Work

Interleukin-6 is a highly prognostic biomarker for POEMS syndrome

Year: 2025

Type: article

Abstract: We performed a retrospective analysis of patients with a new diagnosis of POEMS syndrome seen at Mayo Clinic, Rochester (2011 to 2019) who had IL-6 testing within 90 days of their diagnosis. Of the 117 patients seen during that period, 52 met inclusion criteria. Due to changes in the assay and different laboratories performing the IL-6 assay over the course of the study period, three different upper limits of normal (ULN) for IL-6 were applied based on the laboratory performing the analysis: <17.4 pg/mL from Viracor-IBT Laboratories ( n = 3), <1.8 pg/mL from Mayo Clinic Laboratories ( n = 5), and <5 pg/mL from Quest Diagnostics Nichols Institute, San Juan Capistrano ( n = 44). IL-6 was classified as elevated if above the lab-specific ULN. Hematologic, VEGF and positron emission tomography (PET) responses were assessed as previously described [ 2 , 7 ]. Differences between groups were analyzed using Chi square. Because there are imperfect definitions for progression in POEMS syndrome, we calculated event-free survival (EFS), where an event was defined as disease progression, initiation of new therapy, or death. Planned treatment institution, such as administering maintenance after ASCT, was not considered an event. EFS and OS were assessed using Kaplan-Meier analysis, with survival differences compared using the log-rank test. Univariate analysis evaluated prognostic impact of IL-6, albumin and the Wang score [ 3 , 8 ]. P values less than 0.05 were considered significant. JMP (SAS) was used for statistical analysis. The study was approved by the Mayo Clinic Institutional Review board. Among the 52 patients, 21 (40.4%) had elevated baseline IL-6 levels, (median value 3 times ULN; IQR 1.3, 4.4). The median (range) VEGF plasma levels across the cohort were 342.5 pg/ml (<31–2780) [normal range 31–86 pg/ml]. Elevated IL-6 was associated with male sex, hepatomegaly, ascites, cherry angiomata, low albumin, endocrinopathy, reduced Diffusing Capacity of the Lung for Carbon Monoxide (DLCO), mixed sclerotic/ lytic bone lesions, and higher light chain values (Table 1 ). IL-6 elevation was not associated with VEGF level (data not shown). Upon the formal diagnosis of POEMS syndrome, most patients received systemic therapy including upfront high dose melphalan with autologous stem cell rescue ( n = 18; 35%), lenalidomide or proteosome inhibitor based regimens ( n = 14; 27%), non-transplant alkylator based therapy (cyclophosphamide) ( n = 8; 15.%). One person with associated Castleman’s disease received IL-6 inhibitor, Siltuximab. Localized therapies were implemented in patients with limited bone lesions; radiation ( n = 10; 19.2%) and cryoablation ( n = 1; 1.9%). We found that 29 patients were misdiagnosed as either chronic inflammatory demyelinating polyradiculoneuropathy (CIDP) or solitary plasmacytoma and received treatment prior to their formal diagnosis of POEMS syndrome and VEGF and IL-6 evaluations. The treatments administered prior to diagnosis of POEMS were steroids ( n = 6, 21%) and intravenous immunoglobulin ( n = 13, 46%) or immunosuppressive therapy such as mycophenolate mofetil, azathioprine or cyclophosphamide ( n = 5, 18%). Others received radiation ( n = 3, 11%) for presumed solitary plasmacytoma and one received low dose melphalan ( n = 1, 4%) for possible but not confirmed POEMS. Baseline IL-6 levels did not differ significantly based on prior treatment ( p = 0.12). There was a trend for a higher likelihood of hematologic response (partial response or better) among those with elevated IL-6 (38% versus 23%, p = 0.09), but achievement of at least partial VEGF (PR V ) and PET responses (PR PET ) did not differ ( p = 0.89, p = 0.85). With median follow up of 83 months, 33 patients had events; 27 patients progressed and started a new therapy, while 6 patients died. The median EFS was 41 months (95%CI:16–93). Elevated IL-6 was associated with shorter median EFS (19 months, 95%CI:7.6–41 vs 90 months, 95%CI:17.9 to NR, p = 0.036) (Fig. 1a ). Among those with elevated baseline IL-6 levels, lack of partial heme response (PR H ) or better was associated with inferior median EFS (not reached versus 16 months versus 6.1 months for ≥ PR H , < PR H and not evaluable, respectively, p = 0.085) (Fig. 1b ). For patients with baseline elevated IL-6 levels, t here was no significant interaction between EFS and PR V or PR PET ( ≥ PR V vs < PR V , 23.1 versus 13.3 months, p = 0.14 and ≥ PR PET vs < PR PET , 29.4 vs 7.6 months; p = 0.36). a Event-free survival (EFS) stratified by baseline IL-6 level (normal vs. elevated); b EFS stratified by depth of hematologic response (≥partial response vs. no response) among patients with elevated IL-6; c overall survival (OS) for the full cohort; d OS stratified by baseline IL-6 level. Median OS was not reached; 8-year OS was 85% (Fig. 1c ). Elevated baseline IL-6 was associated with shorter OS (8-year OS of 64.4% and 97%, respectively, p = 0.005) (Fig. 1d ) with a hazard ratio (HR) for death of 16.3 (95%CI 1.98-134.1; p = 0.0096). On univariate Cox proportional hazards analysis elevated IL-6, hypoalbuminemia, and a high Wang prognostic score were significantly associated with worse OS. Although confidence levels overlapped, IL-6 had the strongest effect, with a HR of 17.29 (95% CI: 2.1–142.3, p = 0.008), suggesting a markedly increased risk of mortality. Hypoalbuminemia was also a significant predictor, with an HR of 5.18 (95% CI: 1.28–20.8, p = 0.016), indicating that patients with lower albumin levels had poorer survival. Additionally, a high Wang prognostic score was associated with worse survival (HR 4.88, 95% CI: 1.21–19.6, p = 0.033), reinforcing its prognostic value. Given the small number of events, robust multivariate modeling to determine IL-6’s independence from other prognostic factors could not be performed.

Source: Leukemia

DOI: 10.1038/s41375-025-02659-7