Adjudicated myocarditis and multisystem illness trajectory in healthcare workers post-COVID-19.

BACKGROUND: We investigated the associations of healthcare worker status with multisystem illness trajectory in hospitalised post-COVID-19 individuals. METHODS AND RESULTS: One hundred and sixty-eight patients were evaluated 28-60 days after the last episode of hospital care. Thirty-six (21%) were healthcare workers. Compared with non-healthcare workers, healthcare workers were of similar age (51.3 (8.7) years vs 55.0 (12.4) years; p=0.09) more often women (26 (72%) vs 48 (38%); p<0.01) and had lower 10-year cardiovascular risk (%) (8.1 (7.9) vs 15.0 (11.5); p<0.01) and Coronavirus Clinical Characterisation Consortium in-hospital mortality risk (7.3 (10.2) vs 12.7 (9.8); p<0.01). Healthcare worker status associated with less acute inflammation (peak C reactive protein 48 mg/L (IQR: 14-165) vs 112 mg/L (52-181)), milder illness reflected by WHO clinical severity score distribution (p=0.04) and shorter duration of admission (4 days (IQR: 2-6) vs 6 days (3-12)).In adjusted multivariate logistic regression analysis, healthcare worker status associated with a binary classification (probable/very likely vs not present/unlikely) of adjudicated myocarditis (OR: 2.99; 95% CI (1.01 to 8.89) by 28-60 days postdischarge).After a mean (SD, range) duration of follow-up after hospital discharge of 450 (88) days (range 290, 627 days), fewer healthcare workers died or were rehospitalised (1 (3%) vs 22 (17%); p=0.038) and secondary care referrals for post-COVID-19 syndrome were common (42%) and similar to non-healthcare workers (38%; p=0.934). CONCLUSION: Healthcare worker status was independently associated with the likelihood of adjudicated myocarditis, despite better antecedent health. Two in five healthcare workers had a secondary care referral for post-COVID-19 syndrome. TRIAL REGISTRATION NUMBER: NCT04403607.

A Multicenter, Scan-Rescan, Human and Machine Learning CMR Study to Test Generalizability and Precision in Imaging Biomarker Analysis.

BACKGROUND: Automated analysis of cardiac structure and function using machine learning (ML) has great potential, but is currently hindered by poor generalizability. Comparison is traditionally against clinicians as a reference, ignoring inherent human inter- and intraobserver error, and ensuring that ML cannot demonstrate superiority. Measuring precision (scan:rescan reproducibility) addresses this. We compared precision of ML and humans using a multicenter, multi-disease, scan:rescan cardiovascular magnetic resonance data set. METHODS: One hundred ten patients (5 disease categories, 5 institutions, 2 scanner manufacturers, and 2 field strengths) underwent scan:rescan cardiovascular magnetic resonance (96% within one week). After identification of the most precise human technique, left ventricular chamber volumes, mass, and ejection fraction were measured by an expert, a trained junior clinician, and a fully automated convolutional neural network trained on 599 independent multicenter disease cases. Scan:rescan coefficient of variation and 1000 bootstrapped 95% CIs were calculated and compared using mixed linear effects models. RESULTS: Clinicians can be confident in detecting a 9% change in left ventricular ejection fraction, with greater than half of coefficient of variation attributable to intraobserver variation. Expert, trained junior, and automated scan:rescan precision were similar (for left ventricular ejection fraction, coefficient of variation 6.1 [5.2%-7.1%], P=0.2581; 8.3 [5.6%-10.3%], P=0.3653; 8.8 [6.1%-11.1%], P=0.8620). Automated analysis was 186× faster than humans (0.07 versus 13 minutes). CONCLUSIONS: Automated ML analysis is faster with similar precision to the most precise human techniques, even when challenged with real-world scan:rescan data. Assessment of multicenter, multi-vendor, multi-field strength scan:rescan data (available at www.thevolumesresource.com) permits a generalizable assessment of ML precision and may facilitate direct translation of ML to clinical practice.

Global longitudinal strain is associated with heart failure outcomes in hypertrophic cardiomyopathy.

OBJECTIVE: We hypothesised that abnormal global longitudinal strain (GLS) would predict outcome in hypertrophic cardiomyopathy (HCM) better than current echocardiographic measures. METHODS: Retrospective analysis of risk markers in relation to outcomes in 472 patients with HCM at a single tertiary institution (2006-2012). Exclusion criteria were left ventricular (LV) hypertrophy of other origin, patients in atrial fibrillation, lost to follow-up and insufficient image quality to perform strain analysis. Standardised echocardiogram recordings were reviewed and standard variables and LV GLS were measured. The primary end-point included all cardiac deaths, appropriate defibrillator shocks and heart failure (HF) admissions. The secondary end-point was death by HF and admissions related to HF. RESULTS: Mean age was 50.0±15.0 years; 322 (68%) were men. At a median of 4.3 years (IQR 0.1-7.8) follow-up, 21 (4.4%) patients experienced cardiovascular death: 6 (1.3%) died from HF, 13 (2.7%) had sudden cardiac death and 2 (0.4%) died secondary to stroke. Four (0.8%) patients experienced appropriate defibrillator shock, and 13 (2.7%) were admitted for HF. On multivariate Fine-Gray proportional hazard analyses, GLS was significantly associated with the primary end-point (HR=0.90, 95% CI 0.83 to 0.98, p=0.018) independently of age, maximal provoked LV outflow-tract gradient and LV end-systolic volume. Moreover, GLS was particularly associated with the secondary end-point (HR=0.82, 95% CI 0.75 to 0.90, p<0.0001) independently of age, previous atrial fibrillation, New York Heart Association (NYHA) class III-IV, LV end-systolic volume, E/E', and outflow-tract gradient. Survival curves confirmed that GLS was associated with HF events (GLS <15.6%, p=0.0035). CONCLUSIONS: In patients with HCM, reduced GLS is an independent factor associated with poor cardiac outcomes, and particularly HF outcomes.