Abstract
Rationale: Reverse triggering (RT) occurs when respiratory effort begins after a mandatory breath is initiated by the ventilator. RT may exacerbate ventilator-induced lung injury and lead to breath stacking.
Objectives: We sought to describe the frequency and risk factors for RT among patients with acute respiratory distress syndrome (ARDS) and identify risk factors for breath stacking.
Methods: We performed a secondary analysis of physiologic data from children on synchronized intermittent mandatory pressure-controlled ventilation enrolled in a single-center randomized controlled trial for ARDS. When children had a spontaneous effort on esophageal manometry, waveforms were recorded and independently analyzed by two investigators to identify RT.
Results: We included 81,990 breaths from 100 patient-days and 36 patients. Overall, 2.46% of breaths were RTs, occurring in 15/36 patients (41.6%). A higher tidal volume and a minimal difference between neural respiratory rate and set ventilator rate were independently associated with RT (P = 0.001) in multivariable modeling. Breath stacking occurred in 534 (26.5%) of 2,017 RT breaths and in 14 (93.3%) of 15 patients with RT. In multivariable modeling, breath stacking was more likely to occur when total airway Δpressure (peak inspiratory pressure − positive end-expiratory pressure [PEEP]) at the time patient effort began, peak inspiratory pressure, PEEP, and Δpressure were lower and when patient effort started well after the ventilator-initiated breath (higher phase angle) (all P < 0.05). Together, these parameters were highly predictive of breath stacking (area under the curve, 0.979).
Conclusions: Patients with higher tidal volume who have a set ventilator rate close to their spontaneous respiratory rate are more likely to have RT, which results in breath stacking >25% of the time.
Clinical trial registered with ClinicalTrials.gov (NCT03266016).
Keywords: acute respiratory distress syndrome, patient–ventilator asynchrony, reverse triggering, respiratory entrainment
Patient–ventilator asynchrony (PVA) has been associated with prolonged length of mechanical ventilation (1) and intensive care unit (ICU) and hospital mortality (2). Reverse triggering (RT) is as an underrecognized form of PVA that occurs when the patient has respiratory effort after a mandatory breath is initiated by the ventilator (3). RT can lead to double triggering, breath stacking, and higher-than-desired driving pressure (4), which may exacerbate ventilator-induced lung injury (VILI) (5).
RT likely occurs when the patient’s respiratory control system resets the respiratory rhythm such that the neural respiratory activity corresponds with machine inflations. This concept of respiratory entrainment has been described for years, and animal studies (6–8) suggest more entrainment occurs when machine-delivered tidal volume is higher and flow is low. Animal data also suggest that entrainment is mediated by the Hering-Breuer reflex, as it is lost after bilateral vagotomy, and is secondarily mediated by rapidly adapting receptors and vagal C fibers together with cortical and subcortical influence (5). Human data corroborate the importance of the Hering-Breuer reflex in the physiology of respiratory entrainment (9).
Contemporary data are lacking regarding the frequency with which RT occurs in critically ill patients, the risk factors for RT, and the frequency with which RT results in potentially injurious ventilator settings. These data are crucial to better understanding the magnitude of the problem, identifying high-risk patients, and ultimately devising effective strategies to prevent RT. We sought to describe the frequency with which RT occurs in pediatric patients with acute respiratory distress syndrome (ARDS), describe differences in characteristics between patients who have RT and those who do not, and identify factors that make it more likely for RT breaths to result in breath stacking.
Methods
We performed secondary analysis of physiologic data from esophageal pressure, airway pressure, and flow from ventilated children in an ongoing single-center randomized clinical trial testing a lung and diaphragm–protective ventilation strategy (REDVent [Real-time Effort Driven VENTilator Management], NIH/NHLBI R01HL124666, Clinical Trials.gov NCT03266016) (10) in the ICUs at the Children’s Hospital Los Angeles. On the basis of previous literature, we hypothesized that higher tidal volume and a set ventilator rate close to neural respiratory rate would be risk factors for RT and that breath stacking would be very common when RT occurred.
All patients were enrolled in the REDvent study, which included children from 1 month to 18 years of age who were within 48 hours of initiation of mechanical ventilation with pediatric ARDS (PARDS). Children were excluded if they had a contraindication to an esophageal catheter, severe lower airway obstruction, critical airway, intracranial hypertension, extracorporeal life support, do-not-resuscitate status, severe chronic respiratory failure, spinal cord injury, or cyanotic heart disease.
As part of the REDvent trial, children were randomized to usual-care ventilation versus ventilator management with a computerized decision-support tool to achieve lung and diaphragm–protective ventilation. Esophageal catheters were placed in all children and were used to measure the effort of breathing continuously for titration of ventilator support in the intervention arm and were used for once daily monitoring and data collection in the control arm. Synchronized intermittent mandatory ventilation (SIMV) pressure control–pressure support (PC-PS) was the mode of ventilation used in both arms. Physiologic waveforms of esophageal pressure (Pes), flow, and airway pressure were recorded daily for patients in both intervention and control arms. For this secondary analysis, we included the subset of children who had evidence of spontaneous respiratory effort on a given day, as determined by esophageal manometry. Data were recorded for 30 minutes, and no changes to ventilator settings were performed during this time.
Hardware and Software
Airway pressure was measured with a proximal sampling line together with a self-calibrating pneumotachometer (Variflex 51000–40094; Viasys). One of three esophageal pressure catheters were used on the basis of the size of the patient (Avea SmartCath 6,7, or 8Fr; Carefusion). The amount of air to be inflated into the esophageal balloon was determined using a previously validated calibration algorithm (11). All sensors were connected to a custom-built hardware device (New Life Box; Applied Biosignals), which recorded data at a frequency of 200 Hz. Data were subsequently postprocessed in a custom-built software program for analysis (designed in C# language).
Other clinical data include demographics, risk factors and severity of lung injury, respiratory parameters (observed respiratory rate, ventilator rate, spontaneous [i.e., neural] respiratory rate, peak inspiratory pressure (PIP), positive end-expiratory pressure (PEEP), ΔPes [the change in Pes], transpulmonary pressure [airway pressure − esophageal pressure], pressure × time product, pressure × rate product, blood gas data), sedation and analgesia (cumulative dose of drugs; State Behavioral Scale; and Face, Legs, Activity, Cry, Consolability scale). All doses of opioids were converted to fentanyl equivalents, which were represented as the cumulative dose per day (12, 13). We categorized each patient-day as “spontaneous breathing present” when at least 10% of the minute ventilation resulted from patient-triggered breaths. Because these patients are in pressure-control modes of ventilation, tidal volume will vary from breath to breath and will also be affected by asynchrony and RT. For this reason, we separated tidal volume into one of the following four groups for analysis: 1) mandatory ventilator breath without RT, 2) synchronous patient-triggered breath, 3) asynchronous patient-triggered breath, and 4) reverse-triggered breath.
PVA Definitions
All recordings were independently analyzed by two investigators (T.S. and B.Y.), and each breath was labeled as synchronous or asynchronous using definitions modified from previously validated definitions (14, 15) but that also incorporated esophageal pressure. If a breath was labeled asynchronous, then it was categorized into one of eight possible PVA events. Brief descriptions of the different types of PVA are highlighted here, with full definitions provided in Table E1 in the online supplement. The types include 1) RT (patient effort stimulated by lung inflation from a controlled ventilator breath), 2) ineffective triggering (patient’s spontaneous breathing efforts do not result in a ventilator trigger), 3) double triggering (the delivery of two consecutive ventilator insufflations during one single inspiratory effort) 4) autotriggering (the ventilator is triggered in the absence of a patient’s spontaneous breathing effort), 5) trigger delay (an excessive time between the initiation of the patient’s spontaneous breathing effort and the ventilator trigger), 6) flow asynchrony (mismatch between the ventilator flow and the patient’s demand on a patient-triggered breath), 7) delayed termination (Longer mechanical ventilator cycle times compared with the patient’s inspiratory time), and 8) premature termination (shorter mechanical ventilator cycle times compared with the patient’s inspiratory time).
Training and Interrater Variability
Upon initiation of this study, four authors simultaneously reviewed 10 patient files to collectively identify all the asynchrony events, classify them according to the established schema, and identify key areas of potential misclassification. The classification schema (definitions) were subsequently clarified, and visual examples were added to each for use during future review. Subsequently, two investigators independently reviewed all files, classified each breath as synchronous or asynchronous, and, when asynchronous, labeled the breaths according to the type of PVA. After completion of the files, the investigators subsequently met and identified breaths with discordance of classification. Typically, disputes were able to be resolved by the two investigators through discussion. Breaths that could not be resolved were highlighted, and all four investigators met again to discuss disputed breaths and arrive at a consensus interpretation. There were approximately 150 breaths (of nearly 81,000) that were not resolved by the two investigators and underwent discussion among the four investigators, and all breaths were ultimately able to be classified according to the described schema. The focus of this investigation was on RT, which was defined as patient effort that occurs after a mandatory breath from the ventilator (16). Patient effort was determined using the Pes waveform, mandating a minimum negative deflection of Pes of 2 cm H2O. In addition, if the RT breath resulted in the delivery of an additional breath from the ventilator, it was labeled as breath stacking, using previously published criteria (Figure 1) (17). For analysis, files that contained at least three RT breaths in the 30-minute recording were classified as having “RT.” The pattern of RT was classified as 1:1, 1:2, 1:3, or “other” on the basis of the ratio of breaths to RT breaths. A 1:1 pattern, for example, indicates that every mandatory ventilator breath resulted in an RT breath (Figure 3).
Figure 1.
The waveform of non–breath stacking (left) and breath stacking (right). Black arrows show airway pressure and flow at initiating spontaneous efforts. The breath on the left has reverse triggering (RT) without a breath stack but still generated a tidal volume of approximately 10 ml/kg (predicted body weight, 15 kg). The breath on the right has RT with a breath stack and resulted in a tidal volume of 16.8 ml/predicted body weight (right). Δesophageal pressure (Pes) shows the amount of change between peak and minimum Pes.
Figure 3.
The patterns of reverse triggering (RT) include irregular RT (no clear entrainment pattern), regular RT (clear entrainment pattern of 1:1). (A) Clear entrainment pattern of 1:1. (B) Irregular reverse triggering (no clear entrainment pattern). Red shadings show reverse triggers.
Any patient who had a daily file that met these criteria was classified as a patient with RT. Additional indices that were calculated included the asynchrony index (total asynchrony events/total breaths ×100) and RT index (RTI) (total RT breaths/total breaths ×100). Neural respiratory rate was defined as the time between consecutive patient inspiratory efforts (negative deflection of the Pes signal) divided by 60 seconds. In addition, to compare the relationship between the time the ventilator-initiated a breath and the neural respiratory activity, we calculated the phase angle, which was defined as (time of ventilator-initiated breath − time of esophageal trigger)/ventilator cycle duration ×360° (5, 9, 18). Furthermore, we labeled PawInit and flowInit as the values for airway pressure and flow when the spontaneous negative effort (from esophageal manometry) was initiated (Figures 1 and 2).
Figure 2.
Phase angle = phase difference/ventilator cycle duration × 360°. This is the standard way of expressing the relationship between machine and neural respiratory activity onset.
Statistical Analysis
Medians between groups were compared with the Mann-Whitney U test. The correlation between variables was calculated using the Spearman rank correlation coefficient. The coefficient of variation was calculated as the standard deviation/mean ×100. Comparisons were made between patients on the basis of classification of the patterns of RT, as follows: none, irregular RT (no clear entrainment pattern), and regular RT (clear entrainment pattern of 1:1, 1:2, or 1:3). Additional analysis was stratified by the RTI. Mixed logistic regression models were created to first control for patient-level effects with repeated measurements per patient, evaluating variables that retained an independent association with RT or breath stacking. R version 3.5.0 (2018 The R Foundation for Statistical Computing Platform) was used for statistical analysis, and all tests were considered to be statistically significant when the P value was <0.05. Because of the ongoing nature of the parent clinical trial, outcomes such as mortality are not reported.
Results
Risk Factors for RT
The parent REDvent trial is ongoing and has enrolled approximately 115 of the targeted 276 patients. From the first 60 patients enrolled in the parent trial, 36 had evidence of spontaneous breathing while on SIMV PC-PS and were included in this this study. Patients on pressure support only were not included because RT is not possible on pressure-support breaths. This yielded 81,990 breaths from 100 patient-days (Figure E1). In the entire cohort, 2.4% of breaths were reverse triggers, and 15/36 patients (41.6%) had at least 1 day on which RT occurred. All patients met diagnostic criteria for PARDS (19), and the most common risk factor was pneumonia (Table 1).
Table 1.
The demographics of patients with RT and patients without RT
| Patients without RT | Patients with RT | P Value | |
|---|---|---|---|
| Patients, n (% | 21 (58.4) | 15 (41.6) | — |
| Age, median (IQR), mo | 79.2 (27.1–193.0) | 86.5 (23.5–174.1) | 0.87 |
| Sex, M, n (%) | 11 (52.3) | 7 (46.6) | 0.62 |
| Body weight, median (IQR), kg | 18.7 (10.8–39.5) | 24.5 (11.0–47.8) | 0.55 |
| Length of mechanical ventilator, median (IQR), d | 7.0 (4.0–9.0) | 6.0 (4.5–8.0) | 0.68 |
| Total breaths analyzed, n | 59,323 | 22,667 | — |
| Patient–ventilator asynchrony, median (IQR) | |||
| RT, % | 0.0 (0.0–0.0) | 9.5 (2.4–15.8) | <0.0001 |
| Asynchrony index, % | 3.3 (0.4–38.4) | 38.8 (16.8–60.2) | 0.019 |
| ARDS severity and triggers, n (%) | |||
| Risk factor for lung injury | |||
| Pneumonia | 15 (71.4) | 11 (73.3) | — |
| Sepsis | 1 (4.7) | 3 (20.0) | — |
| Other | 5 (23.9) | 1 (6.7) | 0.40 |
| Severity (PALICC) | |||
| Mild | 4 (19.1) | 4 (26.7) | — |
| Moderate | 9 (42.8) | 4 (26.7) | — |
| Severe | 8 (38.1) | 7 (46.6) | 0.83 |
Definition of abbreviation: ARDS = acute respiratory distress syndrome; IQR = interquartile range; PALICC = Pediatric Acute Lung Injury Consensus Conference; RT = reverse triggering.
To adjust for multiple days of data recorded per patient, we selected the day when RT occurred most frequently from patients with RT and randomly selected 1 day from patients without RT for comparison of ventilator settings and physiologic parameters (Table 2). Patients with RT had a lower median respiratory rate than patients without RT (20.0 [15.5–24.0] vs. 27.0 [24.0–34.0]; P < 0.0001), higher tidal volume on spontaneous synchronous breaths (8.1 [7.7–9.8] ml/kg vs. 6.9 [5.3–7.8] ml/kg; P = 0.017), higher tidal volume on spontaneous asynchronous breaths (9.8 [8.4–11.6] ml/kg vs. 6.9 [5.4–8.5] ml/kg; P = 0.001), and a spontaneous (or neural) respiratory rate closer to the set ventilator rate (difference of 5.1 [0.1–8.3] bpm vs. 13.5 [8.4–24.3] bpm; P < 0.0001) (Table 2). Patient demographics, ARDS severity, ARDS risk factors, blood gas results, effort of breathing, and other ventilator settings were similar between patients with RT and patients without RT. There was a trend that patients with RT received higher doses of opioids (P = 0.08), although they had a trend of being less sedated (P = 0.06) than patients without RT (Table 2). In multivariable modeling, a smaller difference between neural respiratory rate and set ventilator rate (odds ratio [OR], 0.87 [0.82–0.94]; P = 0.0002) and higher tidal volume/kg during spontaneous synchronous breaths (OR, 1.52 [1.1–2.1]; P = 0.015) remained independently associated with a patient having a day with RT.
Table 2.
Each parameter of patients with RT and patients without RT on the representative day
| Patients without RT [median (IQR)] | Patients with RT [median (IQR)] | P Value | |
|---|---|---|---|
| Observed respiratory rate | 27.0 (24.0 to 34.0) | 20.0 (15.5 to 24.0) | <0.01 |
| Set ventilator rate | 16.0 (12.0 to 22.0) | 14.0 (12.0 to 15.5) | 0.38 |
| RRneural − RRvent | 13.5 (8.4 to 24.3) | 5.1 (0.1 to 8.3) | 0.02 |
| Tidal volume of spontaneous synchronous breath/pbw | 6.9 (5.4 to 8.5) (n = 14,824 breaths, 18 patients) | 8.1 (7.7 to 9.8) (n = 4,401 breaths, 13 patients) | 0.017 |
| Tidal volume of spontaneous asynchronous breath/pbw | 6.9 (5.3 to 7.8) (n = 4,685 breaths, 17 patients) | 9.8 (8.4 to 11.6) (n = 2,579 breaths, 13 patients) | 0.001 |
| Tidal volume of ventilatory mandatory breath/pbw | 7.5 (7.2 to 7.8) (n = 3,460 breaths, 4 patients) | 9.9 (8.2 to 11.6) (n = 2,052 breaths, 2 patients) | 0.32 |
| Tidal volume of reverse triggering breath/pbw | NA | 9.4 (8.5 to 13.7) (n = 1,142 breaths, 15 patients) | NA |
| Coefficient of variation in phase angle, % | NA | 20.2 (13.6 to 25.1) | NA |
| Pplateau − PEEP | 13.5 (10.5 to 18.4) | 12.5 (7.8 to 15.0) | 0.30 |
| PEEP | 8.4 (7.2 to 10.1) | 8.8 (8.1 to 10.7) | 0.27 |
| ΔPes | 8.0 (4.5 to 10.0) | 8.0 (4.5 to 12.5) | 0.67 |
| PRP | 207 (152 to 297) | 150 (110.5 to 271) | 0.34 |
| PTP | 85.0 (50.5 to 109.0) | 77.0 (52.0 to 145.5) | 0.54 |
| pH | 7.43 (7.38 to 7.49) | 7.41 (7.35 to 7.45) | 0.50 |
| PaCO2 | 52.0 (38.7 to 63.0) | 45.0 (40.5 to 52.0) | 0.37 |
| HCO3 | 33.0 (29.7 to 36.5) | 28.0 (26.0 to 29.5) | 0.06 |
| Sedation and pain | |||
| Cumulative dexmedetomidine dose, kg/d | 21.8 (12.3 to 40.2) | 21.3 (7.3 to 36.5) | 0.82 |
| Cumulative fentanyl dose, kg/d | 25.0 (23.7 to 53.8) | 55.3 (35.1 to 69.1) | 0.08 |
| State Behavioral Scale | −1.0 (−1.0 to −1.0) | 0.0 (−1.0 to 0.5) | 0.06 |
| Face, Legs, Activity, Cry, Consolability | 1.0 (0.0 to 5.0) | 1.0 (0.0 to 2.5) | 0.49 |
Definition of abbreviations: IQR = interquartile range; NA = not applicable; PaCO2 = arterial carbon dioxide tension; PEEP = positive end-expiratory pressure; Pes = esophageal pressure; Pplateau = plateau pressure; PRP = pressure-rate product; PTP = transpulmonary pressure; pbw = predicted body weight; RRneural = spontaneous or neural respiratory rate; RRvent = ventilator rate; RT = reverse triggering.
RT occurred with a 1:1 pattern of entrainment in 1/15 patients with RT, accounting for 10.5% of the RTs. No other clear discernable pattern of RT (i.e., 1:2 or 1:3) occurred (Tables E2 and E3). Hence, 89.5% of RT breaths in 14/15 patients with RT were classified as an irregular pattern of RT, yet these patients had a median of 9.2% reverse-trigger breaths, and nearly 39% of breaths were asynchronous. The patient with a regular pattern of RT had 73% reverse-trigger breaths and 94% asynchronous breaths.
To understand the dose response, we evaluated the RTI daily. Including all patients, as the RTI increased, the median tidal volume during spontaneous synchronous breaths and spontaneous asynchronous breaths increased (correlation = 0.435 and 0.336; P < 0.0001 and 0.003) (Figures 4C and 4D), with a similar trend for mandatory breaths (correlation = 0.453; P = 0.068). As the RTI increased, the difference between neural respiratory rate and set ventilator rate decreased (correlation −0.574; P < 0.0001) (Figure 4A). There was less of a dose response, although still a clear relationship, between RTI and respiratory rate (correlation −0.418; P < 0.0001) (Figure 4B).
Figure 4.
The boxplots of each variable. (A) The reverse triggering index (RTI) was higher with RRneural getting closer to set RRvent. (B–D) The relationship between RTI and observed respiratory rate (B), tidal volume (Vt)/predicted body weight (pbw) of spontaneous synchronous breath (C) and Vt/pbw of spontaneous asynchronous breath (D). Sixty-five files of RTI = 0, 29 files of 0 < RTI < 20, and six files of RTI > 20. rpm = revolution per minute; RRneural = neural respiratory rate; RRvent = ventilator rate.
Restricting the analysis to the 15 patients who had a day with reverse triggering, RTI appeared to be highest on the first day or the day before patients had at least 10% spontaneous breathing (Figure 5).
Figure 5.
The trend of reverse triggering index and day. The x-axis is the day when spontaneous breathing appeared, which was defined as minute volume of triggered breath/total minute volume >10%.
Risk Factors for Breath Stacking When RT Occurs
Of 15 patients with RT, 14 (93.3%) had at least one episode of breath stacking, which occurred in 534 (26.5%) of 2,017 RT breaths. The median tidal volume when breath stacking occurred as a result of an RT breath was 12.2 ml/kg, compared with 9.8 ml/kg when RT did not result in breath stacking. In univariable analysis (Table 3), breath stacking was more common when the onset of patient effort occurred well after the ventilator-initiated breath (higher phase angle), when flow or inspiratory pressure was lower at the time of patient effort, when inspiratory time was shorter, and when set PIP, Set PEEP, and set Δpressure (PIP − PEEP) were lower. In multivariable modeling, breath stacking was more likely to occur when total airway Δpressure (PIP − PEEP) at the time patient effort began, set PIP, set PEEP, and set Δpressure were lower and when the onset of patient effort occurred well after the ventilator-initiated breath (higher phase angle) (all P < 0.05; Table 4). Together, these parameters were highly predictive of breath stacking (area under the curve, 0.979) (Figure E2).
Table 3.
The comparison between breath stacking and non–breath stacking in the breath-by-breath analysis of RT breaths only
| Non–Breath Stacking Waveform | Breath-Stacking Waveform | P Value | |
|---|---|---|---|
| n (%) | 1,483 (73.5) | 534 (26.5) | — |
| Vt/pbw, median (IQR) | 9.8 (9.1–11.1) | 12.2 (9.1–14.0) | <0.01 |
| ΔPes, median (IQR) | 18.4 (15.9–21.2) | 17.7 (11.8–22.6) | 0.14 |
| Phase angle, median (IQR) | 29.1 (21.6–42.8) | 75.7 (57.3–98.1) | <0.01 |
| FlowInit, median (IQR) | 35.7 (35.0–41.0) | 9.0 (7.0–15.6) | <0.01 |
| PawInit-PEEP, median (IQR) | 13.6 (13.1–16.3) | 11.5 (9.5–13.1) | <0.01 |
| Set PEEP, median (IQR) | 10.0 (10.0–10.0) | 8.0 (8.0–10.0) | <0.01 |
| Set peak inspiratory pressure, median (IQR) | 26.0 (18.0–30.0) | 22.0 (18.0–22.0) | <0.01 |
| Set Δpressure, median (IQR) | 16.0 (8.0–20.0) | 11.0 (10.0–14.0) | <0.01 |
| Ventilator inspiratory time, median (IQR) | 1.0 (1.0–1.0) | 0.8 (0.8–0.9) | <0.01 |
Definition of abbreviations: FlowInit = flow when the spontaneous negative effort (from esophageal manometry) was initiated; IQR = interquartile range; PawInit = airway pressure when the spontaneous negative effort (from esophageal manometry) was initiated; pbw = predicted body weight; PEEP = positive end-expiratory pressure; Pes = esophageal pressure; RT = reverse triggering; Vt = tidal volume.
Table 4.
Mixed-effect logistic regression model (dependent variable: breath stacking)
| Odds Ratio | Lower 95% CI | Upper 95% CI | P Value | |
|---|---|---|---|---|
| PawInit − PEEP | 0.83 | 0.71 | 0.96 | 0.01 |
| Set PIP | 0.84 | 0.80 | 0.89 | <0.01 |
| Set PEEP | 0.62 | 0.53 | 0.71 | <0.01 |
| Set Δpressure (PIP − PEEP) | 0.83 | 0.74 | 0.92 | <0.01 |
| Phase angle | 1.08 | 1.08 | 1.11 | <0.01 |
| Intercept | 4,935.75 | 73.89 | 5.51 × 105 | <0.01 |
Definition of abbreviations: CI = confidence interval; PawInit = airway pressure when the spontaneous negative effort (from esophageal manometry) was initiated; PEEP = positive end-expiratory pressure; PIP = peak inspiratory pressure.
Discussion
We have found that RT occurs in nearly half of patients with PARDS on SIMV. RT seems to occur more often when patients have a neural respiratory rate close to the set ventilator rate and in patients who achieve a higher median tidal volume on non-RT breaths. RT results in breath stacking approximately 25% of the time, and, in these cases, median tidal volume exceeds 12 ml/kg. Breath stacking was more likely to occur when total airway Δpressure (PIP − PEEP) at the time patient effort began, set PIP, set PEEP, and set Δpressure were lower and when the onset of patient effort occurred well after the ventilator-initiated breath began (higher phase angle).
We found that the majority of RT breaths were not associated with a repetitive pattern of respiratory entrainment (only 10% in our cohort), unlike the findings of other studies (5, 6, 20). We speculate that these irregular patterns of RT may be related to the SIMV PC-PS mode used in this study, in which time-cycled breaths (pressure control) are mixed with flow cycled breaths (pressure support). This may predispose to an irregular pattern when spontaneous breathing is present because both respiratory cycle duration and tidal volume may differ on the basis of the type of breath delivered. Hence, even if the patient is frequently breathing spontaneously, a mandatory breath may be delivered, and this lung insufflation leads to later patient effort, with a reverse trigger. Indeed, SIMV has previously been associated with high asynchrony and breath stacking, particularly when the set ventilator rate is more than 10 breaths/min (21). The use of SIMV in our study may also explain the higher proportion of patients with RT in our cohort (40%) than in other reported studies, which were conducted on assist-controlled modes of ventilation. In fact, we found that only one patient (3%) had RT with a reproduceable entrainment pattern, which is consistent with other estimates (22). Nevertheless, it appears as though these patients with irregular patterns of RT have many similarities with patients with regular patterns of RT (Table E3) and are at high risk for adverse events such as breath stacking.
When thinking about the mechanisms of RT, an important phenomenon relates to entrainment of the patient’s respiratory rhythm to periodic lung inflations delivered by mechanical ventilation (9). Respiratory entrainment (7, 9, 18) can occur in wakeful, sleeping, and anesthetized humans on mechanical ventilation (9). Many studies in animals have shown that periodic lung inflation during controlled mechanical ventilation may entrain the respiratory rhythm to the ventilation frequency close to the intrinsic respiratory rate (7, 8, 23, 24). Entrainment with a mechanical ventilator can occur with negative, zero, or positive phase angles. If the phase angle is zero or negative, we would classify this as synchronous ventilation. RT constitutes entrainment with a positive phase angle, in which neural respiration occurs after lung inflation from the ventilator. It is therefore not surprising that we found that RT is much more common when the patient’s neural respiratory rate was close to the set ventilator rate. RT also occurred most on the first day or day before significant spontaneous breathing, when, inherently, the neural respiratory rate may be less than or equal to the set ventilator rate. We presume that the higher doses of opioids in patients with RT (although not statistically significant) may be contributing to this potential for entrainment, given that opioids are known to reduce minute ventilation primarily by reducing respiratory rate, rather than tidal volume. Interestingly, although these patients had a higher dose of opioids, they had similar, if not higher, amounts of wakefulness on the State Behavioral Scale. Previous literature highlights differential phase relationships with entrainment as a function of the state of wakefulness of the patient (9). This illustrates the complex interplay between respiratory control of breathing, sedation, and pain in critically ill patients.
Furthermore, previous work (9) has identified that RT was more common when the ventilator cycle length was less than the respiratory cycle length (i.e., ventilator rate > spontaneous rate). This may be because ventilator inflations begin late during neural exhalation, which augments lung inflation during neural inspiration, invoking a Hering-Breuer reflex. Our data seem consistent with this as a potential mechanism, as higher tidal volume, which can augment Hering-Breuer inflation reflexes by facilitating relaxation of respiratory muscles, was a risk factor for RT (25). Furthermore, flow and pressure applied by the ventilator activate stretch receptors in the upper airways, lungs, and chest wall. Neural feedback by these receptors through vagal fibers causes the respiratory control center to match the phase and frequency of the external stimulus, producing a repetitive respiratory pattern (26). It is plausible that patients who achieve a higher median tidal volume have more stimulation of these stretch receptors (27), making it more likely for them to entrain with the set respiratory rate on the ventilator, leading to RT. These mechanisms may explain our finding that patients with RT received 8–10 ml/kg tidal volume, compared with 6–8 ml/kg for patients without RT.
Although it is important to describe and understand risk factors for RT for physiologic and academic reasons, RT may contribute to VILI, particularly when it results in breath stacking and excessive tidal volume or driving pressure. In fact, we identified that the median tidal volume on breath-stacked RT breaths was 12 ml/kg, which is known to be injurious in ARDS. We found that in pressure-control ventilation, breath stacking is more likely in patients in whom the set peak pressure (or Δpressure) is lower. We speculate that this passive lung inflation triggers the patient breath (i.e., reverse trigger), but the patient senses that the flow being delivered by the ventilator is inadequate, which may prolong the patient’s inspiratory effort beyond the inspiratory time on the time-cycled breath, resulting in a second breath that is patient triggered. Because this is pressure-control ventilation, the speed and magnitude of inspiratory flow is most dependent on the set pressure. Some have reported being able to abolish RT by increasing tidal volume (or increasing the set Δpressure in pressure-control ventilation), which may be explained by the Hering-Breuer deflation reflex (28, 29). However, that would likely be dangerous in these patients with ARDS who already have high tidal volume. We also found, not surprisingly, that when the patient effort occurs late in the course of the mandatory breath (higher phase angle), breath stacking is more common. Therefore, perhaps changing inspiratory time would be beneficial to prevent breath stacking or perhaps reducing the set ventilator rate (making the set rate less than the spontaneous rate) could prevent RT. Of course, these changes may lead to other forms of asynchrony, which should be monitored. These alterations are speculation, as our study was observational and did not test whether these strategies improve RT or breath stacking. Methods focused on eliminating breath stacking from RT should be a focus of investigation.
It is unclear whether RT in and of itself is dangerous if it does not lead to breath stacking. RT preserves some spontaneous effort and may prevent ventilator-induced diaphragm dysfunction. What is clear, however, is that tidal volume and overall transpulmonary pressure is higher in patients with RT and is even higher when RT occurs, regardless of whether it leads to breath stacking. This may increase the risk for VILI even without breath stacking, but this risk must be balanced against the potential benefits in preventing ventilator-induced diaphragm dysfunction. This should be a focus of future investigation.
Our study has limitations. First, our patients were on SIMV PC-PS ventilation from a single center. External validity should be explored in other datasets. Second, this was a sample of patients enrolled in an ongoing randomized controlled trial (RCT) on ARDS. This will exclude a subset of patients who may have RT from different mechanisms (such as patients with intracranial hypertension). Third, because the RCT is ongoing, no outcome analysis or analysis of the potential impact of the intervention on RT was undertaken. This can be done after completion of the RCT. Fourth, we obtained only 30 minutes’ worth of data each day on enrolled patients for this asynchrony analysis, which makes it a single snapshot of the types of asynchrony that occur. Because asynchronies can cluster related to sedation and daily care, it does not fully represent the burden of asynchrony in the ICU (2). However, our study has many strengths, including the use of Pes to definitively label RT and characterize asynchrony, the rigorous approach to labeling all breaths, and a sample size of both breaths and patients that is larger than any other reported cohort on RT.
Conclusions
RT is more likely to occur in patients with PARDS when the set ventilator rate is close to the patient’s intrinsic neural respiratory rate and when tidal volume is higher. Moreover, RT results in breath stacking over 25% of the time, which may make it an important risk factor for VILI. Future research is needed to understand methods to prevent RT, which may include reducing the mandatory ventilator rate below the patient’s neural respiratory rate.
Supplementary Material
Acknowledgments
Acknowledgment
The authors thank the REDvent research team and the staff in the intensive care units at the Children’s Hospital Los Angeles.
Footnotes
Supported by U.S. National Institutes of Health/National Heart Lung Blood Institute grant R01HL124666.
Author Contributions: T.S. and R.G.K.: study conception and design. J.C.H. and R.G.K.: acquisition, analysis, and interpretation of data. T.S., B.Y., M. Kyogoku, M. Kyo, S.O., C.J.L.N., J.C.H., N.S., and R.G.K. manuscript drafting and revision for intellectual content, approval of the final manuscript, and agreement to be accountable for the accuracy and integrity of the work.
This article has an online supplement, which is accessible from this issue’s table of contents at www.atsjournals.org.
Author disclosures are available with the text of this article at www.atsjournals.org.
References
- 1.Thille AW, Rodriguez P, Cabello B, Lellouche F, Brochard L. Patient-ventilator asynchrony during assisted mechanical ventilation. Intensive Care Med. 2006;32:1515–1522. doi: 10.1007/s00134-006-0301-8. [DOI] [PubMed] [Google Scholar]
- 2.Blanch L, Villagra A, Sales B, Montanya J, Lucangelo U, Luján M, et al. Asynchronies during mechanical ventilation are associated with mortality. Intensive Care Med. 2015;41:633–641. doi: 10.1007/s00134-015-3692-6. [DOI] [PubMed] [Google Scholar]
- 3.Blokpoel RGT, Wolthuis DW, Koopman AA, Kneyber MCJ. Reverse triggering: a novel type of patient-ventilator asynchrony in mechanically ventilated children. Am J Respir Crit Care Med. 2019;200:e4–e5. doi: 10.1164/rccm.201809-1781IM. [DOI] [PubMed] [Google Scholar]
- 4.Su HK, Loring SH, Talmor D, Baedorf Kassis E. Reverse triggering with breath stacking during mechanical ventilation results in large tidal volumes and transpulmonary pressure swings. Intensive Care Med. 2019;45:1161–1162. doi: 10.1007/s00134-019-05608-y. [DOI] [PubMed] [Google Scholar]
- 5.Akoumianaki E, Lyazidi A, Rey N, Matamis D, Perez-Martinez N, Giraud R, et al. Mechanical ventilation-induced reverse-triggered breaths: a frequently unrecognized form of neuromechanical coupling. Chest. 2013;143:927–938. doi: 10.1378/chest.12-1817. [DOI] [PubMed] [Google Scholar]
- 6.Benchetrit G, Muzzin S, Baconnier P, Bachy JP, Eberhard A. Neurobiol Control Breath. 1986. Entrainment of the respiratory rhythm by repetitive stimulation of pulmonary receptors: effect of CO2; pp. 263–268. [Google Scholar]
- 7.Muzzin S, Baconnier P, Benchetrit G. Entrainment of respiratory rhythm by periodic lung inflation: effect of airflow rate and duration. Am J Physiol. 1992;263:R292–R300. doi: 10.1152/ajpregu.1992.263.2.R292. [DOI] [PubMed] [Google Scholar]
- 8.Petrillo GA, Glass L, Trippenbach T. Phase locking of the respiratory rhythm in cats to a mechanical ventilator. Can J Physiol Pharmacol. 1983;61:599–607. doi: 10.1139/y83-092. [DOI] [PubMed] [Google Scholar]
- 9.Simon PM, Zurob AS, Wies WM, Leiter JC, Hubmayr RD. Entrainment of respiration in humans by periodic lung inflations: effect of state and CO(2) Am J Respir Crit Care Med. 1999;160:950–960. doi: 10.1164/ajrccm.160.3.9712057. [DOI] [PubMed] [Google Scholar]
- 10.Khemani RG, Hotz JC, Klein MJ, Kwok J, Park C, Lane C, et al. A phase II randomized controlled trial for lung and diaphragm protective ventilation (real-time effort driven VENTilator management) Contemp Clin Trials. 2020;88:105893. doi: 10.1016/j.cct.2019.105893. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Hotz JC, Sodetani CT, Van Steenbergen J, Khemani RG, Deakers TW, Newth CJ. Measurements obtained from esophageal balloon catheters are affected by the esophageal balloon filling volume in children with ARDS. Respir Care. 2018;63:177–186. doi: 10.4187/respcare.05685. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Jakob SM, Ruokonen E, Grounds RM, Sarapohja T, Garratt C, Pocock SJ, et al. Dexmedetomidine for Long-Term Sedation Investigators. Dexmedetomidine vs midazolam or propofol for sedation during prolonged mechanical ventilation: two randomized controlled trials. JAMA. 2012;307:1151–1160. doi: 10.1001/jama.2012.304. [DOI] [PubMed] [Google Scholar]
- 13.Hanks GW, Conno F, Cherny N, Hanna M, Kalso E, McQuay HJ, et al. Expert Working Group of the Research Network of the European Association for Palliative Care. Morphine and alternative opioids in cancer pain: the EAPC recommendations. Br J Cancer. 2001;84:587–593. doi: 10.1054/bjoc.2001.1680. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Mauri T, Yoshida T, Bellani G, Goligher EC, Carteaux G, Rittayamai N, et al. PLeUral pressure working Group (PLUG—Acute Respiratory Failure section of the European Society of Intensive Care Medicine) Esophageal and transpulmonary pressure in the clinical setting: meaning, usefulness and perspectives. Intensive Care Med. 2016;42:1360–1373. doi: 10.1007/s00134-016-4400-x. [DOI] [PubMed] [Google Scholar]
- 15.Blokpoel RGT, Burgerhof JGM, Markhorst DG, Kneyber MCJ. Patient-ventilator asynchrony during assisted ventilation in children. Pediatr Crit Care Med. 2016;17:e204–e211. doi: 10.1097/PCC.0000000000000669. [DOI] [PubMed] [Google Scholar]
- 16.de Haro C, López-Aguilar J, Magrans R, Montanya J, Fernández-Gonzalo S, Turon M, et al. Asynchronies in the Intensive Care Unit (ASYNICU) Group. Double cycling during mechanical ventilation: frequency, mechanisms, and physiologic implications. Crit Care Med. 2018;46:1385–1392. doi: 10.1097/CCM.0000000000003256. [DOI] [PubMed] [Google Scholar]
- 17.Beitler JR, Sands SA, Loring SH, Owens RL, Malhotra A, Spragg RG, et al. Quantifying unintended exposure to high tidal volumes from breath stacking dyssynchrony in ARDS: the BREATHE criteria. Intensive Care Med. 2016;42:1427–1436. doi: 10.1007/s00134-016-4423-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Graves C, Glass L, Laporta D, Meloche R, Grassino A. Respiratory phase locking during mechanical ventilation in anesthetized human subjects. Am J Physiol. 1986;250:R902–R909. doi: 10.1152/ajpregu.1986.250.5.R902. [DOI] [PubMed] [Google Scholar]
- 19.Pediatric Acute Lung Injury Consensus Conference Group. Pediatric acute respiratory distress syndrome: consensus recommendations from the Pediatric Acute Lung Injury Consensus Conference. Pediatr Crit Care Med. 2015;16:428–439. doi: 10.1097/PCC.0000000000000350. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Muzzin S, Trippenbach T, Baconnier P, Benchetrit G. Entrainment of the respiratory rhythm by periodic lung inflation during vagal cooling. Respir Physiol. 1989;75:157–172. doi: 10.1016/0034-5687(89)90060-1. [DOI] [PubMed] [Google Scholar]
- 21.Robinson BRH, Blakeman TC, Toth P, Hanseman DJ, Mueller E, Branson RD. Patient-ventilator asynchrony in a traumatically injured population. Respir Care. 2013;58:1847–1855. doi: 10.4187/respcare.02237. [DOI] [PubMed] [Google Scholar]
- 22.Bourenne J, Guervilly C, Mechati M, Hraiech S, Fraisse M, Bisbal M, et al. Variability of reverse triggering in deeply sedated ARDS patients. Intensive Care Med. 2019;45:725–726. doi: 10.1007/s00134-018-5500-6. [DOI] [PubMed] [Google Scholar]
- 23.MacDonald SM, Song G, Poon CS. Nonassociative learning promotes respiratory entrainment to mechanical ventilation. PLoS One. 2007;2:e865. doi: 10.1371/journal.pone.0000865. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Baconnier PF, Benchetrit G, Pachot P, Demongeot J. Entrainment of the respiratory rhythm: a new approach. J Theor Biol. 1993;164:149–162. doi: 10.1006/jtbi.1993.1145. [DOI] [PubMed] [Google Scholar]
- 25.Hayden MJ, Sly PD, Devadason SG, Gurrin LC, Wildhaber JH, LeSouëf PN. Influence of driving pressure on raised-volume forced expiration in infants. Am J Respir Crit Care Med. 1997;156:1876–1883. doi: 10.1164/ajrccm.156.6.9609003. [DOI] [PubMed] [Google Scholar]
- 26.de Vries HJ, Jonkman AH, Tuinman PR, Girbes ARJ, Heunks LMA. Respiratory entrainment and reverse triggering in a mechanically ventilated patient. Ann Am Thorac Soc. 2019;16:499–505. doi: 10.1513/AnnalsATS.201811-767CC. [DOI] [PubMed] [Google Scholar]
- 27.Modl M, Eber E.Forced expiratory flow-volume measurements Hammer J, Eber E.editorsPaediatric pulmonary function testingVol. 33Basel, Switzerland: Karger; 200534–43. [Google Scholar]
- 28.Yu J. Deflation-activated receptors, not classical inflation-activated receptors, mediate the Hering-Breuer deflation reflex. J Appl Physiol (1985) 2016;121:1041–1046. doi: 10.1152/japplphysiol.00903.2015. [DOI] [PubMed] [Google Scholar]
- 29.Jochmans S, Mazerand S, Mercier des Rochettes E, Abdallah RI, Freeman SA, Monchi M. Reverse triggering with breath stacking in ARDS patients: the “optimum” can be the enemy of the “good”. Minerva Anestesiol. 2018;84:871–872. doi: 10.23736/S0375-9393.18.12506-5. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.