Abstract
BACKGROUND
Previous epidemiological studies have generated inconsistent results regarding the associations between dietary fat intakes and risk for Parkinson’s disease (PD). We therefore prospectively examined these associations in the NIH-AARP Diet and Health Study.
METHODS
A 124-item food frequency questionnaire was administered at baseline in 1995–1996, and PD diagnosis was self-reported at the follow-up survey in 2004–2006. A total of 1,087 cases with a PD diagnosis between 2000 and 2006 and 299,617 controls were included in the analyses.
RESULTS
Overall, intakes of fats and other macronutrients were not associated with PD risk. However, we found a weak positive association between n-6 polyunsaturated fatty acids (PUFA) and the risk for PD. After adjusting for potential confounders, the odds ratio (OR) and 95% confidence interval (CI) between extreme quintiles of n-6 PUFA intake was 1.23 (95% CI=1.02–1.49, P for trend=0.02). A similar association was observed for the intake of linoleic acid. Results were similar among men and among women.
CONCLUSIONS
Our study suggests that fat intake in general is not related to the risk for PD. The weak positive association between intake of n-6 PUFA and PD risk needs further investigation.
Keywords: Dietary fat intake, Parkinson’s disease, cohort study
Parkinson's disease (PD) is the second most common neurodegenerative disease and affects more than one million Americans 1. Although the exact mechanisms underlying PD pathogenesis are yet to be defined; oxidative stress, mitochondrial dysfunction, and inflammation may contribute to this process2–4. The brain is particularly susceptible to oxidative damage because of its high demand for oxygen and the presence of abundant iron and unsaturated fatty acids 5. In addition, polyunsaturated fatty acids (PUFA), which are essential components of neuronal and glial cell membranes, regulate the production of pro/anti-inflammatory cytokines that may also contribute to neurodegenerative diseases such as PD 6.
Murine studies have shown that brain concentrations of fatty acids, such as PUFA, may be affected by dietary intake7, 8. Epidemiological studies on dietary fat intake and PD have generated inconsistent results, ranging from positive associations 9–12, to null 13, 14 or even inverse associations 15–20. We therefore prospectively examined the relationships between PD risk and intakes of total fat, specific types of fat, and other macronutrients in the NIH-AARP Diet and Health Study.
Patients and Methods
Study population and PD case identification
The details of the NIH-AARP Diet and Health Study have been described previ-ously21. Briefly, in 1995–1996, a total of 566,398 members of AARP (formerly known as the American Association of Retired Persons), aged 50 to 71 years, were recruited into a large prospective study to evaluate roles of diet and lifestyle in the development of cancers and other chronic diseases. Participants were from six US states (California, Florida, Louisiana, New Jersey, North Carolina, and Pennsylvania) and two metropolitan areas (Atlanta, Georgia, and Detroit, Michigan). At enrollment, they reported a wide range of dietary intakes of macro- and micronutrients 21. From 2004 to 2006, a follow-up survey was conducted among surviving participants to update lifestyle exposures and to ascertain the occurrence of major chronic diseases, including PD. Participants were asked whether they had been diagnosed with PD by a physician and, if so, the year of diagnosis in the following categories: before 1985, 1985–1994, 1995–1999, or during or after 2000. A total of 318,257 men and women responded to the follow-up survey and 2,432 participants reported a PD diagnosis.
Between 2007 and 2010, we contacted surviving PD patients to confirm self-reported PD diagnoses. Detailed procedures were published previously 22. Briefly, we first asked patients who self-reported PD to confirm their reports. We then asked their treating physician to complete a diagnostic questionnaire and to provide a copy of the patient’s medical records. The medical records were subsequently reviewed by a movement disorder specialist (X. H.). A case was confirmed if one of the following criteria was met: (1) the diagnosis was confirmed by the treating physician, (2) the medical record included a final diagnosis of PD, or (3) at least two cardinal signs were present with one being resting tremor or bradykinesia, the disease course was progressive, and there was an absence of unresponsiveness to levodopa or other features suggesting an alternative diagnosis. Of the 1,069 physician responses received, 940 (87.9%) PD diagnoses were confirmed. The confirmation rate was similar across years of diagnosis: 83.3% for cases diagnosed before 1985, 92.8% for cases diagnosed in 1985–1994, 87.9% for cases diagnosed in 1995–1999, and 87.2% for cases diagnosed after 2000.
We excluded 15,760 subjects whose baseline questionnaires were completed by proxy from the 566,398 members of the cohort, and then subjects who reported extreme intakes (>2 times the interquartile range above the 75th percentile or below the 25th percentile of log-transformed intake) of total energy (n=3,726) and total fat (n=631). Inclusion of participants with extreme intakes however had little effect on the results (data not shown).
After these exclusions, a total of 546,281 baseline enrollees left and 309,619 of them also participated in the follow-up survey and therefore eligible for the current analyses. We further removed subjects who didn’t answer the questions about PD status in the follow-up questionnaire (n=7,054) or were reporting errors and misdiagnoses (n=905), and PD patients whose diagnosis were denied by themselves or treat physicians in the disease validation process (n=48). Because PD may take years to develop and our previous studies indicated dietary and behavioral changes 2–4 years prior to disease diagnosis 23, 24, we further excluded 908 self-reported cases diagnosed before 2000. This left us with a total of 1,087 cases diagnosed between 2000 and 2006, and 299,617 controls in the primary analyses.
Exposure assessment
All study participants completed a mailed questionnaire at baseline that included a validated 124-item food-frequency questionnaire (FFQ) and questions on demographics and lifestyle. The FFQ asked participants to report their typical consumption frequency and portion size of 124 food items during the past year. Consumption frequency allows for 10 categories, ranging from never to two or more times per day for solid foods and never to six or more times per day for drinks. The portion size had three categories-small, medium, and large. Nutrient intakes were calculated from the 1994–1996 U.S. Department of Agriculture Continuing Survey of Food Intake by individuals 25. The FFQ was calibrated against two nonconsecutive 24-hour dietary recalls that were administered by telephone within a year of the baseline questionnaire (n =2,053) 21. The estimated energy-adjusted Pearson correlation coefficients for men and women, respectively, were 0.72 and 0.62 for total fat, 0.76 and 0.69 for saturated fat, 0.71 and 0.62 for monounsaturated fat and 0.53 and 0.56 for polyunsaturated fat 21, 26. Caffeine intake was estimated from consumption of coffee and other caffeine-containing drinks and foods as reported on the FFQ 27. For smoking status, participants were asked whether they had ever smoked more than 100 cigarettes during their lifetime. Ever smokers were further asked about typical number of cigarettes smoked per day, current smoking status, and, for past smokers, the number of years since they last smoked27. Both current smoking and higher caffeine intake was associated with lower risk of PD in this cohort as expected22, 27.
Statistical analysis
Our primary exposures of interest were dietary intakes of total fat, individual types of fat and individual types of fatty acids. We calculated total n-6 PUFA intake as the sum of intakes of 18:2 and 20:4 fatty acids and total n-3 PUFA intake as the sum of intakes of 18:3, 18:4, 20:5, 22:5, and 22:6 fatty acids 28. To account for total energy intake in the analyses, we used the multivariate nutrient density method 29. Exposures of interest were first expressed as percentage of total energy intake and then categorized into quintiles with the lowest quintile as the reference group. Odds ratios (ORs) and 95% confidence intervals (CIs) were derived from multivariate logistic regression models, adjusting for age (year, continuous variable), gender, race/ethnicity (Non-Hispanic White, Non-Hispanic Black, Hispanic, Asian, Pacific Islander, or American Indian/Alaskan Native and Unknown), caffeine intake (mg/day, continuous variable), smoking status (never, ever and current), and total energy intake (Kcal/day, continuous variable).
Only 56.2% of the original cohort enrollees participated in the cohort’s follow-up survey and their baseline population characteristics were significantly different from those excluded from the analysis (data not shown). We therefore conducted additional analyses using the inverse probability weighting method to examine potential selection biases from death or non-participation in the follow-up survey 30. In doing this, we used a directed acyclic graph 31 to analyze the following potential confounders identified from previous publications: age, gender, race/ethnicity, smoking status, caffeine intake, physical activity, body mass index (BMI), education, marital status, total energy intake, vitamin E intake, iron intake, alcohol use, non-steroidal anti-inflammatory drug use, diabetes diagnosis, and self-reported health status. As a result, we identified two minimally sufficient adjustment sets: (1) age, gender, physical activity, race/ethnicity, smoking, caffeine intake, diabetes, alcohol, BMI, and total energy intake; (2) age, gender, physical activity, race/ethnicity, smoking, caffeine intake, diabetes, and self-reported health status (Supplementary Figure 1). We used the second minimally sufficient adjustment set as the final model because it had less missing data on covariates than the first.
We then used the stabilized inverse probability weights to adjust for covariates in the second minimally sufficient adjustment set and to account for non-participation in the follow-up survey 30, 32. We calculated separate stabilized weights to adjust for potential confounders and potential selection bias, respectively. For stabilized confounding weights, we calculated the numerators of these weights as predicted probabilities of exposure from an intercept only model and the denominators of these weights as predicted probabilities of exposure from a model including the covariates in the second minimally sufficient adjustment set as explanatory variables. For stabilized selection weights, we calculated the numerators of these weights as predicted probabilities of not dropping out of the cohort conditional on the exposure of interest and the denominators of these weights as predicted probabilities conditional on the exposure of interest and the covariates in the second minimally sufficient adjustment set. We then calculated the overall stabilized weight as the product of the two weights. The overall stabilized weight was finally applied to logistic regression models for PD risk that contained the exposure of interest as the only explanatory variable 30. We used robust variance estimates to calculate 95% CIs because using weights induced within-subject correlation 33.
In addition to the primary analyses, we also conducted stratified analyses according to age (< 62 and ≥ 62), sex, smoking status (never and ever) and caffeine intake (< 234 and ≥ 234 mg/day). Statistical interactions were tested by adding a multiplicative interaction term to the logistic regression model. All statistical analyses were performed using Stata 12.0 (College Station, TX, USA) and P < 0.05 was considered as statistically significant.
Standard protocol approvals, registrations and patient consent
Participants consented to the study by returning survey questionnaires. The study protocol was approved by the Institutional Review Board of the National Institute of Environmental Health Sciences and the Special Studies Institutional Review Board of the National Cancer Institute.
Results
Characteristics of study population
Table 1 presents the baseline characteristics of the study population according to quintiles of total fat intake. The median intakes of total fat as percentages of energy intake ranged from 20.1% in the lowest quintile to 39.5% in the highest quintile. Participants with higher fat intake were more likely to be men, non-Hispanic Whites and current smokers. They also were more likely to report higher intakes of caffeine, total energy, and cholesterol, but lower intake of carbohydrate. Further, they on average had a higher BMI and were more likely to report diabetes and poorer health status.
TABLE 1.
Baseline characteristics by quintiles of total fat intake as percentages of energy among participants in NIH-AARP Diet and Health Study.
| Characteristics | Q1 | Q2 | Q3 | Q4 | Q5 | |
|---|---|---|---|---|---|---|
| N participants | 60,141 | 60,141 | 60,141 | 60,141 | 60,140 | |
| Total fat intake (Median), % of total energy | 20.1 | 26.0 | 30.0 | 34.0 | 39.5 | |
| Age, years (Mean ± SD) | 61.6 ± 5.3 | 61.5 ± 5.3 | 61.5 ± 5.3 | 61.4 ± 5.3 | 61.2 ± 5.4 | |
| Men, N (%) | 33,481 (55.7) | 33,485 (55.7) | 35,140 (58.4) | 36,420 (60.6) | 35,765 (59.5) | |
| Race, N (%) | ||||||
| Non-Hispanic White | 55,001 (91.5) | 55,677 (92.6) | 55,741 (92.7) | 55,890 (92.9) | 56,098 (93.3) | |
| Other | 4,481 (7.5) | 3,898 (6.5) | 3,805 (6.3) | 3,640 (6.1) | 3,377 (5.6) | |
| Caffeine intake, mg/day (Mean ± SD) | 297.7 ± 339.6 | 331.1 ± 350.9 | 355.2 ± 356.7 | 386.1 ± 373.3 | 429.7 ± 398.7 | |
| Smoking status | ||||||
| never | 23,584 (41.1) | 23,863 (41.5) | 23,628 (41.1) | 22,975 (39.9) | 21,237 (36.9) | |
| past | 29,845 (52.0) | 29,344 (51.0) | 28,729 (50.0) | 28,254 (49.0) | 27,282 (47.4) | |
| current | 3,922 (6.8) | 4,327 (7.5) | 5,156 (9.0) | 6,379 (11.1) | 9,064 (15.7) | |
| Total energy intake, Kcal/day (Mean ± SD) | 1,737.6 ± 839.7 | 1,710.4 ± 703.7 | 1,792.7 ± 725.7 | 1,892.5 ± 780.1 | 2,005.5 ± 870.5 | |
| Carbohydrate intake, % of energy, (Mean ± SD) | 61.4 ± 11.4 | 56.4 ± 7.5 | 52.8 ± 6.3 | 49.4 ± 5.39 | 43.7 ± 5.6 | |
| Protein intake, % of energy (Mean ± SD) | 14.7 ± 3.6 | 15.6 ± 3.1 | 15.6 ± 3.0 | 15.6 ± 2.9 | 15.6 ± 2.9 | |
| Cholesterol intake, mg/day (Mean ± SD) | 121.5 ± 71.2 | 162.0 ± 83.0 | 194.0 ± 97.1 | 228.6 ± 115.3 | 280.2 ± 150.7 | |
| Body mass index, kg/m2 (Mean ± SD) | 25.8 ± 4.4 | 26.6 ± 4.7 | 27.0 ± 4.8 | 27.4 ± 5.0 | 27.9 ± 5.5 | |
| Self-reported diabetes, N (%) | ||||||
| No | 57,428 (95.5) | 56,612 (94.1) | 55,950 (93.0) | 55,389 (92.1) | 53,768 (89.4) | |
| Yes | 2,713 (4.5) | 3,529 (5.9) | 4,191 (7.0) | 4,752 (7.9) | 6,372 (10.6) | |
| Self-reported health status, N (%) | ||||||
| Excellent | 14,026 (23.3) | 12,079 (20.1) | 10,904 (18.1) | 10,046 (16.7) | 9,482 (15.8) | |
| Very good | 22,712 (37.8) | 23,246 (38.7) | 22,866 (38.0) | 22,401 (37.2) | 20,983 (34.9) | |
| Good | 17,946 (29.8) | 19,119 (31.8) | 20,360 (33.9) | 21,092 (35.1) | 21,900 (36.4) | |
| Fair | 4,195 (7.0) | 4,468 (7.4) | 4,721 (7.9) | 5,354 (8.9) | 6,295 (10.5) | |
| Poor | 444 (0.7) | 436 (0.7) | 494 (0.8) | 537 (0.9) | 737 (1.2) | |
Numbers may not add up to the total due to missing.
Associations between dietary fat intake and PD risk
After adjusting for potential confounders, intakes of total energy and total fat each showed a weak positive association with PD risk with borderline significant dose-response relationship (Table 2). In the inverse probability weighted analyses, the association with total energy intake was slightly stronger whereas the one for total fat intake was modestly attenuated. Among individual types of fat, there were weak positive associations between PD risk and intakes of total PUFA and n-6 PUFA. For example, compared with the lowest quintile, the OR was 1.23 (95% CI=1.02–1.49, P for trend=0.02) for the highest quintile of n-6 PUFA intake. The association was also borderline statistically significant in the inverse probability weighted analyses (P for trend=0.03). When we examined the n-3 to n-6 ratio in relation to PD risk, we found that participants in the fourth quintile had a lower risk for PD as compared with those in the first quintile (OR=0.81, 95% CI=0.38–0.98), but the P for trend was not statistically significant (P for trend=0.12). None of these associations remained statistically significant after Bonferroni correction for multiple comparisons. Intake of other types of fat or macronutrients was not associated with PD risk.
TABLE 2.
Associations between dietary intakes of fats and other macronutrients and PD risk.
| Type of fat | % Total energy |
Case, N |
Age and sex adjusted a | Multivariate adjustedb | Inverse probability weightingc |
|||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) |
P for trend |
OR (95% CI) |
P for trend |
OR (95% CI) |
P for trend |
|||
| Total fat | ≤ 23.5 | 213 | ref | 0.17 | ref | 0.06 | ref | 0.25 |
| 23.5–28.1 | 198 | 0.94 (0.77–1.14) | 0.95 (0.78–1.15) | 0.93 (0.76–1.13) | ||||
| 28.1–32.0 | 218 | 1.02 (0.84–1.23) | 1.04 (0.86–1.26) | 0.98 (0.81–1.20) | ||||
| 32.0–36.4 | 231 | 1.08 (0.89–1.30) | 1.11 (0.92–1.34) | 1.07 (0.88–1.30) | ||||
| ≥ 36.4 | 227 | 1.08 (0.90–1.31) | 1.14 (0.94–1.38) | 1.06 (0.87–1.28) | ||||
| Saturated fat | ≤ 6.81 | 220 | ref | 0.34 | ref | 0.12 | ref | 0.3 |
| 6.81–8.38 | 192 | 0.89 (0.73–1.08) | 0.90 (0.74–1.10) | 0.89 (0.73–1.09) | ||||
| 8.38–9.84 | 221 | 1.02 (0.84–1.22) | 1.04 (0.86–1.26) | 1.02 (0.84–1.24) | ||||
| 9.84–11.62 | 243 | 1.12 (0.93–1.35) | 1.16 (0.97–1.39) | 1.12 (0.93–1.35) | ||||
| ≥ 11.62 | 211 | 0.99 (0.82–1.19) | 1.04 (0.86–1.27) | 0.99 (0.82–1.21) | ||||
| Monounsaturated fatty acids | ≤ 8.6 | 211 | ref | 0.24 | ref | 0.1 | ref | 0.26 |
| 8.6–10.5 | 194 | 0.92 (0.76–1.12) | 0.94 (0.77–1.14) | 0.92 (0.75–1.12) | ||||
| 10.5–12.1 | 229 | 1.07 (0.89–1.29) | 1.09 (0.91–1.32) | 1.04 (0.86–1.26) | ||||
| 12.1–13.9 | 221 | 1.02 (0.85–1.24) | 1.06 (0.87–1.28) | 1.02 (0.84–1.24) | ||||
| ≥ 13.9 | 232 | 1.08 (0.90–1.30) | 1.13 (0.94–1.37) | 1.08 (0.89–1.30) | ||||
| Polyunsaturated fatty acids (PUFA) | ≤ 5.2 | 199 | ref | 0.05 | ref | 0.03 | ref | 0.05 |
| 5.2–6.2 | 216 | 1.09 (0.90–1.32) | 1.10 (0.91–1.33) | 1.07 (0.88–1.31) | ||||
| 6.2–7.3 | 214 | 1.07 (0.88–1.30) | 1.08 (0.89–1.31) | 1.08 (0.90–1.32) | ||||
| 7.3–8.6 | 227 | 1.14 (0.94–1.38) | 1.15 (0.95–1.40) | 1.16 (0.96–1.41) | ||||
| ≥ 8.6 | 231 | 1.20 (0.99–1.45) | 1.23 (1.01–1.48) | 1.19 (0.98–1.44) | ||||
| N-6 PUFA | ≤ 4.6 | 196 | ref | 0.03 | ref | 0.02 | ref | 0.03 |
| 4.6–5.5 | 214 | 1.09 (0.90–1.33) | 1.10 (0.91–1.34) | 1.08 (0.89–1.32) | ||||
| 5.5–6.5 | 213 | 1.08 (0.89–1.31) | 1.09 (0.90–1.32) | 1.09 (0.90–1.33) | ||||
| 6.5–7.7 | 234 | 1.19 (0.99–1.44) | 1.20 (0.99–1.46) | 1.22 (1.00–1.48) | ||||
| ≥ 7.7 | 230 | 1.21 (1.00–1.46) | 1.23 (1.02–1.49) | 1.20 (0.99–1.46) | ||||
| N-3 PUFA | ≤ 0.5 | 218 | ref | 0.73 | ref | 0.59 | ref | 0.86 |
| 0.5–0.6 | 212 | 0.98 (0.81–1.82) | 0.98 (0.81–1.19) | 1.00 (0.82–1.21) | ||||
| 0.6–0.7 | 214 | 0.99 (0.82–1.20) | 1.00 (0.83–1.21) | 1.01 (0.84–1.23) | ||||
| 0.7–0.9 | 258 | 1.22 (1.01–1.46) | 1.22 (1.02–1.47) | 1.24 (1.03–1.49) | ||||
| ≥ 0.9 | 185 | 0.91 (0.75–1.11) | 0.93 (0.76–1.14) | 0.89 (0.73–1.09) | ||||
| n-3:n-6 PUFA ratio | ≤ 0.10 | 254 | ref | 0.13 | ref | 0.12 | ref | 0.07 |
| 0.10–0.11 | 212 | 0.86 (0.71–1.03) | 0.85 (0.71–1.02) | 0.87 (0.72–1.05) | ||||
| 0.11–0.12 | 213 | 0.87 (0.72–1.04) | 0.87 (0.72–1.04) | 0.88 (0.73–1.06) | ||||
| 0.12–0.14 | 198 | 0.82 (0.68–0.99) | 0.81 (0.38–0.98) | 0.80 (0.66–0.97) | ||||
| ≥ 0.14 | 210 | 0.88 (0.73–1.05) | 0.87 (0.73–1.05) | 0.85 (0.71–1.03) | ||||
| Total energy, Kcal/day | ≤ 1,200.5 | 196 | ref | 0.24 | ref | 0.09 | ref | 0.04 |
| 1,200.5–1,524.3 | 191 | 0.90 (0.74–1.10) | 0.90 (0.74–1.11) | 0.88 (0.71–1.08) | ||||
| 1,524.3–1,860.1 | 206 | 0.92 (0.76–1.12) | 0.93 (0.76–1.13) | 0.92 (0.75–1.13) | ||||
| 1,860.1–2,346.2 | 250 | 1.07 (0.88–1.29) | 1.09 (0.90–1.32) | 1.12 (0.92–1.37) | ||||
| ≥ 2,346.2 | 244 | 1.04 (0.85–1.26) | 1.09 (0.90–1.33) | 1.12 (0.92–1.37) | ||||
| Protein | ≤ 12.9 | 212 | ref | 0.98 | ref | 0.83 | ref | 0.19 |
| 12.9–14.6 | 239 | 1.12 (0.93–1.35) | 1.12 (0.93–1.35) | 1.16 (0.96–1.41) | ||||
| 14.6–16.1 | 221 | 1.04 (0.87–1.27) | 1.04 (0.86–1.26) | 1.06 (0.87–1.29) | ||||
| 16.1–17.9 | 208 | 1.01 (0.84–1.23) | 1.00 (0.82–1.21) | 0.97 (0.80–1.19) | ||||
| ≥ 17.9 | 207 | 1.05 (0.87–1.28) | 1.04 (0.85–1.26) | 0.95 (0.78–1.16) | ||||
| Carbohydrate | ≤ 44.9 | 217 | ref | 0.77 | ref | 0.61 | ref | 0.53 |
| 44.9–50.3 | 216 | 1.02 (0.84–1.23) | 0.99 (0.82–1.20) | 0.99 (0.81–1.22) | ||||
| 50.3–55.1 | 239 | 1.15 (0.96–1.39) | 1.11 (0.92–1.34) | 1.11 (0.91–1.36) | ||||
| 55.1–60.7 | 219 | 1.08 (0.89–1.30) | 1.02 (0.84–1.24) | 1.01 (0.82–1.24) | ||||
| ≥ 60.7 | 196 | 1.00 (0.82–1.21) | 0.93 (0.76–1.13) | 0.92 (0.74–1.14) | ||||
| Cholesterol | ≤ 0.07 | 203 | ref | 0.99 | ref | 0.52 | ref | 0.94 |
| 0.07–0.09 | 235 | 1.17 (0.97–1.41) | 1.19 (0.98–1.43) | 1.16 (0.95–1.40) | ||||
| 0.09–0.11 | 237 | 1.20 (1.00–1.45) | 1.26 (1.02–1.48) | 1.18 (0.97–1.43) | ||||
| 0.11–0.14 | 198 | 1.01 (0.83–1.22) | 1.04 (0.85–1.27) | 1.00 (0.82–1.22) | ||||
| ≥ 0.14 | 214 | 1.08 (0.90–1.31) | 1.15 (0.95–1.39) | 1.07 (0.88–1.30) | ||||
Abbreviations: OR: odds ratios; CI: confidence intervals; MUFA: monounsaturated fatty acids, PUFA: polyunsaturated fatty acids
Adjusted for age and sex
Adjusted for age, sex, race, smoking status, caffeine intake and total energy intake.
Adjusted for inverse probability weights with 95% CIs calculated from robust variance estimates.
We further examined specific types of PUFA in relation to PD risk (Table 3). As expected, intake of linoleic acid showed a weak positive association with the risk for PD. No other n-6 fatty acids or n-3 fatty acids were associated with the risk for PD.
TABLE 3.
Associations between dietary intakes of specific n-6 and n-3 PUFA and PD risk.
| Type of fat | % Total energy |
Case, N |
Age and sex adjusted a | Multivariate adjustedb | Inverse probability weightingc |
|||
|---|---|---|---|---|---|---|---|---|
| OR (95% CI) |
P for trend |
OR (95% CI) |
P for trend |
OR (95% CI) |
P for trend |
|||
| n-6 PUFA | ||||||||
| Linoleic acid (18:2) | ≤ 4.5 | 196 | ref | 0.04 | ref | 0.03 | ref | 0.04 |
| 4.5–5.5 | 215 | 1.10 (0.91–1.34) | 1.11 (0.91–1.34) | 1.09 (0.90–1.33) | ||||
| 5.5–6.4 | 214 | 1.09 (0.89–1.32) | 1.09 (0.90–1.33) | 1.10 (0.90–1.34) | ||||
| 6.4–7.6 | 233 | 1.19 (0.98–1.44) | 1.20 (0.99–1.45) | 1.21 (1.00–1.47) | ||||
| ≥ 7.6 | 229 | 1.20 (0.99–1.46) | 1.23 (1.01–1.48) | 1.19 (0.98–1.45) | ||||
| Arachidonic acid (20:4) | ≤ 0.03 | 226 | ref | 0.75 | ref | 0.56 | ref | 0.71 |
| 0.03–0.04 | 222 | 0.97 (0.81–1.17) | 0.98 (0.82–1.18) | 0.99 (0.82–1.20) | ||||
| 0.04–0.05 | 209 | 0.92 (0.76–1.11) | 0.94 (0.77–1.13) | 0.94 (0.78–1.14) | ||||
| 0.05–0.07 | 206 | 0.94 (0.77–1.13) | 0.95 (0.79–1.15) | 0.91 (0.75–1.10) | ||||
| ≥ 0.07 | 224 | 1.05 (0.87–1.27) | 1.08 (0.90–1.30) | 1.00 (0.83–1.21) | ||||
| n-3 PUFA | ||||||||
| α-Linolenic acid (18:3) | ≤ 0.5 | 225 | ref | 0.88 | ref | 0.68 | ref | 1 |
| 0.5–0.6 | 212 | 0.95 (0.79–1.15) | 0.96 (0.79–1.16) | 0.97 (0.80–1.18) | ||||
| 0.6–0.7 | 206 | 0.93 (0.77–1.13) | 0.94 (0.78–1.14) | 0.94 (0.78–1.15) | ||||
| 0.7–0.8 | 255 | 1.17 (0.98–1.40) | 1.18 (0.99–1.42) | 1.19 (0.99–1.43) | ||||
| ≥ 0.8 | 189 | 0.90 (0.74–1.10) | 0.93 (0.76–1.12) | 0.89 (0.73–1.08) | ||||
| Eicosapentaenoic acid (20:5) | ≤ 0.005 | 214 | ref | 0.16 | ref | 0.22 | ref | 0.13 |
| 0.005–0.009 | 197 | 0.95 (0.79–1.16) | 0.96 (0.79–1.17) | 0.90 (0.74–1.10) | ||||
| 0.009–0.014 | 215 | 1.02 (0.85–1.24) | 1.03 (0.85–1.24) | 1.01 (0.83–1.23) | ||||
| 0.014–0.025 | 221 | 1.04 (0.87–1.26) | 1.04 (0.86–1.26) | 1.02 (0.85–1.24) | ||||
| ≥ 0.025 | 240 | 1.11 (0.92–1.34) | 1.10 (0.91–1.32) | 1.10 (0.91–1.33) | ||||
| Docosapentaenoic acid (22:5) | ≤ 0.000 | 214 | ref | 0.34 | ref | 0.53 | ref | 0.59 |
| 0.000–0.005 | 221 | 1.07 (0.88–1.29) | 1.01 (0.82–1.24) | 1.12 (0.93–1.37) | ||||
| 0.005–0.007 | 208 | 1.00 (0.82–1.21) | 0.97 (0.80–1.18) | 0.99 (0.81–1.21) | ||||
| 0.007–0.010 | 204 | 0.99 (0.82–1.20) | 0.96 (0.79–1.16) | 0.97 (0.79–1.18) | ||||
| ≥ 0.010 | 240 | 1.14 (0.95–1.38) | 1.09 (0.91–1.32) | 1.13 (0.94–1.37) | ||||
| Docosahexaenoic acid (22:6) | ≤ 0.016 | 228 | ref | 0.22 | ref | 0.27 | ref | 0.3 |
| 0.016–0.024 | 200 | 0.89 (0.74–1.08) | 0.90 (0.74–1.08) | 0.91 (0.75–1.11) | ||||
| 0.024–0.034 | 208 | 0.94 (0.77–1.13) | 0.94 (0.78–1.14) | 0.95 (0.79–1.16) | ||||
| 0.034–0.052 | 209 | 0.95 (0.79–1.14) | 0.95 (0.78–1.14) | 0.96 (0.79–1.16) | ||||
| ≥ 0.052 | 242 | 1.10 (0.92–1.32) | 1.09 (0.91–1.31) | 1.09 (0.90–1.31) | ||||
Abbreviations: OR: odds ratios; CI: confidence intervals; MUFA: monounsaturated fatty acids, PUFA: polyunsaturated fatty acids
Adjusted for age and sex
Adjusted for age, sex, race, smoking status, caffeine intake and total energy intake.
Adjusted for inverse probability weights with 95% CIs calculated from robust variance estimates
Interaction analyses
Overall, we did not observe significant interactions between intakes of these fatty acids and age, sex, smoking or caffeine intake (Supplementary Table 1). The positive associations observed in the main analyses were however apparently limited to never smokers and heavy coffee drinkers.
Discussion
In this large prospective study among older US adults, we found that overall fat intake was not associated with the risk for PD. There were weak positive associations of PD risk with higher intakes of the n-6 PUFA and linoleic acid; however these associations did not persist adjustments for multiple comparisons. To the best of our knowledge, this study is to date the largest prospective analysis on dietary fat and PD risk. Further, we excluded PD cases diagnosed prior to 2000 from the analysis with the hope to reduce concern about reverse causation due to changing dietary habits of PD patients in the years leading to a PD diagnosis.
Dietary fat intake may affect PD risk because fats, particularly PUFA, may contribute to oxidative stress and neuroinflammation. Neural membranes are rich in PUFA and are therefore sources of oxygen radicals through lipid peroxidation 34. Oxidation of membrane PUFA may cause mitochondrial dysfunction and damage cellular components, thus contribute to neuronal degeneration 35. In addition, PUFA are essential components of glial cell membranes and regulate the production of prostaglandins and proinflammatory cytokines 6. Exacerbated or prolonged inflammatory processes in the brain may play a key role in neurodegeneration 36. Furthermore, PUFA may promote α-synuclein oligomerization and aggregation in cultured dopaminergic cells and increase its deposit into intraneuronal Lewy-like inclusions 37. The majority of membrane PUFA is synthesized from dietary n-6 linoleic and n-3 α-linoleic acids, which are also precursors for the synthesis of long chain PUFA by desaturation and elongation reactions 38. The n-3 PUFA are anti-inflammatory, while n-6 PUFA are precursors of proinflammatory prostaglandins. Therefore our finding that n-6 PUFA intake was associated with higher PD risk is biologically plausible.
To the best of our knowledge, the associations between dietary fat intake and PD have been investigated in approximately a dozen epidemiological studies, including both case-control studies and prospective studies. Several earlier case-control studies reported a positive association with intake of total fat 9, 11 or animal/saturated fat 9, 10, but these findings were largely not confirmed in later studies 12, 13, 18, 19. For specific types of dietary fats, the findings are largely null 12, 18, although some interactions were reported between specific types of dietary fats (saturated or PUFA) and iron intake 18 or pesticide exposure 19. Most of these case-control studies used prevalent PD cases that had the diseases for years at the time of the study 9–13, 18, 19 and a couple of studies did not specifically query dietary intake prior to disease diagnosis 12, therefore these findings might have been affected by recall bias and reverse causation that dietary intakes might have changed as a result of the disease.
Prospective cohort studies assessed diet prior to disease diagnosis and thus are relatively less prone to reverse causality and recall bias. However, diet in prospective studies is often evaluated at one point in mid- or late- adulthood, which may not capture the time windows that are relevant to disease etiology. Further, prospective studies may not be immune from reverse causation because patients may suffer from premotor symptoms which may affect dietary intakes. Some cohort studies, but not all, suggested lower risk of PD with higher intakes of one or more types of fatty acids, although the results were not always statistically significant nor consistent for specific types of fatty acids 14–17, 20. Only one previous study evaluated the dietary intake repeatedly and conducted lag analyses that excluded the first several years of follow-up 16. The Health Professionals Follow-up Study and the Nurses’ Health Study included 359 incident PD cases, nearly the total number of cases of the other four cohorts combined 16. This study did not identify associations between PD risk and total fat or individual types of fatty acids with the exception of a borderline inverse association with arachidonic acid 16.
Compared to previous prospective studies, our current study was substantially larger with 1,087 PD cases and 299,617 controls. This large sample size provided better power to detect associations and facilitated adjustment for multiple covariates and gender-specific analyses. Our analyses only included PD cases diagnosed after 2000, at least four to five years after exposure assessment, which might have further alleviated concern about reverse causality. This large population also encompassed a large range of fat intakes. Fat intakes in this cohort have been associated with the risk of various types of chronic diseases, such as prostate cancer, chronic liver disease and hepatocellular carcinoma and pancreatic cancer in a manner consistent with their metabolic effects 28, 39, 40.
Our study also had several limitations. First, PD cases were self-reported and thus reporting and diagnostic errors were inevitable. We confirmed 88% of diagnoses in patients for whom we obtained medical information from their treating neurologists and we further removed patients with identified reporting errors and misdiagnoses from the analyses. However, we were unable to include PD cases who did not report their diagnosis on the follow-up questionnaire as cases in the analysis, and this might have biased the analyses if fat intakes of these cases were systematically different from cases who did report their diagnoses. Further, as the exact date of diagnosis was only available for some cases, we were unable to perform time to event survival analyses. Secondly, fat intakes were assessed as part of a dietary history questionnaire at study baseline; we could not exclude the possibility of measurement errors and residual confounding. Thirdly, PD may have a long prodromal period and we only had dietary assessment once at baseline. Although we only included PD cases diagnosed after 2000, we could not exclude the possibility of reverse causality if some PD patients’ dietary habits had been affected early in the prodromal period. Fourthly, the analysis was performed among participants of the follow-up survey, and therefore the possibility of selection bias cannot be excluded. Nevertheless, we used the inverse probability weighting analysis to account for this potential selection bias and found similar results. Finally, the cohort was predominantly non-Hispanic white, so results may not be readily generalizable to other populations.
In summary, we conducted the largest prospective study to date examining the associations between dietary fat intake and PD risk. Dietary intakes of fat in general were not found to be associated with the risk for PD. The weak positive association between intake of n-6 PUFA and PD needs confirmation and further investigation.
Supplementary Material
Acknowledgements
The authors are grateful to the continuous contribution and support of investigators from the NIH-AARP Diet and Health Study. This study was supported by the Intramural Research Program of the NIH, the National Institute of Environmental Health Sciences (Z01-ES-101986) and the National Cancer Institute (Z01-CP010196-02), and Award Number T42OH00867302 from the National Institute for Occupational Safety and Health.
Footnotes
Disclosure statement
The authors have no conflicts of interest.
Reference
- 1.Bach JP, Ziegler U, Deuschl G, Dodel R, Doblhammer-Reiter G. Projected numbers of people with movement disorders in the years 2030 and 2050. Mov Disord. 2011;26:2286–2290. doi: 10.1002/mds.23878. [DOI] [PubMed] [Google Scholar]
- 2.Jenner P. Oxidative stress in Parkinson's disease. Annals of neurology. 2003;53(Suppl 3):S26–S36. doi: 10.1002/ana.10483. discussion S36-28. [DOI] [PubMed] [Google Scholar]
- 3.Wullner U, Klockgether T. Inflammation in Parkinson's disease. J Neurol. 2003;250(Suppl 1):I35–I38. doi: 10.1007/s00415-003-1107-x. [DOI] [PubMed] [Google Scholar]
- 4.Schapira AH. Mitochondrial dysfunction in Parkinson's disease. Cell death and differentiation. 2007;14:1261–1266. doi: 10.1038/sj.cdd.4402160. [DOI] [PubMed] [Google Scholar]
- 5.Noseworthy MD, Bray TM. Effect of oxidative stress on brain damage detected by MRI and in vivo 31P-NMR. Free Radic Biol Med. 1998;24:942–951. doi: 10.1016/s0891-5849(97)00383-3. [DOI] [PubMed] [Google Scholar]
- 6.Laye S. Polyunsaturated fatty acids, neuroinflammation and well being. Prostaglandins Leukot Essent Fatty Acids. 2010;82:295–303. doi: 10.1016/j.plefa.2010.02.006. [DOI] [PubMed] [Google Scholar]
- 7.Levant B, Ozias MK, Carlson SE. Specific brain regions of female rats are differentially depleted of docosahexaenoic acid by reproductive activity and an (n-3) fatty acid-deficient diet. The Journal of nutrition. 2007;137:130–134. doi: 10.1093/jn/137.1.130. [DOI] [PubMed] [Google Scholar]
- 8.Bowen RA, Clandinin MT. Dietary low linolenic acid compared with docosahexaenoic acid alter synaptic plasma membrane phospholipid fatty acid composition and sodium-potassium ATPase kinetics in developing rats. Journal of neurochemistry. 2002;83:764–774. doi: 10.1046/j.1471-4159.2002.01156.x. [DOI] [PubMed] [Google Scholar]
- 9.Logroscino G, Marder K, Cote L, Tang MX, Shea S, Mayeux R. Dietary lipids and antioxidants in Parkinson's disease: a population-based, case-control study. Annals of neurology. 1996;39:89–94. doi: 10.1002/ana.410390113. [DOI] [PubMed] [Google Scholar]
- 10.Anderson C, Checkoway H, Franklin GM, Beresford S, Smith-Weller T, Swanson PD. Dietary factors in Parkinson's disease: the role of food groups and specific foods. Movement disorders : official journal of the Movement Disorder Society. 1999;14:21–27. doi: 10.1002/1531-8257(199901)14:1<21::aid-mds1006>3.0.co;2-y. [DOI] [PubMed] [Google Scholar]
- 11.Johnson CC, Gorell JM, Rybicki BA, Sanders K, Peterson EL. Adult nutrient intake as a risk factor for Parkinson's disease. International journal of epidemiology. 1999;28:1102–1109. doi: 10.1093/ije/28.6.1102. [DOI] [PubMed] [Google Scholar]
- 12.Miyake Y, Sasaki S, Tanaka K, et al. Dietary fat intake and risk of Parkinson's disease: a case-control study in Japan. Journal of the neurological sciences. 2010;288:117–122. doi: 10.1016/j.jns.2009.09.021. [DOI] [PubMed] [Google Scholar]
- 13.Hellenbrand W, Boeing H, Robra BP, et al. Diet and Parkinson's disease. II: A possible role for the past intake of specific nutrients. Results from a self-administered food-frequency questionnaire in a case-control study. Neurology. 1996;47:644–650. doi: 10.1212/wnl.47.3.644. [DOI] [PubMed] [Google Scholar]
- 14.Tan LC, Koh WP, Yuan JM, et al. Differential effects of black versus green tea on risk of Parkinson's disease in the Singapore Chinese Health Study. Am J Epidemiol. 2008;167:553–560. doi: 10.1093/aje/kwm338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Abbott RD, Ross GW, White LR, et al. Environmental, life-style, and physical precursors of clinical Parkinson's disease: recent findings from the Honolulu-Asia Aging Study. J Neurol. 2003;250(Suppl 3):III30–III39. doi: 10.1007/s00415-003-1306-7. [DOI] [PubMed] [Google Scholar]
- 16.Chen H, Zhang SM, Hernan MA, Willett WC, Ascherio A. Dietary intakes of fat and risk of Parkinson's disease. American journal of epidemiology. 2003;157:1007–1014. doi: 10.1093/aje/kwg073. [DOI] [PubMed] [Google Scholar]
- 17.de Lau LM, Bornebroek M, Witteman JC, Hofman A, Koudstaal PJ, Breteler MM. Dietary fatty acids and the risk of Parkinson disease: the Rotterdam study. Neurology. 2005;64:2040–2045. doi: 10.1212/01.WNL.0000166038.67153.9F. [DOI] [PubMed] [Google Scholar]
- 18.Powers KM, Smith-Weller T, Franklin GM, Longstreth WT, Jr, Swanson PD, Checkoway H. Dietary fats, cholesterol and iron as risk factors for Parkinson's disease. Parkinsonism & related disorders. 2009;15:47–52. doi: 10.1016/j.parkreldis.2008.03.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Kamel F, Goldman SM, Umbach DM, et al. Dietary fat intake, pesticide use, Parkinson's disease. Parkinsonism & related disorders. 2014;20:82–87. doi: 10.1016/j.parkreldis.2013.09.023. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Kyrozis A, Ghika A, Stathopoulos P, Vassilopoulos D, Trichopoulos D, Trichopoulou A. Dietary and lifestyle variables in relation to incidence of Parkinson's disease in Greece. European journal of epidemiology. 2013;28:67–77. doi: 10.1007/s10654-012-9760-0. [DOI] [PubMed] [Google Scholar]
- 21.Schatzkin A, Subar AF, Thompson FE, et al. Design and serendipity in establishing a large cohort with wide dietary intake distributions : the National Institutes of Health-American Association of Retired Persons Diet and Health Study. Am J Epidemiol. 2001;154:1119–1125. doi: 10.1093/aje/154.12.1119. [DOI] [PubMed] [Google Scholar]
- 22.Chen H, Huang X, Guo X, et al. Smoking duration, intensity, and risk of Parkinson disease. Neurology. 2010;74:878–884. doi: 10.1212/WNL.0b013e3181d55f38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Chen H, Zhang SM, Hernan MA, Willett WC, Ascherio A. Weight loss in Parkinson's disease. Ann Neurol. 2003;53:676–679. doi: 10.1002/ana.10577. [DOI] [PubMed] [Google Scholar]
- 24.Chen H, Zhang SM, Schwarzschild MA, Hernan MA, Ascherio A. Physical activity and the risk of Parkinson disease. Neurology. 2005;64:664–669. doi: 10.1212/01.WNL.0000151960.28687.93. [DOI] [PubMed] [Google Scholar]
- 25.Subar AF, Midthune D, Kulldorff M, et al. Evaluation of alternative approaches to assign nutrient values to food groups in food frequency questionnaires. Am J Epidemiol. 2000;152:279–286. doi: 10.1093/aje/152.3.279. [DOI] [PubMed] [Google Scholar]
- 26.Thompson FE, Kipnis V, Midthune D, et al. Performance of a food-frequency questionnaire in the US NIH-AARP (National Institutes of Health-American Association of Retired Persons) Diet and Health Study. Public Health Nutr. 2008;11:183–195. doi: 10.1017/S1368980007000419. [DOI] [PubMed] [Google Scholar]
- 27.Liu R, Guo X, Park Y, et al. Caffeine intake, smoking, and risk of Parkinson disease in men and women. Am J Epidemiol. 2012;175:1200–1207. doi: 10.1093/aje/kwr451. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Thiebaut AC, Jiao L, Silverman DT, et al. Dietary fatty acids and pancreatic cancer in the NIH-AARP diet and health study. Journal of the National Cancer Institute. 2009;101:1001–1011. doi: 10.1093/jnci/djp168. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Willett WC, Howe GR, Kushi LH. Adjustment for total energy intake in epidemiologic studies. The American journal of clinical nutrition. 1997;65:1220S–1228S. doi: 10.1093/ajcn/65.4.1220S. discussion 1229S-1231S. [DOI] [PubMed] [Google Scholar]
- 30.Cole SR, Hernan MA. Constructing inverse probability weights for marginal structural models. American journal of epidemiology. 2008;168:656–664. doi: 10.1093/aje/kwn164. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Greenland S, Pearl J, Robins JM. Causal diagrams for epidemiologic research. Epidemiology. 1999;10:37–48. [PubMed] [Google Scholar]
- 32.Hernan MA, Hernandez-Diaz S, Robins JM. A structural approach to selection bias. Epidemiology. 2004;15:615–625. doi: 10.1097/01.ede.0000135174.63482.43. [DOI] [PubMed] [Google Scholar]
- 33.Hernan MA, Brumback B, Robins JM. Marginal structural models to estimate the causal effect of zidovudine on the survival of HIV-positive men. Epidemiology. 2000;11:561–570. doi: 10.1097/00001648-200009000-00012. [DOI] [PubMed] [Google Scholar]
- 34.Farooqui AA, Horrocks LA. Lipid peroxides in the free radical pathophysiology of brain diseases. Cell Mol Neurobiol. 1998;18:599–608. doi: 10.1023/a:1020625717298. [DOI] [PubMed] [Google Scholar]
- 35.Shchepinov MS, Chou VP, Pollock E, et al. Isotopic reinforcement of essential polyunsaturated fatty acids diminishes nigrostriatal degeneration in a mouse model of Parkinson's disease. Toxicology letters. 2011;207:97–103. doi: 10.1016/j.toxlet.2011.07.020. [DOI] [PubMed] [Google Scholar]
- 36.Venters HD, Dantzer R, Kelley KW. A new concept in neurodegeneration: TNFalpha is a silencer of survival signals. Trends Neurosci. 2000;23:175–180. doi: 10.1016/s0166-2236(99)01533-7. [DOI] [PubMed] [Google Scholar]
- 37.Assayag K, Yakunin E, Loeb V, Selkoe DJ, Sharon R. Polyunsaturated fatty acids induce alpha-synuclein-related pathogenic changes in neuronal cells. The American journal of pathology. 2007;171:2000–2011. doi: 10.2353/ajpath.2007.070373. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Youdim KA, Martin A, Joseph JA. Essential fatty acids and the brain: possible health implications. Int J Dev Neurosci. 2000;18:383–399. doi: 10.1016/s0736-5748(00)00013-7. [DOI] [PubMed] [Google Scholar]
- 39.Pelser C, Mondul AM, Hollenbeck AR, Park Y. Dietary fat, fatty acids, and risk of prostate cancer in the NIH-AARP diet and health study. Cancer Epidemiol Biomarkers Prev. 2013;22:697–707. doi: 10.1158/1055-9965.EPI-12-1196-T. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Freedman ND, Cross AJ, McGlynn KA, et al. Association of meat and fat intake with liver disease and hepatocellular carcinoma in the NIH-AARP cohort. J Natl Cancer Inst. 2010;102:1354–1365. doi: 10.1093/jnci/djq301. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.