Letter: "Albumin-titratable bilirubin" as an index to bilirubin binding.

ogists on the ABA-100, we found that about 13% of the samples (22 of 169) gave results which differed by more than 10 mg/liter. This prompted us to test the assumption concerning the effectiveness of the blanking wavelength. The test wavelength is 650 nm and the blank wavelength is 550 nm for the phosphorus procedure on the ABA-100. We Selected two patients’ samples that had given discrepant results and one that had not, and diluted them 26-fold with physiological saline. These solutions were scanned (vs. saline) from 700 to 400 nm with a ratio-recording spectrophotometer (DK-2A; Beckman Instruments, Fullerton, Calif. 92634). Table 1 shows the absorbance differences for these samples. The patients’ diagnoses were, respectively, cervical carcinoma, gunshot wound, and leukemia. All of these samples absorbed more at the blank wavelength than at the test wavelength and the absorbance difference varied with the patient. Results for phosphorus assays performed on these samples are also shown in the Table. These three examples serve to warn that extreme caution may have to be exercised whenever a bichnomatic analyzer is used for endpoint measurements. It may not be readily apparent how much comparison work is needed before a new procedure can be used with confidence. Rather, the amount of comparison work appears to be a function of the source(s) of samples that are being assayed by the particular laboratory; that is, of whether the patient population is relatively normal or is a combination of patients who are relatively normal and those who are receiving tertiary care. Another point also worth considening is that most of the method interferences documented to date are for procedures in which the blank and test mixtures are read at the same wavelength, information that may not be directly transferable to a bichromatic procedure. Central Laboratory University of Colorado Medical Center Denver, Cob. 80220