Blood pressure behavior during mechanical thrombectomy and drugs used for conscious sedation or general anesthesia

ABSTRACT Background: The optimal blood pressure (BP) during mechanical thrombectomy for acute ischemic stroke is currently unclear. Objective: To investigate BP behavior during mechanical thrombectomy in patients with acute ischemic stroke and its relationship with drugs used for sedation or general anesthesia. Additionally, we investigated the association between BP oscillation during mechanical thrombectomy and recanalization status, and with functional outcome at discharge. Methods: Consecutive patients treated with mechanical thrombectomy for acute ischemic stroke were evaluated in a tertiary hospital from December/2009 to December/2015. Maximum, minimum, and mean systolic and diastolic BP, and mean arterial pressures were collected during the procedure. Sedative drugs were also reviewed. Results: Fifty-three patients with a mean age of 71.9 years (60.4% men) were treated with mechanical thrombectomy. The mean reduction in systolic BP and mean arterial pressure from hospital admission to mechanical thrombectomy were respectively 42 and 36 mmHg. During the procedure, oscillations were 50.4 mmHg for systolic, and 33.2 mmHg for diastolic BP. Patients treated with neuromuscular blocking drugs had more oscillation in systolic BP from hospital admission to procedure (51.1 versus 26.2 mmHg, P =0.06). The use of cisatracurium (43.9 versus 29.6 mmHg, P =0.02) and succinylcholine (44.7 versus 29.3 mmHg, P =0.01) were associated with a significant drop in BP during the procedure. Conclusions: Significant BP oscillation occurs during mechanical thrombectomy. Drugs used for conscious sedation or general anesthesia, specifically neuromuscular blocking agents, might have an influence upon BP levels.


iNtrODUctiON
Mechanical thrombectomy (MT) has become the standard therapy for patients with stroke caused by large vessel occlusions (LVO) up to 24 hours from time last-knownwell (LKW) [1][2][3][4] . Among several factors associated with higher recanalization rates, preservation of collateral flow plays an important role in improving functional outcomes in patients treated with mechanical thrombectomy. Indeed, collateral blood flow could be affected by management of blood pressure (BP) 5,6 , and there is a specific concern about BP drops due to anesthesia during the procedure [7][8][9][10][11][12][13] . The optimal BP during mechanical thrombectomy for acute ischemic stroke is currently unclear, and most of the institutions do not have a standardized protocol for BP management 14 .
The role of conscious sedation versus general anesthesia (GA) in functional outcome is still a matter of debate. Although GA allows a more controlled procedure, BP drops during the administration of anesthetics could lead to failure of the collateral circulation [7][8][9][10][11][12][13] . Similarly, small decreases in BP during conscious sedation also seem to affect clinical outcome after thrombectomy 15 . On the other hand, higher BP levels following mechanical thrombectomy were independently associated with an increased likelihood of mortality and functional dependence in patients with LVO 16 .
The objective of this study was to evaluate BP behavior during mechanical thrombectomy in patients with acute ischemic stroke and its relationship with drugs used for sedation or GA during the procedure.

Patients and data collection
We evaluated consecutive patients treated with mechanical thrombectomy for acute ischemic stroke in a tertiary hospital certified by the Joint Commission International as a Primary Stroke Centre in São Paulo, Brazil, from December 2009 to December 2015. Large vessel occlusion was diagnosed on pretreatment CT angiography (CTA) using standard criteria. Data collected included demographics, stroke risk factors, National Institutes of Health Stroke Scale (NIHSS) score at admission, neuroimaging characteristics (admission ASPECTS), and modified Rankin Scale score at discharge. The NIHSS scores were documented by certified neurologists.
Risk factors were considered if noted on the patient's chart or if medications for known risk factors were used before hospital admission or at discharge. BP levels were documented at least every 15 minutes during mechanical thrombectomy. We collected maximum, minimum, and mean values of systolic and diastolic BP, and mean arterial pressures during the procedure, the time from symptom onset to groin puncture, and post-intervention recanalization assessment on the final digital subtraction angiography.
The devices used included the Solitaire stent retriever or the Penumbra aspiration system. The choice of procedural anesthesia was decided by a consensus between the neurointerventional radiologist and the neurologist. Drugs used for either conscious sedation or GA were reviewed. Recanalization status was determined with the modified Treatment in Cerebral Infarction (mTICI) score, categorized as 0, I, IIa, IIb, and III. Complete reperfusion at the end of endovascular procedure was defined by TICI scores of IIb or III. This study was approved by the Local Ethics Committee of the Hospital Israelita Albert Einstein and the following procedures were in accordance with institutional guidelines.

Statistical analysis
Continuous variables were presented as mean and standard deviation (± SD) or median and interquartile range (± IQR). Data were analyzed for normal distribution. Statistical comparisons for categorical variables between two groups were performed using the χ 2 test. Continuous variables were compared by unpaired (when normal data between two groups were being compared) and paired t test (when normal data from the same patient in two different moments were being compared) or the Mann-Whitney U test (when nonnormal data from different groups were being compared).
Analyses were performed with a two-sided alpha level of 0.05. The Statistical Package for Social Science version 20.0 for Windows (SPSS Inc., Chicago, IL) was used for statistical analyses.
Stent retrievers without distal aspiration (most common technique) were used in 34 patients (64.1%). A Direct Aspiration First Pass Technique (ADAPT) was used in 12 cases (22.6%) while a combination of distal aspiration and stent retriever was used in seven patients (13.2%). The median time from door to arterial puncture was 120 minutes [IQR 97. 2-166.7]. Complete reperfusion was achieved in 34 patients (64.1%). The median NIHSS at hospital discharge was 5 [2,11] (median NIHSS change from hospital admission -8 [-4, -11], P<0.01). GA was used in 47 patients (88.7%) and six (11.3%) patients were managed with conscious sedation. The frequency of the drugs used during the procedure is described in Table 2.
BP behavior during mechanical thrombectomy is depicted in Table 3 and Figure 1. The mean reduction in systolic BP and mean arterial pressure from hospital admission to mechanical thrombectomy (minimal systolic BP during procedure) were 42.0 ± 36.0 and 36.0 ± 26.1 mmHg, respectively. During mechanical thrombectomy, oscillations in BP (maximal minus minimal BP, during the procedure) were 50.4 mmHg (± 26.8) for systolic BP, and 33.2 mmHg (± 19.3) for diastolic BP. Patients treated under GA had a trend to have more oscillation in BP during the procedure when compared to patients treated under conscious sedation (34.8 ± 19.2 mmHg in GA versus 21.1 ± 16.9 mmHg in conscious sedation, P=0.11).
There were no differences is BP variation when using different drugs for sedation (propofol, midazolam, etomidate, fentanyl and remifentanil). Patients treated with neuromuscular blocking drugs had a trend towards more oscillation in systolic BP from hospital admission to mechanical thrombectomy (51.1 ± 38.8 versus 26.2 ± 35.3 mmHg, P=0.06). Considering the use of specific neuromuscular blocking agents, cisatracurium (43.9 ± 17.5 versus 29.6 ± 18.8 mmHg, P=0.02) and succinylcholine (44.7 ± 18.6 versus 29.3 ± 18.2 mmHg, P=0.01) were associated with a significant drop in BP during the procedure. BP oscillation during mechanical thrombectomy was not associated with recanalization status nor with functional outcome at discharge.

DiScUSSiON
In this study, baseline characteristics of patients treated with mechanical thrombectomy for acute ischemic stroke matched similar findings previously described in the setting of this tertiary hospital service and nationwide stroke population 17 . There was only a consideration for higher NIHSS scores in this study compared to a previous publication (17 versus 3) and for frequency of atrial fibrillation (30 versus 15.5%) 17 . Indeed, this would be expected in a large vessel occlusion population.
There is a special concern for potential features that could worsen tissue perfusion during mechanical thrombectomy. Decreases in cerebral perfusion pressure could be directly related to drops in mean arterial BP [18][19][20][21] . This study suggests that significant BP oscillation occurs during mechanical thrombectomy. The reduction of the mean systolic BP during the procedure was more than 40 mmHg, which might worsen tissue perfusion, especially in patients in whom recanalization was not achieved [18][19][20][21] . On the other hand, high maximum systolic BP following MT was previously addressed as an independent risk factor for 3-month mortality and functional dependence in LVO patients 16 . Thereby, this could mean that there is a U-shaped association of BP and functional stroke outcomes 22 .
There were few patients who were treated with mechanical thrombectomy under conscious sedation (six patients), and comparison with general anesthesia was limited. However, significant changes in BP were seen in both groups, raising the question whether conscious sedation is protective or not [7][8][9] . Furthermore, drugs used for conscious sedation were similar to those used in GA, except for the doses applied.
Another important difference between GA and conscious sedation is the use of neuromuscular blockers in the former 23,24 . Cisatracurium is an intermediate long-acting nondepolarizing neuromuscular blocking agent while succinylcholine is a depolarizing neuromuscular-blocking drug 25 . The hemodynamic effects of neuromuscular blockers are well described in the literature, and such effects may be attributed to histamine release and/or acetylcholine-like effects 26,27 . Interestingly we found a significant drop in BP in patients treated with cisatracurium and succinylcholine.
This study had some limitations. Firstly, this was an observational single-center study with relatively small sample size and non-matched control group, which limited analysis to multivariable adjustments. Furthermore, past medical history might have influenced the choice of the drugs used for GA or sedation, which could be a bias in our analysis. Data about the use of vasopressors during the procedure, the doses used of each drug, the doses of specific neuromuscular agents per kilogram used for each patient, as well as the data about blood pressure oscillation in patients that used a combination of drugs were not collected. Finally, there is no data on the influence of BP drop upon long-term functional outcome. Nonetheless, this study is the first to our knowledge to raise the question of the role of specific drugs used for GA or sedation upon BP level in patients treated with mechanical thrombectomy. Further studies are required to better understand the role of sedatives, anesthetics, and neuromuscular blocking agents on functional outcome of patients submitted to mechanical thrombectomy in a setting of acute ischemic stroke.