Preoperative frailty is predictive of complications after major lower extremity amputation

Preoperative frailty is predictive of complications after major lower extremity amputation

From the Society for Vascular Surgery Preoperative frailty is predictive of complications after major lower extremity amputation Zachary B. Fang, BS,...

405KB Sizes 31 Downloads 70 Views

From the Society for Vascular Surgery

Preoperative frailty is predictive of complications after major lower extremity amputation Zachary B. Fang, BS,a Frances Y. Hu, BA,a Shipra Arya, MD,a Theresa W. Gillespie, PhD,b and Ravi R. Rajani, MD,a Atlanta, Ga

ABSTRACT Objective: Preoperative clinical frailty is increasingly used as a surrogate for predicting postoperative outcomes. Patients undergoing major lower extremity amputation (LEA) carry a high risk of perioperative morbidity and mortality, including high 30-day mortality and readmission rates. We hypothesized that preoperative frailty would be associated with an increased risk of postoperative mortality and readmission. Methods: A retrospective review was performed for all patients who underwent transfemoral or transtibial amputation for any indication within a multi-institution system during a 5-year period. Standard demographics and all components of the Modified Frailty Index (mFI) were used to determine preoperative frailty status for each patient. The primary outcome was 30-day mortality, with secondary outcomes of 30-day readmission, unplanned revision, and composite adverse events. Results: Among 379 patients who underwent LEA, the overall readmission and mortality rates for the group were 22.69% and 6.06%, respectively. Readmission rates increased with increasing mFI score: rates were 8.6%, 13.5%, 16.3%, 19.7%, 31.4%, and 37.0% for mFI scores of 0, 1, 2, 3, 4, and $5, respectively (P ¼ .015). On multivariate logistic regression, only mFI (odds ratio, 1.49, 95% confidence interval, 1.24-1.77) and sex (odds ratio, 1.81, 95% confidence interval, 1.00-2.98) were significant predictors of 30-day readmission. Conclusions: Preoperative clinical frailty is associated with an increased 30-day readmission rate in patients undergoing LEA and should be incorporated into preoperative counseling and risk stratification, as well as postoperative planning and care. (J Vasc Surg 2017;65:804-11.)

Patients who undergo major lower extremity amputation (LEA) often have multiple comorbidities, and the prognosis is frequently poor, with mortality rates of 22% at 30 days and 44% at 1 year.1 The long-term outlook is perhaps equally dismal: only 40% of patients who undergo amputation at the transtibial regain full mobility after 2 years, and 5-year mortality rates are as high as 77%.1,2 Previous investigations of prognostic factors have been limited to comparisons between From the Division of Vascular and Endovascular Surgery, Department of Surgery,a and Department of Surgery, Department of Hematology and Medical Oncology, Winship Cancer Institute,b Emory University. This study was supported by the National Center for Advancing Translational Sciences of the National Institutes of Health under award number UL1-TR000454. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Author conflict of interest: none. Presented in the poster competition at the 2016 Vascular Annual Meeting of the Society for Vascular Surgery, National Harbor, Md, June 8-11, 2016. Correspondence: Zachary B. Fang, BS, c/o Ravi R. Rajani, Department of Surgery, Glenn Bldg, 69 Jesse Hill Dr #304, Atlanta, GA 30303 (e-mail: zbfang@ emory.edu). The editors and reviewers of this article have no relevant financial relationships to disclose per the JVS policy that requires reviewers to decline review of any manuscript for which they may have a conflict of interest. 0741-5214 Copyright Ó 2016 by the Society for Vascular Surgery. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvs.2016.10.102

804

surgical subspecialty fields and basic demographics. This is a potentially heterogeneous group of patients, many of whom suffer from chronic wasting, dysvascular syndromes, or other underlying conditions that contribute to an overall frail state. Frailty in medicine is scientifically defined as a biological syndrome that reflects a state of decreased physiologic reserve.3 The exact pathophysiology is currently unknown, but proposed mechanisms include dysregulation of hormones and cytokines in the aging body, accumulation of insults to different organ systems caused by disease, and lifelong wear and tear.3 For surgical fields, frailty is rapidly emerging as a potential method of risk-stratifying patients, and research is being performed to validate frailty assessments across a growing variety of subspecialties, operations, and populations.4-7 To date, frailty status has been identified as a predictor of poor outcomes in colorectal, cardiovascular, and gynecologic surgical procedures. Outcomes such as mortality, increased 30-day readmission rates, and a variety of other postoperative complications have been consistently correlated with an increase in frailty across different studies.8-17 These findings have led to recognition across a variety of fields that frail patients have worse outcomes than the nonfrail and that early identification of frailty is an important step in

Journal of Vascular Surgery

Fang et al

805

Volume 65, Number 3

determining treatment, predicting results, and designing effective interventions. One method that can potentially be used to evaluate frailty is the Modified Frailty Index (mFI), a previously validated retrospective tool that was developed using data from the Canadian Study on Health and Aging.18 Although a substantial portion of the patient population that undergoes LEA may meet most frailty criteria, no studies have provided evidence for a relationship between frailty and perioperative outcomes after amputation. In view of the significant 30-day and 1-year mortality rates associated with LEA overall, an assessment of frailty in these patients would be an important step toward the future use of frailty indices to risk stratify these patients and design potentially effective interventions to improve outcomes. Ideally, interventions that address the underlying causes of frailty could be implemented. Alternatively, information regarding postoperative mortality and morbidity may be of clinical significance at the time of patient counseling: a frail patient with a predicted poor outcome may be better served by end-of-life care discussions with subsequent hospice services. Thus, the objective of this study was to investigate the prevalence of frailty and to examine the relationship between preoperative frailty status and postoperative morbidity and mortality in patients undergoing LEA.

METHODS Study population. All patients who underwent transtibial or transfemoral amputations at a multihospital academic institution between December 2010 and March 2015 were identified. Patients aged <18 years were excluded. Patient medical records were interrogated to generate a frailty score using the mFI. Other sociodemographic variables collected included age, race, ethnicity, insurance status, employment status, and gender. Data sources. Data were collected retrospectively by medical record review from prospectively maintained data sets. This represents composite data from a single vascular surgery academic group that covers five hospitals, including a major academic center, multiple community hospitals, and a large urban community hospital. Institutional Research Board approvals, along with a Health Insurance Portability and Accountability Act waiver, were obtained and maintained in active status throughout the conduct of this project. Institutional Review Board waivers for consent were obtained from all sites for the duration of this study. Analysis. As previously described, the 11 historical parameters of the mFI were used to generate a frailty score: each component of the mFI is worth 1 point, and the maximum score is 11 (Table I).18 Briefly, the mFI is a retrospective method of assessment that measures frailty as an accumulation of deficits and can be

ARTICLE HIGHLIGHTS d

d

d

d

d

d

Significance: This is a novel manuscript that describes frailty assessments in a single-center cohort of major lower extremity amputations. Type of Research: Retrospective multicenter observational cohort study. Take Home Message: Increasing preoperative Modified Frailty Index in 379 patients with major amputations predicted higher 30-day readmission rates but not a higher 30-day mortality. Recommendation: The authors suggest that the Modified Frailty Index may help to risk stratify patients who are undergoing major lower extremity amputations and help with postoperative planning and care. Strength of Recommendation: 2. Weak. Level of Evidence: C. Low or very low.

performed in 2 minutes. It has been previously validated in large, national databases as well as in vascular surgery patient populations, making it an ideal instrument for this study. Components of the mFI were classified as present if they were documented in the medical record. Patients were determined to have an impaired functional status component if it was documented explicitly or if they were documented to live in a facility that provided support for activities of daily living (such as a nursing home). Impaired sensorium included acute and chronic delirium. The primary outcome was all-cause mortality #30 days from a patient’s last amputation, or whichever amputation was intended to be his or her last (as noted in the medical record). Secondary outcomes included unplanned revision, surgical site infection (SSI), stroke, renal failure, prolonged ventilation, sepsis, deep vein thrombosis, myocardial infarction, 30-day readmission, and 1year mortality. Unplanned revision was defined as a return to the operating room for any reason related to the amputation stump. Outcomes other than 1-year mortality were classified as existing if they occurred #30 days from amputation. The definition for 30-day readmission was an unplanned readmission to any participating sites #30 days of discharge. One-year mortality data were collected from the medical record. A composite adverse events end point of SSI, stroke, renal failure, prolonged ventilation, sepsis, deep vein thrombosis, and myocardial infarction was generated to account for low event rates of each of the included outcomes. Patients who were considered lost to follow-up were censored at the time of their last recorded contact with the health care system. Data were analyzed by descriptive and univariate statistics and then by inferential statistics. The differences in specific outcomes among subject groups based on

806

Journal of Vascular Surgery

Fang et al

March 2017

Table I. Components of the Modified Frailty Index (mFI)a CSHA component

mFI component

Point value

Impaired functional statusdpartially or totally dependent

1

History of COPD or current pneumonia

1

Congestive heart failure

Congestive heart failure present in past 30 days

1

Myocardial infarction

History of myocardial infarction

1

Cardiac problems Cardiac disease

History of cardiac surgery History of percutaneous coronary intervention Angina in past 30 days

1

Hypertension

Hypertension requiring medication

1

Clouding or delirium History relevant to cognitive impairment or loss Family history relevant to cognitive impairment

Impaired sensorium

1

Cerebrovascular problems

History of transient ischemic attack

1

History of stroke

Cerebrovascular accident or stroke with neurologic deficit

1

Decreased peripheral pulses

PVD or rest pain History of revascularization

1

History of diabetes mellitus

Diabetes mellitus

Total

Total

Problems Problems Problems Problems Problems

with with with with with

dressing bathing personal grooming cooking going out alone

Chronic/acute respiratory problems Lung problems

1 11

CSHA, Canadian Study of Health and Aging; COPD, chronic obstructive pulmonary disease; PVD, peripheral vascular disease. a Velanovich et al.18

demographics, clinical characteristics, facility, and other variables were examined using c2 and t-tests. The risk of specific outcome variables (eg, readmission and mortality) were also examined, and logistic regression models were used to identify factors associated with patient outcomes. Multivariate logistic regression was performed comparing mFI score, sex, age, race, employment status, insurance status, amputation type, and the components of the mFI. Multivariate models were generated using stepwise selection with entry and exit criteria of 0.05. The effect of mFI on significant outcomes was assessed as an ordinal and as a categoric variable to confirm the linearity of the relationship. Kaplan-Meier survival curves were constructed to examine crude survival differences between individuals with differing mFI scores. The use of mFI >2 as a cutoff for frailty has been described previously.19 This concept was applied to the Kaplan-Meier curves in an effort to provide a more easily interpretable survival analysis as well as to increase the statistical power by combining multiple mFI levels.

RESULTS A total of 379 patients (64.0% male) who underwent major LEA were identified and included in the analysis (Table II). The mean patient age was 59.1 6 15.0 years. The mean number of points scored on the mFI was 2.9 6 1.7 (range, 0-8; mode, 2; median, 3). The most common indications for amputation were ischemia/tissue

loss (64.6%) and infection (18.2%). A transtibial amputation was performed in 51.8% of patients as the definitive treatment for their presenting clinical condition (Table II). When a truncated mFI from 0 to 5 was used to account for low numbers of high-scoring patients, the most common mFI score was 2 (n ¼ 98), followed by $5 (n ¼ 73), 4 (n ¼ 70), 3 (n ¼ 66), 1 (n ¼ 37), and 0 (n ¼ 35; Table III). The most common mFI component present was hypertension (78.9%), followed by diabetes mellitus (54.9%) and peripheral vascular disease (PVD; 54.1%). Association between the mFI and 30-day mortality. The overall mortality at 30 days was 22.5%. Multivariate logistic regression found that only a history of chronic obstructive pulmonary disease (odds ratio [OR], 6.382; 95% confidence interval [CI], 2.194-18.564; P ¼ .0007) and impaired sensorium (OR, 4.605; 95% CI, 1.704-12.441; P ¼ .0026) were significant predictors of 30-day mortality. Univariate and multivariate analysis did not show an association between mFI and 30-day mortality. Association between the mFI and additional outcomes. The overall 30-day readmission rate for the cohort was 22.7%, with increasing mFI score corresponding to increasing rates of readmission (c2 ¼ 18.1158; P ¼ .0028; Table IV). Nonwound-related issues resulted in 71.1% of readmissions, followed by wound-related issues in 28.9%. Among the wound-related issues were SSI, wound dehiscence, and poor wound healing. On

Journal of Vascular Surgery

Fang et al

807

Volume 65, Number 3

Table II. Baseline demographics of patients undergoing major lower extremity amputation (LEA) from 2010 to 2015 stratified by Modified Frailty Index (mFI) score mFI score Parameter

Totals (N ¼ 379)

0 (n ¼ 35)

1 (n ¼ 37)

2 (n ¼ 98)

3 (n ¼ 66)

4 (n ¼ 70)

$5 (n ¼ 73)

Age, years

59 6 15

37 6 14

54 6 15

56 6 12

62 6 13

66 6 12

66 6 12

a

Male

243 (64)

24 (69)

30 (81)

65 (66)

46 (70)

39 (53)

44 (60)

Race Caucasian

104 (27)

16 (46)

11 (30)

26 (27)

16 (24)

18 (24)

19 (26)

African American

265 (70)

18 (51)

26 (70)

70 (71)

47 (71)

54 (73)

57 (78)

10 (3)

1 (3)

0 (0)

2 (2)

3 (5)

2 (3)

2 (3)

Uninsured

83 (22)

14 (40)

10 (27)

28 (29)

11 (17)

10 (14)

10 (14)

Medicaid

64 (17)

5 (14)

9 (24)

17 (17)

13 (20)

9 (12)

11 (15)

Medicare

28 (38)

Other Insurance status

108 (29)

0 (0)

5 (14)

26 (27)

24 (36)

30 (41)

Private

72 (19)

15 (43)

11 (30)

18 (18)

8 (12)

14 (19)

8 (11)

Medicaid/Medicare

52 (14)

1 (3)

2 (5)

9 (9)

10 (15)

11 (15)

21 (29)

Employment status Employed

61 (16)

16 (46)

10 (27)

17 (17)

7 (11)

9 (12)

4 (5)

Unemployed

133 (35)

12 (34)

19 (51)

43 (44)

21 (32)

16 (22)

23 (32)

Retired

135 (36)

1 (3)

5 (14)

27 (28)

26 (39)

39 (53)

42 (58)

41 (11)

2 (6)

3 (8)

10 (10)

9 (14)

10 (14)

8 (11)

9 (2)

4 (11)

0 (0)

1 (1)

3 (5)

0 (0)

1 (1)

179 (47)

18 (51)

19 (51)

33 (34)

26 (40)

42 (57)

47 (64)

200 (53)

17 (49)

18 (49)

65 (66)

40 (60)

32 (43)

31 (42)

Disabled Other Operation Transfemoral Transtibial a

Continuous data are shown as the mean 6 standard deviation and categoric data as number (%).

Table III. Baseline Modified Frailty Index (mFI) components of patients undergoing major lower extremity amputation (LEA) from 2010 to 2015 stratified by the mFI score mFI score

Totals (N ¼ 379), No. (%)

0 (n ¼ 35), No. (%)

1 (n ¼ 37), No. (%)

2 (n ¼ 98), No. (%)

3 (n ¼ 66), No. (%)

4 (n ¼ 70), No. (%)

$5 (n ¼ 73), No. (%)

57 (15)

0 (0)

0 (0)

6 (6)

5 (8)

19 (27)

32 (44)

COPD or current pneumonia

28 (7)

0 (0)

3 (8)

2 (2)

7 (11)

8 (11)

11 (15)

Percutaneous coronary intervention/cardiac surgery/angina

68 (18)

0 (0)

1 (3)

2 (2)

4 (6)

25 (36)

38 (52)

Transient ischemic attack

10 (3)

0 (0)

0 (0)

1 (1)

1 (2)

2 (3)

7 (10)

Cerebrovascular accident

60 (16)

0 (0)

1 (3)

1 (1)

10 (15)

18 (26)

34 (47)

Congestive heart failure

81 (21)

0 (0)

1 (3)

3 (3)

16 (24)

23 (33)

40 (55)

Myocardial infarction

49 (13)

0 (0)

0 (0)

2 (2)

4 (6)

12 (17)

33 (45)

299 (79)

0 (0)

18 (49)

87 (89)

62 (94)

67 (96)

72 (99)

43 (11)

0 (0)

0 (0)

2 (2)

8 (12)

15 (21)

22 (30)

Parameter Impaired functional status History of

Hypertension Impaired sensorium PVD

205 (54)

0 (0)

4 (11)

32 (33)

39 (59)

64 (91)

66 (90)

Diabetes mellitus

208 (55)

0 (0)

9 (24)

58 (59)

42 (64)

45 (64)

61 (84)

COPD, Chronic obstructive pulmonary disease; PVD, peripheral vascular disease.

multivariate logistic regression, only mFI (OR, 1.49; 95% CI, 1.24-1.77) and sex (OR, 1.81; 95% CI, 1.00-2.98) were significant predictors of 30-day readmission (C ¼ 0.671). A multivariable logistic regression model that included

mFI, sex, and age increased the OR estimate for mFI to 1.51 (95% CI, 1.25-1.83; C ¼ 0.671; Table V). Congestive heart failure (OR, 2.21; 95% CI, 1.27-3.85) and PVD (OR, 1.95; 95% CI, 1.15-3.31) were found to be the main drivers of 30-day

808

Journal of Vascular Surgery

Fang et al

March 2017

Table IV. Outcomes after major lower extremity amputation (LEA) stratified by points scored on the Modified Frailty Index (mFI) mFI score Parameter

Totals, No. (N ¼ 379)

0 (n ¼ 35), No. (%)

1 (n ¼ 37), No. (%)

2 (n ¼ 98), No. (%)

3 (n ¼ 66), No. (%)

4 (n ¼ 70), No. (%)

$5 (n ¼ 73), No. (%)

P valuea .4282

30-day mortality

23

2 (5.7)

3 (8.1)

4 (4.1)

3 (4.6)

8 (11.4)

3 (4.1)

30-day readmission

86

3 (8.6)

5 (13.5)

16 (16.3)

13 (19.7)

22 (31.4)

27 (37.0)

.0013

Unplanned revisions

105

10 (28.6)

13 (35.1)

24 (24.5)

13 (19.7)

23 (32.9)

22 (30.1)

.4414

Adverse eventsb

278

23 (65.7)

28 (75.7)

68 (69.4)

44 (66.7)

56 (80.0)

59 (80.8)

.2126

P values correspond to c2 test for homogeneity or the Fisher exact test, where appropriate. Values in bold indicate statistical significance (P < .05). b Adverse events was defined as any postoperative incidence of 30-day mortality, 30-day readmission, unplanned revisions, surgical site infection (SSI), stroke, renal failure, prolonged ventilation, sepsis, deep vein thrombosis, and myocardial infarction. a

Table V. Univariate and adjusted odds ratios (ORs) for the effect of Modified Frailty Index (mFI) score on outcomes OR for mFI (95% CI) Outcome

Univariate

P valuea

Adjustedb

P valuea

.8136

0.867 (0.630-1.193)

.3802

<.0001

1.510 (1.245-1.832)

<.0001

30-day mortality

1.033 (0.787-1.358)

30-day readmission

1.447 (1.219-1.717)

Unplanned revisions

1.022 (0.884-1.181)

.7731

1.156 (0.975-1.371)

.0960

1.155 (0.997-1.338)

.0550

1.274 (1.069-1.517)

.0068

Adverse events

CI, Confidence interval. a Values in bold indicate statistical significance (P < .05). b Adjusted for age and sex.

readmission using a multivariate analysis with the components of the mFI instead of the composite score. To determine whether the increase in 30-day readmission was nonlinear with respect to the mFI score, a multivariate analysis with mFI as a categoric variable, age, and sex was also performed. Using this model resulted in an increase in the OR to 1.65 (95% CI, 0.35-7.81) from 0 to 1 point, 1.36 (95% CI, 0.46-4.07) from 1 to 2 points, 1.26 (95% CI, 0.55-2.87) from 2 to 3 points, 2.13 (95% CI, 0.95-4.78) from 3 to 4 points, and 1.24 (95% CI, 0.61-2.50) from 4 to $5 points. Consistent with an expected linear sample, these findings were not statistically significant at the a ¼ .05 level. The effect of mFI on outcomes was assessed as both an ordinal and categoric variable to confirm the linearity of the relationship. Multivariate analysis was used to assess the increase in effect between each level of mFI (ie, as mFI increased from 1 to 2, from 2 to 3, and so on) and confirmed that effect sizes did not differ significantly. The overall unplanned revision rate was 27.7%. No association was found between mFI and the occurrence of unplanned revisions on univariate analysis. Multivariate logistic regression found that only age (OR, 0.981; 95% CI, 0.966-0.996) and previous myocardial infarction (OR, 2.420; 95% CI, 1.292-4.534) were significant predictors of unplanned revisions. Univariate analysis revealed no association between mFI and composite adverse events. On multivariate analysis, however, previous myocardial infarction (OR, 4.897; 95% CI, 1.452-16.523) and a previous

cardiac procedure (OR, 2.317; 95% CI, 1.038-5.172) were both associated with the composite adverse events outcome. Survival analysis. Kaplan-Meier survival curves of major LEA patients stratified by mFI score showed decreased survival for patients who scored 3, 4, or 5 points compared with 1, 2, or 3 points (Wilcoxon test statistic ¼ 14.2491; P ¼ .0141; Fig 1). An adjusted survival curve separating patients into those with mFI #2 and >2 showed decreased survival in the mFI >2 group (Wilcoxon test statistic ¼ 8.3731; P ¼ .0038; Fig 2).

DISCUSSION This large single-center series of 379 patients found preoperative frailty status was a strong predictor of 30-day readmission after major LEA. With the current focus on frailty as a risk-stratification tool, these findings add to the growing body of evidence across multiple surgical specialties that the effects of frailty are measurable and meaningful as related to postoperative outcomes. The association between frailty and 30-day readmission is consistent with previous findings in patients undergoing colorectal or cardiac procedures.15 Not all of these findings, however, are consistent with the literature: multiple studies focused on vascular surgery populations have found significant relationships between frailty and 30-day mortality and composite adverse outcomes.5,7,19,20

Journal of Vascular Surgery

Fang et al

809

Volume 65, Number 3

Fig 1. Survival analysis of major lower extremity amputation (LEA) patients stratified by the Modified Frailty Index (mFI) score. Survival time is measured in days. Wilcoxon test statistic ¼ 14.2491; P ¼ .0141. mFI_trunc, Truncated mFI score. For mFI ¼ 0, there were 35 patients at 0 days, 33 patients at 30 days, and 19 patients at 365 days. The number of censored patients was 14 at the end of the time period. For mFI ¼ 1, there were 37 patients at 0 days, 33 patients at 30 days, and 21 patients at 365 days. The number of censored patients was 12 at the end of the time period. For mFI ¼ 2, there were 98 patients at 0 days, 93 patients at 30 days, and 61 patients at 365 days. The number of censored patients was 28 at the end of the time period. For mFI ¼ 3, there were 66 patients at 0 days, 62 patients at 30 days, and 39 patients at 365 days. The number of censored patients was 18 at the end of the time period. For mFI ¼ 4, there were 70 patients at 0 days, 61 patients at 30 days, and 33 patients at 365 days. The number of censored patients was 18 at the end of the time period. For mFI $5, there were 73 patients at 0 days, 70 patients at 30 days, and 38 patients at 365 days. The number of censored patients was 20 at the end of the time period. The standard error did not exceed 10% for the specified time period.

A possible reason for the lack of detected effect is variability in the definition of adverse events among the studies, especially the inclusion of events such as 30-day mortality and the breakdown of outcomes by different classification schemes. Several variables of interest that could have an effect on the relationship between mortality and mFI were not available in our data set, including presence of infection at the time of surgery, wound size, hemoglobin A1c, smoking status, and nutritional status. Our analysis also did not account for bilateral LEAs because they represented <5% of our population. These missing variables could also explain the mild protective effect of age against unplanned revision. To our knowledge, this is the first study to examine the effect of frailty on postoperative outcomes in the population undergoing major LEA. These data do not depend on a large national patient data series and thus may not be reflective of national trends; consequently, the presented findings may be more applicable in a similar academic setting. Although large databases have significant value in increased power and sample size, a single

academic center offers granular detail that may not be available in a national study. We consider the diverse indications for amputation to be one of the strengths of this study. Although a relatively small percentage of the study population underwent amputation for trauma-related reasons, almost half of the overall population had mFI scores of #2, mostly attributed to hypertension or diabetes. The traumatic amputation group was included to provide a better representation of the population undergoing amputations as a whole as well as to reflect that vascular, general, and orthopedic surgeons at our medical center perform these operations. For similar reasons, we decided not to limit our analysis to only patients with PVD. We cannot rule out the possibility that some of the patients with diabetes also had PVD (and vice versa). Although this could be interpreted as a weakness of using a retrospective measure of frailty such as the mFI, we believe that it is actually a strength: rather than simply comparing dysvascular to nondysvascular patients, we acknowledge that this is a heterogeneous

810

Fang et al

Journal of Vascular Surgery March 2017

Fig 2. Survival analysis of major lower extremity amputation (LEA) patients stratified by the modified Frailty Index (mFI) score #2 and >2. Survival time is measured in days. Wilcoxon test statistic ¼ 8.3731; P ¼ .0038. bi_mfi of 0 signifies a mFI less than or equal to 2, while bi_mfi of 1 signifies a mFI greater than 2. For mFI #2, there were 170 patients at 0 days, 160 at 30 days, and 103 patients at 365 days. The number of censored patients was 52. For mFI >2, there were 209 patients at 0 days, 193 at 30 days, and 112 at 365 days. The number of censored patients was 54. The standard error did not exceed 10% for the specified time period.

patient population and use one unifying concept to assess risk: frailty. This study adds to the growing body of evidence that frailty is a viable method of risk stratification, even in populations considered to be somewhat saturated with high-risk patients. The retrospective nature of this study places some constraints on the definitions of several variables of the mFI, most notably patient independence and recent angina. These factors were taken into account when noted in the medical record, but a prospective study would conceivably have much greater fidelity of patient factors. The ease with which the mFI can be implemented, however, is a powerful argument for its widespread adaptation in medicine. Instead of the 15 to 30 minutes required to assess a patient’s grip strength and walk time, which other scales use, the mFI score can be calculated in w2 minutes or less. Furthermore, because historical data are used to calculate mFI, a physician with a busy clinic could assess a patient at his or her convenience. However, the mFI needs to be validated in prospective trials before it can be used in such a setting. Whether the accumulation of deficits model of frailty or one using phenotypic measurements will emerge as the most effective method of frailty measurement remains unclear at this time, but a simple, easy, and quickly executed method is more likely to be widely adopted.

As stated above, the relatively small sample size of patients available may have contributed to lack of significance in several expected relationships and to inadequate power for subgroup analysis. The short follow-up period available in the data set limits the strength of inferences beyond 30 days. Another concern is that the mFI does not account for differing levels of severity within components. For example, a patient with mild memory loss would receive the same score as a delirious patient. This could cause the mFI to have decreased discriminatory power when most of a patient’s morbidities are cognitive in nature. PVD is highly prevalent in the population of patients undergoing LEA, and these patients often have much comorbidity. It is thus possible that the results are somewhat skewed towards worse outcomes, even though this patient sample included traumatic amputations as well as amputations done for diabetic foot infections. Another drawback to the retrospective nature of this study is inconsistency in the medical record. Because different surgeons from various specialties may perform amputations at our institution, there were sometimes incomplete or different noted indications for patients between operations. However, all amputations were included for a more complete assessment of the patient population.

Journal of Vascular Surgery

Fang et al

811

Volume 65, Number 3

CONCLUSIONS The association between frailty and 30-day readmission should be given extra weight when confronted with the frail patient who requires an amputation. Although it is widely known that these patients bear some of the heaviest burdens in terms of comorbidities, there is often little special consideration of the needs of new amputees while in the hospital. Easing the transition to home, either with enhanced recovery programs or some other focused follow-up initiative, should be considered by surgeons before discharging such patients. In addition, the evidence added by this study strengthens the argument that reimbursement guidelines should take patient frailty status into account because frail patients clearly have a higher likelihood of returning in need of further medical care. Surgeons who appropriately operate on high-risk patients should not be penalized for doing so, but identification of such patients will be a challenge moving forward. To date, reimbursement levels for operations with expected complication rates have been procedurally generated, but in reality, preoperative factors may play a much more important role in the outcomes than technical aspects. Ideally, these data will aid physicians in determining which patients will benefit most from additional postoperative support as well as assist those who set reimbursement levels for specific patient populations and procedures.

AUTHOR CONTRIBUTIONS Conception and design: ZF, SA, TG, RR Analysis and interpretation: ZF, FH, SA, TG, RR Data collection: ZF, TG, RR Writing the article: ZF, FH, TG, RR Critical revision of the article: ZF, FH, SA, TG, RR Final approval of the article: ZF, FH, SA, TG, RR Statistical analysis: ZF, FH Obtained funding: ZF, TG Overall responsibility: ZF

REFERENCES 1. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG, et al. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). J Vasc Surg 2007;45(Suppl S):S5-67. 2. Fortington LV, Geertzen JH, van Netten JJ, Postema K, Rommers GM, Dijkstra PU. Short and long term mortality rates after a lower limb amputation. Eur J Vasc Endovasc Surg 2013;46:124-31. 3. Afilalo J, Alexander KP, Mack MJ, Maurer MS, Green P, Allen LA, et al. Frailty assessment in the cardiovascular care of older adults. J Am Coll Cardiol 2014;63:747-62. 4. Makary MA, Segev DL, Pronovost PJ, Syin D, BandeenRoche K, Patel P, et al. Frailty as a predictor of surgical outcomes in older patients. J Am Coll Surg 2010;210:901-8. 5. Arya S, Kim SI, Duwayri Y, Brewster LP, Veeraswamy R, Salam A, et al. Frailty increases the risk of 30-day mortality, morbidity, and failure to rescue after elective abdominal

aortic aneurysm repair independent of age and comorbidities. J Vasc Surg 2015;61:324-31. 6. Partridge JS, Harari D, Dhesi JK. Frailty in the older surgical patient: a review. Age Ageing 2012;41:142-7. 7. Karam J, Tsiouris A, Shepard A, Velanovich V, Rubinfeld I. Simplified frailty index to predict adverse outcomes and mortality in vascular surgery patients. Ann Vasc Surg 2013;27: 904-8. 8. Adams P, Ghanem T, Stachler R, Hall F, Velanovich V, Rubinfeld I. Frailty as a predictor of morbidity and mortality in inpatient head and neck surgery. JAMA Otolaryngol Head Neck Surg 2013;139:783-9. 9. Hodari A, Hammoud ZT, Borgi JF, Tsiouris A, Rubinfeld IS. Assessment of morbidity and mortality after esophagectomy using a modified frailty index. Ann Thorac Surg 2013;96: 1240-5. 10. Courtney-Brooks M, Tellawi AR, Scalici J, Duska LR, Jazaeri AA, Modesitt SC, et al. Frailty: an outcome predictor for elderly gynecologic oncology patients. Gynecol Oncol 2012;126:20-4. 11. Dasgupta M, Rolfson DB, Stolee P, Borrie MJ, Speechley M. Frailty is associated with postoperative complications in older adults with medical problems. Arch Gerontol Geriatr 2009;48:78-83. 12. McAdams-DeMarco MA, Law A, Salter ML, Chow E, Grams M, Walston J, et al. Frailty and early hospital readmission after kidney transplantation. Am J Transplant 2013;13:2091-5. 13. Keller DS, Bankwitz B, Nobel T, Delaney CP. Using frailty to predict who will fail early discharge after laparoscopic colorectal surgery with an established recovery pathway. Dis Colon Rectum 2014;57:337-42. 14. Tan KY, Kawamura YJ, Tokomitsu A, Tang T. Assessment for frailty is useful for predicting morbidity in elderly patients undergoing colorectal cancer resection whose comorbidities are already optimized. Am J Surg 2012;204: 139-43. 15. Robinson TN, Wu DS, Pointer L, Dunn CL, Cleveland JC, Moss M. Simple frailty score predicts postoperative complications across surgical specialties. Am J Surg 2013;206: 544-50. 16. Obeid NM, Azuh O, Reddy S, Webb S, Reickert C, Velanovich V, et al. Predictors of critical care-related complications in colectomy patients using the National Surgical Quality Improvement Program: exploring frailty and aggressive laparoscopic approaches. J Trauma Acute Care Surg 2012;72:878-83. 17. Revenig LM, Canter DJ, Master VA, Maithel SK, Kooby DA, Pattaras JG, et al. A prospective study examining the association between preoperative frailty and postoperative complications in patients undergoing minimally invasive surgery. J Endourol 2014;28:476-80. 18. Velanovich V, Antoine H, Swartz A, Peters D, Rubinfeld I. Accumulating deficits model of frailty and postoperative mortality and morbidity: its application to a national database. J Surg Res 2013;183:104-10. 19. Brahmbhatt R, Brewster LP, Shafii S, Rajani RR, Veeraswamy R, Salam A, et al. Gender and frailty predict poor outcomes in infrainguinal vascular surgery. J Surg Res 2016;201:156-65. 20. Ehlert BA, Najafian A, Orion KC, Malas MB, Black JH, Abularrage CJ. Validation of a modified Frailty Index to predict mortality in vascular surgery patients. J Vasc Surg 2016;63:1595-601.

Submitted Aug 18, 2016; accepted Oct 17, 2016.