- Email: [email protected]
Ureteral obstruction after abdominal aortic surgery
Ureteral Obstruction After Abdominal Aortic Surgery Moshe Schein, FCS (SA), Roger Saadia, Obstructive uropathy following abdominal aortic surgery can no longer be considered a rarity. Early bydronepbrosis, developing in the first postoperative year, occurs in 10% to 20% of patients; it usually runs a benign, self-limiting course. The incidence of delayed ureteral obstruction, which develops or persists after the first postoperative year, is unknown because it is asymptomatic in most cases. Although spontaneous resolution is possible, it seems that this late form is more likely to persist. The diagnosis of postoperative bydronephrosis is not an indication for urologic intervention. This should seldom be necessary; it should be resewed only for patients with evidence of worsening obstruction or deteriorating renal function. Early and particularly, delayed hydronepbrosis seems to be a marker for present or impending graft complications, such as infection or false aneurysms. A prolonged follow-up is therefore mandatory whenever the diagnosis is established as it may improve long-term survival and limb salvage. The need for routine screening for this condition remains to be established. With the availability of noninvasive methods, such a task could be easily accomplished.
From the Department of Surgery, University of the Witwatersrand, Johannesburg, South Africa. Requests for reprints should be addressed to Moshe Schein, Department of Surgery, Medical School, York Road, 2193 Parktown, Johannesburg, South Africa. Manuscript submitted September 25,1989, and accepted in revised form January 24, 1990.
86
THE AMERICAN
JOURNAL OF SURGERY
FRCS (Ed), Johannesburg,South Africa
his review is concerned only with ureteral obstruction T developing after abdominal aortic grafting procedures. Rre-existing ureteral obstruction associated with inflammatory aneurysms [Z] probably constitutes a different entity beyond the scope of this discussion. We also exclude ureteric involvement in idiopathic retroperitoneal fibrosis [2], a condition of presumably vasculogenic origin [3], found in 2% of patients with abdominal aortic aneurysms [I]. In 1962, Jacobson et al [4] reported the first case of ureteral obstruction after aortic bifurcation bypass grafting. An additional 95 cases have been reported to date in the English-language literature [2,5-371. INCIDENCE The incidence is difficult to determine since most patients with unilateral or partial bilateral obstruction remain asymptomatic. Furthermore, in most instances the obstruction is transient. Two retrospective and three prospective studies address this issue. Kaufman et al [29] screened 57 patients with ultrasound scans, 2 to 24 months after the insertion of aortic bifurcation grafts; bilateral hydronephrosis was detected in only one patient. In another retrospective study, Shubart et al [34] examined ultrasonographically, one or more times, 96 patients with aortic bifurcation grafts. Hydronephrosis was observed in 19 patients (20%) and was bilateral in 5. The mean interval between the vascular reconstruction and diagnosis was 46 months, with half of their patients identified when they presented with graft complications; in the remainder, the diagnosis of hydronephrosis was made at routine follow-up examinations of the retroperitoneum and the upper anastomosis. Frusha et al [31] performed excretory urograms on 53 patients, 4 to 100 months (mean interval: 32 months) after aortic surgery; mild to moderate hydronephrosis was elicited in 7 patients (13%). In these studies [29,32,34], the number of cases occurring within the first postoperative year was not mentioned. Heard and Hinde [IO] subjected 20 patients to routine intravenous pyelograms at 2 weeks and at 12 months after aortic reconstruction: 4 patients (20%) had hydronephrosis at the initial examination, which persisted in 2 patients (10%) at 1 year. In a recent study, Goldenberg et al [33] performed serial ultrasound examinations on 101 patients, at 1 week, 3 months, and 1 year postoperatively: hydronephrosis developed in 11 patients (12%); it was bilateral in 4 and asymptomatic in all I 1. The obstruction resolved spontaneously in 10 patients, in 9 of them within 3 months of onset; only a single case persisted at 1 year. ETIOLOGY Great emphasis has been placed on mechanical factors. Early reports suggested that the placement of a graft
VOLUME 162 JULY 1991
URETERAL OBSTRUCTION AFTER ABDOMINAL AORTIC SURGERY
limb anterior to the ureter was responsible for ureteral obstruction by compression of the ureter between the graft and the native artery [5-7,9]. Although not specifically mentioned in the literature, the distal anastomotic site of the aortic bifurcated graft is likely to have a bearing on the frequency of this complication; aortoiliac grafting obviously results in less disturbance to the ureter than aortobifemoral reconstruction. Sant et al [24] showed that this complication occurred irrespective of the anatomic position of the graft: in their collective series, the ureter lay anterior to the graft in half of the cases. The entrapment of the ureter between an old, thrombosed graft posteriorly and an fresh prosthesis placed anterior to the ureter was also incriminated [27]; this is particularly prone to happen, since the ureter, which is usually adherent to the peritoneum and is reflected anteriorly when a retroperitoneal tunnel is created, cannot be mobilized because of fibrosis from prior surgery [29]. Other mechanical factors have been suggested, such as the continuous pulsation of the graft against the ureter [38] or the infrequent mass effect of adjacent anastomotic aneurysms [24]. The view that retroperitoneal inflammation followed by fibrosis is the most common etiology is widely held [24,33]. A fibrotic reaction surrounding the prosthesis was demonstrated in a canine aortic graft model [31]. The precise pathogenesis of this phenomenon is unknown, but several authors [12,28,34] have agreed that it represents a nonspecific desmoplastic tissue reaction to the graft. Excessive postoperative bleeding into the retroperitoneum, previous or concomitant retroperitoneal procedures requiring wide dissection, or previous pelvic irradiation have been implicated as contributing factors [24]. Since aneurysm surgery usually requires a more complex dissection, the incidence of ureteral obstruction was twice that occurring after operations for occlusive disease [33]. Similar mechanisms were described in the genesis of idiopathic retroperitoneal fibrosis [39]. Gaylis [40] hypothesized that periaortic lymphedema followed by fibrosis is the cause of inflammatory aneurysms: the development of fibrosis around a “weeping” Dacron prosthesis is a possibility. An unrecognized intraoperative injury to the ureter, complicated by extravasation of urine, may lead to hydronephrosis or result in localized retroperitoneal fibrosis [41], but this is unlikely to account for more than the anecdotal case. In 1963, Shaw and Baue [36] were the first to observe that a hydroureter or a hydronephrosis after an aortic grafting procedure are commonly associated with subtle or overt graft infection; they reported four cases of graft sepsis presenting, a few months after the vascular procedure, with urinary tract symptoms and hydronephrosis. That the association of graft sepsis and obstructive uropathy is not coincidental is supported by other reports. Shubart et al [34] pointed to the high incidence of graft complication in 17 of their 19 cases of ureteral obstruction; 9 patients had graft sepsis and, although 2 of them developed overt graft infection years after the diagnosis of hydronephrosis, 7 presented with a hydronephrosis concomitant with graft infection, supporting the hypothesis THE AMERICAN
that perigraft retroperitoneal inflammation is triggered by an infected prosthesis. Other investigators have reported two cases of aortoenteric fistulas occurring a year subsequent to the diagnosis of the ureteric complication
[W The formation of multiple and recurrent false aneurysms, without evidence of graft infection, is another graft-associated complication [34]. In some of these patients, there was poor graft incorporation with a surrounding serum collection. This observation raises the possibility that ingrowth failure and perigraft fibrosis involving the adjacent ureter may share a common etiologic factor. The question of whether these cases represent a very low-grade graft infection [35] or graft reaction [42,43] remains unresolved. CLINICAL FEATURES AND NATURAL HISTORY In the symptomatic cases, the clinical features are variable. The patient may seek medical help for nonspecific symptoms such as back pain or may present with urologic symptoms such as loin pain, hematuria, or prostatism. It is possible that, in some of these cases, another concomitant urologic disease was responsible for these symptoms. Occasionally, frank renal failure is present right from the outset [24,2q. In most patients, however, the obstructive uropathy is asymptomatic. In one study [34], 84% of the patients were asymptomatic, the hydronephrosis being an incidental finding on ultrasound or angiogram. In two prospective studies [10,33], all the patients were asymptomatic. It is difficult to evaluate the impact of this hydronephrosis on the renal function. In the major studies [l&33,34], ureteral obstruction rarely compromised renal function, but it was often diagnosed early. When it developed, renal failure was due to the complete obstruction of the ureter of a single kidney or nonhydronephrotic complications such as renal artery occlusion [34]. Indeed, in many of the isolated case reports, the presence of an obstructive uropathy is not sufficient proof to incriminate it as a cause of renal dysfunction. Hydronephrosis complicating abdominal aortic grafting has been classified as “early and transient,” “early and persistent,” or “delayed” [33]. Early onset, in the first postoperative year, seems to result from perigraft inflammation or fibrosis; it usually runs a benign course with little clinical repercussions and resolves spontaneously within a few months, in most instances following the maturation of the retroperitoneal process [ 10,33,34]. Although the association of early hydronephrosis with graft sepsis was reported [38], the early form is usually not diagnosed concomitantly with graft complications [34]. Occasionally, early hydronephrosis may persist beyond 12 months and predispose to the delayed form [10,33]. Because all the reports of delayed uropathy are retrospective, the natural history of persistent hydronephrosis and its possible role in the genesis of the delayed form is unclear. Even in the absence of renal dysfunction, the late diagnosis of ureteral obstruction has an important clinical implication: it serves as a marker of present or impending JOURNAL
OF SURGERY
VOLUME
162
JULY
1991
87
SCHEIN AND SAADIA
graft complications, often forming part of a triad consisting of hydronephrosis, anastomotic aneurysms, and a perigraft inflammatory reaction subsequent to low-grade graft infection. The long-term outcome of the delayed obstructive uropathy is not known. Although the putative etiologic factors (graft sepsis or perigraft desmoplastic reaction) could suggest the persistence of the obstruction, a spontaneous resolution may occur. SCREENING
It is unclear whether systematic screening in search for an incipient obstructive uropathy is warranted [44]. Routine intravenous pyelograms both pre-and postoperatively are advocated by some authors [10,18]. After the preoperative angiography, a delayed abdominal film often provides sufficient visualization of the upper urinary tracts to serve as a baseline for subsequent comparison [23]. The optimal timing for screening is an important consideration. Sant et al [24] have speculated that 2 weeks postoperatively may be too early, since retropcritoneal edema and swelling would still be present; they suggest that screening be performed at about 6 to 8 weeks postoperatively as this may uncover the true incidence of early ureteral obstruction, and additional examination at 1 year would reveal the incidence of the delayed form. Goldenberg et al [33] recommend periodic ultrasound studies, beyond the first postoperative year, in all cases of early transient hydronephrosis since these may develop a protracted form. The strong association between the delayed form of hydronephrosis and graft complications calls for a routine evaluation of the urinary tract [34], whether or not vascular reintervention is contemplated. TREATMENT
In view of the uncertainty about the natural history of obstructive uropathy, it is not surprising that the therapeutic indications, especially operative treatment, are poorly defined. There is an obvious trend in the literature: when an incidental postoperative diagnosis of hydronephrosis is made, a conservative approach is favored [10,33,34]; in contrast, when symptoms are present, particularly when these are of a urologic nature (and it is not always proven that they are the direct consequence of the obstructive uropathy), an operative procedure tends to be performed [24,26j. In the absence of a strict stratification of this condition into pathogenic subtypes, the claim that operative treatment affords good results (for example, an 85% success rate in a collective series [24]) must be regarded with caution. The presently acknowledged propensity for the early hydronephrosis to be a self-limiting condition, and the predictably good results of conservative management, shed some degree of suspicion on the necessity of some of the interventions undertaken in the past. The same criticism may apply to the suggestion, in the absence of a prospective trial, that steroids may hasten the resolution of the uropathy [23]. It is interesting to note the effects of the failure to act on the putative etiologic factors of the uropathy: patients obtained surgical relief of their hydronephrosis, only to die later of an aortocnteric fistula related to graft sepsis [28]. 88
THE AMERICAN
JOURNAL OF SURGERY
When operation is indicated, the urologic options are numerous and beyond the scope of this review. The reconstructive operation performed usually consists of ureterolysis either alone or combined with some form of ureteral division and reconstruction. The position of the ureter in relation to the graft must be ascertained preqeratively, either by reference to the previous operative report or by a combined excretory urogram and aortogram [24]. Two issues are debated by urologists [24,26]: whether, on completion of the operation, the ureter must always be anterior to the graft, and whether extravasation of urine is so detrimental that division and reanastomosis of the graft must be preferred to that of the ureter. In any case, the interposition of peritoneum or omentum between the two structures is recommended. CONCLUSION
Because it is often subclinical, obstructive uropathy following abdominal aortic grafting procedures is less rare than is reported. Early hydronephrosis, developing in the first postoperative year, occurs in loo/o to 20% of patients. It is usually an asymptomatic, benign, and selflimiting condition. The incidence of delayed ureteral ob struction, developing or persisting beyond the first postoperative year, is unknown. Renal function is rarely affected. Although spontaneous resolution is possible, it seems that this form of hydronephrosis may be more permanent. The discovery of postoperative hydronephrosis by means of contrast studies or ultrasound scanning is not an indication for urologic intervention; this should be seldom necessary and reserved only for patients with evidence of worsening obstruction and risk of deteriorating renal function. Early and particularly, delayed hydronephrosis appears to be a marker of present or impending graft complications. Beyond the need for renal function preservation, these patients should be closely evaluated for the presence of asymptomatic graft infection or false aneurysms; if limb salvage and survival are to be improved, a prolonged follow-up is mandatory. The need for routine screening for this complication remains to be determined. With the widespread availability of ultrasound scanning, such a task could be easily accomplished. ADDENDUM
Since acceptance of this review, Wright et al [45] have reported a series of 42 patients with hydronephrosis out of 3,580 patients undergoing aortoiliac reconstruction; graft complications affected 55% of them. The incidence of graft complications was 12% in the patients without ureteral complications. These data reinforce the suggestion that hydronephrosis may be a harbinger of graft complications.
REFERENCES 1. Crawford JL, Stowe CL, Safi HJ, Hallman CH, Crawford ES. Inflammatory aneurysms of the aorta. J Vast Surg 1985; 2: 113-24. 2. Koep L, Zuidema GD. The clinical significance of retroperitoneal fibrosis. Surgery 1977; 81: 250-7.
VOLUME 162 JULY 1991
URETERALOBSTRUCTlONAFTERABDOMINALAORTICSURGERY
3.Bullock N. Idiopathic retroperitoneal fibrosis. Br Med J 1988; 297: 240-l. 4. Jacobson ME, Mastio GL, Be&as EM. Aortic grafts. Ureteral obstruction as a late complication of abdominal aneurysm resettion. J Kansas Med Sot 1962; 63: 516-8. 5. Lytton B. Ureteral obstruction following aortofemoral bypass grafts. Surgery 1966; 59: 918-22. 6. Dorfman LE, Thomford NR. Unusual ureteral injury following aortoiliic by-pass graft. J Urol 1969; 101: 25-7. 7. Thomford NR, Dorfman LF. Ureteral obstruction caused by an aortofemoral bypass prosthesis. Am J Surg 1968; 115: 394-6. 8. Erlich RM, Hecht HL, Veenema RJ. Chyluria following aortoiliac bypass graft. A unique method of radiological diagnosis and review of the literature. J Urol 1972; 107: 302-3. 9. Petrone AF, Dudzinski PJ, Maniatis W. Ureteral obstruction secondary to aortic femoral bypass. Ann Surg 1974; 179: 192-6. 10. Heard G, Hmde G. Hydronephrosis complicating aortic reconstruction. Br J Surg 1975; 62: 344-7. 11. Wallijn E, Renders G, Vereecker L. Urological complications following aortofemoral bypass graft. Br J Urol 1975; 47: 617-21. 12. Vandendris M, Coregoir W. Ureteric obstruction secondary to aortic femoral bypass graft. Eur Urol 1976; 2: 135-6. 13. Chantrie M, Vandendris M, Sehulman CC. Management of ureteric obstruction in the solitary kidney by a segmental suspended ureteric prosthesis. Eur Urol 1976; 2: 286-7. 14. Rhind JR. Retroperitoneal fibrosis following aortic surgery. Br J Urol 1977; 49: 522. 15. Antkowiak JG, D’Altorio RA. Ureteral obstruction secondary to bifurcation aortic grafts. Arch Surg 1979; 114: 853-6. 16. Hagedorn CH, Tannenbaum JS, Stonee WJ. Ureteral obstruction following aortofemoral or aortoiliac bypass graft. J Term Med Assoc 1979; 72: 339-40. 17. Rastogi SP, Reid IS. Bilateral ureteral obstruction following aortic bypass surgery. Clin Nephrol 1980; 14: 250-5. 18. Bouterie RL, Harbach LB. Ureteral obstruction after aortofemoral bypass surgery. Urology 1979; 14: 273-5. 19. Fourcroy JL, Azoury B, Miller HC. Bilateral ureteral obstruction as a complication of vascular graft surgery. Urology 1980; 15: 556-8. 20. Rajfer J, Smith GW. Bilateral ureteral obstruction after repair of aortic aneurysm. J Urol 1979; 122: 255-6. 2 1. Carmichael DH, Barnes WF. Obstruction of the ureter due to a false aneurysm. Surgery 1979; 86: 769-7 1. 22. Schapira HE, Gribetz M, Wulfsohn MA, Brendler H. Ureteral injuries during vacular surgery. J Urol 1981; 125: 293-7. 23. Huben RP, Schellhammer PF. Steroid therapy for ureteral obstruction after aortoiliac graft surgery. J Urol 1981; 125: 881-3. 24. Sant GR, Heaney JA, Parkhurst EC, Blaivas JG. Obstructive uropathy-a potentially serious complication of reconstructive vascular surgery. J Urol 1983; 129: 16-22. 25. Futrell W, Kaczala 0, Wolkoff J. Obstructive uropathy secondary to false aneurysm twelve years after abdominal aneurysmecto-
my. South Med J 1975; 68: 607-8. 26. Bergqvist D, Takolander R. Ureteral obstruction as a complication in aorto-iliac reconstructive surgery. Stand J Nephrol 1983; 17: 391-3. 27. Tracy DA, Eisenberg RL, Hedgcock MW. Ureteral obstruction resulting from vascular prosthetic graft surgery. AJR Am J Roentgen01 1979; 132: 415-8. 28. Gibbons RP, Mason JT, Correa RJ, Cummings KB. Management of ureteral obstruction secondary to prior aortic bypass grafts. Am Surg 1981; 47: 465-9. 29. Kaufman JE, Parsons CL, Gosink BB, Schmidt JD. Retrospective study of ureteral obstruction following vascular bypass surgery. Urology 1982; 19: 278-85. 30. Henry LG, Bernhard VM. Ureteral pathology associated with aortic surgery: report of three unusual cases. Surgery 1987; 83: 464-9. 31. Frusha JD, Porter JA, Batson RC. Hydronephrosis following aorta-femoral bypass grafts. J Cardiovasc Surg 1982; 23: 371-371. 32. Cerny JC, Scott T. Non-idiopathic retroperitoneal fibrosis. J Urol 1971; 105: 49-55. 33. Goldenberg SL, Gordon PB, Cooperberg PL, McLaughlin MG. Early hydronephrosis following aortic bifurcation graft surgery: a prcupective study. J Urol 1988; 140: 1367-9. 34. Schubart P, Fortner G, Cummings S, ef al. The significance of hydronephrosis after aortofemoral reconstruction. Arch Surg 1985; 120: 377-81. 35. Bandyk DF, Bemi GA, Thiele BL, Yowne JB. Aortofemoral graft infection due to Staphylococcus epidermidis. Arch Surg 1984; 119: 102-8. 36. Shaw RS, Baue AE. Management of sepsis complicating arterial reconstructive surgery. Surgery 1963; 53: 75-86. 37. Almagren B. Conservative management of ureteral obstruction after aortoiliac graft surgery. Br J Urol 1986; 58: 455. 38. Cockett ATK, Nakamvra RM, Rein BI. The role of arterial pulsation in experimental hydronephrosis. Invest Urol 1965; 2: 548-54. 39. Leper H, Walsh PC. Idiopathic retroperitoneal lib&s. J Urol 1979; 122: 1-5. 40. Gaylis H. Etiology of abdominal aortic “inflammatory” aneurysm: hypothesis. J Vast Surg 1985; 2: 643. 41. Mitchinson MJ, Bird DR. Urinary leakage and retroperitoneal fibrosis. J Urol 1971; 109: 56-8. 42. Kaupp HA, Matulewicz TJ, Lattimer GL, Kremen JE, Gelani VJ. Graft infection or graft reaction? Arch Surg 1979; 114: 1419-22. 43. Paes E, Vollman JF, Mohr W, Hamamt H, Brecht-Krauss D. Perigraft reaction: incompatibility of synthetic vascular graft? New aspects on clinical manifestation, pathogenesis, and therapy. World J Surg 1988; 12: 750-5. 44.. G’Mara CS. Editorial comment. J Urol 1983; 129: 21-2. 45. Wright DJ, Ernst CB, Evans JR, et al. Ureteral complications and aortoiliac reconstruction. J Vast Surg 1990; i 1: 29-37.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 162 JULY 1991
89
From the Department of Surgery, University of the Witwatersrand, Johannesburg, South Africa. Requests for reprints should be addressed to Moshe Schein, Department of Surgery, Medical School, York Road, 2193 Parktown, Johannesburg, South Africa. Manuscript submitted September 25,1989, and accepted in revised form January 24, 1990.
86
THE AMERICAN
JOURNAL OF SURGERY
FRCS (Ed), Johannesburg,South Africa
his review is concerned only with ureteral obstruction T developing after abdominal aortic grafting procedures. Rre-existing ureteral obstruction associated with inflammatory aneurysms [Z] probably constitutes a different entity beyond the scope of this discussion. We also exclude ureteric involvement in idiopathic retroperitoneal fibrosis [2], a condition of presumably vasculogenic origin [3], found in 2% of patients with abdominal aortic aneurysms [I]. In 1962, Jacobson et al [4] reported the first case of ureteral obstruction after aortic bifurcation bypass grafting. An additional 95 cases have been reported to date in the English-language literature [2,5-371. INCIDENCE The incidence is difficult to determine since most patients with unilateral or partial bilateral obstruction remain asymptomatic. Furthermore, in most instances the obstruction is transient. Two retrospective and three prospective studies address this issue. Kaufman et al [29] screened 57 patients with ultrasound scans, 2 to 24 months after the insertion of aortic bifurcation grafts; bilateral hydronephrosis was detected in only one patient. In another retrospective study, Shubart et al [34] examined ultrasonographically, one or more times, 96 patients with aortic bifurcation grafts. Hydronephrosis was observed in 19 patients (20%) and was bilateral in 5. The mean interval between the vascular reconstruction and diagnosis was 46 months, with half of their patients identified when they presented with graft complications; in the remainder, the diagnosis of hydronephrosis was made at routine follow-up examinations of the retroperitoneum and the upper anastomosis. Frusha et al [31] performed excretory urograms on 53 patients, 4 to 100 months (mean interval: 32 months) after aortic surgery; mild to moderate hydronephrosis was elicited in 7 patients (13%). In these studies [29,32,34], the number of cases occurring within the first postoperative year was not mentioned. Heard and Hinde [IO] subjected 20 patients to routine intravenous pyelograms at 2 weeks and at 12 months after aortic reconstruction: 4 patients (20%) had hydronephrosis at the initial examination, which persisted in 2 patients (10%) at 1 year. In a recent study, Goldenberg et al [33] performed serial ultrasound examinations on 101 patients, at 1 week, 3 months, and 1 year postoperatively: hydronephrosis developed in 11 patients (12%); it was bilateral in 4 and asymptomatic in all I 1. The obstruction resolved spontaneously in 10 patients, in 9 of them within 3 months of onset; only a single case persisted at 1 year. ETIOLOGY Great emphasis has been placed on mechanical factors. Early reports suggested that the placement of a graft
VOLUME 162 JULY 1991
URETERAL OBSTRUCTION AFTER ABDOMINAL AORTIC SURGERY
limb anterior to the ureter was responsible for ureteral obstruction by compression of the ureter between the graft and the native artery [5-7,9]. Although not specifically mentioned in the literature, the distal anastomotic site of the aortic bifurcated graft is likely to have a bearing on the frequency of this complication; aortoiliac grafting obviously results in less disturbance to the ureter than aortobifemoral reconstruction. Sant et al [24] showed that this complication occurred irrespective of the anatomic position of the graft: in their collective series, the ureter lay anterior to the graft in half of the cases. The entrapment of the ureter between an old, thrombosed graft posteriorly and an fresh prosthesis placed anterior to the ureter was also incriminated [27]; this is particularly prone to happen, since the ureter, which is usually adherent to the peritoneum and is reflected anteriorly when a retroperitoneal tunnel is created, cannot be mobilized because of fibrosis from prior surgery [29]. Other mechanical factors have been suggested, such as the continuous pulsation of the graft against the ureter [38] or the infrequent mass effect of adjacent anastomotic aneurysms [24]. The view that retroperitoneal inflammation followed by fibrosis is the most common etiology is widely held [24,33]. A fibrotic reaction surrounding the prosthesis was demonstrated in a canine aortic graft model [31]. The precise pathogenesis of this phenomenon is unknown, but several authors [12,28,34] have agreed that it represents a nonspecific desmoplastic tissue reaction to the graft. Excessive postoperative bleeding into the retroperitoneum, previous or concomitant retroperitoneal procedures requiring wide dissection, or previous pelvic irradiation have been implicated as contributing factors [24]. Since aneurysm surgery usually requires a more complex dissection, the incidence of ureteral obstruction was twice that occurring after operations for occlusive disease [33]. Similar mechanisms were described in the genesis of idiopathic retroperitoneal fibrosis [39]. Gaylis [40] hypothesized that periaortic lymphedema followed by fibrosis is the cause of inflammatory aneurysms: the development of fibrosis around a “weeping” Dacron prosthesis is a possibility. An unrecognized intraoperative injury to the ureter, complicated by extravasation of urine, may lead to hydronephrosis or result in localized retroperitoneal fibrosis [41], but this is unlikely to account for more than the anecdotal case. In 1963, Shaw and Baue [36] were the first to observe that a hydroureter or a hydronephrosis after an aortic grafting procedure are commonly associated with subtle or overt graft infection; they reported four cases of graft sepsis presenting, a few months after the vascular procedure, with urinary tract symptoms and hydronephrosis. That the association of graft sepsis and obstructive uropathy is not coincidental is supported by other reports. Shubart et al [34] pointed to the high incidence of graft complication in 17 of their 19 cases of ureteral obstruction; 9 patients had graft sepsis and, although 2 of them developed overt graft infection years after the diagnosis of hydronephrosis, 7 presented with a hydronephrosis concomitant with graft infection, supporting the hypothesis THE AMERICAN
that perigraft retroperitoneal inflammation is triggered by an infected prosthesis. Other investigators have reported two cases of aortoenteric fistulas occurring a year subsequent to the diagnosis of the ureteric complication
[W The formation of multiple and recurrent false aneurysms, without evidence of graft infection, is another graft-associated complication [34]. In some of these patients, there was poor graft incorporation with a surrounding serum collection. This observation raises the possibility that ingrowth failure and perigraft fibrosis involving the adjacent ureter may share a common etiologic factor. The question of whether these cases represent a very low-grade graft infection [35] or graft reaction [42,43] remains unresolved. CLINICAL FEATURES AND NATURAL HISTORY In the symptomatic cases, the clinical features are variable. The patient may seek medical help for nonspecific symptoms such as back pain or may present with urologic symptoms such as loin pain, hematuria, or prostatism. It is possible that, in some of these cases, another concomitant urologic disease was responsible for these symptoms. Occasionally, frank renal failure is present right from the outset [24,2q. In most patients, however, the obstructive uropathy is asymptomatic. In one study [34], 84% of the patients were asymptomatic, the hydronephrosis being an incidental finding on ultrasound or angiogram. In two prospective studies [10,33], all the patients were asymptomatic. It is difficult to evaluate the impact of this hydronephrosis on the renal function. In the major studies [l&33,34], ureteral obstruction rarely compromised renal function, but it was often diagnosed early. When it developed, renal failure was due to the complete obstruction of the ureter of a single kidney or nonhydronephrotic complications such as renal artery occlusion [34]. Indeed, in many of the isolated case reports, the presence of an obstructive uropathy is not sufficient proof to incriminate it as a cause of renal dysfunction. Hydronephrosis complicating abdominal aortic grafting has been classified as “early and transient,” “early and persistent,” or “delayed” [33]. Early onset, in the first postoperative year, seems to result from perigraft inflammation or fibrosis; it usually runs a benign course with little clinical repercussions and resolves spontaneously within a few months, in most instances following the maturation of the retroperitoneal process [ 10,33,34]. Although the association of early hydronephrosis with graft sepsis was reported [38], the early form is usually not diagnosed concomitantly with graft complications [34]. Occasionally, early hydronephrosis may persist beyond 12 months and predispose to the delayed form [10,33]. Because all the reports of delayed uropathy are retrospective, the natural history of persistent hydronephrosis and its possible role in the genesis of the delayed form is unclear. Even in the absence of renal dysfunction, the late diagnosis of ureteral obstruction has an important clinical implication: it serves as a marker of present or impending JOURNAL
OF SURGERY
VOLUME
162
JULY
1991
87
SCHEIN AND SAADIA
graft complications, often forming part of a triad consisting of hydronephrosis, anastomotic aneurysms, and a perigraft inflammatory reaction subsequent to low-grade graft infection. The long-term outcome of the delayed obstructive uropathy is not known. Although the putative etiologic factors (graft sepsis or perigraft desmoplastic reaction) could suggest the persistence of the obstruction, a spontaneous resolution may occur. SCREENING
It is unclear whether systematic screening in search for an incipient obstructive uropathy is warranted [44]. Routine intravenous pyelograms both pre-and postoperatively are advocated by some authors [10,18]. After the preoperative angiography, a delayed abdominal film often provides sufficient visualization of the upper urinary tracts to serve as a baseline for subsequent comparison [23]. The optimal timing for screening is an important consideration. Sant et al [24] have speculated that 2 weeks postoperatively may be too early, since retropcritoneal edema and swelling would still be present; they suggest that screening be performed at about 6 to 8 weeks postoperatively as this may uncover the true incidence of early ureteral obstruction, and additional examination at 1 year would reveal the incidence of the delayed form. Goldenberg et al [33] recommend periodic ultrasound studies, beyond the first postoperative year, in all cases of early transient hydronephrosis since these may develop a protracted form. The strong association between the delayed form of hydronephrosis and graft complications calls for a routine evaluation of the urinary tract [34], whether or not vascular reintervention is contemplated. TREATMENT
In view of the uncertainty about the natural history of obstructive uropathy, it is not surprising that the therapeutic indications, especially operative treatment, are poorly defined. There is an obvious trend in the literature: when an incidental postoperative diagnosis of hydronephrosis is made, a conservative approach is favored [10,33,34]; in contrast, when symptoms are present, particularly when these are of a urologic nature (and it is not always proven that they are the direct consequence of the obstructive uropathy), an operative procedure tends to be performed [24,26j. In the absence of a strict stratification of this condition into pathogenic subtypes, the claim that operative treatment affords good results (for example, an 85% success rate in a collective series [24]) must be regarded with caution. The presently acknowledged propensity for the early hydronephrosis to be a self-limiting condition, and the predictably good results of conservative management, shed some degree of suspicion on the necessity of some of the interventions undertaken in the past. The same criticism may apply to the suggestion, in the absence of a prospective trial, that steroids may hasten the resolution of the uropathy [23]. It is interesting to note the effects of the failure to act on the putative etiologic factors of the uropathy: patients obtained surgical relief of their hydronephrosis, only to die later of an aortocnteric fistula related to graft sepsis [28]. 88
THE AMERICAN
JOURNAL OF SURGERY
When operation is indicated, the urologic options are numerous and beyond the scope of this review. The reconstructive operation performed usually consists of ureterolysis either alone or combined with some form of ureteral division and reconstruction. The position of the ureter in relation to the graft must be ascertained preqeratively, either by reference to the previous operative report or by a combined excretory urogram and aortogram [24]. Two issues are debated by urologists [24,26]: whether, on completion of the operation, the ureter must always be anterior to the graft, and whether extravasation of urine is so detrimental that division and reanastomosis of the graft must be preferred to that of the ureter. In any case, the interposition of peritoneum or omentum between the two structures is recommended. CONCLUSION
Because it is often subclinical, obstructive uropathy following abdominal aortic grafting procedures is less rare than is reported. Early hydronephrosis, developing in the first postoperative year, occurs in loo/o to 20% of patients. It is usually an asymptomatic, benign, and selflimiting condition. The incidence of delayed ureteral ob struction, developing or persisting beyond the first postoperative year, is unknown. Renal function is rarely affected. Although spontaneous resolution is possible, it seems that this form of hydronephrosis may be more permanent. The discovery of postoperative hydronephrosis by means of contrast studies or ultrasound scanning is not an indication for urologic intervention; this should be seldom necessary and reserved only for patients with evidence of worsening obstruction and risk of deteriorating renal function. Early and particularly, delayed hydronephrosis appears to be a marker of present or impending graft complications. Beyond the need for renal function preservation, these patients should be closely evaluated for the presence of asymptomatic graft infection or false aneurysms; if limb salvage and survival are to be improved, a prolonged follow-up is mandatory. The need for routine screening for this complication remains to be determined. With the widespread availability of ultrasound scanning, such a task could be easily accomplished. ADDENDUM
Since acceptance of this review, Wright et al [45] have reported a series of 42 patients with hydronephrosis out of 3,580 patients undergoing aortoiliac reconstruction; graft complications affected 55% of them. The incidence of graft complications was 12% in the patients without ureteral complications. These data reinforce the suggestion that hydronephrosis may be a harbinger of graft complications.
REFERENCES 1. Crawford JL, Stowe CL, Safi HJ, Hallman CH, Crawford ES. Inflammatory aneurysms of the aorta. J Vast Surg 1985; 2: 113-24. 2. Koep L, Zuidema GD. The clinical significance of retroperitoneal fibrosis. Surgery 1977; 81: 250-7.
VOLUME 162 JULY 1991
URETERALOBSTRUCTlONAFTERABDOMINALAORTICSURGERY
3.Bullock N. Idiopathic retroperitoneal fibrosis. Br Med J 1988; 297: 240-l. 4. Jacobson ME, Mastio GL, Be&as EM. Aortic grafts. Ureteral obstruction as a late complication of abdominal aneurysm resettion. J Kansas Med Sot 1962; 63: 516-8. 5. Lytton B. Ureteral obstruction following aortofemoral bypass grafts. Surgery 1966; 59: 918-22. 6. Dorfman LE, Thomford NR. Unusual ureteral injury following aortoiliic by-pass graft. J Urol 1969; 101: 25-7. 7. Thomford NR, Dorfman LF. Ureteral obstruction caused by an aortofemoral bypass prosthesis. Am J Surg 1968; 115: 394-6. 8. Erlich RM, Hecht HL, Veenema RJ. Chyluria following aortoiliac bypass graft. A unique method of radiological diagnosis and review of the literature. J Urol 1972; 107: 302-3. 9. Petrone AF, Dudzinski PJ, Maniatis W. Ureteral obstruction secondary to aortic femoral bypass. Ann Surg 1974; 179: 192-6. 10. Heard G, Hmde G. Hydronephrosis complicating aortic reconstruction. Br J Surg 1975; 62: 344-7. 11. Wallijn E, Renders G, Vereecker L. Urological complications following aortofemoral bypass graft. Br J Urol 1975; 47: 617-21. 12. Vandendris M, Coregoir W. Ureteric obstruction secondary to aortic femoral bypass graft. Eur Urol 1976; 2: 135-6. 13. Chantrie M, Vandendris M, Sehulman CC. Management of ureteric obstruction in the solitary kidney by a segmental suspended ureteric prosthesis. Eur Urol 1976; 2: 286-7. 14. Rhind JR. Retroperitoneal fibrosis following aortic surgery. Br J Urol 1977; 49: 522. 15. Antkowiak JG, D’Altorio RA. Ureteral obstruction secondary to bifurcation aortic grafts. Arch Surg 1979; 114: 853-6. 16. Hagedorn CH, Tannenbaum JS, Stonee WJ. Ureteral obstruction following aortofemoral or aortoiliac bypass graft. J Term Med Assoc 1979; 72: 339-40. 17. Rastogi SP, Reid IS. Bilateral ureteral obstruction following aortic bypass surgery. Clin Nephrol 1980; 14: 250-5. 18. Bouterie RL, Harbach LB. Ureteral obstruction after aortofemoral bypass surgery. Urology 1979; 14: 273-5. 19. Fourcroy JL, Azoury B, Miller HC. Bilateral ureteral obstruction as a complication of vascular graft surgery. Urology 1980; 15: 556-8. 20. Rajfer J, Smith GW. Bilateral ureteral obstruction after repair of aortic aneurysm. J Urol 1979; 122: 255-6. 2 1. Carmichael DH, Barnes WF. Obstruction of the ureter due to a false aneurysm. Surgery 1979; 86: 769-7 1. 22. Schapira HE, Gribetz M, Wulfsohn MA, Brendler H. Ureteral injuries during vacular surgery. J Urol 1981; 125: 293-7. 23. Huben RP, Schellhammer PF. Steroid therapy for ureteral obstruction after aortoiliac graft surgery. J Urol 1981; 125: 881-3. 24. Sant GR, Heaney JA, Parkhurst EC, Blaivas JG. Obstructive uropathy-a potentially serious complication of reconstructive vascular surgery. J Urol 1983; 129: 16-22. 25. Futrell W, Kaczala 0, Wolkoff J. Obstructive uropathy secondary to false aneurysm twelve years after abdominal aneurysmecto-
my. South Med J 1975; 68: 607-8. 26. Bergqvist D, Takolander R. Ureteral obstruction as a complication in aorto-iliac reconstructive surgery. Stand J Nephrol 1983; 17: 391-3. 27. Tracy DA, Eisenberg RL, Hedgcock MW. Ureteral obstruction resulting from vascular prosthetic graft surgery. AJR Am J Roentgen01 1979; 132: 415-8. 28. Gibbons RP, Mason JT, Correa RJ, Cummings KB. Management of ureteral obstruction secondary to prior aortic bypass grafts. Am Surg 1981; 47: 465-9. 29. Kaufman JE, Parsons CL, Gosink BB, Schmidt JD. Retrospective study of ureteral obstruction following vascular bypass surgery. Urology 1982; 19: 278-85. 30. Henry LG, Bernhard VM. Ureteral pathology associated with aortic surgery: report of three unusual cases. Surgery 1987; 83: 464-9. 31. Frusha JD, Porter JA, Batson RC. Hydronephrosis following aorta-femoral bypass grafts. J Cardiovasc Surg 1982; 23: 371-371. 32. Cerny JC, Scott T. Non-idiopathic retroperitoneal fibrosis. J Urol 1971; 105: 49-55. 33. Goldenberg SL, Gordon PB, Cooperberg PL, McLaughlin MG. Early hydronephrosis following aortic bifurcation graft surgery: a prcupective study. J Urol 1988; 140: 1367-9. 34. Schubart P, Fortner G, Cummings S, ef al. The significance of hydronephrosis after aortofemoral reconstruction. Arch Surg 1985; 120: 377-81. 35. Bandyk DF, Bemi GA, Thiele BL, Yowne JB. Aortofemoral graft infection due to Staphylococcus epidermidis. Arch Surg 1984; 119: 102-8. 36. Shaw RS, Baue AE. Management of sepsis complicating arterial reconstructive surgery. Surgery 1963; 53: 75-86. 37. Almagren B. Conservative management of ureteral obstruction after aortoiliac graft surgery. Br J Urol 1986; 58: 455. 38. Cockett ATK, Nakamvra RM, Rein BI. The role of arterial pulsation in experimental hydronephrosis. Invest Urol 1965; 2: 548-54. 39. Leper H, Walsh PC. Idiopathic retroperitoneal lib&s. J Urol 1979; 122: 1-5. 40. Gaylis H. Etiology of abdominal aortic “inflammatory” aneurysm: hypothesis. J Vast Surg 1985; 2: 643. 41. Mitchinson MJ, Bird DR. Urinary leakage and retroperitoneal fibrosis. J Urol 1971; 109: 56-8. 42. Kaupp HA, Matulewicz TJ, Lattimer GL, Kremen JE, Gelani VJ. Graft infection or graft reaction? Arch Surg 1979; 114: 1419-22. 43. Paes E, Vollman JF, Mohr W, Hamamt H, Brecht-Krauss D. Perigraft reaction: incompatibility of synthetic vascular graft? New aspects on clinical manifestation, pathogenesis, and therapy. World J Surg 1988; 12: 750-5. 44.. G’Mara CS. Editorial comment. J Urol 1983; 129: 21-2. 45. Wright DJ, Ernst CB, Evans JR, et al. Ureteral complications and aortoiliac reconstruction. J Vast Surg 1990; i 1: 29-37.
THE AMERICAN JOURNAL OF SURGERY
VOLUME 162 JULY 1991
89