Ninety preterm newborns (less than 37 weeks) were the subjects of this case-control research, which was approved by the Faculty of Medicine, South Valley University’s Ethics Committee (ethical approval code: SVU-MED-PED025-4-25-1-16). Written informed consent was received from the participant patients' caregivers. The current research has been conducted in accordance with the Declaration of Helsinki. Of these, 60 were admitted to the neonatal intensive care unit (NICU) at Qena University Hospital, and 30 were outpatient healthy preterm neonates whom their caregivers came for routine follow-up in the outpatients’ pediatric clinics. The research was conducted between January 2022 and June 2024. Three groups of preterm newborns were included in the present study based on whether they had neonatal RDS or not, and the RDS severity as follow: group A consisted of 30 preterm neonates with mild to moderate RDS. Group B contained 30 preterm newborns with severe RDS (Supplementary Table 1). As a control group, group C comprised 30 healthy preterm newborns that were matched with groups A and B for weight, sex, and gestational age (GA) and represent the control group.

Inclusion criteria: during the study period, preterm neonates with a GA of less than 37 weeks were either admitted or evaluated as outpatients in the NICU at Qena University Hospital, regardless of whether they were healthy or experiencing RDS. RDS-diagnosed neonates (identified by combining radiological evidence from a chest x-ray with clinical symptoms of respiratory distress.

Exclusion criteria: Newborns who are small for GA, have congenital defects, or have early-onset neonatal sepsis. babies of mothers who had gestational diabetes mellitus or diabetes mellitus or who have a history of chronic diseases such hypertension, autoimmune disorders, or renal failure or who experienced premature rupture of the membranes before their due date.

To attain 80% power and 5% confidence (type 1 error), we modify the sample size to be 60 patients.

History from the caregivers of the included neonates was taken (model of delivery, GA, sex, birth weight, antenatal steroids, age at admission, and duration of admission).

Respiratory pattern, rate, and depth, intercostal and subcostal retraction, grunting, and auscultation for additional sounds are all used to diagnose RDS. RDS is graded using Downes’ score [12], Silverman-Anderson score [13], x-ray grading [14] (Supplementary Table 1), the type of assisted ventilation (continuous positive airway pressure, or mechanical ventilation [15], and the length of hospital stay.

Five milliliters of blood were collected from each participant and separated into 2 portions. The first portion, 2 milliliters, was transferred into tubes containing ethylenediaminetetraacetic acid (EDTA). Following a full blood count (complete blood count [CBC]), the sample was immediately frozen at -80°C until the time of genetic analysis. Meanwhile, the remaining portion (3 mL) was evacuated into a separator gel tube and centrifuged at 822 g for 15 minutes. For further biochemical assays of MIF and GDF-15 levels, the separated sera were split into aliquots using 1-mL cryotubes and stored at -20°C.

Abbott Diagnostics' Cell Dyn 1800, Germany was used to measure the complete blood counts. The enzyme-linked immune-sorbent assay (ELISA) multiskan EX micro-platephotometer (STAT FAX-2100; Thermo Scientific, USA) was used to evaluate the levels of serum MIF and GDF-15. A commercial ELISA assay kit from Chongqing Biospes Co., Ltd. (China) was used (Catalogue No. BEK1151, and BEK1290 respectively). The sandwich ELISA technology served as the basis for the applied ELISA technique.

Following the manufacturer's instructions, genomic DNA was extracted from whole EDTA blood samples using the G-spinTM Total DNA Extraction Kit (iNtRON Biotechnology, Inc., Korea). Following that, the DNA samples were kept at -80°C until genetic analysis.

In accordance with previous methods, the restriction fragment length polymorphism-polymerase chain reaction (PCR) approach was carried out to detect MIF (rs755622 G>C) and GDF-15 (rs4808793 C>G) SNPs utilizing certain primer sequences as follow: The forward primer was 5′-CT CAAACACACAAGCTCACGCATGCG-3′ and the reverse primer was 5′-ACCACTGTGGTCCCGCCTTTTGTGAC-3′ [16], and the forward primer was 5′- GCAACAGAGCGAGACTC CA-3′, and the reverse primer was 5′-CCACGCCGGTCGGATTAAAACT-3′ [17], respectively. One microliter of each primer, one microliter of genomic DNA (~100 ng/mL), 10 μL of 2X Prime Taq Premix, and 7 μL of dd H2O were added into every 0.20-mL reaction. The PCR condition for MIF (rs755622 G>C) were as follow: 5 minutes of initial denaturation at 95°C, 30 cycles of denaturation at 95°C for 30 seconds, 30 seconds of annealing at 67°C, 30 seconds of extension at 72°C, and 5 minutes of final extension at 72°C which will yield PCR amplification products at 439 bp (Fig. 1A), while for GDF-15 (rs4808793 C>G) were as follow: 35 cycles of the following temperature profile and 10 minutes of initial denaturation at 94°C: 5 minutes for the last extension, 30 seconds at 94°C, 30 seconds at 65°C, and 30 seconds at 72°C, which will yield PCR amplification products at 665 bp (Fig. 2). FastDigest AluI (FD0014, Lot:00147479, Thermo Fisher Scientific, USA) was used for digestion of the PCR amplification products (10 μL) of rs755622, while The PCR products of rs4808793were digested using a restriction enzyme (AlwNI), which was provided by New England Biolabs and has catalogue number R0514S. The final digestion products were separated on 2.0% agarose gels, then stained with ethidium bromide and observed under ultraviolet light. The band size pattern of various genotypes of the studied SNPs were as follow: for MIF (rs755622 G>C), wild homozygous GG (produces bands at 439 bp), mutant homozygous CC (produces bands at 255, and 184 bp), and mutant heterozygous GC (produces bands at 439 bp, 255, and 184 bp) as showed in (Fig. 1B); while for GDF-15 (rs4808793 C>G), The CC genotype of the wild samples displayed an intact band on the 665 region, while the heterozygous mutant genotype (CG) produced 3 distinct bands on the 665, 422, and 243 bp regions, while the mutant homozygous (GG) showed 2 distinct bands on the 243-bp and 422-bp regions (Fig. 2).

IBM SPSS Statistics ver. 26.0 (IBM Co., USA) was used to analyze the data. The Shapiro-Wilk and Kolmogorov-Smirnov tests of normality were conducted. While continuous data were expressed as mean±standard deviation (SD) or median and interquartile range, categorical data were defined by number and percentage. To compare categorical variables, the chi-square test was used. The independent samples t test for parametric data and the Mann-Whitney U test for nonparametric data were used to identify differences between each of the 2 groups. Pearson and Spearman correlation coefficients are used for studying the correlation in parametric and nonparametric data respectively. The Hardy-Weinberg equation was followed by the SNPs under study. A statistically significant result was defined as a 2-tailed P<0.05.

For the cases and control groups, the mean±SD GAs were 32.68±2.02 weeks and 33.00±2.35 weeks, respectively. Additionally, there were 15 males (50%) to 15 females (50%) in the control group and 30 males (50%) to 30 females (50%) in the cases group. Additionally, the study groups' mean±SD weight values were 2.04±0.31 kg for the control group and 2.04±0.36 kg for the patients. Weight, sex distribution, and GA were all matched between cases and controls, as seen by the negligible discrepancies between the 2 groups, P>0.05 for all. 29 patients (48.3%) did not receive prenatal steroids, whereas 31 patients (51.7%) did. Eight of the control group (26.7%) did not receive prenatal steroids, whereas 22 (73.3%) did, with a considerably higher percentage than in the cases (P=0.049). When comparing several CBC values between patients and controls, the following were found to be substantially lower in cases than in controls: White blood cells (WBCs), lymphocytes, and platelets, all of which had P<0.05: The study groups' mean WBC±SD values were 9.01±3.94 ×10³/μL for the case group and 13.18±2.71 ×10³/μL for the control group. The study groups' mean±SD lymphocyte counts were 3.58±1.65 ×10³/μL for the patients and 6.45±2.76 ×10³/μL for the controls. The study groups' mean± SD platelet counts were 344.6±62.08 ×10³/μL for the control group and 229.85±103.73 ×10³/μL for the patients. There were insignificant differences between cases and controls regarding to the remaining CBC parameters (hemoglobin, neutrophils count %, and lymphocyte %), P>0.05 for all (Table 1).

When comparing the 2 patients subgroups, those with mild to moderate RDS (group A) versus those with sever RDS (group B), there were significantly lower mean GA and weight among group B (31.9±1.99 weeks, and 1.84±0.34 kg, respectively) compared to group A (33.47±1.76 weeks, and 2.25±0.24 kg, respectively), P<0.05 for both. 60% of group A were females, while 60% of group B were males. There was insignificant difference between the 2 patients groups regarding the history and number of doses of antenatal steroids, P>0.05. The mean value of Downes' score and Silverman-Anderson retraction score were significantly higher among group B (6.7±0.95 and 7.27±0.79, respectively) than in group A (4.33±0.48 and 4.67±0.66, respectively), P<0.001 for both. Regarding chest x-ray grading of RDS, grades III and IV were only presented in group B, while grade I was only presented in group A with significantly higher percentage of grade II among group A compared to group B (P<0.001). There was significantly longer mean duration of NICU admission among group B (11.8±4.6 days) than in group A (4.33±1.61), P<0.001. Regarding CBC parameters, there was only significantly lower mean value of neutrophils count among group B (3.86±2.64) compared to group A (5.28±3.13), P=0.036, with insignificant differences between the 2 groups regarding the other CBC parameters, P>0.05 for all (Table 1).

There was insignificantly higher median value of MIF (μg/L) among cases (11.7) compared to the controls (11.00), P=0.581 with significantly higher median MIF blood levels among patients group B (17.32) compared to patients group A (5.50), P<0.001. Serum GDF-15 levels (pg/mL) showed insignificant differences between cases and controls (P=0.774), and its median level was significantly higher in patients group B (3.19) than in patients group A (0.71) (P=0.021) (Table 2).

There was significantly positive correlation between serum MIF levels and each of serum GDF-15 levels (r=0.563, (P<0.001) as showed in (Fig. 3A), and radiological severity of RDS (r=0.393, P=0.002) among preterm neonates with RDS (Fig. 3B). Lack of any significant associations between each of MIF or GDF-15 blood levels with GA, weight, Downes' score and Silverman-Anderson retraction score, duration of admission, and various CBC parameters, P>0.05 for all.

The frequencies of mutant homozygous (CC) and mutant heterozygous (GC) genotypes of MIF rs755622 G>C SNP were significantly higher among cases (11.7% and 51.7%, respectively) compared to the controls (6.7% and 13.3%, respectively), with significantly higher frequency of wild homozygous (GG) genotype among controls (80.0%) in comparison with the cases (36.7%), P<0.001for all. Furthermore, there was significantly higher frequency of mutant C-allele among cases (37.5%) compared to the controls (13.3%), P=0.001 indicating that C-allele could be considered as a genetic risk factor for RDS among preterm neonates with odds ratio (OR)=0.256 and 95% confidence interval (CI), 0.112–0.589 (Table 3). There were insignificant differences regarding both genotypes and alleles frequencies of MIF rs755622 G>C SNP in preterm neonates with mild to moderate RDS compared to those with severe RDS, P>0.05 for all. This indicates lack of MIF rs755622 G>C SNP with the severity of RDS among preterm neonates. Also, there were insignificant differences regarding the median MIF blood levels in terms of genotypes of MIF rs755622 G>C SNP among the included patients, P>0.05 for all.

Regarding to GDF-15 (rs4808793 C>G SNP, there were significantly higher frequency of mutant heterozygous (CG) genotype among cases (65.0%) compared to the controls (20.0%) with significantly higher frequency of wild homozygous (CC) genotype among controls (60.0%) compared to the patients (18.3%), P<0.001. Additionally, there was significantly lower frequency of normal C-allele among cases (50.8%) compared to the controls (70.0%), P=0.014 with higher frequency of mutant G-allele among cases (49.2%) in comparison with the controls (30.0%), indicating that G-allele could be considered as a genetic risk factor for RDS among preterm neonates with OR=0.443 and 95% CI, 0.229–0.856 (Table 3). There were insignificant differences regarding both genotypes and alleles frequencies of GDF-15 (rs4808793 C>G SNP when comparing preterm neonates with mild to moderate RDS versus those with severe RDS, P>0.05 for all. This indicates lack of GDF-15 (rs4808793 C>G SNP with the severity of RDS among preterm neonates. Also, there were insignificant differences regarding the median GDF-15 blood levels in terms of genotypes of GDF-15 (rs4808793 C>G SNP among the included patients, P>0.05 for all.