412 Cortisol cut off point to diagnose adrenal insufficiency (AI) using a monoclonal antibody immunoassay

Abstract

OBJECTIVES/GOALS: AI is diagnosed when peak cortisol level after a cosyntropin stimulation test is <18 mg/dL using polyclonal antibody (pAb) immunoassay. However, the polyclonal assay is being replaced by a specific monoclonal antibody (mAb) immunoassay which yields lower cortisol levels, leading to the over diagnosis of AI and use of unnecessary steroid use. METHODS/STUDY POPULATION: We obtained 36 samples from patients undergoing 1 mcg cosyntropin stimulation tests for diagnosis of AI. Samples were analyzed using pAb immunoassay (Abbott Architect Cortisol), mAb immunoassay (Roche Elecsys Cortisol II), and mass spectrometry (MS). AI was diagnosed if serum cortisol level was <18 using the pAb immunoassay. Measurements by MS and mAb immunoassay were individually used in simple logistic regression models to predict AI. For each model, we calculated a cortisol level corresponding to a 50% probability (median) of AI and used the delta method to determine the standard error and 95% confidence interval of the median. We used receiver operator characteristic (ROC) curve, area under the curve, sensitivity, and specificity to evaluate the potential of the median values as thresholds for each predictor. RESULTS/ANTICIPATED RESULTS: Data showed a mean cortisol level of 17 mcg/dL using the pAb immunoassay, 12 mcg/dL using the mAb immunoassay, and 12.96 mcg/dL using MS. The mean difference in cortisol level between the mAb immunoassay and the pAb immunoassay was 5.12 mcg/dL (p-value <0.01). The ROC curve model indicated an area under the curve of 0.997 with a median value of 11.2 mcg/dL for the mAb immunoassay. This provides a sensitivity of 95%, specificity of 95%, positive predictive value of 95%, and negative predictive value of 94%. This new threshold has a Kappa coefficient of 0.89 when compared to the pAb immunoassay. DISCUSSION/SIGNIFICANCE: New and highly specific mAb immunoassays are being used more widely but yield lower cortisol results. This reflects the need for further studies to determine new cut off points for highly specific cortisol immunoassays. A cut off level of 11.2 mcg/dL would provide a sensitivity of 95% and specificity of 95%.