Predictive role of systemic immune-inflammation index (SII) in tuberculosis infection morbidity and ICU tuberculosis patient mortality: an observational study

Introduction

Tuberculosis (TB), an infectious disease caused by Mycobacterium tuberculosis (MTB), is responsible for infecting approximately a quarter of the global population and causing 1 million deaths in 2020, according to a World Health Organization (WHO) TB report (1). MTB is transmitted via the inhalation of airborne droplet nuclei expelled by individuals with active tuberculosis and subsequently inhaled by susceptible healthy persons (2). Following exposure to MTB, the host’s innate immune response is immediately activated to eliminate the bacilli (3,4). However, among individuals initially infected with TB, 5–10% progress to active TB and become infectious (5). Even after recommended treatment for tuberculosis infection (TBI), TB may recur in their lifetime with the probability of approximately 20% (6). Therefore, achieving the WHO’s “End TB Strategy”, which aims to reduce TB incidence and progression, requires not only treating active TB but also focusing on the early diagnosis and prevention of TBI (7,8).

The current diagnostic tests for TBI include the tuberculin skin test (TST) and interferon-gamma release assays (IGRAs), such as the QuantiFERON-TB Gold In-Tube test and the T-SPOT.TB assay (9-11). Although both have high sensitivity and specificity for identifying TBI in different situations, they cannot be used for everyone owing to economic, technological, and instrumental limitations. Thus, some studies have focused on the basic clinical and laboratory variables that may be associated with TBI. For instance, Van Rie et al. found that a low CD4 count combined with a low body mass index (BMI) was highly associated with early-incident TB (12). In addition, other researchers have noted that extreme levels of monocytes and lymphocytes in the peripheral blood are risk factors for active TB and are associated with TB treatment outcomes (13,14).

The systemic immune-inflammation index (SII), an inflammatory biomarker first reported in 2014 by Hu et al., plays an important role in predicting the prognosis of patients with hepatocellular carcinoma (15). As an index reflecting the status of the immune response and systemic inflammation, the SII has not only been used to predict and evaluate the prognosis of solid tumor diseases, such as gastric or esophageal cancer (16,17), cervical cancer (18) and non-small cell lung cancer (19), but also inflammatory diseases, such as chronic kidney disease (20) and cardiovascular disease (21). According to previous studies, the occurrence of TBI and the development of active TB are closely correlated with the immune status of the human body (1,12). Thus, we hypothesized that SII values could play a significant role in these processes.

There are two specific aims in this observational study. We first used adult samples from the National Health and Nutrition Examination Survey (NHANES) to investigate the correlation between the SII and the morbidity of TBI. Subsequently, we analyzed clinical data from in-hospital ICU patients with TB to investigate the association between SII levels and TB patient mortality. We present this article in accordance with the TRIPOD reporting checklist (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2028/rc).

Methods

Data sourcing and collection

A TBI survey was conducted using data from the NHANES from 2011 to 2012, and the dataset was downloaded from its website (https://www.cdc.gov/nchs/index.htm). For TB skin testing, Tubersol, a tuberculin-purified protein derivative (PPD) product with five international units, was injected into the arm or shoulder area of the participants to identify individuals with TB. In this study, participants with an indentation count >10 mm at 72 h were identified as individuals with TB infection according to the criteria used by the NHANES and other studies (6,11,22). Active TB participants were excluded if they answered “yes” to the question, “Have you ever been told had active TB?” in the TB questionnaire. A total of 6,128 participants underwent PPD testing with recorded induration results. Participants were excluded if they met any of the following criteria: (I) age <18 years; (II) pregnancy; and (III) a diagnosis of human immunodeficiency virus (HIV), tumor, or arthritis. Ultimately, 2,947 participants were enrolled in our study, of which 304 had TBI and 2,643 did not.

Variables

From the demographic and questionnaire data of the NHANES, 12 basic variables were selected for this study. These included age, race, education level, marital status, ratio of family income to poverty, hypertension, diabetes, cigarette and alcohol status, BMI (kg/m²), and living in a household with a TB-infected person. The ratio of family income to poverty was used as an index to represent economic status, while living in a household with a TB-sick person reflected close TB exposure. These two variables have been confirmed as clinical risk factors for TB (23,24).

In the laboratory dataset from the NHANES, 13 assay indices were selected for further research: white blood cell count, lymphocyte count, segmented neutrophil count, monocyte count, platelet count, hemoglobin, alanine aminotransferase, aspartate aminotransferase, albumin, glucose (serum), blood urea nitrogen, creatinine, and total calcium. Aside from the above indexes, two deriving indices were included: the SII, calculated as neutrophils count × platelet count / lymphocyte count, 1,000 cells/µL), and the prognostic nutritional index, calculated as 10× the serum level of albumin + 0.005 × lymphocyte count).

Based on previous studies (25), participants were divided into four groups based on the quartile distribution of the SII: low SII (<306.72, Q1), middle-low SII (306.72–430.54, Q2), middle SII (430.55–624.02, Q3), and high SII (>624.02). The frequency of TB-infected individuals in each SII group was visualized using a bar plot with the “ggplot2” package.

Sample size rationale

For the NHANES development cohort, we included all eligible participants (N=2,947 after exclusions). For the ICU prognostic cohort, 97 consecutive tuberculosis patients were included, among whom 36 deaths occurred within 28 days. Although no formal a priori sample-size calculation was performed, we followed the recommended events-per-variable (EPV) considerations: with 36 outcome events, we restricted the number of candidate predictors to ensure at least 10 events per variable, and applied Least Absolute Shrinkage and Selection Operator (LASSO) regression for variable selection to further minimize overfitting. A post-hoc precision assessment indicated that, for an AUC of 0.715, the 95% confidence interval (CI) width would be approximately ±0.09, which is acceptable for exploratory prognostic modeling.

Data sourcing

In the ICU cohort, patients were enrolled in Anhui Chest Hospital from July 01, 2021, to December 20, 2023, active tuberculosis was defined as microbiologically confirmed disease. The inclusion criteria were as follows: (I) patients admitted to the ICU; (II) adults aged >18 years; and (III) patients with evidence of MTB infection confirmed via culture or Xpert from sputum. Patients with tumors, pregnant women, and those with severe malnutrition were excluded from the study. The patients’ 28-day mortality was considered as the observational endpoint, and telephone follow-up was performed if the patients were discharged before 28 days. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Ethics Committee of the Anhui Chest Hospital (No. KJ2022-098). Informed consent was obtained from all participants prior to their inclusion in the study.

Statistical analysis

R software (version 4.3.2) was used for data analysis and visualization. Numerical variables, such as age and most laboratory indices, were presented as mean ± standard deviation (SD) and compared using Z-tests between TBI and non-TBI groups. Qualitative variables were reported as frequencies and compared using χ² tests. While the NHANES dataset had complete laboratory data, missing clinical data were handled using the “mice” package for multiple imputations (26,27). Validation datasets contained no missing values. Statistical significance was set at P<0.05.

LASSO regression was applied using the “glmnet” package to address high-dimensional variables and multicollinearity. Fixed random seeds ensured reproducibility. A ten-fold cross-validation determined the λ value that minimized mean square error within one standard error (28). Variables selected by LASSO regression were analyzed via univariate logistic regression, followed by multivariable logistic regression using the “autoReg” package.

To evaluate the sensitivity and specificity of the prediction model for TBI, receiver operating characteristic (ROC) curves were generated with the “pROC” package for both training and validation datasets (split using the “caTools” package). A nomogram was constructed from the prediction model, and decision curve analysis (DCA) assessed the clinical benefit of key variables. For ICU patients with TB, SII levels were compared between survivors and non-survivors. The predictive value of SII was assessed using ROC curves, and Cox proportional-hazards models were used to analyze the association between SII levels and survival, visualized using the “Survival” package.

Results

Participant flow

For the development cohort (NHANES 2011–2012), a total of 6,128 adults were screened. After excluding participants aged <18 years, pregnant women, and those with HIV infection, autoimmune diseases, or missing key laboratory data, 2,947 participants were included for model development. For external validation, 97 consecutive ICU patients with microbiologically confirmed tuberculosis admitted to Anhui Chest Hospital between July 2021 and December 2023 were analyzed. Among them, 36 deaths occurred within 28 days of ICU admission, and 61 patients survived to day 28. The median follow-up time was 28 days for all patients. A detailed flow of participants, including inclusion and exclusion counts, is presented in Figure 1.

Figure 1 Flow of participants through the study. HIV, human immunodeficiency virus; ICU, intensive care unit; NHANES, National Health and Nutrition Examination Survey; TB, tuberculosis; TBI, tuberculosis infection.

Baseline features of participants from the NHANES

A total of 2,974 participants were enrolled in this study. The average age of individuals with TB infection (47.2±15.7 years) was higher than those without TB infection (40.3±16.5 years). Among non-TB-infected participants, 53.6% were men, compared to 61.5% in the TB-infected group. Regarding racial demographics, non-Hispanic White individuals were less susceptible to TB infection compared to Hispanic participants. Additionally, individuals with lower education levels were more likely to have TB infection. Married individuals and those living with a TB-infected person had a higher risk of TBI, indicating a possible familial association with the disease (29). Other factors, such as diabetes, hypertension, smoking or alcohol use, BMI, and income-to-poverty ratio, showed no significant differences between the two groups. In terms of laboratory variables, individuals with TB infection exhibited higher lymphocyte counts, serum glucose levels, and alanine aminotransferase levels, but lower neutrophil counts compared to non-infected individuals (Table 1).

SII values in the TBI group and non-TBI group

In this study, the mean SII value in the non-TBI group was 508±338.8 (1,000 cells/µL), while in the TBI group, it was 449±261.4 (1,000 cells/µL). As shown in Figure 2A, the difference in SII levels between the two groups was statistically significant (P<0.05). Using quartile-based SII subgroups, no significant differences were observed between the low-middle and middle SII groups across TBI and non-TBI participants. However, there were a higher proportion of individuals with low SII levels in the TB-infected group (Table 1, Figure 2B). This suggests that a higher SII level may serve as a protective factor against TBI.

Figure 2 Difference of SII between TBIs and non-TBIs. (A) The counts of SII in different groups was shown by boxplot. (B) Stacking diagram showed the percentage of SII subgroups between the two groups. SII, systemic immune-inflammation index; TBI, tuberculosis infection.

SII levels were closely associated with TBI

LASSO regression analysis was used to identify variables distinguishing TBI from non-TBI groups, with a ten-fold cross-validation approach (Figure 3A,3B). Five variables were identified: age, race, education level, living with a TB-infected household member, and SII, the latter being the only laboratory variable included. Univariable and multivariable logistic regression analyses confirmed these variables were significantly associated with TBI. Non-Hispanic Black participants were less likely to have TBI, as shown by both univariable [odds ratio (OR) 0.11, P<0.001] and multivariable (OR 0.12, P<0.001) analyses (Table 2).

Figure 3 LASSO regression for variables selection and the construction of nomogram. (A,B) 11 different variables were included in LASSO regression and ten-fold cross validation was used for parameters selection. (C) The final model used by variables obtained in LASSO regression were constructed and visualized by nomogram. LASSO, Least Absolute Shrinkage and Selection Operator; SII, systemic immune-inflammation index; TB, tuberculosis.

Multivariable regression analysis further revealed that SII, as an independent laboratory factor, was closely related to TBI in the unadjusted model (OR 0.84, 95% CI: 0.75–0.94). The OR values for the SII values in adjusted models 1, 2, and 3 were 0.85, 0.84, and 0.85, respectively, highlighting the index’s significance in the occurrence of TBI. However, after adjusting for all variables, SII was no longer statistically significant in TBI prediction (Table 3).

Using the five identified variables, a comprehensive predictive model for TBI was constructed. A nomogram demonstrated that the model had a 60% probability of predicting TBI, with SII values playing key roles (Figure 3C). Validation through ROC curve analysis yielded area under the curve (AUC) values of 0.746 and 0.734 for the training and validation sets, respectively, indicating good model performance (Figure 4A,4B). DCA suggested that increasing SII levels or reducing TB exposure could improve clinical outcomes and contribute to reducing TBI cases (Figure 4C).

Figure 4 The efficiency of the comprehensive model in the prediction of TBI. (A,B) The ROC curves showed the predictive value of final model constructed in both training set and validation set. (C) DCA curve was depicted to evaluate the clinical usefulness by timely interfering the index in the model. TB exposure represented the “Lived in TB household sick person”. AUC, area under the curve; CI, confidence interval; DCA, decision curve analysis; ROC, receiver operating characteristic; SII, systemic immune-inflammation index; TB, tuberculosis; TBI, tuberculosis infection.

SII values were associated with the mortality of ICU tuberculosis patients

To investigate the relationship between SII levels and mortality in TB, clinical data from ICU patients were analyzed (Table S1). The mean SII level was higher in patients who died compared to survivors (Figure 5A). ROC curve analysis showed an AUC of 0.715 for predicting mortality among ICU patients with TB (Figure 5B). The optimal SII cutoff value was 1,460.25 (1,000 cells/µL), which stratified patients into two groups. Cox proportional-hazards analysis demonstrated that higher SII levels were significantly associated with poor prognosis in ICU patients with TB (Figure 5C). These findings suggest a strong association between elevated SII levels and increased mortality in ICU patients with TB, highlighting the prognostic importance of SII in severe TB cases.

Figure 5 The different level of SII index in ICU TB patients. (A) Boxplot showed the variance of SII value in ICU alive TB patients compared to dead patients. (B) The ROC curve showed the prediction of SII in the mortality of ICU TB patients. (C) The survival curve showed the high level of SII had a poor prognosis of ICU TB patients. ICU, intensive care unit; ROC, receiver operating characteristic; SII, systemic immune-inflammation index; TB, tuberculosis.

Discussion

In this observational study using two datasets, we investigated the predictive role of the SII in TBI and its prognostic significance in the mortality of ICU patients with TB. Our results revealed that the SII level was lower in the TBI group compared to the non-TBI group and was an independent risk factor for TBI. Additionally, we constructed a predictive model combining SII with non-laboratory variables such as age, race, education level, and TB exposure, which demonstrated good performance in identifying TBI. Notably, a high SII level, rather than a low level, was associated with increased mortality in ICU patients with TB.

TBI represents the initial stage of active TB disease, during which the host immune response to MTB infection plays a critical role. This involves complex biochemical processes and signaling pathways, including the activation of NF-κB signaling, type I interferon signaling, and autophagy (30-32). While laboratory variables such as lymphocyte and neutrophil counts can reflect the immune and inflammatory status of the body and aid in clinical decision-making for active TB treatment, they are not reliable for distinguishing individuals with TBI, as demonstrated by our study and previous research (12,33).

The SII, a composite index derived from neutrophil, lymphocyte, and platelet counts, reflects the balance between immune response and inflammation. It has been considered a promising prognostic indicator in various diseases (34,35). Neutrophils constitute one of the earliest responders following MTB exposure, and they restrict bacterial replication through phagocytosis, production of reactive oxygen species, and release of antimicrobial peptides and neutrophil extracellular traps (36). Research indicates have demonstrated that lower circulating neutrophil counts are associated with impaired early innate immunity and increased susceptibility to TB acquisition. Lymphocytes, especially CD4⁺ T cells, are essential for granuloma formation, macrophage activation, and long-term containment of MTB (37-39). A decrease in lymphocyte numbers or function has been linked to increased TB susceptibility The SII incorporates lymphocyte count as the denominator; thus, low lymphocyte levels augment SII during systemic inflammation but reduce SII when adaptive immunity is suppressed, as may occur in early TBI (40). Platelets are increasingly recognized as active immune mediators. Beyond hemostasis, they amplify inflammatory responses through cytokine release and interactions with monocytes and neutrophils. Platelet activation contributes to immune cell recruitment to MTB-infected tissues, and altered platelet indices have been associated with TB disease severity (41). These three individual indicators are each closely associated with TBI and disease progression. Our study further reveals that their combination—SII is also linked to tuberculosis and may serve as a more precise predictive biomarker for the immune status of the disease.

In this study, we found that a low SII level was strongly associated with TBI. This was attributed to decreased neutrophil counts in TBI (Table 1), reflecting impaired host immunity—a critical factor in TBI pathogenesis. Besides, neutrophils, as innate immune cells, play a vital role in early MTB infection by secreting bactericidal enzymes and antimicrobial peptides (42). Supporting this, Martineau et al. reported a higher risk of TB in individuals with lower peripheral blood neutrophil counts (43). Furthermore, recent studies highlight the synergy between neutrophils, dendritic cells, and T cells during early TB infection, forming a critical defense alliance. This alliance facilitates effective T-cell responses against MTB, and a decline in neutrophils significantly weakens this defense mechanism. Additionally, we observed that a high SII level was associated with increased mortality in ICU patients with TB, consistent with findings from other studies on inflammation markers in ICU settings. For example, Nalbant et al. demonstrated that high SII scores predict ICU admission in coronavirus disease 2019 (COVID-19) patients (44). Similarly, Liang et al. reported that elevated SII levels are linked to higher mortality in patients with severe bacterial infections in the ICU (45). These findings suggest that high SII scores indicate an uncontrolled immune response during inflammatory diseases, although the underlying biological mechanisms require further investigation.

In summary, a low SII may represent insufficient innate and adaptive immune activation, predisposing individuals to TBI. The second part, the patients in ICU with active TB, markedly elevated SII likely reflects overwhelming systemic inflammation, neutrophil-driven tissue injury, and dysregulated host response, all of which contribute to poor prognosis. This dual pattern—low SII in infection susceptibility but high SII in severe disease—is biologically plausible and consistent with TB’s biphasic immunopathology.

There are several limitations in this study. First, although IGRA is now increasingly used to complement TST in the diagnosis of TBI, the NHANES 2011–2012 dataset contains missing IGRA data and there are very few participants with paired TST-IGRA results. Using IGRA combined definition would therefore introduce substantial selection bias and markedly reduce statistical power. For these reasons, TBI was defined using TST in this study. Nevertheless, we acknowledge that misclassification remains possible, and future studies incorporating complete TST and IGRA data would enhance diagnostic accuracy. Second, the predictive model for TBI was based on NHANES data and may only be applicable to U.S. adults. Larger and more diverse population studies over longer periods are needed to validate the findings. Third, while adjustments were made for relevant confounders, the possibility of residual confounding cannot be ruled out. Finally, as this was a cross-sectional study, causality or temporality could not be established.

Conclusions

Our study demonstrated that SII has strong predictive value for TBI morbidity based on NHANES data. Additionally, analysis of ICU patients with TB indicated a close relationship between high SII levels and increased mortality. These findings suggest that SII could play a crucial role in understanding the occurrence and progression of TB.

Acknowledgments

We thank the anonymous reviewers for constructive comments.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2028/rc

Data Sharing Statement: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2028/dss

Peer Review File: Available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2028/prf

Funding: This study was funded by Anhui Chest Hospital Project (Nos. AHXK2025Z1004, AHXK20251005).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://jtd.amegroups.com/article/view/10.21037/jtd-2025-aw-2028/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. This study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. This study was approved by the Academic Ethics Committee of Anhui Chest Hospital (No. KJ2022-098). All enrolled patients in this study were informed and signed the consent before sample collection.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Chakaya J, Khan M, Ntoumi F, et al. Global Tuberculosis Report 2020 - Reflections on the Global TB burden, treatment and prevention efforts. Int J Infect Dis 2021;113:S7-S12. [Crossref] [PubMed]
2. Migliori GB, Wu SJ, Matteelli A, et al. Clinical standards for the diagnosis, treatment and prevention of TB infection. Int J Tuberc Lung Dis 2022;26:190-205. [Crossref] [PubMed]
3. Houben RM, Dodd PJ. The Global Burden of Latent Tuberculosis Infection: A Re-estimation Using Mathematical Modelling. PLoS Med 2016;13:e1002152. [Crossref] [PubMed]
4. Vu A, Glassman I, Campbell G, et al. Host Cell Death and Modulation of Immune Response against Mycobacterium tuberculosis Infection. Int J Mol Sci 2024;25:6255. [Crossref] [PubMed]
5. Huang J, Xia J, Hong B, et al. Risk factors and prevalence of latent tuberculosis infection in rheumatic patients: a meta-analysis. BMC Infect Dis 2025;25:1094. [Crossref] [PubMed]
6. Goletti D, Delogu G, Matteelli A, et al. The role of IGRA in the diagnosis of tuberculosis infection, differentiating from active tuberculosis, and decision making for initiating treatment or preventive therapy of tuberculosis infection. Int J Infect Dis 2022;124:S12-9. [Crossref] [PubMed]
7. Ganmaa D, Uyanga B, Zhou X, et al. Vitamin D Supplements for Prevention of Tuberculosis Infection and Disease. N Engl J Med 2020;383:359-68. [Crossref] [PubMed]
8. Vasiliu A, Martinez L, Gupta RK, et al. Tuberculosis prevention: current strategies and future directions. Clin Microbiol Infect 2024;30:1123-30. [Crossref] [PubMed]
9. Barcellini L, Borroni E, Brown J, et al. First evaluation of QuantiFERON-TB Gold Plus performance in contact screening. Eur Respir J 2016;48:1411-9. [Crossref] [PubMed]
10. Wang SH, Rajaram MVS, Trollip A, et al. Novel Automation of an Enzyme-Linked Immunosorbent Spot Assay Testing Method: Comparable Diagnostic Performance of the T-SPOT.TB Test Using Manual Density Gradient Cell Isolation versus Automated Positive Selection with the T-Cell Select Kit. J Clin Microbiol 2022;60:e0055122. [Crossref] [PubMed]
11. Gualano G, Mencarini P, Lauria FN, et al. Tuberculin skin test - Outdated or still useful for Latent TB infection screening? Int J Infect Dis 2019;80S:S20-2. [Crossref] [PubMed]
12. Van Rie A, Westreich D, Sanne I. Tuberculosis in patients receiving antiretroviral treatment: incidence, risk factors, and prevention strategies. J Acquir Immune Defic Syndr 2011;56:349-55. [Crossref] [PubMed]
13. Yang Y, Chi Y, Yuan S, et al. The relationship between ventilatory ratio (VR) and 28-day hospital mortality by restricted cubic splines (RCS) in 14,328 mechanically ventilated ICU patients. BMC Pulm Med 2022;22:229. [Crossref] [PubMed]
14. Wang J, Yin Y, Wang X, et al. Ratio of monocytes to lymphocytes in peripheral blood in patients diagnosed with active tuberculosis. Braz J Infect Dis 2015;19:125-31. [Crossref] [PubMed]
15. Hu B, Yang XR, Xu Y, et al. Systemic immune-inflammation index predicts prognosis of patients after curative resection for hepatocellular carcinoma. Clin Cancer Res 2014;20:6212-22. [Crossref] [PubMed]
16. Ding P, Guo H, Sun C, et al. Combined systemic immune-inflammatory index (SII) and prognostic nutritional index (PNI) predicts chemotherapy response and prognosis in locally advanced gastric cancer patients receiving neoadjuvant chemotherapy with PD-1 antibody sintilimab and XELOX: a prospective study. BMC Gastroenterol 2022;22:121. [Crossref] [PubMed]
17. Ishibashi Y, Tsujimoto H, Hiraki S, et al. Prognostic Value of Preoperative Systemic Immunoinflammatory Measures in Patients with Esophageal Cancer. Ann Surg Oncol 2018;25:3288-99. [Crossref] [PubMed]
18. Holub K, Biete A. Impact of systemic inflammation biomarkers on the survival outcomes of cervical cancer patients. Clin Transl Oncol 2019;21:836-44. [Crossref] [PubMed]
19. Xu F, Zhu H, Xiong D, et al. Tertiary lymphoid structures combined with biomarkers of inflammation are associated with the efficacy of neoadjuvant immunochemotherapy in resectable non-small cell lung cancer: A retrospective study. Thorac Cancer 2024;15:172-81. [Crossref] [PubMed]
20. Lai W, Xie Y, Zhao X, et al. Elevated systemic immune inflammation level increases the risk of total and cause-specific mortality among patients with chronic kidney disease: a large multi-center longitudinal study. Inflamm Res 2023;72:149-58. [Crossref] [PubMed]
21. Orhan AL, Şaylık F, Çiçek V, et al. Evaluating the systemic immune-inflammation index for in-hospital and long-term mortality in elderly non-ST-elevation myocardial infarction patients. Aging Clin Exp Res 2022;34:1687-95. [Crossref] [PubMed]
22. Lu P, Liu Q, Zhou Y, et al. Predictors of Discordant Tuberculin Skin Test and QuantiFERON-TB Gold In-tube Results in Eastern China: A Population-based, Cohort Study. Clin Infect Dis 2021;72:2006-15. [Crossref] [PubMed]
23. Underwood JM, Pampati S, Everett Jones S, et al. School-Level Poverty and Rurality Associated With Differences in Sexual Risk Behaviors Among U.S. Public High School Students. J Adolesc Health 2021;69:964-9. [Crossref] [PubMed]
24. Malik AA, Amanullah F, Siddiqui S, et al. Quantified TB Exposure Can Predict Prevalent TB Disease in Child Contacts. Pediatr Infect Dis J 2021;40:e208-9. [Crossref] [PubMed]
25. Xu M, Chen R, Liu L, et al. Systemic immune-inflammation index and incident cardiovascular diseases among middle-aged and elderly Chinese adults: The Dongfeng-Tongji cohort study. Atherosclerosis 2021;323:20-9. [Crossref] [PubMed]
26. Austin PC, White IR, Lee DS, et al. Missing Data in Clinical Research: A Tutorial on Multiple Imputation. Can J Cardiol 2021;37:1322-31. [Crossref] [PubMed]
27. Morris TP, White IR, Royston P. Tuning multiple imputation by predictive mean matching and local residual draws. BMC Med Res Methodol 2014;14:75. [Crossref] [PubMed]
28. Krstajic D, Buturovic LJ, Leahy DE, et al. Cross-validation pitfalls when selecting and assessing regression and classification models. J Cheminform 2014;6:10. [Crossref] [PubMed]
29. Karumbi J, Garner P. Directly observed therapy for treating tuberculosis. Cochrane Database Syst Rev 2015;2015:CD003343. [Crossref] [PubMed]
30. Poladian N, Orujyan D, Narinyan W, et al. Role of NF-κB during Mycobacterium tuberculosis Infection. Int J Mol Sci 2023;24:1772. [Crossref] [PubMed]
31. Zhou J, Lv J, Carlson C, et al. Trained immunity contributes to the prevention of Mycobacterium tuberculosis infection, a novel role of autophagy. Emerg Microbes Infect 2021;10:578-88. [Crossref] [PubMed]
32. Olson GS, Murray TA, Jahn AN, et al. Type I interferon decreases macrophage energy metabolism during mycobacterial infection. Cell Rep 2021;35:109195. [Crossref] [PubMed]
33. Suryana K, Dharmesti NWW, Rai IBN. High Pretreatment Level of Neutrophil to Lymphocyte Ratio, Monocyte to Lymphocyte Ratio and Other Factors Associated with Delayed Sputum Conversion in Patients with Pulmonary Tuberculosis. Infect Drug Resist 2022;15:5455-62. [Crossref] [PubMed]
34. Yang YL, Wu CH, Hsu PF, et al. Systemic immune-inflammation index (SII) predicted clinical outcome in patients with coronary artery disease. Eur J Clin Invest 2020;50:e13230. [Crossref] [PubMed]
35. Mahemuti N, Jing X, Zhang N, et al. Association between Systemic Immunity-Inflammation Index and Hyperlipidemia: A Population-Based Study from the NHANES (2015-2020). Nutrients 2023;15:1177. [Crossref] [PubMed]
36. Chai Q, Lu Z, Liu CH. Host defense mechanisms against Mycobacterium tuberculosis. Cell Mol Life Sci 2020;77:1859-78. [Crossref] [PubMed]
37. Blomgran R, Ernst JD. Lung neutrophils facilitate activation of naive antigen-specific CD4+ T cells during Mycobacterium tuberculosis infection. J Immunol 2011;186:7110-9. [Crossref] [PubMed]
38. Morel C, Badell E, Abadie V, et al. Mycobacterium bovis BCG-infected neutrophils and dendritic cells cooperate to induce specific T cell responses in humans and mice. Eur J Immunol 2008;38:437-47. [Crossref] [PubMed]
39. Simonson AW, Zeppa JJ, Bucsan AN, et al. Intravenous BCG-mediated protection against tuberculosis requires CD4+ T cells and CD8α+ lymphocytes. J Exp Med 2025;222:e20241571. [Crossref] [PubMed]
40. Sakai S, Mayer-Barber KD, Barber DL. Defining features of protective CD4 T cell responses to Mycobacterium tuberculosis. Curr Opin Immunol 2014;29:137-42. [Crossref] [PubMed]
41. Fox KA, Kirwan DE, Whittington AM, et al. Platelets Regulate Pulmonary Inflammation and Tissue Destruction in Tuberculosis. Am J Respir Crit Care Med 2018;198:245-55. [Crossref] [PubMed]
42. Hilda JN, Das S, Tripathy SP, et al. Role of neutrophils in tuberculosis: A bird's eye view. Innate Immun 2020;26:240-7. [Crossref] [PubMed]
43. Martineau AR, Newton SM, Wilkinson KA, et al. Neutrophil-mediated innate immune resistance to mycobacteria. J Clin Invest 2007;117:1988-94. [Crossref] [PubMed]
44. Nalbant A, Demirci T, Kaya T, et al. Can prognostic nutritional index and systemic immune-inflammatory index predict disease severity in COVID-19? Int J Clin Pract 2021;75:e14544. [Crossref] [PubMed]
45. Liang R, Chen Z, Yang S, et al. A diagnostic model based on routine blood examination for serious bacterial infections in neonates-a cross-sectional study. Epidemiol Infect 2023;151:e137. [Crossref] [PubMed]