Incidence of Delirium and Risk Factors: A Prospective, Observational, Unicentric Study

Abstract

Purpose: To determine the Incidence of delirium over a 10 month observation period and investigate its association with age, alcohol consumption, severity of illness on admission, sepsis, time spent on ventilator and use of continuous midazolam, fentanil or both. Method(s): Prospective observational study. All patients admitted to the 10 bed mixed academic adult ICU during the observation period were evaluated. Inclusion criteria were age over 18 years and ICU hospitalization for at least 48h. Exclusion criteria were neurological diagnosis with GCS <13; previous psychiatric illness; severe hemodynamic instability or MODS; and terminal illness. Patiens or surrogates were asked for consent before data collection. After enrollment, the CAMICU was perfomed daily untill ICU discharge or death. We used SAPS II to assess severity of illness. Age, type and dose of sedative employed, the presence of sepsis, and ventilator days were evaluated daily, and information about alcohol consumption was obtained with the patient's family. Nominal data are presented as frequencies, and continuous variables as means plus SDs. Normality of distribution was assessed with onesample KolmogorovSmirnof test. Comparisons between categorical variables was made with the Chi-squared test, and between continuous variables with the Mann-Whitney U test. Result(s): A total of 445 patients were evaluated for the study. Of these, 291 were excluded and 17 did not give consent. A total of 147 patients compose the final sample. Delirium was observed in 54 patients (36,7%). Alcohol abuse was more common in patients with delirium (57,1% vs 32,1%, p=0,004), as was sepsis (68,9% vs 14,2%, p<0,001) and use of continuous fentanil, midazolam or both (61,7% vs 25%, p<0,001). Patients with delirium also were sicker with mean SAPS II scores of 54,2+-18,3 vs 33+-13,5 (p<0,001) and spent more time on ventilator (mean 5,4+-8,2 vs 1,8+-7,1, p<0,001). We did not find any difference between age (mean 52,9+-12,3 vs 63,7+-15,5, p=0,102) according to the endpoint. Conclusion(s): We found delirium to be common, with incidence comparable with reports of other brazilian as well as foreign ICUs. Alcohol abuse, sepsis and deep sedation with fentanil and/or midazolam were more common in patients who developed delirium. These patients also were sicker as defined by SAPS II scores and spent more time on ventilatory support. Clinical Implications: Improve knowledge about delirium in the ICU and identify targets for improvement.