Diabetes mellitus is associated with the occurrence of well-described microvascular complications, including retinopathy, nephropathy and peripheral neuropathy. The concept of central neuropathy has been controversial for more than 80 years now. As early as 1922 it was recognised that diabetes can lead to cognitive dysfunction [1]. The prevalence of cognitive dysfunction is difficult to estimate as it depends heavily on the way it is assessed. The reported prevalence is about 40% in long-standing or poorly controlled diabetes [2]. Cognitive dysfunction in diabetes is characterised by lowered performance on several cognitive domains, most notably slowing of mental speed and diminished flexibility [3]. The magnitude of these cognitive deficits appears mild to moderate, but can significantly hamper daily functioning, adversely affecting quality of life [4]. Cognitive decline in diabetic patients treated with insulin has so far largely been attributed to recurrent episodes of hypoglycaemia, rather than to hyperglycaemia, although there is little evidence to support this notion [5, 6]. In trying to describe cognitive impairment in diabetes as a complication of the disease, the term ‘diabetic encephalopathy’ was introduced in 1950 [7]. In 1965, rather characteristic pathological changes were found in brains from 16 long-term juvenile diabetic patients who had died from vascular complications of diabetes [8]: diffuse degenerative abnormalities, pseudocalcinosis, severe angiopathy of cerebral vessels, atrophy of the dentate nucleus, demyelinisation of cranial nerves and fibrosis of the leptomeninges. The authors stated that this histological pattern justifies the term ‘diabetic encephalopathy’, because it differs from that seen in any other clinical condition. However, for several reasons, the term ‘encephalopathy’ has not been widely accepted. Firstly, it has strong negative connotations and does not seem to match the mild cognitive problems usually seen in (non-demented) diabetic patients. Secondly, and most importantly, the diagnosis diabetic encephalopathy lacks clear criteria and is therefore difficult to ascertain. Historically, the term diabetic encephalopathy applies to type 1 diabetic patients only. Other terms found in the literature to describe cognitive dysfunction in diabetes include functional cerebral impairment and central neuropathy. To facilitate research into this area and to increase recognition of the disorder, we propose a new term—‘diabetes-associated cognitive decline’ (DACD). This term is not suggestive of a particular pathogenesis, but merely describes a state of mild to moderate cognitive impairment, in particular psychomotor slowing and reduced mental flexibility, not attributable to other causes. Pathogenic research should focus on type 1 diabetes patients first, as in type 2 diabetes bias is introduced by several coexistent risk factors for cognitive dysfunction other than hyperglycaemia (e.g. hypertension, hyperlipidaemia). For further validation of the term DACD, we propose the operational research criteria summarized in the text box. As the relation between cognitive complaints and objective dysfunction is weak, we do not regard subjective complaints about cognitive functioning as a prerequisite for diagnosis of DACD. As cognitive dysfunction in type 1 diabetes is usually mild to moderate [3], we chose a performance >1.5 SD below that of controls as a criterion for DACD. Future research may refine the currently proposed criteria.

Research criteria for diabetes-associated cognitive decline (DACD)

Lowered performance by a patient with established type 1diabetes mellitus (below − 1.5 SD, compared to age- andeducationally matched control subjects) on validated neuropsychologicaltests in one or more of the following cognitivedomains:

 •intelligence

 •psychomotor efficiency

 •cognitive flexibility

 •visual attention

 •visual perception [3, 9].

2. Cognitive deficits are determined as chronic, i.e. not related toacute hypo- or hyperglycaemia.

3. Usually, but not necessarily accompanied by subjective cognitive complaints.

4. Not meeting criteria for dementia, and excluding other likelycauses e.g. head trauma, intoxication, alcohol abuse, Parkinson’sdisease, multiple sclerosis, stroke and other primary brain diseases.

We hope that by offering this simple descriptive definition, research into the pathogenesis, consequences and potential therapeutic interventions of DACD will be stimulated. In addition, the extent to which diabetes-related cognitive decline affects day-to-day functioning and diabetes self-care needs to be assessed, to further expand our understanding of the clinical implications of this potentially burdensome complication. The reliability of this assessment is fully dependent on the accuracy of the diagnostic tests.