Table 1. Summary of studies on the relationship between serum lipids and symptoms of schizophrenia.
| Publication, Country, Study Design | Demographics | Diagnosis | Antipsychotic Medication | Follow-up Duration | Serum Lipid Measures | Symptom Measures | Controlling for | Relationship between Serum Lipids and Symptoms |
|---|---|---|---|---|---|---|---|---|
| Pande et al., 2002 [9]; Canada; case study | N = 1 (26 yr; male) |
SCZ | Clozapine | 26 weeks | TC, TG | BPRS | N/A | 1) Clozapine-associated increases in TC and TG coincided with symptom improvement. 2) Atorvastatin-associated decreases in TC and TG coincided with relapse. 3) Increases in TG and TC after atorvastatin discontinuation coincided with symptom improvement. |
| Garyfallos et al., 2003 [11]; Greece; prospective study |
N = 50 (mean age range: 31.5 ± 6.1–31.8 ± 6.5 yr; 68% male) |
SCZ spectrum disorders | Olanzapine or risperidone | 8 weeks | TC, TG | PANSS | None | 1) Significant inverse association of Δ TG with Δ PANSS total in olanzapine group (n = 25; r = 0.71, p < 0.001).* 2) No significant relationship in risperidone group (n = 25). |
| Atmaca et al., 2003 [12]; Turkey; prospective study | N = 64 (mean age range: 27.9 ± 7.8–32.1 ± 6.2 yr; 45% male) |
SCZ | Clozapine, olanzapine, quetiapine, or risperidone | 12 months | TG | PANSS | None | 1) Significant inverse association of Δ TG with Δ PANSS total in clozapine (n = 14; r = 0.60, p < 0.05) and olanzapine (n = 14; r = 0.58, p < 0.05) groups.* 2) No significant relationship in quetiapine (n = 14) and risperidone (n = 14) groups. |
| Huang et al., 2005 [13]; Taiwan; prospective study | N = 97 (mean age: 32.3 ± 9.4 yr; 51% male) |
SCZ | FGA: haloperidol, loxapine, or sulpiride; SGA: clozapine, olanzapine or risperidone | 3 weeks | TC, TG, HDL-c, LDL-c, VLDL-c | PANSS (responder: ≥50% reduction in PANSS total score) |
none | 1) Responders to SGAs (n = 32): significant increases in TC, TG, HDL-c, VLDL-c. 2) Responders to FGAs (n = 36): no significant changes in lipid profiles. 3) Nonresponders (n = 29): no significant changes in lipid profiles. |
| Procyshyn et al., 2007 [8]; Canada; RCT |
N = 55 (mean age: 37.06 ± 9.80 yr; 75% male) |
SCZ | Clozapine + risperidone/placebo | 8 weeks | TC, TG | PANSS (responder: ≥20% reduction in PANSS total score) |
Δ body weight | 1) Significant inverse association of Δ TC with Δ PANSS negative (B = -2.36, p = 0.007). 2) Significant inverse associations of Δ TG with Δ PANSS total (B = -3.74, p = 0.037) and negative (B = -1.57, p = 0.017). 3) Responders: significant increases in TC and TG compared with nonresponders. |
| Hermes et al., 2011 [14]; USA; prospective study | N = 865 (mean age: 40.6 ± 11.1 yr; 74% male) |
SCZ | FGA: perphenazine; SGA: olanzapine, quetiapine, risperidone, or ziprasidone | 3 months | TC, TG | PANSS | Baseline serum lipid and PANSS values, antipsychotic use, investigator site, age, duration of illness | No significant relationship between Δ serum lipids and Δ PANSS (data not reported). |
| Lally et al., 2013 [15]; UK, observational study | N = 49 (mean age: 37.4 ± 9.3 yr; 67% male) |
SCZ | Clozapine | 6 months | TC, TG, HDL-c, LDL-c | PANSS | Δ body weight, Δ waist circumference, Δ serum clozapine levels | Significant inverse association of Δ TG with Δ PANSS total (B = 9.33, p < 0.001), positive (B = 2.85, p = 0.001), and negative (B = 1.93, p = 0.02).* |
| Publication, Country, Study Design | Demographics | Diagnosis | Antipsychotic Medication | Follow-up Duration | Serum Lipid Measures | Symptom Measures | Controlling for | Relationship between Serum Lipids and Symptoms |
| Terevnikov et al., 2013 [16]; Finalnd; RCT |
N = 36 (mean age range: 43.4–48.2 yr; 51% male) |
SCZ | FGA (various) + mirtazapine/placebo | Weeks 0–6 (Phase I); Weeks 6–12 (Phase II) | TC | PANSS | Duration of illness, antipsychotic dose, Δ body weight, Δ fasting glucose | 1) Significant inverse association of Δ TC (increase of 1 mmol/L) with Δ PANSS total (reduction of 7 points, p = 0.001), positive (reduction of 1.7 points, p = 0.03), negative (reduction of 1.8 points, p = 0.004) in FGA-mirtazapine group. 2) No significant relationship in FGA-placebo group (based on linear regression). 3) No significant relationship in either group in Phase II. |
| Chen et al., 2014 [17]; Taiwan | N = 372 (mean age: 49.2 ± 9.7 yr; 75% male) |
SCZ | FGA/SGA | 2 years | TC, TG, HDL-c, LDL-c | PANSS | Age, sex, antipsychotic use | 1) Significant inverse associations of Δ TG with baseline PANSS total (Estimate = -0.60, p = 0.007) and PANSS negative (Estimate = -1.74, p < 0.001) 2) Significant positive association of Δ HDL-c with baseline PANSS negative (Estimate = 0.21, p = 0.004). |
| Sharma et al., 2014 [18]; India; prospective study | N = 100 (mean age range: 28.1 ± 7.07–31.3 ± 8.1 yr; 46% male) |
SCZ spectrum disorders | SGA (≥ 82%): risperidone (65.3%), olanzapine (16.7%), other (18.1%) | 2–4 weeks (first follow-up); 8–12 weeks (second follow-up) | TC, TG, HDL-c, LDL-c, VLDL-c | BPRS | Δ appetite | 1) Significant inverse association of Δ TG with Δ BPRS at first follow-up (R2 = 0.16, p = 0.001). 2) Significant inverse association of Δ TG at first follow-up with Δ BPRS at second follow-up (R2 = 0.17, p = 0.009). |
| Chukhin et al., 2016 [19]; Finland; RCT | N = 35 (mean age: 37.3 ± 18.3 yr; 100% male) |
97% SCZ | Clozapine/olanzapine with orlistat/placebo | 16 weeks | TC, TG, HDL-c, LDL-c | PANSS | None | No significant relationship between Δ serum lipids and Δ PANSS in either orlistat (n = 18) or placebo (n = 17) group (data not reported). |
| Solberg et al., 2016 [20]; Norway; naturalistic study | N = 55 (mean age: 26.5 ± 6.1 yr; 69% male) |
SCZ or SAD | Unmedicated, n = 11; FGA, n = 5; SGA, n = 39 | 5 years | TC, TG | PANSS | None | Significant positive association of baseline TG with follow-up PANSS total (r = 0.28, p = 0.04). |
| Gjerde et al., 2017 [21]; Norway; prospective study | N = 132 (mean age: 26.7 ± 7.6 yr; 64% male) |
FEP | SGA (88%; various) with/without FGA (16%; various) | 12 months | TC, TG, HDL-c, LDL-c | PANSS | Age, sex, antipsychotic use, BMI | Significant inverse association of Δ HDL-c with Δ PANSS negative (B = -0.54, p = 0.02). |
| Luckhoff et al., 2018 [22]; South Africa; longitudinal study | N = 106 (mean age: 24.2 yr; 73% male) |
FEP | Flupenthixol decanoate (100%) | 12 months | TC, TG, HDL-c, LDL-c | PANSS | Age, sex, ethnicity, substance use, antipsychotic dose, treatment duration, other co-medications | 1) Significant inverse association of Δ TG with Δ disorganized symptoms (r = -0.29; p = 0.040). 2) Δ TG was not a significant predictor in a linear regression model. |
Abbreviations: BMI: body mass index; BPRS: Brief Psychiatric Rating Scale; FEP: first-episode psychosis FGA: first-generation antipsychotic; HDL-c: high-density lipoprotein cholesterol; LDL-c: low-density lipoprotein cholesterol; PANSS: Positive and Negative Syndrome Scale; SAD: schizoaffective disorder; SCZ: schizophrenia; SGA: second-generation antipsychotic; TC: total cholesterol; TG: triglycerides; VLDL-c: very low-density lipoprotein cholesterol. *PANSS reduction was included in the analyses as clinical improvement.