Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature

Brain & Development xxx (2017) xxx–xxx www.elsevier.com/locate/braindev

Original article

Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature Claire Bar a,⇑, Emmanuel Cheuret b, Pierre Bessou c, Jean-Michel Pedespan a a

Service de Neurologie Pe´diatrique, Hoˆpital Pellegrin-Enfants, CHU de Bordeaux, France b Service de Neurologie Pe´diatrique, Hoˆpital des Enfants, CHU de Toulouse, France c Service d’imagerie ante´natale, de l’enfant et de la femme, Hoˆpital Pellegrin-Enfants, CHU de Bordeaux, France Received 28 February 2017; received in revised form 11 May 2017; accepted 18 May 2017

Abstract Objectives: To describe the clinical course, neuroimaging findings and functional outcome of idiopathic spinal cord infarction (SCI) in adolescents. Methods: Retrospective and descriptive analyses of seven patients with idiopathic SCI and 50 additional cases from the literature were included. Data collected concerned clinical presentation, MRI findings, initial diagnosis, treatments and functional outcome at the last medical visit. Results: Mean age at presentation was 13.2 years (range 13–15). All patients presented a sudden and painful acute myelopathy with <24 h time to maximal symptoms manifestation. A suspected trigger related to a minor effort was reported in 3/7 cases. Six patients presented with paraplegia, one with paraparesis. All had bladder dysfunction needing catheterization. Three patients had an initial misdiagnosis. Initial MRI was considered as normal in 2 cases. In the 5 other cases, T2-weighted-MR images showed hyperintensity within the thoracolumbar spinal cord, affecting mostly the anterior spinal artery territory. Evidence for associated spinal growth dystrophy were present in 6/7 cases. Mean follow-up time was 27.4 months (range 3–46): 2 patients recovered autonomous ambulation, 4 patients regained walking ability with aids and one child (the shortest follow-up) remained wheelchairdependent. A neurogenic bladder was still reported in 6/7 children at the last visit. Complementary analyses with literature cases were consistent with the findings obtained in our cohort. Conclusion: Idiopathic SCI typically occurs in adolescence with a rapid onset and painful acute myelopathy. The MRI shows a T2-hyperintense signal within the spinal cord and provides evidence for an ischemic mechanism. Etiology remains unclear in most cases even though some specific risk factors for this age must play an important role in the pathogenesis, such as mechanical constraints on the immature spine. Ó 2017 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Keywords: Spinal cord infarction; Ischemic stroke; Anterior spinal artery syndrome; Idiopathic; Magnetic resonance imaging (MRI); Spinal growth dystrophy; Child

1. Introduction

⇑ Corresponding author at: Service de Neurologie pe´diatrique,

Hoˆpital Pellegrin-enfants, CHU de Bordeaux, Place Ame´lie RabaLe´on, 33076 Bordeaux cedex, France. E-mail address: [email protected] (C. Bar).

Spinal cord infarction (SCI) is a dramatic acute event, associated with long-term sequelae impacting a patient’s quality of life over the long term [1,2]. No epidemiologic data are available for childhood SCI even

http://dx.doi.org/10.1016/j.braindev.2017.05.009 0387-7604/Ó 2017 The Japanese Society of Child Neurology. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

2

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

though it is considered to be much less frequent than a cerebral ischemic stroke. Due to its rarity, knowledge of the natural history of childhood SCI is mostly based on reports of pediatric cases or studies that mix children and adults. Blood supply to the spinal cord is mostly provided by one anterior and two posterior spinal arteries. The anterior spinal artery (ASA) supplies about two thirds of the cross-sectional area of the spinal cord, including the anterior and lateral corticospinal tracts, the spinothalamic tracts and the anterior horn cells [3]. Therefore, clinical manifestations of SCI depend upon the vascular territory involved. The anterior territory is more vulnerable than the posterior region, which contains many anastomoses. Anterior spinal artery syndrome (ASAS) is an acute myelopathy with sudden back pain followed by the appearance of a rapid bilateral motor deficit and sensory impairment, as well as bladder and/or bowel dysfunction. Typically, temperature and pain sensation are altered, whereas vibration, light touch, and proprioceptive sensations are preserved (due to sparing of the lemniscus pathway) [4]. Magnetic resonance imaging (MRI) usually confirms diagnosis with hyperintense signal on T2-weighted images restricted to a vascular territory [5]. As for cerebrovascular ischemic stroke, the pathogenesis of SCI is different between children and adults. Apart from traumatic causes and common etiologies (including aortic surgery, systemic hypotension, parainfectious vasculopathies, cardioembolic disease, tumor embolism or radiotherapy), it can also occur in an otherwise healthy child without any explanation despite extensive investigation [1,6]. A minor and unnoticed trauma is sometimes reported and thought to constitute as a risk factor [1,7–10], just as the presence of an isolated thrombotic risk factor [11–15]. No study has focused on these idiopathic cases of SCI so that the clinical course, MRI findings and the pathogenesis remain largely unknown. In this study, we described 7 cases of idiopathic SCI in children. We also carried out an exhaustive review of the literature that permitted, along with our 7 cases, to analyse 57 cases of acute idiopathic SCI. As it may be misdiagnosis with other causes of acute myelopathy, and especially acute transverse myelitis, the purpose of this retrospective study was to point out typical features and specificities of the clinical presentation and MRI findings of SCI. We also studied the follow-up data of these children and discussed the potential underlying pathogenesis.

of four different French university hospitals, representing tertiary reference centers (Bordeaux, Toulouse, Tours, Montpellier). Inclusion criteria were: (1) Clinical diagnosis of acute transverse myelopathy, total or partial, including sudden and rapid-onset (<72 h) of sensory, motor, or autonomic dysfunction attributable to the spinal cord [4], (2) MRI evidence for ischemic lesions on imaging (vascular territory, extension of the lesion, restricted diffusion), (3) Age one month to 18 years. Patients were excluded from the study if the SCI was related to a specific etiology known to possibly induce systemic hypotension (aortic diseases, any kind of surgery, arteriovenous malformation) or any other obvious cause like inflammatory diseases (infectious, systemic) or radiotherapy. Thrombotic disorders were considered as a risk factor, not as a direct cause, and were thus included in the study. 2.2. Data collection

2. Patients and methods

The local neuro pediatricians were asked to retrieve all cases of SCI by data base search. For each patient included, we collected information from medical charts regarding past history, suspected trigger, clinical manifestation at presentation, initial diagnosis, imaging and other diagnostic investigations, treatment and functional outcome. As a rapid progression of symptoms (<4 h) argues in favor of ischemia, the delay between onset of deficit and maximal severity was also collected [4]. Functional status was evaluated from the clinical report at the last visit taking into account the level of motor improvement (complete, partial, none), ambulation/locomotion (independent, with aids, or wheelchair), sensory impairment (persistent or not) and bladder disorder (need for catheterization or not). All reports of MRI scans were collected and images were all reviewed by the same radiologist (PB). We tabulated the timing of the scan from the onset of deficits, signal changes on axial and sagittal T2-weighted sequences and diffusionweighted imaging, the precise location of the lesion, the number of spinal levels involved, the presence of enhancement after chelate of gadolinium administration and associated spine modifications. Premature disc degeneration was detected as disk space narrowing with decrease of signal intensity within the nucleus pulposus, related to an altered disk hydration [16]. Schmorl’s node corresponds to a central herniation of the nucleus pulposus throughout the cartilaginous and bony end plate into the body of the adjacent vertebra [16]. This study has been approved by the local Ethical Committee. Informed consent from all patients and/or their legal representatives has been obtained.

2.1. Identification of patient study

2.3. Literature review

Patient cases were retrospectively identified by survey among the clinicians of pediatric neurology departments

The Online database Pubmed.com was used to search additional cases by combining the terms ‘‘spinal cord

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

ischaemia” or ‘‘spinal cord ischemia” or ‘‘spinal cord infarct*” or ‘‘ischemic myelopathy” or ‘‘anterior spinal artery syndrome” or ‘‘spinal cord stroke” or ‘‘acute transverse myelopathy” and ‘‘child*” or ‘‘pediatric”. Titles and abstracts of articles (English or French language) were screened and the appropriate articles were reviewed when they could be accessed. Additional reports cited in the reference lists of relevant articles were also reviewed. Inclusion and exclusion criteria were the same as those for the study patients, as were the specific information collected through reports. 3. Results 3.1. Study patients 3.1.1. Demographic data and clinical manifestations Seven patients from two hospitals (Bordeaux and Toulouse, France) were included in this study. Clinical presentation of these patients is summarized in Table 1. Mean age was 13.8 years (range 13–15) with a sex ratio of 1 male to 6 females. Medical histories were unremarkable, except for a treated asthma (Pat. 1) and an early puberty treated by hormonal injection few years before (Pat. 5). There was no family history of any ischemic events such as juvenile myocardial or cerebral infarction. The patients’ symptoms began during the day, and specific events at time of onset were reported in Patients 1, 3, and 6 (during warming-up before handball training, during a lifting effort, few hours after going down the stairs four steps at a time). In all cases, the progression of symptoms was rapid, lasting less than 24 h. Pain was the first symptom in all cases, radiating from the back or legs, rapidly followed by deficit onset in few minutes (Pat. 1, 2, 3, 4, 5, and 7). In patient 6, there was a symptom-free interval of four hours between pain and motor deficit installation. The motor impairment consisted in a symmetric paraplegia in four patients (Pat. 2, 3, 4, and 6), an asymmetric deficit in two patients (Pat. 1 and 5) and a distal paraparesis in one patient (Pat. 7). Concerning sensory impairment, descriptions were very variable, ranging from the total absence of sensory impairment to a deficit of all sensory modalities. Sphincter disorders were present in all cases, with need of catheterization for neurogenic bladder and sometimes associated to bowel disorder. 3.1.2. Initial laboratory data Cerebrospinal fluid was examined in 6 patients and was normal in all cases. Blood routine tests were also normal, without any sign of inflammation. Serologic tests and cerebrospinal fluid polymerase chain reaction for neurotropic viruses were negative when performed (Pat. 1, 2, 5, and 6), as well as various assays for antibodies (Pat. 1, 2, 3, 5, 6, and 7). Laboratory tests for

3

thrombotic risk factors came back normal; these included the activated partial thromboplastin time, prothrombin time and protein C and S (Pat. 1, 2, 3, 4, 6), antithrombin (Pat. 1 and 3), Factor V Leiden and Prothrombin G21210A mutation (Pat. 1, 4, 6), antiphospholipid antibodies (Pat. 1, 2, 3, 4, 5) and blood homocysteine level (Pat. 3, 6). 3.1.3. MRI findings MRI scans were done at admission for all but one patient, who had the first MRI at one month due to an initial misdiagnosis of Guillain–Barre´ syndrome (Pat. 4). Acute myelitis was the initial diagnosis in two other children, rectified 2 weeks later (Pat. 7) and 1 month later (Pat. 5) on subsequent MRI. The initial MRI, done less than 24 h after onset of symptoms, was interpreted as normal in two cases whose diagnosis of SCI was made on a subsequent MRI a few hours later (Pat. 1) and two days later (Pat. 6). However, re-reading of the first MRI of patient 1 revealed a discrete intramedullary hyper-intensity signal on diffusion-weighted images. MRI findings are presented in Table 1. On axial imaging, typical lesions of anterior spinal artery territories, coined ‘‘snake-eyes” or ‘‘owl’s eyes” (Fig. 1A–D), were found in four patients (Pat. 1, 2, 4, and 5). On sagittal imaging, all patients exhibited vertical T2hyperintensities over several segments at lower thoracic and higher lumbar levels (Fig. 2A–F). There was no enhancement after I.V. contrast injection in any patient. Moreover, after reviewing imaging, most patients presented disc modifications in several segments (Fig. 2A– E) and adjacent to the cord infarction in only one patient (Pat. 5, Fig. 2E). These were interpreted as signs of spinal growth dystrophy (i.e. Scheuermann’s disease) of various degrees. 3.1.4. Therapies and outcomes Five patients initially received steroid injection because of the difficulty to eliminate acute transverse myelitis (Pat. 1, 2, 5, 6, and 7), without any short-term improvement observed. Two patient had intravenous immunoglobulin due to misdiagnosis of Guillain–Barre syndrome (Pat. 4) or transverse myelitis (Pat. 5). The latter was also treated with plasmapheresis. Prophylactic aspirin was administrated to five patients, for variable durations (Pat. 1, 2, 3, 6, and 7). Of note, four patients were given antalgics for neuropathic pains (Pat. 1, 2, 3, and 4). All patients underwent extensive physical therapy. Mean time of follow-up was 27.4 months (±14.1, range 3–46). One patient did not show any motor improvement but his follow-up was the shortest. All the other cases regained substantial motor function, with walking ability. Gait aids consisted in using two crutches (Pat. 1 and 3) or only one crutch outside the

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

T12 Diffuse 13/F 7

<12 h

15/M <4 h 6

T2WI: T2-weight image, F: Female, M: Male, R: Right, L: Left, Temp: temperature, Mod: modalities, CM: Conus medullaris, SN: Schmorl’s nodes, PDD: Premature disc degeneration.

No None Complete/ independent 21

Yes None/wheelchair Persistent PDD/ SN

Bladder bowel Bladder bowel

T11-L1

none

Yes Partial/with aids Unknown PDD / SN CM

Anterior in upper level, diffuse below Diffuse

Right L. monoplegia R. Pain, temp, light leg monoparesis touch / T8-T10 Legs Paraplegia Vibration, light touch, pinprick/T10 Left Distal paraparesis None leg 14/F 5

<4 h

None Paraplegia Legs 13/F 4

<24 h

3

Yes Partial/with aids None PDD / SN CM

36

Yes PDD T8-T12

Anterior in upper level, diffuse below Anterior

Bladder bowel Bladder bowel bladder All mod. /unknown Paraplegia 13/F 3

few hours

14/F 2

minutes

36

Yes

Partial/ Persistent independent Partial/with aids Persistent T7-T11 Anterior

30

Yes Partial/with aids None

L4-L5 disc 20 protrusion / SN PDD 46 T11-L1 Anterior

R. monoplegia L. None Bladder monoparesis bowel Paraplegia Pain, temp/unknown Bladder

Legs, back Back, legs Back <24 h 15/F 1

T2WI Associated disc extension modification T2WI localization Time to max symptoms

Pain

Motor impairment Sensation impairment/ sensory level

Sphincter disorder

Confirmatory MRI findings Clinical presentation

No Age Sex

Table 1 Summary of clinical, imaging and follow-up characteristics of study patients.

Time Motor recovery/ Sensory Neurogenic (months) ambulation impairment bladder

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

Follow-up

4

house due to fatigability after a certain distance (Pat. 4 and 5). Sphincter disorder was still present in 6 patients, with a neurogenic bladder requiring catheterization several times a day (Pat. 1, 3, 4, 5, and 6) or anticholinergic treatment (Pat. 2). 3.2. Study patients and literature cases Fifty additional reports were found in the literature [1,6–15,17–36]. In addition to our 7 patients, these literature cases were analyzed for clinical, imaging and follow-up data. 3.2.1. Demographic data and clinical presentation Mean age at presentation was 13.2 years (±3.2; range 1.5–17) with 30 females (53%) and 27 males (47%). Fig. 3 shows the distribution of cases by sex and age at onset. Most cases were previously healthy-children, except for a known hemoglobin SC disease, two children with obesity, one with Down’s syndrome, one mild idiopathic hypertension, one patient treated with beta blockers for palpitations and one congenital strabismus. A suspected trigger, a temporal correlation with a minor or even unnoticed incident, was reported in 30 patients (53%). It consisted mainly of exercise without traumatism, a minor fall, lifting efforts, sudden movement, and prolonged or forced position. Installation time of maximum symptoms from onset was available in 52 patients and was less than 4 h in 25 patients (48%) and less than 24 h in 47 patients (90%). An initial misdiagnosis was reported in 14 patients (24,5%) and included acute transverse myelitis, conversion disorder, Guillain–Barre´ syndrome, hemiplegic migraine, cord compression, viral illness and acute stroke. Main clinical manifestations are summarized in Table 2. 3.2.2. MRI findings MRI were performed in 56 patients; one diagnosis was made by performing an autopsy on a patient dead of inhalation secondary to a bronchial aspiration. First MRI, all performed the day of admission, was interpreted as normal in 13 patients (23%), even though one should bear in mind that none of them underwent initial Diffusion-Weighted Images (DWI) acquisition. The localization of radiological lesion was available for 49 patients and involved the territory of anterior spinal artery, that is the anterior and/or central part of the spinal cord in 44 patients (90%). One patient had infarction limited to the posterior spinal artery territory and 4 patients had more diffuse lesions. Extension of the lesion was available for 55 patients. The upper part of MRI abnormality was located in the thoracic region for most patients (65%), followed by cervical (25%) and lumbar regions (9%). More particularly, infarction

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

5

Fig. 1. ‘‘Snake-eyes” sign on magnetic resonance imaging (MRI) of the spinal cord. Axial T2-weighted images show an hyperintense signal in the anterior horns of the grey matter in patient 1 (A), patient 2 (B), patient 4 (C) and patient 5 (D).

Fig. 2. Sagittal T2-weighted images of thoracolumbar MRI showing increased signal intensity within the spinal cord from T11 to L1 in patient 1 (A), from T7 to T11 in patient 2 (B), from T8 to T12 in patient 3 (C), in the conus medullaris in patient 4 (D), patient 5 (E) and patient 7 (F). Schmorl’s nodes (arrow), Premature disc degeneration (arrowhead).

was located in thoraco-lumbar region, below T7 level to the conus medullaris, in 30 patients (55%). The radiological lesions extended over several levels in almost all patients (96%). DWI was done in 19 patients (34%) and revealed restricted diffusion with hyperintense signal of the spinal cord in all cases. Degenerative disc disease was reported in 21 patients (38%), included mainly disc herniation, Schmorl’s nodes, disc dehydration and vertebral end-plate signal changes. In these cases, the disc modifications were adjacent to the hypersignal of the spinal cord in 13 patients.

3.2.3. Identification of risk factors and therapy While no obvious cause was identified in any cases, some thrombotic disorders were present in 12 patients (21%). One patient had a known hemoglobin SC sickle cell disease with only 2 pain crises as associated complication. Methylenetetrahydrofolate reductase (MTHFR) gene polymorphism have been revealed by laboratory coagulation testing in 6 patients, but blood levels of homocysteine was reported for only 2 patients and were within normal limits. The other detected abnormalities were one heterozygosis for factor V Leiden mutation,

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

6

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

two heterozygosis for the prothrombin variant, one hereditary heterozygotic protein S deficiency and one diagnosis of primary antiphospholipid syndrome. Concerning reported therapies, 22 patients had steroid injections (39%) and 16 patients followed prophylactic aspirin treatment (28%). 3.2.4. Follow-up Time of follow-up was reported in 44 patients with a mean time of 20.7 months (±16.4; range 1.5–76). Clinical findings at follow-up are summarized in Table 3. 4. Discussion SCI is a rare cause of acute myelopathy in children [37]. Compared to reports from adults in which the most frequent causes is aortic disease [38], the etiology remains mainly unknown in children [1,6,22]. This leads to many questions from parents and adolescents, especially on the long-term prognosis. However, there is very sparse published information available on these idiopathic cases. For the first time, we described clinical presentation, MRI findings and outcome of 7 patients combined with 50 literature cases. 4.1. Clinical features of idiopathic SCI in children Age of onset was congruent with the literature review, occurring during adolescence with a peak at 15 years-old. It was also consistent with previous study of Stettler et al. [1] that included all causes of nontraumatic SCI (with some of those patients included in this study). There was a female predominance in our study group of 7 patients, even though, combined with the literature cases, sex ratio was almost 1:1.

Table 2 Clinical manifestations of study patients and literature cases. n = 57 Pain reported

44

(77%)

Motor symptoms Paraplegia Paraparesis Quadriplegia Quadriparesis Hemiparesis Hemiplegia Diparesis of arms

36 7 5 3 2 3 1

(63%) (12%) (9%) (5%) (4%) (5%) (2%)

Sensory symptoms All modalities Pain and temperature Mixed None Not reported

16 25 1 4 11

(28%) (44%) (2%) (7%) (19%)

Sensory level (n = 53) Cervical Thoracic Lumbar Not reported

2 29 8 14

(4%) (55%) (15%) (25%)

Sphincter disorder Bladder +/ bowel Priapism None Not reported

51 5 1 5

(89%) (9%) (2%) (9%)

Clinical manifestations were similar to adult SCI [2,38] or pediatric SCI with identified cause [1,6]. Typical presentation begun with a sudden pain in a previous healthy child, in lower limbs or in the back, immediately followed by onset of motor weakness or after a short symptom-free interval (from a few minutes to four hours based on our study). Motor deficit mostly consisted of paraplegia/paraparesis occurring less than 24 h after

Fig. 3. Age and sex distribution of 57 cases (study patients and literature cases). M: Male, F: Female.

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

onset of first symptoms. Bladder and/or bowel dysfunction was almost always present. Sensory impairments at the thoracic level were presents in half of cases with affecting pain and temperature in 44% of patients. These symptoms were related to suffering of spinothalamic pathways in anterior spinal artery syndrome. In the presence of this acute myelopathy syndrome, it may be difficult for clinician to differentiate it from other causes of myelopathy, especially acute transverse myelitis. Based on our study and literature cases, an initial misdiagnosis was frequent (24%) and probably underreported. The sudden onset with back pain, a clinical topography consistent with a particular vascular territory and normal lumbar puncture should indicate a diagnosis of ischemia [4,37], but, in our literature study, 10% had more prolonged symptom installation (>24 h) or asymmetric presentation with hemiplegia or hemiparesis (10%). MRI of the spinal cord is essential in this context to provide additional information for the assessment of ischemic changes and help in diagnosis differentiation [5]. 4.2. Radiological features of idiopathic SCI in children Typical radiological findings were present on T2weight images, as a ‘‘pencil-like” hyperintense signal extended over several levels within the spinal cord in sagittal sequences. Compared to other causes of SCI, the anterior territory of the spinal cord was more frequently affected in idiopathic SCI (90% in our study, 41% in the study of Stettler et al. [1]), probably due to more extensive damage in surgery-related infarction or systemic hypotension. On axial sequences, these signals can form a typical pattern of ‘‘snake-eyes” [4,25,39],

Table 3 Clinical findings at follow-up in study patients and literature cases.

7

highly evocative of SCI. Based on our study of 7 patients and compared to adults [5], it seemed difficult to correlate clinical syndrome and infarct territory. First, sensitivity examination may have been incomplete or unreliable in children. Moreover, the exact localization may be difficult to visualize due to the relative small transverse area of the spinal cord in children and limited resolution [4]. Early anatomical imaging was normal in 23%. Therefore, if clinical suspicion remains high, MRI should be repeated. DWI sequences can provide early evidence for ischemia [21,23,25], even though it also may prove difficult to interpret in the spinal cord due to artifact resulting from cord movement and cerebrospinal fluid flow artifacts. Main clinical, MRI and cerebrospinal fluid differences between idiopathic SCI and acute transverse myelitis are listed in Table 4 [40]. 4.3. Therapeutic strategies There is no consensus with regards to therapeutic strategies targeting idiopathic SCI. The high prevalence of steroid administration (39%) highlights the challenge of differentiating ischemia from inflammatory mechanisms. However, this does not appear to aggravate SCI. Given this difficulty and the dearth of treatment options, the injection of steroids should be considered even in cases of doubt, as it can promote recovery in children suffering from acute myelitis. If a thrombotic mechanism was suspected, empirical prophylactic aspirin treatment was often administrated (28%), despite the absence of recurrent cases in the available literature. Because it occurs suddenly in seemingly healthy children, there is currently no known way of preventing it. The most effective treatment involves supportive care and long-term intensive rehabilitation. 4.4. Outcome of idiopathic SCI in children

n= Motor recovery Complete Partial None Not reported

6 45 3 3

(11%) (79%) (5%) (5%)

Ambulation Independent With aids Wheelchair Not reported

16 14 7 20

(28%) (25%) (12%) (35%)

Sensory impairment None Persistent Not reported

1 29 27

(2%) (51%) (47%)

Sphincter disorder Complete recovery Neurogenic bladder Not reported

16 24 17

(28%) (42%) (30%)

Prognosis analyses and comparison with other causes of SCI are limited by relative importance of unreported data about functional outcomes and the variation of follow-up times among case-reports. Nonetheless, idiopathic SCI seems to have a better motor recovery than SCI associated with a known etiology (complete or partial improvement in 90% of our study, 71% in the study of Stettler et al. [1]), which was consistent with walking abilities, with more children walking without aids in our study (43% of reported cases vs 32% in Stettler et al.). In adults, the intensity of neurologic deficits at onset is correlated with long term morbidity [2,39]. No statistical data are available in children but, as in Stettler et al. study [1], our only patient who completely recovered was the one with the less severe clinical examination and the more restricted lesion on MRI. Prognosis on

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

8

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

long-term sphincter disorder was similar to other causes of SCI with around 40% of complete recovery and 60% of neurogenic bladders [1]. 4.5. Pathophysiological mechanism of idiopathic SCI The pathogenesis of idiopathic SCI remains largely unknown. Hypoperfusion of the spinal cord may result from various ischemic mechanisms, including thrombosis, embolic phenomena or local vasospasm. Angiography of the spinal cord is usually not performed in children, as in any of our 7 cases, because of a risk of procedure-related infarction. Spinal arteries are fed by radicular arteries (anterior or posterior) that run along the ventral or dorsal nerve root to the spinal cord. As in previous studies on pediatric or adults SCI [1,2,38], the thoraco-lumbar region was the most affected level. This is likely due to the reliance of this region on only one dominant anterior radicular artery, i.e. arteria radicularis magna or artery of Adamkiewicz, a branch from the intercostal artery [5]. Compared to the study of Stettler et al. that reported lesions involving the ASA territory in 41% of patients [1], it reached 90% in our study. Although systemic hypotension or aortic disease lead to transverse infarction by global hypoperfusion of the spinal cord [3,5], idiopathic SCI is probably caused by a lesion of the radicular artery and spinal arteries with better collateral blood supply in the posterior territory of the spinal cord [5]. Moreover, infarction is often restricted to the anterior horns (‘‘snake-eyes” sign on MRI) because of the higher vulnerability to anoxia of motoneurons of the grey matter [5]. One of the mechanisms frequently suggested in idiopathic SCI is the embolization of fibrocartilage of the nucleus pulposus secondary to a minor traumatism. This hypothesis, first described in 1961 by Naiman et al. [41], found its origin in post-mortem examinations revealing multiple fibrocartilaginous emboli from the nucleus pulposus in spinal cord vessels. These are thought to occur

secondary to axial loading forces on the spinal cord leading to increased intradiscal pressure and fibrocartilaginous emboli that could reach the spinal cord vasculature by a venous or arterial retrograde route [42]. This was the most frequent explanation encountered in our literature review, It was confirmed by histology in one case [34]. In the other cases [6–10,18,19,21,27,28,30,32, 33], it was only based on the following diagnostic criteria: (1) a temporal relation with any event that could cause axial loading, and/or prolonged Valsalva and (2) presence of degenerative disc disease, especially protrusions or Schmorl’s nodes at or near the infarction [42]. In our study, a minor trauma with possible axial loading was reported in 3 of our 7 patients. But, in these otherwise unexplained SCI, a memory bias is likely to overestimated these events. Moreover, 6 of our 7 patients had early degenerative changes of inter-vertebral disc, interpreted as signs of spinal growth dystrophy (Scheuermann’s disease). Scheuermann’s disease is an enthesopathy of the disco-apophyseal junction during growth and the most common cause of structural spinal lesion in adolescence [43]. Its prevalence ranges from 5 to 10%, but can affect up to 50% of adolescents depending on the definition [16]. Interestingly, mechanical factors like repetitive micro traumatisms involving axial loading of the immature spine are known to participate in its pathogenesis, but only very rarely lead to neurological complications [43]. This raises the question if these radiological disc changes are not signs of mechanical stress that could predispose adolescents to SCI rather than providing evidence for a supposed underlying fibrocartilaginous embolism. Without histological confirmation, neither possibility can be confirmed and further research is therefore needed. Various thrombotic disorder was associated with SCI in 21% of our literature cases [1,6,11–15,23,26,35]. However, the role of prothrombotic risks as predisposing factors for stroke is still controversial in childhood [44]. For example, MTHFR gene polymorphisms were most

Table 4 Main clinical, MRI and CSF differences between idiopathic SCI and acute transverse myelitis. Characteristics

Idiopathic SCI

Acute transverse myelitis

Clinical Age Context Pain Time of progression

Adolescent Minor traumatism Acute onset <24 h

Variable Prodromal illness Subacute >48 h

MRI Enhancement Spinal cord T2-hyperintensity Localization DWI Disc modification CSF

No ASA territory Thoracolumbar Restricted Frequent Normal

Frequent Diffuse Cervicothoracic Variable No Inflammatory markers

SCI: Spinal Cord Infarction, DWI: Diffusion Weighted-Images, CSF: CerebroSpinal Fluid.

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

frequently found but, without associated hyperhomocysteinemia, these has not been correlated with higher risk for childhood arterial ischemic stroke [45]. In summary and despite various hypotheses, the exact pathophysiological mechanism of idiopathic SCI is not yet known and may be the result of different predisposing factors in adolescence. 5. Conclusion SCI is a rare condition in childhood, and its etiology remains unclear in most cases. Based on our study group of 7 patients in addition to cases found in the literature, we described for the first time the clinical presentation, MRI findings and functional outcomes of these idiopathic SCI. Typically, children exhibit a rapid onset and painful acute myelopathy with a T2-hyperintense signal on MRI limited to a vascular territory, affecting the anterior spinal artery in most cases. However, it can be challenging for clinicians to differentiate SCI from other causes of acute myelopathy, and consultation with a radiologist is essential to avoid delay or misdiagnosis. Despite the potential recovery of ambulation after long and intense rehabilitative physical therapy, longterm morbidity is high with the possibility of sphincter disorder, which impact the quality of life. The exact pathophysiology is not yet understood, but since it occurs almost exclusively in adolescence, some specific risk factors for this age must play an important role in the pathogenesis, such as mechanical stresses on the immature spine.

Acknowledgements We gratefully thank Pr. Jean-Franc¸ois Chateil for kindly reviewing all the MRI images. Authors also wish to thank collaborators who help for retrieve cases and/ or permitted to include their patients in this study: Dr Fre´deric Villega, Dr Marie Husson, Dr Caroline Espil-Taris, CHU Bordeaux, France; Dr. Jean-Claude Netter, CH Tarbes, France; Dr Delphine Peaureaux, CHU Toulouse, France; Dr Laurent Pradeaux, CH Pe´rigueux, France; Dr Emmanuelle Lagrue, Pr Pierre Castelnau, CHU Tours, France; Pr Franc¸ois Rivier, CHU Montpellier, France. References [1] Stettler S, El-Koussy M, Ritter B, Boltshauser E, Jeannet P-Y, Kolditz P, et al. Non-traumatic spinal cord ischaemia in childhood – clinical manifestation, neuroimaging and outcome. Eur J Paediatr Neurol 2013;17:176–84. [2] Robertson CE, Brown Jr RD, Wijdicks EFM, Rabinstein AA. Recovery after spinal cord infarcts: long-term outcome in 115 patients. Neurology 2012;78:114–21.

9

[3] Novy J, Carruzzo A, Maeder P, Bogousslavsky J. Spinal cord ischemia: clinical and imaging patterns, pathogenesis, and outcomes in 27 patients. Arch Neurol 2006;63:1113–20. [4] Masson C. Acute transverse myelopathy: inflammatory or ischemic? Presse Med 2005;34:869–77 [in French]. [5] Weidauer S, Nichtweiß M, Hattingen E, Berkefeld J. Spinal cord ischemia: aetiology, clinical syndromes and imaging features. Neuroradiology 2015;57:241–57. [6] Nance JR, Golomb MR. Ischemic spinal cord infarction in children without vertebral fracture. Pediatr Neurol 2007;36:209–16. [7] Nelson JA, Ho CY, Golomb MR. Spinal cord stroke presenting with acute monoplegia in a 17-year-old tennis player. Pediatr Neurol 2016;56:76–9. [8] Bansal S, Brown W, Dayal A, Carpenter JL. Posterior spinal cord infarction due to fibrocartilaginous embolization in a 16-year-old athlete. Pediatrics 2014;134:e289–92. [9] Reisner A, Gary MF, Chern JJ, Grattan-Smith JD. Spinal cord infarction following minor trauma in children: fibrocartilaginous embolism as a putative cause. J Neurosurg Pediatr 2013;11:445–50. [10] Tan K, Hammond ER, Kerr D, Nath A. Fibrocartilaginous embolism: a cause of acute ischemic myelopathy. Spinal Cord 2009;47:643–5. [11] Tejero-Ferna´ndez V, Ferna´ndez-Rodrı´guez I, Membrilla-Mesa MD, Arroyo-Morales M. Spinal cord infarction in carriers of methylenetetrahydrofolate reductase-polymorphism-like unique risk factor: report of two cases. Spinal Cord 2014;52(Suppl. 3): S8–S10. [12] Meinicke H, Moske-Eick O, Sitzberger AN, Zieger B, Kirschner J. Anterior spinal artery syndrome in a 13-year-old boy 8 days after taekwondo-fight: vascular obliteration due to vessel lesion or thrombophilia? Klin Pa¨diatr 2011;223:182–6. [13] Hakimi KN, Massagli TL. Anterior spinal artery syndrome in two children with genetic thrombotic disorders. J Spinal Cord Med 2005;28:69–73. [14] Ramelli GP, Wyttenbach R, von der Weid N, Ozdoba C. Anterior spinal artery syndrome in an adolescent with protein S deficiency. J Child Neurol 2001;16:134–5. [15] Young G, Krohn KA, Packer RJ. Prothrombin G20210A mutation in a child with spinal cord infarction. J Pediatr 1999;134:777–9. [16] Chateil J-F, Carlier R-Y, Marty-Poumarat C, Brun M. Maladie de Scheuermann et dystrophie rachidienne de croissance. Image Rhumatol. Orthope´dique 2013;1:946–61 Sauramps Medical. [17] Mathias E, Sethuraman U. Ischemic stroke of the spinal cord: a pediatric emergency in an otherwise healthy child. J Emerg Med 2016;51:73–6. [18] Eid R, Raj A, Farber D, Puri V, Bertolone S. Spinal cord infarction in hemoglobin SC disease as an amusement park accident. Pediatrics 2016;138:e20154020. [19] Rengarajan B, Venkateswaran S, McMillan HJ. Acute asymmetrical spinal infarct secondary to fibrocartilaginous embolism. Childs Nerv Syst 2015;31:487–91. [20] Spencer SP, Brock TD, Matthews RR, Stevens WK. Three unique presentations of atraumatic spinal cord infarction in the pediatric emergency department. Pediatr Emerg Care 2014;30:354–7. [21] Manara R, Calderone M, Severino MS, Citton V, Toldo I, Laverda AM, et al. Spinal cord infarction due to fibrocartilaginous embolization: the role of diffusion weighted imaging and short-tau inversion recovery sequences. J Child Neurol 2010;25:1024–8. [22] Sohal AS, Sundaram M, Mallewa M, Tawil M, Kneen R. Anterior spinal artery syndrome in a girl with down syndrome: case report and literature review. J Spinal Cord Med 2009;32:349–54.

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009

10

C. Bar et al. / Brain & Development xxx (2017) xxx–xxx

[23] Beslow LA, Ichord RN, Zimmerman RA, Smith SE, Licht DJ. Role of diffusion MRI in diagnosis of spinal cord infarction in children. Neuropediatrics 2008;39:188–91. [24] Parazzini C, Rossi L, Righini A, Bianchini E, Mastrangelo M, Spreafico F, et al. Spinal cord and vertebral stroke: a paediatric case. Neuropediatrics 2006;37:107–9. [25] Thurnher MM, Bammer R. Diffusion-weighted MR imaging (DWI) in spinal cord ischemia. Neuroradiology 2006;48:795–801. [26] Cooper D, Magilner D, Call J. Spinal cord infarction after weight lifting. Am J Emerg Med 2006;24:352–5. [27] Raghavan A, Onikul E, Ryan MM, Prelog K, Taranath A, Chennapragada M. Anterior spinal cord infarction owing to possible fibrocartilaginous embolism. Pediatr Radiol 2004;34:503–6. [28] Han JJ, Massagli TL, Jaffe KM. Fibrocartilaginous embolism – an uncommon cause of spinal cord infarction: a case report and review of the literature1. Arch Phys Med Rehabil 2004;85:153–7. [29] Horowitz SH, Patel N. Peripheral neurophysiology of acute distal spinal cord infarction. Pediatr Neurol 2003;28:64–5. [30] Davis GA, Klug GL. Acute-onset nontraumatic paraplegia in childhood: fibrocartilaginous embolism or acute myelitis? Childs Nerv Syst 2000;16:551–4. [31] Wilmshurst JM, Walker MC, Pohl KR. Rapid onset transverse myelitis in adolescence: implications for pathogenesis and prognosis. Arch Dis Child 1999;80:137–42. [32] Tosi L, Rigoli G, Beltramello A. Fibrocartilaginous embolism of the spinal cord: a clinical and pathogenetic reconsideration. J Neurol Neurosurg Psychiatry 1996;60:55–60. [33] McLean JM, Palagallo GL, Henderson JP, Kimm JA. Myelopathy associated with fibrocartilaginous emboli (FE): review and two suspected cases. Surg Neurol 1995;44:228–34. [34] Toro G, Roman GC, Navarro-Roman L, Cantillo J, Serrano B, Vergara I. Natural history of spinal cord infarction caused by nucleus pulposus embolism. Spine 1994;19:360–6.

[35] Hasegawa M, Yamashita J, Yamashima T, Ikeda K, Fujishima Y, Yamazaki M. Spinal cord infarction associated with primary antiphospholipid syndrome in a young child. J Neurosurg 1993;79:446–50. [36] Vandertop WP, Elderson A, van Gijn J, Valk J. Anterior spinal artery syndrome. AJNR Am J Neuroradiol 1991;12:505–6. [37] Verhey LH, Banwell BL. Inflammatory, vascular, and infectious myelopathies in children. Handb Clin Neurol 2013;112:999–1017. [38] Rigney L, Cappelen-Smith C, Sebire D, Beran RG, Cordato D. Nontraumatic spinal cord ischaemic syndrome. J Clin Neurosci 2015;22:1544–9. [39] Masson C, Pruvo J, Meder JF, Cordonnier C, Touze E, de la Sayette V, et al. Spinal cord infarction: clinical and magnetic resonance imaging findings and short term outcome. J Neurol Neurosurg Psychiatry 2004;75:1431–5. [40] Absoud M, Greenberg BM, Lim M, Lotze T, Thomas T, Deiva K. Pediatric transverse myelitis. Neurology 2016;87:S46–52. [41] Naiman JL, Donohue WL, Prichard JS. Fatal nucleus pulposus embolism of spinal cord after trauma. Neurology 1961;11:83–7. [42] AbdelRazek MA, Mowla A, Farooq S, Silvestri N, Sawyer R, Wolfe G. Fibrocartilaginous embolism: a comprehensive review of an under-studied cause of spinal cord infarction and proposed diagnostic criteria. J Spinal Cord Med 2016;39:146–54. [43] Bezalel T, Carmeli E, Been E, Kalichman L. Scheuermann’s disease: current diagnosis and treatment approach. J Back Musculoskelet Rehabil 2014;27:383–90. [44] Heller C, Becker S, Scharrer I, Kreuz W. Prothrombotic risk factors in childhood stroke and venous thrombosis. Eur J Pediatr 1999;158(Suppl. 3):S117–21. [45] Morita DC, Donaldson A, Butterfield RJ, Benedict SL, Bale Jr JF. Methylenetetrahydrofolate reductase gene polymorphism and childhood stroke. Pediatr Neurol 2009;41:247–9.

Please cite this article in press as: Bar C et al. Childhood idiopathic spinal cord infarction: Description of 7 cases and review of the literature. Brain Dev (2017), http://dx.doi.org/10.1016/j.braindev.2017.05.009