MATERIALS AND METHODS: A systematic literature search was performed through SCOPUS database and Google Scholar from January till March 2018. All published articles which developed stature estimation from different types of bone, methods and type of statures (i.e. living stature, forensic stature and cadaveric stature) were included in this study. Risks of biases were also assessed. Population studies with no regression equations were excluded from the study.
RESULTS: Seven studies that met the inclusion criteria were identified. In the South-East Asia region, regression equations for stature estimation were developed in Thailand and Malaysia. In these studies, bone measurements were done either by radiography, direct bone measurement, or palpation on body surface for anatomical bony prominence. All of these studies used various parts of bones for stature estimation.
CONCLUSION: The most widely used regression equations for stature estimation in South-East Asian population were from the Thailand population. Further research is recommended to develop regression equations for other South-East Asian countries.
MATERIALS AND METHODS: We retrospectively assessed 107 cadavers that had undergone conventional autopsy and PMCT. We made 5 measurements from the PMCT that included cervical length (CL), thoracic length (TL), lumbosacral length (LS), total column length of the spine, excluding the sacrum and coccyx (TCL), and ellipse line measurement of the whole spine, excluding the sacrum and coccyx (EL). We compared these anthropometric PMCT measurements with AL and correlated them using linear regression analysis.
RESULTS: The results showed a significant linear relationship existed between TL and LS with AL, which was higher in comparison with the other parameters than the rest of the spine parameters. The linear regression formula derived was: 48.163 + 2.458 (TL) + 2.246 (LS).
CONCLUSIONS: The linear regression formula derived from PMCT spine length parameters particularly thoracic and lumbar spine gave a finer correlation with autopsy body length and can be used for accurate estimation of cadaveric height. To the best of our knowledge, this is the first ever linear regression formula for cadaveric height assessment using only post mortem CT spine length measurements.
DESIGN: Data on length/height-for-age percentile values were collected. The LMS method was used for calculating smoothened percentile values. Standardized site effects (SSE) were used for identifying large or unacceptable differences (i.e. $\mid\! \rm SSE \!\mid$ >0·5) between the pooled SEANUTS sample (including all countries) and the remaining pooled SEANUTS samples (including three countries) after weighting sample sizes and excluding one single country each time, as well as with WHO growth references.
SETTING: Malaysia, Thailand, Vietnam and Indonesia.
SUBJECTS: Data from 14202 eligible children were used.
RESULTS: From pair-wise comparisons of percentile values between the pooled SEANUTS sample and the remaining pooled SEANUTS samples, the vast majority of differences were acceptable (i.e. $\mid\! \rm SSE \!\mid$ ≤0·5). In contrast, pair-wise comparisons of percentile values between the pooled SEANUTS sample and WHO revealed large differences.
CONCLUSIONS: The current study calculated length/height percentile values for South East Asian children aged 0·5-12 years and supported the appropriateness of using pooled SEANUTS length/height percentile values for assessing children's growth instead of country-specific ones. Pooled SEANUTS percentile values were found to differ from the WHO growth references and therefore this should be kept in mind when using WHO growth curves to assess length/height in these populations.
METHODOLOGY: A cross-sectional study with a universal sampling of children and adolescents with special needs aged 2-18 years old, diagnosed with cerebral palsy, down syndrome, autism and attention-deficit/hyperactivity disorder was conducted at Community-Based Rehabilitation in Central Zone Malaysia. Socio-demographic data were obtained from files, and medical reports and anthropometric measurements (body weight, height, humeral length, and mid-upper arm circumference) were collected using standard procedures. Data were analysed using IBM SPSS version 26. The accuracy of the formula was determined by intraclass correlation, prediction at 20% of actual body weight, residual error (RE) and root mean square error (RMSE).
RESULT: A total of 502 children with a median age of 7 (6) years were enrolled in this study. The results showed that the Mercy formula demonstrated a smaller degree of bias than the Cattermole formula (PE = 1.97 ± 15.99% and 21.13 ± 27.76%, respectively). The Mercy formula showed the highest intraclass correlation coefficient (0.936 vs. 0.858) and predicted weight within 20% of the actual value in the largest proportion of participants (84% vs. 48%). The Mercy formula also demonstrated lower RE (0.3 vs. 3.6) and RMSE (3.84 vs. 6.56) compared to the Cattermole formula. Mercy offered the best option for weight estimation in children with special needs in our study population.