Suboptimal Cerebral Perfusion is Associated with Ischemia after Intracerebral Hemorrhage

Abstract

Introduction:

Remote ischemic lesions on diffusion-weighted imaging (DWI) occur in one-third of patients with intracerebral hemorrhage (ICH) and are associated with worse outcomes. The etiology is unclear and not solely due to blood pressure reduction. We hypothesized that impaired cerebrovascular autoregulation and hypoperfusion below individualized lower limits of autoregulation are associated with the presence of DWI lesions.

Methods:

This was a retrospective, single-center study of all primary ICH with intraparenchymal pressure monitoring within 10 days from onset and subsequent MRI. Pressure reactivity index (PRx) was calculated as the correlation coefficient between mean arterial pressure (MAP) and intracranial pressure (ICP). Optimal cerebral perfusion pressure (CPPopt) is the cerebral perfusion pressure (CPP) with the lowest corresponding pressure reactivity index (PRx). DeltaCPP (difference between CPP and CPPopt), time spent below the lower limit of autoregulation (LLA), and time spent above the upper limit of autoregulation (ULA) were calculated using mean hourly physiologic data. Univariate associations between physiologic parameters and DWI lesions were analyzed using binary logistic regression.

Results:

505 hours of artifact-free data from seven patients without DWI lesions and 479 hours from six patients with DWI lesions were analyzed. Patients with DWI lesions had higher ICP (17.50 vs 10.92 mmHg; OR 1.14, CI 1.01-1.29) but no difference in MAP or CPP compared to patients without DWI lesions. The presence of DWI lesions was significantly associated with a greater percentage of time spent below the LLA (49.85% vs 14.70%, OR 5.77, CI 1.88-17.75). No significant association was demonstrated between CPPopt, deltaCPP, ULA, LLA, or time spent above the ULA between groups.

Conclusions:

Blood pressure reduction below the lower limits of autoregulation is associated with ischemia after acute ICH. Individualized, autoregulation-informed target for blood pressure reduction may provide a novel paradigm in acute management of intracerebral hemorrhage and requires further study.

Introduction

Remote cerebral infarction detected on diffusion-weighted imaging (DWI) may be seen in approximately one-third of acute primary intracerebral hemorrhage (ICH) (1). The typical appearance of these lesions is punctate and distant from the site of the hemorrhage(2). Several studies have demonstrated the association of secondary ischemia with features of chronic microangiopathy such as cerebral microbleeds and leukoaraiosis in addition to larger hematoma volume, higher initial blood pressure, younger age, male sex, and black race(1,3). The presence of DWI lesions is an independent predictor of poor functional outcome(1,4,5). The underlying pathophysiology remains uncertain; several theorized mechanisms have been proposed including over-aggressive blood pressure reduction, withdrawal of antithrombotic agents, neuroinflammation, metabolic failure, or acute microvascular dysfunction(6).

Cerebrovascular autoregulation enables the cerebrovasculature to maintain adequate cerebral blood flow to the brain over a range of cerebral perfusion pressures. When outside of this range, cerebral ischemia or edema may occur secondary to hypoperfusion or hyperemia respectively. The optimal cerebral perfusion pressure of an individual may vary in the context of vascular autoregulatory function, microangiopathy, and pre-morbid blood pressure range. Impairment of cerebral autoregulation is a well-documented occurrence in a number of acute brain injuries such as subarachnoid hemorrhage, traumatic brain injury, and ICH (7–9).

We hypothesized that the development of DWI lesions was associated with impaired cerebral autoregulation and cerebral perfusion pressures (CPP) maintained below calculated optimal cerebral perfusion pressure (CPPopt). We investigated the relationship between the presence of DWI lesions and the degree of impaired cerebrovascular autoregulation measured by the pressure reactivity index (PRx), time spent below the lower limit of autoregulation, and the difference between CPP and CPPopt (deltaCPP) in patients with primary ICH.

Materials and Methods

Study Population

All patients with primary ICH admitted to the Columbia University Irving Medical Center Neurological Intensive Care Unit from June 2006 to June 2020 who underwent invasive intracranial multimodality monitoring were screened from the Intracerebral Hemorrhage Outcomes Project (ICHOP), a single-center prospective ICH observational cohort study (10). The diagnosis of primary ICH was based upon non-contrast CT head and the absence of secondary etiologies such as vascular malformations or tumor. Institutional criteria for invasive intracranial monitoring required a Glasgow Coma Scale (GCS) of 8 or lower within the first week of ICU admission without expected improvement in exam, clinical deterioration, or neurological deficits out of proportion to radiographic imaging. Only patients with subsequent brain magnetic resonance imaging during hospitalization were included to determine the presence of DWI lesions remote from site of the primary ICH. The study was approved by the institutional review board of Columbia University Medical Center and was performed in accordance with the ethical standards as outlined in the 1964 Declaration of Helsinki and its amendments. Informed consent was obtained from the patient representative.

Management Plan

Management of ICH was in accordance with contemporaneous guidelines from the American Heart Association for medical and surgical treatment(11). Patients were assessed for impaired coagulation parameters and recent administration of anticoagulant or antithrombotic therapy. Appropriate emergent reversal was performed in accordance with an institutional standardized protocol. CT angiogram and/or digital subtracted angiography was performed when a vascular lesion was suspected. External ventricular drainage was performed in cases of symptomatic intraventricular hemorrhage or intracranial hypertension. Surgical hematoma evacuation and/or decompressive hemicraniectomy was performed at the discretion of the neurocritical care and neurosurgical teams in cases of refractory intracranial hypertension or neurological deterioration secondary to hematoma mass effect. Blood pressure targets were based upon contemporary guideline recommendations with a systolic target between 140-180 mmHg and CPP greater than 60 mmHg (11–13). All patients had invasive arterial catheter placement for continuous blood pressure monitoring calibrated at the phlebostatic axis. The head of bed was elevated between 30-45 degrees for all patients except during circumstances necessitating the head to be temporarily flat (i.e. imaging, transport, nursing care, etc.). Interventions to reduce intracranial pressure (ICP) were instated if sustained at more than 20mmHg for greater than 5 minutes. Partial pressure of carbon dioxide in the blood (PaCO2) was targeted to 30 to 40 mmHg, and partial arterial oxygen pressure was maintained above 80 mmHg.

Intracranial Monitoring

Intracranial parenchymal pressure monitoring utilized Integra Neuroscience catheters or Raumedic NEUROVENT catheters. Parenchymal monitors allow for continuous recording of intracranial pressure waveforms, allowing for accurate correlation plots between MAP and CPP. Patients with solely external ventricular drains were not included due to the institutional application of continuous drainage, preventing continuous ICP waveform data. Invasive intracranial parenchymal monitoring was performed in select cases with a Glasgow Coma Scale (GCS) of 8 or lower in the setting of an unreliable or confounded neurological examination or in cases of neurologic deficits out of proportion to that expected by the treating physicians. Exclusion criteria for intracranial pressure monitoring included immediate surgical hematoma evacuation, absent brainstem reflexes, ictus of hemorrhage greater than one week prior, and active coagulopathy. Placement selection was preferentially in the peri-lesional region or non-dominant frontal lobe depending on proximity to eloquent tissue. The placement of invasive intracranial monitoring was agreed upon by the treating neurosurgical and neurointensivist team with consent from the patient’s surrogate representative. Standard institutional safety precautions performed prior to placement included normalization of coagulation parameters, hemodynamic stability, and procedural sterility. Intracranial monitor removal was based upon neurologic stability, probe function, resolution of intracranial hypertension, necessary additional neurosurgical intervention, and at the discretion of the clinical team.

Data Collection

Physiologic digital data was obtained from General Electric Solar 8000i monitors (Milwaukee, WI) and acquired at a sampling frequency of 240Hz using BedmasterEX (Excel Medical Electronics, Jupiter, FL) from years 2006-2012. Beyond year 2012, digital data was obtained from Philips Intellivue monitors (Amsterdam, the Netherlands) and acquired at a sampling frequency of 125 Hz using BedmasterEX from 2012-2014, ICM Plus (Cambridge Enterprise, United Kingdom) from 2014-2019, and Philips Data Warehouse Connect from 2019-2020. Detailed data collection methodologies have been previously reported for the ICHOP database(10).

Definition of DWI lesions

All magnetic resonance images (MRI) were reviewed by a vascular neurologist with critical care competency and a neuroradiologist. Imaging was obtained after removal of intracranial monitors and was performed on a 1.5T or 3T scanner with diffusion-weighted imaging, apparent diffusion coefficient (ADC), fluid-attenuated inversion recovery (FLAIR), and gradient recalled echo (GRE) or susceptibility-weighted imaging (SWI) sequences. The presence of DWI lesions was defined as a high intensity signal on the DWI sequence with a corresponding low intensity signal on ADC. All lesions had to be at least 1 cm remote from the hematoma site. A representative case is illustrated in figure 1.

Figure 1:

Representative computerized tomography (CT) and magnetic resonance imaging (MRI) with susceptibility weighted imaging (SWI), diffusion-weighted imaging (DWI), and apparent diffusion coefficient (ADC) sequences. Upper left: non-contrast head CT demonstrating right frontal intracranial monitor and left frontal external ventricular drain. Upper right: SWI sequence with right thalamic intraparenchymal hemorrhage with intraventricular extension and scattered deep microhemorrhages. Bottom left: DWI sequence demonstrating scattered areas of diffusion restriction remote from the site of hemorrhage. Bottom right: apparent diffusion coefficient sequence with hypointense lesions correlating with DWI lesions.

Cerebrovascular autoregulation indices

The status of cerebrovascular autoregulation may be continuously calculated using the pressure reactivity index (PRx). The PRx was calculated as the Pearson correlation coefficient between MAP and ICP using 10 second average over a 5-minute variable window(14). When cerebrovascular autoregulation is intact, there should be minimal correlation between slow wave changes in the arterial blood pressure and ICP. However, when outside the limits of an individual’s autoregulatory threshold, the changes in arterial blood pressure will be reflected in ICP with a resultant higher correlation. PRx may be calculated continuously in patients with continuous arterial and intracranial monitoring(15). A critical PRx threshold greater than 0.25 is used to define impaired cerebrovascular autoregulation and has been shown to predict outcome in other acute brain injury(16).

With extended periods of monitoring, optimal CPP can be calculated by fitting a parabolic, U-shaped curve across 16 bins of observed CPP ranging from 40 to 120 mmHg against corresponding PRx over 2–8-hour variable windows. The CPP within the curve nadir represents the optimal CPP within an individual’s autoregulatory range(14,17). The difference between the observed CPP and calculated CPPopt is termed deltaCPP. Upper (ULA) and lower (LLA) limits of autoregulation are determined as CPP corresponding to the PRx threshold of 0.25 along the higher and lower end of the fit parabolic curve respectively. The percentage of time an individual spent above the upper limits of autoregulatory (PRx greater than 0.25 with an observed CPP higher than CPPopt) and the percentage of time spent below the lower limits of autoregulation (PRx greater than 0.25 with an observed CPP lower than CPPopt) were calculated.

All raw data files were reviewed prior to analysis, and visually identified artifacts were removed. Any data files with less than two hours of recorded data were not included in the study. Only data points with mean arterial pressures (MAP) between 40 to 210 mmHg, CPP between 20 to 210 mmHg, and ICP between 0 to 80 mmHg were considered valid. Hourly average values of all physiologic parameters were calculated from raw data files for each subject. Given prior literature on the temporal evolution of impaired cerebral autoregulation after ICH onset, we included data for up to 10 days after the estimated ictus of hemorrhage(18).

Statistical Analysis

Demographic and comorbidity variables were compared using chi-square for categorical variables and Mann-Whitney U test for continuous variables. All statistical tests were two-tailed with a significance set at p-value <0.05. Missing data was excluded. The association between continuous physiologic parameters and the presence of DWI lesions was performed using binary logistic regression analysis. Univariate associations for each parameter were calculated; a multivariate analysis was not performed due to the low number of observed outcomes. The generalized estimating equations procedure was utilized to account for repeated measures. A sensitivity analysis repeated the analysis for data recorded from days 0-5 and days 6-10 from ICH ictus to determine if any associations were modified by time from ICH ictus. All calculations were performed in SPSS version 29 (SPSS, Chicago, IL).

Results

A total of 845 patients with intracerebral hemorrhage were screened. Invasive intracranial monitoring was performed in 32 patients; eighteen patients with invasive intracranial monitoring had an MRI. Two cases were excluded due to the identification of a causative vascular malformation. Three patients were excluded due to missing or inadequate duration of continuous physiologic data. Thirteen patients with ICH met the final inclusion criteria (figure 2). Six patients demonstrated remote DWI lesions on MRI and seven had no lesions. Baseline patient demographics and clinical characteristics were not significantly different (table 1). Compared to patients without lesions, those with DWI lesions had an average age of 52.50 years vs 64.43 years and 50.0% vs 28.6% were female respectively. There were no statistically significant differences in prevalence of medical comorbidities, hematoma volume (36.33 cc vs 42.14 cc), hemorrhage location (83.3% deep vs 57.1% deep), degree of acute blood pressure reduction (delta SBP 43.17 mmHg vs 21.53 mmHg), or specific imaging features (cerebral microbleeds, leukoaraiosis, or presence of intraventricular hemorrhage).

Figure 2:

Patient inclusion flowchart illustrating selection criteria

Table 1:

Baseline demographic, clinical characteristics, and outcomes

Characteristic	DWI Absent (N=7)	DWI Present (N=6)	p-value
Age, mean (SD)	64.43 (14.75)	52.50 (11.81)	0.181
Female n (%)	2(28.6)	3 (50.0)	0.429
Race n (%)			0.722
Black	1(14.3)	0(0)
White	2 (28.6)	3(50.0)
Asian	1 (14.3)	1(16.7)
Unknown	3 (42.9)	2(33.3)
Hypertension n (%)	6 (85.7)	6 (100)	0.335
Diabetes n (%)	2 (28.6)	0 (0)	0.155
Hyperlipidemia n (%)	3 (42.9)	0 (0)	0.067
Atrial fibrillation n (%)	0 (0)	1(16.7)	0.261
Prior ischemic stroke n (%)	0 (0)	1(16.7)	0.261
Antithrombotic use n (%)	4 (57.1)	2 (33.3)	0.391
ICH score1, median (min, max)	2 (2,3)	3 (2,4)	0.181
Initial SBP2 mmHg, mean (SD)	152.00 (37.30)	186.17 (55.83)	0.234
Initial DBP3 mmHg, mean (SD)	87.14 (20.61)	105.00 (43.40)	0.445
deltaSBP4 mmHg, mean (SD)	20.86 (21.53)	43.17 (42.71)	0.445
SBP Target mmHg			0.529
140	2 (28.6)	3 (50.0)
160	4 (57.1)	3 (50.0)
180	1 (14.3)	0 (0)
Hematoma location			0.308
Lobar n (%)	3 (42.9)	1 (16.7)
Deep n (%)	4 (57.1)	5 (83.3)
IVH5 present n (%)	6(85.7)	6 (100)	0.335
Day of MRI	7.71 (7.89)	10.50 (9.23)	0.445
Hematoma volume, mean mL (SD)	42.14 (28.45)	36.33 (29.43)	0.731
Cerebral microbleeds, n(%)	3 (50)	2 (28.6)	0.429
Leukoariosis, n(%)	1 (16.7)	1 (14.3)	0.906
90-day mRS6, median (min, max)	5 (4,5)	5 (0,6)	0.836
90 day mortality, n(%)	0 (0)	1 (16.67)	0.261

⁴Delta systolic blood pressure (difference between systolic blood pressure at presentation and 24 hours),

³diastolic blood pressure,

¹intracerebral hemorrhage score,

⁵intraventricular hemorrhage,

⁶modified Rankin scale,

²systolic blood pressure

A total of 1,135 hours of physiologic data were recorded during the first 10 days after ictus. Data was missing or artifactual for 13.3% of MAP and 7.9% of ICP. A total of 984 hours of artifact free data was included for analysis with 505 cumulative hours of data from patients without DWI lesions and 479 hours of physiological data from those with DWI lesions (supplementary table 1). PRx was able to be calculated in 98.98%, CPPopt in 70.22%, and limits of autoregulation in 69.72% of the artifact-free data. No physiologic data was available within the first 24 hours from ictus. The average time from hemorrhage ictus to the onset of ICP monitoring was 3.58±1.77 days, and patients were recorded for an average of 97.6±79.7 hours. Supplementary figures 2 and 3 illustrate distribution plots of monitoring duration for individual subjects and for each day from ICH ictus respectively. MRI was acquired an average of 9.0± 8.3 days from hemorrhage ictus.

Over the first 10 days from hemorrhage onset for the entire cohort, the average MAP was 93.78 ± 13.99 mmHg, ICP was 14.12±8.02 mmHg, and CPP was 78.50±18.13 mmHg. The average autoregulatory values were PRx of 0.24 ± 0.24, CPPopt of 82.04±13.61 mmHg, Delta CPP of −1.71±11.06 mmHg, ULA 86.64±16.63 mmHg, and LLA 75.61±11.58 mmHg. The average time spent below LLA was 32.07%, and the average time spent above ULA was 17.78%.

Compared to those without DWI lesions, patients with DWI lesions demonstrated a significantly higher ICP (17.50±8.26 mmHg vs 10.92±6.29 mmHg; OR 1.14, CI 1.01-1.29). No significant differences were demonstrated between MAP and CPP (table 2). Regarding autoregulatory measures, the percentage of time spent below LLA was significantly associated with the presence of DWI lesions. In patients with DWI lesions, 49.85% of the monitoring period was below LLA compared to 14.70% in patients without DWI lesions (OR 5.77, CI 1.88-17.75). No significant associations were demonstrated between the presence of DWI lesions and PRx, CPPopt, DeltaCPP, ULA, LLA, or time spent above ULA (table 3). Figure 3 illustrates differences in time spent within, above, and below autoregulatory limits between patients with or without DWI lesions.

Table 2:

Physiologic parameters and association with presence of DWI lesions.

	Total	DWI Absent	DWI Present	OR (CI)	P-value
MAP1
Mean mmHg (SD)	93.78 (13.99)	96.65 (14.89)	90.75 (12.28)	0.97 (0.92-1.02)	0.232
ICP2
Mean mmHg (SD)	14.12 (8.02)	10.92 (6.29)	17.50 (8.26)	1.14 (1.01-1.29)	0.033
CPP3
Mean mmHg (SD)	78.50 (18.13)	85.91 (17.49)	73.23(15.82)	0.95 (0.91-1.00)	0.061

³Cerebral perfusion pressure,

²Intracranial pressure,

¹mean arterial blood pressure.

Table 3:

Autoregulation metrics and association with presence of DWI lesions.

	Total	DWI Absent	DWI Present	OR (CI)	P-value
PRx1
Mean (SD)	0.24 (0.24)	0.17 (0.21)	0.32 (0.24)	18.24 (0.21-1556.43)	0.201
CPPopt2
Mean mmHg (SD)	82.04 (13.61)	84.46 (13.93)	79.52 (12.82)	0.97 (0.92-1.03)	0.355
DeltaCPP3
Mean mmHg (SD)	−1.71 (11.06)	0.10 (11.29)	−3.59 (10.51)	0.97 (0.93-1.01)	0.087
LLA4
Mean mmHg (SD)	75.61 (11.58)	77.06 (10.83)	75.07 (11.80)	0.99 (0.93-1.05)	0.625
Time spent below LLA4
Hours (Percentage of recording)	220/686 (32.07%)	51/347 (14.70%)	169/339 (49.85%)	5.77 (1.88-17.75)	0.002
ULA5
Mean mmHg (SD)	86.64 (16.63)	87.60 (15.58)	86.09 (17.18)	1.00 (0.94-1.05)	0.843
Time spent above ULA5
Hours (Percentage of recording)	122/686 (17.78%)	39/347 (11.24%)	169/339 (24.48%)	2.56 (0.73-9.01)	0.143

³Delta cerebral perfusion pressure,

⁴lower limit of autoregulation,

²optimal cerebral perfusion pressure,

¹pressure reactivity index,

⁵upper limit of autoregulation

Figure 3:

Percent of time spent within, above, or below limits of cerebral autoregulation in patients with and without DWI lesions.

Sensitivity analysis segregating the data into recordings from days 0-5 from ICH ictus and days 6-10 from ictus found no modification of the association between time spent below LLA and DWI lesion development in both the early (47.56% vs 7.45%; OR 5.74 CI 1.04-31.66) and later period (50.58% vs 17.39%; OR 4.86 CI 1.23-19.21) (supplementary tables 2 and 3). In the first 5 days from ictus, a more negative deltaCPP (−3.40± 9.48 mmHg vs 0.91±13.11 mmHg; OR 0.97 CI 0.94-0.99) and greater time spent above ULA (37.80% vs 9.57%; OR 5.74 CI 1.04-31.66) were also associated with DWI lesion development. In the later period, lower CPP (69.96±15.82 mmHg vs 88.93±17.13 mmHg; OR 0.93 CI 0.88-0.98) and higher ICP (18.75±8.84 mmHg vs 10.34±5.46 mmHg; OR 1.20 CI 1.04-1.38) were associated with DWI lesions. Differences in time spent within, above, and below autoregulatory limits segregated by time from ICH onset are depicted in supplementary figure 1.

Discussion

To our knowledge, this is the first study to demonstrate an association between suboptimal individualized cerebral perfusion pressure with the development of DWI lesions after ICH. This provides an evidentiary basis towards a plausible pathophysiologic mechanism of secondary ischemic lesions after primary ICH.

After ICH, patients are at risk for clinical and subclinical ischemia during both the acute and subacute period. ICH confers an elevated risk of ischemic stroke, and the presence of DWI lesions on MRI increases the risk by more than two-fold (19,20). This short-term elevated stroke risk is highest within the first month from hemorrhage(19). Additionally, while DWI lesions are seen acutely in one third of cases, new ischemic lesions occur in a quarter of patients between MRIs obtained at baseline and one month after ictus (21). The underlying mechanism of ischemic injury remains unclear with multiple competing etiologies such as antithrombotic withdrawal, activation of prothrombotic factors, local inflammatory effects, or ongoing microvascular hypoperfusion. Multiple studies have investigated the impact of acute blood pressure reduction on the development ischemia after ICH with contradictory results between various cohorts (2,22–24). While early blood pressure reduction is associated with a reduction in hematoma expansion, overly aggressive reduction in those who present extremely hypertensive may result in secondary ischemic injury (24,25). A pooled analysis of several large prospective studies found that while higher presenting systolic blood pressure is associated with DWI lesions develop, it failed to demonstrate an association between the degree of blood pressure reduction and DWI lesions(1). These results may be explained by heterogeneity in autoregulatory ranges among patients with various severity of pre-morbid hypertension and degrees of underlying vasculopathy. The shift of the lower limits of cerebral autoregulation in patients with chronic hypertension is well established and informs the intensity of blood pressure reduction that can be instated without inducing cerebral hypoperfusion(26). There is substantial evidence that patients with features of chronic microangiopathy such as leukoaraiosis and dilated perivascular spaces have higher prevalence of DWI lesions(1,3,22). Cerebral amyloid angiopathy is known to impair cerebral autoregulation as well, which may explain the similar rates of DWI lesion development irrespective of hemorrhage etiology(27,28). Furthermore, the impact of increased intracranial pressure on cerebral perfusion is also variable between patients, depending on several factors such as hematoma volume, perihematomal edema, degree of brain atrophy, the presence of intraventricular hemorrhage, and hemorrhage location. In our cohort, despite maintaining a cerebral perfusion pressure target of greater than 60 mmHg, we found a markedly greater proportion of the monitored period spent below the lower limits of autoregulation in individuals who developed ischemia on MRI. Autoregulation-informed, individualized blood pressure target selection presents a novel approach to reduce the risk of cerebrovascular ischemic events and explain the overall lack of significant improvement in outcomes seen in multiple randomized controlled trials enforcing a monolithic approach to blood pressure target selection (29).

Impairment in cerebrovascular autoregulation after ICH is poorly studied. Small studies of static autoregulation utilizing SPECT, PET, or CT perfusion imaging have demonstrated conflicting results as to whether acute blood pressure reductions decrease cerebral blood flow after ICH(30–32). However, static measures of autoregulation fail to capture the evolving and responsive properties of the cerebrovasculature in reaction to individual hemodynamic changes. Alternatively, dynamic assessments of autoregulation elucidate the ability of cerebrovasculature to maintain adequate cerebral blood flow within a specific range of arterial blood pressures over a broad temporal window. A variety of techniques to assess dynamic cerebral autoregulation have been described including invasive modalities such as intracranial pressure monitoring as well as non-invasive metrics using transcranial doppler (TCD). TCD based techniques of dynamic cerebral autoregulation using either transfer function analysis or correlation coefficient between arterial pressure and middle cerebral artery velocity oscillations have demonstrated autoregulation impairments in certain patients which may be associated with worsened outcomes(33,34). Previous studies utilizing invasive metrics of autoregulation similar to our cohort have demonstrated elevated PRx to be associated with poor outcome(7,35). In the largest series of thirty-eight subjects with severe ICH, the amount of time spent above a critical PRx threshold of 0.2 correlated with worsened mRS at 3 months(36). Within the same study, mortality was lowest in patients with CPP maintained closest to the calculated CPPopt(36). In our small cohort, patients with DWI lesions demonstrated a non-significant trend towards elevated PRx and a more negative deltaCPP (table 3). It is uncertain whether the association of impaired autoregulatory indices with DWI lesions is merely a reflection of a greater severity of baseline small vessel disease or if impaired autoregulation is directly causative in the pathogenesis of DWI lesions. Future investigations combining non-invasive autoregulatory indices with serial MRIs are necessary to demonstrate whether individualized blood pressure ranges are predictive of future ischemic events.

The interpretation of this study is limited by several factors. Due to selection criteria and rarity of invasive intracranial monitoring use, the sample size of the study was small and prohibitory for multivariate analysis of autoregulatory and other physiologic parameters. The generalizability of the results is further limited as patients who receive intracranial monitoring have poor clinical exams and reflect greater severity than what is seen in the overall ICH population. Although there was no significant difference in baseline demographics or clinical characteristics between the two groups, imbalances in risk factors for DWI lesions after ICH may have collinearity with impairment in cerebrovascular autoregulation. Since this was a retrospective study over several years, other unaccounted-for confounders limit the interpretation of the results to hypothesis generation such as differences in MRI imaging equipment and protocols, inter-provider differences in management, and changes in medical care over the duration of enrollment. Due to the need for patient consent, hematoma stability, coagulopathy reversal, and initial stabilization of patients upon first arrival, no patients had placement of invasive monitoring within the first 24 hours, and there was heterogeneity in the timing from ictus to placement and total duration of monitoring. Although we found a similar association between greater time spent below LLA and DWI lesion development in both days 0-5 and days 6-10 from ICH, the impact of impaired cerebrovascular autoregulation on the development of secondary ischemia may be time dependent. Due to the inherent constraints in calculation of CPPopt, the limits of autoregulation were unable to be determined in 30.3% of the artifact-free recording time. Continuous arterial pressures were derived from arterial catheters calibrated at the phlebostatic axis rather than the tragus which is preferred in measurement of CPP. Since patients in our institution spent most of the recording with their head elevated to 30 degrees, the absolute value observed CPP and CPPopt is likely over-estimated by 8-10 mmHg(37). Although absolute values of CPP and CPPopt may be skewed, this does not impact autoregulatory indices of deltaCPP, PRx, or time outside of autoregulatory limits.

Conclusions

Blood pressure target selection based upon assessment of individual cerebral autoregulatory limits is a potential treatment paradigm to improve outcomes after intracerebral hemorrhage. Given the association of poor outcomes with DWI lesions after ICH and the relatively high prevalence from prior studies, prevention of the development of these lesions presents a possible target to improve outcomes after ICH. Prospective study of the impact of individualized, autoregulation-informed blood pressure targets on ischemic injuries after ICH is warranted.