**2. Methods**

The present review follows the recommendations for a narrative review [23]. Hence, we intended to depict not only the extent, but also the range and nature, of these studies, especially with respect to the extent to which the respective findings are integratable in a conceptual frame derived from research on other pain sites. In particular, we were interested in weighing the findings against the background of psychosomatic and psychotraumatologic concept formation. We searched Medline (1966–2022), as of January 2022. Our search strategy included the following terms mapped to the appropriate MeSH subject headings: ("childhood trauma" OR "PTSD" OR "dissociation" OR "amnesia" OR "derealisation" OR "depersonalization") AND ("surgery" OR "postoperative maladaptation" OR "perioperative maladaptation" OR "depression and anxiety" OR "posttraumatic distress" OR "perioperative negative affectivity" OR "somatization"). The following terms were also included: ("comorbidity", "personality", "borderline", and emotional lability). Additional terms used in the search were "lateralized", "hemisphere", and "contralateral", in order to account for the fundamental principles of central pain processing.

To be included in the review, papers needed to measure or focus on specific dimensions of psychological influence on shoulder pain. Peer-reviewed journal papers were included if they were: written in English, involved human participants, and described a measure for psychological influence. Quantitative, qualitative, and mixed-method studies were included in order to consider different aspects of measuring psychological influence. In particular, we sought to identify psychosomatic and further etiologic concepts which are suitable for the integration of physical and psychic symptoms alike. Among those concepts is that of lateralization, referring to the central processing of pain that not only includes the activation of the contralateral hemisphere, but also the functional lateralization of mood, potentially causing syndromes such as anxiety to take side with contralateral peripheral pain for anatomic reasons.

#### **3. Results**

After duplicates were removed, a total of >499 citations were identified. After exclusion based on title/abstract and following assessment of eligibility based on full texts, we considered 48 studies eligible for this review. Our Medline search revealed a heterogenous set of endpoints and independent variables reflecting psychological disturbances. Although the hospital anxiety and depression scale was often reported, other measures of anxiety and depression were also deployed. In addition, specific fears such as pain catastrophizing and kinesiophobia are reported as predictors of the algofunction by three studies. Another important approach to the study question was the analysis of patient expectations as exemplified by Oh et al. [24]. Fewer studies reported personality characteristics as predictors of the algofunction in shoulder impingement. Table A1 (see Appendix A) gives an overview of those studies. As we hypothesized, the literature reviewed here was in line

with the assumption of a posttraumatic pathway of maladaptation after shoulder surgery, and, in addition, the findings did not contradict the assumption of a lateralized pattern of association between shoulder pain and psychiatric syndromes. To make matters even more complex, central sensitization may not only involve lateralization, but also affect patterns of immunologic response. Thus, the present review could mark a new field of study with respect to the (psycho-)dynamics of shoulder pain.

#### **4. Discussion**

#### *4.1. Psychosocial Correlates of Chronic Shoulder Pain*

In clinical settings, the perception of patients with shoulder pain is often characterized by the impression of neurotic alignment, tenseness, and a lower pain threshold [25]. Moreover, the more peculiar the patient is in terms of the emotional presentation, the longer the duration, and the greater the severity of his or her disability was found to be [26]. Since shoulder impingement syndromes are among the top representatives of causes for chronic pain, the associations of chronic pain with psychosomatic features are likewise manifest in shoulder impingement, as well. Accordingly, Cho et al. [27] reported the preoperative association between depression and joint dysfunction as well as quality of life, all of them assessed preoperatively. Similarly, as for the prospective perspective, Dekker et al. [28] found higher scores on the hospital anxiety and depression scale connected to less postoperative satisfaction after 6 months. Park et al. [29] found early (up to 6 months) postoperative pain and the range of motion (RoM) affected by the preoperative presence of anxiety and depression. They concluded that psychological factors would delay the recovery as far as shoulder disorders are concerned. In addition, Cho et al. [30] showed the prediction of the joint function by preoperative depression scores, highlighting the functional influence of depression. Similarly, Martinez-Calderon et al. [31], in their review on psychosomatic influences on shoulder pain, report a relationship between depression, anxiety, emotional distress, and shoulder pain. On this note, depression and anxiety also predicted shoulder pain after 3 months in Debeer et al.'s [32] study, and this study showed improvement in psychological well-being to be linked to less physical pain, whereas the opposite, i.e., prediction of more pain by psychological deterioration, was not shown. This finding is interesting because it may point to a "somato-genic" nature of psychosocial findings in relation to shoulder pain and corresponds to the assumption of pain lateralization [22].

The fear avoidance model of chronic pain posits pain to be modulated by sensory amplification as a result of the attention drawn to its perception by specific fears, mostly pain catastrophizing. As a result, kinesiophobia, that is, the fear of motion and re-injury, rises and makes withdrawal from social and work-related contexts act as a contraphobic compromise, relieving from fear on the one hand, but leading to chronification on the other, especially as regards surgical treatments [12,33]. Less participation and activity, however, prompt less healthy lifestyles and increase the burden of pain. Accordingly, kinesiophobia predicted shoulder pain in Debeer et al.'s study [30], and, in addition, the authors report kinesiophobia to be more stable in men than in women over the study period. Likewise, Martinez-Calderon et al. [30] showed an association between preoperative concerns, fear avoidance, and chronic shoulder pain. In their recent review, DeBaets et al. [33] conclude fear avoidance to predict treatment outcomes only when the treatment was surgical, whereas, otherwise, outcome expectancies and self-efficacy predicted the respective outcomes. This finding begs the question of how much the operative setting poses a specific challenge for coping resources different from conservative settings. DeBaets et al.'s [33] conclusion might therefore be in line with the suggestion of a posttraumatic pathway of postoperative maladaptation after total knee arthroplasty that is hypothesized to result from prior traumatization, setting the stage for re-traumatization by the operation [12].

Moreover, Menendez et al. [34] found psychological factors (i.e., pain catastrophizing and insufficient coping) linked to shoulder pain, and report the same regarding social circumstances such as (un-)employment. In a prospective study, Thorpe et al. [35] identified a cluster of patients with surgery for rotator cuff repair, characterized by restricted psychological health, and whose pain and function of the shoulder were worse than in those without psychological problems. Potter et al. [36], however, found mild and moderate psychological distress not to correlate with the one-year outcomes of arthroscopic rotator cuff repair. Contrarily, Cho et al. [26] report a postoperative decrease in psychologic symptoms along with increased quality of life within a 12-month follow-up. Thus, as much as shoulder pain may induce depression, its reduction might act as an antidepressant. Counterintuitively, however, pain and injury of the dominant shoulder (limb) are associated with less hysteria and hypochondriasis [37]. The same authors highlight the relevance of the dominant upper limb for ambulation, body care, and movement [37]. As regards the impinged shoulder, this translates to dependency on others for daily activities and personal needs, shorter walking range, as well as less speed, and the authors underscore the finding of heightened hysteria as well as hypochondriasis in those with a lesion of the non-dominant limb, who are therefore considered to have a higher somatic awareness. However, this finding might reflect the more general fact that unilateral pain leads to the activation of the contralateral hemisphere. Ji et al. [38] have supposed a differential pattern of nociception in the left and right amygdala, and, in humans, the right hemispheric lateralization of amygdala function is linked to negative emotions [39]. Not least, shoulder disorders are heavily associated with problems returning to work, especially if coinciding with depression and anxiety. Thus, financial problems and disadvantageous prospects in the labor market often affect those with shoulder disorders disproportionately [40].

Considering the aforementioned involvement of specific fears and negative affect in the pathogenesis and maintenance of chronic pain, associations of shoulder pain with psychiatric disorders are to be expected. Accordingly, Bot et al. [41] found prevalences of depression (21%), anxiety (26%), schizophrenia (24%), as well as dementia (29%) elevated in candidates for shoulder arthroplasty. Apart from being, partly excessively, overrepresented, those entities were, except for schizophrenia, also linked to a higher risk of adverse events, including anemic states and longer institutionalized treatment for shoulder pain. Vice versa, depression increases the risk of rotator cuff tear and rotator cuff repair surgery remarkably [42]. That aside, psychiatric comorbidity is linked to increased cost and opioid use in relation to shoulder rotator cuff repair [43]. In addition to psychiatric disorders, sleep problems are present in 70–89% of the patients with rotator cuff tendinopathy [44]. Karels et al. [45] found somatization, kinesiophobia, and pain catastrophizing to predict the persistence of complaints over a 6-month follow-up, and classify their findings as corroborating the fear avoidance model. Similarly, Engebretsen et al. [46] report the association between pain-specific fears and shoulder pain, but [47] no significant effect of self-efficacy on either disability or on return to work. Several studies [48–51] report coping styles, in particular avoidant coping, to be associated with pain and disability of the shoulder.

The understanding of psychosomatic reactions to shoulder pain is connected to the question of causality, given that depression or anxiety could be a reaction to shoulder pain. This stance, however, is called into question by reports of non-linearity of the relationship between the extent of shoulder pain and depression [52]. As Badcock et al. [52] hypothesize, ceiling effects may preclude a further worsening impact of an increased load of depressive symptoms. Furthermore, these authors report the levels of disability to modulate those of depression, e.g., through problems sleeping. However, affective disorders and personality disorders are linked to arthritis, leading some authors to speculate about an essential relationship between those phenomena that, according to this stance, may be the symptoms of a single entity rather than representing different entities [24].

#### *4.2. Recovery Expectancies*

Another avenue of research investigates outcome expectancies in relation to the factual outcomes of therapy. On this note, Oh et al. [15] report outcome expectancies associated with the preoperative dysfunction of the joint, and Henn III et al. found preoperative positive expectations linked to more favorable postoperative results [53] after a 1 year follow-up. Likewise, Martinez-Calderon et al. [30] could show high levels of self-efficacy, resilience and

expectations of recovery to be linked to levels of pain and disability albeit based on heterogenous studies involving different end-points and measures. Chester et al. [54], studying non-surgically managed shoulder pain, report that the prediction of pain (1/2 year) was best by the initial levels of shoulder pain, but strongly mediated by positive recovery expectations and optimism. Interestingly, the positive nature of the expectation may outweigh pain as a predictor. Accordingly, O'Malley et al. [55] found expectations to contribute to the short-term (3 months) functional outcome of shoulder disorders, explaining the interplay of functional improvements with functional expectations in terms of a specific capacity of negative expectations to undermine functional outcomes. In addition, Chester et al. [56] reported self-efficacy as a factor protective against shoulder pain. In addition, Henn III et al. [53], investigating a sample with primary surgical repair of a chronic rotator cuff tear, found positive expectations associated with the actual outcome with respect to function, even after controlling for a set of confounding variables including age, gender, smoking, workers' compensation status, symptom duration, number of previous operations, number of comorbidities, tear size, and repair technique.

Notably, Bandura [57] understood expectations of self-efficacy as the extent to which an individual will strive to cope with a certain health condition. Therefore, expectations may be linked to more or less favorable ways of coping, prevailing mood, sickness behavior, and compliance, and thus impact the course of a disease effectively. That said, a patient´s motivation for treatment is largely guided by her or his expectancies. The motivation for treatment comprises a cognitive, as well as an affective, component [58], the latter being the subjective suffering and the secondary gain from illness, and the former referring to the disease-related concept. The individual connotation of these components may be inclined to more or less predominance of medical, psychic, or social factors, as far as symptoms and expected treatments are concerned. In osteoarthritis, illness perception is predictive of disability, especially the perception of the level of perceived control and consequences of osteoarthritis (OA) [16]. In turn, as disability progresses and the prognosis deteriorates, particularly the judgement of the individual affection with OA in terms of the number of symptoms, the belief about their negative impact, and chronicity, grows ever more pessimistic. Hence, poor illness perceptions seem to function as a self-fulfilling prophecy.

#### *4.3. Higher-Order Factors: Temperament, Personality, and Posttraumatic Pathways*

In that same sense, expectations reflect the tendencies that constitute the personality. For example, Basat et al. [59] found depressive temperament (operationalized as having withdrawn, the presence of self-blaming features, and the absence of steadiness) linked to a worse algofunctional outcome during a follow-up of almost 2 years. Likewise, Bru et al. [60] reported neuroticism and extraversion as well as trait anxiety correlated to pain ratings and concluded from their cross-sectional, retrospective study that personality traits would be more involved in shoulder pain than in back pain (p. 491), and Chiaramonte et al. report an association between primary adhesive capsulitis and perfectionism, novelty seeking (negative), and harm avoidance [61]. Not least, another study [62] found the effect of pain catastrophizing on joint function (though not pain) moderated by optimism.

With respect to such findings, however, Coronado et al. [62] also note the tendency of weakness that the associations between the outcomes of rotator cuff repair and psychological variables display, and the bias towards the study of surgical treatments in that respect as opposed to conservative therapies. Likewise, Sheikhzadeh et al. [63] in their formidable review have shown that psychosocial factors unfold their potential to predict pain in a more clear-cut manner with respect to surgical therapies as opposed to conservative ones. This result is in line with research highlighting the traumatic and interpersonal potential of surgery based on the suggestion that the violation of bodily integrity might bear the risk of re-traumatization for the traumatized [12].

Likewise, as to the link between chronic pain and psychological trauma, shoulder disorders are also overrepresented in veterans or (other) individuals with PTSD [61]. On this note, Wang et al. proposed a substantial overlap between pain, PTSD, and emotional factors, including strong feelings of anger, hatred, and aggression. Interestingly, some authors [33] note that the muscles of the back and the shoulder would be the first to react to tension [64].
