ORIGINAL PAPER

Renal function, attributes and coagulation treatment in atrial fibrillation (R-FACT Study): Retrospective, observational, longitudinal cohort study of renal function and antithrombotic treatment patterns in atrial fibrillation patients with documented eGFR in real-world clinical practices in Germany D. Wu,1 G. Mansoor,2 C. Kempf,3 M.-S. Schwalm,4 J. Chin4

1

Global Health Outcomes, Merck & Co., Inc., Whitehouse Station, NJ, USA 2 Clinical Research, Merck & Co., Inc., Whitehouse Station, NJ, USA 3 Cegedim Strategic Data, Paris, France 4 Cegedim Strategic Data, Jersey City, NJ, USA Correspondence to: David Wu, PhD, Global Health Outcomes, Merck & Co., Inc., One Merck Drive, WS2E47, Whitehouse Station, NJ 08889, USA Tel.: + 1 908 423 3755 Fax: + 1 908 735 1688 Email: [email protected] Disclosures David Wu and George Mansoor are employees of Merck & Co., Inc., and may own stock or hold stock options in the company. Christian Kempf, Marie-Sophie Schwalm and Jiyoung Chin are employees of Cegedim Strategic Data, which received payment from Merck & Co., Inc., for conducting the statistical analysis and for writing and reviewing the manuscript.

SUMMARY

Introduction Atrial fibrillation (AF) is among the most common arrhythmias diagnosed in routine medical practice (1,2), and independently increases the risk of ischemic stroke by 4–5%, depending on a patient’s demographic and clinical characteristics (3,4). AF is common in patients with end-stage renal disease and impaired kidney function is associated with increased risk for several cardiovascular complications including death, myocardial infarction (MI), stroke and

714

What’s known

Aims: This retrospective, observational, longitudinal study aimed to document the distribution, changes in renal function [measured by estimated glomerular filtration (eGFR)] and antithrombotic treatment pattern in atrial fibrillation (AF) patients in real-world settings managed by general practitioners in Germany. Methods and results: Data were extracted from the German Longitudinal Patient Database. A total of 15,900 patients with AF were identified. Among 1660 having eGFR available at baseline, 3.4% had severely impaired eGFR, 9.7% and 25.6% had moderate severe decrease and moderate decrease in eGFR, respectively, and 61.3% had mildly decreased/normal eGFR. Patients with moderately and severely decreased eGFR tended to be older. The proportion of patients with a CHADS2 score ≥ 2 was 92.9% in those with severely decreased eGFR, and 87.0% and 79.1% in those with moderately severe and moderately decreased eGFR. During follow up, 52.1% of patients with severely decreased eGFR, and 26.3% to 23.7% of patients with moderately decreased eGFR were not treated by antithrombotic. When comparing baseline with follow-up eGFR, 55.0% of patients showed decreased eGFR. Age, diabetes, dyslipidaemia and history of myocardial infarction were identified as significant predictors for renal function deterioration based on results from multivariate Cox regression model. Conclusions: Moderate-to-severe renal dysfunction is prevalent (~38%) in German AF patients with documented eGFR managed in actual clinical practices. The risk of stroke, as measured by the CHADS2 score, was associated with decreased renal function. Treatment with anticoagulation therapies decreased with decreasing renal function, despite increasing risk of stroke. Anticoagulation treatments remain suboptimal during the 12-month follow up in patients with moderate or severe renal impairment.

Atrial fibrillation is a frequently diagnosed arrhythmia and is common in patients with end-stage renal disease. Moreover, a progressively lower level of estimated glomerular filtration rate is associated with a graded, increased risk of stroke, and chronic kidney disease independently increases the risk of thromboembolism in atrial fibrillation.

What’s new Renal dysfunction is a prevalent condition in German atrial fibrillation patients managed by general practitioners. Decreasing renal function over time was prevalent in these patients, resulting in more patients with moderately severe or severe renal impairment during the 12-month follow up. Treatment with antithrombotic therapy decreased as renal dysfunction increased, despite increasing risk of stroke. In addition, age, diabetes, dyslipidaemia and MI were independent significant contributors to renal deterioration.

congestive heart failure (5–10). However, few data are available on the prevalence, progression and antithrombotic treatment patterns among adults with AF and longitudinally documented renal function in real-world clinical practice (5,7). Previous research indicated that a progressively lower level of estimated glomerular filtration rate (eGFR) is associated with a graded, increased risk of stroke, and a recent metaanalysis of 21 randomised clinical trials and observational studies concluded that eGFR < 60 ml/min/ 1.73 m² is independently related to the incidence of ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724. doi: 10.1111/ijcp.12379

Renal function, attributes and coagulation treatment in atrial fibrillation

stroke (8–10). Furthermore, previous results also showed that chronic kidney disease increased the risk of thromboembolism in AF independently of other risk factors (8,9). The objectives of this retrospective, observational, longitudinal cohort study conducted in Germany were to estimate the distribution of eGFR in patients with AF and longitudinally documented eGFR in a real-world clinical practice setting; examine/compare the demographic and clinical characteristics according to renal function status; investigate the pharmacological treatment patterns for stroke prevention in these patients according to eGFR stages and investigate the change in eGFR over the follow up and identify potential risk factors for the deterioration of renal function.

Methods Design This was a retrospective longitudinal cohort study consisting of three distinct periods: a 36-month baseline period beginning 1 January 2006, a 36-month enrolment period beginning 1 January 2007 and a 35-month follow-up period beginning 1 January 2007. For each patient, the diagnosis date of AF during the enrolment period was defined as the index date with a minimum baseline period of 12 months before index date. Each patient was followed from their index date until the study end date: 30 November 2010 (Figure 1).

Data sources Data were extracted between 1 January 2006 and 30 November 2010 from the German Longitudinal Patient Database (GLPD), a computerised panel of 550 volunteer general practitioners (GPs) who contribute exhaustive anonymous data on patients’ consultations, medical exams, diagnosis as well as prescriptions to a centralised database. The panel was selected to be nationally representative according to their geographical area, age and gender. The database included routinely collected electronic health records

Figure 1 Schematic diagram of R-FACT study ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

for more than 600,000 unique patients per year collected by the GP at the time of consultation. Analyses performed on the GLPD were approved by the ‘Landesdatenschutzbeauftragter des Landes Hessen’. Information on disease status and medication prescription was encoded using the 10th revision of the International Classification of Diseases (ICD-10) classification system (11). Prescription data contained the dispensed drug name (commercial and international common denomination), the Anatomical Therapeutic Chemical (ATC) classification category, dose regimens and prescription duration and concomitant medication.

Study population Patients were selected if they were ≥ 18 years old at index date; had a diagnosis of AF (ICD-10 code: I-48) during the enrolment period; had at least one visit per year to the GP during the follow-up period and had a minimum of 1 year of presence in the GLPD prior to index date. Patients diagnosed with mitral stenosis (ICD-10: I05.0, I05.2, I34.2 and Q23.2), vascular repair or replacement (ICD-10: Z95.5, Z85.8 and Z95.9) and concomitant hyperthyroidism (ICD-10: E05.x) were not included (Figure 2). Data extracted at baseline included age, gender, smoking status and medical history [presence of AF, hypertension, coronary heart disease, heart failure, left ventricular dysfunction, hypertension, stroke, transient ischemic attack (TIA), peripheral artery disease, aortic plaque, type II diabetes mellitus (T2DM), dyslipidaemia and thromboembolism]. The following laboratory or clinical measurements were extracted taking into account at baseline the closest measurement before or within 90 days after the index date, and at follow up the closest result before the end of study date: eGFR as calculated by the local laboratory estimated from serum creatinine, and entered directly by the GP, weight, height, HbA1c (glycosylated haemoglobin), systolic and diastolic blood pressure (SBP and DBP), low-density lipoprotein cholesterol (LDL-C), high-density lipoprotein cholesterol (HDL-C), triglyceride values and

715

716

Renal function, attributes and coagulation treatment in atrial fibrillation

Patients with diagnosis of AF during the study enrolment period n = 41,857

Patients aged 18 years and older at index date n = 41,612

Patients with at least one GP visit during the 12 month prior to index date (whatever the reason of the visit to GP) n = 36,405

At least one visit to the GP per year during the follow-up period from index date (whatever the reason of the visit to GP) n = 23,192 Patients with an exclusion criterion: • Diagnosis of mitral stenosis • Vascular repair or replacement • Concomitant hyperthyroidism n = 3648 Eligible patients (previous visit + follow-up visit) n = 15,900

eGFR value at baseline: n = 1660

eGFR value at follow-up: n = 4535

eGFR value at baseline + follow-up: n = 1374

Figure 2 Flow chart of patients’ enrolment

international normalised ratio (INR), including the frequency of INR test and the time in the (INR 2–3) therapeutic range (TTR) calculated according to Rosendaal (12). Prescription data on preventive treatments for stroke [VKA (ATC: B01AA), antiplatelets (AP, ATC: B01AC), direct thrombin inhibitors (ATC: B01AE), other antithrombotic agents (ATC: B01AX)] and non-steroidal anti-inflammatory drugs (NSAIDs, ATC: M01A) were extracted for the whole study period.

Assessment of renal function Renal function was assessed through eGFR value. Patients with eGFR values available were stratified by renal function stages according to the National Kidney Foundation Practice Guidelines for Chronic Kidney Disease (13), shown in Table 1. eGFR values were calculated directly by the local laboratory using the modification of the diet in renal disease formula (14)

Table 1 Classification of renal dysfunction according to the National Kidney Foundation Practice Guidelines for chronic kidney disease (13)

Stage of eGFR value Renal (ml/min/1.73 m²) Dysfunction Description

≥ 90

I

60–89

II

45–59

IIIA

30–44

IIIB

15–29

IV

< 15 (or dialysis)

V

Normal or increased renal function Mildly decreased renal function Moderately decreased renal function Moderately severe decreased renal function Severely decreased renal function Kidney failure

ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

Renal function, attributes and coagulation treatment in atrial fibrillation

and entered in the GLPD by the GP. Baseline eGFR was defined based on the nearest measurement before the index date or up to 90 days post–index date if no prior measurement was available; the follow-up eGFR was defined based on the nearest measurement before the study end date.

Statistical methods Demographic and other baseline data were summarised overall and by different classes of eGFR (i.e. eGFR < 30, 30–44, 45–59 and ≥ 60 ml/min/1.73 m²) at baseline in terms of mean, standard deviation, median, range for continuous variables and absolute and relative frequencies for categorical variables. Two-sided 95% confidence intervals (95% CI) were determined using normal approximation. Results were compared between eGFR classes by means of one-way fixed-effect analyses of variance for continuous variables and v2 tests for categorical variables. Univariate Cox regression analysis was used to test the relationship between deterioration of renal function and potential risk factors, including demographic and clinical characteristics (Table 2). The deterioration of renal function was defined as the decrease in eGFR from baseline to follow up (yes/ no). The final multivariate Cox model was carried out taking into account all significant variables with a p-value ≤ 0.15 identified by the univariate Cox regression analysis. All statistical calculations were performed on the entire study population and per eGFR class, using SAS software, version 9.1 (SAS Institute, Cary, NC).

Results Baseline A total of 15,900 patients with AF were identified in the database; 1660 of them had an eGFR value at baseline and 1374 had at least one eGFR value both at baseline and during follow up. Table 2 shows clinical and demographic data for both populations. Among patients with baseline eGFR values, 56 patients (3.4%; 95% CI [2.5%; 4.2%]) had stage IV or V (i.e. severely decreased renal function or renal failure) 161 (9.7%; 95% CI [8.3%; 11.1%]) had stage IIIB (i.e. moderately severe decreased renal function), 425 (25.6%; 95% CI [23.5%; 27.7%]) had stage IIIA (i.e. moderately decreased renal function) and 1018 (61.3%; 95% CI [59.0%; 63.7%]) had stage II or I (i.e. mildly decreased or normal renal function). The total proportion of moderately or severely decreased renal function among AF patients with documented eGFR at baseline was 38.7%. The proportion of male and female patients with documented eGFR at baseline was similar (male ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

n = 855; 51.5%). The median age was 75 (range 30– 97) years in the total population, the oldest subgroups being those with severe and moderately severe renal dysfunction: median age [range] was 80 [35; 92] years in patients with stage IV or V renal function, and the median age [range] was 81 [59; 93] years in patients with stage 3B renal function. The median body mass index ranged from 26.4 to 27.9 kg/m², with comparable values across eGFR classes. The median values of SBP/DBP ranged from 126.5 to 135 mmHg and 75 to 80 mmHg, respectively, across eGFR classes. The most frequent comorbidity was hypertension, which affected 72.9–84.5% of the patients across eGFR classes (76.4% overall), followed by diabetes (38.6–60.7%, 43% overall), coronary heart disease (32.5–48.2%, 37% overall), history of heart failure (20.3–48.2%, 27% overall), peripheral vascular disease (10–32.1%, 13.3% overall) and prior stroke (8– 11.3%, 9.2% overall). Other comorbidities included aortic plaque, which affected 28.6% and 23.6% of the population in stage ≥ 4 and stage 3B patients, respectively. The proportion of patients with a CHADS2 (15) score ≥ 2 at baseline was 67.3% overall; 92.9% in patients with stage IV or V renal function; 87.0%, 79.1% and 58.0% in patients with stage 3B, stage 3A and stage II or I renal function, respectively. The differences were statistically significant across eGFR classes (p < 0.001; Table 2). At baseline the median INR values were almost identical in all the subgroups, ranging from 2.0 to 2.1. The median values of TTR (given by the percentage of the time INR is within the target of 2–3) were 44.7, 48.4, 51.0 and 42.5 in severe (N = 56), moderate/severe (N = 161), moderate (N = 425) and almost normal (N = 1018) renal dysfunction patients, respectively. Overall, the patients were within the TTR range (INR 2–3) about 45.1% of the time on average.

Treatments during follow up The proportion of patients who were treated with vitamin K antagonists (VKA) during follow up (57.4% overall) was significantly different (p < 0.001) between renal function subgroups, the lowest proportion (27.1%) being in patients with severe renal dysfunction; the proportion of patients with moderate renal dysfunction on VKA was comparable between stages 3B and 3A subgroups (55.5% and 58.1%, respectively; Table 3). The proportions of patients treated with AP (26.3% overall) were similar between renal function subgroups, ranging from 25.4% for patients with mild renal dysfunction to 28.2% for patients with moderate renal dysfunction (stage 3A).

717

Comorbidities, n (%) Hypertension History of heart failure Prior stroke Prior hemorrhagic stroke Prior TIA Diabetes CAD MI Peripheral vascular disease Aortic plaque

17 (10.6) 0 (0.0)

6 (3.7) 85 (52.8) 71 (44.1) 14 (8.7) 33 (20.5)

14 (8.7)

6 (10.7) 0 (0.0)

0 (0.0) 34 (60.7) 27 (48.2) 7 (12.5) 18 (32.1)

9 (16.1)

6.4 [5.3, 9.8] 48.39 [0.00, 100.00]

6.3 [4.4, 9.3] 44.65 [0.00, 86.99]

136 (84.5) 68 (42.2)

26.40 [21.14, 36.67] 130.0 [81, 200] 75 [53, 100] 2 (2.5)

27.34 [23.39, 35.94] 126.5 [110, 154] 76 [60, 85] 0 (0.0)

44 (78.6) 27 (48.2)

27.15 [18.89, 43.24] 132.5 [90, 200] 80 [54, 105] 8 (1.7)

27.93 [20.55, 39.39] 135.0 [92, 182] 80 [53, 126] 2 (1.0)

116 (72.0)

42 (75.0)

SBP, mmHg [range] DBP, mmHg [range] Smoking, n (%) [range] HbA1c [range] TTR [range]

354 (34.8)

266 (62.6)

62 (38.5) 81 [59, 93]

42 (9.9)

22 (5.2) 202 (47.5) 185 (43.5) 37 (8.7) 68 (16.0)

48 (11.3) 4 (0.9)

347 (81.6) 146 (34.4)

6.3 [4.6, 10.7] 51.04 [0.00, 100.00]

164 (38.6) 78 [50, 97]

82 (8.1)

36 (3.5) 393 (38.6) 331 (32.5) 54 (5.3) 102 (10.0)

81 (8.0) 11 (1.1)

742 (72.9) 207 (20.3)

6.3 [4.1, 13.6] 42.51 [0.00, 100.00]

601 (59.0) 72 [30, 96]

1018 (61.33)

28 (50.0) 80 [35, 92]

425 (25.60)

161 (9.70)

56 (3.37)

eGFR ≥ 60

eGFR value distribution, n (%) Male, n (%) Age, years [range] Age > 75 years, n (%) BMI (kg/m²) [range]

45 ≤ eGFR < 60

30 ≤ eGFR < 45

eGFR < 30

Parameter

eGFR available at baseline (n = 1660)

Table 2 Clinical characteristics of patients at baseline, by eGFR class at baseline

147 (8.9)

64 (3.9) 714 (43.0) 614 (37.0) 112 (6.7) 221 (13.3)

152 (9.2) 15 (0.9)

1269 (76.4) 448 (27.0)

6.3 [4.1, 13.6] 45.09 [0.00, 100.00]

27.34 [18.89, 43.24] 132.5 [81, 200] 80 [53, 126] 12 (0.72)

778 (46.9)

855 (51.5) 75 [30, 97]

1660 (100)

Total

27.43 [23.39, 35.94] 122.5 [110, 154] 76 [60, 85] 0 (0.0)

0.508

0.173

0.209 < 0.001 < 0.001 0.02 < 0.001

0.2 0.505

< 0.001 < 0.001

0.127

0.42

0.68

< 0.001

8 (16.7)

0 (0.0) 30 (62.5) 21 (43.8) 7 (14.6) 14 (29.2)

5 (10.4) 0 (0.0)

38 (79.2) 23 (47.9)

6.25 [5.2, 9.3] 48.96 [0.00, 86.99]

35 (72.9)

< 0.001

0.624

25 (52.1) 80 [35, 92]

48 (3.49)

eGFR < 30

< 0.001 < 0.001

p-value

12 (8.8)

4 (2.9) 74 (54.0) 62 (45.3) 12 (8.8) 30 (21.9)

15 (10.9) 0 (0.0)

117 (85.4) 59 (43.1)

36 (10.1)

20 (5.6) 175 (48.9) 157 (43.9) 33 (9.2) 61 (17.0)

40 (11.2) 4 (1.1)

296 (82.7) 128 (35.8)

6.3 [5.0, 10.7] 52.11 [0.00, 100.00]

27.97 [20.55, 37.50] 135 [92, 182] 80 [53, 126] 1 (0.6)

26.35 [21.14, 34.67] 132.5 [81, 200] 76 [53, 100] 1 (1.4) 6.3 [5.3, 9.8] 53.82 [0.00, 100.00]

236 (65.9)

132 (36.9) 78 [52, 97]

358 (26.06)

45 ≤ eGFR < 60

104 (75.9)

58 (42.3) 80 [59, 93]

137 (9.97)

30 ≤ eGFR < 45

70 (8.4)

29 (3.5) 348 (41.9) 285 (34.3) 46 (5.5) 86 (10.3)

69 (8.3) 9 (1.1)

613 (73.8) 174 (20.9)

6.3 [4.1, 13.6] 44.52 [0.00, 100.00]

27.24 [18.89, 43.24] 132 [99, 177] 80 [54, 105] 6 (1.6)

320 (38.5)

497 (59.8) 72 [32, 92]

831 (60.48)

eGFR ≥ 60

eGFR available both at baseline and follow up (n = 1374)

126 (9.2)

53 (3.9) 627 (45.6) 525 (38.2) 98 (7.1) 191 (13.9)

129 (9.4) 13 (0.9)

1064 (77.4) 384 (27.9)

6.3 [4.1, 13.6] 48.3 [0.00, 100.00]

27.47 [18.89, 43.24] 132.5 [81, 200] 80 [53, 126] 8 (1.2)

695 (50.6)

712 (51.8) 75 [32, 97]

1374 (100)

Total

0.245

0.142 0.001 0.003 0.018 < 0.001

0.401 0.563

< 0.001 < 0.001

0.194

0.486

0.697

0.007

0.531

0.444

< 0.001

< 0.001 < 0.001

p-value

718 Renal function, attributes and coagulation treatment in atrial fibrillation

ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

eGFR < 30

Thromboembolism 0 (0.0) CHADS2 score, n (%) 0 0 (0.0) 1 4 (7.1) 2 20 (35.7) 3 21 (37.5) 4 9 (16.1) 5 1 (1.8) 6 1 (1.8) CHADS2 52 (92.9) score ≥ 2, n (%) CHA2DS2-VASc score (27), n (%) 0 0 (0.0) 1 1 (1.8) 2 0 (0.0) 3 9 (16.1) 4 17 (30.4) 5 12 (21.4) 6 13 (23.2) 7 3 (5.4) 8 0 (0.0) 9 1 (1.8) INR Value 2 [range] [0.90, 3.20]

Parameter 7 (1.6) 21 (4.9) 68 (16.0) 124 (29.2) 128 (30.1) 61 (14.4) 16 (3.8) 7 (1.6) 336 (79.1)

4 (0.9) 9 (2.1) 33 (7.8) 68 (16.0) 104 (24.5) 102 (24.0) 66 (15.5) 26 (6.1) 12 (2.8) 1 (0.2) 2 [0.86, 6.20]

1 (0.6) 20 (12.4) 47 (29.2) 47 (29.2) 38 (23.6) 4 (2.5) 4 (2.5) 140 (87.0)

0 (0.0) 2 (1.2) 5 (3.1) 24 (14.9) 36 (22.4) 44 (27.3) 32 (19.9) 13 (8.1) 3 (1.9) 2 (1.2) 2 [0.82, 5.40]

45 ≤ eGFR < 60

4 (2.5)

30 ≤ eGFR < 45

eGFR available at baseline (n = 1660)

Table 2 Continued

32 (3.1) 106 (10.4) 197 (19.4) 217 (21.3) 235 (23.1) 134 (13.2) 60 (5.9) 28 (2.8) 7 (0.7) 2 (0.2) 2 [0.81, 6.28]

120 (11.8) 308 (30.3) 309 (30.4) 184 (18.1) 69 (6.8) 19 (1.9) 9 (0.9) 590 (58.0)

13 (1.3)

eGFR ≥ 60

36 (2.2) 118 (7.1) 235 (14.2) 318 (19.2) 392 (23.6) 292 (17.6) 171 (10.3) 70 (4.2) 22 (1.3) 6 (0.4) 2 [0.81, 6.28]

142 (8.6) 400 (24.1) 500 (30.1) 380 (22.9) 177 (10.7) 40 (2.4) 21 (1.3) 1118 (67.3)

24 (1.4)

Total

0.783

< 0.001 0 (0.0) 1 (2.1) 0 (0.0) 9 (18.8) 15 (31.3) 8 (16.7) 11 (22.9) 3 (6.3) 0 (0.0) 1 (2.1) 2 [1, 3]

0 (0.0) 4 (8.3) 18 (37.5) 15 (31.3) 9 (18.8) 1 (2.1) 1 (2.1) 44 (91.7)

< 0.001

< 0.001

0 (0.0)

eGFR < 30

0.5

p-value

0 (0.0) 2 (1.5) 5 (3.6) 20 (14.6) 30 (21.9) 36 (26.3) 29 (21.2) 11 (8.0) 3 (2.2) 1 (0.7) 2 [1, 5]

1 (0.7) 17 (12.4) 40 (29.2) 38 (27.7) 35 (25.5) 3 (2.2) 3 (2.2) 119 (86.9)

4 (2.9)

30 ≤ eGFR < 45

4 (1.1) 7 (2.0) 24 (6.7) 53 (14.8) 84 (23.5) 92 (25.7) 58 (16.2) 25 (7.0) 10 (2.8) 1 (0.3) 2 [1, 6]

19 (5.3) 48 (13.4) 106 (29.6) 109 (30.4) 56 (15.6) 14 (3.9) 6 (1.7) 291 (81.3)

6 (1.7)

45 ≤ eGFR < 60

21 (2.5) 79 (9.5) 159 (19.1) 178 (21.4) 195 (23.5) 115 (13.8) 51 (6.1) 25 (3.0) 6 (0.7) 2 (0.2) 2 [1, 6]

91 (11.0) 237 (28.5) 261 (31.4) 160 (19.3) 56 (6.7) 18 (2.2) 8 (1.0) 503 (60.5)

11 (1.3)

eGFR ≥ 60

eGFR available both at baseline and follow up (n = 1374)

25 (1.8) 89 (6.5) 188 (13.7) 260 (18.9) 324 (23.6) 251 (18.3) 149 (10.8) 64 (4.7) 19 (1.4) 5 (0.4) 2 [1, 6]

111 (8.1) 306 (22.3) 425 (30.9) 322 (23.4) 156 (11.4) 36 (2.6) 18 (1.3) 957 (69.7)

21 (1.5)

Total

0.62

< 0.001

< 0.001

< 0.001

0.425

p-value

Renal function, attributes and coagulation treatment in atrial fibrillation 719

720

Renal function, attributes and coagulation treatment in atrial fibrillation

Table 3 Stroke prevention treatment during follow up, by eGFR class at baseline (N = 1374)

Parameter

No antithrombotic treatment, n (%) Vitamin K antagonist (VKA), n (%) Warfarin, n (%) Median daily dose, mg [range] Median duration, days [range] Phenprocoumon, n (%) Median daily dose, mg [range] Median duration, days [range] Antiplatelet, n (%) Heparin Group, n (%) NSAIDs, n (%) Cox-2 inhibitors, n (%) Change in treatment during follow up, n (%) Median INR value [range] Median TTR

eGFR < 30 (N = 48)

30 ≤ eGFR < 45 (N = 137)

45 ≤ eGFR < 60 (N = 358)

eGFR ≥ 60 (N = 831)

Total (N = 1374)

p-value

25 (52.1)

36 (26.3)

85 (23.7)

218 (26.2)

364 (26.5)

< 0.001*

13 (27.1)

76 (55.5)

208 (58.1)

491 (59.1)

788 (57.4)

< 0.001*

0 (0.0)

1 (1.3) 5.00 [5.00; 5.00]

3 (1.4) 5.00 [5.00; 5.00]

1 (0.2) 5.00 [5.00; 5.00]

5 (0.6) 5.00 [5.00; 5.00]

0.237 < 0.001*

100.00 [100.00; 100.0] 100.00 [58.33; 100.0] 100.00 [100.00; 100.0] 100.00 [58.33; 100.0] 0.833 13 (100.0%) 75 (98.7%) 3.00 [1.50; 3.00] 3.00 [1.50; 4.29]

207 (99.5%) 3.00 [0.50; 3.00]

491 (100.0%) 3.00 [0.50; 9.00]

786 (99.7%) 3.00 [0.50; 9.00]

0.165 0.017*

100 [99; 100]

100 [79; 150]

100 [49; 303]

100 [20; 303]

100 [20; 303]

0.993

13 (27.1) 7 (14.6) 14 (29.2) 1 (2.1) 2 (8.3)

37 (27.0) 32 (23.4) 53 (38.7) 4 (2.9) 15 (14.9)

101 (28.2) 110 (30.7) 138 (38.5) 25 (7.0) 44 (16.2)

211 (25.4) 253 (30.4) 329 (39.6) 41 (4.9) 69 (11.5)

362 (26.3) 402 (29.3) 534 (38.9) 71 (5.2) 130 (13.1)

0.783 0.041* 0.552 0.183 0.216

2.40 [0.90; 5.11] 64.00

2.26 [0.84; 4.90] 59.52

2.30 [0.84; 6.40] 60.63

0.264 0.240

2.45 [0.91; 5.01] 2.34 [0.94; 6.40] 58.11 55.06

*statistically significant value. CAD, coronary artery disease.

Overall, 26.5% of the patients had no antithrombotic treatment. This proportion was 52.1% of patients with severe renal dysfunction and 26.3% and 23.7%, respectively, in patients with moderately severe and moderate renal dysfunction. These results may reflect the reluctance of physicians to use VKAs as renal function decreases even facing the increasing stroke risk. Among patients treated with VKAs, a small proportion (0.6%) was treated by warfarin with a median daily dose of 5 mg for a median duration of 100 days, whereas most of them (99.7%) were treated by phenprocoumon with a median daily dose of 3 mg for a median duration of 100 days. During follow up, the median INR values were almost identical in all subgroups, ranging from 2.26 to 2.45. The median values of TTR were 58.11, 55.06, 64.00 and 59.52 in severe (N = 48), moderately severe (stage 3B; N = 137), moderate (stage 3A;

N = 358) and almost normal (stage 2; N = 831) renal dysfunction patients, respectively.

Evolution of renal function over time The pattern of decreasing renal function (i.e., eGFR) over time was examined. As shown in Figure 4, the majority patients experienced decreasing eGFR during the follow-up period and an increasing trend of deteriorating renal function over time was observed. There were 1374 patients with eGFR values available at baseline and follow up. When compared with baseline eGFR, the mean (SD) eGFR value at 12-month follow up decreased by 2.0 (13.3) ml/min/1.73 m², and the majority (756 patients, 55.0%) showed a decrease in eGFR value. As a consequence, compared with baseline values, the proportion of patients with moderately severe decreased renal function (IIIB) or severely decreased renal function (IV or V) increased by relative 25% at the 12-month follow up (13.5% vs.

Table 4 Number (%) of patients with shift of eGFR between baseline and end of follow up (N = 1374)

Baseline, n (%) Follow up, n (%)

eGFR < 30

30 ≤ eGFR < 45

45 ≤ eGFR < 60

eGFR > 60

Total

p-value

48 (3.49) 67 (4.88)

137 (9.97) 165 (12.01)

358 (26.06) 329 (23.94)

831 (60.48) 813 (59.17)

1374 (100) 1374 (100)

0.01

ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

Renal function, attributes and coagulation treatment in atrial fibrillation

idaemia all remained significantly associated with deterioration of renal function, while age did not (Table 6). Similarly to the CHADS2 scoring system, the CHA2DS2-VASc score was not significantly associated with deterioration of renal function. Comparably, when the model was run with age and the CHA2DS2-VASc score as continuous variables, the presence of diabetes or dyslipidaemia and a history of MI were significantly associated with deterioration of renal function, while age was not (Table 6).

Discussion

Figure 3 Proportion of patients having a CHADS2 score ≥ 2 at baseline and during the follow up (n = 1374).

16.9%). The distribution of eGFR classes showed a statistically significant difference (p = 0.01) between baseline and follow-up time points (Table 4). Figure 3 shows the proportions of patients with a CHADS2 score ≥ 2 for the different classes of eGFR values at baseline and follow up. These proportions were significantly different at both time points (p < 0.001). In addition, the multivariate Cox regression model found that age, presence of diabetes, dyslipidaemia and history of MI were independent and statistically significantly associated with deterioration of renal function (Table 5). After controlling for other confounders, older patients (i.e., age > 75 years) were 1.19 times more likely to experience renal deterioration; the presence of diabetes or dyslipidaemia, respectively, conferred 1.83 times and 1.16 times higher probability of renal deterioration and a history of MI increased the chance by 1.35 times. When replacing the CHADS2 score in the model with the CHA2DS2-VASc score, the presence of diabetes, history of MI and the presence of dyslip-

In this study, we found that among German AF patients with documented eGFR who were managed in routine clinical practices, about 13% had severe or moderately severe renal dysfunction, which is consistent with previous studies from Western countries (16). In addition to AF, most of these patients had other comorbid conditions, the most prevalent being hypertension, diabetes and coronary artery disease. As patients aged, the number of comorbidities increased and renal function decreased as measured by eGFR. The percentage of patients with moderate-to-high risk of stroke (i.e. CHADS2 ≥ 2) increased as renal function decreased, and more than half of theses patients’ function deteriorated during the 12-month follow-up period. Previous research has suggested that decreased renal function is associated with an increased risk of stroke (9,17–19). The results of this study have shown that the proportions of patients with severe and moderate renal dysfunction with a CHADS2 score ≥ 2 increased from baseline to follow up. Moreover, at baseline as well as at follow-up measurements, the CHADS2 scores were significantly different across eGFR classes, with a linear trend observed between the increased risk of stroke (which was estimated by the CHADS2 score in this study)

Table 5 Multivariate Cox analysis of clinical variables for renal function deterioration (N = 1374)

Variables

Events*

HR

CI 95%

Global p-value

Age class (> 75/≤ 75) TIA (yes/no) Heart failure (yes/no) Diabetes (yes/no) Coronary heart disease (yes/no) MI (yes/no) Hypertension (yes/no) Dyslipidaemia (yes/no) CHADS2 (≥ 2/< 2)

353 24 225 360 298 59 581 431 540

1.195 0.850 1.047 1.83 1.053 1.35 0.91 1.16 0.944

[1.01; [0.57; [0.86; [1.48; [0.88; [1.01; [0.73; [1.01; [0.74;

0.037 0.42 0.65 < .0001 0.60 0.04 0.43 0.04 0.65

1.41] 1.26] 1.28] 2.25] 1.28] 1.81] 114] 1.35] 1.21]

*Decreased eGFR value between baseline and follow-up eGFR. HR, hazard ratio; CI, confidence interval.

ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

721

722

Renal function, attributes and coagulation treatment in atrial fibrillation

Table 6 Multivariate Cox analysis of clinical variables for renal function deterioration (N = 1374)

Variables

Events†

HR

CI 95%

p-value

Age class (> 75/≤ 75)* TIA (yes/no) Heart failure (yes/no) Diabetes (yes/no) Coronary heart disease (yes/no) MI (yes/no) Hypertension (yes/no) Dyslipidaemia (yes/no) CHA2DS2-VASc (≥ 2/< 2)* Age‡ TIA (yes/no) Heart failure (yes/no) Diabetes (yes/no) Coronary heart disease (yes/no) MI (yes/no) Hypertension (yes/no) Dyslipidaemia (yes/no) CHA2DS2-VASc‡

353 24 225 360 298 59 581 431 754 817 24 225 360 298 59 581 431 817

1.13 0.83 1.05 1.80 1.062 1.35 0.86 1.35 1.32 1.005 0.836 1.044 1.788 1.08 1.353 0.895 1.162 1.023

[0.97; [0.56; [0.86; [1.45; [0.88; [1.01; [0.70; [1.01; [0.98; [0.99; [0.57; [0.88; [1.51; [0.93; [1.03; [0.74; [1.01; [0.95;

0.12 0.35 0.65 < .0001 0.54 0.04 0.16 0.04 0.06 0.3477 0.363 0.6173 < .0001 0.3124 0.0281 0.2557 0.0414 0.5317

1.30] 1.23] 1.27] 2.13] 1.30] 1.81] 1.06] 1.81] 1.78] 1.01] 1.23] 1.24] 2.11] 1.25] 1.77] 1.08] 1.34] 1.10]

*Categorical variables, †Decreased eGFR value between baseline and follow-up eGFR, ‡Continuous variables. HR, hazard ratio; CI, confidence interval.

Figure 4 Percentage of decreasing eGFR values according to length of follow up (n = 1374)

and the decrease in eGFR value, suggesting an association of decreasing renal function with increased risk of stroke over time. Our study identified and confirmed several cardiovascular disease risk factors associated with severe renal dysfunction, most of them taken into account in the CHADS2 and CHA2DS2-VASc scores, including age ≥ 75 years and several comorbidities such as T2DM, previous history of heart failure, stroke or TIA or presence of aortic plaques. In addition, we found that the proportion of patients treated with antithrombotic medication for stroke prevention during the 12-month follow up decreased with decreasing renal function, and nearly half of the patients with severe renal dysfunction had no antithrombotic therapies prescribed by their physicians. These results

indicate underutilisation of potential life-saving stroke prevention therapies in these high-risk patients with the highest need (20). As this is a retrospective database study, we can only speculate about the reasons why physicians underprescribe VKAs in these highrisk patients. Renal dysfunction is not a contraindication for VKA treatment. Previous studies report poor compliance with warfarin in AF patients without contraindication, as well as undertreatment with VKA (6,20–22). Our study demonstrated that at least comparable treatment gaps existed in AF patients with and without renal dysfunction. We speculate that the therapeutic limitations associated with VKA, such as a narrow therapy window, drug-to-drug interactions, drug-to-food interactions and the presence of multiple comorbidities in renal dysfunction patients may ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

Renal function, attributes and coagulation treatment in atrial fibrillation

result in a more cautious approach by physicians in their utilisation of VKAs despite an increasing risk of stroke. Nevertheless, our study suggested treatment gaps in stroke prevention in AF patients with renal dysfunction, particularly in those with severe renal dysfunction with elevated risk of stroke. It has been shown that patients with renal impairment and AF are at high risk for bleeding when anticoagulated (17,23,24) and this could be a contributor to the decrease in prescribing as kidney function decreases in AF patients. After the study period, a few innovative therapies (e.g. dabigatran, rivoroxaban, apixaban) entered the marketplace. As there is lack of clinical evidence of these medications in patients with severe renal impairment, it is not clear how these new medications can address the aforementioned treatment gaps. Further research will be needed. During the follow-up period, the majority of patients (55%) had a slight decrease in renal function. The results of this study showed that a decrease in renal function tended to accelerate with longer followup period. These results are consistent with those of Roldan et al. who showed that renal function does not remain static in AF patients, but rather can progress from mild to severe in only 2 years (25). This study identified age, presence of diabetes or dyslipidaemia and history of MI as significant risk factors contributing to the deterioration of renal function in AF patients after controlling for other potential clinical confounding factors. Of note, it has been long accepted that renal impairment is a risk factor for stroke and thromboembolism (17). However, renal impairment has not been included in either the CHADS2 or the CHA2DS2-VASc scoring systems because of the lack of data validating its utility. Banerjee et al. found that renal impairment was not an independent predictor of stroke, and that adding it to either scoring system did not improve their predictive ability (26). Because the combination of AF and severe renal impairment may confer greater thromboembolic or stroke risk, which may then lead to treatment with anticoagulants, bleeding risk may also be increased. Therefore, renal function should be carefully monitored in AF patients throughout treatment. There are several limitations of our study. As it is not a mandatory requirement for participating physicians to enter all patients’ information into the database, our data source may not have collected all available eGFR results. Among all eligible study patients (n = 15,900), there were only 10% with documented eGFR values. This may have some negative impact on the generalisability of our study findings, rendering many of the observed eGFR values biased by indication, but this limitation is inherent to any observational electronic medical record database study, as ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724

physicians record the information deemed necessary for their routine practices. On the other hand, as the data were collected in a natural way without any interventions to healthcare providers, this type of study may more accurately reflect real-world physician practices. Another potential limitation is that all physicians contributing to the database are GPs, and as a result, our study may miss some German AF patients with severe renal or cardiac dysfunction routinely managed by specialists, further limiting the generalisability of our study findings to the German AF population. However, as our study objective was to examine German AF patients managed by GPs in their routine practices, this potential limitation does not compromise the value of this study. Finally, assessing bleeding in addition to thromboembolic events may be an important component to management of AF patients. However, the HAS-BLED score could not be determined in this group of patients based on the information available in the database. Information on concomitant use of alcohol was not collected, and baseline abnormal liver function may have been underrecorded, both of which are key components of the score. Although we cannot make population-wide projections based on this study, study limitations do not preclude us from drawing conclusions within the scope of the study’s objectives. Here, we reported on the important issue of renal deficiency in routine clinical practices in Germany, and highlighted the importance of further research and advancement of quality treatment in this underserved subpopulation.

Conclusions Our study found that renal dysfunction, as measured by eGFR, is a prevalent condition in German AF patients with documented eGFR managed by their GPs in routine clinical practices. The risk of stroke, as measured by the CHADS2, is significantly associated with decreased renal function. Treatment with antithrombotic therapy decreases as renal dysfunction increases, despite the increasing risks of stroke. Decreasing renal function over time was prevalent in these patients, resulting in more patients with moderately severe or severe renal impairment during the minimum 12-month follow-up period compared with the baseline. Based on the results from the multivariate Cox regression model, after controlling for other relevant risk factors, age, diabetes, dyslipidaemia and MI were found to be independent and statistically significant contributors to renal deterioration in German AF patients managed by their routine clinical practices. Decreasing renal function in patients with AF presents an additional risk of stroke. In our study, we found that German AF patients with severe renal

723

724

Renal function, attributes and coagulation treatment in atrial fibrillation

dysfunction appear inadequately protected regarding the risk of stroke; numerous patients who are qualified to receive preventive treatment by VKA remain untreated. There is still a significant level of unmet medical need in preventing the risk of stroke in a remarkable number of patients with atrial fibrillation and renal dysfunction. Further research is needed to understand the risk/benefit of utilisation of anticoagulant medications in renal deficiency patients.

Acknowledgments Merck & Co., Inc., provided funding for the study. Mary E Hanson, PhD, of Merck & Co., Inc., provided writing and editorial assistance. Jennifer Rotonda, PhD, of Merck & Co., Inc., provided editorial assistance. Fei Cao Ghoul, PhD, of Cegedim Strategic Data and Alain Y. Platel, PhD, of APMW provided editorial assistance. This assistance was funded by Merck & Co., Inc.

References 1 Ryder KM, Benjamin EJ. Epidemiology and significance of atrial fibrillation. Am J Cardiol 1999; 84: 131R–8R. 2 Ananthapanyasut W, Napan S, Rudolph EH et al. Prevalence of atrial fibrillation and its predictors in nondialysis patients with chronic kidney disease. Clin J Am Soc Nephrol 2010; 5: 173–81. 3 Wolf PA, Abbott RD, Kannel WB. Atrial fibrillation as an independent risk factor for stroke: the Framingham Study. Stroke 1991; 22: 983–8. 4 Lip GY, Boos CJ. Antithrombotic treatment in atrial fibrillation. Heart 2006; 92: 155–61. 5 Deo R, Katz R, Kestenbaum B et al. Impaired kidney function and atrial fibrillation in elderly subjects. J Card Fail 2010; 16: 55–60. 6 Go AS, Hylek EM, Borowsky LH, Phillips KA, Selby JV, Singer DE. Warfarin use among ambulatory patients with nonvalvular atrial fibrillation: the anticoagulation and risk factors in atrial fibrillation (ATRIA) study. Ann Intern Med 1999; 131: 927–34. 7 Baber U, Howard VJ, Halperin JL et al. Association of chronic kidney disease with atrial fibrillation among adults in the United States: REasons for Geographic and Racial Differences in Stroke (REGARDS) Study. Circ Arrhythm Electrophysiol 2011; 4: 26–32. 8 Go AS, Fang MC, Udaltsova N et al. Impact of proteinuria and glomerular filtration rate on risk of thromboembolism in atrial fibrillation: the anticoagulation and risk factors in atrial fibrillation (ATRIA) study. Circulation 2009; 119: 1363–9. 9 Horio T, Iwashima Y, Kamide K et al. Chronic kidney disease as an independent risk factor for new-onset atrial fibrillation in hypertensive patients. J Hypertens 2010; 28: 1738–44. 10 Lee M, Saver JL, Chang KH, Liao HW, Chang SC, Ovbiagele B. Low glomerular filtration rate and risk of stroke: meta-analysis. BMJ 2010; 341: c4249.

Author contributions DW conceived, designed or planned the study, interpreted the results, provided substantive suggestions for revision to the manuscript and obtained funding; GM interpreted the results, provided substantive suggestions for revision to the manuscript, obtained funding and provided administrative/logistic support; CK interpreted the results, wrote sections of the initial draft and provided substantive suggestions for revision to the manuscript and provided statistical expertise; MSS conceived, designed or planned the study, performed or supervised analyses and provided substantive suggestions for revision to the manuscript; JC collected or assembled the data, performed or supervised analyses, provided substantive suggestions for revision to the manuscript and provided statistical expertise. All authors reviewed and approved the final version of the manuscript.

11 World Health Organization. International Statistical Classification of Diseases and Related Health Problems 10th Revision. Geneva, Switzerland: WHO Press, 2011. 12 Rosendaal FR, Cannegieter SC, van der Meer FJ, Briet E. A method to determine the optimal intensity of oral anticoagulant therapy. Thromb Haemost 1993; 69: 236–9. 13 Levey AS, Coresh J, Balk E et al. National Kidney Foundation practice guidelines for chronic kidney disease: evaluation, classification, and stratification. Ann Intern Med 2003; 139: 137–47. 14 Levey AS, Bosch JP, Lewis JB, Greene T, Rogers N, Roth D. A more accurate method to estimate glomerular filtration rate from serum creatinine: a new prediction equation. Modification of Diet in Renal Disease Study Group. Ann Intern Med 1999; 130: 461–70. 15 Gage BF, Waterman AD, Shannon W, Boechler M, Rich MW, Radford MJ. Validation of clinical classification schemes for predicting stroke: results from the National Registry of Atrial Fibrillation. JAMA 2001; 285: 2864–70. 16 Alonso A, Lopez FL, Matsushita K et al. Chronic kidney disease is associated with the incidence of atrial fibrillation: the Atherosclerosis Risk in Communities (ARIC) study. Circulation 2011; 123: 2946–53. 17 Marinigh R, Lane DA, Lip GY. Severe renal impairment and stroke prevention in atrial fibrillation: implications for thromboprophylaxis and bleeding risk. J Am Coll Cardiol 2011; 57: 1339–48. 18 Olesen JB, Lip GY, Kamper AL et al. Stroke and bleeding in atrial fibrillation with chronic kidney disease. N Engl J Med 2012; 367: 625–35. 19 Vazquez E, Sanchez-Perales C, Garcia-Garcia F et al. Atrial fibrillation in incident dialysis patients. Kidney Int 2009; 76: 324–30.

20 Le Heuzey JY, Paziaud O, Piot O et al. Cost of care distribution in atrial fibrillation patients: the COCAF study. Am Heart J 2004; 147: 121–6. 21 Stafford RS, Singer DE. Recent national patterns of warfarin use in atrial fibrillation. Circulation 1998; 97: 1231–3. 22 Mazzaglia G, Filippi A, Alacqua M et al. A national survey of the management of atrial fibrillation with antithrombotic drugs in Italian primary care. Thromb Haemost 2010; 103: 968–75. 23 Manzano-Fernandez S, Cambronero F, Caro-Martinez C et al. Mild kidney disease as a risk factor for major bleeding in patients with atrial fibrillation undergoing percutaneous coronary stenting. Thromb Haemost 2012; 107: 51–8. 24 Roldan V, Marin F, Muina B et al. Plasma von Willebrand factor levels are an independent risk factor for adverse events including mortality and major bleeding in anticoagulated atrial fibrillation patients. J Am Coll Cardiol 2011; 57: 2496–504. 25 Roldan V, Marin F, Fernandez H et al. Renal impairment in a “real-life” cohort of anticoagulated patients with atrial fibrillation (implications for thromboembolism and bleeding). Am J Cardiol 2013; 111: 1159–64. 26 Banerjee A, Fauchier L, Vourc’h P et al. Renal impairment and ischemic stroke risk assessment in patients with atrial fibrillation: the Loire Valley Atrial Fibrillation Project. J Am Coll Cardiol 2013; 61: 2079–87. 27 Odum LE, Cochran KA, Aistrope DS, Snella KA. The CHADS(2)versus the new CHA2DS2-VASc scoring systems for guiding antithrombotic treatment of patients with atrial fibrillation: review of the literature and recommendations for use. Pharmacotherapy 2012; 32: 285–96.

Paper received July 2013, accepted December 2013

ª 2014 John Wiley & Sons Ltd Int J Clin Pract, June 2014, 68, 6, 714–724