Editor, In a recently published, retrospective observational study of vascular surgery patients, Valentijn et al.1 showed that preoperative haemoglobin levels, postoperative haemoglobin levels and haemoglobin decrease were all related to an increased risk of 30-day postoperative cardiovascular adverse events. Strengths of this study include the large sample of patients, and the use of multivariable and sensitivity analyses to adjust, and to reduce, the influence of confounding variables on study endpoints. However, other than the limitations described in the discussion of the article, we have noted further issues of this study making the interpretation of their results questionable. First, health status, type of surgery and comorbidities are important determinants of postoperative cardiovascular adverse events. This study showed that patients with postoperative cardiovascular adverse events had more risk factors of the Revised Cardiac Risk Index, more comorbidities, lower preoperative or postoperative haemoglobin levels and larger haemoglobin decreases. Furthermore, they were older, more likely to undergo open surgery and to receive transfusion. In our opinion, no matter how refined the adjustment is for differences in health status, surgery burden and comorbidities, it is never possible to ensure a complete adjustment for differences between patients with and without postoperative cardiovascular adverse events. Additionally, preoperative anaemia is actually associated with a number of other comorbidities and known risk factors that can significantly affect postoperative cardiovascular adverse events, such as advanced age, poor left ventricular ejection fraction, renal dysfunction, congestive heart failure, myocardial infarction, unstable angina, and so on. This can further make it difficult to identify the specific role of preoperative anaemia, and even sophisticated multivariable analyses are probably inadequate to clarify whether anaemia is a risk factor or simply a marker of a more severe and complex preoperative clinical pattern. Recently, a retrospective study including 574 860 non-cardiac surgical patients has shown that preoperative anaemia was only associated with baseline diseases that markedly increased 30-day mortality and morbidity after surgery, whereas anaemia per se was only a weak and independent predictor of postoperative mortality and morbidity.2 Second, this study included and adjusted preoperative usage of many cardiovascular medications including β-blockers, but the authors did not specify whether patients taking these drugs before surgery continued their therapies after surgery. A systematic review of the literature showed that β-blockers lowered the perioperative risk of myocardial ischaemia and cardiovascular death among patients with clinical risk factors undergoing major vascular surgery.3 Furthermore, withdrawal of chronic β-blockade following vascular surgery has been shown to be associated with increased perioperative mortality.4 Third, this study did not include intraoperative haemodynamic data. In patients undergoing non-cardiac surgery, intraoperative hypotension, tachycardia and hypertension have been shown to be independently associated with postoperative myocardial injury and cardiac adverse events.5 Even short periods of an intraoperative mean arterial pressure less than 55 mmHg can result in myocardial injury, and there is an independent graded relationship between the duration of intraoperative hypotension and postoperative myocardial injury and cardiac complications.6 Finally, postoperative haemoglobin levels were measured within 0 to 3 days after surgery, but the authors did not specify the times when postoperative cardiovascular adverse events occurred and the times of postoperative troponin-T measurement. In high-risk patients undergoing vascular surgery under electrocardiographic monitoring, the majority of ischaemic events (67%) including those culminating in myocardial infarction, were reported to start between 50 min before and 60 min after the end of surgery and during the emergence from anaesthesia.7 Thus, in some patients who experienced postoperative cardiovascular adverse events, their postoperative haemoglobin levels might have been measured after occurrence of the events. Under these circumstances, great caution must be taken when interpreting the relationship of the measured postoperative haemoglobin levels with postoperative cardiovascular adverse events. Acknowledgements related to this article Assistance with the letter: none. Financial support and sponsorship: none. Conflicts of interest: none.