Dement Geriatr Cogn Disord 2015;40:44–53 DOI: 10.1159/000381142 Accepted: February 17, 2015 Published online: April 25, 2015
© 2015 S. Karger AG, Basel 1420–8008/15/0402–0044$39.50/0 www.karger.com/dem
Original Research Article
Hippocampal Subfield Atrophy in Multi-Domain but Not Amnestic Mild Cognitive Impairment Carl Fredrik Eliassen a, b Per Selnes b Ina Selseth Almdahl b Ivar Reinvang a Tormod Fladby b Erik Hessen a, b a Department
of Psychology, University of Oslo, Oslo, and b Department of Neurology, Akershus University Hospital, Lørenskog, Norway
Key Words Hippocampal subfields · Perirhinal cortex · Entorhinal cortex · Amnestic mild cognitive impairment · Atrophy Abstract Background/Aims: To investigate differences in hippocampal (HP) subfields and the adjoining perirhinal and entorhinal cortices (PRC and ERC) between amnestic mild cognitive impairment (aMCI) and multi-domain amnestic MCI (mdMCI) patients, and controls. Methods: Nineteen patients characterized as aMCI were compared with 24 mdMCI patients and 31 controls by means of an automatic HP segmentation procedure. Results: We found significant atrophy of the PRC and ERC in aMCI relative to controls, whereas a more pronounced pattern of atrophy in most subfields, including total HP volume, was found in the mdMCI group. The mdMCI group also had a significant cornu ammonis sector 4 region with dentate gyrus, subiculum and total HP atrophy relative to aMCI. Conclusion: The aMCI group showed atrophy in the PRC and ERC, whereas significantly more affection of the HP subfields was evident in mdMCI. The mdMCI group may thus represent clinical progression relative to aMCI coupled © 2015 S. Karger AG, Basel with HP subfield affection.
Introduction
Carl Fredrik Eliassen Department of Psychology, University of Oslo PO Box 1094, Blindern NO–0317 Oslo (Norway) E-Mail c.f.a.eliassen @ medisin.uio.no
Downloaded by: University of Pittsburgh 198.143.32.65 - 1/27/2016 4:50:23 AM
Hippocampal (HP) and subhippocampal (entorhinal and perirhinal cortices, ERC and PRC) structures in the medial temporal lobe (MTL) are essential for learning and memory [1]. Increasing evidence suggests that MTL structures are particularly vulnerable to early pathophysiological change in incipient dementia [2–4]. Mild cognitive impairment (MCI) with
45
Dement Geriatr Cogn Disord 2015;40:44–53 DOI: 10.1159/000381142
© 2015 S. Karger AG, Basel www.karger.com/dem
amnestic memory deficits and HP atrophy has been identified as a reliable sign of preclinical Alzheimer’s disease (AD) [5]. The HP formation contains subfields with distinct neuronal cell types and specialized connectivity subserving learning and memory, visuospatial functions and other cognitive modalities [6, 7]. Findings indicate that subtle pathological changes within the HP may precede global HP atrophy [8–11], and these substructures may be sequentially affected according to an anterior to posterior gradient of vulnerability [12]. Specifically, atrophy should first be evident in HP input structures, such as the ERC, before a regional atrophic pattern including the HP head, body and tail follows [12, 13]. Several recent findings indicate that the HP cornu ammonis sector 1 (CA1) region is one of the earliest affected regions in MCI patients [8, 11, 14]. One study found atrophy of the CA1 in combination with the subiculum to be associated with later amnestic MCI (aMCI) classification. At follow-up, subsequent atrophy in the CA2-3 region was associated with progression to AD [8]. Both the CA1 and the CA3 region receive input from the ERC, but more to the CA3 region and the dentate gyrus (DG). Whereas the CA3 region displays a high proportion of reverberating neural connectivity, the CA1 region displays fewer such recurrent connections and, together with the subiculum, constitutes the main output region of the HP [15–18]. The CA2-3 region in combination with the subiculum and presubiculum has been reported to be atrophic in aMCI relative to controls [19, 20]. In contrast, a recent study utilizing 7-tesla MR images and manual delineation of HP subfields did not find significant volume differences to any subfields or ERC in MCI, but significant volume differences between early-stage AD versus controls, and early-stage AD versus MCI in most HP subfields, including the ERC [21]. Inconsistent findings as to which HP region is affected in MCI across studies may be related to the variability in segmentation techniques [14, 19, 22]. However, another possible caveat concerns the criteria for selecting patient samples. There is a tradition for classification of MCI patients into subtypes according to which cognitive domains are affected [23]. Especially variants that display memory impairment alone or in combination with other cognitive symptoms have been coupled with later AD diagnosis [24–28]. Most studies of the HP subfield volumetry to date categorize MCI patients according to the Petersen criteria [29, 30], where only one neuropsychological memory measure is required to fall under a certain cutoff value [29]. This has implications for the diagnostic validity as impaired test scores seem to frequently occur among healthy elderly [31]. Jak et al. [32] found ‘comprehensive criteria’, which require reduced scores on two tests within a cognitive domain to be superior to the more commonly used Petersen criteria [29, 30] with respect to reducing false-positives and representing better ecological validity [32–34]. In the present study, we investigated potential differences in atrophy in HP subfields between control subjects and MCI patients classified according to neuropsychological criteria resembling the comprehensive criteria [32] as this classification scheme is more valid [33] and reliable [32, 34] than the traditional approach, which is more frequent in HP subfield volume studies to date. The combination of impairments in several cognitive domains has been shown to be a stronger predictor of conversion to dementia than impairment within a single cognitive domain [24, 25, 35, 36]. HP subfields may become compromised sequentially, where anterior structure atrophy precedes global HP atrophy [12]. Multi-domain amnestic MCI (mdMCI) putatively represents a more advanced stage of pre-dementia than singledomain aMCI [37]. Therefore, the group classified as mdMCI may present with atrophy on par with aMCI in adjacent HP input structures, but the former with more atrophy in the HP and subfields. We suggest that an alternative to the traditional neuropsychological MCI classification criteria may reduce patient sample heterogeneity with respect to cognitive symptoms [32, 33, 38]. We hypothesized that, relative to controls, patients classified as mdMCI would display greater atrophy in a wider range of selected HP subfields than aMCI. Specifically, we expected
Downloaded by: University of Pittsburgh 198.143.32.65 - 1/27/2016 4:50:23 AM
Eliassen et al.: Hippocampal Subfield Atrophy in Multi-Domain but Not Amnestic Mild Cognitive Impairment
46
Dement Geriatr Cogn Disord 2015;40:44–53 DOI: 10.1159/000381142
© 2015 S. Karger AG, Basel www.karger.com/dem
Eliassen et al.: Hippocampal Subfield Atrophy in Multi-Domain but Not Amnestic Mild Cognitive Impairment
PRC and ERC atrophy, and CA2-3, subiculum and presubiculum subfield atrophy in both aMCI and mdMCI as these have previously been reported to be atrophic with similar HP segmentation approaches [19, 20]. Furthermore, we expected a relatively more pronounced atrophy in the HP subfields, including total HP volume in the mdMCI subgroup as this group may represent a more advanced MCI stage than aMCI [37].
Methods Research Participants Participants were recruited consecutively from referrals to the memory clinic at Akershus University Hospital in Oslo between 2005 and 2013. Patients with major depressive and other severe psychiatric disorders were excluded as were patients with neurological disease, a history of traumatic brain injury, cardiovascular events, other severe somatic illness or current substance abuse. For inclusion, subjective (verified by an informant) cognitive complaints lasting more than 6 months were required in addition to objective cognitive symptoms according to the following criteria: preserved general function, no or very mild deficits in activities of daily living, and a Global Deterioration Scale score of 3 [39]. Screening tests included parameters 13–20 (memory, disorientation, abstract thinking, visuospatial ability, language, sensory aphasia, visual agnosia and apraxia) from the stepwise comparative status analysis [40], word fluency, interference and numeral-letter items from the I-Flex [41], items from the Neurobehavioral Cognitive Status Examination – Cognistat [42] as well as Mini-Mental State Examination (MMSE) [43] and a Clinical Dementia Rating (CDR) score of 0.5 [44]. The information for CDR was gathered from both the subject and an informant. Subjects without symptoms according to the checklists or CDR were not included since their cognitive impairment was considered too benign; neither were subjects with more than two cognitive symptoms on the described checklists and/or a score
© 2015 S. Karger AG, Basel 1420–8008/15/0402–0044$39.50/0 www.karger.com/dem
Original Research Article
Hippocampal Subfield Atrophy in Multi-Domain but Not Amnestic Mild Cognitive Impairment Carl Fredrik Eliassen a, b Per Selnes b Ina Selseth Almdahl b Ivar Reinvang a Tormod Fladby b Erik Hessen a, b a Department
of Psychology, University of Oslo, Oslo, and b Department of Neurology, Akershus University Hospital, Lørenskog, Norway
Key Words Hippocampal subfields · Perirhinal cortex · Entorhinal cortex · Amnestic mild cognitive impairment · Atrophy Abstract Background/Aims: To investigate differences in hippocampal (HP) subfields and the adjoining perirhinal and entorhinal cortices (PRC and ERC) between amnestic mild cognitive impairment (aMCI) and multi-domain amnestic MCI (mdMCI) patients, and controls. Methods: Nineteen patients characterized as aMCI were compared with 24 mdMCI patients and 31 controls by means of an automatic HP segmentation procedure. Results: We found significant atrophy of the PRC and ERC in aMCI relative to controls, whereas a more pronounced pattern of atrophy in most subfields, including total HP volume, was found in the mdMCI group. The mdMCI group also had a significant cornu ammonis sector 4 region with dentate gyrus, subiculum and total HP atrophy relative to aMCI. Conclusion: The aMCI group showed atrophy in the PRC and ERC, whereas significantly more affection of the HP subfields was evident in mdMCI. The mdMCI group may thus represent clinical progression relative to aMCI coupled © 2015 S. Karger AG, Basel with HP subfield affection.
Introduction
Carl Fredrik Eliassen Department of Psychology, University of Oslo PO Box 1094, Blindern NO–0317 Oslo (Norway) E-Mail c.f.a.eliassen @ medisin.uio.no
Downloaded by: University of Pittsburgh 198.143.32.65 - 1/27/2016 4:50:23 AM
Hippocampal (HP) and subhippocampal (entorhinal and perirhinal cortices, ERC and PRC) structures in the medial temporal lobe (MTL) are essential for learning and memory [1]. Increasing evidence suggests that MTL structures are particularly vulnerable to early pathophysiological change in incipient dementia [2–4]. Mild cognitive impairment (MCI) with
45
Dement Geriatr Cogn Disord 2015;40:44–53 DOI: 10.1159/000381142
© 2015 S. Karger AG, Basel www.karger.com/dem
amnestic memory deficits and HP atrophy has been identified as a reliable sign of preclinical Alzheimer’s disease (AD) [5]. The HP formation contains subfields with distinct neuronal cell types and specialized connectivity subserving learning and memory, visuospatial functions and other cognitive modalities [6, 7]. Findings indicate that subtle pathological changes within the HP may precede global HP atrophy [8–11], and these substructures may be sequentially affected according to an anterior to posterior gradient of vulnerability [12]. Specifically, atrophy should first be evident in HP input structures, such as the ERC, before a regional atrophic pattern including the HP head, body and tail follows [12, 13]. Several recent findings indicate that the HP cornu ammonis sector 1 (CA1) region is one of the earliest affected regions in MCI patients [8, 11, 14]. One study found atrophy of the CA1 in combination with the subiculum to be associated with later amnestic MCI (aMCI) classification. At follow-up, subsequent atrophy in the CA2-3 region was associated with progression to AD [8]. Both the CA1 and the CA3 region receive input from the ERC, but more to the CA3 region and the dentate gyrus (DG). Whereas the CA3 region displays a high proportion of reverberating neural connectivity, the CA1 region displays fewer such recurrent connections and, together with the subiculum, constitutes the main output region of the HP [15–18]. The CA2-3 region in combination with the subiculum and presubiculum has been reported to be atrophic in aMCI relative to controls [19, 20]. In contrast, a recent study utilizing 7-tesla MR images and manual delineation of HP subfields did not find significant volume differences to any subfields or ERC in MCI, but significant volume differences between early-stage AD versus controls, and early-stage AD versus MCI in most HP subfields, including the ERC [21]. Inconsistent findings as to which HP region is affected in MCI across studies may be related to the variability in segmentation techniques [14, 19, 22]. However, another possible caveat concerns the criteria for selecting patient samples. There is a tradition for classification of MCI patients into subtypes according to which cognitive domains are affected [23]. Especially variants that display memory impairment alone or in combination with other cognitive symptoms have been coupled with later AD diagnosis [24–28]. Most studies of the HP subfield volumetry to date categorize MCI patients according to the Petersen criteria [29, 30], where only one neuropsychological memory measure is required to fall under a certain cutoff value [29]. This has implications for the diagnostic validity as impaired test scores seem to frequently occur among healthy elderly [31]. Jak et al. [32] found ‘comprehensive criteria’, which require reduced scores on two tests within a cognitive domain to be superior to the more commonly used Petersen criteria [29, 30] with respect to reducing false-positives and representing better ecological validity [32–34]. In the present study, we investigated potential differences in atrophy in HP subfields between control subjects and MCI patients classified according to neuropsychological criteria resembling the comprehensive criteria [32] as this classification scheme is more valid [33] and reliable [32, 34] than the traditional approach, which is more frequent in HP subfield volume studies to date. The combination of impairments in several cognitive domains has been shown to be a stronger predictor of conversion to dementia than impairment within a single cognitive domain [24, 25, 35, 36]. HP subfields may become compromised sequentially, where anterior structure atrophy precedes global HP atrophy [12]. Multi-domain amnestic MCI (mdMCI) putatively represents a more advanced stage of pre-dementia than singledomain aMCI [37]. Therefore, the group classified as mdMCI may present with atrophy on par with aMCI in adjacent HP input structures, but the former with more atrophy in the HP and subfields. We suggest that an alternative to the traditional neuropsychological MCI classification criteria may reduce patient sample heterogeneity with respect to cognitive symptoms [32, 33, 38]. We hypothesized that, relative to controls, patients classified as mdMCI would display greater atrophy in a wider range of selected HP subfields than aMCI. Specifically, we expected
Downloaded by: University of Pittsburgh 198.143.32.65 - 1/27/2016 4:50:23 AM
Eliassen et al.: Hippocampal Subfield Atrophy in Multi-Domain but Not Amnestic Mild Cognitive Impairment
46
Dement Geriatr Cogn Disord 2015;40:44–53 DOI: 10.1159/000381142
© 2015 S. Karger AG, Basel www.karger.com/dem
Eliassen et al.: Hippocampal Subfield Atrophy in Multi-Domain but Not Amnestic Mild Cognitive Impairment
PRC and ERC atrophy, and CA2-3, subiculum and presubiculum subfield atrophy in both aMCI and mdMCI as these have previously been reported to be atrophic with similar HP segmentation approaches [19, 20]. Furthermore, we expected a relatively more pronounced atrophy in the HP subfields, including total HP volume in the mdMCI subgroup as this group may represent a more advanced MCI stage than aMCI [37].
Methods Research Participants Participants were recruited consecutively from referrals to the memory clinic at Akershus University Hospital in Oslo between 2005 and 2013. Patients with major depressive and other severe psychiatric disorders were excluded as were patients with neurological disease, a history of traumatic brain injury, cardiovascular events, other severe somatic illness or current substance abuse. For inclusion, subjective (verified by an informant) cognitive complaints lasting more than 6 months were required in addition to objective cognitive symptoms according to the following criteria: preserved general function, no or very mild deficits in activities of daily living, and a Global Deterioration Scale score of 3 [39]. Screening tests included parameters 13–20 (memory, disorientation, abstract thinking, visuospatial ability, language, sensory aphasia, visual agnosia and apraxia) from the stepwise comparative status analysis [40], word fluency, interference and numeral-letter items from the I-Flex [41], items from the Neurobehavioral Cognitive Status Examination – Cognistat [42] as well as Mini-Mental State Examination (MMSE) [43] and a Clinical Dementia Rating (CDR) score of 0.5 [44]. The information for CDR was gathered from both the subject and an informant. Subjects without symptoms according to the checklists or CDR were not included since their cognitive impairment was considered too benign; neither were subjects with more than two cognitive symptoms on the described checklists and/or a score
