- Email: [email protected]
Readmission rates after lower extremity bypass vary significantly by surgical indication
Readmission rates after lower extremity bypass vary significantly by surgical indication Caroline E. Jones, MD, Joshua S. Richman, MD, PhD, Daniel I. Chu, MD, Allison A. Gullick, MSPH, Benjamin J. Pearce, MD, and Melanie S. Morris, MD, Birmingham, Ala Objective: Readmission rates after vascular surgery are among the highest within surgical specialties, and lower extremity bypass has the highest readmission rate of vascular surgery procedures. We analyzed how 30-day readmissions and risk factors for readmissions vary by indication for lower extremity bypass. Methods: We queried the 2012-2014 American College of Surgeons National Surgical Quality Improvement Program procedure-targeted vascular cohort to identify all patients who underwent lower extremity bypass. Emergent procedures and planned readmissions were excluded. Patients were stratified by surgical indication: claudication, critical limb ischemia rest pain (CLI RP), critical limb ischemia tissue loss (CLI TL), and other. The c2 and Wilcoxon rank sum tests were used to test the differences between categorical and continuous variables, respectively. Logistic regression was used to estimate odds ratios for predictors of readmission adjusted for preoperative factors that were selected a priori. Results: The overall 30-day readmission rate among the 6112 patients who underwent lower extremity bypass was 14.8%. Readmission rates varied significantly on the basis of the indication for surgery. In unadjusted comparisons, 18.8% of patients with CLI TL were readmitted compared with 16.5% with CLI RP, 9.4% with claudication, and 8.2% with other indications (P < .001). After adjustment for preoperative factors, 30-day readmissions were higher for patients with CLI TL (odds ratio, 1.67; 95% confidence interval, 1.35-2.06) and CLI RP (odds ratio, 1.70; 95% confidence interval, 1.38-2.09) compared with patients with claudication. Conclusions: The 30-day readmission rates after lower extremity bypass vary significantly by surgical indication. Because lower extremity bypasses are performed for multiple indications, if readmission rates are publically reported and hospitals can be penalized for higher than expected readmission rates, the expected readmission rates should be adjusted for surgical indication. (J Vasc Surg 2016;-:1-7.)
Unplanned readmission rates after vascular surgery are among the highest within surgical specialties. Jencks et al reported a 23.9% unplanned 30-day readmission rate after vascular surgery, substantially higher than the overall surgical readmission rate of 15.6%.1 Vascular surgery readmissions ranked third highest of any diagnosis-related group, behind only heart failure (26.9%) and psychoses (24.6%).1 Among vascular surgical procedures, lower extremity bypass is associated with the highest rate of unplanned readmissions.2 Postoperative readmissions are costly and associated with worse outcomes. The estimated cost of readmissions is $12 billion per year and accounts for approximately 25% of all Medicare expenditures for inpatient care. The average cost per readmission after lower extremity bypass is $10,000 and adds an additional 39% to the index admission cost.1,3 Patients readmitted after lower extremity From the Department of Surgery, University of Alabama-Birmingham. Author conflict of interest: none. This abstract was selected for an oral presentation at the Academic Surgical Congress, Jacksonville, Fla, February 2-4, 2016. Correspondence: Melanie S. Morris, MD, Department of Surgery, University of Alabama-Birmingham, 1922 7th Ave S, Kracke Bldg, Birmingham, AL 35294 (e-mail: [email protected]). 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.2016.03.422
bypass also experience worse outcomes with significantly lower rates of limb salvage at 1 year (78.3% vs 89.6%), higher rates of reoperation (47.4% vs 6.8%), and increased mortality (1.9% vs 0.3%).2 Current literature recognizes wound infection and graft complication as the main reasons for readmission after lower extremity bypass.4 Other recognized risk factors include female gender, smoking, obesity, high American Society of Anesthesiologists (ASA) class, history of cardiac events, chronic obstructive pulmonary disease, renal failure requiring dialysis, and preoperative functional status.2,5 However, the association between the indication for a lower extremity revascularization procedure and risk for readmission is unknown. This relationship is important as previous evidence suggests a potential correlation between preoperative symptoms and complications.6 We examined the relationship between readmission and the main indications for lower extremity bypass: claudication, critical limb ischemia rest pain (CLI RP), and critical limb ischemia tissue loss (CLI TL). We hypothesized that readmission rates after lower extremity bypass differ on the basis of the indication for surgery after adjustment for covariates. METHODS Data set. We queried the 2012-2014 American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) procedure-targeted vascular cohort. The ACS NSQIP reports perioperative variables 1
JOURNAL OF VASCULAR SURGERY --- 2016
2 Jones et al
and surgical outcomes for >400 participating academic and community hospitals within the United States. The ACS NSQIP is a well-validated prospective database that accurately captures readmissions with highly effective data collection and inter-rating reliability.7,8 Clinical nurse training and audit procedures ensure robust data that have improved in reliability every year.9 Patient informed consent was determined to be unnecessary, and Institutional Review Board approval was waived. Patient cohort. Within the ACS NSQIP proceduretargeted vascular cohort, we identified all patients who underwent a lower extremity bypass by the procedural terminology code for lower extremity bypassdopen. This procedural code encompasses all femoral-distal, femoralpopliteal, and popliteal-distal bypasses with vein or prosthetic conduit, femoral endarterectomies, and profundoplasties using Current Procedural Terminology codes 35556, 35566, 35571, 35583, 35585, 35587, 35656, 35666, and 35671. We did not exclude patients with concomitant inflow procedures. We excluded all emergent cases, patients with an ASA class of 5, patients experiencing postoperative length of stay >10 days, and patients who died within 30 days of the index operation. We excluded all planned readmissions and all patients who underwent amputation during their readmission, as this most likely represented a planned readmission awaiting tissue demarcation. Patients were stratified by the NSQIPreported symptoms claudication, CLI RP, CLI TL, and other, including diagnoses of aneurysm, embolism, and technical cause (ie, graft complication). Outcomes. Primary outcome was 30-day unplanned readmission. Secondary outcomes were 14-day readmission and 30-day postoperative complications. Unplanned readmission was defined by the ACS NSQIP as “unplanned readmission (to the same or another hospital) for a postoperative occurrence likely related to the principal surgical procedure within 30 days of the procedure.” The surgical clinical reviewer categorized readmissions as planned or unplanned on the basis of whether the readmission was known at time of discharge from index hospitalization. It is especially important to target unplanned readmissions in this cohort because approximately 25% of readmissions in vascular surgery patients are planned.10 Postoperative complications included those identified by the ACS NSQIP and limited to 30 days postoperatively. Statistical analysis. The results were examined with procedure as the unit of analysis. Analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC). The c2 and Wilcoxon rank sum tests were used to determine the differences among categorical and continuous variables, respectively. The association between readmission and surgical indication was further assessed by logistic regression adjusted for the following preoperative factors that were selected a priori: age, sex, race, body mass index, hypertension, renal failure requiring dialysis, congestive heart failure (CHF), chronic obstructive pulmonary disease, diabetes, aspirin use, beta-blocker use, statin use, functional status, ASA classification, wound classification,
preoperative ankle-brachial index, prior ipsilateral open or endovascular surgery, work relative value units, and length of stay. Functional status was assessed by the NSQIP to determine a patient’s ability to perform activities of daily living 30 days before surgery and was categorized as independent, partially dependent, and totally dependent, depending on the amount of assistance required. We did not include the NSQIP variable on smoking status in our logistic regression models because it has been unreliable in our experience. All models output was examined for multicollinearity. A P value of # .05 was considered significant. RESULTS Patient and procedural characteristics. The 20122014 ACS NSQIP query identified 6112 patients who underwent open lower extremity revascularization. The baseline patient characteristics are shown in Table I. The most common indication was CLI TL (35.7%), followed by claudication (31.2%), CLI RP (29.1%), and other (4.0%). Most patients were male (65.8%) and white (72.5%). Hypertension, renal failure requiring dialysis, CHF, and diabetes were more common in patients with CLI TL (P < .001). Patients with CLI TL were also more likely to be functionally dependent, to have ASA class of 4 representative of a life-threatening condition, and to have longer lengths of stay (P < .001). Postoperative complications. Postoperative complications stratified by surgical indication are illustrated in Table II. In unadjusted analyses, patients with CLI TL were more likely to be readmitted at 14- and 30-day intervals (P < .001) and to experience cardiac complications (P < .001), respiratory complications (P < .001), and sepsis (P ¼ .012). Patients with CLI RP were more likely to experience wound complication (P ¼ .002) and major bypass reintervention (P ¼ .002). Reasons for readmission. The majority of reasons for 30-day readmissions were related to the patient’s index operation, with a minority related to other health conditions. On review, >50% of readmissions were attributed to wound complication, either wound infection or wound breakdown/nonhealing, followed by graft complication (8%), atherosclerosis (8%), gangrene (3%), sepsis (2%), bleeding (2%), and CHF (2%). Myocardial infarction, pulmonary embolus, pneumonia, urinary tract infection, renal failure, and stroke individually accounted for <1% of readmissions. Overall predictors of readmission. Unadjusted and adjusted odds of readmission were estimated by multivariate analysis for all patients undergoing lower extremity bypass and are summarized in Table III. After adjustment for preoperative factors, patients are more likely to be readmitted if they underwent lower extremity bypass for CLI TL (odds ratio [OR], 1.67; 95% confidence interval [CI], 1.35-2.06) or CLI RP (OR, 1.70; 95% CI, 1.382.09) compared with patients with claudication or other indications. Patient characteristics associated with increased 30-day readmission are obesity (OR, 1.41; 95% CI,
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Jones et al 3
Table I. Preoperative factors by indication Parameter Age, years Sex Male Female Race White Black Other BMI Hypertension Current dialysis CHF COPD Diabetes Insulin-dependent Noninsulin-dependent Aspirin use Beta-blocker use Statin use Functional status Independent Partially dependent Totally dependent ASA classification 1: No disturbance 2: Mild disturbance 3: Severe disturbance 4: Life-threatening Wound classification 1: Clean 2: Clean/contaminated 3: Contaminated 4: Dirty/infected ABI #0.39 0.40-0.89 0.90-1.29 $1.30 None, “not palpable” None, DP “palpable” None, unknown Anatomic factors None Prior ipsilateral bypass, same segment Prior ipsilateral PCI, same segment Work RVUs Length of stay, days Days to unplanned readmission
All (N ¼ 6112; Claudication CLI RP CLI TL Other 100%) (n ¼ 1907; 31.20%) (n ¼ 1778; 29.09%) (n ¼ 2184; 35.73%) (n ¼ 243; 3.98%) P value 67.32 (11.07)
65.90 (9.99)
66.76 (11.35)
68.94 (11.44)
68.10 (11.74)
<.001
4021 (65.8) 2091 (34.2)
1335 (70.01) 572 (29.99)
1134 (63.78) 644 (36.22)
1364 (62.45) 820 (37.55)
188 (77.37) 55 (22.63)
<.001
4434 1019 98 27.62 5005 297 147 797
(72.5) (16.7) (1.6) (6.05) (81.9) (4.9) (2.4) (13.0)
1506 208 29 28.28 1507 24 18 265
(78.97) (10.91) (1.52) (5.73) (79.02) (1.26) (0.94) (13.90)
1234 341 17 27.14 1451 59 36 229
(69.40) (19.18) (0.96) (6.15) (81.61) (3.32) (2.02) (12.88)
1498 451 47 27.30 1860 202 87 276
(68.59) (20.65) (2.15) (6.14) (85.16) (9.25) (3.98) (12.64)
196 19 5 28.66 187 12 6 27
(80.66) (7.82) (2.06) (6.45) (76.95) (4.94) (2.47) (11.11)
<.001
1543 1111 4963 3586 4258
(25.2) (18.2) (81.2) (58.7) (69.7)
310 321 1580 997 1349
(16.26) (16.83) (82.85) (52.28) (70.74)
351 305 1431 1024 1214
(19.74) (17.15) (80.48) (57.59) (68.28)
835 446 1763 1433 1539
(38.23) (20.42) (80.72) (65.61) (70.47)
47 39 189 132 156
(19.34) (16.05) (77.78) (54.32) (64.20)
<.001
<.001 <.001 <.001 <.001 .497
<.001 <.001 <.001
5728 (93.7) 338 (5.5) 21 (0.3)
1877 (98.43) 28 (1.47) 1 (0.05)
1695 (95.33) 64 (3.60) 7 (0.39)
1923 (88.05) 238 (10.90) 13 (0.60)
233 (95.88) 8 (3.29) 0 (0.00)
<.001
18 320 4567 1207
(0.3) (5.2) (74.7) (19.7)
10 144 1537 216
(0.52) (7.55) (80.60) (11.33)
4 103 1340 331
(0.22) (5.79) (75.37) (18.62)
3 51 1516 614
(0.14) (2.34) (69.41) (28.11)
1 22 174 46
(0.41) (9.05) (71.60) (18.93)
<.001
5829 133 76 74
(95.4) (2.2) (1.2) (1.2)
1880 16 6 5
(98.58) (0.84) (0.31) (0.26)
1732 25 10 11
(97.41) (1.41) (0.56) (0.62)
1991 85 58 50
(91.16) (3.89) (2.66) (2.29)
226 7 2 8
(93.00) (2.88) (0.82) (3.29)
<.001
1033 2064 183 162 927 265 1478
(16.9) (33.8) (3.0) (2.7) (15.2) (4.3) (24.2)
161 884 50 32 214 87 479
(8.44) (46.36) (2.62) (1.68) (11.22) (4.56) (25.12)
437 497 44 34 312 63 391
(24.58) (27.95) (2.47) (1.91) (17.55) (3.54) (21.99)
423 636 58 90 378 85 514
(19.37) (29.12) (2.66) (4.12) (17.31) (3.89) (23.53)
12 47 31 6 23 30 94
(4.94) (19.34) (12.76) (2.47) (9.47) (12.35) (38.68)
<.001
<.001
3826 (62.6) 1317 (21.5)
1268 (66.49) 344 (18.04)
959 (53.94) 494 (27.78)
1429 (65.43) 426 (19.51)
170 (69.96) 53 (21.81)
969 (15.9)
295 (15.47)
325 (18.28)
329 (15.06)
20 (8.23)
26.27 (4.44) 4.46 (2.17) 11.58 (6.94)
25.24 (4.19) 3.61 (1.89) 11.31 (7.21)
26.59 (4.56) 4.50 (2.13) 11.59 (6.98)
26.92 (4.43) 5.23 (2.16) 11.63 (6.81)
26.24 (4.11) 3.93 (2.10) 12.90 (6.69)
<.001 <.001 .797
ABI, Ankle-brachial index; ASA, American Society of Anesthesiologists; BMI, body mass index; CHF, congestive heart failure; CLI RP, critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss; COPD, chronic obstructive pulmonary disease; DP, dorsalis pedis; PCI, percutaneous coronary intervention; RVUs, relative value units. Categorical variables are presented as number (%). Continuous variables are presented as mean (standard deviation).
1.16-1.71) and extreme obesity (OR, 2.10; 95% CI, 1.463.01). Insulin-dependent diabetes was also associated with readmission (OR, 1.27; 95% CI, 1.07-1.52). Finally, patients with more complicated procedures and postoperative courses as represented by work relative value units (OR, 1.02; 95% CI, 1.00-1.04) and length of stay
(OR, 1.04; 95% CI, 1.01-1.08), respectively, were more likely to be readmitted at 30 days. We looked at whether having multiple procedures by different specialties during the same index hospitalization decreased readmission, but results were not significant. Significant predictors of 14-day readmission were similar, and odds of 14-day readmission
JOURNAL OF VASCULAR SURGERY --- 2016
4 Jones et al
Table II. Postoperative complications by surgical indication Complication
All, No. (%)
Thirty-day readmission Fourteen-day readmission Cardiac complication Respiratory complication Renal complication Wound complication Organ/space infection Sepsis Major bypass reintervention
904 588 95 64 26 520 16 86 208
(14.79) (9.62) (1.55) (1.05) (0.43) (8.51) (0.26) (1.41) (3.40)
Claudication, No. (%) 180 115 14 7 6 135 4 15 40
CLI RP, No. (%)
(9.44) (6.03) (0.73) (0.37) (0.31) (7.08) (0.21) (0.79) (2.10)
294 187 26 16 11 187 3 26 82
(16.54) (10.52) (1.46) (0.90) (0.62) (10.52) (0.17) (1.46) (4.61)
CLI TL, No. (%) 410 274 52 41 9 182 8 43 81
(18.77) (12.55) (2.38) (1.88) (0.41) (8.33) (0.37) (1.97) (3.71)
Other, No. (%) 20 12 3 0 0 16 1 2 5
(8.23) (4.94) (1.23) (0.00) (0.00) (6.58) (0.41) (0.82) (2.06)
P value <.001 <.001 <.001 <.001 .367 .002 .591 .012 .002
CLI RP, Critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss.
were higher in patients who underwent lower extremity bypass for CLI. Predictors of readmission by indication. Predictors of readmission varied by indication with some overlap. For all patients, insulin-dependent diabetes was a significant predictor of readmission (CLI TL: OR, 1.269 [95% CI, 1.00-1.61]; CLI RP: OR, 1.43 [95% CI, 1.05-1.97]; claudication: OR, 1.67 [95% CI, 1.13-2.46]). Patients with CLI RP were more likely to be readmitted if they were extremely obese (OR, 3.02; 95% CI, 1.62-5.6), used aspirin (OR, 1.51; 95% CI, 1.06-2.15), or had a longer postoperative length of stay (OR, 1.09; 95% CI, 1.031.16). Among patients with claudication, redo bypasses (OR, 1.96; 95% CI, 1.14- 3.4) and postoperative length of stay (OR, 1.13; 95% CI, 1.05-1.22) were predictive (Table IV). DISCUSSION In this study, 30-day readmission rates after lower extremity bypass varied significantly by surgical indication in both unadjusted and adjusted analyses. In addition, all study complications occurred more frequently among patients with CLI vs those with claudication or other indications. It is not surprising that patients with CLI incurred more complications because of their increased burden of both local and systemic atherosclerotic disease. However, we stress that because 30-day readmission and 30-day complication rates varied by surgical indication, they should be addressed by the Medicare Hospital Readmissions Reduction Program so that hospitals are not disproportionally penalized for case mix.11 Finally, significant predictors of readmission varied by indication, further suggesting that lower extremity bypasses performed for different indications should be treated as separate and distinct entities. The evidence presented in this study supports our hypothesis that 30-day readmission rates after lower extremity bypass differ on the basis of the indication for surgery. Previous reports associated readmissions with vascular surgical operations.5 We took this concept a step further and proposed that 30-day readmissions were also associated with the indication for lower extremity bypass. In 2013, McPhee et al demonstrated an increased risk of readmission
with patients who underwent surgery for tissue loss (OR, 1.62; P ¼ .0004).12 Similarly, the 2004 study of Goshima et al reported that perioperative complications, reoperations within 3 months, and readmissions within 6 months were highest for patients with tissue loss, followed by rest pain, then claudication.6 Given the authors’ focus on the intense effort involved in limb salvage with lower extremity bypass, they chose to limit their statistical analysis to only patients with tissue loss. In addition, their data were limited by a 6-month readmission time, in which they noted one-third of readmissions resulted from complications associated with the patients’ medical comorbidities. Whereas our findings agreed with these two studies, we studied 30-day and 14-day readmissions, a period when readmissions are more likely to be related to the index operation. Multivariate analysis showed that insulin-dependent diabetes was associated with 30-day readmission for all surgical indications. This was likely related to the increased risk of wound infection for diabetic patients and that most unplanned readmissions were related to a patient’s wound (62.9%).12,13 Length of stay was an independent predictor of readmission for patients undergoing lower extremity bypass for CLI RP and claudication. This was likely due to the relationship between length of stay and postoperative complications and between postoperative complications and readmissions, which agreed with current literature.14 Obesity was a significant predictor of readmission for patients with CLI RP, thought to be largely due to its association with poor wound healing and the increased likelihood of comorbidities, both of which were common reasons for readmission.15,16 Because CLI TL was an independent predictor of readmission after adjustment for other preoperative factors, it is important that expected readmission rates be adjusted for surgical indication. We believe that the evidence presented here highlights the potential value for including the Society for Vascular Surgery Lower Extremity Threatened Limb Classification System in both NSQIP data collection and entry into the ACS NSQIP risk calculator. The Society for Vascular Surgery Threatened Limb Classification is a risk assessment developed to stratify patients on the three major factors that affect clinical management: wound, ischemia, and foot infection.17,18 Collecting information
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Jones et al 5
Table III. Predictors of readmission Unadjusted Parameter Age, years Sex Male Female Race White Black Other BMI Underweight Normal weight Overweight Obesity Extreme obesity Hypertension Current dialysis CHF COPD Diabetes None Insulin-dependent Noninsulin-dependent Aspirin use Beta-blocker use Statin use Functional status Independent Partially dependent Totally dependent ASA classification 1: No disturbance 2: Mild disturbance 3: Severe disturbance 4: Life-threatening Wound classification 1: Clean 2: Clean/contaminated 3: Contaminated 4: Dirty/infected ABI #0.39 0.40-0.89 0.90-1.29 $1.30 None, “not palpable” None, DP “palpable” None, unknown Anatomic factors None Prior ipsilateral bypass, same segment Prior ipsilateral PCI, same segment Work RVUs Length of stay, days Days to unplanned readmission Symptoms Claudication CLI RP CLI TL Other
Adjusted
Readmitted
Not readmitted
OR (95% CI)
P value
OR (95% CI)
67.80 (11.74)
67.24 (10.95)
1.005 (0.998-1.01)
.1618 1.004 (1-1.01)
.2379
565 (62.5) 339 (37.5)
3456 (66.4) 1752 (33.6)
Ref. 1.184 (1.02-1.37)
Ref. .024
Ref. 1.105 (0.95-1.29)
Ref. .204
Ref. .011 .414 <.001 .371 Ref. .959 <.001 <.001 <.001 <.001 .004 .385
Ref. 1.101 (0.91-1.34) 0.745 (0.39-1.42)
Ref. .330 .372
0.745 (0.48-1.16) Ref. 1.049 (0.87-1.27) 1.407 (1.16-1.71) 2.097 (1.46-3.01) 1.113 (0.9-1.39) 1.286 (0.95-1.74) 1.319 (0.88-1.99) 1.118 (0.9-1.39)
.188 Ref. .623 <.001 <.001 .336 .102 .184 .307
627 176 11 28.51 26 246 285 290 51 776 69 34 126
(69.4) (19.5) (1.2) (6.66) (2.9) (27.2) (31.5) (32.1) (5.6) (85.8) (7.6) (3.8) (13.9)
3807 843 87 27.46 206 1602 1847 1370 150 4429 228 113 671
(73.1) (16.2) (9.6) (5.93) (4.0) (30.8) (35.5) (26.3) (2.9) (85.0) (4.4) (2.2) (12.9)
1.005 1.378 2.214 1.403 1.805 1.762 1.095
Ref. (1.06-1.52) (0.41-1.45) (1.02-1.04) (0.54-1.26) Ref. (0.84-1.21) (1.15-1.66) (1.57-3.13) (1.15-1.71) (1.37-2.39) (1.19-2.6) (0.892-1.344)
438 308 158 744 590 648
(48.5) (34.1) (17.5) (82.3) (65.3) (71.7)
3020 1235 953 4219 2996 3610
(58.0) (23.7) (18.3) (81.0) (57.5) (69.3)
1.72 1.143 1.103 1.382 1.117
Ref. (1.47-2.02) (0.94-1.39) (0.915-1.33) (1.19-1.602) (0.95-1.31)
Ref. <.001 .181 .306 <.001 .168
1.269 0.954 1.058 1.136 0.993
1.268 0.768 1.028 0.822
Ref. (1.06-1.52) (0.78-1.17) (0.87-1.29) (0.96-1.34) (0.21-13.46)
P value
Ref. .010 .654 .580 .132 .618
828 (91.6) 64 (7.1) 9 (1.0)
4900 (94.1) 274 (5.3) 12 (0.2)
Ref. 1.382 (1.04-1.83) 4.439 (1.86-10.57)
Ref. .024 <.001
Ref. 0.95 (0.71-1.28) 3.502 (1.44-8.53)
Ref. .735 .006
1 29 650 224
(0.1) (3.2) (71.9) (24.8)
17 291 3917 983
(0.3) (5.6) (75.2) (18.9)
Ref. 1.692 (0.22-13.17) 2.818 (0.38-21.19) 3.87 (0.51-29.21)
Ref. 0.615 .314 .189
Ref. 1.695 (0.21-13.46) 2.161 (0.28-16.58) 2.33 (0.3-17.97)
Ref. .618 .459 .417
841 25 21 17
(93.0) (2.8) (2.3) (1.9)
4988 108 55 57
(95.8) (2.1) (1.1) (1.1)
Ref. 1.373 (0.88-2.13) 2.265 (1.36-3.76) 1.769 (1.02-3.06)
Ref. .159 .002 .041
Ref. 1.128 (0.72-1.77) 1.874 (1.11-3.17) 1.377 (0.79-2.41)
Ref. .601 .019 .264
164 263 25 31 160 36 225
(18.1) (29.1) (2.8) (3.4) (17.7) (4.0) (24.9)
869 1801 158 131 767 229 1253
(16.7) (34.6) (3.0) (2.5) (14.7) (4.4) (24.1)
0.774 0.838 1.254 1.105 0.833 0.951
Ref. (0.63-0.96) (0.53-1.32) (0.82-1.92) (0.87-1.4) (0.57-1.23) (0.76-1.19)
Ref. .017 .446 .296 .410 .357 .656
0.933 1.045 1.179 1.13 0.927 1.08
Ref. (0.75-1.16) (0.66-1.67) (0.76-1.83) (0.89-1.44) (0.62-1.38) (0.86-1.36)
Ref. .539 .854 .461 .327 .709 .506
551 210 143 26.77 4.89 11.58
(61.0) (23.2) (15.8) (4.56) (2.14) (6.9)
3275 1107 826 26.19 4.38
(62.9) (21.3) (15.9) (4.41) (2.17) d
0.972 (0.8-1.19) 1.096 (0.87-1.38) Ref. 1.03 (1.01-1.05) 1.108 (1.07-1.14) d
.778 .437 Ref. <.001 <.001 d
1.049 (0.85-1.29) 1.115 (0.88-1.41) Ref. 1.02 (1-1.04) 1.042 (1.01-1.08) d
.650 .365 Ref. .019 .023 d
180 294 410 20
(19.9) (32.5) (45.4) (2.2)
1727 1484 1774 223
(33.2) (28.5) (34.1) (4.3)
Ref. 1.901 (1.56-2.32) 2.217 (1.84-2.67) 0.861 (0.53-1.39)
Ref. <.001 <.001 .542
Ref. 1.698 (1.38-2.09) 1.668 (1.35-2.06) 0.779 (0.48-1.27)
Ref. <.001 <.001 .319
ABI, Ankle-brachial index; ASA, American Society of Anesthesiologists; BMI, body mass index; CHF, congestive heart failure; CI, confidence interval; CLI RP, critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss; COPD, chronic obstructive pulmonary disease; DP, dorsalis pedis; OR, odds ratio; PCI, percutaneous coronary intervention; RVUs, relative value units. Categorical variables are presented as number (%). Continuous variables are presented as mean (standard deviation).
JOURNAL OF VASCULAR SURGERY --- 2016
6 Jones et al
Table IV. Significant adjusted predictors of readmission by indication Parameter Claudication Diabetes Insulin-dependent Noninsulin-dependent Postoperative length of stay Anatomic factors Prior ipsilateral percutaneous intervention of currently treated segment Prior ipsilateral bypass of currently treated segment None/not documented CLI RP Diabetes Insulin-dependent Noninsulin-dependent Postoperative length of stay BMI classification Underweight Normal Overweight Obese Extreme obesity Aspirin use CLI TL Diabetes Insulin-dependent Noninsulin-dependent
OR (95% CI)
c2
1.668 (1.13-2.46) 1.449 (0.97-2.16) 1.130 (1.05-1.22)
6.608 3.293 10.590
.010 .070 .001
Ref. 1.964 (1.14-3.4) 1.375 (0.85-2.24) c ¼ 0.613
Ref. 5.813 1.642 r2 ¼ 0.0292
Ref. .016 .200
1.433 (1.05-1.97) 1.068 (0.75-1.52) 1.091 (1.03-1.16)
5.011 0.137 8.502
.025 .712 <.001
0.611 (0.28-1.32) Ref. 1.084 (0.79-1.5) 1.322 (0.94-1.86) 3.017 (1.62-5.6) 1.508 (1.06-2.15) c ¼ 0.616
1.574 Ref 0.238 2.558 12.217 5.144 r2 ¼ 0.0391
.210 Ref .626 .110 .001 .023
1.269 (1-1.61) 0.771 (0.56-1.06) c ¼ 0.554
3.883 2.615 r2 ¼ 0.011
.049 .106
P value
BMI, Body mass index; CI, confidence interval; CLI RP, critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss; OR, odds ratio.
on these parameters would likely differentiate patients’ risk for further complications and readmissions. Our findings must be interpreted in the context of several limitations. Although the ACS NSQIP is a powerful database, it does not assess some characteristics that we think are likely to play a role in 30-day readmissions, including socioeconomic status and insurance. Second, the status of planned vs unplanned readmissions is determined by NSQIP-trained nurses, and its validity has been questioned. Per validation studies, the NSQIP is accurate in regard to identifying unplanned readmissions with k of 0.67.19 To increase our specificity for unplanned readmissions, we excluded all patients who underwent lower extremity amputation during their readmission. However, we were unable to evaluate ipsilateral vs contralateral events as the reason for procedure to exclude staged bilateral procedures. The ACS NSQIP collects readmissions within 30 days from the index operation rather than at 30 days from discharge. Therefore, patients with long hospital stays have a shorter time span to be readmitted within the 30-day window, leading to a spurious protective association of length of stay with readmission. To minimize this bias, we excluded all patients with length of stay >10 days. Although we addressed many potential confounders, as with other observational studies, there remains the possibility of unmeasured confounding. Finally, because this is an observational study, no causal connections can be inferred.
The 30-day readmission rates after lower extremity bypass vary significantly on the basis of the indication for surgery and remain significant even after adjustment for other preoperative factors. Readmissions after vascular surgery and in particular lower extremity bypass are frequent, costly, and associated with worse outcomes. If the Affordable Care Act’s Readmissions Reduction Program continues to penalize hospitals with higher than expected readmission rates, it should adjust for surgical indication of lower extremity bypass. CONCLUSIONS Lower extremity bypasses are performed for multiple indications, and the risk for readmission varies by the indication for surgery. Operations performed for CLI are associated with the highest rates of readmission. If hospitals are to be penalized for readmission rates that exceed the predictions, the predicted readmission rates should account for the surgical indication. AUTHOR CONTRIBUTIONS Conception and design: CJ, AG, MM Analysis and interpretation: CJ, JR, DC, AG, BP, MM Data collection: CJ, AG Writing the article: CJ Critical revision of the article: CJ, JR, DC, BP, MM Final approval of the article: CJ, JR, DC, AG, BP, MM
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Statistical analysis: CJ, JR, AG Obtained funding: MM Overall responsibility: CJ
REFERENCES 1. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418-28. 2. McPhee JT, Nguyen LL, Ho KJ, Ozaki CK, Conte MS, Belkin M. Risk prediction of 30-day readmission after infrainguinal bypass for critical limb ischemia. J Vasc Surg 2013;57:1481-8. 3. Orr NT, El-Maraghi S, Korosec RL, Davenport DL, Xenos ES. Cost analysis of vascular readmissions after common vascular procedures. J Vasc Surg 2015;62:1281-7.e1. 4. Davenport DL, Zwischenberger BA, Xenos ES. Analysis of 30-day readmission after aortoiliac and infrainguinal revascularization using the American College of Surgeons National Surgical Quality Improvement Program data set. J Vasc Surg 2014;60:1266-74. 5. Gupta PK, Fernandes-Taylor S, Ramanan B, Engelbert TL, Kent KC. Unplanned readmissions after vascular surgery. J Vasc Surg 2014;59: 473-82. 6. Goshima KR, Mills JL Sr, Hughes JD. A new look at outcomes after infrainguinal bypass surgery: traditional reporting standards systematically underestimate the expenditure of effort required to attain limb salvage. J Vasc Surg 2004;39:330-5. 7. American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP). 2013. Available at: http://www.facs.org/ quality-programs/acs-nsqip. Accessed June 2015. 8. Sellers MM, Merkow RP, Halverson A, Hinami K, Kelz RR, Bentrem DJ, et al. Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2013;216:420-7. 9. Shiloach M, Frencher SK Jr, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, et al. Toward robust information: data quality and interrater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2010;210:6-16.
Jones et al 7
10. Jackson BM, Nathan DP, Doctor L, Wang GJ, Woo EY, Fairman RM. Low rehospitalization rate for vascular surgery patients. J Vasc Surg 2011;54:767-72. 11. Weber SM, Greenberg CC. Medicare Hospital Readmission Reduction Program: what is the effect on surgery? Surgery 2014;156:1066-8. 12. McPhee JT, Barshes NR, Ho KJ, Madenci A, Ozaki CK, Nguyen LL, et al. Predictive factors of 30-day unplanned readmission after lower extremity bypass. J Vasc Surg 2013;57:955-62. 13. Ata A, Lee J, Bestle SL, Desemone J, Stain SC. Postoperative hyperglycemia and surgical site infection in general surgery patients. Arch Surg 2010;145:858-64. 14. Damrauer SM, Gaffey AC, DeBord Smith A, Fairman RM, Nguyen LL. Comparison of risk factors for length of stay and readmission following lower extremity bypass surgery. J Vasc Surg 2015;62: 1192-200.e1. 15. Wiseman JT, Fernandes-Taylor S, Barnes ML, Saunders RS, Saha S, Havlena J, et al. Predictors of surgical site infection after hospital discharge in patients undergoing major vascular surgery. J Vasc Surg 2015;62:1023-31.e5. 16. Zhang JQ, Curran T, McCallum JC, Wang L, Wyers MC, Hamdan AD, et al. Risk factors for readmission after lower extremity bypass in the American College of Surgeons National Surgery Quality Improvement Program. J Vasc Surg 2014;59:1331-9. 17. Mills JL Sr, Conte MS, Armstrong DG, Pomposelli FB, Schanzer A, Sidawy AN, et al. The Society for Vascular Surgery Lower Extremity Threatened Limb Classification System: risk stratification based on wound, ischemia, and foot infection (WIfI). J Vasc Surg 2014;59: 220-34. e1-2. 18. Zhan LX, Branco BC, Armstrong DG, Mills JL Sr. The Society for Vascular Surgery lower extremity threatened limb classification system based on Wound, Ischemia, and foot Infection (WIfI) correlates with risk of major amputation and time to wound healing. J Vasc Surg 2015;61:939-44. 19. Sellers MM, Reinke CE, Kreider S, Meise C, Nelis K, Volpe A, et al. American College of Surgeons NSQIP: quality in-training initiative pilot study. J Am Coll Surg 2013;217:827-32. Submitted Dec 26, 2015; accepted Mar 9, 2016.
Unplanned readmission rates after vascular surgery are among the highest within surgical specialties. Jencks et al reported a 23.9% unplanned 30-day readmission rate after vascular surgery, substantially higher than the overall surgical readmission rate of 15.6%.1 Vascular surgery readmissions ranked third highest of any diagnosis-related group, behind only heart failure (26.9%) and psychoses (24.6%).1 Among vascular surgical procedures, lower extremity bypass is associated with the highest rate of unplanned readmissions.2 Postoperative readmissions are costly and associated with worse outcomes. The estimated cost of readmissions is $12 billion per year and accounts for approximately 25% of all Medicare expenditures for inpatient care. The average cost per readmission after lower extremity bypass is $10,000 and adds an additional 39% to the index admission cost.1,3 Patients readmitted after lower extremity From the Department of Surgery, University of Alabama-Birmingham. Author conflict of interest: none. This abstract was selected for an oral presentation at the Academic Surgical Congress, Jacksonville, Fla, February 2-4, 2016. Correspondence: Melanie S. Morris, MD, Department of Surgery, University of Alabama-Birmingham, 1922 7th Ave S, Kracke Bldg, Birmingham, AL 35294 (e-mail: [email protected]). 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.2016.03.422
bypass also experience worse outcomes with significantly lower rates of limb salvage at 1 year (78.3% vs 89.6%), higher rates of reoperation (47.4% vs 6.8%), and increased mortality (1.9% vs 0.3%).2 Current literature recognizes wound infection and graft complication as the main reasons for readmission after lower extremity bypass.4 Other recognized risk factors include female gender, smoking, obesity, high American Society of Anesthesiologists (ASA) class, history of cardiac events, chronic obstructive pulmonary disease, renal failure requiring dialysis, and preoperative functional status.2,5 However, the association between the indication for a lower extremity revascularization procedure and risk for readmission is unknown. This relationship is important as previous evidence suggests a potential correlation between preoperative symptoms and complications.6 We examined the relationship between readmission and the main indications for lower extremity bypass: claudication, critical limb ischemia rest pain (CLI RP), and critical limb ischemia tissue loss (CLI TL). We hypothesized that readmission rates after lower extremity bypass differ on the basis of the indication for surgery after adjustment for covariates. METHODS Data set. We queried the 2012-2014 American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) procedure-targeted vascular cohort. The ACS NSQIP reports perioperative variables 1
JOURNAL OF VASCULAR SURGERY --- 2016
2 Jones et al
and surgical outcomes for >400 participating academic and community hospitals within the United States. The ACS NSQIP is a well-validated prospective database that accurately captures readmissions with highly effective data collection and inter-rating reliability.7,8 Clinical nurse training and audit procedures ensure robust data that have improved in reliability every year.9 Patient informed consent was determined to be unnecessary, and Institutional Review Board approval was waived. Patient cohort. Within the ACS NSQIP proceduretargeted vascular cohort, we identified all patients who underwent a lower extremity bypass by the procedural terminology code for lower extremity bypassdopen. This procedural code encompasses all femoral-distal, femoralpopliteal, and popliteal-distal bypasses with vein or prosthetic conduit, femoral endarterectomies, and profundoplasties using Current Procedural Terminology codes 35556, 35566, 35571, 35583, 35585, 35587, 35656, 35666, and 35671. We did not exclude patients with concomitant inflow procedures. We excluded all emergent cases, patients with an ASA class of 5, patients experiencing postoperative length of stay >10 days, and patients who died within 30 days of the index operation. We excluded all planned readmissions and all patients who underwent amputation during their readmission, as this most likely represented a planned readmission awaiting tissue demarcation. Patients were stratified by the NSQIPreported symptoms claudication, CLI RP, CLI TL, and other, including diagnoses of aneurysm, embolism, and technical cause (ie, graft complication). Outcomes. Primary outcome was 30-day unplanned readmission. Secondary outcomes were 14-day readmission and 30-day postoperative complications. Unplanned readmission was defined by the ACS NSQIP as “unplanned readmission (to the same or another hospital) for a postoperative occurrence likely related to the principal surgical procedure within 30 days of the procedure.” The surgical clinical reviewer categorized readmissions as planned or unplanned on the basis of whether the readmission was known at time of discharge from index hospitalization. It is especially important to target unplanned readmissions in this cohort because approximately 25% of readmissions in vascular surgery patients are planned.10 Postoperative complications included those identified by the ACS NSQIP and limited to 30 days postoperatively. Statistical analysis. The results were examined with procedure as the unit of analysis. Analyses were performed using SAS software (version 9.4; SAS Institute, Cary, NC). The c2 and Wilcoxon rank sum tests were used to determine the differences among categorical and continuous variables, respectively. The association between readmission and surgical indication was further assessed by logistic regression adjusted for the following preoperative factors that were selected a priori: age, sex, race, body mass index, hypertension, renal failure requiring dialysis, congestive heart failure (CHF), chronic obstructive pulmonary disease, diabetes, aspirin use, beta-blocker use, statin use, functional status, ASA classification, wound classification,
preoperative ankle-brachial index, prior ipsilateral open or endovascular surgery, work relative value units, and length of stay. Functional status was assessed by the NSQIP to determine a patient’s ability to perform activities of daily living 30 days before surgery and was categorized as independent, partially dependent, and totally dependent, depending on the amount of assistance required. We did not include the NSQIP variable on smoking status in our logistic regression models because it has been unreliable in our experience. All models output was examined for multicollinearity. A P value of # .05 was considered significant. RESULTS Patient and procedural characteristics. The 20122014 ACS NSQIP query identified 6112 patients who underwent open lower extremity revascularization. The baseline patient characteristics are shown in Table I. The most common indication was CLI TL (35.7%), followed by claudication (31.2%), CLI RP (29.1%), and other (4.0%). Most patients were male (65.8%) and white (72.5%). Hypertension, renal failure requiring dialysis, CHF, and diabetes were more common in patients with CLI TL (P < .001). Patients with CLI TL were also more likely to be functionally dependent, to have ASA class of 4 representative of a life-threatening condition, and to have longer lengths of stay (P < .001). Postoperative complications. Postoperative complications stratified by surgical indication are illustrated in Table II. In unadjusted analyses, patients with CLI TL were more likely to be readmitted at 14- and 30-day intervals (P < .001) and to experience cardiac complications (P < .001), respiratory complications (P < .001), and sepsis (P ¼ .012). Patients with CLI RP were more likely to experience wound complication (P ¼ .002) and major bypass reintervention (P ¼ .002). Reasons for readmission. The majority of reasons for 30-day readmissions were related to the patient’s index operation, with a minority related to other health conditions. On review, >50% of readmissions were attributed to wound complication, either wound infection or wound breakdown/nonhealing, followed by graft complication (8%), atherosclerosis (8%), gangrene (3%), sepsis (2%), bleeding (2%), and CHF (2%). Myocardial infarction, pulmonary embolus, pneumonia, urinary tract infection, renal failure, and stroke individually accounted for <1% of readmissions. Overall predictors of readmission. Unadjusted and adjusted odds of readmission were estimated by multivariate analysis for all patients undergoing lower extremity bypass and are summarized in Table III. After adjustment for preoperative factors, patients are more likely to be readmitted if they underwent lower extremity bypass for CLI TL (odds ratio [OR], 1.67; 95% confidence interval [CI], 1.35-2.06) or CLI RP (OR, 1.70; 95% CI, 1.382.09) compared with patients with claudication or other indications. Patient characteristics associated with increased 30-day readmission are obesity (OR, 1.41; 95% CI,
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Jones et al 3
Table I. Preoperative factors by indication Parameter Age, years Sex Male Female Race White Black Other BMI Hypertension Current dialysis CHF COPD Diabetes Insulin-dependent Noninsulin-dependent Aspirin use Beta-blocker use Statin use Functional status Independent Partially dependent Totally dependent ASA classification 1: No disturbance 2: Mild disturbance 3: Severe disturbance 4: Life-threatening Wound classification 1: Clean 2: Clean/contaminated 3: Contaminated 4: Dirty/infected ABI #0.39 0.40-0.89 0.90-1.29 $1.30 None, “not palpable” None, DP “palpable” None, unknown Anatomic factors None Prior ipsilateral bypass, same segment Prior ipsilateral PCI, same segment Work RVUs Length of stay, days Days to unplanned readmission
All (N ¼ 6112; Claudication CLI RP CLI TL Other 100%) (n ¼ 1907; 31.20%) (n ¼ 1778; 29.09%) (n ¼ 2184; 35.73%) (n ¼ 243; 3.98%) P value 67.32 (11.07)
65.90 (9.99)
66.76 (11.35)
68.94 (11.44)
68.10 (11.74)
<.001
4021 (65.8) 2091 (34.2)
1335 (70.01) 572 (29.99)
1134 (63.78) 644 (36.22)
1364 (62.45) 820 (37.55)
188 (77.37) 55 (22.63)
<.001
4434 1019 98 27.62 5005 297 147 797
(72.5) (16.7) (1.6) (6.05) (81.9) (4.9) (2.4) (13.0)
1506 208 29 28.28 1507 24 18 265
(78.97) (10.91) (1.52) (5.73) (79.02) (1.26) (0.94) (13.90)
1234 341 17 27.14 1451 59 36 229
(69.40) (19.18) (0.96) (6.15) (81.61) (3.32) (2.02) (12.88)
1498 451 47 27.30 1860 202 87 276
(68.59) (20.65) (2.15) (6.14) (85.16) (9.25) (3.98) (12.64)
196 19 5 28.66 187 12 6 27
(80.66) (7.82) (2.06) (6.45) (76.95) (4.94) (2.47) (11.11)
<.001
1543 1111 4963 3586 4258
(25.2) (18.2) (81.2) (58.7) (69.7)
310 321 1580 997 1349
(16.26) (16.83) (82.85) (52.28) (70.74)
351 305 1431 1024 1214
(19.74) (17.15) (80.48) (57.59) (68.28)
835 446 1763 1433 1539
(38.23) (20.42) (80.72) (65.61) (70.47)
47 39 189 132 156
(19.34) (16.05) (77.78) (54.32) (64.20)
<.001
<.001 <.001 <.001 <.001 .497
<.001 <.001 <.001
5728 (93.7) 338 (5.5) 21 (0.3)
1877 (98.43) 28 (1.47) 1 (0.05)
1695 (95.33) 64 (3.60) 7 (0.39)
1923 (88.05) 238 (10.90) 13 (0.60)
233 (95.88) 8 (3.29) 0 (0.00)
<.001
18 320 4567 1207
(0.3) (5.2) (74.7) (19.7)
10 144 1537 216
(0.52) (7.55) (80.60) (11.33)
4 103 1340 331
(0.22) (5.79) (75.37) (18.62)
3 51 1516 614
(0.14) (2.34) (69.41) (28.11)
1 22 174 46
(0.41) (9.05) (71.60) (18.93)
<.001
5829 133 76 74
(95.4) (2.2) (1.2) (1.2)
1880 16 6 5
(98.58) (0.84) (0.31) (0.26)
1732 25 10 11
(97.41) (1.41) (0.56) (0.62)
1991 85 58 50
(91.16) (3.89) (2.66) (2.29)
226 7 2 8
(93.00) (2.88) (0.82) (3.29)
<.001
1033 2064 183 162 927 265 1478
(16.9) (33.8) (3.0) (2.7) (15.2) (4.3) (24.2)
161 884 50 32 214 87 479
(8.44) (46.36) (2.62) (1.68) (11.22) (4.56) (25.12)
437 497 44 34 312 63 391
(24.58) (27.95) (2.47) (1.91) (17.55) (3.54) (21.99)
423 636 58 90 378 85 514
(19.37) (29.12) (2.66) (4.12) (17.31) (3.89) (23.53)
12 47 31 6 23 30 94
(4.94) (19.34) (12.76) (2.47) (9.47) (12.35) (38.68)
<.001
<.001
3826 (62.6) 1317 (21.5)
1268 (66.49) 344 (18.04)
959 (53.94) 494 (27.78)
1429 (65.43) 426 (19.51)
170 (69.96) 53 (21.81)
969 (15.9)
295 (15.47)
325 (18.28)
329 (15.06)
20 (8.23)
26.27 (4.44) 4.46 (2.17) 11.58 (6.94)
25.24 (4.19) 3.61 (1.89) 11.31 (7.21)
26.59 (4.56) 4.50 (2.13) 11.59 (6.98)
26.92 (4.43) 5.23 (2.16) 11.63 (6.81)
26.24 (4.11) 3.93 (2.10) 12.90 (6.69)
<.001 <.001 .797
ABI, Ankle-brachial index; ASA, American Society of Anesthesiologists; BMI, body mass index; CHF, congestive heart failure; CLI RP, critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss; COPD, chronic obstructive pulmonary disease; DP, dorsalis pedis; PCI, percutaneous coronary intervention; RVUs, relative value units. Categorical variables are presented as number (%). Continuous variables are presented as mean (standard deviation).
1.16-1.71) and extreme obesity (OR, 2.10; 95% CI, 1.463.01). Insulin-dependent diabetes was also associated with readmission (OR, 1.27; 95% CI, 1.07-1.52). Finally, patients with more complicated procedures and postoperative courses as represented by work relative value units (OR, 1.02; 95% CI, 1.00-1.04) and length of stay
(OR, 1.04; 95% CI, 1.01-1.08), respectively, were more likely to be readmitted at 30 days. We looked at whether having multiple procedures by different specialties during the same index hospitalization decreased readmission, but results were not significant. Significant predictors of 14-day readmission were similar, and odds of 14-day readmission
JOURNAL OF VASCULAR SURGERY --- 2016
4 Jones et al
Table II. Postoperative complications by surgical indication Complication
All, No. (%)
Thirty-day readmission Fourteen-day readmission Cardiac complication Respiratory complication Renal complication Wound complication Organ/space infection Sepsis Major bypass reintervention
904 588 95 64 26 520 16 86 208
(14.79) (9.62) (1.55) (1.05) (0.43) (8.51) (0.26) (1.41) (3.40)
Claudication, No. (%) 180 115 14 7 6 135 4 15 40
CLI RP, No. (%)
(9.44) (6.03) (0.73) (0.37) (0.31) (7.08) (0.21) (0.79) (2.10)
294 187 26 16 11 187 3 26 82
(16.54) (10.52) (1.46) (0.90) (0.62) (10.52) (0.17) (1.46) (4.61)
CLI TL, No. (%) 410 274 52 41 9 182 8 43 81
(18.77) (12.55) (2.38) (1.88) (0.41) (8.33) (0.37) (1.97) (3.71)
Other, No. (%) 20 12 3 0 0 16 1 2 5
(8.23) (4.94) (1.23) (0.00) (0.00) (6.58) (0.41) (0.82) (2.06)
P value <.001 <.001 <.001 <.001 .367 .002 .591 .012 .002
CLI RP, Critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss.
were higher in patients who underwent lower extremity bypass for CLI. Predictors of readmission by indication. Predictors of readmission varied by indication with some overlap. For all patients, insulin-dependent diabetes was a significant predictor of readmission (CLI TL: OR, 1.269 [95% CI, 1.00-1.61]; CLI RP: OR, 1.43 [95% CI, 1.05-1.97]; claudication: OR, 1.67 [95% CI, 1.13-2.46]). Patients with CLI RP were more likely to be readmitted if they were extremely obese (OR, 3.02; 95% CI, 1.62-5.6), used aspirin (OR, 1.51; 95% CI, 1.06-2.15), or had a longer postoperative length of stay (OR, 1.09; 95% CI, 1.031.16). Among patients with claudication, redo bypasses (OR, 1.96; 95% CI, 1.14- 3.4) and postoperative length of stay (OR, 1.13; 95% CI, 1.05-1.22) were predictive (Table IV). DISCUSSION In this study, 30-day readmission rates after lower extremity bypass varied significantly by surgical indication in both unadjusted and adjusted analyses. In addition, all study complications occurred more frequently among patients with CLI vs those with claudication or other indications. It is not surprising that patients with CLI incurred more complications because of their increased burden of both local and systemic atherosclerotic disease. However, we stress that because 30-day readmission and 30-day complication rates varied by surgical indication, they should be addressed by the Medicare Hospital Readmissions Reduction Program so that hospitals are not disproportionally penalized for case mix.11 Finally, significant predictors of readmission varied by indication, further suggesting that lower extremity bypasses performed for different indications should be treated as separate and distinct entities. The evidence presented in this study supports our hypothesis that 30-day readmission rates after lower extremity bypass differ on the basis of the indication for surgery. Previous reports associated readmissions with vascular surgical operations.5 We took this concept a step further and proposed that 30-day readmissions were also associated with the indication for lower extremity bypass. In 2013, McPhee et al demonstrated an increased risk of readmission
with patients who underwent surgery for tissue loss (OR, 1.62; P ¼ .0004).12 Similarly, the 2004 study of Goshima et al reported that perioperative complications, reoperations within 3 months, and readmissions within 6 months were highest for patients with tissue loss, followed by rest pain, then claudication.6 Given the authors’ focus on the intense effort involved in limb salvage with lower extremity bypass, they chose to limit their statistical analysis to only patients with tissue loss. In addition, their data were limited by a 6-month readmission time, in which they noted one-third of readmissions resulted from complications associated with the patients’ medical comorbidities. Whereas our findings agreed with these two studies, we studied 30-day and 14-day readmissions, a period when readmissions are more likely to be related to the index operation. Multivariate analysis showed that insulin-dependent diabetes was associated with 30-day readmission for all surgical indications. This was likely related to the increased risk of wound infection for diabetic patients and that most unplanned readmissions were related to a patient’s wound (62.9%).12,13 Length of stay was an independent predictor of readmission for patients undergoing lower extremity bypass for CLI RP and claudication. This was likely due to the relationship between length of stay and postoperative complications and between postoperative complications and readmissions, which agreed with current literature.14 Obesity was a significant predictor of readmission for patients with CLI RP, thought to be largely due to its association with poor wound healing and the increased likelihood of comorbidities, both of which were common reasons for readmission.15,16 Because CLI TL was an independent predictor of readmission after adjustment for other preoperative factors, it is important that expected readmission rates be adjusted for surgical indication. We believe that the evidence presented here highlights the potential value for including the Society for Vascular Surgery Lower Extremity Threatened Limb Classification System in both NSQIP data collection and entry into the ACS NSQIP risk calculator. The Society for Vascular Surgery Threatened Limb Classification is a risk assessment developed to stratify patients on the three major factors that affect clinical management: wound, ischemia, and foot infection.17,18 Collecting information
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Jones et al 5
Table III. Predictors of readmission Unadjusted Parameter Age, years Sex Male Female Race White Black Other BMI Underweight Normal weight Overweight Obesity Extreme obesity Hypertension Current dialysis CHF COPD Diabetes None Insulin-dependent Noninsulin-dependent Aspirin use Beta-blocker use Statin use Functional status Independent Partially dependent Totally dependent ASA classification 1: No disturbance 2: Mild disturbance 3: Severe disturbance 4: Life-threatening Wound classification 1: Clean 2: Clean/contaminated 3: Contaminated 4: Dirty/infected ABI #0.39 0.40-0.89 0.90-1.29 $1.30 None, “not palpable” None, DP “palpable” None, unknown Anatomic factors None Prior ipsilateral bypass, same segment Prior ipsilateral PCI, same segment Work RVUs Length of stay, days Days to unplanned readmission Symptoms Claudication CLI RP CLI TL Other
Adjusted
Readmitted
Not readmitted
OR (95% CI)
P value
OR (95% CI)
67.80 (11.74)
67.24 (10.95)
1.005 (0.998-1.01)
.1618 1.004 (1-1.01)
.2379
565 (62.5) 339 (37.5)
3456 (66.4) 1752 (33.6)
Ref. 1.184 (1.02-1.37)
Ref. .024
Ref. 1.105 (0.95-1.29)
Ref. .204
Ref. .011 .414 <.001 .371 Ref. .959 <.001 <.001 <.001 <.001 .004 .385
Ref. 1.101 (0.91-1.34) 0.745 (0.39-1.42)
Ref. .330 .372
0.745 (0.48-1.16) Ref. 1.049 (0.87-1.27) 1.407 (1.16-1.71) 2.097 (1.46-3.01) 1.113 (0.9-1.39) 1.286 (0.95-1.74) 1.319 (0.88-1.99) 1.118 (0.9-1.39)
.188 Ref. .623 <.001 <.001 .336 .102 .184 .307
627 176 11 28.51 26 246 285 290 51 776 69 34 126
(69.4) (19.5) (1.2) (6.66) (2.9) (27.2) (31.5) (32.1) (5.6) (85.8) (7.6) (3.8) (13.9)
3807 843 87 27.46 206 1602 1847 1370 150 4429 228 113 671
(73.1) (16.2) (9.6) (5.93) (4.0) (30.8) (35.5) (26.3) (2.9) (85.0) (4.4) (2.2) (12.9)
1.005 1.378 2.214 1.403 1.805 1.762 1.095
Ref. (1.06-1.52) (0.41-1.45) (1.02-1.04) (0.54-1.26) Ref. (0.84-1.21) (1.15-1.66) (1.57-3.13) (1.15-1.71) (1.37-2.39) (1.19-2.6) (0.892-1.344)
438 308 158 744 590 648
(48.5) (34.1) (17.5) (82.3) (65.3) (71.7)
3020 1235 953 4219 2996 3610
(58.0) (23.7) (18.3) (81.0) (57.5) (69.3)
1.72 1.143 1.103 1.382 1.117
Ref. (1.47-2.02) (0.94-1.39) (0.915-1.33) (1.19-1.602) (0.95-1.31)
Ref. <.001 .181 .306 <.001 .168
1.269 0.954 1.058 1.136 0.993
1.268 0.768 1.028 0.822
Ref. (1.06-1.52) (0.78-1.17) (0.87-1.29) (0.96-1.34) (0.21-13.46)
P value
Ref. .010 .654 .580 .132 .618
828 (91.6) 64 (7.1) 9 (1.0)
4900 (94.1) 274 (5.3) 12 (0.2)
Ref. 1.382 (1.04-1.83) 4.439 (1.86-10.57)
Ref. .024 <.001
Ref. 0.95 (0.71-1.28) 3.502 (1.44-8.53)
Ref. .735 .006
1 29 650 224
(0.1) (3.2) (71.9) (24.8)
17 291 3917 983
(0.3) (5.6) (75.2) (18.9)
Ref. 1.692 (0.22-13.17) 2.818 (0.38-21.19) 3.87 (0.51-29.21)
Ref. 0.615 .314 .189
Ref. 1.695 (0.21-13.46) 2.161 (0.28-16.58) 2.33 (0.3-17.97)
Ref. .618 .459 .417
841 25 21 17
(93.0) (2.8) (2.3) (1.9)
4988 108 55 57
(95.8) (2.1) (1.1) (1.1)
Ref. 1.373 (0.88-2.13) 2.265 (1.36-3.76) 1.769 (1.02-3.06)
Ref. .159 .002 .041
Ref. 1.128 (0.72-1.77) 1.874 (1.11-3.17) 1.377 (0.79-2.41)
Ref. .601 .019 .264
164 263 25 31 160 36 225
(18.1) (29.1) (2.8) (3.4) (17.7) (4.0) (24.9)
869 1801 158 131 767 229 1253
(16.7) (34.6) (3.0) (2.5) (14.7) (4.4) (24.1)
0.774 0.838 1.254 1.105 0.833 0.951
Ref. (0.63-0.96) (0.53-1.32) (0.82-1.92) (0.87-1.4) (0.57-1.23) (0.76-1.19)
Ref. .017 .446 .296 .410 .357 .656
0.933 1.045 1.179 1.13 0.927 1.08
Ref. (0.75-1.16) (0.66-1.67) (0.76-1.83) (0.89-1.44) (0.62-1.38) (0.86-1.36)
Ref. .539 .854 .461 .327 .709 .506
551 210 143 26.77 4.89 11.58
(61.0) (23.2) (15.8) (4.56) (2.14) (6.9)
3275 1107 826 26.19 4.38
(62.9) (21.3) (15.9) (4.41) (2.17) d
0.972 (0.8-1.19) 1.096 (0.87-1.38) Ref. 1.03 (1.01-1.05) 1.108 (1.07-1.14) d
.778 .437 Ref. <.001 <.001 d
1.049 (0.85-1.29) 1.115 (0.88-1.41) Ref. 1.02 (1-1.04) 1.042 (1.01-1.08) d
.650 .365 Ref. .019 .023 d
180 294 410 20
(19.9) (32.5) (45.4) (2.2)
1727 1484 1774 223
(33.2) (28.5) (34.1) (4.3)
Ref. 1.901 (1.56-2.32) 2.217 (1.84-2.67) 0.861 (0.53-1.39)
Ref. <.001 <.001 .542
Ref. 1.698 (1.38-2.09) 1.668 (1.35-2.06) 0.779 (0.48-1.27)
Ref. <.001 <.001 .319
ABI, Ankle-brachial index; ASA, American Society of Anesthesiologists; BMI, body mass index; CHF, congestive heart failure; CI, confidence interval; CLI RP, critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss; COPD, chronic obstructive pulmonary disease; DP, dorsalis pedis; OR, odds ratio; PCI, percutaneous coronary intervention; RVUs, relative value units. Categorical variables are presented as number (%). Continuous variables are presented as mean (standard deviation).
JOURNAL OF VASCULAR SURGERY --- 2016
6 Jones et al
Table IV. Significant adjusted predictors of readmission by indication Parameter Claudication Diabetes Insulin-dependent Noninsulin-dependent Postoperative length of stay Anatomic factors Prior ipsilateral percutaneous intervention of currently treated segment Prior ipsilateral bypass of currently treated segment None/not documented CLI RP Diabetes Insulin-dependent Noninsulin-dependent Postoperative length of stay BMI classification Underweight Normal Overweight Obese Extreme obesity Aspirin use CLI TL Diabetes Insulin-dependent Noninsulin-dependent
OR (95% CI)
c2
1.668 (1.13-2.46) 1.449 (0.97-2.16) 1.130 (1.05-1.22)
6.608 3.293 10.590
.010 .070 .001
Ref. 1.964 (1.14-3.4) 1.375 (0.85-2.24) c ¼ 0.613
Ref. 5.813 1.642 r2 ¼ 0.0292
Ref. .016 .200
1.433 (1.05-1.97) 1.068 (0.75-1.52) 1.091 (1.03-1.16)
5.011 0.137 8.502
.025 .712 <.001
0.611 (0.28-1.32) Ref. 1.084 (0.79-1.5) 1.322 (0.94-1.86) 3.017 (1.62-5.6) 1.508 (1.06-2.15) c ¼ 0.616
1.574 Ref 0.238 2.558 12.217 5.144 r2 ¼ 0.0391
.210 Ref .626 .110 .001 .023
1.269 (1-1.61) 0.771 (0.56-1.06) c ¼ 0.554
3.883 2.615 r2 ¼ 0.011
.049 .106
P value
BMI, Body mass index; CI, confidence interval; CLI RP, critical limb ischemia rest pain; CLI TL, critical limb ischemia tissue loss; OR, odds ratio.
on these parameters would likely differentiate patients’ risk for further complications and readmissions. Our findings must be interpreted in the context of several limitations. Although the ACS NSQIP is a powerful database, it does not assess some characteristics that we think are likely to play a role in 30-day readmissions, including socioeconomic status and insurance. Second, the status of planned vs unplanned readmissions is determined by NSQIP-trained nurses, and its validity has been questioned. Per validation studies, the NSQIP is accurate in regard to identifying unplanned readmissions with k of 0.67.19 To increase our specificity for unplanned readmissions, we excluded all patients who underwent lower extremity amputation during their readmission. However, we were unable to evaluate ipsilateral vs contralateral events as the reason for procedure to exclude staged bilateral procedures. The ACS NSQIP collects readmissions within 30 days from the index operation rather than at 30 days from discharge. Therefore, patients with long hospital stays have a shorter time span to be readmitted within the 30-day window, leading to a spurious protective association of length of stay with readmission. To minimize this bias, we excluded all patients with length of stay >10 days. Although we addressed many potential confounders, as with other observational studies, there remains the possibility of unmeasured confounding. Finally, because this is an observational study, no causal connections can be inferred.
The 30-day readmission rates after lower extremity bypass vary significantly on the basis of the indication for surgery and remain significant even after adjustment for other preoperative factors. Readmissions after vascular surgery and in particular lower extremity bypass are frequent, costly, and associated with worse outcomes. If the Affordable Care Act’s Readmissions Reduction Program continues to penalize hospitals with higher than expected readmission rates, it should adjust for surgical indication of lower extremity bypass. CONCLUSIONS Lower extremity bypasses are performed for multiple indications, and the risk for readmission varies by the indication for surgery. Operations performed for CLI are associated with the highest rates of readmission. If hospitals are to be penalized for readmission rates that exceed the predictions, the predicted readmission rates should account for the surgical indication. AUTHOR CONTRIBUTIONS Conception and design: CJ, AG, MM Analysis and interpretation: CJ, JR, DC, AG, BP, MM Data collection: CJ, AG Writing the article: CJ Critical revision of the article: CJ, JR, DC, BP, MM Final approval of the article: CJ, JR, DC, AG, BP, MM
JOURNAL OF VASCULAR SURGERY Volume -, Number -
Statistical analysis: CJ, JR, AG Obtained funding: MM Overall responsibility: CJ
REFERENCES 1. Jencks SF, Williams MV, Coleman EA. Rehospitalizations among patients in the Medicare fee-for-service program. N Engl J Med 2009;360:1418-28. 2. McPhee JT, Nguyen LL, Ho KJ, Ozaki CK, Conte MS, Belkin M. Risk prediction of 30-day readmission after infrainguinal bypass for critical limb ischemia. J Vasc Surg 2013;57:1481-8. 3. Orr NT, El-Maraghi S, Korosec RL, Davenport DL, Xenos ES. Cost analysis of vascular readmissions after common vascular procedures. J Vasc Surg 2015;62:1281-7.e1. 4. Davenport DL, Zwischenberger BA, Xenos ES. Analysis of 30-day readmission after aortoiliac and infrainguinal revascularization using the American College of Surgeons National Surgical Quality Improvement Program data set. J Vasc Surg 2014;60:1266-74. 5. Gupta PK, Fernandes-Taylor S, Ramanan B, Engelbert TL, Kent KC. Unplanned readmissions after vascular surgery. J Vasc Surg 2014;59: 473-82. 6. Goshima KR, Mills JL Sr, Hughes JD. A new look at outcomes after infrainguinal bypass surgery: traditional reporting standards systematically underestimate the expenditure of effort required to attain limb salvage. J Vasc Surg 2004;39:330-5. 7. American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP). 2013. Available at: http://www.facs.org/ quality-programs/acs-nsqip. Accessed June 2015. 8. Sellers MM, Merkow RP, Halverson A, Hinami K, Kelz RR, Bentrem DJ, et al. Validation of new readmission data in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2013;216:420-7. 9. Shiloach M, Frencher SK Jr, Steeger JE, Rowell KS, Bartzokis K, Tomeh MG, et al. Toward robust information: data quality and interrater reliability in the American College of Surgeons National Surgical Quality Improvement Program. J Am Coll Surg 2010;210:6-16.
Jones et al 7
10. Jackson BM, Nathan DP, Doctor L, Wang GJ, Woo EY, Fairman RM. Low rehospitalization rate for vascular surgery patients. J Vasc Surg 2011;54:767-72. 11. Weber SM, Greenberg CC. Medicare Hospital Readmission Reduction Program: what is the effect on surgery? Surgery 2014;156:1066-8. 12. McPhee JT, Barshes NR, Ho KJ, Madenci A, Ozaki CK, Nguyen LL, et al. Predictive factors of 30-day unplanned readmission after lower extremity bypass. J Vasc Surg 2013;57:955-62. 13. Ata A, Lee J, Bestle SL, Desemone J, Stain SC. Postoperative hyperglycemia and surgical site infection in general surgery patients. Arch Surg 2010;145:858-64. 14. Damrauer SM, Gaffey AC, DeBord Smith A, Fairman RM, Nguyen LL. Comparison of risk factors for length of stay and readmission following lower extremity bypass surgery. J Vasc Surg 2015;62: 1192-200.e1. 15. Wiseman JT, Fernandes-Taylor S, Barnes ML, Saunders RS, Saha S, Havlena J, et al. Predictors of surgical site infection after hospital discharge in patients undergoing major vascular surgery. J Vasc Surg 2015;62:1023-31.e5. 16. Zhang JQ, Curran T, McCallum JC, Wang L, Wyers MC, Hamdan AD, et al. Risk factors for readmission after lower extremity bypass in the American College of Surgeons National Surgery Quality Improvement Program. J Vasc Surg 2014;59:1331-9. 17. Mills JL Sr, Conte MS, Armstrong DG, Pomposelli FB, Schanzer A, Sidawy AN, et al. The Society for Vascular Surgery Lower Extremity Threatened Limb Classification System: risk stratification based on wound, ischemia, and foot infection (WIfI). J Vasc Surg 2014;59: 220-34. e1-2. 18. Zhan LX, Branco BC, Armstrong DG, Mills JL Sr. The Society for Vascular Surgery lower extremity threatened limb classification system based on Wound, Ischemia, and foot Infection (WIfI) correlates with risk of major amputation and time to wound healing. J Vasc Surg 2015;61:939-44. 19. Sellers MM, Reinke CE, Kreider S, Meise C, Nelis K, Volpe A, et al. American College of Surgeons NSQIP: quality in-training initiative pilot study. J Am Coll Surg 2013;217:827-32. Submitted Dec 26, 2015; accepted Mar 9, 2016.