Demand and continuous flow intermittent mandatory ventilation systems

Weaning from Mechanical Ventilation

For most patients who require mechanical ventilation weaning and extubation is simple. In these patients a variety of strategies can be successful. In addition, sim ple criteria may predict when the patient is ready for extubation. For the small group of patients who require prolonged mechanical ventilation, however, contro versy exists about how best to remove ventilator sup port by weaning, and available data are sparse. Much of the controversy has centered on T-piece weaning ver sus intermittent mandatory ventilation. To date no con trolled study has demonstrated the superiority of either intermittent mandatory ventilation or T-piece weaning in difficult-to-wean patients. In the evolution of this con troversy, concern has developed over the potential for increased inspiratory work and expiratory resistance that may be associated with certain intermittent manda tory ventilation systems. The possibility that significant inspiratory work may occur during assist-control venti lation has also been demonstrated. Respiratory muscle weakness and fatigue is likely important in failure to wean. Other possible causes are failure of the cardiovas cular system and impaired ability of the lung to carry out gas exchange. In this article we first examine criteria and techniques for weaning short-term ventilator pa tients. We then examine criteria to begin the weaning process in prolonged ventilation patients, potential causes of failure to wean, and techniques that can be used to remove ventilator support from patients who are difficult to wean. Much literature has been devoted to techniques and criteria for weaning and extubation of patients from mechanical ventilation. For most patients who require ventilatory support, weaning and extuba tion can be easily accomplished by a variety of tech niques [1-4]. At one referral center 77.2% of all surviving patients were weaned from the ventilator within 72 hours of the onset of mechanical ventila tion, and 91% were weaned within 7 days [1]. Less than 10% of ventilated patients potentially posed problems in weaning from mechanical ventilation. Similarly, at a community hospital, few surviving patients required prolonged ventilatory support [2]. In easy-to-wean patients, Sahn and Lakshminarayan [5] described simple criteria that are predictive of successful discontinuation of ventilator support. For the small group of patients who require pro longed mechanical ventilation, however, minimal data are available. In these patients criteria to deter mine weaning ability or which measurements to follow are not clearly defined. Furthermore, no controlled trials are available to compare the differ ent weaning techniques proposed. In this article we first address routine weaning of the patient who has not required prolonged ventilator support. We then examine the difficult-to-wean patient and dis cuss criteria to begin the weaning process, poten tial causes of failure to wean, and available weaning techniques.

Pressure modes of invasive mechanical ventilation

Pressure modes of invasive mechanical ventilation generate a tidal breath by delivering pressure over time. Pressure control ventilation (PC) is the prototypical pressure mode and is patient- or time-triggered, pressure-limited, and time-cycled. Other pressure modes include pressure support ventilation (PSV), pressure-regulated volume control (PRVC, also known as volume control plus [VC+]), airway pressure release ventilation (APRV), and biphasic ventilation (also known as BiLevel). Despite their complexity, modern ventilators respond to patient effort and respiratory system mechanics in a fairly predictable fashion. No single mode has consistently demonstrated superiority in clinical trials; however, empiric management with a pressure mode may achieve the goals of patient-ventilator synchrony, effective respiratory system support, adequate gas exchange, and limited ventilator-induced lung injury.

Effects of flow triggering on breathing effort during partial ventilatory support

The effects of flow triggering (FT) as compared with pressure triggering (PT) on breathing effort have been the focus of several studies, and discrepant results have been reported. In the initial part of our study, a lung model was used to quantify triggering effort (airway pressure-time product, PTPaw) for a range of sensitivity settings in nine new-generation ventilators. A ventilator providing both FT and PT was then used to compare these systems during pressure-support (PSV) and volume-targeted assist-control ventilation (ACV) in eight ventilator-dependent patients, using sensitivity settings (2 L/min for FT and -2 cm H2O for PT) that had proven significantly different in the initial bench study. Indexes of effort included the esophageal and transdiaphragmatic pressure-time products and inspiratory work of breathing per minute (PTPes/min, PTPdi/min, and Wi/min, respectively). The experimental study revealed significant differences between ventilators in PTPaw at commonly used settings. In two of three ventilators featuring both systems, PTPaw was significantly lower with FT than PT (p < 0.001). In the clinical study, FT as compared with PT, was associated with reductions in all indexes of breathing effort during PSV: 16 +/- 6% (p < 0.001), 13 +/- 10% (p < 0.01), and 14 +/- 12% (p < 0.05) for PTPdi/min, PTPes/min, and Wi/min, respectively. By contrast, no differences were found when FT was used during ACV. Although FT reduced triggering effort in both modes (p < 0.001), the effects observed during the post-trigger phase differed, and explained the discrepant results between the two modes. We conclude that FT more effectively reduces breathing effort when used in conjunction with a pressure-targeted mode than with a volume-targeted mode, especially when flow delivery is close to or below demand.

Effects of different triggering systems and external PEEP on trigger capability of the ventilator

OBJECTIVE

The triggering capability of both the pressure and flow triggering systems of the Servo 300 ventilator (Siemens-Elema, Sweden) was compared at various levels of positive end-expiratory pressure (PEEP), airway resistance (R(aw)), inspiratory effort and air leak, using a mechanical lung model.

DESIGN

The ventilator was connected to a two bellows-in-series-type lung model with various mechanical properties. Lung compliance and chest wall compliance were 0.03 and 0.121/cmH2O, respectively. R(aw) was 5, 20 and 50 cmH2O/l/s. Respiratory rate was 15 breaths/min. To compare the triggering capability of both systems, the sensitivity of pressure and flow triggered pressure support ventilation (PSV) was adjusted to be equal by observing the triggering time at 0 cmH2O PEEP and 16 cmH2O of pressure support (PS) with no air leak. No auto-PEEP was developed. In the measurement of trigger delay, the PS level ranged from 16 to 22 cmH2O to attain a set tidal volume (V(T)) of 470 ml at a R(aw) of 5, 20 and 50 cmH2O/l/s. The PEEP level was then changed from 0, 5 and 10 cmH2O at a PS level of 17 cmH2O and R(aw) of 5 and 20 cmH2O/l/s, and the trigger delay was determined. The effect of various levels of air leak and inspiratory effort on triggering capability was also evaluated. Inspiratory effort during triggering delay was estimated by measurements of pressure differentials of airway pressure (Paw) and driving pressure in the diaphragm bellows (Pdriv) in both systems.

MEASUREMENTS AND RESULTS

There were no significant differences in trigger delay between the two triggering systems at the various PEEP and R(aw) levels. At the matched sensitivity level, air leak decreased trigger delay in both systems, and additional PEEP caused auto-cycling. A low inspiratory drive increased trigger delay in the pressure sensing system, while trigger delay was not affected in the flow sensing system. The Paw and Pdriv differentials were lower in flow triggering than in pressure triggering.

CONCLUSIONS

With respect to triggering delay, the triggering capabilities of the pressure and flow sensing systems were comparable with and without PEEP and/or high airway resistance at the same sensitivity level, unless low inspiratory drive and air leak were present. In terms of pressure differentials, the flow triggering system may require less inspiratory effort to trigger the ventilator than that of the pressure triggering system with a comparable triggering time. However, this difference may be extremely small.

Characteristics of the ventilator pressure- and flow-trigger variables

Pressure- and flow-triggering are available in the Puritan Bennett 7200ae and Siemens SV 300. Using a mechanical lung model, we described the characteristics of the pressure- and flow-triggered continuous positive airway pressure (CPAP) of both ventilators. In the Puritan Bennett 7200ae, the pressure-triggered CPAP is characterized by the relatively insufficient flow delivery after the triggering, resulting in a greater lung pressure-time product (total PTP) than the flow-triggered CPAP. Pressure support of 5 cmH2O results in total PTP less than that with flow-triggered CPAP. In the Siemens SV 300, total PTP with pressure- or flow-triggered CPAP is comparable. Total PTP is less with pressure- or flow-triggered CPAP of the Siemens SV 300 than that of the Puritan Bennett 7200ae, respectively. The application of small pressure- or flow-triggered pressure support in the Puritan Bennett 7200ae eliminates the difference. The impact of these differences on patient inspiratory muscle work remains to be determined.

Synchronized intermittent mandatory ventilation with and without pressure support ventilation in weaning patients with COPD from mechanical ventilation

This prospective study compared two weaning modalities in COPD patients requiring mechanical ventilation (MV) for acute respiratory failure. Nineteen patients with COPD were studied when their precipitating illness was controlled. Although they satisfied the conventional bedside weaning criteria, they could not tolerate any reduction in the respirator rate below 10 cycles/min. At this time, patients were randomized into two groups receiving either synchronized intermittent mandatory ventilation (SIMV) with pressure support ventilation (PSV) (group 1) or SIMV alone (group 2). The volumetric support of ventilation (SIMV rate) was progressively decreased in both groups according to the patient's tolerance with a concurrent decrease in the barometric support of ventilation (PSV levels from 15 cm H2O to 6 cm H2O). At each step of SIMV rate, we found no difference between group 1 and group 2 in arterial blood gases, blood pressure, heart rate, airway occlusion pressure, maximal inspiratory pressure, and oxygen cost of breathing (OCB). At each step, however, group 1 patients showed significantly higher spontaneous tidal volume and lower spontaneous breathing frequency than did group 2 patients. We found a slight but not significant tendency to a shorter weaning period with than without PSV, but no difference in the weaning success. We concluded that (1) conventional weaning criteria might be inaccurate in COPD patients, (2) SIMV appeared very useful in weaning COPD patients from MV, (3) PSV marginally reduced the weaning period when added to SIMV, and (4) the OCB was not significantly improved with PSV.

The performance of two intensive care humidifiers at high gas flows

High continuous flow breathing systems are now available to provide fresh gas flows well in excess of 100 l.min-1 in continuous positive airway pressure systems used for respiratory support. The performance of two commonly used intensive care humidifiers, the Kendal Conchatherm and the Fisher and Paykel FP310 have been assessed at flows of 50, 75, 100, 125 and 150 l.min-1. Their performance when using two humidifiers connected in parallel and in series was also studied. At a fresh gas flow of 100 l.min-1 the single Conchatherm gave an absolute water vapour concentration of 15.6 g.m-3 and the single FP310 11.9 g.m-3. At all flows the best results were achieved using two Conchaterm humidifiers in series. It is concluded that with high continuous flow breathing systems the humidification achieved with conventional humidifiers may be inadequate and it may be necessary to combine two humidifiers to obtain clinically useful humidification.

Test of 20 similar intensive care ventilators in daily use conditions--evaluation of accuracy and performances

Infrequent control, aging of components, may compromise the accuracy of ICU ventilators. In order to assess the reliability of ventilators during their clinical use, we bench tested a group of 20 CPU1 ventilators (Ohmeda) sampled at random in several ICU units. We found major leaks in 5 ventilators, attributable to the disposable tubings used in these systems. Mean error in expired tidal volume and corresponding standard deviation (precision) were greater than 100 ml in two. Positive end expiratory pressure measurement comprised a mean error higher than 2 cm H2O in 40% of the ventilators tested. The valve opening pressure threshold was correlated to the inspiratory flow (r = 0.81) contrary to the valve opening delay (average 138 +/- 40 ms). These two parameters did not correlate with the age of the ventilator. Our study addresses the need for periodic control of ventilator performance in order to minimize the risks of errors and malfunctions.

Analysis of resistance to gas flow in nine adult ventilator circuits

We measured the resistance in nine complete ventilator circuits, partial circuits and 7, 8, and 9 mm ID endotracheal tubes at flow rates of 20 to 120 liters per minute. We found a statistically significant (p less than 0.01) increase in resistive pressure with increases in flow rate, as the diameter of the ETT decreased, and as each component of the ventilator circuit was added to the ETT. There was a curvilinear increase in resistive pressure to increase in flow rate. However, when resistances were computed, the Bennett cascade "circuit" created higher resistance at 20 lpm than at flow rates up to 120 lpm. The Bennett cascade humidifier added the greatest resistive pressure, 3.5 to 8.5 cm H2O, the Engstrom Edith, 0.5 to 6.5 cm H2O, and the Conchapak added the least, 0.0 to 2.5 cm H2O at flow rates of 20 to 120 lpm. After all the components of the ventilator circuit were attached to the ETTs, there was approximately a 97 to 450 percent increase in resistive pressure compared to the resistive pressure created by the ETTs alone.

A prospective comparison of IMV and T-piece weaning from mechanical ventilation

Two hundred (200) consecutive medical and surgical patients requiring mechanical ventilation were entered into a prospective randomized trial of weaning by either intermittent mandatory ventilation (IMV) or T-piece. Patients in these groups were of similar age and sex and had the same total ventilation time (TVT). The study design provided equal time for each weaning mode after specific criteria for oxygenation and ventilation were satisfied (PaO2 greater than 55 mm Hg on FIO2 less than 0.5; VE less than 12 L/min and two of the following four parameters: MVV greater than 2 VE, VT greater than 5 ml/kg, FVC greater than 10 ml/kg, NIF less than or equal to -20 cm H2O). Of the original 200 patients 165 were entered into the weaning phase; 35 patients were withdrawn prior to weaning due to the discretion of the attending physician or protocol error. Weaning time was not different between the IMV (5.3 +/- 1.2 h, mean +/- SEM) and T-piece groups (5.9 +/- 1.4 h, p = NS). Of the 165 patients, 155 (93 percent) were weaned successfully by protocol, 79 in the IMV and 76 in the T-piece group. Of 155 patients, 136 (88 percent) were weaned on the first attempt by protocol. Of the 19 who were not weaned, 11 were weaned successfully on the second and five on the third trial; three patients required three-day weans. We conclude that clinically stable patients who require short-term mechanical ventilation and meet standard bedside weaning criteria can be weaned efficiently by protocol using either IMV or T-piece techniques.

Chest wall distortion in patients with flail chest

Ventilators can impose resistive and elastic loads during subject-initiated and spontaneous breaths. Such loads might worsen the chest wall distortion that is characteristic of patients with flail chest. We have tested this expectation in nine patients with flail chest and four normal subjects. All subjects breathed for 3 to 5 min on each of the following modes: assist control, intermittent mandatory ventilation (IMV), continuous positive airway pressure 5 to 10 cm H2O by demand valve and by a high flow system (CPAP-HF), and spontaneously (T-piece). Pressure at the airway opening was evaluated as a measure of ventilator loading, and magnetometric displacements of the major chest wall dimensions were evaluated to assess chest wall distortion. In contrast to the normal volunteers, patients with flail chest displayed chest wall distortion during active inspirations. The patterns of distortion were variable among patients. The degree of distortion varied among ventilator modes; generally, there was a greater degree of chest wall distortion in breaths with greater loading. For example, distortion was greater during the spontaneous breaths taken on the IMV-mode than during spontaneous breaths taken on the T-piece. The CPAP-HF mode resulted in the least distortion, reversing chest wall distortion in five patients, improving it in two, and not changing the distortion in the remaining two. The improvements may be related to positive pleural pressures and to the minimal ventilator-imposed load of the high gas flow system. The distortion imposed by ventilators increases the work of breathing in these patients and may thus contribute to difficulty in weaning.

Flow resistance of expiratory positive-pressure systems

We measured the flow-resistance of five commercially available 10 cm H2O expiratory positive-pressure (EPP) valves (n = five per valve type) at bias flows of between 0 and 2,000 ml/s. We found that individual valves of each type and manufacturer functioned similarly. Different valve types, however, functioned differently: with one type, system pressure was higher than rated (p less than 0.05), and with another type, system pressure was significantly flow-dependent (p less than 0.01). The remaining types of valves had no flow-resistive properties and maintained a system pressure of 10 cmH2O. We conclude that system pressure is not similar in all continuous positive airway pressure (CPAP) systems using bias flow and EPP valves. The work of breathing imposed by CPAP circuits will be increased in systems whose EPP valves have flow-dependent properties.

Inspiratory effort and occlusion pressure in triggered mechanical ventilation

We have studied eleven patients ventilated in the assisted mode during recovery from acute respiratory failure. We have measured the effort required to trigger the pressure demand valve for 3 different ventilators, and have measured the occlusion pressure as an index of neuromuscular inspiratory drive. We found a delay in the opening of the demand valve, as previously described by other authors. We also found a close correlation between the effort required to open the demand valve and the occlusion pressure. We conclude that the inspiratory effort required to open the demand valve, in the assist mode, is greater than the preset trigger level and that it is well correlated with the neuromuscular inspiratory drive. This inspiratory effort against the closed demand valve, allows the measurement of the occlusion pressure.

A high flow turbine CPAP system

A continuous high flow CPAP system incorporating a turbine blower is described. The system achieves inspiratory flow rates of 150 l/min or more by means of reticulated gas flow and inspired oxygen fractions of 0.21-0.95. Positive airway pressure is provided by weighted disc valves and a modified aviation-type CPAP face mask provides electronic communication with the patient. The mobility of the system also enables its use as an intermittent physiotherapy aid. Work of breathing of the system, as assessed by total pressure fluctuations is at a minimum.

Ventilation and ventilators--an update

In the five years which have passed since the previous review, the literature has been concerned more with the ways in which ventilators may be applied to patients and the effects of differing patterns of ventilation than with the design philosophy of the ventilators themselves. This account should be read in conjunction with that of 1982 [1].

Inspiratory work of breathing during spontaneous ventilation using demand valves and continuous flow systems

To diminish the work of breathing, some demand valve systems are equipped with inspiratory pressure support (IPS). The purpose of this study was to evaluate the work performed during spontaneous breathing using the Siemens Servo 900C ventilator (SVC) at the minimal IPS level available, and comparing it with a demand valve ventilator without IPS, the Ohmeda CPU1 (CPU1), and with a home-built continuous flow system (CFS). We found a larger minute ventilation and inspiratory peak flow with the SVC and the CFS than with the CPU1 (p less than 0.05). When the work of breathing was measured at the airway opening, we found that the CFS led to the least amount of work (0.17 +/- 0.05 J.L-1, p less than 0.05). Additionally, this work was strikingly less for SVC than for CPU1 (0.22 +/- 0.06 versus 0.42 +/- 0.10 J.L-1, p less than 0.001) as a result of a higher flow supplied by the SVC. By contrast, the work measured using the esophageal pressure, i.e., including the work dissipated on the lung and airways, was significantly reduced with the CFS (1.34 +/- 0.45 J.L-1, p less than 0.05), but surprisingly not different between SVC and CPU1 (1.49 +/- 0.57 versus 1.54 +/- 0.51 J.L-1). We conclude that the absence of a demand valve in CFS involves the lowest work of breathing. Likewise, the high flow capability in SVC reduces the work necessary to overcome the circuit and demand valve resistances.(ABSTRACT TRUNCATED AT 250 WORDS)

A continuous high flow intermittent mandatory ventilation system incorporating a Downs venturi device and a modified Nuffield Series 200 ventilator

Systems for respiratory support are becoming increasingly expensive and complex. Many systems suffer inadequacies when used for spontaneous ventilation. Some modes on newer ventilators are rarely used because of the complex controls and settings. There is no truly universal ventilator that satisfies every intensivist's wishes. CPAP/IMV is becoming accepted as the standard management of many patients with acute respiratory failure and there would be few intensive care units where CPAP/IMV is not used for part of a patient's respiratory support. We describe a cost-effective system that may be used for respiratory support in the spontaneously breathing mode. This system combines a high flow venturi, an efficient humidifier and an inexpensive reliable ventilator that can be used for adult and paediatric patients. The system, primarily for use in patients breathing spontaneously, functions well in patients requiring full ventilation.

Pressure support compensation for inspiratory work due to endotracheal tubes and demand continuous positive airway pressure

We evaluated the use of pressure support to compensate for the added inspiratory work of breathing due to the resistances of endotracheal tubes and a ventilator demand-valve system for continuous positive airway pressure (CPAP). A mechanical model was used to simulate spontaneous breathing at five respiratory rates through 7-mm, 8-mm, and 9-mm endotracheal tubes with and without a ventilator demand CPAP circuit. Added work was measured as the integral of the product of airway pressure and volume during inspiration. Additional work was a function of the tube's size, and each 1-mm decrease in the tube's diameter resulted in a 67 to 100 percent increase in work. Adding the ventilator CPAP circuit further increased work and was responsible for 30 to 50 percent of the total work resulting from a tube and CPAP circuit together. Pressure support was added to a level at which net work on the airway was zero, and a relationship between mean inspiratory flow (VT/TI) and the optimal level of pressure support was established for each endotracheal tube. The inspiratory work of breathing was then measured in normal subjects breathing with and without each endotracheal tube plus the demand CPAP circuit. Work per liter of minute ventilation due to the endotracheal tube and CPAP circuit was increased from 54 to 240 percent over levels measured while breathing through an open airway. For each endotracheal tube and VT/TI, a level of pressure support (range, 2 to 20 cm H2O) was found which eliminated added work in the spontaneously breathing subject. This level correlated well with that predicted from the data derived using the mechanical model. We conclude that when adjusting for an endotracheal tube's diameter and VT/TI, pressure support can be used to compensate for the added inspiratory work due to artificial airway resistances.

Use of the oxygen cost of breathing as an index of weaning ability from mechanical ventilation

The oxygen cost of breathing (which is the difference in oxygen consumption measured during controlled ventilation and again during spontaneous ventilation) was measured in 30 patients between the ages of 17 and 96 years at the time of commencement of weaning from mechanical ventilation. There was a significant exponential correlation between the oxygen cost of breathing in ml/m2/min and the oxygen cost of breathing as a percentage of total oxygen consumption during spontaneous ventilation (OCB/VO2SV%) and the duration of weaning in days.

Evaluation of the comfort of spontaneous respiration through three ventilator systems

The use of intermittent mandatory ventilation and continuous positive airways pressure systems is widespread. The comfort of spontaneous ventilation through three systems, with a variety of humidifiers, has been evaluated. The use of demand flow systems and the introduction of some humidifiers caused considerable discomfort from fluctuations in the airway pressure. It is concluded that continuous flow systems are preferable and that fluctuations in airway pressure should not exceed 2.5 cm H2O.

Improved efficacy of spontaneous breathing with inspiratory pressure support

During inspiratory pressure support (IPS) ventilation, first a negative airway pressure is produced by the patient to open a demand valve and then a constant positive airway pressure is maintained at a present level while the patient inhales. The aim of this study was to assess the ability of 10 cm H2O IPS to improve the efficacy of spontaneous ventilation. We studied 8 intubated patients recovering from acute respiratory failure, all were breathing spontaneously via 3 different systems: a Servo 900 C ventilator (SCV) without IPS, a Servo 900 C ventilator with 10 cm H2O IPS, and a continuous flow system (CFS). Compared with the CFS, breathing with the SVC without IPS resulted in an increased respiratory rate (RR), increased tidal Volume (VT), increased transdiaphragmatic pressure (Pdi), and no significant change in PaO2 or PaCO2. Ventilation with IPS resulted in significant improvements in VT, PaO2, and PaCO2 with a decreased RR and Pdi when compared with both the other modes of spontaneous ventilation. A significant decrease in the pressure-time index of the diaphragm (i.e., the product of the mean transdiaphragmatic pressure and the inspiratory duty cycle) occurred during IPS. In 2 patients, we recorded diaphragmatic electromyographic activity during both SVC and IPS. In both patients during IPS, an increased VT and a decreased Pdi coincided with a major reduction of electromyographic activity. We conclude that IPS at a level of 10 cm H2O markedly increases the efficacy of spontaneous breathing while reducing the activity of the inspiratory muscles.

Newer modes of mechanical ventilatory support

Recent modes of ventilatory support aim to facilitate weaning and minimise the physiological disadvantages of intermittent positive pressure ventilation (IPPV). Intermittent mandatory ventilation (IMV) allows the patient to breathe spontaneously in between ventilator breaths. Mandatory minute volume ventilation (MMV) ensures that the patient always receives a preset minute volume, made up of both spontaneous and ventilator breaths. Pressure supported (assisted) respiration is augmentation of a spontaneous breath up to a preset pressure level, and is different from 'triggering', which is a patient-initiated ventilator breath. Other modes or refinements of IPPV include high frequency ventilation, expiratory retard, differential lung ventilation, inversed ratio ventilation, 'sighs', varied inspiratory flow waveforms and extracorporeal membrane oxygenation. While these techniques have useful applications in selective situations, IPPV remains the mainstay of managing respiratory failure for most patients.

New generation ventilators

Desirable features of new generation intensive care ventilators include the ability to ventilate a wide range of patient sizes, an uncomplicated control panel, an appropriate but not excessive variety of ventilatory patterns, adequate patient monitoring and alarm functions, and simplicity of cleaning and routine maintenance. Examples of currently available ventilators include the Servo 900-C, CPU-1, Engstrom Erica, Bear 5, Drager EV-A and Hamilton Veolar. The incorporation of microcomputer control into some of these ventilators has resulted in improved flexibility and a limited number of automatic responses to detected patient changes. However, the function of components provided to allow spontaneous ventilation, such as demand valves, requires considerable improvement. Current trends in ventilator design include further refinement of computer control and the provision of graphic displays showing the results of continuous sophisticated analysis of respiratory function. The extent to which these developments will prove clinically useful will require careful evaluation.