Introduction:

venous thromboembolic disease is a frequent complication in hematologic malignancies and has a significant impact on morbidity and mortality. Despite the existence of multiple well-validated scores to stratify the risk of this disease in solid tumors, hematological malignancies are underrepresented in these models.

Objective:

to design a predictive index for the stratification of the thrombotic risk in patients with hematologic malignancies.

Methods:

an analytical observational study of cases and controls was carried out in the “Arnaldo Milián Castro” Hospital from Villa Clara Province, during the period from October 2016 to January 2019, in 285 hospitalized patients with hematological malignancies (94 with thromboembolic disease and 191 without). For univariate analysis, Chi-square test, Odds Ratio for risk estimation and Cramer's V for strength of association were applied. Logistic regression and ROC curve were applied in the multivariate analysis.

Results:

the predictive index of thromboembolic disease was composed of five predictors: hypercholesterolemia, tumor activity, immobility, use of thrombogenic drugs and diabetes mellitus. High risk of thrombosis was defined as a patient who scored four points or more and low risk as a patient who scored less than four points. The index correctly classified 81.10% of the patients, for a sensitivity of 59.57% and a specificity of 92.15%. The positive and negative predictive values were 78.87% and 82.24%, respectively.

Conclusions:

the developed index represented a specific and effective tool for the prediction of thromboembolic disease in hospitalized patients with hematologic malignancy.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )