Multidimensional Epigenetic Clocks Demonstrate Accelerated Aging Across Physiological Systems in Schizophrenia: A Meta-Analysis
===============================================================================================================================

* Zachary M. Harvanek
* Raghav Sehgal
* Daniel Borrus
* Jessica Kasamoto
* Ahana Priyanka
* Michael J. Corley
* Christiaan H. Vinkers
* Marco P. Boks
* Ryan Smith
* Varun B. Dwaraka
* Jessica Lasky-Su
* Albert T. Higgins-Chen

## Abstract

Importance: Schizophrenia is associated with increased age-related morbidity, mortality, and frailty, which are not entirely explained by behavioral factors. Prior studies using epigenetic clocks have suggested that schizophrenia is associated with accelerated aging, however these studies have primarily used unidimensional clocks that summarize aging as a single “biological age” score.

Objective: This meta-analysis uses multidimensional epigenetic clocks that split aging into multiple scores to analyze biological aging in schizophrenia. These novel clocks may provide more granular insights into the mechanistic relationships between schizophrenia, epigenetic aging, and premature morbidity and mortality.

Study selection: Selected studies included patients with schizophrenia-spectrum disorders and non- psychiatric controls with available DNA methylation data. Seven cross-sectional datasets were available for this study, with a total sample size of 1,891 patients with schizophrenia and 1,881 controls.

Data extraction and synthesis: Studies were selected by consensus Meta-analyses were performed using fixed-effect models.

Main outcomes and measures: We analyzed multidimensional epigenetic clocks, including causality- enriched CausAge clocks, physiological system-specific SystemsAge clocks, RetroelementAge, DNAmEMRAge, and multi omics-informed OMICmAge. Meta-analyses examined clock associations with schizophrenia disease status and clozapine use, after accounting for age and sex.

Results: Overall SystemsAge, CausAge, DNAmEMRAge, and OMICmAge scores demonstrated increased epigenetic aging in patients with schizophrenia after strict multiple-comparison testing. Ten of the eleven SystemsAge sub-clocks corresponding to different physiological systems demonstrated increased aging, with strongest effects for Heart and Lung followed by Metabolic and Brain systems. The causality- enriched clocks indicated increases in both damaging and adaptive aging, though these effects were weaker compared to SystemsAge scores. OMICmAge indicated changes in multiple clinical biomarkers, including hematologic and hepatic markers that support system-specific aging, as well as novel proteins and metabolites not previously linked to schizophrenia. Most clocks demonstrated age acceleration at the first psychotic episode. Notably, clozapine use was associated with increased Heart and Inflammation aging, which may partially be driven by smoking. Most results survived strict Bonferroni multiple testing correction.

Conclusions and relevance: These are the first analyses of novel multidimensional clocks in patients with schizophrenia and provide a nuanced view of aging that identifies multiple organ systems at high risk for disease in schizophrenia-related disorders.

## Introduction

Schizophrenia is a severe psychiatric disorder associated with major reductions in life expectancy of 10-20 years1. Much of the increased mortality stems from natural causes2,3 spanning multiple organ systems4 and multiple physical comorbidities5–8. Some treatments, especially clozapine, may decrease mortality from both non-natural and natural causes9. Increased morbidity and mortality in patients with schizophrenia is associated with accelerated aging, including elevated biomarkers of inflammation, oxidative stress, and age-associated proteins that predict mortality10–12.

Epigenetic clocks are commonly utilized biomarkers that estimate biological age using DNA methylation data13–15. Studies of schizophrenia using epigenetic clocks have shown mixed results depending on the specific clock used15,16. Early studies utilizing Horvath’s multi-tissue clock found no changes in epigenetic age in schizophrenia,14,17 likely because the clock’s training included schizophrenia samples and thus the clock is trained to ignore CpGs with altered aging patterns in schizophrenia16.

More recent work has utilized second- and third-generation epigenetic clocks (GrimAge, PhenoAge, and DunedinPACE), which show increased epigenetic aging in schizophrenia13,18. These epigenetic clocks better align with epidemiologic data, are more likely to capture causal mechanisms in the aging process, and are less likely to be due to cohort or period effects19–22.

However, traditional epigenetic clocks generally summarize aging as a single number, treating aging as unidimensional. This is problematic for studies of aging and schizophrenia, which are both complex, multi-faceted, heterogeneous phenomena in that they involve many different biological processes. Traditional clocks do not allow more nuanced questions to be asked about aging in schizophrenia. For example, could a subset of epigenetic changes seen in patients with schizophrenia actually reflect adaptive mechanisms that protect individuals with age-related disease? Do some aging processes constitute specific responses to behavioral or environmental stressors? Is the aging process uniformly accelerated in the entire body, or are specific physiologic systems disproportionately affected?

To capture multidimensionality in aging, novel epigenetic clocks have recently been developed that report a panel of biological age scores. SystemsAge is composed of 11 systems-based epigenetic clocks, developed by relating blood DNA methylation to 133 clinical and functional biomarkers organized by physiological system and then training mortality predictors for each system23. Causality-enriched epigenetic clocks (CausAge) were developed using Mendelian Randomization techniques that identify CpG sites that are potentially causal for age-related traits24, including both detrimental (DamAge) and beneficial (AdaptAge) changes. OMICmAge was trained to predict mortality using DNA methylation data integrating 40 epigenetic biomarker proxies of proteins, metabolites, and clinical biomarkers that provide insight into specific biological processes25. Other interesting clocks include IntrinClock26, intended to capture aging unrelated to changes in cell composition, and RetroelementAge27, intended to capture aging at CpGs linked to retroelements.

Here, we systematically investigate these new multidimensional epigenetic clocks in 7 datasets of patients with schizophrenia and non-psychiatric controls. We hypothesized that due to increased risk of numerous diseases in schizophrenia4,8, most physiological systems would demonstrate increased epigenetic aging, but that specific systems would be most altered to reflect particularly high disease risks in schizophrenia (e.g. Lung for pneumonia, COPD, and smoking, or Brain for dementia risk). We use meta-analyses to assess for schizophrenia disease effects across studies, then further assessed for effects of first-episode psychosis, interactions with age, sex and smoking status, and clozapine use.

## Methods

### Selection of Datasets

We selected available epigenetic datasets of patients with schizophrenia-spectrum disorders and non- psychiatric controls28,29. Most datasets utilized the Illumina Infinium 450K BeadChip, though one dataset utilized the Illumina Methylation EPICv1 BeadChip. Individuals for whom chronologic age was not available were excluded from the analyses.

### Clock Calculation

Details for clock calculations can be found in Supplementary Methods. All clocks used in these analyses can be found in Supplementary Table 1. Briefly, clocks were calculated using the methylCIPHER package in R30 as described by23,24,31 or using code provided by the authors27,32.

### Cell-type composition estimates

Cell-type composition estimates were obtained using EpiDISH33. When accounting for cell-type composition, Neutrophils were dropped to avoid overfitting, and proportions of NK cells, B cells, CD4T cells, CD8T cells, Monocytes, and Eosinophils were included.

### Epigenetic smoking estimates

As smoking data was not available for most datasets, we used the GrimAge component DNAmPACKYRS, which is a proxy of smoking pack-years predicts mortality better than self-reported pack-years20.

### Statistical analyses

To calculate standardized effect sizes for the effect of schizophrenia, epigenetic ages were first regressed onto age, sex, and any other included covariates for the analyses (e.g., cell-type composition, smoking). These residuals were then scaled such that for the controls, the standard deviation for a given study = 1 and residual mean = 0, and then the final model was a multivariable regression of disease status, age, sex, and any other covariates regressed onto the scaled residuals. To evaluate standardized effect sizes for age-by-disease interaction, the same procedure was used. When examining clozapine effects, a similar procedure was used, limiting analyses to only individuals with schizophrenia and data regarding whether they had taken clozapine or not.

Statistical analyses were performed using R version 4.0.234. Meta-analyses were performed using a fixed-effects model in the Metafor package35. Weights for studies were calculated as 1/SE2. Heatmap and scatter plots were created using ggplot236. Statistics represent nominal values for consistency.

However, strict correction for multiple comparisons via the Bonferroni method adjusts for 73 comparisons to account for all clocks and sub-clocks presented in the heatmap in Figure 4 (nominal, unadjusted p of 0.000685 = adjusted p of 0.05). This adjustment is likely overly stringent as clocks are intercorrelated and capture overlapping aging signals.

## Results

We identified seven different cohorts of patients (Table 1) composed of 2,210 patients with schizophrenia-spectrum disorders and 1,936 non-psychiatric controls. 549 individuals were excluded due to missing age data. Age distributions for each cohort are in Supplementary Figure 1. Two cohorts (GEO152026 and GEO152027) include individuals with first-episode psychosis (n = 716). Three cohorts (GEO116379, GEO80417, and GEO84727) include information regarding clozapine treatment (number on clozapine = 225, number confirmed not on clozapine = 406).

View this table:
[Table 1:](http://medrxiv.org/content/early/2024/10/30/2024.10.28.24316295/T1)

Table 1: 
Characteristics of included Datasets

Summary of the seven datasets analyzed here. Age indicates mean +/- SD, range in parentheses. SCZ: Schizophrenia-spectrum disorders. FEP: First-Episode Psychosis. IoPPN: Institute of Psychiatry, Psychology, and Neuroscience

### Multidimensional Clocks are Altered in Patients with Schizophrenia

Meta-analyses demonstrated significantly increased epigenetic age in patients with schizophrenia in all three (DamAge, AdaptAge, and CausAge) causality-enriched clocks, 10 out of 11 systems-based clocks, as well as total SystemsAge, DNAmEMRAge, OMICmAge, IntrinClock, and RetroelementAge (Figure 1, Supplementary Table 1). The largest effects were seen for Heart (β = 0.87, p = 8.5E-121), SystemsAge (β = 0.85, p = 1.1E-114) and Lung (β = 0.82, p = 1.5E-108) clocks. All results except IntrinClock (unadjusted p = 0.00363) and RetroelementAge (unadjusted p = 0.00251) survived strict Bonferroni multiple testing correction. No significant effect was observed for the Hormone system clock (unadjusted p = 0.79). Of the methylation-based predictors included in OMICmAge, the majority were significantly altered (Supplementary Figure 2). When males and females are analyzed separately, we observe similar patterns (Pearson’s r = 0.96 between effect sizes calculated in males and females separately, Supplementary Figure 3).

![Figure 1:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/10/30/2024.10.28.24316295/F1.medium.gif)

[Figure 1:](http://medrxiv.org/content/early/2024/10/30/2024.10.28.24316295/F1)

Figure 1: 
Patients with Schizophrenia have higher epigenetic age compared to controls

Body plot demonstrates comparative effect sizes of meta-analysis in Systems-based clocks, with larger text size and bolder color indicating strongest effects. Forest plots of meta-analysis results, with positive numbers indicating epigenetic age is higher in patients with schizophrenia. 95th percentile confidence intervals are presented. P values represent nominal significance; Bonferroni correction suggests significance (alpha < 0.05) at nominal p < 0.000685.

To confirm previous findings13,14,17,18 using a larger meta-analysis, we repeated our analysis for unidimensional clocks. We replicated previous results demonstrating accelerated aging across almost all traditional clocks, except the Horvath multi-tissue clock as expected (Supplementary Figure 4). GrimAge (both V1 and V2) showed the strongest standardized effect size of 0.98. The principal component versions of clocks31, largely recapitulated findings from the original clocks.

### Multidimensional Clocks are Altered in First-Episode Psychosis less than in Prevalent Schizophrenia

We asked whether these differences in epigenetic aging pre-dated psychotic symptoms by examining individuals with first-episode psychosis. For the causality-enriched clocks, only AdaptAge shows nominally significantly higher epigenetic age in patients with first-episode psychosis compared to controls, while for the systems-based clocks broadly show significantly higher epigenetic age except the Hormone clock and, after Bonferroni correction, the Immune clock (Figure 2, forest plots). Conclusions regarding OMICmAge, DNAmEMRAge, and RetroelementAge remain unchanged when just considering first-episode psychosis. First-episode psychosis is generally associated with smaller effect sizes than non- FEP, with notable exceptions of RetroelementAge and OMICmAge (Figure 2, scatterplot).

![Figure 2:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/10/30/2024.10.28.24316295/F2.medium.gif)

[Figure 2:](http://medrxiv.org/content/early/2024/10/30/2024.10.28.24316295/F2)

Figure 2: 
Patients with first-episode psychosis show higher epigenetic age in systems-based clocks

Forest plots of meta-analysis results in patients with first-episode psychosis, with positive numbers indicating epigenetic age higher in patients with first-episode psychosis compared to controls. 95th percentile confidence intervals are presented. Scatter plot demonstrates comparison of effect sizes in patients with first-episode psychosis to non-first-episode psychosis cases. Red line is the identity line (x=y). P values represent nominal significance; Bonferroni correction suggests significance (alpha < 0.05) at nominal p < 0.000685.

We also asked whether effects of schizophrenia on clocks might change with age. analyses excluded GEO116379 due to a narrow age range, and demonstrated a significant age-by-disease interaction in the DNAmEMRAge (p = 0.000058), Inflammation (p = 0.00059), Heart (p = 2.7E-10), Lung (p = 8.4E-15), and total SystemsAge (9.2E-10) clocks, and nominally significant interactions in the OMICmAge (p = 0.010), Hormone (p = 0.046), and Metabolic (p = 0.0043) clocks. Sensitivity analyses including GEO116379 identify significant interactions in the same clocks.

### Clozapine is associated with age acceleration across most systems-based clocks

We next sought to identify whether specific medications (i.e., Clozapine) might contribute to higher epigenetic age in schizophrenia. Patients with schizophrenia treated with clozapine showed no differences in causality-enriched clocks compared to patients with schizophrenia not treated with clozapine. However, we observed significantly higher epigenetic age in 7 of 11 systems-based clocks, with the largest effects seen in Inflammation and Heart. Smaller effects were seen for Lung, Metabolic, Immune, Kidney, and Brain that did not pass Bonferroni correction. The overall SystemsAge score showed a similar effect as the strongest systems. No differences were observed in DNAmEMRAge, OMICmAge, IntrinClock, or RetroelementAge with clozapine treatment (Figure 3).

![Figure 3:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/10/30/2024.10.28.24316295/F3.medium.gif)

[Figure 3:](http://medrxiv.org/content/early/2024/10/30/2024.10.28.24316295/F3)

Figure 3: 
Clozapine treatment is associated with accelerated aging in the inflammation, heart, metabolic, and lung clocks

Forest plots of sub-analysis comparing patients with schizophrenia on clozapine to patients with schizophrenia not on clozapine. Positive value indicates higher epigenetic age in patients on clozapine. Analsyes of DNAmPACKYR (an epigenetic marker of smoking) demonstrates that patients on clozapine smoke more than patients not on clozapine. Scatter plot demonstrates that smoking broadly reduces effect sizes of clozapine. Red line is the identity line (x=y). P values represent nominal significance; Bonferroni correction suggests significance (alpha < 0.05) at nominal p < 0.000685.

As we observed evidence of higher smoking rates in patients with clozapine in our study (Figure 3), we next asked whether smoking accounted for these differences. Most systems showed weaker effects after accounting for smoking, with Lung showing the greatest attenuation as expected. After accounting for smoking, only the Inflammation, Heart, and Total SystemsAge clocks remain nominally significantly elevated with clozapine use (Figure 3, Scatterplot).

### Neither Cell-type composition nor Smoking account for the differences in Systems-Based Clocks in patients with Schizophrenia

We examined potential drivers of differences in epigenetic age, including cell-type composition37 and smoking. In general, effect sizes were lower after adjusting for cell-type composition, and CausAge was no longer significantly higher in patients with schizophrenia (p = 0.067). The conclusions from the Systems-based clocks, RetroelementAge, DNAmEMRAge, IntrinClock, and OMICmAge were unchanged after accounting for cell-type proportions (See Figure 4A for a summary).

![Figure 4:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/10/30/2024.10.28.24316295/F4.medium.gif)

[Figure 4:](http://medrxiv.org/content/early/2024/10/30/2024.10.28.24316295/F4)

Figure 4: 
The association between epigenetic aging and schizophrenia is robust to cell composition and smoking, and represents a distinct phenotype from prior body health scores

A: Heatmap demonstrating effect size and significance of the association between schizophrenia, schizophrenia-by-age interaction, and clozapine with specific epigenetic clocks. B Scatter plot of SystemsAge standardized effect sizes versus prior reported body health scores. Red line is the identity line (x=y). C: Scatter plot using rank-order of SystemsAge versus prior reported body health scores. Red line is the identity line (x=y). D: Histogram of # of highly accelerated systems-based clocks (age > 1SD from 0) by disease status. individuals with schizophrenia (blue) or controls (red).

When accounting for smoking, DamAge is no longer significantly associated with schizophrenia (p = 0.58). AdaptAge and CausAge remained nominally significant. Of the systems-based clocks, the Lung clock experienced the greatest decrease, with a 8.6-fold reduction of effect size and is no longer statistically significant. Heart also showed a reduction in effect size but is still statistically significant. All other systems remained robustly elevated in schizophrenia after correcting for smoking.

RetroelementAge and OMICmAge were no longer significant after accounting for smoking, while IntrinClock remained nominally significant and DNAmEMRAge remained significant (See Figure 4A for a summary).

### SystemsAge Identifies Deficits in Distinct Physiologic Systems when compared to biochemical, imaging, and physical assessments

We next compared conclusions from SystemsAge to a recent study examining the health of multiple physiologic systems in neuropsychiatric disorders utilized brain imaging, physical assessment, and biochemical assays38. We identified little statistical correlation between these methods in terms of the effects on organ systems in schizophrenia when considering either overall estimated effect size (Figure 4B, p = 0.49, R2 = 0.083) or rank order (Figure 4C, p = 0.65, R2 = 0.036). Notably, patients with schizophrenia are more likely than controls to have increased epigenetic aging in multiple systems- based clocks (Figure 4D, p < 1E-15).

## Discussion

Here, we show that schizophrenia is associated with broad accelerated aging in multidimensional epigenetic clocks. These clocks provide a far more nuanced picture of aging in through subscores that each capture different aspects of aging. These subscores reveal schizophrenia is characterized by both damaging and adaptive age-related changes, accelerated aging in most physiological systems, many age- related metabolites and proteins, as well as retroelements. Critically, the large sample size and meta- analytic design of this study allowed us to obtain robust results that generally pass strict Bonferroni multiple testing correction even when examining numerous clocks simultaneously. Of note, this correction is likely overly stringent since it assumes the tests are independent, but the clocks are intercorrelated and not fully independent. Thus, even results that do not pass multiple testing correction likely are trustworthy findings. Sensitivity analyses indicated the observed effects are generally independent of cell-type composition and sex. Interestingly, smoking, clozapine, and first-episode psychosis each affected specific clock subscores, highlighting the utility of multidimensional clocks in disentangling the effects of different clinical variables on aging.

Importantly, the multidimensional clocks correlate with the known epidemiology of schizophrenia. Large studies including meta-analyses have demonstrated that individuals with schizophrenia have increased disease and mortality rates from natural causes covering every physiological system4,8. Accordingly, 10/11 of the SystemsAge subscores are increased in schizophrenia. Because each SystemsAge subscore has specific associations with outcomes related to that system23 (e.g. lung cancer for Lung, cognition for Brain, diabetes for Musculoskeletal and Metabolic), SystemsAge could help explain the increased disease and mortality risks in schizophrenia. The only score not increased in schizophrenia is Hormone, which was previously reported to be most related to thyroid disease and cancer risk. Accordingly, thyroid disease is not elevated in schizophrenia8, and cancer shows the smallest increased risk in Correll et al4. As expected, patients with schizophrenia are more likely to show multi-system accelerated aging, consistent with the 69% increased risk of multimorbidity in schizophrenia39. Importantly, SystemsAge can better capture the heterogeneity of risk in schizophrenia and thus could identify which diseases individuals with schizophrenia are at greatest risk for.

A recent study by Tian et al. developed system scores using clinical data and examined changes in schizophrenia38. Interestingly, while we found that Heart and Lung were the scores with greatest increase in schizophrenia, Tian et al. found nearly no change in their Cardiovascular score and a modest change in the Pulmonary score. Clinical experience is more consistent with markedly increased heart and lung aging, given the high rates of smoking in schizophrenia40, known adverse effects associated with antipsychotic medications41 (weight gain, myocardial infarction, pneumonia), and mortality due to pneumonia (RR 7), any respiratory cause (RR 3.75) and cardio-cerebrovascular causes (RR 3.47)4. Given the discrepant results from the two methods, it will be interesting to determine if combining clinical and laboratory-based risk factors with epigenetic scores may provide complementary information on risks of comorbidities in patients with schizophrenia.

Increased epigenetic age was noted at first-episode psychosis, suggesting many changes are detectable early in the disease. However, effect sizes were smaller in first-episode psychosis compared to prevalent schizophrenia for nearly all clocks. Interestingly, only a subset of clocks accelerated over time as suggested by age-by-disease interactions - these were the Heart, Lung, Inflammation, full SystemsAge, and DNAmEMRAge clocks. It is possible changes in these systems reflect the cumulative effects of factors associated with psychosis (e.g., smoking, stress) and treatment of psychosis (e.g., atypical antipsychotics). This suggests that older individuals with schizophrenia may be particularly vulnerable to a cardiovascular, pulmonary, and inflammatory age-related diseases.

Clozapine has unique treatment benefits but also greater metabolic side effects and can induce serious cardiac, hematologic, and neurological adverse events. We find that clozapine treatment is associated with acceleration in SystemsAge, multiple SystemsAge subclocks especially Inflammation and Heart but also in Metabolic and Brain, as well as in multiple hematologic markers of OMICmAge. These changes may reflect the side effect profile of clozapine, and indeed longitudinal studies have shown that clozapine can directly impact the methylome 42. However, there are other explanations: treatment- resistant individuals, in whom clozapine is primarily used, may represent a distinct clinical population with unique risks43. Notably, prior meta-analyses have suggested that clozapine is associated with decreased mortality in patients with schizophrenia, despite its known effects on cardiometabolic risk factors4. Further longitudinal studies are needed to determine potential causal relationships between clozapine, epigenetic aging, and mortality.

Our analysis of OMICmAge reveals novel insights into clinical biomarkers, proteins and metabolites altered in schizophrenia. Previously, we showed that DNAm proxies of serum B2M, Cystatin C, GDF-15, TIMP-1, ADM, and PAI-1 (components of GrimAge) are elevated in schizophrenia, which matches the literature concerning these proteins13. This suggests DNAm proxies can be used for discovery. In some cases, the DNAm proxies may even be more useful. Prior results showed stronger associations with mortality for a DNAm proxy of smoking pack-years compared to self-reported smoking pack-years20, and stronger associations with brain health outcomes for a DNAm proxy of CRP compared to directly measured CRP45. OMICmAge predicted changes in multiple clinical biomarkers, especially increases in DNAm proxies of RDW and decreases in hemoglobin, hematocrit, and albumin, consistent with anemia, malnutrition, other comorbidities, and medication effects in schizophrenia46,47. Notably, higher RDW and lower albumin have been associated with increased mortality in the general population in both NHANES and the UK Biobank48. IGFBP2 was found to be increased, consistent with prior literature49. IGFBP-2 may play a role in increased metabolic risk and altered synaptic plasticity in schizophrenia or treatment49,50.

Potentially novel proteins and metabolites that could play a role in features of schizophrenia include N- acetyl-isoputreanine (cognition51), (4-hydroxy)phenylacetylglutamine (heart disease52), vanillactate (heart disease53), carboxypeptidase B2 (thromboembolic disease54), histone H2B type 1-K (cellular senescence55), mimecan (food intake56), and chordin-like protein 1 (cognitive decline57, adipogenesis58). Notably, directionality is not necessarily consistent with prior literature – for example we found increased DNAm-predicted cystine though a prior study found reduced directly measured cystine59, suggesting a complex relationship between cystine, DNAm and schizophrenia. The biology of these proteins and metabolites in schizophrenia represent fertile areas for future investigation.

Limitations of our study include its cross-sectional nature and limited datasets for first-episode psychosis and clozapine. The absence of patients with schizophrenia without medication prevents analyses of general anti-psychotic treatment and epigenetic aging. Future longitudinal and interventional studies with more detailed phenotypic data will be needed to determine whether antipsychotics or other factors associated with schizophrenia contribute to accelerated aging.

## Conclusion

In this meta-analysis, we identify a rich tapestry of accelerated epigenetic aging in schizophrenia- spectrum disorders. Specific physiological systems are particularly affected, changes can be either damaging or adaptive changes, and many age-related metabolites, proteins, and retroelements are affected. These findings are robust after strict multiple testing correction and correcting for covariates. Factors such as smoking, first-episode psychosis, and clozapine have effects on particular subsets of clocks. These clocks may complement clinical data in identifying and preventing aging health risks in patients in schizophrenia.

## Supporting information

Supplementary Materials [[supplements/316295_file06.docx]](pending:yes)

Supplementary Table 1 [[supplements/316295_file07.xlsx]](pending:yes)

Supplementary Figure 1 [[supplements/316295_file08.pdf]](pending:yes)

Supplementary Figure 2 [[supplements/316295_file09.pdf]](pending:yes)

Supplementary Figure 3 [[supplements/316295_file10.pdf]](pending:yes)

Supplementary Figure 4 [[supplements/316295_file11.pdf]](pending:yes)

## Data Availability

All data is available on NCBI GEO ([https://www.ncbi.nlm.nih.gov/geo/](https://www.ncbi.nlm.nih.gov/geo/)) and datasets are listed in Table 1. Information on clozapine status can be found in their respective papers.

[https://www.ncbi.nlm.nih.gov/](https://www.ncbi.nlm.nih.gov/) 

## Data availability

All data is available on NCBI GEO ([https://www.ncbi.nlm.nih.gov/geo/](https://www.ncbi.nlm.nih.gov/geo/)) and datasets are listed in Table 1. Information on clozapine status can be found in their respective papers13,60.

## Code availability

Code to calculate all clocks except for OMICmAge and DNAmEMRAge will be accessible at [https://github.com/HigginsChenLab/methylCIPHER](https://github.com/HigginsChenLab/methylCIPHER) after publication. Code to calculate OMICmAge, DNAmEMRAge and associated algorithms will be accessible via TruDiagnostic’s DNAm Analysis Software after publication. You can request access to the software at [https://www.trudiagnostic.com/softwarerequest](https://www.trudiagnostic.com/softwarerequest) .

## Author contributions

Z.M.H. and A.H.C. designed the study and drafted the initial manuscript. Z.M.H. performed the analyses.

R.S. contributed body plots and code for analyses. Z.M.H., A.H.C., R.S., D.B., J.K., and A.P. contributed to data cleaning and clock calculation. A.H.C. conceived the project and provided supervision. C.H.V. and M.P.B. provided data and conceptual input. M.J.C., V.B.D., J.L.S., and R.S. provided clock code and insights. All authors reviewed and contributed to the manuscript.

## Conflicts of Interest

A.H.C. and R. Sehgal are scientific advisors to TruDiagnostic Inc. and are named as inventors of SystemsAge which has been patented and licensed to TruDiagnostic Inc. R. Sehgal has received consulting fees from the LongevityTech.fund, Healthy Longevity Clinic and Cambrian BioPharma unrelated to this publication. J.L.S. is a scientific advisor to Precion Inc. and TruDiagnostic Inc. V.B.D. and R. Smith are employees of TruDiagnostic Inc. J.L.S., V.B.D., and R. Smith developed OMICmAge. The other authors do not declare any conflicts of interest.

## Key Points

Question: Do novel, multidimensional epigenetic clocks demonstrate accelerated aging in schizophrenia? Findings: In this meta-analysis, patients with schizophrenia-spectrum disorders demonstrated evidence of broadly accelerated aging in multiple types of epigenetic clocks. This age acceleration is particularly evident in the Heart and Lung systems and is already evident by the time of the first psychotic episode. Meaning: Novel epigenetic clocks may help identify patients with schizophrenia-spectrum disorders at risk for multiple other health comorbidities.

## Acknowledgements/Funding

This work was supported by the Yale Physician Scientist Development Award (Z.M.H.) and CTSA (Z.M.H.; NIH UL1 TR001863), grants from the National Institute on Aging (A.H.C.; 1R01AG065403), the National Heart Lung Institute (J.L.S.; R01HL123915, R01HL155742, and 1R01HL169300), and the Impetus Grants (R. Sehgal). Additionally, Dr. Harvanek is a Pepper Scholar with support from the Claude D. Pepper Older Americans Independence Center at Yale School of Medicine (P30AG021342). Dr. Higgins-Chen was supported by a Pilot Grant from the Claude D. Pepper Older Americans Independence Center at Yale School of Medicine (P30AG021342).

*   Received October 28, 2024.
*   Revision received October 28, 2024.
*   Accepted October 30, 2024.


*   © 2024, Posted by Cold Spring Harbor Laboratory

The copyright holder for this pre-print is the author. All rights reserved. The material may not be redistributed, re-used or adapted without the author's permission.

## References

1.  1.Hennekens CH, Hennekens AR, Hollar D, Casey DE. Schizophrenia and increased risks of cardiovascular disease. Am Heart J. 2005;150(6):1115–1121.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.ahj.2005.02.007&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=16338246&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000234485000002&link_type=ISI) 

2.  2.Saha S, Chant D, McGrath J. A systematic review of mortality in schizophrenia: is the differential mortality gap worsening over time? Arch Gen Psychiatry. 2007;64(10):1123–1131.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/archpsyc.64.10.1123&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17909124&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000249844500003&link_type=ISI) 

3.  3.Chan JKN, Correll CU, Wong CSM, et al. Life expectancy and years of potential life lost in people with mental disorders: a systematic review and meta-analysis. EClinicalMedicine. 2023;65:102294.
    
    

4.  4.Correll CU, Solmi M, Croatto G, et al. Mortality in people with schizophrenia: a systematic review and meta-analysis of relative risk and aggravating or attenuating factors. World Psychiatry. 2022;21(2):248–271.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35524619&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

5.  5.Richmond-Rakerd LS, D’Souza S, Milne BJ, Caspi A, Moffitt TE. Longitudinal Associations of Mental Disorders With Physical Diseases and Mortality Among 2.3 Million New Zealand Citizens. JAMA Netw Open. 2021;4(1):e2033448.
    
    

6.  6.Halstead S, Cao C, Høgnason Mohr G, et al. Prevalence of multimorbidity in people with and without severe mental illness: a systematic review and meta-analysis. Lancet Psychiatry. 2024;11(6):431–442.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38642560&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

7.  7.Pizzol D, Trott M, Butler L, et al. Relationship between severe mental illness and physical multimorbidity: a meta-analysis and call for action. BMJ Ment Health. 2023;26(1).
    
    

8.  8.Lu C, Jin D, Palmer N, et al. Large-scale real-world data analysis identifies comorbidity patterns in schizophrenia. Transl Psychiatry. 2022;12(1):154.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35410453&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

9.  9.Solmi M, Croatto G, Fornaro M, et al. Regional differences in mortality risk and in attenuating or aggravating factors in schizophrenia: A systematic review and meta-analysis. Eur Neuropsychopharmacol. 2024;80:55–69.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38368796&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

10. 10.Upthegrove R,  Khandaker GM. Cytokines, Oxidative Stress and Cellular Markers of Inflammation in Schizophrenia. Curr Top Behav Neurosci. 2020;44:49–66.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access\_num=10.1007/7854_2018_88&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31115797&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

11. 11.Hartwig FP, Borges MC, Horta BL, Bowden J, Davey Smith G. Inflammatory Biomarkers and Risk of Schizophrenia: A 2-Sample Mendelian Randomization Study. JAMA Psychiatry. 2017;74(12):1226–1233.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=29094161&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

12. 12.Campeau A, Mills RH, Stevens T, et al. Multi-omics of human plasma reveals molecular features of dysregulated inflammation and accelerated aging in schizophrenia. Mol Psychiatry. 2022;27(2):1217–1225.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34741130&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

13. 13.Higgins-Chen AT, Boks MP, Vinkers CH, Kahn RS, Levine ME. Schizophrenia and Epigenetic Aging Biomarkers: Increased Mortality, Reduced Cancer Risk, and Unique Clozapine Effects. Biol Psychiatry. 2020;88(3):224–235.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32199607&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

14. 14.Voisey J, Lawford BR, Morris CP, et al. Epigenetic analysis confirms no accelerated brain aging in schizophrenia. NPJ Schizophr. 2017;3(1):26.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28871179&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

15. 15.Oblak L, van der Zaag J, Higgins-Chen AT, Levine ME, Boks MP. A systematic review of biological, social and environmental factors associated with epigenetic clock acceleration. Ageing Res Rev. 2021;69:101348.
    
    

16. 16.Harvanek ZM, Boks MP, Vinkers CH, Higgins-Chen AT. The Cutting Edge of Epigenetic Clocks: In Search of Mechanisms Linking Aging and Mental Health. Biol Psychiatry. 2023;94(9):694–705.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.biopsych.2023.02.001&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36764569&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

17. 17.McKinney BC, Lin H, Ding Y, Lewis DA, Sweet RA. DNA methylation age is not accelerated in brain or blood of subjects with schizophrenia. Schizophr Res. 2018;196:39–44.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28988914&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

18. 18.Caspi A, Shireby G, Mill J, Moffitt TE, Sugden K, Hannon E. Accelerated Pace of Aging in Schizophrenia: Five Case-Control Studies. Biol Psychiatry. 2024;95(11):1038–1047.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37924924&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

19. 19.Sluiskes MH, Goeman JJ, Beekman M, Slagboom PE, Putter H, Rodríguez-Girondo M. Clarifying the biological and statistical assumptions of cross-sectional biological age predictors: an elaborate illustration using synthetic and real data. BMC Med Res Methodol. 2024;24(1):58.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38459475&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

20. 20.Lu AT, Quach A, Wilson JG, et al. DNA methylation GrimAge strongly predicts lifespan and healthspan. Aging (Albany NY). 2019;11(2):303–327.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.18632/aging.101684&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30669119&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

21. 21.Nelson PG, Promislow DEL, Masel J. Biomarkers for Aging Identified in Cross-sectional Studies Tend to Be Non-causative. J Gerontol A Biol Sci Med Sci. 2020;75(3):466–472.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1093/gerona/glz174&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31353411&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

22. 22.Schaie KW. Age changes and age differences. Gerontologist. 1967;7(2):128–132.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access\_num=10.1093/geront/7.2_Part_1.128&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=6041912&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

23. 23.Sehgal R, Markov Y, Qin C, et al. Systems Age: A single blood methylation test to quantify aging heterogeneity across 11 physiological systems. bioRxiv. 2024:2023.2007.2013.548904.
    
    

24. 24.Ying K, Liu H, Tarkhov AE, et al. Causality-enriched epigenetic age uncouples damage and adaptation. Nat Aging. 2024;4(2):231–246.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38243142&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

25. 25.Chen Q, Dwaraka VB, Carreras-Gallo N, et al. OMICmAge: An integrative multi-omics approach to quantify biological age with electronic medical records. bioRxiv. 2023:2023.2010.2016.562114.
    
    

26. 26.Tomusiak A, Floro A, Tiwari R, et al. Development of an epigenetic clock resistant to changes in immune cell composition. Commun Biol. 2024;7(1):934.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=39095531&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

27. 27.Ndhlovu LC, Bendall ML, Dwaraka V, et al. Retro-age: A unique epigenetic biomarker of aging captured by DNA methylation states of retroelements. Aging Cell. 2024:e14288.
    
    

28. 28.Boks MP, Houtepen LC, Xu Z, et al. Genetic vulnerability to DUSP22 promoter hypermethylation is involved in the relation between in utero famine exposure and schizophrenia. NPJ Schizophr. 2018;4(1):16.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41537-018-0058-4&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=30131491&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

29. 29.Hannon E, Dempster EL, Mansell G, et al. DNA methylation meta-analysis reveals cellular alterations in psychosis and markers of treatment-resistant schizophrenia. Elife. 2021;10.
    
    

30. 30.Thrush KL, Higgins-Chen AT, Liu Z, Levine ME. R methylCIPHER: A Methylation Clock Investigational Package for Hypothesis-Driven Evaluation & Research. bioRxiv. 2022:2022.2007.2013.499978.
    
    

31. 31.Higgins-Chen AT, Thrush KL, Wang Y, et al. A computational solution for bolstering reliability of epigenetic clocks: implications for clinical trials and longitudinal tracking. Nature Aging. 2022;2(7):644–661.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=36277076&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

32. 32.Chen Q, Dwaraka VB, Carreras-Gallo N, et al. OMICmAge: An integrative multi-omics approach to quantify biological age with electronic medical records. bioRxiv. 2023.
    
    

33. 33.Teschendorff AE, Breeze CE, Zheng SC, Beck S. A comparison of reference-based algorithms for correcting cell-type heterogeneity in Epigenome-Wide Association Studies. BMC Bioinformatics. 2017;18(1):105.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/s12859-017-1511-5&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28193155&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

34. 34.R: A Language and Environment for Statistical Computing [computer program]. R Foundation for Statistical Computing; 2020.
    
    

35. 35.Viechtbauer W. Conducting Meta-Analyses in R with the metafor Package. Journal of Statistical Software. 2010;36(3):1–48.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.18637/jss.v036.i12&link_type=DOI) 

36. 36.Wickham H. ggplot2 : Elegant Graphics for Data Analysis. Use R!,. 2nd ed. Cham: Springer International Publishing : Imprint: Springer,; 2016.
    
    

37. 37.Chen BH, Marioni RE, Colicino E, et al. DNA methylation-based measures of biological age: meta- analysis predicting time to death. Aging (Albany NY). 2016;8(9):1844–1865.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=27690265&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

38. 38.Tian YE, Di Biase MA, Mosley PE, et al. Evaluation of Brain-Body Health in Individuals With Common Neuropsychiatric Disorders. JAMA Psychiatry. 2023;80(6):567–576.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1001/jamapsychiatry.2023.0791&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37099313&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

39. 39.Rodrigues M, Wiener JC, Stranges S, Ryan BL, Anderson KK. The risk of physical multimorbidity in people with psychotic disorders: A systematic review and meta-analysis. J Psychosom Res. 2021;140:110315.
    
    

40. 40.de Leon J, Diaz FJ. A meta-analysis of worldwide studies demonstrates an association between schizophrenia and tobacco smoking behaviors. Schizophr Res. 2005;76(2-3):135–157.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.schres.2005.02.010&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15949648&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000230119000001&link_type=ISI) 

41. 41.Chow RTS, Whiting D, Favril L, Ostinelli E, Cipriani A, Fazel S. An umbrella review of adverse effects associated with antipsychotic medications: the need for complementary study designs. Neurosci Biobehav Rev. 2023;155:105454.
    
    

42. 42.Gillespie AL, Walker EM, Hannon E, et al. Longitudinal changes in DNA methylation associated with clozapine use in treatment-resistant schizophrenia from two international cohorts. Transl Psychiatry. 2024;14(1):390.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=39333502&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

43. 43.Luykx JJ, Gonzalez-Diaz JM, Guu TW, et al. An international research agenda for clozapine- resistant schizophrenia. Lancet Psychiatry. 2023;10(8):644–652.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37329895&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

44. 44.Solmi M, Tiihonen J, Lähteenvuo M, Tanskanen A, Correll CU, Taipale H. Antipsychotics Use Is Associated With Greater Adherence to Cardiometabolic Medications in Patients With Schizophrenia: Results From a Nationwide, Within-subject Design Study. Schizophr Bull. 2022;48(1):166–175.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34286338&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

45. 45.Conole ELS, Stevenson AJ, Muñoz Maniega S, et al. DNA Methylation and Protein Markers of Chronic Inflammation and Their Associations With Brain and Cognitive Aging. Neurology. 2021;97(23):e2340–e2352.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1212/WNL.0000000000012997&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34789543&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

46. 46.Jiang Y, Cai Y, Lu Y, Wu G, Zhang XY. Relationship between anemia and its correlates and cognitive function in Chinese patients with chronic schizophrenia: A large cross-sectional study. Schizophr Res Cogn. 2024;36:100300.
    
    

47. 47.Xu H, Zheng L, Wang L, Gao H, Wei Y, Chen J. Albumin and Associated Biomarkers in Severe Neuropsychiatric Disorders: Acute-Phase Schizophrenia and Bipolar Disorder. Neuropsychiatr Dis Treat. 2023;19:2027–2037.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=37790800&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

48. 48.Hao M, Jiang S, Tang J, et al. Ratio of Red Blood Cell Distribution Width to Albumin Level and Risk of Mortality. JAMA Netw Open. 2024;7(5):e2413213.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38805227&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

49. 49.Khan S. IGFBP-2 Signaling in the Brain: From Brain Development to Higher Order Brain Functions. Front Endocrinol (Lausanne). 2019;10:822.
    
    

50. 50.Nolin MA, Demers MF, Rauzier C, et al. Circulating IGFBP-2 levels reveal atherogenic metabolic risk in schizophrenic patients using atypical antipsychotics. World J Biol Psychiatry. 2021;22(3):175–182.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=32552257&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

51. 51.Shi M, Bazzano LA, He J, et al. Novel serum metabolites associate with cognition phenotypes among Bogalusa Heart Study participants. Aging (Albany NY). 2019;11(14):5124–5139.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=31327759&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

52. 52.Zong X, Fan Q, Yang Q, Pan R, Zhuang L, Tao R. Phenylacetylglutamine as a risk factor and prognostic indicator of heart failure. ESC Heart Fail. 2022;9(4):2645–2653.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=35624536&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

53. 53.Liu G, Nguyen NQH, Wong KE, et al. Metabolomic Association and Risk Prediction With Heart Failure in Older Adults. Circ Heart Fail. 2024;17(3):e010896.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38426319&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

54. 54.Claesen K, Mertens JC, Leenaerts D, Hendriks D. Carboxypeptidase U (CPU, TAFIa, CPB2) in Thromboembolic Disease: What Do We Know Three Decades after Its Discovery? Int J Mol Sci. 2021;22(2).
    
    

55. 55.Contrepois K, Mann C, Fenaille F. H2B Type 1-K Accumulates in Senescent Fibroblasts with Persistent DNA Damage along with Methylated and Phosphorylated Forms of HMGA1. Proteomes. 2021;9(2).
    
    

56. 56.Cao HM, Ye XP, Ma JH, et al. Mimecan, a Hormone Abundantly Expressed in Adipose Tissue, Reduced Food Intake Independently of Leptin Signaling. EBioMedicine. 2015;2(11):1718–1724.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=26870797&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

57. 57.Oh HS, Rutledge J, Nachun D, et al. Organ aging signatures in the plasma proteome track health and disease. Nature. 2023;624(7990):164-172.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/s41586-023-06802-1&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=38057571&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 

58. 58.Ahn J, Suh Y, Lee K. Chordin-like 1, a Novel Adipokine, Markedly Promotes Adipogenesis and Lipid Accumulation. Cells. 2023;12(4).
    
    

59. 59.Yang J, Chen T, Sun L, et al. Potential metabolite markers of schizophrenia. Mol Psychiatry. 2013;18(1):67–78.
    
    [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/mp.2011.131&link_type=DOI) 
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22024767&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom) 
    
    [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000312771600013&link_type=ISI) 

60. 60.Gunasekara CJ, Hannon E, MacKay H, et al. A machine learning case-control classifier for schizophrenia based on DNA methylation in blood. Transl Psychiatry. 2021;11(1):412.
    
    [PubMed](http://medrxiv.org/lookup/external-ref?access_num=34341337&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F10%2F30%2F2024.10.28.24316295.atom)