The correlation between the rate of drug removal by hemodialysis (dialyzability) which was described in previous papers (n=140) and the protein binding rate (n=135), urinary excretion rate of unmetabolized drugs (n=98), n-octanol/water partition coefficient (O/W coefficient; n=81) and molecular weight (n=140) of the drugs was investigated. The highest correlation parameter was the rate of protein binding (r=-0.600, p<0.0001), and the rate of urinary excretion of unmetabolized drugs (r=0.515, p<0.0001), followed by log O/W coefficient; (r=-0.478, p<0.0001), reciprocal of the volume of distribution (1/Vd; r=0.450, p<0.0001), volume of distribution (r=-0.341, p<0.0001). There was no correlation between the dialyzability and O/W coefficient, molecular weight, and log molecular weight (r=-0.157; p=0.163, r=-0.032; p=0.708, r=-0.026; p=0.770, respectively). Furthermore, multiple regression analysis was carried out using the protein binding rate, 1/Vd, molecular weight, and urinary excretion rate of unchanged drugs as independent variables, and drug dialyzability as a dependent variable. As a result, the protein binding rate (β=-0.527, p<0.0001) and 1/Vd (β=0.314, p=0.0001) were significant independent predictors of drug dialyzability. Drugs with a high protein binding rate and a high volume of distribution indicative of a high rate of transit to tissues appeared to be the most important predictors because these two parameters are related to low dialyzability. Drugs that are metabolized in the liver with a low urinary excretion rate of unmetabolized drugs, highly lipid soluble drugs with a high O/W coefficient were also predictors of low dialyzability. However, it became evident that a molecular weight of around 1, 000 daltons does not predict drug dialyzability.
View full abstract