Schizophrenia is a neurodevelopmental disorder with relatively stable prevalence estimated at about 0.5% of the general population.1 There is now a general consensus that schizophrenia is characterized by high clinical heterogeneity in terms of symptomatic manifestation and outcomes. Therefore, there are ongoing efforts that aim to dissect clinical subtypes of schizophrenia in order to develop personalized treatment approaches. In agreement with these considerations, Carpenter et al.2 proposed to differentiate the deficit subtype of schizophrenia (SCZ-D) characterized primary and enduring negative symptoms. Subsequent studies have demonstrated that nearly one-third of individuals with schizophrenia might meet the criteria of SCZ-D.3 Compared to individuals with non-deficit schizophrenia (SCZ-ND), those with SCZ-D tend to show poor premorbid social functioning, worse cognitive functioning, greater neurostructural abnormalities as well as poor clinical and functional outcomes.4–6

To date, several factors have been found to increase a risk of schizophrenia. These include, i.e., familial liability,7 prenatal and perinatal complications,8 winter and spring season of birth,9 a history of childhood maltreatment10 and substance use (including nicotine).11–13 There are studies showing that some of them, including familial liability, obstetric complications and winter–spring season of birth might be more closely related to a risk of SCZ-D than that of SCZ-ND.6 Moreover, there is evidence supported by meta-analysis that male sex is associated with a risk of SCZ-D.14 Less is known about differential contribution of childhood adverse experiences and substance abuse to a risk of SCZ-D and SCZ-ND. On the basis of a meta-analysis, Alameda et al.15 showed that most forms of childhood abuse are associated with greater severity of symptoms across all dimensions of psychopathology in psychosis. However, a history of childhood neglect might be more specifically associated with negative and depressive symptoms. A history of childhood trauma has also been associated with a lack of positive and negative symptoms remission in patients with first-episode psychosis.16 As similar to a history of childhood trauma, none of previous studies investigated the association of substance use with a risk of SCZ-D and SCZ-ND. There is evidence from longitudinal studies that persistent cannabis use in patients with first-episode psychosis might be associated with greater severity of negative symptoms.17 Moreover, Krishnadas et al.18 found that individuals with schizophrenia and severe nicotine dependence show greater severity of negative symptoms, and thus require higher dosage of antipsychotics.

Following these observations, Alabaf et al.19 hypothesized that timing of known risk factors of schizophrenia might underlie their differential impact on the development of SCZ-D and SCZ-ND. The authors tested the association of known risk factors, including family history of psychosis, advanced paternal age, male gender, birth weight <3000g, summer birth, cannabis use, exposure to physical or sexual abuse and/or bullying in childhood as well as crime-related traumatic events with SCZ-D and SCZ-ND. They found that exposures acting in later life (cannabis use and a history of childhood physical or sexual abuse) are more closely related to a risk of SCZ-D than that of SCZ-ND. In this regard, we aimed to replicate these findings and test differential associations of SCZ-D and SCZ-ND with other risk factors of psychosis, including delivery by cesarean section, a history of other childhood adverse experiences (emotional abuse and neglect) and cigarette smoking at psychosis onset.

General characteristics of the sample are presented in Table 1. A total of 44 individuals (37.2%) met the criteria of SCZ-D. Individuals with SCZ-D were significantly more likely to be males. As expected, they had significantly higher severity of negative symptoms. Individuals with SCZ-D had also significantly lower scores of global cognition, social and occupational functioning as well as lower number of education years. The use of mood stabilizers was significantly more frequent in patients with SCZ-ND compared to those with SCZ-D. However, there were no significant between-group differences in the frequency of using antidepressants and the daily dosage of antipsychotics expressed as chlorpromazine equivalents. Moreover, no significant between-group differences were found with respect to age, age of psychosis onset, employment or education status at psychosis onset, current cigarette smoking status and the severity of depressive symptoms.

The distribution of specific risk factors is shown in Table 2. Apart from between-group sex differences, individuals with SCZ-D were significantly more likely to report birth weight <3000g and any childhood adversities. In turn, substance abuse at psychosis onset was significantly more frequent in participants with SCZ-ND. No significant between-group differences were observed with respect to the frequency of positive family history of schizophrenia in first-degree relatives, summer birth, delivery by cesarean section, specific childhood experiences and cigarette smoking before the onset of psychosis. Observed power was lower than 0.80 in the majority of comparisons (except of sex differences).

Results of binary logistic regression analysis with a diagnosis of SCZ-D as the dependent variable are shown in Table 3. We included a history of any childhood adversities, but not that of specific categories of childhood adversities, to avoid potential multicollinearity. This analysis confirmed that male sex and birth weight <3000g are associated with higher risk of SCZ-D, while substance abuse at psychosis onset is associated with higher risk of SCZ-ND (Cox & Snell R2=0.284, Nagelkerke R2=0.383).

Main findings of the present study indicate that male sex and birth weight <3000g are specifically associated with higher risk of SCZ-D, while substance abuse at psychosis onset might be more closely related to higher risk of SCZ-ND. These findings partially replicate those obtained by Alabaf et al.19 who showed that cannabis use at psychosis onset and a history of childhood physical or sexual abuse are associated with higher risk of SCZ-ND. Nevertheless, some discrepancies need to be discussed in light of methodological differences that might impact findings from smaller samples. It is important to note that the study by Alabaf et al.19 was based on patients with schizophrenia receiving clozapine, and thus it is likely that high proportion of patients showed treatment-resistant schizophrenia. There is some evidence that patients with treatment-resistant schizophrenia are more likely to report a history of childhood trauma compared to those who show sufficient treatment response31,32 Prevalence of childhood trauma in the sample reported by Alabaf et al.19 was 68.5% (41.0% in our sample). Similarly, there is some evidence that obstetric complications might be associated with worse response to treatment.33 Prevalence of birth weight <3000g, which might be associated with obstetric complications, also differed across both samples (28.0% in the study by Alabaf et al.19 and 14.8% in our sample). Moreover, the study by Alabaf et al.19 had greater predominance of male participants compared to our sample (78.7% vs. 61.0%).

Nevertheless, both studies indicate that early and late risk factors might be differentially associated with the development of SCZ-D and SCZ-ND and are in line with some previous studies. The predominance of males among individuals with SCZ-D has been documented by meta-analysis, also showing qualitatively stronger evidence for studies based on the use of SDS.14 This observation points to a number of sex differences reported in patients with schizophrenia. Indeed, male patients have been shown to present earlier age of psychosis onset, worse premorbid functioning and functional outcomes, greater neurostructural alterations, lower levels of depressive symptoms and greater severity of negative symptoms.34,35 To date, various biological mechanisms have been proposed to explain these observations, including protective effects of estrogens and genetic differences.35 Our findings also replicate those showing that individuals with SCZ-D have lower severity of current use of alcohol and other substances compared to those with SCZ-ND.36 Indeed, a higher severity of negative symptoms may make individuals with SCZ-D less likely to engage in risky behaviors and substance use.19 In turn, the association of low birth weight with a risk of schizophrenia has been well-documented.37–39 However, the association of low birth weight with SCZ-D and SCZ-ND has not been widely investigated.

There are certain limitations of this study that need to be considered. First, our sample size was relatively low and power calculations indicate the possibility of type II error. Second, we did not record the initial number of patients approached for participation and reasons of non-participation. Therefore, representativeness of the sample cannot be evaluated. Another point is that we used self-reports and retrospective interviews to record risk factors, and thus a recall bias should be taken into consideration. It is also important to note that we were not able to analyze the effects of specific substances being used before the onset of psychosis. Moreover, certain factors, e.g., birth weight <3000g or delivery by cesarean section do not reflect prenatal and perinatal complications associated with a risk of psychosis. Indeed, only complicated or emergency birth by cesarean section can be perceived as obstetric complication increasing a risk of psychosis.8,40 Previous studies have shown that birth by cesarean section itself is associated with earlier age of psychosis onset but not a risk of psychosis.8,41 Moreover, we investigated birth weight <3000g which does not correspond to the definition of low birth weight (<2500g) according to the World Health Organization. However, both cut-offs (<2500g and <3000g) have been associated with psychosis risk.37 It cannot also be excluded that adopting other definitions of certain risk factors, e.g., family history of schizophrenia and summer birth would yield different results. Finally, it should be highlighted that a cross-sectional design of the present study does not allow to make causal conclusions.

In summary, the present findings indicate that substance abuse at psychosis onset might be more closely related to a risk of SCZ-ND, while male sex and birth weight <3000g increase a risk of SCZ-D. However, data obtained in the present study do not allow to conclude which substances of abuse are related to a risk of SCZ-D. Our findings point to the hypothesis that clinical manifestation and outcome of schizophrenia might be related to timing of exposure to risk factors. However, larger studies based on longitudinal observations as well as more comprehensive recording of risk factors and clinical manifestation are needed to confirm our results.

This study received funding from the OPUS grant awarded by National Science Centre, Poland (grant number: 2018/31/B/NZ5/00527).

AC, recruitment and clinical assessment of patients, manuscript writing; PP, recruitment and clinical assessment of patients; JS, study design and manuscript writing; TG, recruitment and clinical assessment of patients; BM, study design, data analysis, manuscript writing and editing

None declared.