To evaluate the neurobehavioral outcomes of preterm infants with intrauterine growth restriction (IUGR), with and without prenatal advanced brain-sparing.
To evaluate the neurobehavioral outcomes of preterm infants with intrauterine growth restriction (IUGR), with and without prenatal advanced brain-sparing.
A cohort of IUGR infants (birth weight < 10th percentile with abnormal umbilical artery Doppler) born before 34 weeks of gestation was compared with a control group of appropriate-for-gestational age infants matched for gestational age at delivery. MCA pulsatility index was determined in all cases within 72 hours before delivery. Neonatal neurobehavior was evaluated at 40 weeks' ( ± 1) corrected age using the Neonatal Behavioral Assessment Scale. The effect of abnormal MCA pulsatility index (< 5th percentile) on each neurobehavioral area was adjusted for maternal smoking status and socioeconomic level, mode of delivery, gestational age at delivery, pre-eclampsia, newborn illness severity score and infant sex by multiple linear and logistic regression.
A total of 126 preterm newborns (64 controls and 62 IUGR) were included. Among IUGR fetuses, the proportion of abnormal MCA Doppler parameters was 53%. Compared with appropriate-for-gestational age infants, newborns in the IUGR subgroup with abnormal MCA Doppler had significantly lower neurobehavioral scores in the areas of habituation, motor system, social-interactive and attention. Similarly, the proportion of infants with abnormal neurobehavioral scores was significantly higher in the IUGR subgroup with abnormal MCA Doppler parameters in the areas of habituation, social-interactive, motor system and attention.
Abnormal MCA Doppler findings are predictive of neurobehavioral impairment among preterm newborns with IUGR, which suggests that this reflects an advanced stage of brain injury with a higher risk of abnormal neurological maturation. Copyright © 2011 ISUOG. Published by John Wiley & Sons, Ltd.
Preterm birth is a health problem associated with a high risk of cognitive1, sensory2, and behavioral3 disabilities, that can lead to academic underachievement and social and emotional difficulties4. Early interventions in preterm infants improve brain development5 as well as cognitive6 and behavioral7, 8 outcomes. Thus, identification of premature infants at higher risk of neurobehavioral problems forms the basis of targeted preventive interventions.
A substantial proportion of premature newborns are subject to intrauterine growth restriction (IUGR)9, which is associated with exposure to chronic hypoxia and undernutrition during fetal life due to placental insufficiency10. Evaluation of placental function by umbilical artery (UA) Doppler examination is the standard diagnostic procedure used to distinguish between IUGR and constitutional smallness11–13. Long-term outcome studies of such infants have revealed a specific profile of neurocognitive difficulties with poor executive functioning, cognitive inflexibility with poor creativity, and language problems14, 15. Studies of preterm IUGR babies have linked these difficulties during childhood with behavioral impairment already present during the neonatal period16–18.
During growth restriction due to placental insufficiency, hemodynamic adaptation occurs with blood flow redistribution preferentially to the brain, i.e. the brain-sparing effect. Controversy remains as to whether this phenomenon indicates a higher risk of brain injury or is a protective mechanism2, 19–23. In early stages, brain-sparing is expressed as a reduction in the Doppler cerebroplacental ratio, which is present in almost all early-growth restricted fetuses with placental insufficiency24. However, as placental insufficiency and hypoxia progress, a further decrease in resistance to blood flow in the middle cerebral artery (MCA) is observed24. Abnormal MCA Doppler findings indicate an advanced stage of brain-sparing, since it is correlated with established hypoxemia25, 26 and with a relative decrease in blood flow in the frontal areas in favor of the basal ganglia27. No information exists on the neurobehavioral consequences of advanced brain-sparing, as defined by abnormal MCA Doppler parameters, in preterm IUGR infants during the neonatal period, a time when environmental influences are still negligible.
This study aimed to evaluate the neurobehavioral outcomes of preterm infants with IUGR, with normal and abnormal prenatal MCA Doppler parameters.
Between November 2008 and April 2010 a cohort of consecutive singleton infants was created with the following inclusion criteria: i) birth weight < 10th percentile according to local standards28; ii) gestational age at birth less than 34 weeks; iii) UA Doppler pulsatility index (PI) > 95th percentile29 in at least two consecutive examinations 24 hours apart; and iv) positive end-diastolic velocities in both the UA and the ductus venosus before delivery. Controls were singleton appropriate-for-gestational age premature babies (birth weight ≥ 10th percentile according to local standards) resulting from spontaneous preterm deliveries in our institution during the same period and matched with cases by gestational age at delivery ( ± 4 days). Cases were grouped according to the presence of abnormal MCA Doppler parameters, defined as a MCA PI < 5th percentile for gestational age30. Exclusion criteria for both cases and controls were: (i) congenital malformations (including chromosomal abnormalities and infections); (ii) placental histological criteria for chorioamnionitis; (iii) infant death before 40 weeks of corrected age; and (iv) one of the following neurological complications before 40 weeks of corrected age: seizures31, intraventricular hemorrhage ≥ Grade III32 or periventricular leukomalacia33. Pregnancies were dated by the first-trimester crown–rump length measurement34. The study protocol was approved by the local ethics committee and parents provided written informed consent.
Prenatal Doppler ultrasound examinations were performed within 72 hours before delivery by one of two experienced operators (R.C.M. or F.F.) using a Siemens Sonoline Antares ultrasound machine (Siemens Medical Systems, Malvern, PA, USA) equipped with a 6–2-MHz linear curved-array transducer. Doppler recordings were performed in the absence of fetal movements and with voluntary suspended maternal breathing. Spectral Doppler parameters were measured automatically from three or more consecutive waveforms, with the angle of insonation as close to 0 as possible. A high-pass wall filter of 70 Hz was used to record low flow velocities and avoid artifacts.
The UA-PI was measured from a free-floating cord loop. The MCA-PI was measured at the level of its origin from the circle of Willis by means of a transverse view of the fetal head. Ductus venosus evaluation was performed in a mid-sagittal or a transverse section of the fetal abdomen, positioning the Doppler gate at its isthmic portion. Doppler parameters were converted into z-scores according to published normal ranges29, 35.
Neurobehavioral performance was evaluated at 40 weeks' ( ± 1) corrected age with the Neonatal Behavioral Assessment Scale (NBAS)36 which assesses both cortical and subcortical functions by evaluating 35 items. Of the 35 items, 27 are rated on a scale of 1 to 9 (9 being the best performance); eight are rated on a curvilinear scale and, according to the manual, are re-scored as linear on a 5-, 6- or 8-point scale. Items are grouped into six areas: habituation (to light, rattle, bell, and tactile stimulation of the foot); motor system (general tone, elicited activity, spontaneous activity, and motor maturity); social-interactive (responses to visual, animate and inanimate auditory stimuli, and alertness); organization of state (irritability, state lability, maximum excitation, and reaction time); regulation of state (self-quieting and hand-to-mouth responses); and autonomic nervous system (tremors, startles, and skin color). The social-interactive cluster was subscored for visual and auditory stimuli. Additionally, as recently reported by the authors of the NBAS37, an aggregation of individual items (alertness, quality of the alert responsiveness, and cost of attention) was used to evaluate attention capacity.
Each evaluation was performed by one of three trained examiners accredited by The Brazelton Institute (Harvard Medical School, Boston, MA, USA), each previously tested for reliability with an inter-rater reliability above 90%. The examiners were blinded to the study group and perinatal outcome. Neonates were assessed in the afternoon, between two feedings, in a small, semi-dark, quiet room with a temperature between 22 and 27 °C, and in the presence of at least one parent.
The behavioral items were converted into percentiles in reference to normal curves for our population38, and development in each area was considered abnormal if the corresponding score was below the 10th percentile.
Student's t-test or one-way analysis of variance (ANOVA) with post-hoc Bonferroni correction for multiple comparisons and Pearson's chi-square or Fisher's exact test were used to compare quantitative and qualitative data, respectively. Following standard methodology39–42, neurobehavioral outcome was adjusted by multiple linear and logistic regression for the following variables: (i) maternal smoking status during pregnancy: no smoking, one to nine cigarette(s)/day, ≥ 10 cigarettes/day; (ii) low socioeconomic level, defined as routine occupations, long-term unemployed, or never worked (UK National Statistics Socio-Economic Classification (http://www.ons.gov.uk/about-statistics/classifications/current/ns-sec/cats-and-classes/analytic-classes/index.html); (iii) mode of delivery (vaginal delivery vs. Cesarean section); (iv) gestational age at delivery; (v) pre-eclampsia (hypertension > 140/90 on two occasions 12 h apart and proteinuria > 300 mg/24 h); (vi) Score for Neonatal Acute Physiology version-II (SNAP-II)43 calculated within 12 h of admission; and, (vii) infant sex. Statistical analysis was performed using the Statistical Package for Social Sciences software (SPSS 15.0; SPSS Inc., Chicago, IL, USA).
A total of 140 preterm newborns were initially included (70 controls and 70 with IUGR). Subsequently, one infant in the control group was excluded because of histological signs of chorioamnionitis. Another control and two IUGR infants died before 40 weeks of corrected age. Two controls (each with Grade III intraventricular hemorrhage) and three IUGR infants (two with Grade III intraventricular hemorrhage and one with periventricular leukomalacia) were excluded because of neurological complications as defined in the exclusion criteria. Finally, the parents of two controls and three IUGR infants declined to participate after initial acceptance, leaving a final population of 126 preterm infants (64 controls and 62 with IUGR). A neurobehavioral assessment was scheduled for each of them at 40 ( ± 1) weeks of corrected age. The habituation category could not be assessed in 30 newborns (15 controls and 15 IUGR) due to the absence of a sleeping period during the evaluation.
Table 1 depicts the maternal characteristics of the population. None of the mothers used drugs other than tobacco and alcohol during pregnancy. Table 2 shows the perinatal outcomes of the groups. As expected, IUGR newborns had a lower birth weight and a smaller head circumference than controls. Pre-eclampsia and delivery by Cesarean section were more common among the mothers in both IUGR subgroups (with normal and abnormal MCA Doppler findings) than among the control mothers. Neonates in both IUGR subgroups had a higher frequency of adverse perinatal outcomes than controls as well as more severe neonatal illness, as reflected by the significantly higher SNAP-II values.
|Characteristic||Controls (n = 64)||Normal MCA (n = 29)||Abnormal MCA (n = 33)||P*|
|Primiparous||28 (43.8)||18 (62.1)||15 (45.5)||0.24|
|Non-Caucasian ethnicity||11 (17.2)||6 (20.7)||11 (33.3)||0.19|
|Maternal age (years)||31.5 ± 5.2||31.8 ± 5.7||31.9 ± 3.7||0.92|
|Body mass index (kg/m2) at admission||23.7 ± 3||23.5 ± 2.4||22.5 ± 1.7||0.12|
|Low socioeconomic level†||23 (35.9)||15 (51.7)||18 (54.5)||0.15|
|Non-smoking||53 (82.8)||23 (79.3)||28 (84.8)|
|1–9 cigarette(s)/day||7 (10.9)||5 (17.2)||4 (12.1)|
|11–19 cigarettes/day||4 (6.3)||0||1 (3)|
|≥ 20 cigarettes/day||0||1 (3.4)||0|
|Alcohol consumption > 170 g/week||2 (3.1)||2 (6.9)||1 (3)||0.65|
|Characteristic||Controls (n = 64)||Normal MCA (n = 29)||Abnormal MCA (n = 33)||P*||P†|
|Gestational age at delivery (weeks)||31.2 ± 2.4||30.9 ± 2.2||31.3 ± 2.6||1||1|
|Birth weight (g)||1674 ± 460||1032.9 ± 404||1075 ± 462||< 0.001||< 0.001|
|Birth-weight centile||44.3 ± 21.6||4.9 ± 3.5||4.4 ± 3.4||< 0.001||< 0.001|
|Head circumference (mm)||289 ± 2.7||263 ± 2.6||261 ± 2.7||< 0.001||< 0.001|
|Antenatal steroids||49 (76.6)||21 (72.4)||26 (78.8)||0.67||0.8|
|Pre-eclampsia||3 (4.7)||19 (65.5)||16 (39.4)||< 0.001||< 0.001|
|Cesarean section||29 (45.3)||28 (96.6)||27 (81.8)||< 0.001||0.001|
|Apgar score < 7 at 5 min||1 (1.6)||4 (13.8)||4 (12.1)||0.032‡||0.044‡|
|Umbilical artery pH < 7.15 at delivery||1 (1.6)||5 (17.2)||5 (15.2)||0.011‡||0.016‡|
|Neonatal unit stay (days)||8.3 ± 12.1||13.4 ± 20.2||13.3 ± 17.6||0.5||0.43|
|Severe respiratory distress syndrome||5 (7.8)||3 (10.3)||3 (9.1)||0.69||0.82|
|SNAP-II||11.9 ± 8.8||20.1 ± 16.4||21.2 ± 13.7||0.01||0.002|
Between both IUGR groups there were no significant differences in the mean z-scores of the UA (2.21 vs. 2.58; P = 0.17) or ductus venosus PI (1.46 vs. 1.54; P = 0.82).
Table 3 and Figure 1 detail the neurobehavioral outcomes by NBAS categories. The neurobehavioral scores of infants in the IUGR subgroup with normal MCA Doppler parameters did not differ significantly from those in the control group, but were significantly lower for infants in the IUGR subgroup with abnormal MCA Doppler, specifically in the areas of habituation, motor system, social-interactive (for auditory and visual stimuli) and attention. Similarly, the proportion of infants with abnormal neurobehavioral scores did not differ significantly between the control group and the IUGR subgroup with normal parameters, but was significantly higher in the IUGR subgroup with abnormal MCA Doppler, specifically in the areas of habituation, social-interactive, motor system and attention. Figure 2 shows the adjusted odds ratios for abnormal scores in each NBAS area in IUGR infants with abnormal MCA Doppler in relation to the control group.
|Assessment category||Controls (n = 64)||Normal MCA (n = 29)||Abnormal MCA (n = 33)||P*||P†|
|Habituation‡||6.52 ± 1.29||6.22 ± 1.25||5.59 ± 1.53||0.27||0.007|
|Motor system||5.39 ± 0.69||5.40 ± 0.5||4.88 ± 0.82||0.93||0.003|
|Social-interactive||5.86 ± 1.48||5.55 ± 1.62||4.79 ± 1.76||0.94||0.012|
|Visual||5.61 ± 1.66||5.23 ± 1.82||4.34 ± 1.87||0.61||0.004|
|Auditory||6.25 ± 1.45||6.00 ± 1.7||5.13 ± 1.74||0.86||0.006|
|State organization||3.87 ± 0.83||3.97 ± 0.64||3.80 ± 0.83||0.44||0.99|
|State regulation||4.45 ± 1.14||4.09 ± 1.42||4.33 ± 1.45||0.4||0.57|
|Autonomic system||5.52 ± 0.91||5.52 ± 0.94||5.50 ± 1.25||0.94||0.27|
|Attention||5.78 ± 1.48||5.68 ± 1.51||5.11 ± 1.86||0.5||0.013|
Premature infants with IUGR and abnormal UA Doppler parameters, a surrogate marker of placental insufficiency, are known to be at high risk of developing neurobehavioral and neurocognitive deficits44–46. Our study shows that, in this group of neonates, prenatal abnormal MCA Doppler parameters are associated with this higher risk, which suggests that this reflects an advanced stage of brain-sparing where disrupted neurological maturation secondary to hypoxia is already present.
During the second half of gestation, profound changes in brain organization occur, involving critical neural connections and myelination of important neural tracts47. It is not known how the susceptibility of the brain changes as such maturation progresses, but it is plausible that even mild degrees of hypoxia induce permanent changes resulting from the adaptation of the developing brain to a hypoxic and undernourished environment. IUGR fetuses with chronic hypoxemia exhibit a delay in development and behavioral milestones48–50. Our findings provide neonatal correlates for these previous observations.
The data provided by our study add to the body of evidence suggesting that increased brain perfusion is not an entirely protective mechanism. In keeping with this contention, we have demonstrated previously that full-term, small-for-gestational age (SGA) infants with Doppler signs of brain-sparing have reduced neurobehavioral competencies as newborns42, 51 and at 2 years of age52 than do their counterparts without Doppler signs of brain-sparing. Similarly, a large cohort study19 carried out in The Netherlands showed that brain-sparing during the third trimester of pregnancy was associated with a 23% higher than normal prevalence of behavioral problems at 18 months. Evidence of the effects of brain-sparing on neurobehavior in preterm infants with early-onset IUGR is scarce. A series including 31 preterm infants with signs of brain-sparing (18 born SGA) and 58 preterm infants without signs of brain-sparing revealed no differences between the groups in neurobehavior at 11 years23. Another study that included 16 preterm babies (13 born SGA) revealed differences between those with and without signs of brain-sparing according to the Mental Developmental Index at 2 years, which includes the assessment of habituation and social competencies53. However, these differences were not significant after adjustment for brain volume. Both studies included in the brain-sparing group a mixture of infants who were appropriately sized and small for gestational age. In addition, brain-sparing was defined as an abnormal cerebroplacental ratio, which is known to become abnormal earlier as placental insufficiency progresses24. Previous research from our group27, 51, 54 suggests that brain-sparing is a continuum, where increased brain perfusion is first detected by power-Doppler techniques followed by a reduction in the cerebroplacental ratio. At the other end of the spectrum, abnormal findings reflect a more advanced stage of brain hypoxia, which coincides with a relative decrease in blood supply to frontal areas in favor of the basal ganglia27. The results of the current study, where the effects of abnormal MCA blood flow on neurobehavior were more marked than in previous series that defined brain-sparing as an abnormal cerebroplacental ratio, are consistent with this concept. It is also noteworthy that IUGR groups with normal and abnormal MCA Doppler parameters did not differ in head circumference at birth, suggesting an independent effect of abnormal MCA blood flow on behavioral maturation. This may seem inconsistent with the long-recognized association between lagging head growth and neurodevelopment in the overall population of SGA babies55, but this association seems more pronounced in term neonates than in preterm neonates56. In fact, studies comparing symmetrical and asymmetrical growth restriction have failed to demonstrate differences in fetal acid–base status at the time of cordocentesis57 and other indices of perinatal outcome58, 59.
From a clinical perspective, the findings of this study are relevant to prenatal and neonatal management of early-onset IUGR. The evolution of medical technology and therapy has led to a substantial decrease in mortality among premature infants60. However, concern is growing with respect to the persistence of high rates of adverse neurobehavioral outcomes, which represent an educational4 and social61 burden. Identifying at-risk infants is essential to understanding the association between fetal wellbeing and later neurodevelopmental problems and forms the basis for possible preventive interventions. During fetal life, MCA Doppler parameters in preterm growth-restricted fetuses could help define management strategies and timely delivery. In the neonatal period, interventions aimed at reducing stress in the premature infant have been shown to improve frontal region cerebral white matter development5. In low-risk premature infants, it has been reported that individualized developmental interventions prevent short-term neurobehavioral dysfunction8. Furthermore, as these infants are at risk of impaired neurodevelopmental outcome they might benefit from early educational intervention, which has been found beneficial in selected cases6, 7.
This study had certain limitations. First, although NBAS is a reference standard for evaluating a newborn's capacity to respond to the environment and therefore reflects brain maturation, it assesses only neurobehavior and not cognitive function62. However, several studies have demonstrated the correlation between neonatal neurobehavior and later neurocognitive development in infants born preterm16, 18, 63 and at full term64, 65. Furthermore, mothers in the two study groups were not matched on some perinatal conditions. Especially relevant is the higher prevalence of pre-eclampsia in the IUGR subgroup with normal MCA Doppler parameters. Our finding of a better neurobehavioral outcome in this subgroup compared to the IUGR subgroup with abnormal MCA Doppler parameters could be explained by a protective effect of pre-eclampsia (prompting delivery because of maternal indications before fetal deterioration occurs), or secondary to the drugs commonly administered in such cases, e.g. magnesium sulphate. Although we adjusted for some of these potential confounders, we cannot rule out some residual confounding effects. Finally, because of the sample size, our study could be underpowered for detection of some associations and for stratification of the sample according to the presence of pre-eclampsia.
In summary, this study shows that abnormal MCA Doppler, as a sign of advanced brain-sparing, is a major contributor to neurobehavioral impairment among preterm infants with IUGR.
This study was funded by the ‘Fondo de Investigación Sanitaria’ (PI/060347) (Madrid, Spain); Cerebra Foundation for the Brain Injured Child (Carmarthen, Wales, UK); and Thrasher Research Fund (Salt Lake City, UT, USA). R.C.M. is supported by a Marie Curie grant from the EC.