**3.3 Multivariate models for predicting BP**

Hierarchical multiple regression analyses were conducted to determine the joint associations of fatness and LP with BP controlling for MO in both genders (**Table 4**). *Association of Fatness and Leg Power with Blood Pressure in Adolescents DOI: http://dx.doi.org/10.5772/intechopen.106279*

For the girls' SBP model, the covariate explained only 4.3% of the variance in step 1. The addition of BMI and VJP in step 2 increased the total variance to 10.4% indicating that both the independent variables explained an additional variance of 6.1%. LP (*p* < 0.001) and MO (*p* = 0.014) were the significant predictors, with VJP presenting greater explanatory capacity. In the model for boys, fatness and LP explained 23.1% with only 8.9% contribution from MO. All variables made significant contributions, but MO presented the greatest explanatory power. In the model for DBP, only MO and BMI made significant (*p* < 0.05) contributions in girls, while in the boys' model, no independent variable made any statistically significant (*P* > 0.005) contribution. Details of the results can be found in **Table 4**.

Results of the logistic regression models (**Table 5**) indicated that in general only VJP and BMI made significant contributions, which were greater in girls. In the girls' model, both fatness (OR = 2.6, 95% CI = 1.29–5.35; *p* = 0.008) and LP (OR = 0.40, 95% CI = 0.25–0.64; *p* < 0.001) were associated with SBP. These results indicate that fat girls were 2.6 times likely to develop risk of HTN compared to their healthy weight peers. Further, the odd of HTN risk in girls with low LP was 0.40 times that of their counterparts with greater LP. As a whole, the model was able to explain between 4–9% of the variance in SBP and correctly classified 90.2% of the cases. In the Boys' model, no variable made any significant contribution. However, the model also explained between 4–9% of the variation in SBP and correctly classified 90.2% of the cases. For the DBP models, only MO (OR = 1.40, 95%CI = 1.15–1.62, *p* < 0.001) in girls made a significant contribution to the model. The model for boys was not significant (*p* < 0.001). Details of the results are in **Table 5**.
