The Reply

Abstract

First of all, we thank Drs Shimoda and Watanabe very much for engaging in a relevant discussion and an interesting proposition, which spurred further discussion and analyses. They refer to our main analyses where an elevated risk of bleeding was found following systemic fluconazole use in patients with atrial fibrillation on apixaban.1Holt A Strange JE Rasmussen PV et al.Bleeding risk following systemic fluconazole or topical azoles in patients with atrial fibrillation on apixaban, rivaroxaban, or dabigatran.Am J Med. 2022; 135: 595-602.e5Abstract Full Text Full Text PDF Scopus (3) Google Scholar Considering that the majority of bleeding incidents related to systemic fluconazole use were gastrointestinal (65%),1Holt A Strange JE Rasmussen PV et al.Bleeding risk following systemic fluconazole or topical azoles in patients with atrial fibrillation on apixaban, rivaroxaban, or dabigatran.Am J Med. 2022; 135: 595-602.e5Abstract Full Text Full Text PDF Scopus (3) Google Scholar Shimoda and Watanabe propose that a disproportionate distribution of proton pump inhibitors (PPI) or histamine H2 receptor antagonist (H2RA) use could explain the findings. PPI use has indeed been linked to a lower risk of gastrointestinal bleeding among patients initiated on non-vitamin K antagonist oral anticoagulants (NOAC),2Ray WA Chung CP Murray KT et al.Association of oral anticoagulants and proton pump inhibitor cotherapy with hospitalization for upper gastrointestinal tract bleeding.JAMA. 2018; 320: 2221-2230Crossref PubMed Scopus (112) Google Scholar,3Cheung KS Leung WK Gastrointestinal bleeding in patients on novel oral anticoagulants: risk, prevention and management.World J Gastroenterol. 2017; 23: 1954-1963Crossref PubMed Scopus (110) Google Scholar making this question highly relevant. If PPI and H2RA use among patients on apixaban was low, the lack of the proposed protection would increase apixaban users’ susceptibility to the combination of apixaban and systemic fluconazole, thus increasing the risk of bleeding. On the contrary, if PPI or H2RA use was considerable and a protective effect was present, any association between systemic fluconazole use and bleeding would be diluted. At inclusion, 29.3%, 26.4%, and 24.2% of patients were on PPI correspondingly among patients on apixaban, rivaroxaban, and dabigatran—<0.1% were on H2RA in each group. Thus, the highest proportion of patients on PPIs was found in patients on apixaban, which—due to the self-controlled design—should ameliorate the risk in this group, compared with patients on rivaroxaban and dabigatran. To further investigate any mitigating abilities of PPI use, we conducted case-crossover analyses as described in the original paper using 30-day exposure windows linking systemic fluconazole use with bleeding in subgroups according to NOAC agent and PPI use (Figure).1Holt A Strange JE Rasmussen PV et al.Bleeding risk following systemic fluconazole or topical azoles in patients with atrial fibrillation on apixaban, rivaroxaban, or dabigatran.Am J Med. 2022; 135: 595-602.e5Abstract Full Text Full Text PDF Scopus (3) Google Scholar,4Maclure M The case-crossover design: a method for studying transient effects on the risk of acute events.Am J Epidemiol. 1991; 133: 144-153Crossref PubMed Scopus (1683) Google Scholar Among patients on apixaban, it seems that incident bleeding and systemic fluconazole use was strongly associated among patients not on PPIs (Figure). However, because systemic fluconazole was used very sparsely among patients with atrial fibrillation on NOACs,1Holt A Strange JE Rasmussen PV et al.Bleeding risk following systemic fluconazole or topical azoles in patients with atrial fibrillation on apixaban, rivaroxaban, or dabigatran.Am J Med. 2022; 135: 595-602.e5Abstract Full Text Full Text PDF Scopus (3) Google Scholar these analyses rely on few events, resulting in wide confidence intervals, and should be interpreted with that in mind. Nevertheless, with growing evidence of PPIs being associated with a protective effect among patients on NOACs,2Ray WA Chung CP Murray KT et al.Association of oral anticoagulants and proton pump inhibitor cotherapy with hospitalization for upper gastrointestinal tract bleeding.JAMA. 2018; 320: 2221-2230Crossref PubMed Scopus (112) Google Scholar,3Cheung KS Leung WK Gastrointestinal bleeding in patients on novel oral anticoagulants: risk, prevention and management.World J Gastroenterol. 2017; 23: 1954-1963Crossref PubMed Scopus (110) Google Scholar PPI could prove a relevant mitigating measure to decrease risk of gastrointestinal bleeding—particularly if prescribing interacting agents that could increase bleeding risk.1Holt A Strange JE Rasmussen PV et al.Bleeding risk following systemic fluconazole or topical azoles in patients with atrial fibrillation on apixaban, rivaroxaban, or dabigatran.Am J Med. 2022; 135: 595-602.e5Abstract Full Text Full Text PDF Scopus (3) Google Scholar,5Holt A Blanche P Zareini B et al.Gastrointestinal bleeding risk following concomitant treatment with oral glucocorticoids in patients on non-vitamin K oral anticoagulants.Heart. 2022; 108: 626-632Crossref Scopus (6) Google Scholar