# Predicting the Need for Advanced Respiratory Support in COVID-19 Patients During the Initial Pandemic Phase: A Retrospective Analysis

## TL;DR

This study identifies key factors that predict which COVID-19 patients will need advanced respiratory support early in the pandemic, helping with timely care and resource planning.

## Contribution

The study introduces a predictive model using clinical and laboratory markers to identify high-risk COVID-19 patients requiring advanced respiratory support.

## Key findings

- Age, CRP levels, ferritin levels, and SIRI were significant predictors of the need for advanced respiratory support.
- The SIRI ROC curve had an AUC of 0.785, indicating moderate predictive accuracy.
- Early identification of high-risk patients can improve resource allocation during pandemics.

## Abstract

Background: The coronavirus disease 2019 (COVID-19) pandemic, caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), led to high morbidity and mortality rates worldwide. It is known that some patients, initially hospitalized in general wards, deteriorate over time and require advanced respiratory support (ARS). This study aimed to identify key risk factors predicting the need for ARS in patients during the pandemic's early months.

Methodology: In this retrospective study, we included patients admitted within the first three months of the pandemic who were diagnosed with COVID-19 via reverse transcription polymerase chain reaction (RT-PCR). The patients who required ARS or invasive mechanical ventilation at admission were excluded. Data on demographics, comorbidities, symptoms, vital signs, and laboratory parameters were collected. Statistical analyses, including multivariate logistic regression and receiver operating characteristic (ROC) curve analysis, were performed to identify independent predictors of ARS and determine the cut-off point.

Results: Among 162 patients, 32.1% required ARS. Key differences between ARS and non-ARS groups included age, body mass index (BMI), coronary artery disease prevalence, neutrophil count, C-reactive protein (CRP), ferritin, D-dimer, troponin T levels, neutrophil-to-lymphocyte ratio (NLR), systemic immune-inflammation response index (SIRI), and symptom-to-admission time. Multivariate analysis revealed that age, elevated CRP levels, elevated ferritin levels, and SIRI were significant predictors for ARS. The ROC curve for SIRI showed an area under the curve (AUC) of 0.785, with a cut-off value of 1.915.

Conclusions: Age, CRP levels, ferritin levels, and SIRI are crucial predictors of the need for ARS in COVID-19 patients. The early identification of high-risk patients is essential for timely interventions and resource optimization, particularly during the early stages of pandemics. These insights may assist in optimizing strategies for future respiratory health crisis management.