Noninvasive Ventilation or CPAP for Postextubation Support in Small Infants

Abstract

BACKGROUND:

Infants with a high risk of extubation failure are often treated with noninvasive ventilation (NIV) or CPAP, but data on the role of these support modalities following extubation are sparse. This report describes our experience using NIV or CPAP to support infants following extubation in our pediatric ICUs (PICUs).

METHODS:

We performed a retrospective study of children < 10 kg receiving postextubation NIV or CPAP in our PICUs. Data on demographics, medical history, type of support, vital signs, pulse oximetry, near-infrared spectroscopy (NIRS), gas exchange, support settings, and re-intubation were extracted from the electronic medical record. Support was classified as prophylactic if planned before extubation and rescue if initiated within 24 h of extubation. We compared successfully extubated and re-intubated subjects using chi-square test for categorical variables and Mann-Whitney test for continuous variables.

RESULTS:

We studied 51 subjects, median age 44 (interquartile range 0.5–242) d and weight 3.7 (3–4.9) kg. There were no demographic differences between groups, except those re-intubated were more likely to have had cardiac surgery prior to admission (0% vs 14%, P = .040). NIV was used in 31 (61%) and CPAP in 20 (39%) subjects. Prophylactic support was initiated in 25 subjects (49%), whereas rescue support was needed in 26 subjects (51%). Twenty-two subjects (43%) required re-intubation. Re-intubation rate was higher for rescue support (58% vs 28%, P = .032). Subjects with a pH < 7.35 (4.3% vs 42.0%, P = .003) and lower somatic NIRS (39 [24–56] vs 62 [46–72], P = .02) were more likely to be re-intubated. The inspiratory positive airway pressure, expiratory positive airway pressure, and F_IO2 were higher in subjects who required re-intubation.

CONCLUSIONS:

NIV or CPAP use was associated with a re-intubation rate of 43% in a heterogeneous sample of high-risk infants. Acidosis, cardiac surgery, higher F_IO2, lower somatic NIRS, higher support settings, and application of rescue support were associated with the need for re-intubation.

Introduction

Infants admitted to the pediatric ICU (PICU) receive invasive mechanical ventilation for a variety of indications.¹ Patients thought to have an increased risk of extubation failure usually receive some form of noninvasive respiratory support at extubation, such as high-flow nasal cannula (HFNC), CPAP, and noninvasive ventilation (NIV). Noninvasive respiratory support may be used prophylactically or as rescue following liberation from invasive mechanical ventilation.¹

Many studies to date have combined NIV and CPAP as a single category but have not demonstrated differences in the subsequent need for re-intubation.^2–6 A recent large randomized non-inferiority trial in children age 0–15 y, whom the treating clinician felt needed noninvasive respiratory support within 72 h of extubation, compared CPAP to HFNC. There were no differences in re-intubation rate, hospital length of stay (LOS), and PICU LOS, although subjects in the CPAP group were liberated from support 5 h sooner.⁶ Subgroup analysis revealed that children < 12 months of age seemed to benefit the most from CPAP.⁶ Based largely on that trial,⁶ current ventilator liberation guidelines recommend the use of CPAP over HFNC in children < 12 months of age requiring respiratory support at extubation (conditional recommendation).⁷

There are limited data on the use of rescue NIV for postextubation respiratory failure in children,⁸ and rescue NIV is not currently recommended in adults.⁹ In addition to a dearth of supportive data, other challenges for the utilization of NIV in small children include its poor tolerance, lack of pediatric-approved devices, and paucity of interfaces specifically designed for children.¹⁰ Current dedicated NIV devices approved for use in children are limited to those weighing > 5 kg or > 10 kg, depending on the circuit and ventilator. Despite these barriers, CPAP and NIV are still used in small children thought to be at high risk of extubation failure or necessitating more support than what can be achieved by standard oxygen therapy or HFNC.¹¹ The purpose of this report is to describe our experience using NIV or CPAP for postextubation support in subjects weighing < 10 kg in our PICUs.

QUICK LOOK.

Current knowledge

Infants with a high risk of extubation failure are often treated with noninvasive ventilation (NIV) or CPAP. NIV and CPAP are challenging in infants < 10 kg due to lack of high-quality data, poor tolerance, lack of pediatric-approved devices, and paucity of interfaces specifically designed for children. Despite these challenges, CPAP and NIV are still used in infants at high risk of extubation failure.

What this paper contributes to our knowledge

NIV and CPAP were associated with a high risk of re-intubation. Several risk factors for re-intubation were identified in univariate analysis, including acidosis, prior cardiac surgery, higher F_IO2, lower somatic near-infrared spectroscopy, and higher NIV or CPAP settings. Rescue support was also associated with a higher risk of re-intubation. There were no differences in successfully extubated versus re-intubated subjects for vital signs, interfaces, devices used, or demographics.

Methods

We performed a retrospective review of all subjects weighing < 10 kg who received NIV or CPAP in our PICU or pediatric cardiac ICU (PCICU) between July 1, 2017–May 31, 2021. This study was approved by the Duke University Institutional Review Board with a waiver of consent. We used the Duke Enterprise Data Unified Content Explorer (DEDUCE) and Epic Clarity to identify the cohort and obtained data from the electronic health record. DEDUCE is a self-service web-based electronic health query system.¹² Epic Clarity is a Microsoft Structured Query Language Server database that contains a large subset of health data extracted from Duke Maestro Care (Epic) application. Data extracted from DEDUCE and Epic Clarity were then uploaded into a REDCap database (hosted at Duke University Medical Center, Durham, North Carolina), and additional variables were extracted manually from the medical record by trained respiratory therapists (RTs). Patients were excluded if they had a diagnosis of sleep-disordered breathing; if NIV or CPAP was started outside the ICU; or if they used NIV, CPAP, or oxygen at home. Subjects were enrolled once per hospital admission, and only the initial course of noninvasive respiratory support was included.

During the study period, NIV or CPAP was managed via a RT-driven protocol. Our primary form of postextubation support is HFNC, with 6% of subjects in our PCICU¹¹ and 11% (unpublished data) in our PICU being extubated directly to NIV or CPAP. The decision to prophylactically extubate to NIV or CPAP is not managed via protocol but is determined by the clinical team and generally is reserved for those known or suspected to carry a high risk of re-intubation. NIV was defined as any modality in which 2 levels of pressure are utilized, regardless of device or interface used. Relative exclusions to receive NIV or CPAP under our protocol included respiratory arrest, clinical instability, inability to protect the airway, lack of cooperation, excessive secretions, recent facial surgery, facial deformities or trauma, and inability to obtain an adequate seal with the interface.

The goals of the NIV and CPAP protocols are to maintain a pH of 7.30–7.40, P_aO2 of 55–80 mm Hg, and S_pO2 of 88–95%. These goals can be modified based on physician order and can be adjusted based on patient-specific factors, such as cyanotic heart disease or single-ventricle physiology. In brief, our protocol calls for setting expiratory positive airway pressure (EPAP) or CPAP initially at 6 cm H₂O and adjust as needed for oxygenation goals in conjunction with F_IO2. Inspiratory positive airway pressure (IPAP) is adjusted to maintain at tidal volume (V_T) 6–8 mL/kg. The backup rate is set at half of the observed spontaneous rate but may be set higher in patients with difficulty triggering the ventilator. CPAP was delivered via bubble, AirLife (Cardinal Health, Dublin, Ohio), Avea (Vyaire Medical, Mettawa, Illinois), Trilogy (Philips, Amsterdam, the Netherlands), or V60 devices (Philips). NIV was delivered via Trilogy, V60, Avea, or Servo-u (Getinge, Göteborg, Sweden) devices. For subjects < 4 kg, NIV could also be delivered by a critical care ventilator (Viasys [Avea] or Servo-u) via nasal mask or prongs. Initial settings were suggested to be a set frequency of 30 breaths/min, IPAP < 18 cm H₂O, and PEEP 5–6 cm H₂O. When subjects were in the weaning phase, the frequency and IPAP were reduced until the subject was ready to transition to CPAP or HFNC. The Avea ventilator does not allow patient triggering while on NIV and was only used with a nasal mask. During the study period, we did not use a full face mask interface with a critical care ventilator, and its use was exclusively with the V60 and Trilogy.

We collected data on demographics, medical history, indication for hospital admission, location of origin, location where respiratory support was initiated, receipt of cardiac surgery during stay, need for intubation, ICU LOS, hospital LOS, and survival. Subjects who underwent cardiac surgery were classified using the Society of Thoracic Surgeons - European Association for Cardio-Thoracic Surgery Congenital Heart Surgery Mortality (STAT) categories. The duration of mechanical ventilation prior to extubation, the total number of extubation readiness tests (ERTs), and number of failed ERTs were recorded and grouped into categories of 0, 1, 2, and ≥ 3 ERTs. ERTs were performed via an RT-driven protocol. All mechanically ventilated subjects were screened daily by RTs based on predefined criteria; in brief, these include spontaneous respiratory drive, PEEP ≤ 7 cm H₂O, Δ pressure ≤ 20 cm H₂O, F_IO2 0.50, and hemodynamic stability. ERT is performed with a fixed pressure support of 5 cm H₂O and a PEEP of 5 cm H₂O. The details of our protocol have been published elsewhere.¹¹

Specific NIV and CPAP data included indication for support, type (NIV or CPAP), interface, device, and settings used. Subjects were classified based on the initial support received. Settings were reported as initial, maximum, and final. Final settings were the last documented prior to re-intubation or liberation from NIV or CPAP. We also collected heart rate, breathing frequency, mean arterial pressure, S_pO2, F_IO2, and near-infrared spectroscopy (NIRS) measurements immediately before NIV or CPAP was started, then at 1, 2, 4, 6, 12, and 24 h following initiation. NIRS monitoring is standard of care in our PCICU for all patients and is used occasionally in the PICU. Blood gas data were collected if available within 12 h of NIV or CPAP initiation. Arterial, venous, and capillary blood gases were combined for analysis. Blood gases were classified as acidemia (pH < 7.35) and respiratory acidosis (pH < 7.35 and P_CO2 > 45 mm Hg).

The primary outcome was need for endotracheal intubation for any reason. Secondary outcomes included ICU and hospital LOS. We compared those who required intubation with those who did not. Continuous outcomes were compared using the Mann-Whitney rank-sum test, and categorical data were compared using the chi-square test. Statistical significance was set at < 0.05, and analyses were run with SPSS V25 (IBM, Armonk, New York).

Results

We identified 145 patients who received NIV or CPAP, of which 51 received support following extubation and were included in the study. Twenty-two (43.1%) subjects required re-intubation, and all survived to hospital discharge. There were no differences for demographics, location, race, cardiac surgery during stay, STAT category, primary interface, device, or need for continuous sedation between subjects who were successfully extubated and those who required re-intubation (Table 1).

Table 1.

Demographics and Medical History

Prior to extubation, there were no differences in time on mechanical ventilation, total number of ERTs, failed ERTs, primary interface, device used, aerosolized medications used, or subjects who had a blood gas within 12 h (Table 2). Forty-two (82.4%) subjects had a blood gas within 12 h of support initiation, and there were no differences between groups. Subjects with an acidemia from any cause (42.1% vs 4.2%, P = .003) or respiratory acidosis (36.8% vs 0%, P = .001) were more likely to be re-intubated. All 7 subjects with a respiratory acidosis prior to NIV or CPAP required re-intubation (Table 2). Of these 7 subjects, all 7 were on HFNC prior to NIV or CPAP, 6 (85.7%) were post cardiac surgery, 5 (71.4%) treated with NIV, and all 7 were in the PCICU.

Table 2.

Extubation Readiness Test, Noninvasive Ventilation or CPAP, and Gas Exchange Data

Prior to initiation of NIV or CPAP, there were no differences between groups for heart rate, breathing frequency, blood pressure, F_IO2, or NIRS (Table 3). Half of the 22 subjects who were re-intubated within 4 h of initiation and nearly three quarters by 24 h. The F_IO2 was higher at 1 and 2 h, and somatic NIRS was lower after 1 h in those who were re-intubated. Breathing frequency was lower at 4, 6, 12, and 24 h after re-intubation (Table 3). There were no differences between groups for hospital LOS, ICU LOS, initial IPAP, maximum IPAP, initial EPAP/CPAP, maximum EPAP/CPAP, initial backup rate, maximum backup rate, or final backup rate and final F_IO2 (Table 4). Those intubated had higher final IPAP, final EPAP, initial F_IO2, and maximum F_IO2.

Table 3.

Vital Signs and Oxygenation Data in the First 24 Hours Postextubation

Table 4.

Outcomes and Noninvasive Support Settings

Discussion

We report an overall postextubation failure rate of 43% following the use NIV or CPAP in children < 10 kg who were either considered to be at high risk of re-intubation or needed rescue support. Prophylactic NIV or CPAP had a lower re-intubation rate when compared to rescue support. Cardiac surgery prior to admission, acidosis, higher F_IO2, lower somatic NIRS, and higher final IPAP/EPAP were associated with need for re-intubation. There were no differences in hospital or ICU LOS. Vital signs were similar between extubation success and failure except for breathing frequency after 4 h was lower in subjects requiring intubation. This difference was due to 50% of re-intubated subjects being re-intubated at 4 h; thus, the lower breathing frequency was from neuromuscular blockade and/or sedation used to facilitate invasive mechanical ventilation. Final IPAP and EPAP were higher in subjects who were re-intubated but still lower than the maximum allowed by our protocol.

Predicting extubation failure for infants supported on NIV or CPAP is a significant challenge. Surprisingly, we did not find any differences in vital signs between groups except for the high failure rate of subjects with acidosis. However, the F_IO2 was higher after 1 and 2 h in subjects who eventually needed to be re-intubated, suggesting that the inability to decrease the F_IO2 and/or the development of acidosis following initiation of postextubation or rescue noninvasive support should raise the level of concern. We noted that all 7 subjects with respiratory acidosis prior to support initiation required intubation; thus, rescue support should be initiated cautiously in this population, although this needs confirmation in a larger sample size. Other studies have also identified NIRS as a potential predictor of extubation failure, primarily in children post cardiac surgery.^13–15 These studies implicate NIRS as potential tool to evaluate patients at risk of extubation failure but did not report outcomes for those treated with NIV or CPAP separately. The role of NIRS as an additional monitoring tool in non–cardiac surgery patients should be investigated.

NIV settings were slightly higher in subjects who required re-intubation but below the maximum values allowed within our protocol. Given the retrospective nature of our study, we were unable to determine why NIV settings were not maximized prior to re-intubation. We speculate that this could be related to clinician reticence to use high inspiratory pressure in infants due to concerns of gastric insufflation, increased leak around the interface, worsening of patient-device synchrony, treatment intolerance, or rapid deterioration of the clinical condition that precluded further escalation of NIV settings. It is also possible that the V_T was within 6–8 mL/kg, although that is less likely due to the relative inaccuracy of V_T measurements during NIV or CPAP in small infants.

NIV or CPAP is used relatively infrequently in our facility, with HFNC being our primary form of postextubation support.¹⁶ We generally reserve prophylactic postextubation NIV or CPAP for patients at high risk, most commonly in the setting of cardiac surgery, and for those with prolonged mechanical ventilation, prior extubation failure, and multiple failed ERTs. A large, multi-center non-inferiority trial comparing HFNC to CPAP in children found a reduced time on noninvasive respiratory support following extubation with CPAP but no difference in re-intubation rates.¹⁷ However, a subgroup analysis revealed a decreased need for re-intubation in subjects < 12 months of age assigned to CPAP.¹⁷ The higher re-intubation rate observed in our study is likely reflective of our subjects being younger, sicker, and being considered high risk by the clinical team.

Despite the high re-intubation rate, we did not observe a difference in hospital or ICU LOS, in contrast to prior studies in which re-intubation was associated with longer ICU LOS, hospital LOS, and mortality.^18–20 When compared to an older, general PICU population, the hospital and ICU LOS in our high-risk infant sample were longer than previously reported^18,19 but similar to that observed by Miura et al²⁰ in a cohort of infants post cardiac surgery. Our results may indicate re-intubation after failing NIV or CPAP is not a risk factor for prolonged hospital stay; however, this may also be explained by the large percentage of infants post cardiac surgery, who may remain hospitalized for non-respiratory reasons such as feeding intolerance, family education, and other factors. Other studies have also noted re-intubation to be associated with mortality,²¹ which we did not note in this cohort.

Another potential explanation for our re-intubation rate was the high proportion (62%) of children recovering from complex cardiac surgery. Preventing re-intubation in this population is critical as extubation failure is associated with an increased risk of adverse events, including cardiac arrest,²⁰ and in-patient mortality.²¹ Our general practice is to extubate these patients to HFNC unless there are significant risk factors for re-intubation.¹¹ Other studies have noted similar re-intubation rates to our study, with a rate of 38% reported by Rolim et al.⁵ Similarly to our study, they also reported that maximum EPAP and F_IO2 were associated with re-intubation.⁵ A large database study in children following cardiac surgery found a 17% re-intubation rate for all encounters but did not report neonatal outcomes separately, and the settings, interfaces, and devices were not described.⁴ Other studies have reported failure rates of 21%²² and 40%²³ using nasopharyngeal tubes as their primary interface with a critical care ventilator. A study that primarily used nasal CPAP noted a re-intubation rate of 22%.²⁴ The differences in reported re-intubation rates are likely explained by differences in clinical practice between centers, criteria for NIV or CPAP use, and subject characteristics but could also be related to differences in devices and interfaces used. This is an important area of future study as it is currently unclear which infants are more likely to benefit from the various types of postextubation respiratory support (ie, HFNC, CPAP, or NIV) and whether prophylactic versus rescue support is preferred. It is also unclear whether a dedicated NIV device or a critical care ventilator should be used to deliver NIV and which interface(s) should be used.

CPAP has been the predominant postextubation support strategy reported in infants, and there are limited data comparing NIV versus CPAP. A small crossover study of infants recovering from cardiac surgery noted NIV with neurally-adjusted ventilatory assist (NAVA) resulted in high levels of synchrony and reduced diaphragm work compared to CPAP.²⁵ This study indicates that NIV may have some advantages compared to CPAP provided patients are synchronous with both triggering and cycling criteria; whether this translates into improvement in patient-oriented outcomes is yet to be determined. A single-center randomized controlled trial compared nasal CPAP to nasal NIV in subjects post low-risk cardiac surgery and noted no differences in re-intubation between groups (15% for NIV vs 12% for CPAP), but subjects receiving NIV exhibited improved gas exchange.²⁶ Those subjects were predominantly recovering from low-risk cardiac surgery, a cohort for which we primarily use HFNC while achieving similar re-intubation rates.¹¹ Comparing NIV to CPAP or HFNC is an important area of study as a recent randomized controlled trial in adults with very high risk for extubation failure found a lower need for re-intubation for NIV compared to HFNC.²⁷ There were no differences between NIV and CPAP in our study. However, there was likely selection bias in both the decision to use NIV and which interface was chosen.

The high rate of extubation failure reported in our study and others could also be related to ventilator asynchrony. Synchrony, particularly triggering, poses a challenge during NIV in infants due to leaks and trigger sensitivity that vary depending on type of device and interface used.²⁸ NAVA has been shown to improve triggering and reduce work of breathing when used in combination with a total face mask in infants and small children.²⁹ A case-control study of infants after cardiac surgery noted no difference in re-intubation with NAVA compared to HFNC, with 26% treated with NAVA requiring re-intubation, although those treated with NAVA were sicker and spent longer on invasive mechanical ventilation.³⁰ Important areas of future study are the evaluation of work of breathing and triggering with different interfaces (total face, oronasal, and nasal) and comparing dedicated NIV machines to critical care ventilators.

Limitations

Our study has several limitations. As a retrospective study, we were limited to data available within the electronic medical record. Whereas NIV and CPAP are managed via an RT-driven protocol, the decision to place a patient on each modality is not part of the protocol and thus subject to selection bias. The high proportion of subjects post cardiac surgery may limit the generalizability to other patient populations. Due to a high proportion of subjects with S_pO2 > 97%, we were unable to include S_pO2/F_IO2. In some cases, total face mask NIV was not possible due to subject size or their facial features. We also used different NIV or CPAP devices, which could have influenced the results. Due to our limited sample size and missing data, we were unable to perform logistic regression to identify factors associated with the need for re-intubation.

Conclusions

The use of NIV or CPAP following extubation was associated with a re-intubation rate of 43% in this heterogeneous sample of high-risk infants with higher re-intubation rates for rescue support compared to prophylactic use. Acidosis, cardiac surgery, higher F_IO2, lower somatic NIRS, higher support settings, and application of rescue NIV or CPAP were associated with re-intubation.