Midterm milestones:
In the first half of the project multiple milestones have been reached, per 1st of June 2024:
- A total of 2000 patients suspected of a CNS infection have been included in the Netherlands;
- Since Jan 2022 the study has been expanded to also include children with suspected encephalitis;
- A total of 15 hospitals now participate in IPACE, of which 7 hospitals include both adults and children;
- The first batch of 300 validation CSF samples from Denmark has been received in 2023;
- A total of 600 CSF and blood validation samples were received from Zambia in 2024.
Clinical Characteristics:
In the I-PACE cohort of patients with suspected encephalitis we analyzed clinical characteristics to see which patients have a poor outcome. We found that patients with autoimmune encephalitis have a worse prognosis because the diagnosis is often delayed and treatment therefore started late. So for new diagnostic tests it is important to focus on this population and facilitate early recognition and treatment. Furthemore, we focused on patients admitted to the hospital in whom there is a suspicion of a brain infection. There is often a discussion on whether a lumbar puncture is really needed in patients with fever of unknown origin to rule out a brain infection. Our data show clearly that a lumbar puncture changed the medical management in half of the patients.
Biomarkers:
We have studied the diagnostic value of both the proportion and absolute number of granulocytes (acute inflammatory cells) in the cerebrospinal fluid, in contrast to the monocytes (late reponding inflammatory cells). We found that the proportion of granulocytes in the cerebrospinal fluid is highest in patients with a bacterial cause of encephalitis. However, as high and low granuloyctes proportion was present in all diagnostic categories, its diagnostic value is not optimal. In comparison to other markers of (bacterial) infection of the brain, it is however one of the best biomarkers.
Furthemore, we studied novel biomarkers in neonates with sepsis and meningitis to differentiate the two clinical entities use cerebrospinal fluid. We found inflammatory markers IL-1RA, TNF-α, and CXCL-10 in CSF to enable good discrimination between the two disease entities. We will prospectively validate these finding in a cohort of neonates with suspected meningitis.
Prediction models:
Over 30 prediction models have been published in the literature, but these often lack validation and also frequently do not focus on the clinically relevant population - patients with a suspected infection of the brain. We used patient data from adults and children included the IPACE cohort to validate the diagnostic prediction rules and found that these models often fail to accurately predict the diagnosis and new biomarkers or AI models are needed to improve diagnostic prediction. We developed and validated our own prediction model for bacterial meningitis, and validated novel models recently published.
