The association of atrial fibrillation with mortality in patients with breast cancer: author reply

This commentary refers to ‘Incidence, risk factors, and mortality of atrial fibrillation in breast cancer: a SEER-Medicare analysis’, by A. Guha et al., https://doi.org/10.1093/eurheartj/ehab745 and the discussion piece ‘The association of atrial fibrillation with mortality in patients with breast cancer’, by J. J. Noubiap, https://doi.org/10.1093/eurheartj/ehac004.

We would like to thank Dr Noubiap for presenting their excellently explained discussion forum¹ regarding our article in European Heart Journal.² Regarding Dr Noubiap’s first point, we have identified an error in the labelling of groups in Figures 2B, 3B and D, in which the labels were reversed in the final published manuscript. We provided a corrected figure (Figure 1), which now is consistent with the results presented in the tables (showing higher mortality in those with atrial fibrillation within 30-day after breast cancer diagnosis.) The second issue is regarding the absolute difference between those with and without atrial fibrillation and the small P-values for those differences. This is an inherent limitation of P-values, which cannot be used to determine the clinical significance of an association and, therefore, attention should be given to actual estimates of association and absolute risks. Finally, Dr Noubiap notes an apparent inconsistency between the P-value in Figure 3D (P = 0.005) and the corresponding association presented in Table 3 (HR 1.27, 95% CI 0.97–1.67). In review, the log-rank P-values which had been presented in Figures 2B, 3B and D with Cox-adjusted P-values wherein the corresponding P-value is 0.06 concordant with value in Table 3.

Figure 1.

Adjusted survival curves showing all-cause mortality (A) and cardiovascular mortality (B) among breast cancer patients who developed atrial fibrillation within 30 days of breast cancer diagnosis compared to those who did not at 1-year follow-up. (C) Adjusted survival curves showing cardiovascular mortality among breast cancer patients who had atrial fibrillation prior to breast cancer diagnosis compared to those who did not at 1-year follow-up. The plots are adjusted for standard demographic features, breast cancer-related features, cardiovascular risk factors for atrial fibrillation, and breast cancer medication.

Funding

This work was supported in part by NIH grants P30 CA016058, K12-CA133250, and K23-HL155890 (D.A.) and by American Heart Association-Robert Wood Johnson Foundation (Harold Amos) grant (D.A.). A.G. is supported by American Heart Association-Strategically Focused Research Network Grant in Disparities in Cardio-Oncology (nos. 847740 and 863620). A.A. is supported by NIH grant K24HL148521.

Conflict of interest: none declared.