Although previous studies have suggested associations between plasma palmitoleic acid and coronary heart disease (CHD) risk factors, including blood pressure, inflammation, and insulin resistance, little is known about the relation of palmitoleic acid and CHD. This ancillary study of the Physicians’ Health Study was designed to examine whether red blood cell (RBC) membrane cis- palmitoleic acid and cis -vaccenic acid—2 fatty acids that can be synthesized endogenously—are associated with CHD risk. We used a risk set sampling method to prospectively select 1,000 incident CHD events and 1,000 matched controls. RBC membrane fatty acids were measured using gas chromatography. The CHD cases were ascertained using an annual follow-up questionnaire and validated by an End Point Committee through a review of the medical records. In a conditional logistic regression analysis adjusting for demographics, anthropometric, lifestyle factors, and co-morbidity, the odds ratios and 95% confidence intervals (CIs) for CHD were 1.0 (referent), 1.29 (95% CI 0.95 to 1.75), 1.08 (95% CI 0.78 to 1.51), 1.25 (95% CI 0.90 to 1.75), and 1.48 (95% CI 1.03 to 2.14) across consecutive quintiles of RBC membrane cis -palmitoleic acid (p for trend = 0.041). The odds ratio associated with each SD higher RBC membrane cis -palmitoleic acid level was 1.19 (95% CI 1.06 to 1.35) in a multivariate-adjusted model. Finally, RBC membrane cis -vaccenic acid was inversely associated with CHD risk (odds ratio 0.79, 95% CI 0.69 to 0.91, per SD increase). In conclusion, our data showed a positive association between RBC membrane cis- palmitoleic acid and CHD risk in male physicians. Furthermore, RBC membrane cis -vaccenic acid was inversely related to CHD.
Despite a decline in mortality from coronary heart disease (CHD) from its peak in the late 1960s, CHD still carries a disproportionately high burden in the United States. Data from published studies on the role of various fatty acids on the risk of developing CHD have been inconsistent or limited, especially for fatty acids that can be synthesized endogenously by de novo lipogenesis (DNL) pathway (also referred to as fatty acid synthesis). DNL reflects the body’s adaptation to handle high-carbohydrate loads through conversion of excess carbohydrate to fatty acids and triglycerides. DNL pathway involves endogenous use of acetyl-coenzyme A (CoA) to produce saturated fatty acids (i.e., 16:0 and 18:0), which can then be elongated and desaturated to generate other fatty acids such as palmitoleic acid (16:1n-7), oleic acid (18:1n-9), or vaccenic acid (18:1n-7). Heightened hepatic DNL can lead to hypertriglyceridemia, with subsequent progression of atherosclerosis and development of CHD. Although emerging data suggest that these fatty acids generated from DNL can influence the risk of sudden death or CHD risk factors, including diabetes, hypertension, and inflammation, only limited and inconsistent data are available on the influence of monounsaturated fatty acids 16:1n-7 and 18:1n-7 on the risk of CHD or CHD mortality. Wang et al reported no association between plasma phospholipids or cholesterol ester fraction of palmitoleic acid and incident CHD. Data from the Atherosclerosis Risk In Communities (ARIC) study found a positive association between plasma palmitoleic acid and incident diabetes. Because diabetes is associated with a two- to fourfold increased risk of CHD, these data suggest that palmitoleic acid could influence the risk of CHD. The present ancillary study sought to examine associations between red blood cell (RBC) membrane cis -palmitoleic acid and cis -vaccenic acid and CHD risk among United States male physicians.
Methods
The participants in these analyses were members of the Physicians Health Study (PHS) I and II who provided blood samples from 1995 to 2001. The PHS I is a completed randomized trial designed to study the effects of low-dose aspirin and β-carotene on cardiovascular disease and cancer. The PHS II is a randomized trial (started in 1997) designed to study the effects of various vitamins on the risk of cardiovascular disease and cancer.
In the present ancillary study, we used a prospective nested case-control design with a density sampling technique to randomly select 1,000 pairs of incident CHD cases and matching controls among the PHS participants who provided blood samples from 1995 to 2001. In brief, for each CHD case, we randomly selected 1 control from among the participants who were alive and free of confirmed CHD at the index case diagnosis and matched by age at blood collection (within 1 year), year of birth (within 2 years), and point of blood collection (within 3 months). Each case was eligible to serve as a control before the CHD diagnosis. Likewise, each control was eligible to later become a CHD case to ensure that the control series was a representative sample of the study base that generated all the cases. Each study participant gave written informed consent, and the Brigham and Women’s Hospital institutional review board approved the study protocol.
We used stored RBCs collected from 1995 to 2001 for the fatty acid assays. All investigators, including the laboratory personnel, were unaware of the case status of all participants. The RBC fatty acid profiles were quantified using an established gas chromatography method, as previously described. In brief, lipids were extracted from the RBC membranes, followed by saponification and methylation. The resultant fatty acid methyl esters were analyzed using an Autosystem XL gas chromatograph (PerkinElmer, Boston, Massachusetts) equipped with a 100 m × 0.25 mm ID (film thickness 0.25 μm) capillary column (SP-2560, Supelco, Sigma Aldrich, St. Louis, Missouri). The peaks of interest were identified by comparison with authentic fatty acid standards (Nu-Chek Prep, Elysian, Minnesota) and are expressed as the molar percentage proportions of total fatty acids. The interassay coefficients of variation were <4.5% for fatty acids present at levels >1 mol% and <7.1% for fatty acids present at levels <1 mol%.
A detailed description of cardiovascular events in the PHS has been previously published. In brief, we used annual follow-up questionnaires to initially obtain information on incident CHD, defined in the present study as nonfatal or fatal myocardial infarction, percutaneous transluminal coronary angioplasty, or coronary artery bypass graft. All CHD events were adjudicated by an End Point Committee by a review of medical records.
The demographic information was obtained through patient self-report. At baseline, each subject provided information on exercise, smoking (never, former, and current), and alcohol intake. The self-reported baseline weight and height were used to compute the body mass index (weight in kilograms divided by height in square meters). Information on co-morbidity (i.e., hypertension, diabetes, hypercholesterolemia, heart failure, atrial fibrillation, left ventricular hypertrophy) was collected at baseline and through follow-up questionnaires. A validated food frequency questionnaire was used to collect data on the macro- and micronutrients at the point of blood collection (1997 to 2001).
We used the distribution of RBC membrane cis -palmitoleic and cis -vaccenic acids in the control series (n = 1,000) to create the respective quintiles of cis -palmitoleic and cis -vaccenic acids. The mean values and percentages of baseline characteristics of the study participants are presented according to these quintiles. We used conditional logistic regression analysis to estimate the relative risk of CHD using the lowest quintile of each main exposure as the reference category. The initial model adjusted for matching variables (e.g., race, age, point of blood collection, and year of birth). A final model also controlled for body mass index, prevalent hypertension, hypercholesterolemia, atrial fibrillation, diabetes, left ventricular hypertrophy, smoking (never, former, current), alcohol intake (<1, 1 to 6, and ≥7 drinks/wk), physical activity (rarely/never, 1 to 3 times/month, 1 to 4 times/week, and ≥5 times/week), RBC membrane 18:0 and 16:0, and marine omega-3 fatty acids (20:5n-3, 22:5n-3, and 22:6n-3). To obtain a p value for linear trend, we used a new continuous variable that was assigned a median value of cis -palmitoleic acid or cis -vaccenic acid in the regression model. We repeated these analyses using cis -palmitoleic or cis -vaccenic acid as a continuous variable and estimated the relative risk of CHD associated with 1 SD increase of log-transformed cis -palmitoleic or cis -vaccenic acid.
In the secondary analyses, we evaluated the association between steaoryl-CoA desaturase activity, estimated by calculating the ratio of the product to the substrate (16:1n-7/16:0 and 18:1n/18:0) and elongase activity (18:1n-7/16:1n-7 ratio) and the risk of CHD, using each SD increase in each exposure. Because excess carbohydrate and protein intake can increase hepatic DNL, we conducted a sensitivity analysis by repeating the main analysis with additional adjustment for energy intake, whole grain intake, energy from carbohydrate and protein (about 326 participants had missing data on macronutrients). All analyses were performed using SAS, version 9.3 (SAS Institute, Cary, North Carolina), and the α level was set at 0.05. All p values were 2-sided.
Results
The mean age ± SD was 68.7 ± 8.7 years (range 50.4 to 92.0) among the 2,000 study participants. In the control series, the median RBC concentration of cis -palmitoleic and cis -vaccenic acids was 0.49% (interquintile range 0.37% to 0.65%) and 1.69% (interqintile range 1.47% to 2.00%) of the total RBC membrane fatty acid levels, respectively. Compared to those in the lowest quintile, those with a higher quintile of cis -palmitoleic acid had a greater body mass index; greater energy intake; greater stearoyl-CoA desaturase activity; greater oleic, palmitic, and cis -vaccenic acid levels; a greater prevalence of hypertension, atrial fibrillation, and hypercholesterolemia and a lower concentration of RBC marine omega-3 fatty acids and elongase activity and were less likely to exercise ( Table 1 ). Greater RBC cis -vaccenic acid levels was associated with a lower prevalence of diabetes and atrial fibrillation ( Table 1 ). In a conditional logistic regression model adjusting for matching factors, vaccenic acid, body mass index, prevalent hypertension, hypercholesterolemia, atrial fibrillation, diabetes, left ventricular hypertrophy, RBC 16:0 and 18:0, 20:5n-3, 22:5n-3, and 22:6n-3, smoking, alcohol intake, and exercise, the odds ratio (OR) for CHD was 1.0 (referent), 1.29 (95% confidence interval [CI] 0.95 to 1.75), 1.08 (95% CI 0.78 to 1.51), 1.25 (95% CI 0.90 to 1.75), and 1.48 (95% CI 1.03 to 2.14) across consecutive quintiles of RBC cis -palmitoleic acid (p for trend = 0.041; Table 2 ). Each SD increase of cis -palmitoleic acid was associated with a 19% greater risk of CHD (95% CI 6% to 35%) in the fully adjusted model ( Table 2 ). Additional adjustment for whole grain intake, energy intake, and energy from carbohydrate and protein did not alter the association (odds ratio 1.24, 95% CI 1.07 to 1.44 per SD increase in RBC cis -palmitoleic acid).
| Characteristic | cis -Palmitoleic Acid (16:1n-7) | cis -Vaccenic Acid (18:1n-7) | ||||
|---|---|---|---|---|---|---|
| Q1 (Low; 0.28 ± 0.05) | Q3 (0.49 ± 0.02) | Q5 (High; 0.93 ± 0.23) | Q1 (Low; 1.28 ± 0.11) | Q3 (1.69 ± 0.06) | Q5 (High; 2.52 ± 0.47) | |
| Subjects (n) | 373 | 377 | 440 | 433 | 407 | 393 |
| Age (years) | 69.4 ± 8.4 | 68.4 ± 8.7 | 67.6 ± 8.4 | 68.0 ± 8.1 | 69.5 ± 8.6 | 68.1 ± 9.5 |
| Body mass index (kg/m 2 ) | 25.3 ± 3.2 | 26.0 ± 3.3 | 26.9 ± 3.9 | 26.2 ± 3.3 | 25.9 ± 3.4 | 26.2 ± 3.9 |
| cis -Vaccenic acid † | 1.59 ± 0.29 | 1.76 ± 0.43 | 2.01 ± 0.66 | 1.28 ± 0.11 | 1.69 ± 0.06 | 2.52 ± 0.47 |
| 16:1n-7/16:0 ratio | 0.013 ± 0.002 | 0.022 ± 0.002 | 0.038 ± 0.009 | 0.021 ± 0.009 | 0.023 ± 0.009 | 0.029 ± 0.011 |
| 18:1n-9/18:0 ratio | 0.766 ± 0.124 | 0.835 ± 0.134 | 0.941 ± 0.190 | 0.817 ± 0.178 | 0.867 ± 0.163 | 0.871 ± 0.142 |
| 18:1n-7/16:1n-7 ratio | 5.83 ± 2.33 | 3.58 ± 0.89 | 2.26 ± 0.83 | 3.11 ± 1.52 | 3.67 ± 1.43 | 4.19 ± 1.64 |
| Oleic acid cis (18:1n-9) | 13.4 ± 1.2 | 14.2 ± 1.4 | 15.4 ± 2.0 | 13.9 ± 1.5 | 14.4 ± 1.79 | 14.8 ± 1.8 |
| Palmitic acid (16:0) † | 21.8 ± 1.7 | 22.9 ± 2.2 | 24.6 ± 2.9 | 22.6 ± 1.9 | 23.3 ± 2.6 | 23.7 ± 2.9 |
| Stearic acid (18:0) † | 17.9 ± 3.2 | 17.3 ± 2.9 | 16.9 ± 3.5 | 17.8 ± 4.4 | 17.0 ± 2.5 | 17.2 ± 2.0 |
| Marine omega-3 † | 6.63 ± 1.91 | 6.06 ± 1.82 | 5.35 ± 1.85 | 6.43 ± 1.75 | 6.12 ± 2.02 | 5.37 ± 1.93 |
| Energy intake (kcal) ‡ | 1,672 ± 534 | 1,685 ± 547 | 1,680 ± 488 | 1,688 ± 498 | 1,682 ± 524 | 1,666 ± 507 |
| Energy from carbohydrate (%) ‡ | 51.8 ± 9.7 | 50.6 ± 9.7 | 49.3 ± 10.4 | 49.7 ± 9.8 | 50.0 ± 9.2 | 50.9 ± 10.9 |
| Energy from protein (%) ‡ | 18.7 ± 3.4 | 18.3 ± 3.3 | 17.9 ± 3.3 | 18.4 ± 3.3 | 18.5 ± 3.1 | 18.3 ± 3.6 |
| Whole grain (servings/week) ‡ | 4.71 ± 1.88 | 4.43 ± 1.85 | 4.56 ± 1.70 | 4.48 ± 1.77 | 4.47 ± 1.94 | 4.53 ± 1.79 |
| Diabetes mellitus | 10.2% | 8.0% | 8.2% | 9.7% | 9.8% | 5.9% |
| Hypertension § | 53.0% | 52.7% | 59.1% | 53.7% | 57.6% | 49.6% |
| Atrial fibrillation | 10.5% | 8.5% | 10.7% | 10.4% | 7.4% | 10.2% |
| Hypercholesterolemia § | 29.5% | 30.2% | 39.3% | 32.3% | 32.2% | 29.0% |
| Heart failure | 1.9% | 2.4% | 2.5% | 0.7% | 2.7% | 3.3% |
| Left ventricular hypertrophy | 0.8% | 3.5% | 2.1% | 1.9% | 3.0% | 2.8% |
| Current smoking | 0.8% | 1.9% | 5.7% | 2.1% | 4.9% | 3.3% |
| Current drinking | 71.3% | 81.4% | 87.0% | 80.6% | 79.4% | 80.7% |
| Current exercise ‡ | 65.0% | 61.9% | 60.5% | 62.6% | 62.2% | 66.1% |
⁎ We used the quintile distribution of fatty acids in the control series.
† Fatty acids are expressed as molar percentage of total red blood cell membrane fatty acids.
‡ Data were missing for energy, carbohydrate, or protein intake (n = 326), whole grain intake (n = 246), and exercise (n = 22).
§ Hypertension was defined as systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg or the use of antihypertensive drugs; hypercholesterolemia was defined as total cholesterol ≥200 mg/dl or the use of cholesterol-lowering drugs.
| Quintiles of cis -Palmitoleic Acid (16:1n-7) | Cases (n) | OR (95% CI) for CHD | |
|---|---|---|---|
| Model 1 ⁎ | Model 2 † | ||
| Q1 (low; 0.04–0.38) ‡ | 173 | 1.0 | 1.0 |
| Q2 (0.39–0.48) | 205 | 1.23 (0.92–1.64) | 1.29 (0.95–1.75) |
| Q3 (0.49–0.59) | 177 | 1.10 (0.82–1.49) | 1.08 (0.78–1.51) |
| Q4 (0.60–0.90) | 205 | 1.29 (0.95–1.74) | 1.25 (0.90–1.75) |
| Q5 (high; >0.91) | 240 | 1.58 (1.16–2.15) | 1.48 (1.03–2.14) |
| p for Trend | 0.003 | 0.04 | |
| Per SD increase of log-palmitoleic acid ( cis 16:1n-7) | 1.20 (1.09–1.32) | 1.19 (1.06–1.35) | |
⁎ Model 1 adjusted for matching variables and cis-vaccenic acid.
† Model 2 adjusted for matching variables plus body mass index, prevalent hypertension, hypercholesterolemia, atrial fibrillation, diabetes, left ventricular hypertrophy, smoking, alcohol intake, physical activity, and red blood cell membrane 18:0 and 16:0, and marine omega-3 fatty acids (20:5n-3, 22:5n-3, and 22:6n-3).
‡ Fatty acids are expressed as percentage of total red blood cell membrane fatty acids.
RBC cis -vaccenic acid was inversely associated with the risk of CHD in a stepwise fashion (p for linear trend = 0.007; Table 3 ). Each SD increase in cis -vaccenic acid was associated with a 21% lower odds of CHD (95% CI 9% to 31%), in a fully adjusted model ( Table 3 ). Additional adjustment for energy, carbohydrate, protein, and whole grain intake did not alter the relation (OR 0.82, 95% CI 0.69 to 0.98). In secondary analyses, we examined the relation of stearoyl-CoA desaturase and elongase activities (estimated using the product/substrate ratio) and observed a positive association for the 16:1n-7/16:0 ratio with CHD. Each SD of stearoyl CoA desaturase activity was associated with a 13% greater risk of CHD (95% CI 2% to 26%) in a fully adjusted model ( Table 4 ). The relation of the 18:1n-9/18:0 ratio with CHD risk did not reach statistical significance in the full model ( Table 4 ). Finally, an inverse association was seen between elongase activity (18:1n-7/16:1n-7 ratio) and CHD risk (OR 0.81, 95% CI 0.73 to 0.91 per SD increase; Table 4 ).
