Neonatal and birth mother retrospective chart review
Initial billing records identified 6160 children with ICD-9-CM codes consistent with medical visits related to food allergic reactions between 2000 and 2005 (Figure
). However, only 235 (3.8%) of those children were born at PSHMC, to permit direct examination of birth records. Of those 235 children (n=125 for males; n=110 for females), a thorough review of medical records confirmed food allergies (positive IgE or SPT) in only 82 (35%) children. In an effort to increase the number of food allergic cases, direct referrals from pediatricians were subsequently accepted; as a result, we expanded the age of children who could be considered as “cases” to 18 years of age based on these direct physician referrals. This protocol modification increased the number of food allergic children (by 17) for whom we could evaluate birth records to 99; three of these children were born in 1998 or 1999. Medical records from 192 non-food allergic children born at the same hospital were also examined as controls. Overall, the mean age of children at the time that charts were reviewed was 5.0 ± 1.5 years, with a mean age of 4.8 ± 1.9 years in the allergic cohort and 5.1 ± 1.3 years (p=0.12) in the non-allergic children.
Identification of Allergic Cohort. HMC, Hershey Medical Center; RAST, radioallergosorbent test; SPT, skin prick test.
Of the 291 children (99 cases; 192 controls) included in our analysis, 167 were male and 124 were female (Table
). There was no difference in gestational age for children with allergies versus non-allergic children (38.3 weeks ± 2.39 weeks; range 27 to 41 weeks versus 38.0 weeks ± 2.86 weeks; range 25 to 42 weeks). There was also no difference in birth weight between the two groups (3.36 kg ± 0.67 kg; range 0.81 to 4.61 kg versus 3.24 kg ± 0.76 kg, range 0.66 to 5.07 kg).
Characteristics of food allergic and non-allergic children whose birth records were retrospectively reviewed
Ethnicity data were available for 282 mothers: 87.6% of mothers self-identified as white (non-Hispanic), whereas the remaining women identified as African American (4.3%), Asian (4.6%) or Hispanic (3.5%). For 45% of the mothers, the children represented by these data were firstborn. In all, 14% of mothers reported a personal history of asthma and 4% of mothers reported food allergies. The likelihood of having a child diagnosed with food allergies was slightly higher for women with a history of atopy compared to women that did not report atopic conditions (43% versus 32%), but this did not reach significance (OR, 1.60; 95% CI, 0.861-2.996; p=0.14). Maternal parity was not associated with an increased incidence of food allergies in offspring, neither did mean maternal age at time of delivery differ significantly between mothers whose children developed food allergies (30.68 ±5.06 years) compared to those that did not (28.55 ± 6.48 years). However, an association was identified between the odds of food allergy diagnosis in childhood and maternal age at delivery (OR =1.05; 95% CI, 1.017 to 1.105; p=0.005). For each advancing year of maternal age at time of delivery, the odds of food allergies being diagnosed in the child increased by 6.0%. This relationship was linear for mothers whose age ranged from under 20 years through 35 years of age (Figure
Figure 3 Incidence of food allergies as a function of maternal age at delivery (years). Correlation of food allergy diagnosis in childhood and maternal age at delivery. For each advancing year of maternal age, the odds of food allergies diagnosed in newborn increases (more ...)
In the cohort of 99 children with allergies, the mean age at the time of allergy diagnosis was 1.5 years of age and males were significantly more likely to be diagnosed with a food allergy than females (OR = 1.80; 95% CI, 1.088-2.985; p=0.02). Allergies to peanuts, eggs, milk, and other nuts were the most common, impacting 53%, 49%, 43%, and 18% of the food allergic children, respectively. Forty-five percent of children with food allergies were allergic to just one food item, with the remaining children allergic to two or more foods. When these children were exposed to allergenic food(s), they most often experienced skin eruptions, including dermatitis and hives (69%). Anaphylaxis was uncommon (n=10; 5.2%).
No differences were observed with respect to development of food allergies and method of delivery (allergy diagnosis in 34% delivered vaginally and 33% delivered via cesarean; OR=0.93; 95% CI, 0.557-1.564; p=0.79), postnatal antibiotics (40% who developed allergies received antibiotics versus 33% who did not receive antimicrobials; OR=1.35; 95% CI, 0.680-2.677; p=0.39), intrapartum maternal antibiotic exposure (allergy diagnosis in 34% of children whose mothers did or did not receive intrapartum antibiotics; OR=1.00; 95% CI, 0.587-1.715; p=0.998), or time spent in the NICU (allergy diagnosis in 34% and 35% of children who did and did not immediately move to the well child nursery; OR=1.04; 95% CI, 0.544-1.977; p=0.91). Vaginal group B Streptococcus (GBS) status was unknown in half the women; for those mothers in whom GBS status was known to be positive, an increased likelihood of developing food allergies was not observed (OR=1.14; 95% CI, 0.556-2.348; p=0.72). Most women (79%) expressed intent to breastfeed their infants; however, intent to breastfeed did not vary between mothers with a positive or negative personal history of atopy (84% versus 78%; p=0.21). A positive correlation between maternal intent to breastfeed and subsequent food allergy development was observed (p<0.005) (Table