Late introduction of fish and eggs is associated with increased risk of allergy development – results from the FARMFLORA birth cohort

Figures & data

Figure 1. Flow chart of subjects eligible for analysis between healthy children and children allergic at 3 years of age. One farm and one non-farm family withdrew from the study due to personal reasons or change of residence. One farm child was not diagnosed at 3 years of age and two farm children and five non-farm children were diagnosed as allergic at 1.5 years of age but not at 3 years of age (transient allergy); hence, these children were neither included in the healthy nor the allergic group.

Figure 2. Cumulative rate of healthy and allergic children for whom formula or food have been introduced at different time-points. Numbers were calculated by χ² tests. Analyses were carried out for introduction of formula or any food (a), any food (b), fish (c), eggs (d) and flour (e). Significant differences between healthy and subsequently allergic children are denoted as stars of significance: *P ≤ 0.05, **P ≤ 0.01.

Table 1. Characteristics of farm versus non-farm children and healthy versus subsequently allergic children.

Variables	Farmers (n = 28)	Non-farmers (n = 37)	P^e	Healthy (n = 44)	Allergic (n = 11)	P^e
Antenatal characteristics
Heredity^a
Mothers	7 (25%)	11 (30%)	0.68	9 (21%)	5 (46%)	0.12
Fathers	1 (4%)	12 (32%)	0.01	7 (16%)	3 (27%)	0.40
Maternal age at delivery, year	33 (21–42)	32 (22–41)	0.46^f	32 (21–42)	34 (22–41)	0.71^f
Education, level^b (1 = lowest, 5 = highest)
Mothers	2 (1–5)	4 (1–5)	0.20^f	4 (1–5)	3 (1–5)	0.78^f
Fathers	2 (1–5)	2 (1–5)	0.02^f	2 (1–5)	2 (2–4)	0.51^f
Smoking during last mo of pregnancy
Mothers	0 (0%)	1 (3%)	1.00	1 (2%)	0 (0%)	1.00
Fathers	1 (4%)	4 (11%)	0.38	2 (5%)	2 (18%)	0.18
Cats or dogs in house at recruitment	21 (75%)	19 (51%)	0.054	31 (71%)	5 (46%)	0.16
Siblings	18 (64%)	17 (46%)	0.15	23 (52%)	6 (55%)	0.89
Birth characteristics
Gestational week^c	40 (37–42)	39 (36–42)	0.13^f	39 (36–42)	39 (36–41)	0.14^f
Cesarean section	3 (11%)	7 (19%)	0.50	5 (11%)	4 (36%)	0.07
Birth weight^d, g	3495 (2780–4740)	3618 (2440–4830)	0.78^f	3560 (2440–4740)	3235 (2730–4830)	0.14^f
Infant characteristics
Male gender	10 (36%)	23 (62%)	0.04	22 (50%)	9 (82%)	0.09
Maternal fish oil intake, 4 months postpartum	2 (11%)	0 (0%)	0.18	2 (7%)	0 (0%)	1.00
Vitamin A and D supplements at age 1 year	17 (61%)	22 (60%)	0.92	30 (68%)	6 (55%)	0.49
Allergic at 3 year of age	1 (4%)	10 (32%)	0.02
Farm group				23 (52%)	1 (9%)	0.02

Data are presented as n (%) or medians (ranges). A lower total number of subjects are shown for the Healthy/Allergic group than the Farm/Non-farm group due to drop-outs and allergy selection (Figure 1).

^aDoctor’s diagnosed asthma, rhinitis, or atopic eczema.

^b1 = Elementary school, 2 = upper secondary school 2–3 years or equivalent, 3 = qualified graduate from upper secondary engineering course, 4 = university ≤1 year, 5 = university >1 year.

^cn = 27 farmers and 36 non-farmers, and n = 43 healthy and 10 allergic subjects, respctively.

^dn = 36 non-farmers, and n = 10 allergic subjects.

^eχ²-test (or Fisher’s exact test).

^fMann-Whitney U test.

Table 2. Duration of breastfeeding and introduction of complementary foods in farm and non-farm children.

	Farmers (n = 28)		Non-farmers (n = 37)
	Median	IQR	Median	IQR
Variables	months		months		P
Exclusive breastfeeding	4	3–4	3	1–4	0.19
Any breastfeeding	8	5–11	8	4–12	0.74
Any food group	4	4–5	4	4–4	0.15
Potatoes	4	4–5	4	4–5	0.33
Vegetables	5	4–6	5	4–5	0.30
Legumes	7	6–8	6	5–8	0.59
Fruit	5	4–6	5	4–5	0.97
Berries	6	5–8	7	6–9	0.25
Peanuts	19	13–19	19	18–19	0.07
Nuts	11	8–16	15	12–19	0.02
Meat	6	5–6	6	5–6	0.71
Eggs	8	6–10	8	6–9	0.97
Fish	6	5–8	6	5–7	0.85
Flour	5	4–6	5	5–6	0.86
Milk	5	4–5	5	4–6	0.68
	Number of food items introduced
Food diversity at age 6 months	8	6–9	8	6–9	0.88

Table 3. Allergy diagnoses at 3 years of age.

	Eczema	Asthma	Food allergy	ARC^a
Farm children
1	x
Non-farm children
1	x
2	x
3	x
4	x
5	x		x
6	x	x
7		x
8		x
9		x		x
10			x

Diagnoses were made by trained pediatricians, using strict predefined protocols.

^aAllergic rhinoconjunctivitis.

Table 4. Duration of breastfeeding and introduction of complementary foods in healthy and subsequently allergic children.

	Healthy (n = 44)		Allergic (n = 11)
	Median	IQR	Median	IQR
Variables	months		months		P
Exclusive breastfeeding	4	2–4	1	0–3	0.01
Any breastfeeding	8	5–12	6	2–7	0.02
Any food group	4	4–4	4	4–5	0.79
Potatoes	4	4–5	4	4–5	0.98
Vegetables	5	4–6	5	4–6	0.83
Legumes	7	6–8	6	5–7	0.16
Fruit	5	4–5	4	4–5	0.34
Berries	6	5–8	9	5–10	0.45
Peanuts	19	15–19	19	17–19	0.66
Nuts	12	11–18	19	11–19	0.13
Meat	6	5–6	5	5–6	0.37
Eggs	8	6–9	8	5–11	0.57
Fish	6	5–7	6	5–10	0.33
Flour	5	5–6	6	5–6	0.08
Milk	5	4–5	5	4–6	0.29
	Number of food items introduced
Food diversity at age 6 months	8	6–9	8	6–9	0.72

Table 5. Logistic regression of allergy development, exclusive breastfeeding and introduction of flour, fish, and eggs.

Model	Potential covariates	Exclusive breastfeeding^b	P	Flour^c	P	Eggs ≤10 mo^d	P	Fish ≤10 mo^d P^f
1	Crude model	0.59 (0.39–0.89)	0.01	0.64 (0.32–1.24)	0.19	0.08^e (0.13–0.54)	0.01	0.01
2	Exclusive breastfeeding			0.42 (0.18–0.95)	0.04	0.08 (0.01–0.64)	0.02	0.02
3	Maternal heredity	0.59 (0.39–0.89)	0.013	0.58 (0.30–1.20)	0.15	0.08 (0.01–0.55)	0.011	0.004
4	Paternal smoking^a	0.59 (0.39–0.89)	0.013	0.60 (0.30–1.19)	0.15	0.09 (0.01–0.59)	0.013	0.02
5	Cats/dogs in house at recruitment	0.51 (0.31–0.82)	0.01	0.64 (0.32–1.28)	0.21	0.09 (0.01–0.60)	0.013	0.003
6	Gestational week	0.61 (0.39–0.94)	0.02	0.64 (0.31–1.30)	0.22	0.09 (0.01–0.63)	0.02	0.01
7	Cesarean section	0.61 (0.40–0.92)	0.02	0.60 (0.30–1.19)	0.14	0.07 (0.01–0.48)	0.01	0.004
8	Birth weight	0.62 (0.41–0.94)	0.02	0.64 (0.31–1.30)	0.22	0.05 (0.01–0.40)	0.01	0.01
9	Male gender	0.60 (0.39–0.93)	0.02	0.42 (0.18–1.00)	0.051	0.04 (0.003–0.46)	0.01	0.004
10	Farm group	0.61 (0.40–0.95)	0.03	0.36 (0.13–1.01)	0.051	0.10 (0.01–0.80)	0.03	0.001

Analyses were carried out on significant or near significant relationships between allergy and breastfeeding or food-introduction practices. Owing to the low number of allergic subjects, each potential covariate was added in separately in the adjusted models, i.e. only one adjustment variable at a time.

n = 44 healthy and 11 subsequently allergic children.

^aSmoking during the last month of pregnancy.

^bEntered as a continous variable: months of exclusive breastfeeding.

^cEntered as a continous variable: months of consumption within 1.5 year of age (the last time-point for registration of food introduction).

^dEntered as a dicotomous variable: 1 = having eggs/fish introduced ≤10 months of age; 0 = having eggs/fish introduced >10 months.

^eHosmer and Lemeshow Test not applicable.

^fLikelihood ratio test (χ²).