Effect of aerobic training on hot flushes and quality of life—a randomized controlled trial

Background and objective. To estimate whether aerobic training has an effect on frequency of hot flushes or quality of life. Design. A randomized controlled trial. Participants and setting. Symptomatic, sedentary women (n = 176), 43–63 years, no current use of hormone therapy. Intervention. Unsupervised aerobic training for 50 minutes four times per week during 6 months. Outcomes. Hot flushes as measured with Women's Health Questionnaire (WHQ) and Health-Related Quality of Life (HRQoL, SF-36), daily reported hot flushes on phone-based diary, cardiorespiratory fitness (CRF), and body composition. Results. Intervention group had larger decrease in the frequency of night-time hot flushes based on phone diary (P for month X group = 0.012), but not on WHQ scale. Intervention group had less depressed mood (P= 0.01) than control women according to change in WHQ score. Changes in WHQ score in depressed mood (P = 0.03) and menstrual symptoms (P=0.01) in the intervention group were significantly dependent on frequency of training sessions. HRQoL was improved among the intervention group women in physical functioning (P= 0.049) and physical role limitation (P= 0.017). CRF improved (P= 0.008), and lean muscle mass increased (P= 0.046) significantly in the intervention group as compared to controls. Conclusions. Aerobic training may decrease the frequency of hot flushes and improve quality of life among slightly overweight women.


Introduction
Hot fl ushes (i.e. vasomotor symptoms) are a common complaint in many women and may occur for years beyond menopausal transition (1 -3). Hormonal and thermoregulatory involvement in the occurrence of hot fl ushes is likely, and many forms of pharmacological as well as non-pharmacological alleviation have been studied (4). Although traditional post-menopausal hormone therapy (HT) is the most effective treatment to alleviate menopausal hot fl ushes, it has been subject to major criticism due to the results of large clinical trials in which the risks of HT have outweighed the benefi ts (5). Women with higher adiposity report more severe hot fl ushes (6), and successful weight loss has been associated with hot fl ush decrease in a randomized trial (7). Because increased physical activity has benefi cial effects on weight management, it could be one alternative treatment to HT in alleviating menopausal symptoms. Furthermore, physical exercise is also known to increase hypothalamic β -endorphin production, which may stabilize thermoregulation known to be disturbed during menopausal hot fl ushes (8).
Evidence concerning whether or not physical training reduces hot fl ushes is still inconclusive. Some studies have suggested that increasing physical activity may possibly decrease the number of hot fl ushes (6), but studies showing no association (9) or an inverse association (10,11) have been published as well. Previous randomized clinical trials on exercise and hot fl ush reduction have either had small or selected populations, symptom reduction as a non-primary outcome, or other limitations such as non-symptomatic women or lack of a control group (12 -16). Our objective was to estimate whether aerobic training decreases hot fl ushes or increases health-related or menopause-specifi c quality of life among symptomatic menopausal women.

Study design and participants
With the approval of Pirkanmaa Hospital District Ethics Committee, we recruited healthy Caucasian women via newspaper advertisements in Pirkanmaa area between January and March 2009. Women who were willing to participate to the trial after reading the newspaper advertisement were screened for inclusion criteria through a telephone interview. If they were eligible and willing to participate, they received information concerning the trial by mail. Participants were not paid for being included, neither in intervention nor control group, since this is not allowed in medical studies performed in Finland. All participants gave written informed consent.
Design was randomized controlled study with allocation ratio of one. Women were assigned to intervention and control groups by computer randomization. Envelopes including information on assigned group were delivered by research nurses. Inclusion criteria for the study were: symptomatic (daily hot fl ushes), age 40 -63 years, no current use or use in the past 3 months of HT, sedentary (aerobic training less than twice weekly), and 6 -36 months since last menstruation (Figure 1). Women who were physically active (exercising Ն 2 times/week, for at least 30 minutes each time), or whose body mass index (BMI) was over 35 kg/m 2 , or who had coronary heart disease, or orthopaedic or other diseases preventing them from exercising, were excluded from the study. In addition, women who were using medication infl uencing heart rate ( β -blockers, sympathomimetics) were excluded, since they would have biased the results concerning heart rate. Menopausal status was verifi ed by assay of plasma follicle-stimulating hormone (FSH), which had to exceed 30 IU/L.
Before initiating the trial, recruited women kept a 2-week diary (pencil and paper) concerning day-time and night-time hot fl ushes. The women were instructed to record as a hot fl ush a sensation of warmth, with or without concomitant sweating. To obtain an estimate of the overall impact of hot fl ushes we calculated weekly and daily frequency of all hot fl ushes before the intervention, and daily frequency of hot fl ushes both day-time and night-time during the intervention. From a total of 351 women, 176 women participated in the study; 175 women were excluded since they did not meet the inclusion criteria. The duration of intervention was 6 months.

Outcomes
The primary outcome measure of the study was hot fl ushes as measured with Women ' s Health Questionnaire (WHQ). Secondary outcomes to be reported in this paper were hot fl ushes reported by phone-based diary, health-related quality of life as measured with SF-36 scale, cardiorespiratory fi tness (as estimated by the Urho Kekkosen Kuntoinstituuttisäätiö (UKK) 2 km Walk Test (17)), and body composition (weight, fat and lean mass; kg). Specifi c questions related to sleep will be reported in a separate article.
The WHQ was used as an instrument to assess menopause-specifi c quality of life (18). The questionnaire has been validated in the Finnish language, and test-retest reliability has been evaluated with a 2-week interval (19). It is composed of 36 items covering mood states, physical sensations and experience, sexual behaviour, and hot fl ushes, with a total of nine sub-scales (depressed mood, somatic symptoms, hot fl ushes, anxiety/fears, sexual behaviour, sleep problems, menstrual symptoms, memory/concentration, attractiveness). Self-reported symptoms are scored on a fi ve-point scale (values from 0 to 4).
Health-related quality of life was measured by using the SF-36 Health Survey questionnaire, which has been validated in Finland (20)

Measurements
Information on day-time and night-time hot fl ushes and aerobic training was collected by using mobile phone questionnaires twice a day. The mobile phone questionnaire included pre-specifi ed questions on hot fl ushes, sleep, and other symptoms. During the morning, the structured-item questionnaire included fi ve specifi c questions on hot fl ushes (yes/no), question on night sweating (yes/no), and two questions on sleep quality (yes/no). During the evening, the structured-item questionnaire included the same questions as in the morning questionnaire, with added (yes/no) questions regarding headache, mood swings, dysphoria, depression, dryness of the vaginal mucosa, and urinary and other symptoms. If the woman did not want to use her own mobile phone, or the phone did not have a 3G technology option, the phones were arranged by the study team. Responses were automatically transferred via 3G technology into digital format. Since the symptom diary method had not been used earlier, it was evaluated by way of a separate usability questionnaire (SUS; system usability scale (21)) immediately after 6 months of intervention. Most of the women participating in the intervention also responded to the usability questionnaire (response rate 72.7%). The score in the usability questionnaire was 75.4 (range 0 -100). The women reported the proportions of both day-time and night-time hot fl ushes, which were compulsory information programmed to the phone. Reporting on severity of hot fl ushes during the trial was an option, not compulsory, and thus lacking in half the responses, although at base-line hot fl ush status was collected from all women using conventional methods (paper). Base-line questionnaires (on paper) included information on employment, life-style, dietary habits, and other health-related issues. The women also received a questionnaire on life-style changes at the 12th week and at the end of the trial (24 weeks). Although the phone-based questionnaire was not validated, the questions used were simple (yes/no) and expressed the existence of symptoms without grading them.
Fat and lean mass and body mass index were measured by dual X-ray absorptiometry (DXA). Body mass index was calculated as the ratio of weight (kg) to height squared (m 2 ). Cardiorespiratory fi tness was estimated by using the UKK Walk Test, which is a validated method for measuring aerobic fi tness of 20 -65-year-old adults who have no illnesses or disabilities that limit brisk walking and who are not on medication that affects heart rate (17). Heart rate was monitored during the walk and registered immediately at the end (Polar Electro, M61, Finland). Measurements of cardiorespiratory fi tness (CRF) were performed twice at base-line due to the effect of learning.

Aerobic training
Women in the intervention group kept physical activity diaries once a week during the fi rst month; thereafter once a month. Women in the control group were asked (via questionnaire on paper) once in the middle of the trial (11 -12 weeks from base-line) whether they had changed any of their physical activities or dietary habits.
The unsupervised aerobic training programme included aerobic training four times per week, with 50 minutes of exercise each time. Ratings of perceived exertion (RPE) were used to check the intensity of aerobic training (22). Participants were instructed to exercise at a level corresponding to 13 -16 on the scale from 6 to 20. This corresponds to about 64% -80% of maximal heart rate (22). At least two sessions had to include walking or Nordic walking, and the other two sessions could include walking, Nordic walking, jogging, cycling, swimming, skiing, aerobics, or other gymnastic exercise. The intervention mainly involved walking, because earlier UKK trials have shown favourable results on health among menopausal and post-menopausal women (23). Adherence to the trial was supported by an option to participate in instructed aerobics or step aerobics sessions at the UKK Institute twice a week.
The intervention group wore heart rate monitor belts (Suunto ® ; Memory Belt, Suunto, Vantaa, Finland) in training sessions. Every second week an aerobic training instructor gave feedback to the participants concerning the training sessions. Data collected from the heart rate belts were transferred to a computer and analysed with a software program, Firstbeat technologies HEALTH (Firstbeat Health ® , Finland). Participants also reported their exercise by mobile phone and received weekly feedback by program message and an additional program message in case responses were not received during 4 days. Frequency of realized aerobic training sessions was based on both heart rate monitor belt information (sessions lasting at least 15 minutes and at most 4 hours, with at most 20% error) and phone-based aerobic training diary.
Both the intervention and control groups attended lectures once or twice per month given by the principal investigator of the study (R.L.). The lectures took 60 -75 minutes and mostly covered topics of physical activity and general health. The women were not compensated fi nancially for participation in the trial, but expenses arising from the use of mobile phones were covered. Aerobic training-based injuries were monitored by means of questionnaires, and no important harms or unintended effects occurred.

Statistical analysis
Base-line characteristics of the groups were tabulated (means and standard deviations, or percentages; Table I). Differences at base-line between the randomized groups were evaluated by using Student ' s t test for continuous variables when they were normally distributed, and the Mann -Whitney U test when non-normally distributed. Normality was evaluated with the Kolmogorov -Smirnov test. Categorical variables were tested using the chi-square test. All analyses were performed in originally assigned groups. Missing values concerning the primary outcome (WHQ) were imputed by using base-line value in case a 6-month value was missing; 22 persons had imputed values in 1 -9 items of WHQ. Multivariate modelling was not used since the study was a randomized trial and there were no differences in base-line confounding factors between the groups.
As an evaluation of dose-response effect, realized aerobic training sessions with suffi cient intensity and duration were divided to three equal-size groups (low, medium, high) to enable analysis of change in WHQ and SF-36 scores from base-line to 6 months. The effect of realized aerobic training on betweengroup differences in WHQ and SF-36 was tested by use of non-parametric methods (Kruskal -Wallis test), because the distributions were not normal.
The groups had different proportions of women responding to the question concerning hot fl ushes yes or no -the proportions developed differently by the group during intervention time. Regression slopes were statistically compared by using mixed models for binary outcome (yes/no hot fl ushes).
Because the responses for the same person are not conditionally independent, a multilevel mixedeffect logistic regression model was fi tted. This model allows for a difference between groups at base-line and linear changes in the log odds of hot fl ushes over time. To estimate the intervention effect, the interaction term (month ϫ group) was included in the model.
The proportion of women with hot fl ushes was expected to be at least 68% (proportion among controls, P c ). The expected difference between intervention and control group in proportion of women with hot fl ushes was at least 30%. Based on these assumptions, the proportion of women with hot fl ushes would be 47.3% among women in the intervention Table I. Base-line characteristics and quality of life in the intervention and control groups. QoL was estimated by means of the Short Form-36 quality of life (SF-36) score and the menopause-specifi c quality of life score (Women ' s Health Questionnaire (WHQ)). group. If alpha is 0.05 and beta 80%, the number needed in the sample is calculated with the following formula: n ϭ (alpha ϩ beta) 2 ϫ {[P c ϫ (1 -P c ) ϩ P i ϫ (1 -P i )]/(P c -P i ) 2 } ϭ 89.7. Thus, 90 women were needed in each group to achieve 80% power.

Base-line
Of the 176 women who started, 159 continued until the end of the study (90% compliance rate). Finally, 154 women (74 in the intervention group, 80 in the control group) completed the UKK Walk Test; 4 (3 in the intervention group, 1 in the control group) were not willing to complete this test. Dropouts ' ( n ϭ 22; 14 in the intervention group, 8 in the control group) main reasons to discontinue the trial were personal reasons ( n ϭ 17) and chronic illness ( n ϭ 3) or initiation of hormone replacement therapy ( n ϭ 2). Drop-out women whose base-line characteristics were available ( n ϭ 17, missing 5) were compared with responders ( n ϭ 159) in regard to age, education, employment status, smoking, weight, BMI, and FSH. Signifi cant differences were found in age and weight -mean age of the respondents was 54.2 years (SD 3.7) and among drop-outs 51.1 years (SD 3.4) ( P ϭ 0.001 in t test). Weight among drop-outs was on average 59.3 kg (SD 6.0) and among responders 71.9 kg (SD 12.5) ( P ϭ 0.045 in t test).
The participating women were on average 54 years of age, with a body mass index of 26 -27 kg/m 2 (Table I). There were no signifi cant differences by group in any base-line values (Table I). At base-line, the mean number of all hot fl ushes (HF) per day independent of severity (mild, medium, or severe HF) among intervention women was 6.3 and among control women 5.9. Mean number of daily mild HF was 1.7 among the intervention women, 1.9 among controls; mean number of medium daily HF 2.9 and 2.8; and mean number of severe daily HF 1.6 and 1.0, respectively. The lowest number of HF a subject had at base-line was 0.11 HF/day. Average time for menopause was 4.5 years among intervention group women and 3.4 years among control women (SD 4.4 and 4.1, respectively) ( P ϭ 0.097).
Response rates did not differ between the exercise group and the control group: morning questionnaire, intervention group 74.4%, control group 69.4%; evening questionnaire, 69.8% and 70%, respectively.

Hot fl ushes
There were no signifi cant between-group differences in hot fl ushes according to WHQ (Table II).
At base-line phone-based diary, 57% of the women in the intervention and 61% of the women in the control group reported night-time hot fl ushes; proportions for day-time hot fl ushes were 72% and 71% (Table III). Decrease in day-time hot fl ushes was 2.9% in the intervention group (from 72.4% to 69.5%) and 4.5% (from 71.1% to 66.6%) in the control group. According to multilevel mixed models, the decrease by month in the number of nighttime ( P for interaction group ϫ month ϭ 0.012) but not day-time hot fl ushes ( P for interaction group ϫ month ϭ 0.61) was signifi cantly larger in the intervention group than in the control group (Table III; Figure 2). However, in day-time hot fl ushes there was a decrease until the 3-month point (Figure 2).

Changes in WHQ and SF-36
Between-group differences in WHQ were signifi cant in depressed mood ( P Ͻ 0.001) and attractiveness ( P ϭ 0.04) and close to signifi cant in sleep problems ( P ϭ 0.07) at the end of the trial. Between-group difference in change from base-line to 6 months was signifi cant in depressed mood ( P ϭ 0.02) and close to signifi cant in attractiveness ( P ϭ 0.06) ( Table II). WHQ scores were signifi cantly correlated with frequency of realized aerobic training sessions in depressed mood ( P ϭ 0.03) and menstrual symptoms ( P ϭ 0.01) and close to signifi cant in anxiety ( P ϭ 0.06) (Table IV). No differences were found in any of the WHQ items by base-line hot fl ush score (not shown in the table).
After 6 months ' training, intervention group women had signifi cantly higher SF-36 scores in vitality ( P ϭ 0.02) and mental health ( P ϭ 0.02) than did control women (Table II). Between-group differences in changes of SF-36 scores were signifi cant in the areas of physical functioning ( P ϭ 0.049) and physical role limitation ( P ϭ 0.017) and close to signifi cant in general health ( P ϭ 0.07) and vitality ( P ϭ 0.06) (Table II). Women with hot fl ush score higher than or equal to 78.5 (mean hot fl ush score at base-line) had signifi cantly larger improvement in SF-36 vitality ( P ϭ 0.042) and mental health ( P ϭ 0.017) than did women with lower hot fl ush score ( Ͻ 78.5) at base-line. Changes in SF-36 score were not signifi cantly related to frequency of aerobic training sessions (Table IV).

Fitness and body composition
The groups did not differ as regards cardiorespiratory fi tness or body composition at base-line (Table V). The women in the intervention group increased their estimated maximal oxygen consumption (VO 2 max) statistically signifi cantly in comparison with the control group ( P ϭ 0.008). Increase in lean mass was signifi cantly higher among intervention group (0.57 kg versus 0.15 kg; P ϭ 0.046) than among control group (Table V).

Other results
One woman in the intervention group and three women in the control group used SSRI/SNRI medication during the trial (difference not significant). Seven women in the intervention group and four women in the control group used natural remedies. When users of SSRI/SNRI and/or natural remedies were excluded from the analyses, change in SF-36 physical functioning was no more significant ( P ϭ 0.085) (Table II), and signifi cance of change in lean muscle mass decreased as well ( P ϭ 0.066) ( Table V). There were no users of excitatory drugs among either group, and the use of liquids including caffeine did not differ between the groups (mean number of cups of coffee 3.93, SD 1.88 intervention; 3.95, SD 1.94 control)

Discussion
In our trial the intervention group had signifi cantly larger decrease in the frequency of night-time hot fl ushes based on mobile phone diary, although in the WHQ scale between-group differences in hot fl ushes were not found. In the WHQ scale aerobic training resulted also in lower scores for depressed mood and higher scores for attractiveness. Between-group differences in WHQ score change were signifi cant in Table IV. Changes in quality of life (WHQ and SF-36 scores) by frequency of aerobic training sessions (from phone-based diary and heart rate belt information). Intervention group women ( n ϭ 78). Signifi cance between the frequency groups tested with non-parametric Kruskal -Wallis test.
Change (base-line to 6 month) in: Low (  women, whereas several other studies have included women based on age criteria (15). The fi rst randomized study was performed in Sweden and reported promising results concerning exercise in spite of small sample size (24). In a larger ( n ϭ 164) 4-month trial the effects of yoga and moderate intensity walking on menopausal symptoms and mental health were assessed (25). Their results showed that both groups had decreases in hot fl ushes. Signifi cant increases in positive affect scores were found in the walking and yoga groups but not among controls. In a Turkish study, 65 menopausal women participated in an aerobic exercise programme for 6 months (16), depressed mood. Changes in WHQ scores were dose-dependent in depressed mood, anxiety, and menstrual symptoms. Health-related quality of life was signifi cantly improved in the areas of physical functioning and physical role limitation. Cardiorespiratory fi tness improved, and lean muscle mass increased in the aerobic training group.

Previous experimental studies
Earlier experimental studies concerning aerobic training and hot fl ushes are few; only two randomized controlled trials have included symptomatic of hot fl ushes and better QoL (15). An observational study by Aiello et al. (13) showed an increased severity of hot fl ushes in the exercise intervention group compared with controls. The women in the study were relatively old (mean age 60.6 years), and the majority of them reported symptoms other than hot fl ushes. Frequent physical exercise has been associated with a risk of greater severity and frequency of hot fl ushes in other non-experimental studies as well (10,11). In a multi-ethnic sample of 3,302 women in the Study of Women ' s Health in the Nation (SWAN), a trend of increasing prevalence of all symptoms with decreasing physical activity was found (31). In the study by Gold et al. (6) women were asked about their perception of activity level but not the type of activity. Thus, observational and experimental studies have confl icting results due to differences in assessment, i.e. questions concerning physical activity and sample characteristics such as size, age, and menopausal status of the women.

Strengths
Strengths of our study include the trial design, use of valid quality of life scales, and inclusion of women with moderate to severe hot fl ushes. The trial design was comprehensive for evaluation of exercise effi cacy as regards both hot fl ushes and menopausal QoL. Furthermore, the primary outcome of the study was occurrence of hot fl ushes, which was measured with a validated instrument (WHQ). According to a literature review of instruments used to assess health-related quality of life during and after menopause (19), both SF-36 and WHQ are reliable and may overlap in mental health areas. The effect was modest on menopausal health, while the benefi ts were more seen in general health, as shown in previous physical activity studies as well (29). Generally, the non-response rate to the paperbased questionnaires was low. Non-adherers differed from adherent women in regard to age and weightthey were younger and lighter. The infl uence of these issues probably did not affect the primary outcome but might have had an effect on phone-based Table V. Aerobic training-related variables (fi tness and body composition) at base-line, after 6 months intervention, and change between base-line and 6 months. Differences between groups were tested by t test.

Base-line intervention
Control P

End intervention
Control P  ,0,and 0;end: 10,5,and 5;change: 11,5,and 5;end: 4,3,and 3;change: 4,3,and 3. DXA ϭ dual X-ray absorptiometry; VO 2 ϭ volume of oxygen consumption . and the intervention group had signifi cantly fewer hot fl ushes and other menopausal symptoms. In our study, women in the intervention group reported a decreased frequency of hot fl ushes, especially during night-time. An increased level of aerobic training probably improved their quality of sleep and may have resulted in fewer awakenings due to nighttime hot fl ushes. Previous randomized studies have showed that adults with moderate sleep complaints can improve their sleep quality by moderate-intensity exercise programmes (26,27), but lower-intensity exercise such as walking or low-intensity yoga programmes may be ineffective (28). Decrease in daytime hot fl ushes was small, even if the downward trend was steeper among the intervention group until half of the intervention time; thereafter the trend was upward among both groups.
Previous intervention population studies have shown that increased physical activity has been associated with improvement of all dimensions of quality of life, but magnitude in improvement of mental health scale was smaller than in the physical health scale (29). In our results the improvements were found both in mental and in physical health items, possibly due to the sample characteristics, sedentariness and daily symptoms. A randomized, controlled trial involving 464 menopausal women showed quality of life (QoL) improvements after 6 months of regular exercise, independent of weight change (30). However, in the study the aim was not to relieve hot fl ushes. Another non-controlled trial for 16 weeks, aimed at increasing QoL by walking three sessions per week 45 minutes at a time, and involving 16 women, showed favourable results (31). In a study carried out by Teoman et al. (32), aerobic fi tness and quality of life were improved after 6 weeks ' training. Our results are parallel with these fi ndings based on smaller samples, confi rming the benefi ts of exercise both on health-related QoL and night-time hot fl ushes.

Previous non-experimental studies
The results of some non-experimental studies have shown associations between exercise and lower rates questionnaire results if imputation of the missing data was not performed.
In outcome assessment we used a new method of daily mobile phone contact. The method of daily responding decreases recall bias, and the questions were simple (yes/no). To our knowledge no previous study has involved the use of similar technology and mobile phone questionnaires in connection with menopausal symptoms. This novel technology enabled minimization of recall bias, since menopausal hot fl ushes were reported twice a day. Measurement of cardiorespiratory fi tness was based on the UKK Walk Test, which has been widely used and is feasible for both scientifi c and practical purposes (17).
An additional strength of the study is a control group programme. The control group attended the lectures during the trial, since otherwise they might have been more likely to drop out from the trial. Thus, the main purpose of the lectures was to motivate and support the control group without aerobic training. Since both groups were given the same lectures, exercise training effect may be clearer than if the control group was not supported at all.

Weaknesses
Regarding weaknesses of our study we acknowledge that the phone-based assessment of hot fl ushes was not based on a validated instrument. Although the frequency of hot fl ushes was accurately reported, more than 60% of the daily mobile phone responses lacked information regarding their severity, which was not compulsory information in the questionnaire. The majority of the women were unsure of the severity of their hot fl ushes and left the area blank. This is a limitation which should be taken into account in future mobile phone questionnaires. Before starting the trial the women kept 2-week diaries of daily symptoms by way of conventional questionnaires, not mobile phone diaries. In the feedback questionnaires after the intervention the women reported diffi culties in judging the severity of symptoms during the mobile phone survey. In spite of these limitations, our study showed high adherence, since only 14 women in the intervention group and 8 in the control group dropped out.
As another limitation in our trial, the women were on average only slightly overweight, and all were of Caucasian origin. Therefore the results may not be generalized to populations with diverse ethnic backgrounds or a higher prevalence of obesity. However, possibly the effect of aerobic training could be at least as large among obese women, perhaps even larger.
Sample size calculations were based on a 30% difference between the groups in hot fl ushes. Based on this assumption and the expected prevalence of 68% of symptomatic women at base-line, we calculated that 154 women would be needed to have 80% power in the trial. However, in our study the difference between the groups at 6 months was less than the expected 30%. Therefore, the power of the study was limited to show differences in hot fl ushes.

Future studies
The results of recent studies (33,34) have suggested that women with hot fl ushes may have adverse vascular changes. It is possible that symptomatic women may derive more benefi t from aerobic training as regards both their general and vascular health, since aerobic training is benefi cial for mental and physical health. Our aim was to study the effi cacy of aerobic training on menopausal symptoms rather than effects on a general population; therefore we recruited sedentary, recently menopausal women. Further studies are needed to fi nd out the most effective exercise (aerobic versus non-aerobic) among suffi ciently active and fi t women with menopausal symptoms. Also long-term compliance with exercise is a challenge to ours as well as to any study concerning exercise and health. Forthcoming studies need also to re-address the dose-response effect of exercise or other possible sham intervention. More experimental studies using mobile phone technology, different exercise training programmes, and fi tness measurements are needed to confi rm our fi ndings in other populations.

Conclusions
Aerobic training may decrease hot fl ushes, but more research is needed. Aerobic training improved menopausal and health-related quality of life among slightly overweight, symptomatic women.