3.4.5 |Studies that include Bos d 8
Nine studies[30-38] have reported on the diagnostic accuracy of Bos d 8 serum levels. The specificity and sensitivity of Bos d8 detection varied across studies (0.79 - 0.98 and 0.34 - 0.90, respectively). In Alessandri’s study[31], diagnostic efficiency was affected by the use of different testing equipment. Using a cut-off point of 0 (kUa/L) (ISAC standard unit), the sensitivity and specificity were 0.54 and 0.81, respectively. However, with a cutoff value of 0.44 (kUa/L) (ImmunoCAP normalized unit), the sensitivity and specificity were 0.82 and 0.63, respectively. In Nieminen’s study [36], the diagnostic sensitivity was higher in 1-2 years old children compared to 3-14 years old children, whose sensitivities and specificities were 0.34 and 0.94, 0.11 and 0.93, respectively.
4 | DISCUSSION
An analysis of CRD accuracy in CMA will certainly promote a better understanding of diagnostic efficacy and contribute to more precise and individualized diagnosis in CMA[39]. Many CRDs on food allergy have been carried out, but only a few studies have focused on milk allergy. A systematic review of the accuracy of CRD for milk allergy was conducted[22], but the results are still unclear due to the limited number of included studies and cases. Based on previous studies, more articles were included in our study. Through a systematic approach, we performed a more comprehensive analysis.
We analyzed the included studies on milk components and found that the diagnostic value of sIgE levels for Bos d 4, 5, 6, and 8 components varied greatly, with most performing poorly overall. Bos d 4 and 8 components consistently showed high specificity (0.78-0.98 and 0.79-0.98, respectively) but variable sensitivity (0.50-0.82 and 0.34-0.90, respectively), while Bos d 5 showed lower specificity (0.58-0.98) than Bos d 4 and 8. The low sensitivity indicates that a negative result for CRD cannot exclude CMA, necessitating oral food provocation. Conversely, the high specificity suggests that the oral food excitation test can be ignored, as the risk of CMA is sufficiently high in this case. Our findings confirm those of a previous DTA systematic review on CRD in milk allergy[22], indicating that sIgE levels against Bos d 4 can provide a highly accurate diagnostic indicator. In the uncuoglu study [38] of the Bos d 5 component, the test had a higher sensitivity (1.0) than specificity (0.70), while in all other studies, specificity was significantly better than sensitivity. These results can be extrapolated to other studies and components analyzed (Table 2), as the determination of individual components with the UniCAP is comparable to the diagnosis of allergy with the oral provocation test, which does not differentiate between different components of the same food. Thus, in the absence of UniCAP detection of IgE against a specific ingredient, a patient may be sensitized to another ingredient of the same food and diagnosed by OFC, resulting in a false-negative result (decreased sensitivity). Conversely, all positive reactions to any component will also be positive in the oral test, resulting in a low number of false positives (increased specificity). However, this is the first review of the diagnostic value of Bos d 6 as a component of cow’s milk-related allergic reactions. It was found to have high specificity but low sensitivity. Previous studies have indicated that Bos d 6 is characterized by high heat stability and cross-reactivity with beef proteins[40], suggesting that children with high levels of Bos d 6 serum IgE may also be allergic to beef. Therefore, caution should be exercised when ingesting beef to prevent severe allergic reactions. In this review, the diagnostic accuracy of the cow’s milk allergy component of the component-resolved diagnosis (CRD) was compared with that of first-line diagnostic tests, including the APT, sIgE, and SPT, as reported in previous systematic reviews. The results varied depending on the diagnostic test. The DTA results for Bos d 4, 5, and 8 for milk were similar to those of the APT, which showed a sensitivity-specificity pair of 44.2% and 86.9%[41]. However, the overall sensitivity of all components was lower than that of sIgE and SPT (87.3% and 87.9%, respectively), and the specificity was higher than that of sIgE and SPT (47.7% and 67.5%, respectively)[42].
Observations were conducted on all studies, with a particular focus on the diagnostic accuracy of the Bos d 4, Bos d 5, Bos d 6, and Bos d 8 components of milk due to the greater number of studies and the significant variability of results. Possible explanations for this variability include 1) Experimental design, which included case-control, prospective cohorts, and retrospective cohorts. Retrospective designs may pose a risk of information bias, leading to artificially maximized results. Additionally, since the oral food provocation test served as the gold standard for all studies analyzed, differences in the design of this test should be considered a source of bias in the results. Only two studies used the DBOFC test as the reference standard[31, 37], while the majority of studies used the OFC (without using double-blind or only single-blind), which carries the risk of interpretation bias and interference of subjective factors in the appearance of symptoms, potentially leading to bias in patient selection. It is also important to mention the heterogeneity due to the use of different versions of devices (ImmunoCAP, ISAC, UniCAP, and 3gAllergy). Standardization of all CRD tests is necessary to ensure comparable results between different tests. The ImmunoCAP ISAC is considered by the World Health Organization as a complementary diagnostic tool to the component-resolved diagnosis[43]. 2) Age of the patient; our results show that there is an effect of age on diagnostic efficacy and cutoff values. In our study, children aged 1-2 years had smaller optimal cutoff values and higher diagnostic efficacy compared with children aged 3-14 years, suggesting that younger children are more sensitive to smaller cutoff values. This finding is consistent with previous studies[44-47]. However, this difference still lacks statistical power because of the limited number of studies included and the relatively small advantage. Moreover, negative results were noted in our pooled analysis. Therefore, to be applied in clinical practice, prospective studies need to be designed to evaluate the use of CRD in diagnostic tests for suspected milk allergy, and the cutoff values for optimal diagnostic efficacy at different ages need to be studied to reach a consensus.
In the scientific literature, there is a growing body of evidence demonstrating the utility of CRD in the workup of milk allergy, suggesting that the sIgE of milk allergen components can aid in the diagnosis of milk allergy. Using a systematic approach, we analyzed the evidence for the diagnostic accuracy of CRD for milk allergy within a range, filling an important research gap in this area. The strengths of our study include a comprehensive methodology, the use of a highly sensitive search strategy, no language restrictions, and the involvement of multiple countries, databases, and clinical trial registries, which allowed for a thorough literature search. Our inclusion criteria were based on clinically relevant information from carefully selected studies, as well as guidelines from reputable organizations such as the European Academy of Allergy and Clinical Immunology and the Food Allergy [48] and Anaphylaxis Guidelines Group[49]. Furthermore, the internal validity of the studies included in our review was strong, as at least 50% of the participants used 0FC as the reference standard.
Our systematic review has several limitations. First, due to the limited and relatively insufficient number of articles included in the study, a larger dataset or a more scientific approach would be needed to more accurately assess diagnostic accuracy. In addition, the included population was exclusively children and there is a lack of studies in adult populations; therefore, the role of milk CRD in adults remains to be investigated. Finally, and most importantly, there was a great deal of heterogeneity in our analyses, constituting an important obstacle to meta-analysis. This may be due to the great variation in study design: a) Differences in the level of the cutoff value chosen in different studies (e.g95% vs. 100%) may significantly change the suggested critical value. b) Methodological quality: type of equipment or test, use of OFC blinding.
5 | CONCLUSIONS
The findings of this review indicate that the measurement of sIgE levels for the Bos d 4, Bos d 6, and Bos d 8 milk components is highly specific but not very sensitive for diagnosing cow’s milk allergy in children. The implementation of CRD for diagnosing CMA in children may potentially decrease the need for OFC.