Comprehensive pediatric reference intervals for 79 hematology markers in the CALIPER cohort of healthy children and adolescents using the Mindray BC‐6800Plus system

Hematological parameters vary significantly throughout growth and development due to physiological processes such as fetal‐to‐adult erythropoiesis and puberty. Pediatric age‐ and sex‐specific reference intervals (RIs) are thus essential for appropriate clinical decision‐making. The current study aimed to establish RIs for both common and novel hematology parameters on the Mindray BC‐6800Plus system.

K E Y W O R D S complete blood count, differential, pediatrics, reference intervals, test interpretation 1 | INTRODUCTION Assessment of erythrocytes, leukocytes, and platelet function is an essential component in the workup of many pediatric disease states. Modern hematology analyzers in pediatric clinical laboratories complete high-level cellular analysis to provide insight into hematopoiesis and hemostasis. 1 Hematological abnormalities are common in childhood and adolescence, and thus the complete blood count is one of the most highly ordered laboratory investigations in pediatrics. Common indications include clinical suspicion of anemia, bacterial, viral, or parasitic infection, malignancy, and/ or clotting dysfunction as well as screening prior to surgery or medical intervention. These objective laboratory data inform clinical decision-making significantly. It is thus essential that the interpretative recommendations provided to clinicians adequately reflect physiological covariates that influence the hematological system in pediatric patients.
Reference intervals (RIs) are health-associated benchmarks used to flag abnormal laboratory test results to prompt further follow-up and investigation. 2,3 Statistically defined as the 2.5th and 97.5th percentiles in a healthy reference population, RIs are challenging to establish in children and adolescents. 2 This is especially true for hematology parameters wherein known dynamic changes occur throughout physiological growth and development and require age-and sexspecific consideration. The neonatal and infantile periods are associated with significant hematological adaptations, including changes in α and β globin chain expression, growth factor production and responsiveness (e.g., erythropoietin, thrombopoietin), iron metabolism, and hematopoietic activity. 1 Development in cellular and humoral immune response occurs in early life to support protection against infectious pathogens. 4 Puberty is also associated with significant changes in sex hormone activity, including both estrogens and androgens, with direct impact on erythropoiesis. 5 Taken together, physiological adaptations integral to growth and development are expected to elicit quantitative changes in commonly measured hematological parameters across the pediatric age continuum. The use of adult-based RIs for test interpretation in infants, children, and adolescents is thus inappropriate and substantially increases the risk of diagnostic error, presenting a major limitation in healthcare delivery.
Few studies have established evidence-based pediatric RIs for hematology parameters. Available studies vary by population studied (e.g., ethnicity, health status, age range), parameters assessed, and analytical methodology. [6][7][8][9][10][11][12][13][14][15] The Canadian Laboratory Initiative on Pediatric Reference Intervals (CALIPER) derived health-associated RIs for hematology parameters in their cohort of healthy children on three modern hematology systems, including Sysmex XN-3000, Beckman DxH 520, and Beckman Dx900 platforms. [13][14][15] Other groups have established pediatric RIs on the Siemens Advia, Abbott Cell-Dyn, Beckman GenST, and Beckman HmX instruments. [6][7][8][9][10][11][12] Most studies have limited sample size in the first few years of life and do not comprehensively cover all hematological indices and emerging research parameters. In addition, lack of certified reference materials and traceable methods prompts investigation of assay-specific RIs for hematology parameters. Few studies have established continuous pediatric RIs using patient data in conjunction with indirect statistical procedures, [16][17][18][19] contributing to available pediatric hematology reference data. However, direct RI establishment remains the gold standard and is most clinically applicable to laboratories.
The current study aims to establish age-and sex-specific pediatric RIs for available hematology parameters on the Mindray BC-6800Plus in the CALIPER cohort of healthy children and adolescents (30 days to 18 years). These data will facilitate improved clinical decision-making in pediatric clinical laboratories with this instrumentation. Study exclusion criteria included: pregnancy, history of chronic illness, regular use of prescribed medication, and history of acute illness within 1 week of sample collection, as previously described. 2 Whole blood samples were collected through venipuncture in K 2 EDTA tubes (BD Vacutainer). Residual whole blood samples were also collected from apparently healthy and metabolically stable infants, children, and adolescents (30 days to 18 years) from outpatient clinics at The Hospital for Sick Children who were not admitted at the time of collection and did not demonstrate abnormal laboratory test results for clinically ordered tests. All residual specimens utilized for this study had preexisting CBC results that were reviewed for abnormal flags prior to study inclusion. Other biochemical data was also reviewed, when available. Patients from critical care departments as well as hematology and oncology, bone marrow transplant, or infectious disease clinics were excluded. Informed consent was not obtained for residual whole blood specimens as per our REB protocol. Self-reported ethnicity information for prospective cohort is reported in Table S1. Ethnicity information was not available for all participants and was defined by self-reported maternal and paternal ethnicity. Manufacturer maintenance, calibration, and quality control metrics were followed throughout study duration (Table S2).

| Reference interval establishment
All data analyses were completed in R Statistical Programming (Version 4.1.1). Data were plotted by age and color-coded by sex.
Upon visual inspection, age-and sex-partitions were determined and confirmed statistically using the method of Harris and Boyd. 3 Outliers were then removed using the Tukey method for normal distributions and the adjusted Tukey method for skewed distributions. 2,3 Normality was assessed quantitatively using the Shapiro Wilks test and qualitatively using quantile-quantile plots. RIs, defined as the 2.5th and 97.5th percentiles, were established as per Clinical Laboratory Standards Institute (CLSI) guidelines on Defining, Establishing, and Verifying Reference Intervals (EP28-A3c). 2,3 The robust method of Horne and Pesce was used for RI establishment when the sample size was greater than 40 but less than 120 individuals. 2,3 The nonparametric rank method was applied for partitions with a sample size of equal or greater than 120 individuals. 2,3

| Comparison to published reference values on other analytical systems
CALIPER previously established age-and sex-specific RIs for key hematology parameters in children and adolescents (30 days to 18 years) on three other analytical systems, including Beckman DxH500 (benchtop), 13 Beckman DxH920, 15 and Sysmex XN-3000. 14 To compare reference value distributions across analytical systems, data from the current study on Mindray assays were plotted against data from the Beckman DxH900 15 and Sysmex XN-3000 14 series for select parameters, including WBC count, RBC count, and platelet count.

| RESULTS
Of 79 hematology parameters assessed, 52 demonstrated statistically significant age-specific differences and required age partitioning. An  Table 1 (routine parameters) and Table S3 (researchuse-only parameters). Participant counts and age-specific reference value distributions for key parameters in the first year of life are provided in Figures S11-S15.

| Erythrocytes
Complex reference value distributions were observed for the majority of erythrocyte parameters ( Figure 1). RBC count, HGB, and HCT were  HDW only demonstrated a small but statistically significant sex difference prior to puberty (2 to <14 years). All pediatric reference values for NRBC count and percentage were equal to the assay limit of detection of zero 10 9 /L and 0%, respectively.  T A B L E 1 Age-and sex-specific reference intervals for main hematology parameters in the CALIPER cohort of healthy children and adolescents.

| Leukocytes
Three main reference value distribution trends were observed for leukocyte parameters (Figure 2): (1) decreasing concentrations with age (i.e., WBC, monocyte, and lymphocyte count), (2) increasing concentrations with age (i.e., neutrophil count), (3) consistent concentrations with age (i.e., basophil count). Reference standards were also established for the 6-part differential as a percentage, demonstrating an inverse relationship with age relative to count measurement (Supplemental Material). No sex-specific differences were observed.

| Platelets
Of the 10 platelet parameters assessed, 8 required partitioning by age and none required partitioning by sex (Figure 3). Platelet count (including optical and impedance determinations) and PCT decreased Reference interval could not be calculated, minimum and maximum displayed. NaN, confidence limit(s) could not be estimated.
throughout the pediatric age range. By contrast, MPV, IPF, and P-LCR increased in participants aged 5 to <19 years relative to early childhood. No significant differences were observed for PDW or P-LCC.

| Reticulocytes
Minimal age-specific differences were observed for reticulocyte parameters (Figure 4)

| Research-use-only
Of the 32 research-use-only hematology parameters assessed, 18 demonstrated statistically significant age-specific differences across the pediatric age range (Supplemental Material). Three demonstrated sex-specific differences (i.e., macrocyte count, RBC count caliperdatabase.org).
Reference value distributions for erythrocyte parameters mirrored previously observed trends. 9,12-15 Specifically, RBC count, HCT, and HGB increased rapidly from 30 days to <1 year followed by stable levels from 1 to <14 years. Statistically significant sex-specific differences were observed in the pubertal period wherein males demonstrated elevated levels relative to females, likely due to increasing metabolic demand and androgen-stimulated erythropoiesis in males. 5 Menstrual loss and lower iron stores in females may also contribute to this phenomenon. 20 Significant age-specific differences were also observed for WBC count and part of the differential. Total WBC counts in healthy children and adolescents were observed to decrease. The quantitative lower and upper limits ranged from 5.6-4.0 Â 10 9 /L and 12.0-9.2 Â 10 9 /L, respectively, across the pediatric age continuum.
Lymphocyte count also significantly decreased throughout the studied age range, requiring four age partitions. By contrast, neutrophil count was positively associated with age, increasing from 30 days to <19 years. These age-related trends have been previously observed and are suggested to be due to immune system immaturity and differential distribution of lymphocyte subsets from early life to adulthood. 9,10,[13][14][15] No sex-specific differences were observed in our pediatric cohort for leukocyte parameters. Previous studies have reported conflicting findings, with some demonstrating statistically significant elevations in neutrophil counts in adolescent females. 10,24 This may be resultant sex hormone modulation in adolescence; however, reported count differences are unlikely to be clinically significant. 10,25 Low basophil, eosinophil, and immature granulocyte counts and percentages were observed in the study cohort, as expected in a healthy population with low incidence of parasitic infection. 10 Platelet counts consistently decreased throughout the pediatric age range in concordance with published studies. 6 10,14,15 However, variation is observed in other study populations and analytical methodologies (e.g., 195-464 Â 10 9 /L 9 ). Genetic factors have been shown to contribute to platelet counts in healthy individuals at the population level and may explain observed discrepancies. 26,27 In addition, no sexspecific differences were observed in platelet counts across the pediatric age continuum. There is currently no consensus on sex-specific differences in platelet function. 6,10,[13][14][15] Reduced iron stores in females have been suggested to modulate platelet count increases relative to males, as platelet production can be stimulated by iron deficiency. 28 Most available studies are based on postmenopausal females and do not provide sufficient data in the adolescent period, warranting further investigation. 28  Reticulocyte parameters were also evaluated in the current study, expanding on previous reports. 10,14 Reticulocyte parameters are important in the assessment of erythropoietin activity, including composition, maturity, and development of cells released from the bone marrow. 30 Reference values for reticulocyte parameters did not demonstrate marked age-and sex-specific changes, requiring few partitions.
However, reticulocyte percentage and IRF were elevated from 30 days to <1 year. Given that reticulocytes are erythroid precursors, it is not surprising that higher prevalence is observed in the infantile period.
Available RIs studies have reported discrepant age-and sex-specific trends in reticulocyte counts in adolescence. [31][32][33][34] In the current study, no age-specific differences were observed beyond 4 years of age.
Other studies have reported higher values in females aged 12 to <19 years. 34 Observed differences may be due to biological and cohort differences between reference populations as well as variation in statistical methods applied to assess age-and sex-specific differences. Overall, data supports previously observed physiological patterns and suggests remarkable concordance between common hematological systems and methods. Common RIs across methods may be feasible based on minimal differences observed across distinct pediatric cohorts prospectively recruited as part of CALIPER.

| LIMITATIONS
The current study has two main limitations. First, our cohort is representative of the diverse multiethnic population in Toronto, Canada wherein the majority of participants were Caucasian (Table S1); however, we were not sufficiently powered to assess ethnic-specific Future work will focus on this special population.

| CONCLUSIONS
The current study established reference standards for 79 parameters on the BC-6800Plus system in a healthy cohort of Canadian children and adolescents. This is the first study to report health-associated pediatric concentration profiles for novel research parameters on the BC-6800Plus system. These data emphasize the complex biological patterns of hematology parameters in childhood, necessitating ageand sex-specific RIs for evidence-based clinical interpretation.

AUTHOR CONTRIBUTIONS
All authors made substantial contributions to the conception or design of the work or the acquisition, analysis, or interpretation of data for the work, contributed to drafting the work or revising it critically for important intellectual content; provided final approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.