Heparin requirements for full anticoagulation are higher for patients on dabigatran than for those on warfarin - a model-based study

Abstract

Purpose: Dabigatran (D) is increasingly used for chronic anticoagulation in place of warfarin (W). These patients may present for catheter-based procedures requiring full anticoagulation with heparin. This study compares the heparin sensitivity of patients previously on dabigatran, on warfarin, or on no chronic anticoagulant during ablation of atrial fibrillation.

Patients and methods: In a retrospective study of patients treated with D, W, or neither drug (N) undergoing atrial ablation, the timing of heparin doses and resulting activated clotting times were collected. First, the initial activated clotting time response to the first heparin bolus was compared. Then, a non-linear mixed effects modelling (NONMEM) analysis was performed, fitting a pharmacokinetic and -dynamic model to the entire anticoagulation course of each patient. Resulting model coefficients were used to compare the different patient groups.

Results: Data for 66 patients on dabigatran, 95 patients on warfarin, and 27 patients on no anticoagulation were retrieved. The last dose of dabigatran or warfarin had occurred 27 hours and 15 hours before the procedure. Groups D and N both responded significantly less (P<0.05) to the initial heparin bolus than Group W (approximately 50%). Likewise, the model coefficients resulting from the fit to each group reflected a significantly lower heparin sensitivity in groups D and N compared to W. Clearances of the heparin effect in the model did not differ significantly among groups.

Conclusion: Patients on warfarin with an average INR of 1.5 or higher are more sensitive to heparin than patients not previously anticoagulated or patients who discontinued dabigatran 27 hours earlier (approximately two half-lives) warfarin.

Keywords: NONMEM; PKPD model; atrial fibrillation; electrophysiology.

Figure 1

Schematic of the one-compartment PKPD model in panel (A). Notes: The heparin blood concentration depends on the rate of heparin infusion, the volume of distribution, V_dist, and the clearance, C_L. The effect on the ACT is determined by the pharmacodynamic constant, k_ACT, which changes depending on whether the patient was on dabigatran, warfarin, or no anticoagulant before the procedure. Based on the known heparin boluses, the model was fitted to the known ACT response of each individual patient as shown for one sample patient panel (B). Abbreviations: ACT, activated clotting time; PKPD model, pharmacokinetic and -dynamic model.

Figure 2

Response of the activated clotting time (∆ACT) to the initial heparin bolus analyzed for patients on warfarin (W_low and W_high), on dabigatran (D), or no anticoagulant (N) shown as medians with interquartile ranges. Notes: The number of females and males for each group are indicated. The groups W_low and W_high were significantly more sensitive to heparin than groups D and N. The only significant sex-based difference was: males in group D were more sensitive than females. All significant differences (P<0.05) are indicated by asterisks (*). The differences between the groups W_low and W_high, and D and N and were not significant. Abbreviation: ACT, activated clotting time.

Figure 3

The PKPD model coefficient indicating sensitivity to heparin, k_ACT, for groups W_low, W_high, D, and N shown as medians with interquartile ranges. Notes: Analysis was repeated for males and females separately with the numbers indicated. Significant differences in heparin sensitivity (P<0.05) were: W_high > W_low > (D and N) as indicated by the brackets with the asterisks. Groups D and N were not significantly different. The only significant sex-based difference was in group D: males were more sensitive to heparin than females. Abbreviations: ACT, activated clotting time; PKPD model, pharmacokinetic and -dynamic model.