Abstract
BACKGROUND: Inhaled nitric oxide (INO) is used in infants as a therapy for elevated pulmonary vascular resistance. When INO is delivered at low tidal volumes, displayed inspiratory and expiratory volumes vary widely. We hypothesize that volume is removed by the sampling line during the ventilation cycle, and this results in a net volume loss at low tidal volumes. This study aimed to measure the volumes delivered and to assess the accuracy of displayed ventilator values using a test lung.
METHODS: A test lung was connected to a ventilator and an INO delivery system. All tests were performed with stable mode settings across volumes of 18, 30, 42, and 60 mL. Flow measured with a pneumotachometer attached between the test lung and the circuit assessed the percent error between inspiratory and expiratory volumes measured by the pneumotachometer measured and displayed on the ventilator under various INO/sample line conditions to determine where and how much volume was being displaced.
RESULTS: Displayed and measured inspiratory volumes had small variations between the INO/sample line conditions and baseline. However, expiratory volumes, with the sample line connected, exhibited large percent error values that increased (−14, −20, −27, and −34) as tidal volume decreased (60, 42, 30, and 18 mL) and error was significantly larger compared to baseline in all tidal volumes (P < .01) with and without INO delivery.
CONCLUSIONS: We concluded that inspiratory volumes were not affected by INO delivery, but additional removal of volume in the expiratory phase of the breath cycle by the sampling line results in a large error in the displayed expiratory volume.
Introduction
Inhaled nitric oxide (INO) causes selective pulmonary vasodilation by increasing cyclic guanosine monophosphate levels that relax smooth muscles. The Food and Drug Administration has approved the use of INO for persistent pulmonary hypertension in newborns, but there are numerous off-label uses of INO, including pulmonary hypertension related to ARDS and congenital heart disease. INO can be safely delivered to infants and children with effective local pulmonary vasodilation without systemic effects.1-3 Although the survival benefits of INO in acute lung injury remain controversial, the use of INO has been shown to reduce pulmonary artery pressure in patients with pulmonary hypertension, reduce pulmonary hypertension crises in high-risk patients after congenital heart defect surgery, and shorten duration of mechanical ventilation in patients with hypoxic respiratory failure.4-6 Therefore, INO delivery will continue to have a clinical indication in infants and requires safe delivery practices at low tidal volume (VT).
Safe delivery of INO requires continuous monitoring of oxygen, NO, and NO2 concentrations.7-9 During INO delivery with the INOMax DSIR (Mallinckrodt, Staines-upon-Thames, United Kingdom), NO is added by the injector module in proportion to the INO setting and the circuit flow to the inspiratory limb of the ventilator circuit prior to the humidifier, and the sample line continuously removes gas (0.23 L/min) for monitoring near the Y of the endotracheal tube, a distance that is required to allow for adequate mixing.10,11 At the bedside, we have observed that measured inspiratory and expiratory volumes vary widely when INO is being used at low VT. This discrepancy raises concern about the accuracy of the volume being delivered to the patient, especially at times when precise ventilation and oxygenation are imperative. Volumes introduced by INO and removed by sampling are estimated by the manufacturer (Ikaria) and would predict a net positive volume balance at high VT (ie, 500 mL predicts a net change of 0.2 L/min).12 However, clinically observed expired volumes are much less than inspired volumes, indicating a net negative volume balance. The manufacturer-provided formula (Net change = [INOMax dose × minute ventilation/INOMax concentration – INOMax dose] – 0.23 L/min) used with low volumes predicts a net volume loss. Therefore, we hypothesize that removal of volume by the sampling line during the respiratory cycle results in a net volume loss at low VT. Furthermore, given the location of the sampling line in the inspiratory limb, it is unclear whether the net volume loss occurs during the inspiratory phase, equally throughout the cycle, or during the expiratory phase, and how this net volume loss affects the minute ventilation of the patient. This was an in vitro study designed to measure the net volume difference at low VT to predict the effects of INO on minute ventilation. The primary outcome is the difference between the ventilator-displayed VT (inspiratory and expiratory VT) and pneumotachometer-measured VT (proximal inspiratory and expiratory VT).
Quick Look
Current Knowledge
Inhaled nitric oxide (INO) delivery devices interface with the ventilator circuit to deliver NO safely. Volume changes during INO delivery are more relevant at low tidal volumes and have not been studied. Infants receiving INO during mechanical ventilation have displayed volumes that vary widely and cause concern for accuracy of delivered volumes and the effects on minute ventilation at low tidal volumes.
What This Paper Contributes to Our Knowledge
On the basis of volumes measured by the ventilator, it would appear that there is a considerable net loss of volume with the INO delivery device, but in fact the patient receives a tidal volume very close to that which is prescribed, and the apparent net loss only occurs during expiratory phase sampling. This is due to continuous removal of volume by the sample line, the bulk of which occurs during the expiratory phase of the respiratory cycle.
Methods
This project was conducted at Arkansas Children’s Hospital Research Institute. A test lung (Bio-Tek Ventilator Tester, Fluke Biomedical, Everett, Washington) with the infant side was used for all tests. The test lung with compliance of 0.02 L/cm H2O and resistance of 8 cm H2O/0.5 L/s) was connected to the Servo-I ventilator (Maquet, Bridgewater, New Jersey) and the INOMax DSIR. The test lung was ventilated using pressure-regulated volume control ventilation, set PEEP of 5 cm H2O, set VT of 18, 30, 42, and 60 mL, and set breathing frequency of 30 breaths/min. The INO dose was set at 20 ppm for all test conditions. Measurements of respiratory flow and pressure waveforms were acquired using the Biopac MP100 System (Biopac Systems, Santa Barbara, California). A 0–35 L/min pneumotachometer (Hans Rudolph, Shawnee, Kansas) was utilized for data collection. The pneumotachometer was attached directly to the test lung in line with the ventilator circuit and the INO sample line (Fig. 1). Flow was calibrated using a flow meter. Volume measurements were obtained through the computer by integrating the flow signal. Volume was verified with a calibrated syringe (Hans Rudolph, Kansas City, Missouri). To monitor pressure, the pneumotachometer was equipped with a pressure hose and a barb-type port that allowed airway-pressure sampling. Pressure was calibrated with a manometer. All output signals were routed via an analog channel box into the Biopac MP100 data acquisition unit to convert them into digital signals that could then be processed by a computer.
Test lung model with Servo-i, INOMax, and pnuemotachograph.
Data were collected for each of the 4 VT settings in each of the 4 conditions: (A) baseline (ie, ventilator only); (B) ventilator with INO and without the sample line; (C) ventilator with the sampling line and without INO; and (D) the clinical scenario (ie, ventilator with INO and sampling line). Flow waveforms obtained with the pneumotachometer were analyzed to measure inspiratory and expiratory VT for 10 consecutive breaths. Values for inspiratory and expiratory VT that were displayed by the ventilator were matched time-wise and compared to measurements obtained with the pneumotachometer. Displayed measurements were collected from the ventilator via a memory card.
The primary outcome of differences between the ventilator-displayed inspiratory and expiratory VT and the pneumotachometer-measured inspiratory and expiratory VT were compared in terms of the inspiratory VT percent error and expiratory VT percent error between INO/sample line categories at each VT, where percent error = ([displayed VT – measured VT]/Measured VT) × 100.
The inspiratory and expiratory VT levels were not normally distributed; therefore, the Kruskal-Wallis test was used to identify differences by INO/sample line category within each tested VT level and by VT within each INO/sample line category. Where the omnibus Kruskal-Wallis test identified a difference between the groups, pair-wise comparisons were made with the Steel-Dwass adjustment for multiple comparisons between the categories and the reference (control) group.
For comparisons between INO/sample line categories within each VT, the control group was the baseline situation (ie, ventilator only, without INO and without the sampling line) and the primary group of interest was the clinical situation seen in practice (ie, ventilator with INO and sampling line). For comparisons between VT within each INO/sample line category, the control group was the VT = 60 mL because it is the highest volume. An alpha of 0.05 was used to assess the significance of all statistical tests. Data were analyzed with SAS 9.3 (SAS Institute, Cary, North Carolina) and JMP 11 (SAS Institute, Cary, North Carolina).
Results
Overall, the differences in volume during the inspiratory phase were small regardless of whether the sampling line was connected, whereas expiratory volume differences were significantly larger when the sampling line was connected.
Table 1 presents the displayed and measured volumes during the inspiratory phase. During inspiration, there was very little difference between displayed and measured VT and within each VT for the 4 conditions. When INO was added to the ventilator circuit (ie, condition B), the increase in VT became proportionally larger (ie, 20.5, 34, 47, and 66.8 mL, respectively) as INO gas was added to the circuit flow proportionally to deliver the desired INO concentration. This outcome was expected, thus increasing the percent error from baseline (Fig. 2). The conditions with the sampling line connected exhibited a VT closer to the ventilator-displayed VT, indicating the removal of a small amount of gas during the inspiratory phase of the ventilation cycle (ie, 1.5–2 mL). Steel-Dwass analysis revealed no significant difference in inspiratory VT percent error between the condition D (ie, INO and sample line) and the baseline condition (ie, ventilator only), except for condition D at VT of 18 mL (P = .032) (Table 1).
Comparisons of the 4 conditions of inspiratory VT error. A: Baseline condition. B: INO delivery without sample line. C: Sampling line without INO delivery. D: Clinical INO delivery. There were no significant differences between tidal volumes or conditions. VT = tidal volume.
Summary of Displayed and Measured Inspiratory VT Measurements
During the expiratory phase of ventilation, larger errors were noted between the volumes measured with the pneumotachometer and those displayed on the ventilator (Table 2). There was a statistically significant difference in percent error between the clinical scenario (ie, condition D) and the baseline condition at all VT settings (all Steel-Dwass P < .001). Thus, the expiratory VT percent error increased (−14%, −20%, −26%, and −34%) as VT decreased (60, 42, 30, and 18 mL), respectively, because the amount of volume removed by the sampling line remained closely fixed and thus had a proportional effect (Table 2). When comparing condition C (ie, sampling line connected without INO) to baseline, the expiratory VT percent error was significantly higher for all VT settings (P = .008, .001, .001, and .001 for VT of 18, 30, 42, and 60 mL, respectively). As VT decreases, the effect of having the sample line connected shows a significantly steeper increase in the expiratory VT percent error than when the sample line is disconnected and capped. When INO was added without the sampling line connected (ie, condition B), the expiratory VT percent error was not significantly different than baseline across all VT settings (Fig. 3).
Summary of Displayed and Measured Expiratory VT Measurements
Comparisons of the 4 conditions of expiratory VT error (%). When the sample line is attached, as in (C) and (D), there is a statistically significant larger error in volume than when it is not attached, as in (A; baseline condition) and (B). VT = tidal volume.
Discussion
Mechanical ventilation and delivery of INO at low VT has its challenges. No previous studies have looked at the volume changes with INO administration at low VT. Continuous sampling would be expected to have a proportionally greater impact at lower VT than at higher VT. Although continuous sampling occurs, our data demonstrate that volume removal is not equally distributed between the inspiratory phase and the expiratory phase of the ventilator cycle. There is a net loss of volume for each respiratory cycle, as predicted by the manufacturer’s formula, but most of that loss occurs during the expiratory phase. Therefore, the volume delivery to the patient is closer to the prescribed inspiratory VT than previously expected and would not expect to have dramatic effects on minute ventilation.
We found that the difference in the volume added by the introduction of INO into the ventilator circuit and the removal of volume by the sampling system resulted in a net gain of volume to the patient during the inspiratory phase of the breath cycle. We noted that there was a difference in displayed VT (18, 30, 42, and 60 mL) and VT measured at baseline, which resulted in a larger VT delivered to the test lung (20, 33, 46, and 65 mL, respectively). This baseline inspiratory VT error was not significant but represents an underestimation by the ventilator.
When examining the effect of the sampling line in isolation (ie, the difference in volume between condition A and condition C), the mean volume removed was 1.5–2 mL across all VT settings tested. During the expiratory phase, 7.5–9 mL were removed for sampling. However, at baseline an error of 3–4 mL was detected. Thus, even when accounting for the baseline error, the remaining volume removed by the sampling line was larger than the volume removed during the inspiratory phase. Therefore, the removal of additional volume in the expiratory phase of the breath cycle by the sampling system results in a percent error between the displayed expiratory VT and the measured expiratory VT that is larger than the inspiratory VT percent error. This is due to unequal sampling during the phases of the respiratory cycle. Sampling during inspiratory flow allows detection of 
, NO2 production in the humidifier, and NO delivery, which is necessary for safe delivery. However, sampling during expiration makes quantifying endotracheal leaks more difficult.
The manufacturer’s formula would predict a volume removal of 210 mL/min at VT = 18 mL (breathing frequency = 30 breaths/min), resulting in a loss of circuit volume of approximately 40%. However, our mea-surements indicate a removal of ∼135 mL/min (ie, 4.5 mL/cycle at the same breathing frequency). For the largest VT studied (ie, 60 mL), the measured volume removed per cycle was 180 mL/min, which was closer to the predicted value of 184 mL/min based on the formula and the predicted percent of loss (ie, 11%).
Therefore, the patient receives a VT close to the prescribed VT during inspiration because the baseline error, the addition of INO volume, and the volume removed from the sampling line result in delivered volumes close to the displayed inspiratory VT. During the expiratory cycle, volume is removed again after it leaves the patient and before it reaches the expiratory cassette. Therefore, the displayed expiratory VT does not indicate loss of prescribed inspiratory VT alone, but accounts for the total volume removed during the whole respiratory cycle, with the expiratory phase contributing to the majority of the lost volume.
We chose an in vitro design to isolate the volume changes with the addition of INO and the removal of volume by the sampling line. The use of a test lung allowed us to have consistent conditions in which to measure flows that would be expected to change during changes in resistance and compliance. We chose VT settings that would deliver 6 mL/kg to pediatric subjects with a weight of 3 kg, 5 kg, 7 kg, or 10 kg, which represents a clinical range in the population of interest. Our goal was to determine the volume change accuracy compared to the manufacturer-provided formula because there have been no previous studies examining net balance of volume at low VT settings. We also set out to determine whether continuous sampling resulted in equal volume removal across the respiratory cycle, which does appear to be the case according to our data.
This study has its limitations. We used only one ventilator, the Servo-I; other mechanical ventilators, modes of ventilation, and circuit designs may allow for different flow and sampling patterns that were not evaluated in our model. In addition, changes in a patients resistance and compliance may alter the pressures required for flow delivery and thus may allow for changes in sampling volumes taken by the sampling line.
Conclusions
Safe and effective use of INO at low VT requires understanding of volume changes that occur with INO proportional injection and continuous sampling. Proximal flow sensors may be the best way to ensure volume accuracy given our data in using a proximal pneumotachograph. More studies on this topic are warranted.
Acknowledgments
All the authors thank Ms. Casey Gormley for help in editing figure.
Footnotes
- Correspondence: Courtney D Ranallo MD, University of Oklahoma for Health Sciences, 1200 Everett Dr, Suite 8305, Oklahoma City, OK 73104. E-mail: courtney-ranallo{at}ouhsc.edu
See the Related Editorial on Page 1786
Dr Ranallo presented a version of this paper at the Pediatric Critical Care Colloquium in 2015 and at the 45th Critical Care Congress, held February 20–24, 2016, in Orlando, Florida.
The authors have disclosed no conflicts of interest.
- Copyright © 2020 by Daedalus Enterprises
