Lifestyle factors and risk
Clinical bottom line
It is possible to take good epidemiological research and express the results as risk - odds of dying, for instance - as well as more statistical ways like relative risk. Lifestyle has a huge impact on our likelihood of living and dying, many many times greater than risks from other behaviours - even riding motorcycles.
Data sources
References to data sources are in the Table. Those chosen are seminal studies concerning the relationship between lifestyle factors - smoking, eating, exercise - and mortality.
What the sources tell us
The studies have different starting populations and look at different behaviours for varying follow up periods. It is not possible to standardise the risks, so the risks given are for the study population and the follow up period.
What they show is to some extent what we already know. Lifestyle choices have a huge effect on length of life - and for that matter quality as well. Behaviours that help hearts generally help prevent cancers, and osteoarthritis, and macular degeneration, and a host of other conditions no sensible person wants. Most of the studies have already appeared in bandolier's pages over the years, but this is a different way of looking at the results.
The first three studies emphasise the effect of lifestyle on risk in the middle aged westerner.
- For younger women followed up for 14 years, the effect of having a low risk as opposed to a high risk lifestyle moves the risk of dying from 1 in 240 (high risk) to 1 in 2,000 (low risk) (Stampfer et al, 2000).
- For older women followed up for six years we see the same effect. With a high risk lifestyle the risk of dying was 1 in 250, whereas for a low risk lifestyle it was 1 in 2,500 (Akesson et al, 2007).
- The UK study following up people aged 45 to 79 years and with an average age of 58 years for 14 years (Khaw et al, 2008) showed a similar effect. A high risk lifestyle came with a risk of dying of 1 in 4, while a low risk lifestyle had a risk of 1 in 20.
Studies of individual factors tend to be less spectacular, but walking has big effects on cancer and heart disease. Showing any effect of a single intervention on death in older people followed up for a longer period is harder, because people eventually die of something, but even overall mortality can be shown to be lower in people who walk.
The big effect is smoking - both on overall mortality, and especially on lung cancer.
Give us the odds
Table 1: Effect of behaviours on risk of dying. The odds quoted are for the population studies and the duration of follow up
| Study |
Risk group |
Baseline population |
Deaths |
Odds of dying over period studied and for cause (1 in ) |
| Stampfer et al. N Engl J Med 2000;343:16-22. Primary Prevention of Coronary Heart Disease in Women Through Diet and Lifestyle
Data from US Nurses' Health Study; population aged 30-55 at start of study; mortality from coronary heart disease only; 14 yrs follow up |
Overall | 84,129 |
296 |
284 |
| 2 or fewer low-risk factors | 64,779 |
272 |
238 |
|
| 3 low-risk factors | 10,684 |
16 |
657 |
|
| 4 low-risk factors | 6,057 |
6 |
962 |
|
| 5 low-risk factors | 2,608 |
1 |
1,988 |
|
Khaw et al. PLoS Medicine 2008;5(1): e12. Combined Impact of Health Behaviours and Mortality in Men and Women: The EPIC-Norfolk Prospective Population Study
Population aged 45-79 (mean age 58.2) at start of study; mortality from all causes; 14 yrs of follow up |
Overall | 20,244 |
1,977 |
10 |
| 0 low-risk factors | 196 |
49 |
4 |
|
| 1 low-risk factors | 1,407 |
277 |
5 |
|
| 2 low-risk factors | 4,568 |
651 |
7 |
|
| 3 low-risk factors | 7,788 |
682 |
11 |
|
| 4 low-risk factors | 6,285 |
318 |
20 |
|
Akesson et al. Arch Intern Med. 2007;167(19):2122-2127. Combined Effect of Low-Risk Dietary and Lifestyle Behaviors in Primary Prevention of Myocardial Infarction in Women
Study population of women aged 48-83 yrs at baseline; mean follow up duration 6.2 yrs |
Overall | 24,444 |
51 |
479 |
| 1 or fewer low risk factors | 9,289 |
38 |
245 |
|
| 2 low risk factors | 7,089 |
7 |
1,019 |
|
| 3 low risk factors | 5,133 |
4 |
1,148 |
|
| 4 low risk factors | 1,711 |
1 |
1,476 |
|
| 5 low risk factors | 1,222 |
0 |
2,460 |
|
Hakim et al. N Engl J Med 1998;338:94-9. Effects of Walking on Mortality Among Nonsmoking Retired Men
Population aged 61 - 81 (mean age 68.9) at start of study; 12 yrs of follow up; all causes of mortality |
Overall | 707 |
208 |
3 |
| Walking < 1 mile/day | 151 |
65 |
2 |
|
| Walking > 1 mile/day | 556 |
143 |
4 |
|
| Walking 1-2 miles/day | 379 |
105 |
4 |
|
| Walking > 2 miles/day | 177 |
38 |
5 |
|
Hakim et al. N Engl J Med 1998;338:94-9. Effects of Walking on Mortality Among Nonsmoking Retired Men
Population aged 61 - 81 (mean age 68.9) at start of study; 12 yrs of follow up; coronary heart disease as mortality outcome |
Overall | 588 |
29 |
20 |
| Walking < 1 mile/day | 122 |
8 |
15 |
|
| Walking > 1 mile/day | 466 |
21 |
22 |
|
| Walking 1-2 miles/day | 321 |
18 |
18 |
|
| Walking > 2 miles/day | 145 |
3 |
48 |
|
Hakim et al. N Engl J Med 1998;338:94-9. Effects of Walking on Mortality Among Nonsmoking Retired Men
Population aged 61 - 81 (mean age 68.9) at start of study; 12 yrs of follow up; cancer as mortality outcome |
Overall | 672 |
62 |
11 |
| Walking < 1 mile/day | 142 |
19 |
7 |
|
| Walking > 1 mile/day | 530 |
43 |
12 |
|
| Walking 1-2 miles/day | 361 |
34 |
11 |
|
| Walking > 2 miles/day | 169 |
9 |
19 |
|
Knoops et al. JAMA. 2004;292(12):1433-1439. Mediterranean Diet, Lifestyle Factors, and 10-Year Mortality in Elderly European Men and Women: The HALE Project
10 yrs of follow up, data from graph, mean age of population at start of study was 74 yrs |
Healthful lifestyle factor score 0-1 | 246 |
145 |
2 |
| Healthful lifestyle factor score 2 | 702 |
309 |
2 |
|
| Healthful lifestyle factor score 3 | 954 |
334 |
3 |
|
| Healthful lifestyle factor score 4 | 437 |
131 |
3 |
|
Doll et al. BMJ 2004;328(7455):1519. Mortality in relation to smoking: 50 years’ observations on male British doctors
50 yrs follow up, population aged from 21-51+ at baseline; all causes of mortality |
Lifelong non-smoker | 52 |
||
| Former smoker | 41 |
|||
| Current smoker | 28 |
|||
| Current smoker (>25 per day) | 22 |
|||
| 50 yrs follow up, population aged from 21-51+ at baseline; lung cancer as only mortality indicator | Lifelong non-smoker | 5,882 |
||
| Former smoker | 1,471 |
|||
| Current smoker | 402 |
|||
| Current smoker (>25 per day) | 240 |