Supplementary Table 1. A comparison of the basic characteristics between the included participants and participants who were excluded due to missing and extreme values of the total leukocyte count.

Characteristics	Included	Excluded
N	26655	5096
Age (years)	61.5±8.1	62.6±7.7
Men (%)	11762 (44.1)	2216 (43.5)
BMI (kg/m2)§	24.2±3.3	24.5±3.4
Education (%)§
Primary school or below	5857 (22.1)	1084 (21.4)
Middle school	9670 (36.5)	1773 (35.0)
High school or beyond	10949 (41.4)	2202 (43.5)
Smoking status (%)§
Current smokers	5102 (19.2)	931 (18.5)
Former smokers	2597 (9.8)	480 (9.5)
Never smokers	18866 (71.0)	3629 (72.0)
Drinking status (%)§
Current drinkers	6607 (24.8)	1088 (21.4)
Former drinkers	1195 (4.5)	232 (4.6)
Never drinkers	18812 (70.7)	3764 (74.0)
Physical activity (%)	18939 (71.1)	3344 (65.6)
Antibiotics use (%)	2052 (7.7)	427 (8.4)
Aspirin use (%)	2581 (9.7)	459 (9.0)
Family history of CVD (%)	2849 (10.7)	509 (10.0)
Hyperlipidemia (%)	10996 (41.3)	1433 (28.1)
Hypertension (%)	13475 (50.6)	2409 (47.3)
Diabetes mellitus (%)	4455 (16.7)	562 (11.0)

Abbreviation: BMI, body mass index; CVD, cardiovascular disease.

Continuous variables were described as mea±SD if normally distributed, or median (interquartile range) if skewed distributed.

Categorical variables were described as number (%).

^§Data were incomplete for these variables. For totally 26655 participants included in the analysis, 357 (1.3%), 179 (0.7%), 90 (0.3%) and 41 (0.2%) of participants had missing data for BMI, education, smoking status and drinking status, respectively. The other variables included in the analyses did not have missing data. For the 5096 participants who were excluded due to missing and extreme values of total leukocyte count, 1585 (31.1%), 37 (0.7%), 56 (1.1%), and12 (2.4%) of participants had missing data for BMI, education, smoking status and drinking status, respectively.