Predictors of 30-day mortality after endovascular thrombectomy for large vessel occlusion in the elderly

Abstract

Objectives

Elderly acute ischemic stroke (AIS) patients (≥80 years) would have dismal clinical outcomes even after successful endovascular revascularization for large vessel occlusion (LVO) in the anterior circulation. We aimed to identify predictors of 30-day mortality after endovascular thrombectomy (EVT) in the elderly.

Materials and Methods

We included older patients who underwent EVT for AIS due to LVO within 6 h after stroke onset in the anterior circulation between 2017 and 2019. Patients due to posterior circulation stroke, with intracerebral hemorrhage (ICH) or pre-stroke modified Rankin Scale (mRS) score of 4 and 5 were excluded. The primary outcome was mortality within 30 days of EVT. The association between clinical, imaging, procedural, follow-up imaging and mortality were analyzed. Successful reperfusion was defined as modified Thrombolysis in Cerebral Infarction (mTICI) score of 2b or 3. Possible predictors of 30-day mortality were assessed by univariate and multivariable logistic regression.

Results

Total 238 AIS patients eligible for EVT were identified with 58 patients aged 80 years or more. 48 patients met inclusion criteria. Median age was 86 years (age range, 82–102 years). Successful reperfusion was achieved in 38 (79.2%) patients. The 30-day and 90-day mortality rate were 25% and 33.3%, respectively. The independent predictors of 30-day mortality were collateral scores <3 on mCTA (adjusted OR, 16.571; 95% CI, 1.041–263.868; p = 0.047) and number of passes (adjusted OR, 2.475; 95% CI, 1.047–5.847; p = 0.039).

Conclusions

Lower collateral scores on mCTA and higher number of passes in thrombectomy were independently predictive of 30-day mortality in the elderly.

Endovascular thrombectomy (EVT) has been the standard of care in patients with acute ischemic stroke (AIS) due to large vessel occlusion (LVO) within 6 h from symptom onset and up to 24 h in selected patients.^1–3 Because only a small proportion of older patients (≧80 years) were included in the initial five randomized trials, a number of studies have started to look more closely at the effectiveness and safety of EVT in patients over the age of 80 years. Although most studies have revealed that patients of an older age may not have as favorable of outcomes as younger patients, one in every four older patients can benefit from the therapy, and the prognosis post EVT has been shown to be superior to that of usual care.^4–6 Therefore, age should not be a discriminating factor while considering endovascular therapy for patients with AIS.^4,5,7

The favorable outcomes reported in studies that scored patients by using modified Rankin Scale (mRS) at 90 days varied from 19.2% to 51.6% (mRS scores of 0–2), and mortality rates at 90 days were reported to be between 16.1% to 51%.^5–8 Prognostic factors, such as severity of stroke before endovascular treatment and degree of revascularization after thrombectomy, have been proposed, but investigation of these factors has produced inconsistent results.^7,9–11 Differing outcomes and prognostic analyses may be explained by the limited number of cases or by the retrospective cohort design that is used in most studies. These problems reflect a need to refine selection criteria or make case-based considerations for EVT for older patients.

Most research on EVT for older patients has focused on identifying patients who may benefit from the endovascular procedure. Studies dedicated to evaluating predictors of mortality in older patients are somewhat limited, especially short-term or 30-day mortality, which has been considered closely related to the acute disease, instead of other factors such as comorbidity or age. ¹² Better understanding the predictors of 30-day mortality after EVT may reduce the likelihood of futile recanalization or recommendations for potentially detrimental procedures. In this retrospective cohort study, we sought to identify predictors of 30-day mortality post EVT in older patients.

Methods

This study was reviewed and approved by the institutional review boards.

Data collection and clinical variables

We retrospectively collected data for patients who received EVT for AIS due to LVO at our institute from January 2017 to December 2019. Included patients were older than 80 years and had ischemic stroke due to anterior circulation occlusion, including the internal carotid artery (ICA), M1 and M2 segment of the middle cerebral artery (MCA) and tandem cervical ICA/MCA. Patients eligible for EVT met the following criteria: an initial National Institutes of Health Stroke Scale (NIHSS) between 8 to 30, an interval of less than 6 h between stroke onset and groin puncture, and a prestroke mRS score between 0 and 3. Patients who had intracranial hemorrhage or posterior circulation strokes on pre-treatment CT were excluded. Clinical variables of patients collected for analysis included the demographic characteristics; the presence of comorbidities such as hypertension, atrial fibrillation, diabetes mellitus and hyperlipidemia; initial scores on the NIHSS; and administration of an intravenous tissue plasminogen activator (IV t-PA).

Imaging protocol and parameters

All AIS patients suspected of anterior circulation LVO underwent plain computed tomography (CT) of brain, followed by multiphase computed tomographic angiography (mCTA), which is a time-resolved cerebral angiogram of the brain vasculature from skull base to vertex in three phases. ¹³ The first phase was a scan from the aortic arch to vertex, which was triggered at the peak of arterial phase of the brain through bolus tracking. The first phase was completed in less than 7 s. The second phase was completed after a 4-s delay during which the table was repositioned from vertex to skull base. The duration of the scan from the skull base to the vertex was 3.4 s for each additional phase. A total of 80 mL of iodinated contrast medium was injected at a rate of 5 mL/s, and this was followed by a 50-mL normal saline flush at 6 mL/s. Thick-section axial maximum intensity projections were reconstructed at 24-mm slice thickness and 4-mm intervals. The collateral score on the mCTA was recorded using a 6-point ordinal pial arterial filling scale, as previously proposed. ¹³ Images were acquired on a 64-slice CT scanner (SOMATOM definition AS; Siemens, Erlangen, Germany). Scanning parameters were the following: collimation = 64 × 0.6 mm, 120 kVp, 200 mAs, pitch = 1.5, and matrix size = 512 × 512. For statistical analysis, patients were dichotomized into two groups: poor collateral status (a score of 0–2) and good collateral status (a score of 3–5).

Imaging and procedure variables

The thrombus occlusion site, Alberta Stroke Program Early CT (ASPECT) scores, and mCTA collateral scores were prospectively obtained. In this study, ASPECT scores, and mCTA collateral scores were determined by the consensus of two experienced interventional neuroradiologists (Y. -J. Lai & W. -J. Lai). If there was no consensus, imaging grading was then evaluated by a third referee (S. -H. Peng). EVT procedure variables were collected, including the method of anesthesia administration, devices used in the procedure, number of aspiration or stent retriever passes recorded, use of balloon angioplasty with or without intracranial stent placement, final reperfusion modified Thrombolysis in Cerebral Infarction (mTICI) scores, and intraprocedural complications. The first follow-up CT was obtained within 24 h of EVT or at clinical deterioration.

Endovascular techniques

Most EVT treatments were performed under conscious sedation at our institute during the anterior circulation stroke to reduce the door-to-puncture time. Patients who were unable to cooperate during the procedure or who had a reduced consciousness level with a Glasgow Coma Scale score of less than 8 received rapid sequence induction and intubation before EVT to provide airway protection. The technique for thrombectomy was selected according to the vascular anatomy and operator's preference. The procedure involved either aspiration alone with a 6Fr aspiration catheter, or followed by the use of stent retriever, or a combination aspiration and stent retrieval technique through a Neuron MAX 088 (Penumbra, Oakland, California, USA) guiding sheath or a 9F Cello balloon guiding catheter (Medtronic Neurovascular, Irvine, California, USA) in a triaxial assembly. Successful recanalization was defined as achieving an mTICI score of 2B or 3.

Clinical outcomes

Mortality within 30 days was used as the primary outcome, and patients were dichotomized into 30-day mortality group and 30-day survival groups. A comparison of the groups’ baseline characteristics, imaging variables, and procedural results were performed. Other follow-up data included any brain hemorrhage on plain CT images, causes of mortality within 30 days, and mortality at 3-month follow-up. The mRS at 3-month was also recorded. Symptomatic intracerebral hemorrhage (sICH) was defined as a parenchyma hematoma exceeding 30% of the infarct volume with neurological deterioration (NIHSS score ≥ 4) within 36 h of endovascular treatment.

Statistical analyses

The demographic, clinical, angiographic, postprocedural, pathological variables of the 30-day survival and the 30-day mortality groups were compared. Continuous variables were reported as medians (interquartile range, IQR), and categorical variables were presented as percentage. Categorical difference were assessed through a chi-square test, and continuous differences were evaluated through Mann-Whitney U test. A binary logistic regression analysis was then performed to identify factors that could be considered independent predictors for 30-day mortality. A probability value of <0.05 was considered significant. Statistical analysis was performed using SPSS software (version 23.0; IBM Corp., Armonk, NY, USA).

Results

In total, 238 patients received EVT within the period of January 2017 to December 2019. Of the 238, 58 (24.4%) were aged 80 years or older, and 10 were excluded due to vascular occlusion in the posterior circulation. A total of 48 patients (31 women and 17 men; median age, 86 years; age range, 80–102 years) met the inclusion criteria for this study. The primary endpoint, namely 30-day mortality, occurred in 12 (25%) patients.

Baseline characteristics, imaging variables and procedural results for the overall study cohort are summarized in Table 1. The median of initial NIHSS and ASPECTS scores were 20 [interquartile range (IQR): 15–25] and 9 (IQR: 8–10), respectively. IV-tPA was administered in 22 (45.8%) patients. Large-vessel occlusions were discovered in the ICA in 3 (6.3%) patients and in the MCA in 31 (64.6%) patients. ICA with accompanying MCA occlusions were identified in 14 (29.2%) patients. The median of collateral score on mCTA was 4 (IQR: 3–4). Local anesthesia with conscious sedation was administered in 43 (89.6%) patients. Successful reperfusion was achieved in 38 (79.2%) patients, and the median number of aspiration or stent retriever passes was 1 (IQR: 1–2). Additional balloon angioplasty or/and stent placement were used in 10 (20.8%) patients during the EVT procedure.

Table 1.

Comparison of baseline characteristics, imaging variables, and procedural results with mortality in 30 days.

	All (n = 48)	30-day mortality (n = 12)	30-day survival (n = 36)	P value
Age, median (IQR)	86 (81–89)	85 (81–92)	86 (81–88)	0.526
Female, n (%)	31 (64.6)	7 (58.3)	24 (66.7)	0.601
Atrial fibrillation, n (%)	33 (68.8)	7 (58.3)	26 (72.2)	0.369
Hypertension, n (%)	36 (75.0)	11 (91.7)	25 (69.4)	0.124
Diabetes mellitus, n (%)	13 (27.1)	3 (25.0)	10 (27.8)	0.851
Hyperlipidaemia, n (%)	10 (20.8)	1 (8.3)	9 (25.0)	0.218
IV-tPA, n (%)	22 (45.8)	5 (41.7)	17 (47.2)	0.738
Initial NIHSS, median (IQR)	20 (15–25)	21 (16–27)	20 (15–25)	0.384
Obstruction side				0.187
Left, n (%)	25 (52.1)	4 (33.3)	21 (58.3)
Right, n (%)	23 (47.9)	8 (66.7)	15 (41.7)
Location of arterial occlusion				0.183
ICA, n (%)	3 (6.3)	2 (16.7)	1 (2.8)
MCA, n (%)	31 (64.6)	6 (50.0)	25 (69.4)
ICA + MCA, n (%)	14 (29.2)	4 (33.3)	10 (27.8)
ASPECT	All (n = 46)	(n = 11)	(n = 35)
ASPECT, median (IQR)	9 (8–10)	8 (6–10)	9 (8–10)	0.201
mCTA collateral score	All (n = 44)	(n = 10)	(n = 34)
Collateral <3, ≥3				0.004
Collateral <3, n (%)	7 (15.9)	5 (50.0)	2 (5.9)
Collateral ≥3, n (%)	37 (84.1)	5 (50.0)	32 (94.1)
Technical details	All (n = 48)	(n = 12)	(n = 36)	P value
Anaesthesia				<0.001
LA, n (%)	43 (89.6)	7 (58.3)	36 (100.0)
GA, n (%)	5 (10.4)	5 (41.7)	0 (0)
Extra-device, PTA/stent, n (%)	10 (20.8)	2 (16.7)	8 (22.2)	1.000
Number of passes	All (n = 31)	(n = 6)	(n = 25)
Number of passes – all, median, (IQR)	1 (1–2)	4 (1–5)	1 (1–2)	0.046
mTICI	All (n = 48)	(n = 12)	(n = 36)
mTICI ≥2b, <2b				0.101
mTICI, 2b, 3, n (%)	38 (79.2)	7 (58.3)	31 (86.1)
mTICI, 0, 1, 2a, n (%)	10 (20.8)	5 (41.7)	5 (13.9)
Haemorrhage	All (n = 48)	(n = 12)	(n = 36)
ICH, n (%)	8 (16.7)	6 (50)	2 (5.6)	0.002
sICH, n (%)	4 (8.3)	4 (33.3)	0 (0)	0.003

IV-tPA: intravenous tissue-type plasminogen activator; NIHSS: National Institutes of Health Stroke Scale; ICA: internal carotid artery; MCA: middle cerebral artery; ASPECT: Alberta Stroke Program Early CT score; mCTA: multiphase CT angiography; LA: local anaesthesia; GA: general anaesthesia; PTA: percutaneous transluminal angioplasty; mTICI: modified Thrombolysis in Cerebral Infarction; ICH: intracranial haemorrhage; sICH: symptomatic intracranial haemorrhage.

The number of aspiration or stent retriever passes to achieve successful recanalization was significantly higher in patients with 30-day mortality (4 [IQR: 1–5] vs. 1 [IQR: 1–2], p = 0.046). Although there was a trend of more patients achieving successful recanalization (mTICI = 2B or 3) in the 30-day survival group, the difference between the 30-day survival group and the 30-day mortality group was not statistically significant (86.1% vs. 58.3%, respectively, p = 0.101). The two groups displayed no differences in age (p = 0.526); gender (p = 0.601); or baseline comorbidities, such as hypertension (p = 0.124), diabetes (p = 0.851), or hyperlipidemia (p = 0.218).

The overall incidence of postprocedural intracranial hemorrhage was 16.7% of patients (8/48), with an incidence of 50% (6/12) in the 30-day mortality group and 5.6% (2/36) in the 30-day survival group. All patients with sICH (8.3%, 4/48) were in the 30-day mortality group. Two of these cases were complicated by procedure-related subarachnoid hemorrhage. In total, 12 patients died within 30 days of the procedure. Of these deaths, four (33.3%) patients were due to sICH, four (33.3%) were due to septic shock, and two (16.6%) were due to a large brain infarction, and two (16.6%) were due to cardiopulmonary failure. The overall mortality at 3-month follow-up was 33%. The 3-month mRS score ≦2 was achieved in 12.5% and ≦3 was observed in 22.9%.

As displayed in Table 2, among the patients, having a unfavorable collateral score (OR, 16.000; 95% CI, 2.414–106.057; p = 0.004) of 30-day mortality, an increased number of aspiration or stent retriever passes (OR, 3.122; 95% CI, 1.306–7.460; p = 0.01) and any intracerebral hemorrhage (ICH) were associated with increased odds (OR, 17.000; 95% CI, 2.753–104.973; p = 0.002) of 30-day mortality in a univariate logistic regression analysis. In a multivariate logistic regression model, collateral scores on an mCTA (adjusted OR, 16.571; 95% CI, 1.041–263.868; p = 0.047) and number of passes (adjusted OR, 2.475; 95% CI, 1.047–5.847; p = 0.039) emerged as independent predictors of mortality within 30 days.

Table 2.

Univariate and multivariate binary logistic regression analysis of predictors of mortality.

	Unadjusted OR	95% CI	P value	Adjusted OR	95% CI	P value
mCTA collateral score <3	16.000	2.414–106.057	0.004	16.571	1.041–263.868	0.047
Number of passes, per 1-pass increase	3.122	1.306–7.460	0.010	2.475	1.047–5.847	0.039
ICH	17.000	2.753–104.973	0.002	–	–	–

OR: odds ratio; CI: confidence interval; mCTA: multiphase CT angiography.

Discussion

An increase in successful recanalization rates in EVT has been reported for AIS with LVO, but older patients have continued to have worse functional outcomes and higher mortality than younger patients.^6,7,14–17 Most studies have attempted to establish a correlation between various factors and functional outcomes including mortality within 90 days.^5–10 However, Garland et al. reported that short-term mortality was predominantly determined by the acute illness, not comorbidity or age. ¹² The factors that influence early mortality (within 30 days) of EVT in older patients could more closely relate to the characteristics of the disease. The present study indicated that a lower mCTA collateral score (<3, score: 0–5) and a higher number of aspiration or stent retriever passes were independently associated with 30-day mortality after EVT in patients aged ≥80 years with AIS in the anterior circulation LVO.

Approximately one-quarter of patients were older than 90 years, with median age of 86 years, in this study cohort. Regarding to successful recanalization (TICI ≥ 2b) and overall mortality rate in 3 months, the results of our study are comparable to those of another systematic review and meta-analysis study of EVT in older patients, with 3-month mortality of 33.3% versus 34%, and successful recanalization of 79.2% versus 78%, respectively. ⁵ Some studies argue that there was no obvious difference in favorable clinical outcomes and mortality after EVT in older patients comparing to younger patients.^10,18 The results of our study are consistent with those of most recent studies,^6,7,11,19 demonstrating that even with a high rate of successful recanalization, older patients tend to have a higher mortality rate after EVT than younger patients. Old age (≥80 years) was independently associated with an increased mortality rate, ⁷ which may indicate that patient selection for EVT in older patients could be optimized through better identification of patients unlikely to benefit from the treatment.

A similar idea was proposed by Ahn et al., ²⁰ who documented how lower baseline diffusion-weighted imaging ASPECTS scores and poor reperfusion status were independently associated with catastrophic outcome (mRS 4–6) at 90 days after EVT in older patients. The long-term (90 days) outcome in the study might be affected by post-stroke medical comorbidities, especially for older patients, who are more likely to have a variety of post-stroke comorbidities that could affect outcomes. In our study, we focused on 30-day mortality after EVT, which might reduce the impact of post-stroke comorbidities. Moreover, CT-based imaging rather than magnetic resonance imaging has been used as a first-line imaging modality to guide patient selection in both intravenous thrombolytic therapy and EVT. In a multivariable model, this study revealed that patients whose mCTA collateral scores were less than 3 had a likelihood of 30-day mortality 16.57 times higher than patients whose mCTA collateral scores were ≥3. Furthermore, each additional 1 pass in EVT therapy was associated with 14.8% higher odds of 30-day mortality.

Similar to computed tomographic perfusion (CTP), mCTA has been considered a reliable tool for imaging selection and predicting regional tissue fates in AIS patients.^13,21 In addition, the use of mCTA collateral assessment may have advantages over CTP due to the fact that it is less likely to be influenced by patient motion, lesser radiation dosage, and fewer time for image postprocessing. ¹³ Collateral circulation provides ischemic brain tissue with vital blood supply after vessel occlusion. The results of our study accord with a meta-analysis study by Qian et al. ²² who demonstrated positive pretreatment collateral status to be associated with favorable functional outcomes and decreased rates of sICH and mortality of EVT. Ischemic brain tissue is only able to sustain a sufficient blood supply through satisfactory collateral circulation; otherwise, an irreversible infarction core soon develops. Due to age-related neuronal loss and the subsequent reduction in their neurological reserve, poor collateral circulation status in older patients may lead to immediate poor outcomes regardless of recanalization after EVT.

Although earlier reports have revealed that initial NIHSS and ASPECT scores are likely predictors of functional outcomes and mortality,^9,15 no significant difference was found in baseline comorbidities, initial NIHSS, and ASPECT scores between the two groups in this study. If these factors have less impact on 30-day mortality needs further investigation due to the retrospective review and a small number of patients in the cohort. Any postprocedural ICH associated with 30-day mortality may reflect that older patients are vulnerable to the sequelae of stroke and procedure-related complications. The postprocedural sICH rate as 8.3% in this study, which is comparable to the results of previous reports, ranging from 3% to 18%.^{7,9,11,14,15,17,20,23} All patients with sICH were in the 30-day mortality group. There was a trend to early mortality in older patients with sICH after EVT, the difference was not significant in this relatively small retrospect study.

We found that none of the patients with sICH received surgical intervention to relieve increased intracranial pressure caused by hematoma in this study, and all the patients succumbed within 10 days of the procedure. This could be a major contributing factor to 30-day mortality in older patients with post procedure sICH. For older patients in this critical condition, their family members typically expressed a preference for medical treatment instead of an invasive, life-prolonging intervention or surgery. Therefore, careful patient selection to avoid in-hospital mortality should be mandatory and realistic when treating acute stroke in older patients.

Limitations

Our study has several drawbacks. First, this was a retrospective study with a single-center setting. In addition, a relatively small number of patients were included and the statistical power may be weak. Our study revealed some possible associations between mortality and prognostic factors. Further study is required to provide more robust evidence.

Conclusion

Older patients who had lower collateral scores on mCTA at presentation exhibited a higher 30-day mortality rate after EVT. During the endovascular procedure, more passes of devices to achieve successful reperfusion increase the risk of 30-day mortality. Our results may guide clinical patient selection to avoid potentially devastating outcomes when considering older patients for EVT.