Predictive value of the sflt 1 and PlGF in women at risk for preeclampsia in the south of vietnam BS hữu trung

Predictive value of the sFlt 1 and PlGF in women at risk for preeclampsia in the south of Vietnam Accepted Manuscript Predictive value of the sFlt 1 and PlGF in women at risk for preeclampsia in the s.

Accepted Manuscript

Predictive value of the sFlt-1 and PlGF in women at risk for preeclampsia in the

south of Vietnam

Trung Huu Nguyen, Thanh Cong Bui, Tuan Minh Vo, Quan Minh Tran, Loan

Thi-Thanh Luu, Tai Duy Nguyen

DOI: https://doi.org/10.1016/j.preghy.2018.07.008

To appear in: Pregnancy Hypertension: An International Journal

of Women's Cardiovascular Health

Received Date: 5 March 2018

Revised Date: 4 July 2018

Accepted Date: 26 July 2018

Please cite this article as: Nguyen, T.H., Bui, T.C., Vo, T.M., Tran, Q.M., Luu, L.T-T., Nguyen, T.D., Predictive

value of the sFlt-1 and PlGF in women at risk for preeclampsia in the south of Vietnam, Pregnancy Hypertension:

An International Journal of Women's Cardiovascular Health (2018), doi: https://doi.org/10.1016/j.preghy 2018.07.008

This is a PDF file of an unedited manuscript that has been accepted for publication As a service to our customers

we are providing this early version of the manuscript The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

913931988)

b

Department of Family and Preventive Medicine, The University of Oklahoma Health Sciences Center Address: 655 Research Parkway, Suite 400, Oklahoma City, Oklahoma, 73104, United States Emails: thanh-c-bui@ouhsc.edu, thanh.bui@aya.yale.edu Telephone: 1-405-271-8001 ext 50559

c

Oxford University Clinical Research Unit, Wellcome Trust Major Overseas Programme, Ho Chi Minh City, Vietnam Email: quantm@oucru.org, tranminhquan12332192@gmail.com Telephone: +84 8 39237954

d

Tu Du Hospital Address: 284 Cong Quynh Street, District 1, Ho Chi Minh City, Vietnam Email: luuloan123@gmail.com Telephone: 084-913935051

Corresponding author:

Thanh Cong Bui, MD, DrPH,

Assistant Professor of Research

Department of Family and Preventive Medicine, College of Medicine,

The University of Oklahoma Health Sciences Center

Address: 655 Research Parkway, Suite 400, Oklahoma City, Oklahoma, 73104, United States Emails: thanh-c-bui@ouhsc.edu, thanh.bui@aya.yale.edu

Telephone: 1-405-271-8001 ext 50559

Suggested running head: Predictive value of the sFlt-1 and PlGF in preeclampsia

Abstract Word Count: 250

Manuscript Word Count: 2589

Number of tables: 5

Number of figures: 1

Abstract

Background: It has been suggested that soluble fms-like tyrosine kinase receptor-1 (sFlt-1) and

placental growth factor (PlGF) play potential roles in preeclampsia diagnosis Nevertheless, studies on the use of sFlt-1, PlGF, and sFlt-1/PlGF ratio in predicting preeclampsia have found contradictory results Thus, more studies in different populations are needed

Objectives: This study aims to (i) examine the associations between sFlt-1, PlGF, or sFlt-1/PlGF

ratio at gestational ages of 24-28 weeks and subsequent preeclampsia, and (ii) estimate predictive values of these markers in southern Vietnamese women

Methods: We used a nested case-control design from a cohort of 490 pregnant women who were

at risk of preeclampsia The total sample size for statistical analysis consisted of 30 cases and 67controls Levels of sFlt-1 and PlGF were quantified by using a fully automated

electrochemiluminescence immunoassay platform (Elecsys®/Cobas®)

Results: The median of sFlt-1 concentration was not statistically different between case and

control groups The median PlGF concentration was lower (349.7 pg/ml versus 534.6 pg/ml,

P<.001) and the median sFlt-1/PlGF ratio was higher in the preeclampsia group (4.3 versus 1.9, P<.001) After adjusting for maternal age, gestational age, nullipara, and body mass index, the

odds of preeclampsia in women with an sFlt-1/PlGF ratio in the fourth quartile were 10 times greater than in women with an sFlt-1/PlGF ratio in the other quartiles (95% confidence interval,

3.2–31.4; P<.001)

Conclusions: This study contributes to the literature that sFlt-1/PlGF ratio measured at

gestational weeks 24–28 in southern Vietnamese women can separate preeclamptic from

normotensive cases

Keywords: pregnancy-induced hypertension, placental growth factor, PlGF, soluble fms-like

tyrosine kinase, sFlt-1, preeclampsia

INTRODUCTION

Preeclampsia, occurring in 2-8% of pregnancies worldwide, remains a common

morbidity and a leading cause of several maternal and fetal complications(1, 2).These

complications include maternal medical problems in organs (e.g., kidney, liver, brain), preterm birth, maternal mortality, and fetal or neonatal mortality In low- and middle-income countries where neonatal intensive care is limited or less accessible, the maternal and neonatal mortality and morbidity are considerably higher Clinical diagnosis of preeclampsia is typically and

principally based on hypertension and proteinuria(3); this is also a current practice in Vietnam It has been suggested that these classical standards are insufficient to diagnose preeclampsia early for better and proper management(3) Other reliable methods and biomarkers to detect

preeclampsia are needed, and thus have been broadly studied

In 2003-2004, Maynard and Levine reported potential roles of soluble fms-like tyrosine kinase receptor-1 (sFlt-1) and placental growth factor (PlGF) in the risk of preeclampsia(4-6) Specifically, PlGF concentration decreased from gestational weeks 13-18, while sFlt-1 and sFlt-1/PlGF ratio increased about 5 weeks before the occurrence of clinical signs of preeclampsia(4) Since then, several case-control and prospective studies have highlighted the relevance of sFlt-1 and PlGF in preeclampsia, and their potential usefulness in diagnosis and prediction of

preeclampsia(7-18) Nevertheless, findings of the associations between sFlt-1 or PlGF and preeclampsia were not consistent across studies, and the magnitude of associations varied

widely(19, 20) Studies on the use of sFlt-1, PlGF, and sFlt-1/PlGF ratio in predicting

preeclampsia also found contradictory results(20, 21) Possible explanations for the differences

in results include gestational age at the time of blood sampling, comorbidity at enrollment,

selected testing methods, and measures of outcomes In addition, the suggested cut-off values and predictive efficacy of these markers varied largely across study countries(19, 21), suggesting that the changed levels of these markers in hypertensive pregnancy disorders may not necessarily

be similar in different ethnic and geographical populations Thus, more studies in different populations are needed

For these reasons, it is important to investigate the roles of sFlt-1, PlGF, and sFlt-1/PlGF ratio in predicting preeclampsia in Vietnam To our knowledge, there is only one study

examining this topic in women at weeks 15-19 of pregnancy in the north of Vietnam(22,

23).Measuring these biomarkers in early weeks of pregnancies, and hence too early prior to the onset of preeclampsia, may not be optimal because the altered magnitudes of the markers may not yet be observable Using a clinic-based population in the south of Vietnam, this study aims to (i) examine the associations between sFlt-1, PlGF, or sFlt-1/PlGF ratio at gestational ages of 24-

28 weeks and subsequent preeclampsia, and (ii) estimate predictive values of these markers

METHODS

Study design, setting, and sample size

We used a nested case-control design We enrolled a cohort of 490 pregnant women who visited the Obstetrics and Gynecology Clinic, Hospital of the University of Medicine and

Pharmacology in Ho Chi Minh City from October 2012 to June 2014 Inclusion criteria were women between 24-28 weeks of pregnancy who did not have preeclampsia at the time of

enrollment, but were at high risk according to the classification of the American Congress of Obstetricians and Gynecologists(24) Specifically, women were considered at risk for

preeclampsia if they had one of the following risk factors: an age of ≥35 years, a duration of ≥10 years between this pregnancy and the last pregnancy, a history of preeclampsia, diagnosis of diabetes before this pregnancy, diagnosis of hypertension before this pregnancy, body mass index (BMI) ≥23, a history of systemic lupus erythematosus, or family history of preeclampsia Exclusion criteria were having two or more offspring, being diagnosed with preeclampsia at the time of enrollment, being diagnosed with a mental health issue, and refusing to participate At the time of enrollment, we obtained a 3-ml venous non-fasting blood sample from all participants and banked the samples (see details below) We provided routine care services to all participants according to the Vietnam national guidelines for reproductive healthcare(25), and followed up with them until 24 hours after delivery At the time of data collection for this study, there was no national specific guidelines for treating women at risk of preeclampsia with aspirin; therefore, none of the participants were treated with aspirin during the study

After follow-up, 24 (4.9%) women were lost to follow-up or did not have all key data about their pregnancies; these women were excluded Thus, the total cohort sample size was 466 Thirty women were diagnosed with preeclampsia during the follow-up; these women became cases Preeclampsia was defined as (i) a systolic blood pressure of at least 140 mm Hg or a diastolic blood pressure of at least 90 mm Hg, recorded on at least two occasions 4-hours apart (after a 5-minute rest), plus (ii) proteinuria (300 mg in a 24-hour excretion) or one or more systemic signs of severe preeclampsia (thrombocytopenia platelet count <100,000/microliter, impaired liver function, renal insufficiency, pulmonary edema, or new-onset cerebral or visual disturbances)(24) Each case was matched with an average of two controls (i.e., a 2:1 ratio) that

research purposes Women diagnosed with preeclampsia were closely monitored or hospitalized for monitoring, and were treated according to the national guidelines(25)

Variables and measurement

The primary outcome was preeclampsia, defined as aforementioned The main predictors

of interest weresFlt-1, PlGF, and sFlt-1/PlGF ratio At the time of enrollment, participants’ blood samples were collected Serum was obtained through centrifugation (2,500 g for 15 minutes) and was stored at -80oC When cases and controls were identified, their stored serums were defrosted

at room temperature Levels of sFlt-1 and PlGF were quantified by using a fully automated electrochemiluminescence immunoassay platform (Elecsys®/Cobas®, Roche Diagnostics)

according to the manufacturer's instructions The detectable ranges of sFlt-1 and PlGF were10–85,000 pg/mL and 3–10,000 pg/mL, respectively

Other clinical data collected included blood pressure, body mass index (BMI), indicators

of liver function (SGOT, SGPT), indicators of kidney function (creatinine), thrombocytes

through blood tests, and proteinuria through urine tests These variables were collected through examination, laboratory test results, and medical records at the clinic Patient interviews with a structured questionnaire were used to obtain data on age, durations in years between pregnancies,

history of hypertension, history of preeclampsia, history of diabetes, history of systemic lupus erythematosus, and family history of preeclampsia

Statistical analysis

Statistical analyses were performed using Stata 14.2 (StataCorp, College Station, TX)

We first performed basic descriptive statistics (percentages, means, medians) to examine each variable in each study group To explore differences in participants’ characteristics between case

and control groups, we used appropriate tests, such as t-test for normally-distributed continuous

variables or chi-squared test for categorical variables To examine differences in the main predictors of interest between case and control groups in bivariate analysis, we used the Mann-Whitney test for median comparison In multivariable analysis, we dichotomized values of sFlt-

1, PlGF, and sFlt-1/PlGF ratio at their third quartiles (Q3), as did previous studies (8, 26-29), so that the results could be more easily interpreted and compared Multivariable logistic regression models were performed to evaluate unadjusted and adjusted associations between each predictor

of interest and preeclampsia, reported as odds ratios (ORs) Variables included in multivariable models for adjustment were selected based on bivariate associations and a priori We also used logistic regression models to build receiver operator characteristic (ROC) curves to evaluate

predictive values of each biomarker All P values were two-tailed and were considered

statistically significant if <.05

RESULTS

Table 1 displays demographic and clinical characteristics of women in the case versus control groups Bivariate analysis results suggest that women who developed preeclampsia had a higher BMI than did women who were normotensive All other demographic and clinical

characteristics, including gestational age at the time of blood collection, were similar in both case and control groups

Table 2 shows descriptive values of sFlt-1, PlGF, and sFlt-1/PlGF ratio in each group The median of sFlt-1 concentration was not statistically different between the two groups

However, the median of PlGF concentration was lower (349.7 pg/ml versus 534.6 pg/ml,

P<.001) and the median of sFlt-1/PlGF ratio was higher in the preeclampsia group (4.3 versus

1.9, P<.001)

Table 3 displays unadjusted and adjusted associations between preeclampsia and

maternal serum concentrations of sFlt-1, PlGF, and sFlt-1/PlGF ratio After adjusting for

maternal age, gestational age, nullipara, and BMI, the odds of preeclampsia in women with an sFlt-1/PlGF ratio in Q4 were 10 times greater than in women with an sFlt-1/PlGF ratio in Q3 or

lower (95% CI, 3.2–31.4; P<.001) A higher sFlt-1/PlGF ratio (in Q4) was also associated with

an earlier onset of preeclampsia in bivariate analysis (Table 4) Due to a small frequency of preeclampsia in our sample, we could not examine the adjusted associations between the

predictors and time to preeclampsia incidence Nevertheless, when we used ordinal logistic regression models to examine its associations with the predictors, controlled for BMI, the results were similar to those of the main outcome Specifically, time to preeclampsia incidence was not

associated with sFlt-1 (P=.433) but was associated with a lower PlGF level (P=.003) and with a higher sFlt-1/PlGF ratio (P<.001) The results of ROC analysis (Table 5) suggested that sFlt-

Our main findings suggest that PlGF value and sFlt-1/PlGF ratio, but not sFlt-1 value alone, were associated with a likelihood of preeclampsia incidence Women with a sFlt-1/PlGF ratio in Q4 (≥4.36) were 10 times more likely to develop preeclampsia than were other women Overall, our results contribute to emergent evidence that alterations in angiogenic and

antiangiogenic balance play an important role in the onset of preeclampsia

Although the directions of changes in sFlt-1, PlGF, and sFlt-1/PlGF ratio in preeclamptic women in our study are similar to most studies worldwide, there are subtle differences in our results First, in our study, the median of sFlt-1 concentration was higher in preeclamptic women than that in normotensive women, yet this difference was not statistically different Although this finding may be due to a moderate sample size, there are some other possible explanations for the non-significant difference As aforementioned, several studies worldwide found a significant increase in sFlt-1 in preeclamptic women, but other studies found no increase(19) The increase

Second, our results are in accordance with previous findings that PlGF decreases and sFlt-1/PlGF increases in women with preeclampsia However, although the median of sFlt-1/PlGF in preeclamptic women in our study (4.3) is higher than that in some studies (e.g., see Sibai et al, 2008(8)), it is generally lower than the values reported in several other studies(19) Similarly, the OR of having preeclampsia in women with sFlt-1/PlGF ratios in Q4 (versus in Q1-

3) in our study is higher than that in some studies (8, 27), yet lower than that in other studies(28, 29) These differences may be due to a modest increase in sFlt-1 level in preeclamptic cases in our sample, with the associated reasons discussed above On another note, that PlGF but not sFlt-

1 is predictive of preeclampsia supports an important pathophysiologic question: whether the decrease in free PlGF circulating during preeclampsia is mediated by excess binding of sFlt-1 to PlGF or a decrease in placental PlGF production and release (i.e., placental impairment or abnormalities)(5, 31, 32) Evidence suggested that sFlt-1 often increases near the time of clinical symptoms and is secondary to placental ischemia or hypoxia occurs.(33, 34) This may explain why we observed a reduction in PlGF concentration but not yet an elevation in sFlt-1 level at weeks 24-28

In our study, sFlt-1/PlGF ratio was generally superior to individual markers in predictive performance With a cut-off point that generated the best AUC, the sFlt-1/PlGF ratio had modest sensitivity (50.0%) and good specificity (86.6%) In a recent prospective, multi-country,

observational study to validate the performance of a sFlt-1/PlGF ratio in predicting the absence

or presence of preeclampsia in the short term, its ability to rule out preeclampsia in the

subsequent week was very high (negative predictive value = 99.3%; 95%CI, 97.9-99.9), while its ability to predict preeclampsia in the next 4 weeks was modest (positive predictive value = 36.7%; 95%CI, 28.4-45.7)(35) These findings suggest that sFlt-1/PlGF may have better

applications in excluding a potential occurrence of preeclampsia than in predicting

preeclampsia, even short-term or in women at risk

Additionally, the change and suggested cut-off value in our study are considerably lower than those in Verlohren’s and Zeisler’s studies in European medical centers, which used the Elecsys system to measure sFlt-1 and PlGF, as we did(26, 35, 36) Our results, however, are similar to findings from Nguyen’s study in the north of Vietnam,which also used Elecsys(22) Together, these findings highlight the possibility that the examined angiogenic and

antiangiogenic biomarkers, even though measured by the same technique, can vary significantly

in Vietnamese or other Asian populations If this is the case, further larger studies are needed to validate the reference and cut-off values of these markers in Vietnamese or other populations prior to their clinical application in preeclampsia diagnosis and management

The main strength of our study is the nested case-control design, which has advantages of

a prospective study and reduces uncertainty regarding the temporal sequence between exposure and disease onset Our study also has some limitations First, the study sample size is moderate

Nevertheless, we were still able to observe significant differences in the main predictors between groups Second, we could measure sFlt-1 and PlGF at only one time-point during pregnancy (weeks 24–28), and thus could not fully evaluate the predictive performance of these biomarkers Finally, there is not a practical way to date the onset of preeclampsia timely and accurately; this limited our ability to utilize other statistical methods (e.g., Cox proportional hazard) that would allow us to better examine the biomarkers’ predictive performance

This study contributes to the literature that sFlt-1/PlGF ratio measured in southern

Vietnamese women at gestational weeks 24–28 can separate preeclamptic from normotensive cases Similar to previous longitudinal studies, the results suggest that the sFlt-1/PlGF ratio has good specificity, yet modest sensitivity Given a rare consistency in literature regarding the manifest differences of the biomarkers in preeclampsia prediction, further large multi-center prospective studies are needed to evaluate the biomarkers’ predictive ability in heterogeneous patient populations in Vietnam and other developing nations

Table1: Clinical Characteristics of Study Sample

(N=67)

n (%)

Preeclampsia (N=30)

Kinh ethnicity (the major ethnicity in Vietnam) 96 (99.0%) 66 (98.5%) 30 (100.0%) -

Vocational school, college, or higher 65 (67.0%) 42 (62.7%) 23 (76.7%)