Skip to main content
Download PDF

Efficacy and Predictors of Success of Noninvasive Ventilation for Prevention of Extubation Failure in Critically Ill Children With Heart Disease

Pediatric Cardiology volume 34, pages 964–977 (2013)Cite this article

Abstract

The study aimed primarily to evaluate the efficacy of noninvasive ventilation (NIV) and to identify possible predictors for success of NIV therapy in preventing extubation failure in critically ill children with heart disease. The secondary objectives of this study were to assess the efficacy of prophylactic NIV therapy initiated immediately after tracheal extubation and to determine the characteristics, outcomes, and complications associated with NIV therapy in pediatric cardiac patients. A retrospective review examined the medical records of all children between the ages 1 day and 18 years who sustained acute respiratory failure (ARF) that required NIV in the cardiovascular intensive care unit (CVICU) at Lucile Packard Children’s Hospital between January 2008 and June 2010. Patients were assigned to a prophylactic group if NIV was started directly after extubation and to a nonprophylactic group if NIV was started after signs and symptoms of ARF developed. Patients were designated as responders if they received NIV and did not require reintubation during their CVICU stay and nonresponders if they failed NIV and reintubation was performed. The data collected included demographic data, preexisting conditions, pre-event characteristics, event characteristics, and outcome data. The outcome data evaluated included success or failure of NIV, duration of NIV, CVICU length of stay (LOS), hospital LOS, and hospital mortality. The two complications of NIV assessed in the study included nasal bridge or forehead skin necrosis and pneumothorax. The 221 eligible events during the study period involved 172 responders (77.8 %) and 49 nonresponders (22.2 %). A total of 201 events experienced by the study cohort received continuous positive airway pressure (CPAP), with 156 responders (78 %), whereas 20 events received bilevel positive airway pressure (BiPAP), with 16 responders (80 %). In the study, 58 events (26.3 %) were assigned to the prophylactic group and 163 events (73.7 %) to the nonprophylactic group. Compared with the nonprophylactic group, the prophylactic group experienced significantly shorter CVICU LOS (median, 49 vs 88 days; p = 0.03) and hospital LOS (median, 60 vs 103 days; p = 0.05). The CVICU LOS and hospital LOS did not differ significantly between the responders (p = 0.56) and nonresponders (p = 0.88). Significant variables identifying a responder included a lower risk-adjusted classification for congenital heart surgery (RACHS-1) score (1–3), a good left ventricular ejection fraction, a normal respiratory rate (RR), normal or appropriate oxygen saturation, prophylactic or therapeutic glucocorticoid therapy within 24 h of NIV initiation, presence of atelectasis, fewer than two organ system dysfunctions, fewer days of intubation before extubation, no clinical or microbiologic evidence of sepsis, and no history of reactive airway disease. As a well-tolerated therapy, NIV can be safely and successfully applied in critically ill children with cardiac disease to prevent extubation failure. The independent predictors of NIV success include lower RACHS-1 classification, presence of atelectasis, steroid therapy received within 24 h after NIV, and normal heart rate and oxygen saturations demonstrated within 24 h after initiation of NIV.

Critically ill children with heart disease can experience acute respiratory failure (ARF) as a result of lung disease or secondary to their underlying cardiac disease. Among children with ARF, those with underlying heart disease comprise a distinct subgroup whose course is frequently characterized by longer and more complicated intensive care unit (ICU) stays and complex cardiopulmonary interactions.

Conventionally, ARF in most children with heart disease is managed with endotracheal intubation and mechanical ventilatory support. The use of noninvasive ventilation (NIV) in pediatric cardiac patients as an alternative ventilatory support mode is not well established. In contrast, numerous reports on adults and children older than 4 years of age support the use of NIV in diverse clinical scenarios that include chronic obstructive pulmonary disease [4, 34], sleep apnea syndrome [4, 24, 34], pneumonia [10, 29], atelectasis [10], obstructive apnea/hypopnea [38], treatment after scoliosis repair [7], pulmonary edema [1], bronchiectasis [31], acute chest syndrome [9], acute leukemia [32], severe immunocompromise [30, 33], ARF from neuromuscular disease [5, 29], and status asthmaticus [6]. However, the literature on the use of NIV therapy for children with heart disease is scarce [43]. Technical problems, especially with regard to the interface and the ventilatory equipment, unknown effects on cardiopulmonary interactions, risk of aspiration, and effects on wound healing, frequently limit the use of NIV for children with heart disease, especially in the acute setting.

This study aimed primarily to evaluate the efficacy of NIV and identify possible predictors for the success of NIV therapy in preventing extubation failure for critically ill children with heart disease. The secondary objectives of this study were to assess the efficacy of prophylactic NIV therapy initiated immediately after extubation and to determine the characteristics, outcomes, and complications associated with NIV therapy in pediatric cardiac patients.

Methods

Setting and Patients

We performed a retrospective observational study in a 20-bed pediatric cardiovascular ICU (CVICU) at an academic children’s hospital during the period January 2008 to June 2010. The Institutional Review Board of the Stanford University Medical Center approved the study, and the need for informed consent was waived.

The study included all critically ill children with heart disease between the ages of 1 day and 18 years who received NIV at any time during their stay in the CVICU. Any patient with active medical or surgical heart disease was included in our study. The patients in the surgical category included patients who had undergone heart surgery (with or without cardiopulmonary bypass), whereas the patients in medical category included those with heart disease but no surgical intervention at the time of NIV initiation.

The study excluded patients with impending respiratory or cardiac failure after extubation who had only a brief trial of NIV (≤20 min) before tracheal reintubation, those receiving NIV only during procedural sedation, those reintubated for a subsequent surgery or radiologic procedure or interventional procedure, and those with a history of medical or surgical heart disease who were admitted to the ICU only for respiratory issues with no active cardiac issues.

Study Definitions

Acute respiratory failure was based on an oxygen requirement of 50 % or more for normal oxygen saturations (≥94 % for acyanotic cardiac lesions or 75–87 % for cyanotic cardiac lesions), an arterial partial pressure of carbon dioxide (PaCO2) at 50 mmHg or higher (or a venous PaCO2 ≥55 mmHg), or evidence of moderate to severe respiratory distress shown by dyspnea worsening from baseline, tachypnea (respiratory rate exceeding two standard deviations for the child’s age normal range), and the use of accessory muscles [25]. An ARF with ventilation–perfusion impairment, hypoxemia, and parenchymal condensations shown on chest radiography was considered as type 1 [38]. An ARF with hypoventilation, hypercapnia without hypoxemia, and absence of parenchymal condensations shown on chest radiography (excluding atelectasis) was considered as type 2 [38].

We assigned patients to a prophylactic group when NIV was started directly after extubation and to a nonprophylactic group when NIV was started after signs and symptoms of ARF developed. We designated patients as responders if they received NIV and did not require reintubation during their CVICU stay and as nonresponders if they failed NIV and tracheal reintubation was performed. We assigned patients to the following four categories for further analyses:

  1. 1.

    Prophylactic responders (PR): patients transitioned onto NIV directly after extubation and those who responded successfully to NIV with no need for reintubation during their CVICU stay.

  2. 2.

    Prophylactic nonresponders (PN): patients transitioned to NIV directly after extubation and those who did not respond successfully to NIV, with resultant reintubation during their CVICU stay.

  3. 3.

    Nonprophylactic responders (NR): patients transitioned to NIV only after evidence of ARF development and those who responded successfully to NIV with no need for reintubation during their CVICU stay.

  4. 4.

    Nonprophylactic nonresponders (NN): Patients transitioned to NIV after evidence of ARF development and those who did not respond successfully to NIV, with resultant reintubation during their CVICU stay.

Extubation Criteria

Tracheal extubation was performed according to a standardized protocol for ventilator weaning in our CVICU. The criteria for extubation included an inspired oxygen fraction (FiO2) of 0.4 or less required to maintain a systemic oxygen saturation of 94 % or more in acyanotic patients or 75–87 % in cyanotic patients, as measured by pulse oximetry, a peak inspiratory pressure (PIP) of 20 cm H2O or lower, a positive end-expiratory pressure (PEEP) of 6 cm H2O or lower, or a successful trial of pressure support at 12 cm H2O or lower for at least 30 min before extubation.

All patients were continuously monitored for respiratory rate (RR), heart rate (HR), and transdermal oxygen saturation. The adequacy of chest wall excursion and the use of accessory respiratory muscles were assessed by a member of the CVICU medical team before extubation.

In our CVICU, extubation readiness is assessed by the team caring for the patient (including the attending physician, the fellow or the resident physician, the bedside nurse, and the respiratory therapist). Our unit does not have a scoring system predicting extubation readiness. As a common clinical practice, patients intubated for 24 h or longer receive glucocorticoids for 24–48 h after extubation.

NIV Technique

The two methods of NIV that we used were continuous positive airway pressure (CPAP) and bilevel positive airway pressure (BiPAP® Vision®, Respironics, Inc., Murrysville, PA). The choice to initiate CPAP or BiPAP was made by the attending physician caring for the patient. Positive end-expiratory pressure (CPAP or EPAP) was initiated with a backup rate of 4–5 cm H2O for all of our patients, increased as necessary up to a maximum of 10–12 cm H2O if no improvement in oxygen saturation or arterial PaO2 was achieved. Inspiratory positive airway pressure (IPAP) was initiated at 6–8 cm H2O and gradually increased to a maximum of 18–20 cm H2O if no improvement in clinical status or PaCO2 was achieved.

The NIV was administered using ventilator circuits with leakage compensation software, an auto-flow track, and a good synchronized trigger. The two circuits used in our study were a dual-limb Y circuit consisting of an inspiratory limb, a patient connector, and an expiratory limb and a single-limb circuit consisting of large-bore tubing and an open exhaust port with a set of slotted vent holes.

We provided NIV to patients using nasal prongs, a nasal mask, or a facial mask fitted appropriately according to the patient’s age and size to achieve maximum comfort and minimum air leak. We placed a nasogastric tube in all patients before initiation of NIV to prevent gastric distension and emesis. A protective patch was placed over the skin of the nasal bridge (Duoderm; Bristol Myers-Squibb, New York, NY, USA) in patients with tight-fitting facial masks to avoid skin breakdown. The head of the bed was elevated to 45° for all patients to reduce the risk of aspiration.

We chose initial NIV settings according to each patient’s age and weight [25, 35]. Further adjustment was made according to the patient’s clinical status, chest radiograph, oxygen saturation, and gas exchange as measured by blood gas analysis.

Reintubation Criteria

The decision to reintubate is made by the CVICU attending physician and the team caring for the patient. The possible reasons for reintubation include clinical signs of respiratory fatigue and severe respiratory distress despite maximum NIV support, NIV intolerance (due to difficulty tolerating the nasal or facial mask or lack of cooperation), worsening hypercarbia (increase of ≥20 % from the baseline value) or hypoxemia (decrease of ≥20 % from the baseline value), inability to clear airway or oral secretions, hemodynamic decompensation, cardiorespiratory arrest, and a Glasgow coma scale lower than 8 or inability to maintain adequate airway patency due to neurologic impairment.

Prophylactic NIV Criteria

In general, we instituted NIV immediately after tracheal extubation for the following patients: those deemed at high risk for extubation failure and those being weaned from conventional mechanical ventilation. The patients considered at high risk for extubation failure included smaller and younger patients; children who received prolonged continuous mechanical ventilation; children with low left ventricular ejection fraction at the time of extubation, evidence of low cardiac output, pulmonary edema, cardiac surgery with a risk-adjusted classification for congenital heart surgery (RACHS-1) score of 4–6, and evidence of acute lung injury or acute respiratory distress syndrome; children with a neuromuscular disorder, and children with suspected/proven diaphragm dysfunction. The decision to initiate prophylactic NIV was made by the CVICU attending physician and the team caring for the patient.

Data Collection

For each patient, the following variables were collected: age, gender, weight, timing and duration of the NIV event, RACHS-1 score [17], and associated diagnoses (cardiomyopathy, dysrhythmia, myocarditis, pulmonary hypertension, immunodeficiency, chromosomal abnormality, underlying neuromuscular disease). The duration of the event was defined as the time from the initiation of NIV to either reintubation or successful transition to a nasal cannula or room air.

The data collected on variables immediately before initiation of NIV included upper airway obstruction/stridor; administration of intravenous glucocorticoids within 24 h after initiation of NIV, HR, or RR; oxygen saturation; atelectasis; inotrope score [21]; time elapsed since extubation (in hours); days of intubation before extubation; ventricular ejection fraction (if available); pulmonary edema; white cell count; platelet count; C-reactive protein (CRP); two or more organ system dysfunctions; sepsis (culture proven or suspected); pneumonia; FiO2; arterial pH; arterial CO2 (PaCO2); and arterial O2 (PaO2). The HR and RR were expressed as age-appropriate normal or abnormal [28]. A mean of two values for these variables before initiation of NIV was taken as the pre-NIV value for study purposes.

Oxygen saturation was expressed as normal if it was 94 % or higher for acyanotic cardiac lesions or 75–87 % for cyanotic cardiac lesions. The HR, RR, and oxygen saturation were recorded approximately 5–20 min before initiation of NIV. The official radiologist interpretation was used to differentiate consolidation from atelectasis on chest radiographs.

Qualitatively good ventricular systolic function or an ejection fraction (EF) of 55 % or greater as determined by an echocardiographer blinded to the two groups constituted a good EF for study purposes. The criteria for organ dysfunction were based on International Pediatric Sepsis Consensus Conference guidelines [13]. The clinical outcome was evaluated for all patients with success or failure of NIV as well as the duration of NIV (hours), the CVICU length of stay (LOS), the hospital LOS, and the mortality rate. The two complications of NIV assessed in the study were nasal bridge or forehead skin necrosis and incidence of pneumothorax.

Statistical Analysis

Continuous variables are presented as the median [Q1, Q3], where Q1 is the 25th percentile, and Q3 is the 75th percentile and/or the mean ± standard deviation, whereas categorical variables are presented as numbers and percentages. Calculation of p values was performed using the Chi square test, Fisher’s exact test of independence, or both for categorical variables and Wilcoxon rank-sum test for continuous variables. Multivariable logisitic regression models were used to assess variables associated with success of NIV in the overall sample as well as within subsets of prophylactic and nonprophylactic events.

Variables with a p value of 0.2 or less in the univariate analysis were entered into the multiple regression model. Any variable with 20 % or more missing values and any rare variable (≤5 subjects) were not considered for inclusion in the multivariate models.

The model results were expressed in terms of adjusted odds ratios, 95 % confidence intervals, and p values. Backward model selection was used to fit a parsimonious model. Several additional multiple logistic analyses were performed to explore variables left out of the model and to achieve a parsimonious model.

The model’s goodness-of-fit was evaluated using the Hosmer–Lemeshow test, and the discrimination of the model was assessed using the area under the receiver operating characteristic curve (ROC). Kaplan–Meier estimates-of-survival curves were used to determine the CVICU and hospital LOS as well as the days of intubation before extubation in various groups and subgroups.

A logistic regression model with responder as the response and days of intubation before extubation as the predictor was fitted to explore the log odds of being a responder as a function of days of intubation before extubation. Harrell’s method of restricted cubic splines with four knots was used to parameterize the days of intubation before extubation in a nonlinear manner. Analyses were performed using STATA/MP, version 11.1 software (Stata; Corp LP, College Station, TX, USA) and Harrell’s RMS package in R (available from biostat.mc.vanderbilt.edu/rms).

Results

During the study period, 255 NIV events occurred, 221 of which were eligible for inclusion in the study (Fig. 1). The study excluded 15 events with an NIV trial lasting 20 min or less before reintubation, 1 event with NIV used only during procedural sedation, and 18 events during which NIV was transitioned to mechanical ventilation for surgical, radiologic, or another procedure. The 221 eligible events involved 172 responders (77.8 %) and 49 nonresponders (22.2 %). In the study cohort, 201 events received CPAP with 156 responders (78 %), whereas 20 events received (BiPAP), with 16 responders (80 %).

Fig. 1
figure 1

Flow diagram depicting the classification of study events

Full size image

In the allocation of events, 58 (26.3 %) were assigned to the prophylactic group and 163 (73.7 %) to the nonprophylactic group. The patients in 200 (90.4 %) of 221 eligible events received NIV in the postoperative period. Of these 200 events, 50 were in RACHS 4–6 category. Only 21 patients with medical disease received NIV in this study. Of these, 2 patients had cardiomyopathy or myocarditis, and the remaining 19 patients received NIV before their heart surgery.

Table 1 summarizes univariate comparisons between responders and nonresponders and between prophylactic and nonprophylactic events. Table 2 provides the evolution of cardiorespiratory parameters after initiation of NIV in this select patient population.

Table 1 Characteristics and outcomes among responders (R) versus nonresponders (NR) and prophylactic (P) versus nonprophylactic (NP) events
Full size table
Table 2 Evolution of cardiorespiratory parameters after initiation of noninvasive ventilation (NIV) in critically ill children with heart disease
Full size table

Responders Versus Nonresponders

The study cohort comprised 172 responder events (77.8 %) and 49 nonresponder events (79 %) (p = 0.43). The baseline demographics were similar in the responder and nonresponder groups (Table 1). Significant variables for identifying a responder included a lower RACHS-1 score (1–3), good left ventricular (LV) EF, prophylactic or therapeutic glucocorticoid therapy within 24 h after initiation of NIV, presence of atelectasis, fewer than two organ system dysfunctions, fewer days of intubation before extubation, no clinical or microbiologic evidence of sepsis, and no history of reactive airway disease.

Table 2 compares the evolution of cardiorespiratory variables in responders and nonresponders after initiation of NIV in our study. Table 3 depicts normalization of heart rate, respiratory rate, and oxygen saturations with time after initiation of NIV for critically ill children with heart disease. Table 4 describes the independent predictors associated with responders to NIV therapy.

Table 3 Normalization of heart rate, respiratory rate, and oxygen saturations over time after initiation of noninvasive ventilation (NIV) in children with heart disease
Full size table
Table 4 Multivariable logistic regression model for being a responder and a nonprophylactic responder
Full size table

Prophylactic NIV Therapy After Extubation

The prophylactic group had 43 responders (74 %), and the nonprophylactic group had 129 responders (79 %) (p = 0.43). Smaller and younger children were more likely to receive NIV directly after extubation. The prophylactic and nonprophylactic groups differed significantly in terms of age (median, 3 months [9, 34] vs median 5 months [24, 30]; p = 0.04), body weight (median, 4.3 kg [2.8, 6.9] vs median, 4.8 kg [3.6, 8.9]; p = 0.03), acyanotic heart defect (90 % vs 75 %; p = 0.02), inotrope score (median, 0 [0, 7.5] vs median, 0 [0, 3.0]; p = 0.003), days of intubation before extubation (median, 6 days [24, 32] vs median 3 days [4, 7]; p < 0.001), and underlying neuromuscular disease (9 % vs 2 %; p = 0.02). None of the cardiac variables differed significantly between the two groups.

Evolution of Clinical Parameters

The majority of responders in the study cohort had a normal heart rate, a normal respiratory rate, and normal O2 saturations before initiation of NIV. Heart rate, respiratory rate, and O2 saturations normalized sooner in the responders than in the nonresponders, mostly in 6–24 h (Tables 2, 3). The FiO2 requirement and the PaCO2 values normalized in responders during the first 48 h after initiation of NIV, whereas both of these parameters worsened in nonresponders (Tables 2, 3).

Length of CVICU and Hospital Stays, Complications, and Mortality

Although the duration of NIV was longer in the prophylactic group than in the nonprophylactic group (median, 44 h [17, 67] vs median, 18 h [9, 60]; p = 0.03), the CVICU LOS (median, 49 days [16, 114] vs median, 88 days [18, 148]; p = 0.03) and hospital LOS (median, 60 days [25, 120] vs median, 103 days [36, 155]; p = 0.05) were significantly shorter in the prophylactic group than in the nonprophylactic group (Fig. 2).

Fig. 2
figure 2

Comparison of intensive care unit (ICU) lengths of stay between prophylactic events and nonprophylactic events

Full size image

The CVICU LOS and hospital LOS did not differ significantly between the responders and nonresponders (p = 0.56 and p = 0.88, respectively). The CVICU mortality rate was significantly higher in nonresponder group (31 %) than in the responder group (4 %) (p < 0.001). However, the mortality rate did not differ between prophylactic and nonprophylactic events (10 % in both; p = 0.9).

Duration of Mechanical Ventilation Before Extubation

The duration of mechanical ventilation was longer for the responders (median, 7.5 days (2.2, 20) vs median, 3 (4, 7); p = 0.004) and those assigned to the prophylactic group (median, 6 days (24, 32) vs median, 3 days [2, 7]; p < 0.001) (Table 1). Figure 3 compares the proportions of patients waiting for extubation between the responder and nonresponder groups.

Fig. 3
figure 3

Days of intubation before extubation compared between responders and nonresponders

Full size image

The nonresponders were intubated for a longer period than the responders (p = 0.006). A trend toward significance was noted between the nonprophylactic responders and the nonprophylactic nonresponders (p = 0.05) and between the prophylactic responders and the prophylactic nonresponders (p = 0.09). The odds of being a responder with increasing days of tracheal intubation before extubation based on an exploratory univariable logistic regression demonstrated 3–4 days as the time for the maximum odds of being a responder.

Prophylactic Responders Versus Prophylactic Nonresponders

During the study period, 15 events occurred in the prophylactic nonresponder group, and 43 events occurred in the prophylactic responder group. Table 5 compares the characteristics and outcomes between the prophylactic responders (PR) and the prophylactic nonresponders (PN). Figure 4 demonstrates the evolution of the prophylactic responders (PR) and the prophylactic nonresponders (PN) with increasing days of intubation. The responders and nonresponders did not differ significantly in hours of NIV therapy (median, 44 h [16, 74] for PR vs median, 45 h [12.5, 67] for PN; p = 0.68). However, the CVICU LOS (median, 32 days [15, 101] for PR vs median, 101 days [49, 133] for PN; p ≤ 0.007), and hospital LOS (median, 55 days [22, 120] for PR vs median, 120 days [58, 134] for PN; p = 0.029) were significantly shorter in the responder group than in the nonresponder group. Also, Fisher’s exact test showed a significant difference in CVICU mortality between the responders and the nonresponders (5 % for PR vs 27 % for PN; p = 0.03).

Table 5 Characteristics and outcomes among prophylactic responder (PR) versus prophylactic nonresponder (PN) and nonprophylactic responder (NR) versus nonprophylactic nonresponder (NN) events
Full size table
Fig. 4
figure 4

Evolution of prophylactic responders (PR), prophylactic nonresponders (PN), nonprophylactic responders (NR), and nonprophylactic nonresponders (NN) with increasing days of intubation

Full size image

Nonprophylactic Responders Versus Nonprophylactic Nonresponders

During the study period, 34 events occurred in the nonprophylactic nonresponder (NN) group, and 129 events occurred in the nonprophylactic responder (NR) group. Table 5 compares the characteristics and outcomes of the NR and NN events. Figure 4 demonstrates the evolution of NR and NN with increasing days of intubation. There was no significant difference in hours of NIV (median, 16 h [9, 48.5] for NR vs median, 27 h [9, 88]; mean, 76.4 ± 99.2 h for NN; p = 0.26), CVICU LOS (median, 109 days [16, 148] for NR vs median, 58 days [29, 116] for NN; p = 0.6), or hospital LOS (median, 120 days [26, 155] for NR vs median, 66.5 days [43, 120] for NN; p = 0.33). However, the responders and nonresponders differed significantly in terms of CVICU mortality days (4 % for NR vs 32 % for NN; p < 0.001).

Discussion

The primary findings of this study can be summarized as follows:

  1. 1.

    NIV can be successfully applied for critically ill children with heart disease to prevent extubation failure.

  2. 2.

    A longer period of intubation and mechanical ventilatory support before extubation is associated with an increased possibility of NIV failure.

  3. 3.

    NIV failure is associated with increased mortality among both prophylactic and nonprophylactic groups.

Our results are comparable with those of other studies in pediatric noncardiac patients, which have reported success rates of 57–92 % [3, 8, 10, 19, 22, 25, 29, 31, 41]. The overall success rate of 77.8 % in the current study strongly supports the use of NIV as an intervention to reduce ARF in critically ill children with heart disease.

The nonresponders had a significantly higher ICU mortality rate. The mortality rates did not differ significantly between the prophylactic and nonprophylactic groups. The reason for the increased mortality among the nonresponders could have been the presence of sicker patients in the nonresponder group, demonstrated by a higher RACHS score and worse heart function. It also is possible that NIV in patients with marginal cardiac function may cause deterioration of their clinical status after its initiation. For hemodynamically unstable patients, NIV should be aborted within a few hours after its initiation if vital signs show no improvement. However, this is only a speculation not supported by data from the study.

Extubation failure in children with heart disease is variable, and the etiology is diverse. The cardiopulmonary effects from removal of positive-pressure ventilation are more pronounced in children with extubation failure and include escalation in the oxygen requirement as well as an increase in mean arterial blood pressure. Extubation failure after prolonged mechanical ventilation in children is quite common, observed in 22–28 % of premature babies, 15–20 % of critically ill children, and 10 % of children after cardiac surgery [2, 11, 15, 18, 20, 40].

In a recently published study from our group, we demonstrated an extubation failure rate of 22 % (14/64) in children after the Norwood operation [14]. Extubation failure requiring emergent reintubation may cause significant hemodynamic instability, unnecessary airway trauma, increased risk for nosocomial infections, and prolonged duration of mechanical ventilation and ICU stay [14, 18].

The possible reasons for extubation failure in children with heart disease include lung disease, cardiac dysfunction, diaphragmatic paralysis, airway edema, and vocal cord paralysis [14, 15]. To prevent extubation failure, NIV may be a useful therapy in certain scenarios such as those involving lung disease, diaphragmatic paralysis, and airway edema. In other scenarios such as those involving vocal cord paralysis, NIV may be dangerous because these patients are at high risk for aspiration. Moreover, the cardiopulmonary effects of NIV in varied cardiac physiologies are unknown in children with heart disease.

Predictors of NIV Success in Children With Heart Disease

A main objective of the study was to identify factors predictive of NIV success in children with heart disease. We found that children with lower RACHS-1 classification (categories 1–3) had a greater likelihood of NIV success.

Administration of glucocorticoids within 24 h of NIV initiation was another independent predictor associated with success of NIV therapy. Several pediatric studies have shown improvement in postextubation stridor and lower reintubation rates with the use of prophylactic dexamethasone [39]. In addition, the presence of atelectasis was an independent factor associated with success of NIV therapy in our study. Findings have shown NIV to improve atelectasis in postsurgical patients and obese patients [12, 16, 36].

The responders in our cohort more commonly had findings consistent with type 1 ARF, that is, higher FiO2 requirements and lower PaCO2 levels (in both cyanotic and acyanotic events) at the time of NIV initiation. Type 1 ARF in this select patient population could have been due to a ventilation–perfusion mismatch, atelectasis or parenchymal consolidation, acute lung injury or acute respiratory distress syndrome, or pulmonary hypertension. It also is possible that type 1 ARF in our patient population was due to adverse hemodynamics leading to intracardiac mixing that caused right-to-left shunting. Our results are contrary to the already published literature on the use of NIV in children, wherein type 1 ARF is associated with NIV failure [8, 19, 25].

In the published literature, the use of NIV for young patients as well as those with neuromuscular disease and immunodeficiency has produced better outcomes [19, 31, 36]. The current study, however, did not demonstrate success of NIV in young patients or those with neuromuscular disorder, chromosomal abnormality, or immunodeficiency.

Joshi and Tobias [19] demonstrated pulmonary parenchymal disease as a risk factor for NIV failure. We observed no association of pulmonary parenchymal disease with NIV success or failure. We also did not show any association of other cardiac factors such as cyanotic heart defects, dysrhythmias, pulmonary hypertension, or inotrope score with NIV failure. Findings have shown that the use of BiPAP ventilation for delivery of inspiratory pressure markedly reduces the work of breathing [10]. Improving ventilation by BiPAP in patients with acute respiratory insufficiency after cardiac surgery is associated mainly with reduced work of breathing and has alleviated the respiratory muscle fatigue [10]. However, in our cohort, only 20 events (9 %) received BiPAP with 16 responders (80 %). A higher proportion of patients in the responder group received steroids than in the nonresponder group (57 % vs 26 %). These results should be interpreted with caution because all the patients receiving steroids may not have been captured due to limited data availability.

Prophylactic NIV Therapy After Extubation

It has been suggested in adult trials that initiation of NIV therapy during the first 48 h after extubation in high-risk patients is more effective than standard medical therapy in preventing postextubation respiratory failure [27]. Our results suggest that application of prophylactic NIV therapy immediately after extubation does not decrease the rate of reintubation. In addition, younger and smaller children were more likely to receive prophylactic therapy for prevention of ARF. It should be noted that patients in the prophylactic group were never allowed to experience respiratory failure before initiation of NIV. The possible reasons for the use of more aggressive NIV therapy with these patients could have been weaker chest wall muscles, a narrower airway, and greater difficulty handling secretions despite adequate respiratory physiotherapy [42].

Our results differ from those published by Mayordomo-Colunga et al. [26], who demonstrated a higher success rate for children who received prophylactic NIV therapy, termed as elective NIV (eNIV) (81 %), than for children in a nonprophylactic group, termed as rescue NIV (rNIV) (50 %) (p = 0.037).

Feasibility of NIV Therapy for Children With Heart Disease

Application of NIV requires a “buy-in” from the bedside nurse and respiratory therapist, especially for smaller children. We speculate that the high success rate for NIV therapy in our study was in part related to the presence of a dedicated, highly trained, motivated group of respiratory therapists and bedside nursing staff in our CVICU.

A significant level of team motivation is crucial, especially for smaller children, who can experience multiple logistical difficulties. Selection of the correct interface (nasal prongs and a variety of nasal masks or facial masks) and ventilatory equipment appropriate for the patient also is of prime importance. Our center also uses commercial nasal masks designed for older children and adults as oronasal interfaces for smaller children. In addition, our nursing staff and respiratory therapists spend a significant amount of time educating and establishing effective channels of communication with patients and families regarding the goals and the therapeutic end point.

Another reason for our high success rate may be the use of newer sedation agents such as dexmedetomidine to improve patient cooperation and to achieve more effective patient–ventilator synchrony [23, 37]. The other risks for NIV in children are aspiration pneumonitis, increased incidence of infections, maintenance of airway patency due to increased secretions, and increased risk of airway hemorrhage [8, 10, 19, 25, 29, 42].

Complications and Mortality

In our cohort, NIV was a well-tolerated therapy. The most common complication was nasal bridge skin necrosis in 8 % of the patients, followed by pneumothorax in 4 % of the patients. All the children with skin necrosis eventually healed without the need for any major intervention. The complication rate in our study was comparable with those in other pediatric studies [8, 10, 19, 25, 29, 42]. The ICU mortality rate was significantly higher for the nonresponders than for the responders in both the prophylactic and nonprophylactic groups. This may have been due to the fact that the children in the nonresponder group were sicker and more hemodynamically unstable than those in the responder group. The mortality rates did not differ significantly between the prophylactic and nonprophylactic groups. However, the mortality rate was higher for the cohort receiving NIV (10 %) than for all the patients admitted to the CVICU during the study period (3 %). For CVICU LOS, we attribute this difference to a higher case mix index among the patients who received NIV therapy.

Limitations of the Study

  1. 1.

    This study was a single-center study, and the results may not be generalizable to all centers. Moreover, the retrospective nature of the study rendered it susceptible to study design flaws and bias.

  2. 2.

    This study lacked a control group of patients with ARF who were not offered NIV to prevent reintubation.

  3. 3.

    The definition of ARF was modified for cardiac patients according to exisiting pediatric ICU literature. However, it still may not be perfect for representing cardiac patients with respiratory failure.

  4. 4.

    Patients receiving a NIV trial of 20 min were excluded from this study, which may have influenced the results. These patients were already experiencing impending respiratory failure that would have needed reintubation irrespective of a NIV trial.

  5. 5.

    Due to the retrospective nature of this study and limited data availability, the study is missing the etiology of extubation failure. Because of this limitation, we were not able to show the efficacy and success of NIV with a particular etiology of extubation failure for this select patient population.

  6. 6.

    This study lacked hemodynamic data such as central venous pressure, right atrial pressure, and left atrial pressure both before and after NIV.

  7. 7.

    The patients who received prophylactic NIV in the study cohort may not have needed positive-pressure ventilation of any sort, and it may have influenced the final results.

  8. 8.

    The patients in the nonresponder group were sicker than those in the responder group. The majority of the responders had a normal heart rate, a normal respiratory rate, and normal oxygen saturations before initiation of NIV. This may have contributed to a higher failure rate for the nonresponders than for the responders.

  9. 9.

    Due to the retrospective nature of the study and limited data availability, the study was missing data on complications associated with NIV such as aspiration pneumonia, infections, airway hemorrhage, and tracheal secretions.

Despite these limitations, this study has laid the foundation for future prospective, randomized, controlled multicenter trials.

Conclusions

As a well-tolerated therapy, NIV can be safely and successfully applied for critically ill children with cardiac disease to prevent extubation failure. The independent predictors for NIV success include a good left ventricular ejection fraction, administration of glucocorticoids within 24 h after initiation of NIV, normal RR, lower RACHS-1 score (1–3), less organ dysfunction, and the presence of atelectasis before initiation of NIV.

References

  1. Akingbola OA, Servant GM, Custer JR, Palmisano JM (1993) Noninvasive bi-level positive-pressure ventilation: management of two pediatric patients. Respir Care 38:1092–1098

    Google Scholar 

  2. Balsan MJ, Jones JG, Watchko JF et al (1990) Measurements of pulmonary mechanics prior to elective extubation of neonates. Pediatr Pulmonol 9:238–243

    PubMed  Article  CAS  Google Scholar 

  3. Bernet V, Hug MI, Frey B (2005) Predictive factors for the success of noninvasive mask ventilation in infants and children with acute respiratory failure. Pediatr Crit Care Med 6:660–664

    PubMed  Article  Google Scholar 

  4. Brochard L, Mancebo J, Wysocki M, Lofaso F, Conti G, Rauss A, Simonneau G, Benito S, Gasparetto A, Lemaire F et al (1995) Noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease. N Engl J Med 333:817–822

    PubMed  Article  CAS  Google Scholar 

  5. Bryan AC, Wohl MB (1986) Respiratory mechanics in children. In: Macklem PT, Mead J (eds) Handbook of physiology. American Physiological Society, Bethesda, pp 179–191

    Google Scholar 

  6. Carroll CL, Schramm CM (2006) Noninvasive positive-pressure ventilation for the treatment of status asthmaticus in children. Ann Allergy Asthma Immunol 96:454–459

    PubMed  Article  CAS  Google Scholar 

  7. Doherty MJ, Millner PA, Latham M, Dickson R, Elliott M (2001) Noninvasive ventilation in the treatment of ventilatory failure following corrective spinal surgery. Anaesthesia 56:235–238

    PubMed  Article  CAS  Google Scholar 

  8. Essouri S, Chevret L, Durand P, Haas V, Fauroux B, Devictor D (2006) Noninvasive positive pressure ventilation: five years of experience in a pediatric intensive care unit. Pediatr Crit Care Med 7:329–334

    PubMed  Article  Google Scholar 

  9. Fartoukh M, Lefort Y, Habibi A, Bachir D, Galacteros F, Godeau B, Maitre B, Brochard L (2010) Early intermittent noninvasive ventilation for acute chest syndrome in adults with sickle cell disease: a pilot study. Intensive Care Med 36:1355–1362

    PubMed  Article  Google Scholar 

  10. Fortenberry JD, Del Toro J, Jefferson LS, Evey L, Haase D (1995) Management of pediatric acute hypoxemic respiratory insufficiency with bi-level positive pressure (BiPAP) nasal mask ventilation. Chest 108:1059–1064

    PubMed  Article  CAS  Google Scholar 

  11. Freely TW, Hedley-Whyte J (1975) Weaning from controlled ventilation and supplemental oxygen. N Engl J Med 292:903–906

    Article  Google Scholar 

  12. Gaszynski T, Tokarz A, Piotrowski D, Boussignac MachalaW (2007) CPAP in the postoperative period in morbidly obese patients. Obes Surg 17:452–456

    PubMed  Article  Google Scholar 

  13. Goldstein B, Giroir B, Randolph A (2005) International pediatric sepsis consensus conference: definitions for sepsis and organ dysfunction in pediatrics. Pediatr Crit Care Med 6:2–8

    PubMed  Article  Google Scholar 

  14. Gupta P, McDonald R, Gossett JM, Butt W, Shinkawa T, Imamura M, Bhutta AT, Prodhan P (2012) A single-center experience of extubation failure in infants undergoing the Norwood operation. Ann Thorac Surg [Epub ahead of print 10 July]

  15. Harrison AM, Cox AC, Davis S, Piedmonte M, Drummond-Webb JJ, Mee RB (2002) Failed extubation after cardiac surgery in young children: prevalence, pathogenesis, and risk factors. Pediatr Crit Care Med 3:148–152

    PubMed  Article  Google Scholar 

  16. Jaber S, Michelet P, Chanques G (2010) Role of noninvasive ventilation (NIV) in the perioperative period. Best Pract Res Clin Anaesthesiol 24:253–265

    PubMed  Article  Google Scholar 

  17. Jenkins KJ, Gauvreau K, Newburger JW, Spray TL, Moller JH, Iezzoni LI (2002) Consensus-based method for risk adjustment for surgery for congenital heart disease. J Thorac Cardiovasc Surg 123:110–118

    PubMed  Article  Google Scholar 

  18. Johnston C, de Carvalho WB, Piva J, Garcia PC, Fonseca MC (2010) Risk factors for extubation failure in infants with severe acute bronchiolitis. Respir Care 55:328–333

    PubMed  Google Scholar 

  19. Joshi G, Tobias JD (2007) A five-year experience with the use of BiPAP in a pediatric intensive care unit population. J Intensive Care Med 22:38–43

    PubMed  Article  Google Scholar 

  20. Khan N, Brown A, Venkataraman ST (1996) Predictors of extubation success and failure in mechanically ventilated infants and children. Crit Care Med 24:1568–1579

    PubMed  Article  CAS  Google Scholar 

  21. Ko WJ, Lin CY, Chen RJ, Wang SS, Lin FY, Chen YS (2002) Extracorporeal membrane oxygenation support for adult postcardiotomy cardiogenic shock. Ann Thorac Surg 73:538–545

    PubMed  Article  Google Scholar 

  22. Larrar S, Essouri S, Durand P, Chevret L, Haas V, Chabernaud JL, Leyronnas D, Devictor D (2006) Effects of nasal continuous positive airway pressure ventilation in infants with severe acute bronchiolitis. Arch Pediatr 13:1397–1403

    PubMed  Article  CAS  Google Scholar 

  23. Le KN, Moffett BS, Ocampo EC, Zaki J, Mossad EB (2011) Impact of dexmedetomidine on early extubation in pediatric cardiac surgical patients. Intensive Care Med 37:686–690

    PubMed  Article  CAS  Google Scholar 

  24. Lin M, Yang YF, Chiang HT, Chang MS, Chiang BN, Cheitlin MD (1995) Reappraisal of continuous positive airway pressure therapy in acute cardiogenic pulmonary edema: short-term results and long-term follow-up. Chest 107:1379–1386

    PubMed  Article  CAS  Google Scholar 

  25. Mayordomo-Colunga J, Medina A, Rey C, Díaz JJ, Concha A, Los Arcos M, Menéndez S (2009) Predictive factors of noninvasive ventilation failure in critically ill children: a prospective epidemiological study. Intensive Care Med 35:527–536

    PubMed  Article  Google Scholar 

  26. Mayordomo-Colunga J, Medina A, Rey C, Concha A, Menéndez S, Los Arcos M, García I (2010) Noninvasive ventilation after extubation in paediatric patients: a preliminary study. BMC Pediatr 10:29

    PubMed  Article  Google Scholar 

  27. Nava S, Gregoretti C, Fanfulla F, Squadrone E, Grassi M, Carlucci A, Beltrame F, Navalesi P (2005) Noninvasive ventilation to prevent respiratory failure after extubation in high-risk patients. Crit Care Med 33:2465–2470

    PubMed  Article  Google Scholar 

  28. Needlman R (2005) The Harriet lane handbook. 17th edn. Elsevier, Mosby, pp 174, 620

  29. Padman R, Lawless ST, Kettrick RG (1998) Noninvasive ventilation via bi-level positive airway pressure support in pediatric practice. Crit Care Med 26:169–173

    PubMed  Article  CAS  Google Scholar 

  30. Pancera CF, Hayashi M, Fregnani JH, Negri EM, Deheinzelin D, de Camargo B (2008) Noninvasive ventilation in immunocompromised pediatric patients: eight years of experience in a pediatric oncology intensive care unit. J Pediatr Hematol Oncol 30:533–538

    PubMed  Article  Google Scholar 

  31. Phua J, Ang YL, See KC, Mukhopadhyay A, Santiago EA, Dela Pena EG, Lim TK (2010) Noninvasive and invasive ventilation in acute respiratory failure associated with bronchiectasis. Intensive Care Med 36:638–647

    PubMed  Article  Google Scholar 

  32. Piastra M, Antonelli M, Chiaretti A, Polidori G, Polidori L, Conti G (2004) Treatment of acute respiratory failure by helmet-delivered noninvasive pressure support ventilation in children with acute leukemia: a pilot study. Intensive Care Med 30:472–476

    PubMed  Article  Google Scholar 

  33. Piastra M, De Luca D, Pietrini D, Pulitanò S, D’Arrigo S, Mancino A, Conti G (2009) Noninvasive pressure-support ventilation in immunocompromised children with ARDS: a feasibility study. Intensive Care Med 35:1420–1427

    PubMed  Article  Google Scholar 

  34. Plant PK, Owen JL, Elliott MW (2000) Early use of noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease on general respiratory wards: a multicenter randomised controlled trial. Lancet 355:1931–1935

    PubMed  Article  CAS  Google Scholar 

  35. Pons M, Cambra A (2003) Noninvasive ventilation. An Pediatr Barc 59:165–172

  36. Pons Odena M, Piqueras Marimbaldo I, Segura Matute S, Balaguer Argallo M, Palomeque Rico A (2009) Noninvasive ventilation after cardiac surgery: a prospective study. An Pediatr (Barc) 71:13–19

    Article  CAS  Google Scholar 

  37. Shehabi Y, Nakae H, Hammond N, Bass F, Nicholson L, Chen J (2010) The effect of dexmedetomidine on agitation during weaning of mechanical ventilation in critically ill patients. Anaesth Intensive Care 38:82–90

    PubMed  CAS  Google Scholar 

  38. Teague WG, Kervin L, Dawadkar V, Scott P (1991) Nasal bi-level positive airway pressure acutely improves ventilation and oxygen saturation in children with upper airway obstruction. Am Rev Respir Dis 143:505

    Article  Google Scholar 

  39. Tellez DW, Galvis AG, Storgion SA, Amer HN, Hoseyni M, Deakers TW (1991) Dexamethasone in the prevention of postextubation stridor in children. J Pediatr 118:289–294

    PubMed  Article  CAS  Google Scholar 

  40. Tobin MJ, Perez W, Guenther SM et al (1986) The pattern of breathing during successful and unsuccessful trials of weaning from mechanical ventilation. Am Rev Respir Dis 134:1111–1118

    PubMed  CAS  Google Scholar 

  41. Yan AT, Bradley TD, Liu PP (2001) The role of continuous positive airway pressure in the treatment of congestive heart failure. Chest 120:1675–1685

    PubMed  Article  CAS  Google Scholar 

  42. Yañez LJ, Yunge M, Emilfork M, Lapadula M, Alcántara A, Fernández C, Lozano J, Contreras M, Conto L, Arevalo C, Gayan A, Hernández F, Pedraza M, Feddersen M, Bejares M, Morales M, Mallea F, Glasinovic M, Cavada G (2008) A prospective, randomized, controlled trial of noninvasive ventilation in pediatric acute respiratory failure. Pediatr Crit Care Med 9:484–489

    PubMed  Article  Google Scholar 

  43. Zhang CY, Tan LH, Shi SS, He XJ, Hu L (2006) Noninvasive ventilation via bilevel positive airway pressure support in pediatric patients after cardiac surgery. World J Pediatr 2:297–302

    Google Scholar 

Download references

Acknowledgments

We thank Teresa Nishikawa for her staff support during the course of this research.

Disclosure

Authors have nothing to disclose with regard to commercial support or financial relationships.

Financial Support for the Study

None.

Author information

Affiliations

  1. Section of Pediatric Cardiology and Critical Care, Department of Pediatrics, University of Arkansas for Medical Sciences, 1 Children’s Way, Slot 512-3, Little Rock, AR, 72202-3591, USA

    Punkaj Gupta & Parthak Prodhan

  2. Department of Medical Education, Stanford University School of Medicine, Palo Alto, CA, USA

    Jacob E. Kuperstock & Sana Hashmi

  3. Department of Respiratory Care Services, Lucile Packard Children’s Hospital, Palo Alto, CA, USA

    Vickie Arnolde

  4. Biostatistics Program, Department of Pediatrics, University of Arkansas for Medical Sciences, Little Rock, AR, USA

    Jeffrey M. Gossett

  5. Department of Critical Care Medicine and Pediatrics, University of Pittsburgh, Pittsburgh, PA, USA

    Shekhar Venkataraman

  6. Division of Pediatric Cardiology, Department of Pediatrics, Stanford University Medical Center, Palo Alto, CA, USA

    Stephen J. Roth

Authors
  1. Punkaj Gupta
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Jacob E. Kuperstock
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Sana Hashmi
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Vickie Arnolde
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jeffrey M. Gossett
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Parthak Prodhan
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Shekhar Venkataraman
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Stephen J. Roth
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Punkaj Gupta.

Rights and permissions

Reprints and Permissions

About this article

Cite this article

Gupta, P., Kuperstock, J.E., Hashmi, S. et al. Efficacy and Predictors of Success of Noninvasive Ventilation for Prevention of Extubation Failure in Critically Ill Children With Heart Disease. Pediatr Cardiol 34, 964–977 (2013). https://doi.org/10.1007/s00246-012-0590-3

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00246-012-0590-3

Keywords

  • Acute respiratory failure
  • Bilevel positive airway pressure
  • Cardiovascular intensive care unit
  • Continuous positive airway pressure
  • Extubation failure
  • Noninvasive ventilation
  • Reintubation
  • Tracheal extubation