Multicomponent recreational team handball training improves global health status in postmenopausal women at the long term – A randomised controlled trial

ABSTRACT We studied the long-term effects of a multicomponent exercise training protocol (recreational team handball training, RTH) on global health status in inactive postmenopausal women. Participants (n = 45; age 65 ± 6 years, stature 157 ± 6 cm, body mass 66.2 ± 9.4 kg, fat mass 41.4 ± 5.5%, VO2peak 25.7 ± 3.6 mL/min/kg) were randomised into a control group (CG; n = 14) and a multicomponent exercise training group (EXG; n = 31, performing two to three weekly 60-min RTH sessions). Attendance was 2.0 ± 0.4 sessions/week (first 16 weeks) and 1.4 ± 0.5 (following 20 weeks) and mean heart rate (HR) loading was 77 and 79% of maximal HR (p = .002) for the first 16 and the following 20 weeks, respectively. Cardiovascular, bone, metabolic health, body composition and physical fitness markers were evaluated at baseline, and after 16 and 36 weeks. An interaction (p ≤ .046) was shown for the 2-h oral glucose tolerance test, HDL, Yo-Yo intermittent endurance level 1 test (YYIE1) and knee strength, in favour of EXG. At 36 weeks, YYIE1 and knee strength were higher (p ≤ .038) for EXG vs CG. Also, within-group improvements (p ≤ .043) were observed after 36 weeks for EXG in VO2peak, lumbar spine bone mineral density, lumbar spine bone mineral content, P1NP, osteocalcin, total cholesterol, HDL, LDL, body mass, android fat mass, YYIE1, knee strength, handgrip strength and postural balance. At 36 comparatively to 16 weeks, EXG showed an increase (p ≤ .036) in fasting blood glucose, HDL, knee strength and handgrip strength, and a decrease (p ≤ .025) in LDL. Collectively, this multicomponent exercise training (RTH) induces beneficial changes in global health status in postmenopausal women. Highlights We evaluated the long-term effects of a recreational team handball-based multicomponent training on broad-spectrum health and physical fitness markers of inactive postmenopausal women. Improvements in VO2peak and aerobic performance achieved after 16 weeks of training were maintained at 36 weeks. The 20-week extension of the training intervention resulted in further improvements in lipid profile markers and physical fitness variables. Recreational team handball could be suggested as an effective and safe strategy to counteract postmenopausal health-related constrains.


Introduction
Women's global health is deteriorated during menopause (El Khoudary et al., 2019). Indeed, cardiovascular risk factors such as elevated blood pressure and metabolic markers such as triglycerides (TRG) and lowdensity lipoprotein cholesterol (LDL) are adversely changed in postmenopausal compared to premenopausal women (Pu et al., 2017). Moreover, bone mineral density (BMD) (Greendale et al., 2012) and muscle mass (Juppi et al., 2020) decrease during menopause transition. Loss of muscle mass can lead to physical disability and increased risk of falls (Thomas et al., 2021) and low bone mass is a risk factor for fractures (Uusi-Rasi et al., 2019). Thus, lowering cardiovascular risk factors and improving bone mass and reducing falls are important strategies for preventing fractures and reducing morbidity and mortality (Teo & Rafiq, 2021;Uusi-Rasi et al., 2019).
Feasible and efficient exercise training regimes targeting global health status have been suggested to counteract the deleterious menopause-related effects (Grindler & Santoro, 2015). Combined or multicomponent exercise training (combining endurance, resistance and balance training) (World Health Organization, 2020) has shown to be more effective in improving cardiometabolic health status, which has been associated with bone health upregulation (Milanović et al., 2022), than singlecomponent modalities (Batrakoulis et al., 2022). Recreational team sports, an example of multicomponent exercise training (Krustrup et al., 2010a), have demonstrated to be feasible and broad-spectrum health-enhancing exercise options for middle-aged and elderly participants, and to assure exercise adherence over the short-term (Castagna et al., 2020a;Milanović et al., 2019). In fact, cardiovascular, metabolic and musculoskeletal improvements, similar to those obtained by conventional exercise programmes, are achieved with low doses of exercise (2 × 60-min sessions/week) (Krustrup et al., 2010a). Additionally, these are highly motivating activities (Milanović et al., 2019), thus overcoming the main barriers to exercise (i.e. lack of time and motivation) (European Comission, 2022). In fact, team sports practice results in lower levels of negative affects (such as angry, sad, anxious, ashamed, guilty, annoyed and worried) than individual sports practice, which can increase motivation to exercise (Barbry et al., 2021).
Nonetheless, apart from football, there is a lack of studies assessing the broad-spectrum health effects and adherence to other recreational team sports-based exercise programmes beyond the frequently used 12to-16-week period (i.e. short-term), particularly in the elderly. Indeed, only two studies have analysed the long-term effects of this particular multicomponent training. In elderly men, enhanced bone health, as well as attenuated age-related effects were shown for maximal oxygen consumption (VO 2max ) and glycated haemoglobin (HbA1c) after more than two years of floorball training (Pedersen et al., 2018). Moreover, 40 weeks of floorball for postmenopausal women were able to preserve body composition, cardiovascular and bone health, and physical fitness enhancements achieved after an initial 12-week period, as well as improved blood glucose regulation (Seidelin et al., 2017). Nevertheless, a 50% dropout was observed in the subgroup assessed at 40 weeks (Seidelin et al., 2017), confirming that long-term exercise adherence is particularly demanding in older populations (Findorff et al., 2009;Rivera-Torres et al., 2019).
Recreational team handball (RTH) has also been assessed as a global health promotion strategy (Castagna et al., 2020a;Pereira et al., 2021). The health effects of 16 weeks of RTH training were analysed in postmenopausal women, showing improvements on cardiorespiratory and physical fitness, bone health, body composition and lipid profile markers (Pereira et al., 2020;Pereira et al., 2021). Nonetheless, resting heart rate (HR), blood glucose regulatory markers, proximal femur and femoral neck BMD, and lean mass showed no changes. Thus, the role of exercise volume (total intervention duration) was put into perspective for future research. In fact, previous studies showed that an increase in leg BMD and in bone mineral content (BMC), both in the legs and in the whole-body, were observed only after 1 year and 40 weeks of football and floorball training, respectively, but not after 12 weeks (Randers et al., 2010;Seidelin et al., 2017). Extending the duration of the interventions beyond 16 weeks was suggested to ascertain if an enhancement in health status would only occur with longer durations, as well as further improvements after those shown after the short-term (Pereira et al., 2020;Pereira et al., 2021). Moreover, since the postmenopausal status is a permanent condition in women's life, the longterm adherence to exercise programmes, namely to this exercise mode, is of interest.
This study aimed at determining the long-term (36 weeks) adherence and the health and physical fitness effects of multicomponent training (RTH) for postmenopausal women without previous experience with the sport, and at comparing them with the alterations shown after the first 16 weeks (short-term).
We hypothesised that long-term RTH would result in positive health and physical fitness adaptations, and that the adherence to this exercise mode would be maintained over the 36 weeks (long-term) compared to the first 16 weeks (short-term). We also hypothesised that the health and physical fitness improvements (VO 2max , BMD and BMC, bone turnover markers, lipid profile, fat mass, aerobic performance, and postural balance) shown after the first 16 weeks would be maintained or further improved at the long-term.

Participants and study design
Recruitment started one month before the beginning of the trial. Eighty participants met the inclusion criteria (i.e. inactive postmenopausal women aged +55 years or younger if with menopause for at least three years) and were invited to participate in the study (Figure 1). Exclusion criteria was to present any medical contraindication for performing moderate-to-vigorous physical activity (PA). Two participants were on hormonal therapy. Some participants were taking antihypertensive (n = 13) and cholesterol-lowering (n = 16) drugs for more than one year before the starting of the intervention. The participants were asked to inform the research team if they changed their medicine intake or started a new medicine therapy. Nine participants dropped out for personal reasons (three before baseline evaluations and six subsequently). Seventy-one participants were stratified by VO 2peak and randomly allocated (3:2 ratio) to a multicomponent exercise training group (EXG) or a control group (CC). Unequal sample sizes (3:2 ratio) were used since a higher dropout was expected in intervention group than in the control group due to exercise-related injuries. An independent researcher generated the allocation sequence (computer-generated random numbers), enrolled, and assigned the participants to the groups. No significant betweengroups differences were observed at baseline (data not shown). EXG performed 36 weeks of a RTH-based exercise programme, while the CG was instructed to keep their regular daily PA. Both groups were asked not to change their dietary habits during the time of the intervention. Evaluations were performed at baseline, 16 and 36 weeks. Before the 16-week evaluations, two participants in the CG dropped out for personal reasons, one participant was excluded because started a new exercise programme and one participant in the EXG was excluded because did not complete the 16 weeks of the exercise programme, resulting in sixty-seven participants being evaluated. After the 16 weeks evaluations, twelve participants in the CG dropped out, namely ten that did not want to continue to be involved in the study and two for medical issues. Additionally, ten participants of EXG dropped out for personal reasons (2), medical issues (5) and injuries (one shoulder pain and two finger subluxations). Only the finger subluxations were related to the exercise programme; the shoulder pain occurred previously to the enrolment in the exercise programme, however, it was further aggravated during the intervention. At 16 weeks, the dropout rate was 5.6% (1.4% EXG and 4.2% CG) and 36.6% at 36 weeks (15.5% EXG and 21.1% CG). Thus, forty-five participants completed the additional 20 weeks of the intervention and were evaluated at that time point.
After being informed in verbal and written form about the purpose, potential risks, and benefits of the study, and that they could withdraw at any time without penalty, the participants signed a written informed consent. This study was conducted in accordance with the Declaration of Helsinki and ethical approval was provided by the local Ethics Committee. This study was registered at clinicaltrials.gov (identifier: NCT05292261).
Peak oxygen uptake (VO 2peak ) was the primary outcome of this study since cardiorespiratory fitness is associated with a lower risk of cardiovascular disease (Kodama et al., 2009), which is the major cause of morbidity and mortality in postmenopausal women (Baber et al., 2016). Bone health, metabolic health, body mass, body composition and physical fitness, as well as the other cardiovascular health markers were considered as secondary outcomes.
Sample size estimation for the main outcome (VO 2peak ) was performed for a sample power of 95% with an effect size of 0.25 at a significance level of 5% and for a dropout rate of 20%, resulting in 72 participants to be recruited.

Exercise programme
Briefly, EXG performed 2-3 × 60-min/week supervised training sessions, with 48 h of rest in-between, for 36 weeks. Each session was performed on an indoor team handball court (40 × 20 m; 34-36 m 2 /player) and started with a standardised warm-up [low intensity running and coordination exercises, combined with articular movements, strength (namely squats, lunges, push-ups and planks), flexibility (upper body stretching), and balance exercises (stand in one leg and hold the position)], followed by three 15-min periods of RTH matches, interspersed by 2-min breaks, played as small-sided games (4v4, 5v5 or 6v6). Other adaptations were made to the official team handball rules: no body contact was allowed, and softer and lighter team handball balls than the official ones were used, as this population had no experience with this sport, and to avoid potential injuries. Other exceptions to the official team handball rules included: no exclusions; no substitutions; no dribbling; the participants rotated positions every 3 min in a random order, including the goalkeeper; and the ball was immediately put back in play by the goalkeeper after a goal. According to procedures previously described (Pereira et al., 2020;Pereira et al., 2021), HR (Firstbeat Technologies Ltd., version 4.7.2.1, Jyväskylä, Finland), rating of perceived exertion (RPE), and fun levels (Grant et al., 1999), 0-10 scale, were monitored in all training sessions and maximal HR (HR max ) was determined.

Measures and test procedures
All the participants were evaluated on two separate days (distanced two to four days apart) at baseline and at the end of 16 and 36 weeks (Faculty of Sport, University of Porto, Portugal, facilities). No exhausting PA was allowed in the 48 h before testing. The participants were instructed to have a light meal not earlier than 3 h before the evaluations. Blood sampling was performed in the morning after a minimum 8-h overnight fasting. On the first evaluation day, resting HR evaluation, dual-energy X-ray absorptiometry (DXA) scans (body, fat and lean mass and BMD and BMC included), isometric knee extension peak torque and rate of force development (RFD) tests and a standardised treadmill incremental protocol to assess VO 2peak and time to exhaustion (TTE) were performed, in that order. On the second evaluation day, blood sampling for metabolic [blood glucose, insulin, total cholesterol (TC), highdensity lipoprotein cholesterol (HDL), LDL, TRG] and bone turnover markers [sclerostin, carboxy-terminal type-1 collagen crosslinks (CTX), procollagen type-1 amino-terminal propeptide (P1NP) and osteocalcin] determination was performed. After a light breakfast, physical fitness [Yo-Yo Intermittent Endurance level 1 test (YYIE1) performance, handgrip strength (handgrip dynamometer, Takei TKK 5401, Niigata, Japan) and postural balance (single-leg flamingo test)] were evaluated. The PA levels were measured using the International Physical Activity Questionnaire (Craig et al., 2003).
Sclerostin was measured in plasma using the TECO-Medical Sclerostin HS EIA assay (Quidel Corporation, San Diego, CA). The sample aliquots were kept frozen at −80°C until further analysis. The analysis was performed immediately after thawing the samples using one single batch of the assay. Assay performance was verified using the manufacturers' control and the intraassay precision was 10% at both the 0.2 and 1.9 ng/mL levels.
Isometric knee extension peak torque and RFD of the dominant leg were measured on a Biodex isokinetic dynamometer (Biodex System 4 Pro, Shirley, NY, USA) following standard procedures. The participants were instructed to sit on the isokinetic dynamometer chair and hold on the chair handles, with their upper body (i.e. torso and waist) and with the dominant leg tightly stabilised by Velcro straps. Before the test, the participants performed a 3-min warm-up in a cycle ergometer at a work rate of 60 rpm and 0.5 Kp. A familiarisation test was performed by all participants that were instructed to exert a submaximal contraction against the pad. After a 1-min rest period, one maximum contraction of 5 s was performed. Participants received verbal encouragement to achieve their maximum effort and were instructed to stop immediately if they felt any pain or discomfort (Gray et al., 2016). Peak toque (Newton-meters, Nm) was recorded. The RFD (N/s) was calculated as the maximum slope of the force-time curve (force/time).
The detailed description of the assessment of the other markers was described elsewhere (Pereira et al., 2020;Pereira et al., 2021).

Statistical analyses
Results are presented as mean ± standard deviations (SD) and 95% confidence intervals (95% CI). To analyse between-group differences and intervention effects, a two-way analysis of variance (ANOVA) for repeated measures with Bonferroni post hoc tests was performed. Differences in absolute and relative between-groups changes were analysed using a Student's unpaired ttest. Differences in training intensity, RPE and fun levels between the two time points were analysed using a student's paired t-test. Practical significance was assessed by calculating Cohen d and interpreted as trivial (<0.2), small (0.2-0.5), medium (0.5-0.8) and large (>0.8) (Cohen, 1988). The Shapiro-Wilk test was used to test the normality of the data. Statistical Package for the Social Sciences (SPSS Inc., version 26) was used for data analysis. Statistical significance was set at p ≤ .05.

Chronological age, anthropometric characteristics, cardiorespiratory fitness and habitual physical activity
Chronological age, anthropometric characteristics and cardiorespiratory fitness of the participants are presented in Table 1. Habitual, vigorous and moderate PA, walking and sedentary behaviour at baseline and after 16 and 36 weeks are presented in Table 2. No differences (p > .05) were observed either between-or withingroups.
Cardiovascular, bone, metabolic health, body mass, body composition and physical fitness markers at baseline, after 16 and 36 weeks of the participants in exercise training and control groups are presented in Table 2.

Training sessions' attendance, intensity and fun
The participants' mean training attendance was 2.0 ± 0.4 (1.9-2.2) sessions/week during the first 16-week period. In the following 20 weeks, the mean attendance decreased (p < .001) to 1.4 ± 0.5 (1.2-1.6) sessions/week. Cardiovascular training intensities and RPE for 16 and 36 weeks are presented in Table 3. There was an increase in mean HR and RPE (p ≤ .002) between the 16 and 36week periods, respectively. Fun levels were 8.3 ± 1.5 and 8.7 ± 1.1 (0-10, AU), for 16 and 36 weeks, respectively, with no differences between the time points.

Discussion
This randomised controlled trial (RCT) evaluated the effects of a 36-week recreational team handball-based multicomponent training intervention on broad-spectrum health and physical fitness markers of inactive postmenopausal women. In general, although improvements in cardiorespiratory fitness (i.e. VO 2peak ), aerobic performance (i.e. YYIE1) and in most of the healthrelated variables achieved after 16 weeks of training were maintained at 36 weeks, the 20-week extension of the training intervention resulted in further improvements in lipid profile markers and other physical fitnessrelated variables, besides YYIE1, compared to the shortterm (16 weeks). Interestingly, these improvements were obtained despite a decrement in weekly training frequency from 2.0 to 1.4 sessions, during the last 20 weeks.
An interaction was shown for YYIE1 performance, isometric knee extension peak torque, 2 h OGTT blood glucose and HDL in favour of team handball training. As no changes were observed in daily PA and sedentary behaviour during the 36-week period for the exercise training group, the observed positive effects on health and fitness profile are related to the multicomponent training per se.
YYIE1 performance increased after 16 weeks and was maintained at 36 weeks in EXG, that showed higher values than the CG at both time points. Similar aerobic performance results were shown in postmenopausal women after playing floorball for 12 and 40 weeks (Seidelin et al., 2017) and in 20-to 43-year-old men playing football after 12 and 64 weeks (Randers et al., 2010), despite the decrease in training frequency from 12 to 40 weeks, supporting the positive effect of recreational team sports practice in aerobic performance in different populations. Being YYIE1 largely associated to VO 2max in untrained individuals (Castagna et al., 2020b), the increase in this aerobic performancerelated parameter is of most importance, as VO 2max is inversely related to the risk of cardiovascular disease and all-cause mortality (Kodama et al., 2009), and low levels of physical fitness are one of the most important risk factors for cardiovascular diseases (Warburton et al., 2006).
Regarding isometric knee extension peak torque, our data showed that 36 weeks of RTH not only prevented putative aging-related strength loss but, more importantly, increased lower body strength in postmenopausal women compared to the pre-training condition. Nonetheless, quadriceps RFD was unchanged, what is in accordance with results shown after one-year of football training for 68-year-old men (Sundstrup et al., 2016). Additionally, a decrease in lower body strength was observed in the CG (Pedersen et al., 2018;Sundstrup et al., 2016), which reinforces the role of recreational team sports in preventing aging-related strength loss. Moreover, an increase in isometric hamstring strength and RFD was observed in the football training group, which could be related with the nature of the highintensity actions that stimulate this muscle (e.g. sprints) that occur in this sport (Sundstrup et al., 2016). Therefore, exploring the impact of RTH practice in strength improvement of different muscle groups, could also be of interest. Besides obvious benefits related to strength improvement on functional ability and decreased risk of falls, our data are important as lower muscle strength, but not muscle mass, is a predictor of mortality in older adults (Newman et al., 2006). As a decrease in muscle mass during menopause transition is reported in postmenopausal women, the increase in muscle strength promoted by RTH, irrespectively of an increase in muscle mass, seems to be extremely crucial. In fact, muscle strength, more than muscle mass, is considered the key factor for sarcopenia diagnosis, which is a muscle disease highly prevalent at older ages (Cruz-Jentoft et al., 2019), and so postmenopausal women at any time point. Interestingly, the combination of improved cardiorespiratory fitness and muscle strength, as opposed to the isolated improvement of each, is suggested as the most effective strategy to reduce all-cause and cardiovascular mortality risk (Kim et al., 2018). Moreover, the related increase in physical fitness would lead to improvements in global health status (Warburton et al., 2006).
Menopause might lead to serious changes in metabolism, predisposing women to diabetes, insulin resistance, and corresponding increased risk of impaired glucose tolerance over the years (Lambrinoudaki et al., 2022;Wu et al., 2001). Moreover, postmenopausal women are prone to increased TC and LDL and reduced HDL concentrations, which ultimately let them more susceptible to cardiovascular diseases (Saha et al., 2013). Thus, the interaction effect observed in some metabolic variables resulting upon this exercise programme must be highlighted. In fact, although low HDL, high waist circumference as well as elevated fasting glucose levels are associated to metabolic syndrome, which has an increased prevalence after menopause (Marchi et al., 2017), the increase in HDL, together with a decrease in android fat mass observed in withingroup differences (abdominal fat) might help to counteract and prevent this cluster of risk factors associated to cardiovascular diseases. Within-group, but not between-group differences, were also shown in VO 2peak , lumbar spine BMD and BMC, P1NP, osteocalcin, fasting blood glucose, TC, LDL, body mass, android fat mass, handgrip strength and postural balance after 16 weeks, but without further improvements at 36 weeks of RTH. This lack of additional improvements at 36 weeks may indicate a need for higher training frequency, as 97% of the participants complied with two to three weekly sessions over the short-term, but only 42% complied with at least two sessions over the long-term. Nevertheless, it is important to note that the improvements obtained after 16 weeks were maintained at 36 weeks, which suggests that just one weekly 60-min session was sufficient to maintain the beneficial effects attained after two weekly training sessions over 16 weeks. This finding is of the most importance, since lack of time is one of the main barriers to exercise (European Comission, 2022). Therefore, further studies addressing the dose-response effects of RTH on physical and health-related outcomes may optimise training prescription regimens. Besides the potential impact of training frequency, intensity has also been considered an important physiological parameter associated to distinct outcomes in postmenopausal women (Kistler-Fischbacher et al., 2021). In fact, in recreational football practice, which working demands are similar to RTH, mean HR was >80% of HR max and at least 10% of total time was spent >90% HR max (Krustrup et al., 2010b), which is slightly higher that our study values. Short-and long-term recreation football-based interventions resulted in improvements in VO 2max , BMD, bone turnover markers, LDL, fat mass, and physical fitness, compared to a non-exercise group (Milanović et al., 2015;Milanović et al., 2019;Milanović et al., 2022). Therefore, higher exercise volume and/or intensity or eventually a higher sample sized (namely in the CG), might have been necessary to reach group differences in our study.
Volume and programme adherence are limitations of this study as training attendance decreased from 16 to 36 weeks (2.0 to 1.4 sessions/week) and the dropout in the CG was 21%. This clearly demonstrates the difficulty of keeping middle-aged/elderly participants in the interventions, especially in the CG. Nonetheless, the overall dropout in the EXG was only 16%, which is much lower than that observed in other exercise interventions (Milanović et al., 2019) and a potential evidence of the motivational advantage of RTH.
Taken together, data from this RCT study responds to the scarcity of long-term interventions using team sports apart from football and adds relevant knowledge to the literature on recreational team sports health-related outcomes. Moreover, it allowed to access health and physical fitness changes after long-term exposure to RTH, which was put in perspective in previous short-term studies using this exercise mode and population. Considering the injury incidence during the intervention (0.8 injuries/1000 h), RTH could be suggested as a safe and effective strategy to counteract postmenopausal health-related constrains. Our data only partially confirmed our hypothesis. In fact, only some, but not all or even most of the health and physical fitness improvements obtained after the first 16 weeks were further improved at the long-term (i.e. 36 weeks). Therefore, maybe due to the distinct "biological nature" of the physiological outcomes in analysis, still several issues should be put into question, regarding the definitive and optimal dose-response equation that would prompt adequate and sustainable general outcomes.
In conclusion, although we cannot be completely assertive regarding the truly impact of the exercise programme, due to the lack of between-group differences in some parameters after 36 weeks, the participants engaged in the programme improved physical fitness and lipid profile compared to the CG. Therefore, RTH may be considered a beneficial multicomponent exercise mode over the long-term for postmenopausal women.