Restenosis and symptom recurrence after endovascular therapy for claudication: Does duplex ultrasound correlate with recurrent claudication?

Abstract

After endovascular therapy, duplex ultrasound surveillance to detect restenosis guides clinical decisions and defines treatment failure. However, the correlation between duplex ultrasound and symptom recurrence remains unclear. We reviewed our institutional experience (2007–2010) to identify patients undergoing endovascular therapy for claudication. The association between post-intervention systolic velocity ratio and patient-reported symptom recurrence was determined. We analyzed 183 follow-up visits following treatment in 88 limbs (femoropopliteal (56%) or iliac (44%) arteries). After femoropopliteal intervention, median systolic velocity ratio was higher in patients with symptom recurrence (2.99 symptomatic vs. 1.69 asymptomatic; p < 0.001). Elevated systolic velocity ratio or occlusion correlated with symptom recurrence (area under receiver operator characteristic curve = 0.82 [95% CI 0.74–0.83]), and systolic velocity ratio >2.5 was 71% sensitive and 72% specific for symptom recurrence. After femoropopliteal endovascular therapy for claudication, duplex ultrasound-detected restenosis is highly associated with clinical deterioration. This validates objective criteria for treatment failure in claudicants and suggests that symptom status can serve as a primary indicator of anatomic restenosis.

Keywords

Endovascular therapy claudication duplex ultrasound surveillance restenosis outcomes

Introduction

Duplex ultrasound (DU) surveillance is commonly used to assess patency following lower extremity bypass (LEB) and successfully identifies failing conduits, facilitating maintenance of patency.^1–3 Following the widespread adoption of endovascular therapy (ET) in the treatment of lower extremity peripheral arterial disease (PAD), vascular specialists have applied DU to assess the success of percutaneous interventions as well.^4,5 In addition to this clinical application, DU-assessed patency is typically a primary endpoint used to assess effectiveness of therapy in clinical studies, including pre-market assessment of novel endovascular devices.^6–10

Despite the critical role of DU as an objective means of assessing patency, criteria for restenosis are derived from studies comparing ultrasound to angiography. However, the relationship between objective assessment of patency by DU and patient-centered outcomes has not been well-defined. After interventions for claudication, the patient's experience of pain-free walking is potentially the most important endpoint.

The objective of this study is to evaluate the correlation between DU-detected restenosis and symptom recurrence following ET for claudication.

Methods

Patients and follow-up

We performed a retrospective review of our institutional experience (2007–2010) to identify patients who underwent ET including angioplasty with or without adjunctive bare metal stenting for claudication. Treatment selection was at the discretion of the operating surgeon. Patient demographics, procedural details, and lesion characteristics were reviewed. Throughout follow-up, surveillance DU encounters were paired with documentation of patient-reported outcomes in the medical record. First post-operative follow-up visits were excluded in order to distinguish residual disease from restenosis.

For purposes of analysis, patients who reported improvement in symptoms or freedom from symptoms were considered “asymptomatic,” while patients who reported ongoing or worsening symptoms were considered “symptomatic.”

Ultrasound measurements

All duplex ultrasonography was performed at our institution's vascular laboratory by registered vascular technologists. A 7-mHz linear probe was used at 60° insonation angle. The peak systolic velocity (PSV) was obtained in an arterial segment proximal to the treated lesion and also within the treated area itself. The systolic velocity ratio (SVR) was calculated as the ratio between the maximal PSV within the previously treated area and the PSV just proximal to the area of maximal PSV. In a small number of visits, SVR could not be obtained because no flow could be detected in the artery. Initially, these were recorded as SVR = 0; however, for the purposes of receiver operator characteristic (ROC) curve analysis, we recoded these SVR as equal to the highest SVR obtained in the dataset.¹¹ This allowed us to account for occluded arteries without compromising the ROC model. When reporting clinical outcomes in patient subgroups, the small number of limbs with occluded arteries were excluded since they represent a clinically distinct problem with regard to outcomes such as reintervention.

Statistical analysis

We compared DU-derived SVRs between symptomatic and asymptomatic patients, testing for significance with Wilcoxon rank sum tests where appropriate. ROC curves were used to compare dichotomous symptom status with SVR in order to determine which SVR threshold correlated best with symptom recurrence. Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated for a range of thresholds. A p value of less than 0.05 was considered statistically significant. Data were compiled and analyzed using STATA version 11.2 (StataCorp, College Station, TX).

Results

Patient demographics and operative details

Overall, 183 valid pairs of DU and symptom status were identified from 71 patients following ET in 88 limbs. Fifty-eight percent of patients were male, with a mean age of 69.3 years (SD ± 10.5). There were high rates of former/current tobacco use (89%), hypertension (82%), dyslipidemia (65%), and coronary artery disease (52%) among included patients (Table 1). The time period over which these data pairs were obtained ranged from 1 month to >24 months post-op: Seven visits were in the first 3 months following intervention, 30 visits were in months 3–6, 16 visits in months 6–12, 40 visits at 12 months follow-up, 27 visits at 18 months follow-up, 33 visits at 24 months follow-up, and 30 visits at >24 months follow-up.

Table 1.

Patient demographics.

Variable	No. (% of patients) or mean (±SD)
Total patients	71
Total limbs	88
Follow-up visits	183

Age	69.3 (±10.5)
BMI	25.7 (±4.8)
Male	41 (58%)

Comorbidities
Coronary artery disease	37 (52%)
Congestive heart failure	3 (4%)
History of MI	11 (15%)
COPD	11 (15%)
Prior or current smoker	63 (89%)
Chronic renal insufficiency	5 (7%)
Hemodialysis-dependent	3 (4%)
Diabetes	21 (30%)
Hypertension	58 (82%)
Dyslipidemia	46 (65%)
Prior CABG	12 (17%)

BMI, body mass index; MI, myocardial infarction; COPD, chronic obstructive pulmonary disease; CABG, coronary artery bypass graft; SD, standard deviation.

The primary level of disease treated were the femoropopliteal arteries in 56% and the iliac arteries in 44% of limbs (Table 2). For a majority of patients (78%), this was the first ipsilateral intervention for claudication. TASC II classification of lesions treated¹² was comparably distributed between TASC A/B lesions (54%) versus TASC C/D lesions (46%). Eighty-two percent of patients underwent angioplasty with adjunctive bare metal stent placement.

Table 2.

Procedure details.

Variable	No. (% of limbs)
Primary intervention	69 (78%)
Site
Femoropopliteal	49 (56%)
Iliac	39 (44%)
TASC II Classification^a
A	23 (30%)
B	18 (24%)
C	22 (29%)
D	13 (17%)
Intervention type
PTA and stent	72 (82%)
PTA	14 (16%)
Atherectomy	2 (2%)

TASC: Trans-Atlantic inter-society consensus; PTA: percutaneous transluminal angioplasty.

TASC II classification available for 76 limbs.

Symptoms and DU findings

Over the course of 183 follow-up visits, patients presented with stable or worsened symptoms 71 times (38.8%) and improved or resolved symptoms 112 times (61.2%). The median SVR in symptomatic visits was 2.88 (range 1.1–9.29), which was significantly higher than the median SVR of 1.86 (range 0.77–9.29) in asymptomatic visits (p < 0.001) (Figure 1).

Figure 1.

Comparison of distribution of SVR between symptomatic and asymptomatic groups following treatment of all lesions (a), and subcategories of femoropopliteal lesions (n = 113) (b), and iliac lesions (n = 70) (c).

Of the 113 visits with interventions primarily associated with femoropopliteal disease, the median SVR in symptomatic visits was 2.99 (range 1.12–9.29), which was also significantly higher than the median SVR of 1.69 (range 0.77–9.29) in asymptomatic visits (p < 0.001). When examining the association between DU and symptoms following iliac artery intervention over 70 follow-up visits, there was no significant difference in the distribution of SVR between symptomatic and asymptomatic patients (2.39 vs 2.0, p = 0.19) (Figure 1).

ROC analysis

ROC curves, which allow for graphical representation of the sensitivity and specificity of a diagnostic test, were used to assess the relationship between SVR and binary clinical status (symptomatic versus asymptomatic). ROC curves permitted assessment of the accuracy of DU (as assessed by the calculated area under the curve), as well as the SVR threshold for restenosis that maximized sensitivity and specificity.

Over all 183 follow-up visits, area under the ROC curve was 0.74 (95% CI 0.67–0.81) (Figure 2), with distinct differences between those undergoing femoropopliteal interventions (ROC area = 0.82, [95% CI 0.74–0.83] versus iliac interventions (ROC area = 0.60 [95% CI 0.46–0.74]) (Figure 2). Next, patients who received a stent were analyzed separately from those who received PTA only. For visits following stent placement (n = 150), area under the ROC curve was 0.74 (95% CI 0.66–0.82), whereas for visits following PTA alone (n = 33), area under the ROC curve was 0.83 (95% CI 0.69–0.93).

Figure 2.

Receiver operator characteristic curve analysis of SVR and symptom status following treatment of all lesions (a), and subcategories of femoropopliteal lesions (n = 113) (b), and iliac lesions (n = 70) (c).

Over all follow-up visits, the most accurate SVR cutoff for predicting symptom status was 2.61 (sensitivity 63.4%, specificity 75.0%) (Table 4). For femoropopliteal lesions, the threshold of 2.5 distinguished symptom status with 70.8% sensitivity and 72.3% specificity, though specificity was slightly improved to 78.5% when using the threshold SVR = 2.61. For iliac lesions, sensitivity and specificity were lower (Table 3).

Table 3.

Sensitivity and specificity of different SVR threshold values for determining symptom status.

SVR threshold	Total (n = 183)		Femoropopliteal (n = 113)		Iliac (n = 70)
SVR threshold	Sensitivity (%)	Specificity (%)	Sensitivity (%)	Specificity (%)	Sensitivity (%)	Specificity (%)
1.5	93	36.6	97.9	40	82.6	31.9
2	78.9	52.7	89.6	55.4	60.9	48.9
2.5	63.4	69.6	70.8	72.3	47.8	67
2.61^a	63.4	75	70.8	78.5	47.8	70.2
3	47.9	82.1	50	86.2	43.5	76.6
3.5	40.9	88.4	47.9	93.9	26.1	78.7

Threshold that showed highest sensitivity and specificity in total and femoropopliteal lesions.

Table 4.

Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) for different SVR thresholds in follow-up for treatment of femoropopliteal lesions (n = 113).

SVR	Sensitivity (%)	Specificity (%)	PPV (%)	NPV (%)
1.5	97.9	40	54.7	96.3
2	87.5	55.4	59.2	85.7
2.5	70.8	72.3	65.4	77
3	50	87.7	75	70.4
3.5	47.9	93.8	85.2	70.9

For femoropopliteal lesions, using an SVR threshold of 2.5 accurately classified symptom status in 71.7% of cases. This was associated with a PPV = 65.4% and an NPV = 77.0% (Table 4).

Patients with discordant DUS and symptoms following treatment of femoropopliteal PAD

Following treatment for femoropopliteal disease, 28% of follow-up visits had discordance of DU and symptoms when an SVR cutoff of 2.5 was used. Symptomatic patients with SVR <2.5 comprised 12% of all visits while asymptomatic patients with SVR ≥2.5 comprised 16%. Further chart review was performed to determine the course of patients with discordance of DU and symptoms.

Eleven limbs were found to be symptomatic with SVR <2.5 over 14 follow-up visits. Of these 11 limbs, four underwent reintervention for their symptoms and seven were followed without intervention. No limbs in this group occluded. Median follow-up for these patients was 19 months (range 7–48 months).

Twelve limbs were asymptomatic with an SVR ≥2.5 over 18 follow-up visits. Of these, seven patients were followed without reintervention and five patients developed symptoms and subsequently were reintervened on. No patients underwent reintervention for abnormal DU in the absence of symptoms and no limbs occluded. Median follow-up for these patients was 17 months (range 8–62 months).

In order to better describe the outcomes of the cohort of patients found to have abnormal DUS, we reviewed the records of all limbs found to have elevated SVR. Following treatment of femoropopliteal PAD in 49 limbs, 26 had at least one documented SVR ≥2.5 (14 symptomatic, 12 asymptomatic). Of these, 14 limbs were not reintervened on (seven symptomatic, seven asymptomatic) over a median follow-up of 32 months (range 1–67 months). During this time, none of the untreated 14 patients with abnormal DUS progressed to occlusion.

Discussion

Angiographic assessment of patency has traditionally been the gold standard by which the success of surgical bypass or ET is gauged. DU-assessed patency has been accepted as a less invasive alternative to angiographic assessment and is widely employed in vascular practice, outcomes assessment, and as an endpoint in clinical trials.^6–10 However, the true “gold standard” by which the patient will determine the success of therapy is the presence or absence of symptoms, particularly following interventions for claudication. The relationship between DU-assessed patency and clinical status is yet to be evaluated. Here, we report the correlation between DU surveillance and symptom recurrence following ET for claudication.

This correlation was strong following intervention for femoropopliteal disease. Using ROC analysis, we determined that the SVR threshold of 2.5, previously derived using an angiographic standard, is very accurate in distinguishing symptomatic from asymptomatic patients after endovascular treatment of femoropopliteal disease. However, after treatment for iliac disease, sensitivity and specificity were inadequate to suggest a strong correlation between DU and symptom status. We also found good correlation between DU and symptom status for limbs treated with stents and those treated with PTA alone.

Several studies have evaluated the correlation between DU and angiographic restenosis following endovascular interventions in the lower extremities, visceral, and cerebral vascular beds.^13–18 Two studies have examined the association between DU and the “gold standard” of angiography following lower extremity ET. Baril et al.¹³ studied 59 limbs following superficial femoral artery endovascular intervention that had DU and angiography within 30 days of each other. They found that an SVR >1.5 accurately predicted ≥50% in-stent stenosis (sensitivity 93%, specificity 89%) and SVR >3.5 predicted ≥80% in-stent stenosis (sensitivity 74%, specificity 94%). Similarly, Shrikhande et al.¹⁴ studied 254 lesions in 103 patients and found that an SVR >2.5 had high sensitivity and specificity in predicting ≥70% restenosis. For this study, a majority of the lesions were in the superficial femoral artery. Both of these studies use angiography as the “gold standard” for determination of treatment failure. Our results add to these studies by indicating that objective determinants of impaired hemodynamics correlate not simply with angiographic restenosis but also with symptom recurrence. Furthermore, the agreement between objective (DU, angiography) and subjective (symptom status) indicators of adequate flow reinforce the utility of the current practice of DU surveillance following ET.

In addition to its use in clinical decision-making, DU is widely used in retrospective studies, and clinical trials including studies mandated for pre-market approval of novel devices, to define success and failure of ET.^6–10 The concordance we found between DU and subjective symptom status also validates the use of DU for defining treatment failure in clinical trials involving femoropopliteal ET for claudicants. It is certainly reassuring that this often-used outcome metric not only correlates with angiography but also the clinical status of the patient.

Further work needs to be done to determine the optimal surveillance program following endovascular intervention. Many believe DUS should be performed routinely following ET,^4,19–22 in part because areas of asymptomatic restenosis may quickly thrombose and occlude, at which point endovascular treatments have limited durability.^5,13 As a result, DU-detected restenosis is thought to be an adequate indication for reintervention even in the absence of recurrent symptoms. In the study by Baril et al.,¹³ DU-detected restenosis without evidence of recurrent symptoms was the indication for reintervention in 37% of patients.

Others have argued that DU surveillance does not add to clinical assessment in patient management.^23–25 Restenotic lesions after ET may follow a different natural history than restenosis after lower extremity bypass in that recurrence of symptoms is more likely to indicate severe restenosis rather than rapid progression to occlusion.²³ As a result, waiting for a patient to develop recurrent symptoms before reintervention may not pose any significant risk of thrombosis. Of the 12 limbs in our series that were asymptomatic with an SVR ≥2.5, none were treated for their abnormal DU alone, as is our practice. Five of the 12 patients eventually developed symptoms and required reintervention at that point, but none of these patients were occluded. This suggests that DU should be interpreted in the context of patient symptom status and that objective and subjective determinants of treatment failure should both be considered in clinical decision-making.⁹

The utility of DU following endovascular intervention in the iliac arteries remains unclear. Some authors advocate for routine DU surveillance,²⁶ while others will perform DU if a patient experiences symptomatic deterioration or if there is abnormal DU of the common femoral artery distal to the treated segment.²⁷ Our study showed that DU following ET for iliac lesions correlates poorly with recurrence of symptoms. This may be due to a more complex physiologic relationship between maintained patency and symptom recurrence for iliac lesions or due to inferior quality of DU when evaluating the iliac arteries. Compared to femoropopliteal lesions, iliac DU is less likely to be useful as a surveillance tool,²⁸ though it is likely to be helpful for targeted assessment of flow when there are other indications of treatment failure.

Limitations include the retrospective nature of the study, and the fact that conclusions drawn from this small series at a single institution are not necessarily widely applicable. For example, the majority of patients underwent bare metal stent placement, and despite our subgroup analysis of patients by treatment type (stents vs. angioplasty), it is unclear whether these findings may be applied to DUS following atherectomy, covered stent placement, or angioplasty alone. Furthermore, symptom status was dichotomized as simply, “symptomatic” or “asymptomatic.” However, a patient's subjective experience of symptom improvement or deterioration is more nuanced than this and would likely be better represented by extensive post-intervention quality of life measures or documentation of pain-free walking distance. Despite this, we believe our dichotomized variable describing symptom status was accurate and found that it correlated well with DUS findings. Another limitation is that we had no strict follow-up protocol in place because practices vary slightly between providers. Despite this, we captured a wide range of time points and all DUS were performed at the same lab, reflecting real-world practice. Finally, though we describe outcomes in patients who had untreated abnormal DUS and in those with discordant DUS and symptoms, the numbers in all subgroups were small, and although 1/3 of patients had follow-up of 2 years or more, the length of follow-up was variable, and in some cases, limited. We believe these results to be compelling but caution against drawing broad conclusions based on these descriptive data.

Conclusion

We have found that following femoropopliteal ET for claudication, DU findings correlate strongly with recurrence of symptoms and that sustained patency is necessary to prevent recurrent claudication. These findings support the use of post-intervention DU surveillance as objective criteria for treatment failure in clinical trials and registries and also suggest that following treatment for claudication, decisions regarding reintervention for treatment failure can be based on symptom status as a primary indicator of anatomic restenosis.

Footnotes

Acknowledgments

The content of this article was originally presented at the Peripheral Vascular Surgery Society (PVSS) winter meeting Jan 31 – Feb 3, 2013 in Park City, Utah, USA.

Conflict of interest

None declared.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

References

Mills

Harris

Taylor

Jr . The importance of routine surveillance of distal bypass grafts with duplex scanning: a study of 379 reversed vein grafts. J Vasc Surg 1990; 12: 379–386. discussion 387–389.

Idu

Blankenstein

de Gier

. Impact of a color-flow duplex surveillance program on infrainguinal vein graft patency: a five-year experience. J Vasc Surg 1993; 17: 42–52. discussion 52–53.

Mattos

van Bemmelen

Hodgson

. Does correction of stenoses identified with color duplex scanning improve infrainguinal graft patency? J Vasc Surg 1993; 17: 54–64. discussion 64–66.

Bandyk

Chauvapun

. Duplex ultrasound surveillance can be worthwhile after arterial intervention. Perspect Vasc Surg Endovasc Ther 2007; 19: 354–359. discussion 360–361.

Baril

Marone

. Duplex evaluation following femoropopliteal angioplasty and stenting: criteria and utility of surveillance. Vasc Endovascular Surg 2012; 46: 353–357.

Duda

Bosiers

Lammer

. Drug-eluting and bare nitinol stents for the treatment of atherosclerotic lesions in the superficial femoral artery: long-term results from the SIROCCO trial. J Endovasc Ther 2006; 13: 701–710.

Sabeti

Mlekusch

Amighi

. Primary patency of long-segment self-expanding nitinol stents in the femoropopliteal arteries. J Endovasc Ther 2005; 12: 6–12.

Schlager

Dick

Sabeti

. Long-segment SFA stenting--the dark sides: in-stent restenosis, clinical deterioration, and stent fractures. J Endovasc Ther 2005; 12: 676–684.

Siracuse

Giles

Pomposelli

. Results for primary bypass versus primary angioplasty/stent for intermittent claudication due to superficial femoral artery occlusive disease. J Vasc Surg 2012; 55: 1001–1007.

10.

Tosaka

Soga

Iida

. Classification and clinical impact of restenosis after femoropopliteal stenting. J Am Coll Cardiol 2012; 59: 16–23.

11.

Hajian-Tilaki

Hanley

Joseph

. A comparison of parametric and nonparametric approaches to ROC analysis of quantitative diagnostic tests. Med Decis Making 1997; 17: 94–102.

12.

Norgren

Hiatt

Dormandy

. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). J Vasc Surg 2007; 45(Suppl S): S5–67.

13.

Baril

Rhee

Kim

. Duplex criteria for determination of in-stent stenosis after angioplasty and stenting of the superficial femoral artery. J Vasc Surg 2009; 49: 133–138. discussion 139.

14.

Shrikhande

Graham

Aparajita

. Determining criteria for predicting stenosis with ultrasound duplex after endovascular intervention in infrainguinal lesions. Ann Vasc Surg 2011; 25: 454–460.

15.

AbuRahma

Srivastava

Stone

. Critical appraisal of the carotid duplex consensus criteria in the diagnosis of carotid artery stenosis. J Vasc Surg 2011; 53: 53–59. discussion 59–60.

16.

Baker

Chew

. Application of duplex ultrasound imaging in determining in-stent stenosis during surveillance after mesenteric artery revascularization. J Vasc Surg 2012; 56: 1364–1371. discussion 1371.

17.

Fleming

Davis

Craven

. Accuracy of duplex sonography scans after renal artery stenting. J Vasc Surg 2010; 52: 953–957. discussion 958.

18.

AbuRahma

Abu-Halimah

Bensenhaver

. Optimal carotid duplex velocity criteria for defining the severity of carotid in-stent restenosis. J Vasc Surg 2008; 48: 589–594.

19.

Kiguchi

Marone

Chaer

. Patterns of femoropopliteal recurrence after routine and selective stenting endoluminal therapy. J Vasc Surg 2013; 57: 37–43.

20.

Rybicki

Nallamshetty

Yucel

. ACR appropriateness criteria on recurrent symptoms following lower-extremity angioplasty. J Am Coll Radiol 2008; 5: 1176–1180.

21.

Humphries

Pevec

Laird

. Early duplex scanning after infrainguinal endovascular therapy. J Vasc Surg 2011; 53: 353–358.

22.

Spijkerboer

Nass

de Valois

. Evaluation of femoropopliteal arteries with duplex ultrasound after angioplasty. Can we predict results at one year? Eur J Vasc Endovasc Surg 1996; 12: 418–423.

23.

Bui

Mills

Sr Ihnat

. The natural history of duplex-detected stenosis after femoropopliteal endovascular therapy suggests questionable clinical utility of routine duplex surveillance. J Vasc Surg 2012; 55: 346–352.

24.

Tielbeek

Rietjens

Buth

. The value of duplex surveillance after endovascular intervention for femoropopliteal obstructive disease. Eur J Vasc Endovasc Surg 1996; 12: 145–150.

25.

Sacks

Robinson

Summers

. The value of duplex sonography after peripheral artery angioplasty in predicting subacute restenosis. Am J Roentgenol 1994; 162: 179–183.

26.

Davies

Bismuth

Saad

. Outcomes of reintervention for recurrent disease after percutaneous iliac angioplasty and stenting. J Endovasc Ther 2011; 18: 169–180.

27.

Back

Novotney

Roth

. Utility of duplex surveillance following iliac artery angioplasty and primary stenting. J Endovasc Ther 2001; 8: 629–637.

28.

Al Samaraee

McCallum

Cairns

. The results of high-frequency duplex surveillance after iliac arterial stenting in a single center. Vasc Endovascular Surg 2011; 45: 246–254.