The natural history of IgE-mediated cows

milk allergy
Justin M. Skripak, MD, Elizabeth C. Matsui, MD, MHS, Kim Mudd, RN,
and Robert A. Wood, MD Baltimore, Md

Background: Cows milk allergy (CMA) is the most common

food allergy in infants and young children, affecting 2% to 3%
of the general population. Most studies have shown the
prognosis of developing tolerance to cows milk to be good, with
most outgrowing their allergy by age 3 years.
Objective: To define the natural course of CMA and identify
the factors that best predict outcome in a large referral
population of children with CMA.
Methods: Clinical history, test results, and final outcome were
collected on 807 patients with IgE-mediated CMA. Patients
were considered tolerant after they passed a challenge or
experienced no reactions in the past 12 months and had a cows
milk IgE (cm-IgE) level <3 kU/L.
Results: Rates of resolution were 19% by age 4 years, 42% by
age 8 years, 64% by age 12 years, and 79% by 16 years.
Patients with persistent allergy had higher cm-IgE levels at all
ages to age 16 years. The highest cm-IgE for each patient,
defined as peak cm-IgE, was found to be highly predictive of
outcome (P < .001). Coexisting asthma (P < .001) and allergic
rhinitis (P < .001) were also significant predictors of outcome.
Conclusion: The prognosis for CMA in this population is worse
than previously reported. However, some patients developed
tolerance during adolescence, indicating that follow-up and reevaluation of CMA patients is important in their care. cm-IgE
level is highly predictive of outcome.
Clinical implications: The increasing potential for persistence
of CMA, along with cm-IgE levels effect on prognosis, should
be considered when counseling families regarding expected
clinical course. (J Allergy Clin Immunol 2007;120:1172-7.)
Key words: Cows milk, food allergy, IgE, prognosis, tolerance
Food allergy, anaphylaxis,
dermatology, and drug allergy

Cows milk allergy (CMA) is the most common food

allergy in infants and young children, affecting 2% to 3%
of the general population.1-4 Most studies have shown the
prognosis of developing tolerance to cows milk to be
good, with the majority outgrowing their allergy by age
3 years.1,4 Other studies have found less optimistic results,
From the Department of Pediatrics, Division of Allergy and Immunology,
Johns Hopkins University School of Medicine.
Supported by National Institutes of Health Training Grant #5T32 AI07007 and
the Eudowood Foundation.
Disclosure of potential conflict of interest: R. A. Wood has consulting arrangements with Dey Pharmaceutical, has received grant support from Merck and
Genentech, and is on the speakers bureau for Dey, Merck, and Glaxo. The
rest of the authors have declared that they have no conflict of interest.
Received for publication May 21, 2007; revised August 8, 2007; accepted for
publication August 8, 2007.
Available online October 12, 2007.
Reprint requests: Robert A. Wood, MD, CMSC 1102, Johns Hopkins Hospital,
600 North Wolfe St, Baltimore, MD 21287. E-mail: rwood@jhmi.edu.
0091-6749/$32.00
2007 American Academy of Allergy, Asthma & Immunology
doi:10.1016/j.jaci.2007.08.023

1172

Abbreviations used
CMA: Cows milk allergy
cm-IgE: Cows milk IgE

however, and the prognosis for developing tolerance in

older children with persistent CMA remains less clear.5-8
Saarinen et al6 found 15% of children with previously diagnosed IgE-mediated CMA to have persistent sensitivity
at age 8.6 years. There is limited information available regarding the clinical or laboratory factors that may predict
the development of tolerance to cows milk, although children with nonIgE-mediated disease have consistently
been shown to develop tolerance earlier and more frequently than those with IgE-mediated allergy.
In our clinic population, we have followed a large group
of children with CMA. The purposes of this study were to
define the clinical characteristics of this population, define
the rate of allergy resolution over time, and identify the
clinical and laboratory features that may predict the
outcome of CMA over time.

METHODS
This is a retrospective review of the clinical records of 4117
patients seen by the principal investigator (R.A.W.) at 2 pediatric
allergy clinics, 1 private and 1 university-based, between 1993 and
present. There were 1368 with food allergy, of whom 1073 were
diagnosed with milk allergy. Two hundred thirteen patients with milk
allergy were not included in the analysis because they were only seen
once and the visit was before 2004, making the likelihood of at least
1 follow-up unlikely, and an additional 53 patients with only non
IgE-mediated disease were excluded. There were 807 patients with
IgE-mediated CMA on whom data were collected. Data collected
included sex, other food allergies, other atopic conditions, dietary
history, family history of atopy, age at onset of symptoms, symptoms
associated with exposure, age and symptoms with accidental exposures to milk, results of previous skin tests, cows milkspecific IgE
levels (cm-IgE), food challenge results, the reported outcomes of
home milk introductions, and the outcome of other food allergies.
Patients with milk allergy who are followed in our clinic typically
have food-specific IgE levels checked annually using the Phadia CAP
System FEIA (Phadia, Uppsala, Sweden). The diagnosis of CMA was
made on the basis of a history of symptoms clearly associated with
exposure to milk, a positive oral food challenge, and/or a clear
improvement in eczema or other symptoms with milk avoidance.
IgE-mediated disease was defined as having a skin prick test with a
wheal diameter 3 mm and/or a cm-IgE 0.35 kU/L.
The diagnosis of asthma, eczema, or allergic rhinitis was made by
the investigator. These data were collected on all patients from their
initial visit and then updated from their last visit available. Allergy to

Skripak et al 1173

J ALLERGY CLIN IMMUNOL

VOLUME 120, NUMBER 5

Statistical analysis
All analyses were performed with StataSE 8.0 (College Station,
Tex). cm-IgE levels were recorded as <0.35 kU/L, >100 kU/L, or, if
between these values, the specific value was recorded. For purposes
of statistical analysis, each result of <0.35 kU/L was assigned a value
of the lower limit, or 0.18 kU/L, and each value of >100 kU/L was
assigned a value of 101 kU/L. The highest cm-IgE recorded for each
patient was considered the peak level. Peak cm-IgE levels were
stratified into 6 categories: <2 kU/L, 2 to 4.9, 5 to 9.9, 10 to 19.9, 20
to 49.9, and 50. The log-rank test was used to compare clinical
characteristics in resolved versus persistent CMA. Kaplan-Meier
curves were generated to depict the development of cows milk
tolerance over time. Data for all subjects were censored for KaplanMeier analysis. Cox proportional hazards regression was used to
model relationships between cm-IgE levels and oral tolerance. To
meet assumptions of parallel hazards, 3 strata of cm-IgE levels were
used instead of the 6 strata (<5, 5-19.9, and 201 kU/L). In addition,
other predictors of the development of tolerance (such as other atopic
disease) in bivariate analyses were included in the final multivariate
model if they also met assumptions of parallel hazards. Nonparametric
smoothing was used to depict trends in cm-IgE levels over time
using the lowess command with a bandwidth of 0.8. Multiple cm-IgE
levels from any given patient are depicted in either Fig 2, A, or B,
depending on their final CMA status. These repeated measures are
depicted graphically and the trend line plotted, but results of

TABLE I. Criteria for clinical tolerance

Clinical tolerance
definition

1
2*

Criteria

Pass office challenge or home introduction

Pass office challenge or home introduction
OR
cm-IgE <3 kU/L and no history of clinical
reactivity in previous 12 mo
Pass office challenge or home introduction
OR
cm-IgE <15 kU/L and no history of clinical
reactivity in previous 12 mo

*Outcome definition used for all analyses unless noted.

longitudinal analysis techniques accounting for repeated measures

are not presented. A 2-tailed P value <.05 was considered statistically
significant.

RESULTS
Study population
Eight hundred seven patients with IgE-mediated CMA
were included (Table II). There was a 2:1 male:female ratio, with age at the initial visit ranging from 1 month to 209
months (median, 13 months). The median duration of follow-up was 54 months, and the median number of visits
was 5. Other atopic conditions were common (49% had
asthma, 40% had allergic rhinitis, and 71% had eczema
by the time of their most recent follow-up visit). Most
patients (91%) had at least 1 other food allergy; egg and
peanut were most common, followed by tree nuts, soy,
wheat, shellfish, sesame, beef, and fish.
The most common presenting symptoms of milk
allergy were skin-related reactions (85%), including urticaria, angioedema, eczema, or other unspecified rash.
Gastrointestinal symptoms, including vomiting, diarrhea,
bloody stools, and/or gastroesophageal reflux, occurred in
46%, lower respiratory symptoms (wheezing, cough,
stridor, or difficulty breathing) occurred in 14%, upper
respiratory symptoms (rhinitis or nasal congestion) occurred in 6%, and poor growth (weight <5th percentile for
age) occurred in 6%. The median cm-IgE level at the initial
visit was 7.2 kU/L (interquartile range [IQR], 1.8-31.9),
and the median peak cm-IgE was 13.1 kU/L (IQR, 2.8-59).
Breast-feeding history was recorded for 80% of patients, and formula history was recorded for 81%. Of
children whose formula history was recorded, about half
(51%) received a cows milk formula in infancy, 53% a
soy formula, 40% an extensively hydrolyzed formula, and
15% an amino acid formula. Among those whose breastfeeding history was known, 86% were breast-fed, and the
median breast-feeding duration was 8 months, with a
range from 1 to 46 months.
CMA resolution
The patients underwent a total of 289 milk challenges,
68 of which occurred at home and 221 of which occurred
in the clinic. There was an overall pass rate of 57%. Sixty-

Food allergy, anaphylaxis,

dermatology, and drug allergy

other foods was defined as having had a clear symptomatic reaction to

the food and/or having had a positive SPT or food-specific IgE level.
The primary outcome of interest was acquisition of oral tolerance.
Oral milk challenges were routinely performed when, in the judgment
of the principal investigator, the patient had at least a 50% chance of
passing the challenge.9 Patients who were not likely to have acquired
tolerance, on the basis of either a history of recent reactions or
elevated cm-IgE levels, typically did not undergo oral challenges,
but continued to be followed.
For this study, several definitions of oral tolerance were used to
estimate a range of incidence rates for oral tolerance. The definitions
of oral tolerance were based on criteria that ranged from most
stringent to least stringent (Table I). During analysis of the data, each
definition was applied to the entire population. Under the most stringent set of criteria (criteria 1), only patients who passed a formal milk
challenge or successfully introduced milk or concentrated milk products at home were considered tolerant. All other patients were considered to have persistent milk allergy. To take into account the fact
that some patients who had not undergone a milk challenge at home
or in the clinic could have acquired tolerance, a second definition
(criteria 2) of tolerance was also used. Under this second definition,
patients who had a cm-IgE level <3 kU/L at their last clinic visit and
had no history of clinical reactivity in the previous 12 months were
considered to be tolerant, along with those who had passed home or
office challenges. Therefore, patients with persistent CMA had either
experienced symptoms after accidental exposure or a milk challenge
in the past 12 months or had a cm-IgE level 3 kU/L. This second
set of criteria for oral tolerance is based on a previous study in the
same clinic population that found that 84% of children with cm-IgE
 3kU/L exhibited clinical reactivity on milk challenge.9 For the third,
and least stringent, set of criteria (criteria 3) for oral tolerance, patients
who had a cm-IgE level <15 kU/L at their last clinic visit and had no
history of clinical reactivity in the previous 12 months were considered
to be tolerant, along with those who had passed home or office challenges. The cutoff point was set at 15 kU/L for this third definition
of oral tolerance because this threshold has been found to have a
95% positive predictive value for symptomatic allergy in a previous
study.10

1174 Skripak et al

J ALLERGY CLIN IMMUNOL

NOVEMBER 2007

TABLE II. Patient characteristics

Characteristic

Total number of patients

Age at initial visit, median (range)
Duration of follow-up, median (range)
Follow-up visits, median (range)
Sex, n (%)
Male
Female
Other atopic conditions, n (%)
Asthma
Allergic rhinitis
Eczema
Any food allergy besides cows milk, n (%)
Egg
Peanut
Tree nut
Soy
Wheat
Other
Initial symptoms of CMA, n (%)
Skin
Gastrointestinal
Lower respiratory
Upper respiratory
Poor Growth
1 System affected
2 Systems affected
31 Systems affected
cm-IgE levels (kU/L), median (IQR)
Initial (n 5 804)
Peak (n 5 804)
Resolved CMA, n (%)
Definition 1*
Definition 2
Definition 3

TABLE III. Incidence of CMA resolution

Number

Criteria for outgrown allergyy

Definition 1

807
13 mo (1-209)
54 mo (4-285)
5 (1-25)
527 (65)
280 (35)

Age
(y)

393
326
572
732
634
592
412
331
290
457

(49)
(40)
(71)
(91)
(79)
(73)
(51)
(41)
(36)
(57)

2
4
6
8
10
12
14
16
18
Total no.
outgrown

687
369
115
47
48
402
285
94

(85)
(46)
(14)
(6)
(6)
(50)
(35)
(12)

7.2 (1.8-31.9)
13.1 (2.8-59)
120 (15)
307 (38)
439 (54)

*Passed office or home milk challenge.

Passed challenge OR cm-IgE <3 kU/L and no history of clinical reaction.
Passed challenge OR cm-IgE <15 kU/L and no history of clinical reaction.

Food allergy, anaphylaxis,

dermatology, and drug allergy

seven percent of clinic challenges were passed, and 24%

of home challenges were passed. Three sets of criteria
were created to define the acquisition of milk tolerance.
Each set of criteria was applied separately to the entire
population (n 5 807). When tolerance was defined using
the most stringent criteria, as passing a milk challenge, we
found that only 5% outgrew their allergy by age 4 years,
21% by age 8 years, 37% by age 12 years, and 55% by age
16 years. When tolerance was defined as passing a
challenge or a cm-IgE <3 kU/L and no reaction in the
past 12 months, the rates of resolution were 19% at age
4 years, 42% by age 8 years, 64% by age 12 years, and
79% by 16 years. When we also considered children with a
cm-IgE <15 kU/L and no reaction in the past 12 months to
be tolerant, we found 26% tolerant by age 4 years, 56% by
age 8 years, 77% by age 12 years, and 88% by 16 years
(Table III; Fig 1).

Predictors of prognosis
There were 2498 cm-IgE levels recorded. Patients with
persistent allergy had higher cm-IgE levels in the first 2

Passed
No. of
food
subjects challenge*

708
473
289
174
105
60
29
18
6
120

<1%
5%
12%
21%
29%
37%
44%
55%
70%

(0-1)
(3-7)
(9-15)
(17-26)
(24-35)
(31-45)
(35-53)
(44-66)
(55- 83)

Definition 2

Definition 3

Passed food
Passed food
challenge
challenge
OR cm-IgE
OR cm-IgE
<3 kU/L AND <15 kU/L AND
no symptoms no symptoms
in 12 mo*
in 12 mo*

6% (4-7)
19% (16-22)
32% (29-37)
42% (39-47)
52% (47-58)
64% (59-70)
71% (65-77)
79% (72-86)
88% (80-94)
307

9% (7-11)
26% (23-29)
44% (40-48)
56% (51-60)
64% (62-71)
77% (72-81)
83% (78-87)
88% (83-92)
93% (89-97)
439

*Incidence rates of acquisition of oral tolerance by age. 95% CIs are

indicated in parentheses.
Initial n 5 807.

years of life than those who developed tolerance (median

19.0 kU/L vs 1.8 kU/L; P < .001), and this difference was
maintained up through age 18 years (Table IV). Trends in
IgE levels also differed, with the group with persistent allergy exhibiting increasing levels over the first 3 to 4 years
of life, followed by a plateau, and then a gradual decrease
up to age 18 years, whereas the resolved group showed
a decrease over the first 1 to 2 years of life, after which
cm-IgE levels remained stable (Fig 2).
We further examined the relationship between the
acquisition of oral tolerance and cm-IgE levels by calculating incidence rates of oral tolerance over time for each
of 6 peak cm-IgE categories (<2, 2-4.9, 5-9.9, 10-19.9, 2049.9, and 50 kU/L; Fig 3). In general, the higher the peak
cm-IgE was, the lower the likelihood of developing tolerance. Children with peak cm-IgE levels less than 5 kU/L
had the best prognosis, children with peak cm-IgE levels
from 5 to 19.9 kU/L had a somewhat worse prognosis,
and children with peak cm-IgE levels of 20 kU/L or greater
had the worst prognosis. For example, by 4 years, 57%
with a peak cm-IgE <2 kU/L, 37% with a peak level of
2 to 4.9 kU/L, 20% with a peak level of 5 to 9.9 kU/L,
8% with a peak level of 10 to 19.9 kU/L, 3% with a
peak level of 20 to 49.9 kU/L, and <1% with a peak level
of 50 kU/L or greater had become tolerant. By 10 years,
87% with a peak cm-IgE <2 kU/L and 5% with a peak
cm-IgE of 50 kU/L or greater had become tolerant.
Other predictors of developing oral tolerance included
asthma, allergic rhinitis, and a history of being fed
formula (Table V). Eczema, other food allergies, sex,
and breast-feeding were not predictive of acquisition of
oral tolerance. After adjusting for asthma and rhinitis in
a proportional hazards model, peak cm-IgE remained a
strong predictor of acquiring tolerance. To meet assumptions of proportional risk required for this statistical

Skripak et al 1175

J ALLERGY CLIN IMMUNOL

VOLUME 120, NUMBER 5

TABLE IV. cm-IgE levels in patients with persistent vs

resolved CMA
Age range
(mo)

0-23
24-47
48-71
72-95
96-119
120-143
144-167
168-216

model, peak cm-IgE was stratified into 3 categories (<5

kU/L, 5-19.9 kU/L, and >20 kU/L), and each step up
in peak cm-IgE category was associated with a 68%
reduction in the likelihood of acquiring oral tolerance
(hazard ratio [95% CI], 0.32 [.27-.38]).

DISCUSSION
In this referral population of children with milk allergy,
the prognosis for developing tolerance is worse than
previously estimated. Using 3 sets of increasingly broad
criteria to define tolerance, incidence rates of tolerance at 4
years ranged from <1% to 26% in our study, substantially
lower than previously reported. Our findings stand in
marked contrast to the study that is most often quoted,
which found that 75% of children with IgE-mediated milk
allergy were tolerant by the age of 3 years.4 One positive
finding is that patients did continue to achieve tolerance
well into adolescence. This contradicts previous data suggesting that CMA is unlikely to be lost if it has persisted
into their school-age years5,6 and clearly indicates that
there is no age at which outgrowing CMA is impossible.
In addition, we found that the peak cm-IgE level was an
important predictor of acquisition of oral tolerance and
could be used to provide important information for patients and their families regarding their long-term prognosis of CMA.
Defining tolerance as passing a challenge or having a
cm-IgE <3 kU/L and no reactions over the previous year,
we found rates of tolerance of 19% by age 4 years, 42% by
age 8 years, 64% by age 12 years, and 79% by age 16
years. This definition of tolerance is arguably the most
accurate of the 3 definitions used in this study, and the
incidence rates using this definition are most likely closest
to this populations true incidence of tolerance. Although
using the most liberal tolerance definition that relies on a
cm-IgE cutoff of 15 kU/L will minimize the number of
children falsely classified as persistently allergic, it almost

19.0
25.5
25.7
21.5
22.8
28.6
17.4
10.2

(420)
(506)
(340)
(172)
(95)
(59)
(37)
(18)

Resolved
Median (kU/L)
(number)

1.8
1.9
2.0
2.5
2.2
2.1
1.6
1.4

(255)
(259)
(154)
(70)
(56)
(36)
(11)
(6)

P value*

<.001
<.001
<.001
<.001
<.001
<.001
<.001
.002

*Mann-Whitney U test.

certainly overestimates the number developing tolerance,

because a majority of the children with cm-IgE levels
between 3 and 15 kU/L are still allergic.9,10 On the other
hand, the most stringent criteria for tolerance, passing a
milk challenge, will underestimate the number of children
developing tolerance because many children who have not
passed a milk challenge may actually be tolerant. No matter which definition of tolerance is used, it is clear that the
prognosis for outgrowing milk allergy in this population is
quite poor.
It is also important to recognize, however, that even our
criteria 2 definition of tolerance may actually underestimate or overestimate the true rate of outgrowing milk
allergy. For example, there may be a lower rate of
resolution in those lost to follow-up than in those still
followed, because it is possible that many children with
persistently high cm-IgE levels would stop coming for
regular evaluations. A second factor that could lead to an
overestimation of the rate of resolution is that many
children with a cm-IgE <3 are certainly still allergic.9 In
fact, in the study by Perry et al,9 58% of children with
cm-IgE between 0.35 and 3 kU/L reacted at a formal
food challenge.
There are also several issues with this study that could
lead to an underestimation of the true rate of resolution of
CMA. Children lost to follow-up might have actually had
a higher rate of resolution than those still followed at any
given age. These children may have tolerated milk in
unsupervised, home-administered challenges, eliminating
the need for follow-up with us. Another factor could be
that children we assumed to still have milk allergy had
actually outgrown their allergy but had never been challenged. Most importantly, in comparing our data to the
general population of CMA, it is likely that these patients
who had been referred to a tertiary care center represent the
patients with most severe allergy who logically might have
the poorest prognosis.
There were several clinical and laboratory features we
found to be significant in predicting the resolution of
CMA. cm-IgE level was by far the most significant, and
probably the most clinically useful. By using each
patients highest cm-IgE, we found that a higher peak
was significantly associated with a reduced chance of

Food allergy, anaphylaxis,

dermatology, and drug allergy

FIG 1. CMA resolution over time. Kaplan-Meier survival curves for

development of tolerance to cows milk over time using 3 definitions for tolerance. The black line is the survival curve for criteria 1,
the red line is the survival curve for criteria 2, and the blue line is
the survival curve for criteria 3.

Persistent
Median (kU/L)
(number)

1176 Skripak et al

J ALLERGY CLIN IMMUNOL

NOVEMBER 2007

FIG 2. Trend in cm-IgE levels over time by final CMA status. Scatter plots of all cm-IgE levels recorded, by age,
to age 18 years (n 5 2498). A, All values in the group with persistent CMA (n 5 1651). B, All values in the group
with resolved CMA (n 5 847). Nonparametric smoothed curves show the trend in cm-IgE levels over time.
These curves approximate the mean value at any given age.

TABLE V. Predictors of oral tolerance to cows milk

Incidence of oral
tolerance

Food allergy, anaphylaxis,

dermatology, and drug allergy

FIG 3. Relationship of peak cm-IgE level to resolution of IgEmediated CMA over the period of the first 18 years of life. Patients
were stratified by peak cm-IgE level, and survival curves for each
stratum of peak cm-IgE level were plotted. The number of patients
in each stratum was as follows: <2 kU/L (n 5 162); 2-4.9 (n 5 106);
5-9.9 (n 5 85); 10-19.9 (n 5 115); 20-49.9 (n 5 107); 501 (n 5 229).

developing tolerance. We realize that peak cm-IgE can be

an impractical measure because it cannot be applied with
the same confidence to younger children. However, higher
cm-IgE levels were associated with poorer prognosis at
all ages and, even in younger children, the peak cm-IgE
categories can be applied as a best case scenario in
counseling families about prognosis. These data are in
agreement with several previous studies that support this
association between increasing cm-IgE level and worse
prognosis.4,5,10,11
The presence of both asthma and allergic rhinitis were
also associated with a lower likelihood of developing
tolerance. These factors remained significant even in the
multivariate analysis when controlling for peak cm-IgE

Asthma
Yes
No
Allergic rhinitis
Yes
No
Eczema
Yes
No
Other food allergy
Yes
No
Breast-fed, ever (n 5 646)
Yes
No
Formula-fed, ever (n 5 655)
Yes
No
Sex
Male
Female

4y

8y

12 y

P value*

6%
33%

15%
65%

50%
84%

<.0001

6%
30%

19%
55%

55%
72%

<.0001

20%
16%

43%
41%

67%
61%

.57

18%
29%

42%
50%

64%
67%

.07

21%
26%

45%
48%

70%
71%

.46

18%
35%

40%
53%

63%
89%

.005

16%
15%

42%
44%

63%
66%

.14

*Log-rank test.

level. In previous studies, markers of atopy or IgEmediated diseasefor example, the presence of urticaria
or IgE-sensitization to certain foods such as egghave
been associated with worse prognosis.6 However, it is important to note that asthma and rhinitis may appear to be
associated with a poorer prognosis because children who
are followed longer are more likely to carry these diagnoses as well as more likely to have retained milk allergy.

Skripak et al 1177

J ALLERGY CLIN IMMUNOL

VOLUME 120, NUMBER 5

significantly worse than what has been previously reported. Sensitivity persists into school age and beyond in
the majority of our patients. cm-IgE is highly predictive of
outcome and should be used in counseling patients on
prognosis. Prospective studies are needed to confirm this
potential increasing persistence of CMA.
We thank Elizabeth Johnson, MS, for her review of the statistical
analyses.
REFERENCES
1. 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.
2. Saarinen KM, Juntunen-Backman K, Jarvenpaa AL, Kuitunen P, Lope L,
Renlund M, et al. Supplementary feeding in maternity hospitals and the
risk of cows milk allergy: a prospective study of 6209 infants. J Allergy
Clin Immunol 1999;104:457-61.
3. Schrander JJ, van den Bogart JP, Forget PP, Schrander-Stumpel CT,
Kuijten RH, Kester AD. Cows milk protein intolerance in infants under
1 year of age: a prospective epidemiological study. Eur J Pediatr 1993;
152:640-4.
4. 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.
5. Hill DJ, Firer MA, Ball G, Hosking CS. Natural history of cows milk
allergy in children: immunological outcome over 2 years. Clin Exp Allergy 1993;23:124-31.
6. Saarinen KM, Pelkonen AS, Makela MJ, Savilahti E. Clinical course and
prognosis of cows milk allergy are dependent on milk-specific IgE status. J Allergy Clin Immunol 2005;116:869-75.
7. Bishop JM, Hill DJ, Hosking CS. Natural history of cow milk allergy:
clinical outcome. J Pediatr 1990;116:862-7.
8. Hill DJ, Davidson GP, Cameron DJ, Barnes GL. The spectrum of cows
milk allergy in childhood. Clinical, gastroenterological and immunological studies. Acta Paediatr Scand 1979;68:847-52.
9. Perry TT, Matsui EC, Kay Conover-Walker M, Wood RA. The relationship of allergen-specific IgE levels and oral food challenge outcome.
J Allergy Clin Immunol 2004;114:144-9.
10. Sampson HA. Utility of food-specific IgE concentrations in predicting
symptomatic food allergy. J Allergy Clin Immunol 2001;107:891-6.
11. Shek LP, Soderstrom L, Ahlstedt S, Beyer K, Sampson HA. Determination of food specific IgE levels over time can predict the development of
tolerance in cows milk and hens egg allergy. J Allergy Clin Immunol
2004;114:387-91.
12. Vanto T, Helppila S, Juntunen-Backman K, Kalimo K, Klemola T, Korpela
R, et al. Prediction of the development of tolerance to milk in children with
cows milk hypersensitivity. J Pediatr 2004;144:218-22.

Food allergy, anaphylaxis,

dermatology, and drug allergy

In this population, the presence of other food allergies was

also associated with a worse prognosis, although this
association was not statistically significant. In addition,
a worse prognosis was observed among patients who
had ever received formula. However, this association
may be biased because these data were missing for 20%
of the population, and these findings should be confirmed
in future studies.
In previous studies, rates of development of tolerance
for IgE-mediated or immediate-onsettype allergy have
varied: 76% by age 3 years,4 74% by age 5 years,6 and
22% by age 18 months to 13 years.5 The wide differences
in these rates are likely related most to the population studied, with the study by Host and Halken4 including an unselected group of children with milk allergy, and the study
by Hill et al5 focusing on a referral population more similar to ours. Most previous studies did not report details on
cm-IgE levels, but from the information available, it
appears that our population has higher levels on average,
with peak cm-IgE levels exceeding 2 kU/L in 80% of
the population, and exceeding 50 kU/L in 30%. Other
markers of the high degree of atopy in our population include 91% having at least 1 other food allergy, 49% with
asthma, 40% with allergic rhinitis, and 71% with eczema,
although these rates of asthma, eczema, and allergic
rhinitis are consistent with those found in previous
studies.6,10,12
Although the poor prognosis demonstrated in this study
may be a result of this highly atopic referral population, it
may also be that the character of CMA has changed over
time, and that CMA may now truly be a more persistent
disease. In fact, many of the previous studies in this area
are now nearly 2 decades old. In our clinic population, we
continue to see an increasing number of children whose
milk allergy persists into school age and even into
adolescence. We speculate that the factors driving the
rising prevalence of food allergy and other atopic conditions may also be contributing to the changing character of
CMA, and potentially other food allergies.
In conclusion, we have found in the largest cohort of
children with milk allergy ever reported that the prognosis
for the resolution of IgE-mediated CMA appears