Abstract
Hippocampal Malrotation Is Associated With Prolonged Febrile Seizures: Results of the FEBSTAT Study.
Chan S, Bello JA, Shinnar S, Hesdorffer DC, Lewis DV, MacFall J, Shinnar RC, Gomes W, Litherland C, Xu Y, Nordli DR, Pellock JM, Frank LM, Moshé SL, Sun S; FEBSTAT Study Team. AJR Am J Roentgenol 2015;205:1068–1074. doi:10.2214/AJR.14.13330.
OBJECTIVE: Hippocampal malrotation is characterized by incomplete hippocampal inversion with a rounded shape and blurred internal architecture. There is still debate about whether hippocampal malrotation has pathologic significance. We present findings from the Consequences of Prolonged Febrile Seizures in Childhood (FEBSTAT) study on the frequency of and risk factors for hippocampal malrotation. SUBJECTS AND METHODS: FEBSTAT is a prospective multicenter study investigating the consequences of febrile status epilepticus in childhood. MRI studies of 226 patients with febrile status epilepticus were analyzed visually by two board-certified neuroradiologists blinded to clinical details and were compared with MRI studies of 96 subjects with first simple febrile seizure. Quantitative analysis of hippocampal volume was performed by two independent observers. RESULTS: Hippocampal malrotation was present in 20 of 226 (8.8%) patients with febrile status epilepticus compared with two of 96 (2.1%) control subjects (odds ratio [OR], 4.56; 95% CI, 1.05–19.92). Hippocampal malrotation was exclusively left-sided in 18 of 22 (81.8%) patients and bilateral in the remaining four patients (18.2%). There was no case of exclusively right-sided hippocampal malrotation. Hippocampal malrotation was more common in boys than in girls (OR, 6.1; 95% CI, 1.7–21.5). On quantitative volumetric MRI analysis, the left hippocampal volume was smaller in patients with hippocampal malrotation than in control subjects with simple febrile seizure (p = 0.004), and the right-to-left hippocampal volume ratio was higher in the hippocampal malrotation group than in the simple febrile seizure group (p < 0.001). CONCLUSION: Hippocampal malrotation is a developmental malformation that predominantly affects the left hippocampus in male patients and is more frequently found in children with prolonged febrile status epilepticus than in control subjects. These data provide further evidence that hippocampal malrotation represents a pathologic error in brain development rather than a normal variant.
Commentary
Temporal lobe epilepsy (TLE) as a result of hippocampal sclerosis (HS) is often associated with a history of prolonged febrile seizures (FS) and febrile status epilepticus (FSE) (1, 2). In TLE and other epilepsies without HS, subtle abnormalities of hippocampal size, shape, architecture, and orientation may be noted, and the radiologists may label it as “normal,” “variant,” “funny looking,” or “anomaly of unclear significance.” The most common such anomaly is hippocampal malrotation (HIPMAL).
In 2012, FEBSTAT (Consequences of Prolonged Febrile Seizures in Childhood) study, a prospective multicenter study to address the relationship between FSE and HS, reported evidence of acute hippocampal injury by increased hippocampal T2 signal (HT2S) following FSE (3). In addition, HIPMAL was found to be the most common anomaly in the FSE group (n = 20, 10.5%) than in controls (n = 2, 2.1%). In this article, an offshoot of the FEBSTAT study, authors studied the clinical significance of HIPMAL and used quantitative hippocampal volumetric analyses to determine risk factors of FSE.
HIPMAL is incomplete rotation of the hippocampus with an abnormally rounded shape, vertical orientation of collateral sulcus angle, and atypical position and type of fornix as shown in Figure 1. In contrast, HS on visual analysis is defined by loss of volume and increased HT2S typically associated with blurring of hippocampal architecture. Blurring of hippocampal architecture may be present in HIPMAL but is often difficult to confirm especially in young children (<2 years) who may not yet have mature myelination.
Febrile status epilepticus in a 2½-year-old boy. (A) T2-weighted coronal image shows medial positioning and globular shape of left hippocampus. (B) T2-weighted coronal image shows medial positioning and globular shape of left hippocampus, with associated vertical orientation of left collateral sulcus. (C) T2-weighted coronal image shows medial positioning and globular shape of left hippocampus, with associated inferior positioning of posterior left crus of fornix.
The FEBSTAT study originally recruited 199 children aged 1 month to 5 years old who presented with an FSE (continuous or intermittent seizures >30 minutes) that met all the criteria of FS (3). A previously recruited group of 96 children (Columbia University, NY) aged 6 months to 5 years, who presented with a first simple febrile seizure (SFS), served as controls. In this study, to gain sample size for risk factor analysis for HIPMAL, authors “enlarged” the original FEBSTAT FSE cohort (n = 191) by adding two predated cohorts from Duke University (n = 23) and Columbia University (n = 12) for a total of 226 children. Two board-certified neuroradiologists, blinded to clinical details, visually analyzed the MRI of 226 subjects and compared them with 96 controls. Quantitative hippocampal volumetric analysis was done separately.
HIPMAL was more frequent in children with FSE than in the SFS control group. Of 226 children with FSE, 20 (8.8%) had HIPMAL compared with 2 (2.1%) of 96 controls, giving an odds ratio of 4.56 (95% CI, 1.05–19.92). Two risk factors for HIPMAL were male sex and FSE >60 minutes. Prematurity, age of FSE, focal seizure semiology, peak temperature, and delayed development were not associated with HIPMAL. Of intrigue, 22 patients with HIPMAL had it only on the left side (18 patients, 82%) or bilaterally (4 patients, 18%), but never on the right. The story gets more complex when one looks at the relationship of HIPMAL (apparently congenital) and increased HT2S (marker of acute excitotoxic injury). HT2S was high in 4 of the 20 patients (20%) with FSE who had HIPMAL versus 18 of the 206 patients (9%) with FSE but without HIPMAL. Of four with increased H2TS and HIPMAL, H2TS was increased bilaterally in two with bilateral HIPMAL: on the left in one with left HIPMAL, and on the right in one but with left HIPMAL. Of 18 with increased HT2S but without HIPMAL, HT2S was right sided in 14 (status unknown in 4). The study could not show statistically significant concordance of increased HT2S with presence versus absence of HIPMAL. No patient with SFS, even the two with HIPMAL had increased HT2S. In clinical practice, preferential occurrence of HS on one side is not seen. Most surgical series of HS do not describe any side preference (4) except for one series that reported preponderance of right HS in patients with a history of prolonged FS (5). The authors explain their results by evoking possible mechanistic differences; the mechanism of HS development is acute FSE-induced excitotoxic injury to the previously normal (likely right side) hippocampus in contrast to a countervailing mechanism producing left hippocampal injury mainly in patients with HIPMAL who get hit by FSE.
On volumetric analysis, mean (±SD) volume of the left hippocampus was smaller in HIPMAL (2222.3 ± 418.0 mm3) than in SFS (2532.6 ± 402.9 mm3; p = 0.004). The mean volume of the right hippocampus was no different in HIPMAL (2493.4 ± 341.3 mm3) and SFS (2549.4 ± 378.7 mm3; p = 0.6). These differences accounted for a high right-to-left ratio of mean hippocampal volumes in the HIPMAL group compared with the SFS controls (1.14 vs 1.01; p < 0.001). The authors argue that the abnormal ratio is due to small volumes of HIPMAL, suggesting a preexisting hippocampal abnormality, which in association with FSE likely represents a pathologic abnormality rather than a variant of unclear significance. In addition, the authors suggest that HIPMAL may have other clinical consequences besides FSE in the form of reported association in 22q11 deletion syndrome where patients may have neuropsychiatric comorbidities (6), and they cite a cohort of epilepsy patients with HIPMAL who had prefrontal dysfunction (7).
Methodologically, one issue is the “enlargement” of the FEBSTAT cohort to achieve an adequate sample. Brain MRIs are often difficult to compare, even qualitatively let alone quantitatively, across different machines and institutions, and this may be the case in the predated cohorts of Duke and Columbia. As such, HIPMAL is more difficult to diagnose in children under the age of 2, a notable subset of the study cohorts. In addition, the FEBSTAT FSE cohort was not “FS naïve” at the time of enrollment. Nineteen percent of the entire cohort and 26% of the HIPMAL subset had prior FS, raising the question of whether early life (infantile) seizures contribute to the appearance of HIPMAL rather than HIPMAL being a purely congenital anomaly with unidirectional future risk of FSE. It also remains to be seen if HIPMAL as an isolated MRI abnormality in children who present with FSE is the only and main association of FSE. The FEBSTAT study excluded enrollment of children with prior neurologic disabilities. Experience tells us that FSE is common in children with neurologic disabilities in whom HIPMAL may be seen in association with a variety of brain malformations such as extensive multi-lobar/hemispheric dysplasias including hemimegalencephaly, as well as genetic conditions that can cause epileptic encephalopathies. These patients do not have pure TLE.
The study raises multifaceted questions. It brings a higher degree of awareness to recognize HIPMAL as a likely pathologic finding associated with FSE; however, it is far from giving any clear answers. Therefore, as is always the case in epilepsy practice, understanding the clinical context of each patient is important and should continue to guide our clinical decisions.