Use of echocardiograms in the Medicare population: a diagnostic tool in the fight against severe aortic stenosis

Of the total Medicare 5% FFS sample from 1 January 2011, through 31 December 2014 (3,383,990 patients), 2,529,497 (75%) had no record of receiving an ECHO anytime in the database (see attrition diagram in Figure 1). For the 854,493 (25%) patients who had a record of one or more ECHOs, 136,672 (16%) met the additional inclusion criteria of a diagnosis of AS; 6 months of continuous enrollment before time zero (the first instance of their AS diagnosis); and being aged 65 years or older at the time of AS diagnosis (Figure 1). Notably, only 7644 (5.6%) patients with a diagnosis of aortic valve disease had aortic valve replacement (AVR [surgical or transcather]) after diagnosis. The final sample size for each of the two cohorts of interest before matching was 136,672 patients for the ECHO-AS cohort and 1,173,299 patients for the no-ECHO cohort.

For both cohorts, patients’ baseline characteristics before matching are provided in Table 1. Patients who received an ECHO and had AS were older (average age of 78 years) and sicker (Elixhauser score of 5.97) than patients without a record of an ECHO (average age of about 74 years; Elixhauser score of 1.34). Patients from the ECHO-AS cohort also had a much higher rate of symptoms than the no-ECHO cohort (Table 2). ECHO-AS patients experienced dyspnea more than any other symptom (61,686 [45.1%]), followed by fatigue (55,869 [40.9%]). Conversely, patients with no-ECHO experienced fatigue more than any other symptom (136,684 [11.7%]), followed by dyspnea (65,735 [5.6%]). Angina was experienced the least by both cohorts (only 4.1% [5614] of ECHO-AS patients and 0.3% of no-ECHO patients). As expected, the ECHO-AS cohort had higher rates for all of the symptoms. Patients from the ECHO-AS cohort also had higher rates of atrial fibrillation, chronic obstructive pulmonary disease (COPD), hypertension and stroke than their no-ECHO–no-AS counterparts. Over half of the patients without a record of an ECHO (before matching) had no record of COPD, diabetes, stroke, heart failure, hypertension or hyperlipidemia.

After propensity matching, 21,609 patients were assigned to each cohort. Patient demographics, clinical symptoms and the Elixhauser scores for each cohort after matching are shown in Table 3. Although the cohorts were similar, small differences persisted; therefore, variables that remained significant after matching were included as covariates in the Cox regression model. Of the 5291 (24.5%) patients in the no-ECHO cohort who died in the post period, almost half (49%) had a record of a cardiovascular event before their death in the post period (Figure 2). Results of the Cox regression are shown in Table 4. The no-ECHO cohort is statistically significant (p = 0.003) with patients in the no-ECHO cohort at higher risk of death than their ECHO-AS counterparts. Statistically significant symptoms, which increase the risk of death are also shown.

This study found that only 25% of patients from the Medicare 5% FFS sample aged 65 years and above from 2011 to 2014 received an ECHO. AS is associated with a large, growing healthcare burden [2,6], and ECHOs are a key diagnostic tool when evaluating AS [4]. Baseline characteristics before matching indicated that patients who received an ECHO and had AS were older and sicker than patients without a record of an ECHO, which is a possible explanation for why the younger, healthier population did not receive ECHOs. However, AS is often asymptomatic [6]. Of elderly patients in the USA aged 75 years and above with severe AS, 24.4% are asymptomatic [2]. Patients may go years without developing symptoms; alternatively, mild or generic symptoms may be mistaken for respiratory disease [6]. ECHOs can help differentiate respiratory conditions like COPD from structural heart disease; and, exercise testing and stress echocardiography are also helpful when evaluating patients with VHD [12]. These supplemental tests are recommended to get a more complete picture of the disease and uncover otherwise masked symptoms in patients with asymptomatic AS [12]. Most patients with asymptomatic AS will develop symptoms within 5 years [13]; and, once symptoms like exertional angina, syncope and dyspnea develop, average survival rates for patients with severe AS decrease substantially, ranging from 2 to 5 years [9].

AVR has been shown to improve survival rates [14,15] and is indicated for patients with moderate and severe AS [1]; however, waiting too long to intervene may put excessive risk on patients. Unfortunately, a substantial number of patients do not have valve replacements, even though AVR is indicated for patients with severe AS [2,7,8,15]. The literature suggests that patients may not undergo AVR or opt for a more conservative treatment path because of comorbidities or patient preference, but more commonly, patients do not receive appropriate or timely care because physicians under recognize symptoms, misappropriate symptoms (e.g., mistake AS for a respiratory illness), or overestimate operative risks (including advanced age or left ventricular dysfunction) [7,9,15]. One study by Freed et al. [7] found that of 106 patients with severe AS, only 33 (31%) had undergone AVR and that the main reason that patients with symptomatic severe AS did not have AVR was their symptoms were thought to be unrelated to AS. Of the 73 patients who did not have AVR, 42 patients (58%) were asymptomatic but only 4% had undergone exercise stress testing (another common diagnostic tool for recognizing symptoms related to AS). Freed et al. also found that within 15 months, 15 (14%) of the 73 patients who did not undergo AVR died.

Moreover, a multicenter survey by Bach [8] in 2011 found that, overall, many patients who would potentially benefit from AVR were not evaluated properly for AS. Bach found that only about half of the patients with severe AS were referred to a cardiothoracic surgeon (referrals were more likely if patients had angina, which this study found to be an uncommon symptom) and only about 40% had AVR. Of the remaining patients who did not have surgery, about a fourth were asymptomatic.

This study supports a theme in the literature that patients with VHD are commonly underdiagnosed and undertreated because of a lack of testing [7] and timely referrals [8]. While physical examinations and auscultation do allow for the detection of asymptomatic VHD, ECHOs remain a key diagnostic tool in identifying VHD; yet, only about 25% of Medicare FFS patients had an ECHO, and of the 5291 patients in the no-ECHO cohort who died in the post period, almost half (49%) had a record of a cardiovascular event before their death in the post period (Figure 2). While it is not possible to confirm that these patients had AS, they had cardiovascular disease, which may have been treated correctly if those patients had received an ECHO.

The total Medicare parts A and B population in 2017 aged 65 years and older was estimated to be 49.8 million [16]. Of the 3,383,990 patients in the Medicare 5% FFS Database from 2011 through 2014, the initial sample for this study, 74.7% (2,529,497) did not receive ECHOs. If that percentage is applied to the total aged Medicare population, then 37,224,977 people did not receive ECHOs. Table 2 shows that the most common clinical symptom at baseline before matching among patients in the no-ECHO cohort is fatigue, affecting 11.6% (136,684/1,173,299) of patients (fatigue is also common in the ECHO-AS cohort; only dyspnea affects a slightly higher percentage of people). The least common symptom (in both cohorts) is angina, which is observed in 0.3% (2986) of no-ECHO patients. So, at the most, 11.6% of the no-ECHO group would have a reason to have an ECHO and, at the least, 0.3% should have an ECHO. Taking 0.3 and 11.6% of the total aged Medicare population who did not receive ECHOs yields a minimum and maximum of 94,736 and 4,336,541 patients, respectively. This range represents patients who, based on their symptoms, perhaps should have received ECHOs, but did not.

AS is most prevalent in the Medicare population, affecting 12.4% of people aged 75 years and above [2]. Increasing the frequency in which Medicare patients receive ECHOs will likely improve patient outcomes and help diagnose AS more accurately and efficiently. The ability to effectively diagnose AS will become increasingly relevant as the overall US population ages and the burden of disease increases [2,5]. Utilizing key diagnostic techniques like ECHOs more frequently and effectively will help identify severe AS in the Medicare population, broadening treatment options and improving clinical outcomes. If AS can be diagnosed earlier and more effectively, treatment options and survival rates will improve, and the future impact on public health and healthcare-resource utilization will likely decrease.

About 25% of patients in the Medicare population receive an ECHO, and aortic valve disease is diagnosed in about one in four of those patients. After matching, almost 50% of patients without a record of an ECHO who died also had a record of an acute cardiovascular event before their death. In the Medicare population, patients aged 65 years or older with increased risk factors for AS like hypertension, diabetes, COPD and stroke should be considered for a diagnostic ECHO.

To view the supplementary data that accompany this paper please visit the journal website at: Supplementary Material

This study was sponsored by Edwards Lifesciences. EA Grossi is a consultant to and holds intellectual property with Edwards Lifesciences and Medtronic. KA Moore is an employee of Edwards Lifesciences. C Gunnarsson is a consultant to Edwards Lifesciences. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

No writing assistance was utilized in the production of this manuscript.

The authors state that they have obtained appropriate institutional review board approval or have followed the principles outlined in the Declaration of Helsinki for all human or animal experimental investigations. In addition, for investigations involving human subjects, informed consent has been obtained from the participants involved.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/