Original Paper Received: March 24, 2014 Accepted after revision: July 27, 2014 Published online: November 6, 2014

Neonatology 2015;107:43–49 DOI: 10.1159/000366153

Influence of Gestational Age on Dead Space and Alveolar Ventilation in Preterm Infants Ventilated with Volume Guarantee Roland P. Neumann a, b Jane J. Pillow b–d Cindy Thamrin f Alexander N. Larcombe e Graham L. Hall e Sven M. Schulzke a, b  

 

 

a

 

 

 

Department of Neonatology, University Children’s Hospital Basel (UKBB), Basel, Switzerland; b Neonatal Clinical Care Unit, King Edward Memorial Hospital, c Centre for Neonatal Research and Education, School of Paediatrics and Child Health, d School of Anatomy, Physiology and Human Biology and e Telethon Kids Institute, University of Western Australia, Perth, and f Woolcock Institute of Medical Research, Glebe, Australia  

 

 

 

 

 

Key Words Very-low birth weight infants · Premature infants · Tidal volume · Artificial respiration

Abstract Background: Ventilated preterm infant lungs are vulnerable to overdistension and underinflation. The optimal ventilator-delivered tidal volume (VT) in these infants is unknown and may depend on the extent of alveolarisation at birth. Objectives: We aimed to calculate respiratory dead space (VD) from the molar mass (MM) signal of an ultrasonic flowmeter (VD,MM) in very preterm infants on volume-targeted ventilation (VT target, 4–5 ml/kg) and to study the association between gestational age (GA) and VD,MM-to-VT ratio (VD,MM/VT), alveolar tidal volume (VA) and alveolar minute volume (AMV). Methods: This was a single-centre, prospective, observational, cohort study in a neonatal intensive care unit. Tidal breathing analysis was performed in ventilated very preterm infants (GA range 23–32 weeks) on day 1 of life. Results: Valid measurements were obtained in 43/51 (87%) infants. Tidal breathing variables were analysed using multivariable linear regression. VD,MM/VT was negatively associated with GA after adjusting for birth weight Z score (p < 0.001, R2 = 0.26). This association was primarily influenced by the

© 2014 S. Karger AG, Basel 1661–7800/14/1071–0043$39.50/0 E-Mail [email protected] www.karger.com/neo

appliance dead space. Despite similar VT/kg and VA/kg across all studied infants, respiratory rate and AMV/kg increased with GA. Conclusions: VD,app rather than anatomical VD is the major factor influencing increased VD,MM/VT at a younger GA. A volume guarantee setting of 4–5 ml/kg in the Dräger Babylog® 8000 plus ventilator may be inappropriate as a universal target across the GA range of 23–32 weeks. Differences between measured and set VT and the dependence of this difference on GA require further investigation. © 2014 S. Karger AG, Basel

Introduction

Despite improvements in perinatal care, bronchopulmonary dysplasia (BPD) remains a major complication of very preterm birth and contributes significantly to neonatal morbidity and mortality [1]. Ventilator-induced lung injury, most notably volutrauma caused by excessively high tidal volume (VT), is an important factor in the pathogenesis of BPD in ventilated preterm infants [2].

This work was performed at the Neonatal Clinical Care Unit, King Edward Memorial Hospital, Perth, Australia.

Dr. Roland Neumann, MD Department of Neonatology Basel University Children’s Hospital (UKBB) Spitalstrasse 33, CH–4031 Basel (Switzerland) E-Mail roland.neumann @ ukbb.ch

Meta-analysis suggests that volume-targeted ventilation reduces the risk for BPD or death, probably due to the controlled delivery of VT that avoids alveolar overdistension or underinflation of the lungs [3]. The optimal VT for ventilated preterm infants remains unclear. Current recommendations arose from observational studies assessing the relationship between delivered VT and resulting arterial carbon dioxide tension (PaCO2) without appreciation of the proportional distribution of this volume to the respiratory dead space (VD) and the alveolar (terminal airspace) compartment or the stage of lung development [4, 5]. Whereas VT is measured at the airway opening, the removal of waste gas, i.e. CO2 is primarily determined by the alveolar minute volume (AMV), calculated as the product of the alveolar VT (VA) and respiratory rate (RR). Arguably, the VA delivered to ventilated preterm infants should increase with advancing maturation to achieve appropriate cyclic distension of the terminal airspaces. The volume of the terminal airspaces increases exponentially with the onset of the saccular and alveolar stages of lung development at around 26 and 34 weeks of gestational age (GA), respectively, with the full complement of alveoli not present until the age of at least 2–4 years [6, 7]. VA is calculated as the difference between VT and VD. VD represents (1) the anatomical VD, i.e. the conducting airways, and (2) the alveolar VD, i.e. the compartment of alveoli that is ventilated but not perfused and therefore does not participate in gas exchange. The sum of anatomical and alveolar VD represents physiological VD [8, 9]. In addition, appliance dead space (VD,app) is introduced into the system under conditions of positive pressure ventilation or during physiological measurements. Understanding changes in VD associated with advancing gestation may increase our understanding of appropriate targets for VT in ventilated preterm infants at different levels of maturity. The VD-to-VT ratio (VD/VT) should change with increasing GA as changes in terminal airspace volume exceed the changes in the volume of the airways over the third trimester. For any given VT setting on the ventilator, changes in the VD/VT with GA may influence the effective cyclic VA, and hence the efficiency of ventilation. Typically, VD is calculated using the Bohr-Enghoff equation (physiological VD) [8] or the Fowler graphical method (anatomical VD and VD,app) [9] using a CO2 expirogram. Reports of VD in neonates are limited to measurements in moderately preterm and term infants due to the large VD,app and the technical difficulties of expiratory plateau estimation in very preterm infants. Consequently, extrapolation of published data from term neonates and 44

Neonatology 2015;107:43–49 DOI: 10.1159/000366153

children do not yield reliable estimations of VD in those very preterm neonates at the highest risk for BPD. An alternative approach is to use the molar mass (MM) signal from a low-VD,app ultrasonic flowmeter as a surrogate for expired CO2. Whereas VD calculated from MM (VD,MM) has been validated against anatomical VD in spontaneously breathing neonates and infants [10], VD,MM values from ventilated preterm neonates have not been reported. We aimed to test the applicability of the ultrasonic flowmeter for VD,MM estimation in ventilated very preterm infants receiving volume-targeted ventilation on day 1 of life, and to investigate the association between GA and VD,MM/VT, VA and AMV. We hypothesized that after adjustment for birth weight Z score and VD,app, GA is negatively associated with VD,MM/VT and positively associated with VA and AMV.

Methods Design This prospective, observational study was performed within the Neonatal Clinical Care Unit of King Edward Memorial Hospital, Perth, Western Australia. The study protocol was approved by the Human Research Ethics Committee of the Women’s and Children’s Health Service in Western Australia. Written informed parental consent was obtained prior to enrollment. Subjects Ventilated very preterm infants (GA range 230–316 weeks) presenting with respiratory distress syndrome were eligible for study on day 1 of life. Endotracheal intubation was at the discretion of the attending neonatal consultant. Routine unit intubation criteria included prolonged apnoea, severe respiratory distress, FiO2 >0.3 and PaCO2 >60 mm Hg. Infants with major thoracic or cardiac malformations, neuromuscular disorders or a significant leak around the tracheal tube (>20%) were excluded. All patients received prophylactic surfactant (Poractant alfa, 200 mg/kg; Chiesi Farmaceutici, Parma, Italy) and were ventilated using patient-triggered ventilation (Dräger Babylog® 8000 plus, Dräger Medical, Lübeck, Germany). Routine clinical ventilation strategy included setting the target VT to 4–5 ml/kg body weight (‘volume guarantee’ mode) and supporting each respiratory effort by assist-control or pressure support ventilation. Positive end-expiratory pressure was usually set at 5 cm H2O to avoid atelectasis. Ventilator settings were adjusted at the discretion of the attending neonatologist. Lung Function Tests We recorded 120 s of tidal breathing during patient-triggered ventilation in quiet, unsedated sleep in a supine position in accordance with international standards for infant lung function testing [11, 12]. Measurements were obtained using a commercially available apparatus incorporating an ultrasonic flowmeter (Exhalyzer D, Ecomedics, Dürnten, Switzerland). Immediately prior to the measurement, the ultrasonic flowmeter was positioned in series between the flow sensor of the ventilator (located at the distal end

Neumann/Pillow/Thamrin/Larcombe/ Hall/Schulzke

Ultrasonic flowmeter

Fig. 1. Schematic of measurement setup. The ventilator wye-piece was connected to the ultrasonic flowmeter which was attached in series distal to the ventilator flow sensor and the endotracheal tube. VD,app of the ultrasonic flowmeter is 1.1 ml and VD of the flow sensor of the ventilator is 0.9 ml.

Ventilator wye-piece

of the endotracheal tube, ETT) and the ventilator wye-piece, introducing an additional VD,app for lung function testing of 1.1 ml (fig. 1). Flow (range ± 0.5 liters/s and resolution 0.6 ml/s) and MM (range 20–45 g/mol and resolution 0.01 g/mol) were sampled at 200 Hz and corrected to body temperature and ambient pressure, and saturated with water vapor parameters. VD,MM was calculated as reported previously [10]. Briefly, VD,MM was estimated as the volume of air that passes through the ultrasonic flowmeter until the MM signal augments to its 10% rise point. VD,MM was calculated breath-by-breath from at least 5 breaths and averaged to obtain the mean VD,MM for the whole measurement. Breaths were excluded in the presence of irregular MM profile or uncertainty of the 10% MM rise point [10]. VA was calculated as VT – VD,MM, as a surrogate for alveolar VT. Analysis algorithms were written in Matlab (The Mathworks Inc., Natick, Mass., USA). Arterial blood gases were obtained immediately after lung function testing from an umbilical catheter. Sample Size Estimates Aiming for 80% power at the 5% significance level, we calculated a minimum required sample size (n = 42) for a multiple linear regression model with 2 independent continuous predictor variables (birth weight Z score, GA) of medium-to-large-effect size (f2 = 0.25) estimating VD,MM [13]. We planned to recruit 51 infants, assuming a dropout rate of about 15% due to ETT leak, absence of regular tidal breathing and technical difficulties in VD,MM estimation.

ETT connector

Ventilator flow sensor

ETT

further using stepwise multivariable linear regression analysis, in which p < 0.05 was considered statistically significant. Multivariable models were compared using the likelihood ratio test. Statistical analyses were performed using Stata 10 software (Stata Corporation, College Station, Tex., USA).

Results

Applicability of the Ultrasonic Flowmeter for Estimation of VD,MM in Ventilated Preterm Infants The study enrolled 51 preterm infants with a mean GA of 27.5 ± 2.6 weeks who were measured at a median [interquartile range (IQR)] age of 14.5 h (8.0–19.0 h) of life. Acceptable data were obtained from 43 out of 51 (87 %) infants. On average, 26 (SD ± 8.3) breaths were used for VD,MM calculation. The median (IQR) coefficient of variation in VD,MM was 19% (15–25%). Data from 8 infants were excluded due to difficulties in VD,MM estimation in the presence of irregular tidal breathing. Demographic and clinical characteristics of included and excluded patients are given in table 1 and online supplementary table 1 (for all online suppl. material, see www.karger.com/ doi/10.1159/000366153).

Statistical Analyses The influence of GA on VD,MM/VT, (VD,MM – VD,app)/VT, VA and AMV was investigated using multivariable linear regression analysis. Explanatory variables explored in univariable regression analysis included sex, ethnicity, postnatal age at test, GA, body weight and length at birth, birth weight Z score (calculated from the British Growth Reference [14]), household and maternal smoking in pregnancy, clinical chorioamnionitis, prenatal corticosteroids, peak inspiratory pressure, mean airway pressure, positive end-expiratory pressure, inspired oxygen concentration, oxygenation index, PaCO2, RR, minute volume (MV), set VT (‘volume guarantee’) and ETT leak. Factors potentially influencing outcome variables in univariable regression analysis (p < 0.1) were explored

Lung Function Results Table  2 shows the mean values for RR, VT, VD,MM, VA,  MV, AMV, VD,MM – VD,app, VD,MM/VT, (VD,MM – VD,app)/VT and PaCO2. Univariable regression analysis (online suppl. table 2; fig. 2) suggested significant negative associations of VD,MM/VT with GA (p = 0.001, R 2 = 0.23) and body size (birth weight, p = 0.001, R 2 = 0.23; birth length, p < 0.001, R 2 = 0.26). VD,MM/VT was not associated with any of the other explanatory variables considered (online suppl. table  2). Online supplemen-

Respiratory Dead Space in Ventilated Very Preterm Infants

Neonatology 2015;107:43–49 DOI: 10.1159/000366153

45

Table 1. Anthropometric and clinical characteristics of included

patients

0.60

Table 2. Results of lung function measurements (n = 43)

Variable

Mean ± SD

RR, breaths/min VT/kg, ml/kg VD,MM/kg, ml/kg VA/kg, ml/kg MV/kg, ml/kg/min AMV/kg, ml/kg/min (VD,MM – VD,app)/kg, ml/kg VD,MM/VT (VD,MM – VD,app)/VT PaCO2, mm Hg

70.6±14.6 6.57±1.54 2.51±0.61 4.06±1.14 427.8±92.8 283.5±70.0 1.43±0.38 0.39±0.06 0.22±0.04 42.9±5.8

tary table 3 summarizes the results of multivariable regression analyses: VD,MM/VT was negatively associated with GA after adjusting for birth weight Z score (p  < 0.001, R 2  = 0.26). The negative association between VD,MM/VT and GA was not significant after correcting for VD,app (p = 0.22, R 2 = 0.07; fig. 2). Using the equivalent model, VA/kg was not associated with GA whereas AMV/kg showed a positive association with GA (p  = 0.03, R 2  = 0.17). VT/kg was negatively associated with GA (p = 0.03, R 2 = 0.20; fig. 3) whereas the association between VD,MM/kg and GA was no longer significant after correction for VD,app (p  = 0.17, R 2  = 0.05; online suppl. fig.  1). Breathing variables positively associated Neonatology 2015;107:43–49 DOI: 10.1159/000366153

0.40

0.30

0.20 24

26

a

28

30

32

30

32

GA (weeks)

0.30

4.1±0.3

Data are presented as n (%) or mean ± standard deviation unless stated otherwise.

46

0.50 VD,MM/VT

43 (56) 28 (65) 14.5 (8.0–19.0) 27.9±2.5 23.6–32.0 1.14±0.38 –0.09±0.8 36.4±4.1 8.4±1.3 23 (54) 40 (93) 18 (42) 12.4±5.2 4.9±0.15 6.9±1.7

(VD,MM – VD,app)/VT

Number of subjects (% males) Caucasian ethnicity Median (IQR) postnatal age at test, h GA, weeks GA range, weeks Birth weight, kg Birth weight Z score Birth length, cm BMI, kg/m2 Household smoking Prenatal steroids Chorioamnionitis Peak inspiratory pressure, cm H2O Positive end-expiratory pressure, cm H2O Mean airway pressure, cm H2O Set VT (‘volume guarantee’), ml/kg body weight

0.25

0.20

0.15

0.10

b

24

26

28 GA (weeks)

Fig. 2. VD/VT over GA. a VD,MM per VT. b (VD,MM – VD,app) per VT. Linear regression analyses demonstrated a significant negative association between VD,MM/VT and GA (p = 0.001, R2 = 0.23). After subtracting VD,app from VD,MM, the association between VD/VT and GA was no longer significant, indicating the more dominant impact of VD,app on total VD in very preterm infants (p = 0.17, R2 = 0.05).

with GA included RR (p < 0.01, R 2 = 0.49) and MV/kg (p = 0.019, R 2 = 0.13; fig. 4). There was no association between PaCO2 and GA (p = 0.65). Discussion

Measurements of VD,MM were highly feasible in ventilated preterm neonates with an overall success rate of 43/51 (87%). As hypothesized, VD,MM/VT and GA were negatively associated but this association was heavily influenced by VD,app. In addition, despite no difference in VA across the spectrum of GA of the studied infants, Neumann/Pillow/Thamrin/Larcombe/ Hall/Schulzke

12 8

VA/kg (ml/kg)

VT/kg (ml/kg)

10

8

6

4

24

26

a

28

30

b

GA (weeks)

kg body weight. Linear regression analyses demonstrated a significant negative association between V T/kg and GA (p = 0.03,

24

26

28 GA (weeks)

30

32

R 2 = 0.20) but no significant association between VA/kg and GA (p = 0.13, R 2 = 0.06).

100

800 MV/kg and AMV/kg (ml/kg)

RR (breaths/min)

4

2

32

Fig. 3. Change in volumes with GA. a V T/kg body weight. b VA /

80

60

40 24

a

6

26

28

30

400

200

0

32

GA (weeks)

600

b

24

26

28 GA (weeks)

30

32

Fig. 4. Relation of breathing variables to GA. a RR. b MV/kg body weight (solid circles) and AMV/kg (hollow open squares). Linear regression analyses demonstrated significant positive associations

between RR and GA (p < 0.001, R2 = 0.49), MV/kg and GA (solid regression line, p = 0.019, R2 = 0.13) and AMV/kg and GA (dashed regression line, p = 0.03, R2 = 0.17).

there was a positive association between GA and AMV/ kg, primarily due to a positive association between GA and RR. Despite the known exponential growth of the parenchymal compartment over the last trimester and despite more linear growth of the major airways, there are no published reports on the association of VD/VT and GA in ventilated preterm neonates. The mean values of VD,MM (2.5 ± 0.6 ml/kg) and VD,MM/VT (0.39 ± 0.06) as well as of VD,MM corrected for VD,app (mean VD,MM – VD,app, 1.4 ± 0.4 ml/kg; (VD,MM – VD,app)/VT 0.21 ± 0.05) in our study

are lower than those reported in other studies in intubated infants [14]. A higher mean physiological VD of 2.9 ± 0.9 ml/kg and VD/VT of 0.56 ± 0.08 by means of the BohrEnghoff method is reported in intubated neonates of 27– 41 weeks GA [15]. This discrepancy might be explained by methodological differences, with the VD,MM from our study reflecting anatomical VD (Fowler VD) while the Bohr-Enghoff VD includes both anatomical and alveolar VD. Moreover, VD,app in the study employing the BohrEnghoff method was substantially larger than that of our setup (2.6 ml vs. 1.1 ml), and the median postmenstrual

Respiratory Dead Space in Ventilated Very Preterm Infants

Neonatology 2015;107:43–49 DOI: 10.1159/000366153

47

age (34.5 vs. 27.9 weeks) and postnatal age at test (14.0 vs. 0.6 days) were higher than in our population. Numa and Newth [16] measured intrathoracic and extrathoracic VD in ventilated term neonates and reported an average total VD of 3.3 ml/kg. Their estimations are likely to be higher than those from our study, given that the value for extrathoracic VD in their study included the volume of mouth, nasopharynx, oropharynx and larynx. Our study extends previous observations as it includes VD calculation in very and extremely preterm ventilated neonates. Data on VD and VD/VT in these infants are extremely rare due to the methodological difficulties of estimating VD using CO2 sensors in very preterm ventilated neonates. The major strengths of our study were: (1) the recruitment of very preterm infants from a GA of 23 weeks, who received contemporary neonatal intensive care including prophylactic surfactant and volume-targeted ventilation, (2) the experimental setup, using a fast ultrasonic flowmeter (200 Hz sampling frequency) that measures flow and MM simultaneously without need for delay correction and (3) the miniaturized lung function equipment which introduced a VD,app of only 1.1 ml, which follows international guidelines that recommend adding a VD,app of 

Influence of gestational age on dead space and alveolar ventilation in preterm infants ventilated with volume guarantee.

Ventilated preterm infant lungs are vulnerable to overdistension and underinflation. The optimal ventilator-delivered tidal volume (VT) in these infan...
208KB Sizes 0 Downloads 7 Views