Posterior fossa subdural hematoma (PFSDH) in neonates is a rare condition with less than 100 cases described in the literature.1-8 A hematoma in the posterior fossa can easily result in death due to compression of the respiratory and cardiovascular centers in the brainstem. Excessive fetal head molding during the delivery process is the most commonly recognized cause.8,9 The treatment modalities for this condition vary in the literature and depend on the patient's clinical state and on the association of brainstem dysfunction or hydrocephalus.2
In this study, we aimed to assess the long-term outcome of the neonates who presented with a PFSDH and the factors associated with their prognosis. The clinical presentation and the management of these patients are described. The mechanism of hemorrhage into the posterior fossa subdural space is considered.
PATIENTS AND METHODS
We performed a retrospective study of the hospital records of patients admitted to the department of neurosurgery of Necker with a PFSDH since 1985.
We used our computer database with keywords “subdural hematoma,” “term newborn,” and “subdural hematoma.”
The demographic and clinical details of the neonate and the mother were obtained from the notes together with the pre-, peri-, and postnatal history. The symptoms on presentation, the radiological findings, and the operative procedure, if undertaken, were noted.
If a patient had not been reviewed in the outpatient department in the past 6 months, a telephone outpatient consultation was undertaken with the patient or parents.
Statistical analysis was performed by using SPSS statistics 20. The χ2 test or the Fisher tests were used for qualitative variables, and the nonparametric Mann-Whitney U test was used for continuous data.
All the patients were admitted in our neurosurgical intensive care unit for close monitoring and surgical assessment. Neonatologists referred them from Ile de France State secondary centers (n = 15) and from our institution (n = 1).
The median delay between birth and onset of symptoms was 10 hours (range, a few minutes to 20 days). We divided the cohort into 2 groups according to this delay as follows: group 1, acute presentation within the first 24 hours (median 2 hours) (n = 9); group 2, subacute presentation (median 4 days) (n = 7). Group 1 was characterized by a predominance of signs of brainstem dysfunction, namely hypotonia (n = 8), drowsiness (n = 7), deglutition disorders (n = 8), respiratory distress (n = 7), and a full anterior fontanel (n = 7). In 5 cases, mechanical ventilation was required. We found mainly signs related with elevated intracranial pressure in group 2, namely macrocrania (n = 6), and hypotonia (n = 4). Data are detailed in Table 1.
All births were vaginal deliveries, and their characteristics are shown in Table 1. Sixteen neonates (11 boys, 5 girls, M/F ratio = 2.2) were studied. In 1 case, the perinatal notes were incomplete. The median age of the mothers was 28 years (range, 22-37 years), and the majority were primiparous (n = 9). There were no medical complications during pregnancy with a median gestational age of 40 weeks (range, 39-41 weeks). The fetus was in the cephalic presentation in 11 cases, and instrument-assisted delivery with forceps was required in 9 cases. At birth, the neonates were unremarkable with regard to weight (median, 3150 g; range, 2585-4100 g), height (median, 50 cm; range, 47-51 cm), head circumference (median, 35 cm; range, 33-38 cm) and the median Apgar score, at 10 minutes (median, 10; range, 1-10).
Transfontanel ultrasound was the first bedside morphological examination in all cases and showed hydrocephalus in 10 cases and mild ventricular distension in 2 cases. The PFSDHs were not recognized on this radiological examination.
Computed tomography (CT) scan confirmed the hydrocephalus in the 10 cases and demonstrated the PFSHD in all cases. The thickness of the clot ranged between 3 and 30 mm (median, 12.5 mm). Its location is indicated in Table 2. There was significant brainstem compression in 11 cases, and the fourth ventricle was compressed in 10 cases, causing triventricular hydrocephalus. In 1 case, supratentorial subarachnoid hemorrhage was also noted.
There were no major bony deformities and no fractures were noted. An magnetic resonance imaging (MRI) was performed in 2 cases. This allowed a better anatomical delineation of the hematoma and an analysis of the brain parenchyma. Figures 1 and 2 show the main characteristics of the hemorrhages.
There was no statistical difference between group 1 (acute presentation) and group 2 (delayed) in terms of clot location (χ2 test, P = .511), thickness of the hematoma (Mann-Whitney U test, P = 1), fetal presentation (χ2 test, P = .511), use of instrumentation during delivery (χ2 test, P = .095), and need for surgery (χ2 test, P = .475).
All of the neonates were referred by a neonatologist and admitted to the intensive care unit. Surgery was performed in 11 of 16 cases, with a median age of 6 days (range, 1-33 days) following delivery (group 1: median, 5 days; group 2: median, 11 days). Interestingly, the management did not differ between group 1 and 2 with either a craniectomy (n = 5 and 4, respectively) or external ventricular drainage alone (n = 1 in both groups). Surgery was indicated when there were clinical persistent signs of brainstem dysfunction or signs of raised intracranial pressure despite appropriate medical management. CT scan was helpful in the decision process, because the patients who were operated on had radiological signs of brainstem compression (Fisher test, P < .001) and hydrocephalus (Fisher test, P = .008). Moreover, the maximal width of the hematoma was significantly larger in the surgery group (median 20 mm vs 10 mm, Mann-Whitney U test, P = .015).
In 9 cases, the hematoma was evacuated via a suboccipital approach, in the sitting position in 7 cases and in the prone position in 2 cases. Preoperative blood tests were always performed (platelet count, prothrombin time, and activated partial thromboplastin time), and vitamin K was systematically given. The aim of surgery was to carefully wash out the hematoma through a small craniectomy while avoiding the cause for the hemorrhage, namely the transverse sinus or vermian veins. Nevertheless, in 1 case, severe intraoperative bleeding from the sinus occurred, which was controlled with compression and absorbable hemostats. On the postoperative CT scan, there was good brainstem decompression with the resolution of hydrocephalus in all cases.
A ventriculoperitoneal shunt was subsequently required in 2 cases: one from group 1 and one from group 2 (Table 2). One of these was following the evacuation of the PFSDH together with an intracerebellar hematoma and the other following an external ventricular drainage. This patient also had a cerebellar parenchymal hematoma. Five patients were not operated on, because they had only mild symptoms with no hydrocephalus on CT (n = 3 and 2 for groups 1 and 2, respectively).
With a mean follow-up of 7.8 years (range, 4 months to 24 years), we observed neurological deficits in 5 patients. Neurological deficits were severe in 1 patient (patient 1) with global developmental delay and axial and peripheral hypotonia (follow-up, 10 months). Another patient (patient 7) had mild delay with gait disturbance and oculomotor palsy, but he nevertheless managed to pass his driving license (follow-up, 24 years). There were minor deficits in 4 patients with hypoacusia (patient 11) and/or oculomotor palsy (patients 5, 10, and 11). Interestingly, 3 patients with intracerebellar hemorrhage had persistent neurological deficiencies on last follow-up (patients 1, 10, and 11).
Regarding long-term functional outcome, 6 patients had a follow-up of more than 10 years (Table 2). Among 4 patients older than 18 years, 3 are employed and independent (patients 4, 6, and 16), and 1 is employed but still lives with his parents and has a limited social life. Two patients aged younger than 18 years follow normal school, without any help.
When looking for risk factors of permanent neurological deficit, it is interesting to note that the 3 patients with intracerebellar hemorrhage had abnormal examination results, because 1 of them had global delay (patient 1), and 2 of them had persistent oculomotor palsy.
In this retrospective and single-center series, the presentation, management, and outcome of children with a perinatal subdural posterior fossa hemorrhage have been described.
Symptomatic PFSDH in the newborn is rare, with fewer than 100 cases reported in the literature (Table 3). The rate of births in our state is 185 000 per year, and, in a 25-year period, we admitted only 16 patients with PFSDH in our department.
Asymptomatic PFSDH is less exceptional, yet rarely diagnosed. Indeed, in a series of 111 asymptomatic term infants, an MRI was systematically performed for epidemiological purposes, and subdural hematoma was noted in 9 cases (infratentorial, n = 7 or supratentorial, n = 2).10 Because the small hematoma volume, and the absence of symptoms, none of these newborns required surgery. The incidence was smaller in another series on 558 neonates who underwent brain MRI for research purposes (n = 144) or because they were admitted to intensive care unit or had an neurological examination with abnormal results (n = 444).11 A total of 422 abnormal scans were identified, in which 20 infants were diagnosed with posterior fossa abnormalities and 14 were PFSDH (6 term infants and 8 preterm infants).
Birth is, to a certain extent, a factor of head traumatism that may sometimes lead to brain hemorrhage, particularly with vaginal delivery. Exceptional cases of posterior fossa hematoma have been reported on antenatal MRIs or ultrasounds, or after cesarean delivery.12,13 The acquisition of bipedalism by humans during phylogenesis has caused significant difficulties with the birthing process. The female's pelvis has evolved from a wide cylinder in the nonhuman primate to a narrow, truncated hourglass. In addition, the mean brain volume has tripled between nonhuman and human primates.14 As distinct to most quadruped species, the birth delivery in humans requires strong uterine contractions resulting in an increase in the amniotic pressure. This change in pressure is transmitted to the deformable fetal skull as it passes through the birth canal.15 The amniotic pressure increased 2-fold during uterine contractions, whereas the head pressure values increased 4-fold (reaching 40 kPA) as a consequence of the rising resistance of the narrowing pelvis and stretched passageway.15 Obstetric observations and biomechanical modeling have revealed that these high pressures lead to a process known as fetal head molding.9,16 During this process, there is a lengthening of the anteroposterior axis, whereas there is a decrease of the biparietal and suboccipital diameters. The lifting of both parietal bones and the occipital protuberance causes the latter.9,16 These mechanical forces can be transmitted to the falx and the tentorium, leading to damage of major venous channels such as the straight and transverse sinuses, vermian veins, and the great vein of Galen. This could lead to the development of subdural and intraparenchymal hemorrhage around the tentorium and in the posterior fossa.7,8 Other etiologies have been suggested, namely exceptional tumors or clotting disorders.8 The authors report in their series of 15 patients 4 cases with abnormal thrombin or partial thromboplastin time.
The numbers of neonates with such a hemorrhage that are referred to neurosurgeons are likely to represent a small percentage, the majority probably going unnoticed. This is supported by the findings of Takagi who described an occurrence of 10% PFSDH among 229 autopsies of intracranial hemorrhages.17
Some risk factors of PFSDH have been described in the literature as primiparity.1,2,5 In the current series of 16 cases, the majority of the mothers were primiparous (9 cases). Other reported risk factors include breech presentation and the use of instrumentation, both of which were present in a significant number of the patients in the current series.8 The fetal head circumference was in a normal range, between 33 and 38 cm, and therefore not considered to be a risk factor in this study.
In this series, it was possible to distinguish 2 types of clinical presentation: an acute presentation in the first few hours of life with signs of brainstem compression (respiratory distress and oculomotor paresis) and, second, a delayed presentation after a few days with signs of hydrocephalus (increased head circumference, tense fontanel, vomiting). This would be in accordance with the results from a previous study.7 A possible explanation for these 2 types of presentation is that the hemorrhage is low-pressure venous bleeding. In the first group presenting acutely, the bleeding stops when the pressure in the posterior fossa is high. This causes brainstem compression and neurological deterioration. In the second group with a delayed presentation, the bleeding stops spontaneously but subsequent swelling causes obstructive hydrocephalus. In our series, there was no difference between these 2 groups in terms of clot location, thickness of the hematoma, fetal presentation, use of instrumentation during delivery, and need for surgery, but a larger group may be required to draw definitive conclusions.
In the current series, ultrasound was useful to diagnose hydrocephalus, but it failed to detect accurately the posterior fossa hemorrhage. In previous studies, ultrasound was very sensitive in detecting a hematoma in the central incisura but not in the retrocerebellar location.4,6 Another study showed that cranial ultrasound is not sensitive enough, because it could detect abnormalities in only 7 of 19 patients with posterior fossa lesions on MRI.11 Therefore, the most useful imaging modality in these emergency cases remains CT. The coronal and sagittal reconstruction, in addition to the standard axial images, allows a precise localization of the hemorrhage and good evaluation of the brainstem compression and hydrocephalus.8 A CT classification was previously proposed with 4 types of hemorrhage, depending on the location of the hematoma at the center or the periphery of the falcotentorial junction, and the occurrence of intracerebellar hemorrhage.5 The authors propose that surgical evacuation of the hematoma is mandatory when there is an intracerebellar hemorrhage (type IV), or when there is intracranial hypertension or hydrocephalus (type II and III). Hematoma limited in the tentorial incisura (type I) should be treated conservatively with serial imaging. This classification was not used to determine the need for surgical intervention in the current series, and we think that 3 imaging criteria should be taken into account in the decision process: radiological signs of brainstem compression, hydrocephalus, and maximal width of the hematoma (all patients with a hematoma >15 mm had surgery). An MRI scan was obtained in only 2 patients in our series and did not add anything to the clinical decision-making process. A delayed MRI scan, however, may help predict future neurological deficits.
In the present series, all of the neonates were admitted for assessment and for possible neurosurgical intervention. Criteria to perform surgery in previously published series and the current series are consensual, namely brainstem dysfunction or obstructive hydrocephalus.2,3,5,7,8 The operative technique in the present series involved a small suboccipital craniectomy and a gentle aspiration of the clot. The aim of surgery was to decompress the mass effect on the brainstem/cerebellum and to restore the cerebrospinal fluid (CSF) pathways. The venous sinuses were not directly visualized.
Five patients in this series did not have any surgical procedure. In these 5 patients, 3 presented within 1 hour of birth and all presented with mild symptoms, for example, an increase in head circumference. It was therefore felt that immediate surgical intervention was not required and that these neonates could be observed expectantly. Previous series have reported successful conservative management of posterior fossa hematomas, when the patient did not present with signs of brainstem compression or hydrocephalus.4
In the current series, only 2 of 16 patients required a ventriculoperitoneal (VP) shunt. In these 2 patients, one had an intracerebellar hemorrhage, as well as the PFSDH, and the other did not require evacuation of the hematoma but required an external ventricular drain. This finding is similar to 2 other large series that report a low incidence of hydrocephalus (2/13 patients for Hayashi et al5 and 2/15 patients for Perrin et al8). Furthermore, these studies note a possible association of supratentorial subarachnoid hemorrhage with the requirement for a VP shunt. A low rate of VP shunt insertion is, however, in contrast to other studies that have reported an incidence of 36% of hydrocephalus and have stated that posterior fossa hematoma evacuation does not prevent its occurrence.2,7
In the present study, 6 patients had a long-term follow-up of at least 10 years. Five of these 6 patients attended a normal school. Three of the 4 patients who were old enough to work were in further education or gainful employment. One of them had a dramatic clinical presentation with brainstem compression resulting in coma with bilateral mydriasis and corneal reflex abolition, yet this patient is now a bank employee. These results are in accordance with those from Hayashi et al who studied the development from 6 months to 3 years after birth and found a normal development in 40 of 48 patients with a PFSDH and 10 of 13 patients who had been operated on.5 Similar results were found in another study with a good outcome for 1 of 15.8
Yet, a few patients have persistent neurological signs or developmental delay (3 in our series). Perrin et al8 suggested that this was linked to supratentorial brain insults like ischemia, but the other studies were not informative. In our study, we did not find clear risk factors for sequelae, mainly because of the small size of our cohort. Interestingly, we note that the 3 patients with intracerebellar hematoma had global delay or persistent neurological signs. We suggest that a more careful neuropsychological follow-up may be necessary in case of intracerebellar injury. This could detect troubles in the development of executive functions, spatial cognition, or language like the “dysmetria of thought” hypothesis reported by Schmahman and colleagues in vascular insults, or more recently in posterior fossa syndrome associated with cerebellar tumors.18-20
Symptomatic PFSDH is a rare condition in the term newborn population. It generally occurs following vaginal delivery, in primiparous women. The clinical presentation can be dramatic because of brainstem compression. Following resuscitation, the evacuation of the posterior fossa hematoma is safe, reopens the CSF pathways, thus minimizing the need for a VP shunt, and allows for a good long-term cognitive and functional outcome in most patients. In the absence of significant neurological impairment, a conservative approach can be safely used.
The authors have no personal financial or institutional interest in any of the drugs, materials, or devices described in this article.
We thank Dr Alain Pierre-Kahn, MD, who operated on 3 patients in this series.
1. Serfontein GL, Rom S, Stein S. Posterior fossa subdural hemorrhage in the newborn. Pediatrics. 1980;65(1):40–43.
2. Menezes AH, Smith DE, Bell WE. Posterior fossa hemorrhage in the term neonate. Neurosurgery. 1983;13(4):452–456.
3. Hernansanz J, Muñoz F, Rodríguez D, Soler C, Principe C. Subdural hematomas of the posterior fossa in normal-weight newborns. Report of two cases. J Neurosurg. 1984;61(5):972–974.
4. Koch TK, Jahnke SE, Edwards MS, Davis SL. Posterior fossa hemorrhage in term newborns. Pediatr Neurol. 1985;1(2):96–99.
5. Hayashi T, Hashimoto T, Fukuda S, Ohshima Y, Moritaka K. Neonatal subdural hematoma secondary to birth injury. Clinical analysis of 48 survivors. Childs Nerv Syst. 1987;3(1):23–29.
6. Huang CC, Shen EY. Tentorial subdural hemorrhage in term newborns: ultrasonographic diagnosis and clinical correlates. Pediatr Neurol. 1991;7(3):171–177.
7. Govaert P, Calliauw L, Vanhaesebrouck P, Martens F, Barrilari A. On the management of neonatal tentorial damage. Eight case reports and a review of the literature. Acta Neurochir (Wien). 1990;106(1-2):52–64.
8. Perrin RG, Rutka JT, Drake JM, et al.. Management and outcomes of posterior fossa subdural hematomas in neonates. Neurosurgery. 1997;40(6):1190–1199; discussion 1199-1200.
9. Lapeer RJ, Prager RW. Fetal head moulding: finite element analysis of a fetal skull subjected to uterine pressures during the first stage of labour. J Biomech. 2001;34(9):1125–1133.
10. Whitby EH, Griffiths PD, Rutter S, et al.. Frequency and natural history of subdural haemorrhages in babies and relation to obstetric factors. Lancet. 2004;363(9412):846–851.
11. Miall LS, Cornette LG, Tanner SF, Arthur RJ, Levene MI. Posterior fossa abnormalities seen on magnetic resonance brain imaging in a cohort of newborn infants. J Perinatol. 2003;23(5):396–403.
12. Folkerth RD, McLaughlin ME, Levine D. Organizing posterior fossa hematomas simulating developmental cysts on prenatal imaging: report of 3 cases. J Ultrasound Med. 2001;20(11):1233–1240.
13. Coker S, Beltran R, Fine M. Neonatal posterior fossa subdural hematoma. Clin Pediatr (Phila). 1987;26(7):375–376.
14. Raynal P, Le Meaux JP, Chéreau E. Anthropologic evolution of women's pelvis. [in French]. Gynecol Obstet Fertil. 2005;33(7-8):464–468.
15. Rempen A, Kraus M. Pressures on the fetal head during normal labor. J Perinat Med. 1991;19(3):199–206.
16. Sorbe B, Dahlgren S. Some important factors in the molding of the fetal head during vaginal delivery—a photographic study. Int J Gynaecol Obstet. 1983;21(3):205–212.
17. Takagi T, Fukuoka H, Wakabayashi S, Nagai H, Shibata T. Posterior fossa subdural hemorrhage in the newborn as a result of birth trauma. Childs Brain. 1982;9(2):102–113.
18. Schmahmann JD, Sherman JC. The cerebellar cognitive affective syndrome. Brain. 1998;121(pt 4):561–579.
19. Schmahmann JD. The role of the cerebellum in cognition and emotion: personal reflections since 1982 on the dysmetria of thought hypothesis, and its historical evolution from theory to therapy. Neuropsychol Rev. 2010;20(3):236–260.
20. Puget S, Boddaert N, Viguier D, et al.. Injuries to inferior vermis and dentate nuclei predict poor neurological and neuropsychological outcome in children with malignant posterior fossa tumors. Cancer. 2009;115(6):1338–1347.
This article by Blauwblomme et al is a substantive addition to the literature of the rarely encountered problem of large posterior fossa bleeds in neonates. Neonatal posterior fossa surgery can be treacherous. The authors have demonstrated that, when indicated, timely and judiciously performed surgery to evacuate posterior fossa hematomas in the newborn can be both safe and efficacious. The long-term outcome of this series is favorable, again highlighting the importance of expedient management at the outset. The authors are to be congratulated. Hopefully, further series such as this will shed light on the etiology and pathophysiology of this condition.
In this retrospective case series, Blauwblomme et al reviewed the case management of neonates with posterior fossa subdural hematomas (PFSDHs). Based on their results, they concluded that surgery could be done safely in neonatal children with clinical and radiological signs of brainstem compression, hydrocephalus, or raised intracranial pressure from PFSDH. They found that aggressive resuscitation and intervention should be performed even in cases of initial severe brainstem dysfunction because of the good outcomes generally for this process if diagnosed and managed early. Although this is a disorder treated ultimately by neurosurgery because of an acute or subacute decline in neurological status, a good outcome for these children depends on the awareness of the entity, clinical recognition in individual basis, and urgency of communication to neurosurgery by the neonatologist involved in the care. Despite PFSDH being a rare condition, and because early aggressive management leads to good outcomes, further study of the use of insonation through the anterior fontanelle with ultrasound in the early neonatal period may be useful as a screening tool to assess for intracranial and intracerebellar/posterior fossa hemorrhages.
P. David Adelson