The present study sought to derive a simple formula for predicting the drug removal rates during hemodialysis. We examined the relationship between drug removal rates during hemodialysis and the molecular weights or pharmacokinetic parameters of injectable drugs (N = 90) obtained from pharmaceutical interview forms in Japan. Stepwise multiple regression analysis with the removal rate by hemodialysis as the objective variable adjusted for molecular weight or pharmacokinetic parameters as explanatory variables, showed that the logarithm of molecular weight (B = -18.87), the protein binding rate (B = -0.40), and the fraction of the unchanged drug excreted into the urine/volume of distribution (B = 0.05) were significantly and independently associated with drug removal rate by hemodialysis (α = 90.78, adjusted R = 0.64, P = 2.2e ). Our data demonstrated that molecular weight, protein binding rate, and volume of distribution were important factors affecting drug removal during hemodialysis, and that our simple regression equation could be used to predict the drug removal rate during hemodialysis.

Background: The interaction between arbekacin and the hemodialysis membrane is unclear. The aim of this study was to evaluate the adsorption of arbekacin (ABK) onto hemodialysis membranes and to establish a simpler system for evaluating the adsorption properties of the drug. Methods: First, small cut pieces of polysulfone or sulfonated polyacrylonitrile (AN69) hollow fiber membranes were stirred in a solution of ABK for the qualitative assessment. Then, we designed the experimental system that was approximately 1/100 of the actual size using a small dialyzer "mini-module" of our original design for the polysulfone or AN69. We circulated ABK solution in this system for the quantitative assessment. Finally, we administered ABK in subjects undergoing hemodialysis as a clinical trial to evaluate the adsorption of ABK onto polysulfone or AN69. Results: In the qualitative assessment, the rates of ABK adsorption onto polysulfone and AN69 were 6.2 ± 2.9 and 49.8 ± 1.8 %, respectively. In the quantitative assessment, although there was almost no change in the circulating ABK concentration with the use of polysulfone, there was about 68 % decrease in the circulating ABK concentration with the use of AN69. In the clinical trial, clearance by the dialyzer using AN69 was high at 10 min after starting dialysis, despite the sieving coefficient being 0. However, reduction of ABK clearance was observed with time; the removal rate of ABK at the completion of the dialysis using polysulfone or AN69 was about 67 %, with no difference between the two membranes. The main limitation of this study was the small sample size in the clinical trial. Conclusions: These findings suggest that ABK is adsorbed onto AN69. Furthermore, the present adsorption experiment with a mini-module was considered useful as an evaluation system because it was easy to handle, using less solvent and drugs compared with a previous report, and it reflected the results of the clinical trial.

Background: The rate of drug removal by hemodialysis needs to be considered when designing drug dosage regimens for patients on hemodialysis. We previously developed a simplified equation to predict the removal rates of intravenously administered drugs by hemodialysis. Here, we addressed shortcomings of this equation and developed a more accurate equation that can also predict the removal rates of orally administered drugs. Methods: A total of 70 drugs with known pharmacokinetic and physical parameters and drug removal rates that were measured during hemodialysis in clinical cases were randomly assigned at a 4:1 ratio to a training data group or a test data group. A prediction equation was developed by performing stepwise multiple regression analyses using the training data (i.e., the removal rate by hemodialysis) as the objective variable and pharmacokinetic parameters as the explanatory variables. The equation was validated using the test data. Results: Multiple regression analyses revealed that molecular weight (MW), protein binding rate, and fraction excreted unchanged in urine relative to the volume of distribution (Vd) were independently correlated with the drug clearance rate (adjusted coefficient of determination, 0.83; p = 2.2e−16). The following equation was obtained: drug removal rate by hemodialysis (%) = −17.32 × [log (MW)] – 0.39 × [protein binding rate (%)] + 0.06 × [fraction excreted unchanged in urine (%)/Vd (L/kg)] + 83.34. Validation of the equation using the test data showed a very high correlation between predicted and measured reduction rate (R = 0.93, p = 1.87e−6). Mean error was −3.34 (95% confidence interval: −10.03, 3.35), mean absolute error was 9.59, and root mean square error was 16.48. Conclusion: The modified equation derived in this study using pharmacokinetic and physical parameters as variables precisely predicted the removal rates of both intravenous and oral drugs by hemodialysis.

No abstract

No abstract