Trends and In-hospital Outcomes of Intravascular Imaging on Single-vessel Coronary Chronic Total Occlusion Treated with Percutaneous Coronary Intervention

Abstract

Intravascular imaging (IVI), including intravascular ultrasound (IVUS) and optical coherence tomography (OCT), improves outcomes of percutaneous coronary intervention (PCI) for chronic total occlusions (CTO). We sought to quantify temporal trends in the uptake of IVI for CTO-PCI in the United States. We identified adults who underwent single-vessel PCI for CTO between 2008 and 2020. We quantified yearly trends in the number of IVUS-guided and OCT-guided single vessel CTO-PCIs by Cochran-Armitage and linear regression tests. We also examined the rates of in-hospital mortality and other prespecified in-hospital outcomes in patients who underwent CTO-PCIs with and without IVI, using logistic regression. Our study included a total of 151,998 PCIs on single-vessel CTOs, with the absolute number of CTO-PCIs decreasing from 12,345 in 2008 to 8,525 in 2020 (P-trend <0.001). IVUS use has increased dramatically from 6% in 2008 to 18% in 2020 for single-vessel CTO-PCIs (P-trend <0.001). Rates of OCT use have increased as well, from 0% in 2008 to 7% in 2020 (P-trend <0.001). There was no difference in in-hospital mortality between patients who underwent CTO-PCI with and without IVI (P-logistic=0.60). In the largest national analysis of single vessel CTO-PCI trends to date, we found that the use of IVUS has risen substantially accompanied by a similar but lesser rise in the use of OCT. There were no differences in rates of in-hospital mortality between patients who underwent single-vessel CTO-CIs with and without IVI.

INTRODUCTION

Chronic total occlusions (CTO) are found in 18–31% of patients with coronary artery disease (CAD) without previous coronary artery bypass grafting (CABG)^1–3, and up to 89% of patients with previous CABG². Percutaneous coronary intervention (PCI) for CTO is technically challenging and considered an ongoing frontier of interventional cardiology. PCI is performed on 10–30% of CTOs^1–3, with the remaining patients with CTO managed medically or referred for CABG. The introduction of intravascular imaging (IVI) guidance using intravascular ultrasound (IVUS) or optical coherence tomography (OCT) over the past decade has provided a valuable tool for operators performing CTO-PCI. IVI has multiple uses in CTO-PCI, including performing image-guided proximal cap puncture in ambiguous anatomies, identifying intraplaque and extraplaque wire passage and troubleshooting issues with reverse-controlled antegrade and retrograde tracking (CART).⁴ The most important applications of IVI in CTO-PCI are still for assessing plaque burden and morphology and vessel sizing to facilitate optimal stent sizing and deployment.

To better understand the uptake of these technologies, we sought to quantify temporal trends in CTO-PCI and IVI. Temporal trends in CTO-PCIs have been examined in Australia⁵, Europe⁶ and Israel⁷. In the U.S. the most recent analysis of temporal trends in CTO-PCI was eight years ago, using National Inpatient Sample data from 2008–2014⁸. In light of the introduction and adoption of CTO techniques such as the retrograde approach, antegrade dissection/re-entry (ADR), and improvement in operator experience,^9,10 as well as the importance of IVI to guide these techniques, we sought to provide an updated analysis of temporal trends in single-vessel CTO-PCIs and IVI in the intervening decade.

METHODS

Data Source

We utilized the National Inpatient Sample (NIS), the largest inpatient healthcare database developed by the Agency for Healthcare Research and Quality (AHRQ) and the Healthcare Cost and Utilization Project (HCUP)¹¹, to examine the trend of intravascular imaging (IVI)-guided percutaneous coronary intervention (PCI) of single-vessel chronic total occlusion (CTO). When weights are applied, the NIS approximates more than 35 million admissions year each, representative of 97% of the United States (US) population. The NIS continually adds more participating hospitals and undergoes yearly updates, the most notable of which include sampling all HCUP-participating hospitals instead of hospitals that retained all discharges in 2012, and switching from International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) to ICD-10-CM in October 2015. All entries are anonymized, thereby protecting patient confidentiality and guaranteeing strictly de-identified patient information. Consequently, our institutional review board exempted our study from approval as we solely used data from the NIS, which is openly available in the public website of the HCUP¹¹.

Study Population and Variables

We planned the following steps to identify hospital admissions in which single-vessel CTO-PCI was performed, as validated by previous studies^8,12. Firstly, we used International Classification of Diseases, Ninth Revision, Procedural Coding System (ICD-9-PCS) and ICD-10-PCS codes to identify all hospital admissions for single-vessel PCI from 2008 to 2020. PCI was defined by the deployment of either a drug-eluting stent or bare-metal stent, or the implementation of balloon angioplasty. We excluded patients under the age of 18 and hospitalizations with missing information on demographics and in-hospital outcomes to ensure the completeness of our data. Secondly, single-vessel PCI hospitalizations with ST-elevation myocardial infarction (STEMI) and non-ST elevation myocardial infarction (NSTEMI) were excluded to isolate stable ischemic heart disease⁸. Finally, only single-vessel PCI hospitalizations with a diagnosis code of CTO were included in our final study cohort. We meticulously selected our cohort to ensure that hospitalizations included in our study reflected the rates and outcomes of single-vessel PCI of the chronically occluded coronary artery⁸.

Guidance by IVI was defined by the use of intravascular ultrasound (IVUS) or optical coherence tomography (OCT) during PCI. To further describe our cohort, we extracted data on demographics (age, sex, race), comorbidities (smoking, hypertension, diabetes mellitus, hyperlipidemia, obesity, heart failure, atrial fibrillation, valvular heart disease, peripheral artery disease, previous stroke, previous coronary artery bypass graft [CABG], previous pacemaker, chronic obstructive pulmonary disease, pulmonary hypertension, chronic kidney disease, end-stage renal disease, liver cirrhosis, history of malignancy, deficiency anemia, malnutrition, dementia, major depression), hospital characteristics (region, bed size, urban location), primary payer, and median income by ZIP code. To imbue granularity, we also collected procedural information on the number of stents, bifurcation lesions, intra-aortic balloon pump, extracorporeal membranous oxygenation, left ventricular assist device, renal replacement therapy, and mechanical ventilation. All the ICD-9 and ICD-10 codes used in this study to identify comorbidities and procedures can be found in Table S1.

Study Outcomes

Primary outcomes of interest included the annuals trends in the relative number and percentage of IVUS-guided as well as OCT-guided single-vessel CTO-PCIs from 2008 to 2020. Secondary outcomes included the trend of in-hospital mortality, acute kidney injury, urgent CABG, non-home discharge, length of stay, and total hospital cost over the same period. Identical secondary outcomes were analyzed when comparing single-vessel CTO-PCIs with and without IVI.

Statistical Analysis

We calculated nationally representative estimates using survey analysis methodology based on weights of hospital-level discharge¹³. Appropriate weights were applied to provide national estimates for trend analyses that are consistent throughout the entire period¹⁴. Categorical and continuous covariates were summarized as percentages and means with standard deviations (SD), respectively. Cost-to-Charge Ratio files, provided by HCUP, were used to calculate total hospital costs from total hospital charges. Total hospital costs were then adjusted for inflation to US dollars based on the medical care component of the US Consumer Price Index in the year 2021.

We employed the Cochran-Armitage trend test and simple linear regression test to examine the temporal trends in categorical and continuous covariates in baseline characteristics, respectively. The same strategy was used in assessing the trend of outcomes. The number of IVUS- and OCT-guided single-vessel CTO-PCIs was divided by the total number of single-vessel CTO-PCIs to derive yearly comparable numbers and percentages. After compiling 13 years of data, we additionally compared in-hospital outcomes of single-vessel CTO-PCIs with and without IVI guidance. Binary outcomes were compared using both simple and multivariable logistic regression to generate crude odds ratios and adjusted odds ratio (aOR) with 95% confidence intervals (CI). Continuous outcomes were compared using linear regression models to produce mean differences and their 95% CIs. Covariates used to adjust the model included age, sex, race, comorbidities, hospital characteristics, primary payer, median income, and procedural characteristics as they were deemed to potentially impact in-hospital outcomes. Subgroup analyses of in-hospital mortality were performed, stratified into subgroups according to sex, previous CABG, chronic kidney disease, hospital teaching status, number of stents, bifurcation lesions, and use of mechanical circulatory support. Interaction analyses were performed to evaluate the interaction between use of IVI and each of the aforementioned subgroups. Sensitivity analysis including years 2016 to 2020 was also performed to decrease the heterogeneity by using data that solely used ICD-10 codes. Finally, similar multivariable logistic regression was used to evaluate the use of IVI in different subgroups. P-values <0.05 were considered as significant, and all tests were 2-sided. All data curation and analyses were conducted using SAS, version 9.4 (SAS Institute, Cary, NC).

RESULTS

Trends in Patient Characteristics

Our study included a total of 151,998 PCIs on single-vessel CTOs in patients with CAD but without active ACS (Figure 1). The absolute number of CTO-PCIs decreased from 12,345 in 2008 to 8,525 in 2020, with a significant downward trend (P-trend <0.001) (Table S2). The percentage of male patients did not significantly change (P-trend=0.869). The proportion of white patients decreased from 81.9% to 73.3% (P-trend <0.001), while that of black, Hispanic, and Asian patients increased (all P-trend <0.001). The prevalence of comorbidities, including diabetes, hyperlipidemia, obesity, heart failure, atrial fibrillation, valvular heart disease, and chronic kidney disease, substantially increased over the same period (all P-trend <0.001).

Figure 1. Flow-chart of this study.

The flow diagram illustrates the process whereby single-vessel percutaneous interventions performed for chronic total occlusion were identified.

Trends in Intra-Vascular Imaging

The number of IVUS-guided single-vessel CTO-PCIs dramatically increased from 56 in 2008 to 178 per 1,000 single-vessel CTO-PCIs in 2020 (P-trend <0.001) (Figure 2). Similarly, the number of OCT-guided single-vessel CTO-PCIs substantially increased from 0 in 2008 to 7 per 1,000 single-vessel CTO-PCIs in 2020 (P-trend <0.001).

Figure 2. Trend of intravascular imaging-guided single-vessel CTO-PCIs.

The graphs show the trend of IVUS-guided (Figure 2A) and OCT-guided (Figure 2B) single-vessel CTO-PCIs from year 2008 to 2020. The blue graphs represent IVUS- or OCT-guided single-vessel CTO-PCIs per 1,000 single-vessel CTO-PCIs, while the red line represents the percentage of IVUS- or OCT-guided single-vessel CTO-PCIs. The dotted red line shows the line of best fit.

Abbreviations: CTO, chronic total occlusion; IVUS, intravascular ultrasound; OCT, optical coherence tomography; PCI, percutaneous coronary intervention

Trends in Lesion and Procedural Characteristics

The proportion of bifurcation lesions increased from 1.7% in 2008 to 3.8% in 2020, with a significant upward trend (P-trend <0.001) (Table S3). Approximately 61% of single vessel CTO-PCIs deployed a single stent (as opposed to two, three, or four or more stents), a proportion which remained stable over time (P-trend 0.089). The use of renal replacement therapy (RRT) during the admission increased from 1.6% in 2008 to 4.9% in 2020, with a significant upward trend (P-trend <0.001). The use of mechanical ventilation increased from 0.6% in 2008 to 3.1% in 2020, with a significant upward trend (P-trend <0.001).

Trends in Periprocedural Complications and Outcomes

Rates of periprocedural complications, including in-hospital mortality, acute kidney injury, urgent CABG, and non-home discharge, each showed increasing trends (all P-trend <0.001) (Table S3), including a rise in in-hospital mortality from 0.4% to 1.9% and acute kidney injury (AKI) from 2.5% to 17.4%. The length of stay increased from a mean of 2.3 days (SD: 3.0) to 4.0 days (SD: 4.4), with a significant upward trend (P-trend < 0.001). The inflation-adjusted total hospital cost also increased from a mean of $22,665 (SD: 14,957) to $29,312 (SD: 22,883), with a significant upward trend (P-trend <0.001).

Lesion and Procedural Characteristics of IVI-Guided and Non-IVI-Guided Procedures

After aggregating 13 years of data, 12,775 IVI-guided single-vessel CTO-PCIs were compared with 139,223 conventional single-vessel CTO-PCIs (Table S4). The IVI group involved a greater proportion of bifurcation lesions (4.7% vs 2.5%; P_T-test < 0.001) and was more likely to involve a higher number of stent placements (P_ChiSq < 0.001).

Complications and Outcomes of IVI-Guided and Non-IVI-Guided Procedures

No differences were seen between single-vessel CTO-PCIs with and without intravascular imaging in the adjusted odds of in-hospital mortality, acute kidney injury, urgent CABG, and non-home discharge (Table 1). Length of hospital stay also did not differ between the two groups. The use of IVI, however, was associated with higher total hospital cost (mean difference $2,712, 95% CI 2,072–3,353, p<0.001). Similar findings were seen in sensitivity analysis including years 2016 to 2020 (Table S5).

Table 1.

Outcomes after single-vessel CTO-PCI with and without intravascular imaging

Outcome	Intravascular Imaging (+)	Intravascular Imaging (−)	Crude Odds Ratio	P-value	Adjusted Odds Ratioa	P-value
In-hospital mortality (%)	1.2	0.9	1.46 (1.01–2.12)	0.046	1.13 (0.72–1.78)	0.602
Acute kidney injury (%)	10.4	8.4	1.27 (1.11–1.45)	0.001	1.04 (0.89–1.22)	0.627
Urgent CABG (%)	1.0	1.2	0.87 (0.58–1.29)	0.478	0.87 (0.57–1.33)	0.525
Non-home discharge (%)	14.9	13.1	1.06 (0.96–1.16)	0.275	0.99 (0.89–1.10)	0.840
Length of stay (days)	3.5	3.2	0.30 (0.15–0.45)b	<.001	0.09 ([−0.04]–0.22)b	0.172
Total hospital costc ($)	30,115	24,308	5,807 (5,064–6,551)b	<.001	2,712 (2,071–3,353)b	<.001

^aAdjusted for sex, age, comorbidities, hospital characteristics, primary payer, median income, and procedural characteristics

^bMean difference with 95% confidence interval

^cInflation-adjusted total hospital cost, rounded to the nearest United States dollar

Abbreviations: CABG, coronary artery bypass graft

Sensitivity and Subgroup Analyses

We performed additional analyses examining the use of IVI in different subgroups (Figure 3). No difference in the use of IVI was seen between the two sexes, but whites had lower odds of undergoing IVI compared with non-whites (OR 0.86, 95% CI 0.78–0.95, p=0.004). Chronic kidney disease had no difference in the use of IVI, but a history of previous CABG was associated with lower odds of undergoing IVI (OR 0.78, 95% CI 0.78–0.97, p=0.001). Medium (OR 1.27, 95% CI 1.09–1.48, p=0.004) and large hospital bed sizes (OR 1.20, 95% CI 1.04–1.38, p=0.019) were associated with higher odds of IVI use compared with small bed-sized hospitals. Third (OR 1.13, 95% CI 1.00–1.27, p=0.049) and fourth (OR 1.33, 95% CI 1.18–1.50, p<0.001) quartile median income was associated with higher odds of IVI use compared with first, the lowest, median income. Bifurcation lesions (OR 1.63, 95% CI 1.33–1.99, p<0.001) and increasing number of stents were also associated with higher odds of IVI use.

Figure 3. Comparison of in-hospital mortality between intravascular imaging and non-intravascular imaging single-vessel CTO-PCIs in different subgroups.

The figure shows the adjusted odds ratio of in-hospital mortality after intravascular imaging-guided versus conventional single-vessel CTO-PCIs in different subgroups. The vertical line represents the adjusted odds ratio while the horizontal line represents the 95% confidence interval. The sizes of the blue squares are inversely proportional to the width of the 95% confidence interval.

Abbreviation: CABG, coronary artery bypass graft; CI, confidence interval; CKD, chronic kidney disease; CTO, chronic total occlusion; OR, odds ratio; PCI, percutaneous coronary intervention

DISCUSSION

In this analysis of single vessel CTO-PCI, which represents the first update since 2014 in the largest inpatient healthcare database in the United States, we found several important trends between 2008 and 2020. The absolute number of CTO-PCIs performed in the U.S. annually is decreasing. The selected patients who undergo CTO-PCI have a higher burden of significant comorbidities including diabetes, obesity, heart failure, atrial fibrillation, valvular heart disease, and chronic kidney disease. In addition, the procedural complexity of CTO-PCIs is likely increasing as well, potentially reflected in the increased proportion of bifurcation lesions. This rise in the comorbid burden and procedural complexity may explain the increase seen in rates of complications such as AKI and periprocedural mortality during the study period. The rates of IVI use have increased dramatically, and more so for IVUS compared with OCT. However, there was no difference in in-hospital mortality between patients who underwent CTO-PCI with IVI compared to patients who underwent CTO-PCI without IVI.

We found that the absolute number CTO-PCIs performed in the United States per year is decreasing. This could also be a reflection of declining rates of PCI overall. Overall PCI rates declined 10% between 2010 and 2017, primarily driven by the decline in elective PCI rates¹⁵, and nearly all CTO-PCI cases are elective. Successive randomized trials in the past decade have showed that routine PCI of CTO lesions yields no improvement in the rates of all-cause mortality, MI, stroke, or repeat revascularization, as evidenced in meta-analysis, although none of these trials have been powered to demonstrate a reduction in clinical outcomes such as death or MI, and there is also no randomized evidence of survival benefit for any PCI in stable CAD.¹⁶ CTO-PCI is now selectively performed for improvement in refractory anginal symptoms.¹⁷ Although the DECISION-CTO trial showed no improvement in anginal symptoms, it was underpowered due to early termination of enrollment, impeded by mild or absent baseline patient symptoms, and diluted by the fact that nearly 1 in 5 patients crossed over from the no PCI arm to the PCI arm within 3 days of randomization.¹⁸ The trial design also made it impossible to discern the clinical impact of CTO-PCI, as half of the patients in each group underwent PCI for non-CTO lesions within the trial protocol. The randomized EURO CTO trial showed an improvement in angina frequency, quality of life, and physical limitation, with no increase in MACE.¹⁹ Another possible contributor to the decreasing trend of CTO-PCIs is that more multivessel interventions which concomitantly intervene on the non-CTO artery may have been excluded from our analysis confined to single-vessels. In addition, the significantly lower number of CTO-PCIs performed in the year 2020 due to COVID-19 impacted the trend.

Our finding that the rate of periprocedural mortality is rising is not entirely surprising when considering the increasing comorbid burden seen in patients treated with CTO-PCI over time. The complexity of CTO lesions attempted has risen with time, with a mean J-CTO score of 1.76 in 2008–2009 to 2.17 in 2014–2015, as tracked the Euro CTO registry.⁶ Operators were willing to attempt CTO lesions of increasing length and increasing elapsed occlusion duration in months, from a mean estimated occlusion duration of 23.2 months in 2008–2009 to 34.7 months in 2014–2015.⁶ While the J-CTO score, CTO length, and CTO duration are not reported in the NIS, the rising complexity of lesions attempted is suggested by the increasing proportion of bifurcation lesions in our population over 2008–2020. In addition, we found that CTO-PCIs are increasingly being attempted on patients who have a greater number of comorbidities, such as diabetes, obesity, heart failure, atrial fibrillation, valvular heart disease, and chronic kidney disease, as well as the periprocedural use RRT and mechanical ventilation more than tripling over 2008–2020. In an analysis of CTO-PCIs in the National Cardiovascular Data Registry (NCDR) CathPCI Registry, the rate of MACE was 1.6% over 2009–2013.²⁰ This is similar to the 0.7–1.1% rate of urgent CABG surgery and 0.5–0.7% rate of in-hospital mortality seen in the 2009–2013 segment of our study. With regard to in-hospital mortality trends with time, the rate of CTO-PCI-associated in-hospital mortality was relatively stable between 0.4% and 0.1% (p for time=0.120; p for operator=0.999) across the 2008–2015 period in the Euro CTO registry,⁶ but it is unknown whether CTO-PCI-associated in-hospital mortality has risen after 2015 as suggested by our data. The counterpart paper from the NCDR registry did not examine trends in CTO-PCI mortality by year, but showed that trends in CTO-PCI MACE by year was relatively stable between 1.9% and 1.3% (p=0.108) across the 2009–2013 period.²⁰ It is unknown whether in-hospital morality has risen after 2013 as suggested by our data, and future follow-up studies on the contemporary rates of mortality and morbidity are warranted.

The rate of non-fatal periprocedural complications in our study, especially that of AKI, may be overestimated as it includes complications throughout the entire hospital stay, not necessarily complications related to the procedure itself. By NCDR data, 7% of patients undergoing PCI experience AKI, which is itself independently associated with increased in-hospital mortality (adjusted OR of 7.8).²¹ CTO-PCI-associated AKI surpassed 7% in 2013, to 17.4% in 2020, which could be a reflection of increasing complexity of cases attempted as described above. Euro CTO registry data from 2008–2015 suggest no significant change in the contrast volume used across those years,⁶ but it is unknown if this persisted beyond 2015 or whether that held true for non-experienced CTO operators. Importantly, our results are concordant with a prior study of the NIS from 2008–2014, which showed an increasing rate of AKI during that time period.⁸ The increasing rate of AKI with time is also consistent with the evolving patient pool, with a concordant rise in comorbidities, as described above. We do not have information on volume of intraprocedural contrast to perform mediation analysis between CTO-PCI and rate of AKI.

Our finding that IVI use increased dramatically over the past decade is in line with the rising use of IVUS guidance in CTO-PCIs in the Euro CTO registry, from 2.1% in 2008–2009 to 12.8% in 2014–2015.⁶ CTO crossing algorithms by multiple international CTO clubs now recommend the use of IVUS, especially for interrogating proximal ambiguity.¹⁰ In the most current meta-analysis of IVUS-guided CTO-PCI versus angiography-guided CTO-PCI, consisting of 2 randomized and 2 observational studies, IVUS guidance was associated with lower risk of stent thrombosis, shorter procedure time, shorter fluoroscopy time, less contrast volume use, lower total stent length and fewer total number of stents.²² It remains to be determined whether the adoption of IVI improves outcomes in CTO-PCI in large-scale registries and databases. In the aforementioned meta-analysis, there was no difference in all-cause mortality, MACE, cardiovascular mortality, MI, target vessel revascularization, or target lesion revascularization between IVUS-guided CTO-PCI and angiography-guided CTO-PCI.²² While our results are descriptive in nature, they are consistent with the results of randomized and meta-analyzed data suggesting no difference in in-hospital outcomes between IVUS-guided CTO-PCI and angiography-guided CTO-PCI. The results were similar in our subgroup and sensitivity analyses. While lesion characteristics are unavailable to account for procedural risks, it is likely that IVI was utilized more in high-risk lesions and to manage intraprocedural complications or suboptimal results. Therefore, these unmeasured confounders may have differentially increased the rates of adverse outcomes in the IVI cohort. In addition, the adverse outcomes of CTO-PCI, such as in-stent restenosis, typically occur after discharge in the subsequent months. Our study was limited to in-hospital outcomes, but we expect less complications and readmissions to occur with IVI in the long run, which will offset the extra cost associated with IVI.

Although use of OCT in CTO-PCI rose between 2008 and 2020, absolute numbers remained low, being deployed in only 0.7% of CTO-PCIs. This is consistent with other estimates. In an analysis of imaging guidance used in CTO-PCIs at seven US medical centers, IVUS was used in 36% of cases, OCT was used in 3% of cases, and both IVUS and OCT were used in 1.5% of cases.⁴ There are several properties of OCT that likely contribute to its lower usage rate compared to IVUS in CTO-PCI, including its lower depth of penetration compared to IVUS, and its need for flushing the blood column in the arterial lumen, unlike IVUS. There are specific reasons that CTO operators may opt not to use OCT, in particular the risk of propagating dissections and extending hematomas before stenting has occurred due to the requirement for robust injections to obtain adequate imaging. In the aforementioned study, there was wide variability across medical centers in the percent of cases using IVI of either sort, with 0%, 2%, 3%, and 6% at four centers and 15%, 26%, and 48% at the remaining three centers.

Limitations

Our study should be considered in light of several limitations. First, because information about specific CTO-PCI techniques, such as the use of an antegrade or retrograde approach, or the use of subintimal tracking, dual injection, wire escalation, or parallel wires, is not included in ICD-9-PCS and ICD-10-PCS, we cannot estimate trends at that level of granularity. Information about lesion characteristics, such as proximal cap ambiguity, lesion length, tortuosity and calcification, presence of collateral circulation, quality of the distal vessel is not available, all of which are variables that could contribute to an explanation of trends in outcomes. Information about technical or procedural success rates are not available either. Second, because IVI guidance was not assigned to procedures in a randomized manner, and propensity score matching was not performed, we cannot attribute differences in IVI and non-IVI group outcomes to the use of IVI. Furthermore, the use of IVI is subject to confounding by indication, as lesions with more complex characteristics – which may not be captured by ICD data – are more likely to prompt operators to employ IVI guidance. Third, because the NIS is an extrapolated administrative database, the NIS is susceptible to coding errors. Fourth, our study is descriptive and retrospective, so the inferences should not be interpreted for causal relation, but rather be used as hypotheses-generating.

Nevertheless, the study contributes valuable information to the literature on real-world trends in CTO-PCI in the past decade. Data from the Euro CTO club operators, which has impressive procedural success rates and low complication rates, are restricted to the cases of fifty-three expert CTO operators in Europe.²³ While data from the NCDR registry is not restricted to PCIs performed by experienced CTO operators, participation in the NCDR is voluntary, and there is no easy mechanism to extrapolate data from participating hospitals to non-participating hospitals. The NIS, while lacking granular data about lesion characteristics, allows for the calculation of nationally representative estimates using multiplication of discharge weights with post-stratification by hospital characteristics and discharge characteristics to accurately represent the range and distribution of hospitals across the U.S.

Conclusion

In this study, which represents largest analysis of single vessel CTO-PCI trends in the U.S. to date, we found that the number of CTO-PCIs performed per year is decreasing while the rates of complications such as AKI and in-hospital mortality are rising, alongside an increase in the comorbid burden of patients undergoing CTO-PCI. We further found that use of IVUS and OCT have risen, although there are no differences in in-hospital mortality between patients who were given IVI and those who were not. Over the past decade, CTO-PCI has benefited from increasing standardization of terminology, techniques, devices, and algorithms, allowing operators to take on increasingly complex cases in patients with increasing comorbidities, as was seen in our analysis. While no mortality difference was observed between patients receiving and not receiving IVI guidance during CTO-PCI, further randomized study is needed to evaluate whether there are differences in other key procedural and clinical outcomes.

HIGLIGHTS.

• A total of 151,998 PCIs on single-vessel CTOs were analyzed from 2008 to 2020.

• Number of IVUS or OCT in single-vessel CTO-PCIs substantially increased.

• Only 9.2% of all single-vessel CTO-PCIs were guided by intravascular imaging.

• Comorbidities and complications also significantly increased over the years.

• In-hospital mortality did not differ with the use of intravascular imaging.

Acronyms and Abbreviations

aOR

Adjusted odds ratio

CABG

Coronary artery bypass graft

CI

Confidence interval

cOR

Crude odds ratio

CTO

Chronic total occlusion

HCUP

Healthcare Cost and Utilization Project

ICD-9-CM

International Classification of Diseases, 9th Revision, Clinical Modification

ICD-10-CM

International Classification of Diseases, 10th Revision, Clinical Modification

ICD-9-PCS

International Classification of Diseases, 9th Revision, Procedural Coding System

ICD-10-PCS

International Classification of Diseases, 10th Revision, Procedural Coding System

IVI

Intravascular imaging

IVUS

Intravascular ultrasound

NIS

National Inpatient Sample

OCT

Optical coherence tomography

PCI

Percutaneous coronary intervention

SD

Standard deviation

US

United States