Abstract
The SWEDEPAD trial reported an unplanned interim analysis to show no difference in the mortality rate between the paclitaxel-coated and uncoated groups (Nordanstig et al., 2020), which contradicts the long-term risk of paclitaxel-coated devices claimed by a meta-analysis (Katsanos et al., 2018). However, there existed significant differences in mortality rates between the SWEDEPAD trial and the trials included in the meta-analysis, which were caused by significant differences in the patient populations. As a result, the SWEDEPAD trial and meta-analysis results are not directly comparable. An updated meta-analysis including the SWEDPEPAD trial and all studies in the meta-analysis (Katsanos et al., 2018) shows marginal differences in mortality rates between the paclitaxel-coated and control groups at two years with Bayesian relative risk (RR) 1.39 (95% credible interval (CrI) [1.01, 2.39]) and frequentist RR 1.16 (95% confidence interval (CI) [0.99, 1.36]) and differences in mortality rates during the entire follow-up period with Bayesian RR 1.29 (95% CrI [1.01, 1.72]) and frequentist RR 1.13 (95% CI [0.99, 1.28]) under random-effects models. Given the relatively short follow-up thus far in the SWEDEPAD trial (with a mean follow-up of 2.49 years) and the paclitaxel-coated risk being long-term (e.g., 4 or 5 years), the interim results on the risk of paclitaxel-coated devices reported by the SWEDEPAD trial warrant further investigation.
Introduction
In vascular intervention, paclitaxel-coated balloons and stents are standard treatments for peripheral artery disease (PAD). However, safety concerns emerged following a 2018 meta-analysis by Katsanos et al., which suggested an association between these devices and increased long-term mortality. This finding was recently challenged by interim results from the SWEDEPAD trial, which found no significant difference in death rates between coated and uncoated devices. This study re-evaluates these conflicting findings by investigating the comparability of the SWEDEPAD trial and the original meta-analysis populations.
Methods
The authors conducted an updated meta-analysis integrating SWEDEPAD interim data with the randomized controlled trials from the 2018 Katsanos study. Two statistical frameworks were employed to ensure robustness:
- Frequentist Random-Effects Model: Calculated Relative Risk (RR) and p-values using the R package
meta. - Bayesian Random-Effects Model: Calculated RR and 95% credible intervals (CrI) using the R package
bayesmeta.
The Bayesian approach was utilized to determine the posterior probability—the actual likelihood that the mortality rate is higher in the paclitaxel group—providing a more nuanced perspective than binary p-values.
Results
The analysis revealed that while short-term (one-year) mortality showed no significant difference, longer-term outcomes suggested a persistent risk. At the two-year mark and across the entire follow-up period, the Bayesian model indicated a high probability of increased risk in the paclitaxel group.
As shown in the table above, the Bayesian posterior probability that the death rate is higher in the paclitaxel-coated arm reached 0.978 at two years and 0.980 for the entire follow-up period. While frequentist p-values were slightly above the traditional 0.05 threshold (0.063 and 0.065, respectively), the Bayesian evidence for risk remained strong.
Discussion
A critical finding of this study is the "population incomparability" between the datasets. The SWEDEPAD population consisted of 65% Chronic Limb Threatening Ischemia (CLTI) patients, a more advanced stage of PAD, whereas the Katsanos meta-analysis population was 89% Intermittent Claudication (IC). These baseline differences in patient health mean that SWEDEPAD's results cannot simply negate previous findings.
Furthermore, paclitaxel-related mortality is typically observed as a long-term risk (4–5 years). With a mean follow-up of only 2.49 years, the SWEDEPAD interim data may not have reached the necessary duration to capture the full extent of the potential risk.
Conclusion
The interim SWEDEPAD results do not definitively dismiss the safety concerns raised by earlier meta-analyses. The updated analysis shows that a long-term mortality risk remains statistically plausible. Continued monitoring and full 5-year follow-up data from SWEDEPAD and other registries are essential to resolve the debate regarding paclitaxel-coated device safety.
Summary updated