Delirium: a guide for the general physician

Editor – I offer several comments on the interesting article from Todd and Teale.¹ Circular reasoning about dementia is the rule rather than the exception in the delirium literature. For example, Todd and Teale state that dementia is the most significant risk factor for delirium. In fact, the opposite is equally common: dementia is falsely counted as delirium. This occurs because 1) the clinicians have not spent at least 30–45 minutes with family members and other informants searching the time course of present symptoms and the behavioural antecedents; 2) clinicians have not performed domain-specific cognitive tests such as digit span forward for inattention. For example, an 83-year-old man with dementia developed sudden agitation, physical and verbal aggression. Based on sudden onset of confusion, agitation and aggression, he was mislabelled as delirium. A more thorough history revealed that the aggression came after he was treated like an imbecile in a specialist clinic. Everyone talked over him as if he did not exist. Aggression stopped on arrival in the emergency department and did not recur. It would be highly unusual for delirium to clear in two hours so behavioural and psychological symptoms of dementia (BPSD) were the problem rather than delirium. Fluctuating course with altered level of consciousness occur in diffuse Lewy body dementia or diffuse Lewy body mild cognitive impairment.

The authors do not highlight the drawbacks in the 4AT in dementia. Months of the year backwards (MOYB) is usually normal in mild cognitive impairment but abnormal in moderate or severe dementia. 4AT also yields false positives after sedatives or antipsychotics in the emergency department or ward.

The authors assert that a confused patient has delirium until proven otherwise. This assumption can lead to unnecessary investigations and hospital admission.

In Box 4, the authors suggest arterial blood gases. Arterial puncture is difficult in agitated or suspicious patients so a venous PCO₂ and finger oximetry will be more patient-friendly. Box 4 omits measuring serum drug levels (eg lithium) for those on psychoactive drugs, troponin and occasionally creatinine kinase if neuroleptic malignant syndrome is suspected.

The Central Coast Australia Delirium Intervention Study² (CADIS, ClinicalTrials.gov NCT01650896) demonstrated a more specific method to identify delirium by adding three features to the Confusion Assessment Method:

1. meticulous exclusion of BPSD by interviews with family, carers and residential care members

2. not using disorganised thinking, fluctuating cognition or impaired level of consciousness for delirium diagnosis when these may have been caused by antipsychotics or sedatives since symptom onset

3. at least a 25% recent decline in attention of 5- and/or 6-digit span forward using imputed values for baselines.

This CADIS phenotype produced 116 delirious patients with extremely rapid onset and almost as rapid reversal of neurocognitive impairments.

I have recently shown that in 405 medical articles on delirium (ie not intensive care or postoperative) with 789,709 patients,³ only 2.7% of articles reported mean onset in days and only 6.2% reported the mean or median days to recovery. In other words, the bulk of delirium articles describe an event of unproven times of onset or recovery. The CADIS phenotype deserves further study and analysis.

Conflicts of interest

The author has no conflicts of interest to declare.