Cardiac Intensive Care
Impact of Positive End-Expiratory Pressure on Cardiac Index Measured by Ultrasound Cardiac Output Monitor* Oscar A. Ingaramo, MD1; Thienkim Ngo, MD2; Robinder G. Khemani, MD, MsCI2; Christopher J. L. Newth, MD, FRCPC2
Objectives: To evaluate the impact of different levels of positive endexpiratory pressure on cardiac index in children receiving mechanical ventilation. To explore the effect of lung recruitment on the relationship between positive end-expiratory pressure and cardiac output. Design: Prospective, single center, and interventional. Setting: PICU in a tertiary care children’s hospital. Patients: Fifty mechanically ventilated, hemodynamically stable children between 1 month and 20 years old. Interventions: Positive end-expiratory pressure was altered to levels of 0, 4, 8, and 12 cm H2O in random order. Cardiac output was measured at different levels of positive end-expiratory pressure by continuous wave Doppler ultrasound (ultrasound cardiac output monitor). Baseline vital signs were recorded, as well as cardiac index and dynamic compliance of the respiratory system at each positive end-expiratory pressure level. Measurements and Main Results: Median cardiac index decreased marginally as positive end-expiratory pressure increased, with a median change in cardiac index of 0.4 (< 10%) between positive end-expiratory pressure of 0 and 12 cm H2O (p < 0.001). There was no difference in heart rate or blood pressure as positive end-expiratory pressure increased (p > 0.5). For a subset of 29 patients (58%) in whom the highest dynamic compliance was at a positive end-expiratory pressure of 4 or 8 cm H2O, there was no *See also p. 84. 1 Department of Pediatric Critical Care, Sunrise Children’s Hospital, Las Vegas, NV. 2 Department of Anesthesiology and Critical Care Medicine, Children’s Hospital Los Angeles, University of Southern California Keck School of Medicine, Los Angeles, CA. This work was performed at Children’s Hospital Los Angeles, Los Angeles, CA. Dr. Ingaramo disclosed that the USCOM Company provided the USCOM ultrasound device for the study. Dr. Khemani received grant support from NIH. Dr. Newth consulted for Philips (respiratory equipment) and received support from NIH grants. Dr. Ngo disclosed that she does not have any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2013 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: 10.1097/PCC.0b013e3182976251
Pediatric Critical Care Medicine
difference in cardiac index between positive end-expiratory pressure 4 below versus positive end-expiratory pressure at highest dynamic compliance, or cardiac index between positive end-expiratory pressure 4 above versus positive end-expiratory pressure at highest dynamic compliance (p > 0.2). Regardless of optimal dynamic compliance, cardiac index decreased as positive endexpiratory pressure increased (p = 0.02). Conclusions: In hemodynamically stable mechanically ventilated children, although there is a statistically significant decrease in cardiac output as positive end-expiratory pressure is increased between 0 and 12 cm H2O, the mean change is less than 10%, and this is likely not clinically significant. In the presence of lung disease, intensive care physicians should feel less reluctant in their use of positive end-expiratory pressure for hemodynamically stable patients. (Pediatr Crit Care Med 2014; 15:15–20) Key Words: cardiac index; cardiac output; Doppler ultrasonography; dynamic compliance; mechanical ventilation; positive end-expiratory pressure
A
ccurate assessment of hemodynamic function is vital to the management of critically ill patients admitted to PICUs, as many have cardiac impairment, cardiac failure, or hemodynamic instability. For these patients, physicians titrate medical therapies, such as inotropes, vasopressors, and mechanical ventilation, with the goal of optimizing oxygen delivery through normalization of cardiac output (CO) and systemic vascular resistance. Positive end-expiratory pressure (PEEP) is part of the management of most patients on mechanical ventilatory support, particularly for those who have significant lung or lower airway disease. However, because PEEP elevates intrathoracic pressure and decreases venous return, clinicians are often reluctant to elevate PEEP for patients with hemodynamic compromise, even in the setting of severe lung disease (1, 2). In reality, PEEP has independent and difficult to predict effects on right and left ventricular (LV) function (3). Although the negative effects of PEEP on CO have been reported in animals and adults (4, 5), the actual impact of PEEP www.pccmjournal.org
15
Ingaramo et al
on CO in children has previously not been reported. Accurate assessment of CO (or if normalized to body surface area, cardiac index [CI]) requires objective measurement. Although thermodilution using a thermistor-tipped pulmonary artery (PA) floatation catheter is a gold standard for CI measurement (6), it is invasive, and there are risks involved, particularly in children (7). Newer technologies, such as continuous wave Doppler echocardiography, are now widely available and provide a noninvasive objective alternative to thermodilution (8). We sought to quantify the effect of PEEP in a clinically applicable range, on CI in children, as measured by continuous wave Doppler ultrasound. We elected to first concentrate on children with relatively normal hemodynamics. We hypothesized that the application of PEEP may affect CO, but that this effect will be clinically insignificant, when PEEP is in a range from 0 to 12 (9–11). We also sought to explore how the effect of lung recruitment modifies the relationship between PEEP and CO.
METHODS Patients This study was approved by the Institutional Review Board at Children’s Hospital Los Angeles. Children were eligible for inclusion if they were admitted to the PICU and were between 1 month and 20 years old and mechanically ventilated. Patients were excluded from the study if there was a leak around the endotracheal tube or tracheostomy greater than 18% (12). Air leak was calculated using the following formula: (VTI – VTE)/ VTI, where VTI is inspiratory tidal volume and VTE is expiratory tidal volume. Patients were also excluded if there was hemodynamic instability, defined as the use of any inotrope or vasopressor except dopamine less than or equal to 5 μg/kg/min and/ or heart rate (HR) more than 160 beats per minute; baseline
respiratory rate (RR) more than 60 breaths per minute; chronic lung disease; moderate or severe mitral or aortic valve regurgitation; arrhythmias; intracardiac shunts; or aortic aneurysm. Informed, written consent was obtained from the parent(s) or legal guardian(s) of the patient after suitable discussion. Ultrasonic Cardiac Output Monitor Device The USCOM 1A cardiac monitor (ultrasound cardiac output monitor, USCOM Pty Ltd, Coffs Harbour, NSW, Australia) has been approved by Food and Drug Administration in the United States since 2005. It uses continuous wave Doppler echocardiography to measure blood flow by the transthoracic approach across the aortic valve from the suprasternal notch and provides beat-to-beat CO and stroke volume. CI can be calculated if the patient’s body surface area is known. Previous studies have validated its interobserver variability and reproducibility (13). Prior to patient enrollment, an USCOM representative trained investigators on the usage of the instrument, and interobserver variability was measured to ensure minimal difference between the two investigators who performed the measurements (O.I.,T.N.). Interuser reliability was demonstrated by having each investigator measure the CI for 10 sample patients. Investigators were blinded to the other investigator’s measurement; three CI measurements were done on each patient and the average was taken for comparison. There was a mean difference less than 5% among investigators on near simultaneous measurements in all the patients.
Study Protocol Demographics and baseline vital signs were gathered for patients while on their clinician-selected ventilatory support. Two ventilators were used in our unit at the time of the study (Avea, Carefusion Corporation, San Diego, CA; and Servo-I, Maquet, Rastatt/Baden-Württemberg, Germany). Body surface area was calculated. PEEP was then altered to levels of 0, 4, 8, and 12 cm H2O in random order. All other ventilator settings (mode, rate, Fio2, pressure support, inspiratory time, or delta P [peak inspiratory pressure – PEEP] in pressure control mode) were held constant throughout the study protocol. All were on pressure-controlled mode. After 5 minutes of steady-state breathing on the new PEEP, CO was measured. Measurements were repeated a total of three times at each Figure 1. Baseline ventilator settings were recorded along with baseline vital signs and variables. Vital signs level to guarantee reliability included heart rate (HR), respiratory rate (RR), systolic blood pressure (BP), diastolic BP, mean arterial blood and reproducibility. Adverse pressure (MBP), and arterial oxygen saturation (Sao2). Variables included end-tidal Co2 (ETco2), dynamic events were tracked through compliance (Cdyn), and cardiac index. Positive end-expiratory pressure (PEEP) was changed in random order for each patient. *Cdyn was measured with a flow sensor at the endotracheal tube only if the ICU team continuous cardiopulmonary was using the device as part of the patient’s care. Otherwise tidal volume was calculated with appropriate monitoring during each of the adjustment for tubing compliance measured at the ventilator. 16
www.pccmjournal.org
January 2014 • Volume 15 • Number 1
Cardiac Intensive Care
Baseline Demographics and Characteristics Table 1.
Patients Demographics, Baseline Median Ventilator Settings, Vital Signs, and (Interquartile Range), Primary Reason for Mechanical Ventilation No. (%)
Age (mo)
16.5 (7.25, 85.5)
Gender (female)
26 (52%)
Weight (kg)
11.6 (7.325, 24.35)
Baseline cardiac index (L/min/m2)
4.45 (3.825, 5.4375)
Baseline ventilator settings Peak inspiratory pressure (cm H2O)
20 (14.25, 23.5) 5 (4, 6)
Positive end-expiratory pressure (cm H2O) Mean airway pressure (cm H2O) Fio2
8.5 (7, 10.75) 0.35 (0.3, 0.4)
Tidal volume (mL/kg)
7.8 (6.5, 9.4)
Baseline vital signs Heart rate (bpm)
128.5 (107.2, 138)
Systolic blood pressure (mm Hg)
100.5 (88.75, 110.75)
Diastolic blood pressure (mm Hg)
49 (42.75, 59.75)
Mean arterial blood pressure (mm Hg)
69 (57.25, 77.75)
Respiratory rate (bpm)
24 (19, 29.5)
Arterial oxygen saturation
100 (99, 100)
End-tidal Co2
39.5 (36, 43)
Dynamic compliance (mL/cm H2O/kg)
0.61 (0.42, 0.79)
Primary reason for mechanical ventilation, n Postsurgical
20
Respiratory
14
Trauma
4
Sepsis
2
Other
10
Median CIs (L/min/m2) and Interquartile Ranges at Different Levels of Positive End-Expiratory Pressure
Table 2.
CI
PEEP 0
PEEP 4
PEEP 8
PEEP 12
4.7 (4, 5.3)
4.6 (4, 5.6)
4.45 (3.7, 5.3)
4.275 (3.8, 5.1)
PEEP = positive end-expiratory pressure. The CI was measured three times at each level of PEEP and the average value was taken from patient.
difference was seen with repeated-measures ANOVA, then a Scheffé test was used for multiple comparisons between the levels of PEEP. To satisfy assumptions of normality, logarithmic transformations were employed. Similar analysis was performed on the raw hemodynamic variables (HR, SBP, DBP, and MBP), using values on a PEEP of 0 cm H2O as baseline. For example, the difference between HR at PEEP of 4, 8, and 12 cm H2O and HR on PEEP of 0 were compared for analysis. For secondary analysis of the impact of lung recruitment on CI, we graphed Cdyn as a function of PEEP for each patient, to determine the PEEP of highest (optimal) compliance. We then graphed CI at the PEEP level of optimal compliance, as well as CI at PEEP 4 cm H2O above and below PEEP of optimal Cdyn. We again used repeated-measures ANOVA, with a Scheffé test for multiple comparisons.
RESULTS Fifty patients were enrolled in the study. The median age was 16.5 months with an even distribution between males and females (Table 1). Baseline CI was 4.45 L/min/m2 with an interquartile range of 3.8–5.4 L/min/m2. Baseline ventilator settings (median) were peak inspiratory pressure of 20 cm H2O, mean airway pressure (MAP) of 8.5 cm H2O, PEEP of 5 cm H2O, inspiratory time of 0.8 seconds, and Fio2 of 0.35. Cdyn was 0.61 mL/cm H2O/kg. Primary reasons for mechanical ventilation are given in Table 1.
study conditions. Baseline characteristics and vital signs were recorded, as well as CI, HR, RR, end-tidal co2 (ETco2), oxygen saturation, and dynamic compliance (Cdyn) of the respiratory system at each PEEP level. When available, a proximal airway flow sensor specific to the individual patient’s ventilator was used for calculation of Cdyn (Fig. 1).
Changes in CI as a Function of PEEP When grouped, median CI decreased marginally as PEEP increased, with a median change in CI of 0.4 (< 10%) between PEEP of 0 and 12 cm H2O (Table 2). When analyzing CI as a function of PEEP, controlling for repeated measures per patient, there was a small but statistically significant effect of PEEP on CI (Fig. 2) (p < 0.001). This effect was seen between PEEP of 0 and 8, 0 and 12, 4 and 8, 4 and 12, and 8 and 12 cm H2O (p < 0.05).
Statistics Descriptive statistics for continuous data are presented as median and interquartile range, as data were not normally distributed. Categorical data are presented as number and frequency. Given that each patient was subject to different study conditions (and therefore measurements were not independent), repeated-measures analysis of variance (ANOVA) was used to detect differences in CI with varying PEEP. If a
Vital Sign Changes When comparing differences in HR and blood pressure, PEEP of 0 cm H2O was used to establish baseline values. Table 3 displays the median difference in HR and blood pressure among all patients as PEEP was increased. After controlling for repeated measures per patient, there was no difference in any of the hemodynamic variables as PEEP increased (p > 0.05). There was also no difference in RR (p > 0.05).
Pediatric Critical Care Medicine
www.pccmjournal.org
17
Ingaramo et al
by the lack of changes in other variables related to CI, such as HR and blood pressure. It implies that in a hemodynamically stable pediatric with overall mild lung disease and no intracardiac shunts or cardiac disease, the practitioner can apply PEEP within the range of 0–12 cm H2O to titrate to the individual patient’s pulmonary needs, with minimal impact on CI. We reiterate that the CO measurements were performed in stable patients and results might not extrapolate to patients with hemodynamic instability. In practice, the effects of PEEP on CI are complicated. PEEP can adversely affect CO by decreasing right ventricular (RV) preload and may also Figure 2. Cardiac index (CI) stratified by positive end-expiratory pressure (PEEP) (cm H2O). Log transformation increase RV afterload if pulwas used to satisfy assumptions of normality when evaluating the CI at differing levels of PEEP. Overall there was a statistically significant decrease in CI as PEEP increased (p < 0.001). Open circle denotes median monary vascular resistance of log values of CI. Asterisk denotes the difference in CI between the following PEEP levels was statistically is increased (14). Apart from significant: PEEP 0 and 8, PEEP 0 and 12, PEEP 4 and 8, PEEP 4 and 12, and PEEP 8 and 12. Hash venous return and RV filling, denotes the difference in CI between PEEP 0 and 4 was not statistically significant. the effects of PEEP on RV outSecondary Analysis: Lung Recruitment put depend on how PEEP changes lung volume relative to norWe determined the PEEP of optimal compliance for each mal functional residual capacity (FRC); the extent to which it can patient, defined as the PEEP where compliance was highest. alleviate hypoxic pulmonary vasoconstriction; and the overall Twenty-nine of the 50 patients (58%) had their optimal Cdyn change in pulmonary arterial pressure (15). An increase in pulat PEEP levels of 4 or 8 cm H2O; the other 21 were excluded monary arterial pressure increases RV afterload and impedes RV ejection. It is generally accepted that lung compliance is from secondary analysis because their optimal Cdyn was at highest at FRC, and pulmonary vascular resistance is lowest. a PEEP level at the lower or upper range (0 and 12, respecIf pulmonary vascular resistance is minimized, then LV filling tively), which would not make it possible for us to analyze may increase. Furthermore, increases in PEEP may decrease their CI above and below their PEEP of optimal Cdyn. AnalyLV afterload, by decreasing the transmural pressure across the sis of the 29 patients’ data showed that we could define optiLV (16). So, theoretically, if one achieves end-expiratory lung mal Cdyn, with statistically significant differences in Cdyn at PEEP 4 cm H2O above and below optimal Cdyn (Fig. 3A). volumes at FRC with PEEP, then RV and LV afterload should be minimized. However, RV filling may still be compromised However, regardless of optimal Cdyn, CI decreased as PEEP increased (Fig. 3B) (p = 0.02). On multiple comparisons with positive pressure ventilation of any sort, by decreasing venous return. analysis using the Scheffé test, this difference was predomiDuring our study, we attempted to identify the PEEP level nately from CI 4 below optimal PEEP compared to 4 above optimal PEEP (p = 0.007). There was no statistically signifi- closest to FRC (with the highest Cdyn). We attempted to cant difference between CI at optimal PEEP compared to 4 use this as a surrogate for where pulmonary vascular resistance would be lowest. Interestingly, CI was not highest at below or 4 above (p > 0.2). the PEEP of optimal compliance, but rather increasing levels of PEEP resulted in decreasing CI, as found by Richardson DISCUSSION et al (17) in studies on cats with induced lung disease. This We have demonstrated that although increases in PEEP between may imply that PEEP’s effect on reducing preload is the most 0 and 12 cm H2O have a statistically significant effect decreas- important. ing CI in hemodynamically stable, mechanically ventilated This is the first study of which we are aware that quantifies children, the magnitude of change of CI is small (
Impact of Positive End-Expiratory Pressure on Cardiac Index Measured by Ultrasound Cardiac Output Monitor* Oscar A. Ingaramo, MD1; Thienkim Ngo, MD2; Robinder G. Khemani, MD, MsCI2; Christopher J. L. Newth, MD, FRCPC2
Objectives: To evaluate the impact of different levels of positive endexpiratory pressure on cardiac index in children receiving mechanical ventilation. To explore the effect of lung recruitment on the relationship between positive end-expiratory pressure and cardiac output. Design: Prospective, single center, and interventional. Setting: PICU in a tertiary care children’s hospital. Patients: Fifty mechanically ventilated, hemodynamically stable children between 1 month and 20 years old. Interventions: Positive end-expiratory pressure was altered to levels of 0, 4, 8, and 12 cm H2O in random order. Cardiac output was measured at different levels of positive end-expiratory pressure by continuous wave Doppler ultrasound (ultrasound cardiac output monitor). Baseline vital signs were recorded, as well as cardiac index and dynamic compliance of the respiratory system at each positive end-expiratory pressure level. Measurements and Main Results: Median cardiac index decreased marginally as positive end-expiratory pressure increased, with a median change in cardiac index of 0.4 (< 10%) between positive end-expiratory pressure of 0 and 12 cm H2O (p < 0.001). There was no difference in heart rate or blood pressure as positive end-expiratory pressure increased (p > 0.5). For a subset of 29 patients (58%) in whom the highest dynamic compliance was at a positive end-expiratory pressure of 4 or 8 cm H2O, there was no *See also p. 84. 1 Department of Pediatric Critical Care, Sunrise Children’s Hospital, Las Vegas, NV. 2 Department of Anesthesiology and Critical Care Medicine, Children’s Hospital Los Angeles, University of Southern California Keck School of Medicine, Los Angeles, CA. This work was performed at Children’s Hospital Los Angeles, Los Angeles, CA. Dr. Ingaramo disclosed that the USCOM Company provided the USCOM ultrasound device for the study. Dr. Khemani received grant support from NIH. Dr. Newth consulted for Philips (respiratory equipment) and received support from NIH grants. Dr. Ngo disclosed that she does not have any potential conflicts of interest. For information regarding this article, E-mail: [email protected] Copyright © 2013 by the Society of Critical Care Medicine and the World Federation of Pediatric Intensive and Critical Care Societies DOI: 10.1097/PCC.0b013e3182976251
Pediatric Critical Care Medicine
difference in cardiac index between positive end-expiratory pressure 4 below versus positive end-expiratory pressure at highest dynamic compliance, or cardiac index between positive end-expiratory pressure 4 above versus positive end-expiratory pressure at highest dynamic compliance (p > 0.2). Regardless of optimal dynamic compliance, cardiac index decreased as positive endexpiratory pressure increased (p = 0.02). Conclusions: In hemodynamically stable mechanically ventilated children, although there is a statistically significant decrease in cardiac output as positive end-expiratory pressure is increased between 0 and 12 cm H2O, the mean change is less than 10%, and this is likely not clinically significant. In the presence of lung disease, intensive care physicians should feel less reluctant in their use of positive end-expiratory pressure for hemodynamically stable patients. (Pediatr Crit Care Med 2014; 15:15–20) Key Words: cardiac index; cardiac output; Doppler ultrasonography; dynamic compliance; mechanical ventilation; positive end-expiratory pressure
A
ccurate assessment of hemodynamic function is vital to the management of critically ill patients admitted to PICUs, as many have cardiac impairment, cardiac failure, or hemodynamic instability. For these patients, physicians titrate medical therapies, such as inotropes, vasopressors, and mechanical ventilation, with the goal of optimizing oxygen delivery through normalization of cardiac output (CO) and systemic vascular resistance. Positive end-expiratory pressure (PEEP) is part of the management of most patients on mechanical ventilatory support, particularly for those who have significant lung or lower airway disease. However, because PEEP elevates intrathoracic pressure and decreases venous return, clinicians are often reluctant to elevate PEEP for patients with hemodynamic compromise, even in the setting of severe lung disease (1, 2). In reality, PEEP has independent and difficult to predict effects on right and left ventricular (LV) function (3). Although the negative effects of PEEP on CO have been reported in animals and adults (4, 5), the actual impact of PEEP www.pccmjournal.org
15
Ingaramo et al
on CO in children has previously not been reported. Accurate assessment of CO (or if normalized to body surface area, cardiac index [CI]) requires objective measurement. Although thermodilution using a thermistor-tipped pulmonary artery (PA) floatation catheter is a gold standard for CI measurement (6), it is invasive, and there are risks involved, particularly in children (7). Newer technologies, such as continuous wave Doppler echocardiography, are now widely available and provide a noninvasive objective alternative to thermodilution (8). We sought to quantify the effect of PEEP in a clinically applicable range, on CI in children, as measured by continuous wave Doppler ultrasound. We elected to first concentrate on children with relatively normal hemodynamics. We hypothesized that the application of PEEP may affect CO, but that this effect will be clinically insignificant, when PEEP is in a range from 0 to 12 (9–11). We also sought to explore how the effect of lung recruitment modifies the relationship between PEEP and CO.
METHODS Patients This study was approved by the Institutional Review Board at Children’s Hospital Los Angeles. Children were eligible for inclusion if they were admitted to the PICU and were between 1 month and 20 years old and mechanically ventilated. Patients were excluded from the study if there was a leak around the endotracheal tube or tracheostomy greater than 18% (12). Air leak was calculated using the following formula: (VTI – VTE)/ VTI, where VTI is inspiratory tidal volume and VTE is expiratory tidal volume. Patients were also excluded if there was hemodynamic instability, defined as the use of any inotrope or vasopressor except dopamine less than or equal to 5 μg/kg/min and/ or heart rate (HR) more than 160 beats per minute; baseline
respiratory rate (RR) more than 60 breaths per minute; chronic lung disease; moderate or severe mitral or aortic valve regurgitation; arrhythmias; intracardiac shunts; or aortic aneurysm. Informed, written consent was obtained from the parent(s) or legal guardian(s) of the patient after suitable discussion. Ultrasonic Cardiac Output Monitor Device The USCOM 1A cardiac monitor (ultrasound cardiac output monitor, USCOM Pty Ltd, Coffs Harbour, NSW, Australia) has been approved by Food and Drug Administration in the United States since 2005. It uses continuous wave Doppler echocardiography to measure blood flow by the transthoracic approach across the aortic valve from the suprasternal notch and provides beat-to-beat CO and stroke volume. CI can be calculated if the patient’s body surface area is known. Previous studies have validated its interobserver variability and reproducibility (13). Prior to patient enrollment, an USCOM representative trained investigators on the usage of the instrument, and interobserver variability was measured to ensure minimal difference between the two investigators who performed the measurements (O.I.,T.N.). Interuser reliability was demonstrated by having each investigator measure the CI for 10 sample patients. Investigators were blinded to the other investigator’s measurement; three CI measurements were done on each patient and the average was taken for comparison. There was a mean difference less than 5% among investigators on near simultaneous measurements in all the patients.
Study Protocol Demographics and baseline vital signs were gathered for patients while on their clinician-selected ventilatory support. Two ventilators were used in our unit at the time of the study (Avea, Carefusion Corporation, San Diego, CA; and Servo-I, Maquet, Rastatt/Baden-Württemberg, Germany). Body surface area was calculated. PEEP was then altered to levels of 0, 4, 8, and 12 cm H2O in random order. All other ventilator settings (mode, rate, Fio2, pressure support, inspiratory time, or delta P [peak inspiratory pressure – PEEP] in pressure control mode) were held constant throughout the study protocol. All were on pressure-controlled mode. After 5 minutes of steady-state breathing on the new PEEP, CO was measured. Measurements were repeated a total of three times at each Figure 1. Baseline ventilator settings were recorded along with baseline vital signs and variables. Vital signs level to guarantee reliability included heart rate (HR), respiratory rate (RR), systolic blood pressure (BP), diastolic BP, mean arterial blood and reproducibility. Adverse pressure (MBP), and arterial oxygen saturation (Sao2). Variables included end-tidal Co2 (ETco2), dynamic events were tracked through compliance (Cdyn), and cardiac index. Positive end-expiratory pressure (PEEP) was changed in random order for each patient. *Cdyn was measured with a flow sensor at the endotracheal tube only if the ICU team continuous cardiopulmonary was using the device as part of the patient’s care. Otherwise tidal volume was calculated with appropriate monitoring during each of the adjustment for tubing compliance measured at the ventilator. 16
www.pccmjournal.org
January 2014 • Volume 15 • Number 1
Cardiac Intensive Care
Baseline Demographics and Characteristics Table 1.
Patients Demographics, Baseline Median Ventilator Settings, Vital Signs, and (Interquartile Range), Primary Reason for Mechanical Ventilation No. (%)
Age (mo)
16.5 (7.25, 85.5)
Gender (female)
26 (52%)
Weight (kg)
11.6 (7.325, 24.35)
Baseline cardiac index (L/min/m2)
4.45 (3.825, 5.4375)
Baseline ventilator settings Peak inspiratory pressure (cm H2O)
20 (14.25, 23.5) 5 (4, 6)
Positive end-expiratory pressure (cm H2O) Mean airway pressure (cm H2O) Fio2
8.5 (7, 10.75) 0.35 (0.3, 0.4)
Tidal volume (mL/kg)
7.8 (6.5, 9.4)
Baseline vital signs Heart rate (bpm)
128.5 (107.2, 138)
Systolic blood pressure (mm Hg)
100.5 (88.75, 110.75)
Diastolic blood pressure (mm Hg)
49 (42.75, 59.75)
Mean arterial blood pressure (mm Hg)
69 (57.25, 77.75)
Respiratory rate (bpm)
24 (19, 29.5)
Arterial oxygen saturation
100 (99, 100)
End-tidal Co2
39.5 (36, 43)
Dynamic compliance (mL/cm H2O/kg)
0.61 (0.42, 0.79)
Primary reason for mechanical ventilation, n Postsurgical
20
Respiratory
14
Trauma
4
Sepsis
2
Other
10
Median CIs (L/min/m2) and Interquartile Ranges at Different Levels of Positive End-Expiratory Pressure
Table 2.
CI
PEEP 0
PEEP 4
PEEP 8
PEEP 12
4.7 (4, 5.3)
4.6 (4, 5.6)
4.45 (3.7, 5.3)
4.275 (3.8, 5.1)
PEEP = positive end-expiratory pressure. The CI was measured three times at each level of PEEP and the average value was taken from patient.
difference was seen with repeated-measures ANOVA, then a Scheffé test was used for multiple comparisons between the levels of PEEP. To satisfy assumptions of normality, logarithmic transformations were employed. Similar analysis was performed on the raw hemodynamic variables (HR, SBP, DBP, and MBP), using values on a PEEP of 0 cm H2O as baseline. For example, the difference between HR at PEEP of 4, 8, and 12 cm H2O and HR on PEEP of 0 were compared for analysis. For secondary analysis of the impact of lung recruitment on CI, we graphed Cdyn as a function of PEEP for each patient, to determine the PEEP of highest (optimal) compliance. We then graphed CI at the PEEP level of optimal compliance, as well as CI at PEEP 4 cm H2O above and below PEEP of optimal Cdyn. We again used repeated-measures ANOVA, with a Scheffé test for multiple comparisons.
RESULTS Fifty patients were enrolled in the study. The median age was 16.5 months with an even distribution between males and females (Table 1). Baseline CI was 4.45 L/min/m2 with an interquartile range of 3.8–5.4 L/min/m2. Baseline ventilator settings (median) were peak inspiratory pressure of 20 cm H2O, mean airway pressure (MAP) of 8.5 cm H2O, PEEP of 5 cm H2O, inspiratory time of 0.8 seconds, and Fio2 of 0.35. Cdyn was 0.61 mL/cm H2O/kg. Primary reasons for mechanical ventilation are given in Table 1.
study conditions. Baseline characteristics and vital signs were recorded, as well as CI, HR, RR, end-tidal co2 (ETco2), oxygen saturation, and dynamic compliance (Cdyn) of the respiratory system at each PEEP level. When available, a proximal airway flow sensor specific to the individual patient’s ventilator was used for calculation of Cdyn (Fig. 1).
Changes in CI as a Function of PEEP When grouped, median CI decreased marginally as PEEP increased, with a median change in CI of 0.4 (< 10%) between PEEP of 0 and 12 cm H2O (Table 2). When analyzing CI as a function of PEEP, controlling for repeated measures per patient, there was a small but statistically significant effect of PEEP on CI (Fig. 2) (p < 0.001). This effect was seen between PEEP of 0 and 8, 0 and 12, 4 and 8, 4 and 12, and 8 and 12 cm H2O (p < 0.05).
Statistics Descriptive statistics for continuous data are presented as median and interquartile range, as data were not normally distributed. Categorical data are presented as number and frequency. Given that each patient was subject to different study conditions (and therefore measurements were not independent), repeated-measures analysis of variance (ANOVA) was used to detect differences in CI with varying PEEP. If a
Vital Sign Changes When comparing differences in HR and blood pressure, PEEP of 0 cm H2O was used to establish baseline values. Table 3 displays the median difference in HR and blood pressure among all patients as PEEP was increased. After controlling for repeated measures per patient, there was no difference in any of the hemodynamic variables as PEEP increased (p > 0.05). There was also no difference in RR (p > 0.05).
Pediatric Critical Care Medicine
www.pccmjournal.org
17
Ingaramo et al
by the lack of changes in other variables related to CI, such as HR and blood pressure. It implies that in a hemodynamically stable pediatric with overall mild lung disease and no intracardiac shunts or cardiac disease, the practitioner can apply PEEP within the range of 0–12 cm H2O to titrate to the individual patient’s pulmonary needs, with minimal impact on CI. We reiterate that the CO measurements were performed in stable patients and results might not extrapolate to patients with hemodynamic instability. In practice, the effects of PEEP on CI are complicated. PEEP can adversely affect CO by decreasing right ventricular (RV) preload and may also Figure 2. Cardiac index (CI) stratified by positive end-expiratory pressure (PEEP) (cm H2O). Log transformation increase RV afterload if pulwas used to satisfy assumptions of normality when evaluating the CI at differing levels of PEEP. Overall there was a statistically significant decrease in CI as PEEP increased (p < 0.001). Open circle denotes median monary vascular resistance of log values of CI. Asterisk denotes the difference in CI between the following PEEP levels was statistically is increased (14). Apart from significant: PEEP 0 and 8, PEEP 0 and 12, PEEP 4 and 8, PEEP 4 and 12, and PEEP 8 and 12. Hash venous return and RV filling, denotes the difference in CI between PEEP 0 and 4 was not statistically significant. the effects of PEEP on RV outSecondary Analysis: Lung Recruitment put depend on how PEEP changes lung volume relative to norWe determined the PEEP of optimal compliance for each mal functional residual capacity (FRC); the extent to which it can patient, defined as the PEEP where compliance was highest. alleviate hypoxic pulmonary vasoconstriction; and the overall Twenty-nine of the 50 patients (58%) had their optimal Cdyn change in pulmonary arterial pressure (15). An increase in pulat PEEP levels of 4 or 8 cm H2O; the other 21 were excluded monary arterial pressure increases RV afterload and impedes RV ejection. It is generally accepted that lung compliance is from secondary analysis because their optimal Cdyn was at highest at FRC, and pulmonary vascular resistance is lowest. a PEEP level at the lower or upper range (0 and 12, respecIf pulmonary vascular resistance is minimized, then LV filling tively), which would not make it possible for us to analyze may increase. Furthermore, increases in PEEP may decrease their CI above and below their PEEP of optimal Cdyn. AnalyLV afterload, by decreasing the transmural pressure across the sis of the 29 patients’ data showed that we could define optiLV (16). So, theoretically, if one achieves end-expiratory lung mal Cdyn, with statistically significant differences in Cdyn at PEEP 4 cm H2O above and below optimal Cdyn (Fig. 3A). volumes at FRC with PEEP, then RV and LV afterload should be minimized. However, RV filling may still be compromised However, regardless of optimal Cdyn, CI decreased as PEEP increased (Fig. 3B) (p = 0.02). On multiple comparisons with positive pressure ventilation of any sort, by decreasing venous return. analysis using the Scheffé test, this difference was predomiDuring our study, we attempted to identify the PEEP level nately from CI 4 below optimal PEEP compared to 4 above optimal PEEP (p = 0.007). There was no statistically signifi- closest to FRC (with the highest Cdyn). We attempted to cant difference between CI at optimal PEEP compared to 4 use this as a surrogate for where pulmonary vascular resistance would be lowest. Interestingly, CI was not highest at below or 4 above (p > 0.2). the PEEP of optimal compliance, but rather increasing levels of PEEP resulted in decreasing CI, as found by Richardson DISCUSSION et al (17) in studies on cats with induced lung disease. This We have demonstrated that although increases in PEEP between may imply that PEEP’s effect on reducing preload is the most 0 and 12 cm H2O have a statistically significant effect decreas- important. ing CI in hemodynamically stable, mechanically ventilated This is the first study of which we are aware that quantifies children, the magnitude of change of CI is small (
