The unwanted response to cardiac surgery: Time for a reappraisal?☆☆☆

Is there epidemiologic evidence that extracorporeal systems enhance adverse outcome? 

Large-scale epidemiologic studies have failed to show time of bypass (<90-120 minutes) as a specific or independent risk factor for the likelihood of dying,5 having respiratory,6 renal,7 or abdominal complications,8 or having a deep sternal wound infection.9 These studies encompass nearly 100,000 procedures.

Further evidence that supports the notion that extracorporeal circulation alone is not the initiator of subsequent problems comes from novel studies in which coronary surgery was done without bypass. As an example, Hernandez and colleagues10 reported on 7867 patients having myocardial revascularization of whom 1741 were operated on without bypass. Results for mortality and morbidity (apart from incidence of atrial fibrillation) were not different between management strategies. Taken together, the epidemiologic data do not support the concept of extracorporeal circulation as the principal villain of the piece but suggest that bypass is safe for the majority of patients.

However, this conclusion does not exclude the possibility that patients who do less well have an exaggerated response to the period of surgery and bypass. There is certainly evidence for great variability in the response of the fibrinolytic system to surgery and bypass and the response of individuals to a given inflammatory stimulus.11 The major research challenge in the sphere of the inappropriate inflammatory response remains the identification of those factors (whether genotypic or phenotypic) contributing to this increased susceptibility. In the article under discussion, a patient had excess endotoxin in the plasma, but without detrimental effect. Does this observation imply that endotoxemia alone is not pivotal in the genesis of an unwanted outcome?

Is there consistent evidence that a specific mediator is causally associated with adverse outcome? 

The earliest villain in the bypass-related inflammatory response was suggested as activation of complement.3 Although these early studies showed convincing evidence for the role of active fragments in adverse outcomes, activation of complement has now been superseded in the literature by other (more recently discovered) markers. The trend away from the role of complement activation may simply reflect that the modern bypass system does not induce such activity. For example, the study by Schmartz and associates shows no change in activated fragments of C3 or C5 (Article Table 1) and a decrease in C4 activation (Article Figure 2) with the extracorporeal system they used.

Continuing research searching for the mediator of the bypass-related inflammatory response has produced a vast amount of biochemical data on an array of potential culprits.

Cytokines are the principal inflammatory marker reported by Schmartz and colleagues.

Whereas, at low concentrations, cytokines seem to be essential for optimal function of the defense and repair systems of the body, the higher concentrations of cytokines measured in association with sepsis, trauma, and heart surgery are singled out as deleterious. The protagonists of this belief suggest modulating this response will be of benefit to the patient. The question therefore is whether the evidence consistently supports this notion.

What do we want a therapeutic intervention to achieve and how can this be related to abnormal biochemistry? 

The holy grail of the SIRS sleuth is to find a way to block or inhibit the “badness,” thereby allowing good to prevail. Reducing the biologic activities of IL-1 and TNF has been accomplished by several different, but highly specific, strategies, which involve neutralizing antibodies, soluble receptors, and receptor antagonists. Thus far, attempts to block IL-1 or TNF have been spectacularly unsuccessful in human beings with sepsis. Agents such as TNF-neutralizing antibodies, soluble TNF receptors, and IL-1 receptor antagonist have been infused into more than 10,000 patients in double-blind, placebo-controlled trials. Although there has been a consistent small increase (2%-3%) in 28-day survivals with anticytokine therapy, the effect has not reached statistical significance.17

In contrast, a recent meta-analysis showed that patients having heart surgery who received high-dose aprotinin therapy had a significant, 2-fold reduction in mortality compared with patients in the placebo or control arm of the studies examined.18 The exact reason why the use of aprotinin therapy is associated with a reduced mortality is currently opaque. Aprotinin has many potential benefits in terms of modulating the inflammatory response,19, 20 but these are not consistently reported and may be dependent on dose.

The confusion surrounding the role of the panoply of biochemical markers in an adverse outcome is not related only to cytokine responses. Similar arguments can be made for virtually all of the “inflammatory markers” reported in the literature and apply to aprotinin as well. For example, in a study investigating the influence of aprotinin on oxidant stress and neutrophil cytotoxicity using a cell culture system, decrease in neutrophil-associated cell death was observed with aprotinin but neutrophil respiratory burst, reflecting neutrophil oxidant output, was enhanced by this intervention.21 These data can be used by the protagonists of aprotinin therapy to show a benefit (reduced toxicity) or by the antagonists to show a potentially deleterious effect (increased oxidant stress).

Where can we go from here? The first logical step is to persuade investigators to stop considering biochemistry in isolation. If the biochemistry is not in some way related to a simple outcome measure, is it really of any clinical, as opposed to intellectual interest? An important and as yet not clearly defined question is, what are the clinically relevant end points? The obvious suggestion is a patient in a state that makes the responsible physician feel comfortable. If the patient is warm, passing urine, requiring no inotropic drugs, breathing spontaneously without distress, and talking rationally and coherently, does it matter that he or she has a raised “something we have an assay for and can measure?” So far, the isolated biochemical approach has not proved of any obvious benefit in helping us predict or understand nonfatal complications. Is it time to have a rethink on the strategy of investigating potential interventions or markers by first defining whether they are associated with improvements or worsening of patient outcomes?

The 50th anniversary of the successful use of the heart-lung machine will be in May of 2003. Many lives have been saved and their quality improved because of this innovation; some have also been lost or the quality reduced. It may be wishful thinking, but by this anniversary it should not be beyond the clinical and scientific communities involved in cardiothoracic surgery to decide on a core group of patient-oriented end points to allow us to define an appropriate or inappropriate outcome. These criteria can then be used by all investigators to ensure that their studies and data can be related to the patient in addition to the test tube. The suggestion that the patient is warm to the periphery, passing adequate volumes of urine, requiring no inotropic agents, and is extubated and talking rationally and coherently within a period of time from the end of surgery is what every clinician anticipates for the patient and therefore seems a good place to start.