Alloimmunisation in transfused patients: serial cross-matching in a population of hospitalised cats

Abstract

Objectives

Cross-matching is currently recommended as part of pre-transfusion testing for repeat transfusions in cats 4 days after having received an initial transfusion. This prospective study determined when and if cats developed positive cross-match (CM) results after having been transfused with AB-compatible blood.

Methods

Donors were selected according to standard transfusion safety protocols. Twenty-one hospitalised anaemic recipients (blood type A: n = 20; blood type B: n = 1) received 1–4 (median 2) whole blood transfusions (WBTs) over 1–6 days (median 2) in 33 transfusion instances. The tube CM method, including major, minor and recipient control, was employed. Macroscopic and microscopic agglutination reactions were evaluated according to a predetermined scale. CM tests with a positive recipient control could not be evaluated.

Results

No signs of an acute transfusion reaction were observed. A total of 63 CMs were performed. In one cat with immune-mediated haemolytic anaemia the CM could not be evaluated (positive recipient control). The minor CM was negative in all cases. Fifteen of 20 cats had a negative major CM (MCM) 1–12 days (median 5) after their first transfusion. A positive MCM was observed in five cases after 2–10 days (median 5) post-first WBT. These five cats had received a total of 1–4 (median 2) WBTs. Cats with a negative MCM had received 1–3 (median 2) WBTs. In 51.5% (17/33) of transfusion instances, the cat’s haematocrit increased as expected, with cats with a positive MCM at 40% (4/10) vs 56.5% (13/23) if MCM was negative.

Conclusions and relevance

Twenty-five percent (5/20) of the feline recipients likely developed alloantibodies against erythrocyte antigens outside of the AB system as early as 2 days post-first WBT. This adds data to the recommendation to include cross-matching in pre-transfusion screening tests.

Introduction

Transfusion medicine has been established as a safe and effective type of patient care. An increased focus on patient safety has helped identify and control various risk factors with which it is associated, limiting the incidence of acute and delayed immunological and non-immunological transfusion reactions. Modern haemovigilance and blood-banking guidelines have been in place and are continually improving in human medicine. Such efforts are being mirrored in veterinary medicine as permitted by economic and logistical considerations, allowing such therapy to play an important role in feline healthcare.^1–4

One aspect of increased safety in transfusion medicine is the establishment of routine pre-transfusion testing. ⁵ A number of point-of-care blood typing systems have been developed and evaluated for use in feline patients.^6–10 However, because blood-typing devices only detect specific known antigens on both recipient and donor red blood cells (RBCs), they can neither account for antigens outside of the AB system nor for alloantibodies present in the recipient.

Post-transfusion alloimmunisation is a common complication in human medicine. It is, to date, unavoidable and can pose a safety hazard for previously transfused human patients. ¹¹ Such alloimmunisation has also been documented for dogs and horses in veterinary medicine.^12–14 Cross-matching is the standard method for the detection of such serological incompatibilities between recipient and donor. ¹⁵

The current indication for a cross-match (CM) in dogs and cats is when a patient has an unknown transfusion history, has shown a previous transfusion reaction and/or if a prior blood transfusion was administered ⩾4 days prior to a planned transfusion.^16–18 Some authors have expressed the notion that routine cross-matching ought to be introduced for all pre-transfusion testing in cats,^15,19 owing to the clinically relevant naturally occurring antibodies in that species, ²⁰ as well as the presence of potentially unknown blood groups, as evidenced by the discovery of the Mik erythrocyte antigen. ²¹

To our knowledge, no studies with serially performed cross-matching have been undertaken that determine the status of alloantibody presence in both previously transfused or not previously transfused feline patients, and neither has a point been determined at which such patients may eventually develop alloantibodies after their first transfusion.

This paper reports on a study designed to document the occurrence of positive CM results, as well as the usefulness of routine cross-matching in feline transfusion patients presented at the Small Animal Clinic at the Free University of Berlin.

Materials and methods

Study population

The subjects of this prospective clinical study were client-owned anaemic cats that received at least one AB-compatible whole blood transfusion (WBT) while hospitalised at the Small Animal Clinic. Owner consent was routinely given at patient intake for the use of surplus samples for research purposes; therefore, no further approval for this study was needed. Inclusion criteria were the availability of a minimum of 250 μl of whole blood for cross-matching from up to 2 days prior to the date of the first transfusion, the availability of donor blood for cross-matching and that none of the patients had received any intravenous blood products prior to their first transfusion at the clinic. Cross-matching took place across 23 months.

Following in-house transfusion safety standards, blood donors were healthy client-owned pets no more than 8 years of age that weighed no less than 4 kg and were kept indoors exclusively. Regular vaccinations and deworming were expected, as was a known medical history. Donor screening included a physical examination with an emphasis on auscultation, routine blood work, AB blood typing and testing for feline leukaemia virus (FeLV) and feline immunodeficiency virus (FIV) (SNAP combo plus FeLV antigen/FIV antibody test; IDEXX Laboratories). Donors were sedated with a combination of ketamine (5–6 mg/kg IM) and midazolam (0.1 mg/kg IM). Whole blood, 6–7 ml per kg body weight, was obtained from the donor by jugular venepuncture while in ventral recumbency. The blood was collected into an open system using antiseptic technique into syringes prefilled with citrate, phosphate, dextrose and adenine (CPDA-1) solution and transferred into 150 ml paediatric blood-collection bags (Fenwal). Blood units were stored for no more than 21 days in a designated refrigerator kept at a temperature of 4 ± 2°C. A 1–2 ml aliquot of the CPDA-1-preserved donor blood was set aside for later cross-matching.

Transfusions and pre-transfusion testing

Pre-transfusion blood typing was routinely performed by trained laboratory personnel on donor and recipient blood samples, using the slide method as described elsewhere.^2,6 An in-house modification of the University of Pennsylvania tube method for blood typing was used in cases where results were unclear or seemed to indicate blood type AB.^6,10

Transfusions were performed on anaemic patients at their attending veterinarian’s discretion rather than strict laboratory cut-offs; both clinical status and, among other laboratory results, haematocrit (Hct) were taken into account. Patients received WBTs through a peripheral venous catheter placed into the cephalic or medial femoral vein or through a central venous catheter placed into the jugular vein. The method of delivery was through a standard, gravity-driven transfusion line with an integrated 200 µm filter (Sangofix; B Braun). Recipients were monitored for signs of acute transfusion reactions by recording heart rate, respiratory rate, rectal temperature and mucous membrane status at regular intervals. ²²

Cross-matching

Routine pre-transfusion testing included cross-matching (major, minor and recipient control) when a patient had received a transfusion more than 3 days prior to a planned transfusion. In all other cases, cross-matching was part of a pilot study evaluating the usefulness of including cross-matching in routine pre-transfusion testing in anaemic patients, in whom monitoring required regular and yet volume-conscious sampling. Cross-matching was performed by the same investigator according to an established in-house protocol, as well as following the most recent published data for a simplified tube procedure common in veterinary medicine.^16–18,23 Whole blood from both donor (EDTA-anticoagulated sample) and recipient (CPDA-1-preserved sample) was washed prior to testing as follows: plasma was first extracted by centrifuging at 385 g for 2 mins. The remaining RBCs were then washed by mixing them with 500 μl phosphate-buffered saline (PBS) solution and subsequently centrifuging the suspension at 96 g for 1 min. This process was repeated three times.

Both recipient and donor plasma was checked for degree of haemolysis according to the following scale: no haemolysis visible to the human eye (0) to haemolysis that does not allow for visual differentiation of RBCs from plasma (4+) (Table 1). Substrates for cross-matching were plasma and RBCs suspended at a 3–5% concentration in PBS solution. The MCM required 40 μl recipient plasma gently mixed with 20 μl of the prepared 3–5% donor RBC solution. The minor CM used 40 μl donor plasma gently mixed with 20 μl prepared 3–5% recipient RBC solution. In order to control for auto-agglutination, 40 μl recipient plasma was likewise mixed with 3–5% recipient RBC solution. The substrates were then incubated at 37°C for 15 mins. After a 15 s burst of centrifugation at 865 g, the supernatant was examined for haemolysis and compared with that of the plasma originally used. The pellet was then resuspended through gentle agitation while checking for agglutination in the tube against a white background, as well as on a slide under a microscope, if macroscopic examination was negative. Macroscopic and microscopic agglutination were recorded according to the following scale: no agglutinates (0) to 1–2 large agglutinates with clear plasma (4+) (Table 2). Microscopic assessment required the examination of the suspended RBCs within 60 s of placing a drop of RBC solution on the slide and agitating the slide in order to view RBCs and possible RBC agglutinates in motion. The scales used here were adapted from a study assessing degree of agglutination in blood typing. ⁶

Table 1.

Scale for the visual assessment of the degree of plasma haemolysis

0	Plasma is nearly transparent
(+)	Plasma is slightly discoloured
1+	Plasma is light red
2+	Plasma is red
3+	Colour of plasma is close to colour of RBCs
4+	Colour of plasma cannot be visually differentiated from RBCs

RBCs = red blood cells

Table 2.

Scale for the microscopic assessment of the agitated red blood cell (RBC) solution in the cross-match tube method

0	No agglutinates visible upon agitation of RBCs
(+)	A few tiny agglutinates with most of the RBCs in suspension
1+	Many small agglutinates along with RBCs in suspension
2+	Some larger agglutinates with many small agglutinates
3+	Several large agglutinates with clear plasma
4+	1–2 large agglutinates with clear plasma

Tests took place, when possible, prior to and serially every 2 days after a blood transfusion, depending on the availability of samples.

A CM was considered positive if haemolysis or an agglutination reaction was observed macroscopically. If no reaction was evident, a microscopic evaluation was always included. If the recipient control displayed auto-agglutination, the CM was considered invalid.

In all cases, the Hct before transfusion was compared with that measured routinely 10–72 h after transfusion and compared with the expected Hct based on a formula suggested elsewhere.^2,24,25

Statistical analysis

Calculations for collected data were performed using commercial data analysis software (IBM SPSS Statistics version 23.0.0.2, Microsoft Excel version 15.20). Descriptive data analysis was performed on the parameters: patient signalment, underlying disease, indication for transfusions, number of transfusions per recipient, number of donors per recipient, storage time for blood units, volume of blood units, number of CMs and number of positive MCMs. A mixed linear regression analysis was used to determine relationships between the occurrence of a positive MCM, as a risk factor, and Hct development, a measure of transfusion success, as the outcome. The individual animal was included as a random effect in the model. A P value <0.05 was considered significant.

Results

Twenty-one feline patients aged 1–16 years (median 8 years) were included in the study. Their body weight at the time of transfusion ranged from 2.5–8.3 kg (median 4.5 kg) and various breeds were represented. The majority were domestic shorthairs (n = 15), in addition to the following pedigree or pedigree-mix cats: British Shorthair (n = 2), Maine Coon (n = 1), Maine Coon mix (n = 1), Selkirk Rex (n = 1) and Turkish Van (n = 1). The patients included 17 males (16 neutered and one intact) and four spayed females. Disease distribution was as follows: blood loss anaemia (regenerative, n = 4), kidney disease with non-regenerative anaemia (n = 4), anaemia of inflammatory disease (non-regenerative, n = 4), FeLV infection with non-regenerative anaemia (n = 3), immune-mediated haemolytic anaemia (IMHA) (non-regenerative, n = 2), neoplasia with non-regenerative anaemia (n = 2), diabetes mellitus with regenerative anaemia (n = 1) and pure red cell aplasia (PRCA) with non-regenerative anaemia (n = 1). Twenty cats belonged to blood type A and one domestic shorthair cat had blood type B.

Pre-transfusion Hct ranged from 6.5–23.5% (median 14.2%). The patients received 1–4 (median 2) AB-compatible WBTs from 1–5 donors (median 2) over 1-–6 days (median 2 days). The volumes administered ranged from 10 to 30 ml (median 20 ml), which were used alone or in combination with other units, amounting to a transfusion volume per patient and day (henceforth labelled as transfusion instance) that ranged from 1.8–18 ml/kg body weight (median 5.6 ml/kg).

No clinical signs of an acute transfusion reaction, as evidenced by changes in heart rate, respiratory rate, rectal temperature and/or mucous membrane status were recorded for the duration of the study.

Sixty-three CMs were performed in total, each CM including minor and major cross-matching, as well as a recipient control. One CM took about 30 mins to complete. The minor CM was negative in all 63 cases.

Fifteen of 21 cats had a negative MCM between 1 and 12 days (median 5 days) after initial transfusion. The CM in one cat with IMHA was not valid because it displayed a positive recipient control. Five of 20 (25%) cases had a positive MCM 2–10 days (median 5 days) after the initial blood transfusion given to transfusion-naive patients, with macroscopic agglutination ranging from 0 to 1+ and microscopic agglutination from (+) to 2+ (Table 3). Positive CMs were not found in transfusion-naive patients. Cats with a positive MCM had anaemia of inflammatory disease (n = 2), blood loss anaemia (n = 1), FeLV infection with non-regenerative anaemia (n = 1), PRCA (n = 1); they received a total of 1–4 (median 2) blood transfusions from 1–4 donors (median 2) (Table 3). Patients with a negative MCM received 1–3 (median 2) blood transfusions (Table 4). Those with a positive MCM received 2.8–9 ml/kg (median 5.8 ml/kg) of whole blood per transfusion instance, while those with a negative MCM received 1.8–18 ml/kg (median 5.1 ml/kg) per transfusion instance. The storage age of units that had been given to positive MCM patients ranged from 0–21 days (median 7 days).

Table 3.

Cases with a positive major cross-match (MCM): diagnosis, number of blood transfusions instances (BT), number of donors, number of cross-matches (CMs), time to first positive MCM, grade of agglutination and post-transfusion haematocrit (Hct) development (upon final transfusion, where more than one transfusion was given)

Case	Underlying or concurrent disease	BT (n)	Donors (n)	CMs (n)	Time between first BT and first positive MCM (days)	Macroscopic agglutination (grade)	Microscopic agglutination (grade)	Hct increase as expected?
3	Haemorrhage	2	2	4	10	1+	2+	Yes
11	FeLV infection	1	1	5	7	0	2+	Yes
13	Inflammatory disease	1	1	2	2	0	1+	No
20	Pure red cell aplasia	2	2	4	4	0	(+)	No
21	Inflammatory disease	4	4	4	5	0	2+	No

FeLV = feline leukaemia virus

Table 4.

Cases that did not display a positive major cross-match (MCM): diagnosis, number of blood transfusions (BT), number of donors, number of cross-matches (CM) and post-transfusion haematocrit (Hct) development (upon final transfusion, where more than one transfusion was given)

Case	Underlying or concurrent disease	BT (n)	Donors (n)	CMs (n)	Hct increase as expected?
1	Neoplasia	1	2	2	No
2	Kidney disease	2	3	3	No
4	Kidney disease	2	2	5	No
5	Kidney disease	1	2	3	Yes
6	Inflammatory disease	3	2	5	Yes
7	Immune-mediated haemolytic anaemia	2	2	3	Yes
8	Haemorrhage	2	2	2	Yes
9	Haemorrhage	2	2	2	No
12	Haemorrhage	2	2	4	Yes
14	FeLV infection	1	1	2	Yes
15	FeLV infection	1	1	2	Yes
16	Inflammatory disease	1	1	2	Yes
17	Neoplasia	2	2	2	Yes
18	Kidney disease	1	1	2	No
19	Diabetes mellitus	1	1	3	No

FeLV = feline leukaemia virus

Patients’ Hct increased as expected in 51.5% (17/33) of the transfusion instances. The Hct changed at an actual range from –3.8 to 15.2 percentage points (median 2.8 percentage points), as compared with the expected increase of 0.9 to 9 percentage points (median 2.8 percentage points). In patients with a positive MCM, the Hct increased as expected at a rate of 40% (4/10) vs 57% (13/23) for those with negative MCM results. The mixed model linear regression analysis did not identify a statistically significant association between inadequate Hct increase and the occurrence of a positive MCM in the course of testing (Table 5). However, patients with a positive MCM achieved an Hct that was, on average, 1.04 less than expected. The Hct measured in those with a negative MCM had an Hct that was, on average, 0.53 more than expected (P = 0.237; Table 5).

Table 5.

Mean deviation of haematocrit values from the expected post-transfusion increase with SE and confidence intervals (CI) in patients with positive and negative cross-match (CM) results

Positive CM	Mean	SE	95% CI
			Lower	Upper
No	0.535	0.721	−0.935	2.005
Yes	−1.042	1.093	−3.271	1.187

P = 0.237

The bilirubin concentration was not available in enough cases to enable a consistent assessment of its development. The volume of intravenous fluids administered to patients was, depending on clinical status, at or below maintenance rate for all but one patient (19/20) with severe dehydration due to diabetic ketoacidosis.

Discussion

This study addressed the need for cross-matching as a measure to avoid potential transfusion reactions in cats. Based on the occurrence of positive MCMs reported here, 25% of patients may have developed alloantibodies against erythrocyte antigens. The earliest occurrence of such alloimmunisation was documented 2 days after the first AB-compatible blood transfusion was administered.

One limitation on cross-matching in the clinical setting is cautious blood sampling from anaemic cats; that is, clinicians responsible for hospitalised anaemic cats tend to sample as little as possible for monitoring Hct development and, as such, the amount left over for cross-matching is often quite small. In our experience, however, as little as 250 μl blood sufficed in order to adequately perform a CM. It is therefore realistic to expect to be able to safely cross-match in anaemic cats, even when samples are taken for monitoring.

Pre-transfusion testing may be affected by the age of samples used. A study in equine medicine reported a higher occurrence of incompatible CMs with acid citrate dextrose preserved donor samples as early as 1 week after collecting; the authors of that study go so far as to recommend fresh donor samples for cross-matching, without, however, suggesting an acceptable time period between obtaining the sample and performing the CM.^23,26 The general recommendation in veterinary medicine regarding the age of specimens for cross-matching is that they should be under 24 h old, with the exception of unit donor segments that may be as old as the unit itself. In the current study, donor samples were taken at the time of blood donation and preserved with CPDA-1. The age of donor samples thus correlated with the age of the units, which ranged from 0–21 days (median 7 days) in our study. We cannot, therefore, exclude an influence of longer storage on our CM results.

Three of the patients in this study were diagnosed with FeLV, an infection that has been associated with false-positive MCM results. ²⁴ Of the FeLV-positive patients cross-matched here, one showed a positive MCM in the course of testing. It is unlikely that this represents such a false-positive, as the positive MCM did not occur until the seventh day after the initial transfusion.

Transfusion patients frequently receive glucocorticoid therapy in immunosuppressive doses when underlying diseases such as IMHA, PRCA, leukaemia and so on are present. It is unclear whether such treatment suppresses the development of alloantibodies in transfused patients. ²⁷ Seven of 20 patients received glucocorticoid treatment, two of which showed a positive MCM after having been on this treatment for 8 and 12 days, respectively. More data are needed in order to be able to make claims regarding the clinical relevance of positive MCMs in immunosuppressed cats.

The efficacy of transfusions is determined by examining various factors, the simplest of which is whether or not an acute transfusion reaction has taken place. Acute transfusion reactions were not observed in the subjects studied here. RBC viability may also be measured by examining post-transfusion Hct development; a study on transfusions in cats reported that in 43.3% of the patients the Hct was close to the value calculated prior to transfusion. ² In the present study, the number is slightly higher at 51.5% for the study population. The group of patients that showed a positive MCM had a lower response rate (40%), which may be an indication of the severity of disease in those patients; however, a delayed transfusion reaction cannot be ruled out. An increase of the bilirubin concentration in a patient’s blood as a parameter that measures post-transfusion haemolysis would aid in determining whether a delayed transfusion reaction has taken place. Unfortunately, this parameter was not available frequently enough as it was not measured at the same intervals as the blood count, owing to volume constraints in anaemic patients. Fluids administered to our patients should not have had an effect on Hct development, as all patients received fluids at or below maintenance rate, except for one severely dehydrated patient that received a higher volume until rehydration was achieved. It should also be noted that strongly regenerative anaemia may have caused the Hct to increase more than expected and that patients with continuing haemorrhage or haemolysis may have displayed a less than expected Hct increase, while the effectiveness of blood transfusions is best assessed in patients with ineffective erythropoiesis.

The small data set of this study precluded statistical significance; however, our numbers indicate that the Hct increases were less marked in those patients with positive MCMs (Table 5), despite our not having administered mismatched blood. This is a result that warrants further investigation.

The possibly deleterious effects of multiple red cell transfusions, because of their ability to increase the likelihood of alloantibody formation, have been described for humans. ¹¹ In the present study, both patients that received more than one transfusion and those that received blood from more than one donor also constitute a majority (3/5) in the group of patients with positive MCMs. Despite our low case number, such results ought to be taken into consideration when weighing the benefits of multiple transfusions for anaemic patients.

Transfusion reactions have been reported to occur in veterinary medicine despite compatible CMs. One study reports on this for feline patients with the occurrence of delayed haemolysis (n = 2) and acute haemolysis (n = 1), ²⁸ and another demonstrates it for horses. ²⁹ Despite such findings, cross-matching remains one of only two pre-transfusion testing methods available to veterinarians. The type of cross-matching used in veterinary medicine has been largely replaced in human medicine by computer cross-matching and the more sensitive antibody screening. ³⁰ However, in feline transfusion medicine, it remains an effective, easy and relatively economical method for detecting profound incompatibilities in the clinical setting. Not only does it overcome the inability of blood typing to detect unknown or rare blood types such as that defined by the Mik-antigen, but it also helps find the best match for patients that are at risk of having formed alloantibodies.^15,16,18 Several authors, therefore, hint at the expediency of introducing cross-matching into routine pre-transfusion testing.^15,18,31

A limitation of this study lies in its clinical design. Retesting our results was not part of the study design, given that only blood samples left over from patient monitoring were available for cross-matching. Also, monitoring patient Hct was not followed according to a schema predetermined in this study, but rather at the discretion of each patient’s clinician; the same applied to the decision to perform transfusions. This is exemplified by the transfusion of one patient with a Hct of 23.5%. This patient had been diagnosed with anaemia of inflammatory disease and therefore was at risk of rapidly worsening, ³² as well as not having responded adequately to the first transfusions.

Conclusions

We have shown that the possibility of a positive MCM exists as early as 2 days after initial transfusion. We have no data supporting the use of CMs in pre-transfusion testing of transfusion-naive cats, as none of our positive CMs occurred prior to the initial administration of blood transfusions. However, there is no direct way to assess the status of non-AB antibodies in cats, and there is no easy way to determine delayed transfusion reactions in hospitalised patients. This, in addition to a lack of evidence for the textbook rule of cross-matching 4 days post-initial transfusion, should encourage the development of more diligent pre-transfusion testing protocols, which could include performing CMs as part of routine pre-transfusion testing in cats. Experimental studies would be needed to provide a solid evidence base for routine pre-transfusion cross-matching.

Although cross-matching, just like blood typing, cannot avoid sensitisation and does not help increase the likelihood of a successful transfusion as evidenced by an increase in Hct, it is the only other means with which the overall risk of transfusions can be lowered and is therefore a tool towards increasing patient safety. Our recommendation is therefore to consider introducing cross-matching into routine pre-transfusion testing protocols as part of a continued effort to increase patient safety, along with diligent documentation, sound blood-banking practices and an emphasis on the indication for individual transfusions.