Intraparenchymal hemorrhage

Introduction
Intraparenchymal hemorrhage (IPH) refers to hemorrhage occurring within the brain parenchyma itself. A few notes on terminology:
 * The term intracerebral hemorrhage (ICH) is often considered synonymous. However, this term is problematic because similar pathophysiology underlies hemorrhages occurring within the cerebrum, the cerebellum, and the brainstem, and therefore the term intraparenchymal hemorrhage is preferred.
 * The term intracranial hemorrhage is nonspecific and refers to any type of hemorrhage, traumatic or otherwise, that occurs within the cranium.
 * The term hemorrhagic stroke can also be problematic. Although this is technically a type of stroke, the pathophysiology is entirely different from that underlying ischemic stroke, and thus its classification as a type of stroke is not particularly useful.  In this sense, the only similarities between intraparenchymal hemorrhage and ischemic stroke are that they are both causes of sudden neurological deficits related to cerebral vasculature.  Moreover, the term hemorrhagic stroke is sometimes used to describe a particular etiology of intraparenchymal hemorrhage, namely hemorrhagic conversion of infarction.

Other risk factors

 * Statin therapy, possibly low LDL-C
 * An analysis of 27,937 women in the Women's Health Study showed a 2-fold increased risk of both intraparenchymal hemorrhage and subarachnoid hemorrhage in patients with low LDL-C. . However, the annual risk is 0.04% per year in this group (0.02% higher than baseline risk), which is still much lower than the cardiovascular risk for patients with higher LDL. Studies now suggest that the hemorrhagic risk may not be due to LDL itself, but may be a statin specific effect. Even with intensive LDL therapy with statins, the risk of hemorrhagic stroke in a large meta-analysis was low, with number needed to harm of 242, absolute risk 1.3% vs 0.9%.
 * Triglycerides ≤ 74 mg/dL
 * An analysis of 27,937 women in the Women's Health Study showed a 2-fold increased risk of both hemorrhagic stroke (1.8-fold for IPH, and 3.3-fold for IPH) for patients with low triglycerides.

Graeb's scale
This scale is a method to grade the severity of IVH.

Guidelines
American Heart Association / American Stroke Association Guidelines for Management of Spontaneous Intracerebral Hemorrhage

Blood Pressure Control
Reasonable to aim for goal systolic blood pressure (SBP) of 130-150 mmHg in most settings.
 * Exceptions:
 * In patients with elevated intracranial pressure (ICP) from IPH, caution is advised before lowering BP acutely, as in the setting of elevated ICP, a drop in systemic BP could cause low cerebral perfusion pressure (CPP) and worsen brain injury.
 * In patients with very high SBP on presentation, it is reasonable to lower them more slowly, initially to <180 mmHg or so, and then lower them further. There is no clear evidence behind the approach, but it makes pathophysiological sense for a patient whose cerebral autoregulation may be shifted due to chronic hypertension.

The most recent guidelines from 2015 state:"1. For ICH patients presenting with SBP between 150 and 220 mmHg and without contraindication to acute BP treatment, acute lowering of SBP to 140 mmHg is safe (Class I; Level of Evidence A) and can be effective for improving functional outcome (Class IIa; Level of Evidence B). (Revised from the previous guideline)""2. For ICH patients presenting with SBP >220 mmHg, it may be reasonable to consider aggressive reduction of BP with a continuous intravenous infusion and frequent BP monitoring (Class IIb; Level of Evidence C). (New recommendation)"However, these guidelines were published after the INTERACT2 trial but prior to the results of the ATACH-2 trial, which has changed our view on BP management. INTERACT-2 suggested that lowering of SBP in acute IPH to a goal of <140 mmHg was associated with a trend towards improvement in outcome with no adverse events. However, ATACH-2 showed no improvement in outcomes and a trend towards harm when SBP was lowered to <140 mmHg. In analyses of the actually achieved blood pressure in each study, it appears that harm tends to occur at SBP <120 or <130, and benefit may occur at BP <150. As such, it is reasonable to lower SBP to 130-150 mmHg in most settings of IPH.

Associated edema
Edema develops around IPH in the subsequent days. The classical teaching is that most of the edema develops within the first 48 hours, but data suggests that edema can continue to worsen up to 2-3 weeks after onset. Importantly, the amount of edema that develops has to do with the overall surface area in contact with the brain parenchyma. The relative surface area to volume ratio is smaller IPH, and therefore smaller IPH has more relative perihemorrhagic edema, even if larger IPH will have more absolute perihemorrhagic edema. Irregularly-shaped hemorrhages, with more surface area in contact with brain, will have larger amounts of associated edema as well.

Risk factors for worsened perihematomal edema:
Edema should not be treated with steroids, which is a Neurocritical Care Society Clinical Performance Measure.
 * Larger surface area of hemorrhage
 * Diminished ipsilateral venous drainage

Intraventricular hemorrhage
The presence of intraventricular hemorrhage (IVH) frequently leads to hydrocephalus and poor outcomes.

Intraventricular fibrinolytcs
In 2017, the CLEAR III trial was published, which randomized 500 patients with intraventricular alteplase (1 mg q8h up to 12 doses) vs saline. They found no difference in good outcome (mRS 0-3) at 6 months, but lower mortality in the intraventricular alteplase group (18% vs 29%, HR 0.60, 955 CI 0.41-0.86). There were a greater proportion of severely debilitated patients (mRS 5) at 180 days in the treatment group (RR 1.99, 95% CI 1.22-3.26). There were no differences in ventriculitis, symptomatic bleeding, or serious adverse events. This trial caused pause as it was improving mortality but leading to survival of patients with severe disability.

A meta-analysis was published in 2020 which similarly showed a clear mortality benefit for intraventricular thrombolysis with RR 0.58 (95% CI 0.47-0.72), less EVD obstruction (RR 0.41, 95% CU 0.22-0.74), and fewer days to IVH clearance.There was a trend towards improved functional outcome but this was not statistically significant, and there was no difference in shunt dependency. With a correction for publication bias, there was a slight increased risk of bleeding (RR 1.67, 95% CI 1.01-2.74) but a slightly decreased risk of ventriculitis (RR 0.61, 95% CI 0.41-0.91).

Atrial fibrillation
A retrospective study by Roquer et al. compared patients with IPH who had atrial fibrillation with those who did not. They found that, as expected, antiplatelet/anticoagulant use before IPH is associated with higher 3-month mortality, but this effect only persisted in patients without atrial fibrillation. This suggests that patients treated with antithrombotic agents for atrial fibrillation who have IPH at least do not do worse than those who were not treated with antithrombotics.

Delirium
Higher NIHSS, higher ICH score, larger ICH volume, the presence of IVH, mechanical ventilation, and impaired arousal, all were risk factors for delirium in one study. The CAM-ICU has poor sensitivity and specificity in this population, while the Intensive Care Delirium Screening Checklist performs better.

Hematoma expansion
A lowest CT attenuation value within the hematoma of ≤31 Hounsfield units was 92% specific (but only 62% sensitive) for hematoma expansion and 93% specific (but only 45% sensitive) for poor outcome in one retrospective study.

Risk factors

 * Patient factors
 * Anticoagulation
 * With warfarin
 * GCS (higher)
 * History of
 * Dementia
 * IPH (prior)
 * Smoking (current)
 * Shorter time from onset to CT
 * Imaging factors
 * Baseline IPH volume
 * Hypodensities within the hematoma or "blend sign"
 * Intraventricular extension
 * SAH

Several scoring systems have been developed to predict IPH expansion.

BRAIN score (2015)
This score was developed based on patients in the INTERACT2 trial and validated with patients from the INTERACT1 trial, to predict patients with IPH growth at 24 hours of ≥6 mL from baseline. AUROC was 0.73 in both the derivation and validation data sets. . It was validated and compared against the HEP and BAT scores in 230 subjects at two centers in Italy, in which it had an AUROC of 0.70 which was the same as the BAT score and slightly better than the HEP score (AUROC 0.65). The BRAIN score had the highest sensitivity of 76% (compared with 59% for HEP and 49% for BAT) but the lowest specificity of 61% (compared with 63% for HEP and 81% for BAT).

Hematoma Expansion (HEP) score (2015)
This score was developed from 237 patients at a single center in the USA to evaluate risk factors for IPH expansion (absolute increased in volume > 6 mL or >33% on follow-up CT). AUROC was 0.76. It was validated and compared against the BRAIN and BAT scores in 230 subjects at two centers in Italy, in which it had an AUROC of 0.65 which was worse than both the BAT score and the BRAIN score (AUROCs 0.70). It had a sensitivity of 59% which was better than the BAT score (49%) but worse than the BRAIN score (76%) and specificity of 63% which was better than the BRAIN score (61%) but worse than the BAT score (81%). However, it had the highest inter-rater reliability of the scores since it does not have many radiographic features.

BAT score (2018)
This score was developed from a cohort of 344 patients at a hospital in the USA to predict hematoma expansion at 24 hours (defined as relative growth > 33% or absolute growth > 6 ml). It was then validated on subjects enrolled in the ATACH-2 trial (n=954) and in the PREDICT study (n=241). In had an AUROC of 0.77 in the development cohort, 0.65 in the ATACH-2 validation cohort, and 0.70 in the PREDICT validation cohort. It was validated and compared against the HEP and BRAIN scores in 230 subjects at two centers in Italy, in which it had an AUROC of 0.70 which was the same as the BRAIN score and slightly better than the HEP score (AUROC 0.65). It was the most specific of the scores (81% vs 61% for BRAIN and 63% for HEP) but the least sensitive (49% vs 76% for BRAIN and 59% for HEP).

Remote DWI lesions
These lesions are of unclear etiology and may be associated with worsened outcomes. Stress-induced hyperglycemia may be a risk factor for their occurrence.

Type of unit
Admission to a dedicated unit (stroke unit, ICU) with neuroscience physician and nursing expertise for all intraparenchymal hemorrhage patients is a Neurocritical Care Society Clinical Performance Measure.

Care bundles
Using a bundle of care with specific targets (reversal of AC within 90 minutes of arrival, goal BP 130-140 mmHg within 60 minutes, and prompt neurosurgical referral for selected patients) led to a significantly lower odds of death (OR 0.4, 95% CI 0.24-0.16). Overall case fatality rate at 30-days was 10.8% less. However, over 1/2 of the association with improved survival was attributable to a reduction in DNR orders within 24 hours.

DNR orders
Avoid early DNR orders in most patients, as they create a self-fulfilling prophecy In a before-and-after study using care bundles, a 10.8% absolute reduction in mortality was seen, of which more than 50% of the effect of improved survival was attributable to a reduction in DNR orders within 24 hours. In centers with a policy of providing full care for IPH patients for the first 5 days after symptom onset, observed mortality in patients with severe ICH (GCS 12 or less) was 20.2% compared with expected mortality of 50% based on the ICH score. Importantly a good outcome (mRS 0-3) was achieved by 30% of people and independence (mRS 0-2) was achieved in ~10% of patients. Several studies have shown that prognostic models that do not take into account limits on care early after IPH (e.g. ICH score) significantly overestimate mortality and poor outcome.

Patient factors associated with worsened outcome

 * Age (higher)
 * APACHE II score (higher)
 * Blood pressure (higher)
 * Coagulopathy
 * Cognitive impairment prior to IPH
 * FOUR score (worse)
 * GCS (worse)
 * Heart rate (higher)
 * IVH
 * Location of IPH (infratentorial and deep worse than lobar)
 * Limb weakness
 * NIHSS (higher)
 * Premorbid level of functioning
 * Pulse pressure (wider or lower )
 * Race (white, compared with black or hispanic)
 * Size of ICH (bigger = worse)
 * Subarachnoid extension
 * Temperature (higher)

Lab predictors

 * Glucose (higher)

Imaging predictors
White matter hyperintensity on MRI at baseline is not clearly associated with outcome after IPH, regardless of blood pressure.
 * CT hypoattenuation within the hematoma: A lowest CT attenuation value within the hematoma of ≤31 Hounsfield units was 93% specific (but only 45% sensitive) for poor outcome (mRS 4-6) in one retrospective study.
 * Midline shift
 * 5-10 mm: OR of unfavorable outcome (mRS 4-6) was 1.95 (95% CI 1.95-9.56)
 * >10 mm: OR of unfavorable outcome (mRS 4-6) was 50.14 (95% CI 6.85-366.92)
 * IVH: OR of unfavorable outcome (mRS 4-6) was 7.12 (95% CI 4.22-11.99)
 * Hydrocephalus: OR of unfavorable outcome (mRS 4-6) was 10.81 (95% CI 4.67-25.02)
 * Herniation: OR of unfavorable outcome (mRS 4-6) was 18.1 (95% CI 5.42-60.45)
 * Hemorrhage in pons: OR of unfavorable outcome (mRS 4-6) was 5.91 (95% CI 1.95-17.93)

Patient factors

 * Age (older)
 * Blood pressure (higher)
 * Coagulopathy
 * Diabetes mellitus
 * Dialysis dependent
 * FOUR score (worse)
 * GCS (worse)
 * Location of IPH (infratentorial)
 * Hypertension (history)
 * IVH
 * NIHSS (higher)
 * Obesity (ABSENCE of - obesity itself may be protective)
 * Pulse pressure (wider)
 * Subarachnoid extension
 * Size of IPH

Lab predictors

 * Serum
 * Glucose (higher)
 * Potassium (higher)
 * WBCs (higher)

Imaging predictors
A combined model using GCS, hemorrhage volume, midline shift, interaction (volume x midline shift), presence/absence of pontine hemorrhage, and age, was 90% specific and 81% sensitive for mortality, for AUROC of 0.95.
 * Midline shift (in-hospital mortality)
 * 2-5 mm (OR 3.12, 95% CI 1.48-6.57)
 * 5-10 mm (OR 2.92, 95% CI 1.52-5.59)
 * >10 mm (OR 15.76, 95% CI 7.95-31.22)
 * IVH: OR 4.97 (95% CI 3.16-7.82) for in-hospital mortality
 * Hydrocephalus: OR 7.1 (95% CI 3.97-12.63) for in-hospital mortality
 * Herniation: OR 6.7 (95% CI 3.52-12.82) for in-hospital mortality
 * Hemorrhage in pons: OR 4.49 (95% CI 2.01-10.02) for in-hospital mortality
 * Subarachnoid extension

Prognostic scores
Numerous prognostic scores have been developed for prognosticating in patients with IPH. Click here for the main article.

Other outcomes

 * Tracheostomy: more commonly needed in obese patients (OR 2.07, 95% CI 1.62-2.66)
 * Complications: higher rates of renal, pulmonary, GI, infectious, and VTE complications in obese patients.
 * Hospital charges: higher in obese patients

Recurrent IPH
Risk factors include:
 * Non-white race. For black patients (HR 1.22, 95% CI 1.01-1.48) and Asian patients (HR 1.29, 95% CI 1.10-1.50), but not for Hispanic patients (HR 0.98, 95% CI 0.84-1.14).
 * Convexity SAH, either adjacent or remote from the IPH (aHR 7.5, 95% CI 2.6-21.1)