Cardiorespiratory fitness levels and associations with physical activity and body composition in young South African adults from Soweto

South Africa is undergoing rapid urbanisation and experiencing consequences typical of a transitioning society, such as increased physical inactivity and obesity [1]. Self-report physical activity data from the 2010 World Health Organisation (WHO) global health status report [2] have shown that 55.7% of female and 46.4% of male adolescents were inactive according to American College of Sports Medicine (ACSM) guidelines (150 min of moderate activity or 60 min of vigorous activity per week). Furthermore, 58.5% of males and 71.8% of females over the age of 20 years were overweight [2]. Soweto is an urban township in South Africa that clearly exhibits these transitions, with women showing particularly high and increasing rates of overweight and obesity. The increasing prevalence of overweight and obesity, as well as physical inactivity in South Africa [3], makes it extremely important to start examining correlates of obesity and physical activity in young adults, which is a time when these trajectories are being set up [4].

Increased cardiorespiratory fitness, has been associated with decreased overall mortality and morbidity, independent of other risk factors such as high body mass index (BMI) [5]. Fitness data in South Africa is limited, particularly in young adults. The South African National Health and Nutrition Examination Survey (SANHANES) data in 2013 show that only 66% of males and 38% of females (18–24 year old) were considered fit [1]. Fitness and physical activity have been associated with improved cardiovascular and metabolic health; as well as reduced adiposity in adults, adolescents and children [6, 7].

The associations between fitness, habitual physical activity levels and BMI in young South African adults is not well understood. Furthermore, the factors that influence fitness in this population have not yet been fully described. Therefore, this study aims to describe fitness and physical activity levels and patterns in a group of young black South African adults from Soweto, and to examine the associations between fitness, habitual physical activity and BMI.

Participant characteristics are presented in Table 1, stratified by sex. Data for all participants with fitness results (n = 409) are presented in all cases except for the physical activity variables where data for only 256 participants with valid accelerometer data are shown. There were no differences between participants with valid physical activity data compared to those without (Supplementary Table 1). Males were significantly taller and had a lower BMI than females. The majority (75%) of the males were classified as having a normal BMI, while 26% of the females were classified as overweight and 9% as obese, compared to only 6% and 2% of the males respectively (p < 0.001).

Males had a higher estimated VO_2max, and significantly more males were classified as active (96% of males vs 83% of females using the CDC guidelines, p < 0.001) compared to females. Time spent in both moderate and vigorous intensity physical activity was significantly higher (both p < 0.001), while sedentary time was significantly lower (p < 0.01), in males than females (Fig. 1). The majority of the participants (90%) met the CDC recommendations for physical activity, however when WHO guidelines, which include activity bouts, were assessed the percentage of participants meeting guidelines decreased to 40% in males and 20% in females. Significantly more males than females were classified as active using both CDC and WHO criteria (both p < 0.001).

Results of the multiple regression are presented in Table 2. Higher MVPA, lower BMI, and sex (being male) were all significantly associated with higher VO_2max, however sedentary time was not associated with fitness (R² for the model =0.60). This analysis was powered at a level of 1.0. Regression means for time in MVPA were 73 min/day for fit individuals, and 58 min/day for unfit individuals, and regression means for BMI were 21.35 kg/m² for fit and 23.41 kg/m² for unfit individuals.

Estimated VO_2max according to BMI category is presented in Fig. 2, for males and females separately. In both males and females, estimated VO_2max was higher in underweight compared to overweight and obese participants (p < 0.05), and higher in normal weight compared to overweight and obese participants (p < 0.0001).

This study aimed to describe fitness and physical activity levels in young black South African adults from an urban township in Johannesburg (Soweto), as well as to determine whether fitness is associated with physical activity levels and BMI in this population. Significant and independent associations were found between MVPA and fitness, as well as between BMI and fitness. The majority of participants (90%) met the physical activity guidelines without prescriptions for bouted activity, yet only 40% met the guidelines for bouted activity. Both fitness, and time spent in moderate and vigorous intensity physical activity, were significantly higher in males than females. Females spent significantly more time in sedentary behaviours, although sedentary time was high (mean of 9 h and 20 min) for all participants. BMI was also significantly higher in females than males, and so it can be concluded that females in this population are at greatest risk of disease consequences associated with reduced fitness, high levels of physical inactivity and a high prevalence of overweight and obesity.

The present study, in concurrence with many others [12, 15, 16], reported that males are more physically active than females. In both males and females, time spent in MVPA was positively associated with fitness, where fit individuals spent 15 more minutes per day in MVPA than unfit individuals, yet sedentary time was not associated with fitness. SANHANES data showed that 66% of 18–24 year old males were fit, while only 38% of females of the same age were classified as fit, as assessed using a three minute submaximal step test [1]. These percentages varied by location and race, and physical activity levels were not reported. Objectively measured energy expenditure data from the Physical Activity and Health Longitudinal (PAHL) study in South Africa reported that 16 year old girls were more active than boys, spending 61 min/day in MVPA compared to 35.0 min/day in boys. However, the results may have been influenced by the higher fat mass present in girls and fitness was not assessed in the PAHL participants [17]. There is a paucity of data examining associations between physical activity and fitness in young adults, specifically in South Africa, and differences in measurements used makes comparison of the available data difficult. Data from the Helena study in younger European adolescents (aged 12–18 years) [18], reported similar findings to ours as higher fitness levels (as assessed using a 20 m shuttle test), were associated with higher physical activity levels and less time spent sedentary. Adults (aged 18–49 years) from the National Health and Nutrition Examination Survey (NHANES) database had similar fitness levels to our participants assessed using a submaximal treadmill test (42.6 vs 41.9 mlO₂/kg/min for males and 35.7 vs 32.6 mlO₂/kg/min for females). In the NHANES data, fitness was associated with MVPA, but not sedentary time [19].

The observed differences in fitness between males and females in the present study, besides being associated with physical activity levels, may also be influenced by differences in BMI. Similarly to what has been shown in other studies, BMI was inversely associated with estimated VO_2max. We have also shown significant differences in BMI between males and females and reported a high prevalence of overweight and obese females; however this prevalence is still lower (8% compared to 10% in males, and 35% compared to 47% in females) than that reported in the SANHANES data on young South African adults (18–24 years) [1]. Approximately 15% of participants in this study were underweight, yet VO_2max was significantly higher in underweight compared to obese males and females. This concurs with current literature which shows that low and normal BMIs have beneficial effects on fitness in adolescents and young adults [20, 21], likely due to the bidirectional associations between functional respiratory impairment and increased energy expenditure requirements from overweight and obesity [22, 23], but also due to the size dependant nature of the measure.

In this cohort of urban black young adults 96% of males and 83% of females met the recommended CDC physical activity guidelines of 150 min MVPA per week or 75 min vigorous activity per week. However, when using the WHO recommendations that require this activity to be performed in bouts of a minimum of 10 min, only 40% of males and 20% of females were classified as physically active. From these data it would suggest that these young adults are accumulating sufficient amounts of MVPA throughout the day, yet are accumulating this activity sporadically rather than in bouts of 10 min or more. Recent studies using NHANES data have shown that when comparing bouts of moderate and vigorous activity to total accumulated moderate and vigorous activity, bouted activity infers greater benefits on adiposity, yet has similar benefits as accumulated activity on cardiometabolic risk factors [24, 25]. Conversely to our findings, self-report data from previous South African studies presented in the WHO global health status report have shown that 46% of male adolescents and 56% of female adolescents 1–2 years younger than our cohort are inactive when using similar guidelines to the CDC recommendations [2]. Self-report SADHS data from 19 year old urban South Africans have also shown that 49% of males and 75% of females were inactive [3] using GPAQ (WHO) classifications. It is possible that the peri-urban setting from which the present cohort was drawn resulted in higher levels of physical activity for commuting purposes contributing to the accumulation of the recommended amount of MVPA per week, yet lower levels of structured leisure time physical activity thus resulting in sporadic accumulation of higher intensity activity throughout the day. Indeed, the SADHS data showed lower inactivity levels in non-urban vs urban females (75% vs 82% inactivity), but reported no differences for males. It is also possible that discrepancies between self-report and objectively assessed data could be affecting these comparisons. Self-report physical activity data is prone to inaccuracy, and the objective assessment of physical activity levels and patterns in the present study should be considered a strength.

Although the majority of this sample met unbouted physical activity recommendations, sedentary time equated to an average of just less than 9 and a half hours per day (>50% of waking hours) for both groups, while time spent in vigorous intensity activity equated to less than 1% of the day. To our knowledge, this is the only data reporting objectively measured sedentary time in this age group in Soweto. One study in older adult females (mean age 41 years) from Soweto showed that sedentary time was lower - on average 3 h per day; and was not different between those women classified as physically active vs inactive [26]. Furthermore 90% of women in this study did not perform any vigorous intensity activity. However, the 2016 Healthy Active Kids South Africa report has shown that children up until the age of 17 are exceeding sedentary recommendations, and that sedentary time (largely assessed as screen time) increases with age [27]. This high sedentary time in our study is similar to what has been found in European adolescents from the HELENA study [18], who spent 9 h per day being sedentary, but higher than findings from American adolescents and young adults who are reported to spend 7–8 h per day sedentary [28]. With regard to physical activity patterns, similar to the current findings, previous self-report data from young (26 ± 7 years) South African women showed no time reported in vigorous intensity activities per week [7]. Vigorous activity is likely to be accumulated during leisure time physical activity or sports participation, rather than during general daily activities. This implies low levels of voluntary high intensity physical activity participation in this population, confirming findings from studies in other South African cohorts [7].

This high sedentary time and low amounts of vigorous intensity activity, in combination with a high prevalence of overweight and obesity; and the relationships with fitness levels puts these young adults at risk of future cardiometabolic disease. This relationship is most likely bidirectional, where high levels of sedentary behaviour and low fitness results in excess adiposity, thereby increasing BMI. Increased adiposity likely results in further increases in sedentary time, potentially at the expense of moderate and vigorous intensity activity and lower fitness (illustrated in Fig. 3). This vicious cycle of deleterious effects requires intervention, seemingly directed at increasing MVPA; and decreasing BMI and sedentary time, which may allow for improvements in fitness and thus cardiometabolic health. Particular focus should be placed on targeting young female adults, who are at increased risk due to their lower physical activity and fitness levels, higher sedentary time, and high prevalence of overweight and obesity. Considerations of the population and gender specific barriers and factors that could affect interventions should also be made [29, 30]. The findings of this study highlight the importance of increasing MVPA for improvements in fitness; and previous studies have also shown increases in fitness in relation to increases in physical activity (particularly vigorous intensity activity) in populations of similar ages [31, 32]. However, the importance of decreasing sedentary time should not be overlooked. In conjunction with walking for transport, which is already common in this population but tends to be performed at lower intensities, young South African women should also be encouraged to participate in structured, high intensity physical activity on most days of the week, and to reduce sedentary time as much as possible. Potentially, this physical activity should be accumulated in bouts of 10 min of more in order to confer maximum beneficial effects on body composition, thus allowing for improved fitness.

Limitations of this study include the cross sectional design, which limits any causality conclusions. Furthermore, the lack of detailed body composition data prevents deeper analysis into possible correlates of fitness in this cohort, although BMI is a commonly used proxy for body composition. Although the objective measurement of physical activity is a strength of this study, more detail into the context of these activity behaviours would have also been beneficial. The limited number of participants with complete accelerometer data decreased the sample size for physical activity analysis, yet these analyses were still sufficiently powered.

In conclusion, the majority of young South African adults in this study are sufficiently active, yet are not accumulating activity in prolonged bouts, which may be detrimental to body composition. Moderate to vigorous intensity physical activity and BMI are both strongly and independently associated with fitness. The large amount of time spent in sedentary behaviour observed in these young adults is of concern and should be a major focus of any intervention; along with increasing structured, higher intensity activity and decreasing BMI, and focus should be placed on young adult females.