Automated Determination of Respiratory Rate

Determination of respiratory rate by direct observation is difficult because of interobserver variability, length of time needed for accurate readings (up to 1 minute), and the confounding effects of upper airway obstruction that produce chest wall motion and apparent breaths in the absence of gas exchange. The reliable automation of this data gathering is valuable, not only to perioperative caregivers—especially those who perform regional or sedation techniques—but also to nonanesthesiology personnel responsible for patient sedation in more remote or ambulatory locations.

Instrumentation to measure respiration may be direct, as in airflow sensors—pinwheels, gas sampling, or thermocouples—or indirect, imputing air flow from changes in thoracic dimension or breath sounds. Indirect measurements of chest volume may be via strain gauges, impedance measurements, or electrocardiograph (ECG) analysis. In a fully equipped operating suite, the most convenient method of measurement is with a capnograph. In the tracheally intubated, anesthetized patient and in the awake or sedated patient, exhaled gas flow may be sampled using a number of devices. In the case of the spontaneously breathing patient, a nasal cannula may be configured to both sample exhaled gas and simultaneously deliver supplemental oxygen.^7,8

Thoracic Impedance and Inductance

An early attempt to automate respiratory rate measurement in the 1960s was based on the theory that the electrical properties of the thorax changed during the respiratory cycle. The resistance to the flow of alternating current, known as impedance, was found to change with the inflation and deflation of the lungs; it is increased with inhalation. Because of this, this impedance value could be continuously measured; when it exceeded a set threshold, the occurrence of a breath was recorded by a monitor. Efforts were made to correlate changes in impedance with tidal volume, but this was less successful in awake patients and in those with respiratory disease.^9,10 Significantly, this technique does not measure actual air entry but only measures the changes in the shape of the thoracic cage. Accordingly, it could be fooled by respiratory effort in the presence of upper airway obstruction.¹¹ In anesthetized and ventilated surgical patients, however, the technique was much more valuable, producing accurate respiratory rate measurements and high correlation of tidal volumes (r = 0.89) with Wright’s respirometer.¹² In addition, the technique was somewhat sensitive to both airway occlusion and circuit disconnects.

Work has been done to determine the optimal placement of electrode pairs for maximum sensitivity. As expected, maximal separation of electrode pairs provides the most sensitive signal.¹³ Interference between respiratory monitors that use thoracic impedance and implanted respiratory rate–sensitive pacemakers has been reported, producing unwanted heart rate changes. The only solution was to disable the respiratory rate monitoring function in the physiologic monitor.¹⁴

Inductive plethysmography is superficially similar to impedance plethysmography in that electrically active sensors detect chest wall movement. However, the inductive technique uses direct-current sensors in the form of coils of wire around the chest and/or abdomen. The changes in self-inductance (stretch) of these sensors are electrically transmitted and analyzed. These signals can reflect respiratory rate and, after calibration, tidal volume.¹⁵ In a study of infants at risk for apnea, Brouillette and colleagues¹⁶ found that inductive plethysmography was superior to impedance-based measurement in that it recorded fewer cardiac artifacts and was better able to discern upper airway obstruction.

Electrocardiograph-Based Respiratory Monitoring

After the use of continuous ECG monitoring became widespread, attempts were made to use ECG-derived data to record respiratory rate. The ECG may be used through the analysis of changes in both the amplitude of the ECG signal and the changes in the axis of the QRS complex. Amplitude changes in the ECG are caused by the change in overall electrical resistance because of the increased proportion of air, rather than tissue, within the chest cavity during inspiration. This is reflected in a decrease in ECG signal voltage, and thus QRS wave size, that is amenable to analysis.

Inflation of the lungs also produces a change in the axis of the heart, as reflected in the multilead ECG. Using two limb leads, measurements are made of the area under a normal QRS complex. By using these areas as amplitudes of known direction, a vector can be derived as the sum of the two values. These measurements are routinely made by most arrhythmia detection software in current use.¹⁷

Other Techniques

Acoustic techniques actually “listen” for the sound of airflow at the face, pretracheal neck, or chest. One method uses the sound of air impinging on open-ended tubes positioned near the nostrils and mouth.¹⁸ Although lightweight and immune to upper airway obstruction artifact, this method may suffer from interference from ambient noise. Also, it is a pure respiratory frequency monitor, without any other measure of ventilatory adequacy. A variation of this approach uses a pyroelectric polymer mounted inside a face mask that senses the temperature increase from exhaled air. This signal is electrically amplified and displayed. Inexpensive and relatively simple, it can be incorporated into the construction of a face mask or nasal cannula.¹⁹

A sophisticated, noninvasive sensor was described by Zhu and colleagues in 2006.²⁰ A head pillow containing fluid-filled tubes is connected to two sensitive pressure transducers. The pressure signal is processed with a derivation of an algorithm known as wavelet transformation and is filtered to produce both pulse and respiratory signals. The technique is, as expected, less effective in the presence of motion artifact and poor head positioning. However, in 13 healthy subjects, both the sensitivity and the positive predictive value of the sensor were above 95% for respiration, compared with simultaneous nasal thermistor recordings.²⁰ The data were even more favorable with respect to detection of pulse rate. In combination with oximetry, this technique will likely prove valuable both in ambulatory monitoring and in the diagnosis of sleep disorders.

The Rainbow Acoustic Monitoring system (Masimo, Irvine, CA) noninvasively and continuously measures respiration rate using an innovative adhesive sensor with an integrated acoustic transducer that is easily and comfortably applied to the patient’s neck. Using acoustic analysis, the respiratory signal is separated and processed to continuously display respiration rate.²¹ Box 9-1 summarizes the various methods for automated detection and recording of respiratory rate.

Principles Overview

Pressure is defined as the force per unit area exerted on a surface by an external substance, usually a gas or liquid in medical applications. A molecule of a gas can be envisioned as confined within a cube. This molecule of gas transfers its momentum to a vessel wall during a collision equal to 2 × mv, where m is mass and v is mean velocity. After bouncing off the opposite wall, the molecule then hits the first wall again after a time equal to 2 × d/v, where d is the distance between walls, exerting a force of mv²/d. With N molecules, one third will hit each pair of sides of the box, yielding a force (F) on each wall:

Because pressure is force exerted over an area (d²), total pressure (P) is:

or, because d³ is volume (V), then:

This gives the useful fact that, at constant temperature, pressure is inversely proportional to volume (Boyle’s law) and directly proportional to the square of molecular velocity.²²

Determination of atmospheric pressure was the first practical application of pressure gauges. Initially, the force exerted by the weight of the atmosphere was balanced against a column of fluid, either water or mercury. This is an example of an absolute pressure gauge, indicating approximately 30 inches (or 760 mm) of mercury, 101 kPa, or 1.01 bars of pressure at sea level. In contrast, relative or “gauge” pressure indicators read zero at atmospheric pressure and indicate only additional pressure sensed.

Sensors

The history of airway pressure measurement parallels the development of sensitive gauges and meters for atmospheric pressure. Beginning with fluid columns, devices for the estimation of gas pressure rapidly evolved into mechanical linkages translating physical motions—of an evacuated capsule, for example—into displacements of an indicating needle. The Bourdon gauge (Fig. 9-1) uses this principle of change in pressure, translated into motion of an evacuated compartment or tube, and is still a robust and inexpensive mechanism. This mechanical gauge, and others similar to it, require no power, are always on, and have high reliability. Although they do not communicate with data systems or alarms, they are useful backups for modern electronic systems.

Newer electronic sensors use the principle of piezoelectricity, the property of certain crystals, usually quartz, to produce a small electric current when compressed. This output may be calibrated and amplified into a usable signal. A metal or semiconductor whose resistive properties change with strain, a so-called piezoresistive effect, also may be used to sense pressure. This additional technologic development has facilitated the evolution and miniaturization of pressure sensors that may be adapted to almost any application. In general, these sensors do require periodic calibration against atmospheric pressure (zero) and occasionally require one other known data point (span) to ensure linearity across their operational range. They are small, lightweight, and generally much more sensitive to small changes.

Applications

Airway pressure is a key component in measuring ventilation. Detecting ventilator-related events (VREs) helps anesthesia providers identify and avoid a majority of serious anesthesia-related events.^23,24 Disconnection of the circuit without occlusion is a significant anesthesia-related event that may lead to adverse patient outcomes.

Pressure sensors have been used to detect other mechanical problems, including tracheal tube disconnection with occlusion, kinking of the inspiratory limb, fresh gas hose kink or disconnection, circuit leaks, high and low scavenging system pressure, continued high circuit pressure, and kinking in the circuit pressure-sensing hose.²⁵ However, the ability to detect these conditions, especially with older monitors, is not as reliable.

Before the introduction of these sensors, previous techniques included lung auscultation, measures of chest wall movement, use of a precordial or esophageal stethoscope, and measurement of physiologic variables such as pulse, blood pressure, respiration, and skin color. These other monitoring methods, absent end-tidal carbon dioxide and pulse oximetry, often resulted in delayed or failed detection of events.

Ventilation failure events can occur because positive pressure does not always equal tidal volume delivered. Sensors in different locations along the breathing circuit, especially with pressure-based ventilator modes, can give disparate pressure measurements and are thus susceptible to both false-negative and false-positive readings.

Breathing Circuit Low-Pressure Alarms

At a minimum, a low-pressure alarm with an audible alert must be placed in the patient breathing circuit in accordance with the current American Society of Anesthesiologists (ASA) Standards for Basic Anesthetic Monitoring. Pressure-sensitive alarms monitor airway or breathing circuit pressure through a side port and compare it with a preset low-pressure alarm limit. The primary purpose of this alarm is the prompt identification of breathing circuit disconnection. This class of malfunction has been disproportionately associated with adverse outcomes in the intraoperative period and may be almost completely avoided when this relatively simple monitor is used. Although low-pressure alarms are primarily intended to warn of a breathing circuit disconnection when a positive-pressure ventilator is used, it is important to realize that they do not detect some partial disconnections and will likely not detect misconnections or obstructions. However, because 70% of all disconnections occur at the Y-piece,²⁶ it is an excellent addition to the other sensors used to monitor ventilation.

Sensor Position Within the Breathing Circuit

The location of airway pressure sensors is important to both detect circuit disconnects and accurately reflect distal airway pressures. Sampling location, kinetic energy transfer, and Bernoulli effects all influence final measured pressure. Circuit pressure should be measured at an appropriate site for the particular need and at a right angle to axial flow.²⁷ The ideal location is at or near the patient Y-piece, but secretions and condensation may cause sensor malfunction. If it is not close enough to the Y-piece, an undetected disconnection (false-negative) may occur as a result of circuit resistance between the measurement site and the distal, now open, end. This will be compounded if a descending bellows ventilator is used, or if the disconnection site is embedded in sheets, blankets, or some other obstructing material.²⁶

A sensor inside the circuit, upstream of the inspiratory valve, eliminates any risk of contamination but becomes insensitive under certain conditions. The combination of an inspiratory limb low-pressure monitor with an in-line humidifier, capnometer cuvette, angled tracheal tube connector, or corrugated catheter mount²⁸ can create backpressure with medium to high flows and may prevent the alarm from being activated. Increased backpressure has also been seen in the Bain (or C-Pram) breathing circuit as a result of circuit tapering and extension of the fresh gas outlet to near the connector at the patient. Lower flows can also prevent the alarm from being activated if moisture builds in the in-line humidifier.²⁹

One proposed solution is to place the low-pressure alarm on the expiratory limb to avoid backpressure-induced false-negative events.³⁰ However, expiratory limb pressure sensors are not without problems. Apnea volume alarms have also been sounded, without the pressure alarm sounding, as a result of loose retaining rings in the expiratory limb not detected during machine leak testing. Interestingly, reverse-flow and minute volume low alarms may trigger without a low-pressure alert.³¹ In older anesthesia delivery systems without fresh gas compensation, expiratory valve anemometers routinely overestimated tidal volumes because of high flows of fresh gas passing through the circuit during the expiratory phase.

Pressure sensors within the ventilator obviously are insensitive in machines with a manual/automatic selector switch in the manual position.²⁶ The placement of pressure sensors in both the circuit and ventilator may provide a solution, but this has been implemented only in the Medical Anesthesia Delivery Unit (ADU) anesthesia machine (GE Healthcare, Waukesha, WI).³² Recently, ventilation pressure sensors have been placed in both the inspiratory and expiratory circuit limbs.^33,34 The inspiratory limb sensor provides feedback during volume-controlled ventilation. Stress fractures in these sensors may trigger alarms without directing the provider to the site of the fault. The small fracture size prevents detection through low- and high-pressure leak tests. Options include detection through visual inspection, using alcohol-wet hands to detect a small gas leak,³⁵ or using a second reservoir bag as a test lung to verify sensor integrity.

To improve the rate of detection of true-positive events, the low-pressure limit should be set just below the normal peak airway pressure, and subsequent machine testing should establish that a disconnection would cause the alarm to sound. The disconnection should be performed at the distal (patient) end to ensure that backpressure, flows, and monitor limits are not such that the alarm would fail to sound in the event of a disconnection.²⁸

The clinical response to a low-pressure alarm should be rapid and organized. The anesthesiologist must systematically evaluate the gas flow to ensure it is adequate, the breathing circuit to ensure it is not disconnected or leaking, and the ventilator, which may not be driving the bellows or may have a setting error. Working outward from the adjustable pressure-limiting (APL) valve toward the patient helps the anesthesiologist rapidly identify and correct the issue.²⁶

Airway Versus Alveolar Pressures

With controlled ventilation, a transpulmonary inflation gradient is provided by positive airway pressure. Generated by a ventilator, this positive pressure gradient may be constant or may vary during the inspiratory phase. The instantaneous pressure during the ventilation cycle is best measured at the distal end of the ventilator circuit, near the connection to the patient.²⁷ Division of the airway pressure-time product (area under the curve) by total cycle time yields mean airway pressure. In most cases,this average pressure approximates mean alveolar pressure.³⁶ Mean airway pressure also correlates with alveolar ventilation, arterial oxygenation, hemodynamic performance (venous return), and risk of barotrauma.²⁷

Alveolar pressure is clinically of interest, but direct measurement generally is not practical. The measurement requires measurement of airway pressure at the distal portion of the endotracheal tube, as close as possible to the alveoli, whereas clinical airway pressure measurements commonly are at the proximal portion of the endotracheal tube.³⁷ Measured airway and actual alveolar pressures differ because of proximal dissipation of the frictional inspiratory airway pressure component and the expiratory alveolar pressure resistive contributions. Positive end-expiratory pressure (PEEP) adds to the external circuit pressure throughout the cycle, proportionally increasing the mean airway pressure.²⁷

Chest wall and lung compliance, secretions, partial occlusions, and tracheal tube resistance may also affect peak airway pressures. Plateau pressures are measured during an inspiratory pause (no flow) and are affected only by secretions, partial occlusions, and tracheal tube resistance.

Ventilation Mode Effects on Airway Pressures

Volume control will deliver a set tidal volume, and pressure will increase until that volume is reached. In low lung compliance conditions, such as acute respiratory distress syndrome (ARDS), high pressures can be administered in attempts to deliver a set volume.

Pressure control will deliver a set pressure, and the tidal volume will depend on airway resistance and lung and chest wall compliance. Pressure control may be better tolerated in some patients, improving both ventilation and oxygenation. However, if lung compliance decreases, and the same pressure is applied, the tidal volume will decrease. Appropriate volume alarm settings will signal low volumes and avoid undetected hypoventilation.

Pressure support mode delivers a positive pressure that is rapidly achieved and maintained throughout inspiration whenever the patient makes an inspiratory effort. The volume delivered depends on the pressure setting, inspiratory time and effort, and airway resistance and compliance.³⁸

Pressure Alarms

Principles and Requirements

In the management of controlled ventilation, it is important to monitor minute ventilation—the volume of gas delivered to the airway in 1 minute—as a measure of ventilatory “adequacy.” This measure is best obtained by measuring tidal (breath) volume and then multiplying by respiratory frequency. Because arterial partial pressure of carbon dioxide is inversely proportional to alveolar minute ventilation,⁴³ this measurement is important both for initial ventilator setup and subsequent adjustments. Measurements may be made intermittently with manual methods; more commonly, they are made with a device integrated into a ventilator circuit and electronically monitored. This integration may include the derivation and display of minute ventilation, flow-volume loops, and determinations of pulmonary compliance.

Dead Space

Because not all tidal ventilation delivered reaches the alveoli, effective gas exchange occurs only in a fractional part of the tidal volume. The remaining gas occupies the trachea, bronchi, and bronchioles and is referred to as anatomic dead space. This volume is relatively fixed in each patient; thus, as tidal volume decreases, it occupies a larger fraction of each breath. At a tidal volume roughly equal to the anatomic dead space, effective alveolar ventilation ceases, with the same gas being transported distally from bronchi to alveoli that was most recently expelled. Also, the distal breathing circuit and patient airway contribute additional mechanical or apparatus dead space to the system (Fig. 9-2).

Measurement of dead space is important to minimize its contribution to total minute ventilation and clearly calculate the relationship between alveolar minute ventilation and arterial carbon dioxide tension. A combination of precise volume measurement and capnography may be used to calculate anatomic dead space using Fowler’s method. Initially described using nitrogen as the reference gas, the expired volume was measured from the beginning of exhalation until the onset of the “plateau” of alveolar gas.⁴³

Mechanical Devices

The simplest device for the measurement of gas volume is based on a rotating vane or propeller calibrated against a specific density of gas. The total rotation of the attached shaft correlates with the volume of gas (air) that has passed by. The basic principle is that a force is transmitted to the vanes by the impact of gas molecules, and this force is converted to a rotational (angular) momentum and spins the pinwheel. Optical or mechanical transducers count the rotations and convert the value to an equivalent volume. Critical variables are the gas density—which depends on gas composition, humidity, and altitude—and temperature. At very low flows or volumes, the device may be less accurate as a result of the finite mass and inertia of the vanes.

Principles

The behavior of fluids and gases depends on whether their flow is laminar (smooth) or turbulent. Flow through a tube with a length greater than the diameter usually is laminar; however, at a critical velocity (V_C), a transition to turbulent flow occurs. This is shown in the following equation, in which r is the radius of the tube, ⍴ is the density, η is the viscosity of the gas, and k is the critical Reynolds number, which is different for each gas:

This transition to turbulent flow greatly increases the resistance to flow and may be seen in many common scenarios, such as a kink occurring in an otherwise smooth tracheal tube. For flow through an orifice, in which length is less than the radius, turbulent flow is a given. Measured in units of length cubed per unit of time, volume flow may be calculated by detecting a velocity change across a fixed cross-sectional area. Bernoulli’s equation makes this possible by measuring a pressure change across this transition:

where P is pressure, g is gravitational acceleration, and h is the height above a reference plane. Note that with turbulent flow, only the gas density affects the relationship between pressure and velocity. Application of a continuity condition, in which gas in equals gas out, gives the following:

In this case, P₁ and V₁ represent initial pressure and velocity, and P₂ and V₂ represent final values. Rearrangement of this equation allows determination of gas velocity by measurement of pressure change (drop); this is converted into flow by knowing the cross-sectional area (A) of the passage:

Given that flow is proportional to velocity, from the continuity equation we know that flow is also proportional to the square root of the pressure difference (P) across an orifice.²²

However, for laminar flow, where r and L are the radius and length of the tube, the Hagen-Poiseuille law offers this relationship between flow (F) and pressure drop (P₁ − P₂):

This equation allows the laminar flow of a gas to be calculated by determining the pressure drop across a resistance element, not an orifice. Here, in contrast to Bernoulli’s equation, gas viscosity is the factor that may contribute to miscalibration or erroneous readings. Also, flow is directly proportional to the pressure difference, rather than to the square root, as with turbulent flow.

Flowmeters

The concept of the flowmeter encompasses two major areas: the metering of gas supplies, usually to a delivery system, and the measurement of tidal (periodic) ventilation in breathing circuits. The use of relatively constant flowmeters requires precise and reproducible settings to control, rather than measure, gas flows. The primary requirement for flowmeters in human breathing is accurate measurement of nonsteady flows.⁴⁹

Acoustic/Ultrasonic Flow Sensors

A more sensitive flowmeter used for monitoring of variable flows uses ultrasound reflection from moving columns of gas or liquid. Similar to the familiar ultrasound applications to tissue and blood flow imaging, the intrinsic accuracy and compact size of these sensors makes them ideal in the operating room (OR) environment. Initial work on measuring human breathing with ultrasonic flowmeters appeared in the 1980s and was rapidly adopted for measuring many types of ventilation and human populations in and out of the OR.^66,67

Acoustic/ultrasonic flow sensors use clamp-on or wet transducers and can be applicable to liquids, gases, and multiphase mixtures. Their use has limits, but application to monitoring ventilation can be accomplished with the right combination of technology (Box 9-2).⁶⁸

Within the acoustic domain, various techniques of measurement may be used. Single versus multiple acoustic paths, vortex shedding, and passive and active principles, to name a few, are available, although no single “best” method exists.

Several ultrasonic flowmeters are in current use, including the BRDL flowmeter (Birmingham Research and Development Ltd., Birmingham, UK) and the Spiroson flowmeter (ECO Medics AE, Dürnten, Switzerland). The former has two PVC flowmeters with diagonally opposed flow transducers, and the latter has a single flowhead with two rectangular flow channels, again with diagonally opposed flow transducers. The advantage of these flowmeters is minimal airflow obstruction over a large flow range compared with the traditional Fleisch pneumotachometers (Fig. 9-10).^48,69

The ultrasonic flow sensor is used on certain models of Draeger anesthesia workstations, such as the Narkomed 6400. Ultrasonic flow measurement uses the transit time principle, by which opposite sending and receiving transducers are used to transmit signals through the gas flow. The signal travels faster when moving with the flow stream rather than against it, and the difference between the two transit times is used to calculate the gas flow rate. The device is sensitive to gas flow direction and has no moving parts, and accuracy is independent of gas flow composition.

Spirometry

Curves

Curves are a useful depiction of two parameters in a visual format. Traditionally, ventilatory curves use time along the x-axis, and the y-axis can be volume, flow, or pressure.

Loops

Apnea and Disconnect Monitoring

As a circuit disconnect alarm, continuous end-tidal carbon dioxide monitoring is considered the most sensitive method to detect a disconnection or cessation of ventilation. It is slower to alarm than the pressure alarms but is an excellent redundant monitor for such a purpose. During general anesthetics and moderate sedation, it alerts faster than other vital sign monitors—such as pulse oximetry and other, later indicators of extubation and obstruction—compared with capnography.

Capnography frequently is used by personnel other than anesthesiologists for monitoring during moderate sedation. Studies conducted in locations such as the endoscopy suite and the emergency department have shown that capnography identifies hypoventilation and apnea that go undetected by care providers before the onset of hypoxia.^76,77

Continuous Carbon Dioxide Monitoring

Volumetric Capnography and Oxigraphy

With the ability to measure CO₂ concentration and minute ventilation in near real time comes the ability to calculate CO₂ production (VCO₂). The measurement of inspired and expired oxygen concentrations also allows for volumetric oxigraphy and oxygen consumption (VO₂) determination. Dividing the two (VCO₂/VO₂) yields the respiratory exchange ratio (see Chapter 8).

Ventilator-Integrated Displays

Most modern anesthesia and critical care ventilators provide, at the very least, a numeric display of commonly used variables such as rate, tidal volume, and inspiratory/expiratory (I:E) ratio. Many also provide a display of airway pressure over time, perhaps also depicting the location of high- and low-pressure alarm limits. These are stand-alone displays, in that they do not depend on other monitors or external devices for their function.

Monitor-Integrated Displays

The next step up in display sophistication integrates the digital output from a ventilator into a combined display, usually in combination with hemodynamic data. Physiologic monitors with modular construction, such as those from Philips Respironics (Wallingford, CT) and Hewlett-Packard (Dayton, OH), allow the insertion of a ventilator monitoring module and the display of volumes and pressures. In addition, they offer the option of visual display of pressure versus time, flow versus time, or other curves. These modular links are specific to the physiologic monitor and normally accept data from the most common anesthesia ventilators (Fig. 9-21).

Data Link from Ventilators

Even today, the output from ventilators, ventilator monitors, and respiratory mechanics sensors proceeds at a relatively low data rate. The most common interfaces are simple serial outputs that typically run at 19,200 Baud or less.⁷⁸ This common interface makes data transmission and interface design relatively straightforward and simple. Experiments with Ethernet connectivity have been done, but the future seems to lie with wireless data transmission.⁷⁹ The challenges involve data security, verification, and accurate patient identification and tracking as both patients and hardware become more mobile.

1. Hunter J. Proposals for the recovery of people apparently drowned. Philosophical Transactions of the Royal Society of London. 1776;66:412–425.

2. Guedel A.R. Inhalation anesthesia: a fundamental guide. New York: Macmillan; 1937.

3. Baker S.P., Hitchcock F.A. Immediate effects of inhalation of 100% oxygen at one atmosphere on ventilation volume, carbon dioxide output, oxygen consumption and respiratory rate in man. J Appl Physiol. 1957;10(3):363–366.

4. Kory R.C. Routine measurement of respiratory rate: an expensive tribute to tradition. JAMA. 1957;165(5):448–450.

5. Krieger B., Feinerman D., Zaron A., Bizousky F. Continuous noninvasive monitoring of respiratory rate in critically ill patients. Chest. 1986;90(5):632–634.

6. Bowdle T.A. Adverse effects of opioid agonists and agonist-antagonists in anaesthesia. Drug Safety. 1998;19(3):173–189.

7. Ibarra E., Lees D. Mass spectrometer monitoring of patients with regional anesthesia. Anesthesiology. 1985;63:572–573.

8. Fukuda K., Ichinohe T., Kaneko Y. Is measurement of end-tidal CO₂ through a nasal cannula reliable? Anesthesia Progress. 1997;44(1):23–26.

9. Ashutosh K., Gilbert R., Auchincloss J.H., Erlebacher J., Peppi D. Impedance pneumograph and magnetometer methods for monitoring tidal volume. J Appl Physiol. 1974;37(6):964–966.

10. Houtveen J.H., Groot P.F., de Geus E.J. Validation of the thoracic impedance-derived respiratory signal using multilevel analysis. Int J Psychophysiol. 2006;59(2):97–106.

11. Wilkinson J.N., Thanawala V.U. Thoracic impedance monitoring of respiratory rate during sedation: is it safe? Anaesthesia. 2009;64(4):455–456.

12. Freundlich J.J., Erickson J.C. Electrical impedance pneumography for simple nonrestrictive continuous monitoring of respiratory rate, rhythm and tidal volume for surgical patients. Chest. 1974;65(2):181–184.

13. Khambete N., Metherall P., Brown B., Smallwood R., Hose R. Can we optimize electrode placement for impedance pneumography? Ann N Y Acad Sci. 1999;873:534–542.

14. Network E.U.E. Thoracic impedance measurements can interfere with impedance-based rate-responsive pacemakers. Health Devices. 1997;26:393–394.

15. Tobin M.J., Jenouri G., Lind B., et al. Validation of respiratory inductive plethysmography in patients with pulmonary disease. Chest. 1983;83(4):615–620.

16. Brouillette R.T., Morrow A.S., Weese-Mayer D.E., Hunt C.E. Comparison of respiratory inductive plethysmography and thoracic impedance for apnea monitoring. J Pediatr. 1987;111(3):377–383.

17. Moody G.B., Mark R.G., Zoccola A., Mantero S. Derivation of respiratory signals from multi-lead ECGs. Comput Cardiol. 1985;12:113–116.

18. Hok B., Wiklund L., Henneberg S. A new respiratory rate monitor: development and initial clinical experience. Int J Clin Monit Comput. 1993;10(2):101–107.

19. Dodds D., Purdy J., Moulton C. The PEP transducer: a new way of measuring respiratory rate in the non-intubated patient. J Accid Emerg Med. 1999;16:26–28.

20. Zhu X., Chen W., Nemoto T., et al. Real-time monitoring of respiration rhythm and pulse rate during sleep. IEEE Trans Biomed Eng. 2006;53(12 Pt 1):2553–2563.

21. Macknet M.R., et al. Accuracy and tolerance of a novel bioacoustic respiratory sensor in pediatric patients. Available at www.rcjournal.com/abstracts/2007/?id=aarc07_199.

22. Macintosh R. Physics for the anaesthetist, ed 4. Boston: Blackwell Scientific; 1987.

23. Abstracts of the 1998 Joint Meeting of the Society for Technology in Anesthesia and the Rochester Simulator Symposium: Simulation in Anesthesia. J Clin Monit Comput. 1998;14(7):511–539.

24. Winter A., Spence A.A. An international consensus on monitoring? Br J Anaesth. 1990;64(3):263–266.

25. McEwen J., Small C., Jenkins L. Detection of interruptions in the breathing gas of ventilated anaesthetized patients. Can J Anaesth. 1988;35(6):549–561.

26. Raphael D.T., Weller R.S., Doran D.J. A response algorithm for the low-pressure alarm condition. Anesth Analg. 1988;67(9):876–883.

27. Marini J.J., Ravenscraft S.A. Mean airway pressure: physiologic determinants and clinical importance. Part 2: Clinical implications. Crit Care Med. 1992;20(11):1604–1616.

28. Campbell R.M., Sheikh A., Crosse M.M. A study of the incorrect use of ventilator disconnection alarms. Anaesthesia. 1996;51(4):369–370.

29. Milligan K.A. Disablement of a ventilator disconnect alarm by a heat and moisture exchanger. Anaesthesia. 1992;47(3):279.

30. Slee T.A., Pavlin E.G. Failure of low pressure alarm associated with the use of a humidifier. Anesthesiology. 1988;69(5):791–792.

31. Chung D.C., Ho A.M.H., Tay B.A. “Apnea-volume” warning during normal ventilation of the lungs: an unusual leak in the Narkomed 4 anesthesia system. J Clin Anesth. 2001;13(1):40–43.

32. Snyders S., Mitton M. Failure of a flow sensor of a Datex Ohmeda S/5 Aespire. Anaesthesia. 2005;60(9):941–942.

33. Hari M., Jennings M., Mitton M. Flow sensor fault causing ventilator malfunction. Anaesthesia. 2005;60(10):1049–1051.

34. Aldridge J. Leak on Datex Aestiva/5 anaesthetic machine. Anaesthesia. 2005;60(4):420–421.

35. Dhar P., George I., Mankad A., Sloan P. Flow transducer gas leak detected after induction. Anesth Analg. 1999;89(6):1587.

36. Bowes W.A., 3rd., Corke B.C., Hulka J. Pulse oximetry: a review of the theory, accuracy, and clinical applications. Obstet Gynecol. 1989;74(3 Pt 2):541–546.

37. Bardoczky G.I., Engelman E.E. Comparison of airway pressures from different measuring sites. Anesth Analg. 1996;83(4):887.

38. Lian J.X. Understanding ventilator waveforms and how to use them in patient care. Nurs Crit Care. 2009;4(1):43–55.

39. Spielman F.J., Sprague D.H. Another benefit of the subatmospheric alarm. Anesthesiology. 1981;54(6):526.

40. Walker T. Another problem with a circle system. Anaesthesia. 1996;51(1):89.

41. Bashein G.M.D., MacEvoy B. Anesthesia ventilators should have adjustable high-pressure alarms. Anesthesiology. 1985;63(2):231–233.

42. Campos C., Naguib S.S., Chuang A.Z., Lemak N.A., Khalil S.N. Endobronchial intubation causes an immediate increase in peak inflation pressure in pediatric patients. Anesth Analg. 1999;88(2):268–270.

43. West J. Respiratory physiology: the essentials, ed 8. Baltimore: Lippincott, Williams & Wilkins; 2008. pp 17–19

44. Wright B.M. A respiratory anemometer. J Physiol (London). 1955;127:25.

45. Bushman J. Effect of different flow patterns on the Wright respirometer. Br J Anaesth. 1979;51(9):895–898.

46. Ilsley A., Hart J., Withers R., Roberts J. Evaluation of five small turbine-type respirometers used in adult anesthesia. J Clin Monit Comput. 1993;9(3):196–201.

47. Hatch D.J., Jackson E.A. A new critical incident monitor for use with the paediatric T-piece. Anaesthesia. 1996;51(9):839–842.

48. Goto T., Saito H., Nakata Y., et al. Effects of xenon on the performance of various respiratory flowmeters. Anesthesiology. 1999;90(2):555–563.

49. Lynnworth L.C., Korba J.M., Wallace D.R. Fast response ultrasonic flowmeter measures breathing dynamics. IEEE Trans Biomed Eng. 1985;32(7):530–535.

50. Kuppers K.. An improved apparatus for measuring gasesGermany. 1909. Deutsches patent and Markenamt

51. Ball C., Westhorpe R. Maximilian Neu and the first anaesthetic rotameter. Anaesth Intensive Care. 1999;27:333.

52. Foregger R. Early use of rotameter in anaesthesia. Br J Anaesth. 1952;24(3):187–195.

53. Moore J.K., Railton R. Hypoxia caused by a leaking rotameter: the value of an oxygen analyser. Anaesthesia. 1984;39(4):380–381.

54. Gupta B.L., Varshneya A.K. Anaesthetic accident caused by unusual leakage of rotameter [letter]. Br J Anaesth. 1975;47(7):805.

55. McQuillan P.J., Jackson I.J. Potential leaks from anaesthetic machines: potential leaks through open rotameter valves and empty cylinder yokes. Anaesthesia. 1987;42(12):1308–1312.

56. Slater E.M. Transposition of rotameter bobbins. Anesthesiology. 1974;41(1):101.

57. Cooke R. Transposition of rotameter tubes. Br J Anaesth. 1998;80(5):699.

58. Fleisch A. The pneumotachometer [in German]. Arch Ges Physiol. 1925;209:713–722.

59. Yeh M.P., Adams T.D., Gardner R.M., Yanowitz F.G. Effect of O₂, N₂, and CO₂ composition on nonlinearity of Fleisch pneumotachograph characteristics. J Appl Physiol. 1984;56(5):1423–1425.

60. Meriläinen P., Hänninen H., Tuomaala L. A novel sensor for routine continuous spirometry of intubated patients. J Clin Monit Comput. 1993;9(5):374–380.

61. Bardoczky G.I., deFrancquen P., Engelman E., Capello M. Continuous monitoring of pulmonary mechanics with the sidestream spirometer during lung transplantation. J Cardiothorac Vasc Anesth. 1992;6(6):731–734.

62. Søndergaard S., Kárason S., Lundin S., Stenqvist O. Evaluation of a pitot type spirometer in helium/oxygen mixtures. J Clin Monit Comput. 1998;14(6):425–431.

63. Tham R., Oberle M. How do flow sensors work? APSF Newsletter. 2008:10–12.

64. Kann T., Hald A., Jørgensen F.E. A new transducer for respiratory monitoring: a description of a hot-wire anemometer and a test procedure for general use. Acta Anaesth Scand. 1979;23(4):349–358.

65. Kiełbasa J. Measurement of gas flow velocity: anemometer with a vibrating hot wire. Rev Sci Instrum. 81(1), 2010. 015101

66. Brusasco V., Beck K.C., Crawford M., Rehder K. Resonant amplification of delivered volume during high-frequency ventilation. J Appl Physiol. 1986;60(3):885–892.

67. Eberhart R.C., Weigelt J.A. Respiratory monitoring: current techniques and some new developments. Bull Eur Physiopathol Respir. 1985;21(3):295–300.

68. Lynnworth L.C., Liu Y. Ultrasonic flowmeters: half-century progress report, 1955-2005. Ultrasonics. 2006;44(Suppl 1):e1371–e1378.

69. Kästner S.B., Marlin D.J., Roberts C.A., Auer J.A., Lekeux P. Comparison of the performance of linear resistance and ultrasonic pneumotachometers at rest and during lobeline-induced hyperpnoea. Res Vet Sci. 2000;68(2):153–159.

70. Pierce R. Spirometry: an essential clinical measurement. Aust Fam Physician. 2005;34(7):535–539.

71. Bardoczky G.I., Engelman E., D’Hollander A. Continuous spirometry: an aid to monitoring ventilation during operation. Br J Anaesth. 1993;71(5):747–751.

72. Weissman C. Flow-volume relationships during spontaneous breathing through endotracheal tubes. Crit Care Med. 1992;20:615–620.

73. Lumb A.B. Elastic forces and lung volumes. Nunn’s applied respiratory physiology. Woburn, MA: Butterworth Heinemann; 2000. p 47

74. Cheifetz I.M., Myers T.R. Should every mechanically ventilated patient be monitored with capnography from intubation to extubation? Respir Care. 2007;52(4):423–438.

75. Eberhard P. The design, use, and results of transcutaneous carbon dioxide analysis: current and future directions. Anesth Analg. 2007;105:S48–S52.

76. Lightdale J.R., Goldmann D.A., Feldman H.A., et al. Microstream capnography improves patient monitoring during moderate sedation: a randomized, controlled trial. Pediatrics. 2006;117(6):e1170–e1178.

77. Deitch K., Chudnofsky C.R., Dominici P. The utility of supplemental oxygen during emergency department procedural sedation with propofol: a randomized, controlled trial. Ann Emerg Med. 2008;52(1):1–8.

78. Respironics Flo-Trak module. Accessed 2012 Oct 5 at http://oem.respironics.com/Downloads/4100306-FloTrak.pdf.

79. Cisco Systems: 802.11ac: The Fifth Generation of Wi-Fi. Available at www.cisco.com/en/US/prod/collateral/wireless/ps5678/ps11983/white_paper_c11-713103_ns767_Networking_Solutions_White_Paper.html. Accessed 4 Nov 2012