‘The Troponin Story’
Dr Michael Stewart, Middlesbrough
The meeting was opened by the President of the LMI, Mr. Graham Lamont. LSA President, Dr Ewen Forrest, then introduced the speaker, Dr Michael Stewart, a Consultant Cardiologist at the James Cook University Hospital, Middlesbrough.
Dr Stewart began by explaining the pathophysiology of myocardial infarction (MI), stating there are two possible mechanisms for a perioperative cardiac event, the vast majority being type 1 (ruptured plaque with thrombus on it) or type 2 (pre-existing coronary artery stenosis exposed to bleeding, hypotension or tachycardia, which increases myocardial oxygen demand, causing myocardial damage.) Historically, he said, the type 2 events were thought to be the problem in ischaemic heart disease and that optimising patients undergoing non-cardiac surgery pre-operatively would reduce the risk of MI, but the evidence for this was scant.
Dr Stewart talked about subsequent clinical data that did not fit with this hypothesis, the most widely quoted being the POISE and CARP trials. So the alternative hypothesis was actually that the stress of surgery was causing type 1 events.
Dr Stewart explained that there has been a general acceptance that many perioperative events are due to plaque rupture, so pre-operative tests will not identify the lesion causing the risk and therefore the practice of routine non-invasive stress testing in the absence of symptomatic angina other than in the highest risk surgical cases has a more limited role than previously thought.
He said that, in terms of prevalence, the authors of the POISE trial (Lancet 2008; 371: 1839–47) pointed out that these events might be more common than thought. They routinely measured troponin for 3 days post-operatively to see whether beta blockers reduced the rate of MI. It showed that although beta blockade did reduce the rate, the incidence of hypotension was greater which was associated with a higher mortality. In those that did have an MI, 75% occurred within 48 hours of surgery but 2/3 of that group had no symptoms. The only reason that they were picked up was because their troponin levels were measured. When they compared those with symptomatic ischaemic events and those without ischaemic symptoms, there was a similar 30 day mortality of 11.6%.
Dt Stewart said that as a result, the VISION trial was set up. Troponin levels were measured for the 1st 3 days post-operatively. Troponin was shown to be a strong marker of 30-day mortality and a ‘rule of squares’ was designed which predicted mortality risk (JAMA June 6 2012 Vol 307 No 21). Even low levels of troponin release marked people out with higher mortality rates, but only 45% of deaths were deemed to have a cardiovascular cause.
Dr Stewart went onto say that the second paradigm shift is that perioperative myocardial damage is more common than we realised and many events are asymptomatic. Troponin release seems to be a very strong marker of subsequent prognosis, independent of patient symptoms. He said a new term ‘Myocardial Injury after Non-cardiac Surgery’ (MINS) has been introduced. If we differentiate MINS from MI, it might help us to study it and not label patients as having an MI unnecessarily. MINS is troponin release after surgery with no other clear explanation for it (e.g. pulmonary embolism, sepsis, cardiac arrhythmia and anaemia and there are no symptoms or ECG evidence to suggest an MI.) However MINS patients have a mortality rate twice that of patients with MI presenting in the Emergency Department. Once MINS has been identified, the important point was to ensure that patients had good basic medical care in a well monitored environment with the possible addition of aspirin and a statin.
Dr Stewart explained his study and their protocol, for which he received an innovation grant. They took the same VISION type criteria and initially did blind troponin testing. These results were not released to the clinicians, but recorded and compared to event rates and outcomes. They produced a management protocol for surgical teams which said troponin release should be a trigger to look for underlying causes e.g. sepsis, PE, bleeding. If these were excluded, the patient was started on a statin and aspirin. The cardiologist was only informed if there were any ECG changes or the patient had ongoing ischaemic symptoms.
Of the patients to date who were eligible for the study and had a complete data set, 36.6% had an elevated troponin of which 58% were undergoing emergency surgery. The overall 30-day mortality rate was 2.8% (95% of whom had undergone emergency surgery) and a 10% 6-month mortality rate (75% being in the emergency group). He said that an elevated troponin was a marker of increased mortality rate. Dr Stewart saw a MINS level that was twice as high as the VISION study with a 30-day mortality that was 1.5 times higher (albeit with much smaller numbers). He said that after having undertaken a more detailed retrospective analysis, patients with elevated troponin levels are often associated with wider pathophysiology such as sepsis, bleeding, renal impairment or other critical illness. Excluding these patients, they found that there were no MINS patients who had had an unexplained death within 30 days. After univariate and multivariate analysis, they found that all of these variables are predictors of an adverse outcome. Adding troponin to them creates a slightly bigger hazard ratio. Abnormal renal function, sepsis and emergency admission were all indicators of risk and only very high troponin levels (>1000) are independently associated with an excess mortality rate. Troponin is a marker of increased mortality risk but not independent of other established risk factors. That is why interventions targeted exclusively at coronary artery disease such as aspirin and beta blockers, have so far failed to show any benefit in perioperative medicine.
Dr Stewart summarised that troponin release peri-operatively is more common than previously recognised and is associated with poorer outcomes when associated with other pathophysiology. However a raised Troponin, as a stand-alone event, was not associated with a higher mortality. Troponin release appeared to be multi-factorial and less likely to be a separate clinical entity. He pondered whether we should we be measuring troponin routinely post-operatively because currently there is no evidence that specific interventions aimed at reducing Troponin release can modify outcomes. Troponin however may be a good surrogate marker of prognosis for other interventions to try and reduce perioperative morbidity and mortality e.g. changing an inotrope strategy. Observational data may support the initiation of aspirin and a statin to reduce late events.
Dr Stewart took some questions from the floor and Dr Richard Nelson gave the vote of thanks.
Dr Gemma Redmond