[Predictive factors for failure of non-invasive positive pressure ventilation in immunosuppressed patients with acute respiratory failure]

Abstract

Objective: To evaluate the predictive factors for failure of non-invasive positive pressure ventilation (NIPPV) in immunosuppressed patients with acute respiratory failure (ARF).

Methods: The clinical data of 118 immuno-deficient patients treated with NIPPV in the respiratory and intensive care unit (RICU) of the First Affiliated Hospital of Xinjiang Medical University from January 2012 to August 2017 were retrospectively analyzed. The patients were divided into a non-endotracheal intubation (ETI) group (n = 62) and ETI group (n = 56) according to whether ETI was performed during the hospitalization period or not. Each observed indicator was analyzed by univariate analysis, and factors leading to failure of NIPPV were further analyzed by Logistic regression. Receiver operating characteristic (ROC) curve was plotted to evaluate the predictive value of risk factors for failure of NIPPV in immunosuppressed patients with ARF.

Results: The non-intubation rate for NIPPV in immunosuppressed patients was 50.8% (60/118). Compared with the non-ETI group, the body temperature, pH value in the ETI group were significantly increased, the partial pressure of arterial carbon dioxide (PaCO₂) was significantly decreased, the ratio of oxygenation index (PaO₂/FiO₂) < 100 mmHg (1 mmHg = 0.133 kPa), acute physiology and chronic health evaluation II (APACHE II) score ≥ 20, and the number of cases requiring catecholamine were significantly increased, the mortality was significantly increased. Multivariate Logistic regression analysis showed that the APACHE II score ≥ 20 [odds ratio (OR) = 15.274, 95% confidence internal (95%CI) = 2.175-107.252, χ² = 7.516, P = 0.006], PaO₂/FiO₂ < 100 mmHg (OR = 0.075, 95%CI = 0.014-0.408, χ² = 8.968, P = 0.003), and need for catecholamine (OR = 35.736, 95%CI = 6.974-183.124, χ² = 18.400, P < 0.001) were independent risk factors for failure of NIPPV. ROC curve analysis showed that the APACHE II score ≥ 20 and PaO₂/FiO₂ < 100 mmHg could predict failure of NIPPV, the area under ROC curve (AUC) of the APACHE II score ≥ 20 was 0.787, the sensitivity was 83.93%, the specificity was 69.35%, the positive predict value (PPV) was 71.21%, the negative predict value (NPV) was 82.69%, the positive likelihood ratio (PLR) was 2.74, the negative likelihood ratio (NLR) was 0.23, and Youden index was 0.53; the AUC of PaO₂/FiO₂ < 100 mmHg was 0.757, the sensitivity was 80.65%, the specificity was 66.07%, the PPV was 68.18%, the NPV was 78.85%, the PLR was 2.38, the NLR was 0.29, and Youden index was 0.47.

Conclusions: 50.8% of immunocompromised and ARF patients treated with NIPPV did not require ETI, which is independent of the etiology of ARF. APACHE II score ≥ 20, PaO₂/FiO₂ < 100 mmHg, and the need for catecholamine are predictive factors for failure of NIPPV in immunocompromised patients.