Our findings suggest that persons with MCI do have clinically significant functional impairments, particularly on an informant-based measure that quantifies error behaviors, and these functional difficulties are associated with verbal learning performance. Given that MCI participants with early functional changes experience more rapid functional decline6
and convert to dementia faster7
than their functionally intact MCI peers, identifying MCI participants with early functional changes is clinically important. Our results suggest that the development and implementation of more sensitive functional assessments can assist with earlier detection. These findings and their implications are discussed in more detail below.
Participants with MCI differed from NC participants on an error-based functional measure, such that the MCI participants performed more than 1.5 standard deviations worse on the FC-ADL than the NC participants. In contrast, a between-group difference that did not reach statistical significance was observed using a traditional measure of IADLs, though this difference appears clinically negligible. As expected, no difference was observed using a traditional measure of BADLs. These findings suggest that MCI participants may not differ from NC participants on global informant-based ratings of ADLs that focus on functional dependence; however, quantitative assessment of functional errors may yield subtle early functional changes that are clinically meaningful.
Our findings are consistent with prior findings documenting poorer informant-rated functional status in individuals with preclinical dementia and MCI compared with their NC peers.8,17
The effect size generated in the present study for the FC-ADL between-group difference (i.e., d
= 0.84) appears larger than some effect sizes previously reported for other functional measures.8
Our findings extend the extant functional literature by suggesting that traditional and novel functional rating measures differ in their detection of early ADL changes associated with preclinical dementia. Moreover, our data emphasize that functional impairment among individuals with MCI is subtle, which in turn suggests that functional measures to evaluate functional status or the trajectory of functional decline among individuals with MCI need to be more qualitative in nature. This approach is in contrast to evaluating functional trajectory in dementia, where global measures assessing IADL and BADL dependence are appropriate and sensitive to changes.18,38
Although we hypothesized that executive function and memory measures would both predict functional abilities in persons with MCI, our findings were mixed. Only a serial list learning variable significantly correlated with the FC-ADL total score at the a priori significance level. However, for the IADL subscale score, correlations that approached statistical significance included an executive functioning measure of judgment and a delayed recall memory measure. Both of these findings are supported by prior cognitive-ADL research, such that past studies have related memory16,17
and elements of executive functioning17,23
to functional ratings of MCI participants. Furthermore, neuroimaging data support frontal lobe39
and hippocampal changes40
in MCI, which likely mediate, in part, the cognitive difficulties underlying these early functional changes.
The finding that two measures of functional skills were sensitive to disparate cognitive functions in the same sample of MCI participants suggests that informant-based functional measures may be differentially sensitive to cognitive functions. This finding could account for variation in cognitive-ADL relations in the literature.18,19
Additional research is warranted to better understand how functional measures are differentially sensitive to cognitive functioning and how these inconsistencies play out in various diagnostic groups (e.g., Alzheimer disease versus vascular dementia patients) and samples with cognitive severity differences (e.g., NC participants versus dementia patients).
Our between-group item analysis of the FC-ADL yielded several significant differences on items assessing memory, perseveration, and fatigue, though the clinical significance of these differences may be marginal. Of the few differences noted, half of these were memory in nature, which corresponds to the correlational findings that linked a serial list learning variable to FC-ADL performance. The majority of our MCI participants (i.e., 82%) were categorized as either single or multiple domain amnestic MCI. In light of the correlation between the FC-ADL scale and verbal learning, the FC-ADL item analysis findings are likely related to the hallmark memory impairment associated with the MCI research diagnostic criteria3,26
and the preponderance of amnestic MCI participants in our sample. An item analysis on the FC-ADL is warranted between NC and MCI participants with more nonamnestic cognitive profiles.
The current study has a number of strengths including the well-characterized MCI sample with clinical status based on comprehensive work-up and multidisciplinary consensus diagnosis. Our implementation of a sensitive, error-based measure of functional status (FC-ADL) in conjunction with a more traditional, global measure sheds light on discrepancies among prior cognitive-ADL research findings. Finally, the robust NC and MCI between-group difference on the FC-ADL measure supports the inclusion of more sensitive functional measures in the clinical evaluation of MCI and preclinical dementia.
Despite these strengths, there are limitations to the current study that restrict the generalizability of our findings. In particular, our cohort was, on average, college-educated and predominantly white of European ancestry. The sample sizes of the two groups may not have been sufficiently large to detect smaller effects on the more traditional ADL measure, the ADL/neuropsychological correlations, or the between-group FC-ADL item analysis. Although our study employed a rigorous selection process to ensure that participants with MCI were classified according to current research diagnostic criteria, known issues with MCI as a construct apply to our study (e.g., instability of classification over time). Longitudinal follow-up of the current cohort will elucidate relations between MCI status at baseline and long-term clinical outcome with specific emphasis on conversion to dementia. Additional future research directions include a more detailed investigation of functional status by MCI subtype (e.g., amnestic versus nonamnestic MCI, single versus multiple domain) and longitudinal tracking of cognitive and functional changes over time. Qualitative investigation of the specific functional abilities that decline earliest or most rapidly with aging and dementia would also extend the current literature.