Original Article

Red Cell Distribution Width as a Predictor of

Persistent Pulmonary Hypertension of the
Newborn
Setareh Sagheb, MD1 Mahdi Sepidarkish, PhD2 Sayyed Ourmazd Mohseni, BSE3
Amirhossein Movahedian, MD1 Ziba Mosayebi, MD1

1 Tehran University of Medical Sciences, Tehran, Iran
2 Department of Epidemiology and Reproductive Health,
Reproductive Epidemiology Research Center, Royan Institute for
Reproductive Biomedicine, ACECR, Tehran, Iran
3 Ira A. Fulton Schools of Engineering, School of Biological and Health
Systems Engineering, Arizona State University, Tempe, Arizona
Address for correspondence Sayyed Ourmazd Mohseni, BSE, Ira A.
Fulton Schools of Engineering, School of Biological and Health
Systems Engineering, Arizona State University, Tempe, AZ
(e-mail: o.mohseni@outlook.com).

Am J Perinatol

Abstract Objective Persistent pulmonary hypertension of the newborn (PPHN) is a critical

condition with high mortality and morbidity rates in neonatal intensive care unit (NICU)
admitted neonates due to severe hypoxemia. The aim of this study was to evaluate red
cell distribution width (RDW) as a biomarker of hypoxemia and determine the optimal
cutoff point of RDW for identifying neonates with PPHN.
Study Design All PPHN diagnosed, NICU admitted term infants with hypoxemia after
birth from May 2014 to September 2016 were enrolled as case control and healthy term
infants with nonhemolytic jaundice who were admitted for phototherapy on the
second or third day of birth were the control group. Blood samples were collected.
Multiple logistic regression modeling was used to examine the association between
Keywords PPHN and RDW.
persistent pulmonary Result Receiver-operating characteristics (ROC) curve analysis was used to determine
hypertension of the the optimal cutoff point of RDW for identifying neonates with PPHN. RDW was higher in
newborn the PPHN group compared with the control group (p < 0.001). Signicant predictors of
hypoxemia PPHN were mothers underlying disease (p 0.01) and RDW (p < 0.001). The optimal
neonatal intensive RDW cut point for prediction of PPHN by the ROC curve analysis was 17.9 (sensitivity
care unit 85.71%). RDWs area under the curve was 0.9197 (p < 0.001).
red cell distribution Conclusion RDW may be a simple, valuable, accessible marker for predicting PPHN
width before performing echocardiography in hypoxemic NICU admitted neonates.

Persistent pulmonary hypertension of the newborn (PPHN)
with the denition of persistent right-to-left shunting of
blood through foramen ovale and ductus arteriosus or both
after birth leads to severe hypoxemia.1 It occurs in 10% of
term and near-term infants. This is a critical condition with
high mortality and morbidity rates in neonatal intensive care
unit (NICU) admitted neonates.2,3 Lack of access to inhaled
nitric oxide and extracorporeal membrane oxygenation in
Iran and other regional and developing countries can lead to
an increase of mortality and morbidity especially neurode-
velopmental impairment.
Red cell distribution width (RDW) index is a component of
the complete blood cell (CBC) count and is mainly used for
the differential diagnosis of microcytic anemia. Higher RDW

received Copyright by Thieme Medical DOI https://doi.org/

April 4, 2017 Publishers, Inc., 333 Seventh Avenue, 10.1055/s-0037-1604246.
accepted after revision New York, NY 10001, USA. ISSN 0735-1631.
June 12, 2017 Tel: +1(212) 584-4662.
RDW as a Predictor of PPHN of the Neonate Sagheb et al.
values indicate an increase in variations of red blood cell
(RBC) volume. Previously, this was calculated by dividing the
standard deviation of RBC volume to the mean corpuscular
volume and multiplying the product by 100, nowadays this is
automatically analyzed.4
Recently, RDW as an accessible predictive marker of mor-
tality in different diseases in adults including coronary artery,5
heart failure,6,7 stroke,8 bacterial infections and septic
shock,9,10 and diabetes11 has been studied. Moreover, the
correlation between RDW and mortality in older adult pa-
tients has been demonstrated.12,13 Though the mechanism of
increased RDW is not known, inammatory processes14 and
hypoxic events15 are suggested as the mechanisms of these
occurrences.
Limited information is available about the association of
RDW and neonatal diseases. RDW has been demonstrated as
a prognostic tool in neonatal sepsis.16
The aim of this study was to evaluate the relationship
between RDW and PPHN as a biomarker of hypoxemia and
to determine whether it can serve as an accessible marker for
prediction of PPHN in NICU admitted neonates and determine
the optimal cutoff point of RDW for identifying PPHN in
neonates. According to our knowledge, this is the rst time
that this correlation has been studied during neonatal periods.
Methods
Study Design and Population
This casecontrol study was conducted during 28 months from
May 2014 to September 2016 on all full-term newborns (with
the gestational ages of  37 weeks) with diagnosis of PPHN
who were admitted to the NICU of Arash Hospital in Tehran,
soon after birth. These neonates constituted the case group.
Healthy term infants with unconjugated hyperbilirubinemia,
who were admitted to our neonatal ward only for photother-
apy on the second or third day after birth, were included as
control group. The study was approved by the Research
Committee of Faculty of Medicine, Tehran University of Med-
ical Sciences and the local ethical committee and written
consent were obtained from parents before enrollment.
Inclusion criteria in the case group are all term infants
with hypoxemia who were intubated within the rst hour of
birth with diagnosis of PPHN based on real-time echocardio-
graphy combined with Doppler ow imaging (echocardio-
graphy, MyLab7, Esaote, Italy). This was done by neonatal
cardiologists within 24 hours after birth. Estimation of right
ventricle pressure (RVP), using assessments of tricuspid
regurgitation jet is compared with systemic blood pressure
and estimated RVP greater than one-half systemic blood
pressure suggests PPHN. Inclusion criteria in control group
are all healthy, full-term exclusively breast fed newborns
with nonhemolytic jaundice.
Exclusion criteria are (1) infants born to mothers with
chorioamnionitis, (2) major congenital abnormalities at
birth, (3) preterm infants, (4) congenital heart disease, (5)
early onset sepsis with positive blood culture, (6) congenital
diaphragmatic hernia, (7) hemolytic anemia, (8) hypoglyce-
mia, and (9) polycythemia.
American Journal of Perinatology
Mothers underlying disease during pregnancy (diabetes
mellitus, urinary tract infection, preeclampsia, and hypothyr-
oidism) and drug consumption during pregnancy were also
registered. To avoid any confounding effects, blood samples for
the CBC count (including RDW) were collected at the time of
NICU admission. CBC count was calculated by the automated
hematology analyzer XE-1200 (Sysmex, Japan). Arterial blood
gas analyses and other tests such as serum biochemical analysis
were also conducted. Blood gas analyses were performed in
Radiometer device ABL 900 ex sensor cassette and solution
package. In this study, metabolic acidosis was dened as
pH < 7.2 and base decit in the extracellular uid > 12.0
mmol/L in umbilical artery accordance with previous studies.17
Laboratory tests for management of hyperbilirubinemia
in term infants including CBC count were also conducted
based on the American Academy of Pediatrics guideline.18
RDW was reported as the coefcient of variation (percen-
tage) of RBC volume as measured with an automated analy-
zer (Beckman Coulter LH780 hematology analyzer; Beckman
Coulter, Brea, CA).
Statistical Analysis
Categorical and continuous variables were expressed as num-
ber (percentage) and mean  (standard deviation [SD]),
respectively. A chi-square test or Fishers exact test was
performed to compare categorical variables. Students t-test
was used to compare parametric continuous variables. The
MannWhitneys U-test was used to compare nonparametric
continuous variables. Multiple logistic regression modeling
was used to examine the association between PPHN and RDW.
Results are presented as odds ratio (OR) with 95% condence
intervals. Variables were considered in the model for adjust-
ment in the following order: gestational age, gender, weight,
the rst- and fth-minute Apgar score. Variable selection was
done based on the available evidence about the PPHN risk
factors as much as possible, without overlapping with other
measurements to avoid colinearity in the logistic model. Since
the data had several unbalanced and highly predictive risk
factors (complete separation problem), the multiple logistic
regression model was performed using rthlogit to examine
possible association between outcome of interest (PPHN) and
independent variables. The presence of the aforementioned
problems in logistic regression models can result in bias of OR
estimates away from 1. The rthlogit command did not use
maximum log likelihood but penalized log likelihood instead
of reducing bias.
Furthermore, receiver-operating characteristics (ROC)
curve analysis was used to determine the optimal cutoff
point of RDW for identifying neonates with PPHN. Areas
under the curve (AUC) and its standard error were calculated.
Data analysis was undertaken using Stata statistical soft-
ware, release 13.0 (StataCorp, College Station, TX).
Results
A total of 652 preterm and term infants were admitted to our
hospital, of which 138 (25%) fullled the eligible criteria, out
of which 63 (11.4%) with diagnosis of PPHN were considered
 RDW as a Predictor of PPHN of the Neonate Sagheb et al.

Table 1 Baseline characteristics of patients Table 2 Adjusted OR for relationship of PPHN and its predictors

Parameters Cases Control p-Value Variables Adjusted SE 95% CI p-Value

(n 63) (n 75) OR
Gestational 37.58  37.69  0.53 Gestational 1.02 0.27 0.262.03 0.56
age (wk) (1.12) (0.79) age (wk)
Weight (g) 2,896.11  3,190.6  0.002 Weight (g) 0.99 0.0004 0.991.0005 0.91
(525.42) (564.92)
RDW (%) 2.11 0.31 1.572.83 < 0.001
RDW (%) 19.99  (2) 16.32  (2.19) < 0.001
Gender 0.74 0.38 0.262.03 0.56
Gender
Mothers 41.71 61.98 2.26767.52 0.01
Boy 28 (44.4) 39 (52) 0.37 underlying
disease
Girl 35 (55.6) 36 (48)
Metabolic acidosis Abbreviations: CI, condence interval; OR, odds ratio; PPHN, persistent
pulmonary hypertension of the newborn; RDW, red blood cell distri-
Yes 47 (74.6) 0 (0) < 0.001 bution width; SE, standard error.
No 16 (25.4) 74 (100)
Mothers underlying disease
Yes 25 (39.7) 0 (0) < 0.001 The optimal RDW cut point for prediction of PPHN by the
No 38 (60.3) 75 (100) ROC curve analysis was 17.9, which yielded sensitivity
85.71% and specicity 85.33%. Overall, the AUC of RDW
Abbreviation: RDW, red blood cell distribution width. was 0.9197 (95% CI: 0.87430.965, p < 0.001) (Fig. 1).
Note: Values given as mean  standard deviation, or number (percen-
Positive likelihood ratio and negative likelihood ratio were
tage) unless otherwise indicated.
5.84 and 0.16, respectively.

as the study group, while 75 healthy term infants with

Discussion
unconjugated hyperbilirubinemia without hemolysis were
included in the control group. PPHN is a critical condition with high mortality and morbidity
Clinical and demographic ndings of both groups are rates in NICU admitted neonates due to severe hypoxemia.
described in Table 1. There was no signicant difference Physical examination is usually associated with some restric-
between maternal age in the case group (25.78  5.17) and tions for diagnosis of PPHN. Access to RDW index before
control group (25.64  6.25). In the case group, 25 (39.7%) of carrying out echocardiography in these situations will be
mothers had underlying diseases, 10 of which were taking accompanied by early intervention and aggressive treatment.
levothyroxine due to hypothyroidism, 3 were taking insulin, Based on our knowledge, this is the rst time that the
and the others were on a diabetic diet. value of RDW as a predictor marker of critical diseases has
In the case group, the mean (SD) Apgar score was been studied during the neonatal period excluding sepsis.
5.73  2.49 in the rst minute and 7.69  2.13 in the The present study evaluated the correlation of RDW and
fth minute. The mean (SD) weight of cases was PPHN and shows that RDW, as a simple accessible test,
2,896.11  (525.42) which was signicantly lower than the increases during PPHN with any etiologies in comparison
mean (SD) weight of controls, 3,190.6  (564.92) (p 0.002). with healthy term infants. The RDW reference interval in
Compared with neonates with PPHN, neonates without PPHN
had higher levels of rst-minute Apgar score (p 0.02) and
fth-minute Apgar score (< 0.001). The etiologies of the PPHN
were pneumonia (n 18), birth asphyxia (n 17), respiratory
distress syndrome (n 11), meconium aspiration syndrome
(n 9), and malignant transient tachypnea of the newborn
(n 8). Three deaths were observed in the study group while
there was no death in the control group. RDW was higher in
cases in comparison with the control group (19.99  2 vs.
16.32  2.19, respectively; p < 0.001).
After performing the multivariate model of logistic regres-
sion, signicant predictors of PPHN were mothers underlying
disease (OR 41.72, 95% CI: 2.26767.52, p 0.01) and RDW
(OR 2.11, 95% CI: 1.572.83, p < 0.001). Table 2 presents
the results of multivariate logistic regression. The goodness-
of-t test was performed for the nal version, which showed
a good t of the model (HosmerLemeshow chi-square Fig. 1 ROC curve analysis for persistent pulmonary hypertension.
test 2.64, p 0.95). ROC, receiver-operating characteristic.

American Journal of Perinatology

RDW as a Predictor of PPHN of the Neonate Sagheb et al.
term healthy infants at birth was reported between 15.5 and
20% by Christensen et al with the lower reference limit of
15.5%.19 In a retrospective study which was done by Garofoli
et al, mean ( SD) RDW values measured within the rst
3 days after birth were 15.65  1.18% in full-term new-
borns.20 In our study, the mean of RDW in term healthy
infants was 16.32  2.19 which was the same as Chen et als
study (16.04  1.25).16
This study also shows that the elevation of the RDW index
can be an independent predictor of PPHN. Recently, a prog-
nostic roll of RDW in pulmonary hypertension21 and follow-
up in ICU patients22 and the severity of obstructive sleep
apnea and carotid intima23 have been suggested. Although
the mechanism of increased RDW is not known, inamma-
tory processes 14 and hypoxic events15 are suggested as the
mechanisms of these occurrences.
Based on the Yas et als erythropoietin risk-pathway
model, hypoxia, due to any disease, increases the RDW index
and this can be a biomarker of hypoxemia in adults.15 PPHN
due to persistent right-to-left shunting of blood through fora-
men ovale and ductus arteriosus or both after birth leads to
severe systemic hypoxemia. As a result, the most probable
hypothesis is that, in PPHN infants, persistent hypoxia induces
an increase in RDW. The predictive role of RDW as an acces-
sible predictive marker of mortality in different diseases in
adults has been studied as well.
In the study by Aksoy et al, a negative correlation was
observed between RDW and gestational age (r 0.51;
p < 0.001). In their study, higher RDW in premature infants
within the rst 3 days of their life has been also associated
with increased rates of mortality (p < 0.0001) and late-onset
sepsis (p <0.005).24 However, in the study by Aksoy et al,
RDW did not predict mortality in very low birth weight
infants.
Chen et al showed that there was a negative correlation
between the neonatal critical illness score and RDW, while
there was a positive correlation with mortality rate in the
septic shock group (p <0.001). The increase of the RDW index
was associated with the severity of diseases as a result, they
considered RDW as a probable prognostic factor of neonatal
sepsis.16 In their study, they did not mention the sensitivity
and the specicity of RDW and the optimal RDW cut point.
In our study, there was no relationship between RDW and
the mortality rate; this might be due to the low number of
dead infants.
The optimal RDW cut point for prediction of PPHN by the
ROC curve analysis was 17.9, which yielded sensitivity
85.71% with the AUC of RDW 0.9197 (p <0.001). There is
no previous study during neonatal periods that can be
compared with our study. In the study by Abul et al, the
AUC of RDW for prediction of chronic thromboembolic
pulmonary hypertension after pulmonary embolism was
0.735 with 75% sensitivity and they considered RDW as an
independent predictor of chronic thromboembolic pulmon-
ary hypertension in adults.25 In comparison with Abul et als
study, it seems that RDW might help the prediction of PPHN
with a higher sensitivity value.
American Journal of Perinatology
Study Limitations and Strengths
A casecontrol and cross-sectional study was the main
limitation of this study. It is better if cohort studies in
multicenter hospitals are done due to the very low incidence
rate of this condition. Further investigations need to evaluate
this marker for each disease separately and to estimate the
morbidity outcome during the neonatal period.
Despite the limitations, this was the rst study that was
conducted during neonatal periods in hypoxemic events,
thus distinguishing this article from others.
Conclusion
Based on this study, RDW as a simple, accessible marker, with
a high predictable value may be used to predict PPHN before
performing echocardiography in NICU admitted neonates
with hypoxemia for better and faster treatment.
Funding
None.
Conict of Interest
None.
References
1 Steinhorn RH. Neonatal pulmonary hypertension. Pediatr Crit
Care Med 2010;11(2, Suppl):S79S84
2 Konduri GG, Solimano A, Sokol GM, et al; Neonatal Inhaled Nitric
Oxide Study Group. A randomized trial of early versus standard
inhaled nitric oxide therapy in term and near-term newborn
infants with hypoxic respiratory failure. Pediatrics 2004;113
(3 Pt 1):559564
3 Konduri GG, Vohr B, Robertson C, et al; Neonatal Inhaled Nitric
Oxide Study Group. Early inhaled nitric oxide therapy for term
and near-term newborn infants with hypoxic respiratory failure:
neurodevelopmental follow-up. J Pediatr 2007;150(03):235240,
240.e1
4 Bain BJ, Bates I, Laffan M, Lewis SM. Basic haematological tech-
niques. In: Dacie and Lewis Practical Haematology. 11th ed. Chap
3. Philadelphia, PA: Churchill Livingstone/Elsevier; 2012
5 Tonelli M, Sacks F, Arnold M, Moye L, Davis B, Pfeffer M; for the
Cholesterol and Recurrent Events (CARE) Trial Investigators.
Relation between red blood cell distribution width and cardio-
vascular event rate in people with coronary disease. Circulation
2008;117(02):163168
6 Felker GM, Allen LA, Pocock SJ, et al; CHARM Investigators. Red
cell distribution width as a novel prognostic marker in heart
failure: data from the CHARM Program and the Duke Databank.
J Am Coll Cardiol 2007;50(01):4047
7 Allen LA, Felker GM, Mehra MR, et al. Validation and potential
mechanisms of red cell distribution width as a prognostic marker
in heart failure. J Card Fail 2010;16(03):230238
8 Ani C, Ovbiagele B. Elevated red blood cell distribution width
predicts mortality in persons with known stroke. J Neurol Sci
2009;277(1-2):103108
9 Ku NS, Kim HW, Oh HJ, et al. Red blood cell distribution width is an
independent predictor of mortality in patients with gram-nega-
tive bacteremia. Shock 2012;38(02):123127
 RDW as a Predictor of PPHN of the Neonate
10 Jo YH, Kim K, Lee JH, et al. Red cell distribution width is a
prognostic factor in severe sepsis and septic shock. Am J Emerg
Med 2013;31(03):545548
11 Malandrino N, Wu WC, Taveira TH, Whitlatch HB, Smith RJ.
Association between red blood cell distribution width and macro-
vascular and microvascular complications in diabetes. Diabeto-
logia 2012;55(01):226235
12 Patel KV, Ferrucci L, Ershler WB, Longo DL, Guralnik JM. Red blood
cell distribution width and the risk of death in middle-aged and
older adults. Arch Intern Med 2009;169(05):515523
13 Martnez-Velilla N, Ibez B, Cambra K, Alonso-Renedo J. Red
blood cell distribution width, multimorbidity, and the risk of
death in hospitalized older patients. Age (Dordr) 2012;34(03):
717723
14 Lippi G, Targher G, Montagnana M, Salvagno GL, Zoppini G, Guidi
GC. Relation between red blood cell distribution width and
inammatory biomarkers in a large cohort of unselected out-
patients. Arch Pathol Lab Med 2009;133(04):628632
15 Yas JW, Horrow JC, Horne BD. Persistent increase in red cell size
distribution width after acute diseases: a biomarker of hypox-
emia? Clin Chim Acta 2015;448:107117
16 Chen J, Jin L, Yang T. Clinical study of RDW and prognosis in sepsis
newborns. Biomed Res 2015;25(04):576579
17 Norn H, Carlsson A. Reduced prevalence of metabolic acidosis at
birth: an analysis of established STAN usage in the total popula-
tion of deliveries in a Swedish district hospital. Am J Obstet
Gynecol 2010;202(06):546.e1546.e7
Sagheb et al.
18 American Academy of Pediatrics Subcommittee on Hyperbilirubi-
nemia. Management of hyperbilirubinemia in the newborn infant
35 or more weeks of gestation. Pediatrics 2004;114(01):297316
19 Christensen RD, Yaish HM, Henry E, Bennett ST. Red blood cell
distribution width: reference intervals for neonates. J Matern
Fetal Neonatal Med 2015;28(08):883888
20 Garofoli F, Ciardelli L, Mazzucchelli I, et al. The red cell distribution
width (RDW): value and role in preterm, IUGR (intrauterine growth
restricted), full-term infants. Hematology 2014;19(06):365369
21 Hampole CV, Mehrotra AK, Thenappan T, Gomberg-Maitland M,
Shah SJ. Usefulness of red cell distribution width as a prognostic
marker in pulmonary hypertension. Am J Cardiol 2009;104(06):
868872
22 zdemir R, Mutlu NM, zdemir M, et al. The importance of red
cell distribution width (RDW) in patient follow up in intensive
care unit(ICU). Acta Med Mediter 2016;32:349354
23 Gunbatar H, Sertogullarindan B, Ekin S, Akdag S, Arisoy A, Sayhan
H. The correlation between red blood cell distribution width
levels with the severity of obstructive sleep apnea and carotid
intima media thickness. Med Sci Monit 2014;20:21992204
24 Aksoy HT, Eras Z, Canpolat E, Dilmen U. Does red cell distribution
width predict mortality in newborns with early sepsis? Am J
Emerg Med 2013;31(07):1150
25 Abul Y, Ozsu S, Korkmaz A, Bulbul Y, Orem A, Ozlu T. Red cell
distribution width: a new predictor for chronic thromboembolic
pulmonary hypertension after pulmonary embolism. Chron
Respir Dis 2014;11(02):7381
American Journal of Perinatology