Study objectives: To correlate serum propylene glycol concentration with osmol gap, serum lactate concentration, and amount of propylene glycol administered to mechanically ventilated patients receiving continuous infusions of lorazepam (80% propylene glycol by weight), and to characterize the prevalence of hyperosmolality and range of serum propylene glycol concentrations in this patient population.
Design: Prospective, controlled, observational study.
Setting: Adult surgical and cardiothoracic intensive care units (ICUs) of a 1200-bed, urban, tertiary care, teaching hospital.
Patients: Sixty-four consecutively enrolled intensive care patients requiring mechanical ventilation and pharmacologic sedation.
Intervention: Thirteen patients received continuous infusions of high-dose lorazepam (> or = 6 mg/hr) for a minimum of 36 hours, and 26 received continuous infusions of low-dose lorazepam (2-5.99 mg/hr) for 36 hours. Twenty-five control patients received sedatives that did not contain propylene glycol.
Measurements and main results: Serum propylene glycol and lactate concentrations, osmolality, and basic metabolic profiles were obtained 72-108 hours after ICU admission. Clinical data, drug administration, and severity of illness scores were recorded. Osmol gap and the amount of propylene glycol administered before serum sampling predicted propylene glycol concentrations (r(2)=0.692, p<0.05). Osmol gap alone also predicted serum propylene glycol concentrations (r(2)=0.532, p<0.05). Serum lactate concentrations did not correlate with serum propylene glycol concentrations. Unlike the low-dose and control patients, eight (62%) of 13 high-dose patients had osmol gaps above 10. All 13 high-dose patients had serum propylene glycol concentrations previously associated with toxicity.
Conclusion: Osmol gap can be used as a surrogate marker for serum propylene glycol concentration. In critically ill patients receiving lorazepam for sedation, an osmol gap above 10 was associated with concentrations previously reported to cause toxicity.