Abstract
Background
Prior studies have shown benefit for endovascular therapy (EVT) in patients with large‐vessel occlusion and severe deficits, as captured by the National Institutes of Health Stroke Scale (NIHSS). However, the benefit of EVT in patients with NIHSS score of <6 is unclear.
Methods
We queried the National Inpatient Sample (2018–2020) for patients with a large‐vessel occlusion of the internal carotid or middle cerebral artery with an NIHSS score of <6, and compared outcomes between patients treated with EVT versus best medical management, using propensity score matching. The primary outcome was routine discharge (home or self‐care). Secondary outcomes were in‐hospital mortality, intracerebral hemorrhage, and length of stay. Primary and secondary outcomes were evaluated using multivariable regression adjusted for baseline characteristics, stroke severity, and treatment with thrombolysis.
Results
Of the 212 515 patients with an internal carotid artery/middle cerebral artery stroke, 49 115 met the inclusion criteria for our study. There were 3490 (7.1%) patients treated with EVT and 45 625 (92.9%) treated with best medical management. Patients treated with EVT had no difference in ofdds of routine discharge (adjusted odds ratio (OR), 0.99; 95% CI, 0.82–1.19; P = 0.91), but there were longer lengths of hospital stay (adjusted β, 1.26; 95% CI, 0.78–1.74; P<0.001) and higher rates of death (adjusted OR, 2.11; 95% CI, 1.20–3.70; P = 0.01), compared with patients treated with best medical management. In multivariable models, intravenous thrombolysis was independently associated with higher odds of routine discharge (adjusted OR, 1.78; 95% CI, 1.57–2.01; P<0.001) and shorter hospital stay (adjusted β, −0.41; 95% CI, −0.63 to −0.19; P<0.001), but no difference in in‐hospital mortality (adjusted OR, 0.70; 95% CI, 0.39–1.24; P = 0.22). These relationships persisted in the propensity‐matched cohort.
Conclusion
Patients treated with EVT compared with best medical management had worse clinical outcomes, and this may be attributable to residual confounding (eg, deterioration following an initially low NIHSS score) or procedural factors. Patients treated with intravenous thrombolysis had greater odds of routine discharge. Our findings suggest potential real‐world benefits for intravenous thrombolysis in patients with occlusion of the internal carotid artery/middle cerebral artery and low NIHSS score.
Keywords: National Institutes of Health Stroke Scale, stroke, thrombectomy
Nonstandard Abbreviations and Acronyms
- BMM
best medical management
- EVT
endovascular therapy
- HCUP
Healthcare Cost and Utilization Project
- ICA
internal carotid artery
- ICH
intracerebral hemorrhage
- IVT
intravenous thrombolysis
- LVO
large vessel occlusion
- LOS
length of stay
- MCA
middle cerebral artery
- mCCI
modified Charlson Comorbidity Index
- NIHSS
National Institutes of Health Stroke Scale
- NIS
National Inpatient Sample
- PS
propensity score
- SMD
standardized mean differences
Clinical Perspectives
Endovascular thrombectomy for patients with anterior large‐vessel occlusion and National Institutes of Health Stroke Scale score of <6 was not associated with greater odds of routine discharge or shorter hospital stays.
These findings reflect nonrandomized treatment allocation and require validation with randomized trials.
Intravenous thrombolysis for patients with low National Institutes of Health Stroke Scale score was associated with increased odds of routine discharge and shorter hospital stays.
Anterior circulation strokes account for nearly 70% of large‐vessel occlusions (LVOs). Previous clinical trials have shown the benefit of endovascular therapy (EVT) in patients with disabling anterior circulation strokes compared with best medical management (BMM); however, most patients treated in these trials presented with severe deficits and a National Institutes of Health Stroke Scale (NIHSS) score of >6. 1 , 2 , 3 , 4 , 5 , 6 , 7 The MR CLEAN‐LATE (Endovascular Treatment of Acute Ischemic Stroke in the Netherlands for Late Arrivals) trial reported significant benefit with EVT in patients with occlusion of the internal carotid or proximal middle cerebral arteries (M1 or M2) and NIHSS score of >2. In the subgroup with NIHSS score of ≤6, the benefit of EVT over BMM was preserved, although the CI crossed 1 (adjusted common odds ratio [OR], 1.79; 95% CI, 0.95–3.37; P interaction = 0.08). As many as 1 in 8 patients with a proximal intracranial occlusion will have an NIHSS score of <6. 8 Nearly 1 in 3 patients with low NIHSS score treated without EVT may remain dependent on others after 3 months 9 , 10 ; therefore, endovascular reperfusion may be of benefit for these patients. 11 In this analysis of the US National Inpatient Sample (NIS), we used a large population‐based data set to analyze short‐term outcomes, length of hospital stay (LOS), and early mortality rates between patients with anterior proximal LVO treated with EVT versus BMM with NIHSS score of <6.
Methods
Data Source
All data used in this study were from the NIS. The data are publicly available and can be found at www.hcup‐us.ahrq.gov. Data from the NIS contain ≈20% of all US hospitals (totaling to >7 million patients per annum). Using complex sampling weights provided by the Healthcare Cost and Utilization Project (HCUP), data from the NIS can be extrapolated to the entire US population. Inferences drawn from these analyses are representative of the US population. Because this study uses a publicly available data set containing only deidentified patients, it was exempt from Cooper University Health Institutional Board Review.
Patient Population
Patients with a primary diagnosis of internal carotid artery (ICA) or middle cerebral artery (MCA) occlusion were identified from the NIS based on the International Classification of Diseases, Tenth Revision (ICD‐10), codes (I63.331, I63.41, I63.03, and I63.13). The International Classification of Diseases, Tenth Revision, Clinical Modification (ICD‐10‐CM), Procedure Coding System 03CG3ZZ and 03CG3Z7 codes were used to identify patients who underwent EVT, and represented “extirpation of matter from an intracranial artery,” as had been used in prior analyses of the NIS. 12 The ICD‐10 procedure code 3E03317 and diagnosis code Z92.82 were used to identify patients who received intravenous thrombolytic treatment. Patients with an intracranial hemorrhage were identified using the ICD‐10‐CM codes I69.0 and I69.1. The inclusion criteria for our study were patients who were admitted between the years of 2018 and 2020 who had a primary diagnosis of stroke attributable to ICA or MCA occlusion with an NIHSS score of <6. Only adult patients (defined as aged ≥21 years) with complete covariate information (below) were included in the analysis. Comorbidities assessed and required by the analysis were hypertension, diabetes, dyslipidemia, atrial fibrillation, overweight, coronary disease, tobacco use, and a modified Charlson Comorbidity Index (mCCI), as has been previously described. 13 Although history of stroke or transient ischemic attack is represented in the NIS, these variables were excluded from the covariates for analysis, and excluded from the mCCI to reduce confounding given that such diagnoses may overlap with the primary diagnosis of ICA or MCA occlusion. 13
Statistical Analysis
To reduce confounding by indication for EVT, propensity score (PS) matching was performed using the nearest‐neighbor method, without replacement, with a caliper of 0.1 to create a 1:1 matched cohort of EVT versus BMM. For each patient, a PS was calculated using a binary logistic regression model to estimate the individual propensity for EVT, accounting for age, sex, race, median household income quartile of zip code, hospital bed size, hospital location/teaching status, hospital region, hypertension, diabetes, dyslipidemia, atrial fibrillation, overweight, coronary disease, tobacco use, mCCI, and baseline NIHSS score. Standardized mean differences (SMDs) were used to evaluate balance between treatment and control groups. An SMD >[0.1] was considered to represent significant imbalance.
To ensure that our 1:1 match with the nearest‐neighbor algorithm was not biased, we replicated PS matching using radius, kernel, and local linear regression modeling to estimate the association between EVT and the primary outcome. A second PS‐matched model was built, including the covariates from the primary model, with the addition of intravenous thrombolysis (IVT). This model was built to reduce the imbalance in IVT treatment rates between groups. Because of the high number of patients treated with BMM over EVT, a sensitivity analysis for the primary outcome was repeated using 5‐group stratification. In this analysis, patients were categorized into quintiles based on their individual PS for EVT, which is thought to significantly reduce the treatment bias related to EVT. 14
The primary outcome in this study was selected a priori as routine discharge (defined as discharge to home or self‐care). This outcome was selected as a surrogate for good early functional outcome as it has been strongly correlated with functional recovery in stroke. 12 , 15 , 16 Secondary outcomes in this study were LOS, intracerebral hemorrhage (ICH), and in‐hospital mortality. ICH was selected as a safety end point, although it is known this diagnosis is insensitively captured in the NIS because of underascertainment. 17 That said, there was no reason to believe ICH rates would be disproportionately coded among patients treated with EVT versus BMM. Symptomatic intracerebral hemorrhage was not assessed because of event rates <10 for many subgroups (also likely because of underascertainment), which cannot be reported per HCUP requirements. The outcomes of routine discharge, ICH, and in‐hospital mortality were analyzed using logistic regression with adjustment for the same covariates used in PS matching: age, sex, race, median household income quartile of zip code, hypertension, diabetes, dyslipidemia, atrial fibrillation, overweight, coronary disease, tobacco use, mCCI, and NIHSS score. In addition, treatment with IVT was added to all multivariable models. LOS was assessed using a multivariable linear regression model, with adjustment for the covariates mentioned above, including IVT.
All tests were conducted with a 2‐sided significance level of 0.05. All statistical analyses conducted in this study were performed using STATA 17.0 (StataCorp, College Station, TX). Patients with missing data were excluded from PS matching, and no imputations were performed. No sample size calculations were performed as analyses were exploratory based on available data from the NIS.
Results
Of the 1 601 840 total patients in the NIS between 2018 and 2020, 212 515 (13.2%) were diagnosed with an ICA or MCA occlusion, and 49 115 with an NIHSS score of <6 met the inclusion criteria for our study. Of these patients, 3490 (7.1%) were treated with EVT and 45 625 (92.9%) were treated with BMM (Figure 1).
Figure 1.
Inclusion diagram. ICA/MCA indicates internal carotid artery/middle cerebral artery occlusion; NIHSS, National Institutes of Health Stroke Scale; and NIS, National Inpatient Sample.
There were notable differences in baseline demographics between those treated with EVT versus BMM, with proportionally more patients treated with BMM having diabetes (SMD = 0.21) and using tobacco (SMD = 0.12). Patients treated with EVT were more overweight (SMD = −0.10) and had atrial fibrillation (SMD = −0.15). There was disproportionate treatment across hospital regions (SMD = −0.12), age groups (SMD = 0.18), income groups (SMD = −0.12), and mCCI groups (SMD = −0.20). Proportionally more patients with EVT received concomitant IVT (SMD = −0.36) and had higher baseline NIHSS score (SMD = −0.40), despite a maximum cutoff of 5 for both groups. After PS matching, only concomitant IVT remained different between treatment groups (SMD = −0.29; Table 1).
Table 1.
Baseline Demographic and Clinical Variables
| Variable | Unmatched cohort | Propensity score–matched cohort | ||||
|---|---|---|---|---|---|---|
|
Best medical management (n = 45625) |
Thrombectomy (n = 3490) |
SMD |
Best medical management (n = 3490) |
Thrombectomy (n = 3490) |
SMD | |
| Age group, n (%) | 0.18 | 0.03 | ||||
| 21–44 y | 2605 (5.7) | 320 (9.2) | 275 (7.9) | 320 (9.2) | ||
| 45–64 y | 13 135 (28.8) | 1190 (34.1) | 1180 (33.8) | 1190 (34.1) | ||
| ≥65 y | 29 885 (65.5) | 1980 (56.7) | 2035 (58.3) | 1980 (56.7) | ||
| Female sex, n (%) | 21 430 (47.0) | 1660 (47.6) | −0.01 | 1710 (49.0) | 1660 (47.6) | 0.03 |
| Race or ethnicity, n (%) | −0.03 | −0.02 | ||||
| White | 33 225 (72.8) | 2550 (73.1) | 2560 (73.4) | 2550 (73.1) | ||
| Black | 6785 (14.9) | 495 (14.2) | 510 (14.6) | 495 (14.2) | ||
| Hispanic | 3110 (6.8) | 220 (6.3) | 225 (6.4) | 220 (6.3) | ||
| Asian/Pacific Islander | 1225 (2.7) | 95 (2.7) | 80 (2.3) | 95 (2.7) | ||
| Native American | 175 (0.4) | * | * | * | ||
| Other | 1105 (2.4) | 120 (3.4) | 115 (3.3) | 120 (3.4) | ||
| Median household income quartile of zip code, n (%) | −0.12 | −0.06 | ||||
| First | 12 295 (26.9) | 735 (21.1) | 780 (22.3) | 735 (21.1) | ||
| Second | 11 585 (25.4) | 920 (26.4) | 975 (27.9) | 920 (26.4) | ||
| Third | 11 770 (25.8) | 980 (28.1) | 950 (27.2) | 980 (28.1) | ||
| Fourth | 9975 (21.9) | 855 (24.5) | 785 (22.5) | 855 (24.5) | ||
| Hospital bed size, n (%) | −0.06 | −0.07 | ||||
| Small | 18 785 (41.2) | 1380 (39.5) | 1500 (43.0) | 1380 (39.5) | ||
| Medium | 11 424 (25.0) | 810 (23.2) | 770 (22.1) | 810 (23.2) | ||
| Large | 15 415 (33.8) | 1300 (37.2) | 1220 (35.0) | 1300 (37.2) | ||
| Hospital location/teaching status, n (%) | 0.03 | −0.07 | ||||
| Rural | 10 085 (22.1) | 865 (24.8) | 975 (27.9) | 865 (24.8) | ||
| Urban nonteaching | 11 830 (25.9) | 810 (23.2) | 790 (22.6) | 810 (23.2) | ||
| Urban teaching | 23 710 (52.0) | 1815 (52.0) | 1725 (49.4) | 1815 (52.0) | ||
| Hospital region, n (%) | −0.09 | −0.02 | ||||
| Northeast | 9595 (21.0) | 615 (17.6) | 590 (16.9) | 615 (17.6) | ||
| Midwest | 10 550 (23.1) | 840 (24.1) | 845 (24.2) | 840 (24.1) | ||
| South | 18 095 (39.7) | 1370 (39.3) | 1495 (42.8) | 1370 (39.3) | ||
| West | 7384 (16.2) | 665 (19.1) | 560 (16.0) | 665 (19.1) | ||
| Medical history, n (%) | ||||||
| Hypertension | 25 195 (55.2) | 1860 (53.3) | 0.04 | 1830 (52.4) | 1860 (53.3) | −0.02 |
| Overweight | 7680 (16.8) | 730 (20.9) | −0.10 | 775 (22.2) | 730 (20.9) | 0.03 |
| Atrial fibrillation | 12 405 (27.2) | 1200 (34.4) | −0.15 | 1240 (35.5) | 1200 (34.4) | 0.02 |
| Diabetes | 15 535 (34.0) | 860 (24.6) | 0.21 | 850 (24.4) | 860 (24.6) | −0.01 |
| Dyslipidemia | 27 675 (60.7) | 1990 (57.0) | 0.07 | 2015 (57.7) | 1990 (57.0) | 0.01 |
| Coronary artery disease | 10 930 (24.0) | 750 (21.5) | 0.06 | 665 (19.1) | 750 (21.5) | −0.06 |
| Tobacco use | 20 605 (45.2) | 1365 (39.1) | 0.12 | 1270 (36.4) | 1365 (39.1) | −0.06 |
| Modified Charlson Comorbidity Index, n (%) | −0.20 | 0.03 | ||||
| 0 | 5395 (11.8) | 250 (7.2) | 260 (7.4) | 250 (7.2) | ||
| 1 | 5985 (13.1) | 300 (8.6) | 240 (6.9) | 300 (8.6) | ||
| 2 | 9695 (21.2) | 850 (24.4) | 855 (24.5) | 850 (24.2) | ||
| ≥3 | 24 550 (53.8) | 2090 (59.9) | 2135 (61.2) | 2090 (59.9) | ||
| NIHSS score, median (IQR) | 2 (1–4) | 3 (2–4) | −0.4 | 3 (2–4) | 3 (2–4) | −0.03 |
| Intravenous thrombolysis, n (%) | 6985 (15.3) | 1050 (30.1) | −0.36 | 625 (17.9) | 1050 (30.1) | −0.29 |
IQR indicates interquartile range; NIHSS, National Institutes of Health Stroke Scale; and SMD, standardized mean difference. Asterisks indicate insufficient sample for reporting per Healthcare Cost and Utilization Project guidelines. The race group ‘Other’ is reported according to local hospital and/or state practices, which may vary from year to year.
Among unmatched patients, EVT was associated with a similar rate of favorable outcome (routine discharge) compared with BMM (47.3% versus 49.7%; P = 0.24). A similar trend was found in PS‐matched patients (47.3% versus 47.1%; P = 0.96). Table S1 lists the rates of routine discharge at different levels of EVT and IVT. In multivariable logistic regression, treatment with EVT was associated with no difference in odds of routine discharge (adjusted OR [aOR], 0.99; 95% CI, 0.82–1.19; P = 0.91) compared with those treated only with BMM. This persisted with PS matching, with a similar effect estimate (aOR, 0.90; 95% CI, 0.71–1.14; P = 0.37; Table 2, Figure 2). In the PS‐matched model, other significant predictors of favorable outcome included age of at least 65 years (aOR, 0.21; 95% CI, 0.13–0.33), Black race (aOR, 0.67; 95% CI, 0.47–0.96), baseline NIHSS score (aOR, 0.83; 95% CI, 0.77–0.89), mCCI of at least 3 (aOR, 0.39; 95% CI, 0.25–0.63), and female sex (aOR, 0.64; 95% CI, 0.19–0.81).
Table 2.
Effect Estimates for Routine Discharge
| Variable | Unmatched patients | Propensity score–matched patients | ||||
|---|---|---|---|---|---|---|
| Logistic regression for routine discharge | Logistic regression for routine discharge | |||||
| Unadjusted OR (95% CI) | P value | Adjusted OR (95% CI) | P value | Adjusted OR (95% CI) | P value | |
| Treatment | ||||||
| Best medical management | Reference | Reference | Reference | |||
| Thrombectomy | 0.91 (0.77–1.06) | 0.24 | 0.99 (0.82–1.19) | 0.91 | 0.90 (0.71–1.14) | 0.37 |
| Age group, y | ||||||
| 21–44 | Reference | Reference | Reference | |||
| 45–64 | 0.66 (0.54–0.81) | <0.001 | 0.69 (0.55–0.86) | 0.001 | 0.69 (0.43–1.11) | 0.13 |
| ≥65 | 0.23 (0.19–0.28) | <0.001 | 0.25 (0.20–0.31) | <0.001 | 0.21 (0.13–0.33) | <0.001 |
| Female sex | 0.63 (0.58–0.68) | <0.001 | 0.64 (0.58–0.70) | <0.001 | 0.64 (0.51–0.81) | <0.001 |
| Race or ethnicity | ||||||
| White | Reference | Reference | Reference | |||
| Black | 0.94 (0.84–1.05) | 0.25 | 0.82 (0.72–0.94) | 0.01 | 0.67 (0.47–0.96) | 0.03 |
| Hispanic | 1.16 (0.98–1.38) | 0.09 | 1.09 (0.90–1.32) | 0.37 | 1.29 (0.76–2.19) | 0.34 |
| Asian or Pacific Islander | 1.12 (0.87–1.45) | 0.38 | 1.18 (0.89–1.56) | 0.26 | 0.62 (0.28–1.38) | 0.24 |
| Native American | 0.87 (0.46–1.63) | 0.67 | 0.88 (0.46–1.67) | 0.69 | 0.39 (0.03–4.62) | 0.46 |
| Other | 1.07 (0.84–1.36) | 0.60 | 0.99 (0.75–1.31) | 0.94 | 0.91 (0.48–1.73) | 0.78 |
| Median household income quartile of zip code | ||||||
| First | Reference | Reference | Reference | |||
| Second | 1.05 (0.94–1.17) | 0.39 | 1.07 (0.94–1.20) | 0.3 | 0.82 (0.58–1.17) | 0.28 |
| Third | 0.98 (0.88–1.09) | 0.66 | 1.01 (0.89–1.14) | 0.94 | 1.08 (0.75–1.54) | 0.69 |
| Fourth | 1.04 (0.92–1.17) | 0.53 | 1.05 (0.92–1.20) | 0.47 | 1.12 (0.78–1.62) | 0.55 |
| Medical history | ||||||
| Hypertension | 1.22 (1.12–1.33) | <0.001 | 0.98 (0.89–1.08) | 0.7 | 0.88 (0.68–1.13) | 0.33 |
| Overweight | 1.10 (0.99–1.23) | 0.07 | 1.01 (0.9–1.14) | 0.87 | 0.95 (0.70–1.29) | 0.75 |
| Atrial fibrillation | 0.61 (0.56–0.67) | <0.001 | 0.81 (0.73–0.90) | <0.001 | 0.90 (0.70–1.17) | 0.45 |
| Diabetes | 0.71 (0.65–0.78) | <0.001 | 0.96 (0.87–1.07) | 0.46 | 0.89 (0.65–1.20) | 0.43 |
| Hyperlipidemia | 0.93 (0.86–1.01) | 0.11 | 1.06 (0.96–1.16) | 0.24 | 0.88 (0.69–1.14) | 0.33 |
| Coronary artery disease | 0.74 (0.67–0.81) | <0.001 | 0.97 (0.87–1.08) | 0.59 | 1.02 (0.76–1.37) | 0.88 |
| Tobacco use | 1.34 (1.24–1.44) | <0.001 | 1.23 (1.12–1.34) | <0.001 | 1.14 (0.90–1.45) | 0.26 |
| Modified Charlson Comorbidity Index | ||||||
| 0 | Reference | Reference | Reference | |||
| 1 | 0.74 (0.62–0.88) | <0.001 | 0.82 (0.68–0.98) | 0.03 | 1.26 (0.69–2.30) | 0.45 |
| 2 | 0.58 (0.50–0.68) | <0.001 | 0.61 (0.52–0.71) | <0.001 | 0.69 (0.42–1.13) | 0.14 |
| ≥3 | 0.30 (0.26–0.34) | <0.001 | 0.34 (0.29–0.40) | <0.001 | 0.39 (0.25–0.63) | <0.001 |
| NIHSS score | 0.74 (0.72–0.76) | <0.001 | 0.74 (0.71–0.76) | <0.001 | 0.83 (0.77–0.89) | <0.001 |
| Intravenous thrombolysis | 1.37 (1.22–1.52) | <0.001 | 1.78 (1.57–2.01) | <0.001 | 1.62 (1.24–2.12) | <0.001 |
NIHSS indicates National Institutes of Health Stroke Scale; and OR, odds ratio.
Figure 2.
Margins plots for routine discharge, with 95% CIs. BMM indicates best medical management; EVT, endovascular therapy; and NIHSS, National Institutes of Health Stroke Scale.
Among unmatched patients, IVT was associated with a higher rate of favorable outcome (routine discharge) compared with patients not given IVT (56.0% versus 48.2%; P<0.0001). A similar difference was found in PS‐matched patients (53.1% versus 44.3%; P = 0.01). In multivariable logistic regression, treatment with IVT was associated with an increased odds of routine discharge (aOR, 1.78; 95% CI, 1.57–2.01; P<0.001) compared with those not treated with IVT. This persisted with PS matching, with a similar effect estimate (aOR, 1.62; 95% CI, 1.24–2.12; P<0.001; Table 2, Figure 2).
In the linear regression model for LOS, unmatched patients who underwent EVT had a longer LOS (adjusted β, 1.26; 95% CI, 0.78–1.74; P<0.001). EVT was also associated with a longer LOS in the PS matched group (adjusted β, 1.39; 95% CI, 0.82–1.96; P<0.001; Table 3).
Table 3.
Effect Estimates for LOS
| Variable | Unmatched patients | Propensity score–matched patients | ||||
|---|---|---|---|---|---|---|
| Linear regression for length of stay | Linear regression for length of stay | |||||
| Unadjusted β (95% CI) | P value | Adjusted β (95% CI) | P value | Adjusted β (95% CI) | P value | |
| Treatment | ||||||
| Best medical management | Reference | Reference | Reference | |||
| Thrombectomy | 1.53 (1.05 to 2.00) | <0.001 | 1.26 (0.78 to 1.74) | <0.001 | 1.39 (0.82 to 1.96) | <0.001 |
| Age group, y | ||||||
| 21–44 | Reference | Reference | Reference | |||
| 45–64 | −0.03 (−0.44 to 0.37) | 0.87 | 0.01 (−0.39 to 0.42) | 0.94 | 0.21 (−0.77 to 1.19) | 0.68 |
| ≥65 | −0.24 (−0.61 to 0.12) | 0.19 | −0.18 (−0.56 to 0.20) | 0.35 | 0.43 (−0.54 to 1.41) | 0.38 |
| Female sex | −0.07 (−0.24 to 0.11) | 0.45 | −0.07 (−0.26 to 0.10) | 0.40 | −0.27 (−0.91 to 0.37) | 0.41 |
| Race or ethnicity | ||||||
| White | Reference | Reference | Reference | |||
| Black | 0.91 (0.59 to 1.24) | <0.001 | 0.56 (0.25 to 0.88) | <0.001 | 0.90 (−0.3 to 1.82) | 0.06 |
| Hispanic | 0.44 (0.07 to 0.81) | 0.02 | 0.28 (−0.08 to 0.64) | 0.13 | −0.62 (−1.39 to 0.14) | 0.11 |
| Asian or Pacific Islander | 0.58 (−0.04 to 1.20) | 0.07 | 0.46 (−0.13 to 1.05) | 0.13 | 0.30 (−0.99 to 1.59) | 0.65 |
| Native American | 0.76 (−1.30 to 2.83) | 0.47 | 0.30 (−1.67 to 2.28) | 0.76 | 15.71 (−4.96 to 36.4) | 0.14 |
| Other | 0.96 (0.08 to 1.84) | 0.03 | 0.79 (−0.07 to 1.65) | 0.07 | 1.44 (−1.98 to 4.86) | 0.41 |
| Median household income quartile of zip code | ||||||
| First | Reference | Reference | Reference | |||
| Second | −0.41 (−0.68 to −0.14) | 0.003 | −0.25 (−0.51 to 0.017) | 0.06 | −0.40 (−1.33 to 0.54) | 0.40 |
| Third | −0.40 (−0.66 to −0.14) | 0.003 | −0.20 (−0.46 to 0.05) | 0.11 | −0.51 (−1.46 to 0.44) | 0.29 |
| Fourth | −0.55 (−0.81 to −0.29) | <0.001 | −0.28 (−0.53 to −0.03) | 0.03 | −0.59 (−1.60 to 0.41) | 0.25 |
| Medical history | ||||||
| Hypertension | −0.76 (−0.93 to −0.59) | <0.001 | −0.36 (−0.54 to −0.18) | <0.001 | −0.20 (−0.76 to 0.37) | 0.49 |
| Overweight | 0.40 (0.11 to 0.69) | 0.007 | 0.25 (−0.03 to 0.52) | 0.08 | −0.03 (−0.59 to 0.53) | 0.92 |
| Atrial fibrillation | 0.06 (−0.11 to 0.24) | 0.51 | −0.03 (−0.20 to 0.15) | 0.76 | −0.68 (−1.22 to −0.15) | 0.01 |
| Diabetes | 0.56 (0.36 to 0.76) | <0.001 | −0.04 (−0.23 to 0.16) | 0.73 | −0.59 (−1.19 to 0.01) | 0.053 |
| Hyperlipidemia | −0.15 (−0.33 to 0.04) | 0.11 | −0.16 (−0.34 to 0.03) | 0.10 | −0.07 (−0.70 to 0.55) | 0.83 |
| Coronary artery disease | 0.29 (0.10 to 0.49) | 0.004 | 0.04 (−0.17 to 0.25) | 0.70 | 0.06 (−0.64 to 0.75) | 0.87 |
| Tobacco use | −0.33 (−0.50 to −0.15) | <0.001 | −0.39 (−0.57 to −0.22) | <0.001 | −1.16 (−1.74 to −0.57) | <0.001 |
| Modified Charlson Comorbidity Index | ||||||
| 0 | Reference | Reference | Reference | |||
| 1 | 0.68 (0.46 to 0.91) | <0.001 | 0.62 (0.39 to 0.84) | <0.001 | 0.48 (−0.23 to 1.18) | 0.19 |
| 2 | 0.89 (0.68 to 1.10) | <0.001 | 0.69 (0.48 to 0.90) | <0.001 | 1.16 (0.46 to 1.87) | 0.001 |
| ≥3 | 2.04 (1.85 to 2.23) | <0.001 | 1.69 (1.47 to 1.91) | <0.001 | 2.19 (1.54 to 2.83) | <0.001 |
| NIHSS score | 0.36 (0.31 to 0.42) | <0.001 | 0.29 (0.22 to 0.34) | <0.001 | 0.29 (0.14 to 0.45) | <0.001 |
| Intravenous thrombolysis | −0.14 (–0.35 to 0.07) | 0.19 | −0.41 (–0.63 to −0.19) | <0.001 | −1.01 (−1.59 to −0.43) | 0.001 |
LOS indicates length of hospital stay; and NIHSS, National Institutes of Health Stroke Scale. The race group ‘Other’ is defined uniquely accordingreported according to local hospital and/or state practices, which may vary from year to year.
In the linear regression model for LOS, unmatched patients who underwent IVT had a shorter LOS (adjusted β, −0.41; 95% CI, −0.63 to −0.19; P<0.001). IVT was also associated with shorter LOS in the PS matched group (adjusted β, −1.01; 95% CI, −1.59 to −0.43; P = 0.001; Table 3).
For safety end points, the rate of ICH was low in the unmatched cohort in both cohorts (<0.1%). The small counts did not allow the t‐test to be reliable. The rate of ICH was low in both treatment groups; however, because of reporting requirements of the NIS (which preclude reporting event rates ≤10), raw counts are not provided. In unmatched patients, the overall in‐hospital mortality rate was <1%. Patients receiving EVT had higher odds of death versus those treated with BMM (EVT versus BMM, 2.2% versus 0.9%; P = 0.02; aOR, 2.11; 95% CI, 1.20–3.70; P = 0.01). With PS matching, this trend persisted (EVT versus BMM, 2.2% versus 0.9%; P = 0.03; aOR, 3.22; 95% CI, 1.19–8.68; P = 0.02).
In the unmatched cohort, IVT was associated with no change in odds of death (0.9% versus 1.1%; P = 0.44, aOR, 0.70; 95% CI, 0.39–1.24; P = 0.22). This persisted in the matched cohort (1.5% versus 1.6%; P = 0.87; aOR, 0.93; 95% CI, 0.33–2.60; P = 0.89).
Across patient subgroups, there was minimal interaction between EVT and patient‐level factors. However, with the dichotomization of NIHSS score as 0 to 2 versus 3 to 5, EVT was shown to be associated with routine discharge for patients with scores of 3 to 5 (aOR, 1.27; 95% CI, 1.03–1.57; P interaction = 0.01) in the unmatched cohort, with a trend toward benefit with PS matching (aOR, 1.18; 95% CI, 0.87–1.59; P interaction = 0.02). In the unmatched and matched cohorts, the addition or absence of IVT did not have a change in the odds of routine discharge in patients treated with EVT, when compared with patients treated with BMM (Table 4).
Table 4.
Subgroup Analysis for the Primary Outcome of Routine Discharge
| Subgroup | Unmatched cohort | Propensity score–matched cohort | ||
|---|---|---|---|---|
| Adjusted OR (95% CI) a | P value for interaction † | Adjusted OR (95% CI) a | P value for interaction † | |
| NIHSS category | 0.01 | 0.02 | ||
| NIHSS score 0–2 | 0.61 (0.47–0.79) | 0.61 (0.41–0.90) | ||
| NIHSS score 3–5 | 1.27 (1.03–1.57) | 1.18 (0.87–1.59) | ||
| Age category, y | 0.50 | 0.88 | ||
| 21–44 | 1.28 (0.68–2.43) | 1.41 (0.58–3.40) | ||
| 45–64 | 1.05 (0.76–1.44) | 0.95 (0.64–1.42) | ||
| ≥65 | 0.91 (0.72–1.15) | 0.86 (0.62–1.18) | ||
| Sex | 0.13 | 0.18 | ||
| Female | 0.94 (0.72–1.21) | 1.01 (0.72–1.41) | ||
| Male | 1.06 (0.83–1.35) | 0.85 (0.60–1.19) | ||
| Medical history | ||||
| No hypertension | 1.04 (0.79–1.36) | 0.09 | 0.88 (0.62–1.26) | 0.45 |
| Hypertension | 0.95 (0.73–1.22) | 0.92 (0.66–1.28) | ||
| Not overweight | 1.01 (0.83–1.23) | 0.56 | 0.92 (0.71–1.19) | 0.51 |
| Overweight | 0.95 (0.64–1.39) | 0.88 (0.53–1.47) | ||
| No atrial fibrillation | 1.00 (0.79–1.24) | 0.69 | 0.93 (0.69–1.25) | 0.98 |
| Atrial fibrillation | 0.99 (0.72–1.34) | 0.89 (0.60–1.33) | ||
| No diabetes | 0.93 (0.75–1.15) | 0.63 | 0.81 (0.61–1.06) | 0.63 |
| Diabetes | 1.20 (0.84–1.71) | 1.59 (0.97–2.60) | ||
| No dyslipidemia | 0.98 (0.73–1.30) | 0.62 | 0.69 (0.47–1.01) | 0.17 |
| Dyslipidemia | 1.01 (0.80–1.28) | 1.11 (0.81–1.52) | ||
| No coronary artery disease | 0.95 (0.77–1.17) | 0.85 | 0.84 (0.64–1.10) | 0.63 |
| Coronary artery disease | 1.19 (0.83–1.71) | 1.42 (0.82–2.46) | ||
| No tobacco use | 0.84 (0.65–1.07) | 0.06 | 0.65 (0.48–0.87) | 0.03 |
| Tobacco use | 1.23 (0.93–1.61) | 1.50 (1.01–2.22) | ||
| Modified Charlson Comorbidity Index | 0.58 | 0.55 | ||
| 0 | 0.77 (0.41–1.46) | 0.92 (0.29–2.92) | ||
| 1 | 1.35 (0.74–2.49) | 0.78 (0.29–2.10) | ||
| 2 | 1.11 (0.77–1.59) | 1.17 (0.74–1.86) | ||
| ≥3 | 0.93 (0.74–1.18) | 0.81 (0.60–1.10) | ||
| Intravenous thrombolysis | 0.47 | 0.32 | ||
| No | 1.11 (0.90–1.37) | 0.79 (0.78–1.34) | ||
| Yes | 0.78 (0.57–1.07) | 0.63 (0.37–1.09) | ||
NIHSS indicates National Institutes of Health Stroke Scale; and OR, odds ratio.
Effect estimates are adjusted for age group, sex, race group, hypertension, diabetes, dyslipidemia, atrial fibrillation, overweight, coronary disease, tobacco use, modified Charlson Comorbidity Index, NIHSS score, and treatment with intravenous thrombolysis.
† P values for interaction with endovascular therapy are shown.
The alternative PS‐matching strategies (radius, kernel, and local linear regression algorithms) also demonstrated that EVT was associated with higher odds of discharge, shorter hospital stays, and no difference in in‐hospital mortality rates (data not otherwise shown). To ensure that IVT was confounding the insignificant effect of EVT on the primary end point (routine discharge), we ran a secondary PS match where IVT was included in the matching algorithm. After including IVT in the match, there were a total of 3490 patients in each of the treatment groups. A total of 1050 patients (30.1%) and 1010 patients (28.9%) were given IVT in the EVT and BMM groups, respectively (SMD = −0.03). In multivariable logistic regression for routine discharge in this secondary analysis, there remained no difference in the odds of routine discharge with EVT (aOR, 0.94; 95% CI, 0.74–1.19; P = 0.59). In that multivariable model, IVT remained independently associated with a higher probability of routine discharge (aOR, 1.53; 95% CI, 1.22–1.79; P<0.001). In the sensitivity analysis for the primary outcome using 5‐group stratification, there remained no significant increase in the odds of the primary outcome with EVT (Table 5).
Table 5.
Odds of Routine Discharge in a 5‐Group Stratification Propensity Matching Scheme
| Quintile of propensity score | Raw frequency of routine discharge with EVT, % (n/total) | Raw frequency of routine discharge without EVT, % (n/total) | Odds of routine discharge if given EVT, aOR (95% CI); P value |
|---|---|---|---|
| First | 66.0 (155/235) | 64.7 (6205/9591) | 3.90 (0.96–15.88); P = 0.06 |
| Second | 59.0 (230/390) | 54.2 (5110/9431) | 0.44 (0.18–1.06); P = 0.07 |
| Third | 48.1 (315/655) | 51.5 (4720/9171) | 0.67 (0.36–1.27); P = 0.22 |
| Fourth | 40.6 (345/850) | 39.4 (3535/8971) | 0.71 (0.43–1.17); P = 0.18 |
| Fifth | 44.5 (605/1360) | 36.2 (3060/8461) | 1.14 (0.78–1.66); P = 0.51 |
aOR indicates adjusted odds ratio; and EVT, endovascular therapy.
Discussion
In this real‐world analysis of patients with proximal anterior circulation LVO with mild deficits, according to the NIHSS (scores of <6), use of EVT was not associated with better short‐term functional outcomes or shorter LOS when compared with BMM. In subgroup analyses, we observed a modest association between EVT and the primary outcome for patients with NIHSS scores of 3 to 5. Although this association persisted in the unmatched analysis, it lost statistical significance in the PS‐matched analysis. Moreover, we observed a higher odds of early mortality with EVT in this patient population in both the adjusted unmatched and matched cohorts. Overall mortality rates were low (1%–2%) in this cohort, but there was a significant increase in the odds of in‐hospital death among patients treated with EVT. Because of the low reporting of ICH in the NIS (likely because of underascertainment) and lack of discrete data on technical complications of EVT, it is unclear if the higher mortality rate with EVT is associated with procedural factors or if it is related to unmeasured confounders. For example, lack of data on patient clinical deterioration, infarct volume, or perfusion imaging estimation of tissue‐at‐risk precludes any mediation analysis that may partially explain the early mortality risk. Regardless, the unclear benefit of EVT in this population coupled with the potentially higher risk of early mortality should be interpreted with caution and warrants validation in ongoing randomized clinical trials.
Although these results suggest no significant benefit of EVT in patients with anterior proximal LVO with low NIHSS score, the NIHSS score is an imperfect measure of disability. A patient may experience severe aphasia without weakness or vision impairment and have a baseline NIHSS score of <6. Such an aphasia may be disabling and could potentially disqualify a patient from EVT on the basis of simplified treatment criteria using the NIHSS score, according to published guidelines from the American Heart Association/American Stroke Association and Society of Vascular and Interventional Neurology. 18 , 19 That said, the Society of Neurointerventional Surgery recommended thrombectomy for NIHSS score of <6 when symptoms are disabling. 20 Prior studies have been equivocal in determining the benefit for EVT in this patient population. 11 , 21 , 22 The ongoing questions about EVT efficacy may partially account for the low percentage of patients treated with EVT in our cohort (7.1%). Despite our analysis of a large population data set not showing benefits of EVT, we await clinical trial results for further confirmation. Currently, the ENDOLOW (Endovascular Therapy for Low NIHSS Ischemic Strokes) and the MOSTE (Minor Stroke Therapy Evaluation) clinical trials are ongoing. 23 , 24
Other cohort studies have been equivocal in determining the benefit for EVT in this patient population. In 1 matched cohort study of 2 prospective registries of patients with anterior LVO and NIHSS score of <6, Haussen et al found EVT to be significantly and independently associated with greater improvement in NIHSS score at discharge and lower disability at discharge and 3 to 6 months. 9 Similarly, infarct volume and rates of early mortality were lower with EVT in a separate cohort reported by Abbas et al. 25 That said, larger cohort studies, including this 1, have failed to identify benefit (or have reported harm) with EVT in those with low NIHSS score. 26 In the largest cohort to date, investigators from the Swiss Stroke Registry used PS matching to estimate the odds of a favorable 90‐day outcome (modified Rankin scale [mRS] score, 0–1) with EVT±IVT versus IVT alone. 10 There was no difference in the primary or secondary outcomes (including mortality) with EVT when added to IVT. In contrast to that study, our analysis included patients not treated with IVT, and still found no significant difference in our primary functional outcome with EVT. In a separate analysis of overlapping years in the NIS, restricting inclusion of patients to those with basilar artery occlusion and NIHSS score of <10, we also found no independent association between EVT and discharge to home, although there was a significant benefit for IVT. 12 Despite differences between our analysis of NIS data and analysis from the Swiss Stroke Registry, including inclusion criteria (with the Swiss Stroke Registry including only patients treated with IVT), differences in EVT decision‐making between US and Swiss centers, or differences in selection of a “good outcome” (where the Swiss Stroke Registry used 90‐day mRS score, whereas this analysis of the NIS used discharge to home), both studies found no significant difference in functional outcomes.
Because this data set did not contain mRS score, routine discharge was used as a surrogate marker for good functional outcomes. Prior studies have shown that favorable discharge outcome correlates to an mRS score of ≤2 at 90 days. 27 , 28 In our analysis, approximately half of all patients were routinely discharged irrespective of treatment, with half being discharged to destinations where ongoing therapy and other services are necessary. That said, the relatively low rates of routine discharge (based on low NIHSS scores) in our sample may further provide evidence that NIHSS scores should not be solely used to guide clinical decision‐making. Furthermore, an oversimplified quantification of neurologic deficits on the basis of NIHSS score may confound prognostication, as patients with similar low NIHSS scores may have vastly different clinical outcomes depending on their symptoms. For example, patients with similarly low NIHSS scores and deficits in motor function had lower rates of early mortality and morbidity compared with those who had language and consciousness deficits, according to 1 study. 29 As such, we caution the use of NIHSS score alone to guide EVT decision‐making.
Furthermore, in our study, we found EVT was not associated with shorter LOS (compared with those who were treated with BMM), but treatment with IVT was. Among patients with more severe presenting deficits (NIHSS score of ≥10), previous clinical trial data from Endovascular Therapy Following Imaging Evaluation for Ischemic Stroke 3 (DEFUSE‐3) showed patients treated with EVT had shorter LOS and increased time at home in the first 90 days following stroke. 30 It is unclear from this data set whether EVT was associated with technical complications, neurologic deterioration, or both, which would result in longer hospitalizations. Alternatively, the longer LOS among patients with EVT may be related to residual confounding. For example, those treated with EVT may have experienced fluctuations or deteriorations that would not have been captured by the initial NIHSS score. Furthermore, patients treated with EVT may have had more disabling symptoms (eg, severe aphasia or neglect) than those treated with BMM, despite having a similar NIHSS score.
In this analysis, IVT was strongly and independently associated with a shorter LOS. 31 This may be related to the reperfusion advantage of IVT at the site of LVO or via surrounding microcirculation; however, follow‐up imaging of infarct volume and persistence of occlusion are unavailable in the NIS. Alternatively, treatment with IVT may be a surrogate indicator for earlier clinical presentation, with presumably more effective short‐term treatment, including the use of reperfusion therapies as well as other interventions before the ischemic injury becomes irreversible. In addition to these potential benefits of IVT, we report a low rate of ICH and early mortality with IVT, which supports the safety of this intervention in patients with a low NIHSS score. 32 In a secondary analysis in which IVT was included in the PS matching, there remained no significant difference in routine discharge with EVT. This demonstrates that the unequal balance of IVT in the original PS match algorithm (which did not match on the basis of IVT use) did not explain the insignificant effect of EVT on routine discharge. Furthermore, IVT remained strongly associated with routine discharge in both the primary and the secondary PS‐matched analyses. The lack of signal of harm coupled with the greater probability of a favorable discharge, and earlier discharge, provide strong arguments to consider IVT in patients with low NIHSS score, particularly when symptoms may be fluctuating or disabling.
Other notable factors associated with decreased odds of favorable discharge outcome included elderly age and higher mCCI scores. Elderly patients are known to have a poorer functional reserve; and despite achieving similar reperfusion and symptomatic intracerebral hemorrhage rates as younger counterparts, they experience significantly poorer clinical outcomes. 33 In addition, among patients with stroke, a 1‐point increase in the Charlson Comorbidity Index has been shown to result in a 15% increase in the odds of a poor outcome at discharge and a 29% increase in the odds of death by 1 year, according to prior data. 34
Limitations
Despite the positive findings, our study does have several limitations. First, the data set does not have imaging data, which may be critical in determining eligibility for EVT. 35 The lack of imaging in this database can be a significant confounder. Imaging data would provide critical details about a patient's condition severity. The lack of imaging data can explain the lack of treatment effect on routine discharge and the higher mortality associated with EVT in this data set. Further elements, such as time last known well, time to arterial puncture, prestroke mRS score, 36 clinical deterioration, and 90‐day mRS score, were not available within this data set (as such, surrogates of mRS score were used in the analysis). In addition, because this analysis was limited to patients with low NIHSS score, there may be a higher rate of misclassification of patients who had prior or chronic infarcts, and prior LVO as new stroke diagnoses. It is possible that many patients included in this analysis may have had preexisting deficits from a chronic occlusion, and their admission NIHSS score may have represented preexisting neurologic symptoms. Fluctuations in NIHSS score were also not captured in the NIS, nor were perfusion imaging estimates, collateral status, more detailed occlusion locations or recanalization rates related to IVT, 37 or other indicators of tissue at risk, so we cannot report the unique details of thrombectomy decision‐making in this cohort.
Conclusions
Overall, in patients presenting with low NIHSS score and LVO of the proximal anterior circulation, we did not find better outcomes for patients treated with EVT compared with BMM. However, IVT was associated with increased odds of routine discharge and deceased LOS. Our findings illustrate real‐world outcomes for EVT and IVT in patients with ICA/MCA strokes who have a low NIHSS score. We await clinical trial results to further provide data (both clinical and imaging factors) to determine if these treatments may be beneficial in this patient population.
Sources of Funding
None.
Disclosures
Dr Nguyen reports research support from SVIN and Medtronic; and advisory board participation with Idorsia and Brainomix. Dr Siegler reports research support from Medtronic, Philips, and Viz.ai. Thanh N. Nguyen serves on the Editorial Board of S:VIN. Editorial Board Members are not involved in the handling or final disposition of submissions. James Siegler serves on the Digital Strategies Team of S:VIN. Digital Strategies Team Members are not involved in the handling or final disposition of submissions. Shadi Yaghi is a Guest Associate Editor for S:VIN and was not involved in the handling or final disposition of this article. Disclosures provided by Shadi Yaghi in compliance with American Heart Association's annual Journal Editor Disclosure Questionnaire are available at https://www.ahajournals.org/editor-coi-disclosures. The remaining authors have no disclosures.
Supporting information
Online Table S1.
Acknowledgments
None.
Clinical Trial Registration Information: Not applicable.
This article is a republication of the retracted article (https://www.ahajournals.org/doi/10.1161/SVIN.123.000998). The errors in the original article have been corrected and a letter of explanation is provided as a supplemental file.
References
- 1. Berkhemer OA, Fransen PSS, Beumer D, Van Den Berg LA, Lingsma HF, Yoo AJ, Schonewille WJ, Vos JA, Nederkoorn PJ, Wermer MJH, et al. A randomized trial of intraarterial treatment for acute ischemic stroke. N Engl J Med. 2015;372:11–20. [DOI] [PubMed] [Google Scholar]
- 2. Goyal M, Demchuk AM, Menon BK, Eesa M, Rempel JL, Thornton J, Roy D, Jovin TG, Willinsky RA, Sapkota BL, et al. Randomized assessment of rapid endovascular treatment of ischemic stroke. N Engl J Med. 2015;372:1019–1030. [DOI] [PubMed] [Google Scholar]
- 3. Campbell BCV, Mitchell PJ, Kleinig TJ, Dewey HM, Churilov L, Yassi N, Yan B, Dowling RJ, Parsons MW, Oxley TJ, et al. Endovascular therapy for ischemic stroke with perfusion‐imaging selection. N Engl J Med. 2015;372:1009–1018. [DOI] [PubMed] [Google Scholar]
- 4. Saver JL, Goyal M, Bonafe A, Diener H‐C, Levy EI, Pereira VM, Albers GW, Cognard C, Cohen DJ, Hacke W, et al. Stent‐retriever thrombectomy after intravenous t‐PA vs. t‐PA alone in stroke. N Engl J Med. 2015;372:2285–2295. [DOI] [PubMed] [Google Scholar]
- 5. Jovin TG, Chamorro A, Cobo E, de Miquel MA, Molina CA, Rovira A, San Román L, Serena J, Abilleira S, Ribó M, et al. Thrombectomy within 8 hours after symptom onset in ischemic stroke. N Engl J Med. 2015;372:2296–2306. [DOI] [PubMed] [Google Scholar]
- 6. Nogueira RG, Jadhav AP, Haussen DC, Bonafe A, Budzik RF, Bhuva P, Yavagal DR, Ribo M, Cognard C, Hanel RA, et al. Thrombectomy 6 to 24 hours after stroke with a mismatch between deficit and infarct. N Engl J Med 2018;378:11–21. [DOI] [PubMed] [Google Scholar]
- 7. Albers GW, Marks MP, Kemp S, Christensen S, Tsai JP, Ortega‐Gutierrez S, McTaggart RA, Torbey MT, Kim‐Tenser M, Leslie‐Mazwi T, et al. Thrombectomy for stroke at 6 to 16 hours with selection by perfusion imaging. N Engl J Med. 2018;378:708–718. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Heldner MR, Zubler C, Mattle HP, Schroth G, Weck A, Mono M‐L, Gralla J, Jung S, El‐Koussy M, Lüdi R, et al. National Institutes of Health stroke scale score and vessel occlusion in 2152 patients with acute ischemic stroke. Stroke. 2013;44:1153–1157. [DOI] [PubMed] [Google Scholar]
- 9. Haussen DC, Lima FO, Bouslama M, Grossberg JA, Silva GS, Lev MH, Furie K, Koroshetz W, Frankel MR, Nogueira RG. Thrombectomy versus medical management for large vessel occlusion strokes with minimal symptoms: an analysis from STOPStroke and GESTOR cohorts. J Neurointerv Surg. 2018;10:325–329. [DOI] [PubMed] [Google Scholar]
- 10. Manno C, Disanto G, Bianco G, Nannoni S, Heldner M, Jung S, Arnold M, Kaesmacher J, Müller M, Thilemann S, et al. Outcome of endovascular therapy in stroke with large vessel occlusion and mild symptoms. Neurology. 2019;93:e1618–e1626. [DOI] [PubMed] [Google Scholar]
- 11. McCarthy DJ, Tonetti DA, Stone J, Starke RM, Narayanan S, Lang MJ, Jadhav AP, Gross BA. More expansive horizons: a review of endovascular therapy for patients with low NIHSS scores. J Neurointerv Surg. 2021;13:146–151. [DOI] [PubMed] [Google Scholar]
- 12. Patel K, Taneja K, Obusan MB, Yaghi S, Nguyen TN, Thon JM, Kass‐Hout T, Brorson JR, Prabhakaran S, Siegler JE. Real‐world outcomes for basilar artery occlusion thrombectomy with mild deficits: the national inpatient sample. Stroke. 2023;54:2031–2039 [DOI] [PubMed] [Google Scholar]
- 13. Patel K, Taneja K, Wolfe J, Campellone JV, Farooqui M, Ortega‐Gutierrez S, Siegler JE. Association between race and length of stay among stroke patients: the national US emergency departments data set. Stroke Vasc Interv Neurol. 2023;3:e000591 0:e10.1161/SVIN.12410. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Cochran WG. The effectiveness of adjustment by subclassification in removing bias in observational studies. Biometrics. 1968;24:295–313. [PubMed] [Google Scholar]
- 15. Saber H, Navi BB, Grotta JC, Kamel H, Bambhroliya A, Vahidy FS, Chen PR, Blackburn S, Savitz SI, McCullough L, et al. Real‐world treatment trends in endovascular stroke therapy. Stroke. 2019;50:683–689. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kramer AM, Steiner JF, Schlenker RE, Eilertsen TB, Hrincevich CA, Tropea DA, Ahmad LA, Eckhoff DG. Outcomes and costs after hip fracture and stroke. A comparison of rehabilitation settings. JAMA. 1997;277:396–404. [PubMed] [Google Scholar]
- 17. Moradiya Y, Levine SR. Comparison of short‐term outcomes of thrombolysis for in‐hospital stroke and out‐of‐hospital stroke in United States. Stroke. 2013;44:1903–1908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18. Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis NC, Becker K, Biller J, Brown M, Demaerschalk BM, Hoh B, et al. Guidelines for the early management of patients with acute ischemic stroke: 2019 update to the 2018 guidelines for the early management of acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2019;50:e344–e418. [DOI] [PubMed] [Google Scholar]
- 19. Nguyen TN, Castonguay AC, Siegler JE, Nagel S, Lansberg MG, de Havenon A, Sheth SA, Abdalkader M, Tsai JP, Albers GW, et al. Mechanical thrombectomy in the late presentation of anterior circulation large vessel occlusion stroke: a guideline from the society of vascular and interventional neurology guidelines and practice standards committee. Stroke Vasc Interv Neurol. 2023;e000512000512. https://www.ahajournals.org/doi/abs/10.1161/SVIN.122.000512 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20. Mokin M, Ansari SA, McTaggart RA, Bulsara KR, Goyal M, Chen M, Fraser JF. Indications for thrombectomy in acute ischemic stroke from emergent large vessel occlusion (ELVO): report of the SNIS standards and guidelines committee [Internet]. J Neurointerv Surg. 2019;11:215–220. 10.1136/neurintsurg-2018-014640 [DOI] [PubMed] [Google Scholar]
- 21. Goyal N, Tsivgoulis G, Malhotra K, Ishfaq MF, Pandhi A, Frohler MT, Spiotta AM, Anadani M, Psychogios M, Maus V, et al. Medical management vs mechanical thrombectomy for mild strokes: an international multicenter study and systematic review and meta‐analysis. JAMA Neurol. 2020;77:16–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Nagel S, Bouslama M, Krause LU, Küpper C, Messer M, Petersen M, Lowens S, Herzberg M, Ringleb PA, Möhlenbruch MA, et al. Mechanical thrombectomy in patients with milder strokes and large vessel occlusions. Stroke. 2018;49:2391–2397. [DOI] [PubMed] [Google Scholar]
- 23. Endovascular Therapy for Low NIHSS Ischemic Strokes – Full Text View – ClinicalTrials.Gov. [cited 2023 May 7]. Accessed 31 December 2023. https://clinicaltrials.gov/ct2/show/NCT04167527
- 24. Minor Stroke Therapy Evaluation – Full Text View – ClinicalTrials.Gov. [cited 2023 May 7]. Accessed 31 December 2023. https://clinicaltrials.gov/ct2/show/NCT03796468
- 25. Abbas R, Herial NA, Naamani KE, Sweid A, Weinberg JH, Habashy KJ, Tjoumakaris S, Gooch MR, Rosenwasser RH, Jabbour P. Mechanical thrombectomy in patients presenting with NIHSS score <6: a safety and efficacy analysis. J Stroke Cerebrovasc Dis. 2022;31:106282. [DOI] [PubMed] [Google Scholar]
- 26. Volny O, Zerna C, Tomek A, Bar M, Rocek M, Padr R, Cihlar F, Nevsimalova M, Jurak L, Havlicek R, et al. Thrombectomy vs medical management in low NIHSS acute anterior circulation stroke. Neurology. 2020;95:e3364–e3372. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Qureshi AI, Chaudhry SA, Sapkota BL, Rodriguez GJ, Suri MFK. Discharge destination as a surrogate for Modified Rankin Scale defined outcomes at 3‐ and 12‐months poststroke among stroke survivors. Arch Phys Med Rehabil. 2012;93:1408–1413.e1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28. ElHabr AK, Katz JM, Wang J, Bastani M, Martinez G, Gribko M, Hughes DR, Sanelli P. Predicting 90‐day modified Rankin Scale score with discharge information in acute ischaemic stroke patients following treatment. BMJ Neurol Open. 2021;3:e000177. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Sucharew H, Khoury J, Moomaw CJ, Alwell K, Kissela BM, Belagaje S, Adeoye O, Khatri P, Woo D, Flaherty ML, et al. Profiles of the National Institutes of Health Stroke Scale items as a predictor of patient outcome. Stroke. 2013;44:2182–2187. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Tate WJ, Polding LC, Kemp S, Mlynash M, Heit JJ, Marks MP, Albers GW, Lansberg MG. Thrombectomy results in reduced hospital stay, more home‐time, and more favorable living situations in DEFUSE 3. Stroke. 2019;50:2578–2581. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Kasemsap N, Vorasoot N, Kongbunkiat K, Peansukwech U, Tiamkao S, Sawanyawisuth K. Impact of intravenous thrombolysis on length of hospital stay in cases of acute ischemic stroke. Neuropsychiatr Dis Treat. 2018;14:259. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 32. Khatri P, Kleindorfer DO, Devlin T, Sawyer RN Jr, Starr M, Mejilla J, Broderick J, Chatterjee A, Jauch EC, Levine SR, et al. Effect of alteplase vs aspirin on functional outcome for patients with acute ischemic stroke and minor nondisabling neurologic deficits: the PRISMS randomized clinical trial. JAMA. 2018;320:156–166. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33. Chandra RV, Leslie‐Mazwi TM, Oh DC, Chaudhry ZA, Mehta BP, Rost NS, Rabinov JD, Hirsch JA, González RG, Schwamm LH, et al. Elderly patients are at higher risk for poor outcomes after intra‐arterial therapy. Stroke. 2012;43:2356–2361. [DOI] [PubMed] [Google Scholar]
- 34. Goldstein LB, Samsa GP, Matchar DB, Horner RD. Charlson Index comorbidity adjustment for ischemic stroke outcome studies. Stroke. 2004;35:1941–1945. [DOI] [PubMed] [Google Scholar]
- 35. Abdalkader M, Siegler JE, Lee JS, Yaghi S, Qiu Z, Huo X, Miao Z, Campbell BCV, Nguyen TN. Neuroimaging of acute ischemic stroke: multimodal imaging approach for acute endovascular therapy. J Stroke Cerebrovasc Dis. 2023;25:55–71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36. Haussen DC, Al‐Bayati AR, Mohammaden MH, Sheth SA, Salazar‐Marioni S, Linfante I, Dabus G, Starosciak AK, Hassan AE, Tekle WG, et al. The Society of Vascular and Interventional Neurology (SVIN) mechanical thrombectomy registry: methods and primary results. Stroke Vasc Interv Neurol. 2022;2:e000234. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37. Lau HL, Gardener H, Coutts SB, Saini V, Field TS, Dowlatshahi D, Smith EE, Hill MD, Romano JG, Demchuk AM, et al. Radiographic characteristics of mild ischemic stroke patients with visible intracranial occlusion: The INTERRSeCT Study. Stroke. 2021;53:913‐920. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Online Table S1.