Abstract
Background
Nonvitamin K oral anticoagulants require dose adjustment based on kidney function.The most common estimate of kidney function employed in clinical practice is estimated glomerular filtration rate (eGFR); however, product monographs recommend the use of the Cockcroft‐Gault estimated creatinine clearance (eCrCl) for dose adjustment.
Methods and Results
The authors included patients enrolled in the ORBIT‐AF II (Outcomes Registry for Better Informed Treatment of Atrial Fibrillation AF II) trial. Dosing was considered inappropriate when use of eGFR resulted in a lower (undertreatment) or higher (overtreatment) dose than that recommended by the eCrCl. The primary outcome of major adverse cardiovascular and neurological events was a composite of cardiovascular death, stroke or systemic embolism, new‐onset heart failure, and myocardial infarction. Among 8727 in the overall cohort, agreement between eCrCl and eGFR was observed in 93.5% to 93.8% of patients. Among 2184 patients with chronic kidney disease (CKD), the agreement between eCrCl and eGFR was 79.9% to 80.7%. Dosing misclassification was more frequent in the CKD population (41.9% of rivaroxaban users, 5.7% of dabigatran users, and 4.6% apixaban users). At 1 year, undertreated patients in the CKD group had significantly greater major adverse cardiovascular and neurological events (adjusted hazard ratio, 2.93 [95% CI, 1.08–7.92]) compared with the group with appropriate nonvitamin K oral anticoagulants dosing (P=0.03).
Conclusions
The prevalence of misclassification of nonvitamin K oral anticoagulants dosing was high when using eGFR, particularly among patients with CKD. Among patients with CKD, potential undertreatment due to inappropriate and off‐label renal formulae may result in worse clinical outcomes. These findings highlight the importance of using eCrCl, and not eGFR, for dose adjustment in all patients with AF receiving nonvitamin K oral anticoagulants.
Keywords: atrial fibrillation, nonvitamin K oral anticoagulant, ORBIT‐AF II, renal dose adjustment
Subject Categories: Atrial Fibrillation, Arrhythmias
Nonstandard Abbreviations and Acronyms
- CKD‐EPI
Chronic Kidney Disease Epidemiology Collaboration
- eCrCl
estimated creatinine clearance
- ISTH
International Society on Thrombosis and Hemostasis
- MACNE
major adverse cardiovascular and neurological events
- MDRD
Modified Diet in Renal Disease
- NOAC
nonvitamin K oral anticoagulant
- ORBIT‐AF II
Outcomes Registry for Better Informed Treatment of Atrial Fibrillation AF II
Clinical Perspective.
What Is new?
While clinical trials and product monographs recommend the use of the Cockcroft‐Gault estimated creatinine clearance equation for dose adjustment, estimated glomerular filtration rate is the most commonly used renal formula in clinical practice.
Using estimated glomerular filtration rate instead of the Cockcroft‐Gault creatinine clearance equation, there is up to 11.5% misclassification in the general population and up to 42% misclassification in the chronic kidney disease population.
Undertreated patients in the chronic kidney disease group had significantly greater major adverse cardiovascular and neurological events as compared with the group with appropriate nonvitamin K oral anticoagulant dosing.
What Are the Clinical Implications?
These findings highlight the importance of using the Cockcroft‐Gault creatinine clearance equation, and not estimated glomerular filtration rate, for dose adjustment in all patients with atrial fibrillation receiving nonvitamin K oral anticoagulants.
Atrial fibrillation (AF) is associated with a 5‐fold increased risk of stroke and systemic embolism compared with the age‐matched population. 1 , 2 Over the past decade, nonvitamin K antagonist oral anticoagulants (NOACs) have emerged as the preferred anticoagulant. Compared with warfarin, NOACs have an equal or lower bleeding risk and are at least equally effective in thromboembolism prevention. 3 However, NOACs require renal dose adjustment. Up to 1 in 8 patients may receive doses without proper renal adjustment, 4 and both overdosing and underdosing of NOACs are associated with worse clinical outcomes. 5 Perception of higher bleeding risk, frailty, cognitive status, older patient age, presence of renal failure, and lack of familiarity with dosing guidelines have been proposed to explain this observation. 6 , 7
An alternative explanation is the inadvertent misdosing based on assessment of renal function. In routine clinical practice, estimated glomerular filtration rate (eGFR) is automatically reported by laboratories and is commonly used for renal medication dosing 8 using either the Modified Diet in Renal Disease (MDRD) or Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) formulae. 9 In contrast, the pivotal NOAC clinical trials used the Cockcroft‐Gault estimated creatinine clearance (eCrCl) equation for renal dosing, 10 , 11 , 12 which, in turn, is recommended in product monographs and by regulatory bodies. Differences in renal function between eGFR and eCrCl may result in unintentional, inappropriate renal dosing of NOACs as well as worse clinical outcomes. 13 , 14 In fact, a recent study among patients with chronic kidney disease (CKD) with AF showed that using MDRD and CKD‐EPI eGFR instead of the eCrCl resulted in a misclassification rate of 36%, with a majority being undertreated. 14 However, whether potential misclassification leads to worse outcomes is unknown.
In this study, we aim to: (1) assess the potential misclassification of NOAC renal dosing using eGFR compared with the eCrCl standard in a contemporary registry database of patients with AF receiving NOACs; (2) assess the clinical outcomes for patients who received appropriate renal dosing versus those who were undertreated or overtreated for each of the NOACs; and (3) determine whether the rates of misclassification or outcomes differ depending on baseline renal function, including those with CKD.
Methods
The data that support the finding of this study are available from the corresponding author upon reasonable request. This study has been approved by an institutional review committee and all of the patients have given informed consent.
Study Population and Design
Our study included patients enrolled in the ORBIT‐AF II (Outcomes Registry for Better Informed Treatment of Atrial Fibrillation AF II) registry, a national prospective registry of patients with AF who were enrolled from February 2013 to July 2016 at 243 sites across the United States. We excluded all patients who were not taking NOACs, did not have baseline renal function data, were undergoing dialysis, had an eCrCl or eGFR <15 mL/min, had mechanical valve replacement or severe mitral stenosis, or did not have follow‐up data.
Renal Function Estimation
The eCrCl was calculated using the Cockcroft‐Gault equation: (140 – age) × weight (kg) × [0.85 if female]/72×[serum creatinine (mg/dL)]. MDRD eGFR was calculated using the formula: (186×[serum creatinine (mg/dL)]−1.153×[age]−0.203 [×1.212 if Black] [× 0.742 if female]). 15 CKD‐EPI was calculated using eGFR [CKD‐EPI]=(141×min[serum creatinine [mg/dL]/k, 1]α) × max(serum creatinine [mg/dL]/k, 1)−1.209×0.993Age [×1.018 if female] [×1.159 if Black]. The “k” is 0.7 for female and 0.9 for male, “α” is −0.329 for female and −0.411 for male, “min” is the minimum of serum creatinine/k or 1, and “max” is the maximum of serum creatinine/k or 1. 16
NOAC Dose Adjustment
NOAC dose adjustment and medication eligibility is based on eCrCl, as studied in the landmark trials resulting in regulatory approval. While rivaroxaban uses the eCrCl as criteria for dose adjustment and medication eligibility, apixaban and dabigatran use the eCrCl as a cutoff for determining only medication eligibility. Specifically, rivaroxaban 20 mg once daily is used for patients with eCrCl >50 mL/min and rivaroxaban 15 mg once daily is used for patients with an eCrCl of 30 to 50 mL/min, with rivaroxaban being contraindicated for eCrCl <30 mL/min. 17 For dabigatran and apixaban, there was no dose adjustment based on the eCrCl. Instead, these agents were contraindicated with eCrCl <30 mL/min and eCrCl <25 mL/min, respectively. 18 , 19
Exposure of Interest
Agreement was defined as an eGFR (by MDRD or CKD‐EPI) and eCrCl recommending the same NOAC dose. Undertreatment was defined as when eGFR (by MDRD or CKD‐EPI) recommended a lower NOAC dose than would be recommended by the eCrCl equation. Overtreatment was defined as the opposite. For apixaban and dabigatran, renal function was used as a cutoff for assessing eligibility. Therefore, overtreatment was defined as an eGFR recommending NOAC use when it would be contraindicated based on the eCrCl criteria. The index time of classification is at the time of enrollment.
Outcomes
The primary outcome was major adverse cardiovascular and neurological events (MACNE), a composite of cardiovascular death, stroke or systemic embolism/transient ischemic attack, and new‐onset heart failure or myocardial infarction. The secondary outcomes included all‐cause death, cardiovascular death, stroke or systemic embolism/transient ischemic attack, International Society on Thrombosis and Hemostasis (ISTH) major bleeding, new‐onset heart failure, and all‐cause hospitalization. These outcomes were assessed at 1‐year follow‐up.
Statistical Analysis
Descriptive statistics for categorical variables are presented as counts (percentages) and differences between the groups are assessed by Pearson χ2 test. Continuous variables are presented as median (quartile 1–quartile 3) and differences between the groups are assessed by Kruskal‐Wallis test. Patient characteristics were described separately for all patients—those with agreement between eGFR and eCrCl, those with undertreatment, and those with overtreatment. The 95% CI for the percent agreement, undertreated, and overtreated was calculated simultaneously based on the Goodman Method for multinomial proportions. Cumulative incidence curves were generated to estimate the incidence of MACNE by renal function misclassification using unadjusted outcomes data. The cumulative incidence function was used, which accounts for the competing risk of noncardiovascular death, and differences between groups were compared using Gray test.
For 1‐year clinical outcomes, event rates per 100 patient‐years are presented by agreement, undertreatment, and overtreatment categories. Cox proportional hazards models were used to test the association of renal function misclassification and outcomes through 1‐year follow‐up. A robust covariance estimate was included in order to account for variation within sites. All models (including unadjusted models) were stratified by the type of NOAC. Multivariable modeling were used to adjust for clinically relevant patient and site characteristics, which included age, sex, weight, antiplatelet use, history of heart failure, left ventricular function, prior gastrointestinal bleeding, anemia, diabetes, hypertension, history of stroke/transient ischemic attack, coronary artery disease, liver disease, and alcohol abuse. Missing data were handled with multiple imputation, and imputed values were obtained by Markov chain Monte Carlo methods. Combined results using Rubin rules from 5 imputed data sets are presented.
Given the potential for increased misclassification at the lower extreme of renal function, we also performed a sensitivity analysis for rates of misclassification and clinical outcomes among patients with CKD (defined as an eCrCl <60 mL/min). All analyses were performed using SAS software version 9.4 (SAS Institute Inc) and a 2‐tailed P value <0.05 was considered significant for all statistical tests.
Results
Baseline Clinical Characteristics
A total of 8727 patients were included in the analysis. The baseline characteristics of the study cohort and subgroups stratified by misclassification using MDRD eGFR are included in Table 1. Baseline characteristics using CKD‐EPI eGFR are included in Table 2. The median age of the population was 71 years (64–78 years) with a median CHADS2 score of 2 and median HAS‐BLED score of 1. For NOAC use, 48.9% of patients were taking rivaroxaban, 44.6% were taking apixaban, and 6.5% were taking dabigatran. There was concurrent antiplatelet use in 27.6% of patients. The baseline median eCrCl was 82.6 mL/min (59.9–110.5 mL/min), the median eGFR by MDRD was 73.7 mg/dL (59.3–88.4 mg/dL), and the median eGFR by CKD‐EPI was 71.3 mg/dL (56.3–85.7 mg/dL). Patients with undertreatment or overtreatment had statistically significant lower eCrCl and eGFR than patients in the agreement group. The median eCrCl and eGFR by MDRD for patients with agreement in NOAC dosing was 85.3 mL/min and 75.3 mg/dL compared with 61.3 mL/min and 44.9 mg/dL in the undertreatment group and 41.4 mL/min and 56.4 mg/dL in the overtreatment group. Patients had a higher median age in the overtreatment group at 85 years as compared with 71 years for the undertreatment, agreement, and the overall populations. Patients with overtreatment had a statistically lower median weight of 61 kg compared with patients in the agreement group, with a median weight of 90 kg, and the undertreatment group, with a median weight of 104.5 kg.
Table 1.
Baseline Characteristics Sorted by Potential Renal Function Misclassification Using MDRD eGFR
| Overall (N=8727) | Agreement (n=8184) | Undertreatment (n=202) | Overtreatment (n=341) | P value | |
|---|---|---|---|---|---|
| Demographics | |||||
| Age, y | 71.0 (64.0–78.0) | 71.0 (63.0–78.0) | 71.0 (65.0–76.0) | 85.0 (80.0–89.0) | <0.0001 |
| Sex | |||||
| Men | 5117 (58.6) | 4907 (60.0) | 104 (51.5) | 106 (31.1) | <0.0001 |
| Women | 3610 (41.4) | 3277 (40.0) | 98 (48.5) | 235 (68.9) | |
| Race and ethnicity | |||||
| White | 7685 (88.1) | 7214 (88.1) | 184 (91.1) | 287 (84.2) | 0.0001 |
| Black | 377 (4.3) | 361 (4.4) | 5 (2.5) | 11 (3.2) | |
| Hispanic | 372 (4.3) | 345 (4.2) | 8 (4.0) | 19 (5.6) | |
| American Indian/Alaska Native | 14 (0.2) | 14 (0.2) | 0 (0.0) | 0 (0.0) | |
| Asian | 145 (1.7) | 127 (1.6) | 0 (0.0) | 18 (5.3) | |
| Native Hawaiian/Pacific Islander | 3 (0.0) | 3 (0.0) | 0 (0.0) | 0 (0.0) | |
| Other/not reported | 131 (1.5) | 120 (1.5) | 5 (2.5) | 6 (1.8) | |
| Medical history | |||||
| Congestive heart failure | 1865 (21.4) | 1717 (21.0) | 66 (32.7) | 82 (24.0) | 0.0002 |
| Prior cerebrovascular events | 992 (11.4) | 918 (11.2) | 22 (10.9) | 52 (15.2) | 0.0696 |
| Hypertension | 7052 (80.8) | 6594 (80.6) | 182 (90.1) | 276 (80.9) | 0.0031 |
| Hyperlipidemia | 5837 (66.9) | 5463 (66.8) | 147 (72.8) | 227 (66.6) | 0.1978 |
| Diabetes | 2260 (25.9) | 2113 (25.8) | 89 (44.1) | 58 (17.0) | <0.0001 |
| History of CAD | 2379 (27.3) | 2205 (26.9) | 67 (33.2) | 107 (31.4) | 0.0319 |
| Anemia | 729 (8.4) | 650 (7.9) | 33 (16.3) | 46 (13.5) | <0.0001 |
| Chronic kidney disease | 2252 (25.8) | 1826 (22.3) | 202 (100.0) | 224 (65.7) | <0.0001 |
| Cognitive impairment/dementia | 136 (1.6) | 122 (1.5) | 0 (0.0) | 14 (4.1) | 0.0001 |
| Alcohol abuse | 329 (3.8) | 320 (3.9) | 6 (3.0) | 3 (0.9) | 0.0132 |
| Smoking history | 4124 (47.3) | 3895 (47.6) | 92 (45.5) | 137 (40.2) | 0.0239 |
| Gastrointestinal bleed | 357 (4.1) | 324 (4.0) | 10 (5.0) | 23 (6.7) | 0.0323 |
| Stroke risk scores | |||||
| CHADS2 score | 2.0 (1.0–3.0) | 2.0 (1.0–3.0) | 2.0 (1.0–3.0) | 2.0 (2.0–3.0) | <0.0001 |
| CHADS2 score | |||||
| Low: 0 | 829 (9.5) | 813 (9.9) | 10 (5.0) | 6 (1.8) | <0.0001 |
| Medium: 1 | 2636 (30.2) | 2541 (31.0) | 47 (23.3) | 48 (14.1) | |
| High: 2+ | 5262 (60.3) | 4830 (59.0) | 145 (71.8) | 287 (84.2) | |
| Bleeding risk scores. | |||||
| HAS‐BLED score | 1.0 (1.0–2.0) | 1.0 (1.0–2.0) | 2.0 (1.0–2.0) | 2.0 (1.0–2.0) | <0.0001 |
| HAS‐BLED score | |||||
| Low: 0 | 1426 (16.3) | 1401 (17.1) | 24 (11.9) | 1 (0.3) | <0.0001 |
| Medium: 1 or 2 | 6194 (71.0) | 5777 (70.6) | 143 (70.8) | 274 (80.4) | |
| High: 3+ | 1101 (12.6) | 1000 (12.2) | 35 (17.3) | 66 (19.4) | |
| Vital signs and laboratory data | |||||
| Height, cm | 172.0 (163.0–180.0) | 173.0 (163.0–180.0) | 172.0 (163.0–180.0) | 162.0 (155.5–170.0) | <0.0001 |
| Weight, kg | 89.0 (75.0–106.0) | 90.0 (76.0–106.0) | 104.5 (92.0–122.0) | 61.0 (54.0–69.0) | <0.0001 |
| Serum creatinine, mg/dL | 1.0 (0.8–1.2) | 1.0 (0.8–1.1) | 1.5 (1.3–1.7) | 1.1 (0.9–1.4) | <0.0001 |
| Hemoglobin, g/dL | 13.7 (12.5–14.9) | 13.8 (12.5–14.9) | 13.3 (12.0–14.5) | 12.9 (11.6–14.0) | <0.0001 |
| Creatinine clearance, Cockcroft‐Gault, mL/min | 82.6 (59.9–110.5) | 85.3 (63.4–113.2) | 61.3 (53.6–71.3) | 41.4 (30.4–46.3) | <0.0001 |
| eGFR, MDRD, mg/dL | 73.7 (59.3–88.4) | 75.3 (62.2–89.2) | 44.9 (39.1–48.2) | 56.4 (50.7–63.2) | <0.0001 |
| eGFR, CKD‐EPI | 71.3 (56.3–85.7) | 73.1 (58.9–86.7) | 42.5 (37.4–45.4) | 51.7 (45.0–57.5) | <0.0001 |
| Current antithrombotic medications | |||||
| Rivaroxaban | 4270 (48.9) | 3778 (46.2) | 188 (93.1) | 304 (89.1) | <0.0001 |
| Apixaban | 3891 (44.6) | 3846 (47.0) | 13 (6.4) | 32 (9.4) | <0.0001 |
| Dabigatran | 566 (6.5) | 560 (6.8) | 1 (0.5) | 5 (1.5) | <0.0001 |
| Aspirin | 2233 (25.6) | 2110 (25.8) | 56 (27.7) | 67 (19.6) | 0.0308 |
| Clopidogrel | 263 (3.0) | 242 (3.0) | 10 (5.0) | 11 (3.2) | 0.2548 |
| Prasugrel | 11 (0.1) | 11 (0.1) | 0 (0.0) | 0 (0.0) | 0.6940 |
| Ticagrelor | 13 (0.1) | 12 (0.1) | 0 (0.0) | 1 (0.3) | 0.6765 |
| Any antiplatelet | 2408 (27.6) | 2272 (27.8) | 61 (30.2) | 75 (22.0) | 0.0461 |
Values refer to the numeric values/numbers listed for each category. “Other” race refers to patients who do not identify with any of the 5 races listed or those who are multi‐racial (multiple races/mixed), and also includes patients with no reported race.
CAD indicates coronary artery disease; CKD‐EPI, Chronic Kidney Disease Epidemiology Collaboration; eGFR, estimated glomerular filtration rate; INR, international normalized ratio; and MDRD, Modification of Diet in Renal Disease.
Table 2.
Baseline Characteristics Sorted by Potential Renal Function Misclassification Using CKD‐EPI eGFR
| Overall (N=8727) | Agreement (n=8156) | Undertreatment (n=308) | Overtreatment (n=263) | P value | |
|---|---|---|---|---|---|
| Demographics | |||||
| Age, y | 71.0 (64.0–78.0) | 71.0 (63.0–78.0) | 73.0 (67.0–77.0) | 85.0 (80.0–89.0) | <0.0001 |
| Sex | |||||
| Men | 5117 (58.6) | 4885 (59.9) | 175 (56.8) | 57 (21.7) | <0.0001 |
| Women | 3610 (41.4) | 3271 (40.1) | 133 (43.2) | 206 (78.3) | |
| Race and ethnicity | |||||
| White | 7685 (88.1) | 7192 (88.2) | 271 (88.0) | 222 (84.4) | 0.0002 |
| Black | 377 (4.3) | 352 (4.3) | 18 (5.8) | 7 (2.7) | |
| Hispanic | 372 (4.3) | 343 (4.2) | 14 (4.5) | 15 (5.7) | |
| American Indian/Alaska Native | 14 (0.2) | 14 (0.2) | 0 (0.0) | 0 (0.0) | |
| Asian | 145 (1.7) | 130 (1.6) | 0 (0.0) | 15 (5.7) | |
| Native Hawaiian/Pacific Islander | 3 (0.0) | 3 (0.0) | 0 (0.0) | 0 (0.0) | |
| Other/not reported | 131 (1.5) | 122 (1.5) | 5 (1.6) | 4 (1.5) | |
| Medical history | |||||
| Congestive heart failure | 1865 (21.4) | 1694 (20.8) | 102 (33.1) | 69 (26.2) | <0.0001 |
| Prior cerebrovascular events | 992 (11.4) | 921 (11.3) | 33 (10.7) | 38 (14.4) | 0.2653 |
| Hypertension | 7052 (80.8) | 6561 (80.4) | 277 (89.9) | 214 (81.4) | 0.0002 |
| Hyperlipidemia | 5837 (66.9) | 5433 (66.6) | 234 (76.0) | 170 (64.6) | 0.0021 |
| Diabetes | 2260 (25.9) | 2083 (25.5) | 133 (43.2) | 44 (16.7) | <0.0001 |
| History of CAD | 2379 (27.3) | 2192 (26.9) | 110 (35.7) | 77 (29.3) | 0.0022 |
| Anemia | 729 (8.4) | 642 (7.9) | 48 (15.6) | 39 (14.8) | <0.0001 |
| Chronic kidney disease | 2252 (25.8) | 1798 (22.0) | 308 (100.0) | 146 (55.5) | <0.0001 |
| Cognitive impairment/dementia | 136 (1.6) | 126 (1.5) | 0 (0.0) | 10 (3.8) | 0.0012 |
| Alcohol abuse | 329 (3.8) | 315 (3.9) | 11 (3.6) | 3 (1.1) | 0.0730 |
| Smoking history | 4124 (47.3) | 3882 (47.6) | 143 (46.4) | 99 (37.6) | 0.0061 |
| Gastrointestinal bleed | 357 (4.1) | 326 (4.0) | 14 (4.5) | 17 (6.5) | 0.1275 |
| Stroke risk scores | |||||
| CHADS2 score | 2.0 (1.0–3.0) | 2.0 (1.0–3.0) | 2.0 (1.5–3.0) | 2.0 (2.0–3.0) | <0.0001 |
| CHADS2 score | |||||
| Low: 0 | 829 (9.5) | 812 (10.0) | 13 (4.2) | 4 (1.5) | <0.0001 |
| Medium: 1 | 2636 (30.2) | 2540 (31.1) | 64 (20.8) | 32 (12.2) | |
| High: 2+ | 5262 (60.3) | 4804 (58.9) | 231 (75.0) | 227 (86.3) | |
| Bleeding risk scores | |||||
| HAS‐BLED score | 1.0 (1.0–2.0) | 1.0 (1.0–2.0) | 2.0 (1.0–2.0) | 2.0 (1.0–2.0) | <0.0001 |
| HAS‐BLED score | |||||
| Low: 0 | 1426 (16.3) | 1400 (17.2) | 26 (8.4) | 0 (0.0) | <0.0001 |
| Medium: 1–2 | 6194 (71.0) | 5763 (70.7) | 219 (71.1) | 212 (80.6) | |
| High: 3+ | 1101 (12.6) | 987 (12.1) | 63 (20.5) | 51 (19.4) | |
| Vital signs and laboratory data | |||||
| Height, cm | 172.0 (163.0–180.0) | 173.0 (163.0–180.0) | 172.0 (164.0–180.0) | 160.0 (155.0–167.0) | <0.0001 |
| Weight, kg | 89.0 (75.0–106.0) | 90.0 (76.0–106.0) | 101.0 (90.0–116.5) | 60.0 (53.0–66.0) | <0.0001 |
| Serum creatinine, mg/dL | 1.0 (0.8–1.2) | 1.0 (0.8–1.1) | 1.5 (1.3–1.7) | 1.0 (0.9–1.3) | <0.0001 |
| Hemoglobin, g/dL | 13.7 (12.5–14.9) | 13.8 (12.5–14.9) | 13.3 (11.9–14.5) | 12.8 (11.6–13.8) | <0.0001 |
| Creatinine clearance, Cockcroft‐Gault, mL/min | 82.6 (59.9–110.5) | 85.4 (63.2–113.3) | 59.7 (52.8–70.4) | 41.6 (29.8–46.4) | <0.0001 |
| eGFR, MDRD, mg/dL | 73.7 (59.3–88.4) | 75.4 (62.4–89.3) | 46.7 (39.2–50.3) | 58.5 (53.4–65.0) | <0.0001 |
| eGFR, CKD‐EPI | 71.3 (56.3–85.7) | 73.1 (59.1–86.7) | 44.1 (37.5–47.7) | 53.9 (50.2–60.3) | <0.0001 |
| Current antithrombotic medications. | |||||
| Any current antithrombotic medications | 8727 (100.0) | 8156 (100.0) | 308 (100.0) | 263 (100.0) | |
| Rivaroxaban | 4270 (48.9) | 3756 (46.1) | 278 (90.3) | 236 (89.7) | <0.0001 |
| Apixaban | 3891 (44.6) | 3837 (47.0) | 29 (9.4) | 25 (9.5) | <0.0001 |
| Dabigatran | 566 (6.5) | 563 (6.9) | 1 (0.3) | 2 (0.8) | <0.0001 |
| Aspirin | 2233 (25.6) | 2093 (25.7) | 92 (29.9) | 48 (18.3) | 0.0054 |
| Clopidogrel | 263 (3.0) | 239 (2.9) | 15 (4.9) | 9 (3.4) | 0.1370 |
| Prasugrel | 11 (0.1) | 11 (0.1) | 0 (0.0) | 0 (0.0) | 0.6801 |
| Ticagrelor | 13 (0.1) | 12 (0.1) | 0 (0.0) | 1 (0.4) | 0.4949 |
| Any antiplatelet | 2408 (27.6) | 2254 (27.6) | 100 (32.5) | 54 (20.5) | 0.0060 |
Values are expressed as number (percentage) or hazard ratio (95% CI). “Other” race refers to patients who do not identify with any of the 5 races listed or those who are multi‐racial (multiple races/mixed), and also includes patients with no reported race. CAD indicates coronary artery disease; CKD‐EPI, Chronic Kidney Disease Epidemiology Collaboration; eGFR, estimated glomerular filtration rate; INR, international normalized ratio; and MDRD, Modification of Diet in Renal Disease.
Agreement Between Renal Formulae
The results of CKD‐EPI eGFR were similar to MDRD eGFR (Figure 1). Figure 2 shows the agreement between eCrCl and eGFR based on the eligibility and dose adjustment cutoffs for patients receiving rivaroxaban (eCrCl cutoffs of <30 and >50). eCrCl and MDRD eGFR were in agreement for 88.48% of patients. Use of MDRD eGFR resulted in 4.4% cases of undertreatment and 7.12% cases of overtreatment. The agreement between eCrCl and eGFR was higher for both dabigatran and apixaban, where eligibility alone was assessed, with agreement ranging from 98.6% to 98.9% (Figures 3 and 4).For dabigatran and apixaban, there was more overtreatment, as eGFR equations defined more patients as eligible for NOACs compared with eCrCl criteria. Additional analysis using κ statistic for agreement between eCrCl and eGFR showed a κ coefficient range of 0.24 to 0.8 (Tables S1 and S2).
Figure 1. Agreement and potential undertreatment/overtreatment in all patients.
Estimated glomerular filtration rate was calculated with Modification of Diet in Renal Disease (MDRD) and Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) formulae and compared with estimated creatinine clearance.
Figure 2. Agreement between estimated creatinine clearance and Modification of Diet in Renal Disease (MDRD) estimated glomerular filtration rate (eGFR)/Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) eGFR based on the eligibility and dose adjustment cutoffs for patients receiving rivaroxaban.
The boxes are colored green for patients in agreement, blue for potential undertreatment, and orange for overtreatment. Agreement=88.48% (95% CI, 87.5%–89.44%) for MDRD; 87.96% (95% CI, 87.0%–88.94%) for CKD‐EPI; undertreated=4.40% (95% CI, 3.79%–5.02%) for MDRD; 6.51% (95% CI, 5.77%–7.25%) for CKD‐EPI; overtreated=7.12% (95% CI, 6.35%–7.89%) for MDRD; and 5.53% (95% CI, 4.84%–6.21%) for CKD‐EPI. CG indicates Cockcroft‐Gault.
Figure 3. Agreement between estimated creatinine clearance and Modification of Diet in Renal Disease (MDRD) estimated glomerular filtration rate (eGFR)/Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) eGFR based on the eligibility and dose adjustment cutoffs for patients receiving dabigatran.
The boxes are colored green for patients in agreement, blue for potential undertreatment, and orange for overtreatment. Agreement=98.9% (95% CI, 98.1%–99.8%); 98.94% (95% CI, 98.1%–99.78%); undertreated=0.18% (95% CI, 0.00%–0.52%); 0.18% (95% CI, 0.00%–0.52%); overtreated=0.883% (95% CI, 0.113%–1.654%); and 0.88% (95% CI, 0.11%–1.65%). CG indicates Cockcroft‐Gault.
Figure 4. Agreement between estimated creatinine clearance and Modification of Diet in Renal Disease (MDRD) estimated glomerular filtration rate (eGFR)/Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) eGFR based on the eligibility and dose adjustment cutoffs for patients receiving apixaban.
The boxes are colored green for patients in agreement, blue for potential undertreatment, and orange for overtreatment. Agreement=98.84% (95% CI, 98.5%–99.18%); 98.61% (95% CI, 98.2%–98.98%); undertreated=0.33% (95% CI, 0.15%–0.52%); 0.75% (95% CI, 0.48%–1.02%); overtreated=0.82% (95% CI, 0.54%–1.11%); and 0.64% (95% CI, 0.39%–0.89%). CG indicates Cockcroft‐Gault.
Clinical Outcomes
Clinical outcomes at 1 year for patients with agreement versus overtreatment or undertreatment based on MDRD eGFR misclassification are presented in Table 3. The clinical outcomes based on CKD‐EPI eGFR are presented in Table 4. The cumulative incidence of MACNE was higher in the overtreatment and undertreatment groups as compared with the agreement group. This difference in MACNE showed early separation that was maintained over time (Figure 5).
Table 3.
Association Between Potential Renal Function Misclassification Using MDRD and 1‐y Outcomes
| Outcome | Event rate | Unadjusted HR (95% CI) | P value | Global P value | Adjusted HR (95% CI) | P value | Global P value |
|---|---|---|---|---|---|---|---|
| MACNE | <0.0001 | 0.5228 | |||||
| Agreement | 319 (4.20) | Reference | Reference | ||||
| Overtreatment | 27 (8.92) | 2.18 (1.48–3.20) | <0.0001 | 1.02 (0.66–1.60) | 0.9174 | ||
| Undertreatment | 12 (6.63) | 1.63 (0.86–3.09) | 0.1334 | 1.44 (0.75–2.78) | 0.2719 | ||
| All‐cause death | <0.0001 | 0.5992 | |||||
| Agreement | 232 (3.01) | Reference | Reference | ||||
| Overtreatment | 27 (8.62) | 3.25 (2.16–4.90) | <0.0001 | 1.04 (0.61–1.75) | 0.8950 | ||
| Undertreatment | 8 (4.29) | 1.65 (0.80–3.40) | 0.1733 | 1.52 (0.67–3.41) | 0.3158 | ||
| Cardiovascular death | 0.0003 | 0.3392 | |||||
| Agreement | 86 (1.12) | Reference | Reference | ||||
| Overtreatment | 12 (3.86) | 3.77 (1.92–7.38) | 0.0001 | 1.40 (0.60–3.26) | 0.4388 | ||
| Undertreatment | 4 (2.16) | 2.13 (0.78–5.83) | 0.1399 | 2.04 (0.68–6.07) | 0.2013 | ||
| First stroke, non‐CNS embolism, or TIA | <0.0001 | <0.0001 | |||||
| Agreement | 96 (1.25) | Reference | Reference | ||||
| Overtreatment | 5 (1.61) | 1.12 (0.42–3.00) | 0.8271 | 0.52 (0.18–1.53) | 0.2370 | ||
| Undertreatment | 0 (0.00) | … | … | … | … | ||
| First major bleeding | 0.0088 | 0.9280 | |||||
| Agreement | 269 (3.55) | Reference | Reference | ||||
| Overtreatment | 23 (7.61) | 1.99 (1.28–3.08) | 0.0021 | 1.05 (0.63–1.75) | 0.8416 | ||
| Undertreatment | 8 (4.39) | 1.14 (0.56–2.31) | 0.7105 | 0.88 (0.43–1.81) | 0.7333 | ||
| New‐onset HF | 0.0029 | 0.1204 | |||||
| Agreement | 111 (1.83) | Reference | Reference | ||||
| Overtreatment | 7 (2.95) | 1.67 (0.80–3.47) | 0.1687 | 0.76 (0.34–1.71) | 0.5017 | ||
| Undertreatment | 7 (5.77) | 3.26 (1.50–7.09) | 0.0028 | 2.32 (1.03–5.25) | 0.0422 | ||
| First all‐cause hospitalization | 0.2009 | 0.8820 | |||||
| Agreement | 2648 (43.16) | Reference | Reference | ||||
| Overtreatment | 123 (50.58) | 1.19 (0.98–1.44) | 0.0774 | 0.99 (0.81–1.22) | 0.9572 | ||
| Undertreatment | 66 (44.81) | 1.07 (0.84–1.37) | 0.5951 | 0.94 (0.73–1.20) | 0.6178 |
Note: Since there were no patients in the undertreatment group who experienced stroke/transient ischemic attack (TIA), no hazard ratios (HRs) or CIs could be estimated.
The event rate presented is the number of events and number of events per 100 patient‐years in parentheses. The global P value is a test for whether the 3 categories, as a whole, are different than would be expected by chance. CNS indicates central nervous system; MACNE, major adverse cardiovascular and neurological events; and MDRD, Modification of Diet in Renal Disease.
Table 4.
Association Between Potential Renal Function Misclassification Using CKD‐EPI and 1‐y Outcomes
| Outcome | Event rate | Unadjusted HR (95% CI) | P value | Global P value | Adjusted HR (95% CI) | P value | Global P value |
|---|---|---|---|---|---|---|---|
| MACNE | 0.0007 | 0.8733 | |||||
| Agreement | 321 (4.25) | Reference | Reference | ||||
| Overtreatment | 21 (9.01) | 2.16 (1.39–3.37) | 0.0007 | 0.99 (0.62–1.58) | 0.9501 | ||
| Undertreatment | 16 (5.74) | 1.38 (0.82–2.34) | 0.2259 | 1.15 (0.67–1.98) | 0.6030 | ||
| All‐cause death | <0.0001 | 0.3772 | |||||
| Agreement | 231 (3.00) | Reference | Reference | ||||
| Overtreatment | 22 (9.14) | 3.47 (2.24–5.37) | <0.0001 | 1.05 (0.60–1.84) | 0.8684 | ||
| Undertreatment | 14 (4.89) | 1.87 (1.08–3.24) | 0.0261 | 1.56 (0.83–2.92) | 0.1645 | ||
| Cardiovascular death | 0.0004 | 0.6150 | |||||
| Agreement | 87 (1.13) | Reference | Reference | ||||
| Overtreatment | 10 (4.17) | 3.98 (1.99–7.96) | <0.0001 | 1.39 (0.60–3.22) | 0.4452 | ||
| Undertreatment | 5 (1.75) | 1.68 (0.67–4.21) | 0.2674 | 1.44 (0.54–3.86) | 0.4676 | ||
| First stroke, non‐CNS embolism, or TIA | 0.8311 | 0.3738 | |||||
| Agreement | 94 (1.23) | Reference | Reference | ||||
| Overtreatment | 4 (1.67) | 1.19 (0.48–2.93) | 0.7107 | 0.55 (0.21–1.41) | 0.2114 | ||
| Undertreatment | 3 (1.06) | 0.75 (0.23–2.43) | 0.6326 | 0.68 (0.21–2.20) | 0.5178 | ||
| First major bleeding | 0.0141 | 0.6363 | |||||
| Agreement | 271 (3.59) | Reference | Reference | ||||
| Overtreatment | 18 (7.74) | 1.98 (1.25–3.14) | 0.0037 | 0.98 (0.59–1.64) | 0.9461 | ||
| Undertreatment | 11 (3.91) | 1.01 (0.55–1.85) | 0.9832 | 0.74 (0.40–1.38) | 0.3424 | ||
| New‐onset HF | 0.0105 | 0.6999 | |||||
| Agreement | 111 (1.84) | Reference | Reference | ||||
| Overtreatment | 7 (4.00) | 2.25 (1.06–4.78) | 0.0358 | 1.05 (0.45–2.46) | 0.9015 | ||
| Undertreatment | 7 (3.74) | 2.11 (0.98–4.55) | 0.0568 | 1.38 (0.63–3.05) | 0.4203 | ||
| First all‐cause hospitalization | 0.2392 | 0.8187 | |||||
| Agreement | 2638 (43.16) | Reference | Reference | ||||
| Overtreatment | 93 (49.65) | 1.17 (0.96–1.43) | 0.1305 | 0.94 (0.76–1.16) | 0.5396 | ||
| Undertreatment | 106 (46.79) | 1.11 (0.89–1.39) | 0.3442 | 0.98 (0.79–1.22) | 0.8764 |
The event rate presented is the number of events and number of events per 100 patient‐years in parentheses. The global value is a test for whether the 3 categories, as a whole, are different than would be expected by chance. CKD‐EPI indicates Chronic Kidney Disease Epidemiology Collaboration; CNS, central nervous system; HF, heart failure; HR, hazard ratio; MACNE, major adverse cardiovascular and neurological events; MDRD, Modification of Diet in Renal Disease; and TIA, transient ischemic attack.
Figure 5. Cumulative incidence curves of major adverse cardiovascular and neurological events (MACNE) stratified by type of renal function assessment used (Modification of Diet in Renal Disease [MDRD] or Chronic Kidney Disease Epidemiology Collaboration [CKD‐EPI]) in the overall population (A and B) and in the chronic kidney disease (CKD) population (C and D).
Sensitivity Analysis for Patients With CKD
A total of 2184 patients were included in the subgroup analysis of patients with eCrCl <60. Baseline clinical characteristics are outlined in Table S3 and S4. The median age of patients was 81 years (76–85 years) with a median CHADS2 score of 2 and median HAS‐BLED score of 2. The median eCrCl was 47.2 mL/min, median MDRD eGFR was 54.3 mg/dL, and median CKD‐EPI eGFR was 49.7 mg/dL. In this group, 52.7% of patients were taking apixaban, 42.4% were taking rivaroxaban, and 4.8% were taking dabigatran.
In the CKD population, the agreement between eGFR and eCrCl for rivaroxaban was 58.1% to 60.6% (Figure S1). With the MDRD eGFR, 32.8% of patients would be overtreated while 9.1% would be undertreated. A similar trend with more overtreatment was seen with the CKD‐EPI eGFR. The agreement of eGFR and eCrCl for dabigatran and apixaban ranged from 94.3% to 97.1%. For dabigatran and apixaban, there was more potential overtreatment with more patients eligible for NOACs based on eGFR as compared with eCrCl criteria (Figures S2 and S3).
Clinical outcomes at 1 year for patients with agreement versus overtreatment or undertreatment based on MDRD and CKD‐EPI eGFR misclassification are presented in Table S5 and S6.The NOAC undertreatment group for patients with CKD using MDRD eGFR had a statistically significant increase in composite major adverse cardiac events (adjusted hazard ratio, 2.93 [95% CI, 1.08–7.92], P=0.03) compared with the group with appropriate dosing (Table S6). There was no statistical significance in secondary outcomes of cardiovascular death, all‐cause death, or major bleeding between the groups. An increase in cumulative incidence of MACNE was again seen in the CKD group but was not statistically significant (Figures 5 and 6).
Figure 6. Agreement and potential undertreatment/overtreatment in patients with chronic kidney disease.
Estimated glomerular filtration rate was calculated with Modification of Diet in Renal Disease (MDRD) and Chronic Kidney Disease Epidemiology Collaboration (CKD‐EPI) formulae and compared with estimated creatinine clearance (eCrCl). CKD indicates chronic kidney disease.
Discussion
In this large, contemporary observational study, we found up to a 12% discrepancy for NOAC dose adjustment using eCrCl versus eGFR with either equation (CKD‐EPI or MDRD) in all patients with AF. This difference is further exaggerated in patients with AF who have renal dysfunction, with up to a 42% discrepancy in our subgroup analysis. This finding extends previously published retrospective study findings in the patient population of CKD with up to a 36% misclassification rate, but our analysis demonstrates that this discrepancy exists even in the overall general AF population. 4 , 11 , 14 Interestingly, contrary to a recent study in Taiwanese patients with AF, which found that almost all of the discordance in NOAC dosing resulted in overtreatment, our study found almost equal numbers of patients with potential overtreatment and undertreatment in the overall AF population. This is likely attributable to the broader inclusion criteria in our study including patients with and without CKD as well as a North American population, who have a higher median body weight (89 kg in our study versus 63 kg in the Asian study).
Overall, clinical outcomes were generally similar between patients who received correct and potential inappropriate NOAC dose adjustment. However, in the subgroup of patients with AF who had renal dysfunction, defined as an eCrCl <60 mL/min, we observed a statistically significant increase in MACNE for patients who had potential undertreatment compared with correct classification using the MDRD formula. Furthermore, the cumulative incidence curves showed clear separation between the overtreatment and undertreatment groups compared with the group with appropriate dosing that was maintained over the 1‐year follow‐up period.
To our knowledge, this is the first registry study demonstrating the influence of kidney function estimations on NOAC dose adjustment and subsequent clinical outcomes in patients with nonvalvular AF. The use of different renal formulae for dose adjustment may in part explain findings from 3 prior large retrospective studies demonstrating worse clinical outcomes for patients with inappropriate NOAC dose adjustment. 4 , 12 Similar observations have been made in the acute coronary syndrome population where estimation of renal function using different formulae resulted in different medication dose adjustments in 20% of the patients. 19 Furthermore, different formulae for assessment of renal function had different prognostic value for assessing mortality in the postacute coronary syndrome and heart failure populations in large registry studies. 20 , 21
This study has several important clinical implications. First, to our knowledge, this is the first large dedicated AF registry study that investigated the use of eGFR as a potential mechanism for inappropriate NOAC dosing and its impact on clinical outcomes. This mechanism for inadvertent NOAC dosing differs from traditional reasons identified from previous studies such as the perception of higher bleeding risk, higher stroke risk, and frailty. 6 , 7 While eCrCl is the gold standard estimation studied in clinical trials and regulatory bodies, eGFR is often used to estimate renal function becauese of its accessibility in clinical practice as it is automatically reported by a majority of laboratories. 8 This mechanism for NOAC misdosing may be inadvertent in nature and underrecognized. It is also easily avoided through education around this discrepancy and by using eCrCl as the gold standard calculation for renal dose adjustment.
Second, we found that renal dysfunction, older age, and lower body weight were associated with potential overtreatment or undertreatment of NOACs. This difference is likely explained by the fact that the eCrCl formula takes into consideration patient weight while eGFR formulae do not. Elderly patients are more likely to have lower body weight and renal dysfunction, putting them at higher risk for overtreatment. 6 , 22 , 23 Previous studies have shown that eGFR overestimates eCrCl in the elderly population, which would lead to overtreatment and potential harm. 22 , 24 Recognizing the differences between eGFR and eCrCl has important implications for ensuring proper NOAC dosing, especially in these higher‐risk groups.
Third, the use of eGFR results in frequent dosing misclassification and potentially worse clinical outcomes. As discussed above, there is a statistically significant increase in overall MACNE in the CKD subgroup of patients who were undertreated in accordance with the eGFR equations. Previous studies using administrative data have shown worse outcomes with patients with potential undertreatment and overtreatment. 4 , 12 However, unlike the current study, this prior work using administrative data calculated renal function based solely on eGFR because eCrCl was not available, which could jeopardize external validity.
Finally, the current analysis implores education around the difference between eCrCl and eGFR, especially among patients with CKD, to improve NOAC dosing and patient safety. With the widespread use of electronic medical records, calculation of eCrCl may be more easily implemented if patient weight data can be linked and updated regularly. Recent randomized trials have demonstrated the utility of electronic medical alerts in helping with clinical decision‐making. 25 Alerting clinicians to changes in eCrCl calculations through an automated electronic medical record system may improve clinical decision‐making and help facilitate more accurate NOAC dosing.
Limitations
This study has some limitations. First, the use of different eCrCl cutoffs in clinical trials impacts the comparability of potential misclassification between the NOACs. In clinical practice, apixaban dose adjustment is based on frailty criteria including serum creatinine (not clearance or eGFR), body weight, and age. In contrast, eCrCl cutoffs are used for both dose adjustment and eligibility with rivaroxaban (eCrCl of <30 mL/min and >50 mL/min) and likely account for the higher misclassification rate as compared with apixaban and dabigatran, where eCrCl is used solely for dosing eligibility. These clinical considerations were not assessed in the current study. Second, although we found that potential undertreatment was associated with a significantly greater risk of MACNE among patients with MDRD calculated, this study is generally underpowered to evaluate for clinical outcomes given the low overall clinical event rate and the wide CIs. Therefore, efficacy and safety outcomes related to potential misclassification attributable to renal formulae calculation are likely underestimated. There is also potential for model overfitting given the number of adjustment covariates compared with the number of events for some of the outcomes. In addition, the follow‐up period for assessment of clinical outcomes was 1 year, which further limited the number of clinical events. Third, a change in NOAC or dosing may have influenced the results during the study follow‐up, but this information was not collected in the ORBIT‐AF II registry. Finally, it is important to note that the outcomes assessed are based on potential misclassifications from using different renal formulae. The actual reasons for inappropriate NOAC dosing may be multifactorial and were not assessed given the retrospective nature of this study.
Conclusions
Using eGFR instead of eCrCl for estimation of renal function has the potential to inadvertently lead to inappropriate NOAC dose adjustment. Inappropriate NOAC dosing may lead to worse clinical outcomes, particularly in the CKD population. These findings emphasize the importance of using eCrCl for calculation of dose adjustment in all patients with AF taking NOACs.
Sources of Funding
The ORBIT‐AF II registry is sponsored by Janssen Scientific Affairs, LLC, Raritan, NJ. The current analysis was supported by an investigator‐initiated study grant to Dr Fordyce from Bayer Canada Inc, which had no role in the design, analysis, or interpretation of the data.
Disclosures
None.
Supporting information
Tables S1–S6
Figures S1–S3
Acknowledgments
We acknowledge Rosalia Blanco for her support and project management at the Duke Clinical Research Institute.
Supplemental Material is available at https://www.ahajournals.org/doi/suppl/10.1161/JAHA.122.026605
For Sources of Funding and Disclosures, see page 13.
References
- 1. Andrade J, Aguilar M, Atzema C, Bell A, Cairns J, Cheung C, Cox J, Dorian P, Gladstone D, Healey J, et al. The 2020 Canadian Cardiovascular Society/Canadian Heart Rhythm Society comprehensive guidelines for the management of atrial fibrillation. Can J Cardiol. 2020;36:1847–1948. doi: 10.1016/j.cjca.2020.09.001 [DOI] [PubMed] [Google Scholar]
- 2. January C, Wann L, Calkins H, Chen L, Cigarroa J, Cleveland J, Ellinor P, Ezekowitz M, Field M, Furie K, 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. Circulation. 2019;140:e125–e151. doi: 10.1161/CIR.0000000000000665 [DOI] [PubMed] [Google Scholar]
- 3. Ruff C, Giugliano R, Braunwald E, Hoffman E, Deenadayalu N, Ezekowitz M, Camm A, Weitz J, Lewis B, Parkhomenko A, et al. Comparison of the efficacy and safety of new oral anticoagulants with warfarin in patients with atrial fibrillation: a meta‐analysis of randomised trials. Lancet. 2014;383:955–962. doi: 10.1016/S0140-6736(13)62343-0 [DOI] [PubMed] [Google Scholar]
- 4. Steinberg B, Shrader P, Thomas L, Ansell J, Gonarow G, Gersh B, Kowey P, Mahaffey K, Naccarelli G, Reiffel J, et al. Off‐label dosing of non‐vitamin K antagonist oral anticoagulants and adverse outcomes: the ORBIT‐AF II registry. J Am Coll Cardiol. 2016;68:2597–2604. doi: 10.1016/j.jacc.2016.09.966 [DOI] [PubMed] [Google Scholar]
- 5. Steinberg B, Blanco R, Ollis D, Kim S, Holmes D, Kowey P, Fonarow G, Ansell J, Gersh B, Go A, et al. Outcomes registry for better informed treatment of atrial fibrillation II: rationale and design of the ORBIT‐AF II registry. Am Heart J. 2014;168:160–167. doi: 10.1016/j.ahj.2014.04.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6. Saczynski J, Sanghai S, Kiefe C, Lessard D, Marino F, Waring M, Parish D, Helm R, Sogade F, Goldberg R, et al. Geriatric elements and oral anticoagulant prescribing in older atrial fibrillation patients: SAGE‐AF. J Am Geriatr Soc. 2020;68:147–154. doi: 10.1111/jgs.16178 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7. Sanghai S, Wong C, Wang Z, Clive P, Tran W, Waring M, Goldberg R, Hayward R, Saczynski J, McManus D. Rates of potentially inappropriate dosing of direct‐acting oral anticoagulants and associations with geriatric conditions among older patients with atrial fibrillation: the SAGE‐AF study. J Am Heart Assoc. 2020;9:e014108. doi: 10.1161/JAHA.119.014108 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8. Smith D, Schneider J, Thorp M, Vupputuri S, Weiss J, Johnson E, Feldstein A, Petrik A, Yang X, Snyder S. Clinician's use of automated reports of estimated glomerular filtration rate: a qualitative study. BMC Nephrol. 2012;13:154. doi: 10.1186/1471-2369-13-154 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Michels W, Grootendorst D, Verduijn M, Elliott E, Dekker F, Krediet R. Performance of the Cockcroft‐Gault, MDRD, and new CKD‐EPI formulas in relation to GFR, age, and body size. Clin J Am Soc Nephrol. 2010;5:1003–1009. doi: 10.2215/CJN.06870909 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Granger C, Alexander J, McMurray J, Lopes R, Hylek E, Hanna M, Al‐Khalidi H, Ansell J, Atar D, Avezum A, et al. Apixaban versus warfarin in patients with atrial fibrillation. N Engl J Med. 2011;365:981–992. doi: 10.1056/NEJMoa1107039 [DOI] [PubMed] [Google Scholar]
- 11. Chan Y, Chao T, Lee H, Yeh Y, Chang S, Kuo C, Lip G, Chen S. Different renal function equations and dosing of direct oral anticoagulants in atrial fibrillation. JACC Asia. 2022;2:46–58. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12. Yao X, Shah N, Sangaralingham L, Gersh B, Noseworthy P. Non‐vitamin K antagonist oral anticoagulant dosing in patients with atrial fibrillation and renal dysfunction. J Am Coll Cardiol. 2017;69:2779–2790. doi: 10.1016/j.jacc.2017.03.600 [DOI] [PubMed] [Google Scholar]
- 13. Hawkins N, Er L, Sandhu R, Kaul P, McAlister F, Levin A, Andrade J. Validity of different dose reduction criteria for apixaban. Am Heart J. 2021;238:12–15. doi: 10.1016/j.ahj.2021.03.004 [DOI] [PubMed] [Google Scholar]
- 14. Andrade J, Hawkins N, Fordyce C, Deyell M, Er L, Djurdjev O, Macle L, Virani S, Levin A. Variability in non–vitamin K antagonist oral anticoagulants dose adjustment in atrial fibrillation patients with renal dysfunction: the influence of renal function estimation formulae. Can J Cardiol. 2018;34:1010–1018. doi: 10.1016/j.cjca.2018.04.019 [DOI] [PubMed] [Google Scholar]
- 15. Levey A, Coresh J, Balk E, Kausz A, Levin A, Steffes M, Hogg R, Perrone R, Lau J, Eknoyan G, et al. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification and stratification. Ann Intern Med. 2003;139:137–147. doi: 10.7326/0003-4819-139-2-200307150-00013 [DOI] [PubMed] [Google Scholar]
- 16. Levey A, Stevens L, Schmid C, Zhang Y, Castro A III, Feldman H, Kusek J, Eggers P, Van Lente F, Green T, et al. A new equation to estimate glomerular filtration rate. Ann Intern Med. 2009;150:604–612. doi: 10.7326/0003-4819-150-9-200905050-00006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17. Patel M, Mahaffey K, Garg J, Pan G, Singer D, Hacke W, Breithardt G, Haperin J, Hankey G, Piccini J, et al. Rivaroxaban versus warfarin in nonvalvular atrial fibrillation. N Engl J Med. 2011;365:883–891. doi: 10.1056/NEJMoa1009638 [DOI] [PubMed] [Google Scholar]
- 18. Connolly S, Ezekowitz M, Yusuf S, Eikelboom J, Oldgren J, Parekh A, Pogue J, Reilly P, Themeles E, Varrone J, et al. Dabigatran versus warfarin in patients with atrial fibrillation. N Engl J Med. 2009;361:1139–1151. doi: 10.1056/NEJMoa0905561 [DOI] [PubMed] [Google Scholar]
- 19. Melloni C, Peterson E, Chen A, Szczech L, Newby L, Harrington R, Gibler W, Ohman E, Spinler S, Roe M, et al. Cockcroft‐Gault versus modification of diet in renal disease: importance of glomerular filtration rate formula for classification of chronic kidney disease in patients with non‐ST‐segment elevation acute coronary syndromes. J Am Coll Cardiol. 2008;51:991–996. doi: 10.1016/j.jacc.2007.11.045 [DOI] [PubMed] [Google Scholar]
- 20. Szummer K, Evans M, Carrero J, Alehagen U, Dahlstrom U, Benson L, Lund L. Comparison of the chronic kidney disease epidemiology collaboration, the modification of diet in renal disease study and the Cockcroft‐Gault equation in patients with heart failure. Open Heart. 2017;4:e000568. doi: 10.1136/openhrt-2016-000568 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21. Szummer K, Lundman P, Jacobson S, Lindback J, Stenestrand U, Wallentin L, Jernberg T. Cockcroft‐Gault is better than the modification of diet in renal disease study formula at predicting outcome after a myocardial infarction: data from the Swedish Web‐system for Enhancement and Development of Evidence‐based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART). Am Heart J. 2010;159:979–986. doi: 10.1016/j.ahj.2010.03.028 [DOI] [PubMed] [Google Scholar]
- 22. Dowling T, Wang E, Ferrucci L, Sorkin J. Glomerular filtration rate equations overestimate creatinine clearance in older individuals enrolled in the Baltimore Longitudinal Study on Aging: impact on renal drug dosing. Pharmacotherapy. 2013;33:912–921. doi: 10.1002/phar.1282 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Manzano‐Fernandez S, Andreu‐Cayuelas J, Marin F, Orenes‐Pinero E, Gallego P, Valdes M, Vicente V, Lip G, Roldan V. Comparison of estimated glomerular filtration rate equations for dosing new oral anticoagulants in patients with atrial fibrillation. Rev Esp Cardiol (Engl Ed). 2015;68:497–504. doi: 10.1016/j.rec.2014.06.026 [DOI] [PubMed] [Google Scholar]
- 24. Lee K, Choi J, Kim Y, Boo K, Kim D, Choi Y, Choi H, Kim D, Lee D, Roh S, et al. Comparison of renal function estimation formulae for dosing direct oral anticoagulants in patients with atrial fibrillation. J Clin Med. 2019;8:2034. doi: 10.3390/jcm8122034 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Awdishu L, Coates C, Lyddane A, Tran K, Daniels C, Lee J, El‐Kareh R. The impact of real‐time alerting on appropriate prescribing in kidney disease: a cluster randomized controlled trial. J Am Med Inform Assoc. 2016;23:609–616. doi: 10.1093/jamia/ocv159 [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Tables S1–S6
Figures S1–S3