پاسخ به استرس در بیماران مبتلا به اختلالات روانی در مقایسه با افراد با سطوح مختلف از آسیب پذیری به جنون، افراد مبتلا به افسردگی و افراد سالم

Abstract Background and Objectives An experimental design was used to test whether self-reported, psychophysiological and symptomatic stress-responses increase as a function of the underlying vulnerability to psychosis as proposed by vulnerability-stress-models. Methods Stress-responses of participants with psychotic disorders (PSY, n = 35) were compared to those of participants with attenuated positive symptoms (AS, n = 29), first-degree relatives of persons with psychotic disorders (REL, n = 26), healthy controls (HC, n = 28) and controls with depression (DEP, n = 30). Using a repeated measures design, participants were assigned to a noise stressor, a social stressor and a no stress condition in random order. Stress-responses were assessed via self-report, salivary cortisol levels, heart rate and skin conductance levels. State-paranoia and depression were assessed with clinical scales. Results PSY reported to be significantly more stressed than HC, AS and REL across all conditions which went along with increased heart rate and decreased overall cortisol release. In contrast, AS showed elevated levels of cortisol. PSY showed a stronger response of self-reported stress to the noise condition compared to the no stress condition than HC, but no stronger response than the other samples. Furthermore, the stressors did not trigger stronger psychophysiological responses or symptom-increases in PSY. Limitations The social stressor was brief and not individualized and did not have an effect on cortisol. Conclusions The findings support the notion that subjective stress-responsiveness increases with vulnerability, but not the assumption that symptoms arise directly as a function of stress and vulnerability. Also, the generally high levels of arousal seem to be more relevant to psychosis than the responsiveness to specific stressors.

Since the late 1970s stress has been ascribed a central role in the pathogenesis of psychotic disorders (Nuechterlein and Dawson, 1984 and Zubin and Spring, 1977). Vulnerability stress models differ in detail, but all assume that people have varying levels of vulnerability, and that the likelihood of psychotic symptoms is a function of the extent of vulnerability and stress that the individual encounters. The mechanisms that translate stress into psychotic symptoms are proposed to be decreased neuropsychological functioning resulting in a breakdown of information processing abilities in earlier models (Nuechterlein and Dawson, 1984 and Zubin and Spring, 1977) and reasoning biases, perceptual anomalities and emotion-processing in more recent model variants for positive symptoms (Blackwood et al., 2001 and Garety et al., 2001). Vulnerability stress models have a high face validity and are used as a basis of psycho-educational approaches that aim to help patients to monitor early symptoms by reducing stress in their every-day lives. Nevertheless, the evidence for the basic notion that stress triggers psychotic symptoms is not fully conclusive.

3.1. Manipulation check The repeated measures ANOVA of time and condition on self-reported stress revealed a time × condition interaction (F(10, 1470) = 11.4, p ≤ .01, η2partial = 0.07). Pairwise comparisons revealed significant differences between the neutral and the two stress conditions and between the noise and the social stress condition (all p ≤ .01) that are depicted in Fig. 2. Fig. 2 also indicates that the noise stressor had an earlier and more constant impact than the social stressor, the latter showing self-reported stress peaks at t4 (beginning of observance) and t6 (before giving speech). Full-size image (79 K) Fig. 2. Manipulation check. Figure options The repeated measures ANOVA on heart rate revealed a significant condition × time interaction (F(4.4; 511.4) = 8.1, p ≤ .01, η2partial = 0.06). Graphical inspection points to a linear reduction of HR over time in the no stress and noise condition, but no substantial reduction in the social condition. Post-hoc tests of differences between the social and the no stress condition were significant at t5 and t6 (both p ≤ .01). The repeated measures ANOVA on SCL revealed a small but significant condition × time interaction (F(3.7; 479.7) = 3.2, p ≤ .05, η2partial = 0.02). Graphical inspection indicates a slightly stronger increase of SCL in the noise compared to the other conditions. Post-hoc tests of condition differences at each time-point were significant at t1 and t2 (no stress > social stress, p ≤ .05) and at t5 and t6 (noise stress larger than social stress, p ≤ .05). The repeated measures ANOVA of cortisol revealed a trend effect of condition × time interaction (F(5.3; 699.4) = 1.8, p = .10, η2partial = 0.01). Graphical inspection points to reduced cortisol in the noise compared to the social stressor. Thus, overall the stress-inductions were effective to increase subjective stress ratings. The social stressor also had an impact on heart rate. However, the effects of the stressors were less straightforward for SCL and absent for cortisol. 3.2. Impact of stressors on self-reported stress The MANOVA of group, condition and time on self-reported stress revealed a significant effect of group (Λ = 0.75, F(8, 284) = 5.6, p ≤ .01, η2partial = 0.14), and a significant group × condition interaction (Λ = 0.83, F(16, 428.4) = 1.7, p ≤ .05, η2partial = 0.05). There was no effect of group × time (Λ = 0.71, F(40, 509.9) = 1.2, p = .18), indicating that the course of self-reported stress over time was comparable across groups. There was also no group × condition × time interaction (Λ = 0.55, F(80, 491.6) = 1.0, p = .14). The Bonferroni corrected univariate analyses indicated that the group effect was significant for VAS-1 (F(4, 143) = 7.6, p ≤ .01, η2partial = 0.18) as was the group × condition effect (F(7.4, 262.7) = 3.0, p ≤ .01 η2partial = 0.08). For VAS-2 the group (F(4, 143) = 7.3, p ≤ .01, η2partial = 0.17) and group × condition effect were also significant (F(7.9, 282.1) = 2.3, p ≤ .05, η2partial = 0.06). We will report Bonferroni's post-hoc tests and interaction effects for the VAS-1, which are graphically depicted in Fig. 3: Post-hoc tests of the group differences revealed that PSY felt significantly more stressed than REL (p ≤ .05), AS, and HC (p ≤ .01). Post-hoc ANOVAS of the condition × group effect revealed trend significant group differences in the increase of self-reported stress from the no stress-to the noise stress-condition (F(4, 143) = 2.7, p = .07, η2partial = 0.07) which was due to a significantly higher increase in PSY as compared to HC (p ≤ .05) and significant differences from the noise to the social stress-condition (F(4, 143) = 4.0, p ≤ .01, η2partial = 0.10), which was due to a significantly higher decrease in PSY compared to HC (p ≤ .05). Full-size image (25 K) Fig. 3. Interactions between stress condition and group on self-reported stress.