Determinants of functional outcome after revascularization for critical limb ischemia: An analysis of 1000 consecutive vascular interventions
Article Outline
Background
When reporting standards for successful lower extremity revascularization were established, it was assumed that arterial reconstruction, patency, and limb salvage would correlate with the ultimate goal of therapy: improved functional performance. In reality, factors determining improvement of ambulation and maintenance of independent living status after revascularization have been poorly studied. The purpose of this study was to assess the important determinants of functional outcome for patients after intervention for critical limb ischemia.
Methods
The results of 1000 revascularized limbs from 841 patients were studied. Indications were rest pain, 41.1%; ischemic ulceration, 35.6%; gangrene, 23.3%; infrainguinal, 70.9%; aortoiliac, 24.2%; and both, 4.9%. Treatment was by endovascular intervention, 35.5%; open surgery, 61.7%; and both, 2.8%. Patient were mean age of 68 ± 12 years, and 56.6% were men, 74.7% were white, 54.2% had diabetes mellitus, 67% were smokers, 13.4% had end-stage renal disease and were on dialysis, and 36% had prior vascular surgery. Patients were treated with conventional therapy by fellowship-trained vascular specialists at a single center and were analyzed according to the type of intervention, the arterial level treated, age, race, gender, presentation, the presence of diabetes, smoking history, end-stage renal disease, coronary disease, hypertension, hyperlipidemia, obesity, chronic obstructive pulmonary disease, previous stroke, dementia, prior vascular surgery, preoperative ambulatory status, limb loss ≤1 year of treatment, and independent living status. The technical outcomes of reconstruction patency and limb salvage as well as the functional outcomes of survival, maintenance of ambulation, and independent living status were measured for each variable using Kaplan-Meier life-table analysis, and differences were assessed using the log-rank test. A Cox proportional hazards model was used to assess independent predictors of outcome and obtain adjusted hazard ratios and 95% confidence intervals.
Results
At 5 years, 72.4% of the entire cohort had a patent reconstruction and 72.1% had an intact limb. Overall 5-year functional outcomes were 41.9% for survival, 70.6% for maintenance of ambulation, and 81.3% for independent living status. Outcome was not significantly affected by the type of treatment (endovascular or open surgery) or by the level of disease treated (aortoiliac, infrainguinal, or both). The most important independent, statistically significant predictors of particularly poor functional outcome were impaired ambulatory ability at the time of presentation (70% 5-year mortality, hazard ratio, 3.34; 39.5% failure to eventually ambulate, hazard ratio, 2.83; 30% loss of independent living status, hazard ratio, 7.97), and the presence of dementia (73% late mortality, hazard ratio, 1.57; 41.2% failure to eventually ambulate, hazard ratio, 2.20; 46.4% loss of independent living status, hazard ratio, 5.44). These factors were even more predictive than limb amputation alone.
Conclusion
Functional outcome for patients undergoing intervention for critical limb ischemia is not solely determined by the traditional measures of reconstruction patency and limb salvage, but also by certain intrinsic patient comorbidities at the time of presentation. These findings question the benefit of our current approach to critical limb ischemia in functionally impaired, chronically ill patients—patients who undoubtedly will be more prevalent as our population ages.
In the case of surgical intervention, favorable outcome is determined by preconceived definitions of success. For example, surgical intervention to treat cancer might be considered a success if the tumor were removed and the patient experienced long-term survival with no recurrent disease. When reporting standards for the treatment of critical limb ischemia were originally established, success was defined as patency of the vascular reconstruction and limb salvage. It was assumed that these important objectives would correlate with the unstated ultimate goal, namely, preservation of ambulation, maintenance of independence, and perhaps survival.
Over the years, however, we have observed that arterial reconstruction patency does not necessarily correlate with limb salvage. Indeed, measured limb salvage rates often exceed patency rates.1 The counterintuitive nature of this finding raises an interesting question: Does limb salvage alone consistently predict successful functional outcome? Although some reports have demonstrated the advantages of limb salvage compared with major limb amputation by using end points important to vascular surgeons,2, 3, 4, 5, 6 the question of whether limb salvage alone, using the basic functional end points of ambulation maintenance and preservation of independence, has not been studied.
The number of older patients is expected to increase with the coming of the baby-boomers, and it would be helpful to know which preoperative variables influence ultimate functional outcome in addition to limb salvage. The purpose of this study therefore was to determine the importance of limb salvage on ultimate functional outcome and to assess other important determinants of functional outcome for patients after intervention for critical limb ischemia.
Methods
The postoperative functional outcome of 841 patients undergoing 1000 consecutive lower extremity vascular interventions between January 1998 and May 2004 by the vascular surgery teaching service of the Greenville Hospital System University Medical Center (GHSUMC) in Greenville, South Carolina was analyzed using 22 perioperative variables, including limb loss ≤1 year of intervention. Seven fellowship-trained vascular surgeons and one interventional fellowship-trained vascular internist on the teaching service performed all interventions to treat critical, limb-threatening ischemia, defined as ischemic rest pain, ulceration, and gangrene, with an ankle pressure < 50 to 70 mm or a toe pressure <30 mm.
The type of intervention was determined by the presenting patient scenario and by the clinical judgment of the attending physician. The 21 perioperative variables analyzed in addition to limb salvage included age, race, gender, presentation (rest pain, ulceration, gangrene), the presence of diabetes mellitus, tobacco status, chronic renal failure and end-stage renal disease, the presence of coronary artery disease (high risk by the Eagle criteria),7 history of hypertension, history of hyperlipidemia, the presence of obesity (body mass index of ≥30 kg/m2), history of chronic obstructive pulmonary disease (COPD), previous stroke, history of previous cerebrovascular disease, the presence of dementia, history of a vascular procedure, ambulatory status at presentation, independent living status at presentation, the type of procedure performed (endovascular, open, both), the level of arterial disease treated (aortoiliac, infrainguinal, both), and the type of intervention by the level of arterial disease treated (aortoiliac–endovascular, aortoiliac–open, aortoiliac–and both, infrainguinal–endovascular, infrainguinal–open, infrainguinal–both, both–endovascular, both–open, both–both).
Preoperative ambulatory status was characterized as ambulatory (independent ambulation out of house), ambulatory/homebound (ambulatory in the home only), or nonambulatory/transfer (eg, uses legs to transfer from bed to chair or from the chair to the commode). In each case, impaired ambulatory status was determined by physical conditions that were thought to be independent of their vascular disease. Impaired ambulatory status was defined as the patient’s functional status immediately before the development of vascular symptoms. With this definition, ambulatory status was usually a function of other medical comorbidities such as arthritis, cerebrovascular and cardiovascular disease, or advanced age.
Preoperative living status was characterized as independent, defined as living in an independent dwelling without external assistance; or non-independent, defined as living in an assisted living environment or a private residence with external assistance for activities of daily living. As with ambulatory status, living status was determined immediately before the development of vascular symptoms.
Postoperative functional outcome measures included mortality; change in ambulatory function, defined as the permanent postoperative change, despite full recovery from the revascularization, in ambulatory classification (eg, ambulatory to nonambulatory/transfer, ambulatory to nonambulatory/bedridden, or nonambulatory/transfer to nonambulatory/bedridden); and a decline in independent living status, defined as a permanent change to an assisted living residence or incorporation of permanent help into the postoperative domicile to enhance functions of daily living. Short-term assisted living during recovery was not considered a loss of independence.
To frame the postoperative functional outcome measures within the context of the traditional measures of technical success, secondary arterial reconstruction patency (the percentage of limbs with a patent arterial reconstruction) and limb salvage were obtained. Patency, as previously described,8 was determined clinically, in the vascular lab, or by arteriography.
The lower extremity database and data collection process
Since 1992 a prospective vascular registry of all cases performed on the GHSUMC Vascular Surgery Service has been maintained. Since 1998, however, the year our endovascular program was initiated, a subset of patients with lower extremity peripheral arterial disease have been closely scrutinized and actively followed with Institutional Review Committee approval. Each procedure is entered on an Excel spreadsheet (Microsoft Corp, Redmond, Wash). Preoperative demographics are obtained at presentation and entered into the database. Functional information (ambulatory status and living situation) and also the lower extremity grading system (LEGS) score are included.
Follow-up has been recorded at each follow-up visit. For patients receiving open infrainguinal bypass, follow-up with noninvasive duplex scan derived graft flow velocities is obtained at 1 month, every 3 months for the first 18 months, and then every 6 months after this. Interventions for failing bypass grafts (isolated graft flow velocity >300 cm/s and distal velocities <45 cm/s) are recorded in the database. Patients receiving all other types of revascularization (aortoiliac open reconstruction, and all endovascular intervention) are followed-up with a patient visit and a noninvasive vascular study at 1 month and then at 6-month intervals.
The database information is updated with each visit. In addition, the database has been scrutinized each summer by independent research workers looking for missing data points or missing patients. Sources used to attain follow-up include the hospital computerized Lifetime Clinical Record, the computerized radiology Picture Archiving Communication System, and the online obituary services of all statewide newspapers.
To date, >1600 revascularizations and 600 major lower extremity amputations have been entered into the database and monitored for functional outcome and living status. As a result of these follow-up methods, 73 of the 841 patients studied were lost to follow-up. For the purpose of the life tables, patients lost to follow-up were censored at the last confirmed follow-up date. At the end of 60 months, 75 patients were left for analysis.
Statistical analysis
Kaplan-Meier survival curves were used to assess reconstruction patency, limb salvage, survival, maintenance of functional ambulatory status, and maintenance of independent living status over time. The proportional hazards assumption was tested for each plot and found to be appropriate. The log-rank test was used to assess differences in these curves. All factors significantly associated with time to measured event (thrombosis, limb loss, death, loss of ambulation, or loss of independence) in bivariate analysis were then entered into a Cox proportional hazards model. Backward stepwise elimination was used to remove nonsignificant factors from the model. Hazard ratios (HR) and 95% confidence intervals from the final model were used to describe event time risk. All analyses were conducted using SAS 8 software (SAS Inc, Cary, NC). Values of P < .05 were considered indicative of statistical significance.
Results
The overall results for the technical measures of arterial reconstruction patency and limb salvage using Kaplan-Meier life table curves are shown for the entire cohort in Fig 1, Fig 2. The mean age for the cohort was 68.1 ± 12.1 years (range, 29 to 97 years). The mean ± standard deviation follow-up was 27.5 ± 21.6 months, and median follow-up was 24 months (range, 0 to 89 months) for 1931 person-years of follow-up; 75% of patients had ≥1 year of follow-up. The overall 5-year secondary patency was 72.4% and the limb salvage rate was 72.1% (ie, the percentage of patients with a patent reconstruction and an intact limb at the time of functional outcome assessment).
Fig 1.
Kaplan-Meier life tables show the overall secondary patency in 1000 revascularized limbs. Whiskers represent 95% confidence intervals.
Fig 2.
Kaplan-Meier life tables show the overall limb salvage in 1000 revascularized limbs. Whiskers represent 95% confidence intervals.
The log-rank test showed there were no statistically significant differences in the Kaplan-Meier curves analyzing secondary patency for all endovascular procedures vs all open procedures vs both: 1-year patency was 76.6% (71.5, 81.0) vs 82.7% (79.3, 85.6) vs 72.9% (51.2, 86.2). There was also no difference in limb salvage for endovascular procedures vs open procedures vs both: 1-year limb salvage was 76.5 % (71.8, 81.3) vs 82.4% (79.3, 85.5) vs 72.9% (55.6, 90.2). There was, however, a statistically significant better secondary patency rate for patients receiving aortoiliac intervention vs infrainguinal intervention, with a 1-year patency of 89.0% (84.1, 92.5) vs 77.9% (74.4, 80.9) (P < .0001). There was also statistical significance with 1-year limb salvage for patients receiving aortoiliac intervention vs infrainguinal intervention (87.3% vs 78.3%; P = .0002) and for patients receiving aortoiliac open intervention vs patients receiving an infrainguinal endovascular intervention (94.4% vs 75.3%, P = .0033).
Cox proportional hazards models were used to derive preoperative factors independently associated with reconstruction thrombosis and limb loss. Patients with gangrene (1-year secondary patency, 66.7%; HR, 2.05, P < .001), impaired preoperative ambulatory status (1-year secondary patency for nonambulatory/homebound, 64.8%; HR, 1.66; P = .002), or ESRD (1-year secondary patency, 57.5%; HR, 2.64; P < .001) had significantly lower secondary patency. Patients presenting with gangrene (1-year limb salvage, 66.7%; HR, 1.78; P = .0019), diabetes (1-year limb salvage, 74.8%; HR, 1.40; P = .0357), end-stage renal disease (1-year limb salvage, 57.3%; HR, 2.51; P < .0001), multilevel arterial disease (1-year limb salvage, 71.7%; HR, 2.67; P = .0015), or significantly impaired preoperative ambulation (1-year limb salvage for nonambulatory/homebound, 67.0%; HR, 1.79; P = .0003) had significantly lower limb salvage rates.
Overall results for survival, maintenance of ambulation, and maintenance of independent living using Kaplan-Meier life-table curves are shown for the entire cohort in Fig 3, Fig 4, Fig 5, respectively. Table I displays the results for each outcome by each of the predictive variables and log-rank P values. At 5 years, overall survival was 41.9%, maintenance of ambulation was 70.6%, and maintenance of independent living status was 81.3%.
Fig 3.
Kaplan-Meier life tables show the cumulative survival in 841 patients who underwent 1000 limb revascularizations. Whiskers represent 95% confidence intervals.
Fig 4.
Kaplan-Meier life tables showing cumulative maintenance of ambulation in 841 patients who underwent 1000 limb revascularizations. Whiskers represent 95% confidence intervals.
Fig 5.
Kaplan-Meier life tables showing cumulative maintenance of independent living status in 841 patients who underwent 1000 limb revascularizations. Whiskers represent 95% confidence intervals.
Table I. Preoperative patient characteristics by functional outcome
| Patient characteristic | Survival | Ambulatory status | Independent living | |||
|---|---|---|---|---|---|---|
| Patients at risk (n) | Died (%) | Patients at risk (n) | Failed ambulation (%) | Patients at risk (n) | Failed independent living (%) | |
| Age group | ||||||
 <50 | 69 | 27.5 | 69 | 21.7 | 69 | 5.8 |
| 50-59 | 136 | 33.1 | 136 | 14.7 | 134 | 5.2 |
| 60-69 | 220 | 36.8 | 220 | 20.5 | 218 | 10.6 |
| 70-79 | 262 | 50.8 | 262 | 21.0 | 252 | 16.3 |
| ≥80 | 154 | 48.1 | 154 | 24.0 | 145 | 18.6 |
| Log-rank P | <.01 | .10 | <.01 | |||
| Race | ||||||
| White | 640 | 56.3 | 640 | 19.1 | 623 | 11.2 |
| Black | 194 | 36.6 | 194 | 25.3 | 188 | 17.0 |
| Other | 7 | 14.3 | 7 | 14.3 | 7 | 0 |
| Log-rank P | .04 | .45 | .20 | |||
| Gender | ||||||
| Male | 479 | 41.3 | 479 | 16.5 | 471 | 10.4 |
| Female | 362 | 42.5 | 362 | 25.7 | 347 | 15.3 |
| Log rank P | .76 | <.01 | .11 | |||
| Presentation | ||||||
| Rest pain | 341 | 39.0 | 341 | 15.8 | 337 | 10.1 |
| Ulceration | 308 | 41.6 | 308 | 18.2 | 294 | 9.2 |
| Gangrene | 192 | 47.4 | 192 | 32.3 | 187 | 21.9 |
| Log-rank P | .02 | <.01 | <.01 | |||
| Diabetes | ||||||
| No | 385 | 36.1 | 385 | 14.0 | 376 | 10.1 |
| Yes | 456 | 46.7 | 456 | 25.9 | 442 | 14.5 |
| Log-rank P | .01 | <.01 | .05 | |||
| Smoking status | ||||||
| Never | 278 | 39.9 | 278 | 23.7 | 267 | 14.2 |
| Former | 218 | 50.9 | 218 | 21.1 | 212 | 16.0 |
| Current | 345 | 37.7 | 345 | 17.4 | 339 | 8.8 |
| Log-rank P | <.01 | .07 | .01 | |||
| ESRD | ||||||
| No | 633 | 36.8 | 633 | 16.0 | 619 | 10.2 |
| Insufficiency | 106 | 50.0 | 106 | 29.2 | 102 | 16.7 |
| Yes | 102 | 64.7 | 102 | 39.2 | 97 | 22.7 |
| Log-rank P | <.01 | <.01 | <.01 | |||
| CAD | ||||||
| No | 362 | 30.9 | 362 | 17.1 | 355 | 9.3 |
| Yes | 479 | 50.1 | 479 | 23.0 | 463 | 14.9 |
| Log-rank P | <.01 | <.01 | <.01 | |||
| Hypertension | ||||||
| No | 159 | 40.9 | 159 | 18.2 | 153 | 9.8 |
| Yes | 682 | 42.1 | 682 | 21.0 | 665 | 13.1 |
| Log-rank P | .40 | .28 | .19 | |||
| Hyperlipidemia | ||||||
| No | 536 | 45.0 | 536 | 22.6 | 519 | 13.7 |
| Yes | 305 | 36.4 | 305 | 16.7 | 299 | 10.4 |
| Log-rank P | .05 | .04 | .18 | |||
| Obesity | ||||||
| No | 711 | 42.8 | 711 | 21.1 | 688 | 13.7 |
| Yes | 130 | 36.9 | 130 | 16.9 | 130 | 6.2 |
| Log-rank P | .44 | .36 | .03 | |||
| COPD | ||||||
| No | 665 | 36.7 | 665 | 20.8 | 646 | 11.0 |
| Yes | 176 | 61.4 | 176 | 19.3 | 172 | 18.0 |
| Log-rank P | <.01 | .72 | <.01 | |||
| CVA | ||||||
| No | 686 | 39.9 | 686 | 19.1 | 670 | 11.2 |
| Yes | 155 | 50.3 | 155 | 26.5 | 148 | 18.2 |
| Log-rank P | <.01 | <.01 | <.01 | |||
| CVD | ||||||
| No | 621 | 39.6 | 621 | 18.8 | 608 | 11.7 |
| Yes | 220 | 48.2 | 220 | 25.0 | 210 | 14.8 |
| Log-rank P | <.01 | <.01 | .07 | |||
| Dementia | ||||||
| No | 773 | 39.1 | 773 | 18.6 | 762 | 10.0 |
| Yes | 68 | 73.5 | 68 | 41.2 | 56 | 46.4 |
| Log-rank P | <.01 | <.01 | <.01 | |||
| Prior vascular procedure | ||||||
| No | 550 | 42.0 | 550 | 18.9 | 533 | 11.6 |
| Yes | 291 | 41.6 | 291 | 23.4 | 285 | 14.0 |
| Log-rank P | .74 | .21 | .40 | |||
| Ambulation | ||||||
| Ambulatory | 664 | 34.5 | 664 | 16.6 | 663 | 7.8 |
| Homebound | 147 | 69.4 | 147 | 39.5 | 135 | 32.6 |
| Transfer | 30 | 70.0 | 30 | 13.3 | 20 | 30.0 |
| Log-rank P | <.01 | <.01 | <.01 | |||
| Living status | ||||||
| Independent | 809 | 40.7 | 809 | 20.1 | 809 | 12.2 |
| Assisted | 9 | 66.7 | 9 | 0 | 9 | 33.3 |
| Nursing home | 23 | 73.9 | 23 | 39.1 | –n/a– | –n/a– |
| Log-rank P | <.01 | <.01 | <.01 | |||
| Intervention Endovascular | 302 | 40.4 | 302 | 20.9 | 288 | 12.2 |
| Open Surgery | 513 | 41.9 | 513 | 20.5 | 505 | 12.3 |
| Both types | 26 | 57.7 | 26 | 15.4 | 25 | 20.0 |
| Log-rank P | <.01 | .06 | .17 | |||
| Limb loss ≤ 1st yr | ||||||
| No | 670 | 38.8 | 670 | 10.9 | 655 | 8.5 |
| Yes | 171 | 53.8 | 171 | 42.1 | 163 | 28.2 |
| Log-rank P | <.01 | <.01 | <.01 | |||
When mortality was examined, preoperative factors found not to be statistically significant in determining late mortality included gender (P = .758), hypertension (P = .402), obesity (P = .435), and prior vascular surgery (P = .742). Cox proportional hazards models were used to derive preoperative factors independently associated with death (Table II). In decreasing order of influence they include nonambulatory/transfer-only status (1-year survival, 90%; HR, 3.45), end-stage renal disease (1-year survival, 67.1%; HR, 2.25), age >70 (1-year survival, 79.7%; HR, 1.85), COPD (1-year survival, 73.8%; HR, 1.76), the dementia (1-year survival, 55.4%; HR, 1.57), and coronary disease (1-year survival, 75.9%; HR, 1.45). Limb loss ≤1 year of intervention was not a statistically significant determinant of mortality.
Table II. Cox proportional hazards model results for time to death after revascularization
| Patient characteristic | HR estimates (95% CI) | P |
|---|---|---|
| Age group | ||
| <50 | Referent | .26 |
| 50-59 | 1.36(0.80,2.34) | .24 |
| 60-69 | 1.36(0.82,2.25) | .01 |
| 70-79 | 1.85(1.13,3.04) | .03 |
| ≥80 | 1.81(1.07,3.09) | |
| Race | ||
| White | Referent | — |
| Black | 0.72(0.55,0.95) | .02 |
| ESRD | ||
| No | Referent | — |
| Insufficiency | 1.25(0.92,1.71) | .16 |
| Yes | 2.25(1.66,3.06) | <.01 |
| CAD | ||
| No | Referent | — |
| Yes | 1.45(1.14,1.84) | <.01 |
| COPD | ||
| No | Referent | — |
| Yes | 1.76(1.38,2.24) | <.01 |
| Dementia | ||
| No | Referent | — |
| Yes | 1.57(1.13,2.17) | <.01 |
| Ambulation | ||
| Ambulatory | Referent | — |
| Homebound | (1.77,3.05) | <.01 |
| Transfer | 3.45(2.18,5.56) | <.01 |
Factors found not to be statistically significant in determining maintenance of ambulation included age (P = .099), race (P = .454), history of smoking (P = .065), hypertension (P = .282), obesity (P = .356), COPD (P = .718), prior vascular surgery (P = .207), or the type of vascular intervention (endovascular vs open surgery) (P = .063). Cox proportional hazards models were used to derive factors independently associated with deterioration of ambulatory status postoperatively (Table III). In decreasing order of influence they include limb loss ≤1 year (ambulatory maintenance at 1 year, 47.9%; HR, 6.75), ambulatory/homebound status (ambulatory maintenance at 1 year, 63.8% HR, 3.01), the dementia (ambulatory maintenance at 1 year, 67%; HR, 2.01), end-stage renal disease (ambulatory maintenance at 1 year, 73.6%; HR, 1.25), female gender (ambulatory maintenance at 1 year, 85.9%; HR, 1.59), and the presence of diabetes (ambulatory maintenance at 1 year, 81.1%; HR, 1.44).
Table III. Cox proportional hazards model results for analysis of ambulatory deterioration/failure after revascularization for time to ambulation failure
| Patient characteristic | HR estimates (95% CI) | P |
|---|---|---|
| Sex | ||
| Male | Referent | — |
| Female | 1.59(1.17,2.15) | <.01 |
| Diabetes | ||
| No | Referent | — |
| Yes | 1.44(1.03,2.012) | .03 |
| ESRD | ||
| No | Referent | — |
| Insufficiency | 1.67(1.10,2.52) | .02 |
| Yes | 1.25(0.80,1.95) | .32 |
| Dementia | ||
| No | Referent | — |
| Yes | 2.01(1.24,3.23) | <.01 |
| Ambulation | ||
| Ambulatory | Referent | — |
| Homebound | 3.01(2.12,4.27) | <.01 |
| Transfer | 0.91(0.33,2.51) | .86 |
| Limb loss ≤1st year | ||
| No | Referent | — |
| Yes | 6.75(4.88,9.33) | <.01 |
When maintenance of independence was examined, 23 patients who were already in nursing homes before their arterial intervention were excluded from analysis. Factors found not to be statistically significant in determining maintenance of independent living status in the remaining patients included race (P = .196), gender (P = .108), diabetes (P = .051), hypertension (P = .187), hyperlipidemia (P = .18), cerebrovascular disease (P = .070), prior vascular surgery (P = .402), or the type of vascular intervention performed (endovascular vs open surgery) (P = .171). Cox proportional hazards models were used to derive factors independently associated with deterioration in independent living status postoperatively (Table IV). In decreasing order of influence they include nonambulatory/transfer status (maintenance of independent living status at 1 year, 68.7%; HR, 6.05), the presence of dementia (maintenance of independent living status at 1 year, 58.7%; HR, 5.95), ambulatory/homebound status (maintenance of independent living status at 1 year, 66.8%; HR, 3.77), age 70 to 79 (maintenance of independent living status at 1 year, 89.1%; HR 3.42), limb loss within 1 year (maintenance of independent living status at 1 year, 77.3%; HR 3.24), age greater than 80 (maintenance of independent living status at 1 year, 87%; HR, 3.09), and history of a stroke (maintenance of independent living status at 1 year, 86.6%; HR, 1.9).
Table IV. Cox proportional hazards model showing preoperative factors associated with failure in independent living
| Patient characteristic | HR estimates (95% CI) | P |
|---|---|---|
| Age group | ||
| <50 | Referent | — |
| 50-59 | 1.04(0.30,3.56) | .96 |
| 60-69 | 1.82(0.62,5.35) | .28 |
| 70-79 | 3.42(1.21,9.71) | .02 |
| ≥80 | 3.09(1.04,9.16) | .04 |
| Presentation | ||
| Rest pain | Referent | — |
| Ulceration | 0.57(0.33,0.97) | .04 |
| Gangrene | 1.47(0.90,2.39) | .13 |
| CVA | ||
| No | Referent | — |
| Yes | 1.90(1.20,3.00) | <.01 |
| Dementia | ||
| No | Referent | — |
| Yes | 5.95(3.64,9.72) | <.01 |
| Ambulation | ||
| Ambulatory | Referent | — |
| Homebound | 3.77(2.36,6.03) | <.01 |
| Transfer | 6.05(2.45,14.95) | <.01 |
| Limb loss within 1st year | ||
| No | Referent | — |
| Yes | 3.24(2.14,4.93) | <.01 |
Discussion
The purpose of this study was twofold. First, we attempted to determine the importance of limb salvage on the ultimate functional outcome of patients undergoing operative treatment for critical limb ischemia secondary to peripheral arterial disease of the lower extremities. Second, we attempted to assess what other factors may be of importance in determining functional outcome after intervention. This was done by retrospectively reviewing 22 perioperative variables, including amputation ≤1 year of intervention. We found that limb loss did indeed result in significant functional impairment, thus confirming the intuitive importance of limb salvage surgery. Limb loss, however, was not the only variable found to be an important determinant of functional outcome.
Our study measured functional outcome using the three parameters of survival, maintenance of ambulation, and maintenance of independent living status. These parameters were chosen because they represent anticipated benefits one would intuitively expect after lower extremity revascularization. Too often our literature has chosen to report end points of success that are probably more important to surgeons than to patients. These technical end points, which include jargon such as primary patency, assisted primary patency, and secondary patency, are useful to the craftsmen who perform the procedures; however, patients and the payers are usually more interested in answers to basic concerns. Will this procedure make me walk better and will it keep me out of the nursing home? On examination of the data, we found that improvement after intervention (ie, improving from ambulatory homebound to ambulatory out of home or progressing from the nursing home back to home) rarely occurred. It became obvious that our best hope for success in many of these patients was simply to prevent deterioration in functional status.
Even when we lowered our expectations, we found deterioration occurred in several patient groups with specific comorbidities despite successful revascularization. Of all perioperative comorbidities examined, 12 were found to independently predict poorer outcome in at least one parameter measured. Only two variables independently predicted poorer functional outcome in all three categories: preoperative dementia and impaired ambulatory status at the time of presentation for arterial intervention. Three variables were independently associated with poorer functional outcome in two functional outcome parameters: old age (>75 years), limb loss ≤1 year of intervention, and end-stage renal disease. Seven variables were independently associated with poorer functional outcome in one category, three associated with increased mortality (the presence of coronary artery disease, COPD, and African American ethnicity), two associated with loss of ambulation (diabetes and female gender), and two associated with loss of independence (gangrene and history of stroke).
Although hazard ratios showed that amputees had a sixfold greater chance of not ambulating and a threefold greater chance of losing their independence, they experienced no decrease in survival. This contrasts with patients who presented with impaired ambulatory status and dementia preoperatively. These patients not only experienced a twofold to threefold risk of ambulatory deterioration and a sixfold risk of losing their independent living status, but they also experienced a 1.5-fold to threefold risk of mortality compared with referent controls. Other high-risk groups included elderly patients, who had a twofold higher risk of mortality and a threefold risk of losing independence, and patients with end-stage renal disease, who had a twofold higher risk of mortality and a 1.25 greater risk of losing independence.
The findings of this study suggest that some patients who present with limb threatening ischemia, meet every classic indication for intervention, and may have a successful procedure with satisfactory technical outcome, including limb salvage, will experience no functional benefit. These are patients with dementia and impaired ambulatory ability at presentation. They are more prone to death, ambulatory failure, and loss of independence than all other groups observed, even the cohort of patients who lose their limbs. These findings challenge the cognitive foundation of any vascular surgeon, a foundation that states it is acceptable to attempt lower extremity revascularization on any patient who shows aptitude toward ambulation and who is capable of surviving intervention. This report suggests we need to reconsider our position.
It is important to recognize the social backdrop in which this study has been performed. In 2006, the first of the baby-boomers will turn 60 years old. If one uses conservative census figures with atherosclerotic prevalence data, we can expect a 66% increase in patients presenting with peripheral arterial disease of the lower extremities during the next 10 years.9 In addition, patients are living longer with coexisting chronic diseases such as end-stage renal disease. Thus, they present for medical treatment of routine illnesses at advanced age with impaired physiology. Couple this with the looming financial crisis that faces medicine, and one can see that our current medical system will not have the resources available to treat all patients who present, especially those with degenerative diseases attributable to aging. Rationing of care, when it occurs, will probably be economically driven. Treatment may be available, but not “covered” by the third-party payer. It can be imagined that expensive treatment such as intervention for peripheral arterial disease will receive close scrutiny and will most likely be reserved for patients whose therapy has proven functional benefit—a benefit capable of promoting independence and avoidance of expensive institutional skilled nursing care. Studies that help define cost-efficient delivery of health care are desperately needed.
With this in mind, we sought to determine whether our interventions maintained ambulation and independence, a totally different perspective from vascular conventional wisdom where patency still reigns as king. Patency in this report was used only as a frame of reference for considering functional outcome, meaning, was the arterial reconstruction open and working at the time of functional assessment? The use of ambulation and independence as end points radically changes our definition of success. For example, major limb amputations with prosthetic rehabilitation are no longer measures of failure but are considered measures of success. This may represent heresy to many vascular surgeons, but we suspect payers may not consider it with such disdain.
Our study was not designed to examine the natural history of untreated critical ischemia as it pertains to maintenance of ambulation and independence, and thus it is conceivable that functional performance would have been significantly worse without vascular intervention. It was a retrospective study performed at a single center. The type of treatment performed was left to the discretion of the treating physician and was not controlled. Although it identified groups of patients who did poorer than the referent population (patients such as older demented individuals with impaired preoperative functional status or end-stage renal disease), it falls short of providing helpful concrete recommendations for how to treat poor performers. It cannot be concluded that all of the identified high-risk patients will do poorly. Examining the data in this report shows only that a higher percentage of patients in these high-risk groups performed worse than their referent cohort. Some patients in these groups performed well.
Despite these limitations, our report is a large population-based study from the only group of vascular specialists within a referral basin of 1 million people. We have concluded that when the major goal of therapy is maintenance of ambulation and independence, some patients will not be helped with revascularization. These patients are often experiencing a systemic deterioration at the end stages of life and may represent up to 20% to 25% of people, as was the case in our series, who present with symptoms of critical limb ischemia. In these cases, toe gangrene may be the inciting event that motivates the patient to seek care, but treating this is the equivalent of repairing a light bulb on a sinking ship. Further prospective study to determine characteristics attributable to patients who eventually performed well and characteristics of those who eventually performed poorly are needed prior to any consideration of rationing of care.
Conclusion
Functional outcomes such as survival, maintenance of ambulation, and independent living after intervention for critical limb ischemia of the lower extremities are not solely determined by the classic outcome measures of intervention patency and limb salvage. There are additional intrinsic patient comorbidities at the time of presentation that are equally reliable determinants of functional outcome. Specifically, this report questions our conventional approach to critical limb ischemia in patients who are mentally and functionally impaired or who have chronic degenerative diseases such as end-stage renal disease, patients who undoubtedly will be more prevalent as our population ages.
Author contributions
Discussion
Dr John Eidt (Little Rock, Ark). I would like to extend my congratulations to Dr. Taylor for a wonderful presentation. It is an important topic and I think it is timely, and I want to thank him and his coauthors for getting me a copy of this data-packed manuscript a week ago, and even though I’ve had a week, I still haven’t really scratched the surface. There is a lot of information here, and when this gets published you really need to read this because I think it has a lot of important data.
I’d like to emphasize this is a review of 1000 consecutive procedures. It is not selected, and no claudicants are included in this. They have been treated by one of the most well-respected groups of vascular surgeons in the south and they have used state-of-the-art open and endovascular techniques. These results are assuredly as good or better than certainly in our practice, and I’m sure in many of our practices.
The study is somewhat limited by the fact that it is retrospective and much of the data is dependent on the correct input of diagnoses codes at hospital level. For example, I find it a little hard to believe that 31% of these patients living in the heart of tobacco country were recorded as never using tobacco. These limitations are overcome to some degree by the sheer size of the database.
Rather than focusing on traditional end points of graft patency and limb salvage that have become the sort of holy grail of vascular surgery, Dr Taylor and his colleagues have tried to answer a more philosophical question: does vascular intervention really improve our patients’ lives? Sadly, it appears that in many cases, vascular intervention may be fruitless and in some cases even harmful. It may be that we just focus too much on limb salvage and patency and too little attention on functional outcome. It looks like in about a third of these patients, vascular intervention, regardless of open or endovascular, was really of limited value as measured by functional outcome. I have three questions.
The first question is, the average age was 68 years. You showed that at 5 years only 40% or so of these patients were alive, about 70% were ambulatory, and 80% were living independently. Do you have any idea how this compares to an age-matched control group? In other words, are we improving or are we making these people’s lives actually worse?
Second, I recall that John Porter, the late John Porter, from Portland, presented data a few years ago that showed that vascular surgery had little impact really on ambulatory status. He found that nonambulatory patients usually remained nonambulatory after vascular surgery and likewise that ambulatory patients tended to remain ambulatory, i.e., there were few crossovers. I wonder, you have sort of focused on the downside of your data but I wonder were you able to identify factors that were predictive of improved ambulatory status or return to baseline? In other words, how do we better select people we should treat because we are going to improve their ambulatory status?
Finally, I had a well-educated patient that survived a fem-pop bypass, a follow-up fem-tib bypass, a bypass revision, and eventual amputation. When I saw him several years later at a cocktail party he told me that if he had realized how good it was to have an amputation he would never have consented to bypass surgery. You showed that two factors in particular had a greater impact on long-term functional outcome than even amputation itself and that was impaired ambulatory status at the time of presentation and dementia. Has your practice been changed by your data to include more primary amputations, particularly in the demented bed-bound nursing home patient?
Again, I want to congratulate you for undertaking this arduous ask of honestly looking at your results. These types of results will I am sure be very important to those responsible for financing health care in this country, and I would like to thank again the organization for the privilege of initiating the discussion of this paper.
Dr Taylor. John, in terms of cigarette smoking, you know Dr Debakey, I think, or I think it was D. Crawford, used to say that everybody with an aneurysm was either a smoker, a past smoker, or a liar. I’m sure that we probably missed a few of the cigarette smokers in 31%.
Age-matched control is a great question. I think the previous presentation is very eye opening. Maybe what we need to do with the last lady that I showed on the slide is send her to the wound care center and see if we can nurse her through all those. While Dr Marston was down on himself saying that taking care of patients for a year with a wound is—I would argue that is probably a very acceptable alternative. You can keep this lady in her functional state at the nursing home and being able to do what she has been doing. Age-matched control is a great question, and I think you heard some data in the previous presentation that would be insightful regarding that.
Who improved? Theoretically you would think that when somebody comes with a bad leg and you are operating on a bad leg they should get better or you should see improvement, but in reality when we looked at the LEGS data, which is the data that we found, we found this was unusual. It was very, very unusual to see people go from a classification as we classified as ambulation up to an improved ambulation. I think the best—I mean it was very, very rare. It was so rare that we did not look at it in this study, but we can certainly do it and probably by the time the manuscript is published, we could probably put it in, and I apologize, I don’t know the numbers but it is a very unusual occurrence. We typically—I believe your best chance (I agree with Dr Porter)—is to probably hope to get or achieve maintenance of your current level of function.
Has this changed our practice? It is impossible not to know the information that we know and not let it subconsciously work into the way you manage patients. I will say this though, we are as aggressive as ever. Patients still get referred for revascularization. They don’t want to hear that they are going to lose their leg. We still try, but I will tell you, we need to look at it and we need to start thinking about what’s the right thing to do in light of this and I think it has and will eventually change our practice.
Dr Thomas Brothers (Charleston, SC). Dr Taylor, I want to lend my congratulations to an excellent presentation. I wonder if you would also agree that freedom from pain and also preserving a sense of body wholeness are other worthwhile goals that we might strive to achieve and ask whether or not you might consider including those elements of overall quality of life at least equally as important as level of functionality as you review your results.
Dr Taylor. We stayed with the parameters that we have stayed with. Your partner and I, Dr Robertson, were talking yesterday. Functional outcomes you can define in a lot of different ways. Being able to stand, to dress. Is that an important thing? Bodily wholeness, freedom from pain—no question about it. These are indications, these are parameters that I would argue probably need to be looked at more so than patency. Of course we don’t have those data, but it is clearly something that would be worthwhile to look at.
Dr James Seeger (Gainesville, Fla). It is just a little bit of a different spin on this that I would ask you to look at and that is that you looked at the predictors of poor outcome, but you didn’t look at the predictors of good outcome. Most of us would assume that good outcome from the patient’s point of view is that they are alive, walking around, and at home. Although we sort of think we can predict those people, the two questions I would have are what percentage of the patients achieved that outcome, all three good outcomes, and then did you have any predictors of that group of patients in which it would be no question that they should have a procedure?
Dr Taylor. Well you know you can flip our Cox models around and basically you are half full or half empty. You’re right, we have taken a pessimistic approach to this, and I think it is biased upon the last case I presented. I don’t know about you—I know you are down in Florida, just as we are. You are inundated with these difficult, difficult patient decisions and you go home and you feel like you haven’t helped anybody and in reality the sobering fact is we probably haven’t helped as many people as we did 10 years ago when I was in the Air Force working with Dr Mills, so it is a reality. The flip side, the paper that we have on for Saturday, the people we probably help more than anything are claudicants. We probably—you had to admit it that the cardiologists may be onto something that we are not, but we probably should—the people that you are going to get the most functional benefit in, and I think there have been data that have been shown certainly at Northwestern and others, is you are going to help the claudicants, get them back working again, and we probably undertreat claudication in my opinion only, my opinion, and probably overtreat limb-threatening ischemia, but you need data like this to kind of see what is what. Not everybody that was end-stage renal disease or not everybody that had poor function did poorly. Which patients did poorly and which patients didn’t do poorly? Why did some do well? These are the questions. This study wasn’t really very helpful at the bedside but it certainly tells areas where we need to go to work at the bedside to find out exactly which patients need to be revascularized and which ones do you not treat and which ones you send to Dr Marston for the wound care clinic. I think there is a lot of work that needs to be done.
References
- . Management of peripheral vascular disease . J Vasc Surg . 2000;31(suppl):S230–S231
- . Surgery for chronic lower extremity ischemia in patients eighty or more years of age: operative results and assessment of postoperative independence . J Vasc Surg . 1993;18:618–626
- . Critical limb ischemia in patients over 80 years of age: options in a district general hospital . Br J Surg . 1985;82:1361–1363
- Preoperative clinical factors predict post-operative functional outcomes after major limb amputation: an analysis of 553 consecutive patients . J Vasc Surg . 2005;42:227–235
- . Quality of life following lower limb amputation for peripheral vascular disease . Eur J Vasc Surg . 1993;7: 448–5
- . Major amputations of the lower leg (The patients two years later) . Acta Chir Belg . 2002;102:248–252
- Guidelines for perioperative cardiovascular evaluation for non-cardiac surgery (Report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on Perioperative Cardiovascular Evaluation for Noncardiac Surgery)) . J Am Coll Cardiol . 1996;27:910–948
- Invasive treatment of chronic limb ischemia according to the lower extremity grading system (LEGS) score: a 6-month report . J Vasc Surg . 2004;39:1268–1276
- . Management of peripheral vascular disease . J Vasc Surg . 2000;31(suppl):S8–S34
Competition of interest: none.CME article
PII: S0741-5214(06)01100-1
doi:10.1016/j.jvs.2006.06.015
© 2006 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.