ABSTRACT
Background: The clinical heterogeneity of COVID-19 suggests the existence of different phenotypes with prognostic implications. We aimed to analyze comorbidity patterns in critically ill COVID-19 patients and assess their impact on in-hospital outcomes, response to treatment and sequelae. Methods: Multicenter prospective/retrospective observational study in intensive care units of 55 Spanish hospitals. 5866 PCR-confirmed COVID-19 patients had comorbidities recorded at hospital admission; clinical and biological parameters, in-hospital procedures and complications throughout the stay; and, clinical complications, persistent symptoms and sequelae at 3 and 6 months. Findings: Latent class analysis identified 3 phenotypes using training and test subcohorts: low-morbidity (n=3385; 58%), younger and with few comorbidities; high-morbidity (n=2074; 35%), with high comorbid burden; and renal-morbidity (n=407; 7%), with chronic kidney disease (CKD), high comorbidity burden and the worst oxygenation profile. Renal-morbidity and high-morbidity had more in-hospital complications and higher mortality risk than low-morbidity (adjusted HR (95% CI): 1.57 (1.34-1.84) and 1.16 (1.05-1.28), respectively). Corticosteroids, but not tocilizumab, were associated with lower mortality risk (HR (95% CI) 0.76 (0.63-0.93)), especially in renal-morbidity and high-morbidity. Renal-morbidity and high-morbidity showed the worst lung function throughout the follow-up, with renal-morbidity having the highest risk of infectious complications (6%), emergency visits (29%) or hospital readmissions (14%) at 6 months (p<0.01). Interpretation: Comorbidity-based phenotypes were identified and associated with different expression of in-hospital complications, mortality, treatment response, and sequelae, with CKD playing a major role. This could help clinicians in day-to-day decision making including the management of post-discharge COVID-19 sequelae. Funding: ISCIII, UNESPA, CIBERES, FEDER, ESF.
ABSTRACT
BACKGROUND: The identification of factors associated with Intensive Care Unit (ICU) mortality and derived clinical phenotypes in COVID-19 patients could help for a more tailored approach to clinical decision-making that improves prognostic outcomes. METHODS: Prospective, multicenter, observational study of critically ill patients with confirmed COVID-19 disease and acute respiratory failure admitted from 63 ICUs in Spain. The objective was to utilize an unsupervised clustering analysis to derive clinical COVID-19 phenotypes and to analyze patient's factors associated with mortality risk. Patient features including demographics and clinical data at ICU admission were analyzed. Generalized linear models were used to determine ICU morality risk factors. The prognostic models were validated and their performance was measured using accuracy test, sensitivity, specificity and ROC curves. RESULTS: The database included a total of 2022 patients (mean age 64 [IQR 5-71] years, 1423 (70.4%) male, median APACHE II score (13 [IQR 10-17]) and SOFA score (5 [IQR 3-7]) points. The ICU mortality rate was 32.6%. Of the 3 derived phenotypes, the A (mild) phenotype (537; 26.7%) included older age (< 65 years), fewer abnormal laboratory values and less development of complications, B (moderate) phenotype (623, 30.8%) had similar characteristics of A phenotype but were more likely to present shock. The C (severe) phenotype was the most common (857; 42.5%) and was characterized by the interplay of older age (> 65 years), high severity of illness and a higher likelihood of development shock. Crude ICU mortality was 20.3%, 25% and 45.4% for A, B and C phenotype respectively. The ICU mortality risk factors and model performance differed between whole population and phenotype classifications. CONCLUSION: The presented machine learning model identified three clinical phenotypes that significantly correlated with host-response patterns and ICU mortality. Different risk factors across the whole population and clinical phenotypes were observed which may limit the application of a "one-size-fits-all" model in practice.