Active Cerebrospinal Fluid Exchange with IRRAflow® in Difficult Cerebrospinal Fluid: A Multicenter Italian Retrospective Cohort Study

Technical Summary: Active Cerebrospinal Fluid Exchange with IRRAflow® in Difficult CSF

Problem Statement
External ventricular drainage (EVD) is the standard of care for cerebrospinal fluid (CSF) diversion and intracranial pressure control in neurocritical care. However, conventional passive drainage systems often fail when CSF is hemorrhagic, protein-rich, debris-laden, or purulent. In these "difficult CSF" scenarios, catheter occlusion and treatment interruptions are frequent, compromising therapeutic continuity and increasing the risk of infection and other device-related complications. While Active Cerebrospinal Fluid Exchange (ACE) using the dual-lumen IRRAflow® system offers controlled irrigation with balanced drainage in a closed circuit, existing literature remains heterogeneous, largely observational, and limited to case series or small cohorts. There is a lack of prospective multicenter data characterizing the safety of ACE across different clinical indications, exposure durations, and local care processes.

Methodology
This study is a multicenter, retrospective observational cohort study conducted across seven Italian neurosurgical and neurocritical care centers between October 1, 2022, and December 30, 2025.

• Participants: 56 adult patients (≥18 years) treated with IRRAflow® for hemorrhagic or infectious CSF pathology.
• Indications: Cases were categorized into two mutually exclusive macro-groups:
  1. Hemorrhagic (69.6%): Including intraventricular hemorrhage, intracerebral hemorrhage with ventricular extension, aneurysmal subarachnoid hemorrhage with ventricular involvement, and a specific subgroup of chronic subdural hematomas (cSDH) managed with irrigation.
  2. Infectious (30.4%): Including ventriculitis, meningitis, or conditions with purulent/debris-laden CSF.
• Intervention: The IRRAflow® system (9 Fr dual-lumen catheter) was used for active irrigation and unidirectional gravity-driven drainage. Settings (cycle patterns, irrigation doses, alarms) were determined by local clinical judgment rather than a standardized protocol, reflecting real-world practice.
• Endpoints:
  • Primary: Occurrence of ≥1 adverse event (AE) during device use (occlusion, unplanned removal, device-related infection, intracranial abscess). AEs were reported as proportions and rates per 100 device-days to account for varying exposure durations.
  • Secondary: Changes in Glasgow Coma Scale (GCS), 30-day clinical status (improved/stable/worsened/death), modified Rankin Scale (mRS), and neuroradiological evolution.
• Analysis: Statistical analysis utilized R, employing descriptive statistics, Mann-Whitney U tests, t-tests, and Fisher's exact tests. Due to sparse events, multivariable modeling was limited, and the study relied heavily on exposure-adjusted reporting.

Key Results

• Cohort Characteristics: The mean age was 60.0 years; 66.1% were male. The median device duration was 7 days (IQR 4–11), with a total exposure of 473 device-days. Infectious cases had a longer median duration (8 days) compared to hemorrhagic cases (6 days).
• Safety Outcomes:
  • Overall, 19.6% (11/56) of patients experienced ≥1 AE.
  • Indication Dependency: Safety varied significantly by indication. AEs occurred in only 2.6% (1/39) of hemorrhagic cases versus 58.8% (10/17) of infectious cases.
  • Exposure-Adjusted Rates: When normalized to 100 device-days, the AE rate was 0.38 for hemorrhagic indications versus 4.78 for infectious indications (Rate Ratio 12.63, p=0.0018).
  • Specific AEs included occlusion/malfunction (7.1%), unplanned removal (7.1%), device-related infection (3.6%), and intracranial abscess (1.8%). No insertion-related hemorrhagic complications were reported.
• Clinical and Radiological Trajectories:
  • Mean GCS improved from 8.92 pre-treatment to 10.64 post-treatment (p=0.001).
  • At 30 days, 58.9% of patients were improved, 17.9% stable, 3.6% worsened, and 19.6% died.
  • Among survivors with documented mRS (n=39), 43.6% achieved a favorable outcome (mRS 0–2).
  • Neuroradiological follow-up (n=38) showed improvement in 57.9%, stability in 13.2%, and worsening in 28.9%. All patients with radiological worsening died within 30 days.
• Subgroup Analysis (cSDH): In the 12 patients treated for chronic subdural hematoma, there were no adverse events and no 30-day mortality, with 90.9% achieving a favorable mRS.
• Acceptability: Clinician propensity to reuse the system was 98.2% (55/56).

Significance and Claims
The authors conclude that IRRAflow®-based ACE is feasible and highly acceptable in real-world neurocritical care settings. However, the study emphasizes that safety is not uniform across all patient populations. Instead, risk is strongly dependent on the clinical context (indication) and device exposure duration.

• Context-Dependent Safety: The data suggests that infectious conditions represent a high-risk environment for any ventricular device due to inflammatory burden and prolonged manipulation, resulting in a significantly higher AE burden even after exposure adjustment. Conversely, hemorrhagic indications appear to be a more favorable operational setting for ACE, with low AE rates per device-day.
• Process-Oriented Approach: The findings support viewing ACE not as a generic device substitution but as a protocol-sensitive intervention. The authors argue that outcomes reflect the interaction between indication, CSF quality, exposure duration, and care-process variables.
• Future Directions: The paper modestly claims that broad effectiveness generalization is premature. Instead, it advocates for a shift toward indication-specific prospective studies that utilize exposure-based safety metrics (events per 100 device-days) and standardized process variables (irrigation dose, net balance, alarm handling). The study positions itself as a pragmatic safety signal that validates the need for more rigorous, protocolized evaluation before causal claims about clinical benefit can be established.