Spontaneous intracerebral hemorrhage (SICH), while comprising only 10–20% of all strokes (1, 2), remains one of the deadliest forms of the disease, with mortality rates approaching 40% at 1 month (3). Long-term survivors of SICH are often saddled with permanent deficits, with up to 75% of patients suffering significant disability or mortality at 1 year (4). Management of SICH patients currently consists primarily of supportive therapies (5), such as airway management, hemodynamic monitoring, and control of intracranial pressure (6), with no treatment options demonstrating significant efficacy despite extensive investigation into the topic (7).

Despite the disappointing results of interventional studies to date, there is reason to be hopeful going forward. Advancements in the understanding of secondary injury after SICH have highlighted opportunities for therapeutic intervention (5). One such opportunity is preventing secondary expansion of hemorrhage after the initial bleed. Such expansion may occur in up to 30% of SICH patients (8, 9) and is associated with significantly worse clinical outcomes (10). This impact on outcome is independent of previously described predictors of outcome in SICH (11), including patient age, Glasgow Coma Scale score, intraventricular extension, hematoma volume, hemorrhage location, anticoagulant use, and medical history (12–14).

This review will discuss the classifications and current animal models of SICH, as well as what is known about the pathophysiology of secondary hematoma expansion. The interaction between bench research and clinical trials will be examined, with a focus on blood pressure control and the hemostatic mechanism – two areas where findings in animal models of SICH have lead to large-scale, randomized controlled trials in humans.

Broadly, SICH is defined as any intracerebral hemorrhage that is non-traumatic in nature; SICH can be further divided into primary and secondary hemorrhage (15). Primary SICH consists of those hemorrhages in which an underlying vascular malformation or coagulopathy is not identified (16). The two most common causes of primary SICH are arteriosclerosis due to chronic hypertension and cerebral amyloid angiopathy, which together account for up to 88% of all primary SICH (17). Chronic hypertension initially leads to proliferation of smooth muscle cells in the small penetrating arterioles of the brain, but eventually smooth muscle cell death occurs, with replacement of muscle in the tunica media layer with collagen (18). This weakening of the arteriolar wall can lead to vessel ectasia – Charcot–Bouchard aneurysms – and subsequent rupture; it occurs primarily in the deep, penetrating arterioles of the brain (19). In cerebral amyloid angiopathy, the progressive deposition of insoluble amyloid protein in the walls of small- and medium-sized vessels leads to increased vessel fragility over time (20). This deposition increases dramatically with age and occurs primarily in the leptomeningeal and cortical vasculature (21). As a result, SICH caused by cerebral amyloid angiopathy is significantly more common in the elderly population and is more commonly seen in a superficial cortical distribution (21). Patients with cerebral amyloid angiopathy are also at higher risk of recurrent hemorrhage (22).

Secondary SICH can be caused by a variety of underlying lesions and pathologies. Vascular malformations that can lead to SICH include arteriovenous malformations (23), cerebral aneurysms (24), dural arteriovenous fistulas (25), and cavernous malformations (26). Patients who have had ischemic strokes can experience hemorrhagic conversion (27), as can up to 50% of cerebral venous thrombosis patients (28). Neoplastic causes of SICH make up a minority of cases, but melanoma, choriocarcinoma, renal cell carcinoma, and thyroid carcinoma are the most prone to bleeding (29). Investigations into secondary hematoma expansion in secondary SICH are fairly limited, with the exception of those evaluating hemorrhages associated with oral anticoagulant use, where secondary expansion is both more common and associated with worse outcomes (30).

Currently, there are two widely used paradigms for modeling of SICH in animals. The first is the intracerebral injection of autologous blood. Initially developed in the 1960s (31), this model has been used in both large animals (32–34) (with injections typically performed into the frontal lobe) and rodents (15, 35) (with injections typically performed into the basal ganglia). This model has the advantage of allowing for control of hemorrhage volume but does not mimic the effects of vessel rupture seen in SICH (36). The second model involves injection of collagenase into the brain, leading to compromise of the extracellular matrix and subsequent vessel rupture. Developed in the 1990s and used primarily in rodents (37, 38), this model replicates the vascular disruption seen in SICH but has more diffuse effects when compared with the injection model and may result in diffuse inflammation and ischemia that is not commonly seen in the disease process (39). Despite these shortcomings, the collagenase injection model is more commonly used in the evaluation of secondary hematoma expansion.

Although the precise mechanism for secondary hematoma expansion has yet to be fully elucidated, two predominant models currently exist. The first is the “persistent bleeding” model, which proposes that the vessel or vessels that initially rupture continue to bleed and lead to an enlarging hematoma (7). The second model, initially proposed by Fisher (40), postulates that secondary expansion results from the mechanical disruption of neighboring vasculature by the initial bleed.

Regardless of the underlying histopathological cause, experimental evidence suggests that two important contributors to secondary hematoma expansion may be derangement of the coagulation cascade (41, 42) (Figure 1) and hypertension. A variety of coagulopathies have been shown to be associated with an increased risk of secondary hematoma expansion. In a murine collagenase injection model of SICH, Illanes et al. (43) demonstrated significantly higher rates of hematoma expansion and larger hematoma volumes in animals treated with warfarin when compared with controls. A variety of newer anticoagulant agents act to inhibit either factor Xa (e.g., rivaroxaban) or thrombin (e.g., dabigatran). By doing so, these agents halt the so-called “thrombin burst,” whereby a single molecule of factor Xa may activate several hundred molecules of thrombin (44). This “thrombin burst” subsequently upregulates myriad other coagulation factors (45). For these novel agents, Zhou et al. (46) found that hematoma volume and secondary expansion were significantly increased in a subgroup of mice treated with rivaroxaban and high-concentration collagenase (47). Surprisingly, Lauer et al. (48) found that secondary hematoma expansion was not dramatically higher in animals treated with dabigatran compared with control animals, although larger hemorrhagic volumes and worsened neurologic function were seen with high-dose parenteral administration of the drug. The myriad disorders, which may affect the coagulation cascade, have led to the development of more refined laboratory testing methods to detect and track its function. One of the most widely adopted of these is thromboelastography (49), a modality that detects variables such as the rate of fibrin formation, the thrombin burst, and the firmness of the resultant clot (50). Such tests may allow for the stratification of patients at risk of hematoma enlargement and represents an area for future study (50).

Non-pharmacologic disruption of hemostatic mechanisms may be equally important. In an autologous blood injection rat model of intracerebral hemorrhage, Liu et al. (51) found a relationship between hyperglycemia and secondary hematoma expansion. They discovered that this relationship may be mediated by the platelet-inhibiting effects of plasma kallikrein, a protein whose action is enhanced by elevated blood glucose. Hyperglycemia has also been shown in animal models to worsen perihematomal neurolysis and edema, likely secondary to increased levels of inflammatory cytokines such as TNF-α (52). Current work on intensive hyperglycemic management in ischemic stroke is ongoing through the multicenter Stroke Hyperglycemia Insulin Network Effort (SHINE) (53), but organized efforts in hemorrhagic stroke are needed.

The relationship of hypertension to secondary hematoma expansion is more controversial. Although some experimental evidence does exist that higher blood pressures may lead to hematoma expansion and larger overall hematoma volumes (54), other investigators have found no relationship between hypertension and hematoma expansion (55). This variability may be related to the acuity of blood pressure rise, with large, rapid changes in blood pressure being less well tolerated than chronic hypertension (56). One rat model of hypertensive hemorrhagic stroke showed increased vascular permeability on MRI in hypertensive rats in the 1–2 weeks prior to SICH (57), suggesting that vessel changes may be detectable prior to catastrophic hemorrhage. Both animal (58) and human (27) studies also suggest that patients with anticoagulation and hypertension may be at further risk of intracerebral hemorrhage, and focused investigation on these commonly overlapping entities is also warranted.

Because of the high mortality and incidence of long-term disability associated with SICH, it remains a significant health care burden globally. Current therapeutic options remain limited, but continued investigations into the underlying mechanisms of secondary injury, such as hemorrhage expansion, have yielded promising targets for intervention that continue to warrant further investigation. Further refinement of animal models of SICH, the discovery of therapeutic pathways in the laboratory setting, and the transition of this knowledge into treatments in humans remain critically needed.

JG and GH conceived of the topic, drafted the manuscript, and did critical review and revision of manuscript.

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. The reviewer JP and handling editor declared their shared affiliation, and the handling editor states that the process nevertheless met the standards of a fair and objective review.