[Clinical study on optimal switching mode in sequential noninvasive-invasive mechanical ventilation for acute...

To explore the switch time of noninvasive-invasive mechanical ventilation sequential treatment for acute exacerbation of chronic obstructive pulmonary disease (AECOPD), and effectively reduce the rate of tracheal intubation.

A retrospective study was performed on patients with AECOPD, who underwent mechanical ventilation in emergency resuscitation room and admitted to department of respiration of Kaifeng Central Hospital Emergency Center from July 2014 to March 2019. The patients who used noninvasive mechanical ventilation (NIV) were included in NIV group (118 cases), and those who used invasive positive pressure ventilation (IPPV) were included in IPPV group (52 cases). The usage of breathing machine time, hospital days and hospital mortality were compared between the two groups. Clinical indicators such as age, gender, body temperature, respiratory rate, body mass index (BMI), mean arterial pressure (MAP), oxygenation index (PaO₂/FiO₂), respiratory index (RI), pH value, D-dimer, hemoglobin (HB), albumin, blood lactate (Lac), brain natriuretic peptide (BNP), C-reactive protein (CRP), procalcitonin (PCT), serum creatinine (SCr), white blood cell count (WBC), Glasgow coma scale (GCS), sputum excretion drainage were collected. The factors influencing the failure of NIV were analyzed by Logistic stepwise regression analysis. The receiver operating characteristic (ROC) curve was used to test the value of the NIV failure risk prediction model.

There was no significant difference in total mechanical ventilation time and hospital mortality between NIV group and IPPV group (hours: 65.6±11.11 vs. 66.9±12.1, 6.8% vs. 9.6%, both P > 0.05), but the hospital time in group NIV was significantly shorter than that in IPPV group (days: 12.3±2.1 vs. 14.2±2.5, P < 0.05). In NIV group, 101 cases completed NIV continuously, 17 cases of NIV failure turned to IPPV, and the failure rate of NIV was 14.4%. There were statistically significant differences in gender, PaO₂/FiO₂, RI, pH value, D-dimer, PCT, WBC, Lac, sputum excretion drainage and GCS score between NIV failure patients and NIV success patients. Logistic regression analysis showed that RI, pH value, WBC and sputum excretion drainage were independent risk factors for NIV failure [RI: odds ratio (OR) = 3.879, 95% confidence interval (95%CI) was 1.258-11.963, P = 0.018; pH value: OR = 3.316, 95%CI was 1.270-8.660, P = 0.014; WBC: OR = 3.684, 95%CI was 1.172-11.581, P = 0.026; sputum excretion drainage: OR = 0.125, 95%CI was 0.042-0.366, P = 0.000]. The NIV failure risk prediction model based on the above independent risk factors had a good goodness of fit (χ ² = 9.02, P = 0.34). ROC curve analysis showed that the NIV failure risk prediction model had a high predictive value for the patients with AECOPD [the area under ROC curve (AUC) was 0.818±0.051, 95%CI was 0.718-0.918, P = 0.000].

If patients with AECOPD have relative contraindications of NIV but still insist on using NIV, further risk stratification of NIV failure is needed. For those with RI, pH value, WBC abnormalities and sputum excretion drainage, the risk of choosing NIV is significantly increased. We need to pay more attention to the change of the condition and switch to IPPV in time to avoid exacerbation of the condition.