Table 1

Overview of additional evidence published after our original review, adding to the evidence base on the association of occupational physical activity and all-cause mortality. The study reference, country of origin, occupational physical activity assessment, control of confounding, and effect sizes are reported

First author and year of publication*	Country	Occupational physical activity assessment	Control for confounding factors	Effect size (HR, 95% CI)†	Support for health effects of OPA‡
Bahls, 2018	Germany	In two cohorts (the SHIP study and the CARLA study), the Baecke questionnaire was used to assess occupational physical activity in arbitrary units ranging from 1 to 5 (with effect sizes depicting an increment of 1 SD on this scale)	Age, gender, education, income, smoking, and body mass index	SHIP study 1.03, 0.71 to 1.49 CARLA study 1.13, 0.56 to 2.27	SHIP: uncertain CARLA: uncertain
Fan, 2018	China	China Kadoorie Biobank questionnaire was used to assess MET-hours/day spent in low, moderate and vigorous intensity activities. A median split was performed to derive low and high OPA categories (effect sizes depict high compared with low OPA)	Age, education, marital status, alcohol use, smoking, diet, body mass index, diabetes, family history of heart attack or stroke, menopausal status (for women only), blood pressure, and all other domains of physical activity. Stratification by gender.	Males 0.83, 0.77 to 0.90 Females 0.92, 0.82 to 1.03	Males: positive Females: uncertain
Hermansen, 2019	Norway	Saltin-Grimby Physical Activity Level Scale was used, with four outcome categories: ‘mostly sedentary’, ‘walking’, ‘walking and lifting’, and ‘heavy manual labour’	Age, gender, smoking, body mass index, angina pectoris, myocardial infarction, cerebral insult, diabetes, antihypertensive medication, and leisure time physical activity	Mostly sedentary: 1.13, 1.04 to 1.22 Walking: 1.08, 1.00 to 1.17 Walking and lifting: Reference Heavy manual labour: 1.14, 1.02 to 1.27	Sedentary: positive Walking: uncertain Heavy manual labour: negative
Mikkola, 2019	Finland	Assessed using participants’ job title, and a validated job exposure matrix. The matrix includes gender-specific information on the percentage of individuals within the occupation performing physically heavy work. Quartiles of these percentages constitute the four OPA categories	Age and years of education. Stratification by gender	Males Q1: Reference Q2: 1.18, 1.00 to 1.39 Q3: 1.36, 1.16 to 1.58 Q4: 1.54, 1.31 to 1.80 Females Q1: Reference Q2: 0.97, 0.78 to 1.19 Q3: 1.07, 0.87 to 1.31 Q4: 1.14, 0.94 to 1.39	Males: negative Females: uncertain
Sakaue, 2018	Japan	Using the Baecke questionnaire OPA was assessed in arbitrary units ranging from 1 to 5 (with effect sizes depicting an increment of one unit on this scale)	Age. Stratification by gender	Males 0.75, 0.61 to 0.91 Females 0.96, 0.75 to 1.24	Males: positive Females: uncertain
Wanner, 2019	Switzerland	In two population-based cohorts (NRP1A and MONICA), OPA was categorised into low, moderate and high, based on the amount of walking and stair-climbing (NRP1A) and six levels of intensity (MONICA) based on self-administered questionnaires with sedentary or mostly sitting work as the reference.	Age, education, nationality, marital status, smoking, diet, transport and leisure time physical activity, body mass index, blood pressure, and blood cholesterol. Stratification by gender.	NRP1A Males Low: Reference Moderate: 1.10, 0.96 to 1.26 High: 1.25, 1.05 to 1.50 Females Low: Reference Moderate: 1.06, 0.86 to 1.32 High: 0.88, 0.59 to 1.31 MONICA Males Low: Reference Moderate: 1.02, 0.87 to 1.20 High: 0.93, 0.78 to 1.11 Females Low: Reference Moderate: 1.03, 0.74 to 1.43 High: 1.11, 0.75 to 1.66	NRP1A Males: negative Females: uncertain MONICA Males: uncertain Females: uncertain

*See online supplementary file 1 for a list of references.
†Effect sizes expressed in HR, with 95% CI. Note, in our systematic review2 we assessed the association of high as compared with low PA with all-cause mortality, whereas in this table other comparisons are also shown.
‡Depicts to what extent the evidence supports any of the health effects of OPA—that is, negative health effects, positive health effects, or uncertain. This was done by using the following criteria. Positive: upper limit of the CI below 1, suggesting a reduced risk of all-cause mortality with a high level OPA; Negative: lower limit of the CI above 1, suggesting an increased risk of all-cause mortality with a high level OPA; Uncertain: CI overlapping 1.
CARLA, Cardiovascular Disease, Living and Ageing in Halle; MET, metabolic equivalent; MONICA, Monitoring trends and determinants in cardiovascular disease; OPA, occupational physical activity; SHIP, Study of Health in Pomerania.