Febrile coma in Malawian children

Ray, Stephen (2022) Febrile coma in Malawian children. PhD thesis, University of Liverpool.

Text 200224800_Feb2022.pdf - Unspecified Access to this file is embargoed until 1 July 2024. Download (38MB)

Abstract

Background Fever and altered consciousness (febrile encephalopathy) in children is a common presentation in sub-Saharan Africa. Historically, most cases were attributed to cerebral malaria (CM), but with the recent drastic reduction in malaria incidence, non-malarial coma is a larger proportion, and determining the aetiologies is often diagnostically challenging, particularly in resource-limited settings. There is a paucity of rigorous epidemiological data on the aetiology and outcomes of febrile encephalopathy in African children, despite the high burden of poor clinical outcomes associated with this presentation. I led a pan-African systematic review and meta-analysis and a large prospective observational study in Malawi to determine robust epidemiological estimates of the pathogenic causes and long-term neurological consequence of febrile coma in African children, to direct further research to improve outcomes. Methods Firstly, I systematically reviewed studies of non-traumatic coma among paediatric populations across Africa. Disease-specific studies were included if outcomes were reported. I calculated pooled estimates of the prevalence of individual aetiologies plus overall and disease specific estimates of mortality and morbidity. Secondly, I led a large prospective cohort study, in Blantyre, Malawi, that recruited between January 2018 - June 2021. It included febrile children aged 3 months to 14 years in deep coma (Blantyre coma score ≤2). I investigated aetiology (qPCR on blood and CSF) and performed MRI brain and EEG. Detailed functional and neuro-developmental outcome was assessed at 6 months post-discharge. Results For the systematic review and meta-analysis, 127 studies of 31,233 children from 29 African countries reported data on aetiology and/or outcome of non-traumatic encephalopathy. A wide spectrum of underlying aetiologies of non-traumatic encephalopathy were identified from 15 studies. Cerebral malaria was most prevalent (56%), followed by unknown encephalopathy (29%), acute bacterial meningitis (ABM) (10%), sepsis (8%), viral CNS infection (5%), non-infectious causes (4%), encephalitis (3%) and toxic/metabolic aetiologies (1%). Pooled mortality for grouped non-traumatic encephalopathy was high at 29% (12 studies) but by individual disease these estimates varied. Cerebral malaria mortality was 17% (n=77 studies), while ABM mortality was 48% (12 studies). Pooled disability for non-traumatic encephalopathy was infrequently reported. For the prospective cohort study, I recruited 352 participants: 231 cases of cerebral malaria (CM) (65.6%) and 121 non-malarial cases (34.3%). Within the CM group, I identified a co-infection in 63 children (27.2%) [CNS n=27 (11.6%), bloodstream n=36 (15.6 %)] and a novel phenotype of post parasitaemic cerebral malaria coma (n=14). Non-malarial aetiologies included acute bacterial meningitis n=48 (13.6%), encephalitis n=24 [immune-mediated n=4] (6.8%), septic encephalopathy n=13 (3.6%) and unknown encephalopathy n=32 (9.1%). A total of 93 pathogens were detected by qPCR methods in the entire cohort. 54 pathogens were identified to support a diagnosis of CNS infection [40 CSF PCR, 14 blood PCR], while 39 supported a diagnosis of a bloodstream infection [all 39 blood PCR]. Streptococcus pneumoniae (n=44), followed by Salmonella Spp (n=22) [S.typhi and non-typhoidal Salmonella combined] were the most frequently identified pathogens in both blood and CSF. Other pathogens detected by qPCR included Klebsiella Spp, E.coli, Mycobacterium tuberculosis, Enterovirus, CMV, HSV1, HSV2, VZV and SARS-CoV-2. Molecular methods and MRI neuroimaging facilitated a syndromic diagnosis in 320/352 (90.9%) and a pathogenic diagnosis in 289/352 (82.1%). Mortality was significantly higher in the non-malarial group at 6 months (28% vs 17.7%, p= 0.03), as was developmental delay (45% vs 16.8% p=0.000) and severe functional neurological impairment (16.7 vs 7.2 p=0.01). A multivariable logistic regression model identified that a deeper level of coma, tachycardia, deep breathing, hyperlactatemia, raised CSF opening pressure, abnormal CSF profile and a diagnosis of cerebral malaria and meningitis dual infection or sepsis were significantly associated with death. Conclusions Malaria remains the most common cause of coma in Malawi, however non-malarial comas had a higher morbidity and mortality burden. Rigorous molecular and radiological diagnostics identified a wide spectrum of parasitic, viral and bacterial aetiologies. I identified in the largest cohort to date that many bacterial or viral infections found were in those originally diagnosed with cerebral malaria. I identified the novel syndrome of post parasitaemic cerebral malaria. Pathogens identified included rarer CNS causes of coma e.g. neurocysticercosis and SARS-CoV-2. My study has highlighted that febrile encephalopathy in sSA presents in critical illness and has catastrophic outcomes. Children were often severely dehydrated but also with severe brain swelling and injury. What remains is a lack of strong evidence base on how to optimally treat acidosis and dehydration with resuscitative fluids, without creating additional brain injury to an already swollen brain, to improve outcomes.

Item Type:	Thesis (PhD)
Divisions:	Faculty of Health and Life Sciences Faculty of Health and Life Sciences > Clinical Directorate
Depositing User:	Symplectic Admin
Date Deposited:	16 Dec 2022 12:03
Last Modified:	18 Jan 2023 20:47
DOI:	10.17638/03162065
Supervisors:	Griffiths, Mike Seydel, Karl Lalloo, David ORCID: 0000-0001-7680-2200 Cornick, Jennifer
URI:	https://livrepository.liverpool.ac.uk/id/eprint/3162065