[Hemorrhagic complications from anticoagulant treatment: analysis of predictive risk factors].

Unlabelled: Anticoagulant therapy has shown its efficacy in the prevention of thromboembolic complications but it is not devoid of bleeding complications. Although the thromboembolic risk of some cardiac diseases may be extrapolated from well-organized clinical trials, the risk of bleeding complications should be determined in the context of the environment in which it is carried out.

Objective: To determine the complications of the patients in anticoagulant therapy, in our environment, and to analyse the of risk factors.

Patients and methods: We have studied the complications suffered by 300 patients who underwent anticoagulation for cardiac diseases, between March-94 and March-96 retrospectively. We have classified the complications in two groups: a) Fatal or intracranial with sequelae. b) Those requiring hospitalization and/or transfusion. Univariate and multivariant analyses were conducted to identify predictors of complications, including the following factors: age, sex, diabetes, hypertension, length of therapy, distance from our Center to their place of residence, INR (> 3 vs 2 to 3) and number of drugs associated with Acenocoumarol (> or = 3 vs 2 or less).

Results: During the follow-up 24 patients died due to non haemorrhagic complications; 3 left the treatment on their cardiologist recommendation; 2 moved their place of residence and 1 was lost in the follow-up. Of 270 remaining 3 (0.55/100 patients-year) had complications of group a and 21 (3.88/100 patients-year) of grub b. INR > 3 and multiple medications were shown as predictors of complications when including some of the complications considered.

Conclusions: 1) During two years of follow-up 1.1% (0.55/100 patients-year) of patients in anticaogulant therapy had bleeding complications resulting in death or neurological sequelae. 2) When including some of the complications considered the percentage rises to 4.44/100 patients-years. 3) Although there were no differences statistically significant, INR > 3 and polymedication have been found as predictors risk factor. 4) Patients with INR of 2 to 3 and non polymedicated presented a low risk of bleeding complications (1.66/100 patients-year).