Interferon-induced protein 35 (IFI35) is associated with COVID-19 severity and clinical outcomes

Abstract

Excessive activation of the immune system by damage-associated molecular patterns (DAMPs) contributes to COVID-19 severity. Interferon-induced protein 35 (IFI35) is a DAMP-related interferon-stimulated gene recently proposed as a biomarker of hyperinflammation. We investigated the association between IFI35 serum levels, gene expression, and COVID-19 severity in 430 hospitalized patients (214 critical and 216 severe) and 112 convalescent controls. Serum IFI35 levels were quantified using ELISA, and IFI35 mRNA expression in PBMCs was assessed using quantitative PCR (qPCR). IFI35 levels were significantly higher in critically ill patients (median 1,003.6 pg/mL) than in severe cases (867.6 pg/mL; p = 0.001) and controls (798.3 pg/mL; p < 0.0001). Gene expression analysis showed a similar pattern (p = 0.014). Higher IFI35 levels were also associated with adverse clinical outcomes, including mortality, invasive mechanical ventilation, acute kidney injury, and cardiac arrest (all p < 0.05). In receiver operating characteristic analysis, adding IFI35 to a model including C-reactive protein and D-dimer improved discriminative performance for mortality. In survival analysis, IFI35 > 1,475 pg/mL was associated with reduced survival in this cohort (p = 0.0062). These findings identify IFI35 as a promising prognostic biomarker associated with COVID-19 severity and adverse outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-46121-9.

Introduction

The Coronavirus Disease 2019 (COVID-19) pandemic has posed unprecedented challenges to global public health. Long-standing neglect and underinvestment in healthcare systems have strained medical services, particularly in developing countries^1,2. Moreover, factors such as climate change, wildlife trade, urbanization, and increasing global interconnectivity are expected to raise the likelihood of new zoonotic coronaviruses emerging in the coming decades. Although substantial progress has been made in understanding SARS-CoV-2 biology and clinical management, many aspects of its pathogenic mechanisms remain unclear. A deeper understanding of viral immunopathogenesis is essential to identify novel therapeutic targets and improve preparedness for future pandemics³.

The inflammatory response is an essential component of an effective immune response, and in the immunopathology of COVID-19, inflammatory cytokines play a central role^4,5. Excessive immune activation driven by Damage-Associated Molecular Patterns (DAMPs) can amplify the production of pro-inflammatory cytokines, promote inflammatory cell recruitment, and trigger tissue damage−ultimately contributing to disease severity⁶. In a subset of patients, SARS-CoV-2 infection induces a sustained hyperinflammatory state, commonly referred to as a “cytokine storm”, which is associated with poor prognosis, multi-organ dysfunction, and increased mortality^4–6. Given this, the identification and characterization of DAMPs may contribute to the regulation of inflammatory responses, promoting tissue protection and homeostasis, and representing an emerging approach in clinical practice⁷.

Interferon-induced protein 35 (IFI35) is a 35 kDa interferon-stimulated protein originally identified through screening of human cDNA libraries from HeLa cells⁸. IFI35, recognized as a DAMP, has emerged as a relevant mediator of cytokine storms during viral infections and plays a central role in modulating the innate immune response by regulating the expression of interferon-stimulated genes (ISGs)^9,10. Experimental evidence suggests that targeting IFI35 may mitigate tissue damage caused by viral infections^11–15. Notably, a recent study from China reported that serum IFI35 levels in patients with SARS-CoV-2 and influenza virus infections correlated with disease severity, highlighting its potential as a biomarker and a possible therapeutic target in experimental settings¹⁵.

Given the gaps in knowledge about IFI35’s role in innate antiviral immunity−particularly in SARS-CoV-2 infection−investigations into its expression and clinical associations are essential for identifying novel diagnostic and therapeutic strategies. Here, we examined whether IFI35 expression and circulating protein levels are associated with clinical severity in COVID-19 and explored their relationship with clinical parameters linked to poor outcomes.

Methods

Ethical considerations

This study was approved by the Ethics Committee of the Hospital das Clínicas of the Federal University of Pernambuco (HC/UFPE) under the number CAAE: 36613520.0.0000.5640. Informed consent was obtained from all participants and/or their legal guardians, in accordance with the requirements of the National Research Ethics Commission (CONEP). It was conducted in accordance with the provisions of the Declaration of Helsinki and the Good Clinical Practice guidelines.

Study population

This retrospective cross-sectional study included 430 individuals over 18 years of age, irrespective of sex, who had a confirmed diagnosis of COVID-19 through molecular testing (RT-qPCR) or immunological assays (rapid antigen or antibody-based tests, IgG/IgM). Hospitalized patients were recruited between August 2020 and July 2021 from two reference centers for COVID-19 treatment in Petrolina, Northeast Brazil: the University Hospital of the Federal University of the São Francisco Valley (HU-UNIVASF) and the Petrolina Municipal Field Hospital.

The control group comprised convalescent individuals who had symptomatic mild disease with laboratory-confirmed COVID-19 and did not require hospitalization. These individuals were recruited within 12 months post-infection through invitations disseminated via social media, radio, and television. No active SARS-CoV-2 infection was present at the time of recruitment.

Hospitalized patients were categorized based on disease severity at the time of admission. Individuals were classified as critical if they presented with severe pulmonary involvement requiring high-flow oxygen therapy, invasive or non-invasive mechanical ventilation (CPAP, BiPAP, or intubation), septic shock, or organ failure necessitating ICU admission. Those who developed pneumonia requiring low-flow oxygen therapy or who were admitted to general medical wards were classified as severe.

Patient classification and clinical data were retrospectively retrieved from electronic medical records, while control group data were obtained via structured questionnaires.

Sample collection and processing

Samples were collected within 24 h of hospital admission for the patient group, while samples from the control group were obtained during recruitment. Blood samples from both groups were collected via vacuum venipuncture and sent to the Multi-User Research Laboratory (LAMUPE) of HU-UNIVASF. The samples were then centrifuged at 3500 rpm for 10 min to separate serum. The collected material was transferred into cryogenic tubes (Greiner Bio-One) and stored at – 80 °C until IFI35 measurement using commercial kits.

IFI35 quantification

IFI35 was measured by enzyme-linked immunosorbent assay (ELISA) in the serum of individuals included in the study, using the commercial Human IFI35 (Interferon-induced 35 kDa protein) ELISA Kit™ FineTest^®, following the manufacturer’s instructions. This kit is based on a double-sandwich ELISA antibody detection method, and the microplate used is pre-coated with anti-IFI35 antibody. The assay was pre-validated by the manufacturer, and a calibration curve was created for each assay using the standard provided by the kit. An automatic microplate washer (Asys Atlantis^®) was used for washing the microplates according to the specifications of the plate and kit. Absorbances were measured at 450 nm using a Multiskan FC microplate reader (Thermo Fisher). Data were analyzed using a four-parameter logistic regression model.

PBMC isolation and gene expression analysis

PBMCs were isolated using a Ficoll density gradient (Ficoll-Paque™ PLUS, Cytiva, Ref.: 17144002), centrifuged at 1500 rpm for 30 min at 20 °C, and resuspended in Trizol (Life Technologies, Ref.: 15596026) for subsequent RNA extraction. Due to the unavailability of PBMC samples from the mild COVID-19 group, we used samples from healthy blood donors from the same geographical region. Importantly, serum levels of IFI35 did not differ significantly between individuals with mild COVID-19 and healthy blood donors (p = 0.9) (Supplementary Figure S1). A total of 40 individuals were included in this analysis based on sample availability: 6 from the severe group, 26 from the critical group, and 8 healthy blood donors.

Total RNA was extracted from PBMC samples using the TRIzol reagent following the manufacturer’s instructions. cDNA synthesis was performed using the High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems™, Thermo Fisher Scientific, Waltham, MA, USA). Relative gene expression analysis was conducted using the GoTaq Probe qPCR Master Mix (Promega, Madison, WI, USA) and TaqMan Gene Expression Assays (Thermo Fisher Scientific, Waltham, MA, USA) with the following probes: IFI35 (P03667963) and ACTB (2149059). qPCR was performed on a QuantStudio 5 Real-Time PCR System (Applied Biosystems™, Thermo Fisher Scientific, Waltham, MA, USA). The ΔΔCt method was applied to determine relative gene expression, with ACTB (β-actin) serving as the endogenous control for normalization. ACTB was selected as the reference gene based on its stable expression across study groups.

Statistical analysis

Statistical analyses were performed using JASP (version 0.18.3.0) and RStudio^® (version 2023.06.0 + 421). Graphs were generated using GraphPad Prism (version 8.0, GraphPad, San Diego, CA, USA). Categorical data were expressed as absolute frequencies and percentages, while continuous variables were presented as medians with interquartile ranges (IQR).

The Kolmogorov–Smirnov test was used to assess the distribution of continuous variables. For comparisons between two groups, the Mann–Whitney U test was applied to non-parametric variables, and the Student’s t-test was used for parametric variables. Categorical variables were compared using the Chi-square test. For comparisons involving more than two groups, one-way ANOVA was applied for parametric data, and the Kruskal–Wallis with Dunn’s correction for multiple test was used for non-parametric data.

A multivariable multinomial logistic regression was employed to assess whether IFI35 levels independently predicted disease severity, after adjusting for clinical covariates that were significantly associated with severity in the univariate analysis. Critical disease was used as the reference category for the dependent variable to allow direct comparison between severity groups. IFI35 was analyzed as a continuous variable and rescaled per 100 pg/mL increase to improve interpretability. Model assumptions were assessed, and multicollinearity was evaluated using variance inflation factors (VIF). No evidence of significant collinearity was observed (all VIF values < 2). Results were reported as odds ratios (OR) with corresponding 95% confidence intervals (CI).

To evaluate the incremental prognostic value of IFI35, receiver operating characteristic (ROC) curve analysis was performed to compare the discriminative performance of models including conventional laboratory biomarkers (C-reactive protein and D-dimer) with models additionally incorporating IFI35.

To evaluate the association between IFI35 levels and mortality, the optimal cutoff point was determined using the maxstat package in RStudio. Survival analysis was then performed using the Kaplan–Meier method, with group comparisons assessed by the log-rank test. Clinical variables with a p-value < 0.05 in univariate analysis were included in a Cox proportional hazards regression model (backward stepwise method) to evaluate the independent association of IFI35 levels with the risk of death and to estimate hazard ratios (HR). Statistical significance was defined as p < 0.05.

Results

The study included a total of 430 hospitalized patients with COVID-19, 216 classified as severe and 214 as critical, along with 112 convalescent individuals serving as controls. The baseline characteristics of the study population are summarized in Table 1. Age increased significantly with disease severity, with median ages of 35 years in the control group, 51 years in the severe group, and 56 years in the critical group (p < 0.0001). The control group had a significantly lower proportion of males (39%) compared to the severe (62%) and critical (61%) groups (p < 0.0001). Critical patients showed a higher prevalence of comorbidities, including diabetes, hypertension, obesity, chronic heart disease, and chronic kidney disease (p < 0.05). All deaths occurred in the critical group, with a mortality rate of 40.18% (p < 0.0001). Additionally, the length of hospital stay was significantly longer in the critical group compared to the severe group (median of 13 days vs. 4 days, p < 0.0001) (Table 1).

Table 1.

Baseline characteristics and clinical evolution of patients included in the study classified according to the severity of COVID-19 and control group.

	Control (n = 112)	Severe (n = 216)	Critical (n = 214)	p a
Demographic
Age (median, [IQR])	35.0 [28.0-43.5]	51.0 [42.0-61.2]	56.5 [44.0–67.0]	< 0.0001
Male gender, n (%)	44 (39.28)	134 (62.03)	132 (61.68)	< 0.0001
Comorbidities
Diabetes mellitus, n (%)	04 (3.57)	52 (24.07)	75 (35.04)	< 0.0001
Hypertension, n (%)	09 (8.03)	97 (44.90)	116 (54.20)	< 0.0001
Obesity, n (%)	04 (3.57)	39 (18.05)	56 (26.16)	< 0.0001
Chronic heart disease, n (%)	0	04 (1.85)	11 (5.14)	0.016
Chronic kidney disease, n (%)	0	02 (0.92)	16 (7.47)	< 0.0001
Asthma, n (%)	07 (6.25)	04 (1.85)	08 (3.73)	0.12
Cancer, n (%)	0	01 (0.46)	05 (2.33)	0.08
COPD, n (%)	0	06 (2.77)	10 (4.67)	0.06
Death, n (%)	0	0	86 (40.18)	< 0.0001
Length stay, days (median, [IQR])	-	04 [3–6]	13 [7–22]	< 0.0001 b

^aIn bold are values statistically significant. ^bSevere vs. Critical. COPD: Chronic obstructive pulmonary disease; IQR: Interquartile Range.

As shown in Fig. 1, IFI35 serum levels increased with disease severity, with critically ill patients exhibiting significantly higher levels (median: 1,004 pg/mL) compared to those with severe (median: 867 pg/mL, p = 0.0009) and the control group (median: 798 pg/mL, p < 0.0001) (Fig. 1).

Fig. 1.

Violin plots showing serum IFI35 concentrations (pg/mL) in COVID-19 patients stratified by disease severity. Statistical significance was assessed using Dunn’s multiple comparison test, and adjusted p-values are indicated. The analysis included 430 hospitalized patients (216 classified as severe and 214 as critical) and 112 convalescent controls.

To address potential demographic confounding, we performed a stratified analysis by sex. Among women, IFI35 levels differed significantly across severity groups (Kruskal–Wallis p < 0.0001), with higher levels observed in critically ill patients than in the control group (p < 0.0001), whereas no significant difference was observed between critical and severe cases (p = 0.15). Among men, IFI35 levels also differed across severity groups (Kruskal–Wallis p = 0.01), with higher levels observed in critical ill patients than in severe cases (p = 0.006), but not in the control group (p = 0.59). These findings suggest that the association between IFI35 and disease severity is observed in both sexes, although with some variability in pairwise comparisons (Supplementary Figure S2).

This trend was also observed in the relative quantification of gene expression, where critically ill individuals exhibited significantly higher IFI35 expression levels compared to those with severe disease and healthy blood donors (p = 0.014) (Supplementary Figure S3).

To determine whether IFI35 levels independently predict COVID-19 severity, a multivariable multinomial logistic regression analysis was performed, incorporating baseline clinical variables that were significantly associated with severity in the univariate analysis. Using critical disease as the reference category, higher IFI35 levels (per 100 pg/mL increase) were independently associated with lower odds of belonging to the control (p = 0.034) and severe groups (p = 0.005), indicating that higher IFI35 levels were observed among critically ill patients. Table 2 summarizes the variables independently associated with COVID-19 severity.

Table 2.

Multivariable multinomial logistic regression analysis of factors associated with COVID-19 severity.

Variable	Control vs. Critical		Severe vs. Critical
	OR (95% CI)	p-value	OR (95% CI)	p-value
IFI35 (per 100 pg/mL increase)	0.93 (0.88–0.99)	0.034	0.94 (0.90–0.98)	0.005
Age	0.92 (0.89–0.94)	< 0.0001	0.99 (0.98-1.00)	0.403
Male sex	0.35 (0.20–0.63)	< 0.0001	0.96 (0.64–1.44)	0.869
Diabetes	0.28 (0.09–0.87)	0.028	0.75 (0.47–1.20)	0.237
Hypertension	0.37 (0.16–0.88)	0.024	0.93 (0.59–1.44)	0.747
Obesity	0.13 (0.04–0.41)	< 0.0001	0.63 (0.39–1.03)	0.071
CHD	-		0.48 (0.14–1.65)	0.250
CKD	-		0.16 (0.03–0.73)	0.019

CHD Chronic heart disease, CKD Chronic Kidney Disease.

Reference dependent category: Critical.

Additionally, we assessed whether IFI35 serum levels at hospital admission could predict clinical outcomes during hospitalization. Higher IFI35 levels were observed in individuals who died (p < 0.0001), required invasive mechanical ventilation (p < 0.0001), developed acute kidney injury (p = 0.01), or experienced cardiac arrest (p = 0.002) (Fig. 2).

Fig. 2.

Violin plots of serum IFI35 concentrations (pg/mL) according to A disease outcome, B use of invasive mechanical ventilation, C occurrence of acute kidney injury, and D cardiac arrest. Statistical significance was determined using the Mann–Whitney test.

To evaluate the prognostic performance of IFI35 in comparison with established laboratory biomarkers used in clinical practice, receiver operating characteristic (ROC) curve analysis was performed. Among the biomarkers tested, C-reactive protein (CRP) and D-dimer showed the highest individual discriminative capacity, with areas under the curve (AUCs) of 0.693 and 0.756, respectively. Based on these findings, a combined model including CRP and D-dimer was constructed, and the incremental prognostic value of IFI35 was assessed by adding it to this model. The inclusion of IFI35 increased the AUC to 0.790, indicating improved discriminative performance (Fig. 3).

Fig. 3.

Receiver operating characteristic (ROC) curves evaluating the prognostic performance of IFI35, C-reactive protein (CRP), D-dimer, and their combined model for predicting mortality in hospitalized COVID-19 patients. The combined model (IFI35 + CRP + D-dimer) demonstrated the highest discriminatory performance (AUC = 0.790), followed by D-dimer (AUC = 0.756), CRP (AUC = 0.693), and IFI35 (AUC = 0.617).

To assess whether IFI35 could serve as a predictor of mortality, we determined the optimal cut-off value to differentiate survivors from deceased patients using the maximally selected rank test and performed a survival analysis (Supplementary Figure S4). Our findings revealed that individuals with IFI35 levels above 1,475 pg/mL had a significantly shorter survival time compared to those with levels ≤ 1,475 pg/mL (24 days vs. 32 days; Log-rank test, p = 0.007) (Fig. 4).

Fig. 4.

Kaplan–Meier survival curve of patients stratified by IFI35 serum levels using the cutoff value of 1,475 pg/mL. Statistical significance was assessed using the log-rank (Mantel–Cox) test. Survival analysis included n = 430 hospitalized patients (214 critical and 216 severe).

Furthermore, IFI35 levels > 1,475 pg/mL remained an independent predictor of COVID-19 mortality in the Cox proportional hazards regression analysis (HR = 2.33, p = 0.001) (Table 3).

Table 3.

Cox proportional hazards regression analysis for the association between IFI35 levels and mortality risk in hospitalized patients with COVID-19.

Variables	HR (95% CI)	p-value
IFI35 > 1475 pg/mL	2.33 (1.41–3.85)	0.001
Obesity	2.09 (1.28–3.41)	0.003
CHD	4.43 (2.05–9.53)	< 0.0001
CKD	2.46 (1.27–4.75)	0.007

CI confidence interval, HR hazard ratio, CHD chronic heart disease, CKD chronic kidney disease. Variables included in the initial model: age, sex, diabetes, hypertension.

Discussion

This study demonstrates that both serum levels and gene expression of IFI35 increase in parallel with COVID-19 severity, and that elevated IFI35 was independently associated with critical illness and mortality. Furthermore, a cutoff value of 1,475 pg/mL was identified as a clinically relevant threshold for predicting adverse outcomes. Together, these findings position IFI35 as a potential biomarker for risk stratification. However, mechanistic studies, including in vivo and in vitro investigations, are required to determine whether IFI35 has a causal role in disease pathogenesis or represents a viable therapeutic target in severe COVID-19.

IFI35 is a 282-amino acid, 35 kDa protein whose expression is induced by type I and type II interferons in various cell types, including macrophages, fibroblasts, keratinocytes, and epithelial cells⁸. As an interferon-stimulated gene (ISG), IFI35 participates in complex immune regulatory networks, involving multiple signaling pathways and interactions with inflammatory mediators⁹. During viral infections, numerous antiviral restriction factors, including type I IFNs, are produced by diverse cell types¹⁶. The innate immune response must be tightly regulated, as excessive IFN production can be deleterious¹⁷. IFI35’s interactions with regulatory proteins highlight its potential role in antiviral host defense, apoptosis, and cytokine signaling^10,17,18. Studies in animal models show that IFI35 can modulate the expression of other ISGs and inhibit viral replication^{10,17,19–21}. Moreover, IFI35 has been suggested as a marker of inflammation in other inflammatory disorders^7,22,23.

Our findings extend previous observations by confirming that elevated IFI35 levels are associated with greater COVID-19 severity. SARS-CoV-2 infection is characterized by an excessive inflammatory response, often driven by heightened cytokine production by the innate immune system. Although DAMPs such as IFI35 are important for initiating host defense, their overproduction can trigger uncontrolled cytokine and chemokine release, worsening disease outcomes⁹. Prior research has shown that IFI35 predicts COVID-19 severity with performance comparable to C-reactive protein (CRP), procalcitonin (PCT), and interleukin-6 (IL-6). Furthermore, neutralizing IFI35 with specific antibodies improved outcomes in SARS-CoV-2 and influenza models, highlighting its potential relevance as a therapeutic target in experimental settings¹⁵.

The regulation of DAMP-mediated inflammation is critical for preserving tissue homeostasis. DAMPs can activate NF-κB and IRF pathways, leading to excessive production of proinflammatory cytokines^5,7. Although we have not investigated these mechanisms directly, experimental data demonstrate that IFI35 and its partner protein NMI activate NF-κB via TLR4 in macrophages, thereby increasing TNF and IL-6 production. In murine sepsis models, suppression of either IFI35 or NMI reduced inflammation and mortality⁷. These findings provide a mechanistic framework that may help explain the association observed between IFI35 levels and clinical outcomes in our cohort.

Gene expression analysis in our cohort revealed that patients with moderate and severe COVID-19 had significantly higher IFI35 expression than controls. This upregulation may reflect an intensified inflammatory state, as IFI35 is directly induced by IFNs, which drive ISG expression and suppress viral replication^19,24. These results align with an Irish study that identified IFI35 as one of 19 dysregulated genes in SARS-CoV-2–positive patients, associated with type I IFN signaling and upregulated in both mild and moderate/severe cases²⁴. Similarly, a French study using airway epithelial models found that SARS-CoV-2 strongly activated ISGs, including IFI35, while damaging ciliated cells and disrupting intercellular communication²⁵. Although our data demonstrated increased IFI35 expression in this cohort, the expression analysis was conducted in a relatively small subset of patients (n = 40), including only six individuals in the severe group, due to limited PBMC availability. In addition, samples from patients with mild COVID-19 were not available for qPCR analysis because standardized immediate processing conditions were not feasible at the time of collection. Consequently, gene expression levels were compared with samples obtained from healthy donors. The reduced sample size, particularly in the severe subgroup, and the absence of mild cases may have limited statistical power and should be considered when interpreting the gene expression findings.

Evidence from other viral infections reinforces the role of IFI35 in severe disease. A 2024 Chinese study on pandemic H1N1 influenza identified IFI35 as one of ten differentially expressed genes in late-stage infection, suggesting a specialized role in host responses to highly virulent viral strains²⁶. These converging data highlight IFI35’s potential relevance beyond COVID-19, particularly in severe respiratory viral syndromes.

Clinically, elevated IFI35 serum levels were associated with a greater need for ventilatory support, cardiovascular impairment, and a higher mortality risk. These findings are similar to those of other laboratory and clinical markers of severe COVID-19, such as elevated AST, creatinine, urea, C-reactive protein (CRP), and D-dimer levels, as well as increased respiratory support requirements²⁷. Beyond its role in SARS-CoV-2 infection, IFI35 has been implicated in a range of inflammatory and viral diseases, underscoring its broader role in innate immunity regulation and its potential as a prognostic marker^{13,14,28–30}. In our cohort, when assessing the risk of death, IFI35 demonstrated incremental prognostic value beyond conventional biomarkers such as C-reactive protein and D-dimer, supporting its potential clinical relevance for risk stratification.

This study has several limitations. The cohort was drawn from a single regional population, which may limit the generalizability of the findings. The prognostic cutoff identified here was derived and tested within the same cohort, introducing a risk of overfitting; therefore, this threshold should be considered exploratory and requires external validation. The control group was younger and had a different sex distribution compared with hospitalized patients. Although age and sex were included as covariates in multivariable models, residual confounding cannot be excluded. Additionally, controls were recruited up to 12 months post-infection, whereas hospitalized patients were sampled at admission, introducing temporal heterogeneity that may influence biomarker comparisons. Gene expression analysis should be considered exploratory, as it was performed in a relatively small subset of patients and relied on a single endogenous control. Despite these limitations, this study includes over 540 participants and represents the largest Brazilian investigation to date linking IFI35 with COVID-19 severity, as well as the first to integrate serum measurements and gene expression analyses with detailed clinical and laboratory parameters.

Our findings indicate that IFI35 is associated with COVID-19 severity and poor outcomes, supporting its role as an inflammatory biomarker. Increased IFI35 gene expression was also associated with clinical severity. The cutoff value of 1,475 pg/mL effectively discriminates patients at higher risk of death in this cohort, although external validation is required. Taken together, these findings support IFI35 as a promising biomarker for disease severity and suggest its potential relevance in other severe viral infections, although this requires further investigation.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

MRF: formal analysis, investigation, and writing—original draft. CAF: formal analysis, investigation, and writing—review and editing. SLPN: formal analysis, investigation, and writing—review and editing. MXB: formal analysis, investigation, and writing—review and editing. CDFS: formal analysis, methodology, visualization, and writing—review and editing. ACA: funding acquisition, project administration, writing—review and editing. RFC: conceptualization, data curation, funding acquisition, methodology, project administration, validation, and writing—review and editing.

Funding

This work was supported by the Fundação de Amparo à Ciência e Tecnologia do Estado de Pernambuco (FACEPE, Brazil; doctoral scholarships IBPG-1653-2.02/21 and IBPG-1896-4.03/21), by the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES, Brazil – Finance Code 001), and by the Health Secretariat of Pernambuco, Brazil (Agreement 10/2021). Rodrigo F Carmo, CarlosDF Souza and Anderson C Armstrong are research productivity fellows of the Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq, Brazil).

Data availability

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.