Studies on the independent role of parity in long-term body weight change in economically developing countries are scarce and inconclusive, and only a few studies have taken into account patterns of breast-feeding. This association was examined in a national cross-sectional survey representative of Brazilian parous women.

The survey conducted in 1996 measured women's height and weight in the household and data on weight prior to the first pregnancy, parity and breast-feeding were recalled.

A sample of 2338 parous women, 15 to 49 years of age, 29 months after last delivery on average, had current body mass index (BMI, in kg m−2) modelled through hierarchical multiple linear regression analysis. Explanatory variables included parity, days of predominant breast-feeding, BMI pre-pregnancy, socio-economic, geographic, demographic and other reproductive variables.

Prevalences of overweight (BMI=25.0–29.9 kg m−2) and obesity (BMI<=30.0 kg m−2) were 25.2% and 9.3%. The overall mean weight gain per year after the first pregnancy was 0.90 kg for an average time since first pregnancy of eight years. BMI pre-pregnancy modified the association between current BMI and parity. Therefore, weight change attributed to parity calculated for a woman of average height (1.56 m) was 0.60 kg greater for primiparous women with a BMI pre-pregnancy of 30 kg m−2, compared with women with BMI pre-pregnancy of 25 kg m−2. This greater weight retention among obese women was 1.21 kg for women with two children and 1.82 kg for women with three or more children. Parity reduced the effect of weight loss associated with lactation (1.75 kg for six months of lactation among primiparous women and 0.87 kg among women with three or more children). For the sub-sample of 793 primiparous women, a weight decrease of 300 g was associated with each month of predominant breast-feeding for all prior BMI levels.

In this study, weight change associated to reproduction was highly dependent on BMI previous to pregnancy and the effects of parity and lactation were small.

To investigate and compare feeding practices among infants of less than 7 months of age in a rural and an urban area in Tanzania.

Cross-sectional, questionnaire-based interview of mothers and focus group discussions with extension workers and community leaders.

Eleven villages in a rural district and 10 wards in an urban district in the Morogoro region, Tanzania, west of Dar es Salaam.

Probability samples of mothers with infants of less than 7 months of age (n = 320 from each area).

Exclusive breast-feeding was rarely practised in either the rural or urban areas investigated. However, the urban mothers initiated breast-feeding earlier, discarded colostrum less frequently, breast-fed exclusively for a longer period, gave breast milk as the first feed more often and delayed the introduction of solid foods for longer than their rural counterparts. The rural mothers, on the other hand, breast-fed their previous infants slightly longer than the urban mothers.

The better performance of urban mothers could be partly due to sustained breast-feeding support in hospital settings and other campaigns which may not have reached the rural areas. In both the rural and urban areas more efforts are needed to encourage exclusive breast-feeding, to avoid premature complementation and, in the case of the urban areas, to protect extended breast-feeding.

At issue is whether weight loss in HIV infection is a cachectic process, characterised by loss of lean body mass with conservation of fat, or a process of starvation. We present data on body composition from 516 persons at different stages of HIV infection as determined by CD4 counts.

Cross-sectional analyses of body composition in relation to CD4 count.

The baseline data from a prospective cohort study of outcomes in HIV/AIDS in relation to nutritional status in Boston, Massachusetts, USA.

The first 516 subjects with HIV/AIDS to enrol in the study.

Differences in weight in relation to CD4 counts were present only at CD4 counts of 600 or less (slope below CD4=600: 1.9 kg per 100 CD4 cells, P<0.0001). On average, 68% of the difference in weight over CD4 counts was fat (slope: 1.3 kg fat per 100 CD4 cells, P<0.0001).

This cross-sectional analysis suggests that weight loss consists principally of fat loss in those persons with adequate fat stores. This observation will need to be confirmed in longitudinal analyses.

To investigate if high cassava production levels indicate high consumption and high dietary cyanide exposure in three villages situated within the area of Nigeria with higher cassava production than predicted by a geographic model for cassava production in Africa.

Exploratory assessment of: cassava production and processing by qualitative research methods and quantification of residual cyanogens in products; cassava consumption by food frequency and weighed food records and dietary cyanide exposure by urinary thiocyanate and linamarin.

Rural communities of Afuze, Ebue and Ofabo in mid-west Nigeria.

110 subjects from 42 households in three villages for food frequency interviews; 118 subjects in nine Ofabo households for weighed food records.

Cassava cultivation was reported to have increased in the preceding 20 years. It was consumed daily by 37 (88%) households, but its mean contribution to daily energy intake was only 13% (SD = 10). The range of residual cyanogens in cassava foods was 0 to 62 mg HCN equivalent/kg dry weight (dw). Ten samples (19%) had levels above the 10 mg HCN equivalent/kg dw FAO/WHO safety limit. Mean urinary thiocyanate and linamarin were 51 (SD = 35) and 20 (SD = 11) μmol/L, indicating low cyanogen intake and dietary cyanide exposure.

High cassava production levels did not result in high consumption and high dietary cyanide exposure levels, therefore cassava production levels cannot be used to predict consumption or cyanide exposure levels in the study area. A large part of the production is explained by intensive sales.

To evaluate the long-term effect on micronutrient status of a β-carotene-, iron- and iodine-fortified biscuit given to primary school children as school feeding.

Children receiving the fortified biscuit were followed in a longitudinal study for 2.5 years (n=108); in addition, cross-sectional data from three subsequent surveys conducted in the same school are reported.

A rural community in KwaZulu-Natal, South Africa.

Children aged 6–11 years attending the primary school where the biscuit was distributed.

There was a significant improvement in serum retinol, serum ferritin, haemoglobin, transferrin saturation and urinary iodine during the first 12 months of the biscuit intervention. However, when the school reopened after the summer holidays, all variables, except urinary iodine, returned to pre-intervention levels. Serum retinol increased again during the next 9 months, but was significantly lower in a subsequent cross-sectional survey carried out directly after the summer holidays; this pattern was repeated in two further cross-sectional surveys. Haemoglobin gradually deteriorated at each subsequent assessment, as did serum ferritin (apart from a slight increase at the 42-month assessment at the end of the school year).

This study has shown that fortification of a biscuit with β-carotene at a level of 50% of the Recommended Dietary Allowance (RDA) was enough to maintain serum retinol concentrations from day to day, but not enough to sustain levels during the long school holiday break. Other long-term solutions, such as local food production programmes combined with nutrition education, should also be examined. The choice of the iron compound used as fortificant in the biscuit needs further investigation.

To assess the dietary intake and the nutritional status of homeless men.

A night emergency shelter in Paris, France.

Dietary survey (48-h) including alcohol intake and a questionnaire on age, duration of homelessness, smoking habits. Subjects were also weighed and measured.

Ninety-seven men aged 18–72 years (mean 43.3), of whom 54% were homeless for more than 18 months, 82% were smokers and 53% were regular and/or excessive drinkers.

The BMI distribution was shifted towards low values, the percentage of wasted persons being four times higher than in the reference population. The mean total energy intake was 2376 kcal (SD = 879) and included a high and highly variable percentage of energy derived from alcohol (12.0% (SD = 18.1)). Among drinkers, the mean ethanol intake was 90 g (SD = 102) and there was a significant negative correlation between ethanol and non-alcoholic energy intakes. The median intakes of potassium, calcium, zinc, vitamins B1, B2, and niacin were lower than European Population Reference Intakes but only the mean intake of vitamin B1 was significantly lower. Eighty percent of non-alcoholic energy was provided by charitable organisations. For most nutrients, the nutritional density of the shelter ration was not significantly different from the density of the foods purchased by the homeless.

These data suggest that the content of some nutrients should be increased in existing food assistance programs for homeless people in France.

To study the ecological variation of intake of cassava foods and dietary cyanide load.

Ecological study design.

Five communities in south-western Nigeria where tropical ataxic neuropathy (TAN) was described as endemic (area A), 11 communities in south-western Nigeria where TAN was described as absent (area B), and five communities in northern Nigeria (area C).

Subjects were randomly sampled from selected communities. Intake of cassava foods was estimated from dietary history and dietary cyanide load was estimated from urine thiocyanate concentrations. Residual cyanogens in cassava food samples from the community markets were determined.

In total, 1272 subjects from 21 communities – 238 from area A, 659 from area B and 375 from area C – were selected. Intake of cassava food per person per week was 17 meals in area A, 10 meals in area B, and one meal in area C. Geometrical mean urine thiocyanate concentrations were 73 μmol l−1, 51 μmol l−1 and 17 μmol l−1 in areas A, B and C, respectively. Mean residual cyanogen content in cassava food samples was 16 mg HCN eq kg−1 (confidence interval (CI) 13–18) in area A, and 13 mg HCN eq kg−1 in area B (CI 11–14).

This study shows that the intake of cassava foods and dietary cyanide load is high in several communities in south-western Nigeria, predominantly in communities where TAN has been reported. Dietary cyanide load in these communities appears to be determined by the combination of frequency of intake and cyanogen content of cassava foods. Measures to improve the effectiveness of removal of cyanogen from cassava roots during processing are needed in the affected communities.