Acta Psychiatr Scand 2015: 132: 293–300 All rights reserved DOI: 10.1111/acps.12388
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA PSYCHIATRICA SCANDINAVICA
Association between serum lipids and membrane fatty acids and clinical characteristics in patients with schizophrenia Solberg DK, Bentsen H, Refsum H, Andreassen OA. Association between serum lipids and membrane fatty acids and clinical characteristics in patients with schizophrenia. Objective: Earlier reports indicate that patients with schizophrenia have altered lipid levels in serum and cell membranes. The purpose of this study was to determine the relationship between clinical characteristics and serum and membrane lipids. Method: Fifty-five patients with schizophrenia and 51 healthy controls were included. The patients were characterized with Positive and Negative Syndrome Scale (PANSS) and Global Assessment of Functioning (GAF). Serum lipids [high- and low-density lipoprotein cholesterol (HDL, LDL) and triglyceride (TG)] and erythrocyte polyunsaturated fatty acids (PUFA) were measured. Results: Among the participants with schizophrenia, there was a significant correlation between serum triglyceride levels and PANSSpositive symptoms (r = 0.28, P = 0.04), GAF-S (r = 0.48, P = 0.001) and GAF-F (r = 0.32, P = 0.01), and between HDL level and GAF-S (r = 0.37, P = 0.008) and GAF-F (r = 0.28, P = 0.04). Long-chain PUFA were significantly associated with PANSS-negative symptoms (r = 0.52, P < 0.001), GAF-S (r = 0.32, P = 0.02), and GAF-F (r = 0.29, P = 0.04). The patients with schizophrenia had significantly higher TG (P < 0.001) and lower HDL (P < 0.001) levels than healthy controls. HDL was also lower in the subgroup (n = 11) not receiving antipsychotic medication (P = 0.02). Conclusion: The results suggest associations between lipid profile and clinical characteristics. This may indicate a role for lipid biology in schizophrenia pathophysiology.
D. K. Solberg1,2, H. Bentsen2,
H. Refsum2, O. A. Andreassen3,4 1 Institute for Military Psychiatry, Norwegian Defense Medical Services, 2Center for Psychopharmacology, Diakonhjemmet Hospital, 3NORMENT, KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, University of Oslo, and 4Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway
Key words: schizophrenia; membrane lipids; dyslipidemia; symptom levels Dag Kristen Solberg, Institute for Military Psychiatry, Norwegian Defense Medical Services, Pb 1550 Sentrum, 0015 Oslo, Norway. E-mail: [email protected]
Accepted for publication December 10, 2014
Significant outcomes
• Higher serum lipid levels are associated with more severe positive psychotic symptoms (triglycerides)
and worse global functioning (triglycerides, HDL cholesterol) in schizophrenia. levels of polyunsaturated membrane fatty acids are associated with more severe negative symptoms and worse functioning in patients with schizophrenia. Patients with schizophrenia have higher levels of serum triglycerides and lower levels of serum HDL cholesterol than healthy controls.
• Higher •
Limitations
• The cross-sectional design of the study is not fit for making conclusions about causality. • The fluctuation of symptoms of schizophrenia over time makes it difficult to detect associations.
293
Solberg et al. Introduction
Schizophrenia is a severe mental illness, characterized by hallucinations and delusions, negative symptoms, and cognitive dysfunction (1, 2). Schizophrenia is regarded as a complex syndrome of developmental defects caused by genetic and environmental factors (3), but the underlying pathological mechanisms are still largely unknown. Patients with schizophrenia have a higher mortality rate than the rest of the population, with standardized mortality rates about 2.0 (4, 5). This leads to reduction in life span, estimated to approximately 20 years, which mainly are caused by cardiovascular disease (CVD) (6). An important CVD risk factor is the metabolic syndrome, which has an increased prevalence in schizophrenia (7, 8). Lipid abnormalities are important factors in the metabolic syndrome (9). Most studies show that patients with schizophrenia have higher levels of serum lipids (cholesterol and triglycerides) than in a healthy population (10, 11). This dyslipidemia has been regarded as a result of antipsychotic medication and lifestyle factors, as it is known that the use of antipsychotic medication, and especially second-generation antipsychotics, increases the risk of metabolic syndrome (12–14). Increased oxidative stress has been associated with several psychiatric diseases, including schizophrenia, as well as in cardiovascular disease, and may be a common predisposing factor for both CVD and psychiatric illness. (15–17). Dyslipidemia have also been demonstrated in untreated patients with schizophrenia (18–20). Further, several lines of evidence have implicated abnormal lipid metabolism in schizophrenia etiology. In a general population study of the Northern Finland 1966 birth-cohort, schizophrenia patients with early onset had higher levels of triglycerides than patients with later onset (21), suggesting a role for lipid metabolism in disease development. Further, findings that serum lipids variations are associated with clinical characteristics, symptom intensity, and treatment prognosis may also suggest a role of lipids in schizophrenia pathology (22). In a study of chronically ill patients with schizophrenia, the levels of triglycerides and cholesterol were negatively related to negative symptoms severity (23). However, the nature of the relation between lipid metabolism and core clinical characteristics in schizophrenia remains mainly unknown. Changes in lipid biosynthesis might not only represent adverse effects of antipsychotic drugs, but can also be beneficial (24). An association between serum cholesterol levels and cognitive functioning was found in schizophrenia patients 294
treated with olanzapine, clozapine, and haloperidol (25). Moreover, an increase in serum triglycerides was reported to be associated with clinical response to treatment with clozapine (26, 27). Cholesterol is essential for the structural integrity of cell membranes (28). Thus, increased lipid levels, possibly induced by increased expression of lipid biosynthesis and lipid transport genes (29), may be a result of treatment with antipsychotic drugs, indistinguishable from the beneficial clinical effects. Altered metabolism of membrane lipids is another aspect of lipid biology that has been suggested to be involved in schizophrenia pathophysiology (30). Lower levels of polyunsaturated fatty acids (PUFA) in cell membranes have been found in schizophrenia (31, 32). These changes have been present both in the acute and chronic phase of the disease (33). Both antipsychotic-na€ıve and patients treated with antipsychotics have been shown to have lower PUFA levels than healthy controls, suggesting an increased risk of metabolic problems (34). PUFA levels have also been shown to correlate with the treatment effect of antipsychotic medication (35). Finally, there may be links between serum dyslipidemia and changes in membrane fatty acids in patients with schizophrenia. Levels of serum triglycerides have been shown to be negatively associated with levels of PUFA in red blood cells (36). However, the clinical significance of altered membrane lipid metabolism and dyslipidemia in patients with schizophrenia is not yet clear. The relationship between serum lipids and membrane lipids has not yet been extensively studied in schizophrenia. Aims of the study
The purpose of this study was to determine the relationship between lipid metabolism and core clinical characteristics (positive and negative symptoms) in patients with schizophrenia, using both measures of serum and membrane lipids. We hypothesized that levels of membrane polyunsaturated fatty acids (PUFA) are lower and serum lipids (TG, LDL, HDL,) higher in patients than in healthy controls. Further that higher levels of membrane PUFA and higher levels of serum lipids are associated with less severe symptom load. As a secondary analysis, we also explored the different lipid measures in unmedicated vs. medicated patients with schizophrenia. Material and methods
In a cross-sectional study, socioeconomic, clinical, and biological data were gathered from a group of
Lipid profile and clinical characteristics patients with schizophrenia and schizoaffective disorders (n = 55) and healthy controls (n = 51). Of the patients, 44 were diagnosed with schizophrenia and 11 with schizoaffective disorder. The patients were recruited from an established cohort of patients admitted to psychiatric emergency wards in southern Norway in 2001–2003 (31). The patients and healthy controls all gave informed consent. Subjects with substance dependence were not included in this study. The patients were examined in 2006–2008. At the time of examination, they were in a chronic/stable phase of the disease, either being out-patients or patients at psychiatric long time care facilities. Healthy controls were recruited among hospital employees, matched for age. The healthy controls had no mental disorder according to the Mini International Neuropsychiatric Interview (MINI). This study was approved by the Regional ethical committee. Clinical assessment
All patients were diagnosed with the Structural Clinical Interview for DSM-IV (SCID). To measure the severity of symptoms, the Positive and Negative Syndrome Scale (PANSS), Structured Interview Version, and Global Assessment of Functioning (GAF), were used. PANSS results are reported in a five-factor model (37). The 30 items were rated from 1 (best) to 7 (worst). Global Assessment of Functioning (GAF) split version was used which includes Symptom (GAF-S) and Functioning (GAF-F) scale, both with ratings from 0 (worst) to 100 (best) (38). All patients were interviewed by the same clinical investigator (DKS). Biochemical assays
Blood for lipid analyses was sampled after overnight fasting. Cholesterol and triglycerides were analyzed at the Department of Clinical Chemistry at Diakonhjemmet Hospital with standard enzymatic methods from Roche Diagnostics Norge AS, Oslo, Norway. For analyses of polyunsaturated fatty acids, washed blood cells were stored at 80° C and sent within 3 months in dry ice to Mylnefield Research Services LTD, Dundee, UK, who did the analysis. The lipids were extracted, converted into fatty acid methyl esters, and analyzed by gas chromatography, yielding fatty acid profiles. In total, 28 species of fatty acids from C14:0 to C24:1 were reported as micrograms per gram of RBC (red blood cells). The sum of omega-3 fatty acids was C18:3 + C18:4 + C20:3 + C20:5 + C22:5 + C22:6. The sum of omega-6 fatty acids was C18:2 + C18:3 + C20:2 + C20:3 + C20:4 + C22:4 +
C22:5. The sum of omega-3 and omega-6 is named polyunsaturated fatty acids (PUFA). The sum of PUFAs with 20 or 22 carbon atoms is named longchain polyunsaturated fatty acids (LCPUFA). Use of antipsychotic medication
Use of antipsychotic medication is thought to affect the levels of serum lipids. The current and previous uses of antipsychotics among the 55 patients with schizophrenia were registered. Eleven patients did not use any antipsychotic medication at inclusion. The duration of non-use was not measured. All patients complied with antipsychotic drug treatment, defined by measuring detectable levels of antipsychotic drug in serum samples drawn for therapeutic drug monitoring. Use of nutritional supplements
Dietary intake of lipids, including supplements, affects the levels of serum lipids and the constitution of the phospholipids in cell membranes. Use of fatty acid supplements the previous 3 months were registered in both patients with schizophrenia and healthy controls. The dose of supplements was not included because of inaccurate estimates. Statistics
All statistical analyses were performed using SPSS version 20 (SPSS Inc, Chicago, IL, USA / IBM, New York, USA). Demographical and clinical variables are presented as mean values or proportions. Parametric or nonparametric tests were chosen depending on the distribution of variables. The Mann–Whitney test was used to analyze lipid and clinical data that did not follow normal distribution, otherwise Student’s t-test was used. Twosided tests were used, and the significance levels were set to P < 0.05. Spearman’s correlation coefficients were used to evaluate the relationship between lipid data and clinical symptoms (PANSS and GAF). Multivariate linear regression analyses were used to investigate the effect of patient status, smoking status, and demographic variables (sex and age) on lipid measures. Serum triglyceride was ln-transformed because residuals in the models were nonnormally distributed. Results
Fifty-five patients with schizophrenia participated in the study. Demographics of patients and healthy controls are described in Table 1. 295
Solberg et al. Table 1. Demographics Patients (n = 55)
Healthy controls (n = 51)
Table 3. Association between lipid levels in serum and cell membranes and clinical characteristics GAF-S
Age PANSS GAF-S GAF-F
Mean (SD)
Range
Mean (SD)
Range
31.3 (5.7) 81.4 (24.1) 47.7 (12.5) 49.2 (12.7)
21–45 34–122 28–78 28–78
33.0 (5.7)
23–46
n Sex (males) Smokers Education Primary School High School University/college
38 29 n = 53 9 30 14
% 69.1 52.7 17 56.6 26.4
n
%
28 9 n = 50 1 8 41
54.9 17.6 2 16 82
PANSS, Positive and Negative Syndrome Scale; GAF, Global Assessment of Functioning (S: Symptoms, F: Functioning); SD, standard deviation.
Serum lipids
The levels of triglycerides were significantly higher among patients with schizophrenia than in healthy controls (P < 0.001). HDL cholesterol levels were significantly lower among patients than in healthy controls (P < 0.001). Further, HDL levels were also significantly lower in the patients with schizophrenia that were not using an antipsychotic medication compared with healthy controls (P = 0.02). The triglyceride levels were also higher in the group not receiving antipsychotic drugs, but this did not reach significance (P = 0.11). Serum total cholesterol did not differ significantly between the groups (P = 0.21) (Table 2). There was no significant difference in serum lipid levels between participants taking and not taking fatty acid supplements, neither among patients with schizophrenia nor healthy controls. Serum triglyceride level was lower among the 11 patients with schizoaffective disorder than patients with schizophrenia Table 2. Lipid levels in serum and cell membranes in patients and healthy controls
Total cholesterol LDL cholesterol HDL cholesterol Triglycerides RBC PUFA RBC LCPUFA
All patients
Unmedicated pts.
Medicated pts.
5.36 (1.20) 3.33 (1.10) 1.23 (0.37)*** 1.33 (1.71)*** 471 (73) 308 (42)
4.95 (1.53) 3.03 (1.53) 1.33 (0.48)* 1.02 (1.42) 477 (93) 300 (47)
5.47 (1.10)* 3.41 (0.97)* 1.21 (0.34)*** 1.37 (1.77)*** 471 (64) 312 (39)
Healthy controls 5.05 (0.84) 2.70 (0.84) 1.63 (0.36) 0.85 (0.53) 470 (69) 307 (49)
*0.01 < P ≤ 0.05, **0.001 < P ≤ 0.01, ***P ≤ 0.001 vs. healthy controls. Cholesterol: normally distributed (mean, standard deviation, Student’s t-test). Triglycerides, PUFA/LCPUFA: non-normally distributed (median, interquartile range, Mann –Whitney test). LCPUFA = omega-3 + omega-6 with 20 or more carbon atoms. PUFA = omega-3 + omega 6 polyunsaturated fatty acids. Cholesterol, triglycerides: mmol/L, PUFA/LCPUFA: lg/g RBC.
296
Triglycerides Total cholesterol HDL cholesterol LDL cholesterol SumPUFA LCPUFA
0.48*** 0.30* 0.37** 0.28* 0.22 0.32*
GAF-F 0.32* 0.23 0.28* 0.25 0.21 0.29*
PANSS Pos 0.28* 0.22 0.13 0.18 0.01 0.11
PANSS Neg 0.26 0.11 0.22 0.12 0.32* 0.52***
PANSS Tot 0.26 0.15 0.19 0.14 0.14 0.31*
*0.01 < P ≤ 0.05, **0.001 < P ≤ 0.01, ***P ≤ 0.001. Spearman’s correlation coefficient is reported. PANSS, Positive and Negative Syndrome Scale; GAF, Global Assessment of Functioning (S: Symptoms, F: Functioning). PANSS-positive Component: items P1+P3+ P5+P6+G9. PANSS-negative Component: items N1+N2+N3+N4+N6+G7+G8+G16. Sum PUFA = omega-3 + omega- 6 polyunsaturated fatty acids in RBC. LCPUFA = omega-3 + omega-6 PUFA with 20 or more carbon atoms, measured in RBC.
(1.17 vs. 2.00, P = 0.006). There was no difference between patients with schizophrenia and schizoaffective disorder regarding HDL-, LDL-, or totalcholesterol. Serum triglyceride levels were negatively correlated with functioning measured by GAF-F (r = 0.32; P = 0.01), as well as general symptoms levels as measured by GAF-S (r = 0.48, P = 0.001), while they were positively correlated with core psychosis symptoms as measured by PANSS positive (r = 0.28, P = 0.04). There was also a trend for positive correlation with severity of other key symptom domains (PANSS negative, r = 0.26, P = 0.06 and PANSS total; r = 0.26, P = 0.06). For total cholesterol, a significant correlation between serum levels and symptom levels was only observed for GAF-S (r = 0.30 P = 0.02). There was a positive correlation between HDL cholesterol and GAF-F (r = 0.28, P = 0.045) and GAF-S (r = 0.37, P = 0.008), while no significant correlations with PANSS scores were observed. LDL cholesterol was significantly correlated only with GAF-S (r = 0.28, P = 0.05). See Table 3 for more details. Membrane lipids
There were no significant differences between patients with schizophrenia and healthy controls regarding levels of PUFA or LCPUFA (Table 2). The levels of red blood cell (RBC) omega-3 fatty acids were higher among patients with schizophrenia and healthy controls taking fatty acid supplements (125.2 vs. 102.4 µg per g of RBC, P = 0.006, and 122.2 vs. 110.4, P = 0.04 respectively). Fatty acid supplement did not influence the levels of total PUFA or LCPUFA in either patients with schizophrenia or healthy controls.
Lipid profile and clinical characteristics In the patient group, there were significant associations between symptom severity levels and PUFA concentrations (Table 3), with significantly positive correlation between LCPUFA and PANSS-negative symptoms (r = 0.52, P = 0.001) and PUFA (r = 0.32, P = 0.02), and LCPUFA and PANSS total (r = 0.31, P = 0.02). There were also significant negative correlations between LCPUFA and general symptom and function levels, as measured with GAF-S ( 0.32, P = 0.02) and GAF-F ( 0.29 P = 0.04). No significant association between membrane lipids (PUFA and LCPUFA) and serum lipids (triglycerides and cholesterol) were found, neither for patients with schizophrenia, healthy controls nor for the entire sample (data not shown). Medication
Eleven patients with schizophrenia used no antipsychotic medication. Five patients used a first-generation antipsychotic, while 39 used a second-generation antipsychotic. The unmedicated patient group showed significant lower serum levels of HDL cholesterol (P = 0.02) and non-significant higher levels of triglycerides (P = 0.11) than healthy controls. No significant differences were observed for the other serum lipids. There were no significant differences in membrane lipid levels between the medicated and unmedicated groups (Table 2). Further, there was no significant difference in symptom intensity, measured by PANSS, between the two groups. The medicated patients had more severe symptom levels measured by GAF-S (45.7 vs. 55.1, P = 0.03). No significant difference in functioning, measured by GAF-F, was found. Regression model
In a multivariate linear regression analysis, serum triglyceride and total cholesterol, respectively, were dependent variables, while gender, smoking habits, age at inclusion, patient status, and use/non-use of fatty acid supplement were independent variables (Table 4a). Serum triglyceride was ln-transformed because residuals in the models were non-normally distributed. In the total sample of patients and controls, male sex and being a patient were associated with higher serum triglyceride levels. No other variables were significantly associated with triglyceride levels. Male sex and higher age were associated with serum total cholesterol. No other variables were associated with serum total cholesterol. To estimate the effect of antipsychotic medication, all patients using such drugs (n = 44) were removed
Table 4. Estimated effect of sex, patient status, age, fatty acid supplements, and smoking on lipid levels. (a) Patients and healthy controls. (b) Unmedicated patients and healthy controls Variable (a) Constant Sex Patient Age Model fit (R2) (b) Constant Sex Patient Age Model fit (R2)
Ln S-Triglyceride
P-value
0.52 ( 1.14, 0.11) 0.31 ( 0.53, 0.09) 0.57 (0.33, 0.80) 0.016 ( 0.003, 0.034) 0.29
0.10 0.007
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd ACTA PSYCHIATRICA SCANDINAVICA
Association between serum lipids and membrane fatty acids and clinical characteristics in patients with schizophrenia Solberg DK, Bentsen H, Refsum H, Andreassen OA. Association between serum lipids and membrane fatty acids and clinical characteristics in patients with schizophrenia. Objective: Earlier reports indicate that patients with schizophrenia have altered lipid levels in serum and cell membranes. The purpose of this study was to determine the relationship between clinical characteristics and serum and membrane lipids. Method: Fifty-five patients with schizophrenia and 51 healthy controls were included. The patients were characterized with Positive and Negative Syndrome Scale (PANSS) and Global Assessment of Functioning (GAF). Serum lipids [high- and low-density lipoprotein cholesterol (HDL, LDL) and triglyceride (TG)] and erythrocyte polyunsaturated fatty acids (PUFA) were measured. Results: Among the participants with schizophrenia, there was a significant correlation between serum triglyceride levels and PANSSpositive symptoms (r = 0.28, P = 0.04), GAF-S (r = 0.48, P = 0.001) and GAF-F (r = 0.32, P = 0.01), and between HDL level and GAF-S (r = 0.37, P = 0.008) and GAF-F (r = 0.28, P = 0.04). Long-chain PUFA were significantly associated with PANSS-negative symptoms (r = 0.52, P < 0.001), GAF-S (r = 0.32, P = 0.02), and GAF-F (r = 0.29, P = 0.04). The patients with schizophrenia had significantly higher TG (P < 0.001) and lower HDL (P < 0.001) levels than healthy controls. HDL was also lower in the subgroup (n = 11) not receiving antipsychotic medication (P = 0.02). Conclusion: The results suggest associations between lipid profile and clinical characteristics. This may indicate a role for lipid biology in schizophrenia pathophysiology.
D. K. Solberg1,2, H. Bentsen2,
H. Refsum2, O. A. Andreassen3,4 1 Institute for Military Psychiatry, Norwegian Defense Medical Services, 2Center for Psychopharmacology, Diakonhjemmet Hospital, 3NORMENT, KG Jebsen Centre for Psychosis Research, Institute of Clinical Medicine, University of Oslo, and 4Division of Mental Health and Addiction, Oslo University Hospital, Oslo, Norway
Key words: schizophrenia; membrane lipids; dyslipidemia; symptom levels Dag Kristen Solberg, Institute for Military Psychiatry, Norwegian Defense Medical Services, Pb 1550 Sentrum, 0015 Oslo, Norway. E-mail: [email protected]
Accepted for publication December 10, 2014
Significant outcomes
• Higher serum lipid levels are associated with more severe positive psychotic symptoms (triglycerides)
and worse global functioning (triglycerides, HDL cholesterol) in schizophrenia. levels of polyunsaturated membrane fatty acids are associated with more severe negative symptoms and worse functioning in patients with schizophrenia. Patients with schizophrenia have higher levels of serum triglycerides and lower levels of serum HDL cholesterol than healthy controls.
• Higher •
Limitations
• The cross-sectional design of the study is not fit for making conclusions about causality. • The fluctuation of symptoms of schizophrenia over time makes it difficult to detect associations.
293
Solberg et al. Introduction
Schizophrenia is a severe mental illness, characterized by hallucinations and delusions, negative symptoms, and cognitive dysfunction (1, 2). Schizophrenia is regarded as a complex syndrome of developmental defects caused by genetic and environmental factors (3), but the underlying pathological mechanisms are still largely unknown. Patients with schizophrenia have a higher mortality rate than the rest of the population, with standardized mortality rates about 2.0 (4, 5). This leads to reduction in life span, estimated to approximately 20 years, which mainly are caused by cardiovascular disease (CVD) (6). An important CVD risk factor is the metabolic syndrome, which has an increased prevalence in schizophrenia (7, 8). Lipid abnormalities are important factors in the metabolic syndrome (9). Most studies show that patients with schizophrenia have higher levels of serum lipids (cholesterol and triglycerides) than in a healthy population (10, 11). This dyslipidemia has been regarded as a result of antipsychotic medication and lifestyle factors, as it is known that the use of antipsychotic medication, and especially second-generation antipsychotics, increases the risk of metabolic syndrome (12–14). Increased oxidative stress has been associated with several psychiatric diseases, including schizophrenia, as well as in cardiovascular disease, and may be a common predisposing factor for both CVD and psychiatric illness. (15–17). Dyslipidemia have also been demonstrated in untreated patients with schizophrenia (18–20). Further, several lines of evidence have implicated abnormal lipid metabolism in schizophrenia etiology. In a general population study of the Northern Finland 1966 birth-cohort, schizophrenia patients with early onset had higher levels of triglycerides than patients with later onset (21), suggesting a role for lipid metabolism in disease development. Further, findings that serum lipids variations are associated with clinical characteristics, symptom intensity, and treatment prognosis may also suggest a role of lipids in schizophrenia pathology (22). In a study of chronically ill patients with schizophrenia, the levels of triglycerides and cholesterol were negatively related to negative symptoms severity (23). However, the nature of the relation between lipid metabolism and core clinical characteristics in schizophrenia remains mainly unknown. Changes in lipid biosynthesis might not only represent adverse effects of antipsychotic drugs, but can also be beneficial (24). An association between serum cholesterol levels and cognitive functioning was found in schizophrenia patients 294
treated with olanzapine, clozapine, and haloperidol (25). Moreover, an increase in serum triglycerides was reported to be associated with clinical response to treatment with clozapine (26, 27). Cholesterol is essential for the structural integrity of cell membranes (28). Thus, increased lipid levels, possibly induced by increased expression of lipid biosynthesis and lipid transport genes (29), may be a result of treatment with antipsychotic drugs, indistinguishable from the beneficial clinical effects. Altered metabolism of membrane lipids is another aspect of lipid biology that has been suggested to be involved in schizophrenia pathophysiology (30). Lower levels of polyunsaturated fatty acids (PUFA) in cell membranes have been found in schizophrenia (31, 32). These changes have been present both in the acute and chronic phase of the disease (33). Both antipsychotic-na€ıve and patients treated with antipsychotics have been shown to have lower PUFA levels than healthy controls, suggesting an increased risk of metabolic problems (34). PUFA levels have also been shown to correlate with the treatment effect of antipsychotic medication (35). Finally, there may be links between serum dyslipidemia and changes in membrane fatty acids in patients with schizophrenia. Levels of serum triglycerides have been shown to be negatively associated with levels of PUFA in red blood cells (36). However, the clinical significance of altered membrane lipid metabolism and dyslipidemia in patients with schizophrenia is not yet clear. The relationship between serum lipids and membrane lipids has not yet been extensively studied in schizophrenia. Aims of the study
The purpose of this study was to determine the relationship between lipid metabolism and core clinical characteristics (positive and negative symptoms) in patients with schizophrenia, using both measures of serum and membrane lipids. We hypothesized that levels of membrane polyunsaturated fatty acids (PUFA) are lower and serum lipids (TG, LDL, HDL,) higher in patients than in healthy controls. Further that higher levels of membrane PUFA and higher levels of serum lipids are associated with less severe symptom load. As a secondary analysis, we also explored the different lipid measures in unmedicated vs. medicated patients with schizophrenia. Material and methods
In a cross-sectional study, socioeconomic, clinical, and biological data were gathered from a group of
Lipid profile and clinical characteristics patients with schizophrenia and schizoaffective disorders (n = 55) and healthy controls (n = 51). Of the patients, 44 were diagnosed with schizophrenia and 11 with schizoaffective disorder. The patients were recruited from an established cohort of patients admitted to psychiatric emergency wards in southern Norway in 2001–2003 (31). The patients and healthy controls all gave informed consent. Subjects with substance dependence were not included in this study. The patients were examined in 2006–2008. At the time of examination, they were in a chronic/stable phase of the disease, either being out-patients or patients at psychiatric long time care facilities. Healthy controls were recruited among hospital employees, matched for age. The healthy controls had no mental disorder according to the Mini International Neuropsychiatric Interview (MINI). This study was approved by the Regional ethical committee. Clinical assessment
All patients were diagnosed with the Structural Clinical Interview for DSM-IV (SCID). To measure the severity of symptoms, the Positive and Negative Syndrome Scale (PANSS), Structured Interview Version, and Global Assessment of Functioning (GAF), were used. PANSS results are reported in a five-factor model (37). The 30 items were rated from 1 (best) to 7 (worst). Global Assessment of Functioning (GAF) split version was used which includes Symptom (GAF-S) and Functioning (GAF-F) scale, both with ratings from 0 (worst) to 100 (best) (38). All patients were interviewed by the same clinical investigator (DKS). Biochemical assays
Blood for lipid analyses was sampled after overnight fasting. Cholesterol and triglycerides were analyzed at the Department of Clinical Chemistry at Diakonhjemmet Hospital with standard enzymatic methods from Roche Diagnostics Norge AS, Oslo, Norway. For analyses of polyunsaturated fatty acids, washed blood cells were stored at 80° C and sent within 3 months in dry ice to Mylnefield Research Services LTD, Dundee, UK, who did the analysis. The lipids were extracted, converted into fatty acid methyl esters, and analyzed by gas chromatography, yielding fatty acid profiles. In total, 28 species of fatty acids from C14:0 to C24:1 were reported as micrograms per gram of RBC (red blood cells). The sum of omega-3 fatty acids was C18:3 + C18:4 + C20:3 + C20:5 + C22:5 + C22:6. The sum of omega-6 fatty acids was C18:2 + C18:3 + C20:2 + C20:3 + C20:4 + C22:4 +
C22:5. The sum of omega-3 and omega-6 is named polyunsaturated fatty acids (PUFA). The sum of PUFAs with 20 or 22 carbon atoms is named longchain polyunsaturated fatty acids (LCPUFA). Use of antipsychotic medication
Use of antipsychotic medication is thought to affect the levels of serum lipids. The current and previous uses of antipsychotics among the 55 patients with schizophrenia were registered. Eleven patients did not use any antipsychotic medication at inclusion. The duration of non-use was not measured. All patients complied with antipsychotic drug treatment, defined by measuring detectable levels of antipsychotic drug in serum samples drawn for therapeutic drug monitoring. Use of nutritional supplements
Dietary intake of lipids, including supplements, affects the levels of serum lipids and the constitution of the phospholipids in cell membranes. Use of fatty acid supplements the previous 3 months were registered in both patients with schizophrenia and healthy controls. The dose of supplements was not included because of inaccurate estimates. Statistics
All statistical analyses were performed using SPSS version 20 (SPSS Inc, Chicago, IL, USA / IBM, New York, USA). Demographical and clinical variables are presented as mean values or proportions. Parametric or nonparametric tests were chosen depending on the distribution of variables. The Mann–Whitney test was used to analyze lipid and clinical data that did not follow normal distribution, otherwise Student’s t-test was used. Twosided tests were used, and the significance levels were set to P < 0.05. Spearman’s correlation coefficients were used to evaluate the relationship between lipid data and clinical symptoms (PANSS and GAF). Multivariate linear regression analyses were used to investigate the effect of patient status, smoking status, and demographic variables (sex and age) on lipid measures. Serum triglyceride was ln-transformed because residuals in the models were nonnormally distributed. Results
Fifty-five patients with schizophrenia participated in the study. Demographics of patients and healthy controls are described in Table 1. 295
Solberg et al. Table 1. Demographics Patients (n = 55)
Healthy controls (n = 51)
Table 3. Association between lipid levels in serum and cell membranes and clinical characteristics GAF-S
Age PANSS GAF-S GAF-F
Mean (SD)
Range
Mean (SD)
Range
31.3 (5.7) 81.4 (24.1) 47.7 (12.5) 49.2 (12.7)
21–45 34–122 28–78 28–78
33.0 (5.7)
23–46
n Sex (males) Smokers Education Primary School High School University/college
38 29 n = 53 9 30 14
% 69.1 52.7 17 56.6 26.4
n
%
28 9 n = 50 1 8 41
54.9 17.6 2 16 82
PANSS, Positive and Negative Syndrome Scale; GAF, Global Assessment of Functioning (S: Symptoms, F: Functioning); SD, standard deviation.
Serum lipids
The levels of triglycerides were significantly higher among patients with schizophrenia than in healthy controls (P < 0.001). HDL cholesterol levels were significantly lower among patients than in healthy controls (P < 0.001). Further, HDL levels were also significantly lower in the patients with schizophrenia that were not using an antipsychotic medication compared with healthy controls (P = 0.02). The triglyceride levels were also higher in the group not receiving antipsychotic drugs, but this did not reach significance (P = 0.11). Serum total cholesterol did not differ significantly between the groups (P = 0.21) (Table 2). There was no significant difference in serum lipid levels between participants taking and not taking fatty acid supplements, neither among patients with schizophrenia nor healthy controls. Serum triglyceride level was lower among the 11 patients with schizoaffective disorder than patients with schizophrenia Table 2. Lipid levels in serum and cell membranes in patients and healthy controls
Total cholesterol LDL cholesterol HDL cholesterol Triglycerides RBC PUFA RBC LCPUFA
All patients
Unmedicated pts.
Medicated pts.
5.36 (1.20) 3.33 (1.10) 1.23 (0.37)*** 1.33 (1.71)*** 471 (73) 308 (42)
4.95 (1.53) 3.03 (1.53) 1.33 (0.48)* 1.02 (1.42) 477 (93) 300 (47)
5.47 (1.10)* 3.41 (0.97)* 1.21 (0.34)*** 1.37 (1.77)*** 471 (64) 312 (39)
Healthy controls 5.05 (0.84) 2.70 (0.84) 1.63 (0.36) 0.85 (0.53) 470 (69) 307 (49)
*0.01 < P ≤ 0.05, **0.001 < P ≤ 0.01, ***P ≤ 0.001 vs. healthy controls. Cholesterol: normally distributed (mean, standard deviation, Student’s t-test). Triglycerides, PUFA/LCPUFA: non-normally distributed (median, interquartile range, Mann –Whitney test). LCPUFA = omega-3 + omega-6 with 20 or more carbon atoms. PUFA = omega-3 + omega 6 polyunsaturated fatty acids. Cholesterol, triglycerides: mmol/L, PUFA/LCPUFA: lg/g RBC.
296
Triglycerides Total cholesterol HDL cholesterol LDL cholesterol SumPUFA LCPUFA
0.48*** 0.30* 0.37** 0.28* 0.22 0.32*
GAF-F 0.32* 0.23 0.28* 0.25 0.21 0.29*
PANSS Pos 0.28* 0.22 0.13 0.18 0.01 0.11
PANSS Neg 0.26 0.11 0.22 0.12 0.32* 0.52***
PANSS Tot 0.26 0.15 0.19 0.14 0.14 0.31*
*0.01 < P ≤ 0.05, **0.001 < P ≤ 0.01, ***P ≤ 0.001. Spearman’s correlation coefficient is reported. PANSS, Positive and Negative Syndrome Scale; GAF, Global Assessment of Functioning (S: Symptoms, F: Functioning). PANSS-positive Component: items P1+P3+ P5+P6+G9. PANSS-negative Component: items N1+N2+N3+N4+N6+G7+G8+G16. Sum PUFA = omega-3 + omega- 6 polyunsaturated fatty acids in RBC. LCPUFA = omega-3 + omega-6 PUFA with 20 or more carbon atoms, measured in RBC.
(1.17 vs. 2.00, P = 0.006). There was no difference between patients with schizophrenia and schizoaffective disorder regarding HDL-, LDL-, or totalcholesterol. Serum triglyceride levels were negatively correlated with functioning measured by GAF-F (r = 0.32; P = 0.01), as well as general symptoms levels as measured by GAF-S (r = 0.48, P = 0.001), while they were positively correlated with core psychosis symptoms as measured by PANSS positive (r = 0.28, P = 0.04). There was also a trend for positive correlation with severity of other key symptom domains (PANSS negative, r = 0.26, P = 0.06 and PANSS total; r = 0.26, P = 0.06). For total cholesterol, a significant correlation between serum levels and symptom levels was only observed for GAF-S (r = 0.30 P = 0.02). There was a positive correlation between HDL cholesterol and GAF-F (r = 0.28, P = 0.045) and GAF-S (r = 0.37, P = 0.008), while no significant correlations with PANSS scores were observed. LDL cholesterol was significantly correlated only with GAF-S (r = 0.28, P = 0.05). See Table 3 for more details. Membrane lipids
There were no significant differences between patients with schizophrenia and healthy controls regarding levels of PUFA or LCPUFA (Table 2). The levels of red blood cell (RBC) omega-3 fatty acids were higher among patients with schizophrenia and healthy controls taking fatty acid supplements (125.2 vs. 102.4 µg per g of RBC, P = 0.006, and 122.2 vs. 110.4, P = 0.04 respectively). Fatty acid supplement did not influence the levels of total PUFA or LCPUFA in either patients with schizophrenia or healthy controls.
Lipid profile and clinical characteristics In the patient group, there were significant associations between symptom severity levels and PUFA concentrations (Table 3), with significantly positive correlation between LCPUFA and PANSS-negative symptoms (r = 0.52, P = 0.001) and PUFA (r = 0.32, P = 0.02), and LCPUFA and PANSS total (r = 0.31, P = 0.02). There were also significant negative correlations between LCPUFA and general symptom and function levels, as measured with GAF-S ( 0.32, P = 0.02) and GAF-F ( 0.29 P = 0.04). No significant association between membrane lipids (PUFA and LCPUFA) and serum lipids (triglycerides and cholesterol) were found, neither for patients with schizophrenia, healthy controls nor for the entire sample (data not shown). Medication
Eleven patients with schizophrenia used no antipsychotic medication. Five patients used a first-generation antipsychotic, while 39 used a second-generation antipsychotic. The unmedicated patient group showed significant lower serum levels of HDL cholesterol (P = 0.02) and non-significant higher levels of triglycerides (P = 0.11) than healthy controls. No significant differences were observed for the other serum lipids. There were no significant differences in membrane lipid levels between the medicated and unmedicated groups (Table 2). Further, there was no significant difference in symptom intensity, measured by PANSS, between the two groups. The medicated patients had more severe symptom levels measured by GAF-S (45.7 vs. 55.1, P = 0.03). No significant difference in functioning, measured by GAF-F, was found. Regression model
In a multivariate linear regression analysis, serum triglyceride and total cholesterol, respectively, were dependent variables, while gender, smoking habits, age at inclusion, patient status, and use/non-use of fatty acid supplement were independent variables (Table 4a). Serum triglyceride was ln-transformed because residuals in the models were non-normally distributed. In the total sample of patients and controls, male sex and being a patient were associated with higher serum triglyceride levels. No other variables were significantly associated with triglyceride levels. Male sex and higher age were associated with serum total cholesterol. No other variables were associated with serum total cholesterol. To estimate the effect of antipsychotic medication, all patients using such drugs (n = 44) were removed
Table 4. Estimated effect of sex, patient status, age, fatty acid supplements, and smoking on lipid levels. (a) Patients and healthy controls. (b) Unmedicated patients and healthy controls Variable (a) Constant Sex Patient Age Model fit (R2) (b) Constant Sex Patient Age Model fit (R2)
Ln S-Triglyceride
P-value
0.52 ( 1.14, 0.11) 0.31 ( 0.53, 0.09) 0.57 (0.33, 0.80) 0.016 ( 0.003, 0.034) 0.29
0.10 0.007
