Oxygen Saturation As A Predictor Of Mortality In Hospitalized Adult ...

Results

A total of 373 medical records of adult patients with COVID-19 admitted at HCH were identified; four patients were excluded as they were transferred or left AMA (Fig 1). Of the 369 patients included in the analysis, 241 (65.31%) were male and the median age was 59 years (IQR: 49–68). Most patients (68.56%) had at least one comorbidity: obesity (42.55%), diabetes mellitus (21.95%) and hypertension (21.68%). A total of 278 (75.34%) were confirmed COVID-19 cases, 250 (89.93%) by a positive serological RDT, 16 (5.75%) by a positive RT-PCR, and 12 (4.32%) by both methods. The median duration of symptoms prior to hospitalization was 7 days (IQR: 5–10) (Table 1).

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Fig 1. Flowchart of adult hospitalized patients included in the cohort and their outcome.

https://doi.org/10.1371/journal.pone.0244171.g001

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Table 1. Characteristics of adult hospitalized patients with COVID-19 at hospital admission.

https://doi.org/10.1371/journal.pone.0244171.t001

Respiratory and heart rates were elevated on admission, with a median of 28 breaths and 108 beats per minute (IQR: 25–32 and 96–118, respectively). Tachypnea was present in 91.27%. Additionally, SaO2 on admission was generally low with a median of 87% (IQR: 77–92), and more than a quarter of the patients (28.46%) were admitted with SaO2 below 80%. The leukocyte count was elevated (11,565 cells/mm3, IQR: 8,260–15,270), and 76.47% of patients had leukocytosis with associated relative lymphopenia (<10% of the leukocyte count). Laboratory parameters of severe COVID-19, such as C-reactive protein (median 96 mg/dL, IQR: 48–192) and lactate dehydrogenase (median 469.5 U/L, IQR: 356–658) were also elevated. Most patients (72.63%) required high-flow oxygen therapy (FiO2 ≥0.36); which included 226 (61.25%) that required use of a non-rebreather (NRB) mask at 15 L/min. Only 10.18% of patients on a NRB mask entered the ICU. Additionally, 67.21% of patients developed ARDS by SaO2/FiO2 ratio during hospitalization. The median hospital stay before discharge was 8 days (IQR: 5–12), with death generally occurring on day 5 of hospitalization (IQR: 2–8). For discharged patients, the median length of stay was inversely proportional to SaO2 values on admission: 7 days (IQR: 4–10) for SaO2 ≥90%, 9.5 days (IQR: 6–13.5) for SaO2 89–85%, 10.5 days (IQR: 8–14) for SaO2 84–80%, and 22 days (IQR: 17.5–22.5) for SaO2 <80% (p <0.001).

One half of patients (183, 49.59%) died during hospitalization. Significant differences were observed in the proportion of patients aged ≥60 years and with comorbidities, mainly obesity and hypertension, with higher frequencies in the deceased group. The SaO2 value on admission was lower in deceased patients compared to discharged patients (78% vs. 91%, p <0.001). This difference was maintained for each SatO2 category and was also constant between different age groups (Fig 2). The proportion of ARDS on admission or developed during hospitalization was greater among deceased participants. Lower absolute lymphocyte counts, higher white blood cell counts, and higher levels of C-reactive protein and lactate dehydrogenase were found among deceased participants.

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Fig 2. Box plot comparing oxygen saturation medians by age group and by outcome.

https://doi.org/10.1371/journal.pone.0244171.g002

Predictive factors of in-hospital mortality were evaluated using Cox proportional hazards regression. Bivariate analysis found that all SaO2 categories <90% on admission were associated with a hazard mortality from 1.93 (95%CI: 1.07–3.49) to 9.13 (95%CI: 5.50–15.14) times higher compared to SaO2 ≥90%. Survival analysis based on SaO2 categories was plotted using Kaplan-Meier curves (Fig 3). In the bivariate analysis, age ≥60 years (cHR 2.83; 95%CI: 2.03–3.93) and hypertension (cHR 1.63; 95%CI: 1.18–2.26) were also associated with in-hospital mortality. In the multivariate analysis, after adjusting for male sex, age ≥60 years and comorbidities (obesity, diabetes, and hypertension), SaO2 values of less than 90% correlated independently with in-hospital mortality, presenting 1.86 (95%CI: 1.02–3.39), 4.44 (95%CI: 2.46–8.02) and 7.74 (95%CI: 4.54–13.19) times greater risk of death for SaO2 of 89–85%, 84–80% and <80%, respectively. Likewise, age ≥60 years was independently associated with in-hospital mortality (aHR 1.88; 95%CI: 1.32–2.69) (Table 2). After secondary analysis, tachypnea was associated with in-hospital mortality both in the bivariate analysis (cHR 4.82; 95%CI: 1.54–15.12) and in the multivariate model when replacing SaO2 categories (aHR 3.99; 95%CI: 1.27–12.54).

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Fig 3. Kaplan-Meier survival curve by different oxygen saturation category at admission.

https://doi.org/10.1371/journal.pone.0244171.g003

Download:

• PPTPowerPoint slide
• PNGlarger image
• TIFForiginal image

Table 2. Baseline predictor factors of mortality in adult hospitalized patients with COVID-19.

https://doi.org/10.1371/journal.pone.0244171.t002