Severe prolonged sedation associated with coadministration of protease inhibitors and intravenous midazolam during bronchoscopy.

STUDY OBJECTIVE

To determine whether human immunodeficiency virus (HIV)-positive patients who received intravenous midazolam during an inpatient bronchoscopy procedure were more likely to experience severe prolonged sedation if they were taking antiretroviral therapy that included a protease inhibitor versus those who were not taking any antiretroviral therapy.

DESIGN

Retrospective cohort study.

SETTING

Tertiary care academic medical center.

PATIENTS

Two hundred forty-one HIV-positive adults who received intravenous midazolam while undergoing bronchoscopy between January 1, 2003, and December 31, 2006, were analyzed; 51 patients were taking an antiretroviral regimen that included a protease inhibitor (exposed group), whereas 190 patients were not taking any antiretroviral agents (nonexposed group).

MEASUREMENTS AND MAIN RESULTS

Patient demographics, medication administration records, and bronchoscopy data were collected from electronic databases and patient medical records. The exposed and nonexposed groups had similar demographic characteristics except that patients in the exposed group had lower HIV viral loads and were less likely to have altered mental status or respiratory distress before bronchoscopy. In addition, the exposed group had a higher proportion of males and patients with hepatitis B or C virus coinfection. The incidence of severe prolonged sedation was 9.80% in the exposed group versus 1.58% in the nonexposed group (relative risk [RR] 6.21, 95% confidence interval [CI] 1.53-25.12). Specific protease inhibitors associated with severe prolonged sedation were atazanavir-ritonavir and lopinavir-ritonavir. Length of hospital stay was approximately 3 days longer in the exposed group compared with the nonexposed group.

CONCLUSION

Although the interaction between intravenous midazolam and protease inhibitors is well known, this study was the first systematic evaluation, to our knowledge, of the risk of severe prolonged sedation in a cohort of hospitalized HIV-positive patients. Coadministration of protease inhibitors with intravenous midazolam was associated with severe prolonged sedation as well as increased length of hospital stay. Therefore, concomitant use of these drugs should be closely monitored, or alternative sedatives for procedural sedation should be considered.