Abstract
Sickle cell disease (SCD) is an inherited red blood cell disorder that leads to vaso-occlusion, endothelial damage, and activation of pro-coagulant pathways. Recent studies have demonstrated that thrombotic episodes occur at a 3- to 100-fold higher rate (PMID: 17000225, 22417249) in SCD versus non-SCD populations but the risk factors for thrombosis are not clear.
We investigated the incidence and predictors for thrombosis in a longitudinal cohort of 1193 SCD patients treated at our institution between 1/2008 and 12/2017. Clinical and laboratory data were extracted from the electronic medical records from the first outpatient encounter during this time period. Linear and categorical variables were compared by the Kruskal-Wallis and Chi-square test, respectively, and Cox proportional hazard models were adjusted for age, sex, SCD genotype, and hydroxyurea use.
With a median follow up of 5.6 years (interquartile range [IQR], 2.3-9.3 years), 210 SCD patients (17.6%) had 347 arterial or venous thromboses (75 stroke, 272 venous thromboembolism [VTE]) that occurred after 1/1/2008. Clinical and laboratory risk factors according to the occurrence of arterial or venous thrombosis are provided in Table 1. In Cox proportional hazards analysis, HbSS/Sβ0-thalassemia (HR 3.0, P<0.0001), hospitalization rate (10/year increment HR 1.3, P<0.0001), and higher serum creatinine (natural log HR 1.6, P=0.01) were independently associated with incident thrombotic events (Figure 1).
A subset of 174 SCD patients experienced the 272 VTEs; 153 (56%) deep vein thrombosis (DVT), 98 (36%) pulmonary embolism, 17 (6%) concurrent DVT and pulmonary embolism, and 4 (1%) right atrial thrombus. In those with a DVT, 116 were upper and 54 were lower extremity DVTs. 92 of the upper extremity DVTs and all 4 right atrial thrombi were catheter-related. In Cox proportional hazards analysis, independent risk factors for the development of VTE during the observation period were greater hospitalization rate (10/year increment HR 1.3, P<0.0001), higher systolic blood pressure (10-mmHg increase HR 1.2, P=0.002), HbSS/Sβ0-thalassemia (HR 2.1, P=0.004), and higher AST (natural log HR 1.6, P=0.01). The anticoagulation used during the initial VTE was warfarin in 115 (66%), oral direct factor Xa inhibitors in 29 (17%), low-molecular weight heparin in 16 (9%), and no therapy in 14 (8%) patients. Of those treated with an anticoagulant, 83 (52%) received initial therapy for 3-6 months, 47 (29%) for >6 months and 30 (19%) for < 3 months. A recurrent VTE occurred in 31 SCD patients (18% of those experiencing an initial VTE) and was not associated with either the type or duration of anticoagulation therapy.
We investigated whether genetic risk variants for thrombosis in African Americans (PMID: 26888256 and 21232005) were associated with thrombotic events in SCD. In a subset of 329 SCD patients with genome-wide genotyping performed using the Affymetrix Pan-African Axiom GeneChip array, we replicated the association with thrombosis of two variants in thrombomodulin (PMID: 26888256), THBD rs2567617 (MAF 0.25, OR 1.5, P=0.049) and rs1998081 (MAF 0.24, OR 1.5, P=0.059) after adjusting for age, sex, SCD genotype, and hydroxyurea. Thrombomodulin is an endothelial cell surface glycoprotein which promotes protein C activation and reduces circulating thrombin levels.
In conclusion, thrombotic events are common, occurring in 18% of SCD patients over a 10-year follow-up period. HbSS/Sβ0-thalassemia, frequent hospitalizations, kidney disease, and higher systolic blood pressures and AST concentrations were risk factors for thrombosis. Genome-wide marker array analysis points to a potential role of thrombomodulin in SCD-related thrombotic events based on replicated risk variants in THBD. Future studies integrating clinical, laboratory and genetic risk factors may improve our understanding for thrombosis and guide intervention practices in SCD.
No relevant conflicts of interest to declare.
