ORIGINAL RESEARCH

published: 12 November 2021

doi: 10.3389/fmed.2021.752508

Automatic Adjustment of the

Inspiratory Trigger and Cycling-Off
Criteria Improved Patient-Ventilator
Asynchrony During Pressure Support
Ventilation
Ling Liu 1 , Yue Yu 1 , Xiaoting Xu 1 , Qin Sun 1 , Haibo Qiu 1 , Davide Chiumello 2,3,4† and
Edited by:
Longxiang Su,
Yi Yang 1*†
Peking Union Medical College 1
Jiangsu Provincial Key Laboratory of Critical Care Medicine, Department of Critical Care Medicine, Zhongda Hospital,
Hospital (CAMS), China
School of Medicine, Southeast University, Nanjing, China, 2 SC Anesthesia and Resuscitation, San Paolo Hospital—University
Reviewed by: Campus, ASST Santi Paolo e Carlo, Milan, Italy, 3 Department of Health Sciences, University of Milan, Milan, Italy,
4
Dan Stieper Karbing, Coordinated Research Center of Respiratory Insufficiency, University of Milan, Milan, Italy
Aalborg University, Denmark
Na Cui,
Peking Union Medical College Background: Patient-ventilator asynchrony is common during pressure support
Hospital (CAMS), China ventilation (PSV) because of the constant cycling-off criteria and variation of respiratory
Fen Liu,
The First Affiliated Hospital of system mechanical properties in individual patients. Automatic adjustment of inspiratory
Nanchang University, China triggers and cycling-off criteria based on waveforms might be a useful tool to improve
*Correspondence: patient-ventilator asynchrony during PSV.
Yi Yang
yiyiyang2004@163.com Method: Twenty-four patients were enrolled and were ventilated using PSV with different
† These
cycling-off criteria of 10% (PS10 ), 30% (PS30 ), 50% (PS50 ), and automatic adjustment
authors have contributed
equally to this work PSV (PSAUTO ). Patient-ventilator interactions were measured.
Results: The total asynchrony index (AI) and NeuroSync index were consistently lower
Specialty section:
This article was submitted to in PSAUTO when compared with PS10 , PS30, and PS50 , (P < 0.05). The benefit of
Intensive Care Medicine and PSAUTO in reducing the total AI was mainly because of the reduction of the micro-AI
Anesthesiology,
but not the macro-AI. PSAUTO significantly improved the relative cycling-off error when
a section of the journal
Frontiers in Medicine compared with prefixed controlled PSV (P < 0.05). PSAUTO significantly reduced the
Received: 03 August 2021 trigger error and inspiratory effort for the trigger when compared with a prefixed trigger.
Accepted: 11 October 2021 However, total inspiratory effort, breathing patterns, and respiratory drive were not
Published: 12 November 2021
different among modes.
Citation:
Liu L, Yu Y, Xu X, Sun Q, Qiu H, Conclusions: When compared with fixed cycling-off criteria, an automatic adjustment
Chiumello D and Yang Y (2021)
system improved patient-ventilator asynchrony without changes in breathing patterns
Automatic Adjustment of the
Inspiratory Trigger and Cycling-Off during PSV. The automatic adjustment system could be a useful tool to titrate more
Criteria Improved Patient-Ventilator personalized mechanical ventilation.
Asynchrony During Pressure Support
Ventilation. Front. Med. 8:752508. Keywords: automatic adjustment system, pressure support ventilation, patient-ventilator asynchrony, cycling-off,
doi: 10.3389/fmed.2021.752508 trigger

Frontiers in Medicine | www.frontiersin.org 1 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

INTRODUCTION and being able to sustain PSV more than 1 h with inspiratory
support ≤ 15 cm H2 O. Patients were excluded if (1) age <
Pressure support ventilation (PSV) is the most widely used partial 18 or >85 years; (2) tracheostomy at time of the study;
mode of assistance to minimize the effort of patients in breathing. (3) sedation level on the Richmond Agitation–Sedation Scale
During PSV, the assist is delivered by means of a pneumatic ≤ −2 or ≥ 2; (4) contraindication for nasogastric tube
signal generated by patient effort and measured in the ventilatory insertion, e.g., history of esophageal varices, gastroesophageal
circuit, i.e., flow or pressure (1). The ventilator usually cycles surgery in the previous 12 months, or gastroesophageal
from inspiration to expiration when the inspiratory flow falls to bleeding in the previous 7 days, international standard ratio
a predetermined fraction of the peak inspiratory flow, which is > 1.5, activated partial thromboplastin time > 44 s, history
the cycling-off criterion (2). Ideally, the ventilator trigger and of leukemia (13); and (5) hemodynamic instability (heart rate
cycling should coincide with the beginning and the end of the > 140 beats/min, vasopressors required with ≥ 5 µg/kg/min
inspiratory effort of the patients (3). However, patient-ventilator dopamine/dobutamine, or ≥ 0.2 µg/kg/min norepinephrine).
asynchrony is common during PSV (4, 5), thereby contributing
to the increased patient effort, increased duration of mechanical Study Protocol
ventilation, and even increased mortality (6). After obtaining consent, enrolled patients were switched to a
During PSV, prefixed pneumatic controllers can become Servo-i ventilator (Maquet, Solna, Stockholm, Sweden). A 16-
progressively less effective, especially when patients have F nasogastric feeding tube (NeuroVent Research Inc., Toronto,
abnormal respiratory mechanics or ventilator over-assist (7). ON, Canada) with electrodes measuring the electrical activity
Delayed or missed triggers are sensed as an uncomfortable of the diaphragm (EAdi) and balloons measuring esophageal
isometric load leading to increased effort intensity and (Pes) pressures was inserted through the nose and secured after
pronounced dyspnea (8). Moreover, with prefixed cycling- confirming positioning according to the recommendations of
off criteria, such as the default value of 30% peak flow the manufacturer. Static respiratory system compliance (CRS ),
in some ventilators, premature cycling is more frequent in resistance (RRS ), and intrinsic positive end-expiratory pressure
patients with restrictive breathing patterns characterized by (PEEPi) were measured during volume control ventilation
low respiratory system compliance and may result in double (without spontaneous drive) (see Supplementary Material).
triggering. Delayed cycling occurs more frequently in patients Then sedation was decreased to maintain light sedation
with an obstructive pattern characterized by high resistance (6, 9). with the Richmond Agitation–Sedation Scale ranging from 0
Different approaches for optimal ventilator triggering and cycling to −2. As spontaneous breathing and EAdi recovered, patients
have been developed to minimize these problems, such as flow- were switched to an SV800 ventilator with IntelliCycleTM2.0
triggering sensitivity and adjustable flow cycling during PSV. which can automatically adjust triggering and cycling-off
It has been demonstrated that a noninvasive method based criteria breath-by-breath, (Mindray, Shenzhen, China) and were
on flow and airway-pressure tracings was effective for detecting ventilated by PSV with the pressure support level adjusted to
asynchrony (10–12). Therefore, an automatic adjustment system a target tidal volume (VT ) of 6 ml/kg (of predict body weight,
(IntelliCycleTM2.0) capable of automatically adjusting, breath by PBW). During the entire recording period, PEEPe and a fraction
breath, the triggering and cycling-off criteria based on pressure- of inspired oxygen (FiO2 ) were maintained as set by the clinician
time and flow-time waveforms during PSV have been developed in charge of the patient.
(see Supplementary Material). During prefixed pneumatically controlled PSV, the inspiratory
The objective of our study was to show a reduction in patient- trigger was set at 1.5 L/min for flow triggering, and the
ventilator asynchrony with the use of an automatic adjustment rate of rise in pressure was set to 0.05 s in all patients. The
system as compared with prefixed trigger and cycling-off criteria cycling-off criteria were set to 10% (PS10 ), 30% (PS30 ), and
in patients with PSV. 50% (PS50 ). During automatic adjustment PSV, the inspiratory
trigger was set as flow-trigger 1.5 L/min, the rate of rise in
METHODS pressure was set to 0.05 s, and the cycling-off criterion was set to
“AUTO” (PSAUTO ). Both the trigger and cycling-off criteria were
This unblinded crossover study was conducted in a 60-bed adjusted by the automatic adjustment system according to an
general intensive care unit of a teaching hospital affiliated with established algorithm based on the pressure-time and flow-time
Southeast University in China. The protocol was approved by waveforms (Supplementary Figures 1, 2). First, patients were
the Institutional Ethics Committee of Zhongda Hospital (number ventilated with four independent modes (PS10 , PS30 , PS50, and
2016ZDSYLL067-P01). Written informed consent was obtained PSAUTO ) applied in randomized order (Supplementary Table 1).
from the legal primary decision-maker, which was the spouse Each independent condition was maintained for 20 min without
of the patient or the parent or child if no spouse. The trial was washout periods (Supplementary Figure 3).
registered at clinicaltrials.gov (NCT04091269).
Data Acquisition and Analysis
Patients Flow, airway pressure (Paw ), esophageal pressure (Pes ), and
Postoperative (abdominal surgery or orthopedic surgery) or EAdi were acquired during the 20-min time window in each
acute respiratory failure patients were eligible when meeting all condition at 100 Hz from the ventilator via an RS 232 interface
the following criteria: receiving invasive mechanical ventilation connected to a computer. Data were stored for later offline

Frontiers in Medicine | www.frontiersin.org 2 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

analysis (NeuroVent Research Inc., Toronto, ON, Canada). or total asynchrony events divided by the neural respiratory rate
To quantify patient-ventilator interaction, all variables were × 100%.
calculated manually breath by breath from a stable 3-min Triggering and cycling-off errors, which were classified as
period in each condition using customized software (NeuroVent either too late (positive values) or too early (negative values)
Research Inc., Toronto, ON, Canada) by two independent (13), breathing pattern, inspiratory effort, and inspiratory effort
researchers who were blinded to the patient number and assigned for triggering were measured (see Supplementary Material). To
order of crossover treatments, and mean values were calculated. estimate the overall extent of asynchrony and dys-synchrony,
In the event of a mismatch, a third researcher was consulted. the NeuroSync index was calculated by averaging the percentage
Six types of asynchrony were analyzed as previously described errors in triggering and cycling-off for all breaths (13). The
by Thille et al. and Lamouret et al. (6, 14). Macro asynchronies primary endpoint was the difference in the total AI between
include ineffective triggering, which is defined by the existence PSAUTO and PSV with prefixed triggering and cycling-off criteria
of a diaphragmatic signal without a respiratory cycle; auto- (PS10 , PS30 , and PS50 ).
triggering is defined by the existence of a ventilator cycle without
a diaphragmatic signal; and double triggering is defined by Statistical Analysis
the presence of two successive inspiratory cycles without an All statistical analyses were carried out using SPSS 20 (Chicago,
intermediate expiration or with an interrupted expiration. Micro- IL, USA). The values are stated as mean ± SD unless specified
asynchronies are defined by a time difference exceeding 200 ms otherwise. Data from two post-hoc subgroups, a restrictive
between the onset of the EAdi and the early initial rise in subgroup defined as having CRS < 40 ml/cm H2 O with RRS
Paw; between the 70% of peak EAdi and early decrease in < 12 cm H2 O/LS, and an obstructive subgroup, defined as
airway pressure (the opening of the expiratory valve)-late cycling; having RRS > 12 cm H2 O/LS with CRS > 40 ml/cm H2 O, were
and between the decrease in airway pressure and 70% of peak analyzed. The normal distribution of continuous variables was
EAdi-premature cycling. For each subtype of asynchrony, a assessed by using the Shapiro–Wilk test. Log-transformation was
percentage of asynchronies was calculated as follows: the number used for skewed data. Variables were compared between modes
of asynchrony events divided by the total neural respiratory rate using repeated-measures ANOVA followed by Bonferroni’s post-
(which corresponds to the total EAdi signals) × 100%. Macro- hoc test. Categorical data were compared by the chi-square
asynchrony index (AI), micro-AI, and total AI were calculated test followed by Bonferroni’s post-hoc test. P-values <0.05 were
as the number of macro asynchrony events, micro-asynchrony, considered significant.

TABLE 1 | Patient characteristics.

Parameter All Obstructive subgroup Restrictive subgroup Other patients

(n = 24) (n = 8) (n = 8) (n = 8)

Sex, male/female 19/5 5/3 7/1 7/1

Age, year 68 ± 17 75 ± 9 65 ± 17 68 ± 23
APACHE II 17.1 ± 5.3 17.9 ± 4.0 18.7 ± 6.0 14.8 ± 6.5
Main diagnosis
Pneumonia, n (%) 4 (16.7%) – 4 (50%)
Extrapulmonary sepsis, n (%) 2 (8.3%) – 2 (25%)
AECOPD, n (%) 8 (33.3%) 8 (100.0%) –
Abdominal surgery 4 (16.7%) – – 4 (50.0%)
Orthopedic surgery, n (%) 4 (16.7%) – – 4 (50.0%)
Severe trauma, n (%) 2 (8.3%) 2 (25%)
RASS 0 (−1, 0) 0 (−1, 0) −1 (−2, 0) 0 (−1, 0)
PBW, Kg 63 ± 8 63 ± 7 59 ± 9 65 ± 7
PaO2 , mm Hg 107 ± 36 96 ± 30 95 ± 16 135 ± 42
PaO2/ FiO2 276 ± 90 362 ± 78 239 ± 76 227 ± 41
PaCO2 , mm Hg 39 ± 11 48 ± 14 36 ± 6 32 ± 4
pH 7.41 ± 0.06 7.39 ± 0.05 7.43 ± 0.03 7.41 ± 0.06
CRS , ml/cm H2 O 45.8 ± 9.7 50.9 ± 5.8 34.1 ± 5.1 52.5 ± 3.5
RRS , cm H2 O/L/S 12.1 ± 4.9 17.9 ± 4.1 9.2 ± 2.1 9.4 ± 1.7
PEEPi, cm H2 O 1.7 ± 2.0 3.6 ± 2.4 0.9 ± 0.2 0.7 ± 0.7

Data are provided as mean ± SD or median (interquartile range).

APACHE II, Acute Physiology and Chronic Health Evaluation II; RASS, Richmond Agitation-Sedation Scale; PBW, predictive body weight; CRS, static compliance of the respiratory
system; RRS , resistance of respiratory system; PEEPi, static intrinsic positive end expiratory pressure.

Frontiers in Medicine | www.frontiersin.org 3 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

FIGURE 1 | Total AI (A), macro-AI (B), and micro-AI (C) in different modes. AI, asynchrony index; PS10, pressure support ventilation with cycling-off criteria set to
10%; PS30, pressure support ventilation with cycling-off criteria set to 30%; PS50, pressure support ventilation with cycling-off criteria set to 50%; PSAUTO, pressure
support ventilation with automatic. Gray lines showed median (interquartile range). Compared with PS10 , a P < 0.05; Compared with PS30 , b P < 0.05; compared with
PS50 , c P < 0.05.

Frontiers in Medicine | www.frontiersin.org 4 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

TABLE 2 | Asynchronies, NeuroSync index, inspiratory effort, and relative timing errors of cycling-off and trigger in different modes.

Parameters PS10 PS30 PS50 PSAUTO P value

Ineffective triggering, % 0.0 (0.0, 2.3) 0.0 (0.0, 0.0) 0.0 (0.0, 1.0) 0.0 (0.0, 1.4) 0.118
Auto-triggering, % 0.0 (0.0, 0.0) 0.0 (0.0, 3.3) 0.0 (0.0, 0.0) 0.0 (0.0, 0.0) 0.039
Double triggering, % 0.0 (0.0, 6.7) 0.0 (0.0, 0.0) 0.0 (0.0, 0.0) 0.0 (0.0, 0.0) 0.569
Premature cycling-off, % 0.0 (0.0, 2.6) 0.0 (0.0, 1.8) 5.4 (0.0, 22.5)b 0.0 (0.0, 1.8)c <0.001
Late cycling-off, % 7.1 (0.0, 28.1) 1.8 (0.0, 20.8) 0.0 (0.0, 0.0)a 0.0 (0.0, 0.0)a <0.001
Inspiratory trigger delay, % 38.3 (22.4, 48.9) 25.9 (11.4, 47.1) 30.3 (16.3, 45.6) 19.0 (5.0, 31.3)abc <0.001
NeuroSync index, % 15.3 ± 8.2 13.3 ± 6.7a 13.1 ± 4.8 9.7 ± 4.4abc <0.001
“Perfect” synchrony breath, % 18.5 (16.4, 20.7) 21.9 (19.5, 24.2)a 19.7 (17.5, 21.9) 42.2 (39.5, 44.9)abc <0.001
“Acceptable” synchrony breath, % (95% CI) 81.1 (78.9, 83.2) 87.9 (86.0, 89.7)a 89.5 (87.8, 91.2)a 94.8 (93.5, 96.0)abc <0.001
a
PTPes−Trig , cmH2 O.S.min −1
−3.1 (−6.0, −1.1) −2.3 (−4.1, −1.1) −2.0 (−3.6, −1.1) −1.9 (−3.7, −0.8)ab <0.001
PTPes , cmH2 O.S.min−1 −17.1 (−88.2, −13.1) −38.7 (−71.3, −10.6) −40.8 (−58.0, −8.9) −37.4 (−61.2, −9.1) 0.802

Data are provided as mean ± SD or median (interquartile range).

NeuroSync index is an overall indicator of patient-ventilator interaction, where 0% error, perfect; 100% error, zero patient-ventilator interaction; PTPes−Trig, Pre-trigger Pes-time product;
PTPes, Pes-time product; PS10, pressure support ventilation with cycling-off criteria set to 10%; PS30, pressure support ventilation with cycling-off criteria set to 30%; PS50, pressure
support ventilation with cycling-off criteria set to 50%; PSAUTO, pressure support ventilation with automatic adjustment system; CI, Confidence interval, “perfect” synchrony, relative
timing errors of triggering and for cycling-off ≤10% of neural timings, “acceptable” synchrony, relative timing errors of triggering and for cycling-off ≤10% of neural timings.
Compared with PS10 , a P < 0.05; Compared with PS30 , b P < 0.05; Compared with PS50 , c P < 0.05.

RESULTS error in PSAUTO was comparable with that in PS50 in the

obstructive subgroup and was comparable with that in PS10 in the
The study included 24 patients, such as eight patients in the restrictive subgroup. PSAUTO significantly shortened the absolute
restrictive subgroup, eight patients in the obstructive subgroup, and relative triggering errors when compared with a prefixed
and eight other patients without obvious acute respiratory failure trigger (PS10 , PS30, or PS50 ; Figure 3). The Absolute and relative
(CRS > 40 ml/cm H2 O with RRS < 12 cm H2 O/LS). Patient triggering errors were significantly lower when compared with
characteristics and lung mechanism are summarized in Table 1. PS10 , PS30 , and PS50 in the obstructive subgroup but not in the
restrictive subgroup (Supplementary Figure 4).
AI
Total AI was consistently lower in PSAUTO when compared Respiratory Drive and Breathing Pattern
with PS10 , PS30 , and PS50 , (P < 0.05). The benefit of PSAUTO Inspiratory effort for triggering determined by PTPes−trig was
in reducing total AI was mainly in the reduction of micro-AI significantly lower in PSAUTO when compared in PS10 and
but not macro-AI (Figure 1). The percentages of all kinds of PS30 ; however, total inspiratory effort determined by PTPes
asynchronies are reported in Table 2. Total AI and micro-AI was not different among modes (Table 3). In the obstructive
were lower in PSAUTO when compared with PS10 and PS30 in the subgroup, PTPes−trig was significantly lower in PSAUTO than
obstructive subgroup and were lower in PSAUTO when compared in PS10 , PS30 , and PS50 (P < 0.05; Supplementary Table 2).
with PS50 in the restrictive subgroup (Supplementary Table 2). Peak airway pressure was higher in PS10 than in other
modes. There was no difference in the respiratory drive
NeuroSync Index between modes (Table 3). Breathing patterns and respiratory
The NeuroSync index (average of the percentage errors of
drive in obstructive and restrictive subgroups are shown
triggering and cycling-off) was consistently lower in PSAUTO
in Supplementary Table 3.
when compared with PS10 , PS30, and PS50 , indicating improved
patient-ventilator interaction (Table 2). Figure 2 shows a plot of
the percentage errors of triggering (X-axis) and cycling-off (Y- DISCUSSION
axis) for every breath. We have inserted a small centered box
suggesting “perfect” asynchrony to be ≤10% of neural timing This study showed that when compared with PSV with
and a larger box suggesting “acceptable” asynchrony to be ≤33% prefixed pneumatic controllers, an automatic adjustment system
of neural timing (15). There were more “Perfect” asynchrony decreased total AI and improved patient-ventilator interaction
breaths and “Acceptable” asynchrony breaths in PSAUTO than in mainly through a decrease of micro-asynchronies. The automatic
the fixed cycling-off criteria mode (PS10 , PS30 , and PS50 , all P < system was associated with the lower cycling-off error, triggering
0.05; Figure 2). error, and triggering effort in PSV patients.

Cycling-Off and Triggering Error AI and NeuroSync Index

Automatic adjustment PSV significantly improved the relative Both AI and the NeuroSync index are indicators that reflect the
cycling-off error when compared with PS10 , PS30 , and PS50 overall patient-ventilator interaction from different perspectives.
in the whole population (Figure 3). The relative cycling-off PSAUTO constantly reduced total AI and the NeuroSync

Frontiers in Medicine | www.frontiersin.org 5 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

FIGURE 2 | Breath density graph for relative trigger (X-axis) and cycling-off (Y -axis) errors, for all breaths in all patients, during each ventilator mode. PS10, pressure
support ventilation with cycling-off criteria set to 10%; PS30, pressure support ventilation with cycling-off criteria set to 30%; PS50, pressure support ventilation with
cycling-off criteria set to 50%; PSAUTO, pressure support ventilation with automatic adjustment system; asynchrony error, breathes inside the box of percentage error
of neural timings.

Frontiers in Medicine | www.frontiersin.org 6 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

FIGURE 3 | Cycling-off error and trigger error in different modes. (A) cycling-off error, (B) relative cycling-off error, (C) trigger error, and (D) relative trigger error, Y -axis
for cycling-off error: positive values indicate late cycling-off, and negative values indicate early cycling off. Magenta line showed median (interquartile range). ms,
millisecond; PS10, pressure support ventilation with cycling-off criteria set to 10%; PS30, pressure support ventilation with cycling-off criteria set to 30%; PS50,
pressure support ventilation with cycling-off criteria set to 50%; PSAUTO, pressure support ventilation with automatic. Compared with PS10 , a P < 0.05; Compared with
PS30 , b P < 0.05; Compared with PS50 , c P < 0.05.

index when compared with PSV with prefixed pneumatic which showed the difference in AI is found only in micro-
controllers, indicating improved patient-ventilator interaction. asynchronies (14).
Given that macro-asynchronies were rare, the benefit of PSAUTO The present study showed a higher total AI (median of 23–
in reducing the total AI was mainly due to the reduction 57% during PS10 , PS30 , PS50 , and PSAUTO ) when compared with
of micro-asynchronies. These findings agree with previous those in previous studies (range from 0 to 27%) (6, 16, 17).
work comparing PSV and neurally adjusted ventilatory assist, Despite the differences among study patients and ventilators

Frontiers in Medicine | www.frontiersin.org 7 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

TABLE 3 | Breathing pattern and respiratory drive in different modes.

Parameter PS10 PS30 PS50 PSAUTO P value

Ppeak, cmH2 O 16.5 ± 4.2 15.4 ± 4.2a 15.4 ± 4.1a 14.6 ± 4.1a < 0.001
PEEP, cmH2 O 6.2 ± 1.7 6.1 ± 1.6 6.3 ± 1.5 6.6 ± 1.5 0.120
Vt, cmH2 O/kg 6.1 ± 0.2 6.1 ± 0.2 6.1 ± 0.3 6.0 ± 0.1 0.169
RRN , breath/min 20.1 ± 5.6 19.3 ± 7.7 19.3 ± 5.9 20.5 ± 6.4 0.331
TiN , s 1.2 ± 0.0 1.1 ± 0.0 1.2 ± 0.0 1.1 ± 0.0 0.373
TeN , s 2.2 ± 0.2 2.4 ± 0.2 2.7 ± 0.3 2.4 ± 0.2 0.129
TiN/TtN, % 37.0 ± 1.4 33.4 ± 1.6 35.7 ± 1.7 36.5 ± 1.5 0.042
Peak EAdi, µV 12.9 ± 1.7 12.0 ± 1.5 12.8 ± 1.9 12.4 ± 1.9 0.611
Peak EAdi, µV 8.1 ± 5.5 9.3 ± 6.1 8.0 ± 2.4 6.8 ± 2.0 0.864

Data are provided as mean ± SD.

PS10, pressure support ventilation with cycling-off criteria set to 10%; PS30, pressure support ventilation with cycling-off criteria set to 30%; PS50, pressure support ventilation with
cycling-off criteria set to 50%; PSAUTO, pressure support ventilation with automatic adjustment system; Ppeak, peak airway pressure; PEEP, positive end expiratory pressure; Vt, tidal
volume; RR, respiratory rate; TiN , neural inspiratory time; TeN , neural expiratory time; Peak EAdi, peak diaphragm electrical activity.
Compared with PS10 , a P < 0.05; Compared with PS30 .

used, the major reason for the apparent differences between Triggering Error
studies might relate to the calculation method for the AI. First, The present study showed the median delay for triggering during
inspiratory trigger delay was included in the calculation of the AI PS10, PS30, PS50 , and PSAUTO ranged from 187 to 130 ms. These
in the present study, which provided about one-third to one-half values fall within the 80–540 ms range of values previously
of the total AI during PSV with prefixed pneumatic controllers. reported for PSV (1, 18, 23). Beloncle et al. reported absolute
However, the previous study did not calculate inspiratory trigger values for trigger delay < 200 ms in almost all patients, which was
delay in the AI (6). Second, we defined asynchrony as an error of lower than that reported in the present study (18). The different
200 ms between the origin of the EAdi and ventilator insufflation, ventilators and flow-trigger used in different studies might be
which was more sensitive than the threshold used in previous one reason, and different types of the enrolled patients might be
studies (6, 16–18). Therefore, the AI in the present study is more another reason for the difference in trigger delay. During PSAUTO ,
sensitive and comprehensive and therefore not comparable to the algorithm will trigger the ventilator to initiate the inspiratory
those in other studies. phase when it detects a sudden increase of flow waveform, which
reflects the inspiratory effort, leading to a reduced triggering
delay. Furthermore, triggering delay was likely reduced as a
Cycling-Off Error consequence of reduced cycling-off delay during PSAUTO , which
Cycling-off asynchrony is dependent on factors, such as the led to a longer expiration time and lower PEEPi, especially in
inspiratory effort, neural inspiratory time, assist levels, the time patients with obstructive conditions (5). Unfortunately, we did
constant of the respiratory system, and cycling-off criteria of the not measure PEEPi during each mode.
patients (3). Consequently, the optimum flow cycling-off criteria Of note, a single flow-trigger level in the present study
vary among patients and can range from very low levels (5%) in made it hard to draw conclusions regarding the effect of the
patients with a restrictive condition (such as acute respiratory PSAUTO mode on inspiratory triggering asynchronies when
distress syndrome) (4, 19) to more than 50% in patients with compared with lower flow-triggering (e.g., 1.0 L/min). From
an obstructive condition (such as chronic obstructive pulmonary this perspective, a fixed flow-trigger of 1.5 L/min might be not
disease) (5, 20, 21). A previous study showed in a mixed sample of sensitive enough. Considering the similar or shorter triggering
patients that the use of a variable, real-time-adjusted termination delay and no obvious auto-triggering during PSAUTO , automatic
criterion improved some indices of patient-ventilator asynchrony adjustment of triggering based on waveforms might be a useful
when compared with a fixed termination criterion (5% of peak tool for making the triggering setting easier.
inspiratory flow) (22). However, a termination criterion of 5% of
peak inspiratory flow was not commonly used clinically during
PSV. Our results showed a significant improvement in relative Inspiratory Effort and Breathing Pattern
cycling-off error during PSAUTO when compared with PSV with Because PSAUTO significantly reduced triggering delay, it was
prefixed cycling-off criteria of 10%, 30%, and 50%. It was not not unexpected that it was associated with lower inspiratory
unexpected that during PSV with prefixed pneumatic controllers, effort for triggering. The present study showed a comparable
PS50 and PS10 were the “best” cycling-off settings with the breathing pattern and respiratory drive between PSAUTO and
lowest relative cycling-off errors in the obstructive and restrictive PSV with prefixed pneumatic controllers. Of interest, neural
subgroups. In each subgroup, relative cycling-off error in PSAUTO expiratory time remained unchanged at the various cycling-
was comparable with the “best” cycling-off setting during PSV off settings in the present study. These findings agree with
with a prefixed cycling-off. previous work in which expiratory time did not change with the

Frontiers in Medicine | www.frontiersin.org 8 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

increase in cycling-off criteria in chronic obstructive pulmonary Hospital. The patients/participants provided their written
disease patients (5, 20). However, the findings contradict those informed consent to participate in this study.
in previous studies which show an increased expiratory time in
the presence of delayed cycling in acute lung injury (24, 25). AUTHOR CONTRIBUTIONS
Therefore, PSAUTO improved the cycling-off criteria, which was
demonstrated to affect the inspiratory time only at high-pressure LL, YYa, DC, and HQ have given substantial contributions to the
support (20). Peak EAdi around 12 µV confirmed the absence of conception or the design of the manuscript. LL, YYu, XX, and QS
over-assistance during PSV in the present study. to acquisition, analysis, and interpretation of the data. All authors
contributed equally to the manuscript and read and approved the
Limitations final version of the manuscript.
There are some limitations that should be noted. First, our study
was conducted in a small group of patients. Second, respiratory
mechanics were evaluated in patients under sedation who were FUNDING
not actively breathing, therefore, the results will be different from
those measured during PSV. Third, patients were maintained at This study was supported by the Clinical Science and
each mode setting for only 20 min, and steady-state conditions Technology Specific Projects of Jiangsu Province [BE2020786
might not have been achieved. However, the duration was in and BE2019749], the National Natural Science Foundation of
line with that of several studies on the effects of cycling criteria China [Grant Numbers 81870066 and 81670074], the Natural
modifications (4, 19). Science Foundation of Jiangsu Province (BK20171271), Jiangsu
Provincial Medical Youth Talent (QNRC 2016807), and Third
Level Talents of the 333 High Level Talents Training Project in
CONCLUSIONS the fifth phase in Jiangsu (LGY2016051). This study received
An automatic adjustment system based on waveform was funding from Mindray (China). The funder was not involved
associated with less patient-ventilator asynchrony when in the study design, collection, analysis, interpretation of
compared with PSV with prefixed pneumatic controllers. Our data, the writing of this article, or the decision to submit it
results indicated that this system might be a useful tool to titrate for publication.
more personalized mechanical ventilation, especially in patients
with a high risk of patient-ventilator asynchrony. ACKNOWLEDGMENTS
DATA AVAILABILITY STATEMENT The authors thank Mr. Jinglei Liu, Xiaoyong Zhou, and Jun
Chen (Shenzhen Mindray Bio-Medical Electronics Co., Ltd.) for
The raw data supporting the conclusions of this article will be helping to making the figures.
made available by the authors, without undue reservation.
SUPPLEMENTARY MATERIAL
ETHICS STATEMENT
The Supplementary Material for this article can be found
The studies involving human participants were reviewed online at: https://www.frontiersin.org/articles/10.3389/fmed.
and approved by Institutional Ethics Committee of Zhongda 2021.752508/full#supplementary-material

REFERENCES 6. Thille AW, Rodriguez P, Cabello B, Lellouche F, Brochard L. Patient-ventilator

asynchrony during assisted mechanical ventilation. Intensive Care Med.
1. Spahija J, Marchie M, Albert M, Bellemare P, Delisle S, Beck J, (2006) 32:1515–22. doi: 10.1007/s00134-006-0301-8
et al. Patient-ventilator interaction during pressure support ventilation 7. Jubran A, Van de Graaff WB, Tobin MJ. Variability of patient-ventilator
and neurally adjusted ventilatory assist. Crit Care Med. (2010) 38:518– interaction with pressure support ventilation in patients with chronic
26. doi: 10.1097/CCM.0b013e3181cb0d7b obstructive pulmonary disease. Am J Respir Crit Care Med. (1995) 152:129–
2. MacIntyre NR. Clinically available new strategies for mechanical ventilatory 36. doi: 10.1164/ajrccm.152.1.7599811
support. Chest. (1993) 104:560–5. doi: 10.1378/chest.104.2.560 8. Gilstrap D, MacIntyre N. Patient-ventilator interactions. Implications for
3. Yamada Y, Du HL. Analysis of the mechanisms of expiratory asynchrony in clinical management. Am J Respir Crit Care Med. (2013) 188:1058–
pressure support ventilation: a mathematical approach. J Appl Physiol. (2000) 68. doi: 10.1164/rccm.201212-2214CI
88:2143–50. doi: 10.1152/jappl.2000.88.6.2143 9. Garofalo E, Bruni A, Pelaia C, Liparota L, Lombardo N, Longhini F, et al.
4. Tokioka H, Tanaka T, Ishizu T, Fukushima T, Iwaki T, Nakamura Y, et al. Recognizing, quantifying and managing patient-ventilator asynchrony in
The effect of breath termination criterion on breathing patterns and the invasive and noninvasive ventilation. Expert Rev Respir Med. (2018) 12:557–
work of breathing during pressure support ventilation. Anesth Analg. (2001) 67. doi: 10.1080/17476348.2018.1480941
92:161–5. doi: 10.1097/00000539-200101000-00031 10. Thille AW, Brochard L. Promoting patient-ventilator synchrony. Clin Pulm
5. Tassaux D, Gainnier M, Battisti A, Jolliet P. Impact of expiratory trigger setting Med. (2007) 14:350–9. doi: 10.1097/CPM.0b013e31815b1234
on delayed cycling and inspiratory muscle workload. Am J Respir Crit Care 11. Giannouli E, Webster K, Roberts D, Younes M. Response of
Med. (2005) 172:1283–9. doi: 10.1164/rccm.200407-880OC ventilator-dependent patients to different levels of pressure support

Frontiers in Medicine | www.frontiersin.org 9 November 2021 | Volume 8 | Article 752508

Liu et al. Automatic Adjustment Pressure Support Ventilation

and proportional assist. Am J Respir Crit Care Med. (1999) 21. Tassaux D, Michotte JB, Gainnier M, Gratadour P, Fonseca S,
159:1716–25. doi: 10.1164/ajrccm.159.6.9704025 Jolliet P. Expiratory trigger setting in pressure support ventilation:
12. Mellott KG, Grap MJ, Munro CL, Sessler CN, Wetzel PA. Patient-ventilator from mathematical model to bedside. Crit Care Med. (2004)
dysasynchrony: clinical significance and implications for practice. Crit Care 32:1844–50. doi: 10.1097/01.CCM.0000138561.11634.6F
Nurse. (2009) 29:41–55. doi: 10.4037/ccn2009612 22. Hoff FC, Tucci MR, Amato MBP, Santos LJ, Victorino JA.
13. Liu L, Xia F, Yang Y, Longhini F, Navalesi P, Beck J. Neural versus pneumatic Cycling-off modes during pressure support ventilation: effects on
control of pressure support in patients with chronic obstructive pulmonary breathing pattern, patient effort, and comfort. J Crit Care. (2014)
diseases at different levels of positive end expiratory pressure: a physiological 29:380–5. doi: 10.1016/j.jcrc.2014.01.015
study. Crit Care. (2015) 19:244. doi: 10.1186/s13054-015-0971-0 23. Meric H, Calabrese P, Pradon D, Lejaille M, Lofaso F, Terzi N. Physiological
14. Lamouret O, Crognier L, Vardon BF, Conil JM, Dilasser C, comparison of breathing patterns with neurally adjusted ventilatory assist
Raimondi T, et al. Neurally adjusted ventilatory assist (NAVA) versus (NAVA) and pressure-support ventilation to improve NAVA settings. Respir
pressure support ventilation: patient-ventilator interaction during Physiol Neurobiol. (2014) 195:11–8. doi: 10.1016/j.resp.2014.01.021
invasive ventilation delivered by tracheostomy. Crit. Care. (2019) 24. Younes M, Kun J, Webster K, Roberts D. Response of ventilator-dependent
23:22. doi: 10.1186/s13054-018-2288-2 patients to delayed opening of exhalation valve. Am J Respir Crit Care Med.
15. Sinderby C, Liu S, Colombo D, Camarotta G, Slutsky AS, Navalesi P, et al. (2002) 166:21–30. doi: 10.1164/rccm.2107143
An automated and standardized neural index to quantify patient-ventilator 25. Kondili E, Prinianakis G, Anastasaki M, Georgopoulos D. Acute effects of
interaction. Crit Care. (2013) 17:R239. doi: 10.1186/cc13063 ventilator settings on respiratory motor output in patients with acute lung
16. Demoule A, Clavel M, Rolland-Debord C, Perbet S, Terzi N, Kouatchet A, injury. Intensive Care Med. (2002) 27:1147–57. doi: 10.1007/s001340101000
et al. Neurally adjusted ventilatory assist as an alternative to pressure support
ventilation in adults: a French multicenter randomized trial. Intensive Care Conflict of Interest: LL and HQ received a grant from Mindray (China).
Med. (2016) 42:1723–32. doi: 10.1007/s00134-016-4447-8
17. Yonis H, Crognier L, Conil JM, Serres I, Rouget A, Virtos M, et al. Patient- The remaining authors declare that the research was conducted in the absence of
ventilator asynchrony in neurally adjusted ventilatory assist (NAVA) and any commercial or financial relationships that could be construed as a potential
pressure support ventilation (PSV): a prospective observational study. BMC conflict of interest.
Anesthesiol. (2015) 15:117. doi: 10.1186/s12871-015-0091-z
18. Beloncle F, Piquilloud L, Rittayamai N, Sinderby C, Rozé H, et al. A Publisher’s Note: All claims expressed in this article are solely those of the authors
diaphragmatic electrical activity-based optimization strategy during pressure and do not necessarily represent those of their affiliated organizations, or those of
support ventilation improves synchronization but does not impact work of
the publisher, the editors and the reviewers. Any product that may be evaluated in
breathing. Crit Care. (2017) 21:21. doi: 10.1186/s13054-017-1599-z
this article, or claim that may be made by its manufacturer, is not guaranteed or
19. Chiumello D, Pelosi P, Taccone P, Slutsky A, Gattinoni L. Effect of different
inspiratory rise time and cycling off criteria during pressure support endorsed by the publisher.
ventilation in patients recovering from acute lung injury. Crit Care Med.
(2003) 31:2604–10. doi: 10.1097/01.CCM.0000089939.11032.36 Copyright © 2021 Liu, Yu, Xu, Sun, Qiu, Chiumello and Yang. This is an open-access
20. Chiumello D, Polli F, Tallarini F, Chierichetti M, Motta G, article distributed under the terms of the Creative Commons Attribution License (CC
Azzari S, et al. Effect of different cycling-off criteria and positive BY). The use, distribution or reproduction in other forums is permitted, provided
end-expiratory pressure during pressure support ventilation in the original author(s) and the copyright owner(s) are credited and that the original
patients with chronic obstructive pulmonary disease. Crit. Care publication in this journal is cited, in accordance with accepted academic practice.
Med. (2007) 35:2547–52. doi: 10.1097/01.CCM.0000287594.801 No use, distribution or reproduction is permitted which does not comply with these
10.34 terms.

Frontiers in Medicine | www.frontiersin.org 10 November 2021 | Volume 8 | Article 752508