Design
Interventional study without control group. Trial setting is GP practices and trial type is treatment. Primary and secondary measurements were assessed at baseline (T0), after 3 months (T1) directly after the intervention, and at the long-term follow-up after mean 22 months (T2). Primary outcome measure: maximal step-up height (MSH). Secondary outcome measures: age, height, bodyweight, body mass index (BMI), waist circumference, aerobic fitness (VO2-max), self-reported physical activity (PA) and health-related quality of life (SF-36 scores), the subscale of physical function (PF). Inter-examiner repeatability test at T2.
Subjects
Out of 214 female patients consecutively referred from primary health care, 178 attended a first test (T0) and 156 participated in a 3-month group exercise intervention program and took part in a second assessment (T1) [28]. Out of these 156 patients, 114 were randomly invited for a third assessment (T2) out of whom 101 agreed to participate. No significant difference in baseline measurements between the 101 participating patients and remaining 77 were found regarding anthropometry, education, sick leave, self-reported pain and/or reduced physical function. However, age differed significantly, with a higher mean age in the 101 participants compared to the 77 who did not participate in T2 (mean 52 years vs. 47 years, p = 0.003).
Intervention with group training
The 3-month group exercise intervention, tailored to participants’ needs, has previously been described in detail [28].
In brief, 3 weekly training sessions—chosen from different types of group exercise—were agreed with each patient when they began the program. The training was supervised by medically trained and experienced coaches and conducted at local gyms, the municipal swimming pool, on outdoor tracks and in school facilities. At each training session, coaches explained the intensity of the exercise, controlled, for instance, by the pace of music or by the duration of sessions of Nordic walking. The different levels of session intensity—classified as very light, light, or moderate—helped the coach to guide the patient to start at a suitable level depending to baseline tests, to minimise any cardiovascular risks, or acute or overuse injuries. The patients were trained to use the 6-20 Borg RPE Scale [30] to record perceived intensity. Level 13-15 was recommended as a goal at each training session once they individually had received the go-ahead from the coach. Healthy lifestyle to promote positive effects of the training on especially muscle mass and strength was recommended by the coach during intervention; no nicotine two hours before and after the exercise, and adequate food intake more than two hours before exercise and directly after training, was recommended.
At the end of the 3-month program, patients—together with their coaches—planned how to continue with similar activities after the program had ended. Both during and after the 3-month intervention period patients were offered standard primary care, including customary follow-up appointments.
Measurements
Physical examination
Body weight was measured in light clothing without shoes to the nearest 0.1 kg using electronic scales (Seca Delta model 707). Height was measured without shoes to the nearest 0.5 cm using a measuring stick fixed to the wall. Body mass index (BMI) was calculated according to the standard formula (kg·m-2). Waist circumference (cm) was measured at the level of the umbilicus according to standard practice, table 1a.
Diagnosis, work status and capacity for work
A standardized and structured protocol was used by GPs to refer patients to the intervention program. It included common diseases, joint pain or discomfort at 14 specified localizations, and a record of medication. The capacity for work, prevalence of sick leave, disability pension and job seeking status were also recorded in the protocol, table 1b.
Physical function score and any limitation score
The Swedish version of SF-36 [31] assessing health-related quality of life was filled out by patients in advance of admission to the program. Where needed, assistance was given at the time of the first visit. All raw scores were transposed onto scores on a weighted 0-100 scale, higher scores representing better health status [32]. The subscale PF was extracted for further item analyses; the items describing doing moderate activities (3b), climbing several flights of stairs (3d), bending, kneeling or stooping (3f) and walking more than 2 km (3g). We analysed: i) all eight scales (0-100); ii) any limitation score in the PF items 3b, 3d, 3f and 3g, respectively, ranging from 1-3 (1 = severely limited, 2 = somewhat limited, 3 = no limitation); and iii) the sum of 3b, 3d, 3f and 3g ranging from 4-12 (from 4 = severely limited in all items to 12 = no limitation in any of the four items), table 1b.
Physical capacity, testing procedures and equipment
Maximal step-up height (MSH) (cm) was assessed by the standardized maximal step-up test (MST) on each leg. MST has previously been described including tests of validity and repeatability [27]. The intra-examiner test-retest (one week between occasions) showed MSH repeatability of 6.9 and 5.9 cm for right and left leg, respectively. Levels are pre-set at 3 cm apart on a tailor-made device called the step-up box. If a patient tried and failed at the lowest MSH level with both legs, MSH mean value for statistical analysis was interpreted as zero. If one leg failed and the other leg managed 18 cm, the mean MSH registered was 9 cm. A missing value was registered for each leg when a patient did not try, and no mean MSH was calculated. The MST method includes information about how to instruct and encourage the patient during the test. In brief, the tester demonstrates how the standardized MST is performed and instructs patients to step up without assistance from the foot that remains on the floor and without support from handrails. Patients are allowed three attempts at the highest level for each leg with verbal encouragement given. The suggested goal for each patient is to manage MSH with 90 degrees at hip and knee while in the starting position.
The VO2-max was estimated as described by Åstrand [33], one of the most commonly used submaximal exercise tests. In the original article reported validation was r=0.78 and the coefficient of variation (CV) was 15% for a mixed sample of subjects. In new research, published after our study was conducted, the validity correlation coefficient for the Ekblom-Bak submaximal exercise test [34] was r=0.91 and CV was 9,3% compared to results from Åstrands test in that study, r=0.68 and CV 18.1% From each participant’s individual heart rate response to a given sub-maximal workload (i.e. 50–150 W, depending on the participant’s weight and self-reported physical activity) using a bicycle ergometer (Monark Exercise AB, Vansbro, Sweden), oxygen consumption was recorded at a steady-state with a heart rate 120-150 beats·min-1, and quantified as peak absolute oxygen consumption per minute (L·min-1) and per kg body weight per minute (mL·kg-1·min-1).
Self-reported exercise and PA levels
At baseline (T0) and at the third assessment (T2), exercise and PA were assessed using a validated questionnaire common in occupational health care in Sweden [35]. The PA question included a definition of exercise, i.e. to allocates time for exercise in order to maintain or improve fitness, health and wellness, and one of five levels were registered, se table 1b [35].
Training and physical activity measurements
Leisure-time PA or sedentary behaviour was not registered during the 3-month group exercise intervention. Patients filled in their own exercise diary at each training session and coaches checked and signed the diary. The self-selected workouts were evaluated in groups of i) mixed aerobic fitness and strength training, ii) aerobic fitness training, and iii) strength training. The dose in times per week (t.p.w), minutes per week (m.p.w) and intensity (6-20 Borg RPE Scale) during the 60-minute workout sessions were recorded. The number of dance therapy and Qigong sessions (replacing other types of workouts and registered as mostly aerobic fitness and strength training, respectively)—often chosen because of fear of movement or to improve balance and coordination—were registered and included in total exercise. Exercise and PA levels were estimated from both patient diaries from 3 months preceding the long-term follow-up and also from questionnaires at long-term follow-up [35]. An effort equivalent to at least 30 minutes of brisk walking corresponding to level light to somewhat hard 11-13 Borg RPE Scale was required for registration as a session.
Statistics
Maximal step-up height was used as the primary end-point. MSH was calculated as mean step-up height of left and right leg. Results from two measurements by two test leaders, with 30-minute interval, at the long-term follow-up was analysed with help of ANOVA to calculate repeatability [36, 37]. We expect 95% of differences between paired observations, i.e. test leaders, to be less than this definition of a repeatability adopted.
MSH and cardio-metabolic main variables and their changes from baseline to 14-30-month follow-up after the 3-month intervention program were analysed using the students T-test. Logistic regression analysis was performed to explore subjects with maintained MSH as outcome (maintained or increased MSH vs. decreased MSH). Mixed aerobic fitness and strength training sessions were dichotomized into high (> 21 sessions during the 3-month training period) or low (≤ 21 sessions). Age (years), time to follow-up (months) and MSH (cm) at baseline were included as covariates, as these were found to be significant in univariate analyses. Results are presented as odds ratio (OR) with 95% confidence interval (95% CI). Model specification was tested by Pearson’s and Hosmer-Lemeshow’s tests.
To describe differences between the two subgroups with the highest MSH maintenance (increased or maintained) and the subgroup with the lowest MSH maintenance (decreased), the Mann-Whitney U-test, 2*1-sided exact p test was used. Furthermore, the data was also split in tertiles of age (two outliers, 81 and 83 years old at baseline, were added to the oldest group). All tests were two-sided and performed on the 0.05 level of significance. No adjustments for multiplicity were done as this is an exploratory study. Overall, we can expect 1/20 statistical significant results to be caused by random for analysis in independent variables. Statistica v 12.0, StatSoft Inc, Tulsa, United States, was used for statistical analyses.