Longitudinal CSF Iron Pathway Proteins in Posthemorrhagic Hydrocephalus: Associations with Ventricle Size and Neurodevelopmental Outcomes

Abstract

Objective: Iron has been implicated in the pathogenesis of brain injury and hydrocephalus after preterm germinal matrix hemorrhage-intraventricular hemorrhage, however, it is unknown how external or endogenous intraventricular clearance of iron pathway proteins affect the outcome in this group.

Methods: This prospective multicenter cohort included patients with posthemorrhagic hydrocephalus (PHH) who underwent (1) temporary and permanent cerebrospinal fluid (CSF) diversion and (2) Bayley Scales of Infant Development-III testing around 2 years of age. CSF proteins in the iron handling pathway were analyzed longitudinally and compared to ventricle size and neurodevelopmental outcomes.

Results: Thirty-seven patients met inclusion criteria with a median estimated gestational age at birth of 25 weeks; 65% were boys. Ventricular CSF levels of hemoglobin, iron, total bilirubin, and ferritin decreased between temporary and permanent CSF diversion with no change in CSF levels of ceruloplasmin, transferrin, haptoglobin, and hepcidin. There was an increase in CSF hemopexin during this interval. Larger ventricle size at permanent CSF diversion was associated with elevated CSF ferritin (p = 0.015) and decreased CSF hemopexin (p = 0.007). CSF levels of proteins at temporary CSF diversion were not associated with outcome, however, higher CSF transferrin at permanent CSF diversion was associated with improved cognitive outcome (p = 0.015). Importantly, longitudinal change in CSF iron pathway proteins, ferritin (decrease), and transferrin (increase) were associated with improved cognitive (p = 0.04) and motor (p = 0.03) scores and improved cognitive (p = 0.04), language (p = 0.035), and motor (p = 0.008) scores, respectively.

Interpretation: Longitudinal changes in CSF transferrin (increase) and ferritin (decrease) are associated with improved neurodevelopmental outcomes in neonatal PHH, with implications for understanding the pathogenesis of poor outcomes in PHH. ANN NEUROL 2021;90:217-226.

Figure 1.

Top panel (A). Relationship between FOHR and ventricular CSF ferritin levels at temporizing and permanent procedure. At the time of permanent CSF diversion, higher CSF levels of ferritin were associated with larger FOHR (p=0.015) when controlling for PMA at birth and IVH grade. Graph lines represent best fit between FOHR and ferritin at each time point not controlling for other factors. Bottom Panel (B). Relationship between Bayley III composite scores and change in ventricular CSF ferritin levels from temporizing to permanent procedure. Ventricular CSF ferritin levels at time of permanent CSF diversion are associated with Bayley III composite cognitive (p=0.04) and motor (p=0.028) when controlling for change in FOHR from temporizing to permanent procedure. Graph lines represent best fit between composite scores and change in ferritin levels not controlling for other factors.

Figure 2.

Relationship between frontal occipital horn ratio (FOHR) and ventricular cerebrospinal fluid hemopexin levels at temporizing and permanent procedure. At the time of permanent CSF diversion, higher CSF levels of hemopexin levels were associated with smaller FOHR (p=0.007) when controlling for PMA at birth and IVH grade. Graph lines represent best fit between FOHR and hemopexin at each time point not controlling for other factors.

Figure 3.

Longitudinal changes in CSF transferrin are associated with outcome in posthemorrhagic hydrocephalus. Top panel (A). No relationship between ventricular CSF transferrin levels vs. Bayley III composite cognitive, language and motor scores at time of temporary CSF diversion. Middle Panel (B). Ventricular CSF transferrin levels at time of permanent CSF diversion are associated with Bayley III composite cognitive scores. Bottom Panel (C). Change in ventricular CSF transferrin levels between temporary and permanent CSF diversion is significantly associated with improved Bayley III composite cognitive (p=0.04), language (p=0.035) and motor (p=0.008) scores. Statistical model controls for PMA at birth and IVH grade. Graph lines represent best fit between composite scores and transferrin not controlling for other factors.