[Major closed-space bleeding in patients on anticoagulation with acenocoumarol (TAO) or non-fractionned heparin(HS):...

To study the characteristics of major bleeding episodes into a closed space (BCS) of patients under chronic anticoagulation with either unfractionated heparin (HS) or coumadin (CM), and to determine the relationship, if any, of anticoagulation parameters (INR, PT and PTT) values at the time of bleeding with the episode. Finally, to determine risk factors for BCS and mortality in this population.

Descriptive epidemiology of all cases of BCS seen in our hospital from 1995 to 2000 was obtained through the records and follow up visits of all patients under anticoagulation (HS or CM) during this period. A matched case-control study to determine risk factors for BCS was carried out. Cases and controls (1:2) were matched for age, gender, anticoagulant treatment and indication for anticoagulation. Cases were patients with a BCS while on anticoagulation (HS OR CM). Controls were patients under anticoagulation (HS or CM) without any bleeding episode during the study period that had anticoagulation parameter values (INR, PT or PTT) determined the very same day than the cases.

During the study period, 225 patients under anticoagulation were prospectively followed (75 cases and 150 controls) amid a total of 1650 patients under anticoagulation, for a 4.5% prevalence of BCS. Reasons for anticoagulation were: atrial fibrillation in 79 (35.3%), valvular heart disease in 59 (25.9%), pulmonary embolism or deep venous thrombosis in 48 (21.4%), dilated cardiomyopathy in 26 (11.6%) and vascular cerebral stroke in 13 (5.8%). Mean age of cases was 70.5 (SD 9.5) years and 41 (55%) were women, values similar to the controls. At the time of BCS 39 patients were on CM and 36 on HS. The mean INR value in the CM group at the time of the episode of BCS was 5.3 (SD + 7.5) while the PTT value was 2,25 (SD 0.95) in the HS group. There was previous antecedent bleeding in 24 (32%) cases. The most common sites of BCS were: muscular (40%), CNS (30.6%), retroperitoneal (18,6%) and articular (10.6%). Muscular (abdominal or thoracic wall) and retroperitoneal BCS were higher in the HS group (10 and 12 in the HS group versus 5 and 2 in the CM group, respectively; p < 0.0001). In contrast, CNS bleeding was commoner in the CM group (20 in CM versus 3 in HS; p < 0.001). BCS related mortality rate was 14.6% (11/75) and higher in the CM group (p = 0.04). Comparative analysis of the case-control study revealed that anticoagulation values in the CM group at the time of bleeding were within the recommended range in 38.5% of cases vs. 75% of the controls (p < 0.001). Also, there were significant differences in mean INR values between cases and controls (5.3 + 7.5 vs. 2.6 + 0.9, p < 0.029) In the HS group no differences were present in PTT values at the time of bleeding between cases and controls. In BCS cases, a previous bleeding episode was more frequent than in the control group (32% versus 1.3%, p < 0.001). Likewise, mortality was higher in cases (18,6%) than in controls (11.4%), p = 0.01.

In our study, the majority of patients under anticoagulation with CM had INR values above the recommended range at the time of BCS, in contrast with those on HS that had a PTT within the therapeutic range at the time of the BCS. A previous bleeding episode was an independent risk factor for a BCS episode. Bleeding was a late complication in the CM group and frequently in the CNS, while BCS was more frequently associated with muscular or retroperitoneal sites in the HS treated group. BCS related mortality was 15%. Close monitoring of INR is crucial to minimize bleeding complications.