Cytokines and Inflammatory Gene Polymorphisms Associated With Nosocomial Pulmonary Infection After Spontaneous Intracerebral Hemorrhage

Summary

This retrospective study presents a reproducible workflow for early post-admission blood collection, measurement of serum cytokines and Toll-like receptors, genotyping of inflammatory gene polymorphisms, and infection-risk assessment in patients with spontaneous intracerebral hemorrhage.

Abstract

Nosocomial pulmonary infection is a frequent complication after spontaneous intracerebral hemorrhage and may worsen neurological recovery, prolong hospitalization, and increase clinical burden. This retrospective clinical-laboratory study presents a reproducible workflow for evaluating inflammatory biomarker and host immune-genetic profiles associated with nosocomial pulmonary infection after primary spontaneous intracerebral hemorrhage. Patients are classified according to whether nosocomial pulmonary infection occurs after admission. Peripheral venous blood is collected in the early post-admission period under standardized pre-analytical conditions. Serum is separated, aliquoted, and stored for enzyme-linked immunosorbent assay measurement of IL-1β, IL-6, IL-10, IL-17, IFN-γ, TNF-α, TLR2, TLR4, and TLR9. In parallel, genomic DNA is extracted from anticoagulated whole blood and used for polymerase chain reaction-restriction fragment length polymorphism genotyping of selected cytokine- and Toll-like receptor-related loci. The workflow also includes quality-control procedures for sample handling, duplicate ELISA measurements, DNA purity assessment, genotype calling, and repeat genotyping. Statistical analysis includes between-group comparison of clinical characteristics and biomarker levels, Hardy-Weinberg equilibrium testing, logistic regression analysis for genotype and allele associations, adjustment for relevant clinical covariates, and false-discovery-rate correction for multiple genetic comparisons. This combined clinical, inflammatory, and immune-genetic workflow may help characterize infection-risk profiles after spontaneous intracerebral hemorrhage, although prospective multicenter validation is still required before routine clinical application.

Introduction

Spontaneous intracerebral hemorrhage (ICH) is a severe subtype of stroke caused by non-traumatic rupture of cerebral vessels and remains a major source of neurological death and disability. Clinically, ICH is commonly classified as primary or secondary. Primary ICH accounts for most cases and is mainly related to hypertensive arteriopathy or cerebral amyloid angiopathy¹. By contrast, secondary ICH is associated with structural or hematologic causes, including aneurysms, arteriovenous malformations, Moyamoya disease, and coagulation disorders. Although ICH accounts for a smaller proportion of all stroke events than ischemic stroke, it carries disproportionately high short-term mortality and long-term functional burden^2,3,4,5.

Among the major in-hospital complications after ICH, nosocomial pulmonary infection is particularly frequent and clinically consequential. Previous studies have shown that pulmonary infection can occur in a substantial proportion of hospitalized patients after stroke or intracerebral hemorrhage and is associated with longer hospitalization, greater medical burden, poorer neurological recovery, and increased mortality^6,7,8. The clinical importance of this complication extends beyond respiratory management alone. Reduced consciousness, impaired swallowing and cough reflexes, immobilization, mechanical airway support, and invasive drainage procedures may all increase the risk of infection during the acute phase after ICH. At the same time, pulmonary infection may further worsen systemic inflammation, impair oxygen delivery, and complicate neurocritical care, thereby amplifying secondary injury processes.

This relationship is biologically plausible because ICH is not only a focal cerebrovascular event but also a trigger of systemic immune dysregulation. Acute cerebral hemorrhage can activate inflammatory cascades, alter innate immune signaling, and disrupt the balance between neuroinflammation and peripheral host defense. In this context, some patients may become more susceptible to hospital-acquired infections even under similar clinical management. Therefore, pulmonary infection after ICH should not be viewed solely as a consequence of bedside exposure or supportive care; it may also reflect differences in host inflammatory activation and immune responsiveness.

Despite the clinical relevance of this complication, practical approaches for early risk stratification remain limited⁹. Existing clinical risk-stratification approaches for post-ICH pulmonary infection generally rely on bedside and admission-related variables, such as neurological severity, age, swallowing impairment, comorbidities, invasive procedures, and length of device exposure. These tools are clinically useful because they can be applied rapidly in emergency, intensive care, or neurocritical care settings. However, clinical variables alone may not fully explain why patients with comparable neurological severity and similar care exposure show different infection outcomes. Inflammatory biomarkers and host immune-genetic factors may therefore provide additional information about biological susceptibility, especially during the early post-admission period when preventive surveillance and respiratory management decisions are being made.

Cytokines are central mediators of immune and inflammatory responses, and Toll-like receptor (TLR) pathways play key roles in pathogen recognition and innate immune activation. Serum cytokine and TLR protein measurements can provide a snapshot of early systemic inflammatory activation, whereas selected immune-gene polymorphisms may reflect more stable host differences in cytokine production, receptor signaling, or inflammatory regulation. Compared with clinical assessment alone, a combined biomarker and polymorphism workflow can capture both the current inflammatory state and each patient's underlying immune response. This approach is not intended to replace bedside clinical assessment, but to complement it by adding laboratory and host-genetic information that may help identify patients who warrant closer infection surveillance after spontaneous ICH.

On this basis, we hypothesized that patients with spontaneous ICH who develop nosocomial pulmonary infection would exhibit a distinct inflammatory profile together with a different distribution of selected immune-related genetic polymorphisms compared with patients who remain free of infection. To test this hypothesis, the present retrospective clinical-laboratory study analyzed early post-admission serum cytokines, serum protein levels of TLR2, TLR4, and TLR9, and selected inflammatory gene polymorphisms in patients with primary spontaneous ICH. The workflow uses peripheral blood collected during acute hospitalization, ELISA-based protein quantification, PCR-RFLP-based genotyping, and statistical comparison between patients with and without nosocomial pulmonary infection. The aim was to provide a reproducible framework for characterizing inflammatory and host-genetic profiles associated with infection risk, thereby supporting earlier recognition of high-risk patients in the acute hospital setting.

Protocol

This study was approved by the Ethics Committee of the Affiliated Hospital of Jiaxing University, The First Hospital of Jiaxing, Jiaxing, Zhejiang, China (Approval No. 2025-LP-599). All procedures involving human participants were conducted in accordance with institutional requirements and the principles of the Declaration of Helsinki. Written informed consent for blood collection and data use was obtained from all participants and/or their legally authorized representatives before sample collection.

All commercial kits, reagents, consumables, instruments, software, manufacturers, catalog numbers, and key operating specifications used in this workflow are listed in the Table of Materials.

Study scope, eligibility, and patient grouping

This study was conducted as a single-center retrospective clinical-laboratory cohort analysis of patients with primary spontaneous intracerebral hemorrhage (ICH) admitted to the Emergency and Critical Care Center or Intensive Care Unit of the Affiliated Hospital of Jiaxing University, The First Hospital of Jiaxing, Jiaxing, Zhejiang, China, between May 2019 and September 2024. The diagnosis of ICH was confirmed by cranial computed tomography or magnetic resonance imaging in accordance with the Chinese Guidelines for Diagnosis and Treatment of Cerebral Hemorrhage (2019)¹⁰.

Patients were eligible if they had imaging-confirmed primary spontaneous ICH, were admitted within 3 days of symptom onset, and had complete key clinical variables, along with available peripheral blood samples for measurement of inflammatory biomarkers and genetic analysis. Patients were excluded if they had traumatic intracranial hemorrhage, secondary ICH, pulmonary infection present before admission or within 48 h after admission, discharge against medical advice, or death within 48 h after admission, a history of malignancy or autoimmune disease, or incomplete key clinical, laboratory, or genotyping records. After screening, 360 patients met the eligibility criteria and were included in the final analysis. Among them, 180 patients developed nosocomial pulmonary infection during hospitalization, and 180 patients did not. Patients were assigned to the pulmonary infection group or the no pulmonary infection group based on whether a nosocomial pulmonary infection occurred after admission.

Before final grouping, imaging diagnosis, admission time, infection timing, blood-sample availability, and completeness of clinical and laboratory records were checked. This step was used to reduce misclassification and ensure that all included cases had complete data for clinical comparison, measurement of inflammatory biomarkers, and genotyping.

Definition of nosocomial pulmonary infection

Nosocomial pulmonary infection was defined on the basis of radiologic evidence combined with clinical criteria¹¹. Radiologic confirmation required a new or progressive infiltrative, consolidative, or ground-glass lesion on chest radiography or computed tomography. In addition, at least two of the following criteria had to be present: body temperature > 38 °C, purulent respiratory secretions, or abnormal peripheral white blood cell count with WBC < 4 × 10⁹/L or > 10 × 10⁹/L.

Pulmonary infection present before admission or diagnosed within the first 48 h after admission was not classified as nosocomial pulmonary infection. Infection classification was verified using radiologic reports, body temperature records, documentation of respiratory secretions, peripheral blood count results, and clinical diagnosis records.

Biosafety and waste disposal

Human blood, serum, and blood-derived DNA were handled under standard biosafety level 2 precautions. Laboratory personnel wore gloves, laboratory coats, masks, and protective eyewear during blood handling, serum aliquoting, ELISA testing, DNA extraction, PCR preparation, gel electrophoresis, restriction digestion, and UV visualization. Blood tubes were checked for leakage or breakage before processing. Centrifugation was performed using capped tubes or, where available, sealed centrifuge buckets.

Sharps were discarded immediately into approved sharps containers. Blood-contaminated tubes, pipette tips, gloves, ELISA plates, and other disposable consumables were discarded as biological laboratory waste. Liquid biological waste was disinfected in accordance with institutional procedures before disposal. Agarose gels, polyacrylamide gels, staining reagents, and electrophoresis-related materials were discarded in accordance with institutional chemical and laboratory waste requirements.

PCR reagent preparation, DNA template addition, amplified product handling, restriction digestion, and gel electrophoresis were separated as far as possible to reduce contamination. Aerosol-resistant pipette tips were used for PCR and DNA handling. UV visualization was performed only with a protective shield or gel documentation system, and direct eye or skin exposure to UV light was avoided.

Peripheral blood collection and sample processing

To reduce pre-analytical variation, 10 mL of peripheral venous blood was collected at 08:00 on the morning after admission, after an overnight fast of at least 12 h. 5 mL of blood was drawn into a plain vacuum tube for serum preparation, and the remaining 5 mL was drawn into a heparinized tube for genomic DNA extraction.

Before processing, each tube was checked for patient identifier, collection time, tube type, blood volume, leakage, visible clotting, and gross hemolysis. Samples with incorrect labeling, insufficient blood volume, tube leakage, or visible clotting in the anticoagulated tube were excluded from downstream analysis. Blood in the plain tube was allowed to clot at room temperature for 60 min, then centrifuged at 1,500 × g for 10 min at 4 °C. The serum supernatant was transferred into sterile low-protein-binding polypropylene tubes without disturbing the cell layer, aliquoted into 0.5 mL fractions, and stored at −80 °C until analysis.

Samples showing visible moderate or severe hemolysis were excluded from ELISA-based quantification. Mildly hemolyzed samples were recorded and were not used for the primary cytokine analysis. Samples with visible turbidity, obvious lipemia, insufficient residual volume, or repeated freeze-thaw exposure were flagged before testing. No serum aliquot underwent more than one freeze-thaw cycle before analysis.

Cytokine quantification by ELISA

Serum concentrations of IL-1β, IL-6, IL-10, IL-17, IFN-γ, and TNF-α were measured using commercially available human sandwich ELISA kits validated for serum samples. Serum aliquots were thawed, gently mixed, and inspected for hemolysis, turbidity, lipemia, insufficient volume, and freeze-thaw history before loading.

Standards, blank wells, and serum samples were loaded in duplicate. Standard curves were generated on each plate using serial dilutions of standard concentrations and a blank control, according to the kit instructions. Optical density was measured at 450 nm, with a reference correction at 570 nm, using a calibrated microplate reader. Concentrations were calculated by four-parameter logistic curve fitting.

Assay results were accepted only when the standard curve showed an appropriate concentration-response pattern, blank wells showed low background, and the duplicate sample coefficient of variation was ≤ 10%. Samples with duplicate CV > 10% were re-assayed from the original serum aliquot when sufficient volume was available. Samples outside the reliable standard-curve range were diluted or repeated according to the kit instructions. Plates showing inconsistent standards, excessive background, or obvious edge effects were repeated.

The lower analytical detection limits used for data acceptance were 2.0 pg/mL for IL-1β, 1.5 pg/mL for IL-6, 2.0 pg/mL for IL-10, 3.0 pg/mL for IL-17, 4.0 pg/mL for IFN-γ, and 2.5 pg/mL for TNF-α. Final cytokine values were checked against sample identifiers, plate maps, duplicate-readout records, and sample-quality notes before statistical analysis.

Quantification of serum TLR2, TLR4, and TLR9 by ELISA

Serum TLR2, TLR4, and TLR9 protein levels were measured using commercially available human ELISA kits validated for serum specimens. Serum samples were thawed, mixed gently, and inspected using the same sample-quality criteria applied to cytokine measurement.

Standards, blank wells, and serum samples were analyzed in duplicate. Plates were read at 450 nm with reference correction at 570 nm, and concentrations were calculated using four-parameter logistic regression. Results were accepted when the standard curve was valid, blank wells were within the expected background range, and the duplicate CV was ≤ 10%. Samples exceeding this duplicate CV threshold were retested when sufficient samples remained.

The lower analytical detection limits used for result acceptance were 0.10 ng/mL for TLR2, 0.12 ng/mL for TLR4, and 0.15 ng/mL for TLR9. Final TLR2, TLR4, and TLR9 values were checked against sample identifiers, plate maps, duplicate-readout records, and sample-quality notes before statistical analysis.

Genomic DNA extraction

Genomic DNA was extracted from heparin-anticoagulated whole blood using a silica column-based blood genomic DNA extraction kit. Before extraction, anticoagulated blood samples were checked for correct labeling, sufficient volume, visible clotting, leakage, and storage condition.

DNA purity and concentration were assessed by spectrophotometry. Only samples with an A260/A280 ratio between 1.80 and 2.00 were accepted for downstream PCR amplification. Samples outside this range were remeasured after gentle mixing. If the abnormal purity ratio persisted, DNA extraction was repeated when sufficient whole blood remained.

DNA was diluted with nuclease-free water to a working concentration of 50 ng/µL and stored at -20 °C until PCR amplification. Before PCR setup, each DNA sample was checked for concentration, purity, sample identifier, and freeze-thaw history.

PCR-RFLP genotyping and genotype calling

Genotyping was performed using polymerase chain reaction-restriction fragment length polymorphism (PCR-RFLP) for IL-1B +3954 C/T, IL-10-1082 G/A, IL-10-819 T/C, TNF-α-308 G/A, TLR2-196 to-174 ins/del, TLR4 Asp299Gly, TLR4 Thr399Ile, TLR9 rs187084, IFN-γ +874 A/T, and IFN-γ +2108 A/G. Locus-specific primer sequences, expected amplicon sizes, restriction enzymes, and genotype-specific digestion patterns are provided in Supplementary Table 1.

PCR amplification was carried out in a final reaction volume of 25 µL containing 2.5 µL of 10 × buffer, 2.0 µL of dNTP mixture at 2.5 mM each, 1.0 µL of forward primer at 10 µM, 1.0 µL of reverse primer at 10 µM, 0.25 µL of Taq DNA polymerase at 5 U/µL, 2.0 µL of genomic DNA at 50 ng/µL, and 16.25 µL of nuclease-free water. PCR master mix was prepared in a clean pre-PCR area, and the DNA template was added last. A no-template negative control was included in each PCR run.

The thermal profile consisted of an initial denaturation at 95 °C for 5 min, followed by 35 cycles of denaturation at 95 °C for 30 s, annealing at 58–62 °C for 30 s depending on the primer pair, and extension at 72 °C for 30 s, with a final extension at 72 °C for 7 min. PCR products were first confirmed on 1.5% agarose gel before restriction digestion.

PCR amplification was accepted only when the expected amplicon band was visible, and the no-template control showed no amplification. Reactions showing weak bands, nonspecific amplification, primer-dimer interference, or contamination in the negative control were repeated from the original DNA template.

PCR products were digested with the corresponding restriction endonuclease at 37 °C for 3 h in a total digestion volume of 15 µL according to the manufacturer's instructions. Digested fragments were resolved on 2.5% agarose gel for most loci. The TLR2-196 to -174 ins/del locus was resolved on 8% polyacrylamide gel to improve fragment discrimination.

Restriction digestion was accepted only when fragment patterns matched the expected genotype-specific digestion profile. Reactions showing incomplete digestion, smeared bands, unexpected fragment sizes, or unclear separation were repeated. Bands were visualized under ultraviolet illumination after nucleic-acid staining. Gel images were saved with the sample identifier, locus name, run date, and gel number.

Genotype assignment was performed independently by two blinded laboratory investigators. Discordant or ambiguous genotype calls were reviewed using the original gel image and repeated from PCR amplification when necessary. Ten percent of randomly selected samples were re-genotyped for quality control, and the concordance rate was 99.2%. Final genotype data were entered into the analysis dataset only after sample identifiers, locus names, genotype calls, and quality-control records had been checked.

Statistical analysis

Statistical analyses were performed using an appropriate statistical analysis software. The clinical, ELISA, and genotype datasets were merged using de-identified patient codes. Patient codes were checked for duplication, mismatch, and missing values before analysis. A complete-case strategy was used because all included patients had available key clinical, inflammatory, and genotyping data.

In an appropriate statistical analysis software, pulmonary infection status was coded as the grouping variable, with the no pulmonary infection group used as the reference category. Continuous variables were first tested for normality using the Shapiro-Wilk test. Normally distributed continuous variables were expressed as mean ± standard deviation and compared using the independent-samples Student's t test. Non-normally distributed continuous variables were reported as medians with interquartile ranges and compared using the Mann-Whitney U test. Categorical variables were expressed as counts and percentages and compared using the chi-square test or Fisher's exact test, as appropriate.

Genotype distributions in the no pulmonary infection group were evaluated for Hardy-Weinberg equilibrium before association analysis. Genotype and allele frequencies were summarized as n (%). Odds ratios and 95% confidence intervals for pulmonary infection risk were estimated using multivariable logistic regression models. Pulmonary infection status was entered as the dependent variable. Genotype, allele, or biomarker variables were entered as independent variables, as specified for each analysis.

Age and sex were entered into the primary model a priori. BMI, chronic obstructive pulmonary disease, cardiac disease, cerebrospinal fluid leakage, indwelling time of the drain, and admission Glasgow Coma Scale category were entered simultaneously into the fully adjusted model because these variables were clinically relevant or showed baseline imbalance between groups. Multicollinearity was assessed using variance inflation factors (VIFs), and VIFs <5.0 were considered acceptable.

Because multiple polymorphisms were tested, Benjamini-Hochberg false discovery rate correction was applied to the genetic association analyses¹². P values from the genetic association tests were ranked from smallest to largest, and corrected significance was determined according to the Benjamini-Hochberg procedure. All tests were two-sided, and P <0.05 was considered statistically significant unless otherwise specified after multiple-comparison adjustment.

Final outputs were checked against the source dataset, ELISA quality-control records, genotype-calling records, and table values before manuscript preparation. Results were interpreted as associations rather than causal effects because the study used a retrospective observational design.

Results

Clinical indicators

Baseline clinical characteristics are summarized in Table 1. A total of 360 patients with spontaneous intracerebral hemorrhage were included in the final analysis, comprising 180 patients who developed nosocomial pulmonary infection during hospitalization and 180 patients who did not. There were no significant between-group differences in age (62.4 years ± 6.8 years vs 63.6 years ± 7.0 years, P > 0.05), sex distribution (142/180, 78.9% vs 137/180, 76.1%, P > 0.05), smoking history (94/180, 52.2% vs 99/180, 55.0%, P > 0.05), drinking history (99/180, 55.0% vs 106/180, 58.9%, P > 0.05), diabetes mellitus (18/180, 10.0% vs 21/180, 11.7%, P > 0.05), or hypertension (21/180, 11.7% vs 24/180, 13.3%, P > 0.05). In contrast, body mass index was significantly higher in the pulmonary infection group than in the no pulmonary infection group (24.9 kg/m² ± 1.9 kg/m² vs 23.8 kg/m² ± 1.8 kg/m², P < 0.001). Compared with the no pulmonary infection group, the pulmonary infection group more frequently had a history of chronic obstructive pulmonary disease (45/180, 25.0% vs 22/180, 12.2%, P < 0.001) and cardiac disease (23/180, 12.8% vs 6/180, 3.3%, P < 0.001), were more likely to have cerebrospinal fluid leakage (135/180, 75.0% vs 99/180, 55.0%, P < 0.001), and had drain indwelling time ≥ 3 days (131/180, 72.8% vs 79/180, 43.9%, P < 0.001). In addition, the pulmonary infection group had a higher proportion of patients with an admission Glasgow Coma Scale score <9 (142/180, 78.9%) than the control group (108/180, 60.0%; P < 0.001). These findings indicate a less favorable baseline clinical profile in patients who developed nosocomial pulmonary infection.

Characteristic	No pulmonary infection , n = 180	Pulmonary infection, n = 180	P value
Age, years	62.4 ± 6.8	63.6 ± 7.0	>0.05
Male, n (%)	142 (78.9)	137 (76.1)	>0.05
BMI, kg/m²	23.8 ± 1.8	24.9 ± 1.9	<0.001
Smokers, n (%)	94 (52.2)	99 (55.0)	>0.05
Drinking history, n (%)	99 (55.0)	106 (58.9)	>0.05
COPD history, n (%)	22 (12.2)	45 (25.0)	<0.001
Cardiac history, n (%)	6 (3.3)	23 (12.8)	<0.001
Diabetes mellitus, n (%)	18 (10.0)	21 (11.7)	>0.05
Hypertension, n (%)	21 (11.7)	24 (13.3)	>0.05
CSF leakage, n (%)	99 (55.0)	135 (75.0)	<0.001
Drain indwelling time < 3 days, n (%)	101 (56.1)	49 (27.2)	<0.001
Drain indwelling time ≥ 3 days, n (%)	79 (43.9)	131 (72.8)	<0.001
GCS at admission ≥ 9, n (%)	72 (40.0)	38 (21.1)	<0.001
GCS at admission < 9, n (%)	108 (60.0)	142 (78.9)	<0.001

Table 1: Baseline clinical characteristics of patients with spontaneous intracerebral hemorrhage stratified by nosocomial pulmonary infection status. The table summarizes demographic variables, vascular risk factors, and major admission-related clinical characteristics, including age, sex, smoking history, drinking history, diabetes mellitus, hypertension, body mass index, chronic obstructive pulmonary disease, cardiac disease, cerebrospinal fluid leakage, indwelling time of drains, and admission Glasgow Coma Scale category. Continuous variables are presented as mean ± SD and were compared using the independent-samples Student's t test or the Mann-Whitney U test, as appropriate according to data distribution. Categorical variables are presented as n (%) and were compared using the chi-square test or Fisher's exact test, as appropriate. All P values are two-sided. Abbreviations: BMI, body mass index; COPD, chronic obstructive pulmonary disease; CSF, cerebrospinal fluid; GCS, Glasgow Coma Scale. Please click here to download this Table.

Serum cytokine and TLR2/TLR4/TLR9 profiles

Serum cytokine and Toll-like receptor protein results are presented in Figure 1(A–I). Blood samples were collected at 08:00 on the morning after admission, whereas pulmonary infections diagnosed within the first 48 h were excluded by definition. Accordingly, these measurements represented early post-admission inflammatory profiles in relation to subsequent nosocomial pulmonary infection classification. Compared with the no pulmonary infection group, the pulmonary infection group showed significantly higher serum concentrations of IL-1β (17.4 pg/mL ± 3.1 pg/mL vs 33.2 pg/mL ± 2.4 pg/mL, P < 0.001), IL-6 (18.1 pg/mL ± 2.2 pg/mL vs 33.8 pg/mL ± 2.5 pg/mL, P < 0.001), IL-10 (17.6 pg/mL ± 2.0 pg/mL vs 34.1 pg/mL ± 2.8 pg/mL, P < 0.001), IL-17 (22.9 pg/mL ± 2.8 pg/mL vs 28.3 pg/mL ± 4.0 pg/mL, P < 0.001), IFN-γ (18.0 pg/mL ± 2.1 pg/mL vs 33.0 pg/mL ± 2.7 pg/mL, P < 0.001), and TNF-α (19.2 pg/mL ± 2.3 pg/mL vs 35.5 pg/mL ± 3.1 pg/mL, P < 0.001). Serum levels of TLR2 (1.94 ng/mL ± 0.29 ng/mL vs 2.71 ng/mL ± 0.36 ng/mL, P < 0.001), TLR4 (2.08 ng/mL ± 0.31 ng/mL vs 3.06 ng/mL ± 0.42 ng/mL, P < 0.001), and TLR9 (1.89 ng/mL ± 0.27 ng/mL vs 2.63 ng/mL ± 0.35 ng/mL, P < 0.001) were also significantly higher in the pulmonary infection group than in the no pulmonary infection group. Overall, patients who developed nosocomial pulmonary infections showed broad elevations in serum inflammatory and innate immune markers.

Figure 1: Serum cytokine concentrations and serum TLR2, TLR4, and TLR9 protein levels in patients with spontaneous intracerebral hemorrhage with or without nosocomial pulmonary infection. (A) IL-1β, (B) IL-6, (C) IL-10, (D) IL-17, (E) IFN-γ, (F) TNF-α, (G) TLR2, (H) TLR4, and (I) TLR9. Data are shown as mean ± SD for the no pulmonary infection group (n = 180) and the pulmonary infection group (n = 80). Group comparisons were performed using the independent-samples Student's t test or the Mann-Whitney U test, as appropriate. *P < 0.05; **P < 0.01. Please click here to view a larger version of this figure.

Gene polymorphism analyses

Genotype and allele distributions were compared between groups using logistic regression models adjusted for age and sex. Sensitivity analyses, additionally adjusting for body mass index, chronic obstructive pulmonary disease, cardiac disease, cerebrospinal fluid leakage, the drain's indwelling time, and admission Glasgow Coma Scale category, reportedly yielded directionally consistent findings.

Toll-like receptor polymorphism results are shown in Table 2. Significant associations with pulmonary infection risk were observed for TLR2-196 to-174 ins/del and TLR9 rs187084. Compared with the I/I genotype, the D/D genotype of TLR2-196 to-174 ins/del was associated with increased risk (OR 1.93, 95% CI 1.15–3.23, P = 0.017), and the D allele was also associated with higher risk (OR 1.57, 95% CI 1.16–2.12, P = 0.004). For TLR9 rs187084, both the TC genotype (OR 2.69, 95% CI 1.59–4.46, P < 0.001) and CC genotype (OR 2.94, 95% CI 1.64–5.27, P < 0.001) were associated with increased pulmonary infection risk, and the C allele was similarly enriched in the pulmonary infection group (OR 1.77, 95% CI 1.32–2.38, P < 0.001). No statistically significant associations were observed for TLR2 rs3804099, TLR2 rs3804100, TLR4 rs10759932, TLR4 rs1927911, TLR4 rs11536889, TLR9 rs352140, or TLR9 rs574836.

TLR Locus	Genotype / Allele	No Pulmonary Infection, n (%)	Pulmonary Infection, n (%)	Adjusted OR (95% CI)a	Pa
TLR2 −196 to −174 ins/del	I/I	57 (31.7)	38 (21.1)	1.00 (Ref)	—
	I/D	53 (29.4)	52 (28.9)	1.47 (0.84–2.58)	0.226
	D/D	70 (38.9)	90 (50.0)	1.93 (1.15–3.23)	0.017
	I	167 (46.4)	128 (35.6)	1.00 (Ref)	—
	D	193 (53.6)	232 (64.4)	1.57 (1.16–2.12)	0.004
TLR2 rs3804099	TT	94 (52.2)	92 (51.1)	1.00 (Ref)	—
	TC	75 (41.7)	72 (40.0)	0.98 (0.64–1.51)	0.982
	CC	11 (6.1)	16 (8.9)	1.49 (0.65–3.37)	0.456
	T	263 (73.1)	256 (71.1)	1.00 (Ref)	—
	C	97 (26.9)	104 (28.9)	1.10 (0.80–1.53)	0.618
TLR2 rs3804100	TT	99 (55.0)	98 (54.4)	1.00 (Ref)	—
	TC	61 (33.9)	62 (34.4)	1.03 (0.65–1.61)	1
	CC	20 (11.2)	20 (11.2)	1.01 (0.51–1.99)	0.885
	T	259 (71.9)	258 (71.7)	1.00 (Ref)	—
	C	101 (28.1)	102 (28.3)	1.01 (0.73–1.40)	1
TLR4 rs10759932	TT	104 (57.7)	108 (60.0)	1.00 (Ref)	—
	TC	61 (33.9)	57 (31.7)	0.90 (0.57–1.41)	0.73
	CC	15 (8.3)	15 (8.3)	0.96 (0.45–2.07)	0.922
	T	269 (74.7)	273 (75.8)	1.00 (Ref)	—
	C	91 (25.3)	87 (24.2)	0.94 (0.67–1.32)	0.796
TLR4 rs1927911	GG	99 (55.0)	98 (54.4)	1.00 (Ref)	—
	GA	61 (33.9)	62 (34.4)	1.03 (0.65–1.61)	1
	AA	20 (11.2)	20 (11.2)	1.01 (0.51–1.99)	0.885
	G	259 (71.9)	258 (71.7)	1.00 (Ref)	—
	A	101 (28.1)	102 (28.3)	1.01 (0.73–1.40)	1
TLR4 rs11536889	GG	92 (51.1)	88 (48.9)	1.00 (Ref)	—
	GC	64 (35.6)	67 (37.2)	1.09 (0.70–1.72)	0.781
	CC	24 (13.3)	25 (13.9)	1.09 (0.58–2.05)	0.918
	G	248 (68.9)	243 (67.5)	1.00 (Ref)	—
	C	112 (31.1)	117 (32.5)	1.07 (0.78–1.46)	0.689
TLR9 rs187084	TT	66 (36.7)	31 (17.2)	1.00 (Ref)	—
	TC	72 (40.0)	91 (50.6)	2.69 (1.59–4.46)	< 0.001
	CC	42 (23.3)	58 (32.2)	2.94 (1.64–5.27)	< 0.001
	T	204 (56.7)	153 (42.5)	1.00 (Ref)	—
	C	156 (43.3)	207 (57.5)	1.77 (1.32–2.38)	< 0.001
TLR9 rs352140	GG	27 (15.0)	27 (15.0)	1.00 (Ref)	—
	GA	102 (56.7)	99 (55.0)	0.97 (0.53–1.77)	0.955
	AA	51 (28.3)	54 (30.0)	1.06 (0.55–2.04)	0.997
	G	156 (43.3)	153 (42.5)	1.00 (Ref)	—
	A	204 (56.7)	207 (57.5)	1.03 (0.77–1.39)	0.88
TLR9 rs574836	TT	27 (15.0)	26 (14.4)	1.00 (Ref)	—
	TC	99 (55.0)	101 (56.1)	1.06 (0.58–1.94)	0.974
	CC	54 (30.0)	53 (29.4)	1.02 (0.53–1.97)	0.911
	T	153 (42.5)	153 (42.5)	1.00 (Ref)	—
	C	207 (57.5)	207 (57.5)	1.00 (0.74–1.34)	0.94

Table 2: Genotype and allele frequencies of TLR2, TLR4, and TLR9 polymorphisms for pulmonary infection in patients with spontaneous intracerebral hemorrhage. The table presents genotype and allele distributions for TLR2-196 to-174 ins/del, TLR2 rs3804099, TLR2 rs3804100, TLR4 rs10759932, TLR4 rs1927911, TLR4 rs11536889, TLR9 rs187084, TLR9 rs352140, and TLR9 rs574836. Genotype and allele distributions are presented as n (%). Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using logistic regression models adjusted for age and sex, with reference categories indicated as Ref. All P values are two-sided. Abbreviations: OR, odds ratio; CI, confidence interval; Ref, reference. ^aAdjusted for age and sex. Please click here to download this Table.

Interleukin-related polymorphism results are detailed in Table 3. Significant associations were observed for IL-1B +3954 C/T, IL-10-1082 G/A, and IL-10-819 T/C. For IL-1B +3954 C/T, the TC genotype (OR 2.52, 95% CI 1.33–4.79, P = 0.007), TT genotype (OR 2.30, 95% CI 1.34–3.96, P = 0.003), and T allele (OR 1.70, 95% CI 1.24–2.32, P = 0.001) were associated with increased pulmonary infection risk. For IL-10-1082 G/A, both the GA genotype (OR 2.60, 95% CI 1.57–4.32, P < 0.001) and AA genotype (OR 2.05, 95% CI 1.22–3.45, P = 0.009) were associated with increased risk, as was the A allele (OR 1.66, 95% CI 1.23–2.23, P = 0.001). For IL-10-819 T/C, the CC genotype (OR 3.54, 95% CI 1.72–7.29, P < 0.001) and the C allele (OR 1.39, 95% CI 1.01–1.91, P = 0.043) were associated with increased pulmonary infection risk, whereas the TC genotype showed an OR below 1.00 (OR 0.51, 95% CI 0.31–0.82, P = 0.007). No significant associations were found for IL-1B-511 C/T, IL-1B-31 C/T, IL-6-174 G/C, IL-6-1363 G/T, IL-6-1363 C/G, IL-8-251 A/T, IL-8 rs2227306, IL-8 rs1126647, IL-10-592 C/A, IL-17 rs2275913, or IL-17 rs763780.

Gene Locus	Genotype / Allele	No Pulmonary Infection, n (%)	Pulmonary Infection, n (%)	Adjusted OR (95% CI)	P Value
IL-1B +3954 C/T	CC	53 (29.4)	27 (15.0)	1.00 (Ref)	—
	TC	35 (19.4)	45 (25.0)	2.52 (1.33–4.79)	0.007
	TT	92 (51.1)	108 (60.0)	2.30 (1.34–3.96)	0.003
	C	141 (39.2)	99 (27.5)	1.00 (Ref)	—
	T	219 (60.8)	261 (72.5)	1.70 (1.24–2.32)	0.001
IL-1B −511 C/T	CC	94 (52.2)	92 (51.1)	1.00 (Ref)	—
	TC	75 (41.7)	75 (41.6)	1.02 (0.66–1.57)	0.991
	TT	11 (6.1)	13 (7.2)	1.21 (0.51–2.83)	0.828
	C	263 (73.1)	259 (71.9)	1.00 (Ref)	—
	T	97 (26.9)	101 (28.1)	1.06 (0.76–1.47)	0.802
IL-1B −31 C/T	CC	89 (49.4)	82 (45.6)	1.00 (Ref)	—
	TC	78 (43.3)	77 (42.8)	1.07 (0.69–1.66)	0.841
	TT	13 (7.2)	21 (11.6)	1.75 (0.82–3.73)	0.199
	C	256 (71.1)	241 (66.9)	1.00 (Ref)	—
	T	104 (28.9)	119 (33.1)	1.22 (0.89–1.67)	0.227
IL-6 −174 G/C	GG	92 (51.1)	88 (48.9)	1.00 (Ref)	—
	GC	75 (41.6)	74 (41.1)	1.03 (0.67–1.59)	0.977
	CC	13 (7.2)	18 (10.0)	1.45 (0.67–3.13)	0.454
	G	259 (71.9)	250 (69.4)	1.00 (Ref)	—
	C	101 (28.1)	110 (30.6)	1.13 (0.82–1.56)	0.512
IL-6 −1363 G/T	GG	92 (51.1)	85 (47.2)	1.00 (Ref)	—
	GT	78 (43.3)	75 (41.7)	1.04 (0.68–1.60)	0.944
	TT	10 (5.6)	20 (11.1)	2.16 (0.96–4.89)	0.091
	G	262 (72.8)	245 (68.1)	1.00 (Ref)	—
	T	98 (27.2)	115 (31.9)	1.25 (0.91–1.73)	0.191
IL-6 −1363 C/G	CC	95 (52.7)	93 (51.1)	1.00 (Ref)	—
	GC	75 (41.7)	71 (40.0)	0.97 (0.63–1.49)	0.967
	GG	10 (5.6)	16 (8.9)	1.63 (0.71–3.79)	0.345
	C	265 (73.6)	257 (71.4)	1.00 (Ref)	—
	G	95 (26.4)	103 (28.6)	1.12 (0.81–1.55)	0.559
IL-8 −251 A/T	AA	27 (15.0)	26 (14.4)	1.00 (Ref)	—
	AT	107 (59.4)	111 (61.7)	1.08 (0.59–1.96)	0.928
	TT	46 (25.6)	43 (23.9)	0.97 (0.49–1.92)	0.93
	A	161 (44.7)	163 (45.3)	1.00 (Ref)	—
	T	199 (55.3)	197 (54.7)	0.98 (0.73–1.31)	0.94
IL-8 rs2227306	CC	27 (15.0)	25 (13.9)	1.00 (Ref)	—
	CT	99 (55.0)	104 (57.8)	1.13 (0.62–2.09)	0.685
	TT	54 (30.0)	51 (28.3)	1.02 (0.52–1.98)	0.911
	C	153 (42.5)	154 (42.8)	1.00 (Ref)	—
	T	207 (57.5)	206 (57.2)	0.99 (0.74–1.33)	1
IL-8 rs1126647	AA	32 (17.8)	36 (20.0)	1.00 (Ref)	—
	AT	89 (49.4)	81 (45.0)	0.81 (0.46–1.42)	0.552
	TT	59 (32.8)	63 (35.0)	0.95 (0.52–1.72)	0.983
	A	153 (42.5)	153 (42.5)	1.00 (Ref)	—
	T	207 (57.5)	207 (57.5)	1.00 (0.74–1.34)	0.94
IL-10 −592 C/A	CC	94 (52.2)	92 (51.1)	1.00 (Ref)	—
	CA	75 (41.7)	75 (41.6)	1.02 (0.66–1.57)	0.991
	AA	11 (6.1)	13 (7.2)	1.21 (0.51–2.83)	0.828
	C	263 (73.1)	259 (71.9)	1.00 (Ref)	—
	A	97 (26.9)	101 (28.1)	1.06 (0.76–1.47)	0.802
IL-10 −1082 G/A	GG	85 (47.2)	50 (27.8)	1.00 (Ref)	—
	GA	47 (26.1)	72 (40.0)	2.60 (1.57–4.32)	< 0.001
	AA	48 (26.7)	58 (32.2)	2.05 (1.22–3.45)	0.009
	G	217 (60.3)	172 (47.8)	1.00 (Ref)	—
	A	143 (39.7)	188 (52.2)	1.66 (1.23–2.23)	0.001
IL-10 −819 T/C	TT	94 (52.2)	99 (55.0)	1.00 (Ref)	—
	TC	75 (41.7)	40 (22.2)	0.51 (0.31–0.82)	0.007
	CC	11 (6.1)	41 (22.8)	3.54 (1.72–7.29)	< 0.001
	T	263 (73.1)	238 (66.1)	1.00 (Ref)	—
	C	97 (26.9)	122 (33.9)	1.39 (1.01–1.91)	0.043
IL-17 rs2275913	GG	94 (52.2)	99 (55.0)	1.00 (Ref)	—
	GA	79 (43.9)	72 (40.0)	0.92 (0.60–1.42)	0.801
	AA	7 (3.9)	9 (5.0)	1.22 (0.44–3.41)	0.904
	G	267 (74.2)	270 (75.0)	1.00 (Ref)	—
	A	93 (25.8)	90 (25.0)	0.96 (0.68–1.34)	0.864
IL-17 rs763780	TT	110 (61.1)	100 (55.6)	1.00 (Ref)	—
	TC	62 (34.4)	70 (38.9)	1.24 (0.80–1.92)	0.388
	CC	8 (4.4)	10 (4.4)	1.38 (0.52–3.62)	0.688
	T	263 (73.1)	270 (75.0)	1.00 (Ref)	—
	C	97 (26.9)	90 (25.0)	0.90 (0.65–1.26)	0.61

Table 3: Genotype and allele frequencies of interleukin-related polymorphisms for pulmonary infection in patients with spontaneous intracerebral hemorrhage. The table presents genotype and allele distributions for IL-1B +3954 C/T, IL-1B-511 C/T, IL-1B-31 C/T, IL-6-174 G/C, IL-6-1363 G/T, IL-6-1363 C/G, IL-8-251 A/T, IL-8 rs2227306, IL-8 rs1126647, IL-10-592 C/A, IL-10-1082 G/A, IL-10-819 T/C, IL-17 rs2275913, and IL-17 rs763780. Genotype and allele distributions are presented as n (%). Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using logistic regression models adjusted for age and sex, with reference categories indicated as Ref. All P values are two-sided. Abbreviations: IL, interleukin; OR, odds ratio; CI, confidence interval; Ref, reference. ^aAdjusted for age and sex. Please click here to download this Table.

IFN-γ polymorphism results are presented in Table 4. Significant associations were observed for +874 A/T but not for +2108 A/G. Compared with the AA genotype, the AT genotype (OR 1.83, 95% CI 1.18–2.83, P = 0.009) and TT genotype (OR 2.97, 95% CI 1.19–7.38, P = 0.029) at +874 A/T were associated with increased pulmonary infection risk, and the T allele was also associated with higher risk (OR 1.57, 95% CI 1.15–2.15, P = 0.006). In contrast, no significant associations were identified for the +2108 A/G variant.

Locus	Genotype / allele	No pulmonary infection, n (%)	Pulmonary infection, n (%)	Adjusted OR (95% CI)a	Pa
+874 A/T	AA	86 (47.8)	58 (32.2)	1.00 (Ref)	—
	AT	86 (47.8)	106 (58.9)	1.83 (1.18–2.83)	0.009
	TT	8 (4.4)	16 (8.9)	2.97 (1.19–7.38)	0.029
	A	258 (71.7)	222 (61.7)	1.00 (Ref)	—
	T	102 (28.3)	138 (38.3)	1.57 (1.15–2.15)	0.006
+2108 A/G	AA	95 (52.8)	90 (50.0)	1.00 (Ref)	—
	AG	63 (35.0)	66 (36.7)	1.11 (0.71–1.73)	0.746
	GG	22 (12.2)	24 (13.3)	1.15 (0.60–2.20)	0.792
	A	253 (70.3)	246 (68.3)	1.00 (Ref)	—
	G	107 (29.7)	114 (31.7)	1.10 (0.80–1.50)	0.628

Table 4: Genotype and allele frequencies of IFN-γ polymorphisms for pulmonary infection in patients with spontaneous intracerebral hemorrhage. The table presents genotype and allele distributions for IFN-γ +874 A/T and IFN-γ +2108 A/G. Genotype and allele distributions are presented as n (%). Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using logistic regression models adjusted for age and sex, with reference categories indicated as Ref. All P values are two-sided. Abbreviations: IFN-γ, interferon gamma; OR, odds ratio; CI, confidence interval; Ref, reference. ^aAdjusted for age and sex. Please click here to download this Table.

TNF-α polymorphism results are shown in Table 5. Among these loci,-308 G/A was significantly associated with pulmonary infection risk, whereas -863 C/A and -238 G/A were not. Compared with the GG genotype, the GA genotype (OR 2.34, 95% CI 1.42–3.88, P = 0.001) and AA genotype (OR 1.84, 95% CI 1.10–3.07, P = 0.028) at-308 G/A were associated with increased risk, and the A allele was also associated with higher risk (OR 1.55, 95% CI 1.15–2.08, P = 0.004).

Locus	Genotype / Allele	No Pulmonary Infection, n (%)	Pulmonary Infection, n (%)	Adjusted OR (95% CI)	P-value
−308 G/A	GG	85 (47.2)	54 (30.0)	1.00 (Ref)	—
	GA	47 (26.1)	70 (38.9)	2.34 (1.42–3.88)	0.001
	AA	48 (26.7)	56 (31.1)	1.84 (1.10–3.07)	0.028
	G	217 (60.3)	178 (49.4)	1.00 (Ref)	—
	A	143 (39.7)	182 (50.6)	1.55 (1.15–2.08)	0.004
−863 C/A	GG	92 (51.1)	85 (47.2)	1.00 (Ref)	—
	GA	78 (43.3)	75 (41.7)	1.04 (0.68–1.60)	0.944
	AA	10 (5.6)	20 (11.1)	2.16 (0.96–4.89)	0.091
	G	262 (72.8)	245 (68.1)	1.00 (Ref)	—
	A	98 (27.2)	115 (31.9)	1.25 (0.91–1.73)	0.191
−238 G/A	CC	32 (17.8)	36 (20.0)	1.00 (Ref)	—
	CA	89 (49.4)	81 (45.0)	0.81 (0.46–1.42)	0.552
	AA	59 (32.8)	63 (35.0)	0.95 (0.52–1.72)	0.983
	C	153 (42.5)	153 (42.5)	1.00 (Ref)	—
	A	207 (57.5)	207 (57.5)	1.00 (0.74–1.34)	0.94

Table 5: Genotype and allele frequencies of TNF-α polymorphisms for pulmonary infection in patients with spontaneous intracerebral hemorrhage. The table presents genotype and allele distributions for TNF-α-308 G/A, TNF-α-863 C/A, and TNF-α-238 G/A. Genotype and allele distributions are presented as n (%). Odds ratios (ORs) and 95% confidence intervals (CIs) were estimated using logistic regression models adjusted for age and sex, with reference categories indicated as Ref. All P values are two-sided. Abbreviations: TNF-α, tumor necrosis factor alpha; OR, odds ratio; CI, confidence interval; Ref, reference. ^aAdjusted for age and sex. Please click here to download this Table.

Taken together, the results showed that patients who developed nosocomial pulmonary infection after spontaneous intracerebral hemorrhage had a more adverse baseline clinical profile, higher circulating inflammatory cytokine and Toll-like receptor levels, and distinct distributions of several immune-related genetic polymorphisms. The most consistent positive associations in this cohort were observed for TLR2-196 to-174 ins/del, TLR9 rs187084, IL-1B +3954 C/T, IL-10-1082 G/A, IL-10-819 T/C, IFN-γ +874 A/T, and TNF-α-308 G/A.

DATA AVAILABILITY:

All raw data supporting the findings of this study are publicly available in the figshare repository. The deposited dataset contains de-identified clinical data, serum cytokine and Toll-like receptor measurements, and genotype data corresponding to Figure 1(A–I), Table 1, Table 2, Table 3, Table 4, and Table 5. The dataset is available as: Yang, Lunyun; Zhou, Hui; Li, Wei; Shen, Weifeng (2026), Cytokines and Inflammatory Gene Polymorphisms Associated with Nosocomial Pulmonary Infection After Spontaneous Intracerebral Hemorrhage, figshare, Dataset, https://doi.org/10.6084/m9.figshare.32101399.v1.

Supplementary Table 1: Primer sequences, expected PCR amplicon sizes, restriction enzymes, and genotype-specific digestion patterns used for PCR-RFLP genotyping. The table summarizes the locus-specific forward and reverse primer sequences, expected PCR product sizes, restriction enzymes, and expected fragment patterns for the inflammatory and innate immune-related polymorphisms analyzed in this study, including TLR2, TLR4, TLR9, IL-1β, IL-10, TNF-α, and IFN-γ loci. These details support the reproducibility of the PCR-RFLP workflow and provide the reference information used for amplification verification, restriction digestion interpretation, and genotype calling.Please click here to download this file.

Discussion

Spontaneous intracerebral hemorrhage (ICH) remains one of the most severe stroke subtypes and continues to impose substantial early mortality and long-term disability burdens¹³. In clinical practice, patients with ICH are especially vulnerable to in-hospital complications because neurological injury can weaken airway protection, reduce cough effectiveness, restrict mobility, and increase exposure to invasive supportive procedures. Among these complications, nosocomial pulmonary infection is particularly important because it may prolong hospitalization, increase resource use, and further compromise neurological recovery through impaired oxygenation and systemic inflammatory stress¹⁴. These clinical realities reinforce the need for earlier identification of patients at increased infection risk and for more focused surveillance during the acute stage of care¹⁵.

In the present study, patients who developed nosocomial pulmonary infection had a less favorable baseline clinical profile, including more frequent chronic obstructive pulmonary disease, cardiac disease, cerebrospinal fluid leakage, longer indwelling time of drains, and lower admission Glasgow Coma Scale scores. This pattern is clinically plausible because reduced cardiopulmonary reserve may weaken tolerance to respiratory stress, whereas impaired consciousness and prolonged device exposure may increase aspiration risk and facilitate airway colonization¹⁶. Rather than functioning as isolated indicators, these factors may reflect a broader vulnerability state in which neurological severity, comorbidity burden, and procedural exposure collectively shape the likelihood of subsequent pulmonary infection¹⁷.

Beyond baseline clinical characteristics, this study found that patients who later developed nosocomial pulmonary infection had higher serum concentrations of IL-1β, IL-6, IL-10, IL-17, IFN-γ, and TNF-α, as well as higher serum protein levels of TLR2, TLR4, and TLR9. This pattern suggests that a more activated inflammatory and innate immune-related profile is present early after admission in patients who subsequently experience pulmonary infection. At the same time, these findings should be interpreted cautiously. Because blood was collected on the morning after admission and infection was defined after the first 48 h of hospitalization, the observed biomarker differences cannot be interpreted as direct evidence of causality¹⁸. They may reflect early host inflammatory activation associated with later infection susceptibility, but they may also partly capture systemic responses to hemorrhagic severity or evolving subclinical pathological processes¹⁹. Even with this limitation, the biomarker pattern remains clinically relevant because it supports the possibility that inflammatory profiling may contribute to earlier risk recognition in hospitalized patients²⁰.

The genetic findings add a further layer of heterogeneity to this picture. In this cohort, significant associations were observed for TLR2-196 to-174 ins/del, TLR9 rs187084, IL-1B +3954 C/T, IL-10-1082 G/A, IL-10-819 T/C, IFN-γ +874 A/T, and TNF-α-308 G/A. Prior studies have shown that variation in cytokine genes and innate immune receptor genes can influence inflammatory responsiveness and host defense across infectious and inflammatory conditions²¹. Evidence related to TLR2 and TLR9 has likewise suggested that these receptors may contribute to inter-individual differences in pathogen recognition and downstream immune signaling²². Comparable observations have also been reported for interleukin-related loci, especially where promoter or coding variation may alter the balance between pro-inflammatory and anti-inflammatory responses²³. Studies of IFN-γ polymorphisms have further indicated that locus-specific differences may affect infection susceptibility in certain populations²⁴, while TNF-α promoter variation has been linked to altered inflammatory intensity and infection-related risk in previous association studies²⁵. At the same time, not all loci examined in the present study were associated with pulmonary infection risk, which suggests that the host-genetic contribution in this setting is selective rather than generalized²⁶.

A key value of the present work is that it provides a reproducible clinical-laboratory workflow rather than relying only on bedside risk factors or isolated biomarker testing. Existing clinical-only approaches are practical because they use routinely available information such as neurological severity, comorbidities, swallowing impairment, invasive procedures, and device exposure. However, clinical variables alone may not fully capture the biological differences that influence infection susceptibility among patients with similar clinical presentations. Biomarker-only approaches can describe inflammatory activation but may be sensitive to sampling time, specimen handling, hemolysis, storage conditions, and assay variation. Genotyping provides information about a relatively stable host immune response background, but genetic data alone cannot reflect the patient's current inflammatory state. The combined workflow used here, therefore, has a practical advantage: it integrates clinical vulnerability, early serum inflammatory activity, and selected host immunogenetic variation within a single analytical framework.

Several protocol steps are particularly important for the reliability of this workflow. The first is blood-collection timing. In this study, blood was collected at a fixed early post-admission time point, and pulmonary infections diagnosed within the first 48 h after admission were excluded. This timing reduces, although does not eliminate, the possibility that measured biomarkers simply reflect established pulmonary infection. The second critical step is sample handling. Hemolysis, delayed serum separation, repeated freeze–thaw cycles, lipemia, and insufficient residual volume can affect ELISA reliability and should be documented before analysis. The third critical step is ELISA duplicate acceptance. Duplicate sample coefficient of variation should remain within the predefined acceptance threshold, and samples exceeding this threshold should be re-assayed when sufficient serum remains. The fourth critical step is DNA quality control. DNA purity and concentration should be checked before PCR amplification, because poor DNA quality can lead to weak bands, nonspecific amplification, or failed digestion. The fifth critical step is genotype calling. PCR amplification, restriction digestion, gel-band interpretation, independent genotype assignment, and repeat genotyping of selected samples are all necessary to reduce misclassification.

Common troubleshooting points are also important for future use of this workflow. For ELISA testing, high background may indicate incomplete washing, contaminated reagents, excessive incubation, or plate-edge effects. A weak signal may reflect degraded samples, incorrect reagent preparation, or insufficient sample concentration. Inconsistent duplicates should prompt review of pipetting accuracy, plate layout, sample mixing, and residual sample quality. For PCR-RFLP genotyping, failed amplification should first be checked against DNA concentration, primer performance, thermal-cycling conditions, and contamination in the no-template control. Smearing or nonspecific bands may require optimization of annealing temperature, template quantity, or gel concentration. Incomplete restriction digestion may result from insufficient enzyme activity, incorrect buffer, inadequate incubation time, or inhibitors carried over from PCR. Ambiguous gel bands should not be assigned directly; they should be reviewed by two blinded investigators and repeated from PCR amplification when needed.

The workflow also has technical limitations. ELISA measurements can be affected by pre-analytical variation, matrix effects, lot-to-lot differences in kits, standard-curve fitting, and limited dynamic range. A single early blood sample cannot describe the full temporal evolution of inflammatory responses after ICH. PCR-RFLP is accessible and relatively low-cost, but it has lower throughput than sequencing-based methods and depends on clear amplification, complete restriction digestion, and interpretable gel bands. Some polymorphisms may also require alternative genotyping methods if the expected fragment sizes are too close to be reliably separated. In addition, the selected loci represent only part of the immune-genetic background relevant to infection susceptibility. Therefore, the present workflow should be interpreted as a structured, reproducible approach for targeted assessment of biomarkers and polymorphisms, rather than a comprehensive immunogenomic platform.

Taken together, the present findings support a layered interpretation of infection susceptibility after spontaneous ICH. Baseline clinical vulnerability, early inflammatory activity, and host-genetic background do not appear to operate independently; rather, they may represent complementary dimensions of risk. From a translational perspective, practical early-risk assessment may ultimately benefit from a combined framework that integrates clinical characteristics with biomarker and genotype information instead of relying on any single type of indicator alone²⁷. Although the present study was not designed to build or validate a formal prediction model, the standardized workflow, explicit sample-processing checkpoints, ELISA acceptance criteria, genotyping quality-control steps, and open data structure provide a basis for reproducibility in future cohorts²⁸.

Several study-level limitations should also be acknowledged. First, this was a single-center retrospective study, which may limit generalizability and leaves the analysis vulnerable to selection bias and unmeasured confounding. Second, although the sampling design established a temporal gap between admission blood collection and outcome classification, the study still cannot fully separate predictive inflammatory signatures from early evolving disease processes. Third, multiple polymorphic loci were evaluated, and the findings therefore require cautious interpretation until replicated in independent cohorts. Fourth, external validation and mechanistic confirmation were not performed in the present study. These limitations indicate that the study should be understood as providing clinically relevant association signals and a reproducible analytical workflow rather than definitive mechanistic proof.

In summary, patients with spontaneous ICH who developed nosocomial pulmonary infection showed a more adverse clinical profile, higher circulating levels of inflammatory cytokines, higher serum levels of TLR2, TLR4, and TLR9, and distinct distributions of several immune-related genetic polymorphisms. The revised workflow clarifies how early blood collection, sample quality control, ELISA duplicate acceptance, DNA quality assessment, PCR-RFLP verification, genotype calling, and statistical analysis can be combined to support reproducible infection-risk profiling. These findings provide a basis for future multicenter studies to validate clinically usable early-risk stratification strategies after spontaneous ICH.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Human TNF-α ELISA Kit	Invitrogen / Thermo Fisher Scientific	KHC3011	ELISA kit for quantitative measurement of human TNF-α in serum.
Human TLR2 / Toll-like Receptor 2 ELISA Kit	Sigma-Aldrich / Merck	RAB0744-1KT	96-well ELISA kit for measurement of human TLR2 in serum or plasma samples.
Human TLR4 ELISA Kit, colorimetric	Novus Biologicals / Bio-Techne	NBP3-11807	Colorimetric sandwich ELISA kit for measurement of human TLR4 in serum, plasma, or other biological fluids.
Human TLR9 ELISA Kit, colorimetric	Novus Biologicals / Bio-Techne	NBP2-76574	Colorimetric ELISA kit for measurement of human TLR9 in serum samples.
Absorbance microplate reader	Agilent BioTek	800 TS	Microplate reader for ELISA optical-density measurement at 450 nm with reference correction at 570 nm.
QIAamp DNA Blood Mini Kit	QIAGEN	51104	Silica membrane-based kit for genomic DNA extraction from anticoagulated whole blood.
NanoDrop One Microvolume UV-Vis Spectrophotometer	Thermo Scientific	ND-ONE-W	Microvolume spectrophotometer for DNA concentration and A260/A280 purity assessment.
SimpliAmp Thermal Cycler	Applied Biosystems / Thermo Fisher Scientific	A24811	Thermal cycler used for endpoint PCR amplification in PCR-RFLP genotyping.
PCR Master Mix, 2X	Thermo Scientific	K0171	Premixed PCR reagent containing Taq DNA polymerase, dNTPs, MgCl2, and reaction buffer; used for conventional PCR amplification.
Recombinant Taq DNA Polymerase	Thermo Scientific	EP0401	Alternative enzyme source for locus-specific PCR reactions when separate PCR components are used instead of 2X master mix.
Nuclease-free water	Thermo Scientific	R0581	Molecular-biology-grade water used for PCR setup, DNA dilution, and no-template controls.
Custom DNA primers	Sangon Biotech	Custom synthesis	Locus-specific primers for PCR-RFLP genotyping; primer sequences are listed in Supplementary Table S1.
Restriction endonucleases for PCR-RFLP	Thermo Scientific / New England Biolabs	Locus-specific enzymes listed in Supplementary Table S1	Restriction enzymes used for genotype-specific digestion of PCR products. The enzyme selected for each locus should match the digestion pattern listed in Supplementary Table S1.
Agarose I, molecular biology grade	Thermo Scientific	17850	Agarose for preparation of 1.5% and 2.5% gels used to confirm PCR products and resolve most RFLP fragments.
GeneRuler 50 bp DNA Ladder	Thermo Scientific	SM0371	DNA size marker for PCR amplicons and RFLP fragment-size estimation.
SYBR Safe DNA Gel Stain	Invitrogen / Thermo Fisher Scientific	S33102	Nucleic-acid gel stain for visualization of DNA in agarose or acrylamide gels; used as a reduced-hazard alternative to ethidium bromide.
Horizontal agarose gel electrophoresis system	Bio-Rad	Mini-Sub Cell GT system / 1704360 base	Electrophoresis system for agarose gel separation of PCR and RFLP products.
Vertical PAGE electrophoresis system	Bio-Rad	Mini-PROTEAN Tetra Cell / 1658001	Vertical electrophoresis system for 8% polyacrylamide gel separation when higher-resolution fragment discrimination is required.
Gel documentation system with UV or blue-light transillumination	Bio-Rad / equivalent laboratory system	Gel Doc XR+ or equivalent	Imaging system for documenting stained DNA bands after gel electrophoresis. Use UV shielding or blue-light mode where available.
Refrigerated centrifuge	Eppendorf / equivalent laboratory centrifuge	5424 R or equivalent	Refrigerated centrifuge used for serum separation at 1,500 × g and 4 °C.
−80 °C ultra-low temperature freezer	Thermo Scientific / equivalent laboratory freezer	Forma 900 Series or equivalent	Storage of serum aliquots before ELISA analysis.
IBM SPSS Statistics	IBM	Version 25.0	Statistical software used for descriptive statistics, normality testing, group comparisons, logistic regression, and model-output export.
Spreadsheet software	Microsoft	Excel for Microsoft 365	Used for data checking, plate-map organization, duplicate-value review, and preparation of merged analysis files before SPSS analysis.