Reference values of blood cell counts in the first days of life

Giovanni Melioli¹, Francesco Maria Risso^2,5, Andrea Sannia^2,5, Giovanni Serra^2,5, Roberto Bologna¹, Michele Mussap^3,5, Salvatore Mangraviti¹, Patrizia Fortini¹, Fabio Facco¹, Giorgio Reggiardo⁴, Giuseppe Buonocore^5,7, Giovanni Corsello^5,8, Vassilios Fanos^5,6, Antonello Del Vecchio⁹, Claudio Fabris¹⁰, Diego Gazzolo^5,11

1Clinical Pathology Laboratories, G. Gaslini Children's Hospital, Genoa, Italy, ²Department of Maternal, Fetal and Neonatal Medicine G. Gaslini Children's Hospital, Genoa, Italy, ³Laboratory of Clinical Biochemical Medicine, S. Martino, Genoa, Italy, ⁴Biometric Unit, Medi-Service srl, Genoa, Italy, ⁵Neonatal Clinical and Biochemical Italian Research Group from Italian Society of Clinical Biochemistry (SiBioC) and Italian Society of Neonatology (SIN), ⁶Neonatal Intensive Care Unit University of Cagliari, Cagliari, Italy, ⁷Department of Maternal, Fetal, Neonatal and Reproductive Medicine University of Siena, Siena, Italy, ⁸Department of Maternal, Fetal and Neonatal Medicine University of Palermo, Palermo, Italy, ⁹Neonatal Intensive Care Unit Di Venere Hospital, Bari, Italy, ¹⁰Chair of Neonatog, University of Turin, Turin, Italy, ¹¹Neonatal Intensive Care Unit C. Arrigo Children's Hospital, Alessandria, Italy

TABLE OF CONTENTS

1. Abstract

2. Introduction

3. Materials and methods

3.1. Evaluation of sample size

3.2. Blood Cells Collection

3.3. Laboratory Procedure

3.4.Quality Controls

3.5. Statistical Analysis

4. Results

4.1. Quality Controls

4.2. Study Populations

4.3. Hematological Parameters

5. Discussion

6. References

1. ABSTRACT

The lack of updated neonatal reference values for hematological parameters impacts significantly with clinical management of both healthy and sick newborns. The present pilot study was thus aimed at assessing updated hematological Italian reference values in late preterm and term newborns. From January 2004 to December 2008 hematological laboratory tests were performed in 1175 newborns (820 healthy and 355 sick controls) between 33-41 weeks of gestation, during the first four days after birth. Hematological parameters were sorted for gender and gestational age and statistically analyzed. No gender-related differences were observed at different weeks of gestation and no significant differences were found when study population was sub-grouped for late preterm and term newborns. During the first 4 days of life erythrocytes and platelets remained stable whilst white blood cell counts and differentials were significantly modified. This study shares updated reference values for hematological parameters in the early phases after birth and offers additional support for improving the management of sick infants.

2. INTRODUCTION

It has been recently highlighted that the availability of reference values for laboratory parameters in the first days after birth can constitute a relevant tool for an optimal management of post-natal adaptation phases. In this setting, if a normal population cannot be collected as controls, to define the reference intervals, laboratory results from unselected subjects stratified for age and sex can be recorded and further elaborated. From a statistical point of view, the most heterogeneous is the population, the larger should be the sample. In addition, it is commonly accepted to calculate reference values by identifying the 2.5-97.5% ranges obtained in a cohort of patients addressing to a laboratory facility. This is an adequate approach if the population is heterogeneous but can be significantly improved by using the average values and the multiple of standard deviations whether the population had a normal distribution. Conversely, when distribution is asymmetric, a different approach should be applied, such as the Kairisto algorithm (1). Additional problems in post-natal laboratory management regard: i) ethical problems related with blood sampling in healthy donors; ii) ethnic differences among different study-populations, and, iii) clinical and laboratory criteria for classifying healthy preterm/late preterm newborns. This latter point is of relevance since in the first hours after birth, post-natal adaptation can require a high/different capacity of transporting oxygen affecting hemoglobin values and blood concentration. Other confounding factors in the first days of life are breathing activity, perspiratio insensibilis, water and/or milk assumption and active state behavior that may result in blood cell dilution. Thus, any reference values in this period of investigation is expected to be highly variable and therefore the identification of a particularly accurate reference range is eagerly requested in clinical daily practice. In the last decade, technological improvement in clinical pathology laboratories allowed to measure classic blood cell-related parameters such as hemoglobin (Hb), Red Blood Cells (RBC), White Blood Cells (WBC) and its subpopulations, platelets and calculated parameters with great accuracy and precision. However, reference values in use at present are based on historical data (2-13) largely based on low technology instrumentations and only recently, updated reference values were reported (14).

The aim of the present pilot study was thus to evaluate whether a multicentre study, representative of different regions and organizations in Italy, could share novel reference ranges for hematological parameters. To achieve this goal: i) a feasibility study was carried out in a pediatric Institution in which physiological and high risk deliveries/infants were carried out and, ii) a comparison between healthy and sick newborn hematological parameters was conducted.

3. MATERIAL AND METHODS

The study was approved by Italian Society of Neonatology Board, by Local Ethics Committees and the parents of the subjects gave signed and informed consent.

In the period January 2004-December 2008, 1093 laboratory tests, related to 820 newborns (48,5% females and 51,5% males), were performed. One hundred and twenty eight samples were processed in 2004, 197 in 2005, 187 in 2006, 272 in 2007 and 172 in 2008. In 29% of patients, blood was collected at birth, 25% on day 1, 25% on day 2, 12% on day 3 and 9% on day 4. The gestational age distribution for healthy newborns at sampling was 20% for late preterm (33-36 wks), 47% for 37-39 wks and 33% for 40-42 wks.

Gestational age was determined by clinical data and by a first trimester ultrasound scan. Appropriate growth was defined by the presence of ultrasonographic signs (when biparietal diameter and abdominal circumference were between the 10-90th centiles) according to the normograms of Campbell and Thoms (15) and by postnatal confirmation of a birth-weight between the 10-90th centiles according to our population standards after corrections for the mother's height, weight and parity and the sex of the newborn. All healthy infants admitted to the study fulfilled all the following criteria: no maternal illness, no signs of fetal distress, pH more than 7.2 in cord blood or venous blood, Apgar scores at 1 and 5 minutes more than 7 (16). All control newborns were in normal clinical condition and showed no overt neurological syndrome at the discharge from the hospital (Table 1). Controls subjects were 355 newborns whose neonatal outcomes did not fulfill American Academy of Pediatrics criteria for healthy newborns. In these patients 473 blood samples were collected.

3.1. Evaluation of sample size

Analysis of laboratory data expressing a certain heterogeneity requires the accurate definition of the sample size. On these bases, we decided to calculate the results on a number of subjects that was always wider than that formally required. The N-Query software was then used to calculate how many samples, in each experimental condition, should be needed for the calculation of the lower and upper percentiles (2.5-97.5%).

3.2. Blood cells collection

Blood samples were obtained either from healthy or sick children during the first days from delivery. Blood was collected using venipuncture and immediately transferred to the laboratory facility for the tests. All samples were identified using a unique barcode for that individual, that test and that time. Laboratory results were stored in the Laboratory Information System (LIS), using a specific software (Concerto, Metafora, Milan, Italy). Data from the LIS were collected using another software (Frequenza, Metafora, Milan, Italy) that may store in a export file not only the single laboratory but also personal data of the patient and the time of reception in the laboratory. This is particularly useful to sort samples obtained at different time interval from the birth. Of note, if more than one sample was processed for a single patient, only the earliest was included in this study.

3.3. Laboratory procedure

Routine hematological assays are performed using two different platforms. The Siemens Advia 1210 (Munich, Germany) that uses standard procedures for red blood cell count and peroxidase staining of granulocytes and the Beckman Coulter 750 (Brea, CA, USA) that employs a standard approach for red blood cells but performs differential white blood cell count using volume, scatters and conductivity. The two instruments, according to manufacturers' instructions, were daily aligned in order to obtain identical results. The following parameters were analyzed: total number of circulating RBC, Hb, Hematocrit rate (Ht), Mean Cellular Volume of RBC (MCV), Mean Cellular Hb concentration of Hb (MCH), mean cellular Hb concentration (MCHC), RBC Distribution Width (RDW), total circulating WBC, percentage and absolute number of neutrophilic granulocytes (NEU% and NEU), lymphocytes (LYM% and LYM), monocytes (MON% and MON), eosinophils (EOS% and EOS), basophils (BAS% and BAS), and finally platelets (PLT).

3.4. Quality controls

An internal quality control program was performed to evaluate the day-by day performances of instruments and procedure. An external Quality Assurance program was also performed with a standard daily frequency. Both internal and external controls were daily checked by the Quality Control Responsible of the laboratory. In addition, samples from normal donors were also collected to check the stability of the average values of hematological parameters over the months and the years. Data collected for Quality Controls as well as data from healthy adult donors were downloaded from the LIS using the same procedure described for newborn samples. Notably, the Institution has been awarded for excellence by the Joint Commission International in 2006 and the laboratory itself has an ISO 9000 certification since 2003.

3.5. Statistical analysis

Files downloaded from the LIS were analyzed with GraphPad Prism version 3.00 for Windows (GraphPad Software, Inc., San Diego, California, USA), in order to evaluated the following parameters: Average, standard deviation, asymmetry, kurtosis. If the distribution of the single parameter was symmetric, a 2.5-97.5% interval was calculated. On the contrary, in the presence of an asymmetric distribution (usually present for rare events), a different approach (as suggested by Kairisto) was employed. For this, an internal MS-Excel routine was developed, validated and then used. Differences in neonatal outcomes and in hematological parameters among studied groups, after sort by gender, were assessed by using ANOVA on ranks test and post-hoc Tukey's test to isolate the statistical differences. Statistical analysis was performed by using Student's t test for continuous variables and Mann-Whitney U-two sided test when parameters were not normally distributed. Categorical data were analyzed by means of Fisher's exact test or chi-square analysis as appropriate. Linear regression analysis was used to assess correlation between gestational age at sampling and hematological parameters. A P less than 0.05 was considered statistically significant.

4. RESULTS

4.1. Study populations

Neonatal characteristics of the late preterm and term healthy newborns admitted into the study are shown in Table 1. As expected gestational age and weight at birth was significantly (P less than 0.05, for both) lower in late preterm newborns, whilst, according to recruitment criteria, no significant differences (P more than 0.05, for all) regarding Apgar scores at 1^st and 5^th minutes, venous blood pH, partial venous CO₂ and O₂ pressures and base excess values were shown. No significant differences (P more than0.05, for all) were found in neonatal outcomes and admission criteria when evaluated for maternal ethnic and racial characteristics.

4.2. Quality controls

During the period of investigation, a number of quality controls were performed in order to demonstrate that stability and reproducibility of the procedures. All internal and external controls resulted within the expected ranges. To further support this evidence, in Figure 1 are presented the mean values of Hb, obtained in the same period and measured in adult control subjects. As shown, the variability of the mean values was always within 7%, thus suggesting that in the whole period the laboratory performances were under control.

4.3. Hematological parameters

Data obtained in newborns were then analyzed. Figure 2 shows hematological parameters distributions: RBC, Hb, Ht, WBC and PLT counts showed a symmetrical pattern, while other parameters did not. Therefore, reference limits for the first four parameters were calculated using the 2.5-97.5 percentiles, whilst reference values for other parameters were assessed by using same percentiles and the Kairisto procedure for asymmetric data.

Table 2 shows the results of this analysis on the whole population, after sort for gender, independently from the gestational age and the time at sampling. Table 3 shows that there were no gender differences (P more than 0.05, for all) among laboratory parameters in both late preterm and term newborns. In Table 4 laboratory data are associated to different weeks of gestation. No significant differences (P more than 0.05, for all) were found when study population was corrected for gestational age at sampling as well as subgrouping for late preterm (33-36 wks) and term newborns (A: 37-39; B: 40-42 wks, respectively). Table 5 reports hematological parameters changes during the first 4 days from birth. Of note, while RBC and PLT remained virtually unmodified (P more than 0.05, for all), Hb concentrations, Ht% and WBC counts were significantly decreased when compared at birth monitoring time-point. WBC population patterns expressed as percentage values during the first days were of particular interest: neutrophils significantly decreased (P less than 0.05, for all) from 70% to 36%, basophils had a flat trend, while lymphocyte, monocyte and eosinophil percentages showed a slight increase without reaching a statistical significance (P more than 0.05, for all). The finding was noteworthy when WBC populations were expressed as absolute values (Table 6). In details, the mean number of neutrophils was significantly (P less than 0.05) decreased, from about 16000/mmc, at birth, to about 4000/mmc at day 4 (a reduction of about 75%). Lymphocyte, monocyte and eosinophil counts showed a flat pattern, whilst basophil pattern showed a significant decrease from birth to day 4, with an estimated reduction of the 50%.

To further validate the reference values calculated and shown in the abovementioned tables, a representative population of high risk newborns was also studied. As expected, high risk newborns showed lower Ht and Hb concentrations (estimated difference in percentage: 13%, for both) whilst MCV, MCH and MCHC did not differ. Indeed, significant differences (P less than 0.05, for all) were observed for WBC counts expressed as absolute numbers and as percentages.

5. DISCUSSION

The present pilot study offers hematological reference values in the first days after birth in Italian late preterm and term newborns. The finding is of relevance being reference values and their updates eagerly awaited in clinical daily practice of all disciplines especially in neonatology and neonatal critical care medicine. The request is supported not only by recent technological improvement (manual methods vs. low level automation), that has considerably changed laboratory assessment strategies, but mainly by a series of confounding factors related to population study (i.e. ethnic and socioeconomic differences, immigrations rate etc) that can affect the availability of previous observations. In this respect, it is noteworthy that Caucasians do not represent, nowadays, the totality of the newborns attending to Italian Children' hospitals. Of note, in the present series, no significant differences have been found in neonatal outcomes, admission criteria and finally in laboratory results when corrected for maternal ethnic and racial characteristics.

Despite technological improvement characterized by: i) improvemend of measurement performances with excellent reproducibility; ii) assessment in a variety of biological fluids; iii) possible use in longitudinal monitoring because results can be obtained rapidly and by the use of very small blood volumes, the validation on a large number of cases for standard laboratory parameters reference values is highly required. This issue is of utmost importance when these criteria have to be applied in the perinatal period when several confounding factors, participating in the so called post-natal adaptation phase, are still present. To our knowledge, the present results is the first attempt to update laboratory data referring to neonatal period in Italy. Of note, the present findings fit a recent report from a U.S. multicentre study recruiting a cohort of more than 24,000 newborns. In more detail, no significant differences were observed regarding Hb (18.1 g vs 17.5 g) and Ht values (54% vs 51,0%) between the two reports. Further multicenter investigations may be useful in order to verify, in a larger population study, our findings on white blood cells population. However, we did not find any consistent differences with a previous report by our institutes conducted in the late seventies (data not shown).

In the present study we also found no significant differences in laboratory parameters when corrected for gender and gestational age at sampling. This latter issue can be of relevance, bearing in mind the importance of an adequate laboratory monitoring in a such selected population. There is growing evidence that late-preterm population nowadays represent one of the major high-risk newborn population. The main explanation resides in the delivery period (33-36 wks) at a stage when whole organs and especially the central nervous system growth-process are at their higher levels (17). In particular, CNS growth is estimated in 35% for brain weight, 37% for brain volume, and synaptogenesis, dendritic arborization and axonal elongation processes are at their higher levels of activation. Thus, adequate reference curves for laboratory parameters in a such delicate fetal/neonatal period could be especially useful.

In conclusion, our findings showing changes in laboratory parameters in the first days after birth open a new cue for further multicentre investigations aimed at offering, in a larger population study, reference values of the main laboratory parameters currently used in neonatology and critical care medicine.

6. REFERENCES

1. Kairisto V, Hänninen KP, Leino A, Pulkki K, Peltola O, Näntö V, Voipio-Pulkki LM, Irjala K: Generation of reference values for cardiac enzymes from hospital admission laboratory data. Eur J Clin Chem Clin Biochem 32:789-96 (1994)
PMid:7865617

2. American Society for Clinical Laboratory Science: Consumer laboratory testing information page. Available at: www.ascls.org/labtesting/index.php. Accessed December 8, 2008

3. Mugrage ER, Andersen MI: Values for red blood cells of average infants and children. Am J Dis Child 51(4):775-782 (1936)

4. Waugh TR, Merchant FT, Maugham GB: Blood studies on the newborn. Am J Med Sci 198(5):646-652 (1939)
doi:10.1097/00000441-193911000-00009

5. DeMarsh QB, Alt HL, Windle WF: The effect of depriving the infant of its placental blood. JAMA 116(23):2568-2575 (1941)
PMid:14953483

6. Dochain J, Lemage L, Lambrechts A: Principal hematological data in normal newborn infants . Arch Fr Pediatr 9(3):274-278 (1952)
PMid:14377615    PMCid:2011707

7. Marks J, Gairdenr D, Roscoe JD: Blood formation in infancy: III-cord blood. Arch Dis Child 30(150):117-120 (1955)
doi:10.1136/adc.30.150.117
PMid:13410384

8. Guest GM, Brown EW: Erythrocytes and hemoglobin of the blood in infancy and childhood: III-factors in variability, statistical studies. AMA J Dis Child 93(5):486-509 (1957)
PMid:13425654    PMCid:2012042

9. Rooth G, Sjostedt S: Haemoglobin in cord blood in normal and prolonged pregnancy. Arch Dis Child 32(162):91-92 (1957)
doi:10.1136/adc.32.162.91
PMid:4834018    PMCid:1648869

10. Burman D, Morris AF: Cord haemoglobin in low birth weight infants. Arch Dis Child 49(5):382-385 (1974)
doi:10.1136/adc.49.5.382

11)Wintrobe MM: A simple and accurate hematocrit. J Lab ClinMed 15(1):287-289 (1929)

12. Wintrobe MM: Anemia. Arch Intern Med 54(2):256-261 (1934)

13. Jopling J, Henry E, Wiedmeier SE, Christensen RD: Reference ranges for hematocrit and blood hemoglobin concentration during the neonatal period: data from a multihospital health care system. Pediatrics 123(2):e333-7(2009)
PMid:16024340

14. Manconi M, Rolfo A, Cardaroli S, Brini M, Danise P: Hematologic values in healthy and small for gestational age newborns. Lab Hematol 11(2): 152-6 (2005)
doi:10.1532/LH96.04076

15. Campelll S, Thoms A:Measurement of the fetal head to abdomen circumference ratio in the assessment of growth retardation. Ultrasound. Br J Obstet Gynecol 182:650-4. (2000)

16. American Academy of Pediatrics and American College of Obstetricians and Gynecologist (2002. Guidelines for perinatal care, ed 5, Elk Grove Village, Ill, The Academy. (2002)

17. Kinney, HC, Armstrong, DL. Perinatal neuropathology, in Graham DI, Lantos PE eds: Greenfield's Neuropathology ed 7. London, Arnold, 557-559 (2000)

Key-Words: Newborns, Late-Preterm, Hematological Parameters, Reference Curve