Epilepsy and the risk of COVID-19-related hospitalization and death: A population study

1 INTRODUCTION

The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has caused significant morbidity and mortality worldwide.^1-4 Epilepsy is one of the most common neurological conditions, affecting approximately 50 million people worldwide, with significant comorbidities and an increased risk of premature mortality.⁵

The COVID-19 pandemic caused significant changes in health care globally. Health care professionals normally providing epilepsy care were redeployed to COVID-19 work. Epilepsy consultations and investigations were also canceled, making it harder for people with epilepsy to access appropriate care.^{6, 7} In some countries, including Wales, people have been deterred from seeking medical attention for serious non-COVID-19 problems, including epilepsy-related issues, due to perceived COVID-19 risk.^8-10

SARS-CoV-2 is known to directly affect the central nervous system, with high rates of neurological and psychiatric complications.^{11, 12} There is some evidence that people with epilepsy are at increased risk of COVID-19 hospitalization and COVID-19 death.^13-16 This may be in part due to the increased proportion of people with epilepsy with comorbidities, such as dementia and intellectual disability, that increase the risk of severe COVID-19.^{13, 16}

SARS-CoV-2 vaccines have significantly improved COVID-19 outcomes.¹⁷ However, given the significant morbidity and mortality that COVID-19 has caused, and the potential for further COVID-19 and other pandemics with new variants and pathogens, it is important to accurately quantify COVID-19 risks in specific populations. There have been very few population-level, epilepsy-specific studies on serious COVID-19 outcomes to date. In this study, which was part of the Coronavirus and Epilepsy in Wales project, we aimed to use detailed, population-level data to assess whether people with epilepsy were more at risk from severe COVID-19 outcomes, namely COVID-19 hospitalization and death.

2 MATERIALS AND METHODS

We used the Secure Anonymised Information Linkage (SAIL) databank, which contains anonymized individual-level, population-scale, routinely collected electronic health records from multiple different sources. These include hospital admission and demographic data for the complete Welsh population (3.1 million) and primary care records for 86% of the population.^{18, 19} We used the Controlling COVID-19 Through Enhanced Population Surveillance and Intervention project dataset within SAIL (Project 0911).²⁰

We identified people with epilepsy living in Wales before or during the study period (March 1, 2020 to June 31, 2021). We used a previously validated method to define epilepsy as a primary care diagnosis of epilepsy and prescription of at least two antiseizure medications.²¹ To reduce confounding and the dimensionality of regression models, we created a control cohort using exact 5:1 matching on sex, age, and socioeconomic deprivation quintile measured using the Welsh Index of Multiple Deprivation 2019 (WIMD). WIMD uses weighted scores from eight domains to form a score for small geographical areas or Lower Layer Super Output Areas (LSOAs). WIMD scores for each LSOA are then grouped into quintiles, with Quintile 1 being the most deprived and Quintile 5 being the least deprived.²²

We recorded SARS-CoV-2 vaccination status and comorbidities known to increase the risk of COVID-19 hospitalization and death. The comorbidities are based on those used in the QCOVID algorithm (Table 1).^{13, 16, 23} As a sensitivity analysis, we also created an additional control cohort (control cohort 2) using 5:1 exact matching for age, sex, and deprivation with additional propensity matching (logistic regression) based on the nearest propensity score for QCOVID comorbidities. Further to this, we created stratified models for key groups of interest including age (0–18, 18–65, 65+ years) and deprivation (individuals living in areas in the two most deprived quintiles) to highlight differences in these subgroups.

We defined COVID-19 deaths as having International Classification of Diseases, 10th Revision (ICD-10) codes for COVID-19 on death certificates (in any position, e.g., primary or secondary cause of death) or occurring within 28 days of a positive SARS-CoV-2 polymerase chain reaction (PCR) test. In the UK, death certificates have space to record the immediate cause, underlying cause, and conditions contributing to death. COVID-19 hospitalizations were defined as having a COVID-19 ICD-10 code for the reason for admission or occurring within 28 days of a positive SARS-CoV-2 PCR test.

We used Cox proportional hazard models to calculate hazard ratios (HRs) for COVID-19 hospitalization and death. We used Schoenfield residuals to test for proportionality. For some adjusted models, comorbidities we removed due to none/very few people in one or both cohorts having the diagnosis. Details of proportionality tests are given in Table S2, and comorbidities used/removed for each model are indicated in Tables S3 and S4.

We used R (version 4.1.3) for statistical analysis.

We collaborated with Epilepsy Action, a leading UK epilepsy charity, discussing project design, results, and this article with Epilepsy Action and volunteers with epilepsy. We produced a video that showed some of the views of people living with epilepsy on their experiences of the effect of COVID-19.²⁴

This study was approved by the SAIL independent Information Governance Review Panel (Project 0911). The Research Ethics Service has previously confirmed that SAIL projects using anonymized, routinely collected data do not require National Health Service research ethics committee approval.

3 RESULTS

There were 27 279 people in the epilepsy (case) cohort and 136 395 people in the control cohort (epilepsy prevalence of .96% at beginning of study and .95% at end of study). There were 158 (.58%) COVID-19 deaths and 933 (3.4%) COVID-19 hospitalizations in people with epilepsy, and 370 (.27%) deaths and 1871 (1.4%) hospitalizations in controls (Table 2).

Comparing people with epilepsy to controls, the HRs for COVID-19 death and COVID-19 hospitalization were 2.15 (95% confidence interval [CI] = 1.78–2.59) and 2.15 (95% CI = 1.94–2.37), respectively. Adjusted HRs (aHRs; adjusted for comorbidities associated with COVID-19 hospitalization and COVID-19 death) for COVID-19 death and hospitalization were 1.32 (95% CI = 1.08–1.62) and 1.60 (95% CI = 1.44–1.78; Table 3 and Figure 1). Figure 1 shows the changes in hospitalizations and deaths with time during the study period. Steeper gradients on the graphs in Figure 1 correspond to increased rates of COVID-19 hospitalizations (Figure 1A) and deaths (Figure 1B) during the first and second COVID-19 “waves” in the UK during spring 2020 and winter 2020/2021. The alpha SARS-CoV-2 variant corresponds with the second wave. The delta variant was associated with a third wave in the UK in the summer of 2021, which is at the end of our study period and was associated with fewer deaths and hospitalizations in general.

The results were similar when comparing people with epilepsy to control cohort 2 (exact-matched for age, sex, and deprivation and propensity-matched for comorbidities increasing risk of COVID-19 hospitalization and death). HRs (epilepsy compared to control cohort 2) for COVID-19 death and COVID-19 hospitalization were 1.45 (95% CI = 1.22–1.74) and 1.66 (95% CI = 1.51–1.83), respectively.

4 DISCUSSION

In this large population-level study of more than 27 000 people with epilepsy, we found that people with epilepsy were more likely to have a COVID-19-related hospitalization or death when compared to controls during the first 15 months of the pandemic in Wales (crude HRs for hospitalization and death were 2.15 [95% CI = 1.94–2.37] and 2.15 [95% CI = 1.78–2.59]). This was true (but reduced) even when considering 40 comorbidities known to be associated with COVID-19 hospitalization and death (aHRs for hospitalization and death were 1.60 [95% CI = 1.44–1.78] and 1.33 [95% CI = 1.08–1.63]). This means there was an approximately 60% increased chance of hospitalization with COVID and a 33% increased chance of dying with COVID for people with epilepsy when compared to people of the same age, sex, area-based deprivation quintile, and comorbidities during the first 15 months of the COVID-19 pandemic in Wales.

In subgroups (adults aged 18–65 years, adults aged >65 years, people with intellectual disability, and people living in the two most deprived areas), people with epilepsy were more likely to have COVID-19 hospitalizations (Table 3 and Figure 2). There was no significant increase in COVID-19 hospitalizations for children with epilepsy after adjusting for comorbidities. However, the numbers of COVID-19 hospitalizations in children were very small, with only nine hospitalizations for children with epilepsy (15 hospitalizations in the controls).

Our results showed an association between epilepsy and an increased hospitalization rate compared to the matched comparator. This emphasizes that epilepsy might cause an increased risk of severe COVID-19, as with other chronic diseases.¹⁶ The neurotropic effects of Sars-CoV-2 might explain some of this increased risk for people with epilepsy, or epilepsy might be associated with alterations in the immune system, predisposing to more severe COVID-19.^{11, 12, 25} Another explanation is that the increased COVID-19 hospitalization rate reflects a lower threshold for hospitalization for people with epilepsy when presenting with COVID-19 due to perceived risk. COVID-19 may have increased seizures for people with epilepsy, which would have increased the risk of hospitalization and/or death. It is also possible that we have not accounted for a confounding comorbidity in our analysis that increases COVID-19 hospitalization risk and is more common in people with epilepsy.

We have used an established definition for COVID-19 hospitalizations (a hospital admission with a COVID-19 ICD-10 code for the reason for admission, or occurring within 28 days of a positive SARS-CoV-2 PCR test).¹⁶ It is possible that this definition included hospital admissions with COVID-19 but not because of it, for example, seizure- or epilepsy-related admissions.

When considering COVID-19 deaths in subgroups, only adults with epilepsy aged 18–65 years had an increased risk of COVID-19 death when adjusting for comorbidities (Table 3 and Figure 3). There were no COVID-19 deaths in children with epilepsy or children in the control group in our study.

It is established that people with intellectual disability have an increased risk of epilepsy, and almost 16% (4299) of the people with epilepsy in our study had an intellectual disability. Reassuringly, there did not seem to be a significantly increased risk of COVID-19 death in this group, but the numbers of COVID-19 deaths in people with intellectual disability were small. It was also reassuring that people with epilepsy in the most deprived two quintiles were not at increased risk of death when compared to controls.

5 CONCLUSIONS

People with epilepsy were at an increased risk of being hospitalized with, and dying from, COVID-19 when compared to age-, sex-, and deprivation-matched controls in Wales, even when adjusting for a large number of comorbidities. This may have implications for prioritizing future COVID-19 treatments and vaccinations for people with epilepsy.

AUTHOR CONTRIBUTIONS

Huw Strafford performed primary data and statistical analysis. William O. Pickrell and Richard F. Chin conceived, designed and coordinated the study. Michael P. Kerr, Richard F. Chin, Robert Powell, and Ronan A. Lyons provided senior clinical advice. Joe Hollinghurst and Alan Watkins provided statistical advice. Jan Paterson and Daniel Jennings provided public and patient involvement advice. All authors were part of the project team reviewing results and progress. Huw Strafford and William O. Pickrell drafted the manuscript. All authors reviewed and edited the manuscript.

CONFLICT OF INTEREST STATEMENT

M.P.K. has received speaker fees from UCB Pharma and Veriton and is vice chair of SUDEP ACTION. W.O.P. has received speaker fees from UCB Pharma and Angelini Pharma, travel grants from Angelini Pharma, and an unrestricted research grant from UCB Pharma. The authors confirm that they have no conflict of interests. We confirm that we have read the Journal's position on issues involved in ethical publication and affirm that this report is consistent with those guidelines.