Frontiers in Public Health (Jan 2021)

Decreased CO2 Levels as Indicators of Possible Mechanical Ventilation-Induced Hyperventilation in COVID-19 Patients: A Retrospective Analysis

  • Di Hu,
  • Jinpeng Li,
  • Rongfen Gao,
  • Shipei Wang,
  • Qianqian Li,
  • Sichao Chen,
  • Jianglong Huang,
  • Yihui Huang,
  • Man Li,
  • Wei Long,
  • Zeming Liu,
  • Liang Guo,
  • Xiaohui Wu

DOI
https://doi.org/10.3389/fpubh.2020.596168
Journal volume & issue
Vol. 8

Abstract

Read online

Background: Six months since the outbreak of coronavirus disease (COVID-19), the pandemic continues to grow worldwide, although the outbreak in Wuhan, the worst-hit area, has been controlled. Thus, based on the clinical experience in Wuhan, we hypothesized that there is a relationship between the patient's CO2 levels and prognosis.Methods: COVID-19 patients' information was retrospectively collected from medical records at the Leishenshan Hospital, Wuhan. Logistic and Cox regression analyses were conducted to determine the correlation between decreased CO2 levels and disease severity or mortality risk. The Kaplan-Meier curve analysis was coupled with the log-rank test to understand COVID-19 progression in patients with decreased CO2 levels. Curve fitting was used to confirm the correlation between computed tomography scores and CO2 levels.Results: Cox regression analysis showed that the mortality risk of COVID-19 patients correlated with decreased CO2 levels. The adjusted hazard ratios for decreased CO2 levels in COVID-19 patients were 8.710 [95% confidence interval (CI): 2.773–27.365, P < 0.001], and 4.754 (95% CI: 1.380–16.370, P = 0.013). The adjusted odds ratio was 0.950 (95% CI: 0.431–2.094, P = 0.900). The Kaplan-Meier survival curves demonstrated that patients with decreased CO2 levels had a higher risk of mortality.Conclusions: Decreased CO2 levels increased the mortality risk of COVID-19 patients, which might be caused by hyperventilation during mechanical ventilation. This finding provides important insights for clinical treatment recommendations.

Keywords