EXCESSIVE WORK OF BREATHING DURING INTERMITTENT MANDATORY VENTILATION

Br. J. Anaesth. (1986), 58, 1048-1054 EXCESSIVE WORK OF BREATHING DURING INTERMITTENT MANDATORY VENTILATION J. S. MECKLENBURGH, I. P. LATTO, T. A. A. AL-OBAIDI, E. A. SWAI AND W. W. MAPLESON Ventilators with facilities for intermittent mandatory ventilation (IMV) are intended to assist the weaning of patients from controlled ventilation (Downs et al., 1973; Margand and Chodoff, 1975; Browne, 1984). In a perfect IMV ventilator, breathing spontaneously through the breathing system would be indistinguishable from breathing directly from atmosphere apart from the different gas mixture, and assuming that no positive end-expiratory pressure (PEEP) or continuous positive airway pressure (CPAP) has been set. A less than perfect IMV ventilator may hinder spontaneous breathing by imposing additional resistance, or it may assist it if the supply mechanism is such that mouth pressure is more positive during inspiration than during expiration. In the former, weaning may be delayed unnecessarily; in the latter the anaesthetist may be misled into disconnecting the patient from the ventilator prematurely. The present report analyses an instance in which the resistance in the breathing system resulted in difficulty in weaning. The investigation was prompted by our experience with a 72-year-old female patient who had had a mitral valve replacement. Attempts to wean her from the ventilator over a period of 4 days were unsuccessful. When she was allowed to breathe spontaneously, or when her lungs were ventilated at a low IMV rate, an increase in arterial carbon dioxide tension and a decrease in oxygen tension occurred, and the patient quickly became distressed. Eventually an alternative IMV J. S. MECKLENBURGH, B.SC., MJC.J I. P. LATTO, M.B., B.S., F.F.A.R.C.S.J E. A. SWAI, M.B., CH.B, F.FJV.R.C.S.; W. W. MAPLESON, D.SC, F.INST.P.; Department of Anaesthetics, University of Wales College of Medicine, Heath Park, Cardiff CF44XW.T. A. A. AL-OBAmi,B^c.,PHJ).,M.rNST.p.,College of Medicine, Al-Mustansiriyah University, Baghdad (presently Visiting Colleague, University of Wales College of Medicine). SUMMARY Difficulties were experienced in weaning a patient from a ventilator by means of intermittent mandatory ventilation (IMV). The difficulty was overcome by installing an alternative IMV system (Hudson "disposable IMV valve") through which the patient drew her spontaneous breaths. Laboratory measurements showed that the resistance of the ventilator breathing system was much higher than that of the alternative system, mainly as a result of the resistance of the humidifier. It was calculated from measurements with a preset pattern of simulated breathing that the extra, external, work of breathing through the ventilator breathing system was approximately 1.5 times the normal internal mechanical work of breathing for a normal patient; with the alternative system, the extra work was only 0.5 times the normal. It is stressed that the breathing systems of IMV ventilators should be judged with the same rigour as other systems through which the patient is required to breathe spontaneously. It is recommended that manufacturers should pro vide the necessary information. system (Hudson "disposable IMV valve") was connected into the breathing system so that, in spontaneous respiration, gas was drawn from this alternative system instead of through the breathing system of the ventilator. With this arrangement the patient did not become distressed, and the carbon dioxide and oxygen tensions were acceptable. Weaning from mechanical ventilation was accomplished and the trachea extubated. Since it was suspected that the original difficulty in weaning was caused by a high resistance in the ventilator breathing system, a

IMV: WORK OF BREATHING 1049 Alternative IMV system Gas inlet Reservoir tub* One-way valve Pneumotachograph / Pressure relief valve Ventilator Humidifier Filter Flow Modified Starling pump output Pressure tapping FIG. 1. Diagram showing ventilator breathing system, alternative IMV system and method of imitating spontaneous breathing with a modified Starling pump. The alternative IMV system was connected into the ventilator breathing system at the point indicated, when required. laboratory study was undertaken to measure the resistance of both breathing systems. In addition, the work of breathing through the two systems was estimated, because this is a more relevant indicator of the burden imposed on the patient. For instance, if a patient is made to breathe through an external resistance equal to his own respiratory system resistance, this doubles the total resistance. However, during inspiration, the patient has to do work not only on the total resistance, but also on his own compliance. Therefore, a more realistic representation of the burden placed on the patient's respiratory muscles is the percentage increase in the total work of breathing. Incidentally, it should be noted that, although it is customary to speak loosely of the "work of breathing", strictly, the terms should be the "work of one breath" (or one inspiration) or the "mean power of breathing" (the mean rate of working averaged over one or more complete breaths). The "work per litre of total ventilation" has also been used (Mcllroy, Marshall and Christie, 1954). MATERIALS AND METHODS The laboratory experiments were performed on the same ventilator and breathing system as had been used with the patient. The ventilator was a Kontron 3100 with Varicontrol, a Kontron Pearl 3150 humidifier was included in the inspiratory line, and a new dry PALL bacterial filter was attached to the patient connection of the Y-piece (ng. 1). Spontaneous ventilation was simulated using a modified Starling pump, set to deliver a tidal volume of 500 ml at a frequency of 18.75 b.p.m. and with an I:E ratio of 1:1.7. This produced an approximately sine-wave flow pattern in each half-cycle of respiration. Mandatory breaths delivered by the ventilator were allowed to escape to atmosphere through a pressure relief valve, set to open at approximately 7 kpa, since the Starling pump could not imitate lung compliance. The humidifier was dry and unheated during the test, so that the difficulties of measuring the flow of humidified and heated gas were avoided. Careful consideration of the design of the humidifier, the alternative IMV valve and the expiratory valve of the ventilator indicated that the use of warm humidified air would not alter the performance of these components to any great extent. The filter, however, would show an increase in resistance with time as water condensed within it and the use of warm air in this study would, if anything, diminish the difference between the Kontron and alternative systems. When the alternative IMV system (Hudson disposable IMV valve) was connected into the breathing system (fig. 1), inspired gas was drawn from this system instead

1050 of from the ventilator, but expiration was still via the normal expiratory pathway of the ventilator. Although oxygen-enriched, humidified air had been supplied to the alternative IMV system in clinical use, room air was used for the laboratory study. Pressure and flow were measured at the "mouth" using a strain gauge transducer for pressure, and a screen pneumotachograph connected to a Greer differential pressure transducer (Mercury Electronics) for flow, and displayed on a chart recorder (Lectromed). The transducers were calibrated before, and the calibration verified after, the recording using a "series calibrated" flow meter (Fisher Controls) and a liquid-filled manometer. The pressure-flow characteristics of the breathing systems were determined using a range of steady inspiratory and expiratory flows. Pressureflow characteristics were also obtained "dynamically", when spontaneous ventilation was imitated with the Starling pump, by digitizing the flow and pressure signals at 0.04-s intervals and plotting one against the other. When pressure (P) and flow rate (P) are monitored at the mouth the product, P. V, is the instantaneous power ifr at any moment during the respiratory cycle. This was computed at 0.04-s intervals over a complete breath. The work done in any defined time interval r 0 to t 1 is then given by: W Jt-u, Wdt The period over which work is calculated can be the whole or part of the respiratory cycle. Work calculated in this way (using P as the difference between "mouth" and atmospheric pressure) is the work done by the patient on the attached breathing system during spontaneous breathing. This method of determining work done is the one used by Engstrom and Norlander (1962); it avoids the difficulties frequently encountered in interpreting pressure-volume loops and also provides information on the waveform of power. Computation was performed by a BBC microcomputer running BBC Basic and graphical output was to a Hewlett-Packard 7470A graph plotter. In order to determine the sites of power dissipation, experiments were repeated following the removal of individual components of the breathing system. BRITISH OF JOURNAL ANAESTHESIA 8-0.4 - -0.8 - -1 -O.5 Flow (litre s- 1 ) FIG. 2. Pressure-flow characteristics using steady flows for the different breathing system configurations. The curve in the upper right quadrant is for the expiratory limb which was common to all systems. The lower left quadrant shows the characteristics of the inspiratory limb of: A = the complete IMV system (Kontron ventilator, humidifier and filter); B = as in A, but with the humidifier removed; C as in B, but with the bacterial filter removed; D «the alternative IMV system; E = the recommended upper limit of breathing system resistance (Nunn, 1977). RESULTS Pressure-flow characteristics for various configurations of the breathing systems, using steady flows, are shown in figure 2. The slope of each characteristic indicates the resistance. The inspiratory limb of the complete ventilator breathing system (ventilator, humidifier and filter) exhibited the highest resistance (steepest curve, curve A): at a flow of 0.5 litre s" 1 the pressure at the patient connection was 0.6 kpa, that is, a resistance of 1.2 kpa litre" 1 s. The total resistance at 0.5 litre s" 1 was reduced to 0.6 kpa litre" 1 s on the removal of the humidifier (curve B), and to 0.5 kpa litre" 1 s on removal of the filter (curve C). The alternative IMV system had an inspiratory resistance of only 0.4 kpa litre" 1 s (curve D) at 0.5 litre s" 1. The upper limit for resistance recommended by Nunn (1977) is shown for comparison (curve E).

IMV: WORK OF BREATHING 1051 1-1 FIG. 3. Flow, pressure, power and accumulated work during a simulated spontaneous breath (500 ml, 18.75 b.p.m.) for the different breathing system configurations. A = Complete IMV system; B = as in A, but without humidifier; C as in B, but without filter; D alternative IMV system. TABLE I. Work done (mj) on tht two IMV systems during a simulated spontaneous breath of 500 ml tidal volume at 18.75 b.p.m. Kontron breathing system Kontron with the alternative IMV breathing system Inspiratory phase 272 90 Expiratory phase 90 88 Total 362 178 The expiratory pathway was the same for all systems and, therefore, a single pressure-flow curve was obtained. The resistance at aflowof 0.5 litre s" 1 was 0.6 kpa litre" 1 s. Digitized values of pressure and flow obtained during simulated spontaneous breathing gave essentially the same pressure-flow characteristics as from steady flows. The flow, pressure, power and accumulated work during a simulated spontaneous breath are displayed against time for the different breathing system configurations in figure 3. The flow waveforms generated by the Starling pump were similar for all four configurations (fig. 3 a). Pressure waveforms, on the other hand, were different (fig. 3 b) and, therefore, so were the waveforms of power (fig. 3 c) and of accumulated work (fig. 3d). (Accumulated work is the integral of power from the start of inspiratory flow to the current time.) The peak inspiratory power required during a simulated breath was reduced from 0.64 W with the ventilator IMV system, to 0.19 W with the alternative IMV system. Table I gives the calculated work for the inspiratory and expiratory phases for the two main breathing systems. With the alternative IMV system, calculated work in the inspiratory phase was reduced to one-third of that with the original system; expiratory work remained almost the same and total work done (over a whole cycle) was approximately halved. DISCUSSION IMV has potential advantages when weaning a patient from controlled ventilation, but problems may occur with patients whose lung mechanics are

1052 0.8 0.6 0.6 0.5 0.4 0.3 0.2 0.1 c+m-a C+R+D 0 1 2 3 4 Time (s) FIG. 4. A: Power required to ventilate a lung with a total compliance of 1 litre kpa" 1 and a total resistance of 0.6 kpa litre"' 8 derived theoretically using the same flow pattern as during the simulated breaths shown in figure 3. Curves R and C show the power dissipated in the resistance and absorbed and released by the compliance, respectively. B : Accumulated work over simulated spontaneous breaths. Curve C + R shows the work calculated from the power given in A and represents the total work done against the natural compliance and resistance without any additional load. Curves C + R + D and C + R + A show the total work of breathing (see text) when the alternative (D) and ventilator (A) IMV systems are connected to the above compliance and resistance. abnormal (Gilston, 1977). It is clear from the present study that a low breathing system resistance is also of considerable importance during weaning with IMV. The humidifier was the major contributor to the inspiratory resistance of the original IMV system. The Draft International Standard 8185, "Humidifiers for Medical Use", does not specify a maximum resistance to flow, but defines a test for measuring such resistance and states that the resistance toflow must be quoted if the humidifier BRITISH OF JOURNAL ANAESTHESIA is stated to be suitable for use within a breathing system attached to a spontaneously breathing patient. The Draft International Standard 5369, "Breathing Machines for Medical Use", is equally unhelpful; it defines a maximum expiratory resistance for the breathing system of a ventilator, but it does not mention a value for inspiratory resistance. However, the Draft International Standard 8382, " Resuscitators Intended for Use with Humans", does define a maximum inspiratory resistance as a maximum pressure below ambient of 0.5 kpa at the patient connection when an inspiratoryflowof 50 litre min" 1 is drawn from the resuscitator (equivalent to a resistance of 0.6 kpa litre" 1 s). In comparison, normal human respiratory tract resistance is approximately 0.2 kpa litre" 1 s at a flow rate of 0.5 litre s" 1 (Nunn, 1977) and for anaesthetized patients it may be in the range of 0.4-0.6 kpa litre" 1 s at the same flow rate (Bergman, 1969). The ventilator breathing system studied in this communication exhibited a total resistance of 1.2 kpa litre" 1 s at a flow rate of 0.5 litre s" 1, the humidifier contributing about one-half of this (0.64 kpa litre" 1 s). This resistance was about twice the upper limit recommended by Nunn (1977) compare curves A and E in figure 2 whereas the alternative IMV system exhibited an inspiratory resistance well below the limit. However, when the ventilator breathing system, with the humidifier and filter removed, was compared with the alternative IMV system, then there was only a small difference in resistance curves C and D. Therefore, if the alternative IMV system is "teed " into the ventilator breathing system on the ventilator side of the humidifier, as is occasionally done, the reduction in resistance will be negligible. The patient has to do work against his compliance as well as against resistance; therefore, rather than compare just the added resistance with the natural resistance, it is more relevant to compare added work with natural work. Accordingly, the theoretical instantaneous power required to ventilate a lung with characteristics representative of those of a conscious, intubated patient (compliance = 1 litre kpa" 1 ; resistance = 0.6 kpa litre" 1 s) was calculated for theflow pattern used in the experimental measurements. The results are shown in figure 4A where curve R indicates the power dissipated in the resistance and curve C the power associated with the compliance.

IMV: WORK OF BREATHING 1053 The fact that curve C is positive in inspiration and negative in expiration reflects the fact that power is absorbed by the compliance during inspiration and released during expiration, that is, the energy stored within the compliance during inspiration is available to do work during expiration. The accumulated work (on compliance plus resistance) is the sum of the integrals of curves C and R in figure 4A and, for a complete breath, is shown by curve C + R in figure 4B. This "internal" work is the work which the patient must do on his own respiratory system and the tracheal tube when breathing from atmosphere. The reasons for the dip during expiration is that the energy released by the compliance is greater than that required to overcome resistance. This is discussed further below. The accumulated work over the complete respiratory cycle is of the order of 0.2 J and is done mainly during inspiration. When such a lung is connected to a breathing system, the work required to overcome the impedance of the breathing system must be added to the " internal" work. In these circumstances, with the complete ventilator breathing system (curve C + R + A in figure 4B), the total work required was about 0.5 J, an increase of approximately 150% of that done in breathing directly from atmosphere; with the alternative IMV breathing system (curve C + R + D) the work required was about 0.3 J, an increase of only 50 %. Note that curve C + R + D is the sum of curve C + R in figure 4B and curve D in figure 3d. Similarly, curve C + R + A is the sum of curve C + R and curve A. (The expiratory portions of the curves in figure 4B are a little artificial, because the Starling pump produces a somewhat unphysiological waveform of expiratory flow. However, this does not materially affect the above argument, because the main differences between the three curves lie in the inspiratory portion). Although we have argued in favour of expressing the added burden of a breathing system in terms of work instead of resistance, the calculations which we have made for a fixed flow pattern through different loads still do not give a completely fair picture of energy consumed by the muscles because of the following five factors: (1) During expiration, work must be done on respiratory resistance and any external load, but only in extreme circumstances will the expiratory muscles be called into play to provide the necessary energy. Normally, energy stored in the compliance during inspiration will be adequate to do this work. However, an additional complication is that the patient's respiratory muscles normally oppose expiration initially, because the inspiratory muscles relax only gradually during the first part of expiration. Therefore, mechanically, work is then being done by the compliance, not only on the resistance, but also on the respiratory muscles. This would produce a decline in accumulated work in the first part of expiration, much as in curve C + R of figure 4B (between 1 and 2 s) although, there, the decline can be attributed to the model compliance doing work on the model pump. However, it seems most unlikely that the muscles can in any way make use of the energy coming from the compliance and they are almost certainly still consuming oxygen during this gradual relaxation as a result of being in a state of contraction. (2) When a load is imposed on a patient, the waveform of flow during a respiratory cycle may change. (3) In addition, the tidal volume and frequency may change even though the total, or at least alveolar, ventilation may remain much the same. (4) The respiratory muscles also do mechanical work on the circulation the "thoracic pump". Therefore, if the magnitude or pattern of the intrapleural pressure changes with external load, the amount of work done on the circulation may change. (5) Any change in the magnitude or pattern of mechanical work may affect the efficiency of the respiratory muscles and, hence, their oxygen consumption which, rather than the mechanical work done, may be the limiting factor in some patients being weaned from a ventilator. Despite all these caveats, it remains true that estimates of the change in the mechanical work of one breath, offixedtidal volume, duration and I: E ratio (on breathing through an IMV system instead of from the atmosphere) are more informative than changes of total resistance especially since some IMV systems may impose a compliance load or may actually assist the patient's breathing (Bingham, Hatch and Helms, 1986). CONCLUSION During controlled ventilation all the work of breathing is done by the ventilator and, hence, the inspiratory resistance of the breathing system is relatively unimportant. However, as soon as IMV

1054 BRITISH OF JOURNAL ANAESTHESIA is introduced, the patient has to do the work during the spontaneous breaths; thus, the characteristics of the IMV system are just as important as with any breathing system for spontaneous ventilation. Therefore, before using the IMV mode of a ventilator, the anaesthetist should consider the characteristics of its breathing system for spontaneous inspiration and, in particular, note that a humidifier which is perfectly satisfactory during controlled ventilation may be unacceptable for IMV use. It is recommended that manufacturers should publish resistance values of breathing system components if they are intended for use with both controlled and IMV modes of ventilation. REFERENCES Bergman, N. A. (1969). Properties of passive exhalations in ancsthetised subjects. Anesthesiology, 30, 378. Bingham, R. M., Hatch, D. J., and Helms, P. J. (1986). Assisted ventilation and the Servo ventilator in infants. An assessment of three systems used for CPAP/IMV. Anaesthesia, 41, 168. Browne, D. R. G. (1984). Review article, weaning patients from mechanical ventilation. Intent. Care Med., 10, 55. Downs, J. B., Klein, E. F., Desautels, D., Modell, J. H., and Kirby, R. R. (1973). Intermittent mandatory ventilation: A new approach to weaning patients from mechanical ventilators. Chest, 64, 331. Engstrom, G. G., and Norlander, O. P. (1962). A new method for analysis of respiratory work by measurement of the actual power as a function of gas flow, pressure and time. Acta Anaesthesiol. Scand., 6, 49. Gilston, A. (1977). Intermittent mandatory ventilation: Are IMV, MMV, PEEP or sighing advantageous? Anaesthesia, 32, 665. Mcllroy, M. B., Marshall, R., and Christie, R. V. (1954). The work of breathing in normal subjects. Clin. Sci. 13, 127. Margand, P. M. S., and Chodoff, P. (1975). Intermittent mandatory ventilation. An alternative winning technic, a case report. Anesth. Analg., 54, 41. Nunn, J. F. (1977). Applied Respiratory Physiology, 2nd Edn, p. 103. London: Butterworths.