Response

Editor – We would like to thank our colleague for his thoughtful comments and would like to respond point by point.

1. We regret not mentioning the references for the Diagnostic Statistical Manual (DSM) criteria. Although both clinicians and researchers can be expected to be familiar with the formal well defined DSM criteria, it was an error to not put the reference in our reference list. We agree with the fact that acute onset does not mean ‘1 day’, as we adhere to the DSM criteria. Our table did mention ‘acute’ without defining it. We believe that the speed of onset is dependent on the cause of delirium, with postoperative delirium taking around 2 days and sepsis just a few hours.

2. The underdiagnosis of delirium is a frequent problem and might be partly related to the fluctuation of symptoms throughout the day.¹ Missing delirium symptoms could prevent appropriate treatment of the underlying disorder of the patient and could be seen as a medical omission. We agree there is no need for the admission of patients with behavioural and psychological symptoms of dementia (BPSD) to hospital, but this diagnosis is not always easy for a general practitioner (GP) with limited time for observation.

3. Our manuscript aimed to give an overview of delirium by summarising the important aspects and presenting some new insights based on important papers of the recent years. Our review is not exhaustive, and more important highlights have been published recently. We believe the meta-analysis of Witlox has the highest level of evidence on survival and delirium, and we expect delirium researchers of the included studies would have been able to discriminate well between BPSD and delirium.² The confusion assessment method (CAM) is not the ‘gold standard’ test for delirium. The more strict the definition of delirium (according to DSM criteria), the stronger the association with mortality can be expected. This may be an explanation for the lack of association between CAM positive delirium and survival in the Australian cohort.

4. We agree that testing of hearing is important for all diseases that use cognitive testing – not just for delirium, but also dementia and depression. Importantly, hearing loss is also a risk factor for delirium, and this is often underreported. Additionally, there are other important impairments that can influence performance on cognitive functional testing, such as visual impairment and language problems. In general, one could expect that healthcare workers and researchers involved in delirium research take possible impairments into consideration.

• © 2014 Royal College of Physicians