was free of hypertension, diabetes mellitus, or other complications and deliver a healthy infant at term. Twenty-two of the 450 women that were selected initially had to be excluded because of insufficient serum samples for PP13 assays, and 3 other women were excluded because of insufficient clinical data. The final study population therefore consisted of 47 women who experienced preeclampsia, 42 women who experienced IUGR, 46 women who delivered preterm, and 290 control subjects. The study was approved by the Massachusetts General Hospital human research committee, and all subjects provided written informed consent.

PP13 testing Assays for PP13 were performed by technicians who were blinded to pregnancy outcome. The Massachusetts General Hospital Obstetric Maternal Study investigator was not involved in the PP13 assays and was the only member of the research team with knowledge of pregnancy outcomes. Maternal serum concentrations of PP13 were measured with a solid-phase sandwich enzyme-linked immunosorbent assay with a pair of PP13-specific monoclonal antibodies, marked with amplified biotin-extravi-

Obstetrics

www.AJOG.org din-horseradish-peroxidase complex, and developed with tetramethylbenzidine substrate, as previously described.19,23 Optical density was measured at 450 nm and translated quantitatively with a calibration curve made of recombinant PP13 standards (0-500 pg/mL).19 The assay’s operating range is 3-500 pg/mL.23 The analytic detection level was 3 pg/mL, and results that were below the level of detection were assigned a value of 3 pg/mL. The intra- and interassay coefficients of variation for this study were 7.9% and 11.4%, respectively.

Statistical analysis Baseline and delivery characteristics were compared between cases and control subjects with the use of the Fisher exact test for categoric variables and independent t-tests for continuous variables. PP13 levels were not distributed normally; therefore, we compared PP13 levels across the different pregnancy outcomes using the Wilcoxon rank sum test. PP13 levels increase as gestational age increases.19,23 To account for this effect, PP13 levels were examined as multiples of the median (MoM) level among the control subjects at the same gestational age, on the basis of weighted regression of the observed median at each completed week. We also examined maternal weight, height, body mass index (BMI), age, parity, and ethnicity as potential confounders in the adjusted MoM analyses. Statistical modeling was used to evaluate the screening potential of serum PP13. This involved estimation of a woman’s likelihood ratio of the development of each of the pregnancy disorders relative to the chance of a normal outcome that was based on her PP13 level. The 3 sets of control subjects were compared with the use of the Fisher exact test for categoric variables and independent t-tests for continuous variables. There were no significant differences in PP13 levels or any covariates (P ⬎ .6 for any parameter) among the control groups. Therefore, we compared the individual pregnancy disorders with the aggregate group of control subjects (preeclampsia

vs control subjects; IUGR vs control subjects; preterm delivery vs control subjects). Examination of the MoM values showed that they did not fit a Gaussian distribution, either untransformed or after logarithmic transformation. Therefore, logistic regression was used to estimate the odds of being a case vs a control. The inclusion of an individual subject whose risk of a pregnancy disorder that we were aiming to “predict” in the data set that was used to “fit” the regression models could lead to a biased result because of the over-fitting of the models. To avoid such bias, each individual subject’s risk was computed on the basis of a model that excluded that subject (ie, the model did not “learn” from that subject, but only from the others). The likelihood ratio of developing a pregnancy disorder was then calculated as the odds ratio divided by the incidence of the disorder in the overall Massachusetts General Hospital Obstetric Maternal Study cohort between September 1998 and January 2005 (the time span of the current study). During this period, there were 18,858 total deliveries, 943 of which (5%) were complicated by preeclampsia. During the same time period, the incidence rates of IUGR and preterm delivery were 3% and 10%, respectively. Receiver operating characteristic curves were plotted from the likelihood ratios to obtain the area under the curve (AUC) in the prediction of each of the disorders29 and to calculate detection and false-positive rates. Probability values of ⬍.05 were considered statistically significant.

R ESULTS Demographic and delivery characteristics The Table presents the baseline and delivery characteristics of the women according to their pregnancy outcome. Compared with control subjects, women who subsequently experienced preeclampsia demonstrated increased maternal weight, BMI, and blood pressure at baseline. Women with preeclampsia were significantly more likely to deliver smaller fetuses by cesarean delivery at younger gestational ages than were control subjects. Compared with control

Research

subjects, women with IUGR were more likely to be nulliparous and not white and to have had higher systolic blood pressure at enrollment. They delivered earlier; as expected, the birthweight was lower. There were a higher proportion of female-to-male infants, compared with the control subjects. Compared with control subjects, women with preterm delivery were older and less likely to be of Hispanic ethnicity. These women had increased systolic blood pressure at enrollment; the infant’s birthweight was significantly lower than control women, and, as defined, they delivered at significantly earlier gestational ages.

PP13 concentrations and MoMs The median first-trimester PP13 concentration in control pregnancies was 132.5 pg/mL (mean, 243.4 ⫾ 387.3 pg/ mL). By comparison, all 3 pregnancy disorders demonstrated significantly decreased first-trimester PP13 concentrations (Figure 1). The median PP13 value for preeclampsia was 27.2 pg/mL (mean, 31.0 ⫾ 37.9 pg/mL; P ⬍ .001, compared with control subjects); for IUGR, it was 86.6 pg/mL (mean, 109.7 ⫾ 103.9 pg/ mL; P ⬍ .001, compared with control subjects); and for preterm delivery, it was 84.9 pg/mL (mean, 156.2 ⫾ 193.3 pg/mL; P ⫽ .005, compared with control subjects). Concentrations were below the assay detection limit of 3 pg/mL in 27 subjects: 6 control subjects (2%); 14 women with preeclampsia (30%), 2 women with IUGR (5%), and 5 women with preterm delivery (11%). The median concentrations of PP13 among control subjects for each completed week of gestation from 9-12 weeks were 126.5 (n ⫽ 44), 131.0 (n ⫽ 54), 131.2 (n ⫽ 75), and 148.5 (n ⫽ 117), respectively. These medians were used to calculate the MoM values. Among the control subjects, there was no association between PP13 levels that wee expressed as MoM and maternal age (P ⫽ .3), race/ethnicity (P ⫽ .3), systolic blood pressure (P ⫽ .1), diastolic blood pressure (P ⫽ .5), maternal weight (P ⫽ .7), maternal height (P ⫽ .1), BMI (P ⫽ .2), or parity (P ⫽ .9). Consequently, no further adjustments of MoMs were made

JULY 2007 American Journal of Obstetrics & Gynecology

35.e3

Research

Obstetrics

www.AJOG.org

TABLE

Baseline and delivery characteristics according to pregnancy outcome Characteristic

Normal (n ⴝ 289)

Gestational age at enrollment (wk)

11.0 ⫾ 1.3

Nulliparous (%)

53

Maternal age (y)

30.4 ⫾ 5.6

Preeclampsia (n ⴝ 47) 10.6 ⫾ 1.4

IUGR (n ⴝ 45) 10.7 ⫾ 1.5

Preterm (n ⴝ 46) 10.7 ⫾ 1.3

................................................................................................................................................................................................................................................................................................................................................................................

56

80*

57

................................................................................................................................................................................................................................................................................................................................................................................

31.1 ⫾ 5.3

31.4 ⫾ 6.1

33.3* ⫾ 5.0

................................................................................................................................................................................................................................................................................................................................................................................

Race/ethnicity (%)

................................................................................................................................................................................................................................................................................................................................................................................ †

White

69.2

70.8

57.8*

69.6

Black

6.6

4.2

6.7*

2.2

19.4

18.8

11.1*

10.9

Other

2.8

4.2

22.2*

10.9

Height (cm)

163 ⫾ 8

Weight (kg)

64.8 ⫾ 12.7

BMI (kg/m )

................................................................................................................................................................................................................................................................................................................................................................................ † ................................................................................................................................................................................................................................................................................................................................................................................ †

Hispanic

................................................................................................................................................................................................................................................................................................................................................................................ † ................................................................................................................................................................................................................................................................................................................................................................................

162 ⫾ 7

161 ⫾ 7

163 ⫾ 7

78.1* ⫾ 16.9

61.1 ⫾ 10.8

67.4 ⫾ 12.7

24.5 ⫾ 4.6

30.0* ⫾ 6.4

23.5 ⫾ 4.1

25.5 ⫾ 4.6

Systolic blood pressure (mm Hg)

111 ⫾ 9

118 ⫾ 12

114 ⫾ 14

116 ⫾ 12

Diastolic blood pressure (mm Hg)

70 ⫾ 7

75* ⫾ 8

................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ 2 ................................................................................................................................................................................................................................................................................................................................................................................ † † ‡ ................................................................................................................................................................................................................................................................................................................................................................................

72 ⫾ 10

72 ⫾ 8

................................................................................................................................................................................................................................................................................................................................................................................

Tobacco, alcohol drug use (%)

10

10

11

4

Gestational age at delivery (wk)

40.0 ⫾ 1.1

37.9* ⫾ 2.2

38.6* ⫾ 2.1

32.0* ⫾ 3.8

Birthweight (g)

3499 ⫾ 422

3162* ⫾ 755

2430* ⫾ 396

1948* ⫾ 712

................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................ ................................................................................................................................................................................................................................................................................................................................................................................

Delivery mode (%)

................................................................................................................................................................................................................................................................................................................................................................................ ‡

Vaginal

75.4

56.3

73.3

69.6

Cesarean

24.6

43.8

26.7

30.4

Fetal gender (% male)

54.0

41.7

35.6

60.9

................................................................................................................................................................................................................................................................................................................................................................................ ‡ ................................................................................................................................................................................................................................................................................................................................................................................ † ................................................................................................................................................................................................................................................................................................................................................................................

Data are given as median ⫾ SD, unless indicated otherwise. Statistical tests of significance refer to comparisons of each adverse pregnancy outcome group with the healthy control group and are indicated by * P ⱕ .001, †

P ⱕ .05;

‡

P ⱕ .01.

for any of these covariates. The median PP13 level in control subjects was 1 MoM, compared with a median MoM of 0.2 for preeclampsia, 0.6 for IUGR, and 0.6 for preterm delivery (each P ⬍ .01, compared with control subjects).

Receiver operation characteristic curves The following logistic regression equations for each disorder were used: Preeclampsia : Log关p(x) ⁄ (1 ⫺ p关x兴)兴 ⫽ ⫺3.09 ⫺ 3.28 · x IUGR : Log关p(x) ⁄ (1 ⫺ p关x兴)兴 ⫽ ⫺2.10 ⫺ 1.25 · x Preterm delivery : Log关p(x) ⁄ (1 ⫺ p关x兴)兴 ⫽ ⫺1.99 ⫺ 1.09 · x 35.e4

In these equations, x ⫽ MoM of PP13, and p(x) is the probability of the specific disorder, given x ⫽ MoM of PP13. Note that the values in these formulas are based on models that are fitted to the entire data, unlike the models that are used for sensitivity/specificity estimation later, which are based on out-of-sample prediction in which each subject’s value was predicted with the use of a model that was built to excluded that subject. Figure 2 shows the receiver operating characteristic curves for the risks of the development of the 3 disorders that were derived with the use of the likelihood ratios from the odds ratio equations. The AUC for preeclampsia was 0.91 (95% CI, 0.86-0.95). Given a 90% specificity (10% false-positive rate), the MoM cutoff was 0.38. In 38 preeclampsia cases, the PP13 values were below this cutoff, which

American Journal of Obstetrics & Gynecology JULY 2007

yielded a sensitivity of 79% (95% CI, 64%-89%). These sensitivities and specificities are equivalent to an odds ratio of 32.1 (95% CI, 14.5-71.0; ie, a 32-fold greater likelihood of preeclampsia in women with low vs normal PP13 levels). Thus, PP13 testing adds substantial value for the prediction of the subsequent development of preeclampsia, compared with the predictive value that is based on traditional epidemiologic risk factors. In contrast, the AUC for IUGR was 0.65 (95% CI, 0.55-0.74; P ⬍ .002). Given a 90% specificity, the sensitivity was 33% (95% CI, 20%-50%). In 14 of 42 IUGR cases (33%), the PP13 MoM was below the cutoff, and the odds ratio that calculated from the model was 4.3 (95% CI, 2.1-9.1; P ⬍ .01). The AUC for preterm delivery was 0.6 (95% CI, 0.50-0.70). Given a 90% specificity, the sensitivity

Obstetrics

www.AJOG.org was 28% (95% CI, 16%-44%). In 13 of 46 preterm delivery cases (38%), the PP13 MoM was below the cutoff, and the odds ratio that was calculated from the model was 3.4 (95% CI, 1.6-7.2; P ⬍ .01).

FIGURE 1

Median (95% confidence limits) PP13 levels according to pregnancy outcome

200 n = 290

C OMMENT 150

PP13 (pg/ml)

In this prospective nested case-control study, maternal blood testing at 9-12 weeks of gestation demonstrated significantly decreased serum levels of PP13 in women who experienced preeclampsia at ⬎15-25 weeks later. PP13 serum levels were also decreased significantly among women whose pregnancies that were complicated by IUGR and preterm delivery, compared with control pregnancies; however, these levels were still substantially higher than in the women who later had preeclampsia. Although useful in the prediction of IUGR and preterm delivery, the use of PP13 testing at the first prenatal visit to predict subsequent preeclampsia achieved 79% sensitivity with 90% specificity. This compares favorably with other diagnostic tests that have been used to assess the risk of preeclampsia and extends our results from a previous study23 that indicated that PP13 testing at 11-13 weeks of gestation was a useful diagnostic adjunct to Doppler ultrasound in the prediction of earlyonset preeclampsia. The earlier testing that was implemented in this larger study of PP13 enabled us to demonstrate greater diagnostic resolution between preeclampsia and control pregnancy and greater sensitivity for preeclampsia relative to the other pregnancy disorders that we examined. Thus, the enzymelinked immunosorbent assay test kit for PP13 provides a noninvasive, effective, and simple method for the measurement of PP13 levels that can be integrated easily into routine clinical practice in an effort to identify women who are at high risk for morbid pregnancy disorders. The molecular biologic makeup of PP13 provides clues to explain its potential role in the pathogenesis of placental insufficiency and its usefulness as a biomarker for the clinical disorders that arise as a consequence of abnormal placental development. PP13 is a 32-kd dimer protein that shares high-sequence

Research

n = 42

*

n = 46

*

100 n = 47

**

50

0 Control

Preeclampsia

IUGR

PTD

The single asterisk denotes a probability value of ⬍ .005, compared with control subjects; the double asterisk denotes a probability value of ⬍ .001, compared with control subjects. PE, preeclampsia; PTD, preterm delivery.

homologic makeup with the galactan family; 8 of the 16 invariant residues that comprise the carbohydrate recognition/ binding domains in galactans are conserved in PP13.19,20 PP13 is produced in the trophoblast and is packaged into endosomes that are released through “coated pits” through a calcium mobilization process.19 The secreted PP13 binds to sugar residues of extracellular matrix molecules (such as annexin II), which creates a “molecular bridge” for placental implantation in the endometrium.20 PP13 also demonstrates mild lysophospholipase-A activity that leads to the liberation of fatty acid constituents of the plasma membrane, particularly linoleic acid, that might contribute to normal implantation. Furthermore, PP13 increases liberation of prostaglandins, particularly prostacyclin, which are important for trophoblast-stimulated vascular remodeling in the maternal spiral arteries in early placental development19-22 Finally, PP13 binds to betaand-gamma-actin within trophoblasts20 and thus is involved in the migration of

trophoblasts towards the placental bed.21 Although these effects require further study, it is possible that a shortage of PP13 in the first trimester of pregnancy, which is marked by decreased circulating levels either because of decreased transcription or increased catabolism, may impair several critical functions that are required for normal implantation and maternal vascular remodeling. Studies of the potential effects of PP13 mutations are underway (unpublished data).30 For an individual woman, the chance of the development of 1 of the pregnancy disorders that we studied can be estimated from her PP13 level with the use of the logistic regression equations that we have developed. These equations yield likelihood ratios that estimate the risk of disease on the basis of the pretest probability in the overall population and the woman’s individual PP13 level. The approach that was used to calculate the odds ratio of a woman experiencing preeclampsia was based on modeling, taking into account the prevalence of the disease in the population from which the

JULY 2007 American Journal of Obstetrics & Gynecology

35.e5

Research

Obstetrics

www.AJOG.org

FIGURE 2

Receiver operating characteristic curves according to pregnancy outcome

The dotted line represents preeclampsia; the black line represents IUGR; the gray line represents preterm delivery.

samples were taken for this study. This yielded an odds ratio (with the use of a likelihood ratio cut-off of 2.9) of 32 for the prediction of preeclampsia, 4.3 for the prediction of IUGR (with the use of a likelihood ratio cut-off of 8.3), and 3.4 for the prediction of preterm delivery (with the use of a likelihood ratio cut-off of 2.1). In addition, the PP13 likelihood ratio can be applied to a previous risk that was based on factors such as age, parity, history of the disorder,1,2,31-33 which provides the physician with a novel diagnostic tool for assessing a woman’s risk of preeclampsia and IUGR. A limitation of this study is that we measured only PP13 levels and not other predictive biomarkers such as cellular fibronectin,34 Doppler ultrasound of maternal arteries at 11-14 weeks,23 or serum and urine angiogenic factors such as soluble fms-like tyrosine kinase 1 and pla35.e6

cental growth factor.15-17 However, this is among the first studies to examine PP13 in a study that is nested within a large prospective cohort. Furthermore, we currently are investigating whether the combined use of PP13 and PIGF could improve the sensitivity and specificity of either test alone for the prediction of preeclampsia. A second limitation is that we had only a few cases of early onset preeclampsia; therefore, further studies that are dedicated to researching this variant of preeclampsia are needed. Third, we tested the sensitivity and specificity of PP13 in a single study population in whom all of the control subjects completed normal pregnancies. Thus, the control group is somewhat different from the population of women who would be screened in real clinical practice at 9-12 weeks of gestation. For example, we did not include in

American Journal of Obstetrics & Gynecology JULY 2007

the control group women with gestational diabetes mellitus, intrauterine fetal death, or spontaneous abortions. The effect on the sensitivity and specificity of PP13 testing, if it were to be done in an entirely unselected population of all women in real clinical practice, will require further study. We are in the process of performing such studies in a largerscale population to further validate and expand the role of early PP13 testing. A small number of studies have indicated that starting treatment in the first trimester generates a better outcome than the same treatment when given in the second trimester.35,36 Thus, we have started small-scale studies to assess the combination of first-trimester detection for the risk of later development of preeclampsia and early administration of promising preventative means. Whether early assessment of risk with the use of PP13 alone or in combination with other biomarkers can lead to an improvement in pregnancy outcomes or a reduction in health care costs remains unclear. Although women who are found to be high risk could be prescribed more aggressive prenatal care as defined by the American College for Obstetrics Gynecology, there are no data available to demonstrate that these interventions improve outcomes. One reason is that randomized clinical trials of strategies that aim to prevent preeclampsia or mitigate its effects are limited by an inability to select adequately a study population with a high disease event rate because of the lack of reliable biomarkers that are detectable early in pregnancy.37,38 This difficulty increases the likelihood that trials will be underpowered and extremely costly. Therefore, at present, the greatest promise of PP13 testing may be as a critical tool for the research community to help design more efficient randomized trials that will enable interventions to be delivered as early as the first trimester of pregnancy to the women who are at the highest risk. f ACKNOWLEDGMENT We thank Ravi Thadhani, MD, MPH, from the Renal Unit, Department of Medicine, and Department of Obstetrics and Gynecology, Massachusetts General Hospital, Harvard Medical

Obstetrics

www.AJOG.org School, Boston, MA, for his valuable involvement and advice throughout this study.

REFERENCES 1. Roberts JM, Cooper DW. Pathogenesis and genetics of pre-eclampsia. Lancet 2001;357: 53-6. 2. Khong TY, De Wolf F, Robertson WB, Brosens I. Inadequate maternal vascular response to placentation in pregnancies complicated by pre-eclampsia and by small-for-gestational age infants. BJOG 1986;93:1049-59. 3. Lyell DJ, Lambert-Messerlian GM, Giudice LC. Prenatal screening, epidemiology, diagnosis, and management of preeclampsia. Clin Lab Med 2003;23:413-42. 4. Alexander GR, Himes JH, Kaufman RB, Mor J, Kogan M. A United States national reference for fetal growth. Obstet Gynecol 1996;87: 163-8. 5. Rogowski J. Measuring the cost of neonatal and perinatal care. Pediatrics 1999;103 (suppl);329-35. 6. Lewitt EM, Schuurmann Baker L, Corman H, Shiono PH. The direct cost of low birth weight. Future Child 1995;5:35-56. 7. Williams D. Pregnancy: a stress test for life. Curr Opin Obstet Gynecol 2003;15:465-71. 8. Wilson BJ, Watson MS, Prescott GL, et al. Hypertensive diseases of pregnancy and risk of hypertension and stroke in later life: results from cohort study. BMJ 2003; 326:845-51. 9. Gonen R, Perez R, David M, Dar H, Merksamer R, Sharf M. The association between unexplained second-trimester maternal serum hCG elevation and pregnancy complications. Obstet Gynecol 1992;80:83-6. 10. Spencer K, Yu CK, Cowans NJ, Otigbah C, Nicolaides KH. Prediction of pregnancy complications by first-trimester maternal serum PAPP-A and free ␤-hCG and with second-trimester uterine artery Doppler. Prenat Diagn 2005;25:949-53. 11. Dugoff L, Lynch AM, Cioffi-Ragan D, et al, for the FASTER Trial Research Consortium. First trimester uterine artery Doppler abnormalities predict subsequent intrauterine growth restriction. Am J Obstet Gynecol 2005;193: 1208-12. 12. Lambert-Messerlian GM, Silver HM, Petraglia F, et al. Second-trimester levels of maternal serum human chorionic gonadotropin and inhibin A as predictors of preeclampsia in the third trimester of pregnancy. J Soc Gynecol Investig 2000;7:170-4. 13. Florio P, Reis FM, Pezzani I, Luisi A, Severi FM, Petraglia F. The addition of activin A and inhibin A measurement to uterine artery Doppler velocimetry to improve the early prediction of

pre-eclampsia. J Ultrasound Obstet Gynecol 2003;21:165-9. 14. Dugoff L, Hobbins JC, Malone FD, et al, for the FASTER Trial Research Consortium. Firsttrimester maternal serum PAPP-A and freebeta subunit human chorionic gonadotropin concentrations and nuchal translucency are associated with obstetric complications: a population-based screening study (the FASTER Trial). Am J Obstet Gynecol 2004;191:1446-51. 15. Wolf M, Sandler L, Muñoz K, Hsu K, Ecker JL, Thadhani R. First trimester insulin resistance and subsequent preeclampsia: a prospective study. J Clin Endocrinol Metab 2002;87: 1563-8. 16. Thadhani H, Mutter WP, Wolf M, et al. First trimester placental growth factor and soluble Fms-like tyrosine kinase 1 and risk for preeclampsia. J Clin Endocrinol Metab 2004;89: 770-5. 17. Levine RJ, Maynard SE, Qian C, et al. Circulating antigenic factors and the risk of preeclampsia. N Engl J Med 2004;350:672-83. 18. Levine RJ, Thadhani R, Qian C, et al. Urinary placental growth factor and risk of preeclampsia. JAMA 2005;293:77-85. 19. Burger O, Pick E, Zwickel J, et al. Placental protein 13 (PP-13): effects on cultured trophoblasts, and its detection in human body fluids in normal and pathological pregnancies. Placenta 2004;25:608-22. 20. Than NG, Pick E, Bellyei S, et al. Functional analyses of placental protein 13/galectin-13. Eur J Biochem 2004;271:1065-78. 21. Than NG, Sumegi B, Than GN, Berente Z, Bohn H. Isolation and sequence analysis of a cDNA encoding human placental tissue protein 13 (PP13), a new lysophospholipase, homologue of human eosinophil Charcot-Leyden crystal protein. Placenta 1999;20:703-10. 22. Sammar M, Sadan O, Fleischfarb Z, Meiri H, Huppertz B, Golan A. Molecular and cellular impacts of placenta-related pathologies on PP13 in complicated premature deliveries [abstract]. Am J Obstet Gynecol 2005;193(suppl):S77. 23. Nicolaides KH, Bindra, R, Turan OM, et al. A novel approach to first trimester screening for early Pre-eclampsia combining serum PP-13 and Doppler ultrasound. Ultrasound Obstet Gynecol 2005;27;13-7. 24. Thadhani R, Stampfer MJ, Hunter DJ, Manson JE, Solomon CG, Curhan GC. High body mass index and hypercholesterolemia: risk of hypertensive disorders of pregnancy. Obstet Gynecol 1999;94:543-50. 25. Brown MA, Lindheimer MD, de Swiet M, Van Assche A, Moutquin JM. The classification and diagnosis of the hypertensive disorders of pregnancy: statement from the International Society for the Study of Hypertension in Preg-

Research

nancy (ISSHP). Hypertens Pregnancy 2001;20:IX-XIV. 26. Davey DA, MacGillivray I. The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 1988;158: 892-8. 27. National Medical Research Council. Clinical practice guidelines: management of preterm labour. In: Obstetricians and Gynaecologists Academy Of Medicine. Singapore: Ministry of Health; 2001; p 1-22. 28. Catlin A. Thinking outside the box: prenatal care and the call for a prenatal advance directive. J Perinat Neonat Nurs 2005;19:169-76. 29. Hanley JA, McNeil BJ The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982;143:2936. 30. Sammar M, Stalk M, Gebhardt S, Pick E, Meiri H, Huppertz B, Hillerman T RNA splicing and DNA polymorphism leading to two subforms of placenta protein 13 in preeclampsia [abstract]. Am J Obstet Gynecol 2006; 195:S14. 31. Walker JJ. Pre-eclampsia. Lancet 2000; 356:1260-5. 32. Roberts JM. Preeclampsia: is there value in assessing before clinically evident disease? Obstet Gynecol 2001;98:596-9. 33. Myatt L, Miodovnik M. Prediction of preeclampsia. Semin Perinatol 1999;23:45-57. 34. Vhavarria M, Lara-Gonzalez L, GonzalezGleason A, Sogi I, Reyes A. Maternal plasma cellular fibronectin concentrations in normal and preeclamptic pregnancies: a longitudinal study for early prediction of preeclampsia. Am J Obstet Gynecol 2002;187:595-601. 35. Vainio M, Kujansuu E, Iso-Mustajärvi M, Mäenpää J. Low dose acetylsalicylic acid in prevention of pregnancy-induced hypertension and intrauterine growth retardation in women with bilateral uterine artery notches. BJOG 2002;109:161-7. 36. Chiaffarino F, Parazzini F, Paladini D, et al. A small randomised trial of low-dose aspirin in women at high risk of pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2004;112:142-4. 37. Poston L, Briley AL, Seed PT, Kelly FJ, Shennan AH, for the Vitamins in Pre-eclampsia (VIP) Trial Consortium. Vitamin C and vitamin E in pregnant women at risk for pre-eclampsia (VIP trial): randomised placebo-controlled trial. Lancet 2006;367:1145-54. 38. Rumbold AR, Crowther CA, Haslam RR, Dekker GA, Robinson JS, for the ACTS Study Group. Vitamins C and E and the risks of preeclampsia and perinatal complications. N Engl J Med 2006;354:1796-806.

JULY 2007 American Journal of Obstetrics & Gynecology

35.e7