Utility of serum cytokine testing to differentiate complicated common variable immunodeficiency in resource limited settings

Abstract

Background: Common variable immunodeficiency (CVID) is the most common, symptomatic inborn error of immunity (IEI) worldwide. CVID diagnosis requires lymphocyte subset analysis by flow cytometry to delineate risk for autoimmune and inflammatory (AI) disease, known as complicated CVID. In resource-limited settings, reduced access to flow cytometry limits CVID diagnostics.

Objectives: We investigated the utility of serum cytokine testing, as compared to standard-of-care flow cytometry, for the diagnosis of AI disease in IEI and CVID.

Methods: We performed a retrospective review of patients with International Classification of Diseases, Tenth Revision-coded IEI and extracted cytokine levels, tested on a clinically available serum-based multiplex assay of 13 parameters (soluble IL-2 receptor [sIL-2R], IL-1β, IL-2, IL-4, IL-5, IL-6, IL-8, IL-10, IL-12, IL-13, IL-17, TNF-α, IFN-γ). We assessed the association of cytokine levels with AI disease and immunophenotypes using Wilcoxon test or Spearman correlation, statistically adjusted for multiple testing. We compared predictive values of cytokine levels and lymphocyte subsets, measured by the area under the receiver-operating characteristic curve.

Results: In IEI and CVID, higher sIL-2R and IL-10 levels correlated with more AI complications per patient and more severe T-cell immunophenotypes. Composite receiver-operating characteristic curves for sIL-2R and IL-10, as compared to lymphocyte subsets (naive CD4⁺ T cells and class-switched memory B cells), had comparable diagnostic performance for AI disease in patients with IEI and CVID. These findings were validated in an independent IEI patient cohort.

Conclusions: sIL-2R and IL-10 testing had statistically comparable diagnostic performance for complicated CVID as compared to the current standard-of-care using flow cytometry.

Keywords: AI; CVID; CVIDc; Common variable immunodeficiency; IEI; IL-10; autoimmune and inflammatory; autoinflammation; complicated common variable immunodeficiency; cytokines; flow cytometry; inborn error of immunity; resource-limited care setting; soluble IL-2 receptor (sIL-2R); soluble biomarkers.

Fig 1

Elevated serum sIL-2R and IL-10 levels were associated with specific autoimmune and inflammatory end-organ complications in patients with IEI and CVID. AI disease complications as indicated (x-axis) by median cytokine level in serum (y-axis: pg/mL), shown as violin plots with black bar indicating mean cytokine level and dotted line indicating lower limit of normal for sIL-2R (A) and IL-10 (B). Data are from all patients with IEI and any cytokine testing available (n = 201) (left) and all patients with CVID and any cytokine testing available (n = 75) (right). Analysis included all ICD-10 codes meeting criteria for AI disease (reviewed in the Methods) with all statistically significant associations by Wilcoxon test (statistically adjusted for multiple testing) shown for sIL-2R and IL-10; ∗P < .05, ∗∗P < .01, and ∗∗∗P < .001.

Fig 2

Elevated serum sIL-2R and IL-10 levels were associated with an increasing number of autoimmune and inflammatory disease complications per patient with IEI and CVID. Number of AI disease complications per patient (x-axis: 0 = red, 1 = tan, 2 = green, 3 = blue, 4 = purple) by median serum cytokine level (y-axis: pg/mL), shown as violin plots with black bar indicating mean cytokine level and dotted line indicating lower limit of normal cytokine level for sIL-2R (A) and IL-10 (B). Data are from all patients with IEI and any cytokine testing available (n = 201) (left) and all patients with CVID and any cytokine testing available (n = 75) (right). ICD-10 codes meeting criteria for AI disease complication (reviewed in the Methods). Statistically significant associations by Wilcoxon test (statistically adjusted for multiple testing) shown; ∗P < .05, ∗∗P < .01, or as listed.

Fig 3

Elevated serum sIL-2R and IL-10 levels were associated with lower immunoglobulin levels in patients with IEI and CVID. Association by Spearman correlation between cytokine level (x-axis: median, pg/mL in serum) and immunoglobulin level (y-axis: first or median, mg/dL in serum), as indicated. Bolded boxes denote significant associations of P < .05 (statistically adjusted for multiple testing). Spearman correlation coefficient (r) shown for negative (red) and positive (blue) correlations. Data are from all patients with IEI and any cytokine testing available (n = 201) (left) and all patients with CVID and any cytokine testing available (n = 75) (right).

Fig 4

Elevated serum sIL-2R and IL-10 levels were associated with more severe T-cell immunophenotypes by peripheral blood flow cytometry in patients with IEI and CVID. (A) Association by Spearman correlation between cytokine level (x-axis: median, pg/mL in serum) and lymphocyte subset by flow cytometry (y-axis: median, percentage, or absolute count [cell/μL] in blood; regulatory T cells [Treg]). Bolded boxes denote significant associations of P < .05 (statistically adjusted for multiple testing). Correlation coefficient (r) shown for negative (red) and positive (blue) correlations. (B) Spearman correlation between the indicated cytokine level (x-axis: median, pg/mL in serum) and the indicated CD4⁺ T-cell subset by flow cytometry (y-axis: median, percentage, or absolute count [cell/μL] in blood). Dots represent paired labs per unique patient (see Table I). Spearman correlation shown by black line with P value prior to multiple testing and correlation coefficient (r) listed. Data are from all patients with IEI and any cytokine testing available (n = 201) and all patients with CVID and any cytokine testing available (n = 75), as indicated.

Fig 5

Composite sIL-2R and IL-10 levels had comparable predictive value to composite peripheral blood lymphocyte subsets for the diagnosis of AI disease in patients with IEI and CVID. Receiver-operating characteristic (ROC) curves for the diagnosis of AI disease (presence of ≥2 AI disease complications per patient) for cytokines (sIL-2R and IL-10) (A), lymphocyte subsets by flow cytometry in blood (percentage of naive CD4⁺ T cells and percentage of class-switched memory [sm] B cells) (B), or combination of both cytokines and lymphocyte subsets in blood (C), as indicated. Data are generated from the main Beth Israel Lahey Health IEI patient cohort (n = 201 total patients, left panel), the main Beth Israel Lahey Health CVID patient cohort (n = 75 total patients, middle panel), and the Mass General Brigham validation IEI patient cohort (n = 120 total patients, right panel). Individual ROC curves for these parameters are shown in Fig E1. ICD-10 codes meeting criteria for AI disease complication (reviewed in the Methods). There were no statistical differences in AUC measures for the plots shown (P values listed), indicating equal accuracy in the diagnosis of ≥2 AI comorbidities in patients with IEI and those with CVID (CVIDc).