Salivary IL‐8 and sTREM‐1 in chronic urticaria: A diagnostic test accuracy study

Positive predictive value for the combination was 0.727, higher than for sTREM-1 (0.524) or IL-8 (0.600) alone. Negative predictive value was good for the combination (0.824), sTREM-1 (0.875) and IL-8 (0.833). Median values for the basic diagnostic tests in patients with CU were in the normal range for ESR (8.0 mm/h), plasma CRP (2.0 mg/L), C3 (1.285 g/L), C4 (0.27 g/L) and vitamin D (69.5 nmol/L). There were no significant correlations between salivary IL-8 or sTREM-1 and plasma biomarkers in the patient group using Spearman rank cor - relation testing. Although the number of participants enrolled within our timeframe, 43 patients and 34 controls, is less than the 50 pos - itive and 50 negative samples recommended for evaluating the specificity of a test, 9 our data provide the first report of salivary biomarkers with good PPV and NPV in CU not detected by blood tests.


R E S E A R C H L E T T E R Salivary IL-8 and sTREM-1 in chronic urticaria: A diagnostic test accuracy study
To the Editor, Chronic urticaria (CU) is defined as the occurrence of wheals, angioedema or both for more than 6 weeks. 1 CU can present with daily/almost daily signs and symptoms or an intermittent/recurrent course. 1 CU is further classified, based on the role of triggers, as inducible (CIndU) or spontaneous (CSU).CSU, of known or unknown cause, accounts for 60%-90% of all CU cases, has a point prevalence of 0.02%-2.7%and is most prevalent in women. 2 CSU may be caused by type I autoimmunity, with IgE autoantibodies recognising a number of autoantigens and type IIb autoimmunity with IgG autoantibodies recognising high-affinity FcεR1 receptors and IgE.These mast cell-activating autoantibodies are likely key drivers of CSU pathogenesis. 2First-line treatment is with H1-antihistamines and second-line treatment is with omalizumab (anti-IgE). 3Blood-based diagnostic tests for CSU 1,2 are invasive.
We, therefore, investigated salivary levels of the neutrophil che- Saliva samples were collected using the Salivette® system (Sarstedt, Leicester, UK), processed, supplemented with protease inhibitors, fully anonymised and frozen at −80°C on the same day.
However, sTREM-1 analysis found that 59%, 56% and 20% of the samples had levels below the LOD in the healthy controls, OMAtreated and OMA-untreated CU groups, respectively.
For samples with sTREM-1 below the LOD, data were transformed to replace values <LOD with LOD/(sqrt(2)). 7The Shapiro-Wilk test showed that these censored data sets were non-normally distributed.
Censored patient data was grouped according to omalizumab treatment.Kruskal-Wallis ANOVA showed a significant (p = .012)difference between the groups (Figure 1C).Dunn's multiple com- omalizumab-untreated patients and 31 healthy controls with matched IL-8 and sTREM-1 data was performed in R (R Foundation, Vienna, Austria) (Figure 1D).
Both biomarkers were predictive of CU, but the combination did not improve the ability of the tests to confirm the diagnosis.The area under the curve (AUC) for sTREM-1 was 0.749, and with a cutoff value of 62.1 pg/mL the sensitivity was 0.786 and the specificity was 0.677.For IL-8 the AUC was 0.74, and with a cut-off value of 141.9 pg/mL the sensitivity was 0.643 and the specificity was 0.806.
For the combination the AUC was 0.73, the sensitivity was 0.571, and the specificity was 0.903.
• Salivary sTREM-1, but not IL-8, levels may be indicative of response to omalizumab therapy.
moattractant interleukin-8 (IL-8) and the soluble form of triggering receptor expressed on myeloid cells-1 (sTREM-1), a diagnostic and prognostic biomarker of bacterial infection, 4 a potential trigger in CU, as non-invasive diagnostic tests in patients with CU.This pilot case-control study was approved by the Research Ethics Committee, Northern Ireland, (14/NI/1089).Informed written consent was obtained from all participants randomly invited between September 2016 and September 2017 to participate in the study from patients visiting the Immunology and Allergy Clinic at University Hospital Southampton.A detailed history and physical examination confirmed a diagnosis of CU.Patients <18 years, >80 years or on any NSAID were excluded.Our dataset includes 43 with CU for >6 weeks (70% women) and 34 healthy controls (47% women).Twenty-eight CU patients (65%) unresponsive to qds antihistamines were treated with omalizumab.Patient CRP, ESR, C3, C4 and vitamin D values were analysed using routine automated methods.Healthy controls had no known history of CU.

F I G U R E 1
(A) Salivary IL-8 in omalizumab-untreated (n = 14), omalizumab-treated patients (n = 25) with CU and healthy controls.The graph shows individual data points and the horizontal bars represent median values.*p = .0184,***p = .0001compared with healthy controls.(B) sTREM-1 contingency data plot of positive test results for OMA-treated and OMA-untreated CU patients and healthy controls.A significantly (p = .015)higher proportion (80%) of samples from OMA-untreated patients were sTREM-1-positive compared to healthy controls (41%) or OMA-treated CU patients (44%, p < .049).There was no significant difference in the proportion of sTREM-1 positive samples in OMA-treated CU patients and healthy controls (p > .999).(C) Censored sTREM-1 data by treatment group as individual data points.Horizontal bars represent median values.**p = .009comparing samples from OMA-untreated patients with healthy controls.There was no significant difference between controls and OMA-treated patients.(D) CombiROC analysis of salivary sTREM-1 and IL-8, alone and in combination, in CU patients not treated with omalizumab.CU, chronic urticaria; IL-8, interleukin-; sTREM-1, soluble triggering receptor expressed on myeloid cells-1.