Method and Apparatus for Detection of Venous Air Emboli

This invention relates generally to the detection of venous air emboli and more specifically to an apparatus and method for the real time detection of venous air emboli which is based on the monitoring of a patient's inspired and expired respiratory gases.

The occurrence of Venous Air Emboli (VAE) has been well documented in the medical literature. VAE has been shown to be a significant cause of mortality and morbidity in many different types of surgical procedures.

Venous Air Emboli occur when air bubbles enter the blood stream secondary to a negative pressure gradient. Negative pressure gradients can occur any time the surgical site is above the level of the right atrium. When air enters the venous system, small bubbles are formed. These bubbles travel to the right side of the heart, enter the right atrium, pass through the tricuspid valve into the right ventricle, through the pulmonic valve and into the pulmonary artery. As the bubbles traverse the pulmonary capillary tree, they become wedged in the distal pulmonary micro-vascular system. This results in increased pulmonary deadspace.

When gone undetected for sustained periods of time, it is possible for some bubbles to completely traverse the pulmonary capillary tree. These bubbles enter the left side of the heart and ultimately the arterial blood flow. Once inside the arterial circulation, the potential arises for severe intra-operative and/or post operative cardiopulmonary and neurological complications.

Precordial doppler has been shown to be a very specific and accurate detector of VAE. However, there are many problems associated with doppler monitoring. Positioning of the doppler probe over the correct intercostal space, electro-surgical interference, heart murmurs and the requirement of a trained ear make doppler monitoring less than adequate for VAE detection.

Continuous CO.sub.2 monitoring is a sensitive indicator of VAE. However, CO.sub.2 monitoring is very non-specific for VAE. There are many different physiological alterations that lower CO.sub.2 values. Metabolic changes, cardiac output changes, hyperventilation and deadspace changes can lower expired CO.sub.2 values.

Many studies have been performed to analyze conditions and parameters indicative of VAE. Several of these studies are summarized below.

Matjasko et al. (Matjasko, Petrozza and Mackenzie, Sensitivity of End-tidal Nitrogen in Venous Air Embolism Detection in Dogs. Anesthesiology 63:418-423, 1985) describes a study done to compare the sensitivity for the detection of venous air emboli of end-tidal nitrogen (EtN.sub.2) monitoring, precordial Doppler (PD), end-tidal CO.sub.2 (EtCO.sub.2) and pulmonary artery pressure (PAP). This study concluded that: 1) precordial Doppler monitoring is the most sensitive qualitative detector of air entry into the superior vena cava and heart, however, it does not allow quantitative measurement of air entry into and dissipation from the lungs; and 2) changes in EtN.sub.2 following low-dose infusion VAE are less sensitive than changes in EtCO.sub.2, while during bolus VAE they are equally sensitive. Additionally, they expressed concern that shared operating room mass spectrometers would be sensitive enough and sample frequently enough to be of value in the early detection of clinical VAE.

Matjasko et al. (Matjasko, Hellman and Mackenzie, Venous Air Embolism, Hypotension, and End-tidal Nitrogen. Neurosurgery 21:378-382, 1987) describes a study done to compare the sensitivity for the detection of venous air emboli of end-tidal nitrogen (EtN.sub.2) monitoring and end-tidal CO.sub.2 (EtCO.sub.2) monitoring during large bolus and infusion VAE.

Mangan et al. (Mangan, Boucek, Powers and Shadduck, Rapid Detection of Venous Air Embolism by Mass Spectrometry during Bone Marrow Harvesting. Exp. Hematol. 13:639-640, 1985) describes the detection of a venous air embolism by mass spectrometric monitoring of the patient's expired gases. They found that air emboli are suspected when end-tidal CO.sub.2 decreases (down 1.5%) and also when expired nitrogen increases (up 4%).

Russell et al. (Russell, Richard and Snider, Detection of Venous Air Embolism in Dogs by Emission Spectrometry. J Clin Monit. 6:18-23, 1990) describes the detection of a venous air embolism by emission spectrometry monitoring of the end-tidal nitrogen and compares the results with monitoring by mass spectrometry.

Prager et al. (Prager, Gregory, Ascher and Roberts, Massive Venous Air Embolism during Orthotopic Liver Transplantation. Anesthesiology 72:198-200, 1990) reports the detection of a venous air embolism by mass spectrometric monitoring of the patient's end-tidal CO.sub.2 and end-tidal expired nitrogen.

Frankel et al. (Frankel and Holzman, Air Embolism During Posterior Spinal Fusion. Can J Anaesth 35:5:511-514, 1988) presents a list of methods available for the detection of VAE including: precordial doppler, oesophageal stethoscope, measurements of end-tidal CO.sub.2, end-tidal nitrogen, pulmonary artery pressure, transoesophageal echocardiography and transcutaneous O.sub.2 and CO.sub.2 monitoring.

Lew et al. (Lew, Tay and Thomas, Venous Air Embolism During Cesarean Section More Common Than Previously Thought. Anesth Analg 77:448-452, 1993) reports the use of a Raman scattering analyzer to continuously monitor inspired and expired nitrogen, oxygen, halothane, nitrous oxide, and end-tidal carbon dioxide levels via a side stream sampling catheter. Positive evidence of air embolism was defined as an increase in expired nitrogen concentrations of 0.1 vol % from the baseline.

Drummond et al. (Drummond, Prutow and Scheller, A Comparison of the Sensitivity of Pulmonary Artery Pressure, End-Tidal Carbon Dioxide, and End-Tidal Nitrogen in the Detection of Venous Air Embolism in the Dog. Anesth Analg 64:688-692, 1985) reports the results of an experiment seeking to define the relative sensitivities of end-tidal CO.sub.2 analysis, end-tidal nitrogen analysis and pulmonary artery pressure (PAP) monitoring in the detection of VAE. They conclude that, where the capacity to identify increases in expired nitrogen on the order of 0.04% can be achieved, end-tidal nitrogen monitoring will identify VAE events with a sensitivity similar to that of PAP and EtCO.sub.2. They also note that the difficulties inherent in achieving this level of nitrogen sensitivity represents a major limitation in the application of this test.

English et al. (English, Westenskow, Hodges and Stanley, Comparison of Venous Air Embolism Monitoring Methods in Supine Dogs. Anesthesiology 48:425-429, 1978) compares a variety of VAE monitoring methods for the detection of VAE including: precordial Doppler ultrasound frequency; end-tidal CO.sub.2 concentration; mean pulmonary arterial, central venous and arterial pressures; esophageal heart sounds, including high-pitched barely audible tinkling sounds, enhanced second heart sound, systolic and mill-wheel murmurs; the electrocardiogram; arterial blood carbon dioxide and oxygen tensions; heart rate and the presence of cardiac arrhythmias.

None of these studies or the methods described therein disclose a monitoring system that automatically detects and signals the presence of a VAE. For example, the precordial Doppler techniques requires that a trained individual continuously monitor the output of the instrument and interpret the output. Likewise, while it is generally recognized that an increase in the end-tidal nitrogen concentration may be indicative of a VAE, or that a decrease in the end-tidal CO.sub.2 concentration may be indicative of a VAE, neither is considered to be a reliable indicator. Additionally, as with the precordial Doppler, monitoring of end-tidal nitrogen and/or end-tidal CO.sub.2 concentrations requires the attention of an operator to observe the behavior of these parameters and determine when changes therein are significant and whether they are indicative of a VAE or some other factor.

Thus, the ideal monitor for detection of venous air embolism should be: noninvasive; exquisitely sensitive; reliable; automatic; specific for air-induced changes; quantitatively reflect the size of the embolus; and have a duration of positive response accurately reflecting embolus onset and resolution. No single VAE monitor presently available meets all of these requirements.

The present invention is a venous air embolism detector which continuously monitors the respiratory gases of a patient and automatically signals the presence of a VAE. Continuous monitoring of the VAE detector of the present invention by an individual or subjective interpretations of its output are not required.

In one embodiment, the present invention measures and creates an historical average of the end-tidal CO.sub.2 and end-tidal nitrogen concentrations for a predetermined period of time. At the end of that predetermined period of time, the current value of the end-tidal nitrogen concentration is compared to the previously acquired end-tidal nitrogen concentration historical average. If the current value exceeds the highest value of the historical average during the predetermined period of time by at least 0.03 vol %, a nitrogen "TRUE" value is set. Similarly, the current value of the end-tidal CO.sub.2 concentration is compared to the previously acquired end-tidal CO.sub.2 concentration historical average. If the current value is lower than the lowest value of the historical average during the predetermined period of time by at least 1.5 mmHg, a carbon dioxide "TRUE" value is set. When both the nitrogen and carbon dioxide values are "TRUE", the detector signals the presence of a VAE. Thus, the concomitant measurements of end-tidal CO.sub.2 and end-tidal nitrogen concentrations are used to detect a VAE.

In another embodiment, an additional measurement based on the oxygen concentration is included in the test. During the same predetermined time period that the CO.sub.2 and N.sub.2 are being measured, the difference in the patient's inspired oxygen (FiO.sub.2) and expired oxygen (FeO.sub.2) concentrations are measured and averaged. If the current value of (FiO.sub.2 -FeO.sub.2) is lower than the highest value of an historical average during the predetermined period of time by at least 0.02 vol %, an oxygen "TRUE" value is set. When the nitrogen, carbon dioxide and oxygen values are all "TRUE", the detector signals the presence of a VAE. Thus, the concomitant measurements of EtCO.sub.2, EtN.sub.2 and (FiO.sub.2 -FeO.sub.2) concentrations are used to detect a VAE.

In a first embodiment, the present invention is an apparatus for detecting an air embolism comprising: a detector for detecting nitrogen in respiratory gases and determining the concentration of nitrogen in the respiratory gases; and a processor for analyzing the nitrogen concentration comprising: an historical data base of previously acquired nitrogen concentrations; a current data register which contains the current nitrogen concentration; and a comparator which compares the current nitrogen concentration with the historical data base and triggers an alarm signal if a predetermined comparison criteria is satisfied. The detector may be a Raman gas analyzer. Additionally, the processor may further comprise a peak detector which searches the historical data base of previously acquired nitrogen concentrations and determines the maximum value therein; and the comparator may further comprise a difference unit which determines the difference between the maximum historical nitrogen concentration and the value held in the current data register and signals an alarm if the difference exceeds a predetermined difference criteria. In some embodiments, the processor for analyzing the nitrogen concentration determines the end-tidal nitrogen concentration, FeN.sub.2. In yet other embodiments, the comparator may further comprise an averager which determines an average derived from the data contained in the historical data base.

In a second embodiment, the invention is an apparatus for detecting an air embolism comprising: a detector for detecting carbon dioxide and nitrogen in respiratory gases and determining the concentrations of carbon dioxide and nitrogen in the respiratory gases; and a processor for comparing the carbon dioxide and nitrogen concentrations with a predetermined carbon dioxide concentration threshold and a predetermined nitrogen concentration threshold, respectively, the processor having an output which signals the presence of an air embolism when the carbon dioxide and nitrogen concentrations both cross the predetermined carbon dioxide concentration threshold and the predetermined nitrogen concentration threshold, respectively, within a predetermined time window. The detector may be a Raman gas analyzer. In some embodiments, the detector may further comprise means for detecting oxygen and the processor may further comprise a predetermined oxygen concentration threshold and signals the presence of an air embolism when the carbon dioxide, nitrogen and oxygen concentrations cross the predetermined carbon dioxide concentration threshold, the predetermined nitrogen concentration threshold and the predetermined oxygen concentration threshold, respectively, within a predetermined period of time.

In a third embodiment, the invention is for an apparatus for detecting an air embolism comprising: a detector for detecting nitrogen and carbon dioxide in respiratory gases; a gas analyzer for determining the breath-by-breath inspired and expired concentrations of the nitrogen and carbon dioxide in the respiratory gases; an end-tidal carbon dioxide analyzer which receives the measured breath-by-breath carbon dioxide concentrations and determines the value of the end-tidal carbon dioxide concentration, EtCO.sub.2, therefrom for a plurality of breaths; an end-tidal carbon dioxide data memory for storing the plurality of breath-by-breath EtCO.sub.2 values as a function of time; a nitrogen analyzer which receives the measured breath-by-breath nitrogen concentrations and determines the value of the expired nitrogen concentration, FeN.sub.2, therefrom for a plurality of breaths; an expired nitrogen data memory for storing the plurality of breath-by-breath FeN.sub.2 values as a function of time; and a processor which 1) compares the most recent value of EtCO.sub.2 with previously acquired values of EtCO.sub.2 and sets an EtCO.sub.2 TRUE flag if a predetermined EtCO.sub.2 comparison criteria is satisfied; 2) compares the corresponding most recent value of FeN.sub.2 with previously acquired values of FeN.sub.2 and sets an FeN.sub.2 TRUE flag if a predetermined FeN.sub.2 comparison criteria is satisfied; and 3) signals an alarm if the EtCO.sub.2 TRUE flag and FeN.sub.2 TRUE flag both occur for the same current breath. In some versions of this embodiment, the detector may further comprise an oxygen detector which detects oxygen in respiratory gases; the gas analyzer may further comprise an analyzer which determines the breath-by-breath inspired and expired concentrations of the oxygen in the respiratory gases; an oxygen analyzer which receives the measured breath-by-breath oxygen concentrations and determines the value of the inspired oxygen minus the expired oxygen concentration, FiO.sub.2 -FeO.sub.2, therefrom for a plurality of breaths; an expired oxygen data memory for storing the plurality of breath-by-breath FiO.sub.2 -FeO.sub.2 values as a function of time; and the processor may further comprise a processor which 4) compares the most recent value of FiO.sub.2 -FeO.sub.2 with previously acquired values of FiO.sub.2 -FeO.sub.2 and sets an FiO.sub.2 -FeO.sub.2 TRUE flag if a predetermined FiO.sub.2 -FeO.sub.2 comparison criteria is satisfied; and 5) signals an alarm if the EtCO.sub.2 TRUE flag, the FeN.sub.2 TRUE flag and the FiO.sub.2 -FeO.sub.2 TRUE flag all occur for the same current breath. In some versions of this embodiment, the detector is a Raman gas analyzer. This embodiment may further comprise: a maximum value detector which extracts the highest value from the plurality of breath-by-breath FeN.sub.2 values as a function of time which are stored in the expired nitrogen data memory; and a comparator which compares the most recent value of the expired nitrogen concentration FeN.sub.2 with the maximum value and triggers an alarm when the most recent value exceeds the maximum value by at least 0.01 vol %. In other versions, this embodiment may further comprise: a minimum value detector which extracts the lowest value from the plurality of breath-by-breath EtCO.sub.2 values as a function of time which are stored in the end-tidal carbon dioxide data memory; and a comparator which compares the most recent value of the end-tidal carbon dioxide concentration EtCO.sub.2 with the lowest value and triggers an alarm when the most recent value is lower than the lowest value by at least 1.0 mmHg.

A fourth embodiment is for a method of automatically detecting an air embolism comprising the steps of: detecting nitrogen in respiratory gases; determining the breath-by-breath expired nitrogen concentration for a plurality of breaths and storing the plurality of breath-by-breath expired nitrogen concentrations in an historical data base; comparing the most recent value of the expired nitrogen concentration with previously acquired expired nitrogen concentrations in the historical data base; and signalling the presence of an air embolism when the most recent value of the expired nitrogen concentration crosses a baseline threshold derived from the previously acquired historical data base of expired nitrogen concentrations. This embodiment may further comprise the steps of: detecting carbon dioxide in respiratory gases; determining the breath-by-breath end-tidal carbon dioxide concentration for a plurality of breaths and storing the plurality of breath-by-breath end-tidal carbon dioxide concentrations in an historical data base; comparing the most recent value of the end-tidal carbon dioxide concentration with previously acquired end-tidal carbon dioxide concentrations in the historical data base; and signalling the presence of an air embolism when the most recent value of the end-tidal carbon dioxide concentration crosses a baseline threshold derived from the previously acquired historical data base of end-tidal carbon dioxide concentrations and the corresponding most recent value of the expired nitrogen concentration crosses a baseline threshold derived from the previously acquired historical data base of expired nitrogen concentrations. In another version, the step of signalling the presence of an air embolism further comprises the steps of: defining the nitrogen baseline threshold to be the highest expired nitrogen concentration in the nitrogen historical data base; and triggering the alarm when the most recent value of the expired nitrogen concentration exceeds the baseline threshold by at least 0.01 vol %. In yet another version of the method, the step of signalling the presence of an air embolism further comprises the steps of: defining the nitrogen baseline threshold to be the highest expired nitrogen concentration in the nitrogen historical data base; setting a nitrogen trigger when the most recent value of the expired nitrogen concentration exceeds the nitrogen baseline threshold by at least 0.01 vol %. In another version of the method, the step of signalling the presence of an air embolism further comprises the steps of: defining the end-tidal carbon dioxide baseline threshold to be the lowest end-tidal carbon dioxide concentration in the end-tidal carbon dioxide historical data base; setting an end-tidal carbon dioxide trigger when the most recent value of the end-tidal carbon dioxide concentration is lower than the end-tidal carbon dioxide baseline threshold by at least 1.0 mm Hg.

These and other characteristics of the present invention will become apparent through reference to the following detailed description of the preferred embodiments and accompanying drawings.

FIG. 1 shows a schematic diagram of the venous air embolism detector of the present invention;

FIG. 2 is a graphical representation of a typical respiratory gas waveform;

FIG. 3 shows a flow chart of the operation of the venous air embolism detection procedure; and

FIG. 4 is a graphical representation of a breath averaging procedure used in the present invention.

FIG. 1 shows a schematic diagram of a venous air embolism detector 10 of the present invention. A respiratory gas analyzer 20 receives respiratory gases from a patient; analyzes the patient's respiratory gases; and sends signals corresponding to the concentrations of oxygen (O.sub.2), nitrogen (N.sub.2) and carbon dioxide (CO.sub.2) found in the patient's respiratory gases to a data processor 30. The data processor 30 receives directions for analyzing the concentrations of oxygen (O.sub.2), nitrogen (N.sub.2) and carbon dioxide (CO.sub.2) found in the patient's respiratory gases according to a specific procedure from a program source 40. The data processor 30 outputs the results of its analysis to an alarm 50 and a display 60.

The respiratory gas analyzer 20 may be any of a variety of gas analyzers routinely used for respiratory gas analysis including a Raman gas analysis system (e.g. a Rascal.RTM. II Raman Spectrometer available through Ohmeda Inc., Louisville, Colo.) or a mass spectrometry (e.g. an MGA 1100 mass spectrometer available through Marquette Medical Gas Analyzer Co., St. Louis, Mo.). It is preferred for the present invention that a Raman gas analysis system be used since it is less expensive than a mass spectrometer, can more easily be dedicated to a single patient, and has faster response times. Detailed descriptions of Raman gas analysis systems and subsystems are presented in the following U.S. patents, each of which is hereby incorporated herein by reference: 1) U.S. Pat. No. 4,784,486, entitled "MULTI-CHANNEL MOLECULAR GAS ANALYSIS BY LASER-ACTIVATED RAMAN LIGHT SCATTERING", issued to Van Wagenen et al.; 2) U.S. Pat. No. 4,676,639, entitled "GAS CELL FOR RAMAN SCATTERING ANALYSIS BY LASER MEANS", issued to Van Wagenen; 3) U.S. Pat. No. 5,135,304, entitled "GAS ANALYSIS SYSTEM HAVING BUFFER GAS INPUTS TO PROTECT ASSOCIATED OPTICAL ELEMENTS", issued to Miles et al.; 4) U.S. Pat. Re. No. 34,153, entitled "MOLECULAR GAS ANALYSIS BY RAMAN SCATTERING IN INTRACAVITY LASER CONFIGURATION", issued to Benner et al.; and 5) U.S. Pat. No. 5,245,405 entitled "CONSTANT PRESSURE GAS CELL", issued to Mitchell et al.

In one embodiment of the venous air embolism (VAE) detector of the present invention, the detector determines if a VAE is present by analyzing the values of five breath-by-breath gas concentrations extracted from complete breath cycles (one inspiration plus one expiration). These concentrations are: 1) the maximum expired or end-tidal value for carbon dioxide (EtCO.sub.2); 2) the average inspired value for nitrogen (FiN.sub.2); 3) the average expired value for nitrogen (FeN.sub.2); 4) the average inspired value for oxygen (FiO.sub.2); and 5) the average expired value for oxygen (FeO.sub.2).

The manner in which these gas concentration parameters are determined is described by reference to FIG. 2 which is a graphical representation of a typical respiratory gas waveform through a complete breath cycle comprising an expired portion and an inspired portion. A typical CO.sub.2 waveform 70 is shown for a complete breath cycle along with the first derivative of the CO.sub.2 waveform 80. Also shown is a typical N.sub.2 waveform 90 and a typical O.sub.2 waveform 100. The CO.sub.2 waveform 70 is sampled, using a sample and hold procedure, to determine the maximum CO.sub.2 concentration for each expired breath, which is defined to be the end-tidal carbon dioxide value, i.e., EtCO.sub.2. The CO.sub.2 waveform 70 is further analyzed by taking its first derivative 80 and applying a threshold criteria (less than 1% CO.sub.2 /100 msec) to define "steady state" portions of each portion of the complete breath waveform. An expired region A is defined to be the steady state region starting at time t.sub.1 and ending at time t.sub.2. Likewise, an inspired region B is defined to be the steady state region starting at time t.sub.3 and ending at time t.sub.4. The times t.sub.1, t.sub.2, t.sub.3 and t.sub.4 are determined by reference to the first derivative of the CO.sub.2 waveform 80 as shown. The expired region is defined as the steady state region which corresponds to the portion of the CO.sub.2 waveform 70 which exhibits the higher concentration of CO.sub.2 while the inspired region is defined as the steady state region which corresponds to the portion of the CO.sub.2 waveform 70 which exhibits the lower concentration of CO.sub.2.

The average expired value for nitrogen (FeN.sub.2) and the average inspired value for nitrogen (FiN.sub.2) are determined from the N.sub.2 waveform 90. The value of FeN.sub.2 is defined as the average nitrogen concentration during the expired region A of the breath cycle and the value of FiN.sub.2 is defined as the average nitrogen concentration during the inspired region B of the breath cycle. Similarly, the average expired value for oxygen (FeO.sub.2) and the average inspired value for oxygen (FiO.sub.2) are determined from the O.sub.2 waveform 100. The value of FeO.sub.2 is defined as the average oxygen concentration during the expired region A of the breath cycle and the value of FiO.sub.2 is defined as the average oxygen concentration during the inspired region B of the breath cycle.

It is to be understood that the above described method for determining the values of the parameters EtCO.sub.2, FeN.sub.2, FiN.sub.2, FeO.sub.2 and FiO.sub.2 is but one technique which may be used. Other methods may also be used to practice the present invention.

A flow chart of the operation of the venous air embolism detection procedure represented by program source 40 (FIG. 1) is presented in FIG. 3. The process starts at an input block 110 wherein the breath-by-breath gas concentration data, i.e., the values of the parameters EtCO.sub.2, FeN.sub.2, FiN.sub.2, FeO.sub.2 and FiO.sub.2, are input into the data processor 30. Input block 110 proceeds to a decision block 120 where FeN.sub.2 is compared to a predetermined value, in this case 1.5%. If FeN.sub.2 is greater than 1.5%, the process proceeds along the NO path to activity block 130 where the alarm triggers are set to FALSE. The process then proceeds to activity block 140 where all of the input data are saved for the next call and the process returns to the input block 110. This branch and termination of the process is necessary because the background nitrogen concentration is too high and the associated noise on the signal will overwhelm the much lower level FeN.sub.2 signal arising from a VAE event.

In decision block 120, if FeN.sub.2 is less than or equal to 1.5%, the process proceeds along the YES path to activity block 150 where two types of averages, a long term historical average and a recent change average are calculated. Long term historical averages and short term recent change averages are calculated for each of the following parameters: 1) the difference between the FiO.sub.2 and the FeO.sub.2 parameters, i.e., FiO.sub.2 minus FeO.sub.2, also referred to as the Delta O.sub.2 value; 2) the EtCO.sub.2 parameter; and 3) the FeN.sub.2 parameter. One method for determining the historical averages is illustrated in FIG. 4.

In FIG. 4, the data for 15 breaths are shown schematically as 152. The length of each bar represents the value of a particular parameter, i.e. Delta O.sub.2, EtCO.sub.2 or FeN.sub.2, for each of 15 breaths. For the EtCO.sub.2 and FeN.sub.2 parameters, the corresponding long term historical average is an 11 element data base which includes the 11 most recent 2 breath averages. For example, at breath 12, the long term historical average data base for the EtCO.sub.2 parameter includes as its first entry the average of the EtCO.sub.2 parameter at breaths 1 and 2, i.e. i; its second entry is the average of breaths 2 and 3, i.e. ii; its third entry is the average of breaths 3 and 4, i.e. iii; . . . ; its tenth entry is the average of breaths 10 and 11, i.e. x; and its eleventh entry is the average of breaths 11 and 12, i.e. xi. At breath 13, the long term historical average data base is updated by dropping the oldest 2 breath average, i.e. entry i, and adding the most recent 2 breath average, i.e. the average of breaths 12 and 13, xii. In this manner, the historical data base for each parameter is constantly updated to always include the eleven most recent 2 breath averages.

For the Delta O.sub.2 parameter, the corresponding long term historical average is an 11 element data base which includes the 11 most recent 4 breath averages. The procedure for creating and updating the Delta O.sub.2 historical average data base is the same as that previously described for EtCO.sub.2 and FeN.sub.2 except that the averages are for 4 breaths instead of 2.

In conjunction with updating the historical data bases for the Delta O.sub.2, EtCO.sub.2 and FeN.sub.2 parameters in activity block 150 (FIG. 3), the recent change data for each parameter is calculated. The recent change data for EtCO.sub.2 is its most recent 2 breath average. Likewise, the most recent change data FeN.sub.2 is its most recent 2 breath average. However, the recent change data for Delta O.sub.2 is its most recent 4 breath average. After the historical average data bases and the recent change data are calculated in activity block 150, the process proceeds to a decision block 160.

In decision block 160, the most recent change data for each parameter is compared with its respective historical averages data base. Each of the following three conditions must be TRUE in order to proceed along the YES path from decision block 160 to activity block 170. First, the most recent change data for the FeN.sub.2 parameter must be higher by at least 0.03 vol % than the highest of the 11 entries in the FeN.sub.2 historical averages data base. (In some cases the threshold value is 0.25 vol %.) Second, the most recent change data for the EtCO.sub.2 parameter must be lower by at least 1.5 mm Hg than the lowest of the 11 entries in the EtCO.sub.2 historical averages data base. Third, the most recent change data for the Delta O.sub.2 parameter must be lower by at least 0.2 vol % than the highest of the 11 entries in the Delta O.sub.2 historical averages data base. These threshold criteria are summarized in Table I. If any one these conditions is FALSE, the process proceeds along the NO path to activity block 140 where all of the input data are saved for the next call and the process returns to the input block 110.

If all three threshold criteria in decision block 160 are TRUE, then the process proceeds to activity block 170 where an EtCO.sub.2 trigger is set to TRUE, an FeN.sub.2 trigger is set to TRUE, and a Delta O.sub.2 trigger is set to TRUE. The process then proceeds to decision block 180 where the timing of the three TRUE signals is tested. It is noted that the EtCO.sub.2 trigger, the FeN.sub.2 trigger and the Delta O.sub.2 trigger must all be TRUE during the same time period, i.e., for the same breaths, to signal the presence of a VAE. For example, the current data for breaths 14 and 15 (EtCO.sub.2 and FeN.sub.2) and breaths 12, 13, 14 and 15 (Delta O.sub.2) must all meet the three criteria shown in TABLE I. In decision block 180, the relative timing of the EtCO.sub.2 trigger, the FeN.sub.2 trigger and the Delta O.sub.2 trigger is checked to ensure that all three TRUE triggers occurred within the same predetermined time window. If any one of the TRUE triggers occurred outside of the time window, the process proceeds along the YES path to activity block 190 where all of the triggers are reset to FALSE. The process then continues on to activity block 140 where all of the input data are saved for the next call and the process returns to the input block 110.

If the relative timing of the EtCO.sub.2 trigger, the FeN.sub.2 trigger and the Delta O.sub.2 trigger do all occur within the same predetermined time window, the process proceeds from decision block 180 along the NO path to decision block 200. In decision block 200, a disconnect check is performed to determine if there is a disconnection in the airway circuit which would permit air to enter the system. A disconnect is signalled if: 1) the difference between the last FiN.sub.2 value and the current FiN.sub.2 value is greater than 25% of the current FiN.sub.2 ; OR 2) the value of FiN.sub.2 has increased by 2 vol % or more. If either of these conditions is TRUE, a disconnect alarm is activated and the process proceeds along the YES path to activity block 140 where all of the input data are saved for the next call and the process returns to the input block 110.

If neither of the two disconnect criteria are TRUE, then the process continues along the NO path to decision block 210.

If the EtCO.sub.2 trigger, the FeN.sub.2 trigger and the Delta O.sub.2 trigger are all TRUE in decision block 210, the process proceeds along the YES path to activity block 220 where a VAE alarm is activated to signal the presence of a VAE. If any one of the EtCO.sub.2 trigger, the FeN.sub.2 trigger or the Delta O.sub.2 trigger is FALSE in decision block 210, the process proceeds along the NO path to activity block 140 where all of the input data are saved for the next call and the process returns to the input block 110.

While the description of the VAE detector was described using the 11 breath averaging procedure illustrated in FIG. 4, other averaging scenarios are also possible. For example, the historical averages data base may be more or less than the 11 breaths illustrated and the current data may be more than the 2 breath average illustrated.

The above described VAE detection system was tested using a Raman gas analyzer in a study on dogs. In this test, the respiration rate of the dogs provided a breath output approximately every 6 seconds, thus the 11 breaths covered a 66 second time window. The results of this test are summarized in TABLE II.

It will be understood that the apparatus and method of the present invention for a venous air embolism detector may be employed with many different types of respiratory gas analysis systems including Raman and mass spectrometer. Additionally, the selection of a specific averaging technique and specific threshold criteria may vary. Thus, there are numerous other embodiments of the venous air embolism detector which will be obvious to one skilled in the art, including but not limited to changes in the specific averaging technique and threshold criteria, the type of gas analysis system used, the number of gas parameters used, the comparison technique used, for instance comparison of the most recent breath to a running average to establish baseline instead of the highest or lowest value, etc. The apparatus and method of the present invention may be embodied in other specific forms without departing from its spirit or essential characteristics. The described embodiments are to be considered in all respects only as illustrative and not restrictive. The scope of the invention is, therefore, indicated by the appended claims rather than by the foregoing description. All changes which come within the meaning and range of equivalency of the claims are to be embraced within their scope.