Oral food challenges (OFC) are the gold standard diagnostic for food allergy, but not without limitation. Administering incremental doses every 15-30min differs from a real-world exposure where ingestion occurs at a single episode. Blumchen et al. reported a median time to objective symptoms of 55min (range 5-210min)¹; if doses are given every 15min, this could significantly overestimate the reaction threshold.^{1, 2} This could also occur due to incremental dosing causing transient desensitization.³ With OFC increasingly used to determine starting doses for oral immunotherapy and guide dietary avoidance,⁴ we assessed how clinical thresholds and symptoms at OFC compare to an ingestion more representative of real-world consumption.

Seventeen peanut-allergic adults (median age 24 years, range 18–40) underwent initial double-blind, placebo-controlled food challenge (DBPCFC) to peanut, as part of a clinical trial (TRACE Peanut study; ClinicalTrials.gov Identifier: NCT02665793). Detailed methods are described elsewhere.⁵ Doses were given every 30 minutes (using a water-continuous dessert matrix) according to the following schedule: 3μg, 30μg, 300μg, 3mg, 30mg, 100mg, 300mg and 1000mg of peanut protein (or placebo), until stopping criteria (adapted from PRACTALL consensus criteria) were met.⁵ Participants returned for two further DBPCFC, at 8–12 week intervals later. The first was an ‘abbreviated’ DBPCFC using the same matrix, with the first active dose equivalent to the maximum tolerated dose at baseline DBPCFC (see Figure 1 and Table S1); this was done as a safety measure. Subjects allocated (by computer randomization) to placebo: had two initial placebo doses, prior to active doses (Figure 1). The third DBPCFC used the same abbreviated protocol, but with the appropriate dose given as peanut butter (Kraft Foods) mixed into a soya-based spread (Wowbutter) and eaten as a small 3cm sandwich (Kingsmill 50/50 bread). For all challenges, the dosing interval was 30 minutes, although this could be doubled if symptoms were progressing. Triangle testing demonstrated the suitability of Wowbutter for blinding, and prior tolerance to this was demonstrated in all participants. The study was approved by the NHS Human Research Authority (reference 15/LO/0286), and written informed consent from all participants.

At baseline DBPCFC, the median cumulative eliciting dose (cumED) was 133mg (IQR 83.3–433.3mg) peanut protein; 2/17 patients had anaphylaxis (WAO 2020 criteria). Median cumED at abbreviated challenge was 133mg (IQR 33.3–433.3mg) (Figure 2A). The shift in cumED was not significant (p=0.10, Wilcoxon sign-rank test), and there were no major differences in clinical symptoms observed (Fig S1), with 4/17 having anaphylaxis. Fourteen subjects underwent the third DBPCFC using peanut butter sandwiches (one had too low a cumED for the appropriate dose to be accurately measured, and two declined). Median cumED at this challenge was 433mg (IQR 33.3–1433.3mg), representing a non-significant half-log increase in cumED (p>0.05; Figure 2B); 2/14 had anaphylaxis.

In a systematic review and meta-analysis of peanut-DBPCFC, 69% of peanut-allergic individuals show a shift in cumED over time; in 56%, this is limited to a half-log difference, equivalent to 1 dosing interval with a PRACTALL-based semi-log dosing regimen.⁶ Indeed, Dua et al reported a fall of around 0.5-log (equivalent to 1 dosing increment) at subsequent OFC in these same participants.⁵ Therefore, the non-significant shift in cumED with an abbreviated challenge protocol is entirely consistent with the inherent ‘noise’ in determining cumED at OFC. We undertook a post hoc power calculation; our sample size would have been sufficient to detect a 1-log difference in cumED with at least 90% power, that is greater than that due to the inherent intraindividual variability.

In summary, we did not find a significant difference in either cumED or symptoms following DBPCFC with a 30-minutely incremental dosing protocol, compared to an abbreviated challenge which is more representative of a normal consumption episode. In addition, there was no significant difference in cumED between baseline DBPCFC and a more ‘real-world’ exposure to peanut butter in a sandwich. Therefore, using threshold data from OFC (with 30-minute dosing intervals) is a valid approach to individual allergen risk management. Importantly, the impact of cofactors was minimized due to the nature of the study design. In reality, thresholds will vary due to cofactors among other reasons,⁶ so appropriate caution should be exercised in extrapolating challenge thresholds into clinical advice at an individual patient level.

口服食物挑战 (OFC) 是诊断食物过敏的金标准，但并非没有限制。每 15 至 30 分钟增加一次剂量与现实世界的暴露不同，后者在一次发作时摄入。Blumchen 等人。报告出现客观症状的中位时间为 55 分钟（范围 5-210 分钟）¹；如果每 15 分钟给药一次，这可能会显着高估反应阈值。^{1, 2} 这也可能是由于剂量增加导致瞬时脱敏而发生的。³ 随着 OFC 越来越多地用于确定口服免疫治疗的起始剂量和指导饮食避免，⁴我们评估了 OFC 的临床阈值和症状与更能代表真实世界消费的摄入相比如何。

作为临床试验的一部分（TRACE Peanut 研究；ClinicalTrials.gov 标识符：NCT02665793 ）。详细方法在别处描述。根据以下时间表，每 30 分钟给予⁵剂（使用连续的水甜点基质）：3μg、30μg、300μg、3mg、30mg、100mg、300mg 和 1000mg 花生蛋白（或安慰剂），直至停止标准（适应符合 PRACTALL 共识标准）。⁵ 参与者每隔 8-12 周返回两次 DBPCFC。第一个是使用相同矩阵的“缩写”DBPCFC，第一个活性剂量相当于基线 DBPCFC 的最大耐受剂量（参见图 1 和表 S1）；这是作为一项安全措施进行的。分配（通过计算机随机化）安慰剂的受试者：在活性剂量之前有两个初始安慰剂剂量（图 1）。第三个 DBPCFC 使用相同的缩写方案，但将适当剂量的花生酱（卡夫食品）混合到基于大豆的涂抹酱（Wowbutter）中，并作为一个 3 厘米的小三明治（Kingsmill 50/50 面包）食用。对于所有挑战，给药间隔为 30 分钟，但如果症状进展，这可能会加倍。三角测试证明了 Wowbutter 对致盲的适用性，所有参与者都表现出事先对此的容忍度。该研究得到了 NHS 人类研究机构的批准（参考 15/LO/0286），并获得了所有参与者的书面知情同意。

在基线 DBPCFC 中，中位累积诱发剂量 (cumED) 为 133mg (IQR 83.3–433.3mg) 花生蛋白；2/17 名患者出现过敏反应（WAO 2020 标准）。简短攻击时的中位数 cumED 为 133 毫克（IQR 33.3–433.3 毫克）（图 2A）。cumED 的变化不显着（p = 0.10，Wilcoxon 符号秩检验），观察到的临床症状没有重大差异（图 S1），4/17 有过敏反应。十四名受试者使用花生酱三明治进行了第三次 DBPCFC（一名受试者的 cumED 太低而无法准确测量适当的剂量，两名受试者拒绝了）。本次挑战中 cumED 的中位数为 433mg（IQR 33.3–1433.3mg），表示 cumED 的半对数增加不显着（p > 0.05；图 2B）；2/14 有过敏反应。

在对花生-DBPCFC 的系统评价和荟萃分析中，69% 的花生过敏个体表现出 cumED 随时间的变化；在 56% 中，这仅限于半对数差异，相当于基于 PRACTALL 的半对数给药方案的 1 个给药间隔。⁶事实上，Dua 等人报告说，在随后的 OFC 中，这些相同参与者的下降幅度约为 0.5-log（相当于 1 次剂量增量）。⁵ 因此，使用简化的质询协议时 cumED 的非显着变化与 OFC 确定 cumED 的固有“噪音”完全一致。我们进行了事后功率计算；我们的样本量足以以至少 90% 的功效检测 cumED 的 1-log 差异，这大于由于固有的个体内变异性而导致的差异。

总之，与更能代表正常消费事件的简短挑战相比，我们没有发现 DBPCFC 后 30 分钟增量给药方案的 cumED 或症状有显着差异。此外，基线 DBPCFC 和更“真实”的三明治中花生酱暴露之间的 cumED 没有显着差异。因此，使用来自 OFC 的阈值数据（给药间隔为 30 分钟）是个体过敏原风险管理的有效方法。重要的是，由于研究设计的性质，辅因子的影响被最小化。实际上，阈值会因辅助因素等原因而有所不同，⁶ 因此，在将挑战阈值外推到个体患者水平的临床建议中时，应谨慎行事。