Temporal Trends, Practice Variation, and Associated Outcomes With IVUS Use During Peripheral Arterial Intervention

Abstract

BACKGROUND

Intravascular ultrasound (IVUS) has been shown in limited prospective studies to improve procedural outcomes for patients undergoing lower extremity peripheral arterial intervention (PVI).

OBJECTIVES

The authors aimed to study temporal trends, practice variation, and associated outcomes with the use of IVUS during PVI among Medicare beneficiaries.

METHODS

All PVIs performed from 2016 to 2019 among Medicare beneficiaries aged >65 years were included. Temporal trends in IVUS use were stratified by procedural location (inpatient, outpatient, or ambulatory surgery center [ASC]/office-based laboratory [OBL]) and physician specialty. The primary outcome was major adverse limb events (MALE). Inverse probability weighting was used to account for differences in baseline characteristics. Cox regression with competing risks was used to estimate weighted hazard ratios.

RESULTS

During the study period, 543,488 PVIs were included, of which 63,372 (11.7%) used IVUS. A substantial growth in IVUS use was observed, which was driven by procedures performed in ASCs/OBLs (23.6% increase from quarter 1 of 2016 through quarter 4 of 2019). Among operators who used IVUS, there was also notable variation in use (median operator use 5.4% of cases; IQR: 2.2%–15.0%; range, <1%–100%). In weighted analysis, IVUS use during PVI was associated with a lower risk of MALE through a median of 514 days (adjusted hazard ratio: 0.73; 95% CI: 0.70–0.75; P < 0.0001).

CONCLUSIONS

In contemporary nationwide data, IVUS use during PVI has increased since 2016, driven by growth in the ASC/OBL setting. However, there remains substantial variation in operator practice. When used during PVI, IVUS was associated with a lower risk of short- and long-term MALE.

Intravascular ultrasound (IVUS) uses miniaturized ultrasound transducers to provide vessel information beyond what can be obtained by standard angiography.¹ This includes but is not limited to reference vessel size, luminal diameter, luminal cross-sectional area, and characterization of atherosclerotic plaque. Although the original clinical applications of IVUS were in lower extremity peripheral arterial intervention (PVI),² IVUS use during coronary interventions grew rapidly, and consensus documents and best practices endorsing its use have been published.^3–5 The coronary interventional growth has also been supported by randomized trials and observational studies demonstrating improved outcomes compared with traditional angiography alone. For example, several meta-analyses of randomized and observational studies have demonstrated that IVUS-guided coronary stenting is associated with a significantly lower risk of cardiovascular death, myocardial infarction (MI), target lesion revascularization, and stent thrombosis.^6,7

Although the evolution in evidence supporting IVUS use during PVI is more limited,⁸ observational studies have demonstrated improved procedural outcomes with the inclusion of IVUS as part of the revascularization strategy. For instance, multiple cohort studies have demonstrated that IVUS use can improve intraluminal sizing, better evaluate the degree of vessel calcification, interrogate the severity of postintervention dissection, and be used to predict outcomes after stent implantation.^9–12 In one study, IVUS use during below-the-knee angioplasty in critical limb ischemia (CLI) was associated with better wound healing rates.¹³ In addition, in a meta-analysis that included data from 8 observational studies, IVUS-assisted PVI had similar primary patency and reintervention rates compared with angiography-based PVI alone but was associated with significantly lower rates of periprocedural adverse events and vascular complications.¹⁴

Despite the established and emerging data, only little is known about the growth and outcomes associated with IVUS use during PVI in contemporary real-world practice.¹⁵ As such, we aimed to study temporal trends, practice variation, and associated outcomes with the use of IVUS as part of a lower extremity revascularization strategy among Medicare beneficiaries. This study is unique because the Medicare data set allows for the evaluation of use patterns across clinical settings (inpatient, hospital outpatient, and ambulatory surgery center [ASC]/office-based laboratory [OBL]) and physician specialties, as well as the ability to track long-term outcomes.

METHODS

DATA SOURCE, STUDY COHORT, AND EXPOSURE.

The study was approved by the Institutional Review Board of Beth Israel Deaconess Medical Center and was conducted in compliance with the data use agreement in place between the Centers for Medicare and Medicaid Services (CMS) and Beth Israel Deaconess Medical Center. The study cohort included all lower extremity PVIs involving the iliac, femoropopliteal, and tibial arteries performed between January 1, 2016, and December 31, 2019, among fee-for-service Medicare beneficiaries 66 years and older with at least 1 year of enrollment before their index PVI. Inpatient PVIs were identified in the Medicare Provider Analysis and Review inpatient files using International Classification of Diseases-10th Revision (ICD-10) Procedure Coding System claims codes (Supplemental Table 1). Outpatient PVIs were identified by Current Procedural Terminology codes in Medicare carrier fee-for-service files and institutional outpatient files (Supplemental Table 1). Claims codes only identify revascularization procedures and may not include procedures involving unsuccessful revascularization attempts.

Sociodemographics were identified at the time of the index PVI using the master beneficiary summary file. Race/ethnicity was classified based on self-report using categories specified by Medicare at the time of Medicare enrollment. Comorbidities were ascertained using the CMS Chronic Conditions Data Warehouse common chronic conditions and other conditions files. In addition, ICD-10-Clinical Modification claims codes were used to identify CLI, intermittent claudication (IC), and prior amputation during a 1-year look back period (Supplemental Table 2).

The primary exposure of IVUS use during PVI was identified using Current Procedural Terminology and the ICD-10 Procedure Coding System (Supplemental Table 1). Carrier files were used to identify procedure location (inpatient, outpatient, or ASC/OBL) and physician specialty (cardiologist, surgeon, interventional radiologist, or other).

OUTCOMES.

The primary outcome for this analysis was major adverse limb events (MALE), which was defined as a composite of acute limb ischemia (arterial thrombosis/embolism) or major lower extremity amputation (defined as amputation proximal to the level of the forefoot) (Supplemental Table 3). The secondary outcomes included minor amputation (defined as amputation involving the forefoot or phalanges) and any amputation. Only the first procedure during the study period was included in the outcomes analysis. We also used falsification endpoint testing to assess for the presence of unmeasured confounding.¹⁶ The prespecified falsification endpoints were: 1) acute MI requiring hospitalization; 2) congestive heart failure (CHF) requiring hospitalization; and 3) pneumonia requiring hospitalization (Supplemental Table 4).¹⁷

STATISTICAL ANALYSIS.

Categoric variables were reported as counts and percentages and continuous variables as means and SDs. Given the large sample size, standardized mean differences were reported, with a threshold of at least 10% to define statistical significance.¹⁷

Quarterly time trends of the proportion of PVIs that used IVUS were evaluated and plotted from quarter 1 of 2016 to quarter 4 of 2019. This was stratified by procedure location and physician specialty. To evaluate for heterogeneity in practice between operators, each operator’s proportional use of IVUS during PVI for the total study period was assessed and depicted in a histogram.

For the outcomes analysis, the first procedure during the study period for each patient was included. Inverse probability weighting was used as the primary analytical tool for the endpoints to correct for potential confounding caused by imbalances in observed characteristics in baseline characteristics for patients treated with IVUS versus angiography alone.¹⁷ To calculate propensity scores used for weighting, logistic regression models were used to estimate the propensity for being treated with IVUS during PVI. All patient, procedural, and hospital characteristics were used in the model, as listed in Table 1. Each patient was then weighted based on their propensity score. Cumulative incidences at each time point were reported, with cumulative incidence functions estimated by Fine-Gray methods to account for the competing risk of death. Cox regression models were then used to estimate the weighted hazard ratios (HRs) comparing IVUS use during PVI versus no IVUS use. The subdistribution hazard models were constructed based on Fine-Gray methods.¹⁸

TABLE 1.

Preweighting and Postweighting Patient Characteristics (Overall Cohort)

	Preweighting			Postweighting
	Non-IVUS Patients (n = 479,756)	IVUS Patients (n = 63,732)	Standardized Mean Difference	Non-IVUS Patients	IVUS Patients	Standardized Mean Difference
Demographics
Age, y	76.3 ± 7.2	76.4 ± 7.2	−0.8	76.4 ± 7.2	76.3 ± 7.2	0.9
Male	56.7	56.2	1.0	56.7	56.9	−0.4
Race
White	79.6	75.4	10.2	79.2	79.9	−1.8
Black	13.1	14.0	−2.6	13.2	12.9	0.9
Other	7.3	10.6	−11.8	7.6	7.2	1.5
Dual enrollment	22.0	25.2	−7.7	22.3	21.4	2.2
Region
East North Central	14.8	12.4	7.2	14.5	14.2	0.8
East South Central	7.9	3.9	16.9	7.4	8.0	−1.9
West North Central	6.7	4.4	10.0	6.5	6.6	−0.6
West South Central	14.4	15.2	−2.2	14.5	14.5	0.0
Middle Atlantic	11.5	11.5	−0.1	11.5	10.6	2.9
Mountain	4.9	11.5	−24.1	5.7	5.6	0.5
New England	3.8	2.8	5.7	3.7	3.8	−0.4
Pacific	11.1	16.1	−14.8	11.7	11.4	0.8
South Atlantic	24.9	22.2	6.4	24.6	25.4	−2.0
Comorbidities
Acute myocardial infarction	15.3	13.5	5.1	15.1	15.3	−0.4
Alzheimer’s disease	5.8	6.3	−2.0	5.9	5.5	1.3
Alzheimer’s disease and related disorders of senile dementia	20.9	21.8	−2.2	21.0	20.4	1.4
Atrial fibrillation	28.2	25.5	6.0	27.9	28.1	−0.4
Cataract	69.8	70.1	−0.6	69.9	69.7	0.5
Chronic kidney disease	66.5	67.8	−2.8	66.6	66.3	0.7
Chronic obstructive pulmonary disease	50.5	49.1	2.8	50.4	50.5	−0.3
Heart failure	54.2	52.5	3.4	54.0	53.7	0.6
Diabetes	66.6	69.0	−5.2	66.8	65.7	2.3
Glaucoma	26.1	28.1	−4.4	26.3	25.8	1.2
Hip/pelvic fracture	5.8	5.4	1.5	5.7	5.6	0.6
Ischemic heart disease	83.2	82.3	2.5	83.1	83.2	−0.1
Depression	40.7	41.4	−1.4	40.8	40.5	0.7
Osteoporosis	21.0	22.2	−2.9	21.1	20.7	1.1
Rheumatoid arthritis/osteoarthritis	68.4	70.6	−4.8	68.7	68.3	0.9
Stroke/transient ischemic attack	28.1	27.6	1.1	28.1	27.9	0.4
Breast cancer	4.0	4.1	−0.9	4.0	4.0	0.1
Colorectal cancer	3.7	3.7	−0.0	3.7	3.6	0.5
Prostate cancer	7.7	7.8	−0.5	7.7	7.7	0.2
Lung cancer	3.0	2.7	2.0	3.0	3.0	−0.1
Endometrial cancer	0.9	0.8	0.3	0.9	0.9	0.1
Anemia	75.0	76.0	−2.2	75.1	74.7	0.9
Asthma	17.2	18.5	−3.5	17.3	17.1	0.6
Hyperlipidemia	93.8	94.5	−3.0	93.9	93.7	0.8
Benign prostatic hyperplasia	29.0	30.7	−3.9	29.2	29.0	0.5
Hypertension	97.1	97.3	−1.0	97.1	97.1	0.2
Acquired hypothyroidism	30.6	32.6	−4.3	30.8	30.5	0.7
Tobacco use	36.6	35.0	3.3	36.4	36.9	−0.9
History of amputation	7.1	5.0	8.9	6.9	6.7	0.7
Anatomy of disease
Segment (any)
Iliac	23.4	21.4	4.7	23.2	24.7	−3.6
Femoropopliteal	69.1	74.7	−12.4	69.8	69.5	0.5
Tibial	37.7	49.9	−24.7	39.1	36.9	4.6
Segment (exclusive)
Iliac	15.1	11.3	11.1	14.5	16.9	−6.5
Femoropopliteal	40.7	32.3	17.5	39.7	40.3	−1.2
Tibial	15.4	13.7	4.8	15.3	13.3	5.6
Iliac and femoropopliteal	6.5	6.5	−0.0	6.7	6.0	2.9
Iliac and tibial	0.4	0.3	1.4	0.4	0.2	2.9
Femoropopliteal and tibial	20.5	32.6	−27.6	21.8	21.7	0.3
Iliac, femoropopliteal, and tibial	1.4	3.3	−12.3	1.6	1.7	−0.1
PAD severity
Critical limb ischemia	36.2	39.0	−5.9	36.5	35.9	1.3
Intermittent claudication	38.5	42.4	−8.0	38.9	39.0	−0.1
Unspecified PAD	25.4	18.6	16.4	24.6	25.2	−1.3

Values are mean ± SD or %.

CLI = critical limb ischemia; IVUS = intravascular ultrasound; PAD = peripheral artery disease.

Prespecified subgroups included the severity of peripheral artery disease (PAD) (CLI, IC, or unspecified PAD, defined as lower extremity arterial revascularization without an assigned code for CLI or IC), arterial segment (iliac, femoropopliteal, and tibial), and inpatient versus outpatient procedural location.

SENSITIVITY ANALYSES TO EVALUATE FOR UNMEASURED CONFOUNDING.

To evaluate for the impact of unmeasured confounding, 2 additional analyses were performed. First, the association between IVUS use and the falsification endpoints of hospitalization for MI, hospitalization for CHF, and hospitalization for pneumonia were evaluated using similar methods as conducted in the primary analysis.

Second, an instrumental variable analysis was performed using operator preference for IVUS as the instrument and the 2-stage least squares method.^19–21 Because this method does not account for loss to follow-up, only patients with at least 1 year of follow-up were included. Furthermore, death was included as a composite outcome with MALE to account for the competing risk of death because this cannot be accounted for using the 2-stage least squares method. For both stages of the model, patient, procedural, and operator characteristics were included for adjustment. Risk differences with CIs and P values were reported.

All statistical analyses were performed using SAS software, version 9.3 (SAS Institute Inc). A 2-sided P value <0.05 was considered significant.

RESULTS

TEMPORAL TRENDS AND OPERATOR PREFERENCE FOR IVUS.

There were 543,488 patients who underwent PVI identified during the study period, 63,732 (11.7%) of which used IVUS. IVUS use rates by procedure location were as follows: 4.0% of inpatient PVIs, 5.2% of outpatient PVIs, and 26.8% of ASC/OBL PVIs. IVUS use rates by physician specialty included 10.3% of PVIs performed by cardiologists, 11.5% of PVIs performed by surgeons, 17.4% of PVIs performed by interventional radiologists, and 10.4% of PVIs performed by other physician specialties. Rates of IVUS use during PVI increased over time in all procedure locations (Figure 1) and among all physician specialties (Figure 2). The largest procedural location of growth occurred in ASC/OBLs, with a 23.6% increase from quarter 1 of 2016 through quarter 4 of 2019. Additionally, among physician specialties, interventional radiologists demonstrated a 19.7% growth in IVUS use during the study period.

FIGURE 1. Temporal Trends in IVUS Use in Lower Extremity Peripheral Arterial Intervention by Procedure Location.

Rates of intravascular ultrasound (IVUS) use from quarter 1 (Q1) 2016 through quarter 4 (Q4) 2019 during peripheral arterial intervention stratified by procedure location. Q2 = quarter 2; Q3 = quarter 3.

FIGURE 2. Temporal Trends in IVUS Use in Lower Extremity Peripheral Arterial Intervention by Physician Specialty.

Rates of IVUS use from Q1 2016 through Q4 2019 during peripheral arterial intervention stratified by physician specialty. Abbreviations as in Figure 1.

Individual operator’s proportional use of IVUS during PVI was heterogeneous. Among all operators (N = 12,844), the mean use of IVUS during PVI was 6.5%, with a median use of 0% and IQR of 0% to 5% (Figure 3). When restricted to operators who used IVUS at least once during the study period (n = 3,822), the mean use was 13.4%, with a median of 5.4% and an IQR of 2.2% to 15%. Individual operator’s proportional use of IVUS during PVI by location of procedure and by physician specialty are displayed in Supplemental Figures 1 and 2, respectively.

FIGURE 3. Operator Variation in IVUS Use in Lower Extremity Peripheral Arterial Intervention.

The proportion of peripheral arterial interventions among individual operators in which intravascular ultrasound (IVUS) was used during the procedure. The overall preference for IVUS use varied widely across individuals.

PATIENT AND PROCEDURAL CHARACTERISTICS.

When comparing characteristics between procedures performed with and without IVUS, the notable differences included greater use of IVUS among patients residing in the Mountain and Pacific regions. In addition, patients undergoing multilevel PVI (ie, treatment of more than 1 arterial segment) were more often treated with IVUS. Otherwise, there was a substantial balance between patients treated with and without IVUS before weighting, and there were no residual differences after weighting. Patient and procedural characteristics of the non-IVUS and IVUS groups by severity of PAD are shown in Supplemental Tables 5 to 7 and by anatomic location of intervention in Supplemental Tables 8 to 10.

OUTCOMES.

Before weighting, IVUS use with PVI was associated with a 28% reduction in the HR for MALE (95% CI: 0.70–0.75; P < 0.0001) through a median follow-up of 514 days (quarter 1: 212 days, quarter 3: 904 days) (Table 2). This association persisted in the weighted analysis (adjusted HR: 0.73; 95% CI: 0.70–0.75; P < 0.0001) (Central Illustration). There were similar reductions in the components of the composite outcome (acute limb ischemia: adjusted HR: 0.82; 95% CI: 0.78–0.87; P < 0.0001; major amputation: adjusted HR: 0.69; 95% CI: 0.66–0.71; P < 0.0001). Similar findings were observed when stratified by severity of PAD (Supplemental Table 11, Supplemental Figures 3 to 5) and by anatomic location of PVI (Supplemental Table 12, Supplemental Figures 6 to 8). Of note, the reduction in MALE was largest in magnitude among patients with CLI (adjusted HR: 0.76; 95% CI: 0.73–0.79; P < 0.0001) and those with unspecified PAD (adjusted HR: 0.65; 95% CI: 0.61–0.70; P < 0.0001), whereas the overall rates of MALE and the benefit of IVUS were more attenuated among patients with IC. When stratified by inpatient versus outpatient location of intervention, MALE remained lower when IVUS was used during PVI irrespective of place of service (inpatient adjusted HR: 0.92; 95% CI: 0.86–0.99; P = 0.026; outpatient adjusted HR: 0.83; 95% CI: 0.80–0.86; P < 0.0001).

TABLE 2.

Outcomes (Overall Cohort)

	Preweighting		Postweighting
	HR (95% CI)	P Value	HR (95% CI)	P Value
MALE	0.72 (0.70–0.75)	<0.0001	0.73 (0.70–0.75)	<0.0001
Acute Limb ischemia	0.79 (0.75–0.84)	<0.0001	0.82 (0.78–0.87)	<0.0001
Major amputation	0.71 (0.68–0.73)	<0.0001	0.69 (0.66–0.71)	<0.0001
Minor amputation	0.81 (0.78–0.84)	<0.0001	0.80 (0.77–0.83)	<0.0001
Any amputation	0.75 (0.73–0.77)	<0.0001	0.74 (0.72–0.76)	<0.0001
Falsification outcomes
Hospitalization for myocardial infarction	1.02 (0.97–1.09)	0.41	1.03 (0.97–1.09)	0.28
Hospitalization for congestive heart failure	0.91 (0.88–0.94)	<0.0001	0.97 (0.94–1.00)	0.04
Hospitalization for pneumonia	1.06 (1.00–1.12)	0.05	1.06 (1.01–1.12)	0.02

MALE = major adverse limb event(s).

CENTRAL ILLUSTRATION. Cumulative Incidence of the Composite Endpoint of Major Adverse Limb Events and its Components After Lower Extremity Peripheral Arterial Intervention Stratified by Use of IVUS.

ALI = acute limb ischemia; IVUS = intravascular ultrasound; NNT = number needed to treat; MALE = major adverse limb effect.

For the secondary outcomes, the use of IVUS during PVI was associated with lower risks of minor amputation and any amputation, both pre- and postweighting (Table 2). Again, these findings were consistent when stratified by severity of PAD (Supplemental Table 11) and by anatomic location of PVI (Supplemental Table 12).

SENSITIVITY ANALYSES.

Among the falsification endpoints, there was no association between IVUS use and hospitalization for MI, and only a marginal association existed with hospitalization for CHF or pneumonia (Supplemental Figure 9). Similar findings were observed when stratified by severity of PAD (Supplemental Table 11, Supplemental Figures 10 to 12) and by anatomic location of PVI (Supplemental Table 12, Supplemental Figures 13 to 15).

In the instrumental variable analysis, a total of 530,185 patients who underwent PVI and had at least 1 year of follow-up were included. Overall, characteristics were well balanced across increasing quintiles of operator IVUS use, although there were some notable differences in race, region, and arterial segment of PVI (Supplemental Table 13). The instrumental variable’s stage 1 Cragg-Donald Wald F statistic for the composite outcome of MALE or death was 5928.08 (P < 0.001), which is consistent with a strong instrumental variable (Supplemental Table 14). For the composite outcome of MALE or death, IVUS use during PVI was associated with an absolute risk reduction of 7.0% at 1 year (95% CI: 6.3%–7.7%) as well as a 6.7% absolute risk reduction in major amputation or death (95% CI: 6.1%–7.4%) (Supplemental Table 15).

DISCUSSION

In this contemporary analysis of Medicare beneficiaries undergoing PVI from 2016 through 2019, the use of IVUS was nearly 10-fold higher compared with a study from 2006 to 2011²² and grew across procedure location and physician specialty during the study period. The greatest growth that occurred during the study period was in ASC/OBLs and by interventional radiologists. Nonetheless, the use of IVUS during PVI was heterogeneous, with both frequent and infrequent users. Among outcomes, when IVUS was incorporated into a PVI strategy, it was associated with a reduction in MALE. This finding was consistent across arterial segments and severity of PAD and was overall robust to assessments of unmeasured confounding.

The current growth of lower extremity endovascular revascularization has overall been marked and is paralleled by a decline in surgical revascularization.²³ This is likely a result of technological advances in peripheral devices, improvements in procedural techniques, increased global investment in the endovascular field, and greater patient and case complexity. Nonetheless, during this time, amputation rates have remained stagnant,²⁴ with ongoing calls to action to identify strategies to improve limb outcomes.²⁵ One unique aspect of vascular intervention is the diverse specialties and procedural locations involved. This heterogeneity is not observed in fields like coronary revascularization and creates a challenge for harmonizing best practices and creating uniform procedural standards, which may in part be causative of the slow improvement in patient outcomes.

The benefit of IVUS use during coronary revascularization is now well established; however, its use during PVI is still in its infancy. In part, this may be a result of the diversity in experience in the vascular space. This analysis magnifies how the adoption of technology can vary dramatically across specialty and procedural location. For example, the increase in IVUS use was marked among interventional radiologists, although it was more moderate for cardiologists and surgeons. Furthermore, the growth in use in the ASC/OBL setting has outpaced that of other procedural locations. Although this may be driven in part by favorable reimbursement for IVUS use in this clinical setting,¹⁵ it may also reflect changes in operator practices and greater complexity of cases that are moving from hospital systems into private clinics, in particular involving patients with CLI. Irrespective of the underpinnings, the degree of variation in practice patterns among vascular providers requires greater effort in standardizing best practices through multi-specialty guidelines and consensus documents.⁵

Another contributor to the slow and varied growth of IVUS during PVI is the limited evidence base supporting its use. Unlike for PVI, IVUS has become an established, guideline-recommended imaging tool during coronary intervention.²⁶ In multiple randomized trials and observational studies, IVUS has consistently demonstrated reductions in meaningful endpoints, including MI, cardiovascular death, and stent thrombosis. Much of what is used from coronary IVUS images to improve procedural decision making can also be applied to the peripheral interventional space. For instance, IVUS can be used to evaluate the degree of arterial calcification to indicate the need for plaque modification procedures, such as atherectomy. Studies have shown the importance of procedure-related variables (such as balloon inflation time and maximal inflation pressure)²⁷ during peripheral arterial intervention to improve long-term outcomes. IVUS has been shown to help identify the presence of postintervention dissections, improve vessel size assessment, optimize balloon and stent selection and deployment, and evaluate for residual disease. In particular, the use of IVUS for device sizing, especially for drug-coated devices, can have a meaningful impact on reducing restenosis. These findings appear to be translating into meaningful limb outcomes in limited analyses.²⁸ Most recently, a prospective single-center trial of 150 patients undergoing femoropopliteal endovascular intervention that randomized patients to guidance by IVUS and angiography or angiography alone found that restenosis at 12 months was significantly lower in the IVUS-guided revascularization group.²⁹ However, IVUS should be used carefully when sizing balloons/stents to avoid oversizing, in particular when revascularization involves the subintimal space.³⁰

In the current study involving a large cohort of contemporary real-world patients undergoing PVI, IVUS again demonstrated an association with improved limb outcomes when included as part of a PVI procedure. Although amputation is rare among patients with claudication and, as expected, the reduction in limb events was largely driven by patients with CLI and unspecified PAD, there remained a small but meaningful impact on long-term limb outcomes for patients with claudication. Although these observational data may be subject to unmeasured confounding, it is important to note that a similar benefit was observed in a sensitivity analysis using a uniquely different statistical technique with different assumptions. This study overall lends additional support to the emerging evidence that IVUS use when incorporated into a peripheral revascularization strategy may improve longitudinal limb outcomes. As similar supportive data continue to emerge, the next extensions for peripheral IVUS use will require a focus on strategies to improve the uniform adoption across specialties and procedural locations. This will be strengthened by investment in randomized clinical trials and large prospective studies with hard endpoints. In addition, evaluations of the cost-effectiveness of IVUS use during PVI, improvements in workflow strategies, and better education of imaging interpretation will be required.

STUDY LIMITATIONS

First, IVUS use was identified using claims codes, and how it was incorporated into the procedure cannot be discerned. Although the use of claims codes to define the primary exposure can result in misclassification, appropriate and accurate billing is required for reimbursement and likely reduces this risk. Second, CMS data do not include detailed procedural information, including lesion characteristics such as length and severity of disease, degree of calcification, and the presence of total occlusion. Third, the cohort only included patients who experienced revascularization and therefore did not include patients who underwent an unsuccessful endovascular revascularization attempt. Fourth, the instrumental variable analysis also has limitations, particularly if physicians who used IVUS more or less often differ in other unmeasured ways. Fifth, other endpoints like target lesion revascularization could not be evaluated because of the ambiguity of the claims codes. This is a key limitation, and more work is needed to study the impact of IVUS on restenosis and target lesion revascularization rates. One recent randomized trial found that IVUS use during femoropopliteal revascularization resulted in a ~17% reduction in binary restenosis at 1 year,²⁹ supporting the findings of the current analysis. Lastly, despite the evaluation for residual and unmeasured confounding, there remains the possibility that these factored into the findings because of the observational study design.

CONCLUSIONS

In contemporary nationwide data, IVUS use during PVI has increased since 2016. However, there remains substantial variation in operator use. When used as part of a PVI strategy, IVUS use was associated with improved outcomes, including lower rates of incident MALE. Additional prospective data, will be useful to help confirm the association between IVUS use during PVI and improved long-term outcomes as well as help address specific clinical scenarios in which IVUS can be used to make treatment decisions.

PERSPECTIVES.

WHAT IS KNOWN?

IVUS has been established to improve outcomes during percutaneous coronary intervention and has been shown in limited prospective and observational studies to reduce restenosis for patients undergoing lower extremity PVI.

WHAT IS NEW?

In this large, Medicare claims-based observational study, IVUS use during PVI increased from 2016 to 2019, with substantial variation in operator practice. The use of IVUS during PVI was associated with improved long-term clinical outcomes, including a lower risk of MALE.

WHAT IS NEXT?

Additional research, including prospective studies, is needed to understand the mechanisms for this association and help address specific clinical scenarios in which IVUS can be used to make treatment decisions.

ABBREVIATIONS AND ACRONYMS

ASC

ambulatory surgery center

CHF

congestive heart failure

CLI

critical limb ischemia

CMS

Centers for Medicare and Medicaid Services

IC

intermittent claudication

ICD-10

International Classification of Diseases- 10th Revision

IVUS

intravascular ultrasound

MALE

major adverse limb event(s)

MI

myocardial infarction

OBL

office-based laboratory

PAD

peripheral artery disease

PVI

peripheral arterial intervention