Inflammation, anemia, and vitamin D levels associate with infant thrombocytosis risk

Abstract

Background

Thrombocytosis (>500 × 10³ platelets/μL blood) occurs in infants due to infection, inflammation, and/or anemia. Thrombocytosis and extreme thrombocytosis (EXT, >1000 × 10³ platelets/μL blood) can present diagnostic dilemmas, sometimes prompting invasive testing and anticoagulation therapy. We previously identified heightened thrombocytosis rates in hospitalized infants versus older children, but this population largely excluded extremely preterm infants at increased risk for infections and anemia—factors that promote thrombocytosis.

Objectives

To define thrombocytosis and EXT rates, etiologies, and sequelae among infants hospitalized in tertiary neonatal intensive care units (NICUs) to assist clinical decision making and determine factors that associate with thrombocytosis risk in preterm and full-term patients.

Methods

Retrospective analysis of thrombocytosis and EXT cases among 20,818 infants hospitalized in 2 tertiary NICUs from 2011 to 2023, compared to 10,323 patients hospitalized at a quaternary NICU.

Results

Our results revealed thrombocytosis in 3% of all patients (8% of preterm infants). Both estimates were significantly lower than the incidence of thrombocytosis in quaternary NICU patients (20%). EXT was also reduced in our tertiary unit (0.08% vs 0.5% in quaternary NICU). Thrombocytosis was associated with leukocytosis and relative anemia, but not with thrombotic or bleeding complications. Thyroid hormone, liver-derived thrombopoietin, and vitamin D deficiency can drive thrombocytosis in adults. Vitamin D level, but not thyroid hormone or liver function, was inversely correlated with platelet count among infants with thrombocytosis.

Conclusions

Inflammation, anemia, and vitamin D level correlate with infant thrombocytosis and EXT. Liver and thyroid immaturity do not appear to impact thrombocytosis risk. There were no thrombotic complications associated with EXT. These results provide important context for interpreting the origins and appropriate clinical responses to thrombocytosis in preterm infants.

Essentials

• •Infant thrombocytosis (high platelet count) can present diagnostic and therapeutic challenges.
• •We analyzed thrombocytosis in 20,818 hospitalized infants, including preterm and full-term infants.
• •Thrombocytosis affected 3% of infants (8% of preterm) without causing thrombosis or bleeding.
• •Inflammation, anemia, and vitamin D deficiency correlate with infant thrombocytosis risk.

1. Introduction

Thrombocytosis (>500 × 10³ platelets/μL blood) and extreme thrombocytosis (EXT, >1000 × 10³ platelets/μL blood) result from infection, inflammation, anemia, hemolysis, and/or impaired platelet clearance in pediatric patients [[1], [2], [3], [4]]. Primary thrombocytosis can also occur in the context of primary myeloproliferative disorders, but this is virtually nonexistent in pediatric patients [1]. High platelet counts can raise concern for thrombosis, historically prompting anticoagulation therapy and invasive diagnostic testing in some patients [1]. Defining etiologies and sequelae of thrombocytosis and EXT are diagnostically informative and critical to limit unnecessary tests and treatments. Understanding thrombocytosis etiologies can also help infer developmental differences in systems supporting megakaryocyte and platelet production (megakaryothrombopoiesis), which change in early life [5,6].

In previous studies, among pediatric patients at a quaternary hospital, EXT cases were frequently multifactorial, occurring secondary to inflammation, infection, iron deficiency, and/or splenic dysfunction [1,2]. EXT occurred more frequently in the context of critical illness, reflecting more severe pathophysiology. Interestingly, this condition was more common among infants aged <1 year, suggesting that altered megakaryothrombopoiesis in infants, compared with that in older children, could underlie these observations. Indeed, megakaryocytes are less polyploid in neonatal bone marrow compared with that of older children and adults [6]. Neonatal megakaryocytes also have increased proliferation, but each megakaryocyte produces fewer platelets than adult bone marrow megakaryocytes [6]. Additionally, infant platelets differ from those of older individuals [7,8], potentially reflecting developmental changes in the types of megakaryocytes that produce platelets [9,10].

Preterm infants born at <37 weeks gestational age are at risk for infection, nutritional deficiencies, anemia, and organ immaturity that could change thrombocytosis and/or EXT risks [11]. For example, preterm infants are at increased risks of infection and often experience prolonged periods of critical illness requiring respiratory and cardiovascular support. This clinical context should increase thrombocytosis and EXT in preterm infants, but prior studies of neonatal and pediatric thrombocytosis have largely neglected inborn infants hospitalized in tertiary neonatal intensive care units (NICUs) that routinely care for preterm infants [1,4].

We therefore sought to understand how our prior findings in a quaternary NICU population would translate to infants hospitalized in tertiary NICUs. The aims of this study were to clarify the prevalence, etiologies, and sequelae of thrombocytosis, including thrombotic complication risks in preterm infants, to inform diagnostic and therapeutic approaches.

2. Methods

2.1. EXT case identification

We retrospectively identified and profiled 20,818 patient charts for instances of thrombocytosis (>500 × 10³ platelets/μL) among all infants hospitalized in two level 3 (tertiary) NICUs from 2011 to 2023. These NICUs are connected to active labor and delivery wards and care for inborn patients, including many preterm infants. We analyzed all NICU patient data, from birth to discharge, without exclusion. NICU hospitalization can span a few days to several months depending on birth gestational ages and clinical complications.

We identified 1422 instances of thrombocytosis in blood counts from 713 patients. Statistical comparisons and trends were similar when analyzing all blood counts (ie, 1422 instances of thrombocytosis) or just the first instance of thrombocytosis for each patient. Thyroid hormone, albumin, and vitamin D data were collected within 7 days of a blood count that indicated thrombocytosis. We manually reviewed all cases of EXT and near-EXT (>900 × 10³/μL).

We compared tertiary NICU data with published data from 79,618 hospitalized pediatric patients [1], including 10,323 infants hospitalized in a level 4 (quaternary) NICU. This quaternary referral NICU cares for infants requiring critical care or subspecialty management that exceeds birth hospital capabilities. On average, quaternary NICU patients are older with more chronic illness than tertiary NICU patients. For comparisons involving thyroid, liver, and vitamin D, we omitted rare cases of incomplete or missing data (n = 6 from quaternary hospital). This study was approved by the University of Pennsylvania and Children’s Hospital of Philadelphia Institutional Review Boards.

2.2. Statistical analysis and plotting

We used R (v4.2.3) and GraphPad Prism (v10) for statistical analysis and graphical plotting. We calculated linear regression r² values using R and used Bonferroni correction to interpret significance.

3. Results

3.1. Thrombocytosis is rare among preterm infants in a tertiary NICU

In a retrospective analysis of 20,818 hospitalized infants at 2 NICUs, we identified 713 infants (3%) with thrombocytosis (>500 × 10³/μL) and 3 cases of EXT (>1000 × 10³/μL, Figure 1A). No infants were anticoagulated in response to EXT, and there were no thrombotic or bleeding complications directly linked to instances of EXT. The rates of thrombocytosis and EXT observed in this population were significantly less than those of pediatric patients hospitalized in a quaternary NICU [1] (Figure 1B, C). This was surprising, given high rates of sepsis, inflammation, and anemia (which drive pediatric thrombocytosis [1]) among preterm infants.

Figure 1.

Thrombocytosis and extreme thrombocytosis (EXT) rates among infants hospitalized at level 3 (tertiary) versus level 4 (quaternary) NICUs. (A) Blood counts and demographics for infants with thrombocytosis or EXT in our cohort of 20,818 total infants. Blood counts represent values at the time of thrombocytosis or EXT. (B) Thrombocytosis rates among infants hospitalized at our tertiary care unit, including a breakdown of those born preterm (<37 weeks’ gestation at birth). Reference comparison to infants in a level 4 NICU, which was higher than the thrombocytosis rate among all pediatric patients in this referral hospital (15% [1]). (C) EXT rates among infants hospitalized at our tertiary care unit, including a breakdown of those both preterm (<37 weeks’ gestation at birth). Reference comparison to infants in a level 4 NICU [1]. (D) Breakdown of thrombocytosis rates based on gestational age at birth, reflecting the degree of prematurity. HGB, hemoglobin; IQR, interquartile range; NICU, neonatal intensive care unit; PLT, platelet count; RBC, red blood cell count; SD, standard deviation; WBC, white blood cell count. ∗∗∗∗P < .0001.

Many premature infants experience critical illness during prolonged hospitalizations, leading us to hypothesize that increased thrombocytosis rates might be higher among extremely preterm infants and/or those with lowest birth weights. Indeed, thrombocytosis rates were increased among smaller and more premature infants (Figure 1B–D). However, even among the most premature subset, there were fewer EXT cases than expected based on our previous observations among infants in a quaternary NICU (χ² < 0.001 comparing tertiary vs quaternary patient incidence [1]).

3.2. Thrombocytosis among preterm infants reflects inflammation

Multifactorial etiologies often underlie thrombocytosis and EXT, including infection, inflammation, and anemia [1], as well as organ dysfunction related to platelet clearance [2]. To disentangle factors related to thrombocytosis, we evaluated clinical associations between platelet counts, blood traits, and other analytes (Figure 2A). Although there was considerable variability in white blood cell count (WBC), we noted positive associations between platelet counts and WBC (Figure 2B). This suggests concurrent thrombocytosis and leukocytosis in some patients, likely reflecting infection and/or inflammation.

Figure 2.

Platelet count is associated with altered WBC and vitamin D level, with thrombocytosis occurring most frequently at 3 to 4 weeks of life. (A) Heat map depicting r² values with platelet count (PLT) associations for the indicated traits among tertiary NICU patients. Traits with significant correlation, as defined by a line of best fit slope that deviates significantly from 0, are indicated with an asterisk [Pr(>|t|) < 0.05]. Red dots indicate positive correlation and blue dots indicate negative (inverse) correlation. A breakdown of leukocyte parameters is shown below the dashed line. Monocyte (Mono) and Lymphocyte (Lym) counts have significant associations with PLT count. Non-blood traits are shown below the solid line. (B) Scatterplot comparing platelet count versus WBC in infants with thrombocytosis. Line of best fit is shown in red with standard deviation in gray [Pr(>|t|) = 1.8 × 10⁻²]. (C) Histogram showing thrombocytosis cases as a function of postnatal age. The incidence of thrombocytosis peaks in the third week of life in this patient cohort (n = 1422 blood counts indicating thrombocytosis in 713 unique patients). (D) Scatterplot comparing platelet count versus vitamin D level in infants and children with thrombocytosis, including patients in tertiary and quaternary hospitals. Line of best fit is shown in red with standard deviation in gray [Pr(>|t|) = 2.9 × 10⁻¹¹]. Baso, basophils; Eo, eosinophils; HCT, hematocrit; HGB, hemoglobin; MCH, mean corpuscular hemoglobin; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; NICU, neonatal intensive care unit; RBC, red blood cell count; RDW, red cell distribution width; TSH, thyroid stimulating hormone; WBC, white blood cell count.

3.3. Thrombocytosis incidence peaks during the physiologic hemoglobin nadir

Interestingly, there was not a significant association between platelet count and red cell parameters (eg, hemoglobin) among thrombocytosis cases (Figure 2A). This may suggest that anemia was not a major contributing factor to the degree of thrombocytosis among this cohort, but prior transfusions, erythrocyte parameter variability, and/or subclinical iron deficiency confound interpretation.

We next assessed timing of thrombocytosis and EXT in our cohort. Thrombocytosis peaked ∼1 month of life, concurrent with the hemoglobin nadir from physiologic anemia and anemia of prematurity [12] (Figure 2C). Indeed, observed hemoglobin levels at the time of thrombocytosis revealed relative anemia (average hemoglobin, 11.3 ± 2.2) compared with typical hemoglobin levels healthy full-term neonates [13] (Table 1). This finding supports relative anemia and/or subclinical iron deficiency as contributing factors to thrombocytosis among preterm infants, as in other cohorts [1].

Table 1.

Blood counts and clinical context for EXT and near-EXT cases among infants in tertiary NICUs.

Case	Birth gestational age	Sex	Birthweight (kg)	Day of life	Counts at time of EXT
					PLT	WBC	HGB
(Normal range)					(150-400 ×103/μL)	(8.0-15.4 ×103/μL)	(12.5-20.5 g/dL)
1∗	37 wk 2 d	F	2.35	0	1124	26.4	9.3
2	32 wk 0 d	M	1.66	18	948	21	10.9
3	30 wk 6 d	F	1.29	38	1110	17.8	11.2
4	30 wk 4 d	F	0.85	25	904	12.1	8.7
5	28 wk 3 d	F	0.90	27	968	18.4	11
6	25 wk 6 d	F	0.79	25	1053	33.3	7.9

Given a paucity of EXT cases among infants in tertiary NICUs, we included infants with near-EXT (>900 × 10³/μL). Day of life refers to the first instance of EXT or near-EXT. In some cases, EXT was redemonstrated the subsequent day as well. One case (∗) occurred in an infant with a congenital hemolytic anemia, which can predispose to thrombocytosis [3]. PLT, platelet count (×10³/μL); WBC, white blood cell count (×10³/μL); HGB, hemoglobin (g/dL). Normal ranges are shown in parentheses.

EXT, extreme thrombocytosis; NICU, neonatal intensive care unit.

3.4. EXT occurs in multifactorial contexts among preterm infants

To further clarify clinical context for thrombocytosis, we profiled EXT cases in our cohort. Given the remarkable paucity of EXT cases among preterm infants, we analyzed all cases of thrombocytosis exceeding 900 × 10³ platelets/μL (Table 1). Review of these extreme cases highlighted concurrent leukocytosis and anemia in the setting of EXT, with acute infection or inflammation preceding development of EXT. Interestingly, one case also occurred around birth in an infant with a congenital hemolytic disorder, which we and others have previously found to drive thrombocytosis [3]. All cases recovered within a week without direct clinical intervention. None incurred thrombotic complications.

3.5. Vitamin D levels are inversely correlated with platelet count among infants with thrombocytosis

We wanted to ascertain why infants in tertiary NICUs had such low thrombocytosis rates compared with infants in a quaternary NICU (Figure 1). We hypothesized that organ immaturity and/or developmental differences in organ function may be responsible, given the progressive changes in platelet counts that occur in preterm versus term infants [14]. The liver produces thrombopoietin and acute phase reactants. Since these factors are not routinely measured in infants, we analyzed albumin levels as a proxy for liver function. Albumin levels were nearly identical among infants with thrombocytosis in our study cohort versus levels in quaternary hospital patients, suggesting intact liver function (Table 2).

Table 2.

Select clinical test results in temporal relation to thrombocytosis in patients hospitalized at tertiary or quaternary hospitals.

Test [normal range]	Tertiary NICU	Quaternary hospital	P value
Vitamin D [25-80 ng/mL]	53.2 ± 13.26 (5)	36.5 ± 21.4 (8114)	<.05
TSH [<15.0 μIU/mL]	7.02 ± 6.07 (108)	5.95 ± 6.72 (4418)
Thyroxine [>6 μg/dL]	14.22 ± 5.76 (143)	10.0 ± 4.3 (8720)	<.05
Albumin [3.5-5.1 g/dL]	3.45 ± 0.36 (62)	3.4 ± 0.8 (27563)

Statistical comparisons between populations were made by two-sided Welch’s t test. Significant differences are indicated by P values. All parameters were associated with platelet count by linear regression in patients hospitalized in a quaternary hospital (P < .05) but were not significantly correlated with platelet counts in tertiary NICU patients. All values reflect mean ± standard deviation (n unique patients).

NICU, neonatal intensive care unit; TSH, thyroid stimulating hormone.

We then analyzed thyroid hormone levels, since hypothyroxinemia is common among preterm infants [15] and thyroid hormone levels are correlated with adult platelet counts [16]. Levels of thyroid stimulating hormone and thyroxine were within normal range in our cohort and were actually elevated in tertiary NICU patients compared to quaternary hospital patients (Table 2). Thus, thyroid hormone deficiency cannot explain the paucity of thrombocytosis in our cohort.

We next analyzed vitamin D levels in NICU patients with thrombocytosis, given links between vitamin D deficiency and thrombocytosis in adults [17,18]. Vitamin D level and platelet count are inversely correlated in adults, but this trend has not been established in pediatric patients. We analyzed vitamin D levels obtained within 7 days of thrombocytosis occurrence and identified an inverse correlation between vitamin D level and platelet count (Figure 2A and Table 2). We also noted that total vitamin D levels were increased in tertiary NICU vs quaternary patients (53.2 ± 13.3 vs 36.5 ± 21.4 ng/mL, P < .05, Table 2). Higher vitamin D level would be expected to blunt inflammatory responses and thrombocytosis [19,20], despite clinical circumstances that should otherwise drive higher levels of thrombocytosis [1].

4. Discussion

Preterm infant nutrition is closely monitored in NICUs, including vitamin D supplementation. Poor gastrointestinal absorption, medication interactions, or medical complexity may result in lower vitamin D levels among older, more chronically ill infants in our quaternary hospital. Although there is a temporal association between vitamin D and thrombocytosis, we note that vitamin D levels are infrequently measured in infants. Sparse data allows potential bias, although our analysis did comprise >8000 patients. Future targeted evaluation is needed to provide causal links between vitamin D level and thrombocytosis in pediatric patients.

This retrospective study analyzed a large volume of patients over 13 years. Pediatric thrombocytosis occurs most often as a secondary response to an inciting etiology [1]. Among infants hospitalized at our tertiary care NICU, thrombocytosis was associated with a multifactorial combination of infection, inflammation, and relative anemia. Although thrombocytosis and EXT can cause significant clinical concerns, EXT did not cause thrombotic complication and resolved without intervention in all cases in the cohort examined in this study. Our findings show that any degree of thrombocytosis (eg, >500-600 × 10³ platelets/μL) should warrant consideration for an underlying infectious or inflammatory process and that EXT (>1000 × 10³ platelets/μL) may not be physiologically supported in some premature infants.

Although the etiologies and sequelae of thrombocytosis and EXT matched those of other pediatric patients [1], the paucity of thrombocytosis and EXT cases among tertiary NICU patients was surprising (Figure 1). Lower rates of EXT were particularly notable among preterm infants born extremely prematurely or with extremely low birth weight, who are at heightened risk for infection, anemia, and critical illness—factors linked to thrombocytosis and EXT [1,11]. These findings are not overtly linked to liver or thyroid immaturity or developmental differences, although we could not exclude hepatic secretion of key platelet-related cytokines given a lack of routine clinical testing (eg, interleukin-6, thrombopoietin).

Megakaryocyte and platelet production differ in neonates compared with older children and adults, and these differences likely underlie increased rates of thrombocytosis and EXT among infants in some pediatric populations [1]. Although we note that the relatively weak associations between WBC and vitamin D levels with platelet count are of unclear clinical significance, our findings highlight the importance of exogenous systemic factors in determining platelet count that may open exciting avenues for future study. The effects of vitamin D and/or other downstream systemic factors (eg, inflammatory responses) may be harnessed to augment platelet production in vitro or in vivo to support novel cell therapeutic development and used to help understand processes that support megakaryothrombopoiesis in neonates and infants.

We do not anticipate that vitamin D supplementation and relatively reduced propensity for thrombocytosis would have negative health effects. Immunologic roles for platelets are well established, including direct binding of bacteria, immune cell recruitment, and inflammatory cytokine release from activated platelets [8,[21], [22], [23]]. To our knowledge, extremely high platelet levels (EXT) have not been shown superior to modest thrombocytosis (500-1000 × 10³ platelets/μL) in terms of infection clearance or other inflammatory sequelae. Although platelets differ between newborns and adults [8], substantial changes occur during the first weeks of life that “normalize” the hemostatic and immunologic environments [6,24,25]. Most thrombocytosis occurred after this transition in our cohort (∼3 weeks of life, Figure 2C). These findings reinforce the importance of optimal nutrition to support normal organ function in preterm infants, in addition to rigorous assessment in the setting of thrombocytosis to determine clinical management.