1 Introduction
Tuberculosis, caused by
Mycobacterium tuberculosis infection, remains a major global public health threat.
1 Elderly patients are at high risk of severe progression and anti-TB medication adverse events due to physiological decline, weakened immune function, and multiple comorbidities (e.g., cardiovascular diseases, diabetes)—these factors complicate geriatric anti-TB medication management.
2–4 Thus, there is an urgent need for thyroxine level-related biomarkers to guide anti-TB individualized medication, which is critical for optimizing treatment safety and efficacy in the elderly.
In recent years, the role of thyroid function in guiding individualized medication for chronic diseases has gained increasing attention in intelligent pharmacy.
5–7 As the primary active hormones in the thyroid regulatory pathway, T3 (TT3 and FT3) regulate energy metabolism and immune function—key factors mediating medication tolerance and therapeutic response in elderly patients. Compared with T4 and TSH, TT3 and FT3 more directly reflect the body's ability to tolerate anti-TB drugs, making them ideal candidate biomarkers for geriatric pharmaceutical care. Existing studies confirm that low TT3 and FT3 levels correlate with poor prognosis in severe tuberculosis, implying their potential value in thyroxine level-guided anti-TB individualized medication.
8,9The treatment response and medication tolerance of elderly tuberculosis patients differ significantly from young patients, with ‘early acute exacerbation’ as a typical feature: about 70% of severe events occur within one week of admission, making early medication adjustment critical for prognosis.
10–13 However, research on thyroxine level-related biomarkers for guiding early anti-TB individualized medication in this high-risk group is lacking. Based on this gap, this study retrospectively analyzed clinical data of 509 elderly tuberculosis patients admitted to our hospital, to explore the correlation between serum TT3 and FT3 levels and disease severity within one week of admission. It aims to verify the value of TT3 and FT3 as biomarkers for thyroxine level-guided anti-TB individualized medication, providing pharmacists with a precise tool for risk stratification and dose adjustment, and laying a scientific basis for optimizing geriatric anti-TB pharmaceutical care.
2 Materials and methods
2.1 Study population and grouping
This prospective observational study enrolled 509 tuberculosis patients admitted to our hospital (Jan 2022–Dec 2023). Only patients with active pulmonary tuberculosis were included, extrapulmonary or latent TB cases were excluded. Using baseline TT3 and FT3 levels measured within 24 h of admission, patients were followed for 1 week. Those meeting severe criteria (n = 270) within this period were classified into the severe group, with the remainder (n = 239) as non-severe. The ethical approval for this study was obtained from the Beijing Chest Hospital Ethics Committee (ethical approval No. BJKC-2025-KY-48)with informed consent obtained from all participants.
The severe group was defined as patients experiencing severe hypoxemia or acute respiratory failure requiring ventilation support, hypotension, shock, or other organ dysfunction unrelated to tuberculosis. Additionally, patients were further subclassified into an elderly group(aged ≥65 years) and a non-elderly group based on age, to explore the influence of age on the study results.
2.2 Enrollment criteria
Relevant Standards: (1) Tuberculosis Diagnosis: Strictly adhere to the “Diagnostic Criteria for Tuberculosis” (latest version) issued by the National Health Commission to ensure the accuracy and consistency of diagnoses. (2) Thyroid Hormone Level Standards: Determine the normal reference ranges for TT3 and FT3 based on the laboratory standards of our hospital. (3) Diagnostic Criteria for Severe pulmonary tuberculosis:In accordance with the “Definition and Diagnostic Guidelines for Severe pulmonary tuberculosis in Chinese Adults” (2023), which defines Severe pulmonary tuberculosis as active tuberculous lesions in the lung tissue, trachea, bronchi, and pleura, accompanied by severe hypoxemia, requiring mechanical ventilation, hypotension, shock, or other organ dysfunction.
14 (4) Diagnostic Criteria for Elderly Patients: Age ≥65 years.
2.3 Inclusion criteria
(1) Age: Ensure that patients are aged ≥18 years to align with the study scope of adult tuberculosis. (2) Diagnosis: Confirmed diagnosis of tuberculosis through clinical and/or laboratory examinations, and meeting the aforementioned diagnostic criteria for tuberculosis.
2.4 Exclusion criteria
(1) Thyroid Diseases: Previous history of thyroid-related diseases or ultrasound indicating thyroid abnormalities. (2) Pregnancy; (3) Other Severe Diseases: Malignant tumors, severe autoimmune diseases, HIV infection, etc.(4) Recent Surgery or Fasting: History of major surgery or long-term fasting within the past 3 months. (5) Missing Clinical Data:Incomplete clinical data, making it impossible to accurately assess the patient's condition or thyroid hormone status.
2.5 Sample size calculation
This study used a sample size calculation formula based on the effect size reported in previous studies or literature, with an alpha error rate set at 0.05 and a beta error rate set at 0.20 (i.e., power of 80%), to calculate the minimum required sample size. After detailed calculations, at least 474 tuberculosis patients were determined to be included to ensure the reliability and statistical power. Actually, 509 patients were enrolled, meeting the sample size requirement and providing assurance for the accuracy and reliability of the study.
2.6 Data collection
Patient demographic data, including age, sex, weight, height, and medical history of underlying diseases, were systematically recorded. Baseline serum TT3 and FT3 levels were measured within 24 h of admission, with severe disease classification conducted within 7 days after admission to ensure temporal precedence of exposure (thyroid hormone levels) over outcomes (disease severity). Additional laboratory parameters (complete blood count, liver/kidney function tests) and clinical progression data were continuously monitored during the first week, as shown in Fig. 1.
2.7 Statistical methods
All continuous variables underwent normality assessment using the Shapiro–Wilk test. Results confirmed that the primary metrics in this study followed a normal distribution (P > 0.05), providing a robust foundation for subsequent parametric statistical analyses.
Data analysis was performed using SPSS 25.0 statistical software to ensure the scientific validity and accuracy of the statistical methods. Quantitative data were expressed as mean ± standard deviation (x ± s). Since the data passed the normality test and were confirmed to follow a normal distribution, independent samples t-test was used for comparisons between the two groups. Categorical data were presented as rates(%) and compared using the chi-square test or Fisher's exact test.
The Pearson point-biserial correlation coefficient was employed to measure the correlation between continuous variables and dichotomous variables—this study focuses on the short-term dichotomous outcome of “whether severe progression occurs within 1 week after admission”. All patients completed the observation during this period, so differences in the timing of outcome occurrence were not considered; the point-biserial correlation could directly quantify the strength of the association between baseline T3 levels and short-term severe risk.
Furthermore, receiver operating characteristic (ROC) curves were plotted, with corresponding cutoff values calculated to assess the accuracy and value of relevant indicators in predicting disease severity.
3 Research results
3.1 Comparison of general and laboratory data
In this study, we compared the baseline data between non-severe(N = 239) and severe (N = 270) patients. The mean age of the severe group was significantly higher than that of the non-severe group, and the severe group was more prone to having coronary heart disease, renal and liver dysfunction, and malnutrition than the non-severe group. Additionally, TT3 and FT3 levels were significantly lower in the severe group, with statistical significance (P < 0.05). As shown in Table 1.
3.2 Pearson correlation analysis (Point–biserial correlation coefficient form)
Pearson correlation analysis revealed a negative correlation between TT3 and FT3 levels and severe pulmonary tuberculosis. Specifically, the correlation coefficient between TT3 and severe pulmonary tuberculosis was r =−0.541 (P = 0), and the correlation coefficient between FT3 and severe pulmonary tuberculosis was r =−0.423 (P = 0).
Further dividing the patients into an elderly group and a non-elderly group, the results showed that the negative correlation between TT3, FT3, and severe pulmonary tuberculosis remained significant in both subgroups. Among them, the correlation of TT3 in the elderly group was slightly weaker than that in the non-elderly group (elderly group:r =−0.516, P < 0.001; non-elderly group: r =−0.529, P < 0.001);whereas the correlation of FT3 in the elderly group was stronger than that in the non-elderly group (elderly group: r =−0.491, P < 0.001; non-elderly group: r =−0.380, P < 0.001). As shown in Table 2.
3.3 Evaluation of serum TT3 and FT3 for severe pulmonary tuberculosis
Tuberculosis Using Serum TT3 and FT3 Levels ROC curve analysis indicated that both serum TT3 and FT3 levels exhibited high diagnostic accuracy for evaluating severe pulmonary tuberculosis, with areas under the ROC curve of 0.841 and 0.827, respectively. The sensitivity and specificity of the two were similar. After grouping, the serum TT3 and FT3 levels in the elderly group showed higher sensitivity but slightly lower specificity compared to the non-elderly group. See Fig. 2 and Table 3.
4 Discussion
This study conducted a cohort study of 509 tuberculosis patients to investigate the correlation between serum TT3 and FT3 levels and disease severity in elderly tuberculosis patients. Based on the outcome characteristics of 1-week short-term follow-up, Pearson correlation analysis and ROC curve analysis were adopted to quantify the association between baseline T3 levels and early severe risk, evaluate predictive efficacy, and align with the research objectives. The following discussion focuses on differences in TT3 and FT3 levels, correlations, and subgroup analyses.
4.1 Significant differences in TT3 and FT3 levels between severe and non-severe pulmonary tuberculosis patients
In baseline data comparison, severe pulmonary tuberculosis patients had significantly lower TT3 (1.03 ± 0.37 nmol/L) and FT3(2.59 ± 0.93 pmol/L) levels than non-severe patients(1.66 ± 0.59 nmol/L, 4.01 ± 1.99 pmol/L; both
P < 0.001). This finding aligns with previous studies linking low T3 syndrome to poor prognosis and reduced medication tolerance in severe diseases.
15,16 For example, Khan et al. (2006) reported that 50% of pulmonary tuberculosis patients had abnormally low TT3 levels, which correlated with increased drug toxicity.
17 These results confirm that TT3 and FT3 levels are closely associated with both disease severity and anti-TB medication tolerance, laying the foundation for their application in thyroxine level-guided individualized medication.
4.2 Correlation between TT3, FT3 levels and development of severe pulmonary tuberculosis within one week after admission
Pearson correlation analysis (point-biserial coefficient) revealed a significant negative correlation between TT3 and FT3 levels and severe pulmonary tuberculosis (TT3: r =−0.541,
P < 0.001; FT3: r =−0.423,
P < 0.001), indicating that lower TT3 and FT3 levels are associated with more severe disease and higher anti-TB medication risk. This is consistent with clinical observations and previous studies on medication-related prognosis.
8,18–20 ROC curve analysis confirmed high accuracy of TT3 and FT3 for predicting severe disease and medication tolerance risk (AUC = 0.841 for TT3, 0.827 for FT3), providing strong evidence for their utility as biomarkers in thyroxine level-guided anti-TB individualized medication. Additionally, studies by other scholars have suggested that TT3 and FT3 levels are associated with poor prognosis in severe pulmonary tuberculosis,
21,22 which echoes our findings and further emphasizes the role of TT3 and FT3 in assessing and managing pulmonary tuberculosis.
The underlying mechanism may be related to the temporal characteristics of acute inflammation in elderly patients: elderly tuberculosis patients experience a peak of acute inflammatory response within one week after admission, with significantly elevated levels of cytokines such as IL-6 and TNF-α. These factors can rapidly inhibit thyroid hormone synthesis and conversion, leading to a synchronous decrease in T3 levels and progression of severe disease during this period,
15,23–26 suggesting that T3 levels can be used to evaluate disease severity.
4.3 Impact of aging on the correlation between TT3 and FT3 levels and severe pulmonary tuberculosis
Given the key role of age in geriatric pharmaceutical care, we analyzed age-specific associations between TT3 and FT3 levels and severe tuberculosis. Stratified analysis showed that negative correlations persisted in both elderly and non-elderly groups, with critical differences for individualized medication: TT3 correlation was slightly weaker in the elderly (r =−0.516 vs. non-elderly r =−0.529), possibly due to age-related metabolic decline affecting TT3 regulation; while FT3 correlation was stronger in the elderly (r =−0.491 vs. non-elderly r =−0.380). This is because elderly patients have reduced thyroid hormone-binding proteins, and FT3 (free form) is unaffected by protein changes—making it a more stable biomarker for thyroxine level-guided anti-TB medication in the elderly.
27,28ROC curve analysis confirmed age-specific parameters for anti-TB medication guidance: For the elderly group, TT3 had an AUC = 0.799(cutoff = 1.165, sensitivity = 0.760, specificity = 0.717), while FT3 had an AUC = 0.782 (cutoff = 3.220, sensitivity = 0.825, specificity = 0.647); non-elderly groups had different cutoffs (TT3:1.335, FT3: 3.380).
29 The lower TT3 AUC in the elderly may be due to comorbidity-induced metabolic interference, while FT3 maintained high sensitivity—supporting FT3 as the preferred biomarker for thyroxine level-guided anti-TB individualized medication in the elderly, with the cutoff value 3.220 directly applicable for clinical dose adjustment. Comparable studies have also confirmed that total triiodothyronine (TT3) levels significantly decrease with age, with a 9% decrease in the 60–79 age group and a 19% decrease in the 80 and older age group compared to younger adults (20–59 years old). In addition, elderly individuals may also have elevated thyroid-stimulating hormone (TSH)levels. These findings indicate that age affects thyroid hormone levels and should be considered when assessing T3 levels and predicting outcomes in elderly populations.
30Therefore, in clinical practice, we need to fully consider the patient's age and specific circumstances, reasonably select cutoff values for TT3 and FT3 levels, and combine them with other clinical indicators and examination results to improve the accuracy and reliability of predicting severe pulmonary tuberculosis.
4.4 Summary and prospects
This study confirms that with severe disease within 1 week of admission as the endpoint, baseline T3 levels can effectively reflect the early risk in elderly patients. This design not only conforms to their disease course characteristics (predominantly acute progression) but also provides a key basis for clinical early intervention and the setting of follow-up duration in subsequent cohort studies.
Due to the 1-week limited follow-up period (short-term outcome), survival analysis models were not adopted, but the results can still reflect early risk associations. Future studies should further explore the impact of age on thyroid hormone metabolism and regulatory mechanisms, as well as how to more accurately apply TT3 and FT3 levels in different age subgroups to assess the severity of tuberculosis in patients. Meanwhile, we will also focus on other factors that may affect thyroid hormone levels, such as nutritional status and inflammatory responses, to more comprehensively understand the role of thyroid hormones in the pathogenesis and progression of tuberculosis.
4.5 Limitations
This study has the following limitations: The 1-week window used in this study may miss a few cases of delayed severe disease. Future cohort studies can adopt a dual-window design of “1 week and 1 month” to retain early warning value while covering delayed risks. Secondly, the limited sample size may affect the stability and representativeness of the results. Furthermore, the age stratification is relatively coarse and fails to reflect differences among other age groups. We will further optimize the research design to address these issues in future studies.
5 Conclusion
Through a cohort study of 509 tuberculosis patients, we confirmed a close correlation between serum TT3 and FT3 levels and disease severity in the elderly. Pearson correlation and ROC curve analyses verified their potential as biomarkers for thyroxine level-guided anti-TB individualized medication. Subgroup analysis showed age-specific utility: FT3 exhibits stronger correlation with severe disease and medication risk in the elderly (r =−0.491), making it more reliable than TT3 for geriatric pharmaceutical care; the elderly-specific FT3 cut-off (3.220 pmol/L) and TT3 cut-off (1.165 nmol/L) may facilitate severity assessment in elderly pulmonary TB patients; future work should evaluate their utility for individualized anti-TB therapy. These findings align with the goals of intelligent pharmacy, offering a practical tool for optimizing elderly anti-TB treatment safety and efficacy.
2026 Publishing services by Elsevier B.V. on behalf of Higher Education Press and KeAi Communications Co. Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).