To characterize meal patterns across ten European countries participating in the European Prospective Investigation into Cancer and Nutrition (EPIC) calibration study.

Cross-sectional study utilizing dietary data collected through a standardized 24 h diet recall during 1995–2000. Eleven predefined intake occasions across a 24 h period were assessed during the interview. In the present descriptive report, meal patterns were analysed in terms of daily number of intake occasions, the proportion reporting each intake occasion and the energy contributions from each intake occasion.

Twenty-seven centres across ten European countries.

Women (64 %) and men (36 %) aged 35–74 years (n 36 020).

Pronounced differences in meal patterns emerged both across centres within the same country and across different countries, with a trend for fewer intake occasions per day in Mediterranean countries compared with central and northern Europe. Differences were also found for daily energy intake provided by lunch, with 38–43 % for women and 41–45 % for men within Mediterranean countries compared with 16–27 % for women and 20–26 % for men in central and northern European countries. Likewise, a south–north gradient was found for daily energy intake from snacks, with 13–20 % (women) and 10–17 % (men) in Mediterranean countries compared with 24–34 % (women) and 23–35 % (men) in central/northern Europe.

We found distinct differences in meal patterns with marked diversity for intake frequency and lunch and snack consumption between Mediterranean and central/northern European countries. Monitoring of meal patterns across various cultures and populations could provide critical context to the research efforts to characterize relationships between dietary intake and health.

To examine the association between BMI and folate concentrations in serum and red blood cells (RBC) in pregnant women.

A cross-sectional comparison of folate concentrations in serum and RBC sampled simultaneously from the same individual.

The Ottawa Hospital and Kingston General Hospital, Ontario, Canada.

Pregnant women recruited between 12 and 20 weeks of gestation.

A total of 869 pregnant women recruited from April 2008 to April 2009 were included in the final analysis. Serum folate was inversely associated and RBC folate positively associated with BMI, after adjusting for folic acid supplementation, age, gestational age at blood sample collection, race, maternal education, annual income, smoking and MTHFR 677C→T genotype. In stratified analyses, this differential association was significant in women with the MTHFR CC variant. In women with the CT and TT variants, the differential associations were in the same direction but not significant. Folic acid supplementation during pregnancy did not alter the differential association of BMI with serum and RBC folate concentration. This indicates that the current RBC folate cut-off approach for assessing risk of neural tube defects in obese women may be limited.

BMI is inversely associated with serum folate and positively associated with RBC folate in pregnant women, especially for those with the MTHFR CC variant.

The Mediterranean diet (MD) pattern has important health benefits; however, it seems that Spanish school-aged children have been abandoning this healthy pattern recently. We aimed to identify psychopathological, anthropometric and sociodemographic factors that may influence the risk of low MD adherence.

Longitudinal study in three phases. MD adherence was assessed using the Krece Plus food questionnaire and psychopathological symptoms using the Screen for Childhood Anxiety Related Emotional Disorders, Children’s Depression Inventory, Youth’s Inventory-4 and Eating Disorder Inventory-2. Anthropometric data were collected in the first and third phases.

All five representative areas in Reus, Spain.

Adolescents (n 241).

Regardless of past and current BMI, socio-economic status was a protective factor for low MD adherence (OR=0·805, P=0·003) and a risk factor for high BMI (OR=0·718, P=0·002; OR=0·707, P=0·001). Regardless of socio-economic status, depression was involved with risk of low adherence (OR=1·069, P=0·021). Girls with lower MD adherence presented significantly higher scores for eating disorders measured using the Eating Disorder Inventory-2 (low adherence, mean 18·9 (sd 13·5); high adherence, mean 8·9 (sd 9·0), P=0·020) and the Youth Inventory-4 (low adherence, mean 5·2 (sd 4·3); medium adherence, mean 3·6 (sd 3·2), P=0·044). They also presented higher depression symptoms (low adherence, mean 17·7 (sd 9·6); medium adherence, mean 12·3 (sd 7·2), P=0·01) than girls with high adherence.

The results highlight the influence of psychosocial factors on levels of MD adherence. These factors need to be taken into account when developing prevention and health promotion initiatives.

The increasing prevalence of overweight and obesity among children is now an important health problem. This fact, however, does not reflect the scale of the problem. The aim of the present study was to find how much the BMI threshold was exceeded in a population from Kraków.

The study was based on three cross-sectional surveys conducted in 1983, 2000 and 2010. The prevalence of overweight and obesity was estimated based on the International Obesity Task Force cut-off points. In addition to these, an extent of overweight (EOW) index was calculated.

Poland.

Children aged 3–18 years (n 14 534) from Kraków.

Between the populations examined in 1983 and 2010, the EOW index in boys rose by almost 10 %, and the prevalence of overweight and obesity by 39 %. In girls, however, the EOW index decreased by 45 %, while the prevalence of overweight and obesity remained at similar levels. Analyses in separate age groups showed that the EOW index increased only among early adolescents (150 % for boys, 94 % for girls) and late adolescents (390 % and 64 %, respectively).

The observed increased prevalence of overweight and obesity mainly concerned boys and was accompanied by an increase in the amount by which the BMI threshold values were exceeded.

To analyse trends in maternal nutritional status in Bangladesh over a 12-year period and to examine the associations between nutritional status and socio-economic variables.

Maternal nutritional status indicators were height, weight and BMI. Socio-economic variables used were region, residency, education and occupation of the mothers and their husbands, house type, and possession score in the household.

Bangladesh Demographic and Health Surveys (1996, 2000, 2004 and 2007) were the source of data.

A total of 16 278 mothers were included.

All of the socio-economic variables showed significant associations with maternal nutritional status indicators. Regional variation was found to be present; all three indicators were found to be lowest in the Sylhet division. Upward trends in maternal height, weight and BMI were evident from no possessions to four possessions in households, and for no education to higher education of women and their husbands. Bangladeshi mothers measured in 2007 were found to be on average 0·34 cm taller and 3·36 kg heavier than mothers measured in 1996. Between 1996 and 2007 maternal underweight fell from nearly 50 % to just over 30 % while overweight and obesity increased from about 3 % to over 9 % (WHO cut-offs) or from 7 % to nearly 18 % (Asian cut-offs).

The study reveals that over the 12-year period in Bangladesh there has been a substantial reduction in maternal underweight accompanied by a considerable increase in obesity. It is also evident that malnutrition in Bangladesh is a multidimensional problem that warrants a proper policy mix and programme intervention.

To examine how different definitions of meals and snacks can affect the associations of meal frequency (MF) and snack frequency (SF) with dietary intake and adiposity measures.

Based on 7 d weighed dietary record data, all eating occasions providing ≥210 kJ of energy were divided into meals or snacks based on contribution to energy intake (≥15 % or <15 %) or time (06.00–10.00, 12.00–15.00 and 18.00–21.00 hours; other). Diet quality was assessed using the Healthy Diet Indicator (HDI) and Mediterranean Diet Score (MDS).

Great Britain.

British adults aged 19–64 years (n 1487).

MF based on energy contribution was associated with higher intake of dietary fibre, lower intakes of non-milk extrinsic sugars and alcohol, and higher HDI (only men) and MDS. MF based on time was associated with higher HDI and MDS in women only. Conversely, irrespective of the definition of snacks, SF was associated with higher intakes of confectionery and alcohol, lower intakes of cereals, protein, fat and dietary fibre, and lower HDI (except for SF based on energy contribution in women) and MDS. After adjustment for potential confounders, MF based on time, but not MF based on energy contribution, was positively associated with BMI and waist circumference in men only. SF was positively associated with BMI and waist circumference, irrespective of the definition of snacks.

Higher SF was consistently associated with lower diet quality and higher adiposity measures, while associations with MF varied depending on the definition of meals and sex.

To evaluate the impact on food purchasing behaviour of the ‘Change4Life Smart Swaps’ campaign to encourage families to make small changes to lower-fat or lower-sugar versions of commonly eaten foods and drinks.

Quasi-experimental study comparing the proportion of swaps made by an intervention group (267 families who had signed up to the ‘Smart Swaps’ campaign promoted through various media, including television and radio advertising in early 2014) and a comparison group (135 families resident in Wales, signed up for ‘Change4Life’ materials, but not directly exposed to the ‘Smart Swaps’ campaign). During weeks 1, 2 and 3 of the campaign participants were asked to record their purchases of dairy products, carbonated drinks and breakfast cereals, using a mobile phone app questionnaire, when making a purchase within the category.

England and Wales.

Families registered with ‘Change4Life’.

In weeks 2 and 3 a significantly higher percentage of the intervention group had made ‘smart swaps’ than the comparison group. After week 3, 58 % of participants had swapped to a lower-fat dairy product compared with 26 % of the comparison group (P<0·001), 32 % of the intervention group had purchased a lower-sugar drink compared with 19 % of the comparison group (P=0·01), and 24 % had made a change to a lower-sugar cereal compared with 12 % of the comparison group (P=0·009).

In the short term a national campaign to change purchase habits towards healthier products may have some merit but the sustainability of change requires further investigation.

To summarize the best available evidence regarding the short- and long-term health effects of cow’s milk intake in healthy, full-term infants up to 3 years of age.

We conducted a systematic review and meta-analysis.

We searched MEDLINE (via PubMed), EMBASE and the Cochrane Library between 1960 and July 2013 and manually reviewed reference lists of pertinent articles. Two researchers independently reviewed abstracts and full-text articles and extracted relevant data.

We included (randomized/non-randomized) controlled trials and observational studies.

We included data from twenty-three studies (one randomized controlled trial, four non-randomized controlled trials, eight case–control studies and ten cohort studies) for the evidence synthesis. Pooled results of four studies revealed a higher risk of Fe-deficiency anaemia for infants consuming cow’s milk compared with those consuming follow-on formula (relative risk=3·76; 95 % CI 2·73, 5·19). For type 1 diabetes mellitus, six out of seven case–control studies did not show a difference in the risk of developing this disease based on the age of introduction of cow’s milk. We did not find negative associations for other health effects.

Cow’s milk consumption in infancy is associated with an increased risk of developing Fe-deficiency anaemia. Limiting cow’s milk consumption may be important to ensure an adequate Fe intake for infants and toddlers. High-quality patient information for caregivers is needed on how infants’ Fe requirements can be met.

The specific role of major socio-economic indicators (education, occupation, income) in influencing consumer choice of animal foods (AF) intake could reveal distinct socio-economic facets, thus enabling elucidation of mechanisms leading to social inequalities in health. We investigated the independent association of each indicator with intake of different AF and their effect modification.

Cross-sectional study. AF intake was estimated using three 24 h dietary records. Associations between socio-economic factors and AF intake and interactions between socio-economic indicators were assessed using ANCOVA adjusted for age and energy intake. Analyses were performed separately for men and women, since gender interactions were found.

France.

Adults (n 92 036) participating in the NutriNet-Santé Study.

Low educated persons had higher intake of red meat (+9–12 g/d), processed meat (+6–9 g/d) and poultry (for men, +7 g/d) than those with a higher education level. Percentage of fish consumers was lower in individuals of the lowest income class compared with those in higher classes. Manual workers had a higher intake of cream desserts (for men, +14 g/d) than managerial staff. Few significant interactions were found. In stratified analyses, persons with the highest income consumed more yoghurt than those who had lower income, only in low educated individuals.

Socio-economic disparities in AF intake varied according to the socio-economic indicator, suggesting the specific influence of each indicator on AF intake. In particular, lower education was associated with higher intake of red and processed meats and cream desserts, and had an effect modification on the relationship between income and AF intake.

The provision of prescribed vitamin D to all aged-care residents has been implemented in New Zealand as part of a government-led falls prevention programme. To our knowledge, there has been no evaluation of this universal programme on vitamin D status and functional and health outcomes. Thus, we aimed to determine 25-hydroxyvitamin D (25(OH)D) concentrations and their predictors in aged-care residents across the country and to investigate whether the government-funded programme was associated with adequate vitamin D status.

Cross-sectional survey of sociodemographic, biochemical, anthropometric, dietary and health characteristics. Blood samples were analysed for serum 25(OH)D and other biochemical measures. Multiple regression was used to examine predictors of vitamin D status.

Sixteen residential aged-care facilities throughout New Zealand.

Residents aged ≥60 years with residency duration >12 weeks (n 309).

Mean serum 25(OH)D was 89·9 (95 % CI 85·2, 94·5) nmol/l and monthly supplements (1250 µg (50 000 IU)) were taken by 75 % of all residents. Of those not taking a funded supplement, 65·3 % had serum 25(OH)D <50 nmol/l compared with only 1·5 % of supplement users. Being female, residing at lower latitude, increasing duration of aged-care residency and raised serum α1-acid glycoprotein were positively associated with higher 25(OH)D concentrations. Supplemental vitamin D from all sources was the strongest predictor, increasing serum 25(OH)D levels by more than 70 nmol/l. Furthermore, 25 % of participants had serum 25(OH)D levels >125 nmol/l.

Residents taking supplemental vitamin D had adequate vitamin D status; however monitoring of long-term supplementation should be considered, due to the high proportion of participants with high serum 25(OH)D levels.

To investigate the direct and indirect associations between psychological and physical intimate partner violence and the occurrence of common mental disorders (CMD) and how they relate to the occurrence of household food insecurity (HFI).

This was a population-based cross-sectional study. Intimate partner violence was assessed using the Brazilian version of the Revised Conflict Tactics Scale (CTS2) and HFI was assessed using the Brazilian Food Insecurity Scale. The propositional analytical model was based on a review of the literature and was tested using path analysis.

Duque de Caxias, Greater Rio de Janeiro, Brazil (April–December 2010).

Women (n 849) who had been in a relationship in the 12 months preceding the interview.

Both psychological and physical violence were found to be major risk factors of HFI. Psychological violence was associated with HFI indirectly via physical violence and CMD, and directly by an unidentified path. The effects of physical violence seemed to be manifested exclusively through CMD. Most of the variables in the propositional model related to socio-economic position, demographic characteristics, degree of women’s social support and partner alcohol misuse were retained in the ‘final’ model, indicating that these factors contribute significantly to the increased likelihood of HFI.

The results reinforce the importance of considering domestic violence and other psychosocial aspects of family life when implementing interventions designed to reduce/eradicate HFI.

We examined associations between fast-food intake and perceived and objective fast-food outlet exposure.

Information from the Health Behaviours in School-aged Children Study was linked to fast-food outlets in seventy-five school neighbourhoods. We used multivariate multilevel logistic regression analyses to examine associations between at least weekly fast-food intake and perceived and objective fast-food outlet measures.

Data represent 4642 adolescents (aged 11–15 years) in Denmark.

Boys reporting two or more fast-food outlets had 34 % higher odds consuming fast food at least weekly. We detected higher odds of at least weekly fast-food intake among 15-year-old 9th graders (ORall=1·74; 95 % CI 1·40, 2·18; ORboys=2·20; 95 % CI 1·66, 2·91; ORgirls=1·41; 95 % CI 1·03, 1·92), Danish speakers (ORall=2·32; 95 % CI 1·68, 3·19; ORboys=2·58; 95 % CI 1·69, 3·93; ORgirls=2·37; 95 % CI 1·46, 3·84) and those travelling 15 min or less to school (ORall=1·21; 95 % CI 1·00, 1·46; ORgirls=1·44; 95 % CI 1·08, 1·93) compared with 11-year-old 5th graders, non-Danish speakers and those with longer travel times. Boys from middle- (OR=1·28; 95 % CI 1·00, 1·65) and girls from low-income families (OR=1·46; 95 % CI 1·05, 2·04) had higher odds of at least weekly fast-food intake compared with those from high-income backgrounds. Girls attending schools with canteens (OR=1·47; 95 % CI 1·00, 2·15) had higher odds of at least weekly fast-food intake than girls at schools without canteens.

The present study demonstrates that perceived food outlets may impact fast-food intake in boys while proximity impacts intake in girls. Public health planning could target food environments that emphasize a better understanding of how adolescents use local resources.

The objective was to investigate whether irregular eating of meals in adolescence predicts the metabolic syndrome and its components in adulthood, and if any specific meal is of particular importance.

Prospective cohort study with 27 years of follow-up. Information on meals (breakfast, school lunch and dinner with family), lifestyle (alcohol consumption, smoking habits, physical activity, consumption of sweets and pastries) at age 16 years was assessed from questionnaires, and presence or not of the metabolic syndrome and its components were defined at age 43 years in 889 participants (82·1 % of total cohort). Logistic regression was used to calculate odds ratios and confidence intervals.

The Northern Swedish Cohort; all school-leavers of the 9th grade in the town Luleå in 1981.

Adolescents (age 16 years).

Irregular eating of meals at age 16 years was associated with higher prevalence of the metabolic syndrome at age 43 years (OR=1·74; 95 % CI 1·12, 2·71), but this was explained by concurrent unhealthy lifestyle at age 16 years. Poor breakfast at age 16 years was the only meal associated with the metabolic syndrome at age 43 years, independent of other meals, BMI (kg/m2) and lifestyle at age 16 years (OR=1·67; 95 % CI 1·00, 2·80).

Irregular eating of meals in adolescence predicted the metabolic syndrome in adulthood, but not independently of BMI and lifestyle in adolescence. Poor breakfast in adolescence was the only specific meal associated with future metabolic syndrome, even after adjustments. Breakfast eating should be encouraged in adolescence.

In India, national databases indicate anaemia prevalence of 80 % among 6–35-month-old children and 58 % among 36–59-month-old children. The present study aimed to characterise anaemia and the associated factors among infants and pre-schoolers living in rural India.

Multivariate logistic regression analysis of data collected prior to an intervention trial. Fe-deficiency with anaemia (IDA), Fe deficiency with no anaemia (IDNA) and anaemia without Fe deficiency were defined. Serum ferritin, soluble transferrin receptor (sTfR) and sTfR/log ferritin index were used to indicate Fe status.

Twenty-six villages of Nalgonda district, Telangana, India. Data were collected in community sites.

Four hundred and seventy-six infants (aged 6–12 months), 316 pre-schoolers (aged 29–56 months) and their mothers.

Prevalence of anaemia among infants and pre-schoolers was 66·4 and 47·8 %, prevalence of IDA was 52·2 and 42·1 %, prevalence of IDNA was 22·2 and 29·8 %, prevalence of anaemia without Fe deficiency was 14·2 and 5·7 %. Among infants, anaemia was positively associated with maternal anaemia (OR=3·31; 95 % CI 2·10, 5·23; P<0·001), and sTfR/log ferritin index (OR=2·21; 95 % CI 1·39, 3·54; P=0·001). Among pre-schoolers, anaemia was positively associated with maternal anaemia (OR=3·77; 95 % CI 1·94, 7·30; P<0·001), sTfR/log ferritin index (OR=5·29; 95 % CI 2·67, 10·50; P<0·001), high C-reactive protein (OR=4·39; 95 % CI 1·91, 10·06, P<0·001) and young age (29–35 months: OR=1·92; 05 % CI 1·18, 3·13, P=0·009).

Anaemia prevalence continues to be high among infants and pre-schoolers in rural India. Based on sTfR/ferritin index, Fe deficiency is a major factor associated with anaemia. Anaemia is also associated with inflammation among pre-schoolers and with maternal anaemia among infants and pre-schoolers, illustrating the importance of understanding the aetiology of anaemia in designing effective control strategies.

Sociodemographic factors may affect adherence to specific dietary patterns during pregnancy. The present study aimed to identify dietary patterns during pregnancy and associated factors among Brazilian pregnant women.

A cross-sectional analysis. Dietary intake was evaluated with a semi-quantitative FFQ during the first postpartum week; the time frame included the second and third gestational trimesters. Principal component analysis was used to identify dietary patterns during pregnancy. Sociodemographic data were obtained using a structured questionnaire. Multiple linear regressions were applied to test the associations between the sociodemographic factors and dietary patterns.

Mesquita, Rio de Janeiro, Brazil, 2011.

Postpartum women (n 327) who were 18–45 years of age and Mesquita residents.

Three different dietary patterns were identified: ‘healthy’ (mainly comprising legumes, vegetables and fruits), ‘mixed’ (mainly comprising candy, butter and margarine, and snacks) and ‘traditional’ (mainly comprising beans and rice). Women with a higher monthly per capita family income (β=0·0006; 95 % CI 0·0001, 0·001; P=0·011) and women of older age (β=0·021; 95 % CI −0·001, 0·042; P=0·058) were more likely to adhere to the ‘healthy’ dietary pattern. Women with higher parity were less likely to adhere to the ‘healthy’ pattern (β=−0·097; 95 % CI −0·184, −0·009; P=0·030) and were more likely to adhere to the ‘traditional’ pattern (β=0·098; 95 % CI 0·021, 0·175; P=0·012). Although not statistically significant, older women were less likely to adhere to the ‘mixed’ (β=−0·017; 95 % CI −0·037, 0·003; P=0·075) and ‘traditional’ (β=−0·018; 95 % CI −0·037, 0·001; P=0·061) dietary patterns.

Monthly per capita family income, parity and maternal age were factors associated with adherence to a healthy diet during pregnancy.

To investigate the feasibility and acceptability of using wearable cameras as a method to capture the opportunities for food and drink purchasing/consumption that young people encounter on their regular journeys to and from school.

A qualitative study using multiple data-collection methods including wearable cameras, global positioning system units, individual interviews, food and drink purchase and consumption diaries completed by participants over four days, and an audit of food outlets located within an 800 m Euclidean buffer zone around each school.

A community setting.

Twenty-two students (fourteen girls and eight boys) aged 13–15 years recruited from four secondary schools in two counties of England.

Wearable cameras offered a feasible and acceptable method for collecting food purchase and consumption data when used alongside traditional methods of data collection in a small number of teenagers. We found evidence of participants making deliberate choices about whether or not to purchase/consume food and drink on their journeys. These choices were influenced by priorities over money, friends, journey length, travel mode and ease of access to opportunities for purchase/consumption. Most food and drink items were purchased/consumed within an 800 m Euclidean buffer around school, with items commonly selected being high in energy, fat and sugar. Wearable camera images combined with interviews helped identify unreported items and misreporting errors.

Wearable camera images prompt detailed discussion and generate contextually specific information which could offer new insights and understanding around eating behaviour patterns. The feasibility of scaling up the use of these methods requires further empirical work.

Globally, the Na consumption of most people exceeds the WHO recommendation. To be effective, salt reduction programmes require assessment of the dietary sources of Na. Due to methodological difficulties however, comprehensive assessments are rare. Here, we identified Na sources in the Japanese diet using a 4 d diet record that was specifically designed for Na source description.

A cross-sectional study.

Apparently healthy men (n 196) and women (n 196) aged 20–69 years.

The subjects were recruited from twenty-three of forty-seven prefectures in Japan.

The proportion of discretionary Na intake in total Na intake was 52·3 % in men and 57·1 % in women, and was significantly lower in younger subjects. The two major food groups contributing to Na intake were seasonings such as salt or soya sauce (61·7 % of total Na intake in men, 62·9 % in women) and fish and shellfish (6·7 % in men, 6·6 % in women). The third major contributor differed between men and women (noodles in men, 4·9 %; bread in women, 5·0 %). Further, the contribution of each food group to total Na intake differed among age groups.

While individual efforts to decrease Na intake remain important, population approaches to reducing Na content in processed foods are already equally important and will assume greater importance in the future even in Japan, an Asian country facing a rapid Westernization in dietary habits.

The current meta-analysis evaluated the association between vitamin B12 intake and blood vitamin B12 level and colorectal cancer (CRC) risk.

The PubMed and EMBASE databases were searched. A dose–response analysis was performed with generalized least squares regression, with the relative risk (RR) and 95 % CI as effect values.

The meta-analysis included seventeen studies.

A total of 10 601 patients.

The non-linear dose–response relationship between total vitamin B12 intake and CRC risk was insignificant (P=0·690), but the relationship between dietary vitamin B12 intake and CRC risk was significant (P<0·001). Every 4·5 μg/d increment in total and dietary vitamin B12 intake was inversely associated with CRC risk (total intake: RR=0·963; 95 % CI 0·928, 0·999; dietary intake: RR=0·914; 95 % CI 0·856, 0·977). The inverse association between vitamin B12 intake and CRC risk was also significant when vitamin B12 intake was over a dosage threshold, enhancing the non-linear relationship. The non-linear dose–response relationship between blood vitamin B12 level and CRC risk was insignificant (P=0·219). There was an insignificant association between every 150 pmol/l increment in blood vitamin B12 level and CRC risk (RR=1·023; 95 % CI 0·881, 1·187).

Our meta-analysis indicates that evidence supports the use of vitamin B12 for cancer prevention, especially among populations with high-dose vitamin B12 intake, and that the association between CRC risk and total vitamin B12 intake is stronger than between CRC risk and dietary vitamin B12 intake only.

To prospectively assess the associations between lean fish, fatty fish and total fish intakes and risk of stroke in the Spanish cohort of the European Prospective Investigation into Cancer and Nutrition (EPIC-Spain).

Fish intake was estimated from a validated dietary questionnaire. Cox proportional hazards regression models were used to assess the association between the intakes of lean fish, fatty fish and total fish and stroke risk. Models were run separately for men and women.

Five Spanish regions (Asturias, San Sebastian, Navarra, Granada and Murcia).

Individuals (n 41 020; 15 490 men and 25 530 women) aged 20–69 years, recruited from 1992 to 1996 and followed-up until December 2008 (December 2006 in the case of Asturias). Only participants with definite incident stroke were considered as cases.

During a mean follow-up of 13·8 years, 674 strokes were identified and subsequently validated by record linkage with hospital discharge databases, primary-care records and regional mortality registries, comprising 531 ischaemic, seventy-nine haemorrhagic, forty-two subarachnoid and twenty-two unspecific strokes. After multiple adjustments, no significant associations were observed between lean fish, fatty fish and total fish consumption and the risk of stroke in men or women. In men, results revealed a non-significant trend towards an inverse association between lean fish (hazard ratio=0·84; 95 % CI 0·55, 1·29, Ptrend=0·06) and total fish consumption (hazard ratio=0·77; 95 % CI 0·51, 1·16, Ptrend=0·06) and risk of total stroke.

In the EPIC-Spain cohort, no association was found between lean fish, fatty fish and total fish consumption and risk of stroke.