Abdominal aortic aneurysm repair in octogenarians is associated with higher mortality compared with nonoctogenarians

From the Southern Association for Vascular Surgery

Abdominal aortic aneurysm repair in octogenarians is associated with higher mortality compared with nonoctogenarians Caitlin W. Hicks, MD, MS, Tammam Obeid, MD, Isibor Arhuidese, MD, MPH, Umair Qazi, MD, MPH, and Mahmoud B. Malas, MD, MHS, Baltimore, Md Objective: Age is a well-known independent risk factor for death after abdominal aortic aneurysm (AAA) repair. However, there is significant debate about the utility of AAA repair in older patients. In this study, mortality outcomes after endovascular AAA repair (EVAR) and open AAA repair (OAR) in octogenarians (aged $80 years) were compared with younger patients (aged <80 years). Methods: All patients recorded in the Vascular Quality Initiative database (2002-2012) who underwent infrarenal AAA repair were included. Univariable and multivariable statistics were used to compare perioperative (30-day) and 1-year mortality outcomes between octogenarians vs nonoctogenarians for OAR and EVAR. Results: During the study period, 21,874 patients underwent AAA repair (OAR, 5765; EVAR, 16,109), including 4839 octogenarians (OAR, 765; EVAR, 4074) and 17,035 nonoctogenarians (OAR, 5000; EVAR, 12,035). Octogenarians (mean age, 83.0 6 0.1 years) were less frequently male (66% vs 75%) and had a higher prevalence of congestive heart failure (9.9% vs 7.1%), chronic renal insufficiency (12.2% vs 7.5%), and a history of aortic surgery (14.3% vs 7.7%) compared with nonoctogenarians (P < .01 for all). Intraoperative use of blood transfusions and vasopressors was more common in octogenarians for OAR (blood: 3.3 6 4.4 vs 1.8 6 3.7 units; vasopressors: 45.2% vs 32.8%) and EVAR (blood: 0.43 6 1.7 vs 0.31 6 1.6 units; vasopressors: 7.6% vs 5.7%; P < .01 for all). Contrast dye volumes used during EVAR were similar in octogenarians and nonoctogenarians (108 6 71 vs 107 6 68 mL; P [ .18). Perioperative mortality after OAR was 20.1% in octogenarians compared with 7.1% in nonoctogenarians (P < .01). Perioperative mortality after EVAR was 3.8% in the octogenarians compared with 1.6% in nonoctogenarians (P < .01). One-year mortality among octogenarians vs nonoctogenarians was 26% vs 9.7% for OAR and 8.9% vs 4.3% for EVAR (log-rank test, P < .01 for both). Multivariable analysis controlling for baseline and intraoperative differences between groups demonstrated that age $80 years increased the risk of 30-day and 1-year mortality after AAA repair by 223% and 187%, respectively (P < .01 for both). Conclusions: AAA repair should be approached with extreme caution in octogenarians. Perioperative and 1-year mortality rates after OAR are particularly high in the older population, suggesting that the appropriate aneurysm size threshold for OAR might be larger due to the greater operative risk in octogenarian patients. (J Vasc Surg 2016;-:1-10.)

Open (OAR) and endovascular (EVAR) abdominal aortic aneurysm (AAA) repair are common procedures with reasonable outcomes among the general population.1 The risk of death among patients with unrepaired aneurysms increases exponentially with increasing aneurysm size: the 1-year annual estimated risk of rupture is 3% to 15%, 10% to 20%, 20% to 40%, and 30% to 50% for From the Division of Vascular and Endovascular Therapy, Johns Hopkins Medical Institutions. Author conflict of interest: none. Presented as a podium presentation at the Fortieth Annual Meeting of the Southern Association for Vascular Surgery, Cancun, Mexico, January 20-23, 2016. Additional material for this article may be found online at www.jvascsurg.org. Correspondence: Mahmoud B. Malas, MD, MHS, Department of Vascular and Endovascular Surgery, Johns Hopkins Bayview Medical Center, 4940 Eastern Ave, Bldg A/5, Ste 547, Baltimore, MD 21224 (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.2016.03.440

AAAs of 5 to 6 cm, 6 to 7 cm, 7 to 8 cm, and $8.0 cm, respectively, depending on other patient risk factors.2 As such, according the Society for Vascular Surgery (SVS) guidelines, patients should be advised to undergo repair of their AAA when the aneurysm diameter reaches 5.5 cm.2 Although the 5.5-cm AAA threshold was created with the “average” patient in mind, the applicability of the SVS AAA guidelines to higher-risk populations must be considered on a case-by-case basis.2 The appropriateness of these guidelines to an older population is especially unclear. Age is a well-known independent risk factor for death after AAA repair.3-6 Existing reports suggest that older patients have a significantly higher risk after elective AAA repair than younger patients, especially in the perioperative period.6 Although some studies suggest that this risk is not prohibitive to operative intervention, others suggest that, in many cases, it is.7-10 In contrast, nonoperative management of AAA within the older population also carries a substantial risk of death. Two-thirds of AAA ruptures have been demonstrated to occur in patients aged $75 years, with a resulting estimated 30-day mortality risk of 69%.11 As a result, considerable debate is focused on the utility of AAA repairs in the older population.12 1

JOURNAL OF VASCULAR SURGERY --- 2016

2 Hicks et al

In the current study, perioperative and 1-year mortality outcomes after EVAR and OAR were compared between octogenarians (aged $80 years) versus younger patients (aged <80 years). By quantifying the risk of death in the octogenarian vs younger populations, we may be better able to provide patients with an estimate of their surgical risk and inform surgeons about the appropriateness of pursuing aggressive surgical AAA repairs in elderly patients.

Owing to multiple differences in baseline characteristics between the octogenarian and nonoctogenarian groups, we also performed a nearest neighbor 1:1 propensity score matching on 26 clinical and technical covariates for the OAR and EVAR cohorts separately as a sensitivity analysis. Statistical significance was accepted at P < .05. Analyses were performed using Stata 14.1 software (StataCorp LP, College Station, Tex).

METHODS We performed a retrospective analysis of all patients in the Vascular Quality Initiative (VQI) database who underwent infrarenal AAA repair between January 2003 and September 2014. The VQI is a prospectively maintained database approved by the SVS.13 All patient data were obtained from institutions participating in VQI. The Johns Hopkins Institutional Review Board approved the study. Informed consent was waived because the data are publically available through the VQI. Only deidentified information was used in this study. Octogenarians were defined as patients aged $80 years. Outcomes studied included perioperative mortality (death #30 days of surgery) and 1-year all-cause mortality in octogenarians vs nonoctogenarians for OAR and EVAR procedures. Also recorded were postoperative complications, including myocardial infarction, respiratory complications (pneumonia or ventilator dependence), and renal dysfunction (>0.5 mg/dL change in creatinine or need for temporary or permanent dialysis). The VQI data set was linked to the Medicare claims database to obtain more accurate follow-up information. Descriptive analyses were performed using c2 and Student t-tests. Multivariate logistic regression models were constructed to identify predictors of perioperative and 1-year mortality. The models were adjusted for patient age (octogenarian vs nonoctogenarian), gender, race (white vs other), body mass index (kg/m2), EVAR vs OAR approach, urgency (elective, symptomatic, emergency), smoking (ever smoker vs nonsmoker), coronary artery disease, hypertension, diabetes, congestive heart failure, chronic obstructive pulmonary disease, peripheral arterial disease (comprising patients with a history of arterial bypass, peripheral vascular intervention, or carotid endarterectomy, as defined by the VQI data set), chronic kidney disease (defined as preoperative creatinine of >1.78 mg/dL according to the VQI definition), prior aortic repair, preoperative medications (aspirin and statin), baseline hemoglobin level, postoperative vasopressor use, and total number of packed red blood cells transfusion (units). Covariates for the models were chosen based on predictive variables from univariate analyses, prior literature, guidance of likelihood ratio tests, and Akaike information indices. Multiple imputation using multivariate normal data augmentation method14 was use to input values for variables with >5% missing data (baseline hemoglobin and red blood cell transfusion variables). Covariates used in imputation models included patient age, gender, urgency, vasopressor use, creatinine levels, and EVAR or OAR.

RESULTS Study cohort. During the study period, 21,874 patients underwent AAA repair. Of these, 5765 (26%) underwent OAR and 16,109 (74%) underwent EVAR, including 4839 octogenarians (22%; OAR, 765; EVAR, 4074) and 17,035 nonoctogenarians (78%; OAR, 5000; EVAR, 12,035; Table I). Overall, octogenarians (mean age, 83 years) were less frequently male (73% vs 81%), had a lower body mass index (26 vs 28 kg/m2), and a higher prevalence of smoking (90% vs 75%) than nonoctogenarians (P < .01 for all). Maximum AAA diameter at presentation was larger for octogenarians than for nonoctogenarians (60 6 15 mm vs 58 6 20 mm; P ¼ .01). Comorbidities were mostly similar between groups, with the exception of congestive heart failure (13% vs 10%), chronic renal insufficiency (9.4% vs 6.0%), and a history of aortic surgery (6.0% vs 4.6%), which were all higher in octogenarians (P < .01 for all). Octogenarians also had a worse functional status than nonoctogenarians: 2.9% vs 1.6% lived in a nursing home before surgery (P < .01), and 54% vs 58% had an independent ambulatory status (P < .01). These trends were similar for both the OAR and EVAR populations. A comparison of baseline characteristics for octogenarian vs nonoctogenarian patients stratified by type of AAA repair (OAR and EVAR) is reported in Table I. Perioperative factors. Intraoperative use of blood transfusions and vasopressors was more common in octogenarians for both OAR and EVAR (P < .01 for all; Table II). Accordingly, baseline hemoglobin values were lower and estimated blood loss in the operating room was higher for octogenarians than for nonoctogenarians, regardless of operative approach (P # .05 for all; Table II). Octogenarian patients undergoing OAR more frequently required supraceliac clamping (13.1% vs 9.5%; P < .01) and less frequently received heparin (88.7% vs 94.2%; P < .01) compared with nonoctogenarians. Use of cold renal perfusion and renal artery bypass was similar between groups (P ¼ not significant [NS]; Table II). For patients undergoing EVAR, octogenarians more frequently had endoleaks compared with nonoctogenarians (27.7% vs 23.0%; P < .01). Use of renal artery interventions (4.1% vs 3.4%) and contrast dye volumes (108 6 71 vs 107 6 68 mL) were similar between groups (P ¼ NS; Table II). For both OAR and EVAR approaches, octogenarians required more overall blood transfusions during the course of their hospital stay and required longer stays in the

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Hicks et al 3

Table I. Baseline characteristics of octogenarians vs nonoctogenarians undergoing open abdominal aortic aneurysm (AAA) repair (OAR) and endovascular AAA repair (EVAR) OAR (n ¼ 5765; 26%) Variablea Age, years Male White race BMI, kg/m2 ASA class 1 2 3 4 5 AAA diameter, mm Comorbidities Smoking Never Former Current CAD CHF Hypertension Diabetes COPD CKDc ESRD PAD Prior AAA repair Prior aortic surgery Preoperative medications Aspirin Statin b-Blocker ACE inhibitor Platelet antagonist Anticoagulation Living situation Home Nursing home Homeless Ambulation status Independent Assistance Wheelchair-bound Bedbound

Nonoctogenarian (n ¼ 5000; 87%)

Octogenarian (n ¼ 765; 13%)

68.0 6 0.11 75.0 92.5 27.5 6 5.6

83.01 6 0.19 66.0 94.8 25.7 6 4.7

0.5 5.4 55.0 32.7 6.5 62 6 22

0.2 3.6 44.0 40.2 12.0 68 6 18

9.1 43.3 47.6 27.0 7.1 82.5 15.1 33.6 7.5 0.8 10.2 6.6 7.7

24.4 56.4 19.2 28.2 9.9 84.1 15.8 33.3 12.2 0.4 11.0 13.5 14.3

64.3 64.7 67.6 41.1 7.3 8.3

61.3 57.4 64.0 40.0 6.8 18.9

99.3 0.7 0.1

98.0 2.0 0

95.2 3.9 0.4 0.4

88.4 8.2 1.7 1.7

EVAR (n ¼ 16,109; 74%) P valueb <.01 .03 <.01 <.01

<.01

Nonoctogenarian (n ¼ 12,035; 75%)

Octogenarian (n ¼ 4074; 25%)

69.6 6 0.06 82.9 91.5 28.5 6 5.8

83.5 6 0.04 74.8 92.2 26.5 6 5.2

0.5 6.8 65.7 25.2 1.8 56 6 20

0.3 5.2 65.8 27.0 1.7 58 6 14

10.3 50.6 39.1 30.0 10.9 83.6 21.0 33.5 5.4 1.3 9.7 3.7 4.2

25.2 60.9 13.9 30.0 13.6 84.9 18.3 29.4 8.9 1.0 10.4 4.6 5.9

65.4 68.3 61.8 43.5 11.7 10.5

62.3 66.8 61.4 41.3 10.9 16.3

99.0 1.0 0.1

97.5 2.4 0.1

94.0 4.9 0.8 0.4

86.8 11.8 1.4 0.1

<.01

.51 .01 .28 .61 .85 <.01 .26 .49 <.01 <.01 .11 <.01 .05 .69 .56 <.01 <.01

<.01

P valueb <.01 .12 <.01 <.01

<.01 <.01

.93 <.01 .04 <.01 <.01 <.01 .21 .21 .02 <.01 <.01 .09 .63 .05 .18 <.01 <.01

<.01

ACE, Angiotensin converting enzyme; ASA, American Society of Anesthesiologists; BMI, body mass index; CAD, coronary artery disease; CHF, congestive heart failure; COPD, chronic obstructive pulmonary disease; CKD, chronic kidney disease; ESRD, end-stage renal disease; PAD, peripheral arterial disease. a Continuous data are shown as the mean 6 standard deviation and categoric data as percentage. b Bold values indicate statistical significance (P < .05). c Defined as creatinine >1.78 mg/dL per the Vascular Quality Initiative (VQI) database.

intensive care unit than nonoctogenarians (P # .05 for all; Table II). Outcomes. Perioperative (30-day) mortality after OAR was 20.1% in octogenarians and 7.1% in nonoctogenarians (P < .01). Perioperative mortality after EVAR was 3.8% in octogenarians and was 1.6% in nonoctogenarians (P < .01). When stratified by urgency, perioperative mortality following OAR was 40.8% and 6.7% among octogenarians and 19.4% and 2.8% among nonoctogenarians for emergent and elective AAA repairs, respectively (P < .01 for both; Table III).

Similarly, perioperative mortality following EVAR was 17.4% and 1.6% among octogenarians versus 8.2% and 0.6% among nonoctogenarians for emergent and elective AAA repair, respectively (P < .01 for both; Table III). Accordingly, perioperative complications were higher in the octogenarian groups than in the nonoctogenarian groups for OAR (41% vs 30%; P < .01) and EVAR (10% vs 6.5%; P < .01; Supplementary Table, online only). Overall all-cause 1-year mortality among octogenarians vs nonoctogenarians was 26.0% vs 9.7% for OAR (Fig, A)

JOURNAL OF VASCULAR SURGERY --- 2016

4 Hicks et al

Table II. Procedural factors for octogenarians vs nonoctogenarians undergoing elective open abdominal aortic aneurysm (AAA) repair (OAR) and endovascular AAA repair (EVAR) OAR (n ¼ 5765; 26%) Variablea Intraoperative factors Anesthesia General General with epidural Regional Local Baseline hemoglobin, g/dL Estimated blood loss, mL Crystalloid, mL Contrast volume, mL RBCs transfused, units Total procedure time, minutes OAR technical factors Proximal clamp position Supraceliac Above 1 renal artery Above 2 renal arteries Infrarenal Heparin given Mannitol given Cold renal perfusion Concomitant procedure Thromboembolectomy Infrainguinal bypass Renal bypass Other EVAR technical factors Graft body diameter, mm Endoleak Conversion to open repair Concomitant procedure Hypogastric coiling Iliac artery angioplasty/stent Renal angioplasty/stent In-hospital factors Total RBCs, units RBCs transfusion >3 units Vasopressors required ICU length of stay, days

Nonoctogenarian (n ¼ 5000; 87%)

Octogenarian (n ¼ 765; 13%)

56.4 43.6

64.1 35.9

EVAR (n ¼ 16,109; 74%) P valueb

Nonoctogenarian (n ¼ 12,035; 75%)

Octogenarian (n ¼ 4074; 25%)

91.5

88.7

3.8 4.6 13.6 6 2.4 257 6 589 2029 6 1188 107 6 67.6 0.31 6 1.56 142 6 72.9

5.4 5.8 12.7 6 2.0 280 6 596 1954 6 1142 108 6 71.0 0.43 6 1.66 149 6 77.2

<.01 .04 <.01 .18 <.01 <.01

27.8 6 4.3 23.0 0.3 33.0 7.7 12.7 3.4

28.0 6 4.2 27.7 0.3 33.5 7.4 11.3 4.1

<.01 <.01 .51 .59 .48 .02 .05

0.50 6 2.34

0.66 6 2.17

0.93 6 2.74

1.05 6 2.69

<.01 <.01 <.01 .01

<.01

13.3 6 2.2 2013 6 2228 4876 6 2740 e 1.83 6 3.74 235 6 103

12.2 6 2.1 2438 6 2532 4956 6 3019 e 3.26 6 4.42 219 6 95

<.01 <.01 .44 <.01 <.01

P valueb <.01

<.01 9.5 12.3 18.4 59.9 94.2 48.6 6.3 26.0 7.8 7.5 2.6 11.6

2.29 6 4.98 23.6 32.9 4.79 6 8.13

<.01 .88 .76 .51 .77 .44 .18 .96

3.89 6 5.59 41.6 45.2 6.03 6 8.25

<.01 <.01 <.01 <.01

ICU, Intensive care unit; RBCs, red blood cells. a Continuous data are shown as the mean 6 standard deviation and categoric data as percentage. b Bold values indicate statistical significance (P < .05).

and 8.9% vs 4.3% for EVAR (log-rank test, P < .01 for both; Fig, B). When stratified by urgency, 1-year mortality following OAR was 47.8% and 11.9% among octogenarians vs 23.0% and 5.0% among nonoctogenarians for emergent and elective AAA repairs, respectively (P < .01 for both; Table III). Similarly, 1-year mortality following EVAR was 25.5% and 6.2% among octogenarians vs 13.3% and 2.9% among nonoctogenarians for emergent and elective AAA repairs, respectively (P < .01 for both; Table III). Multivariable analyses controlling for baseline and intraoperative differences between groups demonstrated that octogenarians undergoing OAR had a 2.1 (95% confidence interval [CI], 1.54-2.83) higher odds of perioperative death

and a 2.1 (95% CI, 1.60-2.66) higher odds of death #1 year compared with nonoctogenarians. Similarly, octogenarians undergoing EVAR had a 2.5 (95% CI 1.85-3.42) higher odds of perioperative death and a 1.79 (95% CI, 1.512.12) higher odds of death #1 year compared with their younger counterparts. Overall, age $80 years increased the risk of 30-day and 1-year mortality after AAA repair by 223% (95% CI, 1.80-2.76) and 187% (95% CI, 1.632.15), respectively (P < .01 for both). Among other factors significantly associated with death, postoperative vasopressor use, emergent/urgent repair, octogenarian status, and white race were the strongest predictors of perioperative and 1-year mortality after AAA repair regardless of operative approach (Table IV). A subanalysis restricted to only

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Hicks et al 5

Table III. Outcomes for octogenarians vs nonoctogenarians undergoing open abdominal aortic aneurysm (AAA) repair (OAR) and endovascular AAA repair (EVAR) OAR (n ¼ 5765)

EVAR (n ¼ 16,109)

Overall (N ¼ 21,874)

Nonoctogenarian Octogenarian Nonoctogenarian Octogenarian Nonoctogenarian Octogenarian (n ¼ 12,035; (n ¼ 4074; P (n ¼ 17,035; (n ¼ 4839; P (n ¼ 5000; (n ¼ 765; P b b 13%) 25%) value 78%) 22%) valueb 87%) 13%) value

a

Outcome

Perioperative (30-day) mortality Elective repair Emergent repair One-year all-cause mortality Elective repair Emergent repair

7.1

20.1

<.01

1.6

3.8

<.01

3.2

6.4

<.01

2.8 19.4

6.7 40.8

<.01 <.01

0.6 8.2

1.6 17.4

<.01 <.01

1.2 13.2

2.2 25.5

<.01 <.01

9.7

26.0

<.01

4.3

8.9

<.01

5.9

11.6

<.01

5.0 23.0

11.9 47.8

<.01 <.01

2.9 13.3

6.2 25.5

<.01 <.01

3.4 17.7

6.9 33.2

<.01 <.01

a

Data are shown in percentages. Bold values indicate statistical significance (P < .05).

b

0 Number at risk Non-oct 5000 Oct 765

40

Percent Surviving

60 80 100

One- Year Kaplan-Meier Survival Estimates - EVAR

B

0

20

60 20 40 0

Percent Surviving

80 100

One- Year Kaplan-Meier Survival Estimates - OAR

A

2 3567 481

4 6 8 Months after surgery 3367 453

3224 422

Non-oct

3106 399

10

12

2969 378

2812 354

Oct

0 Number at risk Non-oct 12034 Oct 4074

2 8228 2771

4 6 8 Months after surgery 7580 2560

7135 2399

Non-oct

6705 2235

10

12

6254 2066

5713 1876

Oct

Fig. Kaplan-Meier curves for 1-year survival after abdominal aortic aneurysm (AAA) repair. One-year all-cause mortality was significantly worse for octogenarians than for nonoctogenarians following both (A) open AAA repair (OAR) and (B) endovascular AAA repair (EVAR) approaches (log-rank test, P < .001 for both).

patients undergoing open infrarenal AAA repair showed similar results (data not shown). Propensity score-matched outcomes. A sensitivity analysis using propensity matching on 26 baseline, perioperative, and technical variables, demonstrated that the survival differences between octogenarians vs nonoctogenarians described above were reduced, particularly among patients undergoing OAR. Perioperative (30-day) mortality after OAR was 13.0% in the octogenarian group compared with 10.3% in nonoctogenarians (P ¼ .14). Perioperative mortality after EVAR was 2.8% in the octogenarian group compared with 1.9% in nonoctogenarian group (P ¼ .01). When stratified by emergent vs elective AAA repair, perioperative mortality was 26.4% and 6.2% among octogenarians and 18.6% and 6.0% among

nonoctogenarians undergoing OAR, respectively (P ¼ NS for both; Table V). In contrast, perioperative mortality after EVAR was 12.9% and 1.37% among octogenarians vs 8.7% and 0.79% among nonoctogenarians for emergent (P ¼ .05) and elective (P ¼ .06) AAA repairs, respectively (Table V). After propensity matching, overall all-cause 1-year mortality among octogenarians vs nonoctogenarians was 19.1% vs 13.5% for OAR and 7.0% vs 5.0% for EVAR (P # .01 for both). When stratified by urgency, 1-year mortality following OAR was 34.8% and 11.3% among octogenarians vs 23.0% and 8.57% among nonoctogenarians undergoing emergent (P ¼ .01) and elective (P ¼ .23) AAA repair, respectively (Table V). One-year mortality after EVAR was 19.1% and 5.2% among octogenarians vs 14.6% and 3.5% among

JOURNAL OF VASCULAR SURGERY --- 2016

6 Hicks et al

Table IV. Multivariable analysis of factors associated with perioperative (30-day) and 1-year mortality after abdominal aortic aneurysm (AAA) repair Perioperative (30-day) mortalitya Variable Octogenarian (age $80 years) Female gender White race Body mass index (per kg/m2) EVAR (vs OAR) Urgency Elective Symptomatic Emergent Smoking Coronary artery disease Hypertension Diabetes Congestive heart failure COPD Chronic kidney diseased Peripheral arterial disease Prior AAA repair Aspirin Statin Baseline hemoglobin (per g/dL) Postoperative vasopressor requirement Transfusion during hospital stay (units)

OR (95% CI)

P valuec

One-year mortalityb OR (95% CI)

P valuec

(1.80-2.76) (1.03-1.59) (1.39-3.15) (0.97-1.00) (0.63-0.97)

<.01 .03 <.01 .09 .02

1.87 1.03 1.87 0.97 1.04

(1.63-2.15) (0.89-1.20) (1.44-2.44) (0.96-0.98) (0.89-1.22)

<.01 .67 <.01 <.01 .57

Reference 1.22 (0.87-1.70) 4.31 (3.28-5.47) 0.98 (0.74-1.30) 1.21 (1.01-1.59) 1.27 (0.97-1.51) 0.84 (0.64-1.09) 1.46 (1.22-2.12) 1.69 (1.37-2.06) 1.43 (1.10-1.90) 1.57 (1.19-2.09) 1.27 (0.90-1.78) 0.91 (0.73-1.14) 0.95 (0.74-1.16) 0.90 (0.86-0.94) 5.57 (4.69-7.39) 1.13 (1.11-1.16)

.25 <.01 .99 .12 .10 .22 <.01 <.01 <.01 <.01 .17 .43 .49 <.01 <.01 <.01

Reference 1.29 (1.05-1.58) 2.31 (1.90-2.80) 0.98 (0.81-1.18) 1.19 (1.03-1.37) 1.01 (0.85-1.21) 1.05 (0.89-1.24) 1.86 (1.57-2.21) 1.65 (1.45-1.09) 1.91 (1.58-2.30) 1.19 (0.98-1.45) 1.19 (0.93-1.52) 0.83 (0.72-0.95) 0.81 (0.71-0.94) 0.85 (0.83-0.88) 2.53 (2.16-2.97) 1.11 (1.10-1.13)

.02 <.01 .81 .02 .88 .56 <.01 <.01 <.01 .08 .16 <.01 <.01 <.01 <.01 <.01

2.23 1.27 2.07 0.99 0.78

CI, Confidence interval; COPD, chronic obstructive pulmonary disease; EVAR, endovascular aneurysm repair; OAR, open aneurysm repair; OR, odds ratio. a C statistic ¼ 0.92 for multivariable models without imputation. b C statistic ¼ 0.83 for multivariable models without imputation. c Bold values indicate statistical significance (P < .05). d Defined as creatinine >1.78 mg/dL per the Vascular Quality Initiative (VQI) database.

nonoctogenarians for emergent (P ¼ .10) and elective (P < .01) AAA repair, respectively (Table V). DISCUSSION The risks vs benefits of pursuing elective AAA in the elderly population is a topic of considerable debate in the literature. Although the potential for rupture among patients with enlarging aneurysms is real, the risks associated with both OAR and EVAR are not minimal. In the current study, we aimed to quantify this risk. We demonstrated that the risk of both perioperative and 1-year mortality after AAA repair among octogenarians is substantial, at 6.4% and 11.6%, respectively. Furthermore, age $80 years increased the risk of 30-day and 1-year mortality by 223% and 187%, respectively, compared with younger patients. This agerelated risk of death is independent of other patient risk factors. On the basis of these data, the risk of AAA repair among octogenarians may not outweigh the potential benefits of surgery in certain patients. The strongest predictors of death in this cohort were postoperative vasopressor requirements, urgent/emergent status, and octogenarian status. That vasopressor use and urgent/emergent status are associated with higher risk of mortality is not surprising because they are surrogate markers for patient instability. A number of prior reports on AAA outcomes have similarly reported that both

vasopressors15,16 and acuity17,18 are associated with high mortality risks. In contrast, octogenarian status as an independent predictor of mortality is relatively novel. Shahidi et al5 reported a mortality rate of 57% among patients aged $75 years undergoing ruptured AAA repair. Thomson et al19 reported an 8% risk of in-hospital death among patients aged >79 years undergoing elective OAR in a New Zealand population. However, most of the previously published studies on this topic were limited to describing perioperative mortality, did not perform risk adjustment, or use single-institution or nonvascular national databases.6 In the current study, we used the VQI, a vascular-specific national database that was established by the SVS and is maintained by active vascular surgeons,13 to report perioperative and 1-year mortality outcomes among a large group of patients undergoing AAA repair. Generally, AAA repair operations are performed at a point when AAA rupture risk outweighs operative risk. Lederle et al20 reported that the 1-year risk of AAA rupture ranges from 9% to 32% among patients of all ages depending on aneurysm size. However, the average life expectancy among the United States population is only 78.8 years,21 meaning that the risk of death from any cause among an octogenarian cohort is substantial, even without the presence of an AAA. Furthermore, the all-cause mortality among patients aged $80 years in our study was 11.6%

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Hicks et al 7

Table V. Sensitivity analysis: outcomes after propensity score matching on 26 clinical and technical covariates OAR (n ¼ 1066) Outcomea Perioperative (30-day) mortality Elective repair Emergent repair One-year all-cause mortality Elective repair Emergent repair

EVAR (n ¼ 5842)

Nonoctogenarian (n ¼ 533)

Octogenarian (n ¼ 533)

P valueb

Nonoctogenarian (n ¼ 2921)

Octogenarian (n ¼ 2921)

10.3 (55)

13.0 (69)

.14

1.9 (54)

2.8 (83)

6.0 (21) 18.6 (34) 13.5 (72)

6.2 (22) 26.4 (47) 19.1 (102)

.93 .08 .01

0.8 (20) 8.7 (34) 5.0 (146)

1.4 (35) 12.9 (48) 7.0 (204)

.05 .06 <.01

8.6 (30) 23.0 (42)

11.3 (40) 34.8 (62)

.23 .01

5.2 (133) 19.1 (71)

<.03 .10

3.5 (89) 14.6 (57)

P valueb .01

EVAR, endovascular aneurysm repair; OAR, open aneurysm repair; OR, odds ratio. a Data are shown as percentage (No.). b Bold values indicate statistical significance (P < .05).

at 1 year, including 11.9% after OAR and 6.2% after EVAR for elective cases. These data suggest that, for certain older patients who may have a relatively short life expectancy, elective AAA repair may carry a limited survival benefit compared with watchful waiting. As such, we advocate for individualized patient risk assessment when determining the appropriateness for elective AAA repair, particularly within the octogenarian population. The etiology of the disparate outcomes between octogenarians and nonoctogenarians undergoing AAA repair is unclear. Our data suggest that older patients spend more time in the intensive care unit, receive more blood, and require more vasopressors postoperatively than their younger counterparts. It is possible that a worse baseline disease burden within the older population puts them at higher risk for postoperative complications from which they never recover. Certainly, octogenarian patients appear to require more difficult repairs, particularly with the open approach: more octogenarians require supraceliac clamping, and blood loss is higher in this cohort. This may speak to differences in vasculature structure that can occur with age, including the intimal thickening and fragmentation of elastic fibers that has been shown to increase over time.22 Consistent with this notion, adjusting for anatomic differences using propensity matching reduced the differences in mortality between the octogenarian and nonoctogenarian groups undergoing open AAA repair. As such, it appears that technical challenges may be the limiting factor for performing OAR safely in the in the older population. In contrast, we demonstrate that outcomes for EVAR are persistently worse among the octogenarian population even after propensity matching on 26 baseline, perioperative, and technical variables. The worse outcomes among octogenarians after EVAR could speak to a more general frailty of the older population as a whole. Frailty itself has been shown to be a risk factor for death after surgery, even independent of age.23,24 Although the VQI data set does not have a measure of frailty directly, octogenarian patients were significantly more dependent in their ambulatory status and living situation than the younger population.

It is also possible that older patients experience more postoperative complications or are less able to be rescued from those complications than younger patients. Morbidity after AAA repair was recently estimated to be 24% for OAR and 13% for EVAR in matched cohorts.25 Similarly, we report a morbidity rate of 30% and 7% among nonoctogenarians undergoing OAR and EVAR, respectively. Morbidity among octogenarians was significantly higher, occurring in 41% for OAR and in 10% for EVAR. We also note a steep decline in survival within the octogenarian cohort within the first 3 months after AAA repair (Fig), which suggests the differences in mortality that we observe occur early in the postoperative course. These findings deserve further evaluation; if a higher rate of postoperative complications is leading to higher failure-to-rescue rates in the elderly population, this may help explain their overall worse prognosis. No study that we know of to date has examined the difference in failure to rescue after AAA repair among octogenarians vs nonoctogenarians, but this would certainly be informative about ways in which we might better be able to predict and educate patients on their postoperative risk of death after AAA repair. The limitations of this study include our use of an administrative database, inability to account for anatomic AAA differences among patients, and the predefined age cutoff of $80 years. The VQI database is a relatively new database that suffers from poor follow-up in many instances. However, this issue has largely been resolved with the pairing of VQI data to Medicare outcomes data, which allows for a more accurate estimation of mortality.26 Aneurysm anatomy and the subsequent extent of surgical repair are known to play a role in determining postoperative morbidity and mortality.27 Although we do not have precise data on aneurysm angles, burden of atherosclerotic disease, reasons for certain patients undergoing an open vs endovascular approach, or graft types used, we were able to account for aortic cross-clamping location, use of cold renal perfusion or renal stenting, and graft body diameter with the VQI data set. Even after adjusting for

JOURNAL OF VASCULAR SURGERY --- 2016

8 Hicks et al

these factors, the mortality difference between older vs younger patients persisted. Finally, we used age $80 years as the cutoff in our study to define an octogenarian population because debate surrounding the appropriateness of aggressive surgical intervention in this population is prominent.12 It is possible that the balance between the risks vs benefits of operative AAA repair may actually fall at an alternative age cutoff. Age is a well-described risk factor for death after both OAR and EVAR approaches and has previously been defined as both a continuous and categoric variable.3-5 However, no data exist to describe the age at which AAA repair becomes prohibitively high risk. In a study of patients with end-stage renal disease, spline modeling showed that hemodialysis access outcomes differed at an age cutoff of 89 years.28 Similarly, spline modeling analyses may also be able to assist in determining a precise cutoff at which AAA repair becomes too risky to pursue. However, before such a study can be performed, the accepted costs and acceptable predicted risk value associated with AAA repair must be established and agreed upon. CONCLUSIONS The mortality risk associated with both OAR and EVAR is nearly double that for octogenarians compared with nonoctogenarians. Although the precise etiology of these differences may be different for the open and endovascular approaches, we suggest that AAA repair should be approached with extreme caution in the older population. Perioperative and 1-year mortality rates after OAR are particularly high in octogenarians, suggesting that the appropriate aneurysm size threshold for OAR might be larger due to the greater operative risk in these patients. Future studies using spline modeling and cost analyses, as well as a discussion about the acceptable predicted risk associated with operative repair in older patients, would be helpful in supporting the pursuit of formal guidelines on this topic. AUTHOR CONTRIBUTIONS Conception and design: CH, MM Analysis and interpretation: CH, TO, IA, UQ, MM Data collection: CH, TO, IA, UQ, MM Writing the article: CH, TO, MM Critical revision of the article: CH, TO, IA, UQ, MM Final approval of the article: CH, TO, IA, UQ, MM Statistical analysis: CH, TO, IA Obtained funding: Not applicable Overall responsibility: MM REFERENCES 1. Hallin A, Bergqvist D, Holmberg L. Literature review of surgical management of abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 2001;22:197-204. 2. Brewster DC, Cronenwett JL, Hallett JW Jr, Johnston KW, Krupski WC, Matsumura JS, et al. Guidelines for the treatment of

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13. 14. 15.

16.

17.

18.

19.

20.

21.

22.

abdominal aortic aneurysms. Report of a subcommittee of the Joint Council of the American Association for Vascular Surgery and Society for Vascular Surgery. J Vasc Surg 2003;37:1106-17. Ambler GK, Gohel MS, Mitchell DC, Loftus IM, Boyle JR; Audit and Quality Improvement Committee of the Vascular Society of Great Britain and Ireland. The Abdominal Aortic Aneurysm Statistically Corrected Operative Risk Evaluation (AAA SCORE) for predicting mortality after open and endovascular interventions. J Vasc Surg 2015;61:35-43. Tang T, Walsh SR, Prytherch DR, Lees T, Varty K, Boyle JR, et al. VBHOM, a data economic model for predicting the outcome after open abdominal aortic aneurysm surgery. Br J Surg 2007;94:717-21. Shahidi S, Schroeder TV, Carstensen M, Sillesen H. Outcome and survival of patients aged 75 years and older compared to younger patients after ruptured abdominal aortic aneurysm repair: do the results justify the effort? Ann Vasc Surg 2009;23:469-77. Henebiens M, Vahl A, Koelemay MJ. Elective surgery of abdominal aortic aneurysms in octogenarians: a systematic review. J Vasc Surg 2008;47:676-81. Biebl M, Lau LL, Hakaim AG, Oldenburg WA, Klocker J, Neuhauser B, et al. Midterm outcome of endovascular abdominal aortic aneurysm repair in octogenarians: a single institution’s experience. J Vasc Surg 2004;40:435-42. de Leur K, Flu HC, Ho GH, de Groot HG, Veen EJ, van der Laan L. Outcome of elective treatment of abdominal aortic aneurysm in elderly patients. Int J Surg 2015;15:117-23. Visser L, Pol RA, Tielliu IF, van den Dungen JJ, Zeebregts CJ. A limited and customized follow-up seems justified after endovascular abdominal aneurysm repair in octogenarians. J Vasc Surg 2014;59:1232-40. Hughes K, Abdulrahman H, Prendergast T, Rose DA, Ongu’ti S, Tran D, et al. Abdominal aortic aneurysm repair in nonagenarians. Ann Vasc Surg 2015;29:183-8. Howard DP, Banerjee A, Fairhead JF, Handa A, Silver LE, Rothwell PM, et al. Age-specific incidence, risk factors and outcome of acute abdominal aortic aneurysms in a defined population. Br J Surg 2015;102:907-15. Aitken SJ, Naganathan V, Blyth FM. Aortic aneurysm trials in octogenarians: are we really measuring the outcomes that matter [published online ahead of print July 28, 2015]? Vascular http://dx.doi.org/ 10.1177/1708538115597079. Cronenwett JL, Kraiss LW, Cambria RP. The Society for Vascular Surgery Vascular Quality Initiative. J Vasc Surg 2012;55:1529-37. Schafer JL. Analysis of incomplete multivariate data. Boca Raton, FL: Chapman & Hall/CRC Press; 1997. Bush HL Jr, Hydo LJ, Fischer E, Fantini GA, Silane MF, Barie PS. Hypothermia during elective abdominal aortic aneurysm repair: the high price of avoidable morbidity. J Vasc Surg 1995;21:392-400; discussion: 400-2. Kim GS, Ahn HJ, Kim WH, Kim MJ, Lee SH. Risk factors for postoperative complications after open infrarenal abdominal aortic aneurysm repair in Koreans. Yonsei Med J 2011;52:339-46. Rayt HS, Sutton AJ, London NJ, Sayers RD, Bown MJ. A systematic review and meta-analysis of endovascular repair (EVAR) for ruptured abdominal aortic aneurysm. Eur J Vasc Endovasc Surg 2008;36:536-44. Bastos Gonçalves F, Ultee KH, Hoeks SE, Stolker RJ, Verhagen HJ. Life expectancy and causes of death after repair of intact and ruptured abdominal aortic aneurysms. J Vasc Surg 2016;63:610-6. Thomson IA, Goh F, Livingstone V, van Rij AM. How safe is open abdominal aortic aneurysm surgery for octogenarians in New Zealand? ANZ J Surg 2009;79:344-7. Lederle FA, Johnson GR, Wilson SE, Ballard DJ, Jordan WD Jr, Blebea J, et al. Rupture rate of large abdominal aortic aneurysms in patients refusing or unfit for elective repair. JAMA 2002;287:2968-72. Centers for Disease Control and Prevention (CDC). FastStats: life expectancy. Available at: http://www.cdc.gov/nchs/fastats/life-expectancy. htm. Accessed December 2, 2015. Zarkovic K, Larroque-Cardoso P, Pucelle M, Salvayre R, Waeg G, Nègre-Salvayreet A, et al. Elastin aging and lipid oxidation products in human aorta. Redox Biol 2015;4:109-17.

JOURNAL OF VASCULAR SURGERY Volume -, Number -

23. 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. 24. Srinivasan A, Ambler GK, Hayes PD, Chowdhury MM, Ashcroft S, Boyle JR, et al. Premorbid function, comorbidity, and frailty predict outcomes after ruptured abdominal aortic aneurysm repair. J Vasc Surg 2016;63:603-9. 25. Huang Y, Gloviczki P, Oderich GS, Duncan AA, Kalra M, Fleming MD, et al. Outcome after open and endovascular repairs of abdominal aortic aneurysms in matched cohorts using propensity score modeling. J Vasc Surg 2015;62:304-11.e2. 26. Vascular Quality Initiative. M2SÒ and SVSÒ establish a qualified clinical data registry for participants in the Vascular Quality InitiativeÒ. Available at: http://www.vascularqualityinitiative.org/m2s-and-svs-establish-

Hicks et al 9

a-qualified-clinical-data-registry-for-participants-in-the-vascular-qualityinitiative/. Accessed December 2, 2015. 27. Katsargyris A, Oikonomou K, Klonaris C, Topel I, Verhoeven EL. Comparison of outcomes with open, fenestrated, and chimney graft repair of juxtarenal aneurysms: are we ready for a paradigm shift? J Endovasc Ther 2013;20:159-69. 28. Hicks CW, Canner JK, Arhuidese I, Zarkowsky DS, Qazi U, Reifsnyder T, et al. Mortality benefits of different hemodialysis access types are age dependent. J Vasc Surg 2015;61:449-56.

Submitted Dec 20, 2015; accepted Mar 17, 2016.

Additional material for this article may be found online at www.jvascsurg.org.

DISCUSSION Dr Samuel R. Money (Phoenix, Ariz). Thank you Dr Hicks and colleagues for sending me the paper. I especially appreciate how early I received it. This is a well-written paper. I will not fully summarize the paper here; however, it did point out significantly quite high mortality among octogenarians who are undergoing abdominal aortic aneurysm (AAA) repair both by open and endovascular means. This paper hits especially close to home as I have been wondering if we could do this project ourselves, as every year around Christmas time, an octogenarian comes in with a ruptured AAA and we stay up all night, try to save this patient, and in the vast majority of the time, they pass away within a day or two. My first question is you discuss rupture rates and average life expectancy; however, if one were to make it to 85, their life expectancy is not a year or two, but it puts them well over 90. You talk about the mean life expectancy in America being 78 years; however, if one makes it to 85, they are probably going to make it to over 90. Please comment on this. Second question, with a very high mortality rate achieved in patients who underwent open repair, should we even attempt this in a ruptured aneurysm? Should we attempt to do endovascular repair in octogenarians with ruptured aneurysms and if this is not successful, possibly let the patient go. I know this goes against our creed of doing everything we can, but looking at the numbers, the mortality is tremendously high, and my guess is the rate of independent living is not that high in these patients after 1 month, 2, 3, 4, even 6 months following a ruptured AAA done open. Question three. Maybe we should expand that, should we just do endovascular repair on anybody over 80? If they need a supraceliac clamping, should we do a fenestrated graft? Should we basically abandon open AAA repair in octogenarians? In other words, are we simply doing too much surgery in an open fashion in this group? Finally, I was amazed at only 54% to 58% of those 80 or above who undergo AAA repair, whether open or endovascular AAA repair (EVAR) had independent ambulatory status. Approximately half of the patients we operate on over 80 cannot ambulate independently. Are we doing too much surgery? Are we operating on patients who probably would not need surgery? And, are you surprised that approximately half of the AAA patients over 80 could not walk independently? Dr Caitlin W. Hicks. Thank you for your questions. Definitely all of these points are important aspects to be considered. I think the overarching themes that come out of this series of

questions are really, “Who do we operate on?” and “How do we decide when to operate?” I think that the data represented in this study do not tell us one way or another necessarily. It just suggests that risks are higher in the older population group and that when we are choosing operative candidates we need to be aware of that and perhaps counsel these patients accordingly. Certainly, the risk of death after rupture was extremely high in our study, but the risk of death with unrepaired rupture is almost a certainty so that has to be considered as well. I think given the fact that we found mortality to be quite a bit lower following EVAR compared to open AAA repairs within the octogenarian group, you could make an argument that EVAR should definitely be performed preferentially in older patients whenever possible. There are certainly patients that are not candidates for EVARs that need consideration for open AAA repairs, and in those cases, I think it is a patient-based decision; you have to look at the patient’s protoplasm and whether or not they are likely to do well, discuss with the patient and the family all of the potential risks, and then proceed with making a decision about whether to operate based on that discussion. In terms of life expectancy question, it is true that the U.S. life expectancy is 78 years of age. Patients that live past that age are projected to live longer. Again, I think we cannot necessarily capture all of the factors that go into a patient’s condition for surgery just by looking at age. Age is just a single surrogate for fitness, and that is where the recent interest in frailty comes up. I think that using a measure of frailty would be a reasonable way to go about deciding whether an older patient should undergo AAA repair, and that is addressed somewhat by our ambulatory data. We did have a high prevalence of lack of independent ambulation with the octogenarian population in our study. I will say that the Vascular Quality Initiative (VQI) has since abandoned that variable in their AAA modeldit only existed for 2 yearsdso whether it is accurate, I do not actually know, but trying to capture some other means of frailty may be worthwhile in helping to determine operative candidacy. In terms of hypotension being a reasonable indication for EVAR, I do think that is reasonable. Again, we probably need to consider how the patient’s baseline functional status is before we offer an open repair to these higher-risk patients. I think the next step to use these data is to do a spline modeling type of analysis to try to determine exactly where this age cutoff is. We have started to do that, but it is more complex than you think. Given that we have not used the VQI database

10 Hicks et al

before to show that octogenarians do worse after AAA repair than nonoctogenarians, we have no sense of what our baseline is. In this analysis we used age cutoffs of 75, 80, 85, and 90 years. It seemed like at 75 there was a nonsignificant trend in a number of areas, at 85 the findings were also significant, and at 90 you really don’t have enough outcomes to be able to correct for the number of baseline characteristics we needed to so we stuck with our use of age 80 years for the purposes of this study. We need to do a more sophisticated analysis of that to find exactly where the ideal age cutoff would be, but again, I think age is only one surrogate

JOURNAL OF VASCULAR SURGERY --- 2016

value to consider. I do not think we are capturing all aspects of a patient’s appropriateness for surgery using age alone, so having an exact age cutoffdwhether it is 78 or 82dI think is largely irrelevant. This study was imperative before we could do the next study where we look at a measure of frailty and spline modeling, as it forms the basis for future studies to hopefully expand on our current findings and better inform our operative decision making. I think we do need to expand our analyses to develop some sort of frailty model to get a better capture of a global sense of patient status.

JOURNAL OF VASCULAR SURGERY Volume -, Number -

Hicks et al 10.e1

Supplementary Table (online only). In-hospital complications for octogenarians vs nonoctogenarians after open abdominal aortic aneurysm (AAA) repair (OAR) or endovascular AAA repair (EVAR) OAR (n ¼ 5765) Variablea Overall Elective repair Emergent repair Myocardial infarction Elective repair Emergent repair Respiratory complicationc Elective repair Emergent repair Renal dysfunctiond Elective repair Emergent repair a

Nonoctogenarian (n ¼ 5000; 87%) 30.1 23.5 49.1 6.6 4.6 12.3 16.0 10.7 31.6 18.8 14.8 30.5

(1479) (867) (611) (324) (171) (153) (789) (396) (393) (922) (545) (376)

Octogenarian (n ¼ 765; 13%) 41.1 30.2 59.1 8.9 6.5 13.0 25.7 16.0 41.9 26.8 19.9 38.5

(305) (140) (165) (66) (30) (36) (190) (74) (116) (197) (92) (105)

EVAR (n ¼ 16,109) P valueb <.01 <.01 <.02 <.02 .09 .76 <.01 <.01 <.01 <.01 <.01 .01

Data are shown as percentage (No.). Values in bold are statistically significant (P < .05). c Defined as pneumonia or ventilator dependence. d Defined as >0.5 mg/dL change in creatinine or need for temporary or permanent dialysis. b

Nonoctogenarian (n ¼ 12,035; 75%) 6.5 4.2 21.9 1.5 0.8 6.3 2.9 1.4 12.5 4.0 2.8 11.8

(778) (435) (339) (179) (82) (97) (342) (146) (194) (142) (286) (182)

Octogenarian (n ¼ 4074; 25%) 10.4 7.6 27.8 2.5 1.8 6.5 4.3 2.7 14.1 6.6 4.8 17.8

(421) (265) (154) (100) (64) (36) (173) (95) (78) (267) (167) (98)

P valueb <.01 <.01 <.01 <.01 <.01 .84 <.01 <.01 .38 <.01 <.01 <.01