2022 Guideline for the Management of Patients With Spontaneous Intracerebral Hemorrhage: A Guideline From the American Heart Association/American Stroke Association – part. 2

5. Medical and Neurointensive Treatment for ICH

5.1. Acute BP Lowering

Synopsis

Most patients with acute ICH present with elevated BP. Elevated BP on presentation is associated with greater HE, ND, death, and dependency.^151–153 Therefore, it is intuitive to treat high BP during the acute phase of ICH. However, results from randomized clinical trials have been equivocal.^141,146 The current recommendations are based on data from the 2 largest trials (INTERACT2 [Intensive Blood Pressure Reduction in Acute Cerebral Hemorrhage Trial] and ATACH-2 [Antihypertensive Treatment of Acute Cerebral Hemorrhage II]) for early intensive BP lowering (EIBPL) after ICH,^141,146 meta-analyses,^{138,142,144,145,147} and several post hoc analyses of the INTERACT2 and ATACH-2 trials.^139,154,155 As a primary recommendation, lowering systolic BP (SBP) to a target range of 130 to 140 mm Hg is safe and may be reasonable in improving functional outcome in patients presenting with acute ICH of mild to moderate severity and SBP between 150 and 220 mm Hg. Initiating treatment as soon as possible and careful titration of BP-lowering therapy to ensure continuous smooth and sustained control of BP are recommended. Acute lowering of SBP to <130 mm Hg in patients presenting with ICH and elevated BP is potentially harmful and should be avoided.

Recommendation-Specific Supportive Text

1. Several studies have shown that high SBP variability during the hyperacute and acute phases of ICH is associated with poor outcomes.^138,154,156 A post hoc analysis of INTERACT2 found that increased SD of SBP during the first 24 hours had a linear association with death and severe disability at 90 days.¹⁵⁴ A meta-analysis of INTERACT2 and ATACH-2 also showed a continuous association between achieved SBP and lesser variability during the first 24 hours after ICH and the distribution of mRS scores at 90 days, suggesting that avoiding large fluctuations in BP is beneficial.¹³⁸ There is a lack of evidence to guide the choice of BP-lowering agents during the hyperacute phase after ICH, including bolus versus drip management. Intravenous nicardipine was the drug used in ATACH-2, whereas a range of intravenous and oral BP-lowering agents were used in INTERACT2. Any antihypertensive drug with rapid onset and short duration of action to facilitate easy titration and sustained BP control to minimize SBP variability seems appropriate, although venous vasodilators may be harmful because of unopposed venodilation and its effect on hemostasis and ICP.¹⁵⁷
2. The mean time from ICH onset to initiation of EIBPL treatment in ATACH-2 was 182±57 minutes compared with a median of 4 hours (interquartile range, 2.9–5.1 hours) in INTERACT2.^141,146 Evidence suggests that any potential benefit of BP lowering after ICH might be enhanced by earlier reductions in SBP. A subgroup analysis of ATACH-2 found that EIBPL within 2 hours of ICH onset was associated with lower risk of HE and improved 90-day outcomes compared with later time points.¹³⁹ In INTERACT2, reductions in SBP ≥20 mm Hg during the first hour after randomization and maintained for 7 days were associated with lowest risks of death and major disability.¹⁴⁰ Although the window for how long after ICH onset EIBPL remains beneficial has not been studied extensively, it would be expected to extend through the period when there is high risk for further HE.
3. EIBPL in patients with mild to moderate ICH (GCS score ≥5, excluding massive ICH) and SBP >150 to 220 mm Hg to 140 mm Hg appears to be safe. In INTERACT2, EIBPL to a target of 140 mm Hg with cessation of treatment at SBP <130 mm Hg was not associated with increased serious adverse events (SAEs) or mortality and resulted in modest improvement in secondary analyses of functional outcomes and quality of life domains but not in the primary outcome of death or major disability.¹⁴¹ The mean minimum SBP in the EIBPL group was 150 mm Hg. In ATACH-2, EIBPL to a target of 110 to 139 mm Hg was not associated with a lower rate of death or disability compared with standard reduction to 140 to 179 mm Hg.¹⁴⁶ Another study using perfusion CT in patients with small to medium ICHs also found no significant reduction in cerebral blood flow within the perihematomal region with EIBPL to <150 mm Hg.¹⁴³ Several meta-analyses indicated that EIBPL is safe overall and not associated with increased risk for SAEs, severe hypotension, ND, or mortality.^{138,142,144,145,147} A large systematic review and individual patient data meta-analysis including 16 randomized controlled trials (RCTs) with 6221 patients reported that EIBPL within 7 days of ICH onset reduced absolute and relative HE but did not improve functional outcomes.¹⁵⁸ Significant heterogeneity by BP-lowering strategy and agent was a limitation. Although patients with SBP >220 mm Hg were not intentionally included in the trials, it is common practice to take a similar BP-lowering approach in these patients.
4. A post hoc analysis of ATACH-2 trial of 682 participants with moderate to severe ICH (defined as GCS score <13, National Institutes of Health Stroke Scale [NIHSS] score ≥10, ICH volume ≥30 mL, or presence of IVH on presentation) found that EIBPL in this group reduced HE but did not reduce rate of death or disability at 90 days.¹⁴⁸ The safety of EIBPL and target BP for patients with large and more severe ICH and those requiring surgical decompression has less data. In patients with large ICH (>30 mL) requiring ICP monitoring or severe IVH requiring EVD, the burden of low cerebral perfusion pressure (CPP) <60 and <70 mm Hg was associated with increased mortality and poor functional outcomes, respectively, suggesting that BP reduction be accompanied by maintenance of CPP of 60 to ≥70 mm Hg in patients with large ICH, ICP elevation, or compromised CPP.^159,160
5. Compared with INTERACT2, ATACH-2, which did not exclude patients with initial SBP >220 mm Hg, did not show added benefit by lowering SBP to 110 to 139 mm Hg. Although the SBP target of 110 to 139 mm Hg did not worsen neurological outcomes or increase mortality, the additional SBP reduction was associated with increased renal and SAEs during the follow-up period.^141,146 The mean minimum SBP for the EIBPL group in INTERACT2 was 150 mm Hg compared with 129 mm Hg in ATACH-2, implying that EIBPL to <130 mm Hg may negate potential benefits. This is consistent with a secondary analysis of INTERACT2¹⁴⁹ that showed that an achieved mean SBP <130 mm Hg was associated with a modest increase in physical dysfunction and that SBP of 130 to 139 mm Hg was associated with best outcomes and not influenced by baseline BP. Subsequent analysis of ATACH-2 data indicated that elevated baseline serum creatinine, ICH volume ≥25 mL, and higher doses of nicardipine were associated with increased risk for acute renal injury.¹⁵⁵ A post hoc analysis of ATACH-2 in participants with initial SBP ≥220 mm Hg (22.8% of the cohort) reported higher rates of ND at 24 hours and renal adverse events until day 7 or discharge in patients treated with EIBPL compared with standard BP lowering, without any benefit in reducing HE at 24 hours or death or severe disability at 90 days, suggesting that cautious BP lowering may be required in these patients.¹⁵⁰ Similarly, a prospective single-center cohort study of 448 patients with ICH determined that a threshold maximum SBP reduction of 90 mm Hg was significantly associated with acute kidney injury regardless of preexisting chronic kidney disease.¹⁶¹ Acute kidney injury was associated with in-hospital mortality in patients with normal renal function but not in patients with chronic kidney disease.

Knowledge Gaps and Future Research

• The safety and efficacy of EIBPL in patients with SBP >220 mm Hg and those with large and more severe ICHs, who may be more susceptible to cerebral perfusion compromise attributable to high ICP, require more study because these patients were not adequately represented in previous trials. There are insufficient data on which to base a target BP range for large ICH particularly in the absence of ICP monitoring. The rate of BP reduction in patients with excessively high SBP also requires additional study. In INTERACT2, ≈75% of patients presented with mild to moderate ICH <20 mL, median ICH volume was 11 mL, and NIHSS score was 10 in the EIBPL group. In ATACH-2, 91% of patients in the EIBPL group had an ICH volume <30 mL on presentation, median ICH volume was 10 mL, and NIHSS score was 11.
• Only 16% of patients in INTERACT-2 and 9.7% of patients in ATACH-2 had lobar ICH. More research is needed in the lobar subset of patients with ICH to address the different pathophysiology and natural history of lobar compared with deep ICH.
• It remains unclear whether ultraearly BP lowering could be beneficial. In RIGHT-2 (Rapid Intervention With Glyceryl Trinitrate in Hypertensive Stroke Trial), which included 145 of 1149 patients (13%) with ICH and SBP ≥120 mm Hg, transdermal nitroglycerine in the ambulance (median time from ICH onset to randomization, 74 minutes) was associated with worse outcomes and larger hematoma and edema volumes. Interpretation of these findings is limited by the small sample size and the potential confounding venous vasodilator effects of nitroglycerine and inhibition of the early vasoconstriction and platelet plugging phases of hemostasis on HE. The benefits of BP lowering beyond the first 6 hours after symptom onset also remain unclear because INTERACT2 and ATACH-2 required initiation of BP-lowering treatment within 6 and 4.5 hours of symptom onset, respectively.
• More research is also needed to better delineate the importance of various BP measures, including the selection and method (bolus versus drip) of administration of BP-lowering agent, absolute versus relative reduction, and prognostic significance of the magnitude of BP reduction during the first few hours. Secondary analyses of INTERACT2 suggest that large SBP reductions >20 mm Hg within the first hour are associated with lower risks of poor outcomes at 90 days. In contrast, both the individual patient data analysis of INTERACT2 and ATACH-2 and a recent retrospective study in 757 patients found that early SBP reduction of >60 mm Hg in the first hour and between 40 and 60 mm Hg, respectively (compared with <20 mm Hg), were associated with an increased proportion of patients with unfavorable outcome, suggesting caution in lowering BP too quickly in patients with very high BP on arrival. These seemingly disparate findings merit further investigations. The most suitable target for BP reduction in untreated versus treated with controlled hypertension also warrants further exploration.
• Methods of BP measurement for early BP lowering after ICH have not been studied, including noninvasive versus invasive devices and frequency of measurements, which may be defined by studies evaluating targets for minimizing BP variability.

5.2. Hemostasis and Coagulopathy

5.2.1. Anticoagulant-Related Hemorrhage

Synopsis

The risk of HE, rapid deterioration, and poor outcome is increased in patients with ICH on anticoagulation therapy. Management requires emergency reversal of anticoagulation (Figure 2), and protocols and processes of care should be in place.

In general, treatment should be administered when clinically significant anticoagulant levels are suspected on the basis of type and timing of anticoagulant dosing rather than waiting for results of blood tests. Four-factor PCC is superior to plasma for warfarin-associated ICH to rapidly replace vitamin K–dependent coagulation factors¹⁶³ and should be given with intravenous vitamin K to re-establish vitamin K–dependent coagulation factor production. (Note that this guideline uses the term 4-F PCC when the supporting literature specifies this agent and otherwise uses the more general term PCC when the literature does not specify which PCC was used.) Reversal of the anticoagulant effect of direct thrombin inhibitors and factor Xa inhibitors can be performed rapidly with specific reversal agents (idarucizumab¹⁶⁸ and andexanet alfa,¹⁶⁶ respectively). However, there are few clinical data on the effectiveness of these agents in preventing HE or improving functional outcomes, and in real-world situations, clinicians will have to balance the expense against the benefit of these drugs. When specific reversal agents are not available, aPCC or 4-F PCC may promote hemostasis in patients on direct thrombin inhibitors¹⁷⁶ and factor Xa inhibitors.^169–171 RRT may reduce dabigatran concentration.¹⁷⁷ In patients on heparin, protamine reverses the anticoagulant effect.¹⁷⁸

Recommendation-Specific Supportive Text

1. 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
2. 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
3. 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
4. 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.
5. 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
6. 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.
7. 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.
8. 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.
9. 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
10. 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
11. 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
12. 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

Knowledge Gaps and Future Research

• 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.

5.2.2. Antiplatelet-Related Hemorrhage

Synopsis

The effect of antiplatelet agents on the outcome of ICH is uncertain. A systematic review of 25 observational studies found that antiplatelet therapy at the time of the hemorrhage was associated with a 27% increase in mortality but not with functional outcome.²¹¹ In a more recent retrospective cohort study with 3545 patients, antiplatelet use on its own was not independently associated with worse functional outcome, whereas when an antiplatelet was combined with a VKA, there was a reduced chance of favorable outcome, as defined by an mRS score of 0 to 3.²¹² In an RCT, the subset of patients with antiplatelet therapy had more unfavorable functional outcome and higher mortality.²¹³ The studies generally do not provide separate results for different antiplatelet agents, which vary in terms of degree of platelet inhibition, half-life, and reversibility. Platelet transfusions, desmopressin, and TXA have proven effective in reducing bleeding in other clinical indications,^214–216 whereas for spontaneous ICH in patients being treated with antiplatelet agents, no convincing benefit has been demonstrated.^{207–210,213} The exception is emergency craniotomy for hematoma removal, for which reversal of the antiplatelet effect of aspirin with platelet transfusions might reduce postoperative hemorrhage volume.²⁰⁶

Recommendation-Specific Supportive Text

1. 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.
2. 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
3. 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.

Knowledge Gaps and Future Research

• 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.

5.2.3. General Hemostatic Treatments

Synopsis

HE occurs in up to a third of patients after ICH and is associated with poor outcome.¹⁰³ Hemostatic therapy for the prevention of HE remains an attractive therapeutic target after ICH. To date, large RCTs have assessed 2 agents, recombinant factor VIIa and TXA. The modest effects of these agents on limiting HE have not translated into improvement in functional outcome. Presence of the CTA spot sign or other CT indications of possible HE did not predict beneficial response to either hemostatic therapy. ICH expansion most commonly occurs very early after onset, and future studies need to target earlier treatment.

Recommendation-Specific Supportive Text

1. 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
2. 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

Knowledge Gaps and Future Research

• 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.

5.3. General Inpatient Care

5.3.1. Inpatient Care Setting

Synopsis

Patients with ICH who have clinical hydrocephalus, IVH, or infratentorial hemorrhage are best cared for in facilities with neurosurgical and neuro-specialized critical care capabilities. It can be challenging to predict a priori which of these subsets of patients with ICH will require neurosurgical evaluation or management. Therefore, treatment facilities without in-house access to neurosurgical and neurocritical care capabilities should ensure the ability to obtain a consultation for such care or consider transfer to facilities that have these resources. The present guideline recommends that appropriate life-sustaining therapies be initiated before transfer for patients with ICH with an unstable hemodynamic profile, inadequate airway protection, or inadequate gas exchange. Recommendations are intended for patients with no limitation of life-sustaining therapy, the initiation of which should be consistent with the patient’s advance directive information and goals of care. A do not attempt resuscitation (DNAR) order does not itself indicate that the patient should not receive emergency treatment (Section 7.2, Decisions to Limit Life-Sustaining Treatment). Determination of the optimal timing for patients with ICH to transition from the ED to another care environment such as an ICU is complex and may be related to the ability for that ED to manage critically ill patients.^242,243 Depending on the severity of hemorrhage, the appropriate inpatient setting may be an ICU (defined by provision of the full spectrum of critical care and intensive monitoring) or a dedicated stroke unit (licensed by regional or national stroke organizations according to standard of care and round-the-clock stroke expertise).

Recommendation-Specific Supportive Text

1. 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
2. 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
3. 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
4. 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
5. 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.
6. 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.
7. 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
8. 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
9. 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

Knowledge Gaps and Future Research

• 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.

5.3.2. Prevention and Management of Acute Medical Complications

Synopsis

In the first hours and days after ICH, the focus and management goals of physicians and nurses are directed not only at treating the ICH and preventing HE but also at early identification and prevention of acute medical complications. Problems related to impaired swallowing, immobility, hemodynamic response and stability, infection, intensive care delirium, and altered consciousness are among the issues that neuroscience physicians and nurses must address throughout the patient’s hospital course. Medical complications can range in severity but are associated with increased LOS, increased rates of mortality, and worse functional outcomes at 90 days.

Recommendation-Specific Supportive Text

1. 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
2. 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.
3. 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.
4. 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.

Knowledge Gaps and Future Research

• 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

5.3.3. Thromboprophylaxis and Treatment of Thrombosis

Synopsis

Mechanical DVT prophylaxis is rarely contraindicated, and the writing group recommends using IPC devices from the day of diagnosis of ICH on the basis of a large RCT²⁸³ and network meta-analysis of 4 RCTs.²⁷⁶ A meta-analysis of 4 studies demonstrated that heparin or LMWH reduces the risk of PE²⁷⁸ when initiated 48 to 96 hours after onset of the hemorrhage or the diagnosis without a significant increase in hematoma enlargement.^{277,279,281,282} Graduated compression stockings of any length are not effective against symptomatic DVT according to 2 large RCTs and 2 meta-analyses.^{276,278,283,284} The balance between avoidance of recurrent ICH and appropriate treatment of the VTE to evade potentially fatal PE is challenging, especially during the first few days after onset of ICH. In a large registry of patients with VTE, insertion of an inferior vena cava (IVC) filter for those at a high risk of bleeding reduced the risk for PE-related death and for recurrent VTE compared with no IVC filter.²⁸⁵ In 2 retrospective studies on patients with trauma-associated ICH and symptomatic VTE, initiation of therapeutic anticoagulation 1 to 2 weeks after the onset of ICH appeared safe with regard to HE.^286,287

Recommendation-Specific Supportive Text

1. 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
2. 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.
3. 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
4. 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.
5. 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.
6. 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.

Knowledge Gaps and Future Research

• 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.

5.3.4. Nursing Care

Synopsis

Nursing care for the patient with ICH is complex and multifaceted, often requiring critical management of hemodynamics such as BP, fever control, airway management, diagnostic laboratory and radiographic testing, assessment and management of ICP, frequent neurological assessments, and prevention of secondary complications. Nurses must have the education and knowledge to recognize stroke symptoms and the training to activate protocols for prompt assessment and management by the stroke team. Understanding the importance of the “why” (to identify ND early) and “how” (GCS or NIHSS) of neurological assessments will aid nurses in providing focused, quality assessments in a timely manner. Five studies highlight the significant negative impact that early and delayed ND have on patient mortality and functional outcome. As many as 22.6% of patients with ICH had ND in the ED,⁶¹ whereas as many as 70% had ND in the first 24 hours of admission.²⁹² Specialized nurse competency training programs are associated with increased nursing satisfaction and have been shown to improve compliance with stroke evidence-based protocols.²⁹⁶

Recommendation-Specific Supportive Text

1. 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.
2. 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.
3. 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.

Knowledge Gaps and Future Research

• 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.

5.3.5. Glucose Management

Synopsis

Glucose monitoring and management are often considered part of the general care of all patients, including those with ICH. One randomized controlled study of mixed stroke subtypes showed that a bundled care approach, including glycemic control, temperature management, and dysphagia screening, improved outcomes.²⁵⁶ Hyperglycemia on presentation may herald a worse prognosis.^308,309 However, tight glucose control may increase the risk of hypoglycemic events and worsen outcomes.^299–301 The ideal evidence-based approach to glucose management in patients with ICH has remained elusive.³¹⁰

Recommendation-Specific Supportive Text

1. 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
2. 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.
3. 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.

Knowledge Gaps and Future Research

• 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.

5.3.6. Temperature Management

Synopsis

Temperature abnormalities in the setting of acute ICH are common and can occur in >30% of patients with ICH at some point during their hospitalization.^318–321 Fever appears to be associated with both higher clinical severity and worse outcomes³²²; however, evidence for whether treating fever improves outcomes is conflicting.^311,313 The challenge in interpreting this body of literature includes variable but often small sample sizes, few RCTs, different definitions of fever, and different therapeutic approaches addressing fever. Although many empirically treat fever, some data suggest a judicious approach. One study noted that 90% of patients with ICH met systemic inflammatory response syndrome criteria within the first 24 hours of admission. As part of their evaluation, blood cultures were obtained that provided a diagnostic yield of 0.1%, leading to increased costs of care.³²³

Recommendation-Specific Supportive Text

1. 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.
2. 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}

Knowledge Gaps and Future Research

• 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.

5.4. Seizures and Antiseizure Drugs

Synopsis

In this guideline, the writing group uses the International League Against Epilepsy definition of seizure, “a transient occurrence of signs and/or symptoms due to abnormal excessive or synchronous neuronal activity in the brain,”³³² and, in the context of an electrographic seizure, the definition outlined by the American Clinical Neurophysiology Society, “epileptiform discharges averaging >2.5 Hz for ≥10 s (>25 discharges in 10 s) or any pattern with definite evolution and lasting ≥10 s.”^332a New-onset seizures in the context of spontaneous ICH are relatively common (between 2.8% and 28%), and most of these seizures occur within the first 24 hours of the hemorrhage.^{327,333–335} Prophylactic use of antiseizure drugs, however, is of unclear benefit. The optimal approach to monitoring patients with ICH for seizures is unclear. However, data suggest that continuous electroencephalographic monitoring for at least 24 hours is probably reasonable; patients in a coma may require more prolonged monitoring.³²⁷ The relationship among seizures, functional outcomes, and mortality is complex and not well defined. One of the primary challenges in this area is that the studies differ on the definition of seizure and method of detection.^325,329,330 Another consideration is that seizures may be a marker of ICH rather than specifically affecting outcomes.³²⁷

Recommendation-Specific Supportive Text

1. There is uncertainty about the prognostic significance of abnormal electrographic patterns in the setting of ICH.^325,326 The clinical context should therefore be considered in the decision-making process. The recommendation is to initiate antiepileptic medication in the context of an electrographic seizure that is clinically suspected to be contributing to the impaired consciousness in order to improve morbidity (defined as LOS >14 days or discharge to somewhere other than home or a rehabilitation facility).³²⁵ Identifying electrographic seizures can be challenging, however, and may require consultation.
2. There are no large, prospective RCTs to demonstrate the efficacy of treating seizures in the context of ICH. One small randomized trial evaluated the use of prophylactic valproic acid and suggested no difference in mortality or long-term seizure control.³³⁶ Other studies similarly failed to demonstrate a clear mortality benefit from treating seizures in the context of ICH.^337–339 Still others showed better outcomes in patients with post-ICH seizures.^333,334 However, given the inherent limitations in the design of the available studies and the low risk of antiseizure medications in the context of active seizures, the benefits for both abortive and preventive treatment of seizures appear likely to outweigh the risks. Risk scores such as the CAVE score³⁴⁰ can be used to estimate the risk of late seizures (>7 days after ICH). However, in the absence of evidence that antiepileptic medications prevent late seizures after ICH, risk scores should not be used to guide continuation of antiepileptic drugs.
3. The primary focus is on those with possible seizures that are likely contributing to the clinical picture such as patients with ICH with impaired or fluctuating level of consciousness out of proportion to the degree of brain injury or other metabolic abnormalities. These patients may not demonstrate clear and convincing rhythmic movements consistent with typical clinical seizures. If seizures are clinically suspected in this context, it is reasonable to evaluate them with a continuous electroencephalogram for at least 24 hours. One study noted that 28% of those with electrographic seizures were detected after at least 24 hours of continuous monitoring, whereas 94% were detected with at least 48 hours of monitoring. Among patients in a coma, 36% required continuous electroencephalography monitoring for >24 hours to detect the first seizure.³²⁷
4. Earlier studies suggested that prophylactic antiseizure drugs such as phenytoin were associated with worse outcomes in patients with ICH.^335,341 Consequently, the use of alternative prophylactic antiseizure drugs such as levetiracetam may have become more common.³⁴² Recent studies have not consistently identified harm or benefit from the use of prophylactic antiseizure drugs after spontaneous ICH with respect to global functional outcomes,^{328–331,343} but specific domains of abilities such as cognitive function might be negatively affected.³⁴⁴ One meta-analysis (1 RCT, 7 observational studies) found that seizure prophylaxis in patients with ICH was not associated with preventing either early (<14 days from ICH) or long-term seizures.³⁴⁵ Another meta-analysis reported that neither levetiracetam nor phenytoin prophylaxis was associated with worse functional outcomes at the longest follow-up or 90 days, although there was a trend toward better outcomes in populations with higher proportions of patients taking levetiracetam.³⁴⁶

Knowledge Gaps and Future Research

• The relationship between seizures and outcomes and the impact of antiseizure medications, especially when given in a targeted and time-limited manner, on outcome in patients with ICH are not well defined.
• The optimal approach to the patient with ICH with impaired consciousness and an abnormal electroencephalogram is not well defined.
• There is no clear consensus on which abnormal electrographic patterns in patients with ICH and impaired consciousness, with or without seizure, have prognostic significance.

5.5. Neuroinvasive Monitoring, ICP, and Edema Treatment

Synopsis

Limited data exist with respect to the frequency of elevated ICP and its management in the setting of ICH. ICP is typically measured by insertion of ICP monitors into the brain parenchyma or an EVD into the ventricles. The current recommendations on when to use EVD, ICP monitoring, hyperosmolar therapy, and corticosteroids in patients with ICH are based primarily on data from small RCTs, retrospective series, systematic reviews, and meta-analyses. As a primary recommendation, ventricular drainage should be performed in patients with ICH/IVH with hydrocephalus contributing to decreased level of consciousness. The indications for use of ICP monitoring are less clear. In patients with ICH with a GCS score ≤8, ICP monitoring and treatment might be considered to reduce mortality and improve outcomes. Hyperosmolar therapy may be considered for transiently reducing ICP. However, early prophylactic hyperosmolar agents have not demonstrated efficacy in improving outcomes, and their efficacy remains uncertain. Corticosteroids should not be administered for the treatment of elevated ICP in the setting of ICH.

Recommendation-Specific Supportive Text

1. Hydrocephalus (Section 5.3.1, Inpatient Care Setting, Recommendations 3, 6, and 7) is an independent predictor of mortality after ICH.³⁷⁰ EVD is a lifesaving procedure that can rapidly decrease ICP secondary to hydrocephalus.³⁵⁰ A retrospective review of a large series of patients with ICH with IVH demonstrates that EVD placement is an independent predictor of reduced mortality at hospital discharge in patients (GCS score >3) with hydrocephalus at presentation.³⁴⁷ A multi-institutional retrospective analysis suggests that EVD use is associated with lower 30-day mortality rates in patients with greater ICH volumes, higher ICH scores, and lower admission GCS scores.³⁴⁸ A systematic review demonstrates that treatment with ventricular drainage, combined with fibrinolytics, may improve outcome in patients with ICH with intraventricular extension.³⁴⁹ Other studies present conflicting results. In a secondary analysis of the FAST trial (Recombinant Factor VIIa in Acute Intracerebral Haemorrhage), a small number of patients who received EVDs exhibited no overall clinical benefit.³⁷¹ In a retrospective review of primary ICH affecting the thalamus, EVD placement showed no significant correlation with clinical outcomes.³⁷² Small sample sizes and retrospective, post hoc analysis methods introduce significant risk of bias to these studies. Although postventriculostomy hemorrhage is reasonably common in the setting of ICH, it appears to be of minor clinical significance in the majority of patients.³⁷³
2. The frequency at which ICP elevations occur after ICH is unclear. A retrospective analysis of a large institutional cohort demonstrates that intracranial hypertension is common after ICH, especially in younger patients with supratentorial hemorrhage.³⁷⁴ However, an analysis of 2 RCTs suggests that ICP is infrequently elevated during EVD monitoring and drainage in patients with severe IVH.^159,356 A 2019 systematic review and meta-analysis indicates that the prevalence and mortality of intracranial hypertension are high after ICH.³⁵² No randomized studies have addressed the utility of ICP monitoring in patients with ICH. However, multiple retrospective analyses, case series, and secondary analyses examine this topic. Studies including secondary analysis of 1 RCT suggest that increased ICP levels, durations, and variability are associated with poor outcome and mortality.^{159,354–356,370} The impact of ICP monitoring on patient outcome is unclear. A retrospective database analysis suggests that ICP monitoring is beneficial in patients with ICH with moderate to severe ICH/IVH with reduced levels of consciousness, especially those with GCS scores of 9 to 12.³⁵³ Secondary analyses of the ERICH (Ethnic/Racial Variations of Intracerebral Hemorrhage) and MISTIE III (Minimally Invasive Surgery Plus rt-PA for Intracerebral Hemorrhage Evacuation) data do not support the routine use of ICP monitoring in patients with ICH,^351,375 although long-term mortality in MISTIE III was significantly associated with higher proportion of time with high ICP and low CPP in monitored patients.¹⁶⁰ Shortcomings inherent to retrospective studies or secondary analyses such as small sample size and selection biases should be considered in the interpretation of these findings.
3. Prophylactic administration of hyperosmolar agents (including mannitol and hypertonic saline) to attain serum hyperosmolar levels has been studied in patients with ICH. Small retrospective studies suggest a potential benefit of hyperosmolar infusion on cerebral blood flow, edema evolution, and frequency of ICP crises.^376,377 The studies focused on prophylactic use of mannitol infusion have not demonstrated clinical benefit,^357,360,361 whereas prophylactic hypertonic saline infusions have not been well studied. The propensity-matched retrospective analysis from the ERICH study cohort in which 78% of treated cases received only mannitol suggested that hyperosmolar therapy is not associated with better 3-month mRS outcomes.³⁵⁹ A 2007 Cochrane review concluded that there was not enough evidence to determine whether the routine use of mannitol would result in any beneficial or harmful effect.³⁵⁷ A systematic review and meta-analysis conducted in 2018 determined that mannitol could lead to hematoma enlargement and did not recommend routine use in the early stage of supratentorial ICH.³⁶⁰ Although the efficacy of hyperosmolar treatments to attain serum hyperosmolarity is not well established, usual supportive medical care includes treatment of hyponatremia and other post-ICH medical complications.
4. Hyperosmolar therapy is the principal medical strategy in the treatment of cerebral edema.³⁷⁸ A 2011 meta-analysis of randomized clinical trials suggested that mannitol or hypertonic saline, in equiosmolar doses, may be effective in treating acutely elevated ICPs but that hypertonic saline is more effective than mannitol. This meta-analysis included studies of patients undergoing quantitative ICP monitoring regardless of underlying cause.³⁶² A retrospective analysis examined the dose of mannitol needed to reach a stable ICP level in the setting of ICH. The study found that the effect of mannitol on ICP was dose dependent during the period of ICP reduction but not after the ICP had reached a stable level.³⁶³ The optimal mannitol dose required for individual patients with ICH with elevated ICP can be calculated by determining hemorrhage location, hematoma volume, and pretreated ICP measurement.³⁶³ A study of 20 patients with ICH examined mean flow velocities and pulsatility indices in the middle cerebral artery territory. Results suggested that a single bolus of mannitol modifies cerebral hemodynamics (increased flow velocities in affected middle cerebral artery) in patients with ICH.³⁶⁴
5. A 1987 randomized controlled study found that dexamethasone treatment resulted in no beneficial effects and increased complications (principally infections and diabetic complications) in patients with supratentorial ICH.³⁶⁸ A second RCT performed in 1989 demonstrates no differences in outcomes in patients with ICH treated with corticosteroids versus those treated without corticosteroids.³⁶⁵ A 1998 RCT suggests that dexamethasone does not likely cause an unacceptably high rate of complications but also does not provide a benefit.³⁶⁶ More recently, a Cochrane review³⁶⁷ and a meta-analysis³⁶⁹ demonstrated no clear benefit to patients with ICH treated with dexamethasone or glucocorticoids. Taken together, these studies suggest that there may be some risk, in addition to a lack of benefit, for corticosteroid administration in the setting of ICH.

Knowledge Gaps and Future Research

• Because of a paucity of disease-specific data, indications for ICP monitoring in patients with ICH are often derived from the TBI literature. Guidelines suggest ICP monitoring in patients with a GCS score of 3 to 8 and maintenance of an ICP <22 mm Hg and a CPP of 50 to 70 mm Hg, depending on capacity for cerebral autoregulation. Studies focused exclusively on ICH may help to determine specific parameters that can be used to guide the monitoring and treatment of patients with ICH.
• A meta-analysis demonstrated a potential advantage of hypertonic saline over mannitol in lowering ICP across a range of neuropathologies. However, the comparative efficacies of mannitol and hypertonic saline have not been extensively studied in the setting of ICH. Future investigations could determine whether there is a greater benefit of one versus the other for patients with ICH with elevated ICP.
• Neuroinvasive monitoring is advancing rapidly. Multimodality monitoring techniques suggest that fraction of inspired oxygen, mean arterial pressure, and CPP can be used to predict changes in brain tissue oxygen. Elevated glutamate levels are noted in the perihematomal region. Small case series indicate CPP parameters and threshold pyruvate/lactate ratios that are associated with favorable outcomes after ICH. Larger future studies focused on indications and utility of microdialysis and brain tissue oxygenation measurements in the perihematomal region may help to determine optimal tissue oxygenation parameters and metabolic correlates associated with favorable outcomes after ICH.
• Hyperosmolar therapy is typically administered in 4- to 6-hour intervals. However, the duration of transient effects from hyperosmolar therapy in the setting of ICH is unclear. Further studies could determine the effective treatment durations and whether hyperosmolar agents are effective in preventing poor outcomes.

For more information, see the source’s website >>