Baseline characteristics
A total of 19658 patients were included, meeting sepsis 3.0 from the 69619 patients in ICU. 9660 were excluded from the analysis because of brain injury, mental disorders, drug abuse, alcoholism, metabolic diseases, no FiO2 above 21% or no records of oxygen index, etc. 10055 patients for the final analysis. 6714 were assigned to the SAE group and 3341 were assigned to the non-SAE group (Figure 1).
Table1 describe the characteristics of sepsis patients. The incidence of patients with SAE was 66.8%. Compared with the non-SAE group, patients in the SAE group are more likely to be infected by Gram-negative bacteria(18.3% vs 13.7%, P<0.001), including: Klebsiella(6.8% vs 4.8%, P<0.001), Escherichia coli(12.6% vs 10.2%, P<0.001), (16.7% vs 9.0%, P<0.001).The PaCO2[46(40-52) vs 47(42-52), p<0.001], FiO2[50(40-61) vs 50(40-64), p = 0.009], PaO2[93.5(74-125) vs. 98(79-128), p<0.001], SpO2≤99% (98.9% vs 99.7%, P<0.001), SpO2≤98%(96.5% vs 97.4%, P=0.012), SpO2≤97%(92.9% vs 93.8%, P=0.068) in SAE group were lower than non- SAE group. The SAE group had more patients with SpO2≤93% (57.6% vs 53.6%, P<0.001), SpO2≤92% (45.1% vs 39.2%, P<0.001), SpO2≤91% (33.3% vs 26.3%, P<0.001) than the non-SAE group. Patients in the SAE group had higher SAPS II [35(28-46) vs 30(25-35), p<0.001], SOFA [5(3-7) vs 3(2-5), p<0.001], hospital mortality (7.9% vs 7.5%, P=0.476).
Generalized additive models to estimate the optimal oxygen saturation targets in patients with SAE.
Figure 2 shows an association between the incidence, mortality of SAE, and median SpO2, PaO2. SpO2>97%, SpO2<93%, PO2>307mmHg, PaO2<96mmHg were associated with increased incidence of SAE, and SpO2>96%, SpO2<92%, PaO2>350mmHg, PaO2<96mmHg were associated with increased hospital mortality of SAE. Table 2 and Table 3 confirms them. The generalized additive models demonstrated a nonlinear association between them.
Multivariate logistic analysis of risk factors to incidence in patients with SAE
SPO2(93%-97%) [odds ratio (OR): 0.661, 95% confidence Interval (CI): 0.545-0.803, p<0.001], use of vasopressors (OR:0.705 95%CI:0.642-0.775, p<0.001), renal replacement therapy (OR:0.514, 95%CI:0.389-0.680, p<0.001) were protective factors in the incidence of SAE. SPO2≤92% (OR: 1.647, 95%CI:1.260-2.153, p=0.006), SPO2≤91% (OR:1.218, 95%CI:1.048-1.416, p=0.010), the microbiology type of fungus (OR:1.441, 95%CI:1.243-1.671, p<0.001), SAPS II (OR:1.032, 95%CI:1.027-1.037, p<0.001), SOFA (OR:1.275, 95%CI:1.247-1.304, p<0.001) were risk factors in the occurrence of SAE, after adjusted of microbiology type (Klebslella, Escherichia Coli, Gram-negative bacteria), PaCO2, FiO2, SAPS II, and SOFA.
Univariate and multivariate logistic analysis of risk factors to hospital mortality.
SPO2(92%-96%) (OR:0.572, 95%CI: 0.468-0.699, p<0.001) was a protective factor in the occurrence of SAE. Age (OR:1.012, 95%CI: 1.005-1.019, p<0.001), sex (OR:1.213, 95%CI:1.011-1.455, p=0.038), site of infection of lung(OR:1.932, 95%CI:1.399-2.668, p<0.001) were risk factors in the occurrence of SAE, after adjusted of charlson, site of infection (intestinal, urinary, catheter, skin and soft tissue, abdominal cavity), microbiology type (Klebslella, Escherichia Coli, Pseudomonas aeruginosa, Staphylococcus aureus , Gram-negative bacteria, Fungus), PaCO2, FiO2, use of vasopressors, renal replacement therapy, mechanical ventilation, SAPS II, SOFA, and GCS.