Can We Detect Transfusion Benefits in Palliative Care Patients?

Introduction

Red blood cells (RBCs) are a valuable and limited resource and, in situations such as massive blood loss, can have life-saving effects. The cost of preparation and storage of one RBC unit in Australia is estimated at U.S. $284, ¹ and cost to U.S. hospitals ranged from U.S. $210 to U.S. $343, ² with the total cost of a single unit of RBC transfusion between U.S. $522 and U.S. $1183. ³ In palliative care, RBC transfusions are more likely to be used for chronic anemia, with target symptoms such as fatigue and breathlessness, or for general well-being.

However, RBC transfusion is not without risks, including transfusion reactions, fluid overload, and infection. Furthermore, benefits seen in nonpalliative care patients receiving transfusion cannot simply be extrapolated to palliative care patients, given their burden of disease, cachexia, functional state, deconditioning, and treatment-related effects. There is yet to be definitive evidence for the net effect (benefit or harms) of such treatment in the palliative care setting and, as such, clinicians continue to support these patients with RBC transfusions. The aim of this retrospective study was to review the transfusions used in our palliative care unit (PCU), examining evidence of benefit or harms, and preparing for a prospective multisite study.

Methods

This consecutive cohort study of RBC transfusions followed patients between January 1, 2010 and December 31, 2011 in a 15-bed inpatient PCU in Adelaide, Australia. Over this period, there were 693 admissions, a median length of stay of 10 days, with 67% of discharges being due to death. Patient transfusions were decided on by clinical indication, without prescriptive guidelines, by the clinical treating team.

A list of RBC transfusions and patients was generated from a comprehensive transfusion database, and case notes were reviewed for patient demographics, comorbidities, primary disease, and vital status on discharge from any admission, in which there was a transfusion administered. The notes were interrogated for the reason for transfusion, pretransfusion and post-transfusion hemoglobin levels, adverse reactions, and symptomatic response as reported by patient or documented by medical officer.

In addition, Australia-modified Karnofsky Performance Status (AKPS), ⁴ Resource Utilisation Group–Activities of Daily Living (RUG-ADL), ⁵ and Symptom Assessment Scores (SAS) ⁶ for breathing and fatigue were collected prospectively as part of routine care of the national Palliative Care Outcomes Collaboration throughout the study period. ⁷ The AKPS is a 0–100 numerical rating scale of function, with higher values indicating higher function. The RUG-ADL is a functional scale incorporating ratings of bed mobility, toileting, transfers, and eating, scored from 4 to 18, with higher scores indicating lower function. The SAS is a 0–10 numerical rating scale for each of seven common symptoms including breathing and fatigue, with higher scores indicating greater symptom-related distress.

Scores were collected on the day before transfusion, and two and seven days after transfusion (D1, D2, D7, respectively). Responders were defined as an improvement of either by two or more points on RUG-ADL, an increase by 10 or more on AKPS, or a reduction of 2 or more on SAS breathing or fatigue scales. Deterioration was defined as the same magnitude of response, but in the opposite direction.

Given the exploratory and descriptive nature of the study, descriptive statistics are presented and also exploratory group comparisons using chi-squared test. Subgroups considered included baseline hemoglobin, hemoglobin increment, baseline functional state (AKPS), and vital status on discharge. Ethical approval was gained from the Southern Adelaide Clinical Human Research Ethics Committee.

Results

Over the two years, the PCU had 693 admissions, with 378/693 (55%) male, average age 68.3 years (standard deviation [SD] 14.5), and 463/693 (67%) of admissions ending in death. Thirty-one patients received 44 transfusions over the two-year period, totaling 101 RBC units (average 2.3 units per transfusion). Of these patients, 14/31 (45%) were male, average age 63.6 years (SD 13.3), 30/31 (97%) had advanced malignancy, and had a median unweighted Charlson Comorbidity Index of 6. The average hemoglobin pretransfusion was 7.8 g/dL, with hemoglobin ≤8.0 g/dL in 28/44 (64%), and post-transfusion 10.1 g/dL (n = 34), with an average change per RBC transfused of 0.92 g/dL. Of the transfusions, 16/44 (36%) were during admissions ending in death, with 8/16 (50%) and 6/16 (38%) of those transfused occurring within a fortnight and week of death, respectively.

The most common (nonexclusive) indications for transfusion were fatigue/lethargy (41/44, 93%), followed by breathlessness (7/44, 16%), light headedness (3/44, 7%), and ongoing blood loss (6/44, 14%).

Of the 44 transfusions, 39 (89%) were documented as having subjective benefit by clinicians, and of the 31 patients, 29 (94%) reported symptomatic improvement.

There was little improvement in average scores on RUG-ADL, AKPS, or SAS breathing and fatigue scales (Table 1). Using the predefined response criteria, response rates were less than 25% in all scales, with the lowest response in AKPS and highest in fatigue. Considering that a target symptom was not specified, patients had benefit in at least one scale in 22/44 (50%) of cases. Of the 39 transfusions with subjective benefit, 18/39 (46%) had improvement in at least one scale, whereas of the 5 transfusions with no subjective benefit, 4/5 (80%) had improvement in at least one scale. Of the transfusions within the last week and fortnight of life, 3/6 (50%) and 5/8 (62%) have improvement in at least one scale, respectively.

No adverse effects were noted. After transfusion, deterioration using clinical tool criteria was frequently noted, often of the same or greater magnitude of the response rate (Table 1). This was more pronounced in those transfusions during an admission ending in death. Looking at subgroups, baseline hemoglobin, functional status, and vital status at discharge did not influence the likelihood of response (Table 2).

Discussion

In a palliative care population receiving RBC transfusions for symptomatic anemia, clinicians and patients generally perceived benefit; however, routine assessment scales examining symptoms and function did not reproduce these views.

There are a number of possibilities that may explain this. First, it may be that assessment scales used in routine practice do not adequately assess the target symptoms and are not sensitive enough to detect clinician- and patient-reported changes. Here, subjective benefit was not reliably correlated with response using assessment scales. Some have gone on to develop or use more comprehensive tools for the assessment of symptoms such as fatigue. A study of 30 patients receiving RBC transfusion was assessed with both the Brief Fatigue Inventory and FACT F-fatigue subscale, which showed detectable and clinically significant improvements in fatigue when correlating the tools and patient-reported improvement. ⁸ Another group used visual analogue scales, rather than numerical rating scales, and found early and sustained improvements in strength, breathing, and well-being. ⁹ However, another study using numerical rating scales similar to our study to evaluate well-being, fatigue, and breathlessness found short-term benefit in fatigue and breathlessness. ¹⁰ Differences in findings may reflect the sensitivity of the tools, the assessing of a target symptom, and different populations at different stages of the disease trajectory.

Second, it is noted that placebo effect of interventions in palliative care can be large, ¹¹ with a recent study of fatigue in advanced cancer finding a placebo response rate of 56%. ¹² It may be that both patients and clinicians perceive a positive response to a RBC transfusion, which may not be supported by objective measures. Finally, it may be that RBC transfusions in palliative care patients are not very efficacious when used across a broad and diverse population. A recent systematic review examining the benefit of RBC transfusions in palliative care found a lack of high-quality studies to answer this question, with respect to study design and sample sizes. ¹³

Deterioration was also common in this population, and often the rates matched or were greater than the response rates. This is not surprising, given the progressive nature of disease in this population; however, deterioration cannot be attributed to the RBC transfusion. The absence of documented adverse events may go some way to alleviating concerns about potential harms, but the high rate of deterioration still raises questions of net effect of transfusions so very late in life.

However, could there be benefit in particular subgroups? In this exploratory analysis, no difference was found in groups with differing baseline hemoglobin, hemoglobin increment, functional state, and vital status on discharge. The same systematic review found that factors including age, pretransfusion hemoglobin level, functional state, or symptoms did not correlate with clinical benefit. ¹³ It is likely that in many palliative care patients, symptoms such as fatigue are multifactorial, due to a combination of effects from anemia, treatment-induced, disease-related, and physical deconditioning. Not surprisingly, a single intervention targeting one of these factors at one point in time is unlikely to have widespread benefit. What is unclear is whether or not particular target symptoms may receive greater benefit from RBC transfusions.

The limitations of this study include its retrospective nature and small sample size, but the data were collected routinely prospectively and the sample represents all admissions consecutively to a busy regional palliative care inpatient unit over a two-year period. The assessments are made from interpretation of the case notes and review of standard assessment tools. These are neither comprehensive nor sensitive anemia-related tools. Furthermore, this study did not examine RBC transfusions in palliative care patients in settings other than a PCU, such as through oncology services or day infusion centers.

Reviewing long-standing practice, particularly involving a valuable and finite resource, is an important component of modern medicine. Further prospective studies are required to help guide the pragmatic and efficacious use of RBC transfusions—the Australian Palliative Care Clinical Studies Collaborative is currently undertaking such a study, utilizing a prospective registry of consecutive cases as used in their pharmacovigilance series. ¹⁴