PhilSPEN Online Journal of Parenteral and Enteral Nutrition
(Article 10 | POJ_0004.html) Issue January 2010 - January 2012: 75-81
Original Clinical Investigation
The Comparison between percentile and z-score in the BMI based nutrition screening of pediatric patients in the out-patient department of the Institute of Pediatrics in St.Lukes Medical Center, Quezon City, Metro-Manila, Philippines
Submitted: September 10, 2010 | Posted January 21, 2011
Edna P. Llido MD, FPPS (1), Cynthia Aquino MD, FPPS (1), Mary Anne B. Santos MD, DPBCN (1,2), Luisito O. Llido MD, DPBCN (1,2)
Institution where research was conducted:
ABSTRACT: | Back
Objective: To determine which reference to use in the BMI based nutrition screening of pediatric patients: percentile versus z-score.
Methodology: All pediatric patients in the OPD were nutritionally screened using the WHO 2006 to 2007 BMI standards from February to June 2011. Anthropometric data taken were heights in kg and weights in meter using calibrated instruments supervised by a clinical dietitian. A software was designed to determine the percentiles and z-score and the data was saved in the computer database. Differences were compared using T-Test or Wilcoxon signed rank test.
Results: A total of 1,766 patients were included in the study with a male to female ratio of 1.04 to 1. There was no difference between percentile and z-score in the following categories: normal BMI (83%), overweight BMI (7.6%), and obese BMI (2.7%). A difference was seen in the underweight BMI category (percentile = 7.4% vs. z-score = 6.7%, p < 0.05, T-Test) and this was noted to be due to an additional classification of severe underweight in the z-score standard.
Conclusion: There is no difference in the results of BMI based nutrition screening using either z-score or percentile for normal, overweight, and obese nutritional status in pediatric patients except for the underweight due to an additional classification of severe underweight in the z-score category.
INTRODUCTION | Back
Nutrition screening of pediatric patients is part of the over-all care of pediatric patients both in the out-patient and in-patient setting. Since it was established that nutritional status and growth patterns differ between ages and sex from the normal to the extremes of body composition like underweight or obese, standards of growth and development were developed to serve as guides for declaring a pediatric patient as “normal” or “abnormal” in either growth or stature. The normal curves developed by National Center for Health Statistics and Center for Disease Control (CDC)  from the United States and from the World Health Organization (WHO)  were adapted for use in the Philippines and nutrition agencies of the Philippines like the Food and Nutrition Research Institute (FNRI) together with the Philippine Pediatric Society which made initiatives to develop curves and reference values from the local pediatric population . The FNRI standards were then born and after a few more years the International Reference Standards (IRS)  were developed to serve as updates from the FNRI standards. The use of all four were found to be confusing so the Pediatric Department of St. Luke’s Medical Center made a study to determine which of these four reference standards could be used.  It was seen that the WHO-MGRS  and CDC growth standards came closest to depicting the normal growth curve distribution pattern, therefore, the current WHO 2006-7  is now used by the institution for its nutrition screening and nutritional assessment functions. The WHO standards were chosen over the CDC since it covered the normal population and the age range covered was from 0 to 18 years old.
One more issue raised by the pediatricians was: most of them are using the BMI percentile values and there are calls from some sectors to adopt the BMI z-score as the standard to follow. In order to resolve this issue another study was designed to compare the difference(s), if there was any, between the BMI percentiles and the z-score of the WHO 2006-7 growth standards. The objectives of the study are: a) To validate the nutritional status values based on the World Health Organization (WHO) Body Mass Index (BMI) nutritional status charts on pediatric patients of the SLMC Institute of Pediatrics; b) To compare the z-score and percentile cut off results for nutritional status from the pediatric patients of the SLMC Institute of Pediatrics; and c )To decide which criteria to use for the nutrition screening of pediatric patients.
METHODOLOGY | Back
The nutrition screening tool for evaluation is the WHO BMI based nutritional status form modified for use for the Institute of Pediatrics of St. Luke’s Medical Center (Figures 1 and 2). The Inclusion criteria are: all pediatric patients seen at the Out Patient Department of St. Luke’s Medical Center. In order to facilitate the data gathering process, a computer program was developed for both database and computation purposes. These are the data gathered: patient identification (Last Name, First Name, Middle Name), age, sex, height in meters, and weight in kilograms, and Body Mass Index (BMI). The anthropometric measurements and other data were taken by pediatric residents and consultants with rotating assignments at the Out-Patient Department of the Institute of Pediatrics. The weight was taken from DETECTO (Webb City, Missouri) stand weighing scale with height measuring bar and infant weighing scale and length measuring mat by SECA (Hangzhou, China).
Table 1: Comparison between z-score and percentile
Figure 1: Pediatric BMI based Nutrition Screening Form for Boys; Note the integrated curves of both BMI percentiles and Z-score
Figure 2: Pediatric BMI based Nutrition Screening Form for Girls
The standard procedures followed were: a) Height in meter(s) and weight in kilogram(s) were taken from the patients based on correct methods of taking height and weight of pediatric patients as adopted by the nursing department on correct nursing procedures; b) Height in meters and weight in kilograms from which the BMI and z-score are automatically computed by the computer program are taken and encoded; c) the number of patients within all categories are summed up and the difference between all groups are determined.
These are the outcomes assessed: a) Percentage of normal, underweight, severe underweight, overweight and obese in the Filipino pediatric population seen in the OPD of SLMC based on the WHO BMI classification, b) Presence or absence of differences in the number of patients included in the different cut-offs of either z-score or percentiles, classified as severe underweight, underweight, normal, overweight, and obese are determined using paired T-Test or Wilcoxon Signed Rank Test. between all groups are determined.
These are the outcomes assessed: a) Percentage of normal, underweight, severe underweight, overweight and obese in the Filipino pediatric population seen in the OPD of SLMC based on the WHO BMI classification, b) Presence or absence of differences in the number of patients included in the different cut-offs of either z-score or percentiles, classified as severe underweight, underweight, normal, overweight, and obese are determined using paired T-Test or Wilcoxon Signed Rank Test. 
RESULTS | Back
Patient profile: There was a total of 1,766 out-patient pediatric patients included in the study with the male to female ratio at 1.04 to 1. Most of these patients were either well-baby follow ups or regular check-ups at the outpatient clinic.
DISCUSSION: | Back
It is shown in this study that using the BMI percentile or BMI z-score would yield similar results thus either of these two standards can be used for the nutrition screening of pediatric patients. The significant result for the underweight classification (a difference between BMI percentile [131 or 7.4%] and BMI z-score [119 or 6.7%], p<0.05, Wilcoxon Signed Rank Test, table 2) was due to the additional classification of severe underweight in the BMI z-score standards (below -3) which was not present in the BMI percentile classification. However when the underweight and severe underweight in the BMI z-score were combined only as “underweight”, no difference was noted between both groups.
An advantage of using the BMI z-score would be in the determination of severely underweight patients, which classification was not included in the BMI percentile cut-off values. Another advantage was earlier cited by some pediatricians who pointed out that most of the reported BMI values presented in the literature were in the form of z-score. In the overall scheme of things if the goal is merely to determine the nutritional status of a pediatric patient as normal, underweight, overweight, or obese then either the BMI percentile or the BMI z-score can be used. The forms for nutrition screening used in this study (Figures 1 and 2) are thus acceptable for use either in the hospital or community set-up. The results of this study is in full accord with reports from other countries on the use of the WHO 2006-7 BMI based child growth standards (Canada , United States , United Kingdom , Eastern Mediterranean , and Middle East ).
CONCLUSION: | Back
There is no difference in the results of nutrition screening using either z-score or percentile for normal, overweight, and obese. There is a difference in the underweight class due to an additional classification (severe underweight) in the z-score group, however, if these were combined no difference in number between percentile and z-score is noted.
REFERENCES: | Back