Delirium and Benzodiazepines Associated With Prolonged ICU Stay in Critically Ill Infants and Young Children

Heidi A. B. Smith, MD, MSCI; Maalobeeka Gangopadhyay, MD; Christina M. Goben, MD; Natalie L. Jacobowski, MD; Mary Hamilton Chestnut, NP; Jennifer L. Thompson, MPH; Rameela Chandrasekhar, PhD; Stacey R. Williams, NP; Katherine Griffith, NP; E. Wesley Ely, MD, MPH; D. Catherine Fuchs, MD; Pratik P. Pandharipande, MD, MSCI


Crit Care Med. 2017;45(9):1427-1435. 

In This Article


Three hundred patients were enrolled from March 2013 to October 2014. A total of 320 patients were consented, however, 20 patients did not complete the study due to unavailability of research personnel (12 subjects) or baseline developmental delay (eight subjects). Demographic and summary data are presented in Table 1. We previously reported the prevalence of delirium in this cohort of 44%, with rates of delirium as high as 53% in patients less than or equal to 2 years old and 33% in patients between 2–5 years of age.[5] The median (IQR) duration of delirium was 1 day (1–2 d).

Nonnormal mental status (delirium or coma) had a meaningful association with a lower likelihood of ICU discharge on a given day (p = 0.051) (Table 2). The different HRs presented for delirium and coma further describe their individual associations with likelihood of ICU discharge when compared with having a normal mental status. Since this association was greatly modified by age, the above single p value cannot fully describe the depth of the relationship between mental status and likelihood of ICU discharge. Rather, Figure 1 is necessary to illustrate the significant associations between delirium and lower likelihood of ICU discharge among preschool-aged children (> 12 mo old). Benzodiazepine exposure was also significantly associated with a lower likelihood of ICU discharge (p = 0.01). Other risk factors significantly associated with a lower likelihood of ICU discharge included younger age (p = 0.01) and hypoxia (p < 0.001). In contrast, dexmedetomidine exposure increased the likelihood of ICU discharge (p = 0.008), therefore more likely associated with a shorter LOS. The aforementioned relationships were nonlinear; therefore, the HRs presented in Table 2 are specific for the comparison of the 75th to 25th percentiles of each risk factor for likelihood of ICU discharge.

Figure 1.

Delirium and likelihood of ICU discharge modified by age. Delirium was associated with a lower likelihood of ICU discharge (overall p for mental status = 0.05). This relationship was modified by age (interaction p = 0.10) whereby delirium significantly lowered the likelihood of ICU discharge in preschool-aged patients (>12 mo old). The hazard ratios for likelihood of ICU discharge at 60, 36, and 12 months old were 0.17 (95% CI, 0.05–0.61), 0.50 (0.32–0.80), and 0.98 (0.68–1.41), respectively.

An increase in delirium duration was significantly associated with higher benzodiazepine exposure (p = 0.005). Additionally, younger age (p = 0.005) and higher severity of illness (p = 0.007) were also predictors of longer delirium duration.(Table 3) The aforementioned relationships were nonlinear; therefore, the incidence rate ratios (IRRs) presented in Table 3 are specific for the comparison of the 75th to the 25th percentiles of all risk factors assessed. As an example, when considering benzodiazepine exposure, the 75th and 25th percentiles in our cohort were 0.73 and 0 mg/kg/d, respectively, on the day prior to the first delirium assessment. The IRR of 2.47 (95% CI, 1.36–4.49) indicates that, on average, a patient receiving 0.73 mg/kg/d of benzodiazepine will have about 2.5 times the incidence rate of delirium compared with a patient with no benzodiazepine exposure. Furthermore, as a single-point estimate cannot fully describe nonlinear relationships between risk factors and delirium duration, Supplemental Figure 1 (Supplemental Digital Content 1,; legend, Supplemental Digital Content 2, further illustrates the depth of these significant associations.

Benzodiazepine exposure was the sole independent predictor for the development of delirium the day following exposure compared with having a normal mental status (p = 0.02).(Fig. 2) As an illustration, an odds ratio of 2.83 for this relationship means that patients, with approximately 1 mg/kg of benzodiazepine exposure when compared with 0 mg (the 75th and 25th percentiles benzodiazepine exposure) in a 24-hour period, have 2.8 times the odds of developing delirium the following day versus having a normal mental status. There was no statistically significant relationship between age, opioid exposure, mechanical ventilation, or cardiovascular SOFA score on the development of delirium the following day.

Figure 2.

Benzodiazepine exposure versus mental status the following day. We examined the relationships between potential risk factors and mental status the following day using multinomial regression. Drug dosage amounts (mg/kg) were transformed using the cube root transformation in order to mitigate the influence of extremely high values. After adjusting to the median or mode of all other covariates, higher benzodiazepine doses were significantly associated with a higher likelihood of being delirious compared with having a normal mental status (p = 0.02) the day following the exposure.