Effects of Fish Oil Supplements in NIDDM Subjects Controlled Study
The aim of this study was to evaluate the effects of a fish oil preparation (MaxEPA) on hemostatic function and fasting lipid and glucose levels in non-insulin-dependent diabetic (NIDDM) subjects. Eighty NIDDM outpatients aged 55.9 yr (mean SD 11.5 yr) participated in a prospective double-blind placebo-controlled study of MaxEPA capsules (10 g/day) or olive oil (control) treatment over 6 wk. Patients received either MaxEPA or olive oil in addition to preexisting therapy. Metabolic and hemostatic variables were measured before treatment and after 3 and 6 wk. Platelet membrane eicosapentaenoic acid (EPA) content increased in the treatment group (P < 0.001). MaxEPA supplementation was associated with a significant fall in total triglycerides (P < 0.001) but did not affect total cholesterol (P = 0.7) compared with control treatment. Fasting plasma glucose increased after 3 wk (P = 0.01) but not after 6 wk (P = 0.17) treatment with MaxEPA. Spontaneous platelet aggregation in whole blood fell in the MaxEPA group (P < 0.02) after 6 wk, but there were no changes in agonist-induced platelet aggregation, thromboxane generation in platelet-rich plasma, or plasma P-thromboglobulin and platelet factor IV levels. An increase in clotting factor VII (P = 0.02), without changes in fibrinogen or factor X levels, occurred in the MaxEPA group. Similar reductions in blood pressure were observed in both groups. Dietary supplementation with MaxEPA capsules (10 g/day) in NIDDM subjects is associated with improvement in hypertriglyceridemia but with deleterious effects in factor VII and blood glucose
From the Academic Unit of Diabetes and Endocrinology and the Department of Primary Health Care, University College and Middlesex School of Medicine, Whittington Hospital; and the Department of Chemical Pathology and Human Metabolism, Royal Free Hospital, Hampstead, London, United Kingdom. Address correspondence and reprint requests to Timothy J. Hendra, MRCP, Derriford Hospital, Plymouth PL6 8DH, UK. Received for publication 9 October 1989 and accepted in revised form 21 March 1990.
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
Timothy). Hendra, MRCP Mary E. Britton, MRCP David R. Roper, FIMLS Daniel Wagaine-Twabwe, MB lames Y. Jeremy, PhD Paresh Dandona, FRCP Andrew P. Haines, MRCP John S. Yudkin, FRCP
levels. Most indices of platelet function are unaffected by this therapy. Diabetes Care 13:821-29, 1990
N
on-insulin-dependent diabetic (NIDDM) subjects have an increased prevalence of coronary artery disease that is not explicable on the basis of-known risk factors for arteriosclerosis (1,2). There is, however, a marked clustering of such risk factors in these subjects, and abnormalities of blood pressure, lipid metabolism, and hemostatis may contribute to the etiology of the increased cardiovascular risk (3). Abnormalities in lipid levels in diabetic subjects are related to the type of diabetes, sex, and treatment (46). Most studies in NIDDM patients show elevated triglyceride and reduced high-density lipoprotein cholesterol (HDL-chol) levels, but low-density lipoprotein cholesterol (LDL-chol) levels are generally reported to be similar to those in nondiabetic subjects. Abnormalities in lipid levels may be more exaggerated in female diabetic patients, which might help to explain their excess cardiovascular risk compared with male diabetic subjects (4,7). Increases in platelet aggregation have frequently been reported in diabetic patients, and elevated levels of fibrinogen may contribute further to an increased risk of thrombosis (8,9). Some studies have suggested that these hemostatic abnormalities are related to microvascular disease (10), but there are also reports of an association between abnormalities of platelet aggregation (11) and fibrinogen (12) and large-vessel disease. It is, however, not possible to deduce from cross-sectional studies whether such differences are the cause or the consequence of large-vessel disease.
821
FISH OIL SUPPLEMENTATION IN DIABETIC SUBJECTS
An epidemiological association has been described between diets rich in fish and a low prevalence of cardiovascular disease (13). This association may be mediated through the beneficial effects of marine lipids, which are rich in polyunsaturated w-3 fatty acids (o)3FAs), on recognized cardiovascular risk factors. Dietary supplementation with fish oil in nondiabetic subjects lowers triglyceride levels by reducing both the triglyceride and apolipoprotein B (apoB) content of verylow-density lipoproteins (VLDLs; 14,15). A high intake offish oil lowers LDL-chol and LDL-apoB levels by decreasing LDL synthesis (16). Dietary supplementation with 160 mmHg systolic or 95 mmHg diastolic or treatment with antihypertensive medication. Drug histories were obtained for all subjects, with particular reference given to the ingestion of known platelet-active preparations during the 2 wk before study. Each subject was randomized in a double-blind manner to receive twice a day either five MaxEPA capsules (each containing 1 g fish oil; treatment group) or five olive oil capsules (control group). This dose of MaxEPA
822
provided 1.8 g eicosapentaenoic acid (EPA) and 1.2 3 docosahexaenoic acid (DHA) daily, together with smaller amounts of vitamins A, D, and E. Each MaxEPA capsule also contained 6 mg cholesterol. Subjects were instructed not to alter their dietary intake or exercise patterns, although no formal assessment of these variables was performed during the study. Participants were seen at the start of the study and c.t 3 and 6 wk. All subjects attended after a 9-h overnight fast and rested for 15 min before measurement of sitting blood pressure and venesection (with minimal venestasis) through a 19-gauge needle. The height and weight of each subject at each visit were also recorded. The following measurements were performed on blood taken at each visit: 1. Plasma glucose (glucose oxidase method, Beckman, Brea, CA) and full blood count (Technicon particle counter, Tarrytown, NY). 2. Glycosylated hemoglobin (agar gel electrophoresis, Corning, Halstead, Essex, UK). 3. Total serum triglyceride and cholesterol levels were measured enzymatically (29,30). HDL-chol was measured after heparin manganese precip tation and LDL-chol was calculated with the Friede:wald equation (31). Separate analyses of data on LDL-chol were performed in subjects whose level of triglyceride was >4.5 mM (31). 4. Plasma |3-thromboglobulin and platelet factor IV (radioimmunoassay kits from Amersham, Aylej;bury, Bucks, UK, and Abbott, Walsingham, Berks, UK, respectively). To exclude the effects of in vitro activation, data were analyzed only in subjects whose p-thromboglobulin-platelet factor IV ratio was >3.0 (32). All samples from each subject were analyzed in the same batch (within-coefficients of variation [C.V.s] assay 9.0% for p-thromboglobulin and 12.3% for platelet factor IV). 5. Plasma fibrinogen (Von Clauss method, BCL, Lewes, Sussex, UK), factor VII (1-stage assay using beagle plasma and rabbit thromboplastin from National UK Reference Laboratory, Withington Hospital, Manchester, UK), and factor X (Russell's viper venom assay, Diagnostic Reagents, Thame, Oxon, UK). Blood samples for clotting-factor estimation were rapidly frozen to -40°C after venesection so that all samples from each subject could be measured in the same assay (betweenassay C.V.s for fibrinogen 1.4%, factor VII 3.4%, and factor X 3.5%). Although cold activation of factor VII could be responsible for up to a 4-5% increase in activity, because all samples from both groups were not collected on ice before separation, this would not have affected our results. 6. Platelets were isolated from blood collected into EDTA, and platelet membrane phospholipid composition was determined (at 0 and 6 wk) with gas liquid chromatography to assess compliance with the MaxEPA therapy.
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
T.J. HENDRA AND ASSOCIATES
7. Spontaneous platelet aggregation in whole blood over 60 min with the use of the Ultra-Flo wholeblood platelet-counting method (33). C.V.s for the absolute fall in platelet count between 10 and 60 min were between 11.4 and 8.5%,. respectively. 8. In vitro whole blood (Ultra-Flo platelet-counting method; 33) and platelet-rich plasma (34) platelet aggregation studies were performed with four concentrations of each of three agonists, i.e., ADP, epinephrine, and collagen. Platelet sensitivity was expressed in terms of the EC50 (agonist concentration producing 50% of maximal aggregation) for each agonist by method as described previously (33). The C.V.s for agonist EC50 values in the same person on 1 day were 8.5 and 8.3% for ADP in whole blood and platelet-rich plasma, respectively, 6.9 and 7.5% for collagen, and 9.5 and 9.3% for epinephrine. 9. Platelet TXB2 release in platelet-rich plasma 5 min after the addition of ADP (10~5 M), epinephrine (10~5 M), and collagen (1 mg/ml; 35). Statistical analysis. Unpaired Student's t tests or MannWhitney U tests were used to compare between-group differences for each variable at baseline and also to compare changes in each variable in the treatment group between baseline and 3 wk and baseline and 6 wk with changes in control group. Analysis of variance was used to study the independent relationship of treatment with changes in variables while adjusting for differences in baseline characteristics between groups. Values are expressed as mean ± SD for normally distributed data and median (range) for skewed data. Changes are expressed as means (95% confidence interval [Cl]) for normally distributed data and median differences of change (P) for skewed data. The sample size was calculated as sufficient to detect a 24% change in LDL-chol between the groups at the 5% level with a power of 90%, which is the degree of change in LDLchol detected with MaxEPA therapy in a group of insulin-dependent diabetic (IDDM) patients (26). Results at 3 wk are only discussed if they suggest an interpretation at odds with the 6-wk findings.
RESULTS MaxEPA and placebo groups were matched for age, sex, ethnic origin, smoking habits, duration of diabetes, and mode of treatment (Table 1). Male subjects in the MaxEPA group were heavier than those in the control group. There was a similar prevalence of hypertension, peripheral vascular disease, neuropathy, and nephropathy in the two groups. However, coronary artery disease was more common in the MaxEPA group, whereas retinopathy was more frequent in the control group. All variables for which response to MaxEPA were assessed did not differ significantly between the groups at baseline.
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
TABLE 1 Clinical characteristics of subjects Characteristic
MaxEPA
Control
n Age (yr) Weight (kg) Men Women Duration of diabetes (yr) M/F Ethnic origin White Asian Afro-Carribean Smokers Complications Hypertension Coronary artery disease Peripheral vascular disease Diabetic nephropathy Diabetic retinopathy Diabetic neuropathy Treatment Diet alone Sulfonylureas Biguinides Biguinides and sulfonylureas
40 56.0(53.5, 58.5)
40 55.8(52.3, 59.3)
88.8(82.1, 95.5) 76.0 (62.2, 89.4) 6.7 (4.8, 8.5) 30/10
79.9 (74.4, 84.4)* 76.8 (66.6, 77.0) 8.0(5.7, 10.3) 25/15
39 0 1 8
38 1 1 8
4 14
5 3t
9 3 5 9
7 4 14* 10
8 12 8
8 14 3
12
15
Values of continuous variables are means, with 95% confidence intervals in parentheses. *P < 0.05, tP < 0.02.
One patient in the control group died during the trial from an exacerbation of bronchial asthma. One patient in the MaxEPA group withdrew, complaining of flatulence and abdominal pain. Four others (2 MaxEPA, 2 control) withdrew due to difficulty swallowing the capsules. Nine patients (4 MaxEPA, 5 control) were taking platelet-active drugs and did not have platelet function studies performed. In analyzing changes in triglycerides, the results of one patient with gross hypertriglyceridemia were excluded to allow parametric testing of otherwise skewed data. The mean level of triglyceride fell by 30.0% in the treatment group compared to a rise of 10.0% in the control group. There were no significant changes in serum total cholesterol or HDL-chol, but calculated LDL-chol rose within the treatment group. When compared with baseline this rise was significant (mean 0.26 mM, 95% Cl 0.03-0.49, P = 0.03), but when the change in the treatment group was compared with the change in the control group, the difference attributable to treatment failed to reach significance (mean 0.26 mM, 95% Cl - 0 . 0 3 - 0 . 5 6 , P = 0.085; Table 2). These results for LDL-chol were unaffected by excluding subjects with triglyceride levels >4.5 mM. Fasting blood glucose rose within the MaxEPA group at both 3 and 6 wk. However, the mean change in
823
2.0(1.7,2.4) 2.2(1.7,2.6)
-0.90 (-1.30, -0.50) T -» \*
r -»T
-»
Tt -» -*
t -»
v
Fasting plasma glucose HbA,
i I i
I i i
TC
LDL-chol
ft
HDL-chol
Platelets
I ADP-stimulated aggregation i TXB2 4 Spontaneous platelet aggregation
Effect offish oil supplements
t VII -»X,I
Clotting factors
64 .1 (60 .7, 67.6) 63 .5 (59..6, 67.5)
Control
4.59 (0.90, 8. 29) 0.016
62 .9 (59..6, 66.3) 66 .5 (63 .4, 69.6)
MaxEPA
60 min
TC, triglyceride; TXB2, thromboxane B2 generation; u>3FA, w-3 fatty acid; LDL-chol, low-density lipoprotein cholesterol; HDL-chol, high-density lipoprotein cholesterol. Where a control was used effects are in comparison with control, otherwise they are compared to baseline. 'Transient. tMarginal significance (P = 0.07).
n
Refs.
Investigations
Duration (wk)
Dose fish oil (dose o)3FA g/day)
TABLE 6 Fish oil supplementation in non-insulin-dependent diabetes mellitus
Values are means, with 95% confidence intervals in parentheses. Comparisons of changes between groups are compared with unpaired Student's t tests.
0 wk 6 wk Change attributable to treatment P
MaxEPA
10 min
Changes in spontaneous platelet aggregation with MaxEPA and olive oil: percentage of platelets remaining after aggregation
TABLE
FISH OIL SUPPLEMENTATION IN DIABETIC SUBJECTS
fasting blood glucose attributable to MaxEPA treatment was significant only at 3 wk (1.44 mM, 95% Cl 0.34- DISCUSSION 2.5, P = 0.01) and not at 6 wk (0.97 mM, 95% Cl - 0 . 3 9 - 2 . 3 , P = 0.17; Table 3). There were no signifn this study, we demonstrated that dietary suppleicant changes in HbA, in either treatment group. Weight mentation with MaxEPA fish oil capsules for 6 w< was unchanged in both groups, and changes in meain a dose of 10 g/day produces a fall in triglyceride sures of glycemia were unrelated to changes in weight levels and a reduction in spontaneous platelet agin either group. gregation in whole blood. These changes were, howLevels of factor VII rose in the treatment group, with ever, associated with a transient deterioration in blood the change being significantly different from that in the glucose control, an increase in clotting factor VII, and control group at 6 wk (mean 19.3%, 95% Cl 3.9-34.4, a marginal increase in calculated LDL-chol levels. There P = 0.02; Table 3). There was no significant change in were similar significant falls in blood pressure and in levels of fibrinogen or factor X. Systolic and diastolic TXB2 generation in response to collagen in both the blood pressure fell similarly in both groups, with no MaxEPA and olive oil (control groups). No effects were significant difference in these changes between groups observed on standard measures of in vivo platelet function, factor X or fibrinogen levels, or agonist-induced (Table 4). A significant fall in collagen-stimulated TXB2 gener- platelet aggregation with whole blood and platelet-rich ation occurred within the treatment group from baseline plasma techniques. The observed increase in membrane to 6 wk (mean 125 \xg/L [25th-75th percentile 90-230] EPA levels was a measure of subject compliance with vs. 98.5 |Ag/L [59-122], P = 0.003). However, when the MaxEPA capsules and is similar to increases obthis change was compared with the change in the con- served in previous studies in healthy and diabetic subtrol group, the difference failed to reach significance jects (13,26). Studies in nondiabetic subjects have demonstrated (Table 4). ADP and epinephrine-stimulated TXB2 genthat changes in serum lipid levels with fish oil suppleeration was not significantly altered in either group. No significant changes in plasma p-thromboglobulin ments are dose related and that a hypotriglyceridemic or platelet factor IV levels occurred in either group. Ag- effect is seen with low dosages of o>3FAs (3-10 g/day) onist-induced platelet aggregation in either whole blood (14,15,36), whereas dosages of 5-20 g/day are necesor platelet-rich plasma was unaffected by either MaxEPA sary for a hypocholesterolemic effect (16). Some studies or olive oil. Platelet-rich plasma from 24 patients (14 in have suggested that lower intakes of o>3FAs may result MaxEPA group, 10 in control group) did not show sec- in increases in LDL-chol and apoB concentrations, parondary aggregation to epinephrine at any time point, ticularly in subjects with type V hyperlipoproteinemia and in these subjects, no effect of either treatment was (37,38). Because the major catabolic pathway for VLDI. observed on primary aggregation. Spontaneous platelet is via conversion to LDL, this may produce an increase aggregation in whole blood at all time points was sig- in LDL-chol parallel to the fall in triglycerides (39). Alnificantly reduced after 6 wk of MaxEPA therapy (Ta- ternatively, fish oil supplements may lead to hepatic secretion of smaller-than-normal VLDL particles, which ble 5). MaxEPA treatment was associated with a significant are known to be preferentially converted to LDL, comincrease in platelet membrane EPA content (expressed pared to larger triglyceride-rich VLDL (40). In our study, as a percentage of total membrane fatty acids; Table 2). as in that by Haines et al. (26) demonstrating a 24% rise Over 70% of subjects in the MaxEPA group who com- in LDL-chol in IDDM subjects, the estimation of LDLpleted the study had more than a threefold increase in chol concentration was calculated with the Friedewald EPA content at 6 wk compared with baseline. Analysis equation, which assumes a constant ratio of triglyceride of the patients in the two groups combined revealed to cholesterol in VLDL. Because preferential lowering marked negative correlation between changes in EPA of triglycerides in VLDL has been shown to occur with and arachidonic acid (rs = - 0 . 4 2 , n = 50, P < 0.001). MaxEPA, this may artificially elevate calculated LDLThere were also significant negative correlations of chol levels (41). The level of total cholesterol was unchanges in EPA with triglyceride levels (rs = - 0 . 5 0 , changed with MaxEPA treatment in this study, although n = 62, P < 0.001) and collagen-induced TXB2 pro- it was increased by MaxEPA therapy in the study of duction (rs = - 0 . 2 6 , n = 50, P = 0.04). Changes in IDDM subjects. Currently, there have been 10 studies published asspontaneous aggregation at 60 min were correlated negatively with changes in arachidonic acid {rs = - 0 . 4 1 , sessing fish oil supplementation in diabetic subjects n = 42, P = 0.004) and positively with changes in EPA (24-28,42-46), of which 7 are in NIDDM patients (27,28,42-46). The total number of patients treated in (rs = 0.25, n = 52, P = 0.04). The effects of mismatches in body weight and prev- these 7 studies was 73, and only 2 studies used a control alence of retinopathy and ischemic heart disease be- group. These studies are summarized in Table 6. In tween the groups at baseline were assessed with analysis some of these studies, there was evidence of deterior of variance. The significance of changes in all the mea- ration in glycemic control and increases in LDL-chol o sured variables was unaffected when baseline differ- apoB levels, although the dosage of MaxEPA used and duration of treatment differed between investigators ences were incorporated into the analysis.
I
826
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
T.J. HENDRA AND ASSOCIATES
(27,28,45). In 1 study, there is evidence that increases in fasting plasma glucose may be associated with increased basal hepatic glucose output and reduced mealand glucagon-stimulated insulin release (28). However, another study demonstrated improved in vivo insulin sensitivity with MaxEPA supplements (44). We found no evidence supporting previous studies in nondiabetic subjects that have suggested that supplementation with o)3FAs may reduce other cardiovascular risk factors, including fibrinogen levels (21), blood pressure (18,23), and agonist-induced platelet aggregability (18,19,22), in addition to their effects on lipids. However, recent studies in larger numbers of subjects have questioned the effects offish oils on platelet aggregation. The effect of MaxEPA therapy on platelet function has not been investigated previously in NIDDM subjects, but the results of this study are similar to those observed by others in IDDM subjects (24-26), showing reductions in platelet TXB2 generation after MaxEPA therapy (26). However, we were unable to demonstrate reduced platelet aggregation in vivo or in response to agonists in vitro, probably because a greater degree of inhibition of TXB2 synthesis is required to affect in vitro platelet aggregation, a finding that may be common to subjects with atherosclerosis (28). The observed correlations between changes in membrane fatty acids and those in spontaneous platelet aggregation are of uncertain significance, because previous work has indicated that spontaneous platelet aggregation in whole blood may be mediated via non-TXB2-dependent mechanisms (47,48). Factor VII levels were elevated by MaxEPA in this study, which may be deleterious on cardiovascular risk. This finding was unexpected in view of previous studies that have shown that interventions lowering triglyceride levels may be associated with a reduction in factor VII levels (49,50). In this study, olive oil (containing 8-10% polyunsaturated fatty acids) was used as a control in the same dosage as the fish oil. In the olive oil treatment group, there were no significant changes in any of the lipid variables of hemostatic function, with the exception of a fall in collagen-induced TXB2 production. This effect was not associated with any changes in spontaneous platelet aggregation. There is evidence from studies in nondiabetic subjects and animals that olive oil can affect both serum lipids and platelet function, although the dosages used in human studies were much larger than those used herein (51-54). In summary, we have shown that theoretically advantageous changes in serum triglycerides and perhaps collagen-stimulated TXB2 production in platelet-rich plasma occur in NIDDM patients taking 10 g/day of fish oil. The significance of triglycerides as an independent cardiovascular risk factor is debated (55,56), and that of collagen-stimulated TXB2 production is unknown. Potentially deleterious effects of fish oil on other possible risk factors, i.e., blood glucose control, LDL-chol, and factor VII, led us to conclude that fish oil in this dosage
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
cannot be recommended to NIDDM subjects as an aid to reducing cardiovascular risk.
ACKNOWLEDGMENTS
T.J.H., M.E.B, and D.W.-T. received financial support from Duncan Flockhart & Company, Greenford, Middlesex, United Kingdom. We thank J. Wager for technical assistance, Caroline Jackson for computing expertise, and Sharrone McCullough for secretarial assistance.
REFERENCES 1. Jarrett RJ, Keen H, Chakrabarti R: Diabetes, hyperglycemia and arterial disease. In Complications of Diabetes. 2nd ed. Keen H, Jarrett RJ, Eds. London, Arnold, 1982, p. 179-204 2. jarrett RJ: Type 2 (non-insulin-dependent) diabetes mellitus and coronary heart disease: chicken, egg or neither? Diabetologia 26:99-102, 1984 3. Wingard DL, Barrett-Connor E, Criqui MH, Suarez L: Clustering of heart disease risk factors in diabetic compared to nondiabetic adults. Am) Epidemiol 117:19-26, 1983 4. Walden CE, Knopp RH, Wahl PW, Beach KW, Strandness E: Sex differences in the effect of diabetes mellitus on lipoprotein triglyceride and cholesterol concentrations. N Engl) Med 311:953-59, 1984 5. Seviour PW, Teal TK, Richmond W, Elkeles RS: Serum lipids, lipoproteins and macrovascular disease in non-insulin-dependent diabetics: a possible new approach to prevention. Diabetic Med 5:166-71, 1988 6. Howard BV: Lipoprotein metabolism in diabetes mellitus. ) Lipid Res 28:613-28, 1987 7. Rytter L, Troelsen S, Beck-Nielsen H: Prevalence and mortality of acute myocardial infarction in patients with diabetes. Diabetes Care 8:230-34, 1985 8. Hendra TJ, Betteridge DJ: Platelet function, platelet prostenoids, and vascular prostacyclin in diabetes mellitus. Prostaglandins Leukotrienes Essent Fatty Acids 35:197—
212, 1989 9. Fuller JH, Keen H, Jarrett RJ, Omer T, Meade TW, North WRS, Stirling Y: Haemostatic variables associated with diabetes and its complications. Br Med J 2:964-66, 1979 10. Jennings PE, Barnett AH: New approaches to the pathogenesis and treatment of diabetic microangiopathy. Diabetic Med 5:111 -17, 1988 11. Packham MA, Mustard JF: The role of platelets in the development and complications of atherosclerosis. Semin Hemato/23:8-26, 1986 12. Fuller JH: The haemocoagulation system and macroangiopathy in insulin dependent (type 1) diabetes. Ann Clin Res 16:137-41, 1984 13. Leaf A, Weber PC: Cardiovascular effects of n-3 fatty acids. N Engl I Med 318:549-57, 1988 14. Sanders TA, Sullivan DR, Reeve J, Thompson GR: Triglyceride-lowering effect of marine polyunsaturates in patients with hypertriglyceridemia. Arteriosclerosis 5:45965, 1985 15. Nestel PJ, Connor WE, Reardon MF, Connor S, Wong S,
827
FISH OIL SUPPLEMENTATION IN DIABETIC SUBJECTS
16.
17.
18.
19.
20.
21. 22.
23.
24.
25. 26.
27. 28.
29. 30. 31.
828
Boston R: Suppression by diets rich in fish oil of very low density lipoprotein production in man. J Clin Invest 74:82-89, 1984 Illingworth DR, Harris WS, Connor WE: Inhibition of low density lipoprotein synthesis by dietary omega-3 fatty acids in humans. Arteriosclerosis 4:270-75, 1984 Hay CRM, Durber AP, Saynor R: Effect of fish oil on platelet kinetics in patients with ischaemic heart disease. Lancet 1:1269-72, 1982 Lorenz R, Spengler U, Fischer S, Duhm J, Weber PC: Platelet function, thromboxane formation and blood pressure control during supplementation of the western diet with cod liver oil. Circulation 67:504-11, 1983 Ahmed AA, Holub BJ: Alteration and recovery of bleeding times, platelet aggregation and fatty acid composition of individual phospholipids in platelets of human subjects receiving a supplement of cod liver oil. Lipids 19:61724, 1984 Knapp HR, Reilly IAG, Alessandrini P, FitzGerald GA: In vivo indices of platelet and vascular function during fishoil administration in patients with atherosclerosis. N Engl I Med 314:937-42, 1986 Hostmark AT, Bjerkedal T, Kierulf P, Flaten H, Ulshagen K: Fish oil and plasma fibrinogen. Br Med) 297:180-81, 1988 Sanders TAB, Vickers M, Haines AP: Effect on blood lipids and haemostasis of a supplement of cod-liver oil, rich in eicosapentaenoic acid and docosahexaenoic acids in healthy young men. Clin Sci 61:317-24, 1981 Rogers S, James KS, Butland BK, Etherington MD, O'Brien JR, Jones JG: Effects of a fish oil supplement on serum lipids, blood pressure, bleeding time, hemostatic and rheological variables. Atherosclerosis 63:137-43, 1987 Beitz J, Schimke E, Liebaug U, Block H-U, Beitz A, Honigmann G, Sziegoleit W, Muller G, Mest H-J: Influence of a cod liver oil diet in healthy and insulin-dependent diabetic volunteers on fatty acid pattern, inhibition of prostacyclin formation by low density lipoprotein (LDL) and platelet thromboxane. Klin Wochenschr 64:793-99, 1986 Tilvis RS, Rasi V, Viinikka L, Ylikorkala O, Miettinen TA: Effects of purified fish oil on platelet lipids and function in diabetic women. Clin Chim Ada 164:315-22, 1987 Haines AP, Sanders TAB, Imeson JD, Mahler RF, Martin J, Mistry M, Vickers M, Wallace PG: Effects of a fish oil supplement on platelet function, haemostatic variables and albuminuria in insulin-dependent diabetics. Thromb Res 43:643-55, 1986 Schectman G, Kaul S, Kissebah AH: Effect offish oil concentrate on lipoprotein composition in NIDDM. Diabetes 37:1567-73, 1988 Glauber H, Wallace P, Griver K, Brechtel G: Adverse metabolic effect of omega-3 fatty acids in non-insulindependent diabetes mellitus. Ann Intern Med 108:66368, 1988 Roescholau P, Bernt E, Gruber W: Enzymatic determination of total cholesterol in serum. Z Klin Chem Klin Biochem 12:226, 1974 Bucolo G, David H: Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem 19:47682, 1973 Friedewald WT, Levy Rl, Frederickson DS: Estimation of the concentration of low density lipoprotein cholesterol in plasma without use of the preparative centrifuge. Clin Chem 18:449-502, 1972
32. 33.
34. 35.
36. 37.
38.
39.
40.
41. 42.
43.
44.
45.
46.
47.
48.
Kaplan KL, Owen J: Plasma levels of beta-thromboglobulin and platelet factor 4 as indices of platelet activation in vivo. Blood 57:199-202, 1981 Hendra TJ, Oughton J, Smith CCT, Betteridge DJ, Yudkin JS: Exercise-induced changes in platelet aggregation: a comparison of whole blood and platelet rich plasma techniques. Thromb Res 52:443-51, 1988 Born GVR: Aggregation of platelets by adenosine diphos.phate and its reversal. Nature (Lond) 194:927-29, 1962 Mikhailidis DP, Jeremy JY, Barradas MA, Green N, Dardona P: Effect of ethanol on vascular prostacyclin (PGI,) synthesis, platelet aggregation and platelet thromboxane release, fir Med y 287:1495-97, 1983 Gorlin R: The biological actions and potential clinical significance of dietary w-3 fatty acids. Arch Intern Med 148:2043-48, 1988 Sullivan DR, Sanders TAB, Trayner IM, Thompson GR: Paradoxical elevation of LDL apoprotein B levels in hypertriglyceridemic patients and normal subjects ingesting fish oil. Atherosclerosis 61:129-34, 1986 Phillipson PE, Rothrock DW, Connor WE, Harris WS, Illingworth DR: Reduction of plasma lipids, lipoproteins, and apoproteins by dietary fish oils in patients with hypertriglyceridemia. N Engl ] Med 19:1210-16, 1985 Sigurdsson G, Nicoll A, Lewis B: Conversion of very low density lipoprotein to low density lipoprotein: a metabolic study of apolipoprotein B kinetics in human subjects. / Clin Invest 56:1481-90, 1975 Packard CJ, Munro A, Lorimer AR, Gotto AM, Shepherd J: Metabolism of apolipoprotein B in large triglyceriderich very low density lipoprotein of normal and hypertriglyceridemic subjects. ) Clin Invest 75:2178-92, 1984 Huff MW, Telford DE: Dietary fish oil increases conversion of very low density lipoprotein apoprotein B to low density lipoprotein. Arteriosclerosis 9:58-66, 1989 Kasim SE, Stern B, Khilnani S, McLin P, Baciorowski S, Jen K-L: Effects of omega-3 fish oils on lipid metabolism, glycemic control, and blood pressure in type II diabetic patients, j Clin Endocrinol Metab 67:1-5, 1988 Borkman M, Furler S, Storlien L, Kraegen EW, Simons L, Chesterman CN, Chisholm DJ: Effect of fish oil supplementation in non insulin dependent diabetes (NIDD) (Abstract). Diabetes Res Clin Pract 5 (Suppl. 1):S36, 1988 Popp-Snijders C, Schouten JA, Heine RJ, Van Der Meer J, Van Der Veen EA: Dietary supplementation of omega3 polyunsaturated fatty acids improves insulin sensitivity in non-insulin-dependent diabetes. Diabetes Res 4:141 — 47, 1987 Friday KE, Childs MT, Tsunehara CH, Fujimoto WY, Bierman EL, Ensinck JW: Elevated plasma glucose and lowered triglyceride levels from omega-3 fatty acid supplementation in type II diabetes. Diabetes Care 12:276-81, 1989 Albrink MJ, Ullrich IH, Blehschmidt NG, Mike P, Rogers JS: The beneficial effect of fish oil supplements on serum lipids and clotting function of patients with type II diabetes mellitus (Abstract). Diabetes 35 (Suppl. 1):43A, 1986 Fox AC, Burgess-Wilson M, Heptinstall S, Mitchell JRA: Platelet aggregation in whole blood determined using the Ultra-Flo 100 platelet counter. Thromb Haemostasis 48: 327-29, 1982 Saniabadi AR, Lowe GDO, Barbenel JC, Forbes CD: Further studies on the role of blood cells in spontaneous platelet aggregation. Thromb Res 38:225-32, 1985
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
T.J. HENDRA AND ASSOCIATES
49.
Simpson HCR, Meade TW, Stirling Y, Mann Jl, Chakrabarti R, Woolf L: Hypertriglyceridaemia and hypercoagulability. Lancet 1:786-90, 1983 50. Elkeles RS, Chakrabarti R, Vickers M, Stirling Y, Meade TW: Effect of treatment of hyperlipidaemia on haemostatic variables. Br Med j 281:973-74, 1980 51. Siddique HH, El-Hwuegi AS, El-Sagher MR, AN FM, Shaib FA, El-Zarok AA: Effect of olive and sunflower oil on serum lipids and myocardial triglyceride in rats. Planta Mec/:46-49, 1986 52. Mensink RP, Katan MB: Effect of monounsaturated fatty acids versus complex carbohydrates on high-density lipoproteins in healthy men and women. Lancet 1:12225, 1987 53. Barradas MA, Christofides JA, Jeremy JY, Mikhailidis DP, Fry DE, Dandona DP: Oleic acid supplementation alters
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
platelet function and membrane phospholipid arachidonic acid content. Biochem Soc Trans 17:506-507, 1989 54. Sirtori CR, Tremoli E, Gatti E, Montanari R, Sirtori M, Colli S, Gianfranceschi G, Maderna P, Dentone CZ, Testolin G, Gatti C: Controlled evaluation of fat intake in the Mediterranean diet: comparative activities of olive oil and corn oil on plasma lipids and platelets in high risk patients. Am j Clin Nutr 44:635-52, 1986 55. Castelli WP: The triglyceride issue: a view from Framingham. Am Heart J 112:432-37, 1986 56. Cambien F, Jaqueson A, Richard JL, Warnet JM, Ducimetiere P, Claude JR: Is the level of serum triglyceride a significant predictor of coronary death in normocholesterolemic subjects? Am J Epidemiol 124:624-32, 1986
829
The aim of this study was to evaluate the effects of a fish oil preparation (MaxEPA) on hemostatic function and fasting lipid and glucose levels in non-insulin-dependent diabetic (NIDDM) subjects. Eighty NIDDM outpatients aged 55.9 yr (mean SD 11.5 yr) participated in a prospective double-blind placebo-controlled study of MaxEPA capsules (10 g/day) or olive oil (control) treatment over 6 wk. Patients received either MaxEPA or olive oil in addition to preexisting therapy. Metabolic and hemostatic variables were measured before treatment and after 3 and 6 wk. Platelet membrane eicosapentaenoic acid (EPA) content increased in the treatment group (P < 0.001). MaxEPA supplementation was associated with a significant fall in total triglycerides (P < 0.001) but did not affect total cholesterol (P = 0.7) compared with control treatment. Fasting plasma glucose increased after 3 wk (P = 0.01) but not after 6 wk (P = 0.17) treatment with MaxEPA. Spontaneous platelet aggregation in whole blood fell in the MaxEPA group (P < 0.02) after 6 wk, but there were no changes in agonist-induced platelet aggregation, thromboxane generation in platelet-rich plasma, or plasma P-thromboglobulin and platelet factor IV levels. An increase in clotting factor VII (P = 0.02), without changes in fibrinogen or factor X levels, occurred in the MaxEPA group. Similar reductions in blood pressure were observed in both groups. Dietary supplementation with MaxEPA capsules (10 g/day) in NIDDM subjects is associated with improvement in hypertriglyceridemia but with deleterious effects in factor VII and blood glucose
From the Academic Unit of Diabetes and Endocrinology and the Department of Primary Health Care, University College and Middlesex School of Medicine, Whittington Hospital; and the Department of Chemical Pathology and Human Metabolism, Royal Free Hospital, Hampstead, London, United Kingdom. Address correspondence and reprint requests to Timothy J. Hendra, MRCP, Derriford Hospital, Plymouth PL6 8DH, UK. Received for publication 9 October 1989 and accepted in revised form 21 March 1990.
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
Timothy). Hendra, MRCP Mary E. Britton, MRCP David R. Roper, FIMLS Daniel Wagaine-Twabwe, MB lames Y. Jeremy, PhD Paresh Dandona, FRCP Andrew P. Haines, MRCP John S. Yudkin, FRCP
levels. Most indices of platelet function are unaffected by this therapy. Diabetes Care 13:821-29, 1990
N
on-insulin-dependent diabetic (NIDDM) subjects have an increased prevalence of coronary artery disease that is not explicable on the basis of-known risk factors for arteriosclerosis (1,2). There is, however, a marked clustering of such risk factors in these subjects, and abnormalities of blood pressure, lipid metabolism, and hemostatis may contribute to the etiology of the increased cardiovascular risk (3). Abnormalities in lipid levels in diabetic subjects are related to the type of diabetes, sex, and treatment (46). Most studies in NIDDM patients show elevated triglyceride and reduced high-density lipoprotein cholesterol (HDL-chol) levels, but low-density lipoprotein cholesterol (LDL-chol) levels are generally reported to be similar to those in nondiabetic subjects. Abnormalities in lipid levels may be more exaggerated in female diabetic patients, which might help to explain their excess cardiovascular risk compared with male diabetic subjects (4,7). Increases in platelet aggregation have frequently been reported in diabetic patients, and elevated levels of fibrinogen may contribute further to an increased risk of thrombosis (8,9). Some studies have suggested that these hemostatic abnormalities are related to microvascular disease (10), but there are also reports of an association between abnormalities of platelet aggregation (11) and fibrinogen (12) and large-vessel disease. It is, however, not possible to deduce from cross-sectional studies whether such differences are the cause or the consequence of large-vessel disease.
821
FISH OIL SUPPLEMENTATION IN DIABETIC SUBJECTS
An epidemiological association has been described between diets rich in fish and a low prevalence of cardiovascular disease (13). This association may be mediated through the beneficial effects of marine lipids, which are rich in polyunsaturated w-3 fatty acids (o)3FAs), on recognized cardiovascular risk factors. Dietary supplementation with fish oil in nondiabetic subjects lowers triglyceride levels by reducing both the triglyceride and apolipoprotein B (apoB) content of verylow-density lipoproteins (VLDLs; 14,15). A high intake offish oil lowers LDL-chol and LDL-apoB levels by decreasing LDL synthesis (16). Dietary supplementation with 160 mmHg systolic or 95 mmHg diastolic or treatment with antihypertensive medication. Drug histories were obtained for all subjects, with particular reference given to the ingestion of known platelet-active preparations during the 2 wk before study. Each subject was randomized in a double-blind manner to receive twice a day either five MaxEPA capsules (each containing 1 g fish oil; treatment group) or five olive oil capsules (control group). This dose of MaxEPA
822
provided 1.8 g eicosapentaenoic acid (EPA) and 1.2 3 docosahexaenoic acid (DHA) daily, together with smaller amounts of vitamins A, D, and E. Each MaxEPA capsule also contained 6 mg cholesterol. Subjects were instructed not to alter their dietary intake or exercise patterns, although no formal assessment of these variables was performed during the study. Participants were seen at the start of the study and c.t 3 and 6 wk. All subjects attended after a 9-h overnight fast and rested for 15 min before measurement of sitting blood pressure and venesection (with minimal venestasis) through a 19-gauge needle. The height and weight of each subject at each visit were also recorded. The following measurements were performed on blood taken at each visit: 1. Plasma glucose (glucose oxidase method, Beckman, Brea, CA) and full blood count (Technicon particle counter, Tarrytown, NY). 2. Glycosylated hemoglobin (agar gel electrophoresis, Corning, Halstead, Essex, UK). 3. Total serum triglyceride and cholesterol levels were measured enzymatically (29,30). HDL-chol was measured after heparin manganese precip tation and LDL-chol was calculated with the Friede:wald equation (31). Separate analyses of data on LDL-chol were performed in subjects whose level of triglyceride was >4.5 mM (31). 4. Plasma |3-thromboglobulin and platelet factor IV (radioimmunoassay kits from Amersham, Aylej;bury, Bucks, UK, and Abbott, Walsingham, Berks, UK, respectively). To exclude the effects of in vitro activation, data were analyzed only in subjects whose p-thromboglobulin-platelet factor IV ratio was >3.0 (32). All samples from each subject were analyzed in the same batch (within-coefficients of variation [C.V.s] assay 9.0% for p-thromboglobulin and 12.3% for platelet factor IV). 5. Plasma fibrinogen (Von Clauss method, BCL, Lewes, Sussex, UK), factor VII (1-stage assay using beagle plasma and rabbit thromboplastin from National UK Reference Laboratory, Withington Hospital, Manchester, UK), and factor X (Russell's viper venom assay, Diagnostic Reagents, Thame, Oxon, UK). Blood samples for clotting-factor estimation were rapidly frozen to -40°C after venesection so that all samples from each subject could be measured in the same assay (betweenassay C.V.s for fibrinogen 1.4%, factor VII 3.4%, and factor X 3.5%). Although cold activation of factor VII could be responsible for up to a 4-5% increase in activity, because all samples from both groups were not collected on ice before separation, this would not have affected our results. 6. Platelets were isolated from blood collected into EDTA, and platelet membrane phospholipid composition was determined (at 0 and 6 wk) with gas liquid chromatography to assess compliance with the MaxEPA therapy.
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
T.J. HENDRA AND ASSOCIATES
7. Spontaneous platelet aggregation in whole blood over 60 min with the use of the Ultra-Flo wholeblood platelet-counting method (33). C.V.s for the absolute fall in platelet count between 10 and 60 min were between 11.4 and 8.5%,. respectively. 8. In vitro whole blood (Ultra-Flo platelet-counting method; 33) and platelet-rich plasma (34) platelet aggregation studies were performed with four concentrations of each of three agonists, i.e., ADP, epinephrine, and collagen. Platelet sensitivity was expressed in terms of the EC50 (agonist concentration producing 50% of maximal aggregation) for each agonist by method as described previously (33). The C.V.s for agonist EC50 values in the same person on 1 day were 8.5 and 8.3% for ADP in whole blood and platelet-rich plasma, respectively, 6.9 and 7.5% for collagen, and 9.5 and 9.3% for epinephrine. 9. Platelet TXB2 release in platelet-rich plasma 5 min after the addition of ADP (10~5 M), epinephrine (10~5 M), and collagen (1 mg/ml; 35). Statistical analysis. Unpaired Student's t tests or MannWhitney U tests were used to compare between-group differences for each variable at baseline and also to compare changes in each variable in the treatment group between baseline and 3 wk and baseline and 6 wk with changes in control group. Analysis of variance was used to study the independent relationship of treatment with changes in variables while adjusting for differences in baseline characteristics between groups. Values are expressed as mean ± SD for normally distributed data and median (range) for skewed data. Changes are expressed as means (95% confidence interval [Cl]) for normally distributed data and median differences of change (P) for skewed data. The sample size was calculated as sufficient to detect a 24% change in LDL-chol between the groups at the 5% level with a power of 90%, which is the degree of change in LDLchol detected with MaxEPA therapy in a group of insulin-dependent diabetic (IDDM) patients (26). Results at 3 wk are only discussed if they suggest an interpretation at odds with the 6-wk findings.
RESULTS MaxEPA and placebo groups were matched for age, sex, ethnic origin, smoking habits, duration of diabetes, and mode of treatment (Table 1). Male subjects in the MaxEPA group were heavier than those in the control group. There was a similar prevalence of hypertension, peripheral vascular disease, neuropathy, and nephropathy in the two groups. However, coronary artery disease was more common in the MaxEPA group, whereas retinopathy was more frequent in the control group. All variables for which response to MaxEPA were assessed did not differ significantly between the groups at baseline.
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
TABLE 1 Clinical characteristics of subjects Characteristic
MaxEPA
Control
n Age (yr) Weight (kg) Men Women Duration of diabetes (yr) M/F Ethnic origin White Asian Afro-Carribean Smokers Complications Hypertension Coronary artery disease Peripheral vascular disease Diabetic nephropathy Diabetic retinopathy Diabetic neuropathy Treatment Diet alone Sulfonylureas Biguinides Biguinides and sulfonylureas
40 56.0(53.5, 58.5)
40 55.8(52.3, 59.3)
88.8(82.1, 95.5) 76.0 (62.2, 89.4) 6.7 (4.8, 8.5) 30/10
79.9 (74.4, 84.4)* 76.8 (66.6, 77.0) 8.0(5.7, 10.3) 25/15
39 0 1 8
38 1 1 8
4 14
5 3t
9 3 5 9
7 4 14* 10
8 12 8
8 14 3
12
15
Values of continuous variables are means, with 95% confidence intervals in parentheses. *P < 0.05, tP < 0.02.
One patient in the control group died during the trial from an exacerbation of bronchial asthma. One patient in the MaxEPA group withdrew, complaining of flatulence and abdominal pain. Four others (2 MaxEPA, 2 control) withdrew due to difficulty swallowing the capsules. Nine patients (4 MaxEPA, 5 control) were taking platelet-active drugs and did not have platelet function studies performed. In analyzing changes in triglycerides, the results of one patient with gross hypertriglyceridemia were excluded to allow parametric testing of otherwise skewed data. The mean level of triglyceride fell by 30.0% in the treatment group compared to a rise of 10.0% in the control group. There were no significant changes in serum total cholesterol or HDL-chol, but calculated LDL-chol rose within the treatment group. When compared with baseline this rise was significant (mean 0.26 mM, 95% Cl 0.03-0.49, P = 0.03), but when the change in the treatment group was compared with the change in the control group, the difference attributable to treatment failed to reach significance (mean 0.26 mM, 95% Cl - 0 . 0 3 - 0 . 5 6 , P = 0.085; Table 2). These results for LDL-chol were unaffected by excluding subjects with triglyceride levels >4.5 mM. Fasting blood glucose rose within the MaxEPA group at both 3 and 6 wk. However, the mean change in
823
2.0(1.7,2.4) 2.2(1.7,2.6)
-0.90 (-1.30, -0.50) T -» \*
r -»T
-»
Tt -» -*
t -»
v
Fasting plasma glucose HbA,
i I i
I i i
TC
LDL-chol
ft
HDL-chol
Platelets
I ADP-stimulated aggregation i TXB2 4 Spontaneous platelet aggregation
Effect offish oil supplements
t VII -»X,I
Clotting factors
64 .1 (60 .7, 67.6) 63 .5 (59..6, 67.5)
Control
4.59 (0.90, 8. 29) 0.016
62 .9 (59..6, 66.3) 66 .5 (63 .4, 69.6)
MaxEPA
60 min
TC, triglyceride; TXB2, thromboxane B2 generation; u>3FA, w-3 fatty acid; LDL-chol, low-density lipoprotein cholesterol; HDL-chol, high-density lipoprotein cholesterol. Where a control was used effects are in comparison with control, otherwise they are compared to baseline. 'Transient. tMarginal significance (P = 0.07).
n
Refs.
Investigations
Duration (wk)
Dose fish oil (dose o)3FA g/day)
TABLE 6 Fish oil supplementation in non-insulin-dependent diabetes mellitus
Values are means, with 95% confidence intervals in parentheses. Comparisons of changes between groups are compared with unpaired Student's t tests.
0 wk 6 wk Change attributable to treatment P
MaxEPA
10 min
Changes in spontaneous platelet aggregation with MaxEPA and olive oil: percentage of platelets remaining after aggregation
TABLE
FISH OIL SUPPLEMENTATION IN DIABETIC SUBJECTS
fasting blood glucose attributable to MaxEPA treatment was significant only at 3 wk (1.44 mM, 95% Cl 0.34- DISCUSSION 2.5, P = 0.01) and not at 6 wk (0.97 mM, 95% Cl - 0 . 3 9 - 2 . 3 , P = 0.17; Table 3). There were no signifn this study, we demonstrated that dietary suppleicant changes in HbA, in either treatment group. Weight mentation with MaxEPA fish oil capsules for 6 w< was unchanged in both groups, and changes in meain a dose of 10 g/day produces a fall in triglyceride sures of glycemia were unrelated to changes in weight levels and a reduction in spontaneous platelet agin either group. gregation in whole blood. These changes were, howLevels of factor VII rose in the treatment group, with ever, associated with a transient deterioration in blood the change being significantly different from that in the glucose control, an increase in clotting factor VII, and control group at 6 wk (mean 19.3%, 95% Cl 3.9-34.4, a marginal increase in calculated LDL-chol levels. There P = 0.02; Table 3). There was no significant change in were similar significant falls in blood pressure and in levels of fibrinogen or factor X. Systolic and diastolic TXB2 generation in response to collagen in both the blood pressure fell similarly in both groups, with no MaxEPA and olive oil (control groups). No effects were significant difference in these changes between groups observed on standard measures of in vivo platelet function, factor X or fibrinogen levels, or agonist-induced (Table 4). A significant fall in collagen-stimulated TXB2 gener- platelet aggregation with whole blood and platelet-rich ation occurred within the treatment group from baseline plasma techniques. The observed increase in membrane to 6 wk (mean 125 \xg/L [25th-75th percentile 90-230] EPA levels was a measure of subject compliance with vs. 98.5 |Ag/L [59-122], P = 0.003). However, when the MaxEPA capsules and is similar to increases obthis change was compared with the change in the con- served in previous studies in healthy and diabetic subtrol group, the difference failed to reach significance jects (13,26). Studies in nondiabetic subjects have demonstrated (Table 4). ADP and epinephrine-stimulated TXB2 genthat changes in serum lipid levels with fish oil suppleeration was not significantly altered in either group. No significant changes in plasma p-thromboglobulin ments are dose related and that a hypotriglyceridemic or platelet factor IV levels occurred in either group. Ag- effect is seen with low dosages of o>3FAs (3-10 g/day) onist-induced platelet aggregation in either whole blood (14,15,36), whereas dosages of 5-20 g/day are necesor platelet-rich plasma was unaffected by either MaxEPA sary for a hypocholesterolemic effect (16). Some studies or olive oil. Platelet-rich plasma from 24 patients (14 in have suggested that lower intakes of o>3FAs may result MaxEPA group, 10 in control group) did not show sec- in increases in LDL-chol and apoB concentrations, parondary aggregation to epinephrine at any time point, ticularly in subjects with type V hyperlipoproteinemia and in these subjects, no effect of either treatment was (37,38). Because the major catabolic pathway for VLDI. observed on primary aggregation. Spontaneous platelet is via conversion to LDL, this may produce an increase aggregation in whole blood at all time points was sig- in LDL-chol parallel to the fall in triglycerides (39). Alnificantly reduced after 6 wk of MaxEPA therapy (Ta- ternatively, fish oil supplements may lead to hepatic secretion of smaller-than-normal VLDL particles, which ble 5). MaxEPA treatment was associated with a significant are known to be preferentially converted to LDL, comincrease in platelet membrane EPA content (expressed pared to larger triglyceride-rich VLDL (40). In our study, as a percentage of total membrane fatty acids; Table 2). as in that by Haines et al. (26) demonstrating a 24% rise Over 70% of subjects in the MaxEPA group who com- in LDL-chol in IDDM subjects, the estimation of LDLpleted the study had more than a threefold increase in chol concentration was calculated with the Friedewald EPA content at 6 wk compared with baseline. Analysis equation, which assumes a constant ratio of triglyceride of the patients in the two groups combined revealed to cholesterol in VLDL. Because preferential lowering marked negative correlation between changes in EPA of triglycerides in VLDL has been shown to occur with and arachidonic acid (rs = - 0 . 4 2 , n = 50, P < 0.001). MaxEPA, this may artificially elevate calculated LDLThere were also significant negative correlations of chol levels (41). The level of total cholesterol was unchanges in EPA with triglyceride levels (rs = - 0 . 5 0 , changed with MaxEPA treatment in this study, although n = 62, P < 0.001) and collagen-induced TXB2 pro- it was increased by MaxEPA therapy in the study of duction (rs = - 0 . 2 6 , n = 50, P = 0.04). Changes in IDDM subjects. Currently, there have been 10 studies published asspontaneous aggregation at 60 min were correlated negatively with changes in arachidonic acid {rs = - 0 . 4 1 , sessing fish oil supplementation in diabetic subjects n = 42, P = 0.004) and positively with changes in EPA (24-28,42-46), of which 7 are in NIDDM patients (27,28,42-46). The total number of patients treated in (rs = 0.25, n = 52, P = 0.04). The effects of mismatches in body weight and prev- these 7 studies was 73, and only 2 studies used a control alence of retinopathy and ischemic heart disease be- group. These studies are summarized in Table 6. In tween the groups at baseline were assessed with analysis some of these studies, there was evidence of deterior of variance. The significance of changes in all the mea- ration in glycemic control and increases in LDL-chol o sured variables was unaffected when baseline differ- apoB levels, although the dosage of MaxEPA used and duration of treatment differed between investigators ences were incorporated into the analysis.
I
826
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
T.J. HENDRA AND ASSOCIATES
(27,28,45). In 1 study, there is evidence that increases in fasting plasma glucose may be associated with increased basal hepatic glucose output and reduced mealand glucagon-stimulated insulin release (28). However, another study demonstrated improved in vivo insulin sensitivity with MaxEPA supplements (44). We found no evidence supporting previous studies in nondiabetic subjects that have suggested that supplementation with o)3FAs may reduce other cardiovascular risk factors, including fibrinogen levels (21), blood pressure (18,23), and agonist-induced platelet aggregability (18,19,22), in addition to their effects on lipids. However, recent studies in larger numbers of subjects have questioned the effects offish oils on platelet aggregation. The effect of MaxEPA therapy on platelet function has not been investigated previously in NIDDM subjects, but the results of this study are similar to those observed by others in IDDM subjects (24-26), showing reductions in platelet TXB2 generation after MaxEPA therapy (26). However, we were unable to demonstrate reduced platelet aggregation in vivo or in response to agonists in vitro, probably because a greater degree of inhibition of TXB2 synthesis is required to affect in vitro platelet aggregation, a finding that may be common to subjects with atherosclerosis (28). The observed correlations between changes in membrane fatty acids and those in spontaneous platelet aggregation are of uncertain significance, because previous work has indicated that spontaneous platelet aggregation in whole blood may be mediated via non-TXB2-dependent mechanisms (47,48). Factor VII levels were elevated by MaxEPA in this study, which may be deleterious on cardiovascular risk. This finding was unexpected in view of previous studies that have shown that interventions lowering triglyceride levels may be associated with a reduction in factor VII levels (49,50). In this study, olive oil (containing 8-10% polyunsaturated fatty acids) was used as a control in the same dosage as the fish oil. In the olive oil treatment group, there were no significant changes in any of the lipid variables of hemostatic function, with the exception of a fall in collagen-induced TXB2 production. This effect was not associated with any changes in spontaneous platelet aggregation. There is evidence from studies in nondiabetic subjects and animals that olive oil can affect both serum lipids and platelet function, although the dosages used in human studies were much larger than those used herein (51-54). In summary, we have shown that theoretically advantageous changes in serum triglycerides and perhaps collagen-stimulated TXB2 production in platelet-rich plasma occur in NIDDM patients taking 10 g/day of fish oil. The significance of triglycerides as an independent cardiovascular risk factor is debated (55,56), and that of collagen-stimulated TXB2 production is unknown. Potentially deleterious effects of fish oil on other possible risk factors, i.e., blood glucose control, LDL-chol, and factor VII, led us to conclude that fish oil in this dosage
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
cannot be recommended to NIDDM subjects as an aid to reducing cardiovascular risk.
ACKNOWLEDGMENTS
T.J.H., M.E.B, and D.W.-T. received financial support from Duncan Flockhart & Company, Greenford, Middlesex, United Kingdom. We thank J. Wager for technical assistance, Caroline Jackson for computing expertise, and Sharrone McCullough for secretarial assistance.
REFERENCES 1. Jarrett RJ, Keen H, Chakrabarti R: Diabetes, hyperglycemia and arterial disease. In Complications of Diabetes. 2nd ed. Keen H, Jarrett RJ, Eds. London, Arnold, 1982, p. 179-204 2. jarrett RJ: Type 2 (non-insulin-dependent) diabetes mellitus and coronary heart disease: chicken, egg or neither? Diabetologia 26:99-102, 1984 3. Wingard DL, Barrett-Connor E, Criqui MH, Suarez L: Clustering of heart disease risk factors in diabetic compared to nondiabetic adults. Am) Epidemiol 117:19-26, 1983 4. Walden CE, Knopp RH, Wahl PW, Beach KW, Strandness E: Sex differences in the effect of diabetes mellitus on lipoprotein triglyceride and cholesterol concentrations. N Engl) Med 311:953-59, 1984 5. Seviour PW, Teal TK, Richmond W, Elkeles RS: Serum lipids, lipoproteins and macrovascular disease in non-insulin-dependent diabetics: a possible new approach to prevention. Diabetic Med 5:166-71, 1988 6. Howard BV: Lipoprotein metabolism in diabetes mellitus. ) Lipid Res 28:613-28, 1987 7. Rytter L, Troelsen S, Beck-Nielsen H: Prevalence and mortality of acute myocardial infarction in patients with diabetes. Diabetes Care 8:230-34, 1985 8. Hendra TJ, Betteridge DJ: Platelet function, platelet prostenoids, and vascular prostacyclin in diabetes mellitus. Prostaglandins Leukotrienes Essent Fatty Acids 35:197—
212, 1989 9. Fuller JH, Keen H, Jarrett RJ, Omer T, Meade TW, North WRS, Stirling Y: Haemostatic variables associated with diabetes and its complications. Br Med J 2:964-66, 1979 10. Jennings PE, Barnett AH: New approaches to the pathogenesis and treatment of diabetic microangiopathy. Diabetic Med 5:111 -17, 1988 11. Packham MA, Mustard JF: The role of platelets in the development and complications of atherosclerosis. Semin Hemato/23:8-26, 1986 12. Fuller JH: The haemocoagulation system and macroangiopathy in insulin dependent (type 1) diabetes. Ann Clin Res 16:137-41, 1984 13. Leaf A, Weber PC: Cardiovascular effects of n-3 fatty acids. N Engl I Med 318:549-57, 1988 14. Sanders TA, Sullivan DR, Reeve J, Thompson GR: Triglyceride-lowering effect of marine polyunsaturates in patients with hypertriglyceridemia. Arteriosclerosis 5:45965, 1985 15. Nestel PJ, Connor WE, Reardon MF, Connor S, Wong S,
827
FISH OIL SUPPLEMENTATION IN DIABETIC SUBJECTS
16.
17.
18.
19.
20.
21. 22.
23.
24.
25. 26.
27. 28.
29. 30. 31.
828
Boston R: Suppression by diets rich in fish oil of very low density lipoprotein production in man. J Clin Invest 74:82-89, 1984 Illingworth DR, Harris WS, Connor WE: Inhibition of low density lipoprotein synthesis by dietary omega-3 fatty acids in humans. Arteriosclerosis 4:270-75, 1984 Hay CRM, Durber AP, Saynor R: Effect of fish oil on platelet kinetics in patients with ischaemic heart disease. Lancet 1:1269-72, 1982 Lorenz R, Spengler U, Fischer S, Duhm J, Weber PC: Platelet function, thromboxane formation and blood pressure control during supplementation of the western diet with cod liver oil. Circulation 67:504-11, 1983 Ahmed AA, Holub BJ: Alteration and recovery of bleeding times, platelet aggregation and fatty acid composition of individual phospholipids in platelets of human subjects receiving a supplement of cod liver oil. Lipids 19:61724, 1984 Knapp HR, Reilly IAG, Alessandrini P, FitzGerald GA: In vivo indices of platelet and vascular function during fishoil administration in patients with atherosclerosis. N Engl I Med 314:937-42, 1986 Hostmark AT, Bjerkedal T, Kierulf P, Flaten H, Ulshagen K: Fish oil and plasma fibrinogen. Br Med) 297:180-81, 1988 Sanders TAB, Vickers M, Haines AP: Effect on blood lipids and haemostasis of a supplement of cod-liver oil, rich in eicosapentaenoic acid and docosahexaenoic acids in healthy young men. Clin Sci 61:317-24, 1981 Rogers S, James KS, Butland BK, Etherington MD, O'Brien JR, Jones JG: Effects of a fish oil supplement on serum lipids, blood pressure, bleeding time, hemostatic and rheological variables. Atherosclerosis 63:137-43, 1987 Beitz J, Schimke E, Liebaug U, Block H-U, Beitz A, Honigmann G, Sziegoleit W, Muller G, Mest H-J: Influence of a cod liver oil diet in healthy and insulin-dependent diabetic volunteers on fatty acid pattern, inhibition of prostacyclin formation by low density lipoprotein (LDL) and platelet thromboxane. Klin Wochenschr 64:793-99, 1986 Tilvis RS, Rasi V, Viinikka L, Ylikorkala O, Miettinen TA: Effects of purified fish oil on platelet lipids and function in diabetic women. Clin Chim Ada 164:315-22, 1987 Haines AP, Sanders TAB, Imeson JD, Mahler RF, Martin J, Mistry M, Vickers M, Wallace PG: Effects of a fish oil supplement on platelet function, haemostatic variables and albuminuria in insulin-dependent diabetics. Thromb Res 43:643-55, 1986 Schectman G, Kaul S, Kissebah AH: Effect offish oil concentrate on lipoprotein composition in NIDDM. Diabetes 37:1567-73, 1988 Glauber H, Wallace P, Griver K, Brechtel G: Adverse metabolic effect of omega-3 fatty acids in non-insulindependent diabetes mellitus. Ann Intern Med 108:66368, 1988 Roescholau P, Bernt E, Gruber W: Enzymatic determination of total cholesterol in serum. Z Klin Chem Klin Biochem 12:226, 1974 Bucolo G, David H: Quantitative determination of serum triglycerides by the use of enzymes. Clin Chem 19:47682, 1973 Friedewald WT, Levy Rl, Frederickson DS: Estimation of the concentration of low density lipoprotein cholesterol in plasma without use of the preparative centrifuge. Clin Chem 18:449-502, 1972
32. 33.
34. 35.
36. 37.
38.
39.
40.
41. 42.
43.
44.
45.
46.
47.
48.
Kaplan KL, Owen J: Plasma levels of beta-thromboglobulin and platelet factor 4 as indices of platelet activation in vivo. Blood 57:199-202, 1981 Hendra TJ, Oughton J, Smith CCT, Betteridge DJ, Yudkin JS: Exercise-induced changes in platelet aggregation: a comparison of whole blood and platelet rich plasma techniques. Thromb Res 52:443-51, 1988 Born GVR: Aggregation of platelets by adenosine diphos.phate and its reversal. Nature (Lond) 194:927-29, 1962 Mikhailidis DP, Jeremy JY, Barradas MA, Green N, Dardona P: Effect of ethanol on vascular prostacyclin (PGI,) synthesis, platelet aggregation and platelet thromboxane release, fir Med y 287:1495-97, 1983 Gorlin R: The biological actions and potential clinical significance of dietary w-3 fatty acids. Arch Intern Med 148:2043-48, 1988 Sullivan DR, Sanders TAB, Trayner IM, Thompson GR: Paradoxical elevation of LDL apoprotein B levels in hypertriglyceridemic patients and normal subjects ingesting fish oil. Atherosclerosis 61:129-34, 1986 Phillipson PE, Rothrock DW, Connor WE, Harris WS, Illingworth DR: Reduction of plasma lipids, lipoproteins, and apoproteins by dietary fish oils in patients with hypertriglyceridemia. N Engl ] Med 19:1210-16, 1985 Sigurdsson G, Nicoll A, Lewis B: Conversion of very low density lipoprotein to low density lipoprotein: a metabolic study of apolipoprotein B kinetics in human subjects. / Clin Invest 56:1481-90, 1975 Packard CJ, Munro A, Lorimer AR, Gotto AM, Shepherd J: Metabolism of apolipoprotein B in large triglyceriderich very low density lipoprotein of normal and hypertriglyceridemic subjects. ) Clin Invest 75:2178-92, 1984 Huff MW, Telford DE: Dietary fish oil increases conversion of very low density lipoprotein apoprotein B to low density lipoprotein. Arteriosclerosis 9:58-66, 1989 Kasim SE, Stern B, Khilnani S, McLin P, Baciorowski S, Jen K-L: Effects of omega-3 fish oils on lipid metabolism, glycemic control, and blood pressure in type II diabetic patients, j Clin Endocrinol Metab 67:1-5, 1988 Borkman M, Furler S, Storlien L, Kraegen EW, Simons L, Chesterman CN, Chisholm DJ: Effect of fish oil supplementation in non insulin dependent diabetes (NIDD) (Abstract). Diabetes Res Clin Pract 5 (Suppl. 1):S36, 1988 Popp-Snijders C, Schouten JA, Heine RJ, Van Der Meer J, Van Der Veen EA: Dietary supplementation of omega3 polyunsaturated fatty acids improves insulin sensitivity in non-insulin-dependent diabetes. Diabetes Res 4:141 — 47, 1987 Friday KE, Childs MT, Tsunehara CH, Fujimoto WY, Bierman EL, Ensinck JW: Elevated plasma glucose and lowered triglyceride levels from omega-3 fatty acid supplementation in type II diabetes. Diabetes Care 12:276-81, 1989 Albrink MJ, Ullrich IH, Blehschmidt NG, Mike P, Rogers JS: The beneficial effect of fish oil supplements on serum lipids and clotting function of patients with type II diabetes mellitus (Abstract). Diabetes 35 (Suppl. 1):43A, 1986 Fox AC, Burgess-Wilson M, Heptinstall S, Mitchell JRA: Platelet aggregation in whole blood determined using the Ultra-Flo 100 platelet counter. Thromb Haemostasis 48: 327-29, 1982 Saniabadi AR, Lowe GDO, Barbenel JC, Forbes CD: Further studies on the role of blood cells in spontaneous platelet aggregation. Thromb Res 38:225-32, 1985
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
T.J. HENDRA AND ASSOCIATES
49.
Simpson HCR, Meade TW, Stirling Y, Mann Jl, Chakrabarti R, Woolf L: Hypertriglyceridaemia and hypercoagulability. Lancet 1:786-90, 1983 50. Elkeles RS, Chakrabarti R, Vickers M, Stirling Y, Meade TW: Effect of treatment of hyperlipidaemia on haemostatic variables. Br Med j 281:973-74, 1980 51. Siddique HH, El-Hwuegi AS, El-Sagher MR, AN FM, Shaib FA, El-Zarok AA: Effect of olive and sunflower oil on serum lipids and myocardial triglyceride in rats. Planta Mec/:46-49, 1986 52. Mensink RP, Katan MB: Effect of monounsaturated fatty acids versus complex carbohydrates on high-density lipoproteins in healthy men and women. Lancet 1:12225, 1987 53. Barradas MA, Christofides JA, Jeremy JY, Mikhailidis DP, Fry DE, Dandona DP: Oleic acid supplementation alters
DIABETES CARE, VOL. 13, NO. 8, AUGUST 1990
platelet function and membrane phospholipid arachidonic acid content. Biochem Soc Trans 17:506-507, 1989 54. Sirtori CR, Tremoli E, Gatti E, Montanari R, Sirtori M, Colli S, Gianfranceschi G, Maderna P, Dentone CZ, Testolin G, Gatti C: Controlled evaluation of fat intake in the Mediterranean diet: comparative activities of olive oil and corn oil on plasma lipids and platelets in high risk patients. Am j Clin Nutr 44:635-52, 1986 55. Castelli WP: The triglyceride issue: a view from Framingham. Am Heart J 112:432-37, 1986 56. Cambien F, Jaqueson A, Richard JL, Warnet JM, Ducimetiere P, Claude JR: Is the level of serum triglyceride a significant predictor of coronary death in normocholesterolemic subjects? Am J Epidemiol 124:624-32, 1986
829
