Subtle Microscopic Abnormalities in ... - Wiley Online Library

Epilepsia, 45(8):940–947, 2004 Blackwell Publishing, Inc. C 2004 International League Against Epilepsy

Subtle Microscopic Abnormalities in Hippocampal Sclerosis Do Not Predict Clinical Features of Temporal Lobe Epilepsy ∗ Renate M. Kalnins, †¶Anne McIntosh, ‡∗∗ Michael M. Saling, †¶Samuel F. Berkovic, †§¶Graeme D. Jackson, and †§Regula S. Briellmann

Departments of ∗ Anatomical Pathology, †Neurology, and ‡Neuropsychology, §Brain Research Centre and //Epilepsy Research Institute, ¶ Austin Health Departments of Medicine and ∗∗ Psychology, University of Melbourne, Victoria, Australia

Summary: Purpose: Subtle microdysplastic features are found in some patients with hippocampal sclerosis (HS) and refractory temporal lobe epilepsy. The significance of these findings is unknown. We investigated their frequency, relation to the pattern of HS, and clinical associations. Methods: One-hundred forty patients with histologically confirmed HS (mean age at operation, 35 years; 85 women) were analyzed. The presence of HS and subtle structural abnormalities (SSAs) in the mesial temporal lobe and in the lateral neocortical tissue was assessed in detail. Antecedents, seizure characteristics, two verbal memory tests, and outcome in HS patients with and without SSAs were determined. Results: SSAs were found in 60 (43%) of the 140 HS patients, being mesial only in 32 of the 60 cases, and lateral only in nine cases; the remaining 19 cases had both mesial and lateral abnor-

malities. The frequency of SSA was not related to the pattern of HS or other tested variables. Prolonged febrile convulsions were present in 26 (44%) patients with SSAs, and in 26 (34%) patients (not significant) without SSAs. The outcome after surgery did not differ between patients with SSAs (incidence rate ratio for seizure recurrence, 0.9; 95% confidence interval, 0.5–1.6) compared with patients without SSAs (reference ratio, 1). Conclusions: Forty-three percent of HS patients have SSAs in their lobectomy specimens. The presence of SSAs does not predict clinical characteristics, such as presence of prolonged febrile convulsions, postsurgical outcome, or neuropsychological performance, nor does it correlate with the histologic pattern of HS. Key Words: Temporal lobe epilepsy—Morphology— Hippocampal sclerosis—Pathology.

Some patients with refractory temporal lobe epilepsy (TLE) and hippocampal sclerosis (HS) also have radiologically detected dysplastic abnormalities (1,2). Such cases are regarded as examples of dual pathology (1). However, even when magnetic resonance imaging (MRI) or macroscopic examination reveals no evidence of malformation of cortical development, histologic examination may show minor abnormalities. Such histopathologic abnormalities have been labeled microscopic dysplasia (3) and have been reported (although not universally accepted) in different types of epilepsy (3–5), including refractory TLE in ∼30% (3,6,7). These abnormalities, typically involving the architectonic organization of the brain, may reflect a spectrum from a clearly developmental origin to less well defined abnormalities, probably occurring as a consequence of repetitive seizures. To highlight the uncertain origin of these abnormalities, we use the term subtle structural abnormalities (SSAs). In numerous histologic examinations

of patients with HS, we encountered many cases with subtle structural or morphologic abnormalities either within and around the hippocampus (mesial abnormalities) or in the neocortical tissue of the resected temporal lobe (lateral abnormalities). To date it is not clear whether these SSAs are relevant for the patient with epilepsy. In a recent study, no association between clinical features and lateral abnormalities was reported in a small series of HS patients (7). Further to investigate this issue, we assessed a large, relatively homogeneous group of patients, all with histologically confirmed HS, and investigated the relation of SSAs to the severity of HS and clinical characteristics. This is important, as it has been suggested that dysplastic brain tissue predisposes children with febrile convulsions to prolonged or complex convulsions, leading to HS (8,9). It also has been suggested that developmental lesions may lead to reorganization of functions such as memory and therefore have consequences for cognitive outcome after surgery (10–12). Finally, the presence of abnormal tissue extending beyond the margins of resection has been suggested as a risk factor for poor seizure outcome after temporal lobectomy (13,14). Therefore we explored whether SSAs are associated with a history of prolonged febrile

Accepted February 4, 2004. Address correspondence and reprint requests to Prof G. Jackson at Brain Research Centre, Neurosciences Building, Repatriation Campus, Heidelberg, Victoria 3081, Australia. E-mail: [email protected]

940

SUBTLE MICROSCOPIC ABNORMALITIES AND HS convulsions, are a marker for impaired verbal memory performance, or are associated with poor postoperative outcome. METHODS AND SUBJECTS Subjects We initially included 150 consecutive patients with refractory TLE (mean age at operation, 35 years; 84 women, 66 men) and with a pathologic diagnosis of HS. All patients underwent anterior temporal lobectomy (46% rightsided, 54% left-sided resection) and had resection of rostral and midhippocampus. The study was approved by the Hospital Committee on Human Ethics. Pathological examination Pathological assessment was performed in all 150 subjects by the same experienced neuropathologist, blinded to all other study results. Fresh surgical specimens were placed into buffered 20% formaldehyde solution and were well fixed before processing. Seven-micrometerthick paraffin sections were stained with standard hematoxylin and eosin (H&E), cresyl violet, and Kl¨uver Barrerra techniques. In all cases, >90% of the resected lateral temporal lobe was histologically examined, on average requiring 10 to 12 blocks. All mesial temporal tissue was sectioned, usually resulting in a total of about seven slices, apportioned to three to five blocks. The mesial temporal specimen usually consisted of pes hippocampi and anterior hippocampus, together with a portion of the subiculum. Amygdaloid tissue was included if present. However, in none of the cases was the diagnosis of SSA based only on appearances in the amygdaloid tissue. Parahippocampal gyrus was usually not included. The lateral temporal lobectomy specimen comprised neocortex and white matter of the anterolateral temporal lobe, extending caudally from the vicinity of the temporal pole for 35 to 45 mm. Usually included were the middle and inferior temporal gyri and a portion of the lateral occipitotemporal gyrus. Established criteria were used to diagnose HS (15). Thus classic HS was diagnosed when pyramidal neuron loss was seen predominantly in the CA1 and CA3/CA4 hippocampal subfields, with relative sparing of CA2. Classic HS was subdivided into two categories: severe HS was diagnosed in presence of severe neuron loss in CA3/CA4, as well as in CA1; mild/moderate HS was diagnosed in the presence of only moderate or mild neuron loss in CA3/CA4, with severe loss in CA1. A third category, HS plus dentate sclerosis, was diagnosed when severe neuron loss was found in the dentate gyrus granule cell layer, irrespective of the severity of CA1 changes. Of the initially included 150 patients with a histopathologic diagnosis of HS, 10 cases had an additional major pathologic finding (dual pathology). Four of these 10 patients had a dysembryoplastic neuroepithelial

941

tumor (DNET), two patients had a combination of DNET and ganglioglioma, and one patient had a ganglioglioma. The remaining three cases with dual pathology had a hamartoma-like lesion (one case), an old traumatic lesion (one case), and a benign cystic lesion of uncertain etiopathogenesis (one case). These 10 dual-pathology cases were excluded from further analysis. Diagnosis of subtle structural abnormalities Mesial SSA was diagnosed in the presence of the following architectural abnormalities: Hippocampal abnormalities (for orientation, see Fig. 1, which shows a normal hippocampus; the letters a–d indicate the location of abnormalities, as cited): (a) nodular dentate granular neuron displacement into dentate gyrus molecular layer (Fig. 2). The more diffuse displacement commonly seen in HS was not included; (b) more than three large neurons in the stratum radiatum or stratum lacunosum-moleculare in a ×100 microscopic field (Fig. 3); (c) nodularity, or more marked irregularity than usually seen, of the stratum radiatum margin of the CA1 segment of the cornu ammonis (Fig. 4); and (d) malorientation or clumping of neurons in CA1 or prosubiculum (Fig. 5). Extrahippocampal mesial abnormalities were diagnosed in the presence of nodular collections of neurons in mesial temporal white matter. Neocortical, lateral SSAs were diagnosed in the lateral temporal tissue based on the presence of any one of the following: subtle gyral abnormalities, cortical laminar disarray (Fig. 6), aggregated or large molecular layer neurons, abnormal intracortical neuronal aggregates, or grouped neurons in white matter (Fig. 7). Cases with morphologically abnormal cells (giant neurons, balloon cells)

FIG. 1. Section of normal hippocampus with localization of CA1, CA2, CA3. and CA4 labeled. The letters refer to common sites of minor architectural abnormalities and the anatomic localization of Figs. 5–9. Specifically, (a) labels the dentate gyral molecular layer and refers to the location of Fig. 2; (b) localizes the region of the stratum radiatum and lacunosum-moleculare and refers to the location of Fig. 3; (c) shows the stratum radiatum margin of the CA1 segment and refers to Fig. 4; and (d) indicates the pyramidal neuronal layer of hippocampal field CA1 and refers to the location of Fig. 5. Kluver ¨ Barrera stain, magnification ×10.

Epilepsia, Vol. 45, No. 8, 2004

942

R. M. KALNINS ET AL.

FIG. 2. Dentate gyrus showing displacement of enlarged granular neurons into dentate molecular layer (arrow). This forms part of the pathology of hippocampal sclerosis and is not included among the subtle structural abnormalities group. Kluver ¨ Barrera stain, magnification ×160.

were not found, nor were they expected in the absence of MRI abnormalities. Increased numbers of single white matter neurons were not used as a criterion, because quantification is required to separate abnormal cases from normal controls (6,16). Apparent cortical laminar disorder confined to the region of the temporal pole only was not classified as an SSA. In all cases with mesial or neocortical SSAs, more than one microdysplastic feature was present. Based on the localization of SSAs, patients were subdivided into three groups: those with SSAs found only in the mesial tissue (mesial only), those with SSAs found only in lateral tissue (lateral only), and those with SSAs found in both mesial and lateral tissue (both mesial and lateral). We assessed whether the frequency of SSAs correlated with the pattern of HS (severe HS, mild/moderate HS, HS with dentate sclerosis). Further, we assessed whether the clinical

FIG. 3. Strata radiatum and lacunosum-moleculare showing scattered single large neurons (arrows) as well as aberrant myelinated filbers in CA1 (arrowheads). Kluver ¨ Barrera stain, magnification ×100.


FIG. 4. Nodular projection of aberrantly grouped CA1 pyramidal neurons into stratum radiatum. CA1 is inferior; stratum lacunosum-moleculare is superior. Kluver ¨ Barrera stain, magnification ×75.

features differed between patients with or without SSAs, and whether the clinical features depended on the localization of SSAs (mesial SSAs, lateral SSAs, both mesial and lateral SSAs). Clinical characterization Diagnosis of lateralized TLE was based on intensive presurgical investigation, such as video-EEG telemetry, MRI, positron emission tomography (PET), and single-photon emission tomography (SPECT). Major antecedents were noted when the clinical interview indicated the presence of severe perinatal events, moderate to severe head injury, cerebral infections, stroke, and prolonged febrile seizures. A family history of seizures was assigned when first- to third-degree relatives had epileptic seizures (17). Preoperative MRI was performed by using different protocols and scanners, which precluded detailed analysis of the relation between pathologic and MRI abnormalities. Radiologic diagnosis of HS was based on hippocampal

FIG. 5. Aberrant grouping of pyramidal neurons in CA1. Kluver ¨ Barrera stain, magnification ×200.

SUBTLE MICROSCOPIC ABNORMALITIES AND HS

943

FIG. 7. Nodule of ectopic grey matter in lateral temporal white matter. Kluver ¨ Barrera stain, magnification ×150.

a mean of 7.45 on this task (21), whereas our previously studied group with left HS obtained a mean score of 3.6 points (22). Hard paired-associate learning was included here because of its known sensitivity to left HS. Second, we used a 15-word list-learning task, specifically the acquisition phase (trials 1 to 5) of the Rey Auditory Verbal Learning Test (RAVLT). The score is the sum of words correctly recalled over five trials (maximum, 75), with an age-dependent mean score. The effects of left HS on this test are relatively small (22,23), allowing sufficient variance to detect any additional effects of that might be exerted by the presence of SSAs.

FIG. 6. Cerebral neocortex showing disturbed laminar architecture affecting the full cortical thickness (subarachnoid space is superior). Cresyl violet stain, magnification ×32.

atrophy, increased signal, and disturbed internal architecture (18). Neuropsychological data Neuropsychological data were analyzed only in those patients in whom a preoperative neuropsychological assessment had been completed, who spoke English as a first language, and who had a full-scale IQ of >70. Full-scale IQ was estimated by means of the Reynolds short form of the Wechsler Adult Intelligence Test–Revised (19). One hundred and twenty-two (87%) patients fulfilled these criteria. Of these, 39% were men, and 57% had left HS. The sample did not differ from the overall group in terms of gender distribution or laterality of HS. Verbal memory was assessed by means of two tasks. The first was based on the arbitrary paired associates (so-called “hard” pairs) of the Wechsler Memory Scale–Form 1 (PAHARD) (20). The four hard pairs were learned over three trials. The score was the sum of pairs correctly recalled (maximum, 12). A neurologically normal Australian group obtained

Outcome data Outcome was assessed based on semistandardized personal or telephone interviews (A.M., R.S.B.). Outcome was expressed in two separate analyses: As suggested by Engel et al. (24) for patients with a postoperative followup of >24 months (132 patients), patients were separated into class 1 (free from disabling seizures) and classes 2 to 4 (with rare to frequent seizures). Outcome also was expressed as an incidence rate ratio (IRR) to show the relative risk of development of recurrent seizures in patients with SSAs, in comparison with the patients without SSAs (see Statistics section for details of the analysis method). This analysis was performed in 136 patients. Statistics Analysis was done by using χ 2 tests (nominal variables) and t tests (continuous variables). Neuropsychological data were first assessed irrespective of the laterality of HS, by using t tests. As the laterality of HS influences performance on the PALHARD test, a three-way multivariate analysis of variance (MANOVA) was performed with side of HS, presence or absence of SSAs, and gender as covariates. The IRR for seizure recurrence was calculated based on the number of patients with seizure recurrence and the observed person-years in the following groups: without SSAs and with SSAs. The latter group was subdivided Epilepsia, Vol. 45, No. 8, 2004

944


into mesial SSAs only, lateral SSAs only, or both mesial and lateral SSAs. The group of patients without SSAs was used as the reference to assess the relative risk for the other groups. Seizure recurrence was diagnosed when complex partial or generalized tonic–clonic seizures were noted. Seizures occurring within the first 3 weeks after the operation, drug-withdrawal seizures, and auras were not counted. The person-years are calculated as the group sum of the individual time to the first seizure, or, if seizure free, the time to the last review. The 95% confidence interval (CI) is indicated. The level of significance was set at 5%.

RESULTS SSAs, as defined, were found in 60 (43%) of the 140 patients. Of the 60 patients with SSAs, 32 had mesial abnormalities only, nine had lateral abnormalities only, and 19 had both mesial and lateral abnormalities. None of the patients had diagnosis of SSAs based on findings in the amygdala only. Of the 140 patients, 53 patients had mild/moderate HS, 75 patients had severe HS, and the remaining 12 patients had HS plus dentate sclerosis. The frequency of SSAs was not different between these HS subtypes; they were found in 26 (35%) of 75 patients with severe HS, in 25 (47%) of 53 patients with mild/moderate HS, and in eight (66%) of the 12 patients with HS plus dentate sclerosis (p = 0.07, χ 2 ). Subtle structural abnormalities and epilepsy characteristics The 60 HS patients with SSAs showed no differences in clinical variables, such as seizure frequency, or presence and type of antecedent events, compared with the 80 patients without SSAs (Table 1). However, a trend was noted for an earlier onset of seizures in patients with SSAs (7.2 ± 6 years) compared with patients without SSAs (10.2 ± 9 years; p = 0.06). Prolonged febrile convulsions were documented in 26 (44%) of 60 patients with SSAs and in 26 (34%) of 80 patients without SSAs (p = 0.2, χ 2 ). Localization of subtle structural abnormalities Prolonged febrile convulsions were documented in six (32%) of the 19 patients with both mesial and lateral SSAs, in four (44%) of the nine patients with lateral SSAs only, and in 16 (52%) of 32 patients with mesial SSAs only (p = 0.3, χ 2 ). Gender, side of the seizure focus, or severity of the epilepsy did not differ in relation to the localization of SSAs. Subtle structural abnormalities and MRI findings In 134 of the 140 patients, preoperative MRI reports could be reviewed. Preoperative radiologic assessment suggested the presence of classic HS in 130 of these Epilepsia, Vol. 45, No. 8, 2004

TABLE 1. Clinical parameters in HS patients

Age at operation (yr) Gender: male/female Focus side: right/left Sign Antecedents present Prolonged FC present Onset of epilepsy (yr) Weekly no. of CPSs Total no. of GTCSs

With SSAs (n = 60)

Without SSAs (n = 80)

34 (±11) 25/35 (42/58%) 30/30 (50/50%)

37 (±12) 30/50 (38/62%) 32/48 (40/60%)

39 (66%) 26 (44%) 7.2 (±6) 5 (±7) 70 (±137)

44 (56%) 26 (34%) 10.2 (±9) 7 (±14) 76 (±145)

Values are expressed as mean ± SD. Weekly no. of CPSs is estimated for the last year before operation. Not all clinical data were available in all patients. No significant differences were found between two groups. CPSs, complex partial seizures; GTCSs, generalized tonic–clonic seizures; SSAs, subtle structural abnormalities; FCs, febrile convulsions; HS, hippocampal sclerosis.

134 patients. Classic HS with marked volume atrophy, increased T2 -weighted signal, and disturbed internal architecture was reported in 119 cases, whereas seven cases had the diagnosis made based on the presence of two of these three signs, and in four patients, the MRI abnormalities were somewhat less pronounced, so that mild HS was diagnosed. Four patients, despite the histologic diagnosis of HS, had no HS based on radiologic criteria; three of them had no obvious abnormalities, and in one case, a hippocampal tumor was suspected. Of the 130 patients with radiologic diagnosis of HS, no other MRI abnormalities were noted in 105 subjects. Anterior temporal lobe abnormalities were described in seven cases, contralateral hippocampal abnormalities in six cases, and other associated changes, such as temporal lobe atrophy, in 12 cases. Most of the subjects with other MRI changes, apart from unilateral HS, were scanned during the later stages of the study, reflecting the improvement in the radiologic methods used, but precluding any further analysis of the relation of these changes with the histopathology. Subtle structural abnormalities and neuropsychological findings Differences between patients with and without SSAs were assessed irrespective of the laterality of HS. In the PAHARD test, the mean score was 4.8 ± 3.2 for patients without SSAs and 4.7 ± 3.3 for patients with SSA (p = 0.8; t test). In the RAVLT test, the mean total score was 44.9 ± 10 for patients without SSAs, and 45.1 ± 11 for patients with SSAs (p = 0.9; t test). The MANOVA, with PAHARD and RAVLT as the dependent variables, showed that SSAs do not interact with gender or laterality of HS. As expected, an effect of laterality of HS was seen on the PAHARD test, with patients with left HS showing lower scores than patients with right HS (F 1,78 = 5.45; p = 0.02).

SUBTLE MICROSCOPIC ABNORMALITIES AND HS Subtle structural abnormalities and postoperative outcome

945

on histologic examination. In this study, we excluded patients with a clear-cut second abnormality, such as a coexistent DNET or ganglioglioma, and focused on the analysis of SSAs. In our series of 140 TLE patients with pathologically confirmed HS, signs of SSAs of cerebral tissue were found in 43%. This confirmed our impression that SSAs are often present on histopathologic examination of tissue from patients with HS. A recent study analyzed lateral abnormalities only in 25 HS patients and found them in 29% of their sample (7). This is consistent with our findings, in which about half of our patients with SSAs had these changes confined to the mesial temporal lobe. The occurrence of these microdysplastic changes does not appear to relate to the pattern of HS. More important, our study suggests that these subtle abnormalities are not associated with the presence of a history of prolonged febrile convulsions, and they do not predict verbal memory performance or poor outcome after surgery. As with all research approaching the more subtle end of the spectrum of abnormality, the definition of what is included and what is excluded from the category SSA is crucial. Our definitions of what features constitute SSAs are based on our experience, as well as on previous studies, which have aimed to define cases with “microscopic dysplasia” (3–7). Cell counts were not performed in this study, as SSAs are not typically reflected in changes in cell counts, but consist of changes in architectural arrangement of cells. We have used a neutral term, SSA, to highlight the uncertain origins of structural abnormalities, possibilities ranging from developmental origin to reorganization of brain tissue. Reorganization is well documented in the dentate gyrus, where granular neuron dispersion occurs, a feature not included in this study. Diagnosis of SSAs in the mesial temporal structures is difficult, because of the complexity of the architecture and its often startling

Engel classification This analysis was based on 132 patients with a postoperative period longer than 2 years. Moderate outcome (Engel class 2–4) was found in 19 (32%) of 60 patients with SSAs and in 21 (29%) of 72 patients without SSAs (p = 0.4; Table 2). Outcome did not vary in relation to the localization of the SSAs (see Table 2). Incidence rate ratio This analysis was done in 136 patients, 76 without SSAs, and 60 with SSAs. Of the 76 patients without SSAs, 35 (46%) had seizures during the individual observation period, whereas of the 60 patients with SSAs, 28 (47%) had seizures. These proportions were not different (p = 0.9, χ 2 ). Of the 28 patients with SSAs in whom seizures developed, 18 had mesial abnormalities only, four had lateral abnormalities only, and six had both mesial and lateral SSAs (see Table 2). The proportion of patients in each group with seizures did not differ from the proportion that did not have seizures (p = 0.4, χ 2 ). The IRR for development of seizures was calculated. Patients without SSAs were taken as the reference (equal to 1). Compared with those of these patients, the IRRs for patients with SSAs was not different (0.9; 95% CI, 0.5–1.6; p = 0.8). No difference was found if the localization of the abnormalities was factored into the assessment. However, again a tendency was noted for a better outcome in the group with both mesial and lateral SSAs (see Table 2). DISCUSSION Patients with HS, undergoing temporal lobectomy for intractable seizures, often show additional abnormalities

TABLE 2. Outcome in respect to presence or absence of subtle structural abnormalities Microdysplasia SSAs

Engel classification Number of patients Good outcome (Engel class 1) Moderate outcome (Engel class 2–4) Incidence rate ratio Number of patients Patients with seizure recurrence Person-years Incidence rate ratio (95% CI) p Value

Subgroups

SSAs absent Total

Total

Mesial only

Lateral only

Mesial and lateral

72 51 (71%) 21 (29%)

60 41 (68%) 19 (32%)

32 21 (66%) 11 (34%)

9 6 (67%) 3 (33%)

19 14 (74%) 5 (26%)

76 35 (46%) 304 1a 0.8

60 28 (47%) 263 0.9 (0.5–1.6) 0.4

32 18 (56%) 114 1.3 (0.7–2.3) 0.9

9 4 (44%) 33 0.9 (0.3–2.6) 0.1

19 6 (32%) 126 0.5 (0.2–1.2)

Engel’s classification is based on 141 patients with outcome >2 years. Incidence rate ratio is based on 145 patients with follow-up interviews CI, 95% confidence interval, IRR, incidence rate ratio; SSA, subtle structural abnormality. a Reference value. None of the IRRs is different from the reference value (Poisson regression).


946


variations with alterations in the plane of sectioning. Furthermore, with the passing of the era of en bloc resection of temporal lobes, the surgically supplied mesial temporal specimen commonly comprises little other than Ammon’s horn, with no possibility of systematic analysis of the parahippocampal gyrus and often with damage to, or loss of, CA2. Assessment of SSAs also is difficult in neocortex, with a continuum existing between normality and what is considered abnormal. We did not include a control group in this analysis, as healthy mesial and lateral temporal lobe tissue is excised only in extremely rare instances, and we had no access to such tissue. However, we did not focus on the frequency of these abnormalities in patients compared with controls, but assessed the influence of the presence of SSAs within a homogeneous group of HS patients. Despite these limitations, we believe our study sheds light on an important issue, the relation of SSAs to febrile convulsions, memory, and postoperative outcome. It has been postulated that subtle microscopic dysplasia may predispose those patients experiencing febrile convulsions to prolonged events (9). Pedigree studies raise the possibility of an association between febrile convulsions and subtle forms of malformations of cortical development, as characterized by MRI (25). Early childhood convulsions may damage hippocampal tissue and result in an increased propensity for development of a hippocampal seizure focus later in life, as supported by animal data (26), although the causative link between these events has remained controversial. In the presented series, SSAs were not increased in patients with a history of prolonged febrile convulsions. Furthermore, seizure characteristics did not differ between HS patients with or without SSAs. We assessed the effect of SSAs on neuropsychological performance. Verbal learning is impaired in TLE patients with HS, although variability is found in degree, with some patients showing near-normal results (22). These patients are thought to have shifted their function to the contralateral hemisphere. The presence of SSAs did not correlate with overall performance in this large group of HS patients. This may reflect an overriding of any effects of additional SSAs by the major effects of HS. Alternatively, or in addition, regional differences in SSAs might have different effects on neuropsychological performance, which could be obscured in the group analysis. Many studies have attempted to explain the marked variation in postoperative outcomes of patients with partial epilepsy. Most studies relating dysplasia and postoperative outcome have focused on macroscopic dysplasia, and even then, results are not consistent. This controversy might be explained by the heterogeneity of lesions represented. Patients with circumscribed dysplasia may show a favorable postoperative outcome (6,27). However, dysplastic abnormalities in brain structures often occur beyond the limits of the visible lesion (9,28). In such patients, the postoperative outcome might be worse because Epilepsia, Vol. 45, No. 8, 2004

the entire epileptogenic zone might not be removed by a standard temporal lobectomy (29). In this scenario, the presence of additional SSAs could conceivably be an important marker for more widespread epileptogenic tissue. However, our study of 140 patients with histologically confirmed HS suggests that presence of SSAs does not appear to change the risk of seizure recurrence. This result was obtained by using the classification of Engel (24) and by using IRRs. The proportion of patients classified as seizure free differed between the two methods of analysis, with a higher number of patients classified as seizure free by using the Engel classification (see Table 2). This difference is due to methodologic issues and has been recognized previously (30). Another recent pathology study assessed temporal lobe tissue in 31 patients with HS (16). Patients with a seizure-free outcome showed more white matter neurons than did patients with recurrent seizures, although an overlap was seen between the neuron densities in the two groups (16). We excluded this feature from our analysis, as it requires quantification to identify abnormality clearly. Our study does not give a definite answer as to why some HS patients have additional SSAs, and others have not, and did not aim to investigate whether SSAs are a risk factor for epilepsy. In some patients, SSAs may be developmental, whereas in others, they may represent a reaction to an early insult (31) or may even have resulted from ongoing seizure activity. Animal studies and observations in humans support the view that changes as described in this study can develop during postnatal life (32,33) and in response to seizure activity (32,34). After kindling, both apoptotic cell death and proliferation of dentate gyrus neurons have been observed (34). These new neurons may have ectopic localization and have abnormal connections, perhaps contributing to the development of epileptogenesis (32), although the contribution newly formed granule cells make to the network reorganization remains controversial (35). Dysregulated neurogenesis could conceivably lead to development of SSAs later in life and might represent a feature of human epileptic plasticity. In conclusion, despite SSAs being frequently found in the lobectomy specimens of patients with HS, they did not predict clinical characteristics, such as presence of prolonged febrile convulsions, or postsurgical outcome, nor did such abnormalities correlate with the histologic pattern of HS or the MRI features. They may represent a nonspecific finding that may, in some cases, represent a developmental abnormality, whereas in others, it may be a reaction to the ongoing seizure activity.

REFERENCES 1. Raymond AA, Fish DR, Stevens JM, et al. Association of hippocampal sclerosis with cortical dysgenesis in patients with epilepsy. Neurology 1994;44:1841–5.

SUBTLE MICROSCOPIC ABNORMALITIES AND HS 2. Raymond AA, Fish DR, Sisodiya SM, et al. Abnormalities of gyration, heterotopias, tuberous sclerosis, focal cortical dysplasia, microdysgenesis, dysembryoplastic neuroepithelial tumour and dysgenesis of the archicortex in epilepsy. Epilepsy 1995;118:629– 60. 3. Armstrong DD. The neuropathology of temporal lobe epilepsy. J Neuropathol Exp Neurol 1993;52:433–43. 4. Meencke H-J, Janz D. Neuropathological findings in primary generalized epilepsy: a study of eight cases. Epilepsia 1984;25:8– 21. 5. Opeskin K, Kalnins RM, Halliday G, et al. Idiopathic generalized epilepsy: lack of significant microdysgenesis. Neurology 2000;55:1101–6. 6. Hardiman O, Burke T, Philips J, et al. Microdysgenesis in resected temporal neocortex: incidence and clinical significance in focal epilepsy. Neurology 1988;39:1041–7. 7. Kasper BS, Stefan H, Paulus W. Microdysgenesis in mesial temporal lobe epilepsy: a clinicopathological study. Ann Neurol 2003;54:501– 6. 8. Lewis DV. Febrile convulsions and mesial temporal sclerosis. Curr Opin Neurol 1999;12:197–201. 9. Cendes F, Cook MJ, Watson C, et al. Frequency and characteristics of dual pathology in patients with lesional epilepsy. Neurology 1995;45:2058–64. 10. Bell BD, Davies KG. Anterior temporal lobectomy, hippocampal sclerosis, and memory: recent neuropsychological findings. Neuropsychol Rev 1998; 8:25–41. 11. Saling MM, Berkovic SF, O’Shea MF, et al. Lateralisation of verbal memory and unilateral hippocampal sclerosis: evidence of taskspecific effects. J Clin Exp Neuropsychol 1993;15:608–18. 12. Weintrob DL, Saling MM, Berkovic SF, et al. Verbal memory in left temporal lobe epilepsy: evidence for task-related localization. Ann Neurol 2002;51:442–7. 13. Berkovic SF, McIntosh AM, Kalnins RM, et al. Preoperative MRI predicts outcome of temporal lobectomy: an actuarial analysis. Neurology 1995;45:1358–63. 14. Mathern GW, Babb TL, Vickrey BG, et al. The clinical-pathogenic mechanisms of hippocampal neuron loss and surgical outcomes in temporal lobe epilepsy. Brain 1995;118:105–18. 15. Graham D, Lantos P, eds. Greenfield’s neuropathology. 6th ed. London: Arnold, 1997:936. 16. Thom M, Sisodiya S, Harkness W, et al. Microdysgenesis in temporal lobe epilepsy: a quantitative and immunohistochemical study of white matter neurones. Brain 2001;124:2299–309. 17. Briellmann RS, Torn-Broers Y, Jackson GD, et al. Seizures in family members of patients with hippocampal sclerosis. Neurology 2001;57:1800–4. 18. Jackson GD, Berkovic SF, Duncan JS, et al. Optimizing the diagnosis of hippocampal sclerosis using magnetic resonance imaging. Am J Neuroradiol 1993;14:753–62.

947

19. Reynolds CR, Wilson VL, Clark PL. A four test short form of the WAIS-R for clinical screening. Clin Neuropsychol 1983;5:111–6. 20. Wechsler D. A standardized memory scale for clinical use. J Psychol 1945;19:87–95. 21. desRosier G, Ivson D. Paired associate learning: normative data for differences between high and low associate word pairs. J Clin Exp Neuropsychol 1986;8:637–42. 22. Wood AG, Saling MM, O’Shea MF, et al. Components of verbal learning and hippocampal damage assessed by T2 relaxometry. J Int Neuropsychol Soc 2000;6:529–38. 23. Hermann BP, Wyler AR, Steenman H. The interrelationship between language function and verbal learning/memory performance in patients with complex partial seizures. Cortex 1988;24:245–53. 24. Engel JJ. Outcome with respect to seizures. In: Engel JJ, ed. Surgical treatment of the epilepsies. New York: Raven Press, 1987:553–71. 25. Fern`andez G, Effenberger O, Vinz B, et al. Hippocampal malformation as a cause of familial febrile convulsions and subsequent hippocampal sclerosis. Neurology 1998;50:909–17. 26. Dube C, Chen K, Eghbal-Ahmadi M, et al. Prolonged febrile seizures in the immature rat model enhance hippocampal excitability long term. Ann Neurol 2000;47:336–44. 27. Keene DL, Jimenez C-C, Ventureyra E. Cortical microdysplasia and surgical outcome in refractory epilepsy of childhood. Pediatr Neurosurg 1998;29:69–72. 28. Sisodiya SM, Free SL, Stevens JM, et al. Widespread cerebral structural changes in patients with cortical dysgenesis and epilepsy. Brain 1995;118:1039–50. 29. Kuzniecky R, Ho SS, Martin R, et al. Temporal lobe developmental malformations and hippocampal sclerosis: epilepsy surgical outcome. Neurology 1999;52:479–84. 30. McIntosh AM, Wilson SJ, Berkovic SF. Seizure outcome after temporal lobectomy: current research practice and findings. Epilepsia 2001;42:1288–307. 31. Barkovich AJ, Kuzniecky RI, Jackson GD, et al. Classification system for malformations of cortical development: update 2001. Neurology 2002;57:2168–78. 32. Parent JM, Yu TW, Leibowitz RT, et al. Dentate granule cell neurogenesis is increased by seizures and contributes to aberrant network reorganization in the adult rat hippocampus. J Neurosci 1997;17:3727–38. 33. Lombroso CT. Can early postnatal closed head injury induce cortical dysplasia? Epilepsia 2000;41:245–53. 34. Scharfman HE, Goodman JH, Sollas AL. Granule-like neurons at the hilar/CA3 border after status epilepticus and their synchrony with area CA3 pyramidal cells: functional implications of seizureinduced neurogenesis. J Neurosci 2000;20:6144–58. 35. Parent JM, Tada E, Fike JR, et al. Inhibition of dentate granule cell neurogenesis with brain irradiation does not prevent seizureinduced mossy fibre synaptic reorganisation in the rat. J Neurosci 1999;19:4508–19.
