COVID-19 Stroke Apical Lung Examination Study 2: a national prospective CTA biomarker study of the lung apices, in patients presenting with suspected acute stroke (COVID SALES 2)

Abstract

Background: Apical ground-glass opacification (GGO) identified on CT angiography (CTA) performed for suspected acute stroke was developed in 2020 as a coronavirus-disease-2019 (COVID-19) diagnostic and prognostic biomarker in a retrospective study during the first wave of COVID-19.

Objective: To prospectively validate whether GGO on CTA performed for suspected acute stroke is a reliable COVID-19 diagnostic and prognostic biomarker and whether it is reliable for COVID-19 vaccinated patients.

Methods: In this prospective, pragmatic, national, multi-center validation study performed at 13 sites, we captured study data consecutively in patients undergoing CTA for suspected acute stroke from January-March 2021. Demographic and clinical features associated with stroke and COVID-19 were incorporated. The primary outcome was the likelihood of reverse-transcriptase-polymerase-chain-reaction swab-test-confirmed COVID-19 using the GGO biomarker. Secondary outcomes investigated were functional status at discharge and survival analyses at 30 and 90 days. Univariate and multivariable statistical analyses were employed.

Results: CTAs from 1,111 patients were analyzed, with apical GGO identified in 8.5 % during a period of high COVID-19 prevalence. GGO showed good inter-rater reliability (Fleiss κ = 0.77); and high COVID-19 specificity (93.7 %, 91.8-95.2) and negative predictive value (NPV; 97.8 %, 96.5-98.6). In subgroup analysis of vaccinated patients, GGO remained a good diagnostic biomarker (specificity 93.1 %, 89.8-95.5; NPV 99.7 %, 98.3-100.0). Patients with COVID-19 were more likely to have higher stroke score (NIHSS (mean +/- SD) 6.9 +/- 6.9, COVID-19 negative, 9.7 +/- 9.0, COVID-19 positive; p = 0.01), carotid occlusions (6.2 % negative, 14.9 % positive; p = 0.02), and larger infarcts on presentation CT (ASPECTS 9.4 +/- 1.5, COVID-19 negative, 8.6 +/- 2.4, COVID-19 positive; p = 0.00). After multivariable logistic regression, GGO (odds ratio 15.7, 6.2-40.1), myalgia (8.9, 2.1-38.2) and higher core body temperature (1.9, 1.1-3.2) were independent COVID-19 predictors. GGO was associated with worse functional outcome on discharge and worse survival after univariate analysis. However, after adjustment for factors including stroke severity, GGO was not independently predictive of functional outcome or mortality.

Conclusion: Apical GGO on CTA performed for patients with suspected acute stroke is a reliable diagnostic biomarker for COVID-19, which in combination with clinical features may be useful in COVID-19 triage.

Keywords: COVID-19; Diagnostic biomarker; Lung apices; Stroke.

Fig. 1

GGO in the lung apices of a patient with COVID-19.

Fig. 2

(a) Flow diagram demonstrating patient selection for inclusion of patients with CTA in acute suspected stroke stratified by RT-PCR result and the absence or presence of GGO. (b) Flow diagram demonstrating patient selection for inclusion of patients with CTA in acute suspected stroke, stratified by the absence or presence of GGO alone. CTA = CT Angiography. GGO = Ground-Glass Opacification. RT-PCR = Reverse Transcriptase Polymerase Chain Reaction swab test.

Fig. 3

Grotta plots demonstrating functional outcomes at discharge stratified (a) by RT-PCR swab result and (b) in the absence or presence of GGO. The plots demonstrate a shift to dependency on discharge (mRS > 2) in patients with a positive RT-PCR result and in the presence of GGO. GGO = ground-glass opacification. RT-PCR = reverse transcriptase polymerase chain reaction swab test. mRS = modified Rankin Scale.

Fig. 4

Kaplan Meier survival curves demonstrating survival for patients in the absence and presence of GGO at 30 days (a) and 90 days (b).