Arterial development is completed initially in brainstem and cerebellum (20–24 weeks) followed by the basal ganglia and diencephalon by 24–28 weeks and finally the cortex and germinal matrix. Rupture of fragile capillaries in the subependymal germinal matrix can give rise to intraventricular hemorrhage (IVH). Immaturity of the cerebral vasculature in the border zones between vascular territories makes the preterm brain vulnerable to bleeding (1). The immature vasoregulation coupled with rises in arterial pressure due to the stresses of postnatal adaption after extreme premature delivery contributes to the pathogenesis of IVH. IVH occurs 12–72 h after birth (2). IVH causes significant morbidity and mortality in the preterm infant group. IVH occurs in 60–70% of neonates weighing 500–750 g and 10–20% of those weighing 1,000–1,500 g (3). IVH can be divided into four degrees of severity according to Papile (4):

Grade I—bleeding occurs just in the germinal matrix.

Prematurity and low birth weight are the most important risk factors for grade IV hemorrhage. Other risk factors for IVH identified include the absence of prenatal steroid therapy in woman at risk of premature delivery, symptoms of intrauterine infection in the newborn, treatment with fluid boluses and catecholamines for hypotension and treatment for metabolic acidosis, blood clotting disorders, thrombocytopenia, and hypoglycemia. Most of the mentioned risk factors are associated with a systemic inflammatory response involving cytokine release (5). Preterm infants with grade IV IVH are at the high risk of developing neuropathological complications including posthemorrhagic hydrocephalus, periventricular hemorrhagic infarction, germinal matrix destruction, and periventricular leukomalacia. This leads to significant neurodevelopmental morbidity. In one case–control study of 1,472 infants with a history of preterm delivery (23–28 weeks gestational age at birth), those with grade III–IV IVH had significantly increased rates of cerebral palsy (30%), developmental delay (17.5%), deafness (8.6%), and blindness (2.2%). Infants with grade I–II IVH also had compared with infants without IVH increased rates of neurosensory impairment (22 vs 12.1%), cerebral palsy (10.4 vs 6.5%), developmental delay (7.8 vs 3.4%), and deafness (6.0 vs 2.3%) (1). The fact that lower degrees of IVH are also associated with significant neurological morbidity was confirmed by Patra et al. who found that infants who have extremely low birth weight with grades I–II IVH had poor neurodevelopmental outcomes at 20 months than infants with normal cranial ultrasound (6).

The consequences of extravasated blood around blood vessels are evident from what we have learnt about the pathophysiology of brain injury in subarachnoid hemorrhage (SAH) (7, 8). Extravasated red blood cells release on disintegration free hemoglobin which acts as a nitric oxide (NO) scavenger (7). NO, which is generated from l-arginine by endothelial NO synthase stimulates soluble guanylate cyclase in vascular muscle resulting in an increase in the intracellular concentration of guanosine 3′,5′-cyclic monophosphate resulting in relaxation of vascular smooth muscle (9). NO depletion subsequently leads to vasoconstriction causing ischemia and subsequent tissue injury (7).

Free heme and tissue injury trigger a localized inflammatory response involving inflammasome namely, nucleotide-binding and oligomerization domain-like receptor pyrin domain-containing protein 3 (NLRP3) activation (10). Heme activates NLRP3 through P2X receptors, especially the P2X7R and P2X4R. The molecular mechanism by which heme promotes NLRP3 activation in the presence of serum requires Syk phosphorylation, reactive oxygen species, and K⁺ efflux. NLRP3 activation then leads to increased production of interleukin-1 (IL-1) in macrophages. All forms of neuroinflammation that involves production of IL-1 have been, where this has been investigated, shown to be associated with cerebral vasospasm (11).

Interleukin-1 induces the production of IL-6. Increased concentrations of IL-6 in the cerebrospinal fluid were found to be associated with white matter injury in premature infants (12).

Interleukin-1 has been shown to cause vasoconstriction by calcium dependent and independent mechanisms. Investigations in patients with SAH showed that CSF levels of IL-1-beta were increased and correlated with the later development of cerebral vasospasm. IL-1 acts through G-protein-coupled receptors by activation of intracellular phospholipase C and phospholipase mediated protein kinase C (PKC) activation (11):

Phosphatidylinositol trisphosphate produced by phospholipase C activity leads to calcium release that causes myosin light chain (MLC) kinase activation triggering MLC activation and calcium-dependent vasospasm.

In premature newborns, IVH causes cerebral vasospasm and associated neurodisability via heme-induced increased inflammasome-mediated IL-1 production and NO depletion.

Before embarking on trials of intervention cerebral vasospasm in neonatal IVH needs to be documented and assessed in longitudinal studies at the timepoints at which post-IVH cerebral vasospasm has been detected in adults namely at 48 h, between day 4 and 6 and between day 7 and 9 after diagnosis of IVH. To assess the presence of cerebral vasospasm in preterm neonates, non-invasive methods need to be employed. Blood velocity measurements using transcranial Doppler ultrasound can be used in preterm neonates for detection of cerebral vasospasm. The Lindegaard ratio (described earlier) can again be used to diagnose vasospasm (40). The long-term neurological outcome needs to be compared between neonates with and without cerebral vasospasm detected by this method.

Future studies could measure NO metabolites and IL-1 in the CSF of preterms with IVH serially (e.g., daily) to determine whether the increase (IL-1) or decrease (NO metabolites) in concentration over time could predict the crossing of a “vasospasmogenic threshold” and thus predict cerebral vasospasm. The absolute concentration could in this context also be correlated with the time to evolution of the vasospasm. Heme or its metabolite bilirubin could be measured at the same time to assess its predictive value for vasospasm and the need for preventative intervention.

The most important intervention with significant future potential in reduction and prevention of cerebral vasospasm and its associated neurodisability in neonatal IVH is rapid removal of the accumulated blood by all available means. Delayed removal of a perivascular clot in SAH has been demonstrated to be associated with a lack of resolution of vasospasm since clot removal up to 3 days after hemorrhage reduced vasospasm, but after 5 days, it had little influence (41). Drainage, irrigation, and fibrinolytic therapy (DRIFT) in infants with posthemorrhagic ventricular dilatation followed up until 2 years of age showed the following results: of 39 infants assigned to DRIFT, 21 (54%) died or were severely disabled vs 27 of 38 (71%) in the standard group (adjusted odds ratio 0.25 [95% confidence interval: 0.08–0.82]). Among the survivors, 11 of 35 (31%) in the DRIFT group had severe cognitive disability vs 19 of 32 (59%) in the standard group (adjusted odds ratio: 0.17 [95% confidence interval: 0.05–0.57]). Median Mental Development Index was 68 with DRIFT and <50 with standard care. Severe sensorimotor disability was not significantly reduced (42). Some of this effect may have been due to treatment of the evolving hydrocephalus but removal of the ongoing heme sources scavenging NO and causing ongoing release of vasospasm causing IL-1 may have been crucial. To assess the impact of the DRIFT intervention in all grades of IVH, ultrasound Doppler studies assessing correlates of vasospasm before and after the intervention need to be explored and put into relationship to neurological long-term outcome. At the same time, one could measure peripheral blood glial fibrillary acidic protein levels—a sensitive marker of brain injury—before and after the intervention and in controls without the intervention. Neuroendoscopic lavage for clot removal has been successfully and safely employed and shown to be possibly superior to DRIFT related procedures in preventing hydrocephalus and warrants further study (43).

Optimization of cerebral perfusion pressure by hypervolemia with intravenous fluids and induction of hypertension with neosynephrine appears to be an important auxiliary measure to be put into place as anecdotal evidence suggests (see Table 1).

As the listed case reports demonstrate larger cerebral arteries can be involved in IVH-associated vasospasm. Angioplasty has been successful in resolution of vasospasm in SAH and IVH (44, 45) and needs to be explored in established vasospasm in preterm neonates.

Case reports support the future use of peripheral intravenous infusion of the NO donor molsidomine, and intrathecal application of nitroprusside sodium in patients with IVH because they appear to reduce cerebral vasospasm (46).

To support DRIFT once established as an effective intervention intrathecal adjuvant high-dose vasodilator, anti-inflammatory, and hemoglobin and heme scavenging treatments should be employed using slow release preparations where available. Such drugs could include tissue plasminogen activator or NO donors in isolation or in combination (47) and heme scavengers haemopexin and haptoglobin (48, 49). One may argue that conclusions from the data on vasospasm in IVH in adults are not transferable to preterm neonates because inflammation may be different in the fetus and therefore the preterm infant due to the immaturity of the immune system. It is, however, unclear whether this immaturity affects IL-1 production following IVH: experiments employing lipopolysaccharide-induced IL-1 secretion from very low birth weight infants (VLBWI) and from newborns after 30 weeks gestation in comparison to adults showed a diminished ex vivo LPS stimulated IL-1 beta secretion into whole blood in VLBWI (50). In line with these findings Brochu et al. found a weaker IL-1beta response in the immature compared to the more mature term-like brain. In this rat model, as well as in human preterm newborns, the IL-1b response takes place mainly in deep white matter that corresponds to the area of vulnerability to IVH. Prematurity may be associated with an imbalance of the IL-1beta/IL-1receptor antagonist (IL-1-Ra) ratio due to a reduction of upregulation of IL-1Ra expression (51, 52). This may mean a reduced expression of IL-1 in response to stimuli is at least partially offset by an even more reduced anti-IL-1 response. Whole genome analysis of the transcriptome of the fetal inflammatory response syndrome in umbilical cord blood from preterm neonates with fetal inflammatory response syndrome revealed on microarray analysis of leukocyte RNA, a substantial degree of similarity with the systemic inflammatory response syndrome in adults (53).

To inhibit the neuroinflammation associated with heme-induced activation of inflammasomes and with increased IL-1 production leading to an increased risk of vasospasm one could treat the affected newborns with inflammasome inhibitors like telmisartan which has been shown to reduce NLRP3 activity and IL-1 levels and associated cerebral injury (54–56). These anti-inflammatory interventions could be safe in all degrees of severity of IVH but need to be first explored in animal studies. The intervention to reduce vasospasm-associated neurodisability should be employed as soon as an IVH is detected to act as effective preventative. Such animal experiments also need to explore the effect of intra-arterial verapamil, which has been shown to relieve vasospasm in a patient with IVH (see Table 1). All drug treatment interventions including intrathecal application could be assessed in the preterm rabbit model of glycerol-induced IVH, which has been developed for this purpose (57). Such studies could analyze the impact of inflammasome inhibitors and NO donors on transcranial Doppler-ultrasonographic evidence of vasospasm on neurological symptoms and histological pattern of brain injury.

ME conceived the hypothesis, drafted the outline of the review, and wrote the final version of the manuscript. SC made substantial contributions to the acquisition of data and revision. Both authors gave the final approval of the version to be published, agreed to be accountable for all aspects of the work in ensuring that analysis or interpretation of data for the work, and agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.