Thrombocytosis in neonates and young infants: a report of 25 patients with platelet counts of |[ges]|1|[thinsp]|000|[thinsp]|000|[thinsp]||[mu]|l|[minus]|1

The World Health Organization defines thrombocytosis as a platelet count exceeding 600 000 μl−1.1 Thrombocytosis has been reported in neonates and young infants, in which it has variably been defined as a platelet count of >400 000, >600 000 or >900 000 μl−1. 2, 3, 4, 5, 6, 7 We recently reported platelet counts of over 47 000 of the subjects obtained during the first 90 days after birth, and observed that the 95th percentile upper reference range can be as high as 750 000 μl−1.8 Adult patients with very high platelet counts can be at risk for significant thromboembolic and/or hemorrhagic complications,9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 but it is unclear whether similar risks exist for young infants. Moreover, neither the prevalence and causes of thrombocytosis nor the associated conditions or outcomes in infants are well known. To better understand this condition, we undertook a historic cohort study of patients 140 days old (20 weeks), in a multi-hospital health-care system who, in the last 6 years, had a platelet count of 1 000 000 μl−1. During the 6-year inclusion period, 40 471 young infants had one or more platelet count(s) in an Intermountain Healthcare facility, and a platelet count of 1 000 000 μl−1 was identified in 25 patients. The demographic features of these 25 patients are shown in Table 1. One patient had two separate episodes. All 25 patients were hospitalized when extreme thrombocytosis was identified; 4 were identified at the time of hospital admission and 21 had been in-patients for 1 week or more with a previously normal platelet count. None of the patients had phenotypic features of Trisomy 21 or hematologic findings suggestive of a myeloproliferative disorder. The week of life when extreme thrombocytosis was recognized is shown in Figure 1. No episodes were identified in the first week, approximately 40% were observed between the second and fourth weeks and 40% were between weeks 5 and 8. The highest recorded platelet count in these 26 episodes ranged from 1 000 000 to 1 341 000 μl−1. The medical conditions temporally associated with the episodes are shown in Table 2. Infectious diseases were the most frequently associated conditions, occurring in 12. The varieties of infections are shown in Table 3. The onset of infections associated with thrombocytosis occurred during the first week of life in 2 patients (early onset) and after the first week of life in the remaining 10 patients (late onset). Thrombocytosis was recognized in these patients at a median of 9 days after the infection was diagnosed. Postoperative thrombocytosis was identified in eight patients (Table 2); two had surgery for congenital heart disease, two had neurosurgery, two had otolaryngologic/craniofacial surgery and two had bowel-related surgery (Table 4). Thrombocytosis was recognized in these patients at a median of 13 days postoperatively (except one in whom it was noted at the time the patient was readmitted for bowel obstruction secondary to stricture formation, which was 12 weeks after surgery for jejunal atresia and 6 weeks after medical treatment for necrotizing enterocolitis). Four patients were ‘growing’ premature infants with normochromic, normocytic anemia (Table 2). Their gestational ages at delivery ranged from 29 2/7 to 34 1/7 weeks. Out of four, two were receiving oral iron supplementation and two were not and none were receiving human recombinant erythropoietin. Extreme thrombocytosis was detected in these patients at a median age of 26 days after birth. Of the two other patients, one had congenital adrenal hyperplasia (CAH) and the other had methadone withdrawal syndrome (Table 2). Thrombocytosis was noted in these patients on day of life 11 and 10, respectively. The time to a normal platelet count was not accurately discernable because follow-up platelet counts were variably spaced, and post-hospital-discharge platelet counts were not available for five patients. However, for 20 of the 26 episodes, a reduction in platelet count to <1 000 000 μl−1 (mean value=734 000±47 182 μl−1) was noted after 7 days (mean, range 1 to 30 days). No thrombotic complications were clinically suspected and no hemorrhagic episodes were recognized in any of the patients. One patient with complex congenital heart disease was treated with a low-dose of aspirin (20 mg day through discharge from the hospital) because of the concern that the high platelet count might result in thrombotic complications. Thrombocytosis can be classified as either essential (primary) or reactive (secondary). Essential (primary) thrombocytosis is a clonal myeloproliferative disorder with an estimated annual incidence of about 2 per million adults and 0.09 per million children.13, 14, 15, 16, 18, 19, 20 Patients with essential thrombocytosis are at risk for thromboembolic and/or hemorrhagic complications, and many receive anti-platelet or cytoreductive treatment to lower the platelet count and thereby reduce these risks.9, 13, 14, 15, 16, 18, 20 In contrast, reactive (secondary) thrombocytosis is a more common, and generally benign, condition of limited duration.17 The term ‘reactive’ is used to imply that platelet production has increased as a reaction to an underlying condition, generally an infectious or inflammatory disorder.2, 3, 4, 5, 6, 7, 9, 20 Thromboembolic and hemorrhagic complications are rare in patients with reactive thrombocytosis and anti-platelet or cytoreductive treatments are not usually needed, unless other risks are identified.20, 21 Thrombocytosis can also be classified according to severity, as judged by the peak platelet count. To provide reporting consistency, Sutor proposed the following severity classification: mild thrombocytosis=peak count between 500 000 and 700 000 μl−1; moderate thombocytosis=700 000 to 900 000 μl−1; severe thrombocytosis=900 000 to 1 000 000 μl−1; and extreme thrombocytosis 1 000 000 μl−1.5 For this study we used Sutor's classification of extreme thrombocytosis, and after identifying all such among young infants in our electronic records in the past 6 years, we sought to judge whether each case was essential or reactive. We also sought adverse outcomes, specifically looking for thrombotic or hemorrhagic events in the historical record. We also attempted to determine time to resolution of thrombocytosis. Our findings are consistent with those of others who focused on older subjects. For instance, Buss et al.9 identified 280 patients with extreme thrombocytosis, ranging from 12 days to 100 years old, and reported that over 80% of them had reactive thrombocytosis. Schilling22 identified 102 patients with extreme thrombocytosis and judged that 73% were reactive. Denton and Davis23 reported that 1.1% (31/2749) of children admitted to the Pediatric Intensive Care Unit at Bristol Royal Hospital for Children developed extreme thrombocytosis, and all cases were judged to be reactive. In this study 40 471 infants had one or more platelet counts, of which 25 had a count of 1 000 000 μl−1. We judge that all were likely to have been reactive, on the basis that each occurred accompanying a condition known to be associated with reactive thrombocytosis, none had a clinically recognized thrombotic or hemorrhagic event and all resolved. The mechanism responsible for reactive thrombocytosis is almost certain to be increased megakaryopoiesis. Thrombopoietin is the primary humoral regulator of megakaryopoiesis,24 but other cytokines and hematopoietic growth factors, such as interleukin-3 (IL-3), IL-6 and IL-11, can also have contributing roles.25 Increased production of these factors during a period of infection or inflammation likely stimulates megakaryocyte production, resulting in higher platelet counts. A longitudinal analysis, by Ishiguro et al.26 of circulating thrombopoietin and IL-6 concentrations during episodes of infection supports this theory. Most children with reactive thrombocytosis have an infectious disease,21 and many of these involve the respiratory tract.2, 4, 5, 27 Similarly, in our report, infections were associated with 46% of episodes, about half of which were respiratory. Postoperative reactive thrombocytosis is also well known, with peak platelet counts at 7 to 20 days after surgery.2, 3, 4, 5, 6, 7, 9, 10, 28 We found extreme thrombocytosis at a median of 13 days postoperatively. In adults, postoperative thrombocytosis is sometimes considered a risk factor for thrombotic complications,10 but no vaso-occlusive or hemorrhagic events were detected in our patients. The mechanisms responsible for postoperative thrombocytosis have not been delineated, but the work of Folman et al.29 indicates that it is likely similar to infectious-related thrombocytosis, with increased production of IL-6 and thrombopoietin accompanying inflammation. In the current series, anemia was the third most common condition associated with extreme thrombocytosis. Thrombocytosis has been reported in patients with hemolytic anemia or iron deficiency.2, 3, 4, 5, 6, 7, 12, 20, 21, 27, 30, 31 None of our four anemic patients had recognized hemolysis and all were thought to have the anemia of prematurity. Although elevated platelet counts have been observed in anemic patients receiving human recombinant erythropoietin,32 none in our current study had received this medication. One of our patients had CAH. An association between CAH and thrombocytosis has been reported.2, 33 Gasparini et al.33 found elevated platelet counts in 21 infants with CAH and reported a correlation between the levels of 17-OH progesterone and the degree of platelet elevation. Platelet counts of all 21 infants normalized within 4 weeks of diagnosis and treatment. The mechanism leading to thrombocytosis in patients with CAH is unclear. Several studies have reported thrombocytosis in infants born to mothers receiving methadone with or without polydrug abuse.34, 35, 36, 37 Burstein et al. 34 determined that after the first week of life, platelet counts of 33 neonates born to mothers using methadone were higher than controls. The neonate we observed with thrombocytosis after signs of narcotic withdrawal was treated with methadone, beginning on the third day of life. In a murine model, offspring of mice given methadone daily during pregnancy had higher platelet counts and increased bone marrow megakaryocytes.38 The mechanism for this increase is not known. Extreme thrombocytosis is rarely encountered in neonatology practice, but when cases are observed they can be puzzling and concerning. On the basis of 25 affected patients in a multi-hospital health-care system, we speculate that the great majority of neonates and young infants who develop extreme thrombocytosis have reactive (secondary) thrombocytosis, not essential (primary) thrombocytosis. The courses and outcomes of our 25 patients lead us to conclude that this is generally a benign and self-limited condition, and that anti-platelet or cytoreductive treatments are generally not needed. The authors declare no conflict of interest.