Troponomania or an Advancement in Care?
Abstract & Commentary
By Michael H. Crawford, MD, Editor
Sources: van Waes JA, et al. Myocardial injury after noncardiac surgery and its association with short-term mortality. Circulation 2013;127:2264-2271. Beckman JA. Postoperative troponin screening: A cardiac cassandra? Circulation 2013;127:2253-2256.
Elevated troponin levels after non-cardiac surgery have been shown to predict postoperative myocardial infarction (MI) and death. Thus, these investigators from the Netherlands instituted a postoperative troponin monitoring program at one hospital to test the hypothesis that troponin elevations in the first 3 days after surgery would predict 30-day mortality after non-cardiac surgery. Eligible patients had to be ≥ 60 years old undergoing intermediate- to high-risk non-cardiac surgery under anesthesia and who were expected to be in the hospital for at least 24 hours. Data collection was from electronic medical records. Troponin I was measured on the first 3 days after surgery and any abnormal values were reported to the attending physician who took whatever action deemed appropriate. The primary endpoint was 30-day all-cause mortality. The secondary endpoints included postoperative MI (universal definition) and length of stay.
Among the 2216 patients with complete follow-up, 1627 had troponin measured and 907 had all three measures. At least one elevated level was seen in 315 (19%). In about half of these patients (162), the troponin was elevated on the first day. The incidence of an elevated troponin increased with age and was associated with increased mortality as compared to those with normal troponins (8.6% vs 2.2%, P < 0.01). The median time to death with an elevated troponin was 12 days. Also, mortality increased with higher troponin levels (hazard ratio 2.4 with minor elevations and 4.2 with large increases). Multivariate analysis showed that troponin and emergency surgery were the only independent predictors of mortality. Hospital stay doubled with an elevated troponin (10 vs 5 days). Of those with a positive troponin, 40 met universal criteria for MI, 10 (3%) had typical chest pain, and 30 (10%) had ischemic ECG changes, but only one patient (0.3%) showed ST elevation. Cardiology consultation was sought in 35% of the elevated troponin cases, but in the majority no change in management was recommended. Only 6% had coronary angiography and 4% had revascularization. The authors concluded that elevated troponin values early after non-cardiac surgery predicted 30-day all-cause mortality and routine postoperative troponin measurement could improve the care of patients undergoing intermediate- to high-risk non-cardiac surgery.
Although the information is not new, this is the largest study of early postoperative troponin measurements to date and it confirms that elevated levels in the first 3 days independently predict 30-day all-cause mortality. Since the median time to death was 12 days, the authors reasoned that there was plenty of time to potentially reduce mortality postoperatively. The problem is what to do since troponin is predicting all-cause mortality, not cardiac mortality or even MI. In fact, only 2% of those with troponin measured met the universal definition of MI and only one patient had an STEMI by ECG. However, almost 20% had elevated troponin measures and among these patients, only 13% had an MI. Thus, most of the troponin elevations observed were what is now referred to as demand ischemia. Other than optimizing care to improve the myocardial supply-demand equation, there is nothing else to do acutely. We usually evaluate those who do well by stress testing later to be sure they don’t have fixed coronary lesions, but this may be overkill if they are active and asymptomatic postoperatively.
Interestingly, the study gave the attending physicians the troponin results and they could do what they wished with the patient. About one-third of the elevated troponin cases had cardiology consultations and, in 60% of these patients, no change in management was made; only 6% had coronary angiography and 4% had revascularization. In the remaining 30%, changes in medications were made — usually the addition of aspirin or statins.
Thus, it appears that elevated troponin levels are a nonspecific marker of illness. The editorial accompanying this paper comments that this is an underappreciated fact. The Dallas Heart Study of presumable healthy subjects showed elevated troponin T levels in 25% of the subjects at baseline. Abnormal troponins predicted all-cause and cardiac mortality in this study. However, it has been well described that troponin is elevated in pulmonary embolus, respiratory failure, sepsis, subarachnoid hemorrhage, and sick hospitalized patients in general. For these reasons, Beckman argues against routine troponins postoperatively. He believes that the potential harm — believing that the elevation is due to MI and starting on heparin, aspirin, clopidogrel, and possibly doing a coronary intervention, all of which would increase the risk of major bleeding early postoperatively — outweighs the benefit. Also, he believes focusing on the heart may divert attention from discovering the real reason for the troponin elevation. I don’t interpret this as meaning never do a troponin. In patients who are doing poorly postoperatively and in whom cardiac disease is suspected, it is reasonable to measure troponin and act accordingly.