OBJECTIVE: Physical activity is a modifiable risk factor for type 2 diabetes, partly through its action on insulin sensitivity. We report the relation between insulin sensitivity and physical activity measured by accelerometry. RESEARCH DESIGN AND METHODS: This cross-sectional study is of 346 men and 455 women, aged 30 to 60 years, without cardiovascular disease and not treated by drugs for diabetes, hypertension, dyslipidaemia or obesity; they were recruited in 18 clinical centres from 13 European countries. Insulin sensitivity was measured by hyperinsulinaemic euglycaemic clamp. Physical activity was recorded by accelerometry for a median of six days. We studied the relationship of insulin sensitivity with total activity (number of counts/min), percent time spent sedentary, percent tine in light activity and activity intensity (whether the participant recorded some vigorous or some moderate activity). Results: In both men and women, total activity was associated with insulin sensitivity (P < 0.0001). Time spent sedentary, in light activity, and activity intensity were also associated with insulin sensitivity (P < 0.0004/0.01; 0.002/0.03; 0.02/0.004 respectively for men/women), but lost significance once adjusted for total activity. Adjustment for confounders, such as adiposity, attenuated the relations with total activity; there were no interactions with confounders. Even in the 25% most sedentary individuals, total activity was significantly associated with better insulin sensitivity (P < 0.0001) Conclusions: Accumulated daily physical activity is a major determinant of insulin sensitivity. Neither time spent sedentary, in light-activity, nor bouts of moderate or vigorous activity impacted on insulin sensitivity independently of total activity.