Endocrine Abstracts

Study Objective: Sevoflurane is a halogenated inhalational anesthetic increasingly used in the intensive care unit (ICU) for its potentially beneficial effects in patients with acute respiratory distress syndrome. However, safety data on prolonged sevoflurane sedation are lacking and cases of diabetes insipidus (DI) have been observed. The objective of this study was to assess the incidence and risk factors of DI during prolonged sevoflurane sedation.

Design & Setting: A single-center retrospective analysis in the ICU of the University Hospital of Brussels (March 2013 - March 2021).

Patients: One hundred and six (106) ICU patients receiving prolonged sevoflurane sedation were divided into three groups: 1) no DI, 2) suspected for DI (polyuria >40ml/kg/24h and hypernatremia >145 mmol/l but without available urine sampling) and 3) confirmed DI (polyuria and hypernatremia with urine osmolality <300 mOsm/kg [complete DI] or between 300-600 mOsm/kg [partial DI]).

Measurements: The electronic medical records of the University Hospital of Brussels were retrospectively explored. Logistic regression modelling was used to build a predictive model for DI in patients with prolonged sevoflurane sedation.

Main Results: Twenty-two (22/106; 20.8%) patients developed polyuria-hypernatremia. DI was suspected in 13 patients (12.3%) but could not be confirmed due to the absence of urine sampling. DI was confirmed in 9 patients (8.5%), of which 5 were partial and 4 complete. Desmopressin test was performed in 5/9 patients of which none responded, diagnostic for nephrogenic DI. Sevoflurane was administered for a median duration of 96, 155 and 166 hours (p 0.003), at a median end-tidal concentration of 1.11, 1.25 and 1.22% (p 0.09), respectively. The variables identified in the logistic regression model were: higher sevoflurane end-tidal concentration (p 0.02), longer sevoflurane exposure time (p 0.07) and admission for COVID-19 (p 0.16). A cutoff value of >3 for the linear discriminator was predictive for suspected/confirmed DI.

Conclusion: DI occurs frequently during prolonged sevoflurane sedation in the ICU. Intensive care physicians should be vigilant for the development of polyuria-hypernatremia. Monitoring of plasma osmolality, sodium, creatinine, and urine output is advisable for early detection of DI in this setting. This study established a model for DI associated with prolonged sevoflurane sedation, including longer sevoflurane use, higher end-tidal concentrations and COVID-19 infection.