Innovation in interventional cardiology

INTRODUCTION

The limitations of an accurate assessment of the hemodynamic significance of coronary artery lesions through angiographic guidance alone are well-known.1 Instead, the fractional flow reserve (FFR) has proven to be a useful technique to address the coronary physiology and the hemodynamic significance of coronary segments before and after performing an intervention.2-4 Also, measuring FFR post-stenting has proven to be a strong and independent predictor of major adverse cardiovascular events at the 2-year follow-up.3-5

While FFR primarily takes into account the relative luminal narrowing and the amount of viable myocardium perfused by a specific vessel, several factors have been shown to impact the FFR values prior to performing a percutaneous coronary intervention (PCI). Therefore, longer lesion length, high syntax scores, calcifications, and tortuosity are associated with significantly lower FFR values. Conversely, the presence of microvascular dysfunction, chronic kidney disease and female gender have been associated with higher FFR values.6-11

At the present time, there is lack of data on independent predictors of post-PCI FFR. Therefore, the objective of the present study was to assess the patient and procedural characteristics associated with low post-PCI FFR in an all-comer patient population.

METHODS

The FFR-SEARCH study is a prospective single-center registry that assessed the routine distal pressure divided by the aortic pressure (Pd/Pa) and FFR values of all consecutive patients after an angiographically successful PCI. The primary endpoint was to study the impact of post-PCI FFR on the rate of major adverse cardiovascular event at the 2-year follow-up. Accordingly, no further actions were taken to improve post-PCI FFR. The study was performed in full compliance with the Declaration of Helsinki. The study protocol was approved by the local ethics committee. All patients gave their written informed consent to undergo the procedure. Also, anonymous datasets for research purposes were used in compliance with the Dutch Medical Research Act. A total of 1512 patients treated between March 2016 and May 2017 at the Erasmus Medical Center were eligible to enter our study. A total of 504 of these patients were excluded due to hemodynamic instability (156), a rather small distal outflow (129), the operator’s decision not to proceed with post-PCI hemodynamic assessment (148) or other reasons (79). A total of 1000 patients were included in the study. The microcatheter could not cross the treated lesion in 28 patients, technical issues with the catheter prevented post-PCI assessments in 11 patients, and in 2 patients the post-PCI FFR measurements had to be aborted prematurely due to adenosine intolerance. This left 959 patients whose post-PCI FFR values were measured in at least 1 angiographically successfully treated lesion.

Quantitative coronary angiography

The preprocedural lesion type was defined according to the ACC/AHA guidelines12 and divided into 4 categories: A, B1, B2, and C. Comprehensive quantitative coronary angiography analyses were performed pre- and post-stent implantation in all the treated lesions. An angiographic view with minimal foreshortening of the lesion and minimal overlapping with other vessels was selected. Similar angiographic views were used pre- and post-stent implantation. Measurements included pre- and postprocedural percent diameter stenosis, reference vessel diameter, lesion length, and minimum luminal diameter (MLD). In case of a total occlusion in patients presenting with ST-segment elevation myocardial infarction (STEMI) or chronic total coronary occlusion (CTO), the MLD was considered zero and the percent diameter stenosis, 100%. The reference vessel diameter and the lesion length were measured from the first angiographic view with restored flow. All measurements were taken using CAAS for Windows, version 2.11.2 (Pie Medical Imaging, The Netherlands).

Fractional flow reserve measurements

All FFR measurements were acquired using the Navvus RXi system (ACIST Medical Systems, United States), a dedicated FFR microcatheter with optical pressure sensor technology.13,14 Measurements were performed after an intracoronary bolus of nitrates (200 µg). The catheter was advanced while mounted over the previously used guidewire approximately 20 mm distal to the most distal border of the stent. The FFR was defined as the mean distal coronary artery pressure divided by the mean aortic pressure during maximum hyperemia achieved by the continuous IV infusion of adenosine at a rate of 140 µg/kg/min via the antecubital vein. In this study no vessels were assessed using intracoronary adenosine.

Statistical analysis

At baseline, the categorical variables were expressed as counts (percentage) and the continuous ones as mean ± standard deviation. To assess the independent predictors of post-PCI FFR, all the patient and vessel characteristics were primarily assessed through an univariate test using a mixed effects model (LME-model) with a random effect for the patients and a fixed effect for the post-PCI FFR. All variables were subsequently inserted in a multivariate LME-model using the enter method that resulted in all the significant independent predictors of post-PCI FFR values. A forest plot was developed to depict all variables with the corresponding 95% confidence intervals (95%CI). Beta (β) values show the average increase or decrease of the FFR values in the case of dichotomous variables or the increment per unit increase in the case of continuous variables. Statistical analyses were performed using the statistical software package R (version 3.5.1, packages: Hmisc, lme4 and nlme, RStudio Team, United States).

RESULTS

Demographic characteristics

The mean age was 64.6 ± 11.8 years and 72.5% were males. In 959 patients, at least, 1 lesion was measured with an overall 1165 successfully treated and measured lesions. The patient demographics and baseline characteristics are shown on table 1. Up to 70% of the patients presented with an acute coronary syndrome, and 18% had confirmed thrombus as seen on the angiography. Intravascular imaging modalities were used in 9.6% of the patients to guide the procedure. Overall, 1.4 ± 0.6 lesions were treated per patient and in 1.2 ± 0.5 lesions per patient the post-PCI FFR was successfully assessed. The average overall stent length per vessel was 29 mm ± 17 mm with an average stent diameter of 3.2 mm ± 0.5 mm.

Table 1. Baseline patient and vessel characteristics

The mean post-PCI FFR was 0.91 ± 0.07 and 7.7% of vessels had a post-PCI FFR ≤ 0.80. In the LME-model and after adjusting for independent predictors of post-PCI FFR, the left anterior descending coronary artery (LAD) as the measured vessel was the strongest predictor of post-PCI FFR (adjusted β = -0.063; 95%CI, -0.070 to -0.056; P < .0001) followed by the postprocedural MLD (adjusted β = 0.039; 95%CI, 0.015-0.065]; P = .002). Additionally, male sex, in-stent restenosis, CTO, and pre- and post-dilatation were negatively correlated with postprocedural FFR. Conversely, type A lesions, thrombus-containing lesions, postprocedural percent diameter stenosis, and stent diameter were positively correlated with postprocedural FFR. The R2 for the entire model was 53%. Figure 1 shows all significant and non-significant adjusted predictors included in the LME-model. Table 2 shows all adjusted and unadjusted predictors with corresponding β values and 95%CI. The most important predictors are shown on figure 2.

Figure 1. Forest plot of independent predictors of post-PCI FFR. Adjusted beta values with 95% confidence intervals. Triangles indicate significant predictors while circles are indicative of non-significant predictors in the multivariate generalized mixed model to predict post-PCI FFR. ACS, acute coronary syndrome; PCI, percutaneous coronary intervention; STEMI, ST-segment elevation myocardial infarction; LAD, left anterior descending coronary artery; CTO, chronic total coronary occlusion; MLD, minimum lumen diameter.

Table 2. Predictors for post-PCI FFR

Figure 2. Forest plot of most important predictors of post-PCI FFR. Adjusted beta values with 95% confidence intervals. The figure includes all significant predictors from the multivariate generalized mixed model predicting post-PCI FFR except for categorical variables with beta values < 0.02. LAD, left anterior descending coronary artery; CTO, chronic total coronary occlusion; MLD, minimum lumen diameter.

DISCUSSION

This study is the largest report to this day of predictors of post-PCI FFR. Based on data derived from the FFR-SEARCH registry, we could identify several patient and procedural predictors of post-PCI FFR. These predictors will bring more in-depth interpretations of post-PCI FFR values to be able to identify correctly which vessels are prone to future events. At first, male gender appeared to be negatively correlated with postprocedural FFR. This finding is consistent with the findings of former studies that focused on the impact of gender on pre-PCI FFR measurements.6,11,15,16 Compared to females, males are known to have a lower prevalence of microvascular dysfunction.8,17 The concept of FFR is based on drug-induced maximal hyperemia to minimize microvascular resistance. Microvascular dysfunction may hamper this vasodilator response and consequently result in a dampened flow response and high FFR.15 Subsequently, on average, males have larger myocardial masses and myocardial perfusion territories compared to females.18,19 The importance of the latter is illustrated by the second and strongest predictor of post-PCI FFR in this study, the FFR measurements in the LAD. FFR values are associated with the myocardial mass and the outflow territory of the measured vessel. As such, the LAD—the vessel with the largest perfusion area—has previously been associated with lower pre- and postprocedural FFR values.20-22

The diameters of the stents implanted in the RCA are larger, on average, but the outflow territory of the LAD is even larger.23 This discrepancy between luminal dimensions and myocardial mass may explain why the optimal improvement of the FFR measurements in the LAD is difficult to achieve.23

Thirdly, larger stent diameters and larger post-PCI MLDs were associated with higher post-PCI FFR values. However, higher postprocedural percent stenosis was also associated with higher post-PCI FFR values. While these findings may seem contradictory, post procedural percent stenosis was not associated with post-PCI physiology in the DEFINE PCI study either.24

In the intravascular ultrasound substudy of the FFR-SEARCH registry, van Zandvoort et al. showed that evident signs of residual luminal narrowing including focal lesions, underexpansion, and malapposition were present in a significant amount of vessels with post-PCI FFR values ≤ 0.85. These findings were not readily apparent on the comprehensive quantitative coronary angiography.25 Percent diameter stenosis was 20% in the cohort of patients with post-PCI FFR values ≤ 0.85 and > 0.85.26

Together with the latter predictors of post-PCI FFR we identified several others. A dedicated analysis of 26 CTOs recently showed that postprocedural FFR values are typically low initially; however they seem to increase at the 4-month follow-up. The initially low post-PCI FFR values is thought to be due to the microvascular dysfunction of the recently opened vessel, a phenomenon that improves after several months.27 In-stent restenosis and pre- and postdilatation were associated with lower post-PCI FFR values. A finding that is consistent with former studies that showed that, in general, complex lesions are associated with lower post-PCI FFR values.20,21,26,28

Also, it was interesting to see the impact of clinical presentation on post-PCI FFR values in the study population in which most patients presented with acute coronary syndrome. Contrary to former studies that questioned the validity of invasive hyperemic physiological indices in patients with acute coronary syndrome, we could not confirm the impact of clinical presentation on post-PCI FFR values. However, the identification of a thrombus, that often occurs after a ruptured plaque in patients with acute coronary syndrome, was associated with significantly higher FFR values. Despite the restoration of epicardial flow by the PCI, a relatively large number of patients with STEMI have abnormal myocardial perfusion at the end of the procedure.29 This phenomenon is thought to be related to microvascular obstruction due to distal embolization (reperfusion injury) and tissue inflammation due to myocyte necrosis.30,31 The latter may explain the significantly higher post-PCI FFR values reported in patients presenting with thrombus-containing lesions compared to those without such lesions. Conversely, our findings also show that in patients without thrombus-containing lesions the post-PCI FFR may be a valuable diagnostic tool for the identification of patients at a high risk of future adverse cardiac events.

Limitations

This study was conducted with the Navvus microcatheter, a dedicated rapid exchange microcatheter with a mean diameter of 0.022 in that proved its utility in a slight but significant underestimation of the FFR compared to conventional 0.014 in pressure guidewires.32 That is why we cannot directly extrapolate the current findings to wire-based FFR devices.14 Based on the study protocol, no further action was taken in the presence of low post-PCI FFR values. The Target FFR and FFR REACT studies (NCT03259815 and NTR6711) will provide further information on post-PCI FFR and the potential of further actions to improve post-PCI FFR and clinical outcomes.33,34 These studies should also focus on the trade-off of potential benefits and harm when performing additional interventions in order to improve the final FFR values.

CONCLUSIONS

In this substudy of the FFR-SEARCH registry, the largest real-world post-PCI FFR registry conducted to this day, we identified sex, LAD vessels, postprocedural MLD, and several other independent predictors of postprocedural FFR.

CONFLICTS OF INTEREST

L.J.C. van Zandvoort received institutional research support from Acist medical Inc. J. Daemen received institutional research support from Pie Medical, ACIST Medical Inc., PulseCath, Medtronic, Boston Scientific, Abbott Vascular, Pie Medical and speaker and consultancy fees from PulseCath, Medtronic, ReCor Medical, ACIST Medical Inc. and Pie Medical. The remaining authors declared no conflicts of interest.

REFERENCES

INTRODUCTION

Severe aortic stenosis is a common disease in our setting and has high morbidity and mortality rate. Its basic treatment is valve replacement.1 Over the last 2 decades, transcatheter aortic valve implantation (TAVI) has joined the traditional surgical aortic valve replacement (SAVR).2

Data are clear on the association between results and certain characteristics of the centers. The fact that has been most described in the medical literature is that, regarding mortality and complications, better results are obtained in those centers that reach the activity threshold (per center and per operator) for certain processes and procedures,3^-7 including coronary artery bypass graft (CABG)5,8 and primary angioplasty.9^-11 Regarding TAVI, the association between volume and results has been reported in hospitals in the United States.12^-14 In Germany, this association is less obvious.15 In Spain, the association between volume and results has also been reported for CABG.16

There are fewer studies that analyze the structural characteristics of the centers and their association with the characteristics of the healthcare systems of every country and the results obtained. In Spain, Bertomeu et al.17 found a lower mortality rate in patients with acute myocardial infarction (AMI) in high-volume centers with higher complexity. Worner et al.18 described a lower mortality rate in the management of AMI in hospitals with cardiac surgery and intensive care unit (CICU) capabilities. Rodríguez-Padial et al.19 found better results in the management of AMI in hospitals serving large communities. The association between CICU availability and better results has also been reported in our setting for the management of cardiogenic shock due to ST-segment elevation myocardial infarction.20

Our objective was to analyze the structural variables of the treating centers (availability of CICU), the volume of procedures performed and their association with results obtained after aortic valve replacement (whether through TAVI or SAVR).

METHODS

Population and sources of data

This is an observational and retrospective study of all the patients discharged from the hospitals of the Spanish National Healthcare System who underwent a SAVR o a TAVI procedure. The source of data was the minimum basic dataset (MBD) of the Spanish National Healthcare System of 2014 and 2015 (the only years available with a specific code for TAVI in the MBD). The clinical results of the patients transferred were assigned to the centers from which they were eventually discharged. Whenever the same episode was treated through TAVI and SAVR, it was considered as a TAVI treated episode and SAVR as a TAVI related complication. The main result variables were in-hospital mortality, length of the hospital stay, and in-hospital complications. The codes used for the complications seen are shown on table 1 of the supplementary data.

Table 1. Differences in the profile of patients and in the results of transcatheter aortic valve implantation based on the structural characteristics of each center (2014-2015)

Hospital structural variables

To analyze the possible correlation between the hospital structural variables and the results of aortic valve implantation both the volume of procedures performed and the cardiovascular resources available were studied. Hospitals were classified based on the availability of cardiology related resources and according to the RECALCAR criteria21 (table 2 of the supplementary data). Regarding this study, to analyze the inter-hospital differences, only those with cath lab capabilities without (type 3) and with cardiac surgery (type 4) were included. The availability of CICU based on a survey previously conducted by the Spanish Society of Cardiology was also included.22 The characteristics to consider the presence of a CICU were: a) a comprehensive capacity to manage patients in critical condition including invasive mechanical ventilation, and b) the administrative adhesion of the CICU to the cardiology unit.

Table 2. Differences in the profile of patients and in the results of surgical aortic valve replacement based on the structural characteristics of each center (2014-2015)

Statistical analysis

The risk adjustment models were specified based on the Centers for Medicare and Medicaid Services (CMS) methodology. Regarding CABG, the variables included in the 30-day mortality model were considered as independent variables.23 Also, certain variables anticipated by the Society of Thoracic Surgeons score for aortic valve replacement—and that can be identified in the MBD24—were included too. Finally, the CMS model was adjusted to the data structure of the MBD after gathering secondary diagnoses based on clinical categories.25 The multilevel logistics regression models were also adjusted.26,27 Only statistically significant comorbidities and odds ratio (OR) > 1.0 were considered for the adjustment model.

Based on specified models the risk-adjusted standardized mortality ratio (RA-SMR) was estimated.28 To adjust the length of the hospital stay, the Poisson regression model was used including the year of hospital discharge, the sex of the patient, and the degree of severity of groups related by refined diagnosis as risk factors. The expected length of the hospital stay was obtained from the individual predictions of the adjusted model. Also, the risk-adjusted length of stay ratio (RA-LOSR) was estimated as the coefficient between the length of the stay observed and the length of the stay expected.

To distinguish between high and low-volume hospitals (based on the number of episodes treated), a group clustering algorithm was used. To that end, the mathematical model used was developed with two thirds of the database and validated with the remaining third. The algorithm ranked as high-volume centers for TAVI those that performed ≥ 46 procedures, and as high-volume centers for SAVR those that performed ≥ 240 procedures during the study 2 year-period (2014-2015).

Quantitative variables were expressed as means ± standard deviations and the qualitative ones as frequencies and percentages. The correlation among the quantitative variables was analyzed using Pearson correlation coefficient. For comparison purposes, the Student t test for 2 samples and the analysis of variance (ANOVA) were used with correction of the level of significance using the Bonferroni method for ≥ 3 groups. Comparisons among the different categorical variables were conducted using the chi-square test or Fisher’s exact test.

All comparisons were bilateral, and differences were considered statistically significant with P values < .05. Statistical analyses were conducted using the STATA 13 and SPSS v21.0 software package.

RESULTS

A total of 2055 TAVIs and 15146 SAVRs were performed. Back in 2014 a total of 812 TAVIs were performed in 47 centers and in 2015 the number went up to 1243 in 53 centers.

The differences seen in the profile of the patients who underwent TAVI and SAVR are shown on table 1 and table 2, respectively, based on the type of hospital where procedures were performed. No statistically significant differences regarding age and sex were seen in patients who underwent TAVI in any of the 4 groups. Still, comorbidity was significantly higher (higher Charlson index and higher incidence of heart failure) in patients treated in type 3 non-CICU hospitals.

Regarding patients who underwent SAVR, by definition in type 4 hospitals, no statistically significant differences were seen regarding age, sex or presence of comorbidities among patients treated with and without CICU except for a higher prevalence of cardiogenic shock and previous percutaneous coronary interventions in non-CICU hospitals (2.0% vs 1.3%, P < .001; and 4.9% vs 3.9%, P = .004, respectively) (table 2).

The in-hospital mortality adjustment model for surgical aortic valve replacement showed good discrimination capabilities (area under the ROC curve, 0.84; 95% confidence interval [95%CI], 0.82-0.85) and calibration (P < .001). The model median odds ratio was 1.73, indicative of a high inter-hospital variability.

The SAVR specific in-hospital mortality adjustment model also showed excellent discrimination and calibration capabilities too (area under the ROC curve, 0.84; 95%CI, 0.83-0.84; calibration, P < .001) that were slightly lower for the TAVI specific adjustment model (area under the ROC curve, 0.79; 95%CI, 0.74-0.84; calibration, P < .001).

Characteristics of the treating center and TAVI results

Type 4 hospitals had a significantly lower RA-SMR compared to type 3 hospitals (4.04 ± 0.98 vs 4.47 ± 0.79). No statistically significant differences were seen on the RA-LOSR (0.99 ± 0.81 vs 1.07 ± 0.81; P = .278). The presence of a CICU was associated with a slightly lower, but sill statistically significant, RA-SMR (4.03 ± 0.87 vs 4.1 ± 1.07; P < .001). The correlation between CICU and a lower RA-SMR was also found in type 4 (4.03 ± 0.88 vs 4.05 ± 1.08; P < .001) and type 3 hospitals (4.09 ± 0.06 vs 4.59 ± 0.87; P < .001) (table 3).

Table 3. Differences in the results of transcatheter aortic valve implantation based on the characteristics each center (2014-2015)

CICU capable type 4 hospitals had a higher incidence of postoperative shock (1.8% vs 0.6%; P = .017), the same incidence of sepsis (0.8% vs 0.8%; P < .819), and a lower RA-SMR (4.03 ± 0.88 vs 4.05 ± 1.08; P < .001) compared to non-CICU hospitals.

Regarding the volume of procedures performed by the hospitals, the median of TAVI per year was 11 [2-36 for low-volume centers and 33 [9-67] for high-volume hospitals. The RA-SMR was lower in high-volume hospitals (3.95 ± 1.08 vs 4.26 ± 0.72; P < .001) (table 4 and figure 1). The mean adjusted stay did not show any differences between CICU capable and non-CICU hospitals (1.00 ± 0.85 vs 0.98 ± 0.77; P = .581). In general, regarding the crude complication rates, TAVI did not show any statistically significant differences between high and low-volume hospitals (table 4).

Table 4. Differences in the results of transcatheter aortic valve implantation based on the volume of cases of each center (2014-2015)

Figure 1. In-hospital mortality adjusted comparison for transcatheter aortic valve implantation (TAVI) and surgical aortic valve replacement (SAVR). Note that, regarding SAVR, in low-volume centers with CICU only 1 center was excluded. CICU, cardiac surgery and intensive care unit; RA-LOSR, risk-adjusted length of stay ratio; RA-SMR, risk-adjusted standardized mortality ratio.

Characteristics of the treating center and SAVR results

The presence of a CICU turned out to be a protective factor for in-hospital mortality in these patients (OR, 0.79; 95%CI, 0.67-0.93; P = .005). However, the different RA-SMRs seen among various centers with and without CICU capabilities did not show statistically significant differences (5.91 ± 1.49 with CICU vs 5.94 ± 1.72 without it; P = .335) (figure 1). The same thing happened with the RA-LOSR. CICU capable type 4 hospitals had a higher incidence of postoperative shock (2.2% vs 1.3%; P = .024) but a lower incidence of sepsis (1.1% vs 2.3%; P < .001) (table 5).

Table 5. Differences in the results of surgical aortic valve replacement based on the characteristics of each center (2014-2015)

In relation to the volume of procedures performed, the RA-SMR was lower in high-volume hospitals (5.89 ± 1.54 vs 6.27 ± 2.02; P < .001) (table 6) without any statistically significant differences with respect to the RA-LOSR (0.99 ± 0.73 vs 1.06 ± 0.75; P = .463). No statistically significant differences were seen between high and low-volume hospitals in the crude complication rates (table 6).

Table 6. Differences in the results of surgical aortic valve replacement based on the volume of cases of each center (2014-2015)

Association between TAVI and SAVR results

In type 4 hospitals, no statistically significant linear correlations were found between the RA-SMRs of TAVI and those of SAVR (r = 0.21; P = .14). Similarly, the SAVR high-volume variable had a non-statistically significant protective effect when it was introduced in the risk-adjustment model of TAVI related in-hospital mortality (OR, 0.73; 95%CI, 0.33-1.62). The 17 hospitals (1134 episodes identified) that shared the TAVI and SAVR high-volume feature had a TAVI related RA-SMR that was significantly lower compared to centers with the TAVI and SAVR low-volume feature (4 ± 1.1 vs 4.5 ± 0.7; P < .001). A single center (80 episodes) with a high-volume of TAVI and a low-volume of SAVR had the lowest TAVI related RA-SMR of all (2.8 ± 0.3; P < .001 with respect to a high-volume of TAVI and SAVR performed).

DISCUSSION

This study findings that included real-world data in our country, show a consistent correlation between the hospital structural characteristics and the results obtained in aortic valve replacement procedures, both surgical and transcatheter (figure 1). High-volume hospitals with cardiac surgery and intensive care units (CICU) have lower risk-adjusted mortality rates in both procedures.

In relation to the association between volume and results, our study also shows TAVI results that are consistent to those described by the medical literature,10^-14 with mortality rates that are similar to those seen in other countries in the study period (2014-2015) and higher to those published for 2015-2017.14 In Spain, the mortality rate differences seen after adjusting for high and low-volume centers are lower to the ones reported, which may be explained because, actually in those years in Spain, low-volume centers were being compared to very low-volume centers. Therefore, 52 out of the 53 center that performed TAVIs in Spain from 2014 through 2015 were within the range of the 2 lower quartiles (5-54 procedures per year), per volume of procedures performed, in the study conducted by Vemulapalli et al.14. Only 7 of those centers were above the range of the lower tercile in the study conducted by Kaier et al.15.

These data should be interpreted in the context of the learning curve of this technique in our country.29

The correlation between a higher volume and a lower RA-SMR was also found for SAVR. Again in this case, low-volume centers were being compared since only 12 and 10 out of the 42 centers, in 2014 and 2015 respectively, performed > 200 SAVRs, and over 70% of the centers were within the 2 lower quartiles of SAVR volume according to the study conducted by Hirji et al.30.

In this study, TAVI and SAVR high-volume centers had a lower TAVI-adjusted mortality rate compared to low-volume centers for both procedures, which is consistent with the findings reported by Mao et al.31. However, the only hospital identified as a high-volume center for TAVI and a low-volume center for SAVR had excellent TAVI results; since it was a single center with limited number of cases (4% of all TAVIs performed), this finding, suggestive that specific experience is more relevant than global experience in aortic valve replacement procedures, should be studied in the future. However, this is reasonable because it shows that here experience accumulates per processes or specific dedicated teams rather than centers in general.

Since no references were found in the medical literature, the newest finding of this study was the association between the presence of a CICU and the lower mortality rate reported for both techniques. This correlation is even more solid and clinically significant for TAVI rather than SAVR, which seems somehow intuitive, since patients treated with SAVR are often referred to general intensive care units.

The association between CICU availability and optimal results in the management of cardiogenic shock in the AMI setting20 had been described by the Spanish National Healthcare System. However, this association had not been reported in surgical procedures. Medical literature describes a virtuous relation between the volume of SAVRs performed and TAVI results, which is probably associated with the greater experience of the heart team.29,31 The presence of a CICU can be a variable that includes both the cardiologists’ greater experience and higher participation in the management of patients in critical cardiac condition and the experience of the hospital, cardiology unit, and cardiac surgery unit. In both cases, the CICU contributes to a better management of patients treated with interventional procedures (TAVI and SAVR) across the entire healthcare process.

Therefore, the results described may me important to plan healthcare and allocate resources such as teaching and training in the 2 aforementioned procedures.

Limitations

This study is a retrospective analysis of administrative data. However, even with its inherent limitations, the validity of its design has been compared to clinical registries.26,32 Such reliability allows us to compare the results of multiple hospitals33 and has been used specifically to analyze TAVI results.11-13,29,30 However, we should mention that data from the MBD should be interpreted with caution because they were not audited. Finally, this study shows the early experience with TAVI, probably still within the learning curve of this technique in the centers studied, which is why findings should be compared to more recent and larger series.

CONCLUSIONS

There is a correlation between the structural characteristics of the treating centers and the results obtained in aortic valve replacement, both surgical and endovascular, with great heterogeneity among the various centers. Large volume hospitals with cardiac surgery units and CICU capabilities have a lower risk-adjusted mortality rate in both procedures.

FUNDING

This study has been funded by an unconditional grant from the Interhospital Foundation for Cardiovascular Research.

CONFLICTS OF INTEREST

None reported.

ACKNOWLEDGEMENTS

We wish to thank the Health Information Institute of the Spanish National Healthcare System at the Spanish Ministry of Health, Consumer Affairs and Social Welfare for partially disclosing the MBD database.

REFERENCES

INTRODUCTION

The STENTYS Xposition S (STENTYS S.A., Paris, France) is a sirolimus-eluting self-expanding nitinol stent designed to adapt its size to the vessel diameter and facilitate its complete apposition when exerting chronic strength towards the outside. It has long been confirmed that one of the most important factors of stent thrombosis is the incorrect apposition of the stent.1 The characteristics of this stent make it especially useful to revascularize acute coronary syndromes (ACS), especially ST-segment elevation acute coronary syndromes with lesions with high thrombotic load. Also, a potential benefit in ectatic coronary vessels and lesions with great proximal and the distal diameter mismatch has been confirmed, bifurcations (left main coronary artery [LMCA] included), and venous grafts.

The objective of this study was to assess the benefit of this stent in the routine clinical practice by analyzing the type of lesions this stent is used with and the immediate angiographic results and at the clinical follow-up.

METHODS

A cohort of consecutive patients treated with the STENTYS Xposition S stent was analyzed from January 2018 through October 2019 in a tertiary hospital where over 1000 percutaneous coronary interventions are performed each year. All coronary lesions were quantified using a quantitative coronary angiography. Lesions in vessels with changes in size (ectasia or proximal-distal diameter mismatch of the lesion), in a bifurcation, in the presence of a high thrombotic load or in a venous graft were analyzed. The interventional strategy to be followed, imaging modalities included, was left to the operator’s criterion. The clinical and follow-up data were obtained from the electronical clinical records of the healthcare system of our autonomous community. All events were defined in a standard way according to the Academic Research Consortium-2 (ARC-2) consensus document.2

The data analysis was conducted using the IBM SPSS 20.0 statistical software package. Continuous variables were expressed as mean ± standard deviation or median with interquartile range depending on whether they followed a normal distribution or not, respectively. Qualitative variables were expressed as relative percentage. The cumulative incidence of events at the follow-up was estimated.

RESULTS

From January 2018 through September 2019, 1692 percutaneous coronary interventions with stent implantation were performed. The STENTYS Xposition S stent was used in 50 patients (62 lesions). The patients’ median age was 66 years [49-92]. Eighty-eight per cent of the patients were males. Table 1 shows the clinical characteristic of patients and coronary lesions. The most common clinical presentation was ST-segment elevation acute coronary syndrome in 23 patients (46%) followed by non-ST-segment elevation acute coronary syndrome in 22 patients (44%), and stable angina in 5 patients (10%). According to the classification established by the American College of Cardiology/American Heart Association the most common type of lesion was B1 lesion (38.7%). The right coronary artery was the most frequently treated vessel in 33 patients (53.2%).

Table 1 Clinical characteristics of patients and angiographic characteristics of the lesions

Ectasia and great proximal-distal diameter mistmatch at the lesion were the main indication for the use of this stent, in 72.6% of the lesions, with a mean vessel reference diameter of 4.1 mm ± 0.8 mm. A certain size was required to use this type of stent. The percutaneous coronary interventional on a bifurcation was the second most common indication, in 27.4% of the patients (2 of them on the LMCA). The most common type of bifurcation according to the Medina classification was 1-1-0, in 9 cases (52.9%). The secondary branch was damaged in 17% of the patients. The provisional stenting technique was the most widely used in 15 cases (88.2% of bifurcations) re-crossing to the secondary branch in 9 of them (60%). The dilatation of the secondary branch only occurred in 7 patients and only in the other 2 stents were implanted: one in a 0-1-1 bifurcation according to the Medina classification (minicrash technique) and the other in a 1-1-1 bifurcation according to this classification (TAP technique [T-and protrusion technique]). In both cases the STENTYS Xposition S stent was implanted in the main vessel and a non-self-apposing stent in the secondary branch (figure 1).

Figure 1. A: lesion with significant thrombotic load in the mid right coronary artery, which remains after thrombus aspiration. B: 3.5-4.5 mm × 27 mm Xposition S direct stent implantation. C: final angiographic result. D: significant stenosis in distal left main coronary artery. E: 3-3.5 mm × 27 mm Xposition S stent implantation from the proximal left main coronary artery to the left anterior descending coronary artery. F: angiographic result after postdilatation.

A high thrombotic load (Thrombolysis in Myocardial Infarction flow grade 4-5) was seen in 8 lesions. All of them in ectatic coronary vessels or with proximal-distal caliber mismatch. No case of venous graft treated with STENTYS was reported.

Predilatation occurred in 32 lesions (51.6%) and postdilatation in 37 (59.7%). The criterion used for postdilatation was angiography guided visual underexpansion. Intravascular ultrasound was performed in 15 patients (30%) before the implant. It was also used in 2 patients to optimize the percutaneous coronary intervention given the persistent stent underexpansion seen on the angiography. In both cases the minimum lumen area was > 5.5 mm2 with stent expansion > 80% and lack of incomplete apposition (defined as a strut separation of > 0.4 mm axial and 1 mm longitudinal) (figure 2). The optical coherence tomography was performed in a patient with ST-segment elevation acute coronary syndrome before and after the implant. It revealed a high thrombotic load with lack of immediate stent malapposition.

Figure 2. A: acute thrombotic occlusion in left circumflex artery with Thrombolysis in Myocardial Infarction flow grade 0. B: the intravascular ultrasound shows a great deal of thrombus in the lesions despite thrombus aspiration. C: implantation of 2 3.5-4.5 mm × 27 mm Xposition S overlapping stents. D: the intravascular ultrasound performed after stent implantation confirms the good angiographic results and lack of stent malapposition.

Angiographic success was achieved (with the stent properly implanted, a residual lesion ≤ 10%, and Thrombolysis in Myocardial Infarction flow grade 3) in all patients but 1, in whom stent implantation failed in a severely calcified LMCA lesion. In this case, predilatation was first attempted using a conventional balloon and then a cutting balloon on the LMCA severe distal lesion. A 3.3-4.5 mm × 22 mm STENTYS Xposition S stent was implanted with stent loss during retrieval, which remained braced to the guide catheter. Afterwards, a balloon-expandable drug-eluting stent was successfully implanted. The un-crimped stent was retrieved by crossing a guidewire from the femoral access through the stent distal struts. It was finally captured with a snare.

The median score obtained in the PRECISE-DAPT risk calculator (Predicting bleeding complications in patients undergoing stent implantation and subsequent dual antiplatelet therapy) was 16.5 (7-25), and the median score obtained in the DAPT index (Dual antiplatelet therapy) was 1.15 (−2-4). Ticagrelor was the most commonly used P2Y₁₂ inhibitor (58.1%). A 12-month course of dual antiplatelet therapy was prescribed in 48 patients (96%).

After a median follow-up of 373 days (256-439), 1 patient had an acute myocardial infarction 3 months after the intervention. However, the coronary angiography did not reveal coronary artery disease progression but confirmed the good results of the previous intervention. An 84-year-old woman died at admission due to heart failure. Three patients died of non-cardiac causes: 1 due to septic shock at admission, the other patient died 6 months after the percutaneous coronary intervention due to high-grade lymphoma, and the third one 4 months after the percutaneous coronary intervention due to lung cancer. No cases of definitive stent thrombosis or revascularization of the treated lesion were reported. No bleeding was seen either at the follow-up.

DISCUSSION

This type of stent is not widely used in our setting and we believe 2 are the reasons why. The first one is the need for a learning curve to know how to handle this implant. In former iterations of the device, the delivery system had some technical limitations like the jumping phenomenon that could occur right when the sheath was being released due to the elastic properties of nitinol. Unlike its predecessor (STENTYS sirolimus DES), the stent of the new STENTYS Xposition S system, is mounted over a semicompliant balloon and covered by a 0.0032 in-thick sheath. The reason for balloon inflation is not to dilate the stent, but to rupture the external sheath from the distal to the proximal border to allow a proper vessel-wall stent apposition. This has reduced the complexity of the release mechanism.3 However, we should remember that after the implant, the retrieval of both the balloon and the device sheath should be conducted with care by separating the guide catheter from the ostium to avoid deep intubation. The other reason that may explain why this stent is still not widely used can the augmented profile of the device and its rigidity, which both reduce its navigational and crossing capabilities compared to balloon-expandable stents.

Due to the characteristics of the stent and the experienced gained using it, the clinical settings where it can be useful are: ectatic vessel, since the stent reaches 6.5 mm of diameter with the device L size; proximal and distal diameter mismatch due to its adaptative capabilities to the vessel caliber; lesions with high thrombotic load, since this stent self-expanding capabilities facilitate its expansion until it reaches the vessel wall if thrombus reabsorption occurs, which avoids late stent malapposition; and bifurcations with ostial damage and 30º to 70° angles. The stent z-shaped mesh and the presence of small interconnectors facilitate re-crossing the lateral branch and disconnecting the struts without having to use the final kissing balloon technique. Thanks to its self-expanding capabilities, the unconnected struts cover the lateral branch ostium making the double stent technique unnecessary on many occasions.

In the studies published on former iterations of the device, the self-expanding stent proved superior to the balloon-expandable stent regarding better apposition. The randomized APOSSITION II clinical trial,4 conducted among patients with acute myocardial infarction, showed a lower rate of stent malapposition (defined as > 5% of struts per patient as seen on the optical coherence tomography) 3 days after the primary percutaneous coronary intervention. The APOSSITION IV clinical trial,5 also conducted among patients with acute myocardial infarction, showed a significantly lower percentage of stent malapposition at the 4-month follow-up in patients treated with self-expanding stents compared to patients treated with balloon-expandable stents (0.07% vs 1.16%; P = .002). However, no inter-group differences were found at the 9-month follow-up (0.43% vs 0.28%; P = .55) or in the rate of major adverse cardiovascular events (MACE). The clinical repercussions of this improvement in the early apposition of the stent has not been studied thoroughly. The APOSSITTION III trial6 showed that the use of STENTYS BMS in the percutaneous coronary intervention setting was associated with acceptable cardiovascular results at the 2-year follow-up, an overall rate of MACE of 11.2%, and a rate of stent thrombosis of 3.3%. We should mention that this study revealed a significant reduction of adverse events after the systematic implementation of a standard protocol (predilatation, implantation, postdilatation). The data available support the hypothesis of the need for mild postdilatation to avoid early complications probably because the stent does not have enough radial strength to achieve a proper expansion in rigid often calcified lesions, especially when predilatation is not fully effective. Therefore, postdilatation would avoid the incomplete expansion of the stent, which may increase the risk of stent thrombosis.7

Our study with the STENTYS Xposition S stent reached angiographic success in 98.4% of the cases, although we should remember that, from the anatomical point of view, they were not complex lesions (only 30% were type C lesion). Stent implantation failed in 1 severely calcified LMCA lesion; it is precisely in this type of lesions where its use is ill-advised, especially if predilatation is not effective.8

Regarding its use in bifurcations the studies published to this day have also discussed a former iteration of this device with good results. In the observational, multicenter, and prospective OPEN II trial,9 the rate of MACE at the 12-month follow-up was 13% (10.1% at 6 months). This rate of events was basically due to the need for revascularization of the treated lesion, while the rate of stent thrombosis at the 12-month follow-up was 1%. We should also mention that the kissing balloon technique was only used in 21.7% of the patients. Also, there were no significant differences in the rate of MACE between patients in whom the kissing balloon technique was used and those in whom it was not used.9

To this day, the only study published on the new STENTYS Xposition S model is the TRUNC, a prospective and multicenter study that assessed the efficacy and safety profile of this type of stent in the LMCA. Angiographic success was achieved in 96.6% of the patients and the overall rate of MACE was 8.3% at the 12-month follow-up, basically due to revascularization of the lesion treated in 5.4%.10 Here we should mention the preliminary results reported by the SIZING (Worldwide every-day practice registry assessing the Xposition S self-apposing stent in challenging lesions with vessel diameter variance) and WIN (World-wide registry to assess the STENTYS Xposition S for revascularization of coronary arteries in routine clinical practice) registries. Both registries confirm the safety and efficacy profile of the current iteration of the stent in the routine clinical practice.

Limitations

Our study has several limitations. Because of its retrospective, single-center nature and the limited number of cases involved, we cannot draw definitive conclusions on the device safety and efficacy profile. No intracoronary imaging modality was performed systematically to guide the implant, which may have been useful, especially the optical coherence tomography. However, we believe that this study is relevant due to the scarce evidence available on the last iteration of this stent.

CONCLUSIONS

In our series of lesions located in ectatic vessels or with proximal-distal diameter mismatch and in bifurcations, the STENTYS Xposition S stent is a good therapeutic alternative that achieves good immediate angiographic results and good mid-term clinical results.

CONFLICTS OF INTEREST

None declared.

REFERENCES

INTRODUCTION

The physiological assessment of coronary lesions is a routine practice in contemporary cath labs and is strongly recommended by the European guidelines to guide the percutaneous coronary intervention (PCI) decision-making process.1 Unlike fractional flow reserve, the new instantaneous wave-free ratio (iFR) index allows us to analyze the physiological significance of each lesion and each coronary segment.2-5 This has led to the creation of the new and specific SyncVision software package (SyncVision version 4.1.0.5, Philips Volcano, Belgium), that shows the functional compromise of each lesion and predicts the expected iFR improvement after percutaneous treatment.3,4

Few observational studies published have analyzed the reduction in the length of the stent implanted compared to angiography-guided revascularization in long and diffuse coronary lesions.4,5 However, this reduction could be detrimental to the complete coverage of the plaque in this type of lesions, which has proven to be a predictor of major adverse cardiovascular events at the follow-up.6

The objective of our study is to analyze the utility of the iFR and SyncVision software to guide the PCI decision-making process in long, sequential, and diffuse coronary lesions.

METHODS

We have designed a multicenter, randomized, controlled, and open-label trial to compare SyncVision/iFR-guided revascularization to angiography-guided revascularization in patients with long, sequential or diffuse significant angiographic coronary lesions (ClinicalTrials.gov identifier: NCT04283734). All the variables that will be analyzed during the study are shown on table 1.

Table 1. Variables that will be analyzed during the study

Inclusion and exclusion criteria

Patients with the following criteria are being included: a) patients > 18 years old who require percutaneous coronary treatment due to ischemia (silent, stable angina or acute coronary syndrome); b) presence of a vessel with sequential lesions separated by < 10 mm from each other with a total lesion length > 25 mm and a percent diameter stenosis > 60% (as seen on the quantitative coronary angiography assessment) in, at least, 1 segment; or a coronary segment > 30 mm with diffuse disease, and a percent diameter stenosis > 60% (as seen on the quantitative coronary angiography assessment) in, at least, 1 region; c) baseline iFR ≤ 0.89 distal to a potentially randomizable lesion.

We have stablished the following exclusion criteria: a) patients with acute coronary syndrome with non-optimal results in the culprit vessel (final Thrombolysis in Myocardial Infarction (TIMI) flow grade < III, non-reflow phenomenon during treatment, residual coronary dissection, lost or compromise of a major side branch); b) patients with acute coronary syndrome and left ventricular ejection fraction < 45%; c) life expectancy < 12 months; d) patients with severe aortic stenosis; e) contraindication for dual antiplatelet therapy for at least 12 months; f) presence of significant thrombocytopenia (< 10 x 109/L); g) patients with an indication for bypass surgery according to the heart team; h) pregnancy; i) inability to understand the informed consent.

Endpoints

The study primary endpoint is the reduction of the average length of the stent implanted in the SyncVision-guided group measured in millimeters (mm) compared to the angiography-guided group. The study secondary endpoint is a composite of cardiac death, myocardial infarction, definitive or probable stent thrombosis, new target lesion revascularization or new target lesion failure (major adverse cardiovascular events [MACE]); and the assessment of residual ischemia through single-photon emission computed tomography at the 6-month follow-up.

Procedure

After the diagnostic phase, the use of intracoronary vasodilators is mandatory to exclude possible coronary spasms. Lesions will be assessed by 2 expert operators (prior to randomization) to determine the coronary segment to treat when the revascularization is angiography-guided based on current routine clinical practice. Afterwards, the iFR will be determined at baseline. If the obtained iFR is ≤ 0.89, patients will be randomized to the angiography-guided revascularization group (the control group) or to the iFR pullback-guided revascularization group using the SyncVision software (figure 1). Intracoronary imaging can be used in both groups based on the operator’s criteria to optimize the angiographic result.

Figure 1. Summary of randomization, treatment targets, and follow-up of the iLARDI study. iFR, instantaneous wave-free ratio; MACE, major adverse cardiovascular events; PCI, percutaneous coronary intervention.

In the intervention group, a pressure wire (Verrata pressure guidewire, Philips Volcano, Belgium) will be inserted trough a guide catheter towards the vessel ostium to normalize the pressure between the aortic and the vessel ostium. Secondly, the pressure wire will be advanced distally to the lesion. Under stable hemodynamic conditions (without the administration of vasodilators), we will determine the baseline iFR. Afterwards, the wire will be removed under continuous fluoroscopy, and in the same projection. If the iFR at the vessel ostium is 1 ± 0.02, the absence of drift will be confirmed and an angiogram in the same angiographic position will be performed. The SyncVision software can recognize the vessel analyzed and identify the physiological contribution of every lesion and every segment, predicting the improvement of the iFR after treatment. The iFR improvement is depicted as yellow dots. Each yellow dot represents an iFR improvement of 0.01 if that zone was percutaneously treated. The accumulation of many yellow dots suggests that the contribution of that lesion to physiological compromise is high. After performing the physiological assessment of each lesion, the operator would have to treat the minimum segment needed to achieve an iFR of 0.90. Cases without an accumulation of dots have been considered as physiological diffuse disease (defined as the presence of < 20% of the total number of dots) in the coronary segment physiologically assessed. Those cases will be medically treated due to the theoretical absence of benefit of the percutaneous treatment (figure 2 and figure 3).

Figure 2. Flowchart of technical treatment details of patients randomized to the intervention group.

* We consider as optimization the postdilatation of the previous stented area if an in-stent accumulation of yellow dots is seen; or the percutaneous treatment of a new segment with physiological compromise not seen in the baseline iFR-pullback study. iFR, instantaneous wave-free ratio.

Figure 3. Image of iFR co-registration using the SyncVision software in a patient included in the study and randomized to the intervention group with a diffuse lesion in the left anterior descending coronary artery, and the physiological contribution of every segment. The estimated length of the stent to achieve an iFR > 0.89 is 50.6 mm.

Follow-up

Patients will be followed either through phone calls or physical examination at the 3, 6 and 12-month follow-up. At the 6-month follow-up a stress single-photon emission computed tomography (physiological or pharmacological) will be performed in all patients. The composite of cardiovascular death, definitive or probably stent thrombosis, new target lesion failure or new target lesion revascularization will be considered as MACE.

Quantitative coronary measurements

Quantitative coronary measurements will be performed using a validated system (CAAS system, Pie Medica Imaging, Netherlands). The measurements analyzed will be the vessel reference diameter, the vessel minimal lumen diameter, and the percentage of stenosis. All measurements will be taken at baseline and after the PCI.

Statistical analysis

Regarding the statistical analysis, quantitative variables will be expressed as mean ± standard deviation and qualitative variables as absolute numbers and percentages. To determine the relationship among quantitative variables, we will be using the paired Student t test for paired data. To determine the relationship among the qualitative ones, we will use the chi-squared test. In all cases, differences will be considered significant with P values < .05. We will be using the IBM SPSS Statistics software package (version 24.0 for Macintosh, SPSS Corp., United States). To calculate the sample size, we have performed a retrospective analysis of the last 20 patients who were treated at our centre and showed a sequential or diffuse lesion in the coronary vessel analyzed from the iFR-pullback study. The mean length of the stent implanted was 43 ± 9 mm and the reduction of stent length was 12 ± 8 mm on the angiographic analysis. With these data, we have stablished an expected length reduction of 15 mm. The calculated sample size to achieve the primary endpoint with an 80% confidence level and a 5% margin of error was 100 patients.

RESULTS

The recruitment of patients started back in February 2020. After 1 month, we have included the first 7 patients. We expect to complete the recruitment by February 2021 and the follow-up by February 2022.

DISCUSSION

To our knowledge, this randomized study is, the first one to assess the potential benefits of using the SyncVision software in long, sequential or diffuse coronary lesions. Currently, the study is in the recruitment phase and the first patients have already been recruited.

The iFR has proven to be useful in the PCI guide decision-making process.7,8 However, the evidence supporting the use of SyncVision is scarce and controversial in long, sequential or diffuse lesions. On the one hand, the software allows us to know the coronary segments with the highest physiological compromise. This allows us to revascularize only those segments that immediately improve the physiological result with a potential reduction of the length of the stent implanted, which happens to be a predictor of MACE at the follow-up.9 On the other hand, it’s possible that even if we obtain a good immediate physiological result and a reduction of the stent length implanted we won’t be fully covering the plaque in some lesions or coronary segments, which has also proven to be a predictor of MACE.6

A limitation of the study is the sample size, enough to achieve the primary endpoint, but probably inadequate to see differences in MACE. However, we think that it can provide an early insight on the utility of iFR pullback study to guide the PCI decision-making process in this type of lesion. Also, it can be a hypothesis-generator study for future larger-scale studies to show benefits in terms of clinical events reduction.

For these reasons, we believe that the iLARDI is an interesting study that will shows us the potential benefit of SyncVision to guide the PCI decision-making process in long, sequential or diffuse coronary lesions. We intend to complete the results by February 2022.

CONCLUSIONS

The iLARDI study is the first randomized trial to assess the potential utility of SyncVision-guided revascularization in long, sequential and diffuse coronary lesions.

FUNDING

Funds from the Plan Andaluz de Investigación, Desarrollo e Innovación (PAIDI) have been used to pay for the liability insurance associated with clinical research.

CONFLICTS OF INTEREST

F. Hidalgo, S. Ojeda, and J. Segura received personal fees from Philips Volcano. M. Pan received minor fees from Abbott, Philips Volcano, and Terumo. The remaining authors declared no conflicts of interest whatsoever.

REFERENCES

INTRODUCTION

Renal function impairment is associated with poor prognosis in patients with stable or acute coronary syndrome (ACS). One of the most common causes of acute kidney injury (AKI) in hospitalized patients is the nephropathy induced by the IV administration of contrast agents.1Its incidence varies between 1% and 6%, and increases considerably in high-risk conditions like in the ACS setting. The reported frequency of post-coronary angiography contrast-induced nephropathy (CIN) goes from 12% to 46% in patients with ACS.2^,3

There are several potential causes that trigger CIN in patients without a past medical history of kidney failure such as hemodynamic instability, the IV administration of contrast agents, thromboembolic events, and adverse drug reactions, among others. Also, it is important to consider the type of contrast used, its osmolarity, the volume administered, and the lack of preventive measures.4^-6

Because CIN is associated with poor prognosis in hospitalized patients, predictive scores have been designed to identify the most vulnerable patients who can develop this complication. The Mehran score is one of the most popular indices to estimate the chances of post-coronary angiography CIN.7

It is well-established that microalbuminuria is a predictor of kidney dysfunction mainly in diabetic and hypertensive patients.8^-14Also, there is a correlation between high levels of microalbuminuria and the poor outcomes seen in patients with ACS.15^-16 Currently, microalbuminuria is estimated through the dosage of the albumin-to-creatinine ratio (ACR) through a simple urine sample.17

The objective of this study is to calculate microalbuminuria using the ACR as a predictive variable of post-coronary angiography CIN in patients with ACS.

METHODS

Population

Patients with ACS consecutively admitted to the coronary care unit of a community hospital were analyzed. Those undergoing an in-hospital coronary angiography with non-ionic, hyperosmolar IV contrast agents such as iopamidol, optiray or xenetix, were included in the study. The volume of IV contrast for each angiographic study was calculated retrospectively. It was estimated that each injection of contrast material into the left coronary artery required an average 10 cc to 8 cc for the right coronary artery.

Patients with a past medical history of renal failure, macroalbuminuria, treatment with diuretics and patients with secondary angina were excluded from the study.

The urinary ACR was assessed in all patients included in the study using an immunoturbidimetric assay in simple urine samples within the first 24 hours after hospitalization.

Definitions

IV contrast-induced nephropathy(CIN) was defined as an increase in serum creatinine levels ≥ 25% 48 hours after performing the coronary angiography or an absolute increase of ≥ 0.5 mg/dL compared to levels at admission.

Microalbuminuria was defined as an abnormal urinary albumin excretion rate between 30 to 200 mg/min or 30 to 229 mg/day.

The study protocol was approved by the center review board and conducted in compliance with the Declaration of Helsinki, good clinical practice guidelines, and local regulatory requirements. Informed consents were obtained from all patients.

Biochemical considerations

A urine sample collected within the first 24 hours after admission (preferably during morning hours) was centrifuged at 3000 rpm and stored at -20° Celsius until biochemical analysis was conducted. The principle of the ACR test is immunoturbidimetry. This method is based on the reaction of human albumin antibodies to the antigen. Complexes are then measured after agglutination. The COBAS 6000 analyzer (ROCHE, Switzerland) was used to process the sample. The analytical detection limits of the assay were between 3 mg/g and 400 mg/g. The test variation coefficient was 3.8%.

Statistical analysis

The Kolmogorov-Smirnov test was used to analyze the distribution of continuous variables and their kurtosis-skewness measures. Data were expressed as mean and standard deviation or as median with interquartile range (25%-75%) and compared using Student’s t test or Mann-Whitney-Wilcoxon test for independent groups according to their parametric or non-parametric distribution, respectively.

Discrete variables were expressed as percentages and compared using the chi-square test. The cross-product ratio was expressed as odds ratio (OR) with its 95% confidence interval (95%CI). The C-statistic measure was used to detect the best ACR cutoff value associated with the primary endpoint and compare the discrimination capacity of the Mehran score alone and with the ACR combined.

A multivariable regression analysis will be built to predict CIN including ACR and adjusted using the Mehran score.

Both the IBM SPSS Statistics version 19 software and the MedCalc version 11.6.1 software (Mariakerke, Belgium) were used for statistical analysis and to calculate and compare the C-statistic measure. To test the additional predictive value of ACR, the C-statistic measurewas compared using the Mehran score alone and after adding the ACR information obtained.

RESULTS

Out of a total of 397 patients diagnosed with ACS, 148 (59.4%) underwent a coronary angiography during hospitalization and this was the study population. The mean age was 64 ± 12 years; 35% were women, 20% had diabetes, 54% dyslipidemia, 65% hypertension, and 42% were active smokers. The mean blood sugar levels on admission were 110 mg/dL (98-133 mg/dL), the median creatinine clearance rate (estimated using the MDRD) was 86 mL/min (66-107), and the ACR was 5 mg/g (0-14) (table 1). The patient comparison between these groups with or without CIN showed a higher rate of overweight and obesity, left bundle branch block, atrial fibrillation, and AMI Killip and Kimball class III-IV (table 2).

Table 1. Baseline characteristics of the patients

Table 2. Comparison of patients with and without contrast-induced nephropathy

The C-statistic measure showed that the best CIN related ACR cutoff value was 20 mg/g. Twelve patients developed CIN (8.1%) and the ACR of 22% of the patients was >20 mg/g. The rate of ACR > 20 mg/g among patients without CIN was 2.9% and 11.3% (P = .01) among patients with CIN. Contrast-induced nephropathy was significantly higher when the ACR was ≥ 20 mg/g compared to when it was < 20 mg/g (≥ 9.6% vs 1.6%, res­- pectively, P < .001). When the ACR was added to the Mehran score, its predictive power went up to 0.82 (95%CI, 0.76-0.87).(Figure 1).

Figure 1. Effect of albumin-to-creatinine ratio when added to the Mehran score. When the albumin-to-creatinine ratio was added to the Mehran score, its predictive power went up from 0.75 to 0.82 (95%CI, 0.76-0.87).

Using a multivariable regression analysis model the ACR > 20 mg/g turned out to be an independent predictor for CIN: OR, 3.2 (0.7-6.2); P = .01, adjusted by the Mehran score variables (age, women, body mass index, atrial fibrillation, Killip Class III-IV, and creatinine clearance rate).

DISCUSSION

Our study proved the association between the ACR and the development of CIN in patients admitted with ACS.

Acute kidney injury in the ACS setting predisposes to more complications such as in-hospital and long-term mortality; therefore, predicting it is of critical clinical importance. A recent study reported that the rate of AKI was close to 17% in the ACS setting with significant peaks of cardiovascular complications. In this study, the ACR was not used as an early marker of AKI. The development of CIN was not specifically analyzed either as a post-coronary angiography compli­cation.18^-22

Microalbuminuria calculated through the ACR obtained from a simple urine sample is also an established marker of endothelial dysfunction that has been validated to predict cardiovascular events and mortality in different clinical settings. A previous analysis of our group revealed that higher ACR levels are associated with significantly worse outcomes in patients with non-ST-segment elevation ACS, and with a higher rate of hard endpoints like mortality and/or non-fatal acute myocardial infarction at the long-term follow-up (12% vs 2.2%, P =/< .0001).23 Also, other authors proved its utility to assess the risk of developing AKI, mainly in the ACS setting or while being exposed to cardiac surgery.24Tziakas et al confirmed the significant correlation between AMI related higher ACR levels and the development of AKI after this event (area under the ROC curve 0.72; 95%CI, 0.67-0.77). However, the authors did not report on the clinical impact of this complication on the patient’s clinical course or its association with the use of contrast during coronary angiography.25

Special attention should be paid to patients with post-angiographic AKI in the ACS setting. Several studies have shown that CIN negatively impacts the prognosis of hospitalized and long-term patients. In our population, mortality in patients with CIN was significantly higher compared to those without this disease (33% vs 1.8%).

The use of urinary ACR has been less studied in this context. Meng et al. reported that high microalbuminuria levels (ACR in between 30 mg/g and 300 mg/g) were associated significantly with the development of post-contrast acute kidney injury in patients undergoing coronary catheterization (12.1% vs. 5.0%; P = .005). A key point here that distinguishes this study from ours is that they included patients with scheduled coronary angiographies only and out of the ACS setting.26Another relevant point is that the ACR cutoff value to develop CIN was determined from the analysis of the area under the ROC curve, and its value of 20 mg/g was even lower compared to the conventional standard threshold of 30 mg/g, a finding that was consistent with what other clinical studies reported.27

The rate of CIN and its impact on the clinical outcome of coronary patients triggered the development of predictive scores for this disease. One of the best known indices is the Mehran score that includes variables like age > 75 years, hypertension, functional class III/IV heart failure, diabetes mellitus, anemia, use of intra-aortic balloon pump, volume of contrast administered, and past medical history of renal dysfunction and is capable of identifying who the most vulnerable patients are to develop post-coronary angiography CIN (the area under the ROC curve was 0.75). Adding the ACR to this score showed an even greater discriminatory power to predict post-coronary angiography CIN in patients with ACS. This would prove the practical utility of adding this index as a variable to the Mehran score.

CIN, one of the most common causes for acute nephrotoxicity, is a multi-factor event. Among its causes we should mention the direct nephrotoxic effect of the contrast substances used during endovascular procedures on the renal endothelium and the development of acute tubular necrosis. It is estimated that the nephrotoxicity of hyperosmolar contrast enhanced by the hemodynamic alterations produced by the ongoing ACS could alter vascular resistance with changes in the regulation of the release and balance of vasoactive substances like adenosine, endothelin, and nitric oxide. The damage perpetuates the slowing down of renal perfusion, spinal hypoxemia, ischemic injury, and ultimately cell death. In addition to reducing the clearance of oxidative stress products, the lower glomerular filtration rate levels increase the concentration of inflammatory mediators triggering structural alterations at renal tubular epithelium level like edema, vacuolization, and death.28^,29

We believe that these findings could help identify patients at high-risk of developing post-coronary angiography CIN in the ACS setting to promote preventive measures, behaviors, and strategies to avoid this complication.

Limitations

First, one of the main limitations of our work is its single center nature. However, we should mention that the population included was representative and covered the entire spectrum of patients with ACS admitted to our coronary care unit, which secures the internal validity and representativeness of our study. Secondly, the underpowered sample may have conditioned the appearance of false negative results due to its alpha error or lower power and stopped us from performing a proper multivariable analysis. Finally, certain data such as the volume of contrast used in each study was calculated retrospectively with the usual biases of this type of analysis.

CONCLUSIONS

The albumin-to-creatinine ratio, a recognized predictor of renal and endothelial dysfunction, was also a marker of CIN in patients with ACS with an added value when it was included in a widely validated clinical score. These results may be the beginning of a hypothesis-generating study to be confirmed prospectively at a multi-center level.

CONFLICTS OF INTEREST

None declared.

ACKNOWLEDGEMENTS

We wish to thank the entire staff of the Hospital Alemán Coronary Care Unit, particularly the nursing staff who helped collect the urine samples that were crucial to conduct this study.

REFERENCES