Journal Logo

ORIGINAL ARTICLES

Association Between Hospital Volume and Failure to Rescue After Open or Endovascular Repair of Intact Abdominal Aortic Aneurysms in the VASCUNET and International Consortium of Vascular Registries

D’Oria, Mario MD∗,†; Scali, Salvatore MD; Mao, Jialin MD, MS§; Szeberin, Zoltán MD, PhD; Thomson, Ian MB ChB, FRACS||; Beiles, Barry MB BCh, FRACS∗∗; Stone, David MD††; Sedrakyan, Art MD, PhD§; Eldrup, Nikolaj MD, PhD‡‡; Venermo, Maarit MD, PhD§§; Cassar, Kevin MD, FRCS(Ed)¶¶; Altreuther, Martin MD, PhD||||; Boyle, Jonathan R. MB ChB, MD, FRCS∗∗∗; Behrendt, Christian-Alexander MD†††; Beck, Adam W. MD‡‡‡; Mani, Kevin MD, PhD

Author Information
doi: 10.1097/SLA.0000000000005044

Abstract

It is well established that in-hospital mortality after abdominal aortic aneurysm (AAA) repair is lower when high-volume institutions perform the operations.1–3 This has led major international vascular societies from North America and Europe to issue statements on recommended thresholds for procedural volume in clinical practice guidelines, strongly advising that AAA surgery be performed only in institutions which achieve such thresholds.4,5 Although this observation is evident for open aortic repair (OAR), the volume-outcome relationship with endovascular aortic repair (EVAR) is less certain, in part due to the low postoperative mortality rates associated with endovascular procedures.6 Notably, the underlying mechanism through which higher volume centers achieve better outcomes for either OAR or EVAR is poorly understood and several explanations have been proposed.7

Interestingly, postoperative mortality is associated with many factors and stakeholders worldwide have developed aggregate outcome measures that usually focus on processes of care, which may reflect a more sensitive and specific method for measuring health care quality. For example, major complications are known to predict postoperative mortality and the ability to rescue patients experiencing adverse outcomes before death may be a plausible explanation for how higher volume centers achieve improved mortality rates compared to lower volume centers.8 Consequently, failure to rescue (FtR), which is a composite outcome measure defined as postoperative death during the same hospital admission after experiencing at least 1 major complication, could be a major driver of mortality after both OAR and EVAR.

Indeed, FtR can be proposed as quality indicator in elective AAA repair given its potential ability to better differentiate variation in quality of care delivery between centers. This composite outcome suggests that in cases of complications resulting in postoperative mortality, more focus on prevention and/or timely management of such complications may improve outcomes for AAA repair.

The purpose of this study was to investigate the association between complications and mortality after intact AAA repair, with specific focus on determining the relationship between hospital volume and FtR after OAR or EVAR in a multinational dataset from the VASCUNET and International Consortium of Vascular Registries (ICVR).9

METHODS

Data Sources

Data from prospectively maintained vascular surgery quality registries from participating countries were evaluated. As previously described, de-identified patient-level data or aggregate-level data were provided according to state(s) regulations.1 Data on primary elective OAR and EVAR AAA procedures from 2010 to 2016 were combined across all participating countries for analysis. The analysis included intact aneurysm repairs (both asymptomatic and symptomatic) but excluded those with rupture as index presentation.

Study Design

A complete dataset was constructed after combining data from 8 participating national or regional vascular registries that provided adequate data for analysis. Only intact AAA repairs were included in the study. Additional details about individual registries and proportion of national coverage have been previously published.1,3,6, In brief, the registries from Australia, Denmark, Hungary, Malta, New Zealand, and Sweden are national population-based vascular surgical quality registries; the registry from Finland includes all vascular procedures performed in the Helsinki region; and the Vascular Quality Initiative is a national network of regional quality groups made up of >370 North American academic and community hospitals from United States and Canada (https://www.vqi.org/data-analysis). Preoperative characteristics registered included age, sex, history of diabetes, cardiac history, renal disease, and preoperative aneurysm diameter size. Reported major complications included (1) postoperative bleeding requiring a return to the operation room, (2) stroke, (3) acute cardiac event (acute coronary syndrome, myocardial infarction, serious arrhythmia, and/or cardiac failure), (4) respiratory failure requiring mechanical ventilation ≥24-hours and/or re-intubation, (5) new requirement for renal replacement therapy (temporary or permanent), and (6) colonic ischemia requiring laparotomy with or without bowel resection. We excluded procedures that were missing operation type, age, sex, or postoperative mortality information (N = 37).

The analysis of the risks of individual complications and subsequent death following AAA repair was performed using the entire 8-nation cohort (Supplemental Figure 1, https://links.lww.com/SLA/D248). The FtR analysis was not possible for the aggregated 8 registry cohort due to variation in reported complications (not all registries reported data for all of the 6 complications). Subsequently, data from 4 countries (Australia, Hungary, New Zealand, and USA), since each provided information for all 6 major complications following AAA repair, were pooled for the analysis to determine the association between center volume and FtR.

The main exposure variable was hospital volume. The average annual volumes for OAR and EVAR (both elective and nonelective) were determined and categorized into quartiles, as in a previous analysis of this data assessing the volume-outcome relationship for perioperative mortality.1 Mortality was defined as death within the same hospital admission (in-hospital mortality). The primary endpoint was FtR, defined as death during the same hospital admission after experiencing at least 1 of the aforementioned 6 major complications.

Statistical Analysis

Overall cohort characteristics were examined by registry, with categorical variables reported as counts (and percentage) and continuous variables reported as means (and standard deviation). In the aggregated analysis of the 8 participating registries, we examined the risks of individual complications and subsequent death following EVAR and OAR, respectively. For each complication, an unadjusted analysis was performed by including registries that reported this complication. We excluded subjects with missing values for each complication (0.4%–4.6% missing) (Supplemental Table 1, https://links.lww.com/SLA/D249).

Restricting the analysis to the 4 registries that recorded all 6 complications, we examined the proportion of patients having at least 1 complication after EVAR and OAR, overall and by hospital volume quartiles. In the restricted cohort of patients with at least 1 complication for the FtR analysis, patient characteristics were examined by volume groups for EVAR and OAR, respectively. Differences in patient characteristics across volume groups were compared using 1-way analysis of variance tests for age and Chi-square tests for categorical variables.

The analysis of the association between volume and FtR was stratified by procedure type (EVAR or OAR). Unadjusted FtR rates were assessed and compared between hospital volume quartiles using Chi-square tests. The trend of decreasing FtR associated with increasing volume quartile was evaluated using a Cochran-Armitage trend test. We used multivariable logistic regression models to compare FtR across volume groups, adjusting for differences in characteristics of patients in these groups. A random intercept was used to account for clustering within registries. Covariates used for risk-adjusted analysis were age, sex, comorbidities, procedure year, and maximum AAA diameter. Comorbidities included diabetes, history of cardiac disease, and renal dysfunction, whose definitions were previously published (http://www.icvr-initiative.org/media/3954/ICVR%20Risk%20Factors.pdf).

Multiple imputation by fully conditional specification method was used to address arbitrary pattern missing data. Imputation was made for missing comorbidities and aneurysm size based on patient age, sex, procedure type, indication, procedure year, and registry. Twenty imputations were performed, and parameter estimates were combined from the imputations using Rubin rule.10 We performed sensitivity analyses among patients <75 years old and after excluding outlier centers with high complication but low FtR event rates. We also performed a sensitivity analysis examining the impact of centers’ use of EVAR on FtR by additionally adjusting for the proportion of EVAR procedures of all intact AAA repairs at each center.

The association between the number of complications patients experienced and FtR was assessed. The absolute number of complications per patient, and the relative proportion of FtR events stratified by the number of complications were evaluated. A multivariable logistic regression model was used to examine the association between the number of complications patients experienced and FtR when controlling for patient characteristics. Restricting to those who had only 1 complication, we further examined the association between the presence of each individual complication and FtR. The association of hospital volume and FtR was evaluated within each subgroup defined by the number of complications experienced (1, 2, and 3+).

All analyses were performed in SAS 9.3 (Cary, NC). A P value < 0.05 was considered statistically significant.

RESULTS

Study Cohort

A total of 60,273 unique intact AAA repair procedures were included in the study (Supplemental Figure 1, https://links.lww.com/SLA/D248) (EVAR = 43,668; OAR = 16,605). Among all patients, the mean age was 72.9(±8.6) years, a majority were male (81.8%), and the comorbidity frequency was typical for an AAA repair population (Table 1). The most frequently reported complications for both EVAR and OAR were cardiac (EVAR-3.0%, OAR-8.9%) and respiratory failure related events (EVAR-1.0%, OAR-5.7%) (Table 2). Postoperative colonic ischemia resulting in laparotomy was the complication associated with the highest incidence of death after both EVAR (43.6%) and OAR (43.4%).

TABLE 1 - Patient Characteristics by Registry
Australia (N = 11,933) Denmark (N = 3991) Finland (N = 622) Hungary (N = 1722) Malta (N = 122) New Zealand (N = 2065) Sweden (N = 6734) USA (N = 33,084)
Age (Mean/SD) 74.6 (8.2) 71.7 (7.2) 72.9 (9.4) 69.2 (8.7) 73.3 (7.4) 74.4 (7.9) 72.5 (7.4) 72.6 (8.9)
Sex
 Male 10,101 (84.6%) 3315 (83.1%) 546 (87.8%) 1467 (85.2%) 114 (93.4%) 1657 (80.2%) 5718 (84.9%) 26,354 (79.7%)
 Female 1832 (15.4%) 676 (16.9%) 76 (12.2%) 255 (14.8%) 8 (6.6%) 408 (19.8%) 1016 (15.1%) 6730 (20.3%)
Diabetes
 No 10,172 (85.2%) 3499 (87.7%) 483 (77.7%) 1409 (81.8%) 100 (82.0%) 1801 (87.2%) 5622 (83.5%) 26,541 (80.2%)
 Yes 1760 (14.7%) 451 (11.3%) 114 (18.3%) 313 (18.2%) 22 (18.0%) 264 (12.8%) 876 (13.0%) 6499 (19.6%)
 Missing 1 (0.0%) 41 (1.0%) 25 (4.0%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 236 (3.5%) 44 (0.1%)
Cardiac
 No 5502 (46.1%) 2490 (62.4%) 287 (46.1%) 870 (50.5%) 67 (54.9%) 1052 (50.9%) 3911 (58.1%) 18,117 (54.8%)
 Yes 6430 (53.9%) 1421 (35.6%) 306 (49.2%) 852 (49.5%) 55 (45.1%) 1013 (49.1%) 2535 (37.6%) 14,930 (45.1%)
 Missing 1 (0.0%) 80 (2.0%) 29 (4.7%) 0 (0.0%) 0 (0.0%) 0 (0.0%) 288 (4.3%) 37 (0.1%)
Renal disease
 No 10,939 (91.7%) 3254 (81.5%) 478 (76.8%) 0 (0.0%) 110 (90.2%) 1886 (91.3%) 4707 (69.9%) 30,231 (91.4%)
 Yes 993 (8.3%) 146 (3.7%) 26 (4.2%) 0 (0.0%) 12 (9.8%) 179 (8.7%) 234 (3.5%) 2301 (7.0%)
 Missing 1 (0.0%) 591 (14.8%) 118 (19.0%) 1722 (100.0%) 0 (0.0%) 0 (0.0%) 1793 (26.6%) 552 (1.7%)
Procedure year
 10–13 7007 (58.7%) 2231 (55.9%) 358 (57.6%) 914 (53.1%) 65 (53.3%) 1141 (55.3%) 3938 (58.5%) 12,369 (37.4%)
 14–16 4926 (41.3%) 1760 (44.1%) 264 (42.4%) 808 (46.9%) 57 (46.7%) 924 (44.7%) 2796 (41.5%) 20,715 (62.6%)
Approach
 Open 3144 (26.3%) 2516 (63.0%) 200 (32.2%) 1113 (64.6%) 38 (31.1%) 892 (43.2%) 2677 (39.8%) 6025 (18.2%)
 EVAR 8789 (73.7%) 1475 (37.0%) 422 (67.8%) 609 (35.4%) 84 (68.9%) 1173 (56.8%) 4057 (60.2%) 27,059 (81.8%)
Maximum aneurysm diameter (cm)
 Mean/SD 5.9 (1.2) 6.3 (1.4) 6.1 (1.4) 6.5 (1.3) 6.2 (1.2) 6.1 (1.3) 5.7 (1.3)
 Missing 218 (1.8%) 3991 (100.0%) 228 (36.7%) 16 (0.9%) 0 (0.0%) 1 (0.0%) 2 (0.0%) 512 (1.5%)
Included in the final analysis for failure to rescue risk adjusted comparisons.EVAR indicates endovascular aortic repair; SD, standard deviation.

TABLE 2 - Complications After Open and Endovascular Repair of Intact AAA Procedures and Subsequent Risk of Death
EVAR Open
Bleeding
 N total 43,247 16,479
 Event (%) 259 (0.6%) 361 (2.2%)
 Death (% of event) 37 (14.3%) 102 (28.3%)
Stroke
 N total 43,128 16,346
 Event (%) 96 (0.2%) 122 (0.7%)
 Death (% of event) 13 (13.5%) 32 (26.2%)
Cardiac event
 N total 43,443 16,464
 Event (%) 1301 (3.0%) 1469 (8.9%)
 Death (% of event) 118 (9.1%) 215 (14.6%)
Renal replacement
 N total 43,411 16,469
 Event (%) 260 (0.6%) 494 (3.0%)
 Death (% of event) 71 (27.3%) 128 (25.9%)
Respiratory failure
 N total 38,100 11,393
 Event (%) 363 (1.0%) 645 (5.7%)
 Death (% of event) 105 (28.9%) 147 (22.8%)
Ischemic bowel
 N total 39,544 13,873
 Event (%) 101 (0.3%) 274 (2.0%)
 Death (% of event) 44 (43.6%) 119 (43.4%)
AAA indicates abdominal aortic aneurysms.

Among the 4 registries included in the FtR analysis, 4.3% (n = 1628/37,630) of patients undergoing EVAR and 18.5% (n = 2072/11,174) of those undergoing OAR developed at least 1 complication. Among patients experiencing a complication, the FtR rate was 10.3% (n = 168/1628) for EVAR and 15.7% (n = 326/2072) for OAR, respectively (Table 3). The unadjusted rate of postoperative complications increased slightly with increasing volume of intact EVAR procedures (Q1 3.8%; Q4 5.4%; P < 0.001); for OAR, increasing volume was associated with lower rate of postoperative complications (Q1 19.6%; Q4 14.7%, P < 0.001) (Supplemental Figure 2, https://links.lww.com/SLA/D248). Complication rates after adjustment for patient characteristics were higher in Q4 centers after EVAR [odds ratio (OR) = 1.18, 95% confidence interval (CI) 1.01–1.38; P = 0.03), but not after OAR (OR = 0.93, 95% CI 0.75–1.14; P = 0.48) (Supplemental Figure 3, https://links.lww.com/SLA/D248).

TABLE 3 - Failure to Rescue Cohort (Data From Australia, Hungary, New Zealand, and USA)
Number Percent
EVAR N = 37,630
 Complications, died 168 0.4
 Complications, alive 1460 3.9
 No complications, died 70 0.2
 No complications, alive 35,264 93.7
 Missing data 668 1.8
Open N = 11,174
 Complications, died 326 2.9
 Complications, alive 1746 15.6
 No complications, died 76 0.7
 No complications, alive 8868 79.4
 Missing data 158 1.4
EVAR indicates endovascular aortic repair.

Characteristics of the subset of patients in the volume-FtR analysis are depicted in Supplemental Table 2, https://links.lww.com/SLA/D249. There were differences in the characteristics of patients treated at hospitals of the different volume quartiles. Specifically, the highest volume EVAR centers (Q4) repaired greater proportions of female patients (25.2% vs Q1-21.1%, P = 0.26), but fewer subjects with pre-existing renal disease (Q4-14.1% vs Q1-17.3%, P = 0.05). Analogous to elective EVAR, higher volume centers selected OAR patients with renal disease less frequently (Q4-9.0% vs Q1-Q2–10.1%–12.3%, P < 0.001) but no other differences were detected.

FtR Association With Hospital Volume

In the unadjusted comparison, there was a significant trend towards decreased FtR rate with increasing hospital volume for both EVAR and OAR (Fig. 1). Q1 centers (lowest volume quartile) had the highest FtR rate while Q4 centers (highest volume quartile) exhibited the lowest FtR rate (EVAR: 12.7% vs 7.6%; P-trend = 0.03; OAR: 19.7% vs 7.6%; P-trend < 0.001). After adjusting for patient characteristics, the odds of FtR after EVAR in Q4 centers was 46% lower (OR = 0.54 95% CI 0.34–0.87; P = 0.01) than Q1 centers (Fig. 2) (Supplemental Table 3, https://links.lww.com/SLA/D249). For OAR, the odds of FtR in Q4 centers was 80% lower (OR = 0.22, 95% CI 0.11–0.44; P < 0.001) compared to Q1 centers. The association between FtR and hospital volume was confirmed for OAR in sensitivity analyses among patients <75 years old and after excluding outlier centers with high complication but low FtR event rates (Supplemental Table 4, https://links.lww.com/SLA/D249). Notably, no association between the relative proportion of EVAR utilization that accounted for the proportion of intact AAA procedures that a center performed annually and FtR was present (Supplemental Table 5, https://links.lww.com/SLA/D249).

FIGURE 1
FIGURE 1:
Unadjusted failure to rescue rates by hospital volume quartiles.
FIGURE 2
FIGURE 2:
Adjusted odds ratio of failure to rescue by hospital volume quartiles.

FtR Association With Number and Type of Complications

Among patients who experienced a major complication, most subjects experienced 1 (EVAR: 85.1%; OAR: 75.1%) or 2 (EVAR: 11.0%; OAR 17.1%) complications. The predicted adjusted FtR rate increased with the number of complications for both EVAR (from 6% after 1 complication to 70% after ≥3 complications; P < 0.001) and OAR (from 13% after 1 complication to 69% after ≥3 complications; P < 0.001) (Fig. 3A and Table 4). Analyses stratifying by the number of complications in EVAR patients did not show differences in the association between hospital volume and FtR across volume quartiles. For OAR, the association between increasing center volume and reduced FtR rate was evident in patients experiencing a single complication (Q4 vs Q1-Q3: OR 0.65, 95% CI 0.45–0.92; P = 0.02) but not in patients experiencing 3 or more complications (Q4 vs Q1-Q3: OR 0.84, 95% CI 0.42–1.69; P = 0.62).

FIGURE 3
FIGURE 3:
Adjusted predicted risk of failure to rescue by (A) number of complications and (B) type of complication.
TABLE 4 - Association Between Number of Complications and Subsequent Risk of Failure to Rescue
N Patients Adjusted Predicted Risk of FtR (95% CI) OR (95% CI) P Value
EVAR
 #Complications
  1 1385 6% (4%–10%) Ref
  2 179 31% (20%–45%) 6.82 (4.51–10.30) <0.001
  3–5 64 70% (53%–83%) 35.40 (19.35–64.76) <0.001
Open
 #Complications
  1 1556 13% (8%–21%) Ref
  2 354 31% (20%–46%) 3.04 (2.22–4.16) <0.001
  3–5 162 68% (52%–81%) 14.35 (9.79–21.04) <0.001
CI indicates confidence interval; FtR, failure to rescue; OR, odds ratio.

Finally, when examining the odds of subsequent FtR based upon the occurrence of specific types of complications, the hierarchy of complications associated with the greatest subsequent risk of FtR for both EVAR and OAR were notably different. Although colonic ischemia had the highest risk of FtR for both procedures (adjusted predicted risks, EVAR: 27%, 95% CI 14%–45%; OAR: 30%, 95% CI 17%–46%), bleeding requiring a return to the operating room was associated with an 8% risk of FtR after EVAR compared to 25% with OAR. However, respiratory failure and new in-hospital dialysis requirement had analogous rates of FtR for both procedures (∼8%–11%) (Fig. 3B, Supplemental Table 6, https://links.lww.com/SLA/D249).

DISCUSSION

Previous research in different health care systems and surgical specialties including aortic surgery has suggested that presence of preventable postoperative deaths should be a main focus of approaches to quality improvement using FtR as the metric of choice, and the FtR relationship with annual center volume.11–13 This analysis examined the prevalence of complications and their relationship to FtR after both OAR and EVAR of intact AAA using a large multi-national, real-world dataset, which provides multiple unique contributions. This study confirms that highest volume hospitals achieved the lowest FtR rates after elective OAR and EVAR, with the association remaining significant in risk-adjusted analysis. Not surprisingly, morbidity after OAR was more than 3-fold compared to EVAR; however, for both procedures, the occurrence of a major complication was associated with comparable rates of subsequent perioperative mortality (11%–15%), representing FtR events. Finally, FtR risk was associated with both increasing number and specific types of complications.

The detection of a FtR volume-outcome relationship after both elective EVAR and OAR within an aggregated cohort of AAA patients derived from different health care systems within Europe, Australasia, and North America underscores the relevance of these findings internationally. Despite differences among health care delivery models, ethnicities, and socioeconomic parameters, FtR remained significantly associated with specific complications and center volume in adjusted analysis. The results can thus be reasonably regarded as fairly generalizable across health care systems. They also indicate the need for centralization of aortic care to high-volume facilities to obtain the best outcomes for patients, as many aortic procedures may continue to be performed in low-volume institutions.2,14 This might be crucial to help avoid preventable deaths following major operations in vascular patients.

The volume-outcome relationship after AAA surgery is well documented,15–19 although it remains a focus of ongoing controversy and active research.20–22 Because repair of intact AAA is a prophylactic intervention, it is crucial to determine the optimal organization of services that delivers the safest care possible. Importantly, low rates of postoperative mortality following EVAR have led to concerns that all-cause perioperative mortality may be an inadequate quality metric for the evaluation of surgical outcomes. In a previous ICVR study, postoperative mortality improved after OAR when performed at high-volume centers, but no relationship was detected after EVAR, with an annual center volume of 13–16 OAR procedures/year detected as the threshold associated with the greatest risk reduction in postoperative mortality after intact AAA repair.7 The low perioperative mortality rate after EVAR might question the value of this metric as a quality indicator for endovascular procedures. Despite the overall low rate of in-hospital mortality after elective AAA repair, FtR events constitute an important proportion of preventable postoperative deaths for both EVAR and OAR. Notably, given the low rates of periprocedural mortality, FtR could become an important quality metric for short-term outcome evaluation after EVAR. Therefore, these results would further support the concept that FtR might represent a better parameter for discriminating inter-hospital quality of care rather than all-cause postoperative mortality.

Although the current study cannot directly elucidate the cause for why higher procedural volume leads to lower FtR rates, some inferences can be reasonably made. Our results seem to indicate that the primary driver of the inverse association between FtR and volume might not necessarily be the ability to prevent complications, but rather the preparedness to promptly recognize these events and treat them appropriately once detected. In that sense, a potential advantage for using FtR as a quality indicator is its reflection of the team's ability to prevent, detect and/or treat severe complications and avoid subsequent death, making it more dependent on potentially modifiable processes of care rather than patient factors, which are frequently nonmodifiable. Concordantly, efforts should focus not only on developing strategies that prevent complications but also on the early recognition and expedited management of complications when they do occur, thereby improving the ability to rescue patients. The definition of FtR in this study only included patients with recorded complications before death. Of note, some patients undergoing aortic surgery may suffer from sudden death without prior complication, and prevention of such events may depend on both patient selection and peri- and postoperative management.

Our findings emphasize the need for diligent observation of some patient groups even after minimally invasive and seemingly “safer” endovascular treatment. Indeed, EVAR patients are usually elderly with multiple comorbidities, so they represent a uniquely vulnerable population at high risk for in-hospital complications which, in turn, have a significant impact on survival.23 This delineates FtR as a more common and perhaps more important outcome in this population as compared with all-cause mortality.24 When a major in-hospital complication occurs, subsequent mortality risk is elevated irrespective of the treatment modality, and should prompt selective implementation of appropriate care pathways, especially in high-risk subjects. Importantly, the proportion of EVAR procedures that participating institutions used for their total AAA care delivery did not seem to affect the volume-outcome relationship with FtR. The clinical relevance of these finidings would further strengthen current practice guidelines from major vascular societies in USA and Europe as they consider OAR and EVAR together for the identification of volume thresholds for optimal delivery of AAA care.

According to previous studies, FtR drives a large proportion of the variation in mortality rates observed in conjunction with AAA repair.25 This is thought to be predominantly related to hospital attributes that affect timely recognition and prompt management of major complications rather than the operative approach or the patient's comorbid status.26 Several studies have suggested that advanced technology (eg, machine-learning, integrated clinical pathways, and early warning triggers), hospital teaching status, nurse-patient ratio, and increased volume were associated with lower FtR rates across institutions.15,27,28 Although the current analysis is unable to adjust for all these variables, we demonstrated a strong and independent relationship between increasing hospital volume and improved FtR rates.

Despite notable features of FtR as an outcome metric, controversy also exists surrounding its ability to function as a quality indicator. Some stakeholders have argued that if similar volume thresholds for both FtR and all-cause mortality are identified, it would not translate into added benefits for identifying top-performing hospitals. For example, in a national audit from the Netherlands, Lijftogt and colleagues demonstrated that although there was more variation in FtR than all-cause mortality between hospitals, FtR did not identify centers with significantly higher mortality.29

In contrast, a reciprocal decrease in FtR with increasing hospital volume was identified for both EVAR and OAR in this analysis. In fact, a recent analysis of Medicare claims strongly suggests the contrary. Specifically, in an analysis of over 700,000 patients treated within 3,400 hospitals that examined outcomes among AAA, pulmonary resection, colectomy, and pancreatectomy patients from 2005 to 2014, hospitals with the largest reductions in postoperative mortality achieved these improvements primarily through lowering FtR rates and not by reducing serious complications rates.30 Therefore, we would advocate for standardization of incorporating FtR data in outcome reporting to provide enhanced discriminatory ability to identify centers with optimal team performance surrounding major operations.

In addition to being dependent on team performance, the number and severity of complications influence FtR rates and the occurrence of some complications might be inevitable. In fact, this analysis identified that once a patient suffers ≥3 complications, the protective effect of high-volume centers to prevent FtR might be abrogated. This underscores the need to not only prevent major morbidity following invasive procedures but to disrupt the cascade of multiple complications in high-risk patients. Interestingly, Tevis and associates31 examined >470,000 general surgery patients from the ACS-NSQIP database and found that almost half (40%) with complications suffered multiple adverse events. Not surprisingly, these patients were more frail and had greater number of preoperative comorbidities, thereby representing a high-risk subgroup of subjects that could be targeted for prevention and surveillance.

As shown in previous research, colonic ischemia after intact AAA repair, despite being rare, was associated with the highest odds for FtR after both EVAR and OAR.32 However, it is also well known that colonic ischemia may occur relatively unrecognized especially during the early phase (due to lack of specific symptoms), and its rarity might make the threshold for its diagnosis quite high by many physicians. Therefore, it is plausible that specific algorithms tailored to ensure prompt recognition and aggressive management of specific high-impact complications can further improve delivery of AAA care.

Study Limitations

Our study has important limitations and the findings should be interpreted within this context. First, it is not a randomized trial and, although this represents the largest multi-national dataset that has been used to date to assess FtR after EVAR or OAR of intact AAA, the retrospective study design cannot fully account for any selection bias that may exist between high and low volume centers. In addition, other hospital attributes or surgeon-specific variables known to predict mortality were not available for analysis. Similarly, intraoperative details were not available to further stratify EVAR or OAR procedures based on their anatomic complexity, which certainly could have impacted the study results. However, to the extent possible, we have used robust multivariable analyses to account for known residual confounders. Despite the large sample, some analyses were likely underpowered (eg, assessment of volume of complications versus FtR rate), thereby resulting in possible type II statistical error.

Importantly, no predetermined definition of FtR was available and data reporting was heterogeneous across included registries, so our FtR definition may not be able to capture the entire spectrum of possible complications requiring patient rescue. Moreover, specific complications can present on a wide clinical spectrum (eg, colonic ischemia events without need for laparotomy or respiratory failure resulting in tracheostomy and/or admission to a long-term acute care facility), so the current analysis does not have the sensitivity to describe the larger denominator of patients experiencing various permutations of the major complications analyzed and how these events could associate with FtR. Finally, it is well known that all-cause mortality within 6 months of surgery is significantly lower after EVAR, but with longer follow-up time, the pooled hazard estimate favors OAR.33 Therefore, improved outcomes after EVAR performed in high-volume institutions may extend well beyond the immediate postoperative period and be associated with improved long-term treatment durability.34,35 However, this remains a matter to be elucidated in future studies.

CONCLUSIONS

In this multi-national dataset, FtR rate after intact AAA repair with EVAR and OAR is significantly associated with hospital volume. Hospitals in the top volume quartiles achieve the lowest mortality after a complication has occurred. FtR is a common cause of mortality after EVAR and OAR, and can serve as a quality indicator for intact AAA repair. These results can inform policymakers and serve to improve AAA care delivery worldwide.

REFERENCES

1. Scali ST, Beck AW, Sedrakyan A, et al. Hospital volume association with abdominal aortic aneurysm repair mortality: analysis of the international consortium of vascular registries. Circulation 2019; 140:1285–1287.
2. Trenner M, Kuehnl A, Salvermoser M, et al. Editor's Choice - high annual hospital volume is associated with decreased in-hospital mortality and complication rates following treatment of abdominal aortic aneurysms: secondary data analysis of the nationwide German DRG statistics from 2005 to 2013. Eur J Vasc Endovasc Surg 2018; 5:185–194.
3. Budtz-Lilly J, Björck M, Venermo M, et al. Editor's Choice - the impact of centralisation and endovascular aneurysm repair on treatment of ruptured abdominal aortic aneurysms based on international registries. Eur J Vasc Endovasc Surg 2018; 56:181–188.
4. Chaikof EL, Dalman RL, Eskandari MK, et al. The Society for Vascular Surgery practice guidelines on the care of patients with an abdominal aortic aneurysm. J Vasc Surg 2018; 67:2–77.
5. Wanhainen A, Verzini F, Van Herzeele I, et al. Editor's Choice – European Society for Vascular Surgery (ESVS) 2019 clinical practice guidelines on the management of abdominal aorto-iliac artery aneurysms. Eur J Vasc Endovasc Surg 2019; 57:8–93.
6. Beck AW, Sedrakyan A, Mao J, et al. Variations in abdominal aortic aneurysm care: a report from the international consortium of vascular registries. Circulation 2016; 134:1948–1958.
7. Scali S, Beck AW, Sedrakyan A, et al. Optimal threshold for the volume-outcome relationship after open AAA repair in the endovascular era – analysis of the international consortium of vascular registries. Eur J Vasc Endovasc Surg 2021; 61:747–755.
8. Dakour-Aridi H, Paracha NZ, Locham S, et al. Assessment of failure to rescue after abdominal aortic aneurysm repair using the national surgical quality improvement program procedure-targeted data set. J Vasc Surg 2018; 68:1335–1444.
9. Behrendt CA, Venermo M, Cronenwett JL, et al. VASCUNET, VQI and the International Consortium of Vascular Registries – unique collaborations for quality improvement in vascular surgery. Eur J Vasc Endovasc Surg 2019; 58:1–2.
10. Rubin DB. Multiple Imputation for Nonresponse in Surveys. Bridgewater, NJ: John Wiley & Sons; 2004.
11. Scali ST, Giles KA, Kubilis P, et al. Impact of hospital volume on patient safety indicators and failure to rescue following open aortic aneurysm repair. J Vasc Surg 2020; 71:1135–1146.
12. Giles KA, Stone DH, Beck AW, et al. Association of hospital volume with patient selection, risk of complications, and mortality from failure to rescue after open abdominal aortic aneurysm repair. J Vasc Surg 2020; 72:1681–1690.
13. Ahmad T, Bouwman RA, Grigora I, et al. On behalf of the International Surgical Outcomes Study (ISOS) Group. Use of failure-to-rescue to identify international variation in postoperative care in low-, middle- and high-income countries: a 7-day cohort study of elective surgery. Br J Anest 2017; 119:258–266.
14. Mao J, Goodney P, Cronenwett J, et al. Association of very low-volume practice with vascular surgery outcomes in New York. JAMA Surg 2017; 152:759–766.
15. Landon BE, O’Malley AJ, Giles K, et al. Volume-outcome relationships and abdominal aortic aneurysm repair. Circulation 2010; 122:1290–1297.
16. Holt PJ, Poloniecki JD, Gerrard D, et al. Meta-analysis and systematic review of the relationship between volume and outcome in abdominal aortic aneurysm surgery. Br J Surg 2007; 94:395–403.
17. Wanhainen A, Haulon S, Kohl P. Centralization of abdominal aortic aneurysm repair – we can no longer ignore the benefits. Eur J Vasc Endovasc Surg 2020; 60:500–501.
18. Tripodi P, Mestres G, Riambau V. The Vascular Advisory Committee – Catalan Health Service. Impact of centralisation on abdominal aortic aneurysm repair outcomes: early experience in Catalonia. Eur J Vasc Endovasc Surg 2020; 60:531–538.
19. Phillips P, Poku E, Essat M, et al. Procedure volume and the association with short-term mortality following abdominal aortic aneurysm repair in European populations: a systematic review. Eur J Vasc Endovasc Surg 2017; 53:77–88.
20. Sawang M, Paravastu SCV, Liu Z, et al. The relationship between aortic aneurysm surgery volume and peri-operative mortality in Australia. Eur J Vasc Endovasc Surg 2019; 57:510–519.
21. Zettervall SL, Schermerhorn ML, Soden PA, et al. The effect of surgeon and hospital volume on mortality following open and endovascular repair of abdominal aortic aneurysms. J Vasc Surg 2017; 65:626–634.
22. Gray WK, Day J, Horrocks M. Volume-outcome relationship in elective abdominal aortic aneurysm surgery: analysis of the UK hospital episodes statistics database for the getting it right first time (GIRFT) programme. Eur J Vasc Endovasc Surg 2020; 60:509–517.
23. Varkevisser RRB, O’Donnell TFX, Swerdlow NJ, et al. Factors associated with in-hospital complications and long-term implications of these complications in elderly patients undergoing endovascular aneurysm repair. J Vasc Surg 2020; 71:1470–1480.
24. Hicks CW, O’Kelly A, Obeid T, et al. Predicting failure to rescue after abdominal aortic aneurysm repair in elderly patients. J Surg Res 2017; 217:265–270.
25. Sinha S, Ata Ozdemir B, Khalid U, et al. Failure-to-rescue and interprovider comparisons after elective abdominal aortic aneurysm repair. Br J Surg 2014; 101:1541–1550.
26. Gonzalez AA, Dimick JB, Birkmeyer JD, et al. Understanding the volume-outcome effect in cardiovascular surgery: the role of failure to rescue. JAMA Surg 2014; 149:119–123.
27. Hicks CW, Wick EC, Canner JK, et al. Hospital-level factors associated with mortality after endovascular and open abdominal aortic aneurysm repair. JAMA Surg 2015; 150:632–636.
28. Reddy HG, Shih T, Englesbe MJ, et al. Analyzing “failure to rescue”: is this an opportunity for outcome improvement in cardiac surgery? Ann Thorac Surg 2013; 95:1976–1981.
29. Lijftogt N, Karthaus EG, Vahl A, et al. Failure to rescue – a closer look at mortality rates has no added value for hospital comparisons but is useful for team quality assessment in abdominal aortic aneurysm surgery in the Netherlands. Eur J Vasc Endovasc Surg 2018; 56:652–661.
30. Fry BT, Smith ME, Thumma JR, et al. Ten-year trends in surgical mortality, complications and failure to rescue in medicare beneficiaries. Ann Surg 2020; 271:855–871.
31. Tevis SE, Cobian AG, Truong HP, et al. Implications of multiple complications on the postoperative recovery of general surgery patients. Ann Surg 2016; 263:1213–1218.
32. Behrendt CA, Rieß HC, Schwaneberg T, et al. Incidence, predictors and outcomes of colonic ischaemia in abdominal aortic aneurysm repairs. Eur J Vasc Endovasc Surg 2018; 56:507–513.
33. Antoniou GA, Antoniou SA, Torella F. Editor's Choice – endovascular vs open repair for abdominal aortic aneurysm: systematic review and meta-analysis of updated peri-operative and long-term of randomized controlled trials. Eur J Vasc Endovasc Surg 2020; 59:385–397.
34. Lepidi S, D’Oria M. Long-term benefits of EVAR in the modern era: the importance of looking at durability of stent-grafts without forgetting the right pathophysiology. Eur J Vasc Endovasc Surg 2020; 60:956–957.
35. Boyle JR. Is in hospital mortality following EVAR still a valid outcome measure? Eur J Vasc Endovasc Surg 2019; 57:520.
Keywords:

abdominal aortic aneurysm; elective; endovascular aortic repair; failure to rescue; hospital volume; mortality; open aortic repair; postoperative care; quality improvement

Supplemental Digital Content

Copyright © 2021 Wolters Kluwer Health, Inc. All rights reserved.