Secondary Logo

Journal Logo

Perioperative medicine

The efficacy of pre-operative preparation with intravenous iron and/or erythropoietin in anaemic patients undergoing orthopaedic surgery

An observational study

Heschl, Martina; Gombotz, Hans; Haslinger-Eisterer, Bettina; Hofmann, Axel; Böhler, Nikolaus; Meier, Jens

Author Information
European Journal of Anaesthesiology: April 2018 - Volume 35 - Issue 4 - p 289-297
doi: 10.1097/EJA.0000000000000752
  • Free

Abstract

Introduction

Pre-operative anaemia is common among patients undergoing elective orthopaedic surgery. In total hip or total knee replacement (THR, TKR), around 15 to 20% of patients have a haemoglobin (Hb) concentration below WHO thresholds,1,2 and as a consequence, these patients require three to four times more red blood cell (RBC) transfusions during their hospital stay than patients without anaemia.3 Because it has been demonstrated that pre-operative anaemia and peri-operative transfusions correlate with an increase in morbidity and mortality,4–8 a bundle of measures called ‘patient blood management’ (PBM) was introduced some years ago, with the goal to avoid both pre-operative anaemia and peri-operative transfusions.9 PBM is an evidence-based, multidisciplinary, multimodal approach which aims to increase the patient's own RBC mass, reduce peri-operative blood losses and increase tolerance of physiological anaemia (‘three pillar model’ of PBM) with the ultimate goal of improving outcome.

As shown in the literature, the efficacy of different PBM measures varies.10–14 Although some of these measures have become common clinical standards in many hospitals (e.g. cell salvage, restrictive transfusion protocols etc.), others are rarely implemented due to organisational challenges and debatable efficacy. One of the latter is the optimisation of the patient's own RBC mass by pre-operative administration of iron and/or erythropoietin (EPO). Ideally, patients should be examined 4 weeks before surgery, allowing for sufficient time to treat pre-existing anaemia or iron deficiency, which might lead to anaemia later. However, this requires adjustments in patient logistics and workflows. In daily clinical practice, it is often argued that these efforts outweigh the potential positive effect of this measure, and that only little is known about pre-operative administration of iron and/or EPO in terms of long-term mortality.

The sole efficacy of the pre-operative optimisation of RBC mass with iron and/or EPO in combination with postoperative long-term survival analysis has not been studied thoroughly.15 Most studies either included a relatively small number of patients or combined several PBM measures, precluding a differentiated analysis of the effect of pre-operative iron and/or EPO substitution.16 Furthermore, no study exists that analyses the net effect of iron/EPO on postoperative long-term mortality (>1 year) in a clinical setting.

We therefore performed a propensity score-matching analysis of real-world data with the aim of segregating the effect of iron/EPO from other PBM measures on peri-operative transfusion, Hb concentration on discharge and mortality in patients undergoing THR or TKR. Furthermore, the influences on surgery time, in-hospital stay, postoperative survival and Hb concentration at discharge were evaluated.

Methods

There was no need for informed consent, because retrospective data analysis of our protocol-based PBM concept was performed.

This was a single-centre, retrospective study with two cohorts [pre-operative preparation (PREP) with iron and/or EPO vs. no preparation (NOP)]. After approval of our local ethics committee (study number K-64-15), data of patients who received THR or TKR between 1 January 2008 and 30 September 2014 were included. Implementation of PBM started at our institution in 2007. Data including survival were obtained from the patients’ electronic health records. Exclusion criteria were re-operation, missing data on pre-operative blood analysis and pre-operatively and peri-operatively administered RBCs.

In accordance with existing guidelines, our PBM programme is determined by several standard operating procedures (SOPs) and includes a detailed pre-operative patient assessment with a structured interview for bleeding disorders, PREP with iron and/or EPO if needed (see below), intra-operative measures to reduce peri-operative blood losses [e.g. meticulous surgical technique, application of tranexamic acid (20 mg kg−1 body weight for THR and TKR), use of cell saver, massive bleeding protocol] and application of a restrictive transfusion strategy (Hb concentration 7 g dl−1 in healthy patients, 8 g dl−1 in patients with cardiopulmonary restrictions).2,17 Adherence to all PBM measures (PREP, transfusion triggers, coagulation management etc.) was not investigated, but physicians are urged to follow the PBM SOPs thoroughly in every patient at our institution.

According to our PBM SOPs, pre-operative measures of PBM include a complete blood count, iron status (i.e. serum iron, serum ferritin concentration and transferrin saturation) and C-reactive protein (CRP) for all patients in the pre-operative clinic. WHO criteria of 1968 were used to define anaemia (<13 g dl−1 for men, <12 g dl−1 for women). The type of anaemia was determined by analysing serum iron, serum ferritin, transferrin saturation and CRP. Our algorithm is depicted in Fig. 1. As this algorithm has been used since 2007, it is not in complete accordance with algorithms that have been published recently. The distinction between patients with a serum ferritin concentration less than 30 and 30 to 100 μg l−1 is somewhat unnecessary regarding the resulting therapy, but we advised every patient with a serum ferritin concentration less than 30 μg l−1 to have a general anaemia check-up. This algorithm was used throughout the whole study period.

Fig. 1
Fig. 1:
Treatment algorithm for patients with pre-operative anaemia. CRP, C-reactive protein; EPO, erythropoietin; RBC, red blood cell; SF, serum ferritin; TSA, transferrin saturation.

If anaemia could not be classified by the algorithm described, a haematologist was consulted. The expected blood loss was estimated through the Mercuriali algorithm.18 If this algorithm predicted a packed RBC (pRBC) requirement of more than one unit of pRBC for the planned surgical procedure, the patient was assigned to our pre-operative anaemia correction programme, which includes the administration of ferric carboxymaltose (Ferinject®; Vifor Pharma, Bern, Switzerland) and EPO (Retacrit®, Pfizer, Berlin, Germany) according to our SOP. Patients who received iron and or/EPO were assigned to the PREP group retrospectively. Patients who did not receive iron and/or EPO were assigned to the group without pre-operative optimisation of the red cell mass (NOP group). All other measures of PBM (use of cell salvage, transfusion triggers, coagulation management etc.), except PREP with iron and/or EPO, were similar in both groups according to our SOPs.

Long-term survival follow-up was performed by analysis of social insurance data that had been obtained from the electronic records of patients. As a consequence, a complete follow-up could be performed in all patients.

Statistics

For the comparison of PREP and NOP, a bias-reduced subset of the full data set (n=4352) was generated by means of propensity score-matching. The matching variables were determined by multiple regression analyses and Cox regression analysis. Dependent variables were the number of pRBC units administered, Hb on discharge and survival time (these three variables were termed primary endpoints). In addition to PREP, the following factors with suspected influence on the prior-ranking endpoints were used as independent variables: age, sex, weight, BMI, American Society of Anesthesiologists (ASA) physical status classification, New York Heart Association (NYHA) classification, pre-existing hypertension, pre-existing diabetes, pre-existing coronary artery disease, type of surgery and Hb at the pre-operative assessment visit. Table 1 shows the analyses results. According to the influence on the prior-ranking endpoints suggested by P values less than 0.05, age, sex, BMI, ASA classification, NYHA classification, pre-existing hypertension, pre-existing coronary artery disease, type of surgery and Hb at the pre-operative assessment visit were used for propensity score-matching (matching variables).

Table 1
Table 1:
Selection of matching variables (n = 4352)

As the cohorts PREP and NOP were markedly different (allocation of patients by PBM-related algorithms and neither by chance nor by physician's preference), the quality of matching was limited in general. The use of a maximum allowable score difference of 0.10 between paired patients resulted in 331 pairs who fulfilled this criterion. If an absolute standardised score difference of 0.10 would have been chosen as a threshold, the number of suitable pairs would have been reduced to only 46. Thus, it was accepted that matching variables were conspicuous in the cohort comparison (P < 0.05).

Before cohort comparisons, all data of continuous variables were checked for normal distribution (test of normality: Kolmogorov–Smirnov with Lilliefors significance correction, type I error 10%). Normally distributed data were compared by the t test (test for variance homogeneity: Levene test, type I error 5%) for independent samples; metric variables without normally distributed data and variables measured on ordinal scales were analysed by the exact Mann–Whitney U test, categorical variables by exact χ2 or Fisher's exact test. Survival time, depicted by Kaplan–Meier plot, was compared between PREP and NOP by the log-rank test.

Age, BMI, ASA classification, NYHA classification and Hb at the pre-operative visit had P values less than 0.05 and were therefore used as covariates for the further analyses of the primary endpoints.

In addition, the ranking of the operations according to their date was used as a covariate. This variable was not adequate for propensity score-matching due to the extremely unequal distribution of PREP and NOP allocation over the included years. For minimisation of bias, the covariates mentioned above were fed into the cohort comparisons of number of pRBC units administered and Hb on discharge by a nonparametric analysis of covariance (rank analysis of covariance according to Quade). For the survival time, an analogous investigation by a Cox regression approach (survival time as dependent variable; covariates mentioned above and PREP as independent variables) was performed.

The type I error was not adjusted for multiple testing. Therefore, the results of inferential statistics are descriptive only.

Statistical analysis was performed using the open-source R statistical software package, version 3.3.2 (The R Foundation for Statistical Computing, Vienna, Austria).

Results

Between 1 January 2008 and 30 September 2014, a total of 5518 patients underwent THR or TKR surgery at our institution, of whom 4352 were included due to the inclusion/exclusion criteria (Fig. 2). A total of 373 of these patients were treated with iron and/or EPO, and 3979 patients were not. In all, 30.4% of patients treated with iron and/or EPO were stratified as ASA physical status 3, whereas only 19.2% of the others were ASA 3. In both groups, about 50% of patients underwent THR and about 50% underwent TKR. Before propensity score-matching, patients who were not treated with iron and/or EPO had a higher BMI (28.9 ± 5.0 vs. 27.4 ± 4.8 kg m−2, P < 0.001), a higher Hb (14.2 ± 1.2 vs. 12.0 ± 1.1 g dl−1, P < 0.001), a higher serum ferritin concentration (228 ± 215 vs. 152 ± 175 μg l−1, P < 0.001) and a higher transferrin saturation (27 ± 9.6 vs. 20 ± 10%, P < 0.001) at the pre-operative assessment visit than patients who were treated with iron and/or EPO. If PREP was performed, 1244 ± 877 mg iron and 83 000 ± 37 000 IU EPO were administered starting 11.2 ± 8.9 days before surgery in 2.5 ± 1.1 sessions. Before propensity score-matching, the transfusion rate was 9.7% in patients without PREP and 12.3% in patients who received PREP. The total transfusion rate for both groups was 9.9%. The Hb concentration at discharge was higher in patients who received iron and/or EPO (10.9 ± 1.2 vs. 10.7 ± 1.3 g dl−1, P < 0.05), although this difference seems clinically negligible.

Fig. 2
Fig. 2:
Flow chart of inclusion and exclusion criteria and propensity score-matching. NOP, no preparation; PREP, pre-operative preparation; THR, total hip replacement; TKR, total knee replacement.

An overview of the bias-reduced data set after propensity score-matching is shown in Tables 2 and 3. After propensity score-matching in the NOP group and the PREP group, there were 331 patients per group.

Table 2
Table 2:
Demographic and other baseline data for bias-reduced cohorts after propensity score-matching
Table 3
Table 3:
Transfusion data for bias-reduced cohorts after propensity score-matching

After propensity score-matching in both groups, about 50% of patients in each group had undergone THR and TKR (not significant [n.s.]). Patients in the NOP group were comparable with patients in the PREP group in most of the parameters investigated. However, some of the parameters differed despite adequate propensity score-matching. Patients in the PREP group were stratified in higher ASA classifications than patients in the NOP group (P < 0.05), were stratified in higher NYHA classifications (P < 0.05), had a lower BMI (28.7 ± 5.3 vs. 27.4 ± 4.9 kg m−2, P < 0.05) and had a lower Hb concentration at the pre-operative visit (12.9 ± 1.7 vs. 12.2 ± 1.1 g dl−1, P < 0.05). Understandably, patients in the NOP group had a higher serum ferritin concentration (197 ± 14 vs. 152 ± 181 μg l−1, P < 0.001) and a higher transferrin saturation (20 ± 10 vs. 19 ± 8%, P < 0.001) at the pre-operative assessment visit than patients in the PREP group. Three-quarters of the PREP group were treated with intravenous iron and subcutaneous EPO, whereas 25% were treated solely with intravenous iron. In the patients treated with EPO, 28.6% received one dose, 41.8% received two doses and 29.6% received three or more doses of EPO.

The most important of the three primary endpoints of the study, the mean transfusion rate, was higher in the NOP group than in the PREP group (0.5 ± 1.3 vs. 0.2 ± 0.8 U, P < 0.001, for all patients) corresponding to 24% of the patients in the NOP group who received at least one unit of pRBC and 12% in the PREP group, whereas the other primary endpoint, the Hb concentration at discharge, was comparable between groups (10.6 ± 1.1 vs. 10.7 ± 1.3 g dl−1, n.s.) (Fig. 3). There were no differences in survival times between the two study groups (Fig. 4). There was no significant difference in the transfusion rate between patients who received iron alone or iron and EPO (0.36 ± 0.75 vs. 0.19 ± 0.84, n.s.).

Fig. 3
Fig. 3:
Confidence intervals of the mean differences of the two main outcome parameters. Hb, haemoglobin concentration at discharge; pRBC, packed red blood cell (unit).
Fig. 4
Fig. 4:
Long-term survival rates for both study groups. There was no difference between groups. The smaller picture depicts the same numbers on a different scale. NOP, no preparation; PREP, pre-operative preparation.

Performing an additional covariate analysis and a regression analysis, it was demonstrated that after adjustment for age, ASA, BMI, Hb at the pre-operative assessment visit, NYHA classification and date of surgery as covariates, the differences in the number of pRBC units were still present between the two groups (P < 0.001). An analogous covariance-analytical approach still yielded no substantial cohort difference for the haematocrit at discharge (P = 0.39). Furthermore, the implementation of the covariates mentioned into the survival analysis confirmed the results of the unadjusted comparison (regression analysis approach: no substantial influence of iron and/or EPO on survival time, P = 0.42).

Discussion

Consistent with recent literature, we conclude that patients benefit from pre-operative Hb optimisation in terms of a reduced exposure to pRBC transfusions and most likely reduced concomitant side-effects of this treatment. As, after propensity score-matching, long-term mortality was similar in both study groups, we also conclude that PREP with iron and/or EPO can be considered well tolerated, when compared with patients in whom this measure was omitted.

Around the world, PBM is increasingly recognised as an evidence-based, new standard of care. Numerous studies have demonstrated that PBM has the potential to avoid anaemia as well as transfusion.19 Taking these effects together, PBM might be judged as one of the cornerstones of patient safety in the peri-operative setting. A large body of recently published literature explains and analyses the various PBM measures in detail.20,21 Some appear to be easily implemented, such as restrictive transfusion policies or single-unit policies, whereas the implementation of others seems to be cumbersome. However, only the combined application of PBM measures is thought to maximise the effect of improving outcome.22 To date, we are not aware of any other study that analyses the combined use of iron and/or EPO in a real-life setting and additionally describes safety of this approach by analysis of long-term mortality.

Pre-operative optimisation of erythropoiesis with iron and/or EPO has been demonstrated to be very effective in increasing pre-operative Hb concentration, thus reducing the amount of peri-operative transfusion.23–26 However, although the evidence for these PBM measures in an experimental setting are sufficient, observational studies of treatment with iron and/or EPO with long-term follow-up, including a sufficient number of participants, are still missing. We therefore retrospectively analysed the data from our clinic with the aim of describing the efficacy of PREP independently of other measures of PBM.

As patients in the PREP group required only half the amount of transfusion compared with the NOP group, it may therefore be concluded that pre-operative administration of iron and/or EPO is highly effective in reducing peri-operative blood utilisation. Furthermore, Hb levels at discharge were comparable between the PREP and the NOP groups, suggesting similar transfusion thresholds for both groups. Therefore, one can speculate that pre-operative administration of iron and/or EPO results not only in comparable Hb concentrations after surgery, but had a beneficial long-term effect on erythropoiesis.

The aim of our study was to investigate the effect of PREP of patients with iron and/or EPO on peri-operative transfusion needs. However, as PBM is a bundle of measures, other methods might have influenced our results. For example, a restrictive transfusion regimen has a major influence on peri-operative transfusion needs. At our institution, a restrictive transfusion regimen has been implemented since 2007 and was not changed throughout the study period. Another measure that has been demonstrated to be very effective in reducing peri-operative transfusion needs is cell salvage. At our institution, cell salvage was used uniformly during the study period.

One more important confounder for peri-operative transfusion needs is the composition of personnel in the medical team, especially the surgical team. It is well known that, depending on the surgeons, peri-operative bleeding and peri-operative transfusion needs might differ significantly between groups. Due to regulatory limitations, we were not allowed to analyse the role of the medical team in our setting. Therefore, it is difficult to judge whether the outcome of our study has been influenced by the physicians and nurses involved.

As the data were analysed retrospectively, our results are prone to selection bias. We therefore investigated the inclusion rates for every year of our study period in our statistical analysis. Considerably fewer patients were included at the beginning of the study period in the PREP group than in the NOP group (Fig. 5), and therefore, we cannot exclude the possibility that a certain ‘training effect’ of our team might have influenced the results. This holds true also for surgeons and the surgical approach, although the surgical technique did not change throughout the observational time frame. However, with our retrospective approach, we cannot exclude the possibility that the skills of the surgeons involved improved, and thus influenced the results. Because the actual differences are quite low, the impact of this effect is difficult to judge, especially after the propensity score-matching procedure. This holds even more true because an additional covariate analysis and regression analysis demonstrated that, after adjustment for age, ASA physical status classification, BMI, Hb at the pre-operative assessment visit, NYHA classification and date of surgery as covariates, the differences in the number of pRBC units were still present between the two groups However, this inclusion bias might be the most important confounder of our model.

Fig. 5
Fig. 5:
Inclusion rates during the study period for both groups with a slightly higher proportion of patients in the no preparation group during the earlier years of the study period. NOP, no preparation; PREP, pre-operative preparation.

It has to be pointed out that iron therapy in our setting was performed intravenously, which increases treatment efficacy,27,28 but might also carry additional risks. Some authors have demonstrated that pre-operative administration of oral iron might be effective in avoiding peri-operative transfusions in some situations, but the usual time frame of 28 days from patient enrolment to surgery might be too short for oral iron therapy to be effective. Black box warnings for intravenous iron were issued, but the new generation of iron preparations such as carboxymaltose did not trigger any allergic reaction during the whole inclusion period, and patients were monitored for at least 30 min after iron infusion.

Although the administration of EPO is known to be associated with an increased risk of peri-operative ischaemic and thrombo-embolic episodes,29 we did not collect data for this outcome due to the retrospective nature of our study. So far, Stowell et al.29 have published the only randomised controlled trial that evaluated the thrombo-embolic risk associated with EPO therapy as a primary end-point. Adequately powered to demonstrate noninferiority, the authors reported a higher deep vein thrombosis rate associated with pre-operative EPO therapy when compared with standard blood conservation, although in that trial, no pharmacological thromboprophylaxis was allowed. An older article published by de Andrade et al.30 did not demonstrate this effect when adequate pharmacological thromboprophylaxis was provided, and it remains unclear whether EPO represents a real additional thrombotic risk factor in this current era of widespread thromboprophylaxis. Our mortality data suggest that the additional risks induced by the administration of EPO might be negligible, although our study was not powered to detect such differences. Some might judge the EPO doses used too low for our setting. For example, Rosencher et al.31 proposed an algorithm that repeats the administration of EPO until a haematocrit of more than 40% is reached. Whether our rather high transfusion rate could have been reduced by this measure remains open for discussion.

Although we used propensity score-matching to adjust for differences in patients in the NOP and the PREP groups, there remained some minor differences. For example, the mean Hb concentration in the PREP group was slightly lower than in the NOP group. However, both seem to be rather high; patients with very low Hb concentrations at the pre-operative assessment clinic probably received PREP anyway, and as a consequence, the propensity score-matching algorithm was forced to match for patients with rather high Hb concentrations. However, as our transfusion rate was around 10% for both groups, it has to be admitted that even at these high Hb concentrations there is still room for improvement. Whether our results would have even been more pronounced in patients with more profound anaemia remains speculative. In the PREP group, patients were older and had a higher ASA score. Therefore, a higher mortality could be expected irrespective of the treatment. Generally speaking, if parameters were different between the groups, it was always in favour of the NOP group. Nonetheless, we did not observe any differences in mortality between groups, but a reduction of peri-operative transfusions in the PREP group. Whether the higher patient age in the PREP group influenced survival time in such a way that natural life expectancy became a significant confounder in this group was not analysed in our study.

The rather high Hb concentration after the matching procedure in both groups yields another weakness of our study. Taking this fact into account, it has to be stated that we cannot conclude from our data whether patients with higher Hb values would profit more from PREP. However, even with these rather high Hb concentrations, we could demonstrate superiority of PREP with iron and/or EPO. A randomised controlled trial with lower Hb concentrations in both groups seems difficult to justify, because in the control group, nontreatment of pre-operative anaemia is substandard care and therefore probably unethical.32

Patients in our study were treated according to a predefined algorithm. This algorithm was not changed throughout the study period. However, the reason why some of the patients with lower Hb concentrations were prepared pre-operatively whereas other patients with similar Hb levels and other similar properties were not can only be explained by variations in daily clinical practice. It is difficult to judge whether this difference is responsible for some bias which we could not exclude by our statistical approach and indicates a shortcoming of our study.

Three-quarters of patients in the PREP group were treated with intravenous iron and subcutaneous EPO, whereas 25% were treated solely with intravenous iron. This indicates that our preparation algorithm had not been followed thoroughly by all physicians. However, this finding cannot be used to extrapolate adherence to other PBM measures. According to our PREP protocol, no patients received EPO alone, as it has been demonstrated that the application of EPO is effective only if depleted iron stores of the body are filled simultaneously.33,34 We have to point out that pre-operative iron deficiency anaemia is the most common cause of anaemia in our patient population, whereas anaemia of chronic inflammation with or without iron deficiency is less frequent.34 Nevertheless, a rather high number of patients received EPO. Whether PREP is also effective in patients predominantly suffering from anaemia of chronic disease or anaemia of chronic infection cannot be concluded from our data.

Conclusion

In elective knee and hip surgery, algorithm-driven administration of iron and/or EPO results in a 50% reduction of intra-operative blood transfusions without any change in long-term survival. The potential of this measure is most probably additive to other PBM measures. These results clearly demonstrate that the first pillar of PBM is an effective tool that should not be omitted in the daily clinical setting due to the efforts of its implementation.

Acknowledgements relating to this article

Assistance with the study: we would like to thank Wolfgang Schimetta for helping in this research project through statistical support. We are grateful to all those involved in data collection.

Financial support and sponsorship: only departmental resources were used.

Conflicts of interest: none.

Presentation: none.

References

1. Gombotz H, Rehak PH, Shander A, et al. Blood use in elective surgery: the Austrian benchmark study. Transfusion 2007; 47:1468–1480.
2. Goodnough LT, Maniatis A, Earnshaw P, et al. Detection, evaluation and management of preoperative anaemia in the elective orthopaedic surgical patient: NATA guidelines. Br J Anaesth 2011; 106:13–22.
3. Jans Ø, Jørgensen C, Kehlet H, et al. Lundbeck Foundation Centre for Fast-track Hip and Knee Replacement Collaborative Group. Role of preoperative anemia for risk of transfusion and postoperative morbidity in fast-track hip and knee arthroplasty. Transfusion 2014; 54:717–726.
4. Baron DM, Hochrieser H, Posch M, et al. Preoperative anaemia is associated with poor clinical outcome in noncardiac surgery patients. Br J Anaesth 2014; 113:416–423.
5. Carson JL, Duff A, Poses RM, et al. Effect of anaemia and cardiovascular disease on surgical mortality and morbidity. Lancet 1996; 348:1055–1060.
6. Wu W-C, Schifftner TL, Henderson WG, et al. Preoperative hematocrit levels and postoperative outcomes in older patients undergoing noncardiac surgery. JAMA 2007; 297:2481–2488.
7. Shander A, Javidroozi M, Gianatiempo C, et al. Outcomes of protocol-driven care of critically ill severely anemic patients for whom blood transfusion is not an option. Crit Care Med 2016; 44:1109–1115.
8. Musallam KM, Tamim HM, Richards T, et al. Preoperative anaemia and postoperative outcomes in noncardiac surgery: a retrospective cohort study. Lancet 2011; 378:1396–1407.
9. Shander A, Isbister J, Gombotz H. Patient blood management: the global view. Transfusion 2016; 56 (Suppl 1):S94–S102.
10. Meybohm P, Herrmann E, Steinbicker AU, et al. Patient blood management is associated with a substantial reduction of red blood cell utilization and safe for patient's outcome: a prospective, multicenter cohort study with a noninferiority design. Ann Surg 2016; 264:203–211.
11. Rineau E, Chaudet A, Chassier C, et al. Implementing a blood management protocol during the entire perioperative period allows a reduction in transfusion rate in major orthopedic surgery: a before-after study. Transfusion 2016; 56:673–681.
12. Spahn DR, Goodnough LT. Alternatives to blood transfusion. Lancet 2013; 381:1855–1865.
13. Auerbach M, Goodnough LT, Picard D, et al. The role of intravenous iron in anemia management and transfusion avoidance. Transfusion 2008; 48:988–1000.
14. Theusinger OM, Leyvraz P-F, Schanz U, et al. Treatment of iron deficiency anemia in orthopedic surgery with intravenous iron: efficacy and limits: a prospective study. Anesthesiology 2007; 107:923–927.
15. Shander A, Spence RK, Auerbach M. Can intravenous iron therapy meet the unmet needs created by the new restrictions on erythropoietic stimulating agents? Transfusion 2010; 50:719–732.
16. Mercuriali F, Zanella A, Barosi G, et al. Use of erythropoietin to increase the volume of autologous blood donated by orthopedic patients. Transfusion 1993; 33:55–60.
17. Kozek-Langenecker SA, Ahmed AB, Afshari A, et al. Management of severe perioperative bleeding: guidelines from the European Society of Anaesthesiology: first update 2016. Eur J Anaesthesiol 2017; 34:332–395.
18. Mercuriali F, Inghilleri G. Proposal of an algorithm to help the choice of the best transfusion strategy. Curr Med Res Opin 1996; 13:465–478.
19. Muñoz M, Gómez-Ramírez S, Kozek-Langeneker S, et al. ‘Fit to fly’: overcoming barriers to preoperative haemoglobin optimization in surgical patients. Br J Anaesth 2015; 115:15–24.
20. Spahn DR, Zacharowski K. Nontreatment of preoperative anaemia is substandard clinical practice. Br J Anaesth 2015; 115:1–3.
21. Goodnough LT, Shander A. Patient blood management. Anesthesiology 2012; 116:1367–1376.
22. Spahn DR. Patient blood management: success and potential in the future. Ann Surg 2016; 264:212–213.
23. Leal-Noval SR, Muñoz M, Asuero M, et al. Spanish Consensus Statement on alternatives to allogeneic blood transfusion: the 2013 update of the ‘Seville Document’. Blood Transfus 2013; 11:585–610.
24. Muñoz M, Acheson AG, Auerbach M, et al. International consensus statement on the peri-operative management of anaemia and iron deficiency. Anaesthesia 2017; 72:233–247.
25. Hofmann A, Farmer S, Towler SC. Strategies to preempt and reduce the use of blood products: an Australian perspective. Curr Opin Anaesthesiol 2012; 25:66–73.
26. Isbister JP. Clinicians as gatekeepers: what is the best route to optimal blood use? Dev Biol (Basel) 2007; 127:9–14.
27. Rineau E, Chaudet A, Carlier L, et al. Ferric carboxymaltose increases epoetin-α response and prevents iron deficiency before elective orthopaedic surgery. Br J Anaesth 2014; 113:296–298.
28. Lin DM, Lin ES, Tran M-H. Efficacy and safety of erythropoietin and intravenous iron in perioperative blood management: a systematic review. Transfus Med Rev 2013; 27:221–234.
29. Stowell CP, Jones SC, Enny C, et al. An open-label, randomized, parallel-group study of perioperative epoetin alfa versus standard of care for blood conservation in major elective spinal surgery: safety analysis. Spine 2009; 34:2479–2485.
30. de Andrade JR, Frei D, Guilfoyle M. Integrated analysis of thrombotic/vascular event occurrence in epoetin alfa-treated patients undergoing major, elective orthopedic surgery. Orthopedics 1999; 22 (1 Suppl):s113–s118.
31. Rosencher N, Poisson D, Albi A, et al. Two injections of erythropoietin correct moderate anemia in most patients awaiting orthopedic surgery. Can J Anaesth 2005; 52:160–165.
32. Camaschella C, Pagani A, Nai A, et al. The mutual control of iron and erythropoiesis. Int J Lab Hematol 2016; 38 (Suppl 1):20–26.
33. Silverberg DS, Blum M, Agbaria Z, et al. The effect of i.v. iron alone or in combination with low-dose erythropoietin in the rapid correction of anemia of chronic renal failure in the predialysis period. Clin Nephrol 2001; 55:212–219.
34. Dittrich E, Schillinger M, Sunder-Plassmann G, et al. Efficacy of a low-dose intravenous iron sucrose regimen in peritoneal dialysis patients. Perit Dial Int 2002; 22:60–66.
© 2018 European Society of Anaesthesiology