Lifelong Limb Preservation: A Patient-Centered Description of Lower Extremity Arterial Reconstruction Outcomes

Abstract

Background

Life expectancy is short for patients with critical limb ischemia (CLI), many of whom may fear amputation more than death. In light of the reduced life expectancy of these patients, the traditional 5-year freedom from amputation (FFA) statistic may not accurately address their concern. To better answer the question “Will I ever lose my leg?,” we developed a more relevant patient-centered calculation of major amputation risk over a patient’s remaining lifetime.

Methods

We identified all limbs undergoing first time intervention for CLI in a large institutional database from 2005–2013. We calculated the traditional metrics of amputation-free survival (AFS, where failure is death or amputation) and freedom from amputation (FFA, where failure is amputation but deaths are censored and removed from further analysis). Additionally, we propose a new term: lifelong limb preservation (LLP). LLP defines amputation as failure, but deaths are not censored and therefore reflect that LLP has been achieved. All deaths prior to 30-days were considered a failure in all three metrics, reflecting the risk of surgery.

Results

1,006 limbs were identified as having first time intervention for CLI (22% rest pain, 45% ulcer, 27% gangrene; 46% treated by angioplasty+/−stenting, 54% bypass). Using life table analysis, seven-year AFS was 14% (561 events), FFA was 78% (123 events) and LLP was 86% (123 events). LLP was similar between patients undergoing angioplasty/stenting and bypass (seven-year: 86% and 85% respectively). For patients undergoing intervention for rest pain, seven-year rates were: AFS 14%, FFA 84% and LLP 92%. For those undergoing treatment for ulcer, seven-year rates were: AFS 14%, FFA 77% and LLP 86%. Finally, in those with gangrene, rates were: AFS 10%, FFA 67% and LLP 79%. Using LLP, a patient presenting with an ulcer can be told that while we cannot guarantee how long they will live, with revascularization there is approximately an 86% chance they will not lose their leg.

Conclusion

The results above show that the durability of our limb preservation efforts often exceeds the life expectancy of our patients. Using lifelong limb preservation (LLP) as an outcomes assessment provides a more accurate and patient-centered answer to the question, “If I have this procedure, will I ever lose my leg?”

Introduction

Peripheral arterial disease (PAD) affects more than 8.5 million patients over the age of 40 in the United States,^{1, 2} and approximately 1% of all PAD patients will go on to develop critical limb ischemia (CLI).^{3, 4} The diagnosis of CLI has a poor prognosis, with as many as half of all patients ultimately undergoing amputation.^{5, 6} Additionally, up to 20% of CLI patients will die within six months of diagnosis and 5-year survival has been reported to be as low as 50%.^7–12

Previous studies have described the grim prognosis for those with CLI and work has been done to try to stratify patient risk. The Society for Vascular Surgery (SVS) developed a classification system for CLI patients, which classified amputation risk based on three major influences: wound, ischemia and foot infection (WIfi). The SVS WIfi classification system allowed for more precise organization of patients based on their disease burden.¹³ Additionally, work by Conte et al. designated objective performance goals for the treatment of CLI.¹⁴ They defined two important endpoints: amputation-free survival and major adverse limb event (defined as amputation or major reintervention). This provided an additional tool for pre-operative risk assessment in CLI patients regarding the two major outcomes: amputation and death.

Though the current literature provides a framework for categorizing the prognosis and risk of amputation for patients, the estimate given is reflective of those who survive to the end of the analysis. This analysis certainly has its benefits, but it fails to account for those patients who died without ever undergoing an amputation. For many patients, they are aware of their limited lifespan and their priority is to remain amputation-free for their remaining time. Consequently, providing them with a statistic from only the surviving patients may not sufficiently address their concerns.

Our study seeks to improve the current communication with CLI patients. We sought to develop a new risk metric, designed to give patients a more realistic assessment of their prognosis for undergoing amputation. In a patient population where as many as half will not survive five years, communication is essential in helping to guide patients in their treatment decisions. By not censoring out those patients who died over the course of the analysis, we hope to better answer the question, “Will I ever lose my leg?”

Methods

Database

All limbs undergoing first time intervention for critical limb ischemia (CLI) at our institution from 2005–2013 were utilized for this retrospective analysis. Individual medical records were reviewed for all patients within the database. Patients who underwent either lower extremity bypass or percutaneous transluminal angioplasty with or without stenting (PTA/S) were included. There were no hybrid cases in this patient population. Additionally, as this database was designed to compare bypass and PTA/S, the majority of endarterectomy-only cases performed at our institution were not included in the initial query. Therefore, the remaining patients who underwent endarterectomy were subsequently excluded from this analysis. Amputations were counted on a per-limb basis and death on a per-patient basis. There were 1,006 limbs treated for CLI in our series in 902 patients. The Beth Israel Deaconess Medical Center Institutional Review Board approved this study, and patient consent was waived due to the de-identified nature of this dataset.

Variables

Patient demographics, comorbid conditions and operative details were identified for all patients. Indication for procedure included rest pain, ulcer or gangrene, and overall survival was calculated within each of these populations. Our primary endpoints included both of the previously established outcomes in this population, amputation-free survival (AFS) and freedom from amputation (FFA), as well as a novel endpoint, lifelong limb preservation (LLP). Major amputation was defined as below knee trans-tibial or above knee amputation. For amputation-free survival, a terminal event was defined as death or major amputation within the follow-up period. For freedom from amputation, a terminal event was defined as a major amputation with those patients who died without amputation being censored (removed from further analysis without being considered having had an event) at time of death. Lifelong limb preservation (LLP), our novel endpoint, defined a terminal event as a major amputation, however patients who died without an amputation were not censored. Therefore, patients who died with their limb intact within the seven-year interval were considered a success, as they had irrevocably achieved lifelong limb preservation. All deaths prior to 30-days were considered a failure in all three metrics, reflecting the risk of surgery. Seven-year endpoints were selected to take advantage of the extended follow-up in our series of patients and to highlight the differences between these three outcomes.

Statistical Analysis

Significant differences in freedom from each respective event, AFS, FFA and LLP, were analyzed using Kaplan-Meier analysis with the log-rank test. Life-table analysis was used to calculate cumulative rates for each of the three endpoints. Missing data were less than three percent for all variables, except for follow-up time, which was nine percent. Statistical analysis was conducted using SPSS Statistics version 23 (IBM Inc., Chicago, IL).

Results

A total of 1,006 limbs in 902 patients underwent first time interventions for CLI: 46% treated by angioplasty+/‒stenting (N = 460) and 54% with bypass (N= 546). The mean follow-up time was 606 days.

Demographics and Comorbidities

Patient baseline demographics were analyzed for the entire cohort (Table I). The mean age was 72 years old, 58% were male and 80% were white. Twenty-one percent of patients were current smokers. Additionally, coronary artery disease was present in 52%, hypertension in 84%, diabetes mellitus in 74%, chronic renal insufficiency (baseline creatinine > 2.0 mg/dl) in 34%, hemodialysis in 21%, chronic obstructive pulmonary disorder in 12% and congestive heart failure in 32%. An additional 27% of patients had prior myocardial infarction and 25% who had prior coronary artery bypass/percutaneous coronary intervention. Indication for procedure was distributed as follows: rest pain (22%), ulcer (45%) and gangrene (27%). Autogenous conduit was utilized in 87% of bypass patients.

Table 1.

Patient Demographics and Comorbidities

Variable (N = 902)	Percent (%)	Number (N)
Gender (male)	58	584
Race (white)	80	805
Coronary artery disease	52	522
Hypertension	84	843
Diabetes mellitus	74	744
Chronic renal insufficiency	34	338
Hemodialysis	21	211
Hyperlipidemia	58	578
Chronic obstructive pulmonary disease	12	120
Congestive heart failure	32	317
Myocardial infarction	27	273
Coronary artery bypass grafting	25	254
Current smoker	21	213
Mortality	59	596
Indication for Procedure
Rest pain	22	220
Ulcer	45	457
Gangrene	27	275

Outcomes

Using life table analysis, seven-year freedom-from-event rates were calculated for each of our three metrics. Amputation-free survival at seven years was 14% (561 events). Freedom from amputation at seven-years was 78% (123 events) and lifelong limb preservation was 86% (123 events) over the same time interval (Table II and Figure 1). Kaplan-Meier survival curve was performed for the entire cohort (Figure 2). One-year mortality was 16% (N = 164). Thirty-day mortality was 3% (N = 32). Additional perioperative complications for this cohort of patients are detailed in a recent paper from our institution, which utilizes the same database.¹⁵

Table 2.

Life-Table Summary, by Year

	Year	1	2	3	4	5	6	7	Cumulative
Outcome
AFS
Event -Free (%)		74	60	48	37	29	20	14	14
Number of Events (N)		230	107	76	60	40	31	17	561
FFA
Event -Free (%)		90	88	84	84	83	79	78	78
Number of Events (N)		85	15	15	1	2	4	1	123
LLP
Event -Free (%)		91	89	87	87	87	86	86	86
Number of Events (N)		85	15	15	1	2	4	1	123

Figure 1.

Survival Curves for FFA and LLP

Figure 2.

Overall Survival Curve Following Revascularization

Indication for Intervention and Procedure Event Outcomes

Seven-year freedom-from-event rates were further evaluated within the different indication groups, as shown in Table III. For rest pain, freedom-from-event rates were: AFS 14%, FFA 84% and LLP 92%. Among patients with ulcers, they were: AFS 14%, FFA 77% and LLP 86%; and in those with gangrene they were: AFS 10%, FFA 67% and LLP 79%. Survival percentages were then calculated from five- and seven-year mortality events, and found to be, respectively: rest pain 57% and 51%; ulcer 44% and 39% and gangrene 48% and 43% (Table IV).

Table 3.

Life-Table Summary, by Indication for Procedure

Seven-Year Events	Event -Free (%)	Number of Events (N)
Rest Pain (N = 220)
AFS	14	107
FFA	84	14
LLP	92	14
Ulcer (N = 457)
AFS	14	260
FFA	77	56
LLP	86	56
Gangrene (N = 275)
AFS	10	163
FFA	67	51
LLP	79	51

Table 4.

Five- and Seven-Year Survival, by Indication for Procedure

Symptomatology	Proportion Surviving (%)	Number of Events (N)
Rest Pain (N = 220)
5-year survival	57	94
7-year survival	51	107
Ulcer (N = 457)
5-year survival	44	256
7-year survival	39	277
Gangrene (N = 275)
5-year survival	48	143
7-year survival	43	158

Procedure indication rates and freedom-from-event rates were subsequently calculated for angioplasty and bypass. There was an increased percentage of ulcer patients in the angioplasty population over the bypass population (52% vs 40%, respectively); however, the seven-year event rates were comparable between the two (Table V and VI).

Table 5.

Indication for Procedure, by Procedure

Procedure	Percent (%)	Number (N)
Bypass (N = 546)
Rest Pain	23	124
Ulcer	40	216
Gangrene	28	152
Angioplasty (N = 460)
Rest Pain	21	96
Ulcer	52	241
Gangrene	27	123

Table 6.

Life-Table Summary, by Procedure

Seven-Year Events	Event - Free (%)	Number of Events (N)
Bypass (N = 546)
AFS	15	302
FFA	77	68
LLP	85	68
Angioplasty (N = 460)
AFS	13	259
FFA	78	55
LLP	86	55

Discussion

In this study, we sought to develop a novel risk metric, designed to better answer a major concern of CLI patients about whether they will undergo a major amputation over the course of their lifetime. By performing the analysis without any censored deaths, we were able to capture those patients who died without any amputation. This led to an increased overall number of amputation-free patients, giving a more optimistic prognosis than originally achieved using the standard amputation-free survival statistic. Using our new method, a patient can be better informed of the likelihood of remaining amputation-free over the course of their remaining lifespan, with 86% of patients remaining amputation-free, rather than the previously calculated freedom from amputation rate, which was 78% in our sample.

Previous studies have addressed the two major outcomes associated with CLI patients: amputation and death.^{7, 11, 16, 17} With five-year survival rates being less than 50%, and one-year mortality rates in non-revascularized patients ranging from 10–40%, death is a major threat to these patients. Additionally, amputation rates are also high, ranging from 10–40% in the first six months alone. Given these morbid outcomes, communication with these patients over treatment strategies and goal is paramount. However, current metrics account for only surviving patients in the risk of amputation, which does not fully quantify risk among those patients with a short lifespan. As shown in the survival curve found in Figure 2, the majority of patients do have a reduced lifespan and therefore should have a metric that can account for this. LLP provides a more patient-centered metric for evaluating the risk of amputation in the face of known high mortality, and better answers patient questions on their risk of amputation in their remaining lifespan.

As shown in prior work, the prognosis for CLI patients is grim with high mortality.^{5–7, 10–12, 18–22} The results of our study support this, as shown by the low rate of only 14% AFS at seven-years. However, AFS and FFA do not completely answer the patient’s question of whether they will lose their limb in their lifetime. The addition of LLP allows clinicians to be more precise with their patients’ individual risks. As shown in our analysis, a physician can now use a patient’s intervention indication to give them an even more accurate description of their risk of amputation, even before undergoing any procedure. Therefore, a patient presenting with an ulcer can be told that while we cannot guarantee how long they will live, there is approximately an 86% chance they will not lose their leg.

Lifelong limb preservation illustrates several important clinical implications. Prior work has been done to assess CLI patients’ quality of life, which is relevant to the overall clinical outcome.²³ The information in this study will help in preoperative patient education by allowing patients to understand not only their risk of death, but their risk of a life free from amputation. Additionally, by using this novel risk metric, we are able to report a higher proportion of patients who remain amputation-free for the remainder of their lifetime, which demonstrates that these interventions often outlast our patient population. When discussing risk with a CLI patient, the threat of losing a limb may induce more fear than death, and current metrics do not inform patients about their risk of limb loss during their lifetime. By using LLP in addition to FFA, a clinician can improve patient communication by addressing concerns that are most important to a patient.

There are qualifications to this new risk metric. We used a Kaplan-Meier life table analysis for our data. When calculating survival curves, this requires censoring of subjects at the time of death. Here, we retained these subjects in the analysis because our only endpoint was amputation. Death without amputation is treated as a continuing success, which makes it possible to give patients a single percentage, reflective of the overall likelihood of avoiding amputation, alive or dead, at any given time point following intervention. While leaving the dead patients in the analysis is counterintuitive to traditional survival analyses, the purpose of this paper is to provide an additional new metric for improved, simplified, patient-centered communication that directly addresses on of their major concerns: “will I ever lose my leg?”

There are also limitations to this study. First, our data are from a single institution, and may not reflect freedom-from-event rates across the country. For example, almost 75% of our patient population is diabetic and 21% are on dialysis. Both of these values are higher than typically encountered and could certainly affect the outcomes. However, the purpose of this study was to provide the rationale for this new metric and how to use it. Second, our patient population was comprised of bypass and angioplasty patients, but femoral endarterectomy is another method of treatment for CLI, which we chose to exclude. The possibility exists that inclusion of these patients would affect the results. Furthermore, we lack data on CLI patients who elected against intervention, which would provide an important comparison group. Nor did we evaluate other post-operative complications, which could affect a patient’s decision. These are all potential targets for future projects utilizing this new metric.

It is also possible that patients underwent amputation or other procedures at outside institutions, and therefore may not be completely captured, thereby underestimating event rates. Additionally, the time frame of seven years was used as the longest time frame available before the data became unstable due to low numbers and loss to follow-up. Selecting different time frames could change the impact of the study, and could be an area of focus in future studies. Finally, we considered all deaths within the first 30 days perioperatively to be a failure in all three metrics to reflect the risk of surgery. However, there were patients who died shortly after this period. There may also be a subset of patients who died as a result of not undergoing amputation. Though these patients would all be a failure in AFS, they would be censored in FFA and a success in LLP, thus highlighting the need for all three metrics.

Despite these limitations, the concept of lifelong limb preservation has important value in patient communication. Providers are constantly aiming to improve their communication and ensure that patients understand their disease. Patients with CLI are no exception to this, and given their often grim prognosis, it is especially important to be able to provide them with realistic outcomes. LLP provides a more patient-centered metric to provide a better understanding of the risk of amputation. Further research should be done in a larger patient population to determine the significance of this new risk metric. Additionally, incorporation of factors such as the WIfi score into future analysis will allow for even more specific outcomes for patients.

Conclusion

Chronic limb-threatening ischemia is a devastating disease with a grim prognosis. Though our interventions have continued to improve, our ability to predict the long-term survival and the risk of amputation for any patient is often challenging. Therefore, when providing a patient with the traditional risk for amputation following an intervention, the question remains, what if the patient dies before then, and how will that affect these statistics? By using lifelong limb preservation (LLP) as an outcomes assessment, we are more accurately able to provide an answer to the patient’s question, “If I have this procedure, will I ever lose my leg?” We recommend that physicians use this new metric to fully discuss future risks with their vascular patients. Future efforts could be directed at developing a risk score to determine who is most likely to suffer an amputation before death.