Abstract
Background
A recent meta-analysis reported late excess mortality in patients treated with paclitaxel-coated devices (PCDs) for symptomatic femoropopliteal disease. However, this finding is controversial.
Objectives
To investigate the impact on mortality and predictors of repeat exposure to PCDs in patients with lower extremity peripheral arterial disease (LE-PAD).
Methods
We analyzed registry patient-level data from two centers. A total of 214 patients were enrolled, and stratified based on terciles of cumulative dose of paclitaxel. We treated 134 patients with a single PCD exposure and 80 with multiple PCD exposures. We used the follow-up index (FUI) in Kaplan-Meier survival estimates to minimize potential selection bias. We used Cox proportional hazard and splines models to determine the predictors of mortality and assess their relationships with mortality.
Results
The mean cumulative dose of paclitaxel was significantly different among groups (6.40 mg vs. 15.06 mg vs. 38.57 mg, p < 0.001). The 5-year FUI (0.93 ± 0.19 vs. 0.94 ± 0.18 vs. 0.95 ± 0.15, p = 0.836) and survival rates were not different (65.4% vs. 51.9% vs. 72.0%, p = 0.148). There was no dose-response association between paclitaxel dosage and death (p = 0.297). The predictors of death were congestive heart failure, stroke, dialysis dependence, neutrophil-lymphocyte ratio (NLR) > 3, age > 71 years, and body mass index (BMI) < 20 kg/m2. Spline model analysis validated the non-linear associations between mortality, age, BMI, and NLR.
Conclusions
Repeated PCD exposure for LE-PAD did not result in excess late mortality. Predictors of mortality might change over time, and continuous variables had non-linear relationships with death.
Keywords: Mortality, Peripheral artery disease, Predictor
Abbreviations
BMI, Body mass index
BSA, Body surface area
BTK, Below-the-knee
CLTI, Chronic limb-threatening ischemia
CVA, Cerebrovascular accident
EVT, Endovascular therapy
FUI, Follow-up index
LE-PAD, Lower extremity peripheral arterial disease
NLR, Neutrophil-lymphocyte ratio
PCDs, Paclitaxel-coated devices
SAFE-PAD, SAFETY Assessment of Femoropopliteal Endovascular Treatment With Paclitaxel-Coated Devices
SWEDEPAD, Swedish Drug-Elution Trial in Peripheral Arterial Disease
TRENDPAD, Tzuchi Registry of ENDovascular Intervention for Peripheral Artery Disease
INTRODUCTION
A recent meta-analysis reported late excess mortality in patients treated with paclitaxel-coated devices (PCDs) for symptomatic femoropopliteal disease.1 In addition, the Food and Drug Administration issued warnings about the late excess mortality related to PCD exposure and recommended against their use except in high-risk patients.2 However, subsequent studies have not shown a difference in mortality3,4 or improved survival in patients treated with PCDs.5
The aforementioned meta-analysis has also been criticized for lacking long-term, homogeneous, patient-level data identifying confounding factors to explain the observations. Several studies using patient-level data have not shown an association between increased mortality and the use of PCDs.6-8 Despite improved vessel patency using PCDs, a substantial number of patients require repeat exposure to paclitaxel either due to lesion restenosis or due to the involvement of both legs.9,10 The cumulative effect of repeated exposures to paclitaxel is unclear.
This study aimed to assess the outcomes of repeat exposure to PCDs for patients with lower extremity peripheral arterial disease (LE-PAD) using patient-level data from prospective registries in Taiwan.
METHODS
Study population
We identified the primary study cohort from the Tzuchi Registry of ENDovascular Intervention for Peripheral Artery Disease (TRENDPAD), an ongoing, prospective, single-center, observational registry of patients who have undergone endovascular therapy (EVT) for lower limb ischemia since July 2005. We selected 249 legs in 198 patients treated with PCDs for symptomatic LE-PAD between March 2013 and December 2019. Additionally, we recruited 17 legs in 16 patients treated with PCDs for symptomatic femoropopliteal disease at Tainan Municipal Hospital from June 2013 to December 2018.
The inclusion criteria were patients who underwent angiographic interventions for LE-PAD due to de novo lesions and in-stent restenosis lesions, and patients who had pre-existing or re-established adequate vessel run-off to the foot after EVT. Patients who had undergone concomitant iliac or tibial interventions were allowed. The exclusion criteria were: acute thrombotic occlusions, prior use of covered stents, previous bypass graft anastomosis lesions, previous use of PCDs without device information, contraindication to aspirin or clopidogrel, life-threatening infections, follow-up duration < 12 months in surviving patients, and malignancy. Figure 1 shows the flow chart of study participants, and they were stratified based on the terciles of cumulative paclitaxel dose. We obtained informed consent from each patient. The study protocol complied with the Declaration of Helsinki, and the Human Research Ethics Committee of both participating institutions approved the study (09-X-067 for Taipei Tzu-Chi Hospital and 1030705 for Tainan Municipal Hospital).
Figure 1.
Flow chart of study participants. PCDs, paclitaxel-coated devices.
Interventional procedure
The strategy of EVT was left to the discretion of the treating physicians. Four types of paclitaxel-coated balloons, including 430 IN.PACT Admiral or Pacific (3.5 μg/mm2; Medtronic Ireland, Galway, Ireland), 9 Lutonix (2 μg/mm2; Bard, Wexford, Ireland), 23 Passeo-18 Lux (3 μg/mm2; Biotronik AG, Bülach, Switzerland), 114 RangerTM (2 μg/mm2; Boston, Würselen, Germany), and two drug-eluting stents, 30 Zilver PTX (Cook Ireland, Limerick, Ireland) and 7 Eluvia (Boston, Marlborough, MA, USA), were used for the femoropopliteal lesions. We treated below-the-knee (BTK) lesions with three kinds of paclitaxel-coated balloons, including 26 RangerTM, 15 Lutonix, and 10 Passeo-18 Lux. The details of the interventional procedures have been described previously.11,12 In the case of suboptimal angiographic results or flow-limiting dissection after balloon angioplasty as determined by a translesion pressure gradient ≥ 10 mmHg in the femoropopliteal artery or residual diameter stenosis > 50% in the femoropopliteal or infrapopliteal arteries, we implanted provisional bare-metal stents or drug-eluting stents in the femoropopliteal lesions, but only bare-metal stents for the proximal tibial lesions.
The use of PCDs in both legs or repeated use of PCDs in the same leg due to lesion restenosis or lesion progression was defined as repeated exposure to PCDs. The Geriatric Nutritional Risk Index was calculated as (14.89 × albumin [g/dL]) + (41.7 × body weight/ideal body weight),13,14 and the prognostic nutritional index was calculated as 10 × serum albumin (g/dl) + 0.005 × total lymphocyte count (per mm3).15,16 The severity of vessel calcification was graded using the peripheral artery calcification scoring system.17
Follow-up
After EVT, the patients underwent scheduled clinical visits and non-invasive tests, including evaluating the ankle or toe brachial index, transcutaneous oxygen pressure, and duplex ultrasound, as previously reported.12 The indication for reintervention was clinically driven; therefore, asymptomatic patients with restenosis or reocclusion after PCD did not undergo repeat revascularization. The main adverse events were documented during follow-up, including death, amputation, repeat intervention, or other vascular events. The study ended on December 31, 2021.
To reduce the risk of attrition bias and increase the credibility of survival estimates in this observational study, we used the follow-up index (FUI) to examine the actual survival rate, as previously reported.18 The FUI was defined as the ratio of the exact follow-up period to the possible follow-up period at the individual patient level. The outcome of this study was the time between index treatment and death. Potential participants who did not reach the endpoint by five years or were lost to follow-up were censored. The participants in whom we could not ascertain survival information were eventually excluded from the Kaplan-Meier analysis in the verified group.
Statistical analysis
We used SPSS for Windows, version 22.0 (IBM Corp., Chicago, IL, USA) for statistical analyses. Descriptive statistics for continuous variables are presented as mean ± standard deviation and median and interquartile range; descriptive statistics for categorical variables are presented as frequency (percentage). We compared continuous variables using one-way ANOVA and categorical data using the chi-square test. We compared non-parametric continuous and ordinal variables using the Kruskal-Wallis H test. Survival was estimated using Kaplan-Meier analysis and compared using the log-rank test between groups. We used Cox proportional hazards models to identify the independent predictors of mortality at different time points by incorporating adjusted paclitaxel dose by body surface area (BSA), clinically, and anatomically relevant variables. We determined the cut-off values for the predictors of mortality using receiver operating characteristic curves. Associations between mortality with age, body mass index (BMI), and neutrophil-lymphocyte ratio (NLR) were further analyzed using a spline model in R software version 4.0.0.19,20 A p value < 0.05 was considered statistically significant.
RESULTS
We treated 134 legs in 134 patients with a single PCD exposure and 80 patients with repeated exposures to PCDs, including 38 (76 legs), 28 (28 legs), and 14 (28 legs) patients for both legs, lesion restenosis, and both legs and lesion restenosis, respectively. We used femoropopliteal PCDs in 187 patients and 227 legs; and 27 patients with 39 legs were treated with BTK PCDs. Ten patients with 17 legs were treated with PCDs for isolated BTK disease, 10 (15 legs) for concomitant BTK and femoropopliteal lesions, and seven with repeated exposures to PCDs for BTK vessel restenosis, not the femoropopliteal segment. There were 76, 68, and 70 patients in the lower, middle, and upper terciles, respectively.
Table 1 summarizes the baseline demographics of the study participants. The lower tercile patients had the highest proportion of congestive heart failure, and the middle tercile had more patients on dialysis or BMI < 20 kg/m2 than the other groups. There were no significant differences in age, sex, other medical illnesses, and essential laboratory data, including hemogram, lipid profile, C-reactive protein level, serum albumin level, and discharge medication. Lesion and procedural characteristics are detailed in Supplementary Table 1. The patients shared similar interventional characteristics, except that the repeat exposure group had a higher proportion of poor run-off vessels, atherectomy or plaque-modifying devices, and smaller BTK vessels than the single exposure group.
Table 1. Patient characteristics.
| Factors | All patients | Lower tercile | Middle tercile | Upper tercile | p value |
| No. of patients | 214 | 76 | 68 | 70 | |
| Age† (years) | 71.2 ± 11.8 | 70.9 ± 12.9 | 71.1 ± 11.9 | 71.6 ± 10.5 | 0.935 |
| Male# | 113 (53) | 44 (58) | 35 (52) | 34 (49) | 0.511 |
| Diabetes mellitus# | 167 (78) | 60 (79) | 54 (79) | 53 (76) | 0.847 |
| Hypertension# | 180 (84) | 65 (86) | 55 (81) | 60 (86) | 0.677 |
| Coronary artery disease# | 113 (53) | 35 (46) | 37 (54) | 41 (59) | 0.302 |
| Congestive heart failure# | 35 (16) | 18 (24) | 11 (16) | 6 (9) | 0.048* |
| Cerebrovascular accident# | 35 (16) | 17 (22) | 7 (10) | 11 (16) | 0.145 |
| Dialysis dependence# | 85 (40) | 23 (30) | 35 (52) | 27 (39) | 0.033* |
| Smoking history# | 82 (38) | 28 (37) | 25 (37) | 29 (41) | 0.808 |
| Hyperlipidemia# | 133 (62) | 45 (59) | 43 (63) | 45 (64) | 0.799 |
| Atrial fibrillation# | 28 (13) | 10 (13) | 11 (16) | 7 (10) | 0.561 |
| CLTI# | 118 (55) | 48 (63) | 34 (50) | 36 (51) | 0.213 |
| Body surface area† (m2) | 1.57 ± 0.15 | 1.58 ± 0.15 | 1.56 ± 0.14 | 1.58 ± 0.15 | 0.641 |
| Body mass index† (kg/m2) | 23.7 ± 3.41 | 23.9 ± 3.73 | 22.9 ± 3.07 | 24.2 ± 3.27 | 0.068 |
| Body mass index# < 20 kg/m2 | 27 (13) | 8 (11) | 14 (21) | 5 (7) | 0.047* |
| LDL-cholesterol† (mg/dL) | 95 ± 33 | 97 ± 34 | 90 ± 32 | 96 ± 33 | 0.382 |
| HbA1c† (%) | 7.31 ± 1.78 | 7.29 ± 1.82 | 7.17 ± 1.58 | 7.46 ± 1.93 | 0.638 |
| WBC‡ | 6954 (5073, 8978) | 6889 (5688, 9030) | 6810 (5738, 8963) | 7090 (5665, 8958) | 0.769 |
| Hematocrit‡ (%) | 34.8 (30.5, 38.5) | 35.2 (31.6, 38.4) | 34.3 (30.2, 38.6) | 34.1 (29.9, 39.6) | 0.762 |
| Platelet count‡ (103/μL) | 209 (155, 271) | 216 (159, 258) | 196 (140, 277) | 221 (171, 279) | 0.308 |
| NLR‡ | 3.37 (2.35, 5.17) | 3.53 (2.28, 5.15) | 3.80 (2.45, 5.22) | 3.02 (2.29, 5.17) | 0.437 |
| PLR‡ | 141 (104, 205) | 153 (106, 194) | 143 (109, 214) | 147 (91, 238) | 0.915 |
| PNI‡ | 40.5 (35.6, 45.4) | 41.2 (35.7, 45.6) | 39.9 (34.9, 43.7) | 40.8 (36.6, 46.5) | 0.273 |
| CRP level‡ (mg/dL) | 0.58 (0.18, 2.12) | 0.62 (0.19, 4.46) | 0.61 (0.23, 1.70) | 0.51 (0.14, 1.94) | 0.342 |
| Albumin† (g/dL) | 3.23 ± 0.63 | 3.22 ± 0.68 | 3.20 ± 0.60 | 3.27 ± 0.60 | 0.809 |
| GNRI† | 92.4 ± 11.4 | 92.6 ± 12.8 | 90.6 ± 10.6 | 94.0 ± 10.4 | 0.207 |
| Medication | |||||
| Aspirin# | 130 (61) | 42 (55) | 42 (62) | 46 (66) | 0.425 |
| Clopidogrel# | 193 (90) | 64 (84) | 62 (91) | 67 (96) | 0.062 |
| Cilostazol# | 123 (58) | 48 (63) | 40 (59) | 35 (50) | 0.265 |
| ACEI or ARB# | 102 (48) | 42 (55) | 30 (44) | 30 (43) | 0.253 |
| Statin# | 99 (46) | 40 (53) | 28 (41) | 31 (44) | 0.357 |
| Beta-blocker# | 114 (53) | 43 (57) | 29 (43) | 42 (60) | 0.096 |
| Calcium channel blocker# | 86 (40) | 34 (45) | 26 (38) | 26 (37) | 0.597 |
| Insulin# | 83 (39) | 32 (42) | 26 (38) | 25 (26) | 0.726 |
| Follow-up duration‡ (years) | 3.80 (2.25, 5.67) | 3.61 (2.24, 5.71) | 3.40 (1.71, 4.85) | 4.19 (2.37, 6.44) | 0.085 |
ACEI, angiotensin-converting enzyme inhibitor; ARB, angiotensin receptor blocker; CLTI, chronic limb-threatening ischemia; CRP, C-reactive protein; GNRI, geriatric nutritional risk index; HbA1C, glycohemoglobin; LDL, low-density lipoprotein; NLR, neutrophil-lymphocyte ratio; PLR, platelet-lymphocyte ratio; PNI, prognostic nutritional index; WBC, white blood cell.
Lower tercile: 1.10~8.60 mg; middle tercile, 8.60~21.16 mg; upper tercile, 21.17~101.79 mg.
* A p value < 0.05 indicates a statistically significant difference between groups. # Data are presented as n (%). † Data are presented as mean ± SD. ‡ Data are expressed as median and (interquartile ranges; Q1, Q3).
Supplementary Table 1. Lesion and procedural characteristics.
| Factors | All legs | Repetition of PCDs: (-) | Repetition of PCD: (+) | p value |
| No. of affected legs | 266 | 134 | 132 | |
| Target-limb ABI† | 0.53 ± 0.15 | 0.52 ± 0.14 | 0.54 ± 0.16 | 0.362 |
| Numbers of the stiff artery (ABI ≥ 1.4)# | 32 (12.0) | 12 (8.9) | 20 (15.1) | 0.120 |
| Intermittent claudication# | 121 (46) | 65 (49) | 56 (42) | 0.261 |
| Rest pain# | 31 (12) | 19 (14) | 12 (9) | |
| Non-healing ulcer# | 90 (34) | 39 (29) | 51 (39) | |
| Gangrene# | 24 (9) | 11 (8) | 13 (10) | |
| Concomitant intervention | ||||
| Iliac intervention# | 18 (6.8) | 10 (7.5) | 8 (6.1) | 0.649 |
| BTK intervention# | 198 (74.4) | 97 (72.4) | 101 (76.5) | 0.440 |
| Poor BTK runoff# | 190 (71.4) | 88 (65.7) | 102 (77.3) | 0.036* |
| PA involvement# | 137 (51.5) | 72 (53.7) | 65 (49.2) | 0.464 |
| Severe calcification# | 81 (30.5) | 37 (27.6) | 44 (33.3) | 0.311 |
| Occlusion# | 109 (43.0) | 60 (44.8) | 49 (37.1) | 0.204 |
| Lesion type | ||||
| De novo lesions# | 155 (57.4) | 85 (63.4) | 70 (53.0) | 0.116 |
| Restenotic lesions# | 52 (19.5) | 26 (19.4) | 26 (19.7) | |
| In-stent lesions# | 59 (22.2) | 23 (17.2) | 36 (27.2) | |
| Lesion length FP portion† (mm) | 203 ± 118 | 200 ± 113 | 205 ± 122 | 0.748 |
| Lesion length BTK portion† (mm) | 125 ± 66 | 136 ± 83 | 121 ± 61 | 0.525 |
| FP bailout stent# | 114 (44.4) | 58 (44.6) | 56 (44.1) | 0.933 |
| FP bailout stent length† (mm) | 131 ± 78 | 133 ± 74 | 128 ± 82 | 0.703 |
| Length of PCD (mm) | ||||
| FP artery† (mm) | 240 ± 123 | 234 ± 123 | 246 ± 124 | 0.420 |
| BTK artery† (mm) | 151 ± 74 | 147 ± 88 | 152 ± 70 | 0.846 |
| IVUS# | 81 (30) | 36 (27) | 45 (34) | 0.200 |
| Atherectomy or plaque-modifying devices# | 50 (17) | 18 (13) | 32 (24) | 0.024* |
| RVD of the SFA† (mm) | 5.40 ± 0.74 | 5.36 ± 0.83 | 5.45 ± 0.64 | 0.334 |
| Final MLD of the SFA† (mm) | 4.45 ± 0.73 | 4.40 ± 0.80 | 4.51 ± 0.66 | 0.295 |
| Lumen gain in the SFA† (mm) | 3.49 ± 0.97 | 3.53 ± 1.06 | 3.45 ± 0.87 | 0.518 |
| RVD of the PA† (mm) | 4.79 ± 0.73 | 4.81 ± 0.76 | 4.68 ± 0.70 | 0.708 |
| Final MLD of the PA† (mm) | 3.91 ± 0.77 | 3.89 ± 0.78 | 3.94 ± 0.75 | 0.708 |
| Lumen gain in the PA† (mm) | 3.06 ± 1.00 | 3.17 ± 0.97 | 2.95 ± 1.04 | 0.203 |
| Final balloon size of FP artery† (mm) | 5.58 ± 0.74 | 5.52 ± 0.72 | 5.65 ± 0.76 | 0.173 |
| RVD of the BTK vessels† (mm) | 2.90 ± 0.63 | 3.41 ± 0.58 | 2.73 ± 0.56 | 0.002* |
| Final MLD of the BTK artery† (mm) | 2.35 ± 0.58 | 2.94 ± 0.54 | 2.16 ± 0.45 | < 0.001* |
| Lumen gain in the BTK artery† (mm) | 2.01 ± 0.62 | 2.57 ± 0.56 | 1.82 ± 0.52 | < 0.001* |
| Final balloon size of BTK artery† (mm) | 2.85 ± 0.47 | 3.22 ± 0.44 | 2.73 ± 0.41 | 0.003* |
ABI, ankle-brachial index; BTK, below-the-knee; DCB, drug-coated balloon; FP, femoropopliteal; IVUS, intravascular ultrasound; MLD, minimal lumen diameter; PA, popliteal artery; PCD, paclitaxel-coated device; RVD, reference vessel diameter; SD, standard deviation; SFA, superficial femoral artery.
* A p value < 0.05 indicates a statistically significant difference between the groups.
# Data are presented as n (%). † Data are presented as mean ± SD.
Table 2 shows the nominal doses of paclitaxel used in the three groups. The mean doses of paclitaxel in the first EVT (6.04 mg vs. 13.31 mg vs. 18.93 mg, p < 0.001) were significantly different. The cumulative amounts of paclitaxel (6.40 mg vs. 15.06 mg vs. 38.57 mg, p < 0.001) were also significantly different even adjusting by BSA (4.08 mg/m2 vs. 9.78 mg/m2 vs. 24.64 mg/m2, p < 0.001).
Table 2. Nominal dose of paclitaxel.
| Nominal dose of paclitaxel | All patients (N = 214) | Lower tercile (N = 76) | Middle tercile (N = 68) | Upper tercile (N = 70) | p value |
| No. of exposure to paclitaxel† | 1.6 ± 0.9 | 1.1 ± 0.3 | 1.2 ± 0.4 | 2.3 ± 1.2 | < 0.001* |
| Number of drug-coated devices† | 3.1 ± 2.5 | 1.9 ± 1.0 | 2.4 ± 0.9 | 5.6 ± 2.8 | < 0.001* |
| Body surface area† (m2) | 1.57 ± 0.15 | 1.58 ± 0.15 | 1.56 ± 0.14 | 1.58 ± 0.15 | 0.641 |
| First paclitaxel dose† (mg) | 12.57 ± 7.630 | 6.04 ± 2.10 | 13.32 ± 4.36 | 18.93 ± 8.25 | < 0.001* |
| First paclitaxel dose# (mg) | 10.40 (6.94, 17.33) | 6.85 (4.49, 8.13) | 13.62 (10.39, 17.20) | 20.85 (12.76, 25.69) | < 0.001* |
| First paclitaxel dose§ (mg/m2) | 8.07 ± 4.97 | 3.1 ± 2.5 | 3.1 ± 2.5 | 3.1 ± 2.5 | < 0.001* |
| Cumulative paclitaxel dose† (mg) | 19.68 ± 17.11 | 6.40 ± 1.89 | 15.06 ± 3.45 | 38.57 ± 17.64 | < 0.001* |
| Cumulative paclitaxel dose# (mg) | 14.26 (7.72, 25.13) | 6.95 (4752, 8414) | 14.68 (11.89, 17.20) | 33.46 (25.24, 44.68) | < 0.001* |
| Cumulative paclitaxel dose‡ (mg/m2) | 12.61 ± 10.97 | 4.08 ± 1.24 | 9.78 ± 2.51 | 24.64 ± 11.34 | < 0.001* |
Lower tercile: 1.10~8.60 mg; middle tercile: 8.60~21.16 mg; upper tercile: 21.17~101.79 mg.
* A p value < 0.05 indicates a statistically significant difference between the groups. # Data are expressed as median and (interquartile ranges; Q1, Q3). † Data are presented as mean ± SD. ‡ Data are presented as mean ± SD and adjusted by body surface area.
SD, standard deviation.
Follow-up outcomes
During a median follow-up of 3.80 (IQR: 2.25-5.67) years, 74 patients died. Infectious disorders such as sepsis with multi-organ failure and pneumonia were the leading causes of death, followed by cardiovascular disorders (Supplementary Table 2). The cumulative mortality rates from 1 to 5 years were 10.3%, 15.1%, 20.9%, 28%, and 36.2%, respectively. The 3-year and 5-year FUIs were 0.97 ± 0.12 and 0.94 ± 0.18, respectively. The 3-year (79.1% vs. 77.8%, p = 0.748) and 5-year (63.8% vs. 62.2%, p = 0.639) survival rates in the censored and verified groups were similar (Figure 2A and 2B). The 5-year FUI among the three groups was not significantly different (0.93 ± 0.19 vs. 0.94 ± 0.18 vs. 0.95 ± 0.15, p = 0.836); therefore, no survival difference was found between the groups stratified by paclitaxel dose tercile (65.4% vs. 51.9% vs. 72.0%, p = 0.148) (Figure 2C). The middle and upper terciles reached a marginal survival difference at 5 years (p = 0.064). After being adjusted by BSA, the cumulative paclitaxel dose was still not associated with all-cause mortality (p = 0.297).
Supplementary Table 2. Causes of death.
| Factors | All patients (N = 214) | Lower tercile (N = 76) | Middle tercile (N = 68) | Upper tercile (N = 70) | p value |
| Death events | 74 (34.6) | 24 (31.6) | 28 (41.2) | 22 (31.4) | 0.383 |
| Cardiovascular disorder | 26 (12.1) | 9 (11.8) | 11 (16.2) | 6 (8.6) | 0.391 |
| Stroke | 3 (1.4) | 1 (1.3) | 1 (1.5) | 1 (1.4) | 0.997 |
| Pneumonia | 11 (5.1) | 3 (3.9) | 7 (10.3) | 1 (1.4) | 0.052 |
| Sepsis with multi-organ failure | 26 (12.1) | 8 (10.5) | 8 (11.8) | 10 (14.3) | 0.780 |
| Cancer | 6 (2.8) | 3 (3.9) | 1 (1.5) | 2 (2.9) | 0.539 |
| Renal failure decline dialysis therapy | 2 (4.1) | 0 (0) | 0 (0) | 2 (2.9) | 0.125 |
Lower tercile: 1.10~8.60 mg; middle tercile, 8.60~21.16 mg; upper tercile, 21.17~101.79 mg.
Data are presented as n (%).
Figure 2.
Kaplan-Meier estimates rates of freedom from all-cause mortality in three years (A) and five years (B) between the censored (blue line) and verified survivors (red line) and in the groups stratified according to cumulative paclitaxel dose terciles (C). The blue, green, and red lines denote lower, middle, and upper terciles. SE, standard error.
Table 3 shows the independent predictors of mortality identified using the Cox proportional hazard model over five years. The predictors of mortality might change over time. Old age (> 71 years) predicted death across the disease continuum. Dialysis dependence, congestive heart failure, and BMI < 20 kg/m2 were the predictors of later death. In contrast, cerebrovascular accident (CVA)-related mortality became weaker during follow-up. Elevated NLR was significantly associated with 1- to 4-year death, but the power diminished at 5 years. This finding suggested that the status of immune-inflammatory response played an essential role in all-cause mortality beyond traditional risk factors in the earlier period. We used a fully adjusted spline model to test the predictive ability of age, BMI, and NLR for mortality as continuous variables. The curve depicting the relationship between death and NLR or age was wide, flat and J-shaped, and showed a trend towards an increased risk of death at NLR > 3 or age > 71 years at the upper extremes. However, the relationship between BMI and death was biphasic. The risk of mortality increased at BMI < 20 kg/m2, and a high BMI had a survival benefit (Figure 3A-C).
Table 3. Multivariate analysis to identify the predictors of all-cause deaths during the follow-up period.
| Items | One-year mortality | Two-year mortality | Three-year mortality | Four-year mortality | Five-year mortality | |||||
| OR (95% CI) | p value | OR (95% CI) | p value | OR (95% CI) | p value | OR (95% CI) | p value | OR (95% CI) | p value | |
| Age > 71 years | 7.334 (1.938-27.75) | 0.003* | 9.180 (2.804-30.06) | < 0.001* | 7.684 (2.893-20.41) | < 0.001* | 6.213 (2.670-14.46) | < 0.001* | 4.322 (2.121-8.808) | < 0.001* |
| Atrial fibrillation | 1.684 (0.592-4.793) | 0.329 | 2.021 (0.877-4.659) | 0.099 | 2.079 (0.964-4.042) | 0.047* | 1.822 (0.953-3.486) | 0.070 | 1.424 (0.778-2.604) | 0.252 |
| BMI < 20 kg/m2 | 1.568 (0.522-4.712) | 0.423 | 1.676 (0.680-4.131) | 0.262 | 1.931 (0.913-4.086) | 0.085 | 2.009 (1.014-3.981) | 0.046* | 2.297 (1.173-4.498) | 0.015* |
| Coronary artery disease | 1.418 (0.450-4.473) | 0.551 | 1.192 (0.479-2.964) | 0.706 | 1.616 (0.737-3.546) | 0.230 | 1.049 (0.531-2.072) | 0.891 | 1.077 (0.576-2.013) | 0.816 |
| Congestive heart failure | 2.211 (0.855-5.722) | 0.102 | 2.500 (1.116-5.601) | 0.026* | 2.215 (1.099-4.462) | 0.026* | 2.176 (1.160-4.081) | 0.015* | 2.468 (1.335-4.564) | 0.004* |
| CVA | 5.137 (1.793-14.71) | 0.002* | 4.689 (1.934-11.37) | 0.001* | 3.425 (1.562-7.512) | 0.002* | 2.455 (1.147-5.256) | 0.021* | 1.865 (0.941-3.697) | 0.074 |
| CLTI | 1.054 (0.323-3.445) | 0.930 | 1.777 (0.659-4.794) | 0.256 | 1.459 (0.665-3.198) | 0.346 | 1.209 (0.606-2.415) | 0.590 | 1.036 (0.542-1.980) | 0.915 |
| Dialysis dependence | 2.732 (1.000-7.461) | 0.050 | 2.465 (1.057-5.749) | 0.037* | 2.554 (1.258-5.188) | 0.009* | 2.330 (1.257-4.321) | 0.007* | 2.796 (1.519-50146) | 0.001* |
| Sex (Female) | 1.962 (0.701-5.490) | 0.199 | 1.268 (0.563-2.859) | 0.566 | 1.548 (0.779-3.079) | 0.213 | 1.305 (0.703-2.422) | 0.398 | 1.067 (0.622-1.831) | 0.813 |
| C-reactive protein | 0.973 (0.672-1.408) | 0.884 | 0.880 (0.637-1.216) | 0.439 | 0.944 (0.720-1.439) | 0.679 | 0.916 (0.721-1.162) | 0.468 | 0.889 (0.718-1.100) | 0.279 |
| NLR | 2.792 (1.230-6.339) | 0.014* | 2.979 (1.427-6.222) | 0.004* | 3.175 (1.695-5.946) | < 0.001* | 2.179 (1.254-3.785) | 0.006* | 1.282 (0.783-2.100) | 0.323 |
| Albumin | 0.836 (0.334-2.093) | 0.702 | 0.948 (0.439-2.045) | 0.891 | 0.984 (0.514-1.654) | 0.961 | 0.859 (0.477-1.546) | 0.612 | 0.620 (0.360-1.068) | 0.085 |
BMI, body mass index; CI, confidence interval; CLTI, chronic limb-threatening ischemia; CVA, cerebrovascular accident; NLR, neutrophil-lymphocyte ratio; OR, odds ratio.
* A p value < 0.05 was considered statistically significant.
Figure 3.
Non-linear relationships between all-cause mortality and age (A), body mass index (BMI) (B), and neutrophil-lymphocyte ratio (NLR) (C) using spline model analysis adjusted for age, body mass index, neutrophil-lymphocyte ratio, prognostic nutritional index, and geriatric nutritional risk index. The vertical axis denotes the adjusted hazard ratio (HR) versus age 71 (A), BMI 20 (B), and LnNLR 1.1 (or NLR 3.0) (C). The solid and dotted lines represent estimated hazard ratios and 95% confidence intervals.
DISCUSSION
This study is the first to use patient-level data to demonstrate the impact on mortality of repeated exposures to PCDs for patients with symptomatic LE-PAD, including femoropopliteal and BTK disease. The cumulative paclitaxel dose was not associated with death during 105 months of follow-up. The only independent predictor of mortality across the disease continuum was age. Elevated NLR and stroke predicted early and mid-term mortality. Moreover, CHF, dialysis dependence, and low BMI (< 20 kg/m2) were predictors of later death. Non-linear relationships existed between mortality and continuous variables including age, BMI, and NLR.
Safety of repeated exposure to PCDs
Given the limited time since the introduction of PCDs into clinical practice, their long-term safety has not been well established. Katsanos et al. reported more late mortality associated with PCDs than that related to uncoated devices 2-5 years after the index treatment.1 They also showed a positive association between paclitaxel dose and the absolute risk of death. However, patient-level data from industry clinical research, retrospective analysis of administrative databases, subgroup analyses of clinical trials, and interim analysis of an ongoing randomized controlled open-label trial of PCDs have not shown a significant impact of PCDs on survival.3-8 All of these studies are longitudinal observations after a single exposure to PCD; they provide no further information about repeated exposures to PCDs in patients with contralateral LE-PAD or lesion progression in the same leg. The safety of cumulative paclitaxel dosage used during repeated exposures to PCDs remains uncertain. In our previous study, we found no correlation between mid-term mortality and the cumulative dose of paclitaxel in femoropopliteal interventions.21 Compared to single-level treatment, concomitant femoropopliteal and BTK interventions will increase paclitaxel exposure. In the current study, we enrolled 27 patients treated with BTK PCDs for data reanalysis. The mortality outcome was not different with a higher exposure to paclitaxel, either in single EVT with a large-dose or cumulative exposure due to repeated interventions during longer follow-up.
The average dose of paclitaxel ranges from 230 mg to 300 mg for single chemotherapy treatment and up to 1200 mg for multiple chemotherapy treatments. The mean dose of paclitaxel delivered by PCDs in previous clinical trials has ranged from 1 mg to 20 mg based on lesion length, the number of lesions treated, and the technology used.6 In rare circumstances, patients have received up to 70 mg of paclitaxel,22 however, a previous study showed increased systemic side effects, such as moderate neutropenia, sensory neuropathy, and alopecia, at a dose of 70 mg/m2.23
For local drug delivery in vascular interventions, paclitaxel is highly lipophilic, which mediates the rapid uptake of paclitaxel into tissue with high concentrations in the intima layer of arteries and resulting in low plasma concentrations. A prior animal study showed the presence of paclitaxel in local vasculature for up to 60 days,24 and paclitaxel was undetectable in plasma by 24 hours when investigated in LE-PAD after paclitaxel-coated balloons with up to three balloons.25 Unlike the systemic administration of anticancer therapy, the characteristics of high tissue concentration and low plasma level of paclitaxel in vascular interventions attenuate the effect of liver metabolism by cytochromes P450 2C8 and 3A4 and BSA. In addition, our study showed no association between all-cause mortality and paclitaxel dose adjusted by BSA (p = 0.297). The patients who received maximal dose exposure (62.52 mg/m2) remained healthy 8 years after PCD treatment. There is currently no plausible mechanism to link the increase in mortality and the use of PCDs.
Follow-up index
The original RCTs pooled in the meta-analysis were designed to examine limb-related outcomes of paclitaxel.1 Therefore, significant study attrition may have occurred after their primary endpoints were reached, primarily within 1 year. In addition, a considerable amount of missing data may have influenced the results after 2 years. Incomplete follow-up with selectively recorded events may have led to relevant misestimations, especially in longitudinal observational studies.26,27 This is important because it may have remained unnoticed within flawed Kaplan-Meier estimates.
The FUI, a flexible and straightforward measure, is defined by three individual dates readily available for every patient in any clinical-outcome research (i.e., date of inclusion and treatment, date of last contact, and study end date). Using the FUI may reduce bias due to unaccounted follow-up duration and increase the credibility of outcome estimates. Using censored or verified data, our analysis showed no difference in survival at 3 or 5 years. A high FUI strengthens the credibility of actual survival estimates and reduces the methodological limitations in cohort studies.
Predictors of all-cause mortality
Recently, the Swedish Drug-Elution Trial in Peripheral Arterial Disease (SWEDEPAD) and Safety Assessment of Femoropopliteal Endovascular Treatment With Paclitaxel-Coated Devices (SAFE-PAD) study reported high mortality rates in patients with chronic limb-threatening ischemia (CLTI). However, they observed no harm associated with the use of PCDs.3,4 Our study complements these reports, because 54% of our participants had CLTI, with 5-year survival rates of 72.6% in claudicants and 56.4% in patients with CLTI (p = 0.005). Two recent trials of PCDs, mainly composed of claudicants, showed that the predictors of mortality were age, diabetes mellitus, congestive heart failure, renal insufficiency, and stroke.7,28 In patients with CLTI, dialysis dependence, low BMI, Rutherford 6, and ambulatory status were reported to be independent predictors of death after EVT in several studies.29-31 In this study, we also investigated the predictors of death at different time points. Age was the sole predictor of death across the disease continuum. Dialysis dependence, CHF, and low BMI were predictors of later death. These findings may explain the lower 5-year survival rate in the middle tercile group, as this group contained more patients with dialysis dependence and BMI < 20 kg/m2. NLR, a biomarker of the immune-inflammatory response, has been reported to be a survival predictor in patients with LE-PAD.31-33 Elevated NLR reflects both neutrophilia in inflammation and relative lymphopenia, suggesting a more profound imbalance of immune response, which in turn leads to the progression of atherosclerosis, vascular events, and mortality.34
The risk of NLR-related mortality differed over time, with the highest hazard ratio in the third year. The lower NLR level at the index EVT or improved immune-inflammatory status in patients surviving for more than 5 years might be associated with a lesser impact on mortality at 5 years. More attention should be paid to solving the problem of impaired immune-inflammatory response in symptomatic LE-PAD patients in the future.
We further performed analysis using a spline model to test the predictors of mortality and reduce potential bias due to the dichotomy of numerical explanatory variables or common information loss in an observational study.35 Our data demonstrated non-linear associations between mortality and continuous variables, including age, BMI, and NLR. Adjustment for covariates, including age > 71 years, BMI < 20 kg/m2, and NLR > 3, strongly increased the risk of death. Our findings are consistent with the impact of the obesity paradox on survival36 and impaired immune-inflammation status37,38 in patients with LE-PAD.
Study limitations
Our study has several limitations. First, it is a retrospective analysis despite using a prospectively maintained database. Second, selection bias is obvious because the treatment strategy depended on the operator’s discretion, and we only enrolled data from two centers. Although we used the FUI to minimize bias from unaccounted follow-up duration, missing data and incomplete follow-up (65% of surviving patients completed 4 years of follow-up) will influence the actuarial survival estimation. Third, we did not widely apply PCDs to BTK lesions due to the reimbursement policy in Taiwan and doubtful clinical benefits; only 13% of the study participants (15% affected legs) received PCDs for BTK lesions. Finally, we lacked data on changes in NLR over time. Therefore, we could not determine whether the shift in the status of the immune-inflammatory response affected the clinical outcomes.
CONCLUSION
In conclusion, repeated PCDs for LE-PAD intervention did not affect follow-up mortality. There was no dose-response relationship between all-cause mortality and cumulative paclitaxel dosage. Predictors of mortality changed over time, and continuous variables had non-linear associations with death. Large-scale studies are warranted to examine this observation.
Acknowledgments
The authors appreciate Miss I-Ting Wu and the medical staff of the cardiac catheterization laboratory assisting the data collection and the study coordinators who participated in this study.
DECLARATION OF CONFLICT OF INTEREST
All authors disclose no financial or personal relationships with other people or organizations that could inappropriately influence (bias) the work.
FUNDING
This research was supported by grants from Research Projects TCRD-TPE-110-11, Taipei Tzu Chi Hospital, Taiwan.
REFERENCES
- 1.Katsanos K, Spiliopoulos S, Kitrou P, et al. Risk of death following application of paclitaxel-coated balloons and stents in the femoropopliteal artery of the leg: a systematic review and meta-analysis of randomized controlled trials. J Am Heart Assoc. 2018;7:e011245. doi: 10.1161/JAHA.118.011245. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.FDA executive summary: Circulatory System Devices Panel meeting. June 19 and 20, 2019;https://www.fda.gov/media/127698/download [Google Scholar]
- 3.Nordanstig J, James S, Andersson M, et al. Mortality with paclitaxel-coated devices in peripheral artery disease. N Engl J Med. 2020;383:2538–2546. doi: 10.1056/NEJMoa2005206. [DOI] [PubMed] [Google Scholar]
- 4.Secemsky EA, Shen C, Schermerhorn M, et al. Longitudinal assessment of safety of femoropopliteal endovascular treatment with paclitaxel-coated devices among medicare beneficiaries: the SAFE-PAD Study. JAMA Intern Med. 2021;181:1071–1080. doi: 10.1001/jamainternmed.2021.2738. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Behrendt CA, Sedrakyan A, Peters F, et al. Editor’s choice - long term survival after femoropopliteal artery revascularisation with paclitaxel coated devices: a propensity score matched cohort analysis. Eur J Vasc Endovasc Surg. 2020;59:587–596. doi: 10.1016/j.ejvs.2019.12.034. [DOI] [PubMed] [Google Scholar]
- 6.Schneider PA, Laird JR, Doros G, et al. Mortality not correlated with paclitaxel exposure: an independent patient-level meta-analysis. J Am Coll Cardiol. 2019;73:2550–2563. doi: 10.1016/j.jacc.2019.01.013. [DOI] [PubMed] [Google Scholar]
- 7.Gray WA, Jaff MR, Parikh SA, et al. Mortality assessment of paclitaxel-coated balloons. Circulation. 2019;140:1145–1155. doi: 10.1161/CIRCULATIONAHA.119.040518. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Dake MD, Ansel GM, Bosiers M, et al. Paclitaxel-coated Zilver PTX drug-eluting stent treatment does not result in increased long-term all-cause mortality compared to uncoated devices. Cardiovasc Intervent Radiol. 2020;43:8–19. doi: 10.1007/s00270-019-02324-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Hsieh CA, Chou SH, Chen IC, et al. Feasibility and mid-term outcomes of drug-coated balloon angioplasty between intermittent claudication and critical limb ischemia in patients with femoropopliteal disease. Acta Cardiol Sin. 2019;35:308–319. doi: 10.6515/ACS.201905_35(3).20181116A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Meng FC, Chen PL, Lee CY, et al. Real-world comparison of drug-eluting and bare-metal stents in superficial femoral artery occlusive disease with trans-atlantic intersociety consensus B lesions: a 2-year, single-institute study. Acta Cardiol Sin. 2018;34:130–136. doi: 10.6515/ACS.201803_34(2).20171126A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Jang SJ, Chou HH, Juang JJ, et al. Clinical outcomes of repetition of drug-coated balloon for femoropopliteal restenosis after drug-coated balloon treatment. Circ J. 2017;81:993–998. doi: 10.1253/circj.CJ-17-0025. [DOI] [PubMed] [Google Scholar]
- 12.Huang HL, Chou HH, Chen IC, et al. Failure mode and bimodal restenosis of the drug-coated balloon in femoropopliteal intervention. Int J Cardiol. 2018;259:170–177. doi: 10.1016/j.ijcard.2018.02.040. [DOI] [PubMed] [Google Scholar]
- 13.Bouillanne O, Morineau G, Dupont C, et al. Geriatric Nutritional Risk Index: a new index for evaluating at-risk elderly medical patients. Am J Clin Nutr. 2005;82:777–783. doi: 10.1093/ajcn/82.4.777. [DOI] [PubMed] [Google Scholar]
- 14.Tolunay H, Yasar S, Asil S, et al. Prognostic value of nutritional indexes in evaluating the 1-year results after implantation of the carillon mitral contour system. Acta Cardiol Sin. 2022;38:362–372. doi: 10.6515/ACS.202205_38(3).20211222A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Kurtul A, Gok M, Esenboga K. Prognostic nutritional index predicts contrast-associated acute kidney injury in patients with ST-segment elevation myocardial infarction. Acta Cardiol Sin. 2021;37:496–503. doi: 10.6515/ACS.202109_37(5).20210413A. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Pinato DJ, North BV, Sharma R. A novel, externally validated inflammation-based prognostic algorithm in hepatocellular carcinoma: the prognostic nutritional index (PNI). Br J Cancer. 2012;106:1439–1445. doi: 10.1038/bjc.2012.92. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Rocha-Singh KJ, Zeller T, Jaff MR. Peripheral arterial calcification: prevalence, mechanism, detection and clinical implications. Catheter Cardiovasc Interv. 2014;83:E212–E220. doi: 10.1002/ccd.25387. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Von Allmen RS, Weiss S, Tevaearai HT, et al. Completeness of follow-up determines validity of study findings: results of a prospective repeated measures cohort study. PLoS One. 2015;10:e0140817. doi: 10.1371/journal.pone.0140817. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Devlin TF. Spline functions for logistic regression modelling. 1986;In: Proceedings of the 11th Annual SAS Users Group International Conference:646–651. [Google Scholar]
- 20.Stone CJ, Koo CY. Additive splines in statistics. 1985;In: Proceedings of the American Statistical Association:45–48. [Google Scholar]
- 21.Liu KL, Tzeng IS, Chen IC, et al. Midterm mortality between single or multiple exposure to paclitaxel coated devices for the treatment of femoropopliteal artery disease. Eur J Vasc Endovasc Surg. 2022;63:521–522. doi: 10.1016/j.ejvs.2021.09.027. [DOI] [PubMed] [Google Scholar]
- 22.Ng VG, Mena C, Pietras C, et al. Local delivery of paclitaxel in the treatment of peripheral arterial disease. Eur J Clin Investig. 2015;45:333–345. doi: 10.1111/eci.12407. [DOI] [PubMed] [Google Scholar]
- 23.Margolis J, McDonald J, Heuser R, et al. Systemic nanoparticle paclitaxel (nab-paclitaxel) for in-stent restenosis I (SNAPIST-I): a first-in-human safety and dose-finding study. Clin Cardiol. 2007;30:165–170. doi: 10.1002/clc.20066. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Gongora CA, Shibuya M, Wessler JD, et al. Impact of paclitaxel dose on tissue pharmacokinetics and vascular healing: a comparative drug-coated balloon study in the familial hypercholesterolemic swine model of superficial femoral in-stent restenosis. JACC Cardiovasc Interv. 2015;8:1115–1123. doi: 10.1016/j.jcin.2015.03.020. [DOI] [PubMed] [Google Scholar]
- 25.Freyhardt P, Zeller T, Kroncke TJ, et al. Plasma levels following application of paclitaxel-coated balloon catheters in patients with stenotic or occluded femoropopliteal arteries. Rofo. 2011;183:448–455. doi: 10.1055/s-0029-1246028. [DOI] [PubMed] [Google Scholar]
- 26.Kristman V, Manno M, Côté P. Loss to follow-up in cohort studies: how much is too much? Eur J Epidemiol. 2004;19:751–760. doi: 10.1023/b:ejep.0000036568.02655.f8. [DOI] [PubMed] [Google Scholar]
- 27.Greenland S. Response and follow-up bias in cohort studies. Am J Epidemiol. 1977;106:184–187. doi: 10.1093/oxfordjournals.aje.a112451. [DOI] [PubMed] [Google Scholar]
- 28.Böhme T, Noory E, Beschorner U, et al. Evaluation of mortality following paclitaxel drug-coated balloon angioplasty of femoropopliteal lesions in the real world. JACC Cardiovasc Interv. 2020;13:2052–2061. doi: 10.1016/j.jcin.2020.04.050. [DOI] [PubMed] [Google Scholar]
- 29.Iida O, Nakamura M, Yamauchi Y, et al. 3-year outcomes of the OLIVE registry, a prospective multicenter study of patients with critical limb ischemia: a prospective, multi-center, three-year follow-up study on endovascular treatment for infra-inguinal vessel in patients with critical limb ischemia. JACC Cardiovasc Interv. 2015;8:1493–1502. doi: 10.1016/j.jcin.2015.07.005. [DOI] [PubMed] [Google Scholar]
- 30.Soga Y, Iida O, Takahara M, et al. Two-year life expectancy in patients with critical limb ischemia. JACC Cardiovasc Interv. 2014;7:1444–1449. doi: 10.1016/j.jcin.2014.06.018. [DOI] [PubMed] [Google Scholar]
- 31.Brizuela Sanz JA, González Fajardo JA, Taylor JH, et al. Design of a new risk score in critical limb ischaemia: the ERICVA model. Eur J Vasc Endovasc Surg. 2016;51:90–99. doi: 10.1016/j.ejvs.2015.09.025. [DOI] [PubMed] [Google Scholar]
- 32.Spark JI, Sarveswaran J, Blest N, et al. An elevated neutrophil-lymphocyte ratio independently predicts mortality in chronic critical limb ischemia. J Vasc Surg. 2010;52:632–636. doi: 10.1016/j.jvs.2010.03.067. [DOI] [PubMed] [Google Scholar]
- 33.Jhang JY, Tzeng IS, Chou HH, et al. Association rule mining and prognostic stratification of 2-year longevity in octogenarians undergoing endovascular therapy for lower extremity arterial disease: observational cohort study. J Med Internet Res. 2020;22:e17487. doi: 10.2196/17487. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Haumer M, Amighi J, Exner M, et al. Association of neutrophils and future cardiovascular events in patients with peripheral artery disease. J Vasc Surg. 2005;41:610–617. doi: 10.1016/j.jvs.2005.01.013. [DOI] [PubMed] [Google Scholar]
- 35.Gauthier J, Wu QV, Gooley TA. Cubic splines to model relationships between continuous variables and outcomes: a guide for clinicians. Bone Marrow Transplant. 2020;55:675–680. doi: 10.1038/s41409-019-0679-x. [DOI] [PubMed] [Google Scholar]
- 36.Keller K, Hobohm L, Geyer M, et al. Obesity paradox in peripheral artery disease. Clin Nutr. 2019;38:2269–2276. doi: 10.1016/j.clnu.2018.09.031. [DOI] [PubMed] [Google Scholar]
- 37.Huang HL, Tzeng IS, Chou HH, et al. Contemporary cardiovascular outcomes in Taiwanese patients undergoing endovascular therapy for symptomatic lower extremity peripheral arterial disease. J Formos Med Assoc. 2020;119:1052–1060. doi: 10.1016/j.jfma.2019.10.011. [DOI] [PubMed] [Google Scholar]
- 38.Pecoits-Filho R, Lindholm B, Stenvinkel P. The malnutrition, inflammation, and atherosclerosis (MIA) syndrome -- the heart of the matter. Nephrol Dial Transplant. 2002;17:28–31. doi: 10.1093/ndt/17.suppl_11.28. [DOI] [PubMed] [Google Scholar]