Associations of serum glutathione S-transferase pi 1 with NYHA functional class and emergency department disposition in decompensated heart failure: a prospective observational study

Background Glutathione S-transferase pi 1 (GSTP-1) participates in glutathione-dependent detoxification and stress signalling, but its relationship with pre-presentation functional limitation in decompensated heart failure is uncertain. We evaluated serum GSTP-1 across New York Heart Association (NYHA) functional classes and, secondarily, its association with emergency department disposition. Methods This single-centre prospective observational study enrolled 120 adults presenting with decompensated heart failure, either de novo or as worsening of pre-existing heart failure, using a prespecified quota of 30 participants per NYHA class. Eligible consenting patients were enrolled consecutively while the relevant class quota remained open. Eligibility required a compatible decompensated clinical syndrome and objective evidence of congestion. B-type natriuretic peptide, posteroanterior chest radiography, and transthoracic echocardiography informed diagnosis; echocardiography was performed during the index presentation in all 120 patients. Serum GSTP-1 was measured with a research-use-only ELISA. The NYHA comparison was prespecified; admission analyses in 119 patients with known disposition were post hoc and exploratory, using a selected four-variable reference model rather than a validated clinical prediction model. Results GSTP-1 increased stepwise from NYHA I to IV (26.70, 62.44, 85.98, and 126.25 ng/mL; overall p < 0.001). Disposition was available for 119 patients: 36 were discharged and 83 were admitted. GSTP-1 also increased across discharge, ward admission, and intensive care admission (51.10, 66.18, and 94.65 ng/mL; overall p < 0.001). In the exploratory selected-variable model, each 10-ng/mL increase in GSTP-1 was associated with higher odds of admission (adjusted odds ratio 1.36, 95% confidence interval 1.14–1.63; p < 0.001), and adding GSTP-1 improved nested-model fit (likelihood-ratio χ² = 14.47, p < 0.001). NYHA-adjusted estimates were considered consistency checks only because quota construction and sparse within-class outcomes precluded separation of residual confounding. Conclusions Serum GSTP-1 showed a graded association with pre-presentation NYHA functional class and was associated with emergency department disposition in this quota-constructed cohort. The admission findings provide hypothesis-generating, cohort-conditional statistical information; they do not establish haemodynamic severity of the index decompensation, clinical prediction, or decision-making utility. Confirmation in unselected multicentre cohorts using clinically validated assays and serial sampling is needed.