Prediction of 30-day mortality after endovascular repair or open surgery in patients with ruptured abdominal aortic aneurysms

Article Outline

• Abstract
• Methods
  • Patient population
  • Protocol
  • Data collection and definitions
  • Glasgow Aneurysm Score
  • Data and statistical analyses
• Results
  • Patient population
  • Outcomes
  • Validation of GAS
  • Interaction
  • Updated prediction rule
• Discussion
• Conclusion
• Author contributions
• Acknowledgment
• References
• Copyright

Objective

To validate the Glasgow Aneurysm Score (GAS) in patients with ruptured abdominal aortic aneurysms (AAAs) treated with endovascular repair or open surgery and to update the GAS so that it predicts 30-day mortality for patients with ruptured AAA treated with endovascular repair or open surgery.

Methods

In a multicenter prospective observational study, 233 consecutive patients with ruptured AAAs were evaluated; 32 patients did not survive to repair and statistical analysis was performed using collected data on 201 patients. All patients who were treated with endovascular repair (n = 58) or open surgery (n = 143) were included. The GAS was calculated for each patient. The area under the receiver operating characteristics curve (AUC) was used to indicate discriminative ability. We tested for interactions between risk factors and the procedure performed. The GAS was updated to predict 30-day mortality after endovascular repair or open surgery in patients with ruptured AAAs using logistic regression analysis.

Results

Thirty-day mortality was 15/58 (26%) for patients treated with endovascular repair and 57/143 (40%) for patients treated with open surgery (P = .06). The AUC for GAS was 0.69. No relevant interactions were found. The updated prediction rule (AUC = 0.70) can be calculated with the following formula: + 7 for open surgery + age in years + 17 for shock + 7 for myocardial disease + 10 for cerebrovascular disease + 14 for renal insufficiency.

Conclusion

We showed limited discriminative ability of the GAS and therefore updated the GAS by adding the type of procedure performed. This updated prediction rule predicts 30-day mortality for patients with ruptured AAAs treated with endovascular repair or open surgery.

The traditional approach to treat ruptured abdominal aortic aneurysms (AAAs) is open surgery. The Glasgow Aneurysm Score (GAS) is used to predict in-hospital mortality after open surgery for patients with ruptured or unruptured AAA.¹ Several studies have validated this prediction rule in patients with ruptured AAA treated with open surgery. Two validations reported good validity², ³ and two reported poor validity.⁴, ⁵

Since 1994, endovascular repair for ruptured AAAs has been proven to be feasible⁶ and is increasingly being adopted as the treatment of choice.⁷ Several studies showed a reduction in mortality and morbidity rates after endovascular repair compared to rates for open surgery in patients with ruptured AAAs;⁸, ⁹, ¹⁰, ¹¹ however, in other studies this reduction could not be confirmed.¹², ¹³, ¹⁴ Recently, it was suggested that patients at higher risk for periprocedural cardiac complications may benefit more from endovascular repair than from open surgery.¹⁴

Due to the rise of endovascular repair, the patient population receiving open surgery has shifted in recent years and the GAS may no longer be valid in this population. Additionally, the GAS cannot be used to decide whether a patient with a ruptured AAA may benefit more from endovascular repair or from open surgery, as it does not predict outcomes for endovascular repair patients. Several studies investigated whether the GAS was a predictor of outcomes in patients with elective endovascular AAA repair and their results were contradictive.¹⁵, ¹⁶, ¹⁷, ¹⁸ One study attempted to define cut-off values of the GAS to be able to determine if patients with elective AAA should be treated with endovascular or open repair.¹⁹ Furthermore, one study found that the GAS was of limited value in patients with ruptured AAAs treated with open surgery.⁵ Whether the current GAS prediction rule is still valid in predicting 30-day mortality after open surgery and whether it can predict 30-day mortality after endovascular repair needs to be determined. Ideally, the GAS should be modified to identify patients who would be better suited for endovascular repair vs open surgery.

The purpose of our study was to validate the GAS in patients with ruptured AAAs who were treated with endovascular repair or open surgery. In addition, we aimed to update the GAS for prediction of 30-day mortality after endovascular repair or open surgery.

Back to Article Outline

Methods 

Patient population 

In a prospective multicenter observational study, data were collected on 233 consecutive patients between December 22, 2004, and October 31, 2006, in seven institutions in The Netherlands: Atrium Medical Center, Heerlen (45 patients), Catharina Hospital, Eindhoven (25 patients), Erasmus Medical Center, Rotterdam (40 patients), Medical Spectrum Twente, Enschede (24 patients), Medical Center Rotterdam Zuid, Rotterdam (30 patients), University Medical Center, Groningen (37 patients), and University Medical Center, Nijmegen (32 patients). Patients were included if they presented with ruptured AAAs and were treated with endovascular repair or open surgery (n = 201). A total of 32 of 233 patients (14%) were excluded because they died before AAA repair could be initiated; death was caused by severe comorbidity or the patient refused treatment (Fig 1). Rupture of the AAA was confirmed on computed tomography (CT) scan prior to the procedure, or by free blood noted during laparotomy. The Institutional Review Board approved this study and waived the obligation to obtain informed consent due to the acute nature of the clinical problem and the observational nature of this study.²⁰

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Patient flowchart.

Protocol 

In all participating hospitals, endovascular repair was the preferred treatment in patients with ruptured AAA. Except for in one hospital, the vascular surgeon and/or radiologist who performed endovascular repair of ruptured AAAs were available 24 hours a day, 7 days a week. Upon arrival in the hospital, patients who were in a hemodynamically stable condition underwent an abdominal CT-scan to confirm rupture and to assess whether the AAA was anatomically suited for endovascular repair. In 4 patients, an additional aortic angiography was performed to assess anatomic suitability for endovascular repair. If patients were anatomically suitable for endovascular repair, they were treated with this procedure. Hemodynamically unstable patients (n = 37) were immediately transported to the operating room for open surgery. The definition of “hemodynamically stable” was determined upon arrival at the hospital and varied across the participating hospitals. In two hospitals, the attending vascular surgeon made the judgment without explicit criteria. In two other hospitals, the attending vascular surgeon or radiologist considered the patient hemodynamically stable if s/he gave an adequate verbal reply. In one hospital, the attending vascular surgeon defined hemodynamically stable as a systolic blood pressure of 60 millimeter Mercury or higher, whereas in two other hospitals a systolic blood pressure cutoff of 70 millimeter Mercury was used. After CT-scan confirmed the presence of a ruptured AAA, the patient was immediately transported to the operating room where endovascular repair was performed if the AAA was anatomically suitable; otherwise, open surgery was performed. The anatomic inclusion criteria for endovascular repair differed between the participating hospitals since they stocked different endovascular devices. The criteria varied between 7 and 15 millimeters for the proximal neck length, between 30° and 90° for the neck angulation, and between 28 and 32 millimeters for the neck diameter. The endografts used were Talent aortouniiliac (AUI) stent-grafts (Medtronic, Santa Rosa, Calif), Anaconda bifurcated endografts (Vascutek, Renfrewshire, Scotland), Cook endografts (Zenith, Bloomington, Ind), and Excluder endografts (Gore, Flagstaff, Ariz).

Data collection and definitions 

Prospectively collected data included: patient characteristics (ie, age, gender, renal insufficiency, and history of diabetes mellitus, hypertension, angina pectoris, myocardial infarction, congestive heart failure, or cerebrovascular disease), use of medication prior to hospital admission, patients' hemodynamic condition upon presentation to the hospital, shock upon presentation to the hospital, use of CT-scan or angiography prior to the procedure, morphology of the AAA (infrarenal, juxtarenal, or suprarenal), time to repair, and which treatment was performed (endovascular repair or open surgery). Diabetes mellitus was defined as receiving oral medication and/or insulin therapy for diabetes mellitus. Hypertension included patients with receipt of at least one antihypertensive drug. Congestive heart failure included symptoms of congestive heart failure and receipt of medication for this diagnosis. Shock was defined as a systolic blood pressure less than 80 millimeter Mercury based on the lowest blood pressure recorded. Myocardial disease comprised previous myocardial infarction and/or angina pectoris. Cerebrovascular disease included all previous cerebrovascular accidents and transient ischemic attacks. Renal insufficiency referred to a preoperative creatinine value of more than 160 micromols per liter (ie, 1.8 milligrams per deciliter). Data on AAA neck length, neck diameter, and neck angulation was not collected. A standardized form was used to register these data. In order to obtain information about 30-day mortality and the causes of death, medical records, and the computerized database of the participating hospitals were used. The data was prospectively gathered and retrospectively analyzed.

Glasgow Aneurysm Score 

The GAS was originally based on 235 patients treated for AAA between January 1980 and December 1989 at four hospitals in Glasgow, United Kingdom.¹ The GAS was calculated using the following formula: GAS = age in years + 17 for shock + 7 for myocardial disease + 10 for cerebrovascular disease + 14 for renal insufficiency.¹ Patients with a GAS less than 70 are considered to have a low risk of mortality after open surgery for AAA, whereas patients with a GAS more than 85 are considered to have a high risk of mortality after treatment for AAA. In the original paper, ‘shock' was based on clinical information of tachycardia, hypotension, pallor, and sweating. Myocardial disease was defined as previous myocardial infarction and/or angina pectoris. Cerebrovascular disease comprised all grades of stroke including transient ischemic attacks. Renal insufficiency included chronic and acute renal failure.¹

Data and statistical analyses 

Patient data were entered into a database and checked by one of the authors for completeness (J. J. V.). Missing data regarding continuous variables (ie, age and systolic blood pressure) were assumed to be missing at random and entered based on the variable means since the number of missing values is low.²¹ If data regarding patient's medical history or medication were missing, it was assumed that the risk factor was not present or the medication was not used. In total, the proportion of missing data was less then 2% before missing data was entered. Analyses were performed according to the intention-to-treat principle; therefore, intraoperative deaths were also included for analysis in the respective treatment group.

We validated the GAS using receiver operating characteristics (ROC) curves to determine discriminative ability (ie, whether the GAS was higher in patients who died). An area under the ROC curve (AUC) of 0.50 indicates no discriminative ability and the closer the AUC is to 1.0, the better the discriminative ability.

In addition, we tested for interactions between risk factors (as determined in the GAS model) and the specific procedure performed in predicting 30-day mortality using logistic regression analysis. Interaction terms were considered potentially relevant if P < .20.

Based on a previously published approach, the GAS was updated to predict 30-day mortality after either endovascular repair or open surgery.²² In the first step, we estimated new regression coefficients for the GAS variables based on the study data. In the second step, we added the procedure performed (open surgery vs endovascular repair) to the original GAS variables.²² Regression coefficients for the new variable (ie, the procedure) and an intercept term were estimated; the GAS was then multiplied by a calibration slope β_GAS for overall adjustment of the original GAS regression coefficients. The formula we used was (30-day mortality)_{study data} = α + β_procedure * (procedure) + β_GAS * GAS. To calculate the adjusted GAS odds ratios, we used the formula: (adjusted GAS odds ratio) = (original GAS odds ratio) * exp (β_GAS). In the third step, we performed a multivariable logistic regression analysis on 30-day mortality, including the procedure performed (open surgery vs endovascular repair) and all individual GAS variables (ie, age, shock, myocardial disease, cerebrovascular disease, and renal insufficiency), and we estimated new regression coefficients for each variable.²² For each step, the AUC was estimated as a measure of discriminative ability and adjusted for optimism by bootstrapping. We used 200 bootstrap samples drawn with replacement from the original dataset. Bootstrapping means random drawing from the original data with replacement. This is a method to assess statistical accuracy. This validation procedure indicates the performance that may be expected in new, but similar patients.²³ The AUCs from the first, second, and third step were compared, and the prediction rule with the highest AUC is presented as the updated prediction rule.

Analyses were performed using SPSS for Windows Version 11.0.1 (SPSS Inc., Chicago, Ill) and S-Plus Version 6.0 (Insightful Corporation, Seattle, Wash).

Back to Article Outline

Results 

Patient population 

Patient demographics, characteristics, and use of medication before admission are listed in Table I. The proportion of males was somewhat higher in patients treated with endovascular repair compared with patients treated with open surgery (93% vs 83%, P = .05). In addition, the use of statins was somewhat higher in patients treated with endovascular repair compared with patients treated with open surgery (31% vs 19%, P = .06). The other characteristics were similar between the treatment groups. Eighty-three of the 143 patients (58%) treated with open surgery had infrarenal AAAs, 49 (34%) had juxtarenal (ie, no infrarenal neck) AAAs, 4 (3%) had suprarenal AAAs, and in 7 patients (5%) the AAA anatomy was not reported.

Table I. Patient demographics and characteristics

	Endovascular repair	Open surgery
	n = 58	n = 143	P value
Male	54(93%)	118(83%)	.05
Mean age(SD)	73.2(8.6)	73.5(7.5)	.83
Renal insufficiency	8(14%)	16(11%)	.61
Diabetes mellitus	8(14%)	13(9%)	.32
Hypertension	29(50%)	60(42%)	.30
Angina pectoris	6(10%)	18(13%)	.66
Previous myocardial infarction	12(21%)	35(25%)	.57
Heart failure	6(10%)	14(10%)	.91
CVA/TIA	6(10%)	18(13%)	.66
COPD	15(26%)	30(21%)	.45
AAA known before admission	14(24%)	25(18%)	.28
Shock	4(7%)	40(28%)	.001
Medication
Beta-blocker	19(33%)	40(28%)	.50
Anticoagulants	11(19%)	17(12%)	.19
Antiplatelet agents	24(41%)	46(32%)	.21
Statins	18(31%)	27(19%)	.06

SD, Standard deviation; CVA, cerebrovascular accident; TIA, transient ischemic attacks; COPD, chronic obstructive pulmonary disease; AAA, abdominal aortic aneurysm.

Outcomes 

Thirty-day mortality was 15/58 (26%) for patients treated with endovascular repair and 57/143 (40%) for patients treated with open surgery (P = .06). Among patients who were treated with open surgery, 30-day mortality was 39/106 (37%) for those who were hemodynamically stable prior to the procedure and 18/37 (49%) for those who were hemodynamically unstable prior to the procedure (P = .20). Two of the 58 patients (3%) treated with endovascular repair died intraoperatively, while 21 of the 143 patients (15%) treated with open surgery died intraoperatively (P = .02). There was no statistically significant association between the time to repair and 30-day mortality. The causes of 30-day mortality are listed in Table II. The difference in 30-day mortality was approximately similar to the difference in intraoperative mortality. Nine of the 58 patients (16%) initially treated with endovascular repair were converted to open surgery. The reasons for conversion to open surgery were the endografts could not be positioned correctly in the infrarenal neck (n = 1), colon ischemia (n = 1), large hematoma with abdominal compartment (n = 4), and unknown (n = 3).

Table II. Causes of 30-day mortality

	Endovascular repair	Open surgery
	n=58	n=143
Intraoperative	2(3%)	21(15%)
Postoperative
Cardiovascular⁎	3(5%)	8(6%)
Pulmonary†	3(5%)	5(3%)
Renal failure	1(2%)	1(1%)
Sepsis	2(3%)	1(1%)
Shock‡	2(3%)	4(3%)
Coagulopathy	1(2%)	0(0%)
Multiorgan failure	1(2%)	4(3%)
Infection	0(0%)	1(1%)
No treatment due to patients' comorbidity§	0(0%)	3(2%)
Unknown	0(0%)	9(6%)
Total deaths	15(26%)	57(40%)

Validation of GAS 

The GAS was less than 70 in 42 patients; between 70 and 75 in 26 patients; between 76 and 85 in 58 patients; and more than 85 in 75 patients (Table III). The mean GAS among patients who survived 30 days after the initial procedure was 77 for patients treated with endovascular repair and 80 for patients treated with open surgery (P = .14). The mean GAS among patients who died within 30 days after the initial procedure was 87 for patients treated with endovascular repair and 88 for patients treated with open surgery (P = .81). The AUC for the GAS was 0.69 (95%-confidence interval 0.61-0.76).

Table III. GAS and the prediction of 30-day mortality

	Endovascular repair		Open surgery		Odds ratio⁎ (95%-CI)
Score	Number of patients	Mortality (%)	Number of patients	Mortality (%)
< 70	17	2(12%)	25	3(12%)	1.02(0.15-6.9)
70-75	8	2(25%)	18	7(39%)	1.91(0.30-12)
76-85	16	3(19%)	42	17(41%)	2.95(0.73-12)
> 85	17	8(47%)	58	30(52%)	1.21(0.41-3.6)
Total	58	15(26%)	143	57(40%)	1.90(0.97-3.7)

GAS, Glasgow Aneurysm Score; CI, confidence interval.

Interaction 

Testing for interaction between risk factors and the type of procedure performed on 30-day mortality showed no relevant interactions (all P values > .20). This means that no risk factors were associated with a higher mortality for a specific procedure. Instead, endovascular repair always favored open surgery with respect to 30-day mortality given the set of risk factors included in the GAS model.

Updated prediction rule 

In the first step, we estimated new regression coefficients for the GAS variables based on the study data. The AUC adjusted for optimism was 0.67 (Table IV). In the second step, we added one new variable to the original GAS variables: the type of procedure performed (open surgery vs endovascular repair). The estimation of the intercept, regression coefficients, and calibration slope led to the following formula: (30-day mortality)_{study data} = − 5.30 + 0.50 * (procedure) + 0.052 * GAS. The AUC adjusted for optimism was 0.70. The adjusted GAS odds ratios are listed in Table IV. In the third step, the type of procedure performed (open surgery vs endovascular repair) was added to the GAS variables, and new regression coefficients were estimated for each variable. The AUC was adjusted for optimism by bootstrapping and was 0.68. Since the second model had the highest optimism-corrected AUC, we used it to calculate the updated GAS. Multiplication with the weights in the original GAS and rounding gives the following risk score:

Table IV. Multivariable models on 30-day mortality

Variable	OR (95%-confidence interval)
	Step 1	Step 2	Step 3
Procedure performed (open surgery vs endovascular repair)	—	1.65(0.81-3.35)	1.93(0.92-4.03)
Age(per decade)	2.39(1.52-3.75)	2.21(1.18-4.13)	2.44(1.52-3.90)
Shock	1.66(0.80-3.45)	3.82(2.29-6.38)	1.44(0.68-3.05)
Myocardial disease	1.76(0.90-3.44)	1.81(1.13-2.89)	1.74(0.88-3.42)
Cerebrovascular disease	1.13(0.46-2.80)	2.20(1.19-4.07)	1.13(0.45-2.81)
Renal disease	2.06(0.83-5.06)	3.03(1.55-5.89)	2.17(0.87-5.41)
Area under the curve	0.67(0.60-0.74)	0.70(0.62-0.77)	0.68(0.61-0.75)

OR, Odds ratio.

Table V shows an example of how to calculate the 30-day mortality for a patient with a ruptured AAA for endovascular repair or open surgery. Fig 2 shows the 30-day mortality depending on the updated GAS.

Table V. How to calculate 30-day mortality after endovascular repair or open surgery for patients with ruptured AAA; an example of the calculation of 30-day mortality as a function of the updated GAS

AAA, Abdominal aortic aneurysm; GAS, Glasgow Aneurysm Score.

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  The 30-day mortality as a function of the updated Glasgow Aneurysm Score (GAS) for patients with a ruptured abdominal aortic aneurysm undergoing either endovascular or open repair.

Back to Article Outline

Discussion 

Policies for treatment of AAA have changed since the introduction of endovascular repair for patients with ruptured AAA. Patients who are hemodynamically stable receive a CT-scan and some an additional angiography prior to the procedure to assess anatomic eligibility for endovascular repair. Those who are too hemodynamically unstable to undergo imaging are immediately transported to the operating room for open surgery. Therefore, in this prospective multicenter study we aimed to validate the GAS both in patients with ruptured AAAs treated with open surgery and in those treated with endovascular repair. Furthermore, we updated the GAS to predict 30-day mortality after either endovascular repair or open surgery in patients with ruptured AAA. We found that the GAS showed limited discriminative ability in our patient population. In addition, we showed that, considering the included risk factors, 30-day mortality was always lower if patients with ruptured AAA were treated with endovascular repair as opposed to open surgery. This means that no risk factors were associated with a higher mortality for a specific procedure. Furthermore, the advantage for endovascular repair appeared to be caused by difference in intraoperative mortality.

The limited discriminative ability of the GAS may be due to the introduction of endovascular repair in patients with ruptured AAA. When the GAS was developed, open surgery was the only treatment for ruptured AAAs. The limited discriminative ability of the GAS suggests that factors not involved in the GAS influenced mortality after repair for ruptured AAA. In addition, in the evaluation of predictive values of the GAS, we found that patients with a high GAS would benefit less from endovascular repair. This is not consistent with previous findings, suggesting that patients at higher risk for periprocedural cardiac complications would benefit more from endovascular repair than those at lower risk.¹⁴ It should be noted, however, that confidence intervals surrounding the odds ratios were wide.

In the updated GAS, we added one new variable to the GAS variables: the type of procedure performed. It turned out that patients who were in shock prior to the procedure had lower 30-day mortalities than those who were not. This may be due to selection criteria since patients not in shock are expected to have lower 30-day mortalities than those who are in shock and, therefore, were immediately transported to the operating room for open surgery.⁷ Similar to patients not in shock, patients in shock might better undergo endovascular repair than open surgery if the anatomy allows. Therefore, it remains to be clarified whether patients in shock should undergo imaging prior to the procedure, although most patients appear to be sufficiently stable to do so.²⁴

Furthermore, the updated GAS showed that patients who underwent endovascular repair had lower 30-day mortalities than those who underwent open surgery. Again, this may be due to selection criteria since endovascular repair patients were mostly not in shock, while 28% of open surgery patients were. In addition, the interaction terms between GAS variables and the therapeutic procedure performed were not associated with 30-day mortality meaning that no risk factors were associated with a higher mortality for a specific procedure. Instead, endovascular repair always favored open surgery with respect to 30-day mortality given the set of risk factors included in the GAS model. Furthermore, we found no statistically significant association between the time to repair and 30-day mortality. Therefore, the delay due to performing a CT-scan may have limited influence on mortality.

Two studies that validated GAS in patients with ruptured AAA treated with open surgery reported better validity than our study,², ³ and two studies reported worse validity.⁴, ⁵ It should be noted that these studies were performed in patients treated with open surgery, whereas in our study patients treated with open surgery as well as patients treated with endovascular repair were included. Studies that investigated the performance of the GAS in patients with elective AAA did not show consistent results.¹⁵, ¹⁶, ¹⁷, ¹⁸, ¹⁹

Our study had several limitations. The definitions of risk factors were slightly different from the original GAS. In addition, since the intent is for the model to be predictive, the GAS only included patient characteristics that can be known upon patients' presentation to the hospital or shortly thereafter. Consequently, we did not collect data on the AAA anatomy, such as neck length, neck diameter, and neck angulation. Furthermore, our prediction rule is not based on a randomized controlled trial, and the selection for endovascular repair was based on patients' hemodynamic condition and AAA eligibility for endovascular repair. Therefore, selection bias may have affected our results in favor of endovascular repair and the updated GAS may not be generalizable to all practice settings. The data we used, however, were based on patients who were seen consecutively, and our study represents current clinical practice. Also, we had a small sample; therefore, lack of statistical power may have affected our results.

The treatment protocols between the participating hospitals were slightly different. The criteria for patients being hemodynamically (un)stable differed across the hospitals. In addition, the types of endografts used were not the same in all hospitals. As a result, different anatomic criteria were applied across the participating hospitals. It should be noted that in practice, physicians tend to apply more lenient criteria for endovascular repair in case of severe comorbidity in order to avoid open surgery. In order to implement a more uniform treatment policy for patients with ruptured AAAs, similar protocols are needed in the different hospitals. In addition, these protocols enable more precise comparison of endovascular repair with open surgery across the different hospitals.

We recommend ongoing prospective observational and randomized controlled trials in patients with ruptured AAA. Prospective observational studies reflect daily practice and changes in treatment policy over time.²⁵, ²⁶ This is of particular interest in this group of patients, since new types of endografts, which allow for more lenient anatomic criteria, are rapidly becoming available. Randomized controlled trials are needed in order to assess associations between risk factors, the procedure performed, 30-day mortality, and to avoid selection bias. New technologies are only proven in a randomized fashion. If such data are not available, as is the case here, it is ethically correct to randomize patients, since in observational studies biases are likely to be an issue. Furthermore, future studies should investigate which patients should go immediately to the operating room for open surgery and which patients should undergo imaging prior to the therapeutic procedure to determine anatomic eligibility for endovascular repair. In addition, since patient populations may change over time, the development of a prediction tool is an ongoing process; therefore, we encourage further validation and updating of our prediction rule.

Back to Article Outline

Conclusion 

In conclusion, we showed limited discriminative ability of the GAS in patients with ruptured AAAs to be treated with endovascular repair or open surgery. The GAS was updated by adding the type of procedure performed, which predicts 30-day mortality for patients with ruptured AAAs to be treated with either endovascular repair or open surgery.

Back to Article Outline

Author contributions 

Back to Article Outline

The authors thank the members of the Assessment for Radiological Technology (ART) group and Theo Stijnen for their comments and suggestions, Marcel de Wilde from the Erasmus MC Department of Medical Informatics for the development of a digital data entry form, and Mantiva Bandasak for her role in the data entry.

Back to Article Outline

References 

1. Samy AK, Murray G, MacBain G. Glasgow aneurysm score. Cardiovasc Surg. 1994;2:41–44
   • View In Article
   • MEDLINE
2. Leo E, Biancari F, Nesi F, Pogany G, Bartolucci R, De Pasquale F, et al. Risk-scoring methods in predicting the immediate outcome after emergency open repair of ruptured abdominal aortic aneurysm. Am J Surg. 2006;192:19–23
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (86 KB)
   • CrossRef
3. Korhonen SJ, Ylönen K, Biancari F, Heikkinen M, Salenius JP, Lepäntalo M Finnvasc Study Group. Glasgow Aneurysm Score as a predictor of immediate outcome after surgery for ruptured abdominal aortic aneurysm. Br J Surg. 2004;91:1449–1452
   • View In Article
   • MEDLINE
   • CrossRef
4. Tambyraja AL, Fraser SC, Murie JA, Chalmers RT. Validity of the Glasgow Aneurysm Score and the Hardman Index in predicting outcome after ruptured abdominal aortic aneurysm repair. Br J Surg. 2005;92:570–573
   • View In Article
   • MEDLINE
   • CrossRef
5. Tambyraja AL, Lee AJ, Murie JA, Chalmers RT. Prognostic scoring in ruptured abdominal aortic aneurysm: a prospective evaluation. J Vasc Surg. 2008;47:282–286
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (84 KB)
   • CrossRef
6. Yusuf SW, Whitaker SC, Chuter TA, Wenham PW, Hopkinson BR. Emergency endovascular repair of leaking aortic aneurysm. Lancet. 1994;344:1645
   • View In Article
   • CrossRef
7. Visser JJ, van Sambeek MR, Hamza TH, Hunink MG, Bosch JL. Ruptured abdominal aortic aneurysms: endovascular repair versus open surgery–systematic review. Radiology. 2007;245:122–129
   • View In Article
   • CrossRef
8. Franks S, Lloyd G, Fishwick G, Bown M, Sayers R. Endovascular treatment of ruptured and symptomatic abdominal aortic aneurysms. Eur J Vasc Endovasc Surg. 2006;31:345–350
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (102 KB)
   • CrossRef
9. Larzon T, Lindgren R, Norgren L. Endovascular treatment of ruptured abdominal aortic aneurysms: a shift of the paradigm?. J Endovasc Ther. 2005;12:548–555
   • View In Article
   • MEDLINE
   • CrossRef
10. Castelli P, Caronno R, Piffaretti G, Tozzi M, Laganà D, Carrafiello G, et al. Ruptured abdominal aortic aneurysm: endovascular treatment. Abdom Imaging. 2005;30:263–269
    • View In Article
    • MEDLINE
    • CrossRef
11. Alsac JM, Desgranges P, Kobeiter H, Becquemin JP. Emergency endovascular repair for ruptured abdominal aortic aneurysms: feasibility and comparison of early results with conventional open repair. Eur J Vasc Endovasc Surg. 2005;30:632–639
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (256 KB)
    • CrossRef
12. Hinchliffe RJ, Bruijstens L, MacSweeney ST, Braithwaite BD. A randomised trial of endovascular and open surgery for ruptured abdominal aortic aneurysm - results of a pilot study and lessons learned for future studies. Eur J Vasc Endovasc Surg. 2006;32:506–513discussion 514-5.
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (172 KB)
    • CrossRef
13. Peppelenbosch N, Geelkerken RH, Soong C, Cao P, Steinmetz OK, Teijink JA, et al. Endograft treatment of ruptured abdominal aortic aneurysms using the Talent aortouniiliac system: an international multicenter study. J Vasc Surg. 2006;43:1111–1123discussion 1123.
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (280 KB)
    • CrossRef
14. Visser JJ, Bosch JL, Hunink MG, van Dijk LC, Hendriks JM, Poldermans D, et al. Endovascular repair versus open surgery in patients with ruptured abdominal aortic aneurysms: clinical outcomes with 1-year follow-up. J Vasc Surg. 2006;44:1148–1155
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (120 KB)
    • CrossRef
15. Bohm N, Wales L, Dunckley M, Morgan R, Loftus I, Thompson M. Objective risk-scoring systems for repair of abdominal aortic aneurysms: applicability in endovascular repair?. Eur J Vasc Endovasc Surg. 2008;36:172–177
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (194 KB)
    • CrossRef
16. Baas AF, Janssen KJ, Prinssen M, Buskens E, Blankensteijn JD. The Glasgow Aneurysm Score as a tool to predict 30-day and 2-year mortality in the patients from the Dutch Randomized Endovascular Aneurysm Management trial. J Vasc Surg. 2008;47:277–281
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (96 KB)
    • CrossRef
17. Tang TY, Walsh SR, Fanshawe TR, Seppi V, Sadat U, Hayes PD, et al. Comparison of risk-scoring methods in predicting the immediate outcome after elective open abdominal aortic aneurysm surgery. Eur J Vasc Endovasc Surg. 2007;34:505–513
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (305 KB)
    • CrossRef
18. Hirzalla O, Emous M, Ubbink DT, Legemate D. External validation of the Glasgow Aneurysm Score to predict outcome in elective open abdominal aortic aneurysm repair. J Vasc Surg. 2006;44:712–716discussion 717.
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (102 KB)
    • CrossRef
19. Faizer R, DeRose G, Lawlor DK, Harris KA, Forbes TL. Objective scoring systems of medical risk: a clinical tool for selecting patients for open or endovascular abdominal aortic aneurysm repair. J Vasc Surg. 2007;45:1102–1108
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (148 KB)
    • CrossRef
20. Central Committee on Research Involving Human Subjects. http://www.ccmo-online.nl/Accessed: February 10, 2006.
    • View In Article
21. van der Heijden GJ, Donders AR, Stijnen T, Moons KG. Imputation of missing values is superior to complete case analysis and the missing-indicator method in multivariable diagnostic research: a clinical example. J Clin Epidemiol. 2006;59:1102–1109
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (126 KB)
    • CrossRef
22. Steyerberg EW, Borsboom GJ, van Houwelingen HC, Eijkemans MJ, Habbema JD. Validation and updating of predictive logistic regression models: a study on sample size and shrinkage. Stat Med. 2004;23:2567–2586
    • View In Article
    • MEDLINE
    • CrossRef
23. Harrell F. Regression Modeling Strategies: with Applications to Linear Models, Logistic Regression, and Survival Analysis. New York: Springer; 2001;
    • View In Article
24. Lloyd GM, Bown MJ, Norwood MG, Deb R, Fishwick G, Bell PR, et al. Feasibility of preoperative computer tomography in patients with ruptured abdominal aortic aneurysm: a time-to-death study in patients without operation. J Vasc Surg. 2004;39:788–791
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (92 KB)
    • CrossRef
25. Hordijk-Trion M, Lenzen M, Wijns W, de Jaegere P, Simoons ML, Scholte op Reimer WJ, et al. Patients enrolled in coronary intervention trials are not representative of patients in clinical practice: results from the Euro Heart Survey on Coronary Revascularization. Eur Heart J. 2006;27:671–678
    • View In Article
    • MEDLINE
    • CrossRef
26. Lloyd-Williams F, Mair F, Shiels C, Hanratty B, Goldstein P, Beaton S, et al. Why are patients in clinical trials of heart failure not like those we see in everyday practice?. J Clin Epidemiol. 2003;56:1157–1162
    • View In Article
    • Abstract
    • Full Text
    • Full-Text PDF (97 KB)
    • CrossRef