Abstract
This study aims to explore the association between frailty scores (FS) and latent tuberculosis infection (LTBI) and analyze the impact of FS on all-cause mortality in the LTBI population, providing a reference for optimizing screening strategies in high-risk groups. A total of 3520 adults aged 18 to 79 years from the National Health and Nutrition Examination Survey 2011 to 2012 cycle were included, comprising 564 individuals in the LTBI group and 2956 in the non-LTBI group. LTBI was diagnosed through a combination of the tuberculin skin test and interferon-gamma release assays. The FS was constructed based on 53 health deficit indicators covering cognitive function, comorbidities, nutritional metrics, and laboratory parameters. The natural logarithm of FS was generated as natural logarithm of FS (LnFS). A multivariable logistic regression model was used to analyze the association between LnFS and LTBI, while Cox proportional hazards models assessed its impact on all-cause mortality. A mediation analysis was performed with LnFS as the exposure, LTBI as the outcome, and serum globulin concentration as the hypothesized mediator. The FS score was notably greater in the LTBI group than in the non-LTBI group (0.15 ± 0.09 vs 0.13 ± 0.08; P = .003). The association between LnFS and LTBI was stronger (odds ratio = 1.26, 95% confidence interval: 1.05–1.52, P = .016). Globulin partially mediated the link between LTBI and FS (12.5% mediation effect, P = .04). The risk of all-cause mortality was increased by 32% in LTBI patients (hazard ratio = 1.32, 95% confidence interval: 1.05–1.66), and this risk intensified in a dose-dependent manner with increasing FS scores (interaction P = .02). FS are significantly positively correlated with LTBI, and this relationship accelerates the frailty process through chronic inflammatory responses mediated by globulin. FS independently increases the risk of all-cause mortality. The findings provide scientific evidence for early frailty screening and targeted interventions in high-risk LTBI populations, while emphasizing the importance of serum globulin levels.
Keywords: frailty score, latent tuberculosis infection, mediation analysis, mortality
1. Introduction
Tuberculosis (TB) persists as a leading global infectious killer caused by a singular pathogen, continuing to draw significant attention from the academic community due to its epidemiological characteristics and public health challenges. According to the latest data from the World Health Organization, there were 10.8 million new cases of TB globally in 2023, surpassing the mortality rates associated with HIV/AIDS and COVID-19 during the same period.[1] In this dire situation, latent tuberculosis infection (LTBI) serves as a “silent reservoir” within the transmission chain of TB.[2] Data from the National Health and Nutrition Examination Survey (NHANES) 2011 to 2012 indicate that approximately 14% of adults aged 20 years and older are affected by LTBI.[3] Patients with LTBI typically exhibit no significant clinical symptoms and lack radiographic or microbiological evidence of active TB. However, this condition represents an infection state in which the host’s immune system responds to Mycobacterium tuberculosis (MtB) antigens through a cell-mediated immune response.[4] The prolonged presence of LTBI may profoundly affect host health through chronic immune activation and inflammatory responses. Traditional diagnostic methods, such as the tuberculin skin test (TST), are susceptible to interference from Bacillus Calmette-Guérin vaccination and exhibit insufficient specificity. Although molecular diagnostic techniques demonstrate sensitivity comparable to culture methods, their cost and equipment requirements hinder widespread implementation in low-income regions.
As a multidimensional syndrome, frailty involves the deterioration of physiological reserves, lower tolerance to stress, and an elevated risk of unfavorable health consequences.[5] The theoretical foundation of the frailty index (FI) is derived from the health deficit accumulation model proposed by Rockwood et al It measures a person’s health by dividing the number of deficits they have by the total number of deficits evaluated.[5,6] In this study, the frailty score (FS) was constructed using 53 items encompassing cognitive function, activities of daily living, depressive symptoms, comorbidities, healthcare utilization, physical metrics, and key laboratory parameters. This multidimensional approach captures both clinical and subclinical physiological dysregulation, making it particularly relevant for investigating immune-mediated conditions such as LTBI. Previous studies have confirmed that the FI is associated with all-cause mortality,[7,8] cardiovascular diseases, and metabolic disorders.[9] However, the interplay between the FI and LTBI remains unclear.[10] Given that LTBI is influenced by immune status, and frailty, as a comprehensive indicator reflecting overall health vulnerability, it may have a potential relationship with LTBI. The objective of this study is to analyze the association between the FS and LTBI, and to further investigate the impact of the FS on all-cause mortality among individuals with LTBI.
2. Materials and methods
2.1. Study population
We began with the NHANES 2011 to 2012 interview sample (n = 9756).[11] NHANES serves as a cross sectional study and represents a comprehensive survey of the U.S. population, providing extensive information on nutrition and health among the general American populace. The survey data from NHANES is publicly accessible to researchers and users, facilitating a wide range of data analyses. Data collection is conducted by the National Center for Health Statistics in a biennial cycle (https://wwwn.cdc.gov/nchs/nhanes/continuousnhanes). This study was conducted in accordance with the principles outlined in the Declaration of Helsinki. The use of the MIMIC-IV database was approved by the Institutional Review Boards of Beth Israel Deaconess Medical Center and the Massachusetts Institute of Technology. As the MIMIC-IV database provides data freely, informed consent and ethical approval statements were not required for this study.
The inclusion criteria for this study were: individuals aged 18 to 79 years; participants with complete LTBI test results and no missing data on key health behaviors or underlying health conditions. The exclusion criteria were: individuals with missing LTBI-related information; those with missing data on hypertension, diabetes, smoking, alcohol consumption, or globulin levels; individuals aged 80 years or older; and individuals who were HIV positive. Following a systematic selection process, a total of 3520 participants were included in the study, comprising 564 individuals in the LTBI group and 2956 in the non-LTBI group (Fig. 1). The study received ethical clearance from the National Center for Health Statistics’ Institutional Review Board, with all subjects giving written consent, eliminating the necessity for additional ethics evaluations.
Figure 1.
Flowchart of the inclusion pathway of the study population. HIV = human immunodeficiency virus, LTBI = latent tuberculosis infection, NHANES = National Health and Nutrition Examination Survey.
The diagnosis of LTBI was based on positive results from 2 immunological testing methods:[3,12] the TST, defined as a skin induration diameter of ≥10 mm, and the interferon-gamma release assay, specifically using the QuantiFERON-TB Gold In-Tube test, where a positive result was defined as a Nil value ≤8.0 international unit (IU)/mL and a TB antigen value minus Nil value ≥0.35 IU/mL. A participant was classified as having LTBI if either test result was positive. According to the official NHANES report, the QuantiFERON®-TB Gold IT system utilizes specialized blood collection tubes designed for venipuncture to collect whole blood. These tubes include a Nil control tube, a TB antigen tube, and a Mitogen tube (positive control). The tubes are gently shaken to mix the antigens with the whole blood and are incubated at 37°C ± 1°C for 16 to 24 hours. After the incubation period, plasma is harvested, and the amount of interferon-gamma (IFN-γ) produced in response to the peptide antigens is measured using an enzyme-linked immunosorbent assay. Test sample results are reported in IU based on a standard curve generated from testing dilutions of a recombinant human IFN-γ standard.
2.2. Calculation of the frailty index
This research employed a 53-item FI derived from a deficit accumulation model to evaluate frailty in depressed patients.[13,14] Initially, we identified 53 health deficit indicators across 7 dimensions: cognitive function (1 item: memory issues), activities of daily living (20 items assessing functional dependence), depressive symptoms (7 items from the PHQ-9 scale), comorbid conditions (13 items including arthritis, coronary heart disease, diabetes, etc), healthcare utilization (5 items such as hospitalization history and medication use), physical function (1 item: BMI), and laboratory indicators (6 items including HbA1c and hemoglobin). Each indicator was assigned a value of 0 (no deficit) or 1 (complete deficit) based on its presence or severity. The FS was calculated by determining the ratio of actual deficits present to the total number of deficits, yielding a score ranging from 0 to 1 (Table S1, Supplemental Digital Content, https://links.lww.com/MD/R308).
2.3. Research variables
The study variables were systematically extracted from multiple components of the NHANES database, encompassing demographic characteristics, anthropometric measurements, and behavioral factors. Demographic variables, including age (RIDAGEYR), sex (RIAGENDR), race/ethnicity (RIDRETH1), and poverty-income ratio (INDFMPIR), were obtained from the “Demographic Variables and Sample Weights” section. Anthropometric assessment was conducted using body mass index (BMI). Behavioral factors comprised 2 key components: alcohol consumption patterns (ALQ101 for past 12-month consumption, ALQ110 for frequency, ALQ120U for quantity units, and ALQ120Q for quantity amount) extracted from the ALQ questionnaire, and smoking status (SMQ020 for lifetime smoking ≥100 cigarettes and SMQ040 for current smoking status) derived from the SMQ questionnaire. Referring to previous literature, the results of alcohol consumption were divided into 5 levels: never (had <12 drinks in lifetime), former (had ≥12 drinks in 1 year and did not drink last year, or did not drink last year but drank ≥12 drinks in lifetime), mild, moderate, and heavy.[15] Albumin (LBDSALSI), globulin (LBXSGB), lymphocyte count (LBDLYMNO), monocyte count (LBDMONO), and neutrophil count (LBDNENO) were extracted from Laboratory Data. The 2-color endpoint method is used to measure albumin concentration. During the reaction, albumin binds to the bromocresol purple reagent, forming a complex. The system monitors changes in absorbance at 600 nanometers. The change in absorbance is proportional to the concentration of albumin in the sample. The timed rate biuret method is employed to measure the total protein concentration in serum or plasma. Proteins in the sample bind with the reagent to form a copper-protein complex. The rate of absorbance change is monitored by a detector at 545 nanometers. The rate of complex formation is observed to be proportional to the total protein concentration in the sample. Complete blood counts are performed on blood samples using the Beckman Coulter MAXM instrument.[16] Albumin, globulin, lymphocyte, monocyte, and neutrophil counts are included as covariates because their roles in reflecting systemic inflammation, immune function, and nutritional status have been fully proven.
2.4. Statistical analysis
The calculation and analysis refer to the weighting process carried out according to the official guidance of NHANES. Descriptive statistics were used to characterize baseline features, with continuous variables presented as means with standard deviations and categorical variables reported as counts and percentages. Since the FS showed a skewed distribution after the normality test, we performed a natural logarithm and used the natural logarithm of FS (LnFS) as the independent variable for subsequent regression analysis. To evaluate the relationship between LnFS and the likelihood of LTBI, we employed a stepwise logistic regression approach across 3 progressively refined models. The initial analysis (Model 1) examined crude associations without any adjustments. We then introduced controls for age and sex in Model 2 to account for these fundamental demographic factors. Finally, Model 3 incorporated comprehensive adjustments, including both lymphocyte levels and BMI, to provide the most robust assessment of this potential connection. The inclusion of covariates in the multivariate regression analysis took into account multicollinearity, and only indicators with a variance inflation factor <10 were included. The results were expressed as odds ratio (OR) with 95% confidence intervals (CIs). To assess the association between LnFS and all-cause mortality, Kaplan–Meier survival curves were constructed to estimate cumulative survival probabilities across FS quartiles (Q1–Q4), with intergroup differences evaluated by the log-rank test. Subsequently, Cox proportional hazards models were applied with the same covariate adjustments as previously used. The findings are presented as hazard ratios (HR) alongside their corresponding 95% CIs. The proportional hazards assumption was evaluated using Schoenfeld residuals; no significant violations were observed. Mediation analysis was performed to explore the potential mediating role of globulin in the relationship between LnFS and LTBI risk. Restricted cubic spline (RCS) analysis was utilized to examine the nonlinear association between LnFS and both LTBI risk and all-cause mortality, providing visual representation of the dose-response relationship. Subgroup analyses were conducted based on age, BMI, and sex to assess the heterogeneity of associations across different population strata, with interaction terms tested for significance. Given the observational and cross sectional design of this study, we adopted a complete-case analysis approach. Specifically, any participant with missing data on LTBI status, FI components, serum globulin levels, or key covariates (including age, sex, BMI, smoking status, alcohol consumption, hypertension, or diabetes) was excluded from the analysis. Of the 9756 individuals in the NHANES 2011 to 2012 cycle, 3520 adults aged 18 to 79 years had complete data on all variables of interest and were included in the final analytic sample. No imputation methods were applied, as the FS was constructed from 53 interdependent health deficit items, and reliable imputation would require strong assumptions about the missing data mechanism. All analyses incorporated NHANES sampling weights to preserve the national representativeness of the included subsample. All statistical tests were 2-sided, and a P-value < .05 was considered statistically significant. All analyses were carried out using R statistical software (The R Foundation, version 4.2.2, http://www.R-project.org).
3. Results
3.1. Comparison of general characteristics
A total of 3520 participants were included in this study, comprising 564 patients with LTBI and 2956 non-LTBI patients. The average age was 46.52 ± 15.82 years, with LTBI patients being significantly older than their non-LTBI counterparts (49.75 ± 15.27 years vs 46.20 ± 15.83 years, P = .001). Male participants accounted for 49.7% of the total, with a higher proportion in the LTBI group (56.8% vs 49.0%, P = .005). FS and their LnFS were significantly elevated in the LTBI group (0.15 ± 0.09 vs 0.13 ± 0.08, P = .003; −2.13 ± 0.68 vs −2.25 ± 0.67, P = .003). Laboratory findings revealed that LTBI patients had significantly higher lymphocyte counts compared to non-LTBI patients (2.13 ± 0.58 × 103/µL vs 1.99 ± 0.59 × 103/µL, P = .007), and their globulin levels were also significantly elevated (2.89 ± 0.39 g/dL vs 2.76 ± 0.44 g/dL, P < .001) (Table 1).
Table 1.
Comparison of clinical characteristics between LTBI and non-LTBI participants.
| Variables | Overall (n = 3520) | Non-LTBI group (n = 2956) | LTBI group (n = 564) | P value |
|---|---|---|---|---|
| Age (yr) | 46.52 ± 15.82 | 46.20 ± 15.83 | 49.75 ± 15.27 | .001 |
| Sex (n, %) | .005 | |||
| Female | 1713.0 (50.3%) | 1480.0 (51.0%) | 233.0 (43.2%) | |
| Male | 1807.0 (49.7%) | 1476.0 (49.0%) | 331.0 (56.8%) | |
| BMI (kg/m2) | 28.51 ± 5.63 | 28.51 ± 5.63 | 28.49 ± 5.65 | .941 |
| Race/ethnicity (n, %) | <.001 | |||
| Mexican American | 363.0 (7.8%) | 274.0 (6.7%) | 89.0 (19.3%) | |
| Other Hispanic | 369.0 (6.3%) | 258.0 (5.3%) | 111.0 (17.0%) | |
| Other | 2788.0 (85.9%) | 2424.0 (88.1%) | 364.0 (63.7%) | |
| Education (n, %) | <.001 | |||
| <9th grade | 287.0 (4.6%) | 185.0 (3.6%) | 102.0 (15.2%) | |
| 9–11th grade | 480.0 (10.6%) | 381.0 (10.0%) | 99.0 (17.3%) | |
| High school or above | 2753.0 (84.7%) | 2390.0 (86.4%) | 363.0 (67.5%) | |
| Marriage (n, %) | .200 | |||
| Married | 1694.0 (53.0%) | 1376.0 (52.9%) | 318.0 (54.5%) | |
| Widowed | 202.0 (3.8%) | 163.0 (3.6%) | 39.0 (5.5%) | |
| Divorced | 400.0 (11.8%) | 342.0 (11.8%) | 58.0 (11.7%) | |
| Other | 1224.0 (31.4%) | 1075.0 (31.7%) | 149.0 (28.3%) | |
| Hypertension (n, %) | 1443.0 (36.2%) | 1191.0 (35.6%) | 252.0 (42.5%) | .018 |
| Diabetes mellitus (n, %) | <.001 | |||
| DM | 651.0 (13.2%) | 503.0 (12.4%) | 148.0 (21.4%) | |
| IFG | 137.0 (4.2%) | 117.0 (4.3%) | 20.0 (3.5%) | |
| IGT | 149.0 (4.4%) | 117.0 (4.1%) | 32.0 (7.9%) | |
| No | 2583.0 (78.2%) | 2219.0 (79.3%) | 364.0 (67.3%) | |
| Smoke (n, %) | .12 | |||
| Former | 804.0 (24.6%) | 655.0 (24.3%) | 149.0 (28.1%) | |
| Never | 1972.0 (55.2%) | 1680.0 (55.9%) | 292.0 (47.9%) | |
| Current | 744.0 (20.2%) | 621.0 (19.8%) | 123.0 (24.0%) | |
| Alcohol (n, %) | .013 | |||
| Former | 583.0 (14.1%) | 469.0 (13.7%) | 114.0 (18.2%) | |
| Heavy | 745.0 (22.5%) | 635.0 (22.4%) | 110.0 (23.9%) | |
| Mild | 1180.0 (36.9%) | 1001.0 (37.5%) | 179.0 (30.2%) | |
| Moderate | 544.0 (17.1%) | 484.0 (17.4%) | 60.0 (13.9%) | |
| Never | 468.0 (9.4%) | 367.0 (9.0%) | 101.0 (13.7%) | |
| Albumin (g/dL) | 4.32 ± 0.32 | 4.32 ± 0.32 | 4.30 ± 0.33 | .293 |
| RDW (%) | 12.74 ± 0.76 | 12.73 ± 0.76 | 12.80 ± 0.80 | .425 |
| WBC (103/uL) | 6.86 ± 1.81 | 6.85 ± 1.82 | 6.97 ± 1.80 | .378 |
| Lymphocyte (103/uL) | 2.00 ± 0.59 | 1.99 ± 0.59a | 2.13 ± 0.58 | .007 |
| Monocyte (103/uL) | 0.50 ± 0.14 | 0.50 ± 0.14 | 0.49 ± 0.14 | .887 |
| Neutrophils (103/uL) | 4.08 ± 1.43 | 4.08 ± 1.44 | 4.03 ± 1.38 | .851 |
| Hemoglobin (g/dL) | 14.20 ± 1.43 | 14.21 ± 1.43 | 14.13 ± 1.47 | .530 |
| Globulin (g/dL) | 2.77 ± 0.44 | 2.76 ± 0.44 | 2.89 ± 0.39 | <.001 |
| Frailty score | 0.13 ± 0.09 | 0.13 ± 0.08 | 0.15 ± 0.09 | .003 |
| LnFS | −2.24 ± 0.67 | −2.25 ± 0.67 | −2.13 ± 0.68 | .003 |
BMI = body mass index, DM = diabetes mellitus, IFG = impaired fasting glucose, IGT = impaired glucose tolerance, LnFS = natural logarithm of frailty score, LTBI = latent tuberculosis infection, RDW = red cell distribution width, WBC = white blood cell.
3.2. Logistic regression analysis
The LnFS demonstrated a significant association with LTBI. In Model 1, which did not adjust for any covariates, participants with higher LnFS had significantly greater odds of having LTBI at the time of assessment (OR = 1.30, 95% CI: 1.09–1.55, P = .006). After adjusting for age and sex in Model 2, the OR decreased to 1.25 (95% CI: 1.04–1.50, P = .019). Further adjustments for lymphocyte count and BMI in Model 3 resulted in an OR of 1.26 (95% CI: 1.05–1.52, P = .016).
When stratified by quartiles of LnFS, the risk of LTBI was significantly higher in the fourth quartile (Q4, highest quartile) compared to the first quartile (Q1, lowest quartile). The OR for LTBI in Models 1, 2, and 3 were 1.79 (95% CI: 1.27–2.52, P = .003), 1.68 (95% CI: 1.19–2.37, P = .007), and 1.73 (95% CI: 1.18–2.53, P = .009), respectively. Trend tests yielded P-values < .011, indicating a positive correlation between LnFS and the risk of LTBI (Table 2).
Table 2.
Analysis of the association between LnFS and LTBI.
| Model 1 | P value | Model 2 | P value | Model 3 | P value | |
|---|---|---|---|---|---|---|
| LnFS | 1.30 (1.09–1.55) | .006 | 1.25 (1.04–1.50) | .019 | 1.26 (1.05–1.52) | .016 |
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.26 (0.84–1.90) | .234 | 1.23 (0.82–1.83) | .284 | 1.23 (0.82–1.85) | .277 |
| Q3 | 1.26 (0.82–1.91) | .266 | 1.18 (0.77–1.80) | .419 | 1.19 (0.77–1.85) | .392 |
| Q4 | 1.79 (1.27–2.52) | .003 | 1.68 (1.19–2.37) | .007 | 1.73 (1.18–2.53) | .009 |
| P for trend | .003 | .008 | .011 |
Model 1: no covariates were adjusted. Model 2: age and sex were adjusted. Model 3: age, sex, lymphocyte and BMI were adjusted.
LnFS = natural logarithm of frailty score, LTBI = latent tuberculosis infection.
3.3. RCS analysis of the impact of LnFS on LTBI
The overall association was statistically significant (P-overall < .05), while the test for nonlinearity yielded a P-value of .930, indicating no evidence of a nonlinear relationship. The OR (95% CI) for LTBI increased gradually with higher LnFS values, suggesting a positive linear association between LnFS and the odds of LTBI (Fig. 2).
Figure 2.
Restricted cubic spline (RCS) curve depicting the association between the natural logarithm of frailty score (LnFS) and the odds ratio (OR) of latent tuberculosis infection (LTBI). The solid blue line represents the estimated OR of LTBI across different LnFS values, and the light blue shaded area denotes the corresponding 95% confidence interval (CI). A reference line at OR = 1.00 is included to facilitate interpretation. The P-overall value indicates the overall significance of the association, and the P-nonlinear value assesses the evidence for a nonlinear relationship. In this analysis, the overall association was statistically significant (P-overall <.05), whereas the test for nonlinearity showed no significant deviation from linearity (P-non-linear = .930). Thus, the results suggest a positive linear relationship between LnFS and the odds of LTBI.
3.4. Mediation effect analysis
Globulin exerted a contributory intermediary influence in the correlation between LnFS and the likelihood of LTBI. The aggregate impact was calculated to be 0.028 (95% confidence interval: 0.007 to 0.05, P = .02). The average causal mediation effect (ACME) was 0.004 (95% CI: 0.001–0.01, P = .02), while the average direct effect (ADE) was 0.025 (95% CI: 0.003–0.04, P = .04). The proportion of the mediation effect accounted for 12.5% (95% CI: 3.0%-48%, P = .04). (Table 3). These results suggest that LnFS may partially predict the risk of LTBI through its influence on globulin metabolism (Fig. 3).
Table 3.
Glubulin as a mediator variable between LnFS and the risk of LTBI.
| Mediation effect | Estimate | 95% CI lower | 95% CI upper | P value |
|---|---|---|---|---|
| Total effect | 0.028 | 0.007 | 0.05 | .02 |
| ACME | 0.004 | 0.001 | 0.01 | .02 |
| ADE | 0.025 | 0.003 | 0.04 | .04 |
| Proportion mediated | 0.125 | 0.030 | 0.48 | .04 |
ACME = average mediation effect, ADE = average direct effect, LnFS = natural logarithm of frailty score, LTBI = latent tuberculosis infection.
Figure 3.
A mediation analysis framework illustrating the partial mediating role of globulin (M) in the association between logarithmic frailty score (LnFS, X) and latent tuberculosis infection (LTBI) risk (Y). The total effect (TE) of LnFS on LTBI was 0.028 (95% CI: 0.007–0.05, P = .02), decomposed into a direct effect (ADE) of 0.025 (95% CI: 0.003–0.04, P = .04) and an indirect effect (i.e., mediated by globulin) of 0.004 (95% CI: 0.001–0.01, P = .02). The proportion of mediation was 12.5% (95% CI: 3.0%–48%, P = .04). Solid arrows indicate significant pathways, with effect sizes and confidence intervals derived from statistical modeling.
3.5. Kaplan–Meier survival analysis
The weighted mean follow-up time for all participants was 7.9 years (95% CI: 7.7–8.1). When stratified by LTBI status, the mean follow-up duration was 7.91 years (95% CI: 7.70–8.11) in the non-LTBI group and 7.87 years (95% CI: 7.67–8.06) in the LTBI group. Patients were divided into 4 equal groups (Q1–Q4) based on their FI scores. The Kaplan–Meier survival analysis revealed strikingly different outcomes across these groups, with statistical significance reaching well beyond the conventional threshold (P < .0001). The survival curves painted a clear picture: frailty levels dramatically impacted longevity, with each quartile showing progressively worse outcomes. Patients in Q1, representing the lowest quartile of the FI, had the highest overall survival probability throughout the follow-up period. In contrast, patients in Q4, the highest quartile of the FI, demonstrated the lowest overall survival probability (Fig. 4).
Figure 4.
Kaplan–Meier survival curves for overall survival stratified by frailty index quartiles. The y-axis represents the overall survival probability, ranging from 0.00 to 1.00. The x-axis indicates the time in months, with follow-up extending up to 96 months. The 4 colored lines represent different quartiles: Q1 (red), Q2 (cyan), Q3 (green), and Q4 (blue). The P-value for the log-rank test comparing the survival curves among the 4 groups is shown as P < .0001, indicating a statistically significant difference in overall survival. The table below the curve shows the number of patients at risk at different time points (0, 24, 48, 72, and 96 months) for each quartile. The shaded areas around each curve represent the 95% confidence intervals for the survival probabilities.
3.6. Survival analysis
The analysis of HR further elucidates the relationship between the LnFS and all-cause mortality. In the unadjusted model, each 1-unit increase in LnFS was associated with a HR of 6.25 (95% CI: 3.22–12.10, P < .001), indicating a markedly increased risk of mortality. Upon adjusting for age and sex, the HR decreased to 4.58 (95% CI: 2.01–10.5, P < .001) in the second model. Further adjustments for lymphocyte count and BMI in the third model resulted in a HR of 4.13 (95% CI: 1.67–10.2, P = .002). Stratification by quartiles of LnFS revealed that individuals in the highest quartile (Q4) had a significantly elevated risk of all-cause mortality compared to those in the lowest quartile (Q1). The HR for Q4 compared to Q1 were 1.79 (95% CI: 1.27–2.52, P = .003) in the unadjusted model, 1.68 (95% CI: 1.19–2.37, P = .007) in the age- and sex-adjusted model, and 1.73 (95% CI: 1.18–2.53, P = .009) in the fully adjusted model (Table 4).
Table 4.
Analysis of the association between LnFS and all-cause mortality.
| Index | Model 1 (HR, 95 CI%) | P value | Model 2 (HR, 95 CI%) | P value | Model 3 (HR, 95 CI%) | P value |
|---|---|---|---|---|---|---|
| LnFS | 6.25 (3.22–12.10) | <.001 | 4.58 (2.01–10.5) | <.001 | 4.13 (1.67–10.2) | .002 |
| Q1 | Ref | Ref | Ref | |||
| Q2 | 1.26 (0.84–1.90) | .234 | 1.23 (0.82–1.83) | .284 | 1.23 (0.82–1.85) | .277 |
| Q3 | 1.26 (0.82–1.91) | .266 | 1.18 (0.77–1.80) | .419 | 1.19 (0.77–1.85) | .392 |
| Q4 | 1.79 (1.27–2.52) | .003 | 1.68 (1.19–2.37) | .007 | 1.73 (1.18–2.53) | .009 |
| P for trend | .003 | .008 | .011 |
Model 1: no covariates were adjusted. Model 2: age and sex were adjusted. Model 3: age, sex, lymphocyte and body mass index were adjusted.
CI = confidence interval, HR = hazard ratio, LnFS = natural logarithm of frailty score.
3.7. Analysis of RCS regarding the influence of LnFS on overall mortality rates
Cox regression analysis revealed a significant association between LnFS and the outcome event (P-overall < .001), while the non-linear test indicated no evidence of a non-linear relationship (P-non-linear = .812). The RCS curve demonstrated that when LnFS was relatively low (approximately <−2), the HR remained close to 1, with the 95% CI encompassing 1, suggesting a negligible effect of LnFS on the outcome risk within this range. As LnFS increased beyond −2, the HR rose sharply, and the 95% CI progressively moved away from 1, indicating that higher LnFS values were significantly associated with an increased risk of the outcome event (Fig. 5).
Figure 5.
Restricted cubic spline (RCS) curve illustrating the dose-response relationship between natural logarithm transformation frailty score (LnFS) and the hazard ratio (HR) for the outcome event. The x-axis represents the LnFS, while the y-axis denotes the HR with 95% confidence intervals (shaded area). The curve indicates that LogFS had a negligible effect on outcome risk at lower values (LnFS <−2), whereas higher LnFS values were significantly associated with an increased risk of the outcome.
3.8. Subgroup analysis
Subgroup analysis based on logistic regression revealed that the association between the LnFS and LTBI varied across different demographic groups. In the age-stratified analysis, the association was significantly stronger in the group aged ≥65 years (OR = 1.54, 95% CI: 1.10–2.16, P = .023) compared to those aged <65 years (OR = 1.20, 95% CI: 1.00–1.45, P = .073). However, the interaction between age and LnFS was not statistically significant (P = .242) (Table 5).
Table 5.
Subgroup analysis of the correlation between LnFS and LTBI.
| Variables | n (%) | OR (95% CI) | P | P for interaction |
|---|---|---|---|---|
| Age | .242 | |||
| <65 | 2883 (81.9) | 1.20 (1.00–1.45) | .073 | |
| ≥65 | 637 (18.1) | 1.54 (1.10–2.16) | .023 | |
| BMI | .249 | |||
| <25 | 1054 (29.94) | 1.13 (0.81–1.58) | .482 | |
| 25–30 | 1200 (34.09) | 1.25 (0.92–1.69) | .177 | |
| >30 | 1266 (35.97) | 1.52 (1.05–2.20) | .041 | |
| Sex | .508 | |||
| Male | 1713 (48.66) | 1.46 (1.16–1.84) | .006 | |
| Female | 1807 (51.34) | 1.28 (1.02–1.60) | .048 | |
| Marriage | .580 | |||
| Married | 1694 (48.12) | 1.24 (0.98–1.58) | .094 | |
| Widowed | 202 (5.74) | 1.32 (0.42–4.15) | 0638 | |
| Divorced | 400 (11.36) | 1.10 (0.43–2.81) | .849 | |
| Other | 1224 (34.77) | 1.38 (1.11–1.71) | .011 | |
| Smoke | .850 | |||
| Former | 804 (22.84) | 1.15 (0.84–1.57) | .395 | |
| Never | 1972 (56.02) | 1.37 (1.07–1.74) | .023 | |
| Now | 744 (21.14) | 1.09 (0.75–1.58) | .652 |
BMI = body mass index. Adjusted for Lymphocyte, CI = confidence interval, LnFS = natural logarithm of frailty score, LTBI = latent tuberculosis infection, OR = odds ratio.
In the BMI stratification, a significant association was observed only in the group with BMI > 30 kg/m2 (OR = 1.52, 95% CI: 1.05–2.20, P = .041), while no significant associations were found in the BMI < 25 kg/m2 and 25 to 30 kg/m2 groups (P > .05 for both). The interaction between BMI and LnFS was also not significant (P = .249). When stratified by sex, both men (OR = 1.46, 95% CI: 1.16–1.84, P = .006) and women (OR = 1.28, 95% CI: 1.02–1.60, P = .048) demonstrated significant associations; however, the interaction between sex and LnFS was not statistically significant (P = .508). These findings suggest that the association between LnFS and LTBI may be more pronounced in older and obese populations, although the differences across various demographic groups did not reach statistical significance (Table 5).
Cox regression-based survival analysis subgroup results indicated that the association between LnFS and all-cause mortality in LTBI patients also varied by demographic characteristics. The HR for LnFS was significantly higher in the group aged <65 years (HR = 8.25, 95% CI: 2.96–23.04, P < .001) compared to those aged ≥65 years (HR = 3.53, 95% CI: 1.73–7.24, P = .001). In the BMI stratification, the highest mortality risk was observed in the 25 to 30 kg/m2 group (HR = 26.33, 95% CI: 6.50–106.74, P < .001), followed by the <25 kg/m2 group (HR = 5.55, 95% CI: 1.93–15.97, P = .001) and the >30 kg/m2 group (HR = 3.90, 95% CI: 1.59–9.59, P = .003). However, the interaction between BMI and LnFS was not significant (P = .701). Significant effects were observed in both men and women. All subgroup analyses were adjusted for lymphocyte count. These results suggest that the predictive power of LnFS for mortality risk in LTBI patients may be more pronounced in younger, obese, and female populations (Table 6).
Table 6.
Subgroup analysis of the correlation between LnFS and all-cause mortality in LTBI patients.
| Variables | n (%) | HR (95% CI) | P | P for interaction |
|---|---|---|---|---|
| Age | .095 | |||
| <65 | 430 (76.4) | 8.25 (2.96–23.04) | <.001 | |
| ≥65 | 133 (23.6) | 3.53 (1.73–7.24) | .001 | |
| BMI | .701 | |||
| <25 | 176 (31.3) | 5.55 (1.93–15.97) | .001 | |
| 25–30 | 200 (35.5) | 26.33 (6.50–106.74) | <.001 | |
| >30 | 187 (33.2) | 3.90 (1.59–9.59) | .003 | |
| Sex | .263 | |||
| Male | 233 (41.4) | 3.83 (1.14–12.93) | .030 | |
| Female | 330 (58.6) | 8.51 (3.69–19.64) | <.001 | |
| Marriage | .698 | |||
| Married | 317 (56.3) | 8.35 (3.03–22.98) | <.001 | |
| Widowed | 39 (6.9) | 1.16 (0.36–3.69) | .803 | |
| Divorced | 58 (10.3) | 5.64 (0.64–50.05) | .120 | |
| Other | 149 (26.5) | 11.32 (2.81–45.57) | .001 | |
| Smoke | .211 | |||
| Former | 148 (26.3) | 6.32 (3.15–12.68) | <.001 | |
| Never | 292 (51.9) | 2.40 (1.16–4.96) | .018 | |
| Now | 123 (21.8) | 14.67 (3.03–71.06) | .001 |
BMI = body mass index. Adjusted for lymphocyte, LTBI = latent tuberculosis infection, OR = odds ratio.
4. Discussion
This study utilized data from the NHANES 2011 to 2012 cycle to investigate the relationship between LTBI and FS, including their natural logarithm transformation (LnFS). Additionally, it assessed the impact of this association on all-cause mortality. The results indicate that the increased level of LnFS is associated with the occurrence of LTBI and increasing mortality among LTBI patients. Furthermore, serum globulin levels were found to mediate the association between LnFS and LTBI. This finding suggests that the inflammatory processes underlying frailty may play a critical role in the susceptibility to LTBI, highlighting the importance of addressing frailty in clinical settings.
The LTBI remains a significant preventable health issue in the United States.[17] According to the latest recommendations from the U.S. Preventive Services Task Force, LTBI refers to a state in which an individual is infected with MtB, but the bacteria are controlled by the immune system. Individuals with LTBI do not experience symptoms and cannot transmit TB to others; however, they typically test positive on a TST or an interferon-gamma release assay.[18] It is estimated that approximately 5.0% of the population in the United States, or around 13 million individuals, are affected by LTBI.[19] If left untreated, approximately 5% to 10% of healthy individuals with normal immunity may progress to active TB disease. In certain high-risk populations or among individuals with specific comorbidities, this progression rate may be even higher. Within the LTBI population, abnormalities in immune cell function and quantity have been observed; for instance, monocytopenia may impair the ability to clear the bacteria, while abnormal distribution of regulatory T cells may also play a role in the immune modulation associated with LTBI.[20,21] Immunological research has shown that individuals in these high-risk groups exhibit weakened cellular immune responses, particularly with regard to impaired IFN-γ-mediated Th1 responses. This deficiency results in the body’s inability to effectively control the replication of MtB.[22–24] MtB evades host macrophages by escaping or disrupting immune effector functions. The genome of MtB encodes PE/PPE/PE_PGRS proteins, which are inherently disordered, redundant, and antigenic. These proteins primarily function by modulating immune responses, thereby enhancing the virulence of MtB and influencing the immune-mediated clearance of the pathogen.[25] These mechanisms not only elucidate the persistence of LTBI but also provide a theoretical basis for the development of diagnostic biomarkers and vaccine targets.
Frailty is closely associated with the pathophysiological basis of chronic inflammation.[26] Previous studies have indicated that individuals with frailty exhibit significantly elevated levels of serum inflammatory markers, suggesting that chronic low-grade inflammation plays a central role in the onset and progression of frailty.[27] In this study, the FS of patients in the LTBI group was significantly elevated, and the levels of globulins were notably increased among LTBI patients, further supporting the role of chronic inflammation as a bridge between frailty and LTBI. As a persistent low-grade infectious state, LTBI may activate the immune system, triggering a chronic inflammatory response that accelerates muscle wasting, metabolic dysregulation, and organ function decline, ultimately leading to increased FS. Mechanistically, the frail state may heighten the risk of infection through the following pathways: decreased immune cell function and reduced antibody production;[28] metabolic dysregulation: malnutrition and microcirculatory disturbances impair tissue repair and pathogen clearance.[29] Considering that LTBI patients often present with no clinical symptoms, traditional risk assessments may not fully capture their immune status and conversion risk. The FI, by integrating multisystem health deficits, may provide a more comprehensive risk evaluation. Managing LTBI treatment for recent immigrants in primary care settings by general practitioners and community pharmacists is a safe and cost-effective option.[30] The findings of this study suggest that the FI can effectively identify the risk associated with LTBI, providing a valuable reference for enhancing disease management strategies.
Immunosenescence is the term used to describe the slow but steady deterioration of our immune system as we age. This decline manifests through a decrease in the count of immune cells, issues with their functionality, and an imbalance in inflammatory reactions. Previous studies have shown that frail individuals exhibit an imbalance in T cell subpopulations (such as an increase in CD8 + T cells and a decrease in the CD4+/CD8 + ratio), along with reduced natural killer cell activity. These findings suggest a close relationship between immunosenescence and frailty.[31,32] In this study, LTBI patients exhibited a significant increase in lymphocyte counts, which may reflect the immune system’s sustained response to MtB. However, this response may be accompanied by characteristics of immunosenescence. The establishment and maintenance of LTBI depend on the dynamic balance between the host’s immune system and MtB.[33] Following MtB infection, the host primarily controls bacterial replication through T cell-mediated adaptive immunity, as well as innate immunity involving monocytes and macrophages. This immune response leads to the formation of granulomas, which effectively restrict the spread of the bacteria.[34] As the body ages, the thymus becomes less functional, leading to a decline in the generation of naïve T cells. This deterioration is marked by higher concentrations of pro-inflammatory signaling molecules, including IL-6 and TNF-α, while anti-inflammatory cytokines such as IL-10 become increasingly scarce.[35] This state of chronic low-grade inflammation may facilitate the progression of LTBI to active TB disease.
Metabolic dysregulation is 1 of the key pathophysiological foundations of frailty. Individuals with frailty often exhibit abnormalities in metabolic indicators such as blood glucose, lipid levels, and insulin resistance. These abnormalities may contribute to frailty by affecting energy supply, muscle synthesis, and organ function.[36,37] In this study, the prevalence of diabetes was significantly elevated among LTBI patients, and the association between BMI and FS was more pronounced in individuals with higher BMI. This suggests that metabolic dysregulation may play a role in the relationship between LTBI and frailty. Furthermore, frail individuals exhibited significantly lower muscle mass compared to their non-frail counterparts, with muscle wasting closely linked to inflammatory responses and metabolic dysregulation. LTBI may also indirectly contribute to insufficient energy intake and reduced muscle synthesis through mechanisms affecting appetite and nutrient absorption, further exacerbating muscle wasting and frailty.[38] Frail individuals often exhibit dysfunction of the hypothalamic-pituitary-adrenal axis, characterized by elevated cortisol levels and disrupted circadian rhythms. Abnormal hypothalamic-pituitary-adrenal axis function may lead to muscle wasting and frailty through mechanisms that promote catabolism and inhibit anabolism. This study further explores the association between LTBI, a specific infectious state, and frailty, revealing the mediating role of globulin in this relationship. This research not only expands the use of frailty indices but also delves into the lasting health consequences of LTBI. The findings carry substantial scientific weight for public health and clinical endeavors, underscoring the necessity of incorporating frailty evaluations for high-risk LTBI groups to ensure the early detection of vulnerable individuals. Additionally, the study underscores the importance of frailty in clinical prognosis, offering clinicians a new tool for prognostic evaluation. This research also provides new targets for future intervention studies that focusing mediation effect between FS and LTBI induced by globulin levels.
Despite the innovative and rigorous nature of this study, certain limitations must be acknowledged. First, as a cross sectional study, it cannot establish a causal relationship between LTBI and frailty. Future longitudinal studies or randomized controlled trials would be beneficial in more accurately assessing the causal link between the 2. Second, the data for this study were sourced from the NHANES database, which may not fully reflect current health statuses and disease patterns. Third, our reliance on complete-case analysis may limit generalizability if individuals with missing data differed systematically from those included; however, the use of survey weights helps mitigate potential selection bias. Future research should incorporate more recent data to enhance the timeliness and generalizability of the findings. Additionally, future studies should consider a broader range of potential confounding factors, such as genetic background and environmental influences, to provide a more comprehensive evaluation of the association between LTBI and frailty.
5. Conclusion
FS in LTBI patients was significantly higher than that in non-LTBI populations, with a strong association between the LnFS and presence of LTBI. Globulin levels played a partial mediating role in this relationship, accounting for 12.5% of the mediation effect. This study not only confirms a significant positive correlation between FS and LTBI but also reveals the profound implications of this association for all-cause mortality. These findings provide important insights into the biological mechanisms underlying the relationship between frailty and LTBI and emphasize the potential application of serum globulin as a biomarker of inflammatory response in assessing LTBI risk.
Acknowledgments
We acknowledge the contributions of the NHANES program.
Author contributions
Data curation: Mingxia Sun, Jinping Fan, Yan Yang, Aiqing Lin.
Funding acquisition: Mingxia Sun.
Investigation: Mingxia Sun, Jinping Fan, Yan Yang, Aiqing Lin.
Methodology: Mingxia Sun, Aiqing Lin.
Software: Mingxia Sun, Yan Yang, Aiqing Lin.
Validation: Mingxia Sun, Aiqing Lin.
Visualization: Mingxia Sun.
Formal analysis: Jinping Fan, Aiqing Lin.
Supervision: Aiqing Lin.
Writing – original draft: Mingxia Sun, Jinping Fan, Yan Yang.
Writing – review & editing: Aiqing Lin.
Supplementary Material
Abbreviations:
- BMI
- body mass index
- CI
- confidence interval
- FI
- frailty index
- FS
- frailty score
- HR
- hazard ratio
- IFN-γ
- interferon-gamma
- IU
- international unit
- LnFS
- natural logarithm of FS
- LTBI
- latent tuberculosis infection
- MtB
- Mycobacterium tuberculosis
- NHANES
- National Health and Nutrition Examination Survey
- OR
- odds ratio
- RCS
- restricted cubic spline
- TB
- tuberculosis
- TST
- tuberculin skin test
The author(s) declare that financial support was received for the research and/or publication of this article. This work was supported by the Medical and Health Science and Technology Project of Shandong Province (No: 202403020519).
The authors have no conflicts of interest to disclose.
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.
Supplemental Digital Content is available for this article.
How to cite this article: Sun M, Fan J, Yang Y, Lin A. The impact of frailty scores on latent tuberculosis infection and all-cause mortality in this population. Medicine 2026;105:7(e47490).
All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.
Contributor Information
Mingxia Sun, Email: sunmx1983@126.com.
Jinping Fan, Email: fxiaoally@163.com.
Yan Yang, Email: 513098852@qq.com.
References
- [1].De Almeida LLC, Fernandes SP, De Oliveira GD, et al. Harnessing actinobacteria secondary metabolites for tuberculosis drug discovery: historical trends, current status and future outlooks. Nat Prod Bioprospect. 2025;15:52. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Janssen S, Murphy M, Upton C, Allwood B, Diacon AH. Tuberculosis: an update for the clinician. Respirology. 2025;30:196–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Magodoro IM, Wilkinson KA, Claggett BL, Ntusi NAB, Siedner M MJ, Wilkinson RJ. Discordance between measures of Mycobacterium tuberculosis sensitization and type 2 diabetes mellitus in the United States (NHANES): a population-based cohort study. J Infect. 2025;90:106496. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [4].Jonas DE, Riley SR, Lee LC, et al. Screening for latent tuberculosis infection in adults: updated evidence report and systematic review for the US preventive services task force. JAMA. 2023;329:1495–509. [DOI] [PubMed] [Google Scholar]
- [5].Zhou J, Li Y, Zhu L, Yue R. Association between frailty index and cognitive dysfunction in older adults: insights from the 2011-2014 NHANES data. Front Aging Neurosci. 2024;16:1458542. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Rockwood K, Song X, Macknight C, et al. A global clinical measure of fitness and frailty in elderly people. CMAJ. 2005;173:489–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [7].Jayanama K, Theou O, Blodgett JM, Cahill L, Rockwood K. Frailty, nutrition-related parameters, and mortality across the adult age spectrum. BMC Med. 2018;16:188. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [8].Yan Z, Xu Y, Li K, Liu L. The correlation between frailty index and incidence, mortality in obstructive sleep apnea: evidence from NHANES. Heliyon. 2024;10:e32514. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Zhao YC, Wu SQ, Li JK, et al. Predictive value of the combined triglyceride-glucose and frailty index for cardiovascular disease and stroke in two prospective cohorts. Cardiovasc Diabetol. 2025;24:318. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [10].Jayanama K, Theou O, Godin J, et al. Relationship between diet quality scores and the risk of frailty and mortality in adults across a wide age spectrum. BMC Med. 2021;19:64. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [11].Liu X, Wang Y, Shen L, et al. Association between frailty and chronic constipation and chronic diarrhea among American older adults: National Health and Nutrition Examination Survey. BMC Geriatr. 2023;23:745. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Jensen PA, Lambert LA, Iademarco MF, Ridzon R; CDC. Guidelines for preventing the transmission of Mycobacterium tuberculosis in health-care settings, 2005. MMWR Recomm Rep. 2005;54:1–141. [PubMed] [Google Scholar]
- [13].Liu X, Wang Y, Huang Y, et al. Association between frailty index and mortality in depressed patients: results from NHANES 2005-2018. Sci Rep. 2025;15:3305. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [14].Searle SD, Mitnitski A, Gahbauer EA, Gill TM, Rockwood K. A standard procedure for creating a frailty index. BMC Geriatr. 2008;8:24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [15].Wang K, Zhao Y, Nie J, Xu H, Yu C, Wang S. Higher HEI-2015 score is associated with reduced risk of depression: result from NHANES 2005-2016. Nutrients. 2021;13:348. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Selvin E, Manzi J, Stevens LA, et al. Calibration of serum creatinine in the National Health and Nutrition Examination Surveys (NHANES) 1988-1994, 1999-2004. Am J Kidney Dis. 2007;50:918–26. [DOI] [PubMed] [Google Scholar]
- [17].Woodruff R, Miramontes R. Tuberculosis infection among non-US-born persons and persons ≥60 years of age, United States, 2019-2020. Emerg Infect Dis. 2023;29:1470–2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Mangione CM, Barry MJ, Nicholson WK, et al. ; US Preventive Services Task Force. Screening for latent tuberculosis infection in adults: US preventive services task force recommendation statement. JAMA. 2023;329:1487–94. [DOI] [PubMed] [Google Scholar]
- [19].Miramontes R, Hill AN, Yelk Woodruff RS, et al. Tuberculosis infection in the united states: prevalence estimates from the National Health and Nutrition Examination Survey, 2011-2012. PLoS One. 2015;10:e0140881. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Qin S, Chen R, Li M, et al. Changes in immune status of circulating NK cells in patients with latent tuberculosis infection. Cent Eur J Immunol. 2024;49:105–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [21].Wang X, Jiang K, Xing W, et al. Clustering Mycobacterium tuberculosis-specific CD154(+)CD4(+) T cells for distinguishing tuberculosis disease from infection based on single-cell RNA-seq analysis. J Infect. 2025;90:106449. [DOI] [PubMed] [Google Scholar]
- [22].Arora SK, Alam A, Naqvi N, et al. Immunodominant Mycobacterium tuberculosis protein Rv1507A elicits Th1 response and modulates host macrophage effector functions. Front Immunol. 2020;11:1199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Essone PN, Leboueny M, Maloupazoa Siawaya AC, et al. M. tuberculosis infection and antigen specific cytokine response in healthcare workers frequently exposed to tuberculosis. Sci Rep. 2019;9:8201. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [24].Luo J, Zhang M, Yan B, Zhang K, Chen M, Deng S. Imbalance of Th17 and Treg in peripheral blood mononuclear cells of active tuberculosis patients. Braz J Infect Dis. 2017;21:155–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [25].Sharma T, Alam A, Ehtram A, et al. The Mycobacterium tuberculosis PE_PGRS protein family acts as an immunological decoy to subvert host immune response. Int J Mol Sci. 2022;23:525. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Wilkinson TJ, Miksza J, Zaccardi F, et al. Associations between frailty trajectories and cardiovascular, renal, and mortality outcomes in chronic kidney disease. J Cachexia Sarcopenia Muscle. 2022;13:2426–35. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Van Sleen Y, Shetty SA, Van Der Heiden M, et al. Frailty is related to serum inflammageing markers: results from the VITAL study. Immun Ageing. 2023;20:68. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Luo OJ, Lei W, Zhu G, et al. Multidimensional single-cell analysis of human peripheral blood reveals characteristic features of the immune system landscape in aging and frailty. Nat Aging. 2022;2:348–64. [DOI] [PubMed] [Google Scholar]
- [29].Yang Y, Che K, Deng J, et al. Assessing the impact of frailty on infection risk in older adults: prospective observational cohort study. JMIR Public Health Surveill. 2024;10:e59762. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Chhabra S, Koh MCY, Allen DM. The treatment of latent tuberculosis infection in migrants in primary care versus secondary care. Eur Respir J. 2024;64:2401569. [DOI] [PubMed] [Google Scholar]
- [31].Guaraldi G, Zona S, Silva AR, et al. The dynamic association between Frailty, CD4 and CD4/CD8 ratio in people aging with HIV. PLoS One. 2019;14:e0212283. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].Semmarath W, Seesen M, Yodkeeree S, et al. The association between frailty indicators and blood-based biomarkers in early-old community dwellers of Thailand. Int J Environ Res Public Health. 2019;16:3457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [33].Ferluga J, Yasmin H, Al-Ahdal MN, Bhakta S, Kishore U. Natural and trained innate immunity against Mycobacterium tuberculosis. Immunobiology. 2020;225:151951. [DOI] [PubMed] [Google Scholar]
- [34].Lu LL, Chung AW, Rosebrock TR, et al. A functional role for antibodies in tuberculosis. Cell. 2016;167:433–43.e14. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Hall BM, Gleiberman AS, Strom E, et al. Immune checkpoint protein VSIG4 as a biomarker of aging in murine adipose tissue. Aging Cell. 2020;19:e13219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].Tian H, Li YM, Wang CQ, Chen G-Q, Lian Y. Association between non-insulin-based insulin resistance indicators and frailty progression: a national cohort study and Mendelian randomization analysis. Cardiovasc Diabetol. 2025;24:31. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [37].Liu Y, Xin Y, Wang W, et al. The association between metabolic syndrome severity and frailty risk in patients with rheumatoid arthritis: a cross-sectional study. Arthritis Res Ther. 2025;27:145. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Tuan SH, Chang LH, Sun SF, Li C-H, Chen G-B, Tsai Y-J. Assessing the clinical effectiveness of an exergame-based exercise training program using ring fit adventure to prevent and postpone frailty and sarcopenia among older adults in rural long-term care facilities: randomized controlled trial. J Med Internet Res. 2024;26:e59468. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.