Chronic kidney disease (CKD) is a potential independent risk factor for atrial fibrillation (AF). It remains unclear whether anemia is synergistically associated with increased risk of AF onset in subjects with CKD. We evaluated the association of kidney function, hemoglobin (Hb), and their combination with new-onset AF in a population-based cohort study. We conducted a 15-year prospective cohort study of 132,250 Japanese subjects aged 40 to 79 years who participated in annual health checkups from 1993. Kaplan-Meier survival analysis was used to compare freedom from new-onset AF between groups classified by estimated glomerular filtration rate grade, Hb grade, and their combination. Cox proportional hazard model analysis was used to estimate hazard ratios (HRs) for new-onset AF. During a 13.8-year mean follow-up period, 1,232 (0.93%) subjects with new-onset AF were identified. Lower estimated glomerular filtration rate and lower Hb grades were significantly associated with a higher incidence of new-onset AF. Multivariate HRs and 95% confidence intervals (CIs) of new-onset AF were 1.38 (1.21 to 1.56) for mild CKD group, 2.56 (2.09 to 3.13) for CKD group, and 1.50 (1.24 to 1.83) for anemia group. Borderline Hb level was not significantly associated with increased risk for new-onset AF (HR 1.07, CI 0.91 to 1.25, p = 0.4284). In the model with interaction term between CKD and anemia, the risk was significantly higher (p = 0.0343 for the interaction) than that predicted by each factor independently. In conclusion, decreased kidney function and lower Hb level are associated with increased risk for new-onset AF, especially when both are present.
Risk factors for developing atrial fibrillation (AF), such as advanced age, hypertension, and structural heart disease, significantly overlap with risk factors for renal insufficiency. Chronic kidney disease (CKD) has been shown to be a powerful predictor of cardiovascular prognosis, and decreased estimated glomerular filtration rate (eGFR) is clearly associated with an increase in future cardiovascular events. Approximately 4% of elderly patients with CKD have AF, and CKD is associated with the incidence of AF after adjustment for other risk factors. Anemia, a major complication of CKD, is also a risk factor for left ventricular hypertrophy, de novo and recurrent heart failure, and all-cause mortality in patients with CKD. Although several studies have reported on the relation between CKD and AF, the association between anemia and AF onset has not been fully investigated. Moreover, there is no information about whether the combination of CKD with anemia is a synergistic risk factor for AF onset. Therefore, we conducted a large-scale cohort study to determine whether renal dysfunction, anemia, and their interaction are associated with increased risk for new onset of AF in a general Japanese population.
Methods
In 1993, the Ibaraki prefectural government initiated a large community-based cohort study, known as the Ibaraki Prefectural Health Study, to obtain information on health status for the purposes of health education and policy making. The cohort included 194,333 subjects (63,865 men and 130,468 women aged 40 to 79 years) living in Ibaraki Prefecture, Japan, who completed an annual health checkup in 1993.
We excluded 58,022 subjects who did not participate in the 1994 health checkup because they were not followed up for at least 1 year. We also excluded 2,505 subjects from the analysis because of incomplete data. We finally excluded 1,198 subjects with AF and 358 with polycythemia (hemoglobin [Hb] > 18 g/dl in men and Hb > 16 g/dl in women) at baseline. Thus, 132,250 subjects were enrolled in the present study.
Informed consent was obtained from community representatives to conduct an epidemiologic study according to the guidelines of the Council for International Organizations of Medical Sciences. The Ethics Committee of Ibaraki Prefecture approved this study.
Electrocardiograms were obtained at study baseline and at yearly follow-up. All electrocardiograms were read independently by trained physicians, and disagreements were settled by the laboratory supervisor or the investigators. AF was defined as the absence of P waves before each QRS complex, irregular atrial electrical activity with fibrillatory waves varying in size, shape, and timing, and completely irregular RR intervals. As the study end point, new-onset AF was defined as the first presentation of AF during follow-up, which was conducted every year until the end of 2008. The mean follow-up period was 13.8 years.
Levels of Hb and serum creatinine (Cr) and the other risk factors were measured at baseline survey. Cr was measured by the modified method of Jaffe’s reaction using an RX-30 automated analyzer (Nihon Denshi Inc., Tokyo, Japan). eGFR was calculated using the abbreviated equation developed for the Modification of Diet in Renal Disease study as follows: eGFR (mL/min/1.73 m 2 ) = 194 × Cr −1.094 × age −0.287 (0.739 for women). CKD was defined as decreased eGFR <60 mL/min/1.73 m 2 , regardless of the presence of proteinuria. Subjects were divided into 3 groups according to CKD stage as using the guidelines of the National Kidney Foundation classification of CKD as follows: eGFR ≧90 mL/min/1.73 m 2 (normal eGFR group), 60 ≦ eGFR <90 mL/min/1.73 m 2 (mild CKD group), and eGFR <60 mL/min/1.73 m 2 (CKD group).
Hb measurement was performed on a Coulter STKS automated hematology analyzer (Coulter, Fullerton, California). The subjects were divided in to 3 groups according to Hb level as follows: normal group, 15 g/dl ≦ Hb < 18 g/dl in men and 13 g/dl ≦ Hb < 16 g/dl in women; borderline group, 13 g/dl ≦ Hb < 15 g/dl in men and 11 g/dl ≦ Hb < 13 g/dl in women; and anemia, as defined by the World Health Organization criteria of Hb < 13 g/dl for men and Hb < 11 g/dl for women.
We also measured several potential confounders: smoking habit, drinking habit, blood pressure (BP), body mass index (BMI), serum total cholesterol (TC), triglyceride (TG), high-density lipoprotein (HDL) cholesterol, and diabetes mellitus (DM). Subjects were interviewed to ascertain the number of cigarettes smoked per day, usual weekly intake of alcohol, and medical histories of heart and kidney diseases. Past or present smokers were considered smokers, and those drinking alcohol every day were considered alcohol drinkers. DM was determined by a plasma glucose level of ≥126 mg/dl in the fasting state, plasma glucose level of ≥200 mg/dl in the normal daily state, or treatment of DM.
The p values for difference of baseline characteristics according to the eGFR and Hb classifications were calculated by an analysis of variance for age, BP, BMI, TC, TG, HDL, Cr, eGFR, and Hb and by chi-square test for gender, DM, and smoking and drinking habits. AF event-free curves were derived by means of the Kaplan-Meier method and were compared by log-rank test. Hazard ratios (HRs) and 95% confidence intervals (CIs) for new-onset AF were calculated with reference to the normal grade with the Cox proportional hazards regression model. We adjusted for age, gender at baseline, and for the following potential confounders: BP, BMI, TC, TG, HDL, smoking, drinking, and DM. The interaction between CKD and anemia status was tested by introducing a cross product of the 2 dichotomous variables in the model. A p value <0.05 denoted the presence of a statistically significant difference. All statistical analyses were conducted using SAS, version 9.3 (SAS Institute, Inc., Cary, North Carolina).
Results
The baseline characteristics of the subjects according to eGFR and Hb grades are listed in Table 1 . Mean age was significantly higher in subjects with lower eGFR and lower Hb values. BP, BMI, TC, and TG were higher in subjects with lower eGFR and inversely lower in those with lower Hb. Smokers were more predominant and drinkers rather less predominant in the CKD group. DM was also more prevalent in the CKD group but less prevalent in anemia group.
| Variable | Glomerular filtration rate (ml/min/1.73 m 2 ) | P value | Hemoglobin (g/dl) ∗ | P value | ||||
|---|---|---|---|---|---|---|---|---|
| ≧ 90 | 60 ≦ < 90 | < 60 | 15 ≦ < 18 | 13 ≦ < 15 | < 13 | |||
| n = 57,501 (43.5%) | n = 66,787 (50.5%) | n = 7962 (6.0%) | <.0001 | n = 70,412 (53.2%) | n = 54,833 (41.5%) | n = 7005 (5.3%) | <.0001 | |
| Person・year | 343,553.0 | 388,347.0 | 37,623.4 | <.0001 | 414,298.9 | 317,985.1 | 36,239.4 | <.0001 |
| Age (year) | 54.9 ± 10.1 | 61.9 ± 8.4 | 68.9 ± 6.8 | <.0001 | 58.4 ± 9.6 | 60.4 ± 10.1 | 59.1 ± 12.3 | <.0001 |
| Men | 28.1% | 34.1% | 34.2% | <.0001 | 26.0% | 37.0% | 43.0% | <.0001 |
| Systolic blood pressure (mm/Hg) | 130.5 ± 17.5 | 135.3 ± 17.2 | 139.5 ± 17.1 | <.0001 | 134.6 ± 17.4 | 132.3 ± 17.6 | 131.2 ± 18.1 | <.0001 |
| Diastolic blood pressure (mm/Hg) | 77.6 ± 10.7 | 79.4 ± 10.5 | 79.6 ± 10.5 | <.0001 | 79.9 ± 10.6 | 77.3 ± 10.5 | 75.5 ± 10.6 | <.0001 |
| Body mass index (kg/m2) | 23.1 ± 3.0 | 23.7 ± 3.1 | 23.9 ± 3.2 | <.0001 | 24.0 ± 3.1 | 22.9 ± 2.9 | 22.1 ± 3.0 | <.0001 |
| Total cholesterol (mg/dl) | 199.7 ± 34.6 | 206.7 ± 35.0 | 208.5 ± 36.4 | <.0001 | 210.2 ± 34.9 | 198.3 ± 33.6 | 181.7 ± 32.1 | <.0001 |
| Triglyceride (mg/dl) | 127.7 ± 80.1 | 146.4 ± 84.7 | 157.2 ± 84.6 | <.0001 | 150.3 ± 89.6 | 128.2 ± 74.3 | 106.8 ± 63.3 | <.0001 |
| High-density lipoprotein cholesterol (mg/dl) | 57.2 ± 14.4 | 54.5 ± 14.3 | 52.2 ± 14.4 | <.0001 | 55.2 ± 14.3 | 55.9 ± 14.5 | 55.2 ± 14.7 | <.0001 |
| Creatine (mg/dl) | 0.5 ± 0.1 | 0.7 ± 0.1 | 0.9 ± 0.4 | <.0001 | 0.6 ± 0.1 | 0.6 ± 0.1 | 0.7 ± 0.4 | <.0001 |
| eGFR (ml/min/1.73 m2) | 109.5 ± 18.7 | 77.5 ± 8.1 | 53.2 ± 7.3 | <.0001 | 90.2 ± 22.3 | 89.5 ± 22.5 | 91.7 ± 27.4 | <.0001 |
| Hemoglobin (g/dl) | 13.4 ± 1.5 | 13.7 ± 1.3 | 13.3 ± 1.4 | <.0001 | 14.3 ± 1.1 | 13.0 ± 1.0 | 10.8 ± 1.6 | <.0001 |
| Smoker | 25.9% | 28.9% | 29.1% | <.0001 | 24.7% | 30.4% | 35.3% | <.0001 |
| Daily alcohol drinker | 19.7% | 18.8% | 13.7% | <.0001 | 17.6% | 20.6% | 19.0% | <.0001 |
| Diabetes mellitus | 1.8% | 2.8% | 4.9% | <.0001 | 2.8% | 2.1% | 2.1% | <.0001 |
| New-onset AF (incidence rate, % per year) | 409 (0.12%) | 693 (0.18%) | 130 (0.35%) | <.0001 | 578 (0.14%) | 573 (0.18%) | 81 (0.22%) | <.0001 |
∗ These values are for men, these for women as a follows: 13 ≦ Hb < 16; 11 ≦ Hb < 13; Hb < 11.
During the 15-year (mean 13.8 years) follow-up period through 2008, new onset of AF was observed in 1,232 subjects, and the incidence rate was 0.16% per year. AF incidence was significantly higher in the CKD group and anemia group. Kaplan-Meier event-free analysis demonstrated that subjects with CKD were more susceptible to AF (log-rank p <0.0001; Figure 1 ). The subjects with anemia had slightly but significantly higher risk for AF incidence (log-rank p <0.0001; Figure 1 ). Multivariate adjusted HRs for new-onset AF were significantly higher in the mild CKD and CKD groups and were significantly higher in the anemia group, whereas borderline Hb level was not a significant risk for new-onset AF ( Table 2 ).
| Age-adjusted HR (95% CI) | p value | Multivariable HR (95% CI) | p value | |
|---|---|---|---|---|
| Glomerular filtration rate (ml/min/1.73 m 2 ) | ||||
| eGFR ≧ 90 | 1 | 1 | ||
| 60 ≦ eGFR < 90 | 1.51 (1.34-1.71) | <.0001 | 1.38 (1.21-1.56) | <.0001 |
| eGFR < 60 (CKD) | 3.05 (2.50-3.72) | <.0001 | 2.56 (2.09-3.13) | <.0001 |
| Hemoglobin (g/dl) ∗ | ||||
| 15 ≦ Hb < 18 | 1 | 1 | ||
| 13 ≦ Hb < 15 | 1.08 (0.96-1.23) | 0.8637 | 1.07 (0.91-1.25) | 0.4284 |
| Hb < 13 (anemia) | 1.59 (1.32-1.91) | 0.0001 | 1.50 (1.24-1.83) | <.0001 |
∗ These values are for men, these for women as a follows: 13 ≦ Hb < 16; 11 ≦ Hb < 13; Hb < 11. Multivariate analysis adjusted for age, sex, systolic blood pressure, diastolic blood pressure, body mass index, total cholesterol, triglyceride, high-density lipoprotein cholesterol, smoking, alcohol drinking, and diabetes mellitus.
HRs for new-onset AF according to the combined classification with eGFR and Hb grades are listed in Table 3 . Lower Hb level was significantly associated with a higher risk for new-onset AF in each eGFR grade. Even mild CKD complicated with anemia was significantly associated with a higher AF incidence, whereas a borderline level of Hb was not significantly associated with increased risk for incidence of AF in the normal eGFR group. The interaction term was statistically significant for new onset of AF (p = 0.0343). The combination of CKD with anemia was significantly associated with a higher AF incidence than expected from the individual effects ( Table 4 ). Compared with subjects without both anemia and CKD as the reference, subjects with both anemia and CKD were at particularly high risk for new-onset AF. Kaplan-Meier curves of the association of CKD and anemia interaction grade with freedom from AF incidence showed that subjects with both anemia and CKD were at particularly high risk for new-onset AF ( Figure 2 ).
| Estimated Glomerular Filtration Rate | Age-adjusted HR (95% CI) | p value | Multivariable HR (95% CI) | p value |
|---|---|---|---|---|
| ≧90 | ||||
| 15 ≦ Hb < 18 ∗ | 1 | 1 | ||
| 13 ≦ Hb < 15 ∗ | 1.15 (0.94-1.41) | 0.1702 | 1.12 (0.92-1.37) | 0.2702 |
| Hb < 13 (anemia) ∗ | 1.24 (0.84-1.84) | 0.2836 | 1.23 (1.04-1.48) | 0.0180 |
| 60 ≦ 90 | ||||
| 15 ≦ Hb < 18 ∗ | 1.39 (1.16-1.65) | 0.0003 | 1.35 (0.91-2.01) | 0.1415 |
| 13 ≦ Hb < 15 ∗ | 1.85 (1.55-2.20) | <.0001 | 1.59 (1.33-1.89) | <.0001 |
| Hb < 13 (anemia) ∗ | 2.96 (2.11-4.15) | <.0001 | 2.26 (1.60-3.17) | <.0001 |
| <60 (CKD) | ||||
| 15 ≦ Hb < 18 ∗ | 2.52 (1.83-3.46) | <.0001 | 2.11 (1.53-2.90) | <.0001 |
| 13 ≦ Hb < 15 ∗ | 3.93 (3.00-5.15) | <.0001 | 3.00 (1.67-5.38) | <.0001 |
| Hb < 13 (anemia) ∗ | 4.37 (2.43-7.82) | <.0001 | 3.22 (2.43-4.19) | 0.0003 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
