Effects of bacitracin on insulin concentrations of hemolyzed serum samples

Background: Accurate insulin measurements are integral to the calculation of insulin resistance and insulin secretion metrics. Hemolysis of blood samples during collection can lead to release of intracellular insulin degrading enzymes (IDEs) that can cleave insulin into fragments not recognized by current insulin assays, and consequently falsely lower measured insulin concentrations. We hypothesized that adding bacitracin, a known IDE inhibitor in vitro, to blood collection tubes in advance of blood collection could reduce hemolysis-induced insulin degradation. Methods: Samples from 15 healthy adult subjects with obesity enrolled in a clinical trial (NCT02153983) were analyzed. Samples were collected from 2 intravenous (IV) catheters (one placed in each arm) after an overnight fast during an insulin-modified frequently-sampled intravenous glucose tolerance test. 50% dextrose (0.3g/kg) was administered IV at time 0; at the 20- minute timepoint, insulin (0.05U/kg) was injected into the same IV catheter. At multiple time points, blood was drawn simultaneously from each arm via the indwelling catheters. From the arm contralateral to the dextrose/insulin injections, blood drawn into a single syringe was placed into two separate 4mL serum separator tubes (SSTs): a standard (“untreated”) SST and a second SST containing 50μL bacitracin at 20g/L concentration. From the ipsilateral arm, a “backup” blood sample was drawn into an untreated SST. Collected samples were immediately placed on ice and centrifuged within 30min. The factor “time” dichotomized the blood draw time points into pre- and post-insulin injection samples. For each subject, the first “untreated” hemolyzed sample in both “time” groups was selected and compared to the corresponding insulin values in the “bacitracin” and “backup” samples. Results were analyzed using GraphPad Prism 6 using 2-way repeated-measures ANOVA. Results: All 15 subjects had hemolyzed samples from the untreated and bacitracin samples; 12 subjects also had non-hemolyzed samples from the backup IV available from the same timepoints. Insulin values from the bacitracin-treated SSTs were significantly greater than the values from the corresponding untreated hemolyzed samples (mean±SD, 144.5±176.1 vs 116.5±135.9 uIU/mL; p<0.05). Conversely, insulin values from the bacitracin-treated SST were not significantly different from the “backup” SST insulin values (153.4±196.5 vs 161.6±221.2 mIU/mL; p=0.34). Conclusions: This study found pretreatment of SSTs with bacitracin successfully prevented hemolysis-induced insulin degradation. Obtaining insulin measurements in bacitracin-containing SSTs may prove beneficial for metabolic research. Further studies, including a larger sample size and the establishment of a dose-response curve for bacitracin's effects on hemolysis, are warranted.