Professional Documents
Culture Documents
OBS - Accuracy of Invasive and Noninvasive Parameters For Diagnosing Ventilatory Overassistance During PSV - CCM (2017)
OBS - Accuracy of Invasive and Noninvasive Parameters For Diagnosing Ventilatory Overassistance During PSV - CCM (2017)
OBS - Accuracy of Invasive and Noninvasive Parameters For Diagnosing Ventilatory Overassistance During PSV - CCM (2017)
Objective: Evaluate the accuracy of criteria for diagnosing pres- in esophageal pressure during inspiration, and esophageal and air-
sure overassistance during pressure support ventilation. way occlusion pressure. In all definitions, the respiratory rate had the
Design: Prospective clinical study. greatest accuracy for diagnosing overassistance (receiver operat-
Setting: Medical-surgical ICU. ing characteristic area = 0.92; 0.91 and 0.76 for work of breath-
Patients: Adults under mechanical ventilation for 48 hours or more ing, pressure-time product and esophageal occlusion pressure in
using pressure support ventilation and without any sedative for 6 definition, respectively) and always with a cutoff of 17 incursions per
hours or more. Overassistance was defined as the occurrence of minute. In all definitions, a respiratory rate of less than or equal to 12
work of breathing less than 0.3 J/L or 10% or more of ineffective confirmed overassistance (100% specificity), whereas a respiratory
inspiratory effort. Two alternative overassistance definitions were rate of greater than or equal to 30 excluded overassistance (100%
based on the occurrence of inspiratory esophageal pressure-time sensitivity).
product of less than 50 cm H2O s/min or esophageal occlusion Conclusion: A respiratory rate of 17 breaths/min is the parameter
pressure of less than 1.5 cm H2O. with the greatest accuracy for diagnosing overassistance. Respi-
Interventions: The pressure support was set to 20 cm H2O and ratory rates of less than or equal to 12 or greater than or equal to
decreased in 3-cm H2O steps down to 2 cm H2O. 30 are useful clinical references to confirm or exclude pressure
Measurements and Main results: The following parameters were support overassistance. (Crit Care Med 2017; XX:00–00)
evaluated to diagnose overassistance: respiratory rate, tidal vol- Keywords: data accuracy; diagnosis; pressure support ventilation;
ume, minute ventilation, peripheral arterial oxygen saturation, rapid respiration, artificial/methods; respiration, artificial/utilization; work
shallow breathing index, heart rate, mean arterial pressure, change of breathing
N
1
Pulmonary Division, Heart Institute (InCor), Hospital das Clínicas, o precise guidelines are available for bedside adjust-
Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil.
ment of pressure support during pressure support
2
Intensive Care Unit, AC Camargo Cancer Center, São Paulo, Brazil.
ventilation (PSV); therefore, bedside adjustment is
3
Sírio-Libanês Teaching and Research Institute, São Paulo, Brazil.
not as easy as envisaged previously (1). Additionally, studies
Supplemental digital content is available for this article. Direct URL citations
appear in the printed text and are provided in the HTML and PDF versions about the adjustment of the level of pressure support have
of this article on the journal’s website (http://journals.lww.com/ccmjournal). focused on the risk of offering insufficient pressure support
Work was performed at the Heart Institute (InCor), Hospital das Clínicas, (underassistance) because unloading the respiratory muscle is
Faculdade de Medicina da Universidade de São Paulo, São Paulo, Brazil. one of the major objectives of mechanical ventilation (2, 3).
Supported and funded by Fundação de Amparo a Pesquisa do Estado de The adjustment of the level of pressure support during PSV is
São Paulo (Fapesp) (2012/09170-3 and 2010/08947-9), a governmental
nonprofit agency. No restrictions were placed on authors regarding the still a developing issue needing further investigation.
statements made in the manuscript. Adequate pressure support during PSV has not yet been
Dr. de Albuquerque received support for article research from FAPESP defined, probably because few studies have been performed
- Fundação de Amparo à Pesquisa do Estado de São Paulo, and he dis- about excessive pressure support during PSV (overassistance)
closed government work. The remaining authors have disclosed that they
do not have any potential conflicts of interest. and because it is difficult to recognize and diagnose subtle
For information regarding this article, E-mail: [email protected] underassistance. Previous studies that evaluated underassis-
Copyright © 2017 by the Society of Critical Care Medicine and Wolters tance initially measured noninvasive variables, such as min-
Kluwer Health, Inc. All Rights Reserved. ute volume (4–6), tidal volume (4, 5), respiratory rate (4, 5),
DOI: 10.1097/CCM.0000000000002871 vital capacity (6), recruitment of accessory inspiratory muscles
(5, 7, 8), and patient comfort (9), but these variables had little Trumbull, CT) were inserted. A T-piece was inserted between
accuracy for diagnosing underassistance. Other variables, such the ventilator y-connector and the ventilator circuit (Three-way
as work of breathing (WOB) (5, 6), esophageal occlusion pres- T-Shape Stopcock-Type, 2130 Series; Hans Rudolph, Shawnee,
sure (P0.1es) (4, 8) and diaphragmatic electromyography (10) KS) to allow proximal airway pressure monitoring and manual
were evaluated and had greater accuracy, but at the cost of occlusions. During the protocol, we continually recorded the fol-
invasiveness and less suitability to bedside use. lowing variables: proximal airway pressure, esophageal pressure,
Although overassistance may be frequent during PSV gastric pressure, and airway flow. The peripheral arterial oxygen
(11–15), it has not received the same attention as underassis- saturation, heart rate, and systolic and diastolic pressures were
tance, and to our knowledge, no study has evaluated the occur- continuously monitored and recorded from the multiparamet-
rence, diagnosing, and risk factors of overassistance during PSV. ric monitor. Initially, we recorded the variables over 20 minutes,
The probable reasons for that absence are that most patients keeping the ventilator as adjusted by the attending team (base-
receiving overassistance look comfortable, and the complica- line). After the baseline period, the oxygen inspiratory fraction,
tions, such as mechanical ventilation prolongation (16), occur trigger sensitivity, positive end-expiratory pressure (PEEP), and
later or are difficult to perceive, such as ventilator-induced cycling-off criteria were maintained as adjusted by the attending
lung injury (17, 18), hyperinflation (19), diaphragmatic atro- team, but the level of pressure support was set to 20 cm H2O and
phy and dysfunction (20), patient-ventilator asynchrony decreased in 3-cm H2O steps down to 2 cm H2O (20, 17, 14, 11,
(14–16), and sleep disorders (21). Although overassistance
8, 5, and 2 cm H2O). At all pressure support levels, we recorded
may be frequent and has many hazardous adverse effects, there
all variables during a 20-minute period. At any pressure sup-
are no criteria for diagnosing overassistance during PSV.
port level, the protocol could be interrupted before the end of
The primary objective of the present study was to test the
the 20-minute period if the patient had signs of respiratory dis-
accuracy of invasive and noninvasive criteria for diagnosing
tress, defined as the appearance of at least one of the following
overassistance during PSV. The secondary objective was to
signs: heart rate 140 or more beats/min, increased basal respira-
quantify the occurrence of overassistance during PSV.
tory rate of 50% or more, hypotension (systolic blood pressure
< 90 mm Hg) or hypertension (systolic blood pressure > 180 mm
METHODS Hg), use of accessory respiratory muscles, cardiac arrhythmia,
For further details, see supplementary material (Supplemen- diaphoresis, agitation, appearance of abdominal or thoracic par-
tal Digital Content 1, http://links.lww.com/CCM/D56). adoxical movements, and peripheral arterial oxygen saturation
of less than or equal to 90%.
Study Design At the end of each 20-min period or when the patient had
This was a prospective crossover study conducted in a medical showed signs of respiratory distress, we performed five non-
ICU of a tertiary teaching hospital (Respiratory ICU - Hospital consecutive airway occlusion maneuvers with 30- to 60-second
das Clinicas de São Paulo, Brazil). The study was approved by intervals between them. From the occluded cycles, we measured
the local Ethics Committee (0835/11), and a signed informed esophageal (P0.1es) and proximal airway (P0.1aw) occlusion pres-
consent was obtained from next of kin. sure (4, 22–24) and averaged the values of the five measures.
Data were acquired through an analog to digital converter
Patients
(ADS 1000; Lynx Eletronica, São Paulo, Brazil) connected to a
Patients were eligible during the first 48 hours of the transition
personal computer running software designed to acquire the
from a controlled mode (volume or pressure-controlled venti-
data (AqDados 7.2; Lynx Eletronica). Data were analyzed off-line
lation) to PSV and were consecutively recruited. The ICU had
using customized software (Labview 7.1; National Instruments,
no protocol to guide the transit from the controlled mode to
Austin, TX). For each pressure support level, the Labview soft-
an assisted mode of mechanical ventilation, so that transition
ware calculated a representative mean cycle through point-by-
was at the discretion of the attending team.
point averaging of all included respiratory cycles.
Inclusion criteria were age 18 years old or older, mechani-
cal ventilation for 48 hours or more, and absence of any sedative The mean proximal airway pressure, airway flow tidal
drugs for 6 hours or more. Patients were excluded if they had a volume, minute ventilation, respiratory rate, and gastric and
previous neuromuscular disease, a decreased level of consciousness esophageal pressures were calculated from the last 2 minutes of
(Richmond Agitation Sedation Scale score of –3, –4, or –5) or had the 20-minute period or when the patient had signs of respira-
been considered for withdrawal or withholding of life support. tory distress. The inspiratory WOB was calculated using the
At ICU admission, we recorded the demographic character- Campbell diagram (25) and represented the sum of the inspi-
istics, reason for ICU admission, reason for mechanical venti- ratory resistive flow and elastic WOB. The WOB was expressed
lation, and the Simplified Acute Physiology Score. as J per liter. The inspiratory esophageal pressure-time product
(PTPes) was calculated by the integration of esophageal pres-
Monitoring and Variables Recorded sure over inspiratory time, assuring that expiratory pressures
Immediately after patient inclusion, a pneumotachograph were excluded from the analysis of inspiratory PTPes (26, 27).
(MLT3813H-V; Hans Rudolph, Shawnee, KS), an esopha- PTPes was multiplied by the respiratory rate and expressed as
geal, and a gastric air-filled balloon catheter (CooperSurgical, cm H2O s/min (28).
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
TABLE 2. Cutoff Values, Area Under the ROC Curve, Sensitivity, Specificity, Positive
Predictive Value, and Negative Predictive Value of All Criteria Used to Diagnose
Overassistance
Receiver Operator Positive Negative
Characteristics Area Sensitivity Specificity Predictive Predictive
Criterion Cutoff Under the Curve (%) (%) Value (%) Value (%)
probably because the occurrence of underassistance is obvi- low respiratory rate, high tidal volume, and absence of sym-
ous with oxygen arterial desaturation, respiratory distress, pathetic nervous system activation. In addition to the evident
agitation, diaphoresis, and many other unsubtle signals, but difference in the clinical picture, overassistance probably
the occurrence of overassistance is subtle and easily con- did not garner attention because its complications are dif-
founded by comfort because overassistance is expressed by a ficult to diagnose or to associate with overassistance, such as
Figure 3. Receiver operating characteristic (ROC) curve of respiratory rate, rapid shallow breathing index, and esophageal (P0.1es) and proximal airway
(P0.1aw) occlusion pressure. Work of breathing (WOB) in definition (left graphic): ROC area of respiratory rate was significantly higher than the area of the
other parameters (p < 0.01 for the three comparisons). Rapid shallow breathing index vs P0.1es (p = 0.44); rapid shallow breathing index vs P0.1aw
(p = 0.60); P0.1aw vs P0.1es (p = 0.30). Pressure-time product in definition (center graphic): ROC area of respiratory rate was significantly higher than the
area of the other parameters (p < 0.01 for the three comparisons). Rapid shallow breathing index vs P0.1es (p = 0.07); rapid shallow breathing index vs
P0.1aw (p = 0.10); and P0.1aw vs P0.1es (p = 0.47). P0.1es in definition (right graphic): ROC area of respiratory rate was significantly higher than the area of the
other parameters (p <0.01 for rapid shallow breathing index and p = 0.02 for heart rate). p = 0.71 for rapid shallow breathing index vs heart rate.
P0.1es, P0.1aw, and change in esophageal pressure were not tested because P0.1es was in the definition of overassistance.
mechanical ventilation prolongation (16), ventilator-induced definition, a respiratory rate of 12 breaths/min or less confirms
lung injury (17, 18), hyperinflation (19), diaphragmatic atro- overassistance (100% sensitivity), whereas a respiratory rate
phy and dysfunction (20), patient-ventilator asynchrony of 30 or more excludes overassistance (100% sensibility). We
(14–16), and sleep disorders (21). believe that the description of these upper and lower bounds
Previous studies have suggested that the occurrence of over- can be useful clinical references in daily practice.
assistance is high, but they did not quantify it (11, 38). In the
present study, at baseline, overassistance occurred in 37–48% Limitations
of the patients, depending on the definition. Although, to our Our study has limitations. First, it was performed at a single
knowledge, no previous study has evaluated the occurrence center. Our ICU is in a teaching hospital dedicated exclusively to
and diagnosis of overassistance during PSV, many articles have respiratory care and admits patients with different respiratory
evaluated the best adjustment of pressure support during PSV diseases, so other ICUs with different patient or professional
(4–6, 8, 9, 39). These studies included different populations, profiles may have a different percentage of overassistance. Sec-
used different methods to find the best PSV adjustment, and ond, as mentioned previously in the Discussion section, our
aimed to evaluate mainly underassistance. In these articles, definitions of overassistance might have influenced the results;
respiratory rate and P0.1es were the more relevant accurate however, we tested three different definitions based on solid
parameters to adjust the PSV, which is in accordance with our physiologic knowledge and the results were similar because a
results. respiratory rate of 17 was the predictor with higher accuracy
in all definitions. Third, although at the beginning of the pro-
Clinical Relevance tocol, no patient was considered to a spontaneous breathing
The avoidance of overassistance is important to patients test by the attending team, 70% of patients tolerated a pressure
receiving mechanical ventilation, because as mentioned pre- support of 2 cm H2O, signaling that they were able to a sponta-
viously, overassistance is associated with severe mechanical neous breathing test. That discrepancy between the capability
ventilation-induced complications, such as mechanical venti- of the patients and the perception of spontaneous breathing
lation prolongation (16), ventilator-induced lung injury (17, test readiness by the physicians was previously described (40,
18), hyperinflation (19), diaphragmatic atrophy and dysfunc- 41). Fourth, our population was composed of hypercapnic
tion (20), patient-ventilator asynchrony (14–16), and sleep and hypoxemic patients and it is possible that those subgroups
disorders (21). To avoid treatment complications, the first respond differently to overassistance. A final limitation is that
step is awareness of the problem; we believe that the present we did not randomize the different pressure support levels, but
study sheds light on the problem. After awareness, the sec- we applied a descending order of the pressure support. We used
ond step is the capability to diagnose the complications and that strategy on purpose because we wanted to avoid the pos-
again we believe that the present study provides knowledge of sibility that an initial underassistance could cause deteriora-
the invasive and noninvasive parameters that can be used to tion in clinical condition and respiratory muscle fatigue, even
diagnose overassistance. We showed that independently of the a high-frequency fatigue that would hinder the continuation
Copyright © 2017 by the Society of Critical Care Medicine and Wolters Kluwer Health, Inc. All Rights Reserved.
Clinical Investigation
of the evaluation. In fact, other studies have used the same volume, and positive end-expiratory pressure. Am Rev Respir Dis
1988; 137:1159–1164
descending order, probably to avoid the same problems (4).
20. Hudson MB, Smuder AJ, Nelson WB, et al: Both high level pres-
In conclusion, excessive pressure support during PSV is sure support ventilation and controlled mechanical ventilation induce
frequent. In nonsedated or lightly sedated patients in pres- diaphragm dysfunction and atrophy. Crit Care Med 2012; 40:
sure support ventilation, a respiratory rate of 17 ipm is the 1254–1260
parameter with the greatest accuracy for diagnosing overassis- 21. Parthasarathy S, Tobin MJ: Effect of ventilator mode on sleep qual-
ity in critically ill patients. Am J Respir Crit Care Med 2002; 166:
tance and respiratory rates of 12 or less or 30 or more are use- 1423–1429
ful clinical references to confirm or exclude pressure support 22. Ward SA, Agleh KA, Poon CS: Breath-to-breath monitoring of inspi-
overassistance. ratory occlusion pressures in humans. J Appl Physiol Respir Environ
Exerc Physiol 1981; 51:520–523
23. Tobin MJ, Gardner WN: Monitoring of the control of breathing. In:
REFERENCES Principles and Practice of Intensive Care Monitoring. Tobin MJ (Ed).
1. Brochard LJ, Lellouche F: Pressure support ventilation. In: Principles New York, McGraw-Hill, 1998, pp. 415–464
and Practice of Mechanical Ventilation. Tobin MJ (Ed). New York, 24. Sassoon CS, Mahutte CK, Te TT, et al: Work of breathing and airway
McGraw Hill Companies, 2013, pp. 199–227 occlusion pressure during assist-mode mechanical ventilation. Chest
2. Goligher EC, Ferguson ND, Brochard LJ: Clinical challenges in 1988; 93:571–576
mechanical ventilation. Lancet 2016; 387:1856–1866 25. Campbell EJM: The Respiratory Muscles and the Mechanics of
3. Tobin MJ: Advances in mechanical ventilation. N Engl J Med 2001; Breathing. London, Lloyd-Luke, 1958
344:1986–1996 26. American Thoracic Society/European Respiratory Society: ATS/ERS
4. Alberti A, Gallo F, Fongaro A, et al: P0.1 is a useful parameter in set- Statement on respiratory muscle testing. Am J Respir Crit Care Med
ting the level of pressure support ventilation. Intensive Care Med 2002; 166:518–624
1995; 21:547–553 27. Jubran A, Van de Graaff WB, Tobin MJ: Variability of patient-ventilator
5. Banner MJ, Kirby RR, Kirton OC, et al: Breathing frequency and pat- interaction with pressure support ventilation in patients with chronic
tern are poor predictors of work of breathing in patients receiving obstructive pulmonary disease. Am J Respir Crit Care Med 1995;
pressure support ventilation. Chest 1995; 108:1338–1344 152:129–136
6. Van de Graaff WB, Gordey K, Dornseif SE, et al: Pressure support. 28. Sassoon CS, Light RW, Lodia R, et al: Pressure-time product during
Changes in ventilatory pattern and components of the work of breath- continuous positive airway pressure, pressure support ventilation, and
ing. Chest 1991; 100:1082–1089 T-piece during weaning from mechanical ventilation. Am Rev Respir
7. Brochard L, Harf A, Lorino H, et al: Inspiratory pressure support pre- Dis 1991; 143:469–475
vents diaphragmatic fatigue during weaning from mechanical ventila- 29. Cabello B, Mancebo J: Work of breathing. Intensive Care Med 2006;
tion. Am Rev Respir Dis 1989; 139:513–521 32:1311–1314
8. Perrigault PF, Pouzeratte YH, Jaber S, et al: Changes in occlusion 30. Mancebo J, Isabey D, Lorino H, et al: Comparative effects of pres-
pressure (P0.1) and breathing pattern during pressure support venti- sure support ventilation and intermittent positive pressure breath-
lation. Thorax 1999; 54:119–123 ing (IPPB) in non-intubated healthy subjects. Eur Respir J 1995;
9. Vitacca M, Bianchi L, Zanotti E, et al: Assessment of physiologic vari- 8:1901–1909
ables and subjective comfort under different levels of pressure sup- 31. Nava S, Bruschi C, Rubini F, et al: Respiratory response and inspi-
port ventilation. Chest 2004; 126:851–859 ratory effort during pressure support ventilation in COPD patients.
10. Hilbert G, Choukroun ML, Gbikpi-Benissan G, et al: Optimal pressure Intensive Care Med 1995; 21:871–879
support level for beginning weaning in patients with COPD: Measure- 32. Natalini G, Marchesini M, Tessadrelli A, et al: Effect of breathing pat-
ment of diaphragmatic activity with step-by-step decreasing pressure tern on the pressure-time product calculation. Acta Anaesthesiol
support level. J Crit Care 1998; 13:110–118 Scand 2004; 48:642–647
11. Emeriaud G, Larouche A, Ducharme-Crevier L, et al: Evolution of 33. Tobin MJ, Perez W, Guenther SM, et al: The pattern of breathing dur-
inspiratory diaphragm activity in children over the course of the PICU ing successful and unsuccessful trials of weaning from mechanical
stay. Intensive Care Med 2014; 40:1718–1726 ventilation. Am Rev Respir Dis 1986; 134:1111–1118
12. Giannouli E, Webster K, Roberts D, et al: Response of ventilator- 34. Murciano D, Aubier M, Bussi S, et al: Comparison of esophageal,
dependent patients to different levels of pressure support and pro- tracheal, and mouth occlusion pressure in patients with chronic
portional assist. Am J Respir Crit Care Med 1999; 159:1716–1725 obstructive pulmonary disease during acute respiratory failure. Am
13. Leung P, Jubran A, Tobin MJ: Comparison of assisted ventilator modes Rev Respir Dis 1982; 126:837–841
on triggering, patient effort, and dyspnea. Am J Respir Crit Care Med 35. Youden WJ: Index for rating diagnostic tests. Cancer 1950; 3:32–35
1997; 155:1940–1948 36. DeLong ER, DeLong DM, Clarke-Pearson DL: Comparing the areas
14. Thille AW, Cabello B, Galia F, et al: Reduction of patient-ventilator under two or more correlated receiver operating characteristic curves:
asynchrony by reducing tidal volume during pressure-support ventila- A nonparametric approach. Biometrics 1988; 44:837–845
tion. Intensive Care Med 2008; 34:1477–1486 37. Bossuyt PM, Reitsma JB, Bruns DE, et al; Standards for Reporting
15. Thille AW, Rodriguez P, Cabello B, et al: Patient-ventilator asynchrony of Diagnostic Accuracy: Towards complete and accurate reporting of
during assisted mechanical ventilation. Intensive Care Med 2006; studies of diagnostic accuracy: The STARD Initiative. Ann Intern Med
32:1515–1522 2003; 138:40–44
16. de Wit M, Miller KB, Green DA, et al: Ineffective triggering predicts 38. Ducharme-Crevier L, Du Pont-Thibodeau G, Emeriaud G: Interest of
increased duration of mechanical ventilation. Crit Care Med 2009; monitoring diaphragmatic electrical activity in the pediatric intensive
37:2740–2745 care unit. Crit Care Res Pract 2013; 2013:384210
17. The Acute Respiratory Distress Syndrome Network; Brower RG, Mat- 39. Kimura T, Takezawa J, Nishiwaki K, et al: Determination of the optimal
thay MA, Morris A: Ventilation with lower tidal volumes as compared pressure support level evaluated by measuring transdiaphragmatic
with traditional tidal volumes for acute lung injury and the acute respi- pressure. Chest 1991; 100:112–117
ratory distress syndrome. N Engl J Med 2000; 342:1301–1308 40. Epstein SK, Nevins ML, Chung J: Effect of unplanned extubation on
18. Dreyfuss D, Saumon G: Ventilator-induced lung injury: Lessons from outcome of mechanical ventilation. Am J Respir Crit Care Med 2000;
experimental studies. Am J Respir Crit Care Med 1998; 157:294–323 161:1912–1916
19. Dreyfuss D, Soler P, Basset G, et al: High inflation pressure pulmo- 41. Boles JM, Bion J, Connors A, et al: Weaning from mechanical ventila-
nary edema. Respective effects of high airway pressure, high tidal tion. Eur Respir J 2007; 29:1033–1056