The Heritability of Subjective Cognitive Complaints in Older Australian Twins

Abstract

Background:

Subjective cognitive complaints (SCCs) may be a precursor to mild cognitive impairment (MCI) and dementia.

Objective:

This study aimed to examine the heritability of SCCs, correlations between SCCs and memory ability, and the influence of personality and mood on these relationships.

Methods:

Participants were 306 twin pairs. The heritability of SCCs and the genetic correlations between SCCs and memory performance, personality, and mood scores were determined using structural equation modelling.

Results:

SCCs were low to moderately heritable. Memory performance, personality and mood were genetically, environmentally, and phenotypically correlated with SCCs in bivariate analysis. However, in multivariate analysis, only mood and memory performance had significant correlations with SCCs. Mood appeared to be related to SCCs by an environmental correlation, whereas memory performance was related to SCCs by a genetic correlation. The link between personality and SCCs was mediated by mood. SCCs had a significant amount of both genetic and environmental variances not explained by memory performance, personality, or mood.

Conclusion:

Our results suggest that SCCs are influenced both by a person’s mood and their memory performance, and that these determinants are not mutually exclusive. While SCCs had genetic overlap with memory performance and environmental association with mood, much of the genetic and environmental components that comprised SCCs were specific to SCCs, though these specific factors are yet to be determined.

Keywords

Affect Alzheimer’s disease cognitive aging depression human genetics memory disorders neuroticism personality twin study

INTRODUCTION

Older adults commonly complain of having difficulties with their memory and other cognitive abilities. Often, those complaints are not associated with objective decline beyond normal ageing and are therefore known as subjective cognitive complaints (SCCs). Thus, SCCs are the self-experience of decline in cognitive capacity compared to previous ability, which is unrelated to an acute event [1, 2]. SCCs are common in older adults and are often present even when an individual performs within a normal range on cognitive tasks [2].

There is some evidence that SCCs may be a prelude to Alzheimer’s disease (AD). SCCs are associated with an increased risk of future objective cognitive decline [3 –5] and a three- to six-fold increased risk of AD [4 –7]. Several studies have confirmed that individuals with SCCs are more likely to have the early pathological changes associated with AD [8, 9], particularly accumulation of amyloid-β and tau proteins in the brain and cerebrospinal fluid [10 –12]. Therefore, SCCs may be a genuine reaction to a nuanced decline that is too subtle to be detected by neuropsychological tests, but noticeable to the individual experiencing it.

The influence of mood and personality on SCCs

Factors other than true cognitive decline may influence an individual’s complaints. SCCs tend to vary as a function of an individual’s characteristics. Mood disorders, such as depression and anxiety, and personality traits, especially neuroticism, openness, and conscientiousness, have been shown to be associated with SCCs [13 –17]. Thus, some researchers have argued that SCCs should be regarded as simply a proxy for an individual’s tendency towards complaining behaviors rather than their own pre-MCI stage of dementia [18]. However, it has been shown that individuals with SCCs and more depressive symptoms have an increased risk of progression to AD, and that this risk may increase further still if that same individual also scores high on neuroticism [19].

The twin design

The twin design allows researchers to differentiate between genetic and environmental contributions to a phenotype by comparing the concordance of the phenotype in monozygotic (MZ) and dizygotic (DZ) twins raised together. Whereas MZ twins share all their genes, DZ twins only share half. MZ and DZ twins raised together are both assumed to share a common environment. Therefore, the phenotypic variance of a trait can be divided into variance explained by genetic versus environmental effects. The heritability (h ²) of the phenotype can be estimated by comparing the interclass correlations of MZ twin pairs with those of DZ twin pairs.

Structured equation modelling can then be used to differentiate between additive genetic effects (A), environmental effects common to both individuals in a twin pair (C), and environmental effects unique to each individual in a twin pair (E). A model containing all three A, C, and E effects is known as an ACE model, a model containing only genetic and unique environmental effects is known as an AE model, and so forth [20]. Genetic effects are known as additive because the twin design can only estimate the sum of all the genetic influences on a trait and cannot differentiate the effects of specific genes. The aim is not to specify the genes influencing a trait, but the relative contribution of genetics and the environment.

Univariate ACE models estimate the influence of each of these effects on a single phenotypic trait, but it is also possible to estimate how these effects overlap across several phenotypic traits using multivariate models [21]. The most common multivariate ACE model is known as a Cholesky model, which estimates a set of ACE effects for each variable and all other variables downstream [21]. These models allow us to estimate the relative contribution of genetics and the environment across multiple traits at the same time, including how the A, C, and E effects overlap across the different traits.

Genetic relationships between SCCs, mood, and personality

One approach to exploring the relationship between SCCs, cognitive function, mood, and personality is to examine the genetic relationships between these constructs. Mood and personality are both heritable to a certain extent and are both highly correlated with SCCs. There is a well-documented genetic effect on personality traits, as captured by the Five Factor Model [22]. Personality traits appear to be consistent past the age of 50 and are predominantly attributable to genetic influences [23]. Personality is a major factor in wellbeing in aging, as personality traits have been found to predict subsequent changes in individuals’ attitudes toward their own aging as well as mental and physical wellbeing [24].

Mood is also heritable to some extent. Most studies of older twins have demonstrated significant genetic influence on symptoms of depression, reporting modest to moderate heritability ranging from 25% to 37% [25 –29]. However, another study suggested that shared rearing environment, not shared genetics, is the driving force behind the greater similarity in mood between MZ twins than between DZ twins [25].

Therefore, the genetic contributions of personality and mood to SCCs need to be considered. If SCCs were not heritable after accounting for the genetic influence of personality and mood, this would suggest that SCCs are related more to tendencies towards complaints and worry, as opposed to being a marker of very early dementia-related cognitive decline.

Heritability of SCCs using the twin design

Few studies have investigated the heritability of SCCs. Evidence of heritability is important given SCCs are a key criterion of MCI, which is considered an early stage of pre-clinical AD, and the well-established genetic basis for AD. Thus, if SCCs potentially represent the earliest pre-clinical stage of AD, where AD is known to be heritable, one may assume SCCs share a similar genetic contribution. Two studies of Swedish twins aged over 65 years examined the relative contributions of genetic and environmental factors to SCCs. In both studies, SCCs and associated comorbidities (mental, musculoskeletal, respiratory, and urological diseases) were influenced more by environmental factors, such as education, than by shared genetic factors [30, 31]. In contrast, a pilot study of twins from the Older Australian Twins Study (OATS) suggested self-reported SCCs are moderately heritable, controlling for age, sex, and education, but not controlling for personality or mood. Similar to the Swedish studies, though, education had a significant effect on self-reported SCCs [32]. Thus, it is unclear what role genetics play in SCCs.

The aim of the current study was to examine the heritability of SCCs in older adults and to estimate the relative contributions of genetic factors, shared and unshared environmental factors, mood, and personality on SCCs in pairs of MZ and DZ twins. We hypothesized that because SCCs are conceptually a preclinical stage of AD, which is known to have a genetic component, SCCs would therefore also be heritable. Further, we hypothesized that genetic correlations would be found between SCCs, mood, personality, and objective measures of memory performance. Consistent with the hypothesis that SCCs are indeed a pre-MCI stage of dementia, and not merely an indication of certain personality traits and current mood (i.e., a person who is “worried well"), we hypothesized that SCCs would remain heritable over and above the well-known influence of mood and personality as well as SCCs relationship with objective memory performance.

MATERIALS AND METHODS

Study cohort

Participants were 640 participants (320 twin pairs) from the OATS aged 65 to 92 (M = 71.25, SD = 5.66), living in the Australian states of New South Wales, Victoria, and Queensland [33, 34].

Sixteen participants across 14 twin pairs received a diagnosis of dementia. Eight were aware of at least a potential diagnosis at the time of assessment. All 14 twin pairs were excluded from analysis, resulting in a final sample of 612 participants (169 MZ twin pairs and 137 DZ twin pairs).

Measures

Supplementary Figure 1 shows an overview of the key measures of interest collected at each time point for each cohort, including the number of participants with valid data for all independent and dependent variables at each time point.

Subjective cognitive complaints

Participants’ SCCs were measured using two questions asked during a telephone interview: 1) “Have you noticed difficulties with your memory?” and for those who responded yes, 2) “Are you concerned about your memory?” Based on their response to these questions, participants were given an SCCs score of 0 (no difficulties), 1 (noticed difficulties, no concern), or 2 (noticed difficulties and concerned). Participants were asked these questions at each wave or during a follow-up telephone interview 12 months after each wave (W1a; Supplementary Figure 1).

Mood and personality

Mood was assessed at each wave using the Kessler Psychological Distress Scale (K10) [35]. The K10 is a 10-item measure, which asks participants how often they have experienced various emotional states (‘how often did you feel tired for no reason?) on a five-point Likert scale (1 –“none of the time” to 5 –“all of the time”). Scores are summed and range from 10 –50, with higher scores indicating more psychological distress. Scores of 20 or above indicate at least a mild mental disorder, and scores of 30 or above indicate a severe mental disorder [36].

For this study, the majority of K10 scores were collected contemporaneously with SCCs (n = 475); however, for a small subset of participants (n = 114), SCCs were ascertained 12 months (Wave 1a) after the K10 was administered at baseline (Wave 1). There was no contemporaneous measurement of K10 and SCCs for this small subset of participants, so prior mood was used instead.

Personality was assessed using the NEO Five-Factor Inventory NEO-FFI-R [37], administered at Wave 1. The NEO-FFI-R is a 60-item questionnaire that assesses the big five personality traits of Extraversion, Agreeableness, Openness, Neuroticism, and Conscientiousness, where 12 items relate to each of the five personality traits. Participants were asked to rate the degree to which they agree with each statement as it relates to their own beliefs or attributes on a 5-point scale, with higher scores indicating a higher prevalence of each personality trait. One hundred participants did not complete the NEO-FFI-R in our sample.

Memory performance

Participants completed a comprehensive neuropsychological assessment, covering six cognitive domains: Attention/Processing speed, Memory, Verbal Memory, Language, Visuo-spatial, and Executive Function [33]. Scores for memory performance were formed by averaging the z-scores from three episodic memory tests: Logical Memory Story A delayed recall [38, 39]; Rey Auditory Verbal Learning Test (RAVLT) total learning, short-term and long-term delayed recall [39] and Benton Visual Retention Test –Recognition Format [40], followed by calculating the z-score of their mean to form the Memory domain as described in Lee, et al. [41].

Statistical analysis

SCCs were used as an ordinal variable (0, 1, and 2) and all other variables were derived on continuous scale and were inverse normal transformed for behavioral genetics structural equation modelling (SEM). For both univariate and multivariate SEM, the Cholesky model [42] with additive (A), common environmental (C), and unique environmental (E) latent factors model was fitted first and compared with the model restricted to A and E components (AE). The effects of common environment did not add any explanatory value to the model, and so the final model included only the additive genetic effects and unique environmental influences not shared between twins. Akaike information criteria (AIC) [43, 44] and p values were examined to choose the best fit model (ACE versus AE). Ninety-five percent confidence intervals for all the parameter estimates were obtained. Participants’ gender, years of education, and age at the time of assessment were included in all analyses as covariates. All the statistical analyses were performed using the R (4.0.0) package [45]. Estimates of heritability and genetic correlations were obtained using the R statistical package OpenMx version 2.18.1 [46] and umx [42]. T tests and chi-square tests were used to compare sample characteristics between MZ and DZ twins and the p values were obtained based on 5000 permutations of the sample labels.

RESULTS

Table 1 shows a comparison of MZ and DZ twins’ memory performance, Neuroticism, Conscientiousness, K10 mood scores, SCCs, sex, age, recruitment site, and years of education. No significant differences between MZ and DZ twins on all the variables were found (p≥0.05).

Table 1

Participant characteristics

Characteristic	Missing n MZ/DZ	MZ (n = 338) n (%) or mean (SD)	DZ (n = 274) n (%) or mean (SD)	chi-square/t-statistic	p
Female	0/0	208 (61.53%)	180 (65.69%)	0.95	0.277
Age	0/0	70.95 (5.78)	70.79 (5.14)	–0.36	0.718
Years education	0/0	11.70 (3.42)	11.73 (3.56)	0.12	0.908
Memory	14/15	0.12 (0.95)	0.09 (0.98)	–0.25	0.794
Neuroticism	98/74	45.46 (9.46)	45.57 (8.79)	0.12	0.894
Conscientiousness	98/74	49.39 (8.54)	48.85 (9.26)	–0.63	0.535
Extraversion	98/74	50.18 (8.64)	49.94 (7.85)	–0.30	0.764
Openness	98/76	47.70 (6.97)	47.98 (7.46)	0.40	0.690
Agreeableness	98/74	50.18 (7.78)	50.67 (8.07)	0.65	0.520
K10 Mood Score	17/4	11.25 (3.87)	11.06 (3.39)	–0.64	0.521

MZ, monozygotic; DZ, dizygotic. N = 612; 169 MZ twin pairs and 137 DZ twin pairs.

We performed exploratory univariate and pair-wise bivariate analyses of the variables to select the final set of variables for the full multivariate model. The univariate heritability estimates of SCCs, memory performance, NEO-FFI personality subscales, and K10 mood scores are presented in (Table 2). Three NEO subscales (Extraversion, Openness, and Agreeableness) were not significantly associated with SCCs in pairwise comparisons (see Supplementary Table 1) and therefore only Neuroticism and Conscientiousness were included in further analyses. Five-variable univariate and multivariate AE and ACE models were compared for model fit (Table 2 and Supplementary Table 2, respectively). Given its lower AIC, the multivariate AE model was chosen over the ACE model in both instances.

Table 2

Model fit for univariate ACE and AE models and the univariate estimates for heritability ( $h_{a}^{2}$ ), environmental variance ( $h_{e}^{2}$ ), and intraclass correlations for monozygotic twin pairs (MZICC) and dizygotic twin pairs (DZICC), with 95% CI

Variable	AIC for	AIC for	p ACE	MZICC	DZICC	$h_{a}^{2}$	$h_{e}^{2}$
	ACE	AE	versus AE
Memory performance	1488.92	1486.92	1	0.56	0.28	0.56	0.44
				(0.45, 0.66)	(0.22, 0.33)	(0.45, 0.66)	(0.34, 0.55)
Neuroticism	1210.62	1208.62	1	0.39	0.19	0.39	0.61
				(0.24, 0.52)	(0.12, 0.26)	(0.24, 0.52)	(0.48, 0.76)
Extraversion	1234.97	1232.97	1	0.22	0.11	0.21	0.78
				(0.06, 0.37)	(0.03, 0.18)	(0.06, 0.37)	(0.63, 0.94)
Openness	1133.97	1132.20	0.632	0.38	0.19	0.38	0.62
				(0.22, 0.52)	(0.11, 0.26)	(0.22, 0.52)	(0.48, 0.78)
Agreeableness	1221.54	1219.72	0.667	0.26	0.13	0.26	0.74
				(0.09, 0.42)	(0.04, 0.21)	(0.09, 0.42)	(0.58, 0.91)
Conscientiousness	1221.19	1221.23	0.154	0.34	0.17	0.34	0.66
				(0.18, 0.49)	(0.01, 0.24)	(0.18, 0.49)	(0.52, 0.82)
K10 Mood	1671.08	1669.08	1	0.18	0.09	0.18	0.82
				(0.05, 0.31)	(0.02, 0.15)	(0.05, 0.31)	(0.69, 0.95)
SCCs	1117.45	1116.00	0.459			0.33	0.67
						(0.15, 0.49)	(0.51, 0.85)

Estimates in columns 5 to 8 are based on AE model. ACE, Model containing additive genetic (A), common environmental (C) and unique environmental (E) effects; AE, Model containing additive genetic and unique environmental effects; AIC, Akaike information criteria.

Univariate heritability

The univariate estimates of heritability (h²) for memory performance, Neuroticism, Conscientiousness, K10 mood score, and SCCs are shown in Fig. 1. Univariate estimates of environmental variance and intraclass correlations for MZ and DZ twins are shown in Table 2. A third of the variance in SCCs was explained by genetic factors (0.33 [0.15, 0.49]). Both personality subscales, Neuroticism and Conscientiousness, were heritable to a similar extent and to SCCs (0.39 [0.24, 0.52], and 0.34 [0.18, 0.49], respectively). Memory performance was the most heritable of the five variables included in the model, with 56% of the variance explained by genetic factors (0.56 [0.45, 0.66]). K10 mood had the lowest heritability (0.18 [0.05, 0.31]), with 82% of the variance in K10 being explained by unique environmental factors (0.82 [0.69, 0.95]).

Fig. 1

Univariate heritability estimates (h ²) of Memory Performance, Neuroticism*, Conscientiousness*, K10 Mood Score, and SCCs. Error bars represent 95% CI. *NEO-FFI personality subscales

Multivariate AE Cholesky model

The heritability estimates obtained under the five-variable AE model were consistent with the univariate heritability estimates shown in Fig. 1. The five-variable AE Cholesky model path diagram is shown in Fig. 2.

Fig. 2

Path diagram of 5-variable Cholesky AE model. A1 to A5 signify latent additive genetic factors contributing to Memory Performance, Neuroticism, Conscientiousness, K10 Mood Score, and SCCs, respectively. E1 to E5 signify latent unique environmental factors contributing to Memory Performance, Neuroticism, Conscientiousness, K10 Mood Score, and SCCs, respectively. Paths indicate associations between latent factors and variables. Boldface indicates significant associations according to 95% CI. Memory, objective memory performance; NEO_N, neuroticism subscale score of NEO-FFI; NEO_C, conscientiousness subscale score of NEO-FFI; K10, K10 mood score, in which higher scores indicate poorer mood or higher psychological distress.

Significant additive genetic paths were shared between memory performance and SCCs (A1), memory performance and Conscientiousness (A1), and between Neuroticism, Conscientiousness, and K10 mood (A2). The only other significant additive genetic factor in SCCs was unshared with the other variables (A5), meaning SCCs had a significant genetic contribution not explained by memory performance, mood, or personality.

Significant environmental factors were common to K10 mood and SCCs (E4), and between neuroticism, conscientiousness and K10 mood (E2). The only other significant environmental influence on SCCs was the environmental factor that was not shared with other variables (E5). Thus, SCCs had a small genetic association with memory performance and a small environmental association with K10 mood but was mostly explained by genetic and environmental factors specific to SCCs. There was no significant genetic or environmental influence between the two personality factors (neuroticism and conscientiousness) and SCCs.

Phenotypic, environmental, and genetic correlations

Table 3 shows genetic, environmental, and phenotypic correlations between memory performance, Neuroticism, Conscientiousness, K10 mood, and SCCs. The phenotypic correlation between mood and Neuroticism was significant, such that individuals with poorer mood tended to be higher in Neuroticism. The phenotypic correlations between Conscientiousness and Neuroticism and SCCs were all significantly negative, such that individuals who were more conscientious tended to be less neurotic and had lower SCCs. K10 mood was significantly correlated with SCCs, with those with poorer mood (higher K10 scores) having higher SCCs.

Table 3

Genetic, environmental, and phenotypic correlations with 95% confidence intervals

Type of Correlation	Variables	Memory performance	Neuroticism	Conscientiousness	K10 mood
Phenotypic Correlation	Neuroticism	–0.10
		(–0.19, 0.00)
	Conscientiousness	0.05	–0.37^*
		(–0.05, 0.14)	(–0.45, –0.28)
	K10 mood	–0.01	0.34^*	–0.12^*
		(–0.08, 0.08)	(0.25, 0.42)	(–0.21, –0.02)
	SCCs	–0.04	0.08	–0.07	0.19^*
		(–0.14, 0.07)	(–0.03, 0.19)	(–0.18, 0.04)	(0.10, 0.28)
Environmental Correlation	Neuroticism	–0.03
		(–0.20, 0.14)
	Conscientiousness	–0.16	–0.35^*
		(–0.23, 0.02)	(–0.48, –0.20)
	K10 mood	0.06	0.25^*	–0.03
		(–0.08, 0.20)	(0.10, 0.38)	(–0.18, 0.13)
	SCCs	0.16	–0.09	0.09	0.23^*
		(–0.02, 0.14)	(–0.28, 0.11)	(–0.11, 0.28)	(0.07, 0.38)
Genetic Correlation	Neuroticism	–0.17
		(–0.40, 0.08)
	Conscientiousness	0.29^*	–0.40^*
		(0.03, 0.60)	(–0.65, –0.10)
	K10 mood	–0.11	0.61^*	–0.40
		(–0.48, 0.20)	(0.23, 1.00)	(–0.92, 0.05)
	SCCs	–0.28^*	0.36^*	–0.37	0.10^*
		(–0.59, –0.01)	(0.02, 0.77)	(–0.83, 0.02)	(0.00, 0.56)

Asterisks (*) indicate significant at 95% confidence interval.

The environmental correlations between Neuroticism and Conscientiousness and between Neuroticism and K10 mood scores were similar to the phenotypic correlations, suggesting that the environmental factors contributing to Neuroticism also contributed to lower Conscientiousness and poorer mood. The only other significant environmental correlation was between SCCs and K10 mood, which was positive, suggesting that whatever environmental factors contributed to poorer mood also contributed to SCCs.

There were significant genetic correlations between SCCs and memory performance (negative), Neuroticism (positive), and K10 mood score (positive). Therefore, the genetic factors contributing to poorer memory, higher Neuroticism, and poorer mood also contributed to higher SCCs. Memory scores were also positively genetically correlated to Conscientiousness, suggesting that the genetic factors contributing to higher Conscientiousness also contributed to better memory. The genetic correlations between Neuroticism and Conscientiousness, and Neuroticism and K10 mood followed the same pattern as the phenotypic and environmental correlations. However, the genetic correlation between Neuroticism and K10 mood score was numerically higher than the phenotypic and environmental correlations. The genetic factors contributing to higher Neuroticism also contributed to lower Conscientiousness and poorer mood.

DISCUSSION

SCCs can be indicative of future MCI and dementia [1, 2], as well as a symptom of the “worried well” whereby individuals misinterpret normal age-related memory decline that is part of healthy aging due to individual personality and mood factors [18, 19]. These determinants of SCCs are not mutually exclusive and cannot be fully differentiated in a cross-sectional study. Nevertheless, our study shows that SCCs appear to be low to moderately heritable, with contributions from genetic and environmental factors. Memory performance had genetic overlap with the presentation of SCCs, such that the genetic component that increases individuals’ risk of impaired memory also increases their SCCs. Thus, poor memory may indeed be a significant factor in the development of SCCs, which further supports its conceptualization as a pre-dementia syndrome. However, there was a large genetic contribution to SCCs that was specific to SCCs only, indicating that a large proportion of the genetic underpinnings for SCCs are not related to memory, personality, or mood.

We found that objective memory performance is moderately genetically correlated with lower SCCs and that higher Neuroticism and poor mood are genetically linked to SCCs, lower Conscientiousness, and to each other, but not to memory performance. However, the AE Cholesky model showed that the genetic factors influencing high Neuroticism contributed to poorer mood but not SCCs. In addition, there was a significant environmental correlation between SCCs and mood. Therefore, the genetic association between Neuroticism and SCCs may be moderated by mood. These findings, plus the lack of genetic links between memory performance and Neuroticism, suggest that SCCs could be a symptom of older individuals responding negatively to normal age-related cognitive decline. However, like memory performance, this relationship does not fully explain SCCs.

Our findings do not offer any specific insights into the risk of AD based on SCCs. The genetic link between SCCs and memory performance suggests there may be common genetic factors between poor memory and SCCs. However, we cannot say for sure whether or not this relationship is linked to greater risk of AD. In a follow-up study in its early stages, we aim to examine the relationship between SCCs and polygenic risk scores for AD to help us understand whether the genetic underpinnings behind SCCs are the same as those behind future AD.

The environmental association between mood and SCCs also does not provide any specific insights into AD based on SCCs. Chronic stress and poor mood are risk factors for AD, possibly due to the long-term impact of cortisol on the brain [47, 48]. However, our findings are cross-sectional, and are, therefore, not direct evidence supporting this relationship. We cannot extrapolate that our participants with SCCs will develop AD in the future. All we can conclude about our findings is that there does appear to be a relationship between SCCs and poor mood, which appears to be driven by environmental factors.

Strengths and limitations

A strength of this study was the use of the twin design. Twins allow us to differentiate between the contributions of genetic and environmental factors, both common and unique. We found that the AE model, containing only genetic and unique environmental effects, was preferred to a full ACE model additionally containing common environment effects due to a lower AIC and greater parsimony. However, it should be noted that the AE model fit was not significantly better than the ACE model fit. Thus, we cannot conclude that there are no effects of common environment on SCCs; only that including these effects did not give us more explanatory power than that provided by the AE model.

There are limitations to this study that must be acknowledged. SCCs were measured as a three-level ordinal variable, in which “0” indicated no reported memory difficulties, “1” indicated memory difficulties but no concern, and “2” indicated memory difficulties and concern about those difficulties. It is unclear whether the results reflected the experience of memory difficulties or their concern about these difficulties. It is also unclear whether those with concern had more severe memory difficulties, experienced a greater decline, or were simply more emotionally reactive to their memory difficulties than those without concern. Only 62 participants, about 10% of the total study sample, reported difficulties plus concern, compared to 245 participants, about 40%, who reported difficulties without concern. Thus, the sub-sample of “concerned” individuals might not have been adequate to be analyzed separately from those who reported memory difficulties only. The genetic associations between memory performance and SCCs may indicate that it is indeed the memory difficulties, rather than concern that drove the study findings. Given that concern is a major theoretical component in subjective cognitive complaints, the study findings need to be interpreted with this in mind.

In addition, one hundred participants did not complete the NEO-FFI, meaning that the conclusions drawn about the links between personality and other variables were based on smaller sample than the conclusions drawn about the links between memory performance, mood, and SCCs.

The current study also has limitations in terms of statistical power. Looking at the p-values of the comparison between ACE and AE models, for most parameters, our sample has very little power to detect the shared environment component (C) and hence AE model was found to be the best-fitting model. A post-hoc power calculation [49] using the range of observed parameters shows that the power for the C component varies from 1% to 75%. However, for our sample size under AE model, heritability estimates of 0.36 and 0.45 would respectively be detected with 80% power at a level of significances of 0.05 and 0.00625 (accounting for 8 multiple tests).

The statistical power limitations also meant that we were unable to perform longitudinal analysis, even though OATS is a longitudinal study. However, longitudinal analysis, which would demonstrate how SCCs change over time, or to predict whether SCCs later lead to MCI or dementia, was beyond the scope of this study. This study focused on the heritability of SCCs, and the genetic versus environmental influence of mood and personality on SCCs.

This study examined memory performance but no other cognitive domains, such as processing speed and executive function that may also contribute to SCCs [50, 51]. Normal ageing is associated with slowed processing speed and some change in reasoning and executive function, such as working memory, which does not necessarily indicate vulnerability to dementia [51, 52]. Therefore, a definitive study of the heritability of SCCs would require these cognitive domains to be included. However, our questions on SCCs specifically referred to subjective changes in memory, and so memory performance was deemed to be the most relevant measure of objective cognition for our study.

Our method of measuring SCCs differs from other studies that specifically ask about change or decline in memory over time [31] or ask for peer comparisons [53]. The first question “have you noticed difficulties with your memory?” does not imply a decline, only an impairment, or even an acknowledgement of the imperfect nature of healthy memory. However, the follow-up question, asked after a “yes” response to the first question, “are you concerned about your memory?”, captures the information that would be gleaned from asking about either decline or peer comparisons. Arguably, participants would not be concerned about their memory if they felt it had not declined, or if they didn’t believe their memory was worse than age-matched peers’. In fact, concern about impairment is more highly predictive of progression to AD than subjective memory impairment alone [7 , 55]. Concern has also been found to be associated with biomarkers for AD, but subjectively worse memory than age-matched peers has not [56]. Future studies should examine more closely the genetic and biological factors behind SCCs. Examining the relationship between heritability of brain volume, AD biomarkers, and SCCs was beyond the scope of this study. In addition, future studies using the OATS sample may be able to ascertain whether a higher polygenic risk score (PRS) for AD is also found in those with SCCs, considering higher PRS for AD increases the likelihood of having amnestic-MCI [57]. If a similar link can be found for SCCs, it would strengthen the claim that SCCs are a prelude to AD. In combination, these data would further elucidate of the biological basis underpinning SCCs and the theoretical link between SCCs, MCI, and AD.

Conclusions

SCCs appear to be moderately heritable. Although there was a significant genetic contribution to memory performance, a large proportion of the genetic variance in SCCs were not explained by memory performance, personality, and mood. The relationships between SCCs, and personality and mood appear to be shaped primarily by environmental factors. SCCs are a complex phenomenon, which do not appear to have just one cause or a single trajectory. Nevertheless, this paper demonstrates that SCCs are not merely due to a tendency towards complaining behavior in the “worried well”. Biological and genetic processes do appear to be involved in SCCs and need to be teased apart in future studies to determine whether they are linked to increased risk of progression to dementia.

Footnotes

ACKNOWLEDGMENTS

We acknowledge the contribution of the OATS research team () to this study. We thank the participants for their time and generosity in contributing to this research.

This research was facilitated through access to Twins Research Australia, a national resource supported by a Centre of Research Excellence Grant (ID No. 1079102) from the National Health and Medical Research Council.

The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008.

FUNDING

This work was supported by a National Health & Medical Research Council (NHMRC) and Australian Research Council (ARC) Strategic Award Grant of the Ageing Well, Ageing Productively Program (ID No. 401162); NHMRC Project (seed) Grants (ID No. 1024224 and 1025243); NHMRC Project Grants (ID No. 1045325 and 1085606); and NHMRC Program Grants (ID No. 568969 and 1093083).

CONFLICT OF INTEREST

The authors have no conflict of interest to report.

DATA AVAILABILITY

The data supporting the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.

The supplementary material is available in the electronic version of this article: .

References

Jessen

, Amariglio

, Buckley

, van der Flier

, Han

, Molinuevo

, Rabin

, Rentz

, Rodriguez-Gomez

, Saykin

, Sikkes

SAM

, Smart

, Wolfsgruber

, Wagner

(2020) The characterisation of subjective cognitive decline. Lancet Neurol 19, 271–278.

Jessen

, Amariglio

, van Boxtel

, Breteler

, Ceccaldi

, Chetelat

, Dubois

, Dufouil

, Ellis

, van der Flier

, Glodzik

, van Harten

, de Leon

, McHugh

, Mielke

, Molinuevo

, Mosconi

, Osorio

, Perrotin

, Petersen

, Rabin

, Rami

, Reisberg

, Rentz

, Sachdev

, de la Sayette

, Saykin

, Scheltens

, Shulman

, Slavin

, Sperling

, Stewart

, Uspenskaya

, Vellas

, Visser

, Wagner

, Subjective Cognitive Decline Initiative Working Group (2014) A conceptual framework for research on subjective cognitive decline in preclinical Alzheimer’s disease. Alzheimers Dement 10, 844–852.

Slavin

, Sachdev

, Kochan

, Woolf

, Crawford

, Giskes

, Reppermund

, Trollor

, Draper

, Delbaere

, Brodaty

(2015) Predicting cognitive, functional, and diagnostic change over 4 years using baseline subjective cognitive complaints in the Sydney Memory and Ageing Study. Am J Geriatr Psychiatry 23, 906–914.

Geerlings

, Jonker

, Bouter

, Adèr

, Schmand

(1999) Association between memory complaints and incident Alzheimer’s disease in elderly people with normal baseline cognition. Am J Geriatr Psychiatry 156, 531–537.

Numbers

, Crawford

, Kochan

, Draper

, Sachdev

, Brodaty

(2020) Participant and informant memory-specific cognitive complaints predict future decline and incident dementia: Findings from the Sydney Memory and Ageing Study. PLoS One 15, e0232961.

Mitchell

, Beaumont

, Ferguson

, Yadegarfar

, Stubbs

(2014) Risk of dementia and mild cognitiveimpairment in older people with subjective memory complaints:Meta-analysis. Acta Psychiatr Scand 130, 439–451.

Jessen

, Wiese

, Bachmann

, Eifflaender-Gorfer

, Haller

, Kölsch

, Luck

, Mösch

, van den Bussche

, Wagner

, Wollny

, Zimmermann

, Pentzek

, Riedel-Heller

, Romberg

H-P

, Weyerer

, Kaduszkiewicz

, Maier

, Bickel

(2010) Prediction of dementia by subjective memory impairment: Effects of severity and temporal association with cognitive impairment. Arch Gen Psychiatry 67, 414–422.

Verfaillie

SCJ

, Tijms

, Versteeg

, Benedictus

, Bouwman

, Scheltens

, Barkhof

, Vrenken

, van der Flier

(2016) Thinner temporal and parietal cortex is related to incident clinical progression to dementia in patients with subjective cognitive decline. Alzheimers Dement 5, 43–52.

Verfaillie

SCJ

, Pichet Binette

, Vachon-Presseau

, Tabrizi

, Savard

, Bellec

, Ossenkoppele

, Scheltens

, van der Flier

, Breitner

JCS

, Villeneuve

(2018) Subjective cognitive decline is associated with altered default mode network connectivity in individuals with a family history of Alzheimer’s disease. Biol Psychiatry Cogn Neurosci Neuroimaging 3, 463–472.

10.

Buckley

, Maruff

, Ames

, Bourgeat

, Martins

, Masters

, Rainey-Smith

, Lautenschlager

, Rowe

, Savage

, Villemagne

, Ellis

(2016) Subjective memory decline predicts greater rates of clinical progression in preclinical Alzheimer’s disease. Alzheimers Dement 12, 796–804.

11.

Perrotin

, La Joie

, de La Sayette

, Barré

, Mézenge

, Mutlu

, Guilloteau

, Egret

, Eustache

, Chételat

(2017) Subjective cognitive decline in cognitively normal elders from the community or from a memory clinic: Differential affective and imaging correlates. Alzheimers Dement 13, 550–560.

12.

, Teunissen

, Spottke

, Heneka

, Düzel

, Peters

, Li

, Priller

, Buerger

, Teipel

, Laske

, Verfaillie

SCJ

, Barkhof

, Coll-Padrós

, Rami

, Molinuevo

, van der Flier

, Jessen

(2019) Smaller medial temporal lobe volumes in individuals with subjective cognitive decline and biomarker evidence of Alzheimer’s disease—Data from three memory clinic studies. Alzheimers Dement 15, 185–193.

13.

Comijs

, Deeg

DJH

, Dik

, Twisk

JWR

, Jonker

(2002) Memory complaints; the association with psycho-affective and health problems and the role of personality characteristics: A 6-year follow-up study. J Affect Disord 72, 157–165.

14.

Slavin

, Brodaty

, Kochan

, Crawford

, Trollor

, Draper

, Sachdev

(2010) Prevalence and predictors of “subjective cognitive complaints” in the Sydney Memory and Ageing Study. Am J Geriatr Psychiatry 18, 701–710.

15.

Buckley

, Saling

, Ames

, Rowe

, Lautenschlager

, Macaulay

, Martins

, Masters

, O’Meara

, Savage

, Szoeke

, Villemagne

, Ellis

(2013) Factors affecting subjective memory complaints in the AIBL aging study: Biomarkers, memory, affect, and age. Int Psychogeriatr 25, 1307–1315.

16.

Wilson

, Schneider

, Arnold

, Bienias

, Bennett

(2007) Conscientiousness and the incidence of Alzheimer disease and mild cognitive impairment. Arch Gen Psychiatry 64, 1204–1212.

17.

Duchek

, Balota

, Storandt

, Larsen

(2007) The power of personality in discriminating between healthy aging and early-stage Alzheimer’s disease. J Gerontol B Psychol Sci Soc Sci 62, P353–361.

18.

Edmonds

, Delano-Wood

, Galasko

, Salmon

, Bondi

(2014) Subjective cognitive complaints contribute to misdiagnosis of mild cognitive impairment. J Int Neuropsychol Soc 20, 836–847.

19.

Terracciano

, Sutin

, An

, O’Brien

, Ferrucci

, Zonderman

, Resnick

(2014) Personality and risk of Alzheimer’s disease: New data and meta-analysis. Alzheimers Dement 10, 179–186.

20.

Rijsdijk

, Sham

(2002) Analytic approaches to twin data using structural equation models. Brief Bioinformatics 3, 119–133.

21.

South

, Jarnecke

(2015) Methodology of quantitative behavior and molecular genetics. In Genetics of Psychological Well-Being: The role of heritability and genetics in positive psychology, Pluess M, ed. Oxford University Press.

22.

Vukasovic

, Bratko

(2015) Heritability of personality: A meta-analysis of behavior genetic studies. Psychol Bull 141, 769–785.

23.

Johnson

, McGue

, Krueger

(2005) Personality stability in late adulthood: A behavioral genetic analysis. J Pers 73, 523–552.

24.

Kornadt

, Siebert

, Wahl

(2019) The interplay of personality and attitudes toward own aging across two decades of later life. PLoS One 14, e0223622.

25.

Gatz

, Pedersen

, Plomin

, Nesselroade

, McClearn

(1992) Importance of shared genes and shared environments for symptoms of depression in older adults. J Abnorm Psychol 101, 701–708.

26.

Johnson

, McGue

, Gaist

, Vaupel

, Christensen

(2002) Frequency and heritability of depression symptomatology in the second half of life: Evidence from Danish twins over 45. Psychol Med 32, 1175–1185.

27.

McGue

, Christensen

(1997) Genetic and environmental contributions to depression symptomatology: Evidence from Danish twins 75 years of age and older. J Abnorm Psychol 106, 439–448.

28.

Jansson

, Gatz

, Berg

, Johansson

, Malmberg

, McClearn

, Schalling

, Pedersen

(2004) Gender differences in heritability of depressive symptoms in the elderly. Psychol Med 34, 471–479.

29.

McGue

, Christensen

(2003) The heritability of depression symptoms in elderly Danish twins: Occasion-specific versus general effects. Behav Genet 33, 83–93.

30.

Caracciolo

, Gatz

, Xu

, Pedersen

, Fratiglioni

(2012) Differential distribution of subjective and objective cognitive impairment in the population: A nation-wide twin-study. J Alzheimers Dis 29, 393–403.

31.

Caracciolo

, Gatz

, Xu

, Marengoni

, Pedersen

, Fratiglioni

(2013) Relationship of subjective cognitive impairment and cognitive impairment no dementia to chronic disease and multimorbidity in a nation-wide twin study. J Alzheimers Dis 36, 275–284.

32.

Catts

, Thalamuthu

, Lee

TCJ

, Sachdev

(2019) Heritability of subjective cognitive decline in older Australian twins. Behav Genet 49, 516.

33.

Sachdev

, Lammel

, Trollor

, Lee

, Wright

, Ames

, Wen

, Martin

, Brodaty

, Schofield

, team

(2009) A comprehensive neuropsychiatric study of elderly twins: The Older Australian Twins Study. Twin Res Hum Genet 12, 573–582.

34.

Koncz

, Thalamuthu

, Wen

, Catts

, Dore

, Lee

, Mather

, Slavin

, Wegner

, Jiang

, Trollor

, Ames

, Villemagne

, Rowe

, Sachdev

(2022) The heritability of amyloid burden in older adults: The Older Australian Twins Study. J Neurol Neurosurg Psychiatry 93, 303.

35.

Kessler

, Barker

, Colpe

, Epstein

, Gfroerer

, Hiripi

, Howes

, Normand

S-LT

, Manderscheid

, Walters

, Zaslavsky

(2003) Screening for serious mental illness in the general population. Arch Gen Psychiatry 60, 184–189.

36.

(2001) Victorian Population Health Survey. Melbourne, Department of Human Services, Victoria.

37.

McCrae

, Costa

(2004) A contemplated revision of the NEO Five-Factor Inventory. Pers Individ Differ 36, 587–596.

38.

Wechsler

(1997) Wechsler Memory Acale. Third edition manual, The Psychological Corporation, San Antonio.

39.

Strauss

, Sherman

, Spreen

(2006) A compendium of Neuropsychological Tests: Administration, Norms and Commentary, Oxford University Press, New York.

40.

Benton

, Sivan

, Spreen

(1996) Der Benton Test, Huber, Bern.

41.

Lee

, Lipnicki

, Crawford

, Henry

, Trollor

, Ames

, Wright

, Sachdev

, Team

(2013) Leisure activity, health, and medical correlates of neurocognitive performance among monozygotic twins: The Older Australian Twins Study. J Gerontol B Psychol Sci Soc Sci 69, 514–522.

42.

Bates

, Maes

, Neale

(2019) umx: Twin and path-based structural equation modeling in R. Twin Res Hum Genet 22, 27–41.

43.

Akaike

(1998) Information theory and an extension of the maximum likelihood principle. In Selected Papers of Hirotugu Akaike, Parzen E, Tanabe K, Kitagawa G, eds. Springer New York, New York, NY, pp. 199–213.

44.

Akaike

(1974) A new look at the statistical model identification. IEEE Trans Automat Contr 19, 716–723.

45.

R Core Team (2020) R Foundation for Statistical Computing, Vienna, Austria.

46.

Neale

, Hunter

, Pritikin

, Zahery

, Brick

, Kirkpatrick

, Estabrook

, Bates

, Maes

, Boker

(2016) OpenMx 2.0: Extended structural equation and statistical modeling. Psychometrika 81, 535–549.

47.

Ouanes

, Popp

(2019) High cortisol and the risk of dementia and Alzheimer’s disease: A review of the literature. Front Aging Neurosci 11, 43.

48.

Bisht

, Sharma

, Tremblay

M-È

(2018) Chronic stress as a risk factor for Alzheimer’s disease: Roles of microglia-mediated synaptic remodeling, inflammation, and oxidative stress. Neurobiol Stress 9, 9–21.

49.

Visscher

, Gordon

, Neale

(2008) Power of the classical twin design revisited: II detection of common environmental variance. Twin Res Hum Genet 11, 48–54.

50.

Burmester

, Leathem

, Merrick

(2016) Subjective cognitive complaints and objective cognitive function in aging: A systematic review and meta-analysis of recent cross-sectional findings. Neuropsychol Rev 26, 376–393.

51.

Newson

, Kemps

(2006) The nature of subjective cognitive complaints of older adults. Int J Aging Hum Dev 63, 139–151.

52.

Salthouse

(2019) Trajectories of normal cognitive aging. Psychol Aging 34, 17–24.

53.

Tandetnik

, Farrell

, Cary

, Cines

, Emrani

, Karlawish

, Cosentino

(2015) Ascertaining subjective cognitive decline: A comparison of approaches and evidence for using an age-anchored reference group. J Alzheimers Dis 48, S43–S55.

54.

Jessen

, Wolfsgruber

, Wiese

, Bickel

, Mosch

, Kaduszkiewicz

, Pentzek

, Riedel-Heller

, Luck

, Fuchs

, Weyerer

, Werle

, van den Bussche

, Scherer

, Maier

, Wagner

, German Study on Aging, Cognition and Dementia in Primary Care Patients (2014) AD dementia risk in late MCI, in early MCI, and in subjective memory impairment. Alzheimers Dement 10, 76–83.

55.

Verfaillie

SCJ

, Timmers

, Slot

RER

, van der Weijden

CWJ

, Wesselman

LMP

, Prins

, Sikkes

SAM

, Yaqub

, Dols

, Lammertsma

, Scheltens

, Ossenkoppele

, van Berckel

BNM

, van der Flier

(2019) Amyloid-β load is related to worries, but not to severity of cognitive complaints in individuals with subjective cognitive decline: The SCIENCe Project. Front Aging Neurosci 11, 7.

56.

Miebach

, Wolfsgruber

, Polcher

, Peters

, Menne

, Luther

, Incesoy

, Priller

, Spruth

, Altenstein

, Buerger

, Catak

, Janowitz

, Perneczky

, Utecht

, Laske

, Buchmann

, Schneider

, Fliessbach

, Kalbhen

, Heneka

, Brosseron

, Spottke

, Roy

, Teipel

, Kilimann

, Wiltfang

, Bartels

, Düzel

, Dobisch

, Metzger

, Meiberth

, Ramirez

, Jessen

, Wagner

(2019) Which features of subjective cognitive decline are related to amyloid pathology? Findings from the DELCODE study. Alzheimers Res Ther 11, 66.

57.

Logue

, Panizzon

, Elman

, Gillespie

, Hatton

, Gustavson

, Andreassen

, Dale

, Franz

, Lyons

, Neale

, Reynolds

, Tu

, Kremen

(2019) Use of an Alzheimer’s disease polygenic risk score to identify mild cognitive impairment in adults in their 50s. Mol Psychiatry 24, 421–430.