High-dose busulfan is an important component of many bone marrow transplantation-preparative regimens. High busulfan plasma levels have been shown to increase the chance of venoocclusive disease and low levels are associated with recurrence of disease or graft rejection. Currently, busulfan levels are monitored by physical methods that are expensive and time consuming, resulting in relatively low overall use of busulfan testing for dose adjustment. Novel highly selective antibodies for busulfan have been generated and a microtiter plate immunoassay capable of quantifying busulfan levels in plasma has been developed. The assay was configured using a busulfan-horseradish peroxidase (HRP) conjugate as the reporter group and busulfan monoclonal antibodies. The assay requires 30 microL of plasma with no sample preparation. The immunoassay has a standard curve based on busulfan with a range of 75-2000 ng/mL. The time to first result is 30 minutes with up to 40 patient samples in duplicate; multiple plates can be run at once. The coefficient of variation (CV) on signal is <5% for an entire plate, and the 95% confidence interval for negative samples (n = 78) is below the lowest calibrator of 75 ng/mL. Cross-reactivity with the major inactive metabolites (tetrahydrothiophene, tetramethyl sulfone, and tetrahydrothiophene-3-ol-1,1-dioxide) was <0.1%. Results generated with clinical samples (n = 35 and n = 70) correlate well to gas chromatography-mass spectrometry (R = 0.976 and 0.985, respectively) with a slope of 1.05 +/- 0.05. This immunoassay method is suitable for determining levels of busulfan in human plasma. It offers the advantages of using a smaller sample size, does not require sample preparation, and is less labor intensive than other methods. The ability to make 240 determinations per hour enables effective and timely routine monitoring of busulfan levels in clinical practice.