ABC | Volume 113, Nº2, August 2019

Original Article Saad et al. Healthcare utilization and costs after ablation for AF Arq Bras Cardiol. 2019; 113(2):252-257 Table 2 – Monthly resource use and costs before and after the ablation procedure Outcome Before ablation - mean (SD) Before ablation - median (IQR) After ablation - mean (SD) After ablation - median (IQR) p value Number of outpatient office visits 0.05 (0.15) 0 (0 - 0) 0.04 (0.10) 0 (0 - 0) 0.770 Number of emergency department visits 0.17 (0.21) 0.10 (0.04 – 0.23) 0.08 (0.16) 0 (0 – 0.11) < 0.001 Number of emergency department visits - arrhythmic ICD 0.05 (0.07) 0 (0 – 0.09) 0.01 (0.04) 0 (0 – 0) < 0.001 Number of elective hospital admissions 0.01 (0.02) 0 (0 – 0) 0.01 (0.04) 0 (0 – 0) 0.134 Total costs (BRL) 747.75 (1,315.38) 330.95 (142.36 – 754.17) 589.93 (1,779.83) 104.21 (56,35 – 226,51) < 0.001 Outpatient costs (BRL) 156.81 (161.90) 121.48 (56.35 – 206.87) 83.74 (95.17) 62.70 (32.91 – 105.15) < 0.001 Emergency related costs (BRL) 500.95 (1,268.61) 65.21 (3.54 – 433.88) 110.57 (358.86) 0 (0 - 36.98) < 0.001 Elective admissions related costs (BRL) 89.99 (416.33) 0 (0 - 0) 395.61 (1,720.18) 0 (0 - 0) 0.215 SD: standard deviation; IQR: interquartile range; BRL: Brazilian Reais. P values were calculated with non-parametric tests since all variables had a non-normal distribution. Other reports from the literature have also seen the impact of post-ablation cost reduction. In the larger study published to date, Ladapo et al. 11 included 3,194 patients from administrative databases in the US. 11 In that research, the approach was slightly different: they considered that costs can actually increase in the 6 months following the procedure, as a result of the need of reablation in a fraction of the sample, as well as the treatment of peri-procedural complications. Therefore, they analyzed the period from 6 to 36 months after ablation, divided into 6-month cycles. In the time frame of 6-12 months after ablation, mean monthly costs reduced around US$ 800, in comparison with the 6 months immediately before ablation. This number reduced until 18-24 months (where the reduction, compared to before ablation, was around US$ 200), and then increased again to around US$ 800 in the 30-36 months period. However, only 1/3 and 1/10 of patients had at least 24 months and 36 of follow-up time, respectively, making this long-term data more imprecise. Regardless, it seems considerably robust that cost reductions are noted already in the first year, and that it is retained over a longer follow-up period. Some studies in the literature have estimated how long after catheter ablation the procedure would become “cost‑neutral”. In a French retrospective cohort study that included 118 consecutive patients submitted to radiofrequency ablation for paroxysmal AF during a mean follow-up of 32 ± 15 weeks, it was estimated that from the 5 th year onwards, total accumulated costs would be smaller in patients submitted to ablation, as compared to medical treatment. 14 In two Canadian economic models, the cost-neutrality would occur between 2 and 4 years of follow-up. 13,15 These three studies, however, were not fully based on collected data and included some future projections and modelling. Some limitations of our study must be acknowledged. The dataset used for all analyses was based on patient billing information and the patients were made anonymous to the researches. Therefore, direct contact to establish the recurrence was not possible. This could overestimate the success rate because the recurrence was only based on the use of healthcare resources (use of antiarrhytmic drug in the emergency room, cardioversion or repeated procedures) or indirectly by the purchase of antiarrhythmic drug in the pharmacies by the patient. The use of an administrative database carries the risk of bias as any retrospective study, as well as the problems associated with the lack of individual clinical patient information. Moreover, we did not included costs with ambulatory medications, since this information was not available in the patients’ billings information dataset, which did not included out-of-pocket patients expenditures. Finally, the sample size was not large, and the analysis of possible predictors of greater cost reductions after the ablation procedure was probably underpowered. Conclusion In this sample of patients from the Brazilian private healthcare sector, catheter ablation of AF was associated with significantly decreased costs – both ambulatory and hospital-based. Author contributions Conception and design of the research: Saad EB, Tayar DO, Ribeiro RA, Junqueira Jr. SM, Andrade P, d'Avila A; Acquisition of data: Tayar DO; Analysis and interpretation of the data: Saad EB, Tayar DO, Ribeiro RA, Andrade P, d'Avila A; Statistical analysis: Ribeiro RA; Obtaining financing: Junqueira Jr. SM, Andrade P; Writing of the manuscript: Saad EB, Tayar DO, Ribeiro RA, d'Avila A; Critical revision of the manuscript for intellectual content: Saad EB, Tayar DO, Junqueira Jr. SM, Andrade P, d'Avila A. Potential Conflict of Interest Dr. Eduardo Benchimol Saad received lecture fees from Bionese Webster and Biotronik. Dra. Daiane Oliveira Tayar is employed by Johnson and Johnson Medical Brasil (Department of Economics and Access Market). Rodrigo A. Ribeiro received a research grant from Johnson& JohnsonMedical Brazil to conduct the database, statistical analysis and to draft this manuscript. Dr. Silvio Mauro Junqueira Jr. is employed by Johnson and 255