25,501
Views
166
CrossRef citations to date
0
Altmetric
Review Article

The Canadian laboratory initiative on pediatric reference intervals: A CALIPER white paper

, , &
Pages 358-413 | Received 26 Jul 2017, Accepted 12 Sep 2017, Published online: 11 Oct 2017

Abstract

Laboratory investigations provide physicians with objective data to aid in disease diagnosis, clinical decision making, and patient follow up. Clinical interpretation of laboratory test results relies heavily on the availability of appropriate population-based reference intervals (i.e. normative values) or decision limits developed through clinical outcome studies. Although reference intervals are fundamental to accurate laboratory test interpretation, and thus critically important to healthcare, the need for sound evidence-based reference intervals has been largely overlooked, particularly in the pediatric population. In the field of pediatric laboratory medicine, accurate age- and sex-specific reference intervals established using samples from healthy children and adolescents have not been readily available, forcing many clinical laboratories to report adult reference intervals with pediatric test results. When pediatric reference intervals are available, they have often been established with a small sample size, inpatient or outpatient samples, outdated methodologies, and/or inappropriate statistical procedures. To address these unacceptable limitations, several national and global initiatives have begun to close the critical evidence gaps in pediatric reference intervals. Notably, the Canadian Laboratory Initiative on Pediatric Reference Intervals (CALIPER) has made significant strides towards improving pediatric healthcare in Canada and globally. The present report is a white paper summarizing CALIPER, and provides a comprehensive compendium of the data generated through this project over the past decade as a single resource for clinical laboratory specialists, clinicians, and other healthcare workers. CALIPER launched an outreach campaign in 2008 to recruit healthy community children and adolescents, and developed a robust statistical algorithm, in accordance with the Clinical and Laboratory Standards Institute (CLSI) guidelines, to develop accurate age- and sex-specific pediatric reference intervals. The first CALIPER direct reference interval study was published in 2012, with age- and sex-specific reference intervals reported for 40 common biochemical markers. To date, CALIPER has collected health information and blood samples from over 9700 community children and adolescents, and has established a comprehensive database of age- and sex-specific reference intervals for over 100 biomarkers of pediatric disease. CALIPER has also performed a series of transference and verification studies to expand the applicability of the CALIPER database to five major analytical platforms, including Abbott, Beckman, Ortho, Roche, and Siemens. Through novel knowledge translation initiatives, the CALIPER Reference Interval Database has been made freely available online (www.caliperproject.ca) as well as on a mobile application (CALIPER Reference App), and it is used by clinical laboratories across Canada, the United States, and globally. In addition to establishing this comprehensive pediatric reference interval database, CALIPER has also performed a series of sub-studies, including examining how reference intervals are affected by pre-analytical factors (i.e. sample stability at specific storage conditions, fasting status and time of sample collection), biological variation (i.e. intraindividual and interindividual biological variation, reference change values), and ethnicity and pubertal development stage. In this white paper, extensive tables of pediatric reference intervals are provided for easy reference for clinical laboratories worldwide. All data reported have been published in over 20 peer reviewed publications and are also available through the CALIPER Reference Interval Database as well as the CALIPER Reference App for mobile devices.

View correction statement:
Correction

1. Pediatric reference intervals

1.1. Reference intervals as health-associated benchmarks

Reference intervals, sometimes referred to as “normative” or “expected” values, refer to the set of values observed in a reference sample group defined by a specific percentage, most commonly the central 95% () [Citation1]. Reference intervals are fundamental tools used by healthcare and laboratory professionals to interpret patient laboratory test results, ideally enabling differentiation of healthy and unhealthy individuals [Citation2]. However, it is important to note that because population-based reference intervals reflect the range of values expected in a typical community population, laboratory results that fall outside a reference interval do not necessarily indicate a disease, but rather that additional medical follow-up and/or treatment may be warranted. In contrast, decision limits or cutoffs, rather than reference intervals, associated with risk of specific adverse outcomes are used for some laboratory tests, including lipid parameters and bilirubin, to determine diagnosis or treatment. Nevertheless, reference intervals are used extensively in clinical practice to flag results outside the defined normal range and inform physicians if a patient’s result is normal or flagged as abnormal, potentially requiring follow-up testing. Given the critical importance to healthcare, it would be expected that accurate and comprehensive reference intervals are readily available to clinical laboratories and clinicians, but this is not always the case. In fact, the necessity of sound evidence-based reference intervals for accurate interpretation of laboratory test results has been largely overlooked [Citation3]. Instead, much of the effort in improving laboratory quality systems has focused on analytical assay performance. Unfortunately, the tremendous efforts to improve assay performance are wasted when inappropriate reference intervals are used to interpret laboratory test results. This problem is particularly apparent in the field of pediatric laboratory medicine where age-specific reference intervals based on healthy children and adolescents have not been readily available. This has forced many clinical laboratories to use adult reference intervals to interpret pediatric test results, a practice that can lead to erroneous and inaccurate test result interpretation. When pediatric reference intervals are available for clinical use, they are often severely limited, having been established using a small sample size, outpatient or hospitalized patient samples, outdated methodologies, or inappropriate statistical procedures. These limitations are mainly due to challenges in obtaining a sufficient number of healthy pediatric blood samples per partition (120 individuals as per the Clinical and Laboratory Standards Institute (CLSI) guidelines) necessary to establish robust reference intervals for all partitions. The number of partitions required for age- and sex-stratified reference intervals can be quite numerous in the rapidly growing and changing neonatal and pediatric population.

To close the critical evidence gaps in pediatric laboratory medicine, several national and global initiatives over the past decade have begun to establish pediatric reference intervals. Notably, the Canadian Laboratory Initiative on Pediatric Reference Intervals (CALIPER) was started by investigators at The Hospital for Sick Children and University of Toronto as part of the Canadian Society of Clinical Chemists (CSCC) Pediatric Focus Group. This collaborative project among seven pediatric hospitals across Canada (in St. John’s, Montreal, Toronto, Ottawa, Hamilton, Saskatoon, and Vancouver) has made significant strides towards improving pediatric healthcare in Canada and globally.

2. An overview of the CALIPER project

2.1. Pilot studies

To both establish preliminary reference intervals for select laboratory tests and help the CALIPER research team gain practical experience with sample collection, testing and statistical analysis, CALIPER performed several pilot studies. Initially, 15 immunoassays and 24 chemistries were analyzed using the Abbott ARCHITECT ci8200 platform (Abbott Instruments – Abbott Diagnostics, Abbott Park, IL), with results published in 2009 [Citation4]. Pediatric reference intervals were established for five age partitions: birth–12 months, 1–5 years, 6–10 years, 11–14 years, and 15–18 years [Citation4]. However, these age partitions were chosen arbitrarily rather than by using a statistical test, samples were collected from hospital outpatients rather than healthy community children and adolescents, and the recommended number of 120 subjects per group was not always achieved, particularly in the birth–12 month age partition [Citation4].

Subsequently, four additional pilot studies were performed. One study established reference intervals for 25 common analytes on the Ortho VITROS 5600 Integrated System (Ortho Instruments – Ortho Clinical Diagnostics, Rochester, NY) [Citation5] and another established reference intervals for 11 chemistries and 17 immunoassays using Roche cobas 6000 assays (Roche Instruments – Roche Diagnostics, Laval, QC, Canada) [Citation6]. For both these studies, reference intervals were stratified by sex and five arbitrarily-chosen age groups. Due to sample size limitations, the minimum CLSI requirement of 40 reference values per partition was often not met, necessitating the merging of some age partitions. Additionally, these studies were based on serum samples from pediatric outpatient cohorts. CALIPER then determined reference intervals for free thyroxine (FT4) and thyroid stimulating hormone (TSH) for two predetermined age groups: 1–6 years and 7–17 years. However, these reference intervals were not stratified by sex [Citation7]. The final CALIPER pilot study established age- and sex-specific reference intervals for five bone markers [Citation8]. The effects of body mass index, 25-hydroxyvitamin D, and parathyroid hormone (PTH) on the levels of these analytes were also investigated [Citation8].

These initial pilot studies were imperative for assessing the feasibility of the CALIPER project. Based on the success of the pilot studies, the next step for CALIPER was to initiate a major recruitment campaign to collect health information and blood samples from healthy children and adolescents. Directly sampling healthy subjects from the community is essential to determine accurate reference values from which robust reference intervals can be established.

2.2. Direct reference interval studies

Population-based reference intervals can be established using either an indirect or a direct sampling approach. While indirect sampling uses laboratory values obtained from a patient population to estimate reference intervals, direct sampling involves the selection of healthy individuals using defined inclusion and exclusion criteria [Citation1]. Indirect sampling is more commonly used primarily because it is more feasible and less costly. However, the CLSI does not recommend the use of indirect sampling due to the risk of including values from diseased individuals, which can lead to skewed, broader, and thus less sensitive reference intervals [Citation1]. Therefore, following its initial pilot phase, CALIPER began to establish a comprehensive, robust pediatric reference interval database using the direct sampling approach. CALIPER has been extremely successful in overcoming the challenge of obtaining a sufficient number of pediatric blood samples by focusing on recruitment campaigns and blood collection clinics held in the community. To date, CALIPER has created a biobank of healthy pediatric samples from over 9700 community children and adolescents. Using this biobank, CALIPER has now established age- and sex-specific reference intervals for over 100 biomarkers including: common biochemical markers, proteins, lipids, and enzymes [Citation9], endocrine and other chemistry markers [Citation10,Citation11], fertility hormones [Citation12], steroid hormones [Citation13], cancer biomarkers [Citation14], vitamins [Citation15], metabolic disease biomarkers [Citation16], testosterone indices [Citation17], and other biochemical markers [Citation17]. A summary of the findings from each of these studies will be discussed here. The specific methodology used by CALIPER to directly establish pediatric reference intervals is explained in Section 3.

2.2.1. Common biochemical markers

The first comprehensive direct reference interval study by CALIPER was published in 2012 and established age- and sex-specific reference intervals for 40 biochemical markers including common chemistry markers, enzymes, lipids and lipoproteins, and proteins [Citation9]. Serum samples from a total of 1072 males and 1116 females (aged 0 to <19 years) were analyzed using Abbott ARCHITECT c8000 assays. This study highlighted the unique age- and sex-related trends observed for each analyte throughout childhood and adolescence. Overall, age-related changes were found to be more common than sex-associated differences, and only one analyte, lipase, required no age or sex partitions. For creatinine, both the Jaffe and enzymatic methods were used, with the Jaffe method yielding higher results, as expected, because the Jaffe assay is known to be more susceptible to interfering substances. Albumin was also measured using two methods (bromcresol green and bromcresol purple); the values obtained using the bromcresol green method were found to be slightly higher than bromcresol purple. Additional interesting findings included elevated values of several biochemical markers in the neonatal period that rapidly declined after 2 weeks of life. However, some markers were found to have the opposite trend (i.e. low values in the neonatal period that subsequently increased), while others were found to have a wider reference interval during the neonatal period. Overall, these findings highlight the profound importance of establishing accurate reference intervals specific to the neonatal period. Furthermore, as the CALIPER cohort is composed of a multiethnic population, a preliminary investigation into differences in biochemical markers of major Canadian ethnic groups was performed. Ethnic differences were demonstrated in only seven of the 40 analytes including alanine aminotransferase, amylase, immunoglobulin G, immunoglobulin M, magnesium, total protein, and transferrin. Overall, this foundational study established age- and sex-specific reference intervals based on a large number of healthy pediatric subjects of various ethnic backgrounds.

2.2.2. Biochemical and endocrine markers measured by immunoassay

In 2013, CALIPER published reference intervals for an additional 14 analytes, including endocrine and biochemical markers measured by immunoassay [Citation10]. Serum samples from 736 males and 746 females aged 0–<19 years were analyzed using Abbott ARCHITECT i2000 assays. All 14 analytes required age and/or sex partitioning, with several analytes, including α-fetoprotein (AFP), cortisol, ferritin, FT4, troponin I (TnI), and 25-hydroxyvitamin D, requiring additional age partitioning within the first year of life. Of these, AFP, TnI, and ferritin had high concentrations in the neonatal period, decreasing shortly after birth. Several other analytes, including thyroid hormones, had high variance at birth that was reduced by 1 year of age. The majority of analytes had wider reference intervals both during the neonatal period and adolescence, reflecting the influence of growth and development on analyte concentrations.

Although transference studies can be performed to make CALIPER reference intervals applicable to other commonly used analytical platforms (discussed further in Section 4), transference is contingent on a high correlation of assay results between the two analytical systems in question. As there is substantial platform-specific variation for immunoassays, CALIPER is continuing to perform direct reference interval studies for immunoassays on all major analytical platforms. CALIPER established age- and sex-specific reference intervals for 29 endocrine and biochemical markers using the Beckman Coulter DxI 800 Immunoassay System and the Access 2 Immunoassay System (Beckman Coulter instruments – Beckman Coulter Canada Inc, Mississauga, ON, Canada), which were published in 2016 [Citation11]. Overall, similar dynamic patterns in concentration observed using Abbott assays were observed with Beckman assays. One notable difference between the studies was the reduced number of age partitions in the Beckman study, likely due to fewer total participant samples being used [(n = 711 (11) versus n = 1234–1482 [Citation10,Citation12]] although exact sample sizes for each assay vary for a number of reasons, including the number of partitions required and the volume of the sample being used. Additionally, although concentration patterns were largely comparable between Beckman and Abbott assays, differences were observed in concentration ranges for a subset of analytes. This was particularly evident for progesterone, follicle-stimulating hormone (FSH), luteinizing hormone (LH) and PTH. More recently, CALIPER established age- and sex-specific reference intervals for 29 immunoassays using Ortho VITROS 5600 assays [Citation18]. Although the age and sex trends in analyte concentration were similar to those observed using Abbott and Beckman assays, some analyte concentrations were higher or lower overall. This was particularly evident in the neonatal population, with higher concentrations of estradiol, progesterone, testosterone, and TSH obtained when measured on Ortho compared with Abbott and Beckman assays. These observations further highlight the importance of establishing assay-specific reference intervals for immunoassays.

2.2.3. Fertility hormones

In 2013, CALIPER established age-, sex-, and Tanner stage-specific reference intervals for seven fertility hormones [Citation12]. In this study, serum samples from 1234 healthy children and adolescents aged 0–<19 years were analyzed using Abbott ARCHITECT i2000 assays. All fertility hormones demonstrated a complex pattern of change in analyte concentration with age and between sexes, requiring some degree of age and sex partitioning, except for prolactin, which required no sex partitioning. Higher concentrations of estradiol, progesterone, FSH, LH, testosterone, and prolactin were observed in the neonatal and infantile period, requiring age-specific references intervals within the first year of life. Most of the hormones remained stable throughout childhood, but marked increases observed in both sexes during adolescence necessitated multiple age partitions. Estradiol, FSH, LH, and progesterone increased to a greater extent in female adolescents, while testosterone increased to a greater extent in male adolescents. Sex-hormone binding globulin (SHBG), on the other hand, decreased after puberty. Tanner stage-specific reference intervals were also established for these hormones. Each Tanner stage was comprised of a wide range of ages for both males and females. All seven hormones demonstrated similar Tanner stage reference intervals to those determined by sex and age partitioning. Estradiol, FSH, LH, and progesterone markedly increased between Tanner stages I and V in females, but to a lesser extent in males. The opposite was observed for testosterone, with markedly increased concentrations across Tanner stages in males, but to a lesser extent in females. Overall, this study highlighted the complex changes in fertility hormones, particularly throughout adolescence, and underscores the importance of reference intervals specific for the pediatric population.

2.2.4. Steroid hormones

Using a total of 337 serum samples from healthy children and adolescents, CALIPER established age- and sex-specific reference intervals for a panel of eight steroid hormones [Citation13] measured using liquid chromatography tandem mass spectrometry (LC-MS/MS). The effect of diurnal variation on cortisol concentration was examined further; cortisol levels were the highest between 7 and 10 am, then declined to minimal levels and spiked again between 12 and 3 pm, although this peak did not reach the maximum levels seen in the early morning. All subjects aged 0–<19 years were included in the diurnal variation analysis of cortisol; however, its concentration was elevated and more variable in the newborn period, necessitating further studies of diurnal variation for this specific age group. All steroid hormones exhibited dynamic age- and sex-related trends. Elevated and variable levels were noted for all steroid hormones except progesterone. Furthermore, most steroid hormones increased at the onset of puberty, with androstenedione, testosterone, and progesterone exhibiting sex differences. Androstenedione and testosterone were higher in males, while progesterone was higher in females. Overall, this study reported pediatric reference intervals for a newly developed LC-MS/MS method that allows simultaneous measurement of eight steroid hormones using a small volume of serum.

2.2.5. Cancer markers

CALIPER established age- and sex-specific reference intervals for 11 cancer biomarkers using 400-700 serum samples (depending on the analyte in question) from healthy children in 2014 [Citation14]. These samples were measured using Abbott ARCHITECT ci4100 assays. Significant age-related trends were noted for all biomarkers, with the exception of anti-thyroglobulin antibodies. Most cancer markers were elevated in the first week of life, followed by a rapid decrease in concentration that remained relatively stable throughout childhood and adolescence. The exceptions to this age-related trend were the cancer antigens (CA125 and CA15–3) that showed slightly lower concentrations in the first week of life and relatively stable concentrations thereafter. Lastly, both free and total prostate-specific antigen concentrations were very low and stable throughout the pediatric age range, with elevated levels in males during adolescence. Overall, this study highlighted the profound differences observed between pediatric and adult reference intervals for cancer biomarkers. Pediatric cancer biomarker reference intervals are useful not only in clinical practice but also in related research.

2.2.6. Vitamins

CALIPER established age- and sex-specific reference intervals for vitamins A and E, as well as ratios of vitamin E to cholesterol (E:C ratio) and triglycerides (E:T ratio) in 2014 [Citation15]. Serum samples from 307 and 330 healthy children and adolescents were collected for vitamin A and vitamin E, respectively, and were measured using high performance liquid chromatography (HPLC) with a C18 reverse-phase column. No sex partitioning was required for either analyte. Vitamin A had lower concentrations during the first year of life, then gradually increased with age. Vitamin E required a wide reference interval at birth, which subsequently narrowed and remained stable throughout childhood and adolescence. Both E:C and E:T ratios required no age or sex partitioning, with one reference interval spanning the 1–<19 year age range. Overall, these reference intervals will be useful to monitor nutritional status in the pediatric population and highlight the importance of age-specific reference intervals.

2.2.7. Metabolic disease biomarkers

In 2016, 500 CALIPER samples were used to establish age- and sex-specific reference intervals for 37 amino acids using ultra-performance liquid chromatography (UPLC), 32 acylcarnitines, as well as free and total carnitine using tandem mass spectrometry, and β-hydroxybutyrate and free fatty acids using Ortho VITROS 5.1 chemistry analyzer assays [Citation16]. Over 50% of amino acids and over 70% of acylcarnitines required partitioning of reference intervals during the neonatal period. Additionally, 21% of all analytes required partitioning throughout adolescence, and half of these also required sex partitioning. This publication was the first report of sex-specific reference intervals for branched-chain amino acids (valine, leucine, and isoleucine) as well as free and total carnitine for the adolescent population. All three of the branched-chain amino acids required sex-specific reference intervals, with higher levels in males than females in all cases. In addition, free and total carnitine required sex-specific reference intervals in adolescence, with higher levels in females compared to males. Overall, this report will contribute to improved detection and monitoring of several metabolic diseases in the pediatric population.

2.2.8. Testosterone indices

CALIPER established reference intervals for calculated testosterone indices including free testosterone (FT), bioavailable testosterone (BAT), and the free androgen index (FAI) based on a cohort of 471 healthy community children and adolescents in 2015 [Citation17]. Serum samples were measured for total testosterone, SHBG, and albumin using Abbott ARCHITECT ci4100 assays. These parameters were then used to calculate the testosterone indices. FT and BAT were calculated using the Vermeulen equation [Citation19], and FAI was calculated as the percent ratio of total testosterone to SHBG. All calculated testosterone indices exhibited the same age- and sex-related trends throughout the pediatric age range. Levels were initially elevated in males after birth but then declined, with males and females both having low values throughout childhood. At the onset of puberty, all testosterone indices increased to a significantly greater extent in males compared to females.

2.2.9. Other biochemical markers

A total of 1917 healthy CALIPER samples were used to measure serum concentrations of an additional 14 biochemical markers using Abbott ARCHITECT ci4100 assays in 2015 [Citation20]. Again, age- and sex-specific pediatric reference intervals were subsequently established. The majority of analytes exhibited dynamic changes in concentration throughout the pediatric age range and required at least three age partitions. Exceptions to this included cholinesterase, cholinesterase-dibucaine number, and immunoglobulin E which required only two age partitions, and α-1-antitrypsin which required only one age partition. Additionally, eight of the 14 markers, including C-peptide, ceruloplasmin, and insulin, required age partitions within the first year of life. Furthermore, levels of anti-cyclic citrullinated peptide (anti-CCP) and anti-thyroperoxidase (anti-TPO) were below the detection limit of the assay for the entire pediatric age range. These findings further emphasize the need for accurate pediatric reference intervals.

2.3. Knowledge translation strategies

The rigorous process undertaken to directly sample pediatric subjects, perform accurate laboratory analysis, and statistically analyze data to establish age- and sex-specific reference intervals has resulted in a CALIPER database of unprecedented richness, underscoring the importance of research on circulating biomarkers. To ensure that this database of invaluable information is disseminated to healthcare and laboratory professionals, CALIPER has begun several knowledge translation (KT) initiatives. First, all reference intervals established using the CALIPER cohort have been reported in peer-reviewed publications, and have been presented at regional and international conferences and workshops. CALIPER has additionally used innovative KT tools including the CALIPER web database (www.caliperproject.ca) and a novel mobile application of the database. These technological tools allow physicians and laboratory personnel, as well as patients and parents, the ability to view pediatric reference intervals and interpret a child’s laboratory test results based on the CALIPER database. Following the implementation of CALIPER KT strategies, there has been widespread adoption and implementation of CALIPER reference intervals by hospitals across Canada, the United States, and globally. CALIPER has been providing direct support to clinical laboratories interested in implementing CALIPER reference intervals by providing healthy pediatric samples to validate the CALIPER ranges on each local laboratory system and by offering advice and support on data analysis and implementation. A series of transference studies to transfer CALIPER reference intervals from Abbott assays (the instrument used initially to establish CALIPER reference intervals) to assays on Beckman Coulter, Ortho, Roche, and Siemens instruments (Siemens Instruments – Siemens Healthcare Diagnostics Inc, Newark, NJ) have also increased the wide-spread applicability of the CALIPER database to laboratories using any one of these major analytical platforms.

3. CALIPER methodology for determination of direct population-based reference intervals

CALIPER has employed consistent procedures for recruiting pediatric participants, collecting blood samples and health information, analyzing samples, and statistical analysis to determine covariate-specific reference intervals. In this section, the methodology used by CALIPER to establish direct population-based reference intervals will be discussed.

3.1. Participant recruitment and defining a reference population

A schematic of the relationship between a fit-for-purpose laboratory reference interval and the population it serves, as proposed by CLSI, is shown in . Importantly, it highlights that the first step of selecting appropriate reference individuals largely determines the quality of the resulting reference interval [Citation21]. Reference sample groups should be established by direct sampling of reference individuals from the age group(s) of interest using defined inclusion and exclusion criteria. Direct sampling of an appropriately selected reference population is the approach recommended by CLSI, and assessment and subsequent exclusion of any unhealthy individuals increases the validity and appropriateness of the reference interval [Citation1]. Reference individuals should be as similar as possible to the target patient in all aspects other than the disease or condition under investigation.

Figure 1. (A) Reference interval, defined as the central 95% of laboratory test results from a healthy, reference population. A representative normal distribution for analyte concentration is shown. (B) Schematic of establishing reference intervals using the direct method. Adopted from CLSI EP28-A3c guideline [Citation1].

Figure 1. (A) Reference interval, defined as the central 95% of laboratory test results from a healthy, reference population. A representative normal distribution for analyte concentration is shown. (B) Schematic of establishing reference intervals using the direct method. Adopted from CLSI EP28-A3c guideline [Citation1].

3.2. Sample acquisition

CALIPER recruits healthy children and adolescents from birth up to (but excluding) 19 years of age from the greater Toronto and Hamilton areas. CALIPER community clinics are held at schools, daycares, community centers, and camps to collect blood samples from healthy children and adolescents. In addition, CALIPER occasionally recruits participants through information booths at the Hospital for Sick Children and the CALIPER website (www.caliperproject.ca). Interested participants and their parents provide informed consent by completing written consent forms, which are verified by CALIPER staff and volunteers. Once consented, participants complete a health questionnaire about personal and family health history and demographic information. Anthropometric measurements, including height, weight, and waist circumference, are taken. Participants may also be asked to complete an optional self-report on Tanner stage, a measure of sexual maturity. Tanner staging, created by Marshall and Tanner, is the 5-stage standard clinical system used to describe normal pubertal development [Citation22,Citation23]. Blood collection amount is determined by age, with participants aged 0–<1 year donating 3.5 ml of blood, participants aged 1–<11 years donating 7 ml of blood and participants aged 11–<19 years donating 10.5 ml of blood. Blood samples are collected in serum separator tubes, and are centrifuged, separated, and aliquoted within 4 h of collection. Serum aliquots are kept frozen in the CALIPER biobank at −80 °C until testing. CALIPER clinics are held throughout the year on mornings, afternoons, or evenings. Therefore, while date and time of blood collection are recorded, they do vary across samples. Furthermore, fasting is not required, although the time since the last meal is recorded. As remuneration, participants are provided a small cash gift, two volunteer hours (a minimum number of volunteer hours are mandatory for completion of high school in Ontario), and a choice of a CALIPER t-shirt, teddy bear, or storybook.

3.3. Selecting a reference sample group

Upon successful recruitment, a reference sample group is selected to obtain reference values. CALIPER applies several exclusion criteria to subjects when selecting serum samples to analyze. Exclusion criteria include pregnancy, history of chronic illness or metabolic disease, acute illness within the week preceding donation, and use of prescription medication within the 2 weeks preceding donation. Following the application of exclusion criteria, the study population is randomly selected from the remaining healthy participants while ensuring a balanced age and sex distribution. Additionally, efforts are made to ensure that the multiethnic composition of the reference sample reflects that of the province of Ontario as reported by the 2006 Ontario census (Statistics Canada) [Citation24].

3.4. Sample analysis

Batch analysis of samples to obtain reference values is performed after selecting the reference sample group’s frozen aliquots and thawing at 5 °C. Analytical methods are verified according to the manufacturer’s instructions through preventive maintenance, function checks, calibration, and quality control. Calibration is performed after initial assay set-up on the analyzer and each time a reagent kit with a new lot number is used, as per manufacturer's instructions. Quality control procedures are performed for each assay prior to testing, for each day that sample testing is conducted. Quality control results are examined to ensure that the values are within the recommended range for each control level. All samples undergo automated interference analysis for hemolysis, icterus, and turbidity. The majority of direct reference interval studies used assays performed on Abbott ARCHITECT instrumentation, using the c8000 [Citation9], the i2000 [Citation10,Citation12], or the ci4100 [Citation14,Citation17,Citation20] model. Additional direct reference interval studies used LC-MS/MS [Citation13,Citation16], HPLC [Citation15], UPLC [Citation16], or Ortho VITROS 5.1 [Citation16], Ortho VITROS 5600 [Citation18] or Beckman Coulter DxI or Access 2 assays [Citation11].

3.5. Statistical analysis

CALIPER has published several reference interval studies (discussed in the preceding section) [Citation9–17,Citation20]. All direct reference interval studies used a robust statistical method based on CLSI guideline C28-A3 [Citation1] (). Once the reference values of the analyte of interest are determined through analysis of the reference sample group’s serum samples, scatterplots and boxplots are created to visually inspect the data. Extreme outliers are identified by visual inspection and subsequently removed. Visual inspection of scatterplots of age against analyte concentration is used to identify potential age and sex partitions. To determine if these partitions are statistically significant, the Harris and Boyd method is used. This method suggests partitioning if the ratio of the subgroup standard deviations is ≥1.5 [Citation25] and examines not only the differences in means, but also the differences in standard deviation between two subgroups. Statistically significant age and sex partitions are then reviewed with clinical experts and partitions may be collapsed if appropriate. Following confirmation or rejection of suspected partitions, data in each individual partition are transformed using the Box–Cox transformation method to obtain a Gaussian distribution. Following transformation, the normality of each partition is assessed using quantile–quantile (Q–Q) plots and the Shapiro–Wilk test. If data are normally distributed following transformation, outliers are removed using the Tukey test twice [Citation26]. However, if data remain skewed following transformation, outliers are removed using the adjusted Tukey test twice, which accounts for skewness of the data [Citation27]. Following outlier removal, reference intervals are calculated for partitions with a sample size ≥120 using the non-parametric rank method [Citation1], and corresponding 90% confidence intervals around the upper and lower limits are calculated using ranked observations. For partitions with a sample size >40 and <120, the robust method of Horn and Pesce is used for reference interval calculation [Citation28], and 90% confidence intervals of the limits are calculated using percentile bootstrap estimates.

Figure 2. Robust statistical algorithm used by CALIPER to establish pediatric reference intervals using the direct method in accordance with CLSI EP28-A3c guideline [Citation1].

Figure 2. Robust statistical algorithm used by CALIPER to establish pediatric reference intervals using the direct method in accordance with CLSI EP28-A3c guideline [Citation1].

4. CALIPER reference interval transference across analytical assays

When reference intervals are developed, they are specific to the method, instrument and population from which they were established. Rather than creating direct reference intervals for each laboratory, which would be extremely costly and labor-intensive, reference intervals can be transferred from one laboratory (or instrument) to another. The majority of CALIPER reference intervals originally established on Abbott assays have since been transferred to assays from other manufacturers including Beckman Coulter, Ortho, Roche, and Siemens [Citation29–32].

To transfer reference intervals, CALIPER uses an approach in accordance with CLSI guidelines C28-A3c and EP9-A2 [Citation1,Citation33] (). First, 200 patient samples are analyzed on the assay that was used to establish the original reference interval (donor) and the assay that the reference interval is being transferred to (receiver). Samples are selected to ensure that a broad age, sex, and concentration/activity range is covered for each analyte under study.

Figure 3. Statistical algorithm used by CALIPER to transfer and verify pediatric reference intervals in accordance with CLSI EP28-A3c and EP9-A2 guidelines [Citation1,Citation33].

Figure 3. Statistical algorithm used by CALIPER to transfer and verify pediatric reference intervals in accordance with CLSI EP28-A3c and EP9-A2 guidelines [Citation1,Citation33].

Statistical analysis of the analyte concentrations obtained from the donor and receiver assays are plotted against each other. Similar to the method used for reference interval establishment, gross outliers are identified and removed by visual inspection. The line of best fit is determined differently according to the R2 value. Simple linear regression using the least squares approach is used if the R2 ≥ 0.95, and Deming regression is used if 0.70 < R2 < 0.95. An R2 < 0.70 indicates that the correlation between the instruments is inadequate and the reference interval cannot be transferred.

The appropriateness of the linear model is assessed by examining Q–Q plots, standardized residual plots, and Bland–Altman plots. The Q–Q plots are used to assess the normality of the residuals, with points close to the line y = x representing a normal distribution. The standardized residual plot was used to assess the random distribution of residuals. If trends, patterns or clusters were observed, the reference interval was considered non-transferable. Then the Bland–Altman plot was visually inspected, with a random distribution of points, and absence of patterns, clustering or trends suggestive of no bias in the relationship between the methods. If all three graphical analyses demonstrate acceptable method transference, the reference interval is transferred using the line of best fit. Otherwise, transference is deemed inappropriate and a full reference interval study must be performed. The 95% confidence intervals around each reference limit are also determined. Lastly, the transferred reference intervals are verified using approximately 100 healthy CALIPER reference samples. The percentage of reference sample values that fall within the transferred reference intervals and confidence intervals are determined. The reference interval is considered fully verified, according to CLSI guidelines [Citation33], if at least 90% of reference sample values fall within the transferred reference interval. Individual clinical laboratories should verify CALIPER reference intervals on their local platform and population prior to implementing them in clinical practice, as recommended by CLSI [Citation1].

5. CALIPER sub-studies

In addition to establishing direct reference intervals and transferring these to assays on other instruments, CALIPER has performed several sub-studies, including pre-analytical factors, biological variation, and the influence of ethnicity and sexual development on reference values, to further examine areas of interest in pediatric laboratory medicine. Each CALIPER sub-study is summarized here and references to the detailed published studies are provided as resources.

5.1. Preanalytical factors

Preanalytical factors examined by CALIPER include sample stability at specific storage conditions, fasting status and time of sample collection. To examine the effect of storage time on analyte concentration, CALIPER determined the stability of 57 chemistry, protein, and hormone analytes stored at −80 °C on assays for three analytical instruments (Ortho VITROS Chemistry System, Roche cobas Integra 400 Plus, and Siemens Immulite 2500) [Citation34]. Samples from children and adolescents (aged 0–18 years) were analyzed at monthly intervals over 10–13 months and each aliquot was subjected to one freeze-thaw cycle prior to analysis. The majority of analytes examined in this study, with the exception of PTH, were stable when stored at −80 °C [Citation34], suggesting that with appropriate storage, most pediatric samples do not require immediate testing for reference interval determination.

Given the challenging nature of obtaining fasting samples from the pediatric population, particularly from neonates, CALIPER analyzed the change in concentration of 38 routine clinical chemistry analytes measured at fasting, postprandial, and random time points throughout the day [Citation35]. This study was performed on 27 healthy children and adolescents (aged 4–<19 years) by obtaining blood samples after an overnight fast, immediately after a mid-morning breakfast, 2 h after lunch, and in the late afternoon [Citation35]. Data analysis revealed significant differences in concentration for a number of analytes depending on fasting state and time of blood sample collection [Citation35]. These observations confirmed some (e.g. triglycerides), but not all (e.g. glucose), analytes known to vary between the fasting and postprandial state, but also identified some chemistry markers not previously known to significantly change with food intake (e.g. immunoglobulins). It is important for pediatricians to be aware of these physiological variations in concentration for specific analytes to ensure informed decisions are made when ordering fasting, random, postprandial, or serial sample testing for pediatric patients.

5.2. Biological variation

Understanding biological variation is important for both patient monitoring and establishing quality specifications. CALIPER examined short-term biological variation to further understand the physiological changes that occur within and between pediatric subjects for 38 chemistry, lipid, enzyme, and protein analytes [Citation36]. Four blood samples were obtained from 29 healthy children and adolescents (aged 4–<19 years) at different time points throughout an 8-h day [Citation36]. Within- and between-person biological variation, reference change values, and index of individuality were established for these 38 analytes [Citation36]. This study was the first to provide insight into physiological changes that occur within and between pediatric individuals, which quality specifications are suitable for a given test, and how and when to use reference intervals appropriately.

5.3. Indirect versus direct approach to establishing reference intervals

Reference intervals can be established indirectly using historical patient data or directly by testing specimens obtained from healthy subjects. CALIPER compared these two approaches to validate the utility of the indirect approach for use in the pediatric population [Citation37]. CALIPER reference intervals established via direct analysis of healthy child and adolescent specimens were compared with reference intervals established indirectly from hospital-based data using the Hoffman approach. The Hoffman approach aims to identify statistically the portion of a large dataset that represents healthy individuals and to extrapolate these data to determine the 2.5th and 97.5th percentiles [Citation38]. Indirect reference intervals were generally wider than those calculated using the CALIPER direct approach described in detail above [Citation37]. In fact, no reference intervals from the indirect approach fell within the 90% confidence intervals calculated with the direct approach [Citation37]. Thus, although the indirect approach removes the challenge of recruiting healthy individuals, reference intervals robustly established using the direct approach are more accurate and applicable to clinical practice.

5.4. Decision limits

Some biochemical markers require the use of decision limits rather than reference intervals. Decision limits are determined from clinical outcome studies or consensus statements based on expert opinion. When a test value exceeds the decision limit, a diagnostic decision can be made. This contrasts with a reference interval that simply defines the range encompassing 95% of values in a healthy reference population. CALIPER has determined and reported decision limits for two analytes: high sensitivity cardiac troponin I (hs-cTnI) and direct bilirubin [Citation39,Citation40]. The 99th percentile limit of hs-cTnI established from a healthy reference population is used for diagnosing acute myocardial infarction [Citation41]. CALIPER explored the potential sex and age effects on hs-cTnI concentrations in a group of healthy children and adolescents from the CALIPER cohort, as well as the effect of outlier removal on deriving the 97.5th and 99th percentiles as decision limits [Citation39]. This study showed no apparent age or sex difference in hs-cTnI concentration, but concluded that a larger sample size than that recommended by the IFCC (i.e. 300 healthy individuals to establish a 99th percentile) is necessary to prevent inappropriate removal of data as outliers [Citation39]. Total and direct bilirubin are routinely measured for the differential diagnosis of hyperbilirubinemias; however, clinical decision limits for the interpretation of direct bilirubin in pediatrics has been lacking [Citation40]. Total and direct bilirubin were measured in serum samples from 795 healthy children and adolescents; it was concluded that a direct bilirubin concentration of ≥10 µmol/L or >10% of total bilirubin should be used as a decision limit to consider the presence of conjugated hyperbilirubinemia if total bilirubin is also above the reference interval [Citation40].

5.5. Ethnicity and Tanner stage as additional covariates

Several covariates can influence the physiological concentrations of common laboratory markers. CALIPER has thus far partitioned reference intervals by age and sex; however, ethnicity and sexual development may also substantially influence partitioning. Identifying analytes which are significantly affected by these covariates is essential to ensure appropriate reference intervals are established and implemented across populations. CALIPER has performed preliminary analysis on the effect of ethnicity among major ethnic groups including Caucasians, South Asians, East Asians, and others [Citation9,Citation10,Citation12]. Preliminary analysis of the influence of ethnicity was performed in three direct CALIPER reference interval studies that looked at biochemical markers [Citation9], fertility hormones [Citation12], and other chemistry/endocrine markers [Citation10]. Across all three studies, approximately 26% of analytes were found to be significantly different between at least two ethnic groups; however the clinical significance was not assessed. CALIPER is currently performing a more robust and comprehensive analysis of the effect of ethnicity, which will be discussed in “Current and future directions”. The effect of sexual development on reference intervals was examined in Konforte et al. to specifically determine how fertility hormones differ depending on the stage of sexual development [Citation12]. Tanner stages were determined via self-assessment of consenting CALIPER participants. Tanner stage-specific reference intervals were established for fertility hormones.

5.6. Canadian Health Measures Survey (CHMS) collaboration

The CHMS, a program of Statistics Canada, is a nation-wide initiative aimed at addressing limitations within Canada’s health information system [Citation42]. From March 2007 through February 2009, Cycle 1 of the CHMS collected comprehensive health information and blood samples from participants aged 6–<80 years. From August 2009 through November 2011, Cycle 2 of the CHMS collected data from participants aged 3–<80 years. Their sampling was representative of approximately 96% of the Canadian population, and excluded only full-time members of the Canadian Forces, people living on reserves or in other aboriginal settlements, residents of institutions, and people living in certain remote regions. The survey consisted of an initial in-home interview to collect general health information (i.e. nutrition, smoking habits, and medical history), followed by a visit to a mobile examination center to obtain biological samples and direct physical body measures (i.e. height, weight, and blood pressure). Additional details on design and sampling procedures of the CHMS are available at www.statcan.gc.ca/chms.

Through collaboration with Statistics Canada, CALIPER accessed data from the CHMS Cycles 1 and 2, including approximately 12,000 Canadians aged 3–<80 years, to establish age- and sex-specific reference intervals. This collaboration has allowed CALIPER to expand its reference interval database beyond the pediatric population, to include adult and geriatric populations and to determine how representative the CALIPER cohort is of the greater Canadian population. Additionally, as analysis of hematologic parameters requires immediate analysis of whole blood samples, which can be ogistically challenging and costly, this collaboration allowed CALIPER to establish reference intervals for hematology markers for the first time. Overall, CALIPER examined age- and sex-specific changes in biochemical markers [Citation43], endocrine and other chemistry markers [Citation44], and hematology markers [Citation45], and subsequently established a comprehensive database of reference intervals for pediatric, adult, and geriatric Canadian populations.

As the CALIPER cohort has primarily been recruited from the greater Toronto and Hamilton regions and the CHMS participants were recruited at 16 sites across Canada, this provided a valuable opportunity to determine if the CALIPER cohort is similar to that of CHMS and if it is representative of the Canadian population. Extensive similarities were noted between CALIPER and CHMS data, which confirms that the CALIPER cohort is sufficiently representative of the Canadian population. show scatterplots of overlapping CALIPER and CHMS data for alkaline phosphatase (ALP) and creatinine, respectively. The majority of analytes measured in both the CALIPER and CHMS studies also showed great similarities in both concentration and age- and sex-related trends. Additionally, significant differences were noted between the pediatric and adult population for several analytes, further highlighting the importance of establishing reference intervals specific for the pediatric population. Most notable were the significant age-related trends observed for creatinine and uric acid, in which concentrations were much lower in the pediatric compared to adult population [Citation43]. Similarly, ALP and phosphate were also profoundly different across the age range, with significantly higher values in the pediatric compared to adult population [Citation43]. These examples underscore the potential risk to patient safety of using inappropriate reference intervals, as patients can be misclassified due to incorrect interpretation of laboratory test results.

Figure 4. Scatterplots comparing alkaline phosphatase concentration obtained by the CALIPER [Citation9] and CHMS [Citation43] studies for (A) the entire age range covered by both studies (0–<80 years), and (B) the overlapping age range (3–<19 years).

Figure 4. Scatterplots comparing alkaline phosphatase concentration obtained by the CALIPER [Citation9] and CHMS [Citation43] studies for (A) the entire age range covered by both studies (0–<80 years), and (B) the overlapping age range (3–<19 years).

Figure 5. Scatterplots comparing creatinine concentration obtained by the CALIPER [Citation9] and CHMS [Citation43] studies for (A) the entire age range covered by both studies (0–<80 years), and (B) the overlapping age range (3–<19 years).

Figure 5. Scatterplots comparing creatinine concentration obtained by the CALIPER [Citation9] and CHMS [Citation43] studies for (A) the entire age range covered by both studies (0–<80 years), and (B) the overlapping age range (3–<19 years).

6. Current and future directions of CALIPER

Despite CALIPER’s enormous contributions, significant evidence gaps remain in the development of a truly comprehensive national pediatric reference interval database. CALIPER plans to continue its initiative to close the remaining gaps and expand the current database of pediatric reference intervals.

6.1. Ethnic-specific reference intervals

Numerous studies have demonstrated that the lack of consideration for ethnicity-related differences in biomarker levels can pose serious risks to patient safety [Citation46–50]. Therefore, Canada’s multi-ethnic population compels a careful examination of the influence of ethnicity on pediatric reference intervals to minimize diagnostic error. To expand on the preliminary analysis of the changes in biomarker concentration with ethnicity, CALIPER is currently performing a more comprehensive and robust analysis comparing serum concentrations of laboratory biomarkers in four of the major Canadian ethnic groups (i.e. Caucasians, East Asians, South Asians, and Afro-Caribbeans). Ethnicity-specific reference intervals may need to be calculated for a select group of analytes that show statistically and clinically significant differences among ethnic groups examined.

6.2. Neonatal and infantile reference intervals

The CALIPER pediatric reference interval database is extremely robust from 1 to <19 years of age. However, the data available for neonates and infants (0–<1 year) are limited by the fact that these samples were collected primarily from outpatient clinics. CALIPER is initiating a new project to address this limitation by establishing comprehensive neonatal and infant reference intervals through recruitment of expectant mothers to allow monitoring of circulating biomarkers at birth and throughout the first year of life. This will provide, for the first time, complete reference interval data for neonates and infants, which is currently very limited due to difficulties in recruitment and blood collection in this age group. As CALIPER will be recruiting pregnant women, trimester-specific reference intervals, which are also severely lacking, will be established. These planned studies will lead to improved healthcare for neonates and infants in complex, tertiary care pediatric centers, and pregnant women across Canada.

6.3. Fasting versus postprandial blood sampling

Collecting a blood sample in the fasting versus postprandial (i.e. after a meal) state is an important pre-analytical consideration that can introduce test result variability for a subset of analytes. While most analytes are not altered following consumption of a meal, and, therefore, can be measured in a random sample (i.e. anytime, irrespective of hours fasted), some analytes are significantly altered in the postprandial state. Most notably, laboratory tests used to diagnose metabolic abnormalities, including type 2 diabetes and lipid disorders (i.e. glucose and triglycerides, respectively) vary considerably between the fasting and postprandial states and, therefore, have traditionally required patients to fast 10–12 h prior to blood sample collection. However, postprandial rather than fasting measurements may be more informative of the metabolic status of an individual and subsequently may be a more valuable measure of metabolic abnormalities in insulin resistant states and a better predictor of future cardiovascular disease risk. In fact, a prospective study of over 26,000 US women showed that non-fasting triglyceride levels were independently associated with incident cardiovascular events, while fasting levels showed little independent relationship [Citation51]. As several analytes involved in lipid metabolism are secreted following a meal and thus play physiological roles in the postprandial state, it is not surprising that the postprandial profile may be a valuable, or even superior, measure of abnormal lipid metabolism. CALIPER plans to study the postprandial profile of analytes involved in postprandial lipid metabolism in adolescents by collecting blood samples at fasting and specified time points following ingestion of a high-fat drink. Analytes implicated in postprandial lipid metabolism include intestinal apolipoprotein B-48 (apoB-48)-containing triglyceride-rich lipoproteins (i.e. chylomicrons), which carry dietary lipids throughout the circulation; their breakdown products (i.e. remnant lipoproteins), which have been implicated in cardiovascular disease risk (reviewed in [Citation52]); as well as gut peptides (e.g. glucagon-like peptide 1 (GLP-1) and GLP-2), which are secreted following nutrient ingestion and have been recently implicated in regulating lipid metabolism (reviewed in [Citation53,Citation54]). Similar to using the oral glucose tolerance test to determine impaired glucose tolerance and increased risk of type 2 diabetes, the “oral fat tolerance test” may be a helpful tool to determine impaired lipid tolerance and increased risk for cardiovascular events.

6.4. External quality assurance program

To ensure continual relevance of CALIPER reference intervals in clinical practice, now and in the future, it is imperative to monitor instrument-specific reagent formulations and method calibrations in each clinical laboratory. Accordingly, a CALIPER Quality Assurance Program is essential to maintain the accuracy of the CALIPER reference interval database in the face of changing methodologies, reagent formulations, and calibration biases. CALIPER will be developing and instituting an external quality assurance (EQA) program that will be anchored to CALIPER. This will be done in collaboration with CEQAL (www.ceqal.com), a world-leading reference method laboratory specialized in the delivery of accuracy assessment solutions for improving the quality of testing in healthcare. CALIPER’s goal is not only to provide laboratories with appropriate reference intervals but also to encourage harmonization of pediatric reference intervals across all clinical laboratories, particularly those using the same instrumentation/reagents.

6.5. National and global knowledge translation

CALIPER plans to continue its KT campaign to implement pediatric reference intervals into medical laboratories and hospitals across Canada and globally. In addition to continual updates to the online database and mobile device application (app), CALIPER plans to support laboratories in Canada and around the world that wish to implement CALIPER reference intervals in clinical practice. A national KT campaign has already begun in collaboration with senior clinicians and laboratory directors in Canadian hospitals. This campaign will aid in implementing CALIPER reference intervals to achieve Canada-wide harmonization of laboratory test interpretation. Implementation of CALIPER reference intervals will lead to harmonized and evidence-based clinical decision making, will reduce diagnostic errors, and will improve the quality of pediatric healthcare in Canada.

7. Pediatric reference interval tables – the CALIPER database

A compendium of CALIPER reference value data is reported here, providing a single resource for clinical laboratories and clinical programs to access the complete CALIPER database published to-date. Data are included for laboratory markers measured on all major clinical chemistry platforms for routine application and implementation in clinical practice ().

Table 1. Pediatric reference intervals for the Abbott ARCHITECT – age and sex specific.

Table 2. Pediatric reference intervals for the Beckman AU – age and sex specific.

Table 3. Pediatric reference intervals for the Beckman DxC – age and sex specific.

Table 4. Pediatric reference intervals for the Beckman DxI – age and sex specific.

Table 5. Pediatric reference intervals for the Ortho VITROS 5600 – age and sex specific.

Table 6. Pediatric reference intervals for the Roche cobas 6000 – age and sex specific.

Table 7. Pediatric reference intervals for the Roche Modular P – age and sex specific.

Table 8. Pediatric reference intervals for the Siemens Vista – age and sex specific.

Table 9. Pediatric reference intervals for markers of metabolic disease, steroid hormones, and vitamins a and E (HPLC, UPLC, LC-MS/MS).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Abbreviations
AFP=

α-fetoprotein

ALT=

alanine aminotransferase

ALP=

alkaline phosphatase

anti-CCP=

anti-cyclic citrullinated peptide

anti-TPO=

anti-thyroperoxidase

apoB-48=

apolipoprotein B-48

BAT=

bioavailable testosterone

CA125=

cancer antigen 125

CA15-3=

cancer antigen 15-3

CALIPER=

Canadian Laboratory Initiative on Pediatric Reference Intervals

CHMS=

Canadian Health Measures Survey

CLSI=

Clinical and Laboratory Standards Institute

CSCC=

Canadian Society of Clinical Chemists

EQA=

external quality assurance

FAI=

free androgen index

FSH=

follicle-stimulating hormone

FT=

free testosterone

FT4=

free thyroxine

GLP=

glucagon-like peptide

HPLC=

high-performance liquid chromatography

hs-cTnI=

high sensitivity cardiac troponin I

IgG=

immunoglobulin G

IgM=

immunoglobulin M

KT=

knowledge translation

LC-MS/MS=

liquid chromatography tandem mass spectrometry

LH=

luteinizing hormone

PTH=

parathyroid hormone

PSA=

prostate-specific antigen

Q–Q=

quantile–quantile

SHBG=

sex-hormone binding globulin

TnI=

troponin I

TSH=

thyroid stimulating hormone

UPLC=

ultra-performance liquid chromatography

Acknowledgements

The CALIPER program would like to thank thousands of families and children who made the study possible by participating and volunteering to donate blood samples. We also thank hundreds of CALIPER project coordinators, volunteers, clinical chemistry fellows and faculty, as well as industry partners and collaborators across Canada, for their tremendous contributions to the success of the CALIPER study.

Disclosure statement

The authors have no conflict of interest to declare.

Additional information

Funding

The CALIPER study has been supported by operating grants from the Canadian Institutes of Health Research (CIHR), and Sanford Jackson Endowment Funds at the Hospital for Sick Children Foundation, as well as grant and material support from Abbott, Beckman, Ortho, Roche, and Siemens Diagnostics.

References

  • CLSI. Defining, establishing, and verifying reference intervals in the clinical laboratory; Approved Guideline -– Third Edition. CLSI document C28-A3. Wayne (PA): Clinical and Laboratory Standards Institute; 2008.
  • Horn PS, Pesce AJ. Reference intervals: an update. Clin Chim Acta. 2003;334:5–23.
  • Jones GR, Koetsier SD. RCPAQAP first combined measurement and reference interval survey. Clin Biochem Rev. 2014;35:243–250.
  • Chan MK, Seiden-Long I, Aytekin M, et al. Canadian Laboratory Initiative on Pediatric Reference Interval Database (CALIPER): pediatric reference intervals for an integrated clinical chemistry and immunoassay analyzer, Abbott ARCHITECT ci8200. Clin Biochem. 2009;42:885–891.
  • Blasutig IM, Jung B, Kulasingam V, et al. Analytical evaluation of the VITROS 5600 Integrated System in a pediatric setting and determination of pediatric reference intervals. Clin Biochem. 2010;43:1039–1044.
  • Kulasingam V, Jung BP, Blasutig IM, et al. Pediatric reference intervals for 28 chemistries and immunoassays on the Roche cobas 6000 analyzer – a CALIPER Pilot Study. Clin Biochem. 2010;43:1045–1050.
  • Henderson MPA, Grey V. Establishing and evaluating pediatric thyroid reference intervals on the Roche Modular Analytics E 170 using computational statistics and data-mining techniques. Clin Biochem. 2011;44:767–770.
  • Huang Y, Eapen E, Steele S, et al. Establishment of reference intervals for bone markers in children and adolescents. Clin Biochem. 2011;44:771–778.
  • Colantonio DA, Kyriakopoulou L, Chan MK, et al. Closing the gaps in pediatric laboratory reference intervals: a CALIPER database of 40 biochemical markers in a healthy and multiethnic population of children. Clin Chem. 2012;58:854–868.
  • Bailey D, Colantonio D, Kyriakopoulou L, et al. Marked biological variance in endocrine and biochemical markers in childhood: establishment of pediatric reference intervals using healthy community children from the CALIPER cohort. Clin Chem. 2013;59:1393–1405.
  • Karbasy K, Lin DCC, Stoianov A, et al. Pediatric reference value distributions and covariate-stratified reference intervals for 29 endocrine and special chemistry biomarkers on the Beckman Coulter Immunoassay Systems: a CALIPER study of healthy community children. Clin Chem Lab Med. 2016;54:643–657.
  • Konforte D, Shea JL, Kyriakopoulou L, et al. Complex biological pattern of fertility hormones in children and adolescents: a study of healthy children from the CALIPER cohort and establishment of pediatric reference intervals. Clin Chem. 2013;59:1215–1227.
  • Kyriakopoulou L, Yazdanpanah M, Colantonio DA, et al. A sensitive and rapid mass spectrometric method for the simultaneous measurement of eight steroid hormones and CALIPER pediatric reference intervals. Clin Biochem. 2013;46:642–651.
  • Bevilacqua V, Chan MK, Chen Y, et al. Pediatric population reference value distributions for cancer biomarkers and covariate-stratified reference intervals in the CALIPER cohort. Clin Chem. 2014;60:1532–1542.
  • Raizman JE, Cohen AH, Teodoro-Morrison T, et al. Pediatric reference value distributions for vitamins A and E in the CALIPER cohort and establishment of age-stratified reference intervals. Clin Biochem. 2014;47:812–815.
  • Teodoro-Morrison T, Kyriakopoulou L, Chen YK, et al. Dynamic biological changes in metabolic disease biomarkers in childhood and adolescence: a CALIPER study of healthy community children. Clin Biochem. 2015;48:828–836.
  • Raizman JE, Quinn F, Armbruster DA, et al. Pediatric reference intervals for calculated free testosterone, bioavailable testosterone and free androgen index in the CALIPER cohort. Clin Chem Lab Med. 2015;53:e239–e243.
  • Higgins V, Fung AWS, Chan MK, et al. Pediatric reference intervals for 29 Ortho VITROS 5600 immunoassays using the CALIPER cohort of healthy children and adolescents. Clin Chem Lab Med. 2017.
  • Vermeulen A, Verdonck L, Kaufman JM. A critical evaluation of simple methods for the estimation of free testosterone in serum. J Clin Endocrinol Metab. 1999;84:3666–3672.
  • Kelly J, Raizman JE, Bevilacqua V, et al. Complex reference value distributions and partitioned reference intervals across the pediatric age range for 14 specialized biochemical markers in the CALIPER cohort of healthy community children and adolescents. Clin Chim Acta. 2015;450:196–202.
  • Boyd JC. Defining laboratory reference values and decision limits: populations, intervals, and interpretations. Asian J Androl. 2010;12:83–90.
  • Marshall WA, Tanner JM. Variations in pattern of pubertal changes in girls. Arch Dis Child. 1969;44:291–303.
  • Marshall WA, Tanner JM. Variations in the pattern of pubertal changes in boys. Arch Dis Child. 1970;45:13–23.
  • Statistics Canada. Ethnic origins, 2006 counts, for Canada, provinces and territories – 20% sample data [cited 2017 Sep 12]. [Table]. Available from: http://www12.statcan.ca/census-recensement/2006/dp-pd/hlt/97-562/pages/page.cfm?Lang=E&Geo=PR&Code=01&Table=2&Data=Count&StartRec=1&Sort=3&Display=All.
  • Harris EK, Boyd JC. On dividing reference data into subgroups to produce separate reference ranges. Clin Chem. 1990;36:265–270.
  • Tukey J. Exploratory data analysis. Boston (MA): Addison-Wesley; 1977. p. 1–688.
  • Hubert M, Van der Veeken S. Outlier detection for skewed data. J Chemometrics. 2008;22:235–246.
  • Horn PS, Pesce AJ, Copeland BE. A robust approach to reference interval estimation and evaluation. Clin Chem. 1998;44:622–631.
  • Estey MP, Cohen AH, Colantonio DA, et al. CLSI-based transference of the CALIPER database of pediatric reference intervals from Abbott to Beckman, Ortho, Roche and Siemens Clinical Chemistry Assays: direct validation using reference samples from the CALIPER cohort. Clin Biochem. 2013;46:1197–1219.
  • Araújo PAT, Thomas D, Sadeghieh T, et al. CLSI-based transference of the CALIPER database of pediatric reference intervals to Beckman Coulter DxC biochemical assays. Clin Biochem. 2015;48:870–880.
  • Abou El Hassan M, Stoianov A, Araújo PAT, et al. CLSI-based transference of CALIPER pediatric reference intervals to Beckman Coulter AU biochemical assays. Clin Biochem. 2015;48:1151–1159.
  • Higgins V, Chan MK, Nieuwesteeg M, et al. Transference of CALIPER pediatric reference intervals to biochemical assays on the Roche cobas 6000 and the Roche Modular P. Clin Biochem. 2016;49:139–149.
  • CLSI. Method comparison and bias estimation using patient samples; Approved Guideline – Second Edition. CLSI document EP9-A2. Wayne (PA): Clinical and Laboratory Standards Institute; 2002.
  • Brinc D, Chan MK, Venner AA, et al. Long-term stability of biochemical markers in pediatric serum specimens stored at −80 °C: a CALIPER Substudy. Clin Biochem. 2012;45:816–826.
  • Pasic MD, Colantonio DA, Chan MK, et al. Influence of fasting and sample collection time on 38 biochemical markers in healthy children: a CALIPER substudy. Clin Biochem. 2012;45:1125–1130.
  • Bailey D, Bevilacqua V, Colantonio DA, et al. Pediatric within-day biological variation and quality specifications for 38 biochemical markers in the CALIPER cohort. Clin Chem. 2014;60:518–529.
  • Shaw JLV, Cohen A, Konforte D, et al. Validity of establishing pediatric reference intervals based on hospital patient data: a comparison of the modified Hoffmann approach to CALIPER reference intervals obtained in healthy children. Clin Biochem. 2014;47:166–172.
  • Hoffmann RG. Statistics in the practice of medicine. JAMA. 1963;185:864–873.
  • Kavsak PA, Rezanpour A, Chen Y, et al. Assessment of the 99th or 97.5th percentile for cardiac troponin I in a healthy pediatric cohort. Clin Chem. 2014;60:1574–1576.
  • Devgun MS, Chan MK, El-Nujumi AM, et al. Clinical decision limits for interpretation of direct bilirubin–a CALIPER study of healthy multiethnic children and case report reviews. Clin Biochem. 2015;48:93–96.
  • Kimenai DM, Henry RMA, van der Kallen CJH, et al. Direct comparison of clinical decision limits for cardiac troponin T and I. Heart Br Card Soc. 2016;102:610–616.
  • Tremblay M, Wolfson M, Connor Gorber S. Canadian Health Measures Survey: rationale, background and overview. Health Rep. 2007;18(Suppl):7–20.
  • Adeli K, Higgins V, Nieuwesteeg M, et al. Biochemical marker reference values across pediatric, adult, and geriatric ages: establishment of robust pediatric and adult reference intervals on the basis of the Canadian Health Measures Survey. Clin Chem. 2015;61:1049–1062.
  • Adeli K, Higgins V, Nieuwesteeg M, et al. Complex reference values for endocrine and special chemistry biomarkers across pediatric, adult, and geriatric ages: establishment of robust pediatric and adult reference intervals on the basis of the Canadian Health Measures Survey. Clin Chem. 2015;61:1063–1074.
  • Adeli K, Raizman JE, Chen Y, et al. Complex biological profile of hematologic markers across pediatric, adult, and geriatric ages: establishment of robust pediatric and adult reference intervals on the basis of the Canadian Health Measures Survey. Clin Chem. 2015;61:1075–1086.
  • Lim E, Miyamura J, Chen JJ. Racial/ethnic-specific reference intervals for common laboratory tests: a comparison among Asians, Blacks, Hispanics, and White. Hawaii J Med Public Health J Asia Pac Med Public Health. 2015;74:302–310.
  • Boucai L, Surks MI. Reference limits of serum TSH and free T4 are significantly influenced by race and age in an urban outpatient medical practice. Clin Endocrinol (Oxf). 2009;70:788–793.
  • Boucai L, Hollowell JG, Surks MI. An approach for development of age-, gender-, and ethnicity-specific thyrotropin reference limits. Thyroid. 2011;21:5–11.
  • Brewster LM, Mairuhu G, Sturk A, et al. Distribution of creatine kinase in the general population: implications for statin therapy. Am Heart J. 2007;154:655–661.
  • Groesbeck D, Köttgen A, Parekh R, et al. Age, gender, and race effects on cystatin C levels in US adolescents. Clin J Am Soc Nephrol. 2008;3:1777–1785.
  • Bansal S, Buring JE, Rifai N, et al. Fasting compared with nonfasting triglycerides and risk of cardiovascular events in women. JAMA. 2007;298:309–316.
  • Masuda D, Yamashita S. Postprandial hyperlipidemia and remnant lipoproteins. J Atheroscler Thromb. 2017;24:95–109.
  • Xiao C, Dash S, Morgantini C, et al. Gut peptides are novel regulators of intestinal lipoprotein secretion: experimental and pharmacological manipulation of lipoprotein metabolism. Diabetes. 2015;64:2310–2318.
  • Farr S, Taher J, Adeli K. Glucagon-like peptide-1 as a key regulator of lipid and lipoprotein metabolism in fasting and postprandial states. Chddt. 2014;14:126–136.