Stroke is a major cause of mortality and disability worldwide(Reference Feigin, Brainin and Norrving1). Stroke survivors face great life challenges because of severe sequelae (such as paralysis, aphasia, cognitive impairment and psychological disorders)(Reference van Dongen, Hafsteinsdóttir and Parker2), and depression is one of the most common mental diseases after stroke(Reference Jørgensen, Wium-Andersen and Wium-Andersen3). Post-stroke depression further leads to poor quality of life and higher mortality(Reference Bartoli, Di Brita and Crocamo4,Reference Villa, Ferrari and Moretti5) .
The mechanisms of post-stroke depression are multifactorial and not fully understood(Reference Loubinoux, Kronenberg and Endres6,Reference Medeiros, Roy and Kontos7) . The possible mechanisms of post-stroke depression included hypothalamic-pituitary-adrenal axis dysregulation, reduced monoamine levels, abnormal neurotrophic responses(Reference Loubinoux, Kronenberg and Endres6,Reference Medeiros, Roy and Kontos7) , neuronal ischaemic change and neuroinflammation(Reference Medeiros, Roy and Kontos7,Reference Fan, Liu and Sheng8) . In addition, the concentrations of the pro-inflammatory cytokines TNF-α and IL-6 in peripheral blood were significantly elevated in patients with depression(Reference Dowlati, Herrmann and Swardfager9). Cytokines alter the production, metabolism and transport of neurotransmitters that synergistically affect mood and also affect neuronal growth and survival to promote depressive episodes(Reference Kiecolt-Glaser, Derry and Fagundes10). Increasing evidence indicates that foods and nutrients may play a role in depressive episodes because of their anti-inflammatory and pro-inflammatory properties. A meta-analysis showed that an increased intake of total dietary fibre is associated with lower odds of depression(Reference Fatahi, Matin and Sohouli11). Adhere to healthy dietary patterns(Reference Lassale, Batty and Baghdadli12), such as Mediterranean diet(Reference Psaltopoulou, Sergentanis and Panagiotakos13) and traditional Brazilian diet(Reference Canheta, Santos and Souza14) rich in vegetables, fruits, fish, and olive oil are thought to be associated with reduced risk of depression, whereas Western-type dietary patterns are characterised by high consumption of red or processed meat, refined grains, sweets and high-fat dairy products and low intake of fruits and vegetables that aggravate depressive episodes(Reference Li, Lv and Wei15), and a recent meta-analysis also suggested that ultra-processed food was associated with an increased risk of anxiety and depression(Reference Lane, Gamage and Travica16). The association between diet and depression may be partially mediated by inflammatory markers(Reference Lai, Oldmeadow and Hure17).
Studies have investigated the relationship between the inflammatory potential of diet and post-stroke depression(Reference Pereira, da Silva and Hermsdorff18,Reference Xu, Qian and Sun19) . A US population study showed that increasing dietary antioxidant intake may prevent depressive symptoms in adults with stroke(Reference Xu, Qian and Sun19). However, the relationship between certain food antioxidants and depression remains controversial, which may be related to different food questionnaires and methods used to assess dietary intake in previous studies(Reference Lai, Oldmeadow and Hure17,Reference Lin and Shen20) . The dietary inflammatory index (DII) is a dietary index that provides a quantitative means to study the relationship between pro- and anti-inflammatory diets and disease(Reference Shivappa, Steck and Hurley21). A study of postmenopausal women suggested that DII can be used to predict the incidence of menopausal complications(Reference Haghshenas, Baharanchi and Melekoglu22). Ghazizadeh et al. found that DII is significantly associated with major depression in women(Reference Ghazizadeh, Yaghooti-Khorasani and Asadi23). However, the relationship between the DII and depression in adults with stroke remains unclear.
Thus, the objective of this study was to investigate the association between DII and depression in American adults with stroke using data from the National Health and Nutrition Examination Survey (NHANES).
Materials and methods
Data source and study population
The present study was a cross-sectional analysis using data from NHANES 2005–2018. The NHANES is a nationally representative cross-sectional survey of the non-institutionalised US population with data collected in 2-year cycles. During each cycle, the NHANES was conducted based on a stratified multistage probability sampling design and included two components: a household interview and a health examination. This study included individuals older than 18 years who participated in the 2005–2018 NHANES survey cycles. From a total of 1658 participants with stroke, we excluded those with missing data, including the nine-item Patient Health Questionnaire (PHQ-9) score (n 304), DII score (n 59) and participants with missing data on covariates of interest (n 56). The final sample size for the analysis was 1239. All participants provided written informed consent before participating in the survey. The NHANES survey was approved by the Research Ethics Review Board of the National Centre for Health Statistics, and the procedures followed the principles of the Declaration of Helsinki. The NHANES data used in this study are publicly available (https://wwwn.cdc.gov/nchs/nhanes/) and do not require ethical or administrative approval.
Covariates selection
Sociodemographic and lifestyle information, including age, sex, race, marital status, smoking status, education level and sleep status, were obtained through standardised questionnaires. Race was classified as non-Hispanic White, non-Hispanic Black, Mexican American or other. Smoking status was defined as a non-smoker, former smoker or current smoker. Education was classified as less than high school and high school or above. Sleep status was obtained by participants answering the following questionnaire: over the last 2 weeks, how often have you been bothered by trouble falling or staying asleep or sleeping too much? And it was defined as not at all, several days, more than half the days and nearly every day. BMI was provided by the Mobile Examination Centre. A history of hypertension and diabetes was assessed using a combination of questionnaires and examination results. Hypertension was defined as follows: (1) average systolic blood pressure/average diastolic blood pressure ≥ 140/90 mmHg, (2) previous diagnosis by a doctor or health professional or (3) currently taking antihypertensive medications. Diabetes was defined as (1) self-reported diagnosis of diabetes, (2) fasting HbA1c level > 6·4 % or (3) current use of hypoglycaemic drugs. Stroke data were self-reported personal interview data obtained from the medical status section of the NHANES. In this questionnaire, stroke was identified as a stroke diagnosis by a physician or health professional.
Depressive symptom assessment
A validated PHQ-9 was used to assess depressive symptoms. PHQ-9 consists of a nine-item depression module to assess the frequency of depressive symptoms over the past 2 weeks. With each of the nine items ranging from ‘0’ (not at all) to ‘3’ (nearly every day), PHQ-9 has total scores ranging from 0 to 27. A higher PHQ-9 score was associated with more severe depressive symptoms. PHQ-9 scores of 5, 10, 15 and 20 represented mild, moderate, moderately severe and severe depression, respectively(Reference Kroenke, Spitzer and Williams24). Depression was defined as PHQ-9 scores ≥ 10 in clinical practice(Reference Kroenke, Spitzer and Williams24). In this study, we categorised depression into three grades based on the PHQ-9 scores: no clinical depression (PHQ-9: ≤ 9), moderate-to-moderately severe depression (PHQ-9: 10–19) and severe depression (PHQ-9: 20–27).
Dietary data and computation of DII scores
Dietary data were derived from self-reports, and DII was calculated by adding the scores of each dietary component consumed by each participant within a 24-h period. Higher DII scores were associated with a more pro-inflammatory diet, whereas lower DII scores indicated a more anti-inflammatory diet. The method for calculating DII was reported in detail by Shivappa et al. (Reference Shivappa, Steck and Hurley21). DII scores were based on eleven food consumption datasets worldwide(Reference Shivappa, Steck and Hurley21), which provides reliable estimates of the mean and sd of forty-five food parameters. A participant’s exposure relative to the standard global mean was calculated as a z-score, which was calculated by subtracting the mean of the energy-adjusted regionally representative database from each food parameter and dividing this value by the sd of the parameter. These z-scores were converted to central proportion scores by multiplying by 2 and subtracting 1 to reduce the effect of ‘right-skewing’. The final values were multiplied by the overall food parameter-specific inflammatory effect score to obtain a food parameter-specific DII score. DII scores for specific food parameters were added to obtain an overall DII score for each individual. In this study, the NHANES 2008–2015 database provided 28 of the 45 food parameters to calculate DII. Previous studies revealed that the DII scores were still available even if the nutrients used to calculate DII were <30(Reference Shivappa, Steck and Hurley21). These food parameters and other basic information for calculating DII are shown in online Supplementary Table S1.
Statistical analysis
Descriptive data on participants’ characteristics were expressed as means and se or medians, interquartile ranges for continuous variables and numbers and weighted percentages for categorical variables. One-way ANOVA and χ 2 tests were used to compare continuous and categorical variables, respectively. The weight prevalence of different degrees of depression was evaluated in all participants. Multivariate logistic regression analyses were used to estimate the association between DII scores and different degrees of depression in adults with stroke. Model 1 shows the age-adjusted results. Variables were entered in the multivariate logistic regression models if the P value was ≤ 0·10 in the univariable analysis. In the multivariate-adjusted model, we adjusted for baseline age, sex, race, educational status, marital status and smoking status. Restricted cubic spline was used to evaluate the potential non-linear relationship between DII and depression in adults with stroke. Stratification analysis was performed to estimate the relationship between DII and severe depression according to age, sex and smoking status. Data were weighted to ensure that they were representative of US adults using complex survey sampling analysis methods. All data analyses were performed using the R software (version R-4.1.0; Cary, NC, USA). Two-sided P values <0·05 were considered statistical significance.
Results
Characteristics of the study population
This study included 1239 adults with stroke. The mean age of the participants was 63·85 years, and 622 (50·20 %) were women. Among the 1239 participants, the weight prevalence of moderate-to-moderately severe and severe depression was 14·62 and 3·64 %, respectively (Fig. 1).
Percentage (%) of depression status in adults with stroke, National Health and Nutrition Examination Survey 2005–2018.
Table 1 presents the characteristics of the participants and DII score tertiles. Compared with participants in the lowest DII score tertile (tertile 1), those in the highest DII score tertile (tertile 3) were more likely to be female, highly educated and current smokers. There were no differences in the distribution of age, race, marital status, BMI, sleep status or history of diabetes or hypertension.
Baseline characteristics of adults with stroke by the dietary inflammatory index (DII) tertiles (Q) from National Health and Nutrition Examination Survey 2005–2018
* Mean (se).
Associations between DII and depression in adults with stroke
Table 2 shows that the DII score was positively associated with severe depression in adults with stroke. The age-adjusted OR (95 % CI) for severe depression across the DII score tertiles were 1·00 (reference), 0·995 (0·332, 2·976) and 3·853 (1·417, 10·476), respectively (Table 2). The OR for severe depression were similar following additional adjustments for sex, race, education, smoking and marital status (1·00 (reference), 0·942 (0·327, 2·718) and 3·222 (1·150, 9·026) across the DII tertiles). When modelling the DII score continuously, a similar positive association was observed in the fully adjusted models (model 2 in Table 2). A one-unit increase in the DII score was associated with 35·9 % higher odds (P = 0·026) of severe depression in adults with stroke. No significant association between the DII and depression was found for total depression (P = 0·283) or moderate-to-moderately severe depression (P = 0·871). In Fig. 2, the non-linear relationship between DII and severe depression was NS in the restricted cubic spline model (P for non-linearity > 0·05).
Associations of the dietary inflammatory index (DII) with depression in adults with stroke
Model 1: Adjusted for age. Model 2: Adjusted for age, sex, race, education status, marital status and smoking status.
Clinical depression: PHQ-9 ≥ 10; moderate-to-moderately severe depression: PHQ-9 scores:10–19; severe depression: PHQ-9 ≥ 20. PHQ-9, nine-item Patient Health Questionnaire.
Restricted cubic spline of dietary inflammatory index with severe depression.
Stratification analysis
In the stratified analyses (Fig. 3), no interactions were observed between the DII and sex (P-interaction = 0·184) or smoking status (P-interaction = 0·396). However, the DII-severe depression association was stronger in older individuals (P-interaction = 0·010).
Subgroup analysis of the association of the dietary inflammatory index with severe depression in adults with stroke. Results were adjusted for all covariates except the corresponding stratification variable.
Discussion
In this cross-sectional study of a representative US sample, we found that older adults with stroke with higher DII scores were significantly associated with a higher risk of severe depression. No significant association was found in adults with stroke and moderate-to-moderately severe depression. These findings may provide clinical and public health implications for preventing depression in adults with stroke.
The DII level is related to the pro- and anti-inflammatory ability of the diet(Reference Shivappa, Steck and Hurley21), with a higher DII indicating a stronger pro-inflammatory ability. Inflammation is a defence mechanism that prevents harmful substances from entering the body; however, sustained and prolonged inflammation is detrimental to health(Reference Ferrucci and Fabbri25). Increasing evidence indicates that the inflammatory properties of diet are associated with a variety of diseases (such as CHD and various types of cancer)(Reference Marx, Veronese and Kelly26). A pro-inflammatory diet can promote depression by affecting the hypothalamic-pituitary-adrenal axis, oxidative stress, gut-brain axis and other pathways to increase neuroinflammation and change the secretion of neurotransmitters(Reference Marx, Lane and Hockey27–Reference Richards, Yap and McLeod29). In a previous meta-analysis, participants with the highest DII scores had a 23 % higher risk of depression than those with the lowest DII scores(Reference Wang, Zhou and Chen30). A significant association between DII and depression has been confirmed in the general population(Reference Zhao, Sun and Liu31) and in individuals with co-morbid chronic diseases (such as diabetes and CHD)(Reference Jiang, Yin and Liu32). A cohort study from Iran showed that DII is significantly associated with major depression in women(Reference Ghazizadeh, Yaghooti-Khorasani and Asadi23). However, few studies have reported the association between DII scores and depression severity in adults with stroke. We found a significant association between the DII and severe depressive episodes in adults with stroke, whereas this relationship was not observed for moderate-to-moderately severe depression. It should be noted that the DII score was not significantly associated with depression in adults with stroke before the degree of depression was graded, contradicting to the positive results of previous studies in other populations(Reference Wang, Zhou and Chen30–Reference Jiang, Yin and Liu32). Depression is a multifactorial disease(Reference McCarron, Shapiro and Rawles33). Post-stroke depression is directly related to cerebrovascular injury(Reference Medeiros, Roy and Kontos7). Compared with the general population, the composition of risk factors for depression in the stroke population may be different(Reference Medeiros, Roy and Kontos7,Reference Disu, Anne and Griffiths34) , which may lead to different conclusions on the correlation between DII and depression in different populations. In addition, the crude diagnosis of depression may mask the actual relationship between the DII and different degrees of depression. The complex composition of risk factors for depression and the relationship between diet and different levels of depression in different populations need to be further explored.
We further performed a subgroup analysis of the association between the DII score and severe depression in adults with stroke. A higher DII score conferred a higher risk of major depressive episodes in older adults. In line with previous studies, older age was a risk factor for depression(Reference McCarron, Shapiro and Rawles33). Ageing, increased levels of accompanying diseases, lower activity and lower living standards predispose older adults to psychological problems(Reference Zhang, Xiang and Li35), which may be attributed to immune dysregulation, most notably, high blood levels of pro-inflammatory immunogenic stimuli(Reference Ferrucci, Semba and Guralnik36,Reference Fulop, Larbi and Dupuis37) . Additionally, DII promotes the onset of depression by increasing pro-inflammatory factors in the body(Reference Postal and Appenzeller28,Reference Richards, Yap and McLeod29) ; therefore, there may be a synergistic effect between DII and ageing.
Depression plays an important role in the progression and prognosis of stroke(Reference Jørgensen, Wium-Andersen and Wium-Andersen3). Diet appears to be a modifiable risk factor for depression(Reference Canheta, Santos and Souza14,Reference Akbaraly, Brunner and Ferrie38) . Our results extend the association between dietary inflammation potential and depression in the stroke population, with a more detailed classification of the degree of depression and a relative subgroup analysis to identify special populations.
Despite the significance of our findings, this study has a few limitations. First, the NHANES is a cross-sectional observational study; therefore, causality cannot be established, and residual confounding factors cannot be completely ruled out. Second, dietary data were obtained from only one 24-h dietary recall; this may not represent habitual diet, although 24-h dietary recalls have been widely used in previous studies(Reference Zhou, Xie and Liu39,Reference Wu, Shi and Kong40) . Third, the history of antidepressant use and some other characteristic data were not available due to the limitations of the NHANES database. Finally, stroke, depression and dietary recalls were self-reported and may be limited by underlying cognitive difficulties in a stroke population, which may have introduced a bias. However, these questionnaires have been widely used to assess stroke, depression and DII(Reference Iranpour and Sabour41–Reference Chen, Hou and Yao43). Future longitudinal studies are needed to explore the relationship between DII and depression in different populations.
Conclusions
A high DII was associated with a higher likelihood of having severe depression in older adults with stroke. Future studies with prospective designs or clinical trials are needed to confirm the results of this study.
Supplementary material
For supplementary material/s referred to in this article, please visit https://doi.org/10.1017/S0007114524000850.
Acknowledgements
We would like to thank Editage (www.editage.cn) for English language editing.
Financial support
This study was funded by the Clinical Research Award of the First Affiliated Hospital of Xi’an Jiaotong University (grant no. XJTU1AF2018CRF-024) and by the horizontal subject of the First Affiliated Hospital of Xi’an Jiaotong University (grant no. HX201872). The study funders were not involved in the design of the study.
Authorship
Y. G. and Y. R. Z. designed the research; P. P. Z. conducted research and wrote the manuscript; P. P. Z. and Y. R. Z. analysed data; X. X. and Y. B. W. revised the tables and images; Y. G. and Y. R. Z. provided important revisions for the final content. All authors reviewed and approved the final version of the manuscript.
Conflict of interest
None.
