SEARCH

SEARCH BY CITATION

Keywords:

  • Cognitive Development;
  • CP ;
  • Follow-up

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Aim

To study the prognostic value of MRI in preterm infants at term equivalent age for cognitive development at 5 years of age.

Methods

A total of 217 very low birth weight/very low gestational age infants who all received brain MRI at term equivalent age were categorized into 4 groups based on the brain MRI findings. Cognitive development was assessed at 5 years of chronological age by using a short form of Wechsler Preschool and Primary Scale of Intelligence – Revised. This information was combined with neurosensory diagnoses by 2 years of corrected age.

Results

Of all infants 31 (17.0%) had Full Scale Intelligence Quotient (FSIQ) <85, 14 (6.5%) had cerebral palsy and 4 (1.8%) had severe hearing impairment. A total of 41 (22.0%) infants had some neurodevelopmental impairment at 5 years of age. Considering cognitive outcome (FSIQ <85), the positive predictive value of several major MRI pathologies was 43.8%, and the negative predictive value of normal finding or minor pathologies was 92.0% and 85.7%, respectively.

Conclusion

The MRI of the brain at term equivalent age may be valuable in predicting neurodevelopmental outcome in preterm infants by 5 years of age. The findings should always be interpreted alongside the clinical information of the infant. Furthermore, MRI should not replace a long-term clinical follow-up for very preterm infants.

Abbreviations
CP

cerebral palsy

FSIQ

Full Scale Intelligence Quotient

MRI

magnetic resonance imaging

NDI

neurodevelopmental impairment

NPV

negative predictive value

PPV

positive predictive value

VLBW

very low birth weight

VLGA

very low gestational age

V/B

ventricular/brain

Key notes
  • The brain MRI in preterm infants at term equivalent age provides additional information aiding the clinician to identify those infants who are later at a high risk of neurodevelopmental impairment.
  • The interpretations of MRI findings should always be carried out together with the clinical picture of the infant.
  • A long-term follow-up of preterm infants is needed, irrespective of brain imaging findings.

Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

There has been conflicting views of the routine use of brain magnetic resonance imaging (MRI) as a part of the assessment protocol in the care of a very preterm infant [1]. The MRI can be helpful in identifying children for closer follow-up and in directing research for focused risk groups, in quality surveillance purposes of the neonatal care and in providing parents information of either decreased or increased developmental risks. Recently, for example, performing MRI at term without asking parental permission was questioned [2]. Furthermore, there are situations when MRI increases rather than decreases parents' concerns and confusion, if the knowledge of the prognostic significance of the findings is not sufficient. To obtain more accurate information of the prognostic value of MRI findings, we need to differentiate between those findings that are associated with major consequences and those that do not cause significant clinical concern. This requires detailed information of long-term follow-up studies which have evaluated both neurosensory and cognitive outcomes of preterm infants with neonatal MRI [3]. We have previously reported the positive (PPV) and negative (NPV) predictive values of abnormal MRI findings on neurodevelopmental impairments (NDI) at the corrected age of 2 years in this cohort including 182 very low birth weight (VLBW) infants. A major pathological MRI finding at term had 33.3% PPV and normal MRI finding 98.1% NPV on NDI at 2 years of age [4]. These findings are consistent with earlier studies showing that abnormal MRI findings at term equivalent age may predict an adverse neurodevelopmental outcome at 2 years of age [5].

Our aim was to study the prognostic value of MRI at term equivalent age in very preterm infants using cognitive test results at 5 years of age as an end point. We hypothesized that major brain pathologies in MRI in very preterm infants at term equivalent age predict neurodevelopmental impairments at 5 years of age.

Patients and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

Participants

This prospective study is a part of the multidisciplinary PIPARI Study (The Development and functioning of Very Low Birth Weight Infants from Infancy to School Age). The PIPARI Study group consists of VLBW or very low gestational age (VLGA) infants born between 2001 and 2006, in the Turku University Central Hospital. The inclusion criteria from 2001 to the end of 2003 included birth weight ≤1500 grams in preterm infants (born <37 gestational weeks). From the beginning of 2004, the inclusion criteria were expanded to include all infants below the gestational age of 32 weeks at birth, even if the birth weight exceeded 1500 g. In addition, at least one of the parents had to speak either Finnish or Swedish. Fig. 1 shows the flow chart of the participants. All parents provided written consent after receiving oral and written information. The PIPARI Study protocol was approved by the Ethics Review Committee of the Hospital District of the South-West Finland in December 2000.

image

Figure 1. The flow chart of the participants, mean birth weights (BW) and gestational ages (GA) in weeks.

Download figure to PowerPoint

Magnetic resonance imaging of the brain

The MRI of the brain was performed at term. The imaging took place during postprandial sleep without any pharmacological sedation or anaesthesia. The infants were swaddled to calm them and to reduce movement artefacts in the imaging. A pulse oximeter was routinely used during MRI examinations. Ear protection was also used (3M Disposable Ear Plugs 1100; 3M, Brazil and Wurth Hearing protector, Art.-Nr. 899 300 232, Wurth, Austria). For 125 infants born between 2001 and 2003, the MRI equipment was an open 0.23-T Outlook GP (Philips Medical, INC, Vantaa, Finland) equipped with a multipurpose flexible coil fitting the head of the infant, until it was upgraded to the 1.5-T Philips Intera (Philips Medical Systems, Best Netherlands) for the remainder of the study infants (n = 73) born between 2004 and 2006 [6].

Axial T2-weighted images, coronal three-dimensional T1-weighted images and coronal T2-weighted images of the entire brain were obtained when using the 0.23-T equipment. With the 1.5-T equipment, axial T2-weighted, axial T1-weighted and sagittal T2-weighted images were obtained. All of the sequences were optimized for imaging of the term infant brain. The total imaging time was about 25 min. The extracerebral space was measured manually from the MRIs. A cut-off value of 4 mm was used according to the study by McArdle [7]. The width of the extracerebral space was measured in front of the frontal lobe, where the extracerebral fluid space is widest. The group of infants with an extracerebral space of 5 mm was analysed separately, because accuracy of the measurement was 1 mm. V/B ratio refers to [the width of the frontal horns of the lateral ventricles] divided by [the width of the brain tissue at the same plane of cerebral image] [6].

The brain MRI was evaluated by one neuroradiologist (R. Parkkola) blinded to both the clinical data and to the result of the brain ultrasound examinations.

Classification of the study groups

The brain MRI findings were categorized into normal findings, minor pathologies and major pathologies as described previously [6] (Table 1). To evaluate the relation between brain pathology and neurodevelopmental outcome, the infants were categorized into 4 groups, based on the MRI findings at term: (i) normal group, (ii) one or more minor brain pathologies, (iii) one major brain pathology and (iv) several major brain pathologies. If a patient had both minor and major findings, the classification was done according to the number of major pathologies: one major pathology (n = 33, 15.2%) and several major pathologies (n = 25, 11.5%).

Table 1. The classification of MRI findings
Normal findings
Normal brain anatomy (cortex, basal ganglia and thalami, posterior limb of internal capsule, white matter, germinal matrix, corpus callosum and posterior fossa structures)
A width of extracerebral space <5 mm, ventricular/brain ratio <0.35
No ventriculitis
Minor pathologies
Consequences of intraventricular haemorrhages grade 1 and 2
Caudothalamic cysts
A width of the extracerebral space of 5 mm
A V/B ratio of 0.35
Major pathologies
Consequences of intraventricular haemorrhages grade 3 and 4
Injury in cortex, basal ganglia, thalamus or internal capsule, with injury of corpus callosum, cerebellar injury, white matter injury
An increased width of extracerebral space >5 mm
A V/B ratio >0.35, ventriculitis
Other major brain pathology (infarcts)

Neurodevelopmental outcome

The children's cognitive level at 5 years of chronological age (+0–2 months) was evaluated by a psychologist using a short form of Wechsler Preschool and Primary Scale of Intelligence – Revised, Finnish translation (WPPSI-R) [8]: subtests are information, sentences, arithmetic, block design, geometric design and picture completion, and Full Scale Intelligence Quotient (FSIQ) was estimated (normal M = 100, SD 15.0). A quotient of ≥85 (>−1.0 SD) was considered normal intelligence, a quotient of 70–84 (−2.0 SD to −1.0 SD) slightly below normal and a quotient of ≤69 (<−2.0 SD) significantly below normal [9].

In addition, neurodevelopmental impairment (NDI) included at least one of the following findings: FSIQ score <85, cerebral palsy (CP), severe hearing impairment or severe visual impairment. We chose to use a cut-off of 85 for FSIQ, which is close to a −2.0 SD level in our control group, a regional cohort of full-term Finnish children (M = 111.7, SD 14.8) [9]. A diagnosis of CP was determined during a systematic clinical follow-up by 2 years of corrected age. Severe hearing impairment was defined as hearing loss requiring amplification in at least one ear or hearing impairment with a cut-off of 40 dB. Hearing was systematically screened in early infancy (at 1 month of corrected age) by using brain stem auditory evoked potentials (BAEP). Severe visual impairment was categorized as a visual acuity <0.3, or blindness [10].

Statistical analysis

The NPV was defined as the percentage of children with no or minor brain pathology in the MRI, resulting in an average developmental outcome without NDI. The PPV was defined as the percentage of children with one or more major brain pathologies in the MRI, resulting in an abnormal developmental outcome with NDI.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

The neonatal characteristics are shown according to the brain magnetic resonance imaging (MRI) group at term equivalent age in Table 2.

Table 2. Neonatal characteristics are shown according to the brain magnetic resonance imaging (MRI) group at term equivalent age
 Normal findingsOne or more minor pathologiesOne major pathologySeveral major pathologies
Birth weight, mean (SD) [minimum, maximum], g1195.8 (311.4) [580, 2120]1106.5 (347.8) [400, 1940]1012.9 (340.2) [525, 1730]1014.6 (315.7) [560, 1675]
Gestational age at birth, mean (SD) [minimum, maximum], wk29.7 (2.6) [24.0, 35.9]28.5 (2.7) [24.0, 33.0]28.3 (2.5) [23.0, 32.1]28.2 (3.1) [23.4, 35.1]
Males, n (%)63 (52.5)22 (56.4)24 (72.7)13 (52.0)
Females, n (%)57 (47.5)17 (43.6)9 (27.3)12 (48.0)
Cesarean section, n (%)79 (65.8)25 (64.1)21 (63.6)9 (36.0)
Small for gestational age, n (%)45 (37.5)14 (35.9)14 (42.4)8 (32.0)
Bronchopulmonary dysplasia, n (%)7 (5.8)9 (23.1)8 (24.2)5 (20.0)
Sepsis, n (%)18 (15.1)7 (18.4)7 (21.2)3 (12.0)
Operated necrotizing enterocolitis, n (%)4 (3.4)1 (2.6)4 (12.5)1 (4.0)
Laser-treated retinopathy, n (%)1 (0.9)3 (7.9)3 (9.4)1 (4.0)

One hundred and twenty (55.3%) of the infants (n = 217) had normal MRI, 39 (18.0%) one or more minor, 23 (10.6%) one major, 9 (4.2%) several major and 27 (12.4%) minor and major findings in the MRI at term equivalent age.

One hundred and seventy-eight of 217 (82.0%) WPPSI-R assessments were completed. There were 4 (1.8%) children who were too severely handicapped to be assessed. They were included in the analysis as having a significant cognitive delay (FSIQ <70) at the age of 5 years. Of all the infants, 31 (17.0%) had FSIQ<85, 14 (6.5%) had CP, and 4 (1.8%) had severe hearing impairment. There were no children with severe visual impairment. A total of 41 (22.0%) infants had NDI at 5 years of age. (Table 3).

Table 3. The prevalence of Full Scale Intelligence Quotient (FSIQ) <85, cerebral palsy (CP), severe hearing impairment (use of hearing aid) and neurodevelopmental impairment (NDI) (one or more of the three above-mentioned outcomes) are shown according to the brain magnetic resonance imaging (MRI) group at term equivalent age
 FSIQ<85 (n = 31, 17.0%)CP (n = 14, 6.5%)Severe hearing impairment (n = 4, 1.8%)NDI (n = 41, 22.0%)
Normal findingsNPV = 92.0%NPV = 99,2%NPV = 100%NPV = 91.0%
(n = 100, 55.0%)(n = 120, 55.3%)(n = 120, 55.3%)(n = 100, 53.8%)
One or more minor pathologiesNPV = 85,7%NPV = 100%NPV = 100%NPV = 85.7%
(n = 35, 19.2%)(n = 39, 18.0%)(n = 39, 18.0%)(n = 35, 18.8%)
One major pathologyPPV = 35.5%PPV = 6.1%PPV = 6.1%PPV = 38.7%
(n = 31, 17.0%)(n = 33, 15.2%)(n = 33, 15.2%)(n = 31, 16.7%)
Several major pathologiesPPV = 43.8%PPV = 44.0%PPV = 8.0%PPV = 75.0%
(n = 16, 8.8%)(n = 25, 11.5%)(n = 25, 11.5%)(n = 20, 10.8%)

The most common major pathologies were white matter injury (n = 29, 13.4%), capsula interna injury (n = 21, 9.7%), ventriculitis (n = 21, 9.7%) and a ventricular/brain (V/B) ratio >0.35 (n = 21, 9.7%). There were 7 different single major findings (n = 33, 15.2%) (capsula interna injury, white matter injury, ventriculitis, corpus callosum injury, haemorrhage in posterior fossa structures, V/B ratio >0.35, extracerebral space >5 mm) associated with NDI in some children. The minor pathologies were consequences of intraventricular haemorrhages grade 1 (n = 49, 22.6%), an extracerebral space of 5 mm (n = 15, 6.9%) and a V/B ratio of 0.35 (n = 9, 4.1%).

Considering cognitive outcome (FSIQ <85), the PPV of several major findings was 43.8%, and the NPV of normal or minor findings was 92.0% and 85.7%, respectively (Table 3). The distribution of FSIQ at 5 years of age in MRI groups is shown in Table 4. Considering NDI, the PPV of several major findings was 75.0%, and the NPV of normal or minor findings was 91.0% and 85.7%, respectively (Table 3).

Table 4. The mean values (SD, median and interquartile) of Full Scale Intelligence Quotient (FSIQ) at 5 years of age are shown by categories of MRI findings at term equivalent age
 MeanSDMedianInterquartile
Normal findings (n = 100)104.215.0104.593.5–113.0
One or more minor pathologies (n = 35)102.415.6105.092.0–116.0
One major pathology (n = 29)94.618.893.082.0–107.0
Several major pathologies (n = 14)86.823.487.570.0–104.0

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This study shows that brain MRI of very preterm infants at term equivalent age provides additional information in predicting neurodevelopmental outcome at 5 years of age. This information was derived from a regional population of preterm infants with a very high coverage of both MRI and follow-up.

The major MRI pathologies aid clinicians in identifying those infants who are later at a high risk of developmental problems, even though good functional outcome can be achieved despite major lesions. Our findings are in agreement with previous studies with shorter follow-up times [11]. The major lesions seen in the brain MRI indicate that the child needs to be followed by a child neurologist, and the possibilities for early interventions have to be evaluated.

Current practice in neonatal care includes brain ultrasound imaging. Therefore, MR imaging provides additional information to ultrasound images. Ultrasound is more sensitive in detecting transient early findings such as small intraventricular haemorrhages. Brain ultrasound examinations can also be used to identify patients requiring MRI, for example, ventriculomegaly indicates a high risk for additional pathologies seen by MRI [6]. Although some significant parenchymal lesions can be detected by ultrasound [12], MRI is more sensitive in identifying small lesions in the white matter [13-15] which were shown in this study to be significant for the later development of the child. Our findings are supported by a previous study [16] showing an association between diffuse white matter injury and adverse neurodevelopmental outcome at 2 years of age. MRI is also more sensitive in finding cerebellar lesions [17], which can be difficult to discover in a routine neonatal ultrasound examination, although there are specialists who do master ultrasound scanning of the cerebellum. The value of ultrasound examinations increases if carried out repeatedly in short intervals by an experienced specialist. However, not all units have the resources for weekly ultrasound examinations exceeding routine levels. Although the interpretation of MRI also requires an expert neuroradiologist, the readings can be centralized without moving the patient from the closest hospital with MRI capacity.

Normal findings or minor pathologies in MRI provide information for families of the good developmental capacity of their child. The NPV of normal MRI was close to 100%. The PPV for any major developmental problem by 5 years of age was also high (75.0%) when MRI findings were graded by severity. Considering only cognitive outcome, the PPV was lower. This is understandable as cognitive impairments have a multifactorial background. It must also be noted that normal FSIQ does not exclude specific neuropsychological impairments.

An MRI should be offered to families stating the good NPV and the PPV of 44%. From the perspective of the family, normal MRI is likely to be relieving. Major brain pathologies are often detected by ultrasound which increases parental anxiety for their child's future development. The combination of ultrasound imaging and brain MRI at term equivalent age provides better definition of the nature of the brain lesion, which may also be relieving from parental point of view. In our study, the predictive efficacy of the brain MRI was better compared with different available perinatal scoring systems of clinical data [18]. However, the image of brain structures never equals brain function, and therefore, development can be normal despite major pathologies.

We agree that ultrasound and MRI are complementary techniques in neuroimaging for preterm infants [19]. If brain MRI is used routinely, as a part of follow-up protocol [15], the findings have to be interpreted systematically by an experienced neuroradiologist including the most common minor and major pathologies. The clinical significance of the brain MRI should always be interpreted alongside information from ultrasounds, standardized neurological examinations and the medical history of the infant. More research is needed on interventions to optimize the development in those children with major structural brain lesions. Irrespective of brain imaging findings, very preterm infants and their families benefit from a long-term clinical follow-up.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References

This work was supported by a grant from the Foundation for Paediatric Research. The PIPARI Study Group: Mikael Ekblad, MD; Satu Ekblad, RN; Eeva Ekholm, MD, PhD; Leena Haataja, MD, PhD; Mira Huhtala, MD; Pentti Kero, MD, PhD; Hanna Kiiski-Mäki, PhD; Riikka Korja, PhD; Harry Kujari, MD; Helena Lapinleimu, MD, PhD; Liisa Lehtonen, MD, PhD; Tuomo Lehtonen, MD; Virva Lepomäki, MA; Marika Leppänen, MD; Annika Lind, PhD; Hanna Manninen, MD; Jaakko Matomäki, MSc; Jonna Maunu, MD, PhD; Petriina Munck, PhD; Sirkku Setänen, MD; Laura Määttänen, MD; Pekka Niemi, PhD; Anna Nyman, MA; Pertti Palo, MD, PhD; Riitta Parkkola, MD, PhD, Jorma Piha, MD, PhD; Liisi Rautava, MD, PhD; Päivi Rautava, MD, PhD; Milla Ylijoki, MD, PhD; Hellevi Rikalainen, MD, PhD; Katriina Saarinen, Physiotherapist; Elina Savonlahti, MD, Matti Sillanpää, MD, PhD; Suvi Stolt, PhD; Päivi Tuomikoski-Koiranen, RN; Timo Tuovinen, BA; Anniina Väliaho, MA, Tuula Äärimaa, MD,PhD.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Patients and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References