- Email: [email protected]
Exercise Frequency, Health Risk Factors, and Diseases of the Elderly
Archives of Physical Medicine and Rehabilitation journal homepage: www.archives-pmr.org Archives of Physical Medicine and Rehabilitation 2013;94:2046-53
ORIGINAL ARTICLE
Exercise Frequency, Health Risk Factors, and Diseases of the Elderly Wolfgang Kemmler, PhD, Simon von Stengel, PhD From the Institute of Medical Physics, University of Erlangen-Nurnberg (Friedrich-Alexander Universita€t Erlangen-Nu€rnberg), Erlangen, Germany.
Abstract Objective: To determine the effect of exercise frequency on various diseases and risk factors of the elderly. Design: Retrospective analysis of a randomized controlled 18-month exercise trial. Setting: University ambulatory group setting. Participants: Community-dwelling women aged 65 years (NZ162) in the area of Northern Bavaria. Intervention: Mixed, intense aerobic, resistance, and balance protocol for 18 months. Subjects were retrospectively subdivided into 2 groups according to their effective attendance over 18 months (>1e<2 vs 2e4 sessions/wk). Main Outcome Measures: Bone mineral density (BMD), lean body mass, appendicular skeletal muscle mass by dual-energy x-ray absorptiometry, Framingham study-based 10-year coronary heart disease (CHD) risk, and number of falls by calendar method. Results: Significant differences between the low-frequency exercise group (LF-EG) and the high-frequency exercise group (HF-EG) were observed for lumbar spine BMD (HF-EG, 2.4%2.8% vs LF-EG, 0.3%2.2%; P<.001) and proximal femur BMD (HF-EG, 2.4%2.8% vs LF-EG, 0.5%1.6%; PZ.014), lean body mass (1.6%3.4% vs 0.3%2.6%, PZ.053), and appendicular skeletal muscle mass (0.9%4.5% vs 1.3%3.2%, PZ.011). No differences between both exercise groups were observed for 10-year CHD risk (1.94%4.14% vs 2.00% 3.13%; PZ.943) and number of falls (0.951.36 vs 1.031.21 falls/person). Comparing the LF-EG with the less active control group (nZ47), only nonsignificant effects for fall number (PZ.065) and 10-year CHD risk (PZ.178) were evaluated. Conclusions: Although this result might not be generalizable across all exercise types and cohorts, it indicates that an overall exercise frequency of at least 2 sessions/wk may be crucial for impacting bone and muscle mass of elderly subjects. Archives of Physical Medicine and Rehabilitation 2013;94:2046-53 ª 2013 by the American Congress of Rehabilitation Medicine
In Germany, 3 of 4 people aged 65 years and older suffer from 2 and more diseases.1,2 It is generally known that physical activity and in particular exercise positively impacts most risks and is an alternative option to expensive medication with respect to disease prevention and treatment. However, because most exercise trials focus on 1 disease or on risk factors only, the most effective exercise protocol for impacting diverse relevant health problems of the elderly remains to be An audio podcast accompanies this article. Listen at www.archives-pmr.org. Presented in part at the “42. Deutscher Sport€arztekongress,” October 6e8, 2011, Frankfurt, Germany. The article has not been published elsewhere. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.
established. A key decision for an effective exercise protocol is simply how often “exercise” should be performed. Unfortunately, no exercise trial exists yet that focuses on the dose-response effect of exercise frequency on parameters related to fracture risk, sarcopenia, and coronary heart disease (CHD) risk, which may be the most prominent predictors of autonomy and mortality of the elderly.3,4 Based on recent findings,5-8 it was assumed that an exercise frequency generating positive cardiac/metabolic adaptations was ineffective for impacting musculoskeletal parameters. By using data of the Senior Fitness and Prevention (SEFIP) study, this article focuses on the effect of 2 exercise frequencies on musculoskeletal parameters, CHD risk, and falls. The following 4 hypotheses were tested: 1. An average exercise frequency of 2 to 4 sessions/wk over 18 months is significantly more effective for impacting parameters
0003-9993/13/$36 - see front matter ª 2013 by the American Congress of Rehabilitation Medicine http://dx.doi.org/10.1016/j.apmr.2013.05.013
Exercise frequency and diseases related to osteopenia and sarcopenia compared with >1 to <2 sessions/wk. 2. An average exercise frequency of >1 to <2 sessions/wk over 18 months does not significantly impact parameters related to osteopenia and sarcopenia. 3. Both exercise protocols favorably impact CHD risk in elderly females, with no significant differences between groups. 4. Both exercise protocols favorably impact the number of falls in elderly females, with no significant differences between groups.
Methods This retrospective dose-response analysis used participant data of the SEFIP study, a fully registered (www.clinicaltrials.gov) randomized controlled exercise trial. The SEFIP study was conducted between 2006 and 2008 and focused on the effect of exercise on various risk factors among women aged 65 years and older.9,10 To determine the effect of exercise frequency on the given study endpoints, in the present analysis, the exercise group (EG) of the SEFIP study was retrospectively divided into 2 subgroups according to their effective exercise frequency over 18 months.
2047 With relevance for the present analysis, originally 246 women were included and allocated by an external statistician to 2 study arms using computer-generated block randomization: (1) exercise group (EG, nZ123) and (2) wellness control group (WCG, nZ123). After 18 months of intervention, 115 subjects of the EG and 112 subjects of the WCG could be included in the present analysis (details in Kemmler et al9,10). For the present analysis, the weekly attendance rate of the participants for the 18-month intervention was summed, and the EG was subdivided into 2 groups, one subgroup with an overall exercise frequency of <2 sessions/wk (low-frequency EG [LF-EG], nZ36) and another subgroup with an overall exercise frequency of 2 sessions/wk (high-frequency EG [HF-EG], nZ79). To adequately appraise changes among the EGs, the WCG was also subdivided into a more compliant subgroup (20 out of 40 wellness sessions/18mo) and an “inactive” subgroup (<20 sessions/18mo, nZ47). To generate a comparison between exercise and a “nontraining condition,” the following refers exclusively to the results of the “inactive WCG” (see fig 1). Table 1 lists the baseline characteristics of both EGs and the inactive WCG. There were no significant differences between groups with respect to anthropometric, metabolic, or lifestyle parameters that may have confounded the results.
Study endpoints Endpoints of this present dose-response analysis are as follows: Osteopenia (bone mineral density [BMD] at the lumbar spine [LS] and the proximal femur) Fall number (fall rate) Sarcopenia (lean body mass [LBM]; appendicular skeletal muscle mass [ASMM]) CHD risk (Framingham 10-year CHD risk)
Testing procedures Baseline and 18-month follow-up measurements were generally performed at the same time of the day (1h) in a fixed order and by the same researcher. All assessments were carried out in a blinded fashion.
Anthropometry Subjects For a detailed description of the recruitment process, the reader is referred to recent publications of the SEFIP study.9,10 Figure 1 shows the participant flow of the SEFIP study adapted for the present analysis. Briefly, subjects were eligible if they were aged 65 years or older, of Caucasian race, and living independently in the community. Criteria for exclusion were (1) diseases/medication affecting bone metabolism or falls risk, (2) history of stroke or cardiac events, (3) inflammatory diseases, (4) secondary osteoporosis, (5) participation in exercise studies during the last 2 years, and/or (6) athletic history during the last decade.
List of abbreviations: ASMM BMD CHD DXA EG HF-EG LBM LF-EG LS SEFIP tHip WCG
appendicular skeletal muscle mass bone mineral density coronary heart disease dual-energy x-ray absorptiometry exercise group high-frequency exercise group lean body mass low-frequency exercise group lumbar spine Senior Fitness and Prevention (study) total Hip wellness control group
www.archives-pmr.org
Height was determined with a stadiometer, and weight was measured with minimal clothing on digital scales. Body mass index was calculated as weight divided by height squared (kg/m2). Waist circumference was determined as the minimum circumference between the distal end of the rib cage and the top of the iliac crest along the midaxillary line.
Lean body mass, appendicular skeletal muscle mass LBM was assessed with dual-energy x-ray absorptiometry (DXA) technique using the whole-body standard protocol specified by the manufacturer. ASMM index was calculated according to the method described by Baumgartner et al,11 summarizing fat- and bone-free mass of the legs and arms as assessed by wholebody DXA.
Bone mineral density Areal BMD at the LS (L1eL4, p.a.) and the proximal femur (total hip [tHip] region of interest) was determined by DXAa at baseline and after 18 months using standard protocols. Long-term reliability was 0.6% (coefficient of variation) as determined by 251 LS phantom scans conducted during the study period.
2048
W. Kemmler, S. von Stengel
-
-
-
-
-
Fig 1
Adapted flowchart of the present retrospective dose-response study.
Risk of CHD Framingham studyebased 10-year CHD risk was calculated according to Wilson et al.12 Parameters constituting 10-year CHD risk were age, diabetes status, smoking status, total cholesterol categories, low-density lipoprotein cholesterol categories, and blood pressure categories.
Blood parameters Blood was sampled after an overnight fast in the morning (7:00e9:00 AM) in a sitting position from an antecubital vein. Serum samples were centrifuged at 3000rpm for 20 minutes, and total cholesterol and low-density lipoprotein cholesterol levels were analyzed using test kits of Roche Diagnostics GmbH.b Blood pressure was determined in a sitting position after a 5-minute rest with an automatic oscillometric device.c
Falls “Falls” were defined according to the Prevention of Falls Network Europe group.13 Falls were recorded daily using fall calendars compiled by all the subjects. Fall calendars were collected and analyzed monthly. Fallers and nonresponders were also contacted by outcome assessors on a monthly basis.
Intervention In summary, the EG performed the high-intensity aerobic, balance, and resistance exercise program described below continuously over 18 months. To accurately determine subjects’
adherence, attendance lists were kept by the instructors. In addition, attendance, compliance, and motivation of the subjects were monitored by individual training protocols recorded by each participant. Protocols were collected and analyzed every 10 weeks. Apart from the intervention described below, all the participants were asked to maintain their habitual lifestyle. Exercise group The 18-month exercise program consisted of 2 supervised group sessions of 60 minutes and 2 nonsupervised home training sessions of z20 minutes described below. Group sessions The joint session consisted of 4 blocks: (1) endurance; (2) coordination; (3) isometric strength training, functional gymnastics, and flexibility; and (4) dynamic strength training. Each session started with 20 minutes of aerobic dancing at 70% to 85% maximum heart rate, including peak ground reaction forces z3body weight. More intense aerobic dance moves were regularly introduced to produce an increasing amount of highimpact loading. Exercises during the coordination sequence (5min) were performed under progressively unstable postural conditions (eg, reduced base, unstable base, and eyes closed) to enhance static and dynamic balance. The resistance sequence started with functional gymnastics/ floor exercises and stretching and isometric exercises and finished with dynamic exercises. During the functional gymnastic/ isometric exercise section, 1 or 2 sets of 10 to 15 isometric floor exercises that focused on trunk extensors, flexors, hip flexors, and leg abductors/adductors were performed with 6 to 8 seconds of maximum intensity. The participants’ reported rate of perceived www.archives-pmr.org
Exercise frequency and diseases
2049
Table 1 Baseline data (mean SD) of anthropometric parameters and risk factors in EGs and the “inactive” WCG Variable Age (y) Height (cm) Weight (kg) Total body fat (%) Age at menarche (y) Age at menopause (y) Energy intake (kJ/d)* Calcium uptake (mg/d)* Vitamin D uptake (mg/d)* Physical activity (index)y Exercise (min/wk) Serum vitamin-D3 (25-OH) level (ng/mL)z VO2peak (mL,kg1,min1)x Corticosteroids (5mg/d)/thyroxin (75mg/d) >6mo during lifetime (n per group) Smokers/former smokers (n per group)
HF-EG (nZ79)
LF-EG (nZ36)
WCG (nZ47)
68.63.9 69.53.8 69.54.5 161.55.9 162.76.6 159.75.4 67.49.9 69.913.1 69.210.7 36.25.8 36.06.3 37.95.8 14.01.6 13.51.6 13.91.2 49.15.7 48.26.7 49.55.5 66831654 66671803 66381667 895390 1002430 904425 3.85.1 3.74.3 4.04.7 4.21.2 4.30.8 4.61.1 73114 4587 76108 28.89.5 28.611.2 29.910.6 24.14.2
24.43.8
22.83.2
3 (4%)
1 (3%)
2 (6%)
3/15
0/7
1/5
Exercise frequency in the study groups After retrospectively subdividing (<2 vs 2 sessions/wk) the EG (nZ115), weekly exercise frequency averaged 2.660.39 (range, 2.05e3.71) sessions/wk in the HF-EG (nZ79) compared with 1.620.34 (1.04e1.99) sessions/wk in the LF-EG (nZ36). Exercise frequency of the inactive WCG (nZ47) averaged only 0.180.04 wellness sessions/wk (4e19 sessions in 18mo). Of relevance, low attendance rates in the LF-EG and the inactive WCG were not due to longer episodes of diseases/injuries that impact the study endpoints.
Statistical analysis
Abbreviations: 25-OH, 25-Hy25droxy-vitamin-D; VO2peak, peak oxygen consumption. * Four-day dietary analysis. y Based on a scale anchored from 1 (very low) to 7 (very high) according to a subjective assessment of professional, household, and recreational activities. z Cobas.b x Stepwise treadmill test to a voluntary maximum.
exertion was used to introduce more strenuous isometric exercises every 8 to 12 weeks to provide adequate exercise intensity contractions and enhance motivation. During the periods between exercises, flexibility exercises (20e30s of continuous stretching) were performed. Three dynamic exercises for upper-back and upper-limb muscle groups with 2 sets and 12 to 15 repetitions using elastic beltsd were carried out. In addition, 2 sets of 1-legged heel raises, front lunge, and leg abduction were conducted on a 25-cm platform in a circuit mode, with 1 minute of exercise and 1 minute of active rest with stretching exercise. Generally, time under tension for each dynamic resistance exercise was 2 seconds (concentric)e0 second (isometric)e2 seconds (eccentric); thus, 8 to 9 repetitions per leg could be performed during the 60 seconds. Subjects were generally asked to perform at a maximum e2 repetitions exertion level. Intensity was progressively increased by enlarging the amplitude of the movement, changing the velocity of concentric execution, and the introduction of more strenuous exercises. Home training session (EG only) The 20-minute nonsupervised home exercise training consisted of endurance, resistance, balance, and flexibility exercises from the joint workout. Home training exercises were regularly replaced by more intense exercises to consistently provide high exercise intensity. www.archives-pmr.org
Wellness control group As described above, the focus was exclusively on control group participants with low exercise frequency (<20 sessions/18mo, nZ47). Briefly, “exercises” of the WCG focused on well-being, with little or no impact on physical parameters.9,10,14 Sixty minutes of low-intensity physical activity and relaxation training were provided once a week for 10 weeks with 10-week breaks without training. Of importance, volume or intensity of the exercises was not increased during the 10-week blocks or through the study course.
For the present analysis, subjects of the EG were retrospectively allocated to 2 groups with different training frequencies (<2 vs 2 sessions/wk) according to their effective attendance during the 18-month interventional period. For a better interpretation of the results, the corresponding data of an inactive WCG (<20 sessions/ 18mo, nZ47) are also provided. To check for sufficient power to detect a clinically meaningful difference between both EGs, a power calculation based on BMD of the LS and ASMM was performed. In summary, power (1b) to identify a 2.0%2.5% difference between groups (aZ.05%) as significant was z90%. BMD, LBM, and ASMM were log-transformed to obtain normally distributed data. The 10-year CHD risk and fall number could not be transformed to normally distributed data and was therefore analyzed nonparametrically using Mann-Whitney U tests. Although fall parameters are ideally analyzed using more sophisticated methods (negative binominal regression13), this simple approach was selected for consistency of the analysis. Two independent analyses were performed to adequately address the hypotheses. Thus, with respect to the selected study endpoints, (1) the HF-EG (nZ79) was compared with the LF-EG (nZ36) and (2) the LF-EG was compared with the WCG (nZ47). Within-group changes from baseline to follow-up were calculated by using dependent t tests or Wilcoxon rank-sum tests. Effect sizes based on the absolute difference (SD) between baseline and 18-month follow-up were calculated using Cohen’s d.15 All the tests were 2-sided, with P<.05 considered statistically significant. Exact raw P values without further adjusting were presented. SPSS 19.0e was used for all statistical procedures.
Results Study endpoints Tables 2 and 3 provide an overview of the study results. However, each hypothesis will be addressed separately in this section.
2050 Table 2
W. Kemmler, S. von Stengel Changes in the HF-EG vs the LF-EG Mean SD
Variable
HF-EG (nZ79)
Absolute Difference Mean (95% CI)
t (z)
0.9020.121 0.9050.119 0.0020.019
0.9310.173 0.9540.185 0.0230.028
NT NT 0.021 (0.010e0.031)
NT NT 3.96
0.8330.096 0.8290.096 0.0040.013
0.8290.115 0.8350.116 0.0060.020
NT NT 0.015 (0.002e0.018)
NT NT 2.51
43.635.08 44.334.87 0.6931.50
NT NT 0.534 (0.006 to 1.074)
NT NT 1.97
17.602.36 17.762.26 0.1590.803
NT NT 0.402 (0.095e0.708)
1.031.21
0.951.36
10.544.18 8.553.66 2.003.13
10.484.35 8.533.29 1.944.14
LF-EG (nZ36)
P
Effect Size
0.000
NT NT 0.88
0.014
NT NT 0.59
0.053
NT NT 0.39
NT NT 2.60
0.011
NT NT 0.56
0.08 (0.44 to 0.60)
(0.34)
0.757
0.06
NT NT 0.06 (1.56 to 1.67)
NT NT (0.07)
0.943
NT NT 0.02
2
BMD at the LS (mg/cm ) Baseline 18 mo Difference BMD at the proximal femur (tHip) (mg/cm2) Baseline 18 mo Difference LBM (kg) Baseline 18 mo Difference ASMM (kg) Baseline 18 mo Difference Number of falls (n/person) Fall rate 10-year CHD risk according to Wilson et al12 (%) Baseline 18 mo Difference
45.114.87 45.275.34 0.1591.222 18.492.52 18.252.64 0.2420.610
NOTE. For the analysis, changes were adjusted to baseline values. Significance levels are given for between-group differences only. Abbreviations: CI, confidence interval; NT, not tested.
Hypothesis 1: osteopenia and sarcopenia parameters (HF-EG vs LF-EG) BMD at the LS and the proximal femur (tHip) increased significantly in the HF-EG (LS, 2.4%2.8%, P<.001; tHip, 0.7% 2.4%, PZ.021) and maintained or slightly decreased in the LFEG (LS, 0.3%2.2%, PZ.482; tHip, 0.5%1.6%, PZ.083). Between-group differences for both bone parameters were significant (LS, PZ.001; tHip, PZ.014). LBM increased significantly by 1.6%3.4% (P<.001) in the HF-EG and increased slightly by 0.3%2.6% in the LF-EG (PZ.454). ASMM increased nonsignificantly by 0.9%4.5% in the HF-EG (PZ.088) but decreased significantly in the LF-EG (1.3%3.2%, PZ.027). Between-group differences were significant for ASMM (PZ.011) and borderline nonsignificant for LBM (PZ.053). Thus, hypothesis 1 can be clearly confirmed for bone parameters and partially confirmed for muscle parameters.
Hypothesis 2: osteopenia and sarcopenia parameters (LF-EG vs WCG) To clearly determine the net effect of the LF-EG with respect to BMD, LBM, and ASMM, a comparison with data of the inactive WCG was conducted. Eighteen-month BMD changes among this group were 0.2%2.1% (PZ.550) for the LS and 0.6%2.1% (PZ.066) for the proximal femur. On comparing the LF-EG with the inactive WCG, no significant differences (LS, PZ.356; tHip, PZ.945) were determined for BMD.
Accordingly, only slight, nonsignificant effects in favor of the LF-EG were determined for LBM (LF-EG, 0.3%2.6% vs WCG, 0.1%2.8%; PZ.382) and ASMM (LF-EG, 1.3%3.2% vs WCG, 2.1%4.2%; PZ.389). Thus, hypothesis 2 can be confirmed.
Hypothesis 3: CHD risk Framingham study-based 10-year CHD risk according to Wilson et al12 decreased significantly by 1.94%4.14% (P<.001) in the HF-EG and by 2.00%3.13% (P<.001) in the LF-EG. No significant differences (PZ.943) were determined between the 2 EGs. However, because of the significant reductions in the 10-year CHD risk among the inactive WCG (1.043.09, PZ.025), differences between this group and both EGs were nonsignificant (P>.178). Thus, the results confirm hypothesis 3.
Hypothesis 4: falls Rate ratio (falls/person) and risk ratio (fallers/group; not given) did not differ significantly between the HF-EG (0.951.36 falls/ person) and the LF-EG (1.031.21 falls/person). Although the fall rate in the inactive WCG was more than 60% higher (1.691.95 falls/person) than in the larger HF-EG (PZ.015), differences between the LF-EG and the inactive WCG did not reach significance (PZ.065). However, the results clearly confirm hypothesis 4. www.archives-pmr.org
Exercise frequency and diseases Table 3
2051
Changes in the LF-EG vs the “inactive” WCG Mean SD
Variable
WCG (nZ47)
Absolute Difference Mean (95% CI)
t (z)
0.9020.121 0.9050.119 0.0020.019
0.9190.142 0.9170.140 0.0020.020
NT NT 0.004 (0.013 to 0.005)
NT NT 0.93
0.8330.096 0.8290.096 0.0040.013
0.8290.111 0.8240.113 0.0050.017
NT NT 0.000 (0.007 to 0.007)
NT NT 0.07
45.114.87 45.275.34 0.1591.222
43.524.93 43.444.87 0.0771.179
NT NT 0.236 (0.770 to 0.298)
NT NT 0.88
18.492.52 18.252.64 0.2420.610
17.702.43 17.322.35 0.3770.739
NT NT 0.134 (0.441 to 0.174)
LF-EG (nZ36)
P
Effect Size
0.356
NT NT 0.21
0.945
NT NT 0.07
0.382
NT NT 0.20
NT NT 0.87
0.389
NT NT 0.20
0.065
0.41
0.178
NT NT 0.31
2
BMD at the LS (mg/cm ) Baseline 18 mo Difference BMD at the proximal femur (tHip) (mg/cm2) Baseline 18 mo Difference LBM (kg) Baseline 18 mo Difference ASMM (kg) Baseline 18 mo Difference Number of falls (n/person) Fall rate 10-year CHD risk according to Wilson et al12 (%) Baseline 18 mo Difference
1.031.21
1.691.95
0.66 (0.04 to 1.36)
(1.72)
10.544.18 8.553.66 2.003.13
10.895.18 9.854.89 1.043.09
NT NT 0.96 (2.36 to 0.45)
NT NT (1.36)
NOTE. For the analysis, changes were adjusted to baseline values. Significance levels are given for between-group differences only. Abbreviations: CI, confidence interval; NT, not tested.
Discussion The purpose of the research was to evaluate the effect of 2 different exercise frequencies (that were obviously within a feasible and accepted range) on multiple risk factors and diseases of the elderly in an ambulatory group setting. The motivation for the project was based on the relevance for designing exercise programs able to favorably impact a range of important risk factors and diseases of elderly subjects, a cohort that predominantly1,2 features >1 disease or risk factor. Summing up, the results verify the hypothesis that bone and muscle have to be addressed with a frequency of at least 2 sessions/wk, while no corresponding differences were determined for falls and CHD risk factors. Few studies consequently focus on the effect of exercise frequency on clinical endpoints. With respect to bone (BMD), at least 7 studies5-7,16-19 were identified that did actually report heterogeneous results. It is interesting to note that only those trials6,7 that retrospectively structured their subgroups and provided intervention periods of >9 months reported positive results in favor of the more frequent exercise applications. In their meta-analysis, Peterson et al20 did not find any significant effects between low and high exercise frequency with respect to LBM. However, the weak point of this study is that exercise frequency barely differs between the trials (all 2e3 sessions/wk). However, some of the aforementioned “exercise and BMD studies” determined body composition by whole-body DXA.5,6,17-19 Again, only the 4-year study of Cussler et al,6 which retrospectively
www.archives-pmr.org
structured its subgroups according to the effective exercise frequency, reported significant differences for LBM (0.94 sessions/ wk, 0.11.6kg vs 2.11 sessions/wk, 0.91.2kg). Because exercise trials are normally underpowered for focusing on cardiac events (eg, myocardial infarction), most studies determined isolated CHD risk factors that had not been equally sensitive to exercise volume.21 We selected the 10-year CHD risk suggested by Wilson et al12 to determine the overall effect of exercise on CHD risk. However, because of this endpoint and the selected cohort, only 1 of our recent studies8 could be compared with the present study. This exercise trial,8 which also subdivided postmenopausal female exercisers according to their effective exercise frequency over 12 years (<2 vs 2 sessions/wk), reported comparable results, while changes in both groups were significantly more favorable than in an inactive control group. Although no dedicated exercise trial that focuses on the comparison of different exercise frequencies with respect to falls can be identified, a review of the present literature22 does not support the superiority of higher exercise frequencies (or volume). Indeed, studies applying identical (eg, Tai Chi23,24) or different types of exercise with lower (1e2 sessions/wk, eg, Voukelatos et al24 and Weerdesteyn et al25) versus higher exercise frequency (>2 sessions/wk, eg, Li et al,23 Buchner et al,26 and Skelton et al27) both comparably impact fall parameters. The significance of results of present studies is based on the following strong points: (a) the results were derived from a randomized controlled trial with a multicomponent protocol9,10 that positively impacted all the parameters addressed in this study.
2052 (b) Because of the adequate sample size among all groups, the statistical power of the study is high enough to detect relevant differences. (c) The study was longer than most other corresponding exercise studies.22,28,29 This is of importance not only because of the fact that short interventional periods may not reflect the full extent of potential changes (at least with respect to bone) but also because exercise protocols that impact bone initially become less effective after bone adaptation.30 Thus, an effect triggered during the sensitive initial phase does not guarantee a general long-term effectiveness. (d) The study cohort was a relatively homogeneous group of community-living women aged 65 to 80 years with no medications or diseases affecting bone or muscle metabolism.9,10 (e) Strong emphasis was placed on controlling for confounders and monitoring the subjects’ adherence to the study protocol. (f) Group sessions were continuously supervised, and the general intensity of the program was adequately adapted throughout the study course.
Study limitations From a statistical point of view, preplanned, randomized doseresponse studies with subsequent intention-to-treat analysis would have been preferable. However, experience shows that the fundamental weakness of this strategy is that a relevant number of subjects do not exercise with the prescribed frequency of the study arm to which they were allocated. For example, there is considerable variability in subjects’ attendance rate among 2 “preplanned” dose-response studies that focus on bone5,17 (50%e 100%,5 65%e100%17) within their subgroups. Including these subjects in the intention-to-treat analysis profoundly impacts the validation of the proper effect of exercise frequency on the given study endpoint. One may argue that total exercise volume and not exercise frequency may be the more relevant parameter in this topic. For walking, running, cycling, or other types of exercise that do not require relevant instruction, supervision, or equipment, total exercise volume may be indeed the more meaningful parameter, although effects vary depending on whether exercise is taken in 3 sessions of 40 minutes or 1 session of 120 minutes.29 With respect to exercise protocols specifically dedicated to the multimorbidity situation of the elderly,1 the increased demand for personal, material, and financial support results in the requirement of supervised ambulatory group session on specific locations. These classes are fixed with respect to length, rooms, instructors, and number of participants. Thus, exercise frequency, not total exercise volume, is the crucial factor in this setting, because exercise frequency fully reflects exercise volume when the exercise duration (mostly 45e60min in ambulatory classes) is fixed (but not vice versa). Finally, although the subjects’ compliance and attendance were strictly monitored using training logs and attendance lists, the exercise frequency of the unsupervised home training sessions may not be perfectly recorded by the participants.
Conclusions In summary, the results indicate that a rather high exercise frequency of 2 sessions/wk (on average over the year!) has to be applied to relevantly impact musculoskeletal parameters in the elderly, even if the exercise intensity is relatively high. Unfortunately, this exceeds the exercise practice of most (German) elderly
W. Kemmler, S. von Stengel persons,4 the cohort most vulnerable to musculoskeletal diseases. However, falls and CHD risk factors were affected equally by both protocols. While this result might not be generalizable across all exercise types and cohorts, other studies that used mixed exercise protocols applied with higher intensities6,7 fully confirmed our study results, at least with respect to musculoskeletal parameters, for the relevant ambulatory group setting and the vulnerable cohort of postmenopausal women. Thus, to efficiently prevent osteoporosis and sarcopenia with their enormous individual and socioeconomic impact by exercise, it is crucial to provide enough adequate intensity exercise programs and to motivate the elderly to visit exercise classes more frequently.
Perspectives With respect to the currently popular “dose-response” or “minimum effective dose” debate, it is essential for readers and researchers to consider the complexity of exercise training as a whole. In fact, the strong interaction of exercise parameters alone (eg, intensity vs duration vs frequency) irrespective of other confounding factors (eg, exercise principles, compliance, and status of the study participants) makes it very likely that “the” minimum effective dose of an isolated exercise parameter simply does not exist.
Suppliers a. Hologic, Inc, 35 Crosby Dr, Bedford, MA 01730. b. Roche Diagnostics GmbH, Sandhoferstrasse 116, 68305 Mannheim, Germany. c. Bosch, Bahnhofstraße 64, 72417 Jungingen, Germany. d. Thera-Band, 1245 Home Ave, Akron, OH 44310. e. SPSS, Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Keywords Bone; Coronary disease; Exercise; Fall, accidental; Muscle
Corresponding author Wolfgang Kemmler, PhD, Institute of Medical Physics, University of Erlangen-Nuremberg, Henkestrasse 91, 91052 Erlangen, Germany. E-mail address: [email protected]. de.
Acknowledgments We thank Michael Bebenek, PhD, Klaus Engelke, PhD, and Willi A. Kalender, PhD, MD, Institute of Medical Physics, FriedrichAlexander Universit€at Erlangen-N€ urnberg, for their advice and support in statistical and graphical details. Wolfgang Kemmler and Simon von Stengel contributed to the concept, design, analysis and interpretation of data, and preparation of the article.
References 1. Tesch-Ro¨mer C. Gesundheit im Alter: Schicksal, soziale Schicht oder Verhalten? In: Forschung BfBu, editor. Berlin: Gesundheitsberichterstattung des Bundes; 2007.
www.archives-pmr.org
Exercise frequency and diseases 2. Wiesner G, Bittner E. Zur Inzidenz und Pr€avalenz von Mehrfachkrankheiten in Deutschland. ArbeitsmedSozialmedUmweltmed 2005; 40:490-8. 3. Bauer JM, Sieber CC. Sarcopenia and frailty: a clinician’s controversial point of view. Exp Gerontol 2008;43:674-8. 4. Statistisches-Bundesamt. Gesundheit in Deutschland. Berlin: Gesundheitsberichterstattung des Bundes; 2006. 5. Bemben DA, Bemben MG. Dose-response effect of 40 weeks of resistance training on bone mineral density in older adults. Osteoporos Int 2011;22:179-86. 6. Cussler EC, Going SB, Houtkooper LB, et al. Exercise frequency and calcium intake predict 4-year bone changes in postmenopausal women. Osteoporos Int 2005;16:2129-41. 7. Kemmler W, Von Stengel S. Dose response effect of exercise frequency on bone mineral density in postmenopausal, osteopenic females [published online ahead of print November 28, 2012]. Scand J Med Sci Sports. 8. Von Stengel S, Kemmler W. Dosis-Wirkungsbeziehung der Trainingsh€aufigkeit im ambulanten Gruppenrahmen zur Pr€avention von metabolischen und kardialen-Erkrankungen €alterer Frauen. Deutsche Zeitschrift f€ ur Sportmedizin 2013;63:83-9. 9. Kemmler W, von Stengel S, Engelke K, et al. Exercise effects on bone mineral density, falls, coronary risk factors, and health care costs in older women: the randomized controlled Senior Fitness and Prevention (SEFIP) study. Arch Intern Med 2010;170:179-85. 10. Kemmler W, von Stengel S, Engelke K, et al. Exercise, body composition, and functional ability: a randomized controlled trial. Am J Prev Med 2010;38:279-87. 11. Baumgartner RN, Koehler KM, Gallagher D, et al. Epidemiology of sarcopenia among the elderly in New Mexico. Am J Epidemiol 1998; 147:755-63. 12. Wilson PW, D’Agostino RB, Levy D, et al. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-47. 13. Lamb SE, Jorstad-Stein EC, Hauer K, et al. Development of a common outcome data set for fall injury prevention trials: the Prevention of Falls Network Europe consensus. J Am Geriatr Soc 2005;53:1618-22. 14. Kemmler W, Von Stengel S, Engelke K, et al. Exercise decreases the risk of metabolic syndrome in elderly females. Med Sci Sports Exerc 2009;41:297-305. 15. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale: Erlbaum; 1988. 16. Ashe MC, Gorman E, Khan KM, et al. Does frequency of resistance training affect tibial cortical bone density in older women? A randomized controlled trial. Osteoporos Int 2013;24:623-32.
www.archives-pmr.org
2053 17. Bailey CA, Brooke-Wavell K. Optimum frequency of exercise for bone health: randomised controlled trial of a high-impact unilateral intervention. Bone 2010;46:1043-9. 18. Kemmler W, Riedel H. Ko¨rperliche Belastung und Osteoporose Einfluß einer 10monatigen Interventionsmaßnahme auf oss€are und extraoss€are Risikofaktoren einer Osteoporose. Dtsch Z Sportmed 1998;49:270-7. 19. Taaffe DR, Duret C, Wheeler S, et al. Once-weekly resistance exercise improves muscle strength and neuromuscular performance in older adults. J Am Geriatr Soc 1999;47:1208-14. 20. Peterson MD, Sen A, Gordon PM. Influence of resistance exercise on lean body mass in aging adults: a meta-analysis. Med Sci Sports Exerc 2011;43:249-58. 21. Slentz CA, Houmard JA, Kraus WE. Exercise, abdominal obesity, skeletal muscle, and metabolic risk: evidence for a dose response. Obesity (Silver Spring) 2009;17:S27-33. 22. Gillespie LD, Robertson MC, Gillespie WJ, et al. Interventions for preventing falls in older people living in the community. Cochrane Database Syst Rev 2009:CD007146. 23. Li F, Harmer P, Fisher KJ, et al. Tai Chi and fall reductions in older adults: a randomized controlled trial. J Gerontol A Biol Sci Med Sci 2005;60:187-94. 24. Voukelatos A, Cumming RG, Lord SR, et al. A randomized, controlled trial of tai chi for the prevention of falls: the Central Sydney tai chi trial. J Am Geriatr Soc 2007;55:1185-91. 25. Weerdesteyn V, Rijken H, Geurts AC, et al. A five-week exercise program can reduce falls and improve obstacle avoidance in the elderly. Gerontology 2006;52:131-41. 26. Buchner DM, Cress ME, de Lateur BJ, et al. The effect of strength and endurance training on gait, balance, fall risk, and health services use in community-living older adults. J Gerontol A Biol Sci Med Sci 1997; 52:M218-24. 27. Skelton D, Dinan S, Campbell MK, et al. Tailored group exercise (Falls management - FaME) reduces falls in community-dwelling older frequent fallers (an RCT). Age Aging 2005;34:636-9. 28. Howe TE, Shea B, Dawson LJ, et al. Exercise for preventing and treating osteoporosis in postmenopausal women. Cochrane Database Syst Rev 2011:CD000333. 29. Peterson MD, Rhea MR, Alvar BA. Applications of the dose-response for muscular strength development: a review of meta-analytic efficacy and reliability for designing training prescription. J Strength Cond Res 2005;19:950-8. 30. Kemmler W, Lauber D, von Stengel S, et al. Exercise effects on risk factors in early postmenopausal women: 3y EFOPS results. Med Sci Sports Exerc 2005;37:194-203.
ORIGINAL ARTICLE
Exercise Frequency, Health Risk Factors, and Diseases of the Elderly Wolfgang Kemmler, PhD, Simon von Stengel, PhD From the Institute of Medical Physics, University of Erlangen-Nurnberg (Friedrich-Alexander Universita€t Erlangen-Nu€rnberg), Erlangen, Germany.
Abstract Objective: To determine the effect of exercise frequency on various diseases and risk factors of the elderly. Design: Retrospective analysis of a randomized controlled 18-month exercise trial. Setting: University ambulatory group setting. Participants: Community-dwelling women aged 65 years (NZ162) in the area of Northern Bavaria. Intervention: Mixed, intense aerobic, resistance, and balance protocol for 18 months. Subjects were retrospectively subdivided into 2 groups according to their effective attendance over 18 months (>1e<2 vs 2e4 sessions/wk). Main Outcome Measures: Bone mineral density (BMD), lean body mass, appendicular skeletal muscle mass by dual-energy x-ray absorptiometry, Framingham study-based 10-year coronary heart disease (CHD) risk, and number of falls by calendar method. Results: Significant differences between the low-frequency exercise group (LF-EG) and the high-frequency exercise group (HF-EG) were observed for lumbar spine BMD (HF-EG, 2.4%2.8% vs LF-EG, 0.3%2.2%; P<.001) and proximal femur BMD (HF-EG, 2.4%2.8% vs LF-EG, 0.5%1.6%; PZ.014), lean body mass (1.6%3.4% vs 0.3%2.6%, PZ.053), and appendicular skeletal muscle mass (0.9%4.5% vs 1.3%3.2%, PZ.011). No differences between both exercise groups were observed for 10-year CHD risk (1.94%4.14% vs 2.00% 3.13%; PZ.943) and number of falls (0.951.36 vs 1.031.21 falls/person). Comparing the LF-EG with the less active control group (nZ47), only nonsignificant effects for fall number (PZ.065) and 10-year CHD risk (PZ.178) were evaluated. Conclusions: Although this result might not be generalizable across all exercise types and cohorts, it indicates that an overall exercise frequency of at least 2 sessions/wk may be crucial for impacting bone and muscle mass of elderly subjects. Archives of Physical Medicine and Rehabilitation 2013;94:2046-53 ª 2013 by the American Congress of Rehabilitation Medicine
In Germany, 3 of 4 people aged 65 years and older suffer from 2 and more diseases.1,2 It is generally known that physical activity and in particular exercise positively impacts most risks and is an alternative option to expensive medication with respect to disease prevention and treatment. However, because most exercise trials focus on 1 disease or on risk factors only, the most effective exercise protocol for impacting diverse relevant health problems of the elderly remains to be An audio podcast accompanies this article. Listen at www.archives-pmr.org. Presented in part at the “42. Deutscher Sport€arztekongress,” October 6e8, 2011, Frankfurt, Germany. The article has not been published elsewhere. No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.
established. A key decision for an effective exercise protocol is simply how often “exercise” should be performed. Unfortunately, no exercise trial exists yet that focuses on the dose-response effect of exercise frequency on parameters related to fracture risk, sarcopenia, and coronary heart disease (CHD) risk, which may be the most prominent predictors of autonomy and mortality of the elderly.3,4 Based on recent findings,5-8 it was assumed that an exercise frequency generating positive cardiac/metabolic adaptations was ineffective for impacting musculoskeletal parameters. By using data of the Senior Fitness and Prevention (SEFIP) study, this article focuses on the effect of 2 exercise frequencies on musculoskeletal parameters, CHD risk, and falls. The following 4 hypotheses were tested: 1. An average exercise frequency of 2 to 4 sessions/wk over 18 months is significantly more effective for impacting parameters
0003-9993/13/$36 - see front matter ª 2013 by the American Congress of Rehabilitation Medicine http://dx.doi.org/10.1016/j.apmr.2013.05.013
Exercise frequency and diseases related to osteopenia and sarcopenia compared with >1 to <2 sessions/wk. 2. An average exercise frequency of >1 to <2 sessions/wk over 18 months does not significantly impact parameters related to osteopenia and sarcopenia. 3. Both exercise protocols favorably impact CHD risk in elderly females, with no significant differences between groups. 4. Both exercise protocols favorably impact the number of falls in elderly females, with no significant differences between groups.
Methods This retrospective dose-response analysis used participant data of the SEFIP study, a fully registered (www.clinicaltrials.gov) randomized controlled exercise trial. The SEFIP study was conducted between 2006 and 2008 and focused on the effect of exercise on various risk factors among women aged 65 years and older.9,10 To determine the effect of exercise frequency on the given study endpoints, in the present analysis, the exercise group (EG) of the SEFIP study was retrospectively divided into 2 subgroups according to their effective exercise frequency over 18 months.
2047 With relevance for the present analysis, originally 246 women were included and allocated by an external statistician to 2 study arms using computer-generated block randomization: (1) exercise group (EG, nZ123) and (2) wellness control group (WCG, nZ123). After 18 months of intervention, 115 subjects of the EG and 112 subjects of the WCG could be included in the present analysis (details in Kemmler et al9,10). For the present analysis, the weekly attendance rate of the participants for the 18-month intervention was summed, and the EG was subdivided into 2 groups, one subgroup with an overall exercise frequency of <2 sessions/wk (low-frequency EG [LF-EG], nZ36) and another subgroup with an overall exercise frequency of 2 sessions/wk (high-frequency EG [HF-EG], nZ79). To adequately appraise changes among the EGs, the WCG was also subdivided into a more compliant subgroup (20 out of 40 wellness sessions/18mo) and an “inactive” subgroup (<20 sessions/18mo, nZ47). To generate a comparison between exercise and a “nontraining condition,” the following refers exclusively to the results of the “inactive WCG” (see fig 1). Table 1 lists the baseline characteristics of both EGs and the inactive WCG. There were no significant differences between groups with respect to anthropometric, metabolic, or lifestyle parameters that may have confounded the results.
Study endpoints Endpoints of this present dose-response analysis are as follows: Osteopenia (bone mineral density [BMD] at the lumbar spine [LS] and the proximal femur) Fall number (fall rate) Sarcopenia (lean body mass [LBM]; appendicular skeletal muscle mass [ASMM]) CHD risk (Framingham 10-year CHD risk)
Testing procedures Baseline and 18-month follow-up measurements were generally performed at the same time of the day (1h) in a fixed order and by the same researcher. All assessments were carried out in a blinded fashion.
Anthropometry Subjects For a detailed description of the recruitment process, the reader is referred to recent publications of the SEFIP study.9,10 Figure 1 shows the participant flow of the SEFIP study adapted for the present analysis. Briefly, subjects were eligible if they were aged 65 years or older, of Caucasian race, and living independently in the community. Criteria for exclusion were (1) diseases/medication affecting bone metabolism or falls risk, (2) history of stroke or cardiac events, (3) inflammatory diseases, (4) secondary osteoporosis, (5) participation in exercise studies during the last 2 years, and/or (6) athletic history during the last decade.
List of abbreviations: ASMM BMD CHD DXA EG HF-EG LBM LF-EG LS SEFIP tHip WCG
appendicular skeletal muscle mass bone mineral density coronary heart disease dual-energy x-ray absorptiometry exercise group high-frequency exercise group lean body mass low-frequency exercise group lumbar spine Senior Fitness and Prevention (study) total Hip wellness control group
www.archives-pmr.org
Height was determined with a stadiometer, and weight was measured with minimal clothing on digital scales. Body mass index was calculated as weight divided by height squared (kg/m2). Waist circumference was determined as the minimum circumference between the distal end of the rib cage and the top of the iliac crest along the midaxillary line.
Lean body mass, appendicular skeletal muscle mass LBM was assessed with dual-energy x-ray absorptiometry (DXA) technique using the whole-body standard protocol specified by the manufacturer. ASMM index was calculated according to the method described by Baumgartner et al,11 summarizing fat- and bone-free mass of the legs and arms as assessed by wholebody DXA.
Bone mineral density Areal BMD at the LS (L1eL4, p.a.) and the proximal femur (total hip [tHip] region of interest) was determined by DXAa at baseline and after 18 months using standard protocols. Long-term reliability was 0.6% (coefficient of variation) as determined by 251 LS phantom scans conducted during the study period.
2048
W. Kemmler, S. von Stengel
-
-
-
-
-
Fig 1
Adapted flowchart of the present retrospective dose-response study.
Risk of CHD Framingham studyebased 10-year CHD risk was calculated according to Wilson et al.12 Parameters constituting 10-year CHD risk were age, diabetes status, smoking status, total cholesterol categories, low-density lipoprotein cholesterol categories, and blood pressure categories.
Blood parameters Blood was sampled after an overnight fast in the morning (7:00e9:00 AM) in a sitting position from an antecubital vein. Serum samples were centrifuged at 3000rpm for 20 minutes, and total cholesterol and low-density lipoprotein cholesterol levels were analyzed using test kits of Roche Diagnostics GmbH.b Blood pressure was determined in a sitting position after a 5-minute rest with an automatic oscillometric device.c
Falls “Falls” were defined according to the Prevention of Falls Network Europe group.13 Falls were recorded daily using fall calendars compiled by all the subjects. Fall calendars were collected and analyzed monthly. Fallers and nonresponders were also contacted by outcome assessors on a monthly basis.
Intervention In summary, the EG performed the high-intensity aerobic, balance, and resistance exercise program described below continuously over 18 months. To accurately determine subjects’
adherence, attendance lists were kept by the instructors. In addition, attendance, compliance, and motivation of the subjects were monitored by individual training protocols recorded by each participant. Protocols were collected and analyzed every 10 weeks. Apart from the intervention described below, all the participants were asked to maintain their habitual lifestyle. Exercise group The 18-month exercise program consisted of 2 supervised group sessions of 60 minutes and 2 nonsupervised home training sessions of z20 minutes described below. Group sessions The joint session consisted of 4 blocks: (1) endurance; (2) coordination; (3) isometric strength training, functional gymnastics, and flexibility; and (4) dynamic strength training. Each session started with 20 minutes of aerobic dancing at 70% to 85% maximum heart rate, including peak ground reaction forces z3body weight. More intense aerobic dance moves were regularly introduced to produce an increasing amount of highimpact loading. Exercises during the coordination sequence (5min) were performed under progressively unstable postural conditions (eg, reduced base, unstable base, and eyes closed) to enhance static and dynamic balance. The resistance sequence started with functional gymnastics/ floor exercises and stretching and isometric exercises and finished with dynamic exercises. During the functional gymnastic/ isometric exercise section, 1 or 2 sets of 10 to 15 isometric floor exercises that focused on trunk extensors, flexors, hip flexors, and leg abductors/adductors were performed with 6 to 8 seconds of maximum intensity. The participants’ reported rate of perceived www.archives-pmr.org
Exercise frequency and diseases
2049
Table 1 Baseline data (mean SD) of anthropometric parameters and risk factors in EGs and the “inactive” WCG Variable Age (y) Height (cm) Weight (kg) Total body fat (%) Age at menarche (y) Age at menopause (y) Energy intake (kJ/d)* Calcium uptake (mg/d)* Vitamin D uptake (mg/d)* Physical activity (index)y Exercise (min/wk) Serum vitamin-D3 (25-OH) level (ng/mL)z VO2peak (mL,kg1,min1)x Corticosteroids (5mg/d)/thyroxin (75mg/d) >6mo during lifetime (n per group) Smokers/former smokers (n per group)
HF-EG (nZ79)
LF-EG (nZ36)
WCG (nZ47)
68.63.9 69.53.8 69.54.5 161.55.9 162.76.6 159.75.4 67.49.9 69.913.1 69.210.7 36.25.8 36.06.3 37.95.8 14.01.6 13.51.6 13.91.2 49.15.7 48.26.7 49.55.5 66831654 66671803 66381667 895390 1002430 904425 3.85.1 3.74.3 4.04.7 4.21.2 4.30.8 4.61.1 73114 4587 76108 28.89.5 28.611.2 29.910.6 24.14.2
24.43.8
22.83.2
3 (4%)
1 (3%)
2 (6%)
3/15
0/7
1/5
Exercise frequency in the study groups After retrospectively subdividing (<2 vs 2 sessions/wk) the EG (nZ115), weekly exercise frequency averaged 2.660.39 (range, 2.05e3.71) sessions/wk in the HF-EG (nZ79) compared with 1.620.34 (1.04e1.99) sessions/wk in the LF-EG (nZ36). Exercise frequency of the inactive WCG (nZ47) averaged only 0.180.04 wellness sessions/wk (4e19 sessions in 18mo). Of relevance, low attendance rates in the LF-EG and the inactive WCG were not due to longer episodes of diseases/injuries that impact the study endpoints.
Statistical analysis
Abbreviations: 25-OH, 25-Hy25droxy-vitamin-D; VO2peak, peak oxygen consumption. * Four-day dietary analysis. y Based on a scale anchored from 1 (very low) to 7 (very high) according to a subjective assessment of professional, household, and recreational activities. z Cobas.b x Stepwise treadmill test to a voluntary maximum.
exertion was used to introduce more strenuous isometric exercises every 8 to 12 weeks to provide adequate exercise intensity contractions and enhance motivation. During the periods between exercises, flexibility exercises (20e30s of continuous stretching) were performed. Three dynamic exercises for upper-back and upper-limb muscle groups with 2 sets and 12 to 15 repetitions using elastic beltsd were carried out. In addition, 2 sets of 1-legged heel raises, front lunge, and leg abduction were conducted on a 25-cm platform in a circuit mode, with 1 minute of exercise and 1 minute of active rest with stretching exercise. Generally, time under tension for each dynamic resistance exercise was 2 seconds (concentric)e0 second (isometric)e2 seconds (eccentric); thus, 8 to 9 repetitions per leg could be performed during the 60 seconds. Subjects were generally asked to perform at a maximum e2 repetitions exertion level. Intensity was progressively increased by enlarging the amplitude of the movement, changing the velocity of concentric execution, and the introduction of more strenuous exercises. Home training session (EG only) The 20-minute nonsupervised home exercise training consisted of endurance, resistance, balance, and flexibility exercises from the joint workout. Home training exercises were regularly replaced by more intense exercises to consistently provide high exercise intensity. www.archives-pmr.org
Wellness control group As described above, the focus was exclusively on control group participants with low exercise frequency (<20 sessions/18mo, nZ47). Briefly, “exercises” of the WCG focused on well-being, with little or no impact on physical parameters.9,10,14 Sixty minutes of low-intensity physical activity and relaxation training were provided once a week for 10 weeks with 10-week breaks without training. Of importance, volume or intensity of the exercises was not increased during the 10-week blocks or through the study course.
For the present analysis, subjects of the EG were retrospectively allocated to 2 groups with different training frequencies (<2 vs 2 sessions/wk) according to their effective attendance during the 18-month interventional period. For a better interpretation of the results, the corresponding data of an inactive WCG (<20 sessions/ 18mo, nZ47) are also provided. To check for sufficient power to detect a clinically meaningful difference between both EGs, a power calculation based on BMD of the LS and ASMM was performed. In summary, power (1b) to identify a 2.0%2.5% difference between groups (aZ.05%) as significant was z90%. BMD, LBM, and ASMM were log-transformed to obtain normally distributed data. The 10-year CHD risk and fall number could not be transformed to normally distributed data and was therefore analyzed nonparametrically using Mann-Whitney U tests. Although fall parameters are ideally analyzed using more sophisticated methods (negative binominal regression13), this simple approach was selected for consistency of the analysis. Two independent analyses were performed to adequately address the hypotheses. Thus, with respect to the selected study endpoints, (1) the HF-EG (nZ79) was compared with the LF-EG (nZ36) and (2) the LF-EG was compared with the WCG (nZ47). Within-group changes from baseline to follow-up were calculated by using dependent t tests or Wilcoxon rank-sum tests. Effect sizes based on the absolute difference (SD) between baseline and 18-month follow-up were calculated using Cohen’s d.15 All the tests were 2-sided, with P<.05 considered statistically significant. Exact raw P values without further adjusting were presented. SPSS 19.0e was used for all statistical procedures.
Results Study endpoints Tables 2 and 3 provide an overview of the study results. However, each hypothesis will be addressed separately in this section.
2050 Table 2
W. Kemmler, S. von Stengel Changes in the HF-EG vs the LF-EG Mean SD
Variable
HF-EG (nZ79)
Absolute Difference Mean (95% CI)
t (z)
0.9020.121 0.9050.119 0.0020.019
0.9310.173 0.9540.185 0.0230.028
NT NT 0.021 (0.010e0.031)
NT NT 3.96
0.8330.096 0.8290.096 0.0040.013
0.8290.115 0.8350.116 0.0060.020
NT NT 0.015 (0.002e0.018)
NT NT 2.51
43.635.08 44.334.87 0.6931.50
NT NT 0.534 (0.006 to 1.074)
NT NT 1.97
17.602.36 17.762.26 0.1590.803
NT NT 0.402 (0.095e0.708)
1.031.21
0.951.36
10.544.18 8.553.66 2.003.13
10.484.35 8.533.29 1.944.14
LF-EG (nZ36)
P
Effect Size
0.000
NT NT 0.88
0.014
NT NT 0.59
0.053
NT NT 0.39
NT NT 2.60
0.011
NT NT 0.56
0.08 (0.44 to 0.60)
(0.34)
0.757
0.06
NT NT 0.06 (1.56 to 1.67)
NT NT (0.07)
0.943
NT NT 0.02
2
BMD at the LS (mg/cm ) Baseline 18 mo Difference BMD at the proximal femur (tHip) (mg/cm2) Baseline 18 mo Difference LBM (kg) Baseline 18 mo Difference ASMM (kg) Baseline 18 mo Difference Number of falls (n/person) Fall rate 10-year CHD risk according to Wilson et al12 (%) Baseline 18 mo Difference
45.114.87 45.275.34 0.1591.222 18.492.52 18.252.64 0.2420.610
NOTE. For the analysis, changes were adjusted to baseline values. Significance levels are given for between-group differences only. Abbreviations: CI, confidence interval; NT, not tested.
Hypothesis 1: osteopenia and sarcopenia parameters (HF-EG vs LF-EG) BMD at the LS and the proximal femur (tHip) increased significantly in the HF-EG (LS, 2.4%2.8%, P<.001; tHip, 0.7% 2.4%, PZ.021) and maintained or slightly decreased in the LFEG (LS, 0.3%2.2%, PZ.482; tHip, 0.5%1.6%, PZ.083). Between-group differences for both bone parameters were significant (LS, PZ.001; tHip, PZ.014). LBM increased significantly by 1.6%3.4% (P<.001) in the HF-EG and increased slightly by 0.3%2.6% in the LF-EG (PZ.454). ASMM increased nonsignificantly by 0.9%4.5% in the HF-EG (PZ.088) but decreased significantly in the LF-EG (1.3%3.2%, PZ.027). Between-group differences were significant for ASMM (PZ.011) and borderline nonsignificant for LBM (PZ.053). Thus, hypothesis 1 can be clearly confirmed for bone parameters and partially confirmed for muscle parameters.
Hypothesis 2: osteopenia and sarcopenia parameters (LF-EG vs WCG) To clearly determine the net effect of the LF-EG with respect to BMD, LBM, and ASMM, a comparison with data of the inactive WCG was conducted. Eighteen-month BMD changes among this group were 0.2%2.1% (PZ.550) for the LS and 0.6%2.1% (PZ.066) for the proximal femur. On comparing the LF-EG with the inactive WCG, no significant differences (LS, PZ.356; tHip, PZ.945) were determined for BMD.
Accordingly, only slight, nonsignificant effects in favor of the LF-EG were determined for LBM (LF-EG, 0.3%2.6% vs WCG, 0.1%2.8%; PZ.382) and ASMM (LF-EG, 1.3%3.2% vs WCG, 2.1%4.2%; PZ.389). Thus, hypothesis 2 can be confirmed.
Hypothesis 3: CHD risk Framingham study-based 10-year CHD risk according to Wilson et al12 decreased significantly by 1.94%4.14% (P<.001) in the HF-EG and by 2.00%3.13% (P<.001) in the LF-EG. No significant differences (PZ.943) were determined between the 2 EGs. However, because of the significant reductions in the 10-year CHD risk among the inactive WCG (1.043.09, PZ.025), differences between this group and both EGs were nonsignificant (P>.178). Thus, the results confirm hypothesis 3.
Hypothesis 4: falls Rate ratio (falls/person) and risk ratio (fallers/group; not given) did not differ significantly between the HF-EG (0.951.36 falls/ person) and the LF-EG (1.031.21 falls/person). Although the fall rate in the inactive WCG was more than 60% higher (1.691.95 falls/person) than in the larger HF-EG (PZ.015), differences between the LF-EG and the inactive WCG did not reach significance (PZ.065). However, the results clearly confirm hypothesis 4. www.archives-pmr.org
Exercise frequency and diseases Table 3
2051
Changes in the LF-EG vs the “inactive” WCG Mean SD
Variable
WCG (nZ47)
Absolute Difference Mean (95% CI)
t (z)
0.9020.121 0.9050.119 0.0020.019
0.9190.142 0.9170.140 0.0020.020
NT NT 0.004 (0.013 to 0.005)
NT NT 0.93
0.8330.096 0.8290.096 0.0040.013
0.8290.111 0.8240.113 0.0050.017
NT NT 0.000 (0.007 to 0.007)
NT NT 0.07
45.114.87 45.275.34 0.1591.222
43.524.93 43.444.87 0.0771.179
NT NT 0.236 (0.770 to 0.298)
NT NT 0.88
18.492.52 18.252.64 0.2420.610
17.702.43 17.322.35 0.3770.739
NT NT 0.134 (0.441 to 0.174)
LF-EG (nZ36)
P
Effect Size
0.356
NT NT 0.21
0.945
NT NT 0.07
0.382
NT NT 0.20
NT NT 0.87
0.389
NT NT 0.20
0.065
0.41
0.178
NT NT 0.31
2
BMD at the LS (mg/cm ) Baseline 18 mo Difference BMD at the proximal femur (tHip) (mg/cm2) Baseline 18 mo Difference LBM (kg) Baseline 18 mo Difference ASMM (kg) Baseline 18 mo Difference Number of falls (n/person) Fall rate 10-year CHD risk according to Wilson et al12 (%) Baseline 18 mo Difference
1.031.21
1.691.95
0.66 (0.04 to 1.36)
(1.72)
10.544.18 8.553.66 2.003.13
10.895.18 9.854.89 1.043.09
NT NT 0.96 (2.36 to 0.45)
NT NT (1.36)
NOTE. For the analysis, changes were adjusted to baseline values. Significance levels are given for between-group differences only. Abbreviations: CI, confidence interval; NT, not tested.
Discussion The purpose of the research was to evaluate the effect of 2 different exercise frequencies (that were obviously within a feasible and accepted range) on multiple risk factors and diseases of the elderly in an ambulatory group setting. The motivation for the project was based on the relevance for designing exercise programs able to favorably impact a range of important risk factors and diseases of elderly subjects, a cohort that predominantly1,2 features >1 disease or risk factor. Summing up, the results verify the hypothesis that bone and muscle have to be addressed with a frequency of at least 2 sessions/wk, while no corresponding differences were determined for falls and CHD risk factors. Few studies consequently focus on the effect of exercise frequency on clinical endpoints. With respect to bone (BMD), at least 7 studies5-7,16-19 were identified that did actually report heterogeneous results. It is interesting to note that only those trials6,7 that retrospectively structured their subgroups and provided intervention periods of >9 months reported positive results in favor of the more frequent exercise applications. In their meta-analysis, Peterson et al20 did not find any significant effects between low and high exercise frequency with respect to LBM. However, the weak point of this study is that exercise frequency barely differs between the trials (all 2e3 sessions/wk). However, some of the aforementioned “exercise and BMD studies” determined body composition by whole-body DXA.5,6,17-19 Again, only the 4-year study of Cussler et al,6 which retrospectively
www.archives-pmr.org
structured its subgroups according to the effective exercise frequency, reported significant differences for LBM (0.94 sessions/ wk, 0.11.6kg vs 2.11 sessions/wk, 0.91.2kg). Because exercise trials are normally underpowered for focusing on cardiac events (eg, myocardial infarction), most studies determined isolated CHD risk factors that had not been equally sensitive to exercise volume.21 We selected the 10-year CHD risk suggested by Wilson et al12 to determine the overall effect of exercise on CHD risk. However, because of this endpoint and the selected cohort, only 1 of our recent studies8 could be compared with the present study. This exercise trial,8 which also subdivided postmenopausal female exercisers according to their effective exercise frequency over 12 years (<2 vs 2 sessions/wk), reported comparable results, while changes in both groups were significantly more favorable than in an inactive control group. Although no dedicated exercise trial that focuses on the comparison of different exercise frequencies with respect to falls can be identified, a review of the present literature22 does not support the superiority of higher exercise frequencies (or volume). Indeed, studies applying identical (eg, Tai Chi23,24) or different types of exercise with lower (1e2 sessions/wk, eg, Voukelatos et al24 and Weerdesteyn et al25) versus higher exercise frequency (>2 sessions/wk, eg, Li et al,23 Buchner et al,26 and Skelton et al27) both comparably impact fall parameters. The significance of results of present studies is based on the following strong points: (a) the results were derived from a randomized controlled trial with a multicomponent protocol9,10 that positively impacted all the parameters addressed in this study.
2052 (b) Because of the adequate sample size among all groups, the statistical power of the study is high enough to detect relevant differences. (c) The study was longer than most other corresponding exercise studies.22,28,29 This is of importance not only because of the fact that short interventional periods may not reflect the full extent of potential changes (at least with respect to bone) but also because exercise protocols that impact bone initially become less effective after bone adaptation.30 Thus, an effect triggered during the sensitive initial phase does not guarantee a general long-term effectiveness. (d) The study cohort was a relatively homogeneous group of community-living women aged 65 to 80 years with no medications or diseases affecting bone or muscle metabolism.9,10 (e) Strong emphasis was placed on controlling for confounders and monitoring the subjects’ adherence to the study protocol. (f) Group sessions were continuously supervised, and the general intensity of the program was adequately adapted throughout the study course.
Study limitations From a statistical point of view, preplanned, randomized doseresponse studies with subsequent intention-to-treat analysis would have been preferable. However, experience shows that the fundamental weakness of this strategy is that a relevant number of subjects do not exercise with the prescribed frequency of the study arm to which they were allocated. For example, there is considerable variability in subjects’ attendance rate among 2 “preplanned” dose-response studies that focus on bone5,17 (50%e 100%,5 65%e100%17) within their subgroups. Including these subjects in the intention-to-treat analysis profoundly impacts the validation of the proper effect of exercise frequency on the given study endpoint. One may argue that total exercise volume and not exercise frequency may be the more relevant parameter in this topic. For walking, running, cycling, or other types of exercise that do not require relevant instruction, supervision, or equipment, total exercise volume may be indeed the more meaningful parameter, although effects vary depending on whether exercise is taken in 3 sessions of 40 minutes or 1 session of 120 minutes.29 With respect to exercise protocols specifically dedicated to the multimorbidity situation of the elderly,1 the increased demand for personal, material, and financial support results in the requirement of supervised ambulatory group session on specific locations. These classes are fixed with respect to length, rooms, instructors, and number of participants. Thus, exercise frequency, not total exercise volume, is the crucial factor in this setting, because exercise frequency fully reflects exercise volume when the exercise duration (mostly 45e60min in ambulatory classes) is fixed (but not vice versa). Finally, although the subjects’ compliance and attendance were strictly monitored using training logs and attendance lists, the exercise frequency of the unsupervised home training sessions may not be perfectly recorded by the participants.
Conclusions In summary, the results indicate that a rather high exercise frequency of 2 sessions/wk (on average over the year!) has to be applied to relevantly impact musculoskeletal parameters in the elderly, even if the exercise intensity is relatively high. Unfortunately, this exceeds the exercise practice of most (German) elderly
W. Kemmler, S. von Stengel persons,4 the cohort most vulnerable to musculoskeletal diseases. However, falls and CHD risk factors were affected equally by both protocols. While this result might not be generalizable across all exercise types and cohorts, other studies that used mixed exercise protocols applied with higher intensities6,7 fully confirmed our study results, at least with respect to musculoskeletal parameters, for the relevant ambulatory group setting and the vulnerable cohort of postmenopausal women. Thus, to efficiently prevent osteoporosis and sarcopenia with their enormous individual and socioeconomic impact by exercise, it is crucial to provide enough adequate intensity exercise programs and to motivate the elderly to visit exercise classes more frequently.
Perspectives With respect to the currently popular “dose-response” or “minimum effective dose” debate, it is essential for readers and researchers to consider the complexity of exercise training as a whole. In fact, the strong interaction of exercise parameters alone (eg, intensity vs duration vs frequency) irrespective of other confounding factors (eg, exercise principles, compliance, and status of the study participants) makes it very likely that “the” minimum effective dose of an isolated exercise parameter simply does not exist.
Suppliers a. Hologic, Inc, 35 Crosby Dr, Bedford, MA 01730. b. Roche Diagnostics GmbH, Sandhoferstrasse 116, 68305 Mannheim, Germany. c. Bosch, Bahnhofstraße 64, 72417 Jungingen, Germany. d. Thera-Band, 1245 Home Ave, Akron, OH 44310. e. SPSS, Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.
Keywords Bone; Coronary disease; Exercise; Fall, accidental; Muscle
Corresponding author Wolfgang Kemmler, PhD, Institute of Medical Physics, University of Erlangen-Nuremberg, Henkestrasse 91, 91052 Erlangen, Germany. E-mail address: [email protected]. de.
Acknowledgments We thank Michael Bebenek, PhD, Klaus Engelke, PhD, and Willi A. Kalender, PhD, MD, Institute of Medical Physics, FriedrichAlexander Universit€at Erlangen-N€ urnberg, for their advice and support in statistical and graphical details. Wolfgang Kemmler and Simon von Stengel contributed to the concept, design, analysis and interpretation of data, and preparation of the article.
References 1. Tesch-Ro¨mer C. Gesundheit im Alter: Schicksal, soziale Schicht oder Verhalten? In: Forschung BfBu, editor. Berlin: Gesundheitsberichterstattung des Bundes; 2007.
www.archives-pmr.org
Exercise frequency and diseases 2. Wiesner G, Bittner E. Zur Inzidenz und Pr€avalenz von Mehrfachkrankheiten in Deutschland. ArbeitsmedSozialmedUmweltmed 2005; 40:490-8. 3. Bauer JM, Sieber CC. Sarcopenia and frailty: a clinician’s controversial point of view. Exp Gerontol 2008;43:674-8. 4. Statistisches-Bundesamt. Gesundheit in Deutschland. Berlin: Gesundheitsberichterstattung des Bundes; 2006. 5. Bemben DA, Bemben MG. Dose-response effect of 40 weeks of resistance training on bone mineral density in older adults. Osteoporos Int 2011;22:179-86. 6. Cussler EC, Going SB, Houtkooper LB, et al. Exercise frequency and calcium intake predict 4-year bone changes in postmenopausal women. Osteoporos Int 2005;16:2129-41. 7. Kemmler W, Von Stengel S. Dose response effect of exercise frequency on bone mineral density in postmenopausal, osteopenic females [published online ahead of print November 28, 2012]. Scand J Med Sci Sports. 8. Von Stengel S, Kemmler W. Dosis-Wirkungsbeziehung der Trainingsh€aufigkeit im ambulanten Gruppenrahmen zur Pr€avention von metabolischen und kardialen-Erkrankungen €alterer Frauen. Deutsche Zeitschrift f€ ur Sportmedizin 2013;63:83-9. 9. Kemmler W, von Stengel S, Engelke K, et al. Exercise effects on bone mineral density, falls, coronary risk factors, and health care costs in older women: the randomized controlled Senior Fitness and Prevention (SEFIP) study. Arch Intern Med 2010;170:179-85. 10. Kemmler W, von Stengel S, Engelke K, et al. Exercise, body composition, and functional ability: a randomized controlled trial. Am J Prev Med 2010;38:279-87. 11. Baumgartner RN, Koehler KM, Gallagher D, et al. Epidemiology of sarcopenia among the elderly in New Mexico. Am J Epidemiol 1998; 147:755-63. 12. Wilson PW, D’Agostino RB, Levy D, et al. Prediction of coronary heart disease using risk factor categories. Circulation 1998;97:1837-47. 13. Lamb SE, Jorstad-Stein EC, Hauer K, et al. Development of a common outcome data set for fall injury prevention trials: the Prevention of Falls Network Europe consensus. J Am Geriatr Soc 2005;53:1618-22. 14. Kemmler W, Von Stengel S, Engelke K, et al. Exercise decreases the risk of metabolic syndrome in elderly females. Med Sci Sports Exerc 2009;41:297-305. 15. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed. Hillsdale: Erlbaum; 1988. 16. Ashe MC, Gorman E, Khan KM, et al. Does frequency of resistance training affect tibial cortical bone density in older women? A randomized controlled trial. Osteoporos Int 2013;24:623-32.
www.archives-pmr.org
2053 17. Bailey CA, Brooke-Wavell K. Optimum frequency of exercise for bone health: randomised controlled trial of a high-impact unilateral intervention. Bone 2010;46:1043-9. 18. Kemmler W, Riedel H. Ko¨rperliche Belastung und Osteoporose Einfluß einer 10monatigen Interventionsmaßnahme auf oss€are und extraoss€are Risikofaktoren einer Osteoporose. Dtsch Z Sportmed 1998;49:270-7. 19. Taaffe DR, Duret C, Wheeler S, et al. Once-weekly resistance exercise improves muscle strength and neuromuscular performance in older adults. J Am Geriatr Soc 1999;47:1208-14. 20. Peterson MD, Sen A, Gordon PM. Influence of resistance exercise on lean body mass in aging adults: a meta-analysis. Med Sci Sports Exerc 2011;43:249-58. 21. Slentz CA, Houmard JA, Kraus WE. Exercise, abdominal obesity, skeletal muscle, and metabolic risk: evidence for a dose response. Obesity (Silver Spring) 2009;17:S27-33. 22. Gillespie LD, Robertson MC, Gillespie WJ, et al. Interventions for preventing falls in older people living in the community. Cochrane Database Syst Rev 2009:CD007146. 23. Li F, Harmer P, Fisher KJ, et al. Tai Chi and fall reductions in older adults: a randomized controlled trial. J Gerontol A Biol Sci Med Sci 2005;60:187-94. 24. Voukelatos A, Cumming RG, Lord SR, et al. A randomized, controlled trial of tai chi for the prevention of falls: the Central Sydney tai chi trial. J Am Geriatr Soc 2007;55:1185-91. 25. Weerdesteyn V, Rijken H, Geurts AC, et al. A five-week exercise program can reduce falls and improve obstacle avoidance in the elderly. Gerontology 2006;52:131-41. 26. Buchner DM, Cress ME, de Lateur BJ, et al. The effect of strength and endurance training on gait, balance, fall risk, and health services use in community-living older adults. J Gerontol A Biol Sci Med Sci 1997; 52:M218-24. 27. Skelton D, Dinan S, Campbell MK, et al. Tailored group exercise (Falls management - FaME) reduces falls in community-dwelling older frequent fallers (an RCT). Age Aging 2005;34:636-9. 28. Howe TE, Shea B, Dawson LJ, et al. Exercise for preventing and treating osteoporosis in postmenopausal women. Cochrane Database Syst Rev 2011:CD000333. 29. Peterson MD, Rhea MR, Alvar BA. Applications of the dose-response for muscular strength development: a review of meta-analytic efficacy and reliability for designing training prescription. J Strength Cond Res 2005;19:950-8. 30. Kemmler W, Lauber D, von Stengel S, et al. Exercise effects on risk factors in early postmenopausal women: 3y EFOPS results. Med Sci Sports Exerc 2005;37:194-203.