Altered fatty acid metabolism in patients with angiographically documented coronary artery disease

Altered

Fatty Acid Metabolism in Patients With Angiographically Coronary Artery Disease Edward

N. Siguel

and Robert

Documented

H. Lerman

Plasma lipids and fatty acids have been linked to coronary artery disease (CAD), and linoleic acid deficiency has been proposed as a risk factor for cardiovascular disease, but few studies have considered their multivariate effects or found the biochemical shifts associated with abnormal fatty acid metabolism or essential fatty acid (EFA) deficiency. We studied fatty acid patterns associated with CAD using high-resolution capillary column gas-liquid chromatography to analyze fasting plasma from 47 patients with angiographically documented CAD and 56 reference subjects. CAD patients exhibited a shift in fatty acid metabolism similar to that associated with EFA-deficient patients. Compared with reference subjects, CAD patients had (1) reduced percentages of polyunsaturated fatty acids ([PUFA] 45% v 50%, P < ,001). (2) increased monounsaturated fatty acids (26% v 22%. P c ,001). (3) higher ratios of Mead (20:3o9) to arachidonic (20:4w6) acid (0.016 v 0.013, P -c .04), (4) increased levels of 16:107 (2.10% v 1.55%, P c .OOl), and (5) higher concentrations of total fatty acids (356 v 284 mg/dL, P < .OOl), saturated fatty acids (101 v 75 mg/dL, P -z .OOl), monounsaturated fatty acids (91 v 63 mg/dL, P < ,001). PUFA (159 v 143 mg/dL, P c .Ol), 20:3o9 (0.5 ~0.3 mg/dL, P c .Ol) and 16:lw7 (7.7 ~4.5 mg/dL, P < .Ol). On indices of EFA status that depend on percentages or ratios of fatty acids or on the production of abnormal fatty acids, CAD patients were between severely EFA-deficient patients and healthy subjects, a state referred to as EFA insufficiency. Patients had metabolic shifts toward increased production of monounsaturated fatty acids and increased ratios of derivatives to precursors of ~6 fatty acids, shifts that occur when cells are EFA-deficient. Levels of EFAs were negatively correlated with levels of saturated and monounsaturated fatty acids. The percentage of 18:2w6 was positively correlated with high-density lipoprotein (HDL) cholesterol and the ratio of HDL to total cholesterol (r = .58, P c ,001, and r = .61, P < .OOl, respectively) and negatively correlated with triglycerides and total cholesterol (r = .61, P < .OOl, and r = -.24, P < .Ol, respectively). Opposite correlations with these parameters were observed with saturated and monounsaturated fatty acids. Saturated fatty acids, total cholesterol, and indicators of EFA deficiency increased and the HDL to total cholesterol ratio and PUFA decreased the probability of CAD as measured by multivariate linear regression. Reduced plasma percentages of EFAs are associated with CAD and with increased risk for CAD as measured by lipid levels. When plasma EFA percentages decline, monounsaturated fatty acids increase, and the increase is associated with reduced HDL to total cholesterol ratios (I = -.64, P < ,001). Relative EFA insufficiency is a condition with normal or increased whole-plasma concentrations of lipids and fatty acids, reduced concentrations of EFAs in plasma lipid fractions (reduced plasma % of EFAs), and increased plasma concentrations and percentages of markers of EFA deficiency such as 20:3w9 and 16:107. We hypothesize that this condition is caused by excessive levels of saturated fatty acids that saturate or disturb the transport of EFAs. We propose that the EFAs and their derivatives are major indicators of risk for CAD, and the diagnosis of relative EFA insufficiency and its optimal treatment with oils and antioxidants are significant factors in the management of CAD. Copyright (es1994 by W.B. Saunders Company

R

ISK FACTORS FOR CORONARY artery discasc (CAD) include diets high in saturated fat and cholcsterol. as well as elevated plasma cholcstcrol and triglyccrides. Essential fatty acid (EFA) deficiency was proposed as an important factor in the etiology of CAD since the I YhOs. Due to the high prevalence of CAD in the United States. if CAD is associated with EFA deficiencies. biochemical evidence of EFA deficiency should bc widcsprcad in the general population. However. the US Surgeon General’s report states that EFA deticiencics arc reported rarely in the United States’ (p. 58). Because previous studies have not found many of the biochemical changes associated with EFA deficiency such as increased 20:3wY and 16:107. a major link is missing in the heart discasc-diet hypothesis. We propose that EFA dcficicncy is found rarely hccause

982

previously used measures of EFA dcticicncy have lacked sensitivity. and we propose that cardiovascular discaze is associated with insufficient Icvcls of EFAs. To cvaluatc these hypotheses. wc anslyzcd the EFA status of patients with CAD. The major fatty acid groups arc the saturated. monounsaturatcd. and polyunsaturated fatty acids (PUFA). which consist of the EFAs and their dcrivativcs. Saturated and monounsaturated fatty acids arc nonessential bccausc humans can derive them from proteins and carbohydrates.’ Fatty acids undergo dcsaturation and elongation in the body using enzymes apparently shared among fatty acid families: their affinity for thcsc enzymes follows the ordct 03 > w6 > WY fatty acids.3 The precursor or parent fatty acids arc cu-linolcnic. linolcic. and olcic. and their clongation and desaturation products arc rcfcrred to collcctivcly as dcrivatives.J For example. the precursor tu-linolenic acid (IX3w3) has derivatives such as cicosapentaenoic (20:&J) and docosahcxacnoic (22503). The precursors I S:7wh and 1X:303 arc the (parent) EFAs. A useful formula to rcmemher is PUFA = EFA + EFA derivatives = (0-3 + cr)-h precursors) + (o-3 + w-h dcrivativcs) = w.3 + c& (totals). (The cxccptions arc small quantities of PUFA derivatives ot the monounsaturated fatty acids produced in the prcsencc of EFA dcficicncy. such as 70:3wY. and isomers or ~V(~KS ol Metabohsm,

Vol 43, No 8 (August),

1994:

pp 982-993

ESSENTIAL

FATTY

ACID

METABOLISM

IN CAD

the alxwc fatty acids.) It would be more accurate to use PEFA rather than PUFA to rcfcr to the polyunsaturated EFAs and their dcrivativcs, hut we chose to use PUFA to continue with current practice, In this report. the term PUFA includes only w3 and wh c&fatty acids. Modern textbooks of nutrition and medicine consider human EFA deficiency an cxtrcmely rare disorder, usually associated with patients rccciving total parentcral nutrition who have not reccivcd lipid supplements.’ A longprevailing belief has been that the human need for EFAs is met when tissue levels arc sufficient to prcvcnt acute signs of EFA dcficicncy. Inadequate dictarq’ intake of linofeic acid I’cads to reduced concentrations in the plasma and a shift in cellular metabolism from the production of wk family dcrivativcs to unusual WYpolyunsaturated products of olcic acid metabolism. such as 70:.7oY (Mead acid).” Thcsc metabolic shifts lead to a higher pcrccntagc of lO:?w’Y in plasma, and as a result the 10:309i20:4wh (T/T) ratio--the ratio of the concentrations of triene (203c09) to totracnc (20:3wh)-is increased and may reach values cxcceding 1.O.A T/T ratio of 0.2 or more was proposed as a criterion for EFA deficiency of the oh family.’ Using improved chromatographic conditions with greater separation of peaks, enabling more specific detection of X:309, TIT ratios in healthy individuals were rcportcd to be an order of magnitude less than this (mean t SD. O.(II 2: 0.006).“~“~” Fatty acids are integral components of cell membranes, and the fatty acid c(~mposjti~)n of a membrane has profoLlnd ctf~ct:~ on membrane fluidity and function.“’ EFA insuficicncy dcnotcs low Icvcls of EFAs and biochemical evidcncc of EFA deficiency without its obvious classic signshair loss and scborrhcic dermatitis. Because patients with scvcrc: EFA deficiency have numerous other deticiencies, it is difficult to separate clinical signs exclusively due to EFA dcficicncy from those due to a combinati(~n of dcficicnties.” EFA insufficiency could exacerbate cardiovascular disease by affecting the balance of cicosanoids” or impairing mcmhrvnc-based reactions and transport mcchanisms.” These impairments may in turn reduce cell lift. alter physiologic processes. and produce suboptimal tissue functimon. Abnormal platelet aggregation, hypertension, ahnormal glucose met~~bolism, atherosc!erosix, and hyperlipidcmia may he among the consequences of insufficient EFAs.~’ Many studies have linked fatty acids to cardiovascular di~casc~~ and lipoprotein mctabolism.‘5.ih Dietary intcrvcntion studies have provided evidence that EFA levcis modify the risk of cardiovascular disease.” Populations with Ic,w lcvcls of linolcic acid have a higher incidence of CAD,‘” and fatty acid patterns may predict my~)cardial infarction.‘” Since the IYhOs. Kingsbury ct al. among others, have found fatty acid abnormalities in patients with diffcrcnt types of atherosclerosis’“.?’ and myocardial infarction.” Lower levels of most PUFA were found in ph[)sph~)lipids of subjects with sudden cardiac death.” Low dietary intake of linoleic acid predisposes to myocardial infarctionz4 LOW linttlelc acid and 3Nwh in adipose tissue was associated with heart discasc. but there were no differences in adipose tissue 20:3wY bctwccn diseased subjects and controls.”

983

However. these studies did not accurately measure levels of 103~9 and other fatty acids in the small quantities that are indicators of EFA insufficiency (203wY is almost nonexistent in adipose tissue and is better measured in blood). Using indicators of EFA deficiency. wc found a large gap between the upper limit of reference values and those observed in severe EFA deficiency.4” Those studies suggested that there are individuals with insu~cicnt EFA lcvcls substantial enough to account for long-term hyperlipidemia and CAD. It is impossible to conduct experiments whereby subjects follow a diet deficient in EFAs for many years. WC thcrcfore chose to study a cross-section of ;I population with CAD, and found that CAD patients have the abnormal plasma fatty acid profiles consistent with EFA ~nsu~ciency. indicating that plasma EFAa are a significant additional risk factor in CAD. SUBJECTS

AND

Plasma samples were analyzed

METHODS

from -17 patients

(aged 31 to 72

years, 35 men and I:! women) who on cardiac catheterization found to have coronary occlusive disease. Ahout selected at random at one

local

documented

from patients who underwent

hospital. CAD.

not

all of

whom

were

h0 subjects were ~atheteri~dti~~n

had

angiographically

Subjects had been evaluated

due to clinical

evidence of heart disease. We used leftover blood from lipoprotein analyses at a time when practically all patients seen for catheterization had their blood analyzed

for lipids. Subjects selected were

those for whom we found leftover frozen plasma. Subject selection

offatty acid

was independent After

or lipid levels or the extent of CAD.

subjects were selected.

(CAD

group)

this study was restricted

with auhstantive

obstruction

to patients

or stenosis in at least

one major coronary artery. These patients were compared previously

reporteds

reference

group

(aged 25 to 65 years. 29 men and 27 women). of 24 healthy

volunteers

without

disease.

selected suh,jects from the Framingham

Offspring

Study (also known

acids. Mcnn levels of cholesterol, and triglycerides

of prevalent

high-density

were

We

also included

37

Study).

levels of fatty

(HDL)

lipoprotein

1X5* 52. and X0 mg!dL

subjects and 111. 35. and 15X mg:dL

respectively.

and

Cardiovascular

as the Fr~im~ngharn Heart

cycle 3. who provide a rough indication

reference

subjects

This group consisted

any known

randomly

cholesterol.

with a

of 56 non-obese

in one graph

for

for CAD

patients.

the data

from

10

patients [aged 77 to 51 years. six men and four women) with severe EFA

dcticiency

gastrointestinal reference

resulting disease.”

from

chronic

Some

of the

and severe-EFA-deficiency

reported

and are included

nonpaired

fbr equal

or unequal

reference

population

cicnt

to

purposes.‘,” Fatty

as previously described.“.‘“-2’

I test analysis.

as appropriate.

adjusted

to compare

with the subjects. Pearson correlation

analysis. and multivariate

(equivalent

analyses with

A~O~A~c~~v~~riance

graphics analyses were performed sion 3.1). Redmond.

CA)

version

S.(l). An expert

Works.

MO).

further

dummy

analysis).‘” using EXC’EI.

and SYSTAT program.

to

group have been previously

Student‘s

variance

due

acid levels of the

here for comparison

acid and lipid levels were measured We used conventional

malabsorption fatty

(SYSTAT,

“Statistical

assisted in the selection

the

coefii-

variables

Statislical

and

(Microsoft,

ver-

Evanston,

Navigator”

IL.

(Idea-

of statistical

tech-

niques, We calculated various ratios of fatty acids and cholesterols that correct for possible measurement

error+ due to water evapora-

tion from plasma (ie, a ratio of two fatty acid\ <>r cholesterols

in

lip~~pr(~t~in fractions compensates for potential ~~~~~entrati(~n measurement

errors).

To

further

illustrate

group

differences.

the

graphs present data hy group: however. to hilve space. the healthy

SIGUEL

984

reference and Framingham groups are combined as “reference subjects” for statistical analyses and table presentation. All correlations exclude the severe-EFA-deti~ien~~ group. Had that group been included, the correlations would have heen far more significant due to the extreme results of the severe--EFA-deficiency group.

PUFA w3 vie w9

w7 Sat. Fal 24.1~9

differences were found in major plasma fatty acids, and therel’ore, the data were combined for men and women in all groups (differences nonsignificant by t test and Mann-Whitney U test). Table 1 lists the means rt SE of the percentages of the total yield of identified fatty acids. In both groups, most of the 06 fatty acids are in the form of linoleic acid (18:2wh), whereas most of the w3 fatty acids are in the form of ol-linolenic acid (18:3w3) derivatives. Patients with CAD had mean values for iinoleic and cY-linolenic acid that were 84% and Y8% of refercncc values, whereas derivatives of these fatty acids such as arachidonic acid and docosahexaenoic acid were 101?; and 87% of reference values. The percentage of fatty acid as cu-linolenic acid was small in both groups, and individual variations were large in CAD patients; hence, ranges arc broad and differences between groups are not significant. Lower levels of iinoleic acid observed in the plasma of CAD patients (P < X)001) were associated with increases in the percentages of practically all monounsaturated and saturated fatty acids. Patients had percentages of lh:lw7. lX:lw7, 18:109,20:3w9, and I6:O that were 3SZq 19%, 16%. 27%. and 12% greater, respectively, than values obscrvcd in the reference population (all statistically significant [P < .OOl], except 20:309. with P < .O.S due to a high variance caused by low levels of 20:3w9). There was a less than 2% difference in 1X:0 between CAD and reference sex

Table 1. Percentages (%) of C&Fatty

Acids in Plasma of Reference

Subjects and Patients Fatty Acid 16:0

Reference Subjects (n = 56) 18.36

2 0.17

CAD Patients In 7 47) 20.59

rt. 0.18

P

24:oo 22:6w3 225~3 20:5w3 20:4w6 20:3w6 202w6 20:3w9 16:3w3 163w6 16:2W l&IVY7 18:lws 18-W 16:lw7 16:OO

.“ll.--_

I 050

1.00

150

200

Ratios Fig 1. Ratios between means of major types of fatty acids for two groups of subjects, CAD patients and healthy reference subjects [REF). NoPUFA, saturated and monounsaturated fatty acids; Sat. Fat, saturated fat.

groups. In Figure 1 the differences bctwcen gt-oups arc shown as bars proportional to the ratio of key fatty acids. Table 2 presents the totals for groups of fatty acids, and Table 3 shows the ratios for classes of fatty acids. The term PUFA rcfcrs to all the cis-EFAs plus their dcrivativcs. Significantly higher percentages of saturated ~9. and 07 fatty acids (107%. 127’ii. and 115’3 of rtfcrencc values) and significantly lower (P < ,001) values of (110and ~3 fatty

.OOOl

18:0

6.98 + 0.08

6.88 i- 0.15

.05

20:o

0.17

!I 0.01

0.11 rt 0.01

.002

22:o

0.42 i 0.02

0.30 t 0.03

,002

24:0

0.39 2 0.02

0.25 + 0.03

,001

16:lUJ7

1.55i

2.10 i 0.10

.OOOl

18:lw7

1.61 + 0.03

1.91 + 0.05

.OOOl

0.07

LERMAN

NoPUFA

RESULTS No significant

AND

Table 2. Percentages (X) of Groups of Fatty Acids in Plasma of Reference Subjects and Patients

-.....

in = 561

CAD Patients (II = 47)

26.51 z 0.18 3.16 2 0.08 18.85 rt 0.35

28.33 + 0.30 4.01 2 0.13 21.68 ” 0.41

,001 .0001

46.34

41.05

.OOOl

ReferenceSubjects

Fatty Acid

-

P

“~.-_____l_

20: 1 w9

0.13 + 0.01

0.12 + 0.01

.3

Saturated fat w9 W7

20:3ws

0.11 + 0.01

0.14 It 0.01

.05

ru6

22:lwS

0.04 + 0.01

0.04 + 0.01

.9

w3

4.03 + 0.16

3.54 i 0.11

.02

24: 1 w9

0.46 + 0.03

0.30 + 0.04

.Ol

DFAS

0.11 rt 0.01

0.14 L 0.01

.05

18:lwS

18.11

+ 0.34

21.08

-c 0.43

.OOOl

+ 0.48

2 0.57

.OOOl

18:2w6

35.24

r 0.51

29.68

2 0.52

.OOOl

DFA6

18:3w6

0.49

t 0.02

0.48

it 0.03

.7

DFA3

20:2w6

0.22 + 0.01

0.24 2 0.03

.3

Monounsaturatad

22.01

-t 0.41

25.69

i: 0 49

.oao2

20:3w6

1.70 + 0.05

1.84 rt 0.08

.I

PUFA

50.37

L 0.51

44.60

I 0.61

.OOOl

20:4w6

8.48 -t 0.21

8.58 i 0.28

.8

Non-PUFA

48.52

-r 0.52

.OOOl

18:3w3

0.47

+ 0.02

0.46 rt 0.02

.6

54.02 + 0.59 ---

20:5w3 22:5w3

0.62

t 0.04

0.62

-+ 0.03

.O4

0.57

+ 0.02

0.53 L 0.02

.07

22:&J

2.29 i 0.12

1.99 + 0.08

.06

NOTE.

Results are the mean

f SE.

11.10

i 0.21

3.55 r 0.16

NOTE.

fat

Results are the mean

Abbreviations:

DFAi,

t 0.31

.5

3.08 _t 0.10

.02

-t SE.

derivatives

PUFA.

all fatty acids

of families

family

i, non-PUFA,

all saturated

w9 = saturated

11.37

+ moRou~saturated

of fatty

acids,

family

(i - 9, 6, 3);

w3 and w6 = 013 + ~6: uai. fatty acids and fatty fat.

acids

of families

6,7 and

ESSENTIAL

FATTY

Tables 3. Ratios

ACID

METABOLISM

of Cis-Fatty

IN CAD

Acids in Plasma

985

of Reference

Subjects

and Patients Sublects (n = 561

CAD Patients

Reference

Ratios 20:3u,6/20:4w6

In = 47)

0.21 k 0.02

20:5~3/20:4w6

0.077

P

0.23 2 0.02

2 0.006

0.062

.3

2 0.003

.04

20:40,6/18:2w6

0.24 2 0.01

0.29 k 0.01

20:5w3/18:3w3

1.40 * 0.10

1.22 2 0.07

.2

18:1~17116:lo,7

1.12 f 0.04

0.99 -t 0.04

.03

.0005

w3lw6

0.087

k 0.004

0.087

+ 0.003

.9

18:3u,3118:2w6

0.014

k 0.001

0.016

2 0.001

.04

0.008

+ 0.000

0.009

+ 0.001

.03

0.006

k 0.000

0.007

-t 0.000

.5

DFASIDFAG

+ DFA3

DFASIPFAS DFAG/PFAG

0.32

+ 0.01

0.39 -t 0.01

,005

DFA3IPFA3

8.08

k 0.47

7.38 + 0.43

.3

DFA3IDFA6

0.33 k 0.02

0.27 + 0.01

.Ol

PUFAIMONO

2.35 ? 0.07

1.79 -t 0.06

.OOOl

PUFA-transinon-PUFA

1.03 k 0.02

0.81 2 0.02

.OOOl

PUFAinon-PUFA

1.05 2 0.02

0.84 t- 0.02

.OOOl

20:3w9/2O:w46

0.013

2 0.001

0.016

k 0.001

.04

16:lw7/18:2w6

0.046

k 0.003

0.073

k 0.005

.OOOl

of family

i; PFAi,

NOTE.

Results

Abbreviations: precursors

are the mean DFAi.

i SE.

derivatives

of fatty acids of family

of fatty i; PUFA,

and ~6 = w3 + ~06; wi, fatty acids family fattv acids of families

acids

all fatty acids of families

i; non-PUFA,

w7 and w9 = saturated

all saturated

+ monounsaturated

w3 and

fat.

acids (XYYLand 88%) were noted in patients as compared with r’eference subjects. These alterations resulted in lower total percentages, in patients compared with controls (P < .OOOl), of cis-PUFA (89% of reference values) and higher percentages (11%) of non-PUFA, which consist of cis-monounsaturated fatty acids, saturated fatty acids, and polyunsaturated derivatives of w9 (we found no measurable levels of 07 polyunsaturates), The patients had lower percentages of derivatives of cu-linolenic acid and higher percentages of derivatives of oleic acid in their plasma, with no diffcrcnce in derivatives of linoleic acid as compared

with the reference population. Figure 2 shows linoleic derivatives (mostly 20:406) versus linoleic acid. The total levels of 06 derivatives vary little despite wide ranges in precursor linoleic acid. The sum of all the derivatives as a percentage (w3,06, and 09, Table 2) was almost identical in both groups (14.8% in reference group and 14.6% in patients). The ratio of cis-EFAs and derivatives to cir-non-EFAs and derivatives (PUFA/non-PUFA) shown in Tahle 3. provided one of the best discriminating variables based on t test analysis between patients and reference subjects, with a P of less than .OOOl. Patients’ T/T ratios were approximately 23% higher than values for the control population. With severe EFA deficiency, the T/T ratio may increase 100-fold.” Other ratios that were significantly increased in the patients as compared with reference subjects included the ratios of l&206 and 18:3w3 to the sum of other 1X-carbon fatty acids (l&O + 18:lw7 + 18:lwY). and the ratio of lh:lo7 to l&206 (P < .OOOl). In accordance with thcoretical predictions.J the relative pathway activity as measured by the ratio of derivatives to precursors follows the order 03 > w6 > ~9. The ratio of 06 derivatives to precursors is increased in patients compared with reference subjects (0.39 I’ 0.32, P < .Ol), a finding noted with EFA deficiency. Values outside the mean + 2 SD have been proposed as indicative of EFA deficicncy.s Percentages for palmitolcic acid exceeding 2.6% or a percentage of 18206 less than 28% of total fatty acids, as well as ratios of 20:3wY/20:406 greater than 0.025 and 16:107/18:2w6 greater than 0.086. which indicate the presence of EFA cleficiency (primarily 06 deficiency), were present in many of our patients. Using the T/T ratio as an indicator, five of 47 CAD patients had EFA deficiency; using 16:lw7/18:2wh. 13 of 47 had EFA deficiency. Among the reference group, none had EFA deficiency; among the Framingham group. :an approximate

[’ REF ‘z’ FRAM (3 CAD

26

36

31

Linoleic

Acid

41

46

(percent)

Fig 2. Derivatives versus precursors of 18:2w6 (%). REF. a group of healthy reference subjects; FRAM, a random sample from the Framingham Cardiovascular Study, cycle 3; DFAG, the derivatives of linoleic acid.

SIGUEL AND LERMAN

986

them. The inverse relationship between lB:lw7 and lX:?wh is shown in Fig 3. A similar inverse relationship (not shown) exists between the ratios T/T and PUFA~n(~~~-PUFA. WC incorporated the group with severe EFA deficiency in Fig 3 to illustrate the extent of possible linolcic acid dcficicncy in humans. However. the three patients with iinolcic acid Icss than 7% apparently died within months after blood was drawn for this analysis. Figure 4 shows the highly signi~ca~~t inverse relationship (r = -.93. P < ,001) between monounsaturated fatty acids and PIJFA. Significant (t” < .Oi) inverse correlations of linoleic acid were noted with 1h:O fr = -.78f and the m~~n~~unsaturatcd fatty acids fS:lw? (r = -.6X), lX:Iu7 (Y= -50). and IX: 1~09(1.= -.71). Similarly significant negative correlations were found hetwecn ratios that serve as indicator5 of EFA deficiency (20:309/20:4w6 or 1h: 1(07/18:2o)h) and the ratio ~UFA~RO~-PUFA. The ratios of dcrivativos to precursors for the respective fatty acid families arc p~~s~tivc~ycorrtiated with each other (wh t’ w3 derivatives to precursors. I’= .31, P < ,001: w6 \’ 09 derivatives to precursors, Y= 53, P < .OOl) and negatively corrclatcd with EFAs (it, w6 dcrivativcs to precursors 1’ 18:?wh, r = --.72. P < .(I01). These data support the concept that the c~~nversi~~t~of precursors to derivatives increases for att ~lns~itu~~tcd fatty acids when levels of EFAs dccrcasc. The correlations remain signiticant whether they were produced using only the patients or the a_eregatc of the patients and the reference populati(~n.

indicator of fatty acid prevalence in the general population, four of 32 had EFA deficiency when using TIT and two of 32 had EFA deficiency when using lh:lw7/ifKIwh. Mead acid, 20:3w9. is much more difficult to measure than 16: 107 or 18:2w6 because 20:3wY is a very small peak surrounded by other peaks. Altl~ough 16: l~7/1~:2~6 is a mcwc accurate index, 20:3w9 is a more sensitive indicator of EFA deftciency. Table 3 also shows ratios of various cicosanoid precursors (20:3w6. 20:406, and 205~3, precursors of series I, 2, and 3, respectively); the 20:5o3/20:4wh ratio is lower in patients than in controls (P < .04). The total plasma fatty acid concent~ti~~n was higher in patients than in control groups, a reflection of higher choIcstcrol and triglyceride levels. We found higher concentrations of total fatty acids (356 1’ 284 mg/dL, P < .OOl), saturated fatty acids (101 v 75 mg/dL, P < .OOl), monounsaturated fatty acids (91 1’63 mg/dI,. P < .tXtl), PUFA (159 v t43 mg!dL. P < .Ulj3 2&3w9 (OS I*1L3mg!dC, P < .@I1. and 16: lo7 (7.7 t’ 4.5 mg/dL, P < .f)f; rcadcrs can compute concentrations for other fatty acids from the total fatty acid concentration and the percentages shown in other tables). Tables 4 and 5 display correlations between sclccted fatty acids and groups of fatty acids, various ratios, and lipids. There are striking inverse correlations between indicators of levels of EFA status and indicators of EFA deficiency. Percentages of plasma EFAs, PUFA, and ratios proportional to EFAs are inversely correlated with saturated and monou~saturated fatty acids and ratios proportional to

Table 4. Correlations With Ratios of Fatty Acids in Plasma of Reference Subjects and Patients Fatty Acids

16: 1~7

18:2&

.64

-.78

l&O

.87

.64

1.00

-.68

.4S

.70

.43

-.49

.32

.33

-.13

T. Chol.

28

HDL:T.

HDL

Chnl.

.53

-.4?

.02

.I4

.12

.22

.I4

.24

-.73

.50

.39

- .30

.43

-.38

.09

.13

.13

.I7

.41

16:lw7

.71 -.44

.lO

-.32

Triglycerides

-.84

-.40

.-.I8

-.I9

PUFA

-34

f8:O

.31

.41

Mu~~unsatur~t~d Fat

24:0 20:31&J

-.04

Saturated Fat

24: ?wS

-.21

.21

-.I9

- .30

.32

-.36

-.21

.fO

.2cl

X3:2&

-.68

7.00

-.7J

-.79

.90

-- .67

-.24

.58

.61

18:3&

.36

- .23

.I0

.18

--.I7

.05

.I9

.02

.Q4

20:3w6

.I8

- .30

.42

(03

.I9

.I7

.19

.22

29

- .26

.26

-.40

.03

~-.04

20

-.I7

20:4& l&3&

-.I2

.O4

-.I1

.09

.02

20:5w3

-.Of

.02

-.OJ

-25

23

.24

22:6u3

-.37

28

. .23

-.51

.50

.35

.05

.I5

.12

.O6

.Ol

.03

.2?

-19

.33

.36

-.16

-.78

Saturated fat

.49

-.73

1.00

.41

- .71

.56

.23

w9

58

-.75

.38

.98

-.91

.69

.21

w7

.95

-.71

.40

.79

- .76

.53

.32

.98

- .76

-.28

.44

- .33

-.75

w6

-.72

w3

-.28

DFAS

.43

DFAG

-.08

DFA3

-.29

Mo~ounsat~ra~edfat PUFA

.70

.92

- .?2

-.90

.22

-.I8

-.47

-.49

.32

-.23

.05

21 --.79

-.73

.90

-.19 .41 -.71

.33 -24 -.49 1.00 -.93

.43

-.39

.64

.60 -.39

.49

59

‘64

.32

.34

-.38

.09

.13

.I3

17

.I7

.34

- -09

-.Ol

.06

.46

-- .37

-- .93

.?O

.25

.-.59

64

f .oo

.77

- .29

.61

.67

-.17

Triglycerides

.5O

-.61

56

.70

-.?7

1.00

.48

T. Chol.

.39

-24

.23

25

-.29

.48

1.oo

.35

.32

.66

-.49 .Q3

.49

HDL Chol.

-.30

.58

-.39

-.59

.61

.49

.03

1.00

.83

HDL/T. Chol,

-.43

.61

-.43

-.64

.61

.66

.49

.83

1 .oo

NOTE, Significance Abbreviations Abbreviation:

values

general

rules are as follows:

as in previous tables. T. Chol., total cholesterol.

corre-etation Rz > .25 has P < .Ol;

f?z ‘)

.28 has P ( ,005; RZ i

.34 has P

.OQl.

ESSENTIAL

FATTY ACID METABOLISM

IN CAD

987

Table 5. Correlations With Ratios of Fatty Acids in Plasma of Reference Subjects and Patients Fatty Acids

16.lw7

18,2w6

Saturated Fat

20:3w6/20:4w6

.25

20:5co3/20:4&

.05

.lO

p.03

20:4&/16:206

.24

- .61

.32

.05

t

.20 -~.15 .I9

Fat

PUFA

-.31 .13 -.26

Triglycerides

T Chol

HOC

.39

.22

-.19

- .09

.02

.22

.Ol

.03

.23

-.06

p.06

-.24

.21

-.34

-.I1

.14

48

-.48

-.44

51

-.36

--.35

.09

.33

-.33

.34

.37

18:3,~~,3/18:2w6 DFA9IDFA6

Monounsaturated

.36

- .80

20:5w3/18:303 18:1~0,7/16:1<~7

-.I4

DFA3

.42

.47

.22

.25

.I8

.20

.05

01

HDL/T

-.18

Chol

-.29 .15 -.I9 .I6 .24 -.18

.51

- .49

.35

.49

-.52

DFASIPFAS

.24

-.29

.25

.04

- .I3

DFA6lPFA6

.35

-.72

.47

.26

- .38

.I1

.I0

-.30

m.29

-.I0

m.25 .Ol

DFAJ:PFAJ

-.25

.03

- .lO

.37

.33

-.41

- .22

.I7

.24

DFA3JDFAG

--.22

.32

p.18

p.34

.34

-.18

.I0

.33

.30

PUFAiMONO

- .71

.85

52

-97

.96

~ .68

- .26

.62

.67

PUFA-transioon-PUFA

-.?2

.90

p.70

-.93

.99

-.73

- .28

.62

.67

PUFAinon-PUFA

- .72

.91

m.70

.93

.99

- .73

- .27

.62

.67

20:3~.09/2o:w46

.53

- .49

.41

.48

-.54

.30

.20

--20

.26

16:1~.~>7/18:2w6

.97

- .79

.59

.74

-- .80

.57

.39

- .35

.48

NOTE. Significance

correlation

RZ > .25 has P . .Ol; R* > .28 has P < ,005: RZ > .34 has P c ,001.

The cholesterol in HDL and the ratio of HDL to total cholesterol are positively correlated with indicators of EFA status such as 1X:206 and negatively correlated with indicators of EFA deficiency such as 16:lw7. Figure 5 shows the direct relationship between the ratio of HDL to total cholesterot and 18:2w6. Similar curves are obtained plotting the ratio of HDL to total cholesterol versus the ratio PUFA~no~~-PUFA and opposite curves versus the ratios T/T or l~:l~7/1~:~~~ as the X-axes. Triglycerides were negatively correlated with IX:%,&and indices of EFA status such as the ratio PUFA/monounsaturates. and positively corrclatcd with indices of EFA deficiency such as the ratios lh:lw7/18:2~6 and 20:3~9/20:4wh. Figure 6 shows that increased levels of 16:lw7 are associated with increased plasma triglycerides (the opposite relation occurs with 18:7oh in the X-axis). Similar curves for other variables are not shown due to space limitati~~ns. The best linear regres-

sion equations with the ratio of HDL to total cholesterol as the independent variable and PUFA. saturated fatty acids (Satfat). and 18:2w6 (linoleic acid, PlX?h) and 16:lw7 (P1617) fatty acids as independent variables are as follows (in %: statistical significance P in parentheses): HDLitotal

Abbreviations

7.5

values

general

rules are es follows:

as in previous tables.

cholesterol = -0.66 (P < .0007) + 0.02 x PUFA (P < .I X lWT) + 0.44 X Satfat (P < .3X) + 0.02 X PI617 (P < 24). [F ratio = 28. P < .4 x lWrz. multiple correlation R = .68]; HDLitotal cholesterol = -0. IS fP < 39) f o.ni x PI826 (P < .2 x 10-h) - 0.003 x PI617 (P < X5), [F ratio = 30, P < .6 x 10-r”, multiple correlation R = .hl]. Using only 18:2wh as the independent variable, the independent term was -0.17: the coefhcient of 18:2w6 was .Ol and the K was .hl. about the same. A linear regression with l&206. 18:3w3, and 16:O as independent variables had an R of .62 and a negative (signi~cant) coefficient on Ih:O. Many other ~l)mbinations were considered without a suh-

7

R 6.5

1 REF

n A

,> FRAM

5.5

A

A

b

n

s,

8 CAD 2 5

4.5 3.5

c

2.5 c 1.5 i 0.5

c2

7

12

22

17 Linoleic

Acid

27

32

37

42

(I 8:2w6)

Fig 3. Linoleic acid versus palmitoleic acid (%). REF. a group of healthy reference subjects; FRAM, a random sample from the Framingham Cardiovascular Study; SEFAD, patients with severe EFA deficiency.

988

SIGUEL AND LERMAN

Q CAD

i ______--_

_F__

42

37

.-.

57

52

47

_--+----

PUFA (w3 * w6) Fig 4. Monounsaturate~ fatty acids versus PUFA (%). REF. a group of healthy reference Framingham Cardiovascular Study, cycle 3; MONO, monounsaturated fatty acids.

stantial increase in the multiple regression coefficient. Using PUFA and total monounsatured fatty acids, the multiple R was .67. the PUFA coefficient was positive (significant), monounsaturates were negative (sjgni~cant), and the F ratio was 0.41 (P < .I x lo-“). Multivariate regression with an indcpcndent dummy variable, RISK (RISK = 1 for CAD patients, RISK = 0 for reference subjects), which is an indicator of the probability of CAD (values of RISK > 0.5 indicate CAD), produced the following equations: RISK = 2.3 (P < .02) - 0.02 x PI617 (P < .8) + 0.03 x Satfat (P < 33) - 0.05 x PUFA (P < .OOOS),[Fratio = 18,P < 3 x 10mx,multipIeR = ,601; RISK = 2.3 (P < .6 x lo-‘“) - 0.03 x PUFA (P < .008) 2.6 x HDLitotal cholesterol (P < .000003), [F ratio = 44, P < .I x lo-‘j, multiple R = .69]: and RISK = 2.4 (P < .3 x lo-(‘) + 0.00017 x totalcholcst~rol (P < .I?) +

subjects; FRAM. a random sample from the

0.0005 x total triglycerides (P < .61) - 0.05 x PUFA (P < .0002), (F ratio = 19, P < .6 x IV’, multipIeR = .hO]. Similar equations wcrc obtained using concentrations rather than percentages, except that the conc~nt~ti~)Ii of PUFA was not a significant predictor variable and the percentage of PLJFA or 18:2w6 was the most significant independent variable.

DISCUSSION Patients with CAD have fatty acid profiles between thvse found in patients with severe EFA deficiency and those in healthy controls, ie, (1) increased w7 and WYfatty acids; (3) higher 16:107/18:206 ratios; (3) higher 70:3wY/20:4~6 ratios; (4) higher 16:lw7: (5) lower percentages of w6 and ~3 fatty acids; and (6) lower PUFA/noll-PUFA ratios. To

C,

0 REF

0 CAD

25

35

45

Linoieic Acid (I 8:2w6) Fig 5. HDL to total cholesterol ratio vemus linoleic acid. REF. a group of healthy reference subjects; FRAM, a random sample from the Framingham Cardiovascular Study.

ESSENTIAL

FATTY ACID METABOLISM

IN CAD

989

400

n REF 300 0 FRAM

IG .-u f u 200 ,x DQ ‘E F 100

_+_--_--

0 0

2

4

Palmitoleic

(I 6: I w7)

Fig 6. Trjglycerides versus palmitoleic acid (%). REF, a group of healthy reference subjects; FRAM, a random sample from the Framingham Cardiovascular Study. cycle 3.

our knowledge. elevated levels of the w7 and 09 fatty acids and biochemical evidence of EFA deficiency have not been reported in CAD patients to date. These abnormalities, milder than the ones reported in patients with severe EFA deficiency,4J’ may produce subtle clinical changes such as increased lipids, increased platelet aggregation, and suboptimal cell function, including reduced cell fife. One conscquencc of our hypothesis is that red blood cell survival would be reduced in patients with CAD. Although this hypothesis was not tested here, a pilot study in another set of 10 patients who underwent coronary artery bypass surgery for CAD found an increased reticulocyte count and a normal hemoglobin and red blood cell count (E.N. Siguel. unpublished data), suggesting decreased red blood cell survival. In this study we identify a confounding effect heretofore overlooked. EFA plasma levels “alone” are misleading indicators of EFA status. The term “concentration,” when used in the context of plasma lipids, means at lcast two quite diffcrcnt things. One is the total concentration in whole plasma. similar to most nutritional blood tests in medicine. Another is the (relative) concentration of fatty acids within the lipoprotein or lipid fraction of the plasma. This ‘“relative” concentration is equivalent to the amount of each fatty acid as a proportion of total fatty acids, which is what we called “percents” or “percentages.” The percents depend on how many fatty acids are measured and thus may vary from one laboratory to another. Obviously. if we measure more fatty acids. the relative concentration (as a percentage of total fatty acids measured) of any one fatty acid will decline. The relative concentration in lipids or the percent of individual fatty acids is of major importance in determining the supply of EFAs and other fatty acids to cells. 13ecausc cells acquire EFAs from lipoproteins. it is likely that the ability of fatty acids to enter into reactions is determined by their percentages in lipoproteins and at the

interfaces between lipids and membranes rather than in the total aqueous fraction (whole plasma). We propose that the concentration of fatty acids in whole plasma. including the aqueous fraction, measures the efftciency of the fatty acid transport and utilization mechanism. To maintain an adequate supply of PUFA to cells, when the percentage of PUFA decreases in lipoproteins, the body compensates by increasing the total plasma concentration of lipids. Thus, accurate diagnosis requires a concurrent consideration of lipid and fatty acid concentrations in whole plasma and lipoproteins (plasma percents). The percentages of fatty acids in whole fasting plasma reflect the way the body produces different lipids to meet cell needs, ie, it is a summary of the relative amount in each lipid fraction (such as ph~)spholipids~ cholesterol esters, and triglycerides). Until more is known about the specitic mechanisms for regulation of individual lipid h&ions, we propose that whole-plasma fatty acids are the best measures, and our results indicate that they provide meaningful diagnostic information. Two or more dimension relationships among percentages are useful to identify metabolic abnormalities. For example. Fig 2 assists to identify a subject with low linoleic and arachidonic acid possibly due to defective metabolic conversion. Subjects may have high wholeplasma concentrations of PUFA, associated with the hyperlipidemia so common in the United States. and yet have hiochemicaf evidence of EFA deficiency at the cellular level as shown by a decreased plasma EFA percentage and increased 20:3w9. We called this condition relative EFA insufficiency or deficiency. This deficiency could contribute to CAD by forcing cells to make and membranes to incorporate more cholesterol to maintain membrane fuidity.” We propose that hyperlipidcmia is more an indicator of abnormal fatty acid metabolism or EFA deticiency than a mere presence of excessive cholesterol. ltl~identally, WC found significant biochemical evidence of EFA deficiency

990

in our small sample of Framingham Heart Stuciy subjects, cycle 3 (Figs 3 and 4). consistent with the view that EFA deficiency is the major cause of hyperlipidemia and CAD? We need to wait for further analyses of a much larger Framingham sample to determine whether EFA deficiency is a significant problem in the US populatioli. Decreases in plasma linolcic acid were associated with increases in monounsaturated fats and their derivatives. As EFA deficiency becomes more marked, monounsaturatcd fatty acids and their derivatives increase while levels of EFA derivatives remain relatively unchangcd.J These combined findings indicate that with increasing deficiency of EFAs, there is enhanced formation of derivatives of linoleic. cr-linolenic, and oleic acid. Humans may produce monounsaturated fatty acids from other precursors (ic. glucose or saturated fatty acids).j’ Figure 4 suggests that humans have a system to rcgulatc the production of monounsaturates in accordance with the mixture of fatty acids in the diet. The finding that the total percentage of oh fatty acid derivatives remains fairly constant over a wide range of linolcic acid levels further suggests that the body attempts to maintain constant the plasma pcrcentagc (not concentration) of EFA derivatives plus w9 PUFA (~7 PUFA is negligible). Because Ih:lw7 levels are mostly due to endogenous production and arc associated with EFA deficiency, the correlations and linear regressions support the hypothesis that EFA deficiency is associated with a decreased HDL to total cholesterol ratio and increased triglycerides. Our data further indicate that an increased risk for CAD is associated with a decreased HDL to total cholesterol ratio and decreased PUFA. We thus extend to CAD the findings of Miettinen et al’” (serum fatty acids predict myocardial infarction) and Valek et aI”’ (linoleic acid in serum total lipids was the single most effective predictor of cardiovascular death in postinfarction middleaged men}. We found no evidence that increased plasma concentrations or percentages of monounsaturated fatty acids have beneficial effects on CAD. Neither saturated nor monounsaturated fatty acids nor total cholesterol or triglycerides substantially increased the predictability of the model relating lipids to CAD risk as measured by the multiple correlation coefficient R after c~}nsidering the effects of PUFA. Although monounsaturated and saturated fatty acids arc statistically significantly associated with CAD, such effects are minor after adjustment for PUFA, and a casual relationship has not been demonstrated. If the increased plasma levels of monounsaturates were primarily due to increased dietary intake, ie. 1X:1~9 and 18: 1~7, such fats would probably have turned off production of lh:lw7, contrary to our finding of increased 16:1~7. It seems highly unlikely that humans could cat more monounsaturatcs and less polyunsaturates to produce the linear relationship shown in Fig 4. Maintaining fairly constant saturated fatty acid levels over a wide range of mon~)unsaturatcd intake is almost impossible. We therefore propose that humans closely regulate production of monounsaturatcd fatty acids to maintain plasma levels inversely proportional to PUFA levels as a homeostatic mechanism to maintain mcmbranc

SIGUEL AND LERMAN

fluidity. Many currently proposed dietary treatments rcduce both saturated fatty acids and PUFA and incrcasc carbohydrate intake (which arc convcrtcd to saturated fatty acids unless used by the body). Consider two subjects eating isocaloric diets, subject A. a diet high in saturated fatty acids. and subject B. a diet high in carb(~hydr~~t~. It’ A exercises more than B dots or has a faster metabolism, the net effect would be that B accumulates more saturated fatty acids than A. Short-term studies feeding mixtures of differcnt types of fatty acids rarely substantially change totalbody fatty acid composition. Thus. the result5 of crosssectional studies that rclatc body lipid c~)nlp~)siti~~n accumulated over a lifetime arc likely to ditfer from the outcome of studies that feed different types oF l’ats.“’ Hcgsted came to a similar conclusion after reviewing a large number of studies. Hc found that saturated fatty acids increase and PUFA decrease total cholesterol, whereas monollnsatu~tcd fatty acids have an insigni~c~tnt cffcct.” Hegsted also points out that patients with heart disease have reduced levels of linolcic acid and notes that the percentage of 182wh was highest in the adipose tissue of Italian men (when compared with men in North Karclia, Southwest Finland. and Scotland). Our results support the hypothesis by Sinclair that the ratio of EFAs to non-EFAs (eyuivalcnt to our ratio of PUFA/non-PUFA) is “the most important factor in atherosclerotic discasc and in coronary thrombosis.“‘j WC also agree with Sinclair”’ and Ahrcns3 that a low-fat diet. it deficient in EFAs, may not lead to a reduced risk of CAD and may actually increase the risk hccause of the hiochcmical conversion of carbohydrates and protein to saturated fat. We propose that plasma cholesterol and triglyceride levels arc determined or regulated in great part by EFA metabolism, which is in part determined by the total amounts of each type of fatty acid in the body. WC conxidcr it essential that studies to cvaluatc the risk-protecting eticcts of different mixtures of fatty acids in the diet analyze whole-plasma and/or adipose tissue fatty acid composition of subjects on different long-term diets to determine thu net effects of diets varying in carbohydrate and fat and their interaction with cxcrcise, food processing (ie. hydropenation). oxidation, and other variables affecting net c.is-EFA intake. Previous research has shown that 2O:Jwh increases with mild linolcic acid deficiency and decreases with more scvcrc linolcic acid deficiency. Our data show that, on avcragc. CAD patients had higher 70:4wh than controls, but with greater variability (Fig 2). The range of values of ~(7 derivatives is much greater for CAD patients than for the reference group (compare the standard errors in Tahlo 7). consistent with a wider range of biochemical conditions among patients. The ratio 20:3wh/20:4wh is an indicator of EFA status and has been used as a marker in some clinical conditions3” Wood et alZi found reduced levels of 70:3wh in patients with heart disease. WC found opposite results, consistent with incrcascd wh metabolic pathway activity (ie, elongation and desaturation) in the prcscncc of EFA insufficiency. The diffcrcnces are most likelv due to the use by Wood ct atA of older and far less scnsitivc separation

ESSENTIAL

FATTY ACID METABOLISM

IN CAD

techniques. Packed columns or short capillary columns poorly separate 20:206. 20:309, 20:3wh, and 22:O from numerous other peaks in that region. Fatty acids 20:3w6, X406. and 20503 are eicosanoid precursors, and their ratios arc indicators of relative activity of each cicosanoid series (I. 2. or 3). Decreased 20:503/20:406 ratios are associated with decreased bleeding time and increased platelet aggregation. The dccrcased 2O:Sw3/20:4w6 ratios in CAD patients may cause increased platelet aggregation, a risk factor for stroke and CAD. (While one of the authors did a rotation through the clinical laboratory of a major teaching hospital, he found that patients who underwent bypas?, surgery had reduced bleeding times (E.N. Siguel. unpublished observation). There is a huge literature on the roles of w3 and 06 fatty acids in health and disease. 17.ixWe suspect that conflicting results about the roles of PUFA are due to insufficient amounts of antioxidants in diets high in PUFA.“” PUFA, particularly w.3 derivatives, have been proposed for the trcatmcnt of a variety of cardiovascular diseases.“’ The active conversion of EFA precursors, linoleic and a-linolenit acids, to their dcrivativcs (Table 2) suggests that many pcoplc obtain adequate w3 and w6 derivatives from ol-linolcnic and linoleic acid. perhaps with fewer side effects (such as decreased platclct aggregation associated with increased intake of w3 derivatives). Bccausc humans can make arachidonic acid (06) and the elongation cnzymcs arc shared by w3 pathways, it appears that humans can product a-linolenic acid derivatives. A patient fed intravenously only a-linolcnic acid formed a-linolenic acid derivatives.” Howcvcr, little is known about the rate of conversion of precursors to dcrivativcs and under what conditions the conversion is sufficient to meet the body needs. Patients with reduced 20503/20:4w6 ratios are likely to benefit from incrcascd intake of w3 fatty acids. A trial period of soybean or walnut oil (high in both l8:2w6 and 18:3w3) together with appropriate antioxidants may be the best way to increase EFAs and decrease cholesterol (walnuts decreased total cholesterol in a controlled experimentJ’). After soybean oil supplementation for several months, fatty acid analysis helps to determine whether 18:303 is sufficient to incrcasc the 20:503/X:406 ratio or whcthcr fish oils arc necessary. Individuals with reduced w3 derivative to precursor ratios arc likely candidates for w3 from fish oils (rather than from oc-linolcnic acid in vegetable oils). The cause of the decreased enzyme activity for converting 18:3w3 to 205~3 (EPA) could be an unknown nutritional deficiency (such as zinc or copper deficiency). a disease process, or genetic differcnccs. Whatever the rcason, fish oils bypass the metabolic block. It is now well established that 03 derivatives increase HDL cholcstcrol and reduce triglycerides, total cholesterol. plasma fibrinogen. and diastolic blood pressurc.J’ Eating vitamin E in addition to tish oil further reduces triglycerides and fibrinogen levels while reducing the formation of malondialdehydc, suggesting inhibition of fatty acid pcroxidation by vitamin E.j3 Furthermore, w3 and wh derivatives (ic. fish and plant oil extracts such as primrose or borage oil. which contain y-linolenic acid) may meet the EFA requirements with fewer calories than

991

cy-linolenic or linoleic acid (companies arc now placing great emphasis on producing “designer lipids” containing high quantities of EFAs and their derivatives).l’ Individuals with active conversion of a-linolenic to EPA may prefer w3 from vegetable sources. Vegetable oils are cheaper, taste better. are more stable, and may have fcwcr side effects than fish oils. It is now generally acccptcd that optimal treatment requires a mixture of 03 and w6 fatty acids and their derivatives because optimal health dcpcnds on an optimal ratio of 03iw6 fatty acids.” Are Low Lrrds ~~‘EFA.Fa Factor in H~perlipidewia? It is known that increased intake of w3 and/or 06 fatty acids reduces serum cholesterol and triglyceride levels. Does a diet low in w3 + oh fatty acids lead to hyperlipidcmia? It is impractical to answer this question experimentally. Feeding people a diet high (or low) in saturated fatty acids and low (or high) in w3 + w6 for a period long enough (years) to substantially alter fat deposits would be either unethical because such a diet may lead to cardiovascular disease and stroke or impractical due to compliance problems. Our alternative approach studied a cross-section of subjects with a wide range of plasma fatty acid levels and found an inverse relationship between PUFA and hypcrlipidcmia. It is acccptcd that hyperlipidemia is a risk factor for cardiovascular discasc. Thus, our evidence for a rclationship between EFA levels and plasma lipids provides a key link for the nutrition-heart disease hypothesis and a biochemical basis for hyperlipidemia. WC have noted that patients with CAD have low percentages of plasma EFAs, which suggests that EFA insufficiency is a risk factor. Whether cholesterol is the link between EFAs and coronary disease or, as wc hypothesize, EFA metabolism is the more fundamental risk factor, the study of EFAs will help disentangle some of the complexities of lipid metabolism in the etiology of cardiovascular disease. Because humans do synthesize cholesterol and do not synthesize EFAs, but do synthesize fatty acids associated with EFA insufficiency, ie, 20:3w9, which is rarely found in foods. some of the problems of circularity in distinguishing causes from effects arc reduced. The observed increases in 16: lw7 (inversely proportional to 18:206) and 20:3w9 cannot be produced by eating those fatty acids. but must be produced by the body. The fact that lh:lo7 is negatively correlated with the HDL to total cholesterol ratio and positively correlated with triglycerides suggests that low HDL and high triglycerides are caused by low plasma levels of EFAs, which crcatcs a state of EFA deficiency in cells, leading to increased monounsaturate production. Whether the reduced percentages of EFA plasma levels arc due to reduced dietary intake or some hypcrutilization of EFAs by CAD patients does not affect our findings. CAD patients have increased plasma concentrations of EFAs and 20:309 and 16:107. yet have reduced percentages of EFAs, a condition named relative EFA insufficiency.j” Our hypothesis is that relative EFA insufficiency is caused by too much saturated fatty acids, leading to suboptimal transport of EFAs because the transport mechanisms (ie, lipoproteins) are saturated with saturated fatty acids and fail to deliver sufficient EFAs to

SIGUEL AND LERMAN

992

the tissues, causing cells to become marginally deficient thereby increasing production of 16:107 and 20:309.

and

Our results indicate that the body compensates for EFA insufficiency in two ways, (1) by maintenance of percentages of derivatives of w3 and 06 fatty acids, probably the physiologically active EFAs (ic, maintaining constant the concentration of EFA derivatives in lipoproteins); and (2) by accumulation of monounsaturated fatty acids. The increased metabolic activity (fatty acid elongation and desaturation) helps to regulate the amount of EFA dcrivatives available for cell function. and substitutes the need for EFAs with monounsaturates (a monounsaturatc is closer to a polyunsaturate than is a saturated fatty acid). These alterations may take place in EFA deficiency to maintain membrane fluidity. We report that many patients have EFA insufficiency characterized by reduced plasma percentages of EFAs and increased levels of 20:3&J and 16:107 in the presence of normal to elevated concentrations of all fatty acids. The prevalence of such individuals and the clinical significance

of such deficiency remains to bc determined in studies involving a large population such as that of the Framingham Heart Study. Further research is required to identify the optimal treatment mixture of derivatives versus precursors, and w3 versus w6 versus monounsaturated fatty acids and the need for antioxidants. As Kawahara noted. “Exccssivc intake of fatty acids could result in a problem of imbalance of the rcspectivc PUFAS.“~’ In the meantime. our recommended initial treatment is to provide a diet that will bring the plasma fatty acid profile of a patient closer to the profile of a healthy reference population. Diagnosis and treatment of fatty acid abnormalities through nutrition is chcapcr than drugs or surgery and could play a major role in prevention of cardiovascular disease.

ACKNOWLEDGMENT The use for clinical diagnosis of some methods presented in this report is covered by US Patent No. 50751111. The authors express their appreciation to Deeb Salem. MD. Chief. Division of Cardiology. New England Medical Center. for providing access to plasma samples and records of catheterized subjects.

REFERENCES I. The Surgeon General’s Report on Nutrition and Health. Washington. DC. US Department of Health and Human Services. PHS publication no. X8-50210, 1988

2. Mead JF. Alfin-Slater RB. Howton DR. et al: Lipids. Chemistry, Biochemistry and Nutrition. New York, NY, Plenum. 1986 3. Brenner RR: Effect of unsaturated acids on membrane structure and enzyme kinetics. Prog Lipid Res 23:69-96, 19X4 4. Siguel EN, Maclure M: Relative enzyme activity of unsaturated fatty acid metabolic pathways in humans. Metabolism 36664. 669.1987 5. Hirono H. Suzuki H, Yutaka I, et al: Essential fatty acid deficiency induced by total parenteral nutrition and by medium chain triglyceride feeding. Am J Clin Nutr 30:1670-1676. 1977 6. Rivers JPW, Frankel Med Bull 37:5Y-64. 1981

TL: Essential

7. Lundberg WO: The eicosatrienoic acid essential 235.1980

fatty acid deficiency.

Br

significance of cit. cis. ci.c 5. 8. 1I fatty acid deticiency. Nutr Rev 38:233-

8. Siguel EN. Chee KM, Gong J. et al: Criteria for essential fatty acid deficiency in plasma as assessed by capillary column gas-liquid chromatography. Clin Chem 33:1869-1X73, 1987 9. Siguel EN, Blumberg JB. Caesar J: Monitoring the optimal infusion of intravenous lipids: Detection of essential fatty acid deficiency. Arch Pathol Lab Med 110:792-797. 1986 IO. Rivers JP, Frankel Med Bull 37:5Y-64. 1981

TL: Essential

fatty acid deficiency.

Br

11. Siguel EN, Schaefer EJ: Aging and nutritional requirements of essential fatty acids. in Beare J (ed): Dietary Fats. Champaign. IL, American Oil Chemists Society, 1989, pp 163-189 I?. Horrobin DF: The regulation of prostaglandin biosynthesis by the manipulation of essential fatty acid metabolism. Rev Pure Appl Pharmacol Sci 4:339-3X.7. 1983 13. Siguel EN: Cancerostatic Cancer 4:285-289. 1983

effect

of vegetable

14. Schaefer EJ. Rees DG, Siguel EN: Nutrition. and atherosclerosis. Clin Nutr 599-I 11, 1986 15. Harris

WS: Fish

oils and

plasma

lipid

and

diets.

Nutr

lipoproteins. lipoprotein

metabolism in humans: A critical review. J Lipid Res 30:7X5-X07. IYXY 16. Horrohin DF. Manku MS: How do polyunsaturated fatty acids lower plasma cholesterol levels’? Lipids 1X:55X-562. 1YX.i 17. Warren SE. Siguel EN. Gervino E. et al: Efects of cod livtx oil on plasma lipids. eicosanoids. platelet aggregation. and exercise in stable angina pectoris. J Appl Cardiol 3:227-236. IYXX IX. Oliver MF: Diet and coronary heart disease. Br Med Bull 37:4Y-58. 19x1 19. Miettinen TA. Naukkarinen V. Huttunen JK. ct al: Fattyacid composition of serum lipids predicts myocardial infarction. Br Med J 2X5:993-996. 1982 20. Kingsbury KJ. Morgan DM. Stovold R. et al: Polyunaaturated fatty acids and myocardial infarction. Follow-up of patients with aortoiliac and femoropopliteal atherosclerosis. Lancet 2: 1325. 132’). 1969 21. Kingsbury KJ, Brett C, Stovold R. et al: Abnormal fatty acid composition and human atherosclerosis. Postgrad Mrd J iO:425440, 1974 22. Kingsbury KJ: Polvunsaturatcd fatty acid\ and myocardial infarction. Lancet 1:64X-h7h. 1970 23. Luostarinen R. Boherg M. Saldeen T: Fatty acid compo
ESSENTIAL

FATTY ACID METABOLISM

IN CAD

29. Cooper RA: Abnormalities of cell membrane fluidity in the pathogenesis of disease.N Engl J Med 297371.374. 1977 Xl. Grundy SM: Monounsaturated fatty acids. plasma cholesterol. and coronary heart disease. Am J Clin Nutr 45:1168-l 175, lYX7 (suppl5) il. Martin DW Jr. Mayer PA. Rodwell VW. et al: Harper’s Review of Biochemistry (ed 12). Lange Medical, Los Altos. CA.

19x5 32. Valek J. Hammer J. Kohout M, et al: Serum linoleic acid and cardiovascular death in postinfarction middle-aged men. Atherosclerosis 54: Ill- 1 IX. 1985 33. Hegsted DM: Dietary fatty acids, serum cholesterol and coronary heart disease. in Nelson GJ (ed): Health Effects of Dietary Fatty Acids. Champaign, IL. American Oil Chemists Society. lY90 34. Sinclair H: Dietary fats and coronary heart disease. Controversy. L..mcet. I:414415. 1980 35. Ahrena EH: Dietary fats and coronary heart disease: Untinished business. Lancet 2:1345-1348, 1979 36. Christophe AB. Holman RT: Diet and EFA status in cystic tibrosis patients. Inform J Am Oil Chem Sot 3:449-45.5. lY92 37. Sinclair A. Gibson R (eds): Essential Fatty Acids and Eicosanoid5. Champaign, IL. American Oil Chemists Society, 1993 3X. Nelson GJ (ed): Health Effects of Dietary Fatty Acids. Champaign. IL. American Oil Chemistry Society. lY90

993

39. Cunnane SS: Antioxidants, free radicals J Am Oil Chem Sot 4:934-9X. 1993 40. Leaf A. Weber PC: Cardiovascular N Engl J Med 3 18549-557. 1988

and PUFA.

Inform

effects of n-? fatty acids.

41. Sahate J. Fraser GE, Burke K. et al: Effects of walnuts on serum lipid levels and blood pressure in normal men. N Engl J Med 3X:603-607. 1993 42. Haglund 0. Wallin R, Luostarinen R. et al: Effects of a new fluid fish oil concentrate. ESKIMO-3. on triglycerides. cholesterol. fibrinogen and blood pressure. J Intern Med X7:347-353, lY90 43. Haglun 0, Luostarinen R. Wallin R. et al: The effects of fish oil on triglycerides. cholesterol, fibrinogen and malondialdehyde in humans supplemented with vitamin E. J Nutr 121:lhS-169, 1991 44. Mounts lY93

TL (ed):

Designer

Foods.

Inform

J 4:1X+1272.

45. Lands WEM, Libelt B. Morris A. et al: Maintenance of lower proportions of (n-6) eicosanoid precursors in phospholipids of human plasma in response to added dietary (n-3) fatty acids. Biochim Biophys Acta 1180:147-162, 1991 46. Siguel EN: Nutrient Support Serv X:24. 1988

charts:

Essential

fatty

acids.

47. Kawahara Y: Progress Am Oil Chem Sot 3+X-667.

in fats. oils food technology. 1993

Nutr

Inform J