Premenstrual dysphoric disorder and childhood maltreatment, adulthood stressful life events and depression among Lebanese university students: a structural equation modeling approach

This cross-sectional study was conducted between February and March 2021 during the COVID-19 pandemic and the lockdown period imposed by the Lebanese government. Lebanese students were recruited using the snowball technique from all national Universities in Lebanon via an auto-administrated online survey. Participants were above 18 years; excluded were those who refused to participate in the study.

Based on a mean physical abuse score of 23.80 ± 5.51 in healthy controls and 26.48 ± 6.79 in those with PMDD [10], a power of 95% and a risk of error of 5%, the G-power calculated a minimal sample of 232 female university students to be enrolled.

The Arabic self-administered questionnaire, which required approximately 30 min to fill, was anonymous and contained close-ended questions. It was divided to several sections. The first part tackled sociodemographic characteristics including age, educational level, marital status, and household crowding index. The latter was calculated by dividing the number of people in the house by the number of rooms, except the bathrooms and kitchen [37]. The physical activity index was determined by multiplying together the intensity, the time and the frequency of physical activity [38]. Moreover, some questions were dedicated to observe the menstrual activity of the respondents including the number of days between each menstrual cycle, in addition to whether the cycle is usually regular or not. The second part of the questionnaire included the following scales:

All scales, except the LDS and the CASRS, were translated from English to Arabic using one bilingual translator. A backward translation was then performed by a native English-speaking translator, fluent in Arabic and unfamiliar with the concepts of the scales. Minor discrepancies were solved by consensus.

All analyses were carried out using the Statistical Package for Social Sciences (version 24.0 with AMOS; IBM®, Armonk, NY, U.S.A.). We had no missing values in our database since all questions were required in the Google form. Structural equation modeling (SEM) was performed to examine the structural relationship between childhood maltreatment and life’s stressful events taken as independent variables, PMDD as the dependent variable, and depression as the mediator, among university students. In the first step of the SEM, the assessment of normality of the PMDD score was performed by calculation of the skewness and kurtosis; values for asymmetry and kurtosis between − 1 and + 1 are considered acceptable in order to prove normal univariate distribution [47]. These conditions consolidate the assumptions of normality in samples larger than 300 [48]. The root mean square error of approximation (RMSEA) statistic, Tucker Lewis Index (TFI) and the comparative fit index (CFI) were used to evaluate the goodness-of-fit of the model [31]. RMSEA values ≤0.06 or CFI and TFI values > 0.90 indicate a good-fitting model [31]. P < 0.05 was considered statically significant.

The structural relationships between childhood maltreatment and life’s stressful events taken as independent variables, PMDD as the dependent variable, and depression as the mediator, are shown in Fig. 1. There was no multicollinearity between the variables entered in the model. The RMSEA, TFI and CFI values were 0.023 (pclose = 0.722), 0.92 and 0.98 respectively, indicating a good fit of the model. Higher life’s stressful events (Beta = 0.18; p < 0.001), physical (Beta = 0.19; p < 0.001), sexual (Beta = 0.18; p < 0.001) and psychological (Beta = 0.33; p < 0.001) abuse were significantly associated with higher depression. Moreover, higher sexual (Beta = 0.11; p = 0.021) and psychological (Beta = 0.11; p = 0.040) abuse and higher depression (Beta = 0.37; p < 0.001) were significantly associated with higher PMDD. The indirect relationships between psychological abuse/sexual abuse, depression and PMDD showed that depression mediated the association between both psychological (Beta = 0.22; p = 0.001) and sexual (Beta = 0.38; p = 0.004) abuse and PMDD.

Table 2 presents the standardized coefficients with standard errors and p-values of the direct effects of variables on each other.

In terms of association between physical abuse, neglect and depression, our research revealed that more physical abuse and neglect were associated with higher depression, which supports several previous findings. According to a recent study [49], a history of childhood maltreatment was associated with depression in adolescents. All dimensions of childhood maltreatment had a significant association with depression, with psychological violence having the most impact on the occurrence of depression, followed by exposure to violence and physical violence. Moreover, Spinhoven et al. [50] found that emotional neglect is a predicting factor of depression.

Results of our study showed that stressful life’s events were associated with higher depression but not PMDD. This finding supports previous research [2, 15, 31] that highlighted that adulthood stressful life’s events are one of the most common environmental and psychological causes that may trigger depressive symptoms or major depressive disorders. Furthermore, people with a history of notable emotional trauma reported greater rises in depressive symptoms when faced with contingent stressful life’s events [25]. The occurrence of negative life’s events, as well as one’s perception of/response to those events, can influence the production and persistence of depressive symptoms, according to cognitive-behavioral models of depression [51]. Individuals that have a propensity to create negative attributions in relation to the triggers of negative life experiences, about themselves, and about potential outcomes (in accordance with the hopelessness theory of depression) might be more prone to experiencing depression, according to cognitive vulnerability-stress models [52].

In relation to the Lebanese context, recent findings showed that 28.9% of Lebanese adolescents had suicidal ideation [53], 31.7% were alexithymic and 26% had depression in its severe forms [54]. Moreover, 28.9% of Lebanese adults had suicidal ideation [44], 20.8% had alexithymia [55], 32.9 and 26.9% suffered from mild-to-moderate and severe -to-very severe depression [31], whereas 19.9 and 18.4% had moderate and severe anxiety [56] respectively. Such numbers reveal the severity of the situation in Lebanon when it comes to mental health. Our study, which showed the direct effect of major stressful life’s events (such as the pandemic, Beirut Port explosion, and change in the economic situation) on depression, can provide insight to health care providers when dealing with patients directly affected by those major events.

In our study, sexual and psychological abuse were positively associated with the severity of PMDD symptoms among Lebanese university students. These findings support previous research that suggest that childhood maltreatment may play a role in the development of PMDD [57]. Furthermore, patients with PMDD had a higher overall score on the Childhood Trauma Questionnaire, as well as higher scores for physical, emotional, and sexual abuse than healthy controls, according to previous authors [10]. Bertone-Johnson et al. [6] also concluded that childhood maltreatment, with emphasis on emotional and physical abuse, raises the risk of moderate-to-severe PMS (PMDD being the severe type of PMS).

The association between sexual and psychological abuse and PMDD was mediated by depression (depression being a disorder of impaired emotion regulation [29], which is linked to the development of psychopathology later in life) [58, 59]. Emotion regulation is described as the knowledge and comprehension of one’s emotions, acceptance of emotional reactions, and how one reacts and behaves in response to these emotions [60]. Several studies have shown that emotion regulation abilities mediate the connection between childhood trauma and psychopathology. Van der Kolk & Bessel [61] found that one of the most common characteristics of children who have been subject to abuse is their failure to control their emotions in an adaptive fashion. The lack of self-regulatory mechanisms in these children contributes to weak modulated affect, decreased impulse regulation, mistrust, and confusion about others’ reliability and predictability [61, 62]. Hence, the effect of maltreatment on children’s anxious and depressive symptoms was induced by a maladaptive pattern of emotion control [63]. Furthermore, Kim & Cicchetti [64] showed that emotional and/or sexual violence in parent-child relationships are linked to an evolving aberration in the organization of affective systems, which eventually leads to psychopathology. As a result, cyclic fluctuations in mood, affect, and cognition, which happen on a monthly basis, can be more difficult for women with a history of childhood trauma, mainly due to deficiencies in emotion regulation [28]. On another note, childhood maltreatment leads to hypothalamic-pituitary-adrenal axis hyperactivity, developmental changes in the brain, and epigenetic changes in the amygdala and hippocampus. Stress reactions can be heightened as a result of these biological effects [65].

In relation to our findings, healthcare professionals need to take into consideration two main issues while assessing women with PMDD: 1) the mental state of those women, especially those who also manifest symptoms of depression (since it was shown to be a mediating factor of PMDD); and 2) the possibility of childhood maltreatment history (especially emotional and sexual abuse). By taking into account those two factors, health care professionals can assess patients more accurately, in addition to providing adequate interventions that fit their needs and their history.

Moreover, the positive correlation between stressful life’s events and depression is a great insight for healthcare professionals dealing with patients who recently experienced such events, especially in the Lebanese context. During the past couple of years, women living in Lebanon experienced many stressful events as previously mentioned. The parallel occurrence of such stressful events with the sharp rise in mental illness, confirms our findings where life’s stressful events were positively correlated with depression. Thus, healthcare professionals should take into consideration those major events in relation with depression among Lebanese women when designing their approach.

Several limitations can be attributed to this study. First, a screening questionnaire was used to diagnose PMDD, despite the fact that the DSM-5 states that this diagnosis involves the completion of a prospective daily rating for at least two menstrual cycles. Moreover, the PSST is a retrospective questionnaire capturing symptoms in the past year, which is considered far less accurate compared with prospective rating of mood symptoms. Additional diagnosis of PMDD, according to the DSM-5, also includes an increase of at least 30% (relative to the range of the scale used) in the mean self-ratings of negative moods (i.e., irritability, depression, anxiety, and mood swings) in the 7 days before menses as compared with the ratings for the 7 days afterward in at least two of the three base-line cycles. Nevertheless, PSST has been used in previous research since its publication [28, 66]. In addition, the International Society of Premenstrual Disorders has accepted it as a possible diagnostic tool [67]. Second, the research is focused on self-report questionnaires outlining adverse experiences that occurred in the past. Hence, information bias may be present since the data is based on self-reported answers. Due to the cross-sectional design of the study, assumptions about causality remain hypothetical. Moreover, the used scales (except the LDS) are not validated in Lebanon to date. Selection bias is possible due to the snowball technique used to enroll participants. Adding to that, residual confounding bias is also possible since not all factors associated with PMDD have been taking into consideration in this study. For example, several other symptoms (e.g. irritability, sadness, anxiety) are considered to be equally as critical in women with PMDD as depressive symptoms while only depression was used as a mediating factor in our study. Further studies should focus on using those other symptoms as mediators. Finally, the study’s sample size limits the findings’ generalizability. Future research using a longitudinal design and a larger sample could help us better understand these relationships.