Migraine is a complicated brain disorder, defined as a moderate to severe headache(Reference Pescador Ruschel and De Jesus1) that can affect approximately 12 % of population, and is more prevalent in women (> 20 %) than men (> 10 %)(Reference Weatherall2,Reference Burch, Buse and Lipton3) . Migraines can influence hypothalamus, cortical, diencephalic and brainstem nuclei of the brain(Reference Mirzababaei, Khorsha and Togha4), whilst nausea, vertigo, nasal congestion, photophobia, noises and smells, lack of appetite, and gastrointestinal disorders are other symptoms that patients with migraine headache also often report(Reference Weatherall2). In some cases, headache will cease after sleeping(Reference Burstein, Noseda and Borsook5), although lethargy may be present for a prolonged period after headache cessation(Reference Ong and De Felice6). There are numerous reasons for the pathology of migraine disease, for example, a disorder of endogenous pain modulating system, sensitivity to hemostasis alteration and vascular changes(Reference Andreou and Edvinsson7,Reference Mason and Russo8) . Although migraine may originate from hereditary causes, various other internal and external conditions, such as daily diet, certain foods, alcohol, hormonal fluctuations, stressful situations and lifestyle, may affect the intensity and duration of migraine(Reference Weatherall2–Reference Burstein, Noseda and Borsook5). The presence of migraine headaches is typically greater in women than men, where the hypothalamic–hypophyseal–ovarian axis, that regulates oestrogen secretion, could be impacting on cortical excitability or vascular dilation (Reference Sacco, Ricci and Degan9).
The role of diet on migraine headaches, although of contemporary interest, is not well understood. It has been shown that foods that containing tyramine should be eliminated, whilst some vitamins and minerals, such as Mg, B2, and Q10, have been advocated for improving, and recovering from, headache duration and intensity(Reference Gazerani10). Different types of diet, for example, a Western-based diet, may contribute to increased weight gain; consequently, this can affect migraine pathology (nitric oxide (NO), adiponectin and leptin), both directly and indirectly(Reference Bond, Roth and Nash11). Lower consumption of B9 and B12 is associated with methylene tetrahydrofolate reductase (MTHFR) gene mutation(Reference Martin12) and the occurrence of hyperhomocysteinemia, which may worsen the intensity of migraine(Reference Lippi, Mattiuzzi and Meschi13). In addition, vitamin D deficiency has been reported to increase the duration of migraine headaches(Reference Ghorbani, Togha and Rafiee14). Many studies have suggested that an inverse relationship exists between dietary Na and high glycemic carbohydrate and migraine(Reference Razeghi Jahromi, Ghorbani and Martelletti15,Reference Pogoda, Gross and Arakaki16) . Whilst, via prostaglandin reduction, vitamin E consumption has been reported to positively influence migraines(Reference Shaik and Gan17). Moreover, in Shu-Han Meng et al., it was reported that higher intake of Ca and Mg may be inversely associated with migraine in women(Reference Meng, Wang and Kang18).
Nutrient patterns can provide a prospective view between total food intake and disease, where consideration is given to the impact of all dietary nutrients. In addition, studying dietary and nutrient patterns can also encompass both the food synergy and nutrients interactions(Reference Allès, Samieri and Lorrain19). However, many studies have paid attention to the role of specific foods on migraine, whilst few studies, to our knowledge, have explored the impact of dietary patterns on migraine. Therefore, the aim of present study is to investigate the relationship between dietary nutrients patterns and intensity and duration of migraine headaches.
Methods
Study population
This was a cross-sectional study, conducted on 266 women, in March 2016. The participants were diagnosed with migraine and referred to neurology clinics of Sina and Khatam Al-Anbia Hospitals and a professional headache clinic in Tehran, Iran. Inclusion criteria were as follows: 18–50 years of age, BMI between 18·5–30 kg/m2, and visiting the headache clinic for the first time. Having any chronic diseases (CVD, cancer, diabetes, etc) and taking any medicines that can effect on serum lipoprotein concentrations, including atorvastatin and lovastatin, were excluded.
Demographic characteristics
General characteristics, including age, education status, job and marital status, were collected using a demographic questionnaire. The International Physical Activity Questionnaire (IPAQ) was used to assess physical activity and reported as metabolic equivalent (MET) hours per week (MET-h/week)(Reference Wanner, Probst-Hensch and Kriemler20).
Anthropometric measurements
Weight was measured, to the nearest 0·1 kg, with participants unshod and wearing light clothing, using a digital scale (SECA). Height was measured, to the nearest 0·5 cm, with participants unshod and wearing light clothing, using a wall-mounted stadiometer. BMI was calculated. The waist circumference (WC) was measured at the midpoint of the lowest rib and the hip bone. Also, hip circumference was measured using a flexible tape at the largest anterior protrusion. The ratio of waist-to-hip ratio (WHR) was calculated by dividing WC (cm)/height (cm).
Dietary assessment and nutrient dietary patterns
Dietary intake was evaluated using a 147-item semi-quantitative FFQ, considering the past year, via face-to-face interview. The reliability and validity of FFQ have previously been approved in Iran(Reference Mirmiran, Hosseini-Esfahani and Jessri21). Data were analysed by Nutritionist-IV software, and portion sizes were converted to grams. Dietary macronutrients (protein, carbohydrate and fibre), vitamins (A,D,K,B1,B2,B3,B6,B9,B12 andC) and minerals (Ca and Mg) were computed from total intake. Individuals with daily energy intakes lower than 500 kcal or more than 3500 kcal were excluded from the study(Reference Estruch, Martinez-Gonzalez and Corella22).
Migraine diagnosis
Headaches/migraine were categorised according to the International Classification of Headache Disorders (ICHD–III) criteria(23) (beta version) (3–15 migraine days per month for at least 3 months). The validated Migraine Disability Assessment (MIDAS) questionnaire and visual analogue scale (VAS) were used for assessing migraine intensity (Reference Price, McGrath and Rafii24–Reference Ghorbani, Abtahi, Fereidan-Esfahani, Abtahi, Shemshaki, Akbari and Mehrabi-Koushki26). MIDAS has five questions to evaluate the intensity of headaches over the last 3 months, where the questionnaire evaluates the drop in performance caused by migraine. As a result, the patients are divided into four groups: 0–5 (MIDAS Grade I, Little or no disability), 6–10 (MIDAS Grade II, Mild disability), 11–20 (MIDAS Grade III, Moderate disability) or 21+ (MIDAS Grade IV, Severe disability). In addition, the VAS was depicted as a 10-point scale, ranging from: mild pain (Reference Pescador Ruschel and De Jesus1–Reference Burch, Buse and Lipton3), moderate pain (score 4–7) and severe pain (score 8–10). Also, the duration of the migraine headaches was discerned. Duration of headaches were defined as the hours the participants had headache in 1 d in the last month.
Statistical analyses
Data analyses were conducted using SPSS (version 20.0; SPSS Inc.). The normality of variables was assessed using Kolmogorov–Smirnov test. Nutrient patterns were derived using principal component analysis with varimax rotation and based on the correlation matrix. Fifteen nutrients were selected for factor analysis, including vitamin D, vitamin A, vitamin K, vitamin C, vitamin B6, vitamin B2, vitamin B1, vitamin B3, vitamin B9, vitamin B12, carbohydrate, protein, total fibre, Mg and Ca. Factor scores for all participants for each of the extracted factors were calculated by summing the frequency of consumption, multiplied by factor loadings across nutrients. Accordingly, three patterns were extracted. The first nutrient pattern consisted of Ca, vitamin A, vitamin K, vitamin C, vitamin B6, vitamin B2 and Mg. The second nutrient pattern included vitamin B1, carbohydrate, vitamin B3, vitamin B9, protein and total fibre. The third tertile included vitamin D and vitamin B12 based on scree plot (eigenvalue > 1) evaluation. We categorised them into tertiles. The χ 2 test was used to assess the association between the tertiles of nutrient patterns and qualitative variables. In addition, one-way ANOVA was used to assess the association between the tertiles of nutrient patterns and quantitative variables. Linear regression was also applied, in crude and adjusted models (model 1: crude model; model 2: (adjusted for age); model 3: (adjusted for age, physical activity, BMI); model 4: (adjusted for model 3+ education, marital status, job)). P < 0·05 was, a priori, considered to represent statistical significance. Moreover, positive beta values indicated that higher adherence of nutrient patterns increase the changes in the dependent variables and vice versa.
Results
Study population
The baseline characteristics of the participants among tertiles of nutrient patterns are shown in Table 1. Quantitative and qualitative variables across nutrient patterns tertiles did not show any significant differences.
Table 1. General characteristics of participants across tertiles of nutrient patterns scores
WC, waist circumference; HC, hip circumference; WHR, waist-to-hip ratio; PA, physical activity; MET, metabolic equivalents; MIDAS, Migraine Disability Assessment; VAS, visual analogue scale.
P-values were obtained from ANOVA test and χ 2 test. Qualitative variables were showed by number (percentage).
P-values resulted from one-way ANOVA analysis. Quantitative variables were showed by means ± sd. P-value < 0·05 was significant.
Dietary nutrient pattern among food groups and nutrients intake
Food groups and nutrient intakes across tertiles of nutrient patterns are shown in Table 2. Vitamin A, K, C, B6, B2, B9, total fibre, Mg and Ca intakes increased across tertiles of first nutrient pattern (P < 0·05). Vitamin B1, D, B12 and carbohydrate intakes decreased across tertiles of first nutrient pattern (P < 0·05). Vitamin B6, B2, B1, B3, B9, B12, carbohydrate, protein, total fibre and Mg intakes increased across tertiles of second nutrient pattern (P < 0·05). Vitamin D, B3, B12, carbohydrate and protein intakes increased across tertiles of last nutrient pattern (P < 0·05). Vitamin K, C, B9, total fibre and Ca intakes decreased across tertiles of last nutrient pattern (P < 0·05). Three dominant patterns were ascertained (Table 3), which explained a large proportion of the variance. The first nutrient pattern consisted of Ca, vitamin A, vitamin K, vitamin C, vitamin B6, vitamin B2 and Mg. The second nutrient pattern included vitamin B1, carbohydrate, vitamin B3, vitamin B9, protein and total fibre. The third tertile included vitamin D and vitamin B12.
Table 2. Total intakes across tertiles of nutrient patterns scores
RAE, retinol activity equivalents.
P-values were obtained from ANOVA test. Data are presented as mean ± standard deviation.
Table 3. Principal factor loading of nutrients intake
RAE, retinol activity equivalents.
Factor loadings of < 0·5 have been removed to simplify the table. Extraction method: principal component analysis. Rotation method: Varimax with Kaiser Normalization. a. Rotation converged in four iterations.
Nutrient patterns and intensity and duration of migraine headaches
The association between nutrient patterns and MIDAS and VAS among women is demonstrated in crude and adjusted models in Table 4, 5 and 6. There was a significant positive relationship between the second pattern and VAS in the crude model (β 0·27; 95 % CI 0·00, 0·53; P-value = 0·05), which remained significant in model 4, which was adjusted for all confounders (β 0·37; 95 % CI 0·13, 0·61; P-value =< 0·001). In addition, there was an inverse relationship between MIDAS and the first nutrient pattern (β –2·80; 95 % CI –5·20, –0·41; P-value = 0·02), which remained in model 4 (β –3·14; 95 % CI –5·47, –0·81; P-value = 0·01). Duration of headaches were positively related with second and third pattern. This remained after adjusting for confounders in all four models. The present study did not show any statistically significant association between other nutrient patterns and VAS and MIDAS and pain duration.
Table 4. Association of nutrient pattern and VAS among subjects with migraine
PA, physical activity.
* Linear regression; VAS, visual analogue scale; M1, (crud); M2, (adjusted for age); M3, (adjusted for age, PA, BMI); M4, (adjusted for M3+ education, marital status, job).
Table 5. Association of nutrient pattern and MIDAS among subjects with migraine
PA, physical activity.
* Linear regression; MIDAS, Migraine Disability Assessment; M1, (crud); M2, (adjusted for age); M3, (adjusted for age, PA and BMI); M4, (adjusted for M3+ education, marital status and job).
Table 6. Association of nutrient patterns and headache duration among subjects with migraine
PA, physical activity.
* Linear regression; M1, (crud); M2, (adjusted for age); M3, (adjusted for age, PA and BMI); M4, (adjusted for M3+ education, marital status, job).
Discussion
In the current cross-sectional study, there was a significant relationship between the second identified dietary pattern (vitamin B1, carbohydrate, vitamin B3, vitamin B9, protein and total fibre) and VAS, which is used for pain intensity and headache duration. Furthermore, there was a negative relationship between MIDAS, which is used for migraine disability, and the first identified nutrient pattern (Ca, vitamin A, vitamin K, vitamin C, vitamin B6, vitamin B2 and Mg) among women. Furthermore, there we found a significant positive association between vitamin D and B12 (pattern 3) and headache duration.
Several factors may contribute to the incidence of migraine including family history(Reference Russell and Olesen27), age(Reference Lipton and Bigal28), obesity(Reference Bigal, Liberman and Lipton29), sex hormonal change(Reference Finocchi and Strada30), vascular changes(Reference Hamed31), neurotransmitters(Reference Ghosh, Pradhan and Mittal32), low socio-economic status(Reference Winter, Berger and Buring33), head injury(Reference Chuang, Lin and Lin34), alcohol(Reference Panconesi, Bartolozzi and Guidi35), stressful situations and dietary intakes(Reference Hamed31,Reference Tietjen, Brandes and Peterlin36) . Dietary nutrients intakes also have long-standing association with the risk of migraine in both sexes, but especially in women(Reference Askarpour, Sheikhi and Khorsha37). For instance, although of contemporary practical interest, a paucity of studies have evaluated the association between the dietary intake and migraine. Of the available literature, the findings are often contradictory; some cross-sectional studies have demonstrated that high intake of chocolate, caffeine, milk, cheese and alcoholic beverages can increase the risk of migraine(Reference Millichap and Yee38–Reference Rist, Buring and Kurth41). But in our study, we found that Ca intake (rich in milk and cheese) was negatively related with migraine disability. Ca and Mg, together, can used for the development of many neurotransmitters and inflammatory mediators. These two minerals can be helpful for the nervous system function and reduce nerve tension(Reference Meng, Wang and Kang42). Adding to the equivocal findings in the literature, the National Health and Nutrition Examination Surveys (NHANES) of America revealed that high dietary intake of Ca and Mg were inversely associated with migraine in women(Reference Meng, Wang and Kang42). A cross-sectional study also revealed that the Dietary approaches to Stop Hypertension (DASH) adherence, which is rich in Mg and Ca, may be associated with lower headache severity and duration in migraine among female patients(Reference Mirzababaei, Khorsha and Togha43). Furthermore, a systematic literature review revealed that high intake of nuts, citrus fruits, processed meats, monosodium glutamate, aspartame and fatty foods can contribute to migraine(Reference Hindiyeh, Zhang and Farrar44). In line with the aforementioned study, higher consumption of protein (animal- and plant-based) and carbohydrate (simple and complex) may contribute to more headache intensity. A cross-sectional study revealed that high intakes of red meat increases the risk of migraine(Reference Le, Tfelt-Hansen and Skytthe45), whilst, in addition, several studies have reported that high carbohydrates intake causes the migraine among adult population due to insulin level changes and neuronal excitability(Reference Razeghi Jahromi, Ghorbani and Martelletti15,Reference Hindiyeh, Zhang and Farrar44,Reference Midoun, Salem and Akelah46,Reference Bigal and Lipton47) . Total fibre intake, which is high in fruits and vegetables (pattern 2), was positively associated with migraine. Similarly, Silva-Néto et al. found that many plant foods, which are high fibre, can be related with migraine headaches(Reference Silva-Néto, de Almeida Soares and Augusto Carvalho de Vasconcelos48). However, contrary with this study, contemporary research has highlighted that Mediterranean-based diet (rich in vitamin K and C and Mg) adherence was not associated with the risk of migraine(Reference Askarpour, Sheikhi and Khorsha49,Reference Chrysohoou, Panagiotakos and Pitsavos50) . Indeed, our study reinforces the idea that vitamin K may be related with arterial stiffness(Reference Mansour, Ahdab and Daaboul51) and vitamin C may act as a prophylactic agent(Reference Shaik and Gan17).
In the current study, there was a positive association between higher consumption of vitamin B12, B9, and D and duration of migraine headaches. However, total intake of vitamin D was less than the recommended reference intake (RDA), which is 10–20 mcg among participants(Reference Veugelers and Ekwaru52). Rich sources of vitamin D are meat, fish and egg yolk. Red meat may aggregate migraine headaches via tyramine content(Reference Rist, Buring and Kurth41). Further, vitamin D deficiency has been reported to increase the duration of migraine headaches(Reference Ghorbani, Togha and Rafiee14), whilst a cross-sectional study demonstrated that lower consumption of B9 and B12 are associated with increasing the risk of migraine through MTHFR gene mutation(Reference Shaik and Gan17,Reference Schürks, Zee and Buring53) . The main source of vitamin B12 in this population is animal-based proteins, which have been reported to have a positive relationship with migraine headaches. Moreover, a case–control study demonstrated that lower dietary intake of folate was not associated with the risk of migraine(Reference Ferro, Castro and Lemos54). Several studies have reported that supplementation with pyridoxine and folate, not only, reduces the homocysteine levels but also improves the migraine symptoms(Reference Sadeghi, Nasiri and Maghsoudi55,Reference Askari, Nasiri and Mozaffari-Khosravi56) . Concordant with our study, many studies revealed that lower intake of vitamin B6 and B2 is associated with the risk of migraine(Reference Miya Shaik, Lin Tan and Kamal57). Indeed, we found that consumption of foods high in Ca, Mg, vitamin B2, C, B6 and A may reduce the headaches in migraine patients.
To the best of our knowledge, this is the first study to have investigated the relationship between dietary nutrients patterns and intensity and duration of migraine headaches among women in Iran, and very few studies have been conducted pertaining to dietary patterns and migraine(Reference Evans, Lipton and Peterlin58). The present study outcomes confirm the importance of identification of triggers for the management of patients with migraine disorder and emphasises the consumption or avoidance of specific dietary nutrient patterns. This study can help to guide and inform insight into the relationship between nutrients and migraine. However, despite the novel addition to the literature, there were some limitations that warrant consideration. Although the cross-sectional design of the study can show us the relationships between outcome and exposure, the present study design precludes causal inferences being drawn between dietary nutrients patterns and intensity and duration of migraine headaches. In addition, factor analysis (varimax method) was used to identify patterns, and limitations of this method may be responsible for bias in our study. The number of derived factors is optional, and it depends on the decisions of the researchers. Clearly, further prospective longitudinal studies are needed to more firmly establish the impact of, and association between, dietary nutrient patterns and migraine occurrence, severity and duration, respectively.
Conclusion
Our study found a significant relationship between the second identified dietary pattern, which included vitamin B1, carbohydrate, vitamin B3, vitamin B9, protein, and total fibre and VAS and pain duration. Furthermore, we found a relationship between MIDAS and the first nutrient pattern, characterised by Ca, vitamin A, vitamin K, vitamin C, vitamin B6, vitamin B2 and Mg, among women. Additionally, there was a significant association between the third nutrient pattern (vitamin D and B12) and pain duration. Overall, this research demonstrates that dietary nutrients patterns should be monitored closely in individuals suffering with migraine.
Acknowledgements
The authors thank the study participants for their cooperation and assistance in physical examinations. Research number was 95-01-103-31348. Consent forms were collected by all of the participants. The authors also thank those involved in nutritional evaluation and database management, as well as the neurology clinics of Sina and Khatam Alanbia hospitals.
This study was supported by the Tehran University of Medical Sciences, and by grants from Tehran University of Medical Sciences (Grants ID: 95–01–103–31 348 and 93-23-42-1001).
The authors report no conflict of interest.
K.M. and N.B. conceptualised the study. F.K. gathered the data of study. N.B. wrote the first draft of the manuscript and revised the manuscript. H.Y. did the statistical analyses. A.M.B. wrote the first draft of discussion section. A.M. and K.M. were checked the manuscript. C.C.T.C. and N.B. reviewed and edited the paper. All authors read and approved the final version of the manuscript.
