Positive and negative symptoms in methamphetamine-induced psychosis compared to schizophrenia: A systematic review and meta-analysis

Background: The clinical profiles of methamphetamine-induced psychosis (MIP) and schizophrenia are largely overlapping making differentiation challenging. In this systematic review and meta-analysis, we aim to compare the positive and negative symptoms of MIP and schizophrenia to better understand the differences between them. Study design: In accordance with our pre-registered protocol (CRD42021286619), we conducted a search of English-language studies up to December 16th, 2022, in PubMed, EMBASE, and PsycINFO, including stable outpatients with MIP and schizophrenia. We used the Newcastle-Ottawa Scale to measure the quality of cross-sectional, case-control, and cohort studies. Study results: Of the 2052 articles retrieved, we included 12 studies (6 cross-sectional, 3 case-control, and 2 cohort studies) in our meta-analysis, involving 624 individuals with MIP and 524 individuals with schizophrenia. Our analysis found no significant difference in positive symptoms between the two groups (SMD, (cid:0) 0.01; 95%CI, (cid:0) 0.13 to + 0.11; p = 1). However, individuals with MIP showed significantly less negative symptoms compared to those with schizophrenia (SMD, (cid:0) 0.35; 95CI%, (cid:0) 0.54 to (cid:0) 0.16; p = 0.01; I 2 = 54 %). Our sensitivity analysis, which included only studies with a low risk of bias, did not change the results. However, our meta-analysis is limited by its cross-sectional approach, which limits the interpretation of causal associations. Furthermore, differences in population, inclusion criteria, methodology, and drug exposure impact our findings. Conclusions: Negative symptoms are less prominent in individuals with MIP. While both groups do not differ regarding positive symptoms, raises the possibility of shared and partly different underlying neurobiological mechanisms related to MIP and schizophrenia.


Introduction
Amphetamine-type stimulants are a popular street substance, of relatively low cost and highly addictive, and were the most worldwide consumed synthetic illegal stimulant in 2019 (World Drug Report, 2021).Methamphetamine is a psychostimulant drug that acts on the central nervous system through a non-exocytotic mechanism, causing the release of monoamine neurotransmitters, including dopamine, norepinephrine and serotonin (Barr et al., 2006;Panenka et al., 2013).Such an increase in dopamine can produce psychotic symptoms, as revealed by a recent meta-analysis: one-third of methamphetamine users will develop persistent psychotic symptoms, known as methamphetamine-induced psychosis (MIP) (Lecomte et al., 2018).These persistent psychotic symptoms are very similar to those found in schizophrenia (Shoptaw et al., 2009).The impact of methamphetamine and its potential to cause acute psychosis is acknowledged (Murrie et al., 2020).However, the risk of developing a persistent psychotic primary disorder, such as schizophrenia, is still not well defined.Furthermore, nearly half of individuals with schizophrenia will experience substance abuse in their lifetime, with 4.6 times higher odds than the general population (Winklbaur et al., 2006).Moreover, the consequences of methamphetamine on negative and positive symptoms are still unclear.
The latter has received more attention in recent decades, as positive symptoms are easier to recognize and may have a more obvious impact on an individual's life than negative, cognitive, and affective symptoms (Kanchanatawan et al., 2018;Wearne and Cornish, 2018).
However, negative symptoms are a central element of the disease as they are more difficult to treat and have a worse prognosis (Correll and Schooler, 2020).They contribute significantly to long-term disability and poor functional outcomes in individuals with the disorder (Galderisi et al., 2018).Negative symptoms include apathy (avolition, asociality, anhedonia) and diminished expression (alogia and blunted affect).Negative symptoms can be classified as primary or secondary (Kirkpatrick, 2014;Kirschner et al., 2017).Primary negative symptoms are considered intrinsic to schizophrenia, while secondary negative symptoms can be caused by positive symptoms, depression, side effects, and substance abuse.In mid-80, Khantzian suggested the self-medication hypothesis of addictive disorders proposing that a subgroup of individuals is predisposed to addiction, as they used drugs to relieve from psychiatric symptoms (Khantzian, 1985).Indeed, individuals with schizophrenia might abuse psychostimulants to seek relief from their negative symptoms.Potvin and colleagues have addressed this question in a meta-analysis (Potvin et al., 2006) and found that individuals with schizophrenia and substance uses had a lower severity of negative symptoms compared to individuals with schizophrenia without substance use.Nevertheless, this was not supported by another metaanalysis (Large et al., 2014), which did not find lower negative symptoms in individuals with schizophrenia patients comorbid substance use.More recently, Sabe and colleagues focused on the association between cannabis and negative symptoms (Sabe et al., 2020) found that while individuals with schizophrenia who consume cannabis had the same symptoms as nonusers, recently abstinent individuals showed less severe negative symptoms than nonusers.The authors suggested that cannabisusing individuals may be those who are less susceptible to the development of primary negative symptoms.Nevertheless, the authors were not able to differentiate the implication of the two dimensions of negative symptoms, apathy and diminished expression in the symptoms (Kirschner et al., 2016).
To date, the association between negative symptoms and methamphetamine consumption in individuals with MIP, compared to nonusers of drugs with schizophrenia, has not been systematically investigated using a meta-analytic approach.To fill this gap, we have decided to conduct a systematic review and meta-analysis to explore the association between methamphetamine use alone or mixed with other substances and negative and positive symptoms in individuals with MIP and nonusers of drugs with schizophrenia.

Registration
This systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) guidelines (Page et al., 2021) and the Meta-analysis for Observational Studies in Epidemiology (MOOSE) guidelines (Stroup et al., 2000).On July 14th, 2020, the PRISMA protocol was published in the International Prospective Register of Systematic Reviews (registration number CRD42021286619).

Outcomes
The primary aim of the present review is to compare positive and negative symptom severity between MIP patients and patients with schizophrenia or related disorder that are non-drug users.Accordingly, we extracted the mean value and standard deviation at baseline (or the 1-time measure of cross-sectional studies) of negative and positive symptom severity with no restriction on assessment instruments.

Search strategy
PubMed, Embase and PsycINFO databases were searched up to December 16th, 2022, using a combination of MeSH terms and specific search terms (keywords): (schizo* or psychotic or psychosis) and ('amphetamine' or 'crystal meth' or 'methamphetamine' or 'psychoactive substance').
The following search limits were applied: English language, human studies, and an adult population (18 to 65 years old).The titles, abstracts, and/or methods of all papers identified in the literature search were independently inspected by three authors (IG, JC, and MS).Full texts of articles and reviews were obtained, and snowball searches of reference lists were conducted by cross-referencing key papers and other relevant articles identified through electronic searches.

Eligibility criteria
Studies meeting all the following criteria were included: a) available in English; b) observational studies (cohort studies, case-control studies and cross-sectional studies); c) involving inpatients or outpatients aged between 18 and 65 years, diagnosed with schizophrenia, schizoaffective disorder, or related psychotic disorders using DSM (APA, 2022) or ICD (Cooper et al., 1998) diagnosis of schizophrenia, schizoaffective disorder, or related psychotic disorder; d) individuals with current use of methamphetamine or showing a persistent psychotic symptom after the recent cessation of methamphetamine intoxication or acute withdrawal (within one month); e) methamphetamine must be the primary drug of choice or the sole drug used (any form of administration will be accepted).
Studies meeting any of the following criteria were excluded: a) reporting current or recent (within the last two years) use of toxic substances (narcotics, drugs), except for tobacco; b) individuals with a first episode of psychosis and those with mood disorders (bipolar disorder, major depression with psychotic characteristics); c) exclusion of case reports and qualitative studies.

Data extraction and outcomes
Search was conducted in Pubmed, EMBASE and PsycINFO.Three different authors (IG, JC and MS) independently inspected the titles, summaries, target populations and methods using the Rayyan intelligent systematic review (Ouzzani et al., 2016).The following data reported in the supplements were extracted: authors, year of publication, type of study, duration of the study, geographical area, setting, population, age range, percentage of men and women, assessment tools used to diagnose MIP and schizophrenia, inclusion criteria, and exclusion criteria for healthy volunteers.Any disagreements were resolved through consensus.The reasons for study exclusions were recorded and reported in the study flow chart (Supplementary Fig. 1).

Assessment of methodological quality
The Newcastle-Ottawa Scale modified version (NOS) (Wells et al., 2000) (Supplementary Table S1) was used to assess the methodology quality of each retained study (cross-sectional, case control and cohort studies).Two authors (IG and JC) independently conducted assessments.The Newcastle-Ottawa Scale (NOS) utilizes predefined criteria and assigns stars for each study: 4 for the quality of individual selection, 2 for comparability between cases and controls, and 3 stars for the adequate ascertainment of exposure.The highest-quality studies can receive up to nine stars.Studies with fewer than 5 points are identified as representing a high risk of bias (Luchini et al., 2017).

Statistical analysis
Statistical analysis was carried out using the Metafor R package J. Cohen-Laroque et al. (Viechtbauer, 2010).Our primary objective is to investigate the association between negative and positive symptoms and methamphetamine use, designated as the main drug of choice, based on cross-sectional or baseline scores.We defined 'main drug of choice' as either the drug the user reported as preferred or the drug considered to be the principal one by the clinician rater.A detailed examination for each study, reporting the drugs used and whether included individuals fulfilled dependence criteria, is provided in the supplement (Supplementary Table S2).Since the identified studies included small samples, we used the Hedge's g statistic to measure the effect size (Hedges, 1981).Due to the use of heterogeneous scales, we used standardized mean differences for all outcomes, and considering the difference in design, population and substance use, a random-effects model for all analysis.Heterogeneity was calculated with the I 2 statistic (Higgins et al., 2011).The potential for publication bias was examined using funnel plots (Supplementary Figs. 2 and 3).Funnel plots were used to examine publication bias (Sterne et al., 2011).Finally, a meta-regression was conducted, with negative/positive symptoms as the dependent variable and the year of publication of studies as independent variables (Supplementary Figs. 4, 5, 6 and 7).

Literature search
After assessment of the identified publications regarding our inclusion and exclusion criteria, 12 studies were included for qualitative and quantitative analyses (Alexander et al., 2019;Chen et al., 2015;Dawe et al., 2011;Hajebi et al., 2018;Hides et al., 2015;Howells et al., 2018;Mikami et al., 2003;Okada et al., 2016;Srisurapanont et al., 2011;Wang et al., 2016a;Yamamuro et al., 2015;Zhang et al., 2018).A detailed description of each included study can be found in Table 1.Screening of abstract and trial protocol led to 74 exclusions.The detailed reasons for exclusion of screened articles are reported in the flow-chart (Supplementary Fig. 1).Overall, our meta-analysis extracted one timepoint of 6 cross-sectional studies (in which 2 with a case-control design (Chen et al., 2015;L.-J. Wang et al., 2016), 3 case-control studies, 2 cohort studies and 1 prospective study).Two studies including individuals with acute exacerbation of psychosis following intake of methamphetamine were excluded (Medhus et al., 2013;Tomiyama, 1990).Due to the paucity of retrieved studies, we did not perform subgroup analysis.However, for one study (Zhang et al., 2018), the mean age is much lower than usual.

Studies and included individuals' characteristics
A total of 1151 individuals were included in our meta-analysis, including 527 individuals in the 'schizophrenia' arm and 624 in the 'MIP' arm.The population included a greater proportion of male participants in nearly all studies (63.3 % in the 'schizophrenia' group and 67.98 % in the 'MIP' group).The mean age was equivalent for both arms of the meta-analysis, with 32.46 years for the 'schizophrenia' one and 31.35 years for the 'MIP' one.Most studies were conducted in Japan (n = 3), followed by Taiwan (n = 2), Australia (n = 2), China (n = 1), Iran (n = 1), Canada (n = 1), South Africa (n = 1), and 1 study (Srisurapanont et al., 2011) was conducted in various countries: Australia, Japan, the Philippines, and Thailand.The population consisted mainly of outpatients except for 5 both inpatient and outpatient studies (Chen et al., 2015;Hajebi et al., 2018;Mikami et al., 2003;Srisurapanont et al., 2011;Yamamuro et al., 2015).Most studies did not specify if they required minimum weeks of clinical stability prior to admission of patients, besides 5 studies (Chen et al., 2015;Mikami et al., 2003;Srisurapanont et al., 2011;Wang et al., 2016a;Zhang et al., 2018).
Six studies reported concomitant use of other substances (other than methamphetamine), which included alcohol, cannabis, cocaine, heroin, MDMA, opioid and nicotine.Only 1 study specified whether methamphetamine was the main drug of use and only 1 indicated the dose of methamphetamine use in the 'MIP' arm (Chen et al., 2015).The age of first use of methamphetamine was reported in 5 studies for individuals with MIP (Chen et al., 2015;Dawe et al., 2011;Hides et al., 2015;Okada et al., 2016;Yamamuro et al., 2015) with a mean age of first use of 20.3 years.Only 2 studies (Dawe et al., 2011;Hides et al., 2015) mentioned the first age of use for individuals with MIP: the mean age was 19 years.Characteristics of drugs used by participants and characteristics of included individuals can respectively be found in Supplementary Tables S2 and S3.
All included studies used either the DSM-III, the DSM-IV, the DSM-V, the ICD-10, or the MINI to assess methamphetamine dependence or to exclude individuals affected with a primary psychiatric disorder.The main scales used, along with their corresponding scores, can be found in Supplementary Table S4.To assess the negative and positive symptoms, 6 studies used the PANSS scale (Alexander et al., 2019;Hajebi et al., 2018;Howells et al., 2018;Okada et al., 2016;Yamamuro et al., 2015;Zhang et al., 2018), 5 used the BPRS scale (Chen et al., 2015;Dawe et al., 2011;Hides et al., 2015;Mikami et al., 2003;Wang et al., 2016a) and 1 used the Manchester scale (Srisurapanont et al., 2011).Six studies (Alexander et al., 2019;Dawe et al., 2011;Hides et al., 2015;Howells et al., 2018;Mikami et al., 2003;Srisurapanont et al., 2011;Wang et al., 2016a;Wang et al., 2016b) evaluated the depression and anxiety score using the PANSS scale, the BPRS scale, the CDSS scale and the HDRS scale.Details of included studies are reported in Table 1.

Positive symptoms levels for individuals with 'MIP' and 'nondrug users with schizophrenia'
The primary outcome data on positive symptoms severity was available for 12 studies using the PANSS-P scores.When examining the PANSS scale, the range of scores for positive symptoms in individuals with MIP was (14.5-19.8)and was (14-19.8)for individuals with schizophrenia.
The correspondent funnel plot was not in favor of a publication bias (Supplementary Fig. 2).These results suggest that there are no differences regarding positive symptoms for stable individuals with MIP and stable nondrug users' individuals with schizophrenia.

Negative symptoms levels for individuals with 'MIP' and 'nondrug users with schizophrenia'
The primary outcome data on negative symptoms severity was available for 12 studies.
When examining the PANSS scale, the range of scores for negative symptoms in patients with MIP was (7-17), while individuals with schizophrenia exhibited a range of (8.9-19.8).
A significant difference was found between the 'MIP' and nondrug users with schizophrenia groups regarding negative symptoms (SMD, − 0.35; 95CI%, − 0.54 to − 0.16; p = 0.01) in presence of a moderate heterogeneity (I 2 = 54 %).These results suggested that individuals with MIP present fewer negative symptoms that nondrug users with schizophrenia.
The visual inspection of the correspondent funnel plot suggested the presence of at least one outlier (Mikami et al., 2003) (Supplementary Fig. 3).We conducted an Egger regression test (Egger et al., 1997) that indicated the absence of funnel plot asymmetry (t = − 1.09, df = 10, p = 0.303).In addition, we conducted a leave-one-out analysis to assess the impact of potential outliers on the estimation of the overall effect size obtain, but results remained significant (SMD, − 0.254; 95CI%, − 0.37 to − 0.13; p < 0.001), with no reduction of heterogeneity (I 2 = 58.0%) (Supplementary Table S5).
Finally, we extracted from 3 studies the results for diminished expression and amotivation dimension, however results were not significant (Supplementary Fig. 4).

Depressive symptoms for individuals with 'MIP' and 'nondrug users with schizophrenia'
Scores for depressive symptoms were available in six studies and were reported with various scales.

Sensitivity analysis including only studies with a low risk of bias
We conducted a sensitivity analysis by excluding studies that did not present scores on the Newcastle-Ottawa scale and were considered to  have a low risk of bias (total stars >5).However, for both negative and positive symptoms, the results did not differ, although the effect size was numerically higher for negative symptoms, and the confidence interval was large (SMD, − 0.44; 95 % CI, − 0.82 to − 0.07; p = 0.04) (Supplementary Figs.7 and 8).

Exploratory analysis of potential moderators
To explore potential moderators, we conducted a meta-regression, incorporating two potential moderators: the year of publication and the mean age of intervention and control groups as independent variables.When considering negative symptoms as the dependent variable, the test was not significant for both independent variables (Supplementary Figs. 9 and 10).Results were similar when considering positive symptoms as the dependent variable (Supplementary Figs.11 and 12).

Discussion
The main finding of our meta-analysis is that negative symptoms are less prominent in individuals with MIP than in individuals with schizophrenia, while there is no significant difference for positive symptoms.We further discuss below both the differences and similarities between individuals with MIP and individuals with schizophrenia.

Distinct clinical phenomenology
Apart from the presence of less negative symptoms for individuals with MIP -which is an inconsistent feature among included studiesother distinct clinical characteristics were found.Regarding negative symptoms dimensions, Srisurapanont and colleagues (Srisurapanont et al., 2011) proposed that individuals with schizophrenia present more poverty of speech than individuals with MIP, suggesting a higher impact on the diminished expression dimension and similar effects on the apathy dimension.In our findings, negative symptoms seem to be less pronounced in MIP, however we could not verify if the diminished expression dimension is less impacted in MIP (Supplementary Fig. 4).Nevertheless, some results were found in a recent systematic review that reported that negative symptoms did not seem to be prominent in MIP and were even absent in 20 % of their inclusions (Voce et al., 2019).Additionally, a significant proportion of individuals presenting positive symptoms were using other substances than methamphetamines (Voce et al., 2019).
The fact that both populations do not differ regarding positive symptoms is also a major finding of our work.Indeed, several studies have led to the suggestion that MIP does not represent a distinct diagnosis entity but rather that the drug triggers a vulnerability to develop a schizophrenia (Bramness et al., 2012;McKetin et al., 2018;Wearne and Cornish, 2018).Wearne and colleagues (Wearne and Cornish, 2018) gathered evidence that chronic MIP may have clinical similarities to primary psychotic disorders.Several factors might differentiate MIP from schizophrenia such as gender distribution (Ezzatpanah et al., 2014), age of onset (Gan et al., 2018) or premorbid function (Chen et al., 2003).
McKetin and colleagues (McKetin et al., 2018) explored the phenomenological differences between individuals with MIP and individuals with schizophrenia and found the presence of non-persecutory delusions and hallucinations as a marker for persistent MIP or primary psychosis in methamphetamine users.Wang and colleagues (Wang et al., 2016a;Wang et al., 2016b) also suggested that individuals with MIP experience significantly more hallucinations (both visual and tactile) and have less severe negative symptoms than individuals with schizophrenia.Furthermore, Zhang and colleagues proposed more affective symptoms in the individuals with MIP (Zhang et al., 2018).
Another important aspect is the potential impact of methamphetamine on depressive symptoms.Although we did not find significant difference in the few studies included for depressive symptoms, in the Alexander and colleagues' study, individuals with MIP presented fewer depressive symptoms than individuals with schizophrenia, and, individuals with schizophrenia that were methamphetamine consumers also presented score in between those two groups.Nevertheless, we cannot exclude than this group could be a mix of the other group of individuals (Fig. 2).
Finally, many authors (Chen et al., 2015;Yamamuro et al., 2015) have shown that individuals with MIP show more preserved cognitive functions, with possibly less alterations of the prefrontal cortex.As such, a rapid recovery of dopamine levels in early abstinence for methamphetamine users was found (Boileau et al., 2016).

Neurobiology of methamphetamine-induced psychosis
On the neurostructural levels, individuals with MIP presented less alterations, with increased gray matter volumes compared to individuals with schizophrenia (Zhang et al., 2018).In addition, a positive correlation between gray matter volume of the left cerebellum crus and duration of abstinence suggested that prolonged abstinence is beneficial to cognitive function recovery (Zhang et al., 2018).
Furthermore, in schizophrenia cortical GABAergic disturbances have been found (Dienel and Lewis, 2019), specifically in the prefrontal cortex, and are proposed to underlie negative (Egerton et al., 2021).For instance, Kraguljac et al. (2012) found that the severity of negative symptoms was negatively related to the glutamate level (Kraguljac et al., 2012).Although some authors propose a dysfunction of the cortical GABAergic system in individuals with MIP (Hsieh et al., 2014), there are still no evidence to support this hypothesis.
One hypothesis would be that the short-term use of amphetamine leads to a stimulant effect with a protection/reduction of negative symptoms, namely the self-medication hypothesis (Potvin et al., 2006).Another hypothesis would be that negative symptoms arise as a consequence of methamphetamine addiction.However, in both scenarios, it is noteworthy that chronic users are more likely to develop negative symptoms, especially when considering the loss of dopaminergic fibers (Ares-Santos et al., 2013).
Nevertheless, brain circuitries implicated in both disorders are probably distinct, and comparing individuals with MIP and individuals with schizophrenia is an indirect comparison between two distinct disorders.Therefore, most finding should be interpreted with caution.

The methamphetamine sensitization model of schizophrenia
Amphetamine sensitization of the dopaminergic system has been proposed as a rodent model of schizophrenia-like symptoms (Peleg-Fig.2. Forest plot for the comparison between 'MIP' versus 'patients with schizophrenia' that are non-drug users with the outcome negative symptoms.The blue diamond representing the pooled effect is on the left of the line of no effect, indicating less negative symptoms for the MIP group compared to patients with schizophrenia that are non-drug users.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)Raibstein et al., 2009).Therefore, methamphetamine intake can be considered as an external stressor that, in the presence of intrinsic vulnerability, is susceptible to lead to the development of psychotic symptoms or MIP (Goh and Agius, 2010).
As discussed by Bramness and colleagues (Bramness et al., 2012), we hypothesize that the relationship between MIP and primary psychosis can be viewed within the framework of a traditional vulnerability-stress model of schizophrenia.More specifically, we propose the hypothesis of a continuum of vulnerability among methamphetamine users, with on one hand the individual susceptibility to the development of psychosis, and on the other hand the cumulative doses of methamphetamine.Schematically, we can postulate on the distinction of three different populations (Fig. 3).
First, methamphetamine users who will most likely not develop MIP but would still be exposed to transient positive psychotic symptoms during the hours following the intake of high-dose methamphetamine.Second, individuals presenting vulnerability for transition to psychosis (e.g., clinical high-risk state for psychosis) who will most likely develop MIP, depending on the corresponding level of vulnerability and the cumulative dose of the drug.The third group would be individuals with schizophrenia-spectrum disorder who would exhibit exacerbation of their psychotic symptoms in response to stimulants at doses that would not be psychotogenic in normal subjects.
Nevertheless, the question of whether there is a real continuum between these three groups of individuals remains to be answered.

Clinical significance of findings and implication for treatment of individuals with MIP
Identifying vulnerabilities and risk factors for developing MIP can help detect individuals who may be more susceptible to MIP.Various risk factors, mostly associated with the lifetime prevalence of MIP, have been identified, such as the frequency, higher dose, injection, and severity of methamphetamine addiction, concomitant substance use, earlier onset of drug use, and simultaneous personality and mood disorders (Arunogiri et al., 2018;Fiorentini et al., 2021;Grant et al., 2012;Matsumoto et al., 2002).Moreover, a positive family history of methamphetamine abusers multiplies by five the risk of developing psychotic symptoms (Glasner-Edwards et al., 2008).
Furthermore, some genetic factors were also proposed (Nohesara et al., 2016), as some prognostic factors associated to the persistence of MIP: the severity of psychotic symptoms, the duration of methamphetamine use, and prolonged depressive symptoms (Lecomte et al., 2018).
Nevertheless, considering the important rate of lifetime MIP proposed by Lecomte and colleagues (42.7 %), the primary treatment goal should be abstinence of the drug (Herbst et al., 2022).
Primary prevention can target regulation of methamphetamine (amount, duration, and frequency), and interventions on user's lifestyle combine with a follow-up.Cognitive behavioral therapy can also be proposed to motivate abstinence (Roll et al., 2013), treat comorbidity and help maintain follow-up surveys (Chiang et al., 2019).Future study methodologies should particularly focus on the subgroup of individuals with MIP who develop schizophrenia.

Antipsychotic treatment for MIP
According to our results, clinicians should primarily aim to reduce positive symptoms of individuals with MIP.For MIP treatment, antipsychotics that target D2 receptors are classically proposed (Chiang et al., 2019;Fluyau et al., 2019).In case of severe psychotic symptoms, or persistence of MIP after methamphetamine abstinence presented by individuals affected with MIP, antipsychotics should be proposed.Fluyau and colleagues concluded that haloperidol, risperidone, olanzapine and aripiprazole, quetiapine are safe and reduce both positive and negative symptoms of MIP (Fluyau et al., 2019).Of interest, aripiprazole does seem to have a better tolerance as individuals with MIP stay longer in treatment.However, this comes at the price of more akathisia and agitation (Wang et al., 2016b).Furthermore, Temmingh and colleagues studied extra-pyramidal side effects of antipsychotics in individuals with MIP and found that a 4-times more likelihood of experiencing extrapyramidal side effects than individuals with no methamphetamine use disorder (Temmingh et al., 2020).Although existing guidelines do not specify a preferred choice of antipsychotics, in a recent study, Srisurapanont and colleagues proposed that olanzapine and quetiapine may be a preferred antipsychotic for MIP, although the evidence was rated as low-quality (Srisurapanont et al., 2021).

Limitations
Several limitations impact our work and participate to the considerable heterogeneity observed.First, different inclusion criteria and MIP definition were used among the original studies, which contributes to the heterogeneity of the population.Included individuals presented similar global characteristics such as being outpatients or chronic methamphetamine users that were abstinent of this drug of use.However, they were also polysubstance users, including alcohol, cannabis, cocaine, and heroin (Alexander et al., 2019); ketamine, MDMA and opioid (Chen et al., 2015); cannabis (Dawe et al., 2011); alcohol, cannabis and heroin intake (Hides et al., 2015); alcohol and nicotine addiction (Howells et al., 2018;Zhang et al., 2018); nicotine (Srisurapanont et al., 2011).This data of whether included individuals were polysubstance users was not reported in the rest of our included studies (Hajebi et al., 2018;Mikami et al., 2003;Okada et al., 2016;Wang et al., 2016b;Yamamuro et al., 2015).The total number of years of abstinence was also very heterogeneous between studies and not always reported (Supplementary Table S2).Furthermore, the mode of consumption and the 'quality' of methamphetamine was mostly lacking.However, Matsumoto and colleagues did not found difference in the level of psychotic symptoms between individuals that were smoking or injecting methamphetamine (Matsumoto et al., 2002).
The assessment of exposure (average dose of methamphetamine use) was only reported in one study (Chen et al., 2015), as the current dose of antipsychotic medications, which limits our finding.Moreover, the severity of dependence was not clearly reported, nor for methamphetamine, neither for other substances.Furthermore, the definition of 'stable' individuals with MIP or schizophrenia is not reported in most studies, including current medications.
In further studies, authors should clearly report the baseline severity, the stage of the illness, type of drug used, their amount, and the distinction between substance use, substance abuse and addiction to substance.Finally, it must be emphasized that the cross-sectional design of selected studies does not allow for a causal interpretation of the observed associations.

Conclusions
The complex relationship between methamphetamine use and psychosis has received significant attention but remains inadequately explored, particularly when considering the single-study level.This first meta-analysis provides a systematic characterization and comparisons of psychotic symptoms in individuals with MIP and schizophrenia.It found meta-analytic evidence for similar levels of positive symptoms but different levels of negative symptoms, suggesting partly overlapping but also distinct symptom profiles in both populations.Our finding raises questions about the underlying neurobiological effects of such psychostimulants, as well as the treatment of negative symptoms in schizophrenia.Given the prevalence of psychotic symptoms, particularly positive ones, methamphetamine users should be strongly discouraged from continuing consumption.To better understand the natural course of these presentations and the effectiveness of available treatments, detection strategies such as follow-up cohorts are necessary.Additionally, early treatment programs for methamphetamine abuse and prevention strategies should be evaluated.

Role of the funding source
Internal funding only.

Declaration of competing interest
Kaiser S has received royalties form Boehringer Ingelheim company.All other authors declare no conflict of interest.

Fig. 1 .
Fig. 1.Forest plot for the comparison between 'MIP' versus 'patients with schizophrenia' that are non-drug users with the outcome positive symptoms.The blue diamond representing the pooled effect is on the line of no effect, indicating no difference between MIP and patients with schizophrenia that are non-drug users.(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Table 1
Characteristics of included studies.