Surviving a myocardial infarction (MI) can be a stressful experience that entails challenges in daily life during the recovery period [1] . In research on the relations between chronic stress and adverse outcomes post-MI, moderate to high stress was associated with almost double increased risk of dying compared to low levels of stress [2] . Also, persons who experienced moderate to high stress during the months preceding their MI showed increased probabilities of having angina one year after discharge [2] . There is a lack of research on the influence of stress on post-MI fatigue, but associations between perceived stress and fatigue have been identified in the general population [3] . Because post-MI stress seems to be insufficiently explored, it would be worthwhile to validate an easily administered stress symptom assessment measure and its association with fatigue in a sample of persons treated for MI.

Stress is frequently implicated in medically unexplained fatigue [4] , but there is still limited evidence for the role of stress following MI [5] and its association with post-MI fatigue has not yet been fully explored. During the recovery process after coronary heart disease, a large proportion persons have to deal with fatigue [6] . “Fatigue is an unpleasant symptom that incorporates bodily feelings ranging from tiredness to exhaustion, creating an unrelenting overall condition that interferes with persons’ ability to function to their normal capacity” (p. 527) [7] . Fatigue has been described as a bothersome and common symptom post-MI [8] and has been reported in nearly half of the group four months and two years post-MI [9] [10] . Fatigue has generally been conceptualized as a warning sign indicating a harmful accumulation of stress [11] . One study also showed that fatigue played an important mediating role in the relationship between stress and perceived poor health [12] . High perceived stress, mediated by impaired sleep quality, contributed to fatigue in healthy adults [13] . There are many potential stressors that may be related to acute treatment and long-term recovery post-MI. Initially, persons are often admitted to a coronary care unit with symptoms of, for example, chest pain, sweating, nausea, shortness of breath, fatigue, dizziness or collapse [14] . Some persons described these life-threatening feelings as a “sense of dread” [15] . During recovery, some persons had to deal with emotional symptoms related to re-experiencing the MI, such as nightmares and psychological distress when they are reminded of the event [16] . They could experience anxiety related to insecurity about their future health and longevity [17] and emotional distress such as helplessness [18] . Also, acute stress symptoms and pain as well as hostility could predict vulnerability to stress disorder post-MI [19] . Other factors found to be stressful during recovery from MI involved daily struggles, i.e., how well persons managed to return to full-time work, experiences of extreme tiredness, and the fact that they had to stop smoking [20] . The subjective experience of the seriousness of MI rather than the objective seriousness predicted self-reported stress post-MI [21] .

A single-item measure of stress symptoms has previously been used in research on associations between stress and injury among industry employees [22] , and as a screening tool for assessing the degree of perceived stress in relation to physical activity and aerobic fitness [23] . Moreover, in a study on physical activity and mental health, the item was used to define a group with high perceived stress levels [24] .

Given previous findings showing that perceived stress is an important health issue in patients with MI, it is of great interest to identify an easily administered stress measure that can be used for both clinical and research purposes. In the present study, the single-item measure of stress symptoms was validated in a sample of persons treated for MI. We also assessed whether this single measure is related to symptoms of fatigue in a sample of persons treated for MI.

As showed in Table 1, there were fewer women (23%) than men and the mean age of the persons was 63 years ± 8.3 (range 35 - 75). The majority of persons (81%) were cohabiting, and 44% returned to work after MI. Symptom characteristics in the study population are presented in Table 2.

The convergent validity of the single-item measure of stress symptoms was confirmed by significant correlations, 0.315 - 0.665 (p < 0.001), with both the global PSS-10 score and all separate items of the PSS-10; see Table 3. The correlation analyses showed that the single-item measure of stress symptoms was associated, r = 0.512 (CI: 0.37 - 0.66) (p < 0.001), with the global fatigue score and all four fatigue dimension scores (general fatigue, physical fatigue, reduced activity, and mental fatigue), r = 0.414 - 0.600; see Table 4. Using the square correlation coefficient, one obtains a value that can be interpreted in percent. That is, 26% of the variance in the global fatigue score could be explained by stress. The correlation results have also been presented in a box plot diagram, showing a monotonic relationship (y increases as x increases, but the curve is not completely linear); see Figure 1. The response alternative “no stress at all” showed a wide range of variation in fatigue, meaning

PSS: Perceived Stress Scale, 1N = 141.

¹Pearson’s correlation.

¹Pearson’s correlation.

that quite a few persons who scored high on the global fatigue score also reported low stress levels. Moreover, the response alternatives “to some degree” and “quite a lot of stress” showed small variation, but a high level of fatigue, meaning that persons who chose these response alternatives probably experienced high global fatigue. These findings were supported when the single-item measure of stress symptoms was treated as a dichotomous scale with wider spread in low stress levels; see Figure 2. The independent t-test analysis revealed a significant difference in fatigue levels between the group with high stress and the group with low stress, Table 5.

The main results of the present study showed that a single-item measure of stress symptoms was associated with the well-validated PSS questionnaire, indicating that this single-item measure could be used to assess stress in persons treated for MI. There are many reasons for assessing and recognizing stress symptoms after treatment for MI. Stress is obviously a risk factor for coronary heart disease [5] , and it contributes to pathophysiological changes and alterations in cardiac regulation that affect health outcomes [31] . Increased fear of dying and experiences of helplessness predicted stress status better than objective measures of MI severity did [18] . In interviews with women post-MI, it was found that they struggled to maintain control, and thus stress was generated when they experienced loss of control over daily life activities [20] . Also, persons who experienced moderate to high stress during the months preceding their MI showed increased probabilities of having angina one year after discharge and increased odds of having diminished quality of health [2] . Because MI can be a frightening and

distressing experience [32] , it would seem to be of vital importance to identify persons at risk for experiencing stress during recovery from MI. The present results showed that the single-item measure of stress symptoms was

associated with all of the separate PSS-10 items and with the PSS-10 global score, indicating that the single-item measure showed convergent validity. Using this single-item measure of stress symptoms may be advantageous given its simplicity, low cost, and ease of interpretation. The question was interpreted as being appropriately formulated, which should increase the probability that respondents will understand it correctly and respond as requested, in this case with regard to stress symptoms [33] .

Moreover, the results showed that experience of stress measured by the single-item measure of stress symptoms correlated with the global score of the MFI-20. This means that higher stress levels were associated with higher fatigue levels. This is in line with a study of persons in a non-clinical sample, which showed that high stress levels were related to fatigue severity [13] . In the present sample, the single-item measure of stress symptoms was strongly correlated with all four fatigue dimensions of the MFI-20 (general, mental, physical fatigue and reduced activity) two months post-MI. The direction of the relationship between stress and fatigue is not always clear. In the present study, stress has been hypothesized to cause fatigue, but the opposite direction of causality is also possible, i.e., fatigue has been conceptualized as a warning sign indicating harmful accumulation of stress [11] . The mechanisms underlying the association between stress symptoms and fatigue are as yet unknown [34] . The present study indicates that a single-item stress measurement administered two months post- MI could be used to identify persons at risk for negative stress experiences. Studies have shown that stress levels can change over time post-MI, and therefore, it would seem to be worthwhile to measure stress longitudinally during the recovery period [14] . Early and periodic identification of persons who are at risk for experiencing high stress levels could provide opportunities to personalize support by including stress management in the rehabilitation programme. This could also help in preventing the onset of fatigue or an increase in existing fatigue.

In our analyses of the validity of the single-item measure of stress symptoms, we treated the item as both a continuous and a dichotomous variable and found a significant relationship with fatigue in both cases. However, when using the item dichotomously, we identified a risk for neglecting the fact that some persons responding “no stress at all” could nonetheless be experiencing significant fatigue. In addition, 50% of those who responded “stress to some degree” scored above the total sample median for the global fatigue score. However, this problem should be further investigated by, for example, analysing agreement between the measurements using a statistical method for analysis of paired ordinal data [35] . Also, persons recovering from MI may experience several symptoms, and the present study shows that exploring stress and fatigue simultaneously is justified. According to [36] Miaskowski et al. (2004), symptom management research should evaluate more than one symptom at a time as well as the relationships between symptoms. Therefore, future research could focus on the prevalence of post-MI stress and its relationship with other concurrent symptoms in addition to fatigue.

The present study was cross-sectional and this could be seen as a limitation because stress level can change over time post-MI. Therefore, further validation of the single-item measure of stress symptoms is warranted. Also, the relationship between stress and fatigue has not been explored regarding cause and effect, as such an analysis requires a more complex research design. However, one strength of the study is that the sampling procedure was consecutive, that is, all patients with MI admitted to a rural coronary care unit during one year were invited to participate. Consequently, we consider that the present result contributes knowledge about measurement of post-MI stress and its association with fatigue among persons (<75 years) treated for MI in a rural area.

The single-item measure of stress symptoms was found to provide valid measures of post-MI stress. Moreover, the item was useful in measuring associations between stress and fatigue and can therefore indicate fatigue experiences. Further research should focus on stress and its relationship with other concurrent symptoms post-MI, in addition to fatigue.

No conflict of interest has been declared by the authors.

This work was supported by the Centre for Person-Centred Care at University of Gothenburg (GPCC), Sweden. The GPCC is funded by the Swedish Government’s grant for Strategic Research Areas, Care Sciences (Application to Swedish Research Council no. 333-02) and co-funded by the University of Gothenburg, Sweden.