Cost-minimisation analysis of sivelestat for acute lung injury associated with systemic inflammatory response syndrome

Objectives: To conduct a cost-minimisation analysis of sivelestat sodium hydrate treatment for patients in the intensive care unit (ICU) with acute lung injury (ALI) associated with systemic inflammatory response syndrome (SIRS) caused by infection. Design: The analysis was performed based on data from a phase III randomised, multicentre, double-blind, controlled clinical study of up to 14 days treatment with sivelestat, in which the effect of intravenous sivelestat at a high dose (0.20 mg/kg/h; the sivelestat group) was compared with that at a low dose (0.004 mg/kg/ h, effectively a placebo; the control group). Patients: Patients with ALI associated with SIRS caused by infection, who began their treatment under mechanical ventilation management in the ICU. Methods: A four-stage Markov model was constructed to represent the possible conditions of an ALI patient: ICU plus intubated mechanical ventilation; ICU plus weaned from a mechanical ventilator; admission to the general ward; and death. The base-case analysis used a mechanical ventilator weaning daily rate of 2.9% for the control group and 4.0% for the sivelestat group, and the same mortality (1.2%) for both groups at all stages of the Markov model. Medical costs were estimated from standard fees and Japanese National Health Insurance drug prices included fees for hospitalisation within the ICU and general wards, mechanical ventilation, examinations and drug expenditure. Costs were in 2001 values. Sensitivity analyses were performed by varying the weaning rate, mortality, time between weaning and discharge to the general ward, and drug costs. Perspective: Payers of healthcare costs. Main Outcomes: The expected 30-day medical costs per patient in the control and sivelestat groups were Japanese yen (YEN) 4 144 887 and YEN 3 975 451, respectively; a difference of YEN 169 436. Drug expenditure accounted for more than half of the medical costs for each group. The periods under mechanical ventilation management and in the ICU for the sivelestat group were shorter than those for the control group by 2 and 1.8 days, respectively. This was of significance in the reduction of the medical costs. A sensitivity analysis suggested that the expected costs for the sivelestat group exceeded those for the control group when the daily weaning rate for the sivelestat group was < 3.5%, and also when mortality rates were set at 0.9% in the sivelestat group and 1.4% in the control group. Conclusions: This analysis suggests that from the Japanese healthcare payer perspective, treatment with sivelestat may reduce medical costs compared with standard care for patients with ALI associated with SIRS caused by infection.