Reply

Abstract

We appreciate the insightful comments provided by Dr Wang and Dr Yang on our study.1Wang X. Yang L. Clin Gastroenterol Hepatol.https://doi.org/10.1016/j.cgh.2022.09.013Google Scholar They raised questions of whether the higher upper gastrointestinal bleeding (GIB) rates of warfarin compared with direct oral anticoagulants (DOACs) in the current study may have been related to inclusion of patients with reduced-dose DOACs and/or concomitant gastroprotective medication use. Indeed, one of the difficulties in comparing warfarin with DOACs, is that DOACs are administered as a standard or reduced-dose treatment depending on factors that affect its metabolism or elimination, such as kidney function. Meanwhile, the dosing of warfarin is titrated according to international normalized ratio values. Previous observational studies have commonly compared warfarin with a mixed population of standard and reduced-dose DOACs. However, concerns have been raised about the frequency of inappropriate lower dosing of DOACs because of perceived higher bleeding risk. Indeed, a study based on a prospective registry database of DOAC users, demonstrated that 9.4% of DOAC users were inappropriately underdosed, whereas 3.4% were inappropriately overdosed.2Steinberg B.A. et al.J Am Coll Cardiol. 2016; 68: 2597-2604Crossref PubMed Scopus (307) Google Scholar To explore whether inclusion of reduced-dose DOACs may have contributed to our findings, we compared standard and reduced-dose DOACs using the same statistical methods as in the original study.3Ingason A.B. et al.Clin Gastroenterol Hepatol.https://doi.org/10.1016/j.cgh.2022.06.033Google Scholar After accounting for age, gender, treatment indication, comorbidities, concomitant drug use, area of residence, and type of DOAC used, rates of upper GIB were similar between standard and reduced-dose DOACs (hazard ratio, 0.89; 95% confidence interval, 0.50–1.56). This could be expected as reduced-dose DOACs are first and foremost prescribed to achieve therapeutic levels in the setting of reduced drug metabolism or elimination. Differences in upper GIB rates between warfarin and DOACs in our study are therefore unlikely to be explained by inclusion of different DOAC dosing regimens. Although proton pump inhibitors (PPIs) are effective in treating peptic ulcer disease (PUD),4Scally B. et al.Lancet Gastroenterol Hepatol. 2018; 3: 231-241Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar it is debatable whether routine use of PPIs in patients receiving oral anticoagulation results in reduced upper GIB rates in the absence of PUD. An observational study published in 2018 demonstrated that concomitant use of PPIs in oral anticoagulant users was associated with 34% reduction in upper GIB.5Ray W.A. et al.JAMA. 2018; 320: 2221-2230Crossref PubMed Scopus (112) Google Scholar A randomized controlled trial of patients with coronary artery disease receiving low-dose rivaroxaban demonstrated that co-therapy with pantoprazole resulted in reduced rates of PUD-induced GIB (hazard ratio, 0.25; 95% confidence interval, 0.07–0.89).6Moayyedi P. et al.Gastroenterology. 2019; 157: 403-412Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar However, all-cause upper GIB rates were similar between the 2 arms (hazard ratio, 0.93; 95% confidence interval, 0.60–1.47).6Moayyedi P. et al.Gastroenterology. 2019; 157: 403-412Abstract Full Text Full Text PDF PubMed Scopus (81) Google Scholar In the current study, PUD was proportionally a more common cause of upper GIB for DOACs compared with warfarin. However, absolute rates of PUD-induced GIB were similar between warfarin and DOACs. Therefore, the higher rates of upper GIB for warfarin seem to be mostly driven by non-PUD. Importantly, the concomitant use of PPIs was accounted for in our inverse probability weighting model, which reduces the possibility that difference in PPI use explains the observed difference in upper GIB rates between warfarin and DOACs in the current study. In summary, the higher rates of upper GIB for warfarin compared with DOACs in the current study do not seem to be explained by the inclusion of reduced-dose DOACs or differences in concomitant gastroprotective medication use.