Background Prompt revascularization of the ischemic penumbra following an acute ischemic event (AIS) has established benefit within the literature. However, use of the semi-quantitative Alberta Stroke Program Early CT Score (ASPECTS) to evaluate patient suitability for revascularization has been inconsistent in patient risk stratification and selection.
Objective To conduct a meta-analysis to evaluate the available evidence for a clinically valid ASPECTS threshold in assessment of suitability for revascularization following AIS.
Methods Two independent reviewers searched Medline (Ovid) and Cochrane Central Register of Systematic Reviews databases for studies appraising outcomes of endovascular thrombectomy (EVT) in relation to a variably-defined preoperative ASPECTS.
Results A total of 13 articles were included. The pooled good outcome proportion after EVT was 41.4% (95% CI 36.4% to 46.6%; p<0.001), with subjective study-specific definitions of favorable and unfavorable subgroup outcomes of 49.7% (95% CI 44.2% to 55.3%; I2=76.5%; p<0.001) and 33.2% (95% CI 28.5% to 38.3%; I2=33.16%), respectively. Objective trichotomization into low (0–4), intermediate (5–7), and high (8–10) subgroups yielded pooled good outcome proportions of 17.1% (95% CI 6.8% to 36.8%; I2=64.24%; p=0.039), 35.7% (95% CI 30.5% to 41.3%; I2=23.11%; p=0.245), and 49.7% (95% CI 44.2% to 55.3%; I2=76.5%; p<0.001) for low, intermediate, and high ASPECTS, respectively.
Conclusions A subjectively favorable ASPECTS is associated with significantly better outcomes after EVT than an unfavorable ASPECTS, regardless of the cut-off used. EVT is unlikely to be useful in patients with an objectively low ASPECTS and is likely to be useful for those with high ASPECTS; findings in patients with intermediate ASPECTS were equivocal.
Statistics from Altmetric.com
Prompt revascularization of the brain to salvage the ischemic penumbra is of proven benefit for patients with acute ischemic stroke (AIS).1 Endovascular thrombectomy (EVT) has emerged as the standard of care for patients with AIS secondary to emergent large-vessel occlusion (ELVO) within 24 hours of symptom onset.2 3 Several recent large-scale randomized trials have shown that EVT together with intravenous (IV) thrombolysis results in more successful recanalization and better clinical outcomes than thrombolysis alone.4–9 Nevertheless, candidates for EVT must be carefully selected, based on established criteria, to maximize the likelihood of clinical benefit and prevent futile or even harmful reperfusion.10
The Alberta Stroke Program Early CT Score (ASPECTS) is a 10-point semi-quantitative system used to assess early ischemic changes on non-contrast CT after an anterior circulation AIS.11 Although baseline ASPECTS is often obtained before revascularization therapy to identify patients likely to respond poorly, the threshold below which a score is considered unfavorable varies considerably between studies. A cut-off of ASPECTS ≤7 to define a moderate/large core that is associated with an unfavorable response that was proposed by Barber et al in their original publication for use with IV thrombolysis11 was adopted in much of the early literature and subsequent trials of EVT.10 12 13 The SWIFT-PRIME and ESCAPE trials used ASPECTS ≤5 on baseline non-contrast CT as an exclusion criterion. The treatment recommendations of the 2015 American Heart Association/American Stroke Association (Class I; Level of Evidence A) also define patients with unfavorable ASPECTS ≤5.14 Other studies have defined an unfavorable ASPECTS ≤6.8 15
The efficacy of EVT in patients with unfavorable ASPECTS is variable in the literature. Early trials concluded that patients with a large infarct core on MRI diffusion-weighted imaging (DWI) or CT ASPECTS had a significantly worse functional outcome and increased mortality compared with patients with a higher ASPECTS score.10 16 More recently, studies using later-generation endovascular devices, which offer increased speed, better recanalization rates and lower complication rates relative to older technologies,17 18 have suggested that patients with ASPECTS ≤6 may derive similar or equivalent benefit from EVT.15 17 To address this important issue, we performed a comprehensive and updated meta-analysis of EVT data to determine the interaction of baseline ASPECTS and outcome, with a focus on patients with both favorable and unfavorable baseline ASPECTS.
A literature search was conducted through the electronic medical database Medline (Ovid) and the Cochrane Central Register of Systematic Reviews from their commencement until February 2018, in line with recommended meta-analysis guidelines such as PRISMA, to identify eligible articles for this meta-analysis and systematic review. Key MeSH terms and search terms within the search strategy included ‘cerebrovascular accident’, ‘stroke’, ‘ischaemia’, ‘endovascular’, ‘thrombectomy’, and ‘thrombolysis’. The cited references within the relevant articles were also searched for pertinent articles for inclusion in this study.
The inclusion criteria for this study included papers that recruited patients who were eligible for endovascular therapy, described functional outcomes in terms of modified Rankin Scale (mRS) at 90 days, and used contemporary endovascular techniques. Exclusion criteria were duplicate studies, pediatric populations, case reports, abstract papers, and studies with small sample sizes less than 10.
Data extraction and quality assessment
Each retrieved article was critically appraised by two independent reviewers (KP, SS). Clinical data were extracted from text, tables or figures through use of a standardized computerized spreadsheet. Discrepancies or disagreements between the reviewers were resolved by discussion and consensus.
Primary outcomes for this study were performance on mRS, mortality, and spontaneous intracerebral hemorrhage (sICH). These outcomes were stratified by preoperative ASPECTS assessment.
A favorable ASPECTS score has a variable definition in the reviewed literature, ranging from >5 to 7 out of 10. Accordingly, a primary analysis compared outcomes for each of the favorable and unfavorable ASPECTS definitions defined on a per article basis. A secondary analysis was performed with objective subgrouping of reported ASPECTS scores into low (0–4), intermediate (5–7), and high (8–10) subgroups correlated with rates of the same outcome variables.10
A meta-analysis of proportions was performed. Heterogeneity within the sample was anticipated and hence proportions were combined using DerSimonian–Laird random effects models. Subgroup analysis according to ASPECTS score subgroups was performed using meta-regression analysis. Heterogeneity was evaluated using the Cochran Q and I2 test. All analyses were performed using the metaphor package for R version 3.01. P values <0.05 were considered statistically significant.
A total of 667 articles were identified through our electronic database searches. Of these, 13 met the inclusion criteria,5 7 8 15 17 19–26 yielding a total of 2171 patients included in the final analysis. A PRISMA flow diagram describing the process of study selection is presented in figure 1. The methodological and baseline clinical characteristics of the included studies are presented in table 1.
Favorable versus unfavorable ASPECTS
The initial analysis compared outcomes between patients with favorable and unfavorable baseline ASPECTS, individually defined for each study. The results of this analysis are presented in figure 2.
The overall pooled proportion of good outcome (mRS 0–2) after EVT was 41.4% (95% CI 36.4% to 46.6%; p<0.001), although significant heterogeneity was noted in the 13 included studies (I2=78.88%) (figure 2). Subgroup analysis of patients with unfavorable ASPECTS in these studies revealed an estimate of good outcome of 33.2% (95% CI 28.5% to 38.3%; p=0.117), with non-significant heterogeneity (I2=33.16%). Data for the subgroup with favorable ASPECTS were obtained from nine studies and showed that these patients had significantly better functional outcome (49.7%; 95% CI 44.2% to 55.3%; I2=76.5%; p<0.001). The OR of good outcome in patients with an unfavorable ASPECTS was 0.46 (95% CI 0.38 to 0.56; p<0.001). Differences in ASPECTS cut-off, imaging modality, and treatment regimen likely accounted for the heterogeneity observed in this result.
Meta-analysis of the seven studies with available data demonstrated that the overall proportion of sICH was 8.3% (95% CI 5.9% to 11.6%; p=0.04; I2=61.55%) (see online Appendix, supplementary figure 1a). Rates of sICH for unfavorable ASPECTS (10.9%; 95% CI 7.3% to 16.1%; p=0.076; I2=47.56%) were significantly higher than those for favorable ASPECTS (5.9%; 95% CI, 3.8% to 9.1%; p=0.129; I2=47.08%). The OR was found to be 1.75 (95% CI 1.17 to 2.62; p=0.007).
As determined from the eight studies with data available, the overall pooled mortality rate for patients in both subgroups was 19.5% (95% CI 15% to 25%; I2=80.05%; p<0.001) (see online Appendix, supplementary figure 1b). When comparing the two subgroups, mortality after EVT was more likely among patients with unfavorable ASPECTS relative to favorable ASPECTS (25.4%, 95% CI 19% to 33.2%; I2=64.44% vs 13.8%, 95% CI 9.5% to 19.6%; I2=80.63%). The odds of mortality in the unfavorable ASPECTS group was 1.86 (95% CI 1.44 to 2.41; p<0.001). Significant heterogeneity was observed across all three results for this outcome measure.
Low versus intermediate versus high ASPECTS
In the secondary analysis, patients from the 13 studies were trichotomized according to their baseline ASPECTS into low (0–4), intermediate (5–7), and high (8–10), where possible. Subgroups were then compared with respect to the stated outcome measures.
Data on functional outcomes for low ASPECTS patients were available in four studies. The pooled proportion of good outcome (mRS 0–2) was 17.1% (95% CI 6.8% to 36.8%; I2=64.24%; p=0.039) (figure 3). This result was significantly lower than that for intermediate ASPECTS (35.7%, 95% CI 30.5% to 41.3%; I2=23.11%; p=0.245) and high ASPECTS (49.7%, 95% CI 44.2% to 55.3%; I2=76.5%; p<0.001).
Minimal data on the incidence of sICH after EVT were available for stratification by ASPECTS subgroup. sICH rates for high ASPECTS (5.9%, 95% CI 3.8% to 9.1%; I2=47.08%; p=0.129), intermediate ASPECTS (11.9%, 95% CI 5.5% to 23.9%; I2=57.62%; p=0.094) and low ASPECTS (9.1%, 95% CI 1.3% to 43.9%) followed no observable trend (see online Appendix, supplementary figure 2a). For the low ASPECTS subgroup, only data from the post-hoc analysis of the MR CLEAN trial were available, limiting the size of the sample and the power of the analysis.
Mortality after EVT for patients in the low ASPECTS subgroup (42.2%, 95% CI 32.5% to 52.6%; I2=0%; p=0.675) was significantly increased relative to intermediate ASPECTS (24.2%, 95% CI 12.5% to 41.6%; I2=85.79%; p<0.001) and high ASPECTS (13.8%, 95% CI 9.5% to 19.6%; I2=80.63%; p<0.001) (see online Appendix, supplementary figure 2b). The result overall (20.1%, 95% CI 14.3% to 27.5%; I2=86.85%) was significant (p<0.001). Heterogeneity was significant overall, and among the studies analyzed for the upper two subgroups.
Endovascular therapy versus standard care in low ASPECTS cases
We performed a pooled analysis of EVT versus standard care for patients with low ASPECTS scores (0–4). Fifty-one EVT cases were compared with 68 standard care patients. We found no significant difference in good outcome (mRS 0–2) between the two groups after meta-analysis (OR 1.82, 95% CI 0.70 to 4.69, p=0.22) with low heterogeneity (I2=0%).
This meta-analysis shows that a favorable ASPECTS after AIS with ELVO is associated with significantly better functional outcome, lower sICH rates, and lower mortality than an unfavorable ASPECTS. Furthermore, as ASPECTS increases from low to intermediate to high subgroups, functional outcome improves and mortality decreases significantly. A similar pattern is observed for sICH rates, although this trend is non-significant.
Favorable versus unfavorable ASPECTS
The finding in this study that unfavorable ASPECTS is associated with significantly worse outcomes than favorable ASPECTS for all measures, regardless of the cut-off used, is inconsistent with some recent literature regarding EVT. Of the major randomized trials since 2015, all those with data stratified by ASPECTS demonstrated no difference in functional outcome, mortality, or likelihood of parenchymal hematoma between patients above and below the chosen threshold,5–8 a result attributable in part to the increased speed, better recanalization rates, and lower complication rates with second-generation endovascular devices.18 27 However, due to the strict exclusion criteria employed in all but one of these trials, patients with ASPECTS ≤6 were under-represented in the data. Subsequent studies by Kim et al,22 Logan et al15 and Hungerford et al,20 all retrospective reviews of prospectively-maintained databases in which patients with ASPECTS ≤6 were included, also demonstrated similar or equivalent functional outcomes after EVT for patients with unfavorable baseline ASPECTS. Limitations of study design may have affected these results as well. Kim et al excluded patients with DWI-ASPECTS ≤3. In the study by Logan et al, many more patients with an unfavorable ASPECTS (≤6) had ASPECTS 5 or 6 than ASPECTS ≤4. The poor ASPECTS group (≤6) in the analysis by Hungerford et al experienced significantly more loss to follow-up than other groups. These factors are likely to have decreased the chance of detecting a significant difference in outcomes between ASPECTS cohorts. Nevertheless, the results of the present study find support in the literature. Other recent meta-analyses28 29 and several retrospective analyses25 30 confirm that, despite advances in thrombectomy techniques and imaging protocols, patients with lower ASPECTS on pretreatment CT imaging derive significantly less benefit from EVT than those with higher ASPECTS. Consequently, ASPECTS remains a reliable prognostic tool in predicting outcomes for patients with AIS with ELVO who are considered for EVT.
EVT for low ASPECTS (0–4)
The exposure–response relationship for functional outcome and mortality evident on subgroup analysis in this study suggests that patients with baseline low ASPECTS (0–4) are significantly less likely to benefit from EVT. This is consistent with the findings of the HERMES (Highly Effective Reperfusion evaluated in Multiple Endovascular Stroke Trials) collaboration, which pooled data from 1287 patients and concluded that extensive irreversible injury on non-contrast CT of the brain is significantly associated with low rates of good clinical outcome.17 Likewise, Wasser et al demonstrated that patients with an ASPECTS 0–5 obtained from source images on CT angiography had very high rates of poor outcome (86%), in-hospital death (38%), and sICH (7.9%).25 Some of the studies included in this analysis proposed relative benefits of EVT for low ASPECTS patients.19–21 Mourand et al demonstrated good clinical outcomes in patients with DWI-ASPECTS 4–5, and suggested that younger patients (≤70 years) could be considered for EVT even with DWI-ASPECTS 0–3.31 Desilles et al described a positive although non-significant trend towards better functional outcome (23.1% vs 9.5%), lower mortality rate (45.7% vs 57.1%), and lower rate of sICH (23.9% vs 45.5%) in successfully reperfused patients compared with non-reperfused patients from the DWI-ASPECTS 0–4 subgroup.21 Nevertheless, given these patients’ overall poor prognosis, any absolute benefit with current treatment would be likely to be minimal.23 Therefore, ASPECTS ≤4 may offer a viable exclusion criterion in the selection of patients for EVT to prevent futile or even harmful recanalization.
EVT for intermediate ASPECTS (5–7)
There may be a subset of patients with intermediate ASPECTS for whom EVT offers significant clinical benefit after AIS due to ELVO. Results from the subgroup analysis in this study, as well as the published results of studies included in our analysis19 20 22 26 and several other meta-analyses,17 28 29 indicate that patients with ASPECTS 5–7 may achieve similar or equivalent functional outcomes to those with higher ASPECTS, with similar rates of morbidity and mortality. Naturally, not all patients of this ASPECTS subgroup will respond well to EVT.24 Rather, an intermediate ASPECTS score can be considered alongside other criteria such as age group,32 National Institutes of Health Stroke Scale (NIHSS) score, and extent of collateral circulation on CT angiography or CT perfusion.3 Bhole et al33 showed that, among 62 patients undergoing EVT who did not meet top-tier evidence criteria and had baseline ASPECTS <6, 33% achieved mRS ≤2 by 3 months. Similar findings were published more recently by Goyal et al.27 Our results, alongside the findings of these studies, suggest that patients with intermediate ASPECTS represent a distinct clinical subpopulation and subsequently have distinct outcomes when compared with the low ASPECTS subgroup of which they have traditionally been considered a part. This distinction warrants further consideration and study of this intermediate group, which has previously been contraindicated for EVT, and by adopting a more case-by-case selection process for patients with intermediate ASPECTS, institutions can ensure that patients who could potentially benefit from EVT are not excluded from treatment.
Study strengths and limitations
This meta-analysis pools a large amount of data from available trials and observational studies on EVT to achieve a substantial sample size. The inclusion of studies from 2015 onwards is likely to better represent recent endovascular technology and management, although the possibility remains that these studies may have included some cases done with older generation technology. Our study has several limitations. First, the amount of data available for each subgroup after ASPECTS trichotomization varied greatly, with notable heterogeneity. In particular, outcomes with low baseline ASPECTS may have been underpowered to accurately determine clinical benefit in our analysis. Further controlled trials on EVT in patients with large infarct cores using second- and third-generation thrombectomy devices are required. In addition, there were inconsistencies in the imaging modalities used: three of the included studies reported DWI-ASPECTS19 21 22 and two assessed baseline infarct extent on contrast-enhanced CT.15 25 Comparisons between imaging modalities can be misleading, especially given the discrepancy rate of up to 20% between CT ASPECTS and MRI DWI-ASPECTS.34 Once the literature on the association between low or intermediate ASPECTS and outcomes of EVT is more robust, further analyses specific to each modality are warranted. In addition, there is a major limitation in dichotomization given the reported inter-observer variability in assigning ASPECTS, with disagreement in up to 1–4 cases (κ=0.5).35 These limitations of current data suggest an equipoise for intermediate ASPECTS as contraindication for EVT, and indicate a significant potential opportunity for a randomized controlled trial to clarify patient outcomes in this subgroup.
This meta-analysis of recent clinical trials and retrospective studies demonstrates that a favorable ASPECTS obtained after AIS with ELVO is associated with significantly better outcomes after EVT than an unfavorable ASPECTS, regardless of the cut-off used. Since functional outcome and mortality follow an exposure–response pattern relative to ASPECTS, EVT is unlikely to be useful for low (0–4) ASPECTS patients and is useful for patients with high (8–10) ASPECTS. There may be a subset of patients with intermediate (5–7) ASPECTS for whom EVT offers clinical benefit.
Contributors Study inception: KP, HA. Data extraction: KP, SS, AAD, JM. Data analysis: KP, SS, AAD. Interpretation: all authors. Initial draft of manuscript: KP, SS, AAD, HA. Review and editing: all authors. Supervision: DMB, RC, HA.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests None declared.
Patient consent Not required.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All data available in manuscript.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.