Article Text
Abstract
Objective Permissive hypertension may benefit patients with non-recanalized large vessel occlusion (nrLVO) post mechanical thrombectomy (MT) by maintaining brain perfusion. Data evaluating the impact of post-MT blood pressure (BP) levels on outcomes in nrLVO patients are scarce. We investigated the association of the post-MT BP course with safety and efficacy outcomes in nrLVO.
Methods Hourly systolic BP (SBP) and diastolic BP (DBP) values were prospectively recorded for 24 hours following MT in consecutive nrLVO patients. Maximum, minimum, and mean BP levels were documented. Three-month functional independence (FI) was defined as modified Rankin Scale (mRS) scores of 0–2.
Results A total of 88 nrLVO patients were evaluated post MT. Patients with FI had lower maximum SBP (160±19 mmHg vs 179±23 mmHg; P=0.001) and higher minimum SBP levels (119±12 mmHg vs 108±25 mmHg; P=0.008). Maximum SBP (183±20 mmHg vs 169±23 mmHg; P=0.008) and DBP levels (105±20 mmHg vs 89±18 mmHg; P=0.001) were higher in patients who died at 3 months while minimum SBP values were lower (102±28 mmHg vs 115±16 mmHg; P=0.007). On multivariable analyses, both maximum SBP (OR per 10 mmHg increase: 0.55, 95% CI 0.39 to 0.79; P=0.001) and minimum SBP (OR per 10 mmHg increase: 1.64, 95% CI 1.04 to 2.60; P=0.033) levels were independently associated with the odds of FI. Maximum DBP (OR per 10 mmHg increase: 1.61; 95% CI 1.10 to 2.36; P=0.014) and minimum SBP (OR per 10 mmHg increase: 0.65, 95% CI 0.47 to 0.90; P=0.009) values were independent predictors of 3-month mortality.
Conclusions Our study demonstrates that wide BP excursions from the mean during the first 24 hours post MT are associated with worse outcomes in patients with nrLVO.
- blood pressure
- outcome
- systolic blood pressure
- diastolic blood presure
- mechanical thrombectomy
- recanalization
- emergent large vessel occlusion
- stroke
Statistics from Altmetric.com
- blood pressure
- outcome
- systolic blood pressure
- diastolic blood presure
- mechanical thrombectomy
- recanalization
- emergent large vessel occlusion
- stroke
Introduction
Recent landmark randomized controlled clinical trials (RCTs) with successful recanalization (SR) rates ranging from 66% to 88% have provided unequivocal evidence that endovascular reperfusion with mechanical thrombectomy (MT) in patients with large vessel occlusion (LVO) is safe and improves functional outcomes.1–5 MT has become the new standard of care for acute ischemic stroke (AIS) due to LVO and health systems are reorganizing in order to implement this new therapy as quickly and as widely as possible.6
Current international recommendations on MT for AIS inadequately address blood pressure (BP) management before, during and after MT, given the lack of RCT data.7 8 In the absence of specific guidelines, endovascular centers usually target BP to remain below 180/105 mmHg according to current American Heart Association (AHA) recommendations for BP management following intravenous thrombolysis (IVT) for AIS.9 Preliminary data have indicated that higher maximum values of systolic blood pressure (SBP) during the first 24 hours following MT are independently associated with worse functional outcomes, higher mortality, and a higher rate of hemorrhagic complications in patients with LVO.10 11
Theoretically, BP control may be beneficial in LVO patients with SR following MT, since they are at risk of reperfusion hemorrhage following effective clot removal.12–14 Our recent single-center study has provided preliminary data in support of this hypothesis, since it showed that moderate BP control in the first 24 hours following MT was related to lower odds of 3-month mortality compared with permissive hypertension in successfully recanalized LVO patients.11
Conversely, permissive hypertension may be beneficial in patients with non-recanalized LVO (nrLVO) by maintaining adequate brain perfusion pressure.12 Additionally, aggressive BP reduction in the first 24 hours of ictus has been associated with poor outcomes.15 Moreover, it may be argued that allowing the BP to spontaneously rise to optimize collateral blood flow may also assist in clearing emboli from distal vessels.16 However, the data evaluating the impact of permissive hypertension in the post-MT period on outcomes in nrLVO patients are scarce. In view of these considerations, we sought to evaluate the association of BP course in the first 24 hours following MT with safety and efficacy outcomes of patients with nrLVO.
Methods
Study population
Data from consecutive LVO patients who underwent MT at a tertiary care stroke center from July 2013 to December 2016 were prospectively collected as previously described.17–19 At our comprehensive stroke center, the protocol-based practice is to perform MT for patients with CT angiography (CTA)-confirmed LVOs presenting within 6 hours of symptom onset. For the patients presenting 6–12 hours from symptom onset, additional imaging selection criteria are used such as Alberta Stroke Program Early CT Score (ASPECTS) ≥6 on brain CT and/or good collaterals on CTA. LVO was diagnosed on pretreatment CTA using standard criteria.20 Brain CT head and CTA were performed on a 64-slice scanner. CT slice thickness was 1.25 mm with acquisitions in axial, sagittal, and coronal planes with 3D reformatting. A blinded neuroradiologistcalculated ASPECTS and CTA collateral scores as previously described.17 Collateral scores were reported in a dichotomized fashion (good vs poor collaterals) for anterior circulation LVO using a methodology that has been shown to predict clinical outcome.17
Recanalization at the end of the endovascular procedure was defined by modified Thrombolysis in Cerebral Infarction (mTICI) grades,21 which were obtained from the reports of endovascular specialists (LE, AA, DH). For the purpose of this study, successfully recanalized (mTICI grades 2b or 3) LVO patients were excluded from the analyses and only nrLVO patients (mTICI grades 0, 1, 2a) were further evaluated. At our center, AIS patients with emergent LVO are treated according to the AHA guidelines advocating IVT for all eligible patients and subsequent MT in cases with confirmed LVO by CTA. Some of the common reasons for not delivering IVT were delayed time window on presentation, pretreatment with oral anticoagulation therapy, and uncontrolled hypertension. In these situations, patients underwent direct MT without IVT pretreatment. Cerebral edema was determined to be present if both of the following criteria were met on two or more axial diffusion-weighted imaging (DWI) slices of the brain MRI which all patients underwent within 24–48 hours from symptom onset: (1) direct evidence of mass effect of affected gyri; or (2) indirect evidence based on new distortion of adjacent tissue, new midline shift, or new effacement of sulci or lateral ventricle. This two-by-two method was employed to reduce the chance of misclassifying infarct growth as swelling.22
Baseline characteristics
Stroke severity at hospital admission was documented using the National Institute of Health Stroke Scale (NIHSS) score by certified vascular neurologists. Baseline characteristics including demographics, vascular risk factors, admission NIHSS scores, admission ASPECTS, pretreatment with IVT, admission serum glucose, admission serum low density lipoprotein (LDL), admission systolic blood pressure (SBP) and diastolic blood pressure (DBP) levels were recorded as previously described.11 20 23
BP management post MT
Hourly SBP and DBP recordings were obtained in all LVO patients during the first 24 hours following MT using oscillometric non-invasive BP monitoring devices as previously described.9 We documented minimum, mean, and maximum SBP and DBP levels for each individual patient during the BP monitoring period. Standard of care at our tertiary care center is to allow permissive hypertension in nrLVO patients post MT with BP goal of <220/120 mmHg for patients without pretreatment with IVT according to AHA recommendations for AIS management.24 For patients pretreated with IVT, permissive hypertension was set at <180/105 mmHg as advocated by AHA recommendations.24 Intravenous nicardipine was the first-line BP-lowering medication that we used for treating patients with BP levels exceeding the former prespecified cut-offs. Second-line antihypertensive agents were intravenous labetalol or intravenous enalapril. If nrLVO patients developed symptomatic intracranial hemorrhage (sICH) post MT, BP goals were modified according to the treating physician’s preference (<140/90 mmHg or <160/90 mmHg). Institutional Investigation Review Board approval for this study was granted based on the prospectively maintained Acute Ischemic Stroke Database. The Board waived the need for patient consent.
Definition of outcomes
sICH was defined as the presence of a parenchymal hematoma type 2 on brain CT and/or MRI gradient recall echo sequence accounting for deterioration with an increase in NIHSS score of ≥4 points within 36 hours from treatment.25 Functional outcome was evaluated at 3 months using the modified Rankin Scale (mRS) scores by certified vascular neurologists. Functional status was evaluated in person either from a post hospital discharge clinic follow-up or from a hospital visit in all patients. Functional independence (FI) was defined as mRS scores of 0–2.23 Endovascular specialists grading the degree of reperfusion at the end of MT and vascular neurologists assessing NIHSS and mRS scores were unaware of the purposes of the study and performed treatments and assessments as part of their clinical duties. The primary outcome included 3-month FI, while sICH and 3-month mortality represented secondary outcomes. Final infarct volume (FIV) was also calculated by the simplification of the ellipsoid rule A×B×C/2 on delayed (7–14 days) brain MRI using Fluid Averted Inversion Recovery (FLAIR) sequences.20 Delayed brain CT was used in a small number of patients with contraindications to brain MRI.20
Statistical analyses
Continuous variables are presented as mean±SD (normal distribution) and as median with interquartile range (skewed distribution). Categorical variables are presented as percentages with their corresponding 95% CIs. Statistical comparisons for categorical variables between two or three groups were performed using a χ2 test, or in case of small expected frequencies, Fisher’s exact test. Continuous variables were compared using the unpaired t-test, Mann–Whitney U test, one-way analysis of variance or Kruskal–Wallis test as indicated. The correlation of BP measurements with FIV was evaluated using Spearman’s correlation coefficient (r). Univariable and multivariable logistic regression models were used to evaluate associations between BP measurements during the first 24 hours following MT (mean, minimum, and maximum SBP and DBP post MT) with 3-month FI, 3-month mortality and sICH before and after adjustment for the following potential confounders: age, gender, race, vascular risk factors (hypertension, diabetes mellitus, hypercholesterolemia, atrial fibrillation, coronary artery disease, congestive heart failure and current smoking), admission SBP and DBP levels, admission stroke severity, admission serum glucose and LDL levels, pretreatment with IVT, baseline ASPECTS, onset to groin puncture time, type of anesthesia (general anesthesia vs conscious sedation), brain edema, and good collaterals. Associations are presented as OR with corresponding 95% CI. In the initial univariable analyses a P value <0.1 was set as the threshold for inclusion in the multivariable models. Statistical significance was achieved for P<0.05 in the final multivariable analyses. The Statistical Package for Social Science (SPSS Version 22.0 for Windows) was used for statistical analyses.
Results
Our study population consisted of 88 nrLVO patients (mean age 62±15 years, 48% median NIHSS score 16 points (IQR 12–21), median onset to groin puncture time 226 min (IQR 171–290)). The baseline characteristics of our study population are shown in table 1. The location of LVO was the following: 31 in the M1 middle cerebral artery (MCA; 35%), 16 in the intracranial internal carotid artery (ICA; 18%), 12 in the posterior circulation (14%), 10 in M2 MCA (11%). The remaining 19 patients had tandem occlusions (22%). A total of 33 patients underwent MT under general anesthesia (38%), while the remaining 55 patients (62%) received conscious sedation.
In our cohort, a combination of distal aspiration and stent-retriever was used as the most common technique in 47 cases (53%). The second most commonly employed technique was the ADAPT (A Direct Aspiration First Pass Technique), and this was used in 22 cases (25%). Stent-retrievers without distal aspiration were used in 15 cases (17%). The remaining cases were mostly cervical or intracranial carotid occlusions in which acute angioplasty and/or stenting was employed. Systemic thrombolysis prior to MT was administered in 56 patients (64%). The rate of sICH was 9%. Cerebral edema occurred in 32 patients (37%). A total of 27% of patients achieved FI at 3 months, while the 3-month mortality rate was 37%.
The univariable associations of maximal SBP and DBP levels post MT with primary (FI) and secondary (sICH and 3 month mortality) outcomes are summarized in table 2. Patients with FI had lower maximal SBP (160±19 mmHg vs 179±23 mmHg; P=0.001), while maximal DBP did not differ between the two groups (88±14 mmHg vs 96±22 mmHg; P=0.157). Maximal SBP levels (183±20 mmHg vs 169±23 mmHg; P=0.008) and DBP levels (105±20 mmHg vs 89±18 mmHg; P=0.001) post MT were higher in patients who died at 3 months, while no difference was noted in maximal SBP (183±31 mmHg vs 173±22 mmHg; P=0.279) and DBP (97±25 mmHg vs 94±20 mmHg; P=0.692) levels in patients with and without sICH. Increasing maximal SBP measurements strongly correlated to larger FIV (Spearman’s r: 0.534; P<0.001). There was no correlation (P>0.080) between any of the remaining BP measurements and FIV.
Table 2 also depicts the associations of minimum SBP and DBP levels post MT with primary and secondary outcomes. Patients with FI had higher minimum SBP levels (119±12 mmHg vs 108±25 mmHg; P=0.009), while minimal SBP values were lower in patients who died at 3 months (102±28 mmHg vs 115±16 mmHg; P=0.007). No other association was detected between minimal SBP and DBP values post MT with the outcomes of interest. We also failed to document any association between mean SBP and DBP levels post MT and primary or secondary outcomes (table 3). Moreover, all post MT BP measurements did not differ (P>0.2) between patients with good and bad collateral status.
Table 4 shows the univariable and multivariable associations of BP measurements post MT and other baseline characteristics with 3-month FI. In the initial univariable analyses the following variables were related (P<0.1) to FI: admission SBP, admission NIHSS score, maximum SBP post MT and minimum SBP post MT. In the final multivariable models both minimum and maximum SBP were independently associated with FI. More specifically, a 10 mmHg increment in maximum SBP reduced the odds of FI (OR 0.55, 95% CI 0.39 to 0.79; P=0.001), while a 10 mmHg increment in minimum SBP increased the likelihood of FI (OR 1.64, 95% CI 1.04 to 2.60; P=0.033).
Table 5 summarizes the univariable and multivariable associations of BP measurements post MT and other baseline characteristics with 3-month mortality. In the initial univariable analyses the following variables were related (P<0.1) to mortality: admission SBP, admission NIHSS score, general anesthesia (vs conscious sedation), maximum SBP post MT, minimum SBP post MT, and maximum DBP post MT. In the final multivariable models both minimum SBP and maximum DBP were independently associated with mortality. More specifically, a 10 mmHg increment in maximum DBP increased the odds of mortality (OR 1.62; 95% CI 1.10 to 2.38; P=0.014), while a 10 mmHg increment in minimum SBP decreased the likelihood of mortality (OR 0.68, 95% CI 0.49 to 0.95; P=0.022).
We also performed sensitivity analyses (table 6) evaluating the association of post-MT BP measurements with outcomes in the subgroup of nrLVO patients without IVT pretreatment (n=32). Our sensitivity analyses reproduced the significant associations that were documented in the total study sample. Patients with 3-month FI had lower maximal SBP (160±19 mmHg vs 177±17 mmHg; P=0.040), while maximal SBP levels (181±8 mmHg vs 167±22 mmHg; P=0.025) post MT were higher in patients who died at 3 months. In addition, lower minimal SBP levels were also related to 3-month mortality (101±18 mmHg vs 115±18 mmHg; P=0.038).
Discussion
Our study demonstrates that wide BP excursions from the mean during the first 24 hours post MT are associated with worse outcomes in patients with nrLVO. This association was independent of demographics, vascular risk factors, baseline stroke severity, onset to groin puncture time, admission BP, serum glucose and LDL levels, early hypodensity on brain CT quantified by ASPECTS, type of anesthesia, collateral status, and early brain edema.
Successful reperfusion rates post MT have improved from 40–50% with first-generation devices to 70–90% with third-generation devices like stent-retrievers or aspiration catheters during the past 15 years.26 27 Considering the fact that 10–30% of LVO patients still remain non-recanalized post MT carrying a dismal prognosis, it is important to identify other modifiable variables that might improve functional outcomes in this subgroup. Effective post-procedural BP management may be one of these factors. Theoretically, permissive hypertension may be beneficial in patients with nrLVO to maintain brain perfusion pressure. In our cohort of nrLVO patients, we documented that increasing maximum SBP and DBP levels during the first 24 hours following MT were associated with a higher likelihood of 3-month mortality and lower odds of 3-month FI. This finding is partly in line with a recent retrospective observational study that also reported an independent relationship between peak SBP levels during the first 24 hours following MT and worse functional outcomes at 3 months in a broader patient population (both recanalized and non-recanalized LVO patients).10 On the other hand, minimum post-MT BP levels did not affect outcomes in the study by Mistry et al.10 Nevertheless, there are important methodological considerations when comparing the two studies. First, the study population of patients with non-recanalized emergent LVO in our study (n=88) is almost double that in the study by Mistry et al (n=45).10 Second, in the earlier study by Mistry and colleagues only 34 nrLVO patients had 90-day mRS evaluations available.10 Therefore, their analysis was underpowered to reveal independent associations between post-MT BP measurements and 90-day functional outcomes. On the other hand, our study evaluated the association of post-MT BP levels with safety and efficacy outcomes using comprehensive univariable and multivariable logistic regression models.
The association of elevated maximum BP post MT with adverse functional outcomes cannot be attributed to the potential impairment of collateral circulation with extremely elevated BP levels in patients with LVO, since patients with good and bad collaterals in our cohort did not differ in any BP parameter post MT.28 The association of increasing BP levels with increasing final infarct volume seen in the present study, detected both by our group20 and other investigators in previous studies,29 may account for the detrimental effect of higher maximum SBP and DBP post MT on functional outcomes. Alternative explanations may suggest that patients with exceptionally high BP are unhealthier due to coexisting comorbidities (diabetes mellitus, renal failure) or that acutely raised BP levels are a marker of extensive cerebral ischemia and early brain swelling. However, it is noteworthy that brain edema was included as a potential confounder in all analyses without modifying any of the reported associations.
In addition, we detected that lower minimum SBP levels post MT correlated to higher odds of mortality and functional dependence at 3 months. A U-shaped association between acute BP values and outcomes in AIS has been firmly established by multiple observational studies, indicating that extremely low BP levels during the first hours of acute cerebral ischemia are equally as harmful as markedly elevated BP values.29–31 A similar U-shaped relationship between baseline SBP and functional outcome in LVO patients treated with MT has recently been reported by the MR CLEAN (Multicenter Randomized Clinical Trial of Endovascular Treatment of Acute Ischemic Stroke in the Netherlands) investigators.32 The detrimental effect of extremely low BP values following MT in patients with nrLVO may be attributed to impaired cerebral perfusion due to autoregulatory failure leading to infarct expansion and further neurological deterioration.14 33
Nevertheless, the fact that maximum and minimum and not mean BP levels were associated with adverse outcomes indirectly supports the potential detrimental effect of extreme BP fluctuations during the immediate post-procedural period. The findings of our study are corroborated by a recent retrospective study which demonstrated worse clinical outcomes in non-recanalized AIS patients with the highest and the lowest BP values following reperfusion with IVT or endovascular procedures.33 Additional well-designed multicenter studies are required to independently validate our initial observations and to determine the cut-off values above and below which BP should not fluctuate during the first critical hours following MT.
It is plausible to hypothesize that, among patients with nrLVO, those with viable penumbra are most likely to benefit from permissive hypertension. Although there is no easy way to quantify penumbra, high ASPECTS and good collaterals can be used as surrogates for viable hypoperfused cerebral tissue. In the present study, we included in our analyses both ASPECTS and collateral score as potential confounders that may moderate the association of BP measurements with 3-month FI (mRS score of 0–2) and 3-month mortality (tables 4 and 5). However, we failed to detect any significant interaction. Nevertheless, this lack of association may be attributed to a type II error due to the limited sample size (n=88). Future multicenter larger studies should evaluate the potential beneficial effect of permissive hypertension following MT in the subgroup of patients with emergent LVO with adequate collaterals and high ASPECTS.
It is also possible that permissive hypertension may have resulted in small improvements in hand or language functions in nrLVO patients. However, these improvements could not be quantified by the mRS score, which is a course outcome. Unfortunately, the activities of daily living at 3 months were not prospectively evaluated in this dataset. We plan to implement Barthel Index as a standard outcome measure of activities of daily living at 3 months in our future studies. Furthermore, we failed to assess volume status, a potentially essential variable for patients with pressure dependent examination that may benefit the most with permissive hypertension. Nevertheless, it should be acknowledged that the common practice in our center is to maintain euvolemia using fluid administration.
Several other limitations of the present report need to be addressed. First, the modest sample size and retrospective analysis of prospectively collected data are important methodological shortcomings. Second, our methods included primary and secondary outcomes lacking central adjudication. Third, we did not collect data on fluctuation of BP levels during MT and BP variability post MT and could not evaluate their impact on different outcomes. Fourth, the limited sample size prevented us from differentiating between mTICI 0, 1 and 2a recanalization grades when investigating the impact of BP levels post MT on outcomes. We realize that this is an important limitation, however we feel that stratifying the present cohort according to different non-recanalized grades (mTICI 0, 1 and 2a) will further significantly decrease the sample size in each group. This would make it difficult to draw any meaningful results/conclusions from the present study because of type II error. Future studies with a larger number of patients should stratify these groups. Fifth, patients with nrLVO pretreated with IVT had different post-MT BP goals (<180/105 mmHg) from patients treated with direct MT (<220/120 mmHg). Consequently, it would be important to differentiate these two groups in future larger studies, although it is important to note that IVT pretreatment was included as one of the potential confounders in our analyses and did not moderate the association of BP parameters with early functional outcomes and mortality. Sixth, follow-up vascular imaging after the termination of MT is not common practice at our center unless there is unexpected clinical worsening. Thus, we were unable to record the rates of spontaneous subsequent recanalization in the present dataset. Finally, the observational study design did not allow us to establish a cause–effect relationship between BP levels and outcomes in patients with nrLVO treated with MT. The strong correlation that we identified between increasing maximal SBP measurements and more extensive infarct volumes indicates that excessively high BP levels post MT may worsen functional outcome by increasing FIV.
In conclusion, our study demonstrates that nrLVO patients who do poorly are more likely to have had excursions of BP from the mean in the first 24 hours following MT. In the absence of RCT data, the majority of endovascular centers usually allow permissive hypertension for patients with nrLVO. Our preliminary findings indicate that extreme BP variations during the first 24 hours following MT are likely to be harmful and should be avoided. Additional well-designed multicenter studies are required to independently validate our initial observations and determine the actual cut-offs above and below which BP should not fluctuate during the first critical hours following MT. Future research is also necessary to identify the subgroup among nrLVO patients who are most likely to benefit from permissive hypertension (eg, patients with good collaterals and high ASPECTS).
References
Footnotes
Contributors NG: Study concept and design, acquisition of data, analysis and interpretation, critical revision of the manuscript for important intellectual content. GT: Analysis and interpretation, critical revision of the manuscript for important intellectual content. AP: Acquisition of data, critical revision of the manuscript for important intellectual content. KD: Acquisition of data, critical revision of the manuscript for important intellectual content. DA: Acquisition of data, critical revision of the manuscript for important intellectual content. JJC: Critical revision of the manuscript for important intellectual content. BK: Acquisition of data, critical revision of the manuscript for important intellectual content. CN: Acquisition of data, critical revision of the manuscript for important intellectual content. DH: Acquisition of data, critical revision of the manuscript for important intellectual content. KA: Critical revision of the manuscript for important intellectual content. AVA: Critical revision of the manuscript for important intellectual content. ASA: Acquisition of data, critical revision of the manuscript for important intellectual content. LE: Study concept and design, study supervision, critical revision of the manuscript for important intellectual content.
Competing interests DH: consultant for Codman Neurovascular, Medtronic, MicroVention, Penumbra, Sequent, and Stryker; ASA: consultant for Codman, Medtronic, Microvention, Penumbra, Sequent, Siemens, Stryker and has received research support from Sequent and Siemens; LE: consultant for Codman Neurovascular, Medtronic, MicroVention, Penumbra, Sequent, and Stryker.
Ethics approval University of Tennessee Health Sciences Center Acute Ischemic Stroke Database (10-01003-XP).
Provenance and peer review Not commissioned; externally peer reviewed.
Presented at The study findings were partly presented at the Annual Meeting of the Society of Neurocritical Care, 2017.