To assess the magnitude and determinants of vitamin A deficiency (VAD) and coverage of vitamin A supplementation (VAS) among pre-school children.

A community-based cross-sectional study was carried out by adopting a multistage, stratified, random sampling procedure.

Rural areas of eight states in India.

Pre-school children and their mothers were covered.

A total of 71 591 pre-school children were clinically examined for ocular signs of VAD. Serum retinol concentrations in dried blood spots were assessed in a sub-sample of 3954 children using HPLC. The prevalence of Bitot spots was 0·8 %. The total ocular signs were significantly higher (P < 0·001) among boys (2·6 %) compared with girls (1·9 %) and in older children (3–4 years) compared (P < 0·001) with younger (1–2 years), and were also high in children of labourers, scheduled castes and illiterate mothers. The odds of having Bitot spots was highest in children of scheduled caste (OR = 3·8; 95 % CI 2·9, 5·0), labourers (OR = 2·9; 95 % CI 2·1, 3·9), illiterate mothers (OR = 2·7; 95 % CI 2·2, 2·3) and households without a sanitary latrine (OR = 5·9; 95 % CI 4·0, 8·7). Subclinical VAD (serum retinol level <20 μg/dl) was observed in 62 % of children. This was also relatively high among scheduled caste and scheduled tribe children. The rate of coverage of VAS was 58 %.

The study revealed that VAD is a major nutritional problem and coverage of VAS was poor. The important determinants of VAD were illiteracy, low socio-economic status, occupation and poor sanitation. Strengthening the existing VAS programme and focused attention on dietary diversification are essential for prevention of VAD.

To test whether there are differences in diet diversity between children still being partly breast-fed at 9 months and those completely weaned at the same age.

Cross-sectional study.

Cross-sectional study (SKOT cohort) in the area of Copenhagen, Denmark.

Healthy term infants (n 312) at 9 months of age (mean 9·1 (sd 0·3) months).

The infants partly breast-fed (n 168) at 9 months had significantly lower body weight (P < 0·0001), were significantly shorter (P = 0·0022) and were introduced to complementary foods significantly later (P < 0·0001) than completely weaned infants (n 141) of similar age. Furthermore, they had lower intake of energy, both in absolute amount (P < 0·0001) and per kilogram of body weight (P = 0·049). Significantly lower intakes of most energy-yielding nutrients, in absolute amounts and as energy percentages, were seen for the partly breast-fed compared with the completely weaned infants. These differences appear to be caused primarily by differences in the type and amount of milk consumed, as the energy derived from sources other than milk was similar except for fatty spread and vegetables as a side dish. Only small differences were found for absolute intakes of foods between feeding groups, although fatty spread had significantly higher intake rates and consumption (P = 0·031) among partly breast-fed compared with completely weaned infants.

At 9 months the infants partly breast-fed did not eat a less diversified diet compared with those completely weaned at the same age. Despite later introduction to complementary foods compared with the completely weaned, their intake of foods was similar and no delay in their progression towards the family foods was noted.

Little is known about the prevalence of nutritional supplement use in European adolescents. We conducted the present study to analyse the prevalence of nutritional supplement use and factors associated with this use among Slovenian adolescents.

The nutritional supplementation practices of 818 adolescents were studied using an anonymous questionnaire. Information was sought on the type of supplements used, frequency of use and sources of information.

The region of north-west Slovenia.

Schoolchildren from twenty primary schools and twelve secondary schools.

Some 19·3% of all adolescents reported using at least one nutritional supplement and the prevalence of use was significantly higher in adolescents who were members of sports clubs. Multivitamins were the most common nutritional supplement. Older adolescents were significantly more likely to be supplementing with iron, protein and minerals. Less than 16 % of supplement users in our study sought information from health-care professionals. Nearly 62 % obtained information from parents and coaches, and many adolescents appear to decide on nutritional supplementation themselves, without advice. Older adolescents were significantly more likely to combine supplements than younger adolescents.

One-fifth of Slovenian adolescents use nutritional supplements. There are clear differences in supplement use between younger (age 12 years) and older (age 17 years) adolescents. Multiple use of supplements, coupled with self-managed supplementation in older adolescents, is concerning. Hence, there is an urgent need to provide accurate information regarding nutritional supplements, which will help adolescents, their parents and coaches to make informed choices about their use.

The aim of the present study was to assess anthropometric status in South African children and women in 2005 in order to document temporal trends in selected anthropometric parameters.

Heights and weights were measured in a cross-sectional study of children aged 1–9 years and women aged 16–35 years. The WHO reference values and BMI cut-off points were used to determine weight status.

South Africa, representative sample based on census data.

Children (n 2157) and women (n 2403).

Stunting was the most common nutritional disorder affecting 21·7 % of children in 1999 and 20·7 % in 2005. The difference was not statistically significant. Underweight prevalence remained unchanged, affecting 8·1 % of children, whereas wasting affected 5·8 % of children nationally, a significant increase from 4·3 % of children in 1999. Rural children were most severely affected. According to the international BMI cut-off points for overweight and obesity, 10 % of children nationally were classified as overweight and 4 % as obese. The national prevalence of overweight and obesity combined for women was 51·5 %. The prevalence of overweight in children based on weight-for-height Z-score did not change significantly (8·0 % to 6·8 %, P = 0·138), but the combined overweight/obesity prevalence based on BMI cut-off points (17·1 % to 14·0 %, P = 0·02) decreased significantly from 1999 to 2005.

The double burden of undernutrition in children and overweight among women is evident in South Africa and getting worse due to increased childhood wasting combined with a high prevalence of obesity among urban women, indicating a need for urgent intervention.

To investigate the food intake of Slovenian adolescents and to compare it with food-based dietary guidelines developed for children and adolescents, named the Optimized Mixed Diet (OMD). The OMD is a useful tool for the evaluation of food intake of adolescents.

All adolescents completed an FFQ at a regional health centre; a subgroup also completed a 3 d weighed dietary protocol at home.

This study is a part of the first national representative study on the dietary habits of Slovenian adolescents.

This cross-sectional study included a representative sample of 2813 Slovenian adolescents entering high school, aged 14–17 years, from all ten geographical regions of Slovenia.

The greatest deviations from the recommended intakes of the main food groups in the OMD were significantly lower intakes of (P < 0·001, mean): vegetables (179 and 163 g/d in boys and girls, respectively), bread/cereals (271 and 226 g/d), potatoes/rice/pasta (212 and 163 g/d); in boys also a significantly lower intake of fruits (mean: 321 g/d, P < 0·001) and a significantly higher intake of meat/meat products (mean: 126 g/d, P < 0·001). Additionally, the results show too high intake of sugar-sweetened beverages and too low intakes of fish and plant oils in both genders.

The food intake pattern of Slovenian adolescents deviates markedly from a healthy eating pattern. Nutrition education and interventions are needed for Slovenian adolescents.

Examine the association between energy intake and television (TV) viewing in Americans.

Nationally representative, cross-sectional study of 2003–2006 National Health and Nutrition Examination Survey.

Total energy intake was determined by two 24 h recalls. TV viewing was reported as low (≤1 h/d), middle (2–3 h/d), and high (≥4 h/d). Multivariate linear regression models were used to analyse TV viewing and energy intake, adjusted for BMI (percentile for children 2–18 years), age, ethnicity and physical activity.

Pre-school children (2–5 years; n 1369), school-age children (6–11 years; n 1759), adolescents (12–18 years; n 3233) and adults (≥19 years; n 7850) in the USA.

There was a significant association between TV viewing and energy intake for adolescent girls (high v. low: β = 195·2, P = 0·03) and men (high v. low: β = −113·0, P = 0·02; middle v. low: β = −131·1, P = 0·0002). Mean adjusted energy intake for adolescent girls was 7801·0, 8088·5 and 8618·2 kJ/d for low, middle and high TV viewing, respectively. Mean adjusted energy intake for men was 9845·9, 9297·2 and 9372·8 kJ/d for low, middle and high TV viewing.

TV viewing was associated with energy intake in US children and adults only in 12–18-year-old girls and men. For girls, the high TV viewing category consumed more energy daily (816·3 kJ (195 kcal)) than the low category. In men, the middle and high TV viewing categories consumed less energy daily (548·4 kJ (131 kcal) and 473·0 kJ (113 kcal), respectively) than the low category. Our findings support some, but not all previous research. Future research is needed to explore this complicated relationship with rigorous measures of energy intake and TV viewing.

An urgent need in dietary assessment is the development of short tools that provide valid assessments of dietary quality for use in time-limited settings. The present study assessed concurrent and construct validity of the short Diet Quality Screener (sDQS) and brief Mediterranean Diet Screener (bMDSC) questionnaires.

Relative validity was measured by comparing three dietary quality indices – the Diet Quality Index (DQI), the modified Mediterranean Diet Score (mMDS) and the Antioxidant Score (ANTOX-S) – derived from the two questionnaires with those from multiple 24 h recalls over 12 months. Construct validity was demonstrated by correlations between average nutrient intake recorded on multiple 24 h recalls and the DQI, mMDS and ANTOX-S derived by the short screeners.

Both short questionnaires were administered to 102 participants recruited from a population-based survey in Spain.

DQI, mMDS and ANTOX-S correlated (P < 0·001) with the corresponding 24 h recall indices (r = 0·61, 0·40 and 0·45, respectively). Limits of agreement lay between 96 and 126 %, 59 and 144 % and 61 and 118 % for the DQI, ANTOX-S and mMDS, respectively. Dietary intakes of fibre, vitamin C, vitamin E, Mg and K reported on the 24 h recalls were positively associated (P < 0·04) with the DQI, mMDS and ANTOX-S indices.

The sDQS and bMDSC provide reasonable approximations to food-based dietary indices and accurately situate subjects within the indices constructed for the present validation study.

Assessing changes in dietary intake during the transition from adolescence to adulthood is challenging given the need for age-appropriate tools at different developmental stages. The present study investigated the comparability of intake estimates as assessed with the youth/adolescent and adult forms of Willett's FFQ.

Young adults were first asked to complete the adult FFQ as part of a larger study, Project EAT-III (Eating and Activity in Teens and Young Adults). A stratified random sample of respondents was invited to complete the youth/adolescent FFQ by mail within a 3-week period.

Participants were members of a longitudinal cohort who completed baseline surveys (including the adolescent FFQ) at schools in Minneapolis/St. Paul, Minnesota and completed Project EAT-III surveys online or by mail in 2008–2009.

There were ninety-one men and 103 women (median age = 24·6 years) who completed both forms of the FFQ.

The adolescent and adult forms did not provide comparable absolute intake estimates. However, with few exceptions, correlation coefficients between intake estimates were moderate (r = 0·4–0·6). Furthermore, the percentage of individuals classified into the same quartile rank category based on their responses to the adolescent and adult forms was ≥50 % for fibre, vitamins A and E, and servings of fruit (excluding juice), vegetables, dairy, whole grains and soft drinks.

Although responses on the adolescent and adult FFQ cannot be compared to describe changes in absolute intake over time, these tools provide comparable intake rankings and may be used together in longitudinal studies to investigate influences on diet.

To validate an eighty-nine-item semi-quantitative FFQ for measurement of nutrient intakes in elderly women.

FFQ and 3 d food records were filled in by women participating in the Kuopio Fracture Prevention Study (OSTPRE-FPS). Data on intakes of energy, fat, protein, carbohydrate, fibre, Ca, Fe, P, K, Mg, folic acid, vitamin B12, vitamin C, vitamin D and vitamin K from ninety-nine women were available to assess the agreement of the two methods. Validity was assessed using correlation coefficients, cross-classification into quintile categories and Bland–Altman plots. Nutrients relevant to bone health were assessed.

OSTPRE-FPS in Finland.

Elderly women with a mean age 71·3 years.

The FFQ overestimated energy and nutrient intakes as compared with food records by 30–50 %. The highest correlation coefficients of the energy-adjusted nutrient intakes between the methods were observed for fibre (0·60), Mg (0·56) and folic acid (0·49) and the lowest for protein and vitamin D (both 0·19). The cross-classification of energy-adjusted nutrient intakes showed that on average 68 % of the participants (range 62–78 %) were classified into the same or an adjacent quintile category.

The validity of energy and nutrient intakes measured with the FFQ was moderate as compared with 3 d food records in elderly women. The FFQ is a useful tool for the nutrient assessment of elderly women in epidemiological research.

Nutrition-related health problems such as obesity are frequent among children and adolescents of Turkish descent living in Germany, yet data on their dietary habits are scarce. One reason might be the lack of validated assessment tools for this target group. We therefore aimed to validate protein and K intakes from one 24 h recall against levels estimated from one 24 h urine sample in children and adolescents of Turkish descent living in Germany.

Cross-sectional analyses comprised estimation of mean differences, Pearson correlation coefficients, cross-classifications and Bland–Altman plots to assess the agreement between the nutritional intake estimated from a single 24 h recall and a single 24 h urine sample collected on the previous day.

Dortmund, Germany.

Data from forty-three study participants (aged 5–18 years; 26 % overweight) with a traditional Turkish background were included.

The 24 h recall significantly overestimated mean protein and K intake by 10·7 g/d (95 % CI of mean difference: 0·6, 20·7 g/d) and 344 mg/d (95 % CI 8, 680 mg/d), respectively. Correlations between intake estimates were r = 0·25 (P = 0·1) and 0·31 (P = 0·05). Both methods classified 70 % and 69 % of the participants into the same/adjacent quartile of protein and K intake and misclassified 7 % and 7 %, respectively, into the opposite quartile. Bland–Altman plots indicated a wide scattering of differences in both protein and K intake.

Among children and adolescents of traditional Turkish descent living in Germany, one 24 h recall may only be valid for categorizing subjects into high, medium or low consumers.

The relationship between food insecurity, maternal emotional distress and childhood morbidity in resource-poor settings is not well clarified. The present study aimed to assess independent associations between household food insecurity and childhood morbidity and potential modifications by maternal emotional distress.

A cross-sectional survey. A food security scale was used to assess household food insecurity; maternal reports were used to assess recent childhood illness; and the Hopkins Symptom Checklist was used to assess symptoms of emotional distress among mothers.

The Oromia Region, Ethiopia (rural area).

A total of 936 mother–child pairs.

Of 936 children assessed, 22·4 % had experienced diarrhoea, 20·7 % had cough and 21·5 % had fever in the 2 weeks preceding the interview. Household food insecurity was reported by 39 % of mothers. Greater food insecurity and greater maternal emotional distress were each independently associated with higher prevalence of cough and fever. Among mothers with low emotional distress, food insecurity was associated with a 2·3 times greater odds of diarrhoea in their children.

Household food insecurity may increase the risk of childhood illness in rural Ethiopia, and children having mothers with greater emotional distress may be at highest risk. These findings highlight the importance of strengthening policy initiatives aimed at reducing the high prevalence of food insecurity and emotional distress in Ethiopia.

To determine the extent to which weight gain and eating behaviours in infancy predict later adiposity.

Population-based, prospective, longitudinal birth cohort study. Weights collected in infancy were used to calculate Z-scores for weight gain to age 1 year conditional on birth weight (CWG). To avoid multiple significance tests, variables from the parent questionnaire completed at age 1 year describing eating avidity were combined using general linear modelling to create an infancy avidity score. Anthropometry, skinfold thicknesses and bioelectrical impedance data collected at age 7–8 years were combined using factor analysis, to create an adiposity index.

Gateshead, UK.

Members of the Gateshead Millennium Study cohort with data at both time points (n 561).

CWG in infancy significantly predicted adiposity at age 7 years, but related more strongly to length and lean mass. High adiposity (> 90th internal percentile) at age 7 years was significantly associated with high CWG (relative risk 2·76; 95 % CI 1·5, 5·1) in infancy, but less so with raised (> 74th internal percentile) eating avidity in infancy (relative risk 1·87; 95 % CI 0·9, 3·7). However, the majority of children with high weight gain (77·6 %) or avidity (85·5 %) in infancy did not go on to have high adiposity at age 7 years.

Rapid weight gain in infancy and the eating behaviours which relate to it do predict later adiposity, but are more strongly predictive of later stature and lean mass.

To evaluate the associations with chronic disease risk and mortality of the consequences of bean-free diets in Taiwanese adults with regard to gender.

A sub-sample of the National Health Interview Survey (NHIS) in 2001 agreed to physical examination in the subsequent year. This group then took part in the Taiwanese Survey of Hyperglycaemia, Hyperlipidaemia and Hypertension (TwSHHH) in 2002.

Individual records were linked to the eventual death files from 2002 to 2008.

Up to the end of 2008, a total of 2820 men and 2950 women were tracked by death registry over the 6·8 years of follow-up.

Among 38 077 person-years, an average follow-up 6·5 years, 225 all-cause deaths were identified. Generalized linear models showed beans to be favourable for metabolic syndrome (other than for fasting glucose) in men; in women, beans were favourable for waist circumference and HbA1c. Cumulative logistic regression models for the effect of a bean-free diet on metabolic syndrome scores according to the Taiwanese-modified National Cholesterol Education Program–Adult Treatment Panel III (NCEP-tw) gave adjusted odds ratios of 1·83 in men and 1·45 in women. Cox regression models for the bean-free diet showed an increased hazard ratio for all-cause mortality among women (1·98, 95 % CI 1·03, 3·81) but not men (1·28, 95 % CI 0·76, 2·16).

A bean-free diet may play a role in developing the metabolic syndrome in both genders, and is a significant predictor of all-cause mortality in Taiwanese women but not men.

To determine the association between major dietary patterns characterized by factor analysis and risk of depression and anxiety symptoms among adolescents.

Diet and symptoms of depression and anxiety were assessed in a cross-sectional survey among students attending junior high school. Dietary patterns were derived from a self-reported FFQ, which consisted of thirty-eight items. Anthropometric measurements were also performed.

Four junior high schools in Bengbu city, China.

A random sample of 5003 adolescents, 11–16 years of age (mean 13·21 years).

Three major dietary patterns were identified in the study based on factor analysis: ‘snack’, ‘animal food’ and ‘traditional’. The prevalence of depression symptoms, anxiety disorders and the coexistence of both were 11·2 % (560/5003), 14·6 % (732/5003) and 12·6 % (629/5003), respectively. After adjustment for potential confounders, adolescents in the highest tertile of snack dietary pattern scores had a higher odds for ‘pure’ psychological symptoms (‘depression without anxiety’, OR = 1·64; 95 % CI 1·30, 2·06; and ‘anxiety without depression’, OR = 1·87; 95 % CI 1·51, 2·31) compared with coexisting depression and anxiety (OR = 1·93; 95 % CI 1·54, 2·43). Similar to snacks, high consumption of animal foods was associated with a higher risk of psychological symptoms. Compared with low consumption, adolescents in the highest tertile of traditional dietary pattern scores had lower odds for ‘pure’ depression (OR = 0·38; 95 % CI 0·30, 0·49), ‘pure’ anxiety (OR = 0·85; 95 % CI 0·69, 1·04) and coexisting anxiety and depression (OR = 0·50; 95 % CI 0·39, 0·63).

Data from Chinese secondary-school adolescents validated findings from adult populations. Dietary patterns should be considered as important predictors of depression and anxiety among adolescents in further studies.

To assess prevailing food patterns among Balearic Islands’ adolescents, and socio-economic and lifestyle determinants.

Cross-sectional nutritional survey carried out (2007–2008) in the Balearic Islands, a Mediterranean region. Dietary assessment was based on a 145-item semi-quantitative FFQ and two non-consecutive 24 h recalls. Anthropometric measurements and questions related to socio-economic, lifestyle, physical activity and body image were assessed.

Data obtained from a representative sample of all inhabitants living in the Balearic Islands aged 12–17 years.

A random sample (n 1231) of the adolescent population (12–17 years old) was interviewed.

Factor analysis identified two major dietary food patterns: ‘Western’ and ‘Mediterranean’. The ‘Western’ dietary pattern was higher among boys than girls, associated with spending ≥4 h/d on media screen time, but less prevalent among those adolescents who desired a thinner body and those girls who desired to remain the same weight. The ‘Mediterranean’ dietary pattern was mainly followed by girls, and also boys who spent < 2 h/d on media screen time and girls with high parental socio-economic status.

The present study shows the existence of two major dietary patterns among Balearic Islands’ adolescents: ‘Western’ and ‘Mediterranean’, but girls are more ‘Mediterranean’ than boys. This evidence supports that the food pattern of Balearic Islands’ adolescents is in a transitional state characterised by the loss of the traditional Mediterranean dietary pattern towards a Western dietary pattern. Low parental socio-economic status, much leisure-time on sedentary behaviours such as media screen time and body image are factors associated with the ‘Western’ dietary pattern.

We examined the association of metabolic syndrome (MetS) and its components with dietary intakes of Mg in Tehran adults.

In a cross-sectional study, dietary intakes were assessed using a valid and reliable FFQ. MetS was defined according to the modified guidelines of the National Cholesterol Education Program Adult Treatment Panel III. Waist circumference (WC) was coded according to the newly introduced cut-off points for Iranian adults (≥95 cm for both genders).

Participants of the Tehran Lipid and Glucose Study (2006–2008).

Adults (n 2504; 1120 men and 1384 women) aged 18–74 years.

The mean age of participants was 40·8 (sd 14·6) years and 38·2 (sd 13·5) years for men and women, respectively. The reported mean intake of Mg was 349 (sd 109) mg/d. After adjustment for confounding factors, dietary Mg intake was inversely associated with fasting blood glucose (β = −0·08, P = 0·006), TAG (β = −0·058, P = 0·009) and WC (β = −0·013, P = 0·006); however, there were no associations between dietary Mg and diastolic blood pressure, systolic blood pressure or HDL cholesterol. An association was observed between MetS Z-score and Mg intake (crude β = −0·017, P = 0·001), independent of age, gender, smoking, physical activity and BMI; this association was attenuated following further adjustment for dietary factors and menopausal status (β = −0·034, P = 0·061).

Our findings suggest a significant inverse association between dietary Mg, MetS and its components.

To determine the effect of intention to breast-feed on short-term breast-feeding outcomes in women delivering term and preterm infants.

Data from the US Centers for Disease Control and Prevention's Pregnancy Risk Assessment Monitoring System (PRAMS) for three states, Ohio, Michigan and Arkansas, during 2000–2003 were analysed. SAS 9·1·3 and SUDAAN 10 statistical software packages were used for analyses.

Arkansas, Michigan and Ohio, USA.

Mothers of recently delivered infants, selected by birth certificate sampling.

Of 16 839 mothers included, 9·7 % delivered preterm. Some 52·2 % expressed definite intention to breast-feed, 16·8 % expressed tentative intention, 4·3 % were uncertain and 26·8 % had no intention to breast-feed. Overall 65·2 % initiated breast-feeding, 52·0 % breast-fed for ≥4 weeks and 30·8 % breast-fed for ≥10 weeks. Women with definite intention were more likely to initiate (OR = 24·3, 95 % CI 18·4, 32·1), to breast-feed for ≥4 weeks (OR = 7·12, 95 % CI 5·95, 8·51) and to breast-feed for ≥10 weeks (OR = 2·75, 95 % CI 2·20, 3·45) compared with women with tentative intention. Levels of intention did not differ between women delivering preterm and term. Women delivering at <34 weeks were more likely to initiate breast-feeding (OR = 2·24, 95 % CI 1·64, 3·06) and to breast-feed for ≥4 weeks (OR = 2·58, 95 % CI 1·96, 3·41), but less likely to breast-feed for ≥10 weeks (OR = 0·55, 95 % CI 0·44, 0·68), compared with those delivering at term. Women delivering between 34 and 36 weeks were less likely to breast-feed for ≥10 weeks than those delivering at term (OR = 0·63, 95 % CI 0·49, 0·81).

Prenatal intention to breast-feed is a powerful predictor of short-term breast-feeding outcomes in women delivering both at term and prematurely.

To review the health effects of solar radiation, sunbeds and vitamin D.

The literature was searched in the electronic database MEDLINE to indentify published data between 1981 and 2011. Studies were included if they reported relative risk for cutaneous malignant melanoma (CMM) associated with sunbed use, vitamin D and UV effects on human health.

Data from different time periods for populations at different latitudes.

Persons of different ages and ethnic groups.

UV from sun and sunbeds is the main vitamin D source. Young people with white or pigmented skin in northern Europe have a low vitamin D status. A number of health benefits from sufficient levels of vitamin D have been identified. However, UV exposure has been suspected of causing skin cancer, notably CMM, and authorities warn against it.

The overall health benefit of an improved vitamin D status may be more important than the possibly increased CMM risk resulting from carefully increasing UV exposure. Important scientific facts behind this judgement are given.

To assess serum retinol, liver intake patterns, breast-feeding history and anthropometric status in pre-school children of a low socio-economic community where liver is regularly consumed.

Cross-sectional study.

Northern Cape Province, South Africa.

Children aged 1–6 years (n 243) who attended the local primary health-care facility and had not received a vitamin A supplement in the 6 months preceding the study. Non-pregnant female caregivers (n 225), below 50 years of age, were also assessed.

Despite stunting, underweight and wasting being prevalent in 40·5 %, 23·1 % and 8·4 % of the children, only 5·8 % had serum retinol concentrations < 20 μg/dl, which is in sharp contrast to the national prevalence of 63·6 %. None of the caregivers were vitamin A deficient. Liver was eaten by 89·2 % of children, with 87 % of households eating liver at least once monthly and 30 % eating it at least once weekly; liver was introduced into the diet of the children at a median age of 18 months. Ninety-three per cent of the children were being breast-fed or had been breast-fed in the past; children were breast-fed to a median age of 18 months. A significant negative correlation was found between educational level of the caregiver and frequency of liver intake (r = −0·143, P=0·032). There was no correlation between serum retinol and indicators of anthropometric or socio-economic status.

The blanket approach in applying the national vitamin A supplementation programme may not be appropriate for all areas in the country, even though the community may be poor and undernourished.