Prolonged ICU Stay in Severe and Critically-Ill COVID-19 Patients Who Received Convalescent Plasma Therapy

Convalescent plasma administration in severe and critically-ill COVID-19 patients have been proven to not provide improvement in patients' outcome, yet it is still widely used in countries with limited resources due to its high availability and safety. This study aims to investigate its effects on ICU mortality, ICU length of stay (LoS), and improvement of oxygen support requirements.

Data of all severe and critically-ill patients in our COVID-19 ICU was collected retrospectively between May and November 2020. We dichotomized the variables and compared outcome data of 48 patients, who received convalescent plasma to 131 patients, receiving standard of care. Data were analyzed using multiple logistic regression to make prediction models of mortality, length of stay, and oxygen support device requirement.

Overall mortality rate in our COVID-19 ICU was 55.3%, with a median overall length of stay of 8 (4-11) days. Less patients that received convalescent plasma presented with the need for mechanical ventilation on ICU admission (p < 0.001), but with comparable PaO₂ to FiO₂ (P/F) ratio (p=0.95). Factors that confounded mortality were obesity (aOR = 14.1; 95% CI (1.25, 166.7); p=0.032), mechanical ventilation (aOR = 333; 95% CI (4.5,1,000); p < 0.001), higher neutrophil-to-lymphocyte ratio (NLR) (aOR = 7.32; 95% CI (1.82, 29.4); p=0.005), and lower P/F ratio (aOR = 7.70; 95% CI (2.04, 29.4); p=0.003). ICU LoS was longer in patients, who had prior history of hypertension (aOR = 2.14; 95% CI (1.05, 4.35); p=0.036) and received convalescent plasma (aOR = 3.88; 95% CI (1.77, 8.05); p < 0.001). Deceased patients, who received convalescent plasma, stayed longer in the ICU with a mean length of stay of 12.87 ± 5.7 days versus 8.13 ± 4.8 days with a significant difference (U = 434; p < 0.000). The chance of improved oxygen support requirements was lower in obese patients (aOR = 9.18; 95%CI (2.0, 42.1); p < 0.004), mechanically ventilated patients (aOR = 13.15; 95% CI (3.75, 46.09); p < 0.001), patients with higher NLR (aOR = 2.5; 95% CI (1.07, 5.85); p=0.034), and lower P/F ratio (aOR = 2.76; 95% CI (1.1, 6.91); p=0.031).

The length of stay of patients in the convalescent plasma group was significantly longer than the control group. There was no effect of convalescent plasma in ICU mortality and no improvement was observed in terms of oxygen support requirements.