Drug-Coated Balloon-Only Percutaneous Coronary Intervention for the Treatment of De Novo Coronary Artery Disease: A Systematic Review

The use of DCB-only PCI in non-specific clinical scenarios, mainly in SVD, forms the main focus of the current literature. Full details of all study outcomes are provided in Table 3.

PICCOLETO compared the Dior I DCB to the 1st-generation DES, Taxus Liberté [15]. It was stopped prematurely due to clear superiority of the DES. However, certain factors explain these discouraging results. Above all, the Dior I DCB has been shown to elute lower concentrations of paclitaxel compared to subsequent DCB generations. Also, adequate preparation of lesions with POBA before DCB application was only performed in 25% of lesions. Pre-dilatation with POBA and the use of cutting balloons has been shown to further facilitate intimal and medial drug delivery via the formation of microdissections [16]. Moreover, a high rate of bailout stenting was seen and may be attributed to use in type B dissections, which later recommendations do not advocate [14]. Furthermore ‘geographical mismatch’ was not taken into account. This describes stented areas of the vasculature in bailout (with a BMS) that have not been previously treated with a DCB. They are particularly prone to restenosis and are associated with poorer outcomes. PEPCAD I previously also identified this [17, 18]. Geographical mismatch can be avoided through the use of a shorter stent implanted within the DCB treated area. A DES may be used in these cases or a BMS where long-term DAPT is contraindicated [14].

The BELLO study compared the IN.PACT Falcon DCB against the Taxus Liberté DES. It showed a smaller LLL in DCB-treated patients with comparable clinical outcomes to the DES [19, 20]. A sub-analysis of diabetic patients showed similar results [21]. However, issues arise when using LLL to compare stenting strategies with a balloon-only approach as stent placement, by nature, will result in greater acute luminal gain and consequently a greater LLL. Moreover, only a 1st-generation DES was the comparator. Giannini et al. showed comparable clinical outcomes in the BELLO DCB group when propensity score matched against patients treated with a 2nd-generation DES (Xience V or Promus) [22].

Nishiyama et al. [23] report comparable clinical outcomes between patients randomized to receive either SeQuent Please or a 2nd-generation DES (Xience Prime or Xpedition). This was largely due to a lack of adverse events seen in both groups over the short follow-up and the exclusion of DCB patients who received bailout stenting. The use of a Lacrosse non-slip element (NSE) balloon for pre-dilatation in addition to the use of intravascular ultrasound (IVUS) for the evaluation of an optimal result before DCB application, may have also contributed to the good outcomes. Shin et al. [24] compared SeQuent Please against a 2nd-generation DES (Xience Prime or Resolute Integrity). The investigators used a fractional flow reserve (FFR)-guided approach. Following POBA pre-dilatation, if a good FFR was seen (> 0.85) a DCB was used; otherwise a DES was implanted. Excellent clinical, angiographic, and functional results were seen. However, given the reservation of DES for more complex lesions, comparison is limited.

Sinaga et al. [25] retrospectively compared cohorts of SeQuent Please and 2nd/3rd-generation DES (Resolute Integrity, Xience, Promus Element, Biomatrix or Nobori)-treated patients. Comparable clinical outcomes were seen. However, the DCB-only treated group showed a significantly smaller MVD with DES use being associated with more proximal lesions of the major epicardial arteries. This could have confounded clinical outcomes with stenosis of smaller vessels perhaps having a less significant impact. Venetsanos et al. compared large propensity score-matched populations receiving a DCB (SeQuent Please, IN.PACT Falcon or Pantera Lux) against a 2nd/3rd-generation DES (Xience, Promus, Synergy, Resolute, Orsiro or Nobori). DCB treatment was notably associated with a significantly lower occurrence of TLT. This may be related to early discontinuation of DAPT as a minimum of only 6 months was required in the DES group [26]. DCB use was also seen in significantly less complex lesions, despite propensity score matching. A DES was used in the majority of bailout cases with good outcomes. DCB investigators were initially hesitant to do this due to concerns about the vascular effects of combining two drug-eluting devices. Her et al. showed a significantly lower incidence of periprocedural MI in propensity score-matched SeQuent Please treated patients when compared to 1st/2nd-generation DESs (Cypher, Taxus Express and Endeavor). However, 88.5% of DES patients were given only a first-generation device, which may explain the poorer outcomes seen. Periprocedural MI is a complication of up to 30% of DES procedures and troponin release after PCI is classically associated with worse outcomes [27].

Funatsu et al. report a smaller LLL in SeQuent Please-treated patients when compared to POBA (with no significant difference in adverse outcomes) [28]. Sub-analysis showed significantly lower adverse clinical outcomes in patients pre-dilated with a Lacrosse NSE. This may be related to significantly less bailout procedures being performed in this subgroup. A short follow-up period may also explain why POBA results were better than expected. Furthermore, bailout stenting was not indicated in residual stenosis up to 50% (as opposed 30%), with authors commenting that the German Consensus Group recommendations would have resulted in too aggressive a pre-dilatory approach, leading to a higher %bailout. The long-term effects of such an approach are unclear. Her et al. [29] also similarly report superior angiographic and comparable clinical outcomes when comparing SeQuent Please to POBA.

Moreover, in both studies, luminal enlargement was observed in over 50% of patients at follow-up. In a single-armed study, Kleber et al. [30] also found this in 69% of patients. This was most pronounced in areas where plaque burden was highest and was attributed to possible positive vessel remodelling, vascular healing and plaque regression. Cortese et al. [31] found that 94% of patients left dissection (type A–C) had later healed. To better characterize these positive remodeling processes, Ann et al. [32, 33] used IVUS histology and optical coherence tomography (OCT) in two cohorts of DCB-treated patients with a FFR-guided approach. Both studies showed good angiographic, functional, and clinical results. IVUS histological analysis showed the conversion of four thin-cap fibroatheromas to a thick-cap, suggesting possible plaque stabilization. OCT showed an increased mean luminal diameter and volume at follow-up in addition to the sealing of 66% of dissections. The OCTOPUS II study again through an FFR-guided approach with OCT analysis also showed positive luminal gain, healing of dissections and a lack of thin-cap fibroatheromas [34].

Various registries exist to monitor the safety and efficacy of DCBs in the ‘real world’. The majority of these have used SeQuent Please. The SeQuent Please worldwide registry is the largest of these and showed low MACE and TLR rates [35]. Bailout stenting was not associated with adverse outcomes however; diabetes was a significant predictor of TLR. The SeQuent Please Small Vessel PCB-only Registry reported slightly higher (but nonetheless good) MACE and TLR rates [36]. This was attributed to a smaller cohort MVD. Sub-analyses of this registry comparing elderly patients (> 75 years), patients with ACS and Asian versus Western populations have been published. Similar outcomes and %bailout was seen across all groups, despite the presence of more comorbidities in elderly patients and on average smaller vessels with longer lesions in Asian patients [37‐39].

Other smaller SeQuent Please registries include Calé et al. [40], which showed a relatively high incidence of MACE and TLR despite a low frequency of bailout stenting. The authors attributed this to a high-risk population. Benezet et al., The Leipzig Registry, and Hee et al. all show favorable long-term MACE and TLR rates [41‐43]. Cuculi et al. [44] report an IN.PACT Falcon DCB registry showing favorable  %bailout, MACE, and TLR rates. Basavarajah et al’s IN.PACT Falcon registry report a relatively high MACE and TLR rate. This was attributed to the presence of diffuse CAD (> 20 mm) in 80% of patients [45]. Zivelonghi et al. provide the longest follow-up seen in any DCB registry, showing good long-term MACE and TLR rates. Over half the cohort presented with ACS and this was associated with a higher incidence of adverse outcome [46].

Registries of the Dior II DCB include Valentines II. Here, angiographic follow-up was only performed in 34% of patients and showed a relatively high LLL (although low MACE and TLR rates were still seen) [47]. This may be due to recommended pre-dilatation being performed in only 85% of patients. The Spanish Dior Registry similarly reported a high LLL. However, 54% of patients received the less effective Dior I balloon. Again, favorable clinical outcomes were reported and BMS bailout was a predictor of adverse events [48]. DELUX was a Pantera Lux DCB registry which showed comparable outcomes to other DCB registries, with BMS bailout again predicting poorer outcome [49]. The Italian Elutax SV Registry-DCB Rise is a registry of the Elutax SV DCB. It found low DOCE and TLR rates [50].

Study outcomes focusing on the use of DCB-only angioplasty for patients presenting with ACS are summarized in Table 4. In PPCI, the risk of ST occurrence with DES is greater than in elective cases due factors such as incomplete stent apposition and delayed tissue coverage, making the role of a DCB-only approach for this indication interesting to consider. Gobic et al. conducted a randomized trial, in ST elevation MI (STEMI) patients comparing SeQuent Please to the 3rd-generation DES Biomime. They showed a superior LLL in DCB-treated patients (although issues with LLL in stent versus balloon studies have been discussed) and comparable short-term clinical outcomes [51]. Luminal enlargement was also observed. However, DCB patients requiring BMS bailout were excluded from analysis. DCB patients also had a significantly smaller MVD compared to the DES group despite randomization. Nijhoff et al. [52] report an observational comparison of a DCB-only cohort against a previous three-armed randomized trial, DEB-AMI. DCB-only PCI exhibited a comparable LLL versus BMS and DCB + routine BMS, but was inferior to the Taxus Liberté DES. Despite this, comparable clinical outcomes were seen across all four groups. Coronary endothelial dysfunction (tested using acetylcholine) was also least pronounced in the DCB-only group. A lack of acute and late thrombosis indicates DCB-only may be viable in STEMI patients with a long-term DAPT contraindication. PAPPA was a feasibility study of the DCB-only approach (using Pantera Lux) in STEMI showing favorable TLR and MACE rates. However, a very high bailout rate of 41% was seen, due to a high rate of dissections, which may be related to the use of a slightly oversized balloon for pre-dilatation [53]. Ho et al. [54] report acceptable short-term clinical outcomes in STEMI patients treated with SeQuent Please. It should be noted that DCB-only PPCI investigators have commented on the importance of thrombus aspiration where relevant before DCB application to avoid reduced paclitaxel transfer by an interposed mural thrombus.

Table 5 provides a summary of the DCB-only studies which focus on bifurcating lesions. Bifurcation represents around 20% of PCI procedures and is associated with a higher risk of restenosis and complications. Both the interventional approach and operator factors are highly important for a successful procedure. DCB-only PCI in bifurcation theoretically provides certain advantages over stenting strategies. These include the maintenance of a natural flow distribution and avoidance of plaque and carina shift due to the absence of a stent overstretching and straightening the distal vessel, which predisposes to side-branch occlusion or narrowing, leading to adverse clinical events. PEPCAD-BIF compared SeQuent Please against POBA in a randomized study. The DCB group showed a significantly smaller LLL, however comparable clinical outcomes were nonetheless seen [55]. Bifurcating lesions involving the proximal main-branch (Medina class 1,X,X) were, however, notably excluded. Various single-armed studies have also been conducted reporting acceptable MACE and TLR rates [56, 57]. Bruch et al. report a cohort where 75% of patients presented with true bifurcation (Medina 1,1,1, 1,0,1, or 0,1,1). Although not associated with worse clinical outcomes, a high %bailout of 45% was seen and this was associated with lesions of the left anterior descending (LAD) artery and the presence of B2/C lesions. Vaquerizo et al. report clinical and angiographic outcomes in ostial side-branch (SB) lesions (Medina 0,0,1) treated with Dior II and found high MACE and TLR rates [58]. PCI of ostial side-branch lesions is naturally associated with a smaller vessel diameter, greater incidence of recoil, lower acute luminal gain and thus a higher rate of complications. Her et al. found late luminal gain (confirmed on OCT analysis) in both the main-branch (MB) and SB of bifurcating lesions treated with SeQuent Please [59]. No adverse events were reported. Lesions that showed poor image quality due to dissection or artifact on OCT were notably excluded.

Table 6 outlines the studies focusing on DCB-only angioplasty in other clinical scenarios, namely in calcification and CTO. Heavily calcified CAD is associated with poorer clinical outcomes due to the difficulty of adequately deploying a stent due to incomplete stent expansion and strut apposition. Calcification is especially problematic in patients with chronic kidney disease. Ito et al. [60] show acceptable MACE and TLR rates in a feasibility study. Preparation of the calcified lesion required rotational atherectomy in 80% of patients. Chronic hemodialysis (seen in 21% of patients) was associated with an increased risk of adverse events. Comparable clinical and angiographic results were seen when compared to patients with non-calcified lesions. These favorable results may be explained by the exclusion of patients with significant residual stenosis and dissection following lesion preparation as well as the use of IVUS and OCT to aid the procedure and the use of a Lacrosse NSE for pre-dilatation. Rissanen et al. [61] report outcomes in 65 patients with calcified lesions treated with SeQuent Please, following rotational atherectomy. It is thought that rotational atherectomy prior to DCB treatment reduces calcific burden, thus enhancing the penetration of paclitaxel into the vessel wall. This technique has already been established in calcified femoro-popliteal lesions. CTOs are defined as a coronary occlusion without anterograde flow that has been present for at least 3 months. It has a reported incidence of up to 30% and is associated with higher rates of ISR and ST in addition to being technically challenging for interventionalists. In Köln et al.’s [62] feasibility study, favorable angiographic results were seen. A lack of MI, vessel thrombosis, or death at follow-up are also significant findings. However, this study excluded patients who did not achieve satisfactory pre-dilatation and only included patients requiring an anterograde interventional approach.

Comparison of the DCB-only approach against the DES is essential to consider, given the DES forms the current mainstay in de novo CAD treatment. With the exception of the PICCOLETO study, DCB-only and DES PCI show similar clinical outcomes [15]. Despite these encouraging results, the literature is largely lacking in data from randomized trials that compare to a 2nd- or 3rd-generation DES, with Nishiyama et al. [22] providing the only such general de novo CAD study. Longer-term data is also needed. The finding of a reduced risk of thrombosis and peri-procedural MI in comparative observational studies is also of interest [26, 27]. Although reasons for why these results were seen have been speculated, further investigation in a randomized setting would be of use.

Furthermore, deriving conclusions on the angiographic superiority of a DCB-only approach versus the DES is difficult, as the majority of studies used LLL, which naturally favors DCB-only PCI. LLL as an angiographic endpoint should no longer be used due to the larger acute luminal gain seen in DESs, rather studies can use  %DS (which is less influenced by this) or focus on clinical MACE and TLR as primary endpoints. This being said, given the low rate of adverse clinical outcomes occurring in all PCI procedures, comparison between a DCB-only approach and DES may be difficult to characterize based on these alone. As such, angiographic data still have an important role to play. Additionally, the emergence of studies using intravascular imaging such as OCT and IVUS has been useful in further characterizing the benefits of a DCB-only approach on the vasculature. Moreover, their use to guide the DCB-only procedure and to ensure a satisfactory result has been associated with improved outcomes [32‐34]. Future research should continue to adopt these techniques to supplement angiographic data where possible.

Studies comparing the use of DCB-only PCI versus POBA have shown superior angiographic outcomes as expected, however MACE and TLR rates were largely similar and this may be attributed to a short follow-up period in these studies. Although studies comparing POBA are useful to characterize the additional benefits of drug elution, their scope for influencing clinical practice is limited. As such, there should be less emphasis placed on the importance of further such investigation.

There is a wealth of registry data regarding the use of a DCB-only approach for the treatment of de novo CAD. However, their rates of clinical events are highly heterogeneous. This may be attributed to a large variation in follow-up period, patient characteristics, rates of bail-out stenting and experiential and operator factors. A lack of true consensus for the definition of the composite clinical outcome MACE may also be of importance.

Data regarding the use of a DCB-only approach in PPCI are promising, with Gobic et al. [51] showing comparable short-term clinical outcomes to a 3rd-generation DES in a randomized study. Additional longer-term studies are required to further characterize this. Conversely, the use of a DCB-only approach in bifurcation is limited by a lack of data in lesions involving the proximal MB in addition to a lack of randomised studies comparing to DES therapy of any kind. The current strategic mainstay of bifurcation PCI is through MB stenting with provisional stenting of the SB. Ideally, randomized studies specifically comparing these two strategies are needed. The use of DCB-only PCI in the treatment of calcified lesions and CTOs is still in its infancy, with only a small number of single-armed studies that at best point towards possible feasibility as opposed to efficacy.

As expected, due to a lack of foreign body placement in the vasculature, a DCB-only approach was associated with a shorter duration of DAPT of 1–3 months when compared to DESs, which typically required a minimum of 12 months. Given that the majority of studies show comparable clinical outcomes between DCB-only and DES PCI, the shorter duration of DAPT appears to be well tolerated. This presents a key advantage of a DCB-only strategy, especially in cases where long-term DAPT is contraindicated. It should be noted that for the treatment of ACS, DAPT is given for 12 months according to European guidelines, thus the benefit of reducing DAPT for the purposes of PPCI has not been seen [4].