1 Introduction
Noninvasive ventilation (NIV) can be defined as a ventilation modality that supports breathing without the need for intubation or surgical airway. NIV is a popular method of adult respiratory management in both the emergency department and the intensive care unit (ICU), and it has gained increasing support in the care of pediatric patients. Besides avoiding the adverse effects of invasive ventilation, NIV has the added advantage of patient comfort. NIV delivers mechanically assisted breaths without the placement of an artificial airway and has become an important mechanism of ventilator support both inside and outside the ICU.1–3
However recent studies showed that patient who failed NIV, had worse prognosis if intubation was delayed.2,4–8 9
Chest sonography has emerged in recent years as a very promising technique for the high sensibility it has shown in the detection of different lung and pleural pathological states.10,11
2 Aim of work
We hypothesized that lung ultrasound could predict early NIV failure within first 24 hours of NIV initiation.
3 Methods
The study was a prospective observational study on adult patients who were admitted to Critical Care Medicine Department of Cairo university hospital over a period of 7 months starting from 1/6/2019 to 31/12/2019. The ethics committee and institutional review board approved this study with protocol serial number MS-180-2019.
Our study was carried out on 50 patients (Age ≥ 18 years old) who were admitted to ICU with respiratory distress and were indicated for NIV.
3.1 Inclusion criteria
All ICU patients who presented with acute respiratory failure and were indicated for NIV (including COPD with exacerbation, Pneumonia with no or mild secretions, acute lung injury manifested by PaO2 /FiO2 < 200, acute congestive heart failure with pulmonary edema).
3.2 Exclusion criteria
All ICU patient who presented with respiratory failure and NIV was contraindicated including (respiratory arrest or unstable cardiorespiratory status, uncooperative patient, inability to protect airway [impaired swallowing and cough], Trauma or burns involving the face, Facial, esophageal or gastric surgery, Apnea [poor airway drive], Reduced consciousness, Air leak syndrome, extreme anxiety, morbid obesity, copious secretions, need for continuous or nearly continuous ventilation assistance, lack of respiratory drive, and disease with air trapping such as asthma.
All patients were subjected to detailed history taking, full physical examination, laboratory investigations, lactate, Procalcitonin, Pro-Beta Natriuretic Peptide, and C-reactive protein serial measurements (upon NIV initiation and 12 hours later). Severity scores were documented (CURB65, APACHE II, and SOFA). Lactate, Pro-BNP, HACOR score were recorded 12 hours post-NIV initiation to document temporal changes. The total HACOR score is 25 and consists of (HR: 0–1, Acidosis: 0–4, Consciousness: GCS 0–10, Oxygenation; 0–6, and RR: 0–4). HACOR ranged from 0 to 25.
Lung ultrasound was carried out serially 1) upon NIV initiation 2) after 12 hours from application of NIV. Measurements were recorded upon NIV initiation and after 12 hours of NIV application. Lung ultrasound was performed by an experienced operator. Scanning was done to 6 lung zones bilaterally anterior, lateral, and posterior lung zones and each of them was divided into upper and lower zones. Four ultrasound aeration patterns were defined: normal aeration (score 0), moderate loss of lung aeration (score 1), severe loss of lung aeration Score 2), and lung consolidation (score 3). Summation of all scores resulted into lung ultrasound score (LUS), which ranged 0–36.
Early failure of NIV was recognized as worsening pH and arterial pressure of carbon dioxide (paCO2 ), tachypnea (>30 breaths/min), hemodynamic instability, pulse oximeter oxygen saturation (SpO2 ), decreased level of consciousness, inability to clear secretions or inability to tolerate interface within first 24 hours of NIV initiation. Patient who failed to respond to NIV were switched to invasive ventilation.
3.2.1 Statistical analysis
Numerical data was described as mean +/− SD. Categorical data was described as percentage (proportions). Normality of data was checked using the Kolmogorov–Smirnov normality test. Comparison between numerical data was done using Chi Square test. Correlations was plotted if deemed appropriate using Pearson or Spearman correlations. Statistical analysis was done using SPSS version 21.
Sample size was calculated based on the following formula:
N = 2∗P (1 − P )/e∗prevalence
where N is the number required. Z is the level of confidence where z = 1.96 at 5% confidence level. P is the sensitivity we are trying to achieve, and it was settled at 80% and e is the level of margin of error which was set at 10% and incidence of NIV failure was estimated at 15%. N was calculated 41 patients, so we decided to recruit 50 patients to compensate for potential drop out.
4 Results
During the study period, 50 patients received NIV therapy. Mean age of recruited patients was 59 ± 13 years old. Males comprised 58% (29). Patients were categorized into two group according to tolerability of NIV trial: 20 (40%) patients showed good response to NIV, while 30 (60%) patients failed to respond to NIV and were intubated and mechanically ventilated accordingly. The age of the studied group 59 ± 13 years old (21 patients female). Demographics, risk factors, and causes of respiratory failure , severity scores, and outcome are demonstrated in Table 1 .
Table 1 -
Demographics, risk factors, and causes of
respiratory failure , Severity scores and outcome in NIV success versus NIV failure group
NIV success
NIV failure
P
Demographic data
Age
56.2 ± 14.2
61.2 ± 12.0
.180
Male
12 (41.4%)
17 (58.6%)
.525
Risk factors
Smoking
4 (20.0%)
9 (30.0%)
.001
Chronic pulmonary disease
10 (50.0%)
14 (46.7%)
.523
Cardiac disease
7 (35%)
16 (53.3%)
.163
Renal
9 (45%)
12 (40%)
.475
Severity scores
CURB 65
1.4 ± 0.7
1.8 ± 1.1
.249
APACHE II
14.8 ± 5.0
17.9 ± 5.5
.052
SOFA
10.1 ± 2.6
9.1 ± 2.9
.231
Outcome
ICU stay
4.3 ± 1.7
9.4 ± 5.6
<.001
NIV stay hours
8.3 ± 5.2
6.2 ± 9.3
.384
The most common causes of NIV usage were pneumonia (34%), followed by acute exacerbation of COPD 19 (24%) and acute pulmonary edema (20%). Failed NIV was most prominent in 77% of patients who had pneumonia. Failed NIV presented in 67% of patients who had acute exacerbation of COPD. Also, failed NIV was present in 30% of patients who had acute pulmonary edema. There were no significant differences between both groups, regarding clinical severity scores.
Failed NIV trial increased ICU stay (5.1, CI 95% 2.5–7.7, P < .001). Number of patients who required vasopressors during first 24 hours was 22 patients (44%).
Laboratory measurements and clinical scores were demonstrated upon NIV initiation and 12 hours later in Table 2 . Comparison between those who succeeded and those who failed NIV was depicted in Figure 1 .
Table 2 -
Comparison between NIV success and NIV failure group, regarding laboratory values, clinical scores and radiological scoring
NIV success
NIV failure
P
Readings upon initiation of NIV
PH
7.38 ± 0.06
7.34 ± 0.08
.032
Lactate
2.4 ± 0.9
3.6 ± 1.7
.005
PCT
0.9 ± 0.7
1.5 ± 0.6
.004
Pro-BNP (ng/mL)
742.5 ± 743.6
380.0 ± 427.6
.053
Heart rate
122.3 ± 14.9
111.7 ± 11.1
.006
Mean arterial pressure
101.6 ± 17.7
86.5 ± 15.8
.003
Glasgow Coma scale
14.9 ± 0.3
14.2 ± 1.3
.010
HACOR
6.7 ± 1.7
8.1 ± 2.7
.046
Lung ultrasound score
16.4 ± 2.9
20.0 ± 3.6
<.001
Readings 12 h post-NIV initiation
PH
7.38 ± 0.05
7.30 ± 0.07
<.001
Pro-BNP (ng/mL)
460.0 ± 357.8
368.3 ± 601.9
.544
Lactate
1.6 ± 0.6
2.7 ± 1.3
<.001
Heart rate
108.2 ± 16.0
119.4 ± 15.7
.017
Mean arterial pressure
95.9 ± 13.2
83.2 ± 14.7
.003
Glasgow Coma Score
14.9 ± 0.3
14.1 ± 1.5
.005
HACOR
4.5 ± 2.5
10.4 ± 4.8
<.001
Lung ultrasound score
13.0 ± 3.1
19.8 ± 4.3
<.001
Figure 1: Comparisons between those who succeeded (white) and those who failed (black), concerning their lactate, Pro-BNP, HACOR, and LUS score.
In patients who were deemed NIV success, there were significant decremental changes in lactate measurements (P .001), Pro-BNP (P .007), HACOR score: (P .001), and lung ultrasound score (P < .001). While in patients who failed NIV, lactate measurements decreased significantly (P < .001). This was accompanied with significant incremental changes in HACOR score (P .001), and lung ultrasound score (P < .001). Changes in Pro-BNP were insignificant.
Receiver Operating Curve analysis to predict NIV failure for Lactate, HACOR, and LUS upon initiation of NIV and 12 hours later were presented in Tables 3 and 4 . Receiver Operating Curve was plotted for the Lactate level, HACOR score, and LUS score (upon NIV initiation and 12 hours later) to predict NIV failure as shown in Figures 2 and 3 .
Table 3 -
Receiver Operating Curve to predict NIV failure, using lactate, HACOR, and LUS scores upon initiation of NIV
AUC∗
Cut-off
Sensitivity (%)
Specificity (%)
Lactate
0.731
2.6
77.0
75.0
HACOR
0.675
8.0
63.0
70.0
Lung ultrasound score
0.774
18.0
77.0
60.0
Table 4 -
Receiver Operating Curve to predict NIV failure, using lactate, HACOR, and LUS scores 12 h post-initiation of NIV
AUC
Cut-off
Sensitivity (%)
Specificity (%)
Lactate
0.782
2.0
73.0
65.0
HACOR
0.879
6.0
83.0
80.0
Lung ultrasound score
0.891
15
87.0
75.0
Figure 2: Receiver Operating Curve to predict was plotted for the Lactate level, HACOR score, and LUS score, upon initiation of NIV, to predict NIV failure. Lactate was plotted in dotted line. Lung ultrasound was plotted in solid line. HOCAR was plotted in stripped line.
Figure 3: Receiver Operating Curve to predict was plotted for the Lactate level, HACOR score, and LUS score, 12 hours, following NIV initiation, to predict NIV failure. Lactate was plotted in dotted line. Lung ultrasound was plotted in solid line. HOCAR was plotted in stripped line.
Patients were categorized according to their HACOR score and LUS score into two categories (guided by ROC results): group (1) HACOR ≥ 8 combined with LUS ≥ 18, and group (2) either HACOR < 8 or LUS < 18. Time to event (NIV failure) was plotted in Figure 4 . Cox regression showed that a score of HACOR ≥ 8 combined with LUS ≥ 18 increased odds of NIV failure 2.1 times, (OR 2.1, CI 95% 1.0–4.6, P .049).
Figure 4: Kaplan–Meier analysis showing time to event (NIV failure). Patients were categorized according to their HACOR score and LUS score into two categories (guided by ROC results): group (1) HACOR ≥ 8 combined with LUS ≥ 18, and group (2) either HACOR < 8 or LUS < 18. Group (1) was plotted in stripped line. Group (2) was plotted in solid line.
5 Discussion
We tried in our study to probe clinical and radiological predictors of early NIV failure. Patients were categorized into two group according to NIV trial response. Our result showed that 20 patients responded well to NIV trial, and 30 patients failed to respond to NIV trial.
In our study, we showed that NIV success was associated with Lower APACHE II score, lower rates of hemodynamic stability (absence of vasopressors) and shorter ICU stay.
Our study showed that NIV failure was associated with increased serum lactate. This agreed with Zhu ,12 Terzano as well, studied13
Our study also showed that NIV failure was associated with increased procalcitonin as well. In agreement with our finding Zhu 12
Our study also showed that patients with higher HACOR exhibited higher rates of NIV failure and required intubation. These findings agreed with Duan work. Jun Duan 14 Jun Duan also studied15
Our study showed that patients with higher lung ultrasound score exhibited higher rates of NIV failure and required intubation. Cut-off value upon NIV initiation was 18 (sensitivity 77%, specificity 60%, AUC 0.774) and cut off value after 12 hours was 15 (sensitivity 87%, specificity 75%, AUC 0.891) In agreement with our finding Silvia 16
Nobile 17 respiratory failure . He suggested cut-offs for lung reaeration score that could predict NIV failure. Soummer 18
5.1 Limitations
In this study included being a single physician performed the US and being a single center study. Cardiac patients were included in the study which might obscure NIV effects.
6 Conclusion
Our study showed that LUS, HACOR scores, and lactate could be used to predict NIV failure with good predictive value. HACOR and lung ultrasound scores improved among patients who showed good response to NIV while deteriorated among patients who showed NIV failure.
References
[1]. Hostetler MA. Use of noninvasive positive-pressure ventilation in the emergency department. Emerg Med Clin North Am 2008;26 (4):929–939. viii.
[2]. Plant PK, Owen JL, Elliott MW. Early use of non-invasive ventilation for acute exacerbations of chronic obstructive pulmonary disease on general respiratory wards: a multicentre randomised controlled trial. Lancet 2000;355 (9219):1931–1935.
[3]. Moerer O, Harnisch LO, Herrmann P, Zippel C, Quintel M. Patient-ventilator interaction during noninvasive ventilation in simulated COPD. Respir Care 2016;61 (1):15–22.
[4]. Bott J, Carroll MP, Conway JH, et al. Randomized controlled trial of nasal ventilation in acute ventilatory failure due to chronic obstructive airways disease. Lancet 1993;341:1555–1557.
[5]. Brochard L, Mancebo J, Wysocki M, et al. Noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease. N Engl J Med 1995;333 (13):817–822.
[6]. Celikel T, Sungur M, Ceyhan B, Karakurt S. Comparison of noninvasive positive pressure ventilation with standard medical therapy in hypercapnic acute
respiratory failure . Chest 1998;114:1636–1642.
[7]. Dikensoy O, Ikidag B, Filiz A, Bayram N. Comparison of non-invasive ventilation and standard medical therapy in acute hypercapnic
respiratory failure : a randomised controlled trial at a tertiary health centre in SE Turkey. Int J Clin Pract 2002;56:85–88.
[8]. Kramer N, Meyer TJ, Meharg J, Cece RD, Hill NS. Randomized, prospective trial of noninvasive positive pressure ventilation in acute
respiratory failure . Am J Respir Crit Care Med 1995;151 (6):1799–1806.
[9]. Ozyilmaz E, Ugurlu AO, Nava S. Timing of noninvasive ventilation failure: causes, risk factors, and potential remedies. BMC Pulm Med 2014;14:19. doi: 10.1186/1471-2466-14-19.
[10]. Lichtenstein DA, Lascols N, Prin S, Meziere G. The “lung pulse”: an early ultrasound sign of complete atelectasis. Intensive Care Med 2003;29:2187–2192. doi: 10.1007/s00134-003-1930-9.
[11]. Lichtenstein DA, Mezière G, Lascols N, et al. Ultrasound diagnosis of occult pneumothorax. Crit Care Med 2005;33:1231–1238. doi: 10.1097/01.CCM.0000164542.86954.B4.
[12]. Zhu JJ, Liu LJ. Analysis of factors influenced by the effectiveness of non-invasive ventilation in the treatment of acute exacerbation of chronic obstructive pulmonary disease with different severities. Eur Rev Med PharmacolSci 2016;20 (22):4775–4781.
[13]. Terzano C, Di Stefano F, Conti V, et al. Mixed acid–base disorders, hydroelectrolyte imbalance and lactate production in hypercapnic
respiratory failure : the role of noninvasive ventilation. PLoS ONE 2012;7 (4):e35245.
[14]. Duan J, Han X, Bai L, Zhou L, Shicong, Huang S. Assessment of heart rate, acidosis, consciousness, oxygenation, and respiratory rate to predict noninvasive ventilation failure in hypoxemic patients. Intensive Care Med 2017;43 (2):192–199.
[15]. Duan J, Wang S, Liu P, et al. Early prediction of non invasive ventilastion failure in COPD patients: derivation, internal validation, and external validation of a single risk score. Ann Intensive Care 2019;9:108.
[16]. Mongodi S, Santangelo E, Salve G, et al. Lung ultrasound score to monitor non-invasive respiratory support in hypoxemic patients. Eur Respir J 2019;54:A2322.
[17]. Nobile L, Beccaria P, Zambon M, Cabrini L, Landoni G, Zangrillo A. Lung ultrasound reaeration score: a useful tool to predict non-invasive ventilation effectiveness. Crit Care 2014;18 (suppl 1):255. Article number: P255.
[18]. Soummer A, Perbet S, Brisson H, et al. Ultrasound assessment of lung aeration loss during a successful weaning trial predicts postextubation distress. Crit Care Med 2012;40 (7):2064–2072.
