British Journal of Anaesthesia 1990; 65: 245-247

COMPARISON OF FOUR PULSE OXIMETERS: EFFECTS OF VENOUS OCCLUSION AND COLD-INDUCED PERIPHERAL VASOCONSTRICTION J. A. LANGTON, D. LASSEY AND C. D. HANNING

METHOD AND RESULTS

J. A. LANGTON, F.F.A.R.C.S.; D. LASSEY, F.F.A.R.C.S.; C D .

HANKING, B.SC., F.F.A.R.C.S. ; University Department of AnEthics Committee approval was obtained; 20 aesthesia, Leicester General Hospital, Leicester LE5 4PW. healthy subjects gave informed consent to par- Accepted for Publication: February 28, 1990. ticipate in the study. Correspondence to C.D.H.

Downloaded from http://bja.oxfordjournals.org/ at University of Chicago on July 7, 2015

The Ohmeda 3700, Nellcor N100, Nellcor N200 (which has an ECG linkage) and the Datex The ability of four pulse oximeters (the Ohmeda Oscar were evaluated. All the pulse oximeters 3700, Nellcor N100 and N200 and the Datex tested were used in the default mode set at the Oscar) to detect hypoxaemia was determined in minimum time for averaging. The subjects rested the presence of venous obstruction and cold- on a couch and the oximeter probes were applied induced peripheral vasoconstriction. Significant in random order to the fingers of the left hand, the increases in detection time for hypoxaemia were thumb not being used. The probe of an Ohmeda found in both cases. There were no significant Biox III oximeter was applied to the subject's ear differences in detection time between the instru- lobe as a source of reference data and for safety ments, a/though the Ohmeda 3700 displayed purposes. The ECG leads of the Nellcor N200 smaller values of SaOj under certain conditions. were applied in the CM5 configuration. The Peripheral vasoconstriction was induced using oxyhaemoglobin saturation (.Sa^) outputs from three differing methods which gave differing the instruments, calibrated according to the results, thus emphasizing the importance of manufacturer's instructions, were recorded on a methodology in assessments of pulse oximetry. W & W electronic 400 series (model 406) sixchannel chart recorder with a thermowriting system. KEY WORDS All experiments were conducted in an airEquipment: pulse oximeters. conditioned laboratory with a constant room Pulse oximetry is rapidly being adopted as a temperature of 23-25 °C. When stable readings were obtained, a 5-min standard in clinical monitoring [1]. It is clear that period was recorded for baseline values and two instruments from different manufacturers vary in episodes of hypoxaemia were induced, the subject their ability to function under adverse conditions, inhaling a mixture of 10% oxygen in nitrogen in particular poor peripheral perfusion and motion from a pre-filled Douglas bag via a closely fitting artefact. A previous study from this department non-rebreathing mask with normal tidal ven[2] found that venous occlusion and cold-induced tilation. Each episode of hypoxaemia lasted 2 min peripheral vasoconstriction impaired the perform- and was terminated if the reference instrument ance of some of the pulse oximeters under test, so indicated that Sa had decreased to < 75 %. The that episodes of hypoxaemia were not detected subjects breathedOi room air for 3 min between despite apparently normal functioning. The pres- episodes. The time from the first breath of ent study extends the evaluation to further hypoxic gas to the first decrease in the SaOj record oximeters, including one with ECG linkage, and (detection time) and the minimum Sa^ were evaluates different methods of inducing vaso- noted. constriction. SUMMARY

246

BRITISH JOURNAL OF ANAESTHESIA

tween two flat coils of a blood warmer (Fenwal) A standard sphygmomanometer cuff was ap- held gently in place with a crepe bandage. Ice cold plied to the upper arm and inflated to 40 mm Hg water was circulated through the coils using a to induce venous congestion. Five minutes later, roller pump. Thermocouples were inserted into two further episodes of hypoxaemia were induced the circuit to measure the temperature of the water entering and leaving the coils. The subject's as before. hand temperature was measured continuously Vasoconstriction using an Edale digital thermometer with the The test arm was cooled using three different temperature probe applied to the inner aspect of methods on three separate occasions. First, the the index finger. The subject's forearm was cooled subject's forearm was surrounded by a water- to 26 °C, at which point the roller pump was filled plastic envelope that had been cooled stopped. Two episodes of hypoxaemia were overnight in a domestic refrigerator to 4 °C. The induced as before. oximeter probes were allowed to protrude from Differences between the instruments were assesthe envelope to avoid direct cooling. Five minutes sed by analysis of variance and the Wilcoxon later, 12 episodes of hypoxaemia were induced at signed rank test as appropriate. 5-min intervals as the cooling progressed; the One subject was found to have markedly final two episodes were included in the final prolonged control and test detection times of 57 analysis. The total period of induced vaso- and 116 s, respectively, and a reduced skin constriction was 60 min. Second, the subject's temperature of 26.8 °C indicating pre-existing forearm was laid on a water-filled bag which had vasoconstriction. Her results were excluded from been cooled overnight to 4 °C. A second cold bag the detailed analysis. She was a healthy 20-yr-old was then laid across the arm so that the subject's recovery room nurse, with no symptoms of hand and fingers protruded. The sequence of peripheral vascular disease or Raynaud's disease. episodes of hypoxaemia was repeated as above. Venous congestion significantly prolonged the Third, the subject's forearm was wrapped be- detection time in all oximeters compared with Venous occlusion

Detection time (s) Instrument Venous occlusion (n = 13) Datex Oscar Nellcor 200 Ohmeda 3700 Nellcor 100 Biox III Vasoconstriction (n = 15) Datez Oscar Nellcor 200 Ohmeda 3700 Nellcor 100 Biox III

Minimum Sao, (%)

Control

Test

Control

Test

36.2 (2.3) [22-50] 34.3 (2.5) [24-53] 36.5 (2.1) [24-53] 31.1 (1.9) [24-46] 31.9 (2.9) [14-56]

49.1 (5.7)* [24-96] 47.6 (5.5)* [24-84] 48.8 (4.8)** [22-75] 42.6 (5.2)** [19-75] 30.3 (3.2) [19-52]

86.48 (0.73) [78-94] 86.4 (0.94) [78-94] 86.28 (0.8) [78-92] 86.4 (0.9) [79-92] 87.8 (0.7) [80-94]

85.3 (0.87) [78-93] 84.1 (0.87) [77-91] 84.9 (0.8) [78-93] 86.23 (0.79) [76-91] 86.5 (0.64) [80-92]

38.5 (3.2) [22-67] 37.8 (2.6) [24-58] 37.2 (2.8) [24-62] 36.1 (2.9) [24-62] 32.9 (2.9) [14-58]

76.2 (10)** [29-168] 67.8 (8.3)** [29-146] 66.7 (8.2)** [24-161] 66.4 (8.5)** [26-149] 32.0 (4.7) [17-86]

86.5 (0.6) [78-94] 85.7 (0.76) [78-92] 86.3 (0.82) [78-94] 85.8 (0.88) [79-93] 85.7 (0.76) [78-94]

86.4 (0.89) [81-92] 85.9 (0.87) [76-92] 81.1 (0.94) [68-94] 85.5 (0.96) [77-94] 85.9 (0.87) [81-94]

Downloaded from http://bja.oxfordjournals.org/ at University of Chicago on July 7, 2015

TABLE I. Detection time and minimum SaOt for epuodes of induced hypoxaemia during venous occlusion and vasoconstriction (all three methods combined) for each of the pulse oximeters. Mean (SEM) [range] (Wilcoxon signed rank test).*P < 0.05; **P < 0.01

PULSE OXIMETRY AND POOR PERFUSION

The Ohmeda 3700 was common to the present and previous study [2]. The control and test detection times were approximately twice as long in the previous study as in the present, although the methodology was identical. There are two explanations for the differences. First, the envelope used in the previous study may have been filled to such a degree that it also produced venous obstruction in a manner similar to the pumped method in the present study. Second, the subjects may have had poor peripheral perfusion before cooling started. In the present study one subject, who was excluded from the analysis, was found to have a finger temperature of 26.8 °C at the start of the study. Her control and test detection times were 57 s and 116 s, respectively—that is, of the COMMENT same order of magnitude as values found in the The purpose of a pulse oximeter is rapid, accurate previous study. and reliable detection of hypoxaemia. It is inIn contrast with the previous study [2], none of evitable that under certain circumstances a pulse the oximeters failed, and they were comparable in wave is not detectable, and the instruments their responses to venous congestion and vasoshould alert the user immediately. All the oxi- constriction, although the SaOi values displayed meters in the present study functioned reliably by the Ohmeda 3700 were consistently less than and accurately, although with a marked pro- those from the other instruments. This has been longation of detection time. This is in contrast noted by other workers [2, 3]. with our previous study [2] in which one oximeter New oximeters are being introduced frequently failed to detect hypoxaemia despite apparently and it is important that their ability to perform normal functioning. reliably under adverse conditions is tested. The The prolongation of detection time is clinically present study suggests that there have been important, as equivalent degrees of venous con- improvements in the quality of the microprocessor gestion and vasoconstriction may occur in prac- software since the previous study. It also emphatice. The possibility of delayed detection (and sizes the need for consistent methods in inducing thus correction of hypoxaemia) lends weight to vasoconstriction and the importance of avoiding the necessity of monitoring inspired gas concen- inadvertent venous congestion. ECG linkage is trations and adequacy of ventilation. used by some pulse oximeters to scan for a pulse We assessed three different methods of surface wave during a short time period following the R cooling to induce peripheral vasoconstriction. wave on the ECG. It is thought that this may be They may not reproduce exactly the conditions helpful in conditions of poor peripheral perfusion, found in clinical practice, in that they do not but in this study no difference was found between produce concomitant reductions in cardiac out- the instrument with ECG linkage and those put, but they are simple, non-invasive and applied without. easily to volunteers. The envelope and bag REFERENCES methods are convenient and applied easily to the 1. Eichom JH, Cooper JB, Cullen DJ. Standards for patient subject's arm, but there was no control over the monitoring during anesthesia at Harvard Medical School. rate of cooling of the arm. The pump method was Journal of the American Medical Association 1986; 256: devised to give controllable surface cooling, but 1017-1020. probably introduced some venous occlusion, as 2. Wilkins CJ, Moores M, Harming CD. Comparison of the detection times were significantly greater than pulse oximeters: effects of vasoconstriction and venous engorgement. British Journal of Anaesthesia 1989; 62: the other methods. A future modification to this 439-444. technique would be to remove the coils from the 3. Severinghaus JW, Naifeh KH. Accuracy of response of six subject's arm when the cooling was complete and pulse oximeters to profound hypoxia. Anesthesiology 1987; so avoid venous occlusion. 67: 551-558.

Downloaded from http://bja.oxfordjournals.org/ at University of Chicago on July 7, 2015

control (P < 0.05) (table I). There were no differences between the oximeters with respect to detection time. The combined results of all three methods of producing peripheral vasoconstriction showed a significant prolongation of detection time compared with control values (P < 0.01). There were no differences between the oximeters. Both the control and test detection times for the pump method of producing peripheral vasoconstriction were significantly longer than the other two methods. The Ohmeda 3700 gave significantly smaller values of Sao> compared with the other oximeters (P < 0.05).

247

Comparison of four pulse oximeters: effects of venous occlusion and cold-induced peripheral vasoconstriction.

The ability of four pulse oximeters (the Ohmeda 3700, Nellcor N100 and N200 and the Datex Oscar) to detect hypoxaemia was determined in the presence o...
225KB Sizes 0 Downloads 0 Views