Early exposure to solid foods in infancy has been associated with the development of allergic diseases. However, scientific evidence for this is conflicting. The aim of this study was to examine the association between early exposure to solid foods in the infant’s diet and the development of eczema up to 4 years of age. We conducted an etiologic case–control study nested in the PIPO cohort (Prospective Cohort on the Influence of Perinatal Factors on the Occurrence of Asthma and Allergies). In this cohort data on nutrition, environmental exposures and parent-reported eczema were collected prospectively starting from 5 months pregnancy by means of questionnaires administered during two home visits and semi-annual postal questionnaires. In addition, detailed information on the timing of introduction of solid foods at individual food item level was collected at 1 year of age. Adjusted odds ratios and 95% confidence intervals were computed using logistic regression analysis as a measure of association between eczema and the timing of exposure to solid foods. Early introduction (within the first 4 months) of solid foods was inversely associated with eczema up to 4 years of age (adj OR: 0.49; 95% CI: 0.32–0.74). Moreover, we found that early exposure to solid foods was associated with a reduced risk for eczema only among children with allergic parents (adj OR: 0.35; 95% CI: 0.20–0.63), whereas no significant effect was found among children with non-allergic parents (adj OR: 0.69; 95% CI: 0.37–1.29). The results of this study show that early exposure to solid foods is associated with less parent-reported eczema in children, particularly among children with allergic parents. Therefore, the current study does not support a delayed introduction of solid foods for the prevention of eczema in childhood.
Early exposure to solid foods in infancy has been positively associated with the development of allergic diseases (1–4). For that reason, prophylactic nutritional guidelines for the prevention of asthma and allergy have been established for the general population as well as for children at increased risk for asthma and allergy (usually children with a positive family history of asthma and allergy). These guidelines recommend exclusive breastfeeding during the first 4–6 months of life with complementary foods (‘beikost’) after the 5th month (Joint Statement of the Nutritional Committees from the European Society for Paediatric Allergology and Clinical Immunology (ESPACI) and the European Society for Paediatric Gastroenterology, Hepatology and Nutrition (ESPGHAN) (5).
In 2001 the World Health Organization (WHO) tightened these recommendations to 6 months of exclusive breastfeeding and the introduction of solids thereafter (6).
In his review Zeiger compared and reconciled the differences between these recommendations and presented the evidence that has led to their statements (7). The advice of the American College of Allergy, Asthma and Immunology regarding the timing of the introduction of solids to infants, published in July 2006, is more detailed, suggesting to cautiously individualize the introduction of solids into the infant’s diet (8). For all infants weaning can be started from the 6th month, but introduction of eggs, nuts, fish and seafood require caution. ‘The optimal age for the introduction of selected supplemental foods should be 6 months, dairy products 12 months, hen’s egg 24 months and peanut, tree nuts, fish and seafood at least 36 months’. Wide-ranged.
In a recent review by an expert group of the Section on Pediatrics, European Academy of Allergology and Clinical Immunology (SP-EAACI) the recommendation of exclusively breastfeeding for 4–6 months combined with avoidance of solids and cow’s milk for the first 4 months has been reaffirmed (9). However, some reservations have been made in a recent review of these guidelines by the American Academy of Pediatrics (AAP) (10). Although solid foods should not be introduced before the age of 4–6 months, there is no convincing evidence that delaying their introduction beyond this period has a significant protective effect on the development of atopic disease. This includes the introduction of highly allergic foods such as fish, eggs and foods containing peanut protein.
Disagreement on these guidelines may be due to the limited scientific evidence which is based on a few studies only. Some investigators report that early introduction of solid foods is associated with an increased risk for the development of allergy and asthma (1–4), whereas others found no evidence to support a protective effect of delayed solids introduction on the development of eczema, asthma and atopic sensitization (11–13). Furthermore, two studies demonstrated that delayed introduction of solids may even increase the risk of developing wheat allergy (14) and atopy (15).
A systematic review of the literature in Medline, the Cochrane Library and the Drugs and Pharmacology section of Embase with regard to the relationship between early introduction of solid foods to infants and the development of allergic disease by Tarini et al. revealed that early solids introduction may increase the risk of eczema, but is not associated with other allergic diseases (16). In another review related to the timing of solid food introduction and food allergy, Fiocchi et al. concluded that early solids introduction can increase the risk of food allergy whereas avoidance of solids can prevent the development of specific food allergies (8).
It appears that current scarce and inconsistent literature does not allow an authoritative statement to be made regarding the relationship between the timing of solids introduction and the development of allergy. The conflicting results between the different studies may in part be explained by methodological differences, such as selection of study population with respect to family history of allergy or gestational age of the child, different cut-off points for solids introduction and breastfeeding duration, inadequate definition of outcomes and inadequate confounding factors. Furthermore, differences in recall with respect to feeding history may produce bias by differential misclassification. Different duration of follow-up and the ages at which outcomes are assessed may also be relevant for the discrepant findings. Moreover, recent studies have to consider the fact that outcome influences exposure (so-called ‘reverse causality’) (12).
At present, public awareness regarding possible associations between dietary interventions and the development of allergy is very high. Health professionals are advising parents to delay the introduction of solids to infants with early signs of allergy. This may mask a true protective effect or even result in spurious positive associations between delayed solids introduction and allergic outcomes (12). Even in infants without early allergy, a positive family history for allergy by itself may lead to a delayed introduction to solids.
The aim of the current study was to investigate whether early introduction of solid foods is a risk factor for the development of eczema up to 4 years of age in a population based prospective birth cohort.
Study design and population
We conducted an etiologic study using a nested case-control design. Data were collected within the Prospective Cohort on the Influence of Perinatal Factors on the Occurrence of Asthma and Allergies (PIPO cohort), which started in 1997 in Antwerp, Belgium, in children followed from the 5th month of pregnancy. Two thousand pregnant women were invited to participate through four gynecologists between June 1997 and October 2001. Of these, approximately 55% were willing to participate in our study, resulting in a total study population of 1128 children. Data were collected through two home visits and semiannual postal questionnaires. At the age of one and 4 years of age the children and their parents were invited to visit the study center for a medical examination and a blood sample for allergy testing. At present the children are still in follow-up.
Subjects were included in the current study if information on eczema up to 4 years of age and on the timing of solid foods introduction was available. The Committee for Medical Ethics of the University Antwerp granted approval for the assessments. The parents of the participants gave their written informed consent before each assessment.
Detailed data on parental history of allergic diseases and environmental factors during pregnancy, and information about the child’s birth weight and gestational age were collected from questionnaires completed by study nurses during home visits at 5 months pregnancy and 3 months after birth. Information on infant feeding practices, environmental exposures and manifestations of eczema were collected prospectively using semiannual questionnaires between 6 and 48 months of age. Additionally, when the children were 1 year of age, parents provided information on the introduction of 14 solid food items to the child. The question asked was: ‘How old (in weeks) was your child when you gave him/her the following food items to eat or drink? Fruit juice, fruit, vegetables, dairy products, wheat containing product (e.c. bread, biscuit rusk, cake, cereals, crackers, pasta, porridge, etc.), meat, cheese, fish, dried fruits (e.c. raisins), peanut butter, egg (or egg containing food, e.c. cake), chocolate, soft drinks, crisps or salty snacks’.
Serum IgE measurements
In a subsample of 325 children (177 cases and 148 controls) blood was drawn for quantification of specific serum IgE of the following allergens: Dermatophagoïdes pteronyssinus, cat, dog, egg’s white and cow’s milk using CAP FEIA system (Pharmacia Diagnostics, Brussels, Belgium). Detection limits were 0.35 kUa/l and 100 kUa/l for specific serum IgE. Specific IgE measures below 0.35 kUa/l were considered as negative. Children were considered atopic if at least one specific IgE was positive. All blood analyses were performed at the Laboratory of Immunology at the University Hospital Antwerp.
Eczema was assessed as ‘parent-reported eczema’ by an affirmative answer on any of the appropriate questions asked in the questionnaires up to 4 years of age [questions derived from validated questionnaires of the International Study of Asthma and Allergies in Childhood (ISAAC)] (17). The question used for the assessment of parent-reported eczema was: ‘Has your child had eczema during the last 6 months?’ According to the Report of the Nomenclature Review Committee of the World Allergy Organization (October 2003), eczema should be referred to as ‘eczema’ as long as the immunological mechanism of this disease is unclear (18). Cases were defined as children with one or more episodes of parent-reported eczema during the first 4 years of life. Controls were defined as children without parent-reported eczema up to 4 years of age.
The expression ‘solid food’ (or ‘solids’) may not be the most appropriate, because of one liquid food item, but we preferred ‘solid food’ over ‘complementary food’ because it is more widely established in literature. Solids were defined as any one of the food items vegetables, fruits, fruit juice, wheat products, meat, cheese, dairy products (apart from formula milk), egg, fish, dried fruits, peanut butter, chocolate, soft drinks and crisps. Therefore, exposure to solid food was assessed at the level of individual food items, regardless the frequency at introduction. Early exposure to solids was defined as receiving any of these solid foods during the first 4 months.
Cases were considered as exposed if they were introduced to solids during the first 4 months and before the occurrence of eczema. Controls were considered as exposed if they were introduced to solids during the first 4 months. Cases were considered as unexposed if they were introduced to solids after 4 months or during the first 4 months but after the onset of eczema. Controls were considered as unexposed if they were introduced to solids after 4 months.
Ever breast-fed infants were defined as children who had ever consumed breast milk. Parental allergy was defined as having at least one parent with doctor-diagnosed asthma and/or eczema and/or hay fever (vs. both parents non-allergic).
The outcome in the present study was parent-reported eczema of the child up to 4 years of age. Because eczema is a status-type outcome, prevalence odds ratios (OR) with 95% confidence intervals (CI) were computed as a measure of association between exposure to solids and eczema. Crude associations were assessed by contingency tables and evaluated by Pearson’s chi-square tests or Fisher’s exact test, when appropriate. To evaluate the association of eczema with continuous variables independent samples t-tests (for maternal age and child’s birth weight - both distributions were normal) and Mann–Whitney U-test (for breast feeding duration and formula milk introduction) were used. Multiple logistic regression analyses were applied to adjust for potential confounders of the association between early exposure to solids and the occurrence of eczema. Results were expressed as adjusted odds ratios (adj OR) with 95% confidence intervals. Covariates in the multiple logistic regression models were selected based on univariate analyses and evidence from previous studies (12, 14): child’s gender, birth weight and birth order, maternal age, parental allergy and educational level, smoking in pregnancy, passive smoking up to 4 years and breast-feeding. Moreover, interaction between parental allergy and early introduction of solids was investigated by inclusion of interaction terms in the regression models. Statistical analysis was performed using spss for Windows V13.0 (SPSS, Chicago, ILL, USA).
General characteristics of cases and controls
In total, 252 children with eczema (cases) and 305 children without eczema (controls) were included in the study population. Table 1 shows the general characteristics of the two subgroups. The proportion of boys and girls was almost equal. Children with eczema had more often a positive parental history of allergic diseases and highly educated parents, and were also significantly more often atopic. Furthermore, they were more often exposed to environmental tobacco smoke (pre- and post-natal), however, this difference was not statistical significant.
Table 1. General characteristics of the study population for children with eczema (cases) and without eczema (controls) up to 4 years of age
Cases (N = 252), n (%)
Controls (N = 305), n (%)
OR (95% CI)
Comparisons (univariate) between cases and controls expressed as crude odds ratios with 95% confidence interval and p-values.
Descriptive statistics were carried out using all available data for each of the variables.
*Doctor’s diagnosed asthma and/or eczema and/or hay fever.
†N = 325 (177 cases, 148 controls).
‡Excluding cow’s milk formula.
§Exact confidence limits.
Smoking during pregnancy
Passive smoking up to 4 years
Maternal allergic disease(s)*
Paternal allergic disease(s)*
Parental allergic disease(s)*
Atopy (specific IgE in serum)†
Prematurity (≤37 wk)
Educational level mother (high)
Educational level father (high)
Breastfeeding >4 months
Breast feeding duration in weeks, mean (s.e.)
Formula milk first 4 months
Birth weight in grams, mean (s.e.)
Maternal age in years, mean (s.e.)
Solids introduction first 4 months
Exposure to any solids
At 4 months of age, 52.5% of controls and 39.7% of cases had been introduced to any solids (Table 1, lower part). In comparison with their asymptomatic peers, children with eczema were less often introduced to fruit, fruit juice, vegetables and cereals, and slightly more often exposed to meat, fish and dairy products within the first 4 months. Apart from fruit juice, these differences were not statistically significant. Furthermore, none of the cases had been introduced to eggs and cheese up to 4 months, compared with around one percent of the controls.
In addition, almost 80% of cases and 83% of controls were ever breast-fed (crude OR 0.77; 95% CI 0.50–1.18). No differences were found between cases and controls concerning breast-feeding for more than 4 months or mean duration of breast feeding. Also no difference was found between the times of introduction of formula feeding (mean age 10.2 wk and 10.1 wk for cases and controls respectively).
Table 2 shows the type of solid food introduced first among cases and controls during the first year of life. Seventy-seven percent of cases and 74 % of controls received one solid at a time at introduction, whereas the rest of the children received two or more solid food items at introduction. Most often introduced as first solids among both groups were vegetables (around 40%), followed by fruit juice and fruit.
Table 2. Distributions of the types of solids introduced first in cases and controls
Solid foods first introduced
Cases n (%)
Controls n (%)
Introduction of a single food type or combinations of more than one food type within the first week of introduction.
One solid in first week
Two or more solids in first week
Fruit and fruit juice
Fruit juice and vegetables
Fruit and vegetables
Fruit and cereal
Parent-reported eczema at some time within the first 4 years of life was reported in 252 children (cases). Of these, 152 (60.3%) had signs of eczema during the first year, whereas 27.6% of cases had eczema during the fourth year of life. When comparing subjects with and without eczema up to 4 years of age (Table 1), exposure to any solids during the first 4 months was associated with reduced odds of having eczema up to 4 years (crude OR 0.60; 95% CI: 0.43–0.84). After adjustment for confounders, the association between early exposure to solids and eczema remained significant: adj OR with 95% CI: 0.49 (0.32–0.74) (Fig. 1). Furthermore, assessment of adjusted OR with 95% CI for subgroups with and without parental allergy separately showed that early solids exposure was associated with reduced risks for eczema in both subgroups of children, but significant associations were found only among children with allergic parents: adj OR with 95% CI: 0.35 (0.20–0.63) for children with allergic parents and 0.69 (0.37–1.29) for children with non-allergic parents (Fig. 1). However, the interaction term between parental allergy and early introduction of solids was non-significant (p=0.13) and was therefore excluded from the final model.
In addition, the regression model showed that parental allergy was an independent (statistically significant) risk factor for eczema: adj OR (95% CI) 1.73 (1.16–2.59). Smoking during pregnancy, passive smoking of the child up to 4 years of age and having older siblings were associated with increased odds for having eczema, but these associations were not statistically significant: adj OR with 95% CI: 1.49 (0.66–3.38), 1.96 (0.77–4.95) and 1.43 (0.92–2.21) respectively.
In this nested case–control study, we found that early exposure to solid foods is associated with a reduced risk for eczema among children, in particular in children with allergic parents. This association was independent of the effects of parental allergic disease, passive smoking or breast-feeding on the occurrence of eczema.
Evidence that early exposure to solid foods predisposes to the later development of allergies remains controversial. In contrast to the present study, a number of previous observational studies have demonstrated that early weaning is associated with increased risk of allergic diseases. A study by Fergusson et al. in a cohort of 1265 children from New Zealand showed more eczema up to 10 years of age in children introduced to four or more solids before the age of 4 months compared with children not receiving any solids during the first 4 months (1). A Finnish cohort study in 113 high-risk infants compared 6 months exclusively breast fed children with infants who started with solids at the age of 3 months (2). They found a protective effect of solid food avoidance during the first 6 months on the development of eczema and food allergy at the age-of-one, but no difference was seen at 5 years follow-up. A positive association between early introduction of solids and wheezing during childhood was reported by Wilson et al. (3). Furthermore, Morgan et al. found that introduction of solids by or before 17 wk in 257 pre-term children was positively associated with the development of eczema by 12 months of age in particular among male infants (4). However, delayed weaning of pre-term children may have had medical reasons, therefore possibly biasing this association.
On the other hand, few studies found no evidence for a protective effect of delayed introduction of solids on the development of asthma, eczema and atopic sensitization (11–13, 19). Our finding that early exposure to solids has a ‘protective’ effect against eczema is consistent with three previously published studies (14, 15, 20). A longitudinal birth cohort study from the USA by Poole et al. among 1612 children not selected for familial allergy found that delayed introduction of solids does not prevent, but may even increase the risk of developing atopic disease (14). The most common food in the US introduced initially into the infant’s diet was rice cereal, followed by oatmeal cereal. The authors found that delaying exposure to grain cereals until 6 months was associated with an increased risk of developing parent-reported wheat allergy. However, no objective measurement of wheat allergy was used in this study.
A recent study by Mihrshahi et al. (15) has shown that early introduction of solid foods (by 3 months) and earlier cessation of breast-feeding were not associated with the occurrence of asthma or eczema at 5 years, but had a significant protective effect against atopy (as measured by positive skin prick tests) to inhalant or ingested allergens at 5 years of age. This was a prospective study in a cohort of children with high risk of allergic disease, in which all mothers were advised to breast-feed for as long as possible and to avoid introduction of solid foods before 4–6 months. According to these authors, the explanation for a protective effect of early exposure to solids against atopy remains uncertain. However, they suggest that reverse causality (i.e., avoidance of exposure in those who already have symptoms of allergic diseases) may partly explain this association. Another recent study by Zutavern et al. among 2073 children in the LISA birth cohort also found no protective effect of a delayed introduction of solids beyond 4 months of age on the occurrence of asthma, allergic rhinitis and food or inhalant sensitization at the age of 6 years (20). However, for eczema, the results were unclear.
Reverse causality constitutes a problem especially in more recent studies, because of growing public awareness regarding the association of feeding practices with allergic diseases. The first study to investigate the relationship between the timing of solids introduction and atopic dermatitis considering reverse causality was published in 2006 by Zutavern et al. (12). Delayed introduction of solids beyond the 6th month was not associated with the prevention of atopic dermatitis and atopic sensitization. In addition, the authors found strong evidence for reverse causality between ‘early skin or allergic symptoms’ and the timing of solids introduction. However, in our study no significant differences were found in duration of breast feeding or a delayed introduction of solid foods between infants with or without early allergic symptoms. Also, no association existed between parental allergy and the timing of solids introduction and breastfeeding duration (data not shown). Moreover, our finding of a stronger ‘protective’ association between early exposure to solids and eczema among children with allergic parents makes the effect of reverse causality less likely.
At the follow-up of their first study Zutavern et al. accounted for reverse causality by analysing a subgroup of children without early skin or allergic symptoms. We preferred not to exclude any subjects from the analysis. Thus, to avoid reverse causality, we considered children with eczema as exposed under the condition that the onset of eczema occurred after solids were introduced. To our knowledge this is the first study that investigates associations between the timing of exposure to solid foods and eczema taking into account the temporal sequence of exposure and eczema for each child.
An important finding in the present study was that the introduction of solid foods within the first 4 months was ‘protective’ for the occurrence of eczema up to 4 years of age. This effect was found in univariate analysis and remained statistical significant after adjustment for potential confounders. Furthermore, when allergic status of the parents was taken into account, it appeared that early exposure to solids was ‘protective’ in particular among children with allergic parents.
A possible explanation for the ‘protective’ effect of early introduction of solids on the development of eczema may be the induction of oral tolerance for foods. It means that more prolonged exposure to foreign antigens (food proteins) could promote immune tolerance; however, the underlying immunological mechanisms remain unclear. A study by Fleischer showed that the chance of recurrence of peanut allergy (after resolution) is significantly higher in children avoiding peanuts as compared with those who ate peanuts frequently (21). Thus, continuation of exposure of the immune system to certain allergenic food proteins once responsible for allergy, may help to maintain tolerance to this food. Bjorksten has suggested in 2004 that future efforts towards allergy prevention may potentially be based on various ways to increase induction of tolerance for allergens rather than by avoidance regimens (22). Our findings may support this suggestion. However, because we did not quantify the early solids exposures, we cannot confirm the accuracy of this hypothesis. Furthermore, it remains unclear why the effect of early exposure to solids differs between children with and without a family history of allergic disease. Therefore, as the underlying mechanisms are unclear, one should be cautious in interpreting the association between early exposure to solids and eczema as ‘protective’.
The strength of our study is the collection of pre- and post-natal exposure data, and the semiannual assessment of exposure and outcome status, which allowed us to identify more accurately the time sequence of exposure and outcome and, as a result, cases and controls. Moreover, the design of the data collection process allowed us to include infants both with and without parental allergy.
On the other hand, some limitations may have affected our results. Firstly, besides the aforementioned lack of quantification of the solids exposure, the outcome used was ‘parent-reported’ eczema, which was not confirmed by a doctor. Therefore, some of the cases may not be true eczema cases. This misclassification may have led to an underestimation of the associations in our results. Secondly, a recall bias may have overestimated exposure among the cases, by which parents from children with eczema could have been more likely to report early exposure to solids; however, this should have resulted in a positive association between early exposure and outcome. Finally, we could not compare our results with some of the previous studies using a cut-off point for solids introduction of 6 months, because of the small number of children introduced to solids after 6 months.
In conclusion, our study showed that early initial exposure to solid foods is associated with a reduced risk for the development of parent-reported eczema up to 4 years of age, in particular in children from allergic parents. Therefore, we have no evidence to support a delayed introduction of solid foods in the prevention of allergic disease in children. Consequently, our findings add to the small, but increasing number of studies suggesting that solid foods avoidance or elimination in early childhood may not protect from allergic disease.
We thank the families for their participation, and all PIPO-study personnel and nurses for the recruitment and follow-up of the families:
P. Vermeire †, A. Vellinga, H. Van Bever, V. Nelen, C. Bridts, E. Vandervaeren, L. Thijs, M. Duym, G. Meysen, R. Vroom, D. Stappers, K. Vandekerckhove, R. Heyndrickx, F. Willekens, C. Daenen, P. Beyltjens and H. Nafez. We also thank G. François for his advice on the manuscript.
This study was supported by grants from the Fund for Scientific Research Flanders, Belgium (number 7.0015.00) and the Flemish Government (PBO98/26/143).