The Mini-Mental State Examination (MMSE) first appeared in 1975 having been developed as a screening tool for patients with dementia and psychiatric disorders in an inpatient setting.1 Folstein and colleagues did not envisage the global domination that this simple clinical tool would achieve. They clearly state in their seminal paper that ‘the MMS cannot be expected to replace a complete clinical appraisal in reaching a final diagnosis in any individual patient’. But, as frequently happens, this major caveat was quickly forgotten, and, over the course of the next 40 years, this brief 30-item examination gained immense fame among clinicians as a quick and easy method to diagnose dementia. The MMSE is strongly influenced by non-cognitive domains; it does not reliably translate across cultures, as the results are likely to be confounded by language, levels of literacy, and cultural and ethical norms. Despite this, the MMSE has been translated into numerous languages including Cantonese, Arabic, Spanish and Persian.2 ,3 Furthermore, it is unreliable for patients with less than 5 years of education, probably because of the heavy weighting on language and mathematical abilities. A quick search through the 120 000 references related to the MMSE on PubMed quickly reveals these pitfalls, yet it continues to be the cornerstone of dementia diagnosis in many clinical rooms throughout the world.

The MMSE has a role in diagnosing ‘barn door’, well-established dementia and in monitoring patients with moderate dementia, but many clinicians are still unaware of its limitations, even in a Western culture. It has been repeatedly shown that the MMSE is poor at identifying early-stage dementia, especially mild cognitive impairment: such a patient may have severe amnesia and yet score in the normal range on the MMSE. Other worrying features are the lack of components sensitive to executive function and the reliance on just one item (the overlapping pentagons) to screen for visuospatial deficits (figure 1). The MMSE is particularly feeble in assessing patients with frontotemporal dementia, many of whom score within the ‘normal’ range on the test yet cannot function in social or work situations. Another slightly longer test is the Addenbrooke's Cognitive Examination, now in its third incarnation (ACE-III); it takes 15–20 min to administer and is used to screen for mild cognitive impairment, frontotemporal dementia and other less common causes of dementia.4 ,5

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Comparison of the components used to screen for visuospatial deficits in the Mini-Mental State Examination and the Addenbrooke’s Cognitive Examination – III.

In this modern technological era, where medical diagnoses often rely heavily on a combination of highly complex investigations, it seems unthinkable that the diagnosis of a devastating illness such as Alzheimer's disease would hinge on the results of a rough-and-ready 5 min test, the answers to which the patient spent the last 45 min rehearsing in the car on the way to the appointment. Never mind the fact that mood, education status and ethnicity could heavily influence the results. Perhaps even more concerning is that an equally devastating illness such as frontotemporal dementia could be completely missed because the executive and emotional deficits often experienced by patients early in the disease course are not examined by a restrictive tool that measures only the basic memory, attention, language and visuospatial functions.

There is clearly a need to update practice and, while recognising the MMSE's contribution to dementia, it is high time it retired gracefully to make room for newer, more theoretically motivated and, importantly, free assessment tools in dementia. Fortunately, there are now several superior assessment tools available for use in the clinic that can be supported by neuropsychological assessment and, when this is inconclusive, by brain imaging. However, the issue of which type of imaging is another story.