Threshold-Dependent Associations of Pretreatment Serum Albumin and Prealbumin with Drug-Induced Liver Injury Risk During Intensive-Phase Antituberculosis Therapy: A Retrospective Cohort Study

Abstract

Background: Anti-tuberculosis drug-induced liver injury (DILI) is a clinically important adverse event during first-line antituberculosis therapy and can lead to treatment interruption, regimen modification, and poor outcomes. Although several clinical risk factors for DILI have been reported, simple pretreatment biomarkers for identifying high-risk patients remain limited. Serum albumin and prealbumin are routinely available indicators of nutritional and inflammatory status and may reflect baseline hepatic vulnerability; however, their combined and potentially threshold-dependent associations with the risk of DILI have not been fully characterized. Objectives: This study aimed to evaluate the associations between pretreatment serum albumin and prealbumin levels and the incidence of DILI during the 8-week intensive phase of first-line antituberculosis therapy. We further sought to explore potential nonlinear dose-response relationships and candidate threshold effects, assess consistency across clinically relevant subgroups, and examine whether these markers could contribute to preliminary risk stratification in a retrospective cohort of adults with active tuberculosis. Methods: This single-center retrospective cohort study enrolled 250 adults with active tuberculosis who initiated standard first-line antituberculosis treatment between January 2021 and June 2025. Baseline serum albumin and prealbumin levels were measured within 14 days before treatment initiation. The primary outcome was the development of DILI during the 8-week intensive phase, defined according to the adapted American Thoracic Society criteria. Associations were evaluated using multivariable logistic regression, restricted cubic splines, Kaplan-Meier survival analysis, and Cox proportional hazards models. Subgroup and interaction analyses were conducted. A preliminary clinical risk score was derived from independent predictors and evaluated internally within the derivation cohort. Results: The cumulative incidence of DILI was 19.2% (48/250). After multivariable adjustment, each 1 g/dL decrease in albumin was associated with an adjusted odds ratio (OR) of 1.85 (95% CI, 1.32 - 2.59; P < 0.001), and each 10 mg/dL decrease in prealbumin was associated with an adjusted OR of 1.42 (95% CI, 1.11 - 1.82; P = 0.005). Restricted cubic spline analysis showed significant nonlinear threshold effects, with DILI risk increasing sharply below approximately 3.5 g/dL for albumin (P-nonlinearity = 0.008) and below approximately 20 mg/dL for prealbumin (P-nonlinearity = 0.022). Patients with albumin < 3.5 g/dL and prealbumin < 20 mg/dL had a substantially higher DILI incidence (44.8%) and an earlier onset. The association was particularly strong among HIV-positive, elderly, and underweight patients (interaction P < 0.10). A preliminary risk score incorporating albumin < 3.5 g/dL, prealbumin < 20 mg/dL, age > 50 years, and baseline ALT > 40 U/L stratified patients into groups with progressively higher observed DILI risk, with corresponding hazard ratios of 2.95 (95% CI, 1.62 - 5.37) and 6.82 (95% CI, 3.78 - 12.31). In the derivation cohort, the apparent C-index was 0.84, and the bootstrap-corrected C-index was 0.81. Conclusions: In this single-center retrospective cohort study, lower pretreatment albumin and prealbumin levels were independently associated with an increased risk of DILI. These routinely available markers may help identify patients at higher observed risk in similar settings; however, the present findings should be regarded as hypothesis-generating. External validation, multicenter calibration, standardized monitoring protocols, and prospective evaluation are needed before any clinical implementation.