Malapposition: is it a major cause of stent thrombosis?

This editorial refers to ‘Mechanisms of stent thrombosis analysed by optical coherence tomography: insights from the national PESTO French registry’†, by G. Souteyrand et al. on page 1208. Stent thrombosis (ST) is a devastating complication associated with a high morbidity and mortality in patients who have undergone percutaneous coronary intervention (PCI) for coronary heart disease (CHD). In-hospital mortalities of patients suffering from ST have been reported to be as high as 7.9% for acute ST (AST; occurring <24 h after PCI) and subacute ST (SAST; occurring within 30 days after PCI), 3.8% for late ST (LST; occurring within the first year after PCI), and 3.6% for very late ST (VLST; occurring 1 year after PCI).1 Recently, longer duration (30 months) of dual antiplatelet therapy (DAPT) has been shown to reduce, but not eliminate, thrombotic events following drug-eluting stent (DES) implantation as compared with the traditional 12 months of therapy.2 While the introduction of newer generation DES with thinner struts, more biocompatible polymers, and lower drug load have reduced ST rates, it still continues to be a significant cause of morbidity.3 From the SCAAR (Swedish Coronary Angiography and Angioplasty Registry) registry of patients undergoing PCI for acute myocardial infarction, the cumulative rates of definite stent thrombosis up to 3 years were 2.0% for bare metal stents (BMS), 2.1% for first-generation DES, and 1.3% for second-generation DES, while cumulative rates of all-cause mortality up to 3 years were 12.8% for BMS, 10.3% for first-generation DES, and 10.6% for second-generation DES.3 While human pathology examinations of patients who suffered ST have helped us understand associations between various morphological and histological findings within the stented segment and ST, it remains unclear how relevant these findings are to ‘real world’ patients. Moreover, the limited numbers of patients used in these types of analysis might not be sufficient to understand the complex interplay between patient- and stent-related factors that undoubtedly influence risk for ST. Understanding the causes of ST in all its temporal phases in living patients may go a long way towards improving outcomes in patients with CHD who have undergone PCI by allowing us to find methods to eliminate its causes. In this issue of the journal, Souteyrand et al. report on the causes of stent thrombosis from the largest prospective cohort of patients (n = 120) collected to date from 29 French hospitals during January 2013 to October 2014 who presented with acute coronary syndrome (ACS) and had optical coherence tomography (OCT) performed either in the acute setting of ST or deferred during planned angiography.4 The primary endpoint was the underlying mechanism of ST as determined by OCT analysis. The distribution of ST cases revealed that the majority (i.e. 75%) of cases presented VLST, while 15% were SAST, 6% LST, and 4% were AST in presentation. OCT-derived mechanisms of ST were observed in 97% of cases, therefore substantially improving upon the sensitivity of angiography for the detection of the underlying aetiology of ST. The most frequently identified abnormalities were strut malapposition in 34% of cases, followed by neoatherosclerosis (23%), major stent under expansion (11%), coronary evaginations (8%), isolated uncovered struts (8%), edge-related disease progression (8%), intimal hyperplasia (4%), edge dissection (1%), and no cause identified (3%) (Figure 1). Causes of acute, subacute, late, and very late stent thrombosis. (A) Uncovered struts may be present in isolation or in the presence of malapposition, coronary evaginations, and also underexpansion (B). Malapposed stent struts are often covered by fibrin/platelet thrombus (Thr). Evaginations are defined as outward bulges in the luminal contour between struts and are a frequent finding in first-generation drug-eluting stents (DES) and less frequent in second-generation DES. (B) Underexpansion <3.00 mm2 may be a cause of uncovered struts and malapposition. Late stent thrombosis can occur form underlying excessive neointimal hyperplasia, and very late stent thrombosis from neoatherosclerosis. Intriguingly, malapposition was the most frequent finding in both acute and chronic cases, which contradicts previous reports describing mechanistic insights into the underlying causes of ST.5,6 This difference may perhaps be explained by the fact that OCT detects malapposition at much greater sensitivity as compared with angiography. However, the adjudication of malapposition as the dominant cause of ST may be questioned in the current analysis as stents with any major malapposed struts (>200 μm) were defined as malapposed regardless of the number of struts that demonstrated this finding. This issue is further highlighted by the apparent discrepancy between strut level-based analysis of malapposition and its overall adjudication as a dominant cause of ST, since malapposition was observed in nearly all mechanisms. The maximum malapposition distance was 710 μm and the median percentage of malapposed struts was only 12.7%, which raises concern as to whether malapposition was clinically relevant in only 34% of cases. Our pathology data show that malapposition is a greater problem in AST and SAST, with an average of 34% of sections with thrombus showing the presence of malapposition vs. 18% in those with patent stents. Malapposition distance was greater in those with thrombosis (204 μm) as compared with those without (51 μm),7 whereas at late and very late time points the presence of malappostion as an underlying cause of ST was predominantly seen in first-generation DES.8 Thus malapposition might not be a binary variable as treated by Souteyrand et al. but rather the extent of malapposition might be important in determining risk for thrombosis. Neoatherosclerotic plaque rupture was the second most frequent cause of ST (23%; 14% of LST and 29% of VLST cases). Neoatherosclerotic rupture was observed in 17 of 47 (36%) BMS and 10 of 71 (14%) DES patients. However, in our autopsy study, rupture and thrombosis from neoatherosclerosis occurred in 83% of BMS (5 of 6), 15% of first-generation DES (5 of 33), and none of second-generation DES.9 Beyond 3 years, all VLST in BMS (5 of 5) and 33% of VLST in first-generation DES (4 of 12) occurred from neoatherosclerotic plaque rupture.9 Only one patient in the PESTO French registry had LST associated with neoatherosclerosis, and it is unclear whether this occurred within a BMS or DES. Despite the premature onset of neoatherosclerosis in DES (median duration 361 days) as compared with BMS (median duration 721 days), at autopsy the earliest time frame for in-stent rupture to occur was 61 months in BMS and 23 months in DES.9 In patients with in-stent rupture from neoatherosclerosis reported in the PESTO registry, the delay between index PCI and ST was significantly longer in BMS [10.6 years (8.8–14.8)] than in DES [5.3 years (2.4–7.0)], which is consistent with our autopsy findings. Uncovered struts (>20%) were also frequently observed in the PESTO French registry,4 accounting for 74.9% of cases with no cause identified, 49.5% in cases with coronary evaginations, 39.1% in cases with evidence of severe under expansion, 34.5% in cases with isolated uncovered struts (all observed in the VLST group), and 21.2% in cases with malapposition, However, the presence of uncovered stent struts as an isolated finding and dominant mechanism of ST was only reported in 10 cases (8%), and all occurred very late.4 This observation contradicts our pathological findings where we have previously identified uncovered struts as a major cause of LST and VLST.10 In these cases, delayed endothelial coverage was the primary risk factor, with eosinophil and fibrin as associated findings. We reported stents with >30% uncovered struts to be a significant risk factor for LST and VLST, with a nine-fold increased odds ratio for thrombosis compared with a stent with complete coverage.10 In our study, we showed that there was a good correlation between the mean number of uncovered stent struts and the average distance between struts, with most uncovered struts showing less inter-strut distance than covered struts.10 Furthermore, the current study failed to detect a dominant cause of ST in 18% of cases presenting in the acute and subacute phase following the index procedure. Since the majority of stent struts are uncovered in the acute and subacute phase following stenting, their causative role in inducing thrombotic occlusion can only be determined when clinical information regarding anticoagulation therapy is available. Evaginations within the stented segment were first reported in patients with sirolimus-eluting stent (SES) implantation, and have been defined as outward bulges in the luminal contour between struts.11 They have also been reported as a frequent finding in patients with peri-stent contrast staining with incomplete stent apposition.11 Evaginations have been less frequently seen in second-generation DES (37% of everolimus-eluting stents) vs. first-generation DES (71% of sirolimus-eluting stents) at 1 year.12 Coronary evagination/persistent contrast staining has been associated with inflammation and hypersensitivity, with VLST reported in 8% of patients.13 In the PESTO French registry, evaginations were associated with ST in 14% of cases (n = 1) of LST and in 10% (n = 9) of VLST, most of them occurring in DES (n = 9).4 The PESTO French registry is an import first large OCT study performed in >100 patients that identified the underlying mechanism of ST. There is no doubt that the findings need to be corroborated by others as there are too many inconsistences and differences between the OCT findings and our pathological studies, with the most glaring of these being ST from malapposition which occurred at all time points, i.e. early and late after stenting, with greater frequency at late time points. Despite this shortcoming, it is likely that OCT-guided PCI might help prevent early- but not late-acquired malapposition. However, achieving complete apposition may not be realistic in all types of lesions being stented, especially those with a heavy eccentric calcium burden. Malapposition is a frequent finding at the time of implant and has been reported to occur in 26.6% by OCT and 13.6% by intravascular ultrasound (IVUS)14 even during intravascular imaging-guided PCI. Also, the presence of malapposition at baseline is not always associated with late complications as some studies have shown that more than two-thirds of cases of early malapposition resolve during follow up.15 When looking at prospective patient outcomes, the most important shortcoming of the present study relates to the timing of OCT imaging, which was left to the operators' discretion, permitting imaging not only in the acute ST setting but also during deferred PCI. Although this pragmatic approach certainly enabled investigators to acquire OCT pullbacks of sufficient image quality, the impact of time lapse between rescue PCI and deferred imaging procedures may affect the OCT findings. In addition, the present study did not address the interaction between patient- and lesion-related factors. Another drawback is that data on antiplatelet treatment are only provided for the entire cohort, where individual breakdown of therapeutic strategies would be required to understand if differing antiplatelet regimens may have also played a role in the aetiology of ST. Ultimately, to improve our knowledge of ST further, we will need to understand more clearly the mechanisms by which the morphological findings reported by Souteyrand et al. interact with patient-related factors to cause ST. However, the identification of major underlying abnormalities within the stented segment represents an important foundation in our understanding of the mechanisms of ST. Conflict of interest: the CVPath Institute has research grants from Medtronic CardioVascular, Abbott Vascular, Terumo Corporation, Atrium Medical, Boston Scientific, Cordis/Johnson & Johnson, OrbusNeich Medical, and Biosensors International. R.V. has speaking engagements with Merck; receives honoraria from Abbott Vascular, Boston Scientific, Lutonix, Medtronic, and Terumo Corporation; and is a consultant for Biomedical, Abbott Vascular, Medtronic, and W.L.Gore. A.V.F. has sponsored research agreements with Medtronic CardioVascular and Boston Scientific; and is an advisory board member to Medtronic CardioVascular. M.J. is a consultant for Biotronik; receives grant from the European Commission; and has received speaking honoraria from Abbott Vascular, Biotronik, Medtronic, Boston Scientific, Medtronic, and Orbus Neiche. H.M. has received honoraria from Abbott Vascular, Goodman, and Terumo Corporation.