Definition: Vegans were defined as consuming animal products less than once per month, which can be argued to make sense from a nutrition perspective.
Numbers of deaths: 1008 vegans, 4002 lacto-ovo-vegetarians, 1225 pescatarians, 764 semi-vegetarians, and 5516 meat eaters died during the course of the study.
Results regarding mortality risk
Note: These data are only regarding mortality risk, i.e., the statistical risk (probability) of dying of these diseases in this particular population (Adventist participants of this study in the USA)."Table 2. Associations between dietary patterns and all-cause mortality among 88,400 participants in the Adventist Health Study-2":
All vegans (men and women combined):
"
Vegan | Number of deaths: 1008 | Model 1: 0.88 (0.82, 0.95) | Model 2: 0.97 (0.90, 1.05) | Model 3: 0.99 (0.92,1.06) |
"
Male vegans only:
Relative risk of death compared to the meat eater ("non-vegetarian") group:
"
|
Vegan |
Number of
deaths: 408 |
Model 1: |
Model 2: |
Model 3: |
|
At age
65 |
0.60 (0.49, 0.74) |
0.71 (0.57, 0.87) |
0.72 (0.58, 0.88) |
|
|
At age
85 |
0.91 (0.82,
1.03) |
1.06 (0.94,
1.19) |
1.07 (0.95,
1.21) |
"
Note (older vegan men): I don't know why the older group of vegan men had a higher risk. Maybe a higher percentage of these men had vitamin B12 deficiency for a longer period of time (but this is purely speculative). My guess is that it usually all boils down to the ten key nutrients for vegans, especially vitamin B12.
Female vegans only:
Relative risk of death compared to the meat eater ("non-vegetarian") group:
"
Vegan | Number of deaths: 600 | Model 1: 0.90 (0.82, 0.99) | Model 2: 0.97 (0.88, 1.07) | Model 3: 0.98 (0.89, 1.08) |
"
"Model 1: Adjusted for sex (male or female), race (black or non-black), and each region of residence (continuous) (West South Central, Mountain, West North Central, New England Mid-Atlantic, East North Central, South Atlantic, East South Central, Canada, or Pacific). The region variables are considered as continuous, because we also have accounted for the proportion of follow-up in other regions which they have lived and the total of all must add to 1.0.
Model 2: Adjusted for model 1 + marital status (yes or no), education (high school or below, some college, or bachelor’s degree or above), body mass index (>14 to <18.5, ≥18.5 to <25.0, ≥25 to <30, or ≥30 kg/m2), exercise (mins/week); smoking categories varied due to very low number of subjects currently smoking or recently quit in certain subgroup analyses: never, stopped smoking over 30 years ago, then i) 21-30, 11-20, 6-10, <1 to 5 years or current smokers for the combined participants in the vegetarian/non-vegetarian analysis; or ii) 21-30, 11-20, <1 to 10 years or current smokers for sex-specific analyses of vegetarian/non-vegetarian analyses; or iii) <1 to 30 years or current smokers, for dietary pattern-specific analyses; alcohol: wine, beer, and liquor (g/d), and female variables (centralized) (menopausal status, age at menopause, oral contraceptive use, and hormone replacement therapy).
Model 3: Adjusted for model 2 + prevalent diseases: available cardiovascular diagnoses (myocardial infarction, coronary artery bypass surgery or coronary stenting, carotid artery endarterectomy; yes or no), cancer (yes or no), respiratory disease (bronchitis, emphysema; yes or no), and stroke (yes or no)."
What about death from specific causes?
Their Figure 2 (modified with blue colour):
The authors note that a higher risk for stroke, dementia, and Parkinson's disease might be related to vitamin B12 deficiency and/or low intakes and blood levels of EPA and/or DHA.