Biatrial maze procedure versus pulmonary vein isolation for atrial fibrillation during mitral valve surgery: New analytical approaches and end points

OBJECTIVE

To use novel statistical methods for analyzing the effect of lesion set on (long-standing) persistent atrial fibrillation (AF) in the Cardiothoracic Surgical Trials Network trial of surgical ablation during mitral valve surgery (MVS).

METHODS

Two hundred sixty such patients were randomized to MVS + surgical ablation or MVS alone. Ablation was randomized between pulmonary vein isolation and biatrial maze. During 12 months postsurgery, 228 patients (88%) submitted 7949 transtelephonic monitoring (TTM) recordings, analyzed for AF, atrial flutter (AFL), or atrial tachycardia (AT). As previously reported, more ablation than MVS-alone patients were free of AF or AF/AFL at 6 and 12 months (63% vs 29%; P < .001) by 72-hour Holter monitoring, without evident difference between lesion sets (for which the trial was underpowered).

RESULTS

Estimated freedom from AF/AFL/AT on any transmission trended higher after biatrial maze than pulmonary vein isolation (odds ratio, 2.31; 95% confidence interval, 0.95-5.65; P = .07) 3 to 12 months postsurgery; estimated AF/AFL/AT load (ie, proportion of TTM strips recording AF/AFL/AT) was similar (odds ratio, 0.90; 95% confidence interval, 0.57-1.43; P = .6). Within 12 months, estimated prevalence of AF/AFL/AT by TTM was 58% after MVS alone, and 36% versus 23% after pulmonary vein isolation versus biatrial maze (P < .02).

CONCLUSIONS

Statistical modeling using TTM recordings after MVS in patients with (long-standing) persistent AF suggests that a biatrial maze is associated with lower AF/AFL/AT prevalence, but not a lower load, compared with pulmonary vein isolation. The discrepancy between AF/AFL/AT prevalence assessed at 2 time points by Holter monitoring versus weekly TTM suggests the need for a confirmatory trial, reassessment of definitions for failure after ablation, and validation of statistical methods for assessing atrial rhythms longitudinally.

Evolution of association between renal and liver functions while awaiting heart transplant: An application using a bivariate multiphase nonlinear mixed effects model

In many longitudinal follow-up studies, we observe more than one longitudinal outcome. Impaired renal and liver functions are indicators of poor clinical outcomes for patients who are on mechanical circulatory support and awaiting heart transplant. Hence, monitoring organ functions while waiting for heart transplant is an integral part of patient management. Longitudinal measurements of bilirubin can be used as a marker for liver function and glomerular filtration rate for renal function. We derive an approximation to evolution of association between these two organ functions using a bivariate nonlinear mixed effects model for continuous longitudinal measurements, where the two submodels are linked by a common distribution of time-dependent latent variables and a common distribution of measurement errors.

Low pretransplant IgA level is associated with early post-lung transplant seromucous infection

OBJECTIVES

Infection is an important cause of morbidity and mortality after lung transplantation. Immunoglobulins are part of both seromucous (IgA) and serum (IgG) infection defense mechanisms. We therefore hypothesized that lower pretransplant IgA levels would be associated with more early post-lung transplant seromucous infections and greater mortality independent of IgG.

METHODS

From January 2000 to July 2010, 538 patients undergoing primary lung transplantation had pretransplant IgA (n = 429) and IgG (n = 488) measured as a clinical routine. Median IgA was 200 mg·dL^-1 (2% < 70 mg·dL^-1, lower limit of normal); median IgG was 970 mg·dL^-1 (5% < 600 mg·dL^-1). Intensive microbiology review was used to categorize infections and their causative organisms within the first posttransplant year.

RESULTS

In total, 397 seromucous infections were observed in 247 patients, most bacterial. Although IgA and IgG were moderately correlated (r = 0.5, P < .0001), low pretransplant IgA was a strong risk factor (P = .01) for seromucous infections, but pretransplant IgG was not (P ≥ .6). As pretransplant IgA levels fell below 200 mg·dL^-1, the risk of these posttransplant infections rose nearly linearly. Lower pretransplant levels of IgA were associated with greater posttransplant mortality to end of follow-up (P = .004), but pretransplant IgG was not (P ≥ .3).

CONCLUSIONS

Low levels of preoperative IgA, an important immunoglobulin involved in mucosal immunologic defense, but not IgG, are associated with seromucous infections in the year after lung transplantation and increased follow-up mortality. It would appear prudent to identify patients with relative IgA deficiency at listing and to increase vigilance of monitoring for, and prophylaxis against, seromucous infection in this high-risk population.

Probability of atrial fibrillation after ablation: Using a parametric nonlinear temporal decomposition mixed effects model

Atrial fibrillation is an arrhythmic disorder where the electrical signals of the heart become irregular. The probability of atrial fibrillation (binary response) is often time varying in a structured fashion, as is the influence of associated risk factors. A generalized nonlinear mixed effects model is presented to estimate the time-related probability of atrial fibrillation using a temporal decomposition approach to reveal the pattern of the probability of atrial fibrillation and their determinants. This methodology generalizes to patient-specific analysis of longitudinal binary data with possibly time-varying effects of covariates and with different patient-specific random effects influencing different temporal phases. The motivation and application of this model is illustrated using longitudinally measured atrial fibrillation data obtained through weekly trans-telephonic monitoring from an NIH sponsored clinical trial being conducted by the Cardiothoracic Surgery Clinical Trials Network.

Boosted Multivariate Trees for Longitudinal Data

Machine learning methods provide a powerful approach for analyzing longitudinal data in which repeated measurements are observed for a subject over time. We boost multivariate trees to fit a novel flexible semi-nonparametric marginal model for longitudinal data. In this model, features are assumed to be nonparametric, while feature-time interactions are modeled semi-nonparametrically utilizing P-splines with estimated smoothing parameter. In order to avoid overfitting, we describe a relatively simple in sample cross-validation method which can be used to estimate the optimal boosting iteration and which has the surprising added benefit of stabilizing certain parameter estimates. Our new multivariate tree boosting method is shown to be highly flexible, robust to covariance misspecification and unbalanced designs, and resistant to overfitting in high dimensions. Feature selection can be used to identify important features and feature-time interactions. An application to longitudinal data of forced 1-second lung expiratory volume (FEV1) for lung transplant patients identifies an important feature-time interaction and illustrates the ease with which our method can find complex relationships in longitudinal data.

Estimating the prevalence of atrial fibrillation from a three-class mixture model for repeated diagnoses

Atrial fibrillation (AF) is an abnormal heart rhythm characterized by rapid and irregular heartbeat, with or without perceivable symptoms. In clinical practice, the electrocardiogram (ECG) is often used for diagnosis of AF. Since the AF often arrives as recurrent episodes of varying frequency and duration and only the episodes that occur at the time of ECG can be detected, the AF is often underdiagnosed when a limited number of repeated ECGs are used. In studies evaluating the efficacy of AF ablation surgery, each patient undergoes multiple ECGs and the AF status at the time of ECG is recorded. The objective of this paper is to estimate the marginal proportions of patients with or without AF in a population, which are important measures of the efficacy of the treatment. The underdiagnosis problem is addressed by a three-class mixture regression model in which a patient's probability of having no AF, paroxysmal AF, and permanent AF is modeled by auxiliary baseline covariates in a nested logistic regression. A binomial regression model is specified conditional on a subject being in the paroxysmal AF group. The model parameters are estimated by the Expectation-Maximization (EM) algorithm. These parameters are themselves nuisance parameters for the purpose of this research, but the estimators of the marginal proportions of interest can be expressed as functions of the data and these nuisance parameters and their variances can be estimated by the sandwich method. We examine the performance of the proposed methodology in simulations and two real data applications.

Insights Into Timing, Risk Factors, and Outcomes of Stroke and Transient Ischemic Attack After Transcatheter Aortic Valve Replacement in the PARTNER Trial (Placement of Aortic Transcatheter Valves)

Background—

Prior studies of stroke and transient ischemic attack (TIA) after transcatheter aortic valve replacement (TAVR) are limited by reporting and follow-up variability. This is a comprehensive analysis of time-related incidence, risk factors, and outcomes of these events.

Methods and Results—

From April 2007 to February 2012, 2621 patients, aged 84±7.2 years, underwent transfemoral (TF; 1521) or transapical (TA; 1100) TAVR in the PARTNER trial (Placement of Aortic Transcatheter Valves; as-treated), including the continued access registry. Stroke and TIA were identified by protocol and adjudicated by a Clinical Events Committee. Within 30 days of TAVR, 87 (3.3%) patients experienced a stroke (TF 58 [3.8%]; TA 29 [2.7%]; P =0.09), 85% within 1 week. Instantaneous stroke risk peaked on day 2, then fell to a low prolonged risk of 0.8% by 1 to 2 weeks. Within 30 days, 13 (0.50%) patients experienced a TIA (TF 10 [0.67%]; TA 3 [0.27%]; P >0.17). Stroke and TIA were associated with lower 1-year survival than expected (TF 47% after stroke versus 82%, and 64% after TIA versus 83%; TA 53% after stroke versus 80%, and 64% after TIA versus 83%). Risk factors for early stroke after TA-TAVR included more postdilatations, pure aortic stenosis without regurgitation, and possibly more pacing runs, earlier date of procedure, and no dual antiplatelet therapy; high pre-TAVR aortic peak gradient was a risk factor for stroke early after TF-TAVR.

Conclusions—

Risk of stroke or TIA is highest early after TAVR and is associated with increased 1-year mortality. Modifications of TAVR, emboli-prevention devices, and better intraprocedural pharmacological protection may mitigate this risk.

Clinical Trial Registration—

URL: http://www.clinicaltrials.gov . Unique identifier: NCT00530894.

A multiphase non-linear mixed effects model: An application to spirometry after lung transplantation

In medical sciences, we often encounter longitudinal temporal relationships that are non-linear in nature. The influence of risk factors may also change across longitudinal follow-up. A system of multiphase non-linear mixed effects model is presented to model temporal patterns of longitudinal continuous measurements, with temporal decomposition to identify the phases and risk factors within each phase. Application of this model is illustrated using spirometry data after lung transplantation using readily available statistical software. This application illustrates the usefulness of our flexible model when dealing with complex non-linear patterns and time-varying coefficients.

Surgical Versus Percutaneous Femoral Access for Delivery of Large-Bore Cardiovascular Devices (from the PARTNER Trial)

It is unclear if surgical exposure confers a risk advantage compared with a percutaneous approach for patients undergoing endovascular procedures requiring large-bore femoral artery access. From the randomized controlled Placement of Aortic Transcatheter Valve trials A and B and the continued access registries, a total of 1,416 patients received transfemoral transcatheter aortic valve replacement, of which 857 underwent surgical, and 559 underwent percutaneous access. Thirty-day rates of major vascular complications and quality of life scores were assessed. Propensity matching was used to adjust for unmeasured confounders. Overall, there were 116 major vascular complications (8.2%). Complication rates decreased dramatically during the study period. In unadjusted analysis, major vascular complications were significantly less common in the percutaneous access group (35 [6.3%] vs 81 [9.5%] p = 0.032). However, among 292 propensity-matched pairs, there was no difference in major vascular complications (22 [7.5%] vs 28 [9.6%], p = 0.37). Percutaneous access was associated with fewer total in-hospital vascular complications (46 [16%] vs 66 [23%], p = 0.036), shorter median procedural duration (97 interquartile range [IQR 68 to 166] vs 121 [IQR 78 to 194] minutes, p <0.0001), and median length of stay (4 [IQR 2 to 8] vs 6 [IQR 3 to 10] days, p <0.0001). There were no significant differences in quality of life scores at 30 days. Surgical access for large-bore femoral access does not appear to confer any advantages over percutaneous access and may be associated with more minor vascular complications.

Learning curves for transapical transcatheter aortic valve replacement in the PARTNER-I trial: Technical performance, success, and safety

OBJECTIVES

Introduction of hybrid techniques, such as transapical transcatheter aortic valve replacement (TA-TAVR), requires skills that a heart team must master to achieve technical efficiency: the technical performance learning curve. To date, the learning curve for TA-TAVR remains unknown. We therefore evaluated the rate at which technical performance improved, assessed change in occurrence of adverse events in relation to technical performance, and determined whether adverse events after TA-TAVR were linked to acquiring technical performance efficiency (the learning curve).

METHODS

From April 2007 to February 2012, 1100 patients, average age 85.0 ± 6.4 years, underwent TA-TAVR in the PARTNER-I trial. Learning curves were defined by institution-specific patient sequence number using nonlinear mixed modeling.

RESULTS

Mean procedure time decreased from 131 to 116 minutes within 30 cases (P = .06) and device success increased to 90% by case 45 (P = .0007). Within 30 days, 354 patients experienced a major adverse event (stroke in 29, death in 96), with possibly decreased complications over time (P ∼ .08). Although longer procedure time was associated with more adverse events (P < .0001), these events were associated with change in patient risk profile, not the technical performance learning curve (P = .8).

CONCLUSIONS

The learning curve for TA-TAVR was 30 to 45 procedures performed, and technical efficiency was achieved without compromising patient safety. Although fewer patients are now undergoing TAVR via nontransfemoral access, understanding TA-TAVR learning curves and their relationship with outcomes is important as the field moves toward next-generation devices, such as those to replace the mitral valve, delivered via the left ventricular apex.