Postoperative delirium. Part 2

detection, prevention and treatment

Abstract

To target pharmacological prevention, instruments giving an approximation of an individual patient's risk of developing postoperative delirium are available. In view of the variable clinical presentation, identifying patients in whom prophylaxis has failed (that is, who develop delirium) remains a challenge. Several bedside instruments are available for the routine ward and ICU setting. Several have been shown to have a high specificity and sensitivity when compared with the standard definitions according to DSM-IV-TR and ICD-10. The Confusion Assessment Method (CAM) and a version specifically developed for the intensive care setting (CAM-ICU) have emerged as a standard. However, alternatives allowing grading of the severity of delirium are also available. In many units, the approach to delirium follows a three-step strategy. Initially, non-pharmacological multicomponent strategies are used for primary prevention. As a second step, pharmacological prophylaxis may be added. Perioperative administration of haloperidol has been shown to reduce the severity, but not the incidence, of delirium. Perioperative administration of atypical antipsychotics has been shown to reduce the incidence of delirium in specific groups of patients. In patients with delirium, both symptomatic and causal treatment of delirium need to be considered. So far symptomatic treatment of delirium is primarily based on antipsychotics. Currently, cholinesterase inhibitors cannot be recommended and the data on dexmedetomidine are inconclusive. With the exception of alcohol-withdrawal delirium, there is no role for benzodiazepines in the treatment of delirium. It is unclear whether treating delirium prevents long-term sequelae.

Introduction

The first part of this review discussed the pathophysiology and risk factors for postoperative delirium. This second part will address three clinical aspects: detection, prevention and treatment. Many experts agree that all hospitalised elderly patients should benefit from a non-pharmacological multicomponent strategy aimed at the primary prevention of delirium. Such interventions typically aim to reduce several risk factors, often through nursing interventions. In addition, in some patients, pharmacological prophylaxis, most often low-dose antipsychotics, may be considered. This necessitates use of tools to predict the risk in an individual patient based on his/her predisposing and, possibly, precipitating factors. Regular screening for delirium is performed in many units. However, detection of delirium is notoriously difficult¹ and bedside instruments which translate the criteria used in the gold-standard definitions found in DSM-IV-TR (Diagnostic and Statistical Manual of Mental Disorders, 4th edition text revision) or ICD-10 (International Statistical Classification of Diseases and Related Health Problems, 10th revision) into everyday practice for non-psychiatrists are required. Once delirium has developed, the possible causes need to be identified and, if possible, treated. Nevertheless, as delirium is typically multifactorial, a causal treatment may not be possible or immediately effective. Therefore, in many patients, symptomatic treatment will be required. One clinical approach to delirium in elderly surgical patients is illustrated in Fig. 1.

Fig. 1: no caption available.

Detection

Delirium risk scores in surgical patients

In order to target pharmacological prophylaxis at patients who are at risk of developing postoperative delirium, scores that enable prediction of risk are needed. Several models to identify these patients based on predisposing factors or admission criteria are available (Table 1).^2–6

Table 1: Four delirium risk scores for surgical patients

A score based on the presence of a severe illness, cognitive impairment, impaired vision and a high urea/creatinine ratio was originally developed for elderly medical patients² and later validated in a group of patients undergoing orthopaedic surgery.³ Marcantonio et al.⁴ developed and validated a clinical prediction rule specifically for patients undergoing elective non-cardiac surgery. This score includes age, self-reported alcohol abuse, poor cognitive status, poor functional status, a markedly abnormal preoperative sodium, potassium or glucose level, non-cardiac thoracic surgery and aortic surgery. More recently, in elective orthopaedic patients, the ‘Delirium Elderly At-Risk’ (DEAR) instrument has been proposed.⁷ This model incorporates the following risk factors: age, hearing or visual impairment, impairment in any activity of daily living, a low mini mental state examination score or a previous episode of postoperative delirium, and alcohol or benzodiazepine abuse. If two or more risk factors are present, the risk of delirium is greatly increased. However, this score awaits wider application. Finally, for patients undergoing surgery for hip fractures, a score including age, medication history, cognitive performance, albumin concentration, and haematocrit is available.⁵ For cardiac surgery, a specific predictive score is available.⁶Table 1 summarises these risk scores.

A predictive model based on five precipitating factors – use of physical restraints, bladder catheter, more than three medications added, malnutrition and any iatrogenic event – has been validated in a group of hospitalised patients.⁸ A promising model for the calculation of the risk of delirium in patients admitted to an ICU has been validated recently. This model also incorporates data from the first 24 h in the ICU and will become accessible on the internet soon (personal communication, P. Pickkers).

Although these scores may help in identifying patients at high risk, the interpretation of how high the predicted risk has to be in order to justify pharmacological intervention is left to the clinician. Furthermore, most models are applicable to a group of patients undergoing specific procedures only and it is unclear whether it is possible to use the same scoring systems in patients undergoing different types of procedures.

Screening and diagnosis of delirium

In many patients, delirium is easily recognised due to the severe and obvious disturbances in cognition or attention. However, this is not always the case. In particular, the hypoactive form of delirium is notoriously difficult to diagnose. Non-detection rates as high as 66% have been reported.¹ There are two standard definitions of delirium in the literature: the DSM-IV-TR definition and the ICD-10 definition. The two definitions differ. The DSM-IV-TR definition⁹ is focussed on the neuropsychological key features: a disturbance of consciousness with a reduced ability to focus, sustain or shift attention, a change in cognition or the development of a perceptual disturbance that is not better accounted for by a preexisting, established or evolving dementia. The disturbance must develop over a short period of time, that is hours to days and tend to fluctuate during the course of the day, and there must be evidence from the history, physical examination or laboratory findings that the disturbance is caused by the direct physiological consequences of a general medical condition. The ICD-10 definition is more complex.¹⁰ It defines delirium as ‘an etiologically non-specific organic cerebral syndrome characterised by concurrent disturbances of consciousness and attention, perception, thinking, memory, psychomotor behaviour, emotion, and the sleep-wake schedule. The duration is variable and the degree of severity ranges from mild to very severe.’ The ICD-10 definition also subdivides delirium according to the underlying cause.

In clinical practice, these standard definitions are not often used. They are replaced by a multitude of instruments designed to be used by non-psychiatrist clinicians and nurses. Two reviews identifying more than 20 instruments that are used to diagnose delirium on the ward were recently published.^11,12 However, one of the reviews found that only five of the instruments could be recommended as screening tools in lieu of the gold-standard definitions.¹¹ Sensitivity and specificity for the instruments discussed below are shown in Table 2.^11–15 The Confusion Assessment Method (CAM)¹⁶ is a rapidly administered instrument which is available in many languages and can be performed reliably by nurses and physicians. The Delirium Rating Scale (DRS)¹⁷ is a 10-item rating scale intended to be used by a clinician with psychiatric training. However, one of the shortcomings of this test is its inability to distinguish between the hyperactive and hypoactive subtype of delirium. Therefore, a revised version has been developed (DRS-98-R).¹⁸ The Memorial Delirium Assessment Scale (MDAS) consists of 10 items, of which three measure cognition.¹³ Because this scale does not include some important features of delirium, such as variability of symptoms, caution is needed when it is used to screen for delirium. It has been suggested that this scale is best used to quantify the severity of delirium, once the diagnosis has been made with another instrument.¹¹ The NEECHAM Confusion Scale¹⁹ was developed for a rapid bedside assessment by nurses. This scale has been criticised as it measures acute confusion rather than criterion-defined delirium. The Mini Mental Score Examination should not be used to diagnose delirium as, despite a high sensitivity, the specificity is low.¹² As the two gold-standard definitions are not identical, the calculation of sensitivity and specificity for any instrument will differ depending on which one of the two gold-standard definitions is used.

Table 2: Sensitivity and specificity of frequently used instruments to diagnose delirium

Most of these instruments were not developed for ICU use. Therefore, specific tools have been developed for, or validated in, the ICU environment: the Confusion Assessment Method for the ICU (CAM-ICU),^20,21 the Intensive Care Delirium Screening Checklist (ICDSC)¹⁴ and the Nursing Delirium Screening Scale (Nu-DESC).^15,22 The CAM-ICU is based on the same principles as the CAM; however, it was adapted to be used in intubated patients. In a recent comparison between the CAM-ICU and the Nu-DESC, comparable sensitivities were found, but the specificity of the CAM-ICU was greater than that of the Nu-DESC (Table 1).¹⁵ For the ICDSC, sensitivity and specificity were reported as 99 and 62%, respectively, making it a reliable screening tool.¹⁴

What are the criteria that influence the decision about which score to introduce in a specific unit? Apart from sensitivity and specificity, there are some important differences between these instruments. The CAM and the CAM-ICU are probably the most frequently used worldwide and both have excellent specificity. They are suitable to diagnose delirium but require some training and result in a binary score; no information on severity is available. The NEECHAM scale is very popular among nurses and is easy to use, but it is mainly a screening tool.¹¹ If scoring of severity of symptoms is desired, the DRS-98-R or the MDAS or the ICDSC in the ICU setting will be used. Severity may not only be important for research purposes. It has been shown that subsyndromal delirium, a condition that shows some features of delirium but not enough to meet the full diagnostic criteria, is also associated with adverse outcomes.^23,24 In critical care patients, the question of whether sedated patients can be screened using a specific instrument or should not be screened at all has led to extensive discussions and varying incidences of reported delirium.²⁵

Finally, emergence delirium needs to be distinguished from postoperative delirium. In contrast to postoperative delirium which typically develops after a lucid interval of 24–72 h postoperatively, emergence delirium is seen during or immediately after emergence from general anaesthesia and usually resolves within minutes or hours.²⁶ It occurs in all age groups, but seems to be particularly common in children.²⁷ Two recent prospective observational studies in adults reported an incidence of 5% and both identified preoperative benzodiazepine medication as a risk factor.^28,29

Prevention

Non-pharmacological multicomponent strategies

Delirium is rarely caused by a single factor. Rather multiple predisposing and precipitating factors combine to precipitate an episode of delirium. Therefore, the logical approach is to develop preventive strategies which target several common risk factors. The first landmark study successfully using a multicomponent intervention to prevent delirium in hospitalised medical non-ICU patients was published by Inouye et al. in 1999.³⁰ They randomised patients either to a ward using such a multicomponent strategy or to control wards. Six risk factors were targeted by implementing interventions focusing on orientation and therapeutic activities for cognitive impairment, early mobilisation, non-pharmacologic approaches to minimise the use of psychoactive drugs, prevention of sleep deprivation, communication methods, eyeglasses and hearing aids, and early intervention for volume depletion. With this intervention, they were able to show a reduction of delirium from 15% in controls to 10% in the intervention group. This translates to a number needed to treat of 19 to prevent one case of delirium. Although this may not seem very impressive, this value is comparable with interventions in other fields of medicine that are universally accepted. For instance, in a trial of metoprolol for congestive cardiac failure (MERIT-HF),³¹ the number needed to treat to avoid one all-cause death or hospital admission was 17, and to avoid one hospitalisation for worsening heart failure was 22. Once an initial episode of delirium had occurred, however, the intervention had no effect on the severity of delirium or on the likelihood of recurrence. This underscores the importance of primary prevention of delirium. Interestingly, a follow-up analysis 6 months after discharge of those patients who were still alive was unable to identify a lasting beneficial effect of the multicomponent intervention.³²

A randomised clinical trial in patients with hip fractures confirmed the effectiveness of such a multicomponent strategy.³³ In this trial, a daily geriatric consultation targeted 10 domains: oxygen delivery, fluid and electrolyte balance, pain management, reduction in the use of psychoactive and anticholinergic drugs, bowel and bladder function, nutrition, early mobilisation, prevention of postoperative complications, appropriate environmental stimuli and treatment of symptoms of delirium. This prevented one-third of deliria after hip-fracture repair and reduced the number of severe cases of delirium by more than 50%. However, there was no difference between the intervention and control group regarding the number of hospital days with delirium per episode of delirium. This suggests that this intervention had no impact on the duration of delirium once it developed, again stressing the importance of primary prevention.

A study investigating a nurse-led interdisciplinary intervention programme for delirium in elderly hip fracture patients found no effect on the incidence of delirium, but the duration of delirium was shorter and the severity of delirium was lower in the intervention group.³⁴ Data were collected before and after implementation of a programme consisting of training in recognition of delirium, consultation by specialist nurses, standardised analgesia and, when necessary, review of a psychogeriatrician. No effect on activities of daily living rehabilitation was found and the results about mortality were inconclusive.³⁴

Finally, a study investigating cognitively intact patients with hip fractures extended the intervention to the preadmission phase and included obligatory supplemental oxygen in the ambulance, intravenous fluids and blood transfusion in order to maintain normal circulatory parameters and oxygen delivery, adequate pain treatment, reduction in the use of antiemetics and anticholinergics, extranutrition, and improved transfer logistics.³⁵ Patient groups treated before and after implementation of the protocol were compared. This approach led to a reduction in the number of patients who developed delirium during hospitalisation. No long-term outcomes (beyond 30 days postoperatively) were investigated.

Such multicomponent strategies have not been formally tested in the ICU setting. Nevertheless, this approach is frequently used there.³⁶

Pharmacological prevention

Based on the receptor systems thought to be involved in the development of delirium, several groups of drugs targeting the cholinergic, dopaminergic, serotonergic or noradrenergic system could, theoretically, be used to prevent delirium (Fig. 2). However, there is a lack of controlled trials on this topic, particularly in the postoperative setting.

Fig. 2: no caption available.

Antipsychotics

Kalisvaart et al.³⁷ performed a randomised controlled trial including patients aged 70 and older admitted for acute or elective hip surgery who were at intermediate or high risk for postoperative delirium. They used low-dose oral haloperidol (0.5 mg three times a day) as a prophylactic intervention. They were unable to show an advantage of this regimen compared with placebo as far as the incidence of delirium was concerned. However, the severity and duration of delirium were lower in patients receiving haloperidol. In a randomised controlled trial, patients received either placebo or a single oral dose of risperidone 1 mg upon emergence from anaesthesia after elective cardiac surgery.³⁸ They reported a reduction of delirium from 32 to 11%. However, this difference was entirely due to delirium diagnosed on the day of surgery, making it difficult to interpret these data. More recently, in a large double-blind randomised controlled trial, oral olanzapine 10 mg was shown to be an effective prophylaxis in patients of more than 65 years undergoing elective hip or knee replacement surgery.³⁹ The patients received either placebo or 5 mg of olanzapine both immediately before and after surgery. The effect was highly significant: of the almost 500 patients included in this trial, 40% in the placebo arm and 14% in the olanzapine arm developed delirium. The relative advantages of using atypical rather than typical antipsychotics are discussed in the section below on the use of antipsychotics to treat delirium.

Cholinesterase inhibitors

Three randomised controlled trials aiming to prevent delirium with cholinesterase inhibitors have been published. Two groups performed randomised controlled trials in orthopaedic surgery patients using donepezil to prevent delirium without success.^40,41 A third study using prophylactic rivastigmine in elderly patients undergoing elective cardiac surgery was also negative.⁴² The main problem of all these studies was the relatively small sample size (33, 80 and 120 patients, respectively) and the relatively low incidence of delirium (20, 20 and 30%, respectively). Accordingly, they were underpowered to detect a small treatment effect.

Dexmedetomidine

There is considerable interest in the effects of dexmedetomidine on delirium when this agent is used for sedation in the ICU. The results of studies addressing this question are contradictory. Typically, mixed groups of medical and surgical patients are investigated. In a double-blind randomised controlled trial investigating mechanically ventilated patients, Pandharipande et al.⁴³ found a reduction of the combined end-point of delirium-free and coma-free days when dexmedetomidine was used in comparison with lorazepam. However, the difference in delirium-free days was not significant. These results were confirmed in a subgroup of patients with sepsis.⁴⁴ In another double-blind trial, mechanically ventilated patients were randomised either to dexmedetomidine or standard care consisting of propofol or midazolam. The incidence of delirium was higher in the group receiving dexmedetomidine when analysed as the combined end-point of delirium identified by the CAM-ICU and adverse events of delirium and confusion. More CAM-ICU assessments were performed in the dexmedetomidine group. If only CAM-ICU results were analysed, the proportion of positive test results was comparable between the two groups.⁴⁵ A randomised double-blind trial comparing morphine and dexmedetomidine for postoperative sedation in patients undergoing cardiac surgery with cardiopulmonary bypass found no difference in the incidence of delirium diagnosed with the CAM-ICU. However, the duration of delirium was shorter in patients receiving dexmedetomidine.⁴⁶ In contrast, Riker et al.⁴⁷ performed a randomised controlled trial in mechanically ventilated patients and reported a lower prevalence of delirium during treatment when dexmedetomidine was used for sedation instead of midazolam. An open-label trial in cardiac surgery patients also reported less delirium in patients sedated with dexmedetomidine compared with propofol or midazolam.⁴⁸ A meta-analysis incorporating 24 trials, including those reported above, and almost 2500 critically ill patients found no effect of dexmedetomidine on the incidence of delirium.⁴⁹ Although dexmedetomidine offers some unique properties and may reduce mortality, duration of mechanical ventilation,⁵⁰ and length of stay in intensive care,⁴⁹ its effect on delirium remains unclear.

Other interventions

Sleep deprivation is a well known precipitating factor for delirium⁵¹ and several studies using melatonin to prevent delirium have been performed. However, the results are conflicting and there is insufficient evidence to recommend melatonin for the prevention of delirium.⁵² Aizawa et al.⁵³ tested an unusual approach in 40 patients based on the concept that sleep disorders are an important cause of delirium. They used diazepam, flunitrazepam and pethidine administered in the evening and overnight in half of their patients after surgery for gastric or colon cancer and found a reduction in delirium. However, this regimen caused sedation in 40% of the patients in the treatment arm, which may have interfered with the assessment of delirium.

A small randomised trial found that adding gabapentin to the treatment regimen for postoperative pain reduced the incidence of delirium significantly. The authors attributed this effect to an opioid-sparing effect.⁵⁴ In a small trial of patients undergoing cardiac surgery, ketamine (0.5 mg kg⁻¹) at induction was associated with a reduction in postoperative delirium.⁵⁵ In patients receiving ketamine, reduced C-reactive protein concentrations were found on the first postoperative day and the authors suggested that the protective effect of ketamine may have been due, in part, to an anti-inflammatory action of the drug.

In summary, there are no generally accepted pharmacological strategies to prevent delirium.⁵⁶ However, administration of low-dose haloperidol or olanzapine may be an option in selected patients.

Symptomatic treatment of delirium

Before commencing symptomatic treatment of delirium, it is important to manage any precipitating factors. However, treatment of many precipitating factors, such as antibiotics for sepsis or reducing the nitrogen load in the gut of patients with hepatic encephalopathy, will not have an immediate effect on the symptoms of delirium. Therefore, in agitated patients, symptomatic treatment is often indispensible. It has been suggested that it may be necessary to use different treatment strategies for different forms of delirium. However, there is no evidence to support the notion that, for example, hyperactive delirium needs to be treated differently from hypoactive delirium.⁵⁷

Antipsychotics

No data that specifically address the use of antipsychotics for treatment rather than the prophylaxis of postoperative delirium are available. In general, using antipsychotics for treatment decreases the severity of symptoms by 43–70%,⁵⁸ and it is estimated that 50–100% of patients will respond (decrease the severity of symptoms by at least 50%).⁵⁸ Postoperative delirium typically lasts 1–4 days. Whether antipsychotics reduce the duration of delirium is unclear due to a lack of placebo-controlled studies.⁵⁸

An early randomised controlled trial successfully used haloperidol and chlorpromazine, but not lorazepam, to treat the symptoms of delirium in hospitalised AIDS patients.⁵⁹ The relative lack of anticholinergic and sedative effects is an advantage of haloperidol over chlorpromazine. Intravenous haloperidol is frequently used, particularly in the ICU. The intravenous route may confer some advantage over oral dosing with regard to extrapyramidal side-effects.⁶⁰ However, the US Food and Drug Administration recently highlighted the risks of QT-interval prolongation with intravenous haloperidol administration.⁶¹ At the time of writing, the author understands that the licence for the intravenous use of haloperidol is to be withdrawn worldwide (personal communication, Janssen-Cilag AG, Switzerland).

Comparative studies of various atypical antipsychotics with haloperidol show similar beneficial effects.⁵⁷ A randomised trial of enteral olanzapine and haloperidol found similar improvements in delirium symptoms in ICU patients. Haldoperidol was associated with extrapyramidal side-effects.⁶² However, a prospective open label trial in hospitalised cancer patients found that sedation may be a problem with olanzapine use, particularly in patients with hypoactive delirium.⁶³ The factors associated with a poorer response to olanzapine were age more than 70 years, history of dementia, cerebral metastases, hypoxia, and hypoactive and severe delirium.⁶³ A recent review suggested that differences in anticholinergic activity may confer potential advantages of risperidone over olanzapine.⁶⁴ However, a small randomised trial has not been able to confirm this.⁶⁵ In a randomised double-blind placebo-controlled trial, quetiapine has been reported to be a well tolerated and effective treatment of delirium when added to haloperidol in critically ill patients leading to a faster resolution of delirium.⁶⁶ Systematic reviews have identified reduced extrapyramidal side-effects with atypical agents compared with haloperidol, especially when more than 4.5 mg of haloperidol per day is used.^58,67 However, in 2005, the FDA issued a warning that the treatment of behavioural disorders in elderly patients with dementia with atypical antipsychotic medications is associated with increased mortality.⁶⁸

The current guidelines of the Society of Critical Care Medicine (SCCM) and the American College of Critical Care Medicine (ACCM) for the sustained use of sedatives and analgesics in critically ill adults recommend haloperidol as the preferred agent for the treatment of delirium.⁶⁹ Generally, the doses recommended for the treatment of delirium in elderly patients are low (Table 3).^{59,62,63,65,66,70–72}

Table 3: Recommended doses of some antipsychotics for treatment of delirium

Cholinesterase inhibitors

One randomised double-blind placebo-controlled trial investigated whether adding rivastigmine to haloperidol in ICU patients shortens the duration of delirium. However, the trial was terminated prematurely after the safety monitoring board found a higher mortality among patients receiving rivastigmine than in those receiving placebo, although the difference had not reached ‘statistical significance’.⁷³

Dexmedetomidine

Dexmedetomidine was investigated in a small open label study in a mixed group of surgical and medical ICU patients undergoing mechanical ventilation in whom extubation was not possible solely because of agitation.⁷⁴ Delirium diagnosed by the ICDSC was only present in 50% of the patients. Patients were randomised to receive either haloperidol or dexmedetomidine, and dexmedetomidine shortened the time to extubation and decreased length of stay in the ICU.

Benzodiazepines

Benzodiazepines are consistently associated with the development of delirium.^51,75 A recent systematic review concluded that benzodiazepines cannot be recommended for the control of non-alcohol withdrawal related delirium among hospitalised patients.⁷⁶ Administration of benzodiazepines may be followed by paradoxical reactions manifesting as confusion or agitation. This may occur in 1–10% of patients receiving lorazepam.⁷⁷ Elderly patients, in particular, are at risk of such reactions that may be indistinguishable from delirium.

In contrast, benzodiazepines remain the first-line treatment for alcohol-withdrawal delirium.^78,79 In fact, haloperidol is recommended only as an adjunctive therapy to control symptoms such as hallucinations and combativeness.⁷⁸ Several articles describe the use of various other agents in managing alcohol-withdrawal delirium including carbamazepine and other antiepileptic agents, clonidine, dexamethasone and physostigmine. However, there is insufficient evidence of the effectiveness of these drugs.⁷⁹ The use of ethanol is controversial and currently not recommended.⁸⁰ Patients with alcohol dependence are often thiamine deficient, and it has been reported that patients with alcohol-withdrawal delirium have even more substantial deficiencies. Thiamine deficiency is associated with Wernicke's encephalopathy and Wernicke–Korsakoff syndrome. Thiamine administration has a low risk of adverse effects and can prevent the development of these conditions. In particular, it should be given before the administration of intravenous fluids containing glucose which may precipitate acute thiamine deficiency.⁷⁹

Future directions

Even though there exists a large number of clinical trials investigating delirium, summarising and interpreting these data is difficult due to the often small sample size and the heterogeneity of included patients and methods used. Furthermore, there are some key questions that have not been addressed adequately. For example, does prophylaxis or treatment of delirium change the long-term outcome? So far there are no data to support this concept.^32,34 This question is of great importance, as a recent meta-analysis has confirmed that delirium in elderly patients is associated not only with an increased risk of death but also of dementia and institutionalisation. These associations are independent of age, sex, co-morbid illness or illness severity, and presence of dementia at baseline.⁸¹ Another important question is whether we should treat hypoactive delirium, irrespective of whether treatment has an effect on outcome. Hallucinations are also present in hypoactive delirium and patients with hypoactive delirium have been shown to be just as distressed as patients with hyperactive delirium.⁸² This alone may justify treatment. Finally, sub-syndromal delirium has also been shown to be associated with unfavourable long-term outcome.^23,24 However, the question of whether this justifies treatment or not has not been addressed. Based on the pathophysiological concepts reviewed in the first part of this review and provided the hypothesis that neuroinflammation plays an important role in the development of delirium is correct, one obvious question is whether it is possible to prevent or treat neuroinflammation and delirium by the use of minocycline or other drugs.⁸³

Conclusion

The current clinical approach to delirium consists of a three-step approach. Non-pharmacological multicomponent interventions are followed by pharmacological prophylaxis and, in case of failure, symptomatic pharmacological treatment. It is possible to prevent or at least reduce the duration and severity of delirium with non-pharmacological preventive measures and there is evidence supporting the use of pharmacological prevention in selected groups of patients. Most experts agree that regular screening for delirium is a reasonable option. However, in the ICU setting sedation may have a profound impact on the results of the various instruments used for screening for delirium. Therapeutic administration of antipsychotics is effective in the majority of patients to reduce the severity of delirium. Using our current approach reducing the burden of delirium is feasible. This is an important achievement for the individual patient and for the healthcare system as an effect on resource utilisation has been shown.⁸⁴ However, there are no data showing an impact of either prophylactic or symptomatic therapeutic intervention on the long-term consequences of delirium.

Acknowledgements

L.S. is supported by a grant from the Swiss National Science Foundation (grant number: 32003B_121956). No conflicts of interest declared.