Pulse Oximetry Part I by Stephen Colucciello, MD
Sunday, May 9, 2010

Is it necessary to transport arterial samples on ice?
Basically, the answer is “NO.” In the first study, the authors conclude that blood samples analyzed within 20 minutes of arterial puncture do not require storage on crushed ice (with the exception of those obtained in patients with very high white cell counts). In the second paper, the authors suggest that arterial blood samples do not require storage in ice if blood gas analysis is to be performed within 30 minutes.

17. IS IT NECESSARY TO TRANSPORT ARTERIAL SAMPLES ON ICE FOR pH AND GAS ANALYSIS? Nanji, A.A., et al, Can Anaesth Soc J 31(5):568, September 1984 Studies comparing the effect of storage on ice and storage at room temperature on the pH, PaO2, and PaCO2 of arterial blood samples have yielded conflicting results. This study, from the University of British Columbia, compared these parameters in duplicate arterial blood samples obtained from 21 ICU patients. Blood samples of 5-6ml, obtained with syringes having a heparin volume of approximately 0.1 ml, were stored at room temperature or on crushed ice (0øC). Serial determination of pH, PaO2, and PaCO2 was performed up to 60 minutes after the samples were obtained. All of the patients had normal hemoglobin values, and none had an elevated white cell count. The pH values ranged between 7.17 and 7.64. The ranges of the PaO2 and PaCO2 were 54-230mm Hg and 24-77mm Hg, respectively. In both sets of blood samples, increasing storage time was associated with decreases in the pH and PaO2, and an increase in the PaCO2. However, the difference in the magnitude of the changes between the two storage methods was not considered to be clinically significant until 20 minutes after the samples were obtained. in samples stored on ice for 20 minutes, the range of pH changes was 0 to minus 0.02 units; in samples stored at room temperature for 20 minutes, the range was 0 to minus 0.03 units. The range of the changes in PaO2 and PaCO2 after 20 minutes of storage on ice or at room temperature were similar. The authors conclude that blood samples to be analyzed within 20 minutes of arterial puncture do not require storage on crushed ice, with the exception of blood obtained from patients with very high white cell counts, which is associated with an excessive rate of oxygen metabolism and carbon dioxide production. 6 references 3/85-#23

18. STABILITY OF BLOOD GASES IN ICE AND AT ROOM TEMPERATURE Liss, H.P., et al, Chest 103(4):1120, April 1993 BACKGROUND: The reported temperature-dependent changes in arterial blood gas composition that may occur during the interval between blood sampling and analysis have prompted recommendations that blood samples be kept in ice prior to arterial blood gas measurement. METHODS: The authors, from Wright State University School of Medicine in Dayton, OH, compared changes in blood gas composition in samples stored in ice or at room temperature. One hundred nineteen arterial blood samples were analyzed immediately or were stored in ice prior to baseline analysis. After this initial analysis, 60 samples were stored in ice and 59 were stored at room temperature, and further blood gas analysis was performed after 15 and 30 minutes of storage. RESULTS: In both groups of samples, the PaO2 increased significantly after 15 minutes of storage and exhibited a further significant increase after 30 minutes, and the PaCO2 and pH decreased significantly after 15 minutes. In the sample stored at room temperature, but not those stored in ice, the pH exhibited a further significant decrease after 30 minutes. The mean PaO2 increased from 79.7mm Hg at baseline to 82.7mm Hg at 30 minutes in samples stored in ice, and from 78.2mm Hg to 80.6mm Hg in samples stored at room temperature. The PaCO2 decreased from 38.5mm Hg to 38.0mm Hg and from 37.9mm Hg to 37.4mm Hg in the ice and room temperature groups, respectively, and the pH decreased from 7.43 to 7.42 and from 7.41 to 7.40, respectively. CONCLUSIONS: The authors suggest that arterial blood samples do not require storage in ice if blood gas analysis is to be performed within 30 minutes. 16 references 08/93 - #34



What about doing venous gases instead of arterial gases in the setting of metabolic acidosis and for screening of hypercapnia?
In the first paper, the authors suggest that patients with nearly normal venous blood gases are unlikely to have severe acid-base disturbances, while extremely abnormal venous blood gases reflect comparable arterial blood gas values. However, they note that the predicted arterial pH may differ by more than 0.11 units in 5% of cases, the predicted bicarbonate may differ by more than 4.96mEq/l in 5%, and the predicted pCO2 may differ by more than 14.18 torr in 5%.

In the second paper examining acutely ill ED patients, there was a high degree of correlation and agreement between venous and arterial pH.

In the third paper, there was excellent correlation between arterial and venous blood gas samples. However, the emergency physicians who were surveyed felt that the actual differences were too large to allow the routine substitution of venous for arterial gases.

In the fourth paper, the authors found that among patients with acute respiratory illness, venous pH correlates well with arterial pH. However, levels of agreement between the arterial and venous pCO2 were less optimal in many individuals. The authors suggest that a combination of venous blood gas analysis and pulse oximetry can be a useful alternative to routine arterial blood gas analysis in such patients. As a screen for hypercarbia, a venous pCO2 cut-off of 45mm Hg had a sensitivity of 100% and a specificity of 57%.

19. THE USEFULNESS OF PERIPHERAL VENOUS BLOOD IN ESTIMATING ACID-BASE STATUS IN ACUTELY ILL PATIENTS Gennis, P.R., et al, Ann Emerg Med 14(9):845, September 1985 The authors analyzed the relationship between acid-base parameters of paired arterial and venous blood samples in 171 acutely ill adults and twelve cardiac arrest patients treated at the Bronx Municipal Hospital Center in New York. Significant alkalemia was present in 28 of the nonarrest patients and significant acidemia in 20. In the nonarrest patients, the mean venous pH was 0.056+/-0.056 units lower than the mean arterial pH, the mean venous pCO2 was 7.38+/-7.51 torr higher, and the mean bicarbonate was 1.21+/-2.55mEq/l higher than the arterial level. In the cardiac arrest patients, the differences were 0.086+/-0.10 units, 6.58+/- 19.98 torr, and 3.74+/-14.05mEq/L, respectively. Equations were developed that allowed relatively accurate prediction of 68% of arterial pH, pCO2, and bicarbonate values in nonarrest patients based on determination of these parameters in venous blood. However, the authors point out that the predicted arterial pH may differ by more than 0.11 units in 5% of cases, the predicted bicarbonate may differ by more than 4.96mEq/l in 5%, and the predicted pCO2 may differ by more than 14.18 torr in 5%. The authors conclude that their findings do not support the routine utilization of venous blood specimens for assessment of acid-base status, but indicate that inadvertent peripheral venous blood sampling may prove to be useful in the management of acid-base problems in acutely ill patients. They suggest that patients with nearly normal venous blood gases are unlikely to have severe acid-base disturbances, while extremely abnormal venous blood gases reflect comparable arterial blood gas abnormalities. 11 references 1/86-#23

20. VENOUS pH CAN SAFELY REPLACE ARTERIAL pH IN THE INITIAL EVALUATION OF PATIENTS IN THE EMERGENCY DEPARTMENT Kelly A.M., et al, Emerg Med J 18:340, September 2001 BACKGROUND: Although several small studies have reported that the venous pH is a reliable substitute for the arterial pH, most patients requiring evaluation of acid-base status undergo arterial sampling. Arterial puncture is painful and carries a risk of complications for the patient, as well as potential needlestick injury for the healthcare provider. METHODS: This prospective Australian study compared venous and arterial pH values in ED patients considered by their managing physician to require arterial blood gas analysis due to an acute respiratory condition (196) or suspected metabolic derangement (50). RESULTS: The mean arterial pH in these patients was 7.38, with values ranging from 7.05 to 7.61. Alkalosis was present in 17% of the patients, and acidosis in 27%. The correlation between the venous and arterial pH was very close (r=0.92), and the average difference between the two values was only -0.04 units (range, -0.16 to +0.06). Bias plot evaluation also demonstrated good agreement between arterial and venous pH, with the 95% limits of agreement being -0.11 to +0.04 units. CONCLUSIONS: In these acutely ill ED patients, there was a high degree of correlation and agreement between venous and arterial pH. 10 references 3/02 - #20

21. CAN PERIPHERAL VENOUS BLOOD GASES REPLACE ARTERIAL BLOOD GASES IN EMERGENCY DEPARTMENT PATIENTS? Rang, L.C.F., et al, Can J Emerg Med 4(1):7, January 2002 METHODS: In this prospective study, from Queens University and the University of Ottawa in Canada, arterial and venous blood samples were drawn from 218 ED patients requiring blood gas analysis to determine if the venous pH, pCO2 and bicarbonate are predictive of arterial values. Forty-five board-certified emergency physicians were surveyed to determine what values they would consider to represent clinically important differences between venous and arterial pH, pCO2 and HCO3. RESULTS: There was excellent correlation between venous and arterial pH, pCO2 and HCO3. The mean differences between venous and arterial values were small (0.036 units for pH, 6mm Hg for pCO2 and 1.5mEq/l for HCO3). The 26 emergency physicians responding to the survey felt that about 0.05 units for pH, 6.6mm Hg for pCO2 and 1.5mEq/l for HCO3 were clinically significant differences. Sixty-six percent of the pH values, 51% of the pCO2 values and 87% of the HCO3 values fell within these "acceptable" ranges. CONCLUSIONS: There was excellent correlation between arterial and venous blood gas samples in these ED patients. However, the emergency physicians who were surveyed felt that the actual differences were too large to allow the substitution of venous for arterial gases. The authors suggest that venous values might be valid if used with a correction factor, or to follow trends. 12 references 6/02 - #20

22. VENOUS pCO2 AND pH CAN BE USED TO SCREEN FOR SIGNIFICANT HYPERCARBIA IN EMERGENCY PATIENTS WITH ACUTE RESPIRATORY DISEASE Kelly A.M., et al, J Emerg Med 22(1):15, 2002 BACKGROUND: Several small studies have suggested that arterial pH and pCO2 can be estimated from venous blood, but most of these were not conducted in patients with acute respiratory illness. METHODS: This Australian study compared arterial and peripheral venous pH and pCO2 in paired blood samples obtained from 196 ED patients with acute respiratory signs and symptoms. Significant hypercarbia was defined as an arterial pCO2 of 50mm Hg or higher. RESULTS: Significant hypercarbia was identified in 29% of the patients. Arterial pH values ranged between 7.15 and 7.61, and arterial pCO2 ranged between 50-147mm Hg. There was very good agreement between arterial and venous pH, with the latter averaging 0.03 units lower than the former (95% limits of agreement [minus] 0.10 to [plus] 0.04). On average, the venous pCO2 was 5.8mm Hg higher than the arterial pCO2 (95% limits [minus] 8.8mm Hg to [plus] 20.5mm Hg). As a screen for hypercarbia, a venous pCO2 cut-off of 45mm Hg had a sensitivity of 100% and a specificity of 57%. CONCLUSIONS: Among patients with acute respiratory illness, venous pH correlates well with arterial pH. Levels of agreement between the arterial and venous pCO2 were less optimal in many individuals. The authors suggest that a combination of venous blood gas analysis and pulse oximetry can be a useful alternative to routine arterial blood gas analysis in such patients. 14 references 5/02 - #37

How well do venous gases correlate with arterial blood gases in DKA?
In the first study, the author suggests that there is reasonable evidence that the venous and arterial pH are clinically interchangeable in hemodynamically stable DKA patients without respiratory failure. However, the study consisted of only 21 patients.

In the second paper, the authors, from the University of Pennsylvania, reviewed the available literature to determine the clinical utility of ABG measurement in the initial management of DKA. The results suggest that venous pH reliably reflects arterial pH, that differences between the two are not likely to be clinically significant, and that it seems reasonable to substitute venous blood sampling for arterial blood sampling in patients with suspected DKA.

In the third study, the authors suggest that venous blood may be used as an alternative in patients with uremia or DKA, and that the addition or subtraction of the mean differences noted in their study will provide an accurate approximation of arterial pH and bicarbonate levels.

The final study in this section, a convenience sample of 38 patients with 44 episodes of DKA, suggests that peripheral venous blood gases may be a reliable substitute for arterial blood gases in the initial assessment of patients with suspected DKA.



23. THE CASE FOR VENOUS RATHER THAN ARTERIAL BLOOD GASES IN DIABETIC KETOACIDOSIS Kelly, A.M., Emerg Med Australasia 18(1):64, February 2006

BACKGROUND: Although arterial blood gas sampling continues to be recommended for the assessment of acid-base status in patients with diabetic ketoacidosis (DKA), there is some evidence to suggest that pH and, possibly, bicarbonate can be accurately estimated in venous blood samples. METHODS: This Australian author summarized the evidence from six studies (1,022 patients) evaluating the agreement between arterial and venous pH, four of which (784 patients) also evaluated the agreement between arterial and venous bicarbonate. RESULTS: Three of the six studies assessing pH values included patients with DKA. In these studies the weighted average difference between arterial and venous pH was 0.02 units (arterial higher than venous) (95% limits of agreement -0.009 to +0.021 units). In the other three studies that included patients with mixed respiratory and metabolic illness and respiratory failure, the weighted average difference was 0.037 units (95% limits of agreement -0.11 to +0.04 units). Only one of the four studies of agreement in bicarbonate values specified a DKA subgroup (21 patients). The weighted average difference was -1.88mEq/L in the DKA subgroup and -0.99mEq/L in the remaining patients (95% limits of agreement -2.73 to +5.13mEq/L). CONCLUSIONS: The author suggests that there is reasonable evidence that the venous and arterial pH are clinically interchangeable in hemodynamically stable DKA patients without respiratory failure. Results regarding agreement between arterial and venous bicarbonate in DKA patients should be viewed with caution, as they are based on findings in only 21 patients. 15 references 7/06 - #11



24. ARTERIAL BLOOD GAS ANALYSIS: ARE ITS VALUES NEEDED FOR THE MANAGEMENT OF DIABETIC KETOACIDOSIS? Kreshak, A., et al, Ann Emerg Med 45(5):550, May 2005

BACKGROUND: The American Diabetes Association advises an initial arterial blood gas (ABG) in patients with suspected diabetic ketoacidosis (DKA) to determine the severity of acidemia, but also recommends serial measurement of venous pH to monitor the effects of treatment. METHODS: The authors, from the University of Pennsylvania, reviewed the available literature to determine the clinical utility of ABG measurement in the initial management of DKA and whether the venous pH reliably reflects arterial pH as a measure of acid-base status. Four relevant papers were identified. RESULTS: In a study of 200 patients with suspected DKA, the ABG influenced the final diagnosis in 1% of cases, altered treatment in 3.5% (five of these seven involved a change in the route of insulin administration based on pH) and changed disposition in 1%. Venous pH closely reflected arterial pH, typically being 0.015 points lower. The second study compared arterial and venous blood gases in 100 patients with uremia, 21 with DKA and 31 normal controls, and found that venous pH was typically 0.05 points lower than arterial pH. The third study compared pretreatment venous and arterial pH in 44 patients with suspected DKA, and reported a close correlation between the two values and an average difference of 0.03. The fourth study compared arterial and finger capillary blood gas results in 20 patients with DKA, and reported a close correlation between these values (mean difference, 0.03). CONCLUSIONS: The results of this review suggest that the venous pH reliably reflects arterial pH, that differences between the two are not likely to be clinically significant, and that it seems reasonable to substitute venous blood sampling for arterial blood sampling in patients with suspected DKA. 5 references 9/05 - #22



25. COMPARISON OF BLOOD GAS AND ACID-BASE MEASUREMENTS IN ARTERIAL AND VENOUS BLOOD SAMPLES IN PATIENTS WITH UREMIC ACIDOSIS AND DIABETIC KETOACIDOSIS IN THE EMERGENCY ROOM Gokel, Y., et al, Am J Nephrol 20(4):319, July-August 2000

METHODS: The authors of this Turkish study measured both bicarbonate concentrations and pH in simultaneously obtained arterial and venous blood samples in 100 uremic, acidotic patients with end-stage renal failure, 21 patients with DKA, and 31 healthy controls. RESULTS: Among both the uremic and the DKA patients, there was almost perfect correlation between venous and arterial measurements of both pH and bicarbonate levels (r=0.98-0.99). The venous pH was consistently lower than the arterial pH, by a mean of 0.04-0.05 (range, 0.01-0.10), while bicarbonate levels were higher in venous than arterial samples, by a mean of just under 2mmol/L (range 0.6-2.8mmol/L). These relationships were very similar for both groups of patients. In the healthy controls, on the other hand, the correlation between arterial and venous pH and bicarbonate levels was far less strong (r=0.58). CONCLUSIONS: Arterial puncture for the measurement of acid-base status may be associated with complications. The authors suggest that venous blood may be used as an alternative in patients with uremia or DKA, and that the addition or subtraction of the mean differences noted in their study will provide an accurate approximation of arterial pH and bicarbonate levels. 19 references 2/01 - #25



26. COMPARISON OF ARTERIAL AND VENOUS BLOOD GAS VALUES IN THE INITIAL EMERGENCY DEPARTMENT EVALUATION OF PATIENTS WITH DIABETIC KETOACIDOSIS Brandenburg, M.A., et al, Ann Emerg Med 31(4):459, April 1998

BACKGROUND: Patients with diabetic ketoacidosis (DKA) generally undergo venipuncture at the initial evaluation for determination of blood chemistry, as well as arterial puncture for blood gas measurement. Studies in patients without DKA suggest a close correlation between venous and arterial blood gases. METHODS: The authors, from the University of Oklahoma Health Sciences Center, compared venous and arterial blood gas values in a convenience sample of 38 patients with 44 episodes of DKA (defined as an arterial pH below 7.35 or a serum CO2 below 20mmol/L, serum glucose above 250mg/dL, and positive serum ketones). RESULTS: Close correlation was observed between the pH of arterial and venous blood (mean difference, 0.03, range 0-0.11, r=0.97), as well as between arterial and venous HCO3 (r=0.95), and between arterial HCO3 and venous CO2 (r=0.90). The venous pH differed from the arterial pH by more than 0.10 in only one of the 44 episodes (in which the difference was only 0.11 and was not believed to be associated with diagnostic or therapeutic significance). CONCLUSIONS: Findings in this small convenience sample require validation, but suggest that peripheral venous blood gases may be a reliable substitute for arterial blood gases in the initial assessment of patients with suspected DKA. 39 references 8/98 - #19



In the setting of cardiac arrest, can arterial gases be expected to accurately reflect tissue oxygenation?
The first study, from the Chicago Medical School, compared arterial and venous blood gases in 17 episodes of cardiac arrest induced in twelve pigs. They concluded that arterial blood gas analysis does not reflect the acid-base status of systemic tissue during CPR.

The second paper is a clinical study, from the University of Health/Science Chicago Medical School, which compared arterial and mixed venous blood gas parameters during cardiopulmonary resuscitation in 16 patients. The data suggest that respiratory acidosis occurs during cardiopulmonary resuscitation, which may be reflected in mixed venous but not in arterial blood gases. The findings indicate that acid-base management of cardiopulmonary arrest based on ABGs may be inappropriate.

In the final paper of this section, the authors analyzed arterial and venous blood gas changes during CPR in 28 pigs. They found evidence of respiratory alkalosis in arterial blood, reflecting a decrease in pulmonary blood flow, and an acidosis in mixed venous blood, reflecting accumulation of CO2 proximal to the alveolar capillary bed. This mixed venous acidosis may more accurately represent systemic acid-base status.



27. ARTERIOVENOUS CARBON DIOXIDE AND pH GRADIENTS DURING CARDIAC ARREST Grundler, W., et al, Circulation 74(5):1071, November 1986

The American Heart Association has recommended that arterial blood gases be employed to determine the need for bicarbonate therapy during resuscitation from cardiac arrest. This study, from the Chicago Medical School, compared arterial and venous blood gases in 17 episodes of cardiac arrest induced in twelve pigs. The animals were subjected to periods of ventricular fibrillation followed by countershock and mechanical chest compression after the onset of EMD, which occurred in each case. Serial analysis of arterial and venous blood gases demonstrated a decrease in arterial pCO2, from 37.4mm Hg at baseline to 20.1mm Hg two minutes after the onset of EMD, and an increase in pH from 7.46 to 7.54. However, venous pCO2 was observed to increase from 45.2mm Hg to 54.2mm Hg, and venous pH decreased from 7.41 to 7.31. This widening of the normal venoarterial pCO2 and pH gradients was related to decreased pulmonary clearance of CO2 due to reduced pulmonary blood flow. The venoarterial pCO2 and pH gradients normalized in animals that were successfully resuscitated. The venous acidemia observed in the present study is considered to be a reflection of tissue acid-base conditions. Similar findings have also been reported in humans undergoing CPR. It is concluded that arterial blood gas analysis does not appear to reflect the acid-base status of systemic tissue during CPR. The authors also suggest that, on the basis of these findings, bicarbonate administration, which increases the CO2 load, would not appear to be routinely indicated during CPR. 14 references 4/87-#9



28. DIFFERENCE IN ACID-BASE STATUS BETWEEN VENOUS AND ARTERIAL BLOOD DURING CARDIOPULMONARY RESUSCITATION Weil, M.H., et al. N Engl J Med 315(3):153 July 17, 1986

This clinical study, from the University of Health/Science Chicago Medical School, compared arterial and mixed venous blood gas parameters during cardiopulmonary resuscitation in 16 patients hospitalized in an intensive or cardiac care unit. Blood samples were obtained from five to 79 minutes after the onset of the arrest (median, 23 minutes). The average dose of alkali given prior to blood sampling was 130 +/- 30mEq. The average arterial pH and pCO2, values during resuscitation (7.41 and 32mm Hg. respectively) did not differ significantly from prearrest values, which were available in 13 patients (7.37 and 33mm Hg. respectively). In contrast, during resuscitation the average mixed venous pH and pCO2 values were significantly altered from prearrest values (7.31 and 44mm Hg. respectively, prior to the arrest. vs. 7.15 and 74mm Hg. respectively, during resuscitation), and were significantly different from arterial values. Although arterial lactate levels increased significantly (p <0.05), the bicarbonate concentrations in arterial and mixed venous blood did not differ significantly. Overall, the venous-arterial pH gradient increased from 0.06 +/- 0.02 prior to arrest to 0.30 +/- 0.05 during resuscitation (p<0.001), and the pCO2, gradient increased from 11 +/- 2mm Hg to 36 +/- 6mm Hg (p<0.001). These data confirm the findings of previous animal studies, and suggest that respiratory acidosis occurs during cardiopulmonary resuscitation, which may be reflected in mixed venous but not in arterial blood gases. The findings indicate that acid-base management of cardiopulmonary arrest based on ABGs may be inappropriate. 12 references 1/87-#10



29. ARTERIAL BLOOD GASES FAIL TO REFLECT ACID-BASE STATUS DURING CARDIOPULMONARY RESUSCITATION: A PRELIMINARY REPORT Weil, M.H., et al, Crit Care Med 13(11):884, November 1985

The authors, from The Chicago Medical School, analyzed arterial and venous blood gas changes during CPR in 28 pigs. PaCO2 in the animals was maintained at 35-45torr. Determination of arterial and mixed venous blood gases before induction of cardiac arrest and after the onset of EMD demonstrated a significant increase in the venoarterial PCO2 and pH gradients throughout the period of EMD (p<0.001), associated with a progressive increase in arterial lactate levels. In surviving animals, the arterial pH increased from 7.47 at baseline to 7.53 after four minutes of EMD, while the mixed venous pH decreased from 7.42 to 7.26. Arterial pCO2 decreased from 38.3 torr to 22.5 torr while the mixed venous pCO2 increased from 44.8 torr to 57.6 torr. Similar changes were noted in nonsurvivors. After successful resuscitation, reversal of these gradients was observed. These findings confirm observations by the authors during previous CPR and cardiac arrest studies, as well as data obtained in human patients, and indicate that a paradoxical acidosis in mixed venous blood and alkalosis in arterial blood may be noted during resuscitation from cardiac arrest. The evidence of respiratory alkalosis in arterial blood may reflect a decrease in pulmonary blood flow and a subsequent increase in the ventilation/perfusion ratio, while the acidosis in mixed venous blood may reflect accumulation of CO2 proximal to the alveolar capillary bed, which more accurately represents systemic acid-base status. These findings suggest that during CPR administration of sodium bicarbonate based on arterial blood gas values may be inappropriate. 4 references 5/86-#11



How do skin pigmentation, nail polish, and acrylic nails affect pulse oximetry?
The following studies show that:

1. Pulse oximetry recordings are not significantly influenced by dark skin pigmentation in relatively stable but critically ill adults.

2. Nail polish is rarely associated with clinically significant deviations in pulse oximeter readings, although polish removal might be helpful in some situations.

3. Acrylic finger nails may impair the measurement of oxygen saturation, depending on the pulse oximeter used, and may cause significant inaccuracy. Hence, removal of artificial acrylic finger nails may assure an accurate and precise measurement with pulse oximetry.



30. ACCURACY OF PULSE OXIMETRY IN PIGMENTED PATIENTS Bothma, P.A., et al, S Afr Med J 86(5):594, May 1996

BACKGROUND: Prior studies of the influence of skin pigmentation on the accuracy of pulse oximetry have yielded conflicting results. METHODS: This prospective South African study compared the readings of three pulse oximeters (Simed S100e [Simed Co.], Nihon Koden [Nihon Koden Corp.], and Ohmeda 3740 [Ohmeda]) with arterial oxygen saturation by cooximetry in 100 consecutive darkly pigmented critically ill adults. Finger probes were employed with all three pulse oximeters, and an ear probe was also used with the Ohmeda model. RESULTS: Arterial oxygen saturations were between 88-99% (median, 96%), with pulse oximeter readings between 86-100%. The limits of agreement between the two techniques were 3.4-4.5% with the Ohmeda finger probe, 3.8-5.8% with the Ohmeda ear probe, 4.1-5.8% with the Nihon Koden, and 2.6-5.0% with the Simed, all of which were considered to be clinically acceptable. The 95% confidence intervals were small. The precision (standard deviation of the differences) ranged between 1.9 and 2.4%, and bias (mean of the differences) ranged between -1.0 and 1.2%. When the arterial oxygen saturation was below 92%, there was a trend towards lower values with pulse oximetry. Pulse oximetry results tended to be higher than arterial oxygen saturation with the finger probes, and lower with the Ohmeda ear probe. These differences were small and not considered clinically significant. CONCLUSIONS: Pulse oximetry recordings were not significantly influenced by dark skin pigmentation in these relatively stable but critically ill adults. 25 references 1/97 - #35



31. EFFECT OF NAIL POLISH ON OXYGEN SATURATION DETERMINED BY PULSE OXIMETRY IN CRITICALLY ILL PATIENTS Hinkelbein, J., et al, Resuscitation 72:82, January 2007

BACKGROUND: The belief that use of nail polish can be associated with misleading results during pulse oximetry monitoring has not been extensively investigated. METHODS: These German authors examined the effects of nine nail polish colors on pulse oximetry readings in 50 adult Caucasian ICU patients undergoing mechanical ventilation. A different color of polish, ranging over the entire color spectrum from black to clear, was applied on each of nine fingernails and one fingernail without nail polish served as a control. Pulse oximetry readings were obtained using a SIEMENS pulse oximeter monitor SC1281 and a NELLCOR DS-100A Durasensor finger sensor probe. Pulse oximetry readings were compared with arterial blood gas results. RESULTS: There was a close correlation between the mean arterial and control pulse oximetry readings (difference +0.2%). The mean difference between the readings was greatest in the setting of black nail polish (difference 1.6%, 95% confidence interval [CI] -4.1% to +10.6%), purple polish (difference 1.2%, 95% CI -3.1% to +12.5%) and dark blue polish (1.1%, 95% CI -5.7% to +9.1%), and was less than 1% with all other polish colors. For all colors, the mean difference between arterial and pulse oximetry readings was within the oximeter manufacturer's reported deviation of plus/minus 2% in the oxygen saturation range of 70-100%. In general, pulse oximetry tended to underestimate arterial oxygenation results. CONCLUSIONS: Nail polish was rarely associated with clinically significant eviations in pulse oximeter readings, although polish removal might be helpful in some situations. 34 references 6/07 - #34



32. ARTIFICIAL ACRYLIC FINGERNAILS MAY ALTER PULSE OXIMETRY MEASUREMENT Hinkelbein J, Koehler H, Genzwuerker HV, Fiedler F. Resuscitation. 2007 Jul;74(1):75-82. Epub 2007 Mar 13.

INTRODUCTION: Pulse oximetry is the most common technique to monitor oxygen saturation (SpO(2)) during intensive care therapy. However, intermittent co-oximetry is still the "gold standard" (SaO(2)). Besides acrylic nails, numerous other factors have been reported to interfere with pulse oximetry. Data of measurements with artificial finger nails are not sufficiently published. MATERIALS AND METHODS: A prospective clinical-experimental trial in mechanically ventilated and critically ill patients of an ICU was performed. Patients were randomly assigned to either group S (S: Siemens pulse oximeter) or group P (P: Philips pulse oximeter) prior to the measurements. SpO(2) was determined in each patient three times alternately in standard ((N)SpO(2)) and sideways position at the natural nail ((N90)SpO(2)). For the reference measurements oxygen saturation was measured by means of a haemoximeter (co-oximetry). Thereafter, SpO(2) was obtained at the acrylic finger nail in the same way ((A)SpO(2) and (A90)SpO(2)). Bias was calculated as DeltaS=(N)SpO(2)-SaO(2) and DeltaS=(A)SpO(2)-SaO(2). Accuracy (mean difference) and precision (standard deviation) were used to determine the measurement discrepancy. P<0.05 was considered significant. RESULTS: Accuracy and precision without acrylic nails applied were comparable to SaO(2) in both groups (n.s.). With acrylic nails applied a bias of DeltaS=-1.1+/-3.14% for group S (P=0.00522) and a bias of DeltaS=+0.8+/-3.04% for group P was calculated (n.s.). CONCLUSION: Acrylic finger nails may impair the measurement of oxygen saturation depending on the pulse oximeter used and may cause significant inaccuracy. Hence, removal of artificial acrylic finger nails may be helpful to assure an accurate and precise measurement with pulse oximetry.



How do abnormal hemoglobins or infusion of hydroxocobalamin affect oximetry?
Hydroxocobalamin is widely used in Europe as an alternative to other cyanide antidotes for the treatment of patients with cyanide poisoning associated with smoke inhalation. Increasing hydroxocobalamin concentrations are associated with a progressive increase in the oximetric reading for carboxyhemoglobin and methemoglobin, and a corresponding decrease in the oxygen saturation reading.

Unlike standard oximetry, the Rainbow-SET Rad-57 Pulse CO-Oximeter (Masimo) is an eight-wavelength pulse oximeter that is capable of measuring more than two species of human hemoglobin. At the carboxyhemoglobin and methemoglobin levels achieved in healthy volunteers, the Rad-57 device appeared to be reliable. These results cannot be extrapolated to the setting of critical illness or more severe dyshemoglobinemia (or respiratory co- morbidity).



33. POTENTIAL INTERFERENCE BY HYDROXOCOBALAMIN ON COOXIMETRY HEMOGLOBIN MEASUREMENTS DURING CYANIDE AND SMOKE INHALATION TREATMENTS Lee, J., et al, Ann Emerg Med 49(6):802, June 2007

BACKGROUND: Hydroxocobalamin is widely used in Europe as an alternative to other cyanide antidotes for the treatment of patients with cyanide poisoning associated with smoke inhalation. It might be postulated, however, that the intense light absorption produced by hydroxocobalamin can interfere with light source-based co-oximetric blood gas measurements. METHODS: The authors, from the University of California, Irvine, examined the effect of hydroxocobalamin infusion (average dose, 625mg infused over 100 minutes) on oximetry readings in rabbits. RESULTS: Increasing hydroxocobalamin concentrations were associated with a progressive increase in the oximetric reading for carboxyhemoglobin and methemoglobin, and a corresponding decrease in the oxygen saturation reading. Similar patterns were observed in an in vitro study of whole blood samples with different hydroxocobalamin concentrations. Increasing hydroxocobalamin levels were associated with a change of up to (- )7.9% in the oxygen saturation value and (+)14.7% in the carboxyhemoglobin value. CONCLUSIONS: The authors advise caution in the interpretation of hemoglobin fractions measured with co-oximetry when hydroxocobalamin is used for the treatment of cyanide poisoning and smoke inhalation. The effects of this spectral overlap might be particularly relevant in patients receiving continuous hydroxocobalamin infusion and when blood sampling is performed immediately after bolus injection or "upstream" from the hydroxocobalamin infusion site. 10 references 10/07 - #34



34. MEASUREMENT OF CARBOXYHEMOGLOBIN AND METHEMOGLOBIN BY PULSE OXIMETRY: A HUMAN VOLUNTEER STUDY Barker, S.J., et al, Anesthesiology 105(5):892, November 2006

BACKGROUND: Standard pulse oximeters estimate arterial hemoglobin saturation via measurement of tissue light transmission at two wavelengths. Several studies have reported serious errors in estimation of oxygen saturation in the presence of a dyshemoglobinemia. The Rainbow-SET Rad-57 Pulse CO-Oximeter (Masimo) is an eight-wavelength pulse oximeter that is capable of measuring more than two species of human hemoglobin. METHODS: The authors, from the University of Arizona and partially funded by Masimo, examined the performance of the Rad-57 device in ten healthy volunteers exposed to carbon monoxide to produce a carboxyhemoglobin level of 15%, and ten volunteers given an IV infusion of sodium nitrite to induce methemoglobinemia at a level between 5-12%. Pulse oximetry results obtained with the Rad-57 were compared with measurements performed in arterial blood. RESULTS: At a carboxyhemoglobin level of 0-15%, the uncertainty (a measure of test imprecision) of carboxyhemoglobin measurement with the Rad-57 was +/- 2%. At a methemoglobin level of 0-12%, the uncertainty with the Rad-57 was +/- 0.5%. For comparison purposes, the uncertainty of most standard pulse oximeters for measurement of oxygen saturation at values of 70-100% is +/- 2%. CONCLUSIONS: At the carboxyhemoglobin and methemoglobin levels achieved in these healthy volunteers, the Rad-57 device appeared to be reliable. The authors acknowledge that these results cannot be extrapolated to the setting of critical illness of more severe dyshemoglobinemia (or respiratory co- morbidity). 17 references 3/07 - #34



Is room-air pulse oximetry sensitive for hypercapnia?
In this case-control study, the frequency of hypercapnia was inversely related to the oxygen saturation on room air pulse oximetry. At an oxygen saturation cut-off of 96%, room air pulse oximetry appears to be a sensitive screening test for moderate hypercapnia.



35. THE SENSITIVITY OF ROOM-AIR PULSE OXIMETRY IN THE DETECTION OF HYPERCAPNIA Witting, M.D., et al, Am J Emerg Med 23:497, July 2005

BACKGROUND: The suggestion that pulse oximetry does not reliably identify hypercapnia has been based on studies conducted in patients receiving supplemental oxygen. One small study reported that, at a cut-off of 96%, room-air pulse oximetry detected all twelve patients with moderate hypercapnia. METHODS: In this case-control study, from the University of Maryland, the charts of 92 patients with hypercapnia and 257 patients without hypercapnia were reviewed to assess the diagnostic utility of an oxygen saturation of 96% or lower on room-air pulse oximetry for the detection of moderate hypercapnia (PaCO2 above 50mm Hg). All of the patients had pulse oximetry performed on room air followed within eight hours by measurement of arterial blood gases (this interval was less than two hours in 43% of patients and 5-8 hours in 12%). RESULTS: The frequency of hypercapnia was inversely related to the oxygen saturation on room air pulse oximetry. At an oxygen saturation cut-off of 96%, the sensitivity and specificity of room air pulse oximetry for the identification of moderate hypercapnia were 96% and 39%, respectively, the positive likelihood ratio (LR) was 1.6, and the negative LR was 0.1. There were four falsely-negative cases. CONCLUSIONS: The authors acknowledge the limitations of their study design but suggest that, at an oxygen saturation cut-off of 96%, room air pulse oximetry appears to be a sensitive screening test for moderate hypercapnia. 14 references 12/05 - #36



How long does it take for the equilibration of oxygen saturation using pulse oximetry?
For patients placed on supplemental oxygen, the mean interval between initiation of O2 and equilibration was 2.8 minutes. The mean times to equilibration in patients with and without COPD or asthma were 3.5 and 2.5 minutes, respectively.



36. TIME TO EQUILIBRATION OF OXYGEN SATURATION USING PULSE OXIMETRY Gruber, P., et al, Acad Emerg Med 2(9):810, September 1995

BACKGROUND: Although it is generally believed that equilibration of the oxygen saturation after changing the inspired oxygen concentration (FiO2) requires 20-30 minutes, the actual time required for equilibration in patients with active medical problems has not been clearly defined. METHODS: This prospective study, from Long Island Jewish Medical Center in New Hyde Park, NY, examined the actual time required for equilibration of O2 saturation after changing the FiO2 in 51 adults requiring supplemental oxygen (2-4L/min by nasal cannula) in the ED due to acute cardiac and/or pulmonary conditions. Equilibration times were calculated on the basis of O2 saturation measurements made at one-minute intervals for 30 minutes after a change was made in FiO2. RESULTS: For patients placed on supplemental oxygen (43 measurements), the mean room air oxygen saturation was 89.8%, the mean equilibration point was 95.3%, and the mean interval between initiation of O2 and equilibration was 2.8 minutes (95% confidence interval, 2.3-3.3 minutes). The mean times to equilibration in patients with and without COPD or asthma were 3.5 and 2.5 minutes, respectively (p=0.01). Upon discontinuation of supplemental oxygen (18 measurements), the mean initial oxygen saturation was 96.6%, the mean equilibration point was 91.7%, and the mean interval to equilibration was 3.1 minutes (95% CI, 2.3-4.4 minutes), or 4.6 minutes in patients with COPD or asthma compared with 2.7 minutes in patients without these conditions. One patient in each group demonstrated continuous variability in O2 saturation throughout the study. CONCLUSIONS: Times to equilibration after making a change in inspired O2 concentration appear to be substantially shorter than traditionally believed. Equilibration times in patients with COPD or asthma may be longer than in other patients. 15 references 12/95 - #38



Should patients who are hyperventilating be given a paper bag to breathe into?
No. The author, from the University of California in San Francisco, reports three cases in which collapse occurred in patients with myocardial ischemia or hypoxemia who were erroneously believed to be hyperventilating and were instructed to breathe into a paper bag. In a subsequent volunteer study, paper bag rebreathing was associated with a substantial decrease in inspired oxygen.



37. HYPOXIC HAZARDS OF TRADITIONAL PAPER BAG REBREATHING IN HYPERVENTILATING PATIENTS Callaham, M., Ann Emerg Med 18(6):622, June 1989

Patients who are hyperventilating are often managed with paper bag rebreathing. The author, from the University of California in San Francisco, reports three cases in which collapse occurred in patients with myocardial ischemia or hypoxemia who were erroneously believed to be hyperventilating and were instructed to breathe into a paper bag. In two of the cases, this treatment was initiated by paramedics or by a dispatcher, and in the remaining case treatment was initiated by a nurse. Two of the three patients sustained irreversible cardiopulmonary arrests. In a subsequent study, 20 healthy adults hyperventilated to an end-tidal CO2 of 20mm Hg, and were then instructed to hyperventilate into a paper bag containing sensors for the purpose of monitoring gas concentrations. Varying inspired concentrations of CO2 (FiCO2) were achieved with three minutes of bag rebreathing. The FiCO2 did not reach 40mm Hg for 30% of the subjects, and did not reach 35mm Hg for 22%, suggesting that bag rebreathing appears to be an unreliable method of achieving an elevated FiCO2. The inspired concentration of oxygen (FiO2) progressively declined over three minutes of bag rebreathing, to a mean maximal decrease in O2 of 26mm Hg. The decrease in O2 was 34mm Hg in four subjects, and 42mm Hg in one subject. The author suggests that paper bag rebreathing may be associated with a substantial decrease in inspired oxygen that may pose a threat to hypoxic patients. It is further suggested that this technique not be employed unless myocardial ischemia can be ruled out and direct measurement of the patient's oxygenation has been performed. The author cautions against the recommendation and use of this technique by prehospital personnel. 37 references 11/89 - #34



KEY POINTS AND RECOMMENDATIONS
1. Peripheral venous blood gases may be a reliable substitute for arterial blood gases in the initial assessment of patients with suspected DKA.

2. Arterial blood gas analysis does not reflect the acid-base status of systemic tissue during CPR. Respiratory alkalosis in arterial blood reflects a decrease in pulmonary blood flow. Acidosis in mixed venous blood reflects accumulation of CO2 proximal to the alveolar capillary bed. This mixed venous acidosis may more accurately represent systemic acid-base status.

3. Pulse oximetry recordings are not significantly influenced by dark skin pigmentation in stable but critically ill adults.

4. Nail polish is rarely associated with clinically significant deviations in pulse oximeter readings, although polish removal might be helpful in some situations.

5. Acrylic finger nails may impair the measurement of oxygen saturation depending on the pulse oximeter used and may cause significant inaccuracy. Hence, removal of artificial acrylic finger nails may assure an accurate and precise measurement.

6. The Rainbow-SET Rad-57 Pulse CO-Oximeter (Masimo) is an eight-wavelength pulse oximeter capable of measuring carboxyhemoglobin and methemoglobin levels.

7. If someone tries to put a paper bag over your head, don’t let them.



Stephen Colucciello MD

Associate Chair, Department of Emergency Medicine

Carolinas Medical Center