Initial Dosage Optimization of Tacrolimus in Connective Tissue Disease-Associated Interstitial Lung Disease Patients: An Exploratory Study

Abstract

Objective

Tacrolimus has emerged as a promising therapeutic agent for connective tissue disease-associated interstitial lung disease (CTD-ILD) patients. However, its clinical application is substantially hindered by high interindividual pharmacokinetic variability and a narrow therapeutic window, which poses critical challenges particularly for the formulation of the initial dosage regimen.

Methods

A retrospective real-world study was conducted on 22 CTD-ILD patients receiving tacrolimus therapy. Comprehensive clinical data were systematically collected, encompassing demographic characteristics, routine physiological and biochemical parameters, and detailed concomitant medication records. A population pharmacokinetic (PPK) model for tacrolimus in CTD-ILD patients was developed using the nonlinear mixed-effects modeling approach implemented in NONMEM software. Monte Carlo simulation was subsequently performed to recommend the optimal initial dosage regimen of tacrolimus for CTD-ILD patients.

Results

Based on the results of the model, we proposed initial dosing regimens: for CTD-ILD patients, the recommended initial tacrolimus dosage was 0.10 mg/kg across the body weight range of 50–100 kg, which yielded a target attainment probability of 89.1–93.5%, a probability below 5 ng/mL was 6.3–10.9%, a probability above 10 ng/mL was 0–0.2%.

Conclusion

This was an exploratory single-center PPK analysis recommending the initial dosage of tacrolimus for CTD-ILD patients. We proposed the 0.10 mg/kg regimen as a preliminary model-derived starting dose, which warranted prospective validation and therapeutic drug monitoring (TDM). Given the inherent limitations including the relatively small sample size, lack of external validation, further large-scale prospective studies were warranted to confirm these conclusions.

Introduction

Connective tissue diseases (CTDs) constitute a heterogeneous spectrum of immune-mediated systemic inflammatory disorders, with frequent pulmonary involvement as a prominent clinical manifestation.1 Parenchymal lung impairment, particularly interstitial lung disease (ILD), represents a leading driver of morbidity and mortality in the CTD population.2,3 As a recently published meta-analysis encompassing 237 investigations documented by Joy et al, the pooled prevalence of ILD reaches 40% in autoimmune myopathy and systemic sclerosis patients, 17% in Sjögren’s syndrome patients, 11% in rheumatoid arthritis patients, and 6% among individuals diagnosed with systemic lupus erythematosus.4 Therefore, the treatment of CTD-ILD has become a rather challenging issue.

The clinical management of CTD-ILD is a comprehensive strategy aimed at suppressing inflammatory activity, halting fibrotic progression, preserving lung function, and enhancing quality of life. This requires an individualized, multi-disciplinary approach that systematically integrates pharmacotherapy, oxygen supplementation, pulmonary rehabilitation, mechanical ventilation, and, in select advanced cases, lung transplantation.5–15 In terms of pharmacotherapy, it mainly involves using glucocorticoids to suppress acute pulmonary inflammation.16–18 For patients with existing lung structural damage, using anti-fibrotic drugs to slow down the process of lung fibrosis.19,20 Emerging clinical evidence and published literature have substantiated the therapeutic efficacy of nintedanib monotherapy, as well as the tocilizumab/nintedanib combined regimen, in the comprehensive management of CTD-ILD.21,22 Besides, it is important that using immunosuppressants to regulate abnormal immune responses and reduce the attack of autoantibodies on lung tissue, among them, tacrolimus is the first-line immunosuppressant and is commonly used in the treatment of CTD-ILD.23,24 For myositis-associated ILD, tacrolimus serves as a core recommended immunosuppressive agent. In systemic sclerosis-associated ILD and rheumatoid arthritis-associated ILD, it is applied as an adjunctive immunomodulatory regimen in combination with low-dose corticosteroids.

Given the significant pharmacokinetic variability and the narrow therapeutic window, the clinical use of tacrolimus has encountered certain difficulties.25–27 A panel of clinically relevant covariates may have a potential impact on tacrolimus exposure, including CYP3A genotype, hepatic function, hematocrit, albumin, corticosteroid exposure, concomitant antifungal agents, and other CYP3A/P-gp modulators. Usually, therapeutic drug monitoring (TDM) is used to detect the concentrations of tacrolimus to adjust the next dosage of tacrolimus. However, since there is no available tacrolimus concentrations for reference, the initial dosage plan for tacrolimus in CTD-ILD patients is rather challenging. Population pharmacokinetic (PPK) model is a method that analyzes sparse sampling data in the population and combines statistical models to study the metabolic patterns of drugs in the human body. In clinical practice, there have been numerous studies that have comprehensively utilized TDM and PPK to formulate individualized dosage plan for patients.28–30 To date, tacrolimus PPK models have been successfully established in liver transplantation,26 kidney transplantation,25 nephrotic syndrome,31 systemic lupus erythematosus,29 supporting routine individualized dosing in these scenarios. However, due to inherent discrepancies in baseline background immunosuppression regimens, concomitant antifungal prophylaxis, body composition, and underlying organ function status across disease populations, these existing PPK models cannot directly be extrapolated to the unique CTD-ILD patients.

Based on these, the present study aims to develop a PPK model of tacrolimus and provide initial dosage optimization of tacrolimus for CTD-ILD patients.

Methods

Enrolled Patients

This retrospective study enrolled CTD-ILD patients treated with tacrolimus at the Affiliated Hospital of Xuzhou Medical University from August 2022 to September 2025. All included CTD subtypes were diagnosed strictly according to internationally recognized classification criteria. ILD was confirmed via a verification workflow: high-resolution chest CT, standardized pulmonary function testing. Inclusion criteria: 1) confirmed CTD-ILD meeting the above diagnostic standards; 2) received tacrolimus therapy; 3) completed routine tacrolimus TDM. Exclusion criteria: 1) ILD secondary to non-CTD etiologies (idiopathic pulmonary fibrosis, occupational pneumoconiosis, drug-induced lung injury, hypersensitivity pneumonitis); 2) patients with incomplete core clinical records. Extracted data included baseline demographics, physiological and biochemical indicators, concomitant medication profiles, and measured tacrolimus concentrations from TDM records. Tacrolimus concentration measurement was conducted using Emit^® 2000 Tacrolimus Assay (Enzyme-multiplied immunoassay technique, Siemens Healthcare Diagnostics Inc.), whose routine test range covered 2 to 30 ng/mL, and the routine lower limit of quantification (LLOQ) of the assay was 2 ng/mL. This research was approved by the Research Ethics Committee of the Affiliated Hospital of Xuzhou Medical University (No. XYFY2025-KL599-01), where the requirement for written informed consent could be waived since the data were collected retrospectively without patient identifiers. This study adhered to the Declaration of Helsinki.

Modeling

Since all blood samples collected in this study were steady-state trough concentration specimens, data points from the absorption phase were insufficient, which made a reliable estimation of the absorption rate constant Ka impossible. For this reason, we cited previously published PPK studies on tacrolimus conducted in domestic Chinese populations. Apparent oral clearance (CL/F), volume of distribution (V/F), and absorption rate constant (Ka, fixed at 4.48 h−1)32–34 were taken into consideration using non-linear mixed effect modeling (NONMEM, version 7, ICON Development Solutions, Ellicott City, MD, USA) software in the PPK model. Pharmacokinetic parameters and interindividual and intraindividual variabilities were assessed by the FOCE-I method.

The inter-individual variabilities were shown in Formula (1):

(1)

A_i was the individual parameter, TV(A) was the typical parameter, and η_i was the symmetrical distribution.

The random residual variabilities were shown in Formulas (2)-(4):

(2)

(3)

(4)

B_i was the observed concentration, C_i was the individual predicted concentration, and ε₁ and ε₂ were the symmetrical distribution.

The relationship between the parameters and weight was shown in Formula (5):

(5)

D_i was the i-th individual parameter, E_i was the i-th individual weight, E_std was the standard weight of 70 kg, and D_std was the typical parameter. F was the allometric coefficient: 0.75 for the CL/F and 1 for the V/F.35

The continuous or categorical covariates parameters were shown in Formulas (6) and (7), respectively:

(6)

(7)

G_i was the individual parameter, TV(G) was the typical parameter, Y was parameter to be estimated, K_i was the covariate of the i-th individual, K_m was population median for the covariate.

This study employed a two-step method for constructing covariate model. The objective function value (OFV) change was selected as the inclusion criteria, where OFV decrease > 3.84 (P < 0.05) was defined as the inclusion standard, and OFV increase > 6.63 (P < 0.01) was defined as the exclusion standard.

Model Evaluation

The goodness of fit (observations vs population predictions, observations vs individual predictions, conditional weighted residuals (WRES) vs individual predictions, CWRES vs time), individual plots, visual predictive check (VPC) of model, and bootstrap method were used to evaluated the final PPK model of tacrolimus in CTD-ILD patients.

Simulation

The Monte Carlo simulation was used to simulate tacrolimus concentrations in CTD-ILD patients based on different influencing factors. Each set of scenarios simulated 1000 virtual CTD-ILD patients with 6 weight groups (50 kg, 60 kg, 70 kg, 80 kg, 90 kg, and 100 kg) and 8 tacrolimus dosage groups (0.01 mg/kg, 0.02 mg/kg, 0.04 mg/kg, 0.06 mg/kg, 0.08 mg/kg, 0.10 mg/kg, 0.12 mg/kg, 0.14 mg/kg), respectively, where the dosage was split evenly into twice a day. Based on the existing treatment scope for CTD-ILD patients, the therapeutic window range of tacrolimus was 5–10 ng/mL for CTD-ILD patients,23 and 5 ng/mL represented the minimum threshold for effectively inhibiting T cell activation, while 10 ng/mL marked the critical point where the risk of neurotoxicity and nephrotoxicity significantly increased. The probability of reaching the tacrolimus target concentrations was selected as the evaluation criterion.

Results

Patient Information

A total of 22 patients with CTD-ILD, 6 men and 16 women, with ages ranging from 42.00 to 75.00 years old, and weight ranging from 47.50 to 93.00 kg. Among the 22 patients included in this cohort, the CTD subtypes were distributed as follows: idiopathic inflammatory myopathy (n=17), Sjögren syndrome (n=3), and undifferentiated connective tissue disease (n=2). In terms of disease severity, the mean forced vital capacity percentage predicted (FVC % pred) was 72.17 ± 14.40%. Patients were administered oral tacrolimus capsules. A total of 31 tacrolimus concentration values were collected in this study and the proportion below the lower limit of quantification (BLQ) was 3.2%, with 1 to 3 sampling points per patient. For the 3.2% of tacrolimus concentrations (one point, 1.6 ng/mL) that was BLQ, the analytical laboratory provided actual numerical value (rather than simply reporting it as below quantification). Therefore, we included the actual measured value directly in the dataset and treated it as continuous observation in the NONMEM analysis. We believed that retaining the actual measurement, rather than imputing fixed value or excluding the data, provided the most accurate representation of the patient’s drug exposure at this time point. All blood samples were collected prior to the next dosage, ie, the trough concentration values. Table 1 and Table 2 were demographic data of CTD-ILD patients and drug combination, respectively.

Table 1.

Demographic Data of Connective Tissue Disease-Associated Interstitial Lung Disease Patients with Tacrolimus

Characteristic	Mean ± SD	Median (range)
Gender (men/women)	6/16	/
Age (years)	59.29 ± 10.23	55.00 (42.00–75.00)
Weight (kg)	66.07 ± 8.94	66.50 (47.50–93.00)
Albumin (g/L)	40.33 ± 5.04	42.00 (31.00–49.00)
Globulin (g/L)	26.58 ± 5.06	25.80 (18.60–36.20)
Alanine transaminase (IU/L)	32.74 ± 31.83	25.00 (9.00–174.00)
Aspartate transaminase (IU/L)	25.89 ± 15.89	19.00 (11.00–82.00)
Creatinine (μmol/L)	56.33 ± 20.35	53.00 (28.00–121.00)
Urea (mmol/L)	6.55 ± 2.26	6.22 (2.99–10.77)
Total protein (g/L)	67.12 ± 4.90	66.90 (57.00–76.10)
Total cholesterol (mmol/L)	6.49 ± 1.59	6.90 (3.33–8.35)
Triglyceride (mmol/L)	2.44 ± 1.10	2.31 (1.09–5.04)
Direct bilirubin (μmol/L)	3.83 ± 1.17	3.70 (1.90–6.30)
Total bilibrubin (μmol/L)	10.21 ± 4.10	10.20 (3.40–19.60)
Hematocrit (%)	44.07 ± 4.71	43.60 (34.30–53.70)
Hemoglobin (g/L)	144.16 ± 16.07	142.00 (106.00–172.00)
Mean corpuscular hemoglobin (pg)	30.87 ± 1.62	30.60 (28.00–34.30)
Mean corpuscular hemoglobin concentration (g/L)	327.04 ± 8.15	328.00 (309.00–342.00)

Table 2.

Drug Combination in Connective Tissue Disease-Associated Interstitial Lung Disease Patients with Tacrolimus

Drug	Category	N	Drug	Category	N
Bacillus licheniformis granule	0	21	Mycophenolate mofetil capsule	0	20
	1	1		1	2
Caspofungin acetate for injection	0	21	Nintedanib	0	20
	1	1		1	2
Clopidogrel hydrogen sulphate tablet	0	21	Pantoprazole sodium enteric-coated tablet	0	21
	1	1		1	1
Compound sulfamethoxazole tablet	0	10	Posaconazole	0	21
	1	12		1	1
Famotidine for injection	0	21	Prednisone acetate tablet	0	8
	1	1		1	14
Methylprednisolone sodium succinate for injection	0	17	Tigecycline for injection	0	21
	1	5		1	1
Metoprolol succinate extended-release tablet	0	20	Tofacitinib citrate	0	20
	1	2		1	2
Mosapride citrate tablet	0	19
	1	3

Notes: Category, 0: without drug, 1: with drug; N: number of patients.

Modelling

Covariates and drug–drug interaction evaluation process of tacrolimus in CTD-ILD patients was shown in Table S1. Statistically speaking, posaconazole was significant. However, due to a relatively small sample size (one patient with two tacrolimus concentrations in the posaconazole subgroup) and the covariate effect of posaconazole was identified as exploratory and required prospective confirmation. External validation in an independent patient population was required to corroborate this potential association. Thus, although co-administration of posaconazole showed a statistically significant effect during the screening process, it was not retained in the final model. This decision was based on the limited data supporting this estimate (derived from only one patient with two concentrations), ensuring model robustness and avoiding overfitting.

Formulas (8) and (9) were the final tacrolimus PPK model in CTD-ILD patients:

(8)

(9)

Model Evaluation

Figure 1 shows goodness of fit, Figure 2 shows individual plots, Figure 3 shows VPC of model, and Table 3 shows bootstrap (the success rate of the bootstrap procedure was 99.6%, the biases of CL/F, V/F, ω_CL/F, σ₁, σ₂ were −6.90%, −0.78%, −6.87%, −2.02%, 27.64%, respectively), showing tacrolimus PPK model in CTD-ILD patients was credible. In addition, we evaluated inter-individual variability on CL/F and V/F, ultimately only inter-individual variability on CL/F was retained.

Figure 1.

Goodness of fit. (A) Observations vs population predictions. (B) Observations vs individual predictions. (C) Conditional weighted residuals (WRES) vs individual predictions. (D) Conditional WRES vs time. Blue circle: Concentration point. Green dotted line: Trend line.

Figure 2.

Individual plots.

Abbreviations: ID, patient ID number; DV, measured concentration value; IPRED, individual predictive value; PRED, population predictive value.

Figure 3.

Visual predictive check (VPC) of model.

Table 3.

Parameter Estimates of Tacrolimus Final Model and Bootstrap Validation in Connective Tissue Disease-Associated Interstitial Lung Disease Patients

Parameter	Estimate (RSE%)	Bootstrap		Bias (%)
		Median	95% Confidence Interval
CL/F (L/h)	20.3 (12.7)	18.9	[12.2, 24.2]	−6.90
V/F (L)	1920 (35.3)	1905	[990, 6639]	−0.78
Ka (h−1)	4.48 (fixed)	--	--	--
ωCL/F	0.480 (20.8)	0.447	[0.092, 0.617]	−6.87
σ1	0.346 (31.8)	0.339	[0.073, 0.516]	−2.02
σ2	0.322 (100.5)	0.411	[0.010, 1.456]	27.64

Notes: 95% confidential interval was displayed as the 2.5th, 97.5th percentile of bootstrap estimates. CL/F, apparent oral clearance (L/h); V/F, apparent volume of distribution (L); Ka, absorption rate constant (h⁻¹); ω_CL/F, inter-individual variability of CL/F; σ₁, residual variability, proportional error; σ₂, residual variability, additive error. Bias, prediction error, Bias = (Median-Estimate) / Estimate×100%.

Dosage Recommendation

Figure 4 shows the simulated tacrolimus concentrations for CTD-ILD patients. Figure 5 illustrates the probability of achieving the predefined tacrolimus target concentration range in CTD-ILD patients. As summarized in Table 4, for CTD-ILD patients weighing 50–100 kg, the recommended initial tacrolimus dosage was 0.10 mg/kg, which yielded a target attainment probability of 89.1–93.5%, a probability below 5 ng/mL was 6.3–10.9%, a probability above 10 ng/mL was 0–0.2%. The 0.10 mg/kg regimen should be interpreted as a preliminary, model-derived starting dose rather than an established clinical recommendation. Consequently, this dosing strategy required prospective validation and must be implemented in conjunction with rigorous TDM to ensure patient safety.

Figure 4.

Simulated tacrolimus concentrations in CTD-ILD patients. (A) 50 kg CTD-ILD patients. (B) 60 kg CTD-ILD patients. (C) 70 kg CTD-ILD patients. (D) 80 kg CTD-ILD patients. (E) 90 kg CTD-ILD patients. (F) 100 kg CTD-ILD patients. Red dashed lines: 5–10 ng/mL tacrolimus therapeutic window for CTD-ILD patients.

Figure 5.

Probability to achieve the target concentrations of tacrolimus in CTD-ILD patients.

Table 4.

Initial Dosage Recommendation of Tacrolimus in Connective Tissue Disease-Associated Interstitial Lung Disease Patients

Initial Dosage Recommendation of Tacrolimus
Body weight (kg)	Dosage (mg/kg/day)	Probability of achieving the target concentrations 5–10 ng/mL (%)	Probability below 5 ng/mL (%)	Probability above 10 ng/mL (%)
[50–100]	0.10	89.1–93.5	6.3–10.9	0-0.2

Discussion

Tacrolimus, a calcineurin inhibitor, suppresses T lymphocyte activation and IL-2 transcription to attenuate aberrant autoimmune responses, thereby targeting and controlling the pulmonary inflammatory process.36–38 Multiple clinical investigations have validated that tacrolimus combined with glucocorticoids can significantly improve core lung function parameters in ILD patients, slow the progressive course of pulmonary fibrosis, and effectively reduce long-term maintenance glucocorticoid exposure, which further lowers the risk of hormone-related adverse events.23,24,37

However, the clinical utility of tacrolimus is limited by its own characteristics, such as significant pharmacokinetic variability and a narrow therapeutic window.25–27 Subtherapeutic tacrolimus exposure is linked to insufficient therapeutic efficacy, while excessive concentrations carry a high risk of irreversible organ toxicity. Developing individualized dosage regimens for tacrolimus to control its blood concentration within the therapeutic window is a key issue in clinical medication. To date, numerous studies across diverse clinical scenarios, including heart, lung, liver, and kidney transplantation, hematopoietic stem cell transplantation, nephrotic syndrome, myasthenia gravis, and systemic lupus erythematosus, have successfully recommended optimized initial tacrolimus dosages via PPK modeling.29,39–46 In contrast, research focused on initial tacrolimus dosing specifically for the CTD-ILD patient population remains scarce. Against this backdrop, the present study constructs a PPK model of tacrolimus for patients with CTD-ILD based on TDM and real-world clinical data.

This study included a total of 22 patients with CTD-ILD, 6 men and 16 women, with ages ranging from 42.00 to 75.00 years old, and weight ranging from 47.50 to 93.00 kg. The involved concomitant medications included bacillus licheniformis granule, caspofungin acetate for injection, clopidogrel hydrogen sulphate tablet, compound sulfamethoxazole tablet, famotidine for injection, methylprednisolone sodium succinate for injection, metoprolol succinate extended-release tablet, mosapride citrate tablet, mycophenolate mofetil capsule, nintedanib, pantoprazole sodium enteric-coated tablet, posaconazole, prednisone acetate tablet, tigecycline for injection, tofacitinib citrate.

Tacrolimus was mainly metabolized by CYP3A4 and CYP3A5 enzymes in the liver.47–53 Although posaconazole was a known potent CYP3A4 inhibitor and demonstrated a statistically significant effect on tacrolimus clearance in our initial screening, it was not included in the final PPK model. This decision was driven by the sparsity of the data; the estimated effect was derived from only a single patient with two concentration measurements. In PPK modeling, covariate effects supported by such limited sample sizes were often statistically unstable and may represent individual-specific variation rather than a true population trend. Hence, the covariate effect of posaconazole was identified as exploratory and required prospective confirmation.

The typical CL/F of tacrolimus in CTD-ILD patients was 20.3 L/h. This value was comparable to established estimates in solid organ transplant populations (lung: 13.1 L/h;34 kidney: 20.7 L/h;54 liver: 37.6 L/h55) and fell within the typical range for transplant patients. Weight was fixed using the allometric scaling method, which has been reported in many previous studies.34,56

Regarding the handling of the concentration value of 1.6 ng/mL (below the LLOQ of 2 ng/mL), we chose to retain the actual observed value in the model. We believed that, in the context of our specific dataset, this approach offered distinct advantages over the alternative standard methods (deletion, imputation), as detailed below. Compared to the deletion method, the primary disadvantage of deleting the observation was the loss of information. Given that our dataset was relatively small (31 concentrations from 22 patients), excluding even a single data point resulted in a non-negligible loss of pharmacokinetic information. More importantly, deleting a low trough concentration masked the fact that the patient had a very low exposure. Retaining the 1.6 ng/mL value ensured that the model captured this low exposure characteristic, whereas deleting it might bias the clearance estimates by suggesting the drug was not cleared as efficiently. Compared to simple imputation method (setting to LLOQ/2), imputation method introduced artificial error. For instance, replacing it with 1 ng/mL (LLOQ/2) introduced an approximate 37.5% error. Using the actual reported value of 1.6 ng/mL was the most accurate approximation of the true concentration available to us, minimizing the measurement error compared to arbitrary imputation rules. In summary, directly using the 1.6 ng/mL value optimized the use of available real-world evidence by preserving the quantitative signal, avoiding artificial imputation bias, and maintaining model stability in a sparse dataset.

Further, our analysis confirmed that this approach did not materially affect the model estimates compared to the more complex methods (deletion method or simple imputation method). More importantly, the model parameter estimates obtained by retaining the actual value of 1.6 ng/mL fell between those derived from the deletion method and the simple imputation method, suggesting that this approach was closer to the true value and more credible. As shown in Table S2, the CL/F from the deletion method, the retention of 1.6 ng/mL, and the simple imputation method were 19.8 L/h, 20.3 L/h, and 20.5 L/h, respectively. These negligible differences indicated that the handling of this observation had no material impact on the model estimation results.

It was important to note the limitations regarding the estimation of V/F. Due to the nature of the study design, which relied on sparse trough samples, the data provided limited information to precisely characterize the V/F. The bootstrap analysis revealed a wide 95% confidence interval for V/F [990, 6639], indicating substantial uncertainty in this specific parameter. However, since the primary objective of this model was to optimize dosing regimens based on trough concentrations, which were predominantly driven by CL/F, the uncertainty in V/F had a limited impact on the model’s performance for dosage adjustment. We retained V/F as an estimated parameter rather than fixing it to a literature value to avoid introducing external bias. We believed this approach honestly reflected the data’s limitations while maintaining the model’s utility for its intended clinical purpose.

Monte Carlo simulation was a numerical calculation technique based on random sampling and statistical analysis, which was widely used in the formulation of individualized dosage regimens based on PPK.57–60 Based on our established PPK model, Monte Carlo simulations were employed to propose initial dosages for tacrolimus in CTD-ILD patients. The 0.10 mg/kg regimen should be interpreted as a preliminary, model-derived starting dose. Consequently, this dosing strategy required prospective validation and must be implemented in conjunction with TDM to ensure patient safety.

This study had several limitations. First, it was a single-center, retrospective design with a relatively small sample size and the covariate effect of posaconazole was identified as exploratory and required prospective confirmation. Second, real-world clinical data did not include information on genetic polymorphisms, which were known to influence tacrolimus metabolism. While incorporating pharmacogenetic data could enhance model precision, routine genotyping was not yet standard in clinical practice, potentially affecting the model’s immediate translational utility. The developed PPK model, grounded in real-world treatment data, sought to balance scientific rigor with clinical feasibility for tacrolimus dosage individualization. Additionally, the analysis was confined to pharmacokinetic parameters and did not evaluate clinical efficacy or safety outcomes. To facilitate translation into clinical practice, future steps should include prospective external validation in larger, multi-center cohorts, and systematic assessment of the clinical outcomes associated with the proposed dosing regimens.

Conclusion

This was an exploratory single-center PPK analysis recommending the initial dosage of tacrolimus for CTD-ILD patients. Based on model simulations, an initial tacrolimus dosage of 0.10 mg/kg was recommended for CTD-ILD patients weighing 50–100 kg. It was critical to emphasize that the 0.10 mg/kg regimen represented a preliminary starting dose derived from the current population model. Future prospective studies were required to validate this approach, and TDM remained essential for any clinical application.

Funding Statement

This work was supported by The Basic Science (Natural Science) Project of Higher Education Institutions in Jiangsu Province (No. 25KJD310004), The Suzhou Applied Basic Research Science and Technology Innovation Project (No. SYWD2024258), Jiangsu Provincial Pharmaceutical Association-Aosaikang Hospital Pharmacy Scientific Research Fund (No. A202319).

Data Sharing Statement

The datasets used and/or analyzed during the current study are available from the corresponding author (Dongdong Wang) on reasonable request.

Disclosure

The authors report no conflicts of interest in this work.