Objective. The standard technique for positive-pressure ventilation is to regulate the breath size by varying the pressure applied to the bag. Investigators have argued that consistency of peak inspiratory pressure is important. However, research shows that excessive tidal volume delivered with excessive pressure injures preterm lungs, which suggests that inspiratory pressure should be varied during times of changing compliance, such as resuscitation of newborns or treatment after surfactant delivery.
Methods. We modified a computerized lung model (ASL5000 [IngMar Medical, Pittsburgh, PA]) to simulate the functional residual capacity of a 3-kg neonate with apnea and programmed it to change compliance during ventilation. Forty-five professionals were blinded to randomized compliance changes while using a flow-inflating bag, a self-inflating bag, and a T-piece resuscitator. We instructed subjects to maintain a constant inflation volume, first while blinded to delivered volume and then with volume displayed, with all 3 devices.
Results. Subjects adapted to compliance changes by adjusting inflation pressure more effectively when delivered volume was displayed. When only pressure was displayed, sensing of compliance changes occurred only with the self-inflating bag. When volume was displayed, adjustments to compliance changes occurred with all 3 devices, although the self-inflating bag was superior.
Conclusions. In this lung model, volume display permitted far better detection of compliance changes compared with display of only pressure. Devices for administration of positive-pressure ventilation should display volume rather than pressure.
Approximately 10% of the >4000,000 neonates born each year in the United States require some degree of resuscitation to transition from an intrauterine fluid-filled environment to an independent air-breathing existence.1 The most important transitional event is the establishment of air breathing. More than 12% of newborns are born prematurely, with particularly fragile lungs that can be easily damaged by excessive pressure administered during resuscitation.2–4 Defining the optimal respiratory support techniques for neonatal resuscitation has been identified as a priority by the International Liaison Committee on Resuscitation.5
The standard technique for assisting ventilation with positive pressure has been to regulate breath size through the amount of pressure applied to the bag.1, 6 The amount of pressure required to inflate a fluid-filled lung and to facilitate fluid absorption, such as occurs during the pulmonary transition associated with birth, is controversial and probably depends on whether the newborn has initiated spontaneous respirations.7 Several investigators have argued that consistency of peak inspiratory pressure (PIP) is an important variable and have demonstrated that a pressure-regulated T-piece resuscitator delivers more-consistent PIP than either a flow-inflating or self-inflating bag.8–10 However, research with preterm and term animals showed that excessive tidal volume delivered with or without excessive PIP is injurious to the lung,3, 4, 11 which suggests that PIP should be varied during times of changing compliance, such as during resuscitation of newly born infants12 and after surfactant administration.13, 14
A computerized electromechanical lung model (ASL5000 [IngMar Medical, Pittsburgh, PA]) that can be modified to reflect the functional residual capacity (FRC) of the human neonatal lung and can be programmed to change compliance during use has been developed. We programmed this lung model to exhibit compliance changes similar to those presented by a compromised newborn undergoing pulmonary transition. The current study was designed to use the ASL5000 lung model to test the ability of resuscitators to sense changes in pulmonary compliance while ventilating the simulator with a flow-inflating bag, a self-inflating bag, or a T-piece resuscitator.
The equipment used for this study included a standard ASL5000 lung simulator connected to a Dell D620 Latitude computer (Dell, Round Rock, TX), a standard, infant, flow-inflating bag (500-mL Hyperinflation system [Mercury Medical, Clearwater, FL]), a self-inflating bag (160-mL stroke-volume 1st Response [Portex, Keene, NH]), a T-piece resuscitator (Neopuff [Fisher-Paykell, Auckland, New Zealand]), and a cardiopulmonary monitor with neonatal flow sensor (Respironics NICO [Philips, Amsterdam, Netherlands]) that displayed delivered volume and pressure. Gas flow rates for the flow-inflating bag and T-piece resuscitator were set at 10 L/min.
We modified the basic programming of the lung model to simulate the FRC of a 3-kg neonate with apnea (100 mL).15 Two scripts were written, which instructed the model to change compliance according to 1 of 2 randomized patterns (Fig 1) while the model was being ventilated. Subjects were unaware of the pattern design. Both patterns began with a low-normal compliance (1.1 mL/cmH2O), followed by either a low compliance (0.5 mL/cmH2O) consistent with that of a neonate experiencing significant lung disease or a high compliance (1.8 mL/cmH2O) consistent with that of a neonate who had recently received surfactant. Subsequent changes were made as described in Fig 1, with the final epoch returning to the original normal compliance value. Although the model was programmed to remain at a given compliance for 14 breaths, the exact number of breaths at each compliance setting varied somewhat, depending on the rate at which any particular subject ventilated the model. Eight breaths with gradual compliance change to the next level were programmed to occur during transition periods but were not included in the data analysis. The compliance value for each breath was calculated by dividing the expiratory tidal volume by the inspiratory plateau pressure registered by the lung model. Because the subjects ventilated the model at slightly different rates and inspiratory times, the calculated compliance was often slightly different from that which had been programmed. Calculated compliance values and their associated pressures and volumes were then grouped into low-compliance (0.2–0.5 mL/cmH2O) and high-compliance (1.2–1.8 mL/cmH2O) subsets. Data from breaths administered during periods of normal compliance or transitional compliance were ignored.
Forty-five professionals (respiratory therapists, neonatal nurses, and physicians) associated with the University of Virginia NICU, with varying levels of experience, were recruited for the study. Study subjects were required to have participated in providing assisted ventilation during ≥5 neonatal resuscitations. The study was approved by the University of Virginia institutional review board, and informed consent was obtained from all participants. Subjects were told that the simulator would change compliance at unpredictable times and that they should attempt to maintain a constant tidal volume by altering the pressure delivered. They were given an opportunity to familiarize themselves with the model by ventilating it with each device, beginning with the one with which they were most familiar, while the model was set at a constant, low-normal compliance (1.0 cmH2O/mL) and no data were being collected. Most subjects started with the flow-inflating bag, because that is the device most commonly used at our institution. The volume delivered at baseline compliance and a pressure of 15 cmH2O should be ∼15 mL, or 5 mL/kg for a 3-kg neonate with a FRC of 100 mL. Subjects were asked to ventilate with a positive inflation pressure of 15 cmH2O with normal compliance, to get the “feel” of a normal range of tidal volumes with each device. Volumes between 12 and 18 mL (4–6 mL/kg) were considered acceptable. Although maintaining an inflation pressure of 15 cmH2O was the goal during the practice, participants were informed that, when the study began, the compliance would change and the goal would be to maintain constant tidal volumes, not constant inflation pressures. When subjects felt comfortable with the process at baseline compliance, they were asked to ventilate the model with either the flow-inflating bag or the self-inflating bag, according to either script A or script B. The type of bag and the preprogrammed script selected were determined by referring to a randomization list that had been computer-generated in blocks of 6. The subject was then asked to perform a second run using the other bag and a third run using the T-piece resuscitator. The script type was chosen randomly for each run. After subjects had completed 3 runs (ie, using each of the 3 devices) with only pressure displayed during the runs, they were asked to repeat the process with only volume displayed. Therefore, each subject completed 6 runs, that is, 3 with only pressure visible and 3 with only volume visible. Each subject was asked to complete a questionnaire inquiring whether they could detect the change in compliance and whether they could do so more easily with one device than with another.
Data on each breath from the ASL5000 were collected in real time by the computer and later were converted to ASCII format and stored in Excel (Microsoft, Redmond, WA) spreadsheets. These data were subsequently analyzed with SAS 9.1 (SAS Institute, Cary, NC). Inspiratory plateau pressure data were fit via linear mixed models for pressure and volume views, taking into account the repeated observations for each individual. Covariates in these models included compliance, device, compliance-device interaction, and breath number. Careful model selection strategies were implemented, leading to a final covariance model that included a random intercept (with different variances for the 2 compliance categories) and a first-order, autoregressive, within-subject variance. The sandwich estimator of the variance-covariance matrix of the fixed-effects parameters was used for inferences regarding the fixed-effect covariates of interest.16
Professional descriptions and levels of experience are shown in Table 1. There was a wide spectrum of experience, although all participants had been trained in neonatal care, most had participated in >20 resuscitations, and nearly all had completed the Neonatal Resuscitation Program of the American Academy of Pediatrics and the American Heart Association.
Table 2 presents the model-estimated mean pressures delivered for each compliance category with each of the 3 devices, first when only pressure was visible to the subject and then when only volume was visible. If subjects had been able to adjust pressure appropriately to deliver a constant volume, then delivered pressure would have increased significantly during low compliance and decreased during high compliance, to achieve a constant volume delivery of 15 mL. When data obtained when pressure alone was visible to the subjects were combined for all subjects, there was no significant change in delivered pressure when subjects used the flow-inflating bag (P = .727) or the T-piece resuscitator (P = .587). A slight but significant increase in pressure was observed with the self-inflating bag (P < .001), and the change was significantly different from results for the other 2 devices (P < .001).
When volumes were visible, subjects were able to change pressures in the appropriate directions to some degree with all 3 devices (P < .001). Changes with the self-inflating bag were greater than those with the other 2 devices (P < .001). These results are presented graphically in Fig 2.
Analysis of the volumes delivered for each compliance category is presented in Fig 3. When only pressure was visible, the mean volume decreased far below the minimum of the acceptable range during low compliance and far exceeded the maximal limit during high compliance. Individual volumes were as low as 1 mL/kg and as high as 16 mL/kg. When only volume was visible, subjects were able to deliver tidal volumes much closer to the acceptable range, with the self-inflating bag performing significantly better than either of the other devices (Fig 3).
Resuscitation guidelines for neonates focus on establishment of ventilation as a primary goal,5 as opposed to guidelines for resuscitation of adults, which usually focus more on reestablishment of circulation.17 If a neonate is compromised at birth, then current guidelines recommend that he or she first be stimulated to breathe and then be given positive-pressure breaths if ventilation is judged to be inadequate.1 Almost all such neonates can be successfully resuscitated with only these 2 interventions.18 Neonates born preterm are more likely to require positive-pressure ventilation during resuscitation, and there is evidence that even a brief period of excessive pressure or excessive tidal volume can cause significant lung injury.3, 4, 11 The appropriate pressure to use is somewhat controversial.5 Some authors advocate starting with 15 to 20 cmH2O and then increasing as needed to achieve an increase in heart rate.1 Others suggest using sufficient pressure to achieve a visible rise in the chest.19 Several investigators compared the ability of resuscitators to deliver a consistent pressure with various positive-pressure devices and found that pressures delivered with a T-piece resuscitator were more consistent than those delivered with either a flow-inflating or self-inflating bag.8–10 We suggest that, to avoid lung injury and to permit appropriate changes of ventilation to compensate for changes in lung compliance, the consistency and quantity of volume delivered are more important than the consistency of pressure. If pulmonary compliance is the independent variable and if volume is kept constant through changes in pressure, then it is likely that lung injury would be minimized while ventilation remains adequate. Adequate ventilation also can lead to appropriate, constant, carbon dioxide control, which would reduce swings in pulmonary and cerebral blood flow, the latter being particularly important in reducing the risk of germinal matrix hemorrhage.
Some have claimed that the stiffness of the lung and changes in pulmonary compliance can be “sensed” as one assists ventilation by delivering positive-pressure ventilation with a resuscitation bag.1 Two previous attempts to test this concept concluded that anesthesiologists were unable to detect even complete occlusion of an endotracheal tube while an infant with apnea was undergoing hand-ventilation during surgery. However, one study20 was conducted with a large (1000- or 2000-mL) mechanical test lung and either a Mapelson anesthesia circuit or pediatric circle system, both of which interposed a large volume of compressible gas between the person squeezing the bag and the site of obstruction (or decreased compliance). The other study21 also used various types of anesthesia breathing circuits involving large volumes of compressible gas within the circuits. Neither study examined the ability of clinicians to sense compliance changes when a resuscitation bag was connected directly to the model or a patient's endotracheal tube, as would occur during a typical neonatal resuscitation. Before the current study, we thought that, with neonatal bags and small-volume circuits, such as those used to deliver positive-pressure ventilation to neonates, it would be possible to detect compliance changes of the magnitude encountered during pulmonary transition after birth or after administration of surfactant to a neonate with surfactant-deficient lungs. Furthermore, we thought that compliance changes would be more easily detected with a flow-inflating bag, because of the highly compliant wall of the bag interposed between the resuscitator's hand and the gas volume within the bag. The ASL5000 lung model permitted us to test these ideas. The current study found that the average clinician could detect relatively large changes in compliance, but not very well. Furthermore, this ability seemed to occur only with the self-inflating bag, rather than the flow-inflating bag. We speculate that background flow through the flow-inflating bag and the control valve (which is not present in a self-inflating bag) may tend to mask the changing compliance. In addition, the mechanism of a self-inflating bag uses a 1-way valve to define a fixed volume; the volume change in the bag mirrors the volume change in the lung. In comparison, the flow-inflating bag is pressure-, flow-, and resistance-regulated; the volume of gas in the bag may be either delivered to the patient or released through the variable-resistance control valve. Therefore, not all of the volume from the bag is directed to the patient. Finally, the difference in sensitivity between the flow-inflating and self-inflating bags could reflect a difference in the compliance of the bag reservoirs themselves. The T-piece resuscitator has no reservoir, and it is not surprising that detection of changing compliance was difficult until delivered volume was displayed.
It should be noted that some subjects, who had considerable experience with providing respiratory assistance to neonates, were able to detect the compliance changes better than others, some even while using the flow-inflating bag. We plan to explore the possibility that the ASL5000 can be used to train individuals in this technique.
Although we think that the ASL5000 can be accurately programmed to model the changing compliance of neonatal lungs undergoing transition, it should be noted that the model does not permit the person delivering positive-pressure ventilation to use all of the cues that are commonly used during resuscitation (eg, chest rise and improving vital signs). These other signs are very subjective and, although they may be valuable for signaling insufficient ventilation, they may not be sensitive to delivery of excessive pressures and volumes, which have been shown to result in significant lung injury. Also, the model as programmed in the current study does not mimic the retention of lung volume that occurs as the FRC is being established during neonatal transition. Therefore, the volumes that are characteristic of postnatal ventilation and that we selected as target volumes for this study may not be appropriate for the resuscitation situation. Additional studies are needed to define the optimal volumes required to minimize barotrauma during neonatal resuscitation, so that resuscitators, using appropriate equipment, can be appropriately trained.
We decided to include the T-piece resuscitator in the study for several reasons. First, T-piece resuscitators are becoming standard equipment in many delivery areas and some intensive care nurseries. Second, several recent publications have advocated the devices as being preferable to either type of resuscitation bag because of the findings that T-piece resuscitators are associated with more-constant delivery of inflation pressures.8, 9 Because we knew of no way to sense compliance with the T-piece resuscitator and recognized that changing inflation pressures would be more cumbersome with the device, we did not anticipate that subjects would be very successful in adjusting to the changing compliance. Some subjects claimed that they thought they could detect compliance changes on the basis of changes in the sound of gas escaping through the adjustable valve, and it is quite possible that, if a volume display were included on the device and controls for adjusting pressures were made easier and more accessible to the operator, then the T-piece resuscitator might become the more-efficient device.
The results of this study have demonstrated that resuscitators can be far more responsive to compliance changes if they are able to view volumes, rather than simply pressures, when delivering positive-pressure ventilation to changing neonatal lungs. The optimal target volume is not known and likely changes somewhat as the FRC and compliance of the lung change after birth.2, 15, 22 However, tidal volumes exhibited by term newborns have been measured as ranging from 6 to 8 mL/kg,23 and preterm neonates with respiratory distress syndrome have tidal volumes of 4 to 6 mL/kg. Many experts in neonatal assisted ventilation are recommending a “gentle ventilation” strategy, using tidal volumes in the range of 3 to 6 mL/kg for neonates receiving mechanical ventilation. The optimal tidal volumes, inflation times, and rates required to establish a FRC and to achieve normal extrauterine oxygen saturation and Paco2 after birth are not known. We think that more studies evaluating the volume delivered and/or breathed spontaneously during transitional stabilization would help to define this optimal value. Until such data become available, we suggest that manufacturers develop resuscitation equipment for administration of manual ventilation that includes a volume display and that clinicians consider aiming to deliver normal tidal volumes, rather than being guided by prescribed pressure readings, when assisting the ventilation of neonates being resuscitated.
Funding was provided by a Neonatal Resuscitation Program Research Grant from the American Academy of Pediatrics and by the Department of Pediatrics, University of Virginia.
- Accepted December 5, 2008.
- Address correspondence to John Kattwinkel, MD, University of Virginia, Department of Pediatrics, Box 800386, Charlottesville, VA 22908. E-mail:
The authors have indicated they have no financial relationships relevant to this article to disclose.
What's Known on This Subject
Most devices for delivering positive-pressure ventilation are currently equipped with pressure displays only. Rapid changes in compliance occur after birth and after surfactant administration, requiring varying pressures to deliver optimal tidal volumes.
What This Study Adds
Neonatology professionals were unable to detect compliance changes sufficiently to adjust inflation pressures appropriately when only pressure was displayed, but performance improved significantly when volume was displayed. If resuscitation devices had volume displays, then lung injury might be avoided.
- ↵Kattwinkel J, ed. Textbook of Neonatal Resuscitation. 5th ed. Elk Grove Village, IL: American Academy of Pediatrics; 2006
- ↵Avery ME, Fletcher BD. The Lung and Its Disorders in the Newborn Infant. Philadelphia, PA: Saunders; 1974
- ↵International Liaison Committee on Resuscitation. International Liaison Committee on Resuscitation (ILCOR) consensus on science with treatment recommendations for pediatric and neonatal patients: neonatal resuscitation. Pediatrics.2006;117 (5). Available at: www.pediatrics.org/cgi/content/full/117/5/e978
- ↵Lindner W, Vossbeck S, Hummler H, Pohlandt F. Delivery room management of extremely low birth weight infants: spontaneous breathing or intubation? Pediatrics.1999;103 (5):961– 967
- ↵O'Donnell CP, Davis PG, Lau R, Dargaville PA, Doyle LW, Morley CJ. Neonatal resuscitation 2: an evaluation of manual ventilation devices and face masks. Arch Dis Child Fetal Neonatal Ed.2005;90 (5):F392– F396
- ↵Hernandez LA, Peevy KJ, Moise AA, Parker JC. Chest wall restriction limits high airway pressure-induced lung injury in young rabbits. J Appl Physiol.1989;66 (5):2364– 2368
- ↵Verbeke G, Molenberghs G. Linear Mixed Models for Longitudinal Data. New York, NY: Springer; 2000
- ↵Hazinski MF, Nolan JP, Becker LB, Steen PA. Controversial topics from the 2005 International Consensus Conference on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science With Treatment Recommendations. Circulation.2005;112 (suppl I):III-133– III-136
- ↵Upton CJ, Milner AD. Endotracheal resuscitation of neonates using a rebreathing bag. Arch Dis Child.1991;66 (1 spec No.):39– 42
- ↵Smith C, Nelson N, eds. The Physiology of the Newborn Infant. 4th ed. Springfield, IL: Charles C. Thomas; 1976:209
- Copyright © 2009 by the American Academy of Pediatrics