Background/Purpose: The interferon gamma (IFN-G) release assay (IGRA) estimates probability of latent TB infection (LTBI) based on IFN-G released by whole blood after 18h exposure to TB antigen, mitogen or no stimulation (negative control, NC). In a small percentage of individuals, the NC condition results in markedly elevated IFN-G levels, indicating spontaneous IFN-G production (SIP). Here we examine the clinical, laboratory and demographic features of SIP in a large, academic health care system. We hypothesized that SIP is associated with select immunologic and infectious diseases.

Methods: Analysis of the distribution of over 15,000 IGRA NC results obtained between 2010 and 2015 revealed an abnormally distributed population (n = 108) with values ≥ 5 SD above median (≥ 3 IU/mL). Using this discovery cohort, we catalogued all clinical diagnoses noted in the electronic medical record ± 6 wks of IGRA date. Next, we queried a de-identified clinical data repository (DCDR) for occurrences of related ICD9/10 codes and demographic data in all individuals undergoing IGRA testing (~11,820), comparing HIGH NC group (any NC result ≥ 3, n = 83) vs. LOW NC group (no NC ≥ 3, n = 11,740) with univariate (Chi square and Fisher tests) and multivariate (logistic regression) analyses.

Results: Several ICD9/10 codes were significantly (p ≤ 0.001) enriched in the HIGH vs. LOW groups: HIV (27.7% vs. 16.0%), LTBI (22.9% vs. 7.6%), SLE (7.2% vs. 1.3%) and hemophagocyctic lymphohistiocytosis (HLH) (3.6% vs. 0.1%). After controlling for race and other enriched ICD9/10 codes, these relationships remained significant: HIV (OR 2.16, 95%CI 1.31 – 3.58), latent TB (OR 3.9, 95%CI 2.2-6.7), SLE (OR 3.79, 95%CI 1.56-9.19) and HLH (OR 31.8, 95%CI 8.7-116.2). These relationships were also observed analyzing NC levels as a continuous variable. Univariate sub-analysis of subjects with ICD9/10 codes corresponding with SLE suggested a correlation between HIGH NC values and coincident low serum complement levels (p = 0.0356), but not cytopenias or specific autoantibodies.

Conclusion: In individuals undergoing IGRA testing in a large academic health care system, SIP was associated with a limited set of immunologic or infectious diseases. All of the identified diseases are known to involve IFN-G activation. For SLE, SIP may identify a subset of patients with high disease activity and have therapeutic implications. For a very rare disease like HLH, SIP may have diagnostic utility. The importance of SIP in HIV and LTBI remains to be determined.

ACR Meeting Abstracts - https://acrabstracts.org/abstract/correlates-of-spontaneous-cytokine-production-in-individuals-undergoing-interferon-gamma-release-assay-testing/