Accuracy of a first-order model for estimating initial heparin dosage

The accuracy of a first-order pharmacokinetic model for determining initial heparin infusion rates was studied, and factors that could affect the accuracy of the method were investigated. Patients who received an i.v. infusion of heparin for at least 24 hours for treatment of deep-vein thrombosis, pulmonary embolism (PE), or myocardial infarction were identified by retrospective chart review. A therapeutic dosage of heparin was defined by an activated partial thromboplastin time of 45-75 seconds. Heparin dosages were calculated by using estimated blood volume as the heparin volume of distribution, a desired steady-state heparin concentration of 0.30 units/mL, and an elimination rate constant of 0.832 hr-1. The difference between the calculated dosage and the actual therapeutic dosage was calculated. The differences for various patient subgroups were compared, and the estimated dosages were regressed against the actual dosages to determine their predictive value. Data for 49 patients were analyzed. The mean +/- S.E. difference between the actual and calculated dosages was 29.2 +/- 37.1 units/hr. No significant differences were evident according to sex or indication for therapy. Smokers and nonsmokers differed, as did obese and lean patients. The equation appeared to be more accurate in nonsmokers than smokers. The addition of 200 units/hr to the calculated dosage for patients with PE resulted in minor improvement in the predictive capacity of the equation. Moderate agreement was observed between the actual and calculated heparin dosages in non-smokers of various body weights.