Table 4.
Dairy foods, haemorrhagic stroke and sub-arachnoid bleeds
Study | Number of subjects (length of follow-up) | Number and type of disease event | Factors adjusted for | Adjusted RR (95% CI) | Predictive factor and subgroups compared |
---|---|---|---|---|---|
Kinjo et al. [55] A Japanese cohort |
223,170 subjects (16 years) | 4,773 haemorrhagic stroke deaths | Sex, age, area, smoking, alcohol, occupation | 0.74 (0.68–0.80) | Milk four or more times/week vs once/week |
He et al. [57] Health Prof. cohort |
43,732 men (14 years) | 124 haemorrhagic strokes | Smoking, alcohol, BMI, activity, hyperchol., hypertension, aspirin, potassium, multivitamins, vit E, fruit and veg., energy | 1.22 (0.47–3.16) | High-fat dairy once a day or more vs less than once a week |
Umesawa et al. [54] JACC cohort |
21,068 men 32,319 women (10 years) |
113 haemorrhagic strokes 128 haemorrhagic strokes |
Age, sex, BMI, smoking, alcohol, sodium, potassium, fatty acids, area, menopause, hyperchol., diabetes |
0.46 (0.23–0.91) 0.51 (0.28–0.94) |
Top and bottom quintile of dairy calcium intake |
Umesawa et al. [32] JPHC cohort |
41,526 subjects (12.9 years) | 425 haemorrhagic strokes | Age, sex, BMI, smoking, alcohol, sodium, potassium, fatty acids, area, menopause, hyperchol., diabetes | 0.65 (0.43–0.97) | Top quintile of dairy calcium vs lowest quintile |
Larsson et al. [23] ATBC cohort |
26,556 men (13.6 years) | 383 haemorrhagic strokes | Age, smoking, alcohol, BMI, education, total cholesterol, diabetes, IHD, energy intake and activity, intake of various foods and original randomisation |
1.01 (0.82–1.20)a 1.32 (0.89–1.94)a |
Top quintile of a composite of low-fat, whole and sour milk vs lowest quintile Top and lowest quintile of dairy products vs lowest quintile |
0.36M person years: 5,946 haemorrhagic strokes: heterogeneity between studies P = 0.014 Meta-analysis (random effects) RR (95% CI) for highest intake groups 0.75 (0.60–0.94) | |||||
Umesawa et al. [54] JACC cohort |
21,068 men 32,319 women (10 years) |
37 sub-arachnoid bleeds 34 sub-arachnoid bleeds |
Age, sex, BMI, smoking, alcohol, sodium, potassium, fatty acids, area, menopause, hyperchol., diabetes |
0.19 (0.04–0.87) 0.41 (0.17–0.97) |
Top and bottom quintile of dairy calcium intake |
Umesawa et al. [32] JPHC cohort |
41,526 subjects 12.9 years |
217 sub-arachnoid bleeds | Age, sex, BMI, smoking, alcohol, sodium, potassium, fatty acids, area, menopause, hyperchol., diabetes | 0.74 (0.46–1.61) | Top quintile of dairy calcium vs lowest quintile |
Larsson et al. [23] ATBC cohort |
26,556 men 13.6 years |
196 sub-arachnoid bleeds | Age, smoking, alcohol, BMI, education, total cholesterol, diabetes, IHD, energy intake and activity, intake of various foods and original randomisation | 1.26 (1.00–1.52) | Top and bottom quintile: dairy products, all milks (whole + low-fat + sour) |
0.96M person years: 484 sub-arachnoid bleeds: heterogeneity between studies P = 0.004 Meta-analysis (random effects) RR (95% CI) for highest intake groups 0.65 (0.32–1.31) |
Details of cohort studies in which the consumption of dairy food was related to the risk of haemorrhagic stroke, or a sub-arachnoid bleed, with estimates of homogeneity between the studies and the results of meta-analyses
vs = compared with, or relative to
aThe results for haemorrhagic stroke are based on Larsson et al. [23] for milk. Using Larsson et al. [23] for dairy gives P = 0.022, RR = 0.76 (0.58–1.00)