Original Article | Published:

Maternal Nutrition, Infants and Children

Contribution of inappropriate complementary foods to the salt intake of 8-month-old infants

European Journal of Clinical Nutrition volume 66, pages 104110 (2012) | Download Citation

Subjects

Abstract

Background/Objective:

Complementary feeding is a critical developmental time for establishing variety, and appropriate feeding practices are important. These include recommendations for a maximum intake of sodium (salt), because of the potential harmful effect of excessive intake on the developing kidneys and blood pressure in later life. Sodium intakes were examined to establish, if inappropriate complementary feeding practices were contributing to high intakes.

Subjects/Methods:

Data were collected from a cohort of 1178 8-month-old infants born in 1991/92, participating in the Avon Longitudinal Study of Parents and Children, using a 3-day dietary records completed by their mothers. Sodium and energy intakes were calculated and infants were categorised into quartiles of sodium intake.

Results:

Majority of infants were first introduced to solids around 3–4 months, with plain baby rice, rusks and other cereals being the first foods consumed during complementary feeding. In the whole sample, 70% consumed in excess, of 400 mg sodium per day, the maximum UK recommendation for children up to age 12 months. The mean sodium intake for the highest quartile was 1060 mg per day. Infants in this quartile were often consuming cows’ milk as a main drink; eating three times the amount of bread compared with the lowest quartile and used salty flavourings such as yeast extract and gravy.

Conclusions:

Sodium intakes in this cohort of infants were higher than recommendations. Clear and practical education is needed on complementary feeding practice for mothers highlighting what foods to introduce and when. Manufacturers have a responsibility to reduce the sodium content of food products.

Introduction

Substantial evidence highlights that high blood pressure (HBP) is responsible for a considerable number of deaths worldwide (World Health Organisation, 2004), and the link between high sodium intakes and raised BP is well documented (Gilbert and Heiser, 2005). HBP established during childhood can track into adulthood (Lawlor and Davey-Smith, 2005), therefore the lowering of sodium intakes in childhood, could help reduce later HBP. Sodium requirements for healthy infants are low and the maximum intake recommended in the UK is 400 mg per day (less than 1 g of salt) up to age 12 months (Scientific Advisory Committee on Nutrition, 2003). World Health Organisation (WHO) does not provide any specific salt recommendations for children and adolescents (World Health Organisation, 2011) and, the European Society for Paediatric, Gastroenterology, Herpetology and Nutrition (ESPGHAN) committee also do not give any specific recommendations, they state as an overall guideline that additional salt should not be added to foods during infancy (Agostoni et al., 2008). In the short term, too much sodium in the diet can be harmful to the developing kidneys and in the long term infants may acquire a taste for salty foods, which could result in later health problems (Derbyshire and Davies, 2007). Yet, intakes are often higher than the recommendations and levels increase rapidly after the introduction of solids (Derbyshire and Davies, 2007).

Initially, sodium is obtained from milk, either breast or formula, but after complementary feeding (or ‘weaning’), sodium is obtained from a wide range of foods, many of which have a relatively high sodium content such as bread, some breakfast cereals and cheese. Ready-prepared infant foods are, on average, lower in sodium than home-prepared foods (Derbyshire and Davies, 2007). Current WHO advice recommends that breast-fed infants should not be introduced to solid foods before 6 months (World Health Organisation, 2001). Somewhat controversially (Allcutt and Sweeney, 2010), the UK Department of Health has adopted these recommendations for all infants with the additional guidance that no infant should have solids before 17 weeks (Allcutt and Sweeney, 2010; More et al., 2010).

In the UK, the majority of sodium consumed by individuals is added to food during manufacturing, with a relatively small proportion added during cooking or at the table (Cappuccio, 2009). Current intakes in both children and adults are far higher than recommended (National Institute of Clinical Excellence, 2010). There has been resistance from the food industry regarding reformulation; but in recent years work has begun to reduce levels in everyday foods (Cappuccio, 2007; National Institute of Clinical Excellence, 2010). The UK Food Standards Agency has led the way with this work and has published revised salt targets for 2012 for 80 categories of foods (Cappuccio and Capewell, 2010). Additionally, the National Institute of Clinical Excellence (NICE) have produced guidance on the prevention of cardiovascular disease, which seeks to ensure children's intakes do not exceed the age-appropriate recommendations (National Institute of Clinical Excellence, 2010). Supported by the WHO, the UK is leading the development of the European Salt Action Network, which aims to share information and experiences in an attempt to reduce sodium intakes (World Health Organisation Website, 2011).

It is extremely important that mothers receive clear and precise information regarding what foods are appropriate during complementary feeding and when these should be introduced. Many parents may be unaware of the contribution manufactured foods make to sodium intakes. Mothers of low socio-economic status tend to be less well informed about the suitable timings and foods to use, resulting in poorer feeding practices (Savage et al., 1998), and poor eating habits established during childhood may track into adulthood (Fein et al., 2008).

The aim of this study was to examine sodium intakes in the diets of 8-month-old infants, from a well-characterised birth cohort, to establish if inappropriate complementary feeding practices were resulting in high sodium intakes.

Subjects and methods

Subjects

Subjects were participating in the Avon Longitudinal Study of Parents and Children (ALSPAC), an ongoing UK longitudinal cohort study designed to investigate the health and development of children. The study design has been described fully elsewhere (ALSPAC Website; Golding et al., 2001). Briefly, pregnant women were eligible if they had an expected delivery date between April 1991 and December 1992 and were resident in the former Avon Health Authority in South West England. A cohort of 14 541 pregnancies, resulting in 13 988 children alive at 12 months, was established and a 10% random sample of mothers from the final 6 months of recruitment were invited to bring their infants to a research clinic when aged 8 months (n=1453). Ethical approval for the study was obtained from the ALSPAC Law and Ethics Committee and the Local Research Ethic Committees.

Dietary assessment

Dietary data were collected at the 8-month clinic visit in 1993 from the infant's main caregiver. The full method has been described elsewhere (Noble et al., 2001). Briefly, before the clinic visit, caregivers were asked to record in a structured diary (using household measures) all food and drink consumed by the infant over 3 days; 2 weekdays and 1 weekend day (self-selected and not necessarily consecutive), with a description of any leftovers. Mothers of breast-fed infants were asked to record the duration of each feed. At the clinic, caregivers were interviewed by a nutrition fieldworker to clarify any anomalies in the diet diary.

The completed dietary records were transformed into weights and code numbers corresponding to each food and drink consumed. Portion sizes for baby foods were usually described by proportion of a jar eaten or by spoonfuls, thus making weight relatively simple to estimate. Family foods were allocated weights based on food portion sizes (Ministry of Agriculture, Fisheries and Food, 1993) and scaled down appropriately. For breast milk, the duration of each feed was used to estimate the likely volume of milk. A feed lasting 10 min or longer was assumed to be 100 ml in volume and a proportion of this if the feed was of shorter duration (that is, 10 ml per minute). This method has been discussed further in previous analyses of ALSPAC data (Emmett et al., 2000).

Diaries were coded by the nutrition fieldworker using the computer package Microdiet (University of Salford). The nutrient and food group information associated with each food code was obtained from McCance and Widdowson's The Composition of Foods, 5th edition (Holland et al., 1991) and supplements (Holland et al., 1988; Holland et al., 1989) with additional data obtained from the food manufacturers. Contemporary analysis of ready-prepared infant foods was available. All of the records were checked against the original records and any errors identified were corrected. Diaries with very high or very low estimates for selected nutrients were re-checked. An in-house analysis programme was used to generate nutrients for each food the child ate, with average energy and nutrient intakes, weights and sodium content of each food group calculated. The contribution to sodium intake and the amount of sodium per unit energy (MJ) from each food group was calculated.

Anthropometric measures, maternal education and age of introduction

At the 8-month clinic, the infants’ weight was assessed using the Soenhle scale or Seca 724 scales, and length (crown-heel) was measured using Kiddimetre (Raven Equipment Ltd, Essex, UK). Body mass index was calculated using the standard equation weight(kg)/height(m)2.

Highest maternal education level was derived from a questionnaire at 32 weeks of pregnancy. Answers were grouped according to UK standards into: none, CSE (certificate of secondary education; national school exams at 16 years), vocational, O level (national school exams at 16 years, higher than CSE), A-level (national school exams at 18 years) or degree. In this analysis answers were grouped into low (none, CSE or vocational), medium (O level) or high (A-level or degree).

Age of introduction to solid foods was obtained by questionnaire when infants were aged 6 months. The questionnaire contained a list of foods, and mothers were asked to complete the age in months that the food was introduced, and the number of times per week it was currently being consumed.

Statistical methods

Basic descriptive statistics (mean and s.d.) were conducted to establish sodium intakes in the infants’ diet. The infants were categorised into similar sized quartiles depending on the sodium level in their diet. Their energy intake adjusted for body weight was calculated and compared within each quartile by sex to dietary reference values for this age group to ascertain whether these intakes are likely to be plausible (Department of Health, 1991). The Wilcoxon two-sample test was used to compare the mean energy intakes and energy intake per kg body weight in boys and girls separately. A χ2-test was conducted to establish if maternal education status was related to intake of sodium by quartile of intake. Infants’ were also categorised according to the type of milk they consumed as their main drink: breast, formula or cows’ milk. To gain an understanding of the quality of the diets being fed to infants in each quartile of sodium intake, energy intake was held constant by dividing the weight or sodium content of each food by total energy intake (in MJ). All statistical analyses were carried out using SPSS for Windows, version 18 (SPSS Inc., Chicago, IL, USA).

Results

Sample characteristics

The sample consisted of 1178 infants aged 8 months (86% of those invited), and included 644 boys and 534 girls, whose caregiver had completed dietary records. Table 1 presents basic descriptors for the cohort by quartile of sodium intake; these are energy intakes, sodium intakes, maternal educational status and age of introduction to solids. In the whole sample, 70% consumed in excess, of 400 mg per day (data not shown), the maximum recommended for children up to age 12 months (Scientific Advisory Committee on Nutrition, 2003). There were more boys than girls in the top quartile and infants were heavier and longer than those in the bottom quartile. There was no difference in maternal education level between the sodium quartiles.

Table 1: Anthropometric measures, energy and sodium intakes (mean and s.d.) and maternal education and early introduction of solid foods (frequency) in 8-month-old infants for the four quartile groups by sodium intake

Introduction of solids

The majority of infants were first introduced between 3–4 months (n=925), with fewer infants introduced either earlier (n=180) or later (n=32). The UK Department of Health's guidance at the time (1993) stated complementary feeding should occur between 3 and 6 months (Department of Health and Social Security, 1988; Department of Health, 1994). Plain baby rice, flavoured baby rice, plain rusks and other cereals were the first foods consumed during complementary feeding. Infants in the top sodium quartile were much more likely to be introduced to solids early, that is before 3 months.

Energy intakes

Energy intake (mean) in the whole sample was 3.42 MJ, which is close to the estimated average requirements for energy of 3.44 MJ for boys and 3.20 MJ for girls. The bottom quartile had a substantially lower mean intake (2.86 MJ) than the top quartile (4.01 MJ) (Table 1). Energy intakes per kg body weight showed that infants in the bottom sodium group had lower intakes than those in the top group (more than 100 kJ per kg difference). This suggests that many parents of infants in the bottom quartile had recorded implausibly low intakes, thus less food was recorded than actually eaten, and therefore it is likely that this group had higher sodium intakes than have been reported. Conversely, the parents of the top sodium group may have over-reported their child's food intake thus their sodium intake could be artificially high; however, if reduced by a third these intakes would still exceed recommended maximum levels.

Milk consumption

Most infants were receiving either breast or formula milk; however, 13% overall received only cows’ milk (Table 2). Cows’ milk has higher sodium content (55 mg/100 g) than formula (15–30 mg/100 g, depending on type) and breast milk (15 mg/100 g). Infants in the top quartile were much more likely to be taking only cows’ milk as can be seen in Tables 2 and 3. In fact cows’ milk made a substantial contribution to sodium intake in this group. The bottom sodium group was much more in line with recommendations on milk feeding.

Table 2: Types of milk (frequency and percentage) fed to 8-month-old infants in the four-quartile groups of sodium intake
Table 3: Energy-adjusted weight and sodium content (mean (s.d.)) and the percentage sodium contribution of food groups in the four quartiles of sodium intake in 8-month-old infants

Specific foods

Table 3 illustrates energy-adjusted weight and sodium intakes from foods and their percentage contribution to sodium. Approximately 20% of sodium was obtained from ready-prepared infant foods in the whole sample. At the time of the data collection (1993), the average sodium content of baby foods was 71 mg/100 g for dry baby cereal (made-up), 72 mg/100 g for savoury jars and 26 mg/100 g for sweet jars. These levels have been reduced over the intervening years and averages for one baby-food manufacturer in 2010 were 41 mg/100 g, 51 mg/100 g and 15 mg/100 g, respectively (In Practice - Cow and Gate Practical Support for Healthcare Professionals Website, 2011), resulting in a reduction of approximately a third. If infants were eating in a similar way today, the amount of sodium from ready-prepared infant foods would total around 58 mg, a reduction of 35 mg overall. Infants in the top quartile were consuming less than half the amount of ready-prepared infant foods (6% of overall sodium intake) than those in the bottom quartile (34%).

The quartile groups consumed progressively different foods during complementary feeding (Table 3). In the bottom quartile, the majority of sodium was obtained from either milk sources or ready-prepared infant foods. Other food sources contributing to their sodium content were bread and breakfast cereals. Food contributors in the top quartile were primarily family foods; bread, meat, salty flavouring (including marmite) and pasta/rice (containing plain and tinned varieties).

The highest family food contributor to sodium in all quartiles was bread (180 mg/slice). Infants in the top quartile had higher intakes of both bread and breakfast cereals, the most popular of the latter products were weetabix (54 mg/biscuit) and ready brek (2 mg/portion), with a few infants consuming corn flakes/rice krispies (222 mg/portion). Other sources of sodium in this top quartile group were gravy granules (153 mg from one tablespoon of made up gravy) and cheese (67 mg for 10 g of cheddar). Gravy intakes were three times higher in the top quartile compared with the bottom. Intakes of salt flavourings, with contributions predominately from marmite (45 mg from 1 g), were particularly high among the top quartile. The pasta/rice group provided the top quartile with 17 mg sodium per MJ compared with 0.5 mg in the bottom group and this was primarily obtained from canned spaghetti (420 mg/100 g). Work to reduce sodium levels in manufactured foods has continued in recent years, hence it is likely that sodium levels in infant diets will be slightly lower than this currently.

Discussion

This study found strong evidence that poor complementary feeding practices in 8-month-old infants were associated with higher sodium intakes from the diet. Infants classified in the top quartile were consuming inappropriate foods such as cows’ milk as a main drink, salty flavourings and canned pasta during complementary feeding, which contributed to the very high sodium levels in their diet. Sodium intakes exceeded the UK recommendation in around 70% of the infants.

Accounting for body weight, energy intakes in the bottom quartile of sodium intake, suggested under-reporting; sodium intakes in this group are likely to have been higher. The data infer it is difficult to keep an infants’ intake of sodium below the UK recommendation of 400 mg per day. There is a need for appropriate feeding advice regarding the adaptation of family diet for infants to take into account the sodium content of foods and for the manufacturers to reduce further the sodium content of common foods, particularly bread and breakfast cereals.

The inappropriate use of cows’ milk in this age group was contributing to the high sodium intake. ALSPAC data has previously shown that infants who were consuming cows’ milk with no other milk drinks had higher levels of anaemia (Hopkins et al., 2007). Although these data were collected 1993, figures from the latest UK Infant Feeding Survey obtained in 2005 showed consumption of cows’ milk in this age group is still problematic, with 6% (equivalent to 42 000 infants per year) having cows’ milk as a main drink (Bolling et al., 2007), even though this is not recommended before 1 year of age (Agostoni et al., 2008).

During infancy, adding salt to food is inappropriate and gradually introducing infants of 8 months to family foods is recommended. The use of salty flavourings (predominately yeast extract), sauces (including gravy), canned pasta and baked beans were quite common in the top sodium quartile, suggesting that some family food may be inappropriate for infants. The introduction of these items suggests that either parents’ had a lack of information regarding the best foods to use during complementary feeding or were overwhelmed/confused by the information that was available. Whatever the explanation, it suggests that clearer information is needed.

Previous studies have shown that poor feeding practices are associated with low socio-economic status and/or educational attainment (Savage et al., 1998; Hendricks et al., 2006). An American study analysing data from the Infant Feeding Practices Study II found mothers of lower educational status were more likely to adopt poor feeding practices, this included half of these mothers adding salt to their infants’ food (Fein et al., 2008). The current study did not replicate these findings, as no difference was found between the sodium groups in maternal educational status.

Despite initial resistance, the food industry in the UK and other European Countries, have begun to make substantial efforts in recent years to reduce sodium levels in numerous food products. As a result of these changes, average adult salt intakes have been reduced from 9.5 g–8.6 g per day in the UK. Other countries, particularly Portugal and Finland, have made significant improvements reducing levels to 5 g a day through labelling, public education and close collaboration with the food industry (Cappuccio and Capewell, 2010). In this study, breakfast cereals and bread were providing considerable amounts of sodium to the diet. Recently, it has been announced that breakfast cereal manufacturers’ have achieved a 49% reduction in sodium levels of branded breakfast cereals and levels in bread are being addressed (Food Standards Agency Website, 2010). Results of this study show the importance to infant diet of a reduction in the sodium content of both breakfast cereals and bread by the manufacturers. Despite the progress, it is essential that further reductions are made in coming years if we are to achieve the recommended daily maximum target (Gilbert and Heiser, 2005). It is possible that the UK recommended maximum is set too low for this age group, for adults a maximum of 6 g salt a day is recommended this would translate to 253 mg/MJ of dietary energy intake, therefore a level of 850 mg of sodium for an 8-month old consuming 2.75 MJ might be a more appropriate recommendation. Even with a higher recommendation, sodium levels still need to be reduced in infant's diets.

Important strengths of this study are its large sample size, high participation rate, comprehensive dietary intake and anthropometric data. However, there are several limitations; firstly, it was conducted in one geographical area of the UK and therefore results may not be applicable throughout the country, although the cohort was reasonably representative of the UK population at recruitment (ALSPAC Website). Furthermore, it has been shown that dietary intakes in the study are comparable with those of NDNS, a nationally representative cross-sectional sample. Second, these data were collected in 1993 and results may not be reflective of current patterns; the most recent infant feeding survey (Bolling et al., 2007) suggests that although some behaviours have improved slightly, these results still reflect current feeding practices. Sodium intake was estimated using diet diaries with no biological samples (for example, urine collection) to validate the method; however, they do provide a reasonable record of foods consumed and thus can be used to formulate advice.

In conclusion, sodium intakes need to be substantially reduced in children of this age group, as levels were much higher than the recommended maximum. Given that three-quarters of sodium in the diet comes from processed foods, successful salt-reduction strategies can only be achieved with the cooperation of the food industry (World Health Organisation, 2007). To reduce sodium levels in infants’ diets, first, introduce infants to a ‘baby-specific’ diet, using home-prepared foods, which are specifically adapted for the infant and, if desired, use ready-prepared infant foods. Second, clear advice stating what foods are suitable for infants should be given to all mothers including reiteration of advice not to use cows’ milk as a main drink before 12 months of age. Third, through food manufacturers reducing sodium levels in food products further.

References

  1. , , , , , et al. (2008). Complementary feeding: A commentary by the ESPGHAN committee on nutrition. J Pediatr Gastroenterol Nutr 46, 99–110.

  2. , (2010). An exploration of knowledge, attitudes and advice given by health professionals to parents in Ireland about the introduction of solid foods. A pilot study. BMC Public Health, doi:10.1186/1471-24-58-10-201.

  3. ALSPAC Website: .

  4. , , , (2007). Infant Feeding Survey. A survey conducted on behalf of the Information Centre for health and social care and the UK Health Departments by BMRB Social Research. The Information Centre: London.

  5. , (2010). There's no doubt about the health dangers of salt. New Scientist 2758, 22–23.

  6. (2007). Salt and cardiovascular disease - Legislation to cut levels of salt in processed food is necessary and justified. Br Med J 334, 859–860.

  7. (2009). Expert testimony on salt & cardiovascular disease. NICE CVD guidance expert testimony paper 3. NICE: London.

  8. Department of Health (1991). Report on Health and Social Subjects, No 41. Dietary Reference Values for Food Energy and Nutrients for the UK. Report of the Committee on Medical Aspects of Food Policy. HMSO: London.

  9. Department of Health (1994). Weaning and the weaning diet. Rep Health Soc Subj 45, HMSO: London.

  10. Department of Health and Social Security (1988). Present day practice in infant feeding: Third report. HMSO: London.

  11. , (2007). Sodium: can infants consume too much? Nutr Food Sci 37, 400–405.

  12. , , , ALSPAC Study Team (2000). Types of drinks consumed by infants at 4 and 8 months of age: A descriptive study. Public Health Nutr 3, 211–217.

  13. , , , (2008). Selected complementary feeding practices and their association with maternal education. Pediatrics 122, S91–S97.

  14. Food Standards Agency Website (2010): .

  15. , (2005). Salt and health: the CASH and BPA perspective. Nutr Bull 30, 62–69.

  16. , , , ALSPAC Study Team (2001). ALSPAC-The Avon Longitudinal Study of Parents and Children I. Study methodology. Paediatr Perinat Epidemiol 15, 74–87.

  17. , , , (2006). Maternal and child characteristics associated with infant and toddler feeding practices. J Am Diet Assoc 106, S135–S148.

  18. , , (1988). Cereals and cereal products. Third supplement to McCance & Widdowson's The Composition of Foods 4th edn The Royal Society of Chemistry: Cambridge.

  19. , , (1989). Milk products and eggs. Fourth supplement to McCance & Widdowson's The Composition of Foods 4th edn The Royal Society of Chemistry: Cambridge.

  20. , , , , , (1991). McCance & Widdowson's The composition of foods 5th edn The Royal Society of Chemistry: Cambridge.

  21. , , , , , (2007). Infant feeding in the second 6 months of life related to iron status: an observational study. Arch Dis Childhood 92, 850–854.

  22. In Practice - Cow and Gate Practical Support for Healthcare Professionals Website (2011): .

  23. , (2005). Early life determinants of adult blood pressure. Curr Opin Nephrol Hypertens 14, 259–264.

  24. Ministry of Agriculture, Fisheries and Food (1993). Food portion sizes 4nd edn HMSO: London.

  25. , , , (2010). Weaning infants onto solid foods. British Dietetic Association (BDA) Specialist Paediatric Group.

  26. National Institute of Clinical Excellence (2010). Prevention of cardiovascular disease at population level. Public health guidance 25. NICE: London.

  27. , , ALSPAC Study Team (2001). Food and nutrient intake in a cohort of 8-month-old infants in the south-west of England in 1993. Eur J Clin Nutr 55, 698–707.

  28. , , , (1998). Weaning practice in the Glasgow Longitudinal Infant Growth Study. Arch Dis Childhood 79, 153–156.

  29. Scientific Advisory Committee on Nutrition (2003). Salt and Health. The Stationery Office: London.

  30. World Health Organisation (2001). Global Strategy for Infant and Young Child Feeding. WHO: Geneva.

  31. World Health Organisation (2004). Global Strategy on Diet, Phyiscal Activity and Health. World Health Organisation: Geneva.

  32. World Health Organisation (2007). Reducing salt intakes in populations. World Health Organisation: Geneva.

  33. World Health Organisation (2011). Review and updating of current WHO recommendations on salt/sodium and potassium consumption. Call for Public Comments. WHO: Geneva.

  34. World Health Organisation Website (2011): .

Download references

Acknowledgements

We are extremely grateful to all the families who took part in this study, the midwives for their help in recruiting them and the whole ALSPAC team, which includes interviewers, computer and laboratory technicians, clerical workers, research scientists, volunteers, managers, receptionists and nurses. The UK Medical Research Council, the Wellcome Trust and the University of Bristol provide core support for ALSPAC. This publication is the work of the authors and PE will serve as guarantor for the contents of this paper. This research was specifically funded by Danone (Nutricia Ltd) and VC drafted the manuscript, performed the analyses and was involved in interpretation. JW and PE contributed to the interpretation of the data.

Author information

Affiliations

  1. School of Social and Community Medicine, University of Bristol, Bristol, UK

    • V L Cribb
    •  & P M Emmett
  2. DANONE Baby Nutrition (Nutricia Ltd), White Horse Business Park, Trowbridge, UK

    • J M Warren

Authors

  1. Search for V L Cribb in:

  2. Search for J M Warren in:

  3. Search for P M Emmett in:

Competing interests

The authors declare that funding was obtained from Danone (Nutricia Ltd); however, the work was carried out independently.

Corresponding author

Correspondence to P M Emmett.

About this article

Publication history

Received

Revised

Accepted

Published

DOI

https://doi.org/10.1038/ejcn.2011.137

Further reading