Eltrombopag use in surgical procedures in inherited thrombocytopenia: expanding the role of thrombopoietin-receptor agonists

Abstract

Background

Inherited thrombocytopenias (ITs) are a group of rare disorders characterized by low platelet counts and variable bleeding risk. Surgical procedures in these patients often require transient platelet elevation. Thrombopoietin-receptor agonists, such as eltrombopag, offer a potential alternative to platelet transfusions, which carry risks including alloimmunization and refractoriness.

Objectives

To evaluate the efficacy and safety of short-term eltrombopag administration in patients with ITs undergoing elective surgical procedures.

Methods

Nine patients with confirmed ITs (MYH9-related disease, ANKRD26-related thrombocytopenia [-RT], ITGA2B/ITGB3-RT, ACTN1-RT) underwent 15 surgical procedures. Eltrombopag was administered at 50 mg/d, initiated approximately 3 weeks prior to surgery. Platelet counts, bleeding complications, and adverse events were monitored.

Results

A favorable platelet response occurred in 14 of 15 procedures. Major response criteria—platelet count ≥100 × 10⁹/L or ≥100% increase from baseline—were met in 10 cases. Median baseline platelet count was 45 × 10³/μL, with a median increase of 143%, reaching a median peak of 103 × 10³/μL. No bleeding complications or transfusions were required. One patient with ITGA2B/ITGB3-RT failed to respond and required additional hemostatic support. No major adverse events were observed.

Conclusion

Short-term eltrombopag administration was effective in increasing platelet counts and preventing bleeding complications in patients with IT undergoing surgery. The inclusion of the first reports of ACTN1-RT expands the therapeutic scope of thrombopoietin-receptor agonists. Although larger studies remain challenging due to the rarity of these conditions, our results contribute meaningful clinical evidence to guide perioperative management in IT.

Essentials

• •Inherited platelet disorders can increase bleeding risk during surgery.
• •Nine patients received eltrombopag before 15 planned procedures.
• •Platelet counts increased in 14 of 15 procedures without bleeding or major side effects.
• •Eltrombopag may reduce bleeding and need for platelet transfusion in surgery.

1. Introduction

Inherited thrombocytopenias (ITs) are a heterogeneous group of disorders characterized by a reduced number of blood platelets, which can lead to a bleeding tendency of varying severity—from asymptomatic cases to spontaneous bleeding diathesis. Once considered rare, recent advances in diagnostic tools and disease understanding have led to increased recognition of their prevalence and clinical relevance.

Most ITs result from defects in platelet biogenesis and, less frequently, from increased platelet clearance. Platelet biogenesis involves the differentiation of megakaryocytes (Mks) from hematopoietic stem cells, Mk maturation, and the release of platelets from mature Mks [1]. The thrombopoietin (TPO) receptor pathway regulates the entire process of platelet biogenesis, from Mk differentiation to platelet release, with transcription factors playing a key role in Mk development [1]. TPO is the master regulator of platelet production [2]. TPO and thrombopoietin receptor agonists (TPO-RAs) bind to the myeloproliferative leukemia receptor on Mks, causing a conformational change that activates downstream signaling pathways, ultimately enhancing platelet production [3].

Most patients with IT experience mild or no spontaneous bleeding and typically do not require regular medical intervention. However, in specific clinical scenarios—such as surgery or invasive procedures—increasing the platelet count may be necessary [[4], [5], [6]]. Until recently, platelet transfusion was the standard approach in these settings. Despite its efficacy, platelet transfusion has several drawbacks, including the risks of bacterial contamination, acute transfusion reactions, and alloimmunization, which can lead to refractoriness to future transfusions [7,8]. Given these limitations, the emerging evidence that TPO-RAs may be effective in ITs represents a significant therapeutic advance.

TPO-RAs are widely used in the management of immune thrombocytopenia, aplastic anemia, hepatitis C virus-related thrombocytopenia, and, more recently, in patients with chronic liver disease undergoing invasive procedures [9]. Two TPO-RAs have been predominantly used in ITs: eltrombopag and romiplostim. Eltrombopag is a small, nonpeptide molecule administered orally (on an empty stomach), while romiplostim is a recombinant protein requiring subcutaneous injection. Both stimulate platelet production by binding to the TPO receptor on Mks, although at different receptor sites [9].

These agents have been employed in several IT syndromes, with most case reports and small series focusing on MYH9-related disorders (MYH-RDs), Wiskott–Aldrich syndrome, X-linked thrombocytopenia, monoallelic Bernard–Soulier syndrome, and ANKRD26-related thrombocytopenia (ANKRD26-RT) [[10], [11], [12], [13], [14], [15], [16], [17], [18]]. A small phase II, open-label, dose-escalation trial enrolled 24 patients affected by MYH9-RD, ANKRD26-RT, X-linked thrombocytopenia/Wiskott–Aldrich syndrome, monoallelic Bernard–Soulier syndrome, or ITGB3-related thrombocytopenia [ITGB3-RT]) [19]. The study aimed to assess whether eltrombopag could increase platelet counts in patients who are preparing for major surgery and, in those with a more severe bleeding phenotype, whether long-term treatment with the oral drug could suppress bleeding tendency. In this trial, only the patient with ITGB3-RT failed to benefit from treatment [19]. The available evidence on the use of TPO-RAs in IT has been extensively reviewed elsewhere [18].

Although current data are promising, evidence remains limited. Large-scale trials are unlikely due to the rarity of patients requiring these agents and the lack of formal approval for IT indications. Nevertheless, it seems reasonable to propose short-term TPO-RA administration as a first-line option for elective surgery in IT subtypes for which efficacy has been demonstrated [4]. Moreover, their use should be considered in other ITs caused by defects in platelet biogenesis, even in the absence of direct supporting evidence.

Herein, we report a case series of 15 surgical procedures in 9 patients with different forms of ITs (MYH9-RD, ANKRD26-RT, ITGA2B/ITGB3-RT, and ACTN1-related thrombocytopenia [ACTN1-RT]—the latter being the first reported case of TPO-RA use) who received eltrombopag in preparation for surgery.

2. Methods

We performed a retrospective and prospective analysis of the clinical records of patients treated with eltrombopag in preparation for surgical procedures. All patients belonged to families with IT and had a confirmed genetic and phenotypic diagnosis of congenital thrombocytopenia. They were followed at the Congenital Coagulopathies Centre of Centro Hospitalar Universitário de Santo António, Porto, Portugal. Clinical records were reviewed for diagnosis, planned surgical intervention, platelet counts before, during, and after eltrombopag treatment, as well as for bleeding complications and treatment-related adverse effects.

The Institution’s Pharmaceutical and Therapeutic Committee and the Ethics Committee approved this study and the use of off-label eltrombopag. Written informed consent was obtained from all patients or from the parents, in the case of the pediatric patient, prior to treatment. All patients received eltrombopag at an initial dose of 50 mg/d, started approximately 21 days (3 weeks) before the scheduled surgery. None of the patients had been previously exposed to TPO-RAs.

A major response to treatment was defined as either a platelet count exceeding 100 × 10⁹/L or at least a doubling of the baseline platelet count (≥100% increase), together with the absence of surgery-related bleeding complications, as defined in earlier clinical trials in this field [19].

Peripheral blood samples were collected by venipuncture into vacutainer tubes containing tripotassium EDTA. Blood counts were obtained using the automated hematology analyzer Sysmex XE-2100 (TOA Medical Electronics). Complete blood counts were also performed on the ADVIA 2120i Hematology System (Siemens Healthineers). In cases of macrothrombocytopenia, platelet counts were cross-verified using semiquantitative estimates from peripheral blood smears or by flow cytometry. Liver function parameters—including aspartate transaminase, alanine transaminase, and total and fractionated bilirubin levels—were also measured.

Descriptive statistics were calculated using SPSS version 27 (IBM Corp). Data are reported as mean (± SD) for data with normal distribution and median (range and interquartile range) for data with nonnormal distribution. Extreme outliers were excluded using Z-scores.

3. Results

In this study, we included 15 surgical procedures in 9 different individuals (2 females, 7 males): 3 patients with MYH9-RD, 2 patients with ITGA2B/ITGB3-RT, 2 patients with ANKRD26-RT, and 2 patients with ACTN1-RT. The procedures included 1 endoscopic pancreatic biopsy, 2 hernioplasties, 2 hydrocellectomies, 1 total hip replacement, 1 total knee replacement, 1 balloon mitral valvuloplasty, 1 tympanic membrane repair with cartilage graft, 1 cleft palate repair, 1 varicectomy, 2 stapedectomies and 1 cervical corpectomy. One patient was submitted to 24 dental extractions with 12 dental implants in the same procedure.

The surgeries are matched to each individual and disorder in Table 1. The bleeding risk was categorized as assessed by the anesthesiology/surgical team. Baseline bleeding score, as evaluated using the International Society of Hemostasis Bleeding Assessment Tool, is also shown.

Table 1.

Patient age at treatment, bleeding score (ISTH-BAT), surgical procedure and associated bleeding risk, and inherited platelet disorder.

Patient #	Age at surgery (y)	ISTH-BAT	Surgery	Estimated surgery bleeding risk	Platelet disorder
1	57	8	Endoscopic pancreatic biopsy	Variable	ANKRD26-RT
2	71	0	Stapedectomy	Moderate	ANKRD26-RT
	72		Circumcision	Low
	73		Stapedectomy	Moderate
3	60	0	Balloon mitral valvuloplasty	Moderate	ACTN1-RT
4	34	0	Varicectomy	Low	ACTN1-RT
5	70	4	Hernioplasty	Low	MYH9-RD
6	46	8	Total knee replacement	High	MYH9-RD
7	12	5	Hernioplasty	Low	MYH9-RD
	13		Tympanic repair with cartilage graft	Low
	14		Cleft palate repair	High
8	32	1	Hydrocellectomy	Low	ITGA2B/ITGB3-RT
	33		Hydrocellectomy	Low
	39		Multiple dental extractions and implants	Moderate
9	56	13	Cervical corpectomy	High	ITGA2B/ITGB3-RT

Surgical bleeding risk as assessed by the anesthesiology/surgical team.

ISTH-BAT, International Society on Thrombosis and Haemostasis Bleeding Assessment Tool; RD, related disease; RT, related thrombocytopenia.

Following recommendations by others [4], we aimed to start treatment approximately 3 weeks before surgery. Eltrombopag was initiated at a dose of 50 mg once daily, on average 24 days (± 12 days) before the scheduled procedure. Complete blood counts were performed weekly or biweekly and on the day before surgery. In 14 of the 15 procedures, a favorable response to eltrombopag occurred, with platelet counts peaking on average 20 days (± 5 days) after treatment initiation. The major response criteria were met in 10 of 15 instances. The median baseline platelet count was 45 × 10³/μL (range: 18-93 × 10³/μL; IQR 25 × 10³/μL), with a median increase of 143% (range: 64–280%; IQR 131%), resulting in a median peak platelet count of 103 × 10³/μL (range: 50–214 × 10³/μL; IQR 79 × 10³/μL). Data are presented in Table 2.

Table 2.

Patient’s baseline and peak platelet counts, number of days until peaking, and percentage increase.

Patient #	Baseline platelet count (× 103/μL)	Peak platelet count (× 103/μL)	Percentage increase	Days to peak platelet count	Platelet disorder
1	43	138	220.9	28	ANKRD26-RT
2	43	79	83.7	13	ANKRD26-RT
3	93	211	126.9	22/58	ACTN1-RT
4	68	214	214.7	17	ACTN1-RT
5	28	68	142.9	24	MYH9-RD
6	42	160	280.9	28	MYH9-RD
7	46	80	73.9	13	MYH9-RD
	75	146	94.7	17
	18	50	177.8	18
8	42	103	145.2	24	ITGA2B/ITGB3-RT
	48	79	64.6	14
	42	105	150.0	22
9	65	70	7.7	n/a	ITGA2B/ITGB3-RT

Patient’s baseline corresponds to the platelet count at the start of eltrombopag. Peak platelet count corresponds to the surgery day count.

In 3 procedures (patients 2, 7, and 8), only 2 weeks of treatment were completed (13 days in the first 2 patients and 14 in the last patient). Although these 3 cases failed to meet the defined laboratory response criteria—platelet counts ≥100 × 10⁹/L or ≥100% increase from baseline (achieving only 83.7%, 73.9%, and 64.6% increases, respectively), no surgery-related bleeding complications occurred in any of the cases.

Eltrombopag was discontinued 2 to 3 days after the procedures, with platelet counts returning to baseline levels over the following days. Eltrombopag discontinuation coincided with the day from hospital discharge, as defined by the surgical team assessment. No bleeding complications, platelet transfusions, or use of additional hemostatic agents were recorded. Only one procedure (patient #7) included preprocedural administration of tranexamic acid. No changes were observed in liver function parameters in any patient. One minor adverse event (headache) related to eltrombopag was reported. No major adverse events, including thrombosis or bone marrow malignancy, were observed, as expected with short-term treatment. Prophylactic anticoagulation was not used in any patient. In most instances, due to the nature of the procedure, it was not indicated. In procedures for which there is a formal recommendation for prophylactic anticoagulation after surgery (knee arthroplasty and cervical corpectomy), after multidisciplinary discussion, we chose not to use it, as the patients in question had high baseline bleeding scores, and the risks outweighed the benefits.

3.1. Patient-specific outcomes

• •Patient 1 (ANKRD26-RT) initiated eltrombopag while awaiting an endoscopic pancreatic biopsy and potential surgery for suspected pancreatic neoplasia. Six days after starting treatment, he underwent the biopsy with a platelet count of 78 × 10³/μL, without bleeding complications. He continued treatment for 28 days, eventually discontinuing eltrombopag when the diagnosis of pancreatic neoplasia was excluded and surgery deemed unnecessary.
• •Patient 2 (ANKRD26-RT) was submitted to 2 stapedectomies for otosclerosis and a circumcision. This patient’s platelet count was consistently <50 × 10³/μL. In all instances, the platelet count increased higher than the recommended cutoff for the surgery, with peak counts of 75, 79, and 96 × 10³/μL. The time to platelet peak, however, increased with each successive treatment, from 13 days in the first to 33 days for the third procedure.

Both patients (from 2 independent kindreds) were found to harbor a novel heterozygous variant affecting the 5′ untranslated region of ANKRD26, located within a known mutational hotspot, repeatedly implicated in pathogenic gain-of-function dysregulation of ANKRD26 expression [20]. Although initially considered a variant of uncertain significance, according to the American College of Medical Genetics and Genomics (ACMG) and Association of Molecular Pathology (AMP) guidelines, it has been reclassified as likely pathogenic by our laboratory.

• •Patients 3 and 4, a father-son pair with ACTN1-RT due to a previously reported pathogenic variant in ACTN1 [21,22], both responded well to eltrombopag. The father, with platelet counts between 70 and 90 × 10³/μL, was treated to achieve normalization of counts prior to balloon mitral valvuloplasty, as this surgery requires systemic anticoagulation during the procedure and carries high morbidity and mortality in the event of bleeding. He reached a peak platelet count of 211 × 10³/μL after 22 days of treatment, and the surgery was completed without complications. Eltrombopag was continued for 3 days, while the patient was under heparin bridging to warfarin. After this period, the patient remained anticoagulated and closely monitored. No excess bleeding was reported. The son, despite adequate platelet counts for a safenectomy and absent bleeding phenotype, was still proposed to eltrombopag, after multidisciplinary discussion with the surgical team. The decision reflects a precautionary approach rather than a clear-cut indication, as there is a lack of guidelines for most of the mild inherited platelet disorders. The patient had a notable 215% increase in platelet count after 17 days of treatment, without any surgical complications.
• •Patients 5, 6, and 7 belong to a 3-generation pedigree with MYH9-RD. They carry a previously unreported heterozygous in-frame deletion in MYH9, located within a functionally relevant region of the gene, classified as likely pathogenic according to ACMG/AMP criteria. In total, they underwent 5 procedures. Their baseline platelet counts ranged from 18 × 10³/μL to 75 × 10³/μL (Table 1). In this disorder, given the inherent macrothrombocytopenia, with often-inaccurate counts by standard methods and possible associated platelet dysfunction, standard platelet cutoffs may be unreliable. As such, we opted to increase the counts in these patients. A major platelet response was observed in only 2 of 5 treatments. For patient 7, in the first procedure, the increase from baseline was only from 46 to 80 × 10³/μL. This patient received treatment for only 13 days in this instance—the shortest duration to reach peak count in our cohort. However, for the second procedure, there was an increase from 75 to 146 × 10³/μL with 17 days of treatment,. Given the impact from such a small increase in treatment duration, it was surprising that for the third procedure, with 18 days of treatment, the peak was only 50 × 10³/μL. Given the low peak, this was the only procedure in which a hemostatic agent (tranexamic acid) was used as well. Of note, patient 7 was the patient with shorter treatment durations, due to repeated surgery scheduling/drug dispensing issues, which may explain the inadequate response. Patients 5 and 6 showed adequate platelet peaks with proposed 20 to 30 day treatment durations.
• •Patient 8 (ITGA2B/ITGB3-RT) was the first patient with IT to be treated with eltrombopag at our center, in 2017. In previous surgeries, he had experienced adverse events (seizures) with desmopressin and antifibrinolytics. He underwent the same surgical procedure twice without bleeding complications. He achieved a major laboratory response in the first procedure, with a platelet count increase from 42 to 103 × 10³/μL, although treatment was prematurely discontinued due to a severe headache. In the second procedure, we opted for a shorter treatment duration, with only 2 weeks of treatment. Although he did not achieve a major response criterion, there were no bleeding complications. However, he did not present any headaches. As such, he was motivated for a standard treatment for the third procedure, as the headaches did not seem to be related to eltrombopag. He was treated for 22 days with a platelet count increase from 42 to 105 × 10³/μL.
• •Patient 9 showed no hematologic response after 3 weeks of eltrombopag therapy. The dose was subsequently increased to 75 mg for an additional 8 days, resulting in a peak platelet count of only 70 × 10³/μL. The perioperative strategy was influenced by a previous lumbar corpectomy performed 15 years earlier under desmopressin, at a lower platelet count and without bleeding complications. Based on this history, a similar hemostatic approach was selected for the current procedure. As such, management included desmopressin (0.3 μg/kg) and tranexamic acid (1 g bolus). However, following vertebral muscle detachment, the patient developed significant intraoperative bleeding, requiring transfusion of a single unit of platelet concentrate, which resulted in adequate hemostasis. This bleeding episode was unexpected and may reflect differences in surgical complexity or procedure-related factors rather than an intrinsic change in baseline hemostatic status. Ultimately, hemostasis was achieved with the combined use of desmopressin, tranexamic acid, and platelet transfusion.

Both patients, from unrelated kindreds, carry with a previously reported variant in ITGA2B [23,24], considered pathogenic according to the ACMG/AMP guidelines.

4. Discussion

As TPO is the principal regulator of platelet biogenesis, TPO-RAs are expected to benefit patients with ITs caused primarily by defects in platelet production. One of the first proposed uses of TPO-RAs in IT was for disorders associated with impaired proplatelet formation. Most of these ITs stem from mutations in genes encoding components of the actomyosin cytoskeleton and, in some cases, from alterations in membrane glycoproteins that indirectly affect cytoskeletal structure or reorganization [1].

The majority of published reports [[10], [11], [12], [13], [14], [15], [16], [17], [18], [19]] on TPO-RA use in IT focus on MYH9-RD, a disorder characterized by defective proplatelet formation due to mutations in the MYH9 gene, which encodes the heavy chain of nonmuscle myosin IIA [25,26]. These mutations disrupt the late stages of platelet release from mature Mks, despite preserved Mk differentiation and maturation [23]. As such, TPO-RAs have shown efficacy in most reported MYH9-RD cases. In our cohort, all 3 patients with MYH9-RD—who underwent a total of 6 surgical procedures—achieved platelet count increases adequate for the planned surgeries.

Another form of macrothrombocytopenia, ACTN1-RT, results from mutations in ACTN1, which encodes α-actinin-1, a cytoskeletal protein involved in actin filament organization. Patients with ACTN1-RT display abnormalities in proplatelet formation, including reduced numbers of proplatelet tips and enlarged tip size, similar to findings in MYH9-RD [1,22]. Despite the predicted responsiveness to TPO-RAs in this context, no prior reports on their use in ACTN1-RT exist to our knowledge. In our study, both patients with ACTN1-RT demonstrated robust responses, achieving platelet counts >200 × 10³/μL. These findings represent the first reported use of TPO-RAs in ACTN1-RT and suggest their broader applicability in the treatment of cytoskeletal ITs.

In ITGA2B/ITGB3-RT, mutations in membrane glycoprotein receptors are thought to interfere with cytoskeletal dynamics, impairing proplatelet formation. Enhanced α_IIb β₃ signaling may contribute to this defect [27]. Experience with TPO-RAs in this subtype is scarce. In one prior report, a patient with ITGB3-RT failed to achieve a hematologic response with eltrombopag [18], as did our patient 9. In contrast, our patient 8 with ITGA2B/ITGB3-RT, who underwent 3 invasive procedures, showed a favorable platelet response and had no bleeding complications, despite discontinuation of therapy due to headache in the first surgery and suboptimal treatment duration in the second.

Familial thrombocytopenia type 2, caused by point mutations in the 5′ untranslated region of ANKRD26, leads to persistent gene expression during late megakaryopoiesis due to impaired binding of RUNX1 and FLI1 transcription factors. This results in aberrant signaling via the myeloproliferative leukemia pathway and impaired proplatelet formation [28]. In a recent phase 2 study, 9 patients with ANKRD26-RT treated with eltrombopag exhibited either minor (n = 8) or no (n = 1) responses [19]. In contrast, both of our patients responded—one with a major and the other, who received only 2 weeks of treatment, with a minor response—further supporting the variability of response and potential benefit in selected patients.

Collectively, our cohort of patients with MYH9-RD, ACTN1-RT, ITGA2B/ITGB3-RT, and ANKRD26-RT confirms that in all but 1 case, TPO-RAs may be effective in ITs where thrombocytopenia primarily results from impaired proplatelet formation, provided Mk maturation and TPO responsiveness are preserved [1].

Among the most promising applications of TPO-RAs—particularly eltrombopag—is their use as a preoperative measure in patients with IT undergoing elective invasive procedures. To date, 19 such procedures supported by TPO-RAs have been reported, with only 1 documented failure [19]. Our data, which include 15 additional procedures with just 1 further failure, reinforce these findings and broaden the current understanding of their clinical utility in this context.

All patients in our study received eltrombopag at a standardized dose of 50 mg/d, typically initiated 3 weeks prior to surgery. Peak platelet counts were observed around day 20 of treatment. In 3 cases (20%), where eltrombopag was started only 2 weeks before surgery, the platelet response was suboptimal by laboratory criteria but remained sufficient to avoid perioperative bleeding. Based on these results, we recommend initiating eltrombopag at least 3 weeks before major and 2 weeks before minor surgical interventions.

Notably, no bleeding events, platelet transfusions, or use of additional hemostatic agents (except for 1 case of tranexamic acid) in patients with response criteria were required. Only 1 minor adverse event (headache) was recorded, and no major complications such as thrombosis or hepatotoxicity were observed, consistent with short-term use.

A single procedure (in patient 9) was considered a failure, requiring additional hemostatic support with desmopressin, tranexamic acid, and platelet transfusion.

This study has limitations, however, including its retrospective design and small sample size. Larger prospective studies remain challenging due to the rarity of IT and the limited number of patients requiring TPO-RAs. Nevertheless, small-scale studies like ours are crucial to build the evidence base for the safe and effective use of eltrombopag in IT.

5. Conclusion

This cohort expands the reported use of eltrombopag in IT patients undergoing surgery. Notably, this includes the first reported cases in ACTN1-RT, thereby broadening the potential indications of TPO-RAs in congenital thrombocytopenia. Our findings support previous evidence of efficacy and safety, particularly in ITs involving impaired proplatelet formation.