Robert W. Regenhardt, MD, PhD
@rwregen
Throughout my training, I have had several mentors dissuade me from ordering tests that are unlikely to change management during the admission of patients with ischemic stroke. Transthoracic echocardiography (TTE) is one of the studies that is not uncommonly cut, especially if the stroke etiology is clear without it. Indeed, the recent 2018 American Heart Association guidelines do not mandate TTE, but recommend the clinician use his or her judgement. This recent article in Stroke by Yaghi et al. set out to test the yield of TTE in these patients and the utility of troponin levels to improve its yield. They examined 578 patients (mean age 74) admitted to their single center over an 18-month period with ischemic stroke “whose etiologic subtype could be obtained without the need for TTE.” Of these patients, TTE changed clinical management in 11.1%, but identified intracardiac thrombus in only 0.7%. The authors also identified an association between positive troponin levels and TTE changing management (adjusted OR 4.26, 95% CI 2.17-8.34, P<0.001).
As the authors aimed to study specifically the patients whose subtype was determined without the need for TTE (i.e., patients with low pretest probability), they included patients with atrial fibrillation/flutter (55%), small vessel disease (26%), large artery disease (17%), and “other determined mechanisms” (3%). They excluded patients who had an indication for TTE, including patients with known valve replacements and those requiring work-up for embolic strokes of unknown source (low ejection fraction, severe valvular disease, cardiac thrombus, patent foramen ovale). TTE was considered to have “changed clinical management” if the following were identified: regional wall motion abnormality without CAD history (found in 6%), EF<40 without CHF history (4%), severe valvular disease (1.4%), or cardiac thrombus (0.7%). The first three findings prompted referral to cardiology, and the fourth prompted anticoagulation. Positive troponin was quite specific (0.906) for predicting TTE to change management, but not as sensitive (0.339). As above, positive troponin was associated with TTE changing management, including in the subset of patients with noncardioembolic stroke and in sensitivity analyses excluding patients with cardiac thrombus.
It is perhaps surprising that in this cohort of stroke patients with clear etiologies, TTE still changed management in 11.1%. This may reflect the shared risk factors for stroke and cardiovascular disease, and may have little implications for stroke etiology; patients with stroke are likely to also have concomitant covert cardiac disease. Only 0.7% were found to have intracardiac thrombus prompting acute treatment with anticoagulation. Overall, this affirms what many believe in practice: TTEs often do not change management. However, while the other TTE findings accounting for the 11.1% are likely less urgent, it is worth questioning whether they would have been detected at a later time. Would patients’ primary care providers find other evidence of CAD, CHF, or valvular disease? Would delay in diagnosis affect outcomes? Troponin is correlated with TTE changing management and may improve TTE’s diagnostic yield, but with its sensitivity of only 0.339, one may want to also consider additional factors to determine if TTE is, in fact, unnecessary before cutting it from the work-up. As the authors point out, larger studies are necessary. Furthermore, clinical outcome and financial endpoints could also be included in future studies to more thoroughly weigh the risks and benefits of obtaining TTE.
