Résumé
Background: Various factors contribute to the discrepancies observed between the bromocresol green (BCG) and bromocresol purple (BCP) assays of serum concentration. Either a BCG or a modified BCP assay is a routine laboratory for albumin measurement in Japan. High-dose of penicillin G underestimates serum albumin level using a modified BCP method in vitro. Therefore, we examined the serum albumin level in the patients treated with high-dose of penicillin G and also performed the experiments on coincubation with plasma, or albumin product and penicillin G solution in vitro. Methods: The medical records of 71 patients treated with high-dose of Penicillin G collected between 2009 and 2012 were reviewed for age, gender, biochemistry (total protein, albumin and potassium), underlying diseases and usage of albumin product. Patients were divided into 2 groups: BCG group (N = 38) and a modified BCP group (N = 33). We compared serum albumin levels between two groups. We performed the experiments on co-incubation with albumin product or human plasma and penicillin G solution in vitro. Results: Serum albumin levels using a modified BCP assay decreased during the treatment with high dose penicillin G (-0.4 ± 0.1 g/dL), while serum albumin levels by a BCG method did not decrease (0.06 ± 0.05 g/dL). Although only one patient revealed hypoalbuminemia (<2.0 g/dL) by a BCG method (2.6%), ten patients revealed hypoalbuminemia by a modified BCP method (33%). Penicillin G underestimated plasma albumin levels using a modified BCP methods (% underestimation: 42.9 ± 0.0%) more than those using a BCG method (% underestimation: 10.6 ± 0.0%) in co-incubation experiments in vitro. Conclusion: High-dose of penicillin G might cause the underestimation of serum albumin levels using bromocresol dye-binding methods.