| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Department of Psychiatry and Neuropsychology (N.J., P.D., J.v.O., I.M.-G.), South-Limburg Mental Health Research and Teaching Network, EURON, Maastricht University, Maastricht, The Netherlands; the Social, Genetic & Developmental Psychiatry Research Centre (F.R.), Institute of Psychiatry, London, United Kingdom; the Faculty of Medicine (C.D.), Center for Human Genetics, Catholic University Leuven, Leuven, Belgium; the Department of Population Genetics (R.V.), Maastricht University, Maastricht; The Netherlands; the Division of Psychological Medicine (J.v.O.), Institute of Psychiatry, London, United Kingdom; and Mondriaan Zorggroep, Section Social Cognition (I.M.-G.), Heerlen, The Netherlands.
Address correspondence and reprint requests to Inez Myin-Germeys, PhD, Department of Psychiatry and Neuropsychology (loc: KAP2), South-Limburg Mental Health Research and Teaching Network, EURON, Maastricht University, PO BOX 616, 6200 MD Maastricht, The Netherlands. E-mail: i.germeys{at}sp.unimaas.nl
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: Two hundred seventy-five female twin pairs (170 monozygotic and 105 dizygotic) participated in this experience sampling study (ESM). ESM is a validated structured diary technique assessing stressors and mood in daily life. Individual emotional stress reactivity was conceptualized as changes in negative affect in relation to appraised subjective stress in relation to daily events. Structural equation modeling was used to fit univariate models. The best fitting model was chosen based on likelihood and parsimony.
Results: Genetic factors (explaining 12% individual differences) and individual-specific environmental factors (explaining 88%) influenced daily life stress reactivity.
Conclusion: The demonstration of a small genetic influence on the dynamic relationship between minor stress and affective response in the flow of daily life sheds light on the gene-environment interactions that contribute to the risk of developing stress related disorders such as depression.
Key Words: gene-environment interaction • daily life stress • experience sampling method • twin study
Abbreviations: ESM = experience sampling method; PA = positive affect; NA = negative affect; MZ = monozygotic; DZ = dizygotic; RMSEA = root mean squared error approximation; SERT = serotonin transporter gene.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Stress research originally focused on the occurrence and frequency of relatively rare stressors such as major life events (e.g., job loss, death of a loved one, or major financial problems) and demonstrated the negative impact of such events on general health (9,12,13). More recent studies, however, suggested that also minor stressors occurring in the flow of daily life may have an unfavorable effect on psychological health. Although the impact of daily hassles may be smaller compared with major life events, they occur much more frequently and therefore can have a substantial effect on general health. Momentary assessment methods have been developed to assess the dynamic relationship between these minor daily hassles and momentary fluctuation in mood states in the flow of daily life (14,15). Research using such momentary assessment methodologies have provided evidence that minor stressors in daily life are associated not only with decreased psychological well-being and increased somatic symptomatology, but also with stress-related disorders such as depression and psychosis (16–19). Everyday life stress can therefore be considered as a risk factor for reduced mental health. In addition, these studies showed that the individual's emotional appraisal of a stressful event, rather than its occurrence as such, is crucial in determining its subsequent impact (20). For example, affective stress responses, in particular momentary changes in negative affect, were found to occur in parallel with the biological, cortisol-mediated stress response indicating that variability in emotional stress reactivity is a crucial marker of individual differences in sensitivity to stress (21,22). These findings are in line with the proposed importance of momentary variations in negative affect as a genetically influenced trait conferring evolutionary advantage in the face of the ever-changing demands and threats facing the individual in the flow of daily life (23).
The aims of the current study were to quantify individual differences in emotional reactivity to stress in the flow of daily life and to determine their underlying sources of variation. Because individual differences can be caused by either genetic factors or by environmental factors, twin studies that enable the separation of these effects are an appropriate means to this end, in particular because there have been no previous studies of the relative contributions of genes and environment to individual differences in emotional stress reactivity in the flow of daily life. In a sample of 275 female twin pairs, we examined the extent to which emotional reactivity to daily stress can be explained by genetic factors and/or by environmental factors.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Momentary Assessments Method
The experience sampling method (ESM) is a structured diary technique to assess subjects in their daily living environment. It is a valid and reliable way to study immediate effects of stressors on mood, reducing biases in recall (14,15,29). Subjects received a digital wristwatch and a set of ESM self-assessment forms collated in a booklet for each day. The wristwatch was programmed to emit a signal ("beep") at an unpredictable moment in each of 10 90-minute time blocks between 7:30 am and 10:30 pm on 5 consecutive days. After every beep, subjects were asked to stop their activity and to fill out the ESM self-assessment forms previously handed to them, collecting reports of thoughts, current context (activity, persons present, location), appraisals of the current situation, and mood. All self-assessments were rated on seven-point Likert scales. Trained research assistants with experience in momentary assessment techniques explained the ESM procedure to the participants during an initial briefing session, and a practice form was completed to confirm that subjects were able to understand the seven-point Likert scale. Subjects could call a telephone number in case they had questions or problems during the ESM sampling period. Subjects were instructed to complete their reports immediately after the beep, thus minimizing memory distortions, and to record the time at which they completed the form. To know whether the subjects had completed the form within 15 minutes of the beep, the time at which subjects indicated they completed the report was compared with the actual time of the beep. A momentary assessment validation study in this sample showed that random sampling procedures with high sampling loads yielded compliance rates in excess of 90% (30). All reports not filled in within 15 minutes after the beep were excluded from the analysis. Previous work (15) has shown that reports completed after this interval are less reliable and consequently less valid. Subjects with less than 17 valid reports were excluded from the analysis.
Momentary Assessment Measures
Emotional stress reactivity was conceptualized dynamically as mood reactivity to minor disturbances in daily life (19,31). The mood measure and the stress measures were derived from the experience sampling reports as described subsequently.
Assessment of Stress
Conforming with previous work (19,31), stress was conceptualized as the subjective appraised stressfulness of distinctive events that continually happen in the flow of daily life. After each beep, subjects were asked if a negative event happened since the last beep. This event was subsequently rated on a seven-point bipolar scale (–3 = very unpleasant, 0 = neutral, 3 = very pleasant). Responses were recoded to allow high scores to reflect stress (–3 = very pleasant, 0 = neutral, 3 = very unpleasant). Response on this item represented event-related stress.
Assessment of Mood
Using the experiencing sampling method, current mood states were assessed with 15 mood adjectives rated on seven-point Likert scales (from 1 = "not at all" to 7 = "very"). Conforming with previous studies (16,31,32), maximum likelihood factor analysis with oblique rotation identified three factors with eigenvalues of, respectively, 4.3, 1.6, and 0.8. First factor was the positive affect (PA) scale, consisting of the items "cheerful," "satisfied," "energetic," and "enthusiastic" (Cronbach's alpha = 0.86). The second factor, the negative affect (NA) scale, consisted of the mood items "insecure," "lonely," "anxious," "blue," "guilty," and "suspicious" (Cronbach's alpha = 0.76). Three adjectives, "not relaxed," "not calm," and "harried," formed the agitation (AG) scale (Cronbach's alpha = 0.72). The items "irritated" and "feeling controlled" had similar loadings on different scales and were excluded to enhance differentiation between the factors.
Because this study examined stress-induced fluctuations in negative affect, only the NA scale was used. For each subject, the weighted mean on each scale was calculated taking the rotated factor loadings of each item into account (factor loadings were: "insecure" = 0.70, "lonely" = 0.51, "anxious" = 0.62, "blue" = 0.65, "guilty" = 0.53, and "suspicious" = 0.57).
Stress Reactivity Measure
To obtain a measure of stress reactivity for each subject, regression analyses with NA as the dependent variable and event-related stress as the independent variable was conducted for each subject separately (over the 50 reports of each subject). The regression coefficient of this equation represented the individual change in NA associated with the subjective appraisal of minor daily stressors and was used in the analyses as the measure of individual stress reactivity.
Analyses
Correlation Analysis
Comparison between monozygotic (MZ) and dizygotic (DZ) intrapair correlations of the stress reactivity measure gives a first impression about the role of genetic and environmental factors. If the MZ intrapair correlation is about twice as high as the DZ intrapair correlation, additive genetic factors (A, the sum of the average effects of the individual alleles at all loci affecting the phenotype) and individual-specific environmental factors (E, environmental influences that are not shared between family members) are likely to play a role in determining individual differences in the trait. If, on the other hand, the MZ intrapair correlation is much higher than twice the DZ intrapair correlation, additive genetic factors (A), dominant genetic factors (D, effect of interacting genes present at the same chromosomal locus), and individual-specific factors (E) are likely to explain the variation within the variable. If no significant difference between the MZ and DZ intrapair correlation is observed, only common environmental factors (C, environmental influences that are shared between family members) and individual-specific environmental factors are supposed to play in a role in explaining the phenotypic variance.
Structural Equation Model Fitting
Based on the results of the correlation analysis, univariate structural equation models (SEM) were applied using the program Mx (36). The goal of univariate genetic model fitting is to decompose the phenotypic variance into the possible sources A, C, and E.
Model parameters are estimated by minimizing a goodness-of-fit statistic (
2) between observed and model-predicted covariances. Raw data analysis was used to handle missing data problems. When analyzing models with raw data, minus twice the log-likelihood (–2*LL) of the data for each observation is calculated. This implies that there is no overall measure of fit, but rather relative measures of fit, because differences in fit function between submodels are distributed as 2. The goodness-of-fit of, e.g., the full ACE model is measured relative to a perfect fitting (saturated) model. In a saturated model, the maximum number of parameters is estimated to describe the correlational structure between variables. For example, for the covariance matrix in one group (MZ or DZ), that would be five parameters: two variances (of twin 1 and twin 2)—one cross-twin covariance and two means.
A nonsignificant 2 value suggests that the model is consistent with the data, whereas a significant 2 value (p < .05) suggests that the model poorly fits the data and can be rejected. Goodness-of-fit of alternative, nested models were evaluated by changes in 2. For example, the fit of a reduced model (e.g., AE) will be better (i.e., the dropped parameter C will be nonsignificant) if the difference in 2 for 1 degree of freedom does not exceed the critical value (at the 0.05 level) of 3.84. Information about the precision of parameter estimates (and their explained variance) was obtained by likelihood-based confidence intervals (CIs) rather than standard errors (37). In addition, the root mean squared error approximation (RMSEA), a goodness-of-fit index independent of sample size, was calculated for each model. RMSEA values below 0.10 indicate a good fit; values below 0.05 indicate a very good fit.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Stress Reactivity Measure
The stress reactivity measure was not normally distributed and log-transformed [ln (stress reactivity coefficient x 100 + 15)] to reduce abnormality.
The mean of the transformed event-related stress reactivity coefficient equaled 2.85 (SD = 0.28).
Analyses
Correlation Analyses
The Spearman MZ intrapair correlation was 0.14 (p = .08), whereas the Spearman DZ intrapair correlation equaled 0.06 (p = .54), suggesting genetic influence on the phenotype. No difference between MZ versus DZ variance was found (SD MZ = 0.27; SD DZ = 0.30, p = .38).
Structural Equation Model Fitting
The first model tested was the fully saturated phenotypic model to which all genetic models were compared to derive a chi-squared fit-index and p value (Table 1). The second model was a full genetic ACE model. The estimate of C in this model was 0 (95% CI: 0–0.2) and therefore, an AE model was tested. This model fitted as good as the full ACE model (
2 [1] = 0, p = 1) and was more parsimonious. Next, the CE and E model were fitted. Although the fits of these models were not significantly worse than the ACE model (respectively, 2 [1] = 0.43, p < .5 and 2 [2] = 2.1, p = .3), both the AIC and the RMSEA suggested that the fit of the AE model was slightly better compared with the fits of the CE and E model. In the AE model, the standardized effect of the additive genetic factor was estimated at 0.12 (95% CI: 0–0.28) (Figure 1).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Gene-Stress Interaction
This was the first twin study ever to examine individual differences in affective response to small, daily life stress. However, a number of recent studies (39–41) investigated patterns of familial transmission of psychological distress in general population samples. A moderate genetic effect has been found for general psychological distress in an adult general population sample (44% of the phenotypic variation explained by additive genetic effects) and in preschool children (additive genetic effects explaining 43% of the phenotypic variation). Our results are in line with these findings, supporting an etiological genetic component in stress reactivity, the size of which is small to moderate.
Because our stress reactivity measures included the interaction between stressor and emotional response, the indication demonstration of a small genetic influence on this dynamic relationship is indicative of gene-environment interaction: not only the environment, but also genes contribute to individual differences in emotional sensitivity to stress (42). Genes therefore codetermine momentary emotional reactivity to stresses in the immediate environment in the flow of daily life. Approximately 12% of the total variance was explained by genetic factors, indicating that genes have a small impact on differences between people with respect to sensitivity for the negative effect of minor daily events on mood. Increased stress reactivity in daily life is associated with increased risk for depression, bipolar disorder, and nonaffective psychotic disorder (19,32). The current study, thus, suggests that the genes that put subjects at risk to develop these disorders do so in part by increasing stress sensitivity in the flow of daily life.
Because genetic factors were found to play a role, it is important to identify the genes that influence daily life stress reactivity. A possible candidate is a functional polymorphism on the serotonin transporter gene (SERT). It has been demonstrated that SERT sites in the human brain are associated with emotional functions (43). In addition, there is evidence that SERT might be involved in stress-related disorders such as psychotic illness (44) and the tendency to develop depression after exposure to major life events (11). The measures used in the current study are quite different in that not major life events, but small stressors in the flow of daily life, were assessed as the exposure and not clinical depression but momentary changes in negative affect as the outcome. A next important step therefore is to examine the role of SERT in moment-to-moment changes in negative affect in the face of minor stressors in daily life. A positive association between a functional SERT polymorphism and daily life stress sensitivity could help to further unravel the genetics of selected traits that enhance environmental adaptation on the one hand but could also increase the risk of disorders such as depression on the other. A functional polymorphism on the MAO-A gene has also been reported to moderate children's sensitivity to maltreatment (45), which makes MAO-A another suitable candidate gene. The MAO-A gene encodes the MAO-A enzyme, which catabolizes several neurotransmitters among which serotonin, MAO-A, and SERT are, therefore, also possible candidates to examine epistasis or gene-gene interaction. However, Caspi et al. found that the moderation of major life events on depression was specific to the SERT polymorphism and independent of the individual's MAO-A gene status (11). This, however, does not preclude interaction using momen-tary assessment measures of stress and mood reactivity, which are quite different from the measures of major life events and clinical depression used in the work reported by Caspi et al. (11).
Although ours and previous findings need to be replicated, they all show evidence for genotype-environment interaction and they even suggest possible gene-gene interaction. This has implications for traditional molecular genetic studies. Instead of searching for genes directly associated with the phenotype, molecular genetic studies may profit from incorporation of environmental factors in the research design and identify genes that are indirectly (through environmental factors) associated with the phenotype (46).
The major part of the total variance of stress reactivity was explained by individual-specific environmental factors (and measurement error). Social support may be an example of such an individual-specific environmental factor. It is well-known that social support influences the way in which individuals react to stressors (47,48), thus in fact representing an environment-environment interaction.
These results should be interpreted in the light of several methodological limitations. First, the sample consisted of female twin pairs in the general population. Because there is some evidence that affective stress responses to daily life stressors differ for men and women (49), the findings may not generalize to males. Simultaneous analysis of male and female MZ and DZ pairs as well as opposite sex pairs will provide an opportunity to study quantitative and qualitative sex differences in affective stress response to daily life stressors. Second, this study focused on gene-environment interaction in daily life. However, one could argue that there is possibly also gene-environment correlation. Thus, the risk of being exposed to stresses in daily life may rise with increasing genetic liability to react stronger to stress. Kendler and colleagues showed that approximately one third of the association between stressful life events and onset of major depression is explained by gene-environment correlation (50); subjects genetically predisposed to depression select themselves into high-risk environments (13,51). To examine this issue, post hoc analyses were carried out using the cotwin control method (50,52). If the correlation between the risk factor (exposure to daily life stressors) and the outcome (negative affect) is mediated by genetic factors predisposing to both the risk factor and (directly or indirectly) the outcome, the within DZ correlation should be higher than the within MZ correlation. The DZ correlation between minor daily stressors and negative affect equalled 0.52 (p < .001) and was indeed somewhat higher than the MZ correlation (rMZ = 0.47, p < .001), suggesting the possibility of a small degree of genotype-environment correlation.
Third, emotional stress reactivity has been defined in terms of emotional reactivity toward subjective stress. The cross-sectional analyses of the data, however, make it impossible to establish a causal relationship. Therefore, the reverse might be true, in that a worse mood might influence the subjective appraisal of the environment. The overall effect, however, would still be for the individual to experience distress associated with an environmental event.
Fourth, structural equation modeling suggested that the AE model was the best fitted model. However, it must be noticed that the power in this sample was too low to discriminate well between the AE and the other models. However, the fit of the AE model was slightly better compared with the fits of the others models and in addition, the AE model is in accordance with the conclusions drawn from the within-pair correlations. Furthermore, the estimate of the common environmental factor in the full ACE model was zero, suggesting that the familiarity in stress reactivity is caused by genetic factors rather than by environmental factors.
Fifth, the naturalistic diary technique used in this study resulted in a number of missing data (26% of missing data). However, it is important to note that ESM investigates subjects in their normal daily life. This means that missing data are inevitable (e.g., when people are asleep or in noisy environments where they do not hear the beep). In addition, regression analysis was conducted for each subject separately on an average of 37 observations per person (with a minimum of 18 observations) resulting in an accurate and precise measure of stress reactivity.
In conclusion, this was the first twin study ever to examine individual differences in affective response to small, daily life stress. Because it was found that genetic factors explained a part, albeit small, of the phenotypic variance, this study provided evidence for genotype-environment interaction in the flow of daily life. Some people are more sensitive to the effect of small stress because of their genes. The remaining part of the phenotypic variance was explained by individual environmental factors.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Received for publication January 3, 2005; revision received December 6, 2005.
DOI:10.1097/01.psy.0000204919.15727.43
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M. C. Wichers, I. Myin-Germeys, N. Jacobs, G. Kenis, C. Derom, R. Vlietinck, P. Delespaul, R. Mengelers, F. Peeters, N. Nicolson, et al. Susceptibility to Depression Expressed as Alterations in Cortisol Day Curve: A Cross-Twin, Cross-Trait Study Psychosom Med, April 1, 2008; 70(3): 314 - 318. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K. C. Gunthert, T. S. Conner, S. Armeli, H. Tennen, J. Covault, and H. R. Kranzler Serotonin Transporter Gene Polymorphism (5-HTTLPR) and Anxiety Reactivity in Daily Life: A Daily Process Approach to Gene-Environment Interaction Psychosom Med, October 1, 2007; 69(8): 762 - 768. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Wichers, I. Myin-Germeys, N. Jacobs, F. Peeters, G. Kenis, C. Derom, R. Vlietinck, P. Delespaul, and J. Van Os Genetic risk of depression and stress-induced negative affect in daily life The British Journal of Psychiatry, September 1, 2007; 191(3): 218 - 223. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Minerva BMJ, May 13, 2006; 332(7550): 1162 - 1162. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
