71 Hyperferritinemia in patients with pneumonia Secondary analysis with hyperferritinemia defined using 250 ng/ml as cutoff value In a secondary analysis plasma ferritin ≥250 ng/ml (in most routine settings considered the upper limit of normal [6, 21]) was used as cut-off. Ninety CAP patients (52%) and 7 non-infectious controls (14%) had plasma ferritin levels ≥250 ng/ml (Figure 1). CAP patients with ferritin ≥250 ng/ml or <250 ng/ml showed similar demographics. With regard to chronic comorbidities, in CAP patients with ferritin ≥250 ng/ml hematological malignancy, chronic renal disease and immune suppression were more frequent, and COPD less frequent (Supplementary Table 2). Severity scores (PSI, MEWS and qSOFA) and clinical outcomes did not differ between patient groups. Differences in host response biomarker levels between patients with ferritin levels ≥250 ng/ml and <250 ng/ml were largely similar to those in the primary analysis using a cut-off of 500 ng/ml (Supplementary Figures 1-3). Discussion Ferritin is an acute-phase protein elicited by inflammation and infection. Amongst acute conditions associated with hyperferritinemia sepsis is the most common [3]. Knowledge of the frequency of hyperferritinemia in patients with acute infection is almost exclusively derived from ICU settings [2–4]. We here sought to assess the incidence of hyperferritinemia in patients with CAP hospitalized to a general ward and to determine its association with disturbances in key host response pathways implicated in the immunopathology of pneumonia and sepsis. We considered this of interest, as circulating ferritin levels have been and are used as inclusion criterion in trials evaluating immunomodulatory therapies in acute infections and may assist in identifying patients who might benefit from such therapies [13, 23] (clinicaltrials.gov identifiers NCT04530578, NCT04341675, NCT04443881, NCT04990232). We performed two analyses, stratifying patients according to cut-off ferritin levels of ≥500 ng/ml (used in studies in patients with CAP caused by SARS-CoV2: NCT04530578, NCT04341675, NCT04443881) or ≥250 ng/ml (around the normal reference value in most laboratories). While these cut-off levels are arbitrary, both analyses yielded highly similar results, i.e., hyperferritinemia was associated with stronger aberrations in key host response pathways including systemic inflammation, neutrophil activation, cytokine release, endothelial cell activation and dysfunction, and coagulation activation. Several trials use a ferritin level of 500 ng/ml or higher to select patients for antiinflammatory therapy (clinicaltrials.gov identifiers NCT04530578, NCT04341675, NCT04443881). The current study indicates that differences in host response biomarker levels using a ferritin cut-off level of 250 ng/ml were largely similar to those using a cutoff of 500 ng/ml. Notably, both cut-off levels are arbitrary and strict recommendations regarding the optimal ferritin concentration for identification of patients who might benefit from anti-inflammatory therapy cannot be made based on our observational data. Hyperferritinemia was associated with trends toward a longer length of hospital stay and a more frequent transfer to the ICU. Larger cohorts are needed to analyze whether hyperferritinemia in CAP patients admitted to a general ward can be useful in predicting these clinically relevant outcomes. 4
RkJQdWJsaXNoZXIy MTk4NDMw