- Email: [email protected]
Restricted diets in children with reactions to milk and egg perceived by their parents
C LINICAL AND L ABORATORY O BSERVATIONS
Restricted diets in children with reactions to milk and egg perceived by their parents Merete Eggesbø, MD, Grete Botten, MD, PhD, and Hein Stigum, PhD The aim of this study was to investigate the degree to which parents alter the diets of their children on the basis of perceived reactions. From a population-based sample of 2979 2-year old children with reactions to egg or milk perceived by their parents, one third had strict limitations on the intakes of these foods, representing 2.5% of the children in the cohort. In approximately 1 of 6 families the strict diets were initiated without consulting a doctor, and in a substantial proportion the restrictions were unwarranted. High maternal education level and irritability attributed to food were among the risk factors for unwarranted diets. On the other hand, many children, in whom an adverse reaction was verified, lacked appropriate diet restrictions. We conclude that the handling of adverse reactions to food frequently occurs outside the medical care system at the cost of correct diagnosis and appropriate diets. (J Pediatr 2001;139:583-7)
Milk and eggs are important sources of nutrients in childhood, and if they are removed from the diet because of adverse reactions, the child’s nutritional needs may not be met.1-6 Nutritional counseling of the parents and prescription of adequate substitutes for the child may prevent inadequate elimination diets.5,6 Parents may, however, alter the diets on their own initiative, on the basis of perceived reactions, or following advice from a paramedical group. Anec-
dotal reports have been published on children with severe malnutrition resulting from strict elimination diets imposed on them by their parents or paramedical groups.7 Because parents generally tend to overestimate the role of food as the cause of symptoms in their children, these diets are likely to be unwarranted as well. A substantial number of parents perceive that their children have adverse reactions to milk and egg, 11% to 15% and 2% to 4%, respectively, by age 2.8,9
From the National Institute of Public Health, Department of Epidemiology, and the Center for Health Administration, The National Hospital, University of Oslo, Norway.
Funded by grants from the Norwegian Research Council. Submitted for publication July 7, 2000; revision received Feb 20, 2001; accepted May 1, 2001. Reprint requests: Merete Eggesbø, Section of Epidemiology, Department of Population Health Sciences, National Institute of Public Health, Post Box 44 04 Torshov, N-0403 Oslo, Norway. Copyright © 2001 by Mosby, Inc. 0022-3476/2001/$35.00 + 0 9/22/117067 doi:10.1067/mpd.2001.117067
The potential for unguided elimination diets may thus be substantial. However, the degree to which parents actually alter the diet of their children on the basis of perceived reactions to food remains unknown. The aim of this study was to study the occurrence of milkand egg-restricted diets in a populationbased sample of children with perceived reactions to these foods at age 2 and to study the relationship between diet and a verified diagnosis.
METHODS The study population included children born consecutively in Oslo, Norway, in 1992 to 1993. They were recruited at the maternity ward and studied prospectively with postal questionnaires addressed to the parents every 6 months until the child reached age 2; 2979 families (82.2%) responded to the questionnaires when the child was 2 years old. Details of the study design and collection of data have been described elsewhere.9,10 RR OR
Relative risk Odds ratio
Data Collection Information regarding the occurrence and type of any reaction to food was obtained at 12, 18, and 24 months,9 and information regarding type of diet was obtained at age 24 months. Of the 206 families who reported a reaction to milk
583
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS OCTOBER 2001
Table I. Children with strict diet alterations that were unwarranted, among 46 children with parentally perceived reactions to egg or milk which could not be reproduced, according to attributed symptom and socioeconomic factors
Children on an unwarranted strict diet n
n
%
Yes No Yes No Yes No Yes No Yes No
46 18 28 30 16 6 40 9 37 5 41
17 6 11 13 4 6 11 5 12 5 12
37 33.3 39.3 43.3 25 100 27.5 55.6 32.4 100 29.3
Socioeconomic factors High maternal Yes No education† Low maternal age‡ Yes No Yes Nonsmoker§ No Older siblings Yes No
24 22 23 23 38 8 28 18
12 5 9 8 15 2 15 2
50 22.7 39.1 34.8 39.5 25 53.6 11.1
All symptoms Skin Gastrointestinal Vomiting Stomachache Irritability
Adjusted P value*
RR
CI
0.8
(0.4-1.9)
—
1.7
(0.7-4.4)
—
3.6
(2.2-6.0) .01
1.7
(0.8-3.6)
3.4
(2.1-5.5) .03
2.2
(0.9-5.2) .04
1.1
(0.5-2.4) .02
1.6
(0.4-5.6)
4.8
(1.2-18.6) .001
—
—
Two separate regression analyses were performed; the first included symptoms and the second included socioeconomic factors. High income was also analyzed and showed no association with type of diet (data not shown). All variables that reached a significance level of 0.15 (Pearson χ2 test) in the bivariate analysis were included in the multivariate analysis in addition to maternal age, which was included as a confounder. *Significance tested by log exact test. †Maternal education was categorized into 12 years or less and more than 12 years of education. ‡Maternal age was categorized into 30 years or younger and older than 30 years. §Maternal smoking during pregnancy.
or egg in their child at age 2, 138 families were asked to participate in an examination to confirm the reported reaction; 86 families (62%) completed the examination at a mean age of 2.5 years (CI 2.5 to 2.6).11,12 The perceived reaction to egg or milk was considered confirmed when one of the three following criteria were fulfilled: a history of an immediate reaction and corresponding high levels of specific IgE; an objective reaction to an open food challenge; or a reaction verified by double-blind placebo-controlled food challenges.11,12 If the diagnostic criteria were partly fulfilled, the perceived reaction was assessed as “possible.” Children in whom a current adverse reaction 584
was refuted but who had a convincing history of an earlier allergy were classified as “possible earlier allergic.”
Variables MILK- OR EGG-RESTRICTED DIET. The main question, on which the operational definition of the outcome was based, was whether the perceived reaction to food had resulted in removal of any food items or liquids from the child’s diet, in which case the alterations were required to be specified. The diets were categorized into 3 groups: (1) Strict diets: the main product and all or most byproducts had been eliminated; (2) Moderate
diets: only the main product and not the byproducts had been removed; (3) the option “No diet alterations” was marked or the alteration was insubstantial. In some cases when merely “milk” or “eggs” had been stated, information concerning the amount of food needed to elicit a reaction was used to determine whether all products or only the main product was being referred to. The diet was classified as “currently restricted” if it was stated that the food had not been reintroduced again or it had been reintroduced again but had caused the symptoms to recur. BACKGROUND VARIABLES. Information was obtained on whether a physician had been consulted because of the perceived food reaction and whether, because of illnesses of the child, the parents had sought treatment by others than doctors, such as homeopaths and chiropractors. Information on whether the infant was the mother’s first-born child, maternal age in years, maternal smoking habits at the time of the birth, maternal education in years, and family income was obtained from the questionnaire completed by the participating families at the maternity ward.10 The Norwegian Data Inspectorate and the Regional Ethics Committee for Medical Research approved all stages of the study.
Statistical Analysis The main outcome variable was “strict diets” with the categories “yes” and “no.” Two outcomes on inappropriate diets were derived: strict diet among children with perceived reactions that are disproved (Table I) and moderate or no diet, which were combined to comprise the outcome variable “lack of strict diet” among children in whom a food reaction is confirmed (Table II). The normal approximation to the binomial distribution was used for the computation of CIs. All variables that reached a significance level of 0.15, as assessed by means of the Pearson χ2 test in the bivariate analysis, were included in the
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS
VOLUME 139, NUMBER 4 multivariate models. In addition, maternal education and maternal age were always included in the model as potential confounders. Adjusted associations were obtained by multivariate logistic regression. Because of the small numbers, the inappropriate diets were analyzed by exact logistic regression, with a separate analysis for two sets of covariates, attributed symptom and socioeconomic factors. In these analyses the outcome is not rare, and thus the odds ratios are inflated compared with relative risks. Also there are empty cells for some of the categories in Tables I and II. This situation is handled by the exact regression, but we found the estimated odds ratios difficult to interpret. We have therefore chosen not to present them but to show only the significance levels in the adjusted analyses. The analysis were performed with SPSS (Release 6.1; SPSS Inc, Chicago, Ill) and the LogExact software (Cytel Software Corp, Cambridge, Mass).13
RESULTS A prerequisite for initiating a strict diet is to have perceived a food reaction. Thus socioeconomic factors work at two levels: on the parental perception and then on the choice of diet.
Perceived Reactions to Egg and Milk There was no difference in the distribution of maternal education level, maternal age, or maternal smoking habits among parents who perceived a reaction to food compared with those who did not. The proportions of low-income families and children with older siblings were slightly higher among those with perceived reactions to egg or milk than among those without perceived reactions.
The Prevalence of Restricted Diets The cumulative incidence of eggand milk-restricted diets are shown in
Table II. Children without strict diet alterations among 40 children with confirmed reactions to egg or milk, according to attributed symptom and socioeconomic factors
Children lacking indicated restrictions n
n
%
Yes No Yes No Yes No Yes No Yes No
40 30 10 17 23 4 36 3 37 7 33
17 10 7 7 10 1 16 0 17 1 16
42.5 33.3 70 41.2 43.5 25 44.4 0 45.9 14.3 48.5
Socioeconomic factors High maternal Yes No education‡ High Yes No maternal age§ Yes Smoker No
24 16 20 20 7 33
9 8 6 11 5 12
37.5 50 30 55 71.4 36.4
All symptoms Skin Gastrointestinal Vomiting Stomachache† Irritability
RR
CI
Adjusted P value*
0.5
(0.2-0.9)
.07
0.9
(0.5-2.0)
—
0.6
(0.1-3.2)
— 1
0.3
(0.0-1.9)
.7
0.8
(0.4-1.5)
.86
0.5
(0.3-1.2)
.39
2
(1.0-3.8)
.32
Two separate regression analyses were performed, the first included symptoms and the second included socioeconomic factors. High income and older siblings were also analyzed but showed no association with type of diet (data not shown). All variables that reached a significance level of 0.15 (Pearson χ2 test) in the bivariate analysis, were included in the multivariate analysis. Maternal age and education were included as potential confounders. *Significance tested by log exact test. †RR estimates for lack of diet cannot be estimated because of empty cells. ‡Maternal education categorized into 12 years or less and more than 12 years of education. §Maternal age categorized into 30 years or younger and older than 30 years. Maternal smoking during pregnancy.
the Figure. Of the 206 children with parentally perceived reactions to egg or milk at age 2, strict diets had been initiated in 73 children, and both egg and milk had been restricted in 17 of these children. The cumulative incidence of strict milk- or egg-restricted diets by age 2 is thus estimated at 2.5% (CI, 1.9 to 3.0). At age 2 years, 66 children were still on strict diets, giving an estimated point prevalence of 2.1% (CI, 1.6 to 2.6). Among the children on a strict milk-restricted diet, 32.1% and 5.7% had additional restrictions on the intake of egg and fish, respectively. Among the children on a strict egg-restricted diet, the corresponding figures were 45.9% and 21.6%, for additional restrictions on
the intakes of milk and fish, respectively.
Consultations Overall 54.4% had consulted a physician concerning the perceived adverse reaction, independent of the type of food, increasing to 80.6% where reactions were attributed to both egg and milk or 84.9% where strict diet alterations were reported. Among the parents of children with perceived reactions to egg or milk, 21.7% had sought treatment by paramedical groups, as opposed to 11% in families with children without reported reactions to any food (relative risk [RR] = 2; CI, 1.5 to 2.7). Information was not available on 585
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS OCTOBER 2001
Figure. Cumulative incidence of strict and moderate diet restrictions among 206 children with parentally perceived reactions to egg and milk at age 2, in a population-based sample of 2979 children, according to incriminated food.
which illnesses motivated the consultations.
Children Who Had Received Unwarranted Strict Diets Among the 86 children who had been examined with objective diagnostic methods by the researcher, the diagnosis was disproved in 46 children. Among these children 17 had received strict diets (Table I). However, 8 of these 17 children were suspected of having recently outgrown their food reactions, in which case the restricted diets may not have been truly unwarranted. When these 8 children were excluded, a minimum of 20% unwarranted strict diets is estimated. Unwarranted diet restrictions were significantly more common among children in whom irritability and vomiting were attributed to the food compared with other symptoms (Table I). Furthermore, having older siblings, low maternal age, and high maternal education level were significantly associated with unwarranted diets (Table I).
Children with Confirmed Reactions and Lack of Strict Diets In 40 of the 86 examined children the perceived reaction was either confirmed or assessed as possible. Seventeen of these 40 children lacked strict diet restrictions (Table II). The factors associ586
ated with lack of diet restrictions in children in whom it was indicated are shown in Table II.
DISCUSSION This study shows that among children with parentally perceived reactions to egg or milk, one third had had strict limitations on the intake of these foods. Strict alterations were unwarranted in a minimum 1 of 5 children, and high maternal education was a risk factor for unwarranted diets. On the other hand, more than one third of children with confirmed adverse reaction to egg or milk lacked appropriate diet restrictions. The children were not evaluated at the time of the initial report of the perceived reaction, thus allowing for spontaneous recovery. This may result in an overestimation of children on unwarranted diets. When children with a possible earlier allergy were excluded from the multivariate analysis on risk factors associated with unwarranted diets, similar results were obtained. Thus we believe that this misclassification does not affect the validity in predicting risk factors apart from possibly underestimating the effects of the risk factors. The information obtained from the parents was not always sufficient for
certain assessment of the strictness of the elimination. However, these misclassifications are assumed to be nondifferential and should not affect the validity of the results. The cumulative incidence of restricted diets by age 2 is an underestimate because we did not have information on the diets of children in whom a reaction to milk or egg had resolved before age 2.9 The point prevalence estimate, however, is not affected by this uncertainty. The groups were small, and, hence, it is difficult to draw firm conclusions from the study. The sample size is small, making the study prone to type II errors, that is, not detecting factors that may be of importance. On the other hand, the observed risk factors are likely to be of major importance because statistical significance was obtained in spite of the small numbers. Although it is well documented that adverse reaction to food is overreported,8,11,12,14 which factors influence the parental perception of adverse reaction to food is less well understood. The association between low family income and perceived reactions may be explained by an actual difference in the underlying prevalence or a report bias. Maternal education level is generally considered the best measure of socioeconomic class in Norway, and a report bias tied to socioeconomic class would be expected to include maternal education. However, this was not found. In contrast to our results, a study from England reported a strong association between a high maternal education level and the child being perceived as food intolerant.15 Furthermore, children with older siblings are at increased risk of reported adverse reactions. This is opposed to the consistently reported findings of an inverse association between various manifestations of allergy and sibship size,16 but it may be a report bias resulting from increased parental awareness of adverse reactions in subsequent children. In the previous study from England, 67% of parents who perceived a food
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS
VOLUME 139, NUMBER 4 reaction in their children had consulted a doctor,15 which is in keeping with this study. The restricted diets were the parent’s decision in more than half of those who had consulted a doctor.15 In this study we lack information on whether the diet restrictions actually were initiated following doctor’s advice. By age 2, 1.8% and 1.2% of the children had had major restrictions on their dietary content of milk and egg, respectively. This is similar to the findings of the English study in which the corresponding numbers were 1.1% and 0.9%, respectively; although the studies are not completely comparable because of differences in design and the operational definition of restricted diet.15 The high number of children on milkrestricted diets is of great concern. The potential hazards of milk-restricted diets in which the child’s nutritional demands may not be met are well documented.1-5 Although nutritional counseling of the parents and prescription of adequate nutritional substitutes may prevent inadequate diets,5,6 our results indicate that approximately 1 of 6 families initiate strict diets without seeking medical advice. Furthermore, a sufficient diet will depend on whether the doctor actually initiated the diet, a milk-substitute was prescribed and nutritional counseling was given, and whether the child was followed up over time. Diets solely restricted in eggs are less prone to lead to nutritional inadequacy1; however, the egg restrictions were combined with milk restrictions in nearly half of the children, in which case there is increased risk of protein deficiency. Although a less voiced concern, the lack of strict diets in some children with an actual allergy deserves attention too. The scientific basis for advice is generally sparse, and a lack of knowledge among physicians may leave parents to venture into their own decision-making on the strictness of the diet.
Although there may be cultural differences in the perception and parental handling of food reactions, studies so far have shown remarkable similarity between countries. This also applies to the phenomenon of parental overreporting of food reactions compared with what can be verified by objective methods. We are not aware of earlier studies that have actually studied the appropriateness of diets and believe the findings in this study to be of general interest. It is surprising that high maternal education is associated with increased risk of unwarranted diets because a high educational level is otherwise associated with a healthy behavior pattern. The finding may indicate that there is lack of adequate and reliable information concerning adverse reaction to food available to the parents. Furthermore, the finding that only half of the parents consult a doctor when their child is perceived as reacting to food could indicate that the public health system lacks credibility in handling these issues. Attention must be directed toward discouraging parents from restricting their children’s diets unnecessarily, and simultaneous efforts should also be made to ensure that children in need of restricted diets receive these and that they are nutritionally adequate. This can only be achieved by ensuring that perceived reactions to food are objectively confirmed by food challenges, preceding any diet alterations, and that restricted diets are medically supervised.
REFERENCES 1. Henriksen C, Eggesbø M, Halvorsen R, Botten G. Nutrient intake among two-year-old children on cows’ milkrestricted diets. Acta Paediatr 2000;89:272-8. 2. Paganus A, Juntunen Backman K, Savilahti E. Follow-up of nutritional status and dietary survey in children with cow’s milk allergy. Acta Paediatr 1992;81:518-21.
3. Isolauri E, Sutas Y, Salo MK, Isosomppi, Kaila M. Elimination diet in cow’s milk allergy: risk for impaired growth in young children. J Pediatr 1998;132:1004-9. 4. David TJ, Waddington E, Stanton RH. Nutritional hazards of elimination diets in children with atopic eczema. Arch Dis Child 1984;59:323-5. 5. Devlin J, Stanton RH, David TJ. Calcium intake and cows’ milk free diets. Arch Dis Child 1989;64:1183-4. 6. Tiainen JM, Nuutinen OM, Kalavainen MP. Diet and nutritional status in children with cow’s milk allergy. Eur J Clin Nutr 1995;49:605-12. 7. David TJ. The overworked or fraudulent diagnosis of food allergy and food intolerance in children. J R Soc Med 1985;78(Suppl 5):21-31. 8. Bock SA. Prospective appraisal of complaints of adverse reactions to foods in children during the first 3 years of life. Pediatrics 1987;79:683-8. 9. Eggesbø M, Halvorsen R, Tambs K, Botten G. Prevalence of parentally perceived adverse reactions to food in young children. Pediatr Allergy Immunol 1999;10:122-32. 10. Nafstad P, Jaakkola JJ, Hagen JA, Botten G, Kongerud J. Breastfeeding, maternal smoking and lower respiratory tract infections. Eur Respir J 1996;9:2623-9. 11. Eggesbø M, Botten G, Halvorsen R, Magnus P. The prevalence of CMA/ CMPI in young children: the validity of parentally perceived reactions in a population-based study. Allergy 2001; 56:393-402. 12. Eggesbø M, Botten G, Halvorsen R, Magnus P. The prevalence of allergy to egg. A population-based study in young children. Allergy 2001;56:403-11. 13. Cytel Software Corporation CM0U. LogXact for windows. 1996. 14. Host A, Halken S. A prospective study of cow milk allergy in Danish infants during the first 3 years of life: clinical course in relation to clinical and immunological type of hypersensitivity reaction. Allergy 1990;45:587-96. 15. Rona RJ, Chinn S. Parents’ perceptions of food intolerance in primary school children. Br Med J Clin Res Ed 1987;294:863-6. 16. Strachan DP. Allergy and family size: a riddle worth solving [editorial]. Clin Exp Allergy 1997;27:235-6.
587
Restricted diets in children with reactions to milk and egg perceived by their parents Merete Eggesbø, MD, Grete Botten, MD, PhD, and Hein Stigum, PhD The aim of this study was to investigate the degree to which parents alter the diets of their children on the basis of perceived reactions. From a population-based sample of 2979 2-year old children with reactions to egg or milk perceived by their parents, one third had strict limitations on the intakes of these foods, representing 2.5% of the children in the cohort. In approximately 1 of 6 families the strict diets were initiated without consulting a doctor, and in a substantial proportion the restrictions were unwarranted. High maternal education level and irritability attributed to food were among the risk factors for unwarranted diets. On the other hand, many children, in whom an adverse reaction was verified, lacked appropriate diet restrictions. We conclude that the handling of adverse reactions to food frequently occurs outside the medical care system at the cost of correct diagnosis and appropriate diets. (J Pediatr 2001;139:583-7)
Milk and eggs are important sources of nutrients in childhood, and if they are removed from the diet because of adverse reactions, the child’s nutritional needs may not be met.1-6 Nutritional counseling of the parents and prescription of adequate substitutes for the child may prevent inadequate elimination diets.5,6 Parents may, however, alter the diets on their own initiative, on the basis of perceived reactions, or following advice from a paramedical group. Anec-
dotal reports have been published on children with severe malnutrition resulting from strict elimination diets imposed on them by their parents or paramedical groups.7 Because parents generally tend to overestimate the role of food as the cause of symptoms in their children, these diets are likely to be unwarranted as well. A substantial number of parents perceive that their children have adverse reactions to milk and egg, 11% to 15% and 2% to 4%, respectively, by age 2.8,9
From the National Institute of Public Health, Department of Epidemiology, and the Center for Health Administration, The National Hospital, University of Oslo, Norway.
Funded by grants from the Norwegian Research Council. Submitted for publication July 7, 2000; revision received Feb 20, 2001; accepted May 1, 2001. Reprint requests: Merete Eggesbø, Section of Epidemiology, Department of Population Health Sciences, National Institute of Public Health, Post Box 44 04 Torshov, N-0403 Oslo, Norway. Copyright © 2001 by Mosby, Inc. 0022-3476/2001/$35.00 + 0 9/22/117067 doi:10.1067/mpd.2001.117067
The potential for unguided elimination diets may thus be substantial. However, the degree to which parents actually alter the diet of their children on the basis of perceived reactions to food remains unknown. The aim of this study was to study the occurrence of milkand egg-restricted diets in a populationbased sample of children with perceived reactions to these foods at age 2 and to study the relationship between diet and a verified diagnosis.
METHODS The study population included children born consecutively in Oslo, Norway, in 1992 to 1993. They were recruited at the maternity ward and studied prospectively with postal questionnaires addressed to the parents every 6 months until the child reached age 2; 2979 families (82.2%) responded to the questionnaires when the child was 2 years old. Details of the study design and collection of data have been described elsewhere.9,10 RR OR
Relative risk Odds ratio
Data Collection Information regarding the occurrence and type of any reaction to food was obtained at 12, 18, and 24 months,9 and information regarding type of diet was obtained at age 24 months. Of the 206 families who reported a reaction to milk
583
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS OCTOBER 2001
Table I. Children with strict diet alterations that were unwarranted, among 46 children with parentally perceived reactions to egg or milk which could not be reproduced, according to attributed symptom and socioeconomic factors
Children on an unwarranted strict diet n
n
%
Yes No Yes No Yes No Yes No Yes No
46 18 28 30 16 6 40 9 37 5 41
17 6 11 13 4 6 11 5 12 5 12
37 33.3 39.3 43.3 25 100 27.5 55.6 32.4 100 29.3
Socioeconomic factors High maternal Yes No education† Low maternal age‡ Yes No Yes Nonsmoker§ No Older siblings Yes No
24 22 23 23 38 8 28 18
12 5 9 8 15 2 15 2
50 22.7 39.1 34.8 39.5 25 53.6 11.1
All symptoms Skin Gastrointestinal Vomiting Stomachache Irritability
Adjusted P value*
RR
CI
0.8
(0.4-1.9)
—
1.7
(0.7-4.4)
—
3.6
(2.2-6.0) .01
1.7
(0.8-3.6)
3.4
(2.1-5.5) .03
2.2
(0.9-5.2) .04
1.1
(0.5-2.4) .02
1.6
(0.4-5.6)
4.8
(1.2-18.6) .001
—
—
Two separate regression analyses were performed; the first included symptoms and the second included socioeconomic factors. High income was also analyzed and showed no association with type of diet (data not shown). All variables that reached a significance level of 0.15 (Pearson χ2 test) in the bivariate analysis were included in the multivariate analysis in addition to maternal age, which was included as a confounder. *Significance tested by log exact test. †Maternal education was categorized into 12 years or less and more than 12 years of education. ‡Maternal age was categorized into 30 years or younger and older than 30 years. §Maternal smoking during pregnancy.
or egg in their child at age 2, 138 families were asked to participate in an examination to confirm the reported reaction; 86 families (62%) completed the examination at a mean age of 2.5 years (CI 2.5 to 2.6).11,12 The perceived reaction to egg or milk was considered confirmed when one of the three following criteria were fulfilled: a history of an immediate reaction and corresponding high levels of specific IgE; an objective reaction to an open food challenge; or a reaction verified by double-blind placebo-controlled food challenges.11,12 If the diagnostic criteria were partly fulfilled, the perceived reaction was assessed as “possible.” Children in whom a current adverse reaction 584
was refuted but who had a convincing history of an earlier allergy were classified as “possible earlier allergic.”
Variables MILK- OR EGG-RESTRICTED DIET. The main question, on which the operational definition of the outcome was based, was whether the perceived reaction to food had resulted in removal of any food items or liquids from the child’s diet, in which case the alterations were required to be specified. The diets were categorized into 3 groups: (1) Strict diets: the main product and all or most byproducts had been eliminated; (2) Moderate
diets: only the main product and not the byproducts had been removed; (3) the option “No diet alterations” was marked or the alteration was insubstantial. In some cases when merely “milk” or “eggs” had been stated, information concerning the amount of food needed to elicit a reaction was used to determine whether all products or only the main product was being referred to. The diet was classified as “currently restricted” if it was stated that the food had not been reintroduced again or it had been reintroduced again but had caused the symptoms to recur. BACKGROUND VARIABLES. Information was obtained on whether a physician had been consulted because of the perceived food reaction and whether, because of illnesses of the child, the parents had sought treatment by others than doctors, such as homeopaths and chiropractors. Information on whether the infant was the mother’s first-born child, maternal age in years, maternal smoking habits at the time of the birth, maternal education in years, and family income was obtained from the questionnaire completed by the participating families at the maternity ward.10 The Norwegian Data Inspectorate and the Regional Ethics Committee for Medical Research approved all stages of the study.
Statistical Analysis The main outcome variable was “strict diets” with the categories “yes” and “no.” Two outcomes on inappropriate diets were derived: strict diet among children with perceived reactions that are disproved (Table I) and moderate or no diet, which were combined to comprise the outcome variable “lack of strict diet” among children in whom a food reaction is confirmed (Table II). The normal approximation to the binomial distribution was used for the computation of CIs. All variables that reached a significance level of 0.15, as assessed by means of the Pearson χ2 test in the bivariate analysis, were included in the
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS
VOLUME 139, NUMBER 4 multivariate models. In addition, maternal education and maternal age were always included in the model as potential confounders. Adjusted associations were obtained by multivariate logistic regression. Because of the small numbers, the inappropriate diets were analyzed by exact logistic regression, with a separate analysis for two sets of covariates, attributed symptom and socioeconomic factors. In these analyses the outcome is not rare, and thus the odds ratios are inflated compared with relative risks. Also there are empty cells for some of the categories in Tables I and II. This situation is handled by the exact regression, but we found the estimated odds ratios difficult to interpret. We have therefore chosen not to present them but to show only the significance levels in the adjusted analyses. The analysis were performed with SPSS (Release 6.1; SPSS Inc, Chicago, Ill) and the LogExact software (Cytel Software Corp, Cambridge, Mass).13
RESULTS A prerequisite for initiating a strict diet is to have perceived a food reaction. Thus socioeconomic factors work at two levels: on the parental perception and then on the choice of diet.
Perceived Reactions to Egg and Milk There was no difference in the distribution of maternal education level, maternal age, or maternal smoking habits among parents who perceived a reaction to food compared with those who did not. The proportions of low-income families and children with older siblings were slightly higher among those with perceived reactions to egg or milk than among those without perceived reactions.
The Prevalence of Restricted Diets The cumulative incidence of eggand milk-restricted diets are shown in
Table II. Children without strict diet alterations among 40 children with confirmed reactions to egg or milk, according to attributed symptom and socioeconomic factors
Children lacking indicated restrictions n
n
%
Yes No Yes No Yes No Yes No Yes No
40 30 10 17 23 4 36 3 37 7 33
17 10 7 7 10 1 16 0 17 1 16
42.5 33.3 70 41.2 43.5 25 44.4 0 45.9 14.3 48.5
Socioeconomic factors High maternal Yes No education‡ High Yes No maternal age§ Yes Smoker No
24 16 20 20 7 33
9 8 6 11 5 12
37.5 50 30 55 71.4 36.4
All symptoms Skin Gastrointestinal Vomiting Stomachache† Irritability
RR
CI
Adjusted P value*
0.5
(0.2-0.9)
.07
0.9
(0.5-2.0)
—
0.6
(0.1-3.2)
— 1
0.3
(0.0-1.9)
.7
0.8
(0.4-1.5)
.86
0.5
(0.3-1.2)
.39
2
(1.0-3.8)
.32
Two separate regression analyses were performed, the first included symptoms and the second included socioeconomic factors. High income and older siblings were also analyzed but showed no association with type of diet (data not shown). All variables that reached a significance level of 0.15 (Pearson χ2 test) in the bivariate analysis, were included in the multivariate analysis. Maternal age and education were included as potential confounders. *Significance tested by log exact test. †RR estimates for lack of diet cannot be estimated because of empty cells. ‡Maternal education categorized into 12 years or less and more than 12 years of education. §Maternal age categorized into 30 years or younger and older than 30 years. Maternal smoking during pregnancy.
the Figure. Of the 206 children with parentally perceived reactions to egg or milk at age 2, strict diets had been initiated in 73 children, and both egg and milk had been restricted in 17 of these children. The cumulative incidence of strict milk- or egg-restricted diets by age 2 is thus estimated at 2.5% (CI, 1.9 to 3.0). At age 2 years, 66 children were still on strict diets, giving an estimated point prevalence of 2.1% (CI, 1.6 to 2.6). Among the children on a strict milk-restricted diet, 32.1% and 5.7% had additional restrictions on the intake of egg and fish, respectively. Among the children on a strict egg-restricted diet, the corresponding figures were 45.9% and 21.6%, for additional restrictions on
the intakes of milk and fish, respectively.
Consultations Overall 54.4% had consulted a physician concerning the perceived adverse reaction, independent of the type of food, increasing to 80.6% where reactions were attributed to both egg and milk or 84.9% where strict diet alterations were reported. Among the parents of children with perceived reactions to egg or milk, 21.7% had sought treatment by paramedical groups, as opposed to 11% in families with children without reported reactions to any food (relative risk [RR] = 2; CI, 1.5 to 2.7). Information was not available on 585
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS OCTOBER 2001
Figure. Cumulative incidence of strict and moderate diet restrictions among 206 children with parentally perceived reactions to egg and milk at age 2, in a population-based sample of 2979 children, according to incriminated food.
which illnesses motivated the consultations.
Children Who Had Received Unwarranted Strict Diets Among the 86 children who had been examined with objective diagnostic methods by the researcher, the diagnosis was disproved in 46 children. Among these children 17 had received strict diets (Table I). However, 8 of these 17 children were suspected of having recently outgrown their food reactions, in which case the restricted diets may not have been truly unwarranted. When these 8 children were excluded, a minimum of 20% unwarranted strict diets is estimated. Unwarranted diet restrictions were significantly more common among children in whom irritability and vomiting were attributed to the food compared with other symptoms (Table I). Furthermore, having older siblings, low maternal age, and high maternal education level were significantly associated with unwarranted diets (Table I).
Children with Confirmed Reactions and Lack of Strict Diets In 40 of the 86 examined children the perceived reaction was either confirmed or assessed as possible. Seventeen of these 40 children lacked strict diet restrictions (Table II). The factors associ586
ated with lack of diet restrictions in children in whom it was indicated are shown in Table II.
DISCUSSION This study shows that among children with parentally perceived reactions to egg or milk, one third had had strict limitations on the intake of these foods. Strict alterations were unwarranted in a minimum 1 of 5 children, and high maternal education was a risk factor for unwarranted diets. On the other hand, more than one third of children with confirmed adverse reaction to egg or milk lacked appropriate diet restrictions. The children were not evaluated at the time of the initial report of the perceived reaction, thus allowing for spontaneous recovery. This may result in an overestimation of children on unwarranted diets. When children with a possible earlier allergy were excluded from the multivariate analysis on risk factors associated with unwarranted diets, similar results were obtained. Thus we believe that this misclassification does not affect the validity in predicting risk factors apart from possibly underestimating the effects of the risk factors. The information obtained from the parents was not always sufficient for
certain assessment of the strictness of the elimination. However, these misclassifications are assumed to be nondifferential and should not affect the validity of the results. The cumulative incidence of restricted diets by age 2 is an underestimate because we did not have information on the diets of children in whom a reaction to milk or egg had resolved before age 2.9 The point prevalence estimate, however, is not affected by this uncertainty. The groups were small, and, hence, it is difficult to draw firm conclusions from the study. The sample size is small, making the study prone to type II errors, that is, not detecting factors that may be of importance. On the other hand, the observed risk factors are likely to be of major importance because statistical significance was obtained in spite of the small numbers. Although it is well documented that adverse reaction to food is overreported,8,11,12,14 which factors influence the parental perception of adverse reaction to food is less well understood. The association between low family income and perceived reactions may be explained by an actual difference in the underlying prevalence or a report bias. Maternal education level is generally considered the best measure of socioeconomic class in Norway, and a report bias tied to socioeconomic class would be expected to include maternal education. However, this was not found. In contrast to our results, a study from England reported a strong association between a high maternal education level and the child being perceived as food intolerant.15 Furthermore, children with older siblings are at increased risk of reported adverse reactions. This is opposed to the consistently reported findings of an inverse association between various manifestations of allergy and sibship size,16 but it may be a report bias resulting from increased parental awareness of adverse reactions in subsequent children. In the previous study from England, 67% of parents who perceived a food
EGGESBØ, BOTTEN, AND STIGUM
THE JOURNAL OF PEDIATRICS
VOLUME 139, NUMBER 4 reaction in their children had consulted a doctor,15 which is in keeping with this study. The restricted diets were the parent’s decision in more than half of those who had consulted a doctor.15 In this study we lack information on whether the diet restrictions actually were initiated following doctor’s advice. By age 2, 1.8% and 1.2% of the children had had major restrictions on their dietary content of milk and egg, respectively. This is similar to the findings of the English study in which the corresponding numbers were 1.1% and 0.9%, respectively; although the studies are not completely comparable because of differences in design and the operational definition of restricted diet.15 The high number of children on milkrestricted diets is of great concern. The potential hazards of milk-restricted diets in which the child’s nutritional demands may not be met are well documented.1-5 Although nutritional counseling of the parents and prescription of adequate nutritional substitutes may prevent inadequate diets,5,6 our results indicate that approximately 1 of 6 families initiate strict diets without seeking medical advice. Furthermore, a sufficient diet will depend on whether the doctor actually initiated the diet, a milk-substitute was prescribed and nutritional counseling was given, and whether the child was followed up over time. Diets solely restricted in eggs are less prone to lead to nutritional inadequacy1; however, the egg restrictions were combined with milk restrictions in nearly half of the children, in which case there is increased risk of protein deficiency. Although a less voiced concern, the lack of strict diets in some children with an actual allergy deserves attention too. The scientific basis for advice is generally sparse, and a lack of knowledge among physicians may leave parents to venture into their own decision-making on the strictness of the diet.
Although there may be cultural differences in the perception and parental handling of food reactions, studies so far have shown remarkable similarity between countries. This also applies to the phenomenon of parental overreporting of food reactions compared with what can be verified by objective methods. We are not aware of earlier studies that have actually studied the appropriateness of diets and believe the findings in this study to be of general interest. It is surprising that high maternal education is associated with increased risk of unwarranted diets because a high educational level is otherwise associated with a healthy behavior pattern. The finding may indicate that there is lack of adequate and reliable information concerning adverse reaction to food available to the parents. Furthermore, the finding that only half of the parents consult a doctor when their child is perceived as reacting to food could indicate that the public health system lacks credibility in handling these issues. Attention must be directed toward discouraging parents from restricting their children’s diets unnecessarily, and simultaneous efforts should also be made to ensure that children in need of restricted diets receive these and that they are nutritionally adequate. This can only be achieved by ensuring that perceived reactions to food are objectively confirmed by food challenges, preceding any diet alterations, and that restricted diets are medically supervised.
REFERENCES 1. Henriksen C, Eggesbø M, Halvorsen R, Botten G. Nutrient intake among two-year-old children on cows’ milkrestricted diets. Acta Paediatr 2000;89:272-8. 2. Paganus A, Juntunen Backman K, Savilahti E. Follow-up of nutritional status and dietary survey in children with cow’s milk allergy. Acta Paediatr 1992;81:518-21.
3. Isolauri E, Sutas Y, Salo MK, Isosomppi, Kaila M. Elimination diet in cow’s milk allergy: risk for impaired growth in young children. J Pediatr 1998;132:1004-9. 4. David TJ, Waddington E, Stanton RH. Nutritional hazards of elimination diets in children with atopic eczema. Arch Dis Child 1984;59:323-5. 5. Devlin J, Stanton RH, David TJ. Calcium intake and cows’ milk free diets. Arch Dis Child 1989;64:1183-4. 6. Tiainen JM, Nuutinen OM, Kalavainen MP. Diet and nutritional status in children with cow’s milk allergy. Eur J Clin Nutr 1995;49:605-12. 7. David TJ. The overworked or fraudulent diagnosis of food allergy and food intolerance in children. J R Soc Med 1985;78(Suppl 5):21-31. 8. Bock SA. Prospective appraisal of complaints of adverse reactions to foods in children during the first 3 years of life. Pediatrics 1987;79:683-8. 9. Eggesbø M, Halvorsen R, Tambs K, Botten G. Prevalence of parentally perceived adverse reactions to food in young children. Pediatr Allergy Immunol 1999;10:122-32. 10. Nafstad P, Jaakkola JJ, Hagen JA, Botten G, Kongerud J. Breastfeeding, maternal smoking and lower respiratory tract infections. Eur Respir J 1996;9:2623-9. 11. Eggesbø M, Botten G, Halvorsen R, Magnus P. The prevalence of CMA/ CMPI in young children: the validity of parentally perceived reactions in a population-based study. Allergy 2001; 56:393-402. 12. Eggesbø M, Botten G, Halvorsen R, Magnus P. The prevalence of allergy to egg. A population-based study in young children. Allergy 2001;56:403-11. 13. Cytel Software Corporation CM0U. LogXact for windows. 1996. 14. Host A, Halken S. A prospective study of cow milk allergy in Danish infants during the first 3 years of life: clinical course in relation to clinical and immunological type of hypersensitivity reaction. Allergy 1990;45:587-96. 15. Rona RJ, Chinn S. Parents’ perceptions of food intolerance in primary school children. Br Med J Clin Res Ed 1987;294:863-6. 16. Strachan DP. Allergy and family size: a riddle worth solving [editorial]. Clin Exp Allergy 1997;27:235-6.
587