Active smoking is associated with higher rates of incomplete wound healing after endovascular treatment of critical limb ischemia

Abstract

The association between active smoking and wound healing in critical limb ischemia (CLI) is unknown. Our objective was to examine in a retrospective cohort study whether active smoking is associated with higher incomplete wound healing rates in patients with CLI undergoing endovascular interventions. Smoking status was assessed at the time of the intervention, comparing active to no active smoking, and also during follow-up visits at 6 and 9 months. Cox regression analysis was conducted to compare the incomplete wound healing rates of the two groups during follow-up. A total of 264 patients (active smokers: n = 41) were included. Active smoking was associated with higher rates of incomplete wound healing in the 6-month univariate Cox regression analysis (hazard ratio (HR) for incomplete wound healing: 4.54; 95% CI: 1.41–14.28; p = 0.012). The 6-month Kaplan–Meier (KM) estimates for incomplete wound healing were 91.1% for the active smoking group versus 66% for the non-current smoking group. Active smoking was also associated with higher rates of incomplete wound healing in the 9-month univariable (HR for incomplete wound healing: 2.32; 95% CI: 1.11–4.76; p = 0.026) and multivariable analysis (HR for incomplete wound healing: 9.09; 95% CI: 1.06–100.0; p = 0.044). The 9-month KM estimates for incomplete wound healing were 75% in the active smoking group versus 54% in the non-active smoking group. In conclusion, active smoking status at the time of intervention in patients with CLI is associated with higher rates of incomplete wound healing during both 6- and 9-month follow-up.

Introduction

Peripheral artery disease (PAD) affects between 5% and 10% of the adult population (> 40 years old) in the United States.¹ Although smoking prevalence has decreased by more than 50% during the last half-century, close to 20% of the adult population in the United States still actively smokes. Cigarette smoking is the most important preventable risk factor for the development and progression of PAD.^2–6 Smokers have more than double the risk for PAD compared to non-smokers,^2,7–9 while between 50% and 90% of the PAD population are either current or past smokers.^2,10–14 Previous studies have shown that smoking cessation – when achieved – associates with improved limb and cardiovascular outcomes.^15–20

Notwithstanding the plethora of data in the coronary artery disease (CAD) population, for PAD and critical limb ischemia (CLI) data are scarce. A Swedish study of 343 patients with intermittent claudication showed that smoking cessation was associated with a decreased risk for development of CLI.¹⁶ Other studies have shown that active smoking among the CLI population is associated with worse limb outcomes after open surgery.^19,21,22 Smoking cessation in patients with CLI decreases mortality compared to continuation of smoking – to a much higher degree compared to cessation in the general PAD population.¹⁵

Several mechanisms have been reported that explain the association between smoking and poor wound healing. Smoking is associated with lower levels of subcutaneous collagen²³ and significantly reduces cutaneous microvascular function, fostering an environment conducive to the ischemic ulcerations associated with PAD.^24–29 Even after optimal endovascular revascularization, there are still a significant percentage of patients with CLI (40–50%) with either delayed healing or worsening of tissue loss.^25,30,31 Up to 70% of all the lower extremity major amputations are secondary to non-healing wounds in patients with CLI.³² With this study, we assessed the relationship of active smoking status to wound healing for patients with CLI undergoing endovascular revascularization.

Methods

The protocol for this study was approved by The University of California, Davis Medical Center Institutional Review Board, with a waiver of informed consent.

This was a single-center retrospective study including all patients with CLI who underwent lower extremity endovascular intervention from June 2006 to June 2014. Patients who were active smokers at the time of the endovascular intervention were compared to patients who were not active smokers.

Definitions and study outcomes

Patients who had isolated diagnostic angiography, those with presentation suggestive of acute limb ischemia, and those with intermittent claudication were excluded. CLI was classified as Rutherford category 4 to 6 (ischemic rest pain, minor tissue loss, or major tissue loss, respectively). Data collection from the electronic medical records and angiograms was performed by experienced abstractors and verified by board certified endovascular interventionalists. Smoking status was assessed retrospectively based on patient self-report in the clinic appointments, at baseline, and after 6 and 9 months of follow-up. If a patient denied smoking at the last clinic appointment before the endovascular intervention, the patient was registered as a non-active smoker (without a specific cut-off in the time interval from smoking status documentation until the time of the procedure). Other smoking-related variables that were recorded included number of cigarettes smoked per day at the time of the intervention, pack-years at the time of the intervention, and willingness to quit smoking (which was collected from the chart based on the documentation). Since this was a retrospective chart review, data on wound healing were collected by retrospectively looking at the chart documentation and abstracting information regarding the status of the wound (worse, better, similar, complete or incomplete wound healing). No pictures were available for most of the cases. The primary outcomes of the study were incomplete wound healing after 6 and 9 months of follow-up. Secondary outcomes included amputation, bypass, and major adverse limb events (MALE) rates defined as any of the following: major lower extremity limb amputation above the level of the ankle joint, thrombolysis for acute limb ischemia, thrombectomy or bypass surgery.

Statistical analysis

Continuous variables were described with mean ± SD and compared with the Wilcoxon rank sum test, while categorical variables were described with absolute and relative frequencies and compared using a chi-squared or Fisher exact test. The association between baseline demographic, clinical or angiographic characteristics included smoking-related variables and outcomes of interest was examined with a Cox regression analysis. Patients with missing data regarding their smoking habits or the outcomes were not included in the Cox regression analysis. Kaplan–Meier (KM) survival analysis was performed and the KM estimates for 6 and 9 months were calculated. Variables with a significant (p < 0.05) association in the univariable analysis or variables that were considered to have a clinically important association with wound healing were included in the multivariable model. A separate sensitivity analysis was performed among patients who were actively smoking at the time of the procedure. Patients who were still smokers after 6 and 9 months of follow-up were compared after each endpoint to their peers who had stopped smoking. The results are presented as hazard ratios (HR) and 95% CI for the Cox regression analysis. For all tests, p < 0.05 was considered significant. All analyses were performed using Stata software (release 14.1; StataCorp LP, College Station, TX, USA).

Results

In this cohort, 264 patients (active smokers: n = 41; non-active smokers: n = 223) with 554 lesions with intervention between 2006 and 2014 were included in the analysis. Baseline demographic, clinical, and laboratory characteristics are presented in Table 1. This was a largely male (66%), overweight (mean body mass index (BMI): 27.38 ± 6.30) cohort, highly prevalent with diabetes mellitus (DM) (70%), hypertension (HTN) (85%), and CAD (54%). Non-active smokers had a higher mean age (69.9 vs 65.8; p = 0.041). Active smokers were more likely to have chronic obstructive pulmonary disease (COPD) (27% vs 8%; p = 0.001) and to receive multivessel intervention compared to non-active smokers (95% vs 75%; p = 0.004). Stenting rates were equal between the two groups (14% vs 17%). Angiographic and procedural characteristics are presented in Table 2

Table 1.

Baseline characteristics of the included patients (patient level analysis).

Variables	Total, % (n = 264)	Active smokers, % (n = 41)	Non-active smokers, % (n = 223)	p-value
Age, years	69.2 ± 11.5	65.8 ± 10.8	69.9 ± 11.5	0.041
Female sex	34 (90/264)	34 (14/41)	34 (76/223)	0.993
Non-white	57 (133/231)	47 (16/34)	59 (117/197)	0.179
BMI, kg/m2	27.38 ± 6.30	26.16 ± 5.01	27.61 ± 6.49	0.286
Diabetes	70 (181/259)	67 (27/40)	70 (154/219)	0.721
HbA1c	7.8 ± 2.26	8.6 ± 2.80	7.7 ± 2.17	0.388
Hypertension	85 (226/264)	93 (38/41)	84 (188/223)	0.160
SBP, mmHg	134 ± 21.10	138 ± 20.84	133 ± 21.11	0.228
Carotid stenosis	10 (26/249)	15 (6/40)	9 (20/209)	0.303
Coronary artery disease	54 (141/262)	58 (24/41)	53 (117/221)	0.509
Congestive heart failure	34 (90/262)	29 (12/41)	35 (78/221)	0.456
Aspirin use	75 (197/264)	78 (32/41)	74 (165/223)	0.583
Clopidogrel use	43 (114/264)	32 (13/41)	45 (101/223)	0.107
Dyslipidemia	67 (177/262)	68 (28/41)	67 (149/221)	0.913
Triglycerides, mg/dL	127 ± 62.00	141 ± 80.39	125 ± 59.29	0.477
LDL, mg/dL	81.23 ± 37.90	80 ± 27.26	81.40 ± 35.95	0.805
HDL, mg/dL	36.96 ± 13.11	36.30 ± 12.05	37.05 ± 13.31	0.854
Statin use	66 (162/244)	66 (27/41)	66 (135/203)	0.936
ESRD	25 (65/263)	12 (5/40)	27 (60/223)	0.052
Cancer	7 (19/264)	12 (5/40)	6 (14/223)	0.178
COPD	11 (30/264)	27 (11/41)	8 (19/223)	0.001
Stroke history	21 (55/264)	19 (8/41)	21 (47/223)	0.821
Myocardial infarction history	24 (56/228)	25 (7/28)	24 (49/200)	0.954
Procedure type				0.568
Elective procedure	28 (75/264)	34 (14/41)	27 (61/223)
Urgent procedure	67 (177/264)	63 (26/41)	68 (151/223)
Emergent procedure	5 (12/264)	2 (1/41)	5 (11/223)
Multivessel intervention	78 (205/263)	95 (39/41)	75 (166/222)	0.004
Toe pressure, mmHg	40.27 ± 28.73	40.64 ± 32.38	40.22 ± 28.37	0.963
Preprocedural ABI	0.67 ± 0.29	0.64 ± 0.31	0.68 ± 0.29	0.930
Run-off	1.22 ± 0.79	1.52 ± 0.90	1.17 ± 0.76	0.118
Rutherford category at baseline				0.785
Rutherford 4 at baseline	14 (31/225)	18 (5/28)	13 (26/197)
Rutherford 5 at baseline	72 (163/225)	68 (19/28)	73 (144/197)
Rutherford 6 at baseline	14 (31/225)	14 (4/28)	14 (27/197)

Categorical variables are expressed as absolute and relative frequencies. Continuous variables are expressed as mean and SD.

ABI, ankle–brachial index; BMI, body mass index; COPD, chronic obstructive pulmonary disease; ESRD, end-stage renal disease; HDL, high-density lipoprotein, LDL, low-density lipoprotein; SBP, systolic blood pressure.

Table 2.

Angiographic and procedural characteristics (lesion level).

Variables	Total (n = 554)	Active smokers (n = 91)	Non-active smokers (n = 463)	p-value
Chronic total occlusions	41 (170/419)	35 (26/74)	42 (144/345)	0.294
TASC C/D	52 (115/223)	50 (23/46)	52 (92/177)	0.811
Moderate/severe calcification	30 (126/419)	32 (24/74)	30 (102/345)	0.626
Thrombus	5 (19/421)	3 (2/74)	5 (17/347)	0.409
Thrombolysis	3 (11/435)	1 (1/78)	3 (10/357)	0.439
Stent placed	16 (72/437)	14 (11/78)	17 (61/359)	0.734
Maximal stent diameter (mm)	3.52 ± 1.36	3.81 ± 1.39	3.46 ± 1.35	0.029
Flouro time (min)	37.31 ± 25.47	39.38 ± 21.51	36.94 ± 26.13	0.148
Contrast (mL)	180.81 ± 92.68	207.70 ± 112.65	175.71 ± 87.62	0.021
Atherectomy	18 (81/437)	20 (16/78)	18 (65/359)	0.620
Location of the lesion				0.317
Popliteal	2 (12/554)	3 (3/91)	2 (9/463)
Tibial	50 (278/554)	41 (37/91)	52 (241/463)
TP trunk	21 (119/554)	27 (25/91)	20 (94/463)
Peroneal	21 (114/554)	23 (21/91)	20 (93/463)
Dorsalis pedis	5 (26/554)	5 (5/91)	5 (21/463)
Lateral plantar	1 (5/554)	0 (0/91)	1 (5/463)
Procedural success	96 (529/553)	97 (88/91)	95 (441/462)	0.593
Limb loss	12 (63/547)	9 (8/90)	12 (55/457)	0.393
Restenosis	11 (49/437)	10 (8/78)	11 (41/359)	0.768
Post-procedural aspirin use	96 (197/206)	100 (31/31)	95 (166/175)	0.197
Post-procedural clopidogrel use	80 (164/206)	81 (25/31)	79 (139/175)	0.877
ABI 30 days	0.70 ± 0.34	0.78 ± 0.25	0.67 ± 0.37	0.150
ABI 1 year	0.46 ± 0.44	0.60 ± 0.37	0.40 ± 0.45	0.078

Categorical variables are expressed as absolute and relative frequencies. Continuous variables are expressed as mean and SD.

ABI, ankle–brachial index; TASC, TransAtlantic Inter-Society Consensus.

The time to complete wound healing was longer in the active smoking group (269.7 ± 187.5 days vs 208.3 ± 115.7; p = 0.009). There was no difference in the 9-month rates of minor amputations between the two groups (26.5% in the active smoking group vs 24.6% in the non-active smoking group; p = 0.81). Limb loss occurred in 9% and 12% of active smokers and non-active smokers, respectively, during a mean follow-up of 10.5 months, without a statistically significant difference between the two groups. The time to limb loss was similar between active smokers and non-active smokers (93.3 ± 125.0 vs 100.4 ± 134.6 days; p = 0.555). Female gender was associated with lower rates of 12-month lack of wound healing in the multivariate analysis (HR: 0.43; 95% CI: 0.19–0.97; p = 0.041). There was a high loss to follow-up rate in both groups, but without significant differences. For the active smoking group, only 20 out of the 41 patients completed the 9-month follow-up (loss to follow-up rate 51%) vs 116 patients who completed follow-up out of 223 total patients in the non-active smoking group (loss to follow-up rate 48%).

Six-month follow-up

Active smoking (vs non-active smoking) was associated with significantly higher rates of incomplete wound healing in the univariable analysis (HR: 4.54; 95% CI: 1.41–14.28; p = 0.012). No difference was found after adjusting for covariates. The univariable and multivariable Cox regression analyses for incomplete wound healing at 6 months are presented in Table 3. The 6-month KM estimates for incomplete wound healing were 91.1% for the active smoking group versus 66% for the non-active smoking group (Log-rank test p < 0.001). Patients who were still smoking after the intervention did not have a higher risk for incomplete wound healing compared to their peers who had stopped (HR for incomplete wound healing: 1.49; 95% CI: 0.14–16.6; p = 0.74).

Table 3.

Predictors of lack of wound healing (Cox regression analysis).

Variables	6-month lack of wound healing				9-month lack of wound healing				12-month lack of wound healing
	Unadjusted		Adjusted		Unadjusted		Adjusted		Unadjusted		Adjusted
	HR (Cl)	p-value	HR (Cl)	p-value	HR (Cl)	p-value	HR (Cl)	p-value	HR (Cl)	p-value	HR (Cl)	p-value
Current vs no current smokers at baseline	4.54 (1.41–14.28)	0.012	0.00 (0–0)	1.00	2.32 (1.1 1–4.76)	0.026	9.09 (1.06–100)	0.044	1.29 (0.77–2.22)	0.336	3.12 (0.79–12.5)	0.105
Pack-years at baseline	1.01 (1.00–1.02)	0.361	1.00 (0.98–1.02)	0.744	1.00 (1.00–1.01)	0.916	1.00 (0.99–1.02)	0.637	1.00 (1.00–1.01)	0.702	1.01 (0.99–1.02)	0.946
Female vs male	1.54 (0.89–2.63)	0.125	1.16 (0.35–3.85)	0.806	0.66 (0.44–1.00)	0.05	0.44 (0.18–1.09)	0.077	0.75 (0.51–1.09)	0.124	0.43 (0.19–0.97)	0.041
Non-white vs white	0.79 (0.46–1.37)	0.397	0.65 (0.21–2.00)	0.449	1.00 (0.64–1.59)	0.993	0.83 (0.35–1.96)	0.655	1.20 (0.81–1.78)	0.358	0.89 (0.42–1.92)	0.768
Diabetes mellitus	1.02 (0.56–1.85)	0.963	0.61 (0.07–5.26)	0.647	1.1 1 (0.67–1.85)	0.691	0.48 (0.09–2.44)	0.369	1.09 (0.70–1.72)	0.692	0.29 (0.06–1.41)	0.123
Coronary artery disease	0.80 (0.49–1.29)	0.362	1.39 (0.55–3.57)	0.496	0.86 (0.57–1.29)	0.462	1.45 (0.65–3.22)	0.371	1.03 (0.72–1.49)	0.871	1.37 (0.66–2.85)	0.412
Statin	0.84 (0.50–1.41)	0.495	0.99 (0.33–3.03)	0.999	0.87 (0.56–1.35)	0.531	0.81 (0.32–2.04)	0.647	0.80 (0.53–1.19)	0.265	0.83 (0.37–1.89)	0.649
Run-off	0.91 (0.59–1.41)	0.655	3.12 (0.44–1.28)	0.296	0.94 (0.65–1.39)	0.753	0.83 (0.53–1.31)	0.408	0.96 (0.69–1.33)	0.779	0.83 (0.55–1.26)	0.379

HR, hazard ratio.

Nine-month follow-up

Active smoking (vs non-active smoking) was associated with a significant increase in the rates of incomplete wound healing in univariable (HR for incomplete wound healing: 2.32; 95% CI: 1.11–4.76; p = 0.026) and multivariable analysis (HR for incomplete wound healing: 9.09; 95% CI: 1.06–100.0; p = 0.044). The univariable and multivariable Cox regression analysis is presented in Table 3. The 9-month KM estimates for incomplete wound healing were 74.8% in the active smoking group versus 54.0% in the non-active smoking group (log-rank p < 0.001). The KM survival analysis is presented in Figure 1. Patients who were still smoking at 9 months after the intervention had a trend towards higher rates of incomplete wound healing rates, although the difference did not reach levels of statistical significance (HR for incomplete wound healing: 4.54; 95% CI: 0.90–25.0; p = 0.067).

Figure 1. Kaplan–Meier survival analysis curves for 9-month lack of wound healing.

The 9-month lack of wound healing was not higher in the active smoking group in the Cox regression analysis, even if there was a trend towards significance (HR: 4.54; 95% CI: 0.90–25.0; p = 0.067). The Kaplan–Meier estimates for lack of wound healing were higher among current smokers compared to not current smokers (74.8% vs 54.0%).

HR, hazard ratio.

There was no difference between the two groups in the 9-month surgical bypass rates (HR: 0.64; 95% CI: 0.19–2.12; p = 0.465). Figure 2 presents the KM survival curves for 9-month limb loss. There was no difference between the two groups (HR: 0.37; 95% CI: 0.11–1.20; p = 0.098). Figure 3 presents the KM curves for wound healing, freedom from limb loss and amputation-free survival for the whole cohort. In the analysis among patients who were smokers at the time of the procedure, patients who were still smoking at 9 months after the intervention were more likely to experience incomplete wound healing compared to patients who stopped smoking after the intervention (HR: 3.12; 95% CI: 1.12–9.09; p = 0.029) (Figure 4).

Figure 2. Kaplan–Meier survival analysis curves for 9-month freedom from limb loss.

The 9-month limb loss rates were not higher in the active smoking group in the Cox regression analysis (HR: 0.37; 95% CI: 0.11–1.20; p = 0.098). Kaplan–Meier estimates for lack of wound healing were not different among the two groups (95.4% vs 88.9%).

HR, hazard ratio.

Figure 3. Kaplan–Meier survival analysis curves for the whole population of the study.

The 9-month rates for complete wound healing, amputation-free survival, and freedom of limb loss among the whole cohort.

9m-KM, 9-month Kaplan–Meier.

Figure 4. The 9-month lack of wound healing rates among patients who were smoking at baseline.

The 9-month rates for lack of wound healing among patients who were smokers at baseline after the endovascular intervention. Patients who subsequently did not stop smoking were less likely to achieve wound healing in the follow-up (HR for lack of wound healing: 3.12; 95% CI: 1.12–9.09; p = 0.029). The 9-month Kaplan–Meier rates for lack of wound healing were higher among patients who continued smoking versus patients who stopped smoking (88.9% vs 53.9%).

HR, hazard ratio.

Discussion

This was a single-center observational study of 264 patients with CLI who underwent endovascular revascularization procedures from 2006 to 2014 for 554 infrapopliteal lesions. The two groups had similar baseline conditions with notable exceptions including ESRD (higher among non-active smokers) and COPD (higher among smokers). Active smoking at the time of the intervention was associated with higher rates of incomplete wound healing after 6 and 9 months of follow-up. Patients who continued to be active smokers in the follow-up period were less likely to achieve wound healing.

Smoking cessation is strongly recommended in patients with PAD according to both the recent American College of Cardiology (ACC) / American Heart Association (AHA) guidelines (Class I, Level of Evidence: A)¹² and European Society of Cardiology.³³ A previous study from our group showed that smoking cessation in PAD is associated with decreased mortality rates and improved amputation-free survival during 5 years of follow-up.¹⁵ These findings were confirmed in a subgroup analysis among patients with CLI.¹⁵ The cardiovascular risk reduction and reduced incidence of CLI becomes apparent 5 years after smoking cessation.^16,34–36 In our study, 15% of the patients were actively smoking at the time of the intervention. These rates are lower compared to other reports of the PAD literature, which usually show a smoking rate in elective PAD interventions of around 40%.³⁷ However, our cohort included only patients with CLI, commonly treated for PAD for years (and thus, were instructed to discontinue smoking in the past). These rates are not so different from the TAMARIS trial (active smoking rates 15–18%), but less than the Circulase CLI trial (25%) and the EUCLID CLI subgroup analysis (24%).^38–40

Smoking in CLI and PAD populations is also known to be associated with decreased graft patency and acceleration of PAD progression to CLI.^16,19,21,22 However, it is unclear if this is related to baseline smoking at the time of the intervention/surgery or to persistent smoking during follow-up.^41–43 As a consequence, smokers are less likely to undergo bypass compared to non-smokers.⁴⁴ Contrast volume was higher among smokers secondary to the higher rates of multivessel interventions in the active smoking population, while procedural and access-related interventions were more common in the active smoking group. Smoking is known in the surgical literature to be associated with worse procedural, short- and long-term surgical outcomes.^45–47 Earlier studies have shown that smoking intensity may be even more important than smoking status for long-term outcomes such as amputations, MALE, and major adverse cardiovascular events (MACE) after open revascularization for PAD.³⁷ Contrary to intermittent claudication, where it is reasonable to postpone elective interventions until patients achieve smoking cessation, there is no such option in CLI.^48,49 In CLI, interventions are performed more urgently to ensure limb salvage.⁵⁰

Our study is not the first that investigates the effect of smoking in CLI, although it is the first to study smoking’s effect on wound healing in patients with CLI after endovascular revascularizations. It is known that lower extremity wound healing is not only related to the size, severity, depth, and treatment selection, but also to modifiable lifestyle and history-related factors such as diabetes status and smoking.^51–53 A recent meta-analysis among patients with diabetic ulcers, showed a significant association between smoking and delayed healing (healing rates in smokers were 62.1% vs 71.5% in non-smokers).⁵⁴ We showed that active smoking is associated with higher 6-month (91.1% for the active smoking group vs 66% in the non-active smoking group) and 9-month (74.8% in the active smoking group vs 54.0% in the non-active smoking group) incomplete wound healing rates. We also showed that smoking continuation after endovascular revascularization was negatively associated with wound healing, especially after a longer follow-up period (the difference became significant close to the 9-month follow-up, while earlier there was no significant difference).

Smoking may delay wound healing in CLI and diabetic foot ulceration in several ways. First, smoking increases oxidative stress and apoptosis of vascular endothelial cells, increasing vascular dysfunction.⁵⁵ Second, nicotine is a direct vasoconstrictor and can decrease perfusion to the lower extremity tissues, further impairing in this way the process of wound healing. Third, smoking increases carboxyhemoglobin levels, decreasing oxygen transport through hemoglobin to the injured wound tissues.⁵⁶ Fourth, smoking is associated with an increased risk for uncontrolled diabetes, which further delays wound healing in these patients.^15,57

The results of our study extend prior reports in PAD or diabetic foot ulcer populations.¹² Smoking cessation programs should be fully incorporated in multidisciplinary CLI teams and combined with aggressive medical therapy, revascularization, exercise and nutritional support, orthotic devices, and infection control strategies.^50,58,59 However, even if smoking cessation is recommended for all patients with CLI and considered to be the most cost-effective intervention for PAD, reported quit rates and cessation times are suboptimal across the literature.^15,60–62 This is related to a number of different factors, including the lack of evidence-based treatment offered to the patient, limited patient education about the benefits of smoking cessation, difficulties in assessing patients’ smoking addiction severity at baseline, and future compliance with therapeutic interventions.^14,60,63 A preliminary analysis of the Vascular Physician Offer and Report (VAPOR) trial showed that even when physicians try to convince patients to stop smoking through counseling and nicotine replacement therapy prescription, there is no difference compared to the control group.⁶⁴

Limitations

This was a real-world study and thus should be interpreted in the context of observational research and its limitations. Retrospective chart reviews have the disadvantage that questions are asked in a non-uniform way across different patients and providers. Similarly, wound assessment accuracy and uniformity would be improved and standardized in a prospectively designed study and variability in wound care among different patients may have affected our results. We followed the smoking habits of the enrolled patients over time based on clinic appointments but some of the patients may have crossed over from the one group to the other in the follow-up, which we were unable to adjust for with the current analysis. Additionally, the loss to follow-up rate was very high, which can limit the certainty in our findings, while additional events – which are expected in CLI patients, such as amputation and death – contributed to a high number of censored patients. Not accounting for possible overfitting of our model is another limitation. Some of the variables included in the multivariable analysis did not have a significant univariate association, but they were deemed to be clinically significant. Finally, yet importantly, prior data have shown that smoking tends to be under-reported by patients.^37,60 A significant number of patients who deny that they smoke or state that they have quitted smoking, still smoke when biochemical data are tested. Similarly, there is under-reporting in the number of consumed cigarettes. Thus, it is likely that a significant number of our non-active smoking patients were still actively smoking and this could have affected our results.

Conclusion

This is the first study that examines the association between smoking status and wound healing in patients with CLI undergoing endovascular revascularization. We found that active smoking status was associated with increased rates of lack of wound healing after 6 and 9 months of follow-up. Future studies should seek the ideal intervention in order to decrease the smoking rates and try to explain the pathophysiologic mechanism behind this relationship.