In Colombia, paraquat poisoning (PQ) is one of the main causes of self-inflicted pesticide poisoning. There is not antidotal treatment against its toxicity, influencing an uncertain prognosis and changeful survival. Objective: To describe the prognosis of patients with PQ poisoning treated in a high-level hospital and to evaluate the association of mortality with clinical and laboratory test variables, considering its variation over time. Methods: Retrospective study between 2011 and 2018 at the Hospital San Vicente Fundación (Medellín, Colombia). Included cases diagnosed with PQ poisoning, of any age and sex. Diagnoses of concomitant intake of other toxins, underlying lung disease, and trauma were excluded. Sociodemographic and clinical characteristics were described, the effect on mortality and variation over time were analyzed, to search for prognostic research candidates. Results: Cohort of 67 patients, mostly single men, under 30 years of age, located in rural area was recruited. A mortality of 37.3% (95% CI 36.9 to 37.7%) was found, median time to death of 117 h (IQR 62h-238h). The instantaneous risk of dying for the male sex was 1.93; 95% CI 1.44 to 10.26; p = 0.007. Laboratory changes over time: AST (HR = 1.009; 95% CI; 1.007-1.012; p≤0.001), ALT (HR = 1.008; 95% CI 1.006-1.01; p≤0.001), total bilirubin (HR = 1.23; 95% CI 1,173-1.29; p≤0.001), creatinine (HR = 1.53; 95% CI 1,398-1,667; p≤0.001). Conclusion: Factors associated with mortality in PQ poisoning: ingested dose, intentional intake, male sex, and age. The change over time of AST, ALT, creatinine, and total bilirubin is related to death. Design dynamic models with clinical and laboratory test variables that include their evolution during the hospital stay can optimize the prognosis prediction.