A significant percentage of older depressed adults also have cognitive impairment [1
]. Studies based on diagnostic criteria for major and minor depression may miss important associations between depression symptoms and cognitive impairment among older adults. Subtypes of depression in which symptoms are present intermittently or in which sadness and anhedonia are denied may be as important as major depression in older adults because of the association with cognitive impairment. For example, Blazer and colleagues, employing grade-of-membership analysis to analyze the Durham-Piedmont ECA data, found a symptom profile tending to occur in older persons in which cognitive impairment was prominent and accompanied by sleep and appetite disturbance, hopelessness, and thoughts of death [2
]. In addition, proneness to psychological distress has been found to increase risk of Alzheimer's disease [3
]. Futhermore, recent findings by Woo et. al, suggest that subsyndromal depressive symptoms in late-life may impair physiologic processes underlying memory [4
Behavioral genetics may provide information regarding the overlap between depression and cognition. Possessing the ε4 allele of apolipoprotein E (APOE-ε4) has been associated with an increased risk for a range of neuropsychiatric conditions in older adults, perhaps the most widely known of which is Alzheimer's disease [5
]. The APOE-ε4 allele has also been associated with cognitive impairment in non-demented older adults. A review by Savitz et al. [7
] found such a relationship in 32 out of 45 studies examined, associating APOE-ε4 with impaired cognitive functioning both cross-sectionally and longitudinally. APOE-ε4 has been shown to increase risk of cognitive impairment in specific domains of cognition such as memory [8
], attention [10
], and implicit learning [12
]. Specifically, the presence of the APOE-ε4 allele has been shown to increase risk of cognitive impairment when brain reserve is low [13
]. Depression is another neuropsychiatric condition that has been explored for a potential association with APOE-ε4. A similar association between APOE-ε4 and depression has not been clearly established. Initial work done by Krishnan et al. [14
] suggested that APOE-ε4 was associated with late-onset depression in older adults. More recently, an analysis conducted by Yung-Chieh et. al found that APOE-ε4 may be correlated with severe depression in the elderly [15
]. Other investigators found similar associations [15
] although some studies failed to replicate these findings (e.g. [17
Recent studies examining the collective relationships between APOE-ε4, depression and cognitive function have led to new insights. Geda et al. [19
] examined 840 cognitively normal, non-depressed older adults at baseline and then at a 3.5-year follow-up. The authors noted that respondents who developed depression were at a higher risk for developing mild cognitive impairment, but also found an additive interaction in which respondents who developed depression and possessed APOE-ε4 were at a greater risk of developing mild cognitive impairment than could be accounted for by the separate effects of depression and APOE-ε4. Similarly, Hwang et al. [20
] found that depressed older adults with APOE-ε4 showed significantly lower scores on the Mini-Mental State Exam (MMSE) and an increased risk for suicide compared to depressed respondents who did not possess APOE-ε4. These results suggest that a separate class of depression may exist related to the presence of the APOE-ε4 allele and decreased cognitive functioning.
To our knowledge, this is the first investigation to examine the relationship between depressive symptoms, specific domains of cognition, and APOE-ε4 using latent class statistical methods. The underlying principle of latent class analysis is to search for an unobserved latent categorical variable that explains the association among a set of observed variables, in our study depression symptoms and cognitive functioning (). The latent class model has two advantages. First, the model does not demand that we assign patients to a specific class, as each patient receives an estimated probability of class membership. In other words, misclassification error is minimized. Second, the latent class model also lets us, simultaneously, take into account baseline covariates other than the key variables under study in assessing their independent relationship to depressive symptoms and cognitive function, namely APOE-ε4 status.
Figure 1 Latent class analysis of depressive symptoms and cognitive functioning. Note: Data were gathered from the Spectrum Survey, 2001-2003. APOE-ε4 = ε4 allele of apolipoprotein E, MMSE = Mini-Mental State Examination, HVLT = Hopkins Verbal (more ...)
Our purpose was to conduct analysis through the use of latent class models to employ a new way to look at the relationship between patterns of depressive symptoms and specific domains of cognition, and how these patterns might be related to the APOE-ε4 allele. In addition, because APOE genotype was a part of our analysis, we wanted to examine a range of depressive symptoms and not assume the genotype would be associated with meeting the diagnostic criteria for major or minor depression. We recognize the exploratory nature of our analyses but took the opportunity afforded by a primary care study which included genotyping of patients. We hypothesized that we might find a class of depression related to impaired cognitive performance that is associated with the APOE-ε4 allele. Previous work by Yen et al. [15
] used the latent class model to identify potential subtypes of depression in older adults in relation to the APOE-ε4 allele but defined their latent classes using depressive symptoms alone, while we used measures of both depressive symptoms and specific domains of cognition. Our study employed both the Centers for Epidemiologic Studies Depression (CES-D) scale and the Composite International Diagnostic Interview (CIDI) Depression Section, along with four different measures of specific domains of cognition, thereby providing an especially broad picture of respondent symptoms of depression and cognitive functioning. We were able to use the CIDI to define the latent classes while using the CES-D to aid with clinical interpretation of the latent classes. In summary, we carried out a latent class analysis to examine the heterogeneity of the depressive syndrome in late life that pertains to cognitive impairment and how these latent classes might be related to the APOE-ε4 allele.