Authors: Maxence Ficheux, Laure Peyro-Saint-Paul, Dorothée Balayn, Blandine Lecrux, Manon Brossier, Antoine Morin, Antoine Lanot, Chloé Peron, Marion Boulanger, Marie Brionne, Farzin Beygui, Jean-Jacques Parienti, Thierry Lobbedez, C Béchade
Categories: Renal Medicine, Anticoagulation, Dialysis, Randomized Controlled Trial, Protocol, 1728, 1506
Source: BMJ Open
Several randomised controlled trials have demonstrated that novel oral anticoagulants are safer compared with vitamin K antagonists for the management of non-valvular atrial fibrillation (NVAF) to prevent thromboembolic events in the general population. There is a growing interest in the use of apixaban in patients with end-stage renal disease (ESRD) undergoing peritoneal dialysis (PD) but there is a lack of randomised data in this population.
APIDP2 is a prospective parallel, randomised, open-label, blinded endpoint trial involving patients with ESRD undergoing chronic PD who have NVAF. A total of 178 participants will be recruited from 20 French PD centres. Eligible patients will be randomly assigned to receive either apixaban at a reduced dose of 2.5 mg two times per day (dose determined with the previous pharmacokinetic study APIDP1) or dose-adjusted to international normalised ratio (INR) target (2–3) coumadin therapy. Anticoagulation to prevent thromboembolic events will be initiated or changed according to the randomisation for a duration of 1 year. The primary outcome is a major or clinically relevant non-major bleeding from randomisation up to month 12, assessed according to the International Society on Thrombosis and Haemostasis Score. Secondary outcomes encompass an efficacy composite criterion combining stroke or transient ischaemic attack (TIA), cardiovascular death and thrombosis including myocardial infarction cumulated at 12 months. Bleeding events will be also classified according to Global Use of Strategies to Open Occluded Coronary Arteries (GUSTO) and Thrombolysis In Myocardial Infarction (TIMI) criteria and pharmacodynamics outcomes will evaluate the time within the INR target range of 2–3 in the warfarin arm over 1 year, and anti-Xa apixaban activity in case of bleeding events and at 1 month, 6 months and 12 months of follow-up in the apixaban arm. To demonstrate that apixaban is safer than warfarin at 1 year, assuming two interim analyses after 60 and 118 patients, a bilateral alpha risk of 5% and a power of 80%, 178 patients are needed in this randomised trial (effect size found from the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) Study among patients with creatinine clearance 25–30 ml/min), that is, 89 patients per group.
The study has been approved by the ethics committee Comité de Protection des Personnes Sud Est III – Lyon - FRANCE, CT number 2023-507544-37-00. Written informed consent is required for each participant. Findings will be presented at scientific meetings and published in peer-reviewed journals.
ClinicalTrials.gov, NCT06045858; European Clinical Trial System, CT number 2023-507544-37-00.
Keywords: Dialysis, Randomized Controlled Trial, Anticoagulation
The prevalence of atrial fibrillation (AF) in the general population is estimated to range around 1% depending on age, reaching 8% in patients over 80 years old. This prevalence is 20% and 14% in patients on haemodialysis and peritoneal dialysis (PD), respectively.^1^,^3^ In addition, AF is associated with higher morbidity and mortality rates in patients with advanced chronic kidney disease (CKD), in contrast with patients with preserved kidney function, with a higher risk of both bleeding and clotting.^4^
In the general population, novel oral anticoagulants (NOACs) demonstrated a greater safety over warfarin and among 269 patients with creatinine clearance (CrCl) 25–30 mL/min included in the Apixaban for Reduction in Stroke and Other Thromboembolic Events in Atrial Fibrillation (ARISTOTLE) Trial, apixaban caused less bleeding than warfarin.^5^ Furthermore, warfarin has been recognised as a risk factor for calcific uremic arteriolopathy (calciphylaxis), a rare yet serious complication.^6^ Additionally, patients with end-stage renal disease (ESRD) appear to have low adherence to warfarin therapy, with international normalised ratio (INR) time in the target range of only 44%–51% in three randomised controlled trials (RCTs).^7^,^9^
Considering the complex pathophysiology underlying the excess risk of stroke in patients with advanced CKD and the known complications of warfarin therapy, the use of NOACs in these populations may present an appealing alternative. Four significant reviews conducted in 2022 and 2023, based on RCTs and cohort studies, assessed both thrombotic and bleeding outcomes associated with anticoagulant use in advanced CKD. These reviews highlight the necessity for more RCTs.^10^,^13^
The two most significant cohorts to date that compared warfarin with NOACs (1) Siontis et al, who conducted a study involving a large cohort (2351 vs 7053) and found a reduction in major bleeding events with apixaban compared with warfarin, with similar efficacy in dialysis patients;^14^ and (2) Wetmore et al, who included 3130 patients versus 9086, and also found a reduction in major bleeding events with apixaban compared with warfarin in dialysis patients, while maintaining similar efficacy.^15^ In 2023, a third cohort study was published, involving 4313 apixaban users with stage 4 and stage 5 kidney disease. This study found that the use of 5 mg apixaban was associated with a higher risk of bleeding, with no difference in efficacy. These findings supported European dosing recommendations of 2.5 mg two times per day, which differ from the US recommendation of 5 mg two times per day in patients with AF and severe CKD.^16^ Additionally, two recent meta-analyses included the three RCTs^7^,^9^ comparing the safety of NOACs with that of warfarin. These analyses encompassed a total of 341 patients with AF and ESRD, with 176 in the NOAC group and 165 in the warfarin group. The results indicated that safety and efficacy outcomes were comparable, with no significant differences in incidences of stroke, mortality or bleeding.^17 18^
In terms of cardiology recommendations, two meta-analyses involving a total of 10 445 patients with ESRD and 24 335 patients on dialysis, comparing those who received Vitamin K Antagonist (VKA) with those who received no anticoagulation, revealed no significant benefit in preventing ischaemic stroke but did indicate an increased risk of bleeding. Consequently, European Society of Cardiology (ESC) guidelines do not provide any recommendations regarding the use of NOACs in patients with AF who also have ESRD or are undergoing dialysis.^19^
The results obtained from RCTs conducted on haemodialysis patients cannot be extrapolated to PD patients. The two patient populations are distinct, as PD patients do not receive heparin-based anticoagulation, which may increase the risk of bleeding events. Additionally, PD patients do not have arteriovenous fistulas that could potentially lead to bleeding.
Regarding apixaban pharmacokinetics, a preliminary PK study in 2016 by Wang suggested that haemodialysis has a moderate impact on the plasma concentration of apixaban.^20^ Subsequently, Mavrakanas conducted another PK study, revealing that haemodialysis was linked to supratherapeutic levels of apixaban and recommended a dose reduction.^21^
However, as there was a lack of data on the pharmacokinetics of apixaban in patients with ESRD undergoing PD, researchers initiated the APIDP1 Study with 12 PD patients. The results showed a very low PD apixaban clearance, resulting in significantly higher mean area under the curve (AUC) in patients treated with PD, 73% higher (p=0.01) than the mean AUC of healthy controls matched by age, weight and sex.^22^ Fung et al subsequently reported similar pharmacokinetic a proportion of PD patients had supratherapeutic levels even when the dosage was reduced to 2.5 mg two times per day.^23^ Then Lidgard and Shen, in a 2023 editorial, emphasised the importance of reducing the apixaban dosage for stroke prevention in patients with AF undergoing PD treatment, regardless of their age or weight.^24^
The APIDP group now aims to conduct this phase III study (APIDP2) to compare the safety of apixaban at the reduced dose to adjusted warfarin in patients with ESRD treated with chronic PD.
The objective of the APIDP2 Study is to determine whether apixaban 2.5 mg two times per day or dose-adjusted to INR target (2–3) warfarin is safer at 1 year in patients with ESRD treated with chronic PD.
The protocol was prepared in accordance with Standard Protocol Items: Recommendations for Interventional Trials (SPIRIT) guidelines. The SPIRIT checklist is provided as online supplemental material 1.
APIDP2 is a phase III study with parallel enrolment, a prospective randomised open blinded-endpoint design, and 1 year of follow-up. Open-label administration of anticoagulation was chosen because of the need to adjust warfarin to targeted INR. The study flow diagram is provided in figure 1.
Figure 1 Study flow diagram. INR, international normalised ratio.
Recruitment is scheduled to commence in October 2024 with a total of 178 randomised patients to be included. Considering the suitability of the 20 centres, it is anticipated that data collection will be completed around October 2027. The study is promoted by the Centre Hospitalo-Universitaire (University Hospital Center) (CHU) Caen Normandy.
The inclusion and exclusion criteria are shown in box 1.
Patients who are routinely followed by a nephrologist-investigator are registered in a screening list. Prior to the inclusion visit (which occurs 0–60 days before enrolment), the nephrologist-investigator will provide patients with written information about the study and explain its details. During the inclusion visit, eligibility will be determined, and all patients who meet the inclusion criteria and do not meet exclusion criteria will be invited to participate. Potential risks will also be explained, including the possibility of modification of previous anticoagulant treatment or initiation, which could result in ineffectiveness and clotting events, or excessive dosing, which could lead to bleeding events. The primary requirement of the APIDP2 Study is an additional screening visit. To compensate patients for constraints, the sponsor has chosen to provide €100. The informed consent form (online supplemental material 2) will be signed and dated by both the patient and the nephrologist-investigator.
Following written informed consent, patients will be assigned a subject number in the electronic case report form (eCRF). Baseline characteristics such as age, sex, weight, height, CHAD2DS2-VASC Score, nephropathology, residual glomerular filtration rate, PD, concomitant medications and medical history will be documented.
Eligible participants will be randomised 1 by a nephrologist-investigator via an interactive system (EnnovClinical) to apixaban or warfarin with minimisation algorithm including the following centre, age, sex, diabetes mellitus, prior duration of PD, history of transient ischaemic attack (TIA) or stroke, and whether anticoagulation is already initiated or not.
Patients are randomised to receive anticoagulation with adjusted warfarin to targeted INR 2.0–3.0 or apixaban (2.5 mg × 2/day) during 1 year according to the prescription protocol (table 1).
The experimental treatments will be provided by outpatient pharmacies on prescription by the nephrologist-investigator during the following study Initial Visit (V0), Month 1 (M1) and Month 6 (M6). The dispensation of the treatment will be tracked and documented by outpatient pharmacies in the patient’s notebook. Adherence to the treatment will be self-recorded by the patient in their notebook.
Reasons for discontinuation are elective surgical intervention, planned kidney transplantation within the next 3 days, pregnancy and any nephrologist-investigator decision that considers drug discontinuation as a benefit for the patient. All patients who prematurely discontinue the drug are asked to be followed up until the Month 12 (M12).
At M1, M6 and M12, data on outcomes will be bleeding and thromboembolic events, stroke, other AE and serious adverse events (SAEs), all causes of death, blood parameters, changes in anticoagulation prescription, in concomitant treatment. Pharmacodynamic parameters will be recorded (INR or anti-Xa apixaban). Adherence will be assessed by conducting interviews with the patient and entering data from the patient’s notebook.
The study endpoints are listed in box 2. All endpoints will be recorded over the duration of the study, at M1, M6 and M12.
While nephrologists, patients and monitors are not blinded to the allocation, the endpoint adjudication committee remains blinded to randomisation and the identities of the participants.
The nephrologist-investigator will report all events of interest (bleeding, thrombosis, stroke, death) on forms, associated with the patient medical record blinded and anonymised, that will be reviewed by the adjudication committee. The adjudication committee consists of three experts in nephrology, cardiology and neurology. The committee will have the opportunity to ask for further clarification.
The study monitoring is performed by the promotor CHU Caen Normandy, to ensure that data are complete and correctly registered constituent with the source data, in addition to reviewing compliance of the study protocol, laws and regulation. Three monitors are assigned to the project.
Adverse events and SAEs, other that clinical endpoint events, occurring during the study period will be documented and evaluated for severity and causality. Clinical endpoints including major or clinically significative bleeding, systemic embolism, stroke, myocardial infarction and all-cause death (systematically recorded), will be assessed by the nephrologist-investigator at months 1, 6 and 12, and with a final phone call at month 13.
The Data and Safety Monitoring Board (DSMB) is a consultative committee responsible for providing recommendations to the sponsor and investigator regarding the benefit/risk balance and the management of a clinical trial. The DSMB will meet prior to the first inclusion to validate the charter along with operational procedures. The DSMB will convene at month 1 after the twelfth patient in the apixaban arm. The DSMB will analyse pharmacodynamic data and safety data. Specifically, it will analyse the absence of accumulation by comparing apixaban anti-Xa activity between day 7 and month 1 on an intraindividual basis. The DSMB will then provide an opinion on the study’s continuation conditions.
The DSMB will subsequently convene for two interim analyses after 1 year of follow-up, with 60 and then 118 initial inclusions, and/or at any time in case of new safety information. The DSMB will review all adjudicated clinical endpoints and other SAEs, initially in blinded fashion, and then unblinded if necessary. DSMB members include nephrologists, cardiologists and pharmacologists that are not affiliated to the sponsor and free from many conflicts of interest.
The APIDP group formulated a hypothesis that patients undergoing PD may experience fewer bleeding events compared with patients undergoing haemodialysis (without fistula and without heparin bridging). In order to estimate the effect size, we considered the population studied in an RCT that was similar to our cohort, which was the ARISTOTLE population with a CrCl ranging between 25 mL/min and 30 mL/min, with a focus on the 88 patients (64%) in the apixaban arm at the reduced dose of apixaban.^5^ Apixaban 2.5 mg caused less bleeding than warfarin with 17/85 (20%) vs 5/87 (6%) major bleeding or clinically relevant non-major bleeding events, respectively, in the warfarin group versus the apixaban group.
The sample size estimation was computed with PROC SEQDESIGN (SAS software, V.9.4, Cary, North Carolina, USA). With a reference proportion of 20% (ie, in the control group), a two-sided alpha risk of 5%, 80% power, an effect size of −14% in absolute proportion (ie, 6% in the experimental group), two interim and one final analyses, the required total sample size is 60 for the first interim, 118 for the second interim and 178 for the final analysis, by the Haybittle sequential plan with decision rule for better safety as described in figure 2. We did not account for dropouts.
Figure 2 Statistical output of PROC SEQDESIGN in SAS V.9.4 for the APIDP2 Study.
Summary statistics will consist of frequency tables for categorical variables. For continuous variables, summary statistics will consist of the number of available observations, arithmetic mean, SD, median, minimum and maximum unless stated otherwise. Baseline characteristics will be presented by randomised groups in separate tables.
Regarding the primary endpoint, we plan to use a logistic regression model with the primary endpoint as the dependent variable and the group assignment as the independent variable with adjustment made for minimisation factors (ie, age, sex, diabetes mellitus, prior duration of PD, history of TIA or stroke, and whether anticoagulation is already initiated or not), as recommended. The effect size will be reported as an adjusted OR and 95% CI. The main analysis will follow the intent-to-treat principle and a p<0.04826 will be considered to be statistically significant for the final analysis. The first or second interim analyses will stop for better safety if p<0.0027 (after 60 or 118 total inclusions, respectively). The interim and final analyses will be conducted when all the intended participants reach a follow-up of 12 months. Subgroup analyses will be conducted according to age, sex, centre, PD duration (median cut-off), prior stroke, diabetes mellitus and initiation of switch to assess homogeneity of the effect size by the Breslow-Day interaction test.
Regarding efficacy endpoints, we plan to use a composite criterion combining the following events (at least one criterion) to avoid multiple statistical stroke or TIA, cardiovascular death and thrombosis including myocardial infarction cumulated at 12 months, which we will compare between randomised arms under an equivalence hypothesis. Additionally, we plan to perform exploratory sensitivity analyses with time-to-event data for the composite efficacy endpoint and its components separately, as recommended, using Kaplan-Meier curves and Cox models.
Regarding the descriptive safety analysis of the treatment, we plan to use counts (percentages) by randomised arms. We do not anticipate significant missing data, as this population is closely monitored by the healthcare system. In case of missing data, we plan to consider that no primary endpoint is reached in both groups and use multiple imputations for the primary endpoint as a sensitivity analysis. The statistical analysis will be performed by Professor Jean-Jacques Parienti at the University Hospital of Caen using SAS V.9.4 software (SAS Institute, Cary, North Carolina, USA).
The study data will be collected electronically using an eCRF designed with Ennov Clinical software by the data managers at the Clinical Research Unit of CHU Caen Normandy. Data entry will be performed by investigators from each centre and may also be delegated to individuals on the tasks list. The system used to create the eCRF and collect the required data will comply with the current regulations mandated by Commission nationale de l'informatique et des libertés (CNIL) (MR001). Access to the eCRF will be limited to authorised personnel including the sponsor, principal investigators and clinical trial assistants. Rights such as reading, writing and access to specific CRF pages will be allocated based on the individual’s role in the study and their centre. User authentication will be achieved through a unique username and password customised for each user, with all connections to the eCRF logged in an audit trail.
Coherence tests can be scheduled according to the study’s risk level in coordination with the study’s sponsor team. After these tests, queries (requests for clarification) may be sent to investigators to correct or confirm the data entered into the eCRF.
Patients were not involved in the conception of this research. The results will be disseminated to study participants through the European Clinical Trial System (CTIS), ensuring public access to the findings.
Submission to CTIS, CT Number 2023-507544-37-00, approved in May 2024; the study has been approved by the ethics committee Comité de Protection des Personnes Sud Est III – Lyon - FRANCE, CT number 2023-507544-37-00 and National Health Authority in France Agence Nationale de Sécurité du Médicament (Franch National Medicines Safety Agency) on 2024/05/24. Written informed consent (online supplemental material 2) is required for each participant. The full study protocol is provided as online supplemental material 3.
The study is expected to be completed by October 2027. Findings will be presented at scientific meetings and published in peer-reviewed journals.
The authors thank following nephrologist-investigators who will contribute to Elie Zagdoun, Alexandra Botte, Nathalie Maisonneuve-Housieaux, Guillaume Bonnard, Pierre Pfirmann, Ivona Milic, Aurore Barthelemy, Marie-Béatrice Nogier, Stanislas Bataille, Marine Lebreton-Semet, Awena Le Fur, Mack Kabongo, Mathilde Beaumier, Françoise Heibel, Ariane Duval-Sabatier, Hugoline Boulay, Guillaume Seret, Louise Nielsen, Célia Lessore. The authors also thank Danièle Debruyne for having brought its expertise in pharmacokinetics.
Maxence Ficheux, Email: ficheux-m@chu-caen.fr.
Laure Peyro-Saint-Paul, Email: peyrosaintpaul-l@chu-caen.fr.
Dorothée Balayn, Email: balayn-d@chu-caen.fr.
Blandine Lecrux, Email: lecrux-b@chu-caen.fr.
Manon Brossier, Email: brossier-ma@chu-caen.fr.
Antoine Morin, Email: morin-a@chu-caen.fr.
Antoine Lanot, Email: lanot-a@chu-caen.fr.
Chloé Peron, Email: peron-c@chu-caen.fr.
Marion Boulanger, Email: boulanger-m@chu-caen.fr.
Marie Brionne, Email: brionne-m@chu-caen.fr.
Farzin Beygui, Email: beygui-f@chu-caen.fr.
Jean-Jacques Parienti, Email: parienti-jj@chu-caen.fr.
Thierry Lobbedez, Email: lobbedez-t@chu-caen.fr.
C Béchade, Email: bechade-c@chu-caen.fr.