Abstract
Permanent pacemaker implantation (PPMI) is an important complication following transcatheter aortic valve replacement (TAVR). The influence of valvular and subvalvular calcium and its distribution between aortic leaflets on the risk of PPMI following TAVR remains unclear. We performed a systematic review of the aortic valve complex (AVC) calcium by leaflet, left ventricular outflow tract (LVOT) calcium by leaflet, total AVC calcium, total LVOT calcium, and mitral annular calcium and its association with post-TAVR atrioventricular block, left bundle branch block, and new PPMI. The search strategy included five databases identifying 893 articles. A total of 34 studies with 11,528 patients were included for qualitative analysis, and seven studies totaling 1056 patients were suitable for quantitative analysis. On meta-analysis, left coronary cusp calcium and right coronary cusp calcium were significant predictors of PPMI, while noncoronary cusp (NCC) calcium was not predictive (left coronary cusp: mean difference: 21.05 mm 3 , 95% CI: 5.92-36.19, p < 0.001; right coronary cusp: mean difference: 46.02 mm 3 , 95% CI: 1.84-90.21, p = 0.04, and NCC: mean difference: 0.19 mm 3 , 95% CI: −0.32 to 0.50, p = 0.10). On qualitative review, LVOT calcium in the NCC region was the leaflet most commonly predictive of post-TAVR conduction outcomes. Total AVC, total LVOT calcium, and mitral annular calcium had no convincing association with post-TAVR conduction outcomes. The distribution of calcium rather than its total volume was associated with post-TAVR conduction abnormalities. Heterogeneity in methodology and implantation techniques between studies limits the clinical significance of these findings.
Highlights
- •
No prior review has considered all calcium distributions and post-transcatheter aortic valve replacement (TAVR) conduction.
- •
Left and right coronary cusp calcium had an increased risk of pacemaker after TAVR.
- •
In the left ventricular outflow tract, noncoronary calcium had increased permanent pacemaker rates.
- •
Calcium in any distribution did not predict left bundle branch block after TAVR.
- •
Accurate risk assessment of patients can help guide preprocedural planning.
Introduction
Transcatheter aortic valve replacement (TAVR) is an established treatment for patients with symptomatic aortic stenosis. High-grade atrioventricular block (HGAVB), specifically Mobitz II, and third-degree atrioventricular (AV) block have an incidence of 10% to 21% following TAVR and often require permanent pacemaker implantation (PPMI). It has been proposed that the transcatheter heart valve (THV), when deployed may exert a unidirectional mechanical pressure onto the AV conduction system and contribute to new HGAVB after TAVR. High calcium burden in the aortic valve complex (AVC) and left ventricular outflow tract (LVOT) may compound mechanical pressures exerted by the THV and increase the risk of HGAVB. Resolution of initial postprocedural conduction abnormalities in the days after TAVR for some patients supports other theories whereby postprocedural edema and inflammation leads to a transient AV conduction impairment. There is conjecture regarding the individual associations of noncoronary cusp (NCC), right coronary cusp (RCC), and left coronary cusp (LCC) calcium volumes and conduction abnormalities after TAVR. No specific distribution of calcium between leaflets has been established as the highest risk for PPMI requirement after TAVR. Existing studies present mixed findings with no consensus. There has also been recent literature to suggest mitral annular calcification (MAC), while not directly part of the AVC, may contribute to conduction abnormalities after TAVR. It is hypothesized that higher volumes of MAC may increase the risk of conduction abnormalities due to its proximity to the anterior aspect of the AV node. A comprehensive systematic review and meta-analysis of the latest available data has not yet been performed.
Materials and Methods
We performed a systematic review and meta-analysis of published studies that investigated whether there was an association between the distribution of calcium within the AVC and LVOT and the development of HGAVB, left bundle branch block (LBBB), or PPMI after TAVR. As a secondary aim, we performed a systematic review of studies investigating total AVC calcium, total LVOT calcium, and MAC and its association with HGAVB, LBBB, or PPMI after TAVR.
The systematic review and meta-analysis were reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. The review was registered with the International Prospective Register of Systematic Reviews (PROSPERO) (ID CRD42023417976).
Study Selection
A broad search strategy was developed ( Supplementary Figure 1 ) to identify studies that examined preprocedural AVC and/or LVOT calcium and its association with postprocedural conduction disturbances or PPMI in patients after TAVR. Five databases were included: EMBASE, MEDLINE, PUBMED, Scopus, and Wed of Science. The search was performed on 17 April 2023, and restricted to publications available in the English language.
Independent reviewers (P.L. and K.R.) performed title and abstract screening. Potentially eligible articles were reviewed in full text for final inclusion, with disagreements resolved by consensus with a third independent reviewer (R.B.). Abstracts, conference presentations, and review articles were excluded. Studies from 2014 and prior were excluded due to use of older generation THVs no longer commercially available.
Data Extraction and Synthesis
A standardized data collection template was used for data extraction, including recruitment timing, sample size, mean age, sex, bicuspid valve incidence if reported and other baseline demographics, existing conduction abnormalities (left and right bundle branch block [LBBB, RBBB], prior permanent pacemaker [PPM], prior first-degree atrioventricular block [AVB]), type of valve used, calcium volumes and distributions including preset Hounsfield unit (HU) thresholds for measurement and methodology of measurement, outcome measure, follow-up period, whether conduction was the primary outcome, and study findings.
Meta-analysis was performed on studies with quantitative data that was compatible with forest plot creation. Studies with calcium measurements separated based on PPMI status and presented in mean and SDs were able to be included in forest plot analysis. Studies with data presented in medians and interquartile ranges were first assessed for normal distribution or significant skew. Studies with a normal distribution of data were converted to mean and SD values, and studies with skewed data were excluded from meta-analysis. In the case of one study, where calcium measurements were reported for the PPMI group and the whole cohort, raw data values for the non-PPMI group were manually calculated.
Quality of Included Studies and Risk of Bias
The quality of included studies and risk of bias were independently assessed using the Newcastle Ottawa Scale for cohort studies. Studies were graded out of a maximum possible score of 9 and were classified as “Poor” quality (0-2 points), “Fair” quality (3-5 points) or “Good/High” quality (6-9 points).
Statistical Analysis
Data analysis was conducted in Review Manager Version 5.4.1. Mean differences in calcium volumes were compared using random effect and generic inverse variance models. Significance was set as p ≤ 0.05. Interstudy heterogeneity (I 2 statistic) was classified as low (I 2 < 25%), moderate (I 2 : 25%-50%), or high (I 2 > 50%).
Results
Study Selection
Our search strategy yielded a total of 893 articles, and an additional seven articles were identified manually. After duplicates were removed (n = 143), a total of 757 articles were retrieved for title and abstract screening. Seventy articles were determined eligible for full text review. A total of 34 studies were included after full text review for qualitative analysis. The other 36 studies were excluded in full text review for the following reasons: surgical valve, not TAVR (n = 2), abstract or conference presentation (n = 2), study date of 2014 or earlier (n = 2), calcium was not a studied predictor (n = 7), conduction was not a studied outcome (n = 16), conduction was reported within a composite morbidity measure but not individually reported (n = 5), and both calcium and conduction data were reported but not analyzed (n = 2).
Quantitative meta-analysis was able to be performed on seven studies. The other 27 studies were incompatible with meta-analysis for the following reasons: AVC calcium measurements by leaflets were not available (n = 18), calcium data grouped based on conduction outcome were not available (n = 6), and study data were significantly skewed, prohibiting the accurate derivation of mean and SD (n = 3). Study selection is summarized in Figure 1 .
Study Characteristics
Thirty-four studies totaling 11,528 patients were included for review. Characteristics for all included studies were summarized (see Results tables). Pre-existing RBBB was controlled for as a confounder variable in 15 studies, excluded from the study cohorts in two studies, and had no explicit mention in 17 studies. PPMI was the primary conduction outcome in 30 (88%) studies. A subset of 10 studies included the following additional conduction outcomes: LBBB (n = 10), PR and QRS prolongation (n = 2), HGAVB (n = 7), or PPMI dependence for right ventricular pacing at one month follow-up (n = 1). Patients with prior PPM were excluded from the study cohort in 24 studies, not excluded from the study cohort in 6 studies, and had no explicit mention in four studies. All included studies derived their calcium measurements from computed tomography images.
A subset of seven studies totaling 1056 patients were included in the meta-analysis, and detailed patient characteristics of this cohort are summarized in Table 1 . Specifically, all seven studies excluded patients with a prior PPM, and the incidence of pre-existing RBBB was 8.3%. The split of valve types used were mostly balloon-expandable valves (45.9%), followed by self-expandable valves (36.6%), and a smaller cohort of patients who received the no longer commercially available, mechanically expandable valves (17.3%).
| Pooled | Fujita, 2016 | Gonska, 2017 | Kebler, 2017 | Mauri, 2018 | Schaefer, 2021 | Sharma, 2020 | Veulemans, 2020 | |
|---|---|---|---|---|---|---|---|---|
| Sample size | 1056 | 162 | 283 | 183 | 194 | 79 | 25 | 130 |
| Baseline demographics | ||||||||
| Age (y) (M ± SD) | 81.4 ± 6.0 | 82 ± 5 | 79.9 ± 6.2 | 81.1 ± 5.1 | 82.1 ± 5.2 | 79.8 ± 7.1 | 81.6 ± 7.5 | 80.4 ± 6.2 |
| Sex (female) (n [%]) | 559 (52.9%) | 96 (59.3%) | 149 (52.7%) | 103 (56.3%) | 156 (73.6%) | 32 (40.5%) | 12 (49.6%) | 11 (8.5%) |
| BMI (M ± SD) | 26.3 ± 6.0 | 26.1 ± 4.7 | 27.5 ± 4.8 | 26.3 ± 8.1 | 27.6 ± 5.5 | 27.2 ± 5.5 | 21.5 ± 7.5 | 26.6 ± 4.0 |
| EuroScore 1 (M ± SD) | 15.4 ± 12.0 | 21.3 ± 10.6 | 14.2 ± 12.8 | 13.0 ± 11.5 | 14.3 ± 10.4 | 12.1 ± 8.8 | ND | 21.9 ± 14.0 |
| STS-PROM (M ± SD) | 6.4 ± 3.0 | 5.7 ± 2.1 | 6.7 ± 3.9 | 6.6 ± 5.0 | ND | ND | ND | ND |
| EF (%) (M ± SD) | 48.3 ± 13.0 | 51 ± 10 | ND | ND | ND | 52 ± 14 | 39.6 ± 13.5 | ND |
| NYHA III/IV (n [%]) | 355 (33.6%) | ND | 233 (82.7%) | ND | ND | 29 (36.7%) | ND | 93 (71.5%) |
| CAD (n [%]) | 650 (61.5%) | 97 (59.9%) | 168 (59.4%) | 108 (59.0%) | 124 (64.0%) | 43 (54.4%) | 19 (76.0%) | 91 (70.0%) |
| HTN (n [%]) | 522 (49.4%) | 144 (88.9%) | ND | ND | 180 (93.0%) | 59 (74.7%) | 20 (80.0%) | 119 (91.5%) |
| DM (n [%]) | 302 (28.6%) | 43 (26.5%) | 89 (29.7%) | 44 (24.0%) | 64 (33.0%) | 13 (16.5%) | 12 (48.0%) | 37 (28.5%) |
| Baseline conduction | ||||||||
| Prior PPM (n [%]) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | ND | 0 (0.0%) | 0 (0.0%) |
| Prior AF (n [%]) | 314 (29.7%) | ND | 116 (41.0%) | 61 (33.3%) | 72 (34.0%) | ND | 8 (32.0%) | 57 (43.8%) |
| Prior RBBB (n [%]) | 88 (8.3%) | 20 (12.3%) | 20 (7.1%) | 13 (7.1%) | 14 (7.2%) | ND | 10 (40.0%) | 11 (8.5%) |
| Prior LBBB (n [%]) | 94 (8.9%) | 20 (12.3%) | 26 (9.3%) | 17 (9.3%) | 15 (7.7%) | ND | 1 (4.0%) | 15 (11.5%) |
| Valve characteristics | ||||||||
| Self-expandable valve (n [%]) | 387 (36.6%) | 63 (38.9%) | 0 (0%) | 0 (0%) | 194 (100.0%) | 0 (0.0%) | 0 (0.0%) | 130 (100.0%) |
| Balloon-expandable valve (n [%]) | 486 (45.9%) | 99 (61.1%) | 283 (100.0%) | 0 (0.0%) | 0 (0.0%) | 79 (100.0%) | 25 (100.0%) | 0 (0.0%) |
| Mechanically expandable valve (n [%]) | 183 (17.3%) | 0 (0.0%) | 0 (0.0%) | 183 (100.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) | 0 (0.0%) |
| Oversizing mean (M ± SD) | 12.1 ± 10.7 | 15.5 ± 6.5 | 7.9 ± 13.5 | 8.4 ± 9.1 | ND | ND | 9.3 ± 13.8 | 23.1 ± 4.9 |
| Implantation depth (M ± SD) | 5.6 ± 2.2 | 5.2 ± 1.5 | 6.7 ± 2.4 | ND | 5.7 ± 2.0 | ND | ND | 5.1 ± 3.0 |
Quality of Included Studies and Risk of Bias Assessment
All studies scored the highest classification “Good/High” quality on Newcastle Ottawa Scale assessment ( Table 2 ). There were 18 out of 34 studies that recorded the maximum possible score of nine. Variation in score was predominately due to some studies not controlling for confounders or including participants who had the outcome measure at baseline prior to TAVR.
| Study | Selection | Comparability | Outcome | Total score | Classification | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Represents exposed cohort | Selection of the nonexposed cohort | Ascertains exposure | Outcome was not present at start of study | Comparison of cohorts | Assessment of outcome | Long enough follow-up | Adequacy of follow-up | |||
| Abramowitz, 2017 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Al-Azzam, 2017 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Ancona, 2017 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 7 | Good/High | |
| Ancona, 2020 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Brodov, 2019 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Dowling, 2022 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 7 | Good/High | |
| Esposito, 2022 | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 8 | Good/High | |
| Fujita, 2016 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Gamet, 2020 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 7 | Good/High | |
| Gonska, 2017 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Hein-Rothweiler, 2017 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Iacovelli, 2021 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 7 | Good/High | |
| Jochheim, 2019 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Kebler, 2017 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Kim, 2018 | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 8 | Good/High | |
| Lak, 2022 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 6 | Good/High | ||
| Maeno, 2017 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Mahon, 2021 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Mauri, 2016 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Mauri, 2018 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Medranda, 2022 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 6 | Good/High | ||
| Milo, 2023 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 6 | Good/High | ||
| Musallam, 2022 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 6 | Good/High | ||
| Oestreich, 2017 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 7 | Good/High | |
| Okuno, 2021 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Piayda, 2020 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Pollari, 2019 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Pollari, 2022 | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 8 | Good/High | |
| Rodriguez-Olivares, 2016 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Schaefer, 2021 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 6 | Good/High | ||
| Sharma, 2020 | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | ∗ | 7 | Good/High | |
| Spaziano, 2018 | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 8 | Good/High | |
| Veulemans, 2020 | ∗ | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 9 | Good/High |
| Waldschmidt, 2022 | ∗ | ∗ | ∗ | ∗∗ | ∗ | ∗ | ∗ | 8 | Good/High | |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
