Mild cognitive impairment (MCI) is a term used to describe individuals with cognitive impairment that is not severe enough to affect daily functioning (e.g. Petersen, Reference Petersen2004; Winblad et al., Reference Winblad2004). Although MCI has been used to describe cognitive abnormality due to any number of causes that can be progressive, stable, or reversible, it is most often considered to be a transition phase between normal cognition and dementia.
After criteria were developed to classify individuals as having MCI for research purposes, diagnostic criteria for MCI were promulgated to inform clinical practice (Petersen et al., Reference Petersen1999). Following publication of these criteria, MCI was promoted from a descriptive category to a clinical diagnosis. This subtle but important shift led to consequences that are arguably more adverse than beneficial in clinical settings, and their widespread use contributed to a shared understanding of the concept as referring only to the earliest manifestation of Alzheimer's disease (AD). Despite the evolving nature of criteria used to diagnose MCI, ever more complex and ambitious, the perception that MCI refers to a prodrome of AD has been resistant to change.
MCI in current research
This issue of IPG includes three studies concerning MCI, all explicitly or implicitly in the context of its association with AD. These studies have in common novel approaches to detecting MCI or predicting its progression to dementia. Simultaneously, they attest to the conceptual limitations inherent in MCI as well as the utility of studying MCI as it relates to AD.
In the study by Hedman and colleagues (Reference Hedman, Kottorp, Almkvist and Nygard2018), individuals with MCI (using the Petersen, Reference Petersen2004 criteria) were interviewed to determine the degree to which they perceived various everyday technologies (e.g. coffee makers, cell phones, and automated teller machines) as challenging. They found that the perceived challenge level increased over five years for 15% of the technologies, which exceeds that expected by chance. Despite its small sample size, this study adds to findings from other studies demonstrating that MCI is often associated with mild functional decline. With the proliferation of available technologies in everyday life, findings from this study also indicate that quantifying functional ability as it relates to the use of technology is fertile ground for future work.
Unlike the criteria for MCI used by Hedman and colleagues, Beer and colleagues (Reference Beer2018) did not require objective cognitive impairment, but rather, relied on informant interview regarding cognitive decline. They assessed the utility of a memory test that was modified to include semantic interference to classify individuals with MCI who developed dementia four years later. For baseline memory test scores, they established thresholds separately for a clinical sample and a community sample that best predicted progression from MCI to dementia. They found that these thresholds were similar in the two samples, despite their differing demographic characteristics and rates of cognitive progression. Although they do not specifically state their intention to focus on dementia caused by AD, they note that semantic interference is related to “the particular type of cognitive decline seen in AD. . . .” (page 2). Thus, their study was designed with a view toward detection of MCI caused by AD. Findings from their study add to the growing number of studies demonstrating the importance of semantic interference in preclinical AD (e.g. Stark et al., Reference Stark, Yassa, Lacy and Stark2013), and also exemplify the lack of uniformity in the application of criteria for the classification of MCI.
Among the research concerning MCI contained in this issue, one study best illustrates the potential downside of widespread attempts to identify MCI. Park and colleagues (Reference Park2018) aimed to validate a computerized screening measure—the mSTS-MCI – for the detection of amnestic MCI in a community population. This test requires 15 minutes to administer and measures memory, executive functioning, and reaction time. Novel for computerized screening measures, the memory items assess both free recall and recognition for both visual and auditory stimuli. The test produces scores ranging from 0–28 (higher scores indicate better cognitive functioning), of which 18 points are possible for memory items and 10 points are possible for items measuring executive functioning. Reaction time is not part of the total score. They validated the mSTS-MCI against the Korean version of the Montreal Cognitive Assessment, (MoCA-K; Lee et al., Reference Lee2008) and presented the predictive values of cut-off scores for their test (total scores and reaction time) compared to the MoCA-K.
The results from their study would suggest a brief and easy-to-administer method of screening patients for MCI that presumably could be used broadly by healthcare providers. These advantages notwithstanding, computerized cognitive assessment is itself a contentious enterprise. The controversy lies not in the use of the technology per se, but rather, in the inappropriate use of such measures by untrained clinicians. As the diagnosis of MCI is considered to require clinical judgment, measures such as the mSTS-MCI omit this crucial aspect of diagnosis. The ease with which computer-administered tests can be used with little or no knowledge of how to interpret the findings is the most significant liability to this method of screening patients for MCI.
Another area of concern regarding similar methods that incorporate brief computerized screening measures is the inclusion of reaction time. Psychomotor and motor speed reliably increase with normal aging. Thus, patients who are older than those in the sample used to validate the mSTS-MCI (mean age = 74) would necessarily appear to have slow reaction time, when it may well be normal for age. Thus, applying a uniform cut-off score without consideration for age (and other demographic characteristics) is fraught with opportunities for erroneous classification. Use of this test will be problematic in individuals at the ends of the continuum of the adult-age span: application of their cut-off scores will not detect slow reaction times in younger adults and will incorrectly consider individuals in the older age range to demonstrate impaired reaction time.
This study serves as a reminder that as MCI continues to permeate research and clinical practice, efforts to reduce the potential negative consequences of this trend need to be considered.
The popularity of MCI hinders its evolution
Since it was conceived, several sets of diagnostic criteria for MCI have been developed. As well, MCI has been parsed into subtypes, each reflecting different etiologies. However, the ubiquitous use of the term MCI has given rise to a proliferation of information about the concept. As a result, its reputation as a state that represents the very earliest manifestations of AD persists in popular media, clinical settings, and research studies. Indeed, one needs only to perform an internet search of the term MCI to appreciate its connection to AD, and thus the potential for harm to patients who receive a diagnosis of MCI, correctly or not.
Limitations of the concept of MCI have been well articulated and those who question its utility provide a compelling case to abandon the term, at least in clinical practice (Allegri et al., Reference Allegri, Glaser, Taragano and Buschke2008; Saunders et al., Reference Saunders2017). Efforts to address its limitations by modifying its criteria and broadening it to include preclinical states of dementing illnesses other than AD have not led to consistent application of such updates, only fueling the confusion surrounding the concept, even in research settings. Even in the study by Park and colleagues, amnestic MCI was the focus, but this is specified only in the methods section of their paper; only “MCI” is included in the title and abstract of their study. This lack of specificity adds to the misconception that MCI is a unitary concept that refers only to preclinical AD.
Clinical trials aimed at treating AD at its earliest manifestation typically recruit patients with MCI but the heterogenous nature of MCI can lead to inconsistent or even negative results. With a view toward focusing specifically on cognitive decline due AD rather than other causes, the National Institute on Aging and the Alzheimer's Association convened a workgroup to delineate criteria for MCI caused by preclinical AD (Albert et al., 2011). It is of note that the clinical syndrome they describe is not confined to the amnestic MCI subtype that has been described. However, the addition of clinical assessment to establish that the cause is likely to be AD (by ruling out other potential causes of cognitive impairment) increases the likelihood that even patients with non-amnestic MCI caused by AD will be included in this group.
Conclusions
Despite its limitations, the concept of MCI, ingrained in both research and clinical practice, is likely ineradicable at this point. Ambitious efforts to define subtypes of MCI that reflect distinct disease prodromes have met with limited success in gaining consistent application. The three MCI studies in this issue attest to the inconsistent criteria used to diagnose MCI in research studies. At the same time, these studies elucidate various aspects pertaining to the earliest manifestation of progressive cognitive disorders likely due to AD. If the term MCI is here to stay, then restricting its use to refer exclusively to the prodrome of AD has the potential to yield valuable information about the most common cause of dementia in late life. Rather than abandoning the term altogether or increasing its complexity, perhaps embracing its reputation by pruning the term MCI to indicate only preclinical AD is the most judicious course of action. Or in the words of Steve Jobs: “Do not try to do everything. Do one thing well.”
Conflict of interest
None.
You have
Access
Mild cognitive impairment (MCI) is a term used to describe individuals with cognitive impairment that is not severe enough to affect daily functioning (e.g. Petersen, Reference Petersen2004; Winblad et al., Reference Winblad2004). Although MCI has been used to describe cognitive abnormality due to any number of causes that can be progressive, stable, or reversible, it is most often considered to be a transition phase between normal cognition and dementia.
After criteria were developed to classify individuals as having MCI for research purposes, diagnostic criteria for MCI were promulgated to inform clinical practice (Petersen et al., Reference Petersen1999). Following publication of these criteria, MCI was promoted from a descriptive category to a clinical diagnosis. This subtle but important shift led to consequences that are arguably more adverse than beneficial in clinical settings, and their widespread use contributed to a shared understanding of the concept as referring only to the earliest manifestation of Alzheimer's disease (AD). Despite the evolving nature of criteria used to diagnose MCI, ever more complex and ambitious, the perception that MCI refers to a prodrome of AD has been resistant to change.
MCI in current research
This issue of IPG includes three studies concerning MCI, all explicitly or implicitly in the context of its association with AD. These studies have in common novel approaches to detecting MCI or predicting its progression to dementia. Simultaneously, they attest to the conceptual limitations inherent in MCI as well as the utility of studying MCI as it relates to AD.
In the study by Hedman and colleagues (Reference Hedman, Kottorp, Almkvist and Nygard2018), individuals with MCI (using the Petersen, Reference Petersen2004 criteria) were interviewed to determine the degree to which they perceived various everyday technologies (e.g. coffee makers, cell phones, and automated teller machines) as challenging. They found that the perceived challenge level increased over five years for 15% of the technologies, which exceeds that expected by chance. Despite its small sample size, this study adds to findings from other studies demonstrating that MCI is often associated with mild functional decline. With the proliferation of available technologies in everyday life, findings from this study also indicate that quantifying functional ability as it relates to the use of technology is fertile ground for future work.
Unlike the criteria for MCI used by Hedman and colleagues, Beer and colleagues (Reference Beer2018) did not require objective cognitive impairment, but rather, relied on informant interview regarding cognitive decline. They assessed the utility of a memory test that was modified to include semantic interference to classify individuals with MCI who developed dementia four years later. For baseline memory test scores, they established thresholds separately for a clinical sample and a community sample that best predicted progression from MCI to dementia. They found that these thresholds were similar in the two samples, despite their differing demographic characteristics and rates of cognitive progression. Although they do not specifically state their intention to focus on dementia caused by AD, they note that semantic interference is related to “the particular type of cognitive decline seen in AD. . . .” (page 2). Thus, their study was designed with a view toward detection of MCI caused by AD. Findings from their study add to the growing number of studies demonstrating the importance of semantic interference in preclinical AD (e.g. Stark et al., Reference Stark, Yassa, Lacy and Stark2013), and also exemplify the lack of uniformity in the application of criteria for the classification of MCI.
Among the research concerning MCI contained in this issue, one study best illustrates the potential downside of widespread attempts to identify MCI. Park and colleagues (Reference Park2018) aimed to validate a computerized screening measure—the mSTS-MCI – for the detection of amnestic MCI in a community population. This test requires 15 minutes to administer and measures memory, executive functioning, and reaction time. Novel for computerized screening measures, the memory items assess both free recall and recognition for both visual and auditory stimuli. The test produces scores ranging from 0–28 (higher scores indicate better cognitive functioning), of which 18 points are possible for memory items and 10 points are possible for items measuring executive functioning. Reaction time is not part of the total score. They validated the mSTS-MCI against the Korean version of the Montreal Cognitive Assessment, (MoCA-K; Lee et al., Reference Lee2008) and presented the predictive values of cut-off scores for their test (total scores and reaction time) compared to the MoCA-K.
The results from their study would suggest a brief and easy-to-administer method of screening patients for MCI that presumably could be used broadly by healthcare providers. These advantages notwithstanding, computerized cognitive assessment is itself a contentious enterprise. The controversy lies not in the use of the technology per se, but rather, in the inappropriate use of such measures by untrained clinicians. As the diagnosis of MCI is considered to require clinical judgment, measures such as the mSTS-MCI omit this crucial aspect of diagnosis. The ease with which computer-administered tests can be used with little or no knowledge of how to interpret the findings is the most significant liability to this method of screening patients for MCI.
Another area of concern regarding similar methods that incorporate brief computerized screening measures is the inclusion of reaction time. Psychomotor and motor speed reliably increase with normal aging. Thus, patients who are older than those in the sample used to validate the mSTS-MCI (mean age = 74) would necessarily appear to have slow reaction time, when it may well be normal for age. Thus, applying a uniform cut-off score without consideration for age (and other demographic characteristics) is fraught with opportunities for erroneous classification. Use of this test will be problematic in individuals at the ends of the continuum of the adult-age span: application of their cut-off scores will not detect slow reaction times in younger adults and will incorrectly consider individuals in the older age range to demonstrate impaired reaction time.
This study serves as a reminder that as MCI continues to permeate research and clinical practice, efforts to reduce the potential negative consequences of this trend need to be considered.
The popularity of MCI hinders its evolution
Since it was conceived, several sets of diagnostic criteria for MCI have been developed. As well, MCI has been parsed into subtypes, each reflecting different etiologies. However, the ubiquitous use of the term MCI has given rise to a proliferation of information about the concept. As a result, its reputation as a state that represents the very earliest manifestations of AD persists in popular media, clinical settings, and research studies. Indeed, one needs only to perform an internet search of the term MCI to appreciate its connection to AD, and thus the potential for harm to patients who receive a diagnosis of MCI, correctly or not.
Limitations of the concept of MCI have been well articulated and those who question its utility provide a compelling case to abandon the term, at least in clinical practice (Allegri et al., Reference Allegri, Glaser, Taragano and Buschke2008; Saunders et al., Reference Saunders2017). Efforts to address its limitations by modifying its criteria and broadening it to include preclinical states of dementing illnesses other than AD have not led to consistent application of such updates, only fueling the confusion surrounding the concept, even in research settings. Even in the study by Park and colleagues, amnestic MCI was the focus, but this is specified only in the methods section of their paper; only “MCI” is included in the title and abstract of their study. This lack of specificity adds to the misconception that MCI is a unitary concept that refers only to preclinical AD.
Clinical trials aimed at treating AD at its earliest manifestation typically recruit patients with MCI but the heterogenous nature of MCI can lead to inconsistent or even negative results. With a view toward focusing specifically on cognitive decline due AD rather than other causes, the National Institute on Aging and the Alzheimer's Association convened a workgroup to delineate criteria for MCI caused by preclinical AD (Albert et al., 2011). It is of note that the clinical syndrome they describe is not confined to the amnestic MCI subtype that has been described. However, the addition of clinical assessment to establish that the cause is likely to be AD (by ruling out other potential causes of cognitive impairment) increases the likelihood that even patients with non-amnestic MCI caused by AD will be included in this group.
Conclusions
Despite its limitations, the concept of MCI, ingrained in both research and clinical practice, is likely ineradicable at this point. Ambitious efforts to define subtypes of MCI that reflect distinct disease prodromes have met with limited success in gaining consistent application. The three MCI studies in this issue attest to the inconsistent criteria used to diagnose MCI in research studies. At the same time, these studies elucidate various aspects pertaining to the earliest manifestation of progressive cognitive disorders likely due to AD. If the term MCI is here to stay, then restricting its use to refer exclusively to the prodrome of AD has the potential to yield valuable information about the most common cause of dementia in late life. Rather than abandoning the term altogether or increasing its complexity, perhaps embracing its reputation by pruning the term MCI to indicate only preclinical AD is the most judicious course of action. Or in the words of Steve Jobs: “Do not try to do everything. Do one thing well.”
Conflict of interest
None.