Lipid profile and nutritional status of a pediatric population with sickle cell anemia: differences between gender and association with severity markers

Objective: To investigate the associations between nutritional status and lipid profile with biomarkers of hemolysis and inflammation in sickle cell anemia, in addition to considering gender differences. Methods: This cross-sectional study analyzed nutritional, and biochemical data of pediatric patients carrying sickle cell anemia. Results: Vitamin A, apolipoprotein B, total cholesterol, low-density lipoprotein, and non-high-density lipoprotein cholesterol were lower in boys. Hemoglobin was significantly lower, and the white blood cells and lactate dehydrogenase were higher in boys. Body mass index, vitamin A, and triglycerides were associated with hemoglobin levels, while apolipoprotein AI was associated with white blood cells and total bilirubin. Additionally, body mass index and vitamin A were associated with lactate dehydrogenase. Vitamin A showed significant predictive power in hemoglobin alteration and lactate dehydrogenase, while apolipoprotein A-I was able to predict high white blood cells values and total bilirubin. Research, Society and Development, v. 10, n. 10, e344101018934, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i10.18934 2 Conclusion: This study found in a pediatric population with sickle cell anemia that body mass index, vitamin A, triglycerides, and apolipoprotein A-I were associated with biomarkers of hemolysis and inflammation. Boys exhibited the greater nutritional deficit and severity of the disease.


Study populations
This is an observational descriptive cross-sectional study design (Hochman, Nahas, Oliveira Filho, & Ferreira, 2005), which involved demographic, anthropometric, and biochemical data of pediatric patients (aged 5-17 years) carrying stable SCA regardless of hydroxyurea (HU) treatment. It was conducted in an outpatient clinic of the Hemominas Foundation in Minas Gerais State (Brazil) between September 2015 and July 2019. All eligible patients were invited to participate. Exclusion criteria were pregnancy, acute illness, vaso-occlusive crisis (VOC) and blood transfusion in the three months prior to blood collection, as well as the presence of other chronic diseases not related to SCA.
This research was conducted according to the guidelines stated in the Declaration of Helsinki. The study protocol was approved by the Ethics Committees of the Federal University of Juiz de Fora (number 2.146.315) and the Hemominas Foundation (numbers 1.137.657 and 2.521.811).

Sample Size estimation
To obtain adequate statistical power (80%), being conservative and estimating to obtain a moderate correlation (rs around 0.30) between the severity hemolytic and inflammatory biomarkers (hemoglobin, WBC, LDH, and total bilirubin) with nutritional status markers (HA z-score, BMI z-score, vitamin A, vitamin D) and lipid profile associations, the sample size necessary for p<0.05 is 68 participants. Considering eventual losses, we estimated to evaluate at least 75 participants in our study to achieve an adequate statistical power (Cohen, 1988;Portney & Watkins, 2007).

Anthropometric Assessment
Weight and height were measured using an electronic scale (Líder®, Araçatuba, SP, Brazil) and an anthropometer (Altura Exata®, Belo Horizonte, MG, Brazil) with the child wearing light clothes and standing barefoot. Body mass index (BMI) was calculated as weight (kg)/height(m 2 ). Weight and height were transformed into z-scores for height/age (HA z-score) and BMI/age and sex (BMI z-score) and classified using WHO AnthroPlus software (WHO media products, Geneva, Switzerland) (World Health Organization, 2009). Children whose BMI z-score was <−2 SD standard deviation (SD) below the mean were classified as underweight (low); ≥ −1 to ≤ 1 SD as normal weight (acceptable); > 1 SD as overweight or obese (high). Their height was classified as stunted (low) if HA z-score was < −2 SD and normal height (acceptable) if HA z-score ≥ −1 SD. Values between −1EZ and −2EZ were considered at risk (borderline) for both anthropometric indexes.

Biochemical and Hematological Measurements
After overnight (8-12 h) fasting, blood was collected in tubes containing gel separator. Samples were immediately analyzed. Triglycerides, total cholesterol, and high-density lipoprotein cholesterol (HDL-C) were measured using enzymatic kits (Labtest ® , Lagoa Santa, MG, Brazil), according to the manufacturer's instructions. Non-high-density lipoprotein cholesterol (Non-HDL-C) was obtained by the difference between total cholesterol and HDL-C concentrations. Apolipoprotein A-I (ApoA-I) and apolipoprotein (ApoB) were measured by immunoturbidimetric assays accordingly (Abbott ® , Chicago, IL, USA). Total and direct bilirubin were directly measured by the Sims/Horn method using enzymatic kits (Labtest ® , Lagoa Santa, MG, Brazil). Total bilirubin was calculated as the difference between indirect and direct bilirubin. Lactate dehydrogenase (LDH) was measured by the enzymatic colorimetric method according to the manufacturer's instructions. All measurements were performed using the Cobas Mira plus equipment (Roche Diagnostic Systems,Inc., Montclair, NJ, USA).

Statistical analysis
Demographic, anthropometric, biochemical, and hematological data were recorded in a database using IBM SPSS Statistics (version 20; SPSS Inc., Chicago, IL, USA). Shapiro-Wilk statistics were used to test the normality of the distribution.
The demographic characteristics of categorical nature are depicted as frequencies and percentages. Furthermore, normally distributed data are presented as the mean and SD, and non-normally distributed data are presented as the median and first/third quartiles. Two-tailed Student's t-test, Mann Whitney test, and the Chi-squared test were performed to compare differences between the two groups. Partial correlation analysis controlling HU was employed to identify relationships between the severity biomarkers hemolytic and inflammatory (hemoglobin, WBC, LDH, and total bilirubin) with nutritional status markers (HA z-score, BMI z-score, vitamin A, vitamin D) and lipid profile. The variables related to the nutritional status and lipid profile that showed a significant correlation with the severity biomarkers (p ≤ 0.05) were selected to conduct multiple linear regression analysis. For the dependent variables that did not present normal distribution, logarithmic transformation was used. The categorical variable HU (use or not use), age, and gender were included in all regression analyses. Research, Society and Development, v. 10, n. 10, e344101018934, 2021 (CC BY 4.0) | ISSN 2525-3409 | DOI: http://dx.doi.org/10.33448/rsd-v10i10.18934

5
The independent variables that entered the multiple linear regression analysis were also included in the analysis of area under the ROC curve (Receiver Operator Curve) to investigate the accuracy in predicting critical values of interest of the severity biomarkers (hemoglobin, WBC, LDH, and total bilirubin). Cut-off values were set for the 20 th of hemoglobin levels (7.14 g/dL) and the 80 th for WBC (14000 mm 3 ), LDH (673 U/L), and total bilirubin (3.37 mg/dL) levels. The statistical significance level was set at P < 0.05 and 95% confidence interval.

Demographic characteristics and use of hydroxyurea
Of the total number of children with SCA invited to the study, 22 were excluded for the following reasons: refused participate (n = 5), transfusion therapy (n = 13), other exclusion criteria (n = 3), or death (n = 2). Seventy-five children were included in this study, in which 41 (54.7%) were boys. The average age was 11.10 years for boys and 11.82 years for girls.
Age and ethnicity were similar between the boys and girls. Eighteen (43.9%) boys and 21 (61.76%) girls were on HU therapy, without significant difference between groups (p = 0.123).

Nutritional status, lipid profile, and biomarkers of SCA severity
Our results showed a high prevalence of patients with a nutritional deficit (low) or nutritional risk (borderline). We found that 60% of children and adolescents in the study exhibited an adequate nutritional status according to BMI. Considering HA z-score and vitamin A status, less than 50% was considered adequate (acceptable), and only 10.2% had adequate 25-OHD levels ( Table 1).
Most of the patients had adequate levels of total cholesterol, LDL-C, non-HDL-C, and triglycerides. Nonetheless, HDL-C and ApoA-I levels were low in 64.4% and 79.5% of the patients, respectively (Table 1).
Our next step was to analyze the nutritional status and lipid profile, according to gender. The nutritional status markers exhibited no different between females and males, except for the vitamin A levels, which were lower in males. Dyslipidemia characterized by low HDL-C (< 40 mg/dL) was present in both sexes. Nonetheless, ApoB, total cholesterol, LDL-C, and non-HDL-C were higher in girls (Table 2). Severity biomarkers differed between the two gender groups. Hemoglobin was significantly lower, and WBC and LDH were higher in boys (Table 2).

Regression analyses with severity markers, lipid profile and nutrition status
The association between severity biomarkers and variables related to nutritional status and lipid profile in children with SCA was analyzed by multiple regression models. Considering multicollinearity, ApoA-I was used preferentially in relation to HDL-C. BMI z-score, and vitamin A were associated positively with hemoglobin and negatively with LDH, while ApoA-I was negatively associated with WBC and bilirubins. Triglyceride´s level showed a negative association with hemoglobin (Table 3). HA z-score, height-for-age, z-score; BMI z-score, body mass index, z-score; 25-OHD, vitamin D; ApoA-I, apolipoprotein A-I; ApoB, apolipoprotein B; HDL-C, high-density lipoprotein; LDL-C, low-density lipoprotein; Non-HDL-C, non-high-density lipoprotein cholesterol (total cholesterol -HDL-C); triglycerides/HDL-C, triglyceride-to-high-density lipoprotein ratio. Source: Authors.

Discussion
This study showed a high prevalence nutritional deficit or nutritional risk in children with SCA with positive association of BMI z-score and vitamin A with hemolytic marker, hemoglobin, and LDH. In addition, a lipid profile characterized by hypocholesterolemia and predictive power of ApoA-I levels with an inflammation and hemolytic marker, WBC and total bilirubin. HA z-score, height-for-age, z-score; BMI z-score, body mass index, z-score; 25-OHD, vitamin D; ApoA-I, apolipoprotein A-I; ApoB, apolipoprotein B; HDL-C, high-density lipoprotein; LDL-C, low-density lipoprotein; Non-HDL-C, non-high-density lipoprotein cholesterol (total cholesterol-HDL-C); y, years old; triglycerides/HDL-C, triglyceride-to-high-density lipoprotein ratio; WBC, white blood cells; LDH, lactate dehydrogenase. a analyzed by Mann-Whitney U-test and expressed as median (quartiles 1 and 3). b Analyzed by independent sample ttest and expressed as mean (standard deviation). * indicates statistical significance, p <0.05. Source: Authors.
Dyslipidemia has been considered one of the sub-phenotypes in SCD (Adegoke et al., 2016;Akinlade et al., 2014;Teixeira et al., 2019;Valente-Frossard et al., 2020). We observed that dyslipidemia was frequent in our patients (64.4%). The levels of ApoA-I, HDL-C, triglycerides, and triglycerides/HDL-C ratio did not differ between genders. However, ApoB, total cholesterol, LDL-C, and non-HDL-C levels were lower in boys. A study observed in the adult population with SCA, lower levels of total cholesterol, LDL-C, non-HDL-C, HDL-C, and ApoA-I in men than women (Lalanne-Mistrih et al., 2018), and hypocholesterolemia was also reflected in hemolysis and inflammation biomarkers. Boys exhibited worse levels of hemolysis and inflammation biomarkers than girls. This confirms the findings in other populations, in both steady-state and vasoocclusive crisis in SCA, with a proportional relationship between anemia severity and hypocholesterolemia (Akinlade et al., 2014;Ephraim et al., 2016;Teixeira et al., 2019). Variables that entered each analysis: Hemoglobin -BMI z-score, vitamin A, ApoA-I and triglycerides; LDH -BMI z-score, vitamin A; WBC -BMI z-score, ApoA-I, and triglycerides; total bilirubin -ApoA-I. Hydroxyurea, gender, and age were included in all the analyses, but only the hydroxyurea was present in the model. 1 Levels of total bilirubin were log-transformed. * Indicates statistical significance, p < 0.05. Source: Authors.
In our study, ApoA-I was negatively associated and a predictor of WBC and total bilirubin in the ROC curve. ApoA-I is the main protein in the HDL-C particle and responsible for stimulating reverse cholesterol transport, removing the excess from the tissues, and redirecting it to the liver. It may also inhibit the oxidation of LDL-C, assist in removing toxic products, and exhibit anti-inflammatory properties (Navab, Reddy, Van Lenten, & Fogelman, 2011).
Almost half of the individuals evaluated exhibited elevated or borderline triglycerides levels. Triglycerides levels were inversely correlated with hemoglobin level (Mokondjimobe et al., 2012;Valente-Frossard et al., 2020). Increased triglycerides in SCA subjects has previously been demonstrated (Teixeira et al., 2019). A study involving an SCD adult population identified a correlation between higher triglycerides levels and a lower hemoglobin level, higher levels of LDH, arginase, and inflammatory biomarkers, and vascular dysfunction (Zorca et al., 2010). The mechanisms explaining the elevated triglycerides in SCD are not fully understood. More recently, studies have indicated the involvement of angiopoietin-like proteins (ANGPTL) (Hassan, 2017;Zhang, 2016). ANGPTL3, ANGPTL4, and ANGPTL8 (Kersten, 2017;Li & Teng, 2014) inhibit lipoprotein lipase (LPL) activity that, in turn, reduces the triglycerides clearance. Sickle cell hemolysis is believed to generate byproducts that may stimulate the hepatic production of ANGPLT3 and ANGPTL4. Higher ANGPTL3 and ANGPTL4 were observed in adults with SCA compared to the control group (Vendrame, Olops, Saad, Costa, & Fertrin, 2019).
Children with SCA have a more significant nutritional deficit compared to control group (Jesus et al., 2018;Kazadi et al., 2017;Odetunde et al., 2016). Furthermore, differences in nutritional status between girls and boys have been reported, with a greater body mass deficit in boys (Esezobor et al., 2016;Odetunde et al., 2016). Our SCA population displayed lower levels of vitamin A in boys, similar to the result found in American children with SCA (Schall, Zemel, Kawchak, Ohene-Frempong, & Stallings, 2004). In SCA, the prevalence of suboptimal/deficient vitamin A status is high (Brownell et al., 2020;Dougherty et al., 2012). Vitamin A exerts a potent antioxidant effect (Baydas et al., 2002). Sickle cells are more susceptible to oxidative stress (Vona et al., 2021), which seems to be associated with the aseptic inflammatory profile in these individuals (Biswal et al., 2018). Therefore, there is greater demand for vitamin A for neutralization in SCD (Behera, Dixit, Bulliyya, & Kar, 2012).
Additionally, hepatic stellate cells are primarily responsible for the storage of vitamin A, and in case of liver damage, these cells are activated with rapid loss of vitamin A stock (Blaner et al., 2009). Individuals with SCD are exposed to chronic liver injury (Praharaj & Anand, 2021), which associated with more significant oxidative stress could partially explain the high frequency of low levels of vitamin A. Therefore, it is possible to understand the more pronounced profile of hemolysis and inflammation in boys, indicated by lower hemoglobin, leukocytes, and LDH, concomitant with lower levels of vitamin A.
Thus, an inadequate vitamin A status could aggravate the disease or be a consequence of it.
We also identified that BMI z-score and vitamin A levels were positively associated with hemoglobin and inversely with LDH. Similarly, vitamin A showed the predictive power of hemoglobin and LDH levels in SCA subjects. In other studies, lower levels of hemoglobin (Behera et al., 2012;Schall et al., 2004), as well as a higher frequency of hospitalization (Schall et al., 2004), percentage of falcization, and WBC (Behera et al., 2012), were observed in lower levels of vitamin A.
The high prevalence of 25-OHD deficiency observed in our population was also reported in studies with SCD children Oztas, Unal, Eskandari, Tamer, & Ozgunes, 2018) and appears to be influenced by ethnicity . The vitamin D intake by Brazilian children below the daily recommendation seems to be another important factor (Filgueiras, Suhett, Silva, Rocha, & Novaes, 2018). However, the present study does not provide this analysis. There was no difference in vitamin D levels between boys and girls. Although some studies indicate higher levels of vitamin D in boys (Alyahya, 2017;Lee, Licursi, & McMahon, 2015), our results are in agreement with , which investigated SCA children from 4 to 11 years of age. The vitamin D was not correlated with biomarkers of hemolysis and inflammation in our SCA patients. However, the benefit of vitamin D in the prevention of clinical abnormality in SCD is known. For every ten units of increase in levels of this vitamin, there was a 46% reduction in the probability of having acute pain in children and adolescents 3 to 20 years old with SCD (Lee et al., 2015).
This study is limited by its cross-sectional design, making it difficult to understand better the relationship between nutritional status and lipid profile and SCA severity. The size of the population of a single hematology care center also limits the generalization of results. However, our study showed an effort by including all patients with SCA treated at the study site, with a low percentage of rejection. In the analysis of vitamin D status, the season was not considered; however, blood collection occurred in all seasons.

Conclusion
Our findings present the relationship between nutritional status and lipid profile with severity biomarkers of SCA in a pediatric population. The observed results reinforce the need to consider nutritional assessment, including analysis of the lipid profile, in the follow-up of children with SCA. The study shows that lower levels of vitamin A; higher hemolytic and inflammatory patterns; and hypocholesterolemia, are more prominent in boys. Thus, gender is one of the confounding factors and should always be consider in the analysis of the associations of these markers.
Additional and prospective studies that could follow children until adulthood are suggested to verify any long-term reflections in clinical severity of SCA due to their nutritional and lipid profile.