Cerebrospinal fluid haemoglobin as a monitoring biomarker for secondary brain injury after aneurysmal subarachnoid haemorrhage: prospective international multicentre validation study (HeMoVal)
Kevin Akeret, Raphael Matthias Buzzi, Thomas Gentinetta, Moritz Saxenhofer, David Kronthaler, Elisa Colombo, Alexandra Grob, Bart Thomson, Nina Schwendinger, Amr Abdulazim, Joshua Haegler, Gwendoline Canzanella, Vincens Kaelin, Linda Baettig, Lucas Moritz Wiggenhauser, Maria Wostrack, Carolin Albrecht, Matthias Gmeiner, Julia Shawarba, Daniel Couto, Sandra Wymann, Andreas Wassmer, Marlies Illi, Kathrin Bieri, Karl Rössler, Andreas Gruber, Bernhard Meyer, Constantin Roder, Isabel Charlotte Hostettler, Basil Erwin Grüter, Nima Etminan, Luca Regli, Emanuela Keller, Ulrike Held, Dominik J Schaer, Michael HugelshoferObjectives
To validate whether cerebrospinal fluid oxyhaemoglobin (CSF-Hb), measured from external ventricular or lumbar drains, is associated with secondary brain injury (SAH-SBI) after aneurysmal subarachnoid haemorrhage (aSAH) and to assess its value as a real-time monitoring biomarker.
Design
Pre-registered multicentre prospective observational cohort study.
Setting
Eight neurosurgical tertiary centres in Switzerland, Germany and Austria between August 2021 and June 2024.
Participants
366 patients with aSAH (mean age 58 years; 65% women). Of these, 260 provided cerebrospinal fluid (CSF) samples via external ventricular drain (EVD; 2467 samples, median 10 days per patient) and 66 via lumbar drain (LD; 379 samples, median 6 days).
Interventions
Daily CSF samples were collected via EVD or LD from day 1 to day 14 after haemorrhage; no therapeutic interventions were tested.
Main outcome measures
CSF-Hb and its metabolites were analysed post hoc in a blinded manner. The primary outcome was SAH-SBI, defined as a composite of angiographic vasospasm (aVSP), delayed cerebral ischaemia (DCI) and delayed ischaemic neurological deficits (DIND), assessed daily over 14 days. Secondary outcomes included temporal CSF-Hb profiles and associations with aneurysm location, haematoma volume, intraventricular haemorrhage, chronic hydrocephalus and 3 month functional outcome.
Results
CSF-Hb showed a delayed peak pattern: concentrations were low after aSAH, rose to a maximum on day 10 (EVD-derived CSF-Hb median 11.3 µM, IQR 2.64 to 25.90) and then declined. Larger haematoma volume (p<0.001) and intraventricular haemorrhage (p<0.001) were associated with higher EVD-derived CSF-Hb. SAH-SBI occurred in 209/366 patients (57%). Daily EVD-derived CSF-Hb showed no association with SAH-SBI (p=0.25) and only poor prognostic potential for same-day SAH-SBI (area under the curve 0.59, 95% CI 0.56 to 0.63), with substantial between-centre heterogeneity. In a post-hoc exploratory analysis, higher CSF methaemoglobin showed a positive point-estimate of association with SAH-SBI (OR 1.18 per log(µM), 95% CI 1.02 to 1.36). Higher acute-phase EVD-derived CSF-Hb was associated with chronic hydrocephalus and a poor 3 month functional outcome. Catheter-related infection rates were low (2.2%).
Conclusions
In this preregistered multicentre validation study, EVD-derived CSF-Hb did not perform as a robust real-time monitoring biomarker for SAH-SBI, showing limited same-day discrimination and substantial between-centre heterogeneity. These findings argue against clinical implementation of CSF-Hb point-measurement as a single-parameter biomarker. Higher CSF methaemoglobin was associated with SAH-SBI; this hypothesis-generating observation requires prospective confirmation and motivates continued investigation of haemolysis-related pathways. Future work using the HeMoVal biobank will apply multi-marker, pathway-level analyses to define haemolysis-related biomarker signatures and provide a platform for robust external validation of future candidates.
Trial registration number