Clinical , angiographic profile and predictors of in-hospital mortality in percutaneous treatment of lesions in saphenous vein grafts

Background: The angiographic characteristics associated with saphenous vein graft degeneration and the high-risk profile of these patients increase the probability of adverse outcomes during and after percutaneous coronary intervention. This study set out to analyze the clinical and angiographic profile of patients, procedural characteristics, and hospital outcomes of percutaneous coronary intervention performed in saphenous vein grafts, and to investigate predictors of in-hospital mortality in this group. Methods: A retrospective, observational study based on records kept by Central Nacional de Intervenções Cardiovasculares (CENIC) between 2006 and 2016. A comparative analysis of the adverse outcomes – periprocedural acute myocardial infarction, need for urgent coronary artery bypass grafting, and all-cause mortality – was performed according to different time periods (2006-2008, 2009-2011 and 2012-2016). Results: A total of 2,361 patients were included in the analysis. The prevalence of periprocedural acute myocardial infarction and mortality did not differ between time periods. No patient in this sample required urgent coronary artery bypass grafting. Simple logistic regression analysis revealed the following inhospital mortality predictors: advanced age, ST-segment elevation acute myocardial infarction, Killip class 3/4, long lesions, thrombi-containing lesions, three-vessel disease and periprocedural acute myocardial infarction. According to multiple logistic regression analysis, age (OR 1.07; 95%CI 1.021.13; p=0.01), smoking (OR 3.26; 95%CI 1.13 – 9.39; p=0.03), ST-segment elevation acute myocardial infarction (OR 10.36; 95%CI 3.96-27.07; p<0.01) and periprocedural acute myocardial infarction (OR 86.08; 95%CI 15.81-468.63; p<0.01) were correlated with mortality outcomes. Conclusion: Identification of in-hospital mortality predictors may contribute to improve procedural planning for adverse events prevention in patients undergoing percutaneous coronary intervention of saphenous vein grafts.


INTRODUCTION
Saphenous vein grafts (SVG) are extensively used in coronary artery bypass grafting (CABG) procedures and, compared to arterial grafts, are associated with unfavorable outcomes due to graft patency issues, with occlusion of one or more anastomoses occurring in up to 41% of cases within one year of surgery. 1 Therefore, percutaneous coronary intervention (PCI) of SVG is common, accounting to 5 to 10% of all interventions performed in clinical practice. 1ein graft failure may be acute or late (within 30 days or beyond 1 year of CABG, respectively) and pathophysiologic features differ in each case. 2 While factors associated with surgery and preexisting anatomy, such as endothelial denudation during vein graft preparation, diffuse native coronary artery disease, competitive flow and anastomotic suture problems prevail in the acute phase, atherosclerotic degeneration, often diffuse, with friable plaques and high thrombotic load, is more common in the late phase. 3,4Due to these characteristics, PCI for treatment of late SVG failure is therefore more challenging.Distal embolization and slow-flow or no-reflow phenomena may occur in up to 10 to 15% of cases, with greater likelihood of periprocedural acute myocardial infarction (MI) and increased patient morbidity and mortality. 5,6PCI of SVG results in higher rates of restenosis, repeat CABG, acute MI or death, as compared to de novo coronary artery PCI, even when drug-eluting stents and more advanced technologies are used. 1Hence, native coronary artery intervention is preferred (provided the vessel is patent), although not always feasible in these patients.
[9][10] However, robust comparative data of PCI and repeat CABG are lacking.Higher in-hospital mortality rates have been reported in patients undergoing repeat CABG as compared to graft PCI, despite similar long-term clinical outcomes and higher repeat CABG rates in PCI patients. 11,12These findings support the applicability of PCI in this subgroup of patients, particularly in the presence of patent internal mammary artery grafts and favorable anatomy. 2 Optimized clinical treatment is also an alternative for selected cases, and should be the basis of any kind of invasive treatment.
This study set out to analyze the clinical and angiographic profile of revascularized patients, procedural characteristics and hospital outcomes of PCI performed in SVG, and to investigate predictors of in-hospital mortality in this group.

Definitions
Periprocedural acute MI was defined as elevation of markers of myocardial necrosis by more than five times the upper limit of the reference range, or a minimum of 20% rise if previously altered, combined with symptoms suggestive of myocardial ischemia, new ischemic electrocardiogram changes, left branch bundle block, angiographic loss of coronary patency and/or flow, or imaging evidence of loss of viable myocardium, or changes in segmental myocardial contractility. 13Urgent CABG was defined as any repeat target lesion revascularization, occurring during the same hospitalization, due to ischemic signs and/or symptoms, and angiographic luminal diameter reduction ≥50%.Mortality was defined as death occurring during the same hospitalization, including cardiac or non-cardiac causes.Clinical and angiographic characteristics were defined according to current national guidelines. 2

Statistical analysis
Categorical variables were expressed as frequency and percentage.Time periods between 2006 and 2016 were compared using the Chi-square test.The likelihood ratio test was as used as needed.Continuous variables were ex-pressed as mean, standard deviation, median, minimum and maximum values.Time periods were compared using analysis of variance (ANOVA).Multiple comparisons were performed using the Bonferroni correction.Patients were the sampling unit in clinical variable analysis.For patients submitted to more than one procedure, the selection of procedure was made at random.
The contribution of variables of interest to mortality was investigated using simple and multiple regression models.The forward selection procedure was used to investigate the contribution of each independent variable to mortality in order to identify those that best ex plained the occurrence of death.Variables with large amounts of missing data, such as left ventricular dysfunction and presence of collateral circulation, were not considered.The level of significance was set at 5% (p<0.05) for all analyses.

RESULTS
This study included 2,361 patients with previous CABG and submitted to PCI of SVG, listed in CENIC database, totaling up 2,464 procedures in 2,872 vessels, with stent implantation in all cases.The mean number of procedures per patient was 1.04.Mean patient age was 67.4 years; male patients prevailed and 33.4% were diabetic, with no significant differences between the three groups (Table 1).Overall, 14.4% of patients were smokers, with higher smoking prevalence in the 2006-2008 group.The percentage (44.1%) of patients with a history of previous PCI was higher in the 2012-2016 compared to remaining periods (32.5% and 26%, 2009-2011 and 2006-2008, respectively; p<0.01).
Clinical presentation also differed significantly.The per centage of patients with stable angina and ST-segment elevation myocardial infarction (STEMI) was higher between 2006 and 2008 compared to other time periods.Silent ischemia prevailed between 2012 and 2016, while non-STsegment elevation acute coronary syndrome (NSTE-ACS) was more often diagnosed between 2009 and 2011.The percentage of STEMI patients with Killip class 3 or 4 was significantly higher between 2012 and 2016 compared to remaining time periods.
Among the angiographic characteristics displayed in table 2, the incidence of thrombotic lesions differed significantly between groups, with higher rates reported bet ween 2006 and 2008.The prevalence of vessels with Thrombolysis in Myocardial Infarction (TIMI) grade flow 0 or 1 was higher in the 2006-2008 and 2012-2016, as compared to the 2009-2011 group.
Procedural data are shown in table 3. The use of drugeluting stents increased significantly in more recent years, while the use of glycoprotein IIb/IIIa inhibitors dropped.Primary or rescue PCI were more common between 2006 and 2008 compared to remaining time periods.The use of thromboaspiration devices was limited to 1.3% of procedures and was more frequent between 2009 and 2011.Angiographic success was achieved in 96% of procedures, with no differences between groups.Considering in-hospital outcomes, periprocedural acute MI and mortality rates were low and did not differ between the three time periods investigated.Patients in this sample did not require urgent CABG (Table 4).The simple logistic regression model interrogated the contribution of each independent variable to mortality outcomes (Table 5).Patients treated between 2006 and 2008 had 3.74 higher chances of dying compared to those treated from 2009 to 2011 (OR 3.74, 95%CI 1.04-13.45;p=0.043).The risk of death also increased with age -7% increase per additional year of life (OR 1,07; 95%CI 1.02-1.12;p<0.01).As regards clinical presentation, STEMI patients had higher risks of death compared to those with other clinical presentations (OR 11.27; 95%CI 4.73-26.86;p<0.01), as did those progressing to Killip 3 or 4 (OR 42.07; 95%CI 8.49-208.41;p<0.01).Among angiographic characteristics, long lesions (OR 2.51; 95%CI 1.5-5.98;p=0.04), presence of thrombi (OR 2.98; 95%CI 1.23-7.25;p=0.02) and three-vessel disease (OR 5.32; 95%CI 1.23-22.99;p=0.03) were associated with higher mortality rates.Likewise, patients developing periprocedural acute MI (OR 21,5; 95%CI 5.82-79.47;p<0.01) had higher chances of dying during the hospitalization period.
The following in-hospital mortality predictors (   Major in-hospital mortality predictors were age, smoking, STEMI and periprocedural acute MI.It also revealed a predominantly aged population comprising more than 30% of diabetic patients, most of whom suffered from triple vessel coronary disease and left ventricular dysfunction, manifested primarily as acute coronary syndrome.Complex angiographic lesions prevailed in the sample studied.These findings reflect the high-risk clinical profile of patients undergoing PCI of SVG. 14 Compared to native vessel PCI, SVG interventions are associated with higher risks of adverse events in the short term, particularly periprocedural acute MI and related morbidity and mortality. 5,6This may reflect graft degeneration and ensuing atheromatous, thrombotic and friable lesions.SVG interventions were initially associated with 30-day mortality rates of 8%, compared to less than 1% mortality reported in contemporary studies. 15Overall mortality rate in this study was less than 1%, and urgent CABG was not required.
7][18] Elevation of biomarkers of myocardial necrosis has been associated with increased late mortality rates following PCI of SVG. 5,15Despite the low incidence of similar complications in this study, important correlations between periprocedural acute MI and in-hospital mortality were detected, emphasizing the relevance of preventing myocardial necrosis in susceptible populations.Limitations inherent to the data input method adopted in the database employed precluded the analysis of actual rates of embolic protection device use in the population studied.
Age was among the clinical variables with greater impact on in-hospital mortality, with each additional year of life accounting for a poorer prognosis.This finding is in keeping with previous reports of increased 30-day adverse outcome rates with age, in patients submitted to PCI of SVG. 18According to estimates of the American Heart Association (AHA), STEMI in-hospital mortality rates range from 5% to 6%. 19 Patients with previous CABG presenting with STEMI are less prone to acute reperfusion, and tend to have less successful angiographic outcomes and higher 90-day mortality, as compared to patients without previous CABG, particularly when the culprit vessel is the vein graft. 20Of those previously submitted to CABG, STEMI

DISCUSSION
This study involved 2,361 patients submitted to PCI of SVG and revealed high procedural success and low in-hospital complication rates during the time periods considered.

6
patients undergoing PCI had higher chance of death than other clinical presentations in this sample.
Left ventricular dysfunction is thought to be predictor of in-hospital mortality in patients requiring PCI of SVG. 21n this study, left ventricular dysfunction was not associated with higher mortality rates.However, higher chance of death was documented in patients with acute heart failure Killip class 3 or 4, emphasizing the relevance of this classification method for risk stratification in acute coronary syndrome patients. 22omen had a worse prognosis following native artery PCI compared to men.Different from this study, female sex was a predictor of 30-day in-hospital mortality in patients submitted to PCI of SVG in previous investigations. 21Diabetes mellitus did not increase the chance of in-hospital death in the population studied.A different correlation might have been found in the long term, as previously reported in several coronary disease scenarios, even in patients submitted to PCI of SVG. 21Smoking prevalence differed significantly between time periods in this study.This variable was associated with increased in-hospital mortality in this analysis, and was a predictor of 30-day cardiovascular outcomes in patients submitted to PCI of SVG in previous studies. 18ngiographic variables are also important predictors of adverse events in this population.Angiographic measurements of plaque volume, lesion length, presence of thrombi and graft degeneration scores are thought to be correlated with increased 30-day adverse outcome rates. 15,18In this sample, higher mortality rates were observed in patients presenting with lesions longer than 20mm, thrombotic lesions and three-vessel coronary disease.4][25] Increased bleeding risks and the fact that such agents are oftentimes used in patients more likely to present adverse events may explain these findings.Routine use of these drugs in SVG treatment as a mean to prevent distal embolization is contraindicated. 2ence, the use of glycoprotein IIb/IIIa inhibitors was less common in more recent periods in this analysis and was not related to increased in-hospital mortality.
The rate of drug-eluting stent use in the population studied was low and increased significantly over time.This may have reflected limited availability of this technology in Brazilian public health system.Meta-analyses of major studies comparing the use of first generation drug-eluting stents and bare metal stents in PCI of SVG demonstrated superior beneficial effects on late clinical outcomes, with no differences in mortality. 26,27Robust, randomized studies are warranted to evaluate new generation drug-eluting stent outcomes in PCI of SVG.
Aside from the limited use of drug-eluting stents, and the lack of data on use of embolic protection devices and related impacts on outcomes, limitations inherent to the design of observational studies based on volunteered data input sources must be accounted for.Longer follow-up time of the population studied may have led to the identification of other long-term adverse event predictors.Comparative analysis was limited by significant differences between patient populations and procedures at different time periods in this study.These differences may have reflected changes in population characteristics or the improvement of interventional techniques over time.However, in-hospital outcome rates did not differ between the time periods studied.

CONCLUSION
Angiographic characteristics reflecting saphenous vein graft degeneration and the high-risk profile of these patients emphasize the challenging nature of percutaneous coronary intervention of saphenous vein grafts.Periprocedural acute myocardial infarction, urgent coronary artery bypass grafting and mortality rates did not differ between time periods in this observational study.Age, smoking, ST-segment elevation acute myocardial infarction and periprocedural acute myocardial infarction were the major in-hospital mortality predictors in the selected population.
Identification of these variables may contribute to the decision-making process regarding selection of strategies in patients previously submitted to coronary artery bypass grafting, as well as to improved percutaneous coronary intervention planning for prevention of in-hospital adverse cardiac events, with potential impacts on late clinical evolution.

Table 3 .
Characteristics of procedures Results expressed as mean ± standard deviation, or n (%).PCI: percutaneous coronary intervention; TIMI: Thrombolysis in Myocardial Infarction.

Table 6 .
In-hospital mortality predictors by multiple logistic regression analysis

Table 4 .
In-hospital clinical outcomes