Fish Oil Supplementation in Type 2 Diabetes - Diabetes Care

Panagiotis Vreeker | Download | HTML Embed
  • Aug 24, 2000
  • Views: 50
  • Page(s): 9
  • Size: 106.68 kB
  • Report

Share

Transcript

1 Reviews/Commentaries/Position Statements R E V I E W A R T I C L E Fish Oil Supplementation in Type 2 Diabetes A quantitative systematic review VICTOR M. MONTORI, MD PETER C. WOLLAN, PHD Greenland (8). Despite ingesting up to 40% ANDREW FARMER, MA BMBCH SEAN F. DINNEEN, MD, MSC of calories as fat (predominantly of marine origin), this population had a lower inci- dence of coronary heart disease than indi- viduals with similar fat intake on a more conventional diet (9). Further evaluation OBJECTIVE To determine the effects of fish oil supplementation on lipid levels and revealed that dietary fish oil supplementa- glycemic control in patients with type 2 diabetes. tion led to improvement in hypertriglyc- eridemia in nondiabetic individuals RESEARCH DESIGN AND METHODS A comprehensive search of Medline, through lowering VLDL cholesterol syn- Embase, Lilacs, the Cochrane Clinical Trials Registry, bibliographies of relevant papers, and expert input updated through September 1998 was undertaken. All randomized placebo-con- thesis (10,11). However, concern was trolled trials were included in which fish oil supplementation was the only intervention in sub- raised in the initial nonrandomized studies jects with type 2 diabetes. Three investigators performed data extraction and quality scoring in patients with type 2 diabetes that fish oil independently with discrepancies resolved by consensus. Eighteen trials including 823 subjects supplementation was associated with a followed for a mean of 12 weeks were included. Doses of fish oil used ranged from 3 to 18 g/day. deterioration in glycemic control (6,7). The outcomes studied were glycemic control and lipid levels. This concern continues to be mentioned in narrative reviews on the subject (12,13). RESULTS Meta-analysis of pooled data demonstrated a statistically significant effect of fish The aim of the present study was to per- oil on lowering triglycerides (0.56 mmol/l [95% CI 0.71 to 0.41]) and raising LDL cho- form a systematic review of randomized lesterol (0.21 mmol/l [0.02 to 0.41]). No statistically significant effect was observed for fast- controlled trials addressing the effects of ing glucose, HbA1c, total cholesterol, or HDL cholesterol. The triglyceride-lowering effect and the elevation in LDL cholesterol were most marked in those trials that recruited hypertriglyc- fish oil supplementation in patients with eridemic subjects and used higher doses of fish oil. Heterogeneity was observed and explained type 2 diabetes. We were specifically inter- by the recruitment of subjects with baseline hypertriglyceridemia in some studies. ested in the effects of fish oil on lipid levels and glycemic control. CONCLUSIONS Fish oil supplementation in type 2 diabetes lowers triglycerides, raises LDL cholesterol, and has no statistically significant effect on glycemic control. Trials with hard clinical end points are needed. RESEARCH DESIGN AND METHODS Diabetes Care 23:14071415, 2000 Identification and retrieval of primary studies he typical dyslipidemia associated We conducted an electronic literature search T dietary approaches focus on restriction of with type 2 diabetes is a combination saturated fat and limitation of simple car- from 1966 to September 1998 in Medline, of hypertriglyceridemia, low levels of bohydrate and alcohol intake (2). During Embase, Lilacs, Science Citation Index, HDL cholesterol, and abnormal LDL com- the late 1980s, several investigators reported and the Cochrane Controlled Trials Regis- position (1). Several pharmacological on the use of dietary supplementation with ter using a protocol that included the approaches have been used to treat diabetic fish oil as a means of treating diabetic dys- Cochrane Collaborations search strategy dyslipidemia (2). These include use of 3- lipidemia (6,7). for randomized controlled trials (14) and hydroxy 3-methylglutaryl CoA (HMG- The potential role of fish oil in cardio- the following terms: diabetes mellitus, type CoA) reductase inhibitors (3), fibric acid vascular disease risk reduction first came 2 diabetes, non-insulin dependent diabetes derivatives (4), and niacin (5). Standard from observations involving Inuits in (NIDDM), fish oil, n-3 fatty acids, omega- 3 fatty acids, eicosapentaenoic acid (EPA), and docosahexaenoic acid (DHA). Hand From the Mayo Clinic and Foundation (V.M.M., S.F.D.); the Department of Research and Education (P.C.W.), searching of the bibliographic sections of Olmsted Medical Center, Rochester, Minnesota; and the Health Services Research Unit (A.F.), University of Oxford, Oxford, United Kingdom. all relevant articles was undertaken. Address correspondence and reprint requests to Sean F. Dinneen, MD, MSc, Department of Diabetes and Experts were consulted regarding unpub- Endocrinology, Addenbrookes Hospital, Clinic 12, Box 49, Hills Road, Cambridge CB2 2QQ, U.K. E-mail: lished or ongoing studies. Studies were [email protected] included if they were randomized placebo- Received for publication 17 December 1999 and accepted in revised form 3 May 2000. controlled trials that used fish oil supple- Abbreviations: DHA, docosahexaenoic acid; EPA, eicosapentaenoic acid; HMG-CoA, 3-hydroxy 3-methyl- glutaryl CoA. mentation as the only intervention in A table elsewhere in this issue shows conventional and Systme International (SI) units and conversion subjects with type 2 diabetes. Language factors for many other substances. was not an exclusion criterion. DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000 1407

2 Fish oil in type 2 diabetes Outcome measures and data Study (64) was excluded because it is a The dose of fish oil ranged from 3 to extraction nonrandomized nonplacebo-controlled 18 g (1.085.2 g EPA and 0.34.8 g DHA). Study selection, data extraction, and assign- addition to the original study (65). Eighteen The fish oil was usually given in capsules ing of a quality score were undertaken trials met inclusion criteria and were with the exception of 1 study, in which a independently by 3 investigators with dis- included in the review (6582). The effect liquid form was used (70). The dose of crepancies resolved by consensus. When of fish oil on glycemic control and lipid lev- placebo was matched to the dose of fish data were not available in a published els was the focus of 17 of the included tri- oil. The placebo used was a vegetable oil report, efforts were made to contact the pri- als. One study was designed to assess the with the exception of 1 study that used a mary investigators. Data related to the effect of fish oil supplementation on vascu- saline solution (70). effects of fish oil and placebo on glycemic lar physiology; however, these investigators and lipid outcomes were extracted from also reported glycemic and lipid end points Data synthesis each trial. One study reported total glyco- (73). Characteristics of the included studies Glycemic control. Of the 18 trials sylated hemoglobin; a formula was used to are shown in Table 1. included in the review, only 12 reported convert data to HbA1c (15). Study quality their fasting glucose results in a way that was assessed using the score developed by Study characteristics permitted pooling of data (Fig. 1). The Jadad et al. (16), which has a possible range The 18 trials included 7 parallel group pooled weighted mean difference for fast- of 05, with a cutoff of 2 used to designate design and 11 crossover design studies. The ing glucose was 0.26 mmol/l (95% CI studies of high versus low quality. parallel group trials ranged in duration from 0.08 to 0.60). Fifteen trials reported gly- 3 to 24 weeks. The crossover studies had cosylated hemoglobin data, and 11 of these Data analysis and statistical phases that ranged in duration from 2 to 24 were amenable to meta-analysis. The methods weeks. None of the 11 crossover studies pooled weighted mean difference was Extracted data were analyzed using the reported phase-specific data. Four studies 0.15% (0.08 to 0.37). Of the 11 Review Manager 3.1 statistical software had a washout period (38 weeks in dura- crossover studies, 8 reported glycosylated developed by the Cochrane Collaboration. tion) and 1 of these looked for but did not hemoglobin, and 5 had phase duration of A random-effects model was used to pool find a carryover effect (77). Of the 7 studies 8 weeks. data. Effect sizes are presented as weighted that did not have a washout period, 5 looked Lipid profile. Fourteen trials reported data mean differences with 95% CIs. Hetero- for and 2 found a carryover effect (66,72). on triglycerides (Fig. 2). The pooled geneity was assessed using the 2 method. The interrater reliability for the assign- weighted mean difference was 0.56 Publication bias was evaluated using a fun- ment of a quality score was substantial ( = mmol/l (0.71 to 0.41). This effect was nel plot method (17). The method of 0.77). The studies could be classified by most marked in studies that recruited only Rosenthal (18) was used to estimate the their quality scores into 8 studies of low hypertriglyceridemic subjects; the pooled number of unpublished studies with zero- quality (2 points) (6971,76,77,79, weighted mean difference was 0.73 effect that would be required to change a 80,82) and 10 studies of high quality (2 mmol/l (0.95 to 0.51). When studies significant result from the meta-analysis points) (6568,7275,78,81). Because ran- that used the higher doses of fish oil were into a nonsignificant result. Sensitivity domization was an inclusion criterion, all analyzed, the pooled weighted mean dif- analyses were planned a priori and under- studies started with a score of 1. An addi- ference was 0.85 mmol/l (1.44 to taken for the following variables: study tional point was usually assigned for the 0.26). Thirteen trials reported data on design, number of subjects, length of inter- presence of blinding. Most of the articles of total cholesterol. The pooled weighted vention, dose of fish oil, baseline triglyc- low quality failed to describe the method of mean difference was 0.007 mmol/l (0.13 eride level, and study quality. randomization and/or blinding. Some to 0.15). Ten trials reported data on LDL failed to mask the odor of the fish oil sup- cholesterol. The pooled weighted mean dif- RESULTS plement affecting blinding. In all of the tri- ference was 0.21 mmol/l (0.02 to 0.41). als, fish oil was added to the diet rather The increase was most marked in the stud- Search results than being a replacement for some compo- ies that administered the highest doses of We obtained 158 citations, of which 64 nent of the dietary fat intake. fish oil (0.51 mmol/l [0.180.84]) and were deemed relevant by title alone. Subse- A total of 823 subjects were included recruited subjects with baseline hyper- quent evaluation of the abstract and meth- in the 18 trials. The individual study sam- triglyceridemia (0.60 mmol/l [0.161.04]). ods sections led to exclusion of 46 studies. ple size ranged from 8 to 418. The major- Twelve trials reported data on HDL choles- Reasons for exclusion included the follow- ity of participants were men between 55 terol. The pooled weighted mean difference ing: publications that did not assess fish oil and 65 years of age. Most participants had was 0.02 mmol/l (0.010.05). supplementation (15 articles [1933]), non- type 2 diabetes of 510 years duration and Sensitivity analyses. The results for the test randomized studies (11 articles [3444]), were treated with diet or oral hypoglycemic of heterogeneity for all outcomes studied the population had patients without dia- agents. Few had diabetes-related complica- were nonsignificant (P 0.1 [Figs. 1 and betes or patients with type 1 diabetes (8 arti- tions. In 4 studies, all participants were 2]). As indicated above, most of the varia- cles [4552]), duplicate publications (7 hypertriglyceridemic (65,68,71,76). Two tion observed for the triglycerides and LDL articles [5359]), studies that did not other studies included a subset of hyper- cholesterol outcomes could be explained include a placebo arm (3 articles [6062]), triglyceridemic subjects, and these com- by the presence of 2 trials that recruited only and 1 study that did not include human prised 46% (69) and 10% (74) of all hypertriglyceridemic patients and used the subjects (63). The 12-month follow-up participants. Individual study exclusion cri- highest doses of fish oil (71,76). There was report of the Italian Multicenter Fish Oil teria are outlined in Table 1. no association between the studies design, 1408 DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000

3 Montori and Associates Table 1Characteristics of included studies Population characteristics Type 2 Subjects Mean diabetes Diabetes Diabetes Subjects [n (% age duration complications therapy Htg. withdrawn Duration Study Design male)] (years) (years) (%) (n) (%) (n) EC Interventions (weeks) Annuzi et al. Crossover 8 (100) 51 9.8 NA Diet: 1 NA 0 Renal and 1.8/1.2 g EPA/ 2 weeks per 1991 (79) Oral agents: 4 hepatic DHA vs. 10 g phase, no Insulin: 0 failure olive oil washout Axelrod et al. Parallel 20 (NA) 56 7.8 Microvascular (22), Diet: 4 NA 2 Bleeding diathesis, 1.1/1.5 EPA/ 6 weeks 1994 (78) Macrovascular Oral agents: 1 (1 colon anemia, poorly DHA vs. 5 g (22) Insulin: 4 cancer, 1 controlled diabetes, safflower oil noncompliance) proliferative retinopathy, use of ASA, NSAID, or steroids Boberg et al. Crossover 14 (86) 65 NA CAD (7) Diet: 1 NA 0 Renal and 1.8/1.2 g EPA/ 8 weeks per 1992 (66) Oral agents: 13 hepatic DHA vs. 10 g phase, no Insulin: 0 failure, olive oil washout hypothyroidism Borkman et Crossover 10 (70) 57 3.5 CAD (10) Diet: 6 NA 0 Renal or 1.8/1.2 g EPA/ 3 weeks per al. 1989 (77) Oral agents: 4 hepatic DHA vs. 10 g phase, Insulin: 0 failure, safflower oil 3-week microvascular washout Connor et al. Crossover 16 (81) 58 NA NA Diet: 1 100 0 NA 4.1/1.9 g EPA/ 24 weeks per 1993 (76) Oral agents: 10 DHA vs. 15 g phase, no Insulin: 5 olive oil washout Goh et al. Crossover 28 (NA) 57 8 CAD (0) NA NA 0 Heart disease, 1.4/0.88 g EPA/ 12 weeks per 1997 (75)* lipid-lowering DHA vs. 35 mg/kg phase, no agent linseed oil washout Hendra et al. Parallel 80 (75) 56 6.7 Microvascular in Diet: 8 0 0 Pregnancy, oral 1.8/1.2 g EPA/ 6 weeks 1990 (81) control group Oral agents: 32 contraceptives, DHA vs. 10 g (42.5), in fish Insulin: 0 dyslipidemia, olive oil oil (70); CAD recent myocardial in control infarction group (35), in or stroke fish oil (7.5) Luo et al. Crossover 12 (100) 54 6 NA Diet: 2 10 2 Hepatic or 1.08/0.72 g EPA/ 9 weeks per 1998 (74) Oral agents: 8 (protocol renal failure DHA vs. 6 g phase, Insulin: 0 violations) thyroid or sunflower oil 9-week gastrointestinal disorders washout McGrath et Crossover 23 (87) 53 NA Macrovascular Diet: NA NA 0 Renal failure, 1.8/1.2 g EPA/ 6 weeks per al. 1996 (73) (0) Oral agents: NA stroke, cardio- DHA vs. 10 g phase, Insulin: 0 vascular disease, olive oil 6-week hypertension, washout cardiovascular drugs, lipid- lowering agents or vitamins McManus et Crossover 11 (73) 62 7.7 NA Diet: 7 NA 0 No insulin use or 1.8/1.2 g EPA/ 3 weeks, al. 1996 (72) Oral agents: 4 lipid-lowering DHA vs. 35 mg/kg no Insulin: 0 agents linseed oil washout Morgan et al. Parallel 40 (50) 54 9.8 NA Diet: 2 100 0 NA Low dose: 12 weeks 1995 (71) Oral agents: 2 2.6/2.4 Insulin: 16 High dose: 5.2/4.8 vs. 9 or 18 g corn oil Pelikanova et Parallel 20 (100) 51 NA NA Diet: 0 0 0 Obesity, Htg., 15 ml (3 g) 3 weeks al. 1993 (70) Oral agents: 10 renal or hepatic fish oil vs. Insulin: 0 failure 15 ml saline continued on page 1410 DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000 1409

4 Fish oil in type 2 diabetes Table 1Continued Population characteristics Type 2 Subjects Mean diabetes Diabetes Diabetes Subjects [n (% age duration complications therapy Htg. withdrawn Duration Study Design male)] (years) (years) (%) (n) (%) (n) EC Interventions (weeks) Puhakainen et Crossover 9 (44) 53 13 0 Diet: 2 NA 0 Macro-/ 2.16/1.44 g EPA/ 6 weeks per al. 1995 (82) Oral agents: 7 microvascular DHA vs. 6 g phase, no Insulin: 0 complications, corn 6 g washout hepatic or olive oil renal failure, bleeding diathesis, insulin requirement hypothyroidism Schectman et Crossover 13 (69) 52 NA Macrovascular Diet: 2 46 0 Hepatic failure, 2.6/1.4 g EPA/ 4 weeks per al. 1988 (69) (15) Oral agents: 9 renal failure DHA vs. 12 g phase, Insulin: 2 hypothyroidism, safflower oil 4-week poorly controlled washout diabetes, lipid- lowering agents Silvis et Parallel 63 (46) 55 4.8 NA Diet: 0 NA 0 NA 1.4/0.3 g EPA/ 8 weeks per al. 1990 (80) Oral agents: 11 (7 not included DHA vs. 12 g phase, Insulin: 13 in analysis: olive oil 8-week noncompliance washout Sirtori et Parallel 418 (62) 58 5 NA Diet: NA 100 4 Obesity, 1.14/0.8 g EPA/ 24 weeks al. 1997 (65) Oral: NA (from malabsorption, DHA vs. 3 g Insulin: 0 treatment duodenal ulcer, olive oil arm; volunteers noncompliant decision) or unreliable subject, epilepsy, alcoholism, insulin use, history of unstable angina or recent myo- cardial infarction, severe hyperten- sion or severe dyslipidemia Vessby and Crossover 14 (78) 3772 NA NA NA 100 0 Lipid-lowering 1.8/1.2 g EPA/ 8 weeks per Boberg 1990 agents DHA vs. 10 g phase, no (68) olive oil washout Westerveld et Parallel 24 (62.5) 56 6.5 NA Diet: 4 NA 0 Hepatic or 1.8 g EPA 8 weeks al. 1993 (67) Oral agents: 3 renal failure, vs. 1.6 g Insulin: 0 bleeding diathesis, olive oil no cardiovascular disorder in last 3 months, no insulin use *Subjects were classified as a low or a high P/S diet groups; groups were randomized separately. The intervention group had a larger weight; they were divided into 4 groups: 2 doses of fish oil and 2 doses of placebo (10 per group). Obtained from averaging the doses used: 1.5/1 g EPA/DHA for 2 months, then 1/0.7 g EPA/DHA for 6 months. ASA, acetyl salicylic acid; EC, exclusion criteria; Htg., hypertriglyceridemia; n, number randomized; NA, not available; NSAID, nonsteroidal anti-inflam- matory drug. duration, or quality and the direction or tence of such studies. Approximately 253 randomized controlled trials of fish oil sup- magnitude of the outcomes studied. unpublished zero-effect studies would be plementation studying 800 subjects with Publication bias. According to the funnel needed to change the magnitude or direc- type 2 diabetes. In the studies reviewed, fish plot analysis, small studies showing small tion of the effect observed for triglycerides. oil supplementation had a statistically sig- triglyceride-lowering treatment effect or no nificant triglyceride-lowering effect. This net effect are missing from this analysis as a CONCLUSIONS This systematic effect was most marked in studies that result of either publication bias or nonexis- review pools 10 years of evidence and 18 recruited hypertriglyceridemic subjects. 1410 DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000

5 Montori and Associates Fish oil supplementation did not result in any statistically or clinically significant increase in fasting glucose or HbA1c. A sta- tistically significant increase in LDL choles- terol was especially noted in the studies recruiting hypertriglyceridemic subjects and using the highest doses of fish oil. None of the trials examined hard clinical end points such as cardiovascular events or death. Several methodological challenges were encountered in the course of this review. Eleven of 18 trials used a crossover design, and phase-specific data were not available for any of these. There is no accepted method for pooling results from crossover and parallel group design studies. Ideally, individual patient data or at least phase-specific data should be available. In the absence of these data, 3 approaches are possible. The first approach is to not ana- lyze data from crossover studies. The sec- ond is to pool parallel group design and crossover trials separately. The third is to treat data from crossover studies as data coming from parallel group design studies, pool these with data from parallel group design studies, and look for heterogeneity in the analysis. We adopted the latter approach, and our sensitivity analysis did not show any association between study design and direction or magnitude of effect. Use of the crossover design to study fish oil supplementation has other potential draw- backs. Fish oil is incorporated into bio- logic membranes and presumably would require washout periods of appropriate duration to minimize any carryover effect. In our review, only 4 of the 11 crossover studies had a washout period. Despite these limitations, the main findings of the review were similar if crossover studies Figure 1Effect of fish oil supplementation on glycemic control in subjects with type 2 diabetes. Data were included or excluded from the analy- shown represent the weighted mean difference () with 95% CI () using a random effects model. The sis. HbA1c provides an integrated measure overall weighted mean difference () for fasting glucose was 0.26 mmol/l (95% CI 0.08 to 0.60) and of glycemic control over a period of 12 for HbA1c was 0.15% (0.08 to 0.37). These results indicate that fish oil supplementation was associ- weeks. The use of such measurements in ated with a nonsignificant increase in fasting blood glucose and HbA1c. The results for the test of het- studies of short duration will underesti- erogeneity among studies were not significant (P = 0.99 and 0.78, respectively). mate any effects on glycemic control. This may have occurred in several trials included in this review (Table 1). Sensitiv- by Friedberg et al. includes data up to June together, these similar results speak to the ity analysis showed that the study duration 1995. The large Italian Fish Oil Multicenter robustness of the findings that fish oil lowers did not affect the pooled HbA1c estimate. Study, which reported the effects of fish oil in triglycerides without adversely affecting It is interesting to compare the current 418 patients with type 2 diabetes, is glycemic control. systematic review with that of Friedberg et included in our review (65). Our methods Our data are relevant to clinicians man- al. (44). Our review was limited to random- included an assessment for heterogeneity; aging patients with type 2 diabetes. They ized trials involving patients with type 2 the study by Friedberg et al. did not attempt indicate that, in normotriglyceridemic diabetes, whereas the earlier review included to understand or explain heterogeneity. patients, dietary supplementation with fish studies of varying designs and patients with Despite these differences in design, the find- oil leads to a modest lowering of triglyc- both type 1 and type 2 diabetes. Our review ings of the 2 reviews are similar, and both erides without any clinically significant is more current and includes published data findings are keeping with the results of the effect on glycemic control. It is unlikely that up to September 1998, whereas the review largest trial performed in this area. Taken fish oil will be prescribed in normotriglyc- DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000 1411

6 Fish oil in type 2 diabetes eridemic patients, but their use as over-the- counter preparations should not be dis- couraged, provided they are obtained from a reputable source. This recommendation is based partly on the fact that fish oil sup- plementation has been shown to have ben- eficial effects on other biologic systems (e.g., immune function [83]). Fish oil rep- resents a reasonable therapeutic strategy in hypertriglyceridemic individuals. Very few studies have compared fish oil with fibric acid derivatives (62), and we are not aware of any studies assessing the combination of fish oil with other lipid-lowering drugs. The slight increase in LDL cholesterol seen with the use of fish oil can occur with other triglyceride-lowering agents. The mecha- nism of the LDL increase with fish oil has recently been elucidated (84). In addition, large buoyant LDL is known to be less atherogenic than small dense LDL, and this may be the type of LDL produced in response to fish oil (85). Since completion of our review, the Gruppo Italiano per lo Studio delia Sopravvivenza nellinfarto mio- cardico-Prevenzione Trial has published its findings on the administration of fish oil to 11,324 survivors of myocardial infarction (86). Although the analysis for the diabetes subgroup (15% of participants) has not yet been reported, the findings of reduced triglycerides, increased LDL cholesterol, and an overall beneficial effect on survival (relative risk reduction of 10% for the pri- mary end point of death, nonfatal myocar- dial infarction, and stroke) are encouraging. We believe that further long-term studies assessing hard cardiovascular end points in patients with diabetes are needed. References 1. Howard BV: Lipoprotein metabolism in diabetes mellitus. J Lipid Res 28:613628, 1987 2. American Diabetes Association: Manage- ment of dyslipidemia in adults with dia- betes (Position Statement). Diabetes Care 21 (Suppl. 1):S36S39, 1998 3. Pyorala K, Pedersen TR, Kjekshus J, Faerge- man O, Olsson AG, Thorgeirsson G: Cho- lesterol lowering with simvastatin improves prognosis of diabetic patients with coronary Figure 2Effect of fish oil supplementation on the lipid profile of subjects with type 2 diabetes. Data heart disease: a subgroup analysis of the shown represent the weighted mean difference () with 95% CI () using a random effects model. Scandinavian Simvastatin Survival Study Arrowheads represent CIs beyond the scale. The overall weighted mean difference () for total cho- (4S). Diabetes Care 20:614620, 1997 lesterol was 0.007 mmol/l (95% CI 0.13 to 0.15), for LDL cholesterol was 0.21 mmol/l (0.02 to 0.41), 4. Elkeles RS, Diamond JR, Poulter C, Dhan- and for triglycerides was 0.56 mmol/l (0.71 to 0.41). These results indicate that fish oil supple- jil S, Nicolaides AN, Mahmood S, Rich- mentation was associated with a nonsignificant change in total cholesterol levels, a significant triglyc- mond W, Mather H, Sharp P, Feher MD: eride-lowering effect, and a significant increase in LDL cholesterol levels. The results for the test of Cardiovascular outcomes in type 2 dia- heterogeneity among studies were not significant (cholesterol, P = 0.86; LDL cholesterol, P = 0.69; and betes: a double-blind placebo-controlled triglycerides, P = 0.46). study of bezafibrate: the St. Marys, Ealing, 1412 DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000

7 Montori and Associates Northwick Park Diabetes Cardiovascular serum lipids and glycemic control in betic patients. J Clin Endocrinol Metab 67:1 Disease Prevention (SENDCAP) Study. Dia- NIDDM: a randomized controlled study. 4, 1988 betes Care 21:641648, 1998 Diabetes Care 20:913921, 1997 35. Malasanos TH, Stacpoole PW: Biological 5. Garg A, Grundy SM: Nicotinic acid as ther- 21. Urano S, Hoshi-Hashizume M, Tochigi N, effects of omega-3 fatty acids in diabetes apy for dyslipidemia in non-insulin-depen- Matsuo M, Shiraki M, Ito H: Vitamin E mellitus. Diabetes Care 14:11601179, 1991 dent diabetes mellitus. JAMA 264:2994 and the susceptibility of erythrocytes and 36. Herrmann W, Biermann J, Ratzmann KP, 2996, 1990 reconstituted liposomes to oxidative stress Lindhofer HG: Effect of fish oil concentrate 6. Glauber H, Wallace P, Griver K, Brechtel G: in aged diabetics. Lipids 26:5861, 1991 on the lipoprotein profile of patients with Adverse metabolic effect of omega-3 fatty 22. Adler AI, Boyko EJ, Schraer CD, Murphy type II diabetes mellitus. Med Klin 87:1215, acids in non-insulin dependent diabetes NJ: Lower prevalence of impaired glucose 1992 mellitus. Ann Intern Med 108:663668, 1988 tolerance and diabetes associated with daily 37. Schaap GH, Bilo HJG, Beukhof JR, Gans 7. Friday KE, Childs MT, Tsunehara C, Fuji- seal oil or salmon consumption among ROB, Popp-Snijders C, Donker AJM: The moto WY, Bierman EL, Ensinck JW: Ele- Alaska Natives. Diabetes Care 17:1498 effects of short-term omega-3 polyunsatu- vated plasma glucose and lowered triglyc- 1501, 1994 rated fatty acid supplementation in patients eride levels from omega-3 fatty acid 23. Das UN, Kumar KV, Mohan IK: Lipid per- with chronic renal insufficiency. Curr Ther supplementation in type II diabetes. Dia- oxides and essential fatty acids in patients Res Clin Exp 49:10611070, 1991 betes Care 12:276281, 1989 with diabetes mellitus and diabetic 38. Semplicini A, Valle R: Fish oils and their 8. Mouratoff GJ, Carroll NV, Scott EM: Dia- nephropathy. J Nutr Med 4:149155, 1994 possible role in the treatment of cardiovas- betes mellitus in Eskimos. JAMA 199:107 24. Das UN, Kumar KV, Ramesh G: Essential cular diseases. Pharmacol Ther 61:385397, 112, 1967 fatty acid metabolism in south Indians. 1994 9. Bang HO, Dyerberg J, Hjorne N: The com- Prostaglandins Leukot Essential Fatty Acids 39. Sheehan JP, Wei IW, Ulchaker M, Tserng position of food consumed by Greenland 50:253255, 1994 KY: Effect of high fiber intake in fish oil- Eskimos. Acta Med Scand 200:6973, 1976 25. Das UN: Essential fatty acid metabolism in treated patients with non-insulin-depen- 10. Nestel PJ, Connor WE, Reardon MF, Con- patients with essential hypertension, dia- dent diabetes mellitus. Am J Clin Nutr 66: nor S, Wong S, Boston R: Suppression by betes mellitus and coronary heart disease. 11831187, 1997 diets rich in fish oil of very low density Prostaglandins Leukot Essential Fatty Acids 40. Silva JM, Souza I, Silva R, Tavares P, Teixeira lipoprotein production in man. J Clin Invest 52:387391, 1995 F, Silva PS: The triglyceride lowering effect 74:8289, 1984 26. Holler C, Auinger M, Ulberth F, Irsigler K: of fish oils is affected by fish consumption. 11. Harris WS, Connor WE, Illingworth R, Eicosanoid precursors: potential factors for Int J Cardiol 57:7580, 1996 Rothrock DW, Foster DM: Effects of fish oil atherogenesis in diabetic CAPD patients? 41. Stender S, Jensen T, Deckert T: Experience on VLDL triglyceride kinetics in humans. J Perit Dial Int 16:S250S253, 1996 with fish oil treatment with special empha- Lipid Res 31:15491558, 1990 27. Howard WJ: Is it time for a clinical trial of sis on diabetic nephropathy. J Diabetes Com- 12. Garg A: Treatment of diabetic dyslipidemia. dietary fish oil supplementation in indi- plications 4:7071, 1990 Am J Cardiol 81:47B51B, 1998 viduals with NIDDM? Ann N Y Acad Sci 42. Zak A, Zeman M, Tvrzicka E, Stolba P: 13. OBrien T, Nguyen TT, Zimmerman BR: 683:341342, 1993 Effects of fish oils in patients with type 2 Hyperlipidemia and diabetes mellitus. 28. Lee R: Fish oil, essential fatty acids, and diabetes with associated dyslipidaemia. Cas Mayo Clin Proc 73:969976, 1998 hypertension. Can J Physiol Pharmacol 72: Lek Cesk 135:354359, 1996 14. Dickersin K, Scherer R, Lefebvre C: Identi- 945953, 1994 43. Shunto S, Takahashi K, Negishi K, Suzuki fying relevant studies for systematic 29. Morris MC, Manson JE, Rosner B, Buring M, Moritani S, Itabashi A, Katayama S, Ishii reviews. BMJ 309:12861291, 1994 JE, Willett WC, Hennekens CH: Fish con- J, Kusuhara R, Wakabayashi T: Effects of 15. Nutall FQ: Comparison of percent total sumption and cardiovascular disease in the eicosapentaenoic acid on glycemic control GHb with percent HbA1c in people with Physicians Health Study: a prospective and lipid metabolism in healthy and and without known diabetes. Diabetes Care study. Am J Epidemiol 142:166175, 1995 NIDDM subjects. Seraputikku Risachi 13: 21:14751480, 1998 30. Okuda Y, Mizutani M, Tanaka K, Isaka M, 257265, 1992 16. Jadad AR, Moore RA, Carrol D, Jenkinson Yamashita K: Beneficial effects of eicos- 44. Friedberg CE, Janssen M, Heine RJ, C, Reynolds DJM, Gavaghan DJ: Assessing apentaenoic acid for diabetic patients with Grobbee DE: Fish oil and glycemic control the quality of reports of randomized clini- arteriosclerosis obliterans. Diabetes Res Clin in diabetes: a meta-analysis. Diabetes Care cal trials: is blinding necessary? Control Clin Pract 18:139140, 1992 21:494500, 1998 Trials 17:112, 1996 31. Osuka Y, Mizutani M, Ogawa M, Sone H, 45. Stacpoole PW, Alig J, Ammon L, Crockett 17. Egger M, Smith GD, Schneider M, Minder Asano M, Asakura Y, Isaka M, Suzuki S, SE: Dose response effects of dietary marine C: Bias in meta-analysis detected by a sim- Kawakami Y, Field JB, Yamashita K: Long- oil on carbohydrate and lipid metabolism ple, graphical test. BMJ 315:629634, 1997 term effects of eicosapentaenoic acid on in normal subjects and patients with hyper- 18. Rosenthal R: Meta-analytic Procedures for diabetic peripheral neuropathy and serum triglyceridemia. Metabolism 38:946956, Social Research. Beverly Hills, CA, Sage Pub- lipids in patients with type II diabetes mel- 1989 lications, 1984 litus. J Diabetes Complications 10:280287, 46. Eritsland J, Seljeflot I, Abdelnoor M, Arne- 19. Zambon S, Friday KE, Childs MT, Fuji- 1996 sen H, Torjesen PA: Long-term effects of n- moto WY, Bierman EL, Ensinck JW: Effect 32. Prince MJ, Deeg MA: Do n-3 fatty acids 3 fatty acids on serum lipids and glycaemic of glyburide and omega 3 fatty acid dietary improve glucose tolerance and lipemia in control. Scand J Clin Lab Invest 54:273280, supplements on glucose and lipid metabo- diabetics? Curr Opin Lipidol 8:711, 1997 1994 lism in patients with non-insulin-depen- 33. Tonstad S: Indications for lipid-lowering 47. Fasching P, Ratheiser K, Waldhausl W, dent diabetes mellitus. Am J Clin Nutr 56: drugs: unanswered questions. Tidsskr Nor Rohac M, Osterrode W, Nowotny P, Vier- 447454, 1992 Laegeforen 117:674677, 1997 happer H: Metabolic effects of fish-oil sup- 20. Dunstan DW, Mori TA, Puddey IB, Beilin 34. Kassim SE, Stern B, Khilnani S, McLin P, plementation in patients with impaired LJ, Burke V, Morton AR, Stanton KG: The Baciorowski S, Jen K-LC: Effects of omega- glucose tolerance. Diabetes 40:583589, independent and combined effects of aero- 3 fish oils on lipid metabolism, glycemic 1991 bic exercise and dietary fish intake on control, and blood pressure in type II dia- 48. Mackness MI, Bhatnagar D, Durrington DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000 1413

8 Fish oil in type 2 diabetes PN, Prais H, Haynes B, Morgan J, Borth- of non-insulin dependent diabetic patients. Clandinin MT, Ryan EA: A comparison of wick L: Effects of a new fish oil concentrate Diabetes Res Clin Pract 28:3540, 1995 the effects of n-3 fatty acids from linseed oil on plasma lipids and lipoproteins in 61. Shimizu H, Sato N, Tanaka Y, Kashima K, and fish oil in well-controlled type II dia- patients with hypertriglyceridaemia. Eur J Ohtani K-I, Mori M: Effect of eicosapen- betes. Diabetes Care 19:463467, 1996 Clin Nutr 48:859865, 1994 taenoic acid ethyl on urine albumin excre- 73. McGrath LT, Brennan GM, Donnelly JP, 49. Rossing P, Hansen BV, Nielsen FS, Myrup B, tion in NIDDM. Diabetes Care 16:1406 Johnston GD, Hayes JR, McVeigh GE: Effect Holmer G, Parving HH: Fish oil in diabetic 1407, 1993 of dietary fish oil supplementation on per- nephropathy. Diabetes Care 19:12141219, 62. Fasching P, Rohac M, Liener K, Schneider oxidation of serum lipids in patients with 1996 B, Nowotny P, Waldhausl W: Fish oil sup- non-insulin dependent diabetes mellitus. 50. Bonnema SJ, Jespersen LT, Marving J, plementation versus gemfibrozil treatment Atherosclerosis 121:275283, 1996 Gregersen G: Supplementation with olive in hyperlipidemic NIDDM: a randomized 74. Luo J, Rizkalla SW, Vidal H, Oppert JM, oil rather than fish oil increases small arte- crossover study. Horm Metab Res 28:230 Colas C, Boussairi A, Guerre-Millo M, Cha- rial compliance in diabetic patients. Dia- 236, 1996 puis AS, Chevalier A, Durand G, Slama G: betes Nutr Metab 8:8187, 1995 63. Yamada Y, Fushimi H, Inoue T, Matsuyama Moderate intake of n-3 fatty acids for 2 51. Hamazaki T, Takazakura E, Osawa K, Y, Kameyama M, Minami T, Okazaki Y, months has no detrimental effect on glucose Urakaze M, Yano S: Reduction in microal- Noguchi Y, Kasama T: Effect of eicosapen- metabolism and could ameliorate the lipid buminuria in diabetics by eicosapentaenoic taenoic acid and docosahexaenoic acid on profile in type 2 diabetic men: results of a acid ethyl ester. Lipids 25:541545, 1990 diabetic osteopenia. Diabetes Res Clin Pract controlled study. Diabetes Care 21:717724, 52. Lungershausen YK, Howe PR, Clifton PM, 30:3742, 1995 1998 Hughes CR, Phillips P, Graham JJ, Thomas 64. Sirtori CR, Crepaldi G, Manzato E, Mancini 75. Goh YK, Jumpsen JA, Ryan EA, Clandinin DW: Evaluation of an omega-3 fatty acid M, Rivellese A, Paoletti R, Pazzucconi F, MT: Effect of omega 3 fatty acid on plasma supplement in diabetics with microalbu- Pamparana F, Stragliotto E: One-year treat- lipids, cholesterol and lipoprotein fatty acid minuria. Ann N Y Acad Sci 827:369381, ment with ethyl esters of n-3 fatty acids in content in NIDDM patients. Diabetologia 1997 patients with hypertriglyceridemia and glu- 40:4552, 1997 53. Donnelly JP, McGrath LT, Brennan GM: cose intolerance reduced triglyceridemia, 76. Connor WE, Prince MJ, Ullmann D, Riddle Lipid peroxidation, LDL glycosylation and total cholesterol and increased HDL-C M, Hatcher L, Smith FE, Wilson D: The dietary fish oil supplementation in type II without glycemic alterations. Atherosclerosis hypotriglyceridemic effect of fish oil in diabetes mellitus. Biochem Soc Trans 22:34S, 137:419427, 1998 adult-onset diabetes without adverse glu- 1994 65. Sirtori CR, Paoletti R, Mancini M, Crepaldi cose control. Ann N Y Acad Sci 683:337 54. Rivellese AA, Maffettone A, Iovine C, Di G, Manzato E, Rivellese A, Pamparana F, 340, 1993 Marino L, Annuzzi G, Mancini M, Riccardi Stragliotto E: N-3 fatty acids do not lead to 77. Borkman M, Chisholm DJ, Furler SM, Stor- G: Long-term effects of fish oil on insulin an increased diabetic risk in patients with lien LH, Kraegen EW, Simons LA, Chester- resistance and plasma lipoproteins in hyperlipidemia and abnormal glucose tol- man CN: Effects of fish oil supplementation NIDDM patients with hypertriglyceridemia. erance: Italian Fish Oil Multicenter Study. on glucose and lipid metabolism in Diabetes Care 19:12071213, 1996 Am J Clin Nutr 65:18741881, 1997 NIDDM. Diabetes 38:13141319, 1989 55. Pelikanova T, Kohout M, Valek J, Kazdova 66. Boberg M, Pollare T, Siegbahn A, Vessby B: 78. Axelrod L, Camuso J, Williams E, Kleinman L, Karasova L, Base J, Stefka Z: Effect of fish Supplementation with n-3 fatty acids K, Briones E, Schoenfeld D: Effects of a small oil supplementation on insulin secretion reduces triglycerides but increases PAI-1 in quantity of omega-3 fatty acids on cardio- and insulin action in type II (non-insulin- non-insulin-dependent diabetes mellitus. vascular risk factors in NIDDM: a random- dependent) diabetic patients. Cas Lek Cesk Eur J Clin Invest 22:645650, 1992 ized, prospective, double-blind, controlled 131:668672, 1992 67. Westerveld HT, de Graaf JC, van Breugel study. Diabetes Care 17:3744, 1994 56. McVeigh GE, Brennan GM, Johnston GD, HH, Akkerman JW, Sixma JJ, Erkelens DW, 79. Annuzzi G, Rivellese A, Capaldo B, Di McDermott BJ, McGrath LT, Henry WR, Banga JD: Effects of low-dose EPA-E on Marino L, Iovine C, Marotta G, Riccardi G: Andrews JW, Hayes JR: Dietary fish oil aug- glycemic control, lipid profile, lipopro- A controlled study on the effects of n-3 fatty ments nitric oxide production or release in tein(a), platelet aggregation, viscosity, and acids on lipid and glucose metabolism in patients with type 2 (non-insulin-depen- platelet and vessel wall interaction in non-insulin-dependent diabetic patients. dent) diabetes mellitus. Diabetologia 36:33 NIDDM. Diabetes Care 16:683688, 1993 Atherosclerosis 87:6573, 1991 38, 1993 68. Vessby B, Boberg M: Dietary supplementa- 80. Silvis N, Vorster HH, Mollentze WF, Jager J, 57. McVeigh G, Brennan G, Hayes R, John- tion with n-3 fatty acids may impair glu- Huisman HW: Metabolic and haemostatic ston D: Primary nitrate tolerance in dia- cose homeostasis in patients with consequences of dietary fibre and N-3 fatty betes mellitus. Diabetologia 37:115117, non-insulin-dependent diabetes mellitus. J acids in black type 2 (NIDDM) diabetic 1994 Intern Med 228:165171, 1990 subjects: a placebo controlled study. Int Clin 58. McVeigh GE, Brennan GM, Cohn JN, 69. Schectman G, Kaul S, Kissebah AH: Effect Nutr Rev 10:362380, 1990 Finkelstein SM, Hayes RJ, Johnston GD: of fish oil concentrate on lipoprotein com- 81. Hendra TJ, Britton ME, Roper DR, Fish oil improves arterial compliance in position in NIDDM. Diabetes 37:1567 Wagaine-Twabwe D, Jeremy JY, Dandona P, non-insulin-dependent diabetes mellitus. 1573, 1988 Haines AP, Yudkin JS: Effects of fish oil sup- Arterioscler Thromb 14:14251429, 1425 70. Pelikanova T, Kohout M, Valek J, Kazdova L, plements in NIDDM subjects: controlled 59. Maffettone A: Long-term effects (six Base J: Metabolic effects of omega-3 fatty study. Diabetes Care 13:821829, 1990 months) of omega-3 polyunsaturated fatty acids in type 2 (non-insulin-dependent) dia- 82. Puhakainen I, Ahola I, Yki-Jarvinen H: acids on insulin sensitivity and lipid metab- betic patients. Ann N Y Acad Sci 683:272 Dietary supplementation with n-3 fatty acids olism in patients with type 2 diabetes and 278, 1993 increases gluconeogenesis from glycerol but hypertriglyceridemia. Giornale Italiano di 71. Morgan WA, Raskin P, Rosenstock J: A not hepatic glucose production in patients Diabetologia 16:185193, 1996 comparison of fish oil or corn oil supple- with non-insulin-dependent diabetes melli- 60. Shimizu H, Ohtani K, Tanaka Y, Sato N, Mori ments in hyperlipidemic subjects with tus. Am J Clin Nutr 61:121126, 1995 M, Shimomura Y: Long-term effect of eicos- NIDDM. Diabetes Care 18:8386, 1995 83. Connor WE: Importance of n-3 fatty acids apentaenoic acid ethyl (EPA-E) on albuminuria 72. McManus RM, Jumpson J, Finegood DT, in health and disease (Editorial). Am J Clin 1414 DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000

9 Montori and Associates Nutr 71:171S175S, 2000 J Lipid Res 39:388401, 1998 86. GISSI-Prevenzione Investigators: Dietary 84. Fisher WR, Zech LA, Stacpoole PW: 85. Suzukawa M, Abbey M, Howe PR, Nestel PJ: supplementation with n-3 polyunsaturated Apolipoprotein B metabolism in hyper- Effects of fish oil fatty acids on low density fatty acids and vitamin E after myocardial triglyceridemic patients administered either lipoprotein size, oxidizability, and uptake by infarction: results of the GISSI-Prevenzione a fish oil- or vegetable oil-enriched diet. macrophages. J Lipid Res 36:473484, 1995 Trial. Lancet 354:447455, 1999 DIABETES CARE, VOLUME 23, NUMBER 9, SEPTEMBER 2000 1415

Load More