The present study provides data on the prevalence of underweight, normal weight and overweight among adolescents, and the association between body composition and physical fitness separately for boys and girls aged 14 years old adjusted for race and locality in the Tlokwe Local Municipality of the North West Province of South Africa. Body composition is an important component of physical fitness of an individual and provides an indicator of the good well-being [1, 17, 21, 32, 46], hence should be emphasized as a way of healthy lifestyle among adolescents in this area [47, 48]. The present findings reflect the existence of both underweight (44%; 30.7%) and overweight/obesity (8%; 17.3%) respectively for boys and girls. Additionally, the results showed a strong significant positive association (OR, 10.69[95%CI: 2.81-40.73]) between physical fitness and BMI for underweight girls with high physical fitness scores, and overweight girls with high physical fitness scores (OR, 0.11 [95%CI: 0.03-0.50]).
Consistent with other previous studies in sub-Saharan Africa [3–6], and elsewhere [1, 2, 46], our findings showed that underweight was higher in boys than in girls on the one hand while on the other hand overweight/obesity was higher in girls than boys. It should be noted that the consequences of both underweight and obesity are reported to be related to decreased physical exercise/work capability and then reduced health-related physical fitness, such as cardiorespiratory fitness, strength and speed of movement [30–34, 46]. However, because overweight adolescents with low physical fitness in the study, especially girls tended to perform poorly in physical fitness as such physical activity interventions, or follow-up studies recognizing this relationship particularly in the overweight adolescents are needed [36, 48]. Despite that the present sample was less representative of the 2002 South African National Demographics and Health Survey , the rapid increasing of underweight and overweight can still be evident.
With regard to the physical fitness, the normal and underweight adolescent boys respectively, significantly outperformed the overweight boys in explosive strength although the normal group performed better than the other groups. Similarly, underweight girls in the present study performed better in explosive strength than the normal and overweight groups. The observed differences may be explained by the fact that both the normal and underweight groups, compared to the overweight group were able to carry an extra load to be moved during weight-bearing tasks [31–34]. Moreover, boys in the present study performed better than the girls and these findings are similar to the previous findings [23, 32]. The gender differences in physical fitness performance can be explained in part by gender difference in body composition. It is been reported that boys have greater muscle mass, bone density and less body fat than girls across age groups [17, 19, 30, 34, 49]. Previous research studies indicated that excessive fatness have a negative impact on the performance tasks in which the body is projected through space, as in standing broad jump, and on tasks in which the body must be lifted in space, as in bent arm hang [30–32, 38].
Additionally, both the underweight boys and girls were able to hang more seconds in bent arm hang than the overweight group. A study by Artero et al.  indicated that overweight and obese children can perform equally well or even better than children with normal weight in those muscular fitness tests, and such trend was found in the sit-ups test item especially for boys in the present study. No significant differences were found in sit-ups in the normal, underweight and obese groups in girls. Comparatively, the physical fitness performances of adolescents in this study were poorer than that of the AVENA study  and the Republic of Seychelles study .
The results showed strong ORs for the relationship between physical fitness and body composition for moderately fit overweight girls. The observed relationship may be explained by many factors, among others high level of physical fitness which may be attributed to household chores and walking distance to and from school , but unfortunately such information was not assessed in this study. The results show that children with high physical fitness have lower odds for being overweight than children with low physical fitness, and these findings are consistent with other studies. For example, research findings by Shang et al.  revealed that overweight and obese children performed worse in physical fitness exercises compared to normal weight children. Comparable findings from a cross-sectional study in adolescents from the Republic of Seychelles by Bovet et al.  reported a strong inverse relationship between physical fitness and excess body weight. In the present study overweight girls with low physical fitness performed worse in physical fitness than the boys and as such these differences may be explained by the fact that the development of BMI for girls is reported to be greater than that of boys from 12 years to 14.9 years . Since the development is greater it might also have more effect on the capability to perform physical fitness which may explain the stronger association found in girls [19, 49, 50]. From the epidemiological perspectives it is being reported that over-nutrition, which is associated with inadequate physical activity, is assumed to be the product of several risk factors during adolescence [24, 26, 29], hence the need for strategic intervention.
It should be realized that the current study has some limitations which requires caution in the interpretation of the data. It should be noted that physical fitness is a function of both physical activity and non-modifiable factors such as genetics [19, 31], and these factors were however not assessed in the present study and therefore, it was impossible to assess their relative contribution. This study was based on cross-sectional data (baseline measurements) of the PAHLS study, which is set up to be a five year longitudinal study on a group of 14 year-olds who will be followed up until they reach 18 years of age. Available data on measures of strength rather than cardiovascular health is a limitation of the study, which in future studies will be incorporated. In addition, the categorization of adolescents by their BMI may have had an effect on sample size, which as such may have contributed to observed large confidence intervals. Nevertheless, this cross-sectional study will contribute important information about the population of Tlokwe Local Municipality to the body of science. Another limitation of this study might be that the performed analyses are only adjusted for locality of the schools and race. Maturation may one way or the other have affected the results notwithstanding that the study subjects are drawn from a homogeneous age group; unfortunately no reliable data on maturation could be collected within the present study. With the planned longitudinal study we will be able to determine whether the observed prevalence of underweight and overweight as well as its relationship with physical fitness adjusted for other possible confounding factors, such as habitual physical activity, genetics factors, socio-economic status, dietary intake and other modifiable risk factors are consistent over time, and whether the changes in body composition measurements of BMI and %BF are related to changes in physical fitness.