Major ABO-incompatible platelet transfusions are associated with brain ischemia after intracerebral hemorrhage

Abstract

Background:

Major ABO-incompatible platelet transfusions are associated with poor intracerebral hemorrhage (ICH) outcomes, yet drivers for this relationship remain unclear. Brain MRI ischemic lesions after ICH are neuroimaging biomarkers of secondary brain injury and are associated with poor outcomes. Given that ABO-incompatible platelet transfusions can induce immune complex formation, thrombo-inflammation, and endothelial barrier disruption, factors that could exacerbate cerebral ischemia, we explored whether major ABO-incompatible platelet transfusions are risk factors for ischemic lesions on brain MRI after ICH.

Methods:

Adult patients admitted to a tertiary-care academic center between 2009 and 2016 who received a single platelet transfusion within 24 hours of admission after an ICH, had available donor/recipient ABO data, and brain MRI during the hospitalization were analyzed. Adjusted regression models evaluated relationships between major ABO-incompatible platelet units and MRI ischemic lesions.

Results:

A total of 40 patients were included in the study. The mean age was 67.1 (SD 14.1) and 37.5% were female. Twenty percent of patients received a major ABO-incompatible platelet unit. Major ABO-incompatible platelet transfusions were associated with increased odds of MRI ischemic lesions after adjusting for ICH severity (adjusted OR 9.2, 95%CI 1.3–62.7).

Conclusion:

Our exploratory findings suggest that major ABO-incompatible platelet transfusions may contribute to secondary brain injury after ICH. Further work is needed to assess whether avoiding major ABO-incompatible platelet transfusions can prevent secondary brain injury burden and improve ICH outcomes.

Graphical Abstract

Introduction

Although platelets are critical in establishing hemostasis in life-threatening hemorrhage, platelet transfusions in intracerebral hemorrhage (ICH) patients have not uniformly shown benefit and have even demonstrated mortality risk in specific clinical contexts^1,2. Unlike red blood cell and plasma, platelet transfusions are often given irrespective of ABO status given inventory constraints and historic safety data³. However, emerging evidence suggests that major ABO-incompatible platelet units may be associated with reduced transfusion efficacy and mortality risk, particularly in ICH patients^4,5. While the underlying driver(s) for these associations in ICH remain unknown, separate work has demonstrated that ABO-incompatible platelet units can induce immune complex formation, thromboinflammation, and endothelial barrier disruption^6–8. These processes may drive ischemic secondary brain injury burden in ICH patients, which is detectable on brain MRI and consistently associate with poor ICH outcomes^9–12. Yet, it is unknown whether acute major ABO-incompatible platelet transfusion exposures associate with secondary brain ischemia after ICH. To assess potential drivers for platelet transfusion harm in this patient population vulnerable to transfusion medicine practices we explored whether major ABO-incompatible platelet transfusions are associated with MRI-detected brain ischemia after ICH.

Methods

The data that support the findings of this study are available from the corresponding author upon reasonable request. This study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines.

Consecutive spontaneous adult ICH patients admitted to Columbia University Irving Medical Center were prospectively enrolled in the Intracerebral Hemorrhage Outcomes Project (ICHOP) database, a single-center observation cohort study, in an institutional review board-approved study (with consent provided by the patient or family). Patients enrolled between 2009 and 2016 who received a single platelet transfusion within the first 24 hours of admission, had available donor/recipient ABO blood type, and had a brain MRI during hospitalization were included. Patients with secondary ICH (due to vascular malformation, hemorrhagic tumor, or ischemic stroke with hemorrhagic transformation), primary intraventricular hemorrhage, preceding anticoagulant use, or laboratory evidence of baseline coagulopathy were excluded. Patients who received more than one platelet unit during the ICH hospitalization were also excluded.

Patients were managed according to the contemporary American Heart Association guidelines¹³ with local ICH treatment protocols and platelet transfusion practices described previously⁵. ABO blood typing was obtained on admission in anticipation for the need of any urgent blood product use. However, per standard practice, platelet transfusions were not required to be ABO-matched. Platelet and donor ABO types were recorded from the blood bank laboratory information system, and transfusion exposure was classified as either major ABO-incompatible or ABO-compatible. Major ABO incompatibility was defined as a recipient having preformed antibodies against donor platelet ABO antigens (e.g., a type O recipient receiving platelets from a type A donor). ABO-compatible transfusions were defined as either identical donor and recipient blood types or minor mismatches, where donor plasma contained antibodies against recipient red cell antigens (e.g., a type O donor transfused into a type A recipient), consistent with prior studies^5,14. In instances of bidirectional incompatibility, where both major and minor incompatibilities were present from a single platelet unit transfusion (e.g., a type A recipient receiving platelets from a type B donor), units were classified as major ABO-incompatible (see Table 1).

Table 1:

Types of platelet unit incompatibility

		Recipient blood type
		O	A	B	AB
Donor blood type	O	Identical	Minor	Minor	Minor
	A	Major	Identical	Major (bidirectional)	Minor
	B	Major	Major (bidirectional)	Identical	Minor
	AB	Major	Major	Major	Identical

Brain MRIs were obtained per clinical practice, and the clinical MRI obtained following platelet transfusion during the ICH hospitalization was used for analysis. Images were reviewed and adjudicated by blinded study personnel. Remote ischemic lesions were defined as foci of diffusion restriction located >10 mm away from the hematoma on diffusion weighted imaging (DWI) 1000 MRI sequence with corresponding apparent diffusion coefficient (ADC) correlate. MRI ischemic lesion presence was assessed as a binary radiographic outcome and times from admission to MRI acquisition (in days) were recorded.

Given prior work linking acute hemoglobin decrements with MRI ischemic lesions after ICH⁹, as well as separate work showing evidence of hemolysis related to ABO-immune complexes⁸, we also explored hemoglobin decrements as a laboratory outcome. Complete blood count (CBC) data were obtained daily per clinical care for ICH patients and were leveraged to assess serial hemoglobin concentrations. Changes in hemoglobin concentration were quantified and assessed as a continuous laboratory outcome variable calculated from baseline CBC and the CBC on day of MRI acquisition.

Intergroup differences (major ABO-incompatible vs minor ABO-incompatible or ABO-identical platelet transfusion) were determined using analysis of variance or Mann-Whitney U tests for continuous variables and chi-squared tests for categorical variables. Logistic regression models assessed relationships of major ABO-incompatible platelet transfusions with MRI ischemic lesions, adjusting for ICH severity (via ICH score^15,16). Relationships between ABO-incompatible platelet transfusions and hemoglobin decrements were also assessed using linear regression models adjusting for ICH severity. Sensitivity analyses were performed with separate models adjusting for demographics, patient blood type (O vs non-O), and other known covariates of MRI ischemic lesions, including time from admission to brain MRI acquisition and change in systolic blood pressure (SBP) over first 24 hours of admission^11,12,17,18. Additional exploratory analyses assessed platelet transfusion as a categorical variable using ABO-identical platelet units as the reference. Separate logistic regression models were constructed to compare major ABO-incompatible versus ABO-identical units and minor ABO-incompatible versus ABO-identical units, each adjusted for ICH severity. Statistical significance was set at p <0.05. Analyses were performed using SPSS (v28; IBM).

Results and Discussion

A total of 40 ICH patients were included in the analysis (see Figure 1). The mean age was 67.1 (SD 14.1), 37.5% (15 patients) were female, and 20% (8 patients) received a major ABO-incompatible platelet unit. Median time from admission to brain MRI acquisition was 2 days (IQR 1–3.75). No significant differences in baseline demographics or ICH characteristics were noted between patients receiving major ABO-incompatible vs ABO-compatible platelet units (see Table 2). Patients receiving a major ABO-incompatible platelet unit were more likely to have MRI ischemic brain lesions (62.5% (5 patients) vs 21.9% (7 patients)). In multivariable regression analysis, major ABO-incompatible platelet transfusions were associated with ischemic lesions (adjusted OR 9.2, 95%CI 1.3–62.7). Sensitivity analyses adjusting for sex (adjusted OR 6.0, 95%CI 1.1–31.7), race (adjusted OR 8.4, 95%CI 1.3–54.1), patient blood type (adjusted OR 16.7, 95%CI 1.4–204.0), time to MRI acquisition (adjusted OR 5.5, 95%CI 1.0–29.4), and SBP change over the first 24 hours (adjusted OR 4.5, 95%CI 0.80–24.9), did not significantly alter this association. Furthermore, ICH patients receiving major ABO-incompatible platelet transfusions had greater hemoglobin decrements (−3.2 g/dL vs −1.9 g/dL), with a trend toward statistical significance. We continued to identify these relationships in our adjusted linear regression models (coefficient beta −0.29, 95%CI −2.42 to −0.28). However, we did not identify relevant relationships of minor ABO-incompatible platelet units with either MRI ischemic lesions or hemoglobin decrements.

Figure 1. Flow diagram of patient inclusion and exclusion.

ICHOP: Intracerebral Hemorrhage Outcomes Project; ICH: intracerebral hemorrhage; IVH intraventricular hemorrhage; INR: international normalized ratio; MRI magnetic resonance imaging.

Table 2:

Intergroup baseline characteristics

	Major ABO-incompatible platelets N=8	ABO-compatible platelets N=32	p value
Demographics: mean (SD) or N (%)
Age	68.8 (11.6)	66.7 (14.8)	0.713
Female sex	3 (37.5%)	12 (37.5%)	1.000
White race	3 (37.5%)	10 (31.3%)	0.416
Past medical history: N (%)
Hypertension	8 (100.0%)	22 (68.8%)	0.165
Diabetes	3 (37.5%)	6 (18.8%)	0.348
Hypercholesterolemia	3 (37.5%)	7 (22.6%)	0.399
Coronary artery disease	2 (25.0%)	7 (21.9%)	1.000
Medications: N (%)
Anti-platelet use	6 (75.0%)	21 (65.6%)	1.000
ICH characteristics: median (IQR) or N (%)
Hematoma volume (mL)	9.2 (3.0–23.8)	10.5 (4.1–27.5)	0.565
ICH score	1.5 (0.3–2.0)	1.0 (1.0–2.0)	0.916
Location			0.708
Lobar	3 (37.5%)	16 (50.0%)
Basal ganglia or thalamus	5 (62.5%)	12 (37.5%)
Cerebellum	0 (0.0%)	2 (6.3%)
Brainstem	0 (0.0%)	1 (3.1%)
Intraventricular hemorrhage	3 (37.5%)	14 (45.2%)	1.000
Admission laboratory values: mean (SD)
Admission Hemoglobin (g/dL)	13.5 (2.2)	14.0 (1.9)	0.548
Hemoglobin at the time of brain MRI (g/dL)	10.3 (3.01)	12.1 (1.7)	0.031
Hemoglobin decrement (g/dL) between admission and MRI	−3.2 (2.1)	−1.9 (1.7)	0.074
Platelet transfusion characteristics: N (%)
Patient ABO blood type			0.011
A	0 (0.0%)	16 (50.0%)
AB	0 (0.0%)	3 (9.4%)
B	0 (0.0%)	2 (6.3%)
O	8 (100.0%)	11 (34.4%)
MRI characteristics: median (IQR)
Time from admission to brain MRI	3 (1–7.8)	2 (1–2.8)	0.389

Despite platelet transfusion risks², these transfusions remain a key part of acute ICH treatment given the critical importance of rapidly establishing hemorrhage control in the setting of platelet dysfunction^1,19,20. Thus, understanding the risks driving platelet transfusion harm is critical to improving treatment paradigms moving forward. Our exploratory findings provide needed initial insights, as we identified that major ABO-incompatible platelet unit exposures may aggravate ischemic secondary brain injury, a neuroimaging marker known to worsen ICH outcomes^{11,12,17,18,21}. These findings highlight previously unknown potential drivers for prior observations linking ICH morbidity risk with these types of platelet units^2,4,5.

While our single-center findings necessitate replication and further mechanistic insights, it is biologically plausible that our relationships were driven by the major ABO-incompatible transfusion exposure. Platelet transfusions can have pro-inflammatory effects²², with ABO-incompatible platelet units associated with immune complexes that can cause platelet destruction, endothelial barrier disruption, and even hemolysis in non-ICH settings⁸. In our ICH cohort, where central and peripheral thromboinflammation is known to occur and blood-brain barrier breakdown is a hallmark of the disease²³, we noted that major ABO-incompatible platelet transfusions associated with larger hemoglobin decrements, a known driver for brain MRI ischemic lesions⁹. However, we acknowledge that hemoglobin decrements may result from diverse causes and may serve as a shared pathway to ischemia, rather than a mediator directly attributable to ABO-incompatible transfusion. In ICH patients who are susceptible to deleterious inflammation and blood-brain barrier breakdown, the exposure to major ABO-incompatible transfusions may aggravate thromboinflammation peripherally (and potentially centrally) thereby inducing anemic states that impair cerebral oxygen delivery leading to secondary brain injury burden. However, it should be highlighted that previously described hemolysis from ABO-immune complexes are thought to be from minor mismatches. Given that we identified greater hemoglobin decrements and DWI lesions only in our major ABO-incompatible platelet transfusion group, it is uncertain whether our hemoglobin decrements reflect a “bystander” hemolysis or non-hemolysis mediated mechanisms. It could be speculated that the infusion of minor incompatible units is less damaging because there are fewer immune complexes formed, as the anti-A and/or anti-B antibodies (from platelet plasma) per recipient cell is low, compared with major ABO-incompatible platelet transfusions, where the density of antibodies (from the recipient) per transfused platelet is higher and the opportunity for complement activation may be much greater. To this extent, prior ICH work has not identified a clinical impact of minor ABO-incompatible platelet transfusions on ICH outcomes, unlike the large mortality effect sizes seen with major ABO-incompatible transfusions^2,3. Yet, further work is needed to test these relevant hypotheses.

Though our exploratory study had strengths of granular ICH hospitalization data and clinically adjudicated MRI neuroimaging assessments providing a unique investigative platform for insights into this underexplored area, several limitations require mention. First, our small sample size limits causal inference and necessitates further studies to assess the generalizability of our findings across different ICH patient cohorts who receive different types of medical and surgical treatments which may independently affect outcomes²⁴. This also applies to the assessment of different platelet transfusion exposures to various donor blood groups, to evaluate both major and minor ABO-incompatibility effects. Second, our study lacked clinical laboratory tests of interest (e.g., hemolysis) or research-based assessments of thromboinflammation and/or immune complex activation states to clarify underlying mechanisms. Third, the specific clinical indications for platelet transfusion were not systematically recorded and may have influenced outcomes, representing an unmeasured source of bias. Fourth, our study focused on patients undergoing MRI neuroimaging for clinical indications, leaving it unclear whether our findings would generalize across patients who did not receive MRI assessments. Subsequently, future prospective studies with protocolized neuroimaging and laboratory assessments in these patients are needed to clarify the role of major ABO-incompatible platelet units on ICH morbidity and mortality. Finally, it is worth noting that we expectedly identified that major ABO-incompatible platelet units were given to type O ICH patients. While we have not previously identified that type O blood is associated with worse ICH characteristics, complications, or outcomes^25,26, it is known that individuals with type O blood have lower levels of factor VIII and von Willebrand factor²⁷. Thus, it is possible that underlying ABO blood status or other unmeasured factors/confounding played a role in our study observations, necessitating further mechanistic follow-up for our findings.

Our exploratory findings highlight that platelet transfusions may have varying risk profiles based on transfusion characteristics. And because these risks may be particularly pervasive in ICH patients, clinicians (both ICH care providers and transfusion medicine specialists) should recognize them for future efforts to determine whether avoiding major ABO-incompatible platelet units can improve ICH outcomes.

Abbreviations

ADC

apparent diffusion coefficient

CBC

complete blood count

DWI

diffusion-weighted Imaging

ICH

intracerebral himorrhage

ICHOP

Intracerebral Hemorrhage Outcomes Project

INR

international normalized ratio

IVH

intraventricular hemorrhage

MRI

magnetic resonance imaging

SBP

systolic blood pressure

STROBE

Strengthening the Reporting of Observational Studies in Epidemiology