Assessment of hematologic indices and their correlation to hemoglobin A1c among Bosnian children with type 1 diabetes mellitus and their healthy peers

Background Altered levels of many hematological parameters have been directly associated with diabetes in adults, while studies on children with type 1 diabetes mellitus are lacking. The aim of this study was to determine hematological indices in diabetic Bosnian children in comparison to healthy controls as well as to correlate their levels to blood glucose and hemoglobin A1c. Methods 100 healthy and 100 children with type 1 diabetes mellitus (age 1-18) were included in this study. Complete blood count, hemoglobin A1c, and glucose were tested. Results were analysed by IBM SPSS Statistics version 23. Results Significant differences (p<0.05) between healthy and diabetic children were found in relation to HbA1c, glucose, mean platelet volume, the number of white blood cells and erythrocytes, hematocrit, hemoglobin and MCH values. No gender differences or significant age differences were seen for hemoglobin, hematocrit, and MCV, while platelets, MPV, and MCH differed by age only in healthy children. When diabetic children were classified according to HbA1c levels, significant differences were seen for erythrocyte count and hematocrit value (p=0.013 and 0.019, respectively). The number of erythrocytes and white blood cells correlated significantly with HbA1c (p=0.037 and 0.027, respectively). Conclusions Lower levels of erythrocytes, hematocrit, and hemoglobin in diabetic compared to healthy children indicate possible development of anemia, while higher MCV, MCH, and MPV values indicate an alteration in erythrocyte morphology. Hematological indices could be a useful inexpensive tool in the diagnosis and follow up of type 1 diabetes in children.


Introduction
Type 1 diabetes mellitus (T1DM) is one of the most common endocrinological disorders in children and adolescents (1,2). It is caused by various environmental factors interacting with an underlying genetic predisposition resulting in autoimmune destruction of pancreatic beta cells (3,4). The onset of type 1 diabetes mellitus occurs early in life and children affected are at great risk of developing microvascular and macrovascular complications and cardiovascular disease later as a complication of diabetes. Hyperglycemia, dyslipidemia, inflammation, and stress are factors that increase the risk of vascular complications in patients with T1DM (2,8). Those long-term complications are the leading cause of premature mortality in this group of patients (5)(6)(7). T1DM complications that develop in children and adolescents depend on disease duration, metabolic control level, environmental and genetic factors.
T1DM and its complications represent a heavy financial burden on diabetic patients, their families, and society. This is especially true in developing countries. Hematological indices from complete blood counts may be used to overcome these challenges as good indicators and independent predictors for various microvascular and macrovascular complications related to endothelial dysfunction and inflammation. Altered level of many hematological parameters such as red blood cells (RBCs), white blood cells (WBC), and the platelet function and morphology have been observed and shown to be directly associated with diabetes (1-4, 12, 13, 15-18, 25-28) and metabolic syndrome in the adult population, indicating their close relationship with different components of metabolic syndrome, including insulin resistance (9)(10)(11)(12)(13). These studies demonstrated a significant difference in red blood cell distribution width as well as platelet volume between type 2 diabetic patients and controls (both increased in diabetic patients) (9)(10)(11). Differences also manifested at the level of total white blood cells, absolute lymphocyte count, absolute neutrophil count when compared to controls (all three increased significantly in diabetic patients) (10).
Biadgio et al. (10), found that among platelet indices, mean platelet volume, and platelet distribu-tion width were significantly increased in diabetic patients. In their study, some of the parameters of complete blood count (CBC) were found to be useful tools in following type 2 diabetes and its complications.
Milosevic et al. (11) reported independent associations of hematological parameters with insulin resistance and glycemic control. This association was also reported by Uko et al. (14) in type 1 diabetic patients, where changes in packed cell volume, platelet, and total white cell count as well as anemia, thrombocytosis and leukocytosis were reported.
Studies related to the issue of diagnostic significance of different hematological markers in children diabetic population are lacking worldwide due to unrecognized steep rise in the incidence of childhood diabetes affecting all countries in the world, including Bosnia and Herzegovina. Therefore, the aim of this study was to determine hematological indices (RBC count, hemoglobin, hematocrit, MCV, MCH, MCHC, RDW, PLT count, platelet morphology, platelet activation indicator MPV) in Bosnian children with T1DM and compare them to healthy controls of the same age as well as to correlate their levels to blood glucose, HbA1c levels, and anthropometric measurements. Of particular interest was to compare the sex and age differences in the examined parameters. To our knowledge, this is the first study of this kind in Bosnia and Herzegovina.

Patient population and study design
Two hundred children aged between 1-18 years were included in this study. The whole study group (N=200) consisted of 100 (50%) healthy children and 100 (50%) children with type 1 diabetes mellitus treated with insulin at the Department of Paediatrics, University Clinical Centre Sarajevo. The diabetic group was at the later phase divided into three groups according to HbA1c levels.
The control group was age-and sex-matched to the study group. The mean age of children enrolled throcyte count and hematocrit value (p=0.013 and 0.019, respectively). The number of erythrocytes and white blood cells correlated significantly with HbA1c (p=0.037 and 0.027, respectively). Conclusions: Lower levels of erythrocytes, hematocrit, and hemoglobin in diabetic compared to healthy children indicate possible development of anemia, while higher MCV, MCH, and MPV values indicate an alteration in erythrocyte morphology. Hematological indices could be a useful inexpensive tool in the diagnosis and follow up of type 1 diabetes in children.
Written informed consent was obtained from all participants recruited into this study (controls and diabetic patients). The study was approved by the Ethical Committee of the University Clinical Centre Sarajevo and was carried out in compliance with the Helsinki Declaration.
After obtaining detailed histories for all the children involved in the study, they were physically examined by a medical doctor (MD), and data were recorded on forms prepared for this study. Collected data included age (year), sex, height (cm), weight (kg), body mass index (BMI) (kg/m 2 ), family history of chronic diseases, and the laboratory findings. Patients with a history of accompanying chronic disease, current acute infection, alcohol abuse, and atherosclerotic disease or with severe hypoglycemia or diabetic ketoacidosis were excluded from the study.

Laboratory tests
The following laboratory tests were performed on healthy and children with T1DM: white blood cell count (10 12 /L), red blood cell count (10 12 /L), hemoglobin (g/L), hematocrit (%), MCV (fL), MCH (pg), MCHC (g/L), RDW (%CV), platelet count (10 9 /L), MPV (fL), HbA1c (%), glucose (mmol/L). Blood samples were taken from the brachial vein after overnight fasting and were processed within 2 hours after collection, according to the standard operating procedure of our institution. Glucose and HbA1c levels were acquired from blood samples that were centrifuged in gel tubes (3000 rpm for 5 minutes). Red blood cell and platelet morphology parameters were extracted from routinely performed complete blood count results. The blood counts were performed on Cell-Dyn Sapphire (Abbott Diagnostic, USA). Glucose levels were determined with Architect ci8200 (Abbott Diagnostics, USA), and HbA1c was measured using HPLC method on Adams analyzer (Arkray, USA).

Statistical analysis
Statistical evaluation of results was performed by IBM SPSS Statistics version 23 (Chicago, IL, USA). Median, interquartile range (IQR), ratio, and frequency values were used for the descriptive statistics of the data. Data distribution was tested using the Kolmo -gorov-Smirnov test. The parameters were compared within two groups using the Mann-Whitney U test. Spearman's rho correlation analysis was used to assess the relationships between the parameters. The Kruskal-Wallis test was used for the comparison of multiple groups. P-value <0.05 was considered statistically significant.

Results
There were no differences in age, BMI, total platelet count, MCV, MCHC, and RDW between healthy and children with T1DM ( Table I). The Mann-Whitney U test showed a significant difference between the healthy control group and children with T1DM in HbA1c, glucose, MPV, number of white blood cells, number of erythrocytes, hematocrit and hemoglobin values, with p<0.001 for all 7 parameters ( Table I). MCH also shows a significant difference between healthy and children with T1DM, with p=0.035 (Table I).
When healthy and diabetic populations were divided according to gender, statistically significant gender differences were observed only at the level erythrocyte count, hemoglobin, and MCV (p=0.007, 0.048, and 0.012) in the group of diabetic patients (Table II). Healthy controls showed no gender differences in any of the examined parameters (Table II).
In the analysis of the study populations according to their age, hemoglobin, hematocrit, and MCV showed statistically significant differences between compared age groups, both in diabetic children and their healthy peers (Table III). MCHC and RDW were significantly different within compared age groups in diabetic children (p=0.02 and p=0.007, respectively), while platelets, MPV, and MCH showed statistically significant age differences in healthy children (p=0.06, p=0.06, and p=0.01, respectively) ( Table  III). There were no other statistically significant age differences within the study groups except for those expected, such as age and BMI (Table III).
In the next phase, the effects of glycaemia control were tested in the population of diabetic children. After baseline evaluation, diabetic patients were divided into three groups according to their HbA1c levels: Group A consisted of patients with HbA1c levels <7%, group B consisted of patients with HbA1c levels between 7.1-8.4 and group C with HbA1c 8.5% (Table IV). The latest blood A1C targets for children and adolescents with type 1 diabetes, according to the American Diabetic Association 2019 criteria (15) is 7.5%. A lower goal (<7.0%) is reasonable if it can be achieved without excessive hypoglycemia.
All of the healthy children in the control group had HbA1c values <6% (Table IV). According to HbA1c values, 26% of children had an excellent level of disease control, 44% of children showed a some- Table II Gender differences in anthropometric, clinical and hematological characteristics of children with type 1 diabetes mellitus and healthy controls.
Values represent medians (lower-upper quartile), p* -significance between males and females in the group of healthy controls. p** -significance between males and females in children with type 1 diabetes. All differences were tested using the Mann-Whitney U test. BMI, body mass index; HbA1c, hemoglobin A1c; MPV, mean platelet volume; RBC, red blood cells; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; RDW, red blood cell distribution width; WBC, white blood cells.

Parameter
Healthy Controls   what worse level of disease control, while 30% of children with T1DM had essentially uncontrolled disease with HbA1c levels reaching 15% or more (Table IV).
No significant differences were observed at the level of age, BMI, WBC, hemoglobin, MCV, MCH, MCHC, RDW, PLT count, and MPV between compared groups ( Table V). Statistically significant differences between groups were observed only for erythrocyte count and hematocrit value (p=0.013 and 0.019, respectively), in addition to expected differences in glucose and HbA1c levels ( Table V).
Spearman correlation analysis (Table VI) showed significant positive correlations between BMI, HGB, HCT, MCV, RDW, and age in diabetic Bosnian children. MCHC showed a significant negative correlation with age. MPV showed significant negative correlations with platelet and leukocyte counts, and significant positive correlations with hematocrit and RDW. The number of erythrocytes, as well as white blood cells, showed a significant positive correlation with HbA1c. Other correlations of hematological indices with HbA1c and glucose in diabetic children Values represent medians (lower-upper quartile), p* -significance between age groups in healthy controls. p** -significance between age groups in children with type 1 diabetes. All differences were tested using the Kruskal-Wallis test. BMI, body mass index; HbA1c, hemoglobin A1c; MPV, mean platelet volume; RBC, red blood cells; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; RDW, red blood cell distribution width; WBC, white blood cells. Spearman correlation in healthy Bosnian children (Table VII) also showed significant positive correlations between BMI, HGB, HCT, MCV, and age.
MCH and erythrocytes showed positive correlations with age as well. PLT showed a significant negative correlation with age in healthy children. HbA1c showed significant negative correlations with MCH and MCHC, and a significant positive correlation with RDW. Other correlations of hematological indices with HbA1c and glucose, in healthy children, were not significant. The number of erythrocytes and white blood cells did not show a significant correlation with HbA1c.  HbA1c (%) 6.6 (6.1-6.9) 7.9 (7.4-8. WBC (10 12 /L) 5.6 (4.8-6.9) 5.9 (4.9-7.9) 6.9 (5.7-7.9) 0.067

Table VI
Spearman correlation for analysed parameters in children with type 1 diabetes mellitus (N=100).
All correlations were tested using Spearman test. Spearman r (rho), value of Spearman's rank correlation coefficient, where »-« represents negative correlation, »**« correlation is significant at the 0.01 level, »*« correlation is significant at the 0.05 level. p -significance of correlation between parameters.

Discussion
Altered levels of many hematological parameters (red blood cells, white blood cells, platelet function) and their corresponding indices (HGB, HCT, MCV, MPV) have been observed in patients with dia-betes. However, some authors did not report such differences except for higher WBC count in diabetics. In this study, hematological parameters were tested and compared in a group of children with diabetes and their healthy peers. A careful look at our results shows many changes in the diabetic population, one of them also being white blood cell count which was lower in children with type 1 diabetes mellitus, mean value 6.49×10 12 , compared to their healthy peers (mean 7.44×10 12 leukocytes, p<0.001).
Observed differences were neither sex nor age affected. Similar results related to the lower number of leukocytes in pathogenesis of type 1 diabetes were reported by Harsunen et al. (16). The discrepancy in reported leukocyte counts observed in different studies might be explained by their measurements in different stages of diabetes or from the fact that the data are related to different ethnic groups. Therefore, further investigations are required to explain these differences in WBC counts.
From all white blood cells, neutrophils are most involved in the pathogenesis of type 1 diabetes; thus, WBC count alters mainly due to alteration in neutrophil count. Early studies showed that type 1 diabetes patients had higher white blood cells, primarily neutrophil numbers than the controls (17), and increased neutrophil counts were reported to correlate with an augmented cardiovascular complication risk (17). However, more recent studies showed that circulating neutrophil number and white blood cells consequently were decreased in patients with type 1 diabetes and healthy autoantibody positive subjects, which might be associated with cell specific autoimmunity (18)(19)(20)(21). Lower Neutrophil count mediates a decrease in total white blood cell count.
Although our research did not include neutrophil counts, WBC data correlates with the second, more recent studies of Valle et al. (18), Harsunen et al. (19), Wang et al. (20), and Battaglia (21). We found significantly lower WBC values in children with T1DM, compared to healthy controls.
When other hematology data are concerned, descriptive statistics analysis in our study shows a lower number of erythrocytes, lower hematocrit, and hemoglobin levels in patients with type 1 diabetes mellitus when compared to their healthy peers. The Mann-Whitney U test shows a significant difference between healthy group and children with T1DM in number of erythrocytes, hematocrit and hemoglobin values, with p<0.001 for all three parameters analyzed. Interestingly, the number of erythrocytes and the level of hemoglobin were significantly lower in the diabetic female population, whereas MPV was significantly higher when compared to the male diabetic population. Gender differences were not apparent in the healthy population. Hemoglobin, hematocrit, and MCV values also differ by age in both diabetic and healthy children. Mean hemoglobin values in children with type 1 diabetes mellitus in our research were lower when compared to healthy peers, 137.2 vs. 143.4 g/L, respectively. This is in accordance with results reported by Khudhur et al. (22), Mbata et al. (23), and Kothari et al. (24), who reported lower mean values for RBC, HGB, and HCT in diabetics when compared to controls. All of the abovementioned parameters are indicators for the future development of anemia, a common finding in patients with diabetes, particularly in those with nephro pathy and renal impairment as well as retinopathy and macrovascular disease.
When other anemia associated hematological indexes were examined such as MCH and MCHC, differences at the level of different populations as well as gender differences were not evident. However, age differences in values of MCHC and RDW for diabetic, and MCH for healthy children, were evident.
One of the red blood cells indicators often investigated for its association with anemia in diabetic patients is red blood cell distribution with (RDW).
Mean RDW values in our study did not differ significantly between healthy and children with type 1 diabetes, 11.81 vs. 11.96 %, respectively, with p=0.088 determined by the Man-Whitney. Similar results were reported in numerous other studies (29,30). Contrary to these data, Biadgo et al. (10), in a similar comparative study on 296 participants, reported a significant difference in red blood cell distribution width between patients with diabetes and healthy controls. Differences across the study populations, as well as differences in study design and ethnicity, may account for the variability of RDW across studies.
Knowing that diabetes is considered to be a prothrombotic condition where platelets play an active role (33)(34)(35), it was of great interest to examine the differences in platelet size and number between diabetic and healthy populations. In this context, we were especially interested in MPV since this parameter represents a direct marker of platelet function and activity (9). Several studies have reported higher MPV values in individuals with both T1DM and T2DM, and these changes have been linked to metabolic control of the disease (31,32). The average MPV values for healthy Bosnian children were 6.32 fL, while their counterparts with type one diabetes had an average MPV value at 8.32 fL. This difference, as determined by the Mann-Whitney U test, was significant with p<0.001. Similar results as ours were reported in diabetic children by Venkatesh  Interestingly, Uko et al. (14) observed a significant positive correlation between high PLT and WBC count and raised blood sugar level among the diabetic subjects studied (r=0.52 and 0.45), respectively. Other studies showed positive correlations between WBC and HbA1c, such as a recent meta-analysis of 20 studies, which included 90,000 participants where a positive correlation between increased WBC level and diabetes risk was found (45).
In Sulochana et al. (46) study among the 100 type 2 diabetic patients with increased glycated hemoglobin, total leukocyte count was elevated in about 18 of 46 male and 22 of 54 female patients. In Bosnian children with type 1 diabetes, WBC count showed a significant negative correlation with HbA1c and glucose. In the study by Milosevic et al. (11), no such correlation was reported on patients with type 2 diabetes. Reported results are contradictory and require further attention.
The effects of glucose control and levels of HbA1c (indicators of disease control) on hematological parameters were investigated in this study in three groups of children.
We expected significant alterations in RDW and other RBC parameters reported in other studies (47). In our study, statistically significant differences between groups with different levels of disease control were observed only for erythrocyte count and hematocrit value (p=0.013 and 0.019, respectively), excluding expected differences in glucose and HbA1c levels. When assessing the correlation between HbA1c and hematological indices in diabetic children and their healthy peers, in Bosnian children, the number of erythrocytes, as well as white blood cells, showed a significant positive correlation with HbA1c and at the same time negative correlation with MCH and MCHC and positive for RDW in healthy children. Interestingly, some (49) authors reported a significant positive correlation between MPV and HbA1c (R=306, p<0.005), while others on non-diabetic subjects observed an inverse correlation between HbA1c and MCV (r = -0.22, p<0.05), MCH (r = -0.30, p<0.05), and MCHC (r = -0.32, p<0.05) (51). MCV, MCH, and MCHC showed no significant correlation with HbA1c in numerous studies reported (47,50). On the contrary, Jaman et al. (48) reported a statistically significant decrease of HCT in diabetic patients (HbA1c 7.5) when compared with non-diabetic patients (HbA1c 6.5). A decrease in the HCT value of diabetic patients may be due to the dehydration and accumulation of protein.
Ulutaş et al. (42) also reported a positive correlation between MPV level and HbA1c level, which is following our results. Hekimsoy et al. (43), have shown the opposite, while Ersoy et al. (44), reported similar MPV values in children diagnosed with T1DM to those in a control group (8.74 ± 0.96 and 8.49 ± 0.66, respectively). These different results in platelet count could be explained by their dependency on several variables, such as mean platelet survival, platelet production rate, and turnover rate in diabetes mellitus. Our study has some limitations related to the effects of the long-term course of the disease, and these limitations have to be overcome in future studies.

Conclusions
In this study, statistically significant differences in numerous hematological parameters of diabetic Bosnian children were observed when compared to healthy controls. The mean values of erythrocytes, hematocrit, and hemoglobin for diabetic children were found to be within reference ranges but significantly lower than the values of the control group, indicating possible development of anemia in the former group. At the same time, MCV, MCH, and MPV were higher in children with Type 1 diabetes mellitus indicating an alteration in erythrocyte morphology. Hemoglobin, hematocrit, MCV, MCHC, and RDW were age-dependent. Erythrocyte count, hemoglobin, and MCV showed significant gender differences. Age and gender had no effect on MPV, WBC, glucose, or HbA1c levels in children with type 1 diabetes mellitus. The mean value of white blood cells was also found to be lower in diabetic children, supporting a general involvement of the innate immune system in the pathogenesis of type 1 diabetes. Significant differences in mean values of erythrocyte and white blood cells were supported by observed correlations between HbA1c and: WBC and RBC. A significant positive correlation between WBC and HbA1c, as well as RBC and HbA1c, was observed.
Our findings suggest that hematological indices as accessible, simple, inexpensive laboratory parameters could be a useful tool in following up T1DM in children. MPV, in addition to HbA1c, can serve as a prognostic tool for determining the risk of diabetic microvascular and macrovascular complications.