Abstract
Objective
To determine the long‐term outcomes, including mortality and recurrent seizures, among children living with HIV (CLWH) who present with new onset seizure.
Methods
Zambian CLWH and new onset seizure were enrolled prospectively to determine the risk of and risk factors for recurrent seizures. Demographic data, clinical profiles, index seizure etiology, and 30‐day mortality outcomes were previously reported. After discharge, children were followed quarterly to identify recurrent seizures and death. Given the high risk of early death, risk factors for recurrent seizure were evaluated using a model that adjusted for mortality.
Results
Among 73 children enrolled, 28 died (38%), 22 within 30‐days of the index seizure. Median follow‐up was 533 days (IQR 18–957) with 5% (4/73) lost to follow‐up. Seizure recurrence was 19% among the entire cohort. Among children surviving at least 30‐days after the index seizure, 27% had a recurrent seizure. Median time from index seizure to recurrent seizure was 161 days (IQR 86–269). Central nervous system opportunistic infection (CNS OI), as the cause for the index seizure was protective against recurrent seizures and higher functional status was a risk factor for seizure recurrence.
Significance
Among CLWH presenting with new onset seizure, mortality risks remain elevated beyond the acute illness period. Recurrent seizures are common and are more likely in children with higher level of functioning even after adjusting for the outcome of death. Newer antiseizure medications appropriate for co‐usage with antiretroviral therapies are needed for the care of these children. CNS OI may represent a potentially reversible provocation for the index seizure, while seizures in high functioning CLWH without a CNS OI may be the result of a prior brain injury or susceptibility to seizures unrelated to HIV and thus represent an ongoing predisposition to seizures.
Plain Language Summary
This study followed CLWH who experienced a new onset seizure to find out how many go on to have more seizures and identify any patient characteristics associated with having more seizures. The study found that mortality rates continue to be high beyond the acute clinical presentation with new onset seizure. Children with a CNS OI causing the new onset seizure had a lower risk of later seizures, possibly because the trigger for the seizure can be treated. In contrast, high functioning children without a CNS OI were at higher risk of future seizures.
Keywords: central nervous system opportunistic infection, epilepsy, functional status, Lansky score, orphan, seizure recurrence
Key points.
Among CLWH with new onset seizure, mortality remains high even 30 days after the index seizure.
Seizure recurrence risks are highest for children with relatively normal functional status prior to the index seizure.
CLWH who experience new onset seizure due to a CNS OI are less likely to experience recurrent seizure.
Antiseizure medications that can be safely co‐administered with ARVs are desperately needed in the public sector.
1. INTRODUCTION
In 2022, an estimated 1.5 million children under 15 years were infected with human immunodeficiency virus (HIV). 1 Most children living with HIV (CLWH) reside in sub‐Saharan Africa (SSA). 2 Antiretroviral (ARV) therapies are now readily available globally and in SSA HIV has largely transitioned from a fatal infection to a common chronic condition. 3 Studies conducted in the era before highly active antiretroviral therapies were available found that half of CLWH had central nervous system (CNS) involvement with seizures constituting a common presentation. 4 , 5 , 6 , 7 , 8 Whether and when to initiate chronic antiseizure medications (ASMs) in a child in SSA with HIV remains a decision fraught with risks as drug interactions between most of the available ASMs and ARVs may lead to drug toxicities and/or ARV failure. 9 Established risk factors for the development of epilepsy after a first acute symptomatic seizure among HIV uninfected children in SSA include having a symptomatic seizure that was either focal or prolonged, pre‐existing neurodevelopmental delays and malaria as the underlying cause of the seizure. 10 , 11 , 12 , 13 A retrospective case–control study of children in an HIV clinic in South Africa identified risk factors for epilepsy that include bacterial and tubercular meningitis and stroke. 14 Work in Botswana has shown that earlier initiation of ARVs appears to be protective against epilepsy development among CLWH. 15
The cohort of HIV‐associated seizures and epilepsy (CHASE) study enrolled CLWH who presented with new onset seizure. CHASE captured HIV disease and treatment history, seizure semiology, seizure duration, and seizure etiology. Cohort characteristics and 30‐day mortality have previously been reported. 16 Herein, we detail long‐term outcomes from CHASE including mortality after 30 days, recurrent seizures and risk factors for recurrent seizures.
2. METHODS
As previously reported, 16 from April 2016 to June 2019, 73 CHASE children were enrolled from both urban and rural health facilities after presentation with new onset seizure. Demographic and clinical data were obtained at enrolment. Specifically, data were collected on HIV disease and treatment history, any history of previous central nervous system (CNS) insults including cerebral malaria, antenatal injuries or traumatic brain injury, index seizure semiology and severity, index seizure etiology based upon comprehensive review of all available acute clinical data, and functional status prior to the index seizure based upon a Lansky score derived from caregiver history. 17 Repeated seizures occurring during the same acute illness and/or within 30 days of the initial seizure were considered to represent additional acute symptomatic seizures and were not categorized as recurrent seizures or evidence of epilepsy.
2.1. Data collection
2.1.1. Longitudinal follow‐up
After discharge, caregivers were contacted quarterly by phone to administer a brief screening questionnaire to ascertain whether any recurrent seizures had occurred. Twice annually, children were seen in person and assessed with the same screening tool used via phone. If recurrent seizures were suspected and the family had not already sought care, support for transportation was provided and the research team assisted in facilitating a clinical evaluation. For all research visits, evaluations were coordinated to coincide with the child's HIV clinic visits and transportation reimbursement for a parent and the child round trip from their home was provided. Data from HIV clinical visits and any hospital admissions were also reviewed. The primary outcomes of interest were seizure recurrence and death.
2.2. Analysis
Because death was an important competing outcome, adjusted, and unadjusted Cox models were developed to identify potential risk factors for death with p‐values determined from the log rank test comparing the equivalence of the Kaplan–Meier curves except for continuous variables which were compared via quartiles. Potential risk factors examined included age, sex, any history of previous CNS injury, index seizure focality, worst World Health Organization (WHO) HIV stage, index seizure duration, seizure etiology (CNS OI vs other), mother's marital status, and functional status prior to the index seizure via Lansky Score. A Fine‐Gray model was fit to the follow‐up data to evaluate potential risk factors for seizure recurrence adjusting for the important competing outcome of death. The same variables included in the Cox mortality model were evaluated in the Fine‐Gray model for seizure recurrence. Age was categorized into those under 2 years, 2–10 years, and >10 years to capture any differences in risk among the very young relative to mid‐range or older children/teens. Subdistribution hazard ratios with 95% confidence intervals and p‐values were calculated and presented as survival curves stratified by patient characteristics identified as risk factors in the model. For the 4 children lost to follow‐up, outcomes data available through the duration of their abbreviated follow‐up period were considered the final outcome.
2.3. Ethical approval
Ethical approval was obtained from the University of Zambia's Biomedical Research Ethics Committee and the University of Rochester's Research Subjects Review Board. Additional approval was obtained from the National Health Research Authority of Zambia. Written parent or guardian consent was obtained prior to initial enrolment and where appropriate, assent was obtained from the children being followed.
3. RESULTS
Seventy‐three children were enrolled. As previously reported, 16 most had advanced, often untreated HIV. Mean age was 6 years (interquartile range 2.2–10.0), 39 (53%) were male, with infectious etiologies being responsible for over half of the seizures. Children were followed for a median 533 days (IQR 18–957) after their index seizure. See the study flow diagram (Figure 1) for details regarding study activities from participant screening through to the outcomes of recurrent seizure and/or death. Median time from index seizure to death was 16 days (IQR 5–26) and from index seizure to recurrent seizure was 161 days (IQR 86–269). Overall, 28 (38%) children died, 22 within 30‐days of the index seizure. 16 A recurrent seizure occurred in 14 children—19% (14/73) of the overall cohort and 27% (14/51) among those children who survived at least 30 days after their index seizure. One child with seizure recurrence subsequently died several months after the index and recurrent seizures. There was 5% (4/73) lost to follow‐up.
FIGURE 1.
Study flow diagram.
Recurrent seizure risk factors adjusted for the outcome of death using the Fine‐Gray Model are provided in Table 1. Since 0/15 children with CNS OI as the cause for their seizure developed epilepsy, this could not be included in the model. Age, gender, past CNS injury, index seizure focality, index seizure duration, the mother's marital status, and HIV disease severity were not risk factors for recurrent seizures. A higher Lansky score pre‐index seizure, a measure of functional status or performance‐play, was a risk factor for recurrent seizures HR 1.03 (95%CI 1.01–1.05; p = 0.002). See Figure 2 for a cumulative incidence plot depicting time to recurrent seizure stratified by CNS OI as the index seizure etiology.
TABLE 1.
Fine‐gray analysis to identify risk factors for seizure recurrence adjusting for the competing outcome of death (n = 73 with 15 recurrent seizures).
| Subdistribution Hazard ratio | 95% confidence interval | p‐value | |
|---|---|---|---|
| Age | |||
| Less than 2 years (referent) | – | – | – |
| 2 to 10 years | 1.40 | 0.34–5.78 | 0.64 |
| More than 10 years | 0.69 | 0.10–4.76 | 0.71 |
| Female vs male | 1.43 | 0.37–5.53 | 0.60 |
| Previous Central Nervous System (CNS) injury vs no | 2.08 | 0.29–14.98 | 0.47 |
| Focal index seizure vs no | 1.64 | 0.23–11.86 | 0.62 |
| Worst WHO HIV stage | |||
| Stage 1–2 (referent) | – | – | – |
| Stage 3 | 2.55 | 0.31–20.73 | 0.38 |
| Stage 4 | 1.54 | 0.49–4.86 | 0.47 |
| Seizure severity | |||
| Brief (referent) | – | – | – |
| Prolonged (>5, <30 min) | 1.94 | 0.25–15.02 | 0.52 |
| Status epilepticus (≥30 min) | 2.26 | 0.38–13.42 | 0.37 |
| Mother's marital Status | |||
| Married (referent) | – | – | – |
| Divorced | 0.27 | 0.02–3.89 | 0.34 |
| Widowed | 1.35 | 0.16–11.22 | 0.78 |
| Not documented, not in mother's care | 0.96 | 0.19–4.86 | 0.96 |
| Lansky Score (by quartiles) | 1.03 | 1.01–1.05 | 0.002 |
Note: Of 15 children with seizure recurrence, none had a CNS opportunistic infection (OI). Therefore, the risk of seizure recurrence with CNS OI is 0 and the relative risk compared to risk of seizure recurrence without CNS OI is also 0. In this case, the parameter estimate lies on the boundary and the statistical software used to fit the Fine‐Gray model failed to converge for any Fine‐Gray model including CNS OI as an independent variable.
Bolded values were statistically signficant at p < 0.05.
FIGURE 2.
Kaplan–Meier Curve depicting time to epilepsy onset stratified by Index seizure etiology being CNS opportunistic infection vs other cause.
The same characteristics reported in Table 1 as potential risk factors for seizure recurrence are reported as risk factors for death in Table 2. In the unadjusted model, age 2–10 years HR 0.4 (95%CI 0.0.2–0.98; p = 0.045) and having better functional status prior to the index seizure HR 0.99 (95%CI 0.97–0.999; p = 0.03) were protective against death. Children with the worst HIV WHO clinical disease stage were more likely to die HR 4.4 (95%CI 1.03–18.7; p = 0.045). Mortality was also more likely in children whose mothers were widowed or who were in the care of someone other than a parent–HR 4.2 (95%CI 1.2–14.4; p = 0.02) and HR 3.4 (95%CI 1.2–9.6; p = 0.02), respectively. In a univariate Cox model, CNS OI seizure etiology was not associated with death HR 1.8 (95%CI 0.8–4.3; p = 0.18). Younger age, having a widowed mother, and having poorer pre‐seizure functional status remained significant in the adjusted model.
TABLE 2.
Cox models to identify risk factors for death (n = 73 with 28 deaths).
| Unadjusted | Adjusted | |||||
|---|---|---|---|---|---|---|
| Hazard ratio | 95% confidence interval | p‐value | Hazard ratio | 95% confidence interval | p‐value | |
| Age | ||||||
| Less than 2 years (referent) | – | – | – | – | – | – |
| 2 to 10 years | 0.4 | 0.2–0.98 | 0.045 | 0.2 | 0.1–0.8 | 0.02 |
| More than 10 years | 1.0 | 0.4–2.5 | 0.99 | 0.9 | 0.3–2.6 | 0.87 |
| Female vs male | 0.6 | 0.3–1.3 | 0.16 | 0.5 | 0.2–1.3 | 0.15 |
| Previous Central Nervous System (CNS) injury vs no | 1.2 | 0.4–4.1 | 0.72 | 0.8 | 0.2–3.4 | 0.80 |
| Focal index seizure vs no | 0.7 | 0.3–1.5 | 0.37 | 0.8 | 0.3–2.3 | 0.74 |
| Worst World Health Organization HIV stage | ||||||
| Stage 1 or 2 (referent) | – | – | – | – | – | – |
| Stage 3 | 2.6 | 0.4–15.8 | 0.29 | 0.7 | 0.1–6.2 | 0.76 |
| Stage 4 | 4.4 | 1.03–18.7 | 0.045 | 1.9 | 0.4–10.0 | 0.45 |
| Seizure severity | ||||||
| Brief (referent) | – | – | – | – | – | – |
| Prolonged | 0.5 | 0.2–1.2 | 0.11 | 0.6 | 0.2–1.8 | 0.38 |
| Status epilepticus | 0.6 | 0.3–1.6 | 0.32 | 0.6 | 0.2–1.8 | 0.42 |
| Mother's marital status | ||||||
| Married (referent) | – | – | – | – | – | – |
| Divorced | 2.0 | 0.5–7.5 | 0.29 | 3.9 | 0.8–19.4 | 0.09 |
| Widowed | 4.2 | 1.2–14.4 | 0.02 | 5.0 | 1.1–21.8 | 0.03 |
| Not documented, not in mother's care | 3.4 | 1.2–9.6 | 0.02 | 2.8 | 0.9–8.8 | 0.08 |
| Lansky Score (by quartiles) | 0.99 | 0.97–0.999 | 0.03 | 0.98 | 0.96–0.997 | 0.02 |
Bolded values were statistically signficant at p < 0.05.
4. DISCUSSION
Antiretroviral therapies for HIV treatment began nationwide rollout in Zambia over 20 years ago and prevention of mother to child transmission (PMCTC) programs are part of routine pre‐and antenatal care. Despite this, children in the CHASE study presented with new onset seizure almost exclusively in the context of advanced, untreated or erratically treated HIV disease. Zambia may not be unique—in 2023, the United Nations Children's Fund reported that only 57% of CLWH globally were receiving treatment, far short of the 90–90–90 goal. 18
CHASE included rural children at district‐level hospitals and children from an urban academic center, but this study provides little insight into the long‐term outcomes of children with well‐controlled HIV who experience a seizure. Although extensive data were collected regarding pre‐index seizure health status, limited information was available regarding the management of the acute index seizure. Documentation of seizure duration prior to treatment initiation and timing of treatments was poor. This is not unexpected in a low‐income country setting, but does prevent the evaluation of the acute seizure management strategies on long‐term outcomes. Acute seizure treatment options were similar across all study sites with phenobarbital and benzodiazepines being the available agents and little access to ventilatory support should it be needed. It is not possible to know what role, if any, these limited resources played in the high early mortality. Finally, ascertainment of recurrent seizures was largely based upon family report so recurrent seizures that were partial in nature may not have been recognized.
In this prospective cohort study, death was more likely with advanced HIV disease, poorer pre‐seizure functional status, and being an orphan. Being orphaned may be a marker for more intense intraurine HIV exposure and is also associated with poverty and social chaos. Long‐term follow‐up of children in the CHASE study clearly demonstrated high mortality rates extending beyond the first 30‐days previously reported with 11% of 30‐day survivors subsequently succumbing to their disease. Recurrent seizures were extremely common occurring in 27% of children who survived at least 30 days after their index seizure. Having a CNS OI as the cause for the index seizure was protective against subsequent seizures while children with higher functional status who were less impaired before their index seizure were more likely to experience subsequent seizures. CNS OIs represent a potentially reversible cause for the index seizure meaning the index seizure had a provocation amenable to treatment (i.e. treatment of their underlying immune disorder as well as identified infections). In contrast, the index seizure in higher functioning children may reflect a predisposition to seizures unrelated to their HIV status or one related to an earlier HIV‐related brain injury and thus reflect an irreversible, ongoing risk for future seizures None of the 15 children with CNS OIs as the underlying cause for their index seizure had recurrent seizures. This supports focusing upon aggressive management of HIV disease in these children with a watchful waiting approach to ASM initiation. After adjusting for mortality, higher functioning children were at increased risk for recurrent seizures. This suggests that CLWH who are otherwise relatively well when they experience a new onset seizure with no clearly reversible trigger, such as a CNS OI, are likely (>27%) to go on to have additional seizures within a fairly short period of time and would likely benefit from ASM initiation.
Many CLWH who experience recurrent seizures will require chronic treatment with ASMs and medication selection should take into consideration the life‐saving, life‐long ARVs they also require. Unfortunately, the mainstay of ASM available in the public sector in Zambia and many other countries in SSA continues to be phenobarbital and carbamazepine with only very constrained access to valproate outside of tertiary care settings. The co‐usage of enzyme‐inducing ASM and HIV therapies is fraught with hazards. Guidelines over a decade old recommend against the co‐usage of enzyme‐inducing ASMs with ARVS, 19 but few other treatment are available in most Zambian care setting. Access to more optimal ASM options for CLWH are desperately needed.
CONFLICT OF INTEREST STATEMENT
David Bearden served as a paid advisor for Biogen. The remaining authors have no conflict of interest. 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.
ACKNOWLEDGMENTS
This work was funded by the United States National Institute of Health's (US NIH) National Institute of Neurological Disorders and Stroke R01NS094037, R35NS122265 and K23NS117310. The content is solely the responsibility of the authors and does not necessarily represent that of the US NIH.
Birbeck GL, Mwenechanya M, Ume‐Ezeoke I, Mathews M, Bositis CM, Kalungwana L, et al. Long‐term outcomes after new onset seizure in children living with HIV: A cohort study. Epilepsia Open. 2024;9:750–757. 10.1002/epi4.12921
REFERENCES
- 1. UNAIDS . Global HIV and AIDS Statistics Facts Sheet, 2023. Available from: https://www.unaids.org/en/resources/fact‐sheet. Accessed 14 February 2024.
- 2. UNICEF . Sixth stocktaking report: towards an AIDS‐free generation. New York, NY: UNICEF; 2013. [Google Scholar]
- 3. Bradshaw D, Groenewald P, Laubscher R, Nannan N, Nojilana B, Norman R, et al. Initial burden of disease estimates for South Africa, 2000. S Afr Med J. 2003;93:682–688. [PubMed] [Google Scholar]
- 4. Govender R, Eley B, Walker K, Petersen R, Wilmshurst JM. Neurologic and neurobehavioral sequelae in children with human immunodeficiency virus (HIV‐1) infection. J Child Neurol. 2011;26:1355–1364. [DOI] [PubMed] [Google Scholar]
- 5. Kellinghaus C, Engbring C, Kovac S, Moddel G, Boesebeck F, Fischera M, et al. Frequency of seizures and epilepsy in neurological HIV‐infected patients. Seizure. 2008;17:27–33. [DOI] [PubMed] [Google Scholar]
- 6. Power C, Boisse L, Rourke S, Gill MJ. NeuroAIDS: an evolving epidemic. Can J Neurol Sci. 2009;36:285–295. [DOI] [PubMed] [Google Scholar]
- 7. Sinha S, Satishchandra P, Nalini A, Ravi V, Subbakrishna DK, Jayakumar PN, et al. New‐onset seizures among HIV infected drug naive patients from south India. Neurol Asia. 2005;10:29–33. [Google Scholar]
- 8. Tellechea‐Rotta N, Legido A. Acquired immunodeficiency syndrome by vertical transmission: neurological disorders. Rev Neurol. 2003;1–15:255–263. [PubMed] [Google Scholar]
- 9. French JA, Kanner AM, Bautista J, Abou‐Khalil B, Browne T, Harden CL, et al. Efficacy and tolerability of the new antiepileptic drugs, II: treatment of refractory epilepsy: report of the TTA and QSS subcommittees of the American Academy of Neurology and the American Epilepsy Society Epilepsia. 2004;45:410–423. [DOI] [PubMed] [Google Scholar]
- 10. Ba‐Diop A, Marin B, Druet‐Cabanac M, Ngoungou EB, Newton CR, Preux PM. Epidemiology, causes, and treatment of epilepsy in sub‐Saharan Africa. Lancet Neurol. 2014;13:1029–1044. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11. Kariuki SM, Rockett K, Clark TG, Reyburn H, Agbenyega T, Taylor TE, et al. The genetic risk of acute seizures in African children with falciparum malaria. Epilepsia. 2013;54:990–1001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Ngugi AK, Bottomley C, Kleinschmidt I, Wagner RG, Kakooza‐Mwesige A, Ae‐Ngibise K, et al. Prevalence of active convulsive epilepsy in sub‐Saharan Africa and associated risk factors: cross‐sectional and case‐control studies. Lancet Neurol. 2013;12:253–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13. Chomba E, Taylor TE, Hauser WA, Wasterlain C, Organek N, Birbeck GL. Seizure recurrence in rural Zambian children admitted with febrile seizures. Open J Trop Med. 2008;1:101–107. [Google Scholar]
- 14. Burman RJ, Wilmshurst JM, Gebauer S, Weise L, Walker KG, Donald KA. Seizures in children with HIV infection in South Africa: a retrospective case control study. Seizure. 2019;65:159–165. [DOI] [PubMed] [Google Scholar]
- 15. Bearden D, Steenhoff AP, Dlugos DJ, Kolson D, Mehta P, Kessler S, et al. Early antiretroviral therapy is protective against epilepsy in children with human immunodeficiency virus infection in Botswana. J Acquir Immune Defic Syndr. 2015;69:193–199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Ravishankar M, Dallah I, Mathews M, Bositis CM, Mwenechanya M, Kalungwana‐Mambwe L, et al. Clinical characteristics and outcomes after new‐onset seizure among Zambian children with HIV during the antiretroviral therapy era. Epilepsia Open. 2022;7:315–324. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Lansky SB, List MA, Lansky LL, Ritter‐Sterr C, Miller DR. The measurement of performance in childhood cancer patients. Cancer. 1987;60:1651–1656. [DOI] [PubMed] [Google Scholar]
- 18. UNICEF . Pediatric Care and treatment. Available from: https://data.unicef.org/topic/hivaids/paediatric‐treatment‐and‐care/. Accessed 15 January 2024.
- 19. Birbeck GL, French JA, Perucca E, Simpson DM, Fraimow H, George JM, et al. Ad Hoc Task Force Of The Commission On Therapeutic Strategies Of The International League Against Epilepsy Antiepileptic drug selection for people with HIV/AIDS: evidence‐based guidelines from the ILAE and AAN. Epilepsia. 2012;53:207–214. [DOI] [PubMed] [Google Scholar]