Hostname: page-component-78c5997874-94fs2 Total loading time: 0 Render date: 2024-11-16T17:12:21.623Z Has data issue: false hasContentIssue false

Psychosis risk for lesbian, gay, and bisexual individuals: systematic review and meta-analysis

Published online by Cambridge University Press:  25 September 2024

Jean-Paul Selten*
Affiliation:
Department of Psychiatry and Neuropsychology, Faculty of Health, Medicine, and Life Sciences, Mental Health and Neuroscience Research Institute, Maastricht University, Maastricht, The Netherlands
Hussam Alrashed
Affiliation:
Faculty of Life Sciences & Medicine, King's College London, London, UK St. John's Institute of Dermatology, Guy's and St Thomas’ Hospital NHS Foundation Trust, London, UK
Hans Oh
Affiliation:
Suzanne Dworak Peck School of Social Work, University of Southern California, Los Angeles, CA, USA
Gabriëlla A. M. Blokland
Affiliation:
Department of Psychiatry and Neuropsychology, Faculty of Health, Medicine, and Life Sciences, Mental Health and Neuroscience Research Institute, Maastricht University, Maastricht, The Netherlands
*
Corresponding author: Jean-Paul Selten; Email: [email protected]
Rights & Permissions [Opens in a new window]

Abstract

The social defeat hypothesis posits that low status and repeated humiliation increase the risk for psychotic disorders (PDs) and psychotic experiences (PEs). The purpose of this paper was to provide a systematic review of studies on risk of PDs and PEs among lesbian, gay, or bisexual (LGB) people and a quantitative synthesis of any difference in risk. PubMed, PsycINFO, Embase, and Web of Science were searched from database inception until January 30, 2024. Two independent reviewers assessed the eligibility and quality of studies, extracted effect sizes, and noted the results of mediation analyses. Using a random effects model we computed pooled odds ratios (ORs). Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines were followed. The search identified seven studies of PDs and six of PEs. As for PDs, the unadjusted (2.13; 95% confidence interval 0.72–6.34) and covariate-adjusted pooled OR (2.24; 1.72–3.53) were not significantly increased for LGB individuals. After exclusion of a study of limited quality, both the unadjusted pooled OR (2.77; 1.21–6.32) and the covariate-adjusted pooled OR (2.67; 1.53–4.66) were significantly increased. The pooled ORs were increased for PEs: unadjusted, pooled OR = 1.97 (1.47–2.63), covariate-adjusted, pooled OR = 1.85 (1.50–2.28). Studies of PE that examined the mediating role of several variables reported that the contribution of drug abuse was small compared to that of psychosocial stressors. The results of a study in adolescents suggested a protective effect of parental support. These findings suggest an increased psychosis risk for LGB people and support the social defeat hypothesis.

Type
Review Article
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
Copyright © The Author(s), 2024. Published by Cambridge University Press

Introduction

In a recent survey of 30 countries, approximately 9% of respondents identified as lesbian, gay, or bisexual (LGB+), though prevalence varied across countries (https://www.ipsos.com/en/pride-month-2023-9-of-adults-identify-as-lgbt). However, sexual orientation is difficult to estimate, especially in countries where anything other than heterosexuality is prohibited (Pachankis & Bränström, Reference Pachankis and Bränström2019). Furthermore, reports about sexual identity, sexual preference, and sexual behavior often do not align (Patela et al., Reference Patela, Hajat, Schillinger, Blank, Sell and Mostashari2006). There can be no doubt, however, that homophobia and heterosexism cause a lot of stress among LGB people.

People exposed to social stressors are at an increased risk of developing mental health problems and an increasing body of evidence indicates that these stressors are also important in the etiology of schizophrenia spectrum or psychotic disorders (PDs) (e.g. Kirkbride et al., Reference Kirkbride, Anglin, Colman, Dykxhoorn, Jones, Patalay and Griffiths2024; Selten & Cantor-Graae, Reference Selten and Cantor-Graae2005; Selten & Ormel, Reference Selten and Ormel2023). This evidence includes various research findings, including increased risks for subjects with a childhood trauma (Varese et al., Reference Varese, Smeets, Drukker, Lieverse, Lataster, Viechtbauer and Bentall2012), the presence of any psychiatric disorder in childhood or adolescence (Maibing et al., Reference Maibing, Pedersen, Benros, Mortensen, Dalsgaard and Nordentoft2015), a low IQ (Khandaker, Barnett, White, & Jones, Reference Khandaker, Barnett, White and Jones2011), a disadvantaged ethnic minority status (Mirza et al., Reference Mirza, Taft, He, Gooding, Dingwall and Nagel2022; Petrović-van der Deen et al., Reference Petrović-van der Deen, Cunningham, Manuel, Gibb, Porter, Pitama and Lacey2020; van der Ven et al., Reference van der Ven, Olino, Diehl, Nuñez, Thayer, Bridgwater and Anglin2024), a hearing impairment (Linszen, Brouwer, Heringa, & Sommer, Reference Linszen, Brouwer, Heringa and Sommer2016), or gender dysphoria (e.g. Termorshuizen, de Vries, Wiepjes, & Selten, Reference Termorshuizen, de Vries, Wiepjes and Selten2023).

According to the social defeat hypothesis of psychosis the combination of low status, repeated experiences of humiliation, and a poor homeostatic control of dopamine neurons in the midbrain and dorsal striatum lead to increased striatal dopamine activity and thereby to an increased risk of psychosis (Selten & Ormel, Reference Selten and Ormel2023). Psychosis has been conceptualized as a continuum between normality and PD, such that about 7% of the adult population has psychotic experiences (PEs), without necessarily crossing the clinical threshold for PD (Linscott & van Os, Reference Linscott and van Os2013). Given that risk factors for PD overlap with those for PEs, the social defeat hypothesis would also lead us to expect an increased risk of PEs for sexual minorities.

In this study, we focus on LGB people. The aims of this study were to provide (i) a systematic review of studies on risk of PD and (clinical or non-clinical) PEs among LGB people; and (ii) a quantitative synthesis of any difference in risk.

Methods

This systematic review and meta-analysis were conducted following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines (Page et al., Reference Page, McKenzie, Bossuyt, Boutron, Hoffmann, Mulrow and Moher2021). The protocol was prospectively registered on August 26th, 2022 in PROSPERO (CRD42022354853). The final search was performed on January 29th, 2024.

Eligibility

For inclusion in this systematic review, studies were required to (i) compare the presence of PD or PE among homosexual or bisexual individuals relative to heterosexual individuals; (ii) to provide either the odds ratios (ORs) or relative risks (RRs), or the numerators and denominators that would allow for the calculation of such effects. We excluded publications (i) that only reported figures for groups comprising transgender or queer individuals; and (ii) that reported risk for sub-groups of homosexual or bisexual individuals based on organic factors. Examples are samples characterized by the use of illicit drugs or infection with the human immunodeficiency virus.

Search strategy and data extraction

A detailed description of our search strategy is available in the online Supplementary Methods. We conducted the systematic review using PubMed, Embase, PsycINFO, and Web of Science. The following information was extracted from each article: (1) authors, year of publication; (2) country where the study was performed; (3) age, sex, and ethnicity of study population; (4) sexual orientation (heterosexual, homosexual, or bisexual); (5) additional details about sexual orientation (orientation, identity, partnership, behavior/activity, attraction); (6) numbers of psychotic and non-psychotic individuals in groups of LGBs and heterosexuals or general population; (7) classification system, diagnostic interview, or questionnaire used for assessment of psychosis (PD or PE); (8) diagnosis (schizophrenia, mood disorder with psychotic features, PE, etc.); (9) OR, RR, hazard ratio, incidence rate ratio for homosexual and/or bisexual individuals relative to heterosexual individuals; (10) unadjusted and, when available, covariate-adjusted effect estimate (with 95% confidence interval [CI] or standard error (s.e.) accordingly); (11) covariates (age, sex, or other variables).

Two authors (HA and J-PS, or HO and J-PS) independently assessed 401 abstracts of studies for inclusion. The reliability of this decision was fair (Cohen's kappa's 0.44 and 0.69, respectively). HA and J-PS independently reviewed the full text of 20 remaining papers, the reliability of the decision to include or exclude was good (Cohen's kappa = 0.88) (see Fig. 1).

Figure 1. Flow diagram of study selection process.

Three authors extracted effect sizes (HA, HO, J-PS). Any differences were discussed in the presence of the last author (GAMB or HO) and agreement was reached through consultation.

Quality assessment

Since the literature on LGB psychosis risk is still inchoate, we developed our own instrument to assess the quality of each study (see online Supplementary Methods). The quality criteria for studies on PD covered the validity of context (e.g. community survey, case register), sample size, method for assessment of sexual orientation and of PD, classification of PD, inter-rater reliability, and several aspects of statistical analysis. For studies on PEs, the criteria were validity of context (e.g. general population survey, section of population), sample size, method for assessment of sexual orientation, quality of psychosis questionnaire, and statistical analysis.

To check whether the scores obtained correspond to those based on a generally accepted instrument, we also reviewed the studies using the quality assessment tool for observational cohort and cross-sectional studies, developed by the National Institute of Health (NIH; https://www.nhlbi.nih.gov/health-topics/study-quality-assessment-tools).

Two authors (not blind to the identity of the authors) independently assessed the quality of the articles (HA and J-PS). Differences were discussed and reconciled with the opinion of a third author (GAMB or HO) to arrive at a final score.

Statistical analyses

All statistical analyses were performed using R version 4.3.1 (R Core Team, 2022). If ORs were not reported by the study, then they were calculated using the R package ‘metafor’ (Viechtbauer, Reference Viechtbauer2010).

Meta-analyses

Meta-analyses were performed if there were at least two independent cohorts that yielded ORs. Pooled ORs (pooled ORRE) were estimated using the R package ‘metafor’ v4.2.0 (Viechtbauer, Reference Viechtbauer2010). In view of the large differences between the pertinent investigations we considered Random Effects (RE) models. RE estimates were based on the restricted maximum-likelihood method. s.e.s and CIs were calculated as suggested by Knapp and Hartung (Reference Knapp and Hartung2003), because this procedure has much more appropriate false-positive rates than the standard approach. In addition to the standard reported p values, p values were also calculated using a permutation test approach based on Monte Carlo simulation (1000 permutations), which results in more accurate p values compared with standard methods, particularly if the number of studies in a model is small.

The proportion of effect size variation that can be attributed to differences between studies (heterogeneity) was calculated using Cochran's Q-test and presented using I 2. In addition to the pooled ORs and 95% CIs, the number of included effect sizes (k) is reported for each meta-analytic comparison.

Prior to the overall meta-analyses, the covariate-adjusted OR estimates of Bolton and Sareen (Reference Bolton and Sareen2011) were meta-analyzed to generate an overall estimate for males and females combined, and homosexual and bisexual individuals combined, as this study only reported within-subgroup estimates. Likewise, as Lu, Qing, Tu, and Liu (Reference Lu, Qing, Tu and Liu2023) reported only separate results for homosexuals and bisexuals, we conducted a meta-analysis to produce an overall effect estimate for LGBs.

Two studies did not identify a single homosexual or bisexual individual with a PD (Currier et al., Reference Currier, Brown, Walsh, Jager-Hyman, Chaudhury and Stanley2015; Skerrett, Kõlves, & De Leo, Reference Skerrett, Kõlves and De Leo2014). It is difficult, then, to compute an OR because one cannot divide by zero. Thus, to obtain an OR, we replaced zero with one. The covariate-unadjusted OR thus calculated was used for the meta-analysis of both covariate-unadjusted and covariate-adjusted pooled ORs.

Two studies reported results for sexual minorities (Oh, Reference Oh2021; Oh et al., Reference Oh, Koyanagi, Leaune, Zhou, Kelleher and DeVylder2022). Since this concept includes individuals other than LGB people, separate results for LGBs were obtained from the first author and used here.

Two studies, one on PDs and another on PEs, used two different, but overlapping definitions of LGB status, namely sexual orientation and sexual activity (Gevonden et al., Reference Gevonden, Selten, Myin-Germeys, de Graaf, ten Have, van Dorsselaer and Veling2014; Qi, Palmier-Claus, Simpson, Varese, & Bentall, Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020). For the overall meta-analysis we used the definition which was associated with the largest number of individuals (i.e. sexual orientation). One study (Post, Veling, & GROUP Investigators, Reference Post, Veling and GROUP Investigators2021) made a mistake in the calculation of the ORs. We contacted the senior author, who corroborated the error and provided the corrected OR. (The pertinent e-mail is shown in other online Supplementary materials).

Meta-analyses were performed using unadjusted ORs and covariate-adjusted ORs, which resulted in unadjusted pooled ORs and in covariate-adjusted pooled ORs, respectively.

Additional analyses

We conducted additional meta-analyses for (i) homosexual individuals v. bisexual individuals and (ii) males v. females, if at least two studies reported data for these subgroups.

Likewise, if at least two studies reported data, we conducted additional meta-analyses for (i) hallucinations and (ii) delusions.

Publication bias

Publication bias was tested by visually assessing the funnel plot for asymmetry and by performing Egger's test (Egger, Davey Smith, Schneider, & Minder, Reference Egger, Davey Smith, Schneider and Minder1997) using the R package ‘metafor’ (Schwarzer, Carpenter, & Rucker, Reference Schwarzer, Carpenter and Rucker2015; Viechtbauer, Reference Viechtbauer2010). If there were indications for publication bias, the trim-and-fill method (Duval, Reference Duval, Rothsteein, Sutton and Borenstein2005; Duval & Tweedie, Reference Duval and Tweedie2000a; Reference Duval and Tweedie2000b) was used to calculate corrected ORs. In these cases, data from the trim-and-fill method are presented in the text along with the uncorrected estimates, while the presented figures contain the original uncorrected data only.

Results

Study selection

A total of 402 studies were identified in the initial search. One duplicate article was removed and 381 articles were removed after screening titles and abstracts. Seven articles were removed after a full-text review (see Fig. 1). A total of 13 publications, describing 12 unique cohorts, were included: see Table 1 and online Supplementary Table S1 for the details of each study.

Table 1. Characteristics of studies that investigated the relationship between LGB status and risk of PDs or PEs

Notes: AUS, Australia; L, G, or B, lesbian, gay, or bisexual; N.A., not applicable; NLD, The Netherlands.

a Sample sizes per group are reported in online Supplementary Table S1.

b Covariate-unadjusted OR and 95% CI.

c Computed by the authors of this meta-analysis.

d According to International Classification of Diseases, 10th edition.

e OR computed by authors of this meta-analysis by changing the number of psychotic cases in the LGB group from 0 to 1.

f According to Diagnostic and Statistical Manual of Mental Disorders, 4th edition.

g Not applicable. The study did not use dichotomous measures.

All the studies were cross-sectional. Seven studies reported comparisons for PD (Bolton & Sareen, Reference Bolton and Sareen2011; Borgogna, Aita, Trask, & Moncrief, Reference Borgogna, Aita, Trask and Moncrief2022; Chakraborty, McManus, Brugha, Bebbington, & King, Reference Chakraborty, McManus, Brugha, Bebbington and King2011; Currier et al., Reference Currier, Brown, Walsh, Jager-Hyman, Chaudhury and Stanley2015; Post et al., Reference Post, Veling and GROUP Investigators2021; Qi et al., Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020; Skerrett et al., Reference Skerrett, Kõlves and De Leo2014) and six assessed PE (Gevonden et al., Reference Gevonden, Selten, Myin-Germeys, de Graaf, ten Have, van Dorsselaer and Veling2014; Jacob et al., Reference Jacob, Smith, McDermott, Haro, Stickley and Koyanagi2021; Lu et al., Reference Lu, Qing, Tu and Liu2023; Oh, Reference Oh2021; Oh et al., Reference Oh, Koyanagi, Leaune, Zhou, Kelleher and DeVylder2022; Pérez-Albéniz, Lucas-Molina, & Fonseca-Pedrero, Reference Pérez-Albéniz, Lucas-Molina and Fonseca-Pedrero2023). The studies had been conducted in the USA (n = 5), the UK (n = 3), The Netherlands (n = 2), Spain (n = 1), China (n = 1), and Australia (n = 1), countries where homosexuality is legal.

Six studies collected their data from general population samples. The remaining studies included patients diagnosed with a PD (Post et al., Reference Post, Veling and GROUP Investigators2021), college students (Borgogna et al., Reference Borgogna, Aita, Trask and Moncrief2022; Lu et al., Reference Lu, Qing, Tu and Liu2023; Oh et al., Reference Oh, Koyanagi, Leaune, Zhou, Kelleher and DeVylder2022), secondary school students aged 14–18 years (Pérez-Albéniz et al., Reference Pérez-Albéniz, Lucas-Molina and Fonseca-Pedrero2023), individuals who underwent a standardized suicide risk assessment at a psychiatric emergency department (Currier et al., Reference Currier, Brown, Walsh, Jager-Hyman, Chaudhury and Stanley2015), or individuals who died by suicide (Skerrett et al., Reference Skerrett, Kõlves and De Leo2014).

While all of the studies reported on sexual orientation, four studies also provided information on sexual activity, defined as ‘sexual intercourse’, ‘sexual experience’ (Chakraborty et al., Reference Chakraborty, McManus, Brugha, Bebbington and King2011; Qi et al., Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020), ‘touching the genitals of another person in the preceding year’ (Gevonden et al., Reference Gevonden, Selten, Myin-Germeys, de Graaf, ten Have, van Dorsselaer and Veling2014), or ‘sexual experiences in the last 12 months’ (Oh, Reference Oh2021). Questions about PEs concerned the past 12 months (Jacob et al., Reference Jacob, Smith, McDermott, Haro, Stickley and Koyanagi2021; Oh et al., Reference Oh, Koyanagi, Leaune, Zhou, Kelleher and DeVylder2022) or the whole lifetime (Gevonden et al., Reference Gevonden, Selten, Myin-Germeys, de Graaf, ten Have, van Dorsselaer and Veling2014; Lu et al., Reference Lu, Qing, Tu and Liu2023; Oh, Reference Oh2021).

Since the corresponding author of the study conducted in Spain did not respond to our request for group sizes, ORs, or RRs, we could not include the results in our meta-analysis. The study found that adolescents with a homosexual or bisexual orientation reported significantly more symptoms (about 57% more) on the Prodromal Questionnaire-Brief (Loewy, Pearson, Vinogradov, Bearden, & Cannon, Reference Loewy, Pearson, Vinogradov, Bearden and Cannon2011) than their heterosexual peers.

Two studies derived different ORs of probable PD from the Adult Psychiatric Morbidity Survey in the UK (Chakraborty et al., Reference Chakraborty, McManus, Brugha, Bebbington and King2011; Qi et al., Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020). Qi et al.'s (Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020) investigation differed from that of Chakraborty et al. (Reference Chakraborty, McManus, Brugha, Bebbington and King2011) in that it excluded participants about whom information was obtained using proxy interviews and several individuals with missing data. Since it was not evident which result was more valid, we used the study that yielded the lowest OR (Qi et al., Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020).

Quality assessment

The quality assessment results of our instrument were strongly correlated with the tool developed by the NIH (r = 0.68; p < 0.01; online Supplementary Fig. S1 and Tables S2 and S3), supporting the validity of our instrument. The results showed that one study, which examined the number of LGB people in Queensland, Australia, who had committed suicide, was of a distinctly lower quality (Skerrett et al., Reference Skerrett, Kõlves and De Leo2014). The researchers identified only 34 LGBs among 5999 suicide victims (0.57%) and no LGB individual with a diagnosis of a PD. As the authors acknowledged, the study was of limited quality, because the assignment of LGB status was often based on indirect indicators, such as being unmarried and childless, and the proportion of LGB individuals who died by suicide was low. Currier et al.'s (Reference Currier, Brown, Walsh, Jager-Hyman, Chaudhury and Stanley2015) study was of sufficient quality, but its goal was to explore whether patients at an emergency department would disclose their sexual orientation. It screened a small number of subjects (n = 177) and reported no PD among 21 subjects who identified themselves as LGB.

The quality of the remaining five studies on PD was high. The Adult Psychiatric Morbidity Survey in the UK (n = 7403) and the National Epidemiologic Survey on Alcohol and Related Conditions Wave 2 in the USA (n = 34 653) were large investigations of the general population (Bolton & Sareen, Reference Bolton and Sareen2011; Chakraborty et al., Reference Chakraborty, McManus, Brugha, Bebbington and King2011; Qi et al., Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020). Using a dataset of the Healthy Minds Study, Borgogna et al. (Reference Borgogna, Aita, Trask and Moncrief2022) asked 159 students who had reported to have been diagnosed with a PD for their sexual orientation; they then compared the results with those in a control group. A study from the Netherlands reported the proportion of individuals with LGB orientation in a large sample (n = 582) of patients diagnosed with a PD (Post et al., Reference Post, Veling and GROUP Investigators2021).

Among the studies into PEs there was also one that was of lower quality (online Supplementary Fig. S1 and Tables S2 and S3). Lu et al.'s (Reference Lu, Qing, Tu and Liu2023) study was of questionable validity, because the proportion of college students who reported at least one PE was 60.1%. This figure is high if one considers that a systematic review found a median prevalence rate of self-reported PEs of 11.9% (Linscott & van Os, Reference Linscott and van Os2013). It is even higher if one takes into account that Lu et al. dropped 12 items from the analyses because they were very commonly reported. Another problem is that Lu et al. adjusted the results for the quality of relationships with parents, teachers and classmates, while a disturbance in these relationships may contribute to the increased ORs. All studies, except for the Spanish one, reached more than 100 LGB people. One study had a large sample of LGB college students, but the response rate was low (14%) (Oh et al., Reference Oh, Koyanagi, Leaune, Zhou, Kelleher and DeVylder2022).

Meta-analyses

The results of the heterogeneity tests were significant for the covariate-unadjusted ORs derived from all studies on PD, but not for the covariate-adjusted ORs (Fig. 2; online Supplementary Fig. S2A and Table S4). The heterogeneity for the covariate-unadjusted ORs remained significant after exclusion of Skerrett et al.'s paper. As for the studies on PEs, the heterogeneity test was significant for the unadjusted ORs, but not for the adjusted ones. Considering the pros and cons, we decided to use in each analysis an RE model, because the differences among studies were large.

Figure 2. Meta-analysis of studies that examined the relationship between LGB status and risk for PDs.

Given the large variety of covariates for which the results were adjusted, in particular with reference to PD, to ensure comparability, the covariate-unadjusted pooled ORs are presented as the primary result, over the covariate-adjusted pooled ORs.

There was no indication of publication bias in the meta-analyses of covariate-unadjusted ORs from studies on PDs or PEs (online Supplementary Fig. S3 and Table S4). There was an indication (funnel plot asymmetry) of this type of bias in the meta-analysis of covariate-adjusted ORs from all studies on PDs (not after exclusion of Skerrett et al.'s paper), but not in the meta-analysis of all covariate-adjusted ORs for PEs. In cases of a significant bias test indicating funnel plot asymmetry, the trim-and-fill corrected effects (ORtf) are reported in the text.

Main findings for PDs

A first meta-analysis, based on all available studies, found that the unadjusted, pooled OR [95% CI] was not significantly increased for LGB people (pooled ORRE = 2.13 [0.72–6.34]; k = 6) (see Fig. 2A). The covariate-adjusted, pooled OR (pooled ORRE = 2.24 [0.90–5.58]; k = 5) was also not significantly increased and did not become significantly increased after adjustment for publication bias: pooled ORRE-tf = 2.75 [0.94–3.99]; k = 5) (online Supplementary Fig. S2A and Table S4).

After excluding Skerrett et al.'s study, both the unadjusted, pooled OR (ORRE = 2.77 [1.21–6.32]; k = 5) and the covariate-adjusted, pooled OR (ORRE = 2.67 [1.53–4.66]; k = 4), were significantly increased (Fig. 2B; online Supplementary Fig. S2B and Table S4). Qi et al.'s (Reference Qi, Palmier-Claus, Simpson, Varese and Bentall2020) study reported a lower OR when LGB status was based on any homosexual activity (OR = 1.55 [0.30–8.07]. When we used this value, the pooled OR remained significantly increased: unadjusted, pooled ORRE = 2.56 [1.02–6.42]; adjusted, pooled ORRE = 2.59 [1.37–4.89].

From two studies it was possible to derive separate effect sizes for males and females (Bolton & Sareen, Reference Bolton and Sareen2011; Post et al., Reference Post, Veling and GROUP Investigators2021). However, the 95% CI for the unadjusted, pooled ORRE for LGB women was extremely wide (ORRE = 2.23 [0.03–195.06]; k = 2) and, understandably, not significantly different from the corresponding figure for LGB men (unadjusted, pooled ORRE for LGB men = 2.29 [0.54–9.78], k = 2; online Supplementary Table S4).

Two studies distinguished in their analyses between homosexual and bisexual individuals (Bolton & Sareen, Reference Bolton and Sareen2011; Borgogna et al., Reference Borgogna, Aita, Trask and Moncrief2022). The 95% CIs for the pertinent ORs were very wide and overlapping (online Supplementary Table S4).

Main findings for PEs

The meta-analysis revealed that LGB people reported more PEs than heterosexuals. Both the unadjusted, pooled OR (ORRE = 1.97 [1.47–2.63]; k = 6) and the covariate-adjusted, pooled OR (ORRE = 1.85 [1.50–2.28]; k = 6; were significantly increased (Fig. 3A; online Supplementary Fig. S4A and Table S4). After exclusion of the study by Lu et al. (Reference Lu, Qing, Tu and Liu2023), the unadjusted, pooled ORRE (2.07 [1.44–2.96]; k = 5) and the covariate-adjusted pooled ORRE (1.85 [1.50–2.28]; k = 5) became slightly higher (Fig. 3B).

Figure 3. Meta-analysis of studies that examined the relationship between LGB status and risk for PEs.

The OR for individuals who reported some kind of homosexual or bisexual activity was more increased (unadjusted, pooled ORRE = 2.42 [0.93–6.29]; k = 3) than the OR for subjects who reported only a homosexual or bisexual orientation (unadjusted pooled ORRE = 1.91 [1.19–3.08]; k = 3) (Figs 3C and 3D; online Supplementary Figs S4C and S4D and Table S3).

Two papers reported separate estimates for individuals with a homosexual or bisexual orientation (Lu et al., Reference Lu, Qing, Tu and Liu2023; Oh et al., Reference Oh, Koyanagi, Leaune, Zhou, Kelleher and DeVylder2022). In both studies the risk for bisexuals was somewhat higher than that for homosexuals. The results of a meta-analysis showed that the OR for bisexuals was slightly more increased (unadjusted, pooled ORRE = 1.73 [1.68–1.78]; k = 2) than the OR for homosexuals (unadjusted pooled ORRE = 1.35 [1.07–1.69]; k = 2), although CIs just overlapped (online Supplementary Table S3).

Only one study reported separate effect sizes for males and females; the 95% CIs were overlapping (Lu et al., Reference Lu, Qing, Tu and Liu2023).

Mediators

One study conducted a mediation analysis on the association between LGB status and risk of PD (Post et al., Reference Post, Veling and GROUP Investigators2021) and three studies did this for PEs (Gevonden et al., Reference Gevonden, Selten, Myin-Germeys, de Graaf, ten Have, van Dorsselaer and Veling2014; Jacob et al., Reference Jacob, Smith, McDermott, Haro, Stickley and Koyanagi2021; Pérez-Albéniz et al., Reference Pérez-Albéniz, Lucas-Molina and Fonseca-Pedrero2023). They reported significant effects of several variables, including bullying victimization (three studies), discrimination (two studies), and childhood trauma (two studies). Interestingly, two studies of PEs examined the mediating role of drug abuse and reported that the contribution of this variable was small compared to that of psychosocial stressors (online Supplementary Table S5). Finally, the researchers in Spain reported a significant negative interaction between LGB orientation and parental support for PEs (Pérez-Albéniz et al., Reference Pérez-Albéniz, Lucas-Molina and Fonseca-Pedrero2023). In other words, the effect of sexual orientation on reporting of PEs was moderated to an important and significant degree by this type of support. The finding suggests a protective effect.

Limitations of the relevant studies

An important goal of a systematic review is to address shortcomings in the literature. We noted several limitations. First, there were only cross-sectional studies and no longitudinal investigations. However, given the large numbers required for a prospective, longitudinal study, such a study seems hardly feasible. Second, most studies were based on the willingness of individuals to report PEs or to report that they had been given a diagnosis of PD by a health professional. A notable exception is the investigation by Post et al., (Reference Post, Veling and GROUP Investigators2021), which was based on a large sample of patients that had already been diagnosed with a (non-affective) PD. Third, of the studies on PD only one distinguished between affective and non-affective PD (Post et al., Reference Post, Veling and GROUP Investigators2021). Thus, it is not possible to conclude about any difference in the risk for affective and non-affective PD. Fourth, some studies identified substance use as a potential mediator and collecting data on substance use is subject to reporting biases (Gevonden et al., Reference Gevonden, Selten, Myin-Germeys, de Graaf, ten Have, van Dorsselaer and Veling2014; Jacob et al., Reference Jacob, Smith, McDermott, Haro, Stickley and Koyanagi2021). Finally, and there is probably no solution to this problem, measuring LGB status was based on self-identification, which may have resulted in social desirability bias due to the stigma of belonging to a sexual minority group.

Discussion

This systematic review retrieved five studies of good quality that compared the risk of PDs for LGB individuals to that for heterosexual individuals and five studies of good quality that did the same regarding PEs. The odds of PD for LGB individuals were significantly increased and about twice that for heterosexual individuals. Their odds of PEs were also significantly increased, by about the same extent.

To our knowledge this is the first systematic review and meta-analysis that compared the risk of psychotic phenomena for the LGB population to that for the heterosexual population. A strength of this paper is the comprehensive literature survey and the critical assessment of the quality of the studies.

Interpretation

There are many indications that LGB people are more likely to experience stigma, prejudice, and discrimination. The minority stress model posits that this leads to a higher risk to develop common mental disorders, like depression and anxiety (Meyer, Reference Meyer2003), but does not mention a higher risk of non-affective PD. For this reason, the social defeat hypothesis complements the minority stress theory and adds greater specificity by positing that low status and repeated humiliation increase the risk for schizophrenia and other non-affective PDs. This line of thought is further supported by recent findings of increased risks of such disorders for persons with gender dysphoria (e.g. Termorshuizen et al., Reference Termorshuizen, de Vries, Wiepjes and Selten2023).

There are several factors consistent with minority stress theory and the social defeat hypothesis that can be explored in future studies to explain the higher risk of LGB individuals relative to heterosexual individuals. Homophobia and heterosexism are strong drivers of health disparities, resulting in higher exposures to stressors (e.g. discrimination, hate crimes, trauma, adverse childhood experiences) and, in many countries, lower access to health promoting resources (e.g. housing, health care). For some LGB individuals, experiences of discrimination may be linked to high-risk behaviors, such as illicit drug use (Caceres et al., Reference Caceres, Brody, Luscombe, Primiano, Marusca, Sitts and Chyun2017; McCabe, Bostwick, Hughes, West, & Boyd, Reference McCabe, Bostwick, Hughes, West and Boyd2010), which can directly impact or combine with other high-risk behaviors (e.g. chemsex behavior; Maxwell, Shahmanesh, & Gafos, Reference Maxwell, Shahmanesh and Gafos2019) to increase risk for psychosis.

Another approach for further study is a search for protective and resilience factors. Pérez-Albéniz et al. (Reference Pérez-Albéniz, Lucas-Molina and Fonseca-Pedrero2023) found that adolescents with homosexual or bisexual orientation who feel supported by their parents reported fewer PEs. This finding is critical and deserves further research.

Limitations

One limitation is that we excluded papers that reported figures for LGBTQ+ individuals more broadly, which included many LGB individuals. It is unlikely, however, that our more stringent exclusion criteria would have significantly influenced our results, since most of these studies on LGBTQ+ broadly defined reported increased risk for sexual minority groups (e.g. Barr, Bigdeli, & Meyers, Reference Barr, Bigdeli and Meyers2022; Savill, Nguyen, Shim, & Loewy, Reference Savill, Nguyen, Shim and Loewy2022). Future research can explore queer orientation/identity with greater nuance and specificity. Second, despite our efforts to include all relevant papers, we may have missed some studies, which is always a possibility when conducting systematic searches. Finally, it is important to note that the terms ‘lesbian’, ‘gay’, and ‘bisexual’ do not adequately capture the complexity of human sexuality, and future research can unpack these terms, and explore the various facets of sexual identity, desire, and behavior more closely.

Conclusion

The results suggest that the risk of developing psychotic symptoms or PDs is increased for LGB individuals compared to heterosexual individuals and underline the importance of combating discrimination. Future studies can further explore the risk and protective factors, while formulating efforts to reduce this disparity.

Supplementary material

The supplementary material for this article can be found at https://doi.org/10.1017/S0033291724002253.

Acknowledgments

This research received no specific grant from any funding agency, commercial, or not-for-profit sectors.

Competing interests

None.

References

Barr, P. B., Bigdeli, T. B., & Meyers, J. L. (2022). Prevalence, comorbidity, and sociodemographic correlates of psychiatric disorders reported in the all of us research program. JAMA Psychiatry, 79(6), 622628. doi:10.1001/jamapsychiatry.2022.0685CrossRefGoogle ScholarPubMed
Bolton, S. L., & Sareen, J. (2011). Sexual orientation and its relation to mental disorders and suicide attempts: Findings from a nationally representative sample. Canadian Journal of Psychiatry, 56(1), 3543. doi:10.1177/070674371105600107CrossRefGoogle ScholarPubMed
Borgogna, N. C., Aita, S. L., Trask, C. L., & Moncrief, G. G. (2022). Psychotic disorders in college students: Demographic and care considerations. Psychosis, 15(3), 229239.CrossRefGoogle Scholar
Caceres, B. A., Brody, A., Luscombe, R. E., Primiano, J. E., Marusca, P., Sitts, E. M., & Chyun, D. (2017). A systematic review of cardiovascular disease in sexual minorities. American Journal of Public Health, 107(4), e13e21. doi:10.2105/AJPH.2016.303630CrossRefGoogle ScholarPubMed
Chakraborty, A., McManus, S., Brugha, T. S., Bebbington, P., & King, M. (2011). Mental health of the non-heterosexual population of England. British Journal of Psychiatry, 198(2), 143148. doi:10.1192/bjp.bp.110.082271CrossRefGoogle ScholarPubMed
Currier, G. W., Brown, G., Walsh, P. G., Jager-Hyman, S., Chaudhury, S., & Stanley, B. (2015). Screening for sexual orientation in psychiatric emergency departments. Western Journal of Emergency Medicine, 16(1), 8084. doi:10.5811/westjem.2014.12.22254CrossRefGoogle ScholarPubMed
Duval, S. J. (2005). The trim and fill method. In Rothsteein, H. R., Sutton, A. J., & Borenstein, M. (Eds.), Publication bias in meta-analysis: Prevention, assessment, and adjustments (pp. 127144). Chichester: Wiley.CrossRefGoogle Scholar
Duval, S. J., & Tweedie, R. L. (2000a). A nonparametric ‘trim and fill’ method of accounting for publication bias in meta-analysis. Journal of the American Statistical Association, 95(449), 8998.Google Scholar
Duval, S. J., & Tweedie, R. L. (2000b). Trim and fill: A simple funnel-plot-based method of testing and adjusting for publication bias in meta-analysis. Biometrics, 56(2), 455463.CrossRefGoogle ScholarPubMed
Egger, M., Davey Smith, G., Schneider, M., & Minder, C. (1997). Bias in meta-analysis detected by a simple, graphical test. BMJ, 315(7109), 629634. doi:10.1136/bmj.315.7109.629CrossRefGoogle ScholarPubMed
Gevonden, M. J., Selten, J. P., Myin-Germeys, I., de Graaf, R., ten Have, M., van Dorsselaer, S., … Veling, W. (2014). Sexual minority status and psychotic symptoms: Findings from the Netherlands Mental Health Survey and Incidence Studies (NEMESIS). Psychological Medicine, 44(2), 421433. doi:10.1017/S0033291713000718CrossRefGoogle ScholarPubMed
Jacob, L., Smith, L., McDermott, D., Haro, J. M., Stickley, A., & Koyanagi, A. (2021). Relationship between sexual orientation and psychotic experiences in the general population in England. Psychological Medicine, 51(1), 138146. doi:10.1017/S003329171900309XCrossRefGoogle ScholarPubMed
Khandaker, G. M., Barnett, J. H., White, I. R., & Jones, P. B. (2011). A quantitative meta-analysis of population-based studies of premorbid intelligence and schizophrenia. Schizophrenia Research, 132(2–3), 220227. doi:10.1016/j.schres.2011.06.017CrossRefGoogle ScholarPubMed
Kirkbride, J. B., Anglin, D. M., Colman, I., Dykxhoorn, J., Jones, P. B., Patalay, P., … Griffiths, S. L. (2024). The social determinants of mental health and disorder: Evidence, prevention and recommendations. World Psychiatry, 23(1), 5890. doi:10.1002/wps.21160CrossRefGoogle ScholarPubMed
Knapp, G., & Hartung, J. (2003). Improved tests for a random effects meta-regression with a single covariate. Statistics in Medicine, 22(17), 26932710. doi:10.1002/sim.1482, PMID: 12939780.CrossRefGoogle ScholarPubMed
Linscott, R. J., & van Os, J. (2013). An updated and conservative systematic review and meta-analysis of epidemiological evidence on psychotic experiences in children and adults: On the pathway from proneness to persistence to dimensional expression across mental disorders. Psychological Medicine, 43(6), 11331149. doi:10.1017/S0033291712001626CrossRefGoogle ScholarPubMed
Linszen, M. M., Brouwer, R. M., Heringa, S. M., & Sommer, I. E. (2016). Increased risk of psychosis in patients with hearing impairment: Review and meta-analyses. Neuroscience Biobehavioral Reviews, 62, 120. doi:10.1016/j.neubiorev.2015.12.012CrossRefGoogle ScholarPubMed
Loewy, R. L., Pearson, R., Vinogradov, S., Bearden, C. E., & Cannon, T. D. (2011). psychosis risk screening with the Prodromal Questionnaire-brief version (PQ-B). Schizophrenia Research, 129(1), 4246. doi:10.1016/j.schres.2011.03.029CrossRefGoogle ScholarPubMed
Lu, D., Qing, Z., Tu, Y., & Liu, X. (2023). Sexual orientation and psychotic-like experiences among Chinese college students: The role of gender. Frontiers in Psychiatry, 14, 1139484. doi:10.3389/fpsyt.2023.1139484CrossRefGoogle ScholarPubMed
Maibing, C. F., Pedersen, C. B., Benros, M. E., Mortensen, P. B., Dalsgaard, S., & Nordentoft, M. (2015). Risk of schizophrenia increases after all child and adolescent psychiatric disorders: A nationwide study. Schizophrenia Bulletin, 41(4), 963970. doi:10.1093/schbul/sbu119CrossRefGoogle ScholarPubMed
Maxwell, S., Shahmanesh, M., & Gafos, M. (2019). Chemsex behaviours among men who have sex with men: A systematic review of the literature. International Journal of Drug Policy, 63, 7489. doi:10.1016/j.drugpo.2018.11.014CrossRefGoogle ScholarPubMed
McCabe, S. E., Bostwick, W. B., Hughes, T. L., West, B. T., & Boyd, C. J. (2010). The relationship between discrimination and substance use disorders among lesbian, gay, and bisexual adults in the United States. American Journal of Public Health, 100(10), 19461952. doi:10.2105/AJPH.2009.163147CrossRefGoogle ScholarPubMed
Meyer, I. H. (2003). Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychological Bulletin, 129(5), 674697. doi:10.1037/0033-2909.129.5.674CrossRefGoogle ScholarPubMed
Mirza, T., Taft, W., He, V. Y., Gooding, J., Dingwall, K., & Nagel, T. (2022). Incidence of treated first-episode psychosis amongst Aboriginal and non-Aboriginal youth in the Top End of the Northern Territory, Australia. Australasian Psychiatry, 30(4), 513517. doi:10.1177/10398562221075193CrossRefGoogle ScholarPubMed
Oh, H. (2021). Psychotic experiences and non-heterosexuality among Latino and Asian Americans. Psychological Medicine, 51(7), 12311232. doi:10.1017/S0033291720000197CrossRefGoogle ScholarPubMed
Oh, H., Koyanagi, A., Leaune, E., Zhou, S., Kelleher, I., & DeVylder, J. E. (2022). Psychotic experiences, sexual minority status, and suicidal behavior among young adult college students in the United States. Social Psychiatry and Psychiatric Epidemiology, 57(9), 19251929. doi:10.1007/s00127-022-02285-zCrossRefGoogle ScholarPubMed
Pachankis, J. E., & Bränström, R. (2019). How many sexual minorities are hidden? Projecting the size of the global closet with implications for policy and public health. PLoS One, 14(6), e0218084. doi:10.1371/journal.pone.0218084CrossRefGoogle ScholarPubMed
Page, M. J., McKenzie, J. E., Bossuyt, P. M., Boutron, I., Hoffmann, T. C., Mulrow, C. D., … Moher, D. (2021). The PRISMA 2020 statement: An updated guideline for reporting systematic reviews. BMJ, 372, n71. doi:10.1136/bmj.n71CrossRefGoogle ScholarPubMed
Patela, P., Hajat, A., Schillinger, J., Blank, S., Sell, R., & Mostashari, F. (2006). Discordance between sexual behavior and self-reported sexual identity: A population-based survey of New York City men. Annals of Internal Medicine, 145(12), 936. doi:10.7326/0003-4819-145-6-200609190-00005, Erratum in: Ann Intern Med. 2006 Dec 19;145(12):936.Google Scholar
Pérez-Albéniz, A., Lucas-Molina, B., & Fonseca-Pedrero, E. (2023). Parental support and gender moderate the relationship between sexual orientation and suicidal behavior in adolescents. Psicothema, 35(3), 248258. doi:10.7334/psicothema2022.325Google ScholarPubMed
Petrović-van der Deen, F. S., Cunningham, R., Manuel, J., Gibb, S., Porter, R. J., Pitama, S., … Lacey, C. (2020). Exploring indigenous ethnic inequities in first episode psychosis in New Zealand – A national cohort study. Schizophrenia Research, 223, 311318. doi:10.1016/j.schres.2020.09.004CrossRefGoogle ScholarPubMed
Post, D., Veling, W., & GROUP Investigators, (2021). Sexual minority status, social adversity and risk for psychotic disorders – Results from the GROUP study. Psychological Medicine, 51(5), 770776. doi:10.1017/S0033291719003726CrossRefGoogle ScholarPubMed
Qi, R., Palmier-Claus, J., Simpson, J., Varese, F., & Bentall, R. (2020). Sexual minority status and symptoms of psychosis: The role of bullying, discrimination, social support, and drug use – Findings from the Adult Psychiatric Morbidity Survey 2007. Psychology and Psychotherapy, 93(3), 503519. doi:10.1111/papt.12242CrossRefGoogle ScholarPubMed
R Core Team (2022). R: A language and environment for statistical computing. Vienna, Austria: R Foundation for statistical computing.Google Scholar
Savill, M., Nguyen, T., Shim, R. S., & Loewy, R. L. (2022). Online psychosis screening: Characterizing an underexamined population to improve access and equity. Psychiatric Services, 73(9), 10051012. doi:10.1176/appi.ps.202100257CrossRefGoogle ScholarPubMed
Schwarzer, G., Carpenter, J. R., & Rucker, G. (2015). Meta-analysis with R (Use-r!). Switzerland: Springer.CrossRefGoogle Scholar
Selten, J. P., & Cantor-Graae, E. (2005). Social defeat: Risk factor for schizophrenia? British Journal of Psychiatry, 187, 101102. doi:10.1192/bjp.187.2.101CrossRefGoogle ScholarPubMed
Selten, J. P., & Ormel, J. (2023). Low status, humiliation, dopamine and risk of schizophrenia. Psychological Medicine, 53(3), 609613. doi:10.1017/S0033291722003816CrossRefGoogle ScholarPubMed
Skerrett, D. M., Kõlves, K., & De Leo, D. (2014). Suicides among lesbian, gay, bisexual, and transgender populations in Australia: An analysis of the Queensland Suicide Register. Asia-Pacific Psychiatry, 6(4), 440446. doi:10.1111/appy.12128CrossRefGoogle ScholarPubMed
Termorshuizen, F., de Vries, A. L. C., Wiepjes, C. M., & Selten, J. P. (2023). The risk of psychosis for transgender individuals: A Dutch national cohort study. Psychological Medicine, 53(16), 79237932. doi:10.1017/S0033291723002088CrossRefGoogle ScholarPubMed
van der Ven, E., Olino, T. M., Diehl, K., Nuñez, S. M., Thayer, G., Bridgwater, M. A., … Anglin, D. M. (2024). Ethnoracial risk variation across the psychosis continuum in the US: A systematic review and meta-analysis. JAMA Psychiatry, 81(5), e235497. doi:10.1001/jamapsychiatry.2023.5497CrossRefGoogle Scholar
Varese, F., Smeets, F., Drukker, M., Lieverse, R., Lataster, T., Viechtbauer, W., … Bentall, R. P. (2012). Childhood adversities increase the risk of psychosis: A meta-analysis of patient-control, prospective- and cross-sectional cohort studies. Schizophrenia Bulletin, 38(4), 661671. doi:10.1093/schbul/sbs050CrossRefGoogle ScholarPubMed
Viechtbauer, W. (2010). Conducting meta-analyses in R with the metafor package. Journal of Statistical Software, 36(3), 148.CrossRefGoogle Scholar
Figure 0

Figure 1. Flow diagram of study selection process.

Figure 1

Table 1. Characteristics of studies that investigated the relationship between LGB status and risk of PDs or PEs

Figure 2

Figure 2. Meta-analysis of studies that examined the relationship between LGB status and risk for PDs.

Figure 3

Figure 3. Meta-analysis of studies that examined the relationship between LGB status and risk for PEs.

Supplementary material: File

Selten et al. supplementary material 1

Selten et al. supplementary material
Download Selten et al. supplementary material 1(File)
File 90.6 KB
Supplementary material: File

Selten et al. supplementary material 2

Selten et al. supplementary material
Download Selten et al. supplementary material 2(File)
File 37.8 KB
Supplementary material: File

Selten et al. supplementary material 3

Selten et al. supplementary material
Download Selten et al. supplementary material 3(File)
File 1 MB
Supplementary material: File

Selten et al. supplementary material 4

Selten et al. supplementary material
Download Selten et al. supplementary material 4(File)
File 44.6 KB