Diagnostic Decision Points of Specific IgE Titers in Patients With Food Allergy: Are They Appropriate in All Clinical Settings?

In daily clinical practice, it is not so easy to make a confirmative diagnosis of food allergy. Interpretation of skin prick test and serum food-specific IgE (sIgE) test results is influenced by detailed clinical history of an individual patient, but ingested foods which provoke adverse reactions often contain many ingredients and lead to a misdiagnosis of food allergy. Oral food challenge is recommended for the diagnosis of food allergy, but we cannot always perform oral food challenge for various reasons, including recent severe reactions to food, severe atopic dermatitis, patient' and/or parents' refusal and noncooperation of young patients. Thus, from the physician's point of view, cutoff values of sIgE titers that provide high positive and negative predictive values would be very useful in clinical settings because they help determine which patient is more likely to have symptoms in response to a certain food and which are probably nonreactive.

Since Sampson1,2 was first to publish studies on the efficacy of serum specific IgE in determining which food-allergic patient is more likely to fail oral food challenge,3 their diagnostic decision points (DDPs) have been widely used in clinical practice worldwide as well as in Korea.4 Positive predictive value (PPV) represents the proportion of symptomatic individuals to those with positive test results.5 Positive predictive accuracy can be calculated by the following equation: Positive predictive accuracy = sP / sP + (1 - f)(1 - P) where s=sensitivity, f=specificity, and P=prevalence of the condition in the population studied.2 The composition of the population investigated affects the positive predictive accuracy as denoted in the formula. Predictive accuracy is difficult to establish because it varies with the prevalence of disease.2 Sampson et al.2 attempted to minimize the influence of disease prevalence on DDPs among different populations and established DDPs from 90% specificity as an alternative approach. However, the value of the PPV was very low (50.0% in children <24 months of age and 59.3% in those ≥24 months of age, respectively) when the previously established DDPs of egg white-sIgE concentrations for egg allergy were applied in that study population; the value of the PPV was slightly lower than 95% (83.3% in children <24 months of age and 91.7% in those ≥24 months of age, respectively) when the previously established DDPs of cow milk-sIgE concentrations for milk allergy was applied in this study population.6 Similar results have been reported in a recent study of Korean children demonstrating that the sensitivity and specificity of the previously established DDPs for egg white-sIgE are low.7

In the current issue of the Allergy, Asthma & Immunology Research, Kim et al.6 suggest that it is more practical to use higher DDPs of egg white- and cow's milk-sIgE antibodies in Korean children for the diagnosis of egg and cow milk allergies in order to avoid unnecessary diet restriction. Predictive values vary from study to study (Tables 1 and 2).1,2,3,5,6,8,9,10,11,12,13,14,15 These discrepancies may be due to various factors, such as patient age, duration of food allergen avoidance at the time of testing, patient selection, and clinical disorders of patients studied8 as well as the prevalence of disease. DDPs with 95% predictive value are subjected to some limitations: (1) the majority of subjects with atopic dermatitis and elevated total IgE levels showed potentially skewed data and (2) many subjects with higher IgE antibodies to milk or egg did not undergo oral food challenge to confirm the clinical reactivity, thus limiting their diagnostic efficacy.16

Table 1

Positive predictive values of egg white-specific IgE titers reported from various studies

Year	Reporter	Country	Number	Prevalence (%)	% of AD	PPV	Age	sIgE titer (kUA/L)
1997	Sampson1	US	126	73	All	95% PPV	0.6-17.9 yr	6
2001	Sampson2	US	75	80	61	90% specificity (95% PPV)	3 mo-14 yr	7*
2001	Boyano-Martinez9	Spain	81	79	43	94% PPV	<2 yr	0.35
2001	Roehr10	Germany	98	24	all	100% PPV (with APT)	2 mo-11.2 yr	17.5
2002	Boyano-Martinez11	Spain	58	64	50	50th percentile of the frequency distribution	11-24 mo	1.98*
2003	Osterballe12	Denmark	56	64	all	95% PPV	<2 yr	1.5
							>2 yr	1.3
2005	Celik-Bilgili5	Germany	227	67	88	95% PPV	<1 yr	10.9
							≥1 yr	13.2
2007	Komata3	Japan	764	45	74	95% PPV	<1 yr	13.0
							1 yr	23.0
							≥2 yr	30.0
2008	Ando13	Japan	108	62	94	95% PPV	14 mo-13 yr	7.4 (Heated; 30.7)
2015	Kim6	Korea	273	19	81.3	95% PPV	<2 yr	22.9
							≥2 yr	28.1

*The most widely used PPV.

AD, atopic dermatitis; PPV, positive predictive value; sIgE, specific IgE; mo, months; yr, year.

Table 2

Positive predictive values of cow milk-specific IgE titers reported from various studies

Year	Reporter	Country	Number	Prevalence (%)	% of AD	PPV	Age	sIgE titer (kUA/L)
1997	Sampson1	US	109	50	All	95% PPV	0.6-17.9 yr	32
2001	Sampson2	US	62	66	61	90% specificity (95% PPV)	3 mo-14 yr	15*
2001	Garcia-Ara14	Spain	170	44	23	90% PPV	<1 yr	2.5
						95% PPV	<1 yr	5*
2001	Roehr10	Germany	98	41	all	100% PPV (with APT)	2 mo-11.2 yr	0.35
2004	Garcia-Ara15	Spain	66	all	unrevealed	90% PPV	13-18 mo	1.5
							19-24 mo	6
							25-36 mo	14
2005	Celik-Bilgili5	Germany	397	49	88	90% PPV	<1 yr	25.8
							1-16.1 yr	88.8
2007	Komata14	Japan	861	25	74	95% PPV	<1 yr	5.8
							1 yr	38.6
							≥2 yr	57.3
2015	Kim6	Korea	225	23.1	81.3	90% PPV	<2 yr	31.4
							≥2 yr	10.1

*The most widely used PPV.

AD, atopic dermatitis; PPV, positive predictive value; sIgE, specific IgE; mo, month; yr, year.

When determining timing for reintroduction of egg or cow milk to patients who are likely to pass food challenge, physicians should consider patient age, natural history of concerned food allergy, severity of previous symptoms, and reduction in the food-sIgE level in an individual patient.4 Challenge tests can be performed in children when their food-sIgE level decreases to about one-fourth of the diagnostic decision point, unless they have experienced any recent allergic reaction.2 Ando et al.13 have reported that negative predictive value (NPV 84%), based on 95% sensitivity, is 0.6 kUA/L for egg white. Kim et al.6 have suggested a new NPV for Korean children with suspected egg or milk allergy.6 They reported that egg white -sIgE levels of 3.45 kUA/L yield a NPV of 93.6% in children <24 months of age, while a NPV of 1.80 kUA/L yield a NPV of 99.2% in those ≥24 months of age. The cow's milk-sIgE levels of 0.59 kUA/L and 0.94 kUA/L showed NPVs of 100% and 96.9%, respectively, in children <24 months of age and in those ≥24 months of age. These NPVs could provide data for clinicians to select patients who are likely to pass food challenge.

The accuracy of diagnostic tests in food allergy is important because a misdiagnosis of food allergy can lead to serious consequences, including nutritional deficiency.16 The diagnosis of food allergy begins with detailed clinical history taking and thorough physical examination. Both skin prick tests and sIgE measurement alone are not sufficient to diagnose food allergy and should then be guided by the information gathered from the history. The double-blind placebo-controlled food challenge is still the gold standard for diagnosis of food allergy. However, predictive DDPs of sIgE levels could be useful for eliminating the need for unnecessary oral food challenges. Although the universal DDPs are in wide use, alternative DDPs of sIgE levels may improve the diagnostic accuracy in food allergy due to differences in clinical and demographical features of patients among countries or races.