In the present study, we sought to investigate further the instability in self-esteem and affective instability in everyday life in BPD. We compared a sample of remitted BPD participants, i.e. patients with a loss of diagnosis at the time of the e-diary assessment, with acute patients with BPD and HC participants utilizing e-diaries, high-frequency sampling, and various time-sensitive instability indices. We found significantly lower self-esteem instability in the BPD-REM participants compared to the BPD-ACU patients, whereas mainly no significant differences regarding affective instability emerged. On the other hand, the BPD-REM participants consistently showed significantly heightened instability in both self-esteem as well as affect compared to the HC participants. We furthermore addressed potential associations between the strength of the association between changes in affect and changes in self-esteem and the level of functioning as well as the quality of life in the BPD-REM participants. Even though the analyses did not reveal significant associations with the level of functioning nor the quality of life, our results indicate a potential effect of the strength of the intertwinement of changes in tense arousal and self-esteem on the quality of life domains of psychological health and social relationships in the BPD-REM participants. Peculiarly, a greater strength between changes in tense arousal and changes in self-esteem indicated a lower quality of life in these two domains. Taken together, our results suggest that self-esteem instability is lower in remitted BPD, whereas the levels of affective instability are mainly comparable in acute BPD and remitted BPD. However, BPD-REM participants still show heightened self-esteem instability and affective instability in comparison to HC. The strength between changes in tense arousal and changes in self-esteem tentatively hint to a lower quality of life in the domains of psychological health and social relationships in the BPD-REM participants.
Examining the instability of self-esteem seems a promising avenue for future research since our cross-sectional results indicate that it declines after remission from BPD, whereas the affective instability seems to persist. Previous e-diary studies mostly addressed affective instability in BPD (for a review see ). However, those studies including clinical control groups in order to address the specificity of affective dysregulation for BPD, largely failed, e.g. using various instability indices , subcomponents of affective dysregulation , emotional granularity , or emotion sequences . Therefore, it has been suggested that affective instability constitutes a transdiagnostic marker of affective dysregulation. In our study, BPD-REM participants were significantly older than BPD-ACU and HC participants, and it has been discussed whether affective instability declines with older age. A recent comprehensive cross-sectional e-diary study from our group found that global affective instability (i.e. SSD) in acute BPD patients' everyday lives indeed declined with years of age . However, in the current study, we found no differences regarding global affective instability between the, on average younger, acute patients in the BPD-ACU group and the, on average older, remitted participants in the BPD-REM group. Thus, age differences in the BPD-ACU and the BPD-REM groups cannot explain our findings of significantly heightened instability of self-esteem but mainly similar affective instability in the BPD-REM and the BPD-ACU groups.
Furthermore, our finding of comparable levels of affective instability in participants with remitted BPD and patients with acute BPD is in line with results from prospective multi-wave follow-up studies of BPD. Using retrospective self-report measures to assess psychopathology, such as interviews and questionnaires, these studies suggest differential levels of stability across BPD features with the criterion of affective instability persisting over a longer time  and being most prevalent over the follow-ups with higher shares of patients still fulfilling the criterion compared to other criteria such as unstable sense of self (e.g. [28, 30]). Even though these studies examining the course of individual symptoms of BPD demonstrated an overall decrease in all symptoms, results indicate that the diagnostic criterion of unstable sense of self remits more frequently and at a quicker rate than the criterion of affective instability. The criterion of an unstable sense of self in everyday life has been neglected in e-diary studies until lately. However, recent e-diary studies brought new attention to the importance of self-esteem in BPD showing heightened instability in self-esteem in daily life in BPD compared to HC , as well as highlighting its associations with engaging in dysfunctional behaviors . In the study at hand, we cross-sectionally compared acute vs. remitted BPD patients, and no clinical control group was included in the study. Thus, no statement can be made regarding whether our findings of heightened self-esteem instability are BPD-specific or whether they constitute a transdiagnostic phenomenon. A multitude of prior e-diary studies indicates that affective instability constitutes a transdiagnostic marker of dysregulation in the affective system. For a better understanding of the specificity, as a first step, it is of high relevance to examine whether self-esteem instability and affective instability differ between individuals with a current BPD disorder, remitted disorder, and non-clinical controls, i.e. whether they are indicative of clinical group belonging. Studies including clinical controls are clearly warranted to proceed to the next stage in order to examine further whether instability of self-esteem is specific for BPD.
To examine the assumption that affective changes that are accompanied by changes in self-esteem are experienced as more burdensome and threatening, we tested whether the strength of the association between changes of valence and self-esteem and that between changes of tense arousal and self-esteem predict the level of functioning and the quality of life in the BPD-REM participants. We found no association with the level of functioning, i.e. the GAF score, and the strength of the associations between changes in affect and self-esteem. Though the GAF's advantage of simplicity , it comes with a number of limitations. Most importantly, the reliability of the scale tends to be low and not sufficient in the routine clinical setting , as is the validity , especially the predictive validity . Moreover, studies found that the rating score can be influenced by raters' attitude towards the GAF and their knowledge of the patients' day-to-day life, among other confounding variables [59, 60]. Consequently, the GAF was excluded from the DSM-5 .
The associations between changes in affect and changes in self-esteem were not significant predictors of the self-reported overall quality of life or general health. However, even though not significant, results indicate that the association between changes in tense arousal and changes in self-esteem have a potential effect on the domains of psychological health and social relationships of quality of life in the participants in the BPD-REM group. Most likely, these effects are of smaller magnitude, and our study of 35 participants in the BPD-REM group was insufficiently powered to reveal smaller effects. Nonetheless, we would have expected to find an effect of the associations in the two domains of the quality of life questionnaire, in which we found marginal significant effects, i.e. psychological health and social relationships, since large-scale follow-up studies revealed that psychosocial functioning often remains impaired and only a few patients attain social and vocational competence after symptomatic remission (e.g [28, 33, 34]). Future studies should further examine the effects of the associations between changes in affect that are accompanied by changes in self-esteem on participants' well-being and quality of life since they have great clinical significance.
Several limitations of the current study deserve mention. First, our study is limited in that it is based on cross-sectional data. Only longitudinal studies can definitely speak to the trajectories of affective instability and self-esteem instability over the course of BPD and after remission from it, i.e. the loss of diagnosis. However, there are no longitudinal e-diary studies at hand, and our study is the first that compares affective instability and self-esteem instability in acute and remitted BPD. Moreover, attrition rates can bias the results of longitudinal studies through the loss of participants during follow-up, often due to behaviors strongly associated with the disorder itself. To ensure basic comparability between groups in our cross-sectional study, we assessed everyday life symptomatology in patients currently waiting for inpatient treatment on our specialized BPD treatment unit, as well as former patients who underwent this residential treatment on our specialized unit several years ago. Our main findings that self-esteem instability is lower after remission from BPD, whereas affective instability is still heightened and comparable to that of acute BPD patients, are in accordance with prospective multi-wave studies in BPD using retrospective interviews and questionnaires [28, 31]. These findings collectively suggest that the instability of self-esteem seems to remit faster, whereas the affective instability seems to persist for a longer time, even after remission from BPD. This consistency indicates that the findings observed in this study cannot be attributed solely to artifacts such as longer duration of illness, the time elapsed since the loss of diagnosis, or selective mortality. Nonetheless, the results of this study should be replicated in a longitudinal e-diary study before strong conclusions are drawn.
Second, given that only female participants were included in our study, the generalizability of the findings is limited, and the results may not be valid for male patients with BPD. However, the use of an entirely female sample also reduced the heterogeneity of the sample, which may have been useful, given the literature on sex differences in affect  and self-esteem . The results of this study should be replicated in a mixed-sex sample including male patients with BPD.
Third, we defined symptomatic remission as a loss of BPD diagnosis. Patients who had previously met ≥ 5 diagnostic criteria for BPD but dropped below the diagnostic threshold on the continuum of BPD criteria and fulfilled less than five BPD diagnostic criteria within the past year at the time of the e-diary assessment were considered remitted. Thus, our definition of remission is rather liberal. However, since only three BPD-REM participants fulfilled four diagnostic criteria for BPD, a more conservative definition of remission considering only participants fulfilling ≤ 3 diagnostic criteria for BPD did not change our findings regarding instability in self-esteem and affective instability (results available upon request). Moreover, the definitions of symptomatic remission from BPD show substantial variation in previous studies, and no consensus has been reached.
Fourth, the patients in the BPD-ACU group were diagnosed with a variety of co-occuring axis I and axis II disorders. We were unable to address the influence of different comorbid diagnoses on self-esteem instability and affective instability due to our restricted sample size. Thus, no statement can be made regarding whether our findings are independent of any comorbidity. Comorbidity is the rule rather than the exception in BPD . Therefore, a sample of BPD patients with high comorbidity rates constitutes a representative, non-artificial sample. In contrast, BPD patients without comorbid disorders cannot be seen as representative of the BPD population . Additionally, no clinical control group was included in the study. Thus, we cannot make any statement whether our findings of heightened self-esteem instability in acute BPD are BPD-specific or whether they constitute a transdiagnostic phenomenon (as does affective instability). Studies including clinical controls are clearly needed to assess whether self-esteem instability is specific to acute BPD or associated with underlying psychopathology other than BPD.
Fifth, we did not consider emotionally or self-esteem relevant events or triggers (e.g. interpersonal events) that might have influenced participants' ratings during the e-diary assessment period. Because events or triggers might differ between groups and given the growing recognition of the importance of contextual factors in e-diary studies (e.g. ), assessments of relevant events should be included and examined in future studies.
Despite these limitations, the current study significantly deepens our understanding of the unstable psychological processes involved in BPD. By extending prior e-diary research on self-esteem instability and affective instability in BPD, this study conducted in remitted BPD participants' everyday lives builds on growing evidence for the importance of self-esteem instability. Our findings have several clinical implications and provide interesting avenues for consecutive research. E-diaries provide the possibility of in-vivo diagnostic assessments of the severity and time-dependency of dynamic symptoms in daily life and, thus, can help to support clinical diagnoses and decision making. Repeated assessments in participants' daily lives seem especially fruitful for the evaluation of therapeutic interventions such as dialectical behavior therapy , which targets these domains of instability. E-diaries present a promising tool for clinical research by tracking potential changes in self-esteem and affective instability by therapeutic interventions over the course of the psychotherapeutic therapy in the most relevant contexts of all, patients' everyday lives.