Chronic renal artery occlusion: Nephrectomy versus revascularization

Transcription

1 ORIGINAL ARTICLES Chronic renal artery occlusion: Nephrectomy versus revascularization Timothy C. Oskin, MD, Kimberley J. Hansen, MD, Jonathan S. Deitch, MD, Timothy E. Craven, MSPH, and Richard H. Dean, MD,Winston-Salem, NC Purpose: The surgical management of chronic atherosclerotic renal artery occlusion (RA- OCC) was studied. Methods: From January 1987 through December 1996, 397 consecutive patients were treated for atherosclerotic renal artery disease. Ninety-five hypertensive patients (mean blood pressure, 204 ± 31/106 ± 20 mm Hg; mean medications, 3.0 ± 1.1 drugs) were treated for 100 RA-OCCs. Eighty-four (88%) patients had renal dysfunction, defined by serum creatinine levels 1.3 mg/dl (mean serum creatinine level, 2.8 ± 2.0 mg/dl). Demographic characteristics, operative morbidity and mortality, blood pressure/renal function response, and postoperative decline in renal function were examined and compared with that of 302 patients treated for renal artery stenosis (RAS). Results: After operation, there were 5 perioperative deaths (5.2%), 2 (2.8%) after revascularization and 3 (12%) after nephrectomy (P =.11), compared with 12 (4.0%) perioperative deaths in the RAS group (P =.59). After controlling for important covariates, estimated survival and blood pressure benefits did not differ between RA-OCC patients treated by nephrectomy or revascularization (P =.13; 87% vs 92%, P =.54). Excretory renal function was considered improved in 49% of 79 RA-OCC patients with renal dysfunction, including 9 patients removed from dialysis-dependence. Among patients treated for unilateral disease, revascularization for RA-OCC was associated with significant improvement in renal function (P <.01); however, nephrectomy alone did not increase renal function significantly. Improved renal function after operation was associated with a significant and independent increase in survival (P <.01) and dialysis-free survival (P <.01) among patients treated for RA-OCC. In addition, blood pressure benefit, renal function response, and estimated survival did not differ significantly after reconstruction for RA-OCC or RAS. Conclusion: Among hypertensive patients treated for RA-OCC, equivalent beneficial blood pressure response was observed after both revascularization and nephrectomy. In patients who underwent bilateral renal artery revascularization, the change in excretory renal function attributable to repair of RA-OCC cannot be defined. In patients treated for unilateral disease, however, improvement in function was observed only after revascularization. Moreover, improved renal function demonstrated a significant and independent association with improved survival. This experience supports renal revascularization in preference to nephrectomy for RA-OCC in select hypertensive patients when a normal distal artery is demonstrated at operation. (J Vasc Surg 1999;29:140-9.) Renal artery occlusion (RA-OCC) represents the final expression of anatomically progressive atherosclerotic renovascular occlusive disease. When a highgrade stenosis progresses to occlusion, the event may From the Departments of General Surgery and Public Health Sciences, Division of Surgical Sciences, Wake Forest University School of Medicine. Supported in part by a grant from the National Institutes of Diabetes and Digestive and Kidney Diseases, National Institutes of Health, United States Public Health Service (1R01KD47414). Presented at the Twenty-second Annual Meeting of the Southern Association for Vascular Surgery, Rio Grande, Puerto Rico, Jan 21 24, be clinically silent or previously controlled hypertension may worsen and serum creatinine levels may elevate. 1 Given that as much as 35% of mild contralateral renal artery atherosclerotic lesions progress to sig- Reprint requests: Kimberley J. Hansen, MD, Associate Professor of Surgery, Department of General Surgery, Division of Surgical Sciences, Wake Forest University Medical Center, Medical Center Boulevard, Winston-Salem, NC Copyright 1999 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /99/$ /6/93814

2 Volume 29, Number 1 Oskin et al 141 Table I. Demographic characteristics of patients with renal artery stenosis (RAS) and renal artery occlusion (RA-OCC) RAS (n = 302) RA-OCC (n = 95) P value Mean age (years) 62.7 ± ± 9.91 Sex:.15 Male 47% 56% Female 53% 45% Mean SBP (mm Hg) 197 ± ± Mean DBP (mm Hg) 103 ± ± BP medications 2.5 ± ± 1.1 <.01 Mean SCr (mg/dl) 2.0 ± ± 2.0 <.01 Mean EGFR (ml/min/m 2 ) 41 ± ± 18 <.01 Severe renal dysfunction (SCr >2.0) 35% 60% <.01 Dialysis dependence 5% 12%.03 Cardiac disease 42% 51%.15 Angina 23% 25%.71 MI 30% 39%.11 CHF 16% 34% <.01 CABG/PTCA 13% 12%.73 Cerebrovascular disease 28% 23%.34 TIA/CVA 23% 18%.34 CEA 15% 9%.16 SBP, systolic blood pressure; DBP, diastolic blood pressure; SCr, serum creatinine level; EGFR, estimated glomerular filtration rate; MI, myocardial infarction; CHF, congestive heart failure; CABG/PTCA, coronary artery bypass grafting/percutaneous coronary angioplasty; TIA/CVA, transient ischemic attack/cerebrovascular accident; CEA, carotid endarterectomy. nificant stenosis during follow-up, we have preferred renal artery revascularization to nephrectomy in patients with RA-OCC. 2 Despite this preference, issues that determine clinical benefit from revascularization are poorly understood. Past recommendations about surgical reconstruction for RA-OCC have emphasized absolute renal length, demonstration of a reconstituted distal renal artery by means of angiography, and glomerular architecture on kidney biopsy. 3-7 In contrast to those who would recommend nephrectomy based on these parameters, we have reserved nephrectomy for a surgically unreconstructible renal artery to a poorly functioning kidney. To examine this management philosophy, we reviewed our results in patients with RA-OCC treated by revascularization and nephrectomy and compared these results with patients treated for renal artery stenosis (RAS) during the same 10-year period. Specific areas of interest included: (1) the morbidity, mortality, and blood pressure/renal function response after nephrectomy vs revascularization for RA-OCC, and (2) the impact of nephrectomy or revascularization on the subsequent decline in renal excretory function and estimated survival. PATIENTS AND METHODS Patient population. Between January 1987 and January 1997, 397 patients underwent operative repair of atherosclerotic renovascular disease at Wake Forest University Medical Center. Within this group, 117 patients were found to have complete occlusion of at least 1 renal artery. Omitting 22 patients in whom occlusion occurred acutely or after renal artery intervention, 95 consecutive patients were treated for 100 chronic atherosclerotic renal artery occlusions (Table I). This group included 52 women and 43 men with a mean age of 62.6 ± 9 years (range, years). All patients had hypertension (mean blood pressure, 204 ± 31/106 ± 20 mm Hg; mean antihypertension medications, 3.0 ± 1.1 drugs). The mean serum creatinine level was 2.8 ± 2.0 mg/dl (range, mg/dl). Eighty-four patients had renal dysfunction as defined by serum creatinine levels 1.3 mg/dl, including 57 patients with severe renal dysfunction (serum creatinine level, 2.0 mg/dl; 11 patients dependent on dialysis). Site specific end-organ disease (determined by history, physical examination, and the results of perioperative evaluations) was present in 95% of 95 patients and was manifested by renal dysfunction in 88%; cardiac disease in 50%; cerebrovascular disease in 23%; significant aortic disease in 73%; and peripheral vascular disease in 31%. Only 5 patients were free of end-organ manifestations of atherosclerosis. Patient treatment. After complete histories and physical examinations, all patients underwent electrocardiogram. Based on these results, 85 patients underwent further cardiac evaluation, including

3 142 Oskin et al January 1999 Table II. Operative procedures performed in 95 patients with renal artery occlusion (RA-OCC) Number Total number of patients 95 Total number of kidneys 190 Number of kidneys with RA-OCC 100 Revascularizaion 75 Bypass 56 Thromboendarterectomy 14 Reimplantation 5 Nephrectomy 25 Combined aortic reconstruction 39 Aneurysmal disease 14 Occlusive disease 25 resting echocardiography (68 patients), stress echocardiography (29 patients), stress electrocardiogram (25 patients), adenosine-thallium scintigraphy (22 patients), and cardiac catheterization (31 patients). Twelve patients had previously undergone coronary artery bypass grafting. In 3 instances, this procedure was performed within 3 months of renal revascularization. Evaluation of renal vasculature included conventional cut-film angiography in all patients. Distal renal artery reconstitution was demonstrated by means of angiography on delayed films to 56 of 81 kidneys (69%), whereas 54 of 87 kidneys (62%) demonstrated a nephrogram. Neither distal reconstitution nor nephrogram was demonstrated in 21 of 79 kidneys (27%). Preoperative kidney length associated with RA-OCC in 77 patients averaged 8.0 ± 1.5 cm. Isotopic renograms were performed in 43 patients with unilateral RA-OCC. Fourteen patients had more than 25% of their total renal function supplied by the occluded kidney. Conversely, 10 patients had less than 5% of total renal function supplied by the occluded kidney. Operative procedures. Ninety-five patients underwent reconstruction for 100 kidneys with RA- OCC. Seventy-five kidneys in 70 patients underwent revascularization, and 25 nephrectomies were performed for unreconstructible renal artery lesions (Table II). Renal artery reconstruction for RA-OCC included aortorenal bypass in 56 instances, thromboendarterectomy in 14 instances, and reimplantation in 5 instances. Hemodynamically significant contralateral renal artery occlusive disease requiring reconstruction was present in 59 of 90 patients (66%) with unilateral RA-OCC. Of these, 19 of 59 patients required nephrectomy, whereas 40 patients were revascularized on the side of RA-OCC. No nephrectomies were performed for a nonoccluded renal artery, and no hemodynamically significant renal artery lesion was left uncorrected. Simultaneous aortic reconstruction was performed in 39 patients (41%). Fourteen had combined repair for aneurysmal disease, whereas 25 had repair of clinically significant aortic occlusive disease. Data analysis. Within the RA-OCC group, blood pressure/renal function response and postoperative decline in renal function were compared between patients undergoing renal revascularization and patients undergoing nephrectomy. Demographic characteristics, perioperative mortality, blood pressure, and renal function response to operation were compared between the 95 patients treated for RA- OCC and the 302 patients treated for RAS during the same period. Values of serum creatinine were transformed to a calculated measure of estimated glomerular filtration rate (EGFR, ml/min/1.73m 2 ) with the modified formula of Cockroft and Gault 8 : EGFR = [(140 Age)(Weight)/(72)(Serum creatinine level)] [1.73/BSA], where body surface area (BSA) was calculated by BSA = [Height (cm)] [Weight (kg)] For female patients, EGFR was multiplied by 0.85 to correct for sex differences in muscle mass and average rate of creatinine synthesis. Summary statistics (means and standard deviations of continuous data and frequencies and relative frequencies of categorical data) were calculated, and the data were examined to verify that assumptions of statistical tests were met. Statistical comparison of preoperative patient demographics were performed using t tests for continuous factors and chi-square tests for categorical factors. Chi-square tests were used for comparison of blood pressure and graded renal function responses to operation for RA-OCC vs RAS groups and, in RA-OCC patients, nephrectomy vs revascularization groups. In RA-OCC patients, analyses of differences between groups in preoperative EGFR were performed using unpaired t tests; within-group preoperative to postoperative changes in EGFR were assessed by means of paired t tests. Between-group differences in postoperative EGFR were tested by means of analysis of covariance that controlled for preoperative levels of EGFR. Graphic depiction of overall postoperative survival used product-limit estimates 9 of the survival distributions for RA-OCC and RAS patient groups, and in RA-OCC patients, for nephrectomy and revascularization patient groups. Graphic depiction of the influence of renal function response on survival and

4 Volume 29, Number 1 Oskin et al 143 Fig 1. Product-limit estimate of survival among patients with renal artery occlusion by surgery type. Fig 2. Product-limit estimate of survival among patients with renal artery occlusion and patients with renal artery stenosis. Table III. Summary of operative deaths Sex Age Site of RA-OCC Operation Cause of death F 79 Unilateral Bilateral RAB MSOF M 70 Unilateral Bilateral TEA MSOF AFBG F 64 Unilateral Left nephrectomy MSOF Right TEA F 69 Unilateral Right nephrectomy CVA Left RAB; AAA F 67 Unilateral Right nephrectomy MI Left TEA RA-OCC, renal artery occlusion; RAB, renal artery bypass; TEA, thromboendarterectomy; AFBG, aortofemoral bypass graft; AAA, abdominal aortic aneurysm; MSOF, multisystem organ failure; CVA, cerebrovascular accident; MI, myocardial infarction. dialysis-free survival in RA-OCC patients used predicted time to dialysis or death from a proportional hazards regression model 10 so that the change in EGFR could be used as a continuous predictive factor. Tests for associations between factors of interest and survival were performed with proportional hazards regression models. A best subset of factors to predict postoperative survival in RA-OCC and RAS patients combined was selected with a forward stepwise variable selection procedure, 10 in which the following factors were included in the model: age, positive histories of angina, myocardial infarction, congestive heart failure (CHF), coronary artery bypass graft/percutaneous transluminal coronary angioplasty, transient ischemic attack (TIA)/cerebrovascular accident, carotid endarterectomy, peripheral vascular disease or diabetes mellitus, presence of abdominal aortic aneurysm or aortic occlusive disease, preoperative creatinine level, site of disease, preoperative systolic blood pressure (SBP) and diastolic blood pressure (DBP), sex, and preoperative dialysis dependence. Factors were introduced to the model one by one, starting with the most significant, until all factors significant at the 10% alpha-level were included. A similar analysis was performed in the RA-OCC patients after adding preoperative to postoperative change in EGFR as an additional co-variate. RESULTS Postoperative morbidity and mortality Early events. There were 5 (5.3%) perioperative deaths among the 95 patients treated for RA- OCC. A summary of operative deaths is included in Table III. All 5 patients who died required complex procedures, including bilateral revascularization in 5 patients and concomitant aortic procedures in 2 patients. The perioperative mortality rate of 2.8% in the renal artery revascularization group and 12% in the nephrectomy group did not differ statistically (P =.11; Table IV). A similar

5 144 Oskin et al January 1999 Fig 3. Predicted survival for patients with renal artery occlusion who have improved or unimproved renal function after operation. perioperative mortality rate of 4.0% was observed in the RAS group (P =.59; Table V). Late events. During the mean follow-up period of 32 months, the survival rate did not differ significantly within the RA-OCC group whether the patients were treated by nephrectomy or renal artery revascularization (P =.47). A significant and independent association with increased survival and dialysisfree survival among patients with improved postoperative EGFR was demonstrated by means of multivariate analysis, whereas decreased survival was associated with increased preoperative age, serum creatinine level, and history of CHF (Table VI). The productlimit estimate of time to death in the nephrectomy group is depicted by a solid line in Fig 1. During follow-up, there were 24 deaths in the RA-OCC group, compared with 37 in the RAS group (P >.01). Preoperative age, creatinine level, history of CHF, and history of TIA or stroke demonstrated significant and independent association with decreased survival by means of multivariate analysis (Table VII). After controlling for these preoperative factors, estimated survival rates did not differ between patients treated for RA-OCC vs those treated for RAS (P =.39; Table VII). The productlimit estimate of time to death in the RA-OCC group is depicted by a solid line in Fig 2. Graft occlusion occurred in 4 patients during the follow-up period after revascularization of an RA- OCC. The primary procedure on 3 of these patients was an aortorenal bypass with saphenous vein graft in 2 instances and synthetic graft in 1 instance, whereas bilateral thromboendarterectomy was performed in the other patient. All 4 of these patients required secondary procedures. Although 2 patients required secondary nephrectomy, 2 underwent secondary revascularization procedures, including thrombectomy and patch angioplasty in 1 instance and saphenous vein aortorenal bypass in 1 instance. Blood pressure response. Blood pressure and medication requirements were determined at followup examinations at least 2 months after surgery. Patients were considered cured if they had a diastolic blood pressure of 90 mm Hg or less while receiving no antihypertensive medications. The patient was improved if: (1) their diastolic blood pressure was less than 90 mm Hg preoperatively, and control was now obtained with at least 2 fewer medications; (2) their blood pressure was uncontrolled preoperatively, the decrease in their diastolic blood pressure was less than 20 mm Hg associated with postoperative control, and their medications were reduced by at least 1 drug; (3) there was control of blood pressure with a decrease in diastolic blood pressure of at least 20 mm Hg with no increase in the number of medications postoperatively. Patients that were not cured or improved were considered failed. Of the 90 patients surviving surgical repair for RA-OCC, 10 (11%) were considered cured, 72 (80%) were considered improved, and 8 (9%) were considered failed (Table IV). Beneficial blood pressure response (cured or improved) was equivalent after nephrectomy and revascularization (87% vs 92%; P =.54). Renal function response. Postoperative renal function response was determined in patients with preoperative renal insufficiency (ie, a serum creatinine level of 1.3 mg/dl). A significant change in renal function was defined as a 20% change in serum creatinine level or removal from dialysis-dependence. Of 79 patients, 49% had improved renal function, including 9 patients removed from dialysisdependence (Table V). Patients who had improved renal function demonstrated a significant and independent increase in estimated survival (Fig 3) and dialysis-free survival (Fig 4). Evaluation of change in EGFR was determined by site of renal artery disease and extent of operation (Table VIII). A significant increase in postoperative EGFR was observed in 21 patients after unilateral revascularization (preoperative EGFR, 29.6 ± 3.1 ml/min/1.73m 2 ; postoperative EGFR, 36.8 ± 3.7 ml/min/1.73m 2 ; P <.01), 40 patients after bilateral revascularization (preoperative EGFR, 25.6 ± 1.9 ml/min/1.73m 2 ; postoperative EGFR, 33.5 ± 3.1 ml/min/1.73m 2 ; P <.01), and 14 patients after nephrectomy with contralateral revascularization (preoperative EGFR,

6 Volume 29, Number 1 Oskin et al 145 Table IV. Comparison of operative results in patients with renal artery occlusion (RA-OCC) by treatment group Renal artery revascularization Nephrectomy (n = 70) (n = 25) P value Perioperative mortality (%) Hypertension response (%).54 Cured Improved No change 8 13 Renal function response (%)*.31 Improved No change/worsened *Preoperative serum creatinine level of 1.3 mg/dl 31.5 ± 3.2 ml/min/1.73m 2 ; postoperative EGFR, 38.7 ± 4.6 ml/min/1.73m 2 ; P =.02). The change in EGFR observed after nephrectomy alone (preoperative EGFR, 47.3 ± 9.1 ml/min/1.73m 2 ; postoperative EGFR, 50.4 ± 8.4 ml/min/1.73m 2 ; P =.08) was not significant. Based on follow-up EGFR (mean follow-up, 32 months), 16 patients submitted to nephrectomy had an average decline of 0.11 ml/min/1.73m 2 /mo in EGFR, compared with an average decline of ml/min/1.73m 2 /mo in 53 patients after revascularization (P =.64). Mean preoperative renal length was 7.3 ± 1.4 cm (range, cm) in kidneys treated by nephrectomy, compared with 8.2 ± 1.5 cm (range, cm) in kidneys revascularized (P =.01). Of 9 revascularized kidneys less than 7 cm in length, all had a beneficial blood pressure response, and 4 (44%) had improved renal function. In comparison, 8 nephrectomies were performed for kidneys less than 7 cm in length. Beneficial blood pressure response was observed in 6 (75%) patients, whereas 3 (38%) patients had improvement in excretory renal function. When preoperative renal length (8.6 ± 0.3 cm) was compared with postoperative renal length (9.0 ± 0.3 cm) after revascularization of RA-OCC in 31 patients, a significant increase in renal length was observed (P =.04). The change in EGFR in relation to change in renal length was determined according to site of repair. After unilateral repair, the mean EGFR increased by 5.9 ± 8.6 ml/min/m 2, corresponding with a 0.3 ± 1.4 cm increase in renal length (r 2 = 0.54; P =.09). No correlation was found upon analysis of patients after bilateral repair (r 2 = ; P =.98). Of 16 patients with less than 10% of excretory renal function from the affected kidney, 5 (31%) Fig 4. Predicted dialysis-free survival for patients with renal artery occlusion who have improved or unimproved renal function after operation. patients underwent nephrectomy, whereas 11 (69%) patients were revascularized. Of the 11 patients revascularized, 10 (91%) had beneficial blood pressure responses (3 cured), and 5 (45%) had improved renal function. In comparison, 4 of 5 patients (80%) undergoing nephrectomy had a beneficial blood pressure response, whereas 1 (20%) had improved renal function. In the nephrectomy group, a nephrogram was demonstrated by means of preoperative studies in 5 of 22 kidneys (23%), compared with 49 of 65 kidneys (75%) in the revascularized group (P <.01). Reconstitution of the distal renal artery was present in 8 of 21 kidneys (38%) in the nephrectomy group and 48 of 60 kidneys (80%) in the revascularized group (P <.01). Although the absence of a nephrogram or distal reconstitution of the renal artery demonstrated significant associations with nephrectomy in both cases, absence did not preclude revascularization. In kidneys without nephrogram or distal renal artery reconstitution, 48% of renal arteries were successfully repaired. In fact, 8 of 21 patients (38%) with neither preoperative nephrogram nor distal renal artery reconstitution underwent operative renal artery revascularization. Six of these 8 patients (75%) had a beneficial blood pressure response, whereas 6 patients (75%) had improved renal function. Comparisons were made between the 95 patients with RA-OCC and the 302 patients with RAS (Table I). RA-OCC patients tended toward severe hypertension (mean SBP, 204 ± 31 mm Hg vs 197 ± 38 mm Hg; P =.08), required more antihypertensive medications (3.0 ± 1.1 drugs vs 2.5 ±

7 146 Oskin et al January 1999 Table V. Comparison of operative results in renal artery occlusion (RA-OCC) and renal artery stenosis (RAS) groups RA-OCC (n= 95) RAS (n = 302) P value Perioperative mortality (%) Hypertension response (%).52 Cured Improved No change 9 13 Renal function response (%)*.20 Improved No change/worsened *Preoperative serum creatinine level of 1.3 mg/dl Table VI. Results of proportional hazards regression model of follow-up survival for patients with renal artery occlusion (RA-OCC)(n = 90) Variable* Beta Standard error Hazard ratio 95% CI P value Increase in EGFR <.01 Age <.01 Preoperative SCr <.01 History of CHF Preoperative SBP *All variables significant at less than 0.10 included. CI, confidence interval; EGFR, estimated glomerular filtration rate; SCr, serum creatinine level; CHF, congestive heart failure; SBP, systolic blood pressure. Table VII. Results of proportional hazards regression model of follow-up survival for patients with atherosclerotic renovascular disease (n = 380) Variable* Beta Standard error Hazard ratio 95% CI P value RA-OCC Age <.01 CHF TIA/CVA <.01 Dialysis <.01 Preoperative SCr <.01 CI, confidence interval; RA-OCC, renal artery occlusion; CHF, congestive heart failure; TIA/CVA, transient ischemic attack/ cerebrovascular accident; SCr, serum creatinine level. 1.1 drugs; P <.01), had more severe renal dysfunction (60% vs. 35%; P <.01), had more prevalent end-organ failure (CHF [34% vs 16%; P <.01], and more were dialysis dependent [12% vs 5%; P =.03]). In addition, the 90 patients with unilateral RA- OCC required bilateral revascularization to correct hemodynamically significant stenosis in the contralateral nonoccluded kidney more frequently than the RAS patients (65% vs 51%; P =.02). During postoperative evaluation, 87% of patients in the RAS group had a beneficial blood pressure response, compared with 91% of patients in the RA-OCC group (P =.52). In the RAS group, 41% of patients had improved renal function, compared with 49% of patients with improved renal function in the RA- OCC group (P =.20). DISCUSSION As a group, hypertensive patients with RA-OCC have a high rate of renal insufficiency (88%) and a high incidence of site-specific end-organ atherosclerotic disease (95%). Despite the presence of endorgan failure and extrarenal atherosclerosis, the perioperative mortality was equivalent for renal revascularization and nephrectomy. Equivalent beneficial blood pressure response was observed after both

8 Volume 29, Number 1 Oskin et al 147 Table VIII. Change in estimated glomerular filtration rate (EGFR) of patients with renal artery occlusion (RA-OCC) by site of disease and extent of repair Number Preoperative EGFR Postoperative EGFR Adjusted EGFR of patients (mean ± SE) (mean ± SE) (mean ± SE) Unilateral RA-OCC Nephrectomy ± 9.1 (P =.04) 50.4 ± 8.4 (P =.08) 36.3 ± 5.0 (P =.57) Revascularization ± 3.1 (P =.04) 36.8 ± 3.7 (P <.01) 39.5 ± 2.1 (P =.57) Bilateral renal artery disease* Nephrectomy ± 3.2 (P =.12) 38.7 ± 4.6 (P =.02) 34.8 ± 3.6 (P =.99) Revascularization ± 1.9 (P =.12) 33.5 ± 3.1 (P <.01) 35.1 ± 2.1 (P =.99) *Includes 5 patients with bilateral RA-OCC and 49 patients with unilateral RA-OCC and contralateral renal artery disease Nephrectomy of RA-OCC with contralateral renal artery revascularization Revascularization of RA-OCC and contralateral renal artery revascularization and nephrectomy (92% vs 87%); however, only revascularization or revascularization with nephrectomy demonstrated a significant increase in EGFR. In comparison with the patients with RAS, patients with RA-OCC had more severe hypertension, an increased prevalence of CHF, decreased renal excretory function, increased prevalence of preoperative dialysis dependence, and required bilateral revascularization more frequently. Despite these differences, perioperative mortality and the follow-up survival rates were similar for patients treated for RA-OCC and patients treated for RAS. Although nephrectomy and renal artery revascularization were equally effective in the management of renovascular hypertension in the RA-OCC group, only revascularization significantly improved renal function. Neither split renal function tests nor renography were routinely obtained after renal artery repair. The absence of these studies limits conclusions regarding recovery of renal function after repair of RA-OCC, because 54 of 85 operative survivors with RA-OCC were treated simultaneously for contralateral RAS. In this instance, the absolute and relative contribution of the repaired RA-OCC to global renal function cannot be determined. Nevertheless, 21 patients treated for unilateral RA-OCC treated by revascularization demonstrated a significant increase in global excretory function when compared with preoperative renal function. Among the group with RA-OCC, a high prevalence of preoperative renal dysfunction and dialysis dependence was observed. The importance of excretory renal function in the patients with ischemic nephropathy is demonstrated by the decrease in quality and quantity of life associated with progression to end-stage renal disease. Mailloux et al 11 recently reported survival estimates for patients with uncorrected atherosclerotic renovascular disease and renal insufficiency who had progressed to dialysisdependence. Among patients with dialysis-dependent ischemic nephropathy, a 27-month median survival and 5-year survival rate of only 12% was observed. Our experience with surgical management of RA-OCC demonstrated a significant increase in estimated survival associated with improved excretory renal function after operation. A management strategy that emphasizes renal revascularization in preference to nephrectomy appears to provide the best opportunity for dialysis-free survival. In this regard, nephrectomy for RA-OCC is limited to an unreconstructible renal artery to a kidney with negligible renal function (less than 10% of total renal function). Before nephrectomy is performed for RA-OCC in patients with ischemic nephropathy, excretory renal function should be evaluated in the affected kidney. In this study, preoperative isotopic renography was used as a means of determining the relative contribution to renal function from the kidney with RA-OCC. In the case of a small kidney without a demonstrable nephrogram or reconstitution of the distal renal artery on preoperative arteriography and minimal function on renogram, we will explore the renal artery. In 40% of such cases, a normal distal vessel will be found and revascularized. Nevertheless, if atherosclerosis involves the branch vessels and renal function is negligible by means of renogram, we will proceed with nephrectomy. Using the above criteria, 11 patients with RA-OCC and minimal renal function on isotopic renogram were revascularized; 91% had beneficial blood pressure response, and 45% had improved renal function. Because blood pressure response is equivalent after nephrectomy or revascularization and improved

9 148 Oskin et al January 1999 renal function after renal artery repair appears to confer improved follow-up survival, our results support renal artery revascularization for RA-OCC. The practical value of this premise is emphasized by the natural history of atherosclerotic renovascular disease. In this regard, as many as 35% of patients with mild contralateral occlusive lesions will progress to hemodynamically severe occlusive lesions within 5 years. 2 Once nephrectomy has been performed for renovascular hypertension and contralateral renal artery disease progresses, global renal function is threatened and operative intervention may be performed for severe renal insufficiency in the presence of a lesion that is no longer reconstructible. Conversely, overzealous revascularization of a poorly functioning or nonfunctioning kidney will sometimes lead to repair of a kidney in which no functional retrieval can be achieved. In an attempt to avoid this scenario, authors have advised treatment based on distal renal artery reconstitution, presence of a nephrogram, renal length, and the results of renal biopsy, in an attempt to define characteristics that may predict operative response. The absence of distal renal artery reconstitution on preoperative angiography has been considered an indication for preferential nephrectomy. 5,7 Although the absence of reconstitution does make revascularization less likely, a patent distal renal artery was found on exploration in 48% of our patients without reconstitution. Subsequent revascularization in this setting was associated with beneficial blood pressure (83%) and renal function response (58%). The absence of a nephrogram phase during angiography has also been used as an indication for nephrectomy, 12 but 48% of the patients without a nephrogram were successfully revascularized in this study. In this setting, subsequent revascularization was associated with beneficial blood pressure (81%) and renal function response (63%). Finally, renal length of less than 8 to 9 cm has been suggested as a predictor of poor renal functional response, but recent success has been found in revascularization of small kidneys Although kidneys treated by nephrectomy did have a smaller renal size on average, we found that absolute renal length did not preclude revascularization, beneficial blood pressure response, or improved renal function after repair. In fact, of 9 patients with kidneys less than 7 cm in length, a beneficial blood pressure response was observed in all and 4 had improved renal function after renal artery reconstruction. In this latter group of improved patients, an increase from 9% to 43% in relative function of the occluded kidney was demonstrated by means of preoperative and postoperative renography. Although many issues about the recovery of renal function remain unresolved, absent distal renal artery reconstitution or nephrogram and renal length less than 7 cm did not preclude beneficial blood pressure or improved renal function response after repair of RA-OCC. Although the kidney with RA-OCC may receive sufficient collateral flow to prevent infarction, RA- OCC is frequently associated with decrease in size and changes in histology. Often, there is tubular atrophy and hyalinization of glomeruli, with a concomitant decrease in cortical thickness. 16 Zinman and Libertino 5 found preoperative renal biopsy to be a predictor of functional response. Although normal glomerular architecture supports retrieval of excretory function, hyalinization does not preclude the function benefit, because these histologic changes may be focal and biopsy material is not representative of the entire kidney. In this regard, 2 dialysis-dependent patients in this series had preoperative renal biopsies. This material demonstrated at least 60% of biopsied glomeruli senescent or hyalinized. 18 Both patients were permanently removed from dialysis after bilateral renal artery repair. Consequently, we have abandoned renal biopsy to guide renal artery repair for recovery of renal function. Our reliance upon standard cut film arteriography in patients with ischemic nephropathy requires comment. The use of a single midstream flush aortogram requires no more contrast material than that required for multiple intraarterial digital subtraction studies. In addition, information about cortical thickness and renal length and improved clarity of intrarenal artery anatomy are provided by means of standard arteriography. It is widely recognized that arteriography can aggravate renal dysfunction, especially in patients with concomitant azotemia, but we believe the risk is justified in patients with severe/accelerated hypertension and ischemic nephropathy. In these circumstances, the potential benefit derived from identification and correction of significant renovascular occlusive lesions exceeds the risk of arteriography. Finally, the simultaneous management of a significant contralateral renal artery stenosis is based on the operative strategy of complete renal revascularization at a single operation. Previous reports have shown a significant increase in EGFR after bilateral, but not unilateral, procedures for RAS associated with ischemic nephropathy In patients with RA-OCC, however, a significant increase in EGFR was observed, regardless of site of disease or extent of repair. This observation increases our enthusiasm

10 Volume 29, Number 1 Oskin et al 149 for revascularization of the RA-OCC in preference to nephrectomy. Although selection bias inherent to this retrospective report limits the power of our observations, patients with RA-OCC demonstrate equivalent blood pressure response after either revascularization or nephrectomy; however, only revascularization significantly improved excretory renal function. Although the change in global renal function caused by repair of RA-OCC cannot be defined for patients undergoing simultaneous contralateral repair of RAS, increased renal function was associated with significant and independent improvement in estimated survival. Because the lower limits of renal function retrieval are not well defined, these results support renal revascularization in preference to nephrectomy for RA-OCC. We limit nephrectomy to patients with renovascular hypertension with an unreconstructible renal artery to a kidney with negligible excretory renal function. In these latter cases, nephrectomy can provide beneficial blood pressure control without decreasing overall excretory renal function. REFERENCES 1. Dean RH, Tribble RW, Hansen KJ, O Neil E, Craven TE, Redding JF. Evolution of renal insufficiency in ischemic nephropathy. Ann Surg 1991;213: Dean RH, Wilson JP, Burko H, Foster JH. Saphenous vein aortorenal bypass grafts: Serial arteriographic study. Ann Surg 1974;180: Scheft P, Novick AC, Stewart BH, Straffon RA. Renal revascularization in patients with total occlusion of the renal artery. J Urol 1980;124: Libertino JA, Flam TA, Zinman LN, et al. Changing concepts in surgical management of renovascular hypertension. Arch Intern Med 1988;148: Zinman LN, Libertino JA. Revascularization of the chronically occluded renal artery with restoration of renal function. J Urol 1977;118: Lawrie GM, Morris GC, DeBackey ME. Long-term results of treatment of the totally occluded renal artery in 40 patients with renovascular hypertension. Surgery 1980;88(6): Lawson JD, Hollifield JH, Foster JH, Rhamy RK, Dean RH. Hypertension secondary to complete occlusion of the renal artery. Am Surg 1978; Rolin HA, Hall PM, Wei R. Inaccuracy of estimated creatinine clearance for predictors of iothalamate glomerular filtration rate. Am J Kidney Dis 1984;4: Kalbfleisch JD, Prentice RL. The statistical analysis of failure time data. New York: John Wiley and Sons; p SAS Institute. SAS Technical report P-217. In: SAS/Stat Software: The PHREG procedure, version 6. Cary (NC): SAS Institute; p Mailloux LU, Bellucci AG, Mossey RT, et al. Predictors of survival in patients undergoing dialysis. Am J Med 1988; 84: Feltrin GP, Rossi GP, Talenti E, et al. Prognostic value of nephrography in atherosclerotic occlusion of the renal artery. Hypertension 1986;8: Dean RH, Englund R, DuPont WD, et al. Retrieval of renal function by revascularization. Ann Surg 1985;202(3): Whitehouse WM Jr, Kazmers A, Zelenock GB, et al. Chronic total renal artery occlusion: Effects of treatment on secondary hypertension and renal function. Surgery 1980;89(6): Jordan WD Jr, Smith RB, Salam AA, et al. The occluded renal artery: An ongoing surgical challenge. Ann Vasc Surg 1995;9(4): May J, Ross Sheil AG, Horvath J, Tiller DJ, Johnson JR. Reversal of renal failure and control of hypertension in patients with occlusion of the renal artery. Surg Gynecol Obstet 1976;143: Hansen KJ, Starr SM, Sands RE, et al. Contemporary surgical management of renovascular disease. J Vasc Surg 1992; 16: Hansen KJ, Thomason B, Craven TE, et al. Surgical management of dialysis-dependent ischemic nephropathy. J Vasc Surg 1995;21: Hansen KJ, Benjamin ME, Appel RG, Craven TE, Dean RH. Renovascular hypertension in the elderly: Results of surgical management. Geriatr Nephrol Urol 1996;6:3 11. Submitted Jan 29, 1998; accepted Aug 11, 1998.