Evolving TRALI risk across apheresis and whole blood platelet products in the United States: An 18‐year American Red Cross hemovigilance analysis

Abstract

Background

Transfusion‐related acute lung injury (TRALI) remains a leading cause of transfusion‐associated mortality in the United States. Donor‐based mitigation has reduced TRALI risk from plasma and apheresis platelets, but the residual risk associated with whole‐blood‐derived pooled platelets is not well defined.

Study Design and Methods

We retrospectively analyzed transfusion reactions reported to the American Red Cross (ARC) national hemovigilance program from 1 January 2006 through 31 December 2023. Reactions were adjudicated using Canadian Consensus Conference criteria; definite or highly probable TRALI cases (levels 5–6) were included. Component distribution data served as denominators. Trends were evaluated in relation to two mitigation efforts: male‐predominant plasma (2005–2006) and HLA‐antibody screening of previously pregnant apheresis‐platelet donors (2016).

Results

Over 18 years ARC distributed 92.7 million red blood cells, 23.5 million plasma units, and 17 million platelets—including 0.87 million pooled WB‐derived platelets, 11.7 million plasma‐suspended apheresis, 1.5 million PAS apheresis, and 2.9 million pathogen‐reduced (PR) platelet‐additive solution‐3 (PAS) apheresis platelets. We identified 393 high‐probability TRALI cases. Plasma‐associated TRALI declined from 20.0 to <4 cases per million units following male‐predominant policies; RBC risk remained stable (0.7–4.3/million). TRALI incidence for whole blood pooled platelets averaged 13.5/million, exceeding plasma‐suspended (4.4/million), PAS (0.8/million), and PLT PR PAS (0.5/million) apheresis platelets (p = .002 for differences among apheresis subtypes). Compared to apheresis PLT, the TRALI incidence for whole blood PLTS is increased by about 9–13/million.

Conclusion

Current mitigation strategies, which are not required for whole blood derived platelets in the United States, have substantially reduced TRALI risk from apheresis platelets. If whole blood‐pooled platelets production in the United States increases, consideration should be given to the addition of mitigation strategies—such as male‐predominant pooling and donor antibody screening—to align their safety profile.

Abbreviations

PAS

platelet additive solution

PLT

platelets

PR

Pathogen reduced

RBC

red blood cell

TRALI

Transfusion Related Acute Lung Injury

WB

Whole blood

1. INTRODUCTION

Transfusion‐related acute lung injury (TRALI) is an uncommon but serious transfusion complication, marked by sudden‐onset respiratory distress and bilateral pulmonary edema that cannot be fully explained by circulatory overload or other underlying causes. ¹ , ² , ³ , ⁴ The widely accepted “two‐hit” model proposes that TRALI arises from two sequential events. The first hit involves recipient‐specific factors—systemic inflammation, underlying disease, or recent surgical trauma—that prime neutrophils and the pulmonary endothelium. ¹ , ² , ³ , ⁴ The second hit is the transfusion itself: biologically active donor mediators, most often antibodies against human leukocyte antigens (HLA) or human neutrophil antigens (HNA), trigger an acute inflammatory cascade that produces capillary leak, pulmonary edema, and hypoxemia. ¹ , ² , ³ , ⁴

During the past two decades, blood centers have adopted multiple mitigation strategies, especially for plasma and apheresis platelets, to reduce the likelihood that donor units contain HLA or HNA antibodies. ⁵ , ⁶ , ⁷ One well‐established approach is preferential collection from male donors or from females who have never been pregnant or have tested negative for HLA/HNA antibodies. Thanks largely to these policies, the combined incidence of TRALI across all components is now generally below 1 per 100,000 transfused units. ⁷ , ⁸ , ⁹

Less is known about platelet components, whether apheresis or whole‐blood‐derived, because donor screening measures are often layered with manufacturing changes aimed at decreasing immunologic risk. ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ Platelet‐additive solutions (PAS) replace a substantial portion of plasma—often leaving ~35%—and should theoretically lower TRALI risk by decreasing exposure to antibody‐containing plasma. Pathogen‐reduction (PR) technologies such as INTERCEPT further modify platelets by inactivating pathogens and leukocytes. Yet robust, real‐world data comparing TRALI incidence among platelets in PAS, PR platelets, platelets stored entirely in plasma, and pooled whole‐blood‐derived platelets is lacking.

In the United States, pooled platelets derived from whole blood (WB) remain in use at some centers and do not undergo the same stringent donor screening as apheresis platelets or plasma. ¹² Because each pool combines contributions—and residual plasma—from multiple donors, the probability of introducing a clinically significant HLA or HNA antibody may be multiplied. Although fatal TRALI has been reported even with red‐cell units that contain only small amounts of plasma, ¹³ , ¹⁴ it remains unclear what the relative risk of a pool of typically 5‐donor platelets is compared to apheresis platelets manufactured with current mitigation strategies.

The few component‐specific estimates that exist highlight both progress and persistent uncertainty due to relatively small data sets. Vossoughi et al. surveyed five regional blood centers in nine Eastern and Midwestern states, reporting TRALI rates per 100,000 distributed units that fell from 6.4, 2.2, and 3.2 in 2007 for plasma, red blood cells, and platelets, respectively, to 0.43, 0.59, and 1.4 by 2017. ⁷ The authors emphasized that heterogeneous platelet preparations and varied mitigation practices hampered precise risk assessment. Mowla et al. reported relative risk results from only ~215 participating hospitals of the CDC hemovigilance network grouping a range of adjudications from possible to probable TRALI cases from only 1.2 million apheresis units and ~525 k WB platelets. ¹¹

Gaps in knowledge therefore remain: we still lack contemporary, evidence‐based TRALI rates for specific platelet products (pooled vs. apheresis, PAS vs. PR vs. plasma) and for whole‐blood components manufactured under current US practices. ⁹ Leveraging the uniquely large and detailed American Red Cross (ARC) database—encompassing nationwide distribution numbers and linked transfusion‐reaction reports—we aim to (1) determine component‐specific TRALI rates across modern platelet and plasma preparations, (2) quantify the contribution of manufacturing and donor‐selection variables, and (3) identify residual high‐risk scenarios that could inform targeted mitigation.

Accurate, component‐level TRALI estimates will guide clinicians and patients in transfusion decisions, support researchers designing interventional studies, and help policymakers refine donor‐selection criteria and product‐manufacturing standards, ultimately reducing the residual burden of TRALI in US transfusion practice.

2. METHODS

2.1. Data source

American Red Cross hemovigilance program. Between 2006 and 2023, the American Red Cross distributed 92,709,018 units of RBCs, 23,492,563 units of plasma, 2,206,615 units of cryoprecipitate, 16,961,599 units of platelets, of which 875,433 pooled whole‐blood‐derived platelets in plasma (PLT pooled, five donors per pool), which were pooled with five units in each pool, 11,672,537 apheresis‐plasma platelets (PLT plasma), 1,502,332 apheresis PAS platelets (PLT PAS), and 2,911,297 PR platelets in PAS (PLT PAS‐PR). No reports of TRALI related to cryoprecipitate were submitted.

2.2. Definitions and inclusion criteria

Suspected transfusion reactions, reported by hospitals, consistent with TRALI across these blood products were evaluated by the Red Cross physicians (medical directors) according to the Canadian Consensus Conference (CCC) definition and classified as TRALI or possible TRALI on a 6‐point probability scale (P1–P6). ¹⁵ , ¹⁶ Our physicians use both serologic and clinical results from cases to assign the imputability designation to each reported TRALI diagnosis. ¹⁰ These cases were recorded in the American Red Cross hemovigilance program for surveillance of blood donor and transfusion recipient complications. ¹⁰ The operational changes in 2007 that were required for distributing predominantly male‐donor plasma were previously described. ¹⁰ Cases involving transfusion of a single component type (e.g., plasma unit[s] only) are included in the primary analysis. Cases involving transfusion of more than one component type (e.g., RBC and plasma units) were excluded because of the uncertainty in attributing causation to a specific transfusion source when multiple component types (e.g., plasma and RBCs) were transfused within 6 hours of the reaction. ¹⁵ , ¹⁶

High‐probability or definite cases were coded as P5 or P6 by the regional medical directors and assigned the final classification. Low‐probability cases are rescinded by the hospital (P1) or do not support a diagnosis of TRALI or possible TRALI (P2) or more likely reflect other conditions (P3). Indeterminate (P4) cases cannot exclude TRALI or possible TRALI, but the clinical information does not meet the CCC definition (e.g., circulatory overload is present before transfusion, or the volume status is not well defined or documented). High‐probability cases are consistent with TRALI or possible TRALI according to the CCC definition and coded as P5 if a donor was not implicated or a HLA/HNA antibody‐containing component was not involved, or P6 if a donor was implicated (i.e., donor HLA‐antibody is a cognate or cross‐reactive [CREG]‐match for the patient's HLA type or a donor HNA‐antibody specificity is identified). The routine coding system does not distinguish TRALI from possible TRALI, which are collectively referred to as “high‐probability TRALI” or “probable TRALI” although reported clinical details are captured in our system. The presence (possible TRALI) or absence (TRALI) of alternate risk factors for the acute respiratory distress syndrome (ARDS), such as sepsis, pneumonia, cardiopulmonary bypass, or trauma, is described separately for all high‐probability cases.

For the present study on TRALI incidence, only cases classified as P5 and P6—those considered highly probable or definite—were included, ensuring a stronger correlation with transfusion as the underlying cause.

2.3. Platelet components: For the purposes of our analysis, we categorized platelets as follows

1. 100% plasma apheresis platelets (Plasma PLT)

2. PAS platelets (Apheresis stored in platelet‐additive solution)

3. PAS‐PR platelets (Apheresis stored in platelet‐additive solution with pathogen reduction)

4. Pooled platelets (whole‐blood‐derived, with five units combined into one “pool”, each containing a residual plasma of approximately 50 mL)

2.4. Statistical analysis

We calculated the incidence of high‐probability TRALI per million distributions (10⁶ units shipped to hospitals). Data for platelet products were separated into two epochs (2006–2016 and 2017–2023) to reflect the gradual adoption of pathogen reduction and the transition from 100% plasma to PAS as well as aligning results to adoption of required TRALI mitigation strategies. Linear mixed modeling, with product type as the fixed effect and distribution year as the random effect, was applied to detect significant differences in TRALI rates among product types over time.

3. RESULTS

3.1. Component distributions and overall TRALI surveillance

During the 18‐year surveillance period (2006–2023), the American Red Cross (ARC) distributed more than 92.7 million red blood cell (RBC) units, 23.5 million plasma units, and ~ 17 million platelet units (Table 1). Implementation of two sequential mitigation measures occurred during this surveillance period—conversion to predominantly male‐donor plasma in 2005–2006 and introduction of HLA‐antibody testing for all previously pregnant female apheresis‐platelet donors in calendar year 2016. ¹⁰ , ¹²

TABLE 1.

Summary of platelet, red blood cell, and plasma distributions and high‐probability incidence of TRALI (with CI) normalized to distributions over time.

Product	Distributions (2006–2023)	High‐probability TRALI incidence per 106 distributions, (n)
		2006–2016	2017–2023	2006–2023
Platelets
PLT plasma	11,672,537	6.6 (4.1, 9.1) n = 63	0.9 (0, 2.4), n = 4	4.4 (2.3, 6.4), n = 67
PLT PAS	1,502,332	0, n = 0 a	1.5 (0, 3.8), n = 3	0.8 (0, 19.4), n = 4
PLT PAS‐PR	2,911,297	NA b	0.6 (0, 1.5), n = 3	0.5 (0, 1.3), n = 3
PLT Pooled	875,433	14.9 (4.2, 25.6), n = 11	10 (0, 41.7), n = 1	13.6 (4.5, 22.7), n = 12
Red Blood Cells	92,709,018	2.1 (1.7, 2.5), n = 130	2.2 (1.0, 3.4), n = 66	2.1 (1.7, 2.5), n = 196
Plasma	23,492,563	5.6 (1.9, 9.3), n = 103	0.4 (0, 1.4), n = 3	3.6 (1.1, 6.1), n = 106

Note: n = number of TRALI cases identified; Confidence Intervals in parenthesis;

Abbreviations: PLT PAS: platelet‐additive solutions; PLT PAS‐PR: pathogen‐reduced apheresis platelets; PLT plasma: apheresis‐plasma platelets; PLT Pool: whole blood pooled platelet.

^a Excludes outlier value in 2013.

^b No distribution pre‐2016.

ARC physicians investigated 268 suspected TRALI cases in the pre‐mitigation baseline year (2006) and an average of 247 suspected cases per year in the post‐mitigation era (2008–2023). Of these, 393 reports (~76.0%) met the program's “high probability” definition after centralized medical review. Annual high‐probability case counts fell steadily for plasma but remained relatively stable for RBCs and platelets.

3.2. Incidence trends by component

High‐probability TRALI incidence per 1,000,000 distributed units is summarized in Table 1. The rate associated with plasma fell sharply—from 20.1 cases/10⁶ units in 2006 to 5.4/10⁶ cases by 2008—and has remained below 4/10⁶ cases since 2012. No comparable decline was seen for RBCs, whose incidence persisted at 0.7–4.3/10⁶ cases across the entire study interval, consistent with the absence of targeted TRALI‐mitigation strategies for RBC components.

Platelet‐related TRALI demonstrated marked heterogeneity. Before 2016, the hierarchy of risk was pooled whole‐blood‐derived platelets (PLT pooled)—14.9/10⁶ >100% plasma apheresis platelets (PLT plasma)—6.6/10⁶ > distributions of apheresis platelets suspended in platelet‐additive solution (PLT PAS)—0/10⁶ distributions (after excluding an outlier value for 2013). Linear mixed‐model analysis (fixed effect = platelet type; random effect = calendar year) confirmed a significant overall effect of platelet category on incidence (p < .01). Pair‐wise contrasts showed PLT pooled to carry a 9–13‐fold higher risk than any apheresis preparation (all p ≤ .01).

Following universal HLA screening of female platelet donors in 2016 and the progressive shift away from 100% plasma formulations, TRALI incidence from apheresis platelets declined by ~70%. There was a notable decrease in TRALI incidence in apheresis‐platelet products containing less plasma—particularly when pathogen reduction was also employed (PAS‐PR). However, this decrease was not as dramatic as initially anticipated, suggesting that removal of plasma alone confers only incremental safety benefits above and beyond existing donor‐based mitigations (e.g., HLA‐tested female donors, male‐only donors).

However, differences between PLT plasma, PLT PAS, and PLT PAS‐PR preparations were no longer statistically significant in the 2017–2023 sub‐analysis (p‐values for pair‐wise comparison about 0.9 for each comparison, Table 1). In contrast, with PLT pooled‐related TRALI cases reported between 2017 and 2019, its incidence was 10 (0, 41.7) per million units distributed despite shrinking utilization in that interval. Production of PLT pooled ceased entirely in 2020.

3.3. Model‐based pairwise comparisons

To quantify between‐component differences, we fit a mixed‐effects model (fixed effect: component type; random effect: calendar year) and report adjusted absolute risk differences per 10⁶ distributions using pooled whole‐blood platelets (PLT pooled) as the index comparator (Table 2). Compared with PLT pooled, the risk difference was −10.0 (95% CI −17.4 to −2.6; p = .002) for plasma, −11.5 (−18.9 to −4.0; p = .0002) for RBCs, −9.2 (−16.6 to −1.8; p = .006) for apheresis platelets in 100% plasma, −12.8 (−20.8 to −4.8; p = .0001) for apheresis PAS platelets, and −13.1 (−22.4 to −3.8; p = .001) for apheresis PAS‐PR platelets. Interpreted clinically, these differences correspond to ~9–13 fewer high‐probability TRALI cases per million units when transfusing apheresis‐platelet preparations (or plasma/RBCs) rather than PLT pooled over the full 2006–2023 interval. These findings align with the component‐specific incidences in Table 1 (e.g., 13.6/million for PLT pooled vs. ≤4.4/million for apheresis subtypes). In the post‐mitigation era (2017–2023)—after universal HLA screening of previously pregnant apheresis‐platelet donors and broader use of PAS/PR—the pairwise differences among apheresis subtypes were small and not statistically significant (all adjusted p ≈ 0.9), consistent with a floor effect once donor‐based mitigations and reduced plasma volumes were in place (Table 1). By contrast, PLT pooled retained a higher point estimate (10.0 [0, 41.7] per million) despite declining use until production ceased in 2020.

TABLE 2.

Estimates for relative effect measures of TRALI (with CI) among different types of platelet components, RBC, and plasma.

Component	Estimate	Adj upper	Adj lower	Adj P
Plasma vs PLT Pool	−10.0	−2.6	−17.4	0.002
RBC vs PLT Pool	−11.5	−4.0	−18.9	0.0002
PLT PAS vs PLT Pool	−12.8	−4.8	−20.8	0.0001
PLT PR vs PLT Pool	−13.1	−3.8	−22.4	0.001
PLT plasma vs PLT Pool	−9.2	−1.8	−16.6	0.006

Note: Male‐predominant plasma was implemented by our blood center in 2005–2006. HLA testing of female apheresis‐platelet donors was implemented during calendar year 2016. 1 = 100% plasma platelet production peaked in 2012 and dropped sharply after 2020; 2 = PAS platelet production was very low prior to 2017 and dropped sharply after 2020; 3 = PAS‐PR platelet production began in late 2016 and increased sharply after 2020; 4 = pooled whole blood platelet production peaked in 2012 and ceased in 2020 due to reduced hospital demand and pressure from operational challenges to produce a timed product.

Abbreviations: PLT PAS, platelet‐additive solution‐3; PLT PAS‐PR, pathogen‐reduced apheresis platelets in PAS3; PLT plasma, apheresis platelets in plasma; PLT Pool, whole blood pooled platelet.

3.4. Donor‐risk profile of pooled whole‐blood platelets

Table 3 details the distribution of TRALI risk donors within 538,467 pooled platelet units for which complete donor‐health‐questionnaire data were available. At least one donor reporting a recognized TRALI risk factor (prior pregnancy or transfusion) was present in 9.7% of pools; fully mitigated pools containing zero such donors accounted for 90.3% of distributions. Notably, 3–5 high‐risk donors were present in 2.4% of all pools, underscoring the multiplicative antibody‐exposure potential inherent to this component.

TABLE 3.

Summary of pre‐pooled platelet units for which ARC has information on the donor composition of the pool in which one or more donors indicated TRALI risk factor.

Whole blood‐derived platelets in a Pool where donor responded “yes” to TRALI risk in donor health history questionnaire	Pooled platelets distributed to hospitals	Percent
0	486,173	90.3%
1	21,730	4.0%
2	17,774	3.3%
3	9451	1.8%
4	2925	0.5%
5	414	0.1%

In summary, our apheresis‐platelet data suggested that male‐predominant plasma and HLA‐antibody screening of female apheresis‐platelet donors have together reduced plasma‐related TRALI by >90% and apheresis‐platelet‐related TRALI by a further >70%. Nevertheless, pooled whole‐blood platelets—manufactured without analogous donor or manufacturing safeguards—remained the leading residual source of TRALI until their discontinuation in 2020.

4. DISCUSSION

Our 18‐year hemovigilance analysis showed a consistently higher TRALI risk with pooled platelets derived from WB than with any other component per million products distributed (13.6/10⁶ vs. 4.4/10⁶ or less). Apheresis platelets have multiple, overlapping protections—donor HLA‐antibody screening, male‐only or HLA‐tested female plasma, and platelet‐additive solutions (PAS) that replace about 65% of plasma. Although PAS should theoretically lower TRALI by reducing exposure to donor antibodies, our data revealed no meaningful difference in TRALI incidence between PAS platelets and 100% plasma platelets (−3.6/10⁶, p = .9); pathogen‐reduced (PLT PAS‐PR) apheresis platelets likewise showed a nonsignificant decline (−3.9/10⁶, p = .9). These findings imply that donor‐based mitigations already captured most of the achievable risk reduction, leaving little added benefit from simply reducing plasma volume.

By contrast, US policy still permits WB‐derived pools of unscreened donors, each contributing ~50 mL of plasma. ¹² Every additional donor heightens the chance that at least one high‐titer HLA/HNA antibody enters the pool, explaining the elevated TRALI signal we observe. Canada and France have addressed this threat with two complementary safeguards: (1) a male‐predominant pooling rule that requires at least three of four (Canada) or four of five (France) donors in every pool to be male, and (2) systematic HLA‐antibody screening of female donors whose plasma will enter platelet pools. ⁵ , ¹⁷

Our study's strengths include centralized adjudication of all highly probable and definite TRALI cases, a uniform case definition, and linkage to distribution data for every component type, allowing us to track the real‐world impact of evolving mitigations—male‐only plasma, HLA screening, PAS, and PR—over nearly two decades. This is in contrast to published metanalyses of prevalence reports that often segregate reports on passive versus active but may not distinguish what level of certainty was used to report TRALI numbers and will include reports where low‐probability (i.e., possible) cases were included alongside high probability and definite adjudicated cases. ⁹ Limitations of our study include reliance on a single national blood organization (albeit one supplying a large share of the US inventory), passive hospital reporting that likely underestimated true TRALI rates and cannot be used to estimate true absolute rates but provided useful information about relative rates and changes in rates over time. Other limitations are exclusion of multicomponent transfusions, and incomplete visibility into patient mix or local surveillance rigor. Manufacturing and transfusion practice also changed across 18 years which further added heterogeneity.

The recent FDA approval of the Reveos® system (Terumo BCT)—which automates centrifugation and expression of WB into red cells, plasma, and a concentrated platelet product—has rekindled US interest in WB‐derived platelets. ¹⁸ Yet current FDA rules still demand that room‐temperature WB be processed within 8 hours, whereas Canada, France, and several other European nations allow a 24‐hour hold. ¹² Adopting a similar 24‐hour window would ease night‐shift staffing, reduce wastage, and make WB platelets operationally more attractive, provided simultaneous safety measures are considered.

Suggested safety measures may include: (1) routine HLA/HNA antibody testing of all female platelet donors and (2) a male‐predominant pooling rule—at least four of five donors, or all but one, must meet a mitigation criterion in every pool. These steps mirror the successful approaches in Canada and France.

In sum, the marked differential risk between pooled WB and apheresis platelets was likely a true safety signal, not a reporting artifact. Aligning pooled‐platelet safeguards with those already standard for apheresis platelets—and embedding such requirements into any new WB‐platelet workflows—should reduce their TRALI risk. Coupled with ongoing hospital education to recognize and report TRALI, these evidence‐based actions would keep US transfusion practice on par with international best practice and further protect patients from a preventable, potentially fatal reaction.

5. CONCLUSIONS

Despite major advances in mitigating TRALI risk for RBC and plasma transfusions—both now characterized by very low incidence rates—our data highlight that WB‐derived pooled platelets remain a significant source of TRALI risk. Reducing plasma volume through PAS or employing pathogen reduction offers only incremental gains over existing donor screening measures, underscoring the critical need to adopt equally stringent mitigation strategies for pooled platelets. Routine HLA/HNA screening and/or using predominantly male donors for pool formation could align safety measures with those already implemented for apheresis platelets, reducing TRALI risk across all platelet components.

CONFLICT OF INTEREST STATEMENT

The author has disclosed no conflicts of interest.

Young PP, Singh U, Webb J, Looney MR. Evolving TRALI risk across apheresis and whole blood platelet products in the United States: An 18‐year American Red Cross hemovigilance analysis. Transfusion. 2025;65(12):2317–2323. 10.1111/trf.18457

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.