Multinational Analysis of Children Transfused With Pathogen Inactivated Platelets

This is a retrospective multisite cohort study using anonymized data. Each participating site received local institutional ethics approval. The sites gathered data during two time periods, the control period (CP) and the study period (SP). The SP was defined as the time the site began using PI platelets on a widespread basis including all patients at the center, until 2019. The CP was defined as the time period before PI was introduced. The CP and SP were each at least 1 year and designed to be approximately the same amount of time to achieve balanced patient numbers in each. For those sites that have dual inventory (PI and conventional platelets), patients were only included if the patient only received all conventional or all PI platelets. Patients were excluded if they received both PI and conventional platelets. Patients were categorized into age strata by the age of first platelet transfusion: <4 months, 4 months to <2 years, 2 years to <10 years, 10 years to 18 years. To determine platelet utilization and transfusion-related adverse events, patient charts were reviewed. Patients were categorized as having 1 platelet transfusion, 2 platelet transfusions, or >2 platelet transfusions.

The study centers obtained local ethics review and approval. The data were assembled at each site and analyzed in a deidentified manner at Children’s National Hospital after acknowledgment from the IRB for the data usage. Descriptive statistics were used to present the data. Continuous variables were shown with medians along with their accompanying quartile values and SP versus CP was statistically compared with the Wilcoxon rank test. Categorical data were presented with frequencies and either the χ² test or the Fisher’s exact test was used for the SP versus CP comparisons. Because this was an exploratory analysis, we chose not to adjust our P value cut point to account for multiple tests. Rather, we identified those comparisons with P values less than .05 and used these test percentage or medians to evaluate the result for clinical significance. All analyses were performed using SAS version 9.4 (Cary, NC).

The study included 3839 pediatric patients who were 0 to 18 years of age who received >7930 platelet transfusion in total across 4 centers in 3 countries. The study included platelet transfusions between 2013 and 2019. Although the control period did not include PI platelets, product composition was different across sites. Sites prepared platelet concentrates by apheresis and/or the buffy coat method (Table 1). Apheresis platelets are collected from the donor by using an apheresis instrument, whereas buffy coat platelets are made by centrifuging whole blood donations and harvesting the platelets from the buffy layer. Sites prepared platelet concentrates in plasma only or in PAS. The age of patients included in the study spanned 0 to 18 years fairly evenly, with no difference in the age ranges of patients who received platelet transfusions between the control and study period (Table 2).

Overall reaction numbers and percentages and the 9 individual reaction type results for SP and CP, with corresponding P values to assess the differences in their rates are presented in Table 3. When all transfusion reaction types were considered, there was not a difference in the percentage of patients who had transfusion reactions between the time periods (1.0% and 1.1%, P = .803). All of the patients who had adverse events were multitransfused, with some receiving many platelet transfusions. Among those with a transfusion reaction, two diagnoses indicated statically significant differences comparing CP to SP. The first, “mild allergic skin reaction only” results were SP = 0.2%, CP = 0.7%, P value = 0.018, and the absolute risk reduction (SP to CP) was −0.5% with 95% CI (−0.12% to −0.93%). The change in rate is minimal and may or may not represent a clinically significant change. The second, “at least one moderate to severe allergic reaction” results were SP = 0.2%, CP = 0.0%, P value = 0.039, and the absolute risk reduction (SP to CP) was 0.2% with 95% CI (0.001% to 0.46). There were only 4 reactions in this category, 0 in the control and 4 in the study period. The remaining reaction rates comparing SP to CP were not significant, but there were 3 that had higher overall rates than the other. The first, “mild allergic skin only reaction” was found in 18 of the 3839 total patients (0.5%). The second, “febrile non-hemolytic reaction” was found in 14 out of 3839 patients (0.4%). The third “multiple transfusion reactions” was found in 7 out of 3839 patients (0.2%). Overall, these are low rates of transfusion reactions and reflect similar rates published by others.¹⁵ 

Of those patients with a transfusion reaction, there was not a significant difference in the median number of platelet transfusions between control and study period (16 [n = 19; IQR = 7.0–37.5] versus 14 [n = 20; IQR = 2.0–37.0], respectively, P = .580). Of those patients with a transfusion reaction (22 in the control period, 19 in the study period), we found 46.2% had a malignant hematological diagnosis, 10.3% had a benign hematologic, 25.6% had other oncologic diagnoses, 10.3% had undergone a stem cell transplant, 5.1% had sepsis, and 2.6% had a traumatic injury.

This study, which includes over 3839 patients from 3 countries, is the largest known study set of infant and pediatric patients who received PI platelets. The goal of PI is to lower the risk profile of transfused blood products by inactivating any residual infectious pathogens present in the product, such as bacteria or viruses.¹⁶  The 2016 FDA licensure of the PI platelet product included in this study (INTERCEPT Blood System, Cerus Corp, Concord, CA, USA) demonstrated a significant reduction of many classes of pathogens, such as enveloped and nonenveloped viruses, gram-negative and -positive bacteria, parasites, spirochetes, and parasites; although there are certain pathogens that are not sufficiently inactivated by the PI process, such as parvovirus B19 and Hepatitis E.^15,17  Because it is scientifically accepted that the PI treated platelet products have a lower risk profile than conventional platelet products with respect to pathogen transmission, the clinical question about the use of PI have focused on how the therapeutic effectiveness (platelet count increment) and safety profile (acute transfusion reactions) of PI platelets compares to unmanipulated platelet products.

The study demonstrates that the use of PI platelets does not appear to impact the overall probability of pediatric patients having adverse events related to transfusion compared with conventional platelets. Our results are similar to results of other published studies in which the differences between acute transfusion reactions were not observed between conventional and PI platelets and should be considered reassuring for pediatric and neonatal providers who order platelet transfusions for their patients.^3,18  Schultz et al describe the use of conventional and PI platelets (PAS or not PAS) in infants and children aged 0 to 18 years at 1 center.¹⁹  Infants <1 year of age had similar platelet utilization patterns measured by the need for subsequent platelet transfusion 48 hours after the index transfusion. There was a statistically significant increase in platelet transfusions needed 48 hours later after PI platelets in children 1 to 18 years of age. There was no noticeable difference in red cell utilization (used as a surrogate measure for bleeding) or for febrile and allergic reactions between the 2 product types in patients 0 to 18 years of age. Collectively, these results are timely and important because the FDA in the United States has updated its guidance for decreasing the risk of bacterial contamination in platelets by promoting the use of PI platelets as a strategy.²⁰ 

The cost of transfusion reactions and the medications administered to prevent future reactions are substantial.²¹  In a study of >130 000 blood transfusions, given to pediatric and adult patients, the rate of acute transfusion reactions was higher in the pediatric patients compared with the adult patients (6.2 vs 2.1 per 1000 transfusions, respectively).^12,21  Overrepresentation of pediatric patients having acute transfusion reactions held true for 3 of the most common classifications: 1.9 vs 0.47 per 1000 transfusions for febrile nonhemolytic, 0.29 vs 0.078 per 1000 transfusions for hypotension, and 2.7 vs 1.1 per 1000 transfusions for allergic transfusion reactions, respectively.²²  The percent of patients with transfusion reactions was higher for those with platelet transfusions than any other blood product. This may be due to the plasma component of the platelet product, which is associated with transfusion reactions and is lower in platelets in PAS.^13,14,23  In France’s hemovigilance system, there were fewer hypersensitivity reactions after the implementation of universal PI platelets in PAS.²⁴  France also registered a significant difference with apheresis compared with buffy coat platelets, with buffy coat platelets having a significantly lower incidence of reactions. Our study included a mix of PAS PI platelets and plasma PI platelets, as well as a mix of apheresis and buffy coat platelets across the centers. We saw fewer mild allergic reactions in the study period compared with the control period. Although statistically significant, the numbers may not be clinically important. The small difference may be attributable to PI platelets themselves, or that PAS may have been used more consistently in the study period, although we did not measure this specifically. There were so few “moderate to severe allergic reactions” in the study (4 total: 0 in the control time period, 4 in the study time period), it is hard to determine if this represents a clinically meaningful difference. It is worth noting that the overall reaction percentage differed by center; adverse events occurred in 0% of patients in Denmark, 0.3% in Belgium, 4.1% at Stanford, and 1.5% at Vanderbilt (P < .001), with Stanford being the only site not to use PAS and the European sites being the only ones that used buffy coat preparations.

Our study had a number of limitations. Patient-level clinical and demographic data were not available for patients who did not have an adverse event, and only summary data were available for the patients who did have transfusion reactions. This is due to the retrospective nature of the study and the data protections our group was operating under using anonymized data under the General Data Protection Regulation in the European Union. Although study design had to be built to accommodate different nations and healthcare systems’ preexisting data sets, it is important to note that pediatric and neonatal patients were not included in the FDA’s licensing trial of the PI platelet used in this study; thus, alternative study designs are all that we have available to study PI platelets in children. Some pediatricians are concerned about the long-term impacts of using PI platelet products on the health of children who have long life expectancy particularly because of the active ingredient, a psoralen compound. Basic science studies have revealed that amotosalen is covalently bound to platelet membrane constituents by the photochemical reaction.²⁵  These lipid complexes are not removed during the photochemical removal phase of the PI process. Long-term chronic impacts of any treatment are hard to study in any setting, and this study does not measure long-term impacts. Despite the difficulty, it remains important to keep children included in randomized and observational studies such as postmarketing studies in which these impacts may be detectable if present.²⁶  Lastly, transfusion reaction reporting was based on passive reporting, because this is the standard approach in most hospitals and hemovigilance systems. Active and prospective hemovigilance by hospitals, blood centers, independent researchers, and government agencies would potentially find more reactions than are reported through standard passive mechanisms.²⁷ 

This study is a large retrospective case control study of pediatric patients receiving PI platelets. Our results reveal that PI platelets have a similar safety profile when given to children in the immediate posttransfusion time period compared with conventional platelets. There was a slight decrease in mild allergic reactions seen during the study period, which may not be clinically meaningful or could be similar to other studies revealing lower rates of mild allergic reactions when using PAS platelets. Overall, this information is reassuring for clinical decision-making. More research is needed to determine long-term safety and transfusion efficacy in the pediatric population.

This study was done by scientific members and guests of the BEST Collaborative.