Thrombocytopenia (TCP) is the most common hematological complication in patients with chronic liver disease (CLD) (Qamar et al., 2009). Due to the decrease in the platelet count, these patients usually have a higher risk of bleeding during invasive procedures, resulting in increased postoperative complications and mortality (Giannini et al., 2010; Forkin et al., 2018). Traditionally, a growing number of studies have been suggesting that prophylactic platelet (PLT) transfusion before or after invasive procedures can improve the hemostatic potential and reduce the risk of bleeding in CLD patients with severe TCP (platelet count <50 × 10⁹/L) (Maan et al., 2015; Mitchell et al., 2016). This threshold is used as a standard for clinical care and invasive procedures (Scharf, 2021).

Thrombopoietin (TPO) is an important physiological regulator of thrombopoiesis and megakaryocyte maturation, which increases the number of PLTs in the blood by stimulating the production of endogenous functional PLTs. Accumulating studies have revealed that TPO is a therapeutic target that stimulates PLT production (Qureshi et al., 2016; Gilreath et al., 2021).

When considering cost-effectiveness, analysis of increasing PLT values to the same level using traditional PLT infusion may be less costly than TPO agonists. Many centers still use traditional PLT transfusion. However, PLT transfusion may be associated with serious complications, including transfusion reactions and the transmission of infectious agents, which can be fatal in rare cases, and alloimmunization can occur in patients who have received multiple blood product transfusions because recipients form antibodies against donor platelets (Basili et al., 2018; Armstrong et al., 2020; Elsaid et al., 2020; Karlström et al., 2022). In addition, platelet transfusions have only a short-term effect, while TPO agonists have a longer effect. Thus, new and efficient drugs targeting this issue need to be developed and explored in clinical practice.

Recombinant human thrombopoietin (rh-TPO) is generally used for treating thrombocytopenia caused by chemotherapy and immune thrombocytopenia (Vadhan-Raj et al., 2003; Yu et al., 2020). There are no prospective clinical trials to confirm its use in CLD, but multicenter real-world studies have demonstrated its efficacy and safety in CLD patients with TPO (Feng et al., 2022). Avatrombopag, as a novel oral TPO receptor agonist, has been approved recently as an alternative treatment for PLT in the treatment of invasive surgery for CLD patients with severe TCP. Phases I, II, and III studies demonstrated the efficacy and safety of avatrombopag in patients with severe TCP related to CLD (Terrault et al., 2014; Kuter and Allen, 2018; Terrault et al., 2018). Since both drugs are used in this condition, a subsequent issue arises: whether the two drugs can increase platelet count before invasive surgical procedures, whether oral and subcutaneous injections have the same efficacy, and whether there is a difference in adverse drug events effects. However, to date, few studies have focused on the comparison between the efficacy and safety of avatrombopag and rh-TPO.

Thus, in the present study, we analyzed the clinical data of patients with TCP associated with CLD before and after propensity score matching (PSM) to explore the efficacy of avatrombopag compared with rh-TPO in the real-world study.

At present, rh-TPO, glucocorticoids, avatrombopag, and lusutrombopag are commonly used in the clinical treatment of CLD-related TCP. Rh-TPO is a full-length glycosylated TPO purified by gene recombination technology. Rh-TPO and endogenous TPO have similar pharmacological effects in increasing PLT by activating C-MPL (Demetri, 2000). However, problems with the clinical use of rh-TPO include subcutaneous administration and an inability to predict an increase in PLT count. Therefore, new improved drugs or alternative drugs are constantly being explored and developed. Avatrombopag is a novel, oral, small-molecule TPO-R agonist (Terrault et al., 2014; Terrault et al., 2018). It ultimately promotes megakaryocyte proliferation/differentiation and predictably increases the PLT count through the activation of the JAK-STAT and SHC-Ras-Raf-ERK signaling pathways and, therefore, could be used as an alternative to PLT infusion (Bussel et al., 2014; Kuter and Allen, 2018; Długosz-Danecka et al., 2019). Avatrombopag only increases the number of PLT in patients and does not enhance PLT activation (Michelson et al., 2018; Kuter, 2022). To date, there have been few real-world studies of these two drugs in the treatment of TCP associated with CLD, let alone any comparisons between the two drugs.

Two global, multicenter, randomized, double-blind, placebo-controlled phase II (E5501-G000-202, NCT00914927) and phase III clinical trials (ADAPT-1, ADAPT-2) evaluated the efficacy and safety of avatrombopag in patients with severe TCP related to CLD who underwent invasive procedures or surgery. The results suggest that avatrombopag can be used to treat severe TCP related to CLD with high efficacy and safety (Terrault et al., 2014; Terrault et al., 2018). A nationwide multicenter, real-world study has reported the efficacy and safety of rh-TPO in the treatment of TCP caused by cirrhosis (Feng et al., 2022). At the same time, an ethnic sensitivity analysis concluded that avatrombopag was effective and safe in the management of thrombocytopenia in Chinese patients with CLD. Ethnicity does not appear to influence the efficacy and safety of avatrombopag (Lu et al., 2022). Practice Guidelines of the Central European Hepatologic Collaboration (CEHC) on the use of TPO receptor agonists in patients with CLD undergoing invasive procedures were reported. A consensus was agreed that TPO-RAs should be considered for raising PLT count in CLD patients undergoing scheduled abdominal surgery, high-bleeding risk dentistry, endoscopic polypectomy, endoscopic variceal ligation, liver biopsy, liver surgery, liver transplantation, and percutaneous ablation, but it was also agreed that they are less beneficial or not necessary for endoscopy without intervention and paracentesis (Flisiak et al., 2021). Our study included patients with these high-risk procedures, and there were no serious bleeding events in the short term after medication.

In the present real-world study, two important results were found. First, the treatment response rate was 75.18% (106/141) in the avatrombopag group and 68.75% (66/96) in the rh-TPO group. Both before and after PSM, there was no statistically significant difference in treatment efficiency between the two groups. However, in the treatment of Child‒Pugh Class A patients both before and after PSM, the effective rate of the avatrombopag group was higher than that of the rh-TPO group, with a statistically significant difference between the two groups (88.89% vs. 63.41%; 81.33% vs. 61.76%). The use of avatrombopag in the treatment of TCP associated with CLD in Child‒Pugh Class A may achieve better clinical efficacy and significantly reduce the risk of invasive procedures. We thought that this might be because of the way in which it was administered: avatrombopag is administered orally, while rh-TPO is administered subcutaneously. In patients with better liver function, the more effective the drug is absorbed through the digestive system, the better the platelet count will be. For patients with severe liver disease, oral medication is far less effective than injection (Pereira et al., 2005). The clinical response rate of avatrombopag was lower than that of the two phase III clinical trials (75.18% vs. 88.1%/87.9%), which may be due to the difference in the patient populations selected: few patients with liver function in the stage of Child‒Pugh Class C were included, and patients with hepatocellular carcinoma (Barcelona-Clinic Liver Cancer staging classification C or D) were excluded from these clinical trials (Terrault et al., 2014; Terrault et al., 2018). Given the actual clinical situation, these patients excluded from prospective clinical trials will also have application scenarios of the abovementioned drugs in the real world. Second, the efficacy rates of avatrombopag and rh-TPO in patients with initial PLTs >25 × 10⁹/L were higher than those in patients with initial PLTs ≤25 × 10⁹/L both before and after PSM (55.17% vs. 80.36%, 53.85% vs. 78.95%; 55.56% vs. 81.36%, 53.85% vs. 76.47%). Considering the results of ROC analysis (the best threshold of PLT was 26.5 × 10⁹/L both before and after PSM) and the convenience for clinical use, we set the cutoff value as 25 × 10⁹/L.

Regarding safety, there were no serious systemic adverse events caused in the two groups. The most common complications were abdominal pain, fatigue, dizziness, and fever. There was no significant difference in the incidence of abdominal pain, pyrexia, fatigue, indigestion, nausea, diarrhea, peripheral edema, or other adverse events between the two groups. However, the incidence of dizziness in the rh-TPO group was higher than that in the avatrombopag group both before and after PSM (25.0% vs. 7.8%; 24.7% vs. 7.8%). The incidence of abdominal pain and fatigue was higher than that in clinical trials (22.75 vs. 10.3%; 19.9% vs. 6.7%), and there were no significant differences in other adverse events. This may be because we included more patients with Child‒Pugh Class C, whereas clinical trials rarely included such patients (Terrault et al., 2018). These patients have abdominal pain and fatigue symptoms, and it can be difficult to distinguish whether the drug causes these symptoms. In addition, with respect to the delivery route, patients treated with rh-TPO require subcutaneous injection, while patients treated with avatrombopag require oral administration, which is more convenient and safer.

There are still some limitations in our study. First, compared with the prospective clinical trials, it is difficult to monitor patients’ PLT counts dynamically each day due to various issues such as ethics and patient compliance in a real-world study. Second, the change in PLT counts was only directly observed up to 14 days after medication due to the urgency and need for surgical treatment of the patient’s disease, especially in cases where the PLT counts met the minimum requirement for invasive surgery or related treatment (>50 × 10⁹/L). Thus, a long-term result of efficacy and safety needs to be observed in future studies if conditions permit. Third, a limited number of cases were included in the study due to the short time that avatrombopag has been on the market. All of these issues will be focused on and updated further in our future studies.

Taken together, to the best of our knowledge, this is the first real-world study focusing on the comparison of efficacy and safety between avatrombopag and rh-TPO. Our findings proved that avatrombopag showed better clinical efficacy than rh-TPO in the treatment of TCP associated with CLD in Child‒Pugh Class A and showed a lower incidence of dizziness in all patients. The effective rates of avatrombopag and rh-TPO in patients with initial PLTs >25 × 10⁹/L were higher than those in patients with initial PLTs ≤25 × 10⁹/L. In the future, a large cohort or prospective clinical study still needs to be performed to further verify the abovementioned findings.