OBJECTIVE: Rapid noninvasive measurement of total hemoglobin would be extremely useful for various clinical situations. This study determined the clinical accuracy and utility for a pulse total-hemoglobinometer using four wavelengths: 660 nm (reduced hemoglobin), 805 nm (isosbestic point), 940 nm (oxygenated hemoglobin), and 1300 nm (water density). DESIGN: Clinical trial. SETTING: University school of medicine. PATIENTS: Patients were 122 individuals (age, 18-82 yrs; 49.4 +/- 16.0 yrs [mean +/- SD]), including 71 healthy volunteers, 24 patients undergoing surgery, and 27 patients undergoing hemodialysis. INTERVENTIONS: The hemoglobinometer probe, which simultaneously indicated peripheral oxygen saturation, pulse rate, and hemoglobin, was placed on the fingertip similarly to a regular pulse oximeter. The hemoglobin values were compared with those obtained by the co-oximeter or the sodium lauryl sulfate-methemoglobin method. Those hemoglobin values were assigned to either the training set or the validation set for statistical evaluation. MEASUREMENTS AND MAIN RESULTS: Multiple regression analysis including the ratio of the pulsatile optical density (phi(ij)) derived from the four wavelengths and other factors demonstrated that the mean value of the normalized pulse wave obtained from the photodiode at 805 nm (DC805) and the ratios of DC940 and DC1300 (DC940/DC1300) were the pivotal factors in the hemoglobinometer's increased accuracy in the clinically useful range. The coefficient of determination between both methods was r2 = .81 (p < .0001) in the training set and r2 = .75 (p < .0001) in the validation set. When the cutoff value of anemia was set at 10 g/dL, and anemia was defined as <10 g/dL, the respective sensitivity and specificity of hemoglobinometer values to detect anemia in intraoperative patients were 84.3% and 84.6% (n = 20). CONCLUSIONS: The data demonstrated the necessity for consideration of light scattering in red blood cells for pulse-spectrophotometric hemoglobin measurement. This was accomplished with additional factors, such as DC805 and DC940/DC1300. With these improvements, the pulse hemoglobinometer provided noninvasive, clinically acceptable measurement of hemoglobin. The pulse hemoglobinometer is a versatile tool that might be useful for routine health checkups of neonates and young children, intraoperative monitoring of bleeding, and emergency care.
OBJECTIVE: Rapid noninvasive measurement of total hemoglobin would be extremely useful for various clinical situations. This study determined the clinical accuracy and utility for a pulse total-hemoglobinometer using four wavelengths: 660 nm (reduced hemoglobin), 805 nm (isosbestic point), 940 nm (oxygenated hemoglobin), and 1300 nm (water density). DESIGN: Clinical trial. SETTING: University school of medicine. PATIENTS: Patients were 122 individuals (age, 18-82 yrs; 49.4 +/- 16.0 yrs [mean +/- SD]), including 71 healthy volunteers, 24 patients undergoing surgery, and 27 patients undergoing hemodialysis. INTERVENTIONS: The hemoglobinometer probe, which simultaneously indicated peripheral oxygen saturation, pulse rate, and hemoglobin, was placed on the fingertip similarly to a regular pulse oximeter. The hemoglobin values were compared with those obtained by the co-oximeter or the sodium lauryl sulfate-methemoglobin method. Those hemoglobin values were assigned to either the training set or the validation set for statistical evaluation. MEASUREMENTS AND MAIN RESULTS: Multiple regression analysis including the ratio of the pulsatile optical density (phi(ij)) derived from the four wavelengths and other factors demonstrated that the mean value of the normalized pulse wave obtained from the photodiode at 805 nm (DC805) and the ratios of DC940 and DC1300 (DC940/DC1300) were the pivotal factors in the hemoglobinometer's increased accuracy in the clinically useful range. The coefficient of determination between both methods was r2 = .81 (p < .0001) in the training set and r2 = .75 (p < .0001) in the validation set. When the cutoff value of anemia was set at 10 g/dL, and anemia was defined as <10 g/dL, the respective sensitivity and specificity of hemoglobinometer values to detect anemia in intraoperative patients were 84.3% and 84.6% (n = 20). CONCLUSIONS: The data demonstrated the necessity for consideration of light scattering in red blood cells for pulse-spectrophotometric hemoglobin measurement. This was accomplished with additional factors, such as DC805 and DC940/DC1300. With these improvements, the pulse hemoglobinometer provided noninvasive, clinically acceptable measurement of hemoglobin. The pulse hemoglobinometer is a versatile tool that might be useful for routine health checkups of neonates and young children, intraoperative monitoring of bleeding, and emergency care.