Background Many beneficial health outcomes for elderly are associated with sedentary (SED) and physical activity behavior (PA). However, research is scarce regarding how elderly-interventions can improve various SED- and PA-behavior.

Purpose The aim with this study was to determine if eight-weeks of regular supervised exercise would have any effect on movement-patterns in elderly. For comparison, a control group, not receiving supervised exercise, performed similar pre- and end-tests intervals. Reliability of elderly accelerometer measures, in two separate pre-tests was also investigated.

Methods Seventy-eight elderly exercised (65–91 years) and 43 participated in a control group (65–88 years). The exercise group had combined aerobic and strength-training, twice/w à 1 hour. Accelerometer assessments for the exercise group were performed three times; Pre-1 and Pre-2 and one test in the last week of the 8-week-exercise-period. The controls made one pre-test and one end-test at corresponding spring time intervals, without receiving any supervised physical activity.

Results The exercise-group significantly improved from Pre-1 to End-test moderate-to-vigorous-PA-intensity (MVPA +41 min/week=+0.8% of-total-daily-wear-time), light-intensity (LPA +228 min/w=+3.9%), total-SED-time (−254 min/w=−4.7%) and SED-bouts-à−10 min (−320 min/w=−5.7%). The controls showed no significant improvements from pre-1-to-End-tests, except for LPA (+2.3%) and total-SED-time (and −2.7%). Total-daily-counts in the End-test increased for controls (+6,7%) but significantly more for the exercise group (+17,3%). Between the exercise groups´ two pre-tests, no significant difference was observed, except for total-PA (+4.6%) and SED-total (−2.4%).

Conclusion Elderly participating in supervised exercise increase PA- and decrease SED-time, which is of importance because it is linked with improved health for elderly, longevity and potent socioeconomic gains.