[Phenotype proposal for early diagnosis of possibility of metabolic syndrome in school children aged 6 to 15 years]

Abstract

Introduction and objectives: Obesity and metabolic syndrome (MS) continue to be a problem at a socioeconomic level, causing high morbidity and mortality in the adult population. Prevention of risk factors should be carried out from an early age. Currently, there is no consensus on the opportune moment to start an intervention or treatment, regarding metabolic syndrome. The objective of the study is to describe the phenotype to predict early diagnosis of metabolic syndrome in schoolchildren.

Material and methods: Observational, prospective, cross-sectional and analytical study in schoolchildren from 6 to 15 years old, conducted in Guayaquil. Anthropometric measurements and a survey were performed, obtaining signing informed consent. The IBM Watson artificial intelligence (AI) platform with its software Modeler Flow, were used for the analysis.

Results: A population of 1025 students between 6 and 15 years old (mean of 12 years for men and 13 years for women) was examined, of whom 62.3% were men and 37.7% women. 23.9% of the population was overweight and 14% obese. A greater tendency to weight alteration was observed in men than in women (51.37% vs 47.79%), and a lower waist circumference in men (85 cm vs 87 cm, respectively). Males had a higher level of systolic blood pressure (SBP), being within the 90th percentile (mean SBP of 123 mmHg) 61.2%, compared to 38.8% of women, with a p < 0.001. Sedentary lifestyle is similar in both groups, with an average of 4.79 hours in front of the screen and/or video games. A statistically significant correlation was demonstrated between SBP and the waist/height ratio (WHtR) in the 90th percentile and 95th percentile (X² 9.075, p < 0.028, and X² 23,54, p < 0,000 respectively), as well as a relationship between 95th percentile and sex (X² 11.57, p < 0.001). The Modeler Flow software showed us that if WHtR, > 0.46, weight > 56.1 kg and height > 1.61 m, the probability of presenting metabolic syndrome, was of 82.4%. The statistic of this study has a predictive accuracy of 90% (error deviation of 0.009). The importance in the predictors of metabolic syndrome, range from 97.57% to 100%.

Conclusions: A prevalence of 33.9% of metabolic syndrome was observed in schoolchildren from 6 to 15 years old, with pathological cut-off points of: WHtR > 0.46, weight > 56.1 kg, pure sedentary lifestyle > 3 hours in front of the screen/playing video games, and SBP within the 90th percentile (> 123 mmHg). With these four indicators, we can predict a probability of early diagnosis of metabolic syndrome of 97% to 100%.

Keywords: Adolescentes; Adolescents; Artificial intelligence AI; Children; Inteligencia artificial (IA); Metabolic syndrome MS; Niños; Obesidad; Obesity; Overweight; Percentil P; Presión arterial sistólica PAS; Sedentarism; Sedentarismo; Sobrepeso; Síndrome metabólico SM; Waist to height ratio WHrt; Índice cintura/talla I c/t.