- Email: [email protected]
MRI Patterns of brain injury and neurodevelopmental outcomes in neonates with severe anaemia at birth
Early Human Development 105 (2017) 17–22
Contents lists available at ScienceDirect
Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev
MRI Patterns of brain injury and neurodevelopmental outcomes in neonates with severe anaemia at birth Begoña Loureiro a,⁎, Miriam Martinez-Biarge b, Francesca Foti c, Maria Papadaki d, Frances M Cowan b, Courtney J Wusthoff e a
Departamento de Pediatría, Hospital Universitario Cruces, Barakaldo, Spain Department of Paediatrics, Hammersmith Hospital, Imperial College, London, UK Department of Mental Health, Azienda Unita Sanitaria Locale di Reggio Emilia, Emilia-Romagna, Italy d Own private paediatrician practice, Cholargos Attica, Greece e Division of Child Neurology, Stanford University, CA, USA b c
a r t i c l e
i n f o
Article history: Received 22 September 2016 Received in revised form 31 December 2016 Accepted 4 January 2017 Keywords: Neonatal anaemia Feto-maternal haemorrhage Neonatal encephalopathy White matter injury Magnetic resonance imaging Developmental outcomes
a b s t r a c t Aims: To define patterns of brain injury and associated neurodevelopmental outcomes in infants with severe neonatal anaemia. Methods: We studied 20 infants with severe anaemia at birth (haemoglobin b 7 g/dL). Clinical details were analysed for causes of anaemia and co-morbidities. All had early brain magnetic resonance imaging (MRI) scans, which were reviewed for injury pattern. Neurodevelopmental outcomes were assessed at a median age of 24 months. Results: The aetiology of the anaemia was feto-maternal haemorrhage in 17 and antepartum haemorrhage in 3 infants. The predominant site of injury was the white matter, which was affected in all infants, with differing grades of severity and with cystic evolution in 45%. Only one infant showed an injury pattern typical of an acute severe hypoxic-ischaemic insult. Outcomes correlated closely to the severity of MRI findings. Cerebral palsy was seen only with the most severe neuroimaging patterns (n = 6). Global developmental delay, learning or behavioural problems and seizures were common with moderate injury. Visual impairment occurred, particularly with posterior injury. Microcephaly developed in 45%. Interpretation: Severe neonatal anaemia at birth was associated with a white matter predominant pattern of injury, the severity of which was related to neurodevelopmental outcomes. Early MRI and long-term follow-up are advisable following severe neonatal anaemia. © 2017 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Severe anaemia in the immediate neonatal period is a rare but potentially serious condition. Blood loss before, during or after delivery accounts for 5–10% of all severe neonatal anaemia [1]. The more common aetiologies of fetal blood loss are feto-maternal haemorrhage (FMH), antepartum haemorrhage (APH) and twin-to-twin transfusion syndrome (TTTS) [2,3]; less common are haemolytic anaemias and infection such as parvovirus. Some FMH occurs during all pregnancies, but usually in small, clinically insignificant quantities [4]. Severe FMH is a well-established entity with a wide range of reported incidence [5]. Antenatally there is usually no obvious precipitating factor for the FMH and the presenting signs are ⁎ Corresponding author at: Hospital Universitario Cruces, Departamento de Pediatría, Unidad Neonatal, Plaza de Cruces s/n, 48903 Barakaldo, Vizcaya, Spain. E-mail address: [email protected] (B. Loureiro).
http://dx.doi.org/10.1016/j.earlhumdev.2017.01.001 0378-3782/© 2017 Elsevier Ireland Ltd. All rights reserved.
frequently non-specific [6]. Consequently, diagnosis requires a high index of suspicion. Additionally, diagnostic techniques are imperfect. The Kleihauer-Betke screen has been the standard for diagnosis and quantification of FMH, but has several limitations [4]. There is currently no reliable method of estimating the severity of FMH, the rate of fetal blood loss and whether the haemorrhage is acute or chronic. APH is defined as any vaginal bleeding from 20 weeks gestation and remains an important cause of maternal and fetal morbidity and mortality. The most common causes are placenta praevia and placental abruption, accounting for more than half of cases [7]. In contrast, TTTS and other aetiologies of fetal anaemia (e.g. decreased haemoglobin (Hb) production or increased destruction from causes such as Rhesus alloimmunisation and parvovirus infection), can be detected and treated antenatally, reducing the chances of severe anaemia at birth and increasing the overall perinatal survival rate [8,9]. There is accumulating evidence of long-term neurodevelopmental consequences following aetiologies of fetal anaemia such as Rhesus
18
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
alloimmunisation, parvovirus infection or TTTS [2,3]. In contrast, data regarding brain injury and its consequences in the context of FMH or APH are few and based on small studies with relatively limited neuroimaging and outcome data [6,8,9]. A better understanding of MRI patterns following severe neonatal anaemia would be useful not only for clarifying the nature and degree of injury, but also in predicting neurodevelopmental outcomes [10]. The aims of this study were: 1) to describe patterns of brain injury seen on early MRI in infants with severe anaemia due to blood loss not suspected before delivery; and 2) to explore how MRI findings related to clinical factors and later outcomes.
diffusion coefficient maps were analysed visually. An experienced neuroradiologist assessed all images at presentation and a later review by an experienced perinatal neurologist (FMC) confirmed the findings. All images were assessed for anatomy, structural development, evidence of chronic injury (e.g. white matter atrophy or established cysts during the first week), intraparenchymal or extracerebral haemorrhage, and abnormal signal intensity (SI) in the basal ganglia-thalami (BGT), posterior limb of the internal capsule (PLIC), white matter, cortex and brainstem.
2. Patients and methods
All children were offered neurodevelopmental follow-up at 24 months. Children seen by us were assessed using a standardized neurological assessment [16] and the Griffiths' Mental Developmental Scales [17], from which a developmental quotient (DQ) was calculated. For other children information was obtained from local paediatric and developmental services. Cerebral palsy was defined and classified according to published criteria [18,19]. Head growth and occurrence of seizures or epilepsy were noted.
We included singleton infants born or referred to the Hammersmith and Queen Charlotte's Hospitals between 1998 and 2011 for specialist opinion, who had the following: 1) severe anaemia at birth defined as a Hb level b 7 g/dL or clinically apparent severe anaemia requiring immediate transfusion [11]; and 2) brain MRI within 8 weeks of delivery. Where infants required a transfusion before the first Hb could be measured, we estimated their pre-transfusion Hb level from their posttransfusion level based on a formula predicting that a transfusion of 10 mL/kg increases the Hb concentration by approximately 2 g/dL [12]. These estimated values were used as the initial Hb concentrations. Exclusion criteria were: metabolic disorders, congenital malformations or infections, genetic abnormality or a known Hb abnormality. We also excluded infants who had received any fetal intervention, including fetal transfusion or treatment for TTTS. Data were extracted from the patients' medical notes. Perinatal data included maternal age, obstetric and family history, bleeding in any trimester, reduced fetal movements (RFMs), onset and progress of labour, cardiotocographic (CTG) abnormalities, acute intrapartum events, mode of delivery, gestational age (GA), sex, birth weight and head circumference (centiles) [13], Apgar scores, cord pH, resuscitation measures, Kleihauer test results and Hb at birth or first available Hb. Resuscitation was scored on a 0–5 scale (0 no intervention, 1 oropharyngeal suction and stimulation, 2 face-mask oxygen, 3 positive-pressure mask ventilation, 4 intubation, 5 cardiac compressions, drugs or volume expansion). Postnatal data included the clinical course (encephalopathy or seizures) and any co-morbidities, particularly hypoglycaemia. Neonatal encephalopathy (NE) was defined as difficulty initiating and maintaining respiration, altered consciousness, or abnormal tone and reflexes, with or without seizures. Isolated seizures without other neurological symptoms were not considered NE [14]. Hypoglycaemia was defined as ≥ 1 documented episodes of blood/ plasma glucose concentration ≤ 2.5 mmol/L (45 mg/dL) [15]. It was classified as prolonged/recurrent when there was a poor response to the first intervention and as severe if the lowest documented glucose level was ≤1.5 mmol/L. 2.1. Ethical approval Ethical permission for scanning the infants was obtained from the Hammersmith Hospital research ethics committee and individually from the parents of each infant. 2.2. MRI analysis All infants had a brain MRI scan as part of their clinical care. Images were obtained using a 1.0-, 1.5-, or 3-T Philips system (Philips Medical System, Best, Netherlands), depending on the system in use at the time of presentation. Minimal image acquisition included T1-weighted transverse and sagittal images and T2-weighted transverse images. Diffusion-weighted imaging (DWI) was not always performed in infants scanned more than two weeks after birth; when available apparent
2.3. Outcome
3. Results Twenty infants were studied, five inborn and 15 out-born referred for specialist assessment. Seventeen had a Hb b 7 g/dL; the other three had clinical evidence at birth of severe anaemia requiring transfusion before the Hb could be measured and their pre-Hb concentrations were estimated as b 7 g/dL [12]. Based on clinical data, maternal Kleihauer and absence of any other explanation, the anaemia was considered due to FMH in 17 infants and to APH in three. Table 1 gives details of the infants, their perinatal and postnatal events. Mean gestational age was 37 ± 2.5 weeks (range 33–41). Only three infants were b9th centile for weight and none had a head circumference ≤ 9th centile. Six infants were preterm (gestation 33–36 weeks). 3.1. Pre-labour, labour and delivery-related data Two mothers had prior medical problems: one had Crohn's disease and a history of infertility; conception in this pregnancy was by invitro fertilisation. Another mother had rheumatic mitral and aortic valve disease and was on warfarin after suffering a middle cerebral artery infarction at 23 weeks of pregnancy. No other maternal illnesses, family history of seizures or other neurological disorders were identified and no mother had evidence of parvovirus infection. Seventeen infants were born by pre-labour emergency caesarean section (CS), 11 because of RFMs and an abnormal CTG and six because of an abnormal CTG alone. In the 11 mothers who reported RFMs, CTG performed on admission either showed a flat trace (1), a sinusoidal pattern (1), bradycardia (2), decreased variability (1), late decelerations (3) or other non-specific abnormalities (3). Of the six mothers with normal FMs, two attended the hospital reporting early signs of labour but were found to have reduced CTG variability without being in established labour. One mother had abdominal pain lasting 24 h and a sinusoidal CTG. Another mother had vaginal bleeding from placenta praevia and a bradycardic CTG. Two mothers had an abnormal CTG at a routine 33-week gestation antenatal visit (sinusoidal pattern with poor variability in one and poor variability with decelerations in a growth-restricted fetus in the other). Labour started spontaneously in only two mothers, who both reported normal FMs and had initially normal CTGs; one delivered normally and the other delivered vaginally with forceps for later CTG concerns. One infant was born by elective CS, because of a previous CS. Nine infants (45%) had a 5-minute Apgar score ≤ 5. Only three infants had a pH b 7.0, though eight infants had a base deficit ≥ -13. Ten infants (50%) required major resuscitation, though none required drug
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
19
Table 1 Infants' details, perinatal and postnatal events, MRI findings and outcome. Case/sex Aetiology GA (w)
BW (kg)
Apgar 1, 5, 10 min
Resus Cord pH (art)
Base 1st Hb deficit (g/dL)
Lowest glucose level
NE
1,1,3 1,1,3
5 5
7.2 7.2
NA 1
2 3.2
1.5 5.4
Yes Yes 4 days 10 days Yes No (isoelectric EEG) Yes Yes 5 weeks (+8 months)
Seizures
Age at MRI
1F 2F
APH & PA 39 3.5 34+1 2.69 FMHa
3F
FMH
41+3 3.46
6,6
4
7
−14
4.5
NA
4F
FMH
38
2.7
2,3,3
5
6.8
−20
2.7
0.7
Yes Yes
22 days
5M 6F
FMH FMHa
38 34
3.04 2.35
9,10 4,4,6
0 4
7.3 NA
−4 NA
4 2
5.5 NA
No Yes Yes Yes
6 days 8 weeks (+13 months)
7F
APH & PA 41
3.4
5,6,7
2
7.3
NA
4c
0.6
No
4 days & 8 days 6 weeks (+13 months)
8F
FMH
38+3 3.05
2,5
4
7.2
NA
3
4.7
Yes Yes
4 weeks
9F
FMH
Term 2.77
3,8,8
4
7
−13
3.8
0.9
Yes Yes
8 weeks
10 M
FMH
39
3,5,5
3
7
−14
2.3
1,1
Yes Yes
8 days 7 weeks
11 F
FMH
39+5 2.9
3,3,3
4
7
−7
4.4
0.6
Yes Yes
5 days, 3 weeks
12 F
FMH
38+3 3.08
5,8,10
3
6.9
−17
4.2
0.9
No
13 M
FMH
37+1 2.8
6,8,8
5
7.2
−8
3.7c
NA
Yes Yes
14 F
FMH
3.55
Yes
Yes
37
+1
2.94
3,6,9
3
7.1
−12
3.9
5.1
No
Yes
3 days &13 days 8 weeks (+18 months) 5 days &16 days (+5 months) 3 weeks
+1
1.7
7,9
3
7.3
NA
4.6
5.4
No
No
7 weeks
15 F
FMH
33
16 M 17 M 18 M 19 M
FMH FMH FMH APHb
34+6 40+2 33+2 36+6
2.29 2.76 2.03 3.25
2,3,5 3,7 NA 2,2,2
3 3 2 5
7.2 7 7 6.8
−5 −16 −15 −14
3.5 5.5 3 4.6c
4.6 2.1 NA 1.7
No No No Yes
No Yes No Yes
5 8 8 2
20 M
FMH
39+6 2.64
8,8,9
0
7.1
−10
3.2
2.1
No
Yes
8 weeks
weeks weeks weeks weeks
Pattern of MRI injury
Outcome
Severe diffuse Severe diffuse
Death at 7 days Quadriplegia GMCSF IV at 2 years Microcephaly Severe diffuse Quadriplegia GMFCS V at 2 years Visual impairment, Seizures, microcephaly Severe diffuse Quadriplegia GMFCS IV at 2 years Microcephaly, seizures Severe diffuse Death at 11 days Focal/multifocal Quadriplegia GMFCS cystic IV at 2 years Microcephaly Focal/multifocal Mild right hemiplegia cystic DQ 54 at 3 years Visual impairment, squint Microcephaly Focal/multifocal Normal at 5 years cystic Focal/multifocal Learning difficulties at cystic 5 years Microcephaly, seizures Focal/multifocal DQ 83 at 2 years cystic Visual impairment, microcephaly Focal/multifocal DQ 69 at 2 years cystic Microcephaly Focal/multifocal Behaviour problems cystic (DQ 114 at 2 years) Microcephaly Parasagittal Mild right hemiplegia non-cystic DQ 119 at 4 years Parasagittal Language delay at non-cystic 2 years Parasagittal Normal at 22 months non-cystic Mild/minimal Normal at 4 years Mild/minimal Normal at 5 years Mild/minimal Normal at 2 years Mild/minimal DQ 83 at 2.3 years Language delay (sub-quotient 67) Mild/minimal Normal at 2 years
Abbreviations: APH: antepartum haemorrhage; Art: arterial; BW: birth weight; DQ: developmental quotient; F: female; FMH: fetomaternal haemorrhage; GA: gestational age; GMFCS: Gross Motor Function Classification System; Hb: haemoglobin; M: male; NE: neonatal encephalopathy; NA: not available; PA: Placental abruption; PCV: packed cell volume; Post: after transfusion; Resus: resuscitation (0 no intervention, 1 oropharyngeal suction and stimulation, 2 face-mask oxygen, 3 positive-pressure ventilation by mask, 4 intubation, 5 cardiac compressions, drugs or volume expansion). a FMH was diagnosed presumptively in the absence of another explanation, but a Kleihauer test was not available. b APH from placenta praevia. c Estimated haemoglobin [12].
administration. Six infants needed fluids (blood or saline) in the labour ward. The Hb at birth was b7 g/dL in 17 infants, measured either from cord blood or early postnatal blood samples (mean 3.9 ± 2, range 2–5.5 g/dL). The other three infants received emergency transfusion in the labour ward because of severe pallor before the Hb could be measured; their mean post-transfusion Hb was 10.7 ± 2.6 g/dL (range 7.7–12.6) and their mean estimated pre-transfusion Hb was 4.1 ± 0.4 g/dL (range 3.7–4.6) [12]. All 20 infants needed early blood transfusion; the mean number of transfusions per infant was 2.5 (range 1–5). The Kleihauer test was positive in 15 of the 18 mothers tested (estimated fetal blood loss was 91 ± 41mls, range 40–142). In two infants FMH was diagnosed presumptively in view of the absence of other
explanations for the anaemia, but a Kleihauer test was not available. Three infants had APH (placenta praevia in one and placental abruption in two). 3.2. Clinical course Sixteen of the 20 infants were admitted directly to the neonatal intensive care unit (NICU). In three this was mainly because of prematurity of 33–34 weeks (infants 15, 16, 18); none showed signs of NE, developed seizures or became hypoglycaemic. Of the other 13 infants admitted directly to the NICU (three preterm of 34–36 weeks; 10 term), 11 required continuing respiratory support and two had marked pallor requiring urgent transfusion (infants 10,
20
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
14). Eleven had symptoms of moderate-severe NE, and seven also became hypoglycaemic. Two infants (infants 10, 14) were initially thought less severe neurologically though they had abnormal tone and irritability, but deteriorated in the first postnatal hours developing clinical seizures and a marked lactic acidosis. Of the five infants born after hypothermia was introduced as standard therapy for NE, two (infants 11, 19) fulfilled the criteria for moderate-severe hypoxic-ischaemic encephalopathy (HIE) and were cooled. The remaining four infants initially went to the postnatal ward (infants 5, 7, 17, 20), but deteriorated during the first few hours with a progressive lactic acidosis; all developed clinical seizures and three became hypoglycaemic. One infant (infant 5) had good Apgar scores and cord gases but was admitted to the NICU at 2 h with a severe lactic acidosis; he had progressive multiorgan failure and died. Ten infants became hypoglycaemic; six developed severe prolonged hypoglycaemia with a mean lowest glucose of 0.7 mmol/L (range 0.6– 0.9), two had transient hypoglycaemia of 1.5–1.7 mmol/L, associated with NE and another two infants on the 3rd centile for weight had transient glucose levels of 2.1 mmol/L.
3.3. MRI findings The median age at first scan was 21 days; this scan was used to evaluate the injury pattern. Seven infants also had later scans (Table 1). No infant was found to have a developmental brain abnormality and none had evidence of injury that clearly predated the time around birth. The predominant site of injury was the white matter. The most severely affected infants had additional cortical injury adjacent to white matter injury (n = 13, 65%). Only one infant had a pattern of injury
typical of a severe acute hypoxic-ischaemic episode with abnormality on conventional and DWI in the central grey and white matter, cortex and hippocampi. Three other infants also had BGT injury but as part of a devastating global insult involving all tissues. Four main patterns of injury were seen, reflecting the severity of white matter injury. 1. Severe-diffuse injury with generalized swelling or markedly abnormal signal intensity (SI) in the white matter or severe and extensive multicystic encephaloleukomalacia, depending on the postnatal timing of the scan (infants 1–5, Fig. 1a–b). Infants 1, 2, 4 and 5 showed additional BGT injury, which was especially marked in infants 1 and 5 with swelling and abnormal SI also in the brainstem, as well as multiple haemorrhagic foci in the white matter. In infant 3 the MRI at 5 weeks showed atrophy of the BGT, PLIC and brainstem secondary to the white matter injury. Infants 2 and 3 had also extensive subdural haemorrhage. 2. Focal/multifocal cystic injury with localized abnormal white matter SI and cystic evolution (infants 6–12, Fig. 1c). The injury was predominantly parasagittal and posterior in distribution and of varying severity (uni- or bilateral). Fig. 2 illustrates the evolution of this pattern over time in one infant. Infants 6 and 7 also showed some abnormalities in the BGT and PLIC on later scans secondary to the white matter injury. 3. Parasagittal non-cystic White matter injury: marked predominantly posterior abnormal SI, (infants 13–15, Fig. 1d–e). Infant 13 showed reduced thalamic volume on later scans, but normal appearing PLICs. 4. Mild/minimal injury defined as mild increased white matter SI on T2-W images usually in a periventricular or diffuse distribution (infants 16–20, Fig. 1f).
Fig. 1. a: Severe diffuse white matter injury with additional acute basal ganglia (BG) injury (T1-weighted MRI scan in the transverse plane, obtained 4 days after birth). This shows loss of grey matter/white matter differentiation and global swelling. Foci of high signal intensity (SI) are seen in the BG. Multiple foci of haemorrhage are present. This infant (case 1) died; b: Severe diffuse injury, appearing as bilateral cystic encephaloleukomalacia (T2-weighted MRI scan in the transverse plane, obtained 5 weeks after birth). Widespread infarction is seen with cystic breakdown throughout both hemispheres. BGT are preserved in size, but with no myelination in the posterior limb of the internal capsule (PLIC). This infant (case 3) developed a severe spastic quadriplegia, visual impairment and seizures; c: Focal cystic pattern of injury, predominantly posterior. This infant had parasagittal infarction (T2-weighted MRI scan in the transverse plane, obtained 8 weeks after birth). Abnormal SI is present within the posterior white matter, which has evolved into cysts. This infant (case 12) had a normal DQ at 5 years, but developed microcephaly and behavioural problems; d & e: Parasagittal non-cystic infarction (T1-weighted MRI scan in the transverse plane and T2-weighted MRI scan in the sagittal plane, obtained 3 weeks after birth). Abnormal SI in the white matter and cortex, predominantly in posterior parietal lobes. This infant (case 14) had language delay at 2 years; f: Minimal/Mild changes (T2-weighted MRI scan in the transverse plane, obtained 5 weeks after birth). Mild increase in SI in the periventricular white matter. Prominence of extracerebral space anteriorly. This infant (case 16) was normal at 4 years.
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
21
Fig. 2. Parasagittal focal cystic infarction, evolving over time; a: Diffusion weighted images at 3 days; b: T2-weighted MRI scan in the transverse plane obtained at 8 days; c & d: T2weighted MRI scan in the transverse plane and T1-weighted MRI scan in the sagittal plane at 1.5 months. Initially widespread restricted apparent diffusion coefficient (ADC) in the posterior cortex and white matter. The 8 day scan shows very abnormal SI mainly in the posterior white matter, cortex and pulvinar of the thalami. Later scans show evolution to cysts and atrophy and asymmetry of the PLIC and thalami being less developed and smaller on the left. This infant (7) developed a mild right hemiplegia, visual impairment with squint, microcephaly and global developmental delay at 3 years.
3.4. Outcomes Infants 1 and 5 died. Outcomes in survivors were assessed at a median age of 24 months (range 22 months–5 years). Follow-up information from our centre was available for 12 children and from local services for 6 children. Six children had motor impairment (severe spastic bilateral cerebral palsy in four and mild unilateral cerebral palsy in two). Suboptimal head growth (decrease N2SD across centiles, compared with birth HC) was present in nine children (45%). Three children developed later seizures and three were visually impaired. Four children without motor impairment had a DQ ≤ 87 and four showed specific language, learning or behavioural difficulties. Six children had normal outcomes at 2–5 years (Table 1).
3.5. MRI findings and outcomes Outcomes correlated well with MRI findings (Fig. 3). The two infants who died had severe diffuse global injury. Surviving infants with significant motor impairment either had severe bilateral cystic leukoencephalomalacia also with abnormalities in the BGT and PLIC (spastic quadriplegia, infants 2–4) or bilateral parasagittal white matter infarction with secondary changes in the BGT and PLIC on later scans, (spastic quadriplegia, infant 6, and mild unilateral cerebral palsy, infants 7, 13). Global developmental delay, learning or behavioural problems
and microcephaly were frequently found in those with isolated white matter injury. Visual impairment was associated with a posterior bilateral parasagittal injury pattern. Those with mildly abnormal white matter findings had a normal outcome except for one child with mild developmental delay. 4. Discussion In this cohort of infants with severe neonatal anaemia the predominant site of injury was in the white matter, with a wide range of severity; to some extent white matter injury was seen in all cases. Although the most severely affected infants also had injury throughout the cortex, BGT and brainstem, this was never present without white matter involvement. Neurodevelopmental consequences were largely of a severity and type that would be expected from the degree and location of injury on MRI. A recent Dutch study of infants with severe neonatal anaemia also found a predominant pattern of white matter injury in the majority of infants, although in two cases MRI showed isolated BGT injury. In that study, the aetiology of the anaemia included haemolysis in 6% of infants and TTTS in 8%, and other cases in which anaemia could have started in the antepartum period [10]. Patterns of brain injury after TTTS or from pathologies beginning some time antenatally, with a pathophysiology not exclusively related to anaemia, would not necessarily be generalizable to neonatal anaemia following FMH and APH [20].
Fig. 3. Neurodevelopmental outcomes in relation to MRI patterns.
22
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
Significant fetal anaemia may result in poor oxygen delivery, which may lead to compensatory cardiovascular responses to maintain cerebral perfusion. Thus the fetus may compensate in utero for the hypoxia, unless the blood loss is very sudden [4]. The predominant pattern of white matter injury we found following neonatal anaemia may also be seen following “prolonged partial asphyxia”, manifesting as mild neurological symptoms at birth or with a delayed onset, not always meeting full criteria for acute HIE [21]. This pattern of injury is also seen in children with congenital heart disease [22], infection, metabolic disorders or hypoglycaemia, all having a more protracted course than an acute severe asphyctic event [23]. Most infants in this cohort also had symptoms of NE and/or hypoglycaemia. Anaemic infants subjected to chronic/subacute hypoxia may be vulnerable to additional acute injury from “near total asphyxia”, showing a BGT pattern of injury, as was the case in 4 infants. The depletion of glycogen stores secondary to stress in anaemic fetuses may explain the hypoglycaemia [24]. Although the reason is poorly understood, the white matter is the commonest site of injury in neonatal hypoglycaemia, predominantly but not always confined to posterior regions [25]. The severity of brain injury varied significantly, with outcomes similarly ranging from severely affected to normal. Interestingly, we did not find a correlation between the initial Hb level and the severity of brain injury, as also reported by Zonnenberg et al. [10]. We nevertheless found neurodevelopmental outcomes correlated well with MRI findings, which suggests that early MRI can be used in neonatal anaemia due to FMH and APH for parental counselling and planning follow-up. We found that suboptimal head growth and learning or behavioural problems were common after two years. In contrast, Zonnenberg et al. reported that most children were normal at a median age of 19 months, with no differences between children with mild and those with moderate-severe white matter injury. We note however that in their study only 13 children had both neonatal MRI and outcome data, which was obtained at a younger age than in our study. The longer follow-up in our study may explain the difference in findings. For this reason we suggest that long-term follow-up is advisable, even if a child appears to be developing normally in early infancy. We would also suggest an MRI for all children with severe anaemia at birth and that it should definitely be done whenever there is an abnormal neurological examination, encephalopathy, seizures or any co-morbidities, particularly hypoglycaemia. Many of the infants in this study were referred to us for neurological assessment often because of concern about brain injury, and thus the incidence and severity of brain injury in this reported cohort may not apply to all neonates even with severe neonatal anaemia. Despite this selection bias, the strength of our study is that all infants had good quality early MRI, thorough contemporaneous documentation of neurological comorbidities, and a high rate of follow-up. Another limitation of this study is that the total number of infants limits statistical power to analyse the relationship between clinical factors and MRI findings or outcome. Previous works have explored the relationship between reduced FMs, CTG abnormalities, Kleihauer test and first Hb with clinical variables and outcomes [4,6,10], but further studies are needed to examine the significance of these factors in relation to MRI findings. In the absence of such data, and based on our experience and that of others on the lack of correlation between Hb level and brain injury and outcome [10], we suggest that following severe neonatal anaemia, MRI is the best method to determine the presence and extent of injury. This is particularly so given that injury largely affects the white matter and is often worse posteriorly, making it less accessible to imaging using cranial ultrasound. Hypoglycaemia was a common co-morbidity and should routinely be looked for with severe neonatal anaemia. In conclusion, we found that severely anaemic neonates have a high susceptibility to white matter injury, with additional areas of injury present in severer cases. Neurodevelopmental outcomes correlated well with early MRI findings, and reflected known associates of white matter injury [26]. MRI in infants with severe neonatal anaemia is
particularly helpful in counselling parents, and in planning long-term neurodevelopmental follow-up. Conflict of interest None of the authors declares a conflict of interest. Acknowledgements The authors would like to thank Dr. Andrew Chew for his assistance in evaluating children at follow-up, the staff of the Queen Charlotte's neonatal unit, the Robert Steiner MR unit, and the Hammersmith children's outpatients department. We thank the families of the infants and children in this study. Dr. Begoña Loureiro was funded by the Health Department of Basque Government (grant number: 2011011015). References [1] S. Aher, K. Malwatkar, S. Kadam, Neonatal anemia, Semin. Fetal Neonatal Med. 13 (2008) 239–247. [2] J.M.M. van Klink, H.M. Koopman, D. Oepkes, F.J. Walther, E. Lopriore, Long-term neurodevelopmental outcome after intrauterine transfusion for fetal anemia, Early Hum. Dev. 87 (2011) 589–593. [3] E. Lopriore, D. Oepkes, F.J. Walther, Neonatal morbidity in twin-twin transfusion syndrome, Early Hum. Dev. 87 (2011) 595–599. [4] B.J. Wylie, M.E. D'Alton, Fetomaternal hemorrhage, Obstet. Gynecol. 115 (2010) 1039–1051. [5] A. Stroustrup, L. Trasande, Demographics, clinical characteristics and outcomes of neonates diagnosed with fetomaternal haemorrhage, Arch. Dis. Child. Fetal Neonatal Ed. 97 (2012) F405–F410. [6] G.P. Giacoia, Severe fetomaternal hemorrhage: a review, Obstet. Gynecol. Surv. 52 (1997) 372–380. [7] P. Sinha, N. Kuruba, Ante-partum haemorrhage; an update, J. Obstet. Gynaecol. 28 (2008) 377–381. [8] C. Rubob, P. Deruelle, F. Le Goueff, et al., Long-term prognosis for infants after massive fetomaternal hemorrhage, Obstet. Gynecol. 110 (2007) 256–260. [9] Z. Kecskes, Large fetomaternal hemorrhage: clinical presentation and outcome, J. Matern. Fetal Neonatal Med. 13 (2003) 128–132. [10] I.A. Zonnenberg, R.J. Vermeulen, M.W. Rohaan, et al., Severe neonatal anaemia, MRI findings and neurodevelopmental outcome, Neonatology 109 (2016) 282–288. [11] G. Mari, For the collaborative group for Doppler assessment of the blood velocity in anemic fetuses. Noninvasive diagnosis by Doppler ultrasonography of fetal anemia due to maternal red-cell alloimmunization, N. Engl. J. Med. 342 (2000) 9–14. [12] P. Davies, S. Robertson, S. Hegde, R. Greenwood, E. Massey, P. Davis, Calculating the required transfusion volume in children, Transfusion 47 (2007) 212–216. [13] T.J. Cole, J.V. Freeman, M.A. Preece, British 1990 growth reference centiles for weight, height, body mass index and head circumference fitted by maximum penalized likelihood, Stat. Med. 17 (1998) 407–429. [14] F. Cowan, M. Rutherford, F. Groenendaal, et al., Origin and timing of brain lesions in term infants with neonatal encephalopathy, Lancet 361 (2003) 736–742. [15] M. Cornblath, J.M. Hawdon, A.F. Williams, et al., Controversies regarding definition of neonatal hypoglycemia: suggested operational thresholds, Pediatrics 105 (2000) 1141–1145. [16] L. Haataja, E. Mercuri, R. Regev, et al., Optimality score for the neurological examination of the infant at 12 and 18 months of age, J. Pediatr. 135 (1999) 153–161. [17] R. Griffiths, The Abilities of Young Children, Child Development Research Centre, London, United Kingdom, 1970. [18] C. Cans, Surveillance of cerebral palsy in Europe: a collaboration of cerebral palsy surveys and registers, Dev. Med. Child Neurol. 42 (2000) 816–824. [19] R. Palisano, P. Rosenbaam, S. Walter, et al., Development and reliability of a system to classify gross motor function in children with cerebral palsy, Dev. Med. Child Neurol. 39 (1997) 214–223. [20] E. Quarello, M. Molho, Y. Ville, Incidence, mechanism, and patterns of fetal cerebral lesions in twin-to-twin transfusion syndrome, J. Matern. Fetal Neonatal Med. 20 (2007) 589–597. [21] M.J. Barrett, V. Donogue, E.E. Mooney, et al., Isolated acute non-cystic white matter injury in term infants presenting with neonatal encephalopathy, Arch. Dis. Child. Fetal Neonatal Ed. 98 (2013) F158-F60. [22] S.P. Miller, P.S. McQuillen, S. Hamrick, et al., Abnormal brain development in newborns with congenital heart disease, N. Engl. J. Med. 357 (2007) 1928–1938. [23] L.S. de Vries, F. Groenendaal, Patterns of neonatal hypoxic-ischaemic brain injury, Neuroradiology 52 (2010) 555–566. [24] R.C. Vannucci, S.J. Vannucci, Hypoglycemic brain injury, Semin. Neonatol. 6 (2001) 147–155. [25] C. Burns, M.A. Rutherford, J.P. Boardman, F.M. Cowan, Patterns of cerebral injury and neurodevelopmental outcomes after symptomatic hypoglycemia, Pediatrics 122 (2008) 65–74. [26] M. Martinez-Biarge, T. Bregant, C.J. Wusthoff, et al., White matter and cortical injury in hypoxic-ischemic encephalopathy: antecedent factors and 2-year outcome, J. Pediatr. 161 (2012) 799–807.
Contents lists available at ScienceDirect
Early Human Development journal homepage: www.elsevier.com/locate/earlhumdev
MRI Patterns of brain injury and neurodevelopmental outcomes in neonates with severe anaemia at birth Begoña Loureiro a,⁎, Miriam Martinez-Biarge b, Francesca Foti c, Maria Papadaki d, Frances M Cowan b, Courtney J Wusthoff e a
Departamento de Pediatría, Hospital Universitario Cruces, Barakaldo, Spain Department of Paediatrics, Hammersmith Hospital, Imperial College, London, UK Department of Mental Health, Azienda Unita Sanitaria Locale di Reggio Emilia, Emilia-Romagna, Italy d Own private paediatrician practice, Cholargos Attica, Greece e Division of Child Neurology, Stanford University, CA, USA b c
a r t i c l e
i n f o
Article history: Received 22 September 2016 Received in revised form 31 December 2016 Accepted 4 January 2017 Keywords: Neonatal anaemia Feto-maternal haemorrhage Neonatal encephalopathy White matter injury Magnetic resonance imaging Developmental outcomes
a b s t r a c t Aims: To define patterns of brain injury and associated neurodevelopmental outcomes in infants with severe neonatal anaemia. Methods: We studied 20 infants with severe anaemia at birth (haemoglobin b 7 g/dL). Clinical details were analysed for causes of anaemia and co-morbidities. All had early brain magnetic resonance imaging (MRI) scans, which were reviewed for injury pattern. Neurodevelopmental outcomes were assessed at a median age of 24 months. Results: The aetiology of the anaemia was feto-maternal haemorrhage in 17 and antepartum haemorrhage in 3 infants. The predominant site of injury was the white matter, which was affected in all infants, with differing grades of severity and with cystic evolution in 45%. Only one infant showed an injury pattern typical of an acute severe hypoxic-ischaemic insult. Outcomes correlated closely to the severity of MRI findings. Cerebral palsy was seen only with the most severe neuroimaging patterns (n = 6). Global developmental delay, learning or behavioural problems and seizures were common with moderate injury. Visual impairment occurred, particularly with posterior injury. Microcephaly developed in 45%. Interpretation: Severe neonatal anaemia at birth was associated with a white matter predominant pattern of injury, the severity of which was related to neurodevelopmental outcomes. Early MRI and long-term follow-up are advisable following severe neonatal anaemia. © 2017 Elsevier Ireland Ltd. All rights reserved.
1. Introduction Severe anaemia in the immediate neonatal period is a rare but potentially serious condition. Blood loss before, during or after delivery accounts for 5–10% of all severe neonatal anaemia [1]. The more common aetiologies of fetal blood loss are feto-maternal haemorrhage (FMH), antepartum haemorrhage (APH) and twin-to-twin transfusion syndrome (TTTS) [2,3]; less common are haemolytic anaemias and infection such as parvovirus. Some FMH occurs during all pregnancies, but usually in small, clinically insignificant quantities [4]. Severe FMH is a well-established entity with a wide range of reported incidence [5]. Antenatally there is usually no obvious precipitating factor for the FMH and the presenting signs are ⁎ Corresponding author at: Hospital Universitario Cruces, Departamento de Pediatría, Unidad Neonatal, Plaza de Cruces s/n, 48903 Barakaldo, Vizcaya, Spain. E-mail address: [email protected] (B. Loureiro).
http://dx.doi.org/10.1016/j.earlhumdev.2017.01.001 0378-3782/© 2017 Elsevier Ireland Ltd. All rights reserved.
frequently non-specific [6]. Consequently, diagnosis requires a high index of suspicion. Additionally, diagnostic techniques are imperfect. The Kleihauer-Betke screen has been the standard for diagnosis and quantification of FMH, but has several limitations [4]. There is currently no reliable method of estimating the severity of FMH, the rate of fetal blood loss and whether the haemorrhage is acute or chronic. APH is defined as any vaginal bleeding from 20 weeks gestation and remains an important cause of maternal and fetal morbidity and mortality. The most common causes are placenta praevia and placental abruption, accounting for more than half of cases [7]. In contrast, TTTS and other aetiologies of fetal anaemia (e.g. decreased haemoglobin (Hb) production or increased destruction from causes such as Rhesus alloimmunisation and parvovirus infection), can be detected and treated antenatally, reducing the chances of severe anaemia at birth and increasing the overall perinatal survival rate [8,9]. There is accumulating evidence of long-term neurodevelopmental consequences following aetiologies of fetal anaemia such as Rhesus
18
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
alloimmunisation, parvovirus infection or TTTS [2,3]. In contrast, data regarding brain injury and its consequences in the context of FMH or APH are few and based on small studies with relatively limited neuroimaging and outcome data [6,8,9]. A better understanding of MRI patterns following severe neonatal anaemia would be useful not only for clarifying the nature and degree of injury, but also in predicting neurodevelopmental outcomes [10]. The aims of this study were: 1) to describe patterns of brain injury seen on early MRI in infants with severe anaemia due to blood loss not suspected before delivery; and 2) to explore how MRI findings related to clinical factors and later outcomes.
diffusion coefficient maps were analysed visually. An experienced neuroradiologist assessed all images at presentation and a later review by an experienced perinatal neurologist (FMC) confirmed the findings. All images were assessed for anatomy, structural development, evidence of chronic injury (e.g. white matter atrophy or established cysts during the first week), intraparenchymal or extracerebral haemorrhage, and abnormal signal intensity (SI) in the basal ganglia-thalami (BGT), posterior limb of the internal capsule (PLIC), white matter, cortex and brainstem.
2. Patients and methods
All children were offered neurodevelopmental follow-up at 24 months. Children seen by us were assessed using a standardized neurological assessment [16] and the Griffiths' Mental Developmental Scales [17], from which a developmental quotient (DQ) was calculated. For other children information was obtained from local paediatric and developmental services. Cerebral palsy was defined and classified according to published criteria [18,19]. Head growth and occurrence of seizures or epilepsy were noted.
We included singleton infants born or referred to the Hammersmith and Queen Charlotte's Hospitals between 1998 and 2011 for specialist opinion, who had the following: 1) severe anaemia at birth defined as a Hb level b 7 g/dL or clinically apparent severe anaemia requiring immediate transfusion [11]; and 2) brain MRI within 8 weeks of delivery. Where infants required a transfusion before the first Hb could be measured, we estimated their pre-transfusion Hb level from their posttransfusion level based on a formula predicting that a transfusion of 10 mL/kg increases the Hb concentration by approximately 2 g/dL [12]. These estimated values were used as the initial Hb concentrations. Exclusion criteria were: metabolic disorders, congenital malformations or infections, genetic abnormality or a known Hb abnormality. We also excluded infants who had received any fetal intervention, including fetal transfusion or treatment for TTTS. Data were extracted from the patients' medical notes. Perinatal data included maternal age, obstetric and family history, bleeding in any trimester, reduced fetal movements (RFMs), onset and progress of labour, cardiotocographic (CTG) abnormalities, acute intrapartum events, mode of delivery, gestational age (GA), sex, birth weight and head circumference (centiles) [13], Apgar scores, cord pH, resuscitation measures, Kleihauer test results and Hb at birth or first available Hb. Resuscitation was scored on a 0–5 scale (0 no intervention, 1 oropharyngeal suction and stimulation, 2 face-mask oxygen, 3 positive-pressure mask ventilation, 4 intubation, 5 cardiac compressions, drugs or volume expansion). Postnatal data included the clinical course (encephalopathy or seizures) and any co-morbidities, particularly hypoglycaemia. Neonatal encephalopathy (NE) was defined as difficulty initiating and maintaining respiration, altered consciousness, or abnormal tone and reflexes, with or without seizures. Isolated seizures without other neurological symptoms were not considered NE [14]. Hypoglycaemia was defined as ≥ 1 documented episodes of blood/ plasma glucose concentration ≤ 2.5 mmol/L (45 mg/dL) [15]. It was classified as prolonged/recurrent when there was a poor response to the first intervention and as severe if the lowest documented glucose level was ≤1.5 mmol/L. 2.1. Ethical approval Ethical permission for scanning the infants was obtained from the Hammersmith Hospital research ethics committee and individually from the parents of each infant. 2.2. MRI analysis All infants had a brain MRI scan as part of their clinical care. Images were obtained using a 1.0-, 1.5-, or 3-T Philips system (Philips Medical System, Best, Netherlands), depending on the system in use at the time of presentation. Minimal image acquisition included T1-weighted transverse and sagittal images and T2-weighted transverse images. Diffusion-weighted imaging (DWI) was not always performed in infants scanned more than two weeks after birth; when available apparent
2.3. Outcome
3. Results Twenty infants were studied, five inborn and 15 out-born referred for specialist assessment. Seventeen had a Hb b 7 g/dL; the other three had clinical evidence at birth of severe anaemia requiring transfusion before the Hb could be measured and their pre-Hb concentrations were estimated as b 7 g/dL [12]. Based on clinical data, maternal Kleihauer and absence of any other explanation, the anaemia was considered due to FMH in 17 infants and to APH in three. Table 1 gives details of the infants, their perinatal and postnatal events. Mean gestational age was 37 ± 2.5 weeks (range 33–41). Only three infants were b9th centile for weight and none had a head circumference ≤ 9th centile. Six infants were preterm (gestation 33–36 weeks). 3.1. Pre-labour, labour and delivery-related data Two mothers had prior medical problems: one had Crohn's disease and a history of infertility; conception in this pregnancy was by invitro fertilisation. Another mother had rheumatic mitral and aortic valve disease and was on warfarin after suffering a middle cerebral artery infarction at 23 weeks of pregnancy. No other maternal illnesses, family history of seizures or other neurological disorders were identified and no mother had evidence of parvovirus infection. Seventeen infants were born by pre-labour emergency caesarean section (CS), 11 because of RFMs and an abnormal CTG and six because of an abnormal CTG alone. In the 11 mothers who reported RFMs, CTG performed on admission either showed a flat trace (1), a sinusoidal pattern (1), bradycardia (2), decreased variability (1), late decelerations (3) or other non-specific abnormalities (3). Of the six mothers with normal FMs, two attended the hospital reporting early signs of labour but were found to have reduced CTG variability without being in established labour. One mother had abdominal pain lasting 24 h and a sinusoidal CTG. Another mother had vaginal bleeding from placenta praevia and a bradycardic CTG. Two mothers had an abnormal CTG at a routine 33-week gestation antenatal visit (sinusoidal pattern with poor variability in one and poor variability with decelerations in a growth-restricted fetus in the other). Labour started spontaneously in only two mothers, who both reported normal FMs and had initially normal CTGs; one delivered normally and the other delivered vaginally with forceps for later CTG concerns. One infant was born by elective CS, because of a previous CS. Nine infants (45%) had a 5-minute Apgar score ≤ 5. Only three infants had a pH b 7.0, though eight infants had a base deficit ≥ -13. Ten infants (50%) required major resuscitation, though none required drug
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
19
Table 1 Infants' details, perinatal and postnatal events, MRI findings and outcome. Case/sex Aetiology GA (w)
BW (kg)
Apgar 1, 5, 10 min
Resus Cord pH (art)
Base 1st Hb deficit (g/dL)
Lowest glucose level
NE
1,1,3 1,1,3
5 5
7.2 7.2
NA 1
2 3.2
1.5 5.4
Yes Yes 4 days 10 days Yes No (isoelectric EEG) Yes Yes 5 weeks (+8 months)
Seizures
Age at MRI
1F 2F
APH & PA 39 3.5 34+1 2.69 FMHa
3F
FMH
41+3 3.46
6,6
4
7
−14
4.5
NA
4F
FMH
38
2.7
2,3,3
5
6.8
−20
2.7
0.7
Yes Yes
22 days
5M 6F
FMH FMHa
38 34
3.04 2.35
9,10 4,4,6
0 4
7.3 NA
−4 NA
4 2
5.5 NA
No Yes Yes Yes
6 days 8 weeks (+13 months)
7F
APH & PA 41
3.4
5,6,7
2
7.3
NA
4c
0.6
No
4 days & 8 days 6 weeks (+13 months)
8F
FMH
38+3 3.05
2,5
4
7.2
NA
3
4.7
Yes Yes
4 weeks
9F
FMH
Term 2.77
3,8,8
4
7
−13
3.8
0.9
Yes Yes
8 weeks
10 M
FMH
39
3,5,5
3
7
−14
2.3
1,1
Yes Yes
8 days 7 weeks
11 F
FMH
39+5 2.9
3,3,3
4
7
−7
4.4
0.6
Yes Yes
5 days, 3 weeks
12 F
FMH
38+3 3.08
5,8,10
3
6.9
−17
4.2
0.9
No
13 M
FMH
37+1 2.8
6,8,8
5
7.2
−8
3.7c
NA
Yes Yes
14 F
FMH
3.55
Yes
Yes
37
+1
2.94
3,6,9
3
7.1
−12
3.9
5.1
No
Yes
3 days &13 days 8 weeks (+18 months) 5 days &16 days (+5 months) 3 weeks
+1
1.7
7,9
3
7.3
NA
4.6
5.4
No
No
7 weeks
15 F
FMH
33
16 M 17 M 18 M 19 M
FMH FMH FMH APHb
34+6 40+2 33+2 36+6
2.29 2.76 2.03 3.25
2,3,5 3,7 NA 2,2,2
3 3 2 5
7.2 7 7 6.8
−5 −16 −15 −14
3.5 5.5 3 4.6c
4.6 2.1 NA 1.7
No No No Yes
No Yes No Yes
5 8 8 2
20 M
FMH
39+6 2.64
8,8,9
0
7.1
−10
3.2
2.1
No
Yes
8 weeks
weeks weeks weeks weeks
Pattern of MRI injury
Outcome
Severe diffuse Severe diffuse
Death at 7 days Quadriplegia GMCSF IV at 2 years Microcephaly Severe diffuse Quadriplegia GMFCS V at 2 years Visual impairment, Seizures, microcephaly Severe diffuse Quadriplegia GMFCS IV at 2 years Microcephaly, seizures Severe diffuse Death at 11 days Focal/multifocal Quadriplegia GMFCS cystic IV at 2 years Microcephaly Focal/multifocal Mild right hemiplegia cystic DQ 54 at 3 years Visual impairment, squint Microcephaly Focal/multifocal Normal at 5 years cystic Focal/multifocal Learning difficulties at cystic 5 years Microcephaly, seizures Focal/multifocal DQ 83 at 2 years cystic Visual impairment, microcephaly Focal/multifocal DQ 69 at 2 years cystic Microcephaly Focal/multifocal Behaviour problems cystic (DQ 114 at 2 years) Microcephaly Parasagittal Mild right hemiplegia non-cystic DQ 119 at 4 years Parasagittal Language delay at non-cystic 2 years Parasagittal Normal at 22 months non-cystic Mild/minimal Normal at 4 years Mild/minimal Normal at 5 years Mild/minimal Normal at 2 years Mild/minimal DQ 83 at 2.3 years Language delay (sub-quotient 67) Mild/minimal Normal at 2 years
Abbreviations: APH: antepartum haemorrhage; Art: arterial; BW: birth weight; DQ: developmental quotient; F: female; FMH: fetomaternal haemorrhage; GA: gestational age; GMFCS: Gross Motor Function Classification System; Hb: haemoglobin; M: male; NE: neonatal encephalopathy; NA: not available; PA: Placental abruption; PCV: packed cell volume; Post: after transfusion; Resus: resuscitation (0 no intervention, 1 oropharyngeal suction and stimulation, 2 face-mask oxygen, 3 positive-pressure ventilation by mask, 4 intubation, 5 cardiac compressions, drugs or volume expansion). a FMH was diagnosed presumptively in the absence of another explanation, but a Kleihauer test was not available. b APH from placenta praevia. c Estimated haemoglobin [12].
administration. Six infants needed fluids (blood or saline) in the labour ward. The Hb at birth was b7 g/dL in 17 infants, measured either from cord blood or early postnatal blood samples (mean 3.9 ± 2, range 2–5.5 g/dL). The other three infants received emergency transfusion in the labour ward because of severe pallor before the Hb could be measured; their mean post-transfusion Hb was 10.7 ± 2.6 g/dL (range 7.7–12.6) and their mean estimated pre-transfusion Hb was 4.1 ± 0.4 g/dL (range 3.7–4.6) [12]. All 20 infants needed early blood transfusion; the mean number of transfusions per infant was 2.5 (range 1–5). The Kleihauer test was positive in 15 of the 18 mothers tested (estimated fetal blood loss was 91 ± 41mls, range 40–142). In two infants FMH was diagnosed presumptively in view of the absence of other
explanations for the anaemia, but a Kleihauer test was not available. Three infants had APH (placenta praevia in one and placental abruption in two). 3.2. Clinical course Sixteen of the 20 infants were admitted directly to the neonatal intensive care unit (NICU). In three this was mainly because of prematurity of 33–34 weeks (infants 15, 16, 18); none showed signs of NE, developed seizures or became hypoglycaemic. Of the other 13 infants admitted directly to the NICU (three preterm of 34–36 weeks; 10 term), 11 required continuing respiratory support and two had marked pallor requiring urgent transfusion (infants 10,
20
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
14). Eleven had symptoms of moderate-severe NE, and seven also became hypoglycaemic. Two infants (infants 10, 14) were initially thought less severe neurologically though they had abnormal tone and irritability, but deteriorated in the first postnatal hours developing clinical seizures and a marked lactic acidosis. Of the five infants born after hypothermia was introduced as standard therapy for NE, two (infants 11, 19) fulfilled the criteria for moderate-severe hypoxic-ischaemic encephalopathy (HIE) and were cooled. The remaining four infants initially went to the postnatal ward (infants 5, 7, 17, 20), but deteriorated during the first few hours with a progressive lactic acidosis; all developed clinical seizures and three became hypoglycaemic. One infant (infant 5) had good Apgar scores and cord gases but was admitted to the NICU at 2 h with a severe lactic acidosis; he had progressive multiorgan failure and died. Ten infants became hypoglycaemic; six developed severe prolonged hypoglycaemia with a mean lowest glucose of 0.7 mmol/L (range 0.6– 0.9), two had transient hypoglycaemia of 1.5–1.7 mmol/L, associated with NE and another two infants on the 3rd centile for weight had transient glucose levels of 2.1 mmol/L.
3.3. MRI findings The median age at first scan was 21 days; this scan was used to evaluate the injury pattern. Seven infants also had later scans (Table 1). No infant was found to have a developmental brain abnormality and none had evidence of injury that clearly predated the time around birth. The predominant site of injury was the white matter. The most severely affected infants had additional cortical injury adjacent to white matter injury (n = 13, 65%). Only one infant had a pattern of injury
typical of a severe acute hypoxic-ischaemic episode with abnormality on conventional and DWI in the central grey and white matter, cortex and hippocampi. Three other infants also had BGT injury but as part of a devastating global insult involving all tissues. Four main patterns of injury were seen, reflecting the severity of white matter injury. 1. Severe-diffuse injury with generalized swelling or markedly abnormal signal intensity (SI) in the white matter or severe and extensive multicystic encephaloleukomalacia, depending on the postnatal timing of the scan (infants 1–5, Fig. 1a–b). Infants 1, 2, 4 and 5 showed additional BGT injury, which was especially marked in infants 1 and 5 with swelling and abnormal SI also in the brainstem, as well as multiple haemorrhagic foci in the white matter. In infant 3 the MRI at 5 weeks showed atrophy of the BGT, PLIC and brainstem secondary to the white matter injury. Infants 2 and 3 had also extensive subdural haemorrhage. 2. Focal/multifocal cystic injury with localized abnormal white matter SI and cystic evolution (infants 6–12, Fig. 1c). The injury was predominantly parasagittal and posterior in distribution and of varying severity (uni- or bilateral). Fig. 2 illustrates the evolution of this pattern over time in one infant. Infants 6 and 7 also showed some abnormalities in the BGT and PLIC on later scans secondary to the white matter injury. 3. Parasagittal non-cystic White matter injury: marked predominantly posterior abnormal SI, (infants 13–15, Fig. 1d–e). Infant 13 showed reduced thalamic volume on later scans, but normal appearing PLICs. 4. Mild/minimal injury defined as mild increased white matter SI on T2-W images usually in a periventricular or diffuse distribution (infants 16–20, Fig. 1f).
Fig. 1. a: Severe diffuse white matter injury with additional acute basal ganglia (BG) injury (T1-weighted MRI scan in the transverse plane, obtained 4 days after birth). This shows loss of grey matter/white matter differentiation and global swelling. Foci of high signal intensity (SI) are seen in the BG. Multiple foci of haemorrhage are present. This infant (case 1) died; b: Severe diffuse injury, appearing as bilateral cystic encephaloleukomalacia (T2-weighted MRI scan in the transverse plane, obtained 5 weeks after birth). Widespread infarction is seen with cystic breakdown throughout both hemispheres. BGT are preserved in size, but with no myelination in the posterior limb of the internal capsule (PLIC). This infant (case 3) developed a severe spastic quadriplegia, visual impairment and seizures; c: Focal cystic pattern of injury, predominantly posterior. This infant had parasagittal infarction (T2-weighted MRI scan in the transverse plane, obtained 8 weeks after birth). Abnormal SI is present within the posterior white matter, which has evolved into cysts. This infant (case 12) had a normal DQ at 5 years, but developed microcephaly and behavioural problems; d & e: Parasagittal non-cystic infarction (T1-weighted MRI scan in the transverse plane and T2-weighted MRI scan in the sagittal plane, obtained 3 weeks after birth). Abnormal SI in the white matter and cortex, predominantly in posterior parietal lobes. This infant (case 14) had language delay at 2 years; f: Minimal/Mild changes (T2-weighted MRI scan in the transverse plane, obtained 5 weeks after birth). Mild increase in SI in the periventricular white matter. Prominence of extracerebral space anteriorly. This infant (case 16) was normal at 4 years.
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
21
Fig. 2. Parasagittal focal cystic infarction, evolving over time; a: Diffusion weighted images at 3 days; b: T2-weighted MRI scan in the transverse plane obtained at 8 days; c & d: T2weighted MRI scan in the transverse plane and T1-weighted MRI scan in the sagittal plane at 1.5 months. Initially widespread restricted apparent diffusion coefficient (ADC) in the posterior cortex and white matter. The 8 day scan shows very abnormal SI mainly in the posterior white matter, cortex and pulvinar of the thalami. Later scans show evolution to cysts and atrophy and asymmetry of the PLIC and thalami being less developed and smaller on the left. This infant (7) developed a mild right hemiplegia, visual impairment with squint, microcephaly and global developmental delay at 3 years.
3.4. Outcomes Infants 1 and 5 died. Outcomes in survivors were assessed at a median age of 24 months (range 22 months–5 years). Follow-up information from our centre was available for 12 children and from local services for 6 children. Six children had motor impairment (severe spastic bilateral cerebral palsy in four and mild unilateral cerebral palsy in two). Suboptimal head growth (decrease N2SD across centiles, compared with birth HC) was present in nine children (45%). Three children developed later seizures and three were visually impaired. Four children without motor impairment had a DQ ≤ 87 and four showed specific language, learning or behavioural difficulties. Six children had normal outcomes at 2–5 years (Table 1).
3.5. MRI findings and outcomes Outcomes correlated well with MRI findings (Fig. 3). The two infants who died had severe diffuse global injury. Surviving infants with significant motor impairment either had severe bilateral cystic leukoencephalomalacia also with abnormalities in the BGT and PLIC (spastic quadriplegia, infants 2–4) or bilateral parasagittal white matter infarction with secondary changes in the BGT and PLIC on later scans, (spastic quadriplegia, infant 6, and mild unilateral cerebral palsy, infants 7, 13). Global developmental delay, learning or behavioural problems
and microcephaly were frequently found in those with isolated white matter injury. Visual impairment was associated with a posterior bilateral parasagittal injury pattern. Those with mildly abnormal white matter findings had a normal outcome except for one child with mild developmental delay. 4. Discussion In this cohort of infants with severe neonatal anaemia the predominant site of injury was in the white matter, with a wide range of severity; to some extent white matter injury was seen in all cases. Although the most severely affected infants also had injury throughout the cortex, BGT and brainstem, this was never present without white matter involvement. Neurodevelopmental consequences were largely of a severity and type that would be expected from the degree and location of injury on MRI. A recent Dutch study of infants with severe neonatal anaemia also found a predominant pattern of white matter injury in the majority of infants, although in two cases MRI showed isolated BGT injury. In that study, the aetiology of the anaemia included haemolysis in 6% of infants and TTTS in 8%, and other cases in which anaemia could have started in the antepartum period [10]. Patterns of brain injury after TTTS or from pathologies beginning some time antenatally, with a pathophysiology not exclusively related to anaemia, would not necessarily be generalizable to neonatal anaemia following FMH and APH [20].
Fig. 3. Neurodevelopmental outcomes in relation to MRI patterns.
22
B. Loureiro et al. / Early Human Development 105 (2017) 17–22
Significant fetal anaemia may result in poor oxygen delivery, which may lead to compensatory cardiovascular responses to maintain cerebral perfusion. Thus the fetus may compensate in utero for the hypoxia, unless the blood loss is very sudden [4]. The predominant pattern of white matter injury we found following neonatal anaemia may also be seen following “prolonged partial asphyxia”, manifesting as mild neurological symptoms at birth or with a delayed onset, not always meeting full criteria for acute HIE [21]. This pattern of injury is also seen in children with congenital heart disease [22], infection, metabolic disorders or hypoglycaemia, all having a more protracted course than an acute severe asphyctic event [23]. Most infants in this cohort also had symptoms of NE and/or hypoglycaemia. Anaemic infants subjected to chronic/subacute hypoxia may be vulnerable to additional acute injury from “near total asphyxia”, showing a BGT pattern of injury, as was the case in 4 infants. The depletion of glycogen stores secondary to stress in anaemic fetuses may explain the hypoglycaemia [24]. Although the reason is poorly understood, the white matter is the commonest site of injury in neonatal hypoglycaemia, predominantly but not always confined to posterior regions [25]. The severity of brain injury varied significantly, with outcomes similarly ranging from severely affected to normal. Interestingly, we did not find a correlation between the initial Hb level and the severity of brain injury, as also reported by Zonnenberg et al. [10]. We nevertheless found neurodevelopmental outcomes correlated well with MRI findings, which suggests that early MRI can be used in neonatal anaemia due to FMH and APH for parental counselling and planning follow-up. We found that suboptimal head growth and learning or behavioural problems were common after two years. In contrast, Zonnenberg et al. reported that most children were normal at a median age of 19 months, with no differences between children with mild and those with moderate-severe white matter injury. We note however that in their study only 13 children had both neonatal MRI and outcome data, which was obtained at a younger age than in our study. The longer follow-up in our study may explain the difference in findings. For this reason we suggest that long-term follow-up is advisable, even if a child appears to be developing normally in early infancy. We would also suggest an MRI for all children with severe anaemia at birth and that it should definitely be done whenever there is an abnormal neurological examination, encephalopathy, seizures or any co-morbidities, particularly hypoglycaemia. Many of the infants in this study were referred to us for neurological assessment often because of concern about brain injury, and thus the incidence and severity of brain injury in this reported cohort may not apply to all neonates even with severe neonatal anaemia. Despite this selection bias, the strength of our study is that all infants had good quality early MRI, thorough contemporaneous documentation of neurological comorbidities, and a high rate of follow-up. Another limitation of this study is that the total number of infants limits statistical power to analyse the relationship between clinical factors and MRI findings or outcome. Previous works have explored the relationship between reduced FMs, CTG abnormalities, Kleihauer test and first Hb with clinical variables and outcomes [4,6,10], but further studies are needed to examine the significance of these factors in relation to MRI findings. In the absence of such data, and based on our experience and that of others on the lack of correlation between Hb level and brain injury and outcome [10], we suggest that following severe neonatal anaemia, MRI is the best method to determine the presence and extent of injury. This is particularly so given that injury largely affects the white matter and is often worse posteriorly, making it less accessible to imaging using cranial ultrasound. Hypoglycaemia was a common co-morbidity and should routinely be looked for with severe neonatal anaemia. In conclusion, we found that severely anaemic neonates have a high susceptibility to white matter injury, with additional areas of injury present in severer cases. Neurodevelopmental outcomes correlated well with early MRI findings, and reflected known associates of white matter injury [26]. MRI in infants with severe neonatal anaemia is
particularly helpful in counselling parents, and in planning long-term neurodevelopmental follow-up. Conflict of interest None of the authors declares a conflict of interest. Acknowledgements The authors would like to thank Dr. Andrew Chew for his assistance in evaluating children at follow-up, the staff of the Queen Charlotte's neonatal unit, the Robert Steiner MR unit, and the Hammersmith children's outpatients department. We thank the families of the infants and children in this study. Dr. Begoña Loureiro was funded by the Health Department of Basque Government (grant number: 2011011015). References [1] S. Aher, K. Malwatkar, S. Kadam, Neonatal anemia, Semin. Fetal Neonatal Med. 13 (2008) 239–247. [2] J.M.M. van Klink, H.M. Koopman, D. Oepkes, F.J. Walther, E. Lopriore, Long-term neurodevelopmental outcome after intrauterine transfusion for fetal anemia, Early Hum. Dev. 87 (2011) 589–593. [3] E. Lopriore, D. Oepkes, F.J. Walther, Neonatal morbidity in twin-twin transfusion syndrome, Early Hum. Dev. 87 (2011) 595–599. [4] B.J. Wylie, M.E. D'Alton, Fetomaternal hemorrhage, Obstet. Gynecol. 115 (2010) 1039–1051. [5] A. Stroustrup, L. Trasande, Demographics, clinical characteristics and outcomes of neonates diagnosed with fetomaternal haemorrhage, Arch. Dis. Child. Fetal Neonatal Ed. 97 (2012) F405–F410. [6] G.P. Giacoia, Severe fetomaternal hemorrhage: a review, Obstet. Gynecol. Surv. 52 (1997) 372–380. [7] P. Sinha, N. Kuruba, Ante-partum haemorrhage; an update, J. Obstet. Gynaecol. 28 (2008) 377–381. [8] C. Rubob, P. Deruelle, F. Le Goueff, et al., Long-term prognosis for infants after massive fetomaternal hemorrhage, Obstet. Gynecol. 110 (2007) 256–260. [9] Z. Kecskes, Large fetomaternal hemorrhage: clinical presentation and outcome, J. Matern. Fetal Neonatal Med. 13 (2003) 128–132. [10] I.A. Zonnenberg, R.J. Vermeulen, M.W. Rohaan, et al., Severe neonatal anaemia, MRI findings and neurodevelopmental outcome, Neonatology 109 (2016) 282–288. [11] G. Mari, For the collaborative group for Doppler assessment of the blood velocity in anemic fetuses. Noninvasive diagnosis by Doppler ultrasonography of fetal anemia due to maternal red-cell alloimmunization, N. Engl. J. Med. 342 (2000) 9–14. [12] P. Davies, S. Robertson, S. Hegde, R. Greenwood, E. Massey, P. Davis, Calculating the required transfusion volume in children, Transfusion 47 (2007) 212–216. [13] T.J. Cole, J.V. Freeman, M.A. Preece, British 1990 growth reference centiles for weight, height, body mass index and head circumference fitted by maximum penalized likelihood, Stat. Med. 17 (1998) 407–429. [14] F. Cowan, M. Rutherford, F. Groenendaal, et al., Origin and timing of brain lesions in term infants with neonatal encephalopathy, Lancet 361 (2003) 736–742. [15] M. Cornblath, J.M. Hawdon, A.F. Williams, et al., Controversies regarding definition of neonatal hypoglycemia: suggested operational thresholds, Pediatrics 105 (2000) 1141–1145. [16] L. Haataja, E. Mercuri, R. Regev, et al., Optimality score for the neurological examination of the infant at 12 and 18 months of age, J. Pediatr. 135 (1999) 153–161. [17] R. Griffiths, The Abilities of Young Children, Child Development Research Centre, London, United Kingdom, 1970. [18] C. Cans, Surveillance of cerebral palsy in Europe: a collaboration of cerebral palsy surveys and registers, Dev. Med. Child Neurol. 42 (2000) 816–824. [19] R. Palisano, P. Rosenbaam, S. Walter, et al., Development and reliability of a system to classify gross motor function in children with cerebral palsy, Dev. Med. Child Neurol. 39 (1997) 214–223. [20] E. Quarello, M. Molho, Y. Ville, Incidence, mechanism, and patterns of fetal cerebral lesions in twin-to-twin transfusion syndrome, J. Matern. Fetal Neonatal Med. 20 (2007) 589–597. [21] M.J. Barrett, V. Donogue, E.E. Mooney, et al., Isolated acute non-cystic white matter injury in term infants presenting with neonatal encephalopathy, Arch. Dis. Child. Fetal Neonatal Ed. 98 (2013) F158-F60. [22] S.P. Miller, P.S. McQuillen, S. Hamrick, et al., Abnormal brain development in newborns with congenital heart disease, N. Engl. J. Med. 357 (2007) 1928–1938. [23] L.S. de Vries, F. Groenendaal, Patterns of neonatal hypoxic-ischaemic brain injury, Neuroradiology 52 (2010) 555–566. [24] R.C. Vannucci, S.J. Vannucci, Hypoglycemic brain injury, Semin. Neonatol. 6 (2001) 147–155. [25] C. Burns, M.A. Rutherford, J.P. Boardman, F.M. Cowan, Patterns of cerebral injury and neurodevelopmental outcomes after symptomatic hypoglycemia, Pediatrics 122 (2008) 65–74. [26] M. Martinez-Biarge, T. Bregant, C.J. Wusthoff, et al., White matter and cortical injury in hypoxic-ischemic encephalopathy: antecedent factors and 2-year outcome, J. Pediatr. 161 (2012) 799–807.