Preoperative Echocardiography in Hip Fracture Patients: A Waste of Time or Good Practice?

The debate surrounding the value of preoperative testing is as old as the field itself. However, few discussions have highlighted the issues involved as extensively as the perioperative management of patients with hip fractures.1 The affected patient population represents a major challenge to health care systems worldwide both from a medical as well as an economic point of view.2,3 As the current treatment paradigm focuses on early surgical intervention and avoidance of unnecessary delays in an attempt to speed rehabilitation and improve outcomes, the role of traditional preoperative medical evaluation and optimization strategies has been put into question.1,4 Faced with a nonelective procedure that most commonly will need to be performed irrespective of patient comorbidities and risk of complications (unless nonsurgical or end of life care is deemed viable alternatives), a delay in the definitive treatment of a hip fracture patient has been suggested to be unnecessary and often unethical as it does not materially change management or/and may worsen outcome.5 Caught in this dilemma is the anesthesiologist who has to balance the need for expedient surgery on one side and sufficient information gathering and patient status optimization of immediately addressable problems on the other. In this context, it is not uncommon that patients with significant chronic and acute cardiac disease processes are encountered leading to requests for further workup. Echocardiography for routine assessment of suspected cardiac disease is often performed with the goal of identifying pathology that might alter treatment.5,6 However, there is little evidence to date that such practice alters outcomes.7,8 The paucity of data on this topic is partially to blame on the fact that randomized controlled trials may be difficult (although not impossible) to undertake in a heterogeneous population such as hip fracture patients. In an attempt to add to our knowledge, Yonekura et al9 seek to address the association of preoperative echocardiography on perioperative outcomes in hip fracture patients using a population scientific approach. For this purpose, the authors utilized a large nationwide database from Japan and included records from tens of thousands of individuals over a time period of 8 years. In their sample, approximately half of the patients underwent echocardiographic examination preoperatively. As expected, this patient population tended to be older and has higher comorbidity burden compared to those who did not receive the intervention, suggesting the presence of indication bias. After applying propensity score-matching methodology, the investigators failed to show a difference in the rates of in-hospital mortality, complications, and intensive care unit admissions leading them to conclude that there was no association between the use of preoperative echocardiography and in-hospital outcomes.9 While these findings derived from data collected from “real-world” practice without the confines of strict inclusion and exclusion criteria—as is common with randomized clinical trials—are highly important, population-based analyses such as these leave the reader not only with insights but also with unanswered questions.10 As stated by the authors carefully and correctly, one cannot draw firm conclusions from the data beyond determining associations. Causal relationships cannot be definitively determined as residual confounding remains. It therefore remains speculative what lead physicians to request an echocardiographic evaluation of their patients before surgery. We cannot know from these data if routine protocols were followed, if evaluations were based on abnormal clinical exams or history, or if other concerns were drivers for such an examination. However, these data clearly suggest that older age and increased comorbidity burden may have affected this decision. While statistical methodologies, like propensity score matching, are used to control for differences in covariates, matching remains incomplete and relies solely on variables available. Unmeasured variables containing clinical details and decision-making processes are not taken into consideration. The inability to establish causal relationships leaves the possibility of an entirely different interpretation of results open. It is possible that no difference in outcomes was found between those who did and those who did not receive a preoperative echocardiogram, precisely because the use and findings of the echocardiographic intervention lead to care that improved outcomes in a population that would have otherwise fared worse. In this context, it is important to mention that it is not the test that changes outcome but the decisions made based on the findings derived from the examination. It is of note that there is some indication that patients who received an echocardiogram did indeed receive different care. This group was associated with a more frequent use of hemiarthroplasties versus internal fixations, blood transfusions as well as general versus neuraxial anesthesia—all which may be related to the patient’s pathology but nevertheless reported to possibly affect outcomes.11 A longer length of stay was also associated with the use of echocardiography, posing the question if such intervention is a surrogate marker for changes in perioperative care,9 further suggesting the possibility that the use of this test delayed surgery and/or prolonged hospitalization due to logistic reasons. In conclusion, Yonekura et al9 present important information and incrementally add to our current knowledge of the subject. Just as with all population-based studies, where definitive causality cannot be established, one must apply the measure of logical plausibility and examine possible explanations for the various findings. These then become the basis for specific hypotheses for targeted mechanistic studies that can individually address the questions unearthed by observational data. Thanks to investigations such as the one presented by Yonekura et al,9 we gain insights into the realities of daily practice leaving us with the opportunity to study why we see what we see. Until such answers can be produced allowing us to more definitively determine if and which patients benefit from perioperative echocardiography, its utility in patients with hip fractures awaiting surgery remains elusive. However, concerns purely related to cost and logistics related to echocardiography may be addressable with advances in technology and education in the form of point-of-care ultrasonography.12 Anesthesiologists who do not just want to assume the worst when proceeding to the operating room and have reason to suspect cardiac abnormalities that might change their management or raise their level of vigilance may be able to deploy this cost-effective diagnostic tool at the bedside without significant delay. But once again, we will have to wait for evidence for this assumption as well. DISCLOSURES Name: Stavros G. Memtsoudis, MD, PhD, MBA. Contribution: This author helped conceptualize this editorial, prepare the manuscript, and solely responsible for its content. Conflicts of Interest: S. G. Memtsoudis is a one-time consultant for Teikoku and Sandoz. He is a patent holder for Multicatheter Infusion System [US Patent No. US-2017-0361063]. He is a owner of SGM Consulting, LLC. He is a co-owner of FC Monmouth, LLC and FC Monmouth Academy, LLC. None of these relationships have influenced the work presented here. This manuscript was handled by: Nikolaos J. Skubas, MD, DSc, FACC, FASE.