Independent Analysis of Lutonix DCB Data Reveals No Plausible Link Between Paclitaxel and Mortality

(Leipzig, Germany) January 28, 2020 -- The latest insights into the safety and efficacy of drug-coated devices took center stage on Tuesday morning, with a large chunk of the session dedicated to paclitaxel safety.

Speaking during the session was Kenneth Ouriel, Founder, President and CEO of Syntactx (New York, NY), a full-service clinical research organization (CRO) that delivers high-quality clinical research services, including heading up clinical trials for medical device and pharmaceutical companies.

“The need for a CRO is exactly what my talk is about, because it is useful for medical device and pharmaceutical companies to have an independent assessor of data,” Dr Ouriel said. “And the reason for that is more to do with perceived, rather that real, conflicts of interest.”

In other words, CROs ensure that any perceived bias that people may have of company-sponsored trial data is put to bed via robust and independently verified analyses of datasets. 

In his talk, Dr Ouriel presented in-depth independent safety analysis of the Lutonix (BD) drug-coated balloon (DCB) – a 2 µg/mm² paclitaxel-eluting balloon catheter which has been extensively studied in the LEVANT 1, LEVANT 2, and LEVANT Japan series of trials.^1-3

“Lutonix was the first DCB approved by the Food and Drug Administration in the U.S.,” said Dr Ouriel. “Approval was based upon a rigorous preclinical and clinical scientific program that demonstrated both safety and effectiveness.”

The independent Lutonix analysis Dr Ouriel presented is in part a response to the questions raised about paclitaxel safety by the Katsanos et al meta-analysis (2018) which found a late all-cause mortality signal for patients treated with paclitaxel balloons and stents.⁴ As Dr Ouriel noted, Syntactx has since been doing an independent analysis of the Katsanos data as well – a task which has proven challenging indeed. “It’s been a lot of work, but it has been intellectually rewarding,” he said.

“However, we haven’t been able to figure out, when you have two groups of patients, one with an uncoated balloon and one with a paclitaxel-coated balloon, why there is a mortality signal? Some of the smartest people in the field have been asking the same question. There definitely is a signal there, but not a single clinician that I know believes that it's related to the small amount of paclitaxel which is found on the balloon.

“Most of us, myself included, believe that it's probably related to some facet of trial design – more than likely due to missed follow-up visits (that are not at random). As we work towards more complete follow-up on virtually every patient, the mortality signal almost disappears once the vital statistics on patient survival become clear.”

Indeed, several of the trials from leading manufacturers that fed into the meta-analysis had patients lost to follow-up, noted Dr Ouriel, which could have impacted the apparent mortality signal observed. Not least, he added, given that follow-up was different for the DCB cohorts vs non-DCBs.

Crucially, patient-level data was not included in the Katsanos et al meta-analysis, leading many to question if differences in follow-up care in DCB versus percutaneous transluminal angioplasty (PTA) comparators could be driving at least some of the mortality signal.

Focusing back on Syntactx’s independent analysis of Lutonix data, Dr Ouriel relayed that patient-level data was used to compare safety outcomes from 1093 Lutonix and 250 PTA patients across the LEVANT series of trials, including the LEVANT 2 Continued Access cohort – enrolled specifically to assess paclitaxel safety.

“We utilized additional statistical methods to assess the data, including propensity adjustment when pooling the data from the Continued Access cohort,” noted Dr Ouriel. “This is an appropriate way to remove the bias associated with differences in the make-up of patient groups that are not from the same RCT. We also performed time-dependent analyses to account for factors that change over time, and performed multivariable analyses to identify key predictors of mortality.”

Cutting to the chase with regards to whether any mortality signal could be seen with the Lutonix data, Dr Ouriel asserted that, of the 173 deaths seen in the LEVANT 1 and 2 datasets, no deaths were classified as related to paclitaxel based upon the known side effects of the drug. “Mechanistically, if paclitaxel caused death, there should be a disproportionate frequency of mortality in one category or a group of related categories,” explained Dr Ouriel. “That was not observed. Thus, without clustering of death within a category, causation is not supported.”

Staying on the topic of causality, Dr Ouriel underlined the importance of using Bradford Hill criteria – a nine-point system that explores epidemiologic evidence of a causal relationship between a presumed cause and an observed effect. Criteria span consistency, strength of effect, specificity, plausibility, coherence, biological gradient (dose response), coherence, temporality (does mortality increase following index procedure?), and analogy (could the effects be due to immunogenic particulates?).

Save for the last two criteria (temporality is present and particulates have been implicated in other situations), the rest of the criteria can be ruled out, commented Dr Ouriel. “The absence of seven of the nine criteria is consistent with association, but not causation.”

He went on to explore whether there were patient or treatment-related variables associated with increased risk, ie, is there a plausible mechanism for mortality associated with paclitaxel – or with any other biological feature of DCB treatment – even through some unknown mechanism?

Using a propensity-adjusted multivariate analysis of mortality out to five years in the LEVANT 2 data, Dr Ouriel and colleagues identified several variables as predictors of mortality, including age, Rutherford category, left limb, diabetes, anticoagulants at discharge, and prior treatment. However, these variables were shown to be predictors of mortality irrespective of treatment arm, ie, DCB or PTA, thus superseded paclitaxel as predictors of outcome.

As such, the burning questioned remains: “Is there a relationship between additional exposure to paclitaxel and risk of mortality?” said Dr Ouriel.

Looking at the LEVANT 2 RCT and Continued Access data, the effect of initial paclitaxel dose on survival was analyzed in four dose groups: > 0 to ≤ 2 mg; > 2 mg to ≤ 3.5 mg; > 3.5 mg ≤ 5 mg; and > 5 mg. “No significant dose-response relationship was identified,” commented Dr Ouriel, adding that when adjusting for age – the most significant predictor of mortality in both DCB or PTA groups – no identifiable dose-relationship could be seen.

We also looked at the effect of subsequent interventions with paclitaxel devices, which increases drug exposure,” continued Dr Ouriel. “Some of the other analyses that have been reported did not account for reinterventions. Almost 20% of subjects in the LEVANT 2 RCT were treated with a paclitaxel device at some point during their five-year follow-up.

“Subjects in both groups, DCB and PTA, who subsequently underwent an intervention with a paclitaxel device had higher five-year survival rates than those that did not. This finding was confirmed in our other studies, and would be counter-intuitive if additional paclitaxel exposure is indeed harmful in the long run.

 “It should be noted that the mortality rate in both the PTA and DCB groups in the LEVANT 2 study was lower than that of the PAD population as reported in the Swedish Vascular Registry (Sartipy et al, 2018) at five years.”

What this boils down to is that subjects in clinical trials may do better with additional clinical management, said Dr Ouriel, while reducing subsequent interventions is beneficial for patients, it also reduces additional “touch points” with health care providers.

Commenting on the outlook for paclitaxel, Dr Ouriel noted that as more data is added into analyses, the proposed signal for mortality using paclitaxel becomes even weaker, thus he was optimistic that the reputation of paclitaxel would recover. “People have felt reasonably comfortable going back to using paclitaxel devices, although I'm sure the market isn't anywhere near what it was before,” he said.

“I think this has really impressed on people that you do need complete follow-up beyond the primary endpoint, especially given the primary endpoint for many trials was a year or less. Even though the primary endpoint is earlier, people are pretty much doing five-year trials now in the lower extremities.”

He concluded: “There is no significant increase in the hazard ratio for mortality in any analysis of the Lutonix DCB, nor any plausible mechanism for mortality or evidence of paclitaxel causation. Based on all of our analyses to date in a large dataset, the Lutonix DCB continues to offer meaningful benefit relative to risk in indicated patients.”

References

1. Scheinert D, Duda S, Zeller T, et al. The LEVANT I (Lutonix paclitaxel-coated balloon for the prevention of femoropopliteal restenosis) trial for femoropopliteal revascularization: first-in-human randomized trial of low-dose drug-coated balloon versus uncoated balloon angioplasty. JACC Cardiovasc Interv. 2014;7(1):10–19.
2. Rosenfield K, Jaff MR, White CJ, et al. Trial of a Paclitaxel-Coated Balloon for Femoropopliteal Artery Disease. N Engl J Med. 2015;373(2):145–153.
3. ClinicalTrials.gov. LEVANT Japan Clinical Trial. Available at: https://clinicaltrials.gov/ct2/show/NCT01816412. Accessed January 28, 2020.
4. Katsanos K, Spiliopoulos S, Kitrou P, et al. Risk of Death Following Application of Paclitaxel-Coated Balloons and Stents in the Femoropopliteal Artery of the Leg: A Systematic Review and Meta-Analysis of Randomized Controlled Trials. J Am Heart Assoc. 2018;7(24):e011245.