Confidential: For Review Only - BMJ · Confidential: For Review Only 1 1 The impact of restrictive versus liberal transfusion strategies on patient 2 outcomes in patients with cardiovascular

Confidential: For Review O

nly

The impact of transfusion thresholds on mortality and

cardiovascular events in patients with cardiovascular disease (non-cardiac surgery): A Systematic Review and

Meta-analysis

Journal: BMJ

Manuscript ID BMJ.2015.030619

Article Type: Research

BMJ Journal: BMJ

Date Submitted by the Author: 25-Nov-2015

Complete List of Authors: Docherty, Annemarie; University of Edinburgh, Anaesthesia and Critical Care O'Donnell, Robert; Royal Infirmary Edinburgh, Anaesthesia and Critical Care Brunskill, Susan; John Radcliffe Hospital, Systematic Review Initiative, NHS Blood and Transplant Trivella, M; UK Cochrane centre,

Doree, Carolyn; John Radcliffe Hospital, Systematic Review Initiative, NHS Blood and Transplant Holst, Lars; Copenhagen University Hospital, Rigshospitalet, Dept. of intensive care Parker, Martyn; Peterborough and Stamford Hospitals NHS Trust, Orthopaedics Gregersen, Merete; Aarhus University, Geriatrics Pinheiro de Almeida, Juliano; Cancer Institute (ICESP), Hospital das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Surgical Intensive Care Unit and Department of Anesthesiology Walsh, Timothy; University of Edinburgh, Anaesthesia, Critical Care and Pain Medicine

Stanworth, Simon; NHS Blood and Transplant/Oxford Radcliffe Hospitals Trust, Haematology and Transfusion Medicine

Keywords: red cell transfusion, cardiovascular disease, randomised controlled trial, systematic review

https://mc.manuscriptcentral.com/bmj

BMJ


nly

1

The impact of restrictive versus liberal transfusion strategies on patient 1

outcomes in patients with cardiovascular disease excluding those undergoing 2

cardiac surgery: A Systematic Review and Meta-analysis 3

Annemarie B Docherty1,2*, Rob O’Donnell2, Susan Brunskill3, Marialena Trivella3, Carolyn Doree4, 4

Lars Holst5, Martyn Parker

6, Merete Gregersen

7, Juliano Pinheiro de Almeida

8, Timothy S Walsh

1,2, 5

Simon J Stanworth3,9

6

1. Centre for Inflammation Research, University of Edinburgh, Little France Crescent, 7

Edinburgh, EH16 4SA. +44 131 242 6395 8

2. Critical Care Department, Royal Infirmary Edinburgh, Little France Crescent, Edinburgh, 9

EH16 4SA 10

3. Systematic Review Initiative, NHS Blood and Transplant, John Radcliffe Hospital, Oxford, 11

UK 12

4. Centre for Statistics in Medicine, University of Oxford, Oxford, UK 13

5. Department of Intensive Care, , Copenhagen University Hospital, Rigshospitalet, 14

Copenhagen, Denmark 15

6. Department of Orthopaedics, Peterborough and Stamford Hospitals NHS Trust, Peterborough, 16

UK 17

7. Department of Geriatrics, Aarhus University, Aarhus, Denmark 18

8. Surgical Intensive Care Unit and Department of Anesthesiology, Cancer Institute, Hospital 19

das Clinicas da Faculdade de Medicina da Universidade de Sao Paulo, Sao Paulo, Brazil 20

9. Department of Haematology, NHS Blood and Transplant/Oxford University Hospitals NHS 21

Trust, Oxford, UK 22

*Corresponding author. Correspondence to AB Docherty [email protected] 23

Key Words: Red cell transfusion, cardiovascular disease, randomised controlled trial, systematic 24

review, meta-analysis 25

Word count: 4,377 26

Page 1 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

2

ABSTRACT 27

Objectives: To compare patient outcomes with restrictive vs liberal red cell transfusion strategies in 28

patients with cardiovascular disease, excluding those undergoing cardiac surgery. 29

Design: Systematic review with meta-analyses of randomised controlled trials. 30

Data Sources: In-hospital red cell transfusion threshold randomised controlled trials. We searched (to 31

02/11/2015), CENTRAL; MEDLINE; Embase; CINAHL; PUBMED; LILACS; NHSBT Transfusion 32

Evidence Library; ClinicalTrials.gov; WHO International Clinical Trials Registry Platform; ISRCTN 33

Register; EU Clinical Trials Register. We contacted authors for relevant data whenever possible. 34

Trial selection: Published and unpublished randomised controlled trials that evaluated a restrictive vs 35

liberal transfusion threshold and that included patients with cardiovascular disease. 36

Data extraction and synthesis: Data extraction was completed in duplicate. Risk of bias was 37

assessed using Cochrane methodology. Relative risk ratios with 95% CI were presented in all meta-38

analyses. Mantel-Haenzel random effects models were used to pool the risk ratios. 39

Main outcome measures: Thirty-day mortality, and cardiovascular events 40

Results: We identified 41 trials; of these 7 trials included data on patients with cardiovascular disease. 41

Data from a further 4 trials enrolling patients with cardiovascular disease was obtained from the 42

authors. In total, 11 trials enrolling patients with cardiovascular disease (n=3,033) were included for 43

meta-analysis (restrictive transfusion threshold, n=1,514 patients; liberal transfusion threshold, n= 44

1,519). The pooled risk ratio for the association between transfusion thresholds and 30 day mortality 45

was 1.15 (95% CI 0.88 to 1.50, P=0.50) with little heterogeneity (I2=14%). There was an increased 46

risk of acute coronary syndrome (ACS) in patients managed with a restrictive compared with a liberal 47

transfusion threshold (9 trials; RR 1.78, 95% CI 1.18 to 2.70, P=0.01, I2= 0%). 48

Conclusions: The results show that it may not be safe to use a restrictive transfusion trigger below 49

80g/l in patients with ongoing ACS or chronic cardiovascular disease. Effects on mortality and other 50

outcomes are uncertain. These data support the use of a more liberal transfusion threshold (greater 51

Page 2 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

3

than 80g/l) for patients with both acute and chronic cardiovascular disease until adequately powered 52

high quality randomised trials have been undertaken in patients with cardiovascular disease. 53

Trial registration: PROSPERO: CRD42014014251(1); http://www.crd.york.ac.uk/PROSPERO/). 54

Word count: 342 55

56

“What this paper adds” 57

“What is already known on this subject” 58

Restrictive red cell transfusion policies are recommended as safe for the majority of hospital patients 59

with anaemia. 60

Uncertainty exists for patients with cardiovascular disease, in whom the heart may be more 61

susceptible to limited coronary oxygen supply. 62

No previous systematic reviews have specifically compared outcomes for patients with cardiovascular 63

disease outwith the cardiac surgery setting, and guidelines acknowledge the paucity of evidence in 64

this area. 65

“What this study adds” 66

This review indicates that restrictive blood transfusion strategies may not be as safe as more liberal 67

transfusion strategies for patients with co-existing cardiovascular disease in non-cardiac surgery 68

settings. 69

Specifically, we have shown an increased risk of acute coronary syndrome with restrictive triggers 70

below 80g/l. These data support the use of a more liberal transfusion threshold (greater than 80g/l) for 71

patients with both acute and chronic cardiovascular disease, until adequately powered high quality 72

randomised trials have been undertaken in this patient population. 73

Page 3 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

4

INTRODUCTION 74

Approximately 7 million people in the UK have cardiovascular disease(2), and it is a prevalent 75

comorbidity among hospitalised patients. In observational studies anaemia is associated with worse 76

outcomes in patients who have both acute and chronic cardiovascular disease, but it is unclear 77

whether this association is causal or whether correction with red cell transfusions modifies this 78

relationship(3–6). Anaemia both decreases the oxygen content of the blood supplied to the 79

myocardium and may increase myocardial oxygen demand because a higher cardiac output is required 80

to maintain adequate systemic oxygen delivery(7). The heart extracts a high proportion of the oxygen 81

supplied via the coronary arteries, and therefore this circulation is potentially at higher risk from the 82

combination of atheroma-related flow limitation and anaemia. Hypotension, tachycardia and the 83

requirement for catecholamine use, for example during critical illness or major surgery, can further 84

compromise oxygen supply-demand balance resulting in myocardial injury. This has been termed type 85

2 myocardial infarction(8). Troponin release, a biomarker of myocardial injury, is associated with 86

higher mortality in critically ill and perioperative populations(9–11). 87

Systematic reviews of randomised trials of liberal versus restrictive blood transfusion strategies 88

support a general default trigger of around 70g/l for most patient groups(12–14), and this is reflected 89

in recent guidelines advocating restrictive use of blood transfusions(15–18). These have highlighted 90

the lack of evidence and uncertainty regarding best practice for patients with acute or chronic 91

cardiovascular disease(15–18). No systematic reviews have specifically compared outcomes for 92

patients with chronic cardiovascular disease undergoing non-cardiac surgery or other treatments such 93

as intensive care. A recent systematic review restricted to patients undergoing cardiac surgery 94

suggested better outcomes with more liberal transfusions, highlighting the potentially important 95

interaction between anaemia, blood transfusions and outcomes for patients with cardiovascular 96

disease(19). The National Institute for Health and Care Excellence (NICE) blood transfusion 97

guideline, published in November 2015, stated that the optimal transfusion threshold for patients 98

with ongoing acute coronary syndrome was 80-100g/l, but made no specific recommendation for 99

Page 4 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

5

patients with chronic cardiovascular disease and highlighted the need for further research in this 100

specific population(20). 101

We conducted a systematic review and meta-analysis, for the first time, assessing the effect of 102

restrictive vs liberal red cell transfusion strategies on patient outcomes restricted to adult patients with 103

cardiovascular disease excluding patients undergoing cardiac surgery. 104

105

METHODS 106

This systematic review was conducted according to the protocol registered with PROSPERO 107

(registration no: CRD42014014251(1); http://www.crd.york.ac.uk/PROSPERO/). We followed 108

methods defined in the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-109

Analyses) guidelines statement(21). 110

Eligibility criteria 111

We included only randomised controlled trials (RCTs). RCTs were eligible for inclusion if they 112

evaluated the effectiveness of any policy involving the use of a trigger or transfusion threshold based 113

on haemoglobin concentration (including haematocrit) for guiding allogeneic red cell transfusion. 114

Control group patients were required to be transfused at a higher haemoglobin concentration or 115

haematocrit. We considered trials including adult patients (≥18yrs) except those involving cardiac 116

surgery because this is a distinct group of patients, whose cardiovascular risk has been significantly 117

altered by their procedure(1) (Figure 1). We excluded children and neonates due to the low prevalence 118

of cardiovascular disease. 119

We defined Cardiovascular disease in our protocol as: known Coronary Artery Disease (CAD): Acute 120

Coronary Syndrome (ACS), chronic Ischaemic Heart Disease (IHD); or other Cardiovascular Disease 121

- Cerebrovascular Accident (CVA), Transient Ischaemic Attack (TIA), Peripheral Vascular Disease 122

(PVD). We defined Acute Coronary Syndrome as: ST elevation myocardial infarction, non-ST 123

elevation myocardial infarction, or unstable angina. A summary of the definitions for cardiovascular 124

disease used by authors of included trials is presented in Table E1. 125

Page 5 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

6

Search strategy 126

We did not restrict our search by language, date or publication status. We updated a search strategy 127

we conducted in September 2009 reviewing the overall use of red blood cell (RBC) transfusions(22). 128

The present search included two changes: a) in CENTRAL there was a date restriction and b) the 129

following search method was used in MEDLINE and Embase: i) the original search strategies + 130

original RCT filters were re-run up until the end of 2008; ii) the new strategies + new RCT filters 131

were run for all years; iii) the results of i) were then removed from the new search results. Once all the 132

search results had been downloaded into bibliographic software, all previously screened references 133

from the overview of the use of RBC transfusion, along with any duplicates, were removed. Search 134

strategies are available in the online data supplement. 135

The date of the last search was 2nd

November 2015 for the following databases: 136

CENTRAL (The Cochrane Library Issue 8, 2014): publication years from 2009-2014; MEDLINE 137

(1946 onwards); Embase (1974 onwards); CINAHL (1937 onwards); PUBMED (epublications only); 138

LILACS (2009-2014); TRANSFUSION EVIDENCE LIBRARY (1980 onwards); Web of Science 139

(Conference Proceedings Citation Index- Science (CPCI-S) -1990 to present). 140

Ongoing Studies were searched for on five registries: ClinicalTrials.gov; WHO International Clinical 141

Trials Registry Platform (ICTRP); ISRCTN Register; European Union Clinical Trials Register 142

(https://www.clinicaltrialsregister.eu/ctr-search) and the Hong Kong Clinical Trials Registry. All sites 143

were searched on 2nd November 2015. For detailed information regarding search strategies see 144

supplementary appendix 1. 145

Data extraction 146

Trial selection 147

Two authors (AD and RO) independently reviewed all titles and abstracts identified (Figure 1) against 148

the pre-specified eligibility criteria. Any disagreements were resolved by discussion with the other 149

authors. All publications reporting a valid transfusion threshold RCT where inclusion criteria 150

indicated that cardiovascular patients were included were considered. We contacted the authors of all 151

Page 6 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

7

eligible trials for which either cardiovascular subgroups or a high proportion of patients with 152

cardiovascular disease were included. We requested data for the cardiovascular patients in these trials. 153

For the trials that included patients both with and without cardiovascular disease, we looked at 154

whether they stratified their randomisation by the presence or absence of cardiovascular disease. 155

We extracted data using a form piloted before the study. Two authors (AD and RO) independently 156

extracted data on trial characteristics, primary and secondary outcomes, cardiac-specific morbidity 157

and general morbidity. A third author (SS) checked for discrepancies between the independent data 158

extraction, and disagreements were resolved by discussion between the three authors. Our primary 159

outcome was mortality at 30 days. We also extracted mortality data at 60 days, intensive care and 160

hospital mortality, and other mortality as defined by authors. 161

Data on cardiovascular events were categorised as Acute Coronary Syndrome (ACS), Acute 162

Pulmonary Oedema (APO), peripheral ischaemia and thrombotic events wherever possible. The 163

category of ACS included myocardial infarction (MI), acute coronary syndrome, and cardiac arrest. 164

Measures of general morbidity were use of packed RBCs, adverse transfusion reactions, incidence of 165

in-hospital infections, measures of organ dysfunction, duration of ICU/hospital stay, invasive 166

ventilation, haemodynamic support, and renal support. 167

Risk of bias assessment 168

We assessed the risk of bias using the method outlined in the Cochrane Collaboration Handbook for 169

Systematic Reviews of Interventions(23). Risk of bias was assessed as high, low and unclear risk for 170

each of: selection bias, performance bias, detection bias, attrition bias and reporting bias. We 171

specifically assessed blinding for the outcomes of cardiovascular events. 172

Grading quality of evidence 173

We assessed the quality of evidence for mortality, acute coronary syndrome, and acute pulmonary 174

oedema according to GRADE methodology for risk of bias, inconsistency, indirectness, imprecision, 175

and publication bias; classified as very low, low, moderate, or high(24). 176

Page 7 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

8

Data synthesis and analysis 177

All statistical analyses were performed using Review Manager 5 (RevMan)(25). Meta-analysis was 178

undertaken where there were sufficient data. We used a random-effect model as we anticipated that 179

there would be substantial clinical heterogeneity. We reported relative risk ratios (RR) for 180

dichotomous outcomes, with 95% confidence intervals (CI). Where reported, we described non-181

parametric measures with median and interquartile ranges. 182

We included one cluster randomised trial(26). However, we had no information regarding which 183

clusters the patients with cardiovascular disease were in, and the intraclass correlation coefficient 184

(ICC) was 0.001 for mortality, suggesting that only 0.1% of the variance was due to the effect of the 185

trial site, and 99.1% was due to differences between patients. We performed a sensitivity analysis 186

without taking the clustering into account, and this made no difference to our results. We therefore 187

included these data as unique patient data. 188

Assessment of heterogeneity 189

Assessment of clinical heterogeneity included consideration of participant characteristics (eg acute 190

coronary syndrome vs chronic cardiovascular disease), and the clinical setting (critical care vs 191

orthopaedics vs acute coronary syndrome). We undertook a subgroup analysis of patients with chronic 192

cardiovascular disease, excluding trials including patients with ongoing acute coronary syndromes. 193

There was insufficient data to undertake the pre-planned subgroup analyses of critical care trials, or 194

acute coronary syndromes. 195

We assessed statistical heterogeneity of treatment effects between trials using the Chi2 test. We used 196

the I2 statistic to quantify the percentage of variability that was due to heterogeneity (we defined 197

heterogeneity of >50% as moderate heterogeneity and >80% as substantial). 198

Patient Involvement 199

There was no direct patient involvement in this systematic review. 200

201

Page 8 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

9

RESULTS 202

Search Results 203

The search retrieved 9,462 results (of which 283 were ongoing RCTs), which were reduced to 6,520 204

results once duplicates were removed. Previously screened references were removed and 3,955 titles 205

and abstracts (3,832 completed trials and 123 ongoing RCTs), were screened for eligibility. Of these, 206

41 completed trials were eligible for full text screening (with five of the ongoing trials being 207

potentially eligible for inclusion on their completion(27–31)). 208

Thirty trials were ineligible for inclusion. Six trials were ineligible on the basis that they specifically 209

excluded patients with signs or symptoms of heart disease, cardiac disease with New York Heart 210

Association Class (NYHA) II or above, and American Society of Anaesthetists Class (ASA) II or 211

worse(32–37). Two trials targeted pre-operative Haemoglobin S levels in sickle cell anaemia and 212

were therefore not relevant(38,39). Seven full text trials (40–46), and eight abstracts (27,47–53), did 213

not include cardiovascular disease as a baseline characteristic. 214

From the 41 potentially eligible trials, we were able to extract data on patients with cardiovascular 215

disease from 7 trials (n=2,796)(26,54–59). One of these trials published 30 day mortality data for 216

patients with cardiovascular disease, and the authors responded to our request for further data, namely 217

cardiovascular outcomes and general morbidity outcomes (n=32)(59). We were aware, from reported 218

baseline demographic data, that patients with cardiovascular disease were included in eleven further 219

trials(60–70), but we were unable to extract any relevant data directly from the published text. 220

Following contact with relevant authors, seeking any data they had on the patients with cardiovascular 221

disease within their trials, we were able to include data from four trials into this review(66–69). We 222

did not contact the authors of the seven trials and eight abstracts which did not mention cardiovascular 223

disease as a baseline characteristic in their trial(27,40–53). Characteristics of all 29 eligible trials that 224

are not included in this review can be found in the online data supplement (Table E1). In total, 11 225

transfusion threshold trials with patients with cardiovascular disease were included in this review 226

(n=3,033). 227

Page 9 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

10

Trial characteristics 228

The setting of the eleven included trials was varied: three orthopaedic trials(55,66,69), one trial with 229

upper gastrointestinal bleeding (UGIB)(26), two trials with acute coronary syndrome/myocardial 230

infarction(56,57), four trials in critical care(58,59,67,68), and one trial in elective aortic and infra-231

inguinal revascularisation(54). 232

Definitions of cardiovascular disease differed between trials (Table E1). Other than one trial of 233

elective aortic and infra-inguinal revascularisation(54), all trials included patients with the diagnosis 234

of ischaemic heart disease, and all but the two acute MI trials(56,57) included patients with congestive 235

cardiac failure. Other trials also included risk factors for ischaemic heart disease including peripheral 236

vascular disease(54,55,58,66,67,70), cerebrovascular disease(55,66–68), diabetes(66), and 237

hypertension(66). Trials varied from all patients having cardiovascular disease (CVD)(54–57), to pre-238

defined CVD sub-groups(26,58,68), to high proportions of patients with CVD(59,66,67,69). Included 239

trials were both multicentre (n=7)(26,55–59,68) and single centre trials (n=4)(54,66,67,69). 240

Red cell transfusion thresholds varied. The lowest restrictive threshold was 70g/l(58,59,67,68) (n=274, 241

from four trials), through 80g/l(26,55,56) (n=1,125, from three trials) and 90g/l(54) (n=50) to 97(69) 242

(n=34). One trial transfused only with symptoms of anaemia (n=55)(66), and one used haematocrit 243

concentration(57) (24%, n=24). Liberal thresholds also varied considerably: the most common 244

thresholds were 90g/l(58,59,67,68) (n=290, from four trials), and 100g/l(26,54–56,66) (n=1,221 from 245

five trials). Other thresholds were Hb 113g/l (n=25)(69) and 30% haematocrit (n=21)(57). Leucocyte 246

reduced RBCs were used in six out of the eleven trials(55–57,59,68,69). 247

Through data extraction, we were able to identify unique data for patients with cardiovascular disease 248

for 3,033 patients from eleven trials. The sample sizes of these trials varied from n=45(57) to 249

n=2015(55). Of the 3,033 patients with cardiovascular disease, 1,514 were randomised to restrictive 250

thresholds, and 1,519 to liberal thresholds. Six trials that included patients both with and without 251

cardiovascular disease did not stratify their randomisation by the presence of absence of 252

Page 10 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

11

cardiovascular disease(13,26,58,66,67,69). The characteristics of all included trials are described in 253

full in Table 1. 254

Comparison of exposure to transfusion strategy 255

Duration of the intervention from randomisation: one trial maintained the haemoglobin threshold for 256

one year post randomisation(66), two trials for 30 days(69,71), five trials until hospital 257

discharge(26,54–57), three trials until ICU discharge(58,67,68), and one trial for up to 14 258

days(59).The duration of exposure to the two strategies therefore varied considerably. 259

Exposure to allogenic blood: RBC transfusion requirements were extracted from six trials(54–260

57,59,66) (Table E3). For all six trials, patients in the restrictive arm were exposed to considerably 261

less allogeneic blood than patients in the liberal arm. In the restrictive transfusion threshold arms 262

between 20.4%(54) and 84.2%(66) of patients received no blood transfusions compared to a range of 263

0%(57,66) to12%(54) for the liberal transfusion threshold arms. Among patients who did receive 264

RBCs, the number of transfused units was lower in the restrictive transfusion threshold arms (range 265

from median 0 (IQR 0,1),(55) to a mean of 1.6 units (SD 2.0)(57)) compared with the liberal 266

transfusion threshold arms (range from a mean of 1.58 units (SD 1.13)(56) to a mean of 2.5 units (SD 267

1.3)(57)). 268

269

270

Effects on outcomes 271

Mortality 272

Data on mortality was available from all eleven trials. Thirty day mortality was given for all trials 273

except one(26), who reported 28 day mortality. There were 144 deaths (9.5%) in the restrictive 274

transfusion threshold arms, compared to 126 deaths (8%) in the liberal transfusion threshold arms 275

(pooled effect estimate: RR 1.15, 95% CI 0.88 to 1.50, P=0.50, I2=14%, 3033 patients, Figure 2). We 276

performed a subgroup analysis, including only trials where the randomisation was stratified for 277

Page 11 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

12

cardiovascular disease(54–57,59) and for this subgroup, the relative risk was 0.96 (95% CI 0.58 to 278

1.59, P=0.87, I2=14%). The sensitivity analysis in which the two trials including patients with ACS 279

were excluded(56,57) supported the result of the primary analysis (RR 1.10, 95% CI 0.88 to 1.37). 280

The GRADE quality of evidence was judged to be moderate (Table 2). 281

Two trials also presented mortality for all patients at 60 days(55,67), and three at 90 days(66,68,69), 282

however we were able to extract data on cardiovascular patients from only one trial (60 day mortality: 283

Restrictive 66/1007 (6.6%) vs Liberal 76/998 (7.6%))(55). 284

Adverse events: Cardiovascular 285

Nine trials presented data (2,609 patients) on new cardiovascular events(54–59,66–68). The definition 286

of MI varied between trials (Table E4). All trials except two (definition unclear)(58,66) required ECG 287

changes with a rise/fall of cardiac biomarkers using the Third Universal Definition of Myocardial 288

Infarction(8). Five trials also required symptoms consistent with myocardial 289

ischaemia(18,55,57,67,68). The diagnosis of MI was made by investigators in four trials(55,56,67,68), 290

clinicians in three trials(57,59,66), and was unclear in two trials(54,58). The diagnosis was blinded in 291

four trials(55,56,67,68), unblinded in three trials(57,59,66), and unclear in two trials(54,58). The 292

incidence of ACS (Figure 3) ranged from 0%(66,67) to 20.4%(56) in the restrictive transfusion 293

threshold arm and 0%(57,59,66,67) to 11.1%(56) in the liberal transfusion threshold arm. There was 294

evidence of an increased incidence of ACS in patients in the restrictive transfusion threshold arms 295

compared with patients in the liberal transfusion threshold arms (RR 1.78, 95% CI 1.18 to 2.70, 296

P=0.006, I2= 0%, Figure 3A, Restrictive: 59 events/1319 patients vs Liberal: 32 events/1290 patients). 297

This corresponds to 4.6 episodes of ACS per 100 patients when using restrictive strategies, and 2.7 298

per 100 patients when using liberal strategies. The number of patients that would need to be treated 299

with a liberal transfusion strategy in order to prevent one episode of ACS is 52. 300

For the analysis of patients with APO, three trials had a higher incidence of acute pulmonary oedema 301

in the liberal transfusion threshold arms(57,58,66), whereas one trial had a higher incidence in the 302

restrictive transfusion threshold arm(56). There was no evidence of a different risk of APO in the 303

Page 12 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

13

restrictive transfusion threshold arms in comparison to the liberal transfusion threshold arms (RR 0.63, 304

95% CI 0.22 to 1.81, P=0.39, I2=60%, Figure 3B. Restrictive: 24 events/309 patients vs Liberal: 47 305

events/340 patients)). Two trials reported no new episodes of APO(66,67) and there was only one 306

episode of APO in one trial(59). Cerebrovascular and thrombotic events were rare in both restrictive 307

and liberal threshold groups and meta-analysis was not possible. 308

A sensitivity analysis excluding trials that did not stratify randomisation based on cardiovascular 309

disease, found minimal impact on the estimates for the outcomes of ACS and acute pulmonary 310

oedema. The GRADE quality of evidence was judged to be low for ACS mainly because of the 311

serious risk of bias in outcome assessment (Table 2). Sensitivity analysis excluding the two ACS 312

trials(56,57) had minimal impact on the point estimates for this outcome of new ACS (RR 1.76, 1.10 313

to 2.81). The risk of ACS remained higher for the restrictive group on removal of the largest trial 314

which had 2016 participants(55), RR 2.07 (1.02, 4.23). 315

Adverse events: General 316

Non-cardiovascular adverse events were reported across 8 trials(18,54,55,57–59,66,69) (Table E5). 317

These endpoints were described differently in each paper due to the different clinical settings and 318

rationale of the trials. 319

Six trials reported hospital length of stay for patients with cardiovascular disease,(54–59,69). There 320

was no significant difference between restrictive and liberal transfusion threshold arms (mean 321

difference 1.24 days (95% CI -1.0 to 3.48, p=0.28, Figure E2)). Three trials reported in-hospital 322

infection, but the number of events were small,(56,66,71) (Table E2). One trial found no differences 323

in organ support in a post hoc analysis of cardiovascular patients (personal communication from 324

author(68)), and there were no events classified as adverse transfusion reactions. 325

Risk of Bias 326

The risk of bias is summarised in Figures 2 and 3. The main category for high risk of bias was the 327

lack of blinding of participants, clinical staff and research staff (identified in six trials(26,55,57–328

59,68)). The diagnosis of cardiovascular events is difficult in many of the settings in which trials took 329

Page 13 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

14

place, such as during critical illness, increasing the risk of performance bias in conjunction with 330

unblinded outcome assessors. Different definitions was another potential explanation for differing 331

prevalence between trials (Table E4). Diagnosis of cardiovascular events was made by the 332

investigators in five trials, and by unblinded clinicians in 3 trials(57,59,66). The criteria for 333

myocardial infarction was clearly defined in seven trials(54–57,59,67,68), and was unclear in two 334

trials(58,66). There was only one trial where both the definition and the outcome assessment were at 335

high risk of bias(66), but no new cardiovascular events were diagnosed in this trial and its removal did 336

not alter the analysis. 337

DISCUSSION 338

We identified data from eleven randomised trials that enrolled 3,033 patients with cardiovascular 339

disease in whom mortality data were available at 30 days, and nine trials that enrolled 2,609 patients 340

with cardiovascular disease in whom data regarding new cardiovascular events were available. A 341

restrictive transfusion threshold was associated with an increased risk of ACS in patients with 342

cardiovascular disease with low heterogeneity between trials (Moderate quality of evidence as 343

assessed by GRADE). We found no evidence of a difference in 30 day mortality between restrictive 344

and liberal transfusion threshold groups. There was no difference in the incidence of pulmonary 345

oedema between the transfusion thresholds, but heterogeneity existed between trials and the GRADE 346

quality of evidence was judged to be very low. There was no difference in hospital length of stay 347

between restrictive and liberal transfusion strategies, and other outcomes were rare, with inadequate 348

data for meta-analysis. 349

This is the first systematic review to specifically address clinical outcomes for patients with acute and 350

chronic cardiovascular disease managed with restrictive or liberal transfusions not including patients 351

undergoingcardiac surgery. Several well-conducted systematic reviews have been previously 352

published, but these did not examine patient sub-groups with cardiovascular disease(12–14). The 353

inclusion of heterogeneous populations in trials can mask potentially divergent effects in sub-354

populations(72), and effects may be amplified when trials are combined for meta-analyses. The 2012 355

Cochrane review recommended the use of a restrictive transfusion trigger, but suggested caution in 356

Page 14 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

15

patients from high-risk groups such as acute coronary syndrome(14). Similar statements were made 357

by both Holst(13) and Brunskill(12) in their systematic reviews of transfusion thresholds for sepsis 358

and patients undergoing surgery for hip fracture respectively. Evidence is limited by the under-359

representation of patients with cardiovascular disease in many RCTs. For example, only 20% of 360

patients enrolled in a large critical care trial had cardiovascular disease, compared with 29% of 361

excluded patients(58). Similarly, only 14% of patients enrolled in a trial in septic shock had 362

cardiovascular disease(68), whereas observational trials suggest around 25-30% of critical care 363

populations may have co-existing cardiac disease(9,73). 364

The previous reviews in heterogeneous populations suggest overall trends toward lower 30 day 365

mortality with restrictive practices (range of RR 0.85, 95% CI 0.70-1.03(14) to 0.92, 95% CI 0.67-366

1.26(12)), whereas the effect we observed in patients with cardiovascular disease was in the direction 367

favouring liberal transfusion, but without statistical significance (RR 1.10, 95% CI 0.84 to 1.44). We 368

specifically excluded trials in cardiac surgery from our review, because this is a distinct group of 369

patients whose cardiovascular risk has been significantly altered by their procedure. A recent large 370

multicentre RCT in cardiac surgery(74) found no difference in a composite morbidity outcome, but 371

90-day mortality rate was significantly higher in the restrictive transfusion threshold group compared 372

with the liberal group. A recent systematic review and meta-analysis restricted to cardiac surgery 373

trials also found increased mortality with restrictive transfusion thresholds(19). Another systematic 374

review(75) of perioperative RCTs of transfusion practice (including cardiac surgery) also 375

demonstrated higher mortality with a restrictive transfusion threshold trigger, although the prevalence 376

of cardiovascular disease in these trials was uncertain. These data suggest that the presence of 377

cardiovascular disease may significantly modify the effect of transfusion practice on clinical 378

outcomes, and highlight the need for better evidence for this prevalent patient group. 379

We found that new onset ACS occurred more frequently with restrictive transfusion threshold 380

practices. The pooled estimates were 2.7% for liberal versus 4.6% for restrictive practices (number 381

needed to treat approximately 52 to prevent an ACS with more liberal transfusion). The variation in 382

patient populations, transfusion strategies compared, and method of ascertaining ACS create 383

Page 15 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

16

substantial uncertainty in these estimates, but the heterogeneity between trials was low. The estimate 384

of effect was the same when the largest trial was removed. Importantly, for the majority of included 385

cases the restrictive transfusion threshold was 80g/l versus a liberal threshold of 100g/l. These 386

findings suggest that a transfusion threshold of 70g/l, which is widely recommended as the “default” 387

threshold, may not be as safe as higher thresholds for preventing ACS in patients with cardiovascular 388

disease. The safest haemoglobin threshold is uncertain and may be patient specific, but we have 389

shown potential for harm with restrictive triggers below 80g/l. Further trials are needed to inform us 390

on the optimal transfusion strategy in patients with cardiovascular disease. Myocardial injury could 391

impact on other important clinical outcomes such as length of hospital stay, quality of life, longer 392

term mortality, and healthcare costs but few trials measured these outcomes. These outcomes, 393

together with cost effectiveness, should be included in future research particularly as the cost of blood 394

transfusions is relatively low and even in the liberal transfusion threshold arm in this review patients 395

typically received only two to three units. Our review highlights the variability in diagnostic 396

definitions of ACS and the potential for ascertainment bias in clinical trials where blinding of 397

intervention groups is difficult. This resulted in the GRADE assessment as low for the evidence 398

quality, and highlight the need for further high quality research. 399

We found no effect on APO, but the numbers of trials and patients in whom this outcome was 400

reported was small and there was heterogeneity in the findings. APO can result from multiple causes, 401

including Transfusion Associated Circulatory Overload (TACO), and the potential for an effect in a 402

different direction from ACS made it important to consider these outcomes separately rather than 403

include them as a composite. Future trials should standardise diagnostic methods for both ACS and 404

APO and attempt to blind outcome assessors to group allocation. 405

Our review has a number of limitations. There was clinical diversity between trial populations, for 406

example between orthopaedic surgery and critical care. The risk-benefit balance may vary between 407

clinical situations, for example as a result of the degree and duration of physiological stress. The 408

restrictive and liberal transfusion thresholds varied between trials, and the cut-off values actually 409

overlapped (restrictive threshold 70-97g/l; liberal threshold 90-113g/l), which reduces the validity of 410

Page 16 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

17

pooling data across all trials. Exposure to anaemia would have been considerably longer in the four 411

ICU trials than in the four surgical trials, in which exposure to anaemia would have been relatively 412

short. Definitions of cardiovascular disease varied, and inclusion criteria for some trials were 413

restricted to ischaemic heart disease or acute coronary syndrome. However, the direction of effect was 414

consistently in favour of a liberal transfusion threshold for reducing new ACS events across the trials. 415

Finally, not all authors responded to our request for data for cardiovascular patients in their trials, 416

which reduced the precision of our point estimates. 417

In conclusion, this review of available evidence suggests that for anaemic patients with cardiovascular 418

disease the use of restrictive haemoglobin thresholds for blood transfusion (typically 70-80g/l) is 419

associated with higher rates of ACS than more liberal thresholds (typically 90-100g/l). No effects on 420

mortality or other important outcomes were demonstrated. The currently available quality of evidence 421

for all outcomes is low. These data support the use of a more liberal transfusion threshold (greater 422

than 80g/l) for patients with both acute and chronic cardiovascular disease, until adequately powered 423

high quality randomised trials have been undertaken in this patient population. 424

425

ACKNOWLEDGEMENTS 426

Article Information 427

Author Contributions: Dr Docherty had full access to all of the data in the study and takes 428

responsibility for the integrity of the data and the accuracy of the data analysis. 429

Study concept and design: Docherty, O’Donnell, Brunskill, Trivella, Walsh, Stanworth 430

Acquisition, analysis, or interpretation of the data: Docherty, O’Donnell, Brunskill, Doree, Walsh, 431

Stanworth, Holst, Parker, Gregersen, Almeida 432

Drafting of the manuscript: Docherty, O’Donnell, Brunskill, Walsh, Stanworth 433

Statistical Analysis: Docherty, Trivella 434

Obtained fundings: nil 435

Study supervision: Walsh, Stanworth 436

Transparency: Dr Annemarie Docherty affirms that this manuscript is an honest, accurate, and 437

transparent account of the study being reported; that no important aspects of the study have been 438

Page 17 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

18

omitted; and that any discrepancies from the study as planned (and, if relevant, registered) have been 439

explained. 440

"Dr Docherty has the right to grant on behalf of all authors and does grant on behalf of all authors, an 441

exclusive licence (or non exclusive for government employees) on a worldwide basis to the BMJ 442

Publishing Group Ltd to permit this article (if accepted) to be published in BMJ editions and any other 443

BMJPGL products and sublicences such use and exploit all subsidiary rights, as set out in our 444

licence." 445

"All authors have completed the ICMJE uniform disclosure form at 446

www.icmje.org/coi_disclosure.pdf and declare: no support from any organisation for the submitted 447

work; no financial relationships with any organisations that might have an interest in the submitted 448

work in the previous three years; no other relationships or activities that could appear to have 449

influenced the submitted work.” 450

Funding/Support: nil 451

Data sharing statement: no additional data available 452

Ethical approval: This is a systematic review and meta-analysis of anonymised published data, ethical 453

approval was not required 454

Page 18 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

19

REFERENCES 455

456

1. Docherty A, O’Donnell R, Brunskill S, Doree C, Walsh TS, Stanworth S. Transfusion 457

thresholds in patients with Cardiovascular Disease (non-cardiac surgery) [Internet]. 458

PROSPERO. 2014. Available from: 459

http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42014014251 460

2. British Heart Foundation. Cardiovascular Disease Statistics 2015 [Internet]. Oxford; 2015. 461

Available from: https://www.bhf.org.uk/publications/statistics/cvd-stats-2015 462

3. Vincent JL, Baron J-F, Reinhart K, Gattinoni L, Thijs L, Webb A, et al. Anemia and Blood 463

Transfusion in Critically Ill Patients. J Am Med Assoc. 2002;288(12):1499–507. 464

4. Carson J, Duff A, Poses R, JA B, RK S, Trout R, et al. Effect of anaemia and cardiovascular 465

disease on surgical mortality and cardiovascular disease on surgical mortality and morbidity. 466

Lancet. 1996;348:1055–60. 467

5. Wu W-C, Rathore S, Wang Y, Radford M, Krumholz H. Blood transfusion in elderly patients 468

with acute myocardial infarction. N Engl J Med. 2001;345(17):1230–6. 469

6. Horwich TB, Fonarow GC, Hamilton MA, Maclellan WR, Borenstein J. Anemia is associated 470

with worse symptoms, greater impairment in functional capacity and a significant increase in 471

mortality in patients with advanced heart failure. J Am Coll Cardiol. 2002;39(11):1780–6. 472

7. Sabatine MS, Morrow D a, Giugliano RP, Burton PBJ, Murphy S a, McCabe CH, et al. 473

Association of hemoglobin levels with clinical outcomes in acute coronary syndromes. 474

Circulation [Internet]. 2005 Apr 26 [cited 2014 Apr 3];111(16):2042–9. Available from: 475

http://www.ncbi.nlm.nih.gov/pubmed/15824203 476

8. Thygesen K, Alpert J, Jaffe A, Simoons M. Third universal definition of myocardial infarction. 477

JACC Cardiovasc Interv. 2012;60(16):1581–98. 478

9. Ostermann M, Lo J, Toolan M, Tuddenham E, Sanderson B, Lei K, et al. A prospective study 479

of the impact of serial troponin measurements on the diagnosis of myocardial infarction and 480

hospital and six-month mortality in patients admitted to ICU with non-cardiac diagnoses. Crit 481

Care [Internet]. Critical Care; 2014 Jan [cited 2014 Dec 17];18(2):R62. Available from: 482

http://www.pubmedcentral.nih.gov/articlerender.fcgi?artid=4056377&tool=pmcentrez&render483

type=abstract 484

10. Nagele P, Brown F, Gage B, Gibson D, Miller JP, Jaffe AS, et al. High-Sensitivity Cardiac 485

Troponin T in Prediction and Diagnosis of Myocardial Infarction and Long-Term Mortality 486

after Non-Cardiac Surgery. Am Heart J. 2013;166(2):325–32. 487

11. VISION Group. Myocardial Injury after Noncardiac Surgery. Anesthesiology. 2014;120:564–488

78. 489

12. Brunskill S, SL M, Shokoohi A, Pulford E, Doree C, Murphy M, et al. Red blood cell 490

transfusion for people undergoing hip fracture surgery ( Review ). Cochrane Database Syst 491

Rev. 2015;(4). 492

Page 19 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

20

13. Holst LB, Petersen MW, Haase N, Perner A, Wetterslev J. Restrictive versus liberal 493

transfusion strategy for red blood cell transfusion : systematic review of randomised trials with 494

meta-analysis and trial sequential analysis. BMJ. 2015;350:h1354. 495

14. Carson J, Carless P, Herbert P. Transfusion thresholds and other strategies for guiding 496

allogeneic red blood cell transfusion (Cochrane Review). Cochrane Libr. 2012;(2). 497

15. Association of Anaesthetists of Great Britain and Ireland. Blood Transfusion and the 498

Anaesthetist: Red Cell Transfusion 2. London; 2008. 499

16. Carson JL, Grossman BJ, Kleinman S, Tinmouth AT, Marques MB, Fung MK, et al. Clinical 500

Guideline Red Blood Cell Transfusion : A Clinical Practice Guideline From the AABB *. Ann 501

Intern Med. 2012;157(1):49–59. 502

17. Joint UK blood transfusion and tissue transplantation services professional advisory committee. 503

Handbook of Transfusion Medicine. Norfolk DD, editor. TSO; 2013. 504

18. Retter A, Wyncoll D, Pearse R, Carson D, McKechnie S, Stanworth S, et al. Guidelines on the 505

management of anaemia and red cell transfusion in adult critically ill patients. Br J Haematol 506

[Internet]. 2013 Feb [cited 2015 Aug 28];160(4):445–64. Available from: 507


19. Patel NN, Avlonitis VS, Jones HE, Reeves BC, Sterne JA, Murphy GJ. Indications for Red 509

Cell Transfusion in Cardiac Surgery: A Systematic Review and Meta-Analysis of Randomised 510

Controlled Trials and Observational Studies. Lancet Haematol. 2015;2(12):543–53. 511

20. the National Institute for Health and Clinical Excellence. Transfusion: NICE guideline NG24 512

[Internet]. 2015. Available from: http://www.nice.org.uk/guidance/ng24/evidence/full-513

guidance-2177160733 514

21. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews 515

and meta-analyses: the PRISMA statement. PLoS Med [Internet]. 2009 Jul 21 [cited 2014 Jul 516

9];6(7):e1000097. Available from: 517


type=abstract 519

22. Wilkinson KL, Brunskill SJ, Dorée C, Hopewell S, Stanworth S, Murphy MF, et al. The 520

clinical effects of red blood cell transfusions: an overview of the randomized controlled trials 521

evidence base. Transfus Med Rev [Internet]. Elsevier B.V.; 2011 Apr [cited 2015 Sep 522

23];25(2):145–155.e2. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21345644 523

23. Higgins J, Green S, (editors). Cochrane Handbook for Systematic Reviews of Interventions 524

Version 5.1.0 (updated March 2011). The Cochrane Collaboration 2011; 2011. Available from: 525

www.cochrane-handbook.org 526

24. Shunemann H, Oxman A, Vist G, Higgins J, Deeks J, Glasziou P. Interpreting results and 527

drawing conclusions. In: The Cochrane Collaboration, editor. Cochrane Handbook for 528

Systematic Reviews of Interventions Version 510 [updated March 2011]. 2011. 529

25. The Cochrane Collaboration. Review Manager (RevMan). Copenhagen: The Nordic Cochrane 530

Centre; 2014. 531

Page 20 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

21

26. Jairath V, Kahan BC, Gray A, Doré CJ, Mora A, James MW, et al. Restrictive versus liberal 532

blood transfusion for acute upper gastrointestinal bleeding (TRIGGER): a pragmatic, open-533

label, cluster randomised feasibility trial. Lancet [Internet]. 2015 May 5 [cited 2015 Jul 534

10];137–44. Available from: http://www.ncbi.nlm.nih.gov/pubmed/25956718 535

27. Wu Q. Clinical research of restrictive transfusion in elderly patients undergoing orthopedic 536

surgery [Internet]. [cited 2015 Oct 22]. Available from: 537

www.chictr.org.cn/showprojen.aspx?proj=8714 538

28. Tay J, Tinmouth A, Fergusson D, Allan D. Transfusion of red cells in hematopoietic stem cell 539

transplantation (TRIST): study protocol for a randomized controlled trial. Trials [Internet]. 540

BioMed Central Ltd; 2011 Jan [cited 2015 Oct 22];12(1):207. Available from: 541


type=abstract 543

29. Buckstein R. Red Blood Cell Transfusion Thresholds and QOL in MDS (EnhanceRBC) 544

[Internet]. [cited 2015 Oct 22]. Available from: 545

https://clinicaltrials.gov/ct2/show/NCT02099669 546

30. Palmieri TL. A Trial of Restrictive Versus Traditional Blood Transfusion Practices in Burn 547

Patients [Internet]. [cited 2015 Oct 22]. Available from: 548

https://clinicaltrials.gov/ct2/show/NCT01079247 549

31. Matot I. Restrictive Versus Liberal Red Cell Transfusion Strategy in Orthopedic-Oncology 550

Patients Undergoing Surgery - a Randomized Controlled Study [Internet]. [cited 2015 Oct 22]. 551

Available from: https://clinicaltrials.gov/ct2/show/NCT01491308 552

32. Grover M, Talwalkar S, Casbard A, Boralessa H, Contreras M, Brett S, et al. Silent myocardial 553

ischaemia and haemoglobin concentration : a randomized controlled trial of transfusion 554

strategy in lower limb arthroplasty. 2006;105–12. 555

33. Zheng H, Wu J-J, Wang J. Evaluation of effectiveness and analysis of goal-directed blood 556

transfusion in peri-operation of major orthopedic surgery in elderly patients. Exp Ther Med 557

[Internet]. 2013 Feb [cited 2015 Nov 3];5(2):511–6. Available from: 558


type=abstract 560

34. Webert KE, Cook RJ, Couban S, Carruthers J, Lee K, Blajchman MA, et al. A multicenter 561

pilot-randomized controlled trial of the feasibility of an augmented red blood cell transfusion 562

strategy for patients treated with induction chemotherapy for acute leukemia or stem cell 563

transplantation. Transfus Pract. 2008;48(January):81–91. 564

35. Prick BW, Jansen a JG, Steegers E a P, Hop WCJ, Essink-Bot ML, Uyl-de Groot C a, et al. 565

Transfusion policy after severe postpartum haemorrhage: a randomised non-inferiority trial. 566

BJOG [Internet]. 2014 Jul [cited 2015 Feb 6];121(8):1005–14. Available from: 567


36. Nielsen K, Dahl B, Johansson PI, Henneberg SW, Rasmussen LS. Intraoperative transfusion 569

threshold and tissue oxygenation: a randomised trial. Transfus Med [Internet]. 2012 Dec [cited 570

2015 Feb 6];22(6):418–25. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23121563 571

37. Liu D-X, Liu J, Zhang F, Zhang Q-Y, Xie M, Zhu Z-Q. Randomized Controlled Study on 572

Safety and Feasibility of Transfusion Trigger Score of Emergency Operations. Chin Med J 573

Page 21 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

22

(Engl) [Internet]. 2015 Jul 5 [cited 2015 Nov 3];128(13):1801–8. Available from: 574


38. Koshy M, Burd L, Wallace D, Moawad A, Baron J. Prophylactic red-cell transfusions in 576

pregnant patients with sickle cell disease. N Engl J Med. 1988;319(22):1447–52. 577

39. Haberkern BCM, Neumayr LD, Orringer EP, Earles AN, Robertson SM, Black D, et al. 578

Cholecystectomy in Sickle Cell Anemia Patients: Perioperative Outcome of 364 Cases From 579

the National Preoperative Transfusion Study. 2015; 580

40. Bergamin F, Almeida J, Park C, Osawa E, Silva J, Galas F, et al. Transfusion requirements in 581

septic shock patients: a randomized controlled trial. Crit Care. 2014;18(Suppl 1:112. 582

41. Blair S, Janvrin S, Mccollum CN, Greenhalgh RM. Effect of early blood transfusion on 583

gastrointestinal haemorrhage. Br J Surg. 1986;73:783–5. 584

42. Fortune J, Feustel P, Saifi J, Stratton H, Newell J, Shah D. Influence of Hematocrit on 585

Cardiopulmonary Function after Acute Hemorrhage. J Trauma. 1987;27(3):243–9. 586

43. Lotke P, Barth P, Garino J, Cook E. Predonated Autologous Blood Transfusions After Total 587

Knee Arthroplasty Immediate Versus Delayed Administration. J Arthroplasty. 588

1999;14(6):647–50. 589

44. Robertson CS, Hannay HJ, Yamal J-M, Gopinath S, Goodman JC, Tilley BC, et al. Effect of 590

erythropoietin and transfusion threshold on neurological recovery after traumatic brain injury: 591

a randomized clinical trial. Jama [Internet]. 2014 Jul 2 [cited 2014 Dec 2];312(1):36–47. 592

Available from: http://www.ncbi.nlm.nih.gov/pubmed/25058216 593

45. Villanueva C, Colomo A, Bosch A, Concepción M, Hernandez-Gea V, Aracil C, et al. 594

Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med [Internet]. 2013 595

Jan 3 [cited 2014 Jul 9];368(1):11–21. Available from: 596


46. Zygun D a, Nortje J, Hutchinson PJ, Timofeev I, Menon DK, Gupta AK. The effect of red 598

blood cell transfusion on cerebral oxygenation and metabolism after severe traumatic brain 599

injury. Crit Care Med [Internet]. 2009 Mar [cited 2015 Feb 11];37(3):1074–8. Available from: 600


47. Colomo A, Hernandez-Gea V, Muniz-Diaz E, Madoz P, Aracil C, Alarez-Urturi C. 602

Transfusion strategies in patients with cirrhosis and acute gastrointestinal bleeding [abstract]. 603

Hepatology. 2008;48(4 (Suppl)):413A. 604

48. Hochain P, Merle V, Tuil S, Michel P, Ducrotte P, Lerebours E, et al. Transfusion for variceal 605

bleeding in cirrhotic patients [letter]. Gut. 1996;38(1):154. 606

49. So-Osman C. A restrictive transfusion trigger is a method for blood saving in elective 607

orthopaedic surgery [abstract]. Vox Sang. 2004;87(Suppl 3):52. 608

50. Weiss GB, Patten E, Alperin JB, Hokanson JA, Bessman JD, Constanzi JJ, et al. 609

Hypertransfusion for adult acute leukaemia [letter]. Lancet. 1982;319:105. 610

Page 22 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

23

51. Park S, Nam E, Bang S, Cho E, DB S, Lee J. A randomized trial of anemia correction with two 611

different hemoglobin targets in the first-line chemotherapy of advanced gastric cancer. Cancer 612

Chemother Pharmacol. 2008;62(1):1–9. 613

52. Villarejo F, Rizzolo M, Lopez E, Domeniconi G, Arto G, Apezteguia C. Acute anemia in high 614

digestive hemorrhage. Margins of security for their handling without transfusion of red 615

globules. Acta Gastroenterol Latinoam. 1999;29(4):261–70. 616

53. Palmer J, Maciver C, Scott R, Picken M, McClelland D, Keating J, et al. Hip fracture and 617

transfusion trial (HATT) [abstract]. Transfus Med. 1998;8(Suppl 1: 36):52. 618

54. Bush RL, Pevec WC, Holcroft JW. A prospective, randomized trial limiting perioperative red 619

blood cell transfusions in vascular patients. Am J Surg [Internet]. Excerpta Medica, Inc.; 1997 620

Aug [cited 2014 Oct 8];174(2):143–8. Available from: 621

http://linkinghub.elsevier.com/retrieve/pii/S0002961097000731 622

55. Carson J, Terrin M, Noveck H, Sanders D, Chaitman BR, Rhoads GG, et al. Liberal or 623

Restrictive Transfusion in High-Risk Patients after Hip Surgery. N Engl J Med. 624

2011;365:2453–62. 625

56. Carson JL, Brooks MM, Abbott JD, Chaitman B, Kelsey SF, Triulzi DJ, et al. Liberal versus 626

restrictive transfusion thresholds for patients with symptomatic coronary artery disease. Am 627

Heart J [Internet]. Mosby, Inc.; 2013 Jun [cited 2014 Oct 8];165(6):964–971.e1. Available 628

from: 629


type=abstract 631

57. Cooper H a, Rao S V, Greenberg MD, Rumsey MP, McKenzie M, Alcorn KW, et al. 632

Conservative versus liberal red cell transfusion in acute myocardial infarction (the CRIT 633

Randomized Pilot Study). Am J Cardiol [Internet]. Elsevier Inc.; 2011 Oct 15 [cited 2014 Oct 634

8];108(8):1108–11. Available from: http://www.ncbi.nlm.nih.gov/pubmed/21791325 635

58. Hebert PC, Wells G, Blajchman M, Marshall J, Martin C, Pagliarello G, et al. A multicenter, 636

randomized, controlled clinical trial of transfusion requirements in critical care. N Engl J Med. 637

1999;340(6):409–17. 638

59. Walsh TS, Boyd J a, Watson D, Hope D, Lewis S, Krishan A, et al. Restrictive versus liberal 639

transfusion strategies for older mechanically ventilated critically ill patients: a randomized 640

pilot trial. Crit Care Med [Internet]. 2013 Oct [cited 2014 Oct 8];41(10):2354–63. Available 641

from: http://www.ncbi.nlm.nih.gov/pubmed/23939351 642

60. Carson JL, Terrin ML, Barton FB, Aaron R, Greenburg AG, Heck DA, et al. A pilot 643

randomized trial comparing symptomatic vs. hemoglobin-level-driven red blood cell 644

transfusions following hip fracture. Transfus Pract. 1998;38(June):522–9. 645

61. Fan Y-X, Liu F-F, Jia M, Yang J-J, Shen J-C, Zhu G-M, et al. Comparison of restrictive and 646

liberal transfusion strategy on postoperative delirium in aged patients following total hip 647

replacement: a preliminary study. Arch Gerontol Geriatr [Internet]. Elsevier Ireland Ltd; 2014 648

[cited 2015 Feb 6];59(1):181–5. Available from: 649


62. Hebert PC, Wells G, Marshall J, Martin C, Tweeddale M, Pagliarello G, et al. Transfusion 651

Requirements in Critical Care. JAMA. 1995;273:1439–44. 652

Page 23 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

24

63. Foss NB, Kristensen MT, Jensen PS, Palm H, Krasheninnikoff M, Kehlet H. The effects of 653

liberal versus restrictive transfusion thresholds on ambulation after hip fracture surgery. 654

Transfusion [Internet]. 2009 Feb [cited 2014 Oct 8];49(2):227–34. Available from: 655


64. Mazza BF, Freitas FGR, Barros MMO, Azevedo LCP, Machado FR. Blood transfusions in 657

septic shock: is 7.0 g/dL really the appropriate threshold? Rev Bras Ter intensiva [Internet]. 658

2015 [cited 2015 Aug 11];27(1):36–43. Available from: 659


type=abstract 661

65. Nielsen K, Johansson PI, Dahl B, Wagner M, Frausing B, Børglum J, et al. Perioperative 662

transfusion threshold and ambulation after hip revision surgery--a randomized trial. BMC 663

Anesthesiol [Internet]. 2014 Jan [cited 2015 Oct 14];14:89. Available from: 664


type=abstract 666

66. Parker MJ. Randomised trial of blood transfusion versus a restrictive transfusion policy after 667

hip fracture surgery. Injury [Internet]. Elsevier Ltd; 2013 Dec [cited 2015 Feb 6];44(12):1916–668

8. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23756263 669

67. Almeida J, Vincent J, Galas F, Almeida E, Fukushima J, EA O, et al. Transfusion 670

Requirements in Surgical Oncology Patients. Anesthesiology. 2015;122:29–38. 671

68. Holst LB, Haase N, Wetterslev J, Wernerman J, Guttormsen AB, Karlsson S, et al. Lower 672

versus Higher Hemoglobin Threshold for Transfusion in Septic Shock. N Engl J Med 673

[Internet]. 2014 Oct [cited 2014 Oct 1];371(15):1381–91. Available from: 674

http://www.nejm.org/doi/abs/10.1056/NEJMoa1406617 675

69. Gregersen M, Borris LC, Damsgaard EM. Postoperative blood transfusion strategy in frail, 676

anemic elderly patients with hip fracture: the TRIFE randomized controlled trial. Acta Orthop 677

[Internet]. 2015 Jun [cited 2015 Aug 25];86(3):363–72. Available from: 678


type=abstract 680

70. So-Osman C, Nelissen R, Te Slaa R, Coene L, Brand R, Brand a. A randomized comparison of 681

transfusion triggers in elective orthopaedic surgery using leucocyte-depleted red blood cells. 682

Vox Sang [Internet]. 2010 Jan [cited 2014 Oct 8];98(1):56–64. Available from: 683


71. Gregersen M, Damsgaard EM, Borris LC. Blood transfusion and risk of infection in frail 685

elderly after hip fracture surgery: the TRIFE randomized controlled trial. Eur J Orthop Surg 686

Traumatol [Internet]. Springer Paris; 2015 Aug [cited 2015 Aug 11];25(6):1031–8. Available 687

from: http://www.ncbi.nlm.nih.gov/pubmed/25690514 688

72. Deans K, Minneci P, Suffredini A, RL D, Hoffman W, Ciu X, et al. Randomization in clinical 689

trials of titrated therapies: Unintended consequences of using fixed treatment protocols. Crit 690

Care Med. 2007;35(6):1509–16. 691

73. Walsh TS, McClelland DB, Lee RJ, Garrioch M, Maciver CR, McArdle F, et al. Prevalence of 692

ischaemic heart disease at admission to intensive care and its influence on red cell transfusion 693

thresholds: multicentre Scottish Study. Br J Anaesth [Internet]. 2005 Apr [cited 2015 Jul 694


Page 24 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

25

74. Murphy GJ, Pike K, Rogers C a., Wordsworth S, Stokes E a., Angelini GD, et al. Liberal or 696

Restrictive Transfusion after Cardiac Surgery. N Engl J Med [Internet]. 2015 Mar 12 [cited 697

2015 Oct 1];372(11):997–1008. Available from: 698

http://www.nejm.org/doi/abs/10.1056/NEJMoa1403612 699

75. Fominskiy E, Putzu a, Monaco F, Scandroglio a M, Karaskov a, Galas FRBG, et al. Liberal 700

transfusion strategy improves survival in perioperative but not in critically ill patients. A meta-701

analysis of randomised trials. Br J Anaesth [Internet]. 2015 Oct [cited 2015 Oct 702


704

705

Page 25 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

26

Figure Legends: 706

Figure 1: PRISMA Flow Diagram 707

Figure 2: Forest plot of Risk Ratios for 30 day mortality with risk of bias assessment for each study. 708

1.1.1: all studies; 1.1.2: studies that randomised by cardiovascular disease. *Jairath: Additional risk of 709

bias assessed as to completeness of patients recruited into clusters. This was graded as low risk. 710

711

Figure 3: Forest plot with Risk Ratios for Adverse Cardiovascular Events: 712

A: Myocardial infarction, Acute Coronary Syndrome, Cardiac Arrest. Risk of bias assessment 713

included for each study. 714

B: Acute Pulmonary Oedema 715

Page 26 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

27

Table 1: Characteristics of included trials contributing to data-analysis. S: Single-centre, M: 716

Multi-centre trial. CV: Cardiovascular; MI Myocardial Infarction; GI: GastroIntestinal 717

Author Journal Clinical Setting (*leucodepleted)

Trigger Total n CV n (%) Primary endpoint

Almeida

2013 Brazil

(S)

Critical

Care Oncology

Restrictive

(R) 70g/l 101 22 (21.8)

Composite death or severe

complications Liberal (L)

90g/l 97 12 (12.4)

Bush 1997

USA (S)

American

Journal of

Surgery

Elective vascular

surgery

R: 90g/l 50 50 (100) Myocardial ischaemia,

myocardial infarction, death L: 100g/l 49 49 (100)

Carson 2011

USA/Canada

(M)

NEJM

Hip fracture

patients with

Cardiovascular

disease (CVD) or

risk factors for

CVD*

R: 80g/l or

symptoms

of anaemia

1009 1009

(100) 60 day mortality

Walk unaided

L: 10g/dl 1007 1007

(100)

Carson 2013

USA (M)

American

Heart

Journal

Symptomatic

Coronary Artery

Disease*

R: 8g/dl or

symptoms

of anaemia

55 55 (100)

Composite:

all cause mortality,

myocardial infarction or

unscheduled coronary

revascularization L: 10g/dl 55 55 (100)

Cooper 2011

USA (M)

American

Journal of

Cardiology

Acute MI*

R: hct

<24%, 24 24 (100)

Composite: In-hospital

death, recurrent MI, new or

worsening congestive heart

failure L: hct

<30% 21 21 (100)

Gregersen

2015

Denmark (S)

Acta

Orthop.

Frail elderly hip

fracture patients*

R: 97g/l 116 34 (29.3) Recovery from physical

disabilities L: 113g/l 111 25 (22.5)

Hebert

1998 Canada

(M)

NEJM Critical care R: 70g/l 418 160 (38.2)

30 day mortality L: 90g/l 420 197 (46.9)

Holst

2014

Scandinavia

(M)

NEJM Critical care*

R: 70g/l 502 75 (14.9)

90 day mortality L: 90g/l 496 66 (13.3)

Jairath

2015

UK (M)

Lancet Upper GI

haemorrhage

R: 80g/l 403 61 (15%) Feasibility

L: 100g/l 533 76 (14%)

Parker

2013 UK (S) Injury

Hip fracture

patients

R: definite

symptoms

of anaemia

100 50 (50.0)

30d mortality L: raise

Hb to at

least

10.0g/dl

100 37 (37.0)

Walsh

2013 UK

(M)

Critical

Care

Medicine

Critical care* R:70g/l 51 17 (33.3) Feasibility: Difference in

mean Hb during

intervention period. L: 90g/l, 49 15 (30.6)

718

Page 27 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

28

Table 2: Summary of findings for all trials including GRADE quality of evidence assessment. 719

RR – relative risk. 1. Not all participants/clinicians blinded; 2. definition varied between studies; 3. Not 720

all investigators blinded; 4. substantial heterogeneity; 5. low numbers 721

Quality assessment № of patients Effect Quality Importa

nce

№ of

studi

es

Study

design

Risk of

bias

Inconsiste

ncy

Indirectn

ess

Imprecis

ion

Other

considerati

ons

Restrict

ive

Liberal

Transfus

ion

Threshol

d

Relati

ve

(95%

CI)

Absol

ute

(95%

CI)

30 Day Mortality

11 randomised trials

serious

1

not serious not serious

not serious

none 144/1514 (9.5%)

126/1519 (8.3%)

RR

1.15

(0.93

to 1.43)

12

more

per

1000

(from

6

fewer

to 36 more)

⨁⨁⨁

◯

MODERATE 1

6.0% 9 more

per

1000

(from

4

fewer

to 26 more)

Cardiovascular events

8 randomised trials

very

serious

1 2 3

not serious not serious

not serious

none 59/1319 (4.5%)

32/1290 (2.5%)

RR

1.78

(1.18

to

2.70)

19

more

per

1000

(from

4 more

to 42

more)

⨁⨁◯

◯

LOW 1 2 3

1.0% 8 more

per

1000

(from

2 more

to 17

more)

Acute Pulmonary Oedema

4 randomised trials

very

serious

1 2 3

serious 4 not

serious

serious 5 none 24/309

(7.8%) 47/340

(13.8%) RR

0.58 (0.36

to

0.92)

58

fewer

per

1000

(from

11

fewer

to 88

fewer)

⨁◯◯

◯

VERY LOW 1 2 3 4 5

10.7% 45

fewer

per

1000

(from

9

fewer

to 68 fewer)

722

Page 28 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

29

723

Page 29 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

1

TRANSFUSION THRESHOLDS IN PATIENTS WITH CARDIOVASCULAR

DISEASE (NON CARDIAC SURGERY): SYSTEMATIC REVIEW PROTOCOL

Annemarie Docherty1,2 ([email protected])*, Rob O’Donnell

2

(ro’[email protected]), Susan Brunskill3 ([email protected]), Marialena Travella

3

([email protected]), Carolyn Doree3 ([email protected]),

Timothy Walsh1,2 ([email protected]), Simon Stanworth

3,4

([email protected]).

1. Centre for Inflammation Research, University of Edinburgh, Little France Crescent,

Edinburgh, EH16 4SA

2. Critical Care Department, Royal Infirmary Edinburgh, Little France Crescent,

Edinburgh, EH16 4SA

3. Systematic Review Initiative, NHS Blood and Transplant, John Radcliffe Hospital,

Headley Way, Oxford, OX3 9DU

4. Department of Haematology, John Radcliffe Hospital, Headley Way, Oxford, OX3

9DU

*Corresponding author

Page 30 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

2

ABSTRACT

Background

Patients who are acutely ill in hospital are frequently anaemic. Evidence exists from large randomised

controlled trials in critical care that a restrictive blood transfusion threshold of around 70g/l is both

safe, and associated with lower morbidity and mortality than a more liberal transfusion threshold.

However, in patients with cardiovascular disease, there is biological plausibility that anaemia could

compromise the myocardial oxygen supply/demand balance, especially when myocardial work is

increased, but coronary supply decreased (tachycardia; shock) because of the high coronary oxygen

extraction ratio. In addition to this, both of the large published RCTs report a point estimate that

favoured liberal practice for this sub-cohort.

Methods/Design

We will conduct a systematic review with meta-analysis, to compare restrictive versus liberal

transfusion thresholds in patients with cardiovascular disease. We will report our review according to

the recommendations provided by the PRISMA statement. We will include randomised controlled

trials, and the literature search will comprise CENTRAL, MEDLINE, Embase, CINAHL, PUBMED,

LILACS, Transfusion Evidence Library, and Web of Science. Study selection and data extraction will

be conducted by two independent investigators. Our primary outcome will be mortality at thirty days,

secondary outcomes will include other mortality, and measures of both cardiac and non-cardiac

morbidity. We will use the Cochrane risk of bias tool to assess risk of bias of included studies. We

expect that there will be limited quantitative data to include in any statistical analysis, so envisage that

the analysis will be a qualitative summary and interpretation of the data. However, if data allows, we

will undertake a quantitative analysis of the outcome data. We are expecting that there will be

substantial heterogeneity among included studies, and we will therefore use a random-effects model.

If there are a rare number of events we will use the Peto Odds method for pooling the results.

Discussion

Page 31 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

3

This review will combine the available evidence for transfusion thresholds in patients with

cardiovascular disease for the first time. We hope that this will provide stronger guidance for

transfusion in this patient group, as uncertainty defines optimal practice at present.

Systematic Review Protocol registration: PROSPERO: CRD42014014251

Keywords

Red cell transfusion, haemoglobin concentration, critical illness, cardiovascular disease,

randomised controlled trial, systematic review, meta-analysis

BACKGROUND

Description of the condition

Patients who are acutely ill in hospital are frequently anaemic, and have low haemoglobin

concentrations. When not as a result of active bleeding or frequent blood sampling, this is commonly

related to functional iron deficiency, reduced erythropoietin production, and infection (Vincent).

Anaemia is generally well tolerated by many hospital patients and therefore the benefits of any

treatments such as red cell transfusions need to be weighed against the risks of red blood cell

transfusion. Anaemia both decreases the oxygen content of the blood supplied to the myocardium, and

increases myocardial oxygen demand by requiring a higher cardiac output to maintain adequate

systemic oxygen delivery (Sabatine et al., 2005). The myocardium extracts around 75% of arterial

oxygen, and this cannot easily be increased in response to demand.

Patients with cardiovascular disease form a common sub-group of patients admitted to hospital with

critical illness. Around 110 000 patients are treated in Intensive Care Units (ICU) each year in the UK,

and the prevalence of known pre-existing coronary artery disease (CAD) may be as high as 29%;

moreover this proportion may rise as the average age of both the general population and patients

admitted to ICU increase. Patients with cardiovascular disease will have impaired compensatory

mechanisms to enable maximum oxygen delivery to the tissues in the event of anaemia. Global

Page 32 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

4

oxygen delivery is mainly a product of cardiac output, haemoglobin concentration, and oxygen

saturation. To compensate for a fall in haemoglobin, there is therefore an increase in cardiac output,

stressing the heart to increase heart rate and stroke volume. In acute illness, global oxygen demand is

increased, further stressing the heart, and in addition hypotension and tachycardia may reduce

coronary blood flow.

Anaemia has been associated with worse outcomes in patients who have cardiovascular disease

(Shulman, 1997). Zeidman found that anaemia was a significant risk factor in ischaemic heart disease

(Zeidman et al., 2004), correlating with advanced ischaemic heart disease (IHD), chronic heart failure,

rhythm disturbance and higher mortality rate, in comparison to non-anaemic patients. Sabatine

showed that anaemia (haemoglobin concentration <140g/l for ST elevation myocardial infarction (MI),

<110g/l for non-ST elevation MI) was an independent predictor of major adverse cardiovascular

events in patients across the spectrum of acute coronary syndrome (ACS) (Sabatine et al., 2005).

However, it does not necessarily follow that the correction of anaemia by red blood cell transfusion

will improve outcomes.

Description of the intervention

The main treatment option for rapidly raising the haemoglobin concentration in patients with anaemia

remains red cell transfusion. Risks of transfusion are well recognised (although perhaps underreported)

and include acute transfusion reactions, infection with blood-borne pathogens, and Transfusion

Related Acute Lung Injury, as well as immunosuppression (Gerber, 2008; SHOT Steering Group,

2013). Other treatment options for anaemia include erythropoietin, and oral or intravenous iron

therapy. These have been used in chronic anaemia, particularly in renal failure, for many years,

however there is little data within patients with cardiovascular disease (Gerber, 2008).

For patients who are not actively bleeding, trials have looked at randomising patients to either

“restrictive” triggers (patients are transfused only when their haemoglobin concentration falls to

around 70g/l) or “liberal” triggers (patients transfused at a higher haemoglobin concentration of

Page 33 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

5

around 100g/l). Much of the evidence underlying restrictive and liberal thresholds comes from trials

based in Critical Care. In 1999 the Transfusion Requirements in Critical Care (TRICC) Trial found a

similar mortality in patients transfused at a restrictive trigger of less than 70g/l compared with a

liberal trigger of less than 100g/L (Herbert, Wells, & Blajchman, 1999). This finding is supported by

another, more recent, large RBC transfusion trigger trial in critical care: the Transfusion Requirements

in Septic Shock (TRISS) trial (Holst, 2014). They enrolled 1000 cases of septic shock and found that

overall mortality was similar for restrictive (<70g/l) and liberal (<90g/l) transfusion triggers.

Additional randomised trials have been undertaken comparing restrictive vs liberal strategies of red

cell transfusion, and a meta-analysis by Carson et al reports no evidence of benefit with liberal

policies (JL Carson, Carless, & Herbert, 2012) Moreover, recent meta-analyses indicate suggestions

of harm (for example outcomes of infection) in patients transfused more liberally (Rohde et al., 2014).

Therefore, common standards are to treat anaemia in acute illness (excluding haemorrhage) based on

a restrictive transfusion policy.

However, there is uncertainty in the risk-benefit from RBC transfusion to correct moderate-severe

anaemia in patients with cardiovascular disease. A physiological rationale exists for avoiding anaemia

in patients with cardiovascular disease, and many physicians transfuse red cells to maintain higher

haemoglobin concentrations. In patients with chronic IHD, Herbert et al found a non-significant trend

favouring liberal transfusion (30 day mortality difference -4.9%, 95% CI -15.3%, 5.6%). Around 14%

of patients enrolled into TRISS had chronic cardiovascular disease at randomisation, and this group

had a non-significant increased risk of mortality with a restrictive trigger (RR 1.08, 95% CI 0.75,1.40).

Other studies have found an association between anaemia and mortality in critically ill patients with

IHD (JL Carson et al., 1996), and in the general surgical population, particularly in older patients (Wu

et al., 2007). Conversely, both FOCUS (Transfusion Trigger Trial for Functional Outcomes in

Cardiovascular Patients Undergoing Surgical Hip Fracture Repair) and TRACs (Transfusion

Requirements After Cardiac Surgery) found no difference between liberal and restrictive transfusion

groups, in 30 day mortality (Jeffrey Carson et al., 2011; Hajjar et al., 2010).

Page 34 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

6

Regarding patients with acute coronary syndrome, Carson et al found that there were fewer deaths

when a liberal trigger of 100g/l was chosen compared with a restrictive trigger of 80g/l (J. L. Carson

et al., 2013). This did not support the finding from the CRIT trial (Conservative vs liberal red cell

transfusion in acute myocardial infarction pilot trial), which suggested that a higher haematocrit

trigger of 30% vs 24% was associated with worse clinical outcomes (Cooper et al., 2011).

Current guidelines

The majority of current guidelines recommend a restrictive threshold blood transfusion. However, in

the case of patients with cardiovascular disease, the guidelines acknowledge the paucity of evidence,

and recommend a higher threshold. The British Committee for Standards in Haematology (BCSH)

assign Grade 2C to their recommendation that critically ill patients with chronic angina should have a

haemoglobin >=70g/l, but that transfusion to >=100g/l has uncertain benefit (BCSH, 2012). For

patients with acute coronary syndrome, they recommend maintaining a haemoglobin >=80g/l (Grade

2C). The American Association of Blood Banks (AABB) recommends transfusion for hospitalised

patients with pre-existing cardiovascular disease if symptomatic or at a haemoglobin of 80g/l or less

(Grade: weak recommendation; moderate-quality evidence). They were unable to recommend a

threshold for haemodynamically stable patients with acute coronary syndrome (Grade: uncertain

recommendation; very low-quality evidence) (Carson, 2012). The AAGBI state that a haemoglobin of

80-100g/l is safe for patients even with significant cardio-respiratory disease (AAGBI, 2008).British

Columbia Transfusion Medical Advisory Group states that in patients with indications of cardiac

disease, the available evidence suggests that, as a general guide, it may be safer to maintain the

hemoglobin above 90 g/L (BCTMAG, 2003).

How the intervention might work

The rationale for transfusing red blood cells in anaemic patients with cardiovascular disease is to

improve oxygen delivery to the tissues and to the myocardium itself, and reduce the compensatory

work done by the heart to increase cardiac output. In addition to this, acutely ill patients are frequently

hypotensive and tachycardic, which leads to an oxygen supply/demand imbalance (“Type 2 MI”).

Page 35 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

7

However, increasing arterial oxygen content by transfusion may not increase oxygen delivery to the

myocardium distal to an anatomic coronary stenosis.

It may be that although the allogeneic red cells increase the oxygen content of the blood, the effects of

storage on the red cells render them significantly less effective and potentially harmful. The

concentration of 2,3-DPG reduces rapidly, increasing the affinity of the red cells for oxygen, and thus

reducing offloading of oxygen at the tissues. There are well described morphological changes, and the

stiffened cells undergo haemolysis at a higher rate, leaching out free haemoglobin and cytokines into

the circulation, which then cause an inflammatory response and local vasoconstriction. Other possible

adverse effects include loss of red cell nitric oxide production, which is thought to induce local

vasodilatation to facilitate transition, pro-thrombotic effects from factors in the supernatant, and pro-

inflammatory effects from substances in the red cell supernatant.

OBJECTIVES

To assess the effect of a Liberal vs Restrictive Red cell Transfusion Policy in Patients with

Cardiovascular Disease.

METHODS

Criteria for considering studies in this review

Types of studies

We will include randomised controlled trials (RCT).

Types of participants

Trials that include adult patients with cardiovascular disease.

Cardiovascular disease is defined as:

Page 36 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

8

Known Coronary Artery Disease (CAD): Acute Coronary Syndrome

Chronic Ischaemic Heart Disease

other Cardiovascular Disease - Cerebrovascular Accident (CVA), Transient Ischaemic Attack

(TIA), Peripheral Vascular Disease (PVD).

Acute Coronary Syndrome is defined as:

ST elevation myocardial infarction

Non-ST elevation myocardial infarction

Unstable angina

Critical illness is defined as:

Any patient requiring Level 2 or 3 care, as defined by the Department of Health UK (2002).

This will include high risk elective surgical, emergency, and non-operative admissions.

We will exclude patients admitted to hospital for cardiac surgery. This is because they are a distinct

group of patients, whose cardiovascular risk has been significantly altered by their procedure. A

systematic review of transfusion triggers in this population has been performed recently (Curley,

Shehata, Mazer, 2014), which found that there was no increase in mortality or adverse events with a

restrictive trigger. However, Curley concluded that further trials are needed in this area.

Types of interventions

Studies evaluating the effectiveness of any policy involving the use of a trigger or transfusion

threshold based on haemoglobin concentration (including haematocrit) for guiding allogeneic blood

transfusion.

We will then attempt to extract relevant data for our group of interest (patients with cardiovascular

disease) from the studies that we find and combine data where appropriate.

If a distinct group of participants isn’t available for our SR, we will ensure that participants in groups

other than the groups of interest are less than 20% of the total.

Page 37 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

9

Types of outcome measures

Primary:

• Mortality: all cause mortality at 30 days from the day of transfusion , expressed as a risk ratio.

Secondary

• Mortality n(%): all cause mortality at 60 days, Critical Care and hospital mortality, other

mortality defined by authors.

• Morbidity med(IQR): measures of organ dysfunction (eg SOFA Sequential Organ Failure

Assessment, MODS Multiple Organ Dysfunction Score).

• Adverse events n(%): cardiac specific - non-fatal myocardial infarction as classified by the

Universal Definition of Myocardial Infarction (Thygesen, Alpert, Jaffe, & Simoons, 2012),

cardiac event, pulmonary oedema (if classified as cardiogenic); non-cardiac ischaemic events

- peripheral ischaemia, thrombotic events.

• Duration med(IQR): ICU/hospital stay, invasive ventilation, haemodynamic support, renal

support.

• Use of packed red blood cells n(%).

• adverse transfusion reactionsi n(%).

• incidence of in-hospital infections n(%).

Search methods for identification of studies

We anticipate few trials that have taken place in the specific population that this review is exploring:

transfusion triggers in patients with cardiovascular disease. We therefore propose a broad search

initially, to include all in-hospital trials of transfusion triggers. We shall not restrict our search by

language or publication status.

Page 38 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

10

Electronic searches

We will search the following databases:

Cochrane Central Register of Controlled Trials (CENTRAL, The Cochrane Library)

MEDLINE (OvidSP, 1946 to present)

Embase (OvidSP, 1974 to present)

CINAHL (EBSCOHost, 1982 to present)

PubMed (current e-publication only)

LILACS (1982 onwards)

Transfusion Evidence Library (www.transfusionevidencelibrary.com) (1980 to present)

Web of Science: Conference Proceedings Citation Index-Science (CPCI-S) (1990 to present)

We will also search the following databases for ongoing trials (all years):

ISRCTN Register

EU Clinical Trials Register (EUDRACT)

ClinicalTrials.gov

WHO International Clinical Trials Registry Platform (ICTRP)

The search strategies are presented in Appendix 1.

Searching other resources

We will contact experts in the field to identify information, and unpublished studies relevant to the

review. We will search the references of included trials and reviews for further studies. We will

contact the authors of relevant trials to request the raw data for analysis.

DATA COLLECTION AND ANALYSIS

Page 39 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

11

Selection of studies

Two authors will independently screen the titles, abstracts, or both, of the selected reviews and trials

that met the previously defined inclusion criteria. Identification will be discussed until consensus is

reached, in the event of disagreement a third author will adjudicate.

Data extraction and management

The authors will extract the data using a study specific data form (Table 2). We will collect the

following data for patients with cardiovascular disease:

Study characteristics: type of study, study setting and duration, number of participants, patient

demographics at baseline, type of intervention, and leucodepletion status.

Outcomes: number of patients receiving allogeneic blood, amount of allogeneic blood transfused,

mean haemoglobin in each arm of the trial. We will extract data for all cause mortality 30 days, and

also Critical Care mortality, and hospital mortality. We will record cardiac specific morbidity (non-

fatal myocardial infarction, cardiac event, cardiogenic pulmonary oedema, cerebrovascular event),

and general morbidity (thrombotic events, measure of organ dysfunction (eg SOFA, MODS), duration

of invasive ventilation, haemodynamic support, renal support, adverse transfusion reactions, in-

hospital infections). We will also record the main primary outcome of each trial (all patients). Where

necessary, we will contact the authors of the original studies for the raw data. We will enter the data

into the Cochrane Collaboration’s statistical software, Review Manager 2013, and data will be cross-

checked by both authors. We will synthesise the data, and where possible combine it quantitatively.

Assessment of risk of bias in selected studies

Risk of bias tools will be applied to randomised trials.

We will assess the risk of bias using the Cochrane Collaboration Handbook for Systematic Reviews of

Interventions (Higgins, Green, & (editors), 2011). We will assess and categorise the risk of bias into

high, low, or unclear, based on the information reported or retrieved from each trial, for each of the

Page 40 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

12

following domains: selection bias, performance bias, detection bias, attrition bias and reporting bias

(Table 3).

Analysis plan

We are expecting that there will be limited quantitative data to include in any statistical analysis, so

envisage that the analysis will be a qualitative summary and interpretation of the data.

However, if data allows, we will undertake a quantitative analysis of the outcome data. We will report

continuous variables as mean difference. Dichotomous variables will be reported as Relative Risk

ratios. We will use 95% confidence intervals throughout. Quantitative meta-analyses will be

undertaken where there are sufficient data. Meta-analyses will be performed using Review Manager

software. We are expecting that there will be substantial heterogeneity among included studies, and

we will therefore use a random-effects model. If there are a rare number of events we will use the

Peto Odds method for pooling the results.

Unit of analysis issues

The unit of analysis will be the patient, hence we do not expect any unit of analysis issues.

Dealing with missing data

For all outcomes we will carry out the analysis on an intention-to-treat basis. We will contact the

authors of the original papers in the event of missing data. In the event of missing data despite this, we

will conduct a sensitivity analysis to assess the impact on the review. If there is more than 20% of

data missing, we will use real data as this is likely to affect the results substantially.

Assessment of heterogeneity

We will examine heterogeneity using the I2 and the Chi2 statistic as recommended by Higgins

(Higgins et al., 2011). The I2 statistic describes the percentage of variation across the studies that is

due to heterogeneity rather than chance (0% is no observed heterogeneity, >50% is moderate

heterogeneity and > 80% is substantial heterogeneity.

Page 41 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

13

Assessment of reporting biases

We will assess publication bias using funnel plots, as long as there are at least 10 studies for meta-

analysis.

Subgroup analysis and investigation of heterogeneity

If there is significant heterogeneity, we will explore it using subgroup analyses and sensitivity

analyses. We will use the following subgroup analyses:

Acute Coronary Syndrome

Chronic Cardiovascular Disease

Critical Illness

Sensitivity analysis

We will perform a sensitivity analysis considering the quality of the allocation concealment for RCTs,

and carry out the meta-analysis, temporarily excluding the high risk trials

DISCUSSION

The need for this review

There is a need for a comprehensive review to address RBC transfusion practice for anaemic patients

with co-existing CVD. In particular, this review will summarise evidence for the high risk patient

population in whom physiological systems are already stressed, and oxygen demand is increased:

patients with critical illness, and patients with acute coronary syndrome. We will identify all

randomised trials recruiting patients with cardiovascular disease. The ongoing need for ICU care in

hospitals, and the expansion of the elderly population with highest risk for CVD mean questions about

optimal red cell practice is very relevant for clinicians. A Cochrane systematic review update on red

cell transfusion also specifically noted lack of clarity for high risk groups such as those with

CVD/CAD (Carson, 2012).

Page 42 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

14

Implications for practice

This review will combine the available evidence for transfusion thresholds in patients with

cardiovascular disease for the first time. We hope that this will provide stronger guidance for

transfusion in this patient group. It is possible that the review will highlight the paucity of evidence in

patients with cardiovascular disease, and the need for further trials.

Abbreviations

ICU: Intensive Care Unit; CAD: Coronary Artery Disease; IHD: Ischaemic Heart Disease; ACS:

Acute Coronary Syndrome

Competing Interests

Nil

Authors’ Contributions

AD carried out the conceptualisation, literature review, manuscript writing, and consultation with

other team members. RD assisted with the literature review and manuscript writing. SB was primarily

responsible for aspects of the analysis, writing and providing feedback/editing of the manuscript. CD

was primarily responsible for the literature search strategy. MT gave advice on the statistical analysis

plan. TW/SS were the supervising researchers on this project who were primarily responsible for

assisting with the conceptualisation, providing support and guidance with methodology, writing and

editing the manuscript. All authors read and approved the final manuscript.

Page 43 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

15

BIBLIOGRAPHY

AAGBI 2008, Association of Anaesthetists of Great Britain and Ireland. Blood transfusion and the

anaesthetist - red cell transfusion.

http://www.aagbi.org/publications/guidelines/docs/red˙cell˙08.pdf 2008; Vol. June:1–20.

BCTMAG 2003. Guidelines for Red Blood Cell Transfusion. British Columbia Transfusion Medicine

Advisory Group 2003; Vol. November:1–3

Carson, J. L., Brooks, M. M., Abbott, J. D., Chaitman, B., Kelsey, S. F., Triulzi, D. J., … Mark, A.

(2013). Liberal Versus Restrictive Transfusion Thresholds for Patients with Symptomatic

Coronary Artery Disease. American heart journal, 165(6), 964–971. doi:10.1016/j.ahj.2013.03.001.Liberal

Carson, Jeffrey, Terrin, M., Noveck, H., Sanders, D., Chaitman, B. R., Rhoads, G. G., … Hyg, M. S.

(2011). Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery. New England

Journal of Medicine, 365, 2453–2462.

Carson, JL, Carless, P., & Herbert, P. (2012). Transfusion thresholds and other strategies for guiding

allogeneic red blood cell transfusion (Cochrane Review). The Cochrane Library, (2). Carson, JL, Duff, A., Poses, R., JA, B., RK, S., Trout, R., … Strom, B. (1996). Effect of anaemia and

cardiovascular disease on surgical mortality and cardiovascular disease on surgical mortality and

morbidity. Lancet, 348, 1055–1060. Carson JL Grossman BJ Kleinman S Tinmouth AT Marques MB Fung MK Holcomb JB Illoh O

Kaplan LJ Katz LM Rao SV Roback JD Shander A Tobian AAR Weinstein R McLaughlin LGS

Djulbegovic B. 2012. Red blood cell transfusion: a clinical practice guideline from the AABB. Annals of Internal Medicine.157(1):49-58

Cooper, H. a, Rao, S. V, Greenberg, M. D., Rumsey, M. P., McKenzie, M., Alcorn, K. W., & Panza, J.

a. (2011). Conservative versus liberal red cell transfusion in acute myocardial infarction (the

CRIT Randomized Pilot Study). The American journal of cardiology, 108(8), 1108–11.

doi:10.1016/j.amjcard.2011.06.014

Curley, G., Shehata, N., Mazer, C., Hare, G., & Friedrich, J. (2014). Transfusion Triggers for Guiding

RBC Transfusion for Cardiovascular Surgery: A Systematic Review and Meta-Analysis. Critical

care medicine.

Gerber, D. R. (2008). Transfusion of packed red blood cells in patients with ischemic heart disease. Critical care medicine, 36(4), 1068–74. doi:10.1097/CCM.0b013e318169251f

Hajjar, L., Vincent, J., Galas, F., Nakamura, R., Silva, C., Santos, M., … Auler Jr, J. (2010).

Transfusion requirements after cardiac surgery: the TRACS randomized controlled trial. JAMA :

the journal of the American Medical Association, 304, 1559–1567.

Herbert, P., Wells, G., & Blajchman, M. (1999). A Multicenter, Randomized, Controlled Clinical

Trial of Transfusion Requirements in Critical Care. New England Journal of Medicine, 340(6),

409–17. doi:10.1056/NEJM199904013401322 Higgins, J., Green, S., & (editors). (2011). Cochrane Handbook for Systematic Reviews of

Interventions Version 5.1.0 (updated March 2011). The Cochrane Collaboration 2011.

Holst, L., Haase, N., Wetterslev, J., …Perner, A. (2014) Lower versus Higher Hemoglobin Threshold For Transfusion in Septic Shock. N Engl J Med 2014; 371:1381-1391

Retter, A., Wyncoll, D., Pearse, R., Carson, D., McKechnie, S., Stanworth, S., Allard, S., Thomas, D.,

Walsh, T. and British Committee for Standards in Haematology (2013), Guidelines on the management of anaemia and red cell transfusion in adult critically ill patients. British Journal of

Haematology, 160: 445–464. doi: 10.1111/bjh.12143

Rohde, J. M., Dimcheff, D. E., Blumberg, N., Saint, S., Langa, K. M., Kuhn, L., … Rogers, M. a M.

(2014). Health care-associated infection after red blood cell transfusion: a systematic review and

meta-analysis. JAMA : the journal of the American Medical Association, 311(13), 1317–26.

doi:10.1001/jama.2014.2726

Sabatine, M. S., Morrow, D. a, Giugliano, R. P., Burton, P. B. J., Murphy, S. a, McCabe, C. H., …

Braunwald, E. (2005). Association of hemoglobin levels with clinical outcomes in acute

coronary syndromes. Circulation, 111(16), 2042–9. doi:10.1161/01.CIR.0000162477.70955.5F

Page 44 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

16

SHOT Steering Group. (2013). Annual SHOT report 2013. Manchester. Retrieved from

www.shotuk.org/wp-content/uploads/SHOT-Summary-2013-Final.pdf

Shulman, L., Braunwald, E., & Rosenthal, D. (1997). Hematological-oncological disorders and heart

disease. In: Braunwald E/ Heart Disease: A textbook of Cardiovascular Medicine. (5th Editio.,

pp. 1786–1808). Philadelphia. Thygesen, K., Alpert, J., Jaffe, A., & Simoons, M. (2012). Third universal definition of myocardial

infarction. JACC. Cardiovascular interventions, 60(16), 1581–98.

Wu, W., Schifftner, T., Henderson, W., Easton, C., Poses, R., Uttley, G., … Friedmann, P. (2007). Preoperative hematocrit levels and post-operative outcomes in older patients undergoing

noncardiac surgery. JAMA, 297, 2481–2488.

Zeidman, A., Fradin, Z., Blecher, A., Oster, H., Avrahami, Y., & Mittelman, M. (2004). Anemia as a

Risk Factor for Ischemic Heart Disease. Israel Medical Association Journal, 6(January), 16–18.

Page 45 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

17

APPENDIX 1

RBC TRANSFUSION TRIGGERS FOR IHD

– SEARCH NARRATIVE OCTOBER 2014

The following databases were searched for randomised controlled trials on 9.10.14. This was a revised update

search of the search conducted for the SRI RBC Overview (dated Sept 2009). The new strategies were run in all

databases (with date restriction in CENTRAL), although the following search method was used in MEDLINE

and Embase: i) the original search strategies + original RCT filters were re-run up until the end of 2008; ii) the

new strategies + new RCT filters were run for all years; iii) the results of i) were then removed from the new

search results. Once all the search results had been downloaded into bibliographic software, all previously

screened references from the RBC Overview, along with duplicates, were then removed.

The following databases were searched:

CENTRAL (The Cochrane Library Issue 8, 2014): publication years from 2009-2014

MEDLINE (1946 onwards)

Embase (1974 onwards)

CINAHL (1937 onwards)

PUBMED (epublications only)

LILACS (2009-2014)

TRANSFUSION EVIDENCE LIBRARY (1980 onwards)

Web of Science (Conference Proceedings Citation Index- Science (CPCI-S) -1990 to present)

Ongoing Studies:

ClinicalTrials.gov

WHO International Clinical Trials Registry Platform (ICTRP)

ISRCTN Register

EU Clinical Trials Register (https://www.clinicaltrialsregister.eu/ctr-search)

Hong Kong Clinical Trials Registry

The search retrieved 7,270 results, which were reduced to 5,465 results once duplicates were removed, plus 74

ongoing trials. Previously screened references were then removed which reduced the results to 2,900 references,

plus the 74 ongoing trials. These references were sent in screening format to the lead author, Annemarie

Docherty, on 15.10.14.

SEARCH STRATEGIES

CENTRAL

#1 MeSH descriptor: [Erythrocyte Transfusion] explode all trees

#2 MeSH descriptor: [Blood Transfusion] this term only

#3 ((blood or erythrocyte* or "red cell*" or "red blood cell*" or RBC*) near/5 (transfus* or unit* or infus*

or therap*))

#4 ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) near/5 (use* or

usage* or utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or

management or practic* or indicat* or criteri* or standard* or program*)):ab

#5 ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or

requir* or need* or administ* or replac* or support* or strateg* or management or practic* or indicat*

or criteri* or standard* or program*)):ti

#6 (leukodeplet* or leukoreduc* or leucodeplet* or leucoreduc* or leukofiltrat* or leucofiltrat*):ti

#7 ("allogeneic blood" or (unit* near/2 blood) or "allogenic blood" or (blood near/2 exposure) or "donor

blood" or "blood product*" or "blood component*" or "blood support")

#8 (hemotransfus* or haemotransfus* or hypertransfus* or hemotherap* or haemotherap*)

#9 (red cell* or erythrocyte* or blood or RBC*) and transfus*:ti

#10 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9

#11 MeSH descriptor: [Blood Component Transfusion] this term only

#12 MeSH descriptor: [Exchange Transfusion, Whole Blood] explode all trees

#13 MeSH descriptor: [Plasma Exchange] explode all trees

#14 MeSH descriptor: [Platelet Transfusion] explode all trees

#15 MeSH descriptor: [Leukocyte Transfusion] explode all trees

#16 #12 or #13 or #14 or #15

#17 #11 not #16

Page 46 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

18

#18 ("red cell*" or "red blood cell*" or erythrocyte* or RBC*)

#19 MeSH descriptor: [Erythrocytes] this term only

#20 #18 or #19

#21 #17 and #20

#22 #10 or #21

#23 MeSH descriptor: [Hematocrit] this term only

#24 ((h?emoglobin or h?emocrit* or HB or HCT) near/5 (level* or concentration* or target* or maintain*

or rais* or higher or lower or greater or above or below or equal or transfus*))

#25 #23 or #24

#26 #22 and #25

#27 ((transfus* or "red cell*" or "red blood cell*" or RBC*) near/10 (trigger* or thresh?old* or target* or

restrict* or liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or

policies or practic* or standard*))

#28 ((transfus* or "red cell*" or "red blood cell*" or RBC* or h?ematocrit*) and (level* or critical* or

intensive* or h?emorrhag* or bleed*) or hypertransfus*):ti

#29 #26 or #27 or #28 Publication Year from 2008 to 2014

MEDLINE (OVID)

1. randomized controlled trial.pt.

2. controlled clinical trial.pt.

3. randomi*.tw.

4. placebo.ab.

5. clinical trials as topic.sh.

6. randomly.ab.

7. groups.ab.

8. trial.ti.

9. or/1-8

10. exp animals/ not humans/

11. 9 not 10

12. BLOOD TRANSFUSION/

13. ERYTHROCYTE TRANSFUSION/

14. ((blood or erythrocyte* or red cell* or red blood cell* or RBC*) adj5 (transfus* or infus* or unit* or

therap*)).ti,ab.

15. ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) adj5 (use* or usage* or

utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or management or practic*

or indicat* or criteri* or standard* or program*)).ab.

16. ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or requir*

or need* or administ* or replac* or support* or strateg* or management or practic* or indicat* or criteri* or

standard* or program*)).ti.

17. (allogeneic blood or (unit* adj2 blood) or allogenic blood or (blood adj2 exposure) or donor blood or blood

product* or blood component* or blood support).ti,ab.

18. (hemotransfus* or haemotransfus* or hypertransfus* or hemotherap* or haemotherap*).tw.

19. (red cell* or erythrocyte* or blood or RBC*).tw. and transfus*.ti.

20. or/12-19

21. BLOOD COMPONENT TRANSFUSION/

22. EXCHANGE TRANSFUSION, WHOLE BLOOD/ or PLASMA EXCHANGE/ or PLATELET

TRANSFUSION/ or exp LEUKOCYTE TRANSFUSION/

23. 21 not 22

24. ERYTHROCYTES/

25. (red cell* or red blood cell* or erythrocyte* or RBC*).tw.

26. 24 or 25

27. 23 and 26

28. 20 or 27

29. *HEMATOCRIT/

Page 47 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

19

30. ((h?emoglobin or h?ematocrit* or HB or HCT) adj5 (level* or concentration* or target* or maintain* or

rais* or higher or lower or greater or above or below or equal or transfus*)).tw.

31. 29 or 30

32. 28 and 31

33. ((transfus* or red cell* or red blood cell* or RBC*) adj10 (trigger* or thresh?old* or target* or restrict* or

liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or policies or practice*

or standard*)).tw.

34. (((transfus* or red cell* or red blood cell* or RBC* or h?ematocrit*) and (level* or critical* or intensive* or

h?emorrhag* or bleed*)) or hypertransfus*).ti.

35. 32 or 33 or 34

36. 11 and 35

EMBASE (OVID)

1. Randomized Controlled Trial/

2. Randomization/

3. Single Blind Procedure/

4. Double Blind Procedure/

5. Crossover Procedure/

6. Placebo/

7. exp Clinical Trial/

8. Prospective Study/

9. (randomi* or double-blind* or single-blind* or RCT*).tw.

10. (random* adj2 (allocat* or assign* or divid* or receiv*)).tw.

11. (crossover* or cross over* or cross-over* or placebo*).tw.

12. ((treble or triple) adj blind*).tw.

13. or/1-12

14. Case Study/

15. case report*.tw.

16. (note or editorial).pt.

17. or/14-16

18. 13 not 17

19. (animal* or cat or cats or dog or dogs or pig or pigs or sheep or rabbit* or mouse or mice or rat or rats or

feline or canine or porcine or ovine or murine).ti.

20. 18 not 19

21. limit 20 to embase

22. BLOOD TRANSFUSION/

23. ERYTHROCYTE TRANSFUSION/

24. ((blood or erythrocyte* or red cell* or red blood cell* or RBC*) adj5 (transfus* or infus* or unit* or

therap*)).ti,ab.

25. ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) adj5 (use* or usage* or

utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or management or practic*

or indicat* or criteri* or standard* or program*)).ab.

26. ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or requir*

or need* or administ* or replac* or support* or strateg* or management or practic* or indicat* or criteri* or

standard* or program*)).ti.

27. (allogeneic blood or (unit* adj2 blood) or allogenic blood or (blood adj2 exposure) or donor blood or blood

product* or blood component* or blood support).ti,ab.

28. (hemotransfus* or haemotransfus* or hypertransfus* or hemotherap* or haemotherap*).tw.

29. (red cell* or erythrocyte* or blood or RBC*).tw. and transfus*.ti.

30. 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29

31. BLOOD COMPONENT THERAPY/

Page 48 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

20

32. GRANULOCYTE TRANSFUSION/ or LEUKOCYTE TRANSFUSION/ or LYMPHOCYTE

TRANSFUSION/ or PLASMA TRANSFUSION/ or THROMBOCYTE TRANSFUSION/

33. 31 not 32

34. ERYTHROCYTE/

35. (red cell* or red blood cell* or erythrocyte* or RBC*).tw.

36. 34 or 35

37. 33 and 36

38. 30 or 37

39. *HEMATOCRIT/

40. ((h?emoglobin or h?ematocrit* or HB or HCT) adj5 (level* or concentration* or target* or maintain* or

rais* or higher or lower or greater or above or below or equal or transfus*)).tw.

41. 39 or 40

42. 38 and 41

43. ((transfus* or red cell* or red blood cell* or RBC*) adj10 (trigger* or thresh?old* or target* or restrict* or

liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or policies or practice*

or standard*)).tw.

44. (((transfus* or red cell* or red blood cell* or RBC* or h?ematocrit*) and (level* or critical* or intensive* or

h?emorrhag* or bleed*)) or hypertransfus*).ti.

45. 42 or 43 or 44

46. 21 and 45

CINAHL (EBSCOHost)

S1 (MH "Blood Transfusion")

S2 (MH "Erythrocyte Transfusion")

S3 (erythrocyte* or red cell* or red blood cell* or blood or RBC*) N5 (transfus* or infus* or therap* or

unit*)

S4 AB ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) N5 (use* or

usage* or utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or

management or practic* or indicat* or criteri* or standard* or program*))

S5 TI ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or

requir* or need* or administ* or replac* or support* or strateg* or management or practic* or indicat*

or criteri* or standard* or program*))

S6 hemotransfus* or haemotransfus* or hemotherap* or haemotherap* or hypertransfus*

S7 TX (red cell* or erythrocyte* or blood or RBC*) and TI (transfus*)

S8 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7

S9 (MM "Hematocrit")

S10 TX ((hemoglobin or haemoglobin or hematocrit* or haematocrit* or HB or HCT) N5 (level* or

concentration* or target* or maintain* or rais* or higher or lower or greater or above or below or equal

or transfus*))

S11 S9 OR S10

S12 S8 AND S11

S13 TI ((transfus* or red cell* or red blood cell* or RBC* or haematocrit* or hematocrit*) and (level*

or critical* or intensive* or haemorrhag* or hemorrhag* or bleed*))

S14 TX ((transfus* or "red cell*" or "red blood cell*" or RBC*) N10 (trigger* or threshold* or target* or

restrict* or liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or

policies or practice* or standard*))

S15 S12 OR S13 OR S14

S16 (MH Clinical Trials+)

S17 PT Clinical Trial

S18 TI ((controlled trial*) or (clinical trial*)) OR AB ((controlled trial*) or (clinical trial*))

S19 TI ((singl* blind*) OR (doubl* blind*) OR (trebl* blind*) OR (tripl* blind*) OR (singl* mask*) OR

(doubl* mask*) OR (tripl* mask*)) OR AB ((singl* blind*) OR (doubl* blind*) OR (trebl* blind*) OR

(tripl* blind*) OR (singl* mask*) OR (doubl* mask*) OR (tripl* mask*))

S20 TI randomi* OR AB randomi*

S21 MH RANDOM ASSIGNMENT

S22 TI ((phase three) or (phase III) or (phase three)) or AB ((phase three) or (phase III) or (phase three))

S23 (TI (random* N2 (assign* or allocat*)) ) OR (AB (random* N2 (assign* or allocat*)) )

Page 49 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

21

S24 MH PLACEBOS

S25 S16 OR S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24

S26 S12 AND S25

PUBMED (epublications only)

#1 ("erythrocyte transfusion*" OR "red cell transfusion" OR "red cells" OR "red blood cells" OR "red blood

cell transfusion" OR "blood transfusion*" OR RBCs OR "RBC transfusion*" OR hemotransfus* OR

haemotransfus* OR hemotherap* OR haemotherap* OR hypertransfus*)

#2 ((hemoglobin OR haemoglobin OR haematocrit* OR hematocrit* OR HB OR HCT) AND (level* or

concentration* OR target* OR maintain* OR rais* OR higher OR lower OR greater OR above or below OR

equal OR transfus*))

#3 (random* OR blind* OR control group* OR placebo OR controlled trial OR controlled study OR trials OR

systematic review OR meta-analysis OR metaanalysis OR literature OR medline OR cochrane OR embase)

AND ((publisher[sb] OR inprocess[sb]) NOT pubstatusnihms)

#4 #1 AND #2 AND #3

LILACS

(( tw:(transfused OR transfusing OR transfused OR hypertransfusion OR haemoglobin OR hemoglobin OR

haematocrit OR hematocrit) AND (instance:"regional") AND ( db:("LILACS") AND type

of_study:("clinical_trials"))) AND (instance:"regional") AND ( year_cluster:("2009" OR "2010" OR "2011" OR

"2012" OR "2013" OR "2014"))

TRANSFUSION EVIDENCE LIBRARY Subject Area: Blood Components/Red Cells

OR

All fields: trigger OR triggers OR threshold OR thresholds OR haemoglobin OR haemoglobin OR haematocrit

OR hematocrit OR Hb OR HCT

WEB OF SCIENCE – CPCIS database

#1 TS=(("erythrocyte transfusion*" OR "red cell transfusion" OR "red cells" OR "red blood cells" OR "red

blood cell transfusion" OR "blood transfusion*" OR RBCs OR "RBC transfusion*" OR hemotransfus* or

haemotransfus* or hemotherap* or haemotherap* or hypertransfus*)

#2 TS=((hemoglobin or haemoglobin or hematocrit* or haematocrit* or HB or HCT) NEAR/1 (level* or

concentration* or target* or maintain* or rais* or higher or lower or greater or above or below or equal or

transfus*))

#3 #1 AND #2

#4 TS=((transfus* or "red cell*" or "red blood cell*" or RBC*) NEAR/5 (trigger* or threshold* or target* or

restrict* or liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or policies

or practice* or standard*))

#5 #3 OR #4

#6 TS=(randomi* OR randomly OR "random assignment" OR "random allocation" OR blind* OR "control

group*" OR controlled trial OR "controlled study")

#7 #5 AND #6

Ongoing Trials:

ClinicalTrials.gov & WHO ICTRP

Search Terms/Title: randomized OR randomised OR randomly

Intervention: red cell transfusion OR RBC transfusion OR blood transfusion

Title: trigger OR threshold OR target OR restrictive OR liberal OR aggressive OR conservative OR prophylactic

OR limit OR protocol OR policy OR policies OR practice OR standard OR hemoglobin OR hematocrit

ISRCTN & EUCTR [terms searched in combination and individually]

(red cell transfusion OR red blood cell transfusion OR RBC transfusion OR blood transfusion) AND (trigger

OR threshold OR target OR restrictive OR liberal OR aggressive OR conservative OR prophylactic OR limit

OR protocol OR policy OR policies OR practice OR standard OR hemoglobin OR hematocrit OR haemoglobin

OR haematocrit)

Hong Kong Clinical Trials Registry [terms searched individually]

Disease Group: Blood and blood-forming organs OR Circulatory System

Title: randomized OR randomised OR transfusion

Page 50 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

22

Table 1: Levels of Critical Care for Adult Patients. Adapted from Department of Health, 2002. Level 3 is comparable to

Intensive Care, and Level 2 is comparable to High Dependency.

Level of Care Definition

0 Requires hospitalisation - needs can be met through normal ward care

1 Patients at risk of their condition deteriorating

Higher levels of care requiring critical care outreach support

2 Patients requiring single organ system monitoring and support

Pre-operative optimisation or extended postoperative care

Higher levels of monitoring and observation

Major uncorrected physiological abnormality

3 Patients requiring advanced respiratory support alone

Basic respiratory support together with support of at least two organ systems

All complex patients requiring support for multi-organ failure

Page 51 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

23

Table 2: Data extraction fom Trial data

Baseline characteristics Intervention A (lower

trigger)

Intervention B (higher trigger)

Type of study

Study setting and duration (eg ICU alone, hospital) Eg Critical Care, Acute coronary syndrome

Leucodepleted (Y/N)

Number of participants

Patient demographics at baseline (age, sex, medical/surgical

condition)

Type of intervention (trigger/target)

Outcomes

Blood Transfusion Intervention A Intervention B

patients receiving

allogeneic blood

No. analysed

No. included in outcome

No. events

Reasons timepoints

Number of units of blood

each patient received

No. analysed


No. events Reasons

timepoints

Mean Hb (variance)

Mortality Intervention A Intervention B

All cause mortality 30

days

No. analysed


No. events

Reasons

timepoints

Critical Care mortality No. analysed


No. events Reasons

timepoints

Hospital mortality No. analysed


No. events

Reasons

timepoints

Morbidity Intervention A Intervention B

Cardiac specific No. analysed


No. events

Reasons

timepoints

Non-fatal myocardial

infarction

No. analysed


No. events

Reasons

timepoints

Cardiac event No. analysed

No. included in outcome No. events

Reasons

timepoints

Cardiogenic pulmonary oedema

No. analysed No. included in outcome

No. events

Reasons timepoints

Cerebrovascular event No analysed

No included in outcome

No events Reasons

timepoints

Page 52 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

24

Thrombotic events No analysed

No included in outcome

No events Reasons

Timepoints

General Intervention A Intervention B

Measure of organ

dysfunction (eg SOFA,

MODS)

No. analysed


No. events Reasons

timepoints

Duration of invasive

ventilation

No. analysed


No. events

Reasons

timepoints

Haemodynamic support No. analysed


No. events Reasons

timepoints

Renal support No. analysed


No. events

Reasons

timepoints

Adverse transfusion

reactions

No. analysed


No. events

Reasons

timepoints

In-hospital infections No. analysed


No. events Reasons

timepoints

Other

Main outcome of trial (all patients)

Primary outcome No. analysed

No. events (%)

Page 53 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

25

Table 3: Assessment of risk of bias, adapted from the Cochrane Collaboration’s tool for assessing bias

Domain Support for judgement Author Assessment of Bias

No Yes Unclear

Selection bias

Random sequence

generation

Method of generation

Allocation concealment

Performance bias

Blinding participants study

personnel

Outcome assessors

Detection bias

Blinding of outcome

assessment

Attrition bias

Incomplete outcome data

Reporting bias

Selective reporting

Other bias

Other bias

i Adverse transfusion reactions as defined by SHOT: IBCT=incorrect blood component transfused;

ADU=avoidable, delayed or undertransfusion; Anti-D=errors with anti-D immunoglobulin administration;

ATR=acute transfusion reactions; HTR=haemolytic transfusion reactions; TRALI=transfusion-related acute lung

injury; TACO=transfusion-associated circulatory overload; TAD=transfusion-associated dyspnoea;

UCT=unclassifiable complications of transfusion; PTP=post-transfusion purpura ; Others include handling and

storage errors, alloimmunisations and cell salvage (SHOT Steering Group, 2013)

Page 54 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyThe impact of restrictive versus liberal transfusion strategies on patient outcomes in patients with

cardiovascular disease excluding those undergoing cardiac surgery: A Systematic Review and

Meta-analysis

Forest Plot data

Figure 2(A):Forest plot of Risk Ratios for 30 day mortality: all studies

Heterogeneity: Tau=0.03; Chi2=11.58, df=10 (P=0.31); I

2=14%

Test for overall effect Z=1.04 (P=0.30)

Study/Subgroup Restrictive Liberal Risk Ratio

MH RE

95%CI

Events Total Events Total Weight (%)

Almeida 2015 7 22 0 12 0.9 8.48 0.53 to

136.76

Bush 1997 4 49 4 50 3.8 1.02 0.27 to

3.85

Carson 2011 43 1008 52 995 27.7 0.82 0.55 to

1.21

Carson 2013 7 55 1 55 1.6 7.00 0.89 to

55.01

Cooper 2011 2 24 1 21 1.3 1.75 0.17 to

17.95

Gregersen 2015 6 34 3 25 4.0 1.47 0.41 to

5.32

Hebert 1999 29 111 31 146 23.9 1.23 0.79 to

1.91

Holst 2014 33 75 24 66 26.5 1.21 0.80 to

1.82

Jairath 2015 6 49 2 67 2.8 4.10 0.86 to

19.47

Parker 2013 4 70 4 67 3.7 0.96 0.25 to

3.67

Walsh 2013 3 17 4 15 3.8 0.66 0.18 to

1.50

Total 144 1514 126 1519 100 1.15 0.88 to

1.50

Page 55 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyFigure 2(B): Forest plot of Risk Ratios for 30 day mortality: studies that randomised by

cardiovascular disease.

Heterogeneity: Tau=0.06; Chi2=4.67, df=4 (P=0.32); I

2=14%

Test for overall effect: Z=0.17 (p=0.87)

Study/Subgroup Restrictive Liberal Risk

Ratio

MH RE

95%CI

Events Total Events Total Weight

(%)

Bush 1997 4 49 4 50 3.8 1.02 0.27 to

3.85

Carson 2011 43 1008 52 995 27.7 0.82 0.55 to

1.21

Carson 2013 7 55 1 55 1.6 7.00 0.89 to

55.01

Cooper 2011 2 24 1 21 1.3 1.75 0.17 to

17.95

Walsh 2013 3 17 4 15 3.8 0.66 0.18 to

1.50

Total 59 1153 62 1136 100 0.96 0.58 to

1.59

Page 56 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyFigure 3(A): Forest plot with Risk Ratios for Adverse Cardiovascular Events:

Myocardial infarction, Acute Coronary Syndrome, Cardiac Arrest.

Heterogeneity: Tau=0.00, Chi2=0.14, df=3 (P=0.99); I2=0%


Study/Subgroup Restrictive Liberal Risk Ratio

MH RE

95%CI


(%)

Almeida 2015 0 22 0 12 Not

estimable

Bush 1997 2 49 1 59 3.4 2.04 0.19 to

21.79

Carson 2011 38 1008 23 1005 72.5 1.65 0.99 to

2.74

Carson 2013 11 54 6 54 22.3 1.83 0.73 to

4.60

Cooper 2011 1 24 0 21 1.9 2.64 0.11 to

61.54

Holst 2014 6 75 2 66 0.0 2.64 0.55 to

12.64

Parker 2013 0 70 0 67 Not

estimable

Walsh 2013 1 17 0 15 0.0 2.67 0.12 to

60.93

Total 52 1135 30 1130 100 1.71 1.11 to

2.65

Page 57 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyFigure 3(B): Forest plot with Risk Ratios for Adverse Cardiovascular Events: Acute Pulmonary

Oedema



Study/Subgroup Restrictive Liberal Risk

Ratio

MH RE

95%CI


(%)

Carson 2013 7 55 2 55 23.0 3.50 0.76 to

16.11

Cooper 2011 2 24 8 21 24.3 0.22 0.05 to

0.92

Hebert 1999 14 160 35 197 39.1 0.49 0.27 to

0.88

Parker 2013 1 70 2 67 13.6 0.48 0.04 to

5.16

Total 309 340 100 0.63 0.22 to

1.81

Page 58 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyFigure E1: Forest plot with mean difference for hospital length of stay.


Test for overall effect: Z=1.09 (p=0.28)

Study/Subgroup Restrictive Liberal Mean

Difference

IV RE

95%CI

Mean SD Total Mean SD Total Weight

(%)

Bush 1997 11 9 49 10 6 50 19.7 1.00 -2.02

to

4.02

Carson 2011 4 3.9 610 3.7 3.4 610 30.4 0.30 -0.11

to

0.71

Cooper 2011 10.4 7.2 24 4.3 3.3 21 18.8 6.10 2.89

to

9.31

Hebert 1999 28.76 19.5 111 30.6 18.8 146 12.9 -1.84 -6.58

to

2.90

Walsh 2013 53.3 40.1 17 36.3 28.3 15 0.9 17.00 -6.84

to

40.84

Total 845 867 100 1.24 -1.00

to

3.48

Page 59 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyRisk of bias table for forest plots:

Study A B C D E F G H I

Almeida Low Low Low Low Low Low Low Low Unclear

Bush Unclear Unclear Unclear Unclear Low Unclear Unclear Unclear Low

Cars 11 Low Low High Low Low Low Low Low Low

Cars 13 Low Low Unclear Unclear Low Low Low Low Low

Cooper Unclear Low High High Low Low Low High Low

Greg Unclear Low High Low Low Low Unclear

Hebert Low Low High Unclear Low Low Unclear

Holst Low Low High Low Low Low Low Low Low

Jairath Low High High Unclear High Unclear Unclear High High

Parker Unclear Low Unclear Unclear Low Low Unclear High High

Walsh Low Low High Low Low Low Unclear High Low

Risk of bias legend

A: Random sequence generation (selection bias)

B: Allocation concealment (selection bias)

C: Blinding of participants and personnel (performance bias)

D: Blinding of outcome assessment (detection bias)

E: Incomplete outcome data (attrition bias)

F: Selective reporting (reporting bias)

G: Other bias

H: Assessment of Cardiovascular Event

I: Definition of Cardiovascular Event

Page 60 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

1

Online Data Supplement 1

Table E1: Authors’ definition of Cardiovascular Disease 2 Author Definition of Cardiovascular disease

Almeida 2013 Arterial coronary artery disease, congestive heart failure, cerebrovascular disease, peripheral

arterial disease

Blair 1986 Not discussed

Bush 1997 Patients undergoing elective aortic or infrainguinal arterial reconstruction

Carson 2011 history of ischaemic heart disease, electrocardiographic evidence of previous myocardial

infarction, a history or presence of congestive heart failure or peripheral vascular disease, or a

history of stroke or transient ischaemic attack

Carson 2013 ST segment elevation myocardial infarction,

Non ST segment elevation myocardial infarction,

Unstable angina (symptoms of cardiac ischemia at rest with at least one episode lasting 10

minutes AND ST-segment depression of 0.01 mV or more or transient [b30-minute] ST-

segment elevation of 0.1 mV or more in two or more contiguous leads), OR prior documented

coronary artery disease (myocardial infarction, percutaneous cardiac intervention, coronary

artery bypass graft surgery), or age N55 with diabetes mellitus or peripheral arterial disease and

no biomarker elevation

Stable coronary artery disease undergoing a cardiac catheterization (presence of coronary artery

disease (one cardiac artery with at least 70% obstruction by visual inspection based on cardiac catheterization or undergoing a percutaneous cardiac intervention, )during index admission

Cooper 2011 AMI was defined as ischaemic-type chest discomfort lasting 30 minutes and associated with a

creatine kinase-MB (CK-MB) or cardiac troponin level above the upper limit of normal

(determined locally).

Gregersen 2015 Charlson Comorbidity Index on ICD-10. If a patient was diagnosed with one of the first two

boxes (myocardial infarct or congestive cardiac insuffiency/ I21; I22; I23; I50; I11.0; I13.0;

I13.2)

Jairath 2015 History of ischaemic heart disease

Hebert 1999 primary or secondary ICU admission diagnosis of a cardiovascular disease, as well as those

patients with cardiac disease as an important comorbid illness defined as New York Heart

Association class III or IV.

As a second step, we examined all patients who were known to have ischemic heart disease.

The diagnosis most responsible for the patient’s ICU admission was recorded. As many as three

secondary diagnoses and up to eight separate comorbid conditions were identified. The

cardiovascular disease category included all diagnoses related to ischemic heart disease (myocardial infarct, angina, congestive heart failure, and cardiogenic shock), rhythm

disturbances, cardiac arrest, other forms of shock, uncontrolled hypertension, and cardiac and

vascular surgical procedures such as abdominal aortic aneurysm repair and peripheral vascular

surgical procedures.

Holst 2014 history of myocardial infarction, any history of stable or unstable angina pectoris, previous

treatment with nitrates, percutaneous coronary intervention, coronary-artery bypass grafting or

noncoronary vascular interventions, any history of chronic heart failure [defined as New York

Heart Association class III or IV], or any history of cerebral infarction or transitory cerebral

ischemia

Parker 2013 Hypertension

Angina

Previous myocardial infarction

Previous congestive cardiac failure

Other cardiac disease

Walsh 2013 Ischaemic heart disease was defined as evidence from the patient’s records of a previous history

of angina, previous myocardial infarction, or chronic cardiac failure at the time of

randomization.

3

Page 61 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

2

Table E2: Characteristics of all eligible blood transfusion trials not included in quantitative 4

analysis 5 Author Journal Clinical Setting

Trigger Total

n

CV n (%) Primary endpoint Reason for

exclusion

Bergamin

2014

Brazil (S)

Critical Care Critical Care R: 70g/l 73 28 day mortality CVD not baseline

characteristic L: 90g/l 63

Blair

1986

UK (S)

British Journal

of Surgery

GI

haemorrhage

R: 80g/l or

shock

26 Clotting, death CVD not baseline

characteristic

L: 2 units 24

Carson 1998

USA/Scotland

(M)

Transfusion Hip fracture

patients

R: 80g/l or

symptoms of

anaemia

42 19 (45.2) 60 day mortality no reply from

authors

L: 100g/l 42 19 (45.2)

Colomo 2008

Spain (S)

Hepatology

(abstract)

GI

haemorrhage in

cirrhosis

R: 70g/l 109 mortality no further

information

available L: 90g/l 105

Fan 2014

China (S)

Archives of

Gerontology

and Geriatrics

elective Total

hip

replacement

R: 80g/l or

symptoms of

anaemia

41 IHD: 9, BP:

52, CVA:4,

CCF: 2

Postoperative

Delirium (POD)

No reply from

authors

L: 100g/l

42 IHD: 10,

BP: 57,

CVA: 3,

CCF: 3

Fortune 1987

USA (S)

Journal of

Trauma

Traumatic

patients with

class 3-4

haemorrhagic

shock

R: hct near

30%

12 Metabolic stability CVD not baseline

characteristic

L: hct near

40%

13

Foss 2009

Denmark (S)

Transfusion Hip fracture

patients

R: 80g/l 60 28 (46.7) Postoperative

functional mobility

No reply from

authors

L: 100g/l 60 21 (35.0)

Grover 2006

UK (S)

Vox Sanguinis orthopaedic

surgery

R: 80g/l 109 angina: 6

(5.5)

MI: 6 (5.5)

Silent myocardial

ischaemia

ECG abnormalities

excluded

L: 100g/l 109 angina: 8

(7.3)

MI: 6 (5.5)

Haberkern

1997

USA (M)

Blood sickle cell

surgery

R:

conservative

110 Sickle cell events pre-operative HbSS

transfusion target

L: aggressive

120

Not

randomised

134

Hebert 1995

Canada (M)

JAMA Critical care R: 70g/l 33 5 (15.2) Feasibility

Mortality

No reply from

authors

L: 90g/l

36 4 (11.1)

Hochain 1996 Gut (abstract) variceal

bleeding

R: PCV 25 43 Rebleeding CVD not mentioned

L: PCV 32 47

Park H 2008

South Korea

(S)

Cancer Chemo

and Pharm

Gastric

oncology

R: 100g/l Response

Quality of Life

CVD not mentioned

L: 120g/l

Koshy 1988

USA (S)

NEJM sickle cell

pregnancy

R:as

required

36 Perinatal mortality HbSS transfusion

target

Prophylactic

transfusion

36

Liu 2015

China (S)

Chinese

Medical

Journal

Emergency

surgery

R: illness

severity

score

33 Feasibility Exc: patients with

coronary heart

disease

L: standard

care

32

Lotke 1999

USA (S)

Journal of

Arthroplasty

Total knee

arthroplasty

R: 90g/l 62 Not explicit

Hb, LOS, wellbeing

CVD not baseline

characteristic

Page 62 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

3

Author Journal Clinical Setting

Trigger Total

n

CV n (%) Primary endpoint Reason for

exclusion

L: 2 units

starting in

recovery

room

65

Mazza 2015

Brazil (M)

Critical Care R: 70g/l 22 7 (31.8) Effect of

transfusion on

lactate and SvO2

No reply from

authors

L: 90g/l 24 10 (41.7)

Nielsen 2012 Transfusion

medicine

Spinal surgery R: 73g/l 25 subcutaneous

oxygen tension

Exc: cardiac disease

NYHA II or above L: 89g/l 23

Nielsen 2014

Denmark (S)

BMC

Anesthesiology

Elective hip

revision

R: 73g/l 33 5 (15.1) Timed Up and Go-

test

No reply from

authors L: 89g/l 33 7 (21.2)

Palmer 1998

Scotland (S)

Transfusion

Medicine

(abstract)

hip fracture abstract

unavailable

Prick 2013

Netherlands

(M)

BJOG postpartum

haemorrhage

Hb 48-79g/l

R: no

transfusion

262 Physical fatigue Excluded ASA 2-4

patients

L:

transfusion

259

Robertson

2014

USA (M)

JAMA Traumatic brain

injury

R: 70g/l 99 Glasgow Outcome

Scale

CVD not baseline


So-Osman

2004

Netherlands

(S)

Vox Sanguinis

(abstract)

hip/knee

surgery

unclear Blood use CVD not mentioned

So-Osman

2010

Netherlands

(M)

Vox Sanguinis Elective Total

knee or hip

replacement

R: stratified

by risk

299 205 (68.6) Hospital LOS 30 day mortality

not measured

L: standard

care

304 211 (69.4)

Villanueva

2013

Spain (S)

NEJM Upper GI

bleeding

R:70g/l 444 Mortality at 45days CVD not baseline

characteristic,

many CV patients

excluded L: 90g/l 445

Villarejo 1999

Spain (S)

Acta

Gastroenterol

Latinoam.

GI

haemorrhage

R: hct 21% CVD not mentioned

L: hct 28%

Webert 2008

Canada (M)

Transfusion Leukaemia R: 80g/l

29 Bleeding CVD patients

excluded

L: 120g/l 31

Weiss 1982

USA (S)

Lancet (letter) Acute

leukaemia

R: as needed 12 Marrow recovery CVD not baseline


Wu 2011

China

(S)

Intensive Care

Medicine

(abstract only)

liver transplant R: 70g/l 112 30 day mortality CVD not mentioned

L: 100g/l 114

Zheng 2013

China (S)

Experimental

and

Therapeutic

Medicine

Orthopaedic

surgery

elderly patients

Standard 52 blood transfusion

effectiveness

CVD patients

excluded Goal

directed

54

Zygun 2009

UK (S)

Critical Care

Medicine

traumatic brain

injury

R: 8g/dl Change in brain

tissue

oxygenation

CVD not baseline

characteristic Mid: 9g/dl

L: 10g/dl

6

Page 63 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

4

Table E3: Patients receiving allogenic blood. Mean (SD) or Median (IQR) as quoted in orginal 7

paper 8 Restrictive

trigger

Pts

receiving blood

% No. units

transfused

Liberal

trigger

Pts

receiving blood

% No. units

transfused

Bush

1997

90g/l

n=49

39 79.6 1.5 (1.7) 100g/l

n=50

44 88.0 2.4 (2.5)

Carson 2011 80g/l,Sx

anaemia

n=1008

413 41.0 0 (0,1) 100g/l n=1007

973 96.6 2(1,2)

Carson 2013

Acute MI

80g/l n=55

15 27.3 0.49 (1.03)

100g/l n=55

52 94.5 1.58(1.13)

Cooper

2011

Hct <24%

n=24

13 54.2 1.6 (2.0) Hct

<30%

n=21

21 100.0 2.5 (1.3)

Parker Sx anaemia

n=70

11 15.7 0 100g/l

n=67

67 100.0 2 (2,2)

Walsh 70g/l

n=17

12 70.6 1.5

(0.25,3.5)

90g/l

n=15

15 100.0 2 (1,8)

9

Page 64 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

5

Table E4: Diagnosis of Cardiovascular events: 10

I: diagnosed by Investigator, C: diagnosed by Clinical team, U: unclear 11

Blinded

Y/N

Diagnosis

Almeida

2015 (I)

Y Clinical symptoms suggestive of myocardial ischaemia with >=1:

increase or decrease in cardiac troponin I (>=1 value >99th centile URL)

ECG changes: new Q waves, ST-elevation, new LBBB

image based evidence of new loss of viable myocardium

Bush

1997

(U)

? New Q waves and/or CPK elevation (MB fraction >5ng/ml and relative index >2.0)

Carson

2011 (I)

Y Detection of rise or fall of cardiac troponin I with at least one value above the 99th

percentile of the upper reference limit in the context of myocardial ischemia and at least

one of the following:

Symptoms of myocardial ischemia;

New ECG changes indicative of ischemia (eg ST-T changes or new left bundle branch

block [LBBB]) or development of pathological Q waves

Imaging evidence of new loss of viable myocardium or new regional wall motion

abnormality

Autopsy evidence of recent myocardial necrosis

Carson

2013 (I)

Y Rise and/or fall of cardiac biomarkers (preferably troponin) together with evidence of

myocardial ischemia and either symptoms of ischemia or electrocardiogram changes

indicative of new ischemia. Unstable angina was defined as: (1) the absence of elevated

cardiac biomarkers and (2) presence of ischemic symptoms or electrocardiogram changes

indicative of ischemia or (3) chest pain or angina equivalent leading to a coronary artery

intervention (e.g., coronary angioplasty) and (4) hospitalization. Congestive heart failure

required at least one of the following symptoms or signs, new or worsening including:

dyspnea at rest, orthopnea, or paroxysmal nocturnal dyspnea, AND radiological evidence

of heart failure or worsening heart failure AND additional/increased therapy.

Cooper

2011

(C)

N Recurrent ischemic chest discomfort, new ischemic electrocardiographic changes, and

CK-MB increase above the upper limit of normal and increased by 50% over the previous

value. For patients with percutaneous coronary intervention 24 hours previously, CK-

MB >3 times the upper limit of normal and increased by >50% over the previous value

was required. For patients with coronary artery bypass grafting surgery 24 hours

previously, CK-MB >5 times the upper limit of normal and increased by >50% over the

previous value was required.

New or worsening HF was defined as 1 of the following occurring >6 hours after

randomization: cardiogenic shock or a physician’s decision to treat HF with an

intravenous diuretic or intravenous vasoactive drug and evidence of pulmonary vascular

congestion.

Hebert

1999 (I)

? Unclear

Holst

2014 (I)

Y Symptoms, electrocardiographic signs, or elevated biomarker levels resulting in an

intervention

Parker

2013

(C)

N Unclear

Walsh

2013

(C)

N Troponin rise, new ECG change

12

Page 65 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

6

Table E5: Adverse events – General. Mean (SD) 13

Restrictive

trigger

Change

in

MODS

Length of

ICU stay

(d)

LOS

Hospital

(d)

In-hospital

infections

Liberal

trigger

MODS Length of

ICU stay

(d)

LOS

Hospital

(d)

In-hospital

infections

Bush 1997 90g/l 4 (8) days

11 (9)

days

100g/l 4 (4) days

(p>0.6)

10 (6) days

(p>0.6)

Carson 2011

USA

(n=1220)

Canada

(n=791)

80g/l, Sx anaemia

USA

4.0 (3.9)

Canada

12.7 (9.5)

USA 3.7 (3.4)

Canada

12.0 (9.3)

Carson 2013

Acute MI

80g/l 2 100g/l 2

Cooper

2011

Hct <24% 4.3 (3.3)

(CCU)

10.4 (7.2)

Hct

<30%

3.4 (2.3)

(CCU)

4.3 (3.3)

Gregersen 97g/l Hosp

7.8 (5.0)

13

113g/l Hosp

8.4 (8.0)

24

Hebert 1998 70g/l 0.23 +/-

4.2

9.3 (9.7)

28.76

(19.5)

100g/l 1.3 +/-

4.4

10.4 (10.3)

30.6 (18.8)

Parker Sx anaemia 3 100g/l 5

Walsh 70g/l 36.5 (26.7)

53.3 (40.1)

90g/l 25.6 (18.1)

36.3 (28.3)

14

15

Figure Legends: 16

Figure E1: Forest plot with mean difference for hospital length of stay. 17

18

19

20

21

22

23

24

Page 66 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

7

SEARCH STRATEGIES 25 26 CENTRAL 27 #1 MeSH descriptor: [Erythrocyte Transfusion] explode all trees 28 #2 MeSH descriptor: [Blood Transfusion] this term only 29 #3 ((blood or erythrocyte* or "red cell*" or "red blood cell*" or RBC*) near/5 (transfus* or unit* or infus* 30

or therap*)) 31 #4 ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) near/5 (use* or 32

usage* or utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or 33 management or practic* or indicat* or criteri* or standard* or program*)):ab 34

#5 ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or 35 requir* or need* or administ* or replac* or support* or strateg* or management or practic* or indicat* 36 or criteri* or standard* or program*)):ti 37

#6 (leukodeplet* or leukoreduc* or leucodeplet* or leucoreduc* or leukofiltrat* or leucofiltrat*):ti 38 #7 ("allogeneic blood" or (unit* near/2 blood) or "allogenic blood" or (blood near/2 exposure) or "donor 39

blood" or "blood product*" or "blood component*" or "blood support") 40 #8 (hemotransfus* or haemotransfus* or hypertransfus* or hemotherap* or haemotherap*) 41 #9 (red cell* or erythrocyte* or blood or RBC*) and transfus*:ti 42 #10 #1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 43 #11 MeSH descriptor: [Blood Component Transfusion] this term only 44 #12 MeSH descriptor: [Exchange Transfusion, Whole Blood] explode all trees 45 #13 MeSH descriptor: [Plasma Exchange] explode all trees 46 #14 MeSH descriptor: [Platelet Transfusion] explode all trees 47 #15 MeSH descriptor: [Leukocyte Transfusion] explode all trees 48 #16 #12 or #13 or #14 or #15 49 #17 #11 not #16 50 #18 ("red cell*" or "red blood cell*" or erythrocyte* or RBC*) 51 #19 MeSH descriptor: [Erythrocytes] this term only 52 #20 #18 or #19 53 #21 #17 and #20 54 #22 #10 or #21 55 #23 MeSH descriptor: [Hematocrit] this term only 56 #24 ((h?emoglobin or h?emocrit* or HB or HCT) near/5 (level* or concentration* or target* or maintain* 57

or rais* or higher or lower or greater or above or below or equal or transfus*)) 58 #25 #23 or #24 59 #26 #22 and #25 60 #27 ((transfus* or "red cell*" or "red blood cell*" or RBC*) near/10 (trigger* or thresh?old* or target* or 61

restrict* or liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or 62 policies or practic* or standard*)) 63

#28 ((transfus* or "red cell*" or "red blood cell*" or RBC* or h?ematocrit*) and (level* or critical* or 64 intensive* or h?emorrhag* or bleed*) or hypertransfus*):ti 65

#29 #26 or #27 or #28 Publication Year from 2008 to 2014 66 67 MEDLINE (OVID) 68

1. randomized controlled trial.pt. 69

2. controlled clinical trial.pt. 70

3. randomi*.tw. 71

4. placebo.ab. 72

5. clinical trials as topic.sh. 73

6. randomly.ab. 74

7. groups.ab. 75

8. trial.ti. 76

9. or/1-8 77

10. exp animals/ not humans/ 78

11. 9 not 10 79

12. BLOOD TRANSFUSION/ 80

13. ERYTHROCYTE TRANSFUSION/ 81

Page 67 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

8

14. ((blood or erythrocyte* or red cell* or red blood cell* or RBC*) adj5 (transfus* or infus* or unit* or 82

therap*)).ti,ab. 83

15. ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) adj5 (use* or usage* or 84

utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or management or practic* 85

or indicat* or criteri* or standard* or program*)).ab. 86

16. ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or requir* 87

or need* or administ* or replac* or support* or strateg* or management or practic* or indicat* or criteri* or 88

standard* or program*)).ti. 89

17. (allogeneic blood or (unit* adj2 blood) or allogenic blood or (blood adj2 exposure) or donor blood or blood 90

product* or blood component* or blood support).ti,ab. 91

18. (hemotransfus* or haemotransfus* or hypertransfus* or hemotherap* or haemotherap*).tw. 92

19. (red cell* or erythrocyte* or blood or RBC*).tw. and transfus*.ti. 93

20. or/12-19 94

21. BLOOD COMPONENT TRANSFUSION/ 95

22. EXCHANGE TRANSFUSION, WHOLE BLOOD/ or PLASMA EXCHANGE/ or PLATELET 96

TRANSFUSION/ or exp LEUKOCYTE TRANSFUSION/ 97

23. 21 not 22 98

24. ERYTHROCYTES/ 99

25. (red cell* or red blood cell* or erythrocyte* or RBC*).tw. 100

26. 24 or 25 101

27. 23 and 26 102

28. 20 or 27 103

29. *HEMATOCRIT/ 104

30. ((h?emoglobin or h?ematocrit* or HB or HCT) adj5 (level* or concentration* or target* or maintain* or 105

rais* or higher or lower or greater or above or below or equal or transfus*)).tw. 106

31. 29 or 30 107

32. 28 and 31 108

33. ((transfus* or red cell* or red blood cell* or RBC*) adj10 (trigger* or thresh?old* or target* or restrict* or 109

liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or policies or practice* 110

or standard*)).tw. 111

34. (((transfus* or red cell* or red blood cell* or RBC* or h?ematocrit*) and (level* or critical* or intensive* or 112

h?emorrhag* or bleed*)) or hypertransfus*).ti. 113

35. 32 or 33 or 34 114

36. 11 and 35 115

116 EMBASE (OVID) 117

1. Randomized Controlled Trial/ 118

2. Randomization/ 119

3. Single Blind Procedure/ 120

4. Double Blind Procedure/ 121

5. Crossover Procedure/ 122

6. Placebo/ 123

7. exp Clinical Trial/ 124

8. Prospective Study/ 125

9. (randomi* or double-blind* or single-blind* or RCT*).tw. 126

10. (random* adj2 (allocat* or assign* or divid* or receiv*)).tw. 127

11. (crossover* or cross over* or cross-over* or placebo*).tw. 128

12. ((treble or triple) adj blind*).tw. 129

13. or/1-12 130

14. Case Study/ 131

15. case report*.tw. 132

Page 68 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

9

16. (note or editorial).pt. 133

17. or/14-16 134

18. 13 not 17 135

19. (animal* or cat or cats or dog or dogs or pig or pigs or sheep or rabbit* or mouse or mice or rat or rats or 136

feline or canine or porcine or ovine or murine).ti. 137

20. 18 not 19 138

21. limit 20 to embase 139

22. BLOOD TRANSFUSION/ 140

23. ERYTHROCYTE TRANSFUSION/ 141

24. ((blood or erythrocyte* or red cell* or red blood cell* or RBC*) adj5 (transfus* or infus* or unit* or 142

therap*)).ti,ab. 143

25. ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) adj5 (use* or usage* or 144

utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or management or practic* 145

or indicat* or criteri* or standard* or program*)).ab. 146

26. ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or requir* 147

or need* or administ* or replac* or support* or strateg* or management or practic* or indicat* or criteri* or 148

standard* or program*)).ti. 149

27. (allogeneic blood or (unit* adj2 blood) or allogenic blood or (blood adj2 exposure) or donor blood or blood 150

product* or blood component* or blood support).ti,ab. 151

28. (hemotransfus* or haemotransfus* or hypertransfus* or hemotherap* or haemotherap*).tw. 152

29. (red cell* or erythrocyte* or blood or RBC*).tw. and transfus*.ti. 153

30. 22 or 23 or 24 or 25 or 26 or 27 or 28 or 29 154

31. BLOOD COMPONENT THERAPY/ 155

32. GRANULOCYTE TRANSFUSION/ or LEUKOCYTE TRANSFUSION/ or LYMPHOCYTE 156

TRANSFUSION/ or PLASMA TRANSFUSION/ or THROMBOCYTE TRANSFUSION/ 157

33. 31 not 32 158

34. ERYTHROCYTE/ 159

35. (red cell* or red blood cell* or erythrocyte* or RBC*).tw. 160

36. 34 or 35 161

37. 33 and 36 162

38. 30 or 37 163

39. *HEMATOCRIT/ 164

40. ((h?emoglobin or h?ematocrit* or HB or HCT) adj5 (level* or concentration* or target* or maintain* or 165

rais* or higher or lower or greater or above or below or equal or transfus*)).tw. 166

41. 39 or 40 167

42. 38 and 41 168

43. ((transfus* or red cell* or red blood cell* or RBC*) adj10 (trigger* or thresh?old* or target* or restrict* or 169

liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or policies or practice* 170

or standard*)).tw. 171

44. (((transfus* or red cell* or red blood cell* or RBC* or h?ematocrit*) and (level* or critical* or intensive* or 172

h?emorrhag* or bleed*)) or hypertransfus*).ti. 173

45. 42 or 43 or 44 174

46. 21 and 45 175

176 CINAHL (EBSCOHost) 177 S1 (MH "Blood Transfusion") 178 S2 (MH "Erythrocyte Transfusion") 179 S3 (erythrocyte* or red cell* or red blood cell* or blood or RBC*) N5 (transfus* or infus* or therap* or 180

unit*) 181 S4 AB ((red cell* or RBC* or erythrocyte* or red blood cell* or whole blood or transfus*) N5 (use* or 182

usage* or utiliz* or utilis* or requir* or need* or administ* or replac* or support* or strateg* or 183 management or practic* or indicat* or criteri* or standard* or program*)) 184

Page 69 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

10

S5 TI ((red cell* or RBC* or erythrocyte* or blood or transfus*) and (use* or usage* or utiliz* or utilis* or 185 requir* or need* or administ* or replac* or support* or strateg* or management or practic* or indicat* 186 or criteri* or standard* or program*)) 187

S6 hemotransfus* or haemotransfus* or hemotherap* or haemotherap* or hypertransfus* 188 S7 TX (red cell* or erythrocyte* or blood or RBC*) and TI (transfus*) 189 S8 S1 OR S2 OR S3 OR S4 OR S5 OR S6 OR S7 190 S9 (MM "Hematocrit") 191 S10 TX ((hemoglobin or haemoglobin or hematocrit* or haematocrit* or HB or HCT) N5 (level* or 192

concentration* or target* or maintain* or rais* or higher or lower or greater or above or below or equal 193 or transfus*)) 194

S11 S9 OR S10 195 S12 S8 AND S11 196 S13 TI ((transfus* or red cell* or red blood cell* or RBC* or haematocrit* or hematocrit*) and (level* 197 or critical* or intensive* or haemorrhag* or hemorrhag* or bleed*)) 198 S14 TX ((transfus* or "red cell*" or "red blood cell*" or RBC*) N10 (trigger* or threshold* or target* or 199

restrict* or liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or 200 policies or practice* or standard*)) 201

S15 S12 OR S13 OR S14 202 S16 (MH Clinical Trials+) 203 S17 PT Clinical Trial 204 S18 TI ((controlled trial*) or (clinical trial*)) OR AB ((controlled trial*) or (clinical trial*)) 205 S19 TI ((singl* blind*) OR (doubl* blind*) OR (trebl* blind*) OR (tripl* blind*) OR (singl* mask*) OR 206

(doubl* mask*) OR (tripl* mask*)) OR AB ((singl* blind*) OR (doubl* blind*) OR (trebl* blind*) OR 207 (tripl* blind*) OR (singl* mask*) OR (doubl* mask*) OR (tripl* mask*)) 208

S20 TI randomi* OR AB randomi* 209 S21 MH RANDOM ASSIGNMENT 210 S22 TI ((phase three) or (phase III) or (phase three)) or AB ((phase three) or (phase III) or (phase three)) 211 S23 (TI (random* N2 (assign* or allocat*)) ) OR (AB (random* N2 (assign* or allocat*)) ) 212 S24 MH PLACEBOS 213 S25 S16 OR S17 OR S18 OR S19 OR S20 OR S21 OR S22 OR S23 OR S24 214 S26 S12 AND S25 215 216 PUBMED (epublications only) 217 #1 ("erythrocyte transfusion*" OR "red cell transfusion" OR "red cells" OR "red blood cells" OR "red blood 218 cell transfusion" OR "blood transfusion*" OR RBCs OR "RBC transfusion*" OR hemotransfus* OR 219 haemotransfus* OR hemotherap* OR haemotherap* OR hypertransfus*) 220 #2 ((hemoglobin OR haemoglobin OR haematocrit* OR hematocrit* OR HB OR HCT) AND (level* or 221 concentration* OR target* OR maintain* OR rais* OR higher OR lower OR greater OR above or below OR 222 equal OR transfus*)) 223 #3 (random* OR blind* OR control group* OR placebo OR controlled trial OR controlled study OR trials OR 224 systematic review OR meta-analysis OR metaanalysis OR literature OR medline OR cochrane OR embase) 225 AND ((publisher[sb] OR inprocess[sb]) NOT pubstatusnihms) 226 #4 #1 AND #2 AND #3 227 228 LILACS 229 (( tw:(transfused OR transfusing OR transfused OR hypertransfusion OR haemoglobin OR hemoglobin OR 230 haematocrit OR hematocrit) AND (instance:"regional") AND ( db:("LILACS") AND type 231 of_study:("clinical_trials"))) AND (instance:"regional") AND ( year_cluster:("2009" OR "2010" OR "2011" OR 232 "2012" OR "2013" OR "2014")) 233 234 TRANSFUSION EVIDENCE LIBRARY 235 Subject Area: Blood Components/Red Cells 236 OR 237 All fields: trigger OR triggers OR threshold OR thresholds OR haemoglobin OR haemoglobin OR haematocrit 238 OR hematocrit OR Hb OR HCT 239 240 WEB OF SCIENCE – CPCIS database 241 #1 TS=(("erythrocyte transfusion*" OR "red cell transfusion" OR "red cells" OR "red blood cells" OR "red 242 blood cell transfusion" OR "blood transfusion*" OR RBCs OR "RBC transfusion*" OR hemotransfus* or 243 haemotransfus* or hemotherap* or haemotherap* or hypertransfus*) 244

Page 70 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

11

#2 TS=((hemoglobin or haemoglobin or hematocrit* or haematocrit* or HB or HCT) NEAR/1 (level* or 245 concentration* or target* or maintain* or rais* or higher or lower or greater or above or below or equal or 246 transfus*)) 247 #3 #1 AND #2 248 #4 TS=((transfus* or "red cell*" or "red blood cell*" or RBC*) NEAR/5 (trigger* or threshold* or target* or 249 restrict* or liberal* or aggressive* or conservative* or prophylactic* or limit* or protocol* or policy or policies 250 or practice* or standard*)) 251 #5 #3 OR #4 252 #6 TS=(randomi* OR randomly OR "random assignment" OR "random allocation" OR blind* OR "control 253 group*" OR controlled trial OR "controlled study") 254 #7 #5 AND #6 255 256 Ongoing Trials: 257 ClinicalTrials.gov & WHO ICTRP 258 Search Terms/Title: randomized OR randomised OR randomly 259 Intervention: red cell transfusion OR RBC transfusion OR blood transfusion 260 Title: trigger OR threshold OR target OR restrictive OR liberal OR aggressive OR conservative OR prophylactic 261 OR limit OR protocol OR policy OR policies OR practice OR standard OR hemoglobin OR hematocrit 262 263 ISRCTN & EUCTR [terms searched in combination and individually] 264 (red cell transfusion OR red blood cell transfusion OR RBC transfusion OR blood transfusion) AND (trigger 265 OR threshold OR target OR restrictive OR liberal OR aggressive OR conservative OR prophylactic OR limit 266 OR protocol OR policy OR policies OR practice OR standard OR hemoglobin OR hematocrit OR haemoglobin 267 OR haematocrit) 268 269 Hong Kong Clinical Trials Registry [terms searched individually] 270 Disease Group: Blood and blood-forming organs OR Circulatory System 271

Title: randomized OR randomised OR transfusion 272

273

Page 71 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyThe impact of restrictive versus liberal transfusion strategies on patient outcomes in patients

with cardiovascular disease excluding those undergoing cardiac surgery: A Systematic Review

and Meta-analysis

Annemarie B Docherty1,2*, Rob O’Donnell2, Susan Brunskill3, Marialena Trivella3, Carolyn Doree4,

Lars Holst5, Martyn Parker6, Merete Gregersen7, Juliano Pinheiro de Almeida8, Timothy S Walsh1,2,

Simon J Stanworth3,9

PRINT ABSTRACT

Study Question: To compare patient outcomes with restrictive vs liberal red cell transfusion

strategies in patients with cardiovascular disease, excluding those undergoing cardiac surgery.

Methods: This was a systematic review with meta-analyses of in-hospital red cell transfusion

threshold randomised controlled trials that evaluated a restrictive vs liberal transfusion threshold and

that included patients with cardiovascular disease. We performed a comprehensive search of

databases and data extraction in duplicate. We assessed risk of bias using Cochrane methodology,

using a Mantel-Haenzel random effects model to pool the risk ratios. Our outcome measures were

thirty-day mortality, and cardiovascular events. We contacted authors for relevant data whenever

possible.

Study answer and limitations: In total, 11 trials (out of 41 identified eligible trials) enrolling patients

with cardiovascular disease (n=3,033) were included for meta-analysis (restrictive transfusion

threshold, n=1,514 patients; liberal transfusion threshold, n= 1,519). The pooled risk ratio for the

association between transfusion thresholds and 30 day mortality was 1.15 (95% CI 0.88 to 1.50,

P=0.50) with little heterogeneity (I2=14%). There was an increased risk of acute coronary syndrome

(ACS) in patients managed with a restrictive compared with a liberal transfusion threshold (9 trials;

RR 1.78, 95% CI 1.18 to 2.70, P=0.01, I2= 0%).

Page 72 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nlyWhat this study adds: This review indicates that restrictive blood transfusion strategies may not be

as safe as more liberal transfusion strategies for patients with co-existing cardiovascular disease in

non-cardiac surgery settings.

Specifically, we have shown an increased risk of acute coronary syndrome with restrictive triggers

below 80g/l. These data support the use of a more liberal transfusion threshold (greater than 80g/l) for

patients with both acute and chronic cardiovascular disease, until adequately powered high quality

randomised trials have been undertaken in this patient population.

Trial registration: PROSPERO: CRD42014014251(1); http://www.crd.york.ac.uk/PROSPERO/).

Funding: nil

Competing interests: nil

Data sharing: no additional data available

Word count: 299

Suggested figure: Forest plot with Risk Ratios for Adverse Cardiovascular Events. Risk of bias

assessment included for each study.

Page 73 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Page 74 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Page 75 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Page 76 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960


nly

Page 77 of 77


BMJ

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960

Documents

Confidential: For Review Only - BMJ · Confidential: For Review Only 1 1 The impact of restrictive versus liberal transfusion strategies on patient 2 outcomes in patients with cardiovascular