Human and Clinical Nutrition

Supplementation with an Algae Source of Docosahexaenoic Acid Increases (n-3) Fatty Acid Status and Alters Selected Risk Factors for Heart Disease in Vegetarian Subjects1'2
JULIE A. COHQER AND BRUCE J. HOLB3

Department of Human Biology and Nutritional Sciences, University of Guelph, Guelphi, Ontario, Canada NI G 2W1
In contrast to the omnivorous diet, the vegan diet is devoid of the long-chain (n-3) fatty acids, eicosapentae noic acid [EPA,4 20:5(n-3)] and docosahexaenoic acid [DHA, 22:6(n-3)], commonly found in fish and fish oils. Although vegans consume no EPA or DHA (Pan et al. 1993, Sanders et al. 1978), vegetarians consuming eggs (Simopoulos and Salem 1992) consume small amounts of both. Thus, in the vegetarian diet, a-linolenic acid [aLNA, 18:3(n-3)] is the major dietary (n-3) fatty acid. The efficiency of conversion of this fatty acid to DHA, in healthy adults, via desaturation and elongation, ap pears to be ~5% (Emken et al. 1994). Various studies have confirmed that vegans/vegetarians have lower se rum and/or platelet phospholipid levels of DHA than do omnivores (Reddy et al. 1994). DHA is found in high levels in the brain and retina, where it functions in mental performance and visual acuity, respectively (reviewed in Neuringer et al. 1988). Several investigators have shown that dietary fish/fish oils containing EPA plus DHA may help reduce the risk for developing cardiovascular diseases (CVD) (for review see Herold and Kinsella 1986). Fish oils con1Supported by the Heart and Stroke Foundation of Ontario. J.A.C, was a recipient of a Heart and Stroke Foundation of Ontario research fellowship. 2The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact. 3 To whom correspondence and reprint requests should be ad

ABSTRACT The purpose of this double-blind study was to investigate the influence of dietary supplemen tation with an algae source of docosahexaenoic acid [DMA;22:6(n-3)], devoid of any eicosapentaenoic acid [EPA; 20:5(n-3)], on serum/platelet DHA status, the estimated retroconversion of DHA to EPA, and risk factors for heart disease in vegetarian subjects. Healthy vegetarians (12 male, 12 female) consumed nine capsules daily of either DHA (1.62 g/d) or corn oil for 6 wk. Consumption of DHA capsules increased DHA levels in serum phospholipid by 246% (from 2.4 to 8.3 g/100 g fatty acids) and in platelet phospholipid by 225% (from 1.2 to 3.9 g/100 g fatty acids). EPA levels increased in serum phospholipid by 117% (from 0.57 to 1.3 g/100 g fatty acids) and in platelet phos pholipid by 176% (0.21 to 0.58 g/100 g fatty acids) via metabolic retroconversion; the estimated extent of DHA retroconversion to EPA was 11.3 and 12.0%, based on the serum and platelet analyses, respec tively. Arachidonic acid [AA; 20:4(n-6)J levels in serum and platelet phospholipids decreased moderately dur ing the trial period (DHA group) as did both docosapentaenoic acids |22:5(n-6) and 22:5(n-3)|. Although no significant changes were found in the total and LDLcholesterol levels with DHA supplementation, the total cholesterohHDL-cholesterol ratio showed a moderate decrease over time as did the LDL-cholesterol:HDLcholesterol ratio and serum triglycrideconcentra tions. DHA supplementation did not alter the various thrombogenic factors measured. In conclusion, DHA supplementation markedly enhanced the DHA status (of serum and platelets), provided for the formation of substantial EPA, and lowered the total and LDLcholesterohHDL-cholesterol ratios. J. Nutr. 126: 3032-3039,1996.
INDEXING KEY WORDS:

Downloaded from jn.nutrition.org by on October 10, 2009

dressed. 4Abbreviations used: AA, arachidonic acid; ACD, acid citrate dex trose; CVD, cardiovascular disease; DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; aLNA, a-linolenic acid; PPP, plateletpoor plasma; PRP, platelet-rich plasma; PUFA, polyunsaturated fatty acids; TxA2, thromboxane A2; TxB2; thromboxane B2.

humans vegetarian DHA supplementation cholesterol thrombogenic factors

0022-3166/96 $3.00 1996 American Institute of Nutrition. Manuscript received 5 June 1996. Initial review completed 5 July 1996. Revision accepted 22 August 1996.

3032

DMA AND HEART DISEASE RISK FACTORS

3033

taining EPA plus DHA lower serum triglycridelevels in humans (for review see Harris 1989) and reduce blood platelet reactivity (for review see Weaver and Holub 1988). These latter effects have often been attrib uted indirectly to EPA due to its predominance in most fish oil supplements/concentrates studied to date. Iso lated studies with DHA-enriched preparations con taining significant levels of EPA have shown a reduc tion in platelet aggregation (von Schacky and Weber 1985) and a lowering of serum triglycrideconcentra tions in human volunteers (Sanders and Hinds 1992). DHA has been found to inhibit human platelet reactiv ity in vitro (Gaudette and Holub 1991). Although vegetarians tend to be at overall decreased risk for CVD (Dwyer 1988), due partly to their lower serum cholesterol levels, their thrombogenic risk fac tors may be significant (Pan et al. 1993). Considering also the lower DHA status in vegetarians, it was of interest to determine the effect of oral supplementation with a DHA-rich/EPA-free source (from micro-algae, non-fish derived) of triglycride oil (DHASCO) in this group. In addition to evaluating the potential for DHASCO to enhance the DHA and EPA (via retroconversion) status in vegetarians, the influence of di etary DHA on conventional (serum total cholesterol, lipoproteins and triglycrides) and thrombogenic car diovascular risk factors (platelet aggregation, thromboxane production, serum viscosity, factor Vile and fibrinogen levels) was determined. Measures of total cholesterol:HDL-cholesterol and LDL-cholesterol:HDLcholesterol ratios were of particular interest because recent reports indicate that they are superior predictors of coronary heart disease risk compared with total cho lesterol and LDL-cholesterol levels (Kinosian et al. 1994 and 1995).

TABLE

Subject characteristics and estimated dietary intake during docosahexaenoic acid or placebo supplementation! VariableHeight, mWeight, kgBMI, kg/m?Protein, 2 energyFat, of % energyCarbohydrate, % of energyAlcohol, % of % of energyControl1.7

0.067.7 3.422.9 1.011.8 0.625.4 1.758.4 3.14.2 1.9DHASCO1.767.122.910.822.963.60. 1.3

1 Values are reported as means SEM,n = 12 (height, weight, BMI) or n = 10 (protein, fat, carbohydrate, and alcohol as a percentage of energy intake). No significant differences between the groups were found for the above variables. 2 BMI, body mass index.

followed by myristic acid (21%), palmitic acid (16%), oleic acid (11%) and lauric acid (8%); no EPA was de tected. Linoleic acid (53%), oleic acid (21%) and pal mitic acid (10%) predominated in the vegetable oil (corn oil) placebo. Each DHASCO capsule contained 0.5 g oil and was estimated to provide 180 mg DHA. The experimental group consumed nine DHA-con taining capsules per day (total 1.62g DHA/d) with their meals, and the control group consumed nine vegetable oil placebo capsules per day. Each group consumed the capsules for a period of 42 d beginning on d 0. After 42 d of capsule ingestion, both groups completed a washout period for 21 d during which there was no supplementa tion. Subjects were weighed on each visit (d 0, 21, 42 and 63) and height was measured at entry; there were no significant differences in these or other characteris tics between the groups at entry Table1). The amount of fat consumed in encapsulated form (4.5 g/d) repre sented <10% of the dietary fat intake. There were no differences between daily energy intakes (8372 kj/d) or intakes of saturated fat, cholesterol, monounsaturated fat or polyunsaturated fat between the two groups (data not shown). The weight of the subjects was not affected throughout the supplementation period in either group. All dietary records were analyzed by the Can West Diet Analysis-Plus program which includes comparison with the Canadian Recommended Nutrient Intakes (West Publishing, St. Paul, MN). Only 20/24 subjects completed the dietary questionnaire. All subjects com pleted the study. Compliance was monitored by de termining the fatty acid composition of serum and platelet phospholipid at 0, 6 and 9 wk (3 weeks postsupplementation) and from a capsule count at the end of the study. Blood collection. At d 0 (presupplementation), 21 and 42 (supplementation), and 63 (washout), blood was collected by antecubital venipuncture from fasting sub jects into siliconized tubes containing no anticoagulant (for serum isolation), or the anticoagulants acid citrate

Downloaded from jn.nutrition.org by on October 10, 2009

SUBJECTS AND METHODS

Subjects and experimental design. The subjects were 24 healthy vegetarians (12 male, 12 female) se lected from the Guelph community who reported hav ing no meat, poultry or fish for a period of at least 6 mo. Approval for this double-blind study was granted by the Human Ethics Committee of the University of Guelph and written informed consent was obtained from each subject. The 24 subjects (aged 29.6 1.7 y, mean SE)were randomly assigned (12/group) to the two supplementation groups, DHA-supplemented and control (6 males, 6 females each). The DHA-enriched encapsulated triglycride oil, DHASCO (algae de rived), and placebo (corn oil) capsules were kindly pro vided by David Kyle of Martek Biosciences, Columbia, MD. As confirmed by quantitative gas-liquid chromatographic analyses, the DHASCO oil contained DHA as the major fatty acid (39 g/100 g fatty acids)

3034

CONQUER

AND HOLUB

dextrose (ACD) [25 g/L Na citrate, 20 g/L dextrose, 14 g/L citric acid (all from Fisher Chemicals, Nepean, Canada)], orNa citrate (3.2%). Whole blood was treated as follows: Tube without anticoagulant. Whole blood was centrifuged at 1250 x g for 15 min to obtain serum. Serum was used for measurement of plasma total-, HDL- and LDL-cholesterol levels, triglycride levels, serum total phospholipid fatty acid content and serum viscosity. Serum was stored at -20C until all samples were collected and thawed just before analyses of lipids. Serum was stored at room temperature until vis cosity was analyzed (see below). Tubes containing Na citrate. Whole blood was cen trifuged at 200 X g for 17 min to obtain platelet-rich plasma (PRP) which was removed, and the remaining blood was centrifuged at 1250 x g for 15 min to obtain platelet-poor plasma (PPP). The PRP was used in aggre gation studies and the PPP was used for analyses of factor VII and fibrinogen levels. The PPP was stored below -20C until analysis by Med Chem Laboratories (Scarborough, Ontario). Tube containing ACD. Washed platelet suspensions were prepared according to the method of Turini et al. 1994. Platelet aggregation. Platelet aggregation was per formed within 2 h of blood collection. Platelets were counted in PRP using a Coulter Counter, model ZM (Coulter Electronics, Burlington, Canada) and adjusted to a final concentration of 2.0-2.5 x IO11 platelets/L using autologous PPP. Aliquots (0.5 mL) of adjusted PRP were preincubated for 1 min in siliconized cuvettes with stirring at 900 rpm at 37C a dual-channel 800B aggrein gometer (Payton Instruments, Ion Trace, Scarborough, Canada) before addition of the aggregating agent. Colla gen (Hormone-Chemie, Mnchen,Germany) was added at two levels with the final concentrations of 2 and 10 mg/L. The platelets were allowed to aggregate for 2 min following the addition of agonist and then the level of aggregation was measured (Born 1962). Cholesterol and triglycride measurement. Total cholesterol was measured enzymatically with a diag nostic test (Sigma Diagnostics Procedure No. 352, St. Louis, MO). HDL-cholesterol was isolated by using a dextran sulfate and Mg ion solution to precipitate the VLDL and LDL from the serum sample. The HDL frac tion was then assayed by an enzymatic assay (Sigma Diagnostics Procedure No. 352-3). Triglycride was measured enzymatically with a diagnostic test (Sigma Diagnostics Procedure No. 339). LDL-cholesterol was calculated using the formula validated by Friedewald et al. (1972). Analyses of all samples for each volunteer were performed in a single assay. Total phospholipid analysis. The fatty acid compo sitions of total phospholipid from serum and washed platelet suspensions were determined following lipid extraction, TLC, transmethylation of fatty acid resi

dues and gas liquid chromatography by procedures sim ilar to those previously described (Donad-oet al. 1994). Fibrinogen, factor Vile and serum viscosity. Fibrin ogen, factor Vile and serum relative viscosity were ana lyzed by Med-Chem Laboratories LTD. Serum relative viscosity was determined on serum (maintained at room temperature) within 16 h of blood withdrawal. The relative viscosity is the flow time of specimen in seconds/flow time of water in seconds in an Oswald Viscosimeter tube (Gradwohl 1970). The concentration of fibrinogen in plasma samples was determined by a diagnostic test comparing thrombin time values of diluted plasma samples (diluted with Ortho Q.F.A. Buffer, Ortho Diagnostic Systems, Johnson and John son, Raritan, NJ), to which Ortho Q.F.A. Thrombin was added, to a reference curve (Clauss 1957). The log of the thrombin time value is inversely proportional to the log of the fibrinogen concentration. To determine factor Vile levels, prothrombin time was determined using Thromborel S (Behring Diagnostics, Westwood, MA) in conjunction with plasma deficient in factor Vile (Quick 1935). Thromboxane production. Aliquots of PRP (0.5 mL) were stimulated with collagen (2 mg/L) for 2 min and reactions stopped by the addition of 250 mol/L (final concentration) indomethacin (Sigma) and 4.5 g/ L NaCl (final concentration) and were immediately chilled on ice. Samples were centrifuged at 2000 x g for 15 min at 0C.The supernatant was removed and stored at -20C until analysis using the Thromboxane B2 Biotrack enzyme immunoassay system (Amersham Canada, Oakville, Ontario, Canada). Statistical analysis. All data are reported as means SEM.Data that were not normally distributed were transformed (log or square) before analysis to reach nor mality. When observations were missing, least-squared means were calculated so that means could be com pared. Split-plot design, including time and treatment as factors, was used in all analyses before examination of specific differences by Fisher's protected LSD. Spe cific differences for preplanned comparisons between treatment means as well as comparison between di etary intakes and subject characteristics were exam ined by f test (Kirk 1968). Statistical analyses were done using the SAS system (SAS Institute, Cary, NC). Changes within groups as well as comparisons between groups were made.

Downloaded from jn.nutrition.org by on October 10, 2009

RESULTS
The fatty acid compositions of the serum total phos pholipid (controls and DHA-supplemented) at entry (wk 0) and after 3 and 6 wk are given in Table 2. The levels of the various (n-3) polyunsaturated fatty acids (PUFA) were not different between the two groups at

DHA AND HEART DISEASE RISK FACTORS TABLE2

3035

Fatty acid composition of total phospholipid in human serum before and after supplementation with docosahexaenoic acid or placebo1
Control Week Fatty acids g/100 g fatty acids 0.4313.1 16:018:018:118:2(n-6)18:3|n-3)20:3(n-6)20:4|n-6)
0.4a,b14.2 0.6b22.8

DHASCO

0.5a,b 0.4a,b13.2 0.6a,b12.8 0.5a,b12.4 0.4b0.3 0.3a,b 0.3b12.6 0.4a,b13.2 0.4a,b12.3 0.6b22.0 0.4a 0.6a,b0.5b0.040.lc0.3c,d0.05a0.03b0.03bO.OSb0.7a,b21.3 0.6a,b23.90.262.89.00.640.270.200.972.19.50.070.230.5b0.03O.lc0.4c,dO.Oga0.03b0.03b0.06b0.2a0.6b 0.8a,b0.27 0.7a0.25 0.7a,b0.030.2a 0.7a,b0.212.27.31.00.030.030.400.03O.lb0.3bO.lbO.OlaO.OlaO.O 0.033.2 0.052.7 0.2<a9.7 O.lc8.9 (AA)220:5|n-3) 0.2C0.630.270.210.912.29.20.070.08a0.03b0.03b0.08b0.2a0.6bO.Ola0.24 0.4d0.57 0.2a0.2b (EPA)22:4[n-6)22:5(n-6)22:5(n-3)22:6(n-3) 0.07a0.29 0.04C0.29 O.Ola 0.04C0.92 O.Ola 0.06b2.4 O.OSa (DHA)|n-6)/(n-3)EPA/AADHA/AA22:5(n-6)/22:6(n-3)26.4 0.2a0.5b 0.2a8.9 0.4b 0.3b3.3 0.6b0.06 0.2a 0.2a0.14 O.Ola O.Ola0.25 O.Olb1.1 0.02C 0.02a0.10 0.03a0.12 O.Ola 0.06b0.003 0.06C O.Olb26.9 O.Olb26.7 O.Olb27.212.713.023.90.242.79.30.540.250.180.902.19.90.060.230.08 O.OOia O.Olc27.6 O.OOia27.812.311.721.70.212.16.51.30.030.030
Downloaded from jn.nutrition.org by on October 10, 2009

1 Values are reported as means SEM,n = 12. Differing superscripts in a row indicate significant differences, P < 0.05. ~ AA, arachidonic acid; EPA, eicosapentaenoic acid; DHA, docosahexaeneoic acid.

entry and did not change in the control group; however, the DHA rose dramatically (from 2.4 to 8.3 g/100 g fatty acids or 246% overall by wk 6) in the DHASCOsupplemented group as did EPA (from 0.57 to 1.3 g/ 100 g fatty acids). The 22:5(n-3) showed a significant decrease upon DHA supplementation (from 0.92 to 0.43 g/100 g fatty acids) as did all of the 20/22 carbon PUFA of the (n-6) series [arachidonic acid (AA), 20:3, 22:4, 22:5] but not 18:2(n-6) (linoleic acid). In the case of platelet phospholipids (Table 3), the marked rise in DHA with supplementation up to 6 wk (from 1.2 to 3.9 g/100 g fatty acids) and in EPA (from 0.21 to 0.58 g/100 g fatty acids) was also accompanied by significant decreases in 22:5(n-3) as well as AA, 22:4(n-6) and 22:5(n-6) but not 18:2(n-6) or 20:3(n-6). The fatty acid alterations in serum/platelet phospholipid reversed to wards entry (wk 0) values at 3 wk following cessation of the DHA supplementation (data not shown). The effects of DHA supplementation on the serum lipids and lipoproteins are given in Table 4. Although no significant alteration was found in the total- and LDL-cholesterol levels with DHA supplementation, a significant decrease in the total cholesterol:HDL-cholesterol ratio (by 16%) as well as the LDL-cholesterol:HDL-cholesterol ratio (by 22%) was exhibited in 6 wk. Furthermore, the triglycrideconcentrations de creased significantly (P < 0.05) within 3 and 6 wk (by 22 and 16%, respectively) in the DHA-supplemented group.

Collagen-induced platelet aggregation as well as col lagen-induced thromboxane A2 (TxA2) [measured as thromboxane B2 (TxB2)] formation, serum viscosity, fibrinogen and factor Vile levels were not significantly altered by DHA supplementation (Table 5).

DISCUSSION
Despite consuming as much or more plant-derived (n-3) (as aLNA) relative to omnivores (for example, -3.6 vs. 2.7 g/d) (Pan et al. 1993, Rperet al. 1992), vegetarians have a significantly lower DHA status based on measured DHA levels in serum, platelets and breast milk (Reddy et al. 1994, Sanders et al. 1978). The limited conversion of LNA to DHA in humans in vivo (Emken et al. 1994) and the absence of DHA in plantbased foods likely accounts for these differences. The approximate daily intake of DHA in the diet of vegans, ovo-lacto vegetarians, and omnivores is estimated to be 0, 33 (based on 4.5 eggs/wk) and 78 mg, respectively (Ferrier et al. 1995, Pan et al. 1993, Rperet al. 1992); the vast majority of the latter intake is in the form of fish (Raper et al. 1992). Our double-blind trial in vegetarian subjects demon strated that a novel algae-derived source of DHA, free of EPA, which provided 1.62 g of DHA/d could more than triple the DHA content of the serum and platelet

3036

CONQUER

AND HOLUB

TABLE 3 Fatty acid composition of total phospholipid in human platelets before and after supplementation Control Week Fatty acid g/100 g fatty acids 0.417.4 16:018:018:118:2|n-6]20:3|n-6)20:4jn-6| 0.4a17.5 0.5a,b6.1 with docosahexaenoic DHASCO acid or placebo!

0.4a,b16.8 0.3b16.3 0.3a16.9 0.6a17.9 0.3a,b6.7 0.5b6.7 0.3a1.3 0.2b1.3 0.3a,b1.5 0.6a,b24.2 0.1a,b25.5 0.lb21.7 (AA)220:5(n-3) 0.6b0.24 0.4b0.25 0.730.58 (EPA)22:4(n-6)22:5(n-6|22:5|n-3)22:6(n-3) 0.0432.1 0.04a2.1 O.OSb1.2 :0.01 O.la O.lb0.19 O.lb0.17 O.lb0.28 0.04b1.71.111.60.010.04O.lbO.la0.5CO.OaO.Oa0.17 0.05b1.71.111.90.010.04O.lbO.la0.5C0.03O.Oa0.14 0.03C1.71.111.20.010.05O.lbO.la0.5CO.OaO.Oa0.25 0.0030.66 O.Ola0.71 0.0533.8 0.0733.9 (DHA|(n-6)/(n-3)EPA/AADHA/AA22:5(n-6)/22:6(n-3)16.8 0.lb6.7 0.2b6.1 0.3a0.02 0.330.03 O.Ob0.16 O.Oc0.18 O.Ob0.00 O.Oc0.00 0.03b,c17.2 0.04b17.117.416.75.71.124.40.290.4a,b0.4a0.4a0.5aO.la0.8b0.0432.2 O.OOa17.8 0.03b18.016.717.56.81.423.80.481.30.3b0.230.5a,b0.2bO.l3,b0.4b0. 0.03C17.116.917.06.01.323.60.212.40.181.71.210.70.010.050.3aO.Sa0.43,b0.2aO.la,b0.7b0 O.OOa
Downloaded from jn.nutrition.org by on October 10, 2009

1 Values are reported as means SEM,n = 12. Differing superscripts in a row indicate significant differences, P < 0.05. 2 AA, arachidonic acid; EPA, eicosapentaenoic acid; DHA, docosahexaenoic acid.

phospholipids within 3 wk and maintain these levels over the 6 wk supplementation period (Tables 2 and 3). Previous studies have shown that fish oils, which contain EPA plus DHA, raise both EPA and DHA levels in the platelet and serum phospholipids of omnivorous subjects (Herold and Kinsella 1986, Weaver and Holub 1988). The (n-3) fatty acid level in serum phospholipid is regarded as a useful biochemical index of the dietary intake and overall nutritional status for EPA/DHA (Bonaa et al. 1992). The level of DHA attained (8.3 g/100 g total fatty acids in serum phospholipid) is above the levels reported in free-living omnivorous subjects (Bo-

naa et al. 1992). The net rise in EPA in serum phospho lipid (0.7 g/100 g fatty acids) was ~one eighth that of the corresponding DHA rise (Table 2). The EPA rise would reflect fatty acid retroconversion from DHA (Voss et al. 1992) with an estimated efficiency of 11.3 and 12.0% based on the mol/100 mol values for serum and platelets, respectively (net rise in EPA/net rise in EPA + DHA as %) as derived from the data in Tables 2 and 3. Interestingly, the decrease in the (n-6) polyunsaturates [20:4(n-6), 22:4(n-6) and 22:5(n-6)] in platelet and serum phospholipids (Tables 2 and 3), as observed in our study as well as with fish oils having EPA plus

TABLE 4 Effect of supplementation with docosahexaenoic Control Week Variable Total cholesterol, mmol/LHDL-cholesterol, mmol/LTotal HDL-cholesterol, cholesterol: moLtnolLDL-cholesterol, mmol/LLDL-cholesterol: HDL-cholesterol, mol:molTriglycride, mmol/L3.83 0.15a1.35 0.09b3.0 0.2b2.10 0.1831.7 0.15a1.400.1lb,c2.9 O.lOc3.0 0.2a,b2.02 0.18a1.6 0.2b2.29 8b1.7 0.1 0.20a1.20 O.lOa3.2 0.3b1.99 0.1831.8 0.26a1.36 0.12b2.8 0.2a,b1.97 0.2131.6 0.24a1.40 0.13b,c2.7 0.2a1.86 0.171.4 0.20.80 O.lla,b acid or placebo on human serum lipid and lipoprotein levels1 DHASCO

0.2b0.75 0.2C0.96 0.2b0.77 0.2b,c0.92 0.2b,c0.81 0.09a3.63 0.1ic3.67 0.10a,b4.160.19b1.45 0.14b,c3.63 0.08a,b,c3.78

Values are reported as means SEM,n = 12. Differing subscripts in a row indicate significant differences, P < 0.05.

DHA AND HEART DISEASE RISK FACTORS TABLE5 Effect of supplementation with docosahexaenoic acid or placebo on selected thrombogenic risk factors in vegetarians!
Control Week Variable Platelet aggregation,2 % ofmaximum aggregationwith collagenTxBj L 10 mg/ x10 production, ng/2.5 plateletsRelative serumViscosity water,where relative to 1.0Fibrinogen, = water g/LFactor %prothrombin Vile, DHASCO

3037

196.3 1.5 2.3

2.4 21.8 0.1

1.9 24.3 0.0

3.228.2 176.1 1.4 1.9

159.2 1.5 1.9

0.0

119.9

0.1

0.1

0.1

103.1

122.0

time83.5

7.389.5189.11.52.3125.3 11.086.9 4.626.90.00.110.587.3186.01.42.299.3 10.486.1180.61.51.8117.8

2.917.90.10.15.588.7 1.724.30.00.18.1
Downloaded from jn.nutrition.org by on October 10, 2009

1 Values are reported as means SEM,n = 12. No significant differences due to supplementation were found. 2 Values for platelet aggregation are for 2 mg/L collagen as a percentage of maximum aggregation with 10 mg/L collagen. 3 TxB2, thromboxane 82.

DHA, was further accompanied by a marked decrease in docosapentaenoic acid [22:5(n-3)j. The latter fatty acid generally exhibits considerable enrichment in se rum and platelet phospholipids when fish oils [provid ing EPA, DHA and some 22:5(n-3)] are fed (Herold and Kinsella 1986, Weaver and Holub 1989). Thus, dietary DHA itself appears to replace both 22:5(n-6) and 22:5(n3) in circulating and cellular phospholipids in human subjects, likely by competition from DHA-containing precursors, via combined de novo phospholipid biosynthetic pathways and deacylation/reacylation reactions. Clinical trials with fish oils containing EPA and DHA (Harris 1989) have generally shown a significant decrease in serum triglycrideconcentrations and often a rise in LDL-cholesterol. These trials (Harris 1989) have employed fish oil concentrates which usually con tain much more EPA than DHA. Because the EPA-free DHASCO used in this study as well as the EPAcontaining concentrate of DHA used previously (von Schacky and Weber 1985) both lowered serum triglyc ride oncentrations, dietary DHA may produce a tric glyceride-lowering effect independent of EPA. The rela tively minor retroconversion of DHA to EPA is likely insufficient to account for the observed alterations in serum lipids and lipoproteins. In our study, the use of EPA-free DHASCO indicated that dietary DHA itself can provide a moderate reduction in the total-cholesterohHDL-cholesterol ratio and the LDL-cholesterohHDL-cholesterol ratio as well as a slight reduction in serum triglycrideconcentrations of vegetarian sub jects (Table 4). These changes are consistent with a favorable influence of dietary DHA on these recognized lipid/lipoprotein risk factors for cardiovascular disease (Drexel et al. 1994). Recent publications (Kinosian et al.

1994 and 1995) indicate that the total cholesterohHDLcholesterol and LDL-cholesterol:HDL-cholesterol ra tios are better predictors of heart disease than total cholesterol or LDL-cholesterol levels alone. Supple mentation with fish oils containing EPA plus DHA for several weeks has generally not significantly reduced LDL-cholesterol:HDL-cholesterol ratios (Harris 1989). It remains to be studied whether the moderate lipid/ lipoprotein shifts as seen in our vegetarian subjects on relatively low fat diets (23% of energy) are also seen in other population and dietary groups. Although many previous studies, often using intakes of EPA plus DHA ranging from 3 to 6 g/ ,have shown a significant reduction in collagen-induced platelet ag gregation and thromboxane formation, this was not ob served (Table 5) in our subjects consuming 1.62 g DHASCO/d for 6 wk. One previous study, using much higher levels of DHA (6 g/d) over 6 d, reported a reduction in collagen-induced aggregation (von Schacky and Weber 1985). Further, the absence of any significant effect of DHA supplementation on other thrombogenic risk factors (viscosity, fibrinogen, factor Vile) in our trial can be compared with various studies with fish oils (EPA/DHA) in which either no change or some beneficial effects have been reported (Harris 1989, Herold and Kinsella 1986, Saynor and Gillott 1992, Weaver and Holub 1988). Future studies which employ varying dietary levels of DHA, including vary ing EPA/DHA ratios, will be of interest. The marked rise in the DHA level of serum and platelet phospholipids of our subjects is of interest in view of recent epidemiological studies which have shown an inverse relationship between DHA levels in the population (both diet and blood) and the risk of

3038

CONQUER

AND HOLUB
Drexel, H., Amann, F. W., Beran, J., Rentsch, K., Cand-nas,R., Muntwyler, J., Luethy, A., Gasser, T. & Follath, F. (1994) Plasma triglycridesand three lipoprotein cholesterol fractions are inde pendent risk factors of the extent of coronary atherosclerosis. Circulation 90: 2230-2235. Dwyer, J. T. (1988) Health aspects of vegetarian diets. Am. J. Clin. Nutr. 48: 712-738. Emken, E. A., Adlof, R. O. & Gulley, R. M. (1994) Dietary linoleic acid influences desaturation and acylation of deuterium-labelled linoleic and linolenic acids in young adult males. Biochim. Biophys. Acta 1213: 277-288. Ferrier, L. K., Caston, L. J., Leeson, S., Squires, J., Weaver, B. J. & Holub, B. J. (1995 ) a-Linolenic acid- and docosahexaenoic acidenriched eggs from hens fed flaxseed: influence on blood lipids and platelet phospholipid fatty acids in humans. Am. J. Clin. Nutr. 62: 81-86. Friedewald, W. T., Levy, R. I. & Frederickson, D. S. (1972) Estima tion of the concentration of low-density lipoprotein cholesterol in plasma without use of the preparative ultracentrifuge. Clin. Chem. 18: 499-502. Gaudette, D. C. & Holub, B. f. (1991) Docosahexaenoic acid (DHA) and human platelet reactivity. J. Nutr. Biochem. 2: 116-121. Gradwohl, R.B.H. (1970) Gradwohl's Clinical Laboratory Meth ods and Diagnosis. A Textbook on Laboratory Procedures and Their Interpretation (Frankel, S., Reitman, S. & Sonnenwirth, A. C., eds.), 7th d. 663-664. Mosby, Saint Louis, MO. pp. Harris, W. S. (1989) Fish oils and plasma lipid and lipoprotein me tabolism in humans: a critical review. J. Lipid Res. 30: 785-807. Herold, P. M. & Kinsella, J. E. (1986) Fish oil consumption and de creased risk of cardiovascular disease: a comparison of findings from animal and human feeding trials. Am. J. Clin. Nutr. 43: 566-598. Hibbeln, J. R. S. Salem, N. (1995) Dietary polyunsaturated fatty acids and depression: when cholesterol does not satisfy. Am. J. Clin. Nutr. 62: 1-9. Kinosian, B., Click H. & Garland, G. (1994) Cholesterol and coro nary heart disease: predicting risks by levels and ratios. Ann. Intern. Med. 121: 641-647. Kinosian, B., Click, H., Preiss, L. &.Puder, K. L. (1995) Cholesterol and coronary heart disease: predicting risks in men by changes in levels and ratios. J. Invest. Med. 43: 443-450. Kirk, R. E. (1968) Experimental Design: Procedures for the Behav ioral Sciences. Brooks Cole Publishing Company, Belmont, CA. Leng, G. C., Horrobin, D. F., Fowkes, F.G.R., Smith, F. B., Lowe, G.D.O., Donnan, P. T. & Ells, K. (1994) Plasma essential fatty acids, cigarette smoking, and dietary antioxidants in peripheral arterial disease. Arterioscler. Thromb. 14: 471-478. Neuringer, M., Anderson, G. J. & Connor, W. E. (1988) The essen tiality of n-3 fatty acids for the development and function of the retina and brain. Ann. Rev. Nutr. 8: 517-541. Pan, W.-H., Chin, C.-J., Shen, C.-T. &. Lee, M.-H. (1993) Hemostatic factors and blood lipids in young Buddhist vegetarians and omnivores. Am. J. Clin. Nutr. 58: 354-359. Pepe, S. & McLennan, P. L. (1996) Dietary fish oil confers direct antiarrhythmic properties on the myocardium of rats. J. Nutr. 126: 34-42. Quick, A. J. (1935) A study of the coagulation defect in hemo philia and in jaundice. Am. f. Med. Sci. 190: 501. Rper,N. R., Cronin, F. J. & Exler, J. (1992) Omega-3 fatty acid Content of the U.S. food supply. J. Am. Coll. Nutr. 11: 304-308. Reddy, S., Sanders, T.A.B. & Obeid, O. (1994) The influence of maternal vegetarian diet on essential fatty acid status of the new born. Eur. J. Clin. Nutr. 48: 358-368. Sanders, T.A.B., Ellis, F. R. & Dickerson, J.W.T. (1978) Studies of vegans: the fatty acid composition of plasma choline phosphoglycerides, erythrocytes, adipose tissue and breast milk and some indicators of susceptibility to ischemie heart disease in vegans and controls. Am. J. Clin. Nutr. 31: 805-813.

CVD (Leng et al. 1994, Simon et al. 1995). Thus, part of the cardioprotective effect of fish/fish oils containing (n-3) PUFA appears due to DHA in addition to EPA. The cardioprotective effects of (n-3) fatty acids are con sidered to be mediated by a number of physiological and biochemical mechanisms; included among these are studies which indicate that the enrichment of heart tissue in DHA provides an antiarrhythmic effect (Pepe and McLennan 1996), which may account for the re duction in cardiac arrest and sudden cardiac death in those having a higher DHA status. In addition to the above, dietary DHA intakes and increased status in the body have been implicated in favorable effects on attention-deficit disorders (Stevens et al. 1995), depres sion and anxiety disorders (Hibbeln and Salem 1995), as well as protection against breast cancer in postmenopausal women (Zhu et al. 1995). In conclusion, the consumption of 1.62 g of an ani mal-free source of DHA per day by vegetarians readily enhances their DHA status, provides for EPA formation based on serum and platelet phospholipid analysis, and exerts moderately favorable (lowering) effects on the total cholesterol:HDL-cholesterol ratio, the LDL-cholesterol:HDL-cholesterol ratio, as well as serum triglyc ride concentrations. The present studies, conducted with young non-hyperlipidemic vegetarian adults of both genders, should now be extended to other groups (e.g., omnivorous normolipidemics and hyperlipidemics).

Downloaded from jn.nutrition.org by on October 10, 2009

ACKNOWLEDGMENTS
We would like to thank Margaret Berry, Ind Mani and Melinda Gooderham for their help in all aspects of this investigation. We would also like to thank David Kyle and Eeva-Kaarina Koskelo of Martek Biosciences Corporation (Columbia, MD) for supplying both the control and DHA capsules for this investigation. Ap preciation is also extended to Martek for covering the costs of the fatty acid analyses as provided by Lipid Analytical Laboratories, University of Guelph Re search Park (Guelph, Ontario, Canada).

LITERATURE CITED
Bonaa, K. H., Bjerve, K. S. & Nordoy, A. (1992) Habitual fish con sumption, plasma phospholipid fatty acids, and serum lipids: the Tromso study. Am. J. Clin. Nutr. 55: 1126-1134. Born, G.V.R. (1962) Aggregation of blood platelets by adenosine diphosphate and its reversal. Nature (Lond.) 194: 927-929. Clauss, A. (1957) Gerinnungsphysiologische Schnellmethode zur bestimmung des fibrinogens. Acta Haematol. 17: 237. Donadio, J. V., Jr., Bergstrath, E. J., Offord, K. P., Spencer, D. C. & Holley, K. E. (1994) A controlled trial of fish oil in IgA nephropathy. N. Engl. J. Med. 331: 1194-1199.

O,: DHA AND HEART DISEASE RISK FACTORS Sanders, T.A.B. & Hinds, A. (1992) The influence of a fish oil high in docosahexaenoic acid on plasma lipoprotein and vitamin E concentrations and haemostatic function in healthy male volun teers. Br. J. Nutr. 68: 163-173. Saynor, R. & Gillott, T. (1992) Changes in blood lipids and fibrinogen with a note on safety in a long term study on the effects of n-3 fatty acids in subjects receiving fish oil supplements and followed for seven years. Lipids 27: 533-538. Simon, J. A., Hodgkins, M. L., Browner, W. S., Neuhaus, J. M., Bernert, J. T. & Hulley, S. B. (1995| Serum fatty acids and the risk of coronary heart disease. Am. J. Epidemiol. 142: 469-476. Simopoulos, A. P. &.Salem, N., Jr. (1992) Egg yolks as a source of long chain polyunsaturated fatty acids in infant feeding. Am. J. Clin. Nutr. 55: 411-414. Stevens, L. J., Zentall, S. S., Deck, J. L., Abate, M. L., Watkins, B. A., Lipp, S. R. & Burgess, J. R. (1995) Essential fatty acid metabo lism in boys with attention-deficit hyperactivity disorder. Am. J. Clin. Nutr. 62: 761-768.

3039

Turini, M. E., Powell, W. S., Behr, S. R. &. Holub, B. f. (1994) Ef fects of a fish oil and vegetable oil formula on aggregation and ethanolamine-containing lysophospholipid generation in acti vated human platelets and on leukotriene production in stimu lated neutrophils. Am. J. Clin. Nutr. 60: 717-724. von Schacky, C. & Weber, P. C. (1985) Metabolism and effects on platelet function of the purified eicosapentaenoic and docosahex aenoic acids in humans. J. Clin. Invest. 76: 2446-2450. Voss, A., Reinhart, M. & Sprecher, H. (1992) Differences in the inter conversion between 20- and 22- carbon (n-3) and (n-6) polyunsatu rated fatty acids in rat liver. Biochim. Biophys. Acta 1127: 33-40. Weaver, B. J. & Holub, B. J. (1988) Health effects and metabo lism of dietary eicosapentaenoic acid. Prog. Food Nutr. Sci. 12: 111-150. Zhu, Z. R., Agren, J., Mannisto, S., Pietinen, P., Eskelinen, M., Syrjanen, K. &.Uusitupa, M. (1995) Fatty acid composition of breast adipose tissue in breast cancer patients and in patients with be nign breast disease. Nutr. Cancer 24: 151-160.

Downloaded from jn.nutrition.org by on October 10, 2009