monitoring of subjects during avalanche snow...

MONITORING OF SUBJECTS DURING AVALANCHE SNOW BREATHING EXPERIMENTS: POSSIBLE ERRORS

Lenka Horakova1*, Ladislav Sieger2, Karel Sykora3, and Karel Roubik1

1 Czech Technical University in Prague, Faculty of Biomedical Engineering, Kladno, Czech Republic2 Czech Technical University in Prague, Faculty of Electrical Engineering, Prague, Czech Republic 3 Charles University in Prague, Faculty of Physical Education and Sport, Prague, Czech Republic

ABSTRACT: The gas exchange and work of breathing in avalanche burial victims have been a sub-ject of numerous field experiments with healthy volunteers. The vital sign monitors used in monitoring of subjects during these experiments have to work in non-standard ambient conditions, outside the programmed physiological limits and may affect the accuracy of the measurements. In order to assess the reliability of pulse oximetry as a monitoring method of experimental subjects, we used five different pulse oximeters to monitor 12 healthy volunteers during a trial, in which they were breathing into simu-lated avalanche snow. In the course of the trial when desaturation occurred, we observed significant discrepancies in oxygen saturation readings among the oximeters—in the rate of desaturation, and in the lowest recorded values. These observations suggest that relying on a single parameter which is SpO2 as a study endpoint, or as a safety measure, may present a possible safety risk. In future avalanche trials, endpoints should be set on the basis of more parameters and continuous clinical as-sessment.

KEYWORDS: avalanche burial, outdoor experiments, hypoxia, pulse oximetry, technical limits.

1. INTRODUCTIONThe research on pathophysiology of avalanche buried victims relies on experiments with healthy volunteers to a large extent. In numerous field trials, the subjects were breathing into snow with artificial air pockets (Brugger et al., 2003; Roubik et al., 2015; Strapazzon et al., 2017), or using various types of avalanche survival devices (Grissom et al., 2000; Radwin et al. 2001, Wind-sor et al., 2009). During these trials, changes in subject's breathing pattern can be observed— standardly, hyperventilation with worsening hy-poxia and hypercapnia. These observations cor-respond with the statistics from the real avalanche accidents in the Alps, in which most of the victims die due to asphyxia (Hohlrieder et al., 2007)

For the safety of the experimental subjects, the use of standard vital sign monitoring is essential: electrocardiogram, blood pressure measurement and pulse oximetry. However, the vital sign moni-tors used in the breathing experiments have to work in non-standard ambient conditions, outside the programmed physiological limits and may affect the accuracy of the recorded parameters (Horakova et al., 2018).

The oxygen saturation of haemoglobin (SpO2) usually serves as a study endpoint as well. The endpoint is usually set at values of 75% (Brugger et al., 2003; Strappazon et al., 2017) to 85% (Grissom et al., 2000; Radwin et al, 2001). Apart from SpO2, also the analysis of inhaled and ex-haled gases is used to obtain experimental data, to assess the gas exchange of the subject and it can also serve as a study endpoint.

Pulse oximeter is a widely used piece of equip-ment in hospitals and in pre-hospital care. Al-though it can provide indispensable information about patient’s condition, the measuring method has several well recognized limitations. For in-stance, in cold environment, where avalanche accidents and breathing experiments take place, a problem with low perfusion of cold extremities may result in erroneous readings of oxygen satu-ration and affect the results and duration of the trials.

2. OBJECTIVEThe aim of this paper is to assess the reliability of pulse oximetry as a monitoring method for outdoor breathing experiments in simulated avalanche snow. In order to assess this we eval-uate data from five different pulse oximeters used to simultaneously monitor each volunteer. This enables comparison of the values displayed by different devices and investigation of the pos-sible impact on subsequent data analysis and safety of the experimental subjects.

* Corresponding author address:Lenka Horakova, Czech Technical University in Prague, Faculty of Biomedical Engineering, Nam. Sitna 3105, Kladno 27201, Czech Republic;email: [email protected]

Proceedings, International Snow Science Workshop, Innsbruck, Austria, 2018

1409

3. METHODS The data used in this study were recorded during a prospective randomized single-blind crossover breathing trial, conducted in the Krkonose Moun-tains, Czech Republic, at the altitude 762 meters above sea level, in January 2018. The study was approved by the Institutional Review Board of the Faculty of Biomedical Engineering, Czech Technical University in Prague. The study regis-tration number in the ClinicalTrials.gov register is NCT03413878. Written informed consent was obtained from the volunteers before enrollment in the study. The subjects were 12 healthy male volunteers, non-smokers.

During the experiment, each subject was in prone position, ventilating through specially de-signed breathing tubes. In the initial phase, the subject was breathing the ambient air through a mouthpiece, allowing to measure ventilatory pa-rameters, while vital sign monitoring and gas analysis was performed. After that, the tubing was attached to simulated avalanche snow with a 2L air pocket—the main phase. This phase was terminated on subject’s request, or on su-pervising physician's command, or when the study endpoint was reached (end tidal carbon dioxide—EtCO2 62,5 mmHg). The participant was disconnected from the snow and allowed to breathe the ambient air again—the recovery phase.

Throughout all the phases, vital sign monitoring and clinical assessment of consciousness level by the supervising physician were performed. For SpO2 monitoring we used five different finger oxygen saturation probes, placed in all volun-teers on fingers of the right hand in the same order (see Fig. 1 and Table 1).

Figure 1: Subject’s right hand with finger probes of five different pulse oximeters.

Table 1: Pulse oximeters used and their stan-dardised placement on volunteer’s fingers.

We used devices standard for anaesthesia and critical care use: Datex-Ohmeda S/5 anesthesia monitor (Datex-Ohmeda, Madison, WI, USA), CareScape B650 (GE Healthcare, Helsinki, Fin-land), Edan M3B (Edan Instruments, Nanshan Shenzhen, China), including the new generation pulse oximeter Masimo Radical-7 Pulse CO-Oximeter (Masimo, Irvine, CA, USA). On the other hand, we also included a small hand held personal oximeter Nonin PalmSAT 2500 (Nonin Medical Inc., Plymouth, MN, USA). All used mon-itoring devices were validated for clinical use with a valid safety check.

The data from all pulse oximeters and monitors were logged and also the screens of the moni-tors were filmed to document the SpO2 values displayed by all oximeters at the same moment. The response time of all oximeters were set to their minimal values.

To eliminate possible erroneous readings due to low perfusion or motion artefacts, the volunteer’s hand with all probes was placed into a preheated insulated glove and the participants were in-structed to minimise hand and finger movements during experiments.

Apart from continuous monitoring, the volunteers were regularly clinically assessed by the super-vising clinician. To test the consciousness level, the clinician asked the subject to calculate sim-ple mathematical operations and show the re-sults using fingers of the left hand.

4. RESULTS There were observed significant differences in oxygen saturation readings displayed by the five pulse oximeter devices used in this experiment. They varied at the time of desaturation, in the lowest SpO2 value, and also in the recovery phase, when the subject was already breathing

Pulse oximeter device Finger

Datex-Ohmeda S/5 anesthesia monitor (Datex-Ohmeda, Madison, WI, USA)

V.

CareScape B650 (GE Healthcare, Helsinki, Finland)

III.

Edan M3B (Edan Instruments, Nanshan Shenzhen, China)

II.

Masimo Radical-7 Pulse CO-Oximeter (Masimo, Irvine, CA, USA)

IV.

Nonin PalmSAT 2500 (Nonin Medical Inc., Plymouth, MN, USA)

I.


1410

Figu

re 2

: Exa

mpl

e of

five

SpO

2 cu

rves

mea

sure

d by

five

diff

eren

t pul

se o

xim

eter

s. A

-D: f

our

time

poin

ts w

ith m

easu

ring

disp

repa

ncie

s (s

ee te

xt a

nd T

able

2 fo

r fur

ther

det

ails

).


1411

ambient air again and the oxygen saturation was getting back to normal.

An example of oxygen saturations recorded in one subject by five different devices during one trial is presented in Fig. 2. In the graph there are four selected time points labeled A-D and the corresponding SpO2 readings from all pulse oximeters are showed in Table 2.

At the beginning of the trial, shortly after connec-tion to the snow (A), the SpO2 values recorded by al l oximeters were almost identical (99-100%). Gradually, with increasing hypoxia and hypercapnia due to rebreathing of the air contained in snow, desaturation occurred.

The desaturation phase had a significantly dif-ferent waveform presented by each device. If we consider SpO2 75% as a hypothetical experi-mental endpoint (a value used by several studies in the past: Brugger et al., 2003; Strappazon et al., 2017), the first oximeter which displayed this saturation was Nonin PalmSAT. This occurred at the time when other oximeters were showing values from 77% to 90% (B).

The trial was terminated after 6 minutes and 30 seconds, when the study endpoint (EtCO2 over 62,5 mmHg) was reached. Just before the dis-connection (C), the main monitor CareScape B650 displayed SpO2 71%. However, at the same time, Masimo Radical-7 was still showing SpO2 above the considered limit (78%), Edan reached exactly 75% and two other monitors were alarming critical desaturation at 66% and 53% (Datex-Ohmeda and Nonin, respectively). These values of SpO2 were considered only as informative because oxygen saturation was not a study endpoint (unlike in other studies) and clini-cal assessment of the volunteer and his con-sciousness level did not suggest that experiment should be terminated.

Additionally, in the recovery phase, there was also a significant variability in oxygen saturation increase back to normal values (D). Whereas Nonin 2500 and Masimo Radical were already displaying normal values at 100% and 98%, re-spectively, the anaesthetic monitors were still in low saturations (Datex-Ohmeda 67% and CareScape 81%), as well as Edan monitor (78%). Moreover, Datex-Ohmeda S/5 anaesthe-sia monitor had before this point 30 s of loss of readings, despite no apparent problems with finger probe position, peripheral perfusion, or movement artefacts.

Very similar discrepancies among SpO2 readings were recorded in other volunteers in this trial as well.

5. DISCUSSION The main finding of the study is that five simulta-neously monitoring pulse oximeters showed sig-nificantly different oxygen saturation readings in the same volunteer in a course of a simulated avalanche snow breathing experiment. The low-est reached values showed notable variance as well. These findings imply considerable conse-quences for the safety of the experimental sub-jects, and for the experimental data analysis.

If oxygen saturation is set as a breathing study endpoint, the choice of a particular device can affect the duration of the experiment. For in-stance, in our experiment, if we had set the study endpoint at SpO2 75%—a value used in similar studies before (Brugger et al., 2003; Strappazon et al., 2017)—we would have terminated the ex-periment at a different point in time, depending on the selected oximeter. The time difference between the moment when the first pulse oxime-ter showed SpO2 75% (Nonin 2500, 04:40 min after experiment beginning) and the last one (Masimo Radical-7, 06:10 min after experiment beginning) was one and a half minute.

SpO2 in %

A: shortly after connection to snow

B: SpO2 75% reached by the first oximeter

C: just before disconnection from snow

D: recovery phase

Datex-Ohmeda S/5 monitor

100 77 66 67

CareScape B650 99 84 71 81

Masimo Radical 7 100 90 78 98

NONIN 2500 99 75 53 100

Edan M3B 100 84 75 78

Table 2: Four selected time points A-D from Fig. 2 and corresponding SpO2 values in % dis-played by the individual pulse oximeters.


1412

Considering volunteers’ safety, we cannot rely on a single parameter. According to one pulse oximeter the subject is in a critical condition, but another oximeter may show significantly higher number. However, a falsely higher number can cause underestimation of a critical situation, if no proper clinical assessment is done.

The significant differences in displayed SpO2 values may have several reasons. There can be some construction variability among the pulse oximeter devices. The authors speculate that besides measured values of SpO2, the figures displayed by the monitors are modified to some extent by embedded software. This software may contain algorithms that should help to elimi-nate potentionally erroneous SpO2 readings. For instance, Edan M3 monitor was showing oxygen saturations between 73% and 78% throughout the whole desaturation phase, with only minimal fluctuations.

Moreover, as a part of the settings of each de-vice, it is possible to select data averaging and display refreshment time, usually referred to as ‘response’. This means, how fast the displayed value follows the measurement of the parameter. For instance, for SpO2, the monitor can display the values from beat-to-beat, or it can present an average of results from the past time period, e.g. 20 s. The latter is a default setting for Datex-Ohmeda S/5 monitors, because in anaesthesia it helps to eliminate distracting artefacts and false alarms. However, in breathing experiments, we observe much faster changes in volunteers’ physiological parameters and this averaging can give us incorrect information about subject’s state and inaccurate experimental data. In addi-tion, this can present safety risks to the volun-teers.

6. CONCLUSIONThis study documents that even though standard monitoring equipment is used during outdoor breathing trials, it has notable limitations. Relying on a single parameter which is SpO2 as a study endpoint, or as a safety measure, may present a possible safety risk. The irreplaceable role of clinical assessment by a skillful physician should be considered. For future design of avalanche trials we should not depend on a single number, but the endpoints should be set on the basis of more parameters and continuous clinical as-sessment.

ACKNOWLEDGEMENT The study was supported by Czech Technical University in Prague grant No. SGS17/203/OHK4/3T/17.

REFERENCES Brugger, H., G. Sumann, R. Meister, L. Adler-Kastner, P. Mair, H.

C. Gunga, W. Schobersberger, M. and Falk, 2003: Hypoxia and hypercapnia during respiration into an artificial air pocket in snow: implications for avalanche survival. Resuscitation, 58(1), 81-88.

Grissom, C. K., M.I. Radwin, C. H. Harmston, E. L. Hirshberg, and T.J. Crowley, 2000: Respiration during snow burial using an artificial air pocket. Jama, 283(17), 2266-2271.

Hohlrieder, M., H. Brugger, H. M. Schubert, M. Pavlic, J. Ellerton, and P. Mair, 2007: Pattern and severity of injury in avalanche victims. High altitude medicine & biology, 8(1), 56-61.

Horakova, L., K. Sykora, L. Sieger, and K. Roubik, 2018: Breath-ing Experiments into the Simulated Avalanche Snow: Medical and Technical Issues of the Outdoor Breathing Trials. In: Lhot-ska L., L. Sukupova, I. Lacković, and G. Ibbott (eds) World Congress on Medical Physics and Biomedical Engineering 2018. IFMBE Proceedings, 68(1), 711-717.

Radwin, M. I., C. K. Grissom, M. B. Scholand, and C. H. Harm-ston, 2001: Normal oxygenation and ventilation during snow burial by the exclusion of exhaled carbon dioxide. Wilderness & environmental medicine, 12(4), 256-262

Roubik, K., L. Sieger, and K. Sykora, 2015: Work of breathing into snow in the presence versus absence of an artificial air pocket affects hypoxia and hypercapnia of a victim covered with avalanche snow: a randomized double blind crossover study. PloS one, 10(12), e0144332.

Strapazzon, G., P. Paal, J. Schweizer, M. Falk, B. Reuter, K. Schenk, H. Gatterer, K. Grasegger, T. Dal Cappello, S. Malacrida, and L. Riess, 2017: Effects of snow properties on humans breathing into an artificial air pocket–an experimental field study. Scientific reports, 7(1), 17675.

Windsor J. S., E. Hamilton, M. P. Grocott, M. J. O’Dwyer, and J. S. Milledge, 2009: The Snow Snorkel: a proof of concept study. Wilderness and Environmental Medicine, 20, 61-65.


1413

monitoring of subjects during avalanche snow...

Documents