Methods

The institutional review board of the MedStar Health Research Institute approved the present study. The data from consecutive patients with symptomatic, severe aortic stenosis who underwent either TA or TF TAVR from May 2007 to April 2011 were analyzed. Prespecified clinical and laboratory data were collected for all patients on admission, immediately after TAVR, during the index hospitalization, and for ≤1 year. The collected data included medical history, medications at admission, electrocardiographic findings, echocardiographic findings, and cardiac biomarker levels, including cardiac troponin I (cTnI; upper reference limit 0.045 ng/ml) and CK-MB (upper reference limit 3.6 ng/ml).

For TF TAVR, the femoral artery was accessed percutaneously and was closed using 2 Perclose ProGlide vascular suture-mediated closure devices (Abbott Vascular Devices, Redwood City, California). Next, consecutive upsizing of the femoral access site was performed with serial dilators until the appropriate sheath size (22F or 24F) was reached.

For TA TAVR, a left minithoracotomy was performed, and 2 intramural pledget purse-string apical sutures were placed before apical puncture. In all procedures, aortic valvuloplasty during rapid pacing was performed over an Amplatz Super Stiff guidewire (Boston Scientific, Natick, Massachusetts). Next, the bioprosthetic valve was advanced over the same wire and deployed under rapid pacing with fluoroscopic and transesophageal echocardiographic guidance.

The in-hospital outcomes were collected retrospectively and included the length of intensive care unit stay, blood transfusion, acute kidney injury, stroke, bleeding, and in-hospital death. The intensive care unit stay was defined as the interval from the day of the procedure to the day of transfer to medium care. Any neurologic deficit, together with evidence for stroke in neuroimaging, was categorized as a stroke event.

Statistical analysis was performed using SAS, version 9.1 (SAS Institute, Cary, North Carolina). Continuous variables are expressed as the mean ± SD or median (25th to 75th interquartile range), as appropriate according to the variable distribution. Categorical variables are expressed as percentages. Student’s t test was used to compare continuous variables and the chi-square test or Fisher’s exact test to compare categorical variables.

Receiver operating characteristic curves were analyzed to assess the discriminatory ability of the periprocedural cardiac biomarker levels to predict 30-day mortality and to determine the best cutoff value to predict this end point. Thus, the best prognosticator in the receiver operating characteristic curve analyses was the parameter that gave the greatest product of sensitivity and specificity for predicting 30-day mortality. Survival rates ≤1 year were computed using the Kaplan-Meier method, and differences in the parameters were assessed using the log-rank test.

Multivariate analyses were performed to predict 30-day mortality and the high post-TAVR CK-MB using the Cox regression model. The multivariate analyses included the baseline variables and clinical characteristics associated with the outcome of interest on univariate analysis: baseline left ventricular ejection fraction, presence of coronary artery disease, renal insufficiency at baseline, and β-blocker use on admission. A p value <0.05 was considered statistically significant.

Results

A total of 150 patients underwent TAVR from May 2007 to April 2011. Their average age was 85 ± 6 years, and 44% were men. The patients had an average Society of Thoracic Surgeons score of 12 ± 4 and logistic European System for Cardiac Operative Risk Evaluation (EuroSCORE) of 42 ± 22. TF access was used in 103 patients (69%) and TA access in 47 (31%). The TA patients had greater logistic EuroSCOREs, greater rates of atrial fibrillation at baseline, and more often had peripheral vascular disease; however, the other baseline characteristics were comparable between the 2 groups ( Table 1 ).

As shown in Figure 1 , most TAVR patients had significant increases in serum cardiac biomarkers after the procedure; 95% of the TAVR patients had a twofold increase in cTnI beyond the upper reference limit, but only 50% had such increases in CK-MB. The incidence of a 10-fold increase in serum cardiac markers, corresponding to the Valve Academic Research Consortium definition for periprocedural myocardial infarction, was 87.3% for cTnI elevation and 7.3% for CK-MB elevation. A comparison between the TA and TF patients indicated that the x-fold increase in cTnI (6 ng/ml, range 3.6 to 9.3, vs 1.4 ng/ml, range 0.8 to 2.7, respectively) and CK-MB (180 ng/ml, range 101 to 422, vs 27 ng/ml, range 14 to 52, respectively) beyond upper reference limit was significantly greater for TA patients than for TF patients (p <0.0001; Figure 2 ).

Incidence of elevated serum cardiac biomarkers among patients undergoing TAVR.

Magnitude of elevation in serum cardiac biomarkers among TF and TA patients.

Multivariate analysis to predict 30-day mortality among the entire TAVR population (TF and TA) showed that greater peak cTnI and peak CK-MB levels were the only independent predictors for higher mortality ( Table 2 ). However, on receiver operating characteristic curve analysis, both cardiac biomarkers had low prognostic accuracy to predict 30-day mortality for the entire TAVR population ( Figure 3 ). Similar low prognostic accuracy for CK-MB and cTnI was found for the TA patients ( Figure 3 ). However, for TF patients, the CK-MB levels demonstrated high prognostic accuracy for predicting 30-day mortality (area under the curve 0.85, p <0.001; Figure 3 ). From these findings, a CK-MB cutoff value of 7 ng/ml was used to predict 30-day mortality for TF patients, equivalent to a twofold increase greater than the upper reference limit, yielding a sensitivity and specificity of 82% and 72%, respectively.

Receiver operating characteristic curves for predictive power of 30-day outcomes among all TAVR patients (A,B), TA patients (C,D), and TF patients (E,F).

Using a cutoff value of 7 ng/ml, the TF patients (n = 103) were divided into those with high (>7 ng/ml; n = 37) versus low (<7 ng/ml; n = 66) CK-MB. No significant differences were found in the baseline characteristics, apart from the baseline left ventricular ejection fraction, which was slightly greater among patients with high postprocedural CK-MB ( Table 3 ). Additionally, patients who had high versus low postprocedural CK-MB were less frequently taking β blockers at admission (59.5% vs 78.8%, p = 0.036; Table 3 ).

Stay updated, free articles. Join our Telegram channel

Join