TABLE 4.
Meta-analysis of observation studies reporting the effect of apple compared with low dose on CVD risk factors and events1
| Studies (subjects), n | Analysis 1 High dose | Analysis 1 Net change (95% CI) | Analysis 2 High dose | Analysis 2 Net change (95% CI) | Analysis 3 High dose | Analysis 3 Net change (95% CI) | Analysis 4 High dose | Analysis 4 Net change (95% CI) | |
|---|---|---|---|---|---|---|---|---|---|
| Cerebrovascular disease/total stroke | 3 | Knekt 2000, Hansen 2017 (M ≥ 54 g, W ≥ 71 g); Larsson (90 g) | 0.89 (0.83, 0.95) | Knekt 2000, Hansen 2017 (M ≥ 54 g; W ≥ 71 g); Larsson (180 g) | 0.89 (0.83, 0.95) | ||||
| (139, 507) | I 2 = 0.0% | I 2 = 3.5% | |||||||
| Cardiovascular death | 3 | Hertog 1993 (>110 g), Hodgson 2016 (>100 g), Mink (>20 g) | 0.86 (0.78, 0.95) | Hertog 1993 (19–110 g), Hodgson 2016 (5–100 g), Mink (>20 g) | 0.87 (0.79, 0.96) | ||||
| (36, 753) | I 2 = 0.0% | I 2 = 0.0% | |||||||
| Thrombosis or embolia or cerebral infarction | 2 | Knekt 2000 (M ≥ 54 g, W ≥ 71 g); Larsson 2013 (90 g) | 0.76 (0.55, 1.05) | Knekt 2000 (M ≥ 54 g; W ≥ 71 g); Larsson 2013 (180 g) | 0.76 (0.55, 1.06) | ||||
| (84, 169) | I 2 = 50.4% | I 2 = 52.3% | |||||||
| T2DM incidence | 4 | Alperet 2017, Lacoppidan 2015 (>56, >71), Song 2005 (≥ 180); Wedick (>128.6) | 0.86 (0.77, 0.95) | Alperet 2017 (51–77 g), Lacoppidan 2015 (>56, >71), Song 2005 (51.4–154.3 g); Wedick (51.4–102.9 g) | 0.88 (0.80, 0.96) | Alperet 2017, Lacoppidan 2015 (>56, > 71), Song 2005 (≤25.7 g); Wedick (25.7 g) | 0.92 (0.86, 0.99) | Alperet 2017, Lacoppidan 2015 (>56, > 71), Song 2005 (≤25.7 g); Wedick (6–18 g) | 0.93 (0.87, 0.99) |
| (339, 383) | I 2 = 78.0% | I 2 = 84.2% | I 2 = 74.9% | I 2 = 69.14% | |||||
| T2DM incidence (females-only) | 3 | Alperet 2017, Lacoppidan 2015 (>71), Song 2005 (≥ 80); Wedick (>128.6) | 0.81 (0.68, 0.96) | Alperet 2017 (51–77 g), Lacoppidan 2015 (>71), Song 2005 (51.4–154.3 g); Wedick (51.4–102.9 g) | 0.86 (0.74, 0.98) | Alperet 2017, Lacoppidan 2015 (>71), Song 2005 (≤25.7 g); Wedick (25.7 g) | 0.92 (0.82, 1.02) | Alperet 2017, Lacoppidan 2015 (>71), Song 2005 (≤25.7 g); Wedick (6–18 g) | 0.92 (0.83, 1.02) |
| (293, 867) | I 2 = 89.7% | I 2 = 87.9% | I 2 = 81.34% | I 2 = 78.7% | |||||
| Intracerebral hemorrhage | 3 | Knekt 2000, Hansen 2017 (M ≥ 54 g, W ≥ 71 g); Larsson 2013 (90 g) | 0.93 (0.77, 1.12) | Knekt 2000, Hansen 2017 (M ≥ 54 g, W ≥ 1 g); Larsson 2013 (180 g) | 0.92 (0.76, 1.12) | ||||
| (139, 507) | I 2 = 0.0% | I 2 = 0.0% | |||||||
| All-cause mortality | 3 | Knekt 1996 (M ≥ 54 g, W ≥ 71 g); Hodgson 2016 (>100 g); Roswell 2015 (≥ 35.2 g) | 0.85 (0.77, 0.92) | Knekt 1996 (M ≥ 54 g, W ≥ 1 g); Hodgson 2016 (5–100 g); Roswell 2015 (≥ 35.2 g) | 0.85 (0.79, 0.92) | ||||
| (51, 550) | I 2 = 0.0% | I 2 = 0.0% |
1
Meta-analyses were conducted using the random-effects model. I2 is an indicator of between-comparison heterogeneity. I2 >50% was deemed as having significant heterogeneity. T2DM, type 2 diabetes mellitus..