Abstract
Purpose
Taurine (2-aminoethanesulfonic acid), a molecule obtained from diet, is involved in bile acid conjugation, blood pressure regulation, anti-oxidation and anti-inflammation. We performed the first prospective study of taurine and CHD risk.
Methods
We conducted a case-control study nested in the New York University Women’s Health Study to evaluate the association between circulating taurine levels and risk of coronary heart disease (CHD). Taurine was measured in two yearly pre-diagnostic serum samples of 223 CHD cases and 223 matched controls and averaged for a more reliable measurement of long-term taurine levels.
Results
Mean serum taurine was positively related to age and dietary intake of poultry, niacin, vitamin B1, fiber, and iron, and negatively related to dietary intake of saturated fat (all p values ≤ 0.05). There was no statistically significant association between the risk of CHD and serum taurine levels. The adjusted ORs for CHD in increasing taurine tertiles were 1.0 (reference), 0.85 (95% CI, 0.51–1.40), and 0.66 (0.39–1.13; p for trend = 0.14). There was a significant inverse association between serum taurine and CHD risk among women with high total serum cholesterol (>250 mg/dl) (adjusted OR = 0.39 (0.19–0.83) for the third vs. first tertile; p for trend = 0.02) but not among those with low total serum cholesterol (p for interaction = 0.01). The data suggest a possible inverse association of serum taurine with diabetes and hypertension risk.
Conclusions
The findings suggest that high levels of taurine may be protective against CHD among individuals with high serum cholesterol levels.
Keywords: Taurine, serum, coronary heart disease, NYUWHS, epidemiology
Introduction
Coronary heart disease (CHD) is the leading killer of American men and women causing one in five deaths [1]. Large prospective epidemiologic studies have provided evidence that nutritional factors are important modifiable risk factors for CHD. Nutrients such as beta carotene, folate, fiber, and vitamins A, B and C have been shown to be beneficial to the cardiovascular system [2,3]. Identification of novel dietary protective factors may improve the understanding, control, and prevention of CHD.
Taurine (2-aminoethanesulfonic acid) is a sulfur-containing molecule obtained mainly through diet in humans, with miniscule amounts made endogenously. Interest in the health effects of taurine originated from human [4,5] studies in which intakes of foods rich in sulfur-containing amino acids, such as fish, were inversely associated with the risk of hypertension and cardiovascular disease (CVD). In vitro and animal studies have suggested that taurine is involved in a variety of biological processes, including bile salt conjugation, osmoregulation, blood pressure regulation, and antioxidation [6]. Ecological analyses from the WHO Cardiovascular Diseases and Alimentary Comparison (WHO-CARDIAC), a multi-center cross-sectional study, have reported an inverse correlation between group means of urinary excretion of taurine, a surrogate of taurine intake, and mortality rates due to ischemic heart disease (IHD) in both men [7] and women [8]. However, the observed correlations could be the result of ecologic fallacy because factors associated with disease rates at the group level may not be associated with disease at the individual level [9]. In addition, the correlations may be confounded by unmeasured factors related to both taurine intake and IHD at the individual level. Several clinical trials, ecologic studies, and cross-sectional studies have reported that taurine can lower cholesterol [5,10,11] and blood pressure [5,10,12,13]. However, these studies were limited in sample size, short in duration of follow-up, and/or potentially subject to ecologic fallacy [14]. Although taurine has been a popular supplement and ingredient in “energy drinks” in recent years, to date, no prospective epidemiologic studies have evaluated the association between taurine intake and the risk of CHD.
We conducted a prospective case-control study nested in the New York University Women’s Health Study (NYUWHS) to evaluate whether a high concentration of serum taurine is protective against CHD risk. We also assessed the distribution and correlates of serum taurine as well as the associations between serum taurine and risk factors for CHD.
Methods
New York University Women’s Health Study (NYUWHS)
Details of the NYUWHS have been described elsewhere [15]. Briefly, a total of 14,274 women, 34 to 65 years of age, were enrolled between 1985 and 1991 at a breast cancer screening center in New York City. At the time of enrollment, demographic, medical and lifestyle information was collected using a self-administered questionnaire. Active follow-up of the cohort is conducted with questionnaires mailed approximately every two to four years. Lost participants are located using the National Change of Address database of the US Postal Service, the Choice Point credit bureau database, internet-based resources, primarily telephone directories and the Social Security Death Index, the National Death Index (NDI), and telephone calls to contacts. The response rate for the last round of follow-up, completed in 2006, was 83% of the original cohort members.
Physical activity at baseline was assessed in two follow-up questionnaires in which participants were directed to report their activity levels during the baseline period. Cases of hypertension and diabetes were identified using self-report information on physician-diagnosis of the condition, and/or medicine use for these conditions collected at baseline and in all follow-up questionnaires.
Dietary information for the year prior to cohort enrollment was collected at baseline using a self-administered food frequency questionnaire (FFQ) of 70 typical American food items. The questionnaire was similar to the validated Block FFQ [16], with minor modifications to the food list. Details of the dietary assessment in the NYUWHS, including its reproducibility, have been described elsewhere [17]. Daily nutrient intakes were calculated using food composition tables elaborated at the National Cancer Institute [16], with minor modifications.
Non-fasting blood was collected at enrollment, centrifuged and the serum was stored at −80°C for subsequent analysis. Fifty one percent of the participants (n = 7,344) gave blood at more than one visit at roughly yearly intervals. Compared with women who gave blood once, women who donated blood two or more times were older (mean age 53.4 versus 49.3 years old) and more likely to be of European descent (81% versus 77%). Distributions of body mass index (BMI), education, smoking status and multivitamin use were similar in the two groups (data not shown).
The study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human subjects were approved by the IRB of New York University School of Medicine. Written informed consent was obtained from all subjects at time of enrollment.
Ascertainment of CHD endpoints
In each follow-up round of the NYUWHS, participants who reported a CHD event were contacted for further information, including authorization to obtain their medical records. CHD cases were ascertained and verified based on medical chart review and NDI. CHD cases included women who had documented non-fatal or fatal MI, fatal CHD, or surgery to remedy restricted blood flow including coronary artery bypass grafting (CABG) and percutaneous transluminal coronary angioplasty (PTCA). Fatal CHD was classified when: 1) fatal MI was the finding at autopsy, or 2) the underlying cause of death from the NDI was CHD (ICD9 codes 410–414 or ICD10 codes I20–I25), provided that during previous follow-up there was evidence from questionnaires or medical records of pre-existing CHD. Non-fatal MI was confirmed if it met the World Health Organization criteria of symptoms [18] for either diagnostic electrocardiographic changes or elevated levels of cardiac enzymes.
Nested case-control study design and subject selection
In a pilot study [19], we determined that the mean of two yearly measurements would provide an estimate of the long-term taurine serum concentration with temporal reliability similar to that of other biomarkers in epidemiologic studies [20,21], i.e. an intraclass correlation coefficient of 0.65. We therefore selected the first 223 CHD cases (1986–2006) from NYUWHS participants who had donated blood ≥ 2 times during follow-up visits, and who had no history of CVD at the time of blood donation. For each CHD case, one control was selected at random from the NYUWHS subjects who were alive and free of CVD at the date of the coronary event of the case. Cases and controls were matched on age at baseline (± 6 months), menopausal status (pre- and postmenopausal), total number of blood donations (2 or >2), and the dates of both blood donations (± 6 months).
Analysis of serum taurine, total cholesterol and HDL-cholesterol
Laboratory personnel were blinded to the identity and case-control status of study participants. Taurine was determined using a high performance liquid chromatography (HPLC) (Waters, Millford, MA) with precolumn derivitization using o-phthaldialdehyde (Sigma-Aldrich HPLC) and 3-mercaptopropionic acid (Fluka Biochemika) with fluorescence detection. The reagents, sample preparation, and derivitization methods were slightly modified from a previously published protocol [22] as previously described [19]. The coefficient of variation (CV) was 7.3%.
Total cholesterol and HDL-cholesterol measurements were performed by Pacific Biometrics Incorporated, in Seattle, Washington. Total cholesterol was measured using CDC-standardized enzymatic assay with a Trinder end-point reaction. HDL-cholesterol was measured using dextran sulfate precipitation [23] followed by an enzymatic in-house method similar to that used by the CDC Cholesterol Reference Method Laboratory [24].
Statistical Analysis
The distribution of baseline characteristics and established CHD risk factors in CHD cases and controls were compared using conditional logistic regression model which takes into account the matching factors [25]. The mean of the serum taurine in the two yearly samples was the main exposure variable and for simplicity is referred to as serum taurine from here on.
To investigate determinants of serum taurine, linear regression models were used with serum taurine as the dependent variable and participants’ characteristics as the independent variables. Pearson partial correlation coefficients (r) were used to quantify the association between serum taurine and consumption of food groups and intakes of macro- and micro-nutrients, adjusting for variables that may be associated with intakes of taurine and other dietary factors, i.e. age, CHD status, BMI and total calories. The nutrients of interest were those relevant to CHD risk [2,26–30]. Nutrient intakes estimated from the FFQ were energy-adjusted using the residual method [31]. Dietary data from participants who reported unreasonable total caloric intake of less than 500 (n = 9) or greater than 3,500 calories per day (n = 10) were excluded. Serum taurine and dietary intakes were log-transformed to improve the normality of their distributions. The analyses were also conducted in controls only.
To evaluate the association between serum taurine and CHD risk, we estimated odds ratios (ORs) and 95% confidence intervals (CIs) for CHD in relation to tertiles of serum taurine using conditional logistic regression. Cutoff points for tertiles were determined based on the distribution of serum taurine in cases and controls combined. ORs were first adjusted for BMI, smoking status, and serum total cholesterol, which are all major risk factors for CHD. ORs were further adjusted for total calories and potential confounders related to both CHD risk and serum taurine in our data, i.e. energy-adjusted intakes of saturated fat, folate, fiber, and potassium. Tests for linear trend were conducted using an ordered variable for tertile categories of serum taurine.
To examine whether the association between serum taurine and CHD differs by established risk factors for CHD, analyses were stratified by BMI (≤ and > median, i.e. 25 kg/m2), smoking status (never/ever), and total serum cholesterol (≤ and > median, i.e. 250 mg/dl). Multiplicative interaction was tested by using a cross-product term representing the product of a potential effect-modifier and serum taurine.
To assess whether a high concentration of pre-diagnostic serum taurine is protective against subsequent occurrence of hypertension or diabetes, we conducted unconditional logistic regression analyses in cases and controls combined as well as in controls only. Prevalent cases of diabetes or hypertension at baseline were excluded from the respective analyses. All analyses were completed using SAS (version 9.2; SAS Institute Inc, Cary, NC).
Results
Serum samples were stored for an average of 22.7 years. At the time of baseline blood donation, the median age of the study participants was 58, and CHD events occurred at an average age of 70 years. Compared with controls, cases had a significantly higher BMI, and were more likely to have a parental history of heart attack, to be hypertensive, and to have a higher total serum cholesterol as well as a lower serum HDL-cholesterol at baseline (p-values < 0.05) (Table 1). There were no differences between CHD cases and controls by race, level of education, physical activity, or smoking status. The median of serum taurine was similar in cases and controls (126.9 nmol/ml and 130.1 nmol/ml respectively, p = 0.65). The geometric mean of taurine differed significantly by age at baseline (p < 0.01) (Table 1). There were no significant associations between all other baseline characteristics and serum taurine, with p-values ranging from 0.15 to 0.98.
Table 1.
Baseline characteristics of CHD cases and controls, and geometric mean of average taurine
| Baseline Characteristic | n (%) or median (10th & 90th percentiles) | n (%) or median (10th & 90th percentiles) | P-value | Geometric Mean of Taurine a (nmol/mL) | P-value |
|---|---|---|---|---|---|
| Cases (n = 223) | Controls (n = 223) | n = 446 | |||
| Age at baseline (yrs) | |||||
| < 55 | 64 (28.7) | 66 (29.6) | Matched | 117.7 | <0.01 |
| 55 – 59 | 62 (27.8) | 61 (27.4) | 127.8 | ||
| ≥ 60 | 97 (43.5) | 96 (43.0) | 132.2 | ||
| Continuous | 58.9 (47.2–64.2) | 58.6 (47.1–64.3) | |||
| Menopausal status at enrollment | |||||
| Premenopausal | 45 (20.2) | 45 (20.2) | Matched | 126.7 | 0.98 |
| Postmenopausal | 178 (79.8) | 178 (79.8) | 126.6 | ||
| Length of sample storage time (yrs) | |||||
| Continuous | 22.7 (21.7–23.5) | 22.7 (21.7–23.5) | Matched | ||
| Race | |||||
| Caucasian | 182 (86.6) | 178 (84.4) | 0.73 | 126.7 | 0.75 |
| African-American | 17 (8.1) | 17 (8.1) | 128.2 | ||
| Hispanic | 5 (2.4) | 10 (4.7) | 123.8 | ||
| Other | 6 (2.9) | 6 (2.8) | 123.7 | ||
| Education b | |||||
| ≤ High school | 85 (41.9) | 79 (39.1) | 0.37 | 127.0 | 0.92 |
| Some college or equivalent | 62 (30.5) | 45 (22.3) | 123.2 | ||
| College degree | 56 (27.6) | 78 (38.6) | 127.7 | ||
| BMI (kg/m2) | |||||
| ≤ 25 | 97 (43.5) | 126 (56.5) | 0.01 | 127.5 | 0.51 |
| 25.1 – 30 | 83 (37.2) | 68 (30.5) | 126.2 | ||
| > 30 | 43 (19.3) | 29 (13.0) | 124.9 | ||
| Continuous | 25.8 (21.6–32.3) | 24.2 (20.2–30.9) | <0.01 | ||
| Physical activity (MET-hrs/wk) | |||||
| 0 | 13 (6.0) | 7 (3.2) | 0.70 | 140.2 | 0.60 |
| > 0–10 | 91 (42.2) | 101 (46.8) | 125.5 | ||
| > 10–20 | 42 (19.4) | 34 (15.7) | 123.8 | ||
| > 20 | 70 (32.4) | 74 (34.3) | 129.0 | ||
| Continuous | 11.0 (0.6–48.1) | 10.3 (0.8–54.8) | 0.74 | 0. 52 | |
| Smoking status | |||||
| Never | 87 (40.1) | 103 (48.6) | 0.34 | 125.0 | 0.44 |
| Past | 50 (23.0) | 32 (15.1) | 126.4 | ||
| Current | 80 (36.9) | 77 (36.3) | 127.6 | ||
| Parental history of heart attack | |||||
| Yes | 90 (46.2) | 74 (39.6) | 0.01 | 125.2 | 0.25 |
| No | 105 (53.9) | 113 (60.4) | 128.9 | ||
| Prevalent hypertension | |||||
| Yes | 60 (26.9) | 32 (14.4) | <0.01 | 121.7 | 0.15 |
| No | 163 (73.1) | 191 (85.6) | 128.1 | ||
| Serum total cholesterol (mg/dl) | |||||
| < 200 | 16 (7.2) | 34 (15.2) | <0.01 | 125.2 | 0.15 |
| 200–239 | 58 (26.1) | 80 (35.9) | 124.4 | ||
| 240–269 | 62 (27.9) | 50 (22.4) | 125.0 | ||
| ≥ 270 | 86 (38.7) | 59 (26.5) | 130.5 | ||
| Continuous | 257.0 (206.0–312.0) | 239.0 (188.0–297.0) | <0.01 | 0.10 | |
| Serum HDL-cholesterol (mg/dl) | |||||
| < 50 | 30 (13.5) | 15 (6.7) | <0.01 | 123.6 | 0.70 |
| 50–59 | 130 (58.3) | 112 (50.2) | 128.5 | ||
| ≥ 60 | 63 (28.3) | 96 (43.1) | 124.8 | ||
| Continuous | 52.0 (38.0–73.0) | 57.0 (41.0–76.0) | <0.01 | ||
| Serum taurine (nmol/mL) | |||||
| 1st measurement | 127.4 (80.0–180.1) | 128.9 (85.6–173.2) | 0.33 | - | - |
| 2nd measurement | 131.3 (87.0–195.0) | 130.7 (89.1–183.1) | 0.84 | ||
| Average of 1st and 2nd measurement | 126.9 (92.5–173.0) | 130.1 (93.3–165.7) | 0.65 |
P-values for n or the median were obtained from conditional logistic regression excluding those with missing information. Variables with categories were entered as ordered categorical.
Average of two yearly measurements was used to calculate the geometric mean.
P-values for the geometric mean of taurine were obtained from linear regression with taurine as the dependent variable adjusted for age. Variables with categories were entered as ordered categorical.
Geometric mean adjusted for age (except for age variable)
Data were missing on education for 21 cases and 21 controls; on race for, respectively 13 and 12 subjects; on physical activity for 7 and 7 subjects; on parental history of CVD for 28 and 36 subjects; on smoking status at baseline for 6 and 11 subjects; and on total serum cholesterol for 1 case (this was an outlier).
Serum taurine was significantly negatively associated with dietary intake of saturated fat (r = −0.10) and positively related to dietary intakes of niacin (r = 0.11), vitamin B1 (r = 0.11), iron (r = 0.10), and fiber (r = 0.10) (Table 2). Of the food groups, consumption of poultry was significantly positively associated with serum taurine (r = 0.14). There was no apparent correlation between serum taurine and other dietary factors. These correlations were similar among the controls (data not shown).
Table 2.
Pearson partial correlation coefficient of log (average taurine) and log dietary variables
| Dietary Variables | Pearson Partial Correlation Coefficient a | P-value |
|---|---|---|
| Macronutrients b | ||
| Total calories (kcal/d) | −0.02 | 0.68 |
| Saturated Fat (g/d) | −0.10 | 0.04 |
| Carbohydrates (g/d) | 0.06 | 0.20 |
| Total Fat (g/d) | −0.06 | 0.24 |
| Cholesterol (mg/d) | −0.04 | 0.42 |
| Protein (g/d) | 0.02 | 0.69 |
| Micronutrients† | ||
| Niacin (mg/d) | 0.11 | 0.03 |
| Vitamin B1 (mg/d) | 0.11 | 0.03 |
| Fiber (g/d) | 0.10 | 0.05 |
| Iron (mg/d) | 0.10 | 0.05 |
| Folate (μg/d) | 0.09 | 0.06 |
| Potassium (mg/d) | 0.08 | 0.12 |
| Vitamin E (mg/d) | 0.05 | 0.27 |
| Phosphorus (mg/d) | 0.06 | 0.24 |
| Oleic acid (g/d) | −0.05 | 0.31 |
| Linoleic acid (g/d) | 0.04 | 0.40 |
| Vitamin B2 (mg/d) | 0.04 | 0.42 |
| Vitamin C (mg/d) | 0.03 | 0.48 |
| Vitamin A (IU/d) | 0.03 | 0.49 |
| Carotene (μg/d) | 0.01 | 0.86 |
| Retinol (μg/d) | < −0.01 | 0.92 |
| Sodium (mg/d) | < −0.01 | 0.95 |
| Calcium (mg/d) | −0.01 | 0.93 |
| Food Groups (g/d) | ||
| Poultry | 0.14 | < 0.01 |
| Shellfish & mollusks (oysters, clams, shrimp) | −0.11 | 0.06 |
| Fruit | 0.06 | 0.24 |
| Vegetables | 0.05 | 0.34 |
| Meat | −0.03 | 0.51 |
| Fish | 0.01 | 0.86 |
All partial Pearson correlations were adjusted for CHD case/control status, age, BMI and log(total calories) except for the correlation with “total calories” which was which was not adjusted for log(total calories). Participants with caloric intake <500 and >3,500 were excluded.
All macro- and micronutrients were log transformed and energy adjusted using the residual method
Overall, there was no statistically significant association between the risk of CHD and serum taurine levels (Table 3). The adjusted ORs for CHD associated with increasing tertiles of serum taurine were 1.0 (reference), 0.85 (95% CI, 0.51–1.40), and 0.66 (0.39–1.13), respectively (p for trend = 0.14), adjusting for BMI, smoking, and total cholesterol.
Table 3.
Association between mean taurine and CHD overall and stratified
| Taurine Tertile 1 (35.75–115.00 nmol/mL) | Taurine Tertile 2 (115.01–141.71 nmol/mL) | Taurine Tertile 3 (141.72–300.31 nmol/mL) | P for trend | P for interaction | |
|---|---|---|---|---|---|
| Overall | |||||
| Case/control | 81/74 | 73/74 | 69/75 | ||
| Model 1 OR a (95% CI) | 1.00 | 0.90 (0.57–1.43) | 0.82 (0.50–1.34) | 0.42 | |
| Model 2 OR b (95% CI) | 1.00 | 0.85 (0.51–1.40) | 0.66 (0.39–1.13) | 0.14 | |
|
| |||||
| BMI <25 | |||||
| Case/control | 38/45 | 30/38 | 29/43 | ||
| Model 1 OR a (95% CI) | 1.00 | 0.88 (0.45–1.73) | 0.77 (0.40–1.48) | 0.43 | |
| Model 2 OR b (95% CI) | 1.00 | 0.81 (0.40–1.62) | 0.64 (0.32–1.28) | 0.21 | |
| BMI ≥25 | 0.27 | ||||
| Case/control | 43/29 | 43/36 | 40/32 | ||
| Model 1 OR a (95% CI) | 1.00 | 0.84 (0.44–1.60) | 0.82 (0.42–1.59) | 0.55 | |
| Model 2 OR b (95% CI) | 1.00 | 0.81 (0.41–1.58) | 0.83 (0.42–1.63) | 0.59 | |
| Never smoking | |||||
| Case/control | 35/34 | 32/35 | 20/34 | ||
| Model 1 OR a (95% CI) | 1.00 | 0.84 (0.42–1.66) | 0.54 (0.26–1.13) | 0.11 | |
| Model 2 OR b (95% CI) | 1.00 | 0.76 (0.38–1.54) | 0.51 (0.24–1.09) | 0.09 | |
| Ever smoking | 0.25 | ||||
| Case/control | 45/37 | 39/35 | 46/37 | ||
| Model 1 OR a (95% CI) | 1.00 | 0.93 (0.49–1.77) | 1.04 (0.56–1.94) | 0.89 | |
| Model 2 OR b (95% CI) | 1.00 | 0.83 (0.42–1.74) | 0.90 (0.47–1.74) | 0.77 | |
|
| |||||
| Total cholesterol <250 mg/dl | |||||
| Case/control | 35/56 | 36/36 | 29/39 | ||
| Model 1 OR a (95% CI) | 1.00 | 1.63 (0.86–3.06) | 1.19 (0.63–2.27) | 0.52 | |
| Model 2 OR b (95% CI) | 1.00 | 1.69 (0.89–3.24) | 1.19 (0.62–2.30) | 0.52 | |
| Total cholesterol ≥ 250 mg/dl | 0.01 | ||||
| Case/control | 46/18 | 37/38 | 40/36 | ||
| Model 1 OR a (95% CI) | 1.00 | 0.38 (0.18–0.78) | 0.45 (0.22–0.93) | 0.05 | |
| Model 2 OR b (95% CI) | 1.00 | 0.31 (0.15–0.67) | 0.39 (0.19–0.83) | 0.02 | |
ORs for were calculated using unadjusted logistic regression conditional on the matching factors including age, menopausal status, number and dates of blood donations.
ORs were calculated using logistic regression conditional on matching factors with additional adjustment for BMI, smoking, and log(total cholesterol) (except
for variable stratified by in stratified analyses).
The association between serum taurine and CHD risk did not differ appreciably according to baseline BMI (p for interaction = 0.27) (Table 3). Although the inverse association was stronger among never smokers (p for trend = 0.09) than among ever smokers (p for trend = 0.77), there was no statistical evidence of interaction on the multiplicative scale (p for interaction = 0.25). Among women with high total serum cholesterol (>250 mg/dl), those in the highest two tertiles of serum taurine were less likely to develop CHD than those in the lowest tertile (OR = 0.31 (0.15–0.67 and 0.39 (0.19–0.83) for second and third tertile, respectively). There was no association between serum taurine and CHD risk in women with low cholesterol level (p for interaction = 0.01). Further adjustment for total calories and energy-adjusted intakes of saturated fat, folate, fiber, and potassium did not materially change the effect estimates for main effect and interaction analyses (data not shown).
Lastly, we evaluated the associations of serum taurine with the risk of hypertension and diabetes, excluding prevalent cases at baseline. Serum taurine was significantly lower among women who developed hypertension (123.0 nmol/ml) compared with those who did not (130.2 nmol/ml, p = 0.04). In logistic regression models, serum taurine was inversely associated with the risk of hypertension, although the association was not statistically significant (Table 4). Among controls, the association between serum taurine and hypertension risk was significant, with a 45–60% reduction in the risk of hypertension for women in the highest two tertiles of serum taurine compared with those in the first tertile (OR = 0.65 (0.29–1.46) and 0.40 (0.17–0.97) for second and third tertile, respectively (p for trend = 0.04)). Serum taurine was lower among women who developed diabetes (118.5 nmol/ml) compared with those who did not (127.5 nmol/ml, p = 0.07). There was also a non-significant inverse association between serum taurine and risk of diabetes (Table 4). The analysis regarding diabetes risk was not conducted in controls only due the limited numbers of diabetes cases.
Table 4.
Association between taurine and incident hypertension and incident diabetes, overall and in controls
| Hypertension a | Taurine Tertile 1 (35.75–115.00 nmol/mL) | Taurine Tertile 2 (115.01–141.71 nmol/mL) | Taurine Tertile 3 (141.72–300.31 nmol/mL) | P-value |
|---|---|---|---|---|
| N (yes/no) | 45/77 | 36/77 | 32/87 | |
| Model 1 ORb (95% CI) | 1.00 | 0.78 (0.45–1.35) | 0.63 (0.37–1.09) | 0.10 |
| Model 2 ORc (95% CI) | 1.00 | 0.79 (0.45–1.39) | 0.60 (0.34–1.05) | 0.07 |
|
| ||||
| Diabetes a | ||||
| N (yes/no) | 20/126 | 10/129 | 12/129 | |
| Model 1 ORb (95% CI) | 1.00 | 0.47 (0.21–1.06) | 0.57 (0.26–1.21) | 0.13 |
| Model 2 ORc (95% CI) | 1.00 | 0.37 (0.16–0.89) | 0.50 (0.22–1.11) | 0.07 |
|
| ||||
| In controls only | ||||
|
| ||||
| Hypertension a | ||||
| N CHD controls (yes/no) | 14/43 | 13/49 | 7/53 | |
| Model 1 ORb (95% CI) | 1.00 | 0.69 (0.31–1.53) | 0.40 (0.17–0.96) | 0.04 |
| Model 2 ORc (95% CI) | 1.00 | 0.65 (0.29–1.46) | 0.40 (0.17–0.97) | 0.04 |
Prevalent cases of hypertension and diabetes were excluded from the respective analyses
ORs were calculated using unconditional logistic regression adjustment for age.
ORs were calculated using unconditional logistic regression adjusted for age, BMI, case/control status and smoking (no case/control adjustment in controls only analysis).
P-value for trend based on taurine tertiles
Discussion
To the best of our knowledge, this is the first prospective study to examine the relationship between serum taurine and CHD risk. Overall, there was no statistically significant association between the risk of CHD and serum taurine levels (p for trend = 0.14). We found a significant inverse association between serum taurine and CHD risk among women with higher serum total cholesterol which was not observed in women with lower serum total cholesterol.
Although we measured serum taurine in two yearly serum samples to reduce the potential misclassification of taurine levels, it remains possible that taurine has a small effect on CHD risk that could not be detected in the present study, given the sample size of the study. In addition, it is possible that the beneficial effects of taurine on CHD risk may be exerted at a much higher level of taurine intake. In clinical trials of taurine and blood pressure [10,11,13], taurine was given at 3–6 g/day, much higher than the typical daily intake of 40–400 mg/day [32,33].
Our data suggest that cholesterol level may modify the effect of taurine on CHD risk. Taurine may be protective against CHD risk in the presence of adverse effects associated with high cholesterol level. There is evidence linking hypercholesterolemia with increased generation of reactive oxygen species, leading to high levels of circulating oxidized LDL [34] and pro-inflammatory cytokines [35], and numerous cellular and animal studies have demonstrated that taurine has antioxidant and anti-inflammatory properties. Taurine supplementation has been shown to significantly attenuate oxidized-LDL-induced cell injury in Sprague-Dawley rats [36]. In spontaneously hyperlipidimic mice, taurine supplementation significantly decreased levels of lipid oxidation and suppressed the development of atherosclerotic lesions [37]. The evidence suggests that taurine may interact with excessive oxidation and inflammation caused by increased cholesterol concentrations, thus reducing the associated CHD risk. Future epidemiologic studies are needed to confirm our findings.
Our study is also the first epidemiologic study to assess the distribution and correlates of serum taurine. Consistent with studies about taurine contents in foods [38–40], we found a significantly positive association between dietary intake of poultry and serum taurine. However, intake of shellfish and mollusks, the foods with the highest concentration of taurine [38–40], were not significantly correlated with serum taurine level. This is most likely the result of infrequent intake of shellfish and mollusks in our study population, with an average intake of only 14 g/wk (0.49 ounces/wk). Our data indicate that poultry intake, with an average consumption of 161 g/wk (5.68 ounces/wk), was the major source of taurine in our study population. The literature indicates that serum or urinary taurine level is largely determined by dietary taurine intakes, rather than individual ethnic or genetics background [5,41]. However, the content of taurine differs appreciably by type of seafood and cut of meat [42]. Since most of the FFQs, including the one used in our study, do not specify the cut of meat consumed, the association between FFQ-measured dietary food intake and taurine serum level could differ because of the cuts of meat consumed by different populations. The potential beneficial effect of taurine on CHD risk among women with high cholesterol found in the study, if confirmed, may reinforce the role of poultry as part of healthy diet for preventing CHD.
Serum taurine’s negative association with saturated fat and positive associations with niacin, vitamin B1, iron, and fiber may be an indirect result of an overall healthy diet pattern that is associated with high levels of taurine and these nutrients. Women who consumed more chicken, which contains high levels of niacin, iron, and taurine, may have also consumed more vegetables, which are associated with high levels of iron and fiber, and more whole grains, which contain more vitamin B1 and fiber. Importantly, the main effect of taurine on CHD risk and the interaction between taurine and cholesterol did not change after controlling for nutritional factors.
The data also suggest that taurine may be associated with hypertension and diabetes. In ecologic studies, Mizushima et al. [5] found a negative association between urinary taurine excretion and hypertension prevalence comparing two different Japanese populations, and Liu et al. [12] showed a negative correlation between urinary taurine excretion and blood pressure in ethnic Chinese groups. A small clinical trial reported that taurine supplementation decreased blood pressure in 19 borderline hypertensive volunteers [13]. The results of the present study possibly extend the findings of the previous studies because taurine and hypertension status were measured at the individual level and participants were healthy at baseline. The evidence of taurine’s effect on diabetes is limited. A study of taurine supplementation showed a reduction of glucose levels measured in plasma and urine [43]. Future large studies are needed to evaluate the potential beneficial effects of taurine on hypertension, and possibly diabetes.
The present study has several strengths. The prospective nature of the study design with participants free of CVD at the time of blood donation minimized the likelihood that serum taurine was influenced by disease or a change in diet due to disease. The study involved a long-term follow-up, average of 23 years, allowing us to study long-term effects of taurine. The standardization of sample collection, storage and handling procedures [44] are also strengths of this study because measurements made from biological samples can be highly dependent on these factors. Because the study enrollment and blood donation occurred between 1985 and 1991, about a decade before the “energy drink” became popular in the United States, it is certain that the main source of taurine measured in the samples of the NYUWHS participants derived from food, not from enhanced drinks or over the counter supplements. Importantly, the use of repeated samples improved the reliability of the measurement of taurine, and consequently the validity of the findings.
It should be noted that our study population included over 80% Caucasian women, and therefore the generalizability of results to men and other races should be considered with caution. However, there is no evidence that the biological actions of taurine would differ by sex or race. In addition, our results cannot be generalized to indicate health effects of energy drinks, which should be investigated separately, as these drinks often contain not only taurine but also a multitude of other ingredients such as caffeine and ginseng that may influence CHD risk. Although our ascertainment of hypertension status was based on self-reported data, the validity of self-reported hypertension status has previously been shown to have >80% sensitivity and specificity in other epidemiologic studies [45,46]. Lastly, we cannot exclude the role of chance, unmeasured confounders, or biases in our findings, especially those relating subgroup analyses.
In conclusion, the results of this study suggest that although there was no significant overall association between taurine and CHD risk, high levels serum taurine may be protective against CHD risk in those with high serum total cholesterol. The findings also suggest a possible inverse association between serum taurine and the risk of hypertension and diabetes.
Acknowledgments
This research was supported by U.S. grants: NIH grants ES000260, CA16087, CA098661 and American Heart Association grant 0835569D. The authors thank Mr. Alan Bowers and Mrs. Yelena Afanasyeva for their involvement in the study.
Footnotes
Conflict of interest: The authors declare they have no conflict of interest.
Contributor Information
Oktawia P. Wójcik, Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA
Karen L. Koenig, Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA
Anne Zeleniuch-Jacquotte, Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA. New York University Cancer Institute, New York University School of Medicine, New York, NY 10016, USA.
Camille Pearte, Department of Medicine, Division of Cardiology, New York University School of Medicine, New York, NY 10016, USA.
Max Costa, Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA. New York University Cancer Institute, New York University School of Medicine, New York, NY 10016, USA. Department of Pharmacology, New York University School of Medicine, New York, NY 10016, USA.
Yu Chen, Department of Environmental Medicine, New York University School of Medicine, New York, NY 10016, USA. New York University Cancer Institute, New York University School of Medicine, New York, NY 10016, USA.
References
- 1.AHA. 2009 Update At-A-Glance. Dallas, TX: American Heart Association; 2009. pp. 1–36. [Google Scholar]
- 2.Mente A, de Koning L, Shannon HS, Anand SS. A systematic review of the evidence supporting a causal link between dietary factors and coronary heart disease. Arch Intern Med. 2009;169:659–669. doi: 10.1001/archinternmed.2009.38. [DOI] [PubMed] [Google Scholar]
- 3.Visioli F, Hagen TM. Nutritional strategies for healthy cardiovascular aging: focus on micronutrients. Pharmacol Res. 2007;55:199–206. doi: 10.1016/j.phrs.2007.01.008. [DOI] [PubMed] [Google Scholar]
- 4.Yamori Y, Nara Y, Ikeda K, Mizushima S. Is taurine a preventive nutritional factor of cardiovascular diseases or just a biological marker of nutrition? Adv Exp Med Biol. 1996;403:623–629. doi: 10.1007/978-1-4899-0182-8_69. [DOI] [PubMed] [Google Scholar]
- 5.Mizushima S, Moriguchi EH, Ishikawa P, Hekman P, Nara Y, et al. Fish intake and cardiovascular risk among middle-aged Japanese in Japan and Brazil. J Cardiovasc Risk. 1997;4:191–199. [PubMed] [Google Scholar]
- 6.Huxtable RJ. Physiological actions of taurine. Physiol Rev. 1992;72:101–163. doi: 10.1152/physrev.1992.72.1.101. [DOI] [PubMed] [Google Scholar]
- 7.Yamori Y, Liu L, Mizushima S, Ikeda K, Nara Y. Male cardiovascular mortality and dietary markers in 25 population samples of 16 countries. Journal of Hypertension. 2006;24:1499–1505. doi: 10.1097/01.hjh.0000239284.12691.2e. [DOI] [PubMed] [Google Scholar]
- 8.Yamori Y, Liu L, Ikeda K, Miura A, Mizushima S, et al. Distribution of twenty-four hour urinary taurine excretion and association with ischemic heart disease mortality in 24 populations of 16 countries: Results from the WHO-CARDIAC Study. Hypertens Res. 2001;24:453–457. doi: 10.1291/hypres.24.453. [DOI] [PubMed] [Google Scholar]
- 9.Greenland S, Robins J. Invited commentary: ecologic studies--biases, misconceptions, and counterexamples. Am J Epidemiol. 1994;139:747–760. doi: 10.1093/oxfordjournals.aje.a117069. [DOI] [PubMed] [Google Scholar]
- 10.Mizushima S, Nara Y, Sawamura M, Yamori Y. Effects of oral taurine supplementation on lipids and sympathetic nerve tone. Adv Exp Med Biol. 1996;403:615–622. doi: 10.1007/978-1-4899-0182-8_68. [DOI] [PubMed] [Google Scholar]
- 11.Zhang M, Bi LF, Fang JH, Su XL, Da GL, et al. Beneficial effects of taurine on serum lipids in overweight or obese non-diabetic subjects. Amino Acids. 2004;26:267–271. doi: 10.1007/s00726-003-0059-z. [DOI] [PubMed] [Google Scholar]
- 12.Liu L, Liu L, Ding J, Huang Z, He B, et al. Ethnic and environmental differences in various markers of dietary intake and blood pressure among Chinese Han and three other minority peoples of China: Results from the WHO cardiovascular diseases and alimentary comparison (CARDIAC) study. Hepertens Res. 2001;24:315–322. doi: 10.1291/hypres.24.315. [DOI] [PubMed] [Google Scholar]
- 13.Fujita T, Katsuyuki A, Noda H, Yasushi I, Sato Y. Effects of increased adrenomedullary activity and taurine in young patients with borderline hypertension. Circulation. 1987;75:525–532. doi: 10.1161/01.cir.75.3.525. [DOI] [PubMed] [Google Scholar]
- 14.Wójcik O, Koenig KL, Zeleniuch-Jacquotte A, Costa M, Chen Y. The potential protective effects of taurine on coronary heart disease. Atherosclerosis. 2010;208:19–25. doi: 10.1016/j.atherosclerosis.2009.06.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Toniolo PG, Pasternack BS, Shore RE, Sonnenschein E, Koenig KL, et al. Endogenous hormones and breast cancer: a prospective cohort study. Breast Cancer Res Treat. 1991;18:S23–26. doi: 10.1007/BF02633522. [DOI] [PubMed] [Google Scholar]
- 16.Block G, Hartman AM. Issues in reproducibility and validity of dietary studies. Am J Clin Nutr. 1989;50:1133–1138. doi: 10.1093/ajcn/50.5.1133. [DOI] [PubMed] [Google Scholar]
- 17.Riboli E, Toniolo P, Kaaks R, Shore RE, Casagrande C, et al. Reproducibility of a food frequency questionnaire used in the New York University Women’s Health Study: effect of self-selection by study subjects. Eur J Clin Nutr. 1997;51:437–442. doi: 10.1038/sj.ejcn.1600422. [DOI] [PubMed] [Google Scholar]
- 18.Rose GA, Blackburn H, Gillum RF, Prineas RJ. Cardiovascular survey methods. Geneva, Switzerland: World Health Organization; 1982. [Google Scholar]
- 19.Wójcik O, Koenig KL, Zeleniuch-Jacquotte A, Costa M, Chen Y. Reproducibility of HPLC Taurine Measurements in Frozen Serum of Healthy Postmenopausal Women. British Journal of Nutrition. 2010;104:629–632. doi: 10.1017/S000711451000108X. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Al-Delaimy WK, Natarajan L, Sun X, Rock CL, Pierce JP Women’s Healthy Eating and Living (WHEL) Study Group. Reliability of plasma carotenoid biomarkers and its relation to study power. Epidemiology. 2008;19:338–344. doi: 10.1097/EDE.0b013e3181635dc2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Hunter D. Nutritional Epidemiology. 2. New York, NY: Oxford University Press; 1998. Biochemical Indicators of Dietary Intake; pp. 174–243. [Google Scholar]
- 22.Tcherkas YV, Kartsova LA, Krasnova IN. Analysis of amino acids in human serum by isocratic reversed-phase high-performance lipid chromatography with electrochemical detection. Journal of Chromatography A. 2001;913:303–308. doi: 10.1016/s0021-9673(00)01206-1. [DOI] [PubMed] [Google Scholar]
- 23.McNamara JR, Huang C, Massov T, Leary ET, Warnick GR, et al. Modification of the dextran-Mg2+ high-density lipoprotein cholesterol precipitation method for use with previously frozen plasma. Clin Chem. 1994;40:233–239. [PubMed] [Google Scholar]
- 24.Kimberly MM, Leary ET, Cole TG, Waymack PP. Selection, validation, standardization, and performance of a designated comparison method for HDL-cholesterol for use in the cholesterol reference method laboratory network. Clin Chem. 1999;45:1803–1812. [PubMed] [Google Scholar]
- 25.Breslow N, Day NE. Conditional logistic regression for matched sets. In: Davis W, editor. Statistical Methods in Cancer Research Vol 1 The Analysis of Case-Control Studies. Lyon, France: IARC; 1980. pp. 248–279. [Google Scholar]
- 26.Ito M. The metabolic syndrome: pathophysiology, clinical relevance, and use of niacin. Ann Pharmacother. 2004;38:277–285. doi: 10.1345/aph.1D218. [DOI] [PubMed] [Google Scholar]
- 27.Woodside J, McKinley MC, Young IS. Saturated and trans fatty acids and coronary heart disease. Curr Atheroscler Rep. 2008;10:460–466. doi: 10.1007/s11883-008-0072-5. [DOI] [PubMed] [Google Scholar]
- 28.Hu FW, WC Optimal diets for prevention of coronary heart disease. JAMA. 2002;288:2569–2578. doi: 10.1001/jama.288.20.2569. [DOI] [PubMed] [Google Scholar]
- 29.Renaud S, Lanzmann-Petithoty D. Coronary heart disease: dietary links and pathogenesis. Public Health Nutr. 2001;4:459–474. doi: 10.1079/phn2001134. [DOI] [PubMed] [Google Scholar]
- 30.Meyers D. The iron hypothesis -- does iron cause atherosclerosis? Clin Cardiol. 1996;19:925–929. doi: 10.1002/clc.4960191205. [DOI] [PubMed] [Google Scholar]
- 31.Willett WC. Nutritional Epidemiology. New York, NY: Oxford University Press; 1998. p. 514. [Google Scholar]
- 32.Finnegan D. The health effects of stimulant drinks. British Nutrition Foundation Nutrition Bulletin. 2003;28:147–155. [Google Scholar]
- 33.Schuller-Levis GB, Park E. Taurine: new implications for an old amino acid. FEMS Microbiology Letters. 2003;226:195–202. doi: 10.1016/S0378-1097(03)00611-6. [DOI] [PubMed] [Google Scholar]
- 34.Holvoet P, Harris TB, Tracy RP, Verhamme P, Newman AB, et al. Association of high coronary heart disease risk status with circulating oxidized LDL in the well-functioning elderly: findings from the Health, Aging, and Body Composition study. Arterioscler Thromb Vasc Biol. 2003;23:1444–1448. doi: 10.1161/01.ATV.0000080379.05071.22. [DOI] [PubMed] [Google Scholar]
- 35.Wu T, Willett WC, Rifai N, Shai I, Manson JE, Rimm EB. Is plasma oxidized low-density lipoprotein, measured with the widely used antibody 4E6, an independent predictor of coronary heart disease among US men and women? J Am Coll Cardiol. 2006;48:973–979. doi: 10.1016/j.jacc.2006.03.057. [DOI] [PubMed] [Google Scholar]
- 36.Tan B, Jiang D, Huang H, Jia J, Jiang HuC, Li Y. Taurine protects against low-density lipoprotein-induced endothelial dysfunction by the DDAH/ADMA pathway. Vascular Pharmacology. 2007;46:338–345. doi: 10.1016/j.vph.2006.11.006. [DOI] [PubMed] [Google Scholar]
- 37.Matsushima Y, Sekine T, Kondo T, Kameo K, Tachibana M, Murakami S. Effects of taurine on serum cholesterol levels and development of atherosclerosis in spontaneously hyperlipidaemic mice. Clin Exp Pharmacol Physiol. 2003;30:295–299. doi: 10.1046/j.1440-1681.2003.03828.x. [DOI] [PubMed] [Google Scholar]
- 38.Laidlaw SA, Grosvenor M, Kopple JD. The taurine content of common foodstuffs. J Parenter Enteral Nutr. 1990;14:183–188. doi: 10.1177/0148607190014002183. [DOI] [PubMed] [Google Scholar]
- 39.Pasantes-Morales H, Quesada O, Alcocer L, Sanchez Olea R. Taurine content in foods. Nutr Rep Int. 1989;40:793–801. [Google Scholar]
- 40.Lourenço R, Camilo ME. Taurine: a conditionally essential amino acid in humans? An Overview in health and disease. Nutr Hosp. 2002;17:262–270. [PubMed] [Google Scholar]
- 41.Kim ESKJ, Yim MH, Joeng Y, Ko YS, Watanabe T, Nakatsuka H, Nakatsuka S, Matsuda-Inoguchi N, Shimbo S, Ikeda M. Dietary taurine intake and serum taurine levels of women on Jeju Island. Adv Exp Med Biol. 2003:277–283. doi: 10.1007/978-1-4615-0077-3_35. [DOI] [PubMed] [Google Scholar]
- 42.Wójcik OP, Koenig KL, Zeleniuch-Jacquotte A, Costa M, Chen Y. The potential protective effects of taurine on coronary heart disease. Atherosclerosis. 2010;208:19–25. doi: 10.1016/j.atherosclerosis.2009.06.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Elizarova E, Nedosugova LV. First experiments in taurine administration for diabetes mellitus - the effect of erythrocyte membranes. Adv Exp Med Biol. 1996;403:583–588. doi: 10.1007/978-1-4899-0182-8_63. [DOI] [PubMed] [Google Scholar]
- 44.Koenig K, Toniolo P, Bruning P, Bonfrer J, Shore R, Pasternack B. Reliability of serum prolactin measurements in women. Cancer Epidemiol Biomarker Prev. 1993;2:411–414. [PubMed] [Google Scholar]
- 45.Martin L, Leff M, Calonge N, Garrett C, Nelson DE. Validation of self-reported chronic conditions and health services in managed care ppopulations. Am J Prev Med. 2000;18:215–218. doi: 10.1016/s0749-3797(99)00158-0. [DOI] [PubMed] [Google Scholar]
- 46.Okura Y, Urban LH, Mahoney DW, Jacobsen SJ, Rodeheffer RJ. Agreement between self-report questionnaires and medical record data was substantial for diabetes, hypertension, myocardial infarction and stroke but not for heart failure. J Clin Epidemiol. 2004;57:1093–1103. doi: 10.1016/j.jclinepi.2004.04.005. [DOI] [PubMed] [Google Scholar]