Cameron, your comments are appreciated and pose interesting challenges. Although the issues you raise are in relationship to gastrointestinal haemorrhage there are similar concerns in many other clinical settings, especially with an ageing population with potential known and unknown co-morbidities.

Unless a patient is actively bleeding red cell transfusions are prescribed for the specific treatment of severe anaemia that may be having clinical consequences and the urgency of the situation is determined by the severity of the anaemia and the presence of any co-mobidities impacting on oxygen transport. In this setting the transfusion are directly therapeutic to manage the problem and are not prophylactic in the anticipation that there may be a further bleeding or cardiac or respiratory decompensation. In this situation the main questions relate to the quality of the red cells transfusion (ie age and functionality) and the dose to be administered. In otherwise stable patients with a “tolerable” level of anaemia, a level determined by factors such as co-mobidities and age, red cell transfusions are prophylactic. The transfusion is not being prescribed to address an immediately clinically significant defect on oxygen transport and delivery, but in the anticipation that there may be further bleeding. This issue you thus raise Cameron is a very important one as it poses the question as to which patients should receive these “prophylactic” transfusions. Several factors come into this decision making that apply to GIH and many surgical settings.

1.      What co-morbidities need to be considered and can they be specifically managed to minimize the need for red cell transfusions?

2.      What are the predictors that assist in defining the probability of further haemorrhage and likely severity and controllability?

3.      What can be done to minimize the chance of bleeding or re-bleeding?

4.      In giving prophylactic red cell transfusions we are inevitably exposing patients who do not ultimately bleed or re-bleed to the potential adverse effects of transfusion for no definable benefit. This has been the challenging question in the perioperative setting, especially in cardiac surgery and elective orthopaedic surgery where extensive observation studies and some RCT evidence indicates that more patients are exposed to harm than benefit. Red cell transfusions are a risk factor for several adverse clinical outcomes. Some of these patients no doubt needed transfusion while even more did not and suffered adverse consequences. Anaemia and bleeding are both potential factors leading to adverse clinical outcomes, but when are red cell transfusions beneficial and to which patients?

Using the common expression at the and of most publications “More research is needed to answer these questions!”

One factor which may be semi-specific for upper GI bleeders is that (particularly in old patients with lots of cardiovascular comorbidities) there may well be an element of risk of re-bleeding, at times rapid/large volume, even with endoscopically treated lesions. One thing that is always at the back of my mind with bleeding ulcers that are for one reason or another difficult to treat-and difficult to be sure that endoscopic therapy has definitively "fixed", eg gastric varices, ulcers with named vessels in their base, patients with liver disease etc-is not just whether the transfusion has got the patient back to safe territory but also whether brisk rebleeding will be tolerated. Not suggesting the need to deliberately overtransfuse but I do get a bit worried about sailing too close to the wind.

The issue of the impacts of hypotension and anaemia in the elderly patients is complex and no clear answers are available. However, in my view I think is generally possible to make the following comments on the basis of current knowledge and personal experience:-

1.       Hypotension and anaemia are less well tolerated by patients with vascular disease, especially cardiovascular and cerebrovascular, the question is what levels of anaemia can be tolerated and at what levels may transfusion of RBCs improve outcomes.

2.       If transfusion is thought to be indicated, what evidence is there for balancing the risks and the benefits. Obsession with the haemoglobin level as a surrogate endpoint is not appropriate and a broader approach to the role of transfusion in improving oxygen delivery to the tissues is important.

3.       Stored RBCs have a storage lesion that has been documented in vitro and in vivo studies and although the haemoglobin level may rise this does not necessarily indicate that the transfused red cells are functioning in delivering oxygen to the tissues. There is experimental data in hamsters that stored RBC transfusion in the hypotensive context will reconstitute the macrocirculation and improve functional capillary density, and as a result the microcirculation. From this, in the clinical setting, it is probably the patient’s own red cells are effective in transporting and delivering oxygen to the tissues. It takes several hours for the stored red cells to be rejuvinated to fully functional red cells.

4.       The role of fresh verses older stored RBCs remains sub judice and results from RCTs may be available in the not too distant future.

5.       The important issue in the GIH, as it is in many other clinical settings in which RBCs transfusions are given, are clinical outcome being improved. At present it is reasonable to state that in stable moderately anaemic patients there is no evidence that transfusion improves clinical outcomes and patients are exposed to risk without evidence for benefit.

6.       The unanswered question relate to:-

a.       What levels of anaemia can be tolerated and when is anaemia per se a risk factor for adverse clinical outcomes

b.      When is RBC transfusion likely to improve clinical outcomes and what are the indications (triggers) for transfusion

c.       What are the impacts of cardiovascular and cerebrovascular disease.

d.      A precautionary approach should not regard RBC transfusion as the default decision in the context of clinical uncertainty.

 Another confounder in upper GI haemorrhage is that most patients will have ongoing melaena for a day or two after the bleeding has ceased, owing to the passage of "old" blood through the gut. This, together with a falling Hb simply due to post-resuscitation haemodilution, may result in a knee-jerk reaction to transfuse in the mistaken belief that bleeding is ongoing. Accurate endoscopic assessment, including stigmata of recent haemorrhage, helps quantify re-bleeding risk, and correlation with vital signs, esp. BP and pulse. should mitigate unnecessary transfusion.

I see the key advantages of 'tolerated hypotensive resuscitation' in GIH are that it reduces exposure to the adverse immunosuppressive effects of blood and reduces rebleeding and therefore can improve patient outcome. This relationship between transfusion and rebleeding was shown quite nicely in the last British Society of Gastroenterology National Audit of GIH. Matched by Rockall category (severity score for non-variceal bleeding) , a higher proportion of those who were transfused re-bled compared with those who were not transfused. Clearly the younger, fitter patient can tolerate hypotension very well. It becomes trickier in the older patient with vascular disease where coronary and cerebrovascular hyoperfusion is less well tolerated. However, in this group giving blood may not the quick fix we might hope it to be because of the impaired oxygen carrying capacity of stored blood. Do you want to comment James?      

Anne, what are your views on 'tolerated hypotensive resuscition' until the cause of the GIT bleed has been identified and appropriate action taken?

Anne Duggan,

I think we all agree, transfusion has a role in the management of gastro-intestinal haemorrhage. Every decision is a risk benefit assessment. There are patients who do have very impressive gastro-intestinal haemorrhages and the challenge in these and all gastro-intestinal haemorrhages is to apply the strategies that we know can reduce transfusion requirements so that we can reduce exposure to the known risks of transfusion. Endoscopy has clearly moved from a diagnostic to a therapeutic strategy, with banding for varices and technics such as clips and gold probe for vessels exposed by ulceration. These strategies plus medications such as PPI infusions reduce the rates of ongoing bleeding and rebleeding and thus patient's exposure to the adverse effects of transfusion. I see the emerging challenge in gastroenterology as one of  ensuring that patients get the endoscopic and pharmacological treatments in a timely fashion and ensuring that they don't get "transfusion by numbers" post endoscopy and thus exposure unnecessarily to these risks.

I agree excessive transfusion is the important issue. Not being a gastroenterologist I will comment from a haematologist’s perspective. I prefer the concept of critical bleeding as being the focus of attention rather than massive transfusion. By the time the patient is being massively transfused, we are dealing with a different ‘syndrome’ and there is now good evidence that the transfusion (volume and age of blood) are risk factors for adverse outcomes. There is also considerable debate as to what and when in relationship to blood component therapy. As allogeneic blood transfusion may be life saving the risks referred to need to be accepted, but the challenge is to minimize the adverse impact of massive transfusion and/or to minimize transfusion. There are no easy answers and it is fair to state that the management of critical bleeding in general is an area that needs more research.

Accepting that massive haemorrhage presents difficult issues needing further research, lesser volume of GIT haemorrhage probably offer a situation in which transfusion can be avoided or minimized. Anne Duggan I am sure will have helpful input in this respect.

Surely, it is excessive transfusiion that is the problem, not the need for blood per se. If the patient is hosing, blood and blood products will be necessary, but the old idea of keeping the Hb above '10' (old units for old policy) is not appropriate in most patients.

Some patients with acute GI haemorraghe fall into the massive transfusion category. I am getting the distinct impression, having read other comments on this web site that giving a massive transfusion in this clinical context may not be best practice or indeed in the best interest of the patient? I would be interested in comments on this.