Commentary Regarding Katsanos et al's Meta-Analysis: The Lessons Learned Should Impose Prudence and Caution

The meta-analysis presented by Katsanos et al concerns the outcome of 8 randomized studies comparing drug-coated balloons (DCB) vs plain old balloon angioplasty (POBA) in below-the-knee (BTK) intervention for critical limb ischemia (CLI). The analysis performed suggests a higher risk of amputation and death at 1 year with DCB in respect to POBA.  

First of all, the data of the meta-analysis are incomplete for the  ACOART BTK study, which ended the 12-month follow-up in January 2020. A mortality of 7% (3/41) in the DCB arm and 4% (2/44) in the POBA arm is reported in the meta-analysis, while the final 12-month data report a mortality of 4/52 (7.7%) in DCB vs 7/53(13.2%) in POBA patients. Adjusting the meta-analysis with the correct data, the hazard ratio (HR) for death reported in Figure 2 of the manuscript may change significantly.

Mortality risk in CLI patients depends on different factors: age, congestive heart failure, renal insufficiency, diabetes, and coronary artery disease. Although the studies are randomized, the randomization criteria are not focused on mortality risk and random error may occur. In a recent presentation (VEITHsymposium 2019) comparing POBA vs DCB in 1290 patients treated in the last 10 years, with 90% suffering from CLI, no difference in mortality at any time of the follow-up to 6 years was noticed once groups were matched by propensity score analysis.  

Major amputation in CLI patients depends on the patency (foot perfusion) and on wound care. The  studies reported in the meta-analysis have different protocol design and this may explain the difference in the results.

The ACOART BTK, the ACOART II, and the DEBATE-BTK trials have a similar protocol, which included a program of wound healing assessment several times per month until complete healing was achieved and a continuous monitoring of vessel patency with a fast-track strategy for repeat revascularization in case of need. These studies show a major amputation rate more close to 0 than 1 in DCB as well as in POBA patients. On the other hand, the IN.PACT DEEP and the  SINGA-PACLI studies, which together contribute to ≅60% of the data weight in the meta-analysis and trigger the risk of major amputation, share a high rate of major amputation (3-18%) and a higher risk of major amputation in DCB patients, which is hypothetically attributed to paclitaxel crystal embolization. This conclusion appears artificial; otherwise it should have been noticed also in the other studies. It is reasonable to believe that the higher amputation rate in these two studies is related to suboptimal wound care and surveillance, and/or a higher risk scenario that was not well balanced among DCB and POBA patients. In the IN.PACT DEEP study, patients were followed mostly by house physicians and not by a diabetic foot clinic. Patients undergoing amputation were not revascularized and this explains the low target lesion revascularization (TLR) rate (9%) in DCB patients, lower than in POBA (13%), with values much lower than those reported in the DEBATE BTK and ACOART series of studies. In the SINGA-PACLI study, 50% of the patients in both study arms were in end stage renal disease, a risk factor for major amputation due to the association with small artery disease, the distribution of which among groups was not reported and the enrollment of such patients in a randomized trial may be argued as a reason for the poor prognosis linked to the clinical condition. In the same study, baseline lesion pattern was short (9 cm) stenosis, while occlusions were found in only 30% of patients. Single short-vessel stenosis in BTK as the cause of CLI is rare. These lesions are always accompanied by other vessel occlusions. Having a higher rate of stenosis at baseline means leaving a lot of occluded vessel, and the final foot perfusion may be far from optimal and insufficient for wound healing. A similar pattern of baseline lesion is reported in the IN.PACT DEEP study, while in the studies reporting a low major amputation rate (ACOART BTK/II and DEBATE BTK), the mean lesion length was ≅12-20 cm and 70-80% of the lesion were chronic total occlusions (CTOs), a typical scenario of CLI patients.  

In conclusion, the high competing risk of major amputation and death in a CLI scenario and the structural differences among these studies make the assessment of a single causation very problematic, especially within a summary-level analysis, which does not allow adjustment for clinical and angiographic differences and risk between those who experience the event and those who don’t, particularly with an inappropriate patient sample size. Future studies on devices for BTK intervention in CLI patients should share a rigorous methodology in the evaluation of the final result, and on patency and wound healing surveillance. Once the ACOART BTK/II  series and the IN.PACT BTK studies are published, this concept will be reinforced. The lessons learned from the previous meta-analysis on fem-pop treatment for claudication and from the abundance of all further corrections, new analyses, and editorials that followed should impose prudence and caution in releasing similar alarming claims through the same methodology and with potentially the same limitations. 

Disclosure: Dr Liistro reports he is a consultant for Medtronic, BIOTRONIK, Boston Scientific, and Acotec Ltd. Dr Liistro is the principal investigator for the ACOART BTK and THE IN.PACT BTK trials.

Address for correspondence: Francesco Liistro, MD, can be contacted at francescoliistro@hotmail.com