Effect of SARS-CoV-2 Infection on Endovascular Thrombectomy Outcomes—Data from the Florida Stroke Registry

Graphical Abstract

Abstract

BACKGROUND AND PURPOSE:

Endovascular thrombectomy outcomes are impacted by changes in stroke systems of care. During the pandemic, SARS-CoV-2 positive status had major implications on hospital arrival and treatment models of non-COVID-related hospital admissions. Using the Florida Stroke Registry, we compared the rates of in-hospital death and discharge outcomes of patients treated with endovascular thrombectomy who tested positive for SARS-CoV-2 infection during their hospitalization.

MATERIALS AND METHODS:

Data from Get with the Guidelines–Stroke hospitals participating in the Florida Stroke Registry during the COVID pandemic from March 2020 to December 2022 were reviewed to identify endovascular thrombectomy patients with coding for SARS-CoV-2 testing during their hospital stay. Associations between SARS-CoV-2 status and favorable endovascular thrombectomy outcomes of mRS (0–2) at discharge, discharge to home or rehabilitation center, symptomatic intracerebral hemorrhage, in-hospital mortality, and independent ambulation at discharge were examined by using multivariate logistic regression modeling adjusting for demographics, vascular risk factors, and clinical characteristics. Temporal analyses were used to compare outcomes across the study period.

RESULTS:

A total of 8184 patients underwent endovascular thrombectomy (median age 71.1 years, women 49.7%, mean NIHSS 14). Of these, 180 (2.20%) were SARS-CoV-2 positive. Compared with SARS-CoV-2 negative endovascular thrombectomy patients, those who tested positive were younger, more frequently men, but with comparable stroke severity at presentation. In multivariable analysis, adjusting for baseline differences and confounding variables, there was a 33% lower likelihood of being discharged to home/inpatient rehab (OR = 0.67, 95% CI: 0.49–0.93), 65% higher odds of in-hospital death (OR = 1.65, 95% CI: 1.06–2.58), as well as a 91% less chance of having a high mRS (>2) at discharge (OR = 0.15, 95% CI: 0.04–0.60) for patients with positive SARS-CoV-2 infection. However, a similar risk of symptomatic intracerebral hemorrhage was present compared with SARS-CoV-2 negative patients (OR = 0.97, 95% CI: 0.50–1.88). Temporal analysis of SARS-CoV-2 positive patients showed no significant differences.

CONCLUSIONS:

In this large multicenter stroke registry, despite comparable clinical presentation and treatment timelines, SARS-CoV2 positive status negatively impacted thrombectomy outcomes.

Endovascular thrombectomy (EVT) is the standard of care for the treatment of select patients with acute stroke related to a large vessel.¹ Additionally, it is also known that post EVT care in a dedicated stroke unit or neurologic intensive care unit is superior to post-EVT care in other units.^2,3 These treatment paradigms, or systems of care, are essential in ensuring optimal outcomes in patients receiving EVT for large vessel acute ischemic strokes (AIS).

The SARS-CoV-2 pandemic created unprecedented challenges to every aspect of health care, especially in systems of care and resource allocation.⁴ These challenges were amplified in the setting of emergency care and time-sensitive treatments, including EVT for large vessel AIS. Though there has been research examining the relationship between SARS-CoV-2 and AIS, as well as AIS and EVT outcomes, there are limited data reports on outcomes following EVT in patients with AIS who are SARS-CoV-2 infection positive.^5-12

The Florida Stroke Registry (FSR) is a statewide database that collects data on stroke care throughout the state. FSR provides a unique opportunity to examine the impact of SARS-CoV-2 infection status on the outcomes of EVT in patients with AIS. We compared in-hospital and discharge outcomes of endovascularly treated patients with versus without a SARS-CoV-2 positive infection status during their hospitalization.

MATERIALS AND METHODS

Data from Get with the Guidelines–Stroke hospitals participating in the Florida Stroke Registry from March 2020 to December 2022 were reviewed. This study period was selected because, on March 9, 2020, the Florida Governor issued an executive order that declared a state of emergency for the entire state due to COVID-19.^13-16 Figure 1 describes the details of study inclusion. Patients who had an International Classification of Diseases, 10th revision (ICD-10) procedural code for EVT and coding for SARS-CoV-2 testing during their hospital stay were included. EVT was defined as the use of any form of mechanical thrombectomy device for the treatment of an acute intracranial occlusion with or without intra-arterial thrombolysis. These patients were categorized into SARS-CoV-2 positive or negative groups (Supplemental Data). Associations between SARS-CoV-2 status, clinical presentation, and post-EVT hospital outcomes were studied. Outcomes including mRS (0–2), discharge to home or to an acute rehabilitation center, independent ambulation at discharge, symptomatic intracerebral hemorrhage (sICH), and in-hospital mortality were examined by using multivariate logistic regression modeling adjusting for demographics, vascular risk factors, and clinical characteristics with generalized estimating equations. ORs with 95% CIs were calculated by using multivariate modeling with age, sex, race, insurance, smoker status, hypertension, diabetes, dyslipidemia, atrial fibrillation, history of coronary artery disease, previous stroke, arrival mode, IV thrombolysis, region, large vessel occlusion (LVO) location, admission NIHSS, period, onset to arrival, door to CT, door to needle, door to puncture, onset to puncture, onset to puncture <6 hours, onset to puncture 6 to 24 hours, onset to puncture >24 hours, and admitting service as covariates. Two different models were run to elucidate the most impactful covariates and minimize confounding. The first multivariate model used age, sex, insurance, race, EMS delivery mode, time variables, tPA administration, diabetes mellitus status, and diagnosis of atrial fibrillation. The second multivariate model used all previously mentioned variables and also the remaining covariates of smoker status, hypertension, hyperlipidemia, coronary artery disease, prior stroke, LVO location, arrival NIHSS score, admitting service, and admission care. Temporal analyses compared outcomes in SARS-CoV-2 positive patients across the years 2020–2022 by using 2020 as the index year and comparing against 2021 and 2022. Temporal analysis was completed on only the outcomes of good mRS at discharge, discharge to home or an acute rehabilitation center, sICH, and in-hospital mortality due to a low number of patients independently ambulating at baseline. For covariates with informative missingness, the Missing Indicator Method, which adds indicator variables to indicate the missing pattern, was used in the multivariate analysis to improve model performance.

FIG 1.

Details of study participant inclusion flow chart.

RESULTS

A total of 8184 patients underwent EVT during the study period, of which 180 (2.20%) were SARS-CoV-2 positive. Study participant characteristics and hospital characteristics are presented in Supplemental Data.

The patients with SARS-CoV-2 infection were more likely to be younger (67.2 years versus 71.1 years), be men (58.3% men versus 41.7% women), have increased onset to arrival time (229 minutes versus 161 minutes), and have increased onset to puncture time (379 minutes versus 291 minutes).

Overall, 80.78% of all EVT patients had an mRS >2 at discharge, 4670 (57.06%) were independently ambulating, and 4760 (58.16%) were discharged to home or a rehabilitation center. Additionally, overall, EVT patients, 447 (5.46%) had sICH, and 783 (9.57%) died in the hospital. Table 1 details in-hospital mortality, rate of sICH, and favorable discharge outcomes stratified based on SARS-CoV-2 status. As shown in Table 2, adjusting for baseline differences and confounding variables there was a 33% lower likelihood of being discharged to home/inpatient rehab (OR = 0.67, 95% CI: 0.49–0.93), 65% higher odds of in-hospital death (OR = 1.65, 95% CI: 1.06–2.58), as well as a 91% less chance of having a high mRS (>2) at discharge (OR = 0.15, 95% CI: 0.04–0.60) for patients with positive SARS-CoV-2 infection. However, a similar risk of symptomatic intracerebral hemorrhage was present compared with SARS-CoV-2 negative patients (OR = 0.97, 95% CI: 0.50–1.88).

Table 1:

Raw data for investigated patient outcomes^a

Variable	Level	Overall	SARS-CoV-2? No	SARS-CoV-2? Yes	P Value
		n = 8184	n = 8004	n = 180
Discharge location	1-Home	2894 (35.36%)	2839 (35.47%)	55 (30.56%)	.396
	2-Rehab	1866 (22.80%)	1827 (22.83%)	39 (21.67%)
	3-Hosp	1059 (12.94%)	1031 (12.88%)	28 (15.56%)
	4-SNF	1383 (16.90%)	1354 (16.92%)	29 (16.11%)
	5-Other	199 (2.43%)	194 (2.42%)	5 (2.78%)
	6-Died	783 (9.57%)	759 (9.48%)	24 (13.33%)
Discharge ambulation	1-Able to ambulate independently	4670 (57.06%)	4555 (56.91%)	115 (63.89%)	.171
	2-Unable/with assistance	355 (4.34%)	348 (4.35%)	7 (3.89%)
	3-ND/missing	3159 (38.60%)	3101 (38.74%)	58 (32.22%)
Discharge mRS	High (>2)	1488 (80.78%)	1445 (80.41%)	43 (95.56%)	.011
	Low (0–2)	354 (19.22%)	352 (19.59%)	2 (4.44%)
Symptomatic intracerebral hemorrhage <36 hrs	N	7737 (94.54%)	7566 (94.53%)	171 (95.00%)	.783
	Y	447 (5.46%)	438 (5.47%)	9 (5.00%)
In-hospital mortality, %	Y	783 (9.57%)	759 (9.48%)	24 (13.33%)	.197
	N	7390 (90.30%)	7234 (90.38%)	156 (86.67%)

Note:—Hosp indicates Hospital; SNF, Skilled Nursing Facility; ND, Not done/missing.

^aValues are presented as “mean (% of total)” unless otherwise noted.

Table 2:

Outcomes of SARS-CoV-2 positive versus negative patients with AIS undergoing EVT^a

Outcome	Univariate ORs	Multivariate	Multivariate
SARS-CoV-2 positive versus negative EVT patients	(95% CI)	Model 1 ORs (95% CI)	Model 2 ORs (95% CI)
Favorable discharge location (home/rehabilitation vs other)	0.82 (0.616, 1.091), P = .1733	0.66 (0.48, 0.90), P = .0092	0.67 (0.49, 0.93), P = .0146
Good mRS at discharge (0–2)	0.608 (0.054, 6.779), P = .6857	0.17 (0.04, 0.62), P = .0075	0.15 (0.04, 0.60), P = .0071
sICH	0.956 (0.493,1.854), P = .8951	1.00 (0.52, 1.94), P = .9904	0.97 (0.50, 1.88), P = .9321
In-hospital mortality	0.1735 (0.474, 1.128), P = .1569	1.58 (1.02, 2.45), P = .0408	1.65 (1.06, 2.58), P = .0274
Independent ambulation at discharge (restricted to patients who were independently ambulating at presentation)	0.989 (0.664,1.472), P = .9562	0.92 (0.61, 1.39), P = .6899	0.95 (0.62, 1.45), P = .8097

^aMultivariate model 1 adjusts for age, sex, race, insurance, diabetes, atrial fibrillation, arrival mode, IV thrombolysis, and period. Multivariate model 2 adjusts for all variables in Supplemental Data.

Temporal analysis of year-to-year comparisons of outcomes by using 2020 as the index year showed no significant differences in odds of a good mRS at discharge (2020 versus 2021 OR 1.54; 95% CI: 0.89, 2.67 and 2020 versus 2022 OR 1.04; 95% CI: 0.47, 2.28), sICH (2020 versus 2021 OR 0.78; 95% CI: 0.54, 1.11 and 2020 versus 2022 OR 0.87; 95% CI: 0.59, 1.27), and in-hospital mortality (2020 versus 2021 OR 0.89; 95% CI: 0.75,1.05 and 2020 versus 2022 OR 0.92; 95% CI: 0.74,1.14), discharge to home or acute rehab center (2020 versus 2021 OR 0.95; 95% CI: 0.86,1.05 and 2020 versus 2022 OR 1.02; 95% CI: 0.91,1.14) in SARS-CoV-2 positive versus negative patients when comparing the years of 2020 to 2021 and 2020 to 2022 (Table 3).

Table 3:

Temporal analysis of outcomes in SARS-CoV-2 positive patients to examine changes throughout the pandemic (2020–2022)

Outcome	Period	aOR	95% CI
Favorable discharge location (home/rehabilitation vs other)	2020 vs 2021	0.95	(0.86, 1.05)
	2020 vs 2022	1.02	(0.91, 1.14)
Good mRS at discharge (0–2)	2020 vs 2021	1.54	(0.89, 2.67)
	2020 vs 2022	1.04	(0.47, 2.28)
sICH	2020 vs 2021	0.78	(0.54, 1.11)
	2020 vs 2022	0.87	(0.59, 1.27)
In-hospital mortality	2020 vs 2021	0.89	(0.75, 1.05)
	2020 vs 2022	0.92	(0.74, 1.14)

DISCUSSION

Despite younger age and similar timelines of presentation and neurologic severity, when controlling for confounding variables, EVT-treated patients who were SARS-CoV-2 positive had lower odds of achieving favorable discharge outcomes of discharge to home or rehabilitation center, good mRS (0–2) at discharge, and higher odds of in-hospital mortality.

The currently available studies on SARS-CoV-2 status and EVT outcomes for patients with AIS are equivocal about outcomes following EVT.^5-12 Many of the available studies are of smaller cohorts (n < 50), and some are across many countries or regions, leading to the potential for differences in systems of care and treatment guidelines that may not represent true differences in stroke outcomes between SARS-CoV-2 positive and negative patients. Our analysis of Get with the Guidelines–Stroke data across the state of Florida shows that patients with SARS-CoV-2 receiving EVT for LVO strokes are less likely to have a favorable discharge location (home/rehab versus others) and a good mRS (0–2) at discharge, while also having increased risk of in-hospital mortality. Although other groups have posited it that these unfavorable outcomes in patients with SARS-CoV-2 receiving EVT may be the result of the natural course of the SARS-CoV-2 disease, our group believes that there is another prominent factor involved in these differences: changes in systems of care and clinical treatment pathways that were instituted as the SARS-CoV-2 pandemic descended on the world.^17-20 These systemic changes may be reflected in multiple data points in this study, such as the increased time from onset to groin puncture (average 379 versus 291 minutes) shown in Supplemental Data. Increased times were seen in onset to arrival, door to needle, and door to CT time as well. These increases in time may be reflective of the increased time needed to don personal protective equipment, such as respirators, N95 masks, and gowns for both EMS personnel and receiving nurses and physicians in the emergency department and operating rooms. Clinical treatment pathways were also altered as the pandemic began. For example, some institutions require intubation of patients in the emergency department before transportation to the endovascular suite for EVT. These patients also required isolation, especially early in the pandemic, when stopping the spread of the disease was of great importance as no vaccine was available and little was known about the disease. Additionally, some patients may have been more reluctant to seek acute care for stroke symptoms, leading to increased time from symptom onset to groin puncture. This is shown in our data with greater percentages of SARS-CoV-2 patients having onset to puncture times 6–24 hours or >24 hours rather than <6 hours. The reluctance of all people to obtain medical care during the pandemic is well-documented, even among those with acute stroke.²⁰ Fear of spreading the disease and decreased casual interactions that often lead to the discovery of a loved one with a stroke syndrome may have been reasons for these increased times between presentation and treatment in patients with SARS-CoV-2.

The impact of SARS-CoV-2 on the brain has been well-documented, spanning the investigation of cellular endothelial changes in cerebral blood vessels to acute ischemic events in the setting of SARS-CoV-2 predicting worse outcomes in patients with clinical encephalopathy.^21-23 These impacts on brain vasculature and the fact that CNS ischemic events as a cause of acute encephalopathy has been shown to be a predictor of poor outcomes present another explanation for worse outcomes among patients with SARS-CoV-2 undergoing EVT for large vessel AIS.

Another system-of-care factor that may have played a role in the observed differences in outcomes seen here is the admitting service and locations where care was administered. It has been shown that care in a dedicated neurologic unit (neuro ICU, stroke unit) independently improves the outcomes of EVT patients.^2,3 As the pandemic began, patients were being cared for in non-neurologically specific environments by non-neurologic providers to stop the spread of the disease by concentrating SARS-CoV-2 patients in one place, thereby limiting the exposure of other patients and staff. Another explanation for the observed differences in outcomes for SARS-CoV-2 positive patients in this study may be the differences in admitting service/care service; however, we have too large an amount of missing data to answer this question at present.

As this will not likely be the last pandemic, it is relevant and important to keep the SARS-CoV-2/COVID-19 impact on systems of care for AIS in mind. By doing this we can create resiliency in our systems of care to avoid these discrepancies in future outcomes.

Limitations

The retrospective nature of this study is an inherent limitation and potential source of bias. Additionally, significant missing data in some of the variables in the FSR database are a potential source of confounding.

CONCLUSIONS

In this large multicenter stroke registry, we found that despite similar risk factors, neurologic symptom severity, NIHSS scores, and in-hospital treatment timelines, SARS-CoV-2 positive patients were more likely to be younger, be men, have increased onset to arrival time, and have increased onset to puncture time. The major finding of our study was that increased treatment timelines for SARS-CoV-2–positive patients were associated with a lower likelihood of discharge to home or a rehabilitation center, a lower likelihood of mRS <2 at discharge, and a higher likelihood of in-hospital mortality. Temporal assessment showed similar results across the years of 2020, 2021, and 2022. These findings provide novel insight from a large database to add to the emerging literature examining outcomes of concurrent SARS-CoV-2 infection in the setting of EVT for target vessel occlusion AIS. Future studies should examine EVT outcomes by using a larger group of SARS-CoV-2 positive patients, as increased analytic power may provide further insights about the associations reported here.

ABBREVIATIONS:

AIS

acute ischemic stroke

EVT

endovascular thrombectomy

FSR

Florida Stroke Registry

LVO

large vessel occlusion

sICH

symptomatic intracerebral hemorrhage