Association of soy food intake with risk and biomarkers of coronary heart disease in Chinese men

Soy food intake has been associated with a reduced risk of coronary heart disease (CHD) in women, but its association with CHD in men is unclear [1–3]. Several studies have found sex differences in the associations of soy intake with metabolic syndrome, endothelial function, and other cardiovascular risk factors [4,5]. In the Shanghai Men’s Health Study [6], we investigated the association between soy food intake and incident CHD among 55,474 Chinese men (40–74 years) who were free of CHD, stroke, and cancer at baseline (2002–2006). We also examined the associations of soy intake with multiple CHD biomarkers in a subsample of 3,885 men who provided fasting blood samples and had no history of major chronic diseases at recruitment.

Usual dietary intakes were assessed through in-person interviews using a validated food-frequency questionnaire [6]. Energy and nutrient intakes, including soy protein and isoflavones, were calculated based on 2002 Chinese Food Composition Table. Participants who reported extreme energy intake (<800 or >4,200 kcal/day) were excluded from analysis. Dietary intakes were adjusted for total energy by using the residual method. Incident CHD cases (nonfatal myocardial infarction and fatal CHD) were ascertained via follow-up home visits (follow-up rate: 94%) and linkage to the Shanghai Vital Statistics Registry (>99% complete). The diagnosis of cases was further confirmed by reviewing medical records and/or death certificates. Hazard ratios of CHD and 95% confidence intervals were estimated using Cox regression with age as the timescale. Measurements of plasma biomarkers have been described previously and shown reasonable intra-assay and intra-person coefficients of variation [7]. Multivariable-adjusted geometric means of biomarkers by soy protein intake were calculated and compared using linear regression. Written informed consent was obtained from each participant. The study protocol conforms to the ethical guidelines of the 1975 Declaration of Helsinki as reflected in a priori approval by the institutional review boards of all participating institutes.

The median soy protein intake was 9.5 g/day (interquartile range: 6.3–13.7 g/day). During a mean follow-up of 5.4 years, we identified and validated 217 incident CHD cases. Soy intake showed a positive, independent, and dose-response association with risk of CHD. After adjustment for several CHD risk factors and other dietary intakes (Table), the hazard ratios were 1 (referent), 1.23, 1.47, and 1.58 (95% confidence interval: 1.06–2.36) from the lowest to highest quartile of soy protein intake (P for trend =0.02). Restricted cubic spline analysis revealed that CHD risk increased gradually with increasing soy protein intake up to ~15 g/day (80^th percentile); the increase in risk flattened off thereafter (Figure). Positive associations persisted in analyses confined to men without diabetes and family history of cardiovascular diseases and men had no or little change in soy intake in last 5 years. No effect modifications were found for obesity, physical activity, smoking, alcohol consumption, hypertension, soy intake during adolescence, and other dietary factors. Results did not change when the first year of follow-up was excluded. Similar positive associations were observed for soy isoflavones and total soy food consumption.

Table.

Hazard ratio (95% confidence interval) of coronary heart disease by soy food intake in the Shanghai Men’s Health Study (n= 55,474)

	Quartiles of soy food intake				P for trend
	1	2	3	4
Soy protein
Median, g/day	4.3	7.9	11.3	17.3
No. of cases	42	48	60	67
Age- and energy- adjusted model	1.00	1.15 (0.76, 1.74)	1.37 (0.92, 2.03)	1.48 (1.01, 2.18)	0.03
Model 1	1.00	1.22 (0.81, 1.85)	1.45 (0.98, 2.16)	1.55 (1.05, 2.30)	0.02
Model 2	1.00	1.23 (0.81, 1.87)	1.47 (0.99, 2.20)	1.58 (1.06, 2.36)	0.02
Isoflavones
Median, mg/day	13.5	25.5	37.6	58.3
No. of cases	44	43	66	64
Age- and energy- adjusted model	1.00	0.99 (0.65, 1.50)	1.46 (1.00, 2.14)	1.34 (0.92, 1.97)	0.06
Model 1	1.00	1.04 (0.68, 1.59)	1.57 (1.07, 2.30)	1.40 (0.95, 2.07)	0.04
Model 2	1.00	1.04 (0.68, 1.59)	1.58 (1.07, 2.33)	1.42 (0.96, 2.11)	0.04
Soy foods (dry weight)
Median, g/day	11.3	19.6	27.5	41.1
No. of cases	48	41	58	70
Age- and energy- adjusted model	1.00	0.84 (0.55, 1.28)	1.13 (0.77, 1.66)	1.31 (0.91, 1.90)	0.05
Model 1	1.00	0.90 (0.59, 1.36)	1.23 (0.84, 1.81)	1.40 (0.96, 2.03)	0.03
Model 2	1.00	0.91 (0.60, 1.38)	1.25 (0.84, 1.84)	1.43 (0.98, 2.10)	0.02

Model 1: adjusted for age, income, education, physical activity, smoking, alcohol consumption, body mass index, waist-to-hip ratio, use of vitamin E, multivitamin supplement, aspirin, family history of cardiovascular disease, history of hypertension or diabetes, and total energy intake. Model 2: further adjusted for dietary intakes of fruit, fish, red meat, saturated fat and sodium.

Figure.

Risk of coronary heart disease by soy protein intake in the Shanghai Men’s Health Study (— hazard ratio, ---- 95% confidence interval)^a

^a Four knots: 5^th, 35^th, 65^th and 95^th percentiles. Referent: median intake 9.5 g/day. Model was adjusted for all variables listed in the footnote of Table Model 2.

In cross-sectional analysis of biomarkers, soy intake was significantly associated with higher plasma concentrations of interleukin-8 (IL-8) and plasminogen activator inhibitor-1 (PAI-1). In the lowest and highest quartiles of soy protein intake, the multivariable-adjusted geometric means were 1.4 and 1.6 pg/ml for IL-8 and 17.6 and 18.9 ng/ml for PAI-1 (both P for trend <0.01). We did not find significant associations of soy intake with other biomarkers, including blood lipids, glucose, insulin, C-reactive protein, and adipokines.

Few epidemiological studies, mainly from Asian populations, have investigated the relationship between soy food intake and incident CHD. A potential cardioprotective effect of soy foods has been reported in Chinese and Japanese women [1–3]. To our knowledge, our study is the first to focus on soy foods and risk of CHD in men and provide the evidence that habitual high soy intake may have adverse effects on the development of CHD in men.

The mechanisms behind the possible adverse effects of high soy intake on cardiometabolic health in men are largely unknown. Elevated IL-8 and PAI-1 are markers of impaired vascular endothelium and hypercoagulable states. The positive associations of soy intake with these biomarkers observed in our study might provide some explanations. Previous studies have suggested sex-specific associations of soy foods with cardiometabolic health, with favorable associations observed in women and unfavorable associations in men [4,5]. Clinical trials to date overall found no influence of soy protein or isoflavones on endothelial function [8]. Some studies, nevertheless, did observe male-specific adverse effects of soy-supplemented diet on coagulation and fibrinolysis [5,9]. It is also possible that other mechanisms, e.g. sex hormonal pathways, or other components of soy products or metabolites, e.g. trimethylamine-N-oxide, a metabolite from soy lecithin by gut microbiota [10], may be responsible for the observed adverse associations in men.

Major concerns of our study include potential dietary measurement errors and residual confounding. However, because of the prospective design, exposure misclassification is likely to be non-differential and result in an underestimation of the association. Although we cannot rule out the presence of residual confounding, we have adjusted for a range of potential confounders and found similar associations with and without multivariable adjustment.

In conclusion, habitual high soy food intake may be associated with increased risk of incident CHD in middle-aged and older Chinese men; elevated plasma IL-8 and PAI-1 might be potential contributing factors.