Renal Function Outcomes Following Selective Angioembolization for Iatrogenic Vascular Lesions After Partial Nephrectomy

Introduction

Current American Urological Association guidelines encourage nephron-sparing surgery as opposed to radical nephrectomy based on the reported increased risk of cardiac events and chronic kidney disease (CKD) after radical nephrectomy. ^1,2 Partial nephrectomy (PN) is associated with an increased risk of complications, however, including iatrogenic vascular lesions (IVL), which can be a cause of significant morbidity. ³ IVL, which includes both renal artery pseudoaneurysm and arteriovenous fistula, are reported in 0.4% to 4.1% of patients after PN. ^{4 –8} Because of their low incidence, risk factors and outcomes associated with these lesions have not been widely reported.

With an increased emphasis on nephron-sparing surgery along with the ever increasing detection of small renal masses, IVL may become an increasingly encountered complication. ⁴ The vast majority of IVL are managed with angioembolization (AE), which raises the concern of possible added renal injury in this population. For this reason, many centers use selective AE for these lesions in an attempt to conserve renal function. Studies reporting renal function after AE are small, however, with the majority consisting of case reports or small case series. ^5,6 Furthermore, there is a scarcity of literature reporting renal function outcomes beyond the immediate post-treatment period. Within the current study, renal function outcomes for the short and intermediate term were examined before and after PN and AE. Additional factors thought to contribute to renal function outcomes including diabetes and evidence of CKD were also examined. We report the clinical and renal function outcomes in one of the largest series compiled for IVL treated with AE after PN.

Patients and Methods

All patients who underwent PN from 2002 to 2012 were identified at our institution. Those who had AE after PN were selected. Presenting symptoms were identified from hospital and emergency department records. Laboratory values including serum creatinine and hematocrit were recorded before and after PN and AE. eGFR was calculated with the Cockroft-Gault equation using patient's weight, sex, and serum creatinine level. CKD stage was assigned based on the National Kidney Foundation classification. ⁹ Whether a patient underwent a contrast-enhanced CT before AE was also determined. Demographics and comorbid conditions including diabetes, obesity, and CKD classification were recorded. Nephrometry scoring was performed using the R.E.N.A.L (radius; exophytic/endophytic; nearness; anterior/posterior; location) nephrometry score developed by Kutikov and Uzzo. ¹⁰

Results

A total of 849 PN were performed in the study period, with IVL developing in 28 (3.3%) patients. Patient demographics for those in whom an IVL developed are shown in Table 1. The mean age was 55.5 years. At presentation, seven (25%) patients had diabetes mellitus (DM), and three (11%) had significant renal impairment (CKD stage ≥3). Patient symptoms and characteristics before AE are shown in Table 2. The mean drop in hematocrit after PN prior to AE was 6.6±9.9%. Eight (28%) patients needed transfusion. Gross hematuria was a presenting symptom in 20 (71%) patients, flank pain in 10 (36%), and symptomatic anemia (hypotension, tachycardia, or dizziness) in 10 (36%). Contrast-enhanced CT was used to make the diagnosis in nine (32%) patients. The presence of gross hematuria was associated with bypassing a contrast-enhanced CT and proceeding to AE (P=0.01). All patients had identifiable vascular lesions at the time of AE and underwent selective coil embolization. Four (14%) patients needed a second embolization for continued bleeding from an unrecognized IVL at the first AE.

Renal function outcomes are reported in Table 3. Paired changes in eGFR before and after PN were significantly decreased (P<0.01) at a median of 6.5 days after PN, while paired changes before and after AE were not significant at either short-term (mean 2.8 days) or intermediate-term (mean 362 days) follow-up (P=0.50). After PN, 11 patients experienced progression in CKD classification. After AE, four patients experienced CKD progression; however, three demonstrated improved CKD classification. For those with progression after AE, three patients progressed from CKD stage 1 to 2 and one from stage 2 to 3. The mean paired decrease in eGFR for these patients was 7.9 mL/min/1.73 m². Having DM before PN resulted in a greater decrease in eGFR after PN (−23.7 mL/min/1.73 m²) vs not having DM (−13.9 mL/min/1.73 m²), but this was not significant (P=0.25). DM did not correlate with a significant difference in eGFR after AE (P=0.35).

R.E.N.A.L. nephrometry scores where available for 25 of 28 tumors. The tertile breakdown is shown in Table 1. The median nephrometry score was 8 (interquartile range 7–9). Five tumors were completely exophytic and four were completely endophytic. Twenty-one (84%) tumors were adjacent to the renal hilum. Seventeen tumors were <4 cm, 7 were between 4 and 7 cm, and 1 was >7 cm in diameter.

Discussion

The exact method by which IVLs develop after PN is not known precisely. It has been speculated that these result from lacerated branches of the renal artery that initially stop bleeding because of clot formation but rebleed because of clot breakdown and fistulization to the extra vascular space. This can result in hematuria if a communication with the collecting system occurs or a hematoma if a communication to the extrarenal space ensues.

This complication is of concern in the post-PN population because AE is generally indicated and carries the theoretical risk of further renal compromise because of the contrast load with embolization. There was no statistical decrease in eGFR after AE in our study, however (Table 3), which may be the result of using selective AE in the majority of cases (Fig. 1). Using this method, only the contributing branches to the bleeding vessels are embolized at the time of angiography and the noninvolved renal tissue is spared. Several studies have similarly reported preserved renal function in this setting, although at least one study disputes this. ¹¹ Shapiro and coworkers ⁵ reported no decrease in renal function after AE in six patients, although specific data were not given. Hyams and colleagues ⁸ reported the largest series of IVL after minimally invasive PN and noted CKD stage progression in 4 of 20 (20%) patients after AE, although changes in eGFR were not reported. Our series similarly showed 4 of 28 (14%) patients with CKD stage progression, although this was tempered by CKD stage improvement in 3 cases. Furthermore, the average decrease in eGFR in those with stage progression was only 7.9 mL/min/1.73 m² almost 1 year after treatment. These data would suggest that the eGFR for these patients was initially near a CKD stage cutoff and did not likely represent a clinically significant change.

OPEN IN VIEWER

FIG. 1.

Selective angioembolization of a characteristic iatrogenic vascular lesion (pseudoaneurysm) after partial nephrectomy.

Renal function after AE was further evaluated based on several variables. DM is a known adverse risk factor for the development of contrast-induced nephropathy, while CKD is considered one of the biggest risk factors for this type of renal injury. ¹² Those patients who presented with DM were compared with those without DM and were found to have no worse renal function outcomes after AE (P=0.35). A similar comparison was made before and immediately after PN, and although there was a greater decrease in eGFR in patients with DM, this did not reach statistical significance (P=0.25). Again, the statistical calculations are limited by small numbers, but perhaps a difference would be uncovered with a larger population. Preexisting CKD (defined as stage ≥3) only existed in two patients. Its effect on renal function outcomes could not be examined because of its low incidence, although no patients within this group experienced stage progression.

One clinical dilemma in the setting of a suspected IVL after PN is whether a CT scan is necessary to identify the source of hemorrhage. In our institution, CT was not routinely obtained if a patient re-presented with gross hematuria after PN because it was thought that an IVL was the most likely diagnosis. This practice was implemented in an attempt to avoid a double contrast load that would come with performing both a CT and AE. Indeed, there is evidence to suggest contrast volume may be related to contrast-induced nephropathy. ^13,14 Further, an IVL was identified in every case of gross hematuria that proceeded directly to AE. However, this study did not evaluate the overall incidence of gross hematuria after PN and therefore was not equipped to determine a correlation between the presence of gross hematuria and IVL after PN. A study examining this relationship would be of interest because it could determine the likelihood of an IVL given gross hematuria and may help to better guide clinical practice.

The development of IVL may be related to several factors, one being tumor location. It has been hypothesized that endophytic and central lesions pose a greater risk for the development of an IVL. Multiple series have reported IVL in exophytic and peripheral tumors, however. ^7,15 In an effort to describe tumor complexity in our series, R.E.N.A.L. nephrometry scoring was performed. The median nephrometrey score for this series was relatively high, with all but three tumors in the intermediate or high complexity category. Of note, 84% of the tumors were considered to be adjacent to the renal hilum. Although nephrometry scoring was not available for our entire PN cohort, these data may suggest that IVL is more likely to occur in more complex tumors, especially those in close proximity to the hilum.

Finally, this study reinforces the observation that hemorrhage from an IVL often occurs after hospital discharge. The median time to presentation after PN was 10.2 days in our series, which is slightly lower than reported studies, but within the reported range (8.5–14.5 days). ^5,8,11 The wide variation in time to presentation (1–33 days) demonstrates the need for awareness of this complication up to a month after PN. As in all other published series, hematuria was the most common presenting symptom (71%). Other presenting symptoms included flank or abdominal pain, symptomatic anemia, or a combination of both. The variation in presenting symptoms and time to presentation emphasizes the need for a high level of suspicion for any patient presenting with new or unexplained symptoms in the postoperative period after PN.

Several limitations are worth discussing. This study is retrospective and subject to selection bias such that only AEs performed for IVL at our institution were included. Patients who sought follow-up elsewhere could not be identified, and therefore the true incidence may be slightly under-reported. Further, because of the limited number of AEs, significant comparisons between groups (open vs laparoscopic vs robot-assisted PN) were not made. Also, characteristics of the PN population were not known (such as method of closure and nephrometry score), and therefore factors predicting the development of AE could not be identified. Last, hematuria after PN was not specifically studied, such that the association between hematuria and the development of an IVL in the postoperative period could not be made. Future studies examining these factors and their relationship to the development of IVL would be of benefit.

Conclusion

Selective AE for IVL after PN is highly efficacious and confers no significant additional impairment in renal function with short- or intermediate-term follow-up. Because of the low incidence of IVL, the evaluation of the comorbidities that could affect renal function after AE was limited. Larger studies are needed to determine the risk factors associated with IVLs after PN.