Association Between White Blood Cell Count and Atrial Fibrillation Risk　― A Population-Based Prospective Cohort Study ―

Abstract

Background: The incidence and prevalence of atrial fibrillation (AF) are increasing. The white blood cell (WBC) count is an indicator of systemic inflammation and is related to increased cardiovascular disease risk. Using data from the Suita Study, we investigated the association between WBC count and AF risk in the general Japanese population.

Methods and Results: This prospective cohort study included 6,884 people, aged 30–84 years, with no baseline AF. Cox regression was used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for AF incidence by WBC count quintile. Within a median follow-up period of 14.6 years, 312 AF cases were diagnosed. Compared with the lowest WBC count quintile, the highest quintile was associated with an increased AF risk (HR 1.57; 95% CI 1.07–2.29). The association was more pronounced among women than men (HR 2.16 [95% CI 1.10–4.26] and 1.55 [95% CI 0.99–2.44], respectively; P interaction=0.07), and among current than non-smokers (HR 4.66 [95% CI 1.89–11.50] and 1.61 [95% CI 1.01–2.57], respectively; P interaction=0.20). For each 1.0×10⁹-cells/L increment in WBC count, AF risk increased by 9% in men (9% in non-smokers, 10% in current smokers) and 20% in women (13% in non-smokers, 32% in current smokers).

Conclusions: A higher WBC count was positively associated with an elevated AF risk in the general Japanese population, especially in women who smoked.

The increasing incidence and prevalence of atrial fibrillation (AF), especially in aging populations, is a rising public health challenge.^1,2 In addition to its considerable financial and societal burdens,³ AF is a major risk factor for myocardial infarction (MI), stroke, dementia, heart failure, and cardiovascular disease (CVD) mortality.^4–6 Still, AF is not inevitable and could be prevented by managing its risk factors.^7,8

Evidence now suggests that inflammation may be a key process in the development of AF, and this inflammatory process is potentially modifiable.^9–11 The white blood cell (WBC) count is a widely measured inflammatory biomarker. Many studies have reported that an elevated WBC count could be associated with a higher risk of fatal and non-fatal MI, stroke, and CVD.^12–17 In addition, a few studies have investigated the potential association between WBC count and AF risk.^18–20 The Framingham Heart Study and the Norwegian Tromsø Study showed positive associations between the WBC count and incident AF,^18,19 whereas the Atherosclerosis Risk in Communities (ARIC) Study showed no significant association.²⁰

In addition to their scarcity, inconclusive findings, and under-representation of Asian populations, the previous studies did not describe the effect of smoking on the association between the WBC count and AF risk,^18–20 despite evidence of varying associations between WBC count and CVD risk according to smoking status.^12–14 Therefore, in the present study, we used data from the Suita Study to investigate whether the WBC count could influence AF risk in a representative sample of the Japanese population, and stratified the results by smoking status. We hypothesized that the WBC count was associated with AF risk, and that this association is stronger among smokers.

Methods

Study Population and Design

In the present population-based prospective cohort study, people residing in the urban city of Suita in southwest Japan were randomly selected by sex and age group. During the baseline health examination conducted at the National Cerebral and Cardiovascular Center (NCVC) in Suita, participants’ sociodemographic and clinical characteristics were assessed before drawing blood samples and conducting electrocardiograms (ECGs). All participants were asked to attend follow-up visits.^20,21 Of the 8,360 participants initially included in the study, those with AF or atrial flutter at baseline (n=42), those for whom no data about the WBC count were available (n=5), or those lost to follow-up (n=1,429) were excluded, leaving 6,884 participants for analysis.

The Suita Study protocol was approved by the Institutional Review Board of the NCVC (M25-043-4). The study was conducted in accordance with the Declaration of Helsinki, and all participants provided written informed consent before participation.

AF Diagnosis, Determination of the WBC Count, and Assessment of Smoking Behavior

Ascertainment of AF in the Suita Study has been described elsewhere.^22–26 Briefly, during the follow-up visits conducted every couple of years at the NCVC, AF was diagnosed, per the Minnesota Codes (8-3-1 to 4), by 2 trained internists using a standard 12-lead rest ECG. Participants’ hospital records, including AF medication, checkups, and death certificates, were systematically reviewed to ascertain AF diagnosis. Blood samples, obtained from participants during the baseline examination, were collected in tubes containing anticoagulants, and the WBC count was determined using an automated hematology analyzer. Smoking was assessed using a baseline questionnaire administered by trained nurses. The questionnaire asked participants whether they were never, former, or current smokers. Current smokers were asked about their duration of smoking and the number of cigarettes smoked.

Statistical Analysis

Chi-squared and one-way ANOVA tests were used to assess the significance of differences in participants’ categorical and numerical baseline characteristics across WBC count quintiles (Q1–Q5). The Cox proportional hazards models was used to calculate age-, sex-, and multivariable-adjusted hazard ratios (HRs) and 95% confidence intervals (CIs) for AF risk for participants in Q2–Q5 compared with Q1. Then, the results were stratified by sex and smoking status into the following groups: male/non-smoker, male/current smoker, female/non-smoker, and female/current smoker. We further calculated associations in the overall and stratified analyses per 1.0×10⁹-cell/L difference in WBC count. Because of the limited number of incident AF cases among women who smoked (n=12), the associations by quintiles were omitted in this group.

To examine the possible effect of occult diseases and limit the influence of reverse causality, we conducted a sensitivity analysis by excluding participants with WBC counts <3.1 or >8.4×10⁹/L (n=434). In addition, we stratified the results by age, body mass index (BMI), and alcohol drinking categories using variable-specified quintiles, and the corresponding P interaction values were computed. Person-years of follow-up were calculated from the date of baseline examination (between 1989 and 2003) to the date of AF diagnosis, death, censoring, or the last health examination (up to 2015), whichever came first.

After reviewing previous literature^18–20 and the Suita AF risk score,²³ the following baseline characteristics were included in the Cox regression models: age, sex, BMI, smoking and alcohol drinking behaviors, physical activity (PA), hypertension, diabetes, chronic kidney disease (CKD), non-high-density lipoprotein cholesterol (non-HDL-C), the presence of cardiac murmur or valvular disease, arrhythmia other than AF, and history of MI and stroke. Age, sex, smoking, drinking, PA, and histories of MI and stroke were assessed by a questionnaire; BMI, hypertension, cardiac murmurs, and valvular disease were assessed by clinical investigations; diabetes, CKD, and non-HDL-C were assessed by blood testing; and arrhythmia was assessed by ECG.

Statistical analyses were performed using SAS version 9.4 (SAS Institute, Cary, NC, USA).

Results

The highest WBC count quintile (Q5) had a higher proportion of men, obesity, current smokers, heavy drinkers, and participants with diabetes than the lower WBC count quintiles (P<0.001). However, participants in Q5 were relatively younger (P=0.004) and a lower proportion had elevated non-HDL-C (P<0.001; Table 1).

Table 1. Characteristics of Participants by WBC Count

Characteristics	WBC count quintile					P for difference
	Q1	Q2	Q3	Q4	Q5
Mean (±SD) age (years)	55.7±12.8	55.8±13.1	56.5±12.7	56.3±12.5	54.8±12.6	0.004
Men	31.6	39.9	47.5	53.2	63.0	<0.001
Body mass index (kg/m2)
<18	10.9	8.8	7.2	6.7	6.3	<0.001
18–24.9	75.0	73.5	71.5	70.9	68.4
≥25	14.1	17.7	21.3	22.4	25.3
Smoking status
Never	71.7	64.0	55.9	48.6	34.3	<0.001
Former	16.5	16.7	19.3	18.8	13.2
Current
≤20 cigarettes/day	10.1	14.4	18.1	22.2	28.7
>20 cigarettes/day	1.7	4.9	6.7	10.4	23.8
Alcohol intake
Never	53.0	49.5	46.0	45.0	39.4	<0.001
Former	2.0	2.1	2.6	3.1	3.0
Current
<360 mL/day	35.0	37.5	37.2	36.3	37.8
≥360 mL/day	10.0	10.9	14.2	15.6	19.8
Physical activity	41.6	40.0	41.9	39.3	36.4	0.024
Hypertension	26.4	30.0	36.1	35.2	35.4	<0.001
Diabetes	3.5	4.6	4.8	5.4	7.1	<0.001
Chronic kidney disease	8.6	8.4	8.8	8.8	7.9	0.921
Non-HDL-C (mg/dL)
<130	12.7	15.8	16.3	18.3	18.9	<0.001
130–189	54.3	55.8	59.1	58.2	58.5
≥190	33.0	28.4	24.6	23.5	22.6
Cardiac murmur or valvular disease	2.3	2.9	2.2	2.2	2.0	0.522
Arrhythmia other than AF	2.5	1.8	3.1	2.6	2.6	0.320
History of MI	0.6	0.5	1.0	0.8	0.9	0.531
History of stroke	1.8	0.9	1.5	1.7	1.5	0.316

Unless indicated otherwise, data show the percentage of participants in each group. AF, atrial fibrillation; HDL-C, high-density lipoprotein cholesterol; MI, myocardial infarction; WBC, white blood cell.

Within 95,295 person-years (median follow-up 14.6 years), 312 participants developed AF. The distribution of WBC counts overall and according to the presence of AF are shown in Supplementary Figure 1. The incidence of AF, from the lowest (Q1) to highest (Q5) WBC quintiles, was 2.4, 2.5, 3.5, 3.5, and 4.4 per 1,000 person-years (Supplementary Figure 2).

In the model adjusted for age and sex, the highest WBC count quintile (Q5), compared with the lowest (Q1), was associated with an increased risk of AF (HR 1.65; 95% CI 1.15–2.38). Further adjustments for sociodemographic and clinical characteristics modestly attenuated the association, yet it remained significant (HR 1.57; 95% CI 1.07–2.29). The positive association between WBC count and AF risk was more significant among current smokers than non-smokers, with HRs of 4.66 (95% CI 1.89–11.50) and 1.61 (95% CI 1.01–2.57), respectively. In the multivariable-adjusted model, for each 1.0×10⁹-cells/L increment in WBC count, the risk of AF increased by 12% (10% in non-smokers, 15% in current smokers; Table 2).

Table 2. Association Between WBC Count and AF Risk in the Sex-Combined Cohort

	WBC count quintile					P trend	Increase in AF risk per 1.0×109-cells/L increment in WBC count (HR (95% CI))
	Q1	Q2	Q3	Q4	Q5
Overall
WBC count (×109/L)
Range	2.07–4.31	4.32–5.01	5.02–5.70	5.71–6.66	6.67–16.82	–
Median	3.86	4.68	5.34	6.13	7.50	–
Frequency	1,386	1,362	1,380	1,381	1,375	–
AF
New cases	47	47	70	67	81	–	–
Incidence*	2.4	2.5	3.5	3.5	4.4	–	–
HR (95% CI)
Model I	1	0.96 (0.64–1.44)	1.18 (0.82–1.72)	1.25 (0.86–1.82)	1.65 (1.15–2.38)	0.065	1.13 (1.06–1.21)
Model II	1	0.96 (0.64–1.44)	1.12 (0.77–1.63)	1.17 (0.80–1.72)	1.54 (1.06–2.26)	0.146	1.11 (1.04–1.20)
Model III	1	0.95 (0.63–1.43)	1.10 (0.76–1.62)	1.18 (0.80–1.73)	1.57 (1.07–2.29)	0.139	1.12 (1.04–1.20)
Model IV	0.90 (0.62–1.31)	0.86 (0.59–1.25)	1	1.06 (0.76–1.49)	1.41 (1.02–1.97)	–	–
Non-smokers
WBC count (×109/L)
Range	2.07–4.16	4.17–4.79	4.80–5.41	5.42–6.25	6.26–13.33	–	–
Median	3.70	4.49	5.11	5.79	6.93	–	–
Frequency	991	988	982	992	989	–	–
AF
New cases	28	44	45	41	52	–	–
Incidence*	2.0	3.2	3.2	2.9	3.9	–	–
HR (95% CI)
Model I	1	1.60 (0.99–2.57)	1.31 (0.82–2.11)	1.28 (0.79–2.08)	1.67 (1.05–2.64)	0.066	1.11 (1.01–1.23)
Model II	1	1.60 (1.00–2.58)	1.31 (0.81–2.11)	1.20 (0.73–1.95)	1.64 (1.03–2.61)	0.093	1.10 (1.00–1.22)
Model III	1	1.58 (0.98–2.55)	1.28 (0.79–2.06)	1.11 (0.68–1.82)	1.61 (1.01–2.57)	0.133	1.10 (0.99–1.22)
Model IV	0.78 (0.49–1.27)	1.24 (0.81–1.88)	1	0.87 (0.57–1.33)	1.26 (0.84–1.88)	–	–
Current smokers
WBC count (×109/L)
Range	2.63–4.92	4.93–5.72	5.73–6.53	6.54–7.59	7.60–16.82	–	–
Median	4.43	5.33	6.11	7.04	8.50	–	–
Frequency	388	390	387	388	389	–	–
AF
New cases	7	23	22	28	22	–	–
Incidence*	1.4	4.3	4.2	5.4	4.4	–	–
HR (95% CI)
Model I	1	2.67 (1.15–6.23)	2.83 (1.21–6.63)	3.76 (1.64–8.63)	3.43 (1.46–8.06)	0.003	1.12 (1.01–1.24)
Model II	1	2.69 (1.15–6.31)	3.04 (1.29–7.17)	3.78 (1.64–8.70)	3.54 (1.49–8.43)	0.002	1.13 (1.02–1.26)
Model III	1	3.51 (1.45–8.51)	3.92 (1.61–9.56)	5.31 (2.21–12.72)	4.66 (1.89–11.50)	<0.001	1.15 (1.03–1.27)
Model IV	0.26 (0.11–0.62)	0.90 (0.49–1.64)	1	1.35 (0.76–2.40)	1.19 (0.65–2.17)	–	–

*Per 1,000 person-years. Model I, was adjusted for age (continuous) and sex. Model II was further adjusted for body mass index (<18.5, 18.5–24.9, or ≥25 kg/m²), smoking (never, former, current ≤20 cigarettes/day, or current >20 cigarettes/day) in the overall sample only, alcohol (never, former, current <360 mL/day, or current ≥360 mL/day), physical activity (yes/no), hypertension (yes/no), diabetes (yes/no), chronic kidney disease (yes/no), and non-HDL-C (<130, 130–189, or ≥190 mg/dL). Model III was further adjusted for cardiac murmur or valvular disease (yes/no), arrhythmia other than AF (yes/no), history of MI (yes/no), and history of stroke (yes/no). Model IV was adjusted for the same variables as Model III, but Q3 was used as the reference group. The P interaction value for smoking was 0.197. CI, confidence interval; HR, hazard ratio. Other abbreviations as in Table 1.

The WBC count and AF incidence were significantly higher among men than women. However, the association between an increased WBC count and AF risk was more prominent among women than men (HR 2.16 [95% CI 1.10–4.26] vs. 1.55 [95% CI 0.99–2.44], respectively; P interaction=0.074). For each 1.0×10⁹-cells/L increment in WBC count, the AF risk increased by 9% in men and 20% in women (Tables 2,3). Among men who smoked, those in the highest quintile of WBC count (Q5) showed a remarkable rise in AF risk compared with those in the lowest quintile (Q1), whereas the same positive association was not significant among non-smokers. This finding can be explained by the unexpectedly low incidence of AF in the lowest WBC count quintile (reference group) among current smokers. However, each 1.0×10⁹-cells/L increment in WBC count was associated with an increase in AF risk by 9% in non-smokers and 10% in current smokers, suggesting no difference in the association between WBC count and AF risk in men according to smoking status (Table 3). Conversely, women who smoked experienced a higher increase in AF risk than non-smokers with each 1.0×10⁹-cells/L increment in WBC count (32% and 13%, respectively; Table 4). The sex-combined results remained consistent across age, BMI, and drinking categories (P interaction>0.20; Supplementary Table).

Table 3. Association Between WBC Count and AF Risk in Men

	WBC count quintile					P trend	Increase in AF risk per 1.0×109-cells/L increment in WBC count (HR (95% CI))
	Q1	Q2	Q3	Q4	Q5
Overall
WBC count (×109/L)
Range	2.10–4.62	4.63–5.34	5.35–6.07	6.08–7.08	7.09–16.82	–	–
Median	4.15	5.00	5.69	6.48	7.90	–	–
Frequency	649	648	642	650	649	–	–
AF
New cases	39	34	45	39	48	–	–
Incidence*	4.6	3.8	5.2	4.4	5.7	–	–
HR (95% CI)
Model I	1	0.83 (0.53–1.32)	1.15 (0.75–1.76)	1.02 (0.65–1.59)	1.51 (0.99–2.31)	0.323	1.09 (1.01–1.19)
Model II	1	0.86 (0.54–1.36)	1.13 (0.73–1.75)	1.04 (0.66–1.63)	1.54 (0.98–2.42)	0.311	1.09 (1.00–1.19)
Model III	1	0.87 (0.54–1.38)	1.08 (0.69–1.67)	1.05 (0.67–1.66)	1.55 (0.99–2.44)	0.314	1.09 (1.01–1.19)
Model IV	0.93 (0.60–1.44)	0.81 (0.51–1.26)	1	0.98 (0.64–1.51)	1.44 (0.94–2.20)	–	–
Non-smokers
WBC count (×109/L)
Range	2.10–4.32	4.33–5.02	5.03–5.61	5.62–6.39	6.40–12.75	–	–
Median	3.95	4.68	5.32	5.98	7.09	–	–
Frequency	341	339	340	341	340	–	–
AF
New cases	23	23	24	19	26	–	–
Incidence*	5.1	5.3	5.0	4.2	5.8	–	–
HR (95% CI)
Model I	1	0.99 (0.55–1.76)	0.90 (0.51–1.60)	0.80 (0.44–1.47)	1.09 (0.62–1.90)	0.815	1.09 (0.95–1.25)
Model II	1	1.02 (0.57–1.82)	0.88 (0.49–1.57)	0.81 (0.44–1.50)	1.14 (0.64–2.03)	0.880	1.10 (0.96–1.26)
Model III	1	1.03 (0.57–1.85)	0.82 (0.45–1.48)	0.75 (0.40–1.41)	1.12 (0.63–2.01)	0.760	1.09 (0.95–1.25)
Model IV	1.22 (0.68–2.22)	1.26 (0.70–2.26)	1	0.92 (0.50–1.70)	1.37 (0.78–2.42)	–	–
Current smokers
WBC count (×109/L)
Range	2.73–5.02	5.03–5.81	5.82–6.62	6.63–7.64	7.65–16.82	–	–
Median	4.57	5.42	6.20	7.13	8.58	–	–
Frequency	307	308	307	308	307	–	–
AF
New cases	9	22	18	23	18	–	–
Incidence*	2.3	5.2	4.3	5.6	4.4	–	–
HR (95% CI)
Model I	1	2.04 (0.94–4.43)	1.83 (0.82–4.07)	2.50 (1.16–5.40)	2.22 (0.99–4.95)	0.028	1.08 (0.96–1.21)
Model II	1	2.18 (1.00–4.76)	2.01 (0.89–4.53)	2.66 (1.22–5.78)	2.32 (1.02–5.26)	0.019	1.09 (0.97–1.22)
Model III	1	2.74 (1.21–6.22)	2.57 (1.09–6.05)	3.55 (1.54–8.18)	2.94 (1.24–6.98)	0.005	1.10 (0.98–1.24)
Model IV	0.39 (0.17–0.92)	1.07 (0.56–2.02)	1	1.38 (0.74–2.60)	1.14 (0.59–2.22)	–	–

*Per 1,000 person-years. Model I, was adjusted for age (continuous). Model II was further adjusted for body mass index (<18.5, 18.5–24.9, or ≥25 kg/m²), smoking (never, former, current ≤20 cigarettes/day, or current >20 cigarettes/day) in the overall sample only, alcohol (never, former, current <360 mL/day, or current ≥360 mL/day), physical activity (yes/no), hypertension (yes/no), diabetes (yes/no), chronic kidney disease (yes/no), and non-HDL-C (<130, 130–189, or ≥190 mg/dL). Model III was further adjusted for cardiac murmur or valvular disease (yes/no), arrhythmia other than AF (yes/no), history of MI (yes/no), and history of stroke (yes/no). Model IV was adjusted for the same variables as Model III, but Q3 was used as the reference group. Abbreviations as in Tables 1,2.

Table 4. Association Between WBC Count and AF Risk in Women

	WBC count quintile					P trend	Increase in AF risk per 1.0×109-cells/L increment in WBC count (HR (95% CI))
	Q1	Q2	Q3	Q4	Q5
Overall
WBC count (×109/L)
Range	2.07–4.11	4.12–4.74	4.75–5.39	5.40–6.30	6.31–13.33	–	–
Median	3.66	4.44	5.08	5.80	7.00	–	–
Frequency	731	722	733	735	725	–	–
AF
New cases	13	20	24	21	29	–	–
Incidence*	1.2	2.0	2.2	2.0	2.9	–	–
HR (95% CI)
Model I	1	1.62 (0.80–3.25)	1.62 (0.82–3.18)	1.67 (0.84–3.34)	2.42 (1.26–4.67)	0.029	1.20 (1.06–1.36)
Model II	1	1.48 (0.73–2.99)	1.45 (0.73–2.87)	1.43 (0.71–2.89)	2.04 (1.04–3.99)	0.096	1.16 (1.02–1.33)
Model III	1	1.66 (0.81–3.38)	1.53 (0.77–3.15)	1.55 (0.77–3.15)	2.16 (1.10–4.26)	0.061	1.20 (1.06–1.36)
Model IV	0.65 (0.33–1.30)	1.08 (0.59–1.99)	1	1.01 (0.56–1.85)	1.41 (0.81–2.48)	–	–
Non-smokers
WBC count (×109/L)
Range	2.07–4.06	4.07–4.69	4.70–5.30	5.31–6.18	6.19–13.33	–	–
Median	3.61	4.40	5.00	5.70	6.85	–	–
Frequency	647	649	657	641	647	–	–
AF
New cases	13	17	21	22	22	–	–
Incidence*	1.4	1.8	2.2	2.3	2.5	–	–
HR (95% CI)
Model I	1	1.38 (0.67–2.85)	1.47 (0.73–2.93)	1.73 (0.87–3.44)	1.79 (0.90–3.55)	0.094	1.14 (0.98–1.31)
Model II	1	1.35 (0.65–2.79)	1.38 (0.69–2.77)	1.53 (0.77–3.05)	1.60 (0.80–3.21)	0.180	1.12 (0.96–1.30)
Model III	1	1.59 (0.76–3.33)	1.49 (0.74–3.02)	1.69 (0.84–3.41)	1.76 (0.87–3.57)	0.101	1.13 (0.97–1.31)
Model IV	0.67 (0.33–1.36)	1.07 (0.56–2.05)	1	1.13 (0.62–2.08)	1.18 (0.64–2.17)	–	–
Current smokers
WBC count (×109/L)
Range	2.63–4.52	4.58–5.36	5.37–6.14	6.15–7.16	7.18–12.10	–	–
Median	4.13	5.00	5.80	6.63	8.24	–	–
Frequency	81	81	81	83	79	–	–
AF
New cases	0	2	2	4	4	–	–
Incidence*	0.0	1.9	1.8	3.7	4.3	–	–
HR (95% CI)
Model I**	–	–	–	–	–	–	1.37 (1.03–1.84)
Model II**	–	–	–	–	–	–	1.33 (0.99–1.81)
Model III**	–	–	–	–	–	–	1.32 (0.85–2.04)

*Per 1,000 person-years. **The associations were not conducted by quintiles because of the limited number of cases of incident AF. Model I, was adjusted for age (continuous). Model II was further adjusted for body mass index (<18.5, 18.5–24.9, or ≥25 kg/m²), smoking (never, former, current ≤20 cigarettes/day, or current >20 cigarettes/day) in the overall sample only, alcohol (never, former, current <360 mL/day, or current ≥360 mL/day), physical activity (yes/no), hypertension (yes/no), diabetes (yes/no), chronic kidney disease (yes/no), and non-HDL-C (<130, 130–189, or ≥190 mg/dL). Model III was further adjusted for cardiac murmur or valvular disease (yes/no), arrhythmia other than AF (yes/no), history of MI (yes/no), and history of stroke (yes/no). Model IV was adjusted for the same variables as Model III, but Q3 was used as the reference group. Abbreviations as in Tables 1,2.

After excluding participants with WBC counts <3.1 or >8.4×10⁹/L, the sensitivity analysis revealed that each 1.0×10⁹-cells/L increment in WBC count was associated with an 11% increase in AF risk. However, the association among women was attenuated (Table 5). The increasing HR for AF with WBC count is shown using spline curves for the overall and sensitivity samples in the Figure.

Table 5. Association Between WBC Count and AF Risk After Excluding Participants With WBC Counts <3.1 or >8.4×10⁹/L

	WBC count quintile					P trend	Increase in AF risk per 1.0×109-cells/L increment in WBC count (HR (95% CI))
	Q1	Q2	Q3	Q4	Q5
Overall	1	0.89 (0.59–1.34)	1.05 (0.72–1.53)	1.04 (0.70–1.53)	1.30 (0.88–1.91)	0.496	1.11 (1.00–1.24)
	0.95 (0.65–1.39)	0.85 (0.58–1.24)	1	0.99 (0.70–1.40)	1.24 (0.88–1.75)	–	–
Men	1	0.73 (0.44–1.19)	1.06 (0.68–1.65)	0.94 (0.58–1.51)	1.33 (0.83–2.11)	0.708	1.13 (0.99–1.28)
	0.95 (0.61–1.48)	0.69 (0.43–1.11)	1	0.89 (0.57–1.39)	1.26 (0.82–1.93)	–	–
Women	1	1.37 (0.69–2.74)	1.31 (0.67–2.55)	1.30 (0.65–2.59)	1.54 (0.78–3.03)	0.253	1.09 (0.91–1.31)
	0.77 (0.39–1.49)	1.05 (0.56–1.95)	1	0.99 (0.54–1.84)	1.18 (0.65–2.15)	–	–
Non-smokers	1	1.48 (0.92–2.39)	1.16 (0.72–1.87)	1.05 (0.64–1.71)	1.33 (0.82–2.14)	0.349	1.03 (0.90–1.18)
	0.86 (0.54–1.38)	1.28 (0.83–1.96)	1	0.90 (0.58–1.39)	1.14 (0.75–1.74)	–	–
Current smokers	1	2.70 (1.08–6.74)	2.72 (1.11–6.63)	3.78 (1.56–9.19)	4.28 (1.75–10.49)	0.002	1.34 (1.11–1.62)
	0.37 (0.15–0.90)	0.99 (0.51–1.92)	1	1.39 (0.75–2.58)	1.58 (0.84–2.96)	–	–

Data are shown as HRs with 95% CIs in parentheses . HRs and 95% CIs were adjusted for age (continuous), sex, body mass index (<18.5, 18.5–24.9, or ≥25 kg/m²), smoking (never, former, current ≤20 cigarettes/day, or current >20 cigarettes/day), alcohol (never, former, current <360 mL/day, or current ≥360 mL/day), physical activity (yes/no), hypertension (yes/no), diabetes (yes/no), chronic kidney disease (yes/no), non-HDL-C (<130, 130–189, or ≥190 mg/dL), cardiac murmur or valvular disease (yes/no), arrhythmia other than AF (yes/no), history of MI (yes/no), and history of stroke (yes/no). Abbreviations as in Tables 1,2.

Figure.

Spline curves showing the association between the white blood cell (WBC) count and risk of atrial fibrillation (AF). (Left) Overall sample. The reference WBC count is 5.34×10⁹/L (median) and the knot points are 3.48, 4.50, 5.34, 6.37, and 8.29×10⁹/L. (Right) Sensitivity sample. The reference WBC count is 5.30×10⁹/L (median) and the knot points are 3.93, 4.50, 5.30, 6.23, and 7.12×10⁹/L.

Discussion

This prospective cohort study indicated that an increased WBC count was positively associated with an elevated risk of AF among the general population in Japan, and that the association was stronger in women and current smokers than in men and non-smokers.

These results from the Suita Study are comparable to those from previous studies conducted in Western populations. The Framingham Heart Study investigated 936 participants and reported a positive association between WBC count and increased incident AF, with an HR per 1-SD increase in WBC count of 2.16 (95% CI 1.07–4.35).¹⁸ In the Tromsø Study, which included 6,315 participants, the highest vs. lowest WBC count quartile was associated with an increased risk of AF (HR 1.44; 95% CI 1.11–1.88), and every 1-SD increase in WBC count led to a 9% increase in AF risk.¹⁹ In the ARIC Study, which included 14,500 participants, a higher total WBC count was associated with increased AF risk independent of AF risk factors (HR 1.09 [95% CI 1.04–1.15] per 1-SD increase).²⁰ However, this association faded after adjusting for incident CVD.²⁰

The mechanisms by which inflammation can increase AF risk remain obscure. However, it has been suggested that the increased inflammatory biomarkers in AF could signify atrial inflammation, which contributes to electrical and structural atrial remodeling, which then initiates AF.^9–11 In addition, inflammation can alter calcium homeostasis and trigger heterogeneous atrial conduction, factors that could increase AF susceptibility.⁹ Thus, drugs with anti-inflammatory properties, such as steroids, angiotensin-converting enzyme inhibitors, and colchicine, have been investigated as potential remedies for AF and as chemoprophylactics against postoperative AF and AF recurrence.¹⁰

Of note, the association between WBC count and AF risk was more pronounced among women than men. A similar sex difference was shown in a study of 6,756 Japanese residents from the National Integrated Project for Prospective Observation of Non-communicable Disease and Its Trends in the Aged (NIPPON DATA) 90, in which the WBC count was associated with CVD mortality in women but not men.¹⁵ Conversely, the Tromsø Study showed no differences in the association between WBC count and AF risk in the sex-specified analysis.¹⁹ Neither the Framingham Heart Study nor ARIC Study stratified their results by sex.^18,20

We also noted that the association between WBC count and AF risk was greater among current smokers than non-smokers, and this difference was evident in women. In line with our findings, previous reports from the ARIC Study, the Circulatory Risk in Communities Study (CIRCS), and the Malmö Diet and Cancer Study (MDCS) showed that the association between WBC count and the risk of ischemic stroke, a common complication of AF, was more pronounced among current than former and never smokers.^12–14 Smoking is a major risk factor for CVD^27,28 and AF,^29,30 and smokers have a substantially higher WBC count than non-smokers.^31,32 Avoiding smoking is a substantial factor in the Lifelong Health Support 10 (LHS10) scheme proposed by the preventive medicine specialists at the NCVC, based on Japanese epidemiological evidence, to prevent CVD, including AF.³³

Notably, the spline curves (Figure) showed that the linear association between WBC count and AF risk was significantly more prominent among those with WBC levels <3.1 or >8.4×10⁹/L. The Japan Society of Ningen Dock suggested a WBC count between 3.1 and 8.4×10⁹/L as a reference range.³⁴ Within this reference range, a 1.0×10⁹-cells/L increment in the WBC count was associated with an 11% increase in AF risk. Based on the spline curves, it could be assumed that the AF risk may have increased further with the WBC count when the WBC count increased over the reference range or decreased further when the WBC count decreased below the reference range. Still, we could not confirm this assumption because those with WBC counts outside the reference range represented only 6.3% of the whole sample. In addition, this assumption could partly explain why the association between WBC count and AF risk was attenuated among women, who tend to have low WBC counts, when the analysis was confined to those within the reference range.

Despite being beyond the aim of this study, we noticed that the WBC count in the Suita Study was relatively lower than that in previous Western studies. For example, the WBC count quartiles in the Framingham Study (White people) were <5.6, 5.6–6.4, 6.5–7.8, and >7.8×10⁹/L,¹⁸ the quartiles in the Tromsø Study (White people) were 2.2–5.5, 5.6–6.6, 6.7–7.9, and 8.0–34.0×10⁹/L,¹⁹ and the quintiles in the ARIC Study (White people and Black people, who have a relatively lower WBC count) were 3.0–4.6, 4.7–5.4, 5.5–6.2, 6.3–7.4, and 7.5–12.0×10⁹/L.²⁰ The corresponding WBC count quintiles in the Suita Study (Asian people) were 2.07–4.31, 4.32–5.01, 5.02–5.70, 5.71–6.66, and 6.67–16.82×10⁹/L. Such variations may indicate the need to modify the WBC count reference according to ethnicity and highlight the importance of performing more studies to explore the predictive ability of the WBC count in determining other health outcomes among the Asian population.

This study has several strengths: it is one of a few studies, and the first Asian study, to investigate the prospective association between WBC count and AF risk, investigating a randomly selected group representing the urban Japanese population; it used a prospective design, with frequent follow-up visits, to follow participants for an extended period; it ascertained AF diagnosis using medical records to include cases of paroxysmal AF that could have been missed in medical checkups; and it used standard approaches to assess the WBC count and covariates, adjusting the results for most potential confounders.

However, some limitations should also be considered. First, the limited number of incident AF cases may have reduced the statistical power in the stratified analyses. This limited number of cases could be attributed to the relatively young age of our participants and the fact that Asians have lower incidence rates of AF than their White counterparts.³⁵ Second, we did not adjust our results in the primary analysis for high-sensitivity C-reactive protein (hs-CRP), another important inflammatory marker with a potential association with AF,^36,37 because baseline hs-CRP data were available for only 1,003 participants. However, when we confined the analysis to this group, the results before and after hs-CRP adjustment were almost similar (data not shown). Similarly, adjusting for hs-CRP in the Tromsø Study did not materially change the results.¹⁹ Third, we had no data about the differential WBC count. A cross-sectional study from the Fukushima Health Management Survey showed an increased prevalence of AF alongside an increased monocyte count and increased neutrophil/lymphocyte ratio, but not alongside increased neutrophil, lymphocyte, or eosinophil counts.³⁸ The ARIC Study described increased AF risk associated with higher neutrophil counts and a higher neutrophil/lymphocyte ratio, but not higher monocyte, lymphocyte, or eosinophil counts.²⁰

In conclusion, our results provide additional evidence of the inflammatory pathogenesis of AF. The association between the WBC count and AF could be stronger among women and current smokers than among men and non-smokers. The association between the WBC count and AF risk was also seen among people within the reference range of the WBC count (3.1–8.4×10⁹/L). The WBC count could be a valuable biomarker for detecting people at higher risk of AF in the general population. From a preventive perspective, the WBC count should be considered in future AF risk scores and preventive interventions, especially in women who smoke.

Acknowledgments

The authors thank the President of the Suita Medical Association, the members of the Suita City Health Center, the staff of the Preventive Cardiology and Preventive Healthcare departments, and all cohort members.

Sources of Funding

This study was supported by Grants-in-Aid for Scientific Research in Japan (B, No. 16H05252), the Intramural Research Fund for the Cardiovascular Diseases of the National Cerebral and Cardiovascular Center (20-4-9), the Japan Health Research Promotion Bureau (2019-(1)-1), the Japan Science and Technology Agency (JPMJPF2018), the Health and Labour Sciences Research Grants of the Ministry of Health, Labour and Welfare of Japan (20FA1002), and the Meiji Yasuda Research Institute and Life Insurance Company.

Disclosures

None.

IRB Information

The Suita Study protocol was approved by the Institutional Review Board of the National Cerebral and Cardiovascular Center (M25-043-4).

Data Availability

Data will be made available upon reasonable request.

Supplementary Files

Please find supplementary file(s);

http://dx.doi.org/10.1253/circj.CJ-22-0378

References

• 1.   Lippi G, Sanchis-Gomar F, Cervellin G. Global epidemiology of atrial fibrillation: An increasing epidemic and public health challenge. Int J Stroke 2021; 16: 217–221.
• 2.   Dai H, Zhang Q, Much AA, Maor E, Segev A, Beinart R, et al. Global, regional, and national prevalence, incidence, mortality, and risk factors for atrial fibrillation, 1990–2017: Results from the Global Burden of Disease Study 2017. Eur Heart J Qual Care Clin Outcomes 2021; 7: 574–582.
• 3.   Johnsen SP, Dalby LW, Täckström T, Olsen J, Fraschke A. Cost of illness of atrial fibrillation: A nationwide study of societal impact. BMC Health Serv Res 2017; 17: 714.
• 4.   Odutayo A, Wong CX, Hsiao AJ, Hopewell S, Altman DG, Emdin CA. Atrial fibrillation and risks of cardiovascular disease, renal disease, and death: Systematic review and meta-analysis. BMJ 2016; 354: i4482.
• 5.   Emdin CA, Wong CX, Hsiao AJ, Altman DG, Peters SA, Woodward M, et al. Atrial fibrillation as risk factor for cardiovascular disease and death in women compared with men: Systematic review and meta-analysis of cohort studies. BMJ 2016; 532: h7013.
• 6.   Bunch TJ. Atrial fibrillation and dementia. Circulation 2020; 142: 618–620.
• 7.   Kokubo Y, Matsumoto C. Traditional cardiovascular risk factors for incident atrial fibrillation. Circ J 2016; 80: 2415–2422.
• 8.   Ono K, Iwasaki YK, Akao M, Ikeda T, Ishii K, Inden Y, et al. JCS/JHRS 2020 guideline on pharmacotherapy of cardiac arrhythmias. Circ J 2022; 86: 1790–1924, doi:10.1253/circj.CJ-20-1212.
• 9.   Hu YF, Chen YJ, Lin YJ, Chen SA. Inflammation and the pathogenesis of atrial fibrillation. Nat Rev Cardiol 2015; 12: 230–243.
• 10.   Ihara K, Sasano T. Role of inflammation in the pathogenesis of atrial fibrillation. Front Physiol 2022; 13: 862164.
• 11.   Zhou X, Dudley SC Jr. Evidence for inflammation as a driver of atrial fibrillation. Front Cardiovasc Med 2020; 7: 62.
• 12.   Li J, Imano H, Yamagishi K, Tanaka M, Cui R, Muraki I, et al. Leukocyte count and risks of stroke and coronary heart disease: The Circulatory Risk in Communities Study (CIRCS). J Atheroscler Thromb 2022; 29: 527–535.
• 13.   Lee CD, Folsom AR, Nieto FJ, Chambless LE, Shahar E, Wolfe DA. White blood cell count and incidence of coronary heart disease and ischemic stroke and mortality from cardiovascular disease in African-American and White men and women: Atherosclerosis Risk In Communities Study. Am J Epidemiol 2001; 154: 758–764.
• 14.   Zia E, Melander O, Björkbacka H, Hedblad B, Engström G. Total and differential leucocyte counts in relation to incidence of stroke subtypes and mortality: A prospective cohort study. J Intern Med 2012; 272: 298–304.
• 15.   Tamakoshi K, Toyoshima H, Yatsuya H, Matsushita K, Okamura T, Hayakawa T, et al. White blood cell count and risk of all-cause and cardiovascular mortality in nationwide sample of Japanese: Results from the NIPPON DATA90. Circ J 2007; 71: 479–485.
• 16.   Hu ZB, Lu ZX, Zhu F, Jiang CQ, Zhang WS, Pan J, et al. Higher total white blood cell and neutrophil counts are associated with an increased risk of fatal stroke occurrence: The Guangzhou biobank cohort study. BMC Neurol 2021; 21: 470.
• 17.   Brown DW, Giles WH, Croft JB. White blood cell count: An independent predictor of coronary heart disease mortality among a national cohort. J Clin Epidemiol 2001; 54: 316–322.
• 18.   Rienstra M, Sun JX, Magnani JW, Sinner MF, Lubitz SA, Sullivan LM, et al. White blood cell count and risk of incident atrial fibrillation (from the Framingham Heart Study). Am J Cardiol 2012; 109: 533–537.
• 19.   Nyrnes A, Njølstad I, Mathiesen EB, Wilsgaard T, Hansen JB, Skjelbakken T, et al. Inflammatory biomarkers as risk factors for future atrial fibrillation. An eleven-year follow-up of 6315 men and women: The Tromsø study. Gend Med 2012; 9: 536–547.
• 20.   Misialek JR, Bekwelem W, Chen LY, Loehr LR, Agarwal SK, Soliman EZ, et al. Association of white blood cell count and differential with the incidence of atrial fibrillation: The Atherosclerosis Risk in Communities (ARIC) Study. PLoS One 2015; 10: e0136219.
• 21.   Arafa A, Kokubo Y, Sheerah HA, Sakai Y, Watanabe E, Li J, et al. Weight change since age 20 and the risk of cardiovascular disease mortality: A prospective cohort study. J Atheroscler Thromb 2022; 29: 1511–1521, doi:10.5551/jat.63191.
• 22.   Arafa A, Kokubo Y, Sheerah HA, Sakai Y, Watanabe E, Li J, et al. Developing a stroke risk prediction model using cardiovascular risk factors: The Suita Study. Cerebrovasc Dis 2022; 51: 323–330.
• 23.   Kokubo Y, Watanabe M, Higashiyama A, Nakao YM, Kusano K, Miyamoto Y. Development of a basic risk score for incident atrial fibrillation in a Japanese general population: The Suita Study. Circ J 2017; 81: 1580–1588.
• 24.   Arafa A, Kokubo Y, Shimamoto K, Kashima R, Watanabe E, Sakai Y, et al. Stair climbing and incident atrial fibrillation: A prospective cohort study. Environ Health Prev Med 2022; 27: 10.
• 25.   Arafa A, Kokubo Y, Shimamoto K, Kashima R, Watanabe E, Sakai Y, et al. Sleep duration and atrial fibrillation risk in the context of predictive, preventive, and personalized medicine: The Suita Study and meta-analysis of prospective cohort studies. EPMA J 2022; 13: 77–86.
• 26.   Arafa A, Kokubo Y, Kobayashi T, Honda-Kohmo K, Kashima R, Sheerah HA, et al. QRS transitional zone rotations and the risk of atrial fibrillation: The Suita Study. Heart Vessels 2022; 37: 1914–1920, doi:10.1007/s00380-022-02101-3.
• 27.   Arafa A, Lee HH, Eshak ES, Shirai K, Liu K, Li J, et al. Modifiable risk factors for cardiovascular disease in Korea and Japan. Korean Circ J 2021; 51: 643–655.
• 28.   Kokubo Y, Padmanabhan S, Iwashima Y, Yamagishi K, Goto A. Gene and environmental interactions according to the components of lifestyle modifications in hypertension guidelines. Environ Health Prev Med 2019; 24: 19.
• 29.   Aune D, Schlesinger S, Norat T, Riboli E. Tobacco smoking and the risk of atrial fibrillation: A systematic review and meta-analysis of prospective studies. Eur J Prev Cardiol 2018; 25: 1437–1451.
• 30.   Wang Q, Guo Y, Wu C, Yin L, Li W, Shen H, et al. Smoking as a risk factor for the occurrence of atrial fibrillation in men versus women: A meta-analysis of prospective cohort studies. Heart Lung Circ 2018; 27: 58–65.
• 31.   Smith MR, Kinmonth AL, Luben RN, Bingham S, Day NE, Wareham NJ, et al. Smoking status and differential white cell count in men and women in the EPIC-Norfolk population. Atherosclerosis 2003; 169: 331–337.
• 32.   Fröhlich M, Sund M, Löwel H, Imhof A, Hoffmeister A, Koenig W. Independent association of various smoking characteristics with markers of systemic inflammation in men: Results from a representative sample of the general population (MONICA Augsburg Survey 1994/95). Eur Heart J 2003; 24: 1365–1372.
• 33.   Arafa A, Kokubo Y, Kashima R, Teramoto M, Sakai Y, Nosaka S, et al. The Lifelong Health Support 10: A Japanese prescription for a long and healthy life. Environ Health Prev Med 2022; 27: 23.
• 34.   Japan Society of Ningen Dock. Blood test: White blood cells [in Japanese]. https://www.ningen-dock.jp/public/inspection/blood (accessed September 7, 2022).
• 35.   O’Neal WT, Judd SE, Limdi NA, McIntyre WF, Kleindorfer DO, Cushman M, et al. Differential impact of risk factors in blacks and whites in the development of atrial fibrillation: The Reasons for Geographic And Racial Differences in Stroke (REGARDS) Study. J Racial Ethn Health Disparities 2017; 4: 718–724.
• 36.   Yoshikawa T, Hata J, Sakata S, Nagata T, Hirakawa Y, Hirooka Y, et al. Serum high-sensitivity C-reactive protein levels and the development of atrial fibrillation in a general Japanese population: The Hisayama Study. Circ J 2021; 85: 1365–1372.
• 37.   Marott SC, Nordestgaard BG, Zacho J, Friberg J, Jensen GB, Tybjaerg-Hansen A, et al. Does elevated C-reactive protein increase atrial fibrillation risk?: A Mendelian randomization of 47,000 individuals from the general population. J Am Coll Cardiol 2010; 56: 789–795.
• 38.   Suzuki H, Ohira T, Takeishi Y, Sakai A, Hosoya M, Yasumura S, et al. Association between atrial fibrillation and white blood cell count after the Great East Japan Earthquake: An observational study from the Fukushima Health Management Survey. Medicine (Baltimore) 2021; 100: e24177.