New method to predict patients' intravenous heparin dose requirements.

Abstract

OBJECTIVE: To predict intravenous heparin dose requirements of patients treated for thromboembolic disorders. DESIGN: A retrospective cohort study in which we used simple linear regression to predict patients' effective maintenance dose (EMD) of heparin (units/kg/hour needed to achieve and maintain APTT therapeutic range) from patients' "heparin responsiveness" (the APTT increase after the initial 6 hours of heparin treatment per units/kg/hour received). SETTING/PATIENTS: The model was derived from 46 patients treated at one hospital (Hospital A) and then tested in 42 patients treated at another hospital (Hospital B). MEASUREMENTS AND MAIN RESULTS: Among Hospital A patients, there was a strong linear correlation (r = -.880; p < .001) between EMD (mean 16.02 units/kg/hour; 95% CI 14.9, 17.15) and "heparin responsiveness" (HR): EMD = 25.651 - [95.118 x HR]. This model accurately predicted Hospital B patients' EMD: 97% (37/38) fell within the model's 95% prediction interval; the mean absolute difference between predicted and actual EMD was 1.73 units/kg/hour (95% CI 1.39, 2.08); and only 16% of patients had EMD's more than 3 units/kg/hour different from that predicted by the regression model. The model's accuracy was comparable to that of our gold standard, the weight-based heparin dosing nomogram. CONCLUSION: The infusion dose of intravenous heparin effective for an individual patient can be predicted accurately from the patient's body weight and APTT response to the initial 6 hours of treatment. Especially in hospitals where validated heparin dosing nomograms are not used, clinicians may find this simple technique useful in achieving timely therapeutic anticoagulation.