Complete blood count reference intervals for extremely preterm neonates
Christina H. Wolfsberger, Benedikt N. Seidel, Linda Fleck, Markus Herrmann, Diether Kramer, Martin Benesch, Gerhard Pichler, Markus G. Seidel
TL;DR
This study provides detailed blood count reference ranges for preterm babies, helping doctors better interpret their blood test results and identify health issues.
Contribution
The study introduces new, age-stratified blood count reference intervals for preterm neonates and highlights differences between appropriate and small-for-gestational-age infants.
Findings
Reference intervals for blood parameters were established for preterm neonates across gestational and postnatal ages.
Significant differences were found between AGA and SGA infants for leukocytes, thrombocytes, and neutrophilic granulocytes.
Hemoglobin and hematocrit levels were largely similar between AGA and SGA infants, with some exceptions at specific gestational ages.
Abstract
Preterm infants exhibit dynamic hematological profiles during the first postnatal days, necessitating robust reference intervals for clinical interpretation. The aim of this study is to establish comprehensive reference intervals for complete blood counts in preterm neonates during the first five postnatal days, stratified by gestational age and postnatal age. Data from preterm neonates (23 + 0–34 + 0 weeks gestation) admitted to the neonatal intensive care unit Graz (2008–2023) were analyzed. Percentiles (2.5th, 10th, 50th, 90th, 97.5th) were generated for the following hematological parameters included in the full/complete and differential blood count (FBC, CBC): leukocytes, thrombocytes, hemoglobin, hematocrit, neutrophilic granulocytes, and lymphocytes. Stringent exclusion criteria minimized confounding factors. Comparisons were made between appropriate-for-gestational age (AGA) and…
Genes, proteins, chemicals, diseases, species, mutations and cell lines named across the full text — each resolved to its canonical identifier and authoritative record.
Click any figure to enlarge with its caption.
Figure 1
Figure 2- —Medical University of Graz
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsNeonatal Respiratory Health Research · Pregnancy and preeclampsia studies · Neonatal and fetal brain pathology
Introduction
Prematurity, defined as birth before 37 weeks of gestational age, is associated with various physiological and pathophysiological differences compared to term born neonates. Besides morbidities such as bronchopulmonary dysplasia, retinopathy of prematurity, and intraventricular hemorrhage [1], preterm infants also exhibit variations in blood counts and immune function compared to term neonates [2].
Hematopoiesis in preterm neonates differs significantly due to the ongoing maturation of the hematopoietic system after birth. During (embryo- and) fetogenesis, hematopoiesis primarily occurs in the liver and spleen, gradually shifting to the bone marrow as the main site of blood cell production [3]. This transition is often incomplete in preterm neonates, leading to lower blood cell counts. Red blood cell, leukocyte, and thrombocyte production is influenced by gestational age, postnatal age, and medical conditions. Understanding these characteristics is crucial for clinical management and optimizing outcomes [3, 4].
Hematopoiesis and immune development continue postnatally in preterm neonates leading to dynamic changes during the first weeks [2]. Hemoglobin levels decline after birth, due to a physiological transient decrease in erythropoietin, especially in preterm neonates, resulting in reduced regenerative capacity [4]. Platelet counts can be low (thrombocytopenia), though clinically significant thrombocytopenia is less common. Gestational and postnatal age, along with comorbidities, influence platelet counts [5, 6]. Leukocyte counts fluctuate, initially rising then declining [7]. Lymphocyte counts are generally lower in preterm neonates, reflecting a immune immaturity [8], with both number and function of T lymphocytes impaired. This contributes to a higher susceptibility of preterm neonates to infections and underscores the need for protective measures in their care.
Although changes in blood count values in preterm neonates have been reported [9], detailed information on specific components—leukocytes, thrombocytes, hemoglobin, hematocrit, neutrophilic granulocytes, and lymphocytes—in relation to gestational/postnatal age, and intrauterine growth restriction (IUGR) or small-for-gestational age (SGA) status, are still lacking.
This study aimed to define reference intervals for blood cell counts in neonates born before 34 weeks of gestation during the first 5 days after birth, stratified by gestational and postnatal age. We hypothesized that normative values would vary with gestational/postnatal age and differ between neonates born appropriate-for-gestational age (AGA) and SGA/IUGR neonates.
Methods
Study design
The study was designed as a retrospective single-center study, conducted at the Department of Pediatrics and Adolescent Medicine, Medical University of Graz. Data of preterm neonates born between January 2008 and December 2023 were eligible for analysis. The study was approved by Regional Committee on Biomedical Research Ethic of the Medical University of Graz (EC number: 36–215 ex 23/24) and was carried out in accordance with The Code of Ethics of the World Medical Association (Declaration of Helsinki) [10].
Population
Preterm neonates with a gestational age of 34 + 0 weeks or less admitted to the neonatal intensive care unit in Graz were included in this study. Blood samples collected within the first 7 days after birth were analyzed. Exclusion factors are presented in Online Supplement Table 1.
Data analysis
The antepartum medical history (maternal peripartal infection, preterm premature rupture of membranes, IUGR, amniotic fluid infection, gestosis]), antenatal steroids, fetal infections (fetal cytomegalovirus, fetal toxoplasmosis infection, fetal parvovirus B19 infection), fetal proven inborn error of immunity, fetal genetic abnormality, positive family history of severe combined immunodeficiency (SCID), and demographic data (gestational age, gender birth weight, length, diagnosis of SGA [defined as birth weight < 10th centile] or IUGR, Apgar score, umbilical cord pH, mode of delivery and causes of preterm birth) were documented. Received medication during the first 7 days after birth at the NICU, including antibiotics, immunoglobulins, transfusion of platelets/red blood cells, G-CSF, anticonvulsants, steroids, non-steroidal antirheumatic drugs including patent ductus arteriosus-related treatments were documented. Short-term outcome parameters, defined as any diagnosis at 37 to 42 weeks of gestational age or time point of discharge, were assessed. Further, any diagnosis of SCID, infection with opportunistic germs, or of a hematological disease including leukemia/lymphoma or autoimmunocytopenia until an age of corrected 5 years were noted.
Routinely performed blood samples of the preterm neonate were documented within the first 5 days after birth, focusing on leukocytes, thrombocytes, hemoglobin, hematocrit, absolute number of neutrophilic granulocytes, and absolute number of lymphocytes. The documentation included the method of blood sampling, and specifying whether it was an arterial, venous, or capillary sample. All blood sample were measured with a Sysmex XN (Sysmex Austria, Vienna) analyser in our ISO 15189 accredited central laboratory. Intra- and interassay imprecision for all analyses and levels were < 4.7% (1.2% to 4.7%).
Statistical analysis
The statistical analysis was conducted using Python Version 3.10 (Python Software Foundation 2024, Welmington, DE, USA) and SPSS 29.0 (SPSS, Armonk, NY, USA). Percentile calculations were performed with the numpy.percentile() function from NumPy [11], while statistical significance was evaluated using the Kolmogorov–Smirnov test implemented in scipy.stats.kstest() from SciPy [12].
Percentile calculations
Percentile curves for each hematological parameter (leukocytes, thrombocytes, hemoglobin, hematocrit, neutrophilic granulocytes, and lymphocytes) were constructed for the first five postnatal days. Data were stratified by gestational age, with separate calculations for each week from 23 + 0 to 23 + 6 weeks, 24 + 0 to 24 + 6 weeks, and so on, up to 34 + 0 weeks. For each gestational age group and postnatal day, the 10th to 90th percentiles were calculated. These percentiles represent the distribution of values within each group, providing a comprehensive overview of the expected intervals of hematological parameters in preterm neonates during the first week after birth. For the analysis, capillary and arterial blood parameter values were kept together to create the percentiles. As a baseline for future consultation of the data table, we wanted to reduce complexity as much as possible, while also keeping the number of patients per percentile point maximal.
In a sub-analysis, reference intervals (2.5th to 97.5th percentiles) were calculated for the combined gestational age groups of 23 + 0 to 26 + 6 weeks, 27 + 0 to 30 + 6 weeks, and 31 + 0 to 34 + 0 weeks to enhance clinical applicability in a larger cohort with more robust numbers.
In a further sub-analysis, reference intervals (2.5th to 97.5th percentiles) were calculated separately for males and females for the combined gestational age groups of 23 + 0 to 26 + 6 weeks, 27 + 0 to 30 + 6 weeks, and 31 + 0 to 34 + 0 weeks.
Comparison of hematological parameters in appropriate-for-gestational age and small-for-gestational age neonates
The impact of fetal growth on hematological parameters was investigated by comparing preterm neonates born AGA with those born SGA. Preterm neonates with a birth weight below the 10th centile and/or a fetal diagnosis of IUGR were categorized as SGA neonates. For each gestational age group (23 + 0 to 23 + 6 weeks through 34 + 0 weeks) and each hematological parameter, we compared values obtained between 20 and 36 h after birth. There is a known significant difference between arterial and capillary values [13] that could also be found in our overall data. To accommodate for this discrepancy while also keeping the dataset as large as possible for statistical analysis, we calculated the ratio of arterial and capillary datapoints per week, per blood parameter. The Kolmogorov–Smirnov test was used to assess differences in the distributions of these parameters between the AGA and SGA groups. A p value of less than 0.05 was considered statistically significant.
Results
A total of 3288 preterm neonates ≤ 34 + 0 were born between January 2008 and December 2023 and admitted to the NICU Graz. One-hundred-sixty neonates were excluded because they met one or more of the exclusion criteria listed in Online Supplement Table 1 (Online Supplement Fig. 1, Flow chart). The final study population consisted of 3128 neonates with a median (minimum; maximum) gestational age of 32.1 (23.1;34.0) weeks and a birth weight of 1700 (297; 3210) grams. Demographic data, including causes for preterm birth and short-term outcome parameters, are displayed in Table 1 and Online Supplement Table 2. Specific conditions requiring day-based exclusions related to infection, surgery, medication use, transfusions, or intraventricular hemorrhage. The exact number of included neonates for each parameter for each gestational age is displayed in Figs. 1 and 2 and Table 2. Table 1. Demographic data and outcome parameters of included preterm neonates. Values are presented as n (%), or median (minimum; maximum) as appropriateEntire cohort**n = 3128AGA preterm neonatesn = 2792SGA preterm neonatesn = 328**Gestational age (weeks)32.1 (23.1; 34.0)32.1 (23.1; 33.9)32.6 (23.4; 34.0)Birth weight (g)1700 (297; 3210)1750 (376; 3210)1240 (297; 1780)*Female sex n (%)*1407 (44.8%)1244 (44.6%)153 (46.6%)*Spontaneous vaginal delivery n (%)*710 (22.7%)686 (24.6%)21 (6.4%)Apgar 58 (2; 9)8 (2; 9)8 (1; 8)Umbilical cord pH7.31 (7.01; 7.60)7.31 (7.01; 7.60)7.29 (7.05; 7.44)*Prenatal steroids n (%)*2099 (67.1%)1886 (67.6%)208 (63.4%)*Causes for preterm birthPreterm rupture of membranes n (%)*448 (14.3%)431 (15.4%)16 (4.9%)*Amniotic fluid Infection n (%)*243 (7.8%)234 (8.4%)9 (2.7%)*Gestosis n (%)*377 (12.1%)304 (10.9%)73 (22.3%)*Cervical insufficiency n (%)*152 (4.9%)139 (5.0%)12 (3.7%)*Pathological CTG n (%)*428 (13.7%)318 (11.4%)110 (33.5%)*Preterm labor n (%)*983 (31.4%)933 (33.4%)46 (14.0%)Others n (%)497 (15.9%)433 (15.5%)62 (18.9%)AGA appropriate-for-gestational age, CTG cardiotocography, SGA small-for-gestational ageFig. 1Percentile curves (10th, 25th, 50th, 75th, 90th centile) for preterm neonates born between 28 + 0 and 28 + 6 weeks of gestation for each hematological parameter (leukocytes, thrombocytes, hemoglobin, hematocrit, neutrophilic granulocytes, and lymphocytes) during the first five postnatal daysFig. 2Comparison of hematological parameters (leukocytes, hemoglobin, hematocrit, thrombocytes, neutrophilic granulocytes and lymphocytes) in appropriate-for-gestational age (AGA) and small-for-gestational age (SGA) neonates during the first five days after birth. Black asterisks = significant difference between AGA and SGA, difference in proportion of arterial and capillary sampling technique between < 10%; blue asterisks = significant difference between AGA and SGA, difference in proportion of arterial and capillary sampling technique between 10 and 20%; magenta asterisks = significant difference between AGA and SGA, difference in proportion of arterial and capillary sampling technique between 20 and 30%; red asterisks = significant difference between AGA and SGA, difference in proportion of arterial and capillary sampling technique between 30 and 50%Table 210th, 25th, 50th, 75th, and 90th percentiles of hematological parameters (leukocytes, thrombocytes, hemoglobin, hematocrit, neutrophilic granulocytes, lymphocytes) for each gestational age (23 + 0 to 34 + 0 weeks) on the first day after birth (comparable to day 1 in Fig. 1). For certain parameters, including lymphocyte counts and neutrophilic granulocytes at the lowest gestational ages, the number of available samples was limited. These percentiles should therefore be interpreted with cautionGestational agen10th25th50th75th90thn10th25th50th75th90thLeukocytes (cells/µL)****Thrombocytes (cells/µL)23305023894310,76019,43323,02336128,000139,750177,000203,000302,00024746028848012,94021,71025,96185124,000155,000212,000264,000315,800256052997405968514,93822,62367121,400154,000190,000253,000286,60026964860667510,48514,26020,155105105,500141,000187,000246,000280,800271014670741010,10013,20518,45010795,600124,250183,000227,500285,200281515540726010,04014,14018,110160104,900136,000191,000230,000308,3002916256106989943512,15015,677168110,700138,750186,500227,250282,300302055836802010,56014,68017,410212133,100153,750193,500242,250287,800313076666900511,52014,36017,474310111,800153,250191,000238,750284,100325077876993512,84015,77518,760518132,400166,000210,000252,000292,3003369187001079013,51016,80520,270702127,000165,250211,500256,000297,00034175918411,19014,42017,39020,494178122,000168,500212,500255,000304,300Hemoglobin (g/dL)****Hematocrit (%)233612.813.915.016.818.33637.140.543.148.950.4248512.813.815.217.318.48537.439.443.048.452.2256712.513.715.016.318.46735.039.742.346.550.62610512.914.516.017.319.310537.640.944.649.254.32710714.415.817.318.820.410740.944.249.452.056.82816013.915.817.519.420.516039.544.348.953.457.32916814.716.117.619.221.216840.644.949.154.458.83021214.816.417.919.520.721241.445.649.954.057.53131015.416.818.620.121.231042.946.651.655.159.53251814.716.718.219.821.351840.845.750.054.758.23370215.116.618.319.921.470241.545.850.154.558.93417814.816.318.019.721.017840.844.749.053.957.5Neutrophilic granulocytes (cells/µL)****Lymphocytes (cells/µL)**231129004950750014,70016,7001110002100220030003500242419204175750015,95020,3102413302125280040006290251517402700540011,85018,50015152022252700325037602624122025504050660015,180248451538220033005170272733704500550010,50017,70027152019502700450053602838209539005000667511,29038144022002700370051602948260031754375672515,61048144017752500342553303053250032004300920012,4205317202100270036004260319826704125605075509430971600220030003800454032114310045756200857510,880114193026003100390047703317440305250760010,15012,30017022002825350044755310344142006200780011,40014,8003925602950360041504940
Percentile calculations
Percentile curves, representing the 10th to 90th percentiles, were constructed for each hematological parameter (leukocytes, thrombocytes, hemoglobin, hematocrit, neutrophilic granulocytes, and lymphocytes) during the first five postnatal days. Data were stratified by gestational age in weekly increments, ranging from 23 + 0 to 23 + 6 weeks up to 34 + 0 weeks. To present one example, Fig. 1 displays the percentile curves for preterm neonates born between 28 + 0 and 28 + 6 weeks of gestation for each hematological parameter during the first five postnatal days. Table 2 shows the 10th to 90th percentiles for each parameter and gestational week on the first day after birth. Percentile curves for the other gestational ages are presented in Supplemental A. Although fluctuations in individual parameters were observed across different gestational weeks and postnatal ages, a general trend of increasing values with increasing gestational age was noted.
Supplemental B provides combined reference intervals (2.5 to 97.5 percentile) for hematological parameters, stratified by gestational age at birth (23 + 0 to 26 + 6 weeks, 27 + 0 to 30 + 6 weeks, and 31 + 0 to 34 + 0 weeks). Supplemental C offers the same reference intervals further stratified by sex.
Comparison of hematological parameters in appropriate-for-gestational age and small-for-gestational age neonates
A total of 2792 preterm neonates were included in the AGA cohort and 328 neonates in the SGA cohort. The mean gestational age and birth weight of neonates in the AGA cohort was median (minimum; maximum) 32.1 (23.1;33.9) weeks of gestation and 1750 (376; 3210) grams, respectively. In the SGA cohort, the median (minimum, maximum) gestational age and birth weight was 32.6 (23.4; 34.0) weeks of gestation and 1240 (297; 1780) grams, respectively. For each gestational age group (23 + 0–23 + 6 weeks through 34 + 0 weeks) and each hematological parameter, values obtained on the first day after birth were compared between the two groups (Fig. 2). Preterm neonates with IUGR/SGA demonstrate distinct hematological profiles compared to AGA preterm neonates, particularly in terms of reduced leukocytes, neutrophilic granulocytes, and thrombocytes, and a trend towards increased hematocrit and hemoglobin. These differences were most consistently observed from 31 gestational weeks onwards, but in part, already seen in younger gestational ages. To exclude a potential bias from varying proportions of capillary versus arterial-derived blood samples, we carefully noted this variation (see color-coded asterisks in Fig. 2).
Discussion
This study, encompassing a large sample size of very preterm neonates admitted to the NICU Graz between 2008 and 2023, provides robust reference intervals for hematological parameters during the first week after birth. Stringent exclusion criteria ensured a homogenous study population, minimizing the influence of confounders and allowing for a focused analysis of gestational age and postnatal age on hematological parameters. The comprehensive data set, stratified by gestational age in weekly increments, enables clinicians to interpret blood test results with greater precision and identify potential deviations from the established norms. Furthermore, the comparison between SGA and AGA neonates offers valuable insights into the impact of fetal growth on hematological development.
Normal values of complete blood counts in extremely preterm neonates in our cohort differ from previously published data.
Leukocyte counts (50th centile) in our current cohort were lower across all gestational age categories compared to previously reported data [14–17]. In our cohort, the 50th percentile of leukocytes ranged from 7185/µL within the first 24 h after birth in neonates with a gestational age of 25 + 0 to 25 + 6 weeks, to 13,780/µL in neonates with a gestational age of 34 + 0 weeks. In contrast, Scheffer-Mendoza et al. [14] reported higher leukocyte counts, with a median of 17,050/µL at a comparable postnatal age. However, this study [14] focused on full-term neonates delivered vaginally. Overall, these observed differences between preexisting literature and our present study may be attributed to the larger sample size, more detailed stratification by gestational age and postnatal age, and the exclusion of confounding factors in our study. Furthermore, the main difference between our study and the data obtained by, e.g., Scheffer-Mendoza et al. [14] is the gestational age range, as we included preterm neonates from 23 to 34 weeks of gestational age, while Scheffer-Mendoza et al. focused on term neonates. The higher leukocyte values observed in our cohort with more advanced gestational age also highlight the importance of considering gestational age differences. Beside the above-described differences, our findings are largely consistent with the trend of increasing leukocyte counts observed with increasing gestational age, as demonstrated by Wu et al. [18]. Although the absolute values for gestational age categories differed, our study corroborates the general pattern of rising leukocyte counts as gestational age advances. Specifically, the most immature cohort (less than 25 weeks by Wu et al. [18]; 23 + 0 to 23 + 6 weeks in our cohort) exhibited higher leukocyte counts compared to the following week. After this initial peak, a decrease in leukocyte counts was observed, reaching a low point at 29 + 0 to 29 + 6 weeks (Wu et al. [18] and our cohort). Thereafter, a further increase in leukocyte values was seen, with the highest levels occurring at more advanced gestational ages, as mentioned previously.
Comparable to leukocyte counts, lymphocyte values in our cohort are also lower compared to previously reported literature [19]. In our study, the 50th percentile ranged from 2400 lymphocytes/µL in neonates born 23 + 0 to 23 + 6 weeks of gestational age to 4050 lymphocytes/µL, with a gestational age of 34 + 0 weeks. As some gestational age categories included only a small number of neonates, caution is warranted when interpreting these results. Differences were observed when comparing our findings to those of Christensen et al. [19], who reported lymphocytes during the first 3 h after birth, ranging from an estimated 7000/µL in neonates with a gestational age of 23 weeks to 6400/µL in those born at 34 weeks. This discrepancy may be attributed to differences in postnatal age and exclusion criteria. However, our lymphocyte counts were more comparable to the findings of Amatuni et al. [8], who reported counts ranging from 2500/µL in neonates with a gestational age of 22 to 28 weeks to 3450/µL in those born between 32 and 36 weeks.
Our study revealed an initial rise in neutrophilic granulocytes during the first two postnatal days, followed by a decline. This pattern aligns with existing literature, though differences in gestational age must be considered. Prior studies, such as one with a mean gestational age of 38.9 ± 2.4 weeks [20], reported higher absolute values than those in our preterm cohort. For instance, the 25th percentile of neutrophilic granulocytes within the first 24 h in neonates at 34 + 0 weeks was 4560/μL in our data, compared to 6300/μL in term neonates reported by Ianni et al. [17]. This difference was even more notable at the 50th percentile. A potential explanation for this discrepancy is that the study by Ianni et al. included neonates with a mean gestational age of 39.6 ± 1.0 weeks, whereas our analysis focused on extremely and very preterm neonates. Similar trends were observed by Henry and Christensen [9], though their broader inclusion criteria may affect comparability. Despite these differences, our absolute values remain within previously published ranges. A notable similarity with past studies is the low 10th percentile of thrombocyte counts across gestational ages. The lowest 10th percentile value within 24 h post-birth was 85,800/μL in neonates at 29 + 0 to 29 + 6 weeks, comparable to values reported by Henry and Christensen [9]. These counts are well below the adult threshold for thrombocytopenia (150,000/μL) [21], highlighting the need for gestational age-specific reference intervals. While thrombocytopenia may indicate conditions like perinatal asphyxia, sepsis, or NEC [22], our cohort excluded such cases. This strengthens the argument that transient thrombocytopenia may be typical in very preterm neonates, or that current clinical definitions require reassessment.
Hemoglobin and hematocrit values in our cohort of preterm neonates were slightly higher than in term neonates, consistent with previous studies [17]. This is likely due to lower plasma volume and more concentrated erythrocyte levels in preterm neonates [2]. Our results for neonates with a gestational age of 33–35 weeks aligned with Melioli et al. [15], highlighting changes in erythropoiesis that accompany preterm maturation. Hemoglobin remained relatively stable with postnatal age, while hematocrit peaked within the first 24 h across all gestational ages. In neonatal intensive care, red blood transfusions or erythropoietin are commonly used to maintain appropriate hemoglobin and hematocrit levels in preterm infants to prevent anemia. To reduce bias, we excluded factors influencing these values, such as intraventricular hemorrhage, major blood loss, surgery, and transfusion. These exclusions were essential to establish reliable reference intervals that accurately reflect typical hematological profiles in preterm neonates.
Preterm neonates with IUGR/SGA showed notable differences in hematological parameters compared to AGA peers, particularly in leukocytes, thrombocytes, and neutrophilic granulocytes. As Özyürek et al. [23] reported, SGA infants had significantly reduced leukocyte and neutrophil counts through the first week after birth. Similarly, our cohort showed lower thrombocyte counts in SGA neonates—except for those born at 23 + 0 to 23 + 6 weeks. We attribute these reductions to chronic intrauterine stress in SGA cases [24].
Contrary to our expectation that hemoglobin and hematocrit levels might be significantly lower in IUGR/SGA infants due to placental insufficiency and reduced nutrient supply [25], our cohort showed no significant differences between SGA and AGA neonates in these parameters, except for those with a gestational age of 31 + 0 to – 34 + 0 weeks. Lymphocyte differences were only observed in neonates born at 27 + 0 to 27 + 6, 28 + 0 to 28 + 6, 29 + 0 to 29 + 6, and 32 + 0 to 32 + 6 weeks, with lower values in SGA neonates. Özyürek et al. [23] found no differences on day one but did observe them by day seven. Their inclusion of full-term neonates may explain this discrepancy. Our lymphocyte findings should be interpreted with caution due to limited sample size.
Strength and limitation
This study benefits from a large sample size, with 3175 preterm neonates being eligible for analysis. This allowed us to establish robust reference intervals for hematologic parameters in preterm neonates during the first postnatal week, stratified by gestational and postnatal age. Our strict exclusion criteria led to a relatively homogeneous study population, minimizing potential confounders. Some limitations must be considered. One is the combined use of capillary and arterial blood samples for percentile calculations. A comparison between the two methods was not feasible due to the limited and unevenly distributed data in each group (blood value, gestational age, postnatal age) inherent to the retrospective study design. Figure 2 illustrates the representation of each sampling method by group. Differences between arterial and capillary blood sampling methods have been described in the literature [26]. Capillary and arterial blood differ physiologically and in several pre-analytical aspects (collection technique, sample volume, anticoagulant ratio, risk of hemolysis, tissue fluid contamination), all of which may affect hematologic measurements [27]. These differences are particularly relevant for hemoglobin/hematocrit, leukocyte counts, and platelet values. While reflecting clinical reality—where both sample types depends on patient circumstances—this introduces potential variability. In practice, some neonates have arterial, others capillary values, and it is not feasible to use separate percentiles for each variable. While combining both sources increased sample size and enhanced feasibility, it may also introduce variability in the derived reference intervals. Clinicians should therefore apply these intervals with awareness of sample origin, and future well-powered studies should aim to establish separate, standardized reference ranges for capillary and arterial blood in neonates. The second limitation is the retrospective nature of data collection. Since the data were not prospectively collected and controlled, this may introduce biases and unaccounted confounding factors. It also limits causal interpretations and may affect generalizability. Another limitation is the small sample size in selected gestational subgroups. This particularly affects lymphocyte counts and neutrophilic granulocytes at very low gestational ages, where the robustness of percentiles is weaker. Clinicians should therefore interpret values in these groups with caution. Nonetheless, the large overall sample size and strict exclusion criteria help mitigate some inherent limitations of a retrospective approach. Our findings also emphasize the need for future work comparing standardized reference intervals between arterial and capillary blood samples. To date, no large-scale neonatal study has systematically contrasted these sources. Our data demonstrate feasibility, but further studies with adequate power are needed to generate blood source–specific intervals that can guide clinical care.
Conclusion
This study provides comprehensive reference intervals for hematological parameters in preterm neonates during the first 5 days after birth, stratified by gestational age and postnatal age.
Our study addresses several previous gaps: prior reference intervals rarely stratified by gestational week, particularly in extremely preterm infants—our data provide week-specific percentiles. Earlier studies generally did not include both arterial and capillary samples—we provide both, while acknowledging variability between them. Moreover, reliable percentile estimates at very low gestational ages were previously lacking; our data now fill this gap, though caution is needed for parameters with limited numbers. By addressing these limitations, our study offers new insights and contributes original reference values for neonatal hematology.
These normative values can guide clinicians in interpreting blood test results, identifying abnormalities, and monitoring the trajectory of hematological parameters in this vulnerable patient population. The observed differences between AGA and SGA preterm neonates warrant further investigation to understand the potential clinical implications of fetal growth restriction on neonatal hematology. The established reference intervals offer valuable clinical utility for assessing hematological status, enabling early identification of potential deviations and informing appropriate interventions. Future research incorporating longitudinal follow-up could elucidate the long-term implications of these early hematological variations and their impact on later health outcomes. Prospective studies controlling for potential confounding factors could strengthen the current findings and enhance their generalizability, ultimately improving clinical decision-making and patient care.
Supplementary Information
Below is the link to the electronic supplementary material.Supplementary file1 (DOCX 518 KB)Supplementary file2 (DOCX 16 KB)Supplementary file3 (DOCX 21.1 KB)Supplementary file4 (DOCX 80.0 KB)Supplementary file5 (DOCX 31 KB)Supplementary file6 (DOCX 29 KB)