Association between incident cancer and subsequent stroke

Abstract

Objective: A study was undertaken to examine the association between incident cancer and the subsequent risk of stroke.

Methods: Using the Surveillance, Epidemiology, and End Results-Medicare linked database, we identified patients with a new primary diagnosis of breast, colorectal, lung, pancreatic, or prostate cancer from 2001 through 2007. These patients were individually matched by age, sex, race, registry, and medical comorbidities to a group of Medicare enrollees without cancer, and each pair was followed through 2009. Validated diagnosis codes were used to identify a primary outcome of stroke. Cumulative incidence rates were calculated using competing risk survival statistics.

Results: Among 327,389 pairs of cancer patients and matched controls, the 3-month cumulative incidence of stroke was generally higher in patients with cancer. Cumulative incidence rates were 5.1% (95% confidence interval [CI] = 4.9-5.2%) in patients with lung cancer compared to 1.2% (95% CI = 1.2-1.3%) in controls (p < 0.001), 3.4% (95% CI = 3.1-3.6%) in patients with pancreatic cancer compared to 1.3% (95% CI = 1.1-1.5%) in controls (p < 0.001), 3.3% (95% CI = 3.2-3.4%) in patients with colorectal cancer compared to 1.3% (95% CI = 1.2-1.4%) in controls (p < 0.001), 1.5% (95% CI = 1.4-1.6%) in patients with breast cancer compared to 1.1% (95% CI = 1.0-1.2%) in controls (p < 0.001), and 1.2% (95% CI = 1.1-1.3%) in patients with prostate cancer compared to 1.1% (95% CI = 1.0-1.2%) in controls (p = 0.085). Excess risks attenuated over time and were generally no longer present beyond 1 year.

Interpretation: Incident cancer is associated with an increased short-term risk of stroke. This risk appears highest with lung, pancreatic, and colorectal cancers.

Figure 1

Cumulative incidence of stroke in cancer patients compared to matched controls. Cumulative incidence is stratified by cancer type, (A) Breast, (B) Colorectal, (C) Lung, (D) Pancreatic, (E) Prostate. Competing risk survival statistics were used to calculate incidence. Data are shown through the median period of follow-up for cancer patients for each cancer type up to a maximum of 3 years. 95% confidence intervals are noted by dashed lines.