Short-, Medium-, and Long-Term Cardiometabolic Outcomes in First-Episode Psychosis: A Systematic Review and Meta-analysis

Abstract

Background

Individuals with first-episode psychosis (FEP) experience increased cardiometabolic risks, contributing to reduced life expectancy. While metabolic disturbances are well described in chronic schizophrenia, their trajectory from first presentation, before or with minimal antipsychotic exposure, through to long-term follow-up remains incompletely characterized.

Objective

To synthesize evidence on progression of cardiometabolic outcomes in FEP cohorts from before initiation of antipsychotic medication to up to 10 years of follow-up.

Study Design

Following PRISMA and MOOSE guidelines (PROSPERO: CRD42023431072), Medline, Embase, PsycInfo, and CINAHL+ were searched to July 2025 for studies of FEP cohorts with ≤28 days of antipsychotic exposure at baseline. Outcomes included weight, glucose, lipids, blood pressure, metabolic syndrome (MetS), diabetes, and cardiovascular disease (CVD). Random-effects meta-analyses were stratified by follow-up duration. As post-baseline exposure was not quantified, follow-up estimates reflect outcomes under routine care.

Study Results

From 2601 unique articles, 82 studies were included. Obesity and MetS prevalence increased from 5.4% and 8.5%, respectively at baseline, to 26.6% and 18.2% at 1–3 years, whereas type 2 diabetes increased from 0.5% to 5.0% by >5 years. Lipid abnormalities emerged early, while blood pressure changes were minimal, and CVD hospitalization prevalence was 1.4% at 5–10 years. Weight increased by +2.5 kg at 4–8 weeks and + 13.4 kg at >5 years, with parallel increases in BMI and waist circumference.

Conclusions

Cardiometabolic deterioration begins within weeks of FEP onset and accumulates over time, indicating the need for early prevention, sustained monitoring, integrated metabolic care, and patient education throughout the course of care.

Introduction

Individuals with psychosis have an increased risk of several physical health conditions, including cardiovascular diseases, respiratory disorders, metabolic disorders, and cancer.^1,2 These multimorbidities, especially cardiometabolic diseases, contribute to a significantly reduced life expectancy of approximately 10–20 years compared to the general population.^3,4

The increased physical multimorbidity in first-episode psychosis (FEP) is linked to a complex interplay of individual, population, and healthcare system-level factors.⁵ Ethnic minority status and social deprivation, both independently associated with an increased risk of developing psychosis and cardiometabolic disease in the general population, likely contribute to worse physical health outcomes, including increased multimorbidity and cardiovascular mortality in people with psychosis.^6,7 Recent research also suggests shared biological vulnerability, with genetic risk for schizophrenia associated with cardiac structural changes that can worsen cardiac outcomes,⁸ and lipidomic and proteomic changes already observed at age 11 among those who later develop psychotic disorders.^9,10 Previous research shows that two-thirds of individuals with FEP will experience metabolic changes, such as impaired glucose and lipid metabolism, as well as weight gain exceeding 7% during the initial 12 months of treatment.^11,12 Moreover, metabolic syndrome (MeTS), encompassing dyslipidemia, hypertension, impaired glucose regulation, and central obesity, is prevalent in up to 69% of those with chronic schizophrenia.¹³ Although antipsychotic treatment contributes to this observed weight gain and dyslipidaemia,^1,14 altered glucose homeostasis is already observed in antipsychotic-naïve individuals with FEP, suggesting that metabolic disturbances are present from the onset of psychosis.^15–17 In addition, people with psychosis frequently present with pre-existing behavioral and social risk factors, including high rates of smoking, high-fat and low-fiber diets, low physical activity, and obesity and dyslipidemia already at baseline, often in the context of low socioeconomic status and limited access to preventive care.^3,18–21 Despite widespread recognition of these risk factors, national audits and service evaluations consistently show under-assessment and inadequate management of physical health problems in schizophrenia and related disorders.²²

Metabolic changes are clinically significant not only because they contribute to long-term morbidity and mortality, but also because they manifest in day-to-day symptoms, such as shortness of breath after minimal exertion, back pain when standing, or profuse sweating, that can be stigmatizing and limit meaningful social participation. Rapid antipsychotic-associated weight gain in psychosis is experienced with shock, loss of self-worth and control, social withdrawal, and difficulties engaging in weight loss, with people reporting that they did not feel prepared for these changes or supported to manage them.²³ Evidence from guideline-level reviews indicates that multicomponent lifestyle interventions combining behavioral change techniques, dietary modification, and supported physical activity can produce small to moderate improvements in weight, glucose regulation, and lipid profiles, and may also benefit symptoms and sleep in people experiencing severe mental health challenges.²⁴ Recognizing these lived experiences is essential for understanding the full impact of cardiometabolic comorbidity in FEP and for framing recovery in holistic terms.

While prior systematic reviews have reported an increased prevalence of physical health disorders among people with psychosis, these reviews typically pool data from chronic psychosis and FEP cohorts or focus solely on baseline outcomes, limiting insights into the longitudinal progression of physical health outcomes in FEP.^17,25,26 This review addresses these gaps by focusing exclusively on FEP cohorts with longitudinal follow-up data, synthesizing evidence from both cohort studies and antipsychotic intervention studies, which provide detailed baseline and follow-up physical health monitoring. This approach captures the progression of cardiometabolic comorbidities from psychosis onset across short-, medium-, and long-term periods, offering novel insights into the trajectory of metabolic changes in. The primary objective of this review is to systematically examine the existing evidence relating to the short-, medium-, and long-term period prevalence of cardiometabolic outcomes in individuals with FEP.

Methods

Search Strategy

This systematic review and meta-analysis, registered in PROSPERO (CRD42023431072; 17/06/2023), adhered to PRISMA 2020 and MOOSE guidelines.^27,28 We searched Medline, Embase, PsycInfo, and CINAHL+ from inception to July 19, 2025. Although the protocol included grey literature, we prioritized peer-reviewed sources due to resource constraints. The protocol originally included cardiovascular, metabolic, cancer, and respiratory outcomes; owing to volume and heterogeneity, only cardiovascular and metabolic outcomes are reported here, with the remaining outcomes presented elsewhere. A librarian-assisted search used Medical Subject Headings and text words for first-episode psychosis, metabolic, and cardiovascular outcomes. English-language search terms were used, capturing non-English studies with English indexing, but not studies lacking English indexing. Full search details are in the Supplement.

Eligibility Criteria

Studies were included if they met the following criteria: (1) involved individuals with clinician-diagnosed FEP (eg, schizophrenia, schizoaffective disorder), defined according to the Diagnostic and Statistical Manual of Mental Disorders or International Classification of Diseases (ICD) criteria or diagnosis recorded in medical records, at first treatment contact regardless of untreated illness duration; (2) participants with no more than 28 days of antipsychotic exposure at the minimal effective dose, according to The Maudsley Prescribing Guidelines in Psychiatry, 11th edition (table 2.2, pp. 14–15).²⁹ This criterion was applied only at baseline to ensure inclusion at first treatment contact, and sensitivity analyses compared strictly antipsychotic-naïve and mixed-exposure cohorts; (3) reported prevalence or incidence of metabolic or cardiovascular outcomes at specified follow-up time points post-diagnosis, confirmed by clinicians, medical records, or ICD-10/11 criteria. No restrictions were applied for age, publication period, or baseline cardiometabolic status, substance use, or smoking.

Studies were excluded if they: (1) focused solely on organic psychotic disorders; (2) included individuals with multiple episodes of psychosis; (3) measured only baseline physical health outcomes or none at all; (4) were single case studies or case series with fewer than 10 participants. Studies with mixed populations were included if FEP-specific data were extractable.

Outcome Definitions

Follow-up windows were harmonized after study selection based on clustering of reported timepoints and prior reviews: early >4–14 weeks, medium-term 15–52 weeks (≈3–12 months), long-term 1–3 years (53–156 weeks), and very long-term >5 years (≥260 weeks). Continuous outcomes were further subdivided at 4–8 and 9–14 weeks, while proportion analyses used the combined >4–14-week window because of sparse early data. When multiple assessments occurred within a window, the latest was retained.

The primary outcome was period prevalence (or incidence) of cardiometabolic conditions (MeTS, obesity, diabetes, dyslipidemia, hypertension, cardiovascular disease) at predefined windows, consistent with the registered protocol. Secondary outcomes were mean changes in continuous metabolic measures (weight, BMI, waist circumference, glucose, lipids, blood pressure). Across included studies, MetS was defined using four main guidelines. The National Cholesterol Education Program (NCEP) Adult Treatment Panel III (ATP III) definition required three or more of five factors: elevated waist circumference, triglycerides ≥150 mg/dL, low HDL cholesterol (<40 mg/dL in men, <50 mg/dL in women), blood pressure ≥ 130/85 mmHg or treatment, and fasting glucose ≥110 mg/dL.³⁰ The American Heart Association/National Heart, Lung, and Blood Institute ATP III update used the same components but with a lower threshold for fasting glucose (≥100 mg/dL). The International Diabetes Federation (IDF, 2005) criteria required central obesity (using sex- and ethnicity-specific waist cut-offs) plus at least two of the remaining factors, including fasting glucose ≥100 mg/dL or diabetes.³¹ Finally, the ATP III–Asian modification applied the ATP III framework but substituted Asian-specific waist circumference cut-offs.³²

Standard clinical thresholds were applied for obesity (Body Mass Index [BMI] ≥30 kg/m²), ≥7% weight gain, prediabetes (fasting glucose 100–125 mg/dL), diabetes (fasting glucose ≥126 mg/dL or clinical diagnosis), hypertension (≥130/85 mmHg or treatment), dyslipidemia, Homeostatic Model Assessment of Insulin Resistance (HOMA-IR), and QTc prolongation using study-specified cut-offs, because only proportions above that cut-off were available, we could not reapply a higher consensus threshold.³³ HOMA-IR index was reported with thresholds of >3.5 for men and > 3.9 for women in the included studies. QTc prolongation, reported in one cardiovascular study, was defined as a QTc interval > 426.2 ms in men or > 433.4 ms in women based on the Framingham correction. Low high-density lipoprotein (HDL) cholesterol was typically defined as <40 mg/dL in men and < 50 mg/dL in women; elevated low-density lipoprotein (LDL) cholesterol as ≥130 mg/dL (or > 175 mg/dL in one study); hypertriglyceridemia as triglycerides ≥150 mg/dL; and elevated total cholesterol as >240 mg/dL.

Data Extraction

Two authors (AZ and JM) deduplicated studies in EndNote and screened them in Rayyan through title/abstract and full-text rounds, with discrepancies resolved by a third reviewer. Overlapping cohorts were identified by examining recruitment sites and timeframes, selecting the most comprehensive dataset. Initial data extraction commenced on 13/11/2023 by AZ and AJ and was completed in December 2024. Following the rerun of the database search in July 2025, AZ and JM conducted an updated round of data extraction (July–August 2025) to ensure inclusion of the most recent evidence. This was done using a pre-piloted Excel form covering study design, participant characteristics, outcomes, and confounders such as smoking or substance use. Antipsychotic exposure beyond baseline (dose, duration, switching, adherence) and concomitant psychotropics were extracted where reported; because reporting was inconsistent, follow-up estimates reflect usual-care exposure patterns rather than medication-free trajectories.

Risk of Bias Assessment

Risk of bias was assessed using the Joanna Briggs Institute (JBI) Critical Appraisal Checklist for Studies Reporting Prevalence Data.³⁴ AZ and JM independently evaluated nine criteria, with discrepancies resolved by a third reviewer. The JBI Critical Appraisal Checklist for Studies Reporting Prevalence Data focuses on sampling frame, representativeness, outcome measurement, and statistical analysis, which are relevant to both prevalence estimates and within-cohort changes in continuous outcomes derived from the same samples. We therefore applied the same study-level risk of bias ratings when interpreting both prevalence and mean-change results, rather than using separate tools for different outcome types. According to the JBI criteria, 14 studies were rated “Include with caution” and 3 “Exclude” due to limitations such as small sample sizes and large attrition at follow-up (See Supplement). Primary analyses included all eligible studies, while sensitivity analyses excluded both studies marked as “Include with caution” and “Exclude,” retaining only those rated ‘Include. Results from the JBI-restricted analyses were broadly consistent with the main findings and are presented alongside them (see Supplement).

Data Analysis

Analyses were conducted in R (v4.2.3) using metafor (v2.1-0) and the meta package. Random-effects models estimated proportions with 95% CIs for outcomes with k ≥ 2, grouped by follow-up window and outcome type. Proportion meta-analyses used logit-transformed proportions (sm = “PLOGIT”) with a 0.5 correction for zero-event studies; Freeman–Tukey transformation was used only when logit models failed to converge. Results were back-transformed to the proportion scale for interpretation. Heterogeneity was summarized with τ² (heterogeneity variance), I² (percentage of heterogeneity), H², the Q statistic (for overall heterogeneity), and its associated P-value. To evaluate robustness, we re-estimated models using the Paule–Mandel estimator for τ² with Hartung–Knapp CIs and performed leave-one-out influence analyses (refitting the random-effects model while omitting each study in turn). We also repeated the meta-analyses after restricting to studies with an overall JBI decision of “Include,” applying the study-level decision to all subgroups from the same study. Small-study effects were explored with funnel plots and Egger’s test, which was interpreted only when k ≥ 10. Post hoc subgroup analyses examined antipsychotic-naïve status and study design (cohort vs trial/RCT). For outcomes supported by very few studies (typically k ≤ 3), we did not conduct sensitivity analyses, and these pooled estimates should be interpreted with caution. Full subgroup results are presented in the Supplement.

Lived Experience Involvement

A lived experience expert (MJN) contributed to all stages of the review, including protocol development, interpretation of findings, and preparation of the lay summary. This ensured that the review was informed by lived experience perspectives alongside clinical and academic expertise.

Results

Overview of Meta-Analysis

A total of 4970 articles were identified and, after removal of duplicates, 2601 unique articles remained. After screening of title and abstract, 2225 studies were excluded, leaving 376 for full-text review. Of these full texts, 82 were deemed eligible, involving 11 to 80 581 individuals with FEP and 77 unique samples, as some studies reported different health outcomes or follow-up waves for the same sample (see Figure 1). A small number of non-English full texts were retrieved at screening, but none met the inclusion criteria. Out of the included studies, 74 reported on metabolic outcomes, and 8 reported on cardiovascular outcomes (see Supplement). Details of sample characteristics of the included studies are reported in Tables S1 and S2 in the Supplementary Material.

Figure 1.

Prisma Flow Diagram of the Study Selection Process

Study Characteristics

Most studies used second-generation antipsychotics exclusively (n = 41), followed by studies that involved a mixture of first- and second-generation agents (n = 28). Only a small number used first-generation antipsychotics alone (n = 3), and in five studies, the antipsychotic class was not clearly reported. The included studies primarily involved young adults with FEP or first-episode schizophrenia spectrum disorders, with mean ages ranging from 15.9 to 42.3 years (most cohorts in their 20s to early 30s). The studies, conducted between 2002 and 2025, included diverse ethnic groups such as Chinese (n = 21), Caucasian (n = 26), mixed or multi-ethnic samples (n = 13), Indian (n = 5), Spanish (n = 5), and Korean (n = 4). Israeli samples appeared in two studies, while Māori, Taiwanese, Japanese, and Nigerian groups were each represented in one study. Eleven studies did not report ethnicity. The proportion of female participants varied widely (0–100%, median ~ 40%), and study designs included cohorts (n = 46), randomized controlled trials (RCTs) (n = 24), trials (n = 10), and case–control studies (n = 2), with follow-up periods ranging from 2 weeks to 15 years. The majority of samples was completely antipsychotic-naïve at baseline (n = 54), 32 reported pharmaceutical funding, and antipsychotic switching procedures differed across designs (allowed or occurred, n = 33; prohibited, n = 12; not reported, n = 32). A majority of studies also allowed concomitant psychotropic medication use (n = 57), while few explicitly prohibited all concomitant medication (n = 4) or did not report on concomitant prescribing (n = 16). Across studies, the most commonly permitted medication classes were benzodiazepines (n = 42), anticholinergics (n = 39), antidepressants (n = 33), mood stabilizers (n = 22), hypnotics or sleep medications (n = 21), beta-blockers (n = 18), and antiepileptics (n = 6).

Baseline physical-health eligibility criteria varied: 23 studies explicitly excluded participants with cardiometabolic conditions at baseline (for example, diabetes, metabolic syndrome, obesity, hypertension, or dyslipidaemia), 19 applied only generic “serious/unstable medical illness” or “physical illness” exclusions without specifying cardiometabolic abnormalities, 19 did not restrict inclusion based on cardiometabolic status, and in the rest of the included studies the use of cardiometabolic exclusion criteria was not reported (see Supplementary Material).

Between-study heterogeneity was generally high across continuous outcomes (typical I² 80–95%, rising at longer follow-up) and varied for prevalence outcomes, being low to moderate at baseline and early follow-up but often very high at later follow-up. Sensitivity analyses based on AP-naïve status, study design, and JBI-restricted subsets showed similar overall directions across metabolic outcomes, with leave-one-out effects generally small. AP-naïve subgroups and cohort studies tended to report larger mean changes than mixed-exposure samples and RCTs, although confidence intervals often overlapped. Full forest plots for additional outcomes and follow-up windows, as well as sensitivity analyses not shown in the main text, are provided in the Supplementary material.

Anthropometric Outcomes

The proportion of individuals with FEP with obesity (BMI ≥30 kg/m²) was 5.4% (95% CI, 4.0–7.3; k = 3, I² = 0%) at baseline and increased to 26.6% (95% CI, 22.7–31.0; k = 3, I² = 35%) at 1–3 years (see Figure 2). AP-naïve studies showed similar estimates at baseline (5.5%; k = 2) and at 1–3 years (27.2%; k = 2). Sensitivity analyses showed minimal change in pooled estimates when restricted to AP-naïve, JBI-included, or cohort-only studies.

Figure 2.

Forest Plots of Anthropometric Outcomes at 1–3 Years After First Episode Psychosis

The proportion with weight gain ≥7% was 37.6% (95% CI, 28.1–48.1; k = 7, I² = 81%) at 4–14 weeks and increased to 59.9% (95% CI, 21.3–89.2; k = 2, I² = 99%) at 1–3 years. AP-naïve samples showed a substantially higher short-term prevalence (56.5%; k = 3) than minimally treated (25.2%; k = 2) or trial-only subsets (≈24–40%; k = 2–3). Study-design analyses indicated that cohort studies consistently reported higher prevalence than RCTs at equivalent timepoints.

The proportion of individuals with high waist circumference (men ≥102/90 cm, women ≥88/80 cm) was 15.5% (95% CI, 10.9–21.7; k = 2, I² = 39%) at baseline and 33.2% (95% CI, 22.9–45.3; k = 2, I² = 75%) at 1–3 years (see Table 1). Study-design analyses indicated that cohort studies consistently reported higher prevalence than RCTs at equivalent timepoints (see Supplement).

Table 1.

Summary of Cardiometabolic Proportion Outcomes in FEP.

Outcome (Criteria)	Baseline	4–14 weeks	3–12 months	1–3 years	5–10 years
BMI ≥30	5.4% (CI95%, 4.0–7.3, k = 3, I2 = 0%)	−	15.3% (CI95%, 10.4–22.0, k = 1)	26.6% (CI95%, 22.7–31.0, k = 3, I2 = 35%)	36.8% (CI95%, 30.4–43.7, k = 1)
Weight gain ≥7%	−	37.6% (CI95%, 28.1–48.1, k = 7, I2 = 81%)	53.6% (CI95%, 45.6–61.4, k = 5, I2 = 39%)	59.9% (CI95%, 21.3–89.2, k = 2, I2 = 99%)	−
High waist circumference (Men ≥102/90 cm, Women ≥88/80 cm)	15.5% (CI95%, 10.9–21.7, k = 2, I2 = 39%)	−	−	33.2% (CI95%, 22.9–45.3, k = 2, I2 = 75%)	−
Diabetes (FPG ≥126 mg/dL or clinical diagnosis)	0.5% (CI95%, 0.2–1.3, k = 3, I2 = 0%)	0.3% (CI95%, 0.0–2.2, k = 2, I2 = 0%)	−	1.4% (CI95%, 0.4–5.6, k = 1)	5.0% (CI95%, 3.4–7.3, k = 5, I2 = 89%)
Prediabetes (FPG ≥100 mg/dL)	4.8% (CI95%: 3.4–6.8, k = 6, I2 = 55%)	0.6% (CI95%: 0.1–2.8, k = 2, I2 = 0%)	3.3% (CI95%: 1.4–7.8, k = 1)	9.4% (CI95%: 4.0–20.3, k = 4, I2 = 92%)	7.1% (CI95%: 4.3–11.7, k = 1)
Total cholesterol >240 mg/dL	12.6% (CI95%: 3.5–36.3, k = 2, I2 = 93%)	16.7% (CI95%: 6.4–36.9, k = 1)	39.6% (CI95%: 32.3–47.4, k = 1)	19.1% (CI95%: 3.2–62.3, k = 2, I2 = 97%)	41.6% (CI95%: 34.9–48.6, k = 1)
LDL cholesterol >175 mg/dL	11.5% (CI95%: 2.7–37.9, k = 2, I2 = 92%)	12.5% (CI95%: 4.1–32.4, k = 1)	31.6% (CI95%: 24.8–39.3, k = 1)	15.1% (CI95%: 1.9–61.5, k = 2, I2 = 96%)	35.2% (CI95%: 28.3–42.7, k = 1)
Low HDL cholesterol (Men <35, Women <50 mg/dL)	23.4% (CI95%: 13.6–37.3, k = 5, I2 = 95%)	25.0% (CI95%: 11.7–45.6, k = 1)	14.8% (CI95%: 10.1–21.3, k = 1)	33.1% (CI95%: 18.1–52.7, k = 4, I2 = 95%)	26.2% (CI95%: 20.1–33.4, k = 1)
Triglycerides >150 mg/dL	6.2% (CI95%: 4.7–8.1, k = 4, I2 = 0%)	37.5% (CI95%: 20.8–57.8, k = 1)	14.2% (CI95%: 9.5–20.6, k = 1)	21.3% (CI95%: 17.9–25.0, k = 4, I2 = 24%)	24.4% (CI95%: 18.5–31.5, k = 1)
TG/HDL ratio > 3.5	2.2% (CI95%, 0.1–26.1, k = 2, I2 = 93%)	−	14.0% (CI95%, 9.1–20.9, k = 1)	21.2% (CI95%, 17.0–26.1, k = 1)	−
HOMA index (Men >3.5, Women >3.9)	11.5% (CI95%, 8.3–15.7, k = 1)	−	16.7% (CI95%, 12.1–22.5, k = 2, I2 = 0%)	22.0% (CI95%, 17.5–27.1, k = 1)	−
Insulin (Men >15.7, Women >17.3 μU/mL)	12.1% (CI95%, 8.9–16.3, k = 1)	−	15.8% (CI95%, 10.5–23.0, k = 2, I2 = 0%)	18.2% (CI95%, 14.2–23.0, k = 1)	−
Hypertension ≥130/85 mmHg	26.7% (CI95%, 23.7–29.9, k = 3, I2 = 0%)	−	−	25.2% (CI95%, 16.7–36.2, k = 2, I2 = 74%)	−
Metabolic Syndrome	8.2% (CI95%, 5.0–12.9, k = 8, I2 = 77.8%)	13.4% (CI95%, 3.9–36.8, k = 3, I2 = 89.2%)	14.7% (CI95%, 4.5–38.6, k = 2, I2 = 79.5%)	18.0% (CI95%, 12.8–24.8, k = 3, I2 = 68.0%)	36.2% (CI95%, 29.4–43.6, k = 1)
CVD hospitalization (IHD, MI, Stroke)	−	−	−	−	1.3% (CI95%, 1.0–1.9, k = 3, I2 = 93%)

Abbreviations: CVD, cardiovascular disease; HDL, high-density lipoprotein; LDL, low density lipoprotein

Meta-analyses showed consistent increases in body weight, BMI, and waist circumference from baseline across all follow-up periods. Mean body weight increased by 2.47 kg (95% CI, 1.77-3.16; k = 22) at 4–8 weeks, 9.05 kg (95% CI, 7.22-10.88; k = 8) at 1–3 years, and 13.43 kg (95% CI, 9.56-17.30; k = 2) at >5 years. BMI increased by 0.93 kg/m² (95% CI, 0.64-1.22; k = 25) at 4–8 weeks, 3.31 kg/m² (95% CI, 2.82-3.80; k = 10) at 1–3 years, and 4.78 kg/m² (95% CI, 3.79-5.76; k = 3) at >5 years. Waist circumference increased by 3.37 cm (95% CI, 1.53-5.21; k = 7) at 4–8 weeks, 6.28 cm (95% CI, 2.03-10.52; k = 4) at 3 months to 1 year, and 9.88 cm (95% CI, 3.60-16.17; k = 3) at 1–3 years (see Table 2).

Table 2.

Summary of Mean Continuous Metabolic Changes in FEP.

Outcome	4–8 weeks	9–14 weeks	3 months – 1 year	1–3 years	>5 years
Body weight (kg)	2.47 (1.77–3.16, k = 22, I2 = 97.1%)	4.38 (3.56–5.21, k = 15, I2 = 89.4%)	6.02 (4.46–7.58, k = 12, I2 = 95.1%)	9.05 (7.22–10.88, k = 8, I2 = 91.9%)	13.43 (9.56–17.30, k = 2, I2 = 89.2%)
BMI (kg/m2)	0.93 (0.64–1.22, k = 25, I2 = 97.3%)	1.84 (1.55–2.13, k = 15, I2 = 92.7%)	2.84 (2.33–3.34, k = 17, I2 = 96.1%)	3.31 (2.82–3.80, k = 10, I2 = 85.2%)	4.78 (3.79–5.76, k = 3, I2 = 73.1%)
Waist circumference (cm)	3.37 (1.53–5.21, k = 7, I2 = 78.1%)	3.64 (2.17–5.11, k = 4, I2 = 30.1%)	6.28 (2.03–10.52, k = 4, I2 = 88.9%)	9.88 (3.60–16.17, k = 3, I2 = 98.8%)	13.40 (10.77–16.03, k = 1)
Total cholesterol (mg/dL)	12.59 (7.24–17.95, k = 12, I2 = 95.1%)	9.29 (2.90–15.68, k = 10, I2 = 79.6%)	15.11 (4.79–25.44, k = 7, I2 = 89.4%)	8.91 (4.23–13.60, k = 4, I2 = 71.8%)	22.70 (18.63–26.77, k = 1)
LDL cholesterol (mg/dL)	9.34 (5.37–13.30, k = 11, I2 = 87.4%)	10.00 (7.28–12.72, k = 9, I2 = 57.4%)	9.26 (1.11–17.42, k = 5, I2 = 87.0%)	7.00 (4.18–9.81, k = 5, I2 = 49.3%)	14.00 (10.43–17.57, k = 1)
HDL cholesterol (mg/dL)	−0.75 (−2.89–1.38, k = 15, I2 = 98.6%)	−2.00 (−3.48–−0.52, k = 9, I2 = 96.3%)	−2.48 (−4.72–−0.24, k = 7, I2 = 91.1%)	−3.62 (−5.79–−1.46, k = 7, I2 = 77.9%)	−21.50 (−22.97–−20.03, k = 1)
Triglycerides (mg/dL)	32.91 (19.03–46.78, k = 16, I2 = 93.9%)	18.74 (9.73–27.75, k = 12, I2 = 87.7%)	23.20 (7.30–39.10, k = 9, I2 = 89.8%)	28.53 (25.37–31.69, k = 7, I2 = 0%)	45.30 (36.92–53.68, k = 1)
Fasting glucose (mg/dL)	2.03 (0.66–3.39, k = 19, I2 = 98.0%)	1.65 (1.16–2.13, k = 13, I2 = 88.9%)	3.95 (1.34–6.57, k = 7, I2 = 85.6%)	3.20 (1.57–4.84, k = 7, I2 = 71.5%)	7.54 (1.17–13.91, k = 2, I2 = 94.4%)
Fasting insulin (μIU/mL)	5.50 (0.67–10.34, k = 3, I2 = 99.6%)	1.75 (−0.30–3.80, k = 6, I2 = 92.5%)	1.99 (0.48–3.50, k = 5, I2 = 83.5%)	1.46 (0.86–2.06, k = 1)	−4.62 (−8.19–−1.05, k = 1)
HbA1c (%)	−0.20 (−0.21–−0.19, k = 1)	0.67 (−0.77–2.12, k = 2, I2 = 98.7%)	−	0.30 (0.24–0.36, k = 1)	−
HOMA-IR (units)	−	0.38 (−0.23–0.99, k = 3, I2 = 81.6%)	0.29 (−0.16–0.75, k = 4, I2 = 92.4%)	0.96 (−0.07–2.00, k = 2, I2 = 98.9%)	0.50 (0.25–0.75, k = 1)
Systolic BP (mmHg)	−1.86 (−7.87–4.15, k = 5, I2 = 99.0%)	−4.88 (−21.19–11.44, k = 2, I2 = 98.9%)	0.87 (−6.15–7.90, k = 4, I2 = 92.5%)	−2.4 (−7.05–2.10, k = 4, I2 = 96.3%)	10.99 (9.43–12.55, k = 1)
Diastolic BP (mmHg)	−1.00 (−5.48–3.49, k = 4, I2 = 99.0%)	−2.76 (−13.55–8.04, k = 2, I2 = 98.4%)	−0.17 (−5.49–5.15, k = 4, I2 = 97.1%)	−0.56 (−1.26–0.13, k = 4, I2 = 16.0%)	7.14 (6.04–8.24, k = 1)

Abbreviations: FEP, first-episode psychosis; HDL, high-density lipoprotein; LDL, low-density lipoprotein

AP-naïve studies tended to report slightly larger anthropometric increases than mixed or previously treated samples (eg, BMI +2.94 vs +2.66 kg/m² at 6–12 months), and cohort studies consistently produced bigger changes than RCTs. Egger tests were non-significant for weight change at 8–14 weeks (P ≈ 0.34) and 26 weeks to 1 year (P ≈ 0.16), and for BMI change across all eligible timepoints (0–8, 8–14, 24–52, 53–156 weeks; all P>.17). Evidence of small-study effects was observed only for early weight gain at 4–8 weeks (P ≈ 0.01). Despite high heterogeneity (I² often >90%), the direction of weight, BMI, and waist change was stable across AP-naïve, JBI-only, and cohort-only subsets (see Supplement).

Metabolic Syndrome

Definitions of MeTS varied across studies, but the prevalence of MetS consistently increased from 8.2% at baseline (95% CI, 5.0-12.9; k = 8, I² = 77.8%), to 14.7% (95% CI, 4.5–38.6; k = 2, I² = 79.5%) at 3–12 months, and 18.0% (95% CI, 12.8-24.8; k = 3, I² = 68.0%) at 1–3 years (see Table 1). The choice of diagnostic criteria (ATP III vs IDF) explained some variability, but leave-one-out analyses showed comparable prevalence trends across definitions. Most MetS follow-up estimates were based on ≤3 studies, so subgroup evaluations by AP status or study design were underpowered and should be interpreted cautiously. Where available, AP-naïve and JBI-only subsets showed the same upward direction but wider confidence intervals (see Supplement).

Diabetes Prevalence and Glucose-Related Outcomes

The proportion with prediabetes (fasting glucose ≥100 mg/dL) was 4.8% (95% CI, 3.4-6.8; k = 6, I² = 55%) at baseline and increased to 9.4% (95% CI, 4.0-20.3; k = 4, I² = 92%) at 1–3 years. The proportion of individuals with diabetes (fasting plasma glucose ≥126 mg/dL or clinical diagnosis) was 0.5% (95% CI, 0.2-1.3; k = 3, I² = 0%) at baseline and increased to 5.0% (95% CI, 3.4-7.3; k = 5, I² = 89%) at 5–10 years. AP-naïve and JBI-included analyses showed similar patterns, and all contributing studies were observational cohorts; estimates at later timepoints were imprecise, with wide confidence intervals and sparse data. The following findings relating to diabetes were reported narratively due to single-study data: In one study³⁵ the incidence of diabetes was higher with second-generation antipsychotics (HR = 1.71) and first-generation antipsychotics (HR = 1.54) compared to age- and sex-matched general population controls over 12 years. Moreover, another study³⁶ showed a higher diabetes incidence with second-generation antipsychotics (OR = 1.37 for olanzapine+clozapine+aripiprazole; OR = 1.80 for olanzapine+clozapine) and first-generation antipsychotics (OR = 1.45 for low-potency antipsychotics) compared to no antipsychotic treatment over up to 10 years. Furthermore, a 15-year follow-up study³⁷ found that diabetes hospitalization was higher in Māori compared to non-Māori populations (HR = 1.44), while a retrospective cohort study³⁸ found an elevated risk of diabetes in contemporary versus historical cohorts (RR = 21.9) over 6 years.

Mean fasting glucose increased by 2.03 mg/dL (95% CI, 0.66-3.39; k = 19) at 4–8 weeks, 3.20 mg/dL (95% CI, 1.57-4.84; k = 7) at 1–3 years, and 7.54 mg/dL (95% CI, 1.17-13.91; k = 2) at >5 years. AP-naïve studies consistently showed larger glucose increases than mixed samples at 6–52 weeks (e.g., 4.15 vs 0.31 mg/dL at 6–12 months). RCTs reported smaller and more precise glucose changes than cohort studies, which often reported larger effects but with wider CIs. Eligible timepoints (4–8 weeks and 8–14 weeks) showed no evidence of small-study effects (P ≈ 0.81 and P ≈ 0.98, respectively).

Fasting insulin increased by 5.50 μIU/mL (95% CI, 0.67-10.34; k = 3) at 4–8 weeks and HOMA-IR increased by 0.96 units (95% CI, –0.07 to 2.00; k = 2) at 1–3 years. HbA1c changed by –0.20% (95% CI, –0.21 to –0.19; k = 1) at 4–8 weeks and 0.30% (95% CI, 0.24-0.36; k = 1) at 1–3 years (see Table 2). Across fasting insulin, HOMA-IR, and HbA1c, sensitivity analyses were based on very few studies (typically k ≤ 6, often k = 2–3) with high heterogeneity, so most pooled estimates were imprecise, and no Egger tests could be run. Sensitivity analyses did not show any stable or consistent shifts by AP-naïve status, JBI restriction, or study design.

Lipid-Related Outcomes

The proportion of FEP individuals with total cholesterol >240 mg/dL was 12.6% (95% CI, 3.5-36.3; k = 2, I² = 93%) at baseline and increased to 19.1% (95% CI, 3.2-62.3; k = 2, I² = 97%) at 1–3 years. The proportion with LDL cholesterol >175 mg/dL was 11.5% (95% CI, 2.7-37.9; k = 2, I² = 92%) at baseline and increased to 15.1% (95% CI, 1.9-61.5; k = 2, I² = 96%) at 1–3 years. The proportion with low HDL cholesterol (men <35 mg/dL, women <50 mg/dL) was 23.4% (95% CI, 13.6-37.3; k = 5, I² = 95%) at baseline and 33.1% (95% CI 18.1–52.7; k = 4, I² = 95%) at 1–3 years. The proportion with triglycerides >150 mg/dL was 6.2% (95% CI, 4.7-8.1; k = 4, I² = 0%) at baseline and 21.3% (95% CI, 17.9-25.0; k = 4, I² = 24%) at 1–3 years (see Table 1). Sensitivity analyses for prevalence outcomes in dyslipidemia (total cholesterol, LDL, HDL, and triglycerides) were constrained by very small k across subgroups, and no Egger tests were possible. Estimates for total cholesterol >240 mg/dL, LDL >175 mg/dL, and triglycerides >150 mg/dL remained stable across AP-naïve, cohort, and JBI-included models at baseline and 1–3 years, and leave-one-out differences were small.

Lipid outcomes showed increases in total cholesterol, LDL cholesterol, and triglycerides, and decreases in HDL across follow-up, despite high heterogeneity across studies (see Figure 3). Total cholesterol increased by 12.59 mg/dL (95% CI 7.24–17.95; k = 12) at 4–8 weeks and 8.91 mg/dL (95% CI 4.23–13.60; k = 4) at 1–3 years. LDL increased by 9.34 mg/dL (95% CI 5.37–13.30; k = 11) at 4–8 weeks and 7.00 mg/dL (95% CI 4.18–9.81; k = 5) at 1–3 years. Triglycerides increased by 32.91 mg/dL (95% CI 19.03–46.78; k = 16) at 4–8 weeks and 28.53 mg/dL (95% CI, 25.37-31.69; k = 7) at 1–3 years. HDL decreased by 0.75 mg/dL (95% CI, –2.89 to 1.38; k = 15) at 4–8 weeks and 3.62 mg/dL (95% CI, –5.79 to –1.46; k = 7) at 1–3 years (see Table 2).

Figure 3.

Line Plots of Metabolic Changes Over Time (Random-Effects Pooled Means, 95% CI)

Across total cholesterol, LDL, HDL, and triglycerides, cohort studies tended to report larger and more variable lipid changes than RCT and leave-one-out analyses indicated that several early timepoints were influenced by individual high-weight cohorts, yet the direction of change remained stable. AP-naïve cohorts showed slightly larger increases in triglycerides, total and LDL cholesterol, and more pronounced HDL decreases than mixed samples. Small-study effects were only tested where k ≥ 10 (total cholesterol and triglycerides at 4–8 and 8–14 weeks, LDL and HDL at 4–8 weeks, triglycerides 4–8 and 8–14 weeks), and Egger’s test was non-significant at all of these timepoints except triglycerides at 8–14 weeks (Egger P ≈ 0.02).

Blood Pressure

The proportion of FEP individuals with hypertension (≥130/85 mmHg) was 26.7% (95% CI, 23.7-29.9; k = 3, I² = 0%) at baseline and 25.2% (95% CI, 16.7-36.2; k = 2, I² = 74%) at 1–3 years (see Table 1). Longer-term hypertension estimates beyond 3 years were only available from single studies. Baseline estimates were consistent across AP-naïve and cohort subgroups, with no meaningful change when single studies were removed. Additionally, one study³⁵ showed that hypertension incidence was higher with second-generation antipsychotics (HR = 1.23) and first-generation antipsychotics (HR = 1.20) compared to controls over 12 years.

Blood pressure outcomes showed minimal overall change, with high heterogeneity at early follow-up and unstable pooled estimates in sensitivity checks. Mean systolic blood pressure changed by –1.86 mmHg (95% CI, –7.87 to 4.15; k = 5) at 4–8 weeks and –2.48 mmHg (95% CI, –7.05 to 2.10; k = 4) at 1–3 years. Diastolic blood pressure changed by –1.00 mmHg (95% CI, –5.48 to 3.49; k = 4) at 4–8 weeks and –0.56 mmHg (95% CI, –1.26 to 0.13; k = 4) at 1–3 years (see Table 2). Systolic and diastolic BP analyses showed very high heterogeneity at early and medium follow-ups, and leave-one-out sensitivity checks indicated that pooled estimates were unstable. AP-naïve and study-design subsets did not reveal consistent directional differences, and no BP outcome had k ≥ 10 to allow publication-bias testing.

Cardiovascular Outcomes

At 5–10 years, the pooled proportion of CVD hospitalizations in individuals with FEP, including ischemic heart disease, myocardial infarction, and stroke was 1.3% (95% CI, 1.0-1.9; k = 3, I² = 93%). All other cardiovascular outcomes were derived from single studies, of which the findings are reported narratively. One study showed that the proportion of people with FEP with abnormal QTc prolongation was 1.9% at baseline and increased to 5.7% within 4 weeks.³⁹ Ischemic heart disease was reported at 2.6% at 10-year follow-up,⁴⁰ while another study showed that myocardial infarction was reported at 0.2% at 7-year follow-up. Additionally, a further study found a higher incidence of hyperlipidemia incidence with second-generation antipsychotics (HR = 1.44) than first-generation antipsychotics (HR = 1.25) compared to general population controls over 12 years³⁵ (see Table 1).

Discussion

Summary of Main Findings

This systematic review and meta-analysis included 82 studies and examined the progression of cardiometabolic outcomes in individuals with FEP with no or minimal antipsychotic exposure at baseline. Findings show that metabolic changes emerge early, with rapid increases in weight, obesity prevalence, and waist circumference occurring within the first months to years after diagnosis. Glucose and lipid changes were also detectable within the first years. MetS prevalence increased from 8.5% at baseline to 18.2% by 1-3 years, diabetes from 0.5% to 5% at 5-10 years, and hypertriglyceridemia from 6.2% at baseline to 21.3%. at 1-3 years. Hypertension remained relatively stable over the first 3 years, and CVD hospitalizations were low (1.4%) within 5-10 years, suggesting early metabolic changes with potential long-term. Sensitivity analyses showed larger metabolic changes in cohort studies and antipsychotic-naïve individuals than in minimally exposed or trial-based FEP groups, consistent with more naturalistic sampling and fewer exclusions of higher-risk patients.

Comparison to Previous Literature

This review found that cardiometabolic changes in FEP occurred more rapidly than expected based on general population trajectories. Although most included studies did not follow healthy controls in parallel, population cohorts from similar periods show substantially slower metabolic change. Weight and BMI increases that emerged within 1–3 years were of a magnitude typically seen only after decades in community samples,^41–43 and obesity prevalence increased from 5.4% to 26.6% in a few years, reaching levels usually observed in midlife.⁴⁴ Abdominal obesity and lipid changes also progressed more rapidly than population norms, and diabetes prevalence reached midlife-equivalent levels within 5–10 years.^45,46 MeTS prevalence reached 18% by 1–3 years, approximating population rates expected a decade or more later in life.^47–49 While recent generational cohorts show faster weight increases overall, possibly due to changing dietary patterns,⁵⁰ the magnitude and speed of metabolic changes in FEP remain substantially greater.

Compared to prior reviews that included people with chronic or multi-episode cases of psychosis, our estimates of baseline MetS and diabetes were lower, which is consistent with earlier psychosis stages.¹³ Baseline prediabetes (4.8%) and diabetes (0.5%) in this review align with findings by Pillinger et al.,¹⁵ which reported early glucose dysregulation despite limited antipsychotic exposure.^51,52 Prediabetes prevalence in this review was lower at 4–14 weeks than at baseline, then higher again by 1–3 years. One possible explanation is that glycemic indices can vary with symptom severity, which often improves early in treatment before longer-term metabolic changes emerge.⁵³

Early CVD prevalence was low, increasing gradually with time, likely reflecting young age at onset and limited initial risk factors. Long-term risk is shaped by lifestyle, antipsychotics, and healthcare system factors and may not be as evident in this review due to limited data beyond 10 years. Blood pressure appeared stable in pooled analyses, but this may partly reflect sparse data rather than absence of change. The slow increase in diabetes despite weight gain may reflect delayed conversion from dysglycemia or early treatment masking incidence.^15,54 Taken together, these findings reinforce that cardiometabolic risk in psychosis begins early and progresses rapidly, rather than representing a distant consequence of psychosis.

Clinical Implications

Clinically, the results from this review emphasize the need for routine monitoring from FEP onset, particularly for weight, lipids, and blood pressure, within the first few months.³ Early intervention at this stage is both preventive and responsive to risks already present at presentation. Clinical strategies include selecting antipsychotics with lower metabolic risk, integrating diet and physical-activity support, and considering pharmacologic prevention when indicated. Evidence-based approaches, including early metformin initiation⁵⁴ and physical activity support,⁵⁵ can attenuate weight gain and improve treatment adherence, especially in young people.⁵⁶ Current guidelines increasingly recommend structured, multidisciplinary cardiometabolic programs as standard components of early intervention rather than optional add-ons.^3,57

However, studies of physical healthcare in people with serious mental health challenges show that this level of integrated care is often undermined by diagnostic overshadowing, unclear responsibility between services, and delayed investigation of physical symptoms.⁵⁸ In early intervention and community services, responsibility for cardiometabolic monitoring is often nominally shared with primary care, yet roles remain unclear, contributing to fragmented follow-up.^59,60 After time-limited early intervention services, long-term cardiometabolic care usually falls to general practice, where continuity, time constraints, and fragmented pathways increase the risk of loss to follow-up.⁶¹ These findings highlight the need for explicit shared-care agreements that clearly assign responsibility for cardiometabolic screening and treatment across mental health and primary care settings, as well as training and sufficient consultation time for GPs to deliver this long-term physical healthcare.

Equally important is education and shared decision-making. People with lived experience describe feeling unprepared for the visible and distressing effects of antipsychotic-induced weight gain when these are not discussed openly, leading to reduced trust and adherence.^62,63 Guidelines recommend transparent discussion of physical health risks from treatment initiation, supported by tailored prevention advice.^57,64 Such collaborative planning enables individuals to anticipate changes, make informed treatment decisions, and integrate physical health management into recovery.⁶⁵

Integration also means recognizing the subjective and social impact of these physical health changes. People describe the distress of having to adjust to a “new body,” one that feels unfamiliar or outside their control, with rapid changes in shape and appearance affecting self-esteem, body image, and identity.⁶⁵ These visible alterations can intensify stigma, make social situations feel exposing, and reduce willingness to engage with education, work, or relationships.^23,63 Addressing cardiometabolic risk early therefore supports not only physical health but also personal recovery, social participation, and quality of life.

Strengths and Limitations

This review’s strengths include a focus on antipsychotic-naïve and minimally exposed FEP cohorts and synthesized cardiometabolic outcomes across short-, medium-, and long-term follow-up windows that identify critical periods of change. The inclusion of both observational cohorts and interventional studies captured trajectories across routine-care and strict research settings. Importantly, differences by study design and antipsychotic exposure were directionally consistent across outcomes, with cohort studies and antipsychotic-naïve cohorts generally showing larger changes than trial or RCT designs, while confidence intervals often overlapped.

Several limitations should be noted. Between-study heterogeneity was high for many outcomes, reflecting variation in antipsychotic choice, dose and switching, baseline cardiometabolic exclusion criteria, concomitant psychotropic use, follow-up intensity, and healthcare context, which lowers confidence in some pooled estimates even where the direction of change is consistent. For outcomes supported by few contributing studies (typically k ≤ 3), pooled estimates were statistically imprecise and could not be examined with sensitivity analyses or tests of small-study effects, and should therefore be interpreted cautiously. Although most cohorts were antipsychotic-naïve at baseline, we combined strictly antipsychotic-naïve cohorts with those with <28 days of prior exposure to maximize available data; we recognize that even brief exposure can acutely alter glucose and insulin, so early treatment effects are likely to be present even in minimally exposed cohorts.^66–68

Reporting of antipsychotic dose, duration, switching, and particularly adherence was inconsistent, and no study provided adherence in a standardized quantitative format. In addition, many studies permitted concomitant psychotropics, which can independently influence metabolic outcomes, limiting attribution of observed changes to antipsychotics alone. Finally, sparse data beyond five years and concentration of studies in high-income settings limit inference about long-term trajectories and generalizability across healthcare systems and social contexts.

Future Research

Future work should focus on trials that start at treatment initiation in FEP and that evaluate stepped metabolic prevention packages, combining structured lifestyle support, co-commenced metformin, and glucagon-like peptide-1 (GLP-1) receptor agonists for non-response.^14,54,69 Implementation research is required to integrate these strategies into routine FEP care, including scalable multidisciplinary pathways and explicit shared-care arrangements with primary care.⁷⁰ Longer follow-up beyond 5–10 years is also required to clarify how early metabolic changes translate into later diabetes, cardiovascular disease, and premature mortality, as current evidence at later timepoints remains sparse and imprecise. Additionally, future research should examine the comparative trajectories of those who follow medication-based treatment and those who follow non-medication routes to recovery, to clarify how pharmacological and psychosocial approaches differentially shape physical health and functional outcomes. Finally, studies should further examine the lived experience of rapid metabolic change, including impacts on stigma, social participation, and recovery priorities, to ensure prevention strategies are acceptable and aligned with what matters to people receiving care.

Conclusion

In summary, this review shows early and ongoing metabolic change in FEP that points to later cardiometabolic disease. The signal is present at onset (glucose dysregulation), grows within months (weight, lipids), and continues over years. Psychosis affects the whole person, not just mental health. Multidisciplinary care should be established from the outset, with psychiatry collaborating closely with primary care, endocrinology, nursing, dietetics, and exercise professionals to support recovery and physical health.^3,70

Although the findings of this review highlight a concerning trajectory of early and persistent metabolic change, research on metformin has demonstrated its efficacy in preventing antipsychotic-induced weight gain and improving insulin sensitivity when initiated early in treatment.^54,56 Similarly, GLP-1 receptor agonists show promise in addressing obesity and glucose dysregulation in people receiving antipsychotic medication.⁶⁹ These developments show that metabolic complications are not inevitable and that early, proactive intervention can alter the long-term trajectory of physical health in psychosis.

How we Did that

The warning signs of physical illness that we looked at included: high blood pressure, heart attack, disease related to blood sugar, weight and fats in the blood. To examine these, we looked at previous research, undertaken by other researchers with expertise in this area. We then combined these results to get a more complete picture of the latest best evidence in the field that can help Doctors in understanding the physical health changes that can happen due to taking antipsychotic medications.

Key Findings

1. Disease related to blood sugar: Diabetes Type 2

• Finding: This study found that the rate at which people experiencing psychosis get type 2 diabetes increased from 0.5% at baseline to 5.0% at 5–10 years.

• Explanation: This suggests that the likelihood of getting type 2 diabetes does increase over time as a result of first-episode psychosis or the medication used to treat it. This review also found that prediabetes rates increased from 4.8% to 9.5% at 1–3 years, suggesting that type 2 diabetes could be a real risk for people who take antipsychotic medication for long periods of time.

1. Metabolic Syndrome

• Finding: This study found that in people experiencing psychosis, MeTS rates increased from 8.5% at baseline to 18.2% at 1–3 years.

• Explanation: MeTS is a combination of high blood pressure, high blood sugar, excess body fat around the waist, and abnormal cholesterol levels. This increase shows that people who experience psychosis develop these combined risks early, likely influenced by medication, lifestyle changes, or the illness itself.

1. Weight Gain

• Finding: This study found that those who experience psychosis will experience a significant amount of weight gain, particularly when taking antipsychotic medication for a long period of time.

• Explanation: The study found that significant weight gain begins early after a person begins taking antipsychotic medication. This is concerning as weight gain can lead to physical illnesses, such as obesity, heart disease and diabetes. This weight gain occurs as the medication might cause changes to a person’s appetite and the way they process food, causing physical illness. Additionally, the medication can impact the amount of physical activity a person does due to issues relating to obesity, including back pain, etc. However, it is important to note that certain antipsychotics are linked to greater weight gain than others, highlighting that the type of medication used can influence these outcomes. Over one third of people experiencing psychosis gained >7% of their body weight within 1–3 years of starting on antipsychotic medication, showing how quickly this can happen

1. Fats in the Blood and Blood Pressure

• Finding: Both fats in the blood and blood pressure have been shown to increase over time in people experiencing psychosis.

• Explanation: Increased fats in the blood, along with high blood pressure, can cause heart issues for people with a first episode of psychosis, even when there are no antipsychotics in a person’s system. Such changes in fats in the blood and high blood pressure could be from the stress associated with experiencing psychosis, the lack of motivation to exercise, or the changes in diet preferences of those with psychosis.

1. The combination of changes in blood sugar, weight, fats in the blood and blood pressure

• Finding: This study finds that the combination of the above changes in blood sugar, weight, fats in the blood and blood pressure increases over time.

• Explanation: This combination of increased blood sugar, weight, fats in the blood and blood pressure is also called metabolic syndrome. The presence of this suggests that people who experience first episode psychosis may also experience a steady rate of metabolic changes resulting in poor physical health. Such changes are influenced by stress, lifestyle changes, and the introduction of antipsychotic medication. This suggests that Doctors need to examine these areas on a regular basis once they prescribe antipsychotic medications.

1. Heart Health

• Finding: This study found that heart-related issues, like hospitalizations for heart disease, were low at 1.4% after 5–10 years, likely due to the young age of participants (mostly 20s–30s).

• Explanation: This low rate suggests that serious heart problems may develop later, possibly influenced by early weight and fat increases, highlighting the need for long-term monitoring.

Take-Away Points

The findings of this study suggest that changes in physical health, such as weight gain, blood sugar, and blood pressure, are likely to occur from the beginning of psychosis, even with little or no prior use of antipsychotic medications. This suggests that such physical changes are not solely a result of antipsychotic use alone; in fact, they are driven by multiple factors associated with psychosis, such as stress, lifestyle, and psychosis itself. As such, it is important that physical health is monitored from the moment a person is diagnosed with a first episode of psychosis.

Recommendations

• Routine Monitoring: Regular check-ups of weight, blood pressure, fats in the blood, and blood sugar are essential to detect and address any physical health changes quickly.

• Health Interventions: Instead of just treating symptoms of psychosis, doctors need to treat the whole person, including their physical health needs. Links with nutritionists, physical activity specialists, and psychiatrists are essential to reducing the risks associated with first episode psychosis and antipsychotic medication use.

• Patient Education: Raising awareness with those experiencing their first episode of psychosis regarding the potential health challenges that they might experience is crucial in ensuring that these individuals engage with healthcare services.

Author Contributions

A.Z., B.O.D., M.C., K.O.C., and M.J.N. contributed to the protocol development, conceptualization and overall design of the systematic review. Screening was conducted by A.Z. and J.M., and data extraction was completed by A.Z., A.S., and J.M. M.J.N. contributed to the lay summary and interpretation of findings from a lived experience perspective. A.Z. led the drafting of the manuscript, with supervision and feedback from D.C., M.C., K.O.C., M.J.N., B.O.D., and M.C. J.M. and A.S. provided additional feedback on the final manuscript draft.

Funding

This work was supported by the Health Research Board Psychosis Ireland Structured Training and Research (PSI-STAR) Clinical Doctoral award CDA 2021-0005 and by Taighde Éireann – Research Ireland under Grant number 21/RC/10294_P2 at FutureNeuro Research Ireland Centre for Translational Brain Science. The funders did not have any role in the preparation of the manuscript.

Conflicts of Interest

The authors declare no conflict of interest.

Data Availability

The data that support the findings of this study are openly available in figshare: Zierotin, Anna (2025). Short-, Medium-, and Long-term Cardiometabolic Outcomes in First-Episode Psychosis: Study Datasets. Figshare. Dataset. https://doi.org/10.6084/m9.figshare.30316609.

Analytic Code Availability

The analytic code that support the findings of this study are openly available in figshare: Zierotin, Anna (2025). R Code for Systematic Review and Meta-Analyses: Short-, Medium-, and Long-term Cardiometabolic Outcomes in First-Episode Psychosis. figshare. Software. https://doi.org/10.6084/m9.figshare.30329851.v1.

Transparency Declaration

The lead author affirms that the manuscript is an honest, accurate, and transparent account of the study being reported; that no important aspects of the study have been omitted; and that any discrepancies from the study as planned have been explained.