# Challenges in Predicting Intubation in Patients With Severe Coronavirus Disease 2019 (COVID-19) on High-Flow Nasal Cannula: A Focus on Lymphopenia and Conventional Markers

## TL;DR

This study finds that low lymphocyte levels at admission are linked to a higher risk of intubation in severe COVID-19 patients on high-flow nasal cannula, but the predictive power is limited.

## Contribution

The study identifies lymphopenia as an independent predictor of intubation in severe COVID-19 patients on HFNC, despite modest discriminative performance.

## Key findings

- Lower lymphocyte counts and percentages were significantly associated with intubation in severe COVID-19 patients.
- Multivariate analysis confirmed lymphocyte count as an independent indicator of intubation risk.
- Lymphopenia showed modest discriminative power when used alone for predicting intubation.

## Abstract

Background

Intubation risk assessment in severe coronavirus disease 2019 (COVID-19) remains a key clinical challenge, especially for patients treated with high-flow nasal cannula (HFNC). We aimed to clarify whether lymphopenia at admission is associated with the subsequent need for intubation.

Methods

This retrospective study included all consecutive COVID-19 patients admitted to Saitama Red Cross Hospital in Saitama, Japan, from July to September 2021, who required HFNC therapy at presentation. Intubation was performed if the saturation of peripheral oxygen (SpO₂) remained below 90% despite HFNC with a fraction of inspired oxygen (FiO₂) of 80%. We compared clinical and laboratory characteristics between patients who required intubation and those successfully managed on HFNC to identify prognostic factors present at the time of HFNC initiation.

Results

Out of 49 analyzed patients, 15 required intubation during hospitalization and 34 did not. Age and body mass index (BMI) were similar between groups. The Mann-Whitney U tests showed significantly lower lymphocyte counts (intubation: 739.1/μL (interquartile range (IQR) 516.3) vs. non-intubation: 909.0/μL (IQR 423.4); p=0.034), lower lymphocyte percentage (8.0% (IQR 13.0) vs. 15.6% (IQR 9.6); p=0.037), and longer hospital stays (17 days (IQR 18) vs. 14 days (IQR 10); p=0.007) in the intubation group. Multivariate stepwise logistic regression identified lower lymphocyte count (OR 0.997; 95% CI: 0.995-1.000; p=0.048) and longer hospital stay (OR 1.09; 95% CI: 1.01-1.20; p=0.036) as independent indicators of intubation.

Conclusion

In severe COVID-19 patients treated with HFNC, admission lymphopenia was independently associated with subsequent intubation; however, its discriminative performance as a prognostic marker was modest. Lymphocyte counts may aid prognostic risk stratification when interpreted cautiously alongside conventional clinical markers and oxygenation indices, particularly in the context of current and future respiratory viral outbreaks.