This study aimed to shed more light on the frequency and appraisal of LEs in minors at CHR-P, and their interaction with NS. Contrary to our expectations, we found only few and mostly unsystematic differences between CHR-P and the two control groups, especially no indication of a generally significantly higher burden in CHR-P. Compared to GPS, CHR-P reported a similar or higher number of LEs. In doing so, next to the sample-inherent higher number of health-related LEs that CHR-P shared with ClinS, increased numbers of LEs occurred in the interpersonal LE-domains, indicating unstable love relationships, and more conflicts with parents and at work, as well as more experiences of work overload. Furthermore, we found no indication of a LE-reducing effect of NS. On the contrary, especially in CHR-P, the number and presence of any past-year-LEs were positively correlated with the severity of subjective NS as measured by the Adynamia dimension of the SPI-CY.

More interpersonal LEs in CHR-P compared to healthy controls were also found in two other studies [21, 26] that suggested this being caused by a heightened sensitivity to interpersonal issues driven by attenuated psychotic symptoms, e.g., that paranoid ideas may lead to mistrust in the social environment and, consequently, to increased interpersonal LEs [21, 26]. Alternatively, increased interpersonal LEs may increase stress, which is considered a risk factor for psychosis, and may act as a catalyst for further psychotic experiences [26] or, as proposed in cognitive models, the appraisal of interpersonal events may result in negative schemas, leading to the development of attenuated psychotic symptoms [21]. Longitudinal studies are therefore needed to examine the temporal relationship between LEs and CHR-P symptoms.

Interpersonal problems, especially unstable relationships and conflicts due to emotional instability, are core features of borderline personality disorder (BPD) [35, 41], which has been linked to UHR, but not to basic symptom criteria [42,43,44,45,46]. In a mixed-age UHR sample, 62–78% showed affective instability, anger, impulsivity, and unstable relationships per DSM-IV BPD criteria [42]. Thus, the higher rate of interpersonal problems in our CHR-P sample may reflect emerging borderline traits. As personality disorders were not assessed due to the DSM-IV age threshold, future studies should consider their role. With DSM-5 Section III and ICD-11 now emphasizing personality functioning—particularly interpersonal functioning—as central to early personality pathology [47, 48], examining this domain in CHR-P individuals is especially warranted.

Finally, the high number of interpersonal LEs may indicate an erosion of supportive social networks and/or an impediment in the normal growth in social network size across adolescence [17]. Both interpretations that require further longitudinal studies would be in line with the reported smaller social networks overall, whereby social support deficits were specific to relatives other than parents and friends [49]. Since good social networks can act as an environmental protective factor not only for psychosis but also for suicidality in CHR-P states, social relationships were suggested a promising early intervention target in CHR-P, e.g., by expressed emotion interventions or social skills training [49, 50].

Experiences of chronic work overload that were most frequent in CHR-P – even in the absence of significantly increased occurrence of other work/school-situation-related LEs – may be considered the most direct expression of an increased stress susceptibility of all assessed LEs. Yet, as work overload was commonly reported as stressful, no additional significant group differences in stress appraisal were found, despite descriptively higher stress ratings in the two clinical groups. This was also true for most other LEs for that a significantly higher stress appraisal in CHR-P was extremely rare and unsystematic, and even less among past-year-LEs compared to five-year-LEs. Overall, our results therefore do not support a major role of subjective stress appraisal of LEs in terms of a stress-sensitization effect [20, 25,26,27,28,29]. Accordingly, a recent study using Ecological Momentary Assessment [51] reported no contemporaneous or temporal link between stress on CHR-P symptoms but rather a contemporaneous effect of CHR-P symptoms on stress. Thus, in addition to CHR-P symptoms, the higher number of mainly negative LEs might have a cumulative effect on the overall stress level in terms of a stress accumulation that increasingly rises the activity of the HPA and cortisol level during the time of emerging psychosis [52, 53].

Following developmental models of psychosis, increased negative or stressful LEs can lead to the critical stress threshold being exceeded, especially when protective factors including resilience and adaptive coping strategies fail to buffer against these stressors [8, 54]. Accordingly, UHR patients were more likely to exhibit emotional coping strategies that are mainly used in situations perceived as not changeable, while GPS used more task-orientated coping strategies that are mainly used in situations perceived as changeable [24]. This suggested that in themselves stressful feelings of little control over events by CHR-P might be triggered by experiences of low social support [24, 55]. Therefore, treatment strategies that focus on stress management and improving coping skills may help mitigate the negative effects of accumulating, especially normative and therefore largely unavoidable LEs [24].

Contrary to the assumed LE-reducing effect of NS, leading to lower numbers of LEs in CHR-P [14, 19, 24], in CHR-P, we detected neither fewer LEs nor negative correlations, where NS and LEs interacted. Rather, positive correlations between NS and LEs were commonly the most pronounced in CHR-P compared to other groups; only dispute with parents, was mostly related to NS in GPS. A positive association of NS and stressful LEs in UHR was also reported by See et al. [21], with LEs not contributing to a worsening of symptoms over time. Yet, when cannabis use was also taken into account, LEs were no longer significantly associated with NS [21].

Although NS were significantly less severe in ClinS and GPS compared to CHR-P, a slight negative association of LEs and NS was observed in ClinS and GPS only for treatment and presence of any past-year-LEs, again with a positive association in CHR-P. The interaction with treatment might indicate that observable attenuated NS, including affective symptoms that were reported a major reason for help-seeking in CHR-P [56], played a greater role in the decision to seek help in CHR-P compared to ClinS and GPS. Whereas the interaction with presence of past-year-LEs possibly indicate differences in coping with subjective NS [24] that might result in more LEs in CHR-P. Thus, improving coping strategies for subjective NS, e.g., by cognitive-behavioural interventions, may prevent avoidable non-normative LEs, especially interpersonal ones. Yet, more detailed and longitudinal studies are needed to shed more light on LE-NS interactions in the various groups.

Next to the strengths of our study – such as the assessment of the whole range of LEs, not only negative or per se stressful ones, the matching of samples to avoid age-related impact of normative LEs, and assessment of appraisals of LEs, some additional limitations need to be addressed. First, results will be impacted by differences in study design, such as assessment of LEs or of stress and stress tolerance – related to a specific LE or assessed as an independent variable [27, 30, 57], sample composition including the unknown but likely different proportion of CHR-P who will actually develop psychosis [52] and analyses [57]. Second, most LE studies were conducted in adult or mixed-age samples with different normative LEs than present in childhood and early adolescence [17]. Furthermore, parental mental health and parenting behavior also play a much stronger role in the interplay of LEs and mental health [58]; thus, these should be included in future studies of LEs in this young age group. This is also true for other so far not mentioned potential moderators and mediators such as: positive and other symptoms [21, 25, 28], comorbidities, particularly affective and anxiety disorders [58, 59], resilience factors [24], treatments (including different types of medication) [60], and biogenetic parameters of stress [26, 30, 53]. These should be studied using methods that allow studying their interplay, such as structural equation analyses [58], and disentangle cause and effect in longitudinal designs.