Patients with psychotic spectrum disorders have a markedly reduced life expectancy compared to the general population. For instance, patients with schizophrenia have a reduced life span of 15–20 years compared to the general population and also have more somatic co-morbidities (1, 2). This is partially due to the development of the metabolic syndrome (MetS). MetS is a combination of risk factors that can lead to increased mortality and morbidity such as cardiovascular disease and diabetes (3). According to The US National Cholesterol Education Programme Adult Treatment Panel III (ATP III), MetS is diagnosed when a person fulfills at least three of the following criteria: waist size of at least 102 cm for males and at least 88 cm for females; triglycerides of at least 150 mg/dl; HDL cholesterol level of <40 mg/dl for males and <50 mg/dl for females; a blood pressure of more than 130 mmHg systolic or 85 mmHg diastolic; and a fasting glucose of more than 100 mg/dl (4). The International Diabetes Federation (IDF) applies similar criteria but requires the presence of an increased, ethnicity specific waist size plus two or more of the abovementioned factors (5).

Rates of MetS vary significantly between populations. Genetic and geographical environmental differences are known to affect metabolic risk factors. For Europeans the age-adjusted rates are 18.4% for men and 14.4% for women while in South Asians the occurrence in men is 28.8 and 31.8% for women, based on the ATP III MetS definition (6). In Japan, the rate of MetS is 14.2% in the general population (7) while in the United States, the age-adjusted weighted prevalence is 34.3% (8). Compared to the general population, patients with schizophrenia have a 2- to 3-fold increased prevalence of MetS varying per country (9, 10). Given the higher prevalence of MetS among these schizophrenia patients, the syndrome poses a greater health risk to this population. It has a significant impact on morbidity and mortality due to the increased risk of diabetes mellitus type 2 (DM2) and cardiovascular disease (3). Explanations for this higher prevalence are a poor diet, cigarette smoking (11, 12), lack of exercise (12), stress and abnormalities in the hypothalamic-pituitary-adrenal axis (13). On top of this, SGA use is associated with metabolic abnormalities and may exacerbate this condition by causing weight gain, glucose and lipid metabolism deregulation (6, 14–18). Antipsychotics can influence metabolic parameters within 2 weeks of treatment (19). However, the existing body of research suggests that the degree of metabolic dysregulations varies considerably between different SGAs (19). Evidence for weight gain was found for clozapine, zotepine, olanzapine, and sertindole, iloperidone, quetiapine, risperidone and paliperidone, and brexpiprazole. SGA are also associated with glucose abnormalities and development of DM2 (9, 20–23). Pillinger et al. (24), performed a large meta-analysis to compare and rank antipsychotics based on their metabolic side-effects and to identify predictors of antipsychotic-induced metabolic dysregulation. Increased baseline weight, male sex, and non-white ethnicity were found to be predictors of susceptibility to antipsychotic induced metabolic change (24). Increase in fasting-glucose was associated with a higher risk of (cardio)vascular disease and was especially evident in olanzapine, zotepine, and clozapine use (24). Furthermore, several studies have demonstrated lipid disturbances following SGA-use (25). Evidence was found that quetiapine, olanzapine, zotepine, and clozapine are negatively correlated with triglyceride alterations (24). Finally, research indicates that patients using SGA have increased cholesterol levels (26), especially in patients using quetiapine, olanzapine, or clozapine (24). Clozapine is the most effective treatment to improve symptom severity and to reduce the risk of recurrent suicidal behavior in patients with schizophrenia or schizoaffective disorder (27, 28). Meanwhile, together with olanzapine, it is also associated with the highest increases in weight, body mass index (BMI) and total cholesterol, suggesting that the greatest metabolic disturbances are caused by the most efficacious antipsychotics (16, 24, 29).

Qualitative research shows that patients have concerns about the negative long-term effects of antipsychotics on their physical appearance and physical health (30, 31). Multiple studies show that patients using antipsychotic medication consider weight gain, possibly leading to overweight and obesity, as one of the most disturbing adverse events and therefore one of the major reasons for non-adherence to therapy (32, 33). Taking this into consideration, elucidating factors that contribute to the occurrence of MetS in specific antipsychotics is useful for clinical practice.

Although the use of SGA is thus clearly associated with increased risk to develop MetS, specific factors that increase this risk have remained largely elusive. Apart from the effects accounted for by lifestyle and antipsychotic medication, research hints at a shared underlying pathophysiology between schizophrenia and cardiovascular disease (34). The high interindividual variability in the occurrence of MetS suggests that genetic factors influence its risk (35). In previous studies, factors associated with MetS among patients with psychotic spectrum disorders who use SGA have been examined (36–38). These factors, however, have been discussed in isolation and findings remain inconclusive. Furthermore, in the analysis by Pillinger et al. (24), MetS was not examined as an outcome measure; only isolated factors (e.g., dislipidemia, hypercholesteremia) were investigated. We chose to perform a systematic review in which we included only studies that took MetS (the combination of risk factors) as an outcome measure. Since people with MetS have a reduced life span of 15–20 years compared to the general population and also have more somatic co-morbidities (39), we reasoned there would be added value in considering MetS as a conglomerate of factors, instead of its isolated components. Finally, in the study of Pillinger et al. (24), limited clinical and biochemical factors and no genetic factors were investigated, which we here expanded on. Thus, we conducted a systematic review of factors associated with MetS during treatment with SGAs (4, 5).

In this systematic review that examines the clinical, biochemical and/ or genetic factors associated with MetS in patients using SGAs, we found male and female gender, higher age (at initiation of treatment), concomitant use of mood stabilizers, higher baseline BMI, higher current BMI, higher dose and longer duration of treatment to be positively associated with MetS in patients treated with clozapine. In studies with pooled antipsychotics, hypo-adiponectinemia, elevated CRP (≥ 3 mg/L), higher WBC count, female and male gender, older age, higher baseline BMI, higher current BMI, BMI >24, higher dose, longer duration of psychosis, tobacco smoking and HTR2C polymorphism were found to be positively associated with MetS. Overall, higher age, higher baseline BMI, higher current BMI and male as well as female gender were the only factors associated with MetS across all antipsychotics.

A large meta-analysis by Pillinger et al. (24), found male gender to predict greater vulnerability to antipsychotic-induced metabolic dysregulation. In this systematic review gender was also found to be a factor associated with MetS in patients treated with clozapine and in the pooled antipsychotics group. However, the results on this factor are not unequivocal since some studies found male gender while other studies found female gender to be associated with MetS. One possible explanation for these contradictory findings is related to age. In the general population, the risk of MetS increases with age in a gender-specific manner: under 50 years of age the risk is slightly higher in men, while over the age of 50 the risk is higher in women (97). In addition to gender, age was found to be a risk factor in the clozapine studies, as well as in the pooled groups. Given the fact that with older age the risk of MetS in the general population increases, this is not surprising. Due to variations in MetS prevalence between different countries, we expected to find an association between ethnicity and MetS. However, no clear association was found. It cannot be excluded that this incongruence is compounded by socio-economic factors.

Another factor associated with the development of MetS in studies including clozapine and SGAs was higher antipsychotic dosage. It has been suggested that not all SGA have the same propensity to induce metabolic disturbances, whereby clozapine and olanzapine appear to have concentration-dependent metabolic effects (98). Interestingly, drugs with high affinity for the H₁-histamine, muscarinic, and α-adrenergic receptors, also seem to exhibit the strongest off-target metabolic effects, especially with higher doses (98).

As expected, considering the clinical role of CRP as cardiovascular risk indicator, elevated CRP (≥3 mg/L) was found to be positively associated with MetS. The finding of higher WBC, another inflammatory marker, is consistent with previous literature, however, the underlying mechanism explaining this association remains unclear (80). Furthermore, the association between concomitant use of mood stabilizers and MetS in patients treated with clozapine is in line with previous studies showing a higher MetS prevalence in patients receiving polypharmacy vs. monotherapy and studies reporting weight gain and MetS following treatment with mood stabilizers (10, 99–101). Our findings regarding hypoadiponectinemia support results from previous studies (102). Adiponectinemia is thought to have a normalizing effect on metabolic dysregulations (103).

Literature suggests that nicotine can reduce food intake and body weight (104) and might therefore reduce the risk of MetS (46, 74). Interestingly, in our systematic review, tobacco smoking was found to be positively associated with MetS. Tobacco smoking is known to cause upregulation of CYP1A1, CYP1A2, CYP2E1 and UGT, enzymes that are known to be involved in, among others, antipsychotic metabolism (105, 106). Smoking patients therefore would have lower plasma concentrations when comparing to non-smoking patients with the same dose. Nevertheless, physicians might prescribe these patients higher doses in order to reach the same plasma level, which could explain our finding (54, 89, 100).

Previous studies evaluating genetic factors observed inconsistent results on whether certain polymorphisms increase risk of MetS (Supplementary Tables 10, 11). To our knowledge, only the rs1414334 C allele of the HTR2C gene has been associated with MetS in multiple studies, although in pooled groups of antipsychotics. Although the mechanism of action is unclear, Mulder et al. (107), reported this polymorphism also to be significantly related to an increased risk of obesity in psychiatric patients treated with antipsychotics (107). This strengthens the idea that this polymorphism could be relevant.

One of the constituents of MetS is waist circumference (WC). A normal waist circumference differs for specific ethnic groups due to differences in cardiometabolic risk. For example, the relationship between WC and risk factors is such that men and women of South Asian descent present with a more severe metabolic risk profile than those of European descent at the same WC (108). As ethnic descent influences the relationship between WC and metabolic risk factors, current WC data derived from studies in European populations cannot be directly extrapolated to Asians. Furthermore, Asians have increased cardiometabolic risk with lower waist circumferences than other populations (109). Therefore, ethnic background should be considered when using WC as a marker of cardiovascular risk.

Naturally, several limitations may have influenced the results and interpretation of this systematic review. First, the included literature largely consisted of cross-sectional studies. Due to the nature of these studies, no definite conclusions regarding potential causal relationships may be made. Second, the antipsychotic agents used prior to the treatment drug were not recorded in most cross-sectional studies. If these antipsychotics included atypical antipsychotics, the associated factors could not be attributed to the SGA alone. For example, olanzapine prior to the initiation of clozapine leads to adverse metabolic consequences impacting weight, glucose and cholesterol (110). Besides, especially in the pooled groups of antipsychotics, the variation between SGA with their differing propensities to cause metabolic disturbances might have skewed the results. Third, a relatively large part of the studies was conducted in Asian countries. The results of these studies may therefore not be directly extrapolated to the European patient population since there are important metabolic differences partially related to lifestyle and diet between Asian and Caucasian patients on SGA (24). Twenty-seven studies were performed in Asian countries. Furthermore, a substantial number of the studies (40%) was performed in low- or middle-income countries (2 and 21, respectively), which may influence the generalizability of the results. More importantly, the inclusion of studies investigating different ethnicities is a problem regarding the genetic factors. The Chinese population can differ widely in genetic makeup compared to Europeans. This will have clinical implications for genotyping patients based on genetic findings that may not be relevant to the clinical population of interest. Fourth, several factors found to be associated with MetS (see Supplementary Tables 4–11) were only reported in a single study and thus excluded, while other factors were consistently identified in multiple studies. The results of the current study should therefore be interpreted with caution. Finally, we only included studies that used MetS as outcome measure to compare studies adequately. During the suitability screening in our literature search, we noticed that the majority of the initially identified studies investigated the effect of antipsychotics on one or more separate metabolic disturbances, instead of the MetS definition using ATP III, ATP III-A or IDF criteria. For example, the large meta-analysis by Pillinger et al. (24) investigated the effects of various antipsychotics on individual metabolic abnormalities, but not the syndrome as a whole. Therefore, these studies fell outside the scope of this systematic review. Potentially, this influenced the validity of this systematic review, leading to ambiguous and incomplete results.

Despite these limitations, the findings of this study are a promising first step toward the application of using clinical, biochemical and genetic information in personalized medicine. Evidence shows that MetS is highly prevalent among schizophrenia patients. Possibly, several factors interact to increase this risk in schizophrenia patients, including effects of SGAs, unhealthy lifestyle, and genetic and pathophysiological vulnerability. Further research is needed to elucidate how individual risk factors operate to increase this risk as well as how risk factors may interact to further increase MetS risk. Further research is also required to examine whether the contributions of these factors geographically differ. In this context, both clinical and preclinical studies may prove useful in the future to ascertain underlying pathophysiological mechanisms. Further risk factor management strategies are also required, involving pharmaceutical and nonpharmaceutical lifestyle interventions to try and counter the effects of such risk factors on MetS risk profiles in schizophrenia patients.

However, before applying these factors in clinical practice, by determining which patients have a high risk at developing MetS during SGA use, more research is required. Studies are needed using machine learning techniques to identify the exact molecular basis of the identified factors and to individually predict the risk to develop MetS in clinical practice. In this way, the discrepancy between life expectancy of patients with psychotic spectrum disorders and the general population may be reduced.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.

MS and SE performed the literature search. All authors contributed to the writing of the systematic review and read and approved the final manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.