In total,2,272 consecutive patients who were prescribedDOACs for AF (including paroxysmal AF)between September 2011 and January 2016 at Tachikawa General Hospital, Nagaoka, Japan, were retrospectively analyzed. All patients in the DOAC database were included in the analysis. However, patients with (1) valvular disease requiring surgery, (2) prosthetic mechanical heart valve, and (3) mitral stenosis were excluded from the study.

Data on patients’ baseline characteristics including age, sex, body weight, and renal function (creatinine clearance and creatinine levels); history, including comorbidities; and clinical outcomes during 2-year follow-up after DOAC prescription were collected. Furthermore, CHADS₂ score, CHA₂DS₂-VASc score, and HAS-BLED score were determined ¹⁴-¹⁶. The institutional database used in this study was approved by our local ethics committee, and informed consent was obtained from all patients.

Four DOACs (apixaban [n=1,014], edoxaban [n=267], rivaroxaban [n=498], and dabigatran [n=493]) were prescribed for AF at the discretion of the physician. The prescription dose was basically in accordance with themanufacturer’s label recommendations in Japan, including thereduced dose recommendations.For apixaban, a reduced dose is recommended for patients with at least two of the following characteristics: age ≥80 years, weight ≤60kg, or serum creatinine level ≥1.5mg/dL.For edoxaban, a reduced dose is recommended for patients with moderate or severe renal impairment (creatinine clearance of 15–49 mL/min), patients weighing ≤60 kg, and patients being concomitantly treated with interacting drugs (e.g., verapamil). For rivaroxaban, a reduced dose is recommended for patients with moderate renal impairment (creatinine clearance of 15–49mL/min). For dabigatran, a reduced dose is consideredfor elderly patients aged ≥70 years, patients with moderate renal impairment (creatinine clearance of 30–49mL/min), patients concomitantly treated with interacting drugs (e.g., verapamil), and patients at a high risk ofbleeding. However, an inappropriately high or low dose may similarly be prescribed based on the doctor's discretion. Hence, in this study, classifications of DOAC prescriptions, including an appropriate dose and an inappropriate high or low dose, were considered based on the manufacturer’s label recommendations in Japan, as mentioned earlier.

In this study, patients were classified into the following two groups according to a cut-off value of 50 mL/min for creatinine clearance at registration: non-CKD group (n=1460) and CKD group (n=812). A comparison of baseline characteristics, the ratio of the appropriate dose useto the inappropriate dose use, including an inappropriate low dose, andclinical outcomesbetween the two groupswere determinedfor all four DOACs and each type ofDOAC. In this study, if the types of DOAC were changed or the DOAC was discontinued, the patients were excluded from the cohorts but included in the data analysis.

The primary outcome was stroke, including transient ischemic attacks (TIA) and hemorrhagic stroke and systemic embolism.The main safety outcome was major bleeding, defined using the Randomized Evaluation of Long-Term Anticoagulant Therapy criteria ⁵, ¹⁰.Furthermore, the rate of any cause of death was determined.

In this study, the mean age of patients was 72±10 years, with CHADS>₂, CHA₂DS₂-VASc, and HAS-BLED scoresof 1.95±1.32, 3.21±1.67, and 1.89±0.96, respectively. The mean follow-up time was 1.6 years. Overall, the incidence rates of major bleeding and thromboembolic events were 2.3 and 2.1 per 100patient-years, respectively.

Overall, the CKD group had significantly higher mean age and CHADS₂,CHA₂DS₂-VASc, and HAS-BLED scores than the non-CKD group. Nevertheless, body weight was significantly lowerin the CKD groupthan in the non-CKD group. In our database of a single-center registry, 23% of patients with AF treated with a DOAC received an inappropriate dose.Regardingthe analysis of inappropriate dose use, the use of inappropriate low dose was obviously dominant compared to the use of inappropriate high dose (94.4 vs. 5.6%) [Table 3(a)]. On the contrary, the ratio of appropriatedose use was significantly higher in the CKD group than in the non-CKD group (85.7 vs.72.9%, p<0.01).

However, the ratio of appropriate dose usein the CKD group was significantly lower for apixaban than other DOACs [Table 3(b)]. In contrast, the ratio of appropriate dose use in the non-CKD group was significantly higher for edoxaban thanfor other DOACs. Based on the definition of CKD in this study, creatinine clearance was significantly lower in the CKD group than in the non-CKD group (39.3±8.4 vs. 74.2±21.2 mL/min, p<0.01).

Regardingeach type of DOAC, a similar tendency was observed in the comparison of baseline characteristics between the CKD and non-CKD groups—i.e., significantlyhigher values for mean age, appropriate dose use, and CHADS₂, CHA₂DS₂-VASc, and HAS-BLED scores in the CKD group and significantly lower body weight in the CKD group. However, for apixaban, the ratio of appropriate dose use in the CKD group was comparable to that in the non-CKD group (79.8 vs.75.2%, p=0.38).

Overall, the CKD group showed significantly higher incidence rates of death from any cause, including major bleeding and stroke/systemic embolism,than the non-CKD group (death from any cause: 4.5 vs. 1.1 per 100patient-years, p<0.01; major bleeding: 3.3 vs.1.4 per 100patient-years, p<0.01; stroke/systemic embolism:3.6 vs. 1.6 per 100patient-years, p<0.01). The Kaplan–Meier event rate curves indicated that major bleeding and stroke/systemic embolism werefrequentlyobserved in the CKD group [Figure 1]. In addition, multiple logistic regression analysis was performed inboth groups.In the CKD group, age and previous stroke history were independent determinants of stroke/systemic embolism, and low body weight had a tendency to be a determinant of major bleeding [Table 4A].On the contrary, in the non-CKD group, age and prior stroke history were independent determinants of stroke/systemic embolism, and ageand prior bleeding history tended to be determinants of major bleeding [Table 4B]. Concerningeach type of DOAC, for all DOACs excludingapixaban, the incidence rates of major bleeding and stroke/systemic embolism in the non-CKD group were comparable to those in the CKD group. In contrast, for apixaban, the incidence rate of stroke/systemic embolism tended to be higher in the CKD group than in the non-CKD group (3.8 vs. 1.6 per 100patient-years, p=0.03).

Figure 1. (A) Kaplan Meier curve for major bleeding between the chronic kidney disease (CKD) group and the non-CKD group The incidence of major bleeding in the CKD group was significantly higherthan that in the non-CKD group (Logrank: p<0.01). (B) Kaplan Meier curve for stroke/systemic emboli between the CKD group and the non-CKD group The incidence of stroke/systemic emboli in the CKD group was significantly higher than that in the non-CKD group (Logrank: p<0.01).

---

---

Severalstudies have reported a substantial relationship amongAF, CKD, age, and body weight (¹-³, ¹⁰, ¹⁷-¹⁹), andolder age is associated with the occurrence of AF ¹-³.CKD and AF share numerous risk factors and conditionsthat promote their incidence.Therefore, it is established that CKD increases the incidence of AF¹⁰, ¹⁹.In contrast,AF elimination by catheter ablation was associated with improvement in renal function at 1-year follow-up in patients with mild to moderate renal dysfunction ²⁰, indicating that AF equallyaccelerates CKD progression.

Similarly, a relationship between age and CKD has been reported ²¹. The absolute glomerular filtration rate (GFR) value decreased with an increase in age²⁰. Furthermore, Japanese patients with AF are generally small and lean²².According to a previous report, frailty in elderly patients increased with an increase in age ²³. Body loss has equallybeen defined as one of the criteria for frailty²⁴. Thus, there seems to be a relationship amongCKD, age, and bodyweight somewhat, and we have to consider that AF patients with CKD tend to be olderand have lower body weight. In this study, AF patients with CKD were older and had lower body weight than those without, which isconsistentwith the findings of the aforementioned studies.

Several reports have described the association between renal function and major bleeding or stroke/systemic embolism(¹⁰, ²⁵-³²). In addition, older age is considered a predictor of major bleeding events³³, ³⁴ and stroke/systemic embolism³³, ³⁵, ³⁶. In addition, it is actually included as one of the factors forthe CHADS₂, CHA₂DS₂-VASc, and HAS-BLED scores. In this study, age was an independent predictor of stroke/systemic embolism, agreeingwith the abovementioned finding.Low body weight tended to be a predictor of major bleeding in this study. Nevertheless,the relationship between body weight and major bleeding remainsfairly controversial³⁷-⁴⁰. However, several studies have described that lean patients were at higher risk for bleeding complications ⁴¹-⁴⁴. DOACs seemed safer and more effective than vitamin K antagonists, even in patients with low body weight ⁴⁵. However,the risk of increased bleeding by DOACsin underweight patients has been highlighted in previous studies⁴⁶, ⁴⁷.In summary, high age and low body weight, which are specific characteristics of patients with CKD, may be closely associated with an increase in major bleeding or stroke/systemic embolism.

Several previous reports have described thata combination of vitamin K antagonist and antiplatelet therapy is associated with a high annual risk of fatal and non-fatal bleeding episodes ⁴⁸, ⁴⁹. However, an observational study on real-world Asian patients treated with DOAC and antiplatelet agents showed no incremental bleeding risk (of additional antiplatelet therapy to DOAC) in a large pooled population⁵⁰. Therefore, our study findings may be compatible with previous results.

For apixaban, the incidence of stroke was higher in AF patients with CKD than in those without CKD in the present registry, andthis finding might be very suggestive. The prescription of an inappropriate low dose is often observed in the realworld. Theoretically, an inappropriate low dose might lead to increased safety but decreased effectiveness ⁵¹.

The renal clearance ratio of apixaban is approximately 25% ⁵². Asub-analysis of ARISTOTLE trial datarevealed that treatment with apixaban led to an apparent reduction in the incidence of major bleeding in patients with creatinine clearance <50mL/min compared to treatment with vitamin K antagonists. Thus, apixaban prescription might be logicallyeffective in patients with CKD. However, the use of appropriate apixaban dose was significantly lower in the present registry. This is probably because patients with CKDhave been reported to be predisposed to major bleeding,as mentionedearlier; additionally, themanufacturer’s label recommendations for the reduceddose of apixaban (at least two of the following characteristics: age ≥80 years, weight ≤60kg, or serum creatinine level ≥1.5mg/dL) are more intricate than those for other DOACs. In addition, the reduced dose of apixaban is one-half the standard dose (2.5 mg vs. 5 mg). In contrast, the reduced dose of rivaroxaban and dabigatran approximately rangesbetween 70-75% of the standard dose (15 mg vs. 20 mg and 110mg vs.220mg, respectively). Furthermore, the number of patients in the low-dose DOAC groups of the ARISTOTLE trials was not sufficientenough to establish superiority, equivalence, or non-inferiority ⁵³. Though speculative, there might be an association between dose regimens and medication compliance. A systemic review described that the prescribed number of doses per day was inversely related to compliance ⁵⁴.

These findings could explain the possible increase in the incidence of stroke/systemic embolism byan inappropriate low dose of apixaban⁵⁵,⁵⁶.The difference in inappropriate low dose among the four DOACs might have influenced the resultsof this study;nonetheless, there was low statistical power for the evaluation of the effects of inappropriate low dose on clinical outcomes. Therefore, future studies involving larger populations with long-term follow-up will be needed.