Understanding sex differences in long-term outcomes after a first episode of psychosis

The study investigates whether long-reported sex differences in schizophrenia outcomes persist in the long term among first-episode psychosis (FEP) patients after receiving Early Intervention Services (EIS). Prior research suggests women often have later illness onset and initially better outcomes, but these advantages may diminish over time. The authors aim to examine long-term sex differences (clinical, functional, cognitive, treatment) after discharge from EIS to routine care and test the hypothesis that, about 10 years after an FEP, outcomes for women and men would be very similar.

Existing studies on sex differences in FEP outcomes show mixed results. Short-term EIS studies (e.g., Hong Kong cohort) observed better functional outcomes in women during the first 3 years, while others (e.g., Canadian EIS) attribute differences to premorbid adjustment and age at onset. Longitudinal cohorts like OPUS and AESOP-10 indicate outcomes can continue to improve between 5 and 10 years and that men often have worse clinical, social, and service-use outcomes. Some literature suggests women’s early advantages may decline over time, underscoring the need to specifically assess long-term sex differences after EIS discharge.

Design and setting: Prospective cohort leveraging the PAFIP (Programa de Atención a Fases Iniciales de Psicosis) EIS in Cantabria, Spain, with 3-year specialized EIS and long-term follow-up (PAFIP-10). Baseline recruitment of FEP patients occurred between 2001–2008; 10-year reassessments were conducted between June 2014–August 2018 (8–16 years post-enrollment). Ethical approval obtained; informed consent collected. Participants: Inclusion criteria: age 15–60; first episode of psychosis; resident in catchment area; no prior adequate antipsychotic treatment or <6 weeks lifetime; DSM-IV diagnoses of brief psychotic disorder, schizophreniform disorder, schizophrenia, or schizoaffective disorder. Exclusions: DSM-IV drug/alcohol dependence, intellectual disability, neurological disease or head injury. Sample and follow-up: 307 FEP patients assessed at baseline (128 female, 179 male); 297 reassessed at 1 year; 259 at 3 years; 209 reassessed at 8–16 years (10-year cohort; 183 in-person, 26 by phone). A subset of 151 had full clinical/functional/cognitive/treatment data across time; medication analyses included 91 patients. Measures: Baseline sociodemographic and clinical data (including age at onset, DUR of untreated illness and psychosis, premorbid social adjustment, premorbid IQ). Diagnoses via SCID-I at baseline (confirmed/updated at 10 years). Symptom scales: SANS (negative), SAPS (positive), BPRS (general), CDSS (depression). Functioning: Disability Assessment Scale (DAS). Cognition: global cognitive function (GCF) composite from seven cognitive domains (converted to T-scores and deficit scores). Recovery defined as symptomatic remission (SAPS/SANS items ≤2) and adequate psychosocial function (DAS ≤1) at 1, 3, and 10 years. Relapses recorded across 10 years. Antipsychotic treatment recorded as mean daily chlorpromazine equivalents at baseline, 3 months, 6 months, 1 year, 2 years, 3 years, and 10 years. Extrapyramidal symptoms assessed by Simpson–Angus Scale at baseline, 6 weeks, 1 year, 3 years, and 10 years. Statistical analysis: Group differences with t tests and chi-square/Fisher’s exact tests. Longitudinal outcomes tested with repeated-measures ANCOVA (age at first contact as covariate), including within- and between-group-by-time effects with Bonferroni correction. Kaplan–Meier survival analysis and log-rank tests for relapse timing. Significance set at 5% (two-tailed).

- Cohort and follow-up: Of 307 baseline FEP patients, 209 (95 females, 114 males) were reassessed at 8–16 years. Non-completers were more likely to have lower education (61% vs 44% elementary, p=0.005), urban residence (82% vs 68%, p=0.011), unemployment (54% vs 38%, p=0.012), and cannabis use (56% vs 36%, p=0.001); completers had more severe baseline depressive symptoms (F=2.114; p=0.035). Trend toward higher 10-year participation in women and more exitus in men. - Baseline sex differences: Females were older at admission/onset (31 vs 27 years; p<0.001), had better premorbid functioning (t=4.375; p<0.001), higher premorbid IQ (97.7 vs 93.1; p=0.043), higher education, were more often living independently (58% vs 40%; p=0.008), employed (72% vs 53%; p=0.005), partnered (23% vs 13%; p=0.059), and had children (40% vs 14%; p<0.001). Men had higher rates of schizophrenia diagnosis (69.3% vs 51.6%; p=0.009), cannabis (50.9% vs 22.1%; p<0.001) and alcohol use (66.7% vs 33.7%; p<0.001). - Ten-year characteristics: Women more often lived independently (70% vs 45%; p<0.001) and had partners (38% vs 26%; p=0.057). Men more often had schizophrenia diagnosis (84.0% vs 63.9%; p=0.002), tobacco use (61.4% vs 45.3%; p=0.020), and cannabis use (12.3% vs 4.2%; p=0.038). - Symptoms over time: Positive symptoms improved similarly in both sexes (within-group by time F=21.883; p<0.001; no between-group-by-time effect). Negative symptoms improved more in females (within-group by time F=8.135; p<0.001; between-group-by-time F=9.33; p=0.003), with significant timepoint differences at 6 weeks (F=13.15; p<0.001), 3 years (F=6.131; p=0.014), and 10 years (F=5.983; p=0.016). Depressive symptoms improved over time (F=4.665; p=0.032) without sex-by-time differences. - Function and cognition: No significant within- or between-group-by-time effects for DAS or GCF; only years of education showed a significant group-by-time effect for GCF (F=13.33; p=0.001). - Relapses: No sex difference in time to relapse; most relapses occurred in first 3 years (log-rank p=0.792). - Recovery: Women had higher recovery at 1 year (36.6% vs 22.3%; χ²=4.049; p=0.044) and 3 years (50% vs 30.8%; χ²=6.234; p=0.013); at 10 years, recovery rates were not significantly different (46.7% vs 34.4%; χ²=2.656; p=0.103). - Antipsychotic treatment: Females received lower chlorpromazine-equivalent doses at several timepoints: 6 months (F=4.27; p=0.042), 1 year (F=6.838; p=0.01), 2 years (F=11.533; p=0.001), 3 years (F=14.841; p<0.001). Minimal effective doses were reached at 3 years, followed by an increase between 3 and 10 years for both sexes (within-group-by-time F=5.557; p=0.021; age-by-time F=6.46; p=0.013; between-group-by-time F=12.184; p=0.001). No sex differences in extrapyramidal symptoms.

Early advantages in women (later onset, better premorbid functioning, higher education/employment/independent living, lower substance use) were associated with better short-term outcomes during EIS, including higher recovery and response to lower antipsychotic doses. Over approximately 10 years, outcomes between sexes converged—with the notable exception that women sustained a better course and lower severity of negative symptoms. Relapses clustered in the first 3 years with no sex difference thereafter. Importantly, antipsychotic doses increased after discharge from EIS to community services for both sexes, potentially reducing earlier gains, especially for women who may require lower effective doses and be more vulnerable to side effects. The findings suggest that sex-specific factors (biological, psychosocial) and service delivery (duration and dosing practices post-EIS) influence long-term trajectories, raising the possibility that extending EIS and tailoring treatment by sex could improve sustained outcomes.

Women with FEP demonstrated better premorbid profiles and superior short-term outcomes during EIS, including higher recovery rates and lower required antipsychotic doses. By 10 years, most outcomes converged between sexes except for persistently better negative symptom trajectories in women, while antipsychotic dosing increased post-EIS in both groups. The study highlights the need to consider sex-specific treatment strategies and potentially extend EIS duration to sustain gains. Future research should disentangle biological, psychosocial, and environmental contributors to sex differences and optimize functional outcome measures and dosing strategies across long-term care transitions.

- Longitudinal reassessment occurred within an 8–16 year window, introducing variability in follow-up duration. - Potential selection bias: non-completers differed in education, urbanicity, employment, and cannabis use; trend for more exitus among men. - Did not assess biological (e.g., estrogen levels), psychosocial (e.g., resilience), or environmental contributors to sex differences. - Functional assessment via DAS may not fully capture real-world functioning and is criticized as outdated. - Unrecorded factors during community care could have influenced outcomes (e.g., adherence, psychosocial interventions, comorbidities). - Generalizability limited to a single-region Spanish EIS cohort. - Some analyses used subsets with complete longitudinal data, potentially reducing power/representativeness.