Anaemia: A risk factor for death and adverse outcomes following surgery for acute lower limb ischaemia

Abstract

Introduction

Acute limb ischaemia (ALI) forms a significant part of the vascular surgery workload and carries with it high rates of morbidity and mortality. Anaemia is also common amongst vascular surgical patients and has been linked with poor outcomes in some subgroups. We aimed to assess the frequency of anaemia in patients with ALI and its impact on survival and complications following revascularisation to help direct future efforts to optimise outcomes in this patient group.

Methods

A retrospective analysis of prospectively collected departmental data on patients undergoing surgical intervention for ALI between 2014 and 2018 was performed. Anaemia was defined as a pre-operative haemoglobin (Hb) of <120 g/L for women and <130 g/L for men. The primary outcome was overall survival, assessed with the Kaplan–Meier estimator, with application of Cox proportional hazard modelling to adjust for confounding covariates.

Results

There were 158 patients who underwent treatment for ALI: 89 (56.3%) of these were non-anaemic with a mean Hb of 146 (SD = 18.4), and 69 (43.7%) were anaemic with a mean Hb of 106 (SD = 13.4). Anaemic patients had a significantly higher risk of death than their non-anaemic counterparts on univariate analysis (HR = 2.11, 95% CIs, 1.28–3.5, p = 0.0036). There was ongoing divergence in survival up to around 6 months between anaemic and non-anaemic groups. Under the Cox model, anaemia was similarly significant as a predictor of death (HR = 2.15, 95% CIs, 1.17–3.95, p = 0.013), accounting for recorded comorbidities, medication use and blood transfusion.

Conclusions

Anaemia is a significant and independent risk factor for death following revascularisation for ALI and can be potentially be modified. Vascular surgical centres should ensure they have robust pathways in place to identify and consider treating anaemia. There is scope for further work to assess how to best optimise a patient’s levels of circulating haemoglobin.

Keywords

Anaemia acute limb ischaemia

Introduction

Acute limb ischaemia (ALI) contributes 10–16% of the vascular surgical workload¹ with the exact incidence largely unknown. A recent UK-based study placed incidence of ALI at 10 cases per 100,000 population per year.² ALI is a limb- and life-threatening condition with poor 30-day mortality of 9–24% and rates of limb loss in the short term of 8–14%.^2,3 The management of ALI is often a balance between limb viability and patient frailty. Some patients may require a primary amputation due to irreversible ischaemia or comorbidity precluding revascularisation, while some succumb to other simultaneous arterial-bed acute ischaemia. Patients with a salvageable but threatened limb should be offered prompt revascularisation.⁴

The ALI patient cohort typically has a high burden of comorbidities, including anaemia. Anaemia is defined by the World Health Organisation as a circulating haemoglobin (Hb) concentration less than 130 g/L for men and 120 g/L for women, and has a number of aetiologies⁵ with iron deficiency being the most common.⁶ Anaemia is common in patients presenting for surgery (up to a third of vascular surgical patients) and is associated with an increased risk of complications, duration of hospital stay and mortality.^7–10 There is more specifically some limited evidence that it is predictive for mortality following treatment for critical limb ischaemia¹⁰ and amputation for patients hospitalised with peripheral arterial disease.¹¹ The 2020 Commissioning for Quality and Innovation (CQUIN) indicators¹² suggest the investigation and treatment of anaemia in elective vascular surgical patients, but given an absence of specific evidence about the impact of anaemia on acute limb ischaemia outcomes, there are, as yet, no specific recommendations.

In other specialities, studies have demonstrated that non-transfusion pre-operative treatment of anaemia can improve post-operative haemoglobin^13,14 and reduce red blood cell transfusion requirements.¹⁵ There is also some evidence that both length of stay and infection rates may also be reduced, without change in overall mortality.¹⁶ The acute nature of ALI means that the correction of anaemia peri-operatively is limited largely to red blood cell transfusion which may be associated with increases in mortality^17–19 and complication rates²⁰ though some analyses do not demonstrate this effect.²¹

Developing an understanding of the effects of anaemia on outcomes following ALI is therefore important: it is a common vascular presentation with relatively poor outcomes, and if anaemia is contributing to these poor outcomes, there are opportunities for improvement. We therefore aim to assess the effect of anaemia and peri-operative blood transfusion on the outcomes of patients undergoing surgery for ALI.

Methods

Study design

A prospectively maintained database of all vascular procedures at the Northern Vascular Centre, Freeman Hospital, Newcastle-Upon-Tyne (a UK tertiary centre), was retrospectively reviewed over a 5-year period (2014–2018), with any missing data retrospectively acquired by the investigators.

Anaemia was defined as per WHO guidelines, with a threshold set as an Hb lower than 130 g/L for men and 120 g/L for women.²² Baseline characteristics were collected, including demographics, comorbidities, statin, antiplatelet and anticoagulant use, pre-operative Hb, sodium, creatinine, eGFR and units of blood transfused during admission. Hb was collected from admission, last measurement before discharge or death and at 6-week follow-up. Creatinine was measured at admission, day 3 (±24 h), day 7 (±3 days) and at follow-up.

Study population

Patients were identified by searching the database for emergency procedures (open, endovascular or hybrid) performed to resolve acute limb ischaemia, as listed in Table 1. Patients were considered to have ALI if they were identified as having limb-threatening ischaemia, with a clear history of a sudden onset or deterioration of symptoms. Patients who were managed conservatively (systemic anticoagulation or palliation) or who had a primary amputation were not included. Results were then manually verified, and cases not related to the lower limb were excluded; see Table 2 for demographics.

Table 1.

Case selection.

Demographic	Non-anaemic	Anaemic	p
Number (%)	89 (56.3)	69 (43.7)
Age (median (IQR))	69 (26)	73 (18)	0.28
Gender	M 55 : F 34	M 30 : F 39	0.03
Hb (mean (SD))	146 (18.38)	106 (13.42)	<0.001
Admission
Days to surgery (median (IQR))	0 (1)	1 (3)	<0.001
Duration of surgery (min)	120 (60.3)	120 (76)	0.934
Transfused (%)	12 (13.5)	32 (46.4)	<0.001
Comorbidities (%)
IHD	13 (14.6)	16 (23.2)	0.214
COPD	9 (10.1)	7 (10.2)	1
CCF	6 (6.74)	1 (1.45)	0.138
ESRF	1 (2.25)	2 (1.45)	1
Severe CKD	3 (3.37)	3 (4.35)	1
Stroke	8 (8.99)	7 (10.15)	1
Diabetes	16 (17.98)	17 (24.6)	0.33
Hypertension	48 (53.9)	49 (71.0)	0.03
Hypercholesterolaemia	29 (47.7)	42 (42.7)	0.63
Risk factors (%)
Current smoker	30 (34.9)	18 (28.1)	0.48
Ex-smoker	63 (73.3)	43 (67.2)	0.47
Antiplatelet	16 (18.0)	20 (29.0)	0.13
Anticoagulant	11 (12.24)	16 (23.2)	0.09
Revascularisation (%)
Embolectomy	87 (97.8)	67 (97.1)	1
Bypass	1 (1.12)	4 (5.8)	0.17
Stenting	6 (7.25)	5 (7.24)	1
Thrombolysis	5 (5.62)	1 (1.44)	0.23
Fasciotomy	16 (18)	9 (13)	0.51
Angiogram	13 (14.6)	7 (10.2)	0.48
Multiple	29 (32.6)	18 (26.1)	0.39

Table 2.

Demographics.

Outcome	Non-anaemic (n = 89)	Anaemic (n = 69)	OR	95% CI	p
30-day survival (%)	83 (93.3)	57 (82.6)	0.35	0.1–1.064	0.045
Amputation (%)	5 (5.61)	12 (17.4)	3.51	1.08–13.4	0.021
Return to theatre (%)	13 (15.1)	20 (29.4)	2.33	0.99–5.61	0.047
Transfusion (%)	12 (13.4)	32 (43.5)	5.48	2.43–13.1	<0.001
AKI (%)	21 (23.6)	24 (34.8)	1.72	0.81–3.68	0.16
Median days LOS (IQR)	10 (15)	12 (16.5)			0.21

Diabetes was defined by documented medical history, the use of oral antidiabetic agents or insulin or fasting plasma glucose levels of at least 1.26 g/L; hypertension was defined by documented medical history and use of antihypertensive drugs for this purpose, or a systolic blood pressure of at least 140 mmHg or diastolic blood pressure of at least 90 mmHg determined at admission (average of first two measurements). The following other conditions were noted within the documented medical history: ischaemic heart disease, cerebrovascular disease (stroke – both ischaemic or haemorrhagic – and TIA), end-stage renal failure requiring dialysis and chronic obstructive pulmonary disease. A myocardial event was defined as a Type 1 or 2 myocardial infarction as per European Society of Cardiology guidelines.²³ Statin, antiplatelet or anticoagulant usage was noted in the patient’s documented medication list at the time of admission. Smoking status was noted and categorised as non-smoker, ex-smoker or current smoker. eGFR was calculated from serum creatinine using the chronic kidney disease epidemiology formula which has been validated in a broad range of populations.²⁴ Severe chronic kidney disease was defined by GFR categories G4 or G5 (eGFR <30) as per The Renal Association,²⁵ and acute kidney injury was defined by at least Stage 1 AKI as specified in National Institute for Clinical and Healthcare Excellence’s guidance,²⁶ though without reference to urine output as this variable was not reliably recorded.

Follow-up

Information regarding patients’ follow-up visits was entered prospectively into the departmental database. Follow-up was dependent on individual patient requirements. This usually took the form of an initial 6-week out-patient clinical assessment, with further follow-up as appropriate depending on wound healing and symptoms. Follow-up to the point of analysis was supplemented by retrospective review of electronic and paper hospital and general practice records.

Ethics and reporting

The study was approved by the Newcastle-Upon-Tyne Hospital research department and the Newcastle University Review Board. Written informed consent was obtained from all participating subjects according to the Declaration of Helsinki.

The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidance was used to ensure accurate reporting of study findings.²⁷

Outcomes

The primary outcome of interest is all-cause mortality at the end of follow-up.

Secondary outcomes included the length of hospital stay, number of myocardial events within 30 days, post-operative RBC transfusion, renal function, re-intervention rates, limb loss and 30-day mortality.

Length of stay was defined as the duration in days spent at the tertiary arterial centre until discharge home or to a rehabilitation hospital. Follow-up renal function was recorded as eGFR. Overall re-intervention rate was defined as the absolute number of patients who underwent endovascular or open surgical procedure during the follow-up period. Limb loss was defined as any ipsilateral (to the ischaemic event) major lower limb amputation within the follow-up period.

Statistical analysis

Data were recorded in a dedicated database and collected during the dedicated clinic follow-up. Normality was assessed with the Shapiro–Wilk test. Normally distributed data are presented as means (standard deviations (SD)) and hypothesis testing performed with unpaired t-tests (no paired data were assessed). Non-normally distributed data are presented as median (inter-quartile range (IQR)) values with analysis using the Mann–Whitney U test. Categorical data were analysed by Fisher’s exact test. Statistical analysis was performed with Microsoft R Open 3.5.2 (Microsoft, Redmond, USA), using the survival,²⁸ survminer,²⁹ physiology,³⁰ forestmodel³¹ and pROC³² packages. A p value of <0.05 was considered statistically significant. Kaplan–Meier survival curves were used with the log-rank test to compare the overall mortality. Cox regression modelling was used to assess the survival according to the presence of anaemia, with multivariate regression analysis performed with a model designed to adjust for all recorded comorbidities and risk factors, with the inclusion of an interaction term between Hb and transfusion given the likely relationship of transfusion as an intermediate factor between anaemia and adverse outcomes. ROC analysis was performed, with calculation of the Youden Index³³ to identify a threshold Hb to predict mortality.

Results

Over the study period, there were 158 patients who underwent treatment for ALI. 89 (56.3%) of these were non-anaemic with a mean Hb of 146 (SD = 18.4), and 69 (43.7%) were anaemic with a mean Hb of 106 (SD = 13.4).

Patient and admission demographics, comorbidities and risk factors are listed in Table 1. There was a preponderance of females within the anaemic group (p = 0.03), though the ages of both groups were similar (p = 0.28). Other than hypertension (anaemic = 71%, non-anaemic = 53.9%, p = 0.03), there were no significant differences in comorbidities or risk factors between the two groups. Patients who were taking either antiplatelets or anticoagulants were not more likely to be anaemic. Non-anaemic patients were more likely to be operated on sooner (0.5 vs 1 day, p < 0.001), and both groups had a comparable duration of surgery. Anaemic patients were more likely to be transfused (46.4 vs 13.5%, p < 0.001). Most patients across both anaemic and non-anaemic groups had an embolectomy as part of their revascularisation (97.1 vs 97.8%, p = 1), a relatively small proportion had an angiogram (10.2 vs 14.6%, p = 0.48) and some had multiple types of procedure (26.1 vs 32.6%, p = 0.39) with no difference found between groups in the treatment modality utilised.

Primary outcome

Anaemic patients had a significantly worse survival than their non-anaemic counterparts on univariate analysis (HR = 2.11, 95% CIs, 1.28–3.5, p = 0.0036). This is demonstrated in the Kaplan–Meier cumulative survival graph (Figure 1) where there is early and significant divergence in survival between the two groups (p = 0.0029).

Figure 1.

Kaplan–Meier survival graph.

There was no difference in survival detected between groups based on the presence of anaemia post-operatively (p = 0.17), though by this point there had been a significant shift in the proportion of anaemic patients from 43.7% at admission to 76.7%.

Secondary outcomes

Details of secondary outcomes are given in Table 3. Following surgery for ALI, anaemic patients were more likely to have post-operative complications. They had a worse 30-day survival (82.6% vs 93.3%, OR= 0.35, CI = 0.1–10.9, p = 0.045) and were significantly more likely to receive a blood transfusion (43.5% vs 13.4%, OR = 5.48, CI = 2.43–13.1, p < 0.001), return to theatre (29.4% vs 15.1%, OR = 2.33, CI = 0.99–5.61, p = 0.047) or subsequently have a major lower limb amputation (17.4% vs 5.61%, OR = 3.51, CI = 1.08–13.4, p = 0.021). There was no significant difference in either rate of acute kidney injury or myocardial infarction, or length of stay.

Table 3.

Secondary outcomes.

Term
Embolectomy
Thrombectomy
Thrombolysis
Lysis
Endarterectomy

Regression analysis

Cox proportional hazard analysis was performed (Figure 2). Under our model, anaemia was independently significant as a predictor of death (HR = 2.15, 95% CIs, 1.17–3.95, p = 0.013). Age was also significant (HR = 1.03 per year, p = 0.004). No other comorbidities or risk factors were found to significantly impact on survival.

Figure 2.

Cox regression, variables associated with all-cause mortality.

Transfusion analysis

Kaplan–Meier survival analyses were also conducted to compare the survival of patients who were transfused during their admission with ALI against those who were not (Figure 3), demonstrating statistically significant worse survival (25.9 months vs 54.5 months, p = 0.006). There was no difference in survival between patients who were transfused pre-operatively and those transfused at other times during their admission (p = 0.81), though numbers of pre-operative transfusions were small (n = 13). Intraoperative blood loss volumes were unavailable.

Figure 3.

Kaplan–Meier survival graph.

When transfusion was included within the multivariate model along with an interaction term between admission anaemia and transfusion, receiving a blood transfusion became significant as an independent predictor of death, with no significant change in any of the other covariates (Figure 4). The interaction term between anaemia and transfusion borders on statistical significance and may suggest that the increased risk of death associated with transfusion is attenuated if the patient was anaemic at admission (HR = 0.31, CIs = 0.09–1.03, p = 0.056).

Figure 4.

Cox regression including transfusion interaction factor and variables associated with all-cause mortality. Source: reproduced with permission from Richards T, Baikady RR, Clevenger B, et al. Preoperative intravenous iron to treat anaemia before major abdominal surgery (PREVENTT): a randomised, double-blind, controlled trial. Lancet (London, England) 2020; 396(10259): 1353–1361.

ROC analysis

ROC analysis was performed to identify a cut-off haemoglobin for mortality at 3 years following surgery for ALI (Figure 3). This demonstrated a haemoglobin of 124.5 g/L was the most sensitive (66.7%) and specific (62.5%) threshold (AUC = 65.4%).

Discussion

This study demonstrates that anaemia at hospital admission with acute limb ischaemia is a strong predictor for overall mortality, independently accounting for a 2.2-fold increase in risk of death. Presentation with ALI and concurrent anaemia is also associated with an increase in other undesirable outcomes such as requirement for blood transfusion, major lower limb amputation and return to the operating theatre.

Age is an established a risk factor for mortality following vascular surgery,³⁴ so our replication of this finding is not a surprise. Anaemia however is a modifiable risk factor. Whilst the mode of presentation of acute limb ischaemia rarely offers the opportunity to pre-optimise patients as in planned surgery, there is still the chance to investigate and manage patients’ anaemia in the post-operative period. Here, underlying pathology can be identified and treated, and haemoglobin levels corrected. There was an early negative impact of anaemia on survival following admission with ALI; however, this effect continued up to around 6 months, suggesting there is time for post-operative intervention to have benefit. We did find that non-anaemic patients were operated on average 0.5 days sooner, though there is doubt as to the practical significance of this finding as admission times were only recorded to the nearest day.

Data on the severity of ischaemia which patients presented with were not collected. Whilst this will have a significant impact on the prognosis of life and limb, all patients studied had a revascularisation procedure during their admission, implying the presence of significant limb-threatening ischaemia. There was no difference in survival regardless of whether patients were revascularized in the first 24 h of their admission or not, which suggests more significant ischaemia was probably not correlated with worse survival.

We found that blood transfusion was also positively associated with risk of death, which is in agreement with numerous other studies of vascular surgical patients.^17–20 It is notable that these studies assessed the risk of intra and post-operative transfusion which may suggest transfusion reactionary to intraoperative blood loss. This may in turn be associated with more complex or difficult surgical cases. There is a paucity of evidence assessing the effects of pre-operative transfusions which may be given to attempt to optimise Hb levels before patients undergo surgery, and our patient numbers were not large enough to conduct such an analysis. Our analysis suggests, however, that transfusions given to patients with pre-operative anaemia are less harmful, which makes sense: transfusions given to patients who were non-anaemic at admission were likely to be reactionary to higher levels of intraoperative blood loss, procedural complexity and physiological disturbance, which are all also likely to be associated with worse outcomes (though we did not have this data available).

The National Institute for Clinical and Healthcare Excellence recommended in 2016⁴ that anaemia, bleeding risks and anticoagulation be carefully considered for patients undergoing surgery. In addition, there is recognition that an individualised patient blood management (PBM) pathway is an important way to achieve this.³⁵ PBM is a three-pillar WHO-endorsed process, consisting of management of pre-operative anaemia, minimising bleeding and optimising the physiological tolerance of anaemia. We have shown ALI patients are often anaemic; they are actively anticoagulated as part of treatment⁴ (and therefore at greater bleeding risk), and they are frequently transfused, alongside high rates of cardiorespiratory and metabolic comorbidity therefore making all three pillars highly relevant to this already high-risk patient group. Implementing PBM is known to improve patient outcomes,³⁶ but frequently, there are barriers to this.³⁷ In centres where there is no definitive PBM strategy in place for ALI patients, it is therefore likely that implementing one would be highly beneficial. Whilst we did not directly assess concordance with principles of PBM within this dataset, there was no definitive PBM strategy in place at the time of data collection.

A significant question raised by this piece of work is whether anaemia is a cause of poor outcomes or whether it is a marker of an underlying state of disease and poor physiological reserve. One of the limitations of this retrospective review is the lack of data on the underlying cause of anaemia or associated medical conditions such as occult malignancy that may have contributed to the increased mortality¹¹ for these patients. Control and matching of potential confounders are challenging within this observational study which we have tried to address with the use of multivariate analyses.

None the less, the association of pre-operative anaemia with adverse outcome is ignored at our peril. These are some of the only data assessing the effects of anaemia in ALI patients, and our findings are paralleled by studies of patients undergoing lower limb vascular surgery for CLTI, showing similar negative associations with survival, complications and limb loss.^9,11,20 Whilst the acute nature of ALI makes pre-operative optimisation difficult (especially given uncertainty over the benefit of pre-operative transfusion), our analysis demonstrates divergence in survival between anaemic and non-anaemic groups beyond the acute phase of illness. This suggests that there is benefit to be derived from non-transfusion treatment with a latent effect (e.g. iron infusion) and the thorough investigation and treatment of any underlying causes of anaemia in the recovery period. However, this has not been tested in the post-operative setting with pre-operative elective anaemia treatment only improving re-admission rates and not haemoglobin; see Figure 5 for the ROC curve.

Figure 5.

ROC curve.

Areas for future work identified by this study include the effects of transfusion timing and non-transfusion strategies for management of anaemia, and assessment of adherence with PBM principles within vascular centres, including investigation and treatment of anaemia beyond the patient’s initial admission.

Conclusion

This UK tertiary centre study highlights the significant prevalence and impact of pre-operative anaemia in vascular surgical patients presenting with acute lower limb ischaemia. Anaemia is associated with a significant increase in mortality, post-operative complication and blood transfusion requirement. Acute limb ischaemia is a common problem with a high associated morbidity and mortality. Anaemia is a potentially modifiable disease state, and further work should focus on strategies to optimise it in emergency vascular surgical patients to establish if this can lead to improved outcomes.

Footnotes

Acknowledgements

The authors would like to acknowledge the contributions of the vascular consultants Miss Anne Burdess, Mr Mike Clarke, Mr Craig Nesbitt, Mr James McCaslin, Prof G. Stansby, Miss Lucy Wales and Mr Mike Wyatt at the Northern Vascular Centre. In addition, the authors offer special thanks to the research nurses, Deborah Amis, Martin Catterson and Noala Parr.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: SN would like to acknowldege the support of the NIHR who fund his salary.

ORCID iDs

Daniel Perren

Sandip Nandhra

References

Dormandy

Heeck

Vig

. Acute limb ischemia. Semin Vasc Surg 1999; 12(2): 148–153.

Howard

DPJ

Banerjee

Fairhead

, et al. Population-based study of incidence, risk factors, outcome, and prognosis of ischemic peripheral arterial events. Circulation 2015; 132(19): 1805–1815. DOI: 10.1161/CIRCULATIONAHA.115.016424.

Eliason

Wainess

Proctor

, et al. A national and single institutional experience in the contemporary treatment of acute lower extremity ischemia. Ann Surg 2003; 238(3): 382–390. DOI: 10.1097/01.sla.0000086663.49670.d1.

Björck

Earnshaw

Acosta

, et al. Editor's choice - European Society for Vascular Surgery (ESVS) 2020 clinical practice guidelines on the management of acute limb ischaemia. Eur J Vasc Endovascular Surg 2020; 59(2): 173–218. DOI: 10.1016/j.ejvs.2019.09.006.

Hogan

Klein

Richards

. The impact of anaemia and intravenous iron replacement therapy on outcomes in cardiac surgery. Eur J Cardio-Thoracic Surg 2015; 47(2): 218–226. DOI: 10.1093/ejcts/ezu200.

Hung

Ortmann

Besser

, et al. A prospective observational cohort study to identify the causes of anaemia and association with outcome in cardiac surgical patients. Heart 2015; 101(2): 107–112. DOI: 10.1136/heartjnl-2014-305856.

Dunkelgrun

Hoeks

Welten

GMJM.

, et al. Anemia as an independent predictor of perioperative and long-term cardiovascular outcome in patients scheduled for elective vascular surgery. Am J Cardiol 2008; 101(8): 1196–1200. DOI: 10.1016/j.amjcard.2007.11.072.

Baron

Hochrieser

Posch

, et al. Preoperative anaemia is associated with poor clinical outcome in non-cardiac surgery patients. Br J Anaesth 2014; 113(3): 416–423. DOI: 10.1093/bja/aeu098.

Nandhra

Boylan

Prentis

, et al. The influence of preoperative anemia on clinical outcomes after infrainguinal bypass surgery. Ann Vasc Surg 2020; 66: 586–594. DOI: 10.1016/j.avsg.2019.11.043.

10.

Velescu

Clará

Cladellas

, et al. Anemia increases mortality after open or endovascular treatment in patients with critical limb ischemia: a retrospective analysis. Eur J Vasc Endovascular Surg 2016; 51(4): 543–549. DOI: 10.1016/j.ejvs.2015.12.006.

11.

Desormais

Aboyans

Bura

, et al. Anemia, an independent predictive factor for amputation and mortality in patients hospitalized for peripheral artery disease. Eur J Vasc Endovascular Surg 2014; 48(2): 202–207. DOI: 10.1016/j.ejvs.2014.04.005.

12.

NHS England . CCG Indicator Specifications for 2020-2021. https://www.england.nhs.uk/wp-content/uploads/2020/01/FINAL-CQUIN-20-21-Indicator-Specifications-190220.pdf (published 2020).

13.

Wilson

Dekker

Bruns

, et al. Short-term effect of preoperative intravenous iron therapy in colorectal cancer patients with anemia: results of a cohort study. Transfusion 2018; 58(3): 795–803. DOI: 10.1111/trf.14456.

14.

Klein

Chau

Yeates

, et al. Preoperative intravenous iron before cardiac surgery: a prospective multicentre feasibility study. Br J Anaesth 2020; 124(3):243–250. DOI: 10.1016/j.bja.2019.11.023.

15.

Spahn

Schoenrath

Spahn

, et al. Effect of ultra-short-term treatment of patients with iron deficiency or anaemia undergoing cardiac surgery: a prospective randomised trial. The Lancet 2019; 393(10187): 2201–2212. DOI: 10.1016/S0140-6736(18)32555-8.

16.

Shin

Park

Kim

, et al. Efficacy of perioperative intravenous iron therapy for transfusion in orthopedic surgery: a systematic review and meta-analysis. PLoS One 2019; 14(5): e0215427. DOI: 10.1371/journal.pone.0215427.

17.

Obi

Park

Bove

, et al. The association of perioperative transfusion with 30-day morbidity and mortality in patients undergoing major vascular surgery. J Vasc Surg 2015; 61(4): 1000–1009.e1. DOI: 10.1016/j.jvs.2014.10.106.

18.

Bursi

Barbieri

Politi

, et al. Perioperative red blood cell transfusion and outcome in stable patients after elective major vascular surgery. Eur J Vasc Endovascular Surg 2009; 37(3): 311–318. DOI: 10.1016/j.ejvs.2008.12.002.

19.

O'Keeffe

SD.

Davenport

Minion

, et al. Blood transfusion is associated with increased morbidity and mortality after lower extremity revascularization. J Vasc Surg. 2010; 51(3): 616–621.e3. DOI: 10.1016/j.jvs.2009.10.045.

20.

Bodewes

TCF

Pothof

Darling

, et al. Preoperative anemia associated with adverse outcomes after infrainguinal bypass surgery in patients with chronic limb-threatening ischemia. J Vasc Surg 2017; 66(6): 1775–1785.e2. DOI: 10.1016/j.jvs.2017.05.103.

21.

Padmanabhan

Siau

Curtis

, et al. Preoperative anemia and outcomes in cardiovascular surgery: systematic review and meta-analysis. Ann Thorac Surg 2019; 108: 1840–1848. DOI: 10.1016/j.athoracsur.2019.04.108.

22.

World Health Organization . Haemoglobin concentrations for the diagnosis of anaemia and assessment of severity. https://www.who.int/vmnis/indicators/haemoglobin.pdf (accessed 24 March 2016).

23.

Thygesen

Alpert

Jaffe

, et al. Fourth universal definition of myocardial infarction (2018). Revista Espanola de Cardiologia (English ed.) 2019; 72(3): 72–269. DOI: 10.1093/eurheartj/ehy462.

24.

Matsushita

Mahmoodi

Woodward

, et al. Comparison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerular filtration rate. JAMA 2012; 307(18): 1941–1951. DOI: 10.1001/jama.2012.3954.

25.

The Renal Association . CKD stages. https://renal.org/information-resources/the-uk-eckd-guide/ckd-stages/ (published 2018).

26.

NICE CKS . Acute kidney injury. https://cks.nice.org.uk/acute-kidney-injury#!scenario (published 2018).

27.

von Elm

Altman

Egger

, et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies. J Clin Epidemiol 2008; 61(4): 344–349. DOI: 10.1016/j.jclinepi.2007.11.008.

28.

Therneau

Lumley

. Survival: survival analysis. https://CRAN.R-project.org/package=survival (2019, accessed 15 August 2019).

29.

Kassambara

Kosinski

Biecek

, et al.

Survminer: drawing survival curves using “Ggplot2.”

https://CRAN.R-project.org/package=survminer (2019, accessed 15 August 2019).

30.

Wasey

Denney

. Physiology: calculate physiologic characteristics of awake and anesthetized adults, children and infants. https://CRAN.R-project.org/package=physiology (2018, accessed 15 August 2019).

31.

Kennedy

Forestmodel: forest plots from regression models. https://CRAN.R-project.org/package=forestmodel (2020, accessed 7 February 2021).

32.

Robin

Turck

Hainard

, et al. pROC: an open-source package for R and S+ to analyze and compare ROC curves. BMC Bioinformatics 2011; 12(1): 77. DOI: 10.1186/1471-2105-12-77.

33.

Youden

. Index for rating diagnostic tests. Cancer 1950; 3(1): 32–35. DOI: 10.1002/1097-0142(1950)3:1<32::aid-cncr2820030106>3.0.co;2-3.

34.

Back

Leo

Cuthbertson

, et al. Long-term survival after vascular surgery: specific influence of cardiac factors and implications for preoperative evaluation. J Vasc Surg 2004; 40(4): 752–760. DOI: 10.1016/j.jvs.2004.07.038.

35.

Franchini

Marano

Veropalumbo

, et al. Patient Blood Management: a revolutionary approach to transfusion medicine. Blood Transfus 2019; 17(3): 191–195. DOI: 10.2450/2019.0109-19.

36.

Althoff

Neb

Herrmann

, et al. Multimodal patient blood management program based on a three-pillar strategy. Ann Surg 2019; 269(5): 794–804. DOI: 10.1097/SLA.0000000000003095.

37.

Delaforce

Duff

Munday

, et al. Overcoming barriers to evidence-based patient blood management: a restricted review. Implementation Sci 2020; 15(1): 6. DOI: 10.1186/s13012-020-0965-4.