Anaemia is associated with increased mortality in patients undergoing surgery. Similarly, the transfusion of red blood cells (RBC) for the treatment of anaemia is also associated with increased morbidity and mortality. The prevailing recommended strategy to reduce both preoperative anaemia and unnecessary blood transfusions has been termed ‘patient blood management’ (PBM). The implementation of such a PBM program is associated with improved patient outcomes, reduced blood product utilisation, and significant cost savings.1 One of the main pillars of PBM is a restrictive approach to blood transfusions, which is achieved by applying low haemoglobin (Hb) values as ‘transfusion thresholds’.
The recent guidelines of the American Association of Blood Banks recommend an Hb value of 7g/dl as a transfusion threshold for hospitalised adult patients who are haemodynamically stable, including critically ill patients, and a threshold of 8g/dl for patients undergoing orthopaedic or cardiac surgery, and those with pre-existing cardiovascular disease.2 In the recently published First Update 2016 of the European Society of Anaesthesiology guidelines on the management of severe perioperative bleeding, a target Hb concentration of 7–9g/dl is recommended during active bleeding.3
The recommendations for the use of the same pre-determined ‘restrictive’ Hb values as transfusion thresholds for all patients have, however, been criticised because of the limited strength of the supporting evidence, and some new data showing that in some patients a more liberal transfusion strategy may be beneficial.4,5 Even according to the guidelines themselves, the decision to transfuse an individual patient should take into account not only the Hb level, but also factors like the overall clinical context, the rate of decline in Hb, intra-vascular volume status, hypotension or tachycardia unresponsive to fluid challenge, and more.2 However, the extensive discussions about the appropriate transfusion thresholds have been dominated by the desire to minimise blood transfusions and by the need to identify patient populations in whom such restrictive policy may be detrimental,6 but have not seriously addressed the potential impact of haemodilution on the Hb level during dynamic conditions.
The fact that overzealous fluid administration may produce a low Hb in a patient with a normal red cell mass, complicating the interpretation of the Hb concentration in a dynamic clinical situation, may not be adequately recognised.7 Very often, clinicians who order blood transfusions, do not consider the potential impact of haemodilution, which may occur, for example, following the infusion of intravenous fluids in response to anaesthetic agents-induced hypotension.8
The effect of fluid administration on Hb concentration
The Hb concentration is affected by normally occurring changes in plasma volume. For example, Hb and haematocrit (Hct) have significant diurnal oscillations, being higher in the morning and lowest around midnight. Changing from supine to standing position also increases the Hct by about 10% after a relatively short while. In the healthcare setting, increasing the plasma volume by the intravenous administration of fluids may cause significant haemodilution. Haemodilution due to fluid accumulation in critically ill patients has been shown to decrease serum creatinine levels and thus lead to a delay in the diagnosis and underestimation of the severity of acute kidney injury. Similarly, the administration of excessive fluids may cause a relative, but not absolute, reduction in Hb concentration, RBC count, or Hct, which is termed ‘dilutional anaemia’ (DA). Such iatrogenic DA results in a loss of RBC-filled capillaries, leading to a reduction in oxygen-carrying capacity and the possible development of organ dysfunction.9
The decrease in the Hb concentration following fluid administration may lead to an actual decrease in oxygen delivery (DO2). In patients who received close to an additional litre of colloids as part of a perioperative goal-directed therapy (GDT) protocol, the DO2 at the end of the surgical procedure was found to be lower than that of the control group due to an associated mean Hb decrease of 0.9g/dl.10 The administration of 500ml of normal saline to critically ill patients who did not increase their cardiac output (‘non-responders’) led to a mean decrease of 8% in their Hb values with a significant decrease in their DO2.11 A similar paradoxical decrease in DO2 due to haemodilution was also observed in septic patients who did increase their cardiac output in response to a mean of about 800ml of colloids which resulted in a mean decrease in Hb of 1.6g/dl following fluid loading.12
Finally, in patients in septic shock who received large amounts of fluids as part of the original early goal directed therapy (EGDT) protocol, a decrease of 30% in Hct was uniformly observed three hours into the resuscitation, possibly explaining the very high incidence of RBC transfusions in this group of patients.13
The potential impact of haemodilution on blood transfusions
Iatrogenic haemodilution may lead to increased blood transfusion due to dilutional coagulopathy and, in turn, increased surgical bleeding.14 It is well established that differences in volume loading can markedly influence blood product requirements in both cardiac15 and hepatic16 surgery. However, and as already mentioned, fluid administration may also cause Hb levels to decrease below the acceptable transfusion threshold, causing clinicians to misinterpret the situation and order blood transfusions even though no actual significant bleeding is taking place.7,8 This phenomenon may occur especially during the use of perioperative GDT protocols that target hemodynamic goals in order to improve DO2 by using fluids with or without inotropes. Performing a systematic literature search is beyond the scope of this review. However, the results of the following studies do indicate that the administration of more fluids may be associated with more blood transfusions, and that this phenomenon may be occurring more frequently than we may think.
- In one of the largest multicentre, randomised controlled trials (RCT) on GDT in high-risk surgical patients (POM-O), the patients in the intervention group were targeted to achieve their individual preoperative DO2 value by colloids and dobutamine.17 The patients in the GDT group received nearly twice the amount of colloids compared with the control group (a mean of 2.9 vs 1.4ml/kg/h, respectively). Although the transfusion threshold was the same for both groups (Hb > 8g/dl), 22% of the patients in intervention group received blood transfusion, compared with only 11% in the control group. These results were not compared statistically and were not mentioned in the discussion of the article.17
- In another multicentre RCT on GDT in high-risk surgical patients (POEMAS), the intervention group received colloids and dobutamine to achieve a minimal cardiac index (CI) of 2.5l/min/m2.18 The patients in the GDT group received significantly more colloid boluses and significantly more packed PRBC units, compared to the standard therapy group (0.6 ± 1.3 vs 0.2 ± 0.6; p=0.019).18
- In another RCT, where patients were randomised to have their stroke volume (SV) maximised by fluids during major colorectal surgery, the intervention group received an additional mean colloid volume of 1360 ± 446ml. Compared with the control group, the GDT group had significantly higher blood loss (500 [200–1000] vs 250 [100–500], p<0.006) and received significantly more packed RBC (19 vs 8, p<0.03).19
- In another RCT of SV optimisation during elective major abdominal surgery, the GDT group received an additional 956ml of colloids compared with the control group.10 By the end of the operation, the DO2 of the GDT group was significantly lower than that of the control group due to a mean decrease in Hb of 0.9g/dl. SV optimisation was also associated with a trend to increased mean intraoperative blood loss (513 vs 396ml, p=0.090), although the proportion of participants receiving a RBC transfusion during surgery did not differ between groups.
- Sensitivity analysis of postoperative Hb concentration carried out after removal of seven patients that had major bleeding (five in the SV optimisation vs two in the control), the SV optimisation group still had statistically signiﬁcantly lower Hb, suggesting, according to the authors themselves, haemodilution as a mechanism.10
- In another prospective study comparing patients before and after the adoption of a GDT protocol based on pulse pressure variation (PPV), the PPV-guided protocol was associated with less fluid administration, significantly higher Hb values at 8 hours after surgery, fewer blood transfusions and decreased morbidity.20
- In another GDT study, done in patients undergoing total hip arthroplasty, high values of CI and DO2 were targeted by the administration of colloids, crystalloids and dobutamine.21 The patients in the intervention group received significantly more fluids than the control group (6032 ± 1388 vs 2635 ± 346ml, p<0.0001), received more blood intraoperatively (595 ± 316 vs 0ml, p<0.00031), and their blood loss was higher though not statistically significant (1156 ± 679 vs 853 ± 173ml). Overall blood utilisation was the same in both groups, as the control group received more blood in the postoperative period.21
- Similar results were found in another multicentre RCT, performed in high-risk patients scheduled for major abdominal surgery, where the overall amount of RBC transfusion was similar in the control and GDT groups, although the latter received all transfusion earlier during the operation while the control group received their transfusions only postoperatively.22 According to the authors, this earlier transfusion, together with earlier fluid administration and dobutamine administration, may have been responsible for the reduction in organ failures and the shorter hospital stay that were observed in the intervention group.22
In summary, there is a frequent association between the administration of greater amounts of fluids and the amount of blood that is being transfused in the perioperative period. Such association is best explained by haemodilution, namely, the relative, but not absolute, reduction in Hb concentration. Recognising that the decrease in Hb is due to the development of such DA and not due to actual blood loss, may theoretically help in identifying excessive fluid administration and in preventing avoidable blood transfusions.
Continuous non-invasive monitoring of haemoglobin (SpHb) may detect the development of iatrogenic haemodilution
The measurement of Hb concentration continuously and non-invasively, termed SpHb, has become possible with the introduction of multi-wavelength technology (pulse CO-Oximetry) into a new generation of pulse oximeters (Masimo, Irvine, CA).23
The continuous monitoring of SpHb supplements intermittent laboratory Hb measurements by providing real-time visibility to any changes that may occur in the Hb concentration. So far, the recognised potential value of SpHb monitoring has been in the identification of real absolute anemia. In the recently published First Update of the European Society of Anaesthesiology guidelines on the management of severe perioperative bleeding, SpHb monitoring has been recognised as a useful trend monitor for the identification of acute changes in Hb concentration.3
This technique may be of special value when ‘restrictive’ transfusion thresholds are being employed, since, although there is some tolerance to postoperative anaemia among patients without cardiovascular disease, for each 1g/dl decrease in postoperative disease Hb concentration below 7g/dl, mortality has been shown to increase by a factor of 1.5.24 Similarly, the continuous monitoring of SpHb may be useful when repeated Hb measurements are necessary to assess the degree of acute blood loss, as recommended by the recent European guideline on management of major bleeding following trauma.25
However, by offering real-time visibility of changes in Hb levels, the continuous monitoring of SpHb may detect the development of DA in real time. The following case serves as a useful example for the unique ability of SpHb monitoring to identify in real time the development of DA due to the administration of large amounts of fluids (Figure 1). In this case of hepatic surgery, only a minimal amount of fluids was administered during the hepatic resection phase in order to decrease venous pressure and minimise bleeding, as is commonly recommended.16
The development of hypovolaemia during the resection phase is reflected by the gradual increase in the Plethysmographic Variability Index (PVI), which is a dynamic parameter of fluid responsiveness which is also measured continuously and non-invasively by the same pulse oximeter.26 While the PVI gradually increased to very excessive values (close to 40%), the SpHb remained stable at values of 11–12g/dl. At the end of the resection phase (at about 16:29 on the time scale) aggressive fluid rehydration was started.
The sudden and significant increase in blood volume led to the immediate decrease of the PVI to about 10%, a value that is consistent with low or absent fluid responsiveness. Simultaneously, the SpHb decreased from about 11g/dl to about 7.5g/dl (see broken line in Figure 1). This dramatic reduction in SpHb is a clear indication of the development of DA and may serve as warning that excessive amount of fluid has been administered.
This timely diagnosis of DA may help the clinician in the interpretation of the low Hb value that is close or even below the transfusion threshold. It may well be that getting rid of the excess fluids may be a better way to raise the Hb level than giving a blood transfusion. The case clearly demonstrates how can the decision to transfuse be affected by the amount and rate of fluid administration.
It is of interest to note that there is both experimental and clinical evidence that methemoglobin (MetHb) may also reflect the development of DA.27 The suggested physiological explanation for these observations is that DA may lead to up-regulation of perivascular nitric oxide synthase (NOS) and increase NOS-derived nitric oxide (NO), leading to local vasodilation and oxidisation of Hb to MetHb. Hence MetHb may potentially serve as a biomarker of ‘anaemic stress’, which is associated with reduced tissue perfusion during acute haemodilution. A negative correlation was also observed between the change in Hb and MetHb in patients undergoing cardiac surgery.27 Since MetHb can also be measured continuously and non-invasively with the new generation of pulse oximeters (SpMET), it will be of interest to see whether it may serve as another clinical tool for the real-time detection of DA.
Volume expansion by the administration of intravenous fluids may cause dilutional anaemia, a relative lowering of the Hb concentration and a paradoxical decrease in oxygen delivery. Such haemodilution may cause the Hb values to decrease below the acceptable transfusion threshold and thus lead to avoidable blood transfusions. This phenomenon seems to occur more frequently than initially thought, as evidenced by many studies on perioperative GDT in which the administration of more fluids was associated with the administration of more blood transfusions. Hence, continuous non-invasive monitoring of SpHb, in addition to its potential value in detecting actual blood loss, may serve as a real time indicator of the development of haemodilution, help identify fluid overload and prevent avoidable transfusions.
AP serves as a consultant and speaker for Masimo, and is a member of the Medical Advisory Board of Pulsion/Getinge.
1 Leahy MF et al. Improved outcomes and reduced costs associated with a health-system-wide patient blood management program: a retrospective observational study in four major adult tertiary-care hospitals. Transfusion 2017;Feb 02 [Epub ahead of print].
2 Carson JL et al. Clinical Practice Guidelines from the AABB: Red Blood Cell Transfusion Thresholds and Storage. JAMA 2016;316(19):2025–35.
3 Kozek-Langenecker SA et al. Management of severe perioperative bleeding: guidelines from the European Society of Anaesthesiology: First update 2016. Eur J Anaesthesiol 2017;34(6):332–95.
4 Klein HG, Flegel WA, Natanson C. Red blood cell transfusion: Precision vs imprecision medicine. JAMA 2015;314(15):1557–8.
5 Stowell C, Bennett-Guerrero E. The decision to transfuse: One size might not fit all. Crit Care Med 2017;45(5):908–10.
6 Mirski MA et al. Restrictive and liberal red cell transfusion strategies in adult patients: reconciling clinical data with best practice. Crit Care 2015;19:202.
7 Soni N. Intraoperative transfusion, triggers and precision – an oxymoron. Transfus Med 2008;18(4):209–10.
8 Shander A, Lobel GP, Javidroozi M. Anesthesia for patients with anemia. Anesthesiol Clin 2016;34(4):711–30.
9 Ince C. Hemodynamic coherence and the rationale for monitoring the microcirculation. Crit Care 2015;19 Suppl 3:S8.
10 Lai CW et al. Randomized controlled trial of stroke volume optimization during elective major abdominal surgery in patients stratified by aerobic fitness. Br J Anaesth 2015;115(4):578–89.
11 Monnet X et al. Lactate and venoarterial carbon dioxide difference/arterial-venous oxygen difference ratio, but not central venous oxygen saturation, predict increase in oxygen consumption in fluid responders. Crit Care Med 2013;41(6):1412–20.
12 Haupt MT, Gilbert EM, Carlson RW. Fluid loading increases oxygen consumption in septic patients with lactic acidosis. Am Rev Respir Dis 1985;131(6):912–16.
13 Jha V, Gutierrez G. Severe sepsis and septic shock: should blood be transfused to raise mixed venous oxygen saturation? Chest 2007;131(4):1267–8; author reply 8–9.
14 Kozek-Langenecker SA. Fluids and coagulation. Curr Opin Crit Care 2015;21(4):285–91.
15 Vretzakis G et al. Intra-operative intravenous fluid restriction reduces perioperative red blood cell transfusion in elective cardiac surgery, especially in transfusion-prone patients: a prospective, randomized controlled trial. J Cardiothorac Surg 2010;5:7.
16 Ozier Y, Tsou MY. Changing trends in transfusion practice in liver transplantation. Curr Opin Organ Transplant 2008;13(3):304–9.
17 Ackland GL et al. Individualised oxygen delivery targeted haemodynamic therapy in high-risk surgical patients: a multicentre, randomised, double-blind, controlled, mechanistic trial. Lancet Respir Med 2015;3(1):33–41.
18 Pestana D et al. Perioperative goal-directed hemodynamic optimization using noninvasive cardiac output monitoring in major abdominal surgery: a prospective, randomized, multicenter, pragmatic trial: POEMAS Study (PeriOperative goal-directed thErapy in Major Abdominal Surgery). Anesth Analg 2014;119(3):579–87.
19 Challand C et al. Randomized controlled trial of intraoperative goal-directed fluid therapy in aerobically fit and unfit patients having major colorectal surgery. Br J Anaesth 2012;108(1):53–62.
20 Benes J et al. Fluid management guided by a continuous non-invasive arterial pressure device is associated with decreased postoperative morbidity after total knee and hip replacement. BMC Anesthesiology 2015;15:148.
21 Cecconi M et al. Goal-directed haemodynamic therapy during elective total hip arthroplasty under regional anaesthesia. Crit Care 2011;15(3):R132.
22 Donati A et al. Goal-directed intraoperative therapy reduces morbidity and length of hospital stay in high-risk surgical patients. Chest 2007;132(6):1817–24.
23 Barker SJ, Shander A, Ramsay MA. Continuous noninvasive hemoglobin monitoring: A measured response to a critical review. Anesth Analg 2016;122(2):565–72.
24 Goodnough LT et al. Detection, evaluation, and management of preoperative anaemia in the elective orthopaedic surgical patient: NATA guidelines. Br J Anaesth 2011;106(1):13–22.
25 Rossaint R et al. The European guideline on management of major bleeding and coagulopathy following trauma: fourth edition. Crit Care 2016;20:100.
26 Lee HC et al. Pulse oximeter-derived Pleth Variability Index is a reliable indicator of cardiac preload in patients undergoing liver transplantation. Transpl Proc 2016;48(4):1055–8.
27 Hare GM et al. Plasma methemoglobin as a potential biomarker of anemic stress in humans.
Can J Anesth 2012;59(4):348–56.