Search tips
Search criteria 


Logo of jrsocmedLink to Publisher's site
J R Soc Med. 2002 September; 95(9): 450–452.
PMCID: PMC1279992

Can hospital transfusion committees change transfusion practice?


Blood and blood products are commonly over-used in hospital practice. We investigated whether the introduction of a red-cell transfusion trigger (haemoglobin <8 g dL-1) influenced transfusion practice in surgery. Coronary artery bypass grafts (CABGs, n=400), total hip replacements (n=107), colectomies (n=85) and transurethral prostatectomies (TURPs, n=158) were reviewed over two periods of six months, before and after the introduction of the policy by the local hospital transfusion committee.

After introduction of the policy, the proportion of patients transfused fell from 57% to 45% with CABGs (P=0.02) and from 52% to 26% with hip replacements (P=0.006); for colectomies and TURPs there was no change. Hospital stay did not increase in any of the groups. In the second period, haemoglobin concentration on discharge was lower after total hip replacement, by a mean (95% CI) of 0.7 (0.3-1.2) g dL-1 (P=0.002) and after colectomy, by a mean of 0.6 (0.1-1.1) g dL-1 (P=0.03).

Although other factors cannot be excluded, we suggest that the reductions in red-cell transfusion were in large part attributable to the new transfusion policy.


Blood is often given inappropriately1,2,3,4, with consequent risks for patients and wastage of limited and expensive resources. The NHS Executive recently recognized shortcomings in blood product management in the UK and recommended measures to reduce the inappropriate use of blood and blood products5. Hospital transfusion committees—multidisciplinary groups responsible for disseminating local transfusion protocols according to published evidence and national guidelines—were asked to produce a transfusion policy by March 2000.

The transfusion committee at South Manchester University Hospital produced a policy for red-cell transfusion in October 1999, including a standard blood ordering schedule and a transfusion trigger of a haemoglobin concentration <8 g dL-1 in the absence of symptoms. To assess the impact of this trigger on transfusion practice, we audited transfusion requirements in surgery over two periods of six months—before the guidelines, from 1 January 1999 to 30 June 1999; and after the guidelines, from 1 November 2000 to 30 April 2001. The second period was one year after introduction of the transfusion policy to allow full dissemination and implementation.


Four elective surgical procedures were audited—coronary artery bypass grafts (CABGs), total hip replacements, colectomy (right, transverse, left, sigmoid, subtotal or total colectomy or Hartmann's resection) and transrethral prostatectomy (TURP). These were chosen because they are the most frequently performed procedures associated with substantial transfusion requirements. (Although aortic surgery is also common and generally associated with homologous transfusion, in South Manchester these procedures are routinely performed with intraoperative cell salvage and acute normovolaemic haemodilution, which limits homologous blood use. Furthermore, many aortic reconstructions before the transfusion policy were enrolled in trials on transfusion strategies with strict transfusion triggers6,7,8.) Since over 800 CABGs were performed, we decided to limit the analysis to the first 200 in each period. Haematology databases were searched to obtain haemoglobin concentration at admission and before discharge, number of crossmatch requests and details of any red-cell transfusions (including volume transfused).

Proportions were compared with the χ2 test. Continuous variables were expressed as mean and standard deviation (SD) if normally distributed (haemoglobin concentration) and as median and interquartile range (IQR) if skewed. The former were compared by the unpaired t test and the latter by the Mann—Whitney U test.


We studied a total of 400 CABGs, 107 total hip replacements, 85 colectomies and 159 TURPs.

Coronary artery bypass grafts

The median (IQR) number of crossmatch requests was 1 (1-1) in both periods (P=0.53). 114/200 patients (57%) were transfused before and 90/200 (45%) after introduction of the transfusion policy (P=0.02). The median (IQR) volume transfused was 1 (0-2) units before and 0 (0-2) units after introduction of the policy (P=0.12). There was no significant difference in haemoglobin concentration on admission or discharge between the two periods (Table 1). The median (IQR) inpatient stay was 11 (9-15) days in the first period and 10 (8-13) in the second.

Table 1
Haemoglobin concentrations on admission and discharge before and after introduction of the transfusion policy

Total hip replacement

The median (IQR) number of crossmatch requests was 1 (1-2) in the first period and 1 (1-1) in the second (P=0.33). 26/50 patients (52%) were transfused in the first period and only 15/57 (26%) in the second (P=0.006). The median (IQR) volume transfused was 2 (0-3) units before and 0 (0-1.5) units after introduction of the transfusion policy (P=0.003). Haemoglobin concentration on discharge was lower by a mean (95% CI) of 0.7 (0.3, 1.2) g dL-1 (P=0.002) after introduction of the transfusion trigger (see Table 1). The median (IQR) inpatient stay was 10 (8-13) days in the first period and 9 (7-13) days in the second.


The median (IQR) number of crossmatch requests was 1 (1-2) in both periods (P=0.97). 24/45 patients (53%) were transfused in the first period and 22/40 (55%) in the second (P=0.88). The median (IQR) volume transfused was 2 (0-5) units in both periods (P=0.94). Haemoglobin concentrations at admission were similar in the two periods but, on discharge, concentrations were significantly lower after introduction of the transfusion policy (see Table 1). The median (IQR) inpatient stay was 16 (11-31) days before and 16 (12-25) days after institution of the transfusion policy.

Transurethral prostectomy

The median (IQR) number of crossmatch requests was 0 (0-1) in both periods (P=0.39). Only a minority of patients were transfused—12/80 (15%) in the first period and 18/78 (23%) in the second (P=0.2). Haemoglobin concentrations at admission and discharge did not differ in the two periods (see Table 1). The median (IQR) inpatient stay was 7 (5-10) days in the first period and 6.5 (5-9) in the second.


This study showed a reduction in red-cell transfusions after introduction of a transfusion policy. This reduction was evident in cardiac surgery and was particularly impressive in orthopaedic surgery, with transfusions halved and haemoglobin concentrations at discharge substantially lower. Interestingly, although transfusion practice for colectomy did not appear to change after introduction of the transfusion trigger, haemoglobin concentrations at discharge were also lower after introduction of the transfusion policy, suggesting that surgeons were more willing to withhold transfusion. Inpatient stay was not increased by the more stringent transfusion policy—indeed, in the CABG and hip-replacement groups it declined. We cannot, however, comment on outcomes, since we did not measure morbidity or mortality; moreover, other factors may have influenced hospital stay.

A haemoglobin concentration <8 g dL-1 may seem restrictive as a transfusion trigger, but it is in line with recent published work9,10,11 and has lately been used by our group in a large randomized trial evaluating transfusion strategies in aortic surgery6,7,8.

There are several reasons why a reduction in homologous blood use is desirable. Aside from the obvious avoidance of complications, withholding unnecessary transfusions may spare patients the immunomodulatory effect of homologous blood, which is suspected by some of increasing susceptibility to bacterial infection12 or cancer recurrence13. There is also growing concern about future availability of blood stocks in view of the impending routine testing of all homologous blood products for the agent of variant Creutzfeldt—Jakob disease. It is impossible to predict how many will be excluded on the basis of the new test. Furthermore, many existing donors may be unwilling to undergo testing and may therefore be lost to the system.

Were the observed changes attributable to the transfusion policy or to other factors? Changes in personnel, new published data and the introduction of new procedures could have contributed to a change in attitudes to red-cell transfusion. There is greater reason to doubt a direct effect of the transfusion policy on hospital stay, which might have been influenced by factors such as pressure on beds.

Although the NHS Executive has recommended the institution of local transfusion policies, this may be difficult because objective transfusion triggers are lacking. Red-cell transfusion is given to restore or preserve adequate oxygen delivery to the tissues but haemoglobin, the best transfusion trigger currently available14, only measures blood oxygen carrying capacity. Haemoglobin is thus relatively unhelpful, in isolation, in most circumstances where red-cell transfusion is considered15; hospital transfusion committees will therefore need to be flexible in the enforcement of transfusion triggers. In conclusion, this study provides some evidence supporting the role of hospital transfusion committees in formulating local transfusion policies. Further reduction in the number of red-cell transfusions may require identification of alternative transfusion triggers.


1. Sudhindran S. Perioperative blood transfusion: a plea for guidelines. Ann R Coll Surg Engl 1997;79: 299-302 [PMC free article] [PubMed]
2. Kytola L, Nuutinen L, Myllyla G. Transfusion policies in coronary artery bypass—a nationwide survey. Acta Anaesthesiol Scand 1998;42: 178-83 [PubMed]
3. Garrioch MA, Sandbach JA, McIlveney S. The ‘transfusion trigger’ in three large west of Scotland hospitals: medical staff questionnaire survey. Transfus Med 2000;10: 193-8 [PubMed]
4. Capraro L, Nuutinen L, Myllila G. Transfusion thresholds in common elective surgical procedures in Finland. Vox Sang 2000;78: 96-100 [PubMed]
5. Winyard G. Better Blood Transfusion (NHS Executive Clinical Effectiveness circular HSC 1998/224) []
6. Haynes SL, Wong JCL, Torella F, Dalrymple K, Pilsworth L, McCollum CN. The influence of homologous blood transfusion on immunity and clinical outcome in aortic surgery. Eur J Vasc Endovasc Surg 2001;22: 244-50 [PubMed]
7. Torella F, Wong JCL, Haynes SL, McCollum CN. Autologous or homologous blood transfusion in aortic surgery: randomized trial. Br J Surg 2001;88: 604
8. Wong JCL, Torella F, Haynes SL, Dalrymple K, Mortimer AJ, McCollum CN. Autologous versus allogeneic transfusion in aortic surgery (ATIS). A multicenter randomized clinical trial. Ann Surg 2002;235: 145-51 [PubMed]
9. Bracey AW, Radovancevic R, Riggs SA, et al. Lowering the hemoglobin threshold for transfusion in coronary artery bypass procedures: effect on patient outcome. Transfusion 1999;39: 1070-7 [PubMed]
10. Carson JL, Duff A, Berlin JA, et al. Perioperative blood transfusion and postoperative mortality. JAMA 1998;279: 199-205 [PubMed]
11. Hebert PC, Yetisir E, Martin C, et al. Is a low transfusion threshold safe in critically ill patients with cardiovascular disease? Crit Care Med 2001;29: 227-33 [PubMed]
12. Duffy G, Neal KR. Differences in post-operative infection rates between patients receiving autologous and allogeneic blood transfusion: a meta-analysis of published randomized and nonrandomized studies. Transfus Med 1996;6: 325-8 [PubMed]
13. Amato A, Pescatori M. Effect of perioperative blood transfusion on recurrence of colorectal cancer. Meta-analysis stratified on risk factors. Dis Colon Rectum 1998;41: 570-85 [PubMed]
14. Miller RD. Transfusion therapy. In: Miller RD, ed Anesthesia, Vol 1. Philadelphia: Churchill Livingstone, 2000
15. American Society of Anesthesiologists task force on blood component therapy. Practice guidelines for blood component therapy. Anesthesiology 1996;84: 732-47 [PubMed]

Articles from Journal of the Royal Society of Medicine are provided here courtesy of Royal Society of Medicine Press