Article Text

Download PDFPDF
Red cell transfusion benefits in oncology, haematology and palliative medicine populations: a narrative review
  1. Ed Duffy1,
  2. Frances O’Mahony2,
  3. Caroline Burke1,
  4. Aoibheann Conneely2,
  5. Helen O’Connell1 and
  6. Feargal Twomey1
  1. 1Palliative Medicine, Milford Care Centre, Castletroy, Limerick, Ireland
  2. 2Palliative Medicine, Our Lady's Hospice & Care Services, Dublin, Ireland
  1. Correspondence to Dr Ed Duffy, Palliative Medicine, Milford Care Centre, Castletroy, Limerick, Ireland; ed.duffy{at}


Due to the heterogenous nature of the palliative medicine patient population, assessment of benefit, and thus choice of appropriate patient for consideration of transfusion, can be challenging. This can be confounded by the use of both liberal and restrictive transfusion thresholds. The multifactorial nature of many symptoms of anaemia, particularly in patients with advanced malignancy, can further complicate. As such, there is a paucity of data supporting the subjective, objective and clinical benefit of red cell transfusion in the palliative medicine setting. This narrative review summarises the research and evidence surrounding the benefits of red cell transfusion, with a particular emphasis on the oncological, haematological and palliative medicine population. There is a lack of a validated, reproducible patient-reported outcome measures (PROM) to assess response to red cell transfusions in the palliative medicine population with outcome measures varying from objective improvement in haemoglobin level post-transfusion, to subjective response in primary symptom(s). Further investigation is required regarding the development of effective PROMs assessing response to red cell transfusion in the palliative medicine population, to ensure judicious use of this scarce and valuable resource.

  • cancer
  • clinical assessment
  • drug administration
  • haematological disease
  • dyspnoea
  • fatigue

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Key messages

What was already known?

  • Anaemia is a frequently encountered symptom in the palliative medicine population with red blood cell transfusion being a commonly used resource in its management.

What are the new findings?

  • To the authors’ knowledge, this is the first paper specifically collating data regarding clinical benefits of red cell transfusion in the palliative medicine patient population.

  • There is a lack of data regarding choice of appropriate candidate for red blood cell transfusion.

  • There are currently no patient-reported outcome measures for the assessment of response to red blood cell transfusion in the palliative medicine population.

What is their significance?


  • Neither choice of appropriate candidate for red blood cell transfusion in the palliative medicine patient population nor effectice methods of post transfusion assessment are currently supported by effective and robust data


  • Exploration of pre-transfusion criteria to trigger red blood cell transfusion in this population, as well as development of effective and validated tools to assess patient response to transfusion could lead to improved patient blood management and patient outcomes.


The European Cancer Anaemia Study,1 a prospective epidemiological observational study published in 2004, set out to define the incidence, prevalence and management of anaemia in patients with cancer in 748 cancer centres across 24 countries. Between January and July 2001 with a 6-month follow-up period, a population of 13 628 was analysed. Findings demonstrated that the incidence of anaemia was 53.7%, with a mean haemoglobin (Hb) to initiate treatment of 97 g/L.

While palliative medicine is seeing an increase in numbers of patients with non-malignant conditions, the majority of patients receiving palliative medicine input continue to be those diagnosed with a malignancy. According to statistics set out by the Irish Health Service Executive in their Palliative Care Services: Three-Year Development Framework (2017–2019),2 in 2016, 70% and 86% of specialist palliative care patients in the community and inpatient unit settings, respectively, had a diagnosis of cancer.

A review of the research assessing the benefits of red cell concentrate (RCC) transfusions was completed as part of a 2-year review of red cell transfusion practice at an specialist palliative care inpatient unit in Ireland. This review focused on patients with oncological and haematological diagnoses, with emphasis on patients under the care of palliative medicine.

We present a chronological review of the current data assessing the benefits of red blood cell transfusions in patients with cancer who are receiving specialist palliative care.


A 1995 prospective study3 sought to review blood transfusion practice, identify whether benefit was derived in the palliative medicine setting and, if so, which symptoms were most improved. It involved assessment of blood transfusions in malignant disease in eight centres forming part of the South West Thames Palliative Medicine Collaborative Audit Group between September 1992 and August 1992. Indications for transfusion included weakness, dyspnoea, benefit from previous transfusion, acute bleed and prior to either surgery, chemotherapy or radiotherapy. The impact of blood transfusion on dyspnoea, weakness and overall sense of well-being was assessed in 91 patients through use of visual analogue scales (VAS) before and on two occasions after transfusion. It found that the most commonly identified cause for anaemia was chronic anaemia of malignancy. The median Hb pre-transfusion was 79g/L. No correlation was observed between the degree of anaemia and VAS/dependency score, which suggests a multifactorial nature to symptoms. The main outcome finding was that transfusion resulted in most significant improvement in weakness and overall sense of well-being, independent of pretransfusion Hb.

Mercadante et al4 used the findings of Sciortino et al in their 1993 study of home care patients,5 which found no changes in physical measurements post-transfusion, to support their hypothesis that terminally ill patients with cancer can tolerate much lower Hb levels without symptomatic burden. They posited a potential psychological benefit due to improvements in quality of life scores only, post-transfusion. The authors further suggested, however, in view of findings from Gleeson et al3 above that symptomatic relief and improvement in well-being in patients with a diagnosis of advanced malignancy was achieved for weakness, dyspnoea and impaired overall well-being, and that relief of symptoms might persist for up to 2 weeks.

A survey of blood transfusion practice at specialist palliative care inpatient units6 in the Yorkshire Cancer Network assessed between October 2002 and September 2003, sought to assess effects on physical functioning and timing of transfusion before death. Of the patients transfused as inpatients, half received a transfusion within 5 weeks of death. Findings indicated that physical functioning was not improved after transfusion; however, it is noted that this was based on a subset of the cohort (inpatients) and may not be generalisable.

A further study7 by Mercadante et al aimed to assess the effects of RCC transfusion and subsequent increase in Hb levels on anaemia-related symptoms in a cohort of patients with cancer in Italy. The data were gathered between July 2005 and December 2006, with transfusions recommended to patients with Hb 80±5 g/L with a mean target increase of 20 g/L. Of the 61 recruited patients, 133 RCC units were transfused. Of particular interest was the fact that, although significant changes in fatigue and dyspnoea were noted immediately post-transfusion, these were not sustained at 15 days, while increased Hb values were sustained at 2 days and 2 weeks. The authors support their previously posited theory4 that, in advanced cancer, symptoms related to anaemia, which improve with transfusion, may do so, at least partially, on the basis of a psychological response. Therefore, the authors suggest that decisions relating to transfusion of severely ill patients should involve open discussion, taking into account all factors, including psychological ones.

One of the challenges identified with interpretation of these studies was noted in a 2012 Cochrane Review,8 which sought to synthesise clinical evidence and summarise knowledge gaps regarding blood transfusions for anaemia in advanced cancer. Importantly, it was noted that there is currently no gold standard to measure improvement in patients’ condition following a blood transfusion either in the palliative medicine setting or at end of life. At the time of writing, there were no randomised control trials or interrupted time series. Twelve before and after studies were identified; however, all were deemed biased by nature of their design, and as such, provided weak evidence. These showed subjective early response (fatigue, breathlessness and QoL) in 31%–70% between days 2 and 7, which waned at day 14. Overall survival was between 2 and 293 days with a significant proportion (23%–35%) dead within 2 weeks. The mean Hb trigger was 79 g/L with most participants receiving two units.

The authors argue that the potential harms of transfusion close to the end of life need to be distinguished from inappropriate transfusion in dying patients. Studies3 9 have found that pre-transfusion Hb level does not correlate with response to transfusion and maintenance of post-transfusion Hb is not associated with sustained symptomatic improvement. There are few studies employing validated tools to measure response. Further, the authors recommend that prospective studies be initiated, collecting reasons for transfusion and anticipated prognosis, to identify the reason for the high number of deaths within 14 days. Additional areas for research were suggested, including the identification of patients most likely to respond to blood transfusions; to determine duration of response and whether there is a dose response; to identify other supportive methods to manage symptoms in this context avoiding RCC transfusion; identification of causes of early death in those receiving transfusion (expected vs adverse event) and investigation of physicians’ decision-making regarding why they prescribe RCC to some patients, but not others with a similar profile.

A retrospective consecutive cohort study10 of RCC transfusion between 01/01/2010 and 31/12/2011 in a 15-bed palliative medicine unit in Adelaide, Australia looked at reason for transfusion, pre-transfusion and post-transfusion Hb levels, adverse reactions and symptomatic response, as well as Australian Modified Karnovsky Scale (AKPS), Resource Utilisation Groups - Activities of Daily Living (RUG-ADL) and Symptom Assessment Scale (SAS) breathing and fatigue scales. A total of 31 patients (30 of whom had advanced cancer) received 44 transfusions and a total of 101 RCC units. Average pre-transfusion Hb was 78 g/L, with average post-transfusion Hb 101 g/L and an average change per unit of RCC transfused of 9.2 g/L. The most common indications for RCC transfusion were fatigue/lethargy in 93% of cases, breathlessness in 16%, ongoing blood loss in 14% and lightheadedness in 7%. Little improvement was noted in RUG-ADL, AKPS or SAS breathing and fatigue scales, which was in contradiction to clinician and patient perception of benefit. It was suggested that this discordance might relate to the poor sensitivity of the scoring systems, the placebo effect of receiving a RCC transfusion as well as the multifactorial nature of symptoms in advanced cancer. This study is limited by both its retrospective nature and its small sample size, but was considered to be representative of a regional palliative medicine unit.

An international, multisite (17 sites in 5 countries), prospective consecutive cohort study11 assessed symptoms of fatigue, breathlessness, generalised weakness and dizziness in 141 patients pre-RCC transfusion, at day 2 and on an ad-hoc basis up to day 7 between January 2014 and September 2015. A mean of 2.1 units of RCC were transfused, with fatigue being the predominant target symptom in 61% of cases. Benefit in primary target symptom was reported in 49%, with benefit in any symptoms derived in 78%. Harm at day 2 was noted in 12% of patients, and severity was mild, with harms including delayed allergic reaction, infusion related reaction, delirium, constipation, seizure and vomiting. Improvement in the primary target symptom at AKPS of 20–30, 40–50 and 60–70 was 36%, 66% and 41%, respectively, and these were statistically significant. There was disparity between objective improvement and self-reported assessment tools, again without any predictors of response to RCC transfusion. Limitations in this study included the fact that data were not collected beyond 7 days and that data with respect to concurrent treatments were not collected as part of the study.

A 2018 single-centre retrospective chart review12 of 116 deceased haematopoietic stem cell transplant patients sought to determine the mean time difference between the date of last transfusion (RCC or platelet) and death. Between January 2011 and December 2015, results showed that the mean difference in patients in the hospice setting was 45.9 days±66.5 days and 14.6 days±48.1 days in non-hospice patients. Among hospice patients, only 25% of patients received a transfusion within 5 days of death, and none on the day of death. However, 17 non-hospice patients received transfusions on the day of death. It should be noted that the acuity of each particular clinical situation was not documented, that is, whether or not transfusion formed part of a resuscitation effort.

A 2018 systematic review of red blood cell transfusion in the adult palliative medicine setting13 assessed outcomes including symptom relief post transfusion, post-transfusion survival, Hb post-transfusion and adverse events. The authors found that symptom relief was often assessed subjectively, without the use of standardised scales of assessment. In these cases, adverse effects were not often reported. Findings suggested that symptoms were frequently attributed to anaemia; however, it again pointed out that such symptoms may be multifactorial in patients receiving palliative medicine care. One study reviewed14 suggested no significant correlation between Hb level and subjective extent of fatigue. Further, questions were raised with regard to some of the symptom assessment tools (including AKPS and RUG-ADL), as their use has not been validated specifically for anaemia. Another study15 found a mean duration of symptom relief of 18.5 days while improvement at 2, but not at 15 days was noted in a third study, discussed above.7 When discussing potential survival benefit, the question was raised as to whether life prolongation in a terminally ill population is an ethically appropriate therapeutic endpoint, due to the potential for increased suffering.

Chin-Yee et al concluded that low Hb levels are the most common indication for transfusion in patients receiving palliative medicine care and findings suggest that clinician experience, rather than evidence-based practice most often dictates transfusion thresholds. Unsurprisingly, there are more blood transfusions in the acute hospital setting than in hospice. A useful point is made in that the underlying aetiology of anaemia-type symptoms in the palliative medicine setting varies greatly and, as such, a blanket transfusion threshold may not be realistic. Research investigating anaemia in differing patient cohorts might allow for the development of more useful and condition-specific thresholds. The aim of future research should consider patient-reported outcome measures (PROMs) using validated tools.

In 2019, a prospective observational study16 aimed to estimate the effect of RCC transfusion on functional status and quality of life. A total of 208 outpatients 50 years of age and older with at least one haematological/oncological diagnosis were divided into those who had received anticancer treatment (including chemotherapy, biological agents, hormone therapy and/or radiotherapy) in the preceding 4 weeks and those who had not. Primary outcomes included change in 6 minute walk test and fatigue-related Quality of Life (QoL) scores (as measured by Functional Assessment of Chronic Illness Therapy scale version 4) 1-week post-transfusion. A transfusion trigger of <80 g/L was used in most cases. Median change in walk distance was 20 metres and median improvement in fatigue score was 3 points. This change in fatigue score was considered to be consistent with a small but significant improvement in QoL. The authors also found that patients who had undergone recent anticancer treatment and those with intercurrent illness requiring hospitalisation or emergency department visits, responded less well to transfusion.

There have only been two major studies defining current practice in palliative medicine. The first, from Neoh et al, presented a 3-month prospective audit17 assessing red blood cell transfusion practice in UK adult hospices. One-hundred and twenty-one of a total of 210 UK adult hospices participated between September and December 2016. Eighty-three of these administered at least one unit RCC during the 3-month period, which equates to 465 RCC transfusion episodes. A transfusion episode was defined as each patient receiving at least one unit of red blood cells within a 24-hour period. Results showed that 29% of patients had investigations of the cause of anaemia prior to transfusion, the most common causes of anaemia being functional iron deficiency 38%, active bleeding 24% and bone marrow failure 21%. Fifteen per cent of patients had their weight recorded, 5% had a Hb check after each unit transfused and only 28% had a post-transfusion Hb check at any point. Further, only 14% had a post-transfusion performance status documented.

Mean pre-transfusion Hb was 75 g/L with 70% of patients being transfused above the National Institute for Health and Care Excellence (NICE) and European Society for Medical Oncology (ESMO) recommended threshold of 70 g/L, with Hb above 80 g/L in 29%. Indications included low Hb in 51%, combination of low Hb and breathlessness in 40% and fatigue in 16%. Mean and median number of units transfused was 2, with 84% given more than one unit in a transfusion episode, also in contradiction of NICE guidelines. Performance status was measured in 83% of patients transfused; however, no improvement was demonstrated within the 30-day post-transfusion period. Eighteen per cent of patients experienced sustained benefits in their target symptom for up to 30-day post-transfusion. Thirty-two per cent of patients had died at 30-day post-transfusion. This snapshot study seems to indicate that transfusion practices, as set out by both NICE and ESMO, are not being adhered to in >50% of UK adult hospices.

On the basis of these findings, recommendations were developed in conjunction with NHS Blood and Transplant:

  • Thoroughly investigate causes of anaemia.

  • Adopt restrictive transfusion threshold of 70–90 g/L or 80–100 g/L in acute coronary syndrome.

  • Discuss risk and benefits with patients to ensure informed consent.

  • Mandatory weight checks to determine transfusion requirements.

  • Minimise the risk of Transfusion Associated Circulatory Overload, and vigilant assessment for same.

  • Rigorous clinical review of outcome as well as single unit transfusions followed by post-transfusion Hb check.

The second study was a 2019 descriptive, retrospective chart review18 which described patients who received packed red blood cell transfusions in the palliative medicine units of two tertiary hospitals in Canada over a 2-year period. 75.8% of patients had a cancer diagnosis, with gastric cancer the most common (12.5%) followed by leukaemia (12.5%) and colon, oesophageal and lung cancer (8.3%). All patients except one who received a blood transfusion in the palliative medicine inpatient unit died during the same admission. The main symptomatic reasons for transfusion were fatigue (83.3%) and dyspnoea (25%). Transfusion was noted to be beneficial if there was an improvement in Edmonton Symptom Assessment Scale or if narrative comments stated patient-reported improvement in target symptom. Post-transfusion notes included descriptors such as ‘more alert’, ‘less short of breath’ and ‘walking, up out of bed’. A clear effort was made to identify a target symptom(s) to assess pre-transfusion and post-transfusion. Unfortunately, no validated tool currently exists to assess the utility of transfusions, thus leaving a gap in knowledge regarding which cohort of patients might likely benefit from transfusion. The authors suggested that transfusion therapy could be limited to patients with a reasonable life expectancy.

Finally, current blood transfusion practice among palliative medicine doctors was compared with NICE guidance using an online survey19 containing four clinical vignettes of common clinical palliative medicine scenarios. These scenarios involved cases in which red cell transfusion may be considered. A total of 293 of 1070 members of the Association for Palliative Medicine of Britain and Ireland responded. Concordant responses to all four vignettes were selected by less than half of doctors, with 39.3% and 44.8% of doctors in cases 2 and 4, respectively, suggesting two or more units of blood to be transfused at a time, which is contrary to NICE guidelines. However, it should be noted that NICE guidelines were not developed using studies in palliative medicine, thus highlighting the need for palliative medicine specific research to be performed.


Blood transfusions are regularly used in the management of anaemic patients receiving palliative medicine input. However, as demonstrated in the research described above, there is a paucity of data supporting their subjective, objective and/or enduring clinical benefit.

The pathologies affecting patients with oncological and haematological malignancies who might benefit from patient blood management are heterogenous in nature. Allied to the myriad pathologies related to non-malignant conditions that might contribute to a similar need,20 21 it is clear that the assessment of transfusion benefit for each individual patient can be a very challenging task. The benefits of integrating oncology, haematology and palliative care thinking and practice are widely accepted22–24 and also apply here. Similar evidence regarding the integration of palliative care with other specialties is likely soon to follow.25

It is vital, in this combined tumour-directed and host-directed approach, that physicians in palliative medicine take particular note of specialty-specific developments and updates in the management of mutual patient cohorts. It is the pathophysiology of each individual condition, be it oncological, haematological or otherwise, that drives anaemia. Patients must be considered through the lenses and with the perspectives of both palliative medicine and haematology/oncology and so on, via the implementation and frequent re-evaluation of specialty-specific knowledge in the creation of palliative medicine guidance.

Further, as patients approach end-of-life, different criteria may take precedence and the underlying diagnosis can take less prominence. Prognosis, in the context of a scarce resource, can, and should, play a central role in the decision-making process as time becomes shorter. However, as we know, prognostication can be a challenging and inaccurate art in itself, which can further compound the issue. Certainly, the use of the programmes such as the Palliative Care Outcomes Collaboration26 could be helpful in tracking and assessing change in overall patient wellness, while the use of tolls such as PIPS-227 and The Glasgow Prognostic Scale28 and can aid in the assessment of prognosis.

The choice of appropriate candidates for RCC transfusion is central to a discussion of transfusion benefit. ‘Symptoms of anaemia’ such as shortness of breath, fatigue and decreased exercise tolerance do not directly correlate with a low Hb level. There are many other factors, in patients with a terminal diagnosis that can contribute to these symptoms. Distinguishing what proportion of the symptoms described can be ascribed to the low Hb, and which are as a result of, for example, cancer fatigue, adverse effects of medications or any number of other potential concurrent issues, can be a challenging endeavour. It is here again that the value of early palliative medicine intervention is highlighted. The utility of multidisciplinary, drug and non-drug, interventions to address the multifactorial nature of these symptoms, becomes clear.

The importance of cultural norms and expectation cannot be ignored. While clinical and therapeutic decisions should certainly not be made purely on the grounds of geography, the rationale for overall patient management will always, to some degree large or small, be impacted by cultural factors. This is, of course a much more challenging element to incorporate and quantify.

As stated above, anaemia can present clinically in a multitude of non-specific ways including fatigue, shortness of breath and decreased exercise tolerance. Each of these can be, and frequently are, assessed in numerous different ways including verbal responses, visual and numerical scales and more objective functional tests3 11 ,3 Often the presence of a low Hb can trigger a transfusion, even in the absence of significant symptoms of anaemia, or clear evidence that the anaemia is the primary cause of these symptoms. In those receiving a transfusion, the threshold can vary from restrictive to liberal, both interspecialty and intraspecialty.13 17

Post-transfusion outcome measures vary from an objective improvement in Hb post-transfusion to a very subjective improvement in shortness of breath, fatigue or general well-being. While such subjective improvements, often presented in terms of PROMs, may be more clinically desirable in the palliative medicine population in terms of symptom assessment and management, by their nature they are more difficult to quantify and, in the presence of other concurrent treatments (such as oxygen therapy, low dose opioid or psychological support), causality becomes more challenging to demonstrate, even in the presence of a concurrent rise in Hb. There is a clear lack of a validated, reproducible PROM for assessing response to RCC transfusion.

With the ongoing scarcity of RCC as a national and international resource, alongside an increasing body of evidence supporting a restrictive, rather than liberal, transfusion threshold, it is possible that we will see a reduction in the volume of RCC transfusions in the future. This could be further impacted by increased and improved patient education and understanding of the multifactorial nature of these ‘anaemic symptoms’, with greater focus on the other factors that contribute to fatigue, shortness of breath and decreased exercise tolerance, among others.

Finally, with specific emphasis on the Irish situation, and in view of the implementation of SláinteCare,29 there is an aim to shift the locus of patient care from the hospital/hospice to the community, and a shift in the focus of care with a greater emphasis on palliative medicine input. This could further impact on the volume of blood transfusions, leading to improved management of non-haemoglobin centric ‘symptoms of anaemia’.

Areas for further development

  • Consideration of a national audit of patient blood management practices to assess the current palliative medicine practice.

  • Clear criteria to decide whether a patient is a suitable candidate for a blood transfusion, underpinned by up-to-date, specialty-specific and subspecialty-specific literature.

  • Further patient blood management guidelines development regarding restrictive versus liberal transfusion practices, as well as the appropriate use of iron transfusions and other blood sparing products.

  • Methods of assessing effective response to transfusion in the above-determined candidate, including the development of validated transfusion-specific PROMs.

  • Improved prognostic tools.

  • Ongoing integration of palliative medicine into the fields of oncology and haematology30 and expansion into the non-malignant population.


This topic review, which was completed in conjunction with a 2-year review of patient blood management in an inpatient palliative medicine unit in Ireland, demonstrates the dearth of knowledge in the area of RCC transfusions in the palliative medicine setting. Specifically, in terms of assessment of symptoms of anaemia, and thus assessment of the benefits of transfusion, this remains an inexact science. While the need for RCC transfusion in this patient population is undeniable, the choice of the appropriate patient to whom it should be offered, and the method by which its benefit should be assessed, remains unclear, and demands further investigation. Further research should also be directed towards single unit transfusion and blood product alternatives, which are outwith the scope of this review.

Ethics statements

Patient consent for publication



  • Contributors FT, HO’C and FO’M conceived the presented idea. AC, CB and ED collected the data. ED collated the data, drafted and revised the manuscript. FT, AC and FO’M provided editorial oversight. ED is guarantor.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.