ORIGINAL ARTICLE

Asian Pacific Journal of

Allergy and Immunology

IgE-mediated cow’s milk protein allergy

in Singaporean children
Kok Wee Chong,1 Si Hui Goh,1 Seyed Ehsan Saffari,2 Wenyin Loh,1 Isabel Sia,3 Sherilyn Seah,3 Anne Goh1

Abstract

Background: Cow’s milk protein allergy (CMA) is the second most common food allergy in Singapore. However, there is
limited data on local paediatric CMA.

Objective: We aimed to describe the demographics, clinical characteristics, natural history and diagnostic performance
of skin prick test (SPT) and cow’s milk-specific immunoglobulin E (CM-IgE) in Singaporean children diagnosed with
IgE-mediated CMA.

Methods: A retrospective review of medical records was conducted for children with an SPT performed to cow’s milk
between 2011 and 2016.

Results: There were 355 patients included, 313 cow’s milk allergic and 42 cow’s milk tolerant. The median age of reaction
was 6 months (IQR 4-8). The most common allergic presentation was cutaneous reactions, followed by gastrointestinal
reactions. Six patients (1.9%) reported anaphylaxis at initial presentation and 16 children (5.1%) experienced anaphylaxis
to cow’s milk at least once in their lifetime. Most of the CMA patients (81.8%) acquired natural tolerance by 6 years old.
SPT to cow’s milk of ≥ 7 mm and CM-IgE of ≥ 13 kU/L showed good discriminative abilities in predicting a failed oral food
challenge (OFC) outcome.

Conclusion: CMA is a food allergy which commonly presents during infancy, and parents need to be aware of the likeli-
hood of severe allergic reactions, including anaphylaxis. Prognosis for CMA is generally favourable. Future prospective co-
hort studies are required to better understand the natural history and better define the diagnostic cut-off values for allergy
testing in our population.

Key words: anaphylaxis, cow’s milk protein allergy, IgE-mediated food allergy, natural history, Singaporean children

From: Corresponding author:

1
Allergy Service, Department of Paediatric Medicine, KK Women’s and Kok Wee Chong
Children’s Hospital, Singapore Allergy Service, Department of Paediatric Medicine, KK Women’s &
2
Centre for Quantitative Medicine, Duke-NUS Medical School, Children’s Hospital,
National University of Singapore, Singapore 100 Bukit Timah Road, Singapore 229899
3
Yong Loo Lin School of Medicine, National University of Singapore, E-mail: [email protected]
Singapore

Introduction
Cow’s milk protein allergy (CMA) is one of the most com- milk-specific IgE (CM-IgE). Both the SPT and CM-IgE serve
mon food allergies in young children worldwide.1 CMA can be to detect the presence of IgE antibodies but a positive test
classified into either immediate-onset immunoglobulin E (Ig- (SPT wheal size of ≥ 3 mm larger than the negative control
E)-mediated, where the reaction occurs usually within minutes or CM-IgE ≥ 0.35 kU/L) cannot differentiate between sen-
following ingestion or delayed onset non-IgE-mediated where sitization alone and clinical allergy. Hence, the gold standard
the effects develop usually after 2 or more hours.2 The diagno- for diagnosis is still the double-blind placebo-controlled oral
sis of IgE-mediated CMA is often made by obtaining a history food challenge (DBPFC). Strict cow’s milk protein avoidance
suggestive of an immediate reaction to cow’s milk’s exposure, with provision of milk alternatives such as hypoallergenic milk
coupled with evidence of cow’s milk protein sensitization on formula and management of acute allergic reactions remain
either a positive skin prick test (SPT) or the presence of cow’s the mainstay of management in these children.3 Prognosis for
 Asian Pac J Allergy Immunol DOI 10.12932/AP-180219-0496

CMA is favorable as most children outgrow their allergy during Oral Food Challenge
childhood.4 Patients were selected for OFCs during follow up by the
CMA is the second most common food allergy in Singa- attending allergist to determine tolerance acquisition or as a
pore with an estimated prevalence of 0.1-0.44%.5 However, diagnostic challenge in some equivocal cases. Open OFCs were
there is limited local data on the demographics and clinical performed as part of clinical practice while double-blind, pla-
characteristics of CMA. The SPT wheal size and CM-IgE values cebo-controlled food challenges (DBPCFCs) were performed
which can be used as decision points to predict the outcomes as part of an ongoing pharmaceutical trial. All hospital-based
of OFCs have not been evaluated in a Singapore or South-East OFCs were conducted with fresh cow’s milk and the proce-
Asian paediatric population.6 This retrospective study aimed dures and dosage schedules were in accordance with the rec-
to evaluate the demographics, clinical characteristics, natural ommendations set out by the PRACTALL Consensus Report.8
history and diagnostic performance of SPT and CM-IgE in Home-based OFCs were conducted by home graded intro-
Singaporean children diagnosed with IgE-mediated CMA. duction of fresh cow’s milk in low risk patients. Patients were
considered to have passed an OFC if they successfully tolerated
Methods a cumulative dose of 4443mg of cow’s milk protein.8
Study design
In a 5-year period from 2011 to 2016, patients who had a Statistical analysis
SPT to cow’s milk protein were selected for further medical Data were extracted for statistical analysis using SAS soft-
records review. They were identified through our paediatric ware version 9.4 for Windows (SAS, Inc, Cary, NC). Baseline
allergy service’s database at KK Women’s and Children’s Hos- demographic and clinical features were compared between
pital, the main tertiary referral allergy centre in Singapore. CMA and CMT groups using Mann Whitney and Fisher’s ex-
These patients had an SPT performed either due to clinical re- act test for continuous and categorical variables, respectively.
actions to cow’s milk protein, or as part of their work-up for The association between demographics, atopic history, SPT
other food allergies or eczema. wheal size and food-specific IgE concentrations with the risk
Inclusion criteria were 1) patients with proven CMA, as of cow’s milk allergy were tested using univariate and multi-
defined by failing an OFC to cow’s milk, or a documented variable logistic regression analysis. Receiver Operating Char-
immediate reaction to cow’s milk in the preceding 6 months acteristic (ROC) analysis was performed to assess the accura-
coupled with a positive SPT; 2) patients who were cow’s milk cy of SPT and IgE. Sensitivity, specificity, positive predictive
tolerant (CMT), defined by passing an initial diagnostic OFC value (PPV), negative predictive value (NPV) and area under
to cow’s milk, or if they were documented to be taking cow’s the curve (AUC) were calculated for various cut-off points of
milk regularly without any reactions at their first presentation SPT wheal size and CM-IgE concentrations using univariate
to our centre. The diagnosis of anaphylaxis was based on the logistic regression approach. Statistical significance was set at
World Allergy Organization guidelines.7 Exclusion criteria were P < 0.05.
1) patients who had only sensitization found on SPT, without
any prior exposure to cow’s milk or dairy products; 2) patients Ethical approval
with mixed IgE and non-IgE mediated CMA who had primar- The study with waiver of informed consent was approved by
ily delayed (> 1 hour onset) gastrointestinal symptoms. Data SingHealth Centralised Institutional Review Board (reference
on patient demographics, clinical reactions to cow’s milk, SPT, number 2016/2519).
CM-IgE results, and milk alternatives were collected. Personal
history of rhinitis, atopic dermatitis, asthma and other food Results
allergies were diagnosed by the attending allergist; whilst the Demographics and clinical characteristics
rest of personal and family history of atopy were based on A total of 355 patients were included in the study, 313 were
parental reports. cow’s milk allergic and 42 cow’s milk tolerant, at initial pre-
sentation to our unit. The median duration of follow up was
Skin prick test and cow’s milk-specific IgE measurements 1.62 years. The demographics, personal and family history of
The skin was prepped with alcohol and cow’s milk protein atopy were summarised in Table 1. There was no statistical dif-
extract (Stallergenes Greer, Lenoir, LC, USA) was applied to the ference in the racial distribution between the CMA and CMT
skin using a sterile disposable applicator, Duotip-Test® (Lincoln groups, with the CMA group being 65.81% Chinese, 11.82%
Diagnostics, Decatur, IL, USA) by trained technicians. Skin tests Malay, 10.86% Indian and 11.5% others. The first reaction
were performed on the backs of infants and on the forearms of reported among those with CMA (Table 2) were cutaneous
older children. A positive control (histamine) and a negative (rash, angioedema) in 92.4%, gastrointestinal (abdominal pain,
control (saline) solution were also used. SPT wheal size was vomiting, diarrhea) in 28.8%, respiratory (rhinorrhoea, sneez-
measured after 15 minutes. The mean diameter recorded was ing, coughing, stridor, wheezing, shortness of breath) in 7.7%
calculated from the average of the 2 largest measurements that and cardiovascular (hypotension, drowsiness) in 0.6%. Six pa-
were perpendicular to each other. A positive SPT was taken as a tients (1.9%) reported anaphylaxis at initial presentation and
wheal size of ≥ 3 mm compared to the negative control. CM-IgE 16 children (5.1%) experienced anaphylaxis to cow’s milk at
were measured by using the ImmunoCAP System FEIA (Phadia least once in their lifetime.
AB, Uppsala, Sweden).
 Cow’s milk allergy in Singapore children

Table 1. Demographic and clinical characteristics of cow’s milk protein allergy and cow’s milk protein tolerant patients

Variable Cow’s Milk Protein Allergy Cow’s Milk Protein Tolerant p value* Un-adjusted Adjusted
(n = 313) (n = 42) OR** (95% CI) OR** (95% CI)

Demographics

Male Gender 182 (58.15) § 28 (66.67) 0.320 1.42 (0.72, 2.78) 1.34 (0.66, 2.72)

Age at 1st reaction (months) 6 (4 – 8) - - - -

Age at 1st SPT (months) 10.6 (7.3–15.9) 15.1 (10.1–39) < 0.001 0.98 (0.97, 0.99)+ 0.98 (0.97, 0.99)+

Chinese Ethnicity 206 (65.81) 23 (54.76) 0.222 1.59 (0.84, 3.04) 1.85 (0.93, 3.66)

Personal History of Atopy

Rhinitis¶ 106 (33.97) 18 (45) 0.218 0.63 (0.32, 1.21) 0.76 (0.38, 1.54)

Atopic dermatitis¶ 226 (72.44) 30 (75) 0.851 0.90 (0.43, 1.90) 0.84 (0.38, 1.85)

Asthma ¶
34 (10.9) 2 (5) 0.403 1.91 (0.50, 7.32) 2.21 (0.50, 9.82)

Drug allergy# 13 (4.17) 1 (2.5) 1.000 1.19 (0.20, 7.02) 1.17 (0.20, 7.04)

Urticaria/ angioedema# 11 (3.56) 2 (5) 0.651 0.59 (0.14, 2.55) 0.54 (0.12, 2.36)

Other food allergies*** 180 (57.51) 24 (60) 0.865 0.91 (0.47, 1.77) 0.87 (0.44, 1.74)

Family History of Atopy

Rhinitis 148 (47.9) 18 (45) 0.740 1.12 (0.58, 2.16) 0.97 (0.49, 1.90)

Atopic dermatitis 123 (39.81) 16 (40) 1.000 0.98 (0.50, 1.91) 0.85 (0.43, 1.69)

Asthma 95 (30.74) 5 (12.5) 0.016 2.87 (1.13, 7.31) +

2.86 (1.11, 7.34)+

Food Allergy 61 (19.81) 4 (10) 0.194 2.02 (0.72, 5.62) 1.69 (0.60, 4.74)

Investigation Results

SPT wheal size (mm)^ 6 (4–9) 3 (3–4) < 0.001 1.52 (1.28, 1.81)+ 1.66 (1.35, 2.04)+

IgE to cow’s milk (kU/L)^ 6.6 (1.8–24.1) 1 (0.7–1.7) 0.002 1.33 (0.97, 1.81) 1.27 (0.96, 1.68)

Abbreviations: SPT = Skin Prick Test; IgE = Immunoglobulin E; OR = Odds Ratio; CI = Confidence Interval
§
Frequency (%) for categorical variables; Median (IQR) for continuous variables
* Mann Whitney U test and Fisher’s exact test for continuous and categorical variables, respectively
** Univariate and multivariable logistic regression analysis; adjusted for age at 1st SPT and family history of asthma
*** Clinical diagnosis of other food allergy based on clinical history and positive SPT/ IgE
#
Parental report of patient’s history of drug allergy and urticaria/angioedema
¶
Physician diagnosed rhinitis, atopic dermatitis and asthma
^
Odds ratio is reported for the risk of SPT ≥ 3 mm, and serum IgE ≥ 0.35 kU/L
+
Significant at p < 0.05

Table 2. Description of allergic reactions at presentation (n = 313)

Clinical manifestation Percentage (%) Clinical manifestation Percentage (%) Clinical manifestation Percentage (%)

Cutaneous 92.4 Respiratory 7.7 Gastrointestinal 28.8

Any rash 89.1 Runny nose 1.0 Abdominal pain 0.3

Urticarial rash 42.5 Sneezing 1.0 Vomiting 27.2

Maculopapular rash 4.5 Coughing 3.2 Diarrhoea 2.2

Eczematous rash 10.5 Stridor 0.3 Cardiovascular 0.6

Perioral rash 20.8 Wheezing 1.6 Drowsiness 0.6

Rash not described 16.6 Shortness of breath 2.9 Hypotension 0

Angioedema 28.8 Anaphylaxis 1.9

 Asian Pac J Allergy Immunol DOI 10.12932/AP-180219-0496

Natural history 1.0

During follow-up of our CMA children, some achieved nat-

Cow Milk Allergy (Probability)

ural tolerance, as defined by passing a hospital/ home-based 0.8
OFC. The proportions of CMA children who had outgrown
their allergy (defined by either passing a hospital/ home-based
0.6
OFC) by various ages was shown in Figure 1. By 6 years old,
81.8% of them had self-acquired tolerance.
0.4
Milk alternatives
Milk alternatives consumed by the CMA group included 0.2
breastmilk (54.3%), soy milk (44.4%), extensively-hydrolyzed
formula (EHF) (8.6%), amino acid formula (AAF) (9.6%), par-
0.0
tially-hydrolyzed formula (PHF) (9.6%) and goat’s milk (2.9%).
0 5 10 15
27 CMA children (8.6%) were sensitized to soy milk, of which Age (year)
18 (5.7%) had documented allergic reactions to soy.
Figure 1. Cow’s milk allergy resolution over time
Diagnostic performance of SPT and CM-IgE The graph represents the Kaplan-Meier survival curve for de-
There was a total of 121 OFCs conducted in 93 patients, of velopment of tolerance (defined by either passing a hospital or
which 58 were hospital-based OFCs. Table 2 showed the diag- home-based cow’s milk oral food challenge) to cow’s milk over
nostic performance of both the SPT and the CM-IgE tests, using time with the 95% confidence intervals

Table 3. Diagnostic performance for skin prick test and IgE to cow’s milk protein, ROC analysis

Test Cut-off Un-Adjusted p value Sens Spec PPV NPV AUC

Point Odds Ratio
(95% CI)*

SPT (mm) 3 3.59 (0.94, 13.8) 0.0617 87.9 34.8 65.9 66.7 66.3

4 3.32 (1.08, 10.2) 0.0362 72.7 56.5 70.6 59.1 64.8

5 6.58 (1.96, 22.0) 0.0023 66.7 78.3 81.5 62.1 71.8

6 13.1 (2.89, 59.1) 0.0009 60.6 91.3 90.9 61.8 76.3

7 14.1 (2.26, 88.4) 0.0046 48.4 95.7 94.1 56.4 75.3

8 9.88 (1.56, 62.7) 0.0151 39.4 95.7 92.9 52.4 72.6

9 6.71 (1.02, 44.0) 0.0473 30.3 95.7 90.9 48.9 69.9

10 18.2 (0.86, 385) 0.0623 27.3 100.0 100.0 48.9 74.5

Cow’s milk 0.35 13.0 (0.39, 433) 0.1516 100.0 23.1 65.5 100.0 82.8
specific IgE
(kU/L) 1 4.08 (0.80, 20.8) 0.0904 84.2 46.2 69.6 66.7 68.1

2 15.8 (2.68, 93.6) 0.0023 79.0 84.6 88.2 73.3 80.8

3 13.9 (1.90, 101) 0.0095 63.2 92.3 92.3 63.2 77.7

4 9.21 (1.27, 67.0) 0.0283 52.6 92.3 90.9 57.1 74.0

5 9.21 (1.27, 67.0) 0.0283 52.6 92.3 90.9 57.1 74.0

6 9.21 (1.27, 67.0) 0.0283 52.6 92.3 90.9 57.1 74.0

7 9.21 (1.27, 67.0) 0.0283 52.6 92.3 90.9 57.1 74.0

8 7.54 (1.03, 55.2) 0.0468 47.4 92.3 90.0 54.6 72.3

9 7.54 (1.03, 55.2) 0.0468 47.4 92.3 90.0 54.6 72.3

10 6.16 (0.83, 45.7) 0.0754 42.1 92.3 88.9 52.2 70.5

SPT = Skin Prick Test; IgE = Immunoglobulin E; Sens = Sensitivity; Spec = Specificity; PPV = Positive Predictive value; NPV = Negative Predictive Value; AUC = Area
Under the receiver operator characteristic Curve; CI = Confidence Interval
* Univariate logistic regression analysis
 Cow’s milk allergy in Singapore children

Table 3. (Continued)

Test Cut-off Un-Adjusted p value Sens Spec PPV NPV AUC

Point Odds Ratio
(95% CI)*

Cow’s milk 11 6.16 (0.83, 45.7) 0.0754 42.1 92.3 88.9 52.2 70.5
specific IgE
(kU/L) 12 6.16 (0.83, 45.7) 0.0754 42.1 92.3 88.9 52.2 70.5

13 20.0 (0.87, 457) 0.0609 42.1 100.0 100.0 54.2 77.1

14 16.2 (0.69, 382) 0.0840 36.8 100.0 100.0 52.0 76.0

15 16.2 (0.69, 382) 0.0840 36.8 100.0 100.0 52.0 76.0

SPT = Skin Prick Test; IgE = Immunoglobulin E; Sens = Sensitivity; Spec = Specificity; PPV = Positive Predictive value; NPV = Negative Predictive Value; AUC = Area
Under the receiver operator characteristic Curve; CI = Confidence Interval
* Univariate logistic regression analysis

various cut-off values, to predict the outcomes of hospital-based between study populations. In a Danish birth cohort study,
OFCs. SPT to cow’s milk of ≥ 7 mm provided a positive pre- 56% of the patients outgrew their allergy at 1 year, 77% at
dictive value (PPV) of 94.1% in predicting a failed OFC out- 2 years, 87% at 3 years, 92% at 5 and 10 years and 97% at 15
come. CM-IgE of ≥ 13 kU/L yielded 100% PPV and specificity. years of age.14 However, in an Israeli study, less than half of
A further analysis for those 2 years old or older showed that SPT the children diagnosed with IgE-mediated CMA during the
to cow’s milk of ≥ 6 mm provided a PPV of 95.0% in predict- first 9 years of life outgrew it.15 From a more recent study in
ing a failed OFC outcome, with CM-IgE of ≥ 13 kU/L similarly Japan,16 tolerance acquisition rates in children with IgE-medi-
yielding 100% PPV and specificity. Subgroup analysis for those ated CMA were 32.6%, 64.1%, and 84.8% at 3, 5, and 6 years
younger than 2 years old could not be performed due to small of age respectively. Similar to the Japanese, 81.8% of our co-
sample size. hort acquired tolerance by 6 years of age. Variability in natural
tolerance could be a result of heterogeneity in study design,
Discussion outcome measurement, population differences, or a result of a
This is the first and the largest retrospective review of IgE- change in the natural history of CMA over these years.
mediated CMA children in Singapore. The median age of first The mainstay of management for CMA in young children
reaction was 6 months in our cohort, consistent with the ob- is dietary avoidance and replacement with a milk substitute.
servation that CMA often presents during infancy and early Breastfeeding is our first line recommendation for infants di-
childhood, when milk is still the main component of the child’s agnosed with CMA, hence more than half of our CMA cohort
diet. (54.3%) were being breastfed. Whenever possible, mothers
In this cohort, the vast majority of patients (92.4%) reported were encouraged to continue breastfeeding and they were not
cutaneous symptoms at initial reaction and few had respirato- routinely advised for dietary dairy restrictions, unless the in-
ry and cardiovascular symptoms. In our study cohort, anaphy- fants exhibited symptoms whilst being breastfed. A hypoal-
laxis was reported in 16 cases (5.1%), of which 6 cases (1.9%) lergenic formula is one which is tolerated by at least 90% of
reported anaphylaxis at initial presentation. Allergic reactions infants with proven CMA.17 Only EHF and AAF are consid-
to cow’s milk in children are mostly mild to moderate but life ered hypoallergenic by this criterion and are the formulas of
-threatening anaphylaxis can occur.4 In the EuroPrevall birth choice for management of CMA. PHF are not considered hy-
cohort,9 none of the children with suspected CMA reported poallergenic and should not be recommended in infants with
a history of anaphylaxis. Recent studies reviewing triggers for proven CMA. In our CMA cohort, 9.6% were tolerating PHF
food-induced anaphylaxis in Singaporean children reported at the point of diagnosis in the allergy clinic. In a Thai cohort
cow’s milk to be responsible for only 5.2-7.3% of all cases.10,11 of 382 patients diagnosed with CMA, 35.7% of them were
This is in contrast to other recently reported studies, such as a reported to tolerate PHF well.18 Postulations as to why CMA
Korean study, which reported cow’s milk to be the most com- infants tolerated PHF included a possibly milder CMA phe-
mon trigger, accounting for 28.4% of all food-induced ana- notype and variabilities in the degree of allergenicity (extent of
phylaxis in their cohort,12 and a study in New Zealand which hydrolysis) of the PHFs. As the CMA diagnosis of most of our
reported 21% of their food-induced anaphylaxis being second- patients tolerating PHF (86.7%) were not based on OFC, there
ary to cow’s milk.13 These differences could be explained by is also a possibility that some of these patients were not cow’s
variations in patient population, prevalence of type of food milk allergic to begin with. Up to 10-14% of CMA children
allergies (eg. cow’s milk was the most common food allergy also present with soy allergy.19 For the majority of CMA infants
in the Korean study12 while shellfish was most common in the who tolerate soy, soy formulas are nutritionally adequate milk
Singaporean study 10), as well as variations in study designs and alternatives. They are also more palatable and cost significant-
definitions used in anaphylaxis. ly less compared to EHF and AAF in Singapore. In our CMA
The prognosis for CMA is generally favorable with most cohort, close to 95% of our patients were clinically tolerant of
patients outgrowing their allergy during childhood.4 The du- soy and we recommend assessing for soy tolerance in CMA
ration required to acquire natural tolerance varied significantly children.
 Asian Pac J Allergy Immunol DOI 10.12932/AP-180219-0496

Due to the high degree of sequence homology between as some of these patients could have been just sensitized but
cow’s milk and goat’s milk proteins, there is significant cross- not truly allergic to cow’s milk. Our patients did not undergo
allergy and up to 95% of CMA patients would also react regular OFCs to assess for tolerance acquisition but were only
against goat’s milk.4,20,21 A notable finding in this study was selected for OFC when they were deemed by their attending
that there was a small proportion (2.9%) of CMA infants who allergist to be at low risk of a reaction. This could cause a bias
were able to tolerate goat’s milk. Similar findings had been in our results when we analyzed the predictive cut-off values
previously reported in few studies: 1) a small study of 12 CMA for SPT and CM-IgE as those patients with higher values who
patients in Spain, in which 25% of these patients showed might have been CM tolerant, would have been excluded from
adequate oral tolerance and had negative immunological our analysis. Our analysis of tolerance acquisition was also
testing to goat’s milk;22 2) a clinical trial conducted in France based on children who had undergone an OFC and had ex-
also found that 51/55 CMA children tolerated goat’s milk for cluded children who were assumed to be allergic based on
periods ranging from 8 days to 1 year;20 3) a study in Sweden their SPT/ CM-IgE results. We had also not taken into account
showed that all 26 confirmed IgE-mediated CMA patients children who were lost to follow up, who could have possibly
had positive skin test and IgE results to goat’s milk, however outgrown their CMA and hence defaulted subsequent clinic
2/26 passed a double-blind, placebo-controlled OFC to goat’s visits. Hence the natural history gathered from our study can
milk.21 It has been surmised that goat’s milk could be less only be taken as an estimated reference and prospective co-
allergenic than cow’s milk due to its lower alpha-casein con- hort studies are required to assess these better. Another limita-
tent.21 However, we do stress that goat’s milk tolerance in CMA tion was that our diagnosis was made based on an open OFC
infants is an exception and goat’s milk cannot be recommend- instead of the gold standard of a double-blind, placebo-con-
ed as a suitable milk alternative for CMA patients. Similar trolled OFC.
to the discussion on PHF, these CMA patients who tolerated
goat’s milk were not challenge-proven to have CMA and hence Conclusion
a possibility of over-diagnosis. In conclusion, CMA is a food allergy which commonly
A recent systematic review by Cuomo et al concluded that presents during infancy, and parents need to be aware of the
for children < 2 years old, SPT wheal size of ≥ 6 mm or CM- likelihood of severe allergic reactions, including anaphylax-
IgE ≥ 5 kU/L was highly predictive of a diagnosis of CMA.6 is. The mainstay of management includes strict avoidance of
Results for children older than 2 years old had been a lot more cow’s milk protein. Apart from breastfeeding and hypoaller-
heterogenous,6 although a proposed cut-off of SPT wheal size genic formulas, soy formula can be considered for CMA in-
≥ 8 mm or CM-IgE ≥ 15 kU/L by Du Toit G et al had been fants who have been assessed to be soy-tolerant. Goat’s milk
widely accepted.23 Our data showed that SPT to cow’s milk of and PHF should not be recommended as a milk substitute.
≥ 7 mm provided a good PPV and specificity of 94.1% and Most CMA patients acquired natural tolerance by 6 years
95.7% respectively in predicting a failed OFC outcome. A SPT old. Future prospective cohort studies are required to better
result of ≥ 10 mm would provide 100% PPV and specificity. understand the natural history and better define the diagnostic
CM-IgE of level ≥ 13 kU/L yielded 100% PPV and specifici- cut-off values for allergy testing in our population.
ty. Using the conventional cut-off values of SPT < 3 mm and
CM-IgE < 0.35 kU/L yielded high NPV and high sensitivity, at
the expense of specificity. In our cohort, using a CM-IgE cut- Acknowledgements
off of 0.35 kU/L yielded 100% NPV and sensitivity, which is We thank our allergy specialist nurse, Ms Lim Hwee Hoon
clinically useful to rule out CMA in patients presenting with and skin prick test laboratory technicians for their consistent
equivocal or inconsistent history. These cut-off values will be hard work and excellent patient care.
useful in predicting the outcomes of OFCs, potentially re-
ducing the number of OFCs, which can be time and resource Conflict of interest
consuming, and carry with it a risk of severe allergic reactions. The authors have no conflict of interest or sources of finan-
Molecular diagnostic allergy testing involving component re- cial support to report.
solved diagnostics (CRD) is increasingly used in routine clin-
ical practice. However, we had limited data on CRD in our
cohort for analysis.
Authors contributions
• Conceptualization: Kok Wee Chong
The strength of this study was that this is the first and larg-
• Data curation: Kok Wee Chong, Isabel Sia, Sherilyn Seah
est description of our IgE-mediated CMA children in Singa-
• Formal analysis: Seyed Ehsan Saffari
pore, which provided new information on their demographics,
• Manuscript writing: Kok Wee Chong, Si Hui Goh, Isabel Sia,
clinical presentations and natural history. Based on our data,
Sherilyn Seah
we also derived cut-off values for SPT wheal size and CM-IgE,
• Manuscript review: Si Hui Goh, Wenyin Loh, Anne Goh
which will be useful in predicting the outcomes of OFCs in
• Supervision: Anne Goh
our own paediatric population. The main limitations of the
study stem from its retrospective study design. Most of the
CMA patients were diagnosed clinically by the attending al- References
lergist based on a clearly documented immediate reaction to 1. Prescott SL, Pawankar R, Allen KJ, Campbell DE, Sinn JK, Fiocchi A, et al.
A global survey of changing patterns of food allergy burden in children.
cow’s milk coupled with a positive SPT, instead of the diagnos- World Allergy Organ J. 2013;6(1):21.
tic gold standard of an OFC. This is a weakness of the study
 Cow’s milk allergy in Singapore children

2. Johansson SGO, Bieber T, Dahl R, Friedmann PS, Lanier BQ, Lockey 12. Lee SY, Ahn K, Kim J, Jang GC, Min TK, Yang HJ, et al. A multicenter
RF, et al. Revised nomenclature for allergy for global use: Report of the retrospective case study of anaphylaxis triggers by age in Korean children.
Nomenclature Review Committee of the World Allergy Organization, Allergy, Asthma Immunol Res. 2016;8(6):535–40.
October 2003. J Allergy Clin Immunol. 2004;113(5):832–6. 13. Speakman S, Kool B, Sinclair J, Fitzharris P. Paediatric food-induced
3. Fiocchi A, Dahda L, Dupont C, Campoy C, Fierro V, Nieto A. Cow’s milk anaphylaxis hospital presentations in New Zealand. J Paediatr Child
allergy: towards an update of DRACMA guidelines. World Allergy Organ J. Health. 2018;54(3):254–9.
2016;9(1):35. 14. Høst A, Halken S, Jacobsen HP, Christensen AE, Herskind AM, Plesner K.
4. Luyt D, Ball H, Makwana N, Green MR, Bravin K, Nasser SM, et al. BSACI Clinical course of cow’s milk protein allergy/intolerance and atopic diseases
guideline for the diagnosis and management of cow’s milk allergy. Clin Exp in childhood. Pediatr Allergy Immunol. 2002;13(s15):23–8.
Allergy. 2014;44(5):642–72. 15. Levy Y, Segal N, Garty B, Danon YL. Lessons from the clinical course
5. Tham EH, Lee BW, Chan YH, Loo EXL, Toh JY, Goh A, et al. Low of IgE-mediated cow milk allergy in Israel. Pediatr Allergy Immunol.
Food Allergy Prevalence Despite Delayed Introduction of Allergenic 2007;18(7):589–93.
Foods—Data from the GUSTO Cohort. J Allergy Clin Immunol Pract. 16. Koike Y, Sato S, Yanagida N, Asaumi T, Ogura K, Ohtani K, et al. Predictors
2018;6(2):466-475.e1. of Persistent Milk Allergy in Children: A Retrospective Cohort Study. Int
6. Cuomo B, Indirli GC, Bianchi A, Arasi S, Caimmi D, Dondi A, et al. Specific Arch Allergy Immunol. 2018;175(3):177–80.
IgE and skin prick tests to diagnose allergy to fresh and baked cow’s milk 17. Høst A, Koletzko B, Dreborg S, Muraro A, Wahn U, Aggett P, et al. Dietary
according to age: a systematic review. Ital J Pediatr. 2017;43(1):93. products used in infants for treatment and prevention of food allergy.
7. Simons FER, Ardusso LR, Bilò MB, Cardona V, Ebisawa M, El-Gamal YM, Joint Statement of the European Society for Paediatric Allergology and
et al. International consensus on (ICON) anaphylaxis. World Allergy Organ Clinical Immunology (ESPACI) Committee on Hypoallergenic Formulas
J. 2014;7(1):1–19. and the European Society for Paediatric Gastroenterology, Hepatology
8. Sampson HA, Gerth Van Wijk R, Bindslev-Jensen C, Sicherer S, Teuber and Nutrition (ESPGHAN) Committee on Nutrition. Arch Dis Child.
SS, Burks AW, et al. Standardizing double-blind, placebo-controlled oral 1999;81(1):80–4.
food challenges: American Academy of Allergy, Asthma & Immunology 18. Ngamphaiboon J, Chatchatee P, Thongkaew T. Cow’s milk allergy in Thai
-European Academy of Allergy and Clinical Immunology PRACTALL children. Asian Pacific J Allergy Immunol. 2008;26(4):199–204.
consensus report. J Allergy Clin Immunol. 2012;130(6):1260–74. 19. Kattan JD, Cocco RR, Järvinen KM. Milk and soy allergy. Pediatr Clin
9. Schoemaker AA, Sprikkelman AB, Grimshaw KE, Roberts G, North Am. 2011;58(2):407–26.
Grabenhenrich L, Rosenfeld L, et al. Incidence and natural history of 20. Fiocchi A, Brozek J, Schünemann H, Bahna SL, von Berg A, Beyer K,
challenge-proven cow’s milk allergy in European children - EuroPrevall et al. World Allergy Organization (WAO) Diagnosis and Rationale for
birth cohort. Allergy. 2015;70(8):963–72. Action against Cow’s Milk Allergy (DRACMA) Guidelines. Pediatr Allergy
10. Ganapathy S, Lwin Z, Ting DHA, Goh LSH, Chong SL. Anaphylaxis in Immunol. 2010;21:1–125.
children: Experience of 485 episodes in 1,272,482 patient attendances at a 21. Bellioni-businco B, Paganelli R, Lucenti P, Giampietro PG, Perborn H,
tertiary paediatric emergency department from 2007 to 2014. Ann Acad Businco L. Allergenicity of goat’s milk in children with cow’s milk allergy.
Med Singapore. 2016;45(12):542–8. J Allergy Clin Immunol. 1999;103(6):1191–4.
11. Goh SH, Soh JY, Loh W, Lee KP, Tan SC, Heng WJK, et al. Cause and 22. Infante Pina D, Tormo Carnice R, Conde Zandueta M. Use of goat’s milk in
clinical presentation of anaphylaxis in Singapore: From infancy to old age. patients with cow’s milk allergy. An Pediatr (Barc). 2003;59(2):138–42.
Int Arch Allergy Immunol. 2018;175(1–2):91–8. 23. Du Toit G, Santos A, Roberts G, Fox AT, Smith P, Lack G. The diagnosis
of IgE-mediated food allergy in childhood. Pediatr Allergy Immunol.
2009;20(4):309–19.