Improving diagnosis and management of acute brain infections in low and middle-income countries

SUMMARY

Brain infections pose a threat to public health globally. They carry a high burden of mortality and morbidity, with cognitive, physical and neurological sequelae. This is particularly true for low- and middle-income countries (LMICs), which experience large-scale disruptive epidemics but have fewer resources related to diagnosis, management, and aftercare.

The WHO has prepared a global roadmap to defeat meningitis by 2030. In this context, the study by Singh et al.¹, published in The Lancet, is a welcome addition to the literature, particularly from the perspective of less developed regions.

This study aimed to design, implement, and evaluate a tailored, multifaceted intervention in hospitals in Brazil, India, and Malawi to improve the diagnosis and early management of patients with suspected acute brain infections. This was a multicentre before-and-after intervention study conducted at 13 hospital sites, linked to 4 study centres in these 3 countries. The hospitals and countries were chosen to encompass a diverse background of patients from the economic (upper middle, lower middle and low-income countries), geographic (urban, rural and semi-rural), set up (public hospital and private non-profit institutions) and level of care (secondary, tertiary and exclusive specialist referral services) perspective, to improve generalizability of results.

In the pre-intervention phase, the authors designed a multicomponent clinical and laboratory intervention involving hospital stakeholders, policymakers, and patient and public representatives. The intervention was determined after evaluation of routine practice and was tailored for each setting. Detailed real-time analysis of patient journeys (from presentation to discharge) and cerebrospinal fluid (CSF) samples (from collection to analysis) was performed on at least 8 patients per hospital. Important subgroups, based on age, gender, human immunodeficiency virus (HIV) status, timing, and status at presentation to the hospital, were included. A specialist team assessed the hospital laboratory capacity to diagnose brain infections using bespoke tools, based on laboratory and diagnostic checklists issued by the WHO. Based on a modified framework approach, the investigators identified facilitators and challenges to optimal care. The integrated results were presented to all stakeholders to plan the design of the intervention.

The team introduced 3 core intervention components: a diagnostic and management algorithm to provide decision support, a lumbar puncture pack to facilitate optimal CSF collection, transport, and storage of CSF specimens, and a pathogen testing panel based on local epidemiology and laboratory capacity. The microbiological panel used a stepwise approach (prioritized treatable and common pathogens), and the intervention team provided real-time support for clinical interpretation. Staff were also trained and mentored. The incremental cost to the healthcare system was calculated.

The study was conducted between January 2021 and November 2022. Of 10 462 patients screened, the study analyzed 2154 patients with suspected acute brain infection (meningitis, encephalitis or both). Of these, 1330 (62%) were recruited before the intervention and 824 (38%) were recruited after the intervention. 1448 patients (67%) were enrolled in two centres in India (Bengaluru and Vellore). The median age across centres was 23 years (interquartile range 6–44), with 59% being males and 55% residing in an urban area. Adults aged 16 years or older constituted 59%, the rest were children aged between 29 days and 15 years. 10% had an HIV infection at presentation.

The study examined co-primary outcomes: the percentage of patients achieving a syndromic diagnosis and the percentage achieving a microbiological diagnosis––before and after the intervention. The primary outcome was available for 2154 patients. A syndromic diagnosis was achieved in 77% cases before the intervention and 86% after the intervention. A microbiological diagnosis was achieved in 22% of cases before the intervention and in 30% after the intervention. To account for trends of improvement over time, the authors performed an interrupted time series analysis and reported that the post-intervention increase in the percentage receiving a diagnosis was greater than the underlying trend of improvement over time. They also adjusted the results for potential confounders. Results were similar when primary outcomes were analysed for age-based subgroups.

As far as the secondary outcomes were concerned, the authors reported improvement in the percentage of patients having lumbar puncture (10% increase), receiving empirical antimicrobial therapy within a day of admission (9% increase) and time to lumbar puncture (reduced from a median of 13 to 9 hours), post-intervention. The improvement in functional outcomes was modest––the percentage of patients with a good outcome at follow-up (30 days) increased from 59% pre-intervention to 67% post-intervention (the study was not powered to assess this outcome). The study reported that 11 patients should receive the intervention to achieve one additional syndromic diagnosis. The average cost of the intervention per patient varied with the centre (for India between US$ 112 and 291). There was no significant change in the mortality and quality of life.

Although the study did not achieve the designated sample size (the target was 450 patients pre-intervention and 450 patients post-intervention in each centre), it aimed to demonstrate that improvements in brain infection diagnosis and management in limited-resource settings could be achieved through simple optimization of routine hospital care. The authors have reported the most important outcomes, including functional scores, quality of life, mortality, the number needed to treat, and the incremental cost to the healthcare system of each resource used (human, capital, and material). According to the authors, a key attribute that contributed to the success of the intervention was the involvement of local clinicians and administrators. Furthermore, the components of the intervention were simple, structured and algorithm-based, making implementation easier and effective. Training and orientation to the intervention ensured a successful and uniform delivery. The introduction or systematic use of polymerase chain reaction tests for various pathogens also contributed to improvements in the number of diagnoses achieved, as several pathogens were detected more frequently or diagnosed only post-intervention, such as Orientia tsutsugamushi, Streptococcus pneumoniae, Mycobacterium tuberculosis, Enteroviruses, and Chikungunya. The microbiological panel tested common pathogens first (stepwise approach).

The study’s results, although encouraging, should be interpreted with a few caveats in mind. Most of the reported point estimates had wide confidence intervals, raising concerns about the precision and accuracy of results. The graph showing the actual (after intervention) and predicted (if there had been no intervention) trends of diagnosis over time showed considerable overlap in the 95% confidence intervals. Furthermore, the authors did not account for multiple primary outcomes in their sample size or power calculation (using only one primary outcome: ‘microbiological diagnosis’ in the power calculation) or in the result interpretation.

COMMENTS

This study is an attempt to develop evidence for tailored, multifaceted, structured algorithms for the diagnosis and management of brain infections in LMICs, which bear the major brunt of communicable diseases.¹ In 2019, meningitis contributed to 2.5 million cases and 236 000 deaths worldwide.² India alone reported an estimated 610 000 cases of encephalitis (51 900 deaths) and 552 000 cases of meningitis (34 700 deaths), with 1.1% of the total disability adjusted life-years attributable to brain infections.³ Improved diagnosis and optimal management of brain infections are thus a focus of initiatives by the WHO.

Meningitis and encephalitis lead to long-term complications in the form of physical and cognitive sequelae in approximately one in five individuals.⁴ This not only impacts the quality of life but also imposes a major financial burden on individuals, families and the community. The cost of long-term care can be overwhelming in low-resource settings, where people are forced to sacrifice necessities to cover medical expenses. The phrase ‘Time is Brain’, though in common parlance is used in stroke management, also ticks the column in brain infections, particularly acute bacterial meningitis. Delays in lumbar puncture (the classic triad of fever, headache and meningismus may be absent)⁵, in sending the right analysis and in starting appropriate antimicrobial therapy can be devastating as the clinical picture may complicate to cause cranial nerve palsies, hydrocephalus, vasculitic infarcts, brain abscess, subdural empyema, cerebral herniation, septic shock, multiorgan failure and death.

The WHO, in its latest guidelines on the treatment of meningitis, gives a strong recommendation that children and adults with suspected acute meningitis should receive intravenous antimicrobial treatment as early as possible.⁶ A‘one hour window’ post-admission is considered the golden hour for starting treatment for medical emergencies, including bacterial meningitis.⁷ It is good practice to do a lumbar puncture as soon as possible, preferably before starting antimicrobials, unless there are specific contraindications. However, treatment should not be delayed for brain imaging or when lumbar puncture is deferred for some reason. Guidelines also recommend cranial imaging before CSF, if the Glasgow Coma Score is below 10, there are focal neurological signs, cranial nerve deficits, oedema, new-onset seizures (in adults) and/or a severe immunocompromised state.⁶

Although guidelines are in place, the reality on the ground in LMICs is far from perfect. There is a considerable gap in achieving an adequate and timely microbiological diagnosis, as well as limited availability of molecular tests and a lack of standard operating protocols for diagnosis, treatment, and follow-up. This not only leads to worse functional outcomes but also contributes to the development of antimicrobial resistance.

Considering this background, this study conducted in an LMIC setting highlights some important points that can inform and guide the effective formulation and implementation of region-specific guidelines for the diagnosis and management of brain infections. First, multicomponent interventions targeting brain infections are feasible and may potentially improve diagnostic rates and the time to administration of antimicrobial therapy. Second, the importance of tailoring the intervention to local needs and including key stakeholders in the formulation process cannot be undermined, as the ultimate goal is to develop a sustainable approach. Third, an algorithm-based approach, structured protocols for sample collection, transport, and laboratory capacity, access to panel-based pathogen tests (guided by local epidemiology), and capacity building of healthcare personnel are important elements of the intervention.

Future studies in LMICs could consider additional pointers, such as using functional performance as the primary outcome, capacity building in primary care settings (including syndromic diagnosis, timely referral, and initial resuscitation), and the cost-effectiveness of the approach. Progress is expected on this front, as guideline development is already underway for hospital clinicians diagnosing and treating patients with suspected brain infections in India.⁸

Readers can download the Brain Infections Global Diagnostic Toolkit, which includes region-specific clinical algorithms, guidance on lumbar puncture analysis, and a stepwise approach to pathogen testing, from the Brain Infections Global website (available for free).