Evaluation of a new infant nutrition screening tool (Infant Paediatric Yorkhill Malnutrition Score) and its applicability in Iran as compared to the United Kingdom

• Main Author: Milani, Shamsi Afiat
• Published: University of Glasgow 2016
• Subjects: 649; RJ Pediatrics
• Online Access: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.685874

A high prevalence of malnutrition has been reported in paediatric inpatients both in developed and developing countries, using various methodology and criteria. According to national and international guidelines, all inpatients should be screened for risk of malnutrition on admission using a validated screening tool. However, because of the lack of universally accepted definition for malnutrition, there is no consensus on the measures and methods to use for nutritional screening. There is controversy concerning the validity, reliability and practicality of existing paediatric nutrition screening tools. Moreover, current paediatric screening tools have not been designed and validated for infants. The study aimed to: 1) Validate a novel malnutrition screening scheme for infants - the Infant Paediatric Yorkhill Malnutrition Score (iPYMS) and compare its utility in different hospital settings, in UK and Middle East, Iran. 2) Compare the usefulness of various anthropometric measures to predict malnutrition in infants. 3) Determine the factors that correlated with malnutrition in hospitalised infants. 4) Explore the use of body composition measures in sick infants. The Paediatric Yorkhill Malnutrition Score (PYMS) had already been developed in Glasgow for use in children admitted to hospital. It utilized four elements that were reported as recognized predictors of the past, present or future nutrition risk. An audit was carried out at the beginning of the PhD course and aimed to evaluate the effect of PYMS on collection of anthropometric measurements in the wards. Findings suggested that introduction of a screening tool improved the acquisition of anthropometry by nursing staff, but their utilization by medical staff remained poor. Method: The Infant Paediatric Yorkhill Malnutrition Score (iPYMS) was developed by the research team. The score encompasses 4 rated steps that similar to those used for older children: weight < 2nd and 9th centile was used as opposed to BMI, and 3 elements concerning the history of nutritional issues. A score of 1 classifies a patient at medium risk and ≥2 or ≥3 indicates high risk. Infants were studied at admission to two tertiary children's hospital, 210 (0-12 months) in Glasgow, UK and 187 (1-12 months) in Tabriz, Iran. Convenience sampling was used to recruit equal number of patients in each risk group. Four researchers recruited the samples for the UK cohort and one for the Iran cohort. The diagnostic accuracy and validity of iPYMS in both cohorts were assessed by comparing the iPYMS nutritional risk with the Paediatric Subjective Global Nutritional Assessment (SGNA) that determine malnutrition risk and mean skinfolds z-scores (triceps and subscapular) below <-2SD as the benchmark for low fat stores and acute/chronic malnutrition. Discriminant validity was assessed using body composition and anthropometry measurements, with the hypothesis that infants at high risk of malnutrition will have lower fat and possibly lean mass compared with those at low risk. Results: More infants in Iran (32%) were rated as high risk for undernutrition than UK (7%). The diagnostic performance of iPYMS improved with the cut-off ≥ 3, more so in Iran than the UK. In Iran, only, infants who were classified as being at high risk of malnutrition had longer hospital stay. Infants in the iPYMS moderate and high risk groups all had significantly lower mean SD-score for anthropometry. After excluding patients scored high risk based only on low weight z-score (≤-2 SD), the differences in weight and BMI z-scores remained significant. In Iran 76% infants with raised iPYMS had mean skinfolds <-2SD, but only 5% in the UK. The UK infants may thus not actually be malnourished. They may be ill and just at risk of malnutrition. The first step of iPYMS (weight below <9th or 2nd centile) was a strong predictor of malnutrition risk, more so in Iran; in the Iranian cohort, 91% and in the UK 70% of infants above the high risk threshold of ≥ 3 scored as high risk due to the weight below <9th or 2nd centile. ROC Analysis either with SGNA or sum skinfolds z-score as the main outcomes illustrated that admission weight and growth velocity had almost the same predictive value in predicting malnutrition risk. This suggests that weight velocity is no improvement on weight alone as a predictor of malnutrition. Current breast feeding was found to be an independent predictor of malnutrition in Iran. Socioeconomic factors were weak predictors of malnutrition in this population. There is a lack of validated and suitable methods to assess body composition in infants. To determine whether analysing bio-electrical impedance data is practical in our young age range population, this was compared to skinfolds thicknesses and how the two measures of body composition varied relative to SGNA. The WHO standard for skinfolds only starts at 3 months, excluding nearly one third of infants in the Iran cohort and half in the UK. An iPYMS skinfold reference was thus generated using the iPYMS dataset for the UK cohort, as this was a population with low rates of malnutrition risk who had skinfolds levels mainly within the WHO range beyond age 3 months. In Iran, most high SGNA risk infants (72%) had low skinfolds, but in UK there was no association. Iranian infants had much lower mean lean and fat than the UK infants. Fat measured by BIA varied by SGNA rating risk group with both cohorts, but lean differed between risk groups only for Iran cohort. Conclusion: Malnutrition was common in this tertiary children's hospital in Iran. iPYMS might perform well in this setting and could be used by health professionals to identify infants with malnutrition. In contrast, in the UK, iPYMS would mainly identify infants at risk of malnutrition, because of the low prevalence of under-nutrition. On the other hand, we found that weight alone (the first component of iPYMS) is a robust predictor of malnutrition risk. Therefore iPYMS may not add any advantage over the simple measurement of weight alone to identify infants at risk of malnutrition. This is essential where there are limited resources. Studies should be continued to explore a suitable and appropriate gold standard to test the validity of the tools particularly in low prevalence settings as well as the resources and cost of the introducing the tool in clinical practice. Any screening tool for malnutrition can only be considered effective if it results in early intervention and improved clinical outcomes, so the effectiveness of iPYMS needs to be explored in future intervention studies.