In this retrospective case series, we found that there is a temporal association between the commencement of quetiapine therapy and the resolution of refractory hyperactive or mixed delirium. Patients had been in a hyperactive or mixed delirium for a median of 15 days prior to quetiapine initiation, but this state resolved within a median of 4 days after the start of quetiapine treatment. Once quetiapine therapy was commenced, patients were weaned rapidly from most other medications over a median of zero to two days, with propofol treatment continued for a median of six days. The patients who were given quetiapine were considered to have refractory mixed or hyperactive delirium and required the attending clinicians to use challenging management strategies. Quetiapine was well tolerated, with transient hypotension, somnolence and asymptomatic QTc prolongation recorded for only four patients. Quetiapine was ceased in 10 patients before they left the ICU and in another 4 patients before they left the hospital. Given that this is an uncontrolled case series, these data cannot provide proof of efficacy. However, it does encourage the development of randomised, controlled trials of quetiapine in this patient population.
The patients who received quetiapine in this case series are recognised as the group at the highest risk of developing hyperactive or mixed delirium which is refractory to treatment [7
]. In particular, this series contained mixed medical and surgical patients who predominantly presented with sepsis or severe respiratory failure and included a high proportion of patients with drug and/or alcohol dependence. A recent randomised, controlled trial in ICUs that demonstrated the benefits of quetiapine therapy in patients with delirium demonstrated resolution of delirium within a median of one day [25
]; however, that study was performed in a more homogeneous patient group with a shorter duration of delirium and fewer risks for refractory delirium than the current case series. It is therefore encouraging that the delirium in the current series resolved soon after the commencement of quetiapine. This could potentially be associated with the administration of quetiapine itself, the concomitant reduction in opioids and other sedatives or by chance alone [7
The quetiapine doses used in our case series were initially lower than those reported in the recent study by Devlin and colleagues [25
], which commenced at a total of 100 mg daily administered in 12-hourly divided doses. The median maximum total daily dose of 50 mg reported in our case series is also lower than those in other studies, in which median maximum total daily doses of 400 mg were used [28
]. The two patients who received a total daily dose greater than 200 mg in our series had a history of alcohol dependence, which is known to require an increased dose of psychotropic medications [6
]. Although the results of the current series are not conclusive, it is encouraging that lower doses could be prescribed, which may provide guidance for future studies.
Few adverse events were noted overall in the case notes of these patients. Problems were predominantly related to somnolence and postural hypotension, which are similar to those described in the literature [20
]. Despite the presence of hypotension, quetiapine treatment was continued as the attending clinician considered that other causes for hypotension were present and the hypotension resolved spontaneously. One patient developed a prolonged QTc interval, which resulted in discontinuation of the drug. This patient was receiving concomitant fluconazole, an inhibitor of the cytochrome P450 isoenzyme CYP3A4 that also metabolises quetiapine. Concurrent use of fluconazole increases quetiapine levels, thus the QTc prolongation could be considered the result of a drug interaction rather than an adverse effect [29
]. No episodes of extrapyramidal side effects were noted; however, an extrapyramidal side effect rating scale is not routinely used in our institution. The retrospective assessment of adverse events can be very difficult, and it is likely that adverse events were underrecognised, especially given that the rate of adverse events was lower than the rates reported in the literature [28
In this case series, the median daily RASS score, use of more than two medications and a validated case note review were used to evaluate hyperactive or mixed delirium. The use of a validated case note review is an accepted method of establishing the clinician's opinion that a state of delirium exists, although considerable expertise is required to undertake it [27
]. In this case series, the validated chart review was undertaken by a junior doctor and was then confirmed by a senior clinical pharmacist in all cases. The RASS is a subjective 11-point scale that has discrete criteria for levels of sedation and agitation. It has been reported to have high reliability and validity for sedation and agitation in medical, surgical, ventilated, nonventilated, sedated and nonsedated ICU patients and is commonly used in the UK [31
]. The RASS score is routinely measured hourly, providing a longitudinal measure of agitation and sedation [33
]. The RASS is not a delirium scoring system, and agitation may be due to a number of reasons (including pain or physiological dysfunction). However, agitation is a common feature of hyperactive or mixed delirium, and the combination of the RASS score with the validated case note review should identify patients in whom delirium is believed to exist by the attending clinician. Specific delirium scoring systems such as the confusion assessment method (CAM)-ICU have a higher reported sensitivity and specificity and higher interrater reliability (κ = 0.96) for the assessment of delirium [34
]. However, our ICU, like many in the UK, does not use CAM-ICU or any other dedicated delirium screening checklist (for example, the intensive care delirium screening checklist (ICDSC)) in routine clinical practice [32
This study comprised a small, uncontrolled, retrospective, single-centre case series and has significant potential limitations. Selection bias is clearly a problem, given that patients received quetiapine on the basis of the treating physician's opinion. The cohort of patients was small, and no control group was used. Although the data suggest a temporal relationship between the initiation of quetiapine therapy and the resolution of delirium, no causal relationship can be established. It is entirely possible that the patients' delirium was resolving at the same time that quetiapine was commenced. Follow-up was incomplete, and no formal neuropsychological assessment was made following discharge from the ICU; hence no comment can be made regarding the longer-term sequelae of delirium and posttraumatic stress disorder. Given the retrospective, uncontrolled nature of the review, other unrecognised sources of bias might exist. All of these methodological weaknesses could be addressed by a suitably powered, prospective, randomised, placebo-controlled trial. Any prospective evaluation of delirium should include validated tools (ICDSC or CAM-ICU) to support the diagnosis of delirium and the possible effects of quetiapine in the treatment of delirium.