The relationship between a plant-based diet and mental health: Evidence from a cross-sectional multicentric community trial (LIPOKAP study)

Background Dietary patterns emphasizing plant foods might be neuroprotective and exert health benefits on mental health. However, there is a paucity of evidence on the association between a plant-based dietary index and mental health measures. Objective This study sought to examine the association between plant-based dietary indices, depression and anxiety in a large multicentric sample of Iranian adults. Methods This cross-sectional study was performed in a sample of 2,033 participants. A validated food frequency questionnaire was used to evaluate dietary intakes of participants. Three versions of PDI including an overall PDI, a healthy PDI (hPDI), and an unhealthy PDI (uPDI) were created. The presence of anxiety and depression was examined via a validated Iranian version of the Hospital Anxiety and Depression Scale (HADS). Results PDI and hPDI were not associated to depression and anxiety after adjustment for potential covariates (age, sex, energy, marital status, physical activity level and smoking). However, in the crude model, the highest consumption of uPDI approximately doubled the risk of depression (OR= 2.07, 95% CI: 1.49, 2.87; P<0.0001) and increased the risk of anxiety by almost 50% (OR= 1.56, 95% CI: 1.14, 2.14; P= 0.001). Adjustment for potential confounders just slightly changed the associations (OR for depression in the fourth quartile= 1.96; 95% CI: 1.34, 2.85, and OR for anxiety in the fourth quartile= 1.53; 95% CI: 1.07, 2.19). Conclusions An unhealthy plant-based dietary index is associated with a higher risk of depression and anxiety, while plant-based dietary index and healthy plant-based dietary index were not associated to depression and anxiety.


Masoud Lotfizadeh
Results: PDI and hPDI were not associated to depression and anxiety after adjustment for 60 potential covariates (age, sex, energy, marital status, physical activity level and smoking).

61
However, in the crude model, the highest consumption of uPDI approximately doubled the risk 62 of depression (OR= 2.07, 95% CI: 1.49, 2.87; P<0.0001) and increased the risk of anxiety by 63 almost 50% (OR= 1.56, 95% CI: 1.14, 2.14; P= 0.001). Adjustment for potential confounders 64 just slightly changed the associations (OR for depression in the fourth quartile= 1.96; 95% CI: Psychological disorders, including depression and anxiety, are one of the main public health 71 issues globally due to their high prevalence and poor outcomes (1). Depression is a major cause 72 of disease and disability affecting over 300 million people around the world (2). In addition, 73 the percentage of people suffering from mood disorders or anxiety is just over 15% (3). It is 74 well established that psychological disorders adversely influence an individual's health, quality 75 of life, lifespan, and alter dietary behaviors (4).

76
Emerging evidence suggests that diet plays a crucial role in mental health status (5). Amongst 77 various dietary patterns, the health benefits of a vegetarian diet have been widely indicated (6). Similarly, plant-based diet indices (PDIs), calculated using a scoring system, may also be 84 associated with better health status (9-12), but it is of note that all plant-derived foods are not 85 necessarily healthful (e.g. refined grains and potatoes) and possibly increase the risk of various potatoes, refined grains, sweets, desserts, and drinks. All foods derived from animal resources 140 constituted animal foods. All food groups were divided into deciles and a score of 1-10 was 141 assigned to each decile. Accordingly, for plant-based diet indices, scores of 10 and 1 were 142 respectively given to individuals at the highest and the lowest deciles of all plant-based foods, 143 regardless of their health properties. Other deciles proportionally scored between 1 and 10.

144
Animal food groups scored conversely. In other words, participants with the highest intake 145 received a score of 1 and those with the lowest intake score 10. For hPDI, the highest and 146 lowest consumption of healthy plant foods scored 10 and 1, respectively, while the highest and 147 lowest consumption of unhealthy plant foods and animal food items received a score of 1 and 148 10, respectively. For uPDI, participants with higher intake of unhealthy plant-based foods were 149 given higher scores (score 10 for decile 10 and score 1 for decile 1), and healthy plant-and 150 animal-based foods scored reversely (score 1 for decile 10 and score 10 for decile 1). Then all 151 scores were summed to obtain the final score for each index, ranging from 21 to 210. Higher 152 scores represent greater adherence to each score (15).  The final obtained scale will range from 0 (the lowest degree of anxiety and depression) to 21 157 (the greatest degree of anxiety or depression). A scores of ≤7 in each section was considered 158 not to have depression or anxiety symptoms and a score of ≥8 indicated the presence of 159 depression or anxiety symptoms ( 25 ) .

161
All PDI, hPDI, and uPDI were categorized into quartiles. General sample characteristics were 162 expressed as percentage for categorical and mean ± standard deviation (SD) for continuous 163 variables. The differences between quartiles were examined either applying a chi-square test 164 for categorical variables or one-way analysis of variance (ANOVA) for continuous variables 165 and dietary intake. We applied the Kruskal-Wallis or robust Brown-Forsyth tests to evaluate 166 means across quartiles when the assumptions for one-way analysis of variance were not met.

167
The mean depression and anxiety scores across quartiles of PDI, hPDI, and uPDI were 168 compared in crude and multivariable-adjusted models by using Kruskal-Wallis and ANCOVA, 169 respectively. The log transformation was used for non-normal variables in ANVOVA. The 170 crude and multivariable-adjusted odds ratio (OR) and 95% confidence intervals (CIs) for 171 having depression or anxiety across quartiles of PDI, hPDI, uPDI were estimated applying 172 multiple logistic regression. In the first adjusted model, the confounding effects of age, sex and 173 energy intake were controlled. Model 2 was additionally controlled for marital status, physical  Table 1 illustrates the general characteristics of participants according to quartiles of PDI, 180 hPDI, uPDI. People with higher PDI and uPDI scores were younger than those with lower 181 scores of PDI and uPDI. In contrast, the greater adherence to hPDI, the older participants were.

182
Individuals in the fourth quartile of PDI were more probably to be current smoker but those in 183 the highest quartile of hPDI were less likely to be smoker. The distribution of current smokers 184 across the quartiles of uPDI was similar.

185
Participants' dietary intake across quartiles of PDI, hPDI, and uPDI are shown in Table 2.

186
Compared with those with lower scores of PDI, the intake of all food groups was constantly 187 higher in subjects who had greater adherence to the PDI. The one exception to this trend was 188 dairy products which did not significantly differ between PDI categories. In contrast, higher 189 scores of hPDI were associated with lower energy intake and lower consumption of all three 190 macronutrients and different fatty acids, fiber, unhealthy vegetables, nuts, refined grains, meat, 191 fish and sea foods, dairy products, fast foods, sweet desserts and sweet drinks, while the intakes 192 of healthy vegetables, fruit, whole grains, and legumes increased by the sores of hPDI.

193
Likewise, greater adherence to uPDI was associated with lower intakes of most food groups, 194 but not unhealthy vegetables, refined grains, sweet desserts and sweet drinks which were 195 consumed in higher amounts in individuals with higher uPDI scores as compared to individuals 196 with lower uPDI scores.  Our main findings showed that the PDI and hPDI were not associated with depression and 221 anxiety, while higher scores of uPDI were associated with a higher risk of depression and 222 anxiety. Adjustment for potential covariates did not considerably change these associations.

223
This is the first multi-centric study amongst Iranians in this regard which can consider  The direct association between uPDI and mental disorders found in the present study, including 227 depression and anxiety, are in line with previous studies (14,15,26). Despite a null association 228 between an uPDI and stress among young women (17), in an earlier study on apparently 229 healthy, Iranian women, greater adherence to an uPDI was associated with 91% and 31%  (40), dopamine (41) and brain-derived neurotrophic factor. In support of 248 this mechanism, many studies reported a direct link between depression and prolonged 249 psychological distress and inflammatory pathways in the brain (42-44). Lower protein intake 250 may also adversely influence mental health status (45, 46).

251
In stark contrast with some previous evidence, we failed to find any significant association for 252 either PDI or hPDI with depression or anxiety (26,47,48). In a recent study among Iranian 253 women, an inverse link was reported between hPDI and psychological disorders (26).

254
Consistently, a systematic review controlled trials pointed that a PDI could significantly 255 enhance psychological well-being and ameliorate depressive symptoms in patients with type 2 256 diabetes mellitus (49). In the GAZEL cohort, healthy dietary pattern, defined by the 257 consumption of vegetables, was associated with fewer depressive symptoms in men and 258 women (50). Furthermore, in Chinese older people, the highest quartile of ''vegetables-fruits'' 259 pattern score was associated with a decreased risk of incident depression compared to the 260 lowest quartile at baseline (51). For instance, a study has shown that total protein intake and 261 protein intake from milk and milk products were negatively associated with depression (52).  Our study has several strengths. First, our study population is a large multicentric sample of 281 Iranian adults and therefore our results have a significant external validity to extrapolate our 282 findings to the Iranian. Second, since this study was conducted amongst healthy subjects, the 283 confounding effect of diseases on mental health status is eliminated. Third, using a validated 284 FFQ provides us with assessing habitual consumption of most of the food items over a long 285 run and consequently a more precise classification of food groups. Forth, given that various 286 plant foods may exert different health consequences, for instance refined and whole grains, in 287 this study, we categorized them into two healthy and unhealthy groups and examined their 288 impacts and patterns separately in our study population. Fifth, all questionnaires were 289 completed by trained interviewers which provides more reliable and precise responses.

290
There are several limitations of the present study that should be acknowledged. The cross-291 sectional design of this study does not allow us to draw causal inferences. For instance, anxiety 292 may cause people to have greater tendency towards unhealthy food choices. Although all 293 questionnaires were completed by interviewers, it is possible that respondents have answered 294 questions in a manner that they seem psychologically healthy. Moreover, in the present study, 295 only demographic and lifestyle variables were examined and the confounding effect of 296 unmeasured and residual confounders, such as food preferences and adhering to any specific 297 type of diet, cannot be completely ruled out. Finally, the FFQ is a memory-reliable tool and 298 subject to both random and systematic measurement errors, leading to misclassification of 299 dietary intakes, which is an inevitable bias in nutritional epidemiological research. More 300 reliable associations might have been acquired provided that depression, anxiety and dietary 301 intakes were measured by the means of more sophisticated tools, such as physician diagnosis 302 or a combination of instruments (e.g. a 24h recall beside FFQ (56)).

303
In conclusion, our results suggest that the greater adherence to an unhealthy plant-based diet, 304 the higher risk of depression and anxiety. However, neither PDI nor hPDI was pertinent to 305 depression and anxiety. Plant-based diets' composition is similar to a provegetarian diet, 306 identified by . Long-term adherence to a pure vegetarian diet is 307 not convenient for many individuals, however, consuming a greater proportion of daily energy 308 intake from plant-derived foods is an easier goal which can be achieved by many individuals.

309
In addition, due to great heterogeneity in terms of the association of vegetarian diets with 310 mental health status (58), our results can shed light on diets, mainly based on plant foods, and 311 mental health relations. Accordingly, though an uPDI in our study population is a risk factor 312 for depression and anxiety, overall PDI or hPDI cannot be a protective factor against them.

313
However, there is still few data in this regard and our results need to be confirmed by large 314 prospective cohort studies.    (9)