Clinical scoring for the failure of non-invasive ventilation in acute respiratory failure 1 patients in the Emergency Department.

5 Background Non-invasive mechanical ventilation (NIV) has become an alternative to an 6 invasive artificial airway for the management of acute respiratory failure (ARF). NIV failure 7 causes delayed intubation, which eventually has been associated with increased morbidity 8 and mortality. This study aimed to develop the clinical scoring system of NIV failure in ARF 9 patients. 10 Methods This study was a diagnostic, retrospectively cross-sectional, and exploratory model 11 at the Emergency Medicine Department in Ramathibodi Hospital between February 2017 and 12 December 2017. We included all of the acute respiratory failure patients aged >18 years and 13 received non-invasive ventilation (NIV). Clinical factors associated with NIV failure were 14 recorded. The predictive model and prediction score for NIV failure were developed by 15 multivariable logistic regression analysis. 16 Result A total of 329 acute respiratory failure patients have received NIV success (N = 237) 17 and failure (N = 92). This study showed that NIV failure was associated with heart rate > 110 18 bpm, systolic BP < 110 mmHg, SpO2 < 90 %, arterial pH < 7.30 and serum lactate. The 19 clinical scores were classified into three groups: low, moderate, and high. 20 Conclusion We suggested that the novel clinical scoring of the NIV failure in this study may 21 use as a good predictor for NIV failure in the emergency room. 22 Using combination of 5 variables including heart rate > 110 bpm, systolic BP < 110 1 mmHg, SpO2 < 90 %, arterial pH < 7.30 and serum lactate. The clinical scores were classified 2 into three groups: low, moderate, and high.

The study variables were recorded for all eligible patients, including the baseline characteristic 1 factor and potential clinical factors for NIV failure. Clinical factors included gender, age, vital 2 signs at ED arrival (respiratory rate, heart rate, systolic blood pressure, oxygen saturation, body 3 temperature), glasgow coma scale, diagnosis, underlying disease, laboratory test, arterial blood 4 gas, vasoactive agents and qSOFA score. Data were analyzed using STATA version 14.0. All study variables were compared 20 between the NIV success (did not receive intubation) and the NIV failure (receive intubation) 21 groups by using exact probability test for categorical study variables, and T-test in continuous 22 study variables. The predictive power of each variable was calculated using univariable logistic 23 regression and presented as the area under the receiver operating characteristic (AuROC) curve 24 with 95% confidence intervals (CIs). The potential predictors were categorized into three levels 25 by multivariable logistic regression. Regression coefficients for each level of each clinical 26 predictor were divided by the smallest coefficient of the model and rounded to the nearest 27 0 or 0.5, resulting in a scoring scheme. Discrimination of the prediction scores was presented 28 as the AuROC curve, and 95% CIs for the clinical scoring of failure on acute respiratory failure 29 patients received NIV. Calibration of the prediction was presented using the Hosmer-

30
Lemeshow goodness-of-fit test. The number of reports and percentages of each group were 31 presented with the positive likelihood ratio, 95% CIs, and p-value.

5
Due to the lack of the best clinical scoring, this study assessed the novel clinical scoring 6 of the NIV failure in acute respiratory failure patients at the emergency department. This study 7 showed that NIV failure was associated with heart rate > 110 bpm, systolic BP < 110 mmHg, 8 SpO2 < 90 %, arterial pH < 7.30 and serum lactate. The clinical scores were classified into 9 three groups: low, moderate, and high. We suggested that the low group increased the chance 10 of successful NIV in ARF.

11
In this study, serum lactate levels (> 8 mmol/L) was the most relevant variable for 12 predicting NIV failure, with a maximal score of 4 points. Patients with a high lactate level are 13 strongly correlated with increased mortality in various conditions (14-16). Arterial pH (< 7.30) 14 was the second most relevant variable, with a maximal score of 3 points, followed by systolic 15 BP (< 110 mmHg) with a maximal score of 2 points. The pH level, an indicator of the severity 16 of hypercapnia, has been documented as an important predictor to assess NIV success (17). 17 Previous studies have been clearly reported that a lower baseline pH is a risk factor for 18 NIV failure in certain conditions, especially in COPD patients (18,19). COPD patients with mild 19 to moderate acidosis showed that NIV improved patient outcomes exclusively, the baseline 20 pH was ≥7.30 (20). In a certain study, systolic BP < 90 mmHg is considered as a relative 21 contraindication to NIV (21). Heart rate (> 110 bpm) and SpO2 (< 90 %) were less relevant, 22 with a maximal score of 1 point. Generally, a heart rate <110 bpm has been suggested as an 23 indicator to withdrawn NIV in ARF (10). SpO2, an arterial oxygen saturation measured by pulse 24 oximetry, targets should be 88-92%, in hypercapnic ARF patients treated with NIV. A previous 25 study demonstrated that no single variable could predict NIV failure well. On the other hand, 26 a combination of several variables may increase predictive accuracy (13).

27
There are limitations to this study. First, this study was retrospective data collection 28 and conducted in a single center. We now need to validate our results externally to establish 29 the actual value of our risk score for management decisions. Using combination of 5 variables including heart rate > 110 bpm, systolic BP < 110 1 mmHg, SpO2 < 90 %, arterial pH < 7.30 and serum lactate. The clinical scores were classified 2 into three groups: low, moderate, and high.  14 Consent for publication 15 Not applicable. 17 The datasets analysed during the current study are not publicly available due to privacy issues 18 but are available from the corresponding author upon reasonable request.

19
Competing interests 20 The authors declare that they have no competing interests.   24 WL and TT designed this study and protocol development. TT and CJ were responsible for 25 data collection. CY and CJ were responsible for data analysis. WL and TT wrote the manuscript.

Authors' contributions
TT and PN provided the final approval for this version to be published. CY and WL agreed to 1 be accountable for all aspects of the work. All authors read and approved the final manuscript.         in acute respiratory failure patients.