In brief, the post-operative complications of coronary artery bypass graft surgery are as follows:
- Death (3%)
- Respiratory failure
- Atelectasis (70%)
- Pleural effusion (~10%)
- ARDS (rare as hen's teeth)
- Pulmonary hypertension
- Myocardial dysfunction
- Cardiac tamponade (1.5%)
- CABG graft ischaemia (4-5%)
- Neurological complications:
- Focal neurological deficit (3%)
- Impairment of memory or cognition (3%)
- Renal failure (7%)
- Renal failure requiring dialysis (1.25%)
- Mediastinitis (1-4%)
Having your blood pumped artificially in a non-pulsatile fashion is a perverse physiological state which results in a series of post-operative problems. This chapter focuses on the complications of cardiopulmonary bypass which occur in the context of coronary artery bypass graft surgery.
Ray Raper's chapter on post-op management of cardiothoracic surgical patients in the ICU contains some mention of post-operative complications. The 2004 update to the 1999 AHA guidelines relating to the management of patients recovering from coronary artery bypass graft surgery elaborate on this, and are the basis of this summary.
The abovementioned list in the grey box is probably enough to get though the SAQs. Whatever those hypothetical SAQs may be. This material has never appeared in the written paper, and it is included here purely because it occupies a prominent position in Oh's Manual.
As is traditional here, a lengthy digression follows, the value of which is minimal to the time-poor exam candidate.
The AHA include a nice risk assessment scoring system in their publication, which can be used to predict a person's chances of survival. It is based on the analysis of large databases of CABG patients.
Overall, if all patient groups were homogenised, the risk of death is around 3%.
The patient features which predispose one to die after CABG are:
- Age (>65)
- Female gender
- Urgent surgery
- Poor ejection fraction (<40%)
- Left main coronary artery disease (>70% stenosis)
- Multiple coronary arteries with >70% stenosis
There will very likely (70% chance) be some atelectasis, particularly of the left lower lobe.
This occurs for a number of reasons.
Mechanical collapse of the left lung during one-lung ventilation
Typically, the CABG patient is ventilated with one lung, to allow some room for the surgeon to move. And, there is the possibility that the lazy anaesthetist has failed to completely reinflate the left lung after the operation has finished.
Left diaphragm hemiparesis due to phrenic nerve injury
This does not have to be an actually severed phrenic nerve; rather, more likely either the cold fluid during the operation or manual retraction has injured the nerve transiently, and its function will return in time.
Decreased chest wall compliance
This is far from surprising in a ribcage that has recently been sternotomised, cranked open, and then wired back together. The chest wall is not especially compliant now that there are sternal wires. And, its full of sharp pointy drains. This is going to be counterproductive when you try to take a deep full breath (as it will hurt).
There can also be a pleural effusion. Usually, but not always, the drains should take care of this.
Lastly, the effect of being on bypass on the complement cascade tends to result in leaky capillaries, and this may result in a degree of lung injury or ARDS.
Cardiopulmonary bypass results in a degree of pulmonary hypertension. This has several causes:
Atelectasis and hypoxic vasoconstriction:
If 27% of the lung has collapsed and is hypoxic, 27% of the pulmonary vasculature will constrict. The remaining 73% of the pulmonary vessels will receive close to 100% of the right ventricular output. This is a recipe for hypertension all on its own.
Effect of cardiopulmonary bypass of pulmonary vascular endothelium:
Activation of thromboxane-A2 is blamed for some of this; apparently (at least in anaesthetised rabbits) the bypass machinery tends to cause an activation of endothelia everywhere, and this causes neutrophils to aggregate in the microcirculation of the lung. This seems to be associated with an increase in the production of thromboxane-A2, which results in pulmonary vasoconstriction.
Additionally, reduced nitric oxide production seems to play some role. Again, there is experimental evidence for this - this time, in sheep.
Protamine has been implicated in the pathogenesis of post-bypass pulmonary hypertension.
The CABG patient tends to get about 400mg of protamine at the end of the procedure. This is a pretty heroic dose. The protamine molecule is polycationic, and its abundant presence in the bloodstream leads to complement activation, again resulting in neutrophil sequestration in the pulmonary circulation.
Apart from causing pulmonary vasoconstriction, protamine has also been blamed for the vasoplegic syndrome of the post-bypass patient. Protamine is rich in L-arginine, and may cause peripheral vasodilation by increasing the endothelial synthesis of nitric oxide, thereby triggering a widespread GMP-mediated smooth muscle relaxation.
But, more on that later.
There are numerous reasons for cardiac dysfunction after bypass. One can imagine the heart reacts badly to being cut open and having its innards fiddled with.
The risk of having a pericardial effusion after cardiac surgery seems to be around 1.5%. Of these patients, almost half have some evidence of clinically significant tamponade, and require either a return to theatre, or percutaneous drainage. As a differential for hemodynamic stability in the first 24 hours, this should be at the top of one's list, because it can become rapidly catastrophic. There is nothing more embarrassing than discovering tamponade after one has already commenced double-strength noradrenaline and vasopressin in a patient with "unexplainable" cardiovascular collapse.
Impaired contractility: "myocardial stunning"
By definition, this is prolonged but transient postischemic contractile dysfunction of viable myocardium that has been salvaged by reperfusion. There are many situations in which it may occur. It is usually a post-ischaemic affair, with recently reperfused myofibrils contracting poorly for a variety of reasons- poor utilisation of ATP, dysfunctional calcium ion handling, and reperfusion injury causing microvascular dysfunction. We have known that this happens after cardiac surgery for a very long time. In general, the contractility is impaired by as much as 50% post bypass, and reaches its nadir about 4-6 hours after the surgery is finished.
After 48 hours, contractility tends to return to normal. While you wait for this to happen, the patient is bombarded with a fusillade of inotropes and vasopressors; their aortic cavities are pummelled with pulsating helium-filled balloons.
In the immediate post-operative period this is likely to be due to air bubbles in the grafts; in the later hours, it tends to be related to poor perfusion of the arteries distal to the graft. The peak of incidence of this tends to be about 2 hours post reperfusion; thereafter, it becomes less likely. Lastly, graft thrombosis can occur within the first 30 days. Disturbingly, most of it is asymptomatic, and it occurs in about 4-5% of cases.
So, given that the myocardium has recently been cut into and generally mishandled, how do you know that the ECG changes and troponin levels are due to new ischaemia, rather than surgical trauma?
There is thankfully a guidelines statement on the "universal definition of MI" which directs our decisions to some extent.
The following features suggest your post CABG patient is infarcting:
- Troponin level is 5 times the normal maximum,
- New Q waves
- New LBBB
- New regional wall motion abnormality
Troponin cutoff levels are defined by the 99% percentile; CABG patients with a troponin 5 times that tend to have lower survival, and so these patients should probably be watched more carefully anyway.
However, why do troponins in them anyway? You know they will be elevated. Indeed, usually we don't.
The taskforce responsible for the above guidelines confesses shamefully that the "5 times the maximum" cut-off is completely arbitrary, and that the troponin levels should be interpreted as positive ONLY along with other positive findings, like Q waves and suddenly paralysed hypomotile LV walls.
The AHA generally recommends that thrombosed grafts be investigated with angiography and stented, just like any other coronary artery lesion. There does not seem to be any benefit in reopening the chest.
You need more preload than you think you do in the post-cardiotomy patients. The compliance of the left ventricle is impaired; even if you didn't have diastolic dysfunction pre-operatively, you've got it now. Thus, the post-CABG patients can require a PAWP of up to 18-22mmHg just to get to the useful part of the Starling curve.
Also, they may be actually volume-depleted. Some blood has been lost in the bypass machine.
Consider also that the patient has been unusually cold for some time. Hypothermic diuresis has robbed them of some significant amount of intravascular volume, and whatever was left is readily leaking out into the interstitium because of the bypass-associated SIRS response.
In short, fill em up.
Apart form the chronically hypertensive patients who fasted dutifully and omitted their medications, the hypothermic post-bypass patients will tend to have a higher systemic vascular resistance due to hypothermia. To some extent this can be defeated by the routine use of vasodilators such as GTN. One can add to this the afterload-reducing wonderfulness of IABP, and the vasodilation associated with the use of milrinone.
Atrial fibrillation is very common, occurring in up to 30% of post-CABG patients and 40% of valve repair patients. And it can cost you 15-30% of your cardiac output, depending on how the rest of your myocardium is working. It is particularly worrisome if your mitral valve is incompetent, and if your diastolic function is impaired.
Generally speaking, amiodarone is the drug of choice for these people. Digoxin has no better effect than placebo, and beta-blockers would be wonderful if they didn't interfere with the effect of catecholamine inotropes. In view of this, the Europeans recommend beta-blockade as the first choice, and amiodarone as the second.
And of course, the most wonderful feature of this is that the AF will probably resolve within 24 hours.
Somewhere between 5 and 25% of cardiac surgical patients will develop this syndrome, characterised by low systemic vascular resistance and a normal or low cardiac output. Apart from having a poor left ventricle and being on bypass for longer the only other risk factor seems to be ACE-inhibitor therapy - patients who are chronically ACE-inhibited tend to have more vasoplegia post bypass. Additionally, a patient who develops hypotension while on bypass will likely have vasoplegia post-bypass.
The mechanism of this vasodilation are not entirely clear, but there seem to be several aetiological factors at play.
The actual fact that blood is circulating in a non-pulsatile manner seems to lead to microvascular endothelial activation, causing leukocytes to adhere in the small vessels; and of course being exposed to the surfaces of the circuit tends to activate complement. Vasopressin levels are unusually low for a shock state, suggesting that (same as sepsis) there may be some degree of relative vasopressin deficiency. And on top of that the massive protamine dose contributes all this extra nitric oxide. In short, vasodilation is the result.
Noradrenaline seems to be the standard way of dealing with this. A trickle of vasopressin may be added, with reasonably good results. Lastly, methylene blue seems like a sensible way to turn off that nitric oxide production.
A more detailed summary is available of the approach to the hemodynamically unstable post-bypass patient.
Essentially, one can expect some degree of neurological dysfunction post-cardiotomy. These range from mild to severe. The AHA has divided these into Type I and Type 2.
Type 1 complications:
Type 2 complications:
Type 1 complications are essentially embolic and ischaemic. Proximal aortic atheroma is obviously a strong predictor for stroke - this implies that the surgeon, while manipulating the aorta, sends showers of microemboli into the brain. Similarly, the presence of pre-existing cognitive impairment suggests that the cerebral vessels are already diseased, and any additional insult to this circulation will exacerbate the dementia.
Type 2 complications relate more to damage resulting from microvascular insufficiency, and impaired cerebral metabolism. The risk factors for these complications relate to low flow states and therefore can be ameliorated by limiting the patients exposure to periods of hypotension.
The prevention of this sort of morbidity is in the hands of the surgeon, rather than the intensivist. Most of the preventative measures seem to revolve around the early identification of proximal aortic atheroma. if the surgeon is aware of it, they are more likely to tiptoe around it, and cause no strokes.
The AHA document contains in it numerous strategies for the reduction of perioperative stroke risk, and I will not reproduce them here.
Of these, about one fifth go on to require dialysis. Whatever your risk of death preoperatively, once you need dialysis to survive your CABG experience, you mortality risk increases to about 60%.
This is bad news. The major risk factors for developing end-stage renal failure after bypass are advanced age (>70) and pre-existing renal failure. The overall outlook is so bad that if you are over 70 and already have a creatinine over 260, you probably should never have a CABG- your chances of ending up with dead useless kidneys are better than 50%.
The patient recovering from cardiopulmonary bypass is at the same time prone to clotting and bleeding.
This is due to both the fact that the patients blood is constantly in contact with a series of unnatural plastic tubes rather than normal vessels. The massive dose of heparin which is administered is also to blame. And then, when administered in sufficient quantities, the protamine used to reverse the heparin is also on its own a reasonably potent anticoagulant.
The normal process of anticoagulation during bypass is to take a baseline ACT and then to poison the patient with about 300-400 units per kg of heparin, to achieve an ACT of around 400-480 seconds. In comparison, a normal ACT is 100-140, and you can safely perform hemodialysis at 200 or so. In any case, that means the standard 70kg patient will get about 21,000 – 28,000 units of heparin. After the surgeon have finished, the reversal of heparin with protamine is calculated according to the induction dose, and ends up being about 1 mg of protamine per every 100 units of heparin; so the same patient will end up getting about 210-280 mg of protamine. This is quite a lot.
But even assuming the patient comes back from theatre with a relatively normal ACT, the heparin that was hidden by protein-binding may come back out into the circulation and the ACT will increase again. Consider that protamine is a very rapidly degrade product- it has a half-life of around 5 minutes. The patient may become coagulopathic without you noticing, maybe 6 or so hours after coming back from theatre.
Of course, on top of that, the patients are frequently full of platelet inhibitors. And there are fewer platelets, many having been chewed up by the bypass machinery. Confusingly, there is some evidence that platelets are more “activated” post bypass, with a larger number of surface receptors exhibited; and they tend to be resistant to aspirin.
In short, there are multiple reasons as to why the patient might simultaneously clot their grafts and bleed into their pericardium.
Diabetic or not, you seem to develop hyperglycaemia during the time of cardiopulmonary bypass, and it seems resistant to insulin. This is probably the consequence of the catecholamines released during cardiac surgery. Catecholamines tend to act as antagonists to insulin, and they prevent its secretion.
Fortunately, for the majority this does not seem to influence the process of their recovery, and it tends to resolve after 6 or so hours.
An infected mediastinum is no fun. This happens in 1-4% of the patients, and carries a mortality of about 25%. Risk factors for it seem to include obesity, re-sternotomy, and diabetes. Obesity is probably the strongest independent predictor. A controversial association has been found between mediastinitis and the grafting of both internal mammary arteries; it seems this procedure tends to devascularise the sternum, making it more susceptible to infection.
Unsurprisingly, the site of the saphenous graft harvest is more frequently infected. Legs are generally dirty, hairy, frequently awash with human effluent. The risk of the SVG harvest site getting infected seems to range from 1 to 24%.