RESUMEN
Limited data are available regarding the impact of cancer on cerebrovascular accidents in patients with atrial fibrillation (AF). We queried the Nationwide Inpatient Survey Database to identify patients who have diagnostic code for AF. We performed a 1:1 propensity matching based on the CHA2DS2VASc score and other risk factors between patients with AF who had lung, breast, colon, and esophageal cancer, and those who did not (control). The final cohort included a total of 31,604 patients. The primary outcome of in-hospital cerebrovascular accidents (CVA) was lower in the cancer group than in the control group (4% vs 7%, p < 0.001), but with only a weak association (Ñ = -0.067). In-hospital mortality was higher in the cancer group than in the control group (18% vs 11%, p < 0.001; Ñ = -0.099). A subgroup analysis according to cancer type showed similar results with a weak association with lower CVA in breast cancer (4% vs 7%; Ñ = -0.066, p < 0.001), lung cancer (4% vs 6%; Ñ = -0.062, p < 0.001), colon cancer (4% vs 6%; Ñ = -0.062, p < 0.001), and esophageal cancer (3% vs 7%; Ñ = -0.095, p < 0.001) compared with the control groups. A weak association with higher in-hospital mortality was demonstrated in lung cancer (20% vs 11%; Ñ = -0.127, p < 0.001), colon cancer (16% vs 11%; Ñ = -0.076, p < 0.001), and esophageal cancer (20% vs 12%; Ñ = -0.111, p < 0.001) compared with the control groups, but no significant difference between breast cancer and control groups in mortality (11% vs 11%; Ñ = -0.002, p = 0.888). In conclusion, in patients with AF, cancer diagnosis may not add a predictive role for in-hospital CVA beyond the CHADS2VASc score.