Atypical antipsychotics are reported to be associated with severe side effects like weight gain, obesity, and metabolic dysfunction (Li et al., 2018; Luckhoff et al., 2019; Li et al., 2021). It is reported that the prevalence of obesity ranges from 10% to 60% in SZ patients (Naslund et al., 2016; Annamalai et al., 2017). Weight gain/obesity impacts the quality of life and adherence to antipsychotic drugs in SZ (Dibonaventura et al., 2012; Pillinger et al., 2020).

A large of evidence has identified weight gain as a prognosis biomarker of the response to antipsychotic drugs (Ascher-Svanum et al., 2005; Bai et al., 2006; Sharma et al., 2014), underscoring the need for early monitoring of weight in SZ patients. Even in adolescents with SZ, substantial decreases in global psychopathology have been reported to be associated with weight gain after taking antipsychotic medications (Sharma et al., 2014). Studies have demonstrated that risperidone was associated with intermediate weight gain (1.76 kg), compared with olanzapine with the highest (3.45 kg) (De Hert et al., 2011). It is reported that individuals who were younger, female and first-episode and were less exposed to antipsychotics previously were more likely to increase the risk of obesity.

Antipsychotics have different treatment outcomes for patients with SZ (Zhu et al., 2022). Some studies have revealed the prognostic role of weight gain after treatment with antipsychotics in favorable treatment outcomes (Chen et al., 2021). However, considering the independence from confounding factors, such as treatment adherence and duration, the clinical significance of such a relationship has been questioned (Correll et al., 2011; De Hert et al., 2011; Hermes et al., 2011). Notably, a previous study has reported that baseline BMI correlated with antipsychotic-induced weight gain, and it was a predictive biomarker in the therapeutic response to antipsychotics (Verma et al., 2009). Another study in adolescents with SZ revealed that olanzapine-associated weight gain was not independently correlated with therapeutic response to olanzapine, but baseline obesity was related to more olanzapine-associated weight gain and symptomatic outcome (Kemp et al., 2013).

Studies have shown that the first-episode and medication-naive SZ (FEMNS) patients display the advantage of reducing the possible impacts of the use of antipsychotics, the duration of illness, and the impact of comorbidities associated with chronic illness (Xiu et al., 2020). In exploring the link between BMI and response to antipsychotics, thus, we recruited FEMNS patients in the current study. This study was designed to determine the predictive role of baseline BMI in the symptom improvements after 12-week treatment with risperidone in SZ.

The sample included 241 FEMNS patients (128 men, 113 women). Thirty-eight participants were lost before the 2-month follow-up and 53 patients were lost after 2 months of treatment with risperidone. Finally, a total of 91 patients were lost. Table 1 shows the differences between completers and those who dropped out in the follow-up. There were no significant differences in the demographic characteristics and clinical data between completers and drop-outs (all p > 0.05).

After treatment, the mean changes in weight were 2.7 kg (SD = 3.8). According to the criteria of obesity, we identified 34 patients in the high BMI group and 207 patients in the low BMI group. Comparisons of demographic characteristics and clinical data between the high BMI and low BMI groups are shown in Table 1. The mean changes in weight were 0.4 (95% CI: −0.7–1.5) in the high BMI subgroup and 3.1 (95% CI: 2.5–3.6) in the low BMI subgroup. Significant differences were observed in weight gain between the low BMI and high BMI subgroups (p < 0.01).

After treatment with risperidone, clinical symptoms were significantly improved (Bonferroni corrected all p < 0.05) (Table 2). Pearson correlation analysis revealed that baseline BMI was negatively correlated with the improvement in negative symptoms (r = −0.14, p = 0.03). After subgroup analysis, we found that the negative association was only present in the high BMI group (r = −0.43, p = 0.01), but not in the low BMI group (Figure 1). Moreover, baseline BMI was not associated with the improvement in positive symptoms, general psychopathology, and PANSS total score (all p > 0.05). Further linear regression analysis confirmed that baseline BMI (β = −0.15, t = −2.2, p = 0.027) was a predictive factor for negative symptom improvements, while sex, age and years of education were not associated with negative symptom improvement (all p > 0.05).

Further, among those with low BMI, we found 28 patients were underweight or with undernutrition. To rule out their influence on the results, we reanalyzed the data without these 28 patients and there was no difference in the result (Supplementary Tables S1).

We here found that baseline BMI was associated with weight gain and negative symptom improvements in this relatively large sample of FEMNS patients on 12-week treatment with risperidone. However, we did not find an association between female sex, younger age education levels, and symptom improvements.

This study found that baseline BMI was significantly associated with weight gain after 3 months of risperidone monotherapy. Patients with a lower baseline BMI gained more weight after risperidone medication than those with a higher baseline BMI. Our findings are in line with most previous studies (Jones et al., 2001; Leucht et al., 2013; Huhn et al., 2019) and the recent longitudinal cohort study in FEMN patients with SCZ (Vázquez-Bourgon et al., 2022), providing further evidence for a relationship between baseline BMI/bodyweight and antipsychotic-induced weight gain in SCZ (Gentile, 2009; Verma et al., 2009). Notably, previous evidence-based data reveal that a variety of other risk factors, including ethnicity, young age, recent onset of psychotic symptoms, unemployment, unhealthy lifestyle and low income, may contribute to the rate or magnitude of antipsychotic-induced weight gain during long-term treatments (Gentile, 2009). However, in this study, we did not find an association between weight gain and education years, which was well-known to be strongly associated with unemployment, unhealthy diet and low income. This may be due to the differences between the patients with SCZ we recruited in this study and those in previous studies, considering that a majority of previous studies have investigated weight gain in outpatients on long-term antipsychotic medication.

We further found that baseline BMI was an independent predictor for the improvement of negative symptoms after treatment, after controlling for age, sex, and education years. Indeed, there is evidence to show a close relationship between negative symptoms in SZ patients and BMI at baseline (Jones et al., 2001; Leucht et al., 2013; Sharma et al., 2014; Raben et al., 2017; Huhn et al., 2019), which is consistent with our findings. The findings in the current study were also in line with the studies on FEMNS patients (Zipursky et al., 2005; Venkatasubramanian et al., 2010; Kemp et al., 2013). Importantly, the abovementioned studies have all controlled for confounding factors, such as sex, age, antipsychotic medications, and duration of illness.

However, the exact mechanism remains unclear. It may be due to the shared mechanistic pathway between body weight regulation and negative symptom improvements after antipsychotics. Risperidone has been reported to have a relatively high affinity for D₂, 5-HT₂, histamine H₁ receptors, and NE alpha-2 receptors (Nasrallah, 2008). Animal studies also revealed that obesity induced by a high-fat diet and increased food intake correlated with 5-HT₂ deficiency (Mercer et al., 1994). In particular, the use of antipsychotics may lead to the preferential metabolism of carbohydrates over fats and further lead to increased fat storage (Tiwari et al., 2018; Kaar et al., 2019). Leptin, a key signal for determining the size of fat depots in the brain, was found to be increased following atypical antipsychotic medication (Brömel et al., 1998; Kraus et al., 1999). On the other hand, dopamine D₂ receptors play an important role in the reward circuit and in mediating both obesity and therapeutic response to antipsychotics, which may explain the association between BMI and therapeutic benefits in SZ.

Higher BMI was associated with fewer negative symptom changes in this study. It is reported that in patients with higher BMI in the early phase of SZ, brain functional connectivity associated with food cravings and weight control was decreased (Homan et al., 2019). Additionally, the prefrontal cortex of obese patients exhibited altered insulin and DA gene expression, resulting in relatively poorer outcomes than those with lower BMI (Mansur et al., 2018). Altogether, this study suggests that BMI at baseline may be a more sensitive indicator of the therapeutic benefits in the early stage of SZ, as most of the reported correlations between weight gain and therapeutic responses to antipsychotics are mixed.

Some strengths should be mentioned in this study. This is a prospective and longitudinal study examining a well-characterized group of FEMNS patients. Standardized treatment with a single antipsychotic excluded the different impacts of antipsychotic drugs. However, this study has several limitations. First, in this study, obesity-related biomarkers such as glucose, cholesterol, insulin resistance, and lipids were not recorded. Additionally, other metabolic parameters, such as food preference, dietary record, and caloric intake were also not collected. Second, only risperidone was assessed in our study. Therefore, the results of this study may not be generalized to other types of antipsychotics that may be different in metabolic disturbance. Third, the criteria for obesity in our study are only for Chinese, thus the conclusions in this study cannot be generalized to other populations. Fourth, the regulation of body weight may overlap with the pharmacological mechanisms of antipsychotic drugs. Genetic factors are known to be associated with the pharmacodynamics, pharmacokinetics, and adverse effects of antipsychotics. Additional pharmacogenetic analyses may provide new insights into the current findings. Fifth, considering the role of lifestyle, diet, and exercise in explaining the contradictory findings across studies, we did not collect the detailed diet and exercise in the present study.

In conclusion, the current study found that BMI was negatively correlated with the improvements in negative symptoms after treatment with risperidone for 12 weeks in FEMNS patients. In addition, we identified baseline BMI as an independent predictor for negative symptom improvements in SZ. Our study underscores the key role of BMI in the clinical management of patients in the early stage of SZ. These findings provide further evidence that greater efforts should be made to prevent obesity in clinical practice from the early phase of SZ.