Determinants of Poor Outcomes in Acute Poisoning at a Tertiary Care Hospital in an Industrial City of Eastern India: A Retrospective Observational Study

Background: Industrial cities of eastern India combine dense urban populations with proximity to agricultural hinterlands and access to pharmaceutical, household, and agricultural toxic agents, creating a mixed exposure profile. Published data from such settings remain sparse. This study characterised the clinical profile of acute poisoning at an industrial-urban tertiary care centre and evaluated factors associated with escalation of care. Methods: This retrospective observational study included 322 eligible poisoning admissions identified from departmental records at a tertiary care centre in industrial eastern India between January 2023 and March 2025. The prespecified composite endpoint comprised intensive care unit (ICU) admission, mechanical ventilation, vasopressor support, or in-hospital death. Because all patients meeting the composite endpoint were admitted to the ICU, the endpoint was numerically equivalent to ICU admission in this cohort. Multivariable logistic regression and receiver operating characteristic (ROC) analyses were used to evaluate associated factors and assess exploratory discrimination. Results: Median age was 29 years (IQR 21-39; mean 31.6 ± 13.7 years); 182 (56.5%) were female. Suicidal intent was documented in 151 patients (46.9%). Drug overdose (n=100, 31.1%; comprising 99 pharmaceutical overdoses and one alcohol-ingestion case) and corrosive ingestion (n=85, 26.4%) were the leading agent categories. A composite poor outcome was identified in 114 (35.4%) patients. Six patients died, giving an in-hospital mortality of 1.9% (6/322). On multivariable regression, lower Glasgow Coma Scale (GCS) score (odds ratio (OR) 2.73 per unit decrease, 95% confidence interval (CI) 1.64-4.55, p<0.001) and lower peripheral oxygen saturation (SpO₂; OR 1.14 per unit decrease, 95% CI 1.05-1.24, p=0.002) were independently associated with the composite endpoint, which was numerically equivalent to ICU admission in this cohort. The shock index had the largest point-estimate area under the ROC curve (area under the curve (AUC) 0.751, 95% CI 0.693-0.811), followed by heart rate (AUC 0.736) and GCS score (AUC 0.708); no formal pairwise comparison of AUCs was performed. Conclusion: In this single-centre cohort, pharmaceutical overdose and household chemical ingestion were the most frequently recorded poisoning categories, alongside pesticide, rodenticide, and occupational chemical exposures. Lower admission GCS and SpO₂ were associated with the institution-dependent escalation-of-care endpoint. These findings should be considered hypothesis-generating and require validation using objective clinical outcomes in prospective multicentre cohorts.