Pulse oximetry is a noninvasive method of measuring oxyhemoglobin saturation. The validity of pulse oximetry in sickle cell disease (SCD) has been questioned. We evaluated pulse oximetry, arterial blood gas analysis, and co-oximetry in patients with SCD, and we assessed the effect of dyshemoglobin and altered blood-oxygen affinity on their accuracy. Sixteen patients with SCD aged 7-21 years had arterial and venous blood drawn and transcutaneous pulse oximetry performed. Oxyhemoglobin dissociation curves were plotted from the venous blood of 15 patients. Oxyhemoglobin saturation estimated by arterial blood gas analysis (SaO(2)) and measured by pulse oximetry (SpO(2)) were both higher than the saturation by co-oximetry (FO(2)Hb) (mean +/- SD = 96.3 +/- 1.6%, 94 +/- 3.1%, and 89.1 +/- 3.8%, respectively). There was a significant, positive correlation between SpO(2) and FO(2)Hb (r = 0.7, P = 0.002). The patients had elevated levels of methemoglobin (MetHb) and carboxyhemoglobin (COHb) (2.3 +/- 1.4% and 4.7 +/- 1.3%, respectively). The oxyhemoglobin dissociation curves were frequently shifted to the right with oxygen tensions elevated when hemoglobin was 50% saturated with oxygen (P(50)) (32.5 +/- 4.5 mm Hg). There was a strong correlation between the amounts of dyshemoglobin (MetHb + COHb) and the difference between SaO(2) and FO(2)Hb (r = 0.7, P = 0.002). There was no correlation between the difference between SaO(2) and FO(2)Hb and the P(50) (r = 0.27, P = 0.33) There was also a strong positive correlation between SaO(2)-SpO(2) and dyshemoglobin fraction (r = 0.77, P = 0.001). We conclude that pulse oximetry and arterial blood gas analysis overestimate oxygen saturation when compared to co-oximetry, but that SpO(2) is consistently closer than SaO(2) to FO(2)Hb. SpO(2) is partially affected by MetHb and COHb. The discrepancy between SaO(2) and FO(2)Hb is due to the presence of dyshemoglobin and a shifted oxyhemoglobin dissociation curve, but the effect from dyshemoglobin predominates.
Copyright 1999 Wiley-Liss, Inc.