LAUSR

We thank Sun et al. [1] for their interest in our paper regarding postoperative hypoxaemia after minimally invasive colorectal cancer surgery [2]. They raise several clinically important questions; whether using an increase in troponin level below the cut-off of ischaemia is questionable; the appropriate precautions to take to prevent hypoxaemia and steps to take when encountering elevated troponin levels; the association between complications and hypoxaemia; and the association of hypoxaemia with haemoglobin levels, blood loss and obesity. In our study we used a troponin I assay which, according to the manufacturer, has a cut-off value for ischaemia of 45 ng.l . We used an increase in troponin above 15 ng.l 1 as an indicator for myocardial damage. The VISION trial [3], and subsequent studies, have shown that an isolated troponin increase is associatedwith short-and longterm mortality. This is true even for values below the threshold of ischaemia of their respective assays, and for increases from baseline and as fixed values [4, 5]. As correctly stated by Sun et al., the reference values are for our particular assay and cannot be universally applied, but increased risk associated with subclinical elevations in troponin, even below the ischaemic threshold, are generally accepted [6]. An important distinction between myocardial injury after non-cardiac surgery should be made, with regard to both their distinct presentations and the pathophysiological mechanisms: type 1myocardial infarction due to thrombosis and type 2 due to ischaemic imbalance [6]. The vast majority of patients with a troponin increase have no clinical symptoms and no ECG changes and thus are in the type 2 category [7]. Patients with a type 1 myocardial infarction often show both clinical symptoms and/or ECG changes. In our study, the patients of interest were those with type 2 infarction and ischaemic imbalance which could be associatedwith hypoxaemia. In our study, one patient developed a type 1myocardial infarction in the early postoperative period. This patient was subsequently treatedmedically and with invasive radiology. The patient had by far the highest troponin value which was above the cut-off point for ischaemia. We included that patient in our analysis, in order to maintain methodological rigour, but excluding the patient would only have strengthened the association between postoperative troponin I values and hypoxaemia, because that patient had a very limited time period with an oxygen saturation below 88% (as presented in the supplementarymaterial). Our study was purely observational and neither hypoxaemia nor troponin I increase was linked to mortality or long-term morbidity. We do not know whether the associations are linked to an increase in mortality or if interventions would have provided benefit in our study population. In the MANAGE trial [8], dabigatran 110 mg orally twice daily reduced the composite endpoint hazard ratio to 0.72 (95%CI 0.55–0.93) for patients with myocardial injury after non-cardiac surgery. With regard to postoperative complications, we did not observe any association between potential hypoxaemiarelated complications and the time patients spent with oxygen saturation below 88%. However, the frequency of complications was very low and the study was underpowered in this regard.We stated this in ourmanuscript and any conclusionsderived shouldbe cautious. The association between hypoxaemia and haemoglobin levels, blood loss or obesity, was not a part of the initial research protocol and thus not studied. As suggested by Sun et al., future research into these potential confounders/mediators would be of interest.