Participants and study design
This was a prospective, randomized, parallel-group, and single-blind (assessor blind) study. Participants were recruited through announcements on the institutional intranet, posters, the mail distribution list, and kick-off information events. According to the inclusion criteria, women between 45 and 65 years of age, sedentary lifestyle, low regular physical activity (Freiburger Questionnaire <20 metabolic equivalents of task (MET)-hours per week) (14), and employed at Hannover Medical School (Lower Saxony, Germany) were included over a period of 18 months (11/2013-05/2015). The cut-off value of 20 MET-hours per week for the classification of low regular physical activity was chosen according to physical activity recommendations from the German Federal Centre for Health Education . The inclusion was distributed over the calendar year to exclude seasonal influences. Exclusion criteria were acute or chronic infections, coronary heart disease, diabetes mellitus, oncological diseases, joint replacements or any surgery within the last 6 weeks, and any condition that precluded the realization of an exercise intervention.
Throughout the entire period, new participants were randomized 1:1 into an exercise- (EG) and a waiting-control group (CG) using a computer-based list of random numbers generated by an external collaborator (Figure 1). After their assignment, women in the EG took part in an individualized 6-month exercise intervention. At the beginning and after the 6 months, all women underwent a medical examination by a physician and completed questionnaires. Women in the control group were asked to maintain their current physical activity and dietary habits.
The presented data include a secondary analysis of our study. The primary outcome of this trial, changes in telomere length with 6 months of exercise training, has already been published (16).
This study was carried out in accordance with the Declaration of Helsinki and the current guidelines of good clinical practice. The ethics committee of the Hannover Medical School approved the study (ID: 6428), and written informed consent was obtained from participants before their entry into the study.
Anthropometrics and body composition
Height was measured using a stadiometer, and body weight was determined using a calibrated scale (seca 764, seca gmbh & co. kg, Hamburg, Germany). The bioelectrical impedance method was used to estimate fat- and fat-free mass (InBody 720, JP Global Markets GmbH, Eschborn, Germany).
Pre- or postmenopausal status was set according to the definitions recommended by the WHO Scientific Committee in 1980 as the permanent cessation of menstruation resulting from loss of ovarian follicular activity (17). Menopause for our participants was defined as a 6-month absence of menstruation. Everything else was set as premenopausal.
We distributed questionnaires to estimate daily physical activity (Freiburger Physical Activity Questionnaire) and work ability (Work Ability Index [WAI]) (18). The Freiburger Physical Activity Questionnaire was used to assess the total and exercise-related physical activity of adults, both of which are specified as MET-hours per week. The WAI questionnaire (short form) contains 7 items concerning work, work ability and health: WAI item 1 (current work ability compared with the lifetime best, 0-10 points), WAI item 2 (work ability in relation to the demands of the job, 2-10 points), WAI item 3 (number of current diseases diagnosed by a physician, 1-7 points), WAI item 4 (estimated work impairment due to diseases, 1-6 points), WAI item 5 (sick leave during the past year, 1-5 points), WAI item 6 (own prognosis of work ability two years from now, 1-7 points), and WAI item 7 (mental resources, 1-4 points). The sum of these questions results in a total score ranging from 7 to 49 points, with higher values representing greater work ability.
The WAI total score is categorized into four subgroups: 1 = poor (7-27 points), 2 = moderate (28-36 points), 3 = good (37-43 points) and 4 = excellent work ability (44-49 points) (19). To examine the influence of the level of work ability at baseline, the participants were subgrouped into poor-moderate (WAI 1), good (WAI 2) and excellent (WAI 3) groups. The poor WAI subgroup was combined with the poor-moderate subgroup because of only three cases in the poor subgroup.
At baseline and after 6 months, an incremental bicycle exercise test (Ergoline 150P, ergoline GmbH, Bitz, Germany) was performed to measure exercise capacity (maximum workload in Watt) and cardiorespiratory fitness (VO2peak). The test started at 20 Watt (W), and the workload increased by 10 W every minute until the subjects could not maintain the requested 60 rpm pedal frequency (voluntary exhaustion) or the test was prematurely stopped by the physician due to predefined stopping criteria (20). We recorded heart rate and blood pressure, and collected capillary blood samples from the earlobe at rest, 1 min after the start and every 3 min during the test to determine blood lactate concentrations. This was followed by a five-minute recovery period at 20 W (MasterScreen CPx, Carefusion, Höchberg, Germany).
Study intervention: Endurance training
The aim of the training intervention was to perform 210 minutes of endurance training a week (20 - 60 minutes units for at least 3 days per week) over 6 months. This duration was based on positive results from our previous study and general recommendations of a minimum of 30 min of physical activity at 5 days per week but most preferable at all days per week (21-23). We chose endurance exercise at the time of study planning because most evidence pointed on a more pronounced response of endurance training on our primary outcome (telomere length) (24, 25). The participants in the EG were able to complete part of their training during their working hours (full-time staff 60 minutes per week and part-time staff 30 minutes per week) at the in-house health club.
For individualized training, the participants in the EG received heart rate ranges based on the lactate threshold (approx. 60-80% of the estimated max. heart rate) for their respective activities, such as cycling, rowing and walking. These individual heart rate ranges were based on their prior measured cardiorespiratory fitness and a 30-minute constant load test on a bicycle ergometer with 50% of the maximum workload and blood lactate measurements.
To control and document the training heart rates, training content and volume, the participants were equipped with a heart rate monitor, a heart rate belt (PM70, Beurer, Ulm, Germany) and an optional paper or online diary. All participants received an individualized training schedule from an exercise physiologist, which includes advice for specified heart rate ranges, and types and duration of individually performed endurance activities. It was up to the participants whether they completed the endurance activities at home, on commuting to work, in their leisure time or on the exercise machines in the workplace health club. The training schedule also offered the possibility to attend specially created group fitness courses such as Nordic walking, aqua fitness and running. During the whole intervention period, an exercise physiologist supervised and adapted the individual training program.
The normal distribution of data was tested with the Kolmogorov-Smirnov test. Chi-square tests c2 were used to compare sample distributions. Group differences at baseline between the EG and CG were assessed with two-tailed independent t-tests for parametric data or Mann-Whitney U tests for nonparametric data. Data were analyzed by the intention-to-treat (ITT) approach, with the last-observation-carried-forward method for missing data. Within-group differences between baseline and after 6 months were calculated with two-sided dependent t-tests for paired samples for parametric data or Wilcoxon tests for nonparametric data. To analyze the differences between study groups over time, a two-way ANOVA for repeated measures was conducted, where the partial eta-squared (h²) was used as the effect size. To compare the group differences between the three WAI subgroups at baseline, a one-way analysis of variance with Bonferroni post hoc tests for parametric data or a Kruskal-Wallis test with post hoc Mann-Whitney U tests for nonparametric data were used. If not otherwise mentioned, all data were presented as the mean ± standard deviation. Significance was accepted as p<0.05. All tests were performed with SPSS Version 25 (SPSS, IBM Corp, Armonk, NY, USA).