Search for a command to run...
Oral immunotherapy (OIT) is currently the major curative approach for food allergy. Although most individuals receiving OIT develop desensitisation during active treatment, sustained unresponsiveness (SU) or remission, defined as the ability to tolerate food allergens after OIT completion, occurs only in a minority of recipients [1]. Therefore, to predict SU prior to OIT initiation is critical for stratifying patients and advancing precision medicine approaches in food allergy. Here, we present a narrative review with a structured literature search, interrogating recent advances in predicting SU after OIT and comparatively assessing proposed baseline predictors. We exhaustively surveyed publications on PubMed with OIT clinical studies/trials related to predicting SU with baseline profiles. Literature search was performed using search terms “((allergy) OR (food allergy)) AND ((oral immunotherapy) OR (oral immune therapy) OR (OIT)) AND ((sustained unresponsiveness))” and “((allergy) OR (food allergy)) AND ((oral immunotherapy) OR (oral immune therapy) OR (OIT)) AND ((remission))”. Our search end date is 23rd February 2026. Aforementioned search terms retrieved 324 papers. After screening their titles and abstracts in detail, 22 papers with in-depth analysis for SU/remission (referred to as SU hereafter for simplicity) in food allergy OIT were found for further analysis, encompassing 11 clinical studies across 5 common food allergens: peanut, egg, wheat, milk, and fish. Among these reviewed studies, SU predictors can be broadly categorised as: immune parameters, gut microbiota profiles, and clinical measurements. Immune-based markers were most commonly evaluated. Allergen-specific IgE, particularly epitope-specific IgE (e.g., Ara h 2 for peanut), generally exhibited stronger predictive performances, with significant differences reported between the SU and no SU groups, and comparatively higher areas under the curves (AUCs) in receiver operating characteristic (ROC) analyses. Other immune-related parameters, including T cell activation profiles and basophil responsiveness to allergens, showed weaker predictive performance. For example, baseline basophil activation and sensitivity assay results could not predict SU outcomes in [2]. Gut microbiota-based predictors exhibited modest predictive capacity. Özçam et al. reported that microbiota phylogenetic diversity and metabolomic features differentiated SU vs. no SU groups with a lower ROC AUC (0.712) [3]. Clinical measurements, including age, skin prick test wheal size, and allergen cumulative reactive dose, also showed associations with SU, but their predictive performance appeared less potent compared with allergen-specific IgE measures, which seemed to be the best predictors across all reviewed studies. Importantly, applications of advanced mathematical methods like machine learning (ML), where complex data are available, can substantially improve predictive power. In the POISED trial, an ML-based model incorporating allergen- and epitope-specific IgE profiles yielded an ROC AUC of 0.99 [4]. Statistical measurements to evaluate predictive performances vary in the aforementioned publications. To enable cross-sectional comparisons, we re-analysed 3 studies where source data were publicly available using ROC analysis with the pROC (v1.18.0) package based on each parameter and compared their ROC AUCs. With this approach, we confirmed that allergen- and epitope-specific IgE consistently exhibited the best predictive performances, reaching ROC AUCs up to 0.92, in contrast to all other parameters with ROC AUCs below 0.75 (Table 1). IMPACT trial 134w OIT, 26w avoidance POISED trial PN-0: 2y OIT, 13w avoidance timepoint OPIA trial 1y OIT supplemented with HAMSB/LAMS, 6w avoidance (HAMSB, LAMS arms combined for analysis) We also carried out similar comparisons using our own Oral Peanut Immunotherapy with butyrate Adjuvant (OPIA, ACTRN12617000914369) clinical trial cohort [5, 6]. The OPIA study was a randomised double-blinded, placebo-controlled trial led by our team as previously described [5, 6]. The OPIA trial contained 3 arms: no OIT, 12-month OIT with butyrylated high amylose maize starch (HAMSB) as an adjuvant, and 12-month OIT with placebo low amylose maize starch (LAMS). For the current analysis, we focused on the 2 OIT treatment arms (HAMSB and LAMS) and combined them for comparisons. Analyses on the OPIA trial cohort found that clinical measurements like peanut eliciting and cumulative reactive doses seemed to outperform immune parameters like peanut-reactive T cell functional profiles in predicting SU (Table 1). Collectively, we here present the most comprehensive synthesis to date of state-of-the-art advances in predicting SU following OIT. Overall, allergen- and epitope-specific IgE consistently emerged as the strongest predictor of SU, outperforming other immune- and microbiota-related parameters. Clinical measurements, including skin prick testing and food challenges, also exhibited robust predictive power. Importantly, these findings were confirmed through comprehensive reanalyses of 3 publicly available study datasets and in our OPIA clinical trial, although only a subset of representative OPIA data is included here, with additional analyses for more baseline variables (e.g., age and baseline demographics) to be reported separately. As a narrative review, this study is inherently susceptible to bias related to study selection and the nature of the analytical approach. More comprehensive systematic reviews and meta-analyses are therefore warranted to address these limitations. Accurate prediction of SU is of substantial clinical and research relevance. However, feasibility, cost and scalability are also of critical consideration for implementation in routine practice and for research. In this regard, allergen-specific IgE measurement seems to be superior, because it exhibits robust predictive power and is cost-effective and already available. Importantly, the predictive performances of allergen-specific IgE may be further strengthened when integrated with complementary clinical variables, as suggested by prior studies [4]. In addition, treatment-related factors during OIT, such as adherence (e.g., missed doses) and dosing-related adverse outcomes, are likely to influence OIT outcomes as well [7]. Future predictive frameworks may therefore benefit from incorporating both baseline immune markers and longitudinal treatment variables. Currently, most OIT protocols apply limited patient stratifications to fine-tune their dosing strategies and/or treatment durations. Our findings suggest that baseline measurements like allergen- and epitope-specific IgE profiles could facilitate more tailored approaches to move the field towards precision-guided OIT, enabling individualised risk stratification and enhancing SU outcome prediction. Concept and design and drafting of the manuscript: Duan Ni and Ralph Nanan. Acquisition, analysis and interpretation of data: Duan Ni and Ralph Nanan. Both authors read and approved the final manuscript. The OPIA trial was funded by a National Health and Medical Research Council Australia Project Grant (NHMRC 1104134). D.N. was supported by the Norman Ernest Bequest Fund. We acknowledge the contributions of the OPIA study group and help with data management by Katherine Thomson, Matthew Ward, and Ella Ward. We thank the children and their families for participating in the OPIA trial. Open access publishing facilitated by The University of Sydney, as part of the Wiley - The University of Sydney agreement via the Council of Australasian University Librarians. This work was supported by the National Health and Medical Research Council (NHMRC 1104134) and Norman Ernest Bequest Fund. The OPIA trial is funded by a National Health and Medical Research Council Australia Project Grant (NHMRC 1104134). D.N. was supported by the Norman Ernest Bequest Fund. The OPIA trial was registered on 22 June 2017 in the Australian New Zealand Clinical Trials Registry as ACTRN12617000914369. The study was approved by the Human Research Ethics Committee of the Sydney Children's Hospital Network (HREC/16/SCHN/372). Written informed consent was obtained from parents/guardians, and assent was obtained. The authors declare no conflicts of interest. The data that support the findings of this study are openly available in GitHub at https://github.com/Nidane/SustainedResponsivenessPredictor.Online Repository data include a list of publications reviewed in current study and a table summarizing the state-of-the-art advances in predicting SU outcome from OIT with baseline measurements based on 22 papers across 11 clinical trials/studies.