The patient you spoke to and communicated with yesterday on ICU rounds is a little more difficult to rouse this morning. When you approach him, he initially looks surprised and frightened; as you introduce yourself and remind him of yesterday’s conversation, he appears uncertain of who you are and of his surroundings. As you continue speaking, he looks away and appears to stare at something on the ceiling. The nurse reports that a few hours earlier he was grasping at what appeared to be imaginary objects in front of him. He did not sleep last night.
Delirium, with its traditional symptoms described here, frightens patients. The delusions and hallucinations they experience transiently and usually for the first time in their lives are upsetting; paranoid thoughts are common. When cognitive function normalizes in the fluctuating course that characterizes delirium, patients wonder if they are losing their minds and if they will ever return to the way they were before. Because 40% of patients present symptoms only between midnight and 8 am, careful attention to the nurse’s report of a patient’s behavior is helpful in recognizing the problem.
A diagnosis of delirium is based on the reference standard of symptoms described in nonventilated patients deemed clinically delirious, mostly among geriatric patients where the syndrome is prevalent. The DSM-IV criteria require a fluctuating disturbance in cognition and in consciousness in the context of an acute medical illness. Conversely the World Health Organization’s ICD-1015 includes different criteria for clinical and research use, both of which include a broad range of features that capture the phenomenologic complexity of the syndrome. Comparison of these two sets of criteria in a group of 425 nonventilated elderly medical or nursing home patients showed that applying the DSM-IV criteria identified more (n = 106) patients as having delirium than did the ICD-10 criteria, with twenty-five patients considered delirious by both sets of diagnostic criteria (Fig. 54-2).16 Surprisingly, this study reported similar outcomes, in terms of mortality or length of stay, regardless of the criteria applied to diagnose delirium (Table 54-1).
Comparison of different screening tools in the same population of 425 nonmechanically ventilated elderly medical or nursing home patients using DSM-IV criteria and earlier DSM criteria (DSM-III) as well as ICD-10 criteria. More (n = 106) patients were identified as having delirium with the DSM-IV than with any other set of diagnostic criteria. (Adapted, with permission, from Laurila JV, Pitkala KH, Strandberg TE, Tilvis RS. Impact of different diagnostic criteria on prognosis of delirium: a prospective study. Dement Geriatr Cogn Disord. 2004;18:240–244.)
Table 54-1: Outcomes (Mortality or Length Of Stay) in Relationship to the Criteria (DSM IV, DSM III, and ICD-10) Used to Diagnose Delirium in 425 Nonmechanically Ventilated Elderly Medical or Nursing Home Patients (Outcomes Were Similar) |Favorite Table|Download (.pdf)
Table 54-1: Outcomes (Mortality or Length Of Stay) in Relationship to the Criteria (DSM IV, DSM III, and ICD-10) Used to Diagnose Delirium in 425 Nonmechanically Ventilated Elderly Medical or Nursing Home Patients (Outcomes Were Similar)
|Prognostic Variable||DSM-IV (n = 106)||DSM-III (n = 80)||ICD-10 (n = 43)||p-valuea|
|Mortality/1 year, %||34.9||36.3||41.9||n.s.|
|Mortality/2 year, %||58.5||62.5||65.1||n.s.|
|Mean days in institutions/1 yearb||220(150)||230(148)||211(157)||n.s.|
|New admissions to permanent institutional care/2 year, %||55.0[33/60]||60.0[24/40]||64.3[9/14]||n.s.|
|In permanent institutional care or deceased/2 year, %||90.6||93.8||97.7||n.s.|
Thorough descriptions of the natural history of delirium in ventilated patients are lacking, as is a good understanding of its basic scientific and biologic mechanisms. In 2001, two groups reported independently validated delirium screening scales for use in ventilated critically ill patients.3,4 The first, the Intensive Care Delirium Screening Checklist (ICDSC), is an eight-point item scale, where four points or more correspond to a clinical diagnosis of delirium. The scale was validated against the clinical bedside opinion of a psychiatrist. The advantage of using this eight-item scale with a four-point cutoff is its ability to detect subsyndromal delirium in patients with scores of more than 0 but less than 4 items. Subsyndromal delirium is considered clinically significant by psychiatrists17,18 and, in ventilated and nonventilated ICU patients, predicts an intermediate risk of prolonged length of stay and mortality when compared with asymptomatic patients,19 who did better, or delirious patients, whose prognosis was the worst. The second delirium scale was originally adapted as a simplification of the DSM-IV–based Confusion Assessment Method (CAM) scale, so as to make it applicable to ventilated patients; it was validated against the clinical opinion of a geriatrician. This modified scale, the Confusion Assessment Method in Intensive Care Units (CAM-ICU),2 is binary. Other scales, such as the Delirium Detection Scale (DDS) and the Nursing Delirium Screening Scale (NuDESC), have also been assessed in critically ill patients; only the ICDSC and the CAM-ICU are presented here because of their broad application to and validation in ventilated patients, and because of their psychometric quality.
Clinicians are reported to underrecognize delirium.20,21 This has led some critical care professional societies to promote routine critical care delirium screening.22 Canadian governing bodies mandate it for hospital accreditation. Perusal of the literature on screening tools, however, shows that these validated screening tools yield a broad distribution of incidences of this condition. Studies describing the binary CAM-ICU in similar populations reveal delirium incidences that range from 10%23 to over 80%. Sedation may confound the scale measurements.24 In contrast, the range in reported delirium rates using the ICDSC is somewhat narrower: 32%25 to 45%.21 In addition to grading cognitive normalcy, subsyndromal delirium and delirium the ICDSC describes specific symptoms,26 and with them, the prognosis conferred by each one. Whether these two tools screen for the same constellation of symptoms is unclear, given the conflicting results in the publications on the subject.27,28 Because of the numerous methodological issues described above, comparisons of sensitivity and specificity of delirium screening scales to mechanically ventilated patients in and outside the ICU are difficult to address.
Awareness as to which patient is more likely to develop delirium in the ICU will identify patients most likely to benefit from preventive strategies and alert caregivers as to which patients will require more thorough evaluation for delirium symptoms. These risk factors for delirium cannot identify patients who will respond to therapeutic intervention. Different clinical features have been associated with a higher incidence of delirium.29 The incremental likelihood of delirium symptoms increases with severity of illness and with excessive alcohol consumption.29 Hypertensive patients are more likely to become delirious in the ICU,25 as are patients with preexisting dementia and patients who are heavily sedated.25,29 In contrast to patients admitted to a hospital ward, age is not consistently related to the likelihood of developing delirium among ventilated patients,25,30 although there may be an association with age in mostly nonventilated patients older than 65 years when the CAM-ICU delirium detection tool is used.31,32
There are significant gaps in our knowledge regarding delirium in ventilated patients. Despite reports that mechanical ventilation confers a risk for delirium,33 the question of whether the association is with mechanical ventilation per se or an independent feature linked to severity of illness or drugs administered has not been addressed in studies where known risk factors are taken into account. Cardiovascular patients, who are burdened with cognitive dysfunction after cardiac surgery and ICU discharge, remain understudied, as do neurologic patients, and trauma patients (including those with traumatic brain injury), perhaps because of the inherent difficulties in making an assessment as a result of confounding clinical features.
Early physiotherapy and mobilization, when implemented to aid myopathy, significantly reduce delirium rates.34,35 These results raise the question of whether nonpharmacologic interventions, which encourage patients to focus on their autonomy, may prevent or alleviate delirium. Sedatives and analgesics, when carefully titrated and administered according to symptoms, are associated with lower rates of subsyndromal delirium and an increase in the probability of a patient being able to return home.36 Whether the particular type of sedative agent makes any difference to the probability of delirium is not clear. Once delirium develops, however, continuous sedation with dexmedetomidine is associated with lower duration of delirium than continuous intravenous sedation with midazolam, a short-acting benzodiazepine,37 in medical and surgical patients. It is not clear whether these results reflect an inherent problem with midazolam infusions or a therapeutic benefit of dexmedetomidine.38 In cardiac surgery patients, the use of propofol as a sedative did not alter the risk of delirium as compared with dexmedetomidine.39
Although administration of antipsychotics has been a mainstay in the pharmacologic management of delirium in critically ill patients,40 there is no scientific basis for this practice or evidence of benefit of administering antipsychotic agents in delirious critically ill ventilated patients41: neither the duration of delirium nor its severity is reduced. A possible exception is the atypical antipsychotic, quetiapine,42 which, in contrast to other pharmacologic agents,8,41 produced a reduction in the duration of ICU delirium, albeit in a single pilot study.
Descriptions by observers, and stories43 told by individual patients,44 describe the anguish, fear, and harrowing nature of patient perceptions in a delirious state. Some narratives and recall studies describe the positive impact of reassuring or reality-orienting caregivers45 within the critical care setting. The added contribution made by visiting family members and loved ones29 are in keeping with the positive impact of nurse-facilitated family participation in the care of a delirious patient.46
Some investigators have made links between delirium in the ICU and long-term cognitive dysfunction.47 These data should be considered with caution given the frequency of cognitive dysfunction in nondelirious patients,48 given our difficulties in establishing the diagnosis of delirium, and given its many potential confounders, such as the risk of dementia in patients older than 65 years, which doubles every 5 years between the ages of 65 and 85 years.