Significant individual variation in cardiac-cycle-linked cerebrospinal fluid production following subarachnoid hemorrhage

Abstract

Background: Spontaneous subarachnoid hemorrhage (SAH) often results in altered cerebrospinal fluid (CSF) flow and secondary hydrocephalus, yet the mechanisms behind these phenomena remain poorly understood. This study aimed to elucidate the impact of SAH on individual CSF flow patterns and their association with secondary hydrocephalus.

Methods: In patients who had experienced SAH, changes in CSF flow were assessed using cardiac-gated phase-contrast magnetic resonance imaging (PC-MRI) at the Sylvian aqueduct and cranio-cervical junction (CCJ). Within these regions of interest, volumetric CSF flow was determined for every pixel and net CSF flow volume and direction calculated. The presence of acute or chronic hydrocephalus was deemed from ventriculomegaly and need of CSF diversion. For comparison, we included healthy subjects and patients examined for different CSF diseases.

Results: Twenty-four SAH patients were enrolled, revealing a heterogeneous array of CSF flow alterations at the Sylvian aqueduct. The cardiac-cycle-linked CSF net flow in Sylvian aqueduct differed from the traditional figures of ventricular CSF production about 0.30-0.40 mL/min. In 15 out of 24 patients (62.5%), net CSF flow was retrograde from the fourth to the third and lateral ventricles, while it was upward at the cranio-cervical junction in 2 out of 2 patients (100%). The diverse CSF flow metrics did not distinguish between individuals with acute or chronic secondary hydrocephalus. In comparison, 4/4 healthy subjects showed antegrade net CSF flow in the Sylvian aqueduct and net upward CSF flow in CCJ. These net CSF flow measures also showed interindividual variability among other patients with CSF diseases.

Conclusions: There is considerable inter-individual variation in net CSF flow rates following SAH. Net CSF flow in the Sylvian aqueduct differs markedly from the traditional ventricular CSF production rates of 0.30-0.40 mL/min in SAH patients, but less so in healthy subjects. Furthermore, the cardiac-cycle-linked net CSF flow rates in Sylvian aqueduct and CCJ suggest an important role of extra-ventricular CSF production.

Keywords: Cerebrospinal fluid flow; Cerebrospinal fluid production; Cranio-cervical junction; Subarachnoid hemorrhage; Sylvian aqueduct.

Fig. 1

Cardiac-gated flow measurements from phase-contrast MRI of subject #24. (a) The region of interest (ROI) within the Sylvian aqueduct is shown in blue and reference ROIs in red. (b) Flow velocities within the Sylvian aqueduct were determined for every pixel within the ROI. Grey lines show the aqueduct pixel velocities with black line referring to mean velocity. The colored pixel velocities are from the reference ROI with black stippled line showing the mean velocity of the reference ROI. Additional examples from the SAH cohort of flow velocities within the Sylvian aqueduct are presented in Fig. 2. (c) The ROI within the CCJ is shown in blue and the reference ROI in red. (d) Flow velocities within the CCJ were determined for every pixel within the ROI. Grey lines show the CCJ pixel velocities with black line referring to mean velocity. The colored pixel velocities are from the reference ROI with black stippled line showing the mean velocity of the reference ROI

Fig. 2

Cardiac-gated flow measurements within Sylvian aqueduct of SAH patients. Flow velocities within the Sylvian aqueduct were determined for every pixel within the region of interest (ROI). Grey lines show the aqueduct pixel velocities with black line referring to mean velocity. The colored pixel velocities are from the reference ROI with black stippled line showing the mean velocity of the reference ROI. For illustration, CSF flow velocities are shown for patients no. 1 (a) and 11 (b) with SAH < 3 months before, patients no. 4 (c) and 15 (d) with SAH 3–6 months earlier, patients no. 5 (e) and 16 (f) with SAH 6–12 months earlier, and patients no. 20 (g) and 23 (h) with SAH > 12 months before. The individual CSF flow estimates are further detailed in Supplementary Table 1

Fig. 3

Cardiac-cycle-linked net CSF flow (mL/min) in (a) Sylvian aqueduct and (b) cranio-cervical junction. (a) The individual net CSF flow rates in Sylvian aqueduct of patients who had experienced an SAH either < 3 months (n = 4), 3–6 months (n = 8), 6–12 months (n = 4) or > 12 months (n = 8) before, and in four healthy subjects. Negative net flow values refer to antegrade net flow from 3rd to 4th cerebral ventricles, which we consider indicative of ventricular CSF production. Positive net flow values refer to retrograde flow from 4th to 3rd cerebral ventricles, suggesting extra-ventricular CSF production. The red box refers to the traditional concept of ventricular CSF production of 0.30–0.40 mL/min. Analysis of variance (ANOVA) with Bonferroni corrected post-hoc tests showed no differences between the groups (P = 0.79). (b) The individual net CSF flow rates in cranio-cervical junction of two SAH patients and four healthy subjects. Upward net CSF flow rates were seen in all six individuals and refer to net flow from the thecal sac to the intracranial compartment, which we consider indicative of CSF production within the thecal sac. The red box refers to the traditional concept of ventricular CSF production of 0.30–0.40 mL/min

Fig. 4

Correlation between CSF production (mL/min), estimated from Sylvian aqueduct net volumetric flow during the cardiac cycle, and volumes of ventricular size and choroid plexus. While net volumetric flow in the Sylvian aqueduct did not correlate with volume of 4th ventricle (a), there was a significant positive correlation between volume of 3rd ventricle and net flow (b) and the lateral ventricles (c). There was no significant correlation between net flow in Sylvian aqueduct and volume of choroid plexus (d). The fit line and Pearson correlation coefficient are shown. Negative net flow values refer to antegrade net flow from 3rd to 4th ventricle, and positive net flow values refer to retrograde flow from 4th to 3rd ventricle

Fig. 5

Cardiac-cycle-linked net CSF flow (mL/min) in Sylvian aqueduct of different patient categories. The net flow rates were retrieved from our PC-MRI database where PC-MRI acquisitions were done similarly as presented here. Other PC-MRI variables from these patients were previously reported by Eide et al. [16] and Lindstrøm et al. [15]. Patient categories: REF (Reference subjects with no diagnosed CSF disturbance), iNPH (idiopathic normal pressure hydrocephalus), cHC (communicating hydrocephalus other than iNPH), SIH (spontaneous intracranial hypotension), IIH (idiopathic intracranial hypertension), AC (symptomatic arachnoid cyst), PC (symptomatic non-hydrocephalic pineal cyst). The red box refers to the traditional concept of ventricular CSF production of 0.30–0.40 mL/min

Fig. 6

Cardiac-cycle-linked net CSF flow (mL/min) in cranio-cervical junction of different patient categories. The net flow rates were retrieved from our PC-MRI database where PC-MRI acquisitions were done similarly as presented here. Other PC-MRI variables from these patients were previously reported by Eide et al. [16] and Lindstrøm et al. [15]. Patient categories: REF (Reference subjects with no diagnosed CSF disturbance), iNPH (idiopathic normal pressure hydrocephalus), cHC (communicating hydrocephalus other than iNPH), SIH (spontaneous intracranial hypotension), AC (symptomatic arachnoid cyst), and PC (symptomatic non-hydrocephalic pineal cyst). The red box refers to the traditional concept of ventricular CSF production of 0.30–0.40 mL/minNet aqueduct CSF flow changes and secondary hydrocephalus after SAH