Introduction
Major depressive disorder and its subtypes
Major depressive disorder (MDD) is a common but heterogeneous mental disorder. According to the World Health Organization (WHO) report, 322 million people suffered from MDD which accounted for 4.4% of the global population resulting in a surge in suicide rates as well as a huge social and economic burden [Reference Friedrich1]. A diagnosis of MDD in the Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria 5th edition is made if 5 out of 9 symptoms are met, and several of these symptoms are opposites (e.g., weight gain or loss), two people with the same MDD diagnosis may have few symptoms in common [2]. Although DSM-5 contains several specifiers such as depression with atypical or melancholic features, they have not proved sufficient to predict prognosis and treatment responses [Reference Lojko and Rybakowski3, Reference Parker, Fink, Shorter, Taylor, Akiskal and Berrios4]. Because the diagnosis of MDD does not reflect the possible combinations of depressive symptoms, the heterogeneity of MDD significantly hinders the advancement in the etiology and treatment options [Reference Lux and Kendler5, Reference Goldberg6]. This heterogeneity underscores the necessity of an in-depth investigation of clinical symptoms and relevant risk factors [Reference Lynch, Gunning and Liston7].
A different approach to examining the heterogeneity of MDD is through person-based methods, such as latent class analysis (LCA) to identify symptoms that tend to co-occur in patient subgroups [Reference Lubke and Muthen8]. These techniques cluster patients based on the congregation of different depressive symptoms without a pre-conceived hypothesis. Studies have identified an “atypical” subtype characterized by increased sleep and appetite, a “typical” or “melancholic” subtype characterized by decreased sleep and appetite, and the presence of psychomotor symptoms [Reference Lamers, de Jonge, Nolen, Smit, Zitman and Beekman9–Reference Kendler, Eaves, Walters, Neale, Heath and Kessler11]. These subtypes and their most distinguishing symptoms such as appetite and weight gain or loss were also linked with distinct biological and genetic correlates and different neural activities [Reference Lamers, Vogelzangs, Merikangas, de Jonge, Beekman and Penninx12, Reference Simmons, Burrows, Avery, Kerr, Taylor and Bodurka13]. Particular pharmacological interventions may be indicated in some subtypes (e.g., monoamine oxidase inhibitors in atypical depression) highlighting their potential clinical utility [Reference Trivedi, Rush, Crismon, Kashner, Toprac and Carmody14]. However, some promising exceptions exist. Previous studies have yielded inconclusive findings between the increase and decrease in sleep, appetite, weight, and psychomotor symptoms in various population groups, while these distinctions may be crucial in identifying MDD subtypes [Reference Lamers, de Jonge, Nolen, Smit, Zitman and Beekman9, Reference Rodgers, Grosse Holtforth, Muller, Hengartner, Rossler and Ajdacic-Gross10, Reference Alexandrino-Silva, Wang, Carmen Viana, Bulhoes, Martins and Andrade15]. Likewise, studies pointed out that the DSM melancholic or atypical specifiers cannot help to create more homogeneous MDD subgroups [Reference Lorenzo-Luaces, Buss and Fried16, Reference Fried, Coomans and Lorenzo-Luaces17]. To further examine the validity, potential etiological attributes, and clinical usefulness of these subtypes, it is essential to analyze how specific depressive symptoms inform the research of etiology, diagnosis, and the development of targeted treatment [Reference Fried and Nesse18].
Stressors are known for their increased risk of developing MDD
Exposure to lifetime stressors and their related stress responses have been the focus of substantial MDD research [Reference Hammen19–Reference Pego, Sousa, Almeida and Sousa21]. One main reason to study stress in MDD is that adverse childhood experiences and chronic adulthood stressors are common experiences that portend the development of a variety of costly and burdensome mental and physical health outcomes [Reference Richardson, Shaffer, Falzon, Krupka, Davidson and Edmondson22–Reference Steptoe and Kivimaki24]. According to the diathesis-stress theory, exposure to stressors may activate a diathesis or vulnerability, transforming the potential or predisposition into the actuality of psychopathology [Reference Monroe and Simons25]. Exposure to early life adversities and adverse stressful life events serve as general vulnerability factors that make individuals more sensitive to stress challenges and therefore can feed forward into exacerbation of ongoing, or greater susceptibility toward future stress-related disease states, especially as they pertain to negative affect and mental health in general [Reference Tsur and Abu-Raiya26–Reference Ludwig, Pasman, Nicholson, Aybek, David and Tuck28].
Although evidence has suggested that the differential effects of stressors on MDD vary across different stages of life [Reference Su, D’Arcy, Li, O’Donnell, Caron and Meaney29–Reference Teicher and Samson31], little is known about which stressor has the most influential impact on MDD subtypes by contrasting stressors at different life stages, various stressors profiles across the lifespan, and proximal vs. distal stressors. Even less is learned about how these different measures of stressors intersect to increase overall vulnerability to MDD subtypes. Identification of the distinct roles of stressors may be essential in determining the biological bases of depression [Reference Mezulis, Salk, Hyde, Priess-Groben and Simonson32, Reference Carlson and Oshri33]. Treatment guidelines and procedures could then be more personalized when individuals with stressors and those without such stressors but with the same diagnostic labels are differentiated.
To the best of our knowledge, no studies have been conducted or published to comprehensively articulate the complex relationships between stressors across the lifespan and MDD subtypes. In the present study, we aim to untangle the relationships between MDD subtypes and various stressor measures across the lifespan and to explore how stressors intersect in MDD subtypes.
Methods
Study sample
Data were from the ongoing longitudinal cohort Zone d’Épidémiologie Psychiatrique du Sud-Ouest de Montréal (ZEPSOM). It is a community-based population cohort randomly selected from Montreal Southwest, Canada. In 2007, a sample of 2433 participants aged 15 to 65 years was randomly recruited at baseline and the following data were collected at a two-year interval. More details about the study cohort could be found in the previous research [Reference Su, Li, D’Arcy, Caron, O’Donnell and Meng34]. Because this study focused on lifetime stressors and MDD, only those respondents (N = 1351) who had complete information on depressive symptoms at Wave 4, stressful life events at Wave 3, and early life adverse experiences at Wave 5, were selected for the present study. Figure 1 illustrates the process of the study sample selection.
Figure 1. Process of the study sample selection.
Measures
Distal and proximal stressors
The distal stressors were measured by the two following instruments:
Childhood maltreatment
Childhood maltreatment was assessed with the Childhood Trauma Questionnaire-Short Form (CTQ-SF) [Reference Bernstein and Fink35]. It is a 28-item self-report inventory developed to measure five types of abuse or neglect including emotional abuse, physical abuse, sexual abuse, emotional neglect, and physical neglect in childhood or adolescence. The Cronbach’s alpha value was 0.98.
Parental-child bonding
Parental-child bonding was measured with the Parental Bonding Instrument (PBI) which is a 50-item self-administered questionnaire retrospectively measures children’s perceptions of parent-child relationships in terms of parental behaviors and attitudes before the age of 16 years [Reference Parker, Tupling and Brown36]. The questionnaire has two subscales with each consisting of 12 items for assessing care dimension and 13 items for assessing overprotection and control dimension on both maternal and paternal sides. The Cronbach’s alpha value was 0.89 for the maternal care scale and the 0.93 for paternal care scale, and 0.84 for both the maternal and the paternal overprotection subscales.
Stressful life events
Stressful life events that occurred in the past 12 months prior to the data collection (as a proxy of proximal stressors) were measured with the Life’s Events Questionnaire [Reference Laurin37]. It includes 22 items assessing five themes including work and financial issues, romantic relationships, links with family and friend relationships, housing, and experiences of aggression. Its Cronbach’s alpha value in the present study was 0.58.
Depressive symptoms
MDD was assessed using the World Mental Health Survey Composite International Diagnostic Interview (WMH-CIDI), a fully structured diagnostic interview based on the criteria of the DSM- 4th edition and the International Statistical Classification of Diseases and Related Health Problems, 10th revision (ICD-10) [38, 39]. Nine DSM-4 depressive symptoms (including depressed mood, weight loss/gain, increased/decreased appetite, insomnia/hypersomnia, and psychomotor agitation/retardation) listed in the WMH-CIDI were used in the study. We separated these symptoms into a total of 17 symptoms, including sadness, nothing rejoicing, discouragement, pessimism about future, loss of interest, decreased appetite, increased appetite, weight loss, weight gain, insomnia, hypersomnia, psychomotor retardation, psychomotor agitation, fatigue, worthlessness/guilt, poor concentration and suicidal thoughts.
Covariates
Factors, including age, sex, ethnicity, marital status, education attainment, income, immigration status, family history of mental disorders, and comorbidities in mental illnesses were also included. Age was split into four age subgroups: 18 to 29 years, 30 to 44 years, 45 to 59 years, and 60 years and above. Marital status was categorized into single, married/common-law, and separated/divorced/widowed. Education attainment was grouped as less than junior education, high school graduation, and post-high school. Four categories of personal income were generated: less than $10,000, $10,000 to $29,999, $30,000 to $59,999, and $60,000 or more. Respondents were also asked about their family history of mental disorders whether a biological parent, or both, had been diagnosed by a psychiatrist, or hospitalized or experienced any of the following mental health problems including depression, anxiety disorders, delirium or hallucination, substance abuse, and suicide. Comorbidities in mental illnesses were dichotomized into with and without subgroups based on the answers on whether the respondent had any common mental disorder including generalized anxiety disorder, panic disorder, social phobia, agoraphobia, post-traumatic stress disorder, alcohol abuse and dependence, or drug abuse and dependence.
Statistical analysis
Person-centered approaches were used to identify MDD subtypes and stressors profiles across the lifespan. LCA was first used to discover MDD subtypes based on their responses to the WMH-CIDI depressive symptoms. The optimal number of latent classes was determined by the model fit among the four consecutive models with 2-5 classes each, based upon fit indices under the following conditions: lower Akaike Information Criterion (AIC), lower Bayesian Information Criterion (BIC), lower sample size adjusted BIC, higher entropy values, and a significant bootstrap likelihood ratio test (BLRT) [Reference Arminger, Stein and Wittenberg40]. To quantify the impact of stressors across the lifespan, two approaches were used including a composite stress index and cumulative stressors profiles using the latent profile analysis (LPA) [Reference Su, D’Arcy, Li, O’Donnell, Caron and Meaney29]. Similarly, the optimal number of latent stressor profiles was determined by the above-mentioned procedure. The relationships between various stressors measures across the lifespan (the composite stress index, cumulative stressors profiles, childhood maltreatment, parent-child bonding, and stressful life events) and MDD subtypes were then assessed by multinomial logistic regressions considering those covariates (age, sex, ethnicity, marital status, education attainment, income level, immigration status, family history of mental disorders and comorbidities in mental illnesses).
To have an in-depth understanding of how distal and proximal stressors intersected and interacted in different MDD subtypes, we also compared the independent and combined effects of distal or/and proximal stressors across each MDD subtype. The predicted probabilities of each MDD subtype were calculated to estimate the MDD probability an individual exposed to a certain pattern of stressors would have.
Descriptive analysis and regression models were performed using Stata, version 15, and the LPA and LCA were conducted using Mplus, version 8. Significance was set at a 2-sided P < 0.05.
Results
The sociodemographic characteristics of the study samples are presented in Supplementary Table S1. Overall, 63.1% were females and aged 45 or more. Less than half of them were single (44.2%). Most study subjects were white (87.3%) and non-immigrants (80.9%). Approximately two-thirds of them reported receiving a post-high school education (65.4%) and earned more than $30,000 annually (67.8%). A total of 9.7% (N = 131) of the study sample were diagnosed with past 12-month MDD at Wave 4 in the current study.
Latent classes of depression subtypes
The 4-class model of MDD subtypes with the best model fit was chosen for subsequent analyses. BLRT showed no statistical significance between the 4-class model and the 5-class model, indicating that there was no improvement in model specification from the 4-class model to the 5-class model. Details of the model fit indices of different latent class models are presented in Table 1. Figure 2 illustrates various combinations of the studied depressive symptoms based on the four-class LCA model. Class 3 was labeled as “Melancholic depression” for having high intra-class probabilities of depressed mood, decreased appetite, weight loss, insomnia, psychomotor retardation, poor concentration, feeling worthless, and thoughts of death. It was the most prevalent MDD subtype and had a prevalence of 46.5%. Class 1 was named “Severe atypical depression,” which was marked by increased appetite and weight gain. The prevalence of Class 1 was 22.2%. Class 2 (prevalence: 7.1%) was labeled as “Moderate atypical depression” since it had lower probabilities of atypical depressive symptoms than the “Severe atypical depression” group. Class 4 (prevalence: 24.1%) was labeled as “Moderate depression” owing to its less pronounced and moderately severe depressive symptoms of a typical symptom pattern characterized by decreased appetite, weight loss, and insomnia.
Table 1. Model fit indices for a two-class, three-class, four-class, and five-class solutions of major depressive disorder subtypes
Abbreviations: AIC, Akaike’s information criteria; BIC, Bayesian information criteria; ssaBIC, the sample size adjusted Bayesian information criteria.
Figure 2. Latent profiles of major depressive disorder subtypes based on depressive symptoms.
Latent profiles of lifetime stressors
Similarly, a 3-profile model was selected for the lifetime stressors based on the model fit indices (see Table 2). A total of 75.1% participants reported a low level of lifetime stressors, thus was labelled as the “Low stressor” group. There were 6.8% and 18.1% of participants who had scores of high and moderate levels of stressors and were labelled as “High stressor” and “Moderate stressor” groups, respectively.
Table 2. Model fit indices for a one-class, two-class, three-class, and four-class solution for lifetime stressors
Abbreviations: AIC, Akaike’s Information Criteria; BIC, Bayesian Information Criteria; ssaBIC, the sample size adjusted Bayesian Information Criteria.
Note: Lifetime stressors considered childhood maltreatment, child–parent bonding, and stressful life events.
Associations between different stressor measures and MDD subtypes
For early life stressors, the multinomial regression model showed that emotional abuse (OR, 1.11, 95 % CI, 1.02–1.20), sexual abuse (OR, 1.09, 95 % CI, 1.01–1.18), physical neglect (OR, 1.20, 95 % CI, 1.08–1.33) and emotional neglect (OR, 1.09, 95 % CI, 1.00–1.19) were significantly independent risk factors for Severe atypical depression, but not for Moderate atypical depression (p ≥ 0.05) after controlling for age, sex, ethnicity, marital status, education attainment, income level, immigration status, family history of mental disorders and comorbidities in mental illnesses. Likewise, Melancholic depression was associated with a history of childhood maltreatment including physical abuse (OR, 1.10, 95 % CI, 1.02–1.18), emotional abuse (OR, 1.12, 95 % CI, 1.05–1.19), physical neglect (OR, 1.13, 95 % CI, 1.04–1.24) and emotional neglect (OR, 1.10, 95 % CI, 1.03–1.17). However, only emotional abuse was found to be significantly associated with Moderate depression (OR, 1.10, 95 % CI, 1.01–1.20). There was no significant relationship between different dimensions of parent-child bonding and depression subtypes, except that a weak association was found between less maternal care and Melancholic depression (OR, 1.01, 95 % CI, 1.00–1.03).
With regards to the later-on stressful life events, all the five domains of life events were significant risk factors for “Melancholic depression.” Individuals who had work and financial issues, romantic relationship-related stressors, family/friend relationship-related stressors, housing issues, experience of aggression had 1.13 (95 % CI, 1.04–1.22), 1.38 (95 % CI, 1.20–1.60), 1.21 (95 % CI, 1.12–1.31), 1.29 (95 % CI, 1.10–1.51), and 1.22 (95 % CI, 1.13–1.58) times the risk of developing “Melancholic depression,” respectively. Work and financial issues (OR, 1.15, 95 % CI, 1.03–1.27), family/friend relationship-related stressors (OR, 1.17, 95 % CI, 1.04–1.31) and experience of aggression (OR, 1.49, 95 % CI, 1.11–1.98) increased the risk of “Moderate depression,” while the remaining two domains, romantic relationship related stressors (OR, 1.31, 95 % CI, 1.07–1.60) and housing issues (OR, 1.31, 95 % CI, 1.07–1.60) increased the risk of “Severe atypical depression.”
Cumulative stressors increased the risk for “Severe atypical depression” (OR, 1.12, 95 % CI, 1.05–1.19), “Melancholic depression” (OR, 1.12, 95 % CI, 1.07–1.18), and “Moderate depression” (OR, 1.11, 95 % CI, 1.03–1.18). For different stressor profiles, compared to individuals with a “Low stressor” level, individuals with the level of “Moderate stressor” and the level of “High stressor” were 2.83 times (95% CI, 1.04–7.40) and 3.17 times (95% CI, 1.61–6.24) more likely to develop “Melancholic depression,” respectively. Moreover, those in the “High stressor” group were associated with an increased risk of “Severe atypical depression” (OR, 2.82, 95% CI, 1.08–7.36) and “Moderate depression” (OR, 3.39, 95% CI, 1.10–8.20), respectively. In contrast, no significant associations between different stressor measurements and “Moderate atypical depression” were found (see Table 3). All the above-mentioned results are based on the analyses adjusted for covariates.
Table 3. Associations between different psychosocial stress measures and major depressive disorder subtypes*
Abbreviations: CI, confidence interval; LPA, latent profile analysis; OR, odds ratio.
p < 0.05.
* Adjusted for age, sex, marital status, ethnicity, education attainment, income level, immigration status, family history of mental disorders, and comorbidity in mental illnesses including generalized anxiety disorder, panic disorder, social phobia, agoraphobia, post-traumatic stress disorder, alcohol abuse, and dependence, as well as drug abuse and dependence.
Note: OR = Odds ratio; CI = Confidence interval; LPA = Latent profile analysis.
Differences in associations between combinations of distal stressors and proximal stressors and major depressive disorder subtypes
We further compared the intersectionality of proximal and distal stressors in MDD subtypes. Table 4 illustrates differences in associations between various combinations of distal stressors and proximal stressors in four MDD subtypes. Exposure to both distal and proximal stressors increased the risk of all MDD subtypes (p ≤ 0.05) except for “Moderate atypical depression” (see Figure 3 as well). Specifically, for “Severe atypical depression,” it was more likely to occur in the presence of distal stressors (p = 0.03), regardless of whether proximal stressors were present. The proximal and distal stressors were significantly associated with “Melancholic depression” (p = 0.01), separately. Additionally, the co-occurrence of both distal and proximal stressors may have a cumulative effect, resulting in an increased risk of “Melancholic depression.” However, there was no significant difference in “Moderate depression” for those only exposed to distal or proximal stressors.
Table 4. Differences in major depressive disorder subtypes risk for different stressor profiles
Note: + indicates statistically significant difference; – indicates no difference.
Figure 3. Predicted probability of major depressive disorder subtypes across different stressor profiles.
Discussion
The present study provides one of the first pieces of evidence on the differential associations between MDD subtypes and various combinations of stressors across the lifespan in a prospective Canadian community-based cohort. The four MDD subtypes identified in the present study indicate that both symptom patterns and severity of depressive symptoms were sources of MDD heterogeneity. Overall, we found that participants with a level of “high stressor” were more likely to develop different MDD subtypes. Not all the studied stressors were consistently associated with MDD subtypes. Specifically, there was a significant association between proximal stressors (all five subdomains of stressful life events) and “Melancholic depression,” whereas “Severe atypical depression” and “Moderate depression” were only associated with some domains of stressful life events. When compared to depression-free participants, those with “Severe atypical and Melancholic depression” were more likely to have had exposure to distal stressors such as child maltreatment. The combinations of distal and proximal stressors predicted a greater risk of all MDD subtypes except for “Moderate atypical depression.” No specific stressor was associated with “Moderate atypical depression.”
Consistent with the recent review, MDD subtypes were generally differentiated by depression severity and unique combinations of probabilities of depressive symptoms [Reference Ulbricht, Chrysanthopoulou, Levin and Lapane41]. “Melancholic depression” and “Atypical depression” were the common MDD subtypes identified in the review. We discovered four MDD subtypes with various combinations of depressive symptoms and symptoms severity, including “Melancholic,” “Severe atypical,” “Moderate atypical,” and “Moderate depression.” Although no direct comparison can be made for the discovered MDD subtypes across studies as different depressive symptoms considering in each study, the four MDD subtypes identified in the current study demonstrated heterogeneous clinical manifestations.
Stress is found to be a universal process shared by both physiological and psychological systems when people cannot manage the demands or are out of balance [Reference Lupien, McEwen, Gunnar and Heim42, Reference Selye43]. Ample research has consistently shown that early life stressors, major stressful events, as well as their associated stress responses, can activate a cascade of neurochemical and/or inflammatory factors to affect limbic corticotropin-releasing hormone (CRH) functioning [Reference McEwen44]. These changes are linked to alterations of growth/anti-apoptotic factors which in turn, affect neuroplasticity, ultimately leading to psychopathology, i.e., MDD [Reference Hayley, Poulter, Merali and Anisman45]. Adversities in early life can also interfere with children’s normal development and initiate a cascade effect negatively influencing children’s health and functioning [Reference Pearlin46]. Individual variations in the stress responses can be explained by the stressors as well as the differential vulnerability theory, which argues that certain stressors increase a higher risk of mental disorders [Reference Sapolsky47, Reference Turner, Wheaton and Lloyd48]. It is important to specify how different stressors are linked to MDD. Some neurobiological studies have indicated that differences in MDD subtypes might exist in the presence of certain neurochemical disturbances [Reference Hayley, Poulter, Merali and Anisman45]. Accumulating evidence has shown that MDD subtypes can vary with regard to the hypothalamic-pituitary-adrenal (HPA) axis activity, immune function, and treatment response [Reference Lamers, Vogelzangs, Merikangas, de Jonge, Beekman and Penninx12]. For instance, “Melancholic depression” or “Typical depression” was related to CRH neuronal hypoactive HPA axis, whereas “Atypical depression” was aligned with down-regulated HPA activity [Reference Gold and Chrousos49]. Given the complexity of the neurobiological mechanisms involved in MDD subtypes, it is critical to pinpoint the specific associations between each MDD subtype and diversified stressor exposures.
We found significant associations between childhood maltreatment and severe atypical depression as well as “Melancholic depression.” This finding is in line with a review on depression and stress, which also concluded that individuals with a history of childhood abuse or neglect had a higher risk of having “Severe atypical depression” or “Melancholic depression” [Reference Mello, Juruena, Pariante, Tyrka, Price and Carpenter50]. Heim found that changes in the HPA axis could mediate clinical manifestations of depressive symptoms as a response to childhood maltreatment [Reference Heim, Newport, Mletzko, Miller and Nemeroff51]. A history of childhood maltreatment was associated with significantly higher cortisol reactivity and total cortisol exposure. It also showed an opposing pattern of response based on the severity of depression, that is, cortisol levels decreased substantively among people exposed to childhood maltreatment with severe depression [Reference Harkness, Stewart and Wynne-Edwards52]. We also found that emotional neglect predicted “Severe atypical depression,” “Melancholic depression,” and “Moderate depression” but not “Moderate atypical depression.” The higher level of unpleasantness in these MDD subtypes could account for this phenomenon since emotional neglect contributes most to the dimension of anhedonic depression, which can be reflected by the inability to feel pleasure or rejoice [Reference Van Veen, Wardenaar, Carlier, Spinhoven, Penninx and Zitman53]. However, the independent effects of physical abuse and sexual abuse on MDD subtypes have been inconclusive [Reference Spinhoven, Elzinga, Hovens, Roelofs, Zitman and van Oppen54, Reference Gibb, Chelminski and Zimmerman55].
We also discovered that recent major stressful life events increase the risk of melancholic and severe atypical depression. Stressors related to romantic relationships independently increased the risk of “Severe atypical depression,” which manifests a high level of guilt. This finding is consistent with a large-population-based twin study, which found that when depressive symptoms occurred in relation to stressful events, guilt was most common after romantic breakups [Reference Kendler, Myers and Zisook56]. Further, the independent effect of housing stressors on severe atypical depression was also identified in our study. Since “Severe atypical depression” was associated with a disturbing sense of numbness and emptiness [Reference Gold, Gabry, Yasuda and Chrousos57], it is intuitive that housing movement or eviction can be directly related to it. In the same vein, stressors related to work, finance, or family/friend issues predicted “Moderate depression,” which had a moderately depressed mood and higher rates of fatigue and insomnia. These symptom profiles were often found to be associated with stressful events regarding work stressors and family/friend strains, ultimately resulting in the occurrence of “Moderate depression” [Reference Ouellet, Savard and Morin58–Reference Vitaro, Brendgen and Tremblay60]. Harkness and Monroe suggested that those with severe melancholic depression were more likely to have non-severe stressful events prior to depression onset [Reference Harkness and Monroe61]. “Melancholic depression” appears to be an activation and persistence of the normal stress response and has increased post-challenge cortisol levels thus may be sensitized to specific stress [Reference Juruena, Bocharova, Agustini and Young62]. Additionally, the endocannabinoid system may also play an important role in the etiology of “Melancholic depression.” Similarities between “Melancholic depression” and an endocannabinoid deficiency were found including their increased sensitization to stressors [Reference Hill and Gorzalka63]. Although the association between experience of violent aggression and MDD has been well documented [Reference Gorman–Smith and Tolan64], this study provides further evidence to illustrate people with moderate depression and a higher probability of experiencing symptoms like suicidal thoughts, were more likely to report recent aggression experiences. Such experiences might explain the negative outcomes in terms of suicidal ideation [Reference Stephenson, Pena-Shaff and Quirk65].
As expected, we identified that exposures to both distal and proximal stressors substantively increased the risk of having severe atypical depression, melancholic depression as well as moderate depression. The diathesis-stress model suggests that the development of MDD is influenced by the interplay between a pre-existing diathesis, representing a susceptibility, and a later-on stressor can together activate or exacerbate the vulnerability to MDD [Reference Monroe and Simons25]. Our research findings further expand the diathesis-stress model of MDD by indicating that distal and proximal stressors are associated with specific MDD subtypes. In part, this phenomenon could be explained by the fact the stress responses triggered by early life stressors increased stress sensitization and/or proliferation to proximal stressors and thus caused uniform consequences across different psychobiologic mechanisms leading to different MDD subtypes [Reference Depue and Monroe66]. Different combinations of stressors may eventually end up with different MDD subtypes. The intersectional results of distal and proximal stressors support unique combinations of stressors associated with unique depression subtypes. Such findings bridge stress-related etiological findings with clinical management, especially pharmacological interventions had shown promising results in terms of some MDD subtypes being more sensitive to certain antidepressants (e.g., monoamine oxidase inhibitors in atypical depression) [Reference Trivedi, Rush, Ibrahim, Carmody, Biggs and Suppes67]. The potential clinical utility, as well as the practical importance of prevention and intervention, could benefit from more explicitly formulated and tailored strategies to address specific associations between stressors and MDD subtypes.
Strengths and limitations
Built upon person-centered approaches and multiple measures of stressors, this present study discovers specific associations between various stressor exposures across the lifespan and MDD subtypes in a community-based population cohort. Although stressors predicted the increased risk of MDD, different stressors and their combinations were linked to specific MDD subtypes. By utilizing an intersectional approach and the life course perspective, we were able to pinpoint the differential vulnerability to MDD subtypes when facing various exposures to distal stressors and proximal stressors across the lifespan.
There are several limitations to be noted. First, the present study utilized self-report measures to assess various stressors that might be prone to response bias. However, all stressors were assessed by validated and widely used stressor measures, such as CTQ, PBI, and SLE. Second, our study sample had predominantly Caucasians drawn from a high-income country. The generalizability of our findings to other ethnic populations or other population settings needs to be tested. Finally, some MDD subtypes were less prevalent, such as “Moderate atypical depression.” The research findings on “Moderate atypical depression” and various stressors may be impacted by insufficient sample size. Future studies with a larger sample are warranted.
Overall, various stressor profiles across the lifespan demonstrated differential impacts on MDD subtypes, which were characterized by different symptom probabilities and severity. Combinations of distal and proximal stressors dramatically increased the most common MDD subtypes, therefore early identification, targeted prevention programs, and clinical management are urgently called to reduce the negative consequences of such exposures among the most vulnerable populations. Special attention should be paid to addressing differential vulnerability among unique associations between the stressor profiles and MDD subtypes. The findings of the study direct further research into the underlying stress-related neurobiological mechanisms involved in diversified clinical manifestations of MDD.
Supplementary material
The supplementary material for this article can be found at http://doi.org/10.1192/j.eurpsy.2024.1783.
Data availability statement
The ZEPSOM data is not currently freely available to researchers in general due to ethical and data management requirements. Interested researchers can directly contact the Research Team at: [email protected].
Author contribution
J.C. has developed the Longitudinal Montreal South-West Catchment Area Study (ZEPSOM) and has directed and supervised the data collection. X.M. conceived the idea of this study and together with Y.S. and D.L. drafted the manuscript. Y.S. conducted the statistical analyses. M.L. and J.C. provided critical feedback towards developing the research question and results interpretation and were involved in manuscript redrafting. All authors have read and approved the final draft.
Financial support
This work was supported by the Canadian Institute of Health Research (CTP-79839; PJT-148845); and the Canada First Research Excellence Fund provided to McGill University for Healthy Brains for Healthy Lives. These funders had no role in study design, data collection, and analysis, decision to publish.
Ethical standard
The authors assert that all procedures contributing to this work comply with the ethical standards of the relevant national and institutional committees on human experimentation and with the Helsinki Declaration of 1975, as revised in 2008. All procedures involving human subjects were approved by the Institutional Review Board of the Douglas Mental Health University Institute (#IUSMD-18/17).
Competing interest
The authors declare none.
Open access
Comments
No Comments have been published for this article.