The Association Between COVID-19 and Febrile Seizure

Background/Objective

Throughout the pandemic, febrile seizures have resulted from infection secondary to severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2). The objective of this study is to determine if there is an increased association between COVID-19 and febrile seizures as compared with other causes of febrile seizures.

Methods

This was a retrospective case control study. Data were collected from the National Institute of Health (NIH) supported National COVID Cohort Collaborative (N3C). Patients from 6 to 60 months who were tested for COVID-19 were included; cases were defined as COVID-19–positive patients whereas controls were defined as COVID-19–negative patients. Febrile seizures diagnosed within 48 hours of the COVID-19 test were considered to be associated with the test result. Patients were subjected to a stratified gender and date matching design followed by a logistic regression controlling for age and race.

Results

During the study period, 27,692 patients were included. Of those, 6923 patients were COVID-19–positive, among which 189 had febrile seizures (2.7%). After logistic regression, the likelihood of having febrile seizures concurrently with COVID-19 as compared with other causes was 0.96 (P = 0.949; confidence interval, 0.81, 1.14).

Conclusions

There were 2.7% of the patients with COVID-19 that were diagnosed with a febrile seizure. However, when subjected to a matched case control design with logistic regression controlling for confounding variables, there does not appear to be an increased risk of febrile seizures secondary to COVID-19 as compared with other causes.

Febrile seizures have long been a common presentation to emergency departments and are the most common form of childhood seizures, occurring in 2% to 5% of the population. Generally, these are defined as seizures that occur in a pediatric patient 6 months to 60 months of age in the setting of febrile illness and the absence of central nervous system infection. Febrile seizures are generally categorized into either simple or complex seizures. Simple febrile seizures are generalized or non-focal seizures, last less than 15 minutes, and do not recur in the proceeding 24 hours. Complex febrile seizures are any seizures that do not meet the above criteria.¹

The most common cause of febrile seizures includes viral illnesses, bacterial infections, and vaccinations, notably the combined MMR-V vaccine. One of the largest studies to date, the EFES study, was a prospective multicenter study that enrolled patients with febrile seizures and performed a respiratory multiplex array to determine which respiratory viruses were most associated with the condition. Results showed at least one respiratory virus was detected in nearly 83% of patients with febrile seizures, with adenovirus being the highest at 55%. Coronaviruses OC43 and 229 were also tested in the study and found to be the ninth and twelfth most common viruses (7% and 3.5%, respectively). These results, however, were published before the discovery of Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2).²

Early in the SARS-CoV-2 pandemic, case reports documented the association between COVID-19 infection and febrile seizures.^3,4 In addition, studies have examined the general frequency of febrile seizures in COVID-19 patients.⁵ Currently, there are no published studies determining whether there is an increased risk of febrile seizures with COVID-19 infections compared with other causes. The primary aim of this study was to determine whether there is an increased association between COVID-19 infection and febrile seizures as compared with patients who are COVID-19–negative.

METHODS

Study Design

We conducted a retrospective case-control evaluation of COVID-19 infected patients to assess the risk of febrile seizures compared with other causes. Secondary data were collected using the National Institute of Health (NIH) supported National COVID Cohort Collaborative (N3C) database, which is powered by the Palantir platform. The N3C is a partnership among 70 different healthcare institutions across the United States of America that have agreed to share and disseminate clinical data related to COVID-19. Data are obtained from outpatient, emergency department, and inpatient settings. Currently, there are 6.1 million COVID-19–positive patients registered within the database with over 18 billion rows of data available, which include, but are not limited to, laboratory data, vital signs, diagnoses, and mortality.⁶ The project protocol was submitted to our institutional review board. It was deemed not to be human subject research and did not require further review as the data set was deidentified. Our funding was only used to expedite our statistical analysis.

Study Setting, Population, and Protocol

The study population was designed within the enclave, the N3C's individualized secure platform, using SQL coding language. All pediatric patients between the ages of 6 and 60 months with a COVID-19 test performed (both rapid and PCR) from July 1, 2020, to December 31, 2021, were initially included in the study. If they had a past medical history of any of the following conditions they were excluded: epilepsy, cerebral palsy, CNS infection, ventricular shunt, intracranial mass. Concept sets to standardize working definitions and diagnoses were created within ATLAS. These were created for all the exclusion criteria and the diagnosis of febrile seizures (includes both simple and complex).⁷

A stratified matching design using gender and date was employed before statistical analysis. Any patients without a gender listed were excluded. After patients were divided into gender groups, those who tested positive for COVID-19 were cases and the date of their first positive test was designated as the case index date. Only the initial positive test was used in the analysis as each patient may have had multiple tests over the study period. Similarly, patients who tested negative for COVID-19 within the same timeframe (ie, same month and year) were controls, and the date of their first negative test was designated as the control index date. Only the initial index date was used to produce one data point per patient.

Outcome Measures

Diagnosis of febrile seizure was considered to be associated with either the case or control if it occurred 2 days before or 2 days after the index date. This was the primary outcome and was included in the final analysis.

Data Analysis

Within the above strata, each case was randomly matched to three controls (1:3 ratio). After gender and date matching were completed, the data were subjected to a multivariate logistic regression accounting for age and race. Age was viewed as a continuous variable. We did examine age separately as a categorical value, but it did not change the outcome. Thus, it was not included. The regression resulted in an odds ratio for febrile seizures and COVID-19. Furthermore, baseline characteristics were recorded. Febrile seizure types were also compared between COVID-19–positive and –negative patients.

RESULTS

Between July 1, 2020 and December 31^st 2021, 74,052 patients met both inclusion and exclusion criteria. It is important to note that these are not unique patients as some had multiple COVID test results over the time period. 133 patients were excluded as no gender was identified. Patients were then split into month-year groups for matching based on case and control index dates. Ultimately, there were 39 separate month-year and gender groupings.

After matching based on gender and index dates was performed in a 1:3 ratio between COVID-positive and -negative patients and subsequent encounters were removed, 27,692 unique patients were identified and included in the final analysis (Fig. 1). Of these, 12,632 (45.6%) were female and 15,060 (54.4%) were male. The average age of the patients in the study population was 28.4 months (Table 1).

FIGURE 1.

Stratified matching design by gender and COVID test result (month-year).

TABLE 1.

Baseline Demographic Information About the Population

Demographics	Overall (N = 27,692)	COVID (+) (n = 6923)	COVID (−) (n = 20,769)
Age: mean (SD). mo	28.4 (15.2)	27.2 (15.9)	28.7 (14.9)
Categorized age, y
[6M–12M]	4416 (16.0%)	1510 (21.8%)	2906 (14.0%)
(1Y–2Y]	8441 (30.5%)	1,975 (28.5%)	6466 (31.1%)
(2Y–3Y]	6179 (22.3%)	1346 (19.4%)	4833 (23.3%)
(3Y–4Y]	4650 (16.8%)	1060 (15.3%)	3590 (17.3%)
(4Y–5Y]	4006 (14.2%)	1032 (14.9%)	2974 (14.3%)
Gender
Female	12,632 (45.6%)	3158 (45.6%)	9474 (45.6%)
Male	15,060 (54.4%)	3765 (54.4%)	11,295 (54.4%)
Race
African American	6066 (21.9%)	1522 (22.0%)	4179 (20.1%)
White	14,789 (53.4%)	3589 (51.8%)	11,200 (53.9%)
Other	2105 (7.6%)	563 (8.1%)	1542 (7.4%)
Unknown	4732 (17.1%)	1249 (18.0%)	3483 (16.8%)
Ethnicity
Hispanic/Latino	5706 (20.6%)	1527 (22.1%)	4179 (20.1%)
Not Hispanic/Latino	18,568 (67.1%)	4579 (66.1%)	13,989 (67.4%)
Unknown	3418 (12.3%)	817 (11.8%)	2601 (12.5%)

Of the study population, 759 unique patients were diagnosed with febrile seizures. Notably, while a patient could have had multiple febrile seizure ICD-10 codes during the study period, only the initial index date was accounted for in the analysis. In some cases, there were multiple codes for febrile seizures during the 48-hour window around the initial index date. The 759 unique patients accounted for 836 febrile seizure diagnosis codes. One hundred eighty-nine of the 6932 COVID-19–positive patients had a febrile seizure within the window (2.7%), compared with 570 of the 20,769 COVID-19–negative patients diagnosed with febrile seizure (2.7%). Among COVID-19–positive patients, 83.2% were diagnosed with simple febrile seizure and 16.8% with complex febrile seizure. In comparison, 80% of COVID-19–negative patients were diagnosed with simple febrile seizure and 20% with complex febrile seizure.

For the logistic regression, age was viewed as a continuous variable. The odds ratio for febrile seizures based on age in one-month increments was 0.984 (P-value <0.001; CI 0.979, 0.989). African Americans and patients with race identified as “other” were both more likely to be diagnosed with febrile seizures than their White counterparts (OR 1.31, P-value 0.004 [CI 1.09, 1.55] and OR 1.47, P-value 0.003 [CI 1.14, 1.90] respectively). This is independent of COVID test result. The odds ratio for febrile seizures in patients testing positive for COVID was 0.96 (P-value 0.949 [CI 0.81, 1.14]) (Table 2).

TABLE 2.

Logistic Regression Using Age as a Continuous Variable

Characteristics	Reference	Adjusted OR (95% CI)	P
Age, mo		0.984 (0.979, 0.989)	<0.001
Race
African American	White	1.31 (1.09, 1.55)	0.004
Other	White	1.47 (1.14, 1.90)	0.003
Unknown	White	1.16 (0.94, 1.41)	0.161
African American	Other	0.88 (0.67, 1.16)	0.382
African American	Unknown	1.13 (0.90, 1.41)	0.298
Other	Unknown	1.27 (0.95, 1.70)	0.103
COVID
Positive	Negative	0.96 (0.81, 1.14)	0.949

DISCUSSION

The main objective of this multicenter retrospective case control study was to determine if there is an increased association between COVID-19 and diagnosis of febrile seizures versus other common causes of febrile seizures. Based on the logistic regression model, there was no statistically significant association. Thus, it can be inferred that patients with COVID-19 are as likely to be diagnosed with febrile seizures as patients diagnosed with other common respiratory illnesses. It is important to note that the controls in this study are patients that tested negative for COVID-19; based on this design, control patients could have had one of many diagnoses, both viral and bacterial, as the underlying etiology of their febrile seizures. In addition, the controls patients could have simply been screened for COVID-19 and not had active febrile illness. Interestingly, it was found that 189 of the 6,932 COVID-19–positive patients (approximately 2.7%) had febrile seizure within the 48-hour window. In addition, 83% of COVID-19–positive patients had simple febrile seizures.

Both percentages are higher compared with a similar retrospective study, which reported only 0.5% of COVID-19–positive patients being diagnosed with febrile seizures within a three-day window; 68% of those cases were simple febrile seizures. The sample size of COVID-19–positive patients is similar in both studies.⁵ We were unable to evaluate co-infection in our study. For that reason, the true incidence of febrile seizures solely secondary to COVID-19 patients may be somewhat lower. Referring to the prospective EFES study from 2016 in Turkey, coronavirus OC43 and 229 both were found to be associated with a higher proportion of complex febrile seizures.² It is important to mention that, compared with both of the previous studies, selection bias was a key limitation in our study design as only patients tested for COVID-19 were included as cases or controls. This is directly related to the design of the N3C database. However, as expected because of the pandemic, there was a high prevalence of COVID-19 testing during the study, which in turn, reduces the impact of selection bias. As for the distribution of simple versus complex febrile seizures in our study, it is possible that simple febrile seizure cases are overrepresented. Since the 48-hour window is based on COVID-19 test date and time, if a seizure occurs near the end of the window, it may only be recorded as simple. However, it is possible that another seizure occurred within 24 hours of the initial occurrence, which would then qualify as a complex febrile seizure and not have been appropriately accounted for.

Even though there was no increased association seen between COVID-19 and febrile seizure diagnoses, there are some other findings worth noting. Referring to Table 2, African Americans and patients categorized in the “other” racial group are more likely to be diagnosed with febrile seizure compared with their White counterparts, with odds ratios of 1.31 and 1.47, respectively. This contradicts findings from previous studies, where there was no difference among racial groups in hospitalized patients diagnosed with febrile seizures.^8,9 Though the reason for this disparity is unclear, it can be speculated that race may play an underlying role in affecting already established risk factors for febrile seizures, such as genetic predisposition. Alternatively, patients within this cohort may have been incorrectly diagnosed due to implicit bias.

Limitations

There are several limitations to the study, most of which are attributable to the inherent aspects of a retrospective study. Mainly, there are multiple confounders that cannot be controlled for. In addition, the N3C is a novel and constantly updating database. First, medical histories were only included in the Enclave if they were updated after January 2018. Thus, if a patient had a diagnosis before that date that had not been updated in their electronic medical record, it potentially could have been omitted. In addition, we were unable to consider the timing of a patient's medical condition in relation to COVID-19 testing or diagnosis of febrile seizures. For example, if a patient was diagnosed with a brain malignancy months or years after having COVID-19 with a febrile seizure diagnosis, he/she would have been erroneously excluded from the analysis. Second, due to confidentiality purposes, the ages of patients within the N3C database were randomly shifted by up to 6 months at the time of data collection. We chose patients aged 6 months to 5.5 years, which would have theoretically captured any patient from age 0 months to 6 years of age. Lastly, the Palantir platform did not recognize any medical terminology from the ICD03 classification system at the time of data analysis; due to this limitation, 1700+ concepts were dropped from the intracranial mass exclusion criteria concept set created on ATLAS when transferred to Palantir. Some examples of supported classification systems include SNOMED and Nebraska Lexicon. N3C support staff acknowledged the two previous limitations and are expected to remedy these soon.

CONCLUSIONS

The results of this study demonstrate that there is an association between COVID-19 and febrile seizures; 2.7% of COVID-19–positive patients had febrile seizures within a 48-hour window. However, when subjected to a gender-matched case control design with logistic regression to control for confounding variables, there does not appear to be an increased risk of febrile seizures secondary to COVID-19 when compared with other common causes of febrile seizures. The data gathered from this study can assist emergency department providers with evaluation of febrile seizures and the guidance given to families of COVID-19–positive patients. To further assess the risk of COVID-19 causing febrile seizures as compared with other causes a multicenter and prospective study is needed.