Disproportionately high aortic valve calcium scores in atrial fibrillation: implications for transcatheter aortic valve replacement

Abstract

Aims

Doppler mean gradient (MG) can underestimate aortic stenosis (AS) severity in patients with atrial fibrillation (AF) compared with patients with sinus rhythm (SR), potentially delaying intervention in AF. This study compared outcomes in patients with AF and SR following transcatheter aortic valve replacement (TAVR) and investigated delay in TAVR based on computed tomography aortic valve calcium score (AVCS).

Methods and results

Patients who underwent TAVR from 2013 to 2017 for native valve severe AS were identified from an institutional database. Baseline characteristics and overall survival were compared between those with SR and AF. There were 820 patients (mean age 81 years; 41.6% females) included in this study. AF was present in 356 patients. Patients with AF were older (82.2 vs. 80.5, P = 0.003) and had a lower MG compared with patients with SR (42.0 vs. 44.9, P = 0.002) with similar indexed aortic valve area (0.4 vs. 0.4, P = 0.17). Median AVCS was higher in AF (males: AF 2850.0 vs. SR 2561.0, P = 0.044; females: AF 1942.0 vs. SR 1610.5, P = 0.025). Projected AVCS, assuming the same age of diagnosis, was similar between AF and SR. Median survival post-TAVR was worse in AF compared with SR (3.2 vs. 5.4 years, log rank P < 0.001). AF, lower MG, higher right ventricular systolic pressure, dialysis, diabetes, and significant tricuspid regurgitation were associated with higher mortality (P < 0.05 for all).

Conclusion

Older age and higher AVCS in patients with AF compared with those with SR suggest that AS was both underestimated and more advanced at TAVR referral.

Graphical Abstract

Open in new tabDownload slide

Introduction

Aortic stenosis (AS) is a common valvular lesion in the USA, with a prevalence of >7 million among the elderly population (>75 years of age).^1,2 Due to their overlapping risk factors and pathophysiologic associations, AS is comorbid with atrial fibrillation (AF) in up to 50% of patients.^3,4 The presence of AF poses a challenge in grading the severity of AS owing to beat-to-beat variation in left ventricular (LV) stroke volumes affecting transaortic gradients in addition to the lower flow rate from losing atrial contraction.^5,6 Concomitant valvular dysfunctions, including mitral regurgitation (MR) and tricuspid regurgitation (TR), in addition to diastolic dysfunction and increased filling pressures, also contribute to an overall low output state during echocardiography.⁷ Accurate and timely recognition of severe AS is imperative, since it is a swiftly progressive disorder, with procedural intervention being the only effective treatment.¹

Previous studies have shown that conventional echocardiographic techniques underestimate the severity of AS in patients with AF, potentially delaying diagnosis and timely intervention.^5,6 The flow independent aortic valve calcium score (AVCS) is endorsed by guidelines to help gauge severity in patients with low cardiac output and discordant echo parameters such as aortic valve (AV) area and mean gradient (MG).⁸ There is evidence to suggest that patients with AF and low flow low gradient AS (LGAS) have a higher AVCS than their counterparts with sinus rhythm (SR), which could be grounds to rethink our approach to patients with AS who have AF and how to best assess the severity of AS in this population.⁹ Indeed, patients with AF and severe AS experience excess mortality following transcatheter AV replacement (TAVR) compared with patients with SR, perhaps owing to delayed diagnosis and more advanced disease at the time of TAVR, as evidenced by a much higher AVCS. In this study, we aim to (i) characterize AVCS in patients with severe AS and AF compared with those with SR at the time of TAVR, (ii) examine the potential time lag to TAVR in AF compared with SR based on the rate of progression of AVCS, and (iii) compare outcomes after TAVR between these two cohorts.

Methods

Study population

Patients who underwent TAVR from 2013 to 2017 for native valve severe AS with an LV ejection fraction (LVEF) ≥50% and AV area ≤1 cm² were identified from the Tri Site Mayo Clinic Echocardiographic Laboratory database. The patients were divided according to rhythm (SR vs. AF) at the time of echocardiography. Those in paced rhythms during the echocardiogram were excluded. Electrocardiograms of patients with AF during echocardiography were reviewed to avoid misclassification of irregular rhythms that could be mistaken for AF, including SR with marked sinus arrhythmia or frequent premature beats. Additionally, we excluded patients without AVCS. The patients were further categorized by sex, AV area, and MG.

Baseline clinical characteristics and AVCS were collected from the medical records of the patients. Only individuals who granted permission for the use of their medical records for research were included. Vital status and last follow-up were determined from the medical records and Mayo Clinic registration system. The study was approved by the Mayo Clinic Institutional Review Board.

Echocardiography

Echocardiography was performed according to current guidelines, with interrogation of the AV from at least three imaging windows and the use of a non-imaging transducer.¹⁰ Stroke volume was determined from the product of the LV outflow tract (LVOT) area and LVOT pulsed-wave time velocity integral (TVI). The AV area based on TVI was calculated from the Doppler continuity equation. The AV area and dimensionless index of the TVI ratio (LVOT TVI/AV TVI) from the original echocardiogram report were based on averaging representative signals from at least three consecutive cardiac cycles for patients with SR and five consecutive cardiac cycles for patients with AF. Ectopic beats were suspected when the ventricular depolarization (QRS) complex showed a significant change from baseline. Cardiac chamber size and function were determined according to guidelines.¹¹ All echocardiographic images were stored digitally (ProSolv CardioVascular, Inc., Indianapolis, IN, USA) and were retrievable for review.

Computed tomography

Computed tomography (CT) AV calcification was defined as at least four contiguous pixels with a CT density of 130 Hounsfield units or greater. AVCS was reported in arbitrary units (AUs), as described by Agatston et al.¹². AVCS was obtained during routine clinical practice. Sex-specific AVCS thresholds for severe AS are ≥2000 AU for males and ≥1200 AU for females.¹³ Patients were grouped into categories of low AVCS (<2000 AU in men, <1200 AU in women), high AVCS (2000–3000 AU in men, 1200–1600 AU in women), and very high AVCS (>3000 AU in men, >1600 AU in women). In order to examine for the potential time lag to TAVR due to AF, a projected 2-year antedated AVCS in patients with AF was derived by subtracting 304 AU from AVCS at the time of TAVR, assuming an average progression of 152 AU/year.^14,15 A hypothetical time lag of 2 years was chosen based on the nearly 2-year difference in age at TAVR referral among patients with AF and SR in our cohort.

Statistical analysis

Nominal variables were summarized by frequency and percentage. Continuous variables were summarized using median and interquartile range or mean and standard deviation. Univariate tests were conducted to compare patients with AF and SR using the Wilcoxon rank-sum test or two-sample t-test for continuous variables and Fisher’s exact test or χ² test for categorical variables. Kaplan–Meier (KM) estimates were used to estimate the median overall survival. Cox proportional hazards regression models were fit to identify predictors of mortality; variables with a P-value <0.10 in the univariate model were included in the multivariate model. A P-value of <0.05 was considered significant. As part of a sensitivity analysis, we replicated the analysis using a 1:1 matched cohort based on sex and AVCS. R Statistical Software version 4.1.2 was used for the analysis.

Results

See Graphical abstract for a summary of the results.

Demographics and clinical characteristics

A total of 820 patients met inclusion criteria; SR was present in 464 patients (57%) and AF was present in 356 patients (43%). The baseline clinical characteristics of the entire cohort grouped by rhythm are given in Table 1. The mean age was 81.3 years and 41.6% were female. Patients with AF were older (82.2 vs. 80.5 years, P = 0.003). Patients with SR had a higher AV MG (45 vs. 42 mmHg, P = 0.002) and LVEF (59 vs. 55%, P < 0.001). However, there was no significant difference in the indexed valve area between the two groups (SR 0.44 vs. AF 0.43, P = 0.17). Figure 1 shows the AV MG and indexed valve area for patients with AF and SR.

Aortic valve calcium score

Stratified by sex, the AVCS for patients with AF was significantly higher among both males (AF 2850.0 vs. SR 2561.0 AU, P = 0.044) and females (AF 1942.0 vs. 1610.5 AU, P = 0.025) by ∼300 AU, as shown in Figure 2. There was also a significant difference in the calcium score group (low, high, and very high), as shown in Figure 3. Figure 4A and B compare the observed and projected calcium scores for patients with AF and SR by sex.

Determinants of mortality following TAVR

The median follow-up time was 6.6 years (confidence interval 6.41–6.89) for our cohort. Figure 5 shows the KM curves by rhythm status. There was a significant difference in the overall survival by rhythm status (P < 0.0001), where the median survival is longer for patients with SR (5.4 vs. 3.2 years).

Table 2 presents the univariate and multivariate regression results for overall survival. In the univariate analysis, AF, male sex, higher calcium score, dialysis, lower stroke volume index (SVI), lower LVEF, lower AV MG, moderate to severe to severe TR and MR, higher right ventricular systolic pressure (RVSP), and higher left atrial (LA) volume were found to be independent predictors of mortality. In the multivariate analysis, AF, diabetes, dialysis, lower AV MG, moderate to severe to severe TR, and higher RVSP were found to be significant predictors of mortality after TAVR.

Age- and AVCS-matched comparison

The demographics of the sex- and AVCS-matched cohort are displayed in Table 3. A total of 108 patients with SR were excluded, resulting in a cohort of 356 patients with SR and 356 patients with AF. There was still a significant difference in numerous baseline clinical characteristics, including hypertension, ejection fraction, AV MG, and MR/TR severity. Figure 6 shows the KM curves by rhythm status for the sex- and AVCS-matched cohort. There continued to be a difference in the overall survival by rhythm status (P < 0.0001), where the median survival was longer for patients with SR (5.4 vs. 3.2 years).

Discussion

This study showed that patients with SR had significantly higher AV MGs than those with AF at the time of TAVR, despite having a similar valve area. However, AVCSs were significantly higher in both males and females with AF at the time of TAVR compared with their counterparts with SR. Furthermore, there were more patients in the AF group who were classified to have very high AVCS (>3000 AU units for males and >1600 AU in females). It is worth noting that patients with AF were around 2 years older than those with SR when referred for TAVR, which approximately matched the 300 AU higher AVCS in the AF population, further suggesting that there may be a systematic delay in diagnosing severe AS and subsequent referral for TAVR in patients with AF. Overall, long-term survival after TAVR was worse in patients with AF compared with those with SR. In the multivariate regression analysis, AF, along with other markers of advanced structural disease, including moderate and greater TR and high RVSP, were significant predictors of mortality. In a 1:1 sex- and AVCS-matched cohort of patients with AF and SR, long-term survival was still worse in patients with AF, suggesting that rhythm status is a potent predictor of survival for the same degree of AS. Interestingly, we found that a lower MG before the procedure was associated with higher mortality after TAVR even after adjusting for other variables, suggesting that more discordance between haemodynamic and anatomic valve features (a lower MG for similar or higher AVCS as seen in AF) is associated with worse outcomes.

Severe AS is characterized by a peak velocity ≥4 m/s or MG ≥40 mmHg,^16,17 but it is also recognized that both transaortic velocity and MG vary depending on the degree of transaortic forward flow. Thus, the category LGAS describes patients in low forward flow states with a low MG who experience true afterload mismatch from severe AS.⁸ LV systolic dysfunction as a cause for LGAS can be further assessed using dobutamine stress echocardiography.⁸ However, LGAS associated with normal LV function is more challenging, and in patients with AF, is compounded by a loss of atrial contraction, irregular ventricle cycle lengths, diastolic dysfunction, and concomitant valvular regurgitations.^18,19 MG was higher among patients with SR compared with patients with AF across the spectrum of AV areas in our study cohort, which points to the low overall forward flow associated with AF rhythm, as evidenced by a lower SVI.

We found that AVCS was significantly higher in patients with AF compared with SR at the time of TAVR despite a lower overall MG, which suggests that due to flow alterations in AF, conventional MG assessments may be less sensitive in identifying severe AS in these patients. AVCS has emerged as a powerful tool in aiding to parse out severe AS in scenarios with discrepant echocardiographic parameters.⁸ This modality has the advantage of being flow independent, easily reproducible, and highly associated with the rate of AS progression.²⁰ Indeed, by considering AVCS in a cohort of 1541 patients with preserved ejection fraction and a small AV area, Alsidawi et al.⁹ found a high prevalence of anatomically severe AS among patients with AF and a low MG. The authors also showed that the reliance on averaged MG per guidelines may actually underestimate AS in patients with AF compared with their method of recording the single highest MG signal.^5,9 Thus, measurement of AVCS in conjunction with the single highest MG signal has high utility in patients with AF in whom severe AS is suspected, despite a low-averaged MG.

Our finding that patients with AF were referred to TAVR at an older age and those with a much higher AVCS compared with SR underscores the limitations of monitoring AS progression with MG and suggests a possible systematic delay in the diagnosis of severe AS in patients with AF. Tracking of AS progression by calcium score has recently shown promise and is increasingly under study.^14,15,21 Doris et al. demonstrated that AVCS strongly correlated with peak velocity and MG measured by echocardiography. Moreover, the authors reported that AVCS displayed a greater progression to measurement repeatability ratio compared with conventional echocardiographic parameters, which highlights its aptitude for reliably detecting small changes in severity over time. The average AVCS at the time of TAVR for men and women with AF in our study were 2850 and 1942 AU, respectively. We calculated retrospective calcium scores for patients with AF, which extrapolated AVCS 2 years prior to TAVR referral based on the 2-year age difference between patients with AF vs. SR at the time of TAVR and the 152 AU/year rate of progression published by Doris et al. We found that projected calcium score in patients with AF 2 years prior to TAVR referral was nearly identical to AVCS in patients with SR at the time of TAVR. This finding lends supports to the notion of time lag to diagnosis of severe AS in AF, which may in turn influence outcomes.

In our study cohort, AF was associated with significantly worse overall survival following TAVR in comparison with SR, which aligns with prior literature.^4,9,22 Moreover, we found that AF was associated with an increased risk of mortality after adjusting for sex, AVCS, SVI, LVEF, MR and TR, and LA volume. To examine whether the increase in mortality among patients with AF was driven by rhythm status or the extent of AV disease, which we represented as absolute AVCS, we created a 1:1 sex- and AVCS-matched cohort of AF and SR and repeated the survival analysis. Interestingly, after matching for AVCS, patients with AF still had worse survival than patients with SR.

The relationship between the pathologic processes underlying AS and AF is both complex and dynamic. Progressive, unoperated AS leads to altered myocardial contractile function, progressive LV concentric remodelling, impaired LV filling, and diastolic dysfunction.^1,23 In our analysis, structural changes associated with pre-existing AF, such as moderate-severe to severe TR and elevated RVSP,⁴ were significant predictors of mortality. Moreover, previous studies have shown that the presence of preoperative MR in patients undergoing TAVR is more frequent with pre-existing AF and is associated with increased postoperative mortality.^24,25 It is feasible that the presence of AF denotes more advanced structural and myocardial disease, which culminates in worse outcomes from more pronounced pump dysfunction. Indeed, data from the FRANCE-2 study among others suggest that systolic and diastolic heart failure are particularly implicated in excess mortality in patients with AF undergoing TAVR.^4,26,27

Clinical application and future prospective

From a clinical practice standpoint, our research highlights a notable discrepancy between haemodynamic and anatomic assessment of AS in the setting of AF. Anatomically, severe AS is present when the AVCS is ≥2000 in men and ≥1200 in women, as outlined in the guidelines. These cut points were established in patients, of whom around a quarter had AF and the rest had SR.²⁸ Forward flow is lower in AF compared with SR and therefore for the same anatomic AS severity (i.e. same AVCS), the MG will be lower in patients with AF leading to underestimation of AS severity by Doppler MG and delay in referral to AVR.⁹ Consequently, establishing intervention cut-offs for AS based solely on MG in AF becomes challenging. In this study, we are demonstrating that the 2-year gap in age at the time of TAVR matches a 2-year gap in AVCS which means patients with AF should have had AVR 2 years earlier. It has been shown previously that patients with AF are referred less often to AVR because of the higher proportion of LGAS and because clinicians often blame AF for symptoms. Therefore, we propose that when AV area falls below 1 cm² in AF, prioritizing assessment and monitoring of AV stenosis severity through AVCS, which is independent of flow, becomes more important than relying on mean systolic gradient. Unfortunately, we do not have AVCS data in a large cohort of patients with AF and varying degrees of AS (mild, moderate, and severe) by Doppler echocardiography to attempt to establish new cut points MG or AVCS in AF associated with increased risk, so additional studies are needed to examine whether a different set of criteria for intervention are needed for patients with AF and AS.

Limitations

This was a retrospective study with data pooled from three tertiary referral centres. Patients with LVEF <50% were excluded. We observed all-cause mortality and not cardiovascular-specific mortality; however, in this cohort of elderly patients with multiple comorbidities, competing risks for mortality likely confound true cardiovascular death. In calculating the retrospective calcium score, we used the annual rate of progression of AVCS, which assumes linear disease progression, and, therefore, does not account for rapid disease progression in severe cases, although this is consistent with previous studies of disease progression in AS.

Conclusion

Patients with AF had a higher AVCS and worse overall long-term mortality following TAVR compared with patients with SR. This may in part be due to a delay in the diagnosis of severe AS in patients with AF because of altered haemodynamics and low transaortic flow leading to a lower MG despite a small AV area. Therefore, patients with AF likely receive TAVR at more advanced stages of the disease with a lower potential for reverse remodelling following intervention. Although AVCS does not predict outcomes independently of rhythm status, the routine use of AVCS may help to identify patients with severe AS and AF earlier in the disease process before advanced structural heart disease takes hold. Future studies should prospectively assess the benefit of earlier intervention in this population when AVCS meets the guideline criteria for severe AS.

Funding

None declared.

Data availability

The data underlying this article were accessed from the Mayo Clinic Echocardiographic Laboratory Database. The derived data generated in this research will be shared on reasonable request to the corresponding author.

References

Author notes