Accessory pathways with “high-risk” properties confer a small but potential risk of sudden cardiac death. Pediatric guidelines advocate for either risk stratification or ablation in patients with ventricular pre-excitation but do not advocate specific methodology. We sought to compare the cost of differing risk-stratification methodologies in pediatric patients with ventricular pre-excitation in this single institutional, retrospective cohort study of asymptomatic pediatric patients who underwent risk stratification for ventricular pre-excitation. Institutional methodology consisted of stratification using graded exercise testing (GXT) followed by esophageal testing in patients without loss of pre-excitation and ultimately ablation in high-risk patients or patients who became clinically symptomatic during follow-up. A decision analysis model was used to compare this methodology with hypothetical methodologies using different components of the stratification technique and an “ablate all” method. One hundred and two pediatric patients with asymptomatic ventricular pre-excitation underwent staged risk stratification; 73% of patients were deemed low risk and avoided ablation and the remaining 27% ultimately were successfully ablated. The use of esophageal testing was associated with a 23% (p ≤0.0001) reduction in cost compared with GXT stratification alone and a 48% (p ≤0.0001) reduction compared with the “ablate all” model. GXT as a lone stratification method was also associated with a 15% cost reduction (p ≤0.0001) compared with the “ablate all” method. In conclusion, risk stratification of pediatric patients with asymptomatic ventricular pre-excitation is associated with reduced cost. These outcomes of cost-effectiveness need to be combined with the risks and benefits associated with ablation and risk stratification.
A subset of patients with ventricular pre-excitation have increased risk for sudden cardiac death secondary to rapid ventricular response during atrial fibrillation. In patients with ventricular pre-excitation and atrioventricular reentrant tachycardia (AVRT), ablation is effective both in the treatment of clinical symptoms of tachycardia as well as to avoid the associated risk of sudden cardiac death. Because of this, catheter ablation is generally viewed to be first-line therapy in most symptomatic patients with ventricular pre-excitation. Although catheter ablation is recommended for symptomatic patients, treatment for asymptomatic patients is less clear. Catheter ablation has been shown to be effective and relatively safe, but there is still a small but significant procedural risk as well as a cost burden associated with ablation. Recent PACES/HRS guidelines for the care of pediatric patients with asymptomatic ventricular pre-excitation advocated for risk stratification or ablation in all patients. Several methods of stratification exist, but no specific protocol or testing techniques were advocated, and treatment was left to the discretion of the managing physician. The purpose of this study was to compare the cost of varying risk stratification protocols and techniques including graded exercise testing (GXT), transesophageal pacing testing (TEP), transvenous intracardiac testing, and ultimately catheter ablation when indicated versus a treat-all approach. We hypothesize that the use of noninvasive or minimally invasive stratification techniques in a stepwise protocol are less costly than more invasive risk stratification techniques such as intracardiac electrophysiology (EP) study or ablation alone.
Methods
This was a single institutional, retrospective cohort study of pediatric patients with ventricular pre-excitation. This study was performed in accordance with the Cincinnati Children’s Hospital Institutional Review Board (2012-3158). All patients who underwent risk stratification between January 2000 and January 2012 were included in initial data collection. In an effort to analyze a truly asymptomatic population, any patient with a history of AVRT or palpitations before undergoing risk stratification was excluded from the study. Patients were also excluded if they had not completed risk stratification at the time of the study.
Patient demographics and electrophysiologic data were obtained from hospital and EP specific databases. Patient data including age, gender, and race were analyzed. EP-specific data including results of GXT, TEP, and ablation procedures were reviewed. Patient age at time of risk stratification was delineated as the patient age at the time of the GXT, the initial modality used for staged risk assessment.
The staged risk stratification strategy used in this study has been previously described. In brief, all patients in this study initially underwent risk stratification with GXT. In GXT studies with loss of pre-excitation with concomitant lengthening of the PR interval during exercise, patients were considered low risk and typically followed every 2 years with repeat electrocardiogram (ECG) testing. Patients without loss of pre-excitation or with gradual loss of pre-excitation without concomitant lengthening of the PR interval were deemed indeterminate and underwent additional stratification using TEP. During TEP, the shortest pre-excited R-R interval (SPRR) seen during atrial fibrillation was used for risk stratification, and all procedures were performed under conscious sedation. Patients with SPRR intervals ≤240 milliseconds were deemed high risk and were recommended for catheter ablation. Patients with SPRR intervals ≥250 milliseconds were deemed low risk and followed every 2 years with ECG testing. If at any point during or after the risk stratification process the patient developed documented clinical AVRT, the family was offered ablation, which was left up to the discretion of the family and managing electrophysiologist.
A decision analysis model was used to examine cost outcomes achieved using various potential techniques for risk stratification including GXT, TEP, transvenous EP study versus an “ablate all” strategy. Four potential risk stratification models were created. Model 1 (GXT-TEP-catheter ablation [ABL]) represents the actual outcomes and cost using the institutional risk stratification protocol described earlier. This protocol includes initial risk stratification with GXT with additional stratification with TEP or ablation when indicated. Patients with successful stratification either through GXT or TEP were followed every 2 years with a clinic visit and standard ECG. If patients developed clinical symptoms of AVRT, catheter ablation was offered. Following ablation, in either high-risk patients or patients who developed clinical symptoms and underwent ablation for symptoms, patients were seen for 2 subsequent visits with standard ECG. Patients with recurrence of pathway conduction following ablation underwent repeat ablation procedures.
Models 2 through 4 were hypothetical models with cost analysis and outcomes based on the clinical outcomes from Model 1. Model 2 (GXT-EP-ABL) was similar to model 1, with the exception of using transvenous EP study for risk assessment following GXT as opposed to TEP. Model 3 (GXT-ABL) was a hypothetical model in which GXT testing was the only method used for stratification. Patients who were not stratified by exercise testing underwent ablation without further stratification. The estimated number of repeat procedures secondary to pathway recurrence was calculated based on the percentage of repeat procedures required in Model 1 times the number of patients proposed to undergo ablation in Model 3. Model 4 (ABL-ALL) was a hypothetical “treat all model” assuming no use of risk stratification and that all patients with pre-excitation undergo ablation. The estimated number of repeat procedures for recurrence was calculated based on the percentage of repeat procedures required in Model 1 times the number of patients proposed to undergo ablation in Model 4.
Cost comparison was analyzed using a provider model consisting of 66% health maintenance organization (HMO)/commercial insurance and 34% Medicaid payors. Because payor type varies between institutions, total method cost was also calculated using 50% HMO/commercial and 50% Medicaid payors as well as 34% HMO/commercial and 66% Medicaid. Reimbursement charges include total reimbursement including those charges for hospital and professional reimbursement based on International Classification of Diseases, Ninth Revision (ICD-9)/Current Procedural Terminology codes. The specific codes and reimbursement charges used for procedures, clinic follow-up, and ECG charges are shown in Appendices A through D . On the basis of institutional practice, patients who were deemed to be low risk and did not undergo ablation were followed every 2 years. For our modeling, we assumed 10 years of follow-up, which included charges for 5 clinic visits with ECG charges at each visit. Following successful ablation, patients were seen at 2 to 4 weeks postprocedure and 1 year postprocedure. For our modeling, 2 visits with ECGs were included in patients following successful ablation. For the GXT-TEP-ABL model, which used TEP studies as part of the risk stratification protocol, it was assumed that the TEP and ablation procedures were performed on different encounters and incurred costs associated with 2 separate procedures. For the GXT-EP-ABL model, in which transcatheter EP study was used as part of risk stratification, it was assumed that the ablation would be performed during the same encounter as the EP study. As such, only charges for related to an ablation procedure were used in the cost analysis when the EP study was deemed high risk.
Descriptive data were expressed as means with standard deviations (normally distributed continuous variables) or as proportions (categorical and dichotomous variables). The distribution of cost data for each method was tested for normality, and only the ABL-ALL model has a bimodal distribution. For purposes of descriptive representations, potential drivers of cost including gender, race, and age were compared. In the descriptive representation, age was classified into 3 groups including (1) patients aged 5 to 7 years, (2) patients aged 8 to 15 years, and (3) patients aged >16 years. The potential drivers were then compared with respect to their effects on cost using tests of fixed effects. For this comparison, age was used as a continuous variable. Finally, the total cost and cost per patient were compared between the 4 risk stratification models using differences of least squares means.
Mixed models for repeated measures were analyzed to compare the cost of different methods, controlling for the covariates previously listed. Both a fixed effects and interaction model were run to examine the relationship between risk stratification methods and covariates. All data were analyzed using SAS 9.3.
Results
One hundred seventy-four patients with ventricular pre-excitation underwent staged risk stratification over the study period ( Figure 1 ). Four patients were excluded secondary to ongoing risk stratification, and 68 patients were excluded with either clinical symptomatology of palpitations or history of documented AVRT. The remaining 102 patients were analyzed in this study.
Figure 2 demonstrates the patient outcomes of the patient cohort for the GXT-TEP-ABL model as well as the proposed outcomes for the GXT-EP-ABL, GXT-ABL, and ABL-ALL models. In GXT-TEP-ABL and GXT-EP-ABL models, 28 patients ultimately underwent ablation with a single additional ablation secondary to recurrence of pathway conduction. In the GXT-ABL model, in which GXT testing was used as an isolated stratification method, 75 patients underwent ablation with an estimated 3 additional ablation procedures for recurrence. Finally, in the ABL-ALL model, all 102 patients underwent ablation with an additional 4 patients undergoing repeat ablation for estimated recurrence.
Figure 3 demonstrates reimbursement of both the individual testing components as well as method type based on differing payor mixes. Figure 4 demonstrates the average cost per patient as well as the total cost for the entire cohort based on the differing risk stratification models using the 66% HMO/commercial and 34% Medicaid payor mix. The first model, in which patients underwent staged stratification with GXT and TEP demonstrated significantly lower cost compared with other stratification models including a reduction of cost of 12% compared with the GXT-EP-ABL model (p = 0.01), 26% compared with GXT-ABL model (p ≤0.0001) and 48% compared with ABL-ALL model (p <0.0001). Likewise, the GXT-EP-ABL model, in which patients underwent staged stratification with GXT and EP study, demonstrated significantly lower cost compared with GXT-ABL and ABL-ALL models with a reduction in cost of 13% (p = 0.03) and 32% (p ≤0.0001) respectively. Lastly, the GXT-ABL model, using GXT as a lone stratification methodology, demonstrated a 15% decrease compared with ABL-ALL model (p ≤0.0001).
