You are in:  
Prevalence of neonatal ultrasound brain lesions in premature infants with and without intrauterine growth restriction. 

Aim

To compare the prevalence of transient periventricular echodensities (TPE), periventricular leukomalacia (PVL) and hemorrhagic brain lesions (HBL) in singleton intrauterine growth restricted (IUGR) infants and in those appropriate for gestational age (AGA) infants matched for gestational age at delivery.

Introduction

This prospective study compared a consecutive sample of 35 singleton premature infants with a prenatal diagnosis of IUGR diagnosed before 34 weeks of gestation, against 35 AGA infants matched in the neonatal intensive care unit (NICU) for sex and gestational age (GA, +/- 2 weeks) at delivery. IUGR was defined as a US fetal weight estimation below the 10th percentile for gestational age confirmed at birth, and an abnormal blood flow in the umbilical artery (pulsatility index > 2 standard deviations). Indications for delivery included fetal deterioration or maternal decompensation. None of the pregnancies were allowed to continue after 34 weeks as they were considered severely affected. At this gestational age the benefits of delivery outweigh the risks of continuing with the intrauterine insult.

Infants with chromosomal, genetic or structural defects and signs of intrauterine infection or neonatal early onset sepsis, as defined by positive blood culture within the first 72 hours of life, were not eligible for this study.

Prenatal and neonatal data were recorded, including the use of antenatal steroids, gestational age at delivery, birth weight, duration of NICU stay, the presence or absence of respiratory distress syndrome, need for, and duration of oxygen supplementation and mechanical ventilation, use of inotropic drugs, the presence of bronchopulmonary dysplasia, patent ductus arterious, necrotising enterocolitis, hypoglycemia, neonatal early-onset sepsis, index for illness severity (Clinical Risk Index for Babies-CRIB II) on admission, and neonatal death. CRIB II is a marker of illness severity in infants born at J 32 weeks of gestational age, that offers an adjusted estimation of neonatal risk. This index was calculated 1 hour after admission to the NICU and included 5 items: sex, birth weight, gestational age, temperature on admission, and base excess. The CRIB II score ranges from 0 to 27, with the highest value corresponding to the worst condition.

All participants were studied using the same protocol, consisting of brain US scans on the 3rd (US-1) and 14th days (US-II) after delivery and at termequivalent age (US-III). US images were obtained with Acuson Sequoia 512 SHA US equipment (Medisales Inc., Los Alamitos, CA, USA) using an 8-5 MHz sector transducer. Examination included 8 coronal scans, one mid-sagittal and three parasagittal views through the anterior fontanelle. Scans through the posterior and mastoid fontanelles were added to improve visualisation of the most dependent portion of the occipital horns, choroid plexus and particularly, the cerebellum. All images were evaluated in duplicate by two operators (GE and NP) trained in the evaluation of neonatal brain US. In 30 cases (15 normal and 15 abnormal), agreement between observers was calculated.

IVH was classified into four grades according to the Papile description: grade I, IVH confined to the germinal matrix; grade II, IVH with mild or no dilatation of the lateral ventricle; grade III, IVH with patent ventricular dilatation, and grade IV, IVH extending into adjacent brain parenchyma. Cerebellar hemorrhage was suspected when hyper- or hypoechoic lesions were observed in the cerebellum. TPE were defined according to Appleton et al. and Resch at al. as confluent areas of increased echogenicity in the periventricular region with US brightness equal, or similar to that in the choroid plexus and lasting less than 14 days. PVL was classified in four grades: PVL-I, as echodensities persisting for more than 14 days, without cavitation; PVL-II, as echodensities evolving into small cysts; PVLIII, as echodensities evolving into extensive periventricular cystic lesions, and PVLIV, as echodensities evolving to large cystic lesions extending into deep white matter.

In brief, 35 IUGR and 35 AGA singleton infants born between 24 and 34 weeks gestational age were studied. The presence of TPE, PVL, and HBL was assessed with ultrasound (US) at day 3 (US-I), 2 weeks (US-II) after delivery, and at termequivalent age (US-III).

Data Analysis

Comparison between groups was carried out using a two-sided Mann- Whitney U test for medians and Fisher’s exact test for proportions. Agreement between observers was evaluated with Kappa-Cohen test and 95% confidence intervals (95% CI). All statistical analyses were calculated with the SPSS 13.0 (SPSS for Windows, SPSS Inc, Chicago, USA) statistical software. Thirty-five IUGR and 35 AGA singleton infants born between 24 and 34 weeks gestational age were studied. The presence of TPE, PVL, and HBL was assessed with ultrasound (US) at day 3 (US-I), 2 weeks (US-II) after delivery, and at term-equivalent age (US-III).

Results

Neonatal death occurred in 5/35 (14%) of IUGR and in 1/35 (3.0%) of AGA infants. The principal causes of death in IUGR infants included acute respiratory distress syndrome (n=3) and severe sepsis (n=2). In AGA infants the cause of death was cardiovascular organ dysfunction (n=1).

Thirty-five infants in each group underwent US-I; 32 IUGR and 31 AGA underwent US-II, and 29 IUGR and 29 AGA infants underwent US-III. The reduction in the number of evaluated patients was due to death or to parental refusal to continue follow-up. In patients with lesions, US scans were repeated at the discretion of the managing physicians to monitor abnormal features. Agreement between observers showed a Kappa result of 0.64 (95% CI 0.55-0.72) which is considered as good agreement, according to Landis et al.

There were no differences in neonatal outcomes except for reduced birth weight (p=0.001) and increased use of antenatal steroids in IUGR infants compared with the AGA group (p= 0.05). Although the sample size was too small for meaningful comparisons, neonates who received antenatal steroids showed no apparent differences in neonatal outcome, compared with those who did not. There was a non-significant trend in IUGR infants toward a higher prevalence of necrotising enterocolitis, hypoglycemia, bronchopulmonary dysplasia, and patent ductus arteriosus. There was no correlation between late onset sepsis and ultrasound abnormalities in both groups. CRIB II scores were similar in both groups.

The prevalence of TPE was significantly increased in IUGR infants at US-I (18/35 (51.4%) vs. 8/35 (22.9%) p= 0.02). After the first evaluation, progression from TPE to PVL occurred in 8/18 infants in the IUGR group, and in 1/8 in the AGA group. Consequently, the prevalence of PVL was significantly increased in IUGR vs. AGA infants both at US-II (8/32 vs. 1/31 respectively, p=0.03) and at US-III (8/29 vs. 1/29, p=0.02). In the group of IUGR, there were 18/35 TPE lesions at the time of the first US that evolved to PVL in 8 cases at US-III. In the AGA group at US-I there were 8/35 TPE which evolved to PVL in 1 case at US-III. There was no difference in the severity of PVL between both groups, at US-II and III only one IUGR neonate showed PVL II. No infant from both groups showed PVL without TPE.

HBL were decreased in IUGR neonates at US-I (2/35 vs. 8/35 respectively, p=0.08). This difference was less pronounced on US-II (4/32 vs. 2/31 respectively, p= 0.7) and the relation was reverted by US-III, although the difference did not reach statistical significance (5/29 vs. 1/29 respectively, p= 0.2).

In brief, IUGR neonates had an increased prevalence of TPE at US-I (18/35 vs. 8/35p= 0.02) and an increased prevalence of PVL at US-II (8/32 vs. 1/31, p=0.03) and US-III (8/29 vs. 1/29, p= 0.02). No significant differences in the prevalence of HBL were found between the two groups.

Discussion

The results of this study suggest that IUGR infants have an increased prevalence of neonatal brain scans consistent with white matter injury, compared with AGA premature infants. IUGR neonates had a significantly increased prevalence of TPE and PVL. Although the impact of TPE is milder than that of PVL, TPE has been associated with abnormal motor and cognitive neurodevelopment, both suggesting white matter injury. The rate of subsequent development of cerebral palsy in these infants may range from 4% to 8%.

Our findings are consistent with recently published long-term follow-up studies in infants with IUGR. This phenomenon is associated with specific suboptimal development affecting motor and cognitive functions, suggesting white matter abnormalities. In addition, neurostructural studies with US have shown the presence of selective growth restriction in the frontal lobe. A quantitative volumetric three-dimensional magnetic resonance (MR) study demonstrated an early reduction in intracranial volume and in cerebral cortical grey matter, which was correlated with the behavioural response of the term neonate. Finally, studies using MR-diffusion weighted imaging and deformationbased morphometry have shown diffuse white matter injury associated with deep grey matter growth failure, possibly representing a disturbance in corticothalamic connectivity.

Our results apparently differ from those of previous reports, which failed to find significant differences in the prevalence of white matter lesions in IUGR preterm neonates. The differences observed in this study can be mainly related to intrauterine chronic hypoxia, as all cases with signs of intrauterine infection were excluded. In addition, in the majority of cases with signs of late infection there were no US abnormalities, and those presenting either with TPE or PVL were similar in IUGR and AGA neonates.

However, we believe that the design of this study presents two important differences with respect to previous reports. Firstly, we performed serial US scans, which substantially increased the detection rate and consequently, allowed the prevalence of PVL to be established more reliably. Secondly, TPE were actively searched for and recorded, an approach not used in previous studies in IUGR preterm neonates.

Although we found no significant differences in the prevalence of hemorrhagic lesions, we observed an increasing trend in the AGA group. These findings are in line with the results of other studies. A lower presence of hemorrhagic events in the case of IUGR may correspond to different local autoregulatory and vascular reactivity developed by these infants in response to adverse intrauterine environment and chronic hypoxia.

Different patterns of hypoxic ischemic and hemorrhagic lesions in infants with IUGR may correspond to a broad variety of cellular, metabolic, functional and vascular responses that attempt to maintain cerebral homeostasis during chronic hypoxia. Significant changes take place in the cerebral circulation related to the composition and reactivity of the vessels, and to the expression and regulation of neuronal and glial proteins, thus altering the characteristics of the cerebral vessels and tissues in the developing brain.

This study has some limitations. Firstly, the relatively small sample may have prevented statistical differences in some comparisons from being observed. In addition, the diagnosis of TPE is inevitably established by subjective comparison of the US density in the periventricular regions with that of the choroid plexus. This method is inevitably subject to inter-observer variability, an insurmountable problem in any study based on subjective findings. To reduce this variability, we followed the recommendations made by deVries, consisting of well-defined US settings. In addition, the results obtained by the Kappa analysis showed a good agreement, and the confidence intervals reinforce the fact that operators can have a moderate to good agreement following proper training. An advantage of this study is that the study groups were comparable in terms of gestational age at delivery and morbidity, as suggested by similar CRIB II scores, providing further evidence that the observed differences were probably due to the presence, or absence of growth restriction. Moreover, the longitudinal US evaluation allowed us to detect early US findings associated with white brain matter lesions.

Conclusion

The findings of this study suggest that IUGR results in an increased prevalence of white matter damage in preterm infants and that the presence of TPE should be considered in future studies evaluating abnormal findings in IUGR infants. These findings should be confirmed and expanded in larger studies to refine knowledge of the natural history of perinatal brain damage.

The study protocol was approved by the local ethics committee and written informed consent was obtained from the parents or legal guardians of all participants.

This information is not meant to replace the advice of any physician or qualified health professional. The information provided by Cerebra is for information purposes only and is not a substitute for medical advice or treatment for any medical condition. You should promptly seek professional medical assistance if you have concerns regarding any health issue.

Page last updated: 16/12/2011 15:46 
 
 
Downloads
 
© 2012 Cerebra | Registered charity No: 1089812 | Company Limited by Guarantee, registered in England and Wales 4336208