Platelet count and serum thrombopoietin level as predictors for morbidity and/or mortality in thrombocytopenic neonates

Abstract

Background

Serum thrombopoietin in thrombocytopenic infants is largely related to the cause of thrombocytopenia and the underlying disease. Many perinatal factors can affect thrombopoietin level.

Patients and methods

A prospective cross-sectional study on 119 thrombocytopenic neonates: 54 full term and 65 preterm had been conducted. Thrombopoietin assay was done using a qualitative enzyme-linked immunosorbent assay technique. The test was repeated on the change of clinical status (recovery or deterioration).

Results

Lowering of thrombopoietin level was noted on reversal of platelet count to normal (P<0.001). Survival is significantly related to platelet count in full term (P = 0.04), but insignificant among thrombocytopenic preterms. Platelet count is negatively correlated to thrombopoietin level in neonates both in full term and preterm (r = −0.59, −0.69, respectively, P<0.001). Platelet count was found to be the best predictor for duration of recovery of thrombocytopenia in neonates compared with other factors including thrombopoietin level.

Conclusion

Thrombocytopenic neonates had high levels of thrombopoietin. Despite the high thrombopoietin level in neonates died with severe thrombocytopenia, yet, mortality is related to the cause and outcome of thrombocytopenia rather than the serum thrombopoietin level. It is recommended to diagnose and treat the underlying cause of thrombocytopenia rather than to generalize the therapy based on thrombopoietin level.

Keywords:

• Thrombopoietin
• Thrombocytopenia
• Neonates

Introduction

Platelet count in the healthy human fetus reaches adult level of 150–450 × 10⁹/l by the second trimester of pregnancy.¹ However, thrombocytopenia defined as platelet count <150 × 10⁹/l, is the commonest hematological abnormality occurring in 22–35% of all babies admitted to neonatal intensive care units (NICU) reaching to 72% in sick preterm babies.² This high incidence of neonatal thrombocytopenia in preterm babies together with its association with the more severe grades of intracranial hemorrhage makes it one of the most important hematological problems encountered in newborns.³

Treatment of neonatal thrombocytopenia is dependent mainly on repeated platelet transfusions which is associated with several adverse reactions like refractoriness to platelet transfusions, immunological reactions, and volume overload which is a major cause of intracranial bleeding.⁴ Isolation and cloning of thrombopoietin (Tpo) to be used therapeutically were lastly proposed to be the physiological solution of neonatal thrombocytopenia.^5,6 Tpo is proved to promote the proliferation and differentiation of megakaryocytic lineage and can be used to differentiate megakaryocytic and amegakaryocytic thrombocytopenia.⁷

Very limited information is currently available about Tpo levels in healthy term, preterm, or small for gestational age (SGA) babies. In addition, changes in Tpo level in response to thrombocytopenia in the fetus and newborn are not well studied. The available studies have many limitations:^8,9

1.	Normal levels of Tpo in the healthy newborn infants at different gestational ages and levels in premature babies were not stratified according to gestational ages.
2.	Correlations between Tpo levels and platelet counts in thrombocytopenic babies at different gestations were not worked out.
3.	Study of SGA neonates, as a high-risk group for thrombocytopenia, was not adequate. Tpo level was assessed during thrombocytopenia but not compared with SGA babies with normal platelet count.

Assay of Tpo level is needed in infants with neonatal thrombocytopenia.¹⁰

We aimed to evaluate the role of platelet count and serum Tpo level in neonates presented with thrombocytopenia as predictors for morbidity and/or mortality.

Patients and methods

A prospective cross-sectional study had been conducted on 119 thrombocytopenic neonates: they included 54 full term and 65 preterm (mean platelet count 40.13 ± 33.1, gestational age range 28–42 weeks). A thorough history taking including antenatal and perinatal history was done with special stress on the factors known to affect platelet count and Tpo level as premature rupture of membrane, fetor liquor, maternal hypertension, and perinatal asphyxia. Infants with congenital malformations known to be associated with thrombocytopenia or platelets dysfunction and infants who received blood, platelets, or plasma transfusion or underwent exchange transfusion were excluded. Sera from venous blood were stored at −70°C for Tpo assay which was done using a qualitative enzyme-linked immunosorbent assay technique (Quantikine RD Systems, Inc. Minneapolis, MN, USA). The test was repeated after recovery or deterioration of the condition.

Statistical analysis

Data were analyzed using SPSS PC+ for windows version 16.0; data were non-parametric, so we used the median (or mean rank) and interquartiles for description of central tendency and dispersion. Analysis of matched pairs was done using the Wilcoxon mean rank test; however, Mann–Whitney U-test was the test of difference between two unmatched groups. The Spearmann rank correlation was the test of association between two quantitative non-parametric parameters. Linear regression equation was constructed for the prediction of duration of recovery based on a number of relevant parameters.

Results

Exploration of values of platelet count and serum Tpo level revealed a wide dispersion and hence non-normally distributed. Despite this wide variation, there was a well-defined trend to elevation of platelet count with lowering of Tpo level in relation to recovery of clinical bleeding (Table 1).

Table 1. Platelet count and thrombopoietin level in preterm and full-term thrombocytopenic neonates before and after recovery of bleeding (Wilcoxon matched pairs)

	Before	After	Z	P
Preterm neonates
Platelet count (10^3/μl)	150 (110–260)	260 (190–380)	4.19	<0.001
Thrombopoietin (pg/ml)	590 (320–860)	430 (240–620)	3.98	<0.001
Full-term neonates
Platelet count (10^3/μl)	110 (75–220)	200 (140–260)	3.92	<0.001
Thrombopoietin (pg/ml)	720 (530–910)	450 (390–520)	3.73	<0.001

NB: Values provided are medians (interquartiles: 25th and 75th).

Survival was significantly related to the basal platelet count in full-term neonates (higher in alive compared with died neonates; P = 0.04), otherwise no significant difference between alive and died thrombocytopenic neonates was observed (Table 2).

Table 2. Basal platelet count and thrombopoietin level in preterm and full-term thrombocytopenic neonates in relation to survival (Mann–Whitney U-test)

	Died	Alive	Z	P
Preterm neonates
Platelet count (10^3/μl)	210 (170–295)	160 (110–192)	0.16	0.89
Thrombopoietin (pg/ml)	390 (285–410)	460 (315–510)	1.36	0.18
Full-term neonates
Platelet count (10^3/μl )	110 (105–160)	190 (180–310)	1.92	0.04
Thrombopoietin (pg/ml)	490 (410–525)	430 (420–495)	0.64	0.67

Table 3 showed a statistically negative correlation between thrombocytopenia and Tpo level in neonates especially in preterms (P<0.001), although it was not statistically correlated to the duration of recovery in preterms. This could be due to the presence of multiple risk factors for survival in preterm infants. However, duration of recovery in full-term thrombocytopenic neonates showed significant negative correlation to basal platelet count and serum Tpo level (P<0.001).

Table 3. Correlation between basal platelet count versus thrombopoietin and duration of recovery in relation to gestation (Spearmann rank correlation)

	All neonates		Preterm neonates		Full-term neonates
	r	P	r	P	r	P
Thrombopoietin	0.59	<0.001	0.69	<0.001	0.44	0.001
Duration of recovery	0.42	0.005	0.05	0.82	0.73	<0.001

Multiple stepwise regression analysis for prediction of duration of recovery from thrombocytopenia in neonates was summarized in Table 4. It revealed that platelet count was the best predictor when compared with other parameters including mean platelet volume (MPV) and Tpo level (standardized coefficients 0.9, 0.3, and 0.15, respectively).

Table 4. Prediction of duration of recovery in full-term thrombocytopenic neonates (linear regression)

	Unstandardized coefficients		Standardized coefficients beta	T	P
	B	Standard error
Platelet count (10^3/μl)	−0.076	0.016	−0.921	−4.861	<0.001
Mean platelet volume	−0.536	0.262	−0.388	−2.046	0.059
Thrombopoietin (pg/ml)	<0.001	<0.001	0.158	0.998	0.334
Gestational age (weeks)	−0.047	0.061	−0.122	−0.772	0.452
Constant	21.287	3.814		5.581	<0.001

Discussion

Neonatal thrombocytopenia (platelet count <150 × 10³/μl) develops in 25–50% of risk preterm infants admitted to NICU.¹¹ Estimation of Tpo level in healthy and thrombocytopenic newborns is an important step in understanding the pathophysiology of neonatal thrombocytopenia.¹² Murray et al.¹⁰ and Sola et al.¹³ reported different Tpo levels in both normal and thrombocytopenic neonates.

Folman et al.¹⁴ reported that platelets could function as a strong pool for Tpo apart from uptake of Tpo via c-Mpl expressed by platelets and megakaryocytes. Upon stimulation with various platelet agonists, full length biologically active Tpo, which is contained in granules, was released by platelets. Therefore, the Mpl agonists might compete with Mpl-bound Tpo therapy releasing Tpo intravascularly with massive platelet activation. This means that during infection Tpo increases markedly due to increased destruction of platelets and the release of Tpo stored in platelet organelles or due to block of receptors on megakaryocytes and platelets and decrease uptake and metabolism of Tpo. A minor role may be played by increased production by the liver and spleen as an inflammatory response.

The median serum Tpo level among our studied population was higher than that mentioned in other reports,^12,15 possibly due to the difference in selection criteria of the studied groups. Changes in serum Tpo levels among non-thrombocytopenic cases are possibly due to changes in clearance rate rather than production rate.^16,17 The clearance defect in potentially sick newborn infants is suggested to be due to either low megakaryocyte number, receptor consumption or blockage of cellular megakaryoproliferative oncogene (C-MpL).^18,19

Sainio et al.¹⁸ studied fetal Tpo at different gestation ages. They reported that fetal Tpo levels correlate inversely with gestational age and that term fetuses had lowest level of Tpo than preterm. In addition, Tpo level in cord blood collected at term after cesarean section was very high in comparison with fetal level.²⁰ In this study, preterm newborns had lower levels of Tpo after birth.

In our study no statistically significant difference in median serum Tpo level was found between boys and girls, which agrees with other reports.^12,15 Non-thrombocytopenic babies born to mothers with preeclampsia had statistically significant lower Tpo levels compared with normal babies, in agreement with Albert et al.¹²

SGA babies in this study had statistically significant lower Tpo level compared with that appropriate for gestational age (AGA) whether term or preterm, which might be due to the effect of compromised intrauterine growth with or without preeclampsia on serum Tpo production. Potentially infected non-thrombocytopenic babies with antenatal risk factors of infection such as maternal fever and smelly liquor had statistically significant higher Tpo level than others, which is concordant with Colarizi et al.¹⁶ and Wolber et al.²¹ demonstrated that acute inflammation causes significant increase in hepatic Tpo mRNA leading to increased hepatic Tpo production.

MPV (fL) showed negative correlation with Tpo level possibly because of increased megakaryocyte progenitor cells causing increased uptake of Tpo and decreased serum level.

Tpo level in full term and preterm infants is raised in infants with thrombocytopenia.^12,22 In our study, circulating serum Tpo levels among thrombocytopenic infants were widely variable due to the differences in etiology of thrombocytopenia. Colarizi et al.¹⁶ reported very high levels of Tpo among infected infants; these high levels would reflect a defect in Tpo uptake due to blockage of its receptors by antibodies due to immune mechanism or by cortisol as a stress response.²³ Therefore newborns were speculated to be a potential candidate for recombinant human Tpo.²⁴

In this study, SGA thrombocytopenic infants did not show the same ability to increase Tpo level in response to thrombocytopenia when compared with AGA thrombocytopenic infants. There is statistically significant increase in Tpo level in both preterm and full-term babies. Murray et al.¹⁰ reported that thrombocytopenic preterm with intrauterine growth retardation for any reason had low level of megakaryocyte progenitor cells thus causing hypo-regenerative thrombocytopenia. Hiett et al.²⁵ found a significantly smaller number of progenitors in SGA babies compared with those found in AGA group. This may be due to the effect of intrauterine growth failure on Tpo-producing organs as the liver and kidney. Stress could block receptors of megakaryocyte causing reduction of Tpo uptake thus increasing its blood level. Moreover, direct bone marrow depression that occur with stressful conditions causes furthermore suppression of progenitor cells number and its available binding sites for Tpo.²³

Infants with thrombocytopenia due to postnatal problems such as respiratory distress, perinatal asphyxia, and postnatal infection had higher Tpo level compared with thrombocytopenia of unknown etiology.^26–28 In this study, asphyxiated infants had statistically significant high level of Tpo compared with non-asphyxiated babies. Thrombocytopenic infected infants had exceedingly higher levels of circulating Tpo during sepsis than infected non-thrombocytopenic infants; in accordance with Colarizi et al.,¹⁶ possibly due to the inefficient thrombopoiesis with reduced number of megakaryocyte progenitors and reduced uptake. Neonatal thrombocytopenia due to the increased platelet destruction and normal numbers of megakaryocyte was the underlying mechanism responsible for uptake and metabolism of serum Tpo levels rather than the increased production.^16,17,29

Sola et al.¹³ reported great variability in Tpo response in babies with disseminated intravascular coagulopathy, which might be due to the release of Tpo into plasma or due to endotoxemia.³⁰

The median age at onset of thrombocytopenia was 1.8 days (range 1–6 days), and the mean age at recovery was 9.93 ± 3.7 days (range 4–17 days). These findings are similar to that reported in thrombocytopenic infants admitted to other intensive care units.^2,12,13,31 There was no statistically significant difference regarding time course at different gestational ages. The majority of individuals of both groups showed a significant decrease in Tpo with improvement in both full-term and preterm AGA infants. These findings are similar to that described in children³² and adults.⁷ These findings could be either due to an inability to up-regulate Tpo production or to a faster uptake of Tpo by platelet and megakaryocytes. The great variation in Tpo response to thrombocytopenia is consistent with the concept that Tpo level is not directly regulated by platelet count but by the mechanism underlying platelet drop thus stressing the fact that relation between Tpo level and platelet count is not a straightforward one and that there are many contributing factors affecting this level such as gestational age, maternal diseases, and etiology of thrombocytopenia. There are many limitations for the therapeutic use of Tpo for treatment of thrombocytopenic patients e.g. delayed peak platelet response and formation of neutralizing antibodies to the pegylated molecule.^33,34 Recently eltrombopag (non-peptide agonist of the Tpo receptor) was used for the treatment of thrombocytopenia; it causes dose dependent and linear increase of platelet counts without any guarantee against development for the adverse events described with thrombocytopenia, i.e. these adverse effects were the same in both treatment and placebo groups.^35,36

Multiple stepwise regression analysis was carried out to predict variables, which affect duration of recovery and the clinical course of thrombocytopenia. Platelet count has been found to be the best predictor if compared with other parameters like MPV and Tpo level (standardized coefficients 0.9, 0.3, and 0.15, respectively).

Among the thrombocytopenic group, non-survivors had almost the same Tpo response but significantly lower platelet count compared with survivors. Such finding indicates that these babies were able to react to thrombocytopenia by increasing Tpo level and hence their mortality was related to the severity of thrombocytopenia rather than to the level of Tpo.

Moreover, conditions like premature rupture of membranes associated with premature labour, perinatal infection, birth asphyxia, and multiple pregnancies were reported to be risk factors of neonatal morbidity and mortality. All these factors were also associated with high Tpo level in our study.

In conclusion, high Tpo level is a constant finding in thrombocytopenia neonates with different clinical conditions but treatment of thrombocytopenia should be directed to the underlying cause rather than normalization of Tpo level.