Neurology -- Correspondence for Ballard et al., 63 (6) 983-988

Correspondence published:

NINDS AIREN neuroimaging criteria do not distinguish stroke patients with and without dementia: Wiesje van der Flier, Ilse van Straaten, Frederik Barkhof, Philip Scheltens (16 April 2005)

NINDS AIREN neuroimaging criteria do not distinguish stroke patients with and without dementia 16 April 2005

Wiesje van der Flier,
Vrije Universiteit Medical Center
PO Box 7057, 1007 MB Amsterdam, the Netherlands,
Ilse van Straaten, Frederik Barkhof, Philip Scheltens
Send Correspondence to journal:
Re: NINDS AIREN neuroimaging criteria do not distinguish stroke patients with and without dementia

wm.vdflier{at}vumc.nl Wiesje van der Flier, et al.

Ballard et al [1] report that in a group of stroke patients, the neuroimaging component within the NINDS AIREN criteria [2] for vascular dementia did not distinguish between patients with and without post-stroke dementia. In addition, groups did not differ in number or size of infarcts. However, patients with dementia had greater hippocampal atrophy. The authors conclude that the NINDS AIREN neuroimaging criteria may need to be revised, because they do not distinguish between stroke patients with and without dementia. Although we agree with the authors that the criteria may need revision, we would like to make some critical comments concerning the research question of their study.
The NINDS AIREN criteria describe the clinical syndrome of vascular dementia (VaD). Like in any other subtype of dementia, clinical characteristics define whether or not a patient is demented. The neuroimaging component of the criteria only serves to determine the probability that the observed dementia is of vascular origin. In the present sample of patients who all suffered a cerebrovascular event, it was to be expected that the neuroimaging criteria did not reveal any difference between dementia-groups in terms of vascular burden. The authors merely show that a patient may fulfill the radiological criteria for VaD as defined by the NINDS-AIREN, and still not be demented. This is a reversal of the diagnostic process, and reminds us that an MRI scan may never be used in isolation to diagnose VaD. Furthermore, excluding patients who were demented prior to the stroke may have introduced a selection bias, while this group has the highest probability of suffering VaD.
Moreover, the greater hippocampal atrophy of the demented subgroup suggests that these patients may suffer from AD rather than VaD, or have combined pathology. As the authors propose, vascular pathology such as a stroke may interact with preexisting subclinical Alzheimer pathology, resulting in clinical dementia of the Alzheimer type. As the NINDS AIREN criteria were not developed to detect AD, there is no reason to expect that these criteria would be sensitive to detect patients with AD.
Finally, it has been shown that the neuroimaging component of the NINDS AIREN criteria has insufficient reliability [3], and this might partly account for the inability to discriminate between groups. Operational criteria have been put forward and tested, resulting in a considerably improved reliability in experienced readers.
References
1. Ballard CG, Burton EJ, Barber R, et al. NINDS AIREN neuroimaging criteria do not distinguish stroke patients with and without dementia. Neurology 2004; 63(6):983-988.
2. Rom�n GC, Tatemichi TK, Erkinjuntti T, et al. Vascular dementia: diagnostic criteria for research studies. Report of the NINDS-AIREN International Workshop. Neurology 1993; 43(2):250-260.
3. Van Straaten EC, Scheltens P, Knol DL, et al. Operational Definitions for the NINDS-AIREN Criteria for Vascular Dementia. An Interobserver Study. Stroke 2003.