|Home | About | Journals | Submit | Contact Us | Français|
Total intake of trans fat is associated with coronary heart disease (CHD) and recent reports in primarily male populations suggest that blood levels of specific trans isomers may have different effects on risk, particularly risk of sudden cardiac death (SCD).
We prospectively examined the association between dietary intake of trans fat and SCD among 86,762 women from the Nurses’ Health Study. CHD risk factors, including diet and lifestyle factors were updated via questionnaires every 2–4 years, beginning in 1980.
Over 26 years, we documented 317 SCD events. In the primary analysis, we found no significant association between intake of total trans fat, trans-18:1 or trans-18:2 isomers and risk of SCD. Compared to the lowest quintile of intake, the relative risk (RR) (95% CI) of SCD in the highest quintile was 1.28 (0.82, 2.00) for total trans, 1.08 (0.64, 1.83) for trans-18:1 and 1.19 (0.76, 1.88) for trans-18:2. In a secondary pre-specified analysis, total trans fat was significantly related to SCD among women who reported a diagnosis of CHD prior to SCD (RR: 3.24; 95% CI: 1.42, 7.40 for the highest versus lowest quintile, P trend=0.01); however, the test for interaction was not significant (P=0.11)
In this large prospective cohort of women, neither dietary intake of trans fat, nor the individual trans isomers, trans-18:1 and trans-18:2, were significantly associated with risk of SCD. However, trans fat intake may be associated with SCD risk among women with CHD, suggesting that trans fat intake may play a greater role in SCD risk among those with clinically manifest atherosclerosis.
Sudden cardiac death (SCD) accounts for approximately half of all CHD deaths, with over 50% of the cases occurring without any previously diagnosed CHD1. The majority of SCD are due to lethal ventricular arrhythmias2,3, thus factors that influence myocardial electrical stability may influence SCD risk4. Fatty acids have varying effects on the propensity for arrhythmias, influenced by their structural properties, like chain length and saturation5. Omega-3 fatty acids, which have anti-arrhythmic properties6, are associated with lower risk of SCD7.
Intake of trans fat, mainly from hydrogenated vegetable oils, is associated with higher risk of CHD8–10. Trans fatty acids have pro-arrhythmic properties11, which may increase risk of SCD, although this relationship is not well-established. In two recent studies, composed primarily of men, blood levels of trans-18:2 isomers, but not trans-18:1, were positively associated with risk of SCD12,13. Because trans fat may have different effects on SCD risk in women, we prospectively investigated the relationship between intake of total trans fat and the individual trans-18:1 and trans-18:2 isomers and risk of SCD in women.
The Nurses’ Health study is a prospective cohort of 121,700 registered female nurses, aged 30 to 55 y at baseline in 197614. We assess information on medical history, CHD risk factors, lifestyle factors and newly diagnosed disease through self-administered questionnaires biennially. Beginning in 1980, and approximately every 4 years, participants completed a food frequency questionnaire (FFQ). For this analysis, we excluded women who did not complete the 1980 FFQ, left ≥10 food items blank or had implausible energy intake (<600 or >3500 kcal/day). We also excluded women who reported a history of stroke or cancer prior to 1980. The Institutional Review Board of Brigham and Women’s Hospital approved the study protocol.
We collected information on usual diet using a FFQ administered approximately every 4 years. For each food item, a portion size was specified and participants were asked how often, on average, she had consumed that quantity over the past year. Average intake of trans fat and other nutrients was calculated by multiplying the frequency of consumption of each food by its nutrient content and then summing across all foods. Nutrient values were obtained from the Harvard University Food Composition Database15, which accounts for types of margarine and fats used in cooking and baking. Intake of trans fat estimated from the FFQ is significantly correlated with trans fat concentrations in adipose tissue (r=0.51; p<0.001)16 and RBC (r=0.43; p<0.01)17.
On the baseline questionnaire, we collected information on weight, height, smoking status, parental history of MI, menopausal status, use of medications (aspirin and hormone therapy), dietary supplements, physical activity and personal history of CHD and other diseases. This information, with the exception of height and parental history, has been updated on biennial follow-up questionnaires. Details on the validity of the questionnaire has been reported elsewhere18–20.
Details for the classification of SCD have been described elsewhere3. Briefly, cardiac deaths were considered sudden if the death or cardiac arrest occurred within 1 hour of symptom onset as documented by medical records or through reports from next of kin. We included unwitnessed deaths that could have occurred within 1 hour of symptom onset and that had autopsy findings consistent with SCD in our analysis (n=36). To increase the specificity for an “arrhythmic death”, we excluded women with evidence of circulatory collapse (hypotension, exacerbation of congestive heart failure or neurologic dysfunction) before disappearance of the pulse, based on the definition of Hinkle and Thaler2. This rigorous definition of SCD is highly specific for arrhythmic death and systematically excludes most unwitnessed deaths and deaths during sleep. In a secondary analysis, we included unwitnessed deaths and deaths that occurred during sleep where the participant was documented to be symptom free when last observed within the preceding 24 hours and where circumstances suggested that the death could have been sudden (n=104).
For each woman, person-months of follow-up were calculated from the date of return of the 1980 questionnaire until date of death or June 1, 2006, whichever came first. We postulated that if dietary trans fat influenced the risk of SCD, the mechanism of action would most likely be through pro-arrhythmic effects, thus most recent diet most likely influences risk. In primary analysis, we used the most recent diet prior to event21, where 1980 diet was used for the 1980–1984 follow-up period, and 1984 diet was used for the 1984–1986 follow-up period, and so on. Because dietary trans fat may influence SCD risk through more long-term mechanisms related to atherogenesis, in secondary analysis we used the cumulative average of the diet to represent long-term effects and reduce random measurement error21. Because participants may change their diet after diagnosis of intermediate conditions, we stopped updating dietary information in the cumulative average estimate after new diagnoses of nonfatal MI, nonfatal stroke, coronary revascularization, angina, diabetes, hypertension, hypercholesterolemia or transient ischemic attack.
We used Cox proportional hazards models to estimate the relative risk across quintiles of trans fat intake. Models were adjusted for age (in months) and calendar year and adjusted for calories (continuous), smoking (5 categories), BMI (<25, 25–29.9, 30+ kg/m2), family history of CHD (no, prior to 60 years, 60 years or older), menopausal status, use of postmenopausal hormones (current, past, never), aspirin use (<1, 1–6, 7+/week), multivitamin and vitamin E supplements, moderate to vigorous activity (hours/week), alcohol intake (0, 0.1–4.9, 5–14.9 and 15+ g/day), intake of fatty acids (long-chain omega-3, alpha-linolenic and ratio of polyunsaturated to saturated) (in quintiles) and diagnosis of CHD, stroke, diabetes, hypertension and hypercholesterolemia. All variables were included as time-varying covariates. We conducted tests for linear trend by assigning the median value to each quintile and modeling this variable as a continuous variable.
To estimate the impact of substituting a specific percentage of energy from trans fat for the same percentage of energy from carbohydrates, we modeled intake of trans fat, protein and all major types of fat (saturated, monounsaturated, and omega-6 and omega-3 polyunsaturated fats) as continuous variables simultaneously in the model. To estimate the relative risk of substituting energy from trans fat for another fatty acid, we used the difference between the regression coefficients from the same model and calculated the 95% CIs by using the covariance of the regression coefficients22.
In pre-specified secondary analysis, we explored whether the relation between trans fat and SCD differed according to presence or absence of CHD (angina, MI, or coronary revascularization), age (<60, ≥60 y) and intake of alpha-linolenic acid. To test formally for interaction, we modeled the cross-product term between trans fat and the variable of interest (CHD, age or alpha-linolenic acid) and used a likelihood ratio test, comparing models with and without the interaction term. For all stratified analyses, we used the most recent dietary intake. All statistical analysis was performed using SAS software, version 9.1.3 (SAS Institute Inc, Cary NC). This work was funded by grants from the NIH and American Heart Association. The authors are solely responsible for the design, conduct and analyses of the study and the drafting and editing of the manuscript, and its final contents.
The median intake of trans-18:1 isomers was 1.1% of energy [interquartile range (IQR): 0.8%, 1.7%] while the median intake of trans-18:2 isomers was 0.20% of energy (IQR: 0.15%, 0.26%). The median intake (IQR) of trans fat during the first half of follow-up (1980–1994) was 1.7% (1.3%, 2.3%) of total energy, but only 1.2% (0.9%, 1.5%) during the second half of follow-up (1994–2006).
Women who consumed higher amounts of trans fat tended to consume greater amounts of all other types of fat, with the exception of the omega-3 fatty acids docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) (table 1). Women with higher trans fat intake were less likely to report a diagnosis of hypercholesterolemia, take vitamin supplements and exercise and more likely to smoke.
Over 26 years of follow up, we documented 317 cases of sudden cardiac death. Using the simple update method, trans fat intake was positively associated with risk of SCD after adjusting for age and total energy (P-trend=0.01) but this association was attenuated and no longer significant after adjustment for CHD risk factors (table 2). Findings were similar when we used the cumulative average of trans fat intake (multivariate RR: 1.23; 95%CI: 0.81, 1.87). These results were not materially altered when deaths that occurred during sleep or otherwise were unwitnessed and symptom free within 24 hours of death were included (multivariate RR: 1.00; 95%CI: 0.67, 1.48).
When modeled as a continuous variable, the replacement of 2% of energy of carbohydrates with trans fat was not significantly associated with risk of SCD after adjustment for CHD risk factors (Table 3). Results were similar for the isocaloric substitution of most major fat subtypes for trans fat. A nonsignificant trend toward higher risk of SCD was seen with the substitution of omega-3 fatty acids (ALA, EPA and DHA) with trans fat; for each 1% energy, the relative risk of SCD with trans fat was 1.74 (95%CI: 0.83, 3.67). The results were similar in the first and last halves of the study follow-up period. The relative risk of SCD for the substitution of 2% of energy of carbohydrates with trans fat was 1.35 (95%CI: 0.65, 2.79) from 1980–1994 and 0.97 (95%CI: 0.50–1.90) from 1994–2006.
One potential explanation for the lack of a significant association may be the reduction in trans fat in the food supply throughout follow-up. Therefore we estimated the relative risk comparing more extreme intakes of trans fat. Compared to women with <1% of energy from trans fat, the risk of SCD among women with ≥3% of energy from trans fat (4% of the population) was higher, although not statistically significant (multivariate RR: 1.55; 95%CI: 0.80, 3.01).
We found a significant positive association between dietary intake of the trans-18:1 isomer and risk of SCD after adjustment for age and total energy (P-trend=0.01), which was attenuated after adjustment for CHD risk factors (table 4). Dietary intake of the trans-18:2 isomer was not associated with risk of SCD. For each 0.25% of total energy increment, intake of trans-18:1 fat was associated with an adjusted relative risk of 1.00 (95%CI: 0.94, 1.07) while intake of trans-18:2 fat was associated with an adjusted relative risk of 1.28 (95%CI: 0.89, 1.83).
In secondary analyses, total trans fat intake was significantly associated with risk of SCD even after adjustment for CHD risk factors among the sub-group of women with clinically diagnosed CHD (Table 5). In this subgroup, women in the highest quintile of trans fat intake had a relative risk of SCD of 3.24 (95% CI, 1.42, 7.40) compared to women in the lowest quintile of intake. The relative risk for the substitution of 1% of energy from omega-3 fat for trans fat was 5.70 (95%CI: 1.30, 25.0). In contrast, trans fat intake was not associated with risk of SCD among women without a history CHD (Table 5). The statistical power to detect an interaction by history of CHD was limited, with only 100 cases among women with prior CHD (P-interaction=0.11). We found no evidence for interactions between trans fat intake and age (P-interaction=0.31) or intake of alpha-linolenic acid (P-interaction=0.45).
In this large prospective cohort of women, neither intake of trans fat, nor the individual trans isomers, trans-18:1 and trans-18:2, were significantly associated with risk of SCD. Total trans fat intake was associated with a three-fold elevated SCD risk among women with a prior history of CHD, after adjustment for classic CHD risk factors and other important dietary variables; however, the interaction was not significant and this analysis is limited by the small numbers of women with a prior history of CHD in this generally healthy cohort.
Our nonsignificant results for dietary intake appear at odds with the previously published literature on biomarkers of trans fat and SCD risk. In a population-based case-control study, RBC concentration of trans-18:2 isomers, but not trans-18:1, was positively associated with risk of SCD (odds ratio for IQR = 3.1; 95%CI: 1.7, 5.4)13. In the Cardiovascular Health Study, higher plasma levels of trans-18:2 isomers were associated with a doubling in risk (odds ratio for IQR = 2.3; 95%CI: 1.3, 4.3) while plasma trans-18:1 was inversely associated with risk of SCD12. Finally, in an autopsy study, adipose levels of trans-18:1 isomers were inversely associated with SCD while adipose levels of trans-18:2 isomer were not associated with SCD risk; however, the number of cases was small (n=66)23.
There are several potential explanations for our apparent disparate results regarding dietary intake compared to previous studies using biomarkers. First, the trans-18:2 isomer, which was associated with SCD in previous studies, comprises a minority of trans fat in diet, and many of the same food sources contain both trans isomers. Our ability to discriminate fully between intake of trans-18:1 and 18:2 isomers from our dietary questionnaire may be limited. Additionally, the level of trans fats in the cell membranes, such as RBC, is determined by many factors including dietary intake, absorption and metabolism. Biomarkers may approximate better the true biologic effect of trans fat, thus dietary assessment of intake may underestimate the true association.
Secondly, prior studies of trans fat and risk of SCD were conducted in primarily male populations13 or in older populations where the majority of participants had a history of CHD12. Therefore, the underlying disease process in the majority of these cases is most likely CHD, which may not be the case in this younger population of women, where the prevalence of underlying CHD is typically lower3, 24–26. The association between trans fat and SCD was statistically significant in secondary analyses among women with prior CHD, which supports the possibility that trans fat intake plays a greater role in SCD risk in the setting of CHD.
A third potential explanation for the null association may be the lower level of trans fat intake in this population. When we looked at a more extreme comparison of trans fat intake (≥3% v. <1% of energy), the risk of SCD was higher than in the quintile analysis. After a reduction of trans fat in the food supply of Costa Rica, the relationship between trans fat and nonfatal MI was no longer significant27. The effect of trans fat may be even greater in countries like Iran, where average trans fat intake was estimated at 4.2% of energy28. Future research is needed on the association between trans fat and SCD in countries where intake remains high.
Additional limitations of this study warrant discussion. Although we attempted to control for potential confounding variables, the possibility of residual confounding remains. We had more cases of SCD than prior studies, however, the numbers of SCD in this study were lower than the number of CHD events in studies reporting positive associations between trans fat and total CHD8–10, and therefore our power to detect modest associations and examine interactions was limited. However, the association between trans fat and SCD for a substitution of 2% of energy from trans fat (RR=1.20; 95%CI: 0.52, 1.91) was similar to the association with risk of CHD (RR=1.23, 95%CI: 1.11 to 1.37)29. The selective nature of this cohort, US female registered nurses, may limit the generalizability of these findings, although the high level of education and health interest of these participants, and the accuracy of reported dietary information has been well documented30. Another strength is the repeated assessment of diet, which captured the changes in trans fat in the food supply over time.
In conclusion, we found no significant association between trans fatty acid intake and risk of SCD among a large-prospective cohort of women. However, higher dietary intake of trans fat may be associated with an increased risk of SCD among women with established coronary heart disease. Future research is needed to determine whether differences in study results are due to gender differences, other clinical characteristics of participants, or differences in exposure level or endogenous trans-fat metabolism.
This study was supported by NIH grants CA-87969 and HL-34594, a Lerner Research Young Investigator Award and an Established Investigator Award from the American Heart Association (Dr. Albert).
Disclosures: The authors report no conflict of interest
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.