Transparency on Platelet Transfusion in Routine Cancer Care: The Key for Optimal Blood Usage?

Haemato-oncological patients are a vulnerable patient group overall and especially with respect to platelet transfusion. Platelet transfusions are essential to prevent and treat bleeding in thrombocytopenic, immunosuppressed patients. The risk of bleeding not only depends on the platelet count but also on the underlying disease, recent haemorrhage, and complications such as sepsis, uraemia, anaemia, necrosis in tumour tissue, and defects in coagulation function [1]. Given these factors and prophylactic transfusion therapy, this patient group receives approximately 75% of all transfused platelet concentrates (PCs) in Germany [2]. PCs are a scarce, valuable resource, as their provision depends on donations. Demographic developments will cause a decrease of blood donors, which could lead to bottlenecks in the supply of blood products in the future [3‒5]. Consequently, the appropriate use of PCs for the right patient at the right time is indispensable.

Although PC transfusions are part of clinical routine, certain points need to be discussed. A discussion is ongoing about the different PC types (apheresis vs. pool, plasma-suspended vs. additive solution-suspended, ABO-matched transfusions, pathogen-inactivated [PI] vs. non-inactivated PCs) and their impact on safety, effectiveness, and costs. Clinical discussions on PC transfusion practice (e.g., platelet thresholds, prophylactic vs. therapeutic transfusion; diagnosis and handling of PC refractoriness) are also continuing [6‒8]. A possible solution to some of these dialogues could be PI platelets, which have a longer shelf life and fewer transfusion reactions [9, 10]. However, ethics, cost, and feasibility due to the absence of a homogeneous patient population [11] are challenges for randomized studies in this area. Real-world data and advancing digitalization could be an opportunity to improve the quality of evidence [12] in transfusion medicine.

Basic updated information for the abovementioned discussions are patient demographics, clinical characteristics, number of transfusions per patient, transfusion intervals, and guideline adherences. So far, no German study has investigated the data generally collected in routine care and the relevant information for evidence-based discussions on PC supply, optimal blood usage, and health economic considerations. Therefore, this prospective, single-centre observational study evaluated whether data documented in routine inpatient cancer care will provide information to fill the aforementioned information gaps, including PI platelets in mixed transfusion as an example. Based on these real-world data, guideline adherence of PC transfusion was also evaluated. Subsequently, we further sought to create an inventory of essential information as a blueprint for the creation of electronic data collection systems for future comprehensive analyses, which will make it possible to drastically simplify recurring tasks and analyses.

This prospective, non-interventional, single-centre observational study included consecutive haematology/oncology inpatients who received at least one transfusion of PCs in routine clinical care in a German teaching hospital in Munich. The study followed a naturalistic design, and routine clinical platelet transfusion was documented consecutively for all patients over a period of 3 months (March 2015–May 2015). Patients younger than 18 years old were excluded. Except for age, no exclusion criteria were applied.

All PCs transfused in this study were derived from apheresis collection. For transfusion, either conventional (CONV) platelet products or PI products (INTERCEPT® Cerus, Concord, CA, USA) were used on a routine clinical basis. Only PCs of blood groups O and A were used. CONV platelet components were obtained from the Department of Transfusion Medicine, Cell Therapeutics and Haemostaseology of the Ludwig-Maximilians-Universität Hospital of Munich. Exclusively single-donor apheresis platelets were produced according to standard operating procedures and applied after 30 Gy of gamma irradiation. Intercept-treated PI apheresis concentrates were obtained from Haema AG, Leipzig, Germany.

The indication for platelet transfusions during the observational period was determined by the patient’s treating physician. Physicians had no influence on the PC type. The type of PC provided was randomly assigned by the Department of Transfusion Medicine.

Following each transfusion of platelet components, patients were monitored (symptoms, severity, and causality) for adverse events (AEs). The investigator recorded any AEs with an onset within 24 h following the start of the transfusion.

Guideline adherence was evaluated using the Cross-Sectional Guidelines for Therapy with Blood Components and Plasma Derivatives – amended edition 2020 and the risk factors for bleeding complications of the German Medical Association (Bundesärztekammer [BÄK]). Transfusions with a platelet count of more than 10 G/L (Group C patients with acute platelet dysfunction due to chemotherapy include patients with thrombocytopenia in the context of disease or therapy without concomitant risk of bleeding) [13] before PC transfusion. No documentation of at least one of the risk factors for bleeding complications of the BÄK was considered not guideline adherent.

The Blood Centre records and patient medical records served as data sources. All PC transfusion-related treatment patterns, outcomes, and AEs, as well as intervals between transfusions, were recorded. Data were captured electronically on an electronic case report form in the electronic data capture system and in SPSS. Patient confidentiality was protected using coded patient identification numbers assigned by the electronic data capture system. The data were analysed by descriptive statistical methods.

For the inventory of data collection for recurring analyses, a list of all relevant variables to quality, effectiveness, and management of PC transfusion was derived from the literature and the real-world data collected for this study.

During the observation period, 94 patients received at least one platelet transfusion. The population was predominantly male (68.1% vs. 31.9%). The mean age was 54.6 years (22–90 years). The largest subpopulation with respect to primary diagnosis suffered from haematopoietic and lymphatic malignant neoplasms (82 patients; 87.2%). Forty-two patients underwent haematopoietic stem-cell transplantation (SCT) during the observation period, of which 27 received allogeneic SCT and 15 autologous SCT (see Table 1).

The analysis included 942 platelet transfusions, of which 476 (50.5%) were CONV PCs, and 466 (49.5%) were PI PCs. The mean ± SD number of transfused PC per patient was 10.81 ± 9.24 (median, 8; range, 1–44) PCs during the study period. The results show that the distribution of PCs among patients was not uniform. Fifty-five patients received between 1 and 7 transfusions during the observation period and a total of 178 transfusions or 18.9% of the total number of transfusions. Eighteen patients received 8–15 PCs (total, 209 PCs; 22.2%), whereas 9 patients received 16–23 (163 PCs; 17.3%). A small subgroup of 12 patients received more than 23 PCs each (see Table 2). In total, this subgroup received 392 PCs or 41.6% of all transfused PCs (see Fig. 1). Figure 2 shows the transfusion patterns are highly individual in terms of the number of PCs and transfusion intervals but no repetitive pattern for any patient group. The inventory (see Table 3) shows which PC transfusion-specific data need to be collected routinely for scientific evaluations. Of the total of 942 PCs transfused, 174 (18.47%) were not transfused according to the guidelines. This means that 18.47% of the PC transfusions were given at a platelet value above 10 G/L, and no risk factors (according to BÄK) were documented.

The mean ± SD transfusion interval was 1.65 ± 1.1 days, PI PC transfusion interval was 1.64 ± 1.1 days, and CONV PC was 1.65 ± 1.2 days. The mean increment was 12,624 ± 14.7 G/L. CONV platelet increment was 14.01 ± 15.8 G/L, and PI PCs showed an increment of 11.2 ± 13.3 G/L. Five (0.5% per transfusion and per patient) minor AEs were documented.

Bleeding that required PC transfusion was documented for 3 patients (76 PCs). In 2 patients (50 PCs), bleeding was documented before and after PC transfusion, whereas bleeding occurred before PC transfusion in the third patient (26 PCs).

This paper describes the transfusion patterns observed in this study. A total of 942 PCs were transfused to 94 patients over a period of 3 months, and 82% of patients suffered from a haematological malignancy. A previous retrospective study in this centre has shown almost the same number of transfused platelets, a comparable number of patients and demographic, and clinical characteristics within a 3-month period [14]. The data are also in line with other German studies [15, 16], but information on real-life transfusion patterns has not yet been established for Germany. A small number of patients received 42% of all PCs. Most patients who underwent transfusion received an allogeneic SCT, which matched the higher number of PC for patients with allogeneic SCT in the literature [17]. Furthermore, the data show that transfusion in this vulnerable patient group is highly individual due to the heterogeneity of the patient’s clinical characteristics and subsequent needs.

Platelet transfusion is safe, and only a limited number of non-severe transfusion reactions have been documented. Available studies showed similar percentages per transfusion and per patient AEs [18]. However, the distribution of PC in the patient population showed that some patients get particularly many PCs. Due to the high number of transfusions, these patients have a higher risk of infection from contaminated PCs. One way to further reduce this risk for such a vulnerable group is to use PI PCs [19]. Our results showed no evidence of intolerability in the mixed administration of PI and CONV PCs. However, this was not the primary research question of our study; thus, a more detailed investigation may be needed.

As our data demonstrate, the absolute count increment measurement and the transfusion intervals of the PI PC compared with the CONV PC differ only minimally. Similar results were also reported in other European studies [16, 20]. Many studies show that PI PCs provide benefits to patients through an extended shelf life [9], which, unfortunately, is not yet approved in Germany. Fewer AEs, including allergic transfusion reactions, naturally improved safety against transfusion-transmitted infections, especially bacterial, and viral infections. However, from a societal perspective and taking into account the globalization of infectious diseases, the saving of donors, a longer shelf life, and a more stable supply of blood products are important considerations [21].

The increasing complexity and number of treatment options in oncological and haematological diseases and supportive therapy demand for easy-to-apply and high-quality guidelines [22]. In order to support the treating physicians in the indication for a transfusion and for the economic use of PCs, evidence-based guidelines, such as the Cross-Sectional Guidelines for Therapy with Blood Components and Plasma Derivatives – amended edition 2020, have been established [13]. In pursuing the goal of platelet conservation, the guidelines must be adopted in everyday clinical practice. In this study, more than 80% of the transfusions were adherent to the guidelines. Still, 175 PCs were not given according to the guidelines and without documented reasons. A similar study was carried out in the same haematology/oncology ward in 2012. During these 3 months in 2012, 81% of nearly 1,000 PC transfusions were guidelines-compliant. Non-adherence to guidelines can have multiple reasons, like time shortage, ignorance, experience, lack of applicability, and confusing guidelines design [23]. The fact that the results of non-adherent PC transfusions did not change in a period of 3 years should invite a closer examination of the reasons for non-adherence in order to maximize conserved use in PC administration.

The sparse data on bleeding suggest that a structured data collection in daily clinical transfusion routine using the stated inventory, for example, is needed. This inventory could serve as a blueprint for digital clinical software applications. In the future, patient data and variables, such as correct count increment or a steep drop in thrombocytes, could be calculated automatically. Such structured data collection will help clarify the highly individual transfusion events observed in this study. Such data would increase transparency and thereby the possibility of routine scientific data analysis, which enables evidence-based decisions on blood management, routine evaluations, and benchmarking both internally and between hospitals.

This study has shown that a small number of patients received most PC transfusions in a haematology/oncology department of a tertiary centre. The PC transfusion process is highly individual. Furthermore, our study highlights the challenges of platelet transfusion practice in clinical care, such as heterogeneity in the patient population. A guideline adherence of 80% indicates room for improvement. Since not all relevant data were collected routinely, the inventory created in this study, which contained all variables needed for scientific evaluations, could serve as a blueprint for future software-based data collection systems. Digitalization of the healthcare system and of electronic patient records could lead to a more structured data collection and create the possibility of linking to different data resources to address the aforementioned issues and for optimal use and transfusion management.

This non-interventional observational survey was conducted in accordance with the applicable European regulations governing clinical investigations of medical devices (Council Directive 93/42/EEC, EN ISO 14155:2011) and the ICH Guideline for Good Clinical Practice E6 (CPMP/ICH/135/95). The research was conducted ethically in accordance with the World Medical Association Declaration of Helsinki. The study protocol was approved by the Ethics Committee of the Medical Department Ludwig-Maximilians-University Munich (reference number: 034-15 MUC-HV00093). Written informed consent was obtained from participants to participate in the study.

K. Berger receives fees for participation in advisory boards from Cerus Corporation. H. Ostermann receives fees for talks and participation in advisory boards from Cerus Corporation. R. Henschler, V. Kratzer, C. Rieger, and G. Wittmann declare no conflicts of interests.

This study was sponsored by research grants from Cerus^TM Europe B.V., MH Amersfoort, The Netherlands and the Rudolf Marx-Stiftung, Ludwig-Maximilians-University, Munich.

Karin Berger has been involved in the development of the conceptional study design and protocol development, ethical review submission, data collection, data analyses, and writing the draft as well the final version of the manuscript. Reinhard Henschler has been involved in designing the study, interpretation of the results. He contributed to the final version of the publication. Vanessa Kratzer has been involved in data collection, data analyses, drafting, and writing up the final version of the publication. Christina Rieger supported as a clinician on the ward data collection in routine clinical care and contributed to the final version of the publication. Georg Wittmann has been involved in the design the study, interpretation of the results, and contributed to the final version of the publication. Helmut Ostermann has been the medical advisor of the study. He conceived and designed the study, critically reviewed the results, and contributed to the draft and final version of the publication.

The data that support the findings of this study are not publicly available due to their containing information that could compromise the privacy of research participants but are available from Karin Berger.