A Meta-Analytic Review of Polyunsaturated Fatty Acid Compositions in Patients with Depression
Pao-Yen Lin, Shih-Yi Huang, and Kuan-Pin Su
Background: On the basis of evidence from studies showing the antidepressant effects of omega-3 polyunsaturated fatty acids and the inverse relation between fish consumption and the prevalence of depression, the phospholipid hypothesis seems promising in ascertaining the etiology and treatment of depression. Although several studies have shown lower levels of omega-3 (n-3) polyunsaturated fatty acids in depressive patients, the results of individual polyunsaturated fatty acids, including docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA), and the omega-6 (n-6) polyunsaturated fatty acid arachidonic acid (AA), were inconsistent.
Methods: We conducted the meta-analyses of 14 studies comparing the levels of polyunsaturated fatty acids between depressive patients and control subjects. The effect size of each study was synthesized by using a random effects model.
Results: Compared with control subjects, the levels of EPA, DHA, and total n-3 polyunsaturated fatty acids were significantly lower in depressive patients. There was no significant change in AA or total n-6 polyunsaturated fatty acids.
Conclusions: The results showed lower levels of EPA, DHA, and total n-3 polyunsaturated fatty acids in patients with depression, thus implying that n-3 polyunsaturated fatty acids play a role in the pathogenesis of depression. Our findings provide further support to the phospholipid hypothesis of depression and a rationale for using n-3 polyunsaturated fatty acids as an alternative treatment for depression. With these results, future studies examining specific roles of DHA and EPA in different clusters of depressive symptoms are warranted.
Key Words: Arachidonic acid, depression, docosahexaenoic acid, eicosapentaenoic acid, omega-3, polyunsaturated fatty acids
Major depressive disorder (MDD) is a serious affective illness with a high lifetime prevalence rate (1). The World Health Organization estimated that major depressive disorder will become the second leading cause of disability worldwide by 2020, second to ischemic heart disease, and will be the leading cause in developing regions (2). However, the unmet need of pharmacotherapy and high occurrence of somatic symptoms and physical illness in depression imply that the traditional monoamine hypothesis is not enough to approach the etiology of depression (3). Recently, the phospholipid polyunsaturated fatty acids (PUFAs) hypothesis of depression is providing a promising path to discover a new treatment for depression (4-6). For example, it has been observed that societies with high consumption of fish, which is a good source of omega-3 PUFAs, appear to have a lower prevalence of MDD (7-9). In addition, several clinical trials, if not all (10,11), have shown that omega-3 PUFAs were more effective than placebo (12-16), or as effective as conventional antidepressant medication fluoxetine (17), in treating patients with major depression. In terms of treating special populations with depressive disorders, such as pregnant women (18,19), children (16), patients with bipolar depression (15), and patients with Parkinson's disease (20), the usage of omega-3 PUFAs have been found to be beneficial as well. The mechanism of the antidepressant effect of omega-3
From the Department of Psychiatry (P-YAL), Chang Gung Memorial Hospital, Kaohsiung Medical Center, Chang Gung University College of Medicine, Kaohsiung; School of Nutrition and Health Sciences (S-YH, K-PS), Taipei Medical University, Taipei; and Department of Psychiatryand Mind-Body Research Center (MBI-Lab) (K-PS), Institute of Neural and Cognitive Sciences, China Medical University, Hospital, Taichung, Taiwan.
Address correspondence to Kuan-Pin Su, M.D., Ph.D., Department of Psychiatry, China Medical University Hospital, No. 2 Yuh-Der Road, Taichung 404, Taiwan. E-mail: cobolsu@gmail.com.
Received Oct 6,2009; revised Mar 9,2010; accepted Mar 10,2010.
PUFAs has yet to be elucidated. It has been suggested that the abnormal cell membrane fatty acid composition may be of etiologic significance in depression (4,21). The PUFAs are classified mainly into n-3 (or omega-3) and n-6 (or omega-6) groups. EPA and docosahexaenoic acid (DHA), the major bioactive components of n-3 PUFAs, are not synthesized in the human body and should be obtained directly from the diet, particularly in fatty fish (22). The main n-3 PUFAs in the brain is DHA, comprising up to 10% to 20% of total fatty acids composition in the brain, whereas the n-3 PUFAs a-linolenic acid (ALA), EPA, and docosapentaenoic acid (DPA) comprise only .1% of total brain fatty acid composition (23). DHA is associated with neuronal membrane stability and the functions of serotonin and dopamine transmission, which might connect to the etiology of mood and cognitive manifestations of depression (4,12,24). In contrast, EPA is important in the balance of the immune and neuronal functions by antagonizing membrane arachidonic acid (arachidonic acid [AA], an n-6 PUFA) and reducing prostaglandin E2 (PGE2) synthesis (25). For example, animals fed with a high EPA diet could attenuate the sickness behaviors induced by the high AA diets or the PGE2 treatment (26,27). Interestingly, the sickness behaviors, including anorexia, low activity, and a change in sleep pattern and attention, are similar to somatic symptoms of depression (28).
Consistent with the theoretical relevance and findings from epidemiologic data and clinical trials, the abnormal fatty acid compositions in cell membranes in patients with mood disorders have been reported extensively (29-45), and the findings of the differences in individual PUFAs between patients and control groups are inconsistent. In 1996, Maes and colleagues (30) reported that depression was associated with the significantly higher levels of AA in phospholipids, the AA:EPA ratio in both serum cholesteryl esters and phospholipids, and an n-6:n-3 ratio in cholesteryl esters, as well as lower levels of EPA in both serum cholesteryl esters and phospholipids and the total n-3 PUFAs in cholesteryl esters. However, Peet and colleagues (33) reported that the only abnormal erythrocyte PUFAs level was lower DHA, not EPA or AA. Contrary to Maes's previous report (30), the AA
0006-3223/$36.00 doi:10.10l6/j.biopsych.2010.03.018
BIOL PSYCHIATRY 2010;68:140-147 © 2010 Society of Biological Psychiatry
P.-Y. Lin etal.
BIOL PSYCHIATRY 2010;68:140-147  141
level was lower in the depressed patients in their 1999 report (34). In a study of elderly patients with depressive disorders, subjects had lower DHA and higher AA, n-6:n-3, AA:EPA, and AA:DHA ratios than healthy volunteers (36). In samples of patients with acute coronary syndromes, depressed patients had lower DHA, total DHA, and EPA and higher AA and n-6:n-3, AA:EPA, and AA:DHA ratios than those without depression (38). Interestingly, lower DHA levels before starting interferon (IFN)-a therapy predicted IFN-a-induced depression in patients with chronic hepatitis C viral infection (46). The deficit in PUFA levels and abnormal compositions have also been reported in other mood disorders, including lower DHA and total n-3 PUFAs in postpartum depression (35), lower DHA and EPA in social anxiety disorder (47), and lower DHA and AA in bipolar disorder (37).
To understand this discrepancy, we performed a meta-analysis to examine polyunsaturated fatty acid compositions in patients with depression. We pooled results from all case-control studies to analyze individual n-3 and n-6 PUFA compositions in patients with depression.
Methods and Materials
Literature Search
To identify eligible studies for this meta-analysis, a computerized search was performed for all publications available up to August 2009 through PubMed at the National Library of Medicine using the following key words: (depression OR depressive disorder) AND (omega-3 OR EPA or DHA OR polyunsaturated fatty acid), limited to literature in English and human studies. Reference lists from identified articles and relevant reviews were scrutinized for studies not indexed in the electronic databases.
Inclusion Criteria of Studies in the Meta-Analysis
Studies included in this meta-analysis had to meet the following criteria: 1) measured level of any of EPA, DHA, AA, total n-3, or total n-6; 2) used samples from red blood cell (RBC) membrane, blood phospholipids, or cholesteryl esters; 3) included subjects with depression and control subjects; 4) provided enough data to calculate an effect size; 5) were published in peer-reviewed journals; and 6) separate groups of subjects among studies. Studies that included and reanalyzed the same data set as previously published studies were not regarded as independent, and only the study with the highest number of participants was included. When the articles provided data from different sample tissues from the same subjects, we first used data from RBC membrane or blood phospholipids; if both were not available, we used data from blood cholesteryl esters. See Figure SI in Supplement 1 for the flow chart showing the selection of included studies.
Meta-Analytic Methods
In our analysis, the primary outcomes were comparisons of levels of EPA, DHA, AA, total n-3 PUFA, and total n-6 PUFAs between depressed and control subjects for all included studies. The secondary outcomes were comparisons of the PUFA levels between patients with MDD based on DSM criteria (28) and healthy controls, without major systemic diseases or pregnancy.
For each identified study, the effect size (ES) expressing the difference in each of the PUFA indexes between depressive and control subjects was described as the standardized mean difference (SMD) on the basis of Hedges's adjusted g, in which a value greater than 0 indicated that the index was higher in depressive
subjects. The means and standard deviations of each PUFA index of both depressive and control groups were used to derive the ES from each study. When these data could not be retrieved from the publications, we contacted the authors to acquire the data or derived the ES from other measures of variability. The results of individual studies were synthesized by the random effects model (48), by which ESs were pooled and 95% confidence intervals (CIs) were calculated. The significance of the pooled ES was determined by the z test. Sensitivity analyses were performed to determine whether any individual study was responsible for the significant result. Each study was individually removed, and the significance was retested.
A homogeneity test (Q statistic) was performed to determine whether the group of ESs came from a homogeneous source (48). A rejection of homogeneity suggests that there may have been systemic differences existing among the included studies. In addition, publication bias was assessed by linear regression analysis (49), in which the standard normal deviate of the ES was regressed on the precision of the ES (the inverse of the standard error of the ES). When there was no publication bias, the regression line should traverse the origin, and the expected value of the intercept would be zero. The slope (b) of the regression line indicated the size and direction of the association. An examination of publication bias was a test of the null hypothesis that intercept (a) was equal to zero, as determined by the t test. Meta-analyses were conducted by applying Comprehensive Meta-Analysis software, version 2 (Biostat, Englewood, New Jersey). Two-sided p values < .05 were considered statistically significant.
Results
On the basis of the aforementioned search strategy, our initial search identified 1007 studies, of which only 14 were included in current meta-analysis according to the inclusion criteria (29,30, 32-36,38,40-45) (Figure SI in Supplement 1), which included 3318 subjects (648 depressive and 2670 control subjects). The characteristics of these articles are described in Table 1.
The primary outcomes were comparisons of PUFA levels between depressive (defined arbitrarily in individual studies) and control subjects. The results showed a significant but mild decrease in depressive subjects in the level of EPA (ES = —.18, p = .004; Figure 1) and DHA (ES = -.35, p = .0002; Figure 2) and moderately decreased level of total n-3 PUFA (ES = — .51, p < .0001; Figure 3). The levels of AA (p = .95; Figure 4) and total n-6 PUFAs (p = .94; Figure 5) were not significantly different. Sensitivity analyses showed that the significant changes in the levels of EPA, DHA, and total n-3 were not influenced by any single study. Linear regression analysis showed that there was no publication bias in each group of PUFA measurements.
To examine the abnormalities of PUFA levels between patients with DSM-defined MDD and healthy control subjects, we repeated the analyses in the secondary outcomes by excluding studies not using DSM criteria or using subjects with concomitant medical disorders. In these analyses, 318 subjects (191 depressive subjects, 127 healthy control subjects) from six studies (30,32-34, 43,45) were identified and analyzed for the secondary outcomes. Again, the levels were significantly lower in MDD subjects in EPA (ES = -.42, p = .0008; Figure 1), DHA (ES = -.54, p = .0008; Figure 2), and total n-3 (ES = -.85, p < .0001; Figure 3). The levels of AA (Figure 4) and total n-6 PUFAs (Figure 5) were not significantly higher in MDD patients. As in the primary outcomes, the significant difference in EPA, DHA, and total n-3 was not
www.sobp.org/journal
Table 1. Characteristics of Studies Included in This Meta-Analysis
				nclusion Psychiatric	Use of DSM	Types of Control		Sampling
Studies	Patients, n	Controls, n	Patient Source	Disorders	Criteria	Subjects	Country	Tissue
Fehily (1981) (29)	26	26	Not stated	Bipolar and unipolar endogenous depression	No	Healthy controls	UK	Erythrocytes
Maes(1996) (30)	36	24	Psychiatric inpatients	MDD, mean HDRS = 22.2	Yes	Healthy controls	Belgium	Serum
Edwards (1998) (32)	10	14	Not stated	MDD, mean BDI = 26.9	Yes	Healthy controls, mean BDI = 4.9	UK or Canada	Erythrocytes
Peet(1998)(33)	15	15	Not stated	MDD, evaluated by MADRS	Yes	Healthy controls	UK or Canada	Erythrocytes
Maes(1999) (34)	34	14	Psychiatric inpatients	MDD, evaluated by HDRS	Yes	Healthy controls	Belgium	Plasma
DeVriese (2003) (35)	10	38	Healthy pregnant women	Postpartum women with depression	Yes	Postpartum women without depression	Belgium	Serum
Tiemeier 2003 (36)	106	461	Populated-based community subjects	MDD, dysthymia, and minor depression in elderly people	Yes	Elderly people without depression	The Netherlands	Plasma
Frasure-Smith (2004) (38)	54	54	Hospital-based subjects with high risk of ACS	MDD, in patients with recent ACS	Yes	Patients with recent ACS, without depression	Canada	Plasma
Amin (2008) (40)	118	641	Hospital-based subjects with ACS	ACS patients, PHQ > 10	No	ACS patients, PHQ < 10	US	Erythrocytes
Aupperle (2008) (41)	10	28	Outpatient from MS Clinic	MS patients with depression, by CMDI	No	MS patients without depression	US	Erythrocytes
Feart 2008 (42)	117	1273	Populated-based community subjects	Elderly subjects (>65 y) with depressive symptoms, CES-D >17 in men and >23 in women	No	Other elders	France	Plasma
Dinan (2009) (43)	20	24	Not stated	MDD, HADS > 20	Yes	Healthy controls	Ireland	Plasma
Rees (2009) (44)	16	22	Women in the third trimester of pregnancy	pregnant women with MDD, EDS >13, HADS >14or MADRS >25	Yes	Pregnant women without depression, EDS<8	Australia	Plasma
Riemer (in press) (45)	76	36	Psychiatric inpatients	MDD	Yes	Healthy controls	Germany	Serum
ACS, acute coronary syndrome; BDI, Beck Depression Inventory; CES-D, Center for Epidemiologic Depression Scale; CMDI, Chicago Multiscale Depression Inventory; DSM, Diagnostic and Statistical Manual; EDS, Edinburgh Depression Scale; HADS, Hamilton Depression Rating Scale; MADRS, Montgomery Asberg Depression Rating Scale; MDD, major depressive disorder; Ml, myocardial infarction; MS, multiple sclerosis; PHQ, Patient Health Questionnaire.
P.-Y. Lin etal.
BIOL PSYCHIATRY 2010;68:140-147  143
Group by study type
Study name
Statistics for each study
Hedqes's q and 95% CI
	Hedges's g	Lower limit	Upper fimit	Z-Value	p-Value
Mass 1986 (30)	-0.33	-0.90	0.13	-1.48	0.1397
Edwards 1998 (32)	-1.00	-1.83	-0.17	-2.35	0.0187
Peed 999 (33)	-0.58	-1.29	0 13	-1.59	0.1115
Maes 1933 (34)	-0.67	-1.29	-0.04	-2.08	0 0372
Dinar 2009 (43)	-0.23	-0.75	0 23	-0,88	0.3770
Riemer (in press)(45)	-0.24	-0.63	0.16	-1.17	0.2413
DSM-defined MDD us. healthy controls-pooled	-0.42	-0.67	-0.18	-3.36	0.0008
Fehily1981 (29)	[if, I	0.06	I IF.	2.18	0.0296
De Vriese 2003 (35)	-0.59	-1-28	0 11	■1.65	0 0984
Tismeier 2003 (36)	-0 03	-0.30	0 12	-0.84	0.3998
Frasure-Smith 2004 (38)	-0.11	-0.48	0.27	-0.57	0.5684
Amin 2008 (40)	0.00	-0.20	0 20	0 00	1.0000
Aupperle 2008 (41)	-0.07	-0.78	0.6 A	-0.20	0.3435
Feart 2008 (42)	-0.27	-0.46	■0.09	-2.84	0.0045
Rees 2009 (44)	-0.23	-0.86	0.40	-0.71	0.4787
Studies NOT using DSM criteria and healthy controls-poo led	-0.10	-0.25	0.05	-1.32	0.1877
Overall	-0.18	-D.31	-0.06	-2.85	0.Ü044
-2.0D   -1 00   0.D0    1.00 2.00 Control higher   Depression higher
Figure 1. Forest plot showing effect sizes (Hedges's g) and 95% confidence intervals (CIs) from individual studies and pooled results comparing eicosapen-taenoic acid level between depressed patients and control subjects. MDD, major depressive disorder.
contributed by any single study, nor was there publication bias in any of these measurements of PUFAs.
Discussion
The main finding of this meta-analysis confirms that depression is associated with lower levels of total n-3 PUFAs and both types of n-3 PUFAs, EPA and DHA. Previous observational studies have indicated that clinical depression can be accompanied by low levels of n-3 PUFAs in RBC, plasma and,
as more found recently, brain tissue (50). Our findings extend the results from several studies, but not all, that n-3 PUFAs play an important role in depression. It has been reported that the deficit of n-3 PUFAs in rat brain is associated with impaired serotonergic and dopaminergic neurotransmission and in turn leads to an increase in 5-HT2 receptors and a decrease in D2 receptors in the frontal cortex (51-53). This upregulation of 5-HT2 receptors is thought to play a role in the pathophysiology of depression (54).
Group by study type Study name Statistics for each study Hedqes's a. and S5% CI
Hedges's a		Lower limit	Upper limit	Z-Value	p-Value			
Maes 1996 (30) Edwards 1938 (32)	0.04 -0.96	-0.47 -1.79	0.55 -0.13	0.14 -2.27	0.8897 0 0231			
								
Peet1998(33)	-1.00	-1.74	-0.20	-2.65	0.0080			
Maes 1999 (34)	-0.03	-0.65	0 53	-0.11	0.9122			
Dinan 2009 (43)	-0 65	-1.18	-0 13	■2.42	0.0153			
Riemer (in press)(45)	-0.81	-1.22	-0,40	-3.89	0.0001			
DSM-deflned MDD vs. healthy controls-pooled	-0.54	0.85	-0.22	-3.35	0.0G0B			
Fehily1981 (29)	0.59	0.03	1 12	2.07	0.0383			
De Vriese 20 03 ( 35)	-0,99	-1.71	-0,28	-2.73	0.0063			
Tiemeier 2003 (38)	-0.21	-0.42	-0,00	-1.96	0.0499		-o-	
Frasure-Smith 2004 (38)	-0.49	-0.87	-0.11	-2.51	0.0121		—°—	
Amin 2008 (40)	-0.27	-0.47	-0.08	-2.73	0.0053		-°-	
Aupperle 2008 (41)	-0.01	-0.72	0 70	-0.03	0.9800			
Feart 2008 (42) Rees 2009 (44)	-0.06 -0.88	-0.25 •1.54	0 13 -0 22	-0.65 ■2.62	0.5186 0.0039		-c	
								
Studies NOT using DSM criteria and healthy controls-pooled	0.25	-0.48	-0.02	-2.11	0.0347		o	
Overall	-0.35	0.53	-0.15	3.68	0.0002		♦	
-2.00   -1.00   0.00    1.00 2.00 Control higher Depression higher
Figure 2. Forest plot showing effect sizes (Hedges's g) and 95% confidence intervals (CIs) from individual studies and pooled results comparing docosahexae-noic acid levels between depressed patients and control subjects. MDD, major depressive disorder.
www.sobp.org/journal
144  BIOL PSYCHIATRY 2010;68:140-147
P.-Y. Lin et al.
Group by study type      Study name
Statistics for each study
Hedties's o and 95% CI
Overall
	Hedgess a	Lower limit	Upper limit	Z-Value	p-Value
Maes 1 996 (30)	-0.15	-o.ee	0.36	-0.57	0.571S
Edwards 1998 (32)	-0.98	-1.81	-0.15	-2.31	0.0207
Peel 1998 (33)	■0.93	-1.67	-0.2D	-2.49	0.0127
Maes 1 999 (34)	-0.39	-1.01	0.22	-1.25	0 2117
Riernerfti press)[45)	-1.70	-2.15	-1.25	-7.37	0.0000
thy controls-pooled	-0.85	-1.21	-0.49	-4.62	n.onon
Devriese 2003 (35)	■1.04	-1,76	-0.32	-2.85	0.0044
Tiemeier 2003 (36)	-0.20	-0.41	O 01	-1.38	0.0599
Frasure-Smith 2004 (38)	■0.37	-0.75	0 01	-1.92	0.0551
Arnin 2008 (40)	■ 0.35	■0 55	-0.16	■3 53	0.0004
Aupperle 2008(41)	-0.01	-0.71	0.70	-0.02	09859
Feart 2008 (42)	-0.17	-0 36	0 02	-1 78	0 0734
Rees 2009 (44)	-0.67	-1,32	-0.03	-2,04	0.0418
tteria and healthy controls-pooled	■0.35	-0.60	0.09	-2.69	0.0071
	■O.S1	.0.72	0.30	-4.85	0.0000
-2.00    -1.00     0.00     1.00 2.00 Contra! higher   Depression higher
Figure 3. Forest plot showing effect sizes (Hedges's g) and 95% confidence intervals (CIs) from individual studies and pooled results comparing total omega-3 polyunsaturated fatty acid levels between depressed patients and control subjects. MDD, major depressive disorder.
Lower levels in both DHA and EPA may have different physiologic meanings regarding the biological mechanisms of depression. DHA is a major structural component of phospholipids in neuronal cell membranes, whereas EPA is present in neuronal cell membranes in a very small amount. Therefore, it has been proposed that DHA is more important in brain functioning than EPA (55). However, EPA, rather than DHA, appears to be the effective component when treating clinical depression in published studies (56,57). The contradiction between theoretical mechanisms and clinical studies raises questions about different modes of action of DHA and EPA. For example, EPA,
but not DHA, has other important physiologic functions, including a role as a precursor for eicosanoids and a modulator of cytokines (58). It has been proposed that depression is accompanied by increased secretion of eicosanoids, such as prostaglandins, and by an excessive secretion of proinflammatory cytokines (59). EPA can act as the inhibitor of phospholipase A2 to reduce the secretion of eicosanoids and proinflammatory cytokines (26), which might have been associated with the improvement of somatic symptoms in patients with depression (3). In addition, because the designs of previous clinical trials have focused on the augmentation effects of n-3 PUFAs by enrolling only de-
Group by study type         Study name		Statistics for each stud\			
	Hedges's g	Lower limit	Upper limit	Z-Valüe	p-Value
Maes. 1996 (30)	0.00	-0.51	0.51	0.00	1.0000
Peet1998 (33)	-0.71	-1.43	0.01	-1.93	0.0536
Maes 1099 (34)	-0 46	-1.08	0,16	-1.45	0 1463
Dinan 2009 (43)	0,82	0,28	1,35	2,99	0.0028
Riemer (in press)(45)	0.36	-0.04	0.75	1.76	0.0777
DSM-deflned MDD us. healthy controls-pooled	0.10	-0.19	0.39	0.66	0.5066
FehiJy 1981 (29)	0.00	-0.54	0.54	0,00	1.0000
OeVriese 2003 (35)	-0.14	-0.82	0.55	-0.39	0.6956
Tiemeier 2003 (36)	-0.21	-0.42	-0.00	-1.96	0.0499
Frasure-Smith 2004 (38)	0.29	-0.08	0,67	1,52	0.1283
Amin 2008 (40)	-0.08	-0.28	0.12	-0.80	0.4227
Aupperle 2008 (41)	0,24	-0.47	0,95	0,67	0.5056
Feart20D8 (42)	-0.12	D 31	0.O7	-1.21	0.2271
Rees 2009 (44)	0.07	-0.56	0.70	0.21	0.8300
Studies NOT using DSM criteria and healthy controls-pooled	-0.03	-0.22	0.16	-0.35	0.7254
Overall	0.01	-0.15	0.18	0.07	0.9468
Hedge?':: q and '35% CI
-2.00   -1.00    0.00    1.00 2.00 Control higher Depression higher
Figure 4. Forest plot showing effect sizes (Hedges's g) and 95% confidence intervals (CIs) from individual studies and pooled results comparing arachidonic acid levels between depressed patients and controls. MDD, major depressive disorder.
www.sobp.org/journal
P.-Y. Lin etal.
BIOL PSYCHIATRY 2010;68:140-147  145
Group by study type
Study name
Statistics for each study
Hedges's g and 95% CI
	Hedges's g	Lower limit	Upper limit	Z-Value	p-Value
Maes 1996 (30)	0.16	-0.35	0.67	0.62	0.5368
Peet 1998 (33)	-0.08	-1.62	-0.15	-2.37	0.0177
Maes 1999 (34)	-0.56	-1.18	0.06	-1.76	0.0778
DSM defined MDD vs. healthy controls pooled	0.33	0.74	o.os	1.60	0.1098
DeVriese 2003 (35)	0.44	-0.25	1.13	1 25	0.2125
Tlemeier2003 (3B)	0.14	-0.07	Li 35	1,28	0.1998
Frasure-Smith 20C4 (38)	0.11	-0.26	il 49	0.58	0.5831
Amin 2008 (40)	-0.16	-0.35	0.04	-1.57	0.1167
Aupperle 2008 (41)	0.35	-0.36	1.06	0.96	0.3365
Feart 2008 (42)	■0.08	-0.27	0.11	-0 87	0.3852
Rees 2009(44)	0.64	-0.01	1.29	1.94	0.0518
Studies NOT using DSM criteria and healthy controls-pooled D.03		-D.11	0.28	0.84	0.39B3
Overall	0.01	-0.17	0.18	0.07	0.9427
-2.00    -1.00    0.00     1.00 2,00
Control higher Depression higher
Figure 5. Forest plot showing effect sizes (Hedges's g) and 95% confidence intervals (CIs) from individual studies and pooled results comparing total omega-6 polyunsaturated fatty acid levels between depressed patients and control subjects. MDD, major depressive disorder.
pressed patients receiving antidepressant medications simultaneously (56,57), the antidepressant effect of DHA might be hidden if the effects of DHA and the antidepressant medications overlap. According to the results of lower DHA levels in depression in this meta-analytic review, it is too early to exclude DHA's role in depression and its antidepressant effect.
Notably, compared with the primary outcomes of all the 14 studies, the magnitude of ES of differences in the levels of EPA (Figure 1), DHA (Figure 2), and total n-3 (Figure 3) became larger when the analyses were restricted to studies that used DSM criteria for MDD diagnosis, implying that the n-3 PUFA abnormalities were more significant in the more homogeneous groups of major depression. Interestingly, the significance of lower EPA levels in depressed subjects disappeared when we looked only at the studies with depressed subjects other than DSM-defined MDD (ES = -.10, 95% CI = -.25-.05, p = .19; Figure 1), implying that the lower EPA levels were presented only in patients with clinical major depression, but not in subjects with less strictly defined depressive symptoms. This negative finding in these subjects might be an explanation to the previous studies revealing negative findings of n-3 PUFAs' mood-improving effects in nonclinical symptomatic subjects (60,61). Specifically, two meta-analyses (from different groups) using only trials that enrolled patients with diagnosed MDD showed the beneficial effects of n-3 PUFAs, especially EPA, supplementation on depressed mood (56,57). However, the other meta-analyses (from the same group) involving individuals without diagnosis of major depression found no evidence of a beneficial effect of n-3 PUFAs on depressed mood (60,61). Further studies are warranted to investigate the relationship between pretreatment deficits and posttreatment effects of n-3 PUFAs.
The major limitation of this study is that PUFA levels from the clinical studies we analyzed were not directly from the tissues of the central nervous system; hence, the results cannot be applied to brain PUFA levels. Nonetheless, previous studies have revealed that the n-3 PUFA levels from peripheral blood tissues of RBC and plasma might reflect brain levels of n-3 PUFAs in mammals (62-65). For example, in a study measuring PUFAs
from tissues of blood plasma, erythrocytes, liver, muscle, adipose tissue, retina, and brain samples in piglets consuming assigned diets, the levels of EPA and DHA in both plasma and erythrocytes highly correlated to those levels in brain tissue (62). Specifically, the coefficients (ts) of EPA and DHA in plasma in correlation with brain tissue were .78 (p < .001) and .80 (p < .001), respectively. Meanwhile, the rs of EPA and DHA in erythrocyte in correlation with brain were .78 (p < .001) and .80 (p < .001), respectively. Although data from human subjects are not yet available, similar findings of this high brain plasma-erythrocyte correlation of EPA and DHA have been reported in rhesus monkeys (64) and in rats (65). Unlike the consistency among erythrocytes, plasma, and brain tissue in EPA and DHA levels, AA levels in the plasma or erythrocyte might not be as apparent in the brain (62,64). Therefore, caution is warranted in interpreting the finding of nonsignificant differences of AA levels between patients and controls in our study.
There are other methodologic limitations. Although the levels of n-3 and n-6 PUFAs are highly consistent in peripheral blood plasma and erythrocytes (62,65-68), the reliability of combining PUFA levels from plasma and erythrocytes for data analyses is uncertain. In this study, we expect the use of the "percentage" (individual PUFA from the obtained tissues) as the unit of measure might have offered better reliability in combining PUFAs levels for data analyses. Second, although gas chromatography has been used universally as the standard measurement of PUFA levels, the bias from the methodology across laboratories could not be eliminated. Again, we believe the use of percentage as the unit of measure would have better reliability than the use of the "absolute value" in this case. Finally, in our analyses, we excluded Mamalakis's studies (69-73), which measured PUFA levels from human adipose tissues. Although in two of the five studies, the authors also measured PUFA levels from serum phospholipids (71,73), they did not group patients according to clinical diagnosis or rating scales of depression to perform the difference analyses. Therefore, we could not use the data for meta-analyses.
In conclusion, the levels of total n-3 PUFAs, DHA, and EPA
www.sobp.org/journal
146  BIOL PSYCHIATRY 2010;68:140-147
P.-Y. Lin et al.
are lower in patients with depression, implying that n-3 PUFAs play a role in the etiology of depression. This provides further support of phospholipid hypothesis of depression and a rationale for using n-3 PUFAs as an alternative treatment for patients with depression. Because individual PUFAs may have different biological functions, this study might provide essential background to examine specific roles of DHA and EPA in the classification of depressive disorders in future studies.
The work was supported by Grant Nos. NSC 98-262 7-B-039-003, 98-2628-B-039-020-MY3, 98-2320-B-038-018-MY3, and 96-2320-B-038-035-MY2 from the National Science Council in Taiwan; Grant No. 97(2)-TRA-001 from the National Science and Technology Program for Biotechnology and Pharmaceuticals TranslationalMedicine Project in Taiwan; CMU97-336from the China Medical University in Taiwan; and a National Alliance for Research on Schizophrenia and Depression Young Investigator Award.
The authors report no biomedical financial interests or potential conflicts of interest.
Supplementary material cited in this article is available online.
1. Lhe World Health Report (2001). Mental Health: News Understanding News Hope. Geneva: World Health Organization.
2. Murray CJ, Lopez AD (1997): Alternative projections of mortality and disability by cause 1990-2020: Global Burden of Disease Study. Lancet 349:1498-1504.
3. Su KP (2009): Biological mechanism of antidepressant effect of Omega-3 fatty acids: How does fish oil act as a "mind-body interface?" Neuro Signals 17:144-152.
4. Horrobin DF, Bennett CN (1999): Depression and bipolar disorder: Relationships to impaired fatty acid and phospholipid metabolism and to diabetes, cardiovascular disease, immunological abnormalities, cancer, ageing and osteoporosis. Possible candidate genes. Prostaglandins Leu-kotEssent Fatty Acids 60:217-234.
5. Horrobin DF(2001): Phospholipid metabolism and depression:Lhe possible roles of phospholipase A2 and coenzyme A-independent transacy-lase. Hum Psychopharmacol 16:45-52.
6. Su KP, Shen WW, Huang SY (2000): Effects of polyunsaturated fatty acids on psychiatric disorders./Am J Clin Nutr 72:1241.
7. Lanskanen A, Hibbeln JR.Luomilehto J, Uutela A, Haukkala A, Viinamaki H, et al. (2001): Fish consumption and depressive symptoms in the general population in Finland. Psychiatr Serv52:529-531.
8. Lanskanen A, Hibbeln JR, Hintikka J, Haatainen K, Honkalampi K, Viinamaki H (2001): Fish consumption, depression, and suicidality in a general population. Arch Gen Psychiatry 58:512-513.
9. Hibbeln JR (1998): Fish consumption and major depression. Lancet 351: 1213.
10. Marangell LB, Martinez JM, Zboyan HA, Kertz B, Kim HF, Puryear LJ (2003): A double-blind, placebo-controlled study of the omega-3 fatty acid docosahexaenoic acid in the treatment of major depression. Am J Psychiatry 160:996-998.
11. Silvers KM, Woolley CC, Hamilton FC, Watts PM, Watson RA (2005): Randomised double-blind placebo-controlled trial of fish oil in the treatment of depression. Prostaglandins Leukot Essent Fatty Acids 72: 211-218.
12. Su KP, Huang SY.Chiu CC, Shen WW (2003): Omega-3 fatty acids in major depressive disorder. A preliminary double-blind, placebo-controlled trial. Eur Neuropsychopharmacol 13:267-271.
13. Nemets B, Stahl Z, Belmaker RH (2002): Addition of omega-3 fatty acid to maintenance medication treatment for recurrent unipolar depressive disorder./\m J Psychiatry 159:477-479.
14. Peet M, Horrobin DF (2002): A dose-ranging study of the effects of ethyl-eicosapentaenoate in patients with ongoing depression despite apparently adequate treatment with standard drugs. Arch Gen Psychiatry 59:913-919.
15. Frangou S, Lewis M, McCrone P (2006): Efficacy of ethyl-eicosapentae-noic acid in bipolar depression: Randomised double-blind placebo-controlled study. Br J Psychiatry 188:46-50.
16. Nemets H, Nemets B, Apter A, Bracha Z, Belmaker RH (2006): Omega-3 treatment of childhood depression: A controlled, double-blind pilot study. Am J Psychiatry 163:1098-1100.
17. Jazayeri S, Lehrani-Doost M, Keshavarz SA, Hosseini M, Djazayery A, Amini H, et al. (2008): Comparison of therapeutic effects of omega-3 fatty acid eicosapentaenoic acid and fluoxetine, separately and in combination, in major depressive disorder. AustNZ J Psychiatry 42:192-198.
18. Chiu CC, Huang SY, Shen WW, Su KP (2003): Omega-3 fatty acids for depression in pregnancy. Am J Psychiatry 160:385.
19. Su KP, Huang SY, Chiu LH, Huang KC, Huang CL, Chang HC, Pariante CM (2008): Omega-3 fatty acids for major depressive disorder during pregnancy: Results from a randomized, double-blind, placebo-controlled trial. J Clin Psychiatry 69:644-651.
20. da Silva LM, Munhoz RP, Alvarez C, Naliwaiko K, Kiss A, Andreatini R, Ferraz AC (2008): Depression in Parkinson's disease: A double-blind, randomized, placebo-controlled pilot study of omega-3 fatty-acid supplementation. J Affect Disord 111:351-9.
21. Hibbeln JR, Salem N Jr (1995): Dietary polyunsaturated fatty acids and depression: When cholesterol does not satisfy./Am J Clin Wufr62:1-9.
22. Lands WE (1992): Biochemistry and physiology of n-3 fatty acids. FASEB J 6:2530-2536.
23. McNamara RK, Carlson SE (2006): Role of omega-3 fatty acids in brain development and function: Potential implications for the pathogenesis and prevention of psychopathology. Prostaglandins Leukot Essent Fatty Acids 75:329-349.
24. Chalon S (2006): Omega-3 fatty acids and monoamine neurotransmission. Prostaglandins Leukot Essent Fatty Acids 75:259-269.
25. Farooqui AA, Ong WY, Horrocks LA (2006): Inhibitors of brain phospholipase A2 activity: Lheir neuropharmacological effects and therapeutic importance for the treatment of neurologic disorders. Pharmacol Rev 58:591-620.
26. Song C, Leonard BE, Horrobin DF (2004): Dietary ethyl-eicosapentaenoic acid but not soybean oil reverses central interleukin-1-induced changes in behavior, corticosterone and immune response in rats.Stress 7:43-54.
27. Song C, Phillips AG, Leonard BE, Horrobin DF (2004): Ethyl-eicosapentaenoic acid ingestion prevents corticosterone-mediated memory impairment induced by central administration of interleukin-1 beta in rats. Mol Psychiatry 9:630 - 638.
28. American Psychiatric Association (1994): Diagnostic and Statistical Manual of Mental Disorders, 4th ed. Washington, DC: American Psychiatric Association.
29. Fehily AM, Bowey OAM, Ellis FR, Meade BW (1981): Plasma and erythrocyte membrane long chain polyunsaturated fatty acids in endogenous depression. Neurochem Int 3:37-42.
30. Maes M, Smith R, Christophe A, Cosyns P, Desnyder R, Meltzer H (1996): Fatty acid composition in major depression: Decreased omega 3 fractions in cholesteryl esters and increased C20:4 Omega 6/C20:5 omega 3 ratio in cholesteryl esters and phospholipids. J Affect Disord 38:35-46.
31. Adams PB, Lawson S, Sanigorski A, Sinclair AJ (1996): Arachidonic acid to eicosapentaenoic acid ratio in blood correlates positively with clinical symptoms of depression. Lipids 31 (suppl):S157-S161.
32. Edwards R, Peet M, Shay J, Horrobin D (1998):Omega-3 polyunsaturated fatty acid levels in the diet and in red blood cell membranes of depressed patients. J Affect Disord 48:149 -155.
33. Peet M, Murphy B, Shay J, Horrobin D (1998): Depletion of omega-3 fatty acid levels in red blood cell membranes of depressive patients. Biol Psychiatry 43:315-319.
34. Maes M, Christophe A, Delanghe J, Altamura C, Neels H, Meltzer HY (1999): Lowered omega3 polyunsaturated fatty acids in serum phospholipids and cholesteryl esters of depressed patients. Psychiatry Res 85:275-291.
35. De Vriese SR, Christophe AB, Maes M (2003): Lowered serum n-3 polyunsaturated fatty acid (PUFA) levels predict the occurrence of postpartum depression: Further evidence that lowered n-PUFAs are related to major depression. LifeSci 73:3181-3187.
36. Liemeier H, van Luijl HR, Hofman A, Kiliaan AJ, Breteler MM (2003): Plasma fatty acid composition and depression are associated in the elderly:Lhe Rotterdam Study.Am J Clin Nutr 78:40-46.
www.sobp.org/journal
P.-Y. Lin etal.
BIOL PSYCHIATRY 2010;68:140-147  147
37. Chiu CC, Huang SY, Su KP, Lu ML, Huang MC, Chen CC, Shen WW (2003): Polyunsaturated fatty acid deficit in patients with bipolar mania. Eur Neuropsychopharmacol 13:99-103.
38. Frasure-Smith N, Lesperance F, Julien P (2004): Major depression is associated with lower omega-3 fatty acid levels in patients with recent acute coronary syndromes. Biol Psychiatry 55:891-896.
39. Schins A, Crijns HJ, Brummer RJ, Wichers M, Lousberg R, Celis S, Honig A (2007): Altered omega-3 polyunsaturated fatty acid status in depressed post-myocardial infarction patients. Acta Psychiatr Scand 115:35-40.
40. Amin AA, Menon RA, Reid KJ, Harris WS, Spertus JA (2008): Acute coronary syndrome patients with depression have low blood cell membrane omega-3 fatty acid levels. Psychosom Med 70:856-862.
41. Aupperle RL, Denney DR, Lynch SG, Carlson SE, Sullivan DK (2008): Omega-3 fatty acids and multiple sclerosis: Relationship to depression. Jße/iai//Wed31:127-135.
42. Feart C, Peuchant E, Letenneur L, Samieri C, Montagnier D, Fourrier-Reglat A, Barberger-Gateau P (2008): Plasma eicosapentaenoic acid is inversely associated with severity of depressive symptomatology in the elderly: Data from the Bordeaux sample of the Three-City Study. Am J Clin Nutr &7-A 156-1162.
43. Dinan T, Siggins L, Scully P, O'Brien S, Ross P, Stanton C (2009): Investigating the inflammatory phenotype of major depression: Focus on cytokines and polyunsaturated fatty acids. J Psychiatr Res 43:471-476.
44. Rees AM, Austin MP, Owen C, Parker G (2009): Omega-3 deficiency associated with perinatal depression: Case control study. Psychiatry Res 166:254-259.
45. Riemer S, Maes M, Christophe A, Rief W (2009): Lowered omega-3 PUFAs are related to major depression, but not to somatization syndrome. J Affect Disord.
46. Su KP, Huang SY, Peng CY, Lai HC, Huang CL, Chen YC, et al. (2010): Phospholipase A2 and cyclooxygenase 2 genes influence the risk of interferon-alpha-induced depression by regulating polyunsaturated fatty acids levels. Biol Psychiatry 67:550-557.
47. Green P, Hermesh H, Monselise A, Marom S, Presburger G, Weizman A (2006): Red cell membrane omega-3 fatty acids are decreased in non-depressed patients with social anxiety disorder. Eur Neuropsychophar-macol 16:107-113.
48. Shadish WR, Haddock CK (1994): Combining estimates of effect size. In: Cooper H, Hedges LV, editors. The Handbook of Research Synthesis. New York: Russell Sage Foundation, 261-281.
49. Egger M, Davey SG, Schneider M, Minder C (1997): Bias in meta-analysis detected by a simple, graphical test. BMJ 315:629-634.
50. McNamara RK, Hahn CG, Jandacek R, Rider T, Tso P, Stanford KE, Rich-tand NM (2007): Selective deficits in the omega-3 fatty acid docosa-hexaenoic acid in the postmortem orbitofrontal cortex of patients with major depressive disorder. Biol Psychiatry 62:17-24.
51. Delion S,Chalon S,Guilloteau D, Lejeune B, Besnard JC, Durand G (1997): Age-related changes in phospholipid fatty acid composition and mono-aminergic neurotransmission in the hippocampus of rats fed a balanced or an n-3 polyunsaturated fatty acid-deficient diet. J Lipid Res 38:680-689.
52. Delion S, Chalon S, Guilloteau D, Besnard JC, Durand G (1996): Alpha-linolenic acid dietary deficiency alters age-related changes of dopaminergic and serotoninergic neurotransmission in the rat frontal cortex. J Neurochem 66:1582-1591.
53. Delion S, Chalon S, Herault J, Guilloteau D, Besnard JC, Durand G (1994): Chronic dietary alpha-linolenic acid deficiency alters dopaminergic and serotonergic neurotransmission in rats.JWufr 124:2466-2475.
54. Maes M, Meitzer HYM (1995): The serotonin hypothesis of major depression. In: Bloom FE, Kupfer DJ, editors. Psychopharmacology, the Fourth Generation of Progress. New York: Raven, 933-941.
55. Peet M, Stokes C (2005): Omega-3 fatty acids in the treatment of psychiatric disorders. Drugs 65:1051 -1059.
56. Freeman MP, Hibbeln JR, Wisner KL, Davis JM, Mischoulon D, Peet M, et al. (2006): Omega-3 fatty acids: Evidence basis for treatment and future research in psychiatry. J Clin Psychiatry 67:1954-1967.
57. Lin PY, Su KP (2007): A meta-analytic review of double-blind, placebo-controlled trials of antidepressant efficacy of omega-3 fatty acids. J Clin Psychiatry 68:1056 -1061.
58. Fenton WS, Hibbeln J, Knable M (2000): Essential fatty acids, lipid membrane abnormalities, and the diagnosis and treatment of schizophrenia. Biol Psychiatry 47:8 -21.
59. Maes M, Smith RS (1998): Fatty acids, cytokines, and major depression. Biol Psychiatry 43:313-314.
60. Appleton KM, Hayward RC, Gunnell D, Peters TJ, Rogers PJ, Kessler D, Ness AR (2006): Effects of n-3 long-chain polyunsaturated fatty acids on depressed mood: Systematic review of published trials. Am J Clin Nutr 84:1308-1316.
61. Rogers PJ, Appleton KM, Kessler D, Peters TJ, Gunnell D, Hayward RC, etal. (2008): No effect of n-3 long-chain polyunsaturated fatty acid (EPA and DHA) supplementation on depressed mood and cognitive function: A randomised controlled trial. Br J Nutr 99:421-431.
62. Lapillonne A, Demar JC, Nannegari V, Heird WC (2002): The fatty acid profile of buccal cheek cell phospholipids is a noninvasive marker of long-chain polyunsaturated fatty acid status in piglets. J Nutr 132:2319-2323.
63. Makrides M, Neumann MA, Byard RW, Simmer K,Gibson RA (1994): Fatty acid composition of brain, retina, and erythrocytes in breast- and formula-fed infants. Am J Clin Nutr 60:189-194.
64. Connor WE, Neuringer M, Lin DS (1990): Dietary effects on brain fatty acid composition: The reversibility of n-3 fatty acid deficiency and turnover of docosahexaenoic acid in the brain, erythrocytes, and plasma of rhesus monkeys. J Lipid Res 31:237-247.
65. Stark KD (2008): The percentage of n-3 highly unsaturated fatty acids in total HUFAasa biomarker for omega-3 fatty acid status in tissues. Lipids 43:45-53.
66. Arterburn LM, Hall EB,Oken H (2006): Distribution, interconversion,and dose response of n-3 fatty acids in humans. Am J Clin Nutr 83:1467S-1476S.
67. Cao J, Schwichtenberg KA, Hanson NQ, Tsai MY (2006): Incorporation and clearance of omega-3 fatty acids in erythrocyte membranes and plasma phospholipids. Clin Chem 52:2265-2272.
68. Harris WS, Pottala JV, Sands SA, Jones PG (2007): Comparison of the effects of fish and fish-oil capsules on the n 3 fatty acid content of blood cells and plasma phospholipids. Am J Clin Nutr 86:1621-1625.
69. Mamalakis G, Tornaritis M, Kafatos A (2002): Depression and adipose essential polyunsaturated fatty acids. Prostaglandins Leukot Essent Fatty Acids 67:311 -318.
70. Mamalakis G, Kiriakakis M, Tsibinos G, Kafatos A (2004): Depression and adipose polyunsaturated fatty acids in an adolescent group. Prostaglandins Leukot Essent Fatty Acids 71:289-294.
71. Mamalakis G, Kalogeropoulos N, Andrikopoulos N, Hatzis C, Kromhout D, Moschandreas J, Kafatos A (2006): Depression and long chain n-3 fatty acids in adipose tissue in adults from Crete. Eur J Clin Nutr 60:882-888.
72. Mamalakis G, Kiriakakis M, Tsibinos G, Hatzis C, Flouri S, Mantzoros C, Kafatos A (2006): Depression and serum adiponectin and adipose omega-3 and omega-6 fatty acids in adolescents. Pharmacol Biochem Behav 85:474-479.
73. MamalakisG, Kiriakakis M,TsibinosG,Jansen E,Cremers H, Strien C,etal. (2008): Lack of an association of depression with n-3 polyunsaturated fatty acids in adipose tissue and serum phospholipids in healthy adults. Pharmacol Biochem Behav 89:6 -10.
www.sobp.org/journal