Prevalence of MRI abnormalities in people with epilepsy in rural China

Abstract

Objective

To assess the prevalence of brain MRI abnormalities in people with epilepsy in rural China and to compare it with that of individuals in the United Kingdom.

Methods

Brain MRI scans were obtained in people with epilepsy who participated in a rural community-based program in China between July 2010 and December 2012. Individual epileptogenic lesion types were reviewed and their associations with seizure control examined. The MRI findings were compared with 2 previous similar studies in the United Kingdom.

Results

Among the 597 individuals (58% male, median age 38 years) with MRI scans analyzed, 488 (82%) had active epilepsy. The MRI was abnormal in 389 individuals (65%), with potentially epileptogenic lesion in 224 (38%) and nonspecific abnormalities in 165 (28%), and 108 (18%) were potentially resectable. The potentially epileptogenic lesions were less frequently detected in children (<18 years old, 12 of 68, 18%) than in adults (212 of 529, 40%; p < 0.001). In people with potentially epileptogenic lesions, 67% (150 of 224) had failed ≥2 antiseizure medications. They had higher risk of uncontrolled epilepsy than those with normal MRI (risk ratio [RR] 1.25; p < 0.001) and those with nonspecific abnormality (RR 1.15; p = 0.002) after adjustment for age and sex. The diagnostic yield of MRI was similar to that reported in community- and hospital-based studies in the United Kingdom.

Conclusions

More than one-third of people with chronic epilepsy in rural China have potentially epileptogenic lesions identifiable on brain MRI, with two-thirds fulfilling the definition of pharmacoresistance. These findings highlight the magnitude of the unmet needs for epilepsy surgery in China.

The detection of localized pathology on MRI can decisively affect the diagnostic and management pathways in epilepsy, particularly in people refractory to pharmacologic therapy for whom surgery may be a treatment option. Population-based studies on the prevalence of MRI-detected lesions have been reported mostly from high-income countries, examining people with new-onset¹ or established epilepsy.^2,3 In comparison, MRI studies have rarely been performed in low- to middle-income countries, where the majority of people with epilepsy live. Given that the incidence and lifetime prevalence of epilepsy vary with socioeconomic factors,⁴ it is possible that the spectrum and frequency of epileptogenic lesions may also vary, depending on the socioeconomic and environmental settings of the population. A recent hospital-based study of people with new-onset epilepsy in India identified potentially epileptogenic lesions on MRI in almost half of the cohort, mostly due to infection and inflammation.⁵

Six million people are estimated to have active epilepsy in China,⁶ where epilepsy surgery is increasingly performed.⁷ To help estimate the unmet needs for epilepsy surgery in China, we evaluated MRI findings of people with established epilepsy living in the rural communities. The findings were compared with results in previously reported studies from the United Kingdom.^2,3

Methods

Participants

People with epilepsy 2 to 80 years of age being treated in the National Epilepsy Control Program⁸ in rural areas in 4 Chinese provinces (Henan, Hebei, Ningxia, and Shanxi) were recruited between July 1, 2010, and December 31, 2012. People were not included if they had dissociative seizures, alcohol- or drug-related seizures, or progressive or degenerative neurologic or systemic disorders. Those whose seizure control or treatment adherence could not be adequately assessed were also excluded. Pregnant women, people with an intellectual disability unable to provide informed consent, and people in whom MRI is contraindicated, including those with metallic implants or devices and claustrophobia, were not included. Seizure outcomes were assessed at recruitment. Seizure freedom was defined as having no seizures for at least the previous 12 months.

Standard protocol approvals, registrations, and patient consents

The study was approved by the Joint Chinese University of Hong Kong-New Territories East Cluster Research Ethics Committee and by the institutional review board of the Beijing Neurosurgical Institute in China. Written informed consent was obtained from all participants or assent by legal guardians when appropriate.

Clinical assessment tools

As previously described,⁹ predesigned epilepsy history and seizure classification questionnaires were used to collect medical history and seizure information from participants or their caregivers by primary care physicians. The clinical questionnaire consisted of 19 points covering birth, developmental, family, epilepsy, and other medical and drug history. The 33-question seizure classification questionnaire covered a broad range of seizure semiology for classification in accordance with the updated International League Against Epilepsy terminology.¹⁰ The questionnaires were developed on the basis of those previously used for seizure classification^11–13 and were piloted before deployment. Training and standardization workshops for physicians involved were conducted by senior epilepsy specialists (J.W.S. and P.K.).

MRI protocols

Common brain MRI protocols were developed for the participating epilepsy centers. An MRI (1.5T) brain protocol optimized for the investigation of epilepsy that was feasible in the local settings was established. The scanning protocol included a T1-weighted 3D volumetric acquisition sequence with 1-mm isotropic resolution, as well as oblique coronal proton density, T2-weighted, and fluid-attenuated inversion recovery sequences.^14,15

Imaging analysis

The MRI acquisitions were anonymized and identifiable only by the designated individual study identification number. These were systematically evaluated on Osirix PACS (Pixmeo, Geneva, Switzerland)¹⁶ with high-resolution diagnostic monitors by consensus between 2 senior neuroradiologists (I.D. and C.H.) blinded to the demographics, seizure semiology, and all other clinical details. The scans were evaluated for abnormalities that were potentially epileptogenic as previously classified,¹ as well as all other cranial/intracranial abnormalities.

Statistical analysis

Epilepsy was considered active or uncontrolled if the individual had at least 1 seizure in the preceding 12 months at the time of recruitment. For the purpose of analyzing its association with MRI findings, age was grouped into <18, 18 to 34, 35 to 64, and ≥65 years. The Fisher exact test was used to compare categorical variables and proportions of potentially epileptogenic lesions between the rural China cohort and UK cohorts. Multivariable generalized linear models were performed to assess the associations between presence of potentially epileptogenic lesion and epilepsy type, age and sex, and epilepsy control and MRI findings. Statistical significance level was set at p < 0.05. The Holm-Bonferroni method was applied to correct for multiple comparisons. All statistical tests were performed with Stata version 15 (StataCorp, College Station, TX).

Data availability

Anonymized data will be shared by request from any qualified investigator.

Results

Demographics and clinical characteristics

A total of 637 individuals were recruited, of whom 39 were excluded as MRI incompatible and 1 did not complete the questionnaires. Among those 597 who had an MRI, 344 (58%) were male. The median age at entry was 38 (interquartile range [IQR] 27–48) years; median age at epilepsy onset was 14 (IQR 6–25) years; and median duration of epilepsy was 19 (IQR 11–28) years. As previously described,⁹ 525 (88%) individuals had focal, 54 (9.1%) had generalized, and 1 (0.17%) had combined epilepsy; 17 (2.9%) had unknown classification. At recruitment, 524 (88%) were taking antiseizure medications (ASMs), and an additional 46 (7.7%) individuals had been previously treated with ASMs (median number of total ASMs tried 2, IQR 1–3). Among all participants, 488 (82%) individuals had active epilepsy.

MRI findings

Abnormalities are summarized in table 1. Of the 597 individuals, MRI was abnormal in 389 (65%), including potentially epileptogenic lesions in 224 (38%) and nonepileptogenic abnormalities in 165 (28%). Gliosis or encephalomalacia (109 of 597, 18%) was the most common type, which included ulegyria and changes consistent with previous stroke, trauma, and surgery. Mesial temporal sclerosis (83 of 597, 14%) was the second most common epileptogenic lesion found, followed by developmental abnormality (39 of 597, 6.5%). Vascular anomaly (9 of 597, 1.5%) and brain tumors (8 of 597, 1.3%) were relatively uncommon. Twenty-seven individuals had dual pathology, including 19 with mesial temporal sclerosis and gliosis/encephalomalacia, 4 with 2 types of developmental abnormality, 2 with ulegyria and other types of gliosis or encephalomalacia, and 2 with gliosis or encephalomalacia and developmental abnormality.

Table 1.

Summary of MRI findings in the 597 people with epilepsy in rural China

The vast majority (219 of 224, 98%) of people with potentially epileptogenic lesions had focal-onset seizures, 4 (1.8%) had generalized seizures, and 1 (0.45%) had seizures of unknown classification (table 2). Individuals with focal-onset seizures had a significantly higher chance of having a detectable potentially epileptogenic lesion on MRI compared to those who had only generalized-onset seizures (42% vs 7.4%; risk ratio [RR] 5.54, 95% confidence interval [CI] 2.14–14.3; p < 0.001) after adjustment for age and sex.

Table 2.

Frequency of MRI findings according to seizure types and epilepsy classifications

Potentially epileptogenic lesions were less frequently detected in children (<18 years old; 12 of 68, 18%) than in adults (212 of 529, 40%; RR 0.44, 95% CI 0.26–0.74; p < 0.001). Multivariable analysis showed that, compared to children, the risk of detecting potentially epileptogenic lesion was significantly higher in adults 18 to 34 years of age (RR 1.91, 95% CI 1.12–3.27; p = 0.018) and 35 to 64 years of age (RR 2.05, 95% CI 1.22–3.44; p = 0.007) but was not significantly different from those ≥65 years of age (RR 1.17, 95% CI 0.50–2.76; p = 0.71).

Association between MRI findings and seizure control

Epilepsy was controlled in 25% (53 of 208) of those with normal MRI and 22% (36 of 165) in those with nonspecific abnormality, but only 9% (20 of 224) of those with potentially epileptogenic lesion had full seizure control. People who had potentially epileptogenic lesion had significantly higher risk of uncontrolled epilepsy than those who had normal MRI (RR 1.25, 95% CI 1.14–1.36; p < 0.001) and those with nonspecific abnormality in MRI (RR 1.15, 95% CI 1.06–1.26; p = 0.002) after adjustment for age and sex. Among the 224 individuals with potentially epileptogenic lesion, 150 (67%) had active epilepsy despite trials of ≥2 ASMs.

Comparison with previous similar studies

Table 3 illustrates the rates of potentially epileptogenic lesions detected on MRI in rural China and 2 previous studies in the United Kingdom using the same scanning protocol. The early study² was community based and consisted of 174 people with chronic epilepsy who had MRI (1.5T) performed between June 1995 and May 1997. Of the 174, 81 (47%) were male. Their median age was 37 (range 15–73) years; 72% had focal, 25% had generalized, and 3% had combined focal and generalized epilepsy. All had ≥4-year history of epilepsy with at least 1 seizure during the preceding 12 months. The proportion and individual types of potentially epileptogenic lesions were similar to those of our cohort. Another hospital-based study³ of 804 individuals (87% focal, 8% generalized, and 5% unclassified epilepsy) with first attendance for an epilepsy MRI scan (1.5T) between May 1995 and June 2004 also reported a similar potentially epileptogenic lesion rate (36%) and types.

Table 3.

Potentially epileptogenic lesions detected on MRI in people with epilepsy in studies in rural China and United Kingdom

Discussion

We identified potentially epileptogenic lesions in brain MRI in more than one-third of people with epilepsy living in rural China. Gliosis or encephalomalacia accounted for nearly half of the epileptogenic lesions found. The prevalence of mesial temporal sclerosis was also high in the cohort. A number of individuals were found to have various developmental abnormalities. The presence of potentially epileptogenic lesion was associated with focal-onset seizures and uncontrolled epilepsy. The potentially epileptogenic lesions were less prevalent in children compared to adults, likely because a higher proportion of children would have genetic epilepsies while acquired epilepsies are more common in adults. The diagnostic yield and types of lesions were comparable to findings in the United Kingdom.

Notably, a large proportion of the lesions identified are generally considered amenable to resective epilepsy surgery, including mesial temporal sclerosis (14% of the total cohort), vascular malformations (1.5%), focal cortical dysplasia (1.3%), and tumors (1.3%). Combined, these lesions were found in almost a fifth of the participants. These lesions are frequently associated with pharmacoresistance.¹⁷ Indeed, <10% of individuals with potentially epileptogenic lesions had full seizure control in our cohort, and 67% would be considered to have pharmacoresistant epilepsy, having failed ≥2 ASMs.¹⁸

While the mesial temporal sclerosis rate was high in our cohort, it was reported that surgeries for mesial temporal sclerosis in epilepsy centers in Western countries have declined.¹⁹ People with mesial temporal sclerosis in our cohort may have been missed previously due to the lack of radiologic expertise in rural China. Epilepsy surgery also has a shorter history in China, particularly in the rural setting. Furthermore, even when hippocampal sclerosis was detected, people might not have had access to further presurgical evaluation and subsequent intervention. It is also possible that the incidence of mesial temporal sclerosis developing in rural China is higher than in high-income countries. Improvement in pediatric care has been associated with a decline of new cases of mesial temporal sclerosis over the last 30 years in high-income countries. We speculate that these factors might have accounted for the relatively high prevalence of mesial temporal sclerosis detected in this study compared with the declining rate of this pathology in surgical series from Western countries.

These findings have major implications for service provision and highlight the considerable unmet needs in epilepsy management in China, particularly surgical therapy and the required access to specialist neuroimaging. Six million people in China have active epilepsy,⁶ for whom neuroimaging with MRI should be an integral part of evaluation. Extrapolating our findings to the rest of the population, it may be estimated that as many as 1 million individuals with epilepsy may have potentially resectable epileptogenic lesions in China. The most common type of such lesions we identified was mesial temporal sclerosis, for which anterior temporal lobectomy could achieve seizure freedom in >70% of people.²⁰ Identification of these lesions on MRI should therefore trigger presurgical evaluation in epilepsy centers, including optimization of drug treatment, video-EEG monitoring, and neuropsychological assessment. It is almost certain that only a small fraction of people with these lesions has been evaluated for surgery and even fewer have received surgery.²¹

Meeting this demand would require concerted effort in health care resource planning. Because the potentially resectable epileptogenic lesions were more prevalent in people with focal seizures than those with generalized seizures, seizure classification may be used to help prioritize the provision of MRI services. Our findings show that this is possible with the use of standardized questionnaires even in the primary care setting.

Two previous studies in United Kingdom using similar protocols also had high proportions of focal epilepsy (72% and 87%) comparable to the rural China cohort (88%). The 3 studies also had similar MRI diagnostic yields, detecting similar types of potentially epileptogenic lesions and high proportions of mesial temporal sclerosis. Given that these studies used 1.5T MRI scanners, with the advent of high-field magnets at 3T and internationally recommended imaging protocol,²² it is likely that the detection rates of lesions reported in these studies represented an underestimation and the number of cases who should be considered for presurgical evaluation might be even higher.

This study has limitations. There was a bias toward people with convulsive seizures. The reason might be that those with nonconvulsive seizures were less likely to seek medical care in the rural setting. Moreover, most of the individuals in this cohort had chronic and active epilepsy. Therefore, the findings may also not reflect the etiologies of new-onset epilepsy but rather the burden of pathologic lesions in people with active or pharmacoresistant epilepsy.

Through a standardized and locally feasible MRI protocol, we found a high prevalence of potentially epileptogenic lesions in people with epilepsy in rural China, particularly in those with focal seizures. Careful seizure classification, aided by standardized questionnaire on semiology, may help prioritize the use of MRI in resource-limited settings. Our diagnostic yield was comparable to people with chronic epilepsy in high-income countries, suggesting a similar burden of pathological lesions in people with active epilepsy despite different socioeconomic settings. With an estimated 1 million people having potentially resectable epileptogenic lesions, these findings call for urgent actions to address this huge unmet need in epilepsy management in China.

Glossary

ASM

antiseizure medication

CI

confidence interval

IQR

interquartile range

RR

risk ratio

Appendix. Authors

Study funding

This study was supported by a National Institute of Neurological Disorders and Stroke/Fogarty grant (R21NS069223-01) and Key Research Project of the Chinese Ministry of Science and Technology (2016YFC0904400).

Disclosure

I. Davagnanam reports no disclosures relevant to the manuscript. Z. Chen is supported by the Public Health Early Career Fellowship (GNT1156444) from the National Health and Medical Research Council (NHMRC) of Australia and has received research grants from University of Melbourne Early Career Researcher Grant Scheme. C. Hoskote, D. Ding, B. Yang, Y. Wang, T. Wang, and W. Li report no disclosures relevant to the manuscript. J.S. Duncan has received support from Circle Harmony for lecturing. W. Wang reports no disclosures relevant to the manuscript. J.W. Sander has received research grants and honoraria from UCB, Eisai, GW, and Zogenix, which are involved in the manufacturing of antiseizure drugs. P. Kwan is supported by the Medical Research Future Fund Practitioner Fellowship (MRF1136427). He has received research grants from the NHMRC of Australia, Australian Research Council, US NIH, Hong Kong Research Grants Council, Innovation and Technology Fund, Health and Health Services Research Fund, and Health and Medical Research Fund. He/his institution also received speaker or consultancy fees and/or research grants from Biscayne, Eisai Inc, GW Pharmaceuticals, LivaNova, Novartis, UCB Pharma, and Zynerba. Go to Neurology.org/N for full disclosures.