Oxygen Saturation ================= þ Assuming 37øC, PaCO2=40mmHg: SaO2 PaO2@pH7.3 PaO2@pH7.4 PaO2@pH7.5 -------------------------------------------------- 99 171 158 143 98 122 111 101 97 101 92 84 96 89 82 74 95 82 74 68 94 76 69 63 93 72 66 60 92 68 62 57 91 66 60 55 90 63 58 53 88 59 54 49 86 56 51 47 84 53 49 44 82 51 47 42 80 49 45 41 78 47 43 39 76 45 41 38 74 44 40 35 72 42 39 34 þ Oxygen Saturation and CO Poisoning - Summary: Pulse oximeters cannot distinguish between carboxy and oxyhaemoglobin and so give you the total score. Both these agents have the same absorbance profile on the wave lengths used. The PO2 is either measured or calculated depending on your machine. Calculated ones take an HBO2 dissociation curve, plot the SaO2 and then read the result. Measured ones do it directly with a seperate electrode and are therefore accutate. If your machine can measure CO-HB directly it should be able to do the appropriate calcultion of PO2 by firstly subtracting the %sat of CO-Hb. Met-Hb will convert all the SaO2s to 85% when it reaches significant levels. - Pulse ox: reflects oxygenated hemoglobin plus carbon monoxide bound to hemoglobin; can correlate pulse ox and ABG saturation to estimate %COHgB [Buckley RG, Aks SE, Eshom JL, Rhydman R, Schaider J, Shayne P. The pulse oximetry gap in carbon monoxide intoxication. Ann Emerg Med August 1994;24:252-255.] þ Oxygen Saturation >The PaO2 & the oxygen saturation measure two different things. A blood gas >analyzer only measures three variables namely pH, PCO2 & PO2; all other >variables in an ABG report are caculated from these. The PO2 is measured >with a Clark (Platinum/ Silver) electrode & only measures the oxygen >dissolved in plasma. This is basically determined by the inspiratory PO2 & >the net diffusion coefficeint for the alveolar/ capillary & endothelial >cell/ plasma interfaces. In normal circumstances oxygen continues to >diffuse into the RBC & onto HB in is in a dynamic equilibrium with oxygen >dissolved free in plasma. > >Oxygen saturation as determined by pulse oximetry gives an indication of the >oxygen content of blood bound to HB. The pulse oximeter measures the >differential absorption of light at 660 & 940 nm expresses then as a ratio, >correlates this experimentally derived values in hypoxic volunteers & then >expresses the saturation as a percentage. Pulse oximeter values below 80% >are the machines geusstimate of what the oxygen saturation should be because >there is a limit to how hypoxic you make someone. > >So it is possible to have a high PO2 if there is normal diffusion of oxygen >into plasma & a low oxygen saturation is the oxygen cannot diffuse into the >RBC & onto HB > >When there is equal absorption of both 660 & 940nm i.e. the ratio = 1 the >oxygen saturation is 85%. In MethHB states there is equal absorption of >both these frequencies so the pulse oximeter will read 85%, the PO2 will be >high (assuming the patient is receiving supplemental O2) but will mask the >fact that the patient will be profoundly hypoxic. HBCO has a similar O2 >absorbance at 660nm c.f. oxy HB so the pulse oximeter wil be fooled into >over stating the true saturation of the patient . This relates to what has been referred to as a 'saturation gap'. I.e., whenever there is a discrepancy between measured and predicted (based on pO2 and the oxyHb dissociation curve) saturation this implies the presence of a confounding 'substance'. Just as an anion gap implies the existence of unmeasured anions and osmolal gap (usually) implies the presence of unmeasured alcohols so too the saturation gap can be used to infer the presence of carboxy, met or sulf Hb. What makes the issue confusing, however, is the different ways in which saturation is reported. My understanding e.g., is that most blood gas machines measure pO2, pCO2 and pH and CALCULATE the percent saturation from the pO2 using the oxyHb dissociation curve. Some machines, however, actually MEASURE saturation (co-oximeters). In the case of carboxyHb the pulse oximeter will be fooled into reporting the SUM of carboxyHb and oxyHb as the percent saturation (5). Thus someone with carbon moxoxide poisoning will appear to have a normal sat by pulse ox. The measured sat using a co-oximeter will be low, however, and a discrepancy between the pO2 (which will be normal in CO poisoning, baring some other respiratory problem) will be revealed as a saturation gap. In this case the gap is between the predicted saturation based upon the pO2 and the sat measured by co-oximetry (not pulse oximetry). MetHbemia presents a more complicated matter as you have observed. Since it absorbs both wavelengths of light commonly used in pulse oximeters, in a 1:1 ratio, the calibration curve will predict a saturation of 85%. However this is only for the portion of Hb that is oxidized to Fe+3. The remainder of the Hb (Fe+2) will absorb the red and infrared lights in proportion to the percent saturation of the Hb THAT IS AVAILABLE FOR TRANSPORT OF OXYGEN. Thus pulse oximetry will give a reading that is the weighted average of the two: Pulse ox = FMet*0.85 + (1-FMet)*F(avail)O2 (2) [NOTE: see below] where FMet is the fraction (percentage) of reported MetHb and F(avail)O2 is the fraction of reduced Hb that is carrying oxygen. As an example (1) suppose someone has a metHb of 38.8% and a co-oximeter saturation of 61.2%. Pulse oximetry should read Pulse ox = 0.388*0.85 + (1-0.388)*F(avail)O2 We need to calculate F(avail)O2 given the percent of total Hb that is saturated i.e., 61.2%. Thus 0.612 = HbO2/(Hb + HbO2) (this we are given) ? = HbO2/((1-0.388)*Hb + HbO2) (this is what we need (F(avail)O2)) so using the first equation we conclude that Hb = (0.388/0.612)*HbO2 = 0.634*HbO2 substituting this into the equation directly above F(avail)O2 = HbO2/(0.612*0.634*HbO2 + HbO2) = 0.72 so Pulse ox = 0.388*0.85 + (1-0.388)*0.72 = 77%. Not bad considering a measured pulse ox of 80% below (1)! The upshot of these calculations is that pulse oximetry may EITHER over or under estimate the true saturation in MetHbemia If saturation is less than 70% pulse oximetry will overestimate it, if greater than 70% it will underestimate it (3). Another situation where they may be a discrepancy between measured and predicted O2 sat is in situations with congenital Hb with low oxygen affinity (such as HB Kansas--I'm sure there's a Hb Brisbane as well ;)). In this case, as in COHbemia, pO2 will be normal but saturation as measured by co-oximetry will be low. The bottom line in all of this is to look for the presence of a saturation gap. If present it is a clue to the presence of a dyshemoglobin (4). By contrast, a saturation gap does not appear to occur with cyanide poisoning (6). I hope that I haven't confused anybody who thought they understood it all before this. H. Louzon MD "They only told me what was going on after I was about to walk out of the hospital." (1) Kubota M; Soma K; Suzuki M; Hanada N; Takada N; Kusuhara N; Kobayashi H; Yanase N; Abe T; Tomita T [Aniline-induced methemoglobinemia monitored by pulse oximetry] Nippon Kyobu Shikkan Gakkai Zasshi 1993 Jul;31(7):886-9 A case of aniline-induced methemoglobinemia is reported. When the pulse oximeter reading (SpO2) was 80%, the oxygen saturation measured by a co-oximeter (SaO2) was 61.2%, the oxygen saturation calculated from PaO2 values was 98.9% and methemoglobin level was 38.8%. After methylene blue injection, methemoglobin level decreased gradually. With a decrease of methemoglobin level, SpO2 approached SaO2. If disparity between SpO2 and the oxygen saturation calculated from PaO2 values is noted, the presence of methemoglobinemia must be suspected. In clinical situations, the pulse oximeter permits the continuous noninvasive monitoring of oxygen saturation. It is necessary, however, to consider the potential errors in pulse oximetry. (2) Alexander et. al. Principles of Pulse Oximetry: Theoretical and Practical Considerations. Anes Analg 1989;68:368-76 (3) Nijland R; Jongsma HW; Nijhuis JG; Oeseburg B; Zijlstra WG Notes on the apparent discordance of pulse oximetry and multi-wavelength haemoglobin photometry. Acta Anaesthesiol Scand Suppl, 107:1995, 49-52 Multi-wavelength photometers, blood gas analysers and pulse oximeters are widely used to measure various oxygen-related quantities. The definitions of these quantities are not always correct. This paper gives insight in the various definitions for oxygen quantities. Furthermore, the possible influences of dyshaemoglobins and fetal haemoglobin on the accuracy of pulse oximetry are discussed. As pulse oximeters are constructed for the determination of arterial oxygen saturation, they should be validated with sample oxygen saturation values and not with the oxyhaemoglobin fraction. The influence of carboxyhaemoglobin is insubstantial over an oxygen saturation range of 0% to 100%. Through the presence of methaemoglobin, pulse oximetry will give an underestimation above 70% and an overestimation below 70% oxygen saturation. The influence of fetal haemoglobin is insignificant in the neonatal use of pulse oximetry, in the range of 75% to 100% arterial oxygen saturation. However, a pulse oximeter underestimates the arterial oxygen saturation at the 25% level with 5%, if the pulse oximeter has been calibrated in human adults. Such a low level of arterial oxygen saturation can be present in the fetus during labor. (4) Watcha MF; Connor MT; Hing AV Pulse oximetry in methemoglobinemia. Am J Dis Child 1989 Jul;143(7):845-7 Pulse oximetry is a major improvement in the assessment of oxygenation. The device uses plethysmography and light absorbance measurements at two wavelengths to estimate oxygen saturation. It is inaccurate, however, when more than two types of hemoglobin are present. This article describes two infants with methemoglobinemia in whom pulse oximetry overestimated oxygen saturation. We discuss the mechanism of this systematic error and emphasize that pulse oximetry should not be used to estimate true oxygen saturation in the presence of methemoglobin. However, a disparity between oxygen saturation estimates by pulse oximetry and by calculations based on the arterial partial pressure of oxygen and the oxygen-hemoglobin dissociation curve can provide an important clue to the presence of such abnormal types of hemoglobins. Therapy should be based on direct measurements of oxyhemoglobin by cooximetry and not on measurements of oxygen saturation by pulse oximetry or on saturations calculated from the Pao2 and the oxyhemoglobin dissociation curve. (5) Vegfors M; Lennmarken C Carboxyhaemoglobinaemia and pulse oximetry [see comments] Br J Anaesth 1991 May;66(5):625-6 We compared measurements obtained with a pulse oximeter (SpO2) against values obtained from a CO-oximeter in a patient with carbon monoxide poisoning. SpO2 was equal to the sum of the oxyhaemoglobin (HbO) and carboxyhaemoglobin (HbCO) values over the range of HbCO from 30 to 1%. This confirms the experimental findings that pulse oximeters measure HbCO as HbO. The patient was treated with oxygen (FlO2 = 50%) and recovered without any sequelae. Under these circumstances, the half-life of HbCO was approximately 2 h. (6) Curry SC; Patrick HC Lack of evidence for a percent saturation gap in cyanide poisoning. Ann Emerg Med, 20: 5, 1991 May, 523-8 STUDY OBJECTIVES: To determine if toxic concentrations of cyanide in blood result in a difference between calculated and measured percent hemoglobin oxygen saturation (percent saturation gap). DESIGN: An in vitro laboratory study. SETTING: Hospital laboratory. TYPE OF PARTICIPANTS: Arterial blood from five stable patients residing in a tertiary hospital ICU. Venous blood samples obtained from five healthy volunteers. INTERVENTIONS: Two-mL aliquots from each blood sample were placed into four test tubes and mixed with phosphate buffer. Cyanide was added to three of the tubes so that the four tubes contained 0, 6, 12, and 25 mg/L cyanide. MEASUREMENTS AND MAIN RESULTS: The percent saturation gap was calculated by subtracting the percent oxyhemoglobin measured on an oximeter from the percent saturation calculated by a blood gas analyzer. Using two-tailed, paired t tests, we could not demonstrate a difference in mean percent saturation gaps between arterial samples or venous samples with and without cyanide. All blood samples had a normal percent saturation gap (P greater than .99 by Fisher's exact test). We had greater than a 95% chance of demonstrating a difference in mean percent saturation gaps of only 0.46% in arterial blood and of 3.3% in venous blood if such a difference existed. CONCLUSIONS: Our results, as well as a review of the literature, indicate that there are no data supporting the suggestion that a percent saturation gap should imply the diagnosis of poisoning by inorganic cyanide. WARNING: don't waste your time reading this if you have better things to do than witness a demonstration in elementary algebra. Someone wrote to me asking for clarification of the calculation on the pulse ox reading in methemoglobemia. In answering that request I noted a mistake in the original calculation related to the formula directly below. ************************************************************************ >>? = HbO2/((1-0.388)*Hb + HbO2) (this is what we need (F(avail)O2)) so >>using the first equation we conclude that >> ************************************************************************* The formula that I used (Pulse ox = FMet*0.85 + (1-FMet)*F(avail)O2) comes from reference 2. What I did was to apply the numbers provided in the case report (reference 1) to this equation to see if the measured pulse ox agreed with the predicted pulse ox calculated on the basis of this equation. TO take a different example, let's suppose that the methemoglobin percentage is 33%, i.e., FMet = 0.33 (I'm going to use round numbers) and that the Hb is 15. (Note that the actual calculation is really independent of Hb and that if the equations are solved symbolically the Hb cancells out, but I'm going to include it for purposes of concretness.) That means that 0.33*15 = 5 gr/dl of hemoglobin is unavailble to carry oxygen and 15 - 5 = 10 gr/dl is. Let's further assume that O2 saturation as *measured* by a blood gas machine is 60%. What will the pulse ox reading be? Well, the contribution to the pulse ox from Methemoglobin (which is really carrying no oxygen is 85%--this is an artifact of the way light is absorbed by methemoglobin and is *always 85% for that porion of the Hb which is oxidized to MetHb). The pulse ox reading will be a weighted sum of the 'saturation' of MetHb (remember that this is *measured* as 85% even though it is really zero) times the percent MetHb AND the percent saturation of the remainder of the Hb times the fraction of Hb that is available for transport of oxygen i.e., Pulse ox = FMet*0.85 + (1-FMet)*F(avail)O2 The first term becomes FMet*0.85 = 0.33*0.85 = 0.28 Now to calculate the second term we need to know what fraction of 'normal' Hb (NOT the MetHb portion) is saturated under the assumption that the TOTAL Hb has a saturation of 60% that was measured by the ABG machine. SO (1-0.33)*HbO2/(HbO2+ Hb) needs to be calculated (my mistake initially was to use total reduced hemoglobin for Hb when, in fact, it represents only the unsaturated portion of reduced hemoglobin in this equation). If 10 gr of 'normal' Hb is available for transport (as above after subtracting off the 5 gr that isn't) then HbO2 = 0.6*15 = 9 gr Hb = (10-9) = 1 gr (this is the unsaturated amount of normal Hb) thus HBO2/(HbO2 + Hb) = 9/(9+1) = 0.90 That is a 60% saturation of total Hb corresponds to a 90% saturation of that portion of the Hb that is available for transport (10 gr). So (1-0.33)*HbO2/(HbO2+ Hb) = 0.67*9/(9+1) = 0.6 This is the contribution to the saturation made by 'normal' Hb in the second term in the equation. Adding the two: Pulse ox = 0.28 + 0.60 = 0.88 = 88% saturation is what the pulse ox will measure. I now realize that I had made an error in the formula demarcated by '**************' above. Hb in that equation should refer to the portion of reduced Hb that is desaturated and not equal to (1-FMet)*Hb but rather 0 in that particular example since, it turns out, the reduced Hb was 100% saturated. Reduced Hb will be 100% saturated if the measured SaO2 + MetHb = 100% which it was in the first example. The predicted and measured sauration are thus seen to be further apart than origionlly estimated and this is a known limitation of this method of approximation. In fact when I reviewed the formula in that article I was somewhat surprised that the second term was not further simplified. I.e., (1-FMet)*HbO2/(HbO2 + Hb) where Hb = deoxygenated reduced hemoglobin. let measured ABG saturation (of the whole sample) be SaO2 then HbO2 = HbT*SaO2 where HbT is TOTAL hemoglobin Hb would then be the difference between total unsaturated hemoglobin and methemoglobin i.e., Hb = (1-SaO2)*HbT - Fmet*HbT substituting above the second term becomes: (1-FMet)*HbT*SaO2/(HbT*SaO2 + (1-SaO2)*HbT - FMet*HbT) = SaO2 !!!!! Thus the formula could be simplified to Pulse ox = 0.85*FMet + SaO2 Where SaO2 + Fmet < = 100% In all of these calculations I am assuming that when the co-oximeter reorts a total saturation of SaO2 this represents the ratio of TOTAL saturated hemoglobin to total hemoglobin.