RSS Feeds
Effects of n-3 Polyunsaturated Fatty Acids on Glucose Homeostasis and Blood Pressure in Essential Hypertension
A Randomized, Controlled Trial
- Ingrid Toft, MD;
- Kaare H. Bonaa, MD, PhD;
- Ole C. Ingebretsen, MD, PhD;
- Arne Nordoy, MD, PhD; and
- Trond Jenssen, MD, PhD
- From the University of Tromso and Tromso University Hospital, Tromso, Norway. Acknowledgments: The authors thank the staff of the General Clinical Research Centre and thank Jorunn Ekrem, Angstromse Lund Bendiksen, Gro Bolstad, and Hege Iversen for technical assistance. Grant Support: In part by the Norwegian Diabetes Association, Nordic Research Funding, and the Research Council of Norway. Pronova Biocare provided the test medication. Requests for Reprints: Ingrid Toft, MD, Department of Internal Medicine, Tromso University Hospital, N-9038 Tromso, Norway. Current Author Addresses: Drs. Toft, Nordoy, and Jenssen: Department of Internal Medicine, Tromso University Hospital, N-9038 Tromso, Norway. Dr. Bonaa: Institute of Community Medicine, University of Tromso, N-9037 Tromso, Norway. Dr. Ingebretsen: Department of Clinical Chemistry, Tromso University Hospital, N-9038 Tromso, Norway.
Abstract
Objective: To determine whether dietary supplementation with fish oil adversely affects glycemic control in patients with hypertension.
Design: Randomized, double-blind, placebo-controlled study.
Patients: 78 persons with untreated hypertension recruited from a population survey.
Intervention: Participants were randomly assigned to receive eicosapentaenoic and docosahexaenoic acids, 4 g/d, or corn oil placebo, 4 g/d, for 16 weeks.
Measurements: An oral glucose tolerance test; assessments of insulin release, glucose disposal, and insulin sensitivity done using the hyperglycemic clamp technique to keep plasma glucose levels at 10 mmol/L for 180 minutes; assessment of insulin sensitivity done using a euglycemic hyperinsulinemic clamp technique (infusing insulin and glucose to keep plasma glucose levels at 5 mmol/L); assessments of lipid levels and blood pressure. Measurements were done before and after intervention.
Results: Changes in integrated glucose and insulin response after the oral glucose challenge did not differ between the fish oil and corn oil groups after intervention (−0.6 ± 0.7 compared with −1.0 ± 0.6 mmol/L [P > 0.3] for integrated glucose and 143 ± 76 compared with 169 ± 84 pmol/L [P > 0.3] for insulin response). Changes in first-phase insulin release (34 ± 72 pmol/L in the fish oil group compared with 191 ± 112 pmol/L in the corn oil group [P > 0.3]), second-phase insulin release (179 ± 66 pmol/L compared with 257 ± 122 pmol/L [P > 0.3]), and insulin sensitivity index (−0.03 ± 0.01 compared with −0.01 ± 0.01 [µmol/kg · min ÷ pmol/L]; P > 0.3) were also similar in both groups after treatment. Fish oil lowered systolic blood pressure by 3.8 mm Hg more than control (P = 0.04) and lowered diastolic blood pressure by 2.0 mm Hg more than control (P = 0.10). After fish oil treatment, triglyceride levels decreased by 0.28 ± 0.08 mmol/L more than control (P = 0.01), and very-low-density lipoprotein cholesterol levels decreased by 0.13 ± 0.04 mmol/L more than control (P = 0.01).
Conclusion: Fish oil, in doses that reduce blood pressure and lipid levels in hypertensive persons, does not adversely affect glucose metabolism.
Hypertension is a well-documented risk factor for coronary heart disease, but the widespread use of antihypertensive treatment has not resulted in the expected reduction in coronary heart disease mortality [1]. Persons with hypertension tend to have disturbances in glucose and lipid metabolism [2-4] that may contribute to their excess risk for coronary heart disease.
Fish oils rich in polyunsaturated fatty acids of the n-3 family may protect against ischemic cardiovascular disease [5-7]. In hypertensive patients, a modest blood pressure-lowering effect has been shown after fish oil intake in some [8-12] but not all [13-15] studies. Fish oil may favorably affect platelet aggregation [16, 17], hepatic triglyceride and very-low-density lipoprotein (VLDL) cholesterol formation [18-21], and vascular prostaglandin production [9, 16, 22]. It has also been reported to suppress intimal smooth-muscle cell proliferation by inhibiting monocyte and neutrophil chemotaxis [23] and the vascular endothelial production of platelet-derived growth factor-like protein [24]. These antiatherosclerotic effects may be important in preventing the development of coronary heart disease in patients with hypertension [25].
Conflicting results have been published about the effects of fish oil on glucose homeostasis [26-41]. Some [26-31] but not all [32-41] studies have reported that fish oil has detrimental effects on glycemic control in glucose-intolerant persons and in persons with type 2 diabetes. The extent to which the findings from these studies can be generalized is constrained by limitations in study design. Only a few studies [26, 28, 30, 34, 35, 38] have used the classic glucose clamp technique to measure glucose and insulin dynamics, and no studies have examined the effects of fish oil on glucose homeostasis in nondiabetic persons with hypertension. Given the present gaps in the literature, the safety of fish oil supplementation for persons with hypertension has been disputed [42]. We therefore did a randomized, double-blind, placebo-controlled trial in 78 persons with untreated, stable hypertension to study the effects of n-3 polyunsaturated fatty acids on glucose and insulin kinetics, blood pressure, serum lipids, and the incorporation of fatty acid into plasma phospholipids.
Methods
Participants
In 1986-1987, 21 826 persons (81.3% of the population [age range: men, 20 to 61 years; women, 20 to 56 years] living in the municipality of Tromso, Norway, participated in a health survey [43]. On the basis of that survey, 156 hypertensive persons were enrolled in a 10-week trial of dietary supplementation with n-3 polyunsaturated fatty acids [8]. The trial was completed in May 1988. In May 1991 and February 1992, the persons who had participated in the trial were invited to have physical examinations at the Clinical Research Unit of the University Hospital of Tromso as part of recruitment into our study. Of the persons invited, 103 volunteered. Each completed a questionnaire about previous and present illnesses, family history, medication, fish oil intake, physical activity, and smoking and alcohol habits, and each had a laboratory screening that included an oral glucose tolerance test and blood pressure measurements. Fifty-eight participants were receiving no medication and had systolic blood pressure measurements of less than 190 mm Hg and diastolic blood pressure measurements between 90 and 110 mm Hg on three separate occasions. Each had a body mass index of less than 32 kg/m2 body surface area and appeared otherwise healthy. They all participated in the study, as did 26 hypertensive persons recruited from the primary health care services using criteria identical to those described above. Four of these volunteers had been treated with antihypertensive drugs (atenolol, amlodipine, or felodipine); this therapy was discontinued at least 8 weeks before the trial. The 84 participants were encouraged not to change their diets or lifestyles during the study. Those who used cod liver oil supplements were instructed to discontinue this practice 12 months before the study started. The study was approved by the Regional Board of Research Ethics, and each participant gave written informed consent before participation.
Study Design
The participants were randomly assigned to receive either fish oil or corn oil. A person who was not involved in trial management did the randomization using a Statgraphic random number generator [44]. The list of randomization numbers and the codes were sent to the manufacturer of the fish oil and corn oil capsules (Pronova Biocare, Oslo, Norway). The boxes labeled with the randomization numbers were given to the participants in the sequence at which they met. The researchers doing the experiments were blinded to treatment assignments, and the randomization codes were not broken until all laboratory measurements had been done. The fish oil group received 85% eicosapentaenoic acid (C20:5n3) and docosahexaenoic acid (C22:6n3), 4 g/d, as ethyl esters (Omacor, Pronova Biocare, Oslo, Norway). To compensate for the extra daily energy intake received by those assigned to intervention with polyunsaturated fatty acids, the control group was given corn oil, 4 g/d, containing 56% linoleic acid (C18:2n6) and 26% oleic acid (C18:1n9). The fish and corn oils were given in indistinguishable soft gelatin capsules that each contained 1 g of oil. The intervention period lasted 16 weeks. Compliance was checked by counting leftover capsules and by measuring the concentrations of fatty acids in plasma phospholipids before and after intervention. Glucose tolerance studies were done during the last week before treatment and during the last week of intervention. A weight-maintenance diet was held 3 days before the experiments, and participants were asked to abstain from alcohol during this period. All studies were done at 0800 h after an overnight fast. Side effects, compliance, intercurrent diseases, and blood pressure were assessed by interview and physical examination every fourth week during treatment.
Clinical and Laboratory Measurements
Three blood pressure measurements were recorded on each of 2 separate days both before and after intervention by the same investigator using the same stethoscope and mercury sphygmomanometer. The mean of these measurements was used in the analysis. Measurements were done after each patient had rested, comfortably seated, for 10 minutes; Korotkoff test phases 1 and 5 were recorded as systolic and diastolic blood pressures, respectively. Mean arterial pressure was calculated as the diastolic pressure plus one third of the pulse pressure. Waist-to-hip ratio was calculated as the body circumference midway between the inferior border of the rib cage and the superior border of the iliac crest, divided by the maximal body circumference at the buttocks [45].
All participants had an oral glucose tolerance test with 1 g of dextrose per kg body weight or a maximum of 75 g of dextrose. The integrated increase of plasma glucose and insulin levels above baseline measurements after an oral glucose tolerance test was calculated as arbitrary incremental area units over the 2-hour sampling time.
On a separate day, we used a standard hyperglycemic clamp technique to study both glucose-stimulated insulin secretion and insulin sensitivity [46, 47]. Dextrose was infused into an antecubital vein. We measured the blood sugar level every 5 to 10 minutes to keep it stable at 10 mmol/L for 3 hours by variable infusion rates. Blood was drawn from a cannulated dorsal hand vein without stasis; we arterialized the blood by keeping the hand in a heating device at 65 °C [48]. Blood samples for insulin and C-peptide measurements were drawn at −30, 0, 2.5, 5, 7.5, 10, 15, 20, 40, 60, 80, 100, 120, 140, 160, and 180 minutes. First-phase insulin release, which reflects the early insulin peak secreted from the pancreatic β-cells in response to glucose stimulation, was calculated as the area under the insulin curve over the initial 10-minute period of the hyperglycemic clamp technique. Second-phase insulin release, which is a measure of β-cell function under sustained elevated glucose levels, was calculated as the area under the insulin curve from 120 to 180 minutes of the clamp period.
On a third day, we used a euglycemic, hyperinsulinemic clamp technique [46, 47], which is the gold standard for measuring insulin sensitivity. However, this method does not give information about β-cell function. In general, insulin is infused at a rate of 40 mU/m2 body surface area per minute for 180 minutes, inducing plasma insulin levels of about 400 pmol/L [46]. At this plasma insulin level, hepatic glucose production is zero. Plasma glucose level was maintained at 5 mmol/L by a variable glucose infusion. The glucose infusion rate therefore equals the uptake rate of glucose in the body. The insulin sensitivity index can be calculated by using both the hyperglycemic and the euglycemic clamp techniques by dividing the mean glucose infusion rate during the last hour of the clamp period (µmol/kg·min) by the average plasma insulin level in the same period of time (pmol/L). The insulin sensitivity index measures how efficiently plasma insulin induces glucose uptake in insulin-sensitive tissues, such as fat and muscle. The insulin sensitivity index calculated by using the hyperglycemic clamp technique has been shown to be highly correlated with the insulin sensitivity index calculated by using the classic euglycemic, hyperinsulinemic clamp technique [46, 47]. To compare the insulin sensitivity indexes obtained with the two clamp techniques, we used the euglycemic, hyperinsulinemic clamp technique in 31 randomly selected participants on this third day.
Plasma glucose levels were analyzed at the bedside with a Yellow Spring Instruments glucose analyzer (2300 STAT PLUS, Yellow Springs, Ohio). All other blood samples were stored at −70 °C until study completion. Plasma insulin [49] and C-peptide [50] levels were measured by radioimmunoassay methods as previously published. These two assays do not cross-react with each other, and each has a cross-reactivity with proinsulin of less than 15%. Glycosylated hemoglobin A1c levels were measured by a liquid chromatographic procedure (Diamat, Bio-Rad Laboratories, Munich, Germany; reference range, 4.0% to 6.5%). Serum cholesterol and triglyceride levels were measured using a Hitachi 737 Automatic Analyzer with a kit from Boehringer Mannheim [Germany]. High-density lipoprotein (HDL) cholesterol levels were measured after HDL was isolated according to the method described by Burstein and colleagues [51]. Very-low-density lipoprotein (VLDL) cholesterol levels were calculated as 0.46 multiplied by the triglyceride level, and low-density lipoprotein (LDL) cholesterol levels were calculated as total cholesterol level minus the sum of the VLDL and HDL cholesterol levels, according to the formula of Friedewald and colleagues [52]. Apolipoprotein A1 and B levels were measured by rate nephelometry using The Array Protein System (Beckman Instruments, Brea, California). Beckman Instruments delivered the reagents, calibrators, and controls. Serum free fatty acids were analyzed using an acyl-CoA oxydase-based colorimetric kit (Wako Nefa C Kit, Osaka, Japan).
Serum for the analysis of fatty acids in serum phospholipids was prepared by clotting whole blood in a glass tube (Becton Dickinson, Meylan Cedex, France) at room temperature for 1 hour and then centrifuging it at 2000 g for 15 minutes. We transferred 1-mL aliquots of serum into sterile 2-mL cryo vials (Corning, Park Ridge, Illinois), flushed them with N2, and stored them at −70 °C until analysis. Total lipids were extracted from 500 µL of serum according to the method of Folch and coworkers [53]. Phosphatidylcholine diheptadecanoyl was added as internal standard (P-5014 Sigma Chemical Company, St. Louis, Missouri), chloroform-methanol (2:1, v/v) was added as solvent, and butylated hydroxytoluene (75 mg/L) was added as antioxidant. Total phospholipids were separated by solid-phase extraction using NH2 columns (Analytichem Bond Elut LRC, Varian, Harbor City, California) [54] followed by transmethylation using boron trifluoride, and they were evaporated to dryness. The fatty-acid methyl esters were dissolved in hexane and analyzed using a gas-liquid chromatograph (Shimadzu GC-14 A, Shimadzu, Kyoto, Japan) fitted with a capillary column (CP-Sil 88 [length, 50 m, internal diameter, 0.25 mm]; Chrompack, Raritan, New Jersey). Retention times and response factors for each fatty acid were determined using standards obtained from Nu-Chek Prep (Elysian, Minnesota). The results were integrated on a Shimadzu C-R4A integrator, and the fatty acid concentrations were reported as µmol/L of serum.
Statistical Analysis
The study was designed to have at least 85% power for detecting a 20% net difference in change in the insulin sensitivity index at a significance level of 5% (two-sided test), assuming that the standard deviation for change in the insulin sensitivity index was ± 0.01 (calculated using the euglycemic clamp technique) or ± 0.06 (using the hyperglycemic clamp technique). Change was calculated as the value obtained at the end minus the value obtained at the beginning of intervention. All data were checked with regard to frequency distribution and transformed to normal distribution by logarithmic transformation when appropriate. One-sample t-tests were used to compare values obtained before and after fish oil or corn oil supplementation, and two-sample t-tests were used for between-group comparisons. Correlations were tested by linear regression analysis and by computing the Pearson or Spearman correlation coefficients. Analysis of covariance (ANCOVA) was used for adjustments. All results are given as mean ± SE unless otherwise noted. We considered P < 0.05 to be statistically significant. The data were analyzed using the SAS software package (SAS, Cary, North Carolina) [55].
Results
Seventy-eight participants completed the study; 4 persons in the fish oil group and 2 in the corn oil group withdrew from the study. In the fish oil group, 1 person withdrew because of the development of angina pectoris; 1 because of diastolic blood pressure exceeding 110 mm Hg; and 2 for personal reasons. Both persons who withdrew from the corn oil group did so for personal reasons. Mild abdominal discomfort was reported by 6 persons in the fish oil group and 3 persons in the corn oil group. No serious side effects were seen. Compliance according to capsule count was 96.2% in the fish oil group and 97.1% in the corn oil group. Table 1 shows the baseline characteristics of these two groups. The corn oil group had a higher mean body mass index, a higher mean waist-to-hip ratio, and higher mean fasting plasma glucose and insulin levels than the fish oil group. Baseline glucose and insulin levels were not statistically different in the two groups after we adjusted for the difference in body mass index (by using ANCOVA). The mean body weight increased from 74.9 ± 2.3 kg to 76.1 ± 2.2 kg in the fish oil group (P = 0.003) and from 83.5 ± 1.7 kg to 83.8 ± 1.8 kg in the corn oil group (P > 0.3). Because of differences between the groups in body mass index and weight gain, all data were reanalyzed after adjustment for baseline body mass index and change in body weight, to test whether the changes observed during intervention would be significantly affected by these factors. This was not the case, and we present unadjusted data.
Glucose and Insulin Kinetics
Fasting plasma glucose and insulin levels did not change significantly during fish oil or corn oil treatment (Table 2). Glycosylated hemoglobin A1c levels increased slightly in both groups. The difference between the groups was not significant (Table 2). The increase from baseline was not significant after we corrected for change in body weight. The integrated glucose and insulin responses above baseline after the oral glucose tolerance test did not change after intervention (Table 2).
Target plasma glucose levels were sufficiently achieved during the hyperglycemic clamps done before (mean plasma glucose level ±SD, 10.1 ± 0.2 mmol/L, both groups combined) and after (plasma glucose level ±SD, 10.1 ± 0.02 mmol/L) treatment. The corresponding mean insulin plateaus (±SD) were 259 ± 163 pmol/L and 314 ± 202 pmol/L, respectively. During the experiments done using the euglycemic hyperinsulinemic clamp technique, mean steady-state plasma glucose levels before and after intervention (±SD) were 5.0 ± 0.2 mmol/L and 5.0 ± 0.2 mmol/L, respectively. The corresponding mean insulin plateaus (±SD) were 480 ± 141 pmol/L and 353 ± 100 pmol/L, respectively.
First-phase insulin release as measured using the hyperglycemic clamp technique did not change after treatment (Table 2). Second-phase insulin response increased significantly in both groups, but the difference between the groups was not statistically significant (Table 2). The increase in second-phase insulin release persisted after we adjusted for changes in body weight, waist-to-hip ratio, serum triglyceride level, serum free fatty acid level, insulin sensitivity, and blood pressure.
Plasma C-peptide levels did not change during fish oil or corn oil treatment. The average second-phase molar ratio of plasma C-peptide to insulin as determined using the hyperglycemic clamp technique changed from 3.6 ± 0.2 before to 2.9 ± 0.1 after fish oil treatment (P = 0.002) and from 3.3 ± 0.3 to 2.7 ± 0.2 after corn oil treatment (P = 0.02). The change was similar in both groups (P > 0.3).
The values for the insulin sensitivity index obtained using the hyperglycemic clamp technique were higher than the values calculated using the euglycemic clamp technique, but these two sets of values were highly correlated (r = 0.69; P = 0.0001). The baseline insulin sensitivity index was slightly but not significantly higher in the fish oil group than in the corn oil group before intervention. It decreased moderately during fish oil treatment without reaching the level of significance (Table 2). The insulin sensitivity index decreased significantly more in men receiving fish oil than in women receiving fish oil (−0.07 ± 0.02 compared with 0.01 ± 0.01; P = 0.02); this could not be explained by increases in body weight (men, 0.8 ± 0.5 kg; women, 1.6 ± 0.5 kg; P > 0.3). The reduction in the insulin sensitivity index in men after fish oil treatment was also significantly different from the change observed in men after corn oil treatment (−0.07 ± 0.02 compared with −0.003 ± 0.02; P = 0.03). The differences persisted after we corrected for changes in body weight. Changes in the insulin sensitivity index were identical in glucose-intolerant (−0.03 ± 0.02; n = 18) and normoglycemic persons (−0.03 ± 0.07; n = 20) receiving fish oil (P > 0.3), and the glucose-intolerant persons receiving fish oil did not differ from the glucose-intolerant persons receiving corn oil (−0.02 ± 0.01; P > 0.3).
Blood Pressure
The participants receiving fish oil had a significant reduction from baseline levels in systolic blood pressure (4.1 ± 1.4 mm Hg; P = 0.004), diastolic blood pressure (3.3 ± 0.8 mm Hg; P = 0.0008), and mean arterial pressure (3.6 ± 0.9 mm Hg; P = 0.0003) (Table 3). Thus, blood pressure in the fish oil group was statistically different (P = 0.04) from blood pressure in the control group, which did not change. The effect of n-3 fatty acids on blood pressure persisted after adjustment for between-group differences in age, sex, body mass index, waist-to-hip ratio, and change in body weight. The change in blood pressure during fish oil treatment was greater among those persons who had had a low baseline concentration of plasma phospholipid n-3 polyunsaturated fatty acids (Spearman correlation coefficient, r = 0.51; P = 0.001) (Figure 1). The correlations between change in blood pressure and baseline eicosapentaenoic acid or docosahexaenoic acid concentrations were similar.
Plasma Lipids
No changes in total cholesterol and LDL cholesterol levels were seen during treatment (Table 2). In the fish oil group, plasma triglyceride and VLDL cholesterol levels decreased nonsignificantly by 9.2% ± 5.3% (P = 0.11) and 8.9% ± 5.1% (P = 0.11), respectively. In the corn oil group, they increased significantly by 12% ± 5.9% (P = 0.04) and 12.1% ± 4.5% (P = 0.04), respectively. The difference between the two groups was significant for both triglyceride (P = 0.01) and VLDL cholesterol levels (P = 0.01) (Table 2). In the fish oil group, baseline triglyceride levels and change in triglyceride levels were inversely associated (r = −0.40; P = 0.01); this was not seen in the corn oil group (r = 0.20; P = 0.22). Levels of HDL cholesterol decreased by 7.4% ± 3% (P = 0.03) in the corn oil group; this decrease was significantly different from that in the fish oil group (P = 0.04; Table 2). There were no significant changes during the study in levels of apolipoprotein A1 [1.48 ± 0.03 g/L before and 1.47 ± 0.05 g/L after fish oil treatment; 1.44 ± 0.03 g/L before and 1.38 g/L after corn oil treatment], apolipoprotein B (1.12 ± 0.03 g/L before and 1.11 ± 0.04 g/L after fish oil treatment; 1.11 ± 0.04 g/L before and 1.05 ± 0.04 g/L after corn oil treatment), and plasma nonesterified free fatty acids (146 ± 8 mg/L before and 134 ± 8 mg/L after fish oil treatment; 137 ± 8 mg/L before and 149 ± 8 mg/L after corn oil treatment).
Plasma Concentrations of Phospholipid Fatty Acids
As expected, the plasma phospholipid content of eicosapentaenoic acid and docosahexaenoic acid increased after fish oil supplementation (Table 2). In the corn oil group, the linoleic acid level increased by 140.0 ± 47 µmol/L (P = 0.001), and the oleic acid level increased by 67.8 ± 25 µmol/L (P = 0.009).
The plasma phospholipid n-3:n-6 ratio increased by 0.13 ± 0.03 (P = 0.0001) in the fish oil group and decreased by 0.07 ± 0.01 (P = 0.0003) in the corn oil group (Table 2). Subgroup analysis showed that participants with baseline n-3:n-6 ratios in the lowest quartile had the greatest reduction in mean arterial pressure (−8.2 ± 1.2 mm Hg; P = 0.003 compared with the other quartiles). The correlation coefficient (r) between the baseline n-3:n-6 ratio and the change in mean arterial pressure was 0.40 (P = 0.01).
Discussion
In our study, a daily dose of 4 g of eicosapentaenoic and docosahexaenoic acids was effective in lowering blood pressure and triglyceride levels in persons with mild hypertension but did not impair glucose tolerance or insulin secretion. No change in insulin resistance was seen in patients receiving fish oil compared with those receiving corn oil. An apparent increase in the level of glycosylated hemoglobin A1c from baseline values was no longer significant after we corrected for changes in body weight. Even though small changes in glucose excursions may have occurred over the 16-week intervention period, glucose tolerance did not change after treatment according to the oral glucose tolerance test.
Three hours of glucose stimulation (done using the hyperglycemic clamp technique) tended to accentuate the plasma insulin response during the treatment period in both groups, regardless of changes in body weight. Stimulated and unstimulated C-peptide levels were the same before and after treatment. The C-peptide:insulin ratio therefore decreased during both regimens, which suggests that a reduced hepatic clearance of insulin took place after intake of polyunsaturated fatty acids. Fasting [56] and postprandial [57] hyperinsulinemia have been reported to be independent risk factors for the development of cardiovascular disease in middle-aged men. However, the small increase seen in second-phase plasma insulin levels after fish oil treatment occurred during a 3-hour period when the prevailing plasma glucose level was 10 mmol/L. The finding is probably not of clinical importance. Moreover, the integrated insulin response after an oral glucose tolerance test did not increase significantly.
The insulin sensitivity indexes determined by the two techniques correlated significantly (r = 69; P = 0.0001; n = 31). This confirms that the hyperglycemic clamp technique can be used to measure insulin sensitivity in hypertensive persons.
Our finding that fish oil does not alter insulin sensitivity agrees with the results of other studies in which the glucose clamp techniques were used [26, 30, 34]. Eighteen persons in the fish oil group and 23 persons in the corn oil group were glucose intolerant according to World Health Organization criteria [58]. Subgroup analysis did not show any deterioration of glycemic control or impairment in insulin release and action in these persons. On the other hand, in another subgroup analysis, men rather than women tended to become more resistant to insulin after fish oil supplementation. A further analysis showed that those men whose insulin sensitivity indexes were reduced by more than 0.01 (12 out of 21) were those who had been the most sensitive to insulin at the beginning (baseline insulin sensitivity index, 0.31 ± 0.04 compared with 0.11 ± 0.03; P = 0.0004). They tended to have a lower body mass index (25.6 ± 0.71 kg/m2 compared with 27.7 ± 0.88 kg/m2; P = 0.07) and lower fasting plasma glucose levels (5.3 ± 0.1 mmol/L compared with 6.0 ± 0.33 mmol/L; P = 0.09), whereas weight gain during treatment was similar. This may indicate that subgroups can react differently to fish oil treatment, but these findings should be interpreted with great caution because they are the result of a subgroup analysis done on a small number of persons. Further studies are required to address this question in detail. Nevertheless, because the apparent decrease in the insulin sensitivity index occurred in those persons with the lowest risk for developing clinically significant insulin resistance, we feel that it is safe to use n-3 polyunsaturated fatty acids in this group as well.
Until now, most studies on the effects of fish oil in humans were done in persons with non–insulin-dependent diabetes mellitus. Concern has been raised that n-3 polyunsaturated fatty acids could have detrimental effects on glucose homeostasis [26-31] because n-3 polyunsaturated fatty acids inhibit hepatic triglyceride and VLDL cholesterol synthesis [21, 59-61], and the impaired triglyceride synthesis may divert substrates from lipogenesis to gluconeogenesis [62]. In addition, fish oil supplementation has been reported to reduce glucose-stimulated insulin release [26, 28, 29], possibly because n-3 polyunsaturated fatty acids may inhibit the arachidonic acid 5-lipoxygenase pathway for insulin secretion [63-65]. Although we did not observe any impairment in β-cell function or glucose regulation after fish oil treatment, we must emphasize that our study was done in nondiabetic persons and that our data cannot be extrapolated directly to persons with diabetes.
Our results confirm that dietary supplementation with 4 g of eicosapentaenoic and docosahexaenoic acids per day lowers blood pressure in persons with stable, untreated, mild hypertension. As previously shown [8], the greatest decrease in blood pressure occurred in those persons with the lowest baseline plasma concentrations of phospholipid eicosapentaenoic and docosahexaenoic acids. This may indicate that an initially high acceptor-pool for polyunsaturated fatty acids is important for the replacement of n-6 fatty acids by n-3 fatty acids during fish oil intake. Incorporation of n-3 fatty acids in biomembranes may favor the production of the less vasoconstrictive thromboxane A3 rather than the arachidonic acid-derived thromboxane A2[9, 16, 66] or may interfere with the arterial baroreceptor control of vascular resistance [67], and it may thereby contribute to the reduction in blood pressure.
As in previous studies [19], those persons with the highest plasma triglyceride levels had the most pronounced decreases in triglyceride levels after fish oil treatment. The mean pretreatment triglyceride level in the group that received fish oil was only 1.19 ± 0.08 mmol/L, and this may explain why we saw only a modest triglyceride-lowering effect in the group as a whole.
In conclusion, dietary supplementation with n-3 polyunsaturated fatty acids, 4 g/d, reduces blood pressure and triglyceride levels in hypertensive persons without adversely altering glucose metabolism.
- Copyright ©2004 by the American College of Physicians
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.↵
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
Related Clinical Content
