To study risk factors associated with overweight and obesity among adolescents enrolled in private high schools in the city of Pelotas, southern Brazil.

This was a case–control study. The subjects were 264 overweight (body mass index (BMI) ≥85th percentile of the National Center for Health Statistics (NCHS) reference population) and 264 non-overweight (BMI > 5th and <85th NCHS percentiles) adolescents identified by means of an anthropometric survey which included 1608 students. The adolescents were weighed, measured and interviewed about food habits and physical activity at school. Parents provided their own weights and heights when contacted by telephone, and parental BMI was calculated.

Hierarchical multiple conditional logistic regression showed that overweight and obesity were positively associated with mother's (odds ratio (OR) 2.86, 95% confidence interval (Cl) 1.43–5.74) and father's (OR 2.43, 95% Cl 1.37–4.30) BMI ≥ 30 kg m−2, being overweight before 10 years of age (OR 2.26, 95% Cl 1.30–3.90) and the habit of dieting (OR 3.53, 95% Cl 1.76–7.22). Having more than three meals per day was found to be a protective factor against overweight and obesity (OR 0.54, 95% Cl 0.29–1.00).

The present study showed that a family history of obesity, overweight during childhood and the habit of dieting for weight-loss purposes are factors associated with obesity during adolescence. The habit of having more than three daily meals turned out to be a protective factor against overweight. These results suggest the necessity for early intervention at the family and general community levels aimed at the prevention of obesity through actions directed towards the modification of established behaviours.

To determine the prevalences of overweight and obesity in children aged 4 to 12 years, using the standard definitions proposed by the International Obesity Taskforce.

Population prevalence cross-sectional survey involving measurement of height and weight. Data collection took place in 1998.

Schools in Gibraltar.

In total 2994 children, aged 4–12 years, attending these schools (1540 boys, 1454 girls).

Prevalence of overweight in boys was 19.7%, while obesity prevalence was 10.8%. For the girls, overweight prevalence was 21.4%, while obesity prevalence was 10.6%. There were no significant differences in the proportions of overweight or obesity between boys and girls.

This was the first time that the prevalences of overweight and obesity have been estimated in children from Gibraltar. These data provide further information on prevalence rates of overweight and obesity, using defined cut-offs for comparison with data from other countries.

This study aimed at assessing the effectiveness regarding implementation and impact of a take-home nutrition supplementation programme, the Protein Energy Malnutrition (PEM) Scheme, that targets malnourished pre-school children and pregnant and lactating women in the Northern Cape Province of South Africa.

In assessing implementation of the PEM Scheme, a cross-sectional descriptive study was undertaken over a 6-month period in the six regions of the Northern Cape Province. Interviews were conducted with programme managers and health personnel at clinics who were responsible for implementing the PEM Scheme. In assessing the impact of the PEM Scheme on growth, a retrospective review was done of the clinic records (including anthropometric data) of children enrolled in the PEM Scheme over a 1-year period.

About 76% of the budget allocated to the PEM Scheme had been utilised over the 1-year period. The budget for the following financial year was based solely on food supplements purchased in the previous year. Coverage of malnourished pre-school children and eligible pregnant and lactating women for enrolment was estimated to be 50% and 60%, respectively. Eighty-five per cent of health facilities in the province participated in the PEM Scheme. Some of the main problems identified included: lack of training, inappropriate targeting of certain groups, incorrect application especially of discharge criteria for pregnant and lactating women, inadequate assessment for nutrition-related disease, inadequate nutrition counselling and no standardised monitoring. Of the 319 children enrolled over a year, the mean age was 16.2 (standard deviation 16.2) months, 41% had been low-birth-weight and 18% had been diagnosed with tuberculosis. Ten per cent of the children with a weight-for-age Z-score of <−2 moved into the normal Z-score range after being on the PEM Scheme for a mean duration of 8 months. There was an overall improvement in the weight-for-age Z-scores of 25% of the sample, with a significant difference between the mean weight-for-age Z-scores at enrolment and follow-up (t = 4.8, P < 0.001). This was mainly related to significant improvement in the mean weight-for-age Z-scores of children <2 years.

Numerous problems with the PEM Scheme have been identified which could have limited its impact. Recommendations are proposed for improving the effectiveness and impact of the PEM Scheme in the province.

To determine the sociodemographic associations of obesity in Ghana.

A cross-sectional descriptive study was conducted on a sample of 6300 adults aged 25 years and over who were selected by random cluster sampling.

Two urban (high-class and low-class suburbs) and a rural community in Accra, Ghana.

In total, 4731 (1857 males, 2874 females) subjects participated. Demographic data were obtained by a questionnaire and height and weight were determined with subjects in light clothing and without shoes.

The overall crude prevalence of overweight and obesity was 23.4 and 14.1%, respectively. The rates of overweight (27.1 vs. 17.5%) and obesity (20.2 vs. 4.6%) were higher in females than males. Obesity increased with age up to 64 years. There were more overweight and obesity in the urban high-class residents compared with the low-class residents and in urban than rural subjects. Overweight and obesity were highest among the Akan and Ga tribes and relatively low among Ewes. Subjects with tertiary education had the highest prevalence of obesity (18.8%) compared with less literate and illiterate subjects (12.5–13.8%). Subjects whose jobs were of a sedentary nature had higher levels of obesity (15%) than subjects whose jobs involved heavy physical activity (10%). Subjects who did not engage in leisure-time physical activity were more obese than those who had three or more sessions of leisure-time physical activity per week (15.3 vs. 13.5%).

Overweight and obesity are common among residents in the Accra area. Older age, female gender, urban, high-class residence, sedentary occupation and tertiary education were associated with higher levels of obesity. Policies and programmes that promote healthy lifestyles may prove beneficial.

This study aims to analyse the differences, with regard to socio-cultural characteristics and dietary habits, between low and high consumers of yoghurt and other fermented dairy product desserts, and the nutritional significance of these differences.

We analysed the diet of a healthy population (4–65 years), using the 24-hour recall method, on three non-consecutive days. The participants were grouped by age and sex and were also divided into tertiles on the basis of yoghurt consumption. We compared energy and nutrient intakes, educational level and socio-economic status in the low consumption (LC) group and the high consumption (HC) group.

In general there were no significant differences in energy intake or nutritional profile between LC and HC groups. The only significant difference was in the percentage of energy provided by lipids, which was significantly lower in HC women, possibly due to the high number of women in this group who consumed low-fat yoghurt. There were significant differences in the distribution of HC and LC subjects according to the three educational levels (P < 0.05) but n according to socio-economic status.

The fact of being a high consumer of fermented dairy products took place in the framework of other dietary changes that compensated for this high consumption, resulting in the absence of significant differences in energy intake and nutritional profile between HC and LC subjects. The only exception was found in women who consumed low-fat dairy products. There was a relationship between high consumption of fermented dairy products and educational level.

The Bangladesh Integrated Nutrition Programme (BINP) experimented with two models of delivery: the first model uses the Government of Bangladesh's (GOB) own management structure and the second uses the non-government organisations (NGOs) working in the local community. This study compares the relative efficiency of GOB and NGO management in the provision of nutrition services.

A detailed costing survey was carried out to estimate the cost of delivering nutrition services from the Community Nutrition Centres (CNCs). The number of individuals enrolled, the number actually participating in the programme and person-days of service delivered were used as effectiveness measures.

Thirty-five CNCs were randomly selected from five BINP areas, of which 21 were in GOB-run areas and 14 in NGO-run areas.

The cost of providing nutrition services per enrolee was US$ 24.43 for GOB-run CNCs and US$ 29.78 for NGO-run CNCs.

Contrary to the widely held view, the analysis implies that the NGO facilities are not more efficient in the delivery of nutrition services when cost per person-days of service delivered is considered. The food cost component of BINP is so high that, irrespective of the delivery mode, policy makers should examine carefully the components of BINP in order to find the most cost-effective mix of services.

The primary objective of this study was to generate a reference table of food items and average amounts of these items consumed by South Africans, for the Department of Health. The reference table was required to be representative of foods and beverages eaten frequently by children and adults from all age and ethnic groups in order for the Department of Health to test for contaminants in these foods.

The National Food Consumption Survey (NFCS) served as a framework for compiling data on children since this was a national representative survey of 1–9-year-old children undertaken in South Africa in 1999. However, there has never been a national dietary survey on adults in South Africa. Consequently the data had to be extrapolated from existing isolated surveys on adults. Secondary data analysis was conducted on existing dietary databases (raw data) obtained from surveys undertaken on adults in South Africa between 1983 and 2000. Available datasets were regional and independent, and were not individually representative of the South African diet. It was therefore necessary to use different statistical methods, including factor analyses, weighting and correlations, to generate ethnic and geographic representative data for adults. Two methods were used: Method 1, which corresponded with results of the NFCS (over-sampled for low socio-economic status), and Method 2, which was based on ethnic proportions of the population.

The secondary data analyses generated food items most commonly consumed by the South African adult population (Method 1) in descending frequency of usage and average (mean) amount per day: maize porridge (78%/848 g), white sugar (77%/27 g), tea (68%/456 g), brown bread (55%/165 g), white bread (28%/163 g), non-dairy creamer (25%/6 g), brick margarine (21%/19 g), chicken meat (19%/111 g), full-cream milk (19%/204 g) and green leaves (17%/182 g). In 6–9-year-olds, maize porridge (72%/426 g), sugar (76%/23 g), tea (51%/258 g), full-cream milk (35%/171 g) and white bread (33%/119 g) were eaten most frequently. Similarly, in 1–5-year-olds, the foods consumed most frequently were maize porridge (80%/426 g), sugar (76%/21 g), tea (44%/224 g), full-cream milk (39%/186 g) and white bread (24%/83 g). In order to evaluate the validity of the adult data generated, kilojoule values of the individual food items (per capita) were compared with food balance sheets (FBSs). The comparison was favourable except that the FBSs had a higher overall energy intake per capita of between 22 and 28%.

Reference tables of commonly consumed foods and beverages were generated at minimal cost based on secondary data analyses of past dietary surveys in different South African populations.

To describe the living conditions of Senegalese adolescent girls according to their migration status, and to define the main socio-economic and biological determinants of their nutritional and growth status.

Health and living conditions, sexual maturation, and nutritional and growth status of adolescent girls were determined within the framework of a longitudinal study on growth.

The capital city of Senegal (Dakar) and a rural community (Niakhar), 120 km south-east of Dakar.

Three hundred and thirty-one girls, 14.5–16.6 years of age, were recruited from the same villages. Thirty-six per cent of the sample remained in the villages to attend school and/or to help with household subsistence tasks (non-migrants). The remaining (64%) migrated to cities to work as maids (migrants) and lived in two different socio-economic environments: at the home of a guardian during the night and in the house of the employer during the daytime.

Family rural environment and guardian and employer urban environments were socio-economically different (P < 0.001). Living conditions in urban areas were better than in rural areas and the employer's environment was socio-economically more favourable. Migrants had more advanced sexual maturation and higher body mass index (BMI), fat mass index (FMI) and mid-upper arm circumference than non-migrants. However, migrants were smaller than non-migrants. BMI, FMI and weight-for-age were related to socio-economic levels and duration of migration. Schooling was positively related to height and negatively related to BMI.

Migrants enjoyed better living conditions than non-migrants. This could be partly due to the better socio-economic environment of the employer. It was associated with earlier puberty and better nutritional status, but not with a better growth.

Objective: To estimate the role of human milk n-6 and n-3 polyunsaturated fatty acids (PUFA) in term infant growth in two African urban populations.

Observational study. Weight gains at 5 months of age and dietary habits were compared between Congolese infants (n=102) and Burkinabè infants (n=101). Socio-economic status and anthropometry of the mothers were also recorded.

One suburban district in Brazzaville (capital of The Congo) and one in Ouagadougou (capital of Burkina Faso).

Two random samples of nursing mothers and their 5-month-old infants.

All infants were born at term and there was no difference in birth weights. At 5 months of age, infants in Ouagadougou were thinner but not shorter than their counterparts in Brazzaville (average weight gain (standard deviation): 614 (168) g month-1 vs. 720 (176) g month-1; P>0.0001). Drastic differences were found in infant diets with regard to extra fluid intake and n-6 and n-3 PUFA concentrations in breast milk. In Ouagadougou, all infants were given fluids other than milk from birth. Breast milk had highly unbalanced 18:2n-6/18:3n-3 and n-6/n-3 long-chain PUFA ratios (53:1 and 5:1, respectively). In Brazzaville, half of the infants received fluids other than milk, and breast milk showed balanced 18:2n-6/18:3n-3 and n-6/n-3 long-chain PUFA ratios (12:1 and 1:1, respectively). A non-linear relationship between 18:2n-6/18:3n-3 ratio and growth was established in Brazzaville (P=0.0027). The 18:2n-6/18:3n-3 ratio adjusted with covariates had an even more significant effect on weight gain (P=0.0011). Applying the same model in Ouagadougou did not show such a relation.

Data strongly suggest that a balanced ratio of 18:2n-6/18:3n-3 (between 5:1 and 15:1) in breast milk leads to higher weight gain of infants during the first 5 months of life.

To investigate the relationship between different types and levels of physical activity and cardiovascular disease risk factors, including oxidative stress, blood lipids and insulin resistance, in a healthy female population in China.

Healthy women (n = 761) aged 35 to 65 years participated in this study. The habitual physical activity was evaluated by self-administered questionnaire (MOSPA). The dietary intakes of nutrients were calculated from 3-day recall records. Anthropometric data of each subject were measured, fasting blood samples were taken, and erythrocytes and serum were prepared for the measurement of erythrocyte superoxide dismutase activity, serum malondialdehyde, total antioxidant capacity, insulin, glucose and lipids (total cholesterol, triglycerides, apolipoprotein AI (apo A) and apolipoprotein B (apo B)) concentrations.

Low level of physical activity was related to a lower concentration of serum apo B, and higher energy expenditure from household physical activity had a reverse relationship with serum apo B and triglyceride levels. In the group with moderate occupational energy expenditure, the concentration of serum triglycerides was lower, but that of high-density lipoprotein was higher. Moderate energy expenditure (less than 1700 kcal day−1) from leisure-time physical activity was positively related to total antioxidant capacity and insulin sensitivity. However, heavy occupational physical activity may be not beneficial for the cardiovascular system.

This study indicates that leisure-time, moderate occupational and household physical activity levels decreased risk factors for cardiovascular disease.

To investigate underlying causes for food and nutrition insecurity in black South African households and to gain understanding of the factors contributing to better nutrition security, with emphasis on household organisation, gender and intra-household dynamics and social networks.

Within a larger cross-sectional survey that investigated the impact of urbanisation on the health of black South Africans, 166 people, mostly women, were interviewed on household food security. Methods used were structured face-to-face interviews, in-depth interviews, observation, interviews with key informants and a sociodemographic questionnaire. Information was collected from 1998 to 2000 in 15 rural and urban areas of the North West Province, South Africa.

Three-quarters of households in this sample are chronically food-insecure. Families are disrupted, due to migrant work, poverty and increasing societal violence, and half of households are female-headed. Certain categories of female-headed households and households based on partnership relationships, despite more limited resources, achieve a better or an equal economic status and better nutrition security than those households led by men, with the latter often being considered an economic liability. The reliance on and fostering of social ties and networks appear to be of central significance.

Gender and intra-household relations, as well as social networks and income from informal sector activities, are often not uncovered by conventional statistical methods. Qualitative research can reveal the unexpected and furthermore empowers people, as their voices are heard.

There is strong evidence that a high consumption of fruit and vegetables reduces the risk of developing many cancers. This study examined the economic consequences for the healthcare sector if people followed the recommendations and increased their intake of fruit and vegetables.

A life table was used to describe a base case population with respect to life expectancy, cancer incidence and healthcare costs. Relative risks of cancer for a high versus a low intake of fruit and vegetables were obtained from the literature and were used to simulate populations with a higher intake of fruit and vegetables. The empirical data consist of a 20% sample of the Danish population that was followed from 1993 to 1997. Civil registration numbers were used to link various computerised registers, in order to describe each individual in the sample in terms of morbidity, mortality and healthcare costs.

The average daily intake of fruit and vegetables was assumed to be approximately 250g for the general Danish population. Simulated intakes of 400g and 500g increased the life expectancy by 0.8 and 1.3 years, respectively. In addition, it was estimated that 19% and 32% of the cancer incidence could be prevented. The aggregate healthcare costs remained stable, as the resources saved due to a lower cancer incidence were offset by healthcare costs imposed by the fact that healthy people live longer and require more healthcare. However, the variations across age groups and health sectors were substantial.

The study adopted a healthcare sector perspective. Only costs from hospitalisation and primary care were included in the calculations. The costs of changing people's dietary habits, i.e. education, information and promotion as well as other costs that would be relevant from a societal perspective, have not been taken into account. Furthermore, the transition from one level of intake to another is not the focus of the analysis, although it might take decades to observe the full effect of the dietary changes.

Empirical evidence suggests that a considerable fraction of all cancer incidences can be prevented by a higher intake of fruit and vegetables. That may result in improved public health (gain in life years) at no additional cost to the healthcare sector.

To identify the factors associated with childhood undernutrition.

Cross-sectional survey.

Hlabisa health district in KwaZulu/Natal, South Africa.

Eight hundred and sixty-eight children aged 3–59 months.

Questionnaire survey and anthropometric survey of 516 random house-holds with children in the health district. Multivariate analysis took into account the hierarchical relationships between the proposed risk factors. This conceptual model was built up during qualitative studies and with reference to international research in this area.

The mean Z-scores for weight-for-age and height-for-age were −0.52 (95% confidence interval (CI)−0.44 to −0.60) and −1.25 (95% CI −1.15 to −1.35), respectively. Of the children, 26.3% (95% CI 23.3–29.3%) were stunted, 12.0% (95% CI 9.8–4.2%) were underweight-for-age (UWFA) and only 1.3% were wasted. Migrant father, mother's education, literate mother, whether the house was made of traditional materials, the presence of a toilet, whether the last child was breast-fed, duration of breast-feeding and birth weight were found to be significantly related to UWFA, the latter with an odds ratio of more than eight.

While there are individual health and feeding behaviours that need to be addressed, this study has also uncovered the maternal and socio-economic factors which present an array of constraints in the social environment shaping the infant feeding and caring practices employed by women. For the full potential of nutrition promotion activities to be realised, there is a need to address some of these ‘deeper’ constraints.

To study the dynamics of childhood overweight and the influence of dietary intake on tracking of overweight.

A follow-up study conducted in China.

Ninety-five overweight children, 6–13 years old, identified from 1455 children at baseline, were followed over a 2-year period.

Data on anthropometry and 3-day dietary intake were collected at baseline and during follow-up. Overweight was defined using the International Obesity Task Force reference of body mass index (BMI)-for-age. Differences between groups were tested using analysis of variance and Cochran-Mantel-Haenszel tests.

Of the 95 overweight children, 36.8% remained overweight 2 years later (‘tracking group’). Urban boys were the three times more likely than rural boys to remain overweight (63.2% vs. 21.9%). At baseline, the tracking group had higher BMI, body weight and fat intake (% of energy), and lower carbohydrate intake (% of energy), than the non-tracking group (who shifted from overweight to not overweight); they were more likely to have a high-fat or high-meat diet, but less likely to have a diet high in carbohydrate or vegetables and fruit. During the follow-up, the tracking group increased fat intake and reduced carbohydrate intake while the non-tracking group did not; and they also grew slower in height but faster in weight. Tracking of overweight seemed to be related to tracking of high-meat (relative risk (RR) 2.4, 95% confidence interval (CI) 1.0–5.6, P < 0.05) and high-fat (RR 1.5, 95% CI 0.9–2.5, P < 0.1) diets.

Considerable changes in children's overweight status during childhood and adolescence were observed in China, a transitional society. Dietary patterns, particularly dietary composition, seemed to influence the tracking patterns of overweight.

To examine the association between consumption of red and processed meat (RPM) and iron intakes and status in adults.

Further analysis of the Dietary and Nutritional Survey of British Adults, a cross-sectional study of 2197 adults aged 16–64 years carried out in 1986/7.

Adults (836 men and 838 women) with serum ferritin measurements, who were not taking iron supplements, were classified into four groups according to RPM consumption (from 7-day weighed records). Iron absorbed was estimated from equations based on haem and non-haem iron and the influence of iron stores.

Women who ate least meat (<90gday-1) had three times the risk of a low iron intake (below the Lower Reference Nutrient Intake) compared with high consumers of RPM (>140gday-1). Men who ate no RPM also had a higher risk of low iron intake. Using an estimate of minimal values for iron losses, there was a twofold difference in the potential risk of negative iron balance between women non-RPM consumers and high RPM consumers. Status measurements indicated that, among women, anaemia was least prevalent (6%) among high consumers compared with 12–14% among average RPM consumers. Inverse trends were also observed for serum ferritin in both sexes.

Low consumption of RPM has implications for iron intakes and iron status in men and women, since the risk of negative iron balance and its consequences are increased. Dietary messages must consider these implications and provide appropriate advice.

To examine the validity and reproducibility of a self-administered food-frequency questionnaire (FFQ) used for two cohort studies in Japan.

Cross-sectional study.

Two rural towns in the Miyagi Prefecture, in north-eastern Japan.

Fifty-five men and 58 women.

A 40-item FFQ was administered twice, 1 year apart. In the mean time, four 3-day diet records (DRs) were collected in four seasons within the year. We calculated daily consumption of total energy and 15 nutrients, 40 food items and nine food groups from the FFQs and the DRs. We computed Spearman correlation coefficients between the FFQs and the DRs. With adjustment for age, total energy and deattenuation for measurement error with the DRs, the correlation coefficients for nutrient intakes ranged from 0.25 to 0.58 in men and from 0.30 to 0.69 in women, with median of 0.43 and 0.43, respectively. Median (range) of the correlation coefficients was 0.35 (−0.30 to 0.72) in men and 0.34 (−0.06 to 0.75) in women for food items and 0.60 (−0.10 to 0.76) and 0.51 (0.28–0.70) for food groups, respectively. Median (range) of the correlation coefficients for the two FFQs administered 1 year apart was 0.49 (0.31–0.71) in men and 0.50 (0.40–0.64) in women for nutrients, 0.43 (0.14–0.76) and 0.45 (0.06–0.74) respectively for food items, and 0.50 (0.30–0.70) and 0.57 (0.39–0.66) respectively for food groups. Relatively higher agreement percentages for intakes of nutrients and food groups with high validity were obtained together with lower complete disagreement percentages.

The FFQ has a high reproducibility and a reasonably good validity, and is useful in assessing the usual intakes of nutrients, foods and food groups among a rural Japanese population.

The aim of this study was to determine the iron status, and the risk factors for iron deficiency (ID) and iron-deficiency anaemia (IDA), of non-pregnant adult women working in a fruit-packing factory.

A cross-sectional analytical study was done on 338 women, 18 to 55 years of age. Information on demographic data, risk factors for ID, smoking, and the consumption of red meat, chicken and fish was collected by questionnaire. Height and weight were measured and the body mass index (BMI) calculated. A non-fasting venous blood sample was analysed for haemoglobin (Hb), serum ferritin (SF), serum iron, serum transferrin and C-reactive protein; transferrin saturation (TFS) was calculated.

Fruit-packing factory in the Western Cape, South Africa.

The mean value for Hb was 13.06 (standard deviation (SD) 1.16) g dl−1 and for SF 48.0 (SD 47.8) μgl−1 (geometric mean 26.44 μgl−1). Women (n = 325) were categorised on the basis of iron status: 60% had a normal iron status (NIS); 12.6% had low TFS (<16%) but normal Hb (≥12 g dl−1) and SF (≥12 μgl−1) concentrations (LTS); and 27.4% had low iron status (LIS), defined as combinations of low SF (<12 μgl−1 or <20 μgl−1), low TFS (<16%) and low Hb (<12 gdl−1). More than 30% of the women were obese (BMI ≥ 30 kgm−2). The risk ratio for LIS (LIS vs. NIS) was 3.8 (95% confidence interval (CI) 1.9–7.6) if women were still menstruating or 3.2 (95% CI 1.6–6.2) if they were pregnant during the past 12 months. Women with LIS consumed significantly smaller portions of red meat, chicken and fish than did women in the other two groups.

IDA (low Hb, SF and TFS) and ID (low SF and TFS) did not seem to be a major problem. Women who were still menstruating or were pregnant during the past 12 months were at greater risk for ID. The consumption of smaller portions of red meat, chicken and fish was related to LIS. A high prevalence of obesity, which demonstrated the coexistence of both under- and overnutrition, was observed.

To assess the effect of the incorporation of amylase in maize-based flours prepared as gruels on the energy intake and growth of Congolese infants.

A randomised controlled trial. At 18 weeks of age, infants were randomised into either an intervention group, where they were provided with a maize/soya-based flour that contained amylase, or a control group, where they were provided with a similar flour that did not contain amylase.

Urban borough of Poto-Poto in Brazzaville, The Congo.

Eighty infants (40 in each group) were randomised into intervention and control groups. Three infants in the intervention group and two controls subsequently dropped out.

At 24 weeks, the addition of amylase resulted in a significant increase in energy intake (in kJ kg-1 day-1) from gruels (P=0.02) without affecting breast milk consumption. In contrast, total energy intake (in kJ kg-1 day-1) did not differ significantly between groups (P=0.08). After adjustment for morbidity and previous growth, infants in the intervention group showed better growth in length during the trial (+0.22 cm month-1; P=0.04), especially between 24 and 31 weeks of age (+0.51 cm month-1; P>0.01). There were no differences in weight velocity between groups.

The findings of this study suggest that the consumption of amylase-treated gruels allows an increase in energy intake from these gruels without affecting breast milk consumption but has no impact on total daily energy intake. However, if started after 6 months, it could be effective in preventing faltering of infant linear growth.