White blood cell count measured prior to cancer development is associated with future risk of venous thromboembolism--the Tromsø study

Abstract

Background: Elevated white blood cell (WBC) count is associated with risk of venous thromboembolism (VTE) in cancer patients initiating chemotherapy. It is not known whether the risk of VTE by WBC count in cancer patients is causal or merely a consequence of the malignant disease. To address this question, we studied the association between WBC count, measured prior to cancer development, and risk of VTE in subjects who did and did not develop cancer during follow-up in a prospective population-based study.

Methods: Baseline characteristics, including WBC and neutrophil counts, were measured in 24304 initially cancer-free subjects who participated in the Tromsø Study in 1994-1995. Incident cancer diagnosis and VTE events were registered up to September 1, 2007. In the cancer cohort, WBC and neutrophil counts were measured in average 7.1 years before cancer development. Cox-regression models were used to calculate hazard ratios (HRs) for VTE by WBC and neutrophil counts as categorized variables (<40(th), 40-80(th), and >80(th) percentile) with 95% confidence intervals (CIs).

Results: During follow-up, 1720 subjects developed cancer and there were 388 VTE events, of which 116 occurred in the cancer-group (6.9 per 1000 person-years) and 272 in the cancer-free group (1.1 per 1000 person-years). In those who developed cancer, WBC count above the 80(th) percentile (≥ 8.6 x 10(9) cells/L) was associated with a 2.4-fold higher risk (HR 2.36, 95% CI: 1.44-3.87) of VTE compared to WBC count below the 40(th) percentile (<6.4 x 10(9) cells/L). No association was found between WBC count and VTE in those who stayed cancer-free (HR 0.94, 95% CI 0.65-1.36). Similar findings were observed for neutrophils.

Comment: Pre-cancer WBC count was associated with risk of VTE in cancer patients, but not in cancer-free subjects. Our findings suggest that leukocytes may play a causal role in cancer-related VTE rather than only reflecting the low-grade inflammation associated with cancer.

Figure 1. WBC count and risk of venous thromboembolism.

Dose–response relationship between WBC count and risk of VTE in cancer and non-cancer subjects obtained by generalized linear regression. The regression models are adjusted for age, sex, BMI, smoking, self-reported diabetes, physical activity and self-reported CVD. The solid lines show HRs and the shaded areas show 95% CIs. Density plots show the distribution of WBC, and white vertical lines indicate 2.5^th, 25^th, 50^th, 75^th and 97.5^th percentiles.