- Email: [email protected]
Ureteral complications and aortoiliac reconstruction
Ureteral complications and aortoiliac reconstruction Dennis J. Wright, M D , Calvin B. Ernst, M D , James R. Evans, M D , R o g e r F. Smith, M D , Daniel J. Reddy, M D , Alexander D. Shepard, M D , and Joseph P. Elliott, Jr., M D , Detroit, Mich. A 33-year experience with 58 ureteral complications in 50 of 3580 patients undergoing aortoiliac reconstruction was analyzed. Ureteral obstruction was treated before or in conjunction with aneurysm repair in six patients with aneurysmal disease. The remaining 44 patients had 46 ureteral complications after aortic reconstruction; complications included hydronephrosis (42), ureteral leak (3), and ureteral necrosis (1). A high incidence of associated graft complications was noted. Graft thrombosis developed in one of the six patients undergoing prior or simultaneous ureteral procedures, and graft infection developed in another. Thirty-six graft complications developed in 24 (55%) of the 44 patients with postoperative ureteral complications. The complications included 19 anastomotic aneurysms, eight graft limb thromboses, six graft infections, and three aortoenteric fistulas. Twenty-nine of the 44 patients with postoperative ureteral complications underwent ureteral or graft operations or both. These included five patients having ureteral operations alone, seven with a ureteral procedure and subsequent graft operation, eight requiring simultaneous uretera/and graft procedures, and nine undergoing a graft operation with ureteral observation. Six of these 29 patients (21%) died after operation, all from graft complications including aortoenteric fistulas (three), ruptured anastomotic aneurysms (two), and graft infection (one). Graft complications affected 55% of 44 patients with postoperative ureteral complications, compared to 12% of 3536 patients without ureteral complications (p < 0.0001). Patients with postoperative ureteral complications were 4.4 times as likely to have graft complications compared to those without ureteral complications (p < 0.0001). These data suggest that such urologic complications may be markers for recognition of or harbingers for graft complications. (J VASe SURG 1990;11:29-37.) Ureteral complications occur infrequently during management of aortoiliac atherosclerosis. Fewer than 100 cases of preoperative hydronephrosis associated ~vith aortic aneurysms have been reported. In additi~i.~, the incidence of hydronephrosis caused by ureteral obstruction after aortoiliac reconstruction is also low, ranging from 2% to 14% in several prospective studies.l-4 These reports have mainly addressed only the incidence o f ureteral complications, suggesting that progression to persistent ureteral obstruction is rare, especially if hydronephrosis occurs early in the postoperative period. From these analyses it has been assumed that postoperative ureteral complications
have little bearing on ultimate morbidity or mortality after aortic reconstruction. Recently, hydronephrosis caused by ureteral obstruction after aortic reconstruction has been suggested to be associated with graft complications. S To further characterize this relationship and determine the natural history ofureteral complications after aortoiliac reconstruction, a 33-year experience with aortoiliac reconstruction was retrospectively reviewed. The relationships between ureteral and graft complications and outcome of specific management o f these complications were analyzed and form the basis of this article. CLINICAL MATERIAL
.From the Division of Vascular Surgery, Henry Ford Hospital. Presented at the Thirty-seventh ScientificMeeting of the No~.h American Chapter, International Society for Cardiovascular Surgery, New York, N.Y., June 19-20, 1989. Reprint requests: Calvin B. Ernst, MD, Division of Vascular Surgery, Henry Ford Hospital, 2799 W. Grand Blvd., Detroit, MI 48202. 24/6/16662
Beginning in 1954, data have been recorded a n d tracked for patients undergoing vascular surgical procedures at the Henry Ford Hospital by means of a computerized vascular registry. Through periodic patient recall and by following established clinical examination and arteriographic protocols, such data 29
30
Journal of VASCULAR SURGERY
Wright et al.
Table I. Aortic reconstruction in 50 patients with ureteral complications Aneurysmal Occlusive disease disease Total Aortobifemoral reconstruction Aortobiliac reconstruction Other* Total
5 18 1 24
22 1 3 26
27 19 4 50
*One mycotic aneurysm figation; three aortoiliac endarterectomies.
have been regularly updated. 6 Information derived from more than 13,000 vascular surgical procedures have been entered into the registry, and from this database, 3580 patients undergoing 3660 aortoiliac reconstructive procedures were identified. Fifty of the 3580 patients undergoing aortoiliac reconstruction were found to have significant ureteral complications, and their records were scrutinized. Ureteral involvement was documented by urography during aortography, intravenous pyelography (IVP), retrograde pyelography, or operation. Patients were followed by a protocol that included yearly examination. During the first 2 decades of the study, routine translumbar aortography was performed in the immediate postoperative period before hospital discharge in 1958 patients and was repeated at 2- to 3year intervals. All 50 patients were followed through 1988 or until death. Follow-up averaged 7.4 years and ranged from 1 month to 20.3 years. The incidence and type of graft complications were assessed, and the outcome of both the graft and ureteral complication was determined in each patient. The incidence of graft complications was also determined for all 3580 patients entered into the database during the period of this study. Statistical methods. Statistical analysis focused on patients with postoperative ureteral complications. Patients with preoperative ureteral complications were counted as patients without postoperative complications. To assess the association between postoperative ureteral and graft complications, each patient was included as having any type of graft complication or any ureteral complication or both. Thus each patient was counted only once regardless of the number of procedures performed or complications identified. A relative risk with a 95% confidence interval was computed. Log-rank analysis and Kaplan-Meier curves, 7 which adjust for varying lengths of follow-up, were used to compare graft complication rates between patients with (n = 44) and without (n = 3536)
postoperative ureteral complications. In this analysis only graft complications occurring after the first operation were used. The length of follow-up was cal-~ culated as the time from first operation to graft complication, date of last examination, or death. Cox proportional hazards regression analysis 8 was used to adjust for age, sex, race, and length of followup and to compare graft complication rates for those patients with and without postoperative ureteral complications. Again, only the first operation was used in this analysis. Risk estimates and confidence intervals were calculated by use of the results from Cox regression. A p value of less than 0.05 was considered statistically significant. RESULTS
Fifty-eight ureteral complications were identified in 50 patients. There were 29 men and 21 w ~ e n with an average age of 61.3 years, ranging from 42 to 78 years. The overall incidence of ureteral complications was 1.4% (50/3580). Twenty-six patients underwent aortoiliac reconstruction for occlusive disease, and 24 patients underwent reconstruction for aneurysmal disease (Table I). Six patients had preoperative ureteral obstruction, graft complications developed in two. The remaining 44 patients had 46 postoperative ureteral complications manifest by hydronephrosis in 42, ureteral leak in three, and ureterat necrosis in one. Ureteral complications were identiffed from 2 weeks to 12 years postoperatively (average, 18.9 months). Early ureteral complications were arbitrarily defined as those detected within 1 year of aortic reconstruction, and late complications were defined as those detected beyond 1 year. Twenty-four patients (55%) had early ureteral c~mplications, and 20 (45%) had late complications discovered from 14 months to 12 years after operation. Thirty-six graft complications developed in 24 of the 44 patients. Graft complications were identifiec~ from 4 weeks to 12.8 years after surgery (average, 3.3 years). Graft complications were evenly divided between the early ureteral complication group and the late ureteral complication group. Twenty-nine of the 44 patients with postoperative ureteral complications underwent ureteral or graft operations or both. These included five patients undergoing ureteral operations alone, seven undergoing a ureteral procedure and subsequent graft operation, eight undergoing simultaneous ureteral and graft procedures, and nine undergoing graft operation and ureteral observation (Table II). SIX of tb_.7
Volume 11 Number 1 January 1990
Ureteral complications and aortoiliac reconstruction
Table II. Patients undergoing graft and/or ureteral operations (n = 29)
20 Patientsureteral operations
5 Ureteral operations alone 7 Ureteral operations with subsequent graft operataon 8 Simultaneous ureteral/graft operations 9 Graft operations / ureters observed
Table I l L Management of postoperative ureteral complications in 44 patients Management
24 Patientsgraft operations
31
Observation, initially Improved or unchanged Progressive obstruction requiring operation Operation, initially Primary urologic indication Associated with graft revision
No. patients
Ureteral entrapment
28 24 4
0
16 8 8
3 3
Operative mortality 6/29 = 21%.
29 patients (21%) died after operation, all of graft complications, including three aortoenteric fistulas, two ruptured anastomotic aneurysms, and one graft in£f~fion. URETERAL COMPLICATIONS Preoperative ureteral complications. Six patients with preoperative ureteral obstruction included two who were anuric, one with pyohydronephrosis, and three with asymptomatic hydronephrosis. All had evidence of retroperitoneal inflammation at the time of their aortic reconstruction manifest by localized edema, hyperemia, and thickened retroperi~oneal tissues. One patient had an inflammatory abdominal aortic aneurysm, and another had an infected iliac artery aneurysm associated with an aortic aneurysm. One patient underwent nephroureterectomy 1 month before aortic aneurysm repair for pyohydrofiephrosis caused by ureteral compression by the aneu ,,'sm. In four of the remaining five, ureteral operations were performed during aneurysm repair, including two ureterolyses, one segmental ureterectomy and reanastomosis, and one cutaneous ureterostomy. One patient underwent aneurysm repair with ureteral stenting alone. Five of the six required no further urologic operations. The patient who underwent a cutaneous ureterostomy during repair of an aortic and infected iliac artery aneurysm required nephrectomy 1 month after operation for unrelenting urosepsis. All patients had normal postoperative renal fimction except one of the two who were anuric before operation. Postoperative ureteral complications. Fortyfour patients had 46 ureteral complications after aortoiliac reconstruction. Twenty-eight patients, all with asymptomatic hydronephrosis, were initially ob~ . v e d (Table III). In 24 of these, the obstruction
either did not progress in 10, or it resolved spontaneously in 14. Operation for progressive ureteral obstruction was required in the remaining four patients (14%). Three underwent ureterolysis and the fourth underwent segmental ureteral resection and reanastomosis. None of these patients had ureteral entrapment between the graft and host iliac artery. Sixteen patients, eight with no symptoms and eight with symptoms, had 18 ureteral complications and were operated upon when the complication was first recognized (Table III). Eight patients, three of whom were found to have the ureter lying behind the prosthetic graft, underwent ureteral procedures during graft revision, including four ureterolyses and four ureteral resections and reanastomoses. Those with entrapped ureters all underwent ureteral division, transposition, and reanastomosis. The other eight patients underwent urologic operations alone for hydronephrosis, with partial ureteral obstruction in four, hydronephrosis with ureteral leak in three, and an obstructed, nonfimctioning kidney in one. Three of these had ureteral entrapment between the graft and iliac artery. Ureterolysis was performed in four patients, including one with an entrapped ureter who also underwent graft division, ureteral transposition, and graft reanastomosis. Ureteral division and reanastomosis were performed in two patients including one with an entrapped ureter, and nephroureterectomies were performed in two including one with an entrapped ureter. Therefore six of these 16 (38%) patients and six of all 20 (30%) eventually requiring ureteral operations were found to have the ureter lying between the prosthetic graft and host iliac artery (Table III). All patients undergoing ureteral reconstruction had documented improvement by follow-up urographic studies. No patient experienced chronic renal failure. Among 24 of the 44 patients (55%) the ureteral complications were recognized within a year of aortic
32
Journal of VASCULAR SURGERY
Wright et al.
Table IV. Graft complications in 3580 patients undergoing aortoiliac reconstruction with and without postoperative ureteral complications
Table V. Postoperative ureteral complications vs graft complications Graft complication
Complication Patients ~ 44 Patients with postoperative ureteral complications Anastomotic aneurysm Graft thrombosis Graft infection Aortoenteric fistula 3536 Patients without postoperative ureteral complications Anastomotic aneurysm Graft thrombosis Graft infection Aortoenteric fistula
19 8 6 3
16 7 6 3
395 218 25 19
270 202 25 19
~Some patients had more than one graft complication.
reconstruction. Nine (38%) of these required ureteral reconstructions, three of which were performed during aortic graft revision. In 20 patients (45%) the ureteral complications were recognized more than 1 year after aortic reconstruction. Eleven (55%) required ureteral reconstructive procedures, five during aortic graft revisions. These differences were not statistically significant. Graft complications Twenty-four of the 44 patients (55%) with postoperative ureteral complications experienced 36 graft complications, an average of 3.3 years after operation, ranging from 1 month to 12.8 years. Some patients had more than one graft complication. Nineteen anastomotic aneurysms developed in 16 patients, three of whom had two anastomotic aneurysms (Table IV). Eight graft limb thromboses occurred in seven patients. Graft infections developed in six patients, and aortoenteric fistulas developed in three. There were six postoperative deaths in these 24 patients resulting from the graft complications. Graft complications developed in 12 patients with early ureteral complications. In two of these patients, both with anastomotic aneurysms, the ureteral complication was recognized during management of the anastomotic aneurysm. In the remaining 10, ureteral complications were identified before graft complications were noted. Similarly, graft complications developed among 12 patients with late ureteral complications. The ureteral complication preceded the diagnosis of graft complication in five of these patients and was recognized during management of the graft problem in the other seven.
Ureteral complication Yes No Total
Yes 24 438 462
No 20 3098 3118
44 3536 3580
Relative Risk = 4.4; 95% Confidence Interval = (3.3, 5.8); p = <0.0001.
Four hundred thirty-eight (12%) of the 3536 patients undergoing aortoiliac reconstruction who did not have recognized postoperative ureteral complications had 657 graft complications (Table IV). Included were 395 anastomotic aneurysms in 27Q patients, 218 graft thromboses in 202 patients, 25 graft infections in 25 patients, and 19 aortoenteric fistulas in 19 patients. Comparing the development of graft com~'cations among patients with and without postoperative ureteral complications, a patient was 4.4 times more likely to have a graft complication associated with the ureteral complication than without a ureteral complication (p = <0.0001; 95% CI = 3.3,5.8) (Table V). Graft complications developed in patients with postoperative uretcral complications significantly earlier than patients without postoperative ureteral problems (p < 0.0001) (Fig. 1). This difference in graft complication rates bet tween patients with and without postoperative ureteral complications persisted even after adjusting for demographic and follow-up variables (p < 0.0001). Cox regression analysis suggested that patients with postoperative ureteral complications were 3.2 times as likely to have graft complications as patients with~out postoperative ureteral complications ( ~ % CI = 2.0,5.0). DISCUSSION
Postoperative hydronephrosis developing within 1 year of aortic reconstruction has been reported i~1 up to 14% of patients prospectively studied? But the precise incidence of ureteral complications after aortic reconstruction is unknown, because many obstructive ureteral complications may be asymptomatic; and consequently, diagnostic studies are not performed to confirm either early or late problems. The overall 1.4% incidence of ureteral complications in the retrospective study reported herein also represents an underestimation of this problem, since excretory pyelograms were not routinely obtained, and only 55% (1958/3580) of patients underwent post-
Volume 11 Number 1 January 1990
Ureteral complications and aortoiliac reconstruction 33
1.0
0.9 ul
o o ¢L
E o 0
0.8
0.7 0.6 (n=18)
(n=lO) I t __f --
Ureteral Complications
]
o.s I 0.4
=.IQ
F'-
0.3
NoUreteral Complications
(n=207)
,,Q
I I
0.1
0.0 0
I
I
I
|
I
I
I
5
10
15
20
25
30
35
Time in Years p
Fig. 1. Kaplan-Meier curves demonstrate significantlyincreased probability of graft complications developing among patients with ttreteral complications (broken line) than those without ureteral complications (solid line). Note also that graft complications developed significantly earlier in patients with ureteral complications than in those without ureteral complications. operative aortography. Only long-term prospective fbllow-up with either contrast urography or ultrasonography in all patients undergoing aortic reconstruction, a study which seems to be beyond practicality, will define the true incidence of this complication. Causes of ureteral complications associated with ~iortic reconstruction vary. Large aneurysms, especia~,~, those with an iliac component, have been reported to cause uretcral obstruction before operation and in rare cases even lead to arterioureteral f i s t u l a s . 9 ,Under these circumstances inflammation and fibrosis may be major components of the obstruction. " Since only six patients with preoperative ureteral obstruction were identified in this study, definitive conclusions cannot be drawn. However, one finding that appeared common to both patients with preoperative hydronephrosis and those developing it after operation, was the high incidence of associated graft complications. Precautions should be taken in patients with preoperative ureteral obstruction to prevent possible infectious complications of urologic origin. These patients might benefit from preoperative ureteral stent placement to provide decompression of the hydronephrosis with improvement in re-
nal function as well as to facilitate ureteral identification to prevent injury during aortic reconstruction. Management has included ureterolysis, ureteral resection and reanastomosis, or aortic reconstruction alone. Most authors recommend ureterolysis in conjunction with aneurysm repair as was successfully performed in one third of the patients in this study. Postoperative ureteral obstruction, first described by Jacobson et al.,10 has usually been documented by anecdotal case reports or limited sized series. Since Shaw and Baue n first suggested a relationship between graft and ureteral complications, Schubert et al.5 have analyzed a group of 19 patients with postoperative hydronephrosis and the implications of this complication. They reported an extraordinarily high 89% incidence of graft complications associated with hydronephrosis. Graft complications occurred in 55% of patients in this study. Anastomotic aneurysms, graft thromboses, graft infection, and aortoenteric fistulas were significantly more common in patients with postoperative hydronephrosis than those without hydronephrosis (p < 0.0001). Such graft complications also appeared earlier among patients with ureteral complications than those without (p < 0.0001) (Fig.
34
Journal of VASCULAR SURGERY
Wright et al.
1). These findings suggest that postoperative hydronephrosis may signal a graft complication, although an etiologic association is unproven. Initial reports of postoperative ureteral obstruction implicated placement of the graft anterior to the ureter producing ureteral entrapment between the graft and host lilac vessels. 12,13However, subsequent reports have noted entrapment in less than half of the cases of postoperative ureteral obstruction. 14,1s Only six (30%) of the 20 patients undergoing ureteral reconstructive procedures in the present study were found to have ureters lying between the graft and host iliac vessels. These findings suggest that other mechanisms in addition to entrapment are responsible for ureteral obstruction. Prosthetic graft complications may also incite ureteral complications, which in turn may signal an underlying graft complication. 16qs Anastomotic aneurysms may obstruct or rupture into the ureter32,16-1s The inflammatory process accompanying graft infection may obstruct the ureter, which in turn may be a marker for such graft complications. Nine of our patients had synchronous presentation of graft and ureteral complications including eight anastomotic aneurysms and one graft infection. Aside from the obvious causes ofureteral obstruction as a resuk of graft complications, another cause is the retroperitoneal reaction that appears to be stimulated by the implanted graft. ~s,19-22This fibrotic process has been found to be either localized to the site of graft-ureteral contact or as a generalized retroperitoneal process. Although it is attractive to postulate that the graft itself may- initiate this process, it is most likely a combination of retroperitoneal dissection, hematoma resolution, and host response to the prosthesis that produces this fibroproliferative reaction. This analysis confirmed previous reports that suggested that ureteral complications make subsequent graft revisions technically more challenging than those not perturbed by ureteral involvement,s,l° Eight patients required ureteral reconstruction in conjunction with graft revision. In most instances, the intense fibrotic reaction enveloping the ureter made dissection of the graft technically demanding. In four patients ureterolysis provided sufficient exposure of the graft, but in four others, division, intentionally or unintentionally, and reanastomosis of the ureter were required to correct the graft complication. Ureteral involvement associated with a graft complication was identified in only 5% of all graft com-
plications in the entire 33-year experience encompassing 3580 patients (Table V). This small but important association supports the recommendation tq investigate for concomitant ureteral involvement before graft revision. If ureteral involvement is documented, precautions should be taken to avoid ureteral injury during graft revision such as insertion of stents to aid in ureteral identificationY This study also emphasizes that isolated ureteral operations after aortic reconstruction are rarely required, since only 12 patients (0.3%) required such procedures. Hydronephrosis after aortic reconstruction will often spontaneously resolve. However, since it may progress, routine follow-up evaluation by periodic IVP or ultrasonography is suggested to doc= ument resolution or progression. With progression operative correction should be considered. Although there have been anecdotal reports of improvement after steroid therapy, z* operative interventig,q is required when obstruction compromises renal" function. When operation is required for primary ureteral involvement, preoperative selective ureteral catheterization may prove valuable just as when managing combined graft and ureteral complications. Urine bacterial culture data obtained from the obstructed kidney guide organism-specific antimicrobial therapy. Since intense retroperitoneal inflammation encountered during ureteral repair may result in urine extravasation from ureteral injury with possible graft contamination, organism-specific antibiotic therapy is recommended to minimize development of septic complications. REFERENCES
1. Heard G, Hinde G. Hydronephrosis complicating ao~ic reconstruction. Br J Surg 1975;62:344-7. 2. Frusha JD, Porter JA, Batson PC. Hydronephrosis following aortofemoral bypass grafts. J Cardiovasc Snrg 1982;23: 371-7. 3. Goldenberg SL, Gordon PB, Cooperberg PL, McLonghlin MG. Early hydronephrosis following aortic bifurcation graft surgery. J Urol 1988;140:1367-9. 4. Egeblad K, Brochner-Mortensen J, Krarup T, Holstein PE, Bartholdy NJ. Incidence of ureteral obstruction after aortic grafting: a prospective analysis. Surgery 1988;103:411-4. 5. Schubert P, Former G, Cummings D, et al. The significance of hydronephrosis after aortofemoral reconstruction. Arch, Surg 1985;120:377-81. 6. Szilagyi DE, Elliott JP Jr, Smith RE, et al. A thirty-yearsurve~ of the reconstructive surgical treatment of aortoiliac occlusive disease. J VAsc SURG 1986;3:421-36. 7. Lawless JF. Statistical models and methods for lifetime data. New York: John Wiley & Sons Inc, 1982. 8. Cox DR. Regression models and life tables (with discussion) J R Statist Soc B 1972;34:187-220.
Volume 11 Number 1 January 1990
Ureteral complications and aortoiliac reconstruction 35
9. Wheadey JK, Ansley JD, Trulock TS, Campbell D. Ureteroarterial fistula. Urology 1981;13:498-502. 10. Jacobs0n ME, Mastio GJ, Berkas EM. Ureteral obstruction as a late complication of abdominal aneurysm resection. J Kansas Med Soc 1962;63:516-8. 11. Shaw RS, Bane AE. Management of sepsis complicating arterial reconstructive surgery. Surgery 1963;53:75-86. 12. Lytton B. Ureteral obstruction following aortofemoral bypass grafts. Surgery 1966;59:918-22. 13. Thomford NR, Dorfman LF. Ureteral obstruction caused by an aortofemoral bypass prosthesis. Am J Surg 1968;115: 394-6. 14. Antkowiak JG, D'Altorio RA. Ureteral obstruction secondary to bifurcated aortic grafts. Arch Surg 1979;114:853-6. 15. Sant GR, Heaney JA, Parkhurst EC, Blaivas JG. Obstructive uropathy-A potentially serious complication of reconstructive vascular surgery. J Uro! 1983;129:16-22. ~6. Futrell JW, Kaczala O, Wolkoff JS. Obstructive uropathy secondary to a false aneurysm 12 years after abdominal aneurysmectomy. South Med J 1975;68:607-10. 17. Carmichael DH, Barnes WF. Obstruction of the ureter due to a false aneurysm. Surgery 1979;86:769-71.
18. Joost J, Bartsch G, Weimann S, Neurer G. Ilioureteric fistula. Br J Urol 1981;53:477. 19. HenryLG, BemhardVM. Ureteral pathology associated with aortic surgery: a report of three unusual cases. Surgery 1978;83:464-9. 20. Fourcroy JL, Azoury B, Miller HC. Bilateral ureteral obstruction as a complication of vascular graft surgery. Urology 1980;15:556-8. 21. Rhind JR. Retroperitoneal fibrosis following aortic surgery. Br J Urol 1977;49:552. 22. Wallijn E, Renders G, Vereecken L. Urologic complications following aortofemoral bypass graft. Br J Uro11975 ;47:61721. 23. McCarthy WJ, Flinn WR, Carter MF, Yao JST. Prevention and management of ureteral injuries during aortic surgery. In: Bergan J, Yao J, eds. Aortic surgery. 1st ed. Philadelphia: WB Saunders, 1989. 24. Huben RP, ScheUhammer PF. Steroid therapy for ureteral obstruction after aortoiliac graft surgery. J Urol 1981; 125:881-3.
DISCUSSION Dr. K. Wayne Johnston (Toronto, Canada). This ar-
eration is necessary only for those patients who have further complications. This article clarifies the disease associated with delayed onset hydronephrosis, which occurred in 1.2%. In half the cases, delayed nreteric obstruction was due to a dense local fibrotic reaction, the cause of which is unknown. The other half were associated with graft complications. It is logical to expect a relationship between the causes of hydronephrosis and graft infection or an aortoenteric fistula. However, why should there be an association with thrombosis of the limb of the graft or a false aneurysm in the groin? Would the authors recommend that all patients have urologic screening by ultrasonography, CT scanning, or IVP 1 year after aortic surgery to detect those patients with persistent hydronephrosis and half those with delayed ureteric compfications? Should this screening apply only to patients with an aortobifemoral or aortobiiliac graft and not to those with a tube graft? In conclusion, this article has documented that patients with ureteric complications should be investigated for associated graft complications, and conversely, that patients with aortic graft complications should have a fifll urologic workup, since 5% will have an abnormality; I agree. Dr. Calvin B. Ernst. We thank Dr. Johnston for his perceptive review of the manuscript and his in-depth comments. I will try to answer his questions in the order asked. First, you vcanted to know whether ureteral stenting facilitates repair of an inflammatory aneurysm assuming that after operation the retroperitoneal involvement will resolve and thereby relieve the hydronephrosis. Since we
ticle presents the largest detailed analysis of ureteric complications associated with abdominal aortic surgery. First, the authors reported six patients with ureteric obstruction occurring with an abdominal aortic aneurysm. The usual management of ureteric obstruction associated with an inflammatory aneurysm is ureterolysis and aneurysm repair; however, this procedure may be difficult, and there is the risk of ureteric injury and subsequent graft infection. I ask the authors to comment on two other approaches. D o -~ou feel that it is reasonable to simply repair the aneurysm, use temporary indwelling ureteric stents if necessary, and anticipate that the hydronephrosis will resolve spontaneously? I am reluctant to mention the use of steroids, but as documented by others and by three patients who were &herwise inoperable and being treated on our service with steroids, hydronephrosis and the inflammatory component o f the aneurysm may resolve with steroid administration. Would you consider using steroids before surgery to reduce the inflammatory component and ureteric obstruction, especially in a patient with a small aortic aneurysm? Second, the authors described 44 patients with ureteric complications after aortic surgery. There are three clinical types. Early transient hydronephrosis occurs in 10% to 15% of aortic cases because of edema from the dissection and, as reported, is of no clinical significance since it usually resolves. Early persistent hydronephrosis is uncommon, occurring in approximately 1% to 2% of aortic cases. Op-
36
Journal of VASCULAR SURGERY
Wright et al.
had only six patients with preoperative ureteral involvement, only one of whom had an inflammatory aneurysm, we cannot make definitive recommendations. Nonetheless, when encountering an inflammatory abdominal aortic aneurysm with ureteral obstruction, we feel it is important to precisely identify the ureters, and ureteral stenting helps. We place ureteral stents and assume that the periureteral reaction will subside, and the ureteral obstruction will spontaneously resolve after operation. That this invariably occurs is documented by reports in the literature as well as our one patient. The corollary, of course, as you have asked--what is the role of steroids? We do not recommend using steroids in this elderly patient population for obvious reasons. I might also point out that steroids will be most useful, if they are at all, in the early postoperative period when the inflammatory phase of the retroperitoneal reaction is at its maximum. Steroids would probably have no ameliorating effect on the retroperitoneal fibroproliferative reaction, which develops late after operation. You asked about the relationship between aortofemoral bypass complications and ureteral problems. We identified 27 patients who underwent aortobifemoral bypass reconstruction and 19 with aortoiliac reconstruction. Although it is attractive to postulate a relationship, ancL it may well be that there is a similar pathophysiologic defect in healing underlying the retroperitoneal inflammatory reaction that causes these patients to have anastomotic aneurysms, because this was a retrospective analysis we were unable to detect a cause and effect relationship, except that the ureteral complication may serve as a marker for a graft complication. Finally, you asked about urologic screening. Clearly, if one has a graft problem or a symptomatic ureteral problem, these patients should be very thoroughly studied. But we are not entirely sure what the cost/benefit ratio would be of routine screening with ultrasonography, although we certainly would not recommend against it. Knowing the litigious climate in which we live, I am reluctant to recommend a policy of routine postoperative screening based on our experience and this retrospective analysis. The only way to determine the true incidence and natural history of this problem is to prospectively screen all patients, a study which seems to us at least to be beyond practicality. Dr. Ronald J. Stoney (San Francisco, CaliE). My comments and question concern the later ureteral complication leading to hydronephrosis. I would like to make one comment about actually creating the retroureteral tunnel for the placement of graft limbs, since I think that occasionally blind clamping and withdrawing, or drawing the graft limbs through or enlarging the space to place a graft limb probabl3? produced unrecognized injury to the ureter. At least it has been apparent to use in seeing some patients referred with graft complications, some of which were alluded to, that a primary ureteral injury was produced by this mechanism. I think there are ways with opposing fingers to create a tunnel in the retroperitoneum and avoid this and produce a smooth bed for the graft limb to occupy.
Second, in operations or reoperations on the aorta and its branches for a variety of reasons but primarily occlusion and infection, we have had the occasion to remove aortic grafts placed many years earlier, often for infection, and f am surprised to see how few times there is anything more than the dense fibrous encasement of the graft that extends to the ureter. By that I mean the ureter is not really dilated or enlarged, and I was thinking as I came up here that I know personally of only one patient in our own series in the last 2 decades of approaching 200 aortic graft infections, who has had a proven ureteric complication leading to the graft infection. There were others who had the problem of ureteric obstruction, which was brought out by this article today, but to actually have a ureteral contamination or urine contaminating the graft and producing the actual infection or source of infection. I can only recall at least one.
I would like to ask the authors really one question and that is, what do they think the main cause is of the delayed adhesive type reaction between grafts and ureter that leads to this problem and many also be a marker of l a t ~ r a f t infection? Dr. Ernst. Since about a third of our patients ha,d ureteral entrapment, precise placement of the graft was not always accomplished. We quite agree with Dr. Stoney that precise dissection is necessary in making retroperitoneal ttlnnels under direct vision insofar as possible. However, I caution the audience that extensive dissection is contraindicated. We had one patient who had what we thought was overly extensive dissection where the ureter was divascularized and necrosed, ultimately resulting in nephroureterectomy. I think Dr. Stoney is exactly right relative to ureteral complications causing subsequent graft infection, this rarely occurs. However, in two of our patients the ureteral complications were directly related to a subsequent graft infection. I might add, parenthetically, that the retropetitoneal reaction under these particular circumstances creates major technical challenges when managing the graft(-,~mplication. Finally, Dr. Stoney you asked what causes the retroperitoneal reaction? This is an important question, and we really cannot tell you from our analysis. It may result from,. an abnormal fibroprolife'rative reaction in these patients, a defect in healing if you will, which ultimately results iia anastomotic defects and graft incorporation problems as well. Dr. Walter J. McCarthy (Chicago, Ill.). At Northwestern University over the past few years we have attempted to identify patients whose ureters might be adherent in areas where redissection would be necessary du~ing reoperation. In these cases stents have been placed preoperatively to aid in the intraoperative identification of the ureters. Stents have been used to identify ureters in cases where infected grafts needed to be removed, during reoperations for occluded aortofemoral grafts, in cases o£ inflanmaatory aneurysms, and for repair of anastomotic false aneurysms at the iliac bifurcation. We have been ~,~':~-
Volume 11 Number 1 January 1990
eraUy very pleased with this technique, which simplifies and hastens the operation, and we wonder if the authors .have had any experience with this approach. Dr. Ernst. Yes, I think when one is managing graft complications that ureteral stenting is very helpful. We are well aware of the excellent report from the Northwestern group regarding these particular complications. I might also add that it is helpful to make the diagnosis ofconcom-
Ureteral complications and aortoiliac reconstruction 37
kant ureteral involvement before operation. It is difficult, at best, to obtain simultaneous excretory urography and aortography to determine whether ureteral entrapment is present. But with constrast enhanced CT scanning the ureters stand out as little white dots, and if they are obstructed they will be big white dots. Therefore from CT scanning one can determine the relationship between the ureter and the limb of the prosthesis and facilitate reoperation.
have little bearing on ultimate morbidity or mortality after aortic reconstruction. Recently, hydronephrosis caused by ureteral obstruction after aortic reconstruction has been suggested to be associated with graft complications. S To further characterize this relationship and determine the natural history ofureteral complications after aortoiliac reconstruction, a 33-year experience with aortoiliac reconstruction was retrospectively reviewed. The relationships between ureteral and graft complications and outcome of specific management o f these complications were analyzed and form the basis of this article. CLINICAL MATERIAL
.From the Division of Vascular Surgery, Henry Ford Hospital. Presented at the Thirty-seventh ScientificMeeting of the No~.h American Chapter, International Society for Cardiovascular Surgery, New York, N.Y., June 19-20, 1989. Reprint requests: Calvin B. Ernst, MD, Division of Vascular Surgery, Henry Ford Hospital, 2799 W. Grand Blvd., Detroit, MI 48202. 24/6/16662
Beginning in 1954, data have been recorded a n d tracked for patients undergoing vascular surgical procedures at the Henry Ford Hospital by means of a computerized vascular registry. Through periodic patient recall and by following established clinical examination and arteriographic protocols, such data 29
30
Journal of VASCULAR SURGERY
Wright et al.
Table I. Aortic reconstruction in 50 patients with ureteral complications Aneurysmal Occlusive disease disease Total Aortobifemoral reconstruction Aortobiliac reconstruction Other* Total
5 18 1 24
22 1 3 26
27 19 4 50
*One mycotic aneurysm figation; three aortoiliac endarterectomies.
have been regularly updated. 6 Information derived from more than 13,000 vascular surgical procedures have been entered into the registry, and from this database, 3580 patients undergoing 3660 aortoiliac reconstructive procedures were identified. Fifty of the 3580 patients undergoing aortoiliac reconstruction were found to have significant ureteral complications, and their records were scrutinized. Ureteral involvement was documented by urography during aortography, intravenous pyelography (IVP), retrograde pyelography, or operation. Patients were followed by a protocol that included yearly examination. During the first 2 decades of the study, routine translumbar aortography was performed in the immediate postoperative period before hospital discharge in 1958 patients and was repeated at 2- to 3year intervals. All 50 patients were followed through 1988 or until death. Follow-up averaged 7.4 years and ranged from 1 month to 20.3 years. The incidence and type of graft complications were assessed, and the outcome of both the graft and ureteral complication was determined in each patient. The incidence of graft complications was also determined for all 3580 patients entered into the database during the period of this study. Statistical methods. Statistical analysis focused on patients with postoperative ureteral complications. Patients with preoperative ureteral complications were counted as patients without postoperative complications. To assess the association between postoperative ureteral and graft complications, each patient was included as having any type of graft complication or any ureteral complication or both. Thus each patient was counted only once regardless of the number of procedures performed or complications identified. A relative risk with a 95% confidence interval was computed. Log-rank analysis and Kaplan-Meier curves, 7 which adjust for varying lengths of follow-up, were used to compare graft complication rates between patients with (n = 44) and without (n = 3536)
postoperative ureteral complications. In this analysis only graft complications occurring after the first operation were used. The length of follow-up was cal-~ culated as the time from first operation to graft complication, date of last examination, or death. Cox proportional hazards regression analysis 8 was used to adjust for age, sex, race, and length of followup and to compare graft complication rates for those patients with and without postoperative ureteral complications. Again, only the first operation was used in this analysis. Risk estimates and confidence intervals were calculated by use of the results from Cox regression. A p value of less than 0.05 was considered statistically significant. RESULTS
Fifty-eight ureteral complications were identified in 50 patients. There were 29 men and 21 w ~ e n with an average age of 61.3 years, ranging from 42 to 78 years. The overall incidence of ureteral complications was 1.4% (50/3580). Twenty-six patients underwent aortoiliac reconstruction for occlusive disease, and 24 patients underwent reconstruction for aneurysmal disease (Table I). Six patients had preoperative ureteral obstruction, graft complications developed in two. The remaining 44 patients had 46 postoperative ureteral complications manifest by hydronephrosis in 42, ureteral leak in three, and ureterat necrosis in one. Ureteral complications were identiffed from 2 weeks to 12 years postoperatively (average, 18.9 months). Early ureteral complications were arbitrarily defined as those detected within 1 year of aortic reconstruction, and late complications were defined as those detected beyond 1 year. Twenty-four patients (55%) had early ureteral c~mplications, and 20 (45%) had late complications discovered from 14 months to 12 years after operation. Thirty-six graft complications developed in 24 of the 44 patients. Graft complications were identifiec~ from 4 weeks to 12.8 years after surgery (average, 3.3 years). Graft complications were evenly divided between the early ureteral complication group and the late ureteral complication group. Twenty-nine of the 44 patients with postoperative ureteral complications underwent ureteral or graft operations or both. These included five patients undergoing ureteral operations alone, seven undergoing a ureteral procedure and subsequent graft operation, eight undergoing simultaneous ureteral and graft procedures, and nine undergoing graft operation and ureteral observation (Table II). SIX of tb_.7
Volume 11 Number 1 January 1990
Ureteral complications and aortoiliac reconstruction
Table II. Patients undergoing graft and/or ureteral operations (n = 29)
20 Patientsureteral operations
5 Ureteral operations alone 7 Ureteral operations with subsequent graft operataon 8 Simultaneous ureteral/graft operations 9 Graft operations / ureters observed
Table I l L Management of postoperative ureteral complications in 44 patients Management
24 Patientsgraft operations
31
Observation, initially Improved or unchanged Progressive obstruction requiring operation Operation, initially Primary urologic indication Associated with graft revision
No. patients
Ureteral entrapment
28 24 4
0
16 8 8
3 3
Operative mortality 6/29 = 21%.
29 patients (21%) died after operation, all of graft complications, including three aortoenteric fistulas, two ruptured anastomotic aneurysms, and one graft in£f~fion. URETERAL COMPLICATIONS Preoperative ureteral complications. Six patients with preoperative ureteral obstruction included two who were anuric, one with pyohydronephrosis, and three with asymptomatic hydronephrosis. All had evidence of retroperitoneal inflammation at the time of their aortic reconstruction manifest by localized edema, hyperemia, and thickened retroperi~oneal tissues. One patient had an inflammatory abdominal aortic aneurysm, and another had an infected iliac artery aneurysm associated with an aortic aneurysm. One patient underwent nephroureterectomy 1 month before aortic aneurysm repair for pyohydrofiephrosis caused by ureteral compression by the aneu ,,'sm. In four of the remaining five, ureteral operations were performed during aneurysm repair, including two ureterolyses, one segmental ureterectomy and reanastomosis, and one cutaneous ureterostomy. One patient underwent aneurysm repair with ureteral stenting alone. Five of the six required no further urologic operations. The patient who underwent a cutaneous ureterostomy during repair of an aortic and infected iliac artery aneurysm required nephrectomy 1 month after operation for unrelenting urosepsis. All patients had normal postoperative renal fimction except one of the two who were anuric before operation. Postoperative ureteral complications. Fortyfour patients had 46 ureteral complications after aortoiliac reconstruction. Twenty-eight patients, all with asymptomatic hydronephrosis, were initially ob~ . v e d (Table III). In 24 of these, the obstruction
either did not progress in 10, or it resolved spontaneously in 14. Operation for progressive ureteral obstruction was required in the remaining four patients (14%). Three underwent ureterolysis and the fourth underwent segmental ureteral resection and reanastomosis. None of these patients had ureteral entrapment between the graft and host iliac artery. Sixteen patients, eight with no symptoms and eight with symptoms, had 18 ureteral complications and were operated upon when the complication was first recognized (Table III). Eight patients, three of whom were found to have the ureter lying behind the prosthetic graft, underwent ureteral procedures during graft revision, including four ureterolyses and four ureteral resections and reanastomoses. Those with entrapped ureters all underwent ureteral division, transposition, and reanastomosis. The other eight patients underwent urologic operations alone for hydronephrosis, with partial ureteral obstruction in four, hydronephrosis with ureteral leak in three, and an obstructed, nonfimctioning kidney in one. Three of these had ureteral entrapment between the graft and iliac artery. Ureterolysis was performed in four patients, including one with an entrapped ureter who also underwent graft division, ureteral transposition, and graft reanastomosis. Ureteral division and reanastomosis were performed in two patients including one with an entrapped ureter, and nephroureterectomies were performed in two including one with an entrapped ureter. Therefore six of these 16 (38%) patients and six of all 20 (30%) eventually requiring ureteral operations were found to have the ureter lying between the prosthetic graft and host iliac artery (Table III). All patients undergoing ureteral reconstruction had documented improvement by follow-up urographic studies. No patient experienced chronic renal failure. Among 24 of the 44 patients (55%) the ureteral complications were recognized within a year of aortic
32
Journal of VASCULAR SURGERY
Wright et al.
Table IV. Graft complications in 3580 patients undergoing aortoiliac reconstruction with and without postoperative ureteral complications
Table V. Postoperative ureteral complications vs graft complications Graft complication
Complication Patients ~ 44 Patients with postoperative ureteral complications Anastomotic aneurysm Graft thrombosis Graft infection Aortoenteric fistula 3536 Patients without postoperative ureteral complications Anastomotic aneurysm Graft thrombosis Graft infection Aortoenteric fistula
19 8 6 3
16 7 6 3
395 218 25 19
270 202 25 19
~Some patients had more than one graft complication.
reconstruction. Nine (38%) of these required ureteral reconstructions, three of which were performed during aortic graft revision. In 20 patients (45%) the ureteral complications were recognized more than 1 year after aortic reconstruction. Eleven (55%) required ureteral reconstructive procedures, five during aortic graft revisions. These differences were not statistically significant. Graft complications Twenty-four of the 44 patients (55%) with postoperative ureteral complications experienced 36 graft complications, an average of 3.3 years after operation, ranging from 1 month to 12.8 years. Some patients had more than one graft complication. Nineteen anastomotic aneurysms developed in 16 patients, three of whom had two anastomotic aneurysms (Table IV). Eight graft limb thromboses occurred in seven patients. Graft infections developed in six patients, and aortoenteric fistulas developed in three. There were six postoperative deaths in these 24 patients resulting from the graft complications. Graft complications developed in 12 patients with early ureteral complications. In two of these patients, both with anastomotic aneurysms, the ureteral complication was recognized during management of the anastomotic aneurysm. In the remaining 10, ureteral complications were identified before graft complications were noted. Similarly, graft complications developed among 12 patients with late ureteral complications. The ureteral complication preceded the diagnosis of graft complication in five of these patients and was recognized during management of the graft problem in the other seven.
Ureteral complication Yes No Total
Yes 24 438 462
No 20 3098 3118
44 3536 3580
Relative Risk = 4.4; 95% Confidence Interval = (3.3, 5.8); p = <0.0001.
Four hundred thirty-eight (12%) of the 3536 patients undergoing aortoiliac reconstruction who did not have recognized postoperative ureteral complications had 657 graft complications (Table IV). Included were 395 anastomotic aneurysms in 27Q patients, 218 graft thromboses in 202 patients, 25 graft infections in 25 patients, and 19 aortoenteric fistulas in 19 patients. Comparing the development of graft com~'cations among patients with and without postoperative ureteral complications, a patient was 4.4 times more likely to have a graft complication associated with the ureteral complication than without a ureteral complication (p = <0.0001; 95% CI = 3.3,5.8) (Table V). Graft complications developed in patients with postoperative uretcral complications significantly earlier than patients without postoperative ureteral problems (p < 0.0001) (Fig. 1). This difference in graft complication rates bet tween patients with and without postoperative ureteral complications persisted even after adjusting for demographic and follow-up variables (p < 0.0001). Cox regression analysis suggested that patients with postoperative ureteral complications were 3.2 times as likely to have graft complications as patients with~out postoperative ureteral complications ( ~ % CI = 2.0,5.0). DISCUSSION
Postoperative hydronephrosis developing within 1 year of aortic reconstruction has been reported i~1 up to 14% of patients prospectively studied? But the precise incidence of ureteral complications after aortic reconstruction is unknown, because many obstructive ureteral complications may be asymptomatic; and consequently, diagnostic studies are not performed to confirm either early or late problems. The overall 1.4% incidence of ureteral complications in the retrospective study reported herein also represents an underestimation of this problem, since excretory pyelograms were not routinely obtained, and only 55% (1958/3580) of patients underwent post-
Volume 11 Number 1 January 1990
Ureteral complications and aortoiliac reconstruction 33
1.0
0.9 ul
o o ¢L
E o 0
0.8
0.7 0.6 (n=18)
(n=lO) I t __f --
Ureteral Complications
]
o.s I 0.4
=.IQ
F'-
0.3
NoUreteral Complications
(n=207)
,,Q
I I
0.1
0.0 0
I
I
I
|
I
I
I
5
10
15
20
25
30
35
Time in Years p
Fig. 1. Kaplan-Meier curves demonstrate significantlyincreased probability of graft complications developing among patients with ttreteral complications (broken line) than those without ureteral complications (solid line). Note also that graft complications developed significantly earlier in patients with ureteral complications than in those without ureteral complications. operative aortography. Only long-term prospective fbllow-up with either contrast urography or ultrasonography in all patients undergoing aortic reconstruction, a study which seems to be beyond practicality, will define the true incidence of this complication. Causes of ureteral complications associated with ~iortic reconstruction vary. Large aneurysms, especia~,~, those with an iliac component, have been reported to cause uretcral obstruction before operation and in rare cases even lead to arterioureteral f i s t u l a s . 9 ,Under these circumstances inflammation and fibrosis may be major components of the obstruction. " Since only six patients with preoperative ureteral obstruction were identified in this study, definitive conclusions cannot be drawn. However, one finding that appeared common to both patients with preoperative hydronephrosis and those developing it after operation, was the high incidence of associated graft complications. Precautions should be taken in patients with preoperative ureteral obstruction to prevent possible infectious complications of urologic origin. These patients might benefit from preoperative ureteral stent placement to provide decompression of the hydronephrosis with improvement in re-
nal function as well as to facilitate ureteral identification to prevent injury during aortic reconstruction. Management has included ureterolysis, ureteral resection and reanastomosis, or aortic reconstruction alone. Most authors recommend ureterolysis in conjunction with aneurysm repair as was successfully performed in one third of the patients in this study. Postoperative ureteral obstruction, first described by Jacobson et al.,10 has usually been documented by anecdotal case reports or limited sized series. Since Shaw and Baue n first suggested a relationship between graft and ureteral complications, Schubert et al.5 have analyzed a group of 19 patients with postoperative hydronephrosis and the implications of this complication. They reported an extraordinarily high 89% incidence of graft complications associated with hydronephrosis. Graft complications occurred in 55% of patients in this study. Anastomotic aneurysms, graft thromboses, graft infection, and aortoenteric fistulas were significantly more common in patients with postoperative hydronephrosis than those without hydronephrosis (p < 0.0001). Such graft complications also appeared earlier among patients with ureteral complications than those without (p < 0.0001) (Fig.
34
Journal of VASCULAR SURGERY
Wright et al.
1). These findings suggest that postoperative hydronephrosis may signal a graft complication, although an etiologic association is unproven. Initial reports of postoperative ureteral obstruction implicated placement of the graft anterior to the ureter producing ureteral entrapment between the graft and host lilac vessels. 12,13However, subsequent reports have noted entrapment in less than half of the cases of postoperative ureteral obstruction. 14,1s Only six (30%) of the 20 patients undergoing ureteral reconstructive procedures in the present study were found to have ureters lying between the graft and host iliac vessels. These findings suggest that other mechanisms in addition to entrapment are responsible for ureteral obstruction. Prosthetic graft complications may also incite ureteral complications, which in turn may signal an underlying graft complication. 16qs Anastomotic aneurysms may obstruct or rupture into the ureter32,16-1s The inflammatory process accompanying graft infection may obstruct the ureter, which in turn may be a marker for such graft complications. Nine of our patients had synchronous presentation of graft and ureteral complications including eight anastomotic aneurysms and one graft infection. Aside from the obvious causes ofureteral obstruction as a resuk of graft complications, another cause is the retroperitoneal reaction that appears to be stimulated by the implanted graft. ~s,19-22This fibrotic process has been found to be either localized to the site of graft-ureteral contact or as a generalized retroperitoneal process. Although it is attractive to postulate that the graft itself may- initiate this process, it is most likely a combination of retroperitoneal dissection, hematoma resolution, and host response to the prosthesis that produces this fibroproliferative reaction. This analysis confirmed previous reports that suggested that ureteral complications make subsequent graft revisions technically more challenging than those not perturbed by ureteral involvement,s,l° Eight patients required ureteral reconstruction in conjunction with graft revision. In most instances, the intense fibrotic reaction enveloping the ureter made dissection of the graft technically demanding. In four patients ureterolysis provided sufficient exposure of the graft, but in four others, division, intentionally or unintentionally, and reanastomosis of the ureter were required to correct the graft complication. Ureteral involvement associated with a graft complication was identified in only 5% of all graft com-
plications in the entire 33-year experience encompassing 3580 patients (Table V). This small but important association supports the recommendation tq investigate for concomitant ureteral involvement before graft revision. If ureteral involvement is documented, precautions should be taken to avoid ureteral injury during graft revision such as insertion of stents to aid in ureteral identificationY This study also emphasizes that isolated ureteral operations after aortic reconstruction are rarely required, since only 12 patients (0.3%) required such procedures. Hydronephrosis after aortic reconstruction will often spontaneously resolve. However, since it may progress, routine follow-up evaluation by periodic IVP or ultrasonography is suggested to doc= ument resolution or progression. With progression operative correction should be considered. Although there have been anecdotal reports of improvement after steroid therapy, z* operative interventig,q is required when obstruction compromises renal" function. When operation is required for primary ureteral involvement, preoperative selective ureteral catheterization may prove valuable just as when managing combined graft and ureteral complications. Urine bacterial culture data obtained from the obstructed kidney guide organism-specific antimicrobial therapy. Since intense retroperitoneal inflammation encountered during ureteral repair may result in urine extravasation from ureteral injury with possible graft contamination, organism-specific antibiotic therapy is recommended to minimize development of septic complications. REFERENCES
1. Heard G, Hinde G. Hydronephrosis complicating ao~ic reconstruction. Br J Surg 1975;62:344-7. 2. Frusha JD, Porter JA, Batson PC. Hydronephrosis following aortofemoral bypass grafts. J Cardiovasc Snrg 1982;23: 371-7. 3. Goldenberg SL, Gordon PB, Cooperberg PL, McLonghlin MG. Early hydronephrosis following aortic bifurcation graft surgery. J Urol 1988;140:1367-9. 4. Egeblad K, Brochner-Mortensen J, Krarup T, Holstein PE, Bartholdy NJ. Incidence of ureteral obstruction after aortic grafting: a prospective analysis. Surgery 1988;103:411-4. 5. Schubert P, Former G, Cummings D, et al. The significance of hydronephrosis after aortofemoral reconstruction. Arch, Surg 1985;120:377-81. 6. Szilagyi DE, Elliott JP Jr, Smith RE, et al. A thirty-yearsurve~ of the reconstructive surgical treatment of aortoiliac occlusive disease. J VAsc SURG 1986;3:421-36. 7. Lawless JF. Statistical models and methods for lifetime data. New York: John Wiley & Sons Inc, 1982. 8. Cox DR. Regression models and life tables (with discussion) J R Statist Soc B 1972;34:187-220.
Volume 11 Number 1 January 1990
Ureteral complications and aortoiliac reconstruction 35
9. Wheadey JK, Ansley JD, Trulock TS, Campbell D. Ureteroarterial fistula. Urology 1981;13:498-502. 10. Jacobs0n ME, Mastio GJ, Berkas EM. Ureteral obstruction as a late complication of abdominal aneurysm resection. J Kansas Med Soc 1962;63:516-8. 11. Shaw RS, Bane AE. Management of sepsis complicating arterial reconstructive surgery. Surgery 1963;53:75-86. 12. Lytton B. Ureteral obstruction following aortofemoral bypass grafts. Surgery 1966;59:918-22. 13. Thomford NR, Dorfman LF. Ureteral obstruction caused by an aortofemoral bypass prosthesis. Am J Surg 1968;115: 394-6. 14. Antkowiak JG, D'Altorio RA. Ureteral obstruction secondary to bifurcated aortic grafts. Arch Surg 1979;114:853-6. 15. Sant GR, Heaney JA, Parkhurst EC, Blaivas JG. Obstructive uropathy-A potentially serious complication of reconstructive vascular surgery. J Uro! 1983;129:16-22. ~6. Futrell JW, Kaczala O, Wolkoff JS. Obstructive uropathy secondary to a false aneurysm 12 years after abdominal aneurysmectomy. South Med J 1975;68:607-10. 17. Carmichael DH, Barnes WF. Obstruction of the ureter due to a false aneurysm. Surgery 1979;86:769-71.
18. Joost J, Bartsch G, Weimann S, Neurer G. Ilioureteric fistula. Br J Urol 1981;53:477. 19. HenryLG, BemhardVM. Ureteral pathology associated with aortic surgery: a report of three unusual cases. Surgery 1978;83:464-9. 20. Fourcroy JL, Azoury B, Miller HC. Bilateral ureteral obstruction as a complication of vascular graft surgery. Urology 1980;15:556-8. 21. Rhind JR. Retroperitoneal fibrosis following aortic surgery. Br J Urol 1977;49:552. 22. Wallijn E, Renders G, Vereecken L. Urologic complications following aortofemoral bypass graft. Br J Uro11975 ;47:61721. 23. McCarthy WJ, Flinn WR, Carter MF, Yao JST. Prevention and management of ureteral injuries during aortic surgery. In: Bergan J, Yao J, eds. Aortic surgery. 1st ed. Philadelphia: WB Saunders, 1989. 24. Huben RP, ScheUhammer PF. Steroid therapy for ureteral obstruction after aortoiliac graft surgery. J Urol 1981; 125:881-3.
DISCUSSION Dr. K. Wayne Johnston (Toronto, Canada). This ar-
eration is necessary only for those patients who have further complications. This article clarifies the disease associated with delayed onset hydronephrosis, which occurred in 1.2%. In half the cases, delayed nreteric obstruction was due to a dense local fibrotic reaction, the cause of which is unknown. The other half were associated with graft complications. It is logical to expect a relationship between the causes of hydronephrosis and graft infection or an aortoenteric fistula. However, why should there be an association with thrombosis of the limb of the graft or a false aneurysm in the groin? Would the authors recommend that all patients have urologic screening by ultrasonography, CT scanning, or IVP 1 year after aortic surgery to detect those patients with persistent hydronephrosis and half those with delayed ureteric compfications? Should this screening apply only to patients with an aortobifemoral or aortobiiliac graft and not to those with a tube graft? In conclusion, this article has documented that patients with ureteric complications should be investigated for associated graft complications, and conversely, that patients with aortic graft complications should have a fifll urologic workup, since 5% will have an abnormality; I agree. Dr. Calvin B. Ernst. We thank Dr. Johnston for his perceptive review of the manuscript and his in-depth comments. I will try to answer his questions in the order asked. First, you vcanted to know whether ureteral stenting facilitates repair of an inflammatory aneurysm assuming that after operation the retroperitoneal involvement will resolve and thereby relieve the hydronephrosis. Since we
ticle presents the largest detailed analysis of ureteric complications associated with abdominal aortic surgery. First, the authors reported six patients with ureteric obstruction occurring with an abdominal aortic aneurysm. The usual management of ureteric obstruction associated with an inflammatory aneurysm is ureterolysis and aneurysm repair; however, this procedure may be difficult, and there is the risk of ureteric injury and subsequent graft infection. I ask the authors to comment on two other approaches. D o -~ou feel that it is reasonable to simply repair the aneurysm, use temporary indwelling ureteric stents if necessary, and anticipate that the hydronephrosis will resolve spontaneously? I am reluctant to mention the use of steroids, but as documented by others and by three patients who were &herwise inoperable and being treated on our service with steroids, hydronephrosis and the inflammatory component o f the aneurysm may resolve with steroid administration. Would you consider using steroids before surgery to reduce the inflammatory component and ureteric obstruction, especially in a patient with a small aortic aneurysm? Second, the authors described 44 patients with ureteric complications after aortic surgery. There are three clinical types. Early transient hydronephrosis occurs in 10% to 15% of aortic cases because of edema from the dissection and, as reported, is of no clinical significance since it usually resolves. Early persistent hydronephrosis is uncommon, occurring in approximately 1% to 2% of aortic cases. Op-
36
Journal of VASCULAR SURGERY
Wright et al.
had only six patients with preoperative ureteral involvement, only one of whom had an inflammatory aneurysm, we cannot make definitive recommendations. Nonetheless, when encountering an inflammatory abdominal aortic aneurysm with ureteral obstruction, we feel it is important to precisely identify the ureters, and ureteral stenting helps. We place ureteral stents and assume that the periureteral reaction will subside, and the ureteral obstruction will spontaneously resolve after operation. That this invariably occurs is documented by reports in the literature as well as our one patient. The corollary, of course, as you have asked--what is the role of steroids? We do not recommend using steroids in this elderly patient population for obvious reasons. I might also point out that steroids will be most useful, if they are at all, in the early postoperative period when the inflammatory phase of the retroperitoneal reaction is at its maximum. Steroids would probably have no ameliorating effect on the retroperitoneal fibroproliferative reaction, which develops late after operation. You asked about the relationship between aortofemoral bypass complications and ureteral problems. We identified 27 patients who underwent aortobifemoral bypass reconstruction and 19 with aortoiliac reconstruction. Although it is attractive to postulate a relationship, ancL it may well be that there is a similar pathophysiologic defect in healing underlying the retroperitoneal inflammatory reaction that causes these patients to have anastomotic aneurysms, because this was a retrospective analysis we were unable to detect a cause and effect relationship, except that the ureteral complication may serve as a marker for a graft complication. Finally, you asked about urologic screening. Clearly, if one has a graft problem or a symptomatic ureteral problem, these patients should be very thoroughly studied. But we are not entirely sure what the cost/benefit ratio would be of routine screening with ultrasonography, although we certainly would not recommend against it. Knowing the litigious climate in which we live, I am reluctant to recommend a policy of routine postoperative screening based on our experience and this retrospective analysis. The only way to determine the true incidence and natural history of this problem is to prospectively screen all patients, a study which seems to us at least to be beyond practicality. Dr. Ronald J. Stoney (San Francisco, CaliE). My comments and question concern the later ureteral complication leading to hydronephrosis. I would like to make one comment about actually creating the retroureteral tunnel for the placement of graft limbs, since I think that occasionally blind clamping and withdrawing, or drawing the graft limbs through or enlarging the space to place a graft limb probabl3? produced unrecognized injury to the ureter. At least it has been apparent to use in seeing some patients referred with graft complications, some of which were alluded to, that a primary ureteral injury was produced by this mechanism. I think there are ways with opposing fingers to create a tunnel in the retroperitoneum and avoid this and produce a smooth bed for the graft limb to occupy.
Second, in operations or reoperations on the aorta and its branches for a variety of reasons but primarily occlusion and infection, we have had the occasion to remove aortic grafts placed many years earlier, often for infection, and f am surprised to see how few times there is anything more than the dense fibrous encasement of the graft that extends to the ureter. By that I mean the ureter is not really dilated or enlarged, and I was thinking as I came up here that I know personally of only one patient in our own series in the last 2 decades of approaching 200 aortic graft infections, who has had a proven ureteric complication leading to the graft infection. There were others who had the problem of ureteric obstruction, which was brought out by this article today, but to actually have a ureteral contamination or urine contaminating the graft and producing the actual infection or source of infection. I can only recall at least one.
I would like to ask the authors really one question and that is, what do they think the main cause is of the delayed adhesive type reaction between grafts and ureter that leads to this problem and many also be a marker of l a t ~ r a f t infection? Dr. Ernst. Since about a third of our patients ha,d ureteral entrapment, precise placement of the graft was not always accomplished. We quite agree with Dr. Stoney that precise dissection is necessary in making retroperitoneal ttlnnels under direct vision insofar as possible. However, I caution the audience that extensive dissection is contraindicated. We had one patient who had what we thought was overly extensive dissection where the ureter was divascularized and necrosed, ultimately resulting in nephroureterectomy. I think Dr. Stoney is exactly right relative to ureteral complications causing subsequent graft infection, this rarely occurs. However, in two of our patients the ureteral complications were directly related to a subsequent graft infection. I might add, parenthetically, that the retropetitoneal reaction under these particular circumstances creates major technical challenges when managing the graft(-,~mplication. Finally, Dr. Stoney you asked what causes the retroperitoneal reaction? This is an important question, and we really cannot tell you from our analysis. It may result from,. an abnormal fibroprolife'rative reaction in these patients, a defect in healing if you will, which ultimately results iia anastomotic defects and graft incorporation problems as well. Dr. Walter J. McCarthy (Chicago, Ill.). At Northwestern University over the past few years we have attempted to identify patients whose ureters might be adherent in areas where redissection would be necessary du~ing reoperation. In these cases stents have been placed preoperatively to aid in the intraoperative identification of the ureters. Stents have been used to identify ureters in cases where infected grafts needed to be removed, during reoperations for occluded aortofemoral grafts, in cases o£ inflanmaatory aneurysms, and for repair of anastomotic false aneurysms at the iliac bifurcation. We have been ~,~':~-
Volume 11 Number 1 January 1990
eraUy very pleased with this technique, which simplifies and hastens the operation, and we wonder if the authors .have had any experience with this approach. Dr. Ernst. Yes, I think when one is managing graft complications that ureteral stenting is very helpful. We are well aware of the excellent report from the Northwestern group regarding these particular complications. I might also add that it is helpful to make the diagnosis ofconcom-
Ureteral complications and aortoiliac reconstruction 37
kant ureteral involvement before operation. It is difficult, at best, to obtain simultaneous excretory urography and aortography to determine whether ureteral entrapment is present. But with constrast enhanced CT scanning the ureters stand out as little white dots, and if they are obstructed they will be big white dots. Therefore from CT scanning one can determine the relationship between the ureter and the limb of the prosthesis and facilitate reoperation.