Table 2

Randomized controlled trials examining the safety of omega-3 fatty acids.

Number of Patients

Discontinued Treatment

GI Side Effects

Bleeding

Cancer


Study

Fish Oil

Ctrl

Fish Oil

Ctrl

Fish Oil

Ctrl

Fish Oil

Ctrl

Fish Oil

Ctrl


Borchgrevink [24]

100

100

9

9

7

7

NR

NR

0††

1††

Dehmer [31]

43

39

0

NR

7

3

3

1

NR

NR

Grigg [29]

52

56

2

NR

2

1

0

0

NR

NR

Milner [32]

95

99

11

0

24

NR

0

NR

NR

NR

Reis [30] *

124*

51*

NR

NR

59

11

4

0

NR

NR

Nye [39]

36

37

NR

NR

NR

NR

NR

NR

NR

NR

Bairati [33]

59

60

NR

NR

NR

NR

0

0

NR

NR

Bellamy [28]

60

60

3

0

5

0

NR

NR

NR

NR

Kaul [27]

58

49

NR

NR

2

0

0

0

NR

NR

Franzen [34]§

103§

101§

6

7

14

10

0

0

NR

NR

Leaf [25]

226

221

2

7

16

18

7

7

NR

NR

Sacks [37]

41

39

6

1

NR

NR

0

0

NR

NR

Cairns [26]

325

328

NR

NR

122

101

17

38

NR

NR

Eritsland [40]

317

293

41

25

40

30

5

4

NR

NR

Rossing [41]

18

18

4

3

3

1

0

0

1

1

Johansen [42]||

196||

192||

NR

NR

3

2

NR

NR

NR

NR

von Schacky [36]

111

112

14

15

4

3

0

0

0

0

Nilsen [43]

150

150

NR

NR

NR

NR

NR

NR

2**

0**

Durrington [56]

30

29

1

1

1

1

0

0

0

0

Marchioli [44]

5666

5658

1616

687

278

NR

NR

NR

142

134

Maresta [35]††

146††

141††

NR

NR

2

2

6

3

NR

NR

Burr [45]

1571

1543

NR

NR

NR

NR

NR

NR

51**

47**

Calò [46]

79

81

NR

NR

NR

NR

1

1

NR

NR

Leaf [47]

200

202

73

69

NR

NR

NR

NR

NR

NR

Raitt [48]

100

100

17

26

11

12

NR

NR

3

4

Brouwer [49]

273

273

23

18

17

12

NR

NR

4

4

Yokoyama [50]‡‡

9326

9319

1087

673

352

155

105

60

242

218

GISSI-HF[55]

3494

3481

102

104

96

92

NR

NR

NR

NR

OMEGA-Trial [57]

1919

1885

NR

NR

NR

NR

35

34

NR

NR

Relative Risk (95% CrI)

1.23 (0.87, 1.68)

1.29 (1.02, 1.61)

0.85 (0.52, 1.27)

1.02 (0.73, 1.25)


Abbreviations: Ctrl = control; GI = gastrointestinal; NR = not reported.

* Reis et al. [30] randomized 222 patients, 204 of which underwent angioplasty. Of these patients, 14 patients had unsuccessful angioplasties and 4 other patients underwent bypass surgery following their angioplasty. These 18 patients were excluded from further study. The remaining 186 patients were included in the study, including 124 patients in the fish oil group (60 randomized to ethyl ester fish oil, 64 randomized to triglyceride fish oil) and 62 randomized to placebo. Clinical event data were available for all patients however 11 patients in the placebo group did not provide side effect data. Thus, side effect analyses are based on 124 fish oil patients and 51 placebo patients.

The control group involved a total of 60 patients, 53 of which underwent follow-up angiography and 7 of which who did not. Consequently, the sample size of the control group is 53 patients for baseline and restenosis data but 60 patients for all other analyses.

The study by Bellamy et al. [28] did not involve the use of placebo and thus, treatment discontinuation data were not presented for the control group. The study by Marchioli et al. [44] was a factorial design in which the control group was comprised of 2,830 patients randomized to vitamin E and 2,828 patients randomized to control. Discontinuation data were not available for those randomized to control.

§In the study by Franzen and colleagues [34], 204 patients were randomized to fish oil (n = 103) or placebo (n = 101). However, only 175 patients underwent angiographic follow-up (92 fish oil and 83 placebo). Consequently, the sample size for restenosis should be a total of 175 patients whereas the sample size for all other analyses should be 204.

Sacks and colleagues [37] randomized 80 patients to fish oil (n = 41) or control (n = 39). A total of 21 patients were excluded from the study prior to undergoing their follow-up angiogram. Although clinical events were assessed in all 80 randomized patients, restenosis was only assessed in 59 patients (31 fish oil, 28 controls).

||Johansen and colleagues [42] randomized a total of 500 patients to omega-3 fatty acid (n = 250) or placebo (n = 250). Of these patients, 5 were disqualified before their angioplasty (4 omega-3 patients and 1 placebo patient), 103 were disqualified during their angioplasty (n = 49 and 54, respectively), and 4 patients were excluded because they died prior to their follow-up angiogram (1 omega-3 patient and 3 patients randomized to placebo). There were therefore 196 patients randomized to omega-3 fatty acids and 192 patients randomized to placebo who were ultimately included in the study. The cardiac death analyses are based on 197 patients randomized to omega-3 fatty acids and 195 patients randomized to placebo whereas all other analyses are based on 196 and 192 patients, respectively.

**Data obtained from previous Cochrane review [19].

†† Maresta et al. [35] randomized 339 patients (169 to fish oil, 170 to placebo). Of these patients, 52 were disqualified before or at their angioplasty, leaving 287 patients (146 fish oil, 141 placebo). Side effect and clinical event data are based on these patients. An additional 23 patients were unevaluable drop-outs, leaving 264 patients (129 fish oil, 135 placebo) with 6-month angiograms available, and 7 additional patients were excluded because quantitative coronary angiography was not possible, resulting in a total of 257 patients (125 fish oil, 132 placebo) being included in the restenosis analyses.

‡‡The JELIS Trial [50] (conducted by Yokoyama) was performed on a population with low, moderate and high risk cardiovascular patients. Due to the lack of data, cardiovascular instead of all-cause mortality was used for the secondary prevention group. Demographic and cardiovascular outcome data are presented for the high-risk group (n = 3,664) whereas safety data are presented for the entire study population (n = 18,645). The JELIS cardiovascular data were included as part of our pooled analysis. However, due to the heterogeneity of the entire study population, data from the JELIS trial were not included in our pooled analysis of safety data.

Filion et al. BMC Cardiovascular Disorders 2010 10:24   doi:10.1186/1471-2261-10-24

Open Data