All data entry and analysis were performed using SPSS Version 15.0 (SPSS Inc. Chicago IL, United States, 2009). BMI calculated by the formula BMI = weight (kgs) / height (m
2). The reference data on L (lambda), M (mu), S (sigma) values and 3rd, 5th, 10th, 15th, 25th, 50th, 75th, 85th, 90th, 95th and 97th percentiles obtained after smoothing for weight, height and BMI by age data of the girls under study by using the software LMS Chartmaker Pro version 2.43 and the centile curves drawn after smoothing the values. Centiles were calculated by LMS and curves generated by the centile values calculated using the following formula: [
22]
$$ {C}_{100\alpha }(t)=M\ (t){\left[1+L\ (t)S\ (t)\ {Z}_{\alpha}\right]}^{1/L(t)} $$
where
Zα normal equivalent deviate for tail area
α, C100
α weight or height centile corresponding to
Zα,
t age in years, and L(
t) is skewedness (
t), M (
t) is median, S (
t) is coefficient variation and C
100α (
t) shows the corresponding values of each curve at age
t. The weight, height and BMI, edf (equivalent degrees of freedom) parameters used were L3M5S3, L0M7S4 and L2M4S3, in consultation and under guidance of the support team member of LMS Chartmaker Pro Huiqi Pan, the author of software LMS chartmaker Pro version 2.43 [
23] where L stands for Box-Cox power of transformation, M is the median and S is the coefficient of variation, the number after the letter is the edf value. Since weight, height and BMI values changed monotonously with the age therefore, the rescale option used in the LMS chartmaker and the best fitted graphs obtained. The smoothed L values for weight varied between - 1.2 and - 0.3, for height was 1.0 and the value for BMI was between − 0.5 to − 1.5. Data were presented as Mean ± SD for ‘n’ where n representing the number of study participants. The outliers, i.e.
, values above and below 4SD were considered outliers and not included in the analysis. The outliers were identified from the z-score plot of the parameter under consideration and excluded from the dataset. The mean weight, height and BMI data from the present sample were compared with the weight-, height- and BMI-for-age relative to the WHO (2007) and CDC (2000) references. Statistically significant differences between the means of weight-, height- and BMI-for-age relative to CDC and WHO references, and the present study were calculated with Student’s t-test. One-way analysis of variance (one-way ANOVA) followed by Tukey’s multiple comparison
post-hoc test was applied for comparison between multiple groups. The level of significance was accepted with
P < 0.05. The symbols (*and φ) indicated a significant difference between the study population and WHO, CDC references respectively according to Student’s t-test and the hash symbol (#) indicated a significant difference according to ANOVA tests between the groups that were being compared. The single, double and triple symbols exhibited the level of significance for
P < 0.05, 0.01 and 0.001 respectively.