INTRODUCTION

The prevalence of overweight and obesity has increased worldwide in children and adults. Being overweight and obese in childhood is strongly associated with an increased risk of metabolic syndrome in adulthood^1-6. Interestingly, despite the excess fat mass, not all obese children and adolescents present metabolic disorders. Indeed, not all normal-weight individuals are metabolically healthy. Consequently, there are two phenotypes namely "metabolically healthy phenotype (MHP)" and "metabolically unhealthy phenotype (MUP)," which are differentiated by the presence or absence of metabolic disturbances. and that in obese patients they generate more intrigue. The pathogenesis of the MUP is not fully understood. and is characterized by hyperglycemia, hypertension, low high-density lipoprotein cholesterol (HDL) levels, or hypertriglyceridemia and can be accompanied by insulin resistance or low-grade inflammation^7,8. MHP subjects commonly has a reduced inflammatory profile compared to MUP subjects and shows higher adiponectin levels than their metabolically unhealthy counterparts⁸. These findings suggest that lower levels of inflammation could play an essential role in the protective profile of the MHP⁹. In adults, metabolic syndrome has been associated with changes in several hematological parameters, such as red blood cells, platelets, and leucocytes¹⁰. The precise mechanism of hematological disturbances remains unclear but may be related to insulin resistance, changes in cytokine concentration, and chronic low-grade inflammation^5,11,12. The adipose tissue releases local and systemic bioactive molecules like adipokines (e.g., adiponectin and leptin), and proinflammatory cytokines, mainly tumor necrosis factor-alpha (TNF-α) and interleukin 6 (IL-6). Which are related to the cardiometabolic risk profile¹⁰. Interestingly, chronic low-grade systemic inflammation associated with obesity has been suggested to play a significant role in developing cardiometabolic complications^8,13,14.

In children, inflammatory markers are strongly associated with adiposity¹⁵. A strong, significant positive correlation has been described between the body mass index (BMI) and the total number of leukocytes in children, which suggests that the percentage of body fat influences the leukocyte count, especially in neutrophils⁷. Furthermore, obesity in children is related to increased IL-6 and TNF-α, which may contribute to increased circulating leukocytes⁷. An increase in red blood cell distribution width (RDW) positively correlates with BMI and is associated with a higher risk of metabolic syndrome, unfavorable lipid profile, and cardiovascular disease (CVD)^16-19. Therefore, hematologic parameters can identify patients at risk of cardiovascular diseases.

The clinical utility of assessing metabolic status in childhood lies in identifying those children with worse metabolic profiles and a higher risk of developing CVD^2,20. In recent years, an increased interest has been focused on biomarkers associated with obesity in children and adolescents. Currently, limited studies are comparing the distributions of hematological parameters between MHP and MUP in the pediatric population. Hematological parameters such as RDW, Neutrophil-to-Lymphocyte Ratio (NLR), Platelet-to-Lymphocyte Ratio (PLR), Lymphocyte-to-Monocyte Ratio (LMR), and RDW-to-Platelet Ratio (RPR) has been associated with various inflammatory conditions and can be used for the early detection of individuals at risk of CVD^12,14.

Hematologic parameters could constitute an accessible and cost-effective test for screening children and adolescents at risk of presenting the MUP. However, whether they can serve as biomarkers to distinguish MHP from MUP in pediatric populations is unknown. The current study investigated if hematological parameters can differentiate MHP from MUP in children and adolescents. The secondary objective was to explore the correlations of hematological parameters with inflammatory markers, insulin resistance, and adipokines, which could provide insights into the underlying mechanisms of these relationships.

METHODS

A cross-sectional study was conducted from May 2015 to December 2017 using a non-probabilistic sampling of consecutive cases in an outpatient pediatric endocrinology clinic at a secondary care hospital and public schools. The inclusion criteria were age from 5 to 18 years, with presumably good health, without prescription. Participants between 5 and 10 years old were classified into childhood and adolescence between 11 and 18 years²¹. Subjects with an infectious process a month before the study, secondary obesity, insulin-dependent diabetes, congenital metabolic disease, hormonal disorder, or inflammatory diseases were excluded. This study was in agreement with the Declaration of Helsinki and approved by the Ethics Committee of the Instituto Mexicano del Seguro Social (R-2013-785086) and Instituto Nacional de Perinatologia (212250-3310-11402011). Participants and their parents signed a free and informed an assent form and consent form.

Power analysis calculations were based on effect size and power using G*Power software. The power calculation for one-way ANOVA assumed a priori alpha error probability of 0.05, a power of 0.8, a medium effect size (F = 0.25), and the number of groups = 6. According to these assumptions, 216 patients would be required for this study.

RESULTS

Participants characteristics

Analysis of 292 children and adolescents included 98 with normal weight (33.56 %), 88 with overweight (30.14 %), and 106 with obesity (36.3 %). The participant's characteristics are summarized in Table 1. The subject's age mean was 11 years (SD: 2.61) and ranged from 6 to 18 years. Sex, nutritional status, and metabolic phenotypes were similar between children and adolescents. Approximately 52 % of the participants were women. Tanner's classification was unavailable in thirty-six participants (12.3 %). Compared with patients in whom Tanner was determined, these patients did not differ in age, sex, nutritional status, or prevalence of the MUP. The percentage of prepubertal patients was 42.2% in the childhood group and 8.28% in the adolescent group.

Comparison of Hematological parameters between MH phenotype and MU phenotype.

As shown in Figure 1, RDW values were significantly higher in children with MUO compared to children with MUNW (13.67 vs. 12.33; p = 0.01), and in MUO compared to MHO (13.67 vs. 13.15; p = 0.04). In adolescents, the PLR was higher for the MHNW group, with a mean value of 152.60 (SD 62.97) compared to 111.16 (SD 44.12) for the MHO group (Figure 1). The size effect was minimum (η²0.02 and 0.03, respectively). There were no differences between groups regarding NLR, LMR, and RPR values. In the analyses of all participants, there was no significant difference in hematological parameters between the six groups (Table 2).

Figure 1.  A. Level of Red blood cell Distribution Width (RDW) in children stratified by metabolic phenotype. B. Level of Platelet-to-Lymphocyte Ratio (PLR) in adolescents stratified by metabolic phenotype. 

Explanatory variables of unhealthy metabolic phenotype

Table 3 shows results from logistic regression analyses, with metabolic phenotype as the dependent variable. In the unadjusted model, obesity, PLR, and insulin resistance were significantly associated with the MUP. In the multivariate logistic regression analysis, obesity, PLR, and LMR were significantly associated with the MUP. The primary independent factor was obesity (OR 2.45, 95% CI 1.28 to 4.67, p < 0.01). A decrease in PLR was independently associated with a small increased risk of present MUP (OR 0.99; 95% CI, 0.98 to 0.99, p = 0.02). Similarly, a decrease in LMR was associated with a small increased risk of being MUP (OR 0.91; 95% CI 0.84 to 0.99, p = 0.04). Unadjusted and multiadjusted logistic regression analyses suggested no significant association between RDW, NLR, and RPR with MUP.

Since puberty may play an essential role in the risk to MUP, we performed stratified analysis by pubertal stage (Table 4). The multi-adjusted model showed that obesity was associated with an increased likelihood of having MUP (OR 2.53, 95% CI 1.26 to 5.09, p < 0.01). The associations between MUP and hematological parameters were not statistically significant (p for all ≥ 0.05) after adjusting for puberty.

Correlates of hematological indices and inflammatory biomarkers, cytokines, and insulin resistance

Among the hematologic indices and inflammatory biomarkers, NLR showed positive but low correlations with IL-6 (rho = 0.24, 95% CI: 0.12 to 0.34) and uCRP (rho= 0.23, 95% CI: 0.11 to 0.33). Similarly, the LMR showed negative and low correlations with the uCRP (rho = 0.22, 95% CI: -0.32 to -0.10). Positive and low correlations were found between NLR and leptin (rho = 0.26, 95% CI: 0.16 to 0.37) and HOMA-IR (rho = 0.22, 95% CI: 0.12 to 0.34). As shown in Table 5, correlation coefficients revealed no correlation between RDW, RPR, PLR, inflammatory biomarkers, adipokines, and HOMA-IR.

DISCUSSION

Although RDW was increased in MUO in childhood and PLR was reduced in MHO in adolescence, our results did not support the usefulness of hematologic parameters as tests for screening of MUP. After adjusting for age, nutritional status, and puberty, hematologic parameters were not associated with MUP. Therefore, hematologic indices are unreliable biomarkers of MUP in children and adolescents. The confounding effect of obesity and insulin resistance may partly explain this lack of association.

Barazzoni et al. suggested that the association between BMI and hematological parameters is mediated by their associations with insulin resistance markers and abdominal fat²⁵. In this study, Waist Circumference (WC) was not an independent variable for MUP. However, It is essential to acknowledge that WC may not always reflect visceral fat as it cannot differentiate between subcutaneous fat in the abdominal area and visceral fat accumulation²⁶.

Hematologic parameters in obese children are often disturbed. The high RDW and NLR values in obese subjects in this study are consistent with those reported by other researchers^5,11,18. Although RDW is seemingly increased in obese subjects, our results showed that this parameter does not explain the presence of MUP. The mechanisms linking RDW and obesity remain to be elucidated. Several mechanisms have been proposed to increase RDW in obesity, including inflammation, oxidative stress, or iron deficiency^18,27,28. Although it has been proposed that proinflammatory cytokines may produce an increase in RDW by inhibiting erythrocyte maturation¹⁵, our results do not support an association of RDW with three proinflammatory cytokines (IL-1b, IL-6, TNF-α).

The association between NLR and metabolic syndrome in adults presents conflicting results. Two studies showed that the NLR was significantly associated with presence and severity of metabolic syndrome^10,29. Opposite, Wang et al.¹⁹showed no significant differences in the counts of neutrophils between MHO and MUO phenotypes. Similarly, Bahadir et al. reported that the NLR was not associated with metabolic syndrome³⁰.

In children, a study performed in Turkey with 187 children and adolescents (aged 6 to 15 years) showed that NLR was significantly higher in obese adolescents compared with healthy controls⁵. In contrast, three studies showed that NLR did not differ significantly between overweight/obesity and normal weight children and adolescents^14,31,32. Our results revealed that NLR values were similar in subjects with MHP and MUP. Similarly, a study conducted in Italy with 552 obese children and adolescents has shown that the average NLR value did not differ significantly between the obese group with metabolic syndrome compared to the obese group without metabolic syndrome¹⁰.

PLR has been identified as a potential biomarker of inflammation, insulin resistance, the severity of obesity, and a predictor of metabolic syndrome (MetS)^33-35. In our study, PLR was significantly lower than the normal weight group. These findings are different and contradictory to previous studies. In adults, PLR was significantly higher in morbidly obese adult patients than normal-weight patients. In another study, PLR could not predict newly diagnosed MetS³³.In pediatric populations, there was no difference in PLR in obese subjects compared to healthy control groups^5,14. Additionally, the association between PLR and MUP was blurred after adjusting for the age group, puberty, HOMA-IR, and nutritional status in our study. Taking together, these findings indicate that PLR cannot be suggested as a biomarker of MUP in the pediatric population.

In our study, NLR levels correlated positively, although weak, with IL-6, leptin, CRP, and HOMA-IR. This finding might indicate that the increase in NLR in obese children could be an expression of low-grade inflammation and an insulin-resistant state¹⁹. A study in healthy young adults showed that the correlation between NLR, CRP, and fibrinogen was highly significant in obese individuals, reflecting the subclinical inflammation in this group¹³. In addition, Zaldivar et al.⁷reported increased neutrophils in children with obesity. Interestingly, there was a significant positive correlation between NLR and leptin, which can be related to the chemotaxis of neutrophils induced by leptin³. Adipose tissue neutrophils produce chemokines and cytokines, facilitating macrophage infiltration, which could contribute to the development of insulin resistance³.

An unexpected result of this study was the high prevalence of the MUNW phenotype (41.9 %). Our results show that risk factors for cardiovascular disease are present even in childhood and adolescence, regardless of the nutritional status of these subjects. Currently, there is limited data on the prevalence of MUNOW in pediatric populations. Previous studies reported that 14 to 22 % of normal-weight children and adolescents were metabolically unhealthy³⁶.

Damanhoury et al.²⁶reported that approximately 30% of children with overweight/obesity were classified as having the MUP. Marín-Echeverri et al.³⁷evaluated the association between abdominal obesity (AO) and cardiovascular risk factors in 232 children (2-5 years), of whom 50 % had AO, and 88 % of these were overweight or obese. The authors reported that 75.8 % of children with AO had ≥ 1 risk factors for metabolic syndrome. Shashaj et al.³⁸reported at least one metabolic abnormality in 86 (39.3 %) of children who manifested new-onset overweight or obesity, of which dyslipidemia (25.1 %) and hypertension (13.2 %) were the most prevalent alterations³⁸. Guzzetti et al.³⁹showed that dyslipidemia is frequent in children and adolescents with obesity, while hypertension is rare³⁹. These findings agree with our results.

Limitations in this study need to be noted. First, the cross-sectional design does not allow for establishing causal relationships between the variables studied, and the duration of obesity could not be determined. Second, a direct measurement of body fat distribution was not performed (for example, using dual-energy X-ray absorptiometry). Although confounding variables are considered in the study, there are other confounders (e.g., diet and exercise) that were not included, which could influence the results. Finally, the findings in this study could not be generalized to all pediatric patients since most of the participants included in this study were recruited from one reference hospital.

In conclusion, the study demonstrates that hematological parameters are not independently associated with the MUP, and they are unlikely to represent reliable biomarkers of MUP in the pediatric population.