Clinical characteristics and socio-demographic features of psychotic major depression

Psychotic major depression (PMD) is a special subtype of depression with a worse prognosis. Previous studies failed to nd many differences among patients with PMD versus those with non-psychotic major depression (NMD) or schizophrenia(SZ). This study compared psychotic major depression with nonpsychotic major depression and schizophrenia based on sociodemographic factors (including season of conception) and clinical characteristics. We aimed to provide data to inform clinical diagnoses and etiology research. This case–control study used data for patients admitted to Shandong Mental Health Center from June 1, 2016 to December 31, 2017. We analyzed cases that had experienced a PMD episode (International Classication of Diseases, Tenth Revision codes F32.3, F33.3), NMD (F32.0–2/9, F33.0–2/9), and SZ (F20–20.9). Data were collected on sex, main discharge diagnosis, birth date, ethnicity, family history of psychiatric diagnoses, marital status, age at rst onset, educational attainment, allergy history, and existence of trigger events.

Psychotic symptoms are likely to be a risk factor for conversion from unipolar depression to bipolar disorder [11,12] and schizophrenia (SZ) [13]. A longitudinal study showed that within 2 years of their rsthospitalization, 41% of patients initially diagnosed with PMD met the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition criteria for bipolar or schizoaffective disorders [14]. Psychiatric comorbidity has also been reported to be common in PMD [15]. In addition, depression is common in different stages of SZ and may interplay in SZ progression, which has raised questions about the validity of a PMD diagnosis [16,17]. Therefore, further work on the differential etiology of PMD from other psychoses is needed [18].
Few studies have been conducted to clarify the risk factors for PMD, especially early life risk factors [19]. In general, previous studies failed to nd many demographic differences among patients with PMD versus those with non-psychotic major depression (NMD) [20]. However, there were some exceptions, with the main nding being that a family history [19,21] of psychosis increased the risk for PMD. Other ndings included that patients with PMD were less likely to be Caucasian [4,14] and have lower educational attainment [4,19] compared with those with NMD. One study observed more years of education in those with PMD compared with NMD [22]. In terms of early childhood and adolescent risk factors, patients with PMD were signi cantly more likely to report histories of physical or sexual abuse [24] and have lower sports grade in school [21] compared with patients with NMD.
Studies comparing PMD and SZ found a higher proportion of females in the PMD group than in the SZ group [19,23]. Most studies that compared the mean onset age of PMD and SZ reported SZ had an earlier onset age [19,23]. In one study, patients with PMD were less often single compared with those with SZ [25]. A number of psychosocial risk factors have also been associated with a follow-up diagnosis of PMD and SZ, including living alone, having a basic-level quali cation, being unemployed, having less than monthly contact with friends, having no close con dants, and having experienced childhood adversity [18].
In traditional Chinese medicine, the annual rhythm of the dominant seasons is believed to be an important force that maintains the stability of our living environment, with the structure and function of the human body adapting to this rhythm and changing regularly. Seasonal factors can potentially exert an in uence before birth. Accumulating evidence suggests that environmental factors pertaining to early life are associated with alterations in gene expression regulated by epigenetic factors. These factors include placental dysfunction [26], maternal malnutrition during pregnancy [27], psychological distress during pregnancy [26], and infections during the gestational and postnatal periods [28,29]. Therefore, it is possible that these factors also affect susceptibility to neuropsychiatric diseases in later life [26].
Previous studies have investigated potential associations between season of birth and major depression [30,31]. However, few studies have investigated associations between season of birth/conception and psychotic features in patients with major depression. In the present study, inpatients with PMD were compared with those with NMD and SZ using measures of sociodemographic factors (including season of conception) and clinical characteristics. We examined the relative importance of various clinical features of PMD to identify characteristics reliably associated with this diagnosis.

Setting
This paper was based on data from patients admitted to Shandong Mental Health Center from June 1, 2016 to December 31, 2017. The database from which our data were acquired contained patients' admission number, age, sex, admission year, main discharge diagnosis, birth date, ethnicity, family history of psychiatric diagnoses, marital status, age at rst onset, educational attainment, allergy history, and existence of trigger events.

Sample
Cases that had experienced an episode of PMD (International Classi cation of Diseases, Tenth Revision codes F32.3, F33.3), NMD (F32.0-2/9, F33.0-2/9), and SZ (F20-20.9) were included in our analysis. In total, 242 multiple admissions for the same patient were deleted from the included cases. Twenty cases with evidence of psychotic symptoms precipitated by an organic cause or developmental retardation were excluded. Patients who were not born in Shandong province were also excluded (78 patients with SZ, 27 patients with NMD, and ve with PMD). This left, 98 patients with (PMD), 351 with NMD, and 967 with SZ for inclusion in our analysis.

Clinical characteristics
Family history of psychiatric diagnoses was de ned as at least one family member with a mental disorder within the patient's rst-, second-, and third-degree relatives. Age at onset was de ned as the patient's age at the rst episode of the disorder. Impactful negative life events within 1 year before the rst diagnosis with the disorder were de ned as trigger events. A positive allergy history was de ned as having been allergic to something.

Sociodemographic factors
Marital status was determined according to the patient's situation at the time of treatment. For example, if a patient had been divorced but had remarried at the time of the medical record, they were classi ed as "married." Educational attainment was divided into four categories: primary school or below, junior high school or vocational school, senior high school or junior college, and university and above. Patients' ethnicity was classi ed as Chinese Han population and ethnic minorities.
The date of conception was calculated according to patients' date of birth. The expected delivery date is 280 days from the rst day of the last menstrual period of a pregnant woman. We assumed that fertilization occurred during ovulation (i.e., the 14th day of the last menstrual period), meaning the date of birth minus 266 days would be the date of conception. The season of conception was classi ed according to the month of conception (spring: Feb-Apr, summer: May-Aug, autumn: Sep-Oct, winter: Nov-Jan). All factors were self-reported or reported by family members accompanying the patient.

Statistical analyses
Differences in missing data between diagnostic groups were compared using Fisher's exact test. The sociodemographic variables and clinical characteristics of the diagnostic groups (PMD vs. NMD and PMD vs. SZ) were compared using independent samples Mann-Whitney U tests or Pearson's chi-square tests. Variables with a P-value less than 0.15, family history of a psychiatric diagnosis, and season of conception were incorporated into the generalized linear model (GLM). Odds ratios (OR) were calculated using GLMs with binomial distribution and logit link. Cases with missing data were automatically dropped from each analysis by SPSS version 22. The ndings were considered statistically signi cant when the two-tailed analysis resulted in a P-value <0.05. Categories of variables were transformed into serial numbers, including sex (1=male,2=female), marital status (1=umarried,2=married, 3=widowed,4=divorced), family and allergy history(1=yes,0=no),season of conception(1=spring, 2=summer,3=autumn,4=winter), educational attainment(1= Primary school or below,2= Junior high school or vocational school,3= Senior high school or junior college,4= University and above), ethnicity (1=Han,2=minority).

Results
Missing data (Table 1) Table 1 shows the comparison of missing data between the PMD, SZ, and NMD groups. There were no statistically signi cant differences between the diagnostic groups in terms of missing data.

Model for relative utility of clinical and demographic variables between groups
In the GLM comparing PMD with NMD, we included variables with a P-value <0.15 and season of conception. As shown in Table 3, fewer patients with PMD had a primary school or below education compared with the NMD group (P=0.022, OR 0.397, 95% con dence interval [CI]: 0.18-0.874). Not having a family history of a psychiatric diagnosis was a protective factor for PMD (P=0.027, OR 0.557, 95% CI: 0.332-0.937). Season of conception showed no statistically signi cant difference between the groups. A GLM was conducted to examine the relative utility of various clinical and demographic variables in differentiating patients with PMD versus SZ. Demographic and clinical variables that had a P-value <0.15, family history of a psychiatric diagnosis, and season of conception were tested. This included sex (male, female), marital status (unmarried, married, widowed, divorced), family and allergy histories (yes, no), season of conception (spring, summer, autumn, winter), and educational attainment (primary school or below, junior high school or vocational school, senior high school or college, university or above).
Results from the model are presented in Table 4. Educational attainment, existence of trigger events, age at rst onset, and allergy history signi cantly differentiated the diagnostic groups when other variables were controlled in the model (P<0.01). Patients with a primary school or below education and that were unmarried (P<0.05) were less likely to have PMD compared with SZ. Patients with PMD were more likely to have trigger events, be allergic to something, and be older at the age at rst onset compared with those with SZ.

Discussion
Although season of conception was not a factor that differentiated PMD from NMD and SZ in the present study, we did observe interesting ndings. Patients with PMD appeared to be better educated than those with SZ and NMD, which was partly inconsistent with previous studies that reported no signi cant difference (or the opposite) in education between NMD and PMD [4,19,32]. A reason for this could be that previous studies mainly considered college degree versus no college degree [4,32], whereas we divided educational attainment into four groups; only patients with primary school and below education showed a statistically signi cant difference between groups.
The ndings reported in this paper highlighted that people with PMD were more likely to have a family history of a psychiatric diagnosis compared with people with NMD, which was consistent with previous studies [20]. Similar to other studies [19], we found no differences in family history between PMD and SZ. Our ndings concerning onset age in PMD versus SZ were consistent with earlier studies [19,23]. Previous studies comparing the onset age of PMD with NMD were not explicit. In some studies, patients with PMD had an earlier onset age [22], whereas in others, the onset age showed no statistically signi cant differences [33]. In our study, patients with PMD were younger than those with NMD patients at the rst episode of the disorder, but this difference was not statistically signi cant.
Despite differences in onset age and marital status [25] that were previously reported in other studies, we found no statistically signi cant sex differences between the PMD and SZ groups, although there were more female patients in the PMD group (55.1%) than in the SZ group (46.6%). However, differences in the sample size may explain the discrepancies between the studies [19,23].
The analysis of differences between Chinese ethnic minorities and Han population received little attention in past epidemiological research that compared PMD with NMD and SZ. Our nding of no statistically signi cant differences in ethnicity was not surprising as there were few patients recorded as ethnic minorities. A previous study found that experiencing a major life event in the year before illness onset had a substantial effect size, but did not meet statistical signi cance compared PMD (P = 0.058) and SZ (P = 0.056) to a population-based sample of controls without a history of psychosis [18]. In the present study, more patients with PMD reported impactful negative life events within 1 year before the rst onset of the mental disorder compared with those with SZ, indicating a difference in pathological mechanisms between the two disorders. However, it is important to note that patients with SZ may be more likely to have underreported trigger events because the nature of their disease.
To our knowledge, this is the rst study that included allergy history in the analysis. Surprisingly, we found statistically signi cant differences when comparing SZ with PMD. This result requires further veri cation, but may offer some enlightenment for pathological studies focused on PMD.

Limitations
The classi cation of some in uencing factors in this study was a little unclear. For example, family history could have been divided more speci cally according to parents or other relatives and different kinds of psychiatric diseases. The presence of trigger events was mainly determined by patients, and was not assessed and de ned using evaluation tools such as the Life Events and Di culties Schedule.
Another limitation was that we failed to investigate the potential role of comorbid disorders, which probably resulted in bias.

Conclusion
Our results suggest that patients with PMD are similar to those with NMD in terms of demographic variables and clinical characteristics, although they are more likely to have a family history of a psychiatric diagnosis in their rst-, second-, and third-degree relatives and to have obtained a primary school degree. The small number of signi cant factors may point to underlying heterogeneity. There were more differences between patients with PMD and those with SZ, with the difference in allergy history suggesting a direction for further research. In general, major depression is a highly heterogeneous disorder [34]. It is known that there are clinical differences between PMD that has its onset in early adulthood and PMD in old age [19,35]. Therefore, further re nements in screening and treatment are needed for this clinical population.

Declarations
Ethics approval and consent to participate This study was conducted in accordance with the tenets of the Declaration of Helsinki and was approved by the ethical committee of the National Natural Science Foundation of China. The research did not directly involve any human body. Patient data were extracted from monthly reports maintained by Shandong Mental Health Center. All individual-level data were anonymous.

Consent for publication
This manuscript is approved by all authors for publication.

Availability of data and materials
Research data has been uploaded with the paper in supplementary le "data.xlsx" and can be requested from doramaia@sina.com.

Competing Interests
The authors have declared that no competing interests exist.

Funding
This study was supported by the National Natural Science Foundation of China (ID number 81574098). The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript. All authors state that this research was conducted independently of the in uence of funding bodies.