آیا آلکسیتیمیا تحت تاثیر استرس موقعیتی و یک صفت پایدار مربوط به تنظیم احساسات در آن است؟

Abstract Numerous studies have found alexithymia to be negatively associated with psychological distress. However, the nature of such an association remains unclear and controversial. Whereas some authors consider alexithymia as a stable personality trait constituting a vulnerability factor regarding mental disorders, other authors view it as a defensive mechanism secondary to the occurrence of psychological distress. The present prospective study (12 weeks follow-up) examines the stability of alexithymia in the context of acute changes in the level of psychological distress. Contrary to most previous studies that have examined alexithymia stability after a reduction in psychological distress, this study examines alexithymia stability in the context of increase in psychological distress. Four indicators of stability were considered: absolute, mean-level, rank-order and relative stability. Although not absolutely stable, alexithymia was found to show a high degree of relative stability despite the acute increase in psychological distress.

1. Introduction All human beings experience emotions. However, individuals differ in the way they are interested in their emotional life, and the extent to which they are able to differentiate between their feelings and describe them to others. The concept of alexithymia has been proposed to account for this variability. Alexithymia is a multi-faceted construct comprising (a) a difficulty identifying feelings as well as distinguishing between feelings and the bodily sensations of emotional arousal; (b) a difficulty describing feelings to others; (c) a restricted imagination, as evidenced by a paucity of fantasies; and (d) a cognitive style that is literal, utilitarian, and externally oriented (Taylor & Bagby, 2000). There is a large body of evidence showing that alexithymia is associated with a number of psychiatric and psychosomatic disorders (for overviews, see Corcos and Speranza, 2003 and Lumley et al., 1996). However, the nature of the link between alexithymia and those disorders remains unclear: does alexithymia constitute a vulnerability factor for these disorders or is it merely a reaction accompanying them? This question raises the issue of the stability of alexithymia. According to several authors (e.g., Martínez-Sánchez et al., 1998, Martínez-Sánchez et al., 2003, Pinard et al., 1996, Saarijärvi et al., 2001 and Salminen et al., 1994) alexithymia is best conceptualized as a stable personality trait reflecting a deficit in the cognitive processing of emotional information ( Taylor, 2000). This deficit would result in poor emotional regulation and stress management abilities, thus leading to poor mental health as well as to somatic disorders (indeed, unregulated stress leads to a hyper-activation of the corticotrophin axis, which results in a drop in immune defences, leading to higher somatic vulnerability). Arguments in favour of the view of alexithymia as a stable personality trait, related to one’s mental health level but independent of its variations, are mainly based on longitudinal studies on psychiatric outpatients (e.g., Pinard et al., 1996, Saarijärvi et al., 2001 and Salminen et al., 1994), showing that whereas the level of psychological distress significantly dropped after treatment (even for untreated patients) the level of alexithymia did not change significantly. On the other hand, other authors (e.g., Haviland et al., 1988 and Honkalampi et al., 2000) suggest that alexithymia must be better considered as a state-dependent phenomenon; that is to say, a consequence of personal distress (i.e. anxiety, depression). In such a perspective, alexithymia would be merely a coping mechanism protecting the self against emotional distress associated with situations of intense vulnerability ( Corcos & Speranza, 2003). Arguments in favour of this view are based on longitudinal studies on psychiatric (out)patients, in which mean alexithymia scores remained stable despite a significant drop in psychological distress but in which in-depth analyses revealed that only half of the patients remained in the same alexithymia category at follow-up (e.g., Honkalampi et al., 2000). Albeit suggestive and based on appropriate designs, findings from the aforementioned studies suffer from limitations lying essentially in the type of analysis that has been performed on the data. To understand these limitations, it is imperative to distinguish between the various forms of stability. In this respect, four aspects need to be distinguished: absolute stability, mean-level stability, rank-order stability and relative stability (Roberts and DelVecchio, 2000 and Santor et al., 1997). Absolute stability is reached if every single score remains exactly the same across time or situations. Needless to say, this level of stability is seldom achieved. Nevertheless, although varying, the scores may possess clearly defined arithmetical properties conferring them partial stability despite their variations. First, it is possible that variation is small or that variations at the individual level compensate for one another, with the result that the mean of the group remains stable across time. This kind of stability has been called mean-level stability. Second, it is conceivable that albeit varying, the relative differences between individual scores remain the same across time, which case has been referred to as relative stability. Third, it is also possible that despite a variation of scores across time, the position of the individuals among the group remains the same. This situation has been referred to as rank-order stability. It is noteworthy that the three latter kinds of stabilities are quite independent. Consider for instance the following data. In series like: A (1, 2, 4, 8, etc), B (3, 6, 12, 24, etc.), Kendall’s and Pearson coefficients would be 1, although means would be different. On the other hand, in series like A (1, 2, 3, 4, etc.), B (2, 1, 4, 3, etc.), means would be similar, Pearson coefficient would be quite high whilst Kendall’s tau would be drastically lower. As the various forms of stability refer to different realities, it is important that any study investigating the issue of alexithymia consider all indicators. For instance, as shown in a recent study by Luminet, Bagby, and Taylor (2001) on psychiatric outpatients with major depression, alexithymia scores may decrease following the treatment (absolute change), whereas relative differences among individuals may still remain the same (relative stability). The limitation inherent to the studies by Pinard et al., 1996, Saarijärvi et al., 2001 and Salminen et al., 1994 lie in the fact that conclusions have been drawn on the basis of an incomplete examination of the various indicators of stability. Only mean-level stability was considered. However, it is perfectly conceivable that alexithymia demonstrates mean-level stability and yet absolute, relative and rank-order instability. The weakness of Haviland et al., 1988 and Honkalampi et al., 2000 studies rests in the fact that alexithymia has been analysed in a categorical way, although it is best conceptualized as a continuous dimension (Luminet et al., 2001). Moreover, relative and rank-order stability of alexithymia were not assessed. So notwithstanding a growing literature on alexithymia stability, statistical limitations seem to make any conclusions premature. However, determining alexithymia status would constitute an important step forward. Indeed, it covers much more than the question of the construct validity of alexithymia as a personality trait. Establishing alexithymia’ status is essential to resolve the question of whether it might constitute a vulnerability factor regarding mental disorders. To what degree have previous studies provided elements allowing the conclusion that alexithymia might have played a role in the trouble genesis? The common feature of nearly all prior studies is that they have examined the extent to which alexithymia remained stable when mental health improved. We have already pointed out the statistical limitations of these studies. However, even if new analyses concluded in the stability of alexithymia scores at follow-up, this would not constitute evidence that alexithymia is a stable trait, nor that it preceded the mental disorder. Indeed, it is well-known that some somatic diseases (e.g., Crohn’s disease) develop as a result of stress but persist even when stress has eased ( Reynaert, 1995). In the same vein, we cannot formally exclude from the afore-mentioned studies the possibility that alexithymia developed as a defensive mechanism secondary to psychological distress and that it either persisted beyond or that it declined more slowly than psychological distress. Consequently, settling the issue of whether alexithymia could be secondary to psychological disorders would involve collecting data on the level of alexithymia before the onset of the trouble. If alexithymia scores remain stable and do not parallel psychological distress scores, then it can be concluded that alexithymia is not merely a consequence of psychological distress. Only two studies collected information on the level of alexithymia before the onset of psychological distress ( Martínez-Sánchez et al., 1998 and Martínez-Sánchez et al., 2003). These studies concluded that mean-level alexithymia was stable. Unfortunately, samples were too small to cover the entire distribution of alexithymia scores and not all indicators of stability were considered. In line with Martínez-Sánchez et al. (1998), the present study adopted a prospective design, with the level of alexithymia being measured before the emergence of psychological distress. Since a prospective study on potential psychiatric patients was difficult to implement (it is impossible to determine anticipatively who is going to develop mental disorders and when), we examined the various indicators of stability in a stressful context in which a significant increase in psychological distress is expected. Academic exams were chosen as the stressful context for three reasons. First, the exam period has proven to be the main stress-event among university students ( Kohn & Frazer, 1986). Second, it is the only natural stressful event that all university students encounter, making this prospective study both possible and ethical. Finally, it allowed for a standardisation of the triggering factor: the increase in psychological distress was caused by the same event for everybody.

. Results 3.1. Univariate statistics and bivariate correlations among variables Means, standard deviations, Cronbach’s alphas and zero-order correlations between alexithymia and psychological distress are reported in Table 1. As shown in the table, alexithymia is related to psychological distress both within and across time, thus making the question of the stability of alexithymia highly relevant in a context in which changes in psychological distress are expected. Table 1. Descriptive statistics and correlations among variables under consideration Variable 1 2 3 4 5 6 M SD 1. BSI (1) — 87.44 26.98 2. BSI (2) .64⁎⁎⁎ — 104.47 32.16 3. TAS-20 (1) .31 ⁎⁎ .23† — 48.45 11.25 4. TAS-20 (2) .39 ⁎⁎⁎ .34 ⁎⁎ .74⁎⁎⁎ — 49.51 10.85 Notes. Italic coefficients are Cronbach’s alphas; boldface coefficients are test–retest correlations. BSI = Brief Symptom Inventory; TAS-20 = total alexithymia score. (1) Fulfilled at baseline, (2) fulfilled at follow-up. ⁎⁎ p ⩽ 0.01. ⁎⁎⁎ p ⩽ 0.001. † p < 0.1. Table options 3.2. Examining the various indicators of stability 3.2.1. Absolute and relative stability Absolute stability—also referred to as ipsative stability ( Honkalampi et al., 2001)—and relative stability were examined via Pearson correlation coefficients performed respectively on total alexithymia scores at baseline and follow-up, and BSI scores at baseline and follow-up. Positive and significant correlations were found for total alexithymia scores (r = 0.74, p < 0.001). Although indicating that individual scores are not absolutely stable (Pearson correlations differ from 1, indicating that scores did not remain exactly the same across time), these large correlations constitute strong evidence in favour of relative stability. On the other hand, the Pearson correlation for the BSI (r = 0.64, p < 0.001) emphasized the change in psychological distress over time (coefficients under 0.70 are usually considered as reflecting instability: Kline, 1993). 3.2.2. Mean-level stability As indicated earlier, this level of stability pertains to the group as a whole. Mean-level stability was evaluated in psychological distress and alexithymia. The mean BSI score at baseline was 87.44 (SD = 26.98) and at follow-up was 104.47 (SD = 32.16), reflecting a substantial increase in psychological distress during the exam period; mean change score = −17.03, t(69) = −5.59, p < 0.001. At the subscale level, a significant (and expected) increase was observed in depression, anxiety, somatization, hostility, obsession and phobias. There was no significant increase in paranoid and psychotic symptoms. Contrary to what was observed regarding psychological distress, there was no significant increase in alexithymia scores. The mean TAS-20 score at baseline was 48.45 (SD = 11.25) and at follow-up was 49.51 (SD = 10.85); mean change score = −1.06, t(70) = −1.11, p > 0.25. Since t-tests indicate whether the difference between means at baseline and follow-up can be considered as statistically different from zero but do not specify the magnitude of the change, effect sizes (Cohen’s d) were calculated. For alexithymia, the effect size was −0.083 which correspond to an insignificant effect according to the norms established by Cohen for social sciences (1988). For the BSI, the effect size was almost 6 times larger (Cohen’s d = −0.574) and amounted to a medium high effect according to Cohen’s norms. These results clearly indicate that, on average, alexithymia scores remained stable despite a significant increase in psychological distress. However, more information about the stability of individual scores is needed as it is plausible that increases and decreases in individual scores counterbalance one another, particularly since, as previous analyses demonstrated, scores did not show absolute stability. 3.2.3. Rank-order stability Kendall’s tau’s were performed respectively on BSI scores at baseline and follow-up, and alexithymia scores at baseline and follow-up. Kendall’s tau was 0.54 (p < 0.001) for TAS-20 scores, and it was 0.50 (p < 0.001) for the BSI scores. Although significant, these coefficients reflect, nonetheless, substantial rank inversions, and it is of note that there are nearly as many inversions in alexithymia scores as in psychological distress scores. 3.3. Explaining the relative stability of alexithymia scores Given that some degree of relative stability was still observed in measures of psychological distress (BSI), and that measures of psychological distress were correlated with alexithymia at both time periods (see Table 1), we examined next the degree to which relative stability in alexithymia scores might be attributed to relative differences in psychological distress severity among individuals at either or both baseline and follow-up. Assessing whether stability in alexithymia scores might be explained by relative differences among individuals in psychological distress severity requires an examination of the relationship between baseline and follow-up alexithymia scores, after removing any variance in alexithymia scores related to psychological distress severity. This was done by regressing follow-up alexithymia scores on (a) baseline and follow-up psychological distressscores, followed by (b) baseline alexithymia scores. Results are reported in Table 2. Findings showed that although a significant amount of variance of follow-up alexithymia scores could be predicted by psychological distress severity, the greatest part of variance was predicted by baseline alexithymia scores, over and above the variance already explained by psychological distress severity. By showing that the relative stability of alexithymia is not attributable to psychological distress severity, these results constitute strong evidence that alexithymia is much more than a transitory state accompanying psychological distress. Table 2. Relative stability of alexithymia scores accounting for psychological distress severity at baseline and follow-up (hierarchical regression analysis) Criterion variable Hierarchical order Predictor variable Adjusted R square F change (1,65) TAS-20 (2) 1 BSI (1) 11 4.98⁎⁎ 1 BSI (2) 2 TAS-20 (1) 55.5 62.96⁎⁎⁎ Note. BSI = Brief Symptom Inventory; TAS-20 = total score of alexithymia. (1) Fulfilled at baseline (2) fulfilled at follow-up. ⁎⁎ p ⩽ 0.01. ⁎⁎⁎ p ⩽ 0.001. Table options 3.4. Relationship between changes in alexithymia and changes in psychological distress Absolute, rank-order and relative stability analyses demonstrated that in spite of mean-level stability, alexithymia scores still varied slightly at the individual level. Although these variations were small, it was nevertheless important to examine the extent to which these changes might be accounted for by changes in psychological distress. To this end, changes scores in alexithymia were regressed on changes scores in the BSI. Results showed that less than 1% variance in alexithymia change scores could be accounted for by variance in BSI change scores [F(1, 62) = 1.112, p > 0.29]. These findings clearly suggest that the small changes in TAS-20 scores are not attributable to changes in BSI scores.