Literature review and meta-analysis of the efficacy of cilostazol on limb salvage rates after infrainguinal endovascular and open revascularization

Abstract

Background:

Current clinical guidelines recommend the use of cilostazol in the treatment of patients with infrainguinal peripheral artery disease (PAD) who experience intermittent claudication. However, the role of cilostazol therapy in patients with advanced PAD and critical limb ischemia (CLI) remains unclear. To conduct a meta-analysis of randomized controlled trials and cohort studies that evaluated the effect of cilostazol vs standard antiplatelet therapy on limb-related and arterial patency-related outcomes. We also reviewed literature pertinent to the effect of cilostazol on wound healing in patients with advanced PAD.

Methods:

We performed a MEDLINE, EMBASE, COCHRANE (CENTRAL), SCOPUS, and US Clinical Trials database search for all trials and studies since 1999 that compared cilostazol with standard antiplatelet therapy in the setting of infrainguinal PAD revascularization procedures (endovascular or open). Aggregate data was collected from four randomized control trials and six retrospective cohort studies. The end point incidence ratios and treatment effects were generated from each study and reported as hazard ratios (HR) using a random-effect model. We also reviewed 10 studies that evaluated the effect of cilostazol on wound healing in patients with advanced PAD.

Results:

From more than 25,000 total patients, 3136 patients met our inclusion criteria. All patients had at least lifestyle-impacting intermittent claudication, and more than 50% met the definition of CLI (Rutherford class ≥4). Patient age range was 53 to 83 years, and the majority were male (66%). The mean follow-up time averaged 2 years across all studies. Meta-analysis revealed that cilostazol treatment favored amputation-free survival (hazard ratio [HR], 0.79; 95% confidence interval [CI], 0.69–0.91), limb salvage rate (HR, 0.42; 95% CI, 0.27–0.66), decreased repeat revascularization (risk ratio [RR], 0.44; 95% CI, 0.37–0.52), and decreased restenosis (RR, 0.68; 95% CI, 0.61–0.76). Cilostazol treatment also increased freedom from target lesion revascularization (RR, 1.35; 95% CI, 1.21–1.53) with no difference in all-cause mortality. Effective wound healing was found to be an inconsistent outcome measure in patients receiving cilostazol therapy.

Conclusions:

We observed that cilostazol therapy has a beneficial impact on all limb-related and arterial patency-related outcomes, but no effect on all-cause mortality in patients with advanced PAD and CLI undergoing revascularization procedures. Additional studies are needed to evaluate the effect of cilostazol therapy on wound healing in patients with advanced PAD.

Peripheral artery disease (PAD) affects more than 12 million Americans, and its net burden of disease is expected to grow given the increased prevalence of smoking, diabetes, hypertension, and obesity.^1,3 Lower extremity PAD can progress from an asymptomatic status to disabling symptoms of pain with walking consistent with intermittent claudication (IC). Cilostazol, a phosphodiesterase III inhibitor with presumed vasodilatory and mild antiplatelet activity, is demonstrated to be a cost-effective and efficacious medication for alleviating IC, improving quality of life, and increasing pain-free walking distance in patients with PAD.^4,5

A growing body of evidence suggests that in conjunction with endovascular or surgical intervention, cilostazol can significantly improve lower extremity outcomes in patients with advanced PAD.^6,7 The IRONIC trial demonstrated that incorporation of cilostazol therapy perioperatively with open and endovascular revascularization improved maximum walking distance and quality of life measures.⁸ Despite such findings, recent analysis of Medicare populations demonstrated marked underuse of cilostazol in patients with PAD.⁹ Additionally, there is a lack of systematic analyses exploring the potential benefits of cilostazol for limb-related and patency-related outcomes after revascularization.

We performed a systematic review of recent clinical trials to determine whether cilostazol therapy provided added benefit in amputation-free survival, limb salvage, freedom from target lesion revascularization (TLR), and arterial reintervention in patients with advanced PAD and critical limb ischemia (CLI). Based on prior data in patients with IC, we hypothesized that cilostazol therapy may also have important clinical benefits in patients with advanced PAD and CLI with few clinically significant side effects.

METHODS

Search strategy and study selection.

Based on a standard search strategy established by the Cochrane Collaboration, we systematically searched Embase 2020, Ovid Medline 2020, Scopus 2020, Cochrane Central Register of Controlled Trials (CENTRAL), clinicaltrials.gov 1997–2019, PubMed, and Medline databases. The published literature was searched using strategies created for the concepts of “amputation (primary and secondary)” and “cilostazol,” with secondary search string parameters of “endovascular repair,” “endovascular intervention,” “open repair,” “bypass,” and “wound healing” in adults more than 18 years old.¹⁰ The search strategies were established using a combination of standardized terms and key words. The set of studies identified by the primary parameters (“amputation” and “cilostazol”) was sufficient to encompass all studies considered for analysis.

All searches were completed in March 2020, resulting in a total of 210 citations. There were 94 duplicate citations from that pool using Endnote’s automatic duplication finder, leaving 116 unique citations in the project library. Fully reportable searches are presented in the Supplementary Search Methods (online only).

Our search included randomized control trials (RCTs), prospective and retrospective case control studies, and cohort studies. Three independent reviewers screened the titles, abstracts, and main bodies of 116 total searched articles and individually reviewed the eligible literature and extracted data from those that met the preset inclusion criteria. Articles reporting on overlapping or identical population subsets were screened so as to avoid including duplicate data in our analysis. An additional reviewer resolved any disagreement in screening or extraction and consensus was achieved unanimously.

Because standardized wound healing outcome measures in populations treated with cilostazol have historically not been reported, we separately searched these databases for “wound healing,” “repair,” “transcutaneous oxygen,” and “ulcer” with the same initial search strategy to screen for studies with cilostazol, which resulted in additional studies for our analysis.^11–15

Eligibility criteria.

We included all studies that include patients with infrainguinal, lower extremity PAD (Rutherford class ≥2), who were initiated on cilostazol before or at the time of index arterial revascularization (either surgical or endovascular), and a minimum clinical follow-up of at least 3 months after initiation of cilostazol therapy. Patient control groups were allowed to include placebo or any combination of antiplatelet medications excluding cilostazol. These antiplatelet medications included aspirin, clopidogrel, and ticlodipine. Studies that reported at least one of our predetermined end points were included in the analysis. These end points included amputation-free survival, the need for repeat revascularization, arterial restenosis, freedom from TLR, wound healing, and overall survival. Of note, patients with previous arterial repairs to the target vessel or previous primary-assisted patency interventions were excluded from this analysis. Eligibility criteria, mode of prespecified statistical analysis, search strategy, and independent review was approved by all study investigators for this meta-analysis, and the PRISMA guidelines were implemented in our reporting.¹⁶

Outcome reporting.

Outcomes of interest were categorized into limb-related outcomes (amputation-free survival and limb salvage), arterial patency-related outcomes (restenosis, freedom from TLR, and repeat revascularization), and all-cause mortality (Table I). We also summarized adverse events such as drug side effects and bleeding. RCTs were analyzed using an intention-to-treat analysis.

Table I.

Reported outcomes are subdivided into three major categories

Category	Outcome measure	Definition
Limb-related outcomes	Amputation-free survival (HR)	Avoidance of major amputation (above knee, below knee). Minor amputation was considered to be below the metatarsal (toes/distal foot/ray amputation).
	Limb salvage rate (HR)	Successful rate of lower-limb preservation without amputation or need for prosthesis.
Patency-related outcomes	Repeat revascularization rate	Rate of secondary revascularization (either endovascular or open surgical methods) distinct from TLR in the same limb as primary revascularization.
	Restenosis (RR)	Peak systolic velocity ratio ≥2.4 or reduction in >50% of vessel diameter by follow-up ultrasound or angiography.
	Freedom from TLR (RR)	Avoidance of repeat endovascular treatment of target lesion following initial lesion treatment.
Mortality outcomes	Mortality (RR)	All-cause mortality.

HR, Hazard ratio; RR, risk ratio.

Study statistics were verified to specifically fit in these definitions.

Data and statistical analysis.

For RCTs, we used the Cochrane Collaboration’s Risk of Bias Assessment Tool¹⁷ (Supplementary Fig 1, online only). Trials were categorized into low, high, or unclear amount of risk. For the retrospective cohort trials and observational studies, we used the Newcastle-Ottawa Quality Assessment scale¹⁸ (Supplementary Table, online only). This scale weighed studies on population selection, comparability, exposure, and outcomes.

Interstudy variation and heterogeneity was assessed using the I³ statistic with a P value of less than .05 as a delineator for statistical significance. We quantified the effect of heterogeneity by using the I³ index, which reflects the proportion of interstudy variability owing to heterogeneity as opposed to random chance. When three or more studies reported the same outcome measure, we used an I³ of greater than 50% as a statistical delineator for significant heterogeneity. We inspected studies for their effect size, population distribution, and sample size to address publication bias.

We also quantified the risk of publication bias by generating funnel plots for the restenosis outcome measure as it was a key non-mortality measure with a plurality of study capture. The DerSimonian and Laird, random effect model was used for all sensitivity analysis. We used the model’s method of evaluating interstudy variance (tau²) over an external approach. Effect estimates are presented as pooled risk ratios (RR) or hazard ratios (HR) with 95% confidence intervals (CI) where available in studies. We requested information on other end points from respective authors; however, if no response was obtained in a timely manner or data were not available on the desired end point, the study in question was excluded from the outcome analysis. This meta-analysis was conducted using STATA version 15.1 (StataCorp, College Station, Tex).

Ethics.

Trials included in our analysis were evaluated for risk of bias. This study is a meta-analysis and a review of the existing literature, and therefore required no individual or specific party consent. No additional review or ethical approval was necessary for this pooled analysis.

RESULTS

Study and patient capture.

Our initial search strategy identified 664 records of which 38 were duplicate findings. We ultimately excluded 601 based on a screen of abstracts, and 25 studies met our inclusion criteria (Fig 1). This resulted in the capture of 25,062 patients, from which 3136 patients were treated with cilostazol and 21,926 patients were treated with standard antiplatelet therapy without cilostazol. Standard antiplatelet therapy was considered to be at least aspirin as defined by the global vascular guidelines on the management of patients with advanced PAD.¹⁹

Fig 1.

Search strategy and subsequent article screening. RCTs, randomized, controlled trials.

Study and patient demographics.

Of the patients included in the study, 91% were gathered from retrospective reviews (Table II), and the mean follow-up ranged between 3 months and several years. Participants, in aggregate had a more than 95% compliance rate of taking cilostazol at the initiated dose. Table II indicates the timing of cilostazol initiation. Studies were published between 2008 and 2017, and were conducted in several countries including Japan, Spain, and the United States. The mean patient age was 75.6 ± 8.7 years, and 53.7% were male. Revascularization technique was endovascular in 67% of patients, and the remainder of patients underwent open bypass or peripheral endarterectomy procedures (Table III). Indications for cilostazol treatment and procedural revascularization included IC, rest pain, and CLI. One study included patients exclusively with IC (Rutherford class ≤3), and the remaining studies included a wide range of 11% to 100% of the study population consisting of patients with CLI (Rutherford class ≥5; Table III).

Table II.

Study characteristics listed above with key patient demographics used to compare study outcomes

Author	Year	Design	Indication (% CLI)	Procedural intervention	Total No. (% cilostazol)	Age, years	Male, %	Cilostazol dosage	Medical intervention comparison	Medical intervention cointervention	Cilostazol timing	Follow-up range, months
Ilda	2008	RCT	25%	EVT	127 (49.6%)	70 ± 8	27.60%	200 mg/d	Ticlodipine (200 mg/d)	Aspirin (100 mg/d)	Periprocedural	36
Ilda	2013	RCT	11%	EVT	200 (50%)	73 ± 8	68.60%	200 mg/d	None	Aspirin (100 mg/d)	Periprocedural	37
Soga20	2009	RCT	0% (100% IC)	EVT	78 (50.0%)	71 ± 8	32.50%	200 mg/d	None	Aspirin (100 mg/d) and ticlodipine (200 mg/d)	Periprocedural	24
Soga21	2011	Retrospective cohort	100%	EVT	618 (57.6%)	72 ± 11	69.20%	200 mg/d	None	Aspirin (100 mg/d) and clopidogrel (75 mg/d)	Periprocedural	21
Soga22	2012	Retrospective cohort	17%	EVT	861 (57.1%)	72 ± 7	71.00%	200 mg/d	None	Aspirin (100 mg/d) and clopidogrel (75 mg/d)	Periprocedural	25
Soga23	2017	RCT	100%	EVT	50 (50.0%)	73 ± 10	74.00%	200 mg/d	None	Aspirin (100 mg/d)	Periprocedural	3
Ishii24	2010	Retrospective cohort	26%	EVT	174 (37.2%)	66 ± 11	59.20%	200 mg/d	None	Aspirin (100 mg/d)	Periprocedural	37
Neel9	2015	Retrospective cohort	51%	EVT (37.47%), open (62.53%)	22954 (8.7%)	76 ± 8	44.48%	100–200 mg/d	None	Aspirin (100 mg/d)	Preoperatively	12
Perez	2014	Retrospective cohort	0% (100% IC)	None	1317 (14.5%)	66 ± 10	87%	100–200 mg/d	None	None	Currently taking	18
Chen	2011	Retrospective cohort	0% (100% IC)	None	14241 (16.6%)	69.9 ±16	52.30%	100–200 mg/d	None	None	Currently taking	7 years (max)

CLI, Critical limb ischemia; EVT, endovascular therapy; IC, intermittent claudication; RCT, randomized controlled trial.

Table III.

Key demographics and clinical factors from included studies

Author	Year	Indication (IC/CLI)	Total No.		Age, years		Sex			Smoking		Type II diabetes		Rutherford classification
			Cilos	Cont	Cilos	Cont		Cilos	Cont	Cilos	Cont	Cilos	Cont		Cilos	Cont
Ilda	2008	75/25	63 (49.6%)	64 (50.4%)	70 ± 8	70 ± 9	Male: Female:	44 (69.8%) 19 (30.2%)	48 (75%) 16 (25%)	39 (62%)	37 (58%)	46 (73%)	47 (72%)	1–3: 4: 5–6:	45 13 5	50 5 9
Ilda	2013	89/11	93 (48.7%)	98 (51.3%)	72 ± 9	73 ± 8	Male: Female:	64 (64%) 36 (36%)	67 (67%) 33 (33%)	41 (44%)	47 (48%)	52 (56%)	55 (56%)	1–3: 4: 5–6:	84 9 0	87 11 0
Soga20	2009	100/0	39 (50%)	39 (50%)	70 ± 7	72 ± 8	Male: Female:	31 (79.5%) 8 (20.5%)	34 (87.2%) 5 (12.8%)	12 (31%)	16 (41%)	13 (33%)	17 (44%)	1–3: 4: 5–6:	Not available
Soga21	2011	0/100	356 (57.6%)	262 (42.4%)	72 ± 11	71 ± 10	Male: Female:	242 (68%) 114 (32%)	165 (63%) 97 (37%)	258 (72.5%)	205 (78.2%)	107 (30.1%)	61 (23.3%)	1–3: 4: 5–6:	0 96 260	0 62 200
Soga22	2012	83/17	492 (57.1%)	369 (42.9%)	72 ± 9	72 ± 9	Male: Female:	348 (70.7%) 114 (19.3%)	233 (63.1%) 136 (36.9%)	181 (36.9%)	95 (25.9%)	283 (57.5%)	237 (64.6%)	1–3: 4: 5–6:	360 53 79	274 28 67
Soga23	2017	0/100	25 (50%)	25 (50%)	73 ± 10	73 ± 8	Male: Female:	18 (72%) 7 (28%)	19 (76%) 6 (34%)	12 (48%)	15 (60%)	17 (68%)	20 (80%)	1–3: 4: 5–6:	0 9 16	0 10 15
Ishii24	2010	74/26	61 (35.1%)	113 (64.9%)	66 ± 10	66 ± 11	Male: Female:	37 (60.7%) 24 (39.3%)	66 (58.4%) 47 (41.6%)	13 (21.3%)	29 (25.7%)	37 (60.7%)	84 (74.3%)	1–3: 4: 5–6:	35 14 12	63 17 33
Neel9	2015	49/51	1999 (8.7%)	20955 (91.3%)	76 ± 8	76 ± 7	Male: Female:	966 (48.3%) 1033 (51.7%)	9743 (46.5%) 11212 (53.5%)	Not available		732 (36.6%)	8084 (38.6%)	1–3: 4: 5–6:	Not available

Cilos, Cilostazol; CLI, critical limb ischemia; Cont, control; IC, intermittent claudication.

No significant difference were observed in patient sex, smoking status, and type 2 diabetes between treatment groups.

Patient demographics such as age and gender were not significantly different between patients treated with and without cilostazol (Table III). Similarly, rates of clinical comorbidities, such as diabetes, smoking status, and PAD severity were not different between patients treated with or without cilostazol that was reported in individual studies. Notably, in one study all patients had end-stage renal disease.²⁵

Cilostazol treatment demographics.

We observed variation in the cilostazol treatment window, which ranged from an unspecified number of years to the actual day of the index revascularization procedure. All studies offered cilostazol both preoperatively and postoperatively, and one study provided treatment only preoperatively (Table II). Dosing ranged from 100 to 200 mg twice a day (every 12 hours). The majority of studies reported high compliance with cilostazol therapy (>95% compliance). Only one study reported a lower compliance rate of 89.7%, although this rate is reasonable when comparing other retrospective trials looking at cilostazol in the context of PAD.²⁰ In all studies analyzed, only five patients reportedly withdrew from cilostazol therapy owing to adverse medication effects (19, 20).

Limb-related outcomes.

Limb-related outcomes were evaluated from reported clinical follow-up that extended up to 6 months from time of revascularization. Outcome measures included amputation-free survival and limb salvage rates.

Amputation-free survival was reported in three studies, all of which favored the use of cilostazol therapy. The HRs were 0.85 (95% CI, 0.71–0.98), 0.67 (95% CI, 0.49–91), and 0.44 (95% CI, 0.18–0.99), at approximately 1, 2, and 3 years, respectively.^9,21,24 Although this finding presents a trend of decreasing HR, all studies presented statistically significant net benefit of cilostazol compared to treatment with other antiplatelet agents. Using a random effect model, the combined summary estimate yielded a pooled HR of 0.79 (95% CI, 0.69–0.91) for amputation-free survival. The corresponding I³ of this outcome measure was 45.7% (P = .16; Fig 2, A).

Fig 2.

Limb-related outcomes with cilostazol therapy. A, Amputation-free survival was associated with a hazard ratio (HR) of 0.79 that favored cilostazol therapy. B, Limb salvage was associated with a HR of 0.42 that also favored cilostazol therapy.

Limb salvage rates were reported in two studies with HRs of 0.42 (95% CI, 0.25–0.69) and 0.44 (95% CI, 0.17–1.14). The two studies had mean follow-up times of 21 and 25 months and both favored the use of cilostazol for improved limb salvage; however, one of the studies did not yield a statistically significant benefit. The combined measure, using a random effects model, demonstrated a HR of 0.42 (95% CI, 0.27–0.66), suggesting that a net benefit was seen in patients receiving cilostazol therapy (Fig 2, B).

A pooled sensitivity analysis was performed to evaluate the effect of cilostazol with or without adjunct antiplatelet therapy (either monotherapy with aspirin or dual antiplatelet therapy [DAPT] with clopidogrel and aspirin). Our assessment was underpowered to determine a difference in limb-related outcomes between adjunct monotherapy and DAPT. However, amputation-free survival was successfully compared between cilostazol with or without adjunct DAPT. The pooled effect HR for amputation-free survival of patients treated with cilostazol in the setting of DAPT was 0.83 (95% CI, 0.71–0.97) with an I² of 54.8% (P = .137). The only study that evaluated DAPT in this context without cilostazol demonstrated a HR of 0.67 (95% CI, 0.49–0.91; Supplementary Fig 2, online only). These ratios were not statistically different, suggesting that in this limited data set the effect of cilostazol on amputation-free survival was not affected by adjunct use of DAPT.

Patency-related outcomes.

We took repeat revascularization rates as a potential surrogate marker for combined primary-assisted patency and secondary patency. Repeat revascularization rates were reported in one RCT of 78 patients who had IC, and one retrospective cohort of 618 patients who had CLI. Both studies reported similar RR favoring cilostazol of 0.41 (95% CI, 0.19–0.88) and 0.44 (95% CI, 0.37–0.53).^20,21 Using a random effect model, the combined outcome measure yielded a RR of 0.44 ((95% CI, 0.37–0.52; Fig 3, A).

Fig 3.

Arterial patency-related outcomes with cilostazol therapy. A, Repeat revascularization rate was associated with a risk ratio (RR) of 0.44 that favored cilostazol therapy. B, Restenosis rate was associated with a RR of 0.68, and was significantly lowered with cilostazol therapy. C, Freedom from target lesion revascularization (TLR) was associated with a RR also favoring cilostazol therapy.

Peripheral arterial restenosis was reported in six studies, and restenosis was the most common and consistent outcome measure reported in our search. All RCTs that we evaluated demonstrated that patients receiving cilostazol therapy had lower restenosis compared to any antiplatelet therapy over the same clinical follow-up times ranging between 2 and 3 years. Two retrospective cohorts similarly demonstrated significantly less restenosis rates in patients receiving cilostazol.^22,25 Using a random effect model the combined RR for restenosis was 0.68 (95% CI, 0.61–0.76; Fig 3, B).

Additionally, for patients who underwent endovascular treatments, we evaluated freedom from TLR as another end point for patency. All RCTs demonstrated a statistically significant freedom from TLR rates among patients receiving cilostazol therapy. One retrospective study observed a trend toward a benefit, but the 95% CI included 1.0. The random effect model yielded a combined RR of 1.35 (95% CI, 1.21–1.53; Fig 3, C).

A pooled sensitivity analysis was also performed to evaluate the effect of cilostazol with or without adjunct antiplatelet therapy on patency-related outcomes. Only restenosis as a primary outcome could be pooled across two or more studies. In the setting of DAPT, the RR for cilostazol usage was 0.73 (95% CI, 0.61–0.86) with an I² of 35.4% (P = .20). This was compared with a RR of 0.65 (95% CI, 0.55–0.77) for the only study that analyzed the effect of DAPT without cilostazol (Supplementary Fig 3, online only). Again, the difference between these variables was not statistically different. This finding suggests that cilostazol with or without DAPT did not have a large impact on restenosis rates. Given the limited dataset, generalizations about the effect of cilostazol with or without DAPT cannot be made at this time.

Mortality-related outcomes.

All-cause mortality was evaluated in our pooled analysis of the captured studies. No difference was observed between all-cause mortality between patients who received cilostazol or standard antiplatelet therapy independently in any of the studies or in the random effects model (Fig 4, A). Using a Cox proportions model, we also observed no difference in the mortality HRs between patients who received cilostazol vs standard antiplatelet therapy (Fig 4, B).

Fig 4.

All-cause mortality with cilostazol therapy. No difference was observed between control and cilostazol-treated groups in studies that reported either risk ratios (RR) (A) or hazard ratios (HRs) (B).

Adverse outcomes specifically related to cilostazol were reported in four studies. The most common adverse outcome was palpitations (which ranged from 4.8% to 8.0%) and headache (which ranged from 0% to 6.6%).^20,23–25 These four studies mentioned high variance in symptom severity with fewer than 900 patients having any side effects (constituting <1% percentage of total patients pooled). Given the low rate of reported adverse outcomes, our analysis was not adequately powered to determine if differences exist in adverse events between patients receiving cilostazol vs standard antiplatelet therapy.

Wound healing outcomes.

Wound healing outcomes were evaluated qualitatively in this systematic review because the data in the literature are limited. Wound healing outcome measures included time to wound healing or completion of wound healing as determined by an ulcer wound grade. Several low powered prospective and retrospective studies evaluated wound healing in patients with lower extremity PAD over predetermined time periods. These studies suggest that patients who use cilostazol have improved wound healing as measured by ulcer wound grade and/or decreased time to healing.^11–14 However, studies did not uniformly standardize their outcome measures and therefore a further sensitivity analysis could not be performed. Moreover, baseline wound classification or scoring was not performed for most patients in the captured analysis.

The remaining literature consists of case reports^22,24 or focuses on outcomes that may act as proxies to wound healing (eg, transcutaneous oxygen pressure), but not actual wound healing. One prospective study looked at the impact of cilostazol on the skin perfusion pressure (mm Hg) of lower extremities of 14 patients affected by PAD Rutherford class 3 and 4. A 1-month course of cilostazol therapy improved skin perfusion pressure from 24.5 ± 8.88 to 42.8 ± 21.0 mm Hg (P < .01).²⁰ Another prospective, randomized study similarly observed that patients receiving cilostazol had increased skin oxygen supply as measured by transcutaneous oxygen pressure measurement (37.1 ± 11.9 mm Hg vs 42.0 ± 9.7 mm Hg; P < .05).²¹

The majority of data regarding the impact of cilostazol on wound healing is additionally limited by the use of combination therapy (eg, cilostazol/aspirin, cilostazol/clopidogrel) or a multifactorial wound care protocol.¹⁵ Multivariate analysis demonstrates that cilostazol has a significant impact on wound healing.¹⁵ However, high-powered studies that examine the individual role of cilostazol are largely absent.

DISCUSSION

In this systematic review, we provide an updated analysis of modern RCTs and observational studies evaluating the impact of cilostazol therapy on patients with PAD who are undergoing either open or endovascular revascularization procedures. We report that patients who received cilostazol therapy along with open and/or endovascular intervention have overall improved outcomes. Specifically, patients who receive cilostazol have improved amputation-free survival, limb salvage rates, and decreased arterial restenosis rates. Additionally, patients who received cilostazol along with endovascular revascularization procedures also demonstrated higher freedom from TLR. We did not observe an increase in significant major adverse events in patients who received cilostazol, and there was no difference in all-cause mortality between patients who received and did not receive cilostazol therapy. These findings underscore the potential benefits associated with cilostazol therapy in patients with PAD undergoing arterial revascularization procedures, and also offer new avenues for future potential investigation.

Prior studies have also evaluated the effect of cilostazol in patients with PAD.²⁶ In eight well-designed clinical trials with 12- to 24-week follow-ups that involved more than 2000 patients, cilostazol was found to significantly improve the maximum walking distance and pain free walking distance in patients with IC.²⁷ Additional studies also demonstrated that treatment with cilostazol also significantly increased the resting ankle-brachial index by 6% to 11%, and importantly improved patient perceptions of functional status and quality of life compared with placebo.²⁸ Interestingly, withdrawal of cilostazol therapy for as short of a period of 2 weeks was found in a small cohort of 45 patients with IC to significantly decrease walking distance.²⁹ These prior data highlight the important therapeutic benefits of cilostazol in patients with PAD.

Prior studies have evaluated the effect of cilostazol in the perioperative setting.³⁰ A recent meta-analysis evaluated the effect of cilostazol on revascularization outcomes after endovascular therapy.⁷ In a review of more than 3800 patients with mild to moderate PAD (Rutherford classes 2 and 3), it was observed that cilostazol therapy was associated with higher primary patency, higher freedom from TLR, and a lower risk of amputation.⁹ In our study, we also included patients who received open revascularization procedures as well as patients with more advanced forms of PAD and CLI. Unlike patients with IC, patients with advanced PAD and CLI often require more timely revascularization for limb salvage purposes, and in this cohort we observed that cilostazol therapy was beneficial in patients who receive either open or endovascular procedures. Future studies such as the multinational and multicenter clinical trial Best Endovascular vs Best Surgical Therapy in Patients with Critical Limb Ischemia (BEST-CLI) may help to determine whether cilostazol has sustained clinical benefits during the 84-month study follow-up period.³¹

Previous pooled analyses of the effect of cilostazol on patients with advanced PAD after revascularization have included populations with limited demographic heterogeneity.^20,21,25 Initial meta-analyses that evaluated the effect of cilostazol on patients who received endovascular interventions was composed entirely of patients from Japan.^20,21,24,25 Our updated analysis benefits from the addition of more than 22,000 American patients. This analysis vastly expands the racial and demographic profile of the evaluated study population and the impact of cilostazol on a more diverse and generalizable patient population.

The mechanism of action of cilostazol on arterial patency and enhanced limb salvage in patients with PAD undergoing revascularization procedures is not entirely understood.⁴ Prior studies suggest that cilostazol may inhibit neointimal growth and therefore can decrease the risk of in-stent stenosis.³² However, we and others have observed that cilostazol seems to also improve primary patency, as well as arterial patency in patients who undergo open revascularization procedures without stent placement.⁹ Other studies also confirm that cilostazol therapy is associated with favorable outcomes, regardless of adjunct use of oral anticoagulants to assist with sustained arterial patency.²³ Our study adds to this body of evidence and supports the use of cilostazol for enhanced limb salvage independent of type of oral antiplatelet used.

We report that cilostazol has a relatively safe side effect profile, and no major adverse outcomes were uncovered in our systematic review of its use in patients with advanced PAD. Common side effects that include headache, tachycardia and diarrhea have been previously reported at low rates and are rarely life threatening.²⁶ Several studies have also demonstrated the safe use of triple antiplatelet therapy of cilostazol, aspirin, and thienopyridine, without increased risk of bleeding.³³ Bleeding events were not consistently reported in the studies captured for this analysis, and future studies are needed to better evaluate the therapeutic profile of cilostazol with or without clopidogrel, which is now a more commonly used antiplatelet following implantation of bare-metal and/or drug-eluting peripheral arterial stents.³²

Peripheral leg and foot wound healing are important surrogate end points for successful revascularization in patients with advanced PAD and CLI. Prior investigation of cilostazol and wound healing is limited to case reports and small cohort studies.^20–22,24 These studies indicate that cilostazol can enhance periwound oxygen perfusion, improve granulation tissue, and shorten wound-healing periods.^20–22,24 Our systematic review observed how the use of cilostazol enhances wound healing in patients with arterial disease. However, unfortunately, we observed a lack of consistent and comparable wound healing outcome measures.^22,24 Additional studies are needed for a more robust clinical investigation of the effects of cilostazol on wound healing in patients with advanced PAD after revascularization.

Limitations.

We note some limitations in our study. First, we decided to include patients with both IC and CLI to help determine the effects of cilostazol therapy in a broader range of patients affected by symptomatic PAD. We realize that this inclusion criterion may have introduced confounding variables associated with disease severity, distribution of disease, and types of adjunct interventions offered.^{7,9,13,20–25,34} Additionally, the perioperative use of statins, the type of bypass graft conduit (autologous vs prosthetic), and anatomical considerations such as distal bypass target (below the knee vs above the knee) may have influenced our data outcome variables.¹⁹ These factors were not uniformly reported in the studies captured in this review, and therefore could not be controlled for in our analysis.

Second, studies included in our analysis of cilostazol therapy varied by perioperative timing and length of clinical follow-up.^{7,9,13,20–25,34} Although these variables may have affected our study findings, we found low risk of bias in all of the studies captured in our analysis by applying the Cochrane Risk of Bias Assessment Tool and the Newcastle-Ottawa Tool (Supplementary Fig 1, online only). Third, in our evaluation we observed that some studies reported RR and others reported HR. We have included both outcome variables in our reporting but realize that these variables are not mutually exclusive and do not necessarily mirror one another.

Fourth, RCTs included in our analysis did not blind participants to intervention type, which may have a source of performance bias for which we could not control. Similarly, although compliance rates for this large aggregate of patient data were high (>95% compliance), the follow-up itself was variable, ranging from 3 months to 7 years. More long-term follow-up studies would further validate the impact of cilostazol over greater intervals.

Reporting and publication biases were difficult to assess since all studies evaluated demonstrated at least one significant benefit related to cilostazol therapy, and no study at this time has reported a negative outcome. Previous studies have not objectively measured publication bias through funnel plot analysis.^{7,9,13,20–25,34} We generated funnel plots for common outcome measures such as restenosis (Supplementary Fig 4, online only). These analyses again affirm concern for publication bias in prior reporting. Owing to a small effect size in some of our analyses, the CI is observed to be quite large. However, this impact on pooled outcome measures was respectively reflected in their weight.

Last, we observed a limited cohort of authors have indeed studied and published about the role of cilostazol in patients with IC and CLI.^{7,9,13,20–25,34} Although this serves as a potential source of bias, we observed a high level of consistency in reported outcome measures between author groups from different studies. We hope that, despite these limitations, our study provides new perspectives on the role of cilostazol in the treatment of patients with CLI, and highlights key knowledge gaps that remain to be addressed by future clinical investigation.

CONCLUSIONS

Over the past three decades, cilostazol has been demonstrated to be safe, efficacious, and cost effective in the treatment of patients afflicted by IC. This meta-analysis captures the largest and most diverse patient population to date and demonstrates that cilostazol improves limb-related and arterial patency-related outcomes in patients with advanced PAD and CLI who are undergoing endovascular and open revascularization procedures. Further trials are needed to evaluate the effect of cilostazol in patients receiving DAPT with clopidogrel, and the effects of cilostazol on wound healing outcome measures in patients with PAD.