philpan1c

PhilSPEN Online Journal of Parenteral and Enteral Nutrition

(Article 3 | POJ_0003.html) Issue January 2010 - January 2012: 17-32

Original Clinical Investigation

Comparison of standard values of nutrition screening and assessment using BMI percentiles from FNRI-PPS, IRS, CDC 2000, and WHO-MGRS child growth standards in the pediatric population of a tertiary care hospital in the Philippines admitted between years 2000 and 2003

Abstract | Introduction | Methodology | Results | Discussion | Conclusion | References | PDF (1.22 MB) |Back to Articles Page

Submitted: May 8, 2008 | Posted: August 29, 2010

Authors:

Edna P. Llido MD, Mercedita M. Macalintal MD, Ma. Christina Reyes MD, Nellie D. Gundao MD, Mary Anne B. Santos MD, Donabelle Faye I. Navarrete RND

Institution where research was conducted:

Pediatric Clinical Nutrition Section, Obesity and Weight Management Center, St. Luke’s Medical Center, Metro Manila, Philippines

 

ABSTRACT: | Back

Background: Nutrition screening in pediatrics patients commonly uses standards to evaluate the patient’s body composition. The nutrition support group in a private tertiary care hospital in the Philippines planned to develop a BMI based nutrition screening form based on existing standards: FNRI-PPS, IRS, CDC, and WHO-MGRS child growth standards.

Objectives: To identify the standard that best reflected the admitted Filipino pediatric population.

Methodology: Data from 24,957 pediatric patients aged 1 month to 18 years (1.25:1 male to female ratio) were collected from years 2000-2003. BMI percentiles (p5, p50, and p95) were developed from all the reference standards and the number of patients who fell into the following categories: below p5, between p5 and p95, and above p95, were counted and compared as to which come closest to the normal distribution.

Results: Analysis per age group showed that the CDC standard and WHO-MGRS child growth standards values followed normal distribution patterns with the 5th to 95th percentile higher (CDC:63%, WHO-MGRS:62%) compared to below 5th percentile (CDC:16%, WHO-MGRS:14%) and above 95th percentile (CDC:22%, WHO-MGRS:24%), whereas FNRI-PPS and IRS showed a lower distribution in the 5th to 95th percentile (FNRI-PPS:17%, IRS:38%), but higher in the below p5 (FNRI-PPS:26%, IRS:30%) and above 95 percentile (FNRI-PPS:56%, IRS:32%). The values of the WHO-MGRS child growth standards in the one to 12 month group come closer to standard compared to the CDC (5th to 95th: WHO-MGRS [57%] versus CDC [54%]; below 5th: WHO-MGRS [19%] versus CDC [15%]; above 95th: WHO-MGRS [24%] versus CDC [31%]) while similar results were seen in the 2 to 18 year olds (5th to 95th: WHO-MGRS [62%] versus CDC [63%]; below 5th: WHO-MGRS [14%] versus CDC [16%]; above 95th: WHO-MGRS [24%] versus CDC [22%]).

Conclusion: We conclude that the WHO-MGRS child growth standard is the preferred tool for use in BMI-based nutrition screening of pediatric patients in the 1 to 12 month age group, whereas the CDC and WHO-MGRS growth standards yield similar results for the 2 to 18 year olds.

 

KEYWORDS: BMI, percentiles, WHO-MGRS, CDC, FNRI-PPS, IRS, Philippines

INTRODUCTION | Back

Nutrition screening in pediatric patients use anthropometric data, basically height and weight, to get a good evaluation of the patient’s body composition and health. Growth curves based on standards from the normal population are used to track the patient’s progress in each age group which would determine if they are within normal range or not. Among the commonly used standards in the Philippines are the WHO (World Health Organization) child growth curves (1), NHANES and CDC (Center for Disease Control) in the United States (2), and in the Philippines by the FNRI-PPS (Food and Nutrition Research Institute-Philippine Pediatric Society) (3). Full utilization of the growth curves prepared by the FNRI-PPS since 1993 was not realized due to some questions on its reliability especially by some pediatricians, who preferred to use the CDC growth curves. This issue was partially resolved by the development of the IRS (International Reference Standards) in 2003 (4). Currently new curves based on the WHO-MGRS (Multicentre Growth Reference Study) were made available for use (5).

The nutrition support group in St. Luke’s Medical Center needed a simple nutrition screening tool for its pediatric population, which is required for the standards of care for hospitalized patients as mandated by JCAHO (Joint Commission on Accreditation of Health Care Organizations) (6) so it planned to develop its own growth curves adopted from the different existing standards, which would become part of a rapid nutrition screening system for all admitted pediatric patients. To evaluate its reliability, all values from the FNRI, IRS, CDC, and WHO-MGRS were used on a specific sample population from the clinical nutrition database of the center. The objectives of this study were: a) to determine which of these growth standards would yield the pattern closest to the normal distribution and b) to choose which standard will be used for the nutrition screening purposes of this institution.

 

METHODOLOGY | Back

The normal distribution of the BMI values of the pediatric population would follow this pattern: 90% of the total count would fall within the 5th to 95th percentile, 5% are within the below 5th percentile, and 5% are within the above 95th percentile, thus forming the normal bell-shaped curve (7). Placing the resulting percentage values of the study population in a graph using the Microsoft Excel 2003 worksheet and enabling the “smooth line” function of the “format data series” for each standard would yield a smooth curve, thus showing a clear picture of the BMI distribution using the different anthropometric standards compared in this study.

The FNRI-PPS and IRS standards did not have BMI (Body Mass Index) values for use so the group derived it from the percentile values of both height and weight per age and sex and computed for the BMI for each percentile to be compared (P5 or 5th percentile, P50 or 50th percentile, and P95 or 95th percentile). The CDC did not have BMI for the ages 1 month to 12 months so the BMI was also derived from the percentile values of height and weight for the specific age and sex. All reference values are shown in Table 2 for the boys and Table 3 for the girls.

tbl02a

tbl03a

Back to Discussion

Pediatric patients aged 1 month to 18 years were collected from the St. Luke’s Medical Center clinical nutrition database from years 2000 to 2003. The age, sex, height and weight which were converted to BMI values were gathered. Percentile cut off values of the BMI from the different standards were used to segregate the population into lower P5, between P5 and P50, between P50 and P95, and above P95. The lower P5 represents the underweight population, the group within the P5 and P95 values represents the normal population, and the group above the P95 represents the obese population. The percentile grouping was analyzed per age group. Data analyzed were: number and percentages in the different nutritional status groups using the following standards: FNRI-PPS, IRS, CDC, and WHO-MGRS.        

 

RESULTS | Back

A total of 24,957 patients were gathered with 13,887 males and 11,070 females (male to female ratio of 1.25 is to 1, Table 1). The BMI distribution showed skewness to the right (skewness = 2), with a mean of 18 and standard deviation of 4.95 (Figure 1).

fig01atbl_01a

Standards comparison, boys, 1 to 12 months (Figure 2, Table 4 and 6A): The FNRI-PPS standards yielded more patients on the outlying 95th percentile (>p95: 1,990 [54%]), while the less than 5th percentile and 5th to 95th percentile have similar numbers (<p5: 899 [25%] and p5_p95: 785 [21%]). The IRS standards also yielded a similar pattern (>p95:1,373 [37%]; <p5: 1,226 [33%]; p5_p95: 1,077 [29%]). The WHO-MGRS and CDC standards showed the highest numbers in the 5th to 95th percentile (WHO-MGRS: 2,020 [55%]; CDC: 1,899 [52%]) with lower numbers in the below 5th percentile (WHO-MGRS: 668 [18%]; CDC: 452 [12%]) and above 95th percentile values (WHO-MGRS: 986 [27%]; CDC: 1,323 [36%]).

fig02a

tbl04a

tbl06A

Standards comparison, boys, 2 to 18 years (Figure 3, Table 4 and 6B): The FNRI-PPS standards showed similar pattern with the 1 to 12 month boys group having more values in the below 5th percentile (n=2,509 [25%]) and above 95th percentile (n=6,006 [58%]) compared to the central 5th to 95th percentile (n=1,698 [18%]). The CDC and WHO-MGRS showed highest values for the 5th and 95th percentile (CDC: 6,396 [60%]; WHO-MGRS: 6,203 [59%]) and lower values for the below 5th percentile (CDC: 1,634 [15%] versus WHO-MGRS: 1,230 [14%]) and above 95th percentile (CDC: 2,183 [25%] versus WHO-MGRS: 2,780 [27%]). The CDC however had higher values compared to the WHO-MGRS in the 5th to 95th percentile. The IRS had similar pattern with the CDC and WHO-MGRS, but its values were lower in the central p5-p95 group (n=3,978 [36%]) and higher in the <p5 (n=2,915 [30%]) and >p95 (n=3,320 [34%]).

fig03a

tbl06B

Standards comparison, girls, 1 to 12 months (Figure 4, Tables 5 and 6A): The FNRI-PPS and IRS standards showed similar patterns of higher values for the below 5th percentile (FNRI-PPS: 913 [34%] versus IRS: 1,035 [38%]) and above 95th percentile (FNRI-PPS: 1,359 [50%] versus IRS: 941 [35%]) compared to the central 5th to 95th percentile (FNRI-PPS: 430 [16%] versus IRS: 726 [27%]). The CDC and WHO-MGRS, however, showed highest values for the 5th to 95th percentile (CDC: 1,528 [57%]; WHO-MGRS: 1,587 [59%]) and lower values for the below 5th percentile (CDC: 507 [19%]; WHO-MGRS: 544 [20%]) and above 95th percentile (CDC: 667 [25%]; WHO-MGRS: 571 [21%]).

fig04a

tbl05a

Standards comparison, girls, 2 to 18 years (Figure 5, Table 5 and 6B): The FNRI-PPS standard showed lower 5th to 95th percentile values (n=1,399 [17%]) compared to the outlying below 5th percentile (n=2,229 [28%]) and above 95th percentile (n=4,740 [55%]) while the CDC, WHO-MGRS, and IRS showed the opposite (<5th percentile: CDC: 1,310 [16%];  WHO-MGRS: 1,001 [14%], and IRS: 2,377 [31%]; >95th percentile: CDC: 1,220 [17%]; WHO-MGRS: 1,662 [20%], and IRS: 2,292 [29%]; 5th to 95th percentile: CDC: 5,838 [67%], WHO-MGRS: 5,705 [66%], and IRS: 3,699 [40%]).

fig05a

Statistics summary (Table 6A, 6b, 6C, Figures 6 to 8): The CDC and the WHO-MGRS standards showed the highest numbers in the 5th to 95th percentile groups compared to the outlying below 5th percentile and above 95th percentile. In males 59% (WHO-MGRS) to 60% (CDC) of the population is within the 5th to 95th percentile group, 14% (WHO-MGRS) to 15% (CDC) is below the 5th percentile, and 25% (CDC) to 27% (WHO-MGRS) is above the 95th percentile. In females 66% (WHO-MGRS) to 67% (CDC) of the population is within the 5th to 95th percentile, 14% (WHO-MGRS) to 16% (CDC) are below the 5th percentile, and 17% (CDC) to 20% (CDC) is above the 95th percentile. The distribution pattern of the WHO-MGRS in both male and female population in the one to 12 month age group is more symmetrical and well distributed compared to the CDC growth standard, but in the two (2) to eighteen (18) year olds the distribution pattern is similar.

The IRS showed similar pattern as the CDC and WHO-MGRS, but had lower values in the central 5th to 95th percentile and higher values in the outlying below 5th and above 95th percentile. In males 36% of the population is within the 5th to 95th percentile group, 30% is below the 5th percentile, and 34% is above the 95th percentile. In females 40% of the population is within the 5th to 95th percentile, 31% are below the 5th percentile, and 29% is above the 95th percentile.

The FNRI-PPS values are the opposite of the WHO-MGRS, CDC, and IRS based data.  In males 18% of the population is within the 5th to 95th percentile group, 25% is below the 5th percentile, and 58% is above the 95th percentile. In females 17% of the population is within the 5th to 95th percentile, 28% is below the 5th percentile, and 55% is above the 95th percentile.

tbl06C

fig06a

Back to Discussion

fig07a

Back to Discussion

fig08a

DISCUSSION: | Back

The normal pattern of distribution for all the standards would be a higher 5th to 95th percentile number compared to a lower below 5th percentile and above 95th percentile number as would be expected from a normal population (7). The data showed that the FNRI-PPS standards resulted to an opposite pattern while the WHO-MGRS, CDC, and IRS showed the normal pattern. The WHO-MGRS and CDC standards yielded the highest results for the central 5th to 95th percentile group and lowest in the below 5th and above 95th percentile compared to the IRS standards. The WHO-MGRS and CDC standards thus come closest to producing the normal distribution pattern for the pediatric population. There was a difference between these two standards in the one to twelve month age group, where the WHO-MGRS standard yielded a more symmetrical distribution (Figure 6). This difference was not seen in the two to eighteen year old age group however (Figure 7).

The different results from the FNRI-PPS and IRS were analyzed by comparing the values of the four standards (FNRI-PPS, IRS, CDC, and WHO-MGRS) in the 5th, 50th, and 95th percentile groupings. The 5th percentile curves (Figure 9, in this instance showing the values from the female population; Table 3) showed the FNRI-PPS and IRS higher compared to the CDC and WHO-MGRS values. This was due to the assignment of higher cut-off values for both the FNRI-PPS and IRS standards, which resulted to higher “total count” values in the below 5th percentile and smaller “total count” values in the 5th to 95th percentile group in both FNRI-PPS and IRS standards compared to the CDC and WHO-MGRS. The 95th percentile curves (Figure 10, also showing the data from the female population; Table 3) showed higher “total count” values for the CDC and WHO-MGRS indicating lower cut-off values for the FNRI-PPS and IRS for the 95th percentile compared to the CDC and WHO-MGRS, thus assigning more patients in the above 95th percentile. This resulted to lower “total count” values in the 5th to 95th percentile group, but higher “total count” values above 95th percentile numbers for the FNRI-PPS and IRS standards.

fig09a

fig10a

Back to Discussion

CONCLUSION: | Back

The 5th percentile cut off values for the FNRI-PPS and IRS were higher compared to the WHO-MGRS and CDC standards while the 95th percentile cut off values were lower, thus resulting to higher outlying statistics in both below 5th and above 95th percentile groups.

We therefore conclude that the WHO-MGRS child growth standard is the preferred tool for use in BMI-based nutrition screening of pediatric patients in the 1 to 12 month age group, whereas the CDC and WHO-MGRS growth standards yield similar results for the 2 to 18 year olds.

 

REFERENCES: | Back

  1. World Health Organization. Physical status: The use and interpretation of anthropometry. Geneva, Switzerland: World Health Organization 1995. WHO Technical Report Series.
  2. National Center for Health Statistics in collaboration with the National Center for Chronic Disease Prevention and Health Promotion (2000). http://www.cdc.gov/growthcharts.
  3. Florentino RF, Santos-Ocampo P, et al. FNRI-PPS Anthropometric tables and charts for Filipino children. Manila: Food and Nutrition Research Institute, DOST and Philippine Pediatric Society, 1999.
  4. Mendoza TS and Barba C. A Handbook on International References Standards (IRS) growth tables and charts adopted for use in the Philippines. Food and Nutrition Research Institute, DOST and UNICEF, 2003.
  5. De Onis M, Garza C, Victora CG, Bhan MK, and Norum KR. The WHO Multicentre Growth Reference Study (MGRS): Rationale, planning, and implementation. Food and Nutrition Bulletin 2004; 25 (supplement 1): S3-S84.
  6. Standards for nutrition support: hospitalized patients. Joint Commission on Accreditation of Health Care Organizations, ASPEN Board of Directors. Nutr Clin Pract, 1995 Dec; 10 (6):208-19.
  7. Dawson B and Trapp RG, Basic and clinical biostatistics. 4th ed. Lange Medical Books/ McGraw-Hill, 2001: 26-57.

 

Abstract | Introduction | Methodology | Results | Discussion | References | Back to Articles Page