The sample for this study was recruited from a day hospital unit of a specialized psychiatric service for the assessment and treatment of early-onset mood disorders, managed by the Mood Disorders Program team at Bambino Gesù Children’s Hospital in Rome. Subjects were referred to confirm a diagnostic hypothesis of a moderate/severe mood disorder. Referral sources included primary care doctors (pediatricians and general practitioners), secondary care providers (child psychiatrists in national healthcare centers), and the emergency department of Bambino Gesù Children’s Hospital.

Historical data obtained during clinical assessments performed in our outpatient service from June 2013 to March 2022 were extracted from clinical records. We included consecutively referred children and adolescents, aged 6 to 18 years at the time of evaluation. Categorical diagnoses were clinically made and confirmed with the Kiddie-Schedule for Affective Disorders and Schizophrenia for School-aged Children, Present and Lifetime version (K-SADS-PL) (26) following the DSM-5 criteria. We included subjects diagnosed with a current MDE in the context of either an MDD or BD. Subjects diagnosed with persistent depressive disorder with intermittent major depressive episodes, with current MDE, and subjects diagnosed with MDE with mixed symptoms were also included.

We excluded patients with MDE in partial or complete remission at the time of evaluation, those with intellectual disability, autism spectrum disorder, substance-induced mood disorders, or mood disorders due to other medical conditions, as defined by DSM-5. Parents/legal guardians provided written informed consent for anonymous reports of data in aggregate form, in compliance with research ethics. Data collected were stored in individual clinical records. The study has been conducted according to the guidelines established in the Declaration of Helsinki. Considering the retrospective nature of the analysis, the current study did not require the approval of the local ethics committee according to current legislation, but a notification was sent (Comitato Etico Ospedale Pediatrico Bambino Gesù practice number 3380). Data were retrospectively anonymously analyzed in line with personal data protection policies.

All subjects were evaluated over a minimum of three appointments, totaling 9–10 h of clinical assessment. All study participants were assessed by an experienced child and adolescent psychiatrist and an experienced psychologist using the K-SADS-PL (26), a semistructured, clinician-administered diagnostic interview performed with both subjects and their parents or adult legal representatives to assess current and past psychopathological features and psychiatric disorders, including primary mood disorders and comorbidities in juveniles, according to the DSM-5 criteria. Childhood clinical features (syndromal or subsyndromal) preceding the onset of current disorders were systematically assessed and coded both as “traits” if subsyndromal or “disorder” if meeting the criteria for a DSM-5 diagnosis, with particular attention to the time of onset of first mood and non-mood disorder-related psychiatric symptoms (27).

Additionally, depressive and hypo/manic or mixed symptoms were rated by the same experienced clinicians using the Children’s Depression Rating Scale-Revised (CDRS-R) (28) and the K-SADS Mania Rating Scale (KMRS) (29). The CDRS-R is a semistructured interview used to rate depressive symptoms in subjects aged 6–18 years, based on 17 items with a raw score range of 17–120 (sum of the scores of single items, considered positive at scores >30 corresponding to a T-score >55). We analyzed the CDRS-R total score and the single items’ scores (school dysfunctions, difficulties in having fun, difficulties in interpersonal relationships, sleep disorders, appetite disorders, excessive fatigue, psychosomatic complaints, irritability, excessive guilt, low self-esteem, depressive feelings, morbid ideas, suicidal ideation, excessive crying, reduced facial expressions, slow speech, and motor hypoactivity). The KMRS is a structured interview used to rate hypo/manic symptoms in subjects aged 6–18 years, based on 14 items, with a total score range of 1–68 (sum of the scores of 0 to 6 for single items minus 13, considered indicative of significant hypomanic symptoms at scores >12). We analyzed both the total score and the single items’ score (euphoria and expansiveness; irritability and anger; mood lability; reduced need for sleep; crowded thoughts; increased energy; increased activities; motor hyperactivity; grandiosity; rapid or pressured speech; distractibility; impaired judgment; hallucinations; and delusions). Single items’ score of <2 is considered normal, a score of 2–3 is considered borderline, and a score >3 is considered clinically significant (29). Depression severity at the time of evaluation—whether mild, moderate, severe, with or without psychotic symptoms, or in partial/complete remission—was defined based on clinical judgment according to the DSM-5 criteria and CDRS-R cutoffs: a total score of 30 to 44 for mild, 45 to 55 for moderate, 55 or more for severe depression, and 29 or less for remission.

Suicidal ideation and suicidal behaviors were evaluated with the Columbia Suicide Severity Rating Scale (C-SSRS) (30). Suicidal behavior is defined as any self-harming behavior resulting in any kind of damage with non-zero intent to die, declared by the patient or evident from documented circumstances. Non-suicidal self-injury (NSSI) was systematically investigated lifetime and accurately differentiated from suicidal behavior and defined as in section III of DSM-5 (31).

To assess self-reported symptoms, the Children’s Depression Inventory-2 (CDI-2) (32) and the Multidimensional Anxiety Scale for Children-2 (MASC-2) (33) were administered to patients and their parents/caregivers. For each module of CDI-2 and MASC-2, the total T-score was analyzed.

Finally, the Italian version of the Child Behavior Checklist (CBCL) (34) was administered to parents/caregivers to score different behavioral and emotional problems severity. It has strong psychometric properties and is widely used in Italy (35, 36). It provides scores on three behavior rating scales that address internalizing symptoms, externalizing symptoms, and total behavioral problems. Sub-items of these scales include syndromic and DSM-oriented scales.

An overall Deficient Emotional Self-Regulation CBCL profile was calculated by summing scores for attention problems, aggression, and anxious-depressed syndromic scales. This score is reported to be indicative of Deficient Emotional Self-Regulation at scores between 180 and 210 (1 to 2 standard deviations) and as meeting the criteria for a Dysregulation Profile at a >210 (>2 standard deviations) score (37).

Categorical variables were presented as counts and frequency rates, while continuous variables were expressed as means and standard deviations. For scales with missing data, participants were excluded from analyses, and no imputation methods were applied.

Associations between unpaired categorical variables were tested using Pearson’s chi-square test, with Fisher’s exact test applied when more than 20% of cells in the contingency table had expected frequencies <5 (38). Mean differences of continuous variables between the two groups were tested using Student’s t-test for unpaired data.

We performed a logistic regression analysis with female sex as the dependent variable, to assess which of the studied variables best predict the MDE phenotype in female patients, while controlling for other factors. In more detail, our aim was to determine how accurately and independently the most relevant clinician-, self-, and parent-rated measures can distinguish between male and female patients with a current MDE based on psychiatric symptom expression. Of the CDRS-R items, KMRS items, CBCL scales, and MASC-2 and CDI-2 T-scores, variables found significantly different between sexes at previous bivariate analyses were tested as predictors, while female sex was set as the predicted value. Predictor selection was performed using a stepwise backward approach, progressively removing variables with p >0.05. Odds ratios (OR), with their 95% confidence intervals (95% CI), p-values, and Wald test statistics have been presented. All tests were two-tailed. A p-value of 0.05 or less was considered indicative of significance. Analyses were conducted with Microsoft Office 365—Excel and IBM SPSS Statistics V26 software.

After the application of the exclusion criteria, we retrospectively identified 382 patients diagnosed with current MDE who were eligible for inclusion in the study. Of these participants, 73.8% (N = 282) were female patients and 26.2% (N = 100) were male patients. Male patients were significantly younger at the time of evaluation (14.0 years vs. 15.2 years; p = 0.001). They were also younger when their first psychiatric symptom was reported by parents/caregivers (6.32 years vs. 7.75 years; p = 0.003) and when the first symptoms of their mood disorder were reported (11.3 years vs. 12.5 years; p = 0.005).

Bipolar disorder was diagnosed in 13.6% (N = 52) of the participants. This diagnosis was more common in male participants (21% vs. 11%; p = 0.012). Conversely, MDD diagnosis was significantly more frequent among female participants (78% vs. 61%; p < 0.001).

Thirty-seven percent of the subjects had a severe MDE, with a CDRS-R total score >55. Most patients were classified as having a moderate severity depressive episode (CDRS-R total score between 45 and 55). Severe MDEs were more frequently diagnosed in female than in male participants (41.5% vs. 26.0%; p = 0.006); mild MDEs, in turn, were more frequently diagnosed in male participants (33.0% vs. 16.7%; p < 0.001).

The most frequent comorbidities were anxiety disorders and obsessive-compulsive or tic disorders (27%), followed by ADHD/behavior disorders (11.8%), learning disorders (11.3%), substance abuse (8.4%), and eating disorders (6.5%). Only ADHD/behavior disorders were significantly more prevalent among male participants than female participants (20% vs. 8.9%; p = 0.003). No other significant differences in comorbidities were observed between male and female participants. For further details on participants’ demographics and psychiatric diagnoses, see Table 1 .

More than 60% of the participants reported having experienced at least one episode of active suicidal ideation during their lifetime. History of active suicidal ideation was more commonly reported by female than male participants (63.9% of female patients vs. 47.0% of male patients; p < 0.001). Almost one-third of female subjects reported a lifetime history of one or more suicide attempts, while male subjects reported suicide attempts significantly less frequently (29.4% vs. 13.0%; p = 0.001). Consistently, female patients scored higher on the C-SSRS screening version (2.24 ± 1.73 vs. 1.02 ± 1.36; t = 4.73; p < 0.001). Non-suicidal self-injury (current or lifetime) was also reported in approximately half of the sample and was more commonly reported by female patients (58.5% vs. 27.0%; t = 29.3; p < 0.001). For further details, see Table 1 .

Depression severity, measured by the CDRS-R total score, was significantly higher among female than male patients (52.0 ± 12.5 vs. 47.2 ± 12.1; t = 3.35; p < 0.001; Table 2 ). Several depressive symptoms were also significantly more severe in female participants compared to male participants, including excessive weeping (3.51 vs. 2.14; p < 0.001), excessive fatigue (3.85 vs. 2.89; p < 0.001), low self-esteem (4.37 vs. 3.61; p < 0.001), depressive feelings (4.43 vs. 3.67; p < 0.001), suicidal ideation (2.67 vs. 2.11; p = 0.005), appetite disturbance (2.78 vs. 2.18; p < 0.001), physical complaints (3.02 vs. 2.39; p = 0.002), and excessive guilt (3.09 vs. 2.41; p < 0.001; Table 2 ). No depressive symptom was found to be more severe in male participants.

The severity of hypo/manic symptoms, measured by the KMRS total score, was similar in mean values between male and female patients, but significant mixed hypo/manic symptoms (defined by the KMRS total score above the instrument threshold of 12) were identified in more male than in female participants (17% vs. 7.7%; p = 0.010; Table 1 ). Male patients scored higher than female patients in KMRS increased motor activity (1.43 vs. 1.25; p = 0.031) and KMRS distractibility (1.55 vs. 1.37; p = 0.047), whereas KMRS mood lability was significantly higher in female patients (2.62 vs. 2.35; p = 0.037). For further details on clinician-rated symptoms, see Table 2 .

Both self-reported and parent-reported anxiety symptoms, measured by the MASC-2 total score, were significantly higher among female than male patients (69.3 vs. 56.4, p < 0.0001 and 69.0 vs. 63.2, p = 0.025, respectively). Self-reported depressive symptoms, measured by the CDI-2 total score, were also significantly higher among female than male patients (74.8 vs. 66.1; p = 0.001; Table 1 ). Parent reports of depressive symptoms, however, did not show significant differences between groups.

Different internalizing and externalizing parent-rated symptoms and behaviors assessed by the CBCL were different between the sexes. Post-traumatic stress problems (73.7 vs. 70.0; p = 0.004), attention problems (65.1 vs. 62.6; p = 0.025), internalizing problems (71.8 vs. 68.3; p = 0.002), the anxious/depression scale (73.5 vs. 70.5; p = 0.026), the somatic complaints scale (69.9 vs. 66.0; p = 0.011), obsessive-compulsive problems (68.2 vs. 64.8; p = 0.007), and affective problems (76.6 vs. 73.6; p = 0.007) were more severe in female than male patients. Conversely, male participants had significantly higher scores in rule-breaking behaviors compared to female participants (62.6 vs. 59.9; p = 0.012). For further details on self/parent-rated symptoms, see Table 2 .

To determine the relative importance of each symptom in differentiating between male and female participants with an MDE, a logistic regression model was developed using items that showed significant differences between sexes in the bivariate analyses. The model accurately predicted 80% of cases, correctly classifying 92.5% of female participants and 46.9% of male participants.

The following factors were significantly and independently associated with female gender: a) CDRS-R excessive weeping item score (OR 1.53; p < 0.001), b) KMRS lability item score (OR 1.50; p = 0.014), c) CDRS-R excessive fatigue item score (OR 1.38; p = 0.002), d) CDRS-R appetite disturbance item score (OR 1.28; p = 0.041), and e) CBCL attention problems scale score (OR 1.07; p = 0.001). Factors significantly and independently associated with male gender included a) KMRS distractibility item score (OR 0.55; p = 0.009) and b) CBCL conduct problems scale score (OR 0.93; p = 0.001). For further details on the regression model, see Table 3 .

Depressive symptoms were more severe in female participants, as assessed by both clinicians using the CDRS-R and patients using the CDI-2. However, parent-rated depressive symptoms (CDI-2 parent-report module) did not differ between the sexes. Discrepancies between self- and parent-reported measures of depression are well-documented, with self-report generally recommended as the primary source for assessing depressive symptoms in pediatric patients (39). Some studies suggest that nearly half of parents struggle to distinguish between normal mood swings and signs of depression (40), highlighting the challenges in parental recognition of depressive symptoms. However, it must be acknowledged that some authors have questioned the wording of the CDI instrument and the use of discriminative memory required, which can pose difficulties for younger children (41). Given this, we primarily relied on the clinician-rated CDRS-R for our assessment of depression severity. Furthermore, the consistency of results between the CDI-2 self-report module and the CDRS-R strengthens our findings pointing toward a meaningful difference between sexes in both the CDRS-R and the CDI-2 self-report module, while the CDI-2 parent-report module appears less sensitive to detecting these differences.

Higher depressive symptom severity in female participants is in line with previous findings (15) reporting small but significant sex differences in adolescent depression severity and among specific symptoms, including guilt, body image dissatisfaction, concentration problems, depressed mood, fatigue, and health concerns. Our more homogeneous sample, characterized by greater severity and evaluation across a broad range of psychiatric symptoms, allowed us to identify additional differences in affective symptoms, also including mixed features. Additionally, our findings showed greater consistency across raters, and our regression model demonstrated improved fit and classification accuracy.

The higher severity of depressive symptoms in female patients with early-onset mood disorders can be understood within a complex biopsychosocial framework. Several vulnerability factors for depression—such as genetic predispositions, pubertal timing, affective temperament, and cognitive response styles—differ by sex (42). Adolescence brings psychological and physical changes that influence self-perception, interpersonal experiences, and mental health outcomes within a socio-environmental context. Cultural, social, and familial factors are significant, as gender-specific social pressures affect male and female adolescents differently. Girls may also experience a difference in the incidence of adverse childhood events, which we did not investigate, but could partly explain the observed differences (43). Beyond exposure to stressors, the attribution of significance to life events and coping responses differs by sex (44), with relational problems among peers and rumination as a stress-coping strategy more prominent in female subjects (45). This may contribute to the greater expression of core depressive symptoms in female respondents. Recent findings also highlight the role of neurobiological factors in sex-specific susceptibility to depression. Studies of adolescents with depression link gender differences to variations in the hypothalamic–pituitary–adrenal axis response to stress (46). Increased stress reactivity during late puberty and reduced cortisol secretion in early puberty have been identified as risk factors for depression in female adolescents (47). Sex-specific neuroimmune-endocrine factors involved in stress responses and depression have been identified (48), as well as structural nervous system differences observable via magnetic resonance imaging (49). These findings underscore the interplay between biological, psychological, and social factors in explaining the higher depressive symptom severity in female subjects.

Sex differences in appetite disturbances cannot be attributed to differing rates of comorbid eating disorders, which are rare in our sample. This reflects our unit’s referral process, where patients with prominent eating symptoms are directed to a specialized eating disorder unit. Consequently, our homogeneous sample strengthens the study by ensuring that observed differences reflect sex-specific mood disorder expression.

Additionally, clinician- and parent-rated somatic complaints were significantly associated with MDEs in female patients, consistent with previous research (50). We propose that excessive fatigue, appetite disturbances, and somatization in female patients may represent a subset of atypical depressive symptoms. This aligns with studies in adults showing a higher prevalence of atypical MDEs in women (51). However, atypical depression remains understudied in pediatric age.

To our knowledge, sex differences in subthreshold mixed hypomanic symptoms have not been previously studied in children and adolescents with MDE. Our earlier study assessed fewer adolescents with full-threshold mixed symptoms per the DSM-5 criteria (24). In the present study, mixed features (subthreshold or threshold) were systematically rated using the KMRS in a larger population diagnosed with an MDE. Significant mixed hypomanic features (KMRS total score at threshold) were present in ~10% of the sample and were more frequent in male patients. Male participants were also twice as likely to be diagnosed with pediatric BD at their first psychiatric evaluation for MDE. Prior pediatric studies reported no sex differences in pediatric BD diagnosis (23), but in earlier samples of children diagnosed before age 12, 85% were male patients (52). Consistent with our findings, adult BD cohorts show that female respondents more often experience depressive episodes at onset (53), more depressive episodes throughout illness (54), and mixed features during hypomania or mania (55). Adolescent BD cohorts tend to show hypomanic or manic episodes more frequently in male patients (23). Furthermore, in children and adolescents with severe early-onset mood disorders, conversion from MDD to BD is common during follow-up, affecting up to 45% of cases (56). The DSM-5 criteria limit BD diagnosis at first assessment, as follow-up often reveals additional affective episodes meeting the BD criteria over time. Thus, BD diagnoses at first evaluation in our sample may reflect only a subset of cases where the current MDE is neurobiologically linked to BD. This higher rate of early BD diagnosis in male patients suggests that the DSM-5 criteria more readily recognize BD in male patients, who typically present with manic polarity and distinct episodes, compared to female participants, who more often present with severe MDE and persistent depressive symptoms. Additionally, male patients exhibit more recognizable DSM-5 hypo/manic features in mixed states, as noted in our previous work (24).

Consistent with earlier findings (24), male participants in this study more frequently reached the KMRS threshold, while subthreshold mixed hypomanic features were similarly represented in both sexes, with sex differences in mixed symptom expression. Female participants showed more severe emotional lability, whereas male participants exhibited greater motor activity and distractibility, reflecting a pattern where male patients express more “outer” psychomotor symptoms and female patients display more “inner” emotional lability. While motor activation and distractibility align with DSM-5 mania and mixed features criteria, emotional lability does not. Given the prognostic importance of mixed hypomanic symptoms in adolescents with MDE, even when subthreshold (57), we stress the need for high clinical sensitivity in assessing such symptoms in female patients. Future research should include longitudinal studies to verify these findings and explore the utility of alternative tools, such as the Koukopoulos Mixed Depression Rating Scale (58), which may better capture “inner excitation” symptoms and enhance diagnostic sensitivity beyond current DSM-5 criteria.

We observed that the first psychiatric symptoms emerged over a year earlier in boys than in girls. This aligns with the higher incidence of comorbid ADHD in male participants and is consistent with previous findings on psychiatric comorbidities in mood disorders and the well-documented BD-ADHD comorbidity in pediatric populations (27, 37). These findings highlight the importance of assessing potential neurodevelopmental disorder traits in patients with early-onset mood disorders.

Our findings confirm the higher prevalence of suicidal ideation and attempts in female patients compared to male patients, consistent with extant literature (15, 24, 59). This aligns with the “gender paradox” in suicidal behavior, where women are more likely to attempt suicide, while men have higher rates of suicide completion and often use more lethal methods (59). Explanations for this phenomenon often point to greater impulsivity and higher rates of substance abuse in men (60).

A recent review of adolescents highlights this differential pattern, identifying both common and sex-specific risk factors for suicide attempts. Female-specific risk factors include eating disorders, PTSD, BD, dating violence victimization, depressive symptoms, interpersonal problems, and previous abortion. Male-specific factors include disruptive behavior, hopelessness, parental separation/divorce, peer suicide, and access to means (61). Similarly, a cohort of Chinese college students found suicide-related outcomes linked to depressive symptoms in female students and anger in male students (62). In our study, suicidal ideation and attempts were more frequent in female participants and correlated with their higher burden of depression severity, further supporting the interplay between depression and suicidal behavior in female subjects.

Analysis of parent-reported symptoms in the CBCL revealed notable sex differences: female subjects scored higher on internalizing symptoms, while male subjects scored higher on conduct problems. This aligns with the findings in adults with unipolar depression (6) and BD (63), where women are reported to experience comorbid anxiety and post-traumatic symptoms more frequently than men, while men show higher rates of conduct and substance use disorders. Researchers have noted that female participants are more likely to manage emotional problems with internalizing strategies, whereas male participants often adopt externalizing “escaping behaviors” (64).

Of note, no significant sex difference in substance use was observed in our sample, likely due to the study’s focus on early-phase juvenile-onset mood disorders. Such differences may become evident in adults during more chronic phases of these disorders.

Finally, female respondents reported higher levels of self- and parent-rated anxiety symptoms than male respondents, consistent with previous studies in adults (65) and adolescents (66).

This study enriches the literature by providing a detailed, multidimensional analysis of sex differences in severe early-onset mood disorders, a topic rarely explored in comparable populations. A key strength is the use of a single-center sample of referred early-onset mood disorders, ensuring population homogeneity and consistent structured assessments for all participants. However, several limitations should be noted. First, the high severity of depressive episodes in this clinically referred sample may limit the generalizability of the findings to broader clinical populations. Second, while the CDRS-R and KMRS are internationally recognized tools widely used in Italy (24, 27, 67), their validity and reliability in Italian adolescents have not been formally addressed. Reliability analysis in this sample showed acceptable internal consistencies (Cronbach’s alpha: 0.669 for KMRS, 0.796 for CDRS-R). The lower KMRS consistency may reflect the uneven distribution of hypo/manic symptoms, as significant symptoms were confined to participants with subthreshold mixed features, with many participants with a low score of all KMRS items. Third, the study did not assess adverse childhood events, which may vary by sex and influence the findings.

Our study identifies a moderately distinct pattern of sex differences in depressive symptoms among children and adolescents with early-onset mood disorders, partially mirroring adult findings. Consistent with adult literature and limited adolescent studies, we confirm that female participants with MDEs experience more severe depressive symptoms, anxiety, and suicidal behaviors, while male participants more frequently exhibit disruptive behaviors. Moreover, our study highlights sex-based differences in mixed hypomanic symptoms, pediatric BD diagnoses, and early childhood clinical features preceding mood disorder onset.

These findings should be understood within a framework where the interplay of biological, psychological, and social factors begins in childhood and continues into adolescence and adulthood. This interaction shapes sex-specific depression development and intersects with diagnostic systems that often overlook sex and gender differences in key clinical presentations. These observations highlight the importance of integrating a sex- and gender-informed perspective in research, diagnosis, and treatment to enhance outcomes for adolescents with mood disorders.