G*power software was used to estimate the sample size power with the following inputs: 90 COVID-19 patients, 90 control subjects, 0.05 two-tailed α-error p, and 0.5 effect size d. The estimated power of sample size was 0.91, and the statistically accepted power is 0.8 (Kang, 2021).

Baseline laboratory data included ESR and total and absolute counts of WBC (granulocytes, lymphocytes, and monocytes). COVID-19 patients showed a significantly increased median (IQR) ESR compared to HC (45 [37–55] vs. 13 [12,13,14,15,16,17,18,19] mm/h; p < 0.001). Likewise, the total count of WBC was significantly higher in patients than in HC (7.8 [6.2–9.8] vs. 7 [6.0–8.1] × 109/L; p = 0.001), but the count in both groups was within the reference range. Granulocytes were significantly elevated in patients compared to HC (5.2 [4.0–6.9] vs. 3.8 [3.1–4.8] × 109/L; p < 0.001), while lymphocytes were significantly lower in patients (1.8 [1.4–2.4] vs. 2.5 [2.2–3.1] × 109/L; p < 0.001). Monocyte counts showed no significant difference between patients and HC (p = 0.567). GLR was significantly higher in patients than in HC (2.3 [1.8–4.5] vs. 1.6 [1.3–1.8]; p < 0.001), while LMR decreased significantly in patients (3.5 [3.0–4.3] vs. 4.8 [3.7–5.6]; p < 0.001) (Table 2).

HWE analysis demonstrated that genotype frequencies of both SNPs (rs1990760 and rs2111485) did not deviate significantly from HWE in the HC group (p = 0.523 and 0.253, respectively). Association analysis was conducted under five genetic models; allele, recessive, dominant, over-dominant, and co-dominant. For SNP rs1990760, the genotype frequencies of CC, CT, and TT were 4.4, 37.8, and 57.8%, respectively, in COVID-19 patients. The corresponding frequencies in HC were significantly different (24.4, 53.3, and 22.2%, respectively; p < 0.001). Under the allele model (T vs. C), the mutant T allele was associated with a significantly increased risk of COVID-19 (OR 3.44; 95% CI 2.19–5.39; p < 0.001; pc < 0.001). A higher risk of infection with COVID-19 was associated with the mutant TT genotype under recessive (TT vs. CC + CT: OR 4.86; 95% CI 2.52–9.38; p < 0.001; pc < 0.001) and co-dominant (OR 14.59; 95% CI = 4.42–48.19; p < 0.001; pc < 0.001) models. Under dominant model (CT + TT vs. CC), the CT + TT genotypes were also associated with an increased risk of COVID-19 (OR 6.95; 95 CI 2.27–21.28; p = 0.001; pc = 0.005). Under over-dominant model (CT vs. CC + TT), there was no significant risk of COVID-19 associated with SNP rs1990760 (p = 0.035; pc = 0.175) (Table 3).

In the case of SNP rs2111485, genotype frequencies (AA, AG, and GG) showed significant differences between COVID-19 patients and HC (6.7, 64.4 and 28.9% vs. 42.2, 41.1 and 16.7%, respectively; p < 0.001). The mutant G allele (allele model; G vs. A) and GG genotype (co-dominant model; GG vs. AA) were associated with a higher risk of COVID-19 (OR 2.65, 95% CI 1.73–4.05, p < 0.001, pc < 0.001; OR = 10.72, 95% CI 3.66–31.40, p < 0.001, pc < 0.001, respectively). Under dominant (AG + GG vs. AA: OR = 10.15; 95% CI 4.01–25.69; p < 0.001; pc < 0.001) and over-dominant (AG vs. AA + GG: OR 2.62; 95% CI 1.43–4.80); p = 0.002; pc = 0.01) models, a significant increased risk of COVID-19 was also associated with SNP rs2111485. Under recessive model (GG vs. AA + AG), there was no significant risk of COVID-19 associated with SNP rs2111485 (p = 0.063; pc = 0.315) (Table 3).

To further explore the association of IFIH1 SNPs with COVID-19, haplotype analysis of the two SNPs was conducted (haplotype: rs1990760-rs2111485). Four haplotypes were encountered (T-A, T-G, C-A, and C-G) with frequencies of 22.2, 54.4, 16.7 and 6.7%, respectively, in COVID-19 patients, and 20.6, 28.3, 42.2 and 8.9%, respectively, in HC. The T-G haplotype was significantly associated with an increased risk of the disease (OR 3.02; 95% CI 1.95–4.68; p < 0.001; pc < 0.001), while the C-A haplotype was associated with a significantly decreased risk of contracting COVID-19 (OR 0.27; 95% CI 0.18–0.45; p < 0.001; pc < 0.001). The T-A and C-G and haplotypes showed no association with COVID-19 risk (Table 3). LD analysis revealed that the rs1990760 and rs2111485 polymorphisms were moderately linked as indicated by the LD coefficient (D′) of 0.57 and correlation coefficient (R2) of 0.18 (Fig. 1).

Two-locus linkage disequilibrium (LD) plot for the IFIH1 variants rs1990760 and rs2111485 showing the LD coefficient (D′; 0.57; left) and correlation coefficient (R2; 0.18; right). Plots were generated using SHEsis software

IFIH1 levels (median and IQR) were significantly higher in the serum of COVID-19 patients than in HC (15.35 [13.28–17.30] vs. 4.60 [3.17–6.50] ng/mL; p < 0.001) (Fig. 2A). ROC curve analysis demonstrated the potential of IFIH1 in distinguishing between COVID-19 patients and HC as indicated by an AUC of 0.999 (95% CI 0.999–1.0; p < 0.001). At a cut-off point of 10 ng/mL, the sensitivity and specificity of IFIH1 were 100.0 and 97.8%, respectively (Fig. 2B).

IFIH1 serum levels were measured and compared between infected individuals and healthy controls. The ROC curve analysis was performed to indicate the specificity and sensitivity of IFIH1 between the compared groups. A Box and whisker plots of interferon-induced helicase-1 (IFIH1) levels in the serum of: COVID-19 patients (n = 90) and healthy controls (HC; n = 90). The horizontal line inside the boxes indicates the median. Whiskers represent interquartile range (IQR; 25–75%). Significance was determined using Mann–Whitney U test (***p < 0.001). Serum IFIH1 levels were significantly higher in COVID-19 patients than in HC (15.35 [IQR 13.28–17.30] vs. 4.60 [IQR 3.17- 6.50] ng/mL; p < 0.001). B Receiver-operating characteristic (ROC) curve analysis of IFIH1 in COVID-19. ROC analysis demonstrated the potential of IFIH1 in distinguishing between COVID-19 patients and HC as indicated by the area under the curve (AUC) of 0.999

To evaluate the impact of IFIH1 SNPs on IFIH1 serum levels, these levels were stratified by genotypes of SNPs rs1990760 and rs2111485 in all participating individuals (COVID-19 plus HC patients; n = 180). Individuals with the mutant-type genotypes rs1990760 TT (14.2 [8.2–16.9] vs. 3.4 [2.5–4.8] ng/mL; p < 0.001) and rs2111485 GG (13.4 [7.4–16.4] vs. 6.4 [3.4–8.1] ng/mL; p < 0.001) showed significantly elevated levels of IFIH1 compared to the wild-type genotypes (rs1990760 CC and rs111485 AA, respectively) (Fig. 3).

Box and whisker plots of interferon-induced helicase-1 (IFIH1) levels in the serum of: COVID-19 patients plus healthy controls (n = 180) classified by IFIH1 gene variants (A rs1990760; B rs2111485). The horizontal line inside the boxes indicates the median. Whiskers represent interquartile range (IQR; 25–75%). Black circles indicate outliers. Significance was determined using Mann–Whitney U test and the false discovery rate was used to correct probability (p) for multiple comparisons (*p < 0.05; ***p < 0.001; ns: not significant). The mutant-type genotypes of IFIH1 SNPs (rs1990760 TT and rs2111485 GG) were associated with significantly elevated levels of IFIH1 compared to the wild-type genotypes (rs1990760 CC and rs2111485 AA, respectively)

Spearman’s rank-order correlation analysis indicated that IFIH1 was positively correlated with ESR (rs = 0.708; p < 0.001; Fig. 4A), WBC (rs = 0.183; p = 0.014; Fig. 4B), and GLR (rs = 0.463; p < 0.001; Fig. 4C), while it showed a negative correlation with LMR (rs = -0.415; p < 0.001; Fig. 4D).

Spearman's rank-order correlation analysis of interferon-induced helicase-1 (IFIH1) with A erythrocyte sedimentation rate (ESR), B white blood cell (WBC) count, C granulocyte-to-lymphocyte ratio (GLR), and D lymphocyte-to-monocyte ratio (LMR) in all participants (COVID-19 patients plus controls; n = 180). The correlation coefficient (rs) and two-tailed probability (p) are shown. IFIH1 was positively correlated with ESR, WBC, and GLR, while it showed a negative correlation with LMR

STRING analysis demonstrated that IFIH1 protein interacts with five proteins, DHX58 (DExH-box helicase 58), MAVS (Mitochondrial antiviral signaling protein), NLRC5 (NLR family CARD domain containing 5), ISG15 (ISG15 ubiquitin like modifier), and ATG12 (Autophagy related 12) (Fig. 5).

Prediction of the protein–protein interaction network between Interferon-induced helicase-1 (IFIH1) and relevant proteins (interaction score > 0.9) as revealed by STRING analysis (https://string-db.org). Colored nodes indicate IFIH1 and first shell of interactions. Edges indicate IFIH1-protein association. Connections were identified as colored lines. DHX58: DExH-box helicase 58; MAVS: Mitochondrial antiviral signaling protein; NLRC5: NLR family CARD domain containing 5; ISG15: ISG15 ubiquitin like modifier; ATG12: Autophagy related 12