Long-term impacts of COVID-19 on systemic inflammation and control of breathing reflexes: an observational cohort study

Figures

• 

  Fig. 1. Representative raw data traces from rebreathing chemoreflex tests. Raw data for minute ventilation (V̇E), pulse oximetry (SpO2), as well as partial pressures of oxygen (PO2) and carbon dioxide (PCO2) measured at the mouth throughout two rebreathing tests in hyperoxic (target 30% inspired O2) (left) and hypoxic (target end-tidal PCO2 of 50 mmHg) (right) conditions. Recordings were taken from the same individual

• 

  Fig. 2. Rebreathing chemoreflex plots. Minute ventilation (V̇E) plotted as a function of end-tidal PCO2 under hyperoxic (target 30% inspired O2) (top) and hypoxic (target end-tidal PCO2 of 50 mmHg) (bottom) test conditions in the same individual. The results of the estimated VRT point via mcp bootstrap analyses are included as blue density plots on the x axis. Individual data points represent mean minute ventilation subsampled from a running average channel at about 10 samples per minute

• 

  Fig. 3. Principal component analysis of inflammatory profiles across control and recovered groups. A PCA plot with individual participants represented by points, separated by group (control = blue, recovered = orange). B PCA plot with strength and directionality of effect for each inflammatory marker

• 

  Fig. 4. Inflammatory marker expression following COVID-19 recovery. Expression of Cystatin C (A), IGFBP-4 (B), MMP-2 (C), OPN (D), Myoglobin (E), NGAL/Lipocalin-2 (F), and IL-10 (G) in control versus all recovered participants (left panels) and at discrete timepoints following recovery from COVID-19 compared to healthy controls (right panels). Tukey style box plots represent medians with boxes representing the first and third quartiles, and whiskers representing the largest value no further than 1.5 * interquartile range (IQR). Significant differences across groups based on post-hoc pairwise t-tests are provided. Versions with individual data points are provided in the supplemental figures

• 

  Fig. 5. VRT, HCVR, and HVR levels following recovery from COVID-19. Measurements of VRT (A), HCVR (B), HVR taken at ETPCO2 50 mmHg (C), HVR taken at ETPCO2 55 mmHg (D) HVR taken at 3 points above VRT (E) all at discrete timepoints following recovery from COVID-19 compared to healthy controls. Error bars represent 95% confidence intervals. Significant differences across groups based on post-hoc pairwise t-tests are provided

• 

  Fig. 6. Correlation analysis of candidate ventilatory control parameters and inflammatory marker expression. A A matrix visualizing Spearman rank correlations of all inflammatory markers versus multiple ventilatory control parameters. Color represents the directionality and strength of the correlation coefficient. Asterisks represent significant correlations after family-wise multiple comparisons corrections. B Raw data plots representing correlations between candidate inflammatory markers SAA and CRP with each HVR measurement. Data are presented on a log10 scale, with point shapes and colors representing measures within control and recovered groups. A linear fit is included with standard error indicated by the shaded region

• 

  Fig. 7. Within subject comparison of VRT and HCVR. Within subject comparisons of (A), VRT and (B), HCVR show differences in the control of breathing parameters pre- and post-COVID infection. Comparisons are made within both hyperoxic and hypoxic oxygen tensions