Presentation Authors: Jeffrey Sack*, Sarasota, FL, Jayme Sack, Saddle River, NJ, Hari Tunuguntla, Piscataway, NJ
Introduction: The optimal treatment for renal artery in-stent restenosis (RAISR) is unknown. To determine the optimal treatment for RAISR, we conducted a Bayesian network meta-analysis comparing percutaneous balloon angioplasty (PTA), bare metal stent (BMS), drug eluding stent (DES), and cutting balloon angioplasty (CB
Methods: We searched MEDLINE, Cochrane and EMBASE libraries for all studies using pre-specified criteria of endovascular treatments for RAISR without restriction on language between January 1980 to April 2018. We excluded patients who had renal transplant renal artery stenosis. A Bayesian network meta-analysis was performed using WinBUGS (v1.4.3) comparing PTA vs BMS vs DES vs CB using both a fixed and random effect model using non-informative priors. PTA was considered the reference treatment and odds ratios (OR) for 12-month restenosis were calculated for all comparisons along with 95% confidence intervals (CI). There was insufficient data to include covered stents, drug coated balloons and brachytherapy. We calculated the surface under the cumulative ranking (SUCRA) a simple numerical summary of the probabilities. SUCRA is 100% when a treatment is certain to be the best and 0% when the treatment is certain to be the worst. Bias was assessed using heterogeneity plots and the Cochrane tool.
Results: We included 13 studies involving a total of 943 patients. (mean age 65.1 years, 60% male). DES was the most effective treatment for RAISR with an OR=0.26 [95% CI, (0.02-3.04)], compared with PTA and a SUCRA of 81% followed by CB [OR 0.28, (0.06-1.00), SUCRA=70%] and BMS [OR 0.62, (0.20-1.96), SUCRA=37%]. PTA scored the lowest SUCRA of 12%, consistent with the least effect treatment of those compared.
Conclusions: Based on these data, treatment of RAISR with DES appears to be most favored, followed by CB, BMS and then PTA. The lack of randomized controlled trials and the small data sets of the included studies limits our findings.