Progress in abdominal aortic aneurysm surgery: four decades of experience at a teaching center1

ELSEVIER SCIENCE5

PII:SO967-21O9(97)OOOO3-3

Cardiovascular Surgery, Vol. 5, No. 2, pp. 15(L156, 1997 0 1997 The Intemationat Society for Cardiovascular Surgery Published by Elsevier Science Ltd. Printed in Great Britain 0967-2109/97 $17.00 + 0.00

Progressin abdominalaortic aneurysmsurgery: four decades of experienceat a teaching center J. C. Chen,H. D. Hildebrand,A.J. Salvian,Y.N. Hsiangand D.C. Taylor Departmentof Surgery, VancouverHospitaland HealthScienceCentre,Universityof British Columbia,Vancouver,BritishColumbia,Canada The purpose of this study was to examine the changing trends in surgical management of patients with abdominal aortic aneurysms at a tertiary care teaching hospital over the past 40 years, by analysisof demographic data, perioperativevariablesand outcomes on allpatients having abdominal aortic aneurysm surgery between 19!3!5and 1993. Some 1604 abdominal aortic aneurysms were assessed. The annual rate of abdominal aortic aneurysm surgery increased from 17.6 to 67.8 cases per year. The non-ruptured to ruptured abdominal aortic aneuysm ratio increased from 2.4 : 1 in the first decade to 3.4 : 1 in the last 5 years. In nonruptured abdominal aortic aneurysm repairs, the followingvariables changed over the four decades: patients age over 80 years increased (2.4°%J to 8.0%; P< O.04), concomitant lowerlimb occlusivedisease increased (12.2% to 23.7Yo;P
Introduction Modern abdominal aortic aneurysm surgery began in 1951 with the report by Dubost et al. [1] of aneurysm resection and interposition grafting. Much pro-

Comespondenceto: Dr D. C. Taylor, Divisionof Vascular Surgery, Department of Surgery, Vancouver Hospital, University of British Columbia, 910 West IOthAvenue,Vancouver,BC, V5Z 4E3 Canada Presented at the Annual Joint Meeting of the Society for Vascular Surgeryand the North AmericanChapter of the InternationalSociety for CardiovascularSurgery, Chicago, Illinois, 9-13 June 1996

1s0

gress has been made since then in many aspects of abdominal aortic aneurysm surgery with regard to patient selection, perioperative management, and operative technique. In recent years, there has been considerable enthusiasm for endovascular stent grafting as a promising new method of treating such aneurysms [2–8]. Any new therapy, however, must have comparable outcomes with the current standard before it can be completely adopted. In this study, all patients who underwent abdominal aortic aneurysm surgery in the authors’ institution were examined to identify secular trends and to define the CARD1OVASCUMR SURGERY APRIL1997 VOLS NO2

Progress inabdominal aordcaneurysm surgery: J. C.Chen et al. current standard in the perioperative management of these patients.

Patients and methods Patients All patients who had abdominal aortic aneurysm surgery at the Vancouver General Hospital between 1955 and 1993 were reviewed. The latteris a tertiary care university affiliatedteaching hospital. Data were examined from a hospital aneurysm registry that has been maintained since 1955. Patient information was entered concurrently by medical quality appraisers unaware of the purpose of this study. The study period was divided into six sequential time intervals from 1955 to 1966, 1967 to 1973, 1974 to 1978, 1979 to 1983, 1984 to 1988, and 1989 to 1993. The variables recorded for each time period included number of cases per year, 30-day inhospital mortality rates, preoperative variables including demographic data (age and gender), prevalence of concomitant medical conditions (coronary artery disease, congestive heart failure, hypertension, peripheral vascular disease, cerebral vascular disease, diabetes, chronic obstructive lung disease, smoking, renal dysfimction) and size of aneurysms. Intraoperative variables included, duration of operation and clamp time, graft type and material, and frequency of complications (major venous tear, intraoperative hypotension, blood loss >31 for non-ruptured and >41 for ruptured abdominal aortic aneurysm). Postoperative variables recorded were complication rates involving each organ system as well as graft and wound complications. For ruptured aneurysms, additional variables examined were prevalence of preoperative shock (systolic blood pressure <80 mmHg in the emergency room), unconsciousness, >1 h time lapse from the emergency room to operating room, and preinduction hypotension (systolic blood pressure <80 mmHg before induction of anesthesia). Each variable was compared between periods. Non-ruptured and ruptured aneurysms were analyzed separately. Statistical methods Categorical variables were tested using the ~ square test, while continuous variables were tested with the two-tailed t-test. Variables not recorded in every time period were excluded from analysis.Significance was assumed if P< O.05.

Results A total of 1604 patients who had surgery in the past 40 years were studied. There were 1168 patients CARD1OVASCULAR SURGERY APRIL 1997 VOL S NO 2

with non-ruptured abdominal aortic aneurysm, and 436 with ruptured aneurysms. The proportion of male patients was 83.7°/0, and the male : female ratio did not change significantly over the study period. The rate of aneurysm surgery at the authors’ hospital increased from 17.6 cases per year in 1955– 1966 to 67.8 cases per year in 1989–93 (Figure 1). aneurysm aortic Non-ruptured abdominal accounted for the majority; the non-ruptured to ruptured ratio increased from 2.4 : 1 in the first decade to 3.4 : 1 in 1989–1993. The mortality rate for non-ruptured abdominal aortic aneurysm decreased from 17.OYOin the 1955– 1966 period to 3.4’%oin the last period (P<0.0001; Fi&we 2). In contrast, mortality rate for ruptured aneurysms did not change significantly over the entire study period (54.2°/0 in 1955–1966 versus

44.2’Yoin 1989–1993; P= 0.3; Fi&we 2). Non-ruptured aneurysms

Preoperativecharacteristics The proportion of patients over the age of 80 steadily increased. From 1955–1966, 2.4°/0 of non-ruptured patients were aged over 80 years, while in 1989– 1993, the proportion increased to 8% (P< O.05). Patients with concomitant lower-limb occlusive disease increased from 12.20/0in 1955–1966 to 23.70/o in 1989–1993 (P< O.05). Patients with renal dysfunction (preoperative blood urea nitrogen >14 mmol/1) increased from 22’XOin 1955–1966 to 35.99’o in 1989–1993 (P< O.01). The size of aneurysms operated on steadily decreased. From 1955– 1966, only 16.O’%0of aneurysms were <6 cm in diameter, whereas in 1989–1993, 54.2’Yo of aneurysms repaired were <6 cm (P< O.001). The presence of associated conditions such as coronary artery disease (46’Yo), high blood pressure (36%) and cerebral vascular disease (1 1%) did not vary significantly between periods. Intraoperativevariables Length of operation and clamp time shortened over the periods studied. In 1955–1966, the average operating room time was 247 min and 34’%oof the aortic clamp times were >75 min. In 1989–1993, the average operating room time decreased to 208 min (P< O.05) and only 11.5% of aortic clamp times were >75 min (P< O.001). Proportions of patients with >31 blood loss decreased significantly between 1955–1966 and 1967–1973 from 24.4 to 12.6Y0 respectively (P< O.02), and remained steady at that rate for the subsequent periods. The rate of intraoperative graft thrombosis decreased significantly from 4% in 1955–1966 to 1.5% in 1989–1993 (P= 0.01). The proportion of patients with intraoperativehypotension (systolic blood pressure <80 mmHg for 1s1

Progress inabdominal aordc aneurysm surgery: J. C. Chen et al. 70 ORAAA

M

4

mNRAAA

IE3TOTAL

60

55-66

67-73

83-88

79-83

74-78

89-93

PERIOD Figure 1

Ratesof abdominalaortic aneurysmsurgeryfrom 1955 to 1993

60

I

**

50 ** 40 30 20 10 0 55-65

66-73

74-78

79-83

84-88

89-93

PERIOD Z Mortality rates of non-ruptured abdominal aortic aneurysm and ruptured abdominal aortic aneurysm patients from 1955 to 1993. Figure *, P
P=0.3

>5 rein) decreased significantlybetween 1955–1966 and 1989–1993 from 9.0 to 0.7’%0(P< O.0001). The rate of major intraoperativevenous injures remained unchanged at 4 to 6’Yothroughout all six periods. 1s2

Postoperativecomplications The rate of postoperative graft thrombosis causing leg ischemia decreased between 1955–1966 and CARD1OVASCUMR SURGERY APRIL1997 VOLS NO2

Progress in abdominal aortic 1989–1993 from 5.7 to 1.l Yo (P <0.05). Limb ischemia requiring amputation decreased from 3.20/. in 1955–1966 to O’YOin 1989–1993 (F’<0.05). The rate of postoperative hemorrhage requiring re-exploration also decreased significantly from 8.2Y0 in 1955–1966 to O’XO in 1989–1993 (P< O.0001). Postoperative myocardial infarction, renal failure, graft infection, wound infection, and wound dehiscence showed no significant trend with occurrence rates of 3.4°/oj 6.10/., 0.40/., 1.9°/0 and 1.30/., respectively in 1989–1993. These results are shown in Table 1. Ruptured aneurysms

Preoperativecharacteriwics Similar to non-ruptured abdominal aortic aneurysm patients, the proportion of patients with ruptured aneurysms who were over the age of 80 increased during the period studied. Time lapse from emergency room admission to arrivalat the operating room decreased significantly over the years. In 1955–1966, only 8.7’%oof patients had an emergency to operating room time of <1 h, where as in 1989–1993, 55.80/. of patients were in the operating room within 1 h of their admission (P< O.0001). Preinduction hypoten-

aneurysm surgery:

J. C. Chen et al.

sion decreased significantly from 46.7’XOin 1955– 1966 to 19.5’Yoin the 1989–1993 period (P< O.001). There were no significant differences found between periods in the proportion of patients with associated coronary artery disease (42.0’%.), high blood pressure (41. O’XO),cerebral vascular disease (9.6%) or lower-limb occlusive disease (12.0%). Prevalence of shock in the emergency room remained unchanged at 550/..

Intraoperativevan”ables There was a significant decrease in the frequency of intraoperative hypotension from 50.0’7. in 1955– 1966 to 23.5Y0 in 1989–1993 (P< O.001). The proportion of patients with major venous injuries decreased significantlyfrom 19.3% in 1955–1966 to 5.9% in 1989-1993 (P< O.05). The mean proportion of patients with blood loss >41 (58.8’%.), rate of intraoperative distal thrombosis (4%), average operative time (225 rein) and the proportion of clamp time >75 min (52%) did not change significantly over the periods studied.

Postoperativecomplications Rates of graft thrombosis causing leg ischemia decreased between 1955–1966 and 1989–1993 from

Table1 Prevalenceof variablesexaminedin non-ruptured aneurysmpatients: comparingperiods 1955-1966 and 1989-1993 Variable Preoperative Aneurysmsize <6 cm Renaldysfunction(BUN>14 mmoll) Age >80 years Age <60 years Lower-limb occlusivedisease Male gender Coronaryartery disease High blood pressue Cerebralvasculardisease Intraoperative Hypotension Clamptime >75 min Graft thrombosis Blood 10SS >31 Averageoperating room time (rein) Venoustear Postoperative Reoperationfor hemorrhage Wound dehiscence Graft thrombosis Amputation Wound infection Graft infection Myocardialinfarction Renaldysfunction(BUN>35 mmoll) Ischemiccolitis Pulmonarycomplications Mortality

1955-1966

1989-1993 (“fi)

P

16.0 22.0 2.4 18.8 12.2 86.3 42.7 29.0 11.3

54.2 35.9 8.0 7.6 23.7 82.4 46.9 31.3 15.6

<0.001 <0.01 <0.05 <0.05 <0.05 n.s. n.s. n.s. n.s.

9.0 34.0 4.0 24.4 247 6.7

0.7 11.5 1.5 12.6 208 6.1

<0.0001 <0.001 0.01 <0.02 <0.05 n.s.

8.2 3.6 5.7 3.2 2.5 1.0 2.5 10.2 1.8 3.0 17.0

0.0 1.3 1.1 0.0 1.9 0.4 3.4 6.1 1.1 5.3 3.4

<0.0001 <0.01 <0.05 <0.05 n.s. n.s. n.s. n.s. n.s. n.s. <0.0001

(Ye)

BUN,blood urea nitrogen; n.s., not significant.

CARD1OVASCULJIR SURGERY APRIL1997 VOL!3 NO 2

153

Progress in abdominalaortic aneurysmsurgery:J C. Chenet al.

12.9’-XO to 3.2’Yo(RO.05).

Wound dehiscence rates

decreased between the same periods from 7.O’%O to 0’%0 (P< O.05YO). The mean proportion of patients with bleeding requiring reoperation (10.60/0), renal dysfunction (29’Yo), wound infection (1.3%) and myocardial infarction (15.5°/0) did not show significant differences. These results are summarized in Table 2. Graft material In 1955–1966 and 1967–1973, Teflon was the most commonly used graft used, accounting for 80’Yoand 67% of all grafts, respectively. Other graft materials used during the 1955–1966 period included homograft (9’70), nylon (3’Yo), and Dacron (14’Yo). Over the last two decades, Dacron has become the most common graft material, accounting for 99’ZOof the grafts used. In the last decade, there has been a shift from bifurcated to tube grafts such that tube graft use increased from 14°/0 in 1984–1988 to 56°/0 in 1989–1993 (P< O.05).

Discussion This study identified significant changes in abdominal aortic aneurysm surgery at the authors’ insti-

tution over the past four decades. As one of the longest abdominal aorric aneurysm series reported to date, this study offers a unique longitudinal perspective of aneurysm surgery and allow observations of trends and progress. The annual rate of abdominal aortic aneurysm repair has increased steadily at the authors’ hospital, with non-ruptured abdominal aortic aneurysm accounting for the majority of this increase. The reasons for this may be attributed to increased utilization of abdominal ultrasound and computed tomography scans, an increasing population that is also aging, and increased incidence of abdominal aortic aneurysm [9, 10]. Non-ruptured aneurysms The higher proportion of octogenarians in the later intervals likely reflects aging and increased longevity of the referral population, as well as liberalization of indications as the risk of surgery decreases. This was observed in both non-ruptured and ruptured abdominal aortic aneurysm patients. Since the prevalence of abdominal aortic aneurysm increases in older age groups [11], a higher proportion of octogenarians is expected. Older patients have more associated conditions. Concomitant lower-limb

TableZ Prevalenceof variablesexaminedin ruptured aneurysmpatients:comparingperiods 1955-1966 and 1989-1993 Variable Preoperative Emergencyroom to operating room <1 h PreinductionBP <80 mmHg systolic Age >80 years Age <60 years Male gender Shock(BP <80 mmHg systolic) Coronaryartery disease High blood pressure Cerebralvasculardisease Lower-limb occlusivedisease lntraoperative Hypotension Venoustear Blood ]0SS >4 I Graft thrombosis Clamptime >75 min Averageoperating room time (rein) Postoperative Graft thrombosis Wound dehiscence Reoperationfor hemorrhage Wound infection Graft infection Myocardialinfarction Renaldysfunction(BUN>35 mmoL4 Ischemiccolitis Pulmonarycomplications Mortality

1955-1966 (Y.)

1989-1993 (Ye)

P

8.7 46.7 6.5 12.0 89.5 50.0 43.0 41.0 11.0 12.9

55.8 19.5 13.0 2.6 79.2 44.2 33.8 41.6 13.0 7.8

<0.0001 <0.001 <0.05 <0.05 n.s. n.s. n.s. n.s. n.s. n.s.

50.0 19.3 62.0 3.0 57.8

23.5 5.9 58.8 8.0 54.5 230

<0.001 <0.05 n.s. n.s. n.s. n.s.

12.9 7.0 8.0 7.2 1.0 10.5 31.5 13.1 39.5 54.2

3.2 0.0 7.8 1.3 0.0 18.2 22.6 7.8 45.2 44.2

<0.05 <0.05 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.

BP, blood pressure;BUN,blood urea nitrogen: n.s., not significant.

1s4

CARD1OVASCULAR SURGERY APRIL1997 VOL S NO Z

Progress in abdominal aortic aneurysm surgery: J. C. Chen et al.

occlusive disease and renal dysfunction were significantly more common during laterperiods in the nonruptured abdominal aortic aneurysm patients. In a previous study, these were both shown to predict mortality risk [12]. Despite the increased frequency of co-morbid conditions, advances in perioperative management has led to improved operative outcomes. The increased proportion of small aneurysms is likely attributed to improved physician awareness, resulting in early detection of abdominal aortic aneurysm and from increased use of abdominal ultrasound and computed tomography scans for common abdominal symptoms. In the authors’ present practice, atherosclerotic aneurysms <5 cm were managed non-operatively unless they were symptomatic, rapidly expanding or associated with other complications. Significant decreases were seen in intraoperative duration and complications over the decades. These included length of operation, prolonged aortic clamp time, blood loss, hypotensive episodes, and intraoperative distal arterial thrombosis. Although speed is not an essential goal of this operation, prolonged o~erative time correlates well with Door outcome b~cause it reflects intraoperative d&iculties and complications. An early technical advance occurred in the 1955–1966 .~eriod when aneurvsm resection . was abandoned in favor of intraluminalgrafting,with the aneurysm wall left in situ for coverage [13, 14]. This may have accounted for the decrease in operative blood loss and postoperative hemorrhage. The latter was the most common cause of death in the authors’ report of the first decade [15], while in the 1989–1993 period, cardiac events accounted for six of the nine deaths. The routine use of intravenous heparin, staged declamping and flushing techniques advocated by Imparato et al. [16] have likely helped to reduce the incidence of intraoperative as well as postoperative graft thrombosis, distal ischemia and need for amputation. Improvement in mortality rate in non-ruptured abdominal aortic aneurysm patients parallels many of the trends described above and likely is the direct result of advancements in all aspects of aneurysm surgery combined. Ahhough there was not a significant difference in the incidence of postoperative myocardial infarction, a decreasing trend was observed in the last four periods. Ruptured aneurysms In the ruptured abdominal aortic aneurysm patients, the significant increase in the proportion of patients operated on within 1 h of their hospital arrivaldemonstrates improvement in emergency room diagnosis and rapid transfer of these patients to the operating room. The proportion of unstable patients, however, CARDIOVASCULAR SURGERY APRIL1997 VOL5 NO2

as unchanged with advancing periods as the proportion of patients with preoperative shock, preoperative azotemia and co-morbid conditions were not different between periods. The decrease in the proportion of patients with preinduction hypotension, from 50!4. to 24Y0,can be attributed to more effective preoperative fluid resuscitation by the emergency response team. Realizing that the definition of appropriate fluid resuscitation has remained controversial, the authors’ present practice is to resuscitateruptured abdominal aortic aneurysm patients to a systolic blood pressure of 100 mmHg until surgical control is obtained. Improvement in intraoperative management is likely responsible for the decrease in patients with intraoperative hypotension, major venous tears and >41 blood 10SS, Heparin was not used in the majority of the ruptured aneurysm patients, which may explain the steady level of intraoperative graft thrombosis at 4’%.. Despite improvements in various preoperative and intraoperative parameters, significant decreases in postoperative complications were limited to lower post-operative thrombosis and wound infection rates. The rate of death, myocardial infarction, renal failure, and ischemic colitis remained essentially unchanged in this group of patients over the four decades. The mortality rate of ruptured aneurysm patients dipped to its lowest level in 1979–1983 at 33.8’%., and then rose again in 1984–1988 to 480/. such that no unilateral trend was observed. No specific factor could account for this. These mortality rates compare well with rates reported by others [17–26]. It is dishearteningfor the authors not to have seen a decrease in mortality for the surgical treatment of ruptured abdominal aortic aneurysm over four decades of experience. This is in spite of what they feel are advances in prehospital and emergency room care, faster diagnosis, and definitive surgical treatment with improvements in intra- and postoperative care. The authors feel that is at least partly explained by changes in the population to include an older group of patients with more co-morbid conditions. However, it may also be due to more rapid prehospital transport, bringing patients who would previously have died prior to hospital care. The advances documented in this study are the direct result of dedication to patient selection and perioperative management for abdominal aortic aneurysm patients. With endovascular stent grafting looming on the horizon as a promising new technique, vascular surgeons must, once again, lead the way in objectively evaluating its capabilities and defining its roles. Although early results appear promising in selected patients [2–8], widespread recommendation should be withheld until results are shown to be comparable with conventional repair. 1s!3

Progress in abdominal ao~”c aneurysm surgery: J. C. Chen et al.

In conclusion, over the past 40 years, tremendous progress has been made in abdominal aortic aneurysm surgery, resulting in significant decreases in mortality and morbidity associated with elective aneurysm repair. In addition, older patients with smaller aneurysms and more co-morbid conditions were operated on more frequently during this period. Despite these advances, however, mortality for patients with ruptured abdominal aortic aneurysm repair has not changed significantly.

References 1. Dubost, C., Allary,M. and Oeconomos, N., Resectionof an aneurysm of the abdominalaorta: re-establishmentof the continuity by a preserved arterial graft, with result after five months. Archivesof Surgery,1952, 64, 405-408. 2. May, J., White, G. H., Waugh, R. C., Yu, W., Stephen, M. S. and Harris, J. P., Endoluminalrepair of abdominalaortic aneurysms. Medicd30urnal ofAustralia, 1994, 161, 541-543. 3. Moore, W. S. and Vescera, C. L., Repair of abdominal aorric aneurysmby transfemoralendovasculargraft placement. Annals of Surgey, 1994, 220, 331–341. 4. Andrews, S. M., Cuming, R., Macsweeney, S. T., Barrett, N. K., Greenhalgh,R. M. and Nott, D. M., Assessmentof feasibility for endovascularprosthetic tube correction of aortic aneurysm. British~ournalof Surgey, 1995, 82, 917–919. 5. Green, R. M. and Chuter, T. A., Evolution of technologiesin endovasculargrafiing.CardiovascularSurgety,1995, 3, 101–107. 6. Parodi, J. C., Endovascularrepairof abdominalaornc aneurysms and other arterial lesions.Journal of VascularSurgey, 1995, 21, 549–555. 7. Parodi, J. C., Marin, M. L. and Veith, F. J., Transfemoral,endovascular stented graft repair of an abdominal aortic aneurysm. Archivesof Surgey, 1995, 130, 549–552. 8. White, G. H., May, J. and Yu, W., Sydneyexperiencewithendoluminalgratis for the treatment of simpleand complex abdomiof EndovascularSurgey, 1995, 2, nal aonic aneurysms. 3ourncd 220-221. 9. Melton III, L. J., Bickerstaff,L. K. and Hollier, L. H., Changing incidence of abdominal aortic aneurysms: a population based study. AmericanJournal of Epidemiology,1984, 120, 379–386. 10. Stevens, K., The incidenceof abdominalaortic aneurysms.BritishJournal of ClintkzlPractice, 1993, 47, 208–210. 11. Bengtsson, H., Bergqvist, D. and Stemby, N. H., Increasing prevalence of abdominal aortic aneurysms: a necropsy study. EuropeanJournal of Surgey, 1992, 158, 19–23. 12. Chen, J. C., Hildebrand, H. D., Salvian, A. J., Taylor, D. C. and Hsiang Y. N., Predictors of monality in nonruptured and

1S6

rupturedabdominalaortic aneurysms.Journal of VascularSurget-y 1996, 24, 614. 13, Creech, O., Endo-aneurysmorrhaphyand treatment of aortic aneurysm.Annak of Surgey, 1966, 164, 935. 14. Graham, A. L., Najafi, H., Dye, W. S., Javid, H., Hunter, J. A. andJulian, O. C., Rupturedabdominalaorricaneurysm.Archives of Surgey, 1968, 97, 102&1031. 15. Hildebrand,H. D. and Chung, W. B., Abdominalaortic aneurysmectomy: a comparative study of morbidity and monality. AmericanSurgeon,1971, 37, 476481. 16. Imparato, A. M., Berman, I. R., Bracco, A., Kim, G. E. and Beaudet, R., Avoidanceof shockand peripheralembolismduring surgexyof the abdominalaona. Surgey, 1973, 73, 68–73. 17. Johnston, K. W., Muhicenter prospectivestudy of nonruptured abdominal aonic aneurysm: pan II – variablespredicting morbidity and monality. Journal of VascularSurgery, 1989, 9, 437447. 18. Johansen, K., Kohler, R. T., Nicholls, S. C., Zieler, R. E., Clowes, A. W. and Kazmers, A., Ruptured abdominal aonic aneurysm: the Harbotview experience.Journal of VascularSurgey, 1991, 13, 240–245. 19. Harris, L. M., Fagtzioli,G. -- - L.,. Fiedler,. R... Curl., G. R. and Ricotta, J. J., Ruptured abdominalaortic aneurysms:factors affecting mortality rates. Journalof VascularSurgey, 1991, 14, 812– 820. 20. Gloviczki,P., Pairolero, P. C. and Mucha, P., Ruptured aonic aneurysms:repair shouldnot be denied.Journal of VascularSurgey, 1992, 15, 851–859. 21. Ernst, C., Abdominal aortic aneurysm. New EnglandJou~alof Medicine,1993, 328, 1167–1 172. 22. Katz, D. J., Stanley,J. C. and Zelenock, G. B., Operativemortality rates for intact and ruptured abdominal aonic aneurysms in Michigan: an eleven-yearstatewideexperience.Journal of VascularSurgey, 1994, 19, 804–817. 23. Johnston, K. W. and the Canadian Society for VascularSurgery AneurysmStudy Group., Ruptured abdominalaorric aneurysm: six-year follow-up results of a multicenter prospective study. Journal of VascularSurgey, 1994, 19, 888-900. 24. Huber, T. S., Harward, T. R., Flynn, T. C., Albright,J. L. and Seeger, J. M., Operativemortality rates after elective infrarenal aortic reconstructions. Journal of VascularSurgery, 1995, 22, 287–294. 25. Pemeton, J. M., Lsssonde, J. and Laurendeau, F., Ruptured abdominal aonic aneurysm: impact of comorbidity and postoperativecomplicationson outcome. Annalsof VascularSurgery, 1995, 9, 535-541. 26. Hardman, D. T. A., Fisher, C. M., Patel, M. I., Neale, M,, Chambers,J., Lane, R. and Appleberg,M., Ruptured abdominal aorricaneuxysms:who shouldbe offeredsurgery?.Journalof VascularSurgery,1996, 23, 123–129. Paper accepted 31 October 1996

CARDIOVASCULAR SURGERY APRIL1997 VOLS NO2