Author + information
- Received August 15, 2016
- Revision received October 12, 2016
- Accepted October 20, 2016
- Published online January 16, 2017.
- Makoto Utsunomiya, MD, PhDa,∗ (, )
- Mitsuyoshi Takahara, MDb,c,
- Osamu Iida, MDd,
- Yasutaka Yamauchi, MDe,
- Daizo Kawasaki, MDf,
- Yoshiaki Yokoi, MD, PhDg,
- Yoshimistu Soga, MDh,
- Norihiko Ohura, MDi,
- Masato Nakamura, MD, PhDj,
- OLIVE Investigators
- aDivision of Cardiovascular Medicine, Tokyo Rosai Hospital, Tokyo, Japan
- bDepartment of Metabolic Medicine Graduate School of Medicine, Osaka University, Osaka, Japan
- cDepartment of Diabetes Care Medicine, Graduate School of Medicine, Osaka University, Osaka, Japan
- dCardiovascular Division, Kansai Rosai Hospital, Amagasaki, Japan
- eCardiovascular Center, Kikuna Memorial Hospital, Yokohama, Japan
- fCardiovascular Center, Morinomiya Hospital, Osaka, Japan
- gDepartment of Cardiology, Kishiwada Tokushukai Hospital, Kishiwada, Japan
- hDepartment of Cardiology, Kokura Memorial Hospital, Kitakyushu, Japan
- iDepartment of Plastic and Reconstructive Surgery, School of Medicine, Kyorin University, Mitaka, Japan
- jDivision of Cardiovascular Medicine, Ohashi Medical Center, Toho University, Tokyo, Japan
- ↵∗Reprint requests and correspondence:
Dr. Makoto Utsunomiya, Division of Cardiovascular Medicine, 4-13-21 Omori-minami, Ohta-ku, Tokyo, Japan 143-0013.
Objectives This study aimed to assess the optimal angiographic endpoint of endovascular therapy (EVT) for wound healing.
Background Several reports have demonstrated acceptable patency and limb salvage rates following infrapopliteal interventions for the treatment of critical limb ischemia (CLI). However, the optimal angiographic endpoint of EVT remains unclear.
Methods We conducted a subanalysis of the prospective multicenter OLIVE (Endovascular Treatment for Infrainguinal Vessels in Patients with Critical Limb Ischemia) registry investigation assessing patients who received infrainguinal EVT for CLI. We analyzed data from 185 limbs with ischemic ulcerations classified as Rutherford class 5 or 6, managed with EVT alone (i.e., not undergoing bypass surgery). The wound healing rate after EVT was estimated by the Kaplan-Meier method. The association between final angiographic data and wound healing was assessed employing a Cox proportional hazards model.
Results The overall wound healing rate was 73.5%. The probabilities of wound healing in patients with wound blush obtainment was significantly higher than that of those without wound blush (79.6% vs. 46.5%; p = 0.01). In the multivariate analysis, wound blush obtainment was an independent predictor of wound healing.
Conclusion The presence of wound blush after EVT is significantly associated with wound healing. Wound blush as an angiographic endpoint for EVT may serve as a novel predictor of wound healing in patients with CLI.
Endovascular therapy (EVT) is becoming increasingly important as a means of achieving revascularization in patients with critical limb ischemia (CLI) (1). The growing importance of this therapeutic approach is largely attributable to advances in devices, such as wires and balloons, and to improvements in treatment techniques that have enabled treatment of the elderly and patients with systemic comorbidities (2). In EVT, which is associated with lower patency rates than bypass surgery, it is necessary to efficiently deliver blood to a wound (3,4). Thus, revascularization with consideration of angiosomes where the wound is located and assessment of blood flow in arteries not only below the knee (BTK), but also below the ankle (BTA) have recently been proposed to be important on the basis of the angiosome concept (5,6). However, there is no clearly established endpoint regarding angiographic findings that allow discontinuation of EVT for CLI and prediction of ulcer healing. We previously reported that the presence or absence of wound blush (WB), which is defined as an area densely stained with contrast media around the wound on the final angiograms during EVT, correlates significantly with lower limb salvage (7). In addition, there are several reports of studies demonstrating that the angiographic findings after EVT correlates with subsequent lower limb outcomes (8,9). However, all of these studies were retrospective, and no comparisons of these aspects of performing EVT for CLI have been reported. Thus, the present study aimed to analyze the data from the OLIVE (Endovascular Treatment for Infrainguinal Vessels in Patients with Critical Limb Ischemia) registry, a multicenter prospective observational study conducted in Japan, to identify the optimal angiographic endpoints for assessing wound healing in response to EVT.
In this subanalysis employing data from the OLIVE registry, we included patients with tissue loss (Rutherford class 5 to 6). On the final angiograms during EVT, we determined how many BTK vessels (0 to 3) were depicted, how many BTA vessels (0 to 2) were depicted, whether the pedal arch was depicted, whether direct flow based on the angiosome concept was obtained, and whether WB was obtained. Then, these results were analyzed to determine whether they correlated with ulcer healing. To determine whether direct flow based on the angiosome concept was obtained, images of each wound were analyzed by the Wound Core Lab to determine which angiosome was involved in the wound, and the final angiograms were analyzed to determine whether revascularization had been achieved in the angiosomes. In the event of a wound extending over several angiosomes, direct flow was to be determined if revascularization had been achieved in all angiosomes. WB would be considered to be positive if final angiography revealed that contrast media injected from a catheter implanted in the popliteal artery reached the area around the wound even if it was a part. Although patients with evaluable images were enrolled regardless of whether they had received digital subtraction angiography or conventional imaging, patients were excluded if their feet were not imaged, if WB could not be determined due to frequent leg movements during imaging, or if they had undergone carbon dioxide angiography. All angiograms were independently evaluated by an experienced revascularization specialist who was blinded to the clinical courses of the patients.
The judgment of the wound healing was accomplished by the plastic surgeon in the wound core lab.
The OLIVE registry was a prospective study conducted at 19 centers in Japan and the clinical outcomes were assessed at 1, 3, 6, and 12 months after EVT. All patients in the OLIVE registry provided informed consent prior to enrollment, and the study protocol was approved by the institutional review board of each participating center.
During the period from December 2009 to July 2011, patients with de novo CLI (Rutherford class 5 to 6) caused by infrainguinal disease were enrolled in this analysis. The following patients were excluded from the analysis: those without wounds (Rutherford class 4), those with wound images on which angiosomes could not be evaluated because of poor image quality, and those with final images inadequate for WB evaluation because of poor image quality or frequent body movements during imaging.
Intervention procedure and clinical follow-up
The EVT strategy and wound treatment before and after EVT were left to the discretion of the treating physicians at each center. Lesions in the superficial femoral artery were treated with a provisional nitinol stenting strategy, whereas popliteal lesions and BTK-BTA lesions were treated with balloon angioplasty. Post-procedural skin perfusion pressure was obtained within 12 to 48 h after the index procedure. Ulcers were photographed before and after the procedure and again at 1, 3, 6, and 12 months after EVT. All photographs were independently reviewed by an experienced plastic surgeon blinded to the baseline characteristics of the patients and to the EVT strategies used. The time required for 100% epithelialization of the reference wound was determined in each case. In the case with multiple wounds, when all wounds were healed it was judged to be wound healing.
Data are presented as the mean ± SD for continuous variables and number (percentage) for discrete variables, unless otherwise indicated. The wound healing rate after EVT was estimated by the Kaplan-Meier method, based on categorization according to whether or not WB was obtained. The association between final angiographic data and wound healing was assessed by the Cox proportional hazards model. A p value <0.05 was considered to indicate a statistically significant difference. Hazard ratios and their 95% confidence intervals are presented. All statistical analyses were performed using IBM SPSS statistics software version 22 (SPSS Inc., Chicago, Illinois).
From the OLIVE registry investigation, 38 patients were excluded because there was no tissue loss (only resting pain), 10 due to low quality of the wound picture before the procedure, 56 because angiography around the wound was not fully recorded after EVT, and 47 due to the low quality, attributable to excessive patient movement, of the angiography performed to evaluate WB. Of the 199 evaluable patients, 14 were lost to follow up, and the remaining 185 (93%) completed the 1-year follow-up assessment.
Baseline patient characteristics are shown in Table 1. The mean age was 73 years. There were 123 men (67%). Fifty-one patients were unable to walk without assistance (28%). The risk factors were diabetes mellitus in 138 patients (75%), hypertension in 147 patients (80%), and chronic kidney disease in 117 patients (63%). There were 103 patients with a history of dialysis (56%). Cilostazol was used in 93 patients (50%). The average body mass index of 22 and the statin utilization rate of 22% in this study population were lower than those indicated in American and European study. There was not significant difference in baseline patient’s characteristics between WB-positive group and negative group. Limb characteristics are shown in Table 2. According to the Rutherford classification, there were 40 patients with class 6 CLI (22%) and 123 (67%) had wounds involving their toes. Mean baseline ankle-brachial index was 0.72 and is thought to be relatively high. This result could be falsely elevated secondary to medial calcinosis in the patients with dialysis and diabetes. Skin perfusion pressure (SPP) value showed in Table 2 was mean SPP value measured at the dorsal and plantar of the foot. Post-procedural SPP value in positive WB group was higher than negative WB group.
Forty-three cases (23%) received treatment for only the femoropopliteal lesion, and 113 patients received only infrapopliteal intervention. Of these 113 cases, 72 cases (39%) received treatment for 1 BTK vessel, 63 cases (34%) received treatment for 2 BTK vessels, and 7 cases (4%) received treatment for all 3 BTK, vessels. Thirty cases (16%) received additional treatment for the BTA lesions.
The final angiographic data after catheterization are shown in Table 3. WB was positive in 142 patients (76%) and negative in 43 (23%). As to BTK lesions, 72 patients had 1 patent vessel, 70 had 2 patent vessels, and 28 had 3 patent vessels, whereas 15 patients had none. As to BTA lesions, either the plantar or the dorsalis pedis artery was patent in 104 patients, in 64 of whom both arteries were patent. No BTA runoff was obtained in 17 patients. Moreover, the pedal arch was patent in 104 patients. According to the assessment of angiosomes measured by the Wound Core Lab, the wound was determined to be located in the posterior tibial artery angiosome in 51 patients, the posterior tibial artery angiosome in 25, and the peroneal angiosome in 2, whereas the wound was determined to extend over 2 or more angiosomes in 102 patients. When the term angiosome direct was defined as a condition in which all BTK vessels supplying the affected angiosome were patent, then angiosome direct was achieved in 77 patients (41%), whereas angiosome indirect was found in 108 (58%). In the group with positive WB, there were many numbers of the below the knee and below the ankle runoff in the completion angiography. In addition, the ratio that pedal arch was patent was higher, and the ratio of direct flow based on angiosome concept was higher than in the negative WB group.
Kaplan-Meier curves representing the association between WB and the ulcer healing rate are shown in Figure 1. The ulcer healing rate was 79.6% in patients with positive WB but only 46.5% (i.e., significantly lower) in those with negative WB (p = 0.010).
The results of multivariate analysis regarding the associations between angiographic data after EVT and ulcer healing are shown in Table 4. WB was the only independent variable suggesting ulcer healing among the various angiographic variables examined. The results of multivariate analysis adjusted for CLI severity in the lower limbs and general condition are shown in Table 5. Even after adjustment for each of these variables, WB remained an independent predictor of ulcer healing. Representative cases of positive and negative WB are shown in Figure 2.
The present study has confirmed that the delivery of blood to wounds is important for wound healing when EVT is performed for CLI patients with wounds. Specifically, when angiography after EVT demonstrates contrast media reaching areas around the wound, showing WB, ulcer healing can be expected with a high probability. Even after this association was adjusted for the number of patent BTK vessels, BTA runoff status, the presence or absence of the pedal arch, and direct or indirect flow based on the angiosome concept, WB was the only independent predictor of ulcer healing. Moreover, WB was still an independent predictor after adjustment for CLI severity in the affected limb or general condition. These observations suggest that delivery of blood to the wound is extremely important for ulcer healing.
As yet, no angiographic endpoint in EVT for CLI has been established. Although 1 straight-line flow used to be regarded as important, the lower limb outcomes of patients with only 1 straight-line flow, based on the blood flow achieved with EVT, are considerably poorer than in those whose flow was obtained by bypass surgery because of its lower patency rate (10,11). In recent years, there have been several articles highlighting the importance of the angiosome concept (5,8). The angiosome concept states that, with consideration of the angiosomes supplied by BTK vessels, direct revascularization of vessels supplying the wound improves lower limb outcomes. This concept is theoretical and extremely easy to understand. However, there are also many opposing views on this concept. In fact, some studies have found no correlation between the angiosome concept and lower limb outcomes (6,12,13); therefore, the association between the concept and prognosis is still far from being established. This may be attributable, for instance, to the problem that the angiosome concept is based on angiosomes identified by injection of dye into cadaveric lower limbs without vascular insufficiency and does not take collateral circulation in the feet into consideration. Though there are reports regarding the importance of the status of BTA lesions and the association of the presence or absence of pedal arch depiction with prognosis (6,9), WB was found to be more important in the present study. Needless to say, the angiosome concept, the status of BTA lesions, and the pedal arch may be all important, albeit indirectly. However, the present study revealed that a simple index of perfusion at the wound site is the most important index for wound healing.
WB may also be an extremely important index in daily clinical practice. When EVT is performed in CLI patients, it is often difficult to treat all affected vessels and lesions. If WB is obtained when EVT is applied to treatable lesions, ulcer healing can be expected, and treatment can be discontinued. Moreover, if WB is not observed on the final angiograms taken after EVT for treatable vessels, further treatment, such as amputation at a higher level and bypass surgery, should be considered. This index, simple WB, appears to be useful because it differs from indices that have been objectively established. Skin perfusion pressure or other noninvasive assessment tools to evaluate the perfusion around wound was thought to be important to predict wound healing (14). However, in this study, after adjustment, post-EVT SPP value did not correlate with wound healing. In this study, the SPP values collected were measured at dorsal and plantar. And we showed the mean SPP value in the text. Because 67% of enrolled patients had wound in the toes, it might be the cause of this result that a measurement position was far from a wound.
It is a subanalysis of the OLIVE registry data, obtained from a study that was not originally designed to analyze the association between WB and wound healing.
Thus, in the patients in the OLIVE registry, treatment had been discontinued without angiography of the feet in 56 patients. This may have reduced the accuracy of the present study. Because an evaluation of the perfusion around the wound was not regarded as important at the time of OLIVE registry, the evaluation was not required in the protocol. Therefore, it was not able to confirm WB in many cases.
Moreover, we were unable to assess WB in 47 patients with inadequate images due to frequent foot movements during angiography. Involuntary movements of patients, particularly of their feet, make it difficult to assess WB. Because patients with frequent involuntary movements or those unable to follow the orders of operators may have included those with a poor prognosis, the possibility of selection bias cannot be ruled out. However, the selection bias may not be so important because the comparison between the analysis group and the analysis exclusion group showed that the patient background was almost similar in both groups (Online Table 1).
WB should have been judged by independent experts more than 2 to keep objectivity. And we should have stated the observer variability but did not have it in this study.
Risk factors for delayed wound healing recognized so far including ankle-brachial index, SPP, below the knee, and below the ankle runoff were not able to show the significant association in this study. This suggests that sample size might be insufficient. However, WB could predict wound healing significantly in such situation with small sample size of this study. WB may suggest the stronger association with the wound healing than such known risk factors. A study with large sample size will be needed to confirm whether WB is independently related to wound healing.
The presence of WB after EVT is significantly associated with wound healing. WB as an angiographic endpoint in EVT may serve as a novel predictor of wound healing in patients with CLI.
WHAT IS KNOWN? There is no clearly established angiographic endpoint of EVT for CLI.
WHAT IS NEW? WB is the most important endpoint to predict ulcer healing.
WHAT IS NEXT? Additional studies are required to confirm the reliability of wound blush to predict wound healing.
For a supplemental table, please see the online version of this article.
The authors have reported that they have no relationships relevant to the contents of this paper to disclose.
- Abbreviations and Acronyms
- below the ankle
- below the knee
- critical limb ischemia
- endovascular therapy
- skin perfusion pressure
- wound blush
- Received August 15, 2016.
- Revision received October 12, 2016.
- Accepted October 20, 2016.
- American College of Cardiology Foundation
- Shiraki T.,
- Iida O.,
- Takahara M.,
- et al.
- Higashimori A.,
- Iida O.,
- Yamauchi Y.,
- et al.
- Alexandrescu V.,
- Soderstrom M.,
- Venermo M.
- Utsunomiya M.,
- Nakamura M.,
- Nagashima Y.,
- Sugi K.