2022 Guideline for the Management of Patients With Spontaneous Intracerebral Hemorrhage: A Guideline From the American Heart Association/American Stroke Association

1. Complete blood count and coagulopathy studies (prothrombin time/partial thromboplastin time/INR) can help determine hemorrhage type, including spontaneous ICH attributable to extreme thrombocytopenia (eg, platelets <10 000, although platelet counts below higher thresholds also may contribute to ICH), anticoagulant-related hemorrhage, or coagulopathy secondary to malignancy or liver failure. Anticoagulant-related hemorrhages are associated with increased hematoma volume and expansion, as well as increased morbidity and mortality.^84–86 Admission anemia is associated with hemorrhagic expansion and poor outcomes^73,74 and thrombocytopenia is associated with higher mortality for patients taking antiplatelets.⁷⁵ In patients taking warfarin, admission INR value may predict outcome. One study showed a dose response of INR level in warfarin-related hemorrhage associated with poor outcome,⁸⁸ whereas another showed no association.⁸⁹ Elevated troponin on admission for patients with ICH is associated with increased in-hospital mortality for both medical and surgical ICH patient populations.^{67,69,90–92} The association of admission troponin with functional outcomes and 30-day mortality was reported in 1 study⁶⁷ but not in another study after adjustment for confounding factors.⁷¹ Renal failure on admission also is associated with poor functional outcomes,^71,76,77,79 in-hospital mortality,⁸⁰ and 12-month mortality.^76,79 Admission hyperglycemia is associated with unfavorable short- and long-term outcomes,⁷⁰ short-term mortality,^68,78,81 and long-term mortality after ICH.⁶⁸ Additional lifestyle risk factors that should be assessed include tobacco smoking, diet, alcohol, and waist-to-hip ratio.⁹³

Further studies are necessary to determine whether platelet or coagulation activity assays may identify a subgroup of patients who benefit from platelet transfusion, desmopressin acetate, tranexamic acid (TXA), or other acute therapies for ICH.

Although changes in traditional coagulation factors or diluted thrombin time may indicate the presence of DOAC medications, these studies are not reliable enough to determine the level of anticoagulation at the time of presentation with DOAC-related ICH. Specific factor Xa inhibition levels have been developed for the factor Xa inhibitors and thrombin-based assays for dabigatran, but these studies are not widely available and often are not able to be run in an emergency setting quickly enough for decision-making. Specific reliable measurements of these anticoagulants could determine which patients may benefit from reversal of anticoagulation.

Viscoelastic hemostatic assays, including thromboelastography and rotational thromboelastography, allow measurement of both cellular and plasma components of clot formation and fibrinolysis, unlike traditional coagulation tests (prothrombin time/partial thromboplastin time/INR) that reflect an in vitro coagulation pathway. These laboratory values predict significant bleeding and need for transfusions in trauma patients but have not been shown to improve outcome or mortality. Viscoelastic assays detect coagulation abnormalities that do not always appear on traditional coagulation tests in patients with ICH. It is unclear whether the results of these studies correlate with patient outcome. Understanding the significance of these studies in patients with ICH is an area of emerging and active research.

Interpretation of admission ECG and troponin values can be challenging in patients with ICH because these can be either secondary to neurocardiogenic changes or attributable to true myocardial ischemia, which is important in the early evaluation and management of patients with ICH. Interpretation and management of early electrocardiographic changes in patients with ICH is an area of future study.

A prospective, multicenter, observational study of 62 patients presenting within 6 hours of spontaneous ICH reported that the sensitivity, specificity, predictive value, and accuracy of detecting ICH on MRI by experienced readers were 100%.⁹⁴ A similar study in 200 patients in which MRI was done first followed by CT found that MRI and CT were equivalent for detecting acute ICH and that MRI was more accurate for detecting chronic ICH.⁹⁵ A prospective, single-center study in patients with spontaneous ICH reported that MRI was slightly more sensitive than CT for detecting small IVH‚ where MRI sensitivity was 100% compared with 97% for CT.⁹⁶

HE occurs early after ICH and is an independent predictor of ND, mortality, and poor functional outcome.^30,111 A prospective, observational study in 103 patients with spontaneous ICH who had a baseline CT within 3 hours of ICH onset and a repeat CT at 1 and 20 hours after baseline scan found that substantial HE occurred in 26% of patients on the 1-hour scans and in an additional 12% of patients on the 20-hour scans.⁹⁷ HE was associated with ND. In another study, the frequency of HE was greatest among those who underwent the initial CT scan within 3 hours of ICH onset and progressively declined as the time to initial scan was prolonged; 15% of patients exhibited HE between 6 and 12 hours and 6% between 12 and 24 hours. HE after 24 hours was extremely rare (0%).⁹⁸ However, delayed IVH has been reported in 21% of patients with no initial IVH, and infrequently beyond 24 hours,⁹⁹ delayed IVH is more likely to be associated with delayed HE, is independently associated with mortality and poor outcomes, and often requires emergency surgical intervention. Incorporating new IVH appearance and IVH expansion into the definition of HE appears to improve prediction of poor neurological outcome.^113,114 In patients with ICH with stable examination and preserved level of consciousness, follow-up CT scans at ≈6 and 24 hours after onset appear adequate to exclude HE and document final ICH volume.

This recommendation pertains to indications for repeat imaging to detect other downstream effects of recent hemorrhage that may occur beyond the first 24-hour period. Evidence derived from patients with mild TBI, defined as a GCS score ≥13, suggested that routine repeat head CT in neurologically stable patients is of low yield and often unnecessary,^100,115 whereas other evidence indicated that routine serial neuroimaging may have some value in patients with moderate or severe TBI.¹⁰¹ However, these studies included few subjects with ICH (most patients had subarachnoid, subdural, or epidural hemorrhages), and the physiological differences between traumatic and nontraumatic hemorrhage limit the generalizability of these data to primary ICH. A single-center observational study in 239 patients with spontaneous ICH admitted to a neurological intensive care unit (ICU) with a standardized order set, including serial CT at 6, 24, and 48 hours and hourly neurological assessments, found that 35% of patients required emergency neurosurgical interventions after admission; 46% were instigated by imaging findings versus 54% by a change in neurological examination,¹⁰² suggesting that routine serial imaging might be of supplemental value to neurological assessments. Beyond the first 24 hours, serial imaging is generally guided by the clinical picture of the patient.

Although benefits of therapies that target HE have currently not been demonstrated, stratification of patients at risk of HE can influence the triage and intensity of monitoring of these patients and their prognosis. A prospective, multicenter, observational study reported that HE was more frequent in patients with a CTA-positive spot sign than in those without it, although the negative and positive predictive values of the spot sign were not robust.¹⁰⁴ Mortality and poor modified Rankin Scale (mRS) score at 90 days were greater in patients with CTA-positive spot sign. Subsequent meta-analyses^106–108 also suggested that CTA-positive spot sign can predict HE and mortality, although interpretation of these analyses is limited by high heterogeneity of the included studies. A meta-analysis of individual data from 5435 patients reported that the addition of the spot sign provided small improvement in the discrimination of an HE prediction model composed of simple clinical variables (ICH volume, time from ICH onset to imaging, and use of antithrombotic drugs).¹⁰³ The sensitivity and positive predictive values of the spot sign to predict HE are time dependent; they are highest between 0 and 2 hours of ICH onset–to–scan time and decrease as time lapses.¹⁰⁵ CTA also can detect some structural causes of secondary ICH (Section 4.2, Diagnostic Assessment for ICH Pathogenesis). Although CTA does not appear to commonly trigger acute renal injury,¹¹⁶ this risk remains a relevant consideration in obtaining this study.

Previous studies have suggested that signs on NCCT of heterogeneous density within the hematoma or irregularities at its margins (also described in the literature as hypodensities, fluid level, swirl, black hole, blend, island, or satellite signs) can serve as alternatives to the spot sign to predict HE¹¹² (Figures S1 and S2 in the Data Supplement). A meta-analysis of 25 studies including 10 650 patients reported that these NCCT markers are associated with HE and poor functional outcome, although there was substantial heterogeneity and pooled estimates were unadjusted for confounding variables.¹⁰⁶

Routine serial CT after the initial scan, regardless of neurological status, to evaluate for ICH expansion, development of hydrocephalus, or brain swelling is not uncommon in clinical practice. Although the usefulness of this practice has been studied extensively in patients with ICH attributable to TBI, there is a paucity of studies in patients with nontraumatic, spontaneous ICH. Future research should evaluate the cost/benefit implications of serial imaging after ICH and clarify the patient characteristics and conditions under which serial imaging should be considered.

The utility of NCCT signs to predict HE, alone or as part of prediction scores based on clinical variables, and guide decision-making on the triage and monitoring of patients with ICH at high risk for HE is appealing, particularly in low-resource settings where immediate performance and interpretation of CTA are challenging. However, the prognostic yield and clinical relevance of these NCCT signs and scores are yet to be adequately examined in prospective large studies. An important goal of future research is to refine the utility of NCCT signs (defined by standardized criteria) and HE scores to maximize their diagnostic and predictive capabilities and validity.

In patients with anticoagulant-associated ICH and without preexisting limitation of life-sustaining therapies, the anticoagulant should be discontinued immediately and rapid reversal performed as soon as possible after diagnosis of ICH, regardless of whether the INR result is available. In a case series review of warfarin-related ICH, there were significant delays before administration of reversal therapy (mean, 3.3 hours from CT to PCC, 4.8 hours from arrival to reversal agent).¹⁶² Earlier treatment was associated with a trend to better survival after controlling for severity (ICH score). In a large observational multicenter study, earlier (<4 hours) reversal of VKA-related ICH (to a goal INR <1.3) combined with BP control was associated with a significant reduction in HE and lower in-hospital mortality.¹⁷⁹ Time of last dose and renal function are likely to be the most useful tests to guide therapy; results of coagulation assays should not delay initiation of reversal therapy. Specific pathways may reduce time to reversal of anticoagulation. In 1 study, implementation of a bundle of care that included anticoagulation reversal, intensive BP lowering, neurosurgery, and access to critical care was significantly associated with lower 30-day mortality after ICH.¹⁸⁰ Reversal of anticoagulation in the presence of a left ventricular assist device (LVAD) does not appear to be associated with LVAD-related thrombosis on the basis of observational data.^181,182

In an RCT comparing 4-F PCC 30 IU/kg with FFP in patients within 12 hours of onset of VKA-associated ICH and INR >1.9, 4-F PCC was superior at rapidly reversing anticoagulation (67% of 4-F PCC–treated patients achieved INR ≤1.2 within 3 hours of starting treatment versus 9% of FFP-treated patients).¹⁶³ In addition, 4-F PCC was associated with a reduction in HE (18.3% versus 27.1% of FFP-treated patients at 24 hours). No statistically significant difference was observed in functional outcomes, although the study was not powered for this. The study was stopped early by the data safety monitoring board because of concerns in the FFP group, who had a higher rate of HE. There was no difference in SAEs or thromboembolic events. Infusion of PCC was significantly faster than infusion of FFP. An earlier RCT comparing 4-F PCC with FFP in patients with acute major bleeding and INR ≥2.0 included 24 patients with intracranial hemorrhage.¹⁸³ Overall, the study demonstrated noninferiority of 4-F PCC to FFP in hemostatic efficacy and superiority in rapid INR correction. The dose of 4-F PCC is based on INR and body weight (25–50 IU/kg), or fixed-dose regimens (1500 U for intracranial bleeding) are used.⁸⁷ The optimal dosing strategy will require large randomized studies. Although multiple formulations of both PCC and plasma products are available in different settings, most have not been systematically studied for their relative effectiveness.⁸⁷

In a case review of 17 patients with VKA-associated major bleeding, PCC rapidly corrected the INR when given with or without vitamin K.¹⁶⁵ However, in 2 cases when PCCs were administered without vitamin K, despite initial rapid normalization of the INR, there was a rebound increase 12 to 24 hours later. One patient given PCC without vitamin K had hematoma enlargement with clinical deterioration. All participants in the INCH trial (International Normalized Ratio [INR] Normalization in Coumadin Associated Intracerebral Haemorrhage),¹⁶³ which confirmed the superiority of 4-F PCC over FFP, received 10 mg IV vitamin K in addition to 4-F PCC. Intravenous vitamin K at a dose of 5 to 10 mg should be administered regardless of the type of coagulation factor replacement (PCC or plasma) in patients with VKA-related ICH.¹⁶⁴

Patients with VKA-associated ICH with INR <2.0 were excluded from the RCTs confirming superiority of PCC over FFP. A systematic review of the treatment of warfarin-associated bleeding included 318 patients in 12 studies, 3 of which included patients with intracranial hemorrhage. Patients who received PCCs had a more rapid correction of anticoagulation, but whether clinical outcomes were improved was unclear.¹⁶⁴ A case review of 88 patients with warfarin-related ICH and INR >1.2 demonstrated survival benefit of PCC over FFP.¹⁶² Dosing information for 4-F PCC recommends doses for use only when INR ≥2.0. An observational study of 205 patients with VKA-related ICH treated with PCC to an INR ≥1.5 noted increased risk of venous thromboembolism (VTE) with higher doses of PCC (>2000 and 3000 IU).¹⁸⁴ Hence, a lower dose of 10 to 20 IU/kg is suggested when INR is <2.0 but ≥1.3 to achieve rapid correction of INR and limit HE.

Andexanet alfa is a recombinant coagulation factor that reverses the inhibition of factor Xa. In a large multicenter open-label study in patients with major bleeding within 18 hours after administration of a factor Xa inhibitor (apixaban, edoxaban, enoxaparin, rivaroxaban), andexanet alfa significantly reduced anti–factor Xa activity, with a 10% VTE rate and 15% mortality rate.¹⁶⁶ In a subgroup publication of patients with factor Xa inhibitor–associated ICH, excellent or good hemostatic outcome, defined as <35% increase in hematoma volume after 12 hours, was seen in 79% of patients.¹⁶⁷ A number of small single-center case series have described comparable rates of hemostatic efficacy, mortality, and safety comparing andexanet alfa with PCC, although definitions of HE vary.¹⁸⁵ Data comparing outcomes in patients given either andexanet alfa or PCC are limited by baseline imbalances between the groups attributable to selection bias. In a small single-center study, higher rates of hemostatic efficacy and thromboembolism were seen in the andexanet alfa group, and cost was significantly higher when andexanet alfa was used compared with PCC.¹⁸⁶ In another small comparison, andexanet alfa was similar to PCC in terms of stability of hematoma on CT at 6 and 24 hours.¹⁸⁷ Hence, although andexanet alfa can be effective to reverse anti–factor Xa activity, data on safety and clinical outcomes (including functional outcome) from a randomized trial are awaited. Because of the structural similarity of the factor Xa inhibitors, andexanet alfa also likely neutralizes betrixaban and edoxaban in the same manner.¹⁸⁸ The recommended dosing of andexanet alfa depends on the specific factor Xa inhibitor and the time since last dose.¹⁸⁹

Idarucizumab, a monoclonal antibody, binds dabigatran with high affinity and neutralizes its activity. In a large prospective cohort study, in patients taking dabigatran with serious bleeding or undergoing a procedure, including 53 patients with ICH, idarucizumab 5 g (administered as two 2.5-g boluses) rapidly led to complete reversal of dabigatran (based on diluted thrombin time or ecarin clotting time) independently of age, sex, and renal function, with thrombotic events occurring in 5% of the patients with ICH.^168,190 Unfortunately, imaging studies were not mandated in the ICH populations, and there are no data on clinical outcomes in the ICH population except that 17% of the patients with ICH died within the first 5 days. A number of real-world case series in both the United States¹⁹¹ and Europe^175,192,193 have shown similar rates of mortality and acceptable incidence of thrombotic events, suggesting a therapeutic effect of idarucizumab after ICH. The absence of a control group and lack of imaging data limit any conclusions on clinical efficacy.

In healthy volunteers taking factor Xa inhibitors at doses of 37.5 to 50 IU/kg, 4-F PCC reverses coagulation assays.^194–198 A meta-analysis of 10 single-arm case series included 251 patients with factor Xa inhibitor–related ICH given 4-F PCC. Effective hemostasis was seen in most cases with acceptable mortality rates and thrombosis risk.¹⁷¹ In a multicenter observational study including 172 patients with factor Xa inhibitor–related ICH given 4-F PCC or aPCC, a high rate of hemostasis and low risk of thrombotic events (5% with aPCC, 3.3% with 4-F PCC) were seen.¹⁷⁰ In another multicenter retrospective case series, there were no differences in efficacy, mortality, or safety with aPCC and both low- and high-dose 4-F PCC in patients taking apixaban or rivaroxaban presenting with ICH.¹⁶⁹ In another case series, although factor Xa levels on admission were associated with HE, administration of PCC was not associated with differences in HE, mortality, or functional outcomes.¹⁹⁹ aPCC and 4-F PCC have not been directly compared in a randomized trial for factor Xa inhibitor reversal, although there is more evidence from observational studies to support the use of 4-F PCC, which is also more widely available.

In a preclinical in vitro study, adsorption of dabigatran by activated charcoal removes dabigatran from pooled human plasma.¹⁷³ It is advised that charcoal should be given within 2 hours of ingestion before intestinal absorption of dabigatran. In healthy volunteers given activated charcoal 2 and 6 hours after apixaban ingestion, apixaban absorption was reduced and half-life was significantly reduced.¹⁷⁴ Similarly, activated charcoal given to healthy volunteers up to 8 hours after ingestion of rivaroxaban significantly reduced exposure.¹⁷² Thus, activated charcoal is a supplementary treatment option in DOAC-associated ICH when the most recent dose was taken within the previous 2 to 8 hours in order to enhance elimination and neutralize the ongoing anticoagulant effect. Only single case reports exist in patients with hemorrhage, so it is not possible to comment on clinical outcomes.

In a prospective multicenter case series, 5 patients with dabigatran-associated ICH bleeding treated with aPCC (FEIBA, 50 U/kg) were compared with matched cases receiving supportive care. Patients treated with aPCC had favorable outcomes compared with the matched control subjects as assessed by treating physicians with no thromboembolic events.¹⁷⁶ In a small single-center retrospective case series, aPCC (FEIBA) was given to 16 patients with dabigatran-associated bleeding; no clinically significant HE was observed.¹⁷⁵ These case series, although small and limited by lack of control, suggest that aPCC can reverse dabigatran and may be considered when idarucizumab is not available. In vitro thrombin generation studies show that PCCs increase peak thrombin generation with variable effects on kinetic parameters and suggest that PCC at a dose of 50 IU/kg can produce hemostasis at therapeutic dabigatran levels.²⁰⁰ However, in healthy volunteers treated with DOACs, reversal with procoagulant concentrates (PCC or aPCC) did not fully restore levels of fibrin formation in studies with flowing blood, especially for dabigatran, suggesting potential limitations of the nonspecific PCC strategies to reverse DOAC-induced coagulopathy.^201,202

Dabigatran is excreted by the kidneys, and elimination is delayed in those with renal impairment. Therefore, RRT is able to decrease the plasma concentrations of dabigatran, although the effect of RRT on clinical outcomes is unclear. In a systematic review of the literature, including 11 patients with ICH who underwent RRT for dabigatran removal, patients had normal renal function or varying degrees of renal impairment. RRT in the form of hemodialysis (intermittent hemodialysis in 10 patients, continuous veno-veno hemodialysis in1 patient) was effective at reducing dabigatran concentrations. The majority of patients received PCC in addition to RRT. Recovery or rehabilitation was reported in the majority of patients, but a quarter died as a result of progression of intracranial bleeding. Half the patients had rebound of dabigatran concentrations after cessation of RRT.¹⁷⁷

Protamine binds to UFH and thus neutralizes the anticoagulant effect of UFH. Hence, in patients with UFH-associated ICH, intravenous protamine is reasonable to reverse the anticoagulant effect of heparin.¹⁷⁸ However, because UFH has a short half-life and protamine can cause hypersensitivity reactions and is a weak anticoagulant, caution is needed in the selection of the required dose.²⁰³ Intravenous protamine should not exceed 50 mg/10 min because of the risk of hypotension and bronchoconstriction; repeated smaller doses are preferable.¹⁷⁸

In a small retrospective case series of patients on LMWH, protamine only partially reversed the anticoagulant effect. The majority of patients had cessation of bleeding. Protamine only partially affected anti–factor Xa levels, which were of use to assess the amount of anticoagulant present but did not predict the effect of protamine.²⁰⁴ Therefore, intravenous protamine is reasonable to partially reverse the anticoagulant effect of LMWH.¹⁷⁸ Andexanet alfa has also been shown to significantly reduce anti–factor Xa levels in 16 patients taking enoxaparin.¹⁶⁶

Hemostatic expansion remains a therapeutic target after ICH. There is a lack of data on the clinical benefit of reversal of anticoagulation (eg, HE, functional outcome) compared with confirmation of reversal of anticoagulation parameters.

The clinical utility of anticoagulant tests for the DOACs is not established. The role of blood tests (eg, anticoagulant parameters, thromboelastography, point-of-care tests) to target reversal of anticoagulation therapy should be studied.

Choice of reversal agents for anticoagulation therapy-related ICH will continue to evolve as our understanding of efficacy, safety, and thromboembolic risk is better defined. Development and research of new anticoagulant reversal agents is encouraged. One new agent, ciraparantag (aripazine), is designed to be a universal antidote to factor Xa inhibitors, dabigatran, LMWH, and UFH.²⁰⁵ Phase II and III clinical trials are awaited.

There are limited data on when to administer idarucizumab relative to last dose of dabigatran and on the use of idarucizumab with PCCs and other blood products. Rapid hemostasis may not be ensured in patients with existing comorbidities or hypocoagulable states that impair clotting.

The potential synergistic benefits of a bundle of care, including BP lowering and reversal of anticoagulation, should be studied, as well as specific care pathways (eg, keeping reversal agents on the ward, not requiring consultation with hematology, training of nurses). Such pathways may reduce time to reversal of anticoagulants and improve outcome.

One moderate-size RCT studied patients with acute hypertensive basal ganglia hemorrhage and requiring emergency craniotomy for removal of the hematoma who were also receiving aspirin therapy. Results showed that transfusion of 1 U of previously frozen apheresis platelets before surgery, with or without an additional platelet unit 24 hours later, reduced postoperative rate and volume of hemorrhage.²⁰⁶ Platelet transfusion also was associated with higher activities of daily living (ADL) score and lower 6-month mortality. All patients screened were investigated with a platelet aggregation test to exclude those with aspirin resistance. The excluded patients did not receive platelet transfusions; however, their outcomes were similar to those of patients with sensitivity to aspirin and treated with platelet transfusions. Among the methodological limitations of this trial, SAEs were not reported in this population, cases with incomplete hemostasis during operation were excluded, nonuniform surgical procedures were performed, and the methodology of ICH volume determination was below the current standard. Platelet aggregation testing is rarely available on an emergency basis in clinical practice.

In 2 retrospective studies in patients with spontaneous ICH while taking antiplatelet agents,^207,209 treatment with desmopressin (0.3 µg/kg) was associated with reduced expansion of the hematoma in 1 of the studies.²⁰⁷ The latter study included all ICHs, of which 42% were intraparenchymal, but results were not provided for the subsets. In a third retrospective study in patients with spontaneous ICH while on antiplatelet agents, treatment with desmopressin (0.4 µg/kg) in combination with platelet transfusion did not reduce HE or improve functional outcome compared with usual care.²⁰⁸

A moderate-size RCT in patients with spontaneous supratentorial ICH and concomitant antiplatelet therapy who were not planned for surgical evacuation showed that 1 U platelet concentrate (2 U for adenosine-diphosphate receptor blockers) given for the purpose of reducing HE and thereby reducing death or dependence was associated with a shift toward worse functional outcome at 3 months, as measured with the mRS, and a borderline significant increase in risk of any SAE.²¹⁰ There was no reduction in the expansion of the intracerebral hematoma or in 3-month mortality. Although this study was open to enrollment of individuals taking either cyclooxygenase inhibitors (such as aspirin) or ADP receptor blockers (such as clopidogrel), only 5 of the 190 participants were taking ADP blockers alone, limiting the generalizability of the findings to antiplatelet agents other than aspirin. These findings do not apply to preoperative platelet transfusions.

Studies of antiplatelet-related ICH have mostly included only aspirin. It will be important to study the effects of reversal of other antiplatelet agents, especially the ADP receptor P2Y12 inhibitors.

Platelet transfusions appear beneficial for reversal of aspirin before craniotomy and hematoma evacuation, but it is not known whether this effect also pertains to other invasive procedures or surgeries such as EVD and minimally invasive surgery (MIS).

The apparent beneficial effect of platelet transfusions for reversal of aspirin before craniotomy in a Chinese population should be confirmed in other populations with rigorous volumetric data and with adverse event reporting.

The effect of desmopressin is uncertain because of the lack of RCTs, but such trials are ongoing.

Ticagrelor is not reversed by platelet transfusions. A monoclonal antibody against ticagrelor reversed inhibition of platelet function in healthy volunteers.²¹⁷ A phase III trial to evaluate the clinical effectiveness of this antidote is ongoing.

Numerous phase II dose escalation and pilot studies have been performed testing recombinant factor VIIa.^223–226 In a phase IIb RCT, a dose-dependent reduction in HE and significant reduction in poor functional outcome were seen with recombinant factor VIIa. There was no difference in serious thromboembolic events.²¹⁹ However, in a larger phase III study testing recombinant factor VIIa within 4 hours of ICH onset, despite significant similar modest limitation of HE with the 80–µg/kg dose, there was no difference in functional outcome at 3 months compared with placebo. There was a significant increase in arterial thrombotic events.²¹⁸ Meta-analysis of recombinant factor VIIa RCTs showed no benefit on HE, functional outcome, or SAEs.²²⁷ However, a secondary post hoc analysis found trends toward improved outcome with recombinant factor VIIa treatment among younger patients with ICH with smaller hematoma volumes and shorter onset-to-treatment intervals,²²⁸ raising the possibility that future studies might identify meaningful patient subgroups for treatment. In a pooled analysis of 2 RCTs that were halted early after they failed to achieve recruitment targets, the use of the CTA spot sign was not effective in predicting response to factor VIIa 80 µg/kg given within 6.5 hours of ICH onset, and there was no difference in expansion or functional outcome between the treatment groups.²²⁹

In a large phase III RCT, TXA led to a significant but modest reduction in HE and early death (within 7 days) but no significant difference in functional outcome.²²² TXA was safe with no increase in VTE and a reduction in SAE compared with placebo. The study had an 8-hour time window, and most patients were enrolled >3 hours after ICH onset. There was a significant interaction with baseline SBP, showing a favorable shift in outcome with TXA in participants with baseline SBP <170 mm Hg. In 2 small phase II trials in patients with positive CTA spot sign and including black hole sign and blend sign in 1 trial, there was no significant difference in HE or functional outcome at 3 months.^220,221 In a recent meta-analysis including these 2 RCTs, TXA demonstrated a reduction in HE predicted by markers on CT scan but no difference in mortality or functional outcome.²³⁰

The time window for administration of hemostatic therapies remains uncertain. Specifically, it will be important to determine whether rapid administration of hemostatic therapy (such as recombinant factor VIIa or TXA) limits HE, reduces mortality, and improves functional outcome. Larger trials of these hemostatic therapies with earlier treatment windows are underway.

Another important goal is to identify patients at risk of HE (with factors other than time) who may still have the potential to benefit from hemostatic therapies. This could potentially include imaging markers to predict HE (eg, spot sign, blend, black hole sign) or other imaging factors (IVH, volume) or blood tests (eg, thromboelastography, glial fibrillary acidic protein) to select patients most likely to benefit from hemostatic therapies.

It is unknown if there is an ICH volume threshold above which limiting HE does not translate into clinical benefit.

Another important knowledge gap is determining if there is a potential synergistic effect of combined BP lowering and hemostatic therapy.

ICH is a complex clinical event that has been shown to benefit from specially trained, multidisciplinary care. Meta-analysis has shown the benefit of dedicated care such as a stroke unit for both ischemic stroke and ICH. The benefits of multidisciplinary care likely relate to the complex and multifaceted clinical domains affected by ICH. Rehabilitation teams and specially trained nurses working together with physicians familiar with patients with ICH have been shown to improve outcomes and reduce mortality compared with a general medical ward.^231,232

In patients with ICH, initiating coagulopathy reversal and BP control before transfer is recommended to avoid delays in treatment.²³³ However, in cases when transfer is the priority according to clinical assessment, it is best not to delay transfer. A study bundled BP control and correction of coagulopathy among other care initiatives and achieved improvement in early delivery of intensive BP lowering, although no significant change in time to reversal of coagulopathy was achieved. Bundled care did improve 30-day survival, which was mediated predominantly by a decrease in DNAR orders and increased admission to the ICU, both of which are likely to improve attainment of anticoagulation reversal and BP lowering.²³³

Determining whether enlarged ventricles represent ICH/IVH–related hydrocephalus versus unrelated conditions such as central brain atrophy can be challenging. Furthermore, some patients with ICH with expansion of the ventricles will require ventriculostomy, whereas others may not. Clinical hydrocephalus, defined as a worsening clinical examination attributable to acute hydrocephalus from ICH, is associated with worsened prognosis.²⁴⁴ Patients who develop clinical hydrocephalus should be evaluated and treated with ventricular drain placement and ICP monitoring when appropriate. For centers without this level of support, transfer is recommended to reduce mortality.^233,234

Patients with ICH can have myriad issues that span multiple clinical domains and can trigger rapid clinical change, supporting the use of a multidisciplinary team care.^231,235,236 The highest-risk period for neurological decline is within the first 12 hours after the hemorrhage, with deterioration events becoming uncommon after 48 hours.^245,246 The ability to affect the patient’s clinical course often rests on the ability to detect changes in the neurological examination accurately and consistently. Patients with ICH can benefit from neuromonitoring by staff trained in neurological assessment. When detected in a timely manner, these neurological changes can lead to changes in management.^99,102

Patients may require invasive mechanical ventilation to ensure adequate airway protection and adequate gas exchange. It is reasonable to consider patients with decreased level of consciousness (eg, GCS score ≤8) or a declining neurological examination for invasive mechanical ventilation. Even in cases when the patient can demonstrate adequate gas exchange, neurological injury may limit the patient’s ability to protect their airway, heralding a need for invasive support. Not all centers are capable of caring for complex or unstable patients. Stabilization of the patient’s hemodynamic profile and airway takes priority before transfer unless the facility lacks this capability, in which case transfer to a higher level of care should not be delayed. There are no randomized data to support these recommendations on the transfer of patients with ICH. These concepts are not disease specific, but they are reasonable and likely represent the safest approach when considering a patient with ICH for transfer.

Determination of the appropriate level of monitoring for patients with ICH can be challenging.²⁴⁷ Patients with mild to moderate ICH may, under certain conditions, be monitored safely in a dedicated stroke unit or step-down unit. A prospective observational study of 10 811 consecutive patients with spontaneous ICH who were not comatose and did not require mechanical ventilation in the first 24 hours from admission found that treatment in a stroke unit was associated with improved functional outcomes compared with treatment in either an ICU (OR, 1.27 [95% CI, 1.09–1.46]) or a general ward and that mortality was higher in an ICU (OR, 2.11 [95% CI, 1.75–2.55]) or a general ward compared with a stroke unit.²³⁷ In a subgroup analysis of severely affected patients (NIHSS score 10–25), adjusted mortality was not different when stroke units were compared with neurointensive care units, although the odds of a poor outcome (mRS score >3) was significantly lower for patients treated in a neurointensive care unit (OR, 0.45 [95% CI, 0.26–0.79]). Other nonrandomized prospective studies that included patients with ICH have reported reduced fatality and improved outcomes, especially at longer-term follow-up.^238,248 A prospective study of 105 043 stroke admissions from the Swedish Stroke Register (2001–2005) reported decreased mortality and institutional living after 3 months for patients admitted to stroke units compared with other types of wards and a relative benefit for patients with ICH (hazard ratio [HR], 0.61 [95% CI, 0.58–0.65]).²³⁸ A systematic review and meta-analysis of 8 older RCTs (1993–2004) comparing stroke unit care with general ward care found that patients with ICH benefit at least as much as patients with ischemic stroke from stroke unit care in terms of reduced death and dependency.²³¹ Other retrospective studies have identified criteria that predict low to no occurrence of readmission to an ICU after initial admission to a stroke unit or step-down unit.^249–251 Criteria include low ICH volume (<20 mL), low NIHSS score (≤10), high GCS score (≥13), minimal or no IVH, and absence of uncontrolled BP and respiratory failure.

ICH is a complex clinical event that has been shown to benefit from specially trained, multidisciplinary care. Patients with moderate to severe ICH (suggested by volume ≥30 mL), IVH, clinical hydrocephalus, or infratentorial location carry an increased risk of clinical decline. These patients have been shown to benefit from a neuro-specific ICU compared with a general critical care unit in terms of reduced mortality, length of stay (LOS), and duration of mechanical ventilation and improved outcomes.^{235,236,239–241,252,253} The postulated reasons for the improvement in outcomes are varied and range from improved quality metrics to enhanced ability to detect neurological changes with specially trained nursing staff.^235,240 In 1 study, having a full-time intensivist was associated with a lower mortality rate.²³⁹

Neurosurgical intervention can alter the clinical course for a subset of patients with ICH. Several studies sought to provide guidance for which patients are best suited for neurosurgical clinical support. Patients with IVH or infratentorial location may benefit from availability of neurosurgical care.^{102,233,234,254} The difficulty in interpreting these data is that the neurosurgical contribution to care is rarely isolated. One study included neurosurgical consultation in a care bundle with BP control and anticoagulation reversal and found a significant benefit on mortality.²³³

There are several potential indications for neurosurgical evaluation in patients with supratentorial ICH. In an effort to assist clinical decision-making, several studies attempted to highlight clinical factors that conferred a need for neurosurgical evaluation.^{102,233,234,255} Patients with moderate to severe supratentorial ICH (identified in most studies by volume ≥30 mL or GCS score <8) may benefit from neurosurgical evaluation.^233,234,254

More prospective studies are needed to confirm which patients are best cared for in neuro-specific ICUs, stroke units, or step-down units.

Most studies of patients with ICH excluded those with limitations on neurocritical care interventions according to the patient’s goals of care. Limitations in life-sustaining treatments do not necessarily indicate comfort care only (Section 7.2, Decisions to Limit Life-Sustaining Treatment). There is opportunity to define the scope, efficacy, and outcomes for patients who have a priori directives for limited interventions in the context of ICH.

There are no data on ICH-specific recommendations for the optimal timing and care bundles for transfer to facilities with appropriate resources such as defining which patients with ICH need to be intubated before transfer.

From existing data, it has been difficult to ascertain the individual impact of specialized nursing care, neuro/critical care, neurosurgical care, BP control, and reversal of coagulopathy. From a practical perspective, these care clinicians and interventions are bundled in such a way that, for most high-volume treatment centers, it may not matter. However, for smaller centers that see a smaller number of patients with ICH, being able to provide optimal care for a subset of patients with ICH who do not need the entire bundle might be of value.

Caring for severely affected patients with ICH is challenging. There is limited understanding of and methodology for mitigating the distress of caring for patients with ICH on hospital staff.

The use of standardized order sets and protocols for prevention of complications is well established in the literature for all types of patient-specific care. The QASC trial (Quality in Acute Stroke Care) evaluated nurse-driven protocol implementation in 19 Australian acute stroke units from 2005 to 2010.²⁵⁸ This large trial showed evidence that early implementation of treatment protocols (within 72 hours of admission)—monitoring fever, hyperglycemia, and swallowing dysfunction in acute stroke units—was shown to decrease LOS, death, and disability of patients at 90 days,²⁵⁶ with sustained benefits on long-term survival at 4 years (>20%) compared with the control units.²⁵⁸ The use of integrated care pathways or multidisciplinary communication tools such as order sets or protocols improves timely assessments, clinical documentation, and communication and decreases LOS.²⁵⁷ Hospitals with higher use of standardized order sets and adherence to specific pathways are associated with overall decreased complication rates related to infection, pneumonia, and hyperglycemia for patients with stroke.²⁵⁹

The risk of death resulting from pneumonia for patients with stroke is ≈35%.²⁶² The use of a validated swallow assessment tool and standardized dysphagia screening protocols in conjunction with treatment protocols to manage fever and hyperglycemia was associated with reduced death and disability at 90 days in a single-blind cluster RCT.²⁵⁶ The ASSIST (Acute Screening of Swallow in Stroke/TIA) dysphagia screening tool was administered by a trained nurse or a speech pathologist. In a prospective open-label nonrandomized study, guideline-based protocols compared with conventional care also were associated with lower risk of pneumonia, mechanical ventilation, and 90-day mortality.²⁶⁴ Patients with a positive dysphagia screen have a significantly higher 5-year mortality rate,²⁶¹ and for this reason, early identification is key for not only good long-term outcomes but also survival. Comparison of rates of pneumonia between sites with formal swallow screening protocols and sites with no formalized screening found a significant difference of 2.4% (formal screen) versus 5.4% (no formal screen).²⁶² Implementation of a targeted nursing bedside swallow evaluation intervention has been shown to cut rates of pneumonia in half (6.5%–2.8%) at some sites²⁶⁵ and thus supports the need for nurse education in swallow assessment interventions. Two systematic reviews support decreased rates of stroke-associated pneumonia and inpatient deaths with early dysphagia screening, specialist-driven swallow assessment, and formal written protocols that are implemented before any oral intake.^260,263 All studies support early evaluation with a formal dysphagia screening tool.

There is evidence to suggest that patients with stroke may have up to a 30% risk of developing significant cardiac arrhythmias²⁶⁶ during their hospital admission. Evaluation with continuous cardiac monitoring for the first 24 to 72 hours of admission, the time frame in which many of these arrhythmias are seen, is reasonable, depending on the clinical severity of the ICH. Individuals of older age with a larger lesion (>5 cm) had a statistically significant higher likelihood of developing clinically relevant arrhythmias that may require acute intervention.²⁶⁶ Another study found that ≈25% of patients with stroke experienced cardiac arrhythmias and that most of these were in the first 72 hours of admission.²⁶⁷ Neither of these studies were predictive of long-term outcomes or mortality but rather provided information on monitoring and treatment implications for the patient. Common admitting interventions to consider are 12-lead ECG, troponin level, and placing the patient on continues cardiac monitoring on arrival.

Infectious complications are associated with poor long-term outcomes, including readmission within the first 30 days after ICH.^272,273 A study found that patients with IVH (P<0.001) and patients with ICH scores >2 also had higher risk of infectious complications (P=0.0014).²⁷¹ ICH score >2 was found to be a significant risk factor for infectious complications (OR, 1.7 [95% CI, 1.2–2.3]; P=0.02).²⁶⁹ These studies used these findings to guide risk assessment and diagnostic testing, which included chest radiographs; urinalysis; white blood cell counts; serum C-reactive protein; blood, urine, or sputum cultures; and, if indicated, cerebrospinal fluid. It is reasonable for all patients with ICH, especially those with larger hematomas, including IVH, to be monitored closely for fevers and signs of infection throughout the course of their hospital stay to reduce LOS, decrease mortality, and improve long-term functional outcomes. None of the studies provide prescriptive guidance on frequency of specific diagnostic tests or treatment of infectious processes.

Additional diagnostic tests for early identification of infectious processes are not routinely necessary for multiple reasons (eg, cost and patient comfort). There is some evidence to support that early markers such as albumin levels may be early predictors of patients at high risk of infection, but none is yet validated for clinical use.

There is a lack of data on prevention of infectious complications and interventions to reduce hospital-acquired pneumonia, especially in nonventilated patients with ICH.

More studies guiding additional follow-up and therapies in the postacute phase for patients with both ICH and cardiovascular disease would potentially provide benefit because their all-cause long-term mortality may be increased.

Growing evidence suggests that inpatient delirium can affect patients’ LOS and long-term functional outcome, although there are no tools specific to ICH-related delirium and no standards or specific interventions to affect this patient population. Until such ICH-specific delirium studies are performed, clinicians will commonly apply guidance that has been developed for ischemic stroke.²⁷⁴

The in-hospital incidence of thromboembolic complications in patients with ICH is ≈7%,²⁸⁸ and the risk of DVT is 4 times higher than in patients with acute ischemic stroke,²⁸⁹ attributable in part to the fear of worsening hemorrhage and initial contraindication to pharmacological prophylaxis. A network meta-analysis showed that IPC devices were more effective than compression stockings to reduce VTE in patients with acute ICH.²⁷⁶ The CLOTS Trial (Clots in Legs or Stockings After Stroke) 3 was the largest of the RCTs, even when considering the subset of the study population who had hemorrhagic stroke (13%).²⁷⁵ In the entire study population, there was a reduction of symptomatic and asymptomatic proximal DVT compared with control, although the reduction was statistically significant only for asymptomatic DVT. There was a trend toward reduced mortality in the IPC group. In this study, it was also observed that patients in the IPC group had increased risk of skin breaks.²⁷⁵

In a meta-analysis of 2 RCTs and 2 observational studies with a total of 1000 patients with ICH, prophylaxis with any type of heparin versus compression stockings (3 studies) or control (1 study) resulted in a significant reduction of the risk of PE with a nonsignificant increase in the risk of HE and no significant difference in DVT or death.²⁷⁸ The heparin regimens and the time of initiation of pharmacological prophylaxis (24–96 hours from admission) differed between the studies. The duration of follow-up was only 10 days in 1 study²⁷⁷ and was not reported in another study.²⁸⁰ In a more recent meta-analysis of 9 studies and >4000 patients but addressing only safety outcomes, prophylaxis with any type of heparin was not associated with a significant increase of HE or extracranial hemorrhage, an increase in mRS scores of 3 to 5, or an increase in numbers of Glasgow Outcome Scale scores of 2 to 3.²⁷⁹ Only 1 of the 9 studies had a low probability of bias.

There is clear indication for beginning VTE prophylaxis after ICH, with the goal of selecting the optimal post-ICH timing that maximizes benefits of VTE prophylaxis while minimizing risk of promoting ICH expansion. One small RCT²⁷⁷ and 2 larger retrospective studies^281,282 addressed the timing of first dose of UFH or LMWH prophylaxis after ICH in terms of safety. The incidence of rebleeding or HE was not higher in the early start versus the delayed start group in any of the studies. An important point is that, in the retrospective studies, those with larger hematomas tended to be selected for later start times. The early start was 4 days (versus 10 days) after the ICH diagnosis in the RCT,²⁷⁷ a median of 42 hours after admission in the larger retrospective study (comparing initiation of VTE prophylaxis within 48 hours of admission versus >48 hours),²⁸¹ and within 48 hours from symptom onset in the smaller retrospective study.²⁸² The earliest start for any patient in these studies was 25 hours after admission. In a multivariable analysis, the hematoma size, but not timing of prophylaxis, was independently associated with HE.²⁸² It may be reasonable to first document hemorrhage stability on CT if LMWH prophylaxis is started in the 24- to 48-hour window after ICH onset. In another large observational study with start of prophylaxis (UFH or LMWH) 0 to 1 days after CT demonstrating stability, intracranial hemorrhagic complications were observed in 1.7%.²⁹⁰

A meta-analysis of 2 RCTs and 2 observational studies showed that graduated compression stockings, which were used in 3 of the comparator groups, were less effective than pharmacological prophylaxis to reduce PE.²⁷⁸ In a large RCT (CLOTS Trial 1), thigh-length compression stockings were not more effective than control to reduce the risk of DVT in patients with stroke, although only 9% of those had ICH.²⁸³ In a second large RCT, again with only 12% of cases having hemorrhagic stroke, thigh-length stockings were more effective than knee-length stockings in lowering the incidence of DVT, but the reduction was significant only for asymptomatic proximal DVT.²⁸⁴ Compression stockings were less effective than IPC to reduce VTE in a network meta-analysis of 3 studies and a subset from the CLOTS Trial 3, focusing on patients with ICH.²⁷⁶ The design of all the studies included screening with compression ultrasound for DVT, thereby also including asymptomatic events in the efficacy outcome.

In the RIETE registry (Computerized Registry of Patients With Venous Thromboembolism) of >40 000 patients with VTE, a subset of 344 cases with IVC filter insertion attributable to high risk of bleeding were matched with an equal number of patients without IVC filter with the use of propensity scores.²⁸⁵ The 30-day all-cause mortality did not differ between the groups, but those with IVC filter had a lower risk for PE-related death and higher risk for recurrence of VTE. The number of patients with high bleeding risk attributable to recent ICH is not provided.

With regard to treatment of VTE, in a retrospective cohort study of 2902 patients with spontaneous ICH, VTE was diagnosed in 3% of the cases, but this complication was independently associated with an mRS score ≥4 at discharge and at follow-up after 3 months.²⁹¹ In a small retrospective study of patients with traumatic ICH, UFH or LMWH was initiated for treatment of VTE when the neurosurgeon felt this was safe, on average 13 days after admission, and only 1 patient experienced minimal expansion of the ICH.²⁸⁶ In a second, slightly larger retrospective study of traumatic ICH and VTE, those with progression of the hematoma had anticoagulant therapy initiated after a median of 5.5 days from the injury, whereas those without expansion had their anticoagulation started after a median of 10 days.²⁸⁷ In this study, only 40% of the hemorrhages were intraparenchymal. Factors that should go into the consideration of the timing of anticoagulation are size of the hematoma, patient age, and extent of the VTE.

It is unknown whether graduated compression stockings or IPC devices increase the efficacy of VTE thromboprophylaxis when added to pharmacological prophylaxis or allow greater delay in initiating pharmacological prophylaxis in patients with acute ICH.

It is uncertain whether IPC devices reduce the risk of symptomatic DVT or improve functional outcomes in patients with acute ICH.

There is currently insufficient evidence to determine the safety of LMWH prophylaxis during the first 48 hours after ICH onset. A question that should be tested is whether demonstration of stability of the hematoma by repeat imaging is useful for deciding on the safety of initiation of pharmacological prophylaxis 24 to 48 hours after onset of symptoms.

A large prospective study comparing 2 time points for initiation of pharmacological prophylaxis in patients with ICH should be performed.

Prophylactic insertion of IVC filters was shown to lack benefit in a large RCT in trauma patients, but data are lacking for patients with spontaneous ICH.

The effectiveness of IVC filters specifically in patients with ICH and early onset of VTE has not been studied.

The earliest time point for anticoagulant treatment of VTE in patients with spontaneous ICH is not well established because the studies were in trauma-associated ICH. Timing of initiation of anticoagulation for VTE in the presence of an EVD and after surgical decompression also has limited data and high practice variability.

Future studies should address whether anticoagulation for VTE in spontaneous ICH should be started with full therapeutic dose or with gradual increases of the dose.

Future studies should address whether anticoagulation for VTE in spontaneous ICH should be started with UFH, LMWH, or DOACs.

Frequent occurrence of early ND in patients with ICH is well established in multiple studies, and for this reason, ED nursing neurological assessments must be reliable and frequent. There are multiple assessment tools from which to choose, but one of the easiest and most universal is the GCS. The GCS allows straightforward evaluation of mentation and recognition of decline in patients with ICH. Proper training is required to assess this scale. One study identified patients with ICH in the prehospital to early postarrival stage as more likely to have ultraearly neurological decline compared with patients with ischemic stroke (30.8% versus 6.1%).²⁹³ These patients with ultraearly neurological decline and early ND have increased mortality and poor functional outcomes at 90 days. Timely interventions from nurses and physicians are driven by early identification of ND by the nurse through robust, reliable, and frequent GCS examinations. The frequency of neurological assessments depends on both physical location and clinical condition of the patient. The study that identified ultraearly neurological decline performed 3 serial GCS evaluations in the ultraearly time period (first 2.5 hours since onset), during initial prehospital assessment, at initial ED arrival, and early in the ED course.

Frequent neurological and vital sign assessments of patients with ICH are indicated to capture ND and prevent secondary complications. One study found that nursing examination discovered up to 54% of ND leading to intervention (ie, surgery or placement of ventriculostomy) versus 46% of ND identified by neuroimaging changes.¹⁰² These data highlight the opportunity and impact that nursing examinations have on patient care and potential outcomes. Studies indicate that patients are at highest risk of ND in the first 12 to 24 hours of ICH onset and up to 72 hours after admission.^102,292 In a prospective observational study of hourly neurological checks in a neurocritical care unit, change in GCS score within the initial 12 hours was a significant predictor of worse functional outcome at 90 days.²⁴⁵ In the ICU, especially for patients with ICH of higher clinical severity, neurological assessments are typically performed hourly for the first 24 hours or until the ICH is stable. However, around-the-clock nursing interventions run the risk of ICU delirium and sleep deprivation, which may have further negative impact on patient functional outcome, cognition, and quality of life.^246,298 Staff training and care plans should be individualized to illness acuity with consideration of the need for frequent neurological assessments in the acute phase.

Nurse stroke competencies are a hallmark of providing evidence-based care. Growing literature supports the need for standardizing formal training for nurses caring for patients with ICH. To date, few studies have compared outcomes or quality of care between nurses with formal competency training and those without such training. One study found lower death rates among patients with stroke admitted to teaching hospitals, with an increased number of doctor and nursing specialists and increased nursing resources.²⁹⁵ Although the study suggests that nurses at teaching hospitals with more available staff may affect mortality outcomes, it does not clearly define nurse-driven stroke care competencies. Another study highlights increased nursing stroke care knowledge and increased compliance to stroke care guidelines with the introduction of a formal stroke competency program.²⁹⁶ During the analysis of this intervention, it was found that nurses who held specialized certifications scored better in adherence to protocols and knowledge assessment. These data highlight the opportunity for organizations, hospitals, and stroke teams to consider the development of a stroke competency training program and to foster and encourage more nurses to apply for specialized certification.

The benefit versus risk of frequent nursing neurological and vital sign assessments is not well established in the literature, leaving a wide range of recommendations. Opportunities exist for clearer delineation of the time frame in which patients should be receiving hourly nursing assessments and criteria that help to establish when frequent monitoring is no longer of value and may affect recovery.

The effects of nursing intervention on cerebral hemodynamics are poorly understood. Nursing care is multifaceted and wide-ranging, depending on the needs of the patient, and can include position changes, oral care, neurological and physical examinations, and wound care. Research evaluating the impact of clustered nursing care in the ED, ICU, and stroke unit on patient outcomes is needed.

Many studies provide evidence that ND, early or delayed, is prevalent in the ICH patient population. No studies have addressed how nursing actions may help to prevent ND. This is poorly understood and leaves a gap in guiding nursing care in what type of preventive measures may reduce ND in the acute phase of ICH.

Caring for severely affected patients with ICH is challenging. The potential distress of perceived inappropriate care in nurses is an important topic for future research.

Monitoring serum glucose is important because it can provide an opportunity to intervene in the event of hyperglycemic or hypoglycemic events.^256,299 The QASC study, a single-blind cluster RCT, investigated an intervention of treatment protocols to manage fever, hyperglycemia, and swallowing dysfunction compared with no intervention and found that patients in the intervention group were significantly less likely to be dead or dependent at 90 days, although the impact of each specific intervention could not be determined.²⁵⁶

No trials have analyzed the effects of untreated hypoglycemia given the known acute clinical risks. The range of blood sugars outlined (<40–60 mg/dL, 2.2–3.3 mmol/L) reflects the thresholds for treatment for hypoglycemia in studies reviewed for this guideline.^{299–301,309} The NICE-SUGAR trial (Normoglycemia in Intensive Care Evaluation and Surviving Using Glucose Algorithm Regulation) randomly assigned patients in the ICU to either intensive glucose control (target, 81–109 mg/dL) or conventional glucose control (<180 mg/dL) and found that intensive glucose control resulted in increased all-cause mortality at 90 days.²⁹⁹ An important finding was that severe hypoglycemic events (glucose ≤40 mg/dL) were significantly more common in the intensive control group compared with the conventional control group (6.8% versus 0.5%). It was unclear whether lower blood glucose levels, higher administration of insulin, or other factors accounted for this finding. However, in a prospective observational study of patients with severe brain injury, tight systemic glycemic control (80–110 mg/dL) was associated with low cerebral microdialysis glucose and brain energy crisis, which were correlated with increased hospital mortality.³⁰⁰ Balancing the risks of hypoglycemia and hyperglycemia, both of which may worsen outcomes in patients with ICH, may justify treating low blood glucose at higher thresholds than studied in general critical care populations. The risk of treating hypoglycemia is exceedingly low, and treatment is highly recommended despite a low quality of evidence.

No trials have evaluated untreated hyperglycemia, rendering the data for this approach limited. Hyperglycemia appears to be an independent predictor of poor outcomes. However, the relationship among serum glucose, the timing of that measurement, and the presence/absence of comorbid diabetes remains unclear.^78,302–307 The optimal glucose level at which treatment should be initiated and the target range are not clear because the upper limit of tolerable hyperglycemia varies between studies. If carefully approached, the risk of treating moderate to severe hyperglycemia should be relatively low and outweighed by the potential benefit. However, in the NICE-SUGAR trial, in patients receiving general critical care, a blood glucose target of <180 mg/dL was associated with lower mortality than a target of 81 to 108 mg/dL, suggesting that targets for treating hyperglycemia should be less intensive in critically ill adult patients.²⁹⁹ In most studies, hyperglycemia is managed by either a subcutaneous insulin or an intravenous insulin infusion protocol.

Optimized serum glucose targets and the optimal agents for glucose control in patients with ICH have not been defined.

The relationship among serum glucose, diabetes, and functional outcomes in patients with ICH remains unclear.

There is a paucity of data on the impact of postprandial glycemic response in patients with ICH and the effect on outcomes.

Fever in patients with ICH has been associated with worse outcomes.^318,321 Treating fever seems reasonable; however, there is less evidence that therapeutic temperature modulation improves outcomes.^315,316,320 One pilot study of therapeutic temperature modulation with a surface device for fever with a normothermia target observed no improvement in outcomes but reported increased duration of sedation, days of mechanical ventilation, and ICU LOS.³²⁰ Because of the variability in definitions of fever used in the literature (ranging from 37.7 °C/99.5 °F to 38.3 °C/100.9 °F), this guideline uses the term elevated temperature. In addition, there is significant variability in the literature on the approach to addressing fever, for example, pharmacotherapy versus catheter-based thermal management.³¹² A multicenter RCT in patients with stroke (ischemic plus hemorrhagic) randomly assigned treatment with paracetamol for body temperature 36 °C to 39 °C within 12 hours of symptom onset and reported no improvement in expected functional outcome except in a post hoc analysis of patients with baseline temperature of 37 °C to 39 °C.³¹³ In a prospective database, pharmacological treatment of temperatures ≥37.5 °C for 48 hours was associated with an increased likelihood of a good outcome at 3 months.³¹¹ Another RCT comparing catheter-based normothermia with a target temperature of 36.5 °C against conventional step-wise fever management with anti-inflammatory drugs and surface cooling reported a significant reduction in fever burden for catheter-based normothermia but no significant differences in mortality or long-term outcomes.³¹² Therefore, clinical trial evidence does not support a benefit of therapeutic temperature modulation, either surface devices or catheter-based normothermia, although pharmacological treatment of fever may be associated with improved outcomes.

Therapeutic hypothermia (35 °C/95 °F to 36.5 °C/97.7 °F) may be a physiologically reasonable approach to reducing perihematoma edema but has not been demonstrated to be clinically beneficial. Interpretation of the data is limited by small pilot cohorts, historical control subjects, and nongeneralizable samples such as only those with large-volume hemorrhages. Most of the available data were evaluated primarily with descriptive statistics. In 2 small pilot studies, therapeutic hypothermia was associated with high survival rates and maintenance of stable perihematomal edema volume.^314–316 However, therapeutic hypothermia is not without risk and should be considered of unclear benefit.^{315–317,324}

Treating fever in patients with ICH and improving outcomes remains an opportunity for future research. It is possible that some of the early data have been limited in this regard because only recently has research started considering health-related quality of life.

The maintenance of normothermia in patients with ICH has not been demonstrated to clearly improve outcomes and is a potential therapeutic opportunity.

Perihematomal edema remains an important concern in patients with ICH. Whether temperature modulation improves edema or functional outcomes remains unclear.

Mortality, a prespecified secondary analysis in MISTIE III, was significantly lower in the MIS group compared with the standard medical care group at 7, 180, and 365 days, although the trial was neutral on the primary outcome (functional outcome benefit).³⁸¹ Although smaller, likely underpowered RCTs did not always show a mortality benefit for MIS,^{392,400,401,406,408} most meta-analyses comparing stereotactic puncture or endoscopic drainage with standard medical care reported significantly decreased odds of death with any MIS compared with standard medical care.^{380,382,386–390,394} Multiple safety end points were addressed in the MISTIE III trial, including symptomatic hemorrhage within 72 hours after last dose of alteplase and bacterial brain infection, which were similar between groups, indicating that stereotactic aspiration with thrombolysis appears to be safe.³⁸¹ SAEs at 30 days were significantly lower in the MIS group versus the standard medical care group. Only asymptomatic bleeding was higher in the MIS versus the standard medical care group (32% versus 8%). Other RCTs and meta-analyses confirm no significant difference in safety end points (brain rebleeding after treatment and infection) for endoscopy and stereotactic aspiration/craniopuncture techniques compared with standard medical care or craniotomy.^{391,398,400,405,406,408,411–413} Most RCTs enrolled patients <80 years of age, although age did not modify the effect of surgery except in 1 meta-analysis in which improved outcomes from any surgery for ICH were found for patients 50 to 69 years of age.³⁹³

Studies comparing MIS with conventional craniotomy have shown improved outcomes with a less invasive approach, raising the possibility that open craniotomy may damage more brain tissue while removing blood. Both small RCTs^{389,398,399,411,412,414} and all meta-analyses of either clinical trials alone or combined with observational studies from different settings comparing stereotactic puncture or endoscopic drainage with craniotomy have shown significantly decreased odds of functional dependence (or combined with death) and increased odds of good functional outcome with MIS.^{383,385–387,389,390,394,395} A network meta-analysis suggested the highest ranking of favorable prognosis for stereotactic aspiration, followed by endoscopy, then craniotomy, and last standard medical care.³⁸² RCTs comparing MIS with craniotomy have included patients with ICH volume >25 mL and time interval to surgery from <6 to 72 hours after presentation. In the early surgery study, MIS showed a functional outcome benefit compared with craniotomy only if the CTA spot sign was positive but also showed a higher risk of rebleeding.³⁹⁹

Many small RCTs of MIS show a functional outcome benefit from MIS compared with standard medical care at follow-up times of 3 months to 1 year.^{391,392,396–399,401,403,408,412,413} In the MISTIE III trial, stereotactic aspiration plus irrigation with alteplase did not improve functional outcomes at 1 year compared with standard medical care in patients with ICH volume >30 mL.³⁸¹ However, planned exploratory analyses of clot removal showed a significant association between extent of clot removal and both mortality and lower mRS score (0–3), specifically in those patients who achieved the surgical aim (end-of-treatment clot size ≤15 mL). Meta-analyses of this and smaller clinical trials and observational studies from different settings comparing stereotactic puncture or endoscopic drainage with standard medical care have shown improved functional outcomes (alone or together with survival) with MIS.^{379,380,382–390,394} Most RCTs included only ICH volume >20 mL, although several included ICH volumes as low as 10 mL. One meta-analysis found that MIS was most beneficial for patients with hematoma volume between 25 and 40 mL and with a GCS score ≥9,³⁸⁷ whereas MISTIE III and 2 other meta-analyses found that hematoma volume did not modify the effect of surgery.^379,381,384

Current evidence does not support specific recommendations for selecting candidates for surgery. A priori analyses focusing on clinical details, hematoma volume, patient age, GCS score (baseline clinical severity), and follow-up timing would inform future clinical trial design and recommendations.

RCTs of MIS have not addressed a priori questions about timing of surgery and intent to stabilize ICH before surgery. Optimal time to surgical treatment with MIS remains a controversial issue primarily because of the risk of rebleeding, although reducing hematoma volume early (<12 or 24 hours) may reduce secondary brain injury and improve outcomes with no effect on bleeding risk as suggested by observational data. Several RCTs are underway that will address aspects of these questions.

Although a functional outcome benefit of MIS compared with conventional craniotomy is reported for many RCTs, a mortality benefit is uncertain and may reflect the practice to perform craniotomy but not MIS in deteriorating patients. Most small RCTs are underpowered and did not show a mortality benefit of MIS compared with craniotomy; however, most meta-analyses of smaller clinical trials and observational studies comparing stereotactic puncture or endoscopic drainage with conventional craniotomy showed significantly decreased odds of death with MIS.

Currently, no adequately powered clinical trial data compare different devices for MIS in ICH. Although surgeon experience and ability to achieve adequate hematoma removal with low rebleeding risk and acceptable outcomes may prove superior to a single technique, ongoing innovation with the development of new surgical devices will require comparisons of endoscopic and stereotactic techniques with thrombolysis and with potential for intrahematomal delivery of therapeutic agents. Ongoing RCTs will add useful data to these questions.

In patients with moderate to large IVH and higher clinical severity (defined in a propensity score–matched analysis as GCS score <13, ICH volume >11 mL, and Graeb score ≥7 [indicating moderate to severe IVH^348]), EVD placement alone is associated with improved survival compared with conservative treatment.^347–349 In a large retrospective analysis with propensity score matching, EVD use was associated with higher survival in patients with severe ICH as defined above, although not overall.³⁴⁸ A smaller retrospective analysis found a positive association of EVD alone with survival at hospital discharge in patients presenting with hydrocephalus and a GCS score >3 after adjustment for clinical severity.³⁴⁷ There were no age limits on these studies.

In patients with IVH obstructing the third or fourth ventricle and small- to moderate-volume ICH (<30 mL), controlled irrigation with a thrombolytic agent such as alteplase or urokinase improves survival in patients with clinical hydrocephalus requiring a routinely placed EVD. Mortality, a prespecified secondary analysis in CLEAR III, was significantly lower in the EVD plus alteplase group compared with the EVD plus saline group at 180 days.⁴¹⁶ Smaller RCTs also have shown a mortality benefit for IVT,^{356,423–425} and all meta-analyses of RCTs with or without observational data comparing EVD alone or with saline with EVD plus alteplase or urokinase reported significantly decreased odds of death with IVT.^415,417,418 Multiple safety end points were addressed in CLEAR III, including symptomatic hemorrhage, which was not different between study groups. Both bacterial ventriculitis and SAEs were significantly less frequent in the alteplase group versus the saline-administered group, indicating that IVT appears to be safe.⁴¹⁶ Other RCTs and meta-analyses confirm no significant difference in safety end points (rebleeding after treatment and ventriculitis) for IVT compared with EVD alone or with saline irrigation.

It is not clear whether EVD plus IVT improves functional outcomes. In CLEAR III, EVD plus irrigation with alteplase did not improve functional outcomes at 180 days compared with EVD plus saline in patients with obstructive IVH and ICH volume <30 mL.⁴¹⁶ However, a low proportion of participants achieved near-complete clot removal, and functional benefit was reported from removing greater amounts (>85%) of IVH volume. Alternatively, the absence of benefit of IVT on functional outcome in clinical trials might also be attributable to cerebral injury associated with parenchymal hemorrhage. CLEAR III included a high proportion of patients with thalamic ICH, a location with poor prognosis. In CLEAR III, a greater proportion of patients in the IVT arm had severe disability (mRS score 5) at 180 days, suggesting that mortality reduction occurred at the expense of severe morbidity. Meta-analyses of this and smaller clinical trials and observational studies comparing IVT to EVD (with or without saline) have shown heterogeneous effects on functional outcomes from IVT, depending on time of follow-up and functional outcome scale used.^{415,417,419,426} Most RCTs included patients up to 75 or 80 years of age. CLEAR III excluded patients with anticipated early withdrawal of life-sustaining therapies. For patients being considered for IVT, shared decision-making between physicians and family members is recommended to weigh mortality and functional outcome benefits with consideration of patient preferences.

Compared with conservative treatment, there is uncertainty over whether EVD alone improves functional outcomes. In a systematic review of studies including patients with nontraumatic IVH secondary to ICH or subarachnoid hemorrhage and Graeb score >7, EVD alone was not associated with return to an independent lifestyle.³⁴⁹ In a large retrospective analysis with propensity score matching, EVD use was not associated with functional outcome at discharge.³⁴⁸ Subgroup analysis by several clinical severity factors found that patients receiving an EVD had more disability on the mRS compared with patients who did not receive an EVD. In this retrospective cohort, it is possible that patients who received an EVD were more severely disabled at presentation, thus requiring an EVD, versus those who did not. Moreover, patients who may have died without EVD placement also may have worse outcomes. A smaller retrospective analysis found a positive association between EVD alone compared with no EVD and good outcome at hospital discharge.³⁴⁷ However, retrospective studies are unable to evaluate unmeasured confounders that contribute to the decision to place EVDs in patients with IVH.

Endoscopic surgery for hypertensive IVH combined with EVD with or without IVT has been studied in small RCTs and observational studies.^{419,420,427,428} Small RCTs have reported no significant difference in mortality rate, and 2 of them reported improved short-term functional outcomes for the endoscopic group compared with the EVD group.^427,428 One meta-analysis reported higher IVH evacuation rate, lower mortality, improved functional outcomes, and lower permanent shunt rate for endoscopic surgery plus EVD compared with EVD plus IVT.⁴²⁰ No conclusive evidence was provided comparing endoscopic surgery with EVD alone. A network meta-analysis reported improved survival and functional outcomes for endoscopic surgery compared with EVD plus alteplase or urokinase, all of which were superior to EVD alone.⁴¹⁹ Lower rates of permanent shunting, intracranial rebleeding, or infection in the endoscopic surgery group suggest that this intervention seems safe, although no large high-quality RCTs directly comparing these interventions have been performed and risk of publication bias is high.

Current evidence does not support specific recommendations for selecting patients with IVH for EVD in terms of timing or volume of IVH; EVD insertion rates vary widely between hospitals and regions. One retrospective analysis found that small IVH volume (Graeb score ≤2) not associated with obstructive hydrocephalus was not associated with unfavorable outcome or death after ICH, whereas a Graeb score >2 was independently associated with unfavorable outcome and higher mortality.

Exploratory analyses of CLEAR III suggest associations of improved functional outcome in alteplase-treated patients with larger IVH volumes and randomized earlier after symptom onset. A priori analyses focusing on clinical details (IVH volume and time to initiation of thrombolytic treatment) would inform future recommendations.

RCTs of IVT and endoscopy have not addressed a priori questions about adequate removal of IVH and optimal timing of the intervention. Further research is needed to determine functional outcome benefit of near-complete IVH evacuation compared with targeting opening of the lower ventricular system and resolution of hydrocephalus and intracranial hypertension.

Currently, no adequately powered clinical trial data compare different surgical approaches for evacuation of IVH. Are endoscopic techniques superior to EVD plus IVT, and is addition of lumbar drainage superior to EVD alone plus IVT for outcomes or avoidance of permanent shunting?

Craniotomy for ICH of volume >10 mL in patients with significant neurological deficit remains of uncertain benefit compared with conservative management. Both STICH I and STICH II demonstrated no benefit in functional outcome with craniotomy in situations in which the treating neurosurgeon was uncertain about the benefits of either treatment.^429,431 A patient-level data meta-analysis performed contemporaneously suggested that certain cohorts might benefit,³⁹³ and a smaller (n=108) single-center RCT found that craniotomy improved functional outcome.⁴³⁰ Three meta-analyses published in 2020 provide mixed results: 2 meta-analyses suggest a benefit in functional outcome and mortality with any surgery,^382,384 and 1 meta-analysis found no benefit in functional outcome or mortality.³⁸⁰

Despite the unclear value of craniotomy to improve overall functional benefit or mortality, limited data suggest that craniotomy for hematoma evacuation might be considered as a lifesaving measure in patients who are deteriorating. STICH II found a trend toward improved mortality with surgery, despite a 21% crossover rate from medical therapy to surgery, 74% of which were attributable to deterioration.⁴²⁹ Individuals who crossed over had deeper coma with worse neurological deficits than those in the early surgery group and had worse prognosis compared with individuals who did not cross over, but their surgery did not affect trial results, which were analyzed by intention to treat.⁴²⁹ This suggestion of a mortality benefit was further supported by a recent small (n=61) RCT that demonstrated improved mortality with surgery⁴³² and 2 meta-analyses that suggest a possible mortality benefit.^382,384 Therefore, given the crossover attributable to deterioration observed in STICH II and the data suggesting a possible mortality benefit, for patients who are deteriorating, craniotomy for hematoma evacuation may be considered as a life-saving measure.

The potential impact of timing of craniotomy for ICH on outcome remains debated. Although STICH I and II did not identify an early time effect, a significant majority of enrolled patients underwent surgery >12 hours from onset, and those with surgery <12 hours from onset were likely secondary to severe presentation or deteriorating status. A late time threshold, however, was identified in the STICH I and II cohorts, with expectations of worse outcome beyond 62 hours. Only 2 single-center RCTs have been performed that required surgery within ≤12 hours from onset. The study by Morgenstern et al,⁴³⁷ although not powered for efficacy (n=34), found a promising mortality signal when surgery was performed within 12 hours. More encouragingly, Pantazis et al⁴³⁰ (n=108) demonstrated a benefit in functional outcome when surgery was undertaken within <8 hours. Future multicenter research evaluating the benefit of surgery within 12 hours may clarify this knowledge gap.

In an individual-patient data meta-analysis, for patients with spontaneous cerebellar hemorrhage without brainstem extension, hematoma evacuation was not significantly associated with improved functional outcomes at 3 months but was associated with survival benefit at both 3 and 12 months.⁴⁴² Mortality benefit occurred for patients with larger hematoma volumes (>15 mL), whereas volumes <12 mL were associated with lower likelihood of good outcome with surgery. A systematic review of 41 studies (37 retrospective and 4 prospective) reported no significant association of surgical evacuation with either mortality or functional outcomes at 6 months but, because of a large proportion of retrospective studies, suffered from a high risk of bias.⁴⁴³ A large retrospective review found that pooled mortality rates were lower in patients treated with surgery compared with conservative treatment but functional outcomes were more favorable with nonsurgical management.⁴⁴⁴ This may reflect variable indications for surgery. Another study reported mortality reduction with surgery in cases with hydrocephalus, but not without, indicating the importance of treating hydrocephalus.⁴⁵⁰ One retrospective study reported trends for improved mortality and functional outcome for suboccipital decompression and hematoma evacuation compared with evacuation alone.⁴⁴⁸ Most studies support a lifesaving benefit from surgery under conditions of a deteriorating clinical examination, impending brainstem compression, clinical hydrocephalus with fourth ventricle obstruction, and radiographic obliteration of basal cisterns.^{442,445–448,450,451}

A perceived lack of equipoise concerning the lifesaving benefits of surgical evacuation for cerebellar ICH most likely precludes the design of future randomized trials to address the question of surgical versus conservative management. The efficacy of surgical evacuation for improving functional outcomes remains uncertain.

Previous studies have not addressed a priori questions about timing of surgery for cerebellar ICH and specifically whether initial conservative treatment compared with immediate surgical evacuation is preferable in patients with cerebellar ICH >3 cm/15 mL who are in a good clinical condition. For such patients, a retrospective study reported that an initial conservative approach often leads to good outcome and that there may be a subgroup of patients in whom surgery can be safely deferred. The optimal timing and indications of surgical treatment in large cerebellar ICH with good clinical condition are worthy of further study.

Currently, no adequately powered studies have compared different surgical approaches for cerebellar ICH. Several small retrospective studies compared endoscopic evacuation or stereotactic aspiration with standard suboccipital craniectomy, with variable efficacy. Comparison of MIS techniques with suboccipital hematoma evacuation with or without decompression is an important topic for future clinical trials. Further investigation also is needed to determine whether MIS in patients with >15-mL cerebellar ICH volume and good clinical condition improves functional outcome compared with best medical treatment.

Radiological features suggestive of underlying amyloid angiopathy are associated with the highest risk of ICH recurrence. These include a prior lobar ICH (HR, 4.8),⁵⁷² the presence of microbleeds^27,573,574 (in particular strictly lobar microbleeds),^27,575 the number of lobar microbleeds (HR, 1.88 for 1 microbleed, 2.93 for 2–4 microbleeds, 4.12 for >4 microbleeds),⁵⁷² and the presence of disseminated cortical siderosis (HR, 4.69).^576,577 The presence of microbleeds and cortical siderosis can be determined during the etiological workup of ICH (Section 4.1, Diagnostic Assessment of Acute ICH Course). Carriers of apolipoprotein E genotypes associated with amyloid angiopathy are similarly at higher risk of ICH recurrence compared with those with the more common ε3/ε3 genotype; those with the ε2 or ε4 allele have an HR of 3.3 and 2.5 for recurrence, respectively.⁵⁷⁸ Recurrence risk also increases with higher measured outpatient BP⁵⁶³ and age^570,579 (HR, 2.8 in age >65 years) and is higher in those of Black race (HR, 1.22) or Asian race (HR, 1.29) compared with White race (race defined by self-designation, clinicians, or administrative personnel while in hospital).⁵⁶⁸ Association of ICH recurrence with Hispanic ethnicity has been inconsistent.^568,580

There is insufficient evidence to estimate ICH recurrence risk on an individual-patient basis. Deriving and validating a prediction rule incorporating clinical, radiological, and genotype biomarkers and determining the most informative thresholds for categorizing these factors would be helpful to estimate the risk of recurrence.

The mechanism by which race is associated with ICH recurrence, including the likely crucial role of social determinants of health, is unclear. More research into this association is required.

MRI findings suggestive of small vessel disease may reflect an increased risk for ICH recurrence. More research is needed into the recurrence risks associated with T2 hyperintensities, enlarged perivascular spaces, microangiopathic changes, intragyral hemorrhage, and lobar versus nonlobar microbleeds.

In a large prospective cohort study of 1145 patients with primary ICH and a median follow-up of 36.8 months, inadequate BP control was associated with increased risk of both lobar (HR, 3.53 [95% CI, 1.65–7.54]) and nonlobar (HR, 4.23 [95% CI, 1.02–17.52]) ICH recurrence.⁵⁶³ In PROGRESS (Perindopril ProtectionAgainst Recurrent Stroke Study), treatment with perindopril and indapamide reduced mean BP by 10.8/4.4 mm Hg in patients enrolled with ICH and resulted in a relative risk reduction of 42% (95% CI, 14–60) in major vascular events and a number needed to treat of 18 to prevent ICH recurrence over 5 years.⁵⁸¹ The optimal timing for BP lowering after ICH is not known, and a decision to initiate antihypertensive therapy in the acute setting should be in accordance with the recommendations discussed in Section 5.1, Acute BP Lowering.

In the PRoFESS trial (Prevention Regimen for Effectively Avoiding Second Strokes), the risk of ICH during follow-up was higher in subjects with SBP ≥160 mm Hg compared with those with SBP of 130 to 139 mm Hg (HR, 2.07 [95% CI, 1.22–3.51]), with a nonsignificant trend toward lower rates of ICH with SBP <130 mm Hg. Similarly, the risk of ICH was higher in subjects with DBP ≥100 mm Hg compared with DBP of 80 to 89 mm Hg (HR 2.58 [95% CI, 1.50–4.45]).⁵⁸² In a large prospective cohort study of 1145 patients with primary ICH, the risk of ICH recurrence was significantly higher for patients with SBP ≥120 mm Hg and DBP ≥80 mm Hg compared with patients who had SBP <120 mm Hg and DBP <80 mm Hg.^581,584 The relationship between SBP and ICH recurrence was continuous with an HR of 1.33 and 1.54 per 10-mm Hg increase for recurrent lobar and nonlobar ICH, respectively. Although a continuous relationship allows some flexibility with specific BP goals, the ICH evidence supports the ≤130/80-mm Hg target recommended in the 2017 hypertension clinical practice guidelines.⁵⁸⁵

The ideal target BP to prevent ICH recurrence is not known. More research is required to determine whether a more aggressive target of SBP of ≤120 mm Hg is beneficial.

The timing to initiate BP therapy and the optimal class of medication to achieve control are uncertain. Moreover, emerging research suggests that home BP measurements may be a more accurate measure of control. The timing of therapy, best choice of antihypertensive medication, and best approach to outpatient BP monitoring require further study.

It will be important to determine the predominant factors at the individual, systemic, and societal levels that preclude optimal BP control and identify strategies to overcome these barriers.

The balance of prothrombotic risks in patients with ICH and an LVAD or mechanical valves with the recurrent hemorrhagic risk of anticoagulation resumption remains challenging. There are sparse data on the risk and timing of device thrombosis versus worsening hemorrhage, and data remain observational. One study found that in patients with LVAD, anticoagulation resumption with warfarin at a median of 14 days from the index ICH was associated with fewer fatal and nonfatal thrombotic events than the resumption of antiplatelet alone, and there was no significant difference in recurrent ICH rates.⁵⁸⁶ In an observational study of 22 patients with LVAD with ICH, none had evidence of LVAD thrombosis after reversal and holding of anticoagulation for up to 13 days.¹⁸² In patients with mechanical heart valves, 1 study reported that although complications were significantly increased when anticoagulation was resumed before day 14, the composite of hemorrhage and thromboembolic risk suggested that anticoagulation may be considered in those with mechanical valves as early as day 6 from the index ICH.⁵⁸⁷ The decision to restart anticoagulation (eg, at 14 days after ICH for patients with LVAD and potentially earlier for patients with mechanical valves and relatively small ICHs) is therefore reasonable and safe in patients with LVAD or mechanical valves but requires individualized assessment of risk and benefit.

The decision to continue antiplatelet therapy in patients with a history of ischemic vascular events who have an incident ICH is challenging given concerns about the risk of ICH recurrence. One open-label RCT addressed this question.⁵⁸⁹ In 537 patients randomized at a median of 76 days after ICH onset and followed up for a median of 2 years, treatment with antiplatelet medications led to no increased risk of ICH and a reduction in the composite end point of nonfatal myocardial infarction, nonfatal stroke (including ICH and ischemic stroke), and death resulting from a vascular cause. On extended follow-up for up to 7 years, the study found no statistically significant effect of antiplatelet therapy on recurrent ICH or all other major vascular events.⁶⁰³ These results are consistent with a large meta-analysis of 1916 patients with ICH that reported no significant increase in risk of ICH recurrence and a decreased risk of thromboembolic and ischemic events with resumption of antiplatelet therapy.⁵⁸⁸ Important caveats include a scarcity of data on risk differences by location or cause of ICH, lack of blinding, and selection bias in patient enrollment based on clinician assessments of risk. Individual clinician assessment of patients’ risks of recurrent ICH and benefits of antiplatelet therapy is needed, but the available data support that, in appropriate patients, the resumption of antiplatelet therapy is reasonable. The optimal timing for resuming antiplatelet therapy has not been systematically studied.

A number of retrospective analyses have attempted to address the risks and benefits of anticoagulation therapy in patients with both nonvalvular AF and a history of ICH.^{179,590,591,593–595,604} The studies vary by design, including national registries and retrospective and prospective cohorts; have variable inclusion and exclusion criteria and timing to the initiation of anticoagulation; generally study VKA therapy; and include some replication of cohorts across studies. With these limitations, which include systematic differences between anticoagulated and nonanticoagulated individuals attributable to the confounding of choice of therapy by clinician-perceived risk-benefit profile, the published literature suggests a potential reduction in recurrent ischemic events and all-cause mortality with the use of anticoagulation. Anticoagulation may be considered in select patients‚ based on assessments of risk and benefit, and enrollment in ongoing prospective RCTs should be prioritized to address this clinical dilemma. Given the reduced risk of ICH with DOACs compared with VKAs in stroke prevention trials and real-world practice, these may be favored in patients with a history of ICH if anticoagulation is deemed indicated, although data are lacking.

The timing of resumption or initiation of anticoagulation in patients with AF and ICH remains challenging. A study suggests that a composite net benefit of stroke risk reduction and bleeding risk minimization occurs when anticoagulation is started 7 to 8 weeks after ICH.⁵⁹⁷ Before 4 to 8 weeks, there appears to be a significant increase in bleeding risk.^596,597 These studies suggest that the optimal timing of initiation of anticoagulation is ≈8 weeks after the index ICH. However, these studies are limited by confounding by indication and clinician and patient preferences. Therefore, timing should be considered on a case-by-case basis of individual risk assessments of thromboembolism, recurrent ICH, and late ICH expansion.

Left atrial appendage closure is an alternative in patients with AF and ICH who have contraindications to long-term oral anticoagulation. Two meta-analyses of the PROTECT-AF trial (WATCHMAN Left Atrial Appendage System for Embolic Protection in Patients With Atrial Fibrillation) and PREVAIL trial (Evaluation of the WATCHMAN Left Atrial Appendage [LAA] Closure Device in Patients With Atrial Fibrillation Versus Long Term Warfarin Therapy) reported outcomes in patients randomized to left atrial appendage closure or warfarin therapy.^598,599 Rates of ischemic stroke with left atrial appendage closure did not demonstrate noninferiority compared with warfarin, but rates of hemorrhagic stroke and bleeding were lower, and the primary end point of stroke, systemic embolism, and cardiovascular death was similar across the 2 treatment arms.^598,599 In patients with a history of ICH and AF, data from a small, nonrandomized, retrospective cohort showed lower cardiovascular mortality, all-cause mortality, hemorrhagic stroke risk, and major bleeding events with left atrial appendage closure compared with standard medical therapy.⁶⁰¹ Other small retrospective studies reported low event rates similar to rates in the patients without ICH⁶⁰⁰ and no ischemic stroke or ICH within 30 days of left atrial appendage closure among patients diagnosed with CAA.⁶⁰² Application of these results to individual patients with ICH remains unclear because of the potential confounding by patient selection, limited numbers of patients reported, and lack of standardization of time interval to left atrial appendage closure, type of antiplatelet or anticoagulant, and duration of treatment before and after implantation.

In addition to the uncertainty of risk and benefit of anticoagulation in patients with AF and ICH, there is limited evidence for individual selection of optimal timing of anticoagulation resumption in patients for whom anticoagulation will be restarted. Ongoing trials and future studies with stratification based on ICH location, mechanism, and risk factors for recurrence may lead to more informative decisions.

Most analyses evaluating the role of appropriate antithrombotic therapy in patients with ICH have focused on recurrent events. Future studies that incorporate outcomes such as disability or quality of life in addition to clinical events may provide information that is more patient-centric. More research is also needed on the timing of resumption of antiplatelet therapy and the differences in benefits and risks among different agents by different indications and across sex, racial, and ethnic groups.

Prospective data are lacking on the safety and efficacy of left atrial appendage closure in patients with ICH, particularly when performed within 6 months from the index ICH. This is important given that most patients under consideration for device implantation are under a time-sensitive risk-benefit analysis based on thromboembolic risk of untreated AF. Future studies may need to explore earlier timing and better standardized type and duration of antiplatelet therapy or anticoagulation therapy before and after implantation. As for all device-related therapies, future changes in left atrial appendage closure device type may affect patient outcome.

The association of statin use with both acute outcomes and the reduction of recurrent vascular events in patients who have had an ICH has been uncertain. The SPARCL study identified an increased risk of ICH with high-dose atorvastatin use in the setting of very-low-density lipoprotein levels.⁶¹² Post hoc analyses identified entry into the trial with an ICH as the stroke subtype conveying the highest risk for subsequent ICH but did not find an association between ICH and the most recent pre-ICH low-density lipoprotein value.⁶⁰⁶ Additional nonrandomized, observational studies have not found an association with statin use in patients with ICH.^{607,609,613,616,618} The risk may be mediated by complex interactions among genetic risk of recurrent ICH, lipid levels, and ICH location.^614,617 In addition, lipophilic statins may be associated with higher rates of ICH than hydrophilic statins.⁶¹⁵ Other retrospective analyses suggest the potential for improved outcomes after ICH with statin use^605,619,620 and a reduction in short- and long-term mortality with statin use.^{608,618,621–627} However, the results should be interpreted with caution because of selection bias and confounding by indication in these nonrandomized studies. Given this uncertainty, the decision to use statins in patients with ICH depends on risk assessment of ischemic cardiovascular and cerebrovascular events versus recurrent ICH. Enrollment in ongoing randomized clinical trials addressing this question can be encouraged. Clinical trials of the lipid-lowering PCSK9i (proprotein convertase subtilisin/kexin type 9 inhibitors) have thus far not suggested increased risk of first ICH but have not yet examined risk of recurrence in patients with prior ICH.^628–630

The use of NSAIDs is associated with an increased risk of bleeding. Overall event rates of ICH are low in the general population, but a large meta-analysis of observational studies found an increased risk of hemorrhagic stroke with diclofenac and meloxicam use.⁶¹¹ A subsequent large meta-analysis found an increased risk of ICH with any NSAID use.⁶¹⁰ One small study of patients with ICH with short-term outcomes found no association of NSAID use with outcomes and recurrent ICH, but follow-up was limited to 90 days.⁶³¹ Given the increased risk of bleeding with NSAID use, that patients with ICH are at higher risk of recurrent ICH than the general population, and the existence of safer alternatives to NSAID such as acetaminophen for most indications, the regular (eg, daily) use of NSAIDs after an ICH is not recommended, although randomized data and data drawn from individuals with ICH are lacking.

The effect of statins on long-term incident ischemic and hemorrhagic events in patients with ICH is uncertain, as is the effect of statin use on short-term outcomes after ICH. Ongoing and future studies to identify patient populations who may benefit from both short- and long-term statin use or from changing to an alternative lipid-lowering agent such as ezetimibe or a PCSK9i are needed.

Further research on radiological and biological markers that may further refine risk of recurrence such as MRI markers of cerebral small vessel disease, genetic risk, BP, and medication interactions would aid in the risk stratification in patients requiring the use of these medications.

There is growing but still uncertain evidence that some commonly used medications other than the antithrombotics may increase ICH risk.

There are positive effects from multimodal secondary stroke prevention. Secondary prevention includes increased physical activity, smoking and recreational drug cessation, reduction in alcohol consumption, and a healthy diet.^632,640 A healthy diet contains increased levels of fish rich in long-chain omega-3 fatty acids, vegetables and fruit, and whole-grain products, as well as lower levels of red meat, reduced levels of salt and added sugar, and replacement of saturated fats with polyunsaturated or monounsaturated fats.⁶⁴¹ A meta-analysis⁶³² showed positive effects in patients with transient ischemic attack and stroke with lower BP, and positive trends were noted in relation to blood lipids and anthropomorphic measures. Many studies were small and of varied quality, and none were studies of patients with ICH.

Heavy alcohol consumption can lead to intermittently elevated BP, which is particularly unhealthy in people with a prior ICH.^633,634 For those with large alcohol intake, a reduction by half had the strongest impact on BP.⁶³⁵ Heavy alcohol consumption⁶³³ or all alcohol consumption⁹³ is associated with ICH risk in observational studies, although confounding by other lifestyle factors is difficult to exclude.

Lifestyle modifications, in particular increased physical activity, might lead to reduced BP.^637,640 Although the mechanism of action is not fully understood, supervised training and counseling seem to have a significant impact on increasing physical activity. Increased physical activity such as reducing sitting time and taking daily walks has an impact, in particular going from sedentary to some activity level. These activities are feasible for many patients after stroke.⁶³⁶

In patients with stroke, psychosocial education for the caregiver can be beneficial to increase patients’ activity level and participation and/or quality of life.⁶⁴² Psychosocial interventions reduced depressive symptoms not only in the stroke survivors but also in their caregivers⁶³⁸ and may lead to reduced anxiety and improved quality of life and coping.⁶³⁵ This type of intervention was found to be acceptable to caregivers and can be delivered in a group setting and in one-to-one formats.⁶⁴³ As often in rehabilitation, the question of timing requires tailoring to individual needs. Data supporting the use of psychosocial education have come largely from studies of general stroke but not specifically from patients with ICH.

Practical support for the caregiver (such as how to walk safely with the patient) and training (such as how to perform certain exercises) are reasonable and can make performing some rehabilitation exercises at home feasible. This is not as effective as doing exercises with professionals but can lead to improvement in patients’ standing balance.⁶³⁹ Whether this is cost-effective compared with conventional rehabilitation is unclear; however, caregiver burden did not seem to increase. Caregiver-mediated exercise routines may be a promising form of therapy to add to usual care.⁶³⁹ Several factors limit the interpretation of these studies, however. The data are from studies of patients with general stroke rather than specifically patients with ICH, and the positive associations with standing balance are derived from secondary rather than primary outcome analyses.

There are positive effects of lifestyle modification; however, it is not known how to best target this and make the changes sustainable. The presumption is that tailored interventions are better than standardized interventions, but this needs to be investigated. There is also a lack of knowledge about which components of lifestyle modifications have the highest impact and the optimal frequency and content of outpatient follow-up visits.

Patient and caregiver education has been shown to be beneficial; however, it is still unclear how this should be delivered.

Another component of the education of the patient with ICH and caregiver that has not been studied is systematic use of advanced directives to determine preferences in case of recurrent ICH or other major events.