Anaemia management with red blood cell transfusion to improve post-intensive care disability: Protocol for the ABC post-ICU randomised controlled trial

Abstract

Background:

Anaemia is prevalent after intensive care unit (ICU) discharge as a consequence of factors such as blood sampling, concurrent inflammation affecting erythropoiesis, and the use of restrictive ICU red blood cell (RBC) transfusion practice during inpatient stay. ICU survivors experience poor health-related quality of life (HRQoL). Prevalent symptoms include fatigue and weakness, to which anaemia may contribute. There are no trials exploring the effectiveness of treating anaemia with RBC transfusions post-ICU discharge.

Methods and analysis:

The ABC post-ICU trial is a multicentre prospective, parallel group, randomised trial, with embedded moderation and mediation analysis. Participants are adult ICU survivors with anaemia (haemoglobin (Hb) ⩽94 g/L) fit for ICU discharge. Patients are randomised to usual care (default Hb transfusion trigger <70 g/L, target 70–90 g/L) or single-unit RBC transfusions to achieve Hb range 100–120 g/L. The intervention is from randomisation to hospital discharge. Primary outcome is the physical component summary score (PCS) of the 36-item short form (SF-36) health survey, which measures HRQoL, assessed 90 days post-randomisation. Secondary outcomes at 90 days include: hospital length of stay, mortality, fatigue score, activities of daily living, and Mental Component Scale score (MCS) SF-36. Outcomes are also measured at 30 and 180 days. Safety outcomes include: new infections, transfusion-related adverse events, and major adverse cardiac events. Analysis includes a moderation analysis based on baseline recalled PCS SF-36, comorbidity burden, mobility, and systemic inflammation (C-reactive protein (CRP) concentration). A mediation analysis based on 30 days blood samples will explore whether anaemia severity (Hb) or persisting inflammation (CRP) mediates intervention effects. A health-economic analysis over 180 days will be conducted. The sample size is 346, providing 90% power to detect a difference in PCS SF-36 of 5 points, assuming >70% completed SF-36 follow-up.

ClinicalTrials.gov:

NCT04591574.

Keywords

anaemia transfusion critical illness intensive care post-intensive care syndrome recovery health-related quality of life physical function

Strengths and limitations

This trial measures the effect of a more liberal blood transfusion strategy for anaemic ICU survivors on outcomes prioritised by patients, such as health-related quality of life, fatigue severity, and ability to undertake activities of daily living.

The trial has statistical power to detect a minimally clinically important difference in the physical component of quality of life using a validated instrument (SF-36).

The trial includes pre-planned moderation analyses to explore whether some groups of patients benefit more or less from the intervention, and mediation analyses to explore factors which may partially explain how the intervention works.

A limitation is that the trial is not blinded from researchers or participants, because this is difficult for blood transfusion interventions.

An anticipated limitation may be follow-up completion for the primary outcome, which is known to be challenging in this patient population. The sample size is inflated to account for this.

Introduction

Approximately 80% of patients requiring organ support in an intensive care unit (ICU) survive their critical illness. Many survivors subsequently experience poor health as a result of physical, mental, and cognitive impairments collectively called a ‘post intensive care syndrome (PICS)’.¹ New disabilities are often superimposed on pre-existing health impairments, and it is estimated that more than 50% of ICU survivors have pre-existing multiple long term conditions (MLTCs).²

Fatigue is the most prevalent symptom reported by ICU survivors, and is often severe and debilitating.^3–5 Impairment of physical function is common, predominantly affecting patients in the first 2–3 months following ICU discharge. Disability, as described by the World Health Organisation (WHO) International Classification of Functioning, Disability and Health Framework, is common, and includes impaired body function (muscle strength, walking ability), activity (ability to carry out Activities of Daily Living (ADLs)), and participation (including health-related quality of life (HRQoL), and return to work).^6,7 ICU survivors have excess long-term mortality compared to other hospitalised patients, and utilise substantially higher healthcare resources, especially during the first 3–6 months following ICU discharge.⁸ Patients typically spend a further 3 weeks in the acute hospital after ICU discharge, and 25%–30% experience unplanned emergency readmission to hospital within 3 months of hospital discharge.^2,9

Anaemia as a potential mechanism of delayed recovery and poor health after critical illness

Anaemia is highly prevalent during critical illness, because of bleeding, blood sampling, and impaired erythrogenesis, and is frequently present in survivors.^10,11 A key uncertainty is the degree to which anaemia is a risk-modifiable factor for the longer-term physical, mental, and cognitive impairments that characterise PICS. Although many red blood cell (RBC) transfusion threshold trials have been conducted in different clinical populations, including critical illness, uncertainty remains about the association between anaemia and HRQoL, and whether anaemia treatments improve outcomes other than mortality.^12,13 For example, recent trials in acutely brain injured patients suggest more liberal blood transfusion strategies may improve patient-centred outcomes such as HRQoL and functional status.^14,15 Other populations in whom restrictive RBC transfusions may not be preferable to liberal transfusions include patients with acute and chronic cardiovascular disease,^16,17 and older patients¹⁸ in whom mortality and complications may occur more frequently. In previous ICU-based trials deaths mostly occurred before ICU discharge rather than among ICU survivors. Longer-term disability and HRQoL were rarely reported, and the evidence for these longer-term outcomes is weak.¹² This could be important among ICU survivors, because cardiovascular disease is present in up to 25% of ICU patients, pre-existing MLTCs are prevalent, and mean age is 60–65 years.^9,19 Anaemia also persists in many patients following ICU discharge, and is prevalent at hospital discharge.¹¹ Most clinicians do not investigate the aetiology of post-ICU anaemia or actively treat it.²⁰ There is a need for further research on the impact of anaemia treatments on outcomes prioritised by patients, such as HRQoL, fatigue severity, and ability to undertaken activities of daily living.

In ICU, impaired erythrogenesis has multifactorial pathophysiology, but systemic inflammation is central by impairing iron absorption and availability through hepcidin upregulation, inhibiting erythropoietin release, and direct inhibition of marrow RBC production.^10,21 Persisting systemic inflammation is prevalent after ICU discharge²² and is associated with poor functional recovery²³ and HRQoL. This could occur via direct effects (e.g. neuro-muscular dysfunction) or indirect effects (e.g. persisting anaemia).

Existing evidence

A large observational literature demonstrates the high disability burden suffered by ICU survivors, especially during the first 3–6 months, but often lasting many years. This includes physical impairments, reduced Activities of Daily Living, and mental and cognitive disability.^1,6,24

We found no studies specifically exploring the impact of transfusion strategies after ICU discharge on health status among general ICU population survivors. Two ICU based transfusion trials measured longer-term HRQoL: a trial in septic shock patients,²⁵ and a trial in older patients requiring prolonged mechanical ventilation.²⁶ The intervention in both trials occurred during ICU, and ICU mortality was high. Survivor cohorts were small, confirmed markedly reduced HRQoL, but found no differences in HRQoL. These studies had limited statistical power, and did not explore whether baseline health status or other factors influenced any treatment effects. This is relevant, because pre-existing health is an important determinant of longer-term HRQoL in heterogeneous ICU populations.^27,28 Recent clinical transfusion guidelines for critical illness do not make recommendations specific to ICU survivors.^29,30 No rehabilitation trials have considered anaemia treatment either as the main intervention or as part of the intervention.³¹

Alternative approaches to treating post-ICU anaemia include blood conservation strategies, iron therapy, and erythropoietin therapy either individually or in combination. Evidence for these interventions is inconclusive, but several trials are ongoing.^32–35 These approaches may also treat anaemia, but the degree and rate of correction may differ from RBC transfusion. Both approaches require further investigation.

Trial hypothesis and objectives

We hypothesise that in patients discharged from intensive care with moderate to severe anaemia, treatment of anaemia using blood transfusions will improve quality of life, disability, and other measures of well-being.

Primary objective

To determine whether treating anaemia from the time of ICU discharge using blood transfusions (target haemoglobin (Hb) 100–120 g/L) results in an improvement in self-reported physical HRQoL 90 days after intensive care discharge, compared with current usual care which recommends transfusions when Hb is less than 70 g/L to achieve a target Hb 70–90 g/L.

Secondary objectives

To determine the effect of treating anaemia using blood transfusions on fatigue, disability, and death rate compared with current usual care.

To determine the safety of treating anaemia with blood transfusions, compared with current usual care.

To determine which patients may benefit most from treating anaemia with blood transfusions.

To determine whether more active treatment of anaemia using blood transfusions is cost-effective.

Methods and analysis

Overall design

A prospective, parallel group, open, unblinded, randomised controlled intervention trial, with embedded moderation and mediation analyses. This publication summarises protocol version 9.0, 14th March 2025.

Trial registration and sponsor

ClinicalTrials.gov Identifier: NCT04591574; Sponsor, ACCORD (https://www.accord.scot/).

Personalised medicine questions

We aim to explore personalised medicine in relation to the use of blood transfusions in ICU survivors. We will explore baseline patient characteristics that may moderate the effects of RBC transfusions and factors during the intervention that may mediate effectiveness of the intervention on the primary outcome, if an effect is found.

Moderation variables: we will explore the importance of four baseline characteristics on moderating treatment effects, namely: (1) baseline co-morbidity score (Functional Comorbidity Index, FCI),³⁶ (2) recalled (pre-ICU admission) physical HRQoL (Physical Component Summary (PCS) of the 36-item short form (SF-36) health survey),³⁷ (3) mobility at ICU discharge (ICU mobility scale³⁸) and/or (4) baseline severity of systemic inflammation (C Reactive Protein (CRP) concentration at randomisation).

Mediation variables: we will explore the importance of two potential mediators of intervention effects based on variables at 30 days post-randomisation, namely: (1) severity of persisting inflammation (CRP concentration), and (2) persisting anaemia (Hb concentration).

Cost effectiveness questions

We will estimate cost per Quality-Adjusted Life Year (QALY) and cost per life-year over the study follow-up period. Depending on the results of the short-term outcomes, linkage of routinely collected healthcare data may be undertaken after completing the trial to estimate long-term cost-effectiveness.

Patient and public involvement

We consulted with ICU patients and their families, who helped develop our research. This group strongly supported exploring interventions that might reduce fatigue and disabilities, and agreed that anaemia is an important consideration. Patients felt HRQoL was a key patient-centred outcomes and prioritised fatigue as a key symptom. The Trial Steering Committee (TSC) includes a patient member. ‘How can the physical consequences of critical illness be prevented and what is the best way to support recovery from these after intensive care?’ is a James Lind Alliance priority setting partnership ‘top 10’ question for critical care research.

Inclusion and exclusion criteria

Inclusion and exclusion criteria are shown in Table 1. For inclusion criteria, we chose a Hb threshold of ⩽94 g/L based on the mid-point of the ‘moderate’ anaemia range defined by the World Health Organisation at the time of trial design.³⁹ For exclusion criteria, we excluded patients with primary neurological ICU-admission diagnoses because these are likely to dominate the primary outcome. Patients recovering from cardiac surgery, post-variceal haemorrhage with chronic liver disease, those with pre-existing dialysis-dependent renal failure, and patients with chronic transfusion-dependent haematological disorders are excluded because best practice evidence is stronger in these groups.

Table 1.

Inclusion and exclusion criteria.

Inclusion criteria
[1] Patient who received level 3 ICU care at any time point during the current hospital admission (defined as advanced respiratory support and/or at least two organ support) [2] Patient considered ready for discharge from the ICU by the caring clinical team [3] Hb ⩽94 g/L when ready for ICU discharge or during the first 7 days following the decision by the treating clinician that the patient is ready for ICU discharge [4] Age ⩾ 16 years. [5] Patient expected to remain in study hospital until hospital discharge [6] Consent provided (by participant or in accordance with appropriate mental capacity legislation for the site)
Exclusion criteria
[1] Contraindication or objection to RBC transfusion [2] Active bleeding when screened [3] Primary neurological ICU admission diagnosis, including primary brain or spinal cord injury [4] Patients discharged from the ICU following cardiac surgery [5] Currently receiving or planned to receive end-of-life care [6] Not expected by clinical team to survive to hospital discharge. [7] Patient with a proven chronic haematological disease that requires regular RBC transfusion to treat anaemia [8] Patient with dialysis-dependent chronic renal failure prior to ICU admission [9] Patient receiving regular erythropoietin (or any erythropoiesis stimulating agent) treatment for anaemia prior to ICU admission. [10] Unable to obtain consent (from patient or in accordance with appropriate mental capacity legislation for the site) [11] Readmission to ICU during current hospitalisation episode and not enrolled following previous ICU admissions [12] Patient recovering following liver transplantation, kidney transplantation, or combined kidney/pancreas transplantation [13] Patient recovering from variceal bleeding due to chronic liver disease [14] Prisoner [15] Patient due for imminent hospital discharge within the next 24 h

Trial outcomes

Primary and secondary outcomes are shown in Table 2.

Table 2.

Summary of primary and secondary outcomes.

Outcome	Outcome measure	Method and timing of measurement
Primary outcome	Physical HRQoL	Physical Component Scale (PCS) of the SF-36 Health survey at 90 days post-randomisation.³⁷
Secondary outcomes	Patient centred outcomes
	Physical HRQoL	Physical Component Scale (PCS) of the SF-36 Health survey at 30 and 180 days post-randomisation.
	Mental HRQoL	Mental Component Scale (MCS) of the SF-36 Health survey at 30, 90, and 180 days post-randomisation.
	HRQoL	SF-36 Health survey eight sub-domains (i.e., physical functioning, limitations due to physical health, pain, general health, energy/fatigue, social functioning, role limitations due to emotional problems and emotional well-being) at 30, 90, and 180 days post-randomisation.
	Patient fatigue	Fatigue Severity Scale (FSS)⁴⁰ at 30, 90, and 180 days post-randomisation
	Activities of Daily Living (ADLs)	WHODAS questionnaire⁷ at 90 days post-randomisation
	Survival	Patients alive at 30, 90, and 180 days post-randomisation
	Anaemia severity	Haemoglobin concentration at 30 days post-randomisation
	Resource use and cost-effectiveness
	Post-ICU length of hospital stay	Hospital records
	Care costs during 180 days post-randomisation.	Health resource use questionnaire at 30, 90, and 180 days
	Incremental cost per QALY at 180 days	Health economic evaluation
	Process and safety outcomes
	Protocol compliance (during intervention index hospital stay)	Deviations from intervention transfusion practice
	Anaemia severity during index hospital stay	Haemoglobin concentrations during index hospital stay (from routine data at least weekly)
	Red blood cell (RBC) use	Blood transfusions (units of RBCs) from hospital/medical records during 90 days follow-up
	New infections	Infections requiring antibiotic administration during 90 days follow-up
	Transfusion-related adverse events	From hospital/medical records during 90 days follow-up
	Major adverse cardiovascular events (MACE)	From hospital records during 90 days follow-up

Patient screening and consent

Patient screening is at the time they are considered ready for ICU discharge by clinical teams. Members of the responsible clinical team are asked to confirm suitability prior to any approach for consent by research staff. The eligibility window is from the time the patient meets inclusion criteria for up to 7 days following the decision by the treating clinician that the patient is ready for ICU discharge, as long as the patient remains in the acute hospital.

Wherever possible, patients are approached when they have mental capacity. However, when the clinical and/or research teams consider that they lack mental capacity approaches appropriate to the local legal arrangements for Adults with Incapacity are used, including a surrogate decision maker (SDM) where appropriate. Any patients in whom consent is from a SDM are approached for consent to remain in the trial once they regain capacity.

Randomisation

Patients are randomised using a web-based randomisation service managed by the Edinburgh Clinical Trials Unit (ECTU) to ensure allocation concealment. Allocation is in a 1:1 ratio, stratified by centre in permuted random blocks. The randomisation sequence was created using computer-generated pseudo random numbers by an ECTU member of staff not involved in clinical care.

Intervention

Duration of intervention

The intervention period lasts from the time of enrolment until discharge from the acute hospital, and ends when one of the following occurs: patient is discharged home from hospital; patient dies during the acute hospital stay; patient is discharged to a rehabilitation hospital or facility; or, patient is discharged to an intermediate or long-term care facility.

If a participant is discharged to another acute care hospital from the enrolling hospital as part of the same hospitalisation episode, the patient remains in the trial and treatment allocation is maintained wherever possible until discharge from that acute hospital.

Assessment and management during intervention period

Measurement of Hb during the intervention period is at the discretion of clinical teams, but is requested by research teams if not routinely available for 7 days.

Usual care group: Patients are managed according to current National Institute of Health and Care Excellence (NICE) guidance (https://www.nice.org.uk/guidance/ng24). For the purpose of the trial this means: use restrictive RBC transfusion thresholds for patients who need RBC transfusions and who do not have major haemorrhage or have acute coronary syndrome; use a threshold of less than 70 g/L and a Hb target of 70–90 g/L after transfusion, unless the responsible clinician decides to modify this based on the patients’ condition; use single-unit RBC transfusions for adults who do not have active bleeding and re-check Hb within 24–48 h of the transfusion. This guidance should be followed after each Hb measurement undertaken during the intervention period.

Intervention group: patients should be managed as follows following each Hb measurement undertaken during the intervention period: if the Hb is ⩾100 g/L no transfusion should be given; if the Hb is <100 g/L a single-unit RBC transfusion should be given followed by clinical reassessment and re-checking the Hb; if the Hb remains <100 g/L further single unit RBC transfusions are given until the Hb is 100–120 g/L.

For both study groups, the Hb is re-checked after each unit of RBCs administered within 24 h unless patients are bleeding consistent with best practice. (NICE Blood Transfusion Quality Statement 3: www.nice.org.uk/guidance/qs138/chapter/Quality-statements).

Transfusion prescription and patient monitoring during transfusion

Transfusions are commenced as soon as possible after a triggering Hb result and within a maximum of 48 h. Transfusions that do not occur or occur after 48 h are reported as protocol deviations. Bedside prescription checking and monitoring of patients during the transfusion period follow locally agreed protocols and procedures. This may include prescription of additional medication, such as diuretics, at the discretion of the clinical team. If Hb is not re-checked within 48 h this is also reported as a deviation. In the usual care group, administration of a blood transfusion at a trigger Hb ⩾70 g/L without the responsible clinician providing a reason for modifying the transfusion trigger is also recorded as a deviation.

Use of alternatives to blood transfusion

Patients in both study groups can receive alternatives to blood transfusion if directed by the clinical teams. However, these should be consistent with current NICE guidance concerning alternatives to blood transfusion (www.nice.org.uk/guidance/ng24).

Other treatment and interventions during the intervention period

All other treatments and therapies provided to patients in both study groups are at the discretion of the clinical teams.

Data collection

Data collection is summarised in Table 3.

Table 3.

Data collection during the trial.

Time point	Measurements	Comments
Baseline	Demographics: age, sexPre-existing health status: Recalled SF-36 (PCS and MCS) 30 days prior to hospitalisation. Functional Co-morbidity Index (FCI) Clinical Frailty Index prior to the onset of the acute illness that resulted in this hospitalisation ICU Mobility Scale (IMS) at the time of randomisation Baseline health service utilisation questionnaireAcute illness characteristics: ICU admission diagnosis APACHE II score Duration of mechanical ventilation Duration of renal replacement therapy (if required) SOFA score from the 24 h period prior to the participant being ready for ICU discharge.Blood Samples: haemoglobin concentration; reticulocyte count; CRP concentration; blood sample for mechanistic study (if consent provided)	Recalled SF-36 completed as close to the time of enrolment as possible. For patients lacking capacity at enrolment, deferred until patients regain capacity
30 days post randomisation (7 prior or 14 days post visit date)^a (surviving patients)	Physical Component Scale (PCS) of the SF-36 Mental Component Scale (MCS) of the SF-36 Fatigue Severity Scale (FSS) New Infections (defined as any clinically suspected or proven infection occurring after randomisation requiring a new course of antibiotics) Major Adverse Cardiac Events (MACE), comprising: any confirmed stroke, myocardial infarction, or cardiovascular death Transfusion-related adverse events Haemoglobin concentration CRP (in local hospital laboratory) Blood samples for erythrogenesis and inflammatory markers (for patients providing permission for this and where feasible to collect) Survival Health service utilisation questionnaire	Follow-up can be undertaken in any clinic (eg post-ICU follow-up clinic; routine follow-up clinic) or in a clinical Research Facility or in the patient’s home or current residence or in a hospital wardFollow up undertaken by local investigator team
90 days post randomisation (±14 days)^a (surviving patients)	Physical Component Scale (PCS) of the SF-36 Mental Component Scale (MCS) of the SF-36 Fatigue Severity Scale (FSS) WHODAS ADLs questionnaire New Infections (defined as any clinically suspected or proven infection occurring after randomisation requiring a new course of antibiotics) Major Adverse Cardiac Events (MACE), comprising: any confirmed stroke, myocardial infarction, or cardiovascular death Transfusion-related adverse events Survival Health service utilisation questionnaire	Follow up by postal questionnaire followed by telephone consultation if required.Survival determined by local team consulting electronic health record.Follow up undertaken by local investigator team.
180 days post randomisation (±14 days)^a (surviving patients)	Physical Component Scale (PCS) of the SF-36 Mental Component Scale (MCS) of the SF-36 Fatigue Severity Scale (FSS) Survival Health service utilisation questionnaire	Follow up by postal questionnaire followed by telephone consultation if required.Survival determined by local team consulting electronic health record.Follow up by local investigator team.
5 years post randomisation^b	Survival (via data linkage) Health care resource and costs (via data linkage)	5-year outcomes determined by data linkage without needing to contact patients.

If follow up in person is not possible or feasible, or is declined by the patient, then research teams should make efforts to obtain questionnaire data either by postal questionnaire or via telephone.

The decision to undertake data linkage will be determined based on the results of the main trial and any effects seen.

Sample size

Based on previous research^2,41 we estimated mean PCS SF-36 at 90 days will be significantly reduced at 35–40 (normal population average value 50), with standard deviation (SD) of 10–15 points. We defined a minimum clinically important difference as 5 points of the PCS element of the SF-36 at 90 days (smaller differences are of unlikely importance to patients).⁴² This represents a ‘medium’ effect size of 0.42. Assuming the mean (SD) PCS SF-36 will be 38 (SD 12) with restrictive (usual) care, to detect an improvement from RBC transfusion to a mean of 43 at 90% power (two-sided P value 0.05) would require 244 patients.

In this population there is a risk of loss to follow-up for SF-36 (other than death), which in most previous trials has been between 10% and 30%. At trial outset we inflated the sample size allowing for a 20% loss to follow-up at 90 days. This required a sample size of 305 (randomised 1:1 between the two groups). Adjusting for pre-specified baseline covariates is expected to further increase study power and/or protect against extra variability.

In May 2024, follow-up completion for the primary outcome was 75% despite active attempts to maximise this. As part of the strategy to ensure internal validity and adequate study power we increased the sample size to ensure complete follow-up for at least 244 surviving patients. At this time, rates of deaths during follow-up were 12/182 patients (7%). A conservative approach was agreed, assuming loss to follow-up might be as high as 30%. This required an inflation in sample size from 305 to 346 patients to ensure at least 90% power for a difference in the SF-36 PCS of 5 points. We also decided to use change from baseline SF-36 PCS as the primary outcome, adjusting for the baseline value in the analysis, and use imputation for missing outcome data in survivors to maximise inclusion of information in the primary analysis (see below). The final sample size is 346 patients.

Analysis

All details of the statistical approaches are specified in the Statistical Analysis Plan (SAP; Supplemental Appendix 1).

Primary outcome

The primary analysis is an intention-to-treat (ITT) comparison of change from baseline PCS SF-36 at 90 days post-randomisation between randomised groups, using a mixed effects linear model, with site as a random effect, and adjustment for age, Functional Comorbidity Index, pre-ICU recalled PCS SF-36, ICU length of stay, ICU mobility scale, and baseline inflammation (assessed by CRP concentration). Given that 20%–30% of randomised patients are expected to have missing SF-36 data, we will use multiple imputation by chained equations (MICE), employing the passive approach model (or an alternative appropriate model) under the missing at random assumption. We will also conduct sensitivity type analyses to investigate the robustness of the findings to the assumptions behind this model.

In addition to the full ITT approach for the primary endpoint, (i.e. the real-world effect of the intervention amongst all those offered the intervention) we will also estimate the treatment efficacy (the idealised effect achieved by those fully compliant with their randomised treatment and all aspects of the protocol).

Moderation analyses of PCS SF-36 scale

We will conduct pre-specified moderation analyses on the primary endpoint. These will fit interaction terms for the four baseline parameters and treatment group, namely: (a) Co-morbidity score (FCI); (b) pre-ICU recalled SF-36; (c) ICU mobility scale; and (d) baseline level of systemic inflammation (CRP concentration). We will first explore the nature of the potential interaction by fitting each baseline parameter as a continuous interaction with treatment, and if warranted, look at suitable categorisation (e.g. above vs below the median) for ease of clinical interpretation.

Mediation analysis of PCS SF-36 scale

We will conduct two mediation analyses on the principal analysis of PCS SF-36 model. These will explore how these factors may mediate the intervention effect. First, we will examine the role of Hb at 30 days post-randomisation, as a measure of anaemia severity and persisting impaired erythropoiesis. Specifically, we will assess whether the degree of anaemia correction at 30 days, according to group allocation, relates to the endpoint. Second, we will explore whether CRP concentration at 30 days post-randomisation, as an indicator of persistent inflammation within each group, mediates the impact of RBC transfusions on the endpoint. This is relevant because RBC transfusions may be pro- or anti-inflammatory, superimposed on the persisting inflammation that often occurs in the post-ICU period.

We also plan to measure a range of biomarkers that may provide additional information about the mechanisms by which the intervention influences the primary outcome, based on blood sampling at baseline and 30 days. These will include effects of the intervention on iron status, marrow erythrogenesis (red blood cell production), inflammation, and immune suppression. These exploratory analyses will be underpinned by a Directed Acyclic Graph (DAG) and pre-specified analysis plan (Supplemental Appendix 2).

Secondary outcomes

Secondary outcomes will be analysed using models appropriate for the type of outcome (e.g. for continuous outcomes as per the primary endpoint/principal analysis; mortality using a Cox proportional hazards model, binary outcomes with adjusted logistic regressions).

Health economic evaluation

The cost-effectiveness of the intervention compared with ‘usual care’ will take the format of a within-trial cost-effectiveness analysis and use a cost-utility analysis framework. The primary health economic outcome measure will be the incremental cost per QALY at 180 days, and a further analysis will assess the cost per Life Year at 180 days. The estimation of QALYs will rely on the SF-6D, derived from the SF-36.⁴³ If appropriate based on trial findings, the impact of the intervention on the downstream costs, health outcomes and cost effectiveness will be assessed over a 5-year time horizon. The perspective for analysis will be that of the UK National Health Service and Personal Social Services as per NICE guidance (https://www.nice.org.uk/process/pmg36), with a secondary analysis taking a wider societal perspective (e.g. including indirect costs in terms of the costs of lost productivity).

A detailed Health Economic Analysis Plan (HEAP) will be finalised before database lock.

Study within a trial (SWAT): sub-study to explore reasons that patients decline consent to participate

During the trial conduct it has become apparent that consent rates among eligible patients is lower than expected, at around 40% of patients approached for consent. The reason for this unexpected low agreement to participate is uncertain, but important to the validity of this and future trials in ICU survivor populations. With the advice and support of patient collaborators and the Trial Steering Committee, we developed a provisional logic model hypothesising possible reasons for this unexpectedly low consent rate (Supplemental Appendix 3). We designed a mixed methods study including a patient focus group, survey of research staff, and interviews with research staff as a SWAT to increase understanding of patient decision-making in relation to the trial. This is summarised in Supplemental Appendix 3.

Ethics and trial oversight

The trial sponsor is ACCORD (https://www.accord.scot/). Trial oversight is by an independent Trial Steering Committee and Data and Safety Monitoring Committee. Ethical approval is from the Scotland a Research Ethics Committee (REC; Scotland; 19/SS/0105) and the Wales REC4 (England, Wales; 19/WA/0276).

Role of funder and sponsor

The funder had no role in trial design. The sponsor reviewed the trial protocol and ethical submission and approved it (including amendments) prior to submission. Neither will have any role in analysis, interpretation, or the decision to submit for publication.

Current status

Recruitment started in September 2020 at one centre (Edinburgh), but the trial was impacted by the COVID19 pandemic, and wider roll out was postponed. Further site set-up occurred from October 2022. Total sites opened during the trial is 19, across England (15 sites), Scotland (3 sites), and Wales (1 site). Recruitment is currently 313 (April 10th 2025), and is expected to complete by June 30th 2025.

Supplemental Material

sj-docx-1-inc-10.1177_17511437251374884 – Supplemental material for Anaemia management with red blood cell transfusion to improve post-intensive care disability: Protocol for the ABC post-ICU randomised controlled trial

Supplemental material, sj-docx-1-inc-10.1177_17511437251374884 for Anaemia management with red blood cell transfusion to improve post-intensive care disability: Protocol for the ABC post-ICU randomised controlled trial by Timothy S. Walsh, Lydia Emerson, Jo Singleton, Rachel Locherty, David Hope, Stephanie Cholbi, Annabel Giddings, Alix Macdonald, Nazir Lone, Annemarie B. Docherty, Gillian Mead, Simon J. Stanworth, Alexander Drakesmith, Noemi B. A. Roy, Peter Hall, Aileen R. Neilson, Roz Pollock, Aryelly Rodriguez, John Norrie, Christopher J. Weir, Akshay Shah and David Griffith in Journal of the Intensive Care Society

Footnotes

ORCID iDs

Timothy S Walsh

Lydia Emerson

Nazir Lone

Aryelly Rodriguez

Akshay Shah

Ethical considerations

Scotland A Research Ethics Committee (REC; Scotland; 19/SS/0105); Wales REC4 (England, Wales; 19/WA/0276).

Author contributions

Conception and acquiring funding: Walsh; Docherty; Drakesmith; Griffith; Hall; Lone; Mead; Neilson; Norrie; Stanworth; Roy.

Trial design: Walsh; Docherty; Drakesmith; Emerson; Cholbi; Giddings; Griffith; Hall; Hope; Locherty; Lone; Macdonald; Mead; Neilson; Norrie; Pollock; Rodriguez; Shah; Singleton; Stanworth; Roy; Weir.

Statistical design and analysis: Walsh; Griffith; Lone; Norrie; Rodriguez; Weir.

Trial oversight: Walsh; Docherty; Drakesmith; Emerson; Cholbi; Giddings; Griffith; Hall; Hope; Locherty; Lone; Macdonald; Mead; Neilson; Norrie; Pollock; Rodriguez; Shah; Singleton; Stanworth; Roy; Weir.

Health economic evaluation: Hall; Neilson.

All authors reviewed the final manuscript and approved it prior to submission.

Funding

The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the J P Moulton Foundation. The trial is being delivered across the United Kingdom with infrastructure support from the National Institute of Health and Care Research (NIHR) Clinical Research Network.

Declaration of conflicting interests

The authors declared the following potential conflicts of interest with respect to the research, authorship, and/or publication of this article: Akshay Shah has received consultancy fees from Pharmacosmos (UK) outside of the submitted work.

Supplemental Material

Supplemental material for this article is available online.

References

Needham

Davidson

Cohen

, et al. Improving long-term outcomes after discharge from intensive care unit: report from a stakeholders’ conference. Crit Care Med 2012; 40(2): 502–509.

Walsh

Salisbury

Merriweather

, et al. Increased Hospital-based physical rehabilitation and information provision after Intensive Care Unit Discharge: the RECOVER randomized clinical trial. JAMA Intern Med 2015; 175(6): 901–910.

Griffith

Salisbury

Lee

, et al. The burden of specific symptoms reported by survivors after critical illness. Am J Respir Crit Care Med 2018; 197: 269–272.

Nedergaard

Haberlandt

Reichmann

, et al. Patients’ opinions on outcomes following critical illness. Acta Anaesthesiol Scand 2018; 62(4): 531–539.

Bench

Stayt

Shah

, et al. Prevalence and experience of fatigue in survivors of critical illness: a mixed-methods systematic review. Anaesthesia 2021; 76(9): 1233–1244.

Ohtake

Lee

Scott

, et al. Physical impairments associated with post-intensive care syndrome: systematic review based on the World Health Organization’s International Classification of Functioning, Disability and Health Framework. Phys Ther 2018; 98(8): 631–645.

Hodgson

Udy

Bailey

, et al. The impact of disability in survivors of critical illness. Intensive Care Med 2017; 43(7): 992–1001.

Lone

Gillies

Haddow

, et al. Five-year mortality and hospital costs associated with surviving intensive care. Am J Respir Crit Care Med 2016; 194(2): 198–208.

Lone

Lee

Salisbury

, et al. Predicting risk of unplanned hospital readmission in survivors of critical illness: a population-level cohort study. Thorax 2019; 74: 1046–1054.

10.

Walsh

Saleh

EE.

Anaemia during critical illness. Br J Anaesth 2006; 97(3): 278–291.

11.

Walsh

Saleh

Lee

, et al. The prevalence and characteristics of anaemia at discharge home after intensive care. Intensive Care Med 2006; 32(8): 1206–1213.

12.

Pagano

Dennis

Idemudia

, et al. An analysis of quality of life and functional outcomes as reported in randomized trials for red cell transfusions. Transfusion 2023; 63(11): 2032–2039.

13.

Carson

Stanworth

Guyatt

, et al. Red blood cell transfusion: 2023 AABB international guidelines. JAMA 2023; 330(19): 1892–1902.

14.

Turgeon

Fergusson

Clayton

, et al. Liberal or restrictive transfusion strategy in patients with traumatic brain injury. N Engl J Med 2024; 391(8): 722–735.

15.

Taccone

Rynkowski

Møller

, et al. Restrictive vs Liberal transfusion strategy in patients with acute brain injury: the TRAIN Randomized Clinical Trial. JAMA 2024; 332(19): 1623–1633.

16.

Docherty

O’Donnell

Brunskill

, et al. Effect of restrictive versus liberal transfusion strategies on outcomes in patients with cardiovascular disease in a non-cardiac surgery setting: systematic review and meta-analysis. BMJ 2016; 352: i1351.

17.

Carson

Fergusson

Noveck

, et al. Restrictive versus liberal transfusion in myocardial infarction - a patient-level meta-analysis. NEJM Evid 2025; 4(2): EVIDoa2400223.

18.

Simon

Craswell

Thom

, et al. Outcomes of restrictive versus liberal transfusion strategies in older adults from nine randomised controlled trials: a systematic review and meta-analysis. Lancet Haematol 2017; 4(10): e465–e74.

19.

Docherty

Walsh

TS.

Anemia and blood transfusion in the critically ill patient with cardiovascular disease. Crit Care 2017; 21(1): 61.

20.

Shah

Stanworth

Lee

, et al. Prevalence, management and outcomes associated with anaemia in ICU survivors: a retrospective study. Anaesthesia 2021; 76(10): 1421–1423.

21.

Nemeth

Ganz

Hepcidin and iron in health and disease. Annu Rev Med 2023; 74: 261–277.

22.

Griffith

Vale

Campbell

, et al. Persistent inflammation and recovery after intensive care: a systematic review. J Crit Care 2016; 33: 192–199.

23.

Griffith

Lewis

Rossi

, et al. Systemic inflammation after critical illness: relationship with physical recovery and exploration of potential mechanisms. Thorax 2016; 71(9): 820–829.

24.

Oeyen

Vandijck

Benoit

, et al. Quality of life after intensive care: a systematic review of the literature. Crit Care Med 2010; 38(12): 2386–2400.

25.

Rygård

Holst

Wetterslev

, et al. Long-term outcomes in patients with septic shock transfused at a lower versus a higher haemoglobin threshold: the TRISS randomised, multicentre clinical trial. Intensive Care Med 2016; 42(11): 1685–1694.

26.

Walsh

Boyd

Watson

, et al. Restrictive versus liberal transfusion strategies for older mechanically ventilated critically ill patients: a randomized pilot trial. Crit Care Med 2013; 41(10): 2354–2363.

27.

Griffith

Salisbury

Lee

, et al. Determinants of health-related quality of life after ICU: Importance of patient demographics, previous comorbidity, and severity of illness. Crit Care Med 2018; 46(4): 594–601.

28.

Oeyen

Vermeulen

Benoit

, et al. Development of a prediction model for long-term quality of life in critically ill patients. J Crit Care 2018; 43: 133–138.

29.

Vlaar

Oczkowski

de Bruin

, et al. Transfusion strategies in non-bleeding critically ill adults: a clinical practice guideline from the European Society of Intensive Care Medicine. Intensive Care Med 2020; 46(4): 673–696.

30.

Coz Yataco

Soghier

Hébert

, et al. Red blood cell transfusion in critically ill adults: an American College of Chest Physicians Clinical Practice Guideline. Chest 2025; 167(2): 477–489.

31.

Parry

Knight

Connolly

, et al. Factors influencing physical activity and rehabilitation in survivors of critical illness: a systematic review of quantitative and qualitative studies. Intensive Care Med 2017; 43(4): 531–542.

32.

Shah

Chester-Jones

Dutton

, et al. Intravenous iron to treat anaemia following critical care: a multicentre feasibility randomised trial. Br J Anaesth 2022; 128(2): 272–282.

33.

Geneen

Kimber

Doree

, et al. Efficacy and safety of intravenous iron therapy for treating anaemia in critically ill adults: a rapid systematic review with meta-analysis. Transfus Med Rev 2022; 36(2): 97–106.

34.

Litton

French

Herschtal

, et al. Iron and erythropoietin to heal and recover after intensive care (ITHRIVE): a pilot randomised clinical trial. Crit Care Resusc 2023; 25(4): 201–206.

35.

Walsh

TS.

Active prevention and treatment of post-critical illness anemia-is it time for large effectiveness trials?

JAMA Netw Open 2025; 8(3): e252368.

36.

Fan

Gifford

Chandolu

, et al. The functional comorbidity index had high inter-rater reliability in patients with acute lung injury. BMC Anesthesiol 2012; 12: 21.

37.

Ware

Jr. Sherbourne

. The MOS 36-item short-form health survey (SF-36). I. Conceptual framework and item selection. Med Care 1992; 30(6): 473–483.

38.

Tipping

Bailey

Bellomo

, et al. The ICU mobility scale has construct and predictive validity and is responsive. A multicenter observational study. Ann Am Thorac Soc 2016; 13(6): 887–893.

39.

World Health Organisation. Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. WHO. https://www.who.int/publications/i/item/WHO-NMH-NHD-MNM-11.1 (2011, accessed 25 May 2025).

40.

Krupp

LaRocca

Muir-Nash

, et al. The fatigue severity scale. Application to patients with multiple sclerosis and systemic lupus erythematosus. Arch Neurol 1989; 46(10): 1121–1123.

41.

Bateman

McArdle

Walsh

TS.

Time course of anemia during six months follow up following intensive care discharge and factors associated with impaired recovery of erythropoiesis. Crit Care Med 2009; 37(6): 1906–1912.

42.

Samsa

Edelman

Rothman

, et al. Determining clinically important differences in health status measures: a general approach with illustration to the Health Utilities Index mark II. Pharmacoeconomics 1999; 15(2): 141–155.

43.

Brazier

Roberts

Deverill

The estimation of a preference-based measure of health from the SF-36. J Health Econ 2002; 21(2): 271–292.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

4.13 MB