Introduction
The clinical characteristics and treatment received by patients hospitalized with COVID-19 have changed over time. The objective was to analyze the clinicaltherapeutic evolution of patients in the epidemic waves and estimate a predictive model for mortality.
Methods
Retrospective cross-sectional study considering patients admitted with confirmed SARS-CoV-2 infection until March 2022. Sociodemographic variables, comorbidities and treatments were collected and a predictive model for mortality was created using multivariate logistic regression.
Results
1,784 patients were included. Significant differences were found between the epidemic waves with respect to age, sex, arterial hypertension, diabetes mellitus, obesity and chronic kidney disease. Ceftriaxone, azithromycin, hydroxychloroquine, methylprednisolone and lopinavir-ritonavir were the most frequently used drugs in the first wave. Amoxicillin, dexamethasone and tocilizumab were prescribed more frequently in successive waves. The percentage of deaths varied from 5.6% in the fourth wave to 14.1% in the third (p < 0.001). The resulting factors associated with mortality (OR; 95% CI) were ICU admission (56.5; 27.4-121), age (1.09; 1.08-1.11), days of admission (0.98; 0.96-0.99), chronic kidney disease (1.67; 1.16-2.40) and having received treatment with tocilizumab (2.49; 1.43-4.30), dexamethasone (1.58; 1.10-2.26) and methylprednisolone (2.46; 1.63-3.68). The area under the curve achieved by the model was 0.863.
Conclusion
There are significant clinical-therapeutic differences in patients along the first six epidemic waves. Knowledge of mortality risk factors will allow the detection of hospitalized patients at higher risk and early optimization of their therapeutic management.