Introduction

Treating children with psychotropics requires more care than treating adults in part because children’s brains are developing more rapidly than are those of adults.  The younger the child, the larger and more unknown the potential long term behavioral effects from altered synaptic or neural function.  Children also may be more susceptible to the adverse reactions of some psychotropics than are adults.

Making diagnoses in children may also be more difficult than in adults.  ADHD, bipolar disorder, and depression may be much more difficult to reliably distinguish in children than in adults.  Therefore, treatment with agents that normalize neural and psychiatric functions, are normal constituents of the body, and address several behavioral manifestations of psychiatric disorders would be desirable.  As we’ll see, long chain Ω-3 fatty acids (FA’s) may offer these attributes and also may enhance normal development at least in comparison to development engendered by diets currently extant in much of the developed earth.  If a link between Ω-3 FA consumption and some psychiatric disorders were demonstrated, it would have enormous consequences ranging from the potential for dietary interventions to both reduce the incidence of some disorders and treat some as well to a justification for legislation establishing social structures to facilitate extended breast feeding.

To elucidate the potential role of Ω-3 FA’s in treating psychiatric disorders in children, this paper will review:

1. the first detailed hypothesis linking depression with a reduction in dietary Ω-3 FA’s and increased dietary arachidonic acid (AA) content,
2. a comparison of neolithic and “modern” diets,
3. epidemiological evidence for a linkage between Ω-3 FA consumption and depression as well as linkages between both depression and Ω-3 FA consumption and cardiovascular disease,
4. biochemical linkages between the both Ω-3 FA’s and the Ω-3/Ω-6 FA ratio and depression, effects of Ω-3 FA’s in adult psychiatric disorders,
5. the metabolic pathways involved, and
6. the mechanisms by which Ω-3 FA’s and the ratio of Ω-3/Ω-6 FA’s may effect psychiatric function.

Finally, this paper will review some of the potential applications of Ω-3 FA’s in psychiatry as well as their potentially synergistic interactions with other regulatory compounds such as insulin and thyroid hormones.

Initial Hypothesis

In 1991, Smith*1 hypothesized that the drastically increasing rate of depression documented by Klerman and Weissman*2 resulted from an increase in the ratio of Ω−6/ Ω−3 FA consumption.  (Actually, Rudin*55,*55a anticipated a link between Ω-3 FA consumption and depression by a decade, Burton functionally anticipated it by 340 years, and the Egyptians by over 1000 years.*111pgs 454-455)   The extent of the increase in the lifetime prevalence of major depression is enormous.  By age 44, the 1935-1944 cohort experienced roughly 27 times the rate of major depression as did the 1905-1914 cohort*2.  Every cohort from 1905 to 1955-1964 demonstrates an increase in major depression and the onset of major depression begins earlier in each cohort.

These findings are most likely valid because:

1. The authors carefully exclude the possibility of artifacts ( e.g. Changing diagnostic criteria, selective migration, differential mortality, changing attitudes of mental health professionals, etc.) generating such consistent changes between and among cohorts
2. Similar observations were made in other countries (Sweden, Germany, New Zealand, and Canada) *2
3. In other cultures (Puerto Ricans living in Puerto Rico, Mexican-Americans living in Los Angeles, CA., and Koreans living both in Seoul and in rural areas) no such increase occurred.

This difference suggests that the causal agents are a product of our society and possibly amenable to change.  I acknowledge that the lack of change in cohorts of Mexican Americans living in Los Angeles could pose some difficulties but it also potentially offers an opportunity to test various hypotheses.  A more recent epidemiological study on cohorts since 1952 does not demonstrate a continuation in

Smith bases his argument that the increasing rate of depression derives significantly from the increase in vegetable oil and seed consumption (Ω−6 fatty acids) relative to free ranging game consumption ( Ω−3 fatty acids) on 7 observations: 1) Volunteers given monokines develop all the symptoms necessary for the diagnosis of major depression.  2) IL-1 can account for the hormone abnormalities in depression.  3) Diseases and cohorts where macrophage activations is known to occur have high rates of depression.  4) Microglia, which are macrophages residing in the brain, can secrete monokines.  5) Estrogen increases IL-1 secretion by macrophages which is consistent with the higher rates of depression in females cross culturally.*17 (He did note that the low incidence of atherosclerosis in premenopausal women appears inconsistent with macrophage activation causing depression but reasoned that it was their iron status*20 that protects them.  6) Eicosapentaenoic acid (EPA), an Ω−3 fatty acid, suppresses macrophages while linoleic acid, an Ω−6 fatty acid, activates them.  7) Epidemiology is consistent with this hypothesis.

Smith also hypothesized that since the epidemiology of ischemic heart disease (ISD) paralleled that of MDD, a similar mechanism might underlie both.  Since Smith’s publication substantial additional data has been published supporting his hypotheses.  I will review these and additional data as well as human studies supporting the efficacy of Ω−3 fatty acids in modifying human behavior.

Evolutionary Perspective – Neolithic vs “Modern” Diets

The formidable forces of evolution efficiently adapt any species to its environment.  Substantial changes from that environment over short periods of time, inclusive of dietary changes, most likely are detrimental until the species adapts.  For humans, this assertion implies that altering the Paleolithic diet via agriculture has adversely affected our health, inclusive of our mental health.  Arguing that deviations from the Paleolithic diet can be detrimental to the health of contemporary humans requires minimal genetic change over the last 10,000 to 50,000 years.  Eaton et.al.*3 support this assertion:

…. Biomolecular evidence indicating that humans and chimpanzees now differ genetically by just 1.6% even though the hominid-pongid divergence occurred seven million years ago *4, and by dentochronogic data showing that current Europeans are genetically more like their Cro-Magnon ancestors than they are like 20^th century Africans or Asians *5.  Accordingly, it appears that the gene pool has changed little since anatomically modern humans, Homo sapiens sapiens, became widespread about 35,000 years ago and that, from a genetic standpoint, current humans are still late Paleolithic, preagricultural hunter-gatherers.

Given that genetically humans are preagricultural and remain nearly identical to our Paleolithic ancestors*10, the diet that best fits our nutritional needs would have the composition of a hunter gather’s.  Estimates of the composition of their diet vary somewhat but are reasonably consistent*3,*6,*8,*9.  The most recent estimate using the largest data base*6 indicates that our ancestors consumed 37% protein, 41% carbohydrate, 22% fat with a ratio of polyunsaturates to saturates (P:S) of 1.4, 480mg. of cholesterol, and 3500 mg. of potassium vs only 256 mg. of sodium.  Changes from this diet can have profound effects.  For example, because of improved protein availability members of affluent societies are now almost as tall as our Paleolithic predecessors*11,*9 having regained roughly 6 inches of stature lost in the transition to an agricultural society.  Although this diet varies considerably from current diets in industrialized societies (protein 251 gm. vs 90 gm., fat 71 gm. vs 142 gm., P:S 1.4 vs .44)*8 and these differences may be responsible for numerous diseases*3,*6,*13,*14,*63,*64,*65,*66, here I will focus on the probable consequences of the differences in dietary fat composition.

The differences in fat consumption in the analysis above understate the effect of the transition to an industrialized society on the nutritional effects of fat.  Agricultural practices and technology related to agricultural products have deleteriously altered the compositional shifts in fatty acid consumption.  Although neolithic humans consumed much game than do the members of industrialized societies, the composition of that meat differed markedly from the meat we consume today.  The fat of domestic animals reflects increased storage fats which are saturated in contrast to structural fats which are polyunsaturated.  Not only does wild game contain less fat than domesticated animals (3.9% vs >30%)*15, but the fat of wild game contains about three times the proportion of polyunsaturates and game animals have more Ω−3 fatty acid (3% vs 0.4% of total fat) than does beef*12,*9,*13,*13b.  Eggs from free-ranging chickens have an Ω−6/Ω−3 of 1.3 while eggs from farmed chickens have a ratio of 19.4 *13,13a.  The invention of the continuous screw press and steam-vacuum deodorization allowed the industrial production of vegetable oils.  Hydrogenation reduces the linolenic acid (LNA -C18:3 Ω−3) content of soybean oil from 8.5% to 3% *13c pg.33 increasing the ratio of linoleic acid (LA – Ω−6) to LNA (13) and introducing trans fatty acids previously not experienced by humans. During the last 150 years these changes have altered the ratio of Ω−3 to Ω−6 fatty acids in the human diet from about 1:1 to between10 and 25:1.*9,*13; note that Smith*1 based his argument on an Ω−3/Ω−6 of only 5:1.

Fully assessing the effects of changes in our diet on fatty acid consumption requires a detailed analysis.  While LA has long been considered an essential fatty acid (EFA)*14 pg.1, more recently, LNA has also.  Even though LNA can be converted to EPA and DHA (docosahexaenoic acid)in humans, the conversion is inefficient and not clearly sufficient for optimal function suggesting that EPA & DHA functionally may be essential FA’s.  The longer Ω−3 fatty acids, EPA & DHA, have more effect on prostaglandin and leukotriene synthesis, are more easily encorporated into membranes, and may be required in some tissues which may not convert one Ω−3 fatty acid to another.*14 pgs 28-39.  The efficiency of the conversion of LNA to longer chain Ω−3 fatty acids may decline with age*13 which may imply a greater requirement for EPA and DHA in older adults.  A complete analysis would include comparisons of the longer chain Ω−3 fatty acid composition of Paleolithic and Westernized diets but such an analysis is not available.  Humans did not evolve to ingest the proportions of Ω−3 and Ω−6 fatty acids extant in the American diet today.  The evidence that follows indicates that this change has had adverse psychological consequences.

Epidemiology

Since Smith generated his hypotheses, additional epidemiological data has become available which supports his assertion.  Hibbeln correlated the prevalence of major depression*17 with annual fish consumption finding a remarkably linear inverse correlation (r=-0.84, p<0.005, see below) over wide ranges of both depression (0.12% to 5.8%) and fish consumption (~25 lbs to ~147 lbs./person/year).  Fish consumption and serum phospholipid (PL) composition show some correlation since the Japanese have respectively 30 and 80 -100 times more EPA and DHA in their serum PL’s than do Americans.*131 

In addition to the arguments that Hibbeln made supporting the validity of the differences in the prevalence of major depression among countries, the same methodology*17 showed a much smaller difference in the prevalence of bipolar disorder suggesting that various possible biases did not artifactually create the differences in the prevalence of depression.  This reasoning also suggests that the recent finding that in Stirling County the rate of depression has not increased since the 1952 cohort may not extrapolate more widely as also suggested by Blazer and Kaplan.*184  Furthermore, the location of this population, “in Canada, on the eastern side of the Gulf of Maine,” creates the possibility that this population consumes more fish than do other populations which may limit the decrease in Ω−3/Ω−6 FA consumption noted above.  Since the Stirling County Study examined later cohorts than did the Kleman and Weissman study,*2 the increase in depression by cohort may have ceased by the time of the later study.

Since Smith published his paper, substantial evidence has accumulated implicating inflammation*196 and Ω3 FA’s*13,*13c,*14 pgs. 77- 248,*131 in atherosclerosis.  In addition the correlation between depression and cardiovascular disease has become unavoidable.  A meta analysis of 83 studies over 30 years showed that greater severity of depression predicts greater risk of coronary artery disease (r=.25, p<.0000001) and myocardial infarction (r=.25, p<.0000001).*135  Although more general, other supportive data show that men identified as depressed and not suffering from alcoholism suffer twice the expected death rate*187,*186,*185 and at the end of 16 years of study 83% were dead or had been persistently impaired *188,*186,*185.  Since cardiovascular disease causes more deaths than any other disease, it is likely that a significant proportion of these excess deaths are attributable to an atherosclerosic process.  Prospective data from the Balitmore ECA study*189 support this inference.  Interviews conducted 12 years after the original 1981 ECA study revealed that after allowing for coronary risk factors (age, sex, cigarette smoking, history of diabetes, history of hypertension, socioeconomics, and marital status), simply a history of dysphoria independently increased the odds ratio for an MI by a factor of 2 and a history of major depression increased the odds ratio for an MI by a factor of 4.5 independent of tricyclic use.  These increases may understate the role of depression in MI since the investigators could not look at prior mortality from MI*189, since current data suggests that depression is related to a worse prognosis after an MI*193,*194 (i.e. Those who had died from an MI prior to the second interview would have more likely been depressed.), and since anxiety and depression elevate by a factor of about 2 the risk of developing hypertension, itself an independent risk factor for MI.*189  Supporting these findings are data demonstrating a two fold increase in hospitalization for ischemic heart disease among the more severely depressed.*191

These data suggest that depression may precede the development of cardiovascular disease or that both depression and some forms of cardiovascular disease may be different manifestations of the same disease process.  The finding of intimal lesions in all the aortas and more than half of the right coronary arteries of15-19 year olds*195 suggests the latter.

An inverse correlation between fish consumption and depression of course does not prove causality even when that correlation in space at a given time complements a correlation between Ω-3 FA consumption and depression over time.  Demonstrating causality requires showing that adding Ω−3 fatty acids to the population’s diet reduces the rate of depression.  Intervention on this scale would be expensive and not warranted without further supporting data.

Supporting data could include:

1) Demonstrating effects of Ω−3 fatty acids on:
a) Depression & other related psychiatric conditions
b) Human behavior

2) Establishing:
a) Putative mechanisms of action
b) Ω−3 fatty acid dependent changes in electrophysiological  parameters
c) Safety or improvement in other parameters of health

All of this evidence is not available.  At least one trial of Ω−3 fatty acids in depression is underway but the results are not yet available.  Much of the remainder of the above supporting classes of data have become available since Smith’s publication*1.

Effects on Depression

The earliest formal report of the use of Ω-3 FA in depression appeared in a treatise by Robert Burton on melancholy.  He treated melancholy with a diet producing a high ratio of fish oil to other fats.  For severe depression he recommended both a two-week diet of cow brains, of course, an excellent source of DHA and EPA and abstinence from alcohol.  He acknowledges learning of the efficacy of cow brains from early Arabic medical texts.*111 pgs.454-455  More recently Rudin*55,*55a hypothesized that a deficiency in Ω−3 FA’s might generate depression and treated 44 patients with two to six tablespoons of flaxseed oil, containing a high concentration of LNA, linolenic acid, daily.  Some patients responded within a few days, some within hours, and some became manic but were managed by a reduction in the dose of flaxseed oil.  As we know, the induction of mania is a common feature of antidepressant and at least one mood stabilizer, lamotrigine.  Switching to supplemental  Ω−6 FA’s, reactivated the patients’ symptoms.  As we’ll see, despite not being blinded, this study may have produced results replicable in more controlled studies.

The data implicating fatty acids in depression contain some inconsistencies.  Two early studies*26,*27 reported increased levels of DHA in plasma phosphotidylcholine in depressed patients but those studies had substantial limitations inclusive of:  not using research diagnostic criteria, including bipolar as well as endogenous and reactive depressives, not assessing diet, and incorporating as one-third the subjects, patients taking medications which could cause phosholipidosis.*31  Since phospholipids may contain  Ω-3 & Ω-6 FA’s, increasing the content of phospholipids could create a measurement artifact.

In addition to the inverse correlations in time and space between Ω-3 FA consumption and depression noted above, data more recent than the those  suggesting a positive correlation, have supported the inverse correlation between EPA & DHA and depression.  Studies in heart disease and direct measurements in depressed patients support this association.

Studies in Heart Disease

As succinctly reviewed by Hibbeln and Salem*30 and by Hibbeln et. al.,*133 the intake of polyunsaturated fatty acids (PUFA’s) may explain the conflict among reports demonstrating that lowering cholesterol increases depression*33, does not alter mood*34, or reduces depression*35,36.

Lowered Cholesterol & Increased Fish Oil Reduces Depression:

A prospective 5 year study*35 showed that implementing a cholesterol-lowering diet which included increased fish consumption reduced measures of depression (r=-.1, P<.02) and aggressive hostility (r=-.1, P<.08) in the Hopkins Symptom Checklist.  A Finnish study*36 found that lower serum cholesterol was associated with lower mortality due to accident and violence in coastal Western Finland [high fish consumption] but not in inland Eastern Finland [low fish consumption]*133.  Similarly, a Hawaiian study*134 in which the subjects consumed large amounts of fish, low rates of suicide correlated with low serum cholesterol.

Lowered Cholesterol & Decreased Ω−3/Ω−6 Increases Correlate of Depression:

Increased aggression is associated with decreased serotonin levels*51 which are associated with depression.  Aggression in primates increased on a diet that lowered serum cholesterol but that diet also increased the ratio of Ω−6/Ω−3 from ~6:1 on a high fat diet to ~33:1 on the low fat diet.*37  Similarly, violent impulsive offenders*38 exhibit an analogous change in this ratio when compared to diet-matched controls.

Studies in Depressed Patients

In 1996, Maes’ group examined fatty acid composition of serum cholesteryl esters and phospholipids in 36 subjects with major depression, 14 with minor depression, and 24 normal subjects.  Subjects with major depression had a significantly higher arachidonic acid (AA)/EPA ratio in both serum cholesteryl esters and phospholipids and a significantly increased Ω−6/Ω−3 ratio in cholesteryl ester fraction compared to healthy volunteers and subjects with minor depression.  Subjects with major depression had significantly lower LNA in cholesteryl esters than did normal controls.  Subjects with major depression showed significantly lower total Ω−3 PUFA’s in cholesteryl esters and significantly lower EPA in serum cholesteryl esters and phospholipids than either subjects with minor depression or healthy controls.*29

Similarly, Adams*54 showed that in patients with unipolar and not bipolar depression, the AA/EPA ratio in RBC phospholipid correlated with the severity of depression as measured by the Hamilton depression rating scale, HRS, omitting anxiety elements, (r=.472, P<.05) and the linear rating scale, LRS, (r=.729, P<.01).  Plasma AA/EPA correlated with the LRS (r=.527, p<.05).  RBC phospholipid EPA significantly correlated with the LRS (r=-.546, p<.05) as did plasma (but not RBC phospholipid) DHA (r=-.539, p<.05). Interestingly, this group has plasma PUFA values within the normal range*55  The findings did not appear explicable by differences in intake of Ω−6 and Ω−3 PUFA since EPA intake estimated from the food frequency questionnaire did not reveal any significant correlations with plasma or erythrocyte AA/EPA ratios (r=.39, P>.05, r=.1, P>.05 respectively).  These data could not determine if the high ratios of AA/EPA resulted from depression or predated the symptoms.

Low concentrations of CSF 5-HIAA are associated with depressive, suicidal, and violent behavior.*51  CSF 5-HIAA is a metabolite of serotonin and HVA is a metabolite of dopamine; both metabolites are found in the CSF.  A study of both alcoholics and normal volunteers*50 found that plasma levels of DHA positively predicted concentrations of CSF-5-HIAA and HVA in both healthy volunteers and late onset alcoholics, suggesting that dietary intake of this fatty acid may influence the function of the serotonergic nervous system and possibly reduce depressive, suicidal, and violent behavior.*22  However in early onset alcoholics and in a violent group of subjects containing both ETOH dependant and non-alcoholic subjects a negative correlation existed between plasma DHA and CSF 5-HIAA levels.  In this regard, the normal volunteers and late onset alcoholics significantly differed from the early onset alcoholics (p<.0002 for CSF 5-HIAA and p<.0009 for CSF HVA).  These negative correlations appear specifically related to violent and impulsive behavior but not to alcohol consumption and suggest that abnormalities in EFA metabolism *38, transport, or selective brain uptake or in EFA regulation of serotonin synthesis, release, metabolism, or uptake may exist in violent offenders.*50  Since the lesion affects both HVA and 5-HIAA, the abnormality is more likely in the transport, uptake, or metabolism of EFA’s.

Hibbeln et. al. conducted another study of 50 patients hospitalized for recent suicide attempts comparing 30 subjects with a primary diagnosis of depression to 20 subjects with other primary diagnoses.  Only among the 20 subjects without a primary diagnosis of depression, higher plasma concentrations of EPA predicted strikingly lower scores in 6 different psychological rating scales which relate to suicidal risk: the Suicide Assessment Scale ( r= !0.69, p<0.002), CPRS Depression score (r= !0.71, p<0.002), Guilt (r= !0.51, p<0.03) and Impulsivity from the Karolinska Personality Scale, Neuroticism (r= !0.57, p<0.05) and Impulsivity (r= !0.59, p<0.002).  No other fatty acid was related to any of the psychological rating scales.  Among these subjects CSF 5-HIAA was not related to  plasma cholesterol or to any fatty acid.  These data do suggest that some subgroups of suicidal patients may reduce their suicidal risk with the consumption of EPA;*22 however, the lack of correlation of suicidality in depressed patients with plasma levels of EPA is puzzling.  Possibly, RBC membrane levels are more sensitive and specific in the depressed group, or perhaps the lower levels know to exist in depressed patients limit the variability which obscures the relationship.  A more complete analysis awaits the availability of the full study.

Edwards et. al. *23,*24 and Peet et. al. *25 have demonstrated statistically significant reductions in the DHA level in RBC membranes of depressed patients vs. controls.  Edwards studied 10 patients with major depression compared to14 healthy non-depressed controls matched for age, social class, body mass index, number of children, smoking history and alcohol intake. The depressed subjects showed significantly lower levels of DHA, EPA, and total Ω−3FA.  The mean dietary intake of DHA and generally the other Ω−3 FAs were lower in the patients, while the mean intake of the Ω−6 FAs were higher.  There was no difference in energy intake.  Both dietary and RBCM Ω−3FA, DHA, and LNA levels showed a significant negative correlation with the BDI score.  [The correlation of LNA with BDI and in fact its emerging as the single predictor of BDI is puzzling since LNA did not differ between the depressed and the control group (.12 vs .09, p=.41).  It is especially confusing given that the combination of dietary and RCBM data in a forward stepwise multiple regression analysis, DHA (β coefficient = -.92, p=.0003) and the Ω−6FA LA (β=.48, p=.03) predicted the BDI score.  This finding is in general agreement with the hypothesis that the Ω−3/Ω−6 ratio relates to depression.  Although other findings suggest that AA would more likely correlate with the BDI, the correlation of LA with the BDI is not unreasonable since Ω−6FA are preferentially elongated vs. Ω−3 fatty acids.  These correlations contrast to the lack of correlation of EPA with suicidality in the depressed subjects studied by Hibbeln as noted above.

These findings support those of Peet et. al.*25 who studied 15 depressed patients compared to 15 controls matched for age and gender but not for tobacco use which differed between the two groups.  They also showed a significant decrease in total Ω−3 fatty acids in depressed vs controls (5.39 vs 9.04 p<.02) and in DHA (3.11 vs 5.43 p<.009).  Interestingly, the Ω−6/Ω−3 ratio was not significantly different nor were the ratios of arachidonic acid (AA) to either EPA or DHA (p<.06 but .003 required due to multiple comparisons) significantly different after correcting for multiple comparisons.  These investigators also showed that the differences in DHA between the groups was abolished by incubating RBC’s with hydrogen peroxide and hypothesized that oxidative stress or deficient defense mechanisms against oxidation may underlie the decreased DHA levels in the RBCM’s of the depressed patients.

Postpartum Depression  Pregnancy depletes maternal plasma DHA*43 probably to meet the demands of fetal development.  Multiple pregnancies progressively decrease maternal plasma DHA.  One study*42 is underway to determine if supplementing the maternal diet with 200 mg per day of DHA in late stage pregnancy will limit postpartum depression. The study has not yet shown a effect on depression perhaps because it did not supplement with EPA as well.  It has shown that the supplement improved cognitive capacity postpartum.

Clearly, the weight of the evidence supports the involvement of Ω−3 fatty acids and probably the ratio of Ω−3/Ω−6 fatty acids in depression as hypothesized Smith*1.

Effects in Other Related Psychiatric Conditions

Bipolar Disorder

A recent study in bipolar disorder*18 gave 9.6 gm/da. Of Ω−3 fatty acids (6.2 gm. EPA & 3.4 gm. DHA) vs. a placebo of olive oil in addition to “usual” treatment to 30 patients in a 4-month parallel group, placebo-controlled double-blind design measuring the time to exit double-blind treatment due to symptoms of bipolar disorder sufficiently severe to warrant a change in medication.  The differences between the Ω−3 fatty acid treated group and the placebo group showed that Ω−3 fatty acids were highly effective (p=0.002; Mantel-Cox, log-rank statistic, Chi-Square=9.990, df=1).  The study group was unstable as demonstrated by the time to 50% rate of termination of 65 days.  Eight patients entered the study while receiving no other mood stabilizing drugs.  Of these, the four who received Ω−3 monotherapy remained in remission significantly longer than did the four placebo treated patients (p=0.04; Mantel-Cox).  The authors note that although the study was not designed to provide definitive data on antidepressant effects, most of the patients receiving placebo who were considered treatment failures exhibited depressive exacerbations or recurrence.  Even among the few patients treated with only fish oil or placebo, the response to fish oil was sufficient to separate it from placebo.

One caution in interpreting the results of this study is that the olive oil was not augmented to mimic the fishy taste of the active oil as has been done in some other studies.*19  The resultant episodic fishy aftertaste in combination with the perceived clinical response allowed the Ω−3 treated patients to guess their treatment group more correctly than did the controls ( 85.7% vs. 63%).  However, that Hamazaki et. al.*19 found significant differences in aggression when comparing one group given control oil adulterated with fishy smell to another given fish oil suggests this difference in smell probably does not account for the findings.  Stoll’s study was not of conventional design*126 but that does not obviate the significance of the study’s findings.*126,*127

Aggression

Hibbeln et. al.*133 provide an excellent review of some of the literature linking long chain PUFA’s to hostility and depression and demonstrate a highly significant correlation between total long chain PUFA’s (in this paper, the sum of 20 and 22 carbon Ω-3 and Ω−6 FA’s) and CSF levels of 5-HIAA (r=.34, p<.07) and HVA (r=.39, p<.04).

Even among normal young adults, additional DHA can alter behavior.  Hamazaki et. al.*19 demonstrated that adding 1.5 – 1.8 gm per day of DHA to the diet of students substantially reduced their aggression toward others (extraggression) under stress.  In this study, Japanese students were randomized to fish oil or soybean oil plus 3% partially deodorized fish oil to make the capsules indistinguishable.  The students were tested on the P-F Study, Stroop, and Dementia detecting tests at the end of summer vacation and during final exams.  No differences emerged between the groups on either the Stroop or the dementia detecting test.  The control group showed increased extraggression during finals or near the completion date for a graduation thesis as compared to the end of summer vacation (two-way ANOVA, p=0.0022) while extraggression decreased, albeit insignificantly in the DHA treated group.  In additional studies, this group showed that with no stressor component within 30 days from any checkpoints of hostility measurements there were no significant changes in hostility either in the DHA group or in the control group.  In a third study, fourteen students took either DHA capsules or control capsules for two months while the students were under continuous psychological stress (final exams).  The plasma noradrenaline concentration was

significantly decreased in the DHA group (-31%) while unchanged in the control group.  The plasma ratio of adrenaline to noradrenaline was

increased in every DHA subject (+78%) p<0.02), and intergroup differences were significant (p<0.03).  That these results indicate that DHA controls hostility at times of mental stress.  That they were obtained in Japan makes them more remarkable since the average daily intake of DHA by the students at baseline was 220 mg. per day (about 25% of the average intake by members of the general population in Japan) yielding a plasma DHA concentration of 3% at baseline and 6% after supplementation*49 while the typical American intake is about 45 mg. per day yielding a typical plasma DHA concentration of 1% or less.*49  The finding by Mayes et. al.*48,*104 that students facing an academic examination increase the production of inflammatory cytokines (IL₆, TNFα, and IFNγ) coupled with the above study supports Smith’s hypothesis that inflammatory components insufficiently modulated by  Ω-3FA play a role in depression and support studying the application of EPA and DHA to aggressive disorders.

Putative Mechanisms of Action

The potential number of mechanisms involved in the relationship between Ω−3 and Ω−6 fatty acids and depression is large.  This discussion undoubtedly and unintentionally will overlook numerous possibilities perhaps even the most significant ones.  The reader should regard this section only as a superficial introduction to the potential mechanisms of action of these fatty acids on the psyche.  I will not attempt to review the mechanisms by which these EFA’s which exert their effects on other aspects of health (see*14 for an excellent review) despite my suspicion that such effects have psychological sequela (e.g. If a deficiency in Ω-3 fatty acids accelerates atherosclerosis, the compromised coronary vasculature may limit exercise which contributes to depression.*59,*60,*61)

In his original paper,*1 Smith argued that by competing with AA for lipoxygenase, EPA reduced the production of leukotrienes and notes that fish oil also reduced macrophage production of TNF, and IL₁.*102,*103 While the weight of evidence reviewed herein supports Smith’s hypothesis that a decreased Ω-3/Ω−6 ratio contributes to depression, the effects he notes on leukotrienes, TNF, and IL₁ do not give a direct causal mechanism.

Maes and Smith also argue that general inflammatory conditions promote depression.*48  Below note a number of mechanisms by which Ω-3FA’s may affect neural function and thereby ameliorate the course of depression:

1) Limit Effects of Pro-inflammatory Compounds

a) Limit Direct Neuronal Stimulation – The  Ω-3/Ω−6 ratio clearly alters the production of prostaglandins, eicosanoids, and leukotrienes.  Avanzino et. al.*72 have shown direct excitatory and inhibitory electrophysiological effects of small amounts (<10^-5 μmole) of some of these compounds (PGE₁, PGE₂, and PGF_2α as well as linolenic acid) on over 20% of the neurons tested in the medial reticular formation of the medulla.  Thus, immunologically active compounds modulated by the  Ω-3/Ω−6 ratio may directly affect the firing of brain stem neurons and thereby potentially alter mood.

b) Reduce AA – A key in reducing the production of pro-inflammatory prostaglandins and leukotrienes is limiting their precursor, AA.  The synthesis of AA depends on desaturases.  Desaturase is activated by dietary protein, ATP, EFA deficiencies, and insulin and inhibited by EPA, glycerol, glucagon, epinephrine, glucocorticoids, and thyroxines.  Δ5 desaturase converts dihomo-γ-linolenic acid to arachidonic acid (AA).  By inhibiting Δ5 desaturase, EPA can limit the production of AA.

But given the current American diet limiting the production of AA may be problematical.  α-linolenic acid (LNA), an  Ω-3FA, effectively competes with linoleic acid (LA) for Δ6-desaturase which converts linoleic acid to AA.  (Δ6-desaturase has about three times the affinity for LNA as for LA.)  Unfortunately, the contemporary American diet contains 20 to 25 gm of linoleic acid/da. (about 6% to 7% of calories).  (For growing subjects, the requirement for LA is 1% to 2% of calories and is much lower for adults.)   At these levels, the amount of linolenic acid to inhibit the conversion of linoleic acid to AA is unknown.  These levels of linoleic acid suffice to inhibit the conversion of LNA to EPA and DHA and may be associated with heightened or chronic inflammatory states.*95

c) Limit Production of Pro-inflammatory Eicosanoids    -Arachidonic acid (AA) is the precursor for PGE₂, thromboxane A₂, and LTB₄ which constrict blood vessels, clot platelets, release lysosomal enzymes, and increase vascular permeability.  By inhibiting the production of AA (see 3), EPA can limit the production of these inflammatory compounds.

EPA also shifts production to less inflammatory leukotrienes.  It shifts production from LTB₄ to LTB₅ which has 10% of the neutrophil chemotaxic property, 3% of the chemokineseis, 5% of the aggregation, and 1% of the potency in causing lysosomal release as does LTB₄.  EPA also reduces PAF  production by leukocytes by 25%.*94  Substituting Ω-3FA’s for linoleic (Ω−6) acid also would decrease IL₁ and TNFα secretion.

One should not feel free to exceed the paleolithic ratio of Ω-3/Ω-6 FA’s since overzealous substitution of Ω-3 FA’s for Ω-6 FA’s could have unanticipated consequences since some metabolites of AA (epoxyeicosatrienoic acids (EETs) produced via cytochrome P450 epoxygenases) have specific anti-inflammatory properties.  EETs decrease cytokine-induced endothelial cell adhesion molecule expression and prevent leukocyte adhesion to the vascular wall.*147

2) Optimize Membrane Properties

a) Fluidity – “When a cis-double bond is introduced into a fatty acyl chain, it can no longer attain a straight configuration and this intrinsic bend leads to looser packing ….. Or a more fluid or disordered membrane.”  Stubbs*79 shows that the center of lipid bilayers are more fluid with phosphotidylcholine containing DHA than one containing AA.  Others have also found effects of fluidity on membrane function:

Spector and Yorek*78 have observed changes in membrane fluidity in intact cells as a function of fatty acid saturation.  Spector an Yorek*78 found that “the neutral amino acid transport system operated more efficiently when the cells were enriched with PUFA, presumably because it membrane lipid domain became more fluid.” … They also found that the K_m and V_max of the glutamate and taurine transport systems increased with increasing PUFA.  But the complexity of membrane structure and function are reflected in their conclusion, “Some [transport carriers] are completely unresponsive…, others … mediate transport more efficiently, and … others … mediate transport less efficiently.”

Bourre et. al.*144 report that, “In nerve-ending membranes, fluidity is affected by the diet, depending on the membrane region.  Feeding the sunflower oil [contains normal amount of ω-6 but no ω-3 FA] diet compared to the soybean oil [contains ω-3 FA] diet results in less fluidity in the surface polar part of the membranes … But greater fluidity in the apolar part of the membranes.”

b) Bilayer Hydration – The greater the degree of unsaturation (greatest with DHA) the more water penetrates into the bilayer center.  “… Changes in unsaturation leading to changes in the degree of water penetration could lead to large changes in the dielectric constant at the steep part of the gradient.” Obviously, a change in the dielectric constant can fundamentally alter the electrical properties of neurons for which direct evidence exists:

Park and Ahmed*92 cultured diencephalic neurons in 4 media control, not FA supplemented, LA supplemented and LNA supplemented.  LA and LNA supplementation altered specific elements of the Na⁺ channel such that they could generate action potential at a higher frequency.  They believe that the effect results from an increased membrane fluidity.  They did not test for differential effects of DHA, EPA, and AA.  Some of the effects did not reverse when the media were altered which is consistent with the finding that some electrophysiological parameters do not change despite  Ω-3 supplementation when neonates are initially supplied insufficient  Ω-3FA.*14 pg. 252 & 270  The changes in the ERG as a function of dietary  Ω-3 and Ω−6 fatty acid composition supports the hypothesis that these fatty acids can fundamentally alter the electrical properties of neurons.*14 pgs 249-286.

Stubbs continues, “Amino acid side chains of proteins could potentially sense this and transmit the effect to a protein, leading to functional modulation.”

c) Lipid Polymorphic Properties – “One of the effects of increasing unsaturation is to lower the bilayer-to-nonbilayer phase transition temperature of phosphatidylethanolamine (PE).  This effect would be maximal for a DHA containing PE which would stay in the bilayer configuration at physiological temperatures only due to the presence of bilayer stabilizing phosphotidylcholine (PC) and other lipids.  Although the lipids may be in the bilayer configuration, the destabilizing force induced by the PE is still present.  This force has been formalized in the intrinsic bilayer curvature hypothesis.  In essence, it claims that membrane   proteins may sense this force or unstable membrane structure even while the bilayer form is maintained.  … [Experimentally confirming this potential Stubbs states] gramicidin is stored in the bilayer in an energetically unfavorable state that is allowed to move to a more stable state if the bilayer becomes destabilized.  This phenomenon may, in fact, be common in membrane proteins.  One can envisage a membrane protein conformation (e.g. An ion channel or receptor) poised in a metastable state rapidly switching to another state upon destabilization of the bilayer, triggered by areas of the bilayer rich in DHA and other highly unsaturated fatty acyl-containing phospholipids providing a region of low stability.”

All three of these mechanisms suggest that membrane proteins may be sensitive to the level of polyunsaturation and indeed evidence exits for this hypothesis.  Stubbs*79 has shown this for protein kinase C the activity of which increases and then decreases as the number of double bonds/PC increases.  Murphy*80 reviews data demonstrating effects of dietary FA alteration on both Na⁺/K⁺ ATPase and adenylate cyclase*81,*78 both of which can effect cellular function.  Maes et. al.*29 reviews studies indicating that membrane properties altered by PUFA’s affect tryptophan hydroxylase, 5-HT reuptake, monoamine oxidase activity, the brain concentration of 5-HT, 5-HIAA, and catecholamines.  Greenwood et. al.*83 adds cholinesterase to this list.

Others have come to similar conclusions:

Salem and Niebyliski have demonstrated unique properties of DHA and probably EPA in comparison to AA in membrane function.  “Our interpretation of these findings is that the free volume of the acyl chain region increases with the degree of unsaturation and is maximal with 22:6n3[DHA]. … The 22:6n-3 phosphatidylserine was preferentially associated with membrane proteins. … We favour the hypothesis that this species specifically interacts with and regulates the activity of membrane proteins.”*93  Others have shown that alterations in the diet comparable to those between formula an breastmilk, significantly alter the neuronal composition content of phosphatidylethanolamine and phosphotidylcholine ( phosphatidylserine was not measured.).*91

d) Composition – Dietary EPA & DHA increases the concentration of these components in membranes and reduces the concentration of AA, LNA, and 22:4ω6.  Supplying only EPA and not DHA results in an increase in membrane EPA and 22:5ω3 but not in DHA.  Thus not supplying both FA’s in the diet can alter membrane composition and function.  EPA is preferentially incorporated into phosphatidylcholine (PC) located primarily in the outer membrane while DHA is preferentially incorporated into phosphatidylethanolamine (PE) primarily located on the internal membrane.*14

When adequate dietary sources of EPA & DHA are not available, the DHA level in neural tissues drops markedly and docosapentaenoic acid (DPA), 22:5ω6 is substituted*142.  Particularly if the EFA deficiency includes Ω-6 FA’s, 20:3ω-9 and 22:3ω-9 FA’s are substituted.*160  Astroglia and synaptosomes may resist the deprivation more than oligodendrocytes, myelin, and cranial and peripheral nerves.  Substituting AA or DPA for DHA alters neural function.*93  Of great significance is the finding that once improper FA’s are substituted for DHA, proper neural function may not be regained even when dietary EPA & DHA are available.*159

3) Optimize Receptors

a) Function – Spector and Yorek *78, Murphy*80, and Maes et. al.*29 review data demonstrating that altering membrane lipids can significantly modify various properties of opiate, adrenergic, insulin, serotonergic, and catecholaminergic receptors.  Davidson et. al.*86 conclude that optimal function of presynaptic dopamine autoreceptors depends on a balance of  Ω-3 and Ω−6 fatty acids.

b) Density – Delion et. al.*145 have shown in rats that an Ω-3 deficient diet induces a 44% higher 5-HT₂ receptor density in the frontal cortex (P<.05), as measured by [³H] spiperone binding, without a change in endogenous serotonin concentration.  They also showed that the deficient diet lowered the density of D₂ receptors 14%, as measured by [¹²⁵I]iodosulpride binding, in the frontal cortex but not in the striatum.   lowered the concentration of endogenous dopamine.

4) Alter Nerotransmitters

a) Serotonin metabolism and transmission – Maes and Smith*48 also state, “IL₁ and IL₆ administered peripherally or centrally, profoundly alter brain metabolism of 5-HT or 5-HT transporter messenger RNA in the brain.” although they supply no reference.  They also note that 5-HT reuptake inhibitors suppress IL₆, IL_1β, IL₂, TNFα, and IFNγ release by human blood monocytes and T cells*105 providing another mechanism by which some antidepressants may work.  Obviously, effects on 5-HT metabolism could play a role in depression but the mechanism of the effect on 5-HT metabolism is not clear.  However, Delion et. al.*145 showed no change in endogenous serotonin concentrations in rats fed a diet deficient in Ω-3 FA’s

b) Limit Intercellular Glutamate – Glutamate activates NMDA receptors which stimulate phospholipase A₂ to release AA.  AA then inhibits the uptake of glutamate into glial cells.*74  AA also reduces the high affinity uptake of GABA and inhibits membrane-bound (Na⁺ + K⁺)ATPase.*75  Alterations of (Na⁺ + K⁺)ATPase should be considered especially serious since it consumes half the energy used by the brain.*144  Increasing the proportion of Ω-3FA’s (fish oil) reduced these effects in experimental cerebral ischemia.*76  If such supplementation alters the AA inhibition of glutamate reuptake in normal cerebral function, this may be one mechanism by which Ω-3FA’s might treat mood disorders and by which a relative surfeit of Ω−6FA’s might induce mood disorders.  The reduction in AA might be unusually important neonatally since the susceptibility of the CNS to excitatory amino acid-mediated injury peaks near postnatal day 7 in rats and at that time is 60 times greater than that of the adult brain.*82  The risk of a positive feedback cycle exists in this system since free radicals formed from the oxidation of AA may trigger the release of glutamate in the hippocampus.*82 Positive feedback in this system could exacerbate mood disorders and potentially generate seizures.

c) Dopamine and Noradrenaline – While Delion et. al.*145 showed that an α-linolenic deficient diet lowered the concentration of dopamine in the frontal cortex by 55% (P<.05) but not in the striatum or cerebellum.  This diet did not alter the concentrations of 5-HT or noradrenaline in any of the three regions

5) Optimize Signal Transduction

a) Second Messengers – FA’s released for membrane phospholipids or provided by diet can act as second messengers and can even substitute for second messengers of both the inositide phospholipid and the cyclic AMP signal transduction pathways.*155

Hibbeln and Salem*30 review data indicating that diets enriched with  Ω-3FA’s when compared to those enriched with Ω−6FA’s greatly enhance G_s coupling to adenylate cyclase as well as G_s number.  In addition they note, “polyunsaturate composition is the primary component determining the fractional volume of the membrane which predicts kinetics of G_s protein…” They also note “in depressed patients, … a decrease in β₂ adrenergic-stimulated cyclic AMP concentrations, which is G protein-dependent.

In her review of the role of fatty acids in gene expression, Simopoulos notes, “ AA and its metabolites have been identified as a novel group of intracellular second messengers regulating ion channels and the activities of enzymes involved in signal transduction such as GTPase-activating protein or protein kinase.”*142

b) Limit Kindling – Hibbeln and Salem*30 review in detail mechanisms which could lead to repetitive episodes of depression and Stoll et. al. Suggest a kindling mechanism underlying bipolar disorder.*18  This complex mechanism relates to NMDA receptors (see limit glutamate above).  Ω-3FA’s limit kindling by altering the chemical environment of phophatidylinositol-bis phosphate making it more resistant to hydrolysis by phospholipase C.  (This phenomenon and the limitation of glutamate noted above suggests that increasing dietary Ω-3 FA’s has some potential in treating epilepsy.)

c) Alter Calcium Flux – Hibbeln and Salem note that DHA inhibits L-type calcium channels which might decrease stimulated calcium flux and intracellular concentration which would affect a number of cellular processes.*30

d) Alter Enzyme Concentrations – Bourre et. al.*144 have shown that eliminating Ω-3 FA’s reduces Na⁺-K⁺-ATPase by nearly 50% in nerve terminals.  Obviously changes in the concentration of this enzyme could alter at least local polarization levels or the rate of their recovery and thereby alter neurotransmission.

EPA and DHA competitively inhibit both cyclooxygenase and lipoxygenase and thereby reduce conversion of free AA to pro-inflammatory eicosanoids.  This alteration may directly affect neuronal function.  Farooqui and Horrocks*82 review data demonstrating, “AA and its oxygenated metabolites modulate neuronal activity by regulating activities of a number of enzymes including some isozymes of protein kinase C and membrane bound Na⁺,K⁺ ATPases.”

6) Alter Gene Expression

Clarke and Jump*142,*148,*154 show that PUFA’s can directly alter gene expression.  They conclude, “The transcriptional response to dietary PUFA is very rapid, less than 3 h, which suggests that PUFA directly modulate gene transcription rather than exert their influence by modifying membrane lipid fatty acid composition and altering hormone release or signalling.”

Sellmayer et. al.*149 have showed that AA induces expression of early genes by autocrine or paracrine mechanisms via its conversion to PGE₂ which activates protein kinase C dependent intracellular signalling which increases early gene mRNA.  These effects are inhibited by EPA and DHA and may explain their reduction of pathologic growth.*150

DHA & EPA suppress the production of IL-1b mRNA*153 which could alter neural function as noted above.

Ω-3 & Ω-6 FA’s suppress hepatic lipogenesis at higher levels of intake than that needed to fulfill the essential fatty acid requirement for optimal growth.*142  The suppression of fatty acid synthase (FAS) in the rat liver by ingested PUFA’s occurs at the level of gene transcription.*156-*158  Inhibiting FA biosynthesis reduces the substrate for Δ9 desaturase which limits the availability of Ω-9 fatty acids for incorporation into neural membranes and normalizes membrane composition as noted above.

Interactions between genes and environmental stimuli control cortical plasticity.*152 pgs 987-1031  The early genes c-fos and EGR-1 conceivably could mediate some of these interactions.*151  Since metabolites of AA influenced by the  Ω-3/Ω-6 ratio alter the expression of these genes,*142 FA consumption could play a pivotal role in neural development.

The influence Ω-3 and Ω-6 FA’s have over the expression of c-fos m-RNA has some potential for altering neural development via the following mechanism.  Trauth et. al.*165 have shown that prenatal exposure to nicotine induces persistent c-fos elevations in the brainstem and forebrain of the postnatal rat brain.  They argue that since “c-fos overexpression evokes cell death in otherwise healthy cells*166,*167,*168,*169,*170, these results provide a mechanistic link between the early cellular insult caused by nicotine and the later appearance of functional deficits after a period of apparent normality*171,*172,*173,*174,*175.”  If Ω-3 FA’s suppress chronic elevations of c-fos, they might serve a neuroprotective function.  Since the acute expression of c-fos is a component of cellular repair,*176 the ratio of Ω-3/Ω-6 FA’s and especially the ratio of (EPA+DHA)/AA may have substantial influence over neural development and optimal neuronal repair and pruning.  Since these FA’s are present in breast milk, this potential mechanism or others comparably influenced by these FA’s raises substantial concern over their absence in formula and might account for some of the substantial advantages breast feeding holds over formula feeding as noted below.

7) Alter Neurodevelopment

Clearly, altering the neural development of the fetus could predispose to psychiatric disorders inclusive of depression.  This effect would be even more likely if any of the membranes’ structure becomes fixed with FA’s other than those required for optimal function.  As noted below, current data indicates that modern diets have altered fetal and neonatal development and these alterations probably are permanent for at least some structures.

In assessing the literature, one must be cognizant that the desaturase enzymes are more active in rats and mice than in humans.*162  This difference implies that humans are more dependent on dietary sources of long-chain PUFA’s than are rodents.  Altered neurodevelopment induced by PUFA deficiencies in rodents will likely be more extreme in humans and a negative finding in rodents holds little reassurance for humans.*160

Rapid brain growth in humans spans both the pre- and postnatal periods*161 which necessitates adequate supplies of PUFA’s during both periods.

If the consumption of Ω−3 fatty acids are vital to the mental health of adults as noted above, shouldn’t they also be important in fetal and infant development?  And if they are could some of our increasing proclivity for depression derive from abnormal development induced by the lack of adequate Ω−3FA during fetal or neonatal development?  The latter question has not been answered but accumulating evidence noted below indicates that not feeding infants breast milk damages their neural development.  Nettleton*14 pgs. 249-286 has written an excellent, relatively recent review of this topic with 152 references and makes an number of significant observations:

1) Preterm infants fed fish-oil supplemented formula had better visual acuity at 2 and 4 months of age compared with infants fed commercial formula

2) In DHA depleted monkeys,…….In spite of changes in the fatty acid composition of brain with refeeding, electroretinograms of the monkeys remained abnormal.  Quite possibly, …… If critical retinal tissue development was complete and irreversible by the time Ω−3FA were fed, the functional improvements could not be achieved.  {More recently, similar findings have been reported in felines*121}

3) In brain, DHA and AA content increases three to five times in the last trimester and again a s much during the first 3 months of life when neural tissue development is most rapid….  DHA accumulation continues during the first 2 years of life after birth.

4)  Whether the placenta can convert linolenic acid or EPA to DHA is not known.  The placenta of many species lacks the necessary enzymes to convert precursor fatty acids to long-chain Ω−3 and Ω−6FA.  The fetus depends on maternal supplies for Ω−3FA.

5) …strongly suggests that the fetus is unable to desaturate LNA to DHA and therefore cannot effectively utilize LNA as a source of DHA for brain and retinal development.

6) Mothers of the low-birth-weight infants also consumed less linolenic acid and long-chain Ω−3FA (EPA and DHA).  …. Suggesting that low intakes of EFA’s might retard placental development, thereby contributing to fetal growth retardation.

7) The cerebral cortex phospholipid content of infants fed formula showed a significant reduction in DHA compared with those fed breast milk, confirming that brain composition is affected by diet in early life. ….. Preterm infants consuming human milk were compared with those fed infant formula containing LNA from soybean oil as the sole source of Ω−3FA.  After 61/2 weeks, the formula-fed infants had significantly lower levels of DHA in their red blood cell phospholipids than infants fed human milk…. Furthermore, the DHA content of red cell lipids in formula-fed infants declined from birth levels during the feeding period, whereas the DHA of breast-fed infants increased.

8) Supplementing an LNA-containing formula with low levels of fish oil, however, prevents the loss of DHA from red cell phospholipids.

9) Human milk is the only infant food containing fatty acids of 20 or more carbon atoms.  It is also the only infant food containing DHA, the amount depending largely on the mother’s diet and, more specifically, her consumption of fish. … Women who consume fish regularly have twice the level of DHA in their milk as those who do not. …

10) Total fat content increases with the duration of lactation, but the concentration of individual fatty acids varies over time.

11) … Vegetarian women had less than half the amount of DHA [in their breast milk] as omnivores.

12) Presently, the only way to ensure that infants receive DHA, necessary for brain and retina, is to breast feed. …Evidence is accumulating that DHA-depleted or inadequately fed infants have reduced neural responses.

More recent results support this effect of long chain Ω−3 fatty acids on infant behavior.  Even in rats, the dietary content of fatty acids may affect learning.  Yoshida*85 and Yoshida et. al.*88,*90 conclude from experiments on rats fed linolenate sufficient and insufficient diets “that the biochemical characteristics of membrane surfaces of brain microsomes [from the hippocampus] are affected significantly by the learning task itself in a dietary oil-dependent manner.”

Yehuda*39 show that Ω−3 fatty acids alter look duration indicating slowed development of visual information processing:

One specific domain of infant development, visual attention, has shown consistent effects of Ω-3 fatty acid status in several studies in both monkey and human infants. Low Ω-3 fatty acid status is associated with increased look durations to visual stimuli.  Longer fixations are also found in low-birth-weight human infants and in infants exposed prenatally to alcohol.  Fixation duration normally decreases with age, and longer visual fixations have been associated with slower visual processing and with slower disengagement from fixated stimuli.  Thus, Ω-3 fatty acid deficiency may slow the development of visual information processing or impair the ability to shift attention.  Alternatively, longer fixations may reflect increased reactivity to visual stimuli.  Fixation duration was not correlated with visual acuity in either species, so that the effect on visual attention appears to be independent of the effect on acuity development.

Effects of early Ω-3 fatty acid status on other behavioral domains, such as sleep patterns, temperament, emotional reactivity and response to stress, are largely unexplored.  In rhesus monkeys, we have begun to examine the effects of long-term Ω-3 fatty acid deficiency and supplementation in some of these areas.  Juvenile and young adult monkeys with a long-term history of Ω-3 fatty acid deficiency showed increased locomotor activity and stereotyped behaviors compared with those fed high levels of “-linolenic acid.  Deficient monkeys also showed increased reactivity to a variety of social and nonsocial stimuli, including increased approach and interaction.  Juvenile deficient monkeys increased their environmental exploration in a new environment, unlike controls fed diets high in “-linolenic acid or a group supplemented with DHA.  DHA-supplemented monkeys showed higher levels of aggressive displays than the other two groups.  They also showed shorter durations of sleep during daytime hours, virtually never sleeping during times when the other groups often were observed doing so.  More extensive observational studies of behavioral patterns and activity cycles are now underway.

The increased aggression appeared in a group supplemented with DHA which is does not represent the physiological balance of EPA and DHA.  This lack of appropriate balance may explain the conflict of this data with the data and implications of other studies reviewed above.*35,*36,*37,*38,*134.  While a reductionistic approach advances many elements of science, when systems which are as complex as breast milk and evolved over millions of years are compared to alternative systems (e.g. conventional formulas), it is unlikely that statically supplementing conventional formula with one component of the time varying multicomponent system will reproduce the effects of the complex system.  The results of this oversimplification probably are reflected in the above abstract and in that below*40:

The long-chain EFA DHA (docosahexaenoic acid; 22:6n3) is highly concentrated in central nervous system tissues and retinal rod outer segments.  DHA deficiencies during pre- and postnatal development are hypothesized to adversely affect visual acuity, cognitive capabilities, and to potentially correlate with behavioral disorders. Although human and rhesus monkey breast milk contain DHA and arachidonic acid (AA), infant formulas marketed in the United States are virtually devoid of these nutrients. In humans, plasma concentrations of these essential fatty acids predict cerebrospinal fluid concentrations of the serotonin metabolite 5-hydroxyindolacetic acid (5-HIAA).

CSF 5-HIAA levels correlate with aggressive and prosocial behaviors in rhesus monkeys. Nursery-reared (formula fed) monkeys raised with age mates, and with no adults present, average lower levels of CSF 5-HIAA and more impulsive and aggressive behaviors throughout their life span. In addition, when they become adolescents, these nursery-reared monkeys consume higher amounts of alcohol compared to mother-raised (breast-fed) monkeys.  We postulated that the near absence of AA and DHA during early development may influence development of the serotonergic system, and hence social and aggressive behaviors and alcohol consumption, in nursery-reared infants. To isolate the effects of AA and DHA supplementation on physiology and behavior from the effects of maternal deprivation, a long-term study was initiated on nursery-reared rhesus macaques comparing infants fed standard formula to infants fed standard formula supplemented with DHA and AA at concentrations found in breast milk (AA and DHA each at 1%).

Control infants are fed standard formula (a 50:50 blend of Similac and Primilac formulas), whereas supplemented infants receive the standard formula plus supplementation with DHA and AA (each at 220 mg/l, apx. 1% of fat). Multiple behavioral and physiological assessments are being conducted throughout the lifespan, including neonatal reflex and temperament, blood sampling for cortisol levels, cerebrospinal fluid collection for assessment of monoamine metabolites, behavioral observation in individual cages and in social groups, alcohol consumption, and assessment of dominance status.

Results from replications 1 and 2 revealed that plasma concentrations of DHA and AA in standard formula fed infants (7 µg/ml and 30 µg/ml) were low compared to supplemented (27 µg/ml and 36 µg/ml) and mother-raised infants (32 µg/ml and 36 µg/ml). Overall, monkeys fed the supplemented formula exhibited stronger orienting and motor skills on a standard primate neurobehavioral battery than infants fed the standard nursery formula. However, these differences were marked during Days 7 and 14 only, in as much as the mean scores from the two groups converged on Days 21 and 30. This pattern suggests an earlier maturation of specific visual and motor abilities in the supplemented infants and a ceiling effect of the behavioral instrument. The fatty acid supplementation did not affect activity levels or items representative of state control (temperament), indicating that this manipulation does not affect behavioral reactivity in this age infant. In keeping with these findings, no differences between supplemented and control infants were obtained in either basal or postseparation challenge levels of plasma cortisol.

Replicating the two studies just described with actual breast milk could provide very interesting data.  In most of the studies reviewed in relating Ω−3 fatty acids to depression, either fish oil which contains both EPA and DHA or EPA levels alone or in combination with DHA or in proportion to AA were related to the decrement in depression and/or aggression.*35,*36,*37,*38,*134  Thus the above two studies have eliminated a component key to the effects of Ω−3 fatty acids on behavior.  Even raising infants without mothers introduces an exceedingly unphysiological condition.  Others have suggested differential effects of Ω−3 and Ω−6 fatty acids.  Crawford suggests that AA may be related to birth weight and DHA may relate more to the degree of maturity.*41

Additional evidence suggests that not breastfeeding infants may cause lasting neurological or psychological deficits:

1. The increased consumption of alcohol in the nursery reared monkey’s noted above is consistent with the recent report that early weaning in humans contributes to alcoholism at age 30.*45
2. Longer duration of breastfeeding has been associated with higher cognitive functioning*46,*47 pg. 68 and educational achievement in childhood.*46
3. Less breastfeeding is associated with disorders of immune regulation and autoimmune disorders*47 pgs. 249-258 and atherosclerosis*47 pg. 19 both of which are consistent with Smith’s hypotheses and the review by Maes and Smith of the role of cytokines in depression.*48
4. When one group of full-term human infants was randomized to formula supplemented by long-chain polyunsaturated fatty acids including DHA and AA among others but not EPA or to unsupplemented formula and compared at 10 months, the supplemented group significantly outperformed the control group on a three step test of problem solving.*88,*89  The authors speculate that the LCPUFA’s accelerated the development of the frontal cortex.
5. Agostoni et. al. demonstrate that the developmental quotient of full term infants 4 months after birth significantly correlates with erythrocyte phospholipid DHA (r=0.32, p=.01).*117
6. Lanting et. al. found, “Among children fed with formula-milk exclusively from birth or as a supplement to breast-milk within the first 3 weeks of life, the frequency of the neurological abnormality ‘minor neurological dysfunction grade 2’ at age 9 years was about twice as high as that among children fully breastfed at least for the first 21 days of life.”*118
7. Several additional studies adjusted for confounding factors have found that breastfeeding benefits long term cognition in term and preterm infants*98*119,*120 and that breast feeding and long chain ω3 supplementation improve objectively measurable functions:
   a) Plasma docosahexaenoate correlates with psychomotor measurements at 1 year of age.*122
   b) supplementation of preterm infants with docosahexaenoaate improves visual function.*123
   c)Differences exist in the visual performance between breastfed and formula-fed full term infants.*124,*125

d)Even more objectively, Uauy et. al. compared the VEP (visual evoked       potential) in infants fed formula, formula supplemented with fish oil,      and breastmilk.  The breast fed and the supplemented groups were      similar while the formula fed infants measures of VEP were          lower.*110

Lucas et. al. provide further support for the advantages of breast milk for brain development.  They report IQ data on 300 premature infants (mean gestational age =  31.4 weeks) some of whom had consumed breast milk and some who had been fed formula.*98  They conclude:

Children who had consumed mother’s milk in the early weeks of life had a significantly higher IQ at 71/2 – 8 years than did those who received no maternal milk.  An 8.3 point advantage (over half a standard deviation) in IQ remained even after adjustment for differences between groups in mother’s education and social class (p>.0001).  This advantage was associated with being fed mother’s milk by tube rather than with the process of breastfeeding.  There was a dose-response relation between the proportion of mother’s milk in the diet and subsequent IQ.  Children whose mothers chose to provide milk but failed to do so had the same IQ as those whose mothers elected not to provide breast milk.

They also conclude that the components of breast milk that may be important include, “long-chain lipids, which are not present in formulas, are important for the structural development of the nervous system (e.g. Docosahexanoic acid [22:6ω-3], which is accumulated in large amounts in the developing brain and retina).  Human milk also contains numerous hormones and trophic factors*99,*100 that might influence brain growth and maturation.

Even more ominous are the findings of Makrides et. al.*101 who showed reduced visual acuity at 16 and 30 weeks in fullterm infants fed formula vs. breast milk.  Even partial breast feeding did not fully correct this loss.  They also conclude:

Analysis of breastmilk from mothers in their study showed that we had overestimated the proportion of DHA as breastmilk fat and therefore supplemented-formula-fed infants received 0.32% of their fat as DHA, whereas breastmilk contributed 0.21%.  There has been a change in DHA levels in breastmilk over the past decade in Australia from 0.32% to 0.21% in our study and our data demonstrate that this lower level is inadequate to maintain erythrocyte DHA at baseline (day 5) levels.  It may be that the influence of maternal fat composition on infant development needs to be considered.

Holman el. al.*43 showed substantial declines in maternal plasma FA levels during pregnancy and lactation.  At 36 weeks gestation the Ω−3 and Ω−6 group levels were 83% & 57% of normal (p<.001), AA and EPA  levels were 65% & 42% of normal.  DHA was the most suppressed at 35% of normal.  Note that the level of various PUFA’a in normal Americans should not be assumed normal or optimal given the distortions in our natural diet discussed elsewhere in this paper.  Comparing the mean melting point (MMP), a measure of phospholipid fluidity*129 among various disease states reveals the potential significance of these changes in maternal plasma fatty acids (the higher the MMP, the less fluid the phospholipids):

They also conclude:

…our food supply now emphasizes Ω−6 PUFA at the expense of the Ω−3 PUFA which are also essential for tissue and function.*129…This has occurred despite knowledge that LA and LNA are competitive substrates in the metabolic cascade by which 18-carbon PUFA yield 20- and 22-carbon more highly unsaturated PUFA.*130 High LA suppresses the conversion of LNA to 20:5ω3, 22:5ω3, and 22:6ω3.  …Current availability of many oils rich in LA but poor in LNA may be raising the proportion of LA in our diet so high that it suppresses the utilization of the small proportion of LNA present. … The decrease in PUFA during pregnancy appears among both ω6 and ω3 FA, the latter being the greater {decrement}, due to high requirement(s) by the fetus, increased drain of ω3 acids from the mother, or marginal supplies of dietary ω3 acids.  The latter cause is suggested… …It seems reasonable to suggest that the LNA intake be increased during pregnancy, lactation, and infancy, when the requirements of ω3 PUFA are highest, during development of the nervous system, which is rich in lipids containing high proportions of ω3 PUFA.  The mental apparatus of the coming generation is developed in utero, and the time to begin supplementation is before conception.  A normal brain cannot be made without an adequate supply of ω3 PUFA, and there may be no later opportunity to repair the effects of an ω3 fatty acid deficiency once the nervous system is formed.

Even if Jacobson et. al.’s*178 finding that the differences in IQ between breastfed and formula fed infants derive from parenting are not invalidated by future studies (few of the infants in this study were breastfed for the minimum duration recommended by the American Academy of Pediatrics – 6 months of exclusive breastfeeding and continued breastfeeding until at least 12 months postpartum*179), the solid findings of adverse effects on the ERG and VER in the relevant animal models described above argue that the dictum of “at least do no harm”¹⁸² or its close cousin “keep them from harm and injustice”¹⁸¹ applies and requires breastfeeding and minimizing the differences between breastmilk and formula.

I frame these observations as, “Feeding with routine infant formula delays the development of and may permanently damage portions of the infant’s brain.”  The latter frame will more likely generate the requisite social action.

That the developing human brain uses 60% of the energy fed to the fetus and neonate*40 clearly underscores the importance and vulnerability of brain development.  Holdcroft et. al. have shown that the brains of normal mothers shrink during pregnancy.*109  If the mother’s body responds to the demands of the fetus by supplying EPA and DHA from her brain to the fetus in an attempt to protect fetal development, then inadequate EPA and DHA consumption by the mother may underlie this shrinkage and perhaps postpartum depression as noted above.  The effects of both maternal and infant nutrition on development and behavior clearly deserve substantial study although such work is complex and time consuming.  The evidence of damage to the developing nervous system of the fetus and infant by inadequate Ω−3 supply warrants substantial intervention to protect infants.

 Brief Review of Additional Potential Psychiatric Applications

Alcoholism  Although outside of the scope of this paper, the fact that chronic, moderate alcohol consumption depletes DHA from adult feline and rhesus monkey brains*52,*53 coupled with both the findings noted under neonatal development and the depression common among alcoholics suggests that particularly longer chain EFA’s may play a role in treating this disorder.

Dyskinesia  Aspirin antagonizes the reversal of dopamine induced dyskinesia by DGLA supplementation suggesting that eicosanoids may be involved in the action of endogenous dopamine in the striatum.*86  The relationship between  Ω-3 and Ω−6 FA’s and eicosanoid production suggests that investigating their effects on dyskinesia’s is warranted but one study*106 found  supplementation with linoleic and  γ-linolienic acid was marginally significant but not clinically important; however, it produced highly significant improvements in total psychopathology and schizophrenia subscale scores and a significant improvement in memory.  These FA’s are not equivalent to longer chain PUFA’s which have different and more pronounced effects.  These considerations warrant further study of EPA and DHA in dyskinesia.

Alzheimer’s Disease Kalmijn et. al.*116 showed that diets high in LA positively correlate with cognitive impairment and decline and fish oil was negatively correlated with cognitive impairment and decline.  The anti-inflammatory properties of   FA’s and the known inflammatory component of Alzheimer’s disease suggest a possible role for Ω-3 FA’s in preventing or treating Alzheimer’s disease.

Attention Deficit Disorder One study showed that boys (6-10) with ADHD had somewhat lower concentrations of 22:6n3, 20:5n3, and 20:4n6 than controls and behavioral disturbances significantly correlated only with plasma 22:6n3 and duration of breast feeding.*136

Schizophrenia A review of the extensive literature on the role of Ω-3 and Ω−6 FA’s in schizophrenia (see Laugharne et. al.*180 among others) is beyond the scope of this paper.

Conduct Disorder There is a paper on shifts in the P300 evoked response in conduct disorder and another paper on the shifts in this evoked response induced by omega 3 FA.  I have to read these to see if the FA shifts might counteract those in conduct d/o.  Sorry I’ve not had time yet.

Immunologically Mediated Disorders The moderation of the inflammatory response mediated by increasing the dietary ratio of Ω-3 to Ω-6 FA’s argue strongly for their potential in treating and possibly preventing PANDAS and other immunologically mediated psychiatric disorders.

Discussion

The above arguments predict that altering the composition of macronutrients in western societies will reduce the incidence of depression.  Nonetheless,  such broad scale modifications should not be undertaken without verifying the effects of increasing the Ω−3/Ω−6 ratio at least on populations at risk.  Such studies and further exploration of the hypotheses above may reveal some surprises.  For example, I would not have expected that thyroxine and EPA supplementation shared a common mechanism of action, the inhibition of Δ5 desaturase.

The application of the involvement of the Ω−3/Ω−6 FA ratio to psychiatric disorders is of particular relevance to children.  Psychotropics can have unknown long term effects on their brain development.  Therefore, determining if the return of their diet to that anticipated by their genes can ameliorate or perhaps prevent the development of some psychiatric disorders is, in fact, imperative.

These studies may not demonstrate the expected results if conducted exclusively with typical reductionsitic techniques.  For example, in some of the studies noted above, only DHA was used as a supplement.  Since EPA and DHA can have markedly different effects on biological functions,*113 different dose response curves*141, and synergistic effects*141 experiments in this domain should strive for the physiological balance among LNA, DHA, & EPA for which we evolved.  Gaining the full benefit of the increased Ω−3/Ω−6 ratio may require adequate antioxidant consumption and other changes in micro and macronutrients.

Antioxidants- Since Ω−3 fatty acids are preferentially oxidized*14 pg. 8,*160,*96 these studies should include adequate antioxidant coverage*131,*160 (e.g. 400 I.U. of mixed tocopheryls, 250 mg. bid of vitamin C, and perhaps others such as biotin*70,*71) and optimally would be tested in a diet meeting the parameters of a true Paleolithic diet.*6  Supplying the elements of this diet will probably require some supplementation because current activity levels do not permit the consumption of the quantities of food consumed by neolithic humans.  Assuring adequate vegetable and fiber intake may prove problematical.

Exercise- Exercise and the ratio of Ω−6/Ω−3 fatty acids in the diet may not be independent in their protection against depression.  For example, extended exercise increases glucagon which inhibits Δ5 desaturase as does EPA.  Exercise also decreases insulin which activates delta 5 desaturase.  As noted above  Δ5 desaturase converts dihomo-gamma linolenic acid (DGLA) to arachidonic acid (AA), an Ω−6 fatty acid.*67,*94  Since several of the studies reviewed above correlated increased depression with an increased AA/EPA ratio, exercise should enhance the antidepressant action of EPA by supplementing its inhibition of Δ5 desaturase.  When Δ5desaturase is inhibited, DGLA is preferentially converted to PGE₁ which inhibits the clumping of platelets, dilates blood vessels (perhaps obviating the need for Viagra*57,*58, 58a), relaxes bronchial tubes, reduces the liver’s production of cholesterol, and reduces pain and inflammation inclusive of autoimmune diseases.  PGE₁ also stimulates the secretion of some hormones but inhibits the release of insulin from the pancreas.  In contrast, with low levels of glucagon and high levels of insulin, DGLA is metabolized to AA which can be synthesized to PGE₂ which increases pain, to thromboxane A₂ which promotes platelet clumping and vasoconstriction, and to some leukotrienes (e.g. LTB₄, EPA shifts production from LTB₄ to LTB₅ which is far less active than LTB₄.*68)which promote allergic reactions.  These effects are so potent that the fasting insulin level in combination with a few other markers is now considered a significant risk factor for ischemic heart disease.*56  Thus, trials which seek to maximally reduce the rate of depression must combine regular, extended (>40 min.) aerobic exercise with a 1:1 ratio of Ω−3/Ω−6 fatty acids.  This particular interaction between exercise and the Ω−3/Ω−6 fatty acid ratio may not be the only such interaction but a complete discussion of these interactions is beyond the scope of this paper.

Limiting the activity of delta 5 desaturase may be only one of the reasons that exercise has consistently demonstrated efficacy in treating depression*59,*60,*61 despite its being underutilized.*62 Western society is now so far removed from the intensity and rhythm (roughly alternating days of intense physical activity and rest*107) that the typical American would have to add a daily 12 mile walk to match the average energy expenditure of our Paleolithic ancestors.*108.  Much of this decrement in physical activity has occurred recently. For example, between 1956 and 1990, the proliferation of labor saving devices in Britain reduced caloric expenditure in excess of basal metabolism by 65%.*6  We are unlikely to attain full psychological health without adequate aerobic and probably anaerobic exercise.*59,*60,*61,*63,*64

Macronutrients- Limiting carbohydrate intake will be important to reduce insulin secretion which stimulate Δ5 desaturase.  While some fish supply EPA and DHA (e.g. Salmon, mackerel, bluefish, among others) the consumption of others, tuna, shark, and swordfish, should be limited due to mercury contamination*132  Some standard diets should be avoided (e.g. National Cholesterol Education Panel Step II can produce EFA deficiencies.).*137  Low fat, high carbohydrate diets stimulate the synthesis of palmitic acid, one of the most damaging.*138  Obviously, such diets should be avoided in these studies.

However, the more complex the study, the more difficult it is and the more risk of error.  Thus some studies should involve simple alterations of the Ω-3/Ω−6 fatty acid ratio and others should intervene at more levels.

Unfortunately, we cannot expect alterations in the Ω-3/Ω−6 ratio to immediately reduce major depression to the baseline we most likely evolved.  First, note the demonstration that the effects of the early deprivation of Ω−3 fatty acids may not be reversible despite the normalization of their content in the total brain (I.e. When congeners of Ω−3 fatty acids replace Ω−3FA in some neural structures, the congener remains fixed and may not be replaceable by the optimal fat; thus permanently compromising function.). *14 pg 252  This phenomenon implies that at least one generation must elapse before new cohorts experience the full effect of appropriate nutrition.  If the mother does not possess adequate stores of Ω−3FA or is not sufficiently supplemented with these fats during pregnancy, then more than one generation may elapse before realizing the full benefit of improved nutrition.  Unfortunately, the analysis by Makrides*101 reviewed above suggests that at least some mothers’ breastmilk may already supply their infants suboptimal nutrition.

Even though all autoimmune disorders and inflammatory disorders are potentially addressed by altering the fatty acid composition of our diet, exploring this area might lead to treatments for disorders not currently anticipated.  For example, the inhibition of the effects of kindling by mechanisms closely related to some mechanisms by which anti-epileptic medications function*18 suggests that increasing the Ω-3/Ω−6 ratio might treat some forms of epilepsy.  That epinephrine inhibits seizures*114 and inhibits Δ5 desaturase coupled with the knowledge that exercise also inhibits kindling development*115 supports testing this potential application.

Conclusions

The evidence demonstrating that inadequate consumption of Ω-3 FA and excessive consumption of Ω-6 fatty acids contributes substantially to the increasing rates of depression and to other mood disorders and aggression is strong.

The breadth of support for this hypothesis from many disciplines warrants numerous interventions:

1. Public education to increase consumption of long chain Ω-3 FA’s.
2. Research to determine the proper dosages of EPA & DHA and the optimum ratio of these FA’s required to treat various forms of depression and other psychiatric disorders particulary in children.
3. Substantial public education and legislation, exceeding the effort made to decrease tobacco consumption, aimed at:
   a) Supplementing females’ diets with long chain Ω−3 FA’s prior to conception, during pregnancy, and during lactation
   b) Emphasizing the importance of breast feeding
   c) Requiring corporations facilitate breast feeding by their employees who have newborn infants through at least 1 year
   d) Establishing a national system for supplying breast milk to those infants whose mothers cannot breast feed
   e) Conducting research to determine the optimum composition of fatty acids in the mother’s diet before conception, during gestation, and during lactation
   f)Requiring infant formulae include both Ω-6 and Ω-3 FA’s inclusive of long chain Ω-3 & Ω-6 FA’s as recommended by the International Society for the Study of Fatty Acids and Lipids (ISSFAL)*142

These actions should be integrated with efforts informing the public of the broad health benefits*6,*9,*13,*13c,*14,*63,*64,*65,*66,*141 of increasing Ω-3 FA consumption and of an increased Ω3/Ω6 ratio in the diet.

Showing that re-balancing the  Ω-3/Ω−6  ratio in the Western diet ameliorates a wide variety of disorders and then effecting a functional improvment in this diet, will not fully realize the lesson that we can garner from these findings.  The lesson is that our bodies and minds are exceedingly complex and evolved for optimal function under conditions very different from those extant today.  When we deviate from these conditions by substantively altering our nutrition, our level of activity, and our exposure to light and sleep as well as other factors, our minds and bodies will function suboptimally and sometimes dangerously so.  Until we accept our genetic heritage and incorporate it in our lives (e.g. Increased exercise, decreased sodium &increased potassium intake*112, more sustained social groups, less work and more social interaction, etc.) we will not have the enjoyment in life for which we evolved.  We can choose to ignore our heritage and suffer the exorbitant cost of depression and other ills*14 pgs.64-248, 287-354,*63,*64,*65,*66,*177 and impose these ills on our children or we can restructure elements of our society to honor our heritage and live in better emotional balance enjoying improved health.*63,*64, and compare references *65 & *66.

In summary, by altering the  Ω-3/Ω−6 FA ratio from that we evolved to consume, our culture has experimented with a highly tuned system knowing little about the probable effects of that experiment.  Given the efficacy of evolution, we likely suffer from our imprudent adventure.

References