CARBON DIOXIDE CLEARANCE DURING HIGH FREQUENCY JET VENTILATION: Effect of Deadspace in a Lung Model

CARBON DIOXIDE CLEARANCE AND DEADSPACE DURING HIGH FREQUENCY JET VENTILATION: Investigations in the Dog

Carbon dioxide clearance using bubble CPAP with superimposed high-frequency oscillations in a premature infant lung model with abnormal lung mechanics.

The effect of increased apparatus dead space and tidal volumes on carbon dioxide elimination and oxygen saturations in a low-flow anesthesia system

High‐flow nasal therapy – modelling the mechanism

High-Frequency Oscillatory Ventilation for Adult Patients With ARDS

Mechanical ventilation with high tidal volume or frequency is associated with increased expression of nerve growth factor and its receptor in rabbit lungs

What is the central question of this study? The goal of this study was to investigate the effect of alterations in tidal volume and alveolar volume on the elevated physiological dead space and the contribution of ventilatory constraints thereof in heart failure patients during submaximal exercise. What is the main finding and its importance? We found that physiological dead space was elevated in heart failure via reduced tidal volume and alveolar volume. Furthermore, the degree of ventilatory constraints was associated with physiological dead space and alveolar volume. Patients who have heart failure with reduced ejection fraction (HFrEF) exhibit impaired ventilatory efficiency [i.e. greater ventilatory equivalent for carbon dioxide ( ) slope] and elevated physiological dead space (VD /VT ). However, the impact of breathing strategy on VD /VT during submaximal exercise in HFrEF is unclear. The HFrEF (n= 9) and control (CTL, n= 9) participants performed constant-load cycling exercise at similar ventilation ( ). Inspiratory capacity, operating lung volumes and arterial blood gases were measured during submaximal exercise. Arterial blood gases were used to derive VD /VT , alveolar volume, dead space volume, alveolar ventilation and dead space ventilation. During submaximal exercise, HFrEF patients had greater slope and VD /VT than CTL subjects (P= 0.01). At similar , HFrEF patients had smaller tidal volumes and alveolar volumes (HFrEF 1.11± 0.33litres versus CTL 1.66± 0.37 litres; both P≤ 0.01), whereas dead space volume was not different (P= 0.47). The augmented breathing frequency in HFrEF patients resulted in greater dead space ventilation compared with CTL subjects (HFrEF 15± 4lmin-1 versus CTL 10± 5lmin-1 ; P= 0.048). The HFrEF patients exhibited greater increases in expiratory reserve volume and lower inspiratory capacity (as a percentage of predicted) than CTL subjects (both P< 0.05), which were significantly related to VD /VT and alveolar volume in HFrEF patients (all P< 0.03). In HFrEF, the reduced tidal volume and alveolar volume elevate physiological dead space during submaximal exercise, which is worsened in those with the greatest ventilatory constraints. These findings highlight the negative consequences of ventilatory constraints on physiological dead space during submaximal exercise in HFrEF.

STANDARD adult circle systems (with adult ventilator bellows and carbon dioxide absorber) equipped with pediatric circuit hoses, rather than semiclosed partial rebreathing systems or specially designed pediatric circle systems (with small ventilator bellows and carbon dioxide canister), often are used to ventilate infants and children during administration of anesthesia. 1–3If adult circle ventilator systems are used, it is important to understand the possible limitations and to make necessary modifications of adult techniques for use with infants. One limitation of this equipment for infant ventilation is the difficulty in determining how to set the tidal volume (TV) if using a time-cycled, volume-limited mode of ventilation to ensure that the patient receives the desired TV. The large compression volume of the circle system relative to the infant’s lung volume, 4leaks around uncuffed endotracheal tubes, effects of fresh gas flow (FGF) on delivered TV, and the mechanical difficulty of setting a small TV using an adult bellows assembly contribute to a discrepancy between set and delivered TVs.Recently, the Food and Drug Administration approved the use of a circle anesthesia system equipped with an electronic piston ventilator. This system is designed to accurately deliver small TVs by providing easy-to-use ventilator modes and controls, automatic system compliance compensation, and eliminating the interaction between FGF and TV. As a result, time-cycled, volume-limited ventilation of infants with a piston-driven ventilator may be more “user-friendly” and reliable compared with ventilation with most bellows-equipped ventilators. The purpose of our study was to compare the performance of the Drager Narkomed GS ventilator system (North American Drager, Telford, PA), equipped with a traditional ascending bellows ventilator, with the new Drager Narkomed 6000 ventilator system, which uses a circle anesthesia circuit, in ventilating an infant test lung model.A Drager Narkomed GS circle anesthesia system, equipped with a standard adult bellows and carbon dioxide absorber, and a Drager Narkomed 6000, which incorporates a Divan piston-driven ventilator (Dragerwork, AG, Lubeck, Germany), were compared regarding delivery of minute ventilation (V̇E) to an infant test lung. Both ventilator systems were equipped with a disposable pediatric circle circuit (Pediatric King; King Systems Corporation, Noblesville, IN). The test lung model used in this study has been described previously. 5–8V̇Ewas measured using a test lung (Bio-Tek Ventilator Tester; Bio-Tek Instruments, Winooski, VT). The Bio-Tek test lung includes two wire wool–filled metal canisters that simulate lungs that have either normal compliance (0.003 l/cm H2O) or low compliance (0.001 l/cm H2O), as defined by the American National Standards Institute. 9The test lung determines delivered V̇Eby measuring peak inspiratory pressure (PIP) and multiplying by the calculated lung compliance to determine the TV. V̇Eis calculated by multiplying the TV by the respiratory rate (RR). The accuracy of the test lung for V̇Eis ± 4% for TVs of 5–300 ml in the infant mode. The test lung was connected to the ventilator systems by a 3.5-mm endotracheal tube (Mallinckrodt Medical, St. Louis, MO) cut distally (removing the Murphy eye to prevent system leakage), with a 15-mm connector on each end (fig. 1). The test lung was set for ambient barometric pressure, temperature, and humidity before all testing. At the recommendation of the test lung manufacturer, the least restrictive adapter (parabolic restrictor Rp20) connected the endotracheal tube to the test lung. The study was divided into three parts (fig. 1). During all three parts of the study, an inspiratory-to-expiratory ratio of 1:2 was maintained. No significant leaks in either of the ventilator systems were detected before testing. For each condition tested, V̇Ewas measured three times. The average of these three V̇Emeasurements was used for subsequent data analysis. The V̇Edelivered by the ventilator systems to the test lung were compared during time-cycled, pressure-limited and time-cycled, volume-limited ventilation. To simulate a variety of pediatric conditions, V̇Ewas measured with both ventilator systems using a variety of RRs (20, 30, 40, and 50 breaths/min) and with the test lung set in both normal- and low-compliance infant modes. During time-cycled, pressure-limited trials the PIP was adjusted to 20, 30, 40, and 50 cm H2O. Using the GS ventilator system, the desired PIP was achieved by adjusting the drive gas flow to the bellows to medium, adjusting the bellows upward to the maximal setting, and adjusting the inspiratory pressure limit (“pop-off valve”) to the target PIP. With the 6000 ventilator system, the desired PIP was achieved by direct entry of the desired PIP value into the operation control panel. During time-cycled, volume-limited trials, TVs of 30, 40, 50, 100, 200, and 300 ml were set. With the GS ventilator system this was achieved by visual adjustment of the bellows combined with adjustment of the drive gas flow to the bellows from its initial medium setting upward or downward (if needed) until the desired TV was indicated by the machine’s spirometer. In the 6000 ventilator system, TV was adjusted by direct entry of the desired value into the operation control panel. An FGF of 3 l O2/min was used during all part A testing.Minute ventilation was measured before and after an acute decrease in test lung compliance during time-cycled, volume-limited ventilation. V̇Ewas measured starting with an RR of 20 breaths/min and a TV of 50, 100, or 200 ml, with the test lung set to mimic normal-compliance infant lungs (0.003 l/cm H2O). Then, without changing any ventilator settings, V̇Ewas measured after the test lung compliance was decreased to 0.001 l/cm H2O. An FGF of 3 l O2/min was used during all part B tests. PIP limits of 80 cm H2O for the 6000 ventilator system and maximum for the GS ventilator system were set before all testing.The test lung was set in the normal-compliance infant mode (0.003 l/cm H2O) for all testing. To test the effect of an increasing FGF, baseline V̇Emeasurements were made starting at an FGF of 1 l O2/min and a set TV of 50, 100, or 200 ml, with an RR of 20 breaths/min. Without changing ventilator settings, V̇Ewas measured after incremental increases of FGF to 3, 6, and 10 l/min. To test the effect of decreasing FGF, we reversed the procedure, obtaining baseline V̇Emeasurements starting with an FGF of 10 l O2/min, an RR of 20 breaths/min, and a set TV of 50, 100, or 200 ml. Without changing ventilator settings, the FGF was adjusted incrementally downward to 6, 3, and 1 l/min, and V̇Ewas again measured.The multiple regression technique was used to analyze the data for part A: The dependent variable was V̇E; independent variables were the ventilator systems used (GS, 6000), lung compliance, RR, and PIP (pressure-limited data) or TV (volume-limited data). For parts B and C, the repeated-measures analysis-of-variance technique was used to analyze the data. The dependent variable was V̇E; independent variables were the ventilator systems, TV, and lung compliance (part B) or FGF (part C).During time-cycled, pressure-limited ventilation both the GS and the 6000 ventilator systems generated nearly identical V̇Eover the entire range of PIPs studied in both the compliant and noncompliant infant lung models (P = 0.77 and P = 0.33, respectively;fig. 2). During time-cycled, volume-limited trials, the 6000 ventilator system could be set at all TVs; we were not able to set the GS ventilator system to achieve some higher TVs (especially at high RRs in the low-compliance lung model) or any TV less then 50 ml. Thus, no comparison data points were obtained for those TVs. Only TVs of 50, 100, and 200 ml were compared between the two ventilator systems. In the normal-compliance lung model, the 6000 ventilator system delivered slightly higher V̇Ethan the GS ventilator system, but this difference was not statistically significant (P = 0.18). In the low-compliance lung model, the 6000 ventilator system delivered greater V̇Ethan the GS ventilator system (an average increase in V̇Eof 18%;P = 0.024;fig. 3). As lung compliance was decreased from normal to low, both ventilator systems delivered less V̇E(41–58% less) to the test lung at all TVs studied during time-cycled, volume-limited ventilation (P < 0.001) (fig. 4). The 6000 ventilator system was better able to compensate for decreased lung compliance than the GS ventilator system (P < 0.001). The GS ventilator system delivered progressively more V̇Eto the test lung as FGF was increased from 1 to 10 l/min at all TVs studied (P < 0.001 for FGF, P < 0.001 forTV). The GS ventilator system delivered progressively less V̇Eto the test lung as FGF was decreased from 10 to 1 l/min at all TVs studied (P < 0.001 for FGF, P < 0.001 for TV) (fig. 5A). The 6000 ventilator system maintained nearly identical V̇Eas FGF was increased from 1 to 10 l/min at all three TVs studied (P = 0.14 for FGF, P < 0.001 for TV) or as FGF was decreased from 10 to 1 l/min at all three TVs studied (P = 0.07 for FGF, P < 0.001 for TV) (fig. 5B). We observed nearly identical performance of the GS and 6000 ventilator systems during time-cycled, pressure-limited ventilation over a wide range of RRs and PIPs and two test lung compliances. During time-cycled, volume-limited ventilation trials, at TVs that could be obtained in both systems being tested, the 6000 ventilator system delivered slightly greater V̇Ethan the GS ventilator system (an average increase of 18% as measured by the Bio-Tek test lung). This difference between systems may reflect in part how the TV was set on the GS ventilator as well as differences between the spirometers of the two systems. Smaller TVs (50 or 100 ml) were extremely difficult to set accurately with the GS ventilator system, requiring visual setting of the bellows (with 50-ml TV, the bellows was fully “seated” at the bottom of the bellows assembly) followed by adjustment of the driving gas flow to the bellows. TVs below 50 ml were not obtainable at all with the GS ventilator system (the machine’s digital spirometer display of TV does not register such low TVs). Substantial decreases in V̇Ewere observed with both ventilator systems as compliance decreased, at all TVs studied. The 6000 ventilator system was marginally better able to maintain V̇Ewith decreasing lung compliance. Changing FGF during time-cycled, volume-limited ventilation does not effect V̇Ewith the 6000 ventilator system but does influence V̇Ewith the GS system.Based on our in vitro findings, what can we conclude about possible advantages or disadvantages of a piston-driven ventilator system compared with a traditional bellows ventilator system during infant ventilation? Pediatric anesthesiologists often use the ventilator of an adult circle system in the time-cycled, pressure-limited ventilation mode during infant anesthesia because of their familiarity with this type of ventilation and the mechanical ease of setting a target PIP limit with most ventilators. Our study indicates that during time-cycled, pressure-limited ventilation, it does not make a meaningful difference if one uses either a bellows-equipped ventilator or a piston-driven ventilator; the systems deliver equivalent volumes to the infant test lung over a wide range of RRs and PIPs. The near equivalence of the V̇Eproduced by the GS and 6000 ventilator systems during time-cycled, pressure-limited ventilation reinforces observations made in prior publications using this same lung model. 5–8During time-cycled, pressure-limited ventilation, V̇Eis dependent on lung compliance, PIP achieved, and RR, regardless of the compliance of the circuit used, the precise method of achieving a given PIP with the circle system, and the ventilator system used to achieve a given PIP (Mapleson system, circle system, or free-standing ventilator). 5–8One potential disadvantage of time-cycled, pressure-limited ventilation is that if for whatever reason there were a sudden change in lung compliance or resistance in the system, such as a surgeon leaning on the chest of an infant or a kinked endotracheal tube, there would be no change in the PIP, and there would be a decrease in delivered ventilation. Other means of monitoring respirations, however, such as auscultation of breath sounds and measured expired carbon dioxide (including the configuration of the wave form), likely would diagnose such problems. In the above scenario, with time-cycled, volume-limited ventilation there would be a sudden rise in peak inflation pressure as well as changes in breath sounds and the carbon dioxide waveform, but TV might be maintained better. Our study did not assess these clinically relevant means of assessing ventilation but rather just examined the performance of each ventilator system under extreme changes in lung compliance.The GS and 6000 ventilator systems do not perform equally during time-cycled, volume-limited ventilation. The GS ventilator system cannot be used easily as a true volume-limited ventilator. Determination of an appropriate set TV based on patient weight is cumbersome. Badgwell et al. 4have described the nonlinear relationship between patient weight and set TV required for time-cycled, volume-limited ventilation in infants using adult ventilator systems (150–200 ml/kg in a 1-kg infant vs. 25 ml/kg in infants more than 10 kg). Once calculated, the smaller TVs required for infants may be mechanically difficult to set because of the lack of precision of the adult bellows assembly. The lowest TV easily set during our study trials was 200 ml, which is the first mark on the bellows assembly. Thus, those who choose to use adult circle systems for infant ventilation often adjust the volume limit of the bellows slowly upward until the desired chest expansion or target PIP is achieved. An initial PIP of approximately 20 cm H2O usually is chosen, and further adjustment of the TV upward or downward is based on chest expansion, end tidal carbon dioxide concentration, and oxygen saturation. This type of time-cycled, volume-limited ventilation might be more accurately described as time-cycled, volume-limited, pressure-guided ventilation.The 6000 ventilator system, in contrast to the GS system, is easily set for all TVs during time-cycled, volume-limited ventilation, including TVs less than 50 ml. The ability to set accurate infant TVs, along with the consistency in V̇Ewith changing FGF, would appear to be an advantage of the 6000 ventilator system for use with infants. Peters et al. 10reported successful time-cycled, volume-limited ventilation of 20 infants between 2 and 6 kg with TVs of 10 ml/kg at RRs between 25 and 40 breaths/min, using a piston-driven ventilator system. Such low TVs are not easily obtainable using the GS ventilator system. It is possible to set the GS ventilator to very low TVs by setting the bellows at the very lowest limit and then adjusting the drive gas flow to the bellows so that the desired TV is achieved; however, these low TV settings are below the technical limits of the GS spirometer and would necessitate assessment of efficacy purely on a clinical basis. Also, with any change in FGF, set TV would need to be readjusted (i.e. , if FGF were increased, set TV would need to be decreased to maintain constant V̇E; if FGF were decreased, set TV would need to be increased to maintain constant V̇E). The consistency in V̇Eover a wide range of FGFs we observed using the 6000 ventilator system is similar to results reported by Schirmer et al. 11in a test lung study using a piston-driven ventilator: TVs ranging from 20–200 ml were delivered reliably during time-cycled, volume-limited ventilation over a range of FGFs (1–6 l/min). When using the 6000 ventilator system, once an appropriate set TV has been achieved, FGF can be adjusted over a wide range without risk of inadequate ventilation (with a decrease in FGF) or overventilation or barotrauma (with an increase in FGF).Another design difference between the GS and 6000 ventilator systems as used for time-cycled, volume-limited ventilation is the response to a changing lung compliance. Our results show that the 6000 ventilator system has some ability to compensate for compliance changes; the GS ventilator system does not. Because the compensation for a decrease in lung compliance by the 6000 ventilator system is incomplete, the clinical advantage of this compliance feature needs further evaluation; a set TV may require significant upward adjustment to maintain V̇Eas lung compliance decreases, regardless of the ventilator system used. The conclusion we have reached based on our in vitro study is consistent with results reported by Schirmer et al. 11using a piston-equipped ventilator in an animal model. In that study, using newborn piglets, decreased lung compliance was induced by creation of a tension pneumothorax. Although the piston-equipped ventilator had an improved ability to maintain constant ventilation after induction of the pneumothorax, it was not able to maintain normal ventilation in several piglets. 12The actual lung compliance of the piglets may have been different from the compliance settings we studied, so that direct comparison of results is not possible. Further studies would be required to examine the clinical importance of the compensation provided by the 6000 ventilator because the acute changes in compliance that we studied may have been more extreme than those that would be observed in most clinical settings.Our study indicates that the piston-equipped Drager Narkomed 6000 ventilator system performs in a manner similar to the traditional ascending bellows-equipped Drager Narkomed GS ventilator system during time-cycled, pressure-limited ventilation. During time-cycled, volume-limited ventilation, however, the 6000 ventilator system can be set easily to achieve small TVs, but the GS ventilator system cannot. More importantly, the 6000 ventilator system allows maintenance of a constant TV during a wide range of FGFs, without further adjustment. Regardless of the ventilator system used, or the type of ventilation that is chosen (time-cycled, pressure limited vs. time-cycled, volume-limited), significant adjustment of ventilator parameters is required to maintain ventilation if lung compliance changes.It is important to emphasize that this study evaluated only the ability of each ventilator system to deliver V̇Eto a test lung; our study did not address other potential issues related to clinical use of these ventilator systems. Further studies are warranted to evaluate the clinical role of the 6000 ventilator system during infant anesthesia compared with other available systems.

Background: Heliox is a low density gas that enhances aerosol drug delivery, increases laminar flow, and improves carbon dioxide diffusion. High frequency jet ventilation (HFJV) is a time-cycled pressure limited ventilator that delivers less than dead space tidal volumes with transitional gas flows at high celerity. The purpose of this study was to evaluate whether heliox at any admixture affected MAP and Servo pressure on the HFJV. Methods: The QuickLung Breather (IngMar Medical, Pittsburgh, PA) test lung model was used to simulate spontaneous respirations while connected to a LifePulse HFJV (Salt Lake City, UT) and a Servo-U conventional ventilator (CV) (Getinge, Wayne, NJ). The test lung was attached to a 2.5 mm ETT set to a resistance of 50 cm H20 /L/S, frequency 42 breaths/min, 25% inspiratory time (0.36 s), with measured exhaled tidal volumes of 9-10 mL while using CPAP +8 on CV. A LifePort ETT connected the HFJV and CV circuit to the ETT and lung model. A jet rate of 300 breaths/min and I-time 0.02 s, and CPAP +8 (CV) were constant throughout the study. HFJV was powered by wall sourced, 50 PSI oxygen and then added a heliox blender (Precision Medical (Northampton, PA.) for the test separately. Peak inspiratory pressure (PIP) on HFJV was increased to 20, 30, and 40 cm H2O on room air, and 80/20, heliox gas admixture upon achieving a 5-min steady state for changes in PIP and heliox concentration exhaled tidal volumes (CV) and Servo pressure, MAP (HFJV) was recorded. Mean and standard deviations of combined HFJV + CV on room air vs 80/20 heliox powered to the CV and the HFJV were calculated and compared using a t-test with a statistical significance level set at P < .05. Results: There was a statistically significant difference in MAP (P = .035) with changes in heliox compared to room air. In the presence of heliox, neither Servo pressure nor MAP were different. Conclusions: Tidal volume increases as a heliox gas admixture is added to HFJV. Future applications and testing with different Heliox concentrations needs to be studied further.Measured ParameterServo U and HFJV No HelioxHeOx powered by HFJV or CVP valueServo Pressure2.9 (1.05)2.75(0.953)0.8352MAP10.2 (0.95000)10.13(0.1697)0.0700Vt exh48 (2.645)60 (12.5)0.0297

Carbon dioxide clearance during apnoea.

Needle cricothyroidotomy

Can We Finally Take the "VE" Out of THRIVE?

We can't solve problems by using the same kind of thinking we used when we created them –Albert Einstein A defining moment for intensive care occurred during the polio epidemics of the 1950s. The preferred mode of artificial respiratory support was negative pressure tank respirators, often described as 'iron lungs'. The mortality from bulbar or bulbospinal polio with this mode of ventilation was almost 90% in the Blegsdam hospital in Copenhagen. As a result, Hans Lassen, an epidemiologist and senior figure, consulted the freelance anaesthetist Bjørn Ibsen, who suggested tracheotomy and intermittent positive pressure ventilation (IPPV) as a solution. Subsequent implementation of this solution, famously using medical and dental student volunteers to ventilate the lungs of these patients manually, reduced the mortality to 25%. This was a paradigm shift not only in terms of artificial ventilation, but it also acted as the catalyst for the development of intensive care. Since that time, IPPV has remained the major foundation of intensive care, albeit with refinements over the years [1–3]. Many of these refinements, particularly those that have been developed by manufacturers of ventilators, have never been adequately evaluated as would now be the case for the licensing or introduction of new drug therapies. Most of these 'advances', while possibly increasing patient comfort, have had little verifiable effect on truly important outcomes such as mortality. For example, a recent study failed to identify any important benefits of airway pressure release ventilation (APRV) or biphasic positive airway pressure (BIPAP) over the more basic approach of assist-control ventilation commonly utilised in North America [4]. Soni and Williams have described positive pressure ventilation as a 'radical departure from normal physiology [5]. They and others have well described IPPV's consequences for pulmonary, cardiovascular and remote organ function [3, 5–8]. High levels of positive end-expiratory pressure (PEEP) have the potential to exacerbate facets of this physiological trespass. Many refinements or adjuncts to positive pressure ventilation lack sufficient evidence of benefit and are based on contrary, equivocal or preliminary evidence with regard to improved patient outcomes. Some have been insufficiently investigated in the clinical arena to allow meaningful quantification of purported outcome benefits, such as tracheal gas insufflation, permissive hypercapnia, inverse ratio ventilation or inhaled prostacyclin. Some improve physiological surrogates such as oxygenation but appear to have no positive impact on outcomes, a notable example being inhaled nitric oxide [9]. Others can claim an equivocal or hypothesis generating level of evidence such as high-level PEEP, recruitment-manoeuvres, high frequency oscillatory ventilation, prone ventilation and corticosteroid regimens [10–15]. Presently all these treatments remain unproven 'rescue' therapies. Nevertheless, they continue to be employed clinically on the basis of their beneficial effects on physiological surrogates such as Pao2 or for reasons other than important patient-centred outcomes [16]. A new way of achieving gas exchange, using extracorporeal membrane oxygenation (ECMO), had its first success almost 20 years after Ibsen's own radical move from non-invasive negative pressure ventilation to invasive positive pressure ventilation via a tracheotomy [17]. However, Donald Hill's isolated success did not immediately translate into something meaningful for patients when ECMO was subject to a randomised controlled trial comprising 90 patients, published 7 years later in 1979 [18]. The mortality for patients assigned to both conventional ventilation and ECMO groups was over 90%. The concept of lung protective ventilation had not been conceived at that time in the mid 1970s, the technology and experience were limited and the procedure more invasive, so the results are hardly surprising in retrospect. Fifteen years later another small (40 patients) trial of partial ECMO support, primarily to facilitate removal of carbon dioxide, also failed to demonstrate benefit [19]. At that time, the authors' conclusion that 'Extracorporeal support for ARDS should be restricted to controlled clinical trials' seemed quite reasonable. Since those early attempts to utilise ECMO for respiratory support, new concepts of and strategies for protective lung ventilation to diminish ventilator-induced lung injury have been established, and the concept of 'lung resting' during ECMO developed. Technological advances have simplified treatment and reduced adverse sequelae, improving the risk-benefit profile. Clinical proof of effectiveness was subsequently established in neonates in the 1990s [20, 21], but demonstrable benefit for adults had to await publication of the CESAR (Conventional ventilatory support versus Extracorporeal membrane oxygenation for Severe Adult Respiratory failure) trial [22]. This trial was analysed on an intention-to-treat principle and also adhered to a prespecified published protocol [23]. The primary outcome was reduced mortality or severe disability at 6 months. The point estimate for absolute risk reduction of this primary outcome in patients transferred to a centre with the option of ECMO was 16% (p < 0.03) and for death it was 12% (p < 0.07). These risk reductions exceed that obtained by activated protein C in the PROWESS study by a factor of two or more. The cost per quality adjusted life years (QALY) was approximately £20 000, well within the threshold adopted by the National Institute for Health and Clinical Excellence (NICE) for cost-effective therapies that are affordable within the NHS [24]. It is also noteworthy that there have been additional advances in ECMO technology that have simplified delivery and probably increased the safety of this therapy since completion of the trial. The re-evaluation of ECMO by clinicians and government has been accelerated by the recent H1N1 influenza pandemic, where ECMO was perceived as life saving – particularly by the media and the public. The result in Scotland was the formation of a multidisciplinary expert group, tasked by Government to review the role of ECMO. After conducting interviews, visiting ECMO centres and reviewing a wide range of evidence this group concluded 'The Expert Group believes that there is sufficient evidence to support the efficacy, safety and clinical effectiveness of ECMO for adult patients with potentially reversible advanced respiratory failure' [25]. Nevertheless, many clinicians and some authorities remain sceptical of the need to develop ECMO [26], although others are more open minded [27]. Given the current evidence base, imperfect though it is, we believe that ECMO can presently be considered as one treatment modality in severe ARDS that should be contemplated in concert with other adjuncts in those with severe but potentially remediable disease. This view is supported by others [16, 28], although its precise role may be shaped by the results of new trials of more conventional support such as high frequency oscillation [12]. Extracorporeal membrane oxygenation, as provided in the CESAR trial, provided full support to maximise oxygenation and carbon dioxide elimination. Extracorporeal carbon dioxide (ECCO2) elimination can be accomplished with lesser support, and use of such devices has been described and are available commercially [29–31]. It has been postulated that partial support can also allow less aggressive and injurious mechanical ventilation, reduce lung bio-trauma and hence improve outcomes [31]. Although such approaches have intuitive appeal, they have not yet been subject to scrutiny in large clinical trials, and benefits remain to be more precisely quantified. NICE has issued specific guidance with regard to arteriovenous extracorporeal membrane carbon dioxide removal (AVECCO2R), being concerned with the incidence of vascular complications in some case series [32]. What of the future? The technology has improved since the CESAR trial, and ECMO is easier to provide – although specialist expertise is still required. Equipment will eventually become more refined and competition from increased volume sales will reduce costs, making ECMO more affordable and cost-effective. In the intervening years, use of extracorporeal techniques in general intensive care have become commonplace. With the development of haemofiltration, the most common variant being continuous venovenous haemofiltration (CVVH), there has been a transition from heroic fluid resuscitation followed by high-dose furosemide, before resorting to extracorporeal renal replacement therapy (RRT), to a greater early reliance on RRT. Thus, for renal support, there is currently less emphasis on squeezing the remaining renal function from injured tissues with fluids and diuretics, and more emphasis on appropriately timed extracorporeal support. This transition occurred ahead of good evidence that this was the preferred option. There are parallels with respiratory ECMO. It is now, most commonly, a venovenous extracorporeal method that can be accomplished by a single dual-lumen catheter using similar anticoagulation strategies to those used for CVVH. An intrinsic and possibly insurmountable problem with most mechanical ventilation strategies is how to keep diseased lung units open (and that may be injurious too), without over-distending and injuring more normal areas of lung, as well as preventing the unwanted physiological effects on other organs as described by Soni and others. Could an extracorporeal strategy be helpful if applied earlier in acute respiratory distress syndrome (ARDS) rather than just as a late rescue therapy, as well as in a wider range of respiratory diseases? Extracorporeal membrane oxygenation has already been used in conditions other than ARDS, such as acute severe asthma or chronic obstructive airways disease. Partial support may be adequate for such patients and obviate the need for invasive ventilation (Dr Axel Nierhaus, Hamburg – personal communication). Other described indications include support as a bridge to lung transplantation, postcardiac arrest support, and support for potential organ donors. Ambulatory ECMO as a bridge to lung transplantation, avoiding invasive ventilation, is already a strategy that has been used in the UK. In the medium term it is likely that a national centre with satellites will develop to provide adequate coverage for patients with severe life-threatening gas exchange abnormalities. This requires adequate retrieval facilities to transport patients who are close to physiological decompensation. However, in the more distant future, as costs decrease and safety increases through further technological advances, training and experience, might 'oxygen or carbon dioxide dialysis' become a primary mode to regulate gas exchange in patients with severe reversible lung disease? For the last 50 years, IPPV has been the workhorse of intensive care and its judicious use has saved countless lives. However, it has its limitations, as has ECMO. The challenge will be to integrate these different technologies to get the best results for patients. We need to be open-minded to explore and define the place of ECMO in the future, as well as acknowledge that it has a place in current practice. No external funding and no competing interests declared. doi:10.1111/j.1365-2044.2010.06507.x

Carbon dioxide elimination during high-frequency jet ventilation

ESTIMATION OF ARTERIAL PCO2 DURING HIGH FREQUENCY JET VENTILATION: Studies in the Dog