Vol. 30, Issue 6, 676-683, June 2002

Pharmacokinetics of Sodium Nitrite-Induced Methemoglobinemia in the Rat

Michael C. Kohn, Ronald L. Melnick, Frank Ye, and Christopher J. Portier

Laboratory of Computational Biology and Risk Analysis, National Institute of Environmental Sciences, Research Triangle Park, North Carolina

A biologically based mathematical model was created to characterize time and dose-dependent relationships between exposure to nitrite and induction of methemoglobinemia. The model includes mass action equations for processes known to occur: oral absorption of nitrite, elimination from the plasma, partitioning between plasma and erythrocytes, binding of nitrite to hemoglobin and methemoglobin, and the free radical chain reaction for hemoglobin oxidation. The model also includes Michaelis-Menten kinetics for methemoglobin reductase-catalyzed regeneration of hemoglobin. Body weight-scaled rate constants for absorption (k_a) and elimination (k_e), the effective erythrocyte/plasma partition coefficient (P), and the apparent K_m for methemoglobin reductase were the only parameters estimated by formal optimization to reproduce the observed time course data. Time courses of plasma nitrite concentrations and blood levels of hemoglobin and methemoglobin in male and female rats that had received single intravenous or oral doses of sodium nitrite were measured. Peak plasma levels of nitrite were achieved in both sexes approximately 30 min after oral exposure, and peak methemoglobin levels were achieved after 100 min. The model predicts that 10% of the hemoglobin is oxidized to the ferric form after oral doses of 15.9 mg/kg in male rats and 11.0 mg/kg in female rats and after intravenous doses of 8.9 and 7.1 mg/kg in male and female rats, respectively. The t_1/2 for recovery from methemoglobinemia was 60 to 120 min depending on dose and route of administration. A sensitivity analysis of the model was performed to identify to which parameters the predictions of the model were most sensitive and guide attempts to simplify the model. Replacement of the V_max of methemoglobin reductase with a value representative of humans predicted a 10% methemoglobinemia following an intravenous dose of 5.8 mg/kg, in close agreement with an observed value of 5.7 mg/kg for humans.