Previous studies have demonstrated a widely variable incidence of thromboembolism (VTE) in non-Hodgkin lymphoma (NHL) patients, ranging from 3% to 60%. This is in part because many previous studies have considered NHL as a single disease entity, combining a heterogeneous group of histological subtypes, each with a potentially different thrombogenic potential. We sought to clarify differences in VTE incidence between the indolent and intermediate grade forms of NHL, by looking at follicular lymphoma (FL) and diffuse large B-cell lymphoma (DLBCL), respectively. Furthermore, since chemotherapy is a major risk factor for VTE, we compared VTE incidence rate in the pre- and post-chemotherapy time periods.
We assembled a retrospective cohort of the United States veterans diagnosed with DLBCL or FL between 1998 and 2009. VTE was identified by ICD-9 codes, while DLBCL or FL was identified by ICD-O3 codes in the Veterans Health Administration central cancer registry. Patients who did not receive any chemotherapy or had a history of VTE prior to the diagnosis of lymphoma were excluded. VTE rate was compared by Chi-square testing.
A total of 2094 DLBCL patients and 470 FL patients were included. In the period between lymphoma diagnosis and treatment initiation, VTE occurred at a rate of 18. 0 per 100 person-years in DLBCL and 2. 3 per 100 person-years in FL (p<0. 0001). Similarly, in the 6 months following treatment initiation, DLBCL patients had a VTE rate of 12. 3 per 100 person-years, compared to 7. 3 per 100 person-years in FL (p=0. 039). In DLBCL, patients had a higher risk of VTE between diagnosis and treatment compared to the 6 months after treatment initiation (p=0. 038). In FL, treatment was associated with a trend toward higher risk of VTE, although not statistically significant (p=0. 056).
DLBCL is more “thrombogenic” than FL both before and during treatment. To our knowledge, this is the first report of direct comparison of VTE event rates between different histological subtypes of NHL, knowledge of which could influence decisions related to thromboprophylaxis. There are limitations of ICD-9 codes which should be acknowledged, and further research will be required to identify the optimal method to identify patients with VTE. The lower incidence of VTE after treatment initiation in DLBCL patients suggests that the reduced disease burden associated with treatment may actually decrease the overall risk of VTE despite the increased risk associated with chemotherapy. Conversely, in FL the risk of VTE appears to be higher in association with treatment.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.