Table 1 Summary of high quality studies (n = 25)

Shaw et al., 2014 [18]

Level of Living Survey (LNU) and the Swedish Panel Study of Living Conditions of the Oldest Old (SWEOLD)

Sweden

13

1682

44

Structured interview

Index summarising activity over the past 12 months. Respondents engaging in gardening, hunting, dancing, and any type of sport “sometimes” received one point and two points if reported as “often”. Two points were also given for any person who engaged in a sport regardless of frequency. If one had a score of less than two they were defined as inactive.

Age, gender, education, and occupational status. Also, to account for the potential influence of poor health on smoking and physical inactivity behavior, all models controlled for health status in 1981, including measures of circulatory problems mobility problems, and psychological problems.

PA and smoking

All-cause mortality

Prinelli et al., 2015 [14]

Intervention title not mentioned.

Italy

17.4

974

40–74

Not validated

PA was assessed by questionnaire asking about the frequency that one played sports and a binary score was created. Physically active if they were engaged in at least one sport, and physically inactive if they did not play sports.

Age, sex, education level, BMI, time spent watching television, and energy expenditure.

PA, diet, and smoking.

All-cause mortality

Behrens et al., 2013 [29]

NIH-AARP Diet and Health Study.

USA

12.5

170,672

62.5

Not validated

Looked at vigorous and moderate PA. Used a cut-off of 3 times of vigorous PA per week to define adherence to the vigorous PA recommendations. Used a cut-off of 3 h/week of moderate activity to approximate adherence to the moderate PA recommendations. People were assumed to meet the American College of Sports Medicine guidelines if they met 1 of the moderate or vigorous criteria.

Age, marital status, education, race/ethnicity, body height, hip circumference, and alcohol intake.

PA, abdominal leanness, smoking, and diet. Sub-analysis included alcohol consumption

All-cause mortality

Shaw & Agahi, 2012 [30]

Health and Retirement Study (HRS)

USA

10

19,662

Not mentioned because combined multiple cohorts.

Not validated

Were asked, “On average over the last 12 months have you participated in vigorous PA or exercise three times a week or more?” If they answered yes, they were classified as active, and no, inactive.

Age, gender, race, education, household income, and marital status. Also, to control for potential health conditions, they controlled for self-rated health, self-reported presence of 5 serious chronic conditions, a count of functional limitations, and weight status.

PA, smoking, alcohol consumption

All-cause mortality

Carlsson et al., 2013 [31]

Intervention title not mentioned.

Sweden

11

4232

60

Not validated

Leisure time PA during the past year was categorised as inactive, light PA if at least 2 h/week, moderate PA if at least once or twice per week, and intense PA if 3+ times per week. Healthy PA score was at least once per week.

Age (everyone was 60 years old), sex, education level and BMI.

PA, smoking, and diet.

All-cause mortality and CVD

Hulsegge et al., 2016 [39]

The Doetinchem Cohort Study

Netherlands

9.8

5263

46

Validated

Hours/week of spent on PA was summed. Only PA’s with a MET value of 4.0 or higher were included, which is in line with the Dutch PA recommendations.

Age, sex, highest educational level achieved during follow-up, and employment status in model

PA, diet, smoking, and alcohol consumption

All-cause mortality and CVD

Petersen et al., 2015 [32]

Diet Cancer and Health Study

Denmark

14

6768

50–64

Not validated

Examined the average number of hours/week spent in the past year on PA in leisure, moderate, and high intensity time during summer and winter months, which were averaged. Being physically active for at least 30 min/day at moderate intensity was set as the cut-off following the World Cancer Research Fund and Nordic recommendations.

Age, civil status, and length of education.

PA, smoking, alcohol consumption, waist circumference and diet.

All-cause mortality, CVD, and Cancer

Yun et al., 2012 [20]

Intervention title not mentioned.

Korea

10.3

59,941

30–84

Structured interview

Participants activity levels were assessed via interview, and participants were given a score of 1 (binary) if they reported PA (never, < 3 times per/week).

Stratified by sex and adjusted for age, income, and education

PA, smoking, alcohol consumption, BMI. Also, subgroup analysis included fruit and vegetable consumption

All-cause mortality and Cancer

Kvaavik et al., 2010 [19]

Health and Lifestyle Survey (HALS)

United Kingdom

20

4886

43.7

Structured interview

PA was assessed by number of leisure time exercise activities that one had performed during the past fortnight and time spent doing these activities. Poor PA was defined as spending little or no time on PA (< 120 min during 1 week).

Age, sex, occupational social class, BMI, blood pressure, and the other three health behaviours.

PA, smoking, diet, and alcohol.

All-cause mortality, CVD, and Cancer

McCullough et al., 2011 [36]

The CPS-II Nutrition Cohort

USA

14

111,966

Men 63.6 years & women 61.9 years.

Not validated

Participants reported average time per week spent in 7 recreational activities. These were then converted into MET scores. MET-hours/week less than 8.75 received a score of 0, 8.75 to < 17.5 earned a score of 1, and ≥ 17.5 earned a score of two for reaching preferable levels according to the American Cancer Society.

Age, education, smoking history.

PA, BMI, diet, and alcohol consumption

All-cause mortality, CVD, and Cancer

Larsson et al., 2016 [25]

Cohort of Swedish Men (COSM) and the Swedish Mammography Cohort (SMC)

Sweden

13

64,679

59.3

Validated

PA was assessed in time spent on various activities during the previous year. Researchers added up time per week spent engaged in walking/bicycling and exercise. Those participants with ≥150 min/wk. of PA were classified as active.

Age, education, family history of myocardial infarction before 60 years of age, aspirin, and history and diagnosis of hypertension, hypercholesterolemia, diabetes mellitus, and atrial fibrillation before baseline.

PA, smoking, BMI, and diet

Heart failure

Kabat et al., 2015 [15]

NIH-AARP study

USA

10.5

476,396

50–71

Not validated

PA assessment was based on PA was based on 4 levels of strenuous activity (never/rarely/1–3 times per month, 1–2 times per week, 3–4 times per week, and ≥ 5 times per week).

Age, education, race, 14-level smoking variable, marital status. In addition, they examined the association of adherence with risk of cancer at individual sites in men and women separately and, for sites with smaller numbers, in both sexes combined.

PA, BMI, diet, and alcohol consumption.

All-cause mortality and Cancer

Larsson et al., 2014 [33]

Swedish Mammography Cohort

Sweden

10.4

31,696

49–83

Not validated

Women reported on average how many minutes per day they had walked/bicycled, and how many hours per week they had exercised. Binary PA was created and a low risk PA behaviour was considered to include both low to moderate activities (walk/cycling ≥40 min/day), or more vigorous PA (exercise ≥1 h/week).

Age, educational, use of aspirin, diabetes, atrial fibrillation, family history of myocardial infarction before the age of 60 years, total energy intake, and non-recommended foodscore.

PA, diet, alcohol consumption, smoking, and BMI

Stroke

Akesson et al., 2014 [26]

Intervention title not mentioned.

Sweden

11

20,721

45–70

Validated

Low risk PA included men who walked or cycled for ≥40 min/day and exercised ≥1 h/week.

Age, level of education, marital status, family history of MI, use of aspirin, non-Recommended Food Score, and energy intake.

PA, smoking, diet, alcohol consumption and waist circumference.

MI

Chomistek et al., 2013

Women’s Health Initiative Observational Study

USA

12.2

71,018

50–79

Not validated

MET hours/week was computed as a sum of frequency, duration, and intensity of PA. Women were classified into 4 categories: inactive (≤1.7 MET hours/week), low PA (1.8–8.3 MET hours/week), medium PA (8.4–20 MET hours/week), and high PA (> 20 MET hours/week).

Stratified by age, and covariates included race/ethnic group, family income, education, marital status, smoking, parental history of premature MI, depression, alcohol intake, hours of sleep, and intake of total calories, saturated fat, and fiber.

PA and sitting time

CVD

Wang et al., 2011 [35]

Five independent cross-sectional, population-based health examination surveys (FINRISK)

Finland

14.1

38,075

25–64

Not validated

PA was merged into 3 categories: ‘low’, corresponding to light levels of both occupational and leisure time PA; ‘moderate’, corresponding to moderate or high level occupational or leisure time PA; and, ‘high’, corresponding to moderate or high level of both occupational and leisure time PA. This was then dichotomized into PA (low versus moderate or high).

Age, study year, education, systolic blood pressure, total cholesterol, and history of myocardial infarction, valvular heart disease, diabetes, and using antihypertensive drug

PA, smoking, BMI, and vegetable consumption

Heart failure

Kirkegaard et al., 2010 [27]

Diet, Cancer and Health Cohort Study

Denmark

9.9

55,487

50–64

Validated

Participants reported their PA via reporting avg. number of hours/week spent the past year engaged in leisure activity, or occupational activity. Participants were deemed active if they reported > 30 mins/day of activity or had a job with light manual activity or heavy manual activity.

Age, education, use of non-steroidal anti-inflammatory drugs, hormone replacement therapy among women, and history of cancer in first degree relatives.

PA, waist circumference, smoking, alcohol consumption, and diet.

Colorectal cancer

Nomura et al., 2016 [28]

The Black Women’s Health Study

USA

13.9

49,103

21–69

Validated

PA was scored against the World Cancer Research Fund/American Institute for Cancer Research recommendations.

Age, geographic region, caloric intake, smoking, family history of colorectal cancer, education, menopausal status, diabetes, insulin usage, aspirin usage, colonoscopy, and sigmoidoscopy.

PA, smoking, and diet

Colorectal cancer

McKenzie et al., 2015 [37]

European Prospective into Cancer and Nutrition (EPIC) cohort study

Europe (mix of countries)

10.9

242,918

53.2

Not validated

Healthiest behaviour was defined as scoring in the top quintile of PA based on recreational and household MET tasks.

Height, age at menarche, age at full term pregnancy, education, oral contraceptive use, hormone replacement therapy use, breastfeeding, total energy intake excluding alcohol

PA, smoking, alcohol consumption, BMI, and diet.

Breast Cancer

Eriksen et al., 2015 [17]

UK SABRE Study

United Kingdom

21

2096

40–69

Not validated

PA was computed as the total weekly energy expended in sporting activities, cycling, walking, and in other strenuous activity during leisure time. Being active and meeting the cutoff was defined as those undertaking moderate exercise for 5+ hours per week, or vigorous exercise for ≥2.5 h/week.

Age, sex, systolic and diastolic blood pressure, hypertension treatment, total cholesterol, HDL-cholesterol, BMI, social class, employment status, and occupational physical activity.

PA, smoking, alcohol consumption, and fruit and vegetable consumption

CVD

Vergnaud et al., 2013 [38]

European Prospective into Cancer and Nutrition (EPIC) cohort study

Europe (mix of countries)

12.8

121,443

25–70

Not validated

PA was scored against the World Cancer Research Fund/American Institute for Cancer Research recommendations. We looked at men only.

Stratified for sex, age, and center and adjusted for educational, smoking, and intensity of smoking. Additional adjustment for total energy intake.

PA, BMI, diet, and alcohol consumption

All-cause mortality, CVD, and Cancer

Fazel-taber Malekshah et al., 2016 [16]

Golestan Cohort Study (GCS)

Iran

8.0

40,708

40–75

Not validated

For PA, a low risk group was those that reported at least 30 min/day of moderate and vigorous PA on average.

Adjusted for age, gender, resident area, ethnicity, socioeconomic status (Wealth Score), marital status, education, alcohol consumption, and opium use.

PA, smoking, and alternative eating index.

All-cause mortality, CVD, and cancer.

Zhang et al., 2017 [21]

Shanghai Men’s Health Study

China

9.3

59,503

40–74

Validated

People were categorised as healthy if they engaged in > = 150 min/week of moderate-to-vigorous PA.

Adjusted for age, occupation, education, income, and history of diabetes mellitus, and colorectal cancer family history of first degree relatives.

PA, waist hip ratio, smoking, alcohol consumption, and dietary behaviour.

Colorectal cancer

O’Donovan et al., 2017 [22]

Health Survey for England (HSE) and Scottish Health Survey (SHS)

UK

9.4

106,341

47

Validated

Asked questions about PA frequency and duration in activity (domestic PA, light-intensity, moderate-intensity) and type/intensity for sports/exercise. MET-hours were calculated and 3 categories of PA were made (≥60 min/wk., < 60 min/wk., and none).

Adjusted for age, sex, occupation, and longstanding illness.

PA and smoking

All-cause mortality, CVD, and cancer.

Larsson et al., 2017 [23]

Cohort of Swedish Men and Swedish Mammography Cohort

Sweden

15.3 for men and 15.7 for women

33,454 men and 30,639 women

45–79

Validated

PA and sedentary behaviour was calculated by asking about previous year. Participants who had recorded on average at least 150 min per week of activity were categorised as healthy and those with < 150 min per week of activity as unhealthy.

Adjusted for age, education, family history of MI and cancer, asprin use, sedentary time, BMI, and history of diabetes, hypertension and hypercholesterolemia.

PA, smoking, alcohol consumption, and a healthy diet

All-cause mortality