Skip to main content
Download PDF
  • Research article
  • Open access
  • Published:

The impact of kidney function in patients on antithrombotic therapy: a post hoc subgroup analysis focusing on recurrent bleeding events from the AFIRE trial

BMC Medicine volume 20, Article number: 69 (2022) Cite this article

Abstract

Background

The success of antithrombotic therapies is assessed based on thrombotic and bleeding events. Simultaneously assessing both kinds of events might be challenging, and recurrent bleeding events are often ignored. We tried to confirm the effects of kidney function on outcome events in patients undergoing antithrombotic therapy.

Methods

As a post hoc subgroup analysis of the Atrial Fibrillation and Ischemic Events with Rivaroxaban in Patients with Stable Coronary Artery Disease (AFIRE) trial, a randomized clinical trial with a median follow-up of 36 months, patients were divided into high and low estimated glomerular filtration rate (eGFR) groups with a cutoff value of 50 mL/min. The cumulative incidence of bleeding and crude incidence of recurrent bleeding per 100 patient-years were calculated. We used the Cox regression model with multiple failure time data for recurrent bleeding events.

Results

Among 2092 patients, 1386 (66.3%) showed high eGFR. The cumulative bleeding events per 100 patients at 1 year were 5.4 and 6.2 in the high and low eGFR groups, respectively. The difference continued to increase over time. The hazard ratio for time to the first bleeding event in the high eGFR group was 0.875 (95% confidence interval 0.701–1.090, p = .234) and that for the first composite event was 0.723 (95% confidence interval 0.603–0.867, p < .000). The recurrent bleeding events per 100 person-years were 11.3 and 15.3 in the high and low eGFR groups, respectively, with a rate ratio of 0.738 (95% confidence interval 0.615–0.886, p = .001). During the observation period, the risk of bleeding changed with time. It peaked soon after the study enrollment in both groups. It decreased continuously in the high eGFR group but remained high in the low eGFR group.

Conclusions

We reaffirmed that kidney function affects bleeding events in patients on antithrombotic therapy, considering recurrent events. Patients should have detailed discussions with physicians regarding the possible bleeding events when continuing antithrombotic therapy, especially in patients with decreased kidney function.

Trial registration

UMIN Clinical Trials Registry, UMIN000016612. ClinicalTrials.gov, NCT02642419. Registered on 21 October 2015.

Peer Review reports

Background

A vast majority of antithrombotic therapies are available in the market. They have been usually assessed based on thrombotic events for efficacy and based on bleeding events for safety. These events are opposite to each other due to the nature of antithrombotic drugs. Additionally, the frequency and effect on prognosis are distinct. Fatal bleeding cases are relatively rare, while cases of nuisance bleedings are common, and their effect on the prognosis is generally limited [1]. In contrast, ischemic events are relatively less frequent but have a significant effect on the prognosis in the form of irreversible harm and often permanent disability.

Bleeding is common and can influence adherence to therapy or result in its disruption [2, 3]. Furthermore, even its severity might vary for each case, and the discontinuation of these drugs is known to increase the risk of thrombotic events [4]. Some patients who resume antithrombotic therapy may also develop bleeding events again. A consensus report recommended an individualized early resumption of antithrombotic therapy after bleeding events [4].

In addition to expert opinion, those recommendations were primarily based on previous studies. However, their data should be interpreted with caution due to concerns regarding the study settings. For example, composite endpoints are frequently used in clinical studies. However, they have been frequently criticized because each outcome event has a different effect on the overall health [5]. Furthermore, especially for antithrombotic therapy, most previous studies included only the first event as the endpoint [6], and the recurrent bleeding events were often ignored.

When considering therapy for a patient, the individual patient’s characteristics are also important. Although physicians attempt to assess each patient’s risk, the heterogeneity of patient characteristics makes applying a unified quantitative assessment of the absolute benefits and risks challenging [7]. For example, the prevalence of atrial fibrillation (AF) increases with decreasing kidney function [8], and patients with AF and chronic kidney disease have a high risk of both thromboembolic and hemorrhagic complications [9]. Meanwhile, non-vitamin K antagonistic oral anticoagulants (NOACs) are excreted mainly by the kidneys, and therefore, kidney function affects the outcome events in patients on NOACs [10].

In this study, we investigated the effects of kidney function on bleeding and thrombotic cardiovascular events, including recurrent bleeding in patients on antithrombotic therapy with rivaroxaban who were enrolled in the Atrial Fibrillation and Ischemic Events with Rivaroxaban in Patients with Stable Coronary Artery Disease (AFIRE) trial [11].

Methods

Study design and patients

This study was a post hoc subgroup analysis of the data acquired from the AFIRE trial; the study design and primary outcomes of the AFIRE trial have been published previously [11]. Briefly, the AFIRE trial enrolled 2215 patients from 294 centers with stable coronary artery disease with AF, and patients were randomly prescribed rivaroxaban monotherapy or rivaroxaban and antiplatelet agent combination therapy. The primary finding was that rivaroxaban monotherapy was non-inferior to combination therapy for a composite of stroke, systemic embolism, myocardial infarction, unstable angina requiring revascularization, or death. Furthermore, monotherapy was superior to combination therapy in terms of major bleeding.

In addition to the primary analysis stated above, other subgroup analyses have also been described previously [12]. The study protocol of the AFIRE trial itself was approved by the institutional review board (IRB) at each participating institution, and the study was performed in line with the Declaration of Helsinki. Of the 294 study centers, 61 did not have their own IRB, and they used the approval from the central ethics review board. All enrolled patients provided written informed consent.

Outcomes

In this report, we focused on any bleeding events and ischemic cardiovascular events, including myocardial infarction, unstable angina requiring revascularization, ischemic stroke, systemic embolism, and death due to any cause. We also included recurrent bleeding events observed in each patient during the observation period. However, in terms of ischemic cardiovascular events, we only included the first event in each patient. We evaluated these outcomes according to the kidney function as estimated glomerular filtration rate (eGFR), represented by creatinine clearance (CrCl) as estimated using the Cockcroft–Gault (CG) equation. We focused on patients whose eGFR data were available at enrollment in the AFIRE trial (n = 2092). The overall median follow-up was 36 months.

Statistical analysis

All baseline characteristics were assessed at the time of study enrollment. We divided the patients into high and low eGFR groups (eGFR ≥ 50 and < 50 mL/min, respectively) at the time of study enrollment, the cutoff point was determined according to the criteria indicated in the Japanese drug package insert. Continuous variables were compared using unpaired t-test, and the categorical variables were compared using the chi-square test or Fisher’s exact test, as appropriate.

Cumulative incidence of bleeding over time was calculated for the two groups. In this analysis, we considered the first bleeding event and death from any cause as competing risks [13,14,15]. Additionally, we performed a similar analysis with ischemic cardiovascular events, and the first event between bleeding and the ischemic cardiovascular event was considered for analysis.

We performed time-to-first event analysis for bleeding events as well as the first event analysis for composite endpoints of bleeding events, ischemic cardiovascular events, and death. We also adjusted for the differences in the patients’ baseline characteristics.

The crude incidence of recurrent bleeding per 100 patient-years of the follow-up period was calculated by dividing the total number of bleeding events in each eGFR group. The confidence intervals (CIs) were calculated using the quadratic approximation to the Poisson log-likelihood for the log-rate parameter [16]. From these results, the consequent incidence ratios were obtained [17]. Additionally, we used the negative binomial distribution model to modify an estimated rate ratio by assessing the wide variations in patients with respect to their risks of recurrent bleeding. This model used the backward selection method to select the independent covariates, including personal characteristics. Furthermore, we extended the composite of recurrent events to include all bleeding and ischemic cardiovascular events as well as death, i.e., we counted death as an additional event. The consequent rate ratio was the effect of eGFR on the composite of these events.

Since bleeding events might have occurred two or more times in a patient during the study period, we used the Cox regression model with multiple failure time data for bleeding events. We considered not only the first bleeding event but also the subsequent events. In this analysis, we assume that the hazard ratio may change with time; however, time was measured from study enrollment and was independent of the time the last event occurred. For the measured time, we used the time to each event from the time of the previous event using the conditional risk set model [18, 19] and estimated the course with the backward selection method to select independent covariates from the same variables in the previous negative binomial distribution model.

Sensitivity analysis

In this study, we have employed eGFR, as estimated by the CG equation, and used the cutoff point of 50 mL/min. However, other methods such as the Chronic Kidney Disease Epidemiology Collaboration (CKD-EPI) equation are commonly used to estimate GFR, and their accuracy is validated well in Western countries. Furthermore, the CG equation is said to overestimate GFR and exhibit large variability compared with the CKD-EPI equation [20]. As a sensitivity analysis, we also employed the CKD-EPI equation to estimate GFR, and two groups were distinguished by the cutoff at 50 mL/min/1.73 m2 to compare the estimated risk of bleeding events during the observation period. Moreover, we have also employed an additional cutoff point at 45 mL/min/1.73 m2 and 60 mL/min/1.73 m2 as estimated by the CKD-EPI equation since these points are used to define the CKD stage, and they are the closest lower and higher points to 50 mL/min/1.73 m2. We compared these results with the cutoff points at 45 mL/min and 60 mL/min estimated by the CG equation. From these sensitivity analyses, we have attempted to determine the robustness of our results, by different GFR estimation methods and through different cutoff points.

All tests were two-sided, and p values < .05 were considered to indicate a statistically significant difference. All analyses were performed using Stata SE v15.1 (Stata Corp LLC, College Station, TX, USA).

Results

Patient background data

Among the 2092 patients, the high eGFR group included 1386 patients (66.3%; mean eGFR 74.3 ± 21.8 mL/min), whereas the low eGFR group included 706 patients (33.7%; mean eGFR 38.7 ± 7.9 mL/min) (Table 1). There were several differences between the two groups. For example, the high eGFR group had younger patients (71.4 ± 7.6 vs. 80.3 ± 5.8 years, respectively, p < .001), more male patients (85.6% vs. 66.0%, respectively, p < .001), and higher hemoglobin level (13.9 ± 1.6 vs. 12.6 ± 1.7 g%, respectively, p < .001) than the low eGFR group. In contrast, the low eGFR group showed higher rates of heart failure (30.3% vs. 47.6%, respectively, p < .001) and kidney dysfunction (0.3% vs. 1.8%, respectively, p < .001) than the high eGFR group. Additionally, the CHADS2 score was higher in the low eGFR group than in the high eGFR group (median, 2.0 vs. 3.0, respectively, p < .001). There were no differences in terms of the details of the prescribed drugs at study enrollment between the two groups, except for the initial dose of rivaroxaban based on the study protocol.

Table 1 Clinical characteristics of the patients at study enrollment
Full size table

Incidence of outcome events

During the study period, up to 10 episodes of bleeding were observed per patient, and there were no differences in the bleeding incidence between the two groups (p = .546) (Table 2). There were 232 (16.7%) patients with at least one bleeding episode in the high eGFR group compared with 131 (18.6%) patients in the low eGFR group (p = .300). We observed up to 3 severe bleeding episodes, as defined by the International Society on Thrombosis and Haemostasis criteria, and the incidence of severe bleeding between the two groups (p = .096) did not differ significantly. The incidence of the other outcome events between the two groups did not differ significantly, except for systemic embolism and death from any cause, which was higher in the low eGFR group.

Table 2 Outcome events according to estimated glomerular filtration rate
Full size table

Cumulative incidence of bleeding

Figure 1 illustrates the cumulative number of bleeding events per 100 patients in the two groups over time, while death was considered a competing risk. By the end of the first year, the cumulative episodes of bleeding per 100 patients were 5.4 in the high eGFR group and 6.2 in the low eGFR group (difference of 0.8 per 100 patients). Beyond 1 year, the difference increased persistently (1.3 at 2 years and 1.7 at 3 years). Similarly, the cumulative number of bleeding and ischemic cardiovascular events per 100 patients was also calculated. By the end of the first year, the cumulative events per 100 patients were 6.4 in the high eGFR group and 7.8 in the low eGFR group (a difference of 1.4 per 100 patients). Beyond 1 year, the difference increased persistently (2.3 at 2 years and 2.9 at 3 years).

Fig. 1
figure 1

Estimated cumulative number of bleeding events per 100 patients. Time since enrollment (days)

Full size image

Estimation of the effect of the difference in eGFR

Table 3 summarizes the estimated effect of eGFR difference on all event outcomes. The hazard ratio in the high eGFR group compared to that in the low eGFR group was 0.875 [95% confidence interval (CI) 0.701–1.090, p = .234] for the first bleeding event and 0.723 (95% CI 0.603–0.867, p = .000) for the first composite events after adjustments. In the high and low eGFR groups, there were 314 and 198 bleeding events over 2783.4 and 1294.7 person-years of follow-up, respectively. Therefore, the bleeding rate per 100 person-years was 11.3 in the high eGFR group and 15.3 in the low eGFR group, with a rate ratio of 0.738 (95% CI 0.615–0.886, p = .0009) for recurrent bleeding as determined by the Poisson regression. The negative binomial regression model yielded a rate ratio of 0.860 (95% CI 0.680–1.088, p = .208) for the high eGFR group than that for the low eGFR group. When extending the composite outcomes other than bleeding events, the rate ratios remained constantly lower than those of the previous models in the high eGFR and low eGFR groups. The recurrent bleeding event analysis revealed larger differences than the time-to-first event analysis in the same composite outcome settings. Also, we assured that the effect of the interaction term between kidney function and the number of antithrombotic agents for all outcome analyses above was not significant.

Table 3 Comparison of the effect of estimated glomerular filtration rate difference among different endpoints
Full size table

The curves for the risk estimate of bleeding events over time stratified according to eGFR reflect the events’ dynamics (Fig. 2). Visual inspection showed a change in the risk of recurrent bleeding over the study period. It peaked soon after the study enrollment in both groups and decreased continuously in the high eGFR group; however, it remained high in the low eGFR group. The interaction term between antithrombotic therapy (monotherapy vs. combination therapy) and kidney function showed a significant effect in the models for the recurrent events (p = 0.042 for the model of recurrent bleeding events), but none for the first event model. In the recurrent event models, the therapy difference had a large impact on the high eGFR group.

Fig. 2
figure 2

Estimated risk of bleeding events, Cockcroft–Gault equation at 50 mL/min. Time since enrollment (days)

Full size image

Sensitivity analysis

The time course of the estimated risk of bleeding events for the two groups divided at 50 mL/min/1.73 m2, as estimated by the CKD-EPI equation, showed similar results with that from the CG equation at the cutoff point of 50 mL/min (Fig. 2, Fig. S1).

At the cutoff point of 45 mL/min/1.73 m2, as estimated by the CKD-EPI equation, both groups showed a peak soon after the observation was commenced, and another higher peak was observed in the low eGFR group, afterward. On the other hand, the risk of the high eGFR group decreased continuously. The low eGFR group had a higher bleeding risk than the high-eGFR group throughout the observation period. The shapes of these curves were similar to those estimated by the CG equation at the cutoff point of 45 mL/min (Figs. S2 and S3). At the cutoff point of 60 mL/min/1.73 m2, as estimated by the CKD-EPI equation, the estimated bleeding risk had decreased gradually, and increased afterwards in the low eGFR group, while it had decreased from the peak in the high eGFR group. Although the high eGFR group showed a peak of bleeding risk following enrollment, the shapes of the curves were similar between the different estimation methods for GFR (Figs. S4 and S5). Between the different estimation methods for GFR at the same cutoff point, a similar time course of the bleeding risk was also observed.

Discussion

In this study, we demonstrated the effects of kidney function, represented by eGFR, on the outcome events in patients from the AFIRE trial with stable coronary artery disease who were on antithrombotic therapy. We found that eGFR < 50 mL/min was related to a high outcome incidence, and the bleeding risk remained high over time in these patients.

Our findings are consistent with and extended the results of previous studies. Adverse medication-related outcomes in the studies for chronic kidney disease patients could be classified as those leading to kidney damage directly and other metabolic complications such as hyperkalemia and bleeding events [21]. In patients suffering from chronic kidney disease and receiving oral anticoagulant therapy, the reduction in thrombotic risk outweighs the bleeding risk in most cases and should be considered based on the balance between the benefits and risks [22]. For each type of thrombotic and bleeding event, composite endpoints are frequently used to demonstrate the summarized effect of the intervention. Although there are advantages to using composite outcomes, such as statistical power, we should also highlight their rationale. Composite outcomes are based on the assumption that a similar direction of the effect of the intervention will occur for each component of an aggregate outcome [5]. However, composite outcomes may include heterogeneous outcome endpoints with different magnitudes or directions as well as clinical impacts. In some studies, the newly proposed composite variables, such as net clinical benefits and net adverse clinical events, have also been used [23, 24]. Those variables can include both bleeding and thrombotic events, and they can be easily quantified as the summarized measures. Combining endpoints with different types of endpoints and directions can be problematic because the clinical effect of each event varies along with its severity and frequency. Therefore, we should examine the details of a composite event in a clinical context. Weighting methods have also been suggested to overcome these challenges [25, 26]. Each type of outcome event is assigned a specific weight based on the clinical effect on a patient. However, the weight for a type of outcome might not be consistent between patients and studies [27]. These findings suggest that although these new methods are promising, they are still under development for clinical use [28].

The results of the present study can be interpreted as not only an extension of the original AFIRE trial but also as complementary to that study. Our present results further explain the effect of kidney function. We also performed the analysis for recurrent bleeding events. Including recurrent events has been reported to result in greater power and more accurate estimation of the risks [29]. We added the first thrombotic cardiovascular events to the bleeding events to confirm the change in direction and magnitude. Although they should be interpreted with caution for the abovementioned composite outcomes, the comparable and consistent results demonstrated the effects of kidney function on the outcome events. Without including bleeding events, these results would have underestimated the effects of kidney function on the outcome events. The effect on the outcomes was quite large, as evident from the difference of incidence between the groups. It is expected that a bleeding event could have influenced the decision to continue the drugs and that discontinuation of drugs would consequently be related to the thrombotic events. However, our findings for the effect of kidney function on the outcomes would be consistent.

In our study, the bleeding risk increased in both groups after enrollment. However, the risk in the low eGFR group was higher than that in the high eGFR group. Over time, the bleeding risk in both groups decreased, but that in the low eGFR group remained high during this study period. It has been reported that the risk of bleeding in patients on anticoagulant therapy peaks immediately after the initiation of therapy [30]. However, the results from the said study might be biased due to the enrollment of prevalent drug users, in addition to the disease severity [31]. In our study, both groups had similar drug usage in addition to the other background characteristics at enrollment, which would have had little effect on our results. Therefore, we should consider resuming oral anticoagulants soon after bleeding events, especially in patients with preserved eGFR. In contrast, detailed discussions and considerations between patients and physicians are necessary, especially in patients with decreased eGFR. These findings are consistent with and extend expert consensus report, which recommended resuming oral antithrombotic therapy in all situations with clear indications, even in cases of major bleeding, as long as it is not a life-threatening intracranial or extracranial bleeding [4]. Our findings could contribute to facilitating communication between physicians and patients not only to consider the patient’s preferences but also to convey the physicians’ recommendations to each other. For such points, a recent clinical guideline recommended shared decision-making to explore patients’ values, goals, and preferences [32]. In contrast, all clinical guidelines for patients with decreased kidney function suggest dose adjustments for drugs including anticoagulants [33, 34]. Most of them have not suggested resuming anticoagulants after bleeding events. A recent expert consensus report recommended the resumption of anticoagulation after bleeding events [35]. Considering this lack of coherence between recommendations, our study results might be considered as offering significant evidence and providing a direction for future studies on chronic kidney disease patients with anticoagulants.

This study has several limitations. First, we used different analytic methods to assess the composite outcome events, which could include inherent problems. However, our results demonstrated that the effects of the differences in kidney function on the outcome events were consistent with those from previous studies. Second, this study was performed as a subanalysis of the AFIRE trial. In the original study protocol, the resumption of antithrombotic therapy after a bleeding event was expected to be as early as possible [12]. Furthermore, we did not include the dosages of the drugs used in this analysis. We have assumed that the subsequent decision to continue or discontinue the drugs after bleeding episodes would be similar between the two groups. In each case, these decisions were left to the physicians, which might not be consistent in all patients. However, it is unrealistic to use a unified protocol following bleeding events in future trials. Third, the study population included only Japanese patients with stable coronary artery disease, limiting the generalizability of our findings only to Japanese patients. However, most recent antithrombotic drugs called NOACs have similar renal excretion; therefore, our inferences might be consistent for different NOACs. Fourth, we used CrCl to estimate GFR and the cutoff point at 50 mL/min to succeed for the original AFIRE study protocol. Using other measures to show the kidney function classification might have been more relevant, such as CKD-EPI (Chronic Kidney Disease Epidemiology Collaboration) equations and CKD staging of KDIGO (Kidney Disease Improving Global Outcomes). However, given the imprecision in measures for estimating kidney function, individualization of the drug dosing based on each method is reasonable [33]. Moreover, in our sensitivity analysis, we have also employed the CKD-EPI equation and different cutoff points. Accordingly, we had confirmed the robustness of our findings where different methods of estimates ultimately led to similar results. Finally, our present study results were from prespecified subgroup analyses in addition to being a set of post hoc analyses after a randomized controlled clinical trial; it would be inevitable they are subject to inflated false-positive rates, from multiple testing [36]. Therefore, our results should be interpretated with caution, and future research would be expected; despite this, our findings might not necessarily change significantly.

Conclusions

In conclusion, our results reaffirmed that kidney function affects bleeding events in patients on antithrombotic therapy, considering recurrent events. We believe that patients on antithrombotic therapy, especially those with decreased kidney function, should receive detailed explanations from their physicians regarding possible bleeding events when continuing antithrombotic therapy.

Availability of data and materials

Data are available from the corresponding author upon reasonable request.

Abbreviations

AF:

Atrial fibrillation

AFIRE:

Atrial Fibrillation and Ischemic Events with Rivaroxaban in Patients with Stable Coronary Artery Disease trial

CG:

Cockcroft–Gault

CI:

Confidence interval

CKD-EPI:

Chronic Kidney Disease Epidemiology Collaboration

CrCl:

Creatinine clearance

eGFR:

Estimated glomerular filtration rate

NOACs:

Non-vitamin K antagonistic oral anticoagulants

References

  1. Steg PG, Bhatt DL. Is there really a benefit to net clinical benefit in testing antithrombotics? Circulation. 2018;137(14):1429–31.

    Article  Google Scholar 

  2. Roy P, Bonello L, Torguson R, de Labriolle A, Lemesle G, Slottow TLP, et al. Impact of “nuisance” bleeding on clopidogrel compliance in patients undergoing intracoronary drug-eluting stent implantation. Am J Cardiol. 2008;102(12):1614–7.

    Article  Google Scholar 

  3. Ben-Dor I, Torguson R, Scheinowitz M, Li Y, Delhaye C, Wakabayashi K, et al. Incidence, correlates, and clinical impact of nuisance bleeding after antiplatelet therapy for patients with drug-eluting stents. Am Heart J. 2010;159(5):871–5.

    Article  CAS  Google Scholar 

  4. Halvorsen S, Storey RF, Rocca B, Sibbing D, ten Berg J, Grove EL, et al. Management of antithrombotic therapy after bleeding in patients with coronary artery disease and/or atrial fibrillation: expert consensus paper of the European Society of Cardiology Working Group on Thrombosis. Eur Heart J. 2017;38(19):1455–62.

    CAS  PubMed  Google Scholar 

  5. Goldberg R, Gore JM, Barton B, Gurwitz J. Individual and composite study endpoints: separating the wheat from the chaff. Am J Med. 2014;127(5):379–84.

    Article  Google Scholar 

  6. Glynn RJ, Buring JE. Ways of measuring rates of recurrent events. BMJ. 1996;312(7027):364–7.

    Article  CAS  Google Scholar 

  7. Pocock SJ, Stone GW, Mehran R, Clayton TC. Individualizing treatment choices using quantitative methods. Am Heart J. 2014;168(5):607–10.

    Article  Google Scholar 

  8. Soliman EZ, Prineas RJ, Go AS, Xie D, Lash JP, Rahman M, et al. Chronic kidney disease and prevalent atrial fibrillation: the Chronic Renal Insufficiency Cohort (CRIC). Am Heart J. 2010;159(6):1102–7.

    Article  Google Scholar 

  9. Reinecke H, Brand E, Mesters R, Schäbitz W-R, Fisher M, Pavenstädt H, et al. Dilemmas in the management of atrial fibrillation in chronic kidney disease. J Am Soc Nephrol. 2009;20(4):705–11.

    Article  CAS  Google Scholar 

  10. Nielsen PB, Lane DA, Rasmussen LH, Lip GY, Larsen TB. Renal function and non-vitamin K oral anticoagulants in comparison with warfarin on safety and efficacy outcomes in atrial fibrillation patients: a systemic review and meta-regression analysis. Clin Res Cardiol. 2015;104(5):418–29.

    Article  CAS  Google Scholar 

  11. Yasuda S, Kaikita K, Akao M, Ako J, Matoba T, Nakamura M, et al. Antithrombotic therapy for atrial fibrillation with stable coronary disease. N Engl J Med. 2019;381(12):1103–13.

    Article  CAS  Google Scholar 

  12. Yasuda S, Kaikita K, Ogawa H, Akao M, Ako J, Matoba T, et al. Atrial fibrillation and ischemic events with rivaroxaban in patients with stable coronary artery disease (AFIRE): protocol for a multicenter, prospective, randomized, open-label, parallel group study. Int J Cardiol. 2018;265:108–12.

    Article  Google Scholar 

  13. Wolbers M, Koller MT, Stel VS, Schaer B, Jager KJ, Leffondré K, et al. Competing risks analyses: objectives and approaches. Eur Heart J. 2014;35(42):2936–41.

    Article  CAS  Google Scholar 

  14. Competing-risks regression. In: Stata survival analysis reference manual. College Station, TX: Stata Press. StataCorp; 2017. p. 154–179.

  15. Rogers JK, Pocock SJ, McMurray JJV, Granger CB, Michelson EL, Östergren J, et al. Analysing recurrent hospitalizations in heart failure: a review of statistical methodology, with application to CHARM-Preserved. Eur J Heart Fail. 2014;16(1):33–40.

    Article  Google Scholar 

  16. Calculate person-time, incidence rates, and SMR. In: Stata survival analysis reference manual. College Station, TX: Stata Press. StataCorp; 2017. p. 253–260.

  17. Report incidence-rate comparison. In: Stata survival analysis reference manual. College Station, TX: Stata Press. StataCorp; 2017. p. 249–252.

  18. Cox regression with multiplefailure data. In: Stata survival analysis reference manual. College Station, TX: Stata Press. StataCorp; 2017. p. 86–88.

  19. Cleves K. How do I analyze multiple failure-time data using Stata? Analysis of multiple failure-time survival data. Stata FAQ. Available at https://www.stata.com/support/faqs/statistics/multiple-failure-time-data/. Accessed 30 Jan 2022.

  20. Funakoshi Y, Fujiwara Y, Kiyota N, Mukohara T, Shimada T, Toyoda M, et al. Prediction of glomerular filtration rate in cancer patients by an equation for Japanese estimated glomerular filtration rate. Jpn J Clin Oncol. 2013;43(3):271–7.

    Article  Google Scholar 

  21. Whittaker CF, Miklich MA, Patel RS, Fink JC. Medication safety principles and practice in CKD. Clin J Am Soc Nephrol. 2018;13(11):1738–46.

    Article  Google Scholar 

  22. Ha JT, Neuen BL, Cheng LP, Jun M, Toyama T, Gallagher MP, et al. Benefits and harms of oral anticoagulant therapy in chronic kidney disease: a systematic review and meta-analysis. Ann Intern Med. 2019;171(3):181–9.

    Article  Google Scholar 

  23. Bergmark Brian A, Kamphuisen Pieter W, Wiviott Stephen D, Ruff Christian T, Antman Elliott M, Nordio F, et al. Comparison of events across bleeding scales in the ENGAGE AF-TIMI 48 Trial. Circulation. 2019;140(22):1792–801.

    Article  CAS  Google Scholar 

  24. Patrick WS, Mariusz T, Ply C, Rodrigo M, Norihiro K, Kuniaki T, et al. Patient-oriented composite endpoints and net adverse clinical events with ticagrelor monotherapy following percutaneous coronary intervention: insights from the randomised GLOBAL LEADERS trial. EuroIntervention. 2019;15(12):e1090–8.

    Article  Google Scholar 

  25. Armstrong PW, Westerhout CM, Van de Werf F, Califf RM, Welsh RC, Wilcox RG, et al. Refining clinical trial composite outcomes: an application to the Assessment of the Safety and Efficacy of a New Thrombolytic–3 (ASSENT-3) trial. Am Heart J. 2011;161(5):848–54.

    Article  Google Scholar 

  26. Bakal JA, Westerhout CM, Cantor WJ, Fernández-Avilés F, Welsh RC, Fitchett D, et al. Evaluation of early percutaneous coronary intervention vs. standard therapy after fibrinolysis for ST-segment elevation myocardial infarction: contribution of weighting the composite endpoint. Eur Heart J. 2012;34(12):903–8.

    Article  Google Scholar 

  27. Kaul S, Diamond GA. Trial and error. J Am Coll Cardiol. 2010;55(5):415–27.

    Article  Google Scholar 

  28. Armstrong PW, Westerhout CM. Composite end points in clinical research. Circulation. 2017;135(23):2299–307.

    Article  Google Scholar 

  29. Claggett B, Pocock S, Wei LJ, Pfeffer MA, McMurray JJV, Solomon SD. Comparison of time-to-first event and recurrent-event methods in randomized clinical trials. Circulation. 2018;138(6):570–7.

    Article  Google Scholar 

  30. Garcia DA, Lopes RD, Hylek EM. New-onset atrial fibrillation and warfarin initiation: high risk periods and implications for new antithrombotic drugs. Thromb Haemost. 2010;104(12):1099–105.

    Article  CAS  Google Scholar 

  31. Ray WA. Evaluating medication effects outside of clinical trials: new-user designs. Am J Epidemiol. 2003;158(9):915–20.

    Article  Google Scholar 

  32. Hindricks G, Potpara T, Dagres N, Arbelo E, Bax JJ, Blomström-Lundqvist C, et al. 2020 ESC Guidelines for the diagnosis and management of atrial fibrillation developed in collaboration with the European Association for Cardio-Thoracic Surgery (EACTS): the Task Force for the diagnosis and management of atrial fibrillation of the European Society of Cardiology (ESC) Developed with the special contribution of the European Heart Rhythm Association (EHRA) of the ESC. Eur Heart J. 2020;42(5):373–498.

    Article  Google Scholar 

  33. Turakhia MP, Blankestijn PJ, Carrero J-J, Clase CM, Deo R, Herzog CA, et al. Chronic kidney disease and arrhythmias: conclusions from a Kidney Disease: Improving Global Outcomes (KDIGO) Controversies Conference. Eur Heart J. 2018;39(24):2314–25.

    Article  CAS  Google Scholar 

  34. January CT, Wann LS, Calkins H, Chen LY, Cigarroa JE, Cleveland JC, et al. 2019 AHA/ACC/HRS Focused Update of the 2014 AHA/ACC/HRS Guideline for the Management of Patients With Atrial Fibrillation: a report of the American College of Cardiology/American Heart Association Task Force on Clinical Practice Guidelines and the Heart Rhythm Society in Collaboration with the Society of Thoracic Surgeons. Circulation. 2019;140(2):e125–51.

    Article  Google Scholar 

  35. Tomaselli GF, Mahaffey KW, Cuker A, Dobesh PP, Doherty JU, Eikelboom JW, et al. 2020 ACC Expert Consensus Decision Pathway on Management of Bleeding in Patients on Oral Anticoagulants: a report of the American College of Cardiology Solution Set Oversight Committee. J Am Coll Cardiol. 2020;76(5):594–622.

    Article  CAS  Google Scholar 

  36. Wang R, Lagakos SW, Ware JH, Hunter DJ, Drazen JM. Statistics in medicine—reporting of subgroup analyses in clinical trials. N Engl J Med. 2007;357(21):2189–94.

    Article  CAS  Google Scholar 

Download references

Acknowledgements

We thank the AFIRE investigators, Mebix, Inc., and the Japan Cardiovascular Research Foundation.

Funding

This work was supported by the Japan Cardiovascular Research Foundation based on a contract with Bayer Yakuhin, Ltd., who had no role in the design of the trial, collection or analysis of the data, interpretation of the trial results, or writing of the manuscript.

Author information

Authors and Affiliations

  1. Department of General Medicine and Primary Care, Kumamoto University Hospital, 1-1-1 Honjo, Chuo-ku, Kumamoto, Kumamoto, 860-8556, Japan

    Kunihiko Matsui

  2. Department of Cardiovascular Medicine, Tohoku University Graduate School of Medicine, 1-1 Seiryo-machi, Aoba-ku, Sendai, Miyagi, 980-8574, Japan

    Satoshi Yasuda

  3. National Cerebral and Cardiovascular Center, 6-1 Kishibe Shimmachi, Suita, Osaka, 564-8565, Japan

    Satoshi Yasuda

  4. Division of Cardiovascular Medicine and Nephrology, Department of Internal Medicine, Faculty of Medicine, University of Miyazaki, 5200 Kihara, Kiyotake, Miyazaki, 889-1692, Japan

    Koichi Kaikita

  5. Department of Cardiovascular Medicine, Graduate School of Medical Sciences, Kumamoto University, 1-1-1 Honjo, Chuo-ku, Kumamoto, Kumamoto, 860-8556, Japan

    Koichi Kaikita

  6. Department of Cardiology, National Hospital Organization Kyoto Medical Center, 1-1 Mukaihata-cho, Fukakusa, Fushimi-ku, Kyoto, Kyoto, 612-8555, Japan

    Masaharu Akao

  7. Department of Cardiovascular Medicine, Kitasato University School of Medicine, 1-15-1 Kitasato, Minami-ku, Sagamihara, Kanagawa, 252-0373, Japan

    Junya Ako

  8. Department of Cardiovascular Medicine, Faculty of Medical Sciences, Kyushu University, 3-1-1 Maidashi, Fukuoka, Fukuoka, 812-8582, Japan

    Tetsuya Matoba

  9. Division of Cardiovascular Medicine, Toho University Ohashi Medical Center, 2-22-36, Ohashi, Meguro-ku, Tokyo, 153-8515, Japan

    Masato Nakamura

  10. Department of Cardiovascular Medicine, Juntendo Tokyo Koto Geriatric Medical Center, 3-3-20 Shinsuna, Koto-ku, Tokyo, 136-0075, Japan

    Katsumi Miyauchi

  11. Department of Cardiology, Tokyo Women’s Medical University, 8-1, Kawada-cho, Shinjuku-ku, Tokyo, 162-8666, Japan

    Nobuhisa Hagiwara

  12. Cardiovascular Center, Yokohama City University Medical Center, 4-57, Urafune-cho, Minami-ku, Yokohama, Kanagawa, 232-0024, Japan

    Kazuo Kimura

  13. Department of Cardiology Osaka Police Hospital, 10-31 Kitayama-cho, Tennouji-ku, Osaka City, Osaka, 543-0035, Japan

    Atsushi Hirayama

  14. Kumamoto University, 2-39-1 Kurokami, Chuo-ku, Kumamoto, 860-8555, Japan

    Hisao Ogawa

Authors
  1. Kunihiko Matsui
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Satoshi Yasuda
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Koichi Kaikita
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Masaharu Akao
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Junya Ako
    View author publications

    You can also search for this author in PubMed Google Scholar

  6. Tetsuya Matoba
    View author publications

    You can also search for this author in PubMed Google Scholar

  7. Masato Nakamura
    View author publications

    You can also search for this author in PubMed Google Scholar

  8. Katsumi Miyauchi
    View author publications

    You can also search for this author in PubMed Google Scholar

  9. Nobuhisa Hagiwara
    View author publications

    You can also search for this author in PubMed Google Scholar

  10. Kazuo Kimura
    View author publications

    You can also search for this author in PubMed Google Scholar

  11. Atsushi Hirayama
    View author publications

    You can also search for this author in PubMed Google Scholar

  12. Hisao Ogawa
    View author publications

    You can also search for this author in PubMed Google Scholar

Contributions

KMa performed the statistical analyses and drafted the manuscript. SY and HO contributed to the study conception, design, and data acquisition. All authors contributed to the drafting and critical revision of the manuscript for important intellectual content. All authors have read and approved the final version of the manuscript for publication.

Corresponding author

Correspondence to Kunihiko Matsui.

Ethics declarations

Ethics approval and consent to participate

The trial was conducted according to the Declaration of Helsinki and approved by the institutional review board of the National Cerebral and Cardiovascular Center, Japan, along with the institutional review boards of all participating institutions. All participants provided written informed consent before enrollment.

Consent for publication

Obtained.

Competing interests

KMa has no relationships relevant to the contents of this paper to disclose. SY reports grants from Takeda Pharmaceutical, Abbott, and Boston Scientific and personal fees from Daiichi-Sankyo and Bristol-Meyers. KKa reports grants from Grants-in-Aid for Scientific Research (20K08451) from the Ministry of Education, Culture, Sports, Science and Technology of Japan and grants and personal fees from Bayer Yakuhin and Daiichi Sankyo. MA reports grants from the Japan Agency for Medical Research and Development (AMED), personal fees from Bristol-Myers Squibb and Nippon Boehringer Ingelheim, and grants and personal fees from Bayer Yakuhin and Daiichi Sankyo. JA reports personal fees from Bayer Yakuhin and Sanofi and grants and personal fees from Daiichi Sankyo. TM reports grants from Japan Cardiovascular Research Foundation and personal fees from Nippon Boehringer Ingelheim, Daiichi Sankyo, Astra Zeneca, and Bayer Yakuhin. MN reports grants and personal fees from Bayer Yakuhin, Daiichi Sankyo, and Sanofi and personal fees from Bristol-Myers Squibb and Nippon Boehringer Ingelheim. KMi reports personal fees from Amgen Astellas BioPharma, Astellas Pharma, MSD, Bayer Yakuhin, Sanofi, Takeda Pharmaceutical, Daiichi-Sankyo, Nippon Boehringer Ingelheim, and Bristol-Myers Squibb. NH reports grants and personal fees from Bayer Yakuhin, grants from Nippon Boehringer Ingelheim, and personal fees from Bristol-Myers Squibb. KKi reports grants from the Japan Cardiovascular Research Foundation; grants and personal fees from Bayer Yakuhin, Daiichi Sankyo, Sanofi, MSD, and AstraZeneca; and personal fees from Bristol-Myers Squibb and Nippon Boehringer Ingelheim. AH reports grants and personal fees from Boston Scientific Japan, Otsuka Pharmaceutical, Sanofi, Astellas Pharma, Bristol-Myers Squibb, Daiichi Sankyo, and Bayer Yakuhin; grants from Fukuda Denshi, Abbott Japan, Japan Lifeline, Takeda Pharmaceutical, and Sumitomo Dainippon Pharma; and personal fees from Toa Eiyo, Nippon Boehringer Ingelheim, Amgen Astellas BioPharma, and AstraZeneca. HO reports personal fees from Towa Pharmaceutical, Bristol-Meyers Squibb, Pfizer, Toa Eiyo, Bayer Yakuhin, and Novartis Pharma.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Figure S1.

Estimated risk of bleeding events, CKD-EPI equation at 50 ml/min/1.73m2. Time since enrollment (days).

Additional file 2: Figure S2.

Estimated risk of bleeding events, CKD-EPI equation at 45 ml/min/1.73m2. Time since enrollment (days).

Additional file 3: Figure S3.

Estimated risk of bleeding events, Cockcroft-Gault equation at 45 ml/min. Time since enrollment (days).

Additional file 4: Figure S4.

Estimated risk of bleeding events, CKD-EPI equation at 60 ml/min/1.73m2. Time since enrollment (days).

Additional file 5: Figure S5.

Estimated risk of bleeding events, Cockcroft-Gault equation at 60 ml/min. Time since enrollment (days).

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Matsui, K., Yasuda, S., Kaikita, K. et al. The impact of kidney function in patients on antithrombotic therapy: a post hoc subgroup analysis focusing on recurrent bleeding events from the AFIRE trial. BMC Med 20, 69 (2022). https://doi.org/10.1186/s12916-022-02268-6

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: https://doi.org/10.1186/s12916-022-02268-6

Keywords

BMC Medicine

ISSN: 1741-7015

Contact us