High-Risk Extubation Readiness Testing for Children With Cardiac Critical Illness

Abstract

BACKGROUND:

A protocolized extubation readiness test (ERT), including a spontaneous breathing trial (SBT), is recommended for patients who are intubated. This quality-improvement project aimed to improve peri-extubation outcomes by using a high-risk ERT protocol in intubated cardiac patients in addition to a standard-risk protocol.

METHODS:

After baseline data collection, we implemented a standard-risk ERT protocol (pressure support plus PEEP), followed by a high-risk ERT protocol (PEEP alone) in cardiac subjects who were intubated. The primary outcome, a composite of extubation failure and rescue noninvasive respiratory support, was compared between phases. Ventilator duration and use of postextubation respiratory support were balancing measures.

RESULTS:

A total of 213 cardiac subjects who were intubated were studied, with extubation failure and rescue noninvasive respiratory support occurring in 10 of 213 (4.7%) and 8 of 213 (3.8%), respectively. We observed a reduction in the composite outcome among the 3 consecutive phases (5/29 [17.2%], 10/110 [9.1%] vs 3/74 [4.1%]; P = .10), but this did not reach statistical significance. In the logistic regression model when adjusting for admission type, the high-risk ERT protocol was associated with a significant reduction of the composite outcome (adjusted odds ratio 0.20, 95% CI 0.04-0.091; P = .037), whereas the standard-risk ERT protocol was not (adjusted odds ratio 0.48, 95% CI 0.15–1.53; P = .21). This was not accompanied by a longer ventilator duration (2.0 [1.0, 3.0], 2.0 [1.0–4.0], vs adjusted odds ratio 2.0 [95% [1.0–6.0]; P = .99) or an increased use of planned noninvasive respiratory support (10/29 [35.5%], 35/110 [31.8%], vs 25/74 [33.8%]; P > .99).

CONCLUSIONS:

In this quality-improvement project, a high-risk ERT protocol was implemented with improvement in peri-extubation outcomes among cardiac subjects.

Introduction

A protocolized extubation readiness test (ERT), including a spontaneous breathing trial (SBT) is recommended for all patients who are intubated. ¹ For patients who are at standard risk of extubation failure, pressure support augmentation with PEEP, or PEEP alone during SBT is recommended. ¹ For patients at high risk of extubation failure, PEEP alone without pressure support augmentation was recommended. ¹ Physiologic studies have shown that pressure support augmentation significantly underestimates postextubation work of breathing, which may result in premature extubation and increased extubation failure rate. ¹ Although this has not been corroborated in clinical trials, the avoidance of extubation failure is of utmost importance, therefore, the use of a more stringent PEEP-only SBT in patients at high risk is recommended. ¹

Children with cardiac illness are considered at high risk of extubation failure due to abrupt changes to cardiopulmonary interactions that result from the transition from positive-pressure ventilation (PPV) to spontaneous breathing. Mechanical ventilation does more than support the respiratory system in cardiac patients, PPV has direct effects on cardiac function, pulmonary compliance, and pulmonary vascular resistance, which are significant causes of extubation failure in this group of patients. ² However, there is a scarcity of studies on liberation from mechanical ventilation for children with congenital heart disease before or after surgical repair, myocarditis-cardiomyopathy, or other congenital or acquired cardiac conditions. The largest published pediatric randomized controlled trial by Foronda et al ³ on daily evaluation and the impact of SBT on the duration of mechanical ventilation excluded children with congenital heart disease due to their complexity. Extubation in this population can be extremely challenging due to limited cardiopulmonary reserve. ² To date, there is no clear consensus on the best SBT method for children with cardiac illness. ¹

We recently performed a quality-improvement project by applying a standard-risk ERT protocol for all the patients who were intubated in our multidisciplinary pediatric ICU and demonstrated improvement in extubation failure rates. ⁴ However, data from the standard-risk ERT quality improvement indicated cardiac compromise as the main reason for extubation failure (70%) and, therefore, we developed a high-risk ERT protocol for cardiac patients. In line with the mechanical ventilation liberation guidelines, ⁵ our ERT protocol was modified to ensure that patients at high risk were not over-supported during the SBT. ⁵ Our modified ERT protocol uses pressure support of 0 cm H₂O for patients at high risk of extubation failure. This current quality-improvement project aimed to improve extubation failure and rescue noninvasive respiratory support rates by applying a more stringent high-risk ERT protocol in cardiac patients who are intubated, without prolonging the duration of mechanical ventilation.

Quick Look.

Current Knowledge

The use of ERTs has improved outcomes in intubated pediatric patients at standard risk of extubation failure. However, there is limited evidence on the use of ERTs in the subgroup of children with cardiac illness who are at higher risk of extubation failure due to their medical complexity. PEEP without pressure support augmentation during SBTs has been suggested to more closely resemble the work of breathing at extubation and provide a better assessment of extubation readiness.

What This Paper Contributes to Our Knowledge

A high-risk ERT protocol with a PEEP-only SBT was utilized for intubated patients with cardiac illness. The high-risk protocol resulted in reduced extubation failure and rescue noninvasive respiratory support as compared to the standard-risk protocol with pressure support.

Methods

Quality-Improvement Project Design

The project protocol was reviewed and exempted from institutional review board approval because it was a quality-improvement project. We adopted the “Model for Improvement” quality-improvement process to achieve the aims of this study. ⁴ We used three plan-do-study-act cycles over 3.5 years. The outcome measures included the following: (1) development of a high-risk ERT that was acceptable and safe, (2) performance of an ERT in > 70% of all cardiac subjects on mechanical ventilation, and (3) reduction in the composite outcome of an extubation failure and the need for rescue noninvasive respiratory support. Balancing measures included the following: (1) reduction or maintenance of mechanical ventilation duration, and (2) use of planned noninvasive respiratory support. This high-risk ERT quality-improvement initiative included all cardiac (surgical and non-surgical) subjects who were admitted to the pediatric ICU from January 2020 to June 2023 and who required mechanical ventilation regardless of the duration of ventilation and subject’s age. Exclusion criteria project included patients (1) with a tracheostomy tube in place before pediatric ICU admission, and (2) those who died before extubation.

Protocol Development and Implementation

The first plan-do-study-act cycle spanned 7 months, from January 2020 to July 2020, and involved baseline data collection and study of subjects without an ERT (termed phase 0). A respiratory therapist–driven standard-risk ERT protocol was developed for the management of pediatric subjects on conventional invasive mechanical ventilation, as previously described. ⁴ In brief, this standard-risk ERT protocol included a checklist for extubation readiness and a 2-h SBT by using pressure support inversely proportional to endotracheal tube (ETT) size; data were collected for 22 months from August 2020 to May 2022 for subjects on this standard-risk ERT protocol (this second plan-do-study-act cycle was termed phase I). This plan-do-study-act cycle revealed that the clinical group with the highest extubation failure rates were the cardiac subjects. Hence, further modification of this ERT by using a PEEP-only SBT (limited to a maximum of 6 cm H₂O) was done for the cardiac subjects at high risk, and data were collected for 12 months, from June 2022 to May 2023 (this third plan-do-study-act cycle was termed phase II). Phase I took place over a prolonged period of time due to the reduction of elective cardiac surgical cases during the COVID-19 pandemic. Implementation measures were applied as previously described. ⁴

In the high-risk ERT protocol, subjects who fulfil the criteria on the ERT checklist were be exposed to an SBT. In the first 60 min of the SBT, subjects utilized PEEP with pressure support augmentation inversely proportional to the ETT size. If tolerated, then the pressure support augmentation was removed, which left the subject on PEEP (maximum 6 cm H₂O) alone for another 30 to 60 min. This stepwise method was done to identify subjects who may be clinically destabilized early due to the study procedure and to achieve greater acceptance by clinical and nursing stakeholders. The ERT was considered passed if the subject tolerated the entire procedure. Subjects were ventilated with either the Hamilton G5 (Hamilton Medical, Bonaduz, Switzerland) or the SLE6000 (SLE, South Croydon, United Kingdom) during their pediatric ICU stay and were required to be on minimum ventilator settings; Δ P of a maximum 10 cm H₂O with low PEEP of a maximum 6 cm H₂O and ${\text{F}}_{{\text{IO}}_{\text{2}}}$ <0.40 before the initiation of SBT. ^{4 , 6} Similarly, subjects were monitored closely during the SBT and a blood gas was assessed after completion of the SBT. Subjects that failed the SBT were returned to previous ventilator settings, whereas those who passed the SBT remained on SBT ventilator settings until extubation. Although the ERT and SBT findings were communicated to the managing clinical team, the timing for extubation and the need for postextubation respiratory support was at the discretion of the attending physician. ⁴

Data Collection

Data collection was performed prospectively by the quality-improvement team using a standardized data collection tool that included subjects’ demographics (age, sex, weight), clinical characteristics (admission diagnosis, comorbidities), and outcomes. Comorbidities were defined according to the list of chronic complex conditions. ⁷ Extubation failure was defined as the need for non-elective re-intubation within 48 h of extubation. Rescue noninvasive respiratory support was defined as noninvasive respiratory support that was not planned for before extubation or the need to escalate respiratory support to noninvasive respiratory support within 48 h of extubation. Rescue noninvasive respiratory support included the use of any bi-level positive airway pressure, CPAP, or heated high-flow nasal cannula.

Statistical Analysis

Categorical and continuous variables were presented as counts (percentages) and median (interquartile range), respectively, in phase 0, I, and II groups. Statistical comparisons of subject characteristics were made among phase 0, I, and II groups using the Fisher exact test and the Kruskal-Wallis test, as appropriate. Extubation failure, the need for rescue noninvasive respiratory support, and the composite of both extubation failure and rescue noninvasive respiratory support (primary outcome) were compared among all phases. In subjects who required both rescue noninvasive respiratory support and who eventually had a failed extubation, the composite outcome was counted as one event (extubation failure). Logistic regression to determine the association between the ERT protocol for the primary outcome was performed, adjusting for admission type (cardiac surgical vs cardiac non-surgical), and reported as the adjusted odds ratio with 95% CI.

The ERT protocol variable was categorized into none (phase 0 [reference]), standard risk (phase I), and high risk (phase II). The covariate admission type was determined a priori as cardiac surgical (mostly elective) and cardiac non-surgical (mostly non-elective) patients have different clinical profiles and may have a different risk of extubation failure or rescue noninvasive respiratory support. Moreover, there was an anticipated reduction in cardiac surgical admissions during the COVID-19 pandemic (phase I) and statistical adjustment was deemed necessary. In addition to standard statistical analysis, outcomes were analyzed in statistical process control charts. X-bar control charts were used, which included the center line (mean) and upper and lower control limit (± 3 SD above and below the center line. Statistical analysis was performed using STATA software, version 15.1 (StataCorp, College Station, Texas) and control charts were performed using Microsoft Excel version 16.77.1 (Microsoft, Redmond, WA).

Results

A total of 365 children with cardiac conditions were admitted to the pediatric ICU during the course of this quality-improvement initiative between January 2020 and June 2023. Of these, 232 required intubation and 213 subjects were included in the analysis after applying inclusion-exclusion criteria (Fig. 1). The overall median (interquartile range) age of the cohort was 0.8 (0.3–4) years. There were 37 of 213 (17.4%) and 176 of 213 (82.6%) cardiac and cardiac surgical cases, respectively (Table 1). No differences in demographic and clinical characteristics were observed among the subjects included in the 3 quality-improvement phases, except that there were fewer subjects with comorbidities in phase 0 compared with the later phases (4/29 [13.8%], 77/110 [70.0%], vs 46/74 [62.2%]; P < .001).

Fig. 1.

Flow chart.

Table 1.

Characteristics of Subjects in the Baseline, Standard-Risk, and High-Risk ERT Groups

ERT adherance was 79 of 110 (71.8%) and 56 of 74 (75.7%) in phases I and II, respectively. Extubation failure progressively decreased in each consecutive phase (4/29 [13.8%], 4/110 [3.6%] vs 2/74 [2.7%]; P = .08) but this was not statistically significant. The cause of extubation failure was cardiac failure in 6, respiratory failure in 2, and arrhythmia and pulmonary hypertension in one subject, respectively. There were only 4 occasions in which a subject had a failed ERT just before extubation, but, due to clinical consensus, was extubated anyway. These 4 subjects were from phase I and one subject had an extubation failure. Use of rescue noninvasive respiratory support also decreased in phase II (1/29 [3.5%], 6/110 [5.5%] vs 1/74 [1.4%]; P = .35); however, this was not statically significant. Extubation failure was higher in subjects who required rescue versus planned noninvasive respiratory support (3/11 [27.3%] vs 4/61 [6.6%]; P =.004). The composite outcome progressively decreased in each consecutive phase (5/29 [17.2%], 10/110 [9.1%] vs 3/73 [4.1%]; P = .10), but this too was not statically significant (Fig. 2). After adjusting for the admission category, the high-risk ERT protocol was independently associated with a decreased odds of extubation failure and rescue noninvasive respiratory support (adjusted odds ratio 0.20, 95% CI 0.04–0.09; P = .037), but the standard-risk ERT was not (adjusted odds ratio 0.48, 95% CI 0.15–1.53; P = .21) (Table 2).

Fig. 2.

Control chart for (A) extubation failure, (B) rescue NRS and (C) composite of both over the three phases of study. Solid line indicates event rate; Dashed line = center line (CL) reflects the mean; Dotted lines = upper (UCL) and lower (LCL) control limits is calculated by three times standard deviation above and below the center line, respectively.

Table 2.

Uni- and Multivariate Logistic Regression for Composite of Extubation Failure and Rescue Noninvasive Respiratory Support

The total noninvasive respiratory support used (including both rescue and planned noninvasive respiratory support) remained the same across all 3 phases (10/29 [34.5%], 35/110 [31.8%] vs 25/74 [33.8%]; P > .99). Ventilator duration also did not change over the 3 consecutive phases (2.0 [1.0–3.0], 2.0 [1.0–4.0] vs 2.0 [1.0–6.0] days; P = .99). For subjects who underwent the high-risk ERT, all tolerated the SBT with pressure support augmentation and spent a median (interquartile range) of 40 (30–48.8) min on the PEEP-only SBT. None of the subjects were destabilized while on the high-risk ERT protocol and no adverse events were reported.

Discussion

Our results demonstrated that an ERT protocol can be successfully implemented with satisfactory staff adherance and sustained. In this quality-improvement study, the implementation of the ERT protocol reduced extubation failure without prolonging the length of ventilation, specifically, the high-risk ERT protocol (PEEP only) was associated with an improvement in the composite odds of extubation failure and rescue noninvasive respiratory support (adjusted odds ratio 0.20, 95% CI 0.04-0.091; P = .037) among cardiac subjects in multivariable analysis. Extubation failure and rescue noninvasive respiratory support rates progressively decreased in each consecutive phase, while the total noninvasive respiratory support used and ventilator duration remained the same. Hence, the reduction in extubation failure and rescue noninvasive respiratory support was likely not due to an increase in planned noninvasive respiratory support use or extended duration of mechanical ventilation.

Patients with cardiac conditions tend to be at higher peri-extubation risk due to poor tolerance toward the abrupt changes in airway and intrathoracic pressures after transition from PPV to spontaneous breathing. ⁵ During PPV, the right-ventricle preload and left-ventricle afterload decrease, thus lowering left-ventricle filling pressure and limiting pulmonary edema. ⁸ As such, PPV will greatly benefit those with left-sided conditions, for example, left-to-right shunts causing pulmonary over-circulation and left-ventricle dysfunction. ⁸ For such patients, extubation may precipitate and/or aggravate pulmonary edema, manifesting as increased work of breathing and contribute to extubation failure or the need for rescue noninvasive respiratory support. ⁵ Any upper-airway obstruction, by causing larger swings in intrathoracic pressure, will compound the work of breathing. In contrast, PPV may cause an increase pulmonary vascular resistance due to compression of intra-alveolar capillaries, thereby increasing right-ventricle afterload. ^{9 , 10} In this case, right-sided conditions benefit from early extubation provided sufficient time is granted to recover from the effects of cardiopulmonary bypass. Indeed, in the previous quality-improvement study conducted by our team, the majority of extubation failure and rescue noninvasive respiratory support consisted of patients with cardiac conditions. ⁴ Thus, SBTs in these patients should reflect the non-intubated condition as closely as possible to accurately assess readiness for extubation. In the current quality improvement, this was achieved by using a PEEP-only SBT protocol.

In pediatric critical care, the rate of extubation failure is reportedly between 5 and 15%. ¹¹ The use of an ERT has been associated with decreased mechanical ventilation duration in children without increasing the extubation failure rate or the need for noninvasive ventilation. ^{3 , 12} These trials use a pressure support level of 5–10 cm H₂O and PEEP of 5 cm H₂O in their SBT. However, most studies exclude children with cardiac disease due to their complexity and lack of data with regard to the most suitable ERT/SBT method in this cohort. ^{3 , 12} One randomized controlled trial (N = 110) found improved extubation success with the use of SBT versus no SBT in children after cardiac surgery. ¹³ The extubation failure rate in our cohort of cardiac subjects was ∼13% before implementation of the standard or high-risk ERT protocol, which is similar to previous studies. ^{2 , 3} The median age of our cohort was < 1 year, also consistent with previous studies. ^{2 , 3 , 13} The surgical complexity was represented by a good spread of cases that encompass different RACHS categories. Retrospective studies have shown that patients with congenital heart disease are more vulnerable to extubation failure include those who are of younger age, longer mechanical ventilation duration, and genetic syndromes. ^{11 , 14} With no added risk to the patients, we were able to reduce extubation failure rates comparable to other large multi-center studies. ¹⁵ Being a vulnerable population and knowing that an episode of extubation failure itself is associated with poorer outcomes, including higher mortality and increased hospital length of stay and ventilator duration as corroborated by previous studies, ^{2 , 15} it is doubly important to accurately assess readiness for extubation in cardiac patients. An ERT/SBT has a high sensitivity for identifying patients ready to be removed from invasive mechanical ventilation; however, the specificity is limited due to the low event rate. ¹⁶

Physiologic studies indicate that the use of pressure support plus PEEP SBT significantly underestimates the postextubation work of breathing. Higher levels of pressure support were not needed for smaller-diameter ETTs at physiologic inspiratory flows and may lead to excessive support during SBTs. ^{1 , 2} In contrast, there is also concern that the use of PEEP-only SBT may result in increased work of breathing, leading to an increased SBT failure rate and delayed extubation. ^{1 , 17} For a compromise, some investigators use variable pressure support based on ETT size, with the assumption that higher pressure is needed to overcome the resistance of smaller ETTs. ¹ In a population of children with congenital heart disease, a fixed pressure support of 5 cm H₂O versus variable pressure support based on ETT size did not seem to lead to improved extubation failure rates or mechanical ventilation duration. ² Although the debate on the issue of ETT resistance and the appropriateness of PEEP-only SBTs continues, our quality-improvement team has been accustomed to applying PEEP-only SBTs in our practice. In this quality improvement, the high-risk ERT/SBT was implemented for eligible patients in a protocolized manner regardless of the managing physician on service and showed beneficial results without any increase in adverse events.

The main limitation of this study was that, due to the single-center quality-improvement design, the results are poorly generalizable to other centers. This ERT protocol was also driven by respiratory therapists in the pediatric ICU, and some centers without respiratory therapists may not see the same effects. Another limitation was that the use of postextubation noninvasive respiratory support was based on physician’s discretion and not standardized. Based on the sample size and effect size in comparison between the phase 0 and II groups, the statistical power of this study is 74%, which exposes it to potential type-2 error. However, in view of the limited data concerning the ERT/SBT in high-risk populations, we believe reporting this work will be of value. Data on other important factors such as the use of sedatives and the practice of rehabilitation and/or early mobilization, which may impact on the peri-extubation outcomes, were not studied in this quality improvement, further limiting its generalizability. Due to personnel manpower constraint, there is currently no respiratory therapist coverage during nights and weekends, hence educating the nurses and physicians so that they are also able to perform the ERT will increase patient safety and ERT adherence. Patients who require rescue noninvasive respiratory support can also be further studied to identify potential risk factors for extubation failure.

Conclusions

A routine ERT/SBT followed by timely extubation should be adopted for all patients who are intubated as recommended by consensus guidelines. ¹ In this quality-improvement study, as opposed to a standard-risk ERT/SBT (with pressure support augmentation), a high-risk ERT/SBT (PEEP only) protocol was associated with a reduction in extubation failure and rescue noninvasive respiratory support among cardiac subjects.