Table A2.
GRADE Evidence and Summary of Findings (SoF) Table.
| Certainty Assessment | No. of Patients | Effect | Certainty | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| No. of Studies | Study Design | Risk of Bias | Inconsistancy | Indirectness | Imprecision | Other Considerations | Experimental | Control | Relative (95% CI) | |
| Major Cardiovascular Disease | ||||||||||
| 10 | Randomized trials | Not serious a | Not serious | Serious c | Not serious | Publicaiton bias strongly suspected d | 3576/60,337 (5.9%) | 3457/60,499 (5.7%) | RR 1.03 (0.99 to 1.08) |
⨁⨁◯◯ Low |
| Myocardial Infarction | ||||||||||
| 7 | Randomized trials | Not serious a | Not serious | Serious c | Not serious | None | 2190/46,643 (4.7%) | 2212/46,629 (4.7%) | RR 0.99 (0.93 to 1.05) |
⨁⨁⨁◯ Moderate |
| Stroke | ||||||||||
| 6 | Randomized trials | Not serious a | Serious b | Serious c | Not serious | None | 1461/52,890 (2.8%) | 1345/52,885 (2.5%) | RR 1.17 (1.00 to 1.37) |
⨁⨁◯◯ Low |
| Other CVD | ||||||||||
| 4 | Randomized trials | Not serious a | Not serious | Serious c | Not serious | None | 845/10,291 (8.2%) | 799/10,276 (7.8%) | RR 1.06 (0.95 to 1.18) |
⨁⨁⨁◯ Moderate |
| CVD Mortality | ||||||||||
| 12 | Randomized trials | Not serious a | Not serious | Not serious | Not serious | None | 2761/76,244 (3.6%) | 2468/75,396 (3.3%) | RR 1.12 (1.04 to 1.19) |
⨁⨁⨁⨁ High |
| All-cause Mortality | ||||||||||
| 6 | Randomized trials | Not serious a | Serious b | Serious c | Not serious | None | 3516/41,836 (8.4%) | 3193/41,105 (7.8%) | RR 1.08 (1.00 to 1.16) |
⨁⨁◯◯ Low |
| Other Mortality | ||||||||||
| 3 | Randomized trials | Not serious | Serious b | Serious c | Not serious | None | 1144/29,200 (3.9%) | 907/28,325 (3.2%) | RR 1.23 (0.98 to 1.53) |
⨁⨁◯◯ Low |
a Several trials indicated some concerns for the incomplete accounting of patients and outcome events, such as loss to follow-up; however, we did not downgrade the grade level, as the proportion of loss to follow-up was low compared to the event rates, and there were not large differences between controlled and intervention groups. b We downgraded the grade level as the unexplained substantial heterogeneity in the results. c All the included trials in the meta-analysis were from high-income countries. The baseline nutrients status of the populations might differ from people from low- and middle-income countries. This could possibly limit the applicability of the results. d The Egger’s test (p value = 0.0071) detected potential publication bias in the meta-analysis.