High-grade renal injury: non-operative management of urinary extravasation and prediction of long-term outcomes


Jean-Alexandre Long, Urology Department, Grenoble University Hospital, 38 043 Grenoble Cedex 9, France. e-mail: JALong@chu-grenoble.fr


Study Type – Therapy (outcomes)

Level of Evidence 2b

What's known on the subject? and What does the study add?

High-grade renal trauma seems to be eligible for conservative management. Ureteric stent placement raises issues about its usefulness and its timing. Predictive factors of post-trauma function and surgery need to be known.

Urinary extravasation is not associated with poor functional outcome. Ureteric stenting is needed only in case of sepsis and ureteric clot retention. The only independent predictor of long-term renal function is the importance of devascularised renal fragments.


  • • To predict the outcomes of a non-operative approach to managing urinary extravasation after blunt renal trauma.


  • • A prospective observational study was conducted between January 2004 and October 2011. First-line non-operative management was proposed for 99 patients presenting with a grade IV blunt renal injury according to the revised American Association for the Surgery of Trauma (AAST) classification. Among them, 72 patients presented with a urinary extravasation.
  • • Management and outcomes were recorded and compared between patients presenting and those who did not present with urinary leakage. Relative postoperative renal function was assessed 6 months after the trauma using dimercapto-succinic acid renal scintigraphy.
  • • Predictors of the need for endoscopic or surgical management and long-term renal function were evaluated on multivariate analysis.


  • • Among patients with urinary leakage, endoscopic ureteric stent placement and open surgery were required in 37% and 15%, respectively.
  • • On multivariate analysis, fever of >38.5 °C and ureteric clot obstruction were independent predictors of the need for ureteric stent placement. The only predictor of open surgery was the percentage of devitalised parenchyma.
  • • Long-term renal function loss was correlated to the percentage of devitalised parenchyma and associated visceral lesions. Urinary extravasation did not predict surgical intervention or long-term renal function loss.


  • • Urinary extravasation after blunt renal trauma can be successfully managed conservatively and does not predict long-term decreased renal function or surgery requirement.
  • • A devascularised parenchyma volume of >25% predicts a higher rate of surgery and poorer renal function.

American Association for the Surgery of Trauma (classification)


interquartile range


odds ratio


Renal Injury Staging Classification.


The management of high-grade blunt renal trauma is still controversial and has developed over time towards a conservative approach [1]. Non-operative management consists of monitoring patients closely to detect complications early. In case of complication, the patient is then managed as conservatively as possible using minimally invasive techniques first [2].

Management of urinary extravasation raises issues between an early ureteric stent placement and close monitoring [3]. Our hypothesis is that non-operative treatment can be used to manage high-grade blunt renal injuries. The surgeon should identify treatment outcome predictors to quickly recognise cases where a more aggressive strategy is needed.

The present study evaluated the outcomes of patients with a high-grade blunt renal injury managed non-operatively (grade IV according to the 2011 Revised Renal Injury Staging Classification [RISC][4,5] or IV and V according to the 1989 American Association for the Surgery of Trauma [AAST] classification [6]).


Patients admitted to our centre with a high-grade blunt renal trauma between January 2004 and October 2011, were considered for inclusion in this study. Patient data were prospectively collected in an Institutional Review Board-approved renal trauma database.

Renal injury was graded based on CT findings, and classified according to the recent revision of the AAST classification (RISC) [5]. According to RISC, patients with post-traumatic main renal artery thrombosis were excluded. Grade IV injuries were defined as: laceration through the parenchyma into the collecting system, vascular segmental vein or artery injury as well as any injury involving the collecting system (including renal pelvis laceration and ureteric pelvic disruption).

Overall, the database included 273 patients. Only patients with grade IV renal injury (99 patients) were included. Among them, 72 patients presented with urinary leakage on the first or subsequent CT with various rates of renovascular injuries. Consequently, 27 patients presented with isolated segmental vein or artery injury, leading to a limited parenchymal devascularisation. Among the patients with urinary leak, 16 and six also presented with a parenchymal devascularisation involving at least 25% and 50% of the kidney, respectively.


Four senior urologists managed all of the patients according to our institutional first-line non-operative treatment protocol. In the Emergency Room, patients who had active bleeding on CT were referred for immediate angioembolisation, regardless of their haemodynamic status. Emergency surgery was performed only when immediate resuscitation failed. In the Intensive Care Unit, non-operative management consisted of bed rest, hydration and analgesia. No systematic antibiotics were given even when there was urinary leak. Antibiotics were only given in the case of a fever of >38.5 °C after laboratory evaluation that included urine and blood cultures, as well as a chest and abdominal CT.

Retrograde ureteric stent placement was performed in the case of pain caused by ureteric clot obstruction (suspected on the basis of clinical findings and confirmed with CT) or hyperthermia of >38.5 °C not resolving with antibiotics after excluding a pulmonary cause. A stent was also placed in the case of persistent, significant urine leakage on subsequent CT defined by an increasing urinoma or a persistent absence of ureteric opacification.

Follow-up CT imaging was systematically 3–5 days after the trauma. Earlier imaging may also have been performed based on the patients' clinical course.

Delayed surgical intervention was performed if there was persistent sepsis despite antibiotics as well as urinoma and collecting system drainage.


Early evaluation consisted of recording trauma characteristics and initial hospital management (angioembolisation, transfusion, length of stay and endoscopic or surgical management).

Long-term postoperative renal function based on DMSA renal scintigraphy was performed 4–6 months after the trauma. Clinical outcome and blood pressure were evaluated. Relative renal function after nephrectomy was zero. In all, 64 patients completed a 6-month follow-up.


Data are presented as the median and interquartile range (IQR) and frequency (%) for continuous and categorical variables, respectively. Bivariable comparisons were performed using the independent Mann–Whitney U-test for continuous data, and chi-square or Fisher's exact test for categorical data, as appropriate.

First, the patients presenting with a major renal injury and urinary leakage were compared with those without urinary leakage. Multiple linear regression was used to study potential predictors of long-term relative renal function for all patients presenting with a grade IV blunt renal trauma. Multiple logistic regression was used to identify predictors of immediate or delayed intervention and, among patients with urinary leak, to identify predictors of ureteric stent placement and surgery.

Moreover, the study population was subdivided according to the proportion of devitalised parenchyma, with a threshold of 25% of the entire parenchyma.

Variables attaining a P= 0.2 in the univariate analysis were included in the multivariate analysis.



Demographic data, trauma characteristics and patient outcome are summarized in Table 1.

Table 1. Patients' demographic and trauma characteristics, and outcome of Grade IV with or without urinary extravasation
VariableUrinary extravasationNo urinary extravasation P
  1. Values are expressed as median (IQR) for continuous variables and frequency (percentage) for categorical variables.

No. of patients7227 
Median (IQR) age, years22.9 (18.3–35.7)20.8 (17.2–35.6)0.48
N (%):   
 Male gender61 (86)7 (28)0.21
 Right laterality30 (42)16 (59)0.95
  Deceleration injury19 (26)11 (41)0.31
  Direct Impact65 (90)24 (89)0.93
  Gross haematuria68 (94)19 (70)<0.001*
  Haemodynamic instability5 (7)7 (26)0.017*
Median (IQR) haemoglobin, mg/dL121 (105–137)109 (94.8–123)0.003*
Associated lesions, n (%)34 (47)17 (61)0.009*
Median (IQR) uro-haematoma diameter, cm10 (6–14)8 (2–12)<0.04*
N (%)   
 Percentage of devascularised parenchyma:  0.049*
  0–2550 (69)7 (26)
  25–5015 (21)11 (41)
  >507 (10)9 (33)
  Angioembolisation14 (19)11 (41)0.049*
  Surgical intervention11(15)8 (30)0.52
  JJ stent27 (37)0
  Transfusion30 (42)16 (59)0.08
Median (IQR):   
  CT number2 (2–3)2.5 (2–4.2)0.34
  Length of stay, days15 (10–21)14 (10–24.2)0.72
  Relative renal function, %39 (20–44)30 (5–40)0.16

During their hospital stay, 27 (37%) patients required JJ stent placement for the following indications: persistent, significant urinary leak (13 patients), suspected sepsis (nine), pain related to clot obstruction (four), and acute renal failure in a solitary kidney patient (one).

In all, 11 patients underwent surgery, including nephrectomy in six cases (five sepsis cases despite antibiotics and drainage; one haemorrhage despite angioembolisation), renorrhaphy in two patients (for persistent leakage despite stent placement), surgical drainage in three patients (for a urinoma/haematoma).

Conservative management was possible in 92% of cases, including 14 patients requiring angioembolisation (Table 2).

Table 2. Management characteristics
 Urinary extravasation, n (%)No urinary extravasation, n (%)
Non-operative61 (85)19 (70)
Conservative66 (92)21 (78)
Nephrectomy6 (8)6 (22)
Conservative procedures:  
 Ureteric stent270
 Uro-haematoma drainage31


There were no significant differences between the two groups for age, gender, laterality, number of CTs, median length of stay, transfusion rates and management, excluding ureteric stenting, which was performed only in the case of urinary extravasation.

Patients with urinary leakage had a slightly lower rate of associated lesions and were more likely to be haemodynamically stable when compared with those with no leak. They also showed a higher rate of gross haematuria at presentation and a higher haemoglobin level. Median, long-term relative renal function was not significantly different between the two groups (39% vs 30%, P= 0.16; Table 1).


On multivariate analysis, the only independent predictors of ureteric stent placement were suspected sepsis (odds ratio [OR] 40.1, P= 0.002) and ureteric clot obstruction (OR 20.2, P= 0.01; Table 3).

Table 3. Predictors of ureteric stent placement in cases of urinary extravasation
 Univariate analysisMultivariate analysis
JJ stent (n= 27)No (n= 45) P OR (95% CI) P
  1. Values are expressed as median (IQR) for continuous variables and frequency (%) for categorical variables.

Median (IQR) age, years27 (16.0–42.1)22.6 (18.2–33.9)0.89  
Male gender, n (%021 (78)41 (91)0.133.84 (0.56–26.3)0.17
Right laterality, n (%)10 (37)20 (44)0.87  
Haemodynamic instability, n (%)4 (15)2 (5)0.162.38 (0.16–35.7)0.53
Median (IQR) haemoglobin, mg/dL126 (117–138.5)118 (98.5–130)0.0891.02 (0.81–1.23)0.33
Associated lesions, n (%)15 (55)20 (44)0.38  
Median (IQR) uro-haematoma diameter, cm10 (6–14)11 (6–15)0.51  
Percentage of devascularized parenchyma, n (%)  0.140.58 (0.04–7.69)0.68
 <2517 (34)33 (66)   
 25–509 (56)7 (44)   
 >501 (17)5 (83)   
Symptoms in follow-up, n (%):     
 Fever >38.5 °C9 (33)1 (2) <0.001* 40.1 (3.6–333.2) 0.002*
 Clot obstruction5 (18)1 (2) 0.001* 20.2 (2.17- 250.1) 0.01*

The only independent predictor of kidney surgical intervention in patients with urinary leakage was a percentage of devitalised parenchyma of >25% (OR 13.7, P= 0.02). Patients undergoing renal surgery were more likely to have associated visceral lesions (82% vs 43%, P= 0.013). Interestingly, persistent urinary extravasation was not significantly associated with surgery requirement (36% vs 13%, P= 0.11; Table 4).

Table 4. Predictors of surgery after urinary extravasation
 Univariate analysisMultivariate analysis
No surgery (n= 61)Surgery (n= 11) P OR (95% CI) P
  1. Values are expressed as median (IQR) for continuous variables and frequency (percentage) for categorical variables.

Median (IQR) age, years23.1 (18.2–38.2)22.3 (18.2–28.4)0.35  
Male gender, n (%)54 (88)8 (73)0.193.9 (0.4–41.6)0.24
Right laterality, n (%)26 (43)4 (36)0.69  
Haemodynamic instability, n (%)5 (8)1 (9)0.88  
Median (IQR) haemoglobin, mg/dL126 (101.5–138.7)120 (99.7–126.2)0.180.99 (0.97–1.03)0.93
Associated lesions, n (%)26 (43)9 (82)0.013*5.2 (0.68–38.5)0.11
Median (IQR) uro-haematoma diameter, cm10 (3–12.2)10 (8.5–15)0.71  
Percentage of devascularized parenchyma >25%, n (%)  0.025*13.7 (1.46–125)0.02*
 <2546 (75)4 (36)   
 25–5012 (20)4 (36)   
 >503 (5)3 (27)   
Fever >38.5 °C6 (10)3 (27)0.145.5 (0.67–45)0.11
Persistent urinary extravasation >7 days, n (%)8 (13)4 (36)0.086.1 (0.68–54.6)0.11

On multivariate analysis, surgical intervention was predicted by a percentage of devitalised parenchyma of >25% (P= 0.049) and the presence of associated lesions (P= 0.031). Urinary leakage was not an independent predictor of an operative management (Table 5).

Table 5. Multivariate analysis to predict postoperative renal function and need for surgical intervention
Prediction of relative renal function
 Regression coefficient P 95% CI
Urinary extravasation1.130.59−3.17; 5.4
Percentage of devascularised parenchyma >25%−9.82<0.001*−14.4; −5.2
Gender1.830.33−1.99; 5.66
Age0.180.16−0.07; 0.44
Haemodynamic instability at presentation−2.720.19−6.8; 1.45
Associated lesions3.670.14−1.18; 8.5
Operative management vs non-operative−12.3<0.001*−16.4; –8.1
Prediction of kidney surgical intervention
 OR P 95% CI
Urinary extravasation1.010.930.29; 3.35
Percentage of devascularised parenchyma >25%3.380.049*1.01; 12.3
Associated lesions4.380.031*1.26; 20.4
Haemodynamic instability at presentation2.450.180.66; 8.9


In the entire study population, operative management (P< 0.001), percentage of devitalised parenchyma of >25% (P< 0.001), and uro-haematoma diameter (P= 0.036) were independent predictors of long-term renal function, whereas urinary extravasation was not (Table 5).


Urinary extravasation was more likely among patients without extended devascularization (86% vs 54%, P< 0.001). Need for intervention (10% vs 27%, P= 0.033), length of stay (12 vs 18 days, P= 0.008) and relative renal function (40% vs 10%, P< 0.001) were associated with devascularization (Table 6).

Table 6. Influence of parenchymal devascularisation on outcomes
 Devascularisation <25% (n= 58)Devascularisation >25% (n= 41) P
  1. Values are expressed as median (IQR) for continuous variables and frequency (percentage) for categorical variables.

Median (IQR) age, years22.8 (12.4–41.2)22.0 (18.6– 28.3)0.33
Male gender, n (%)49 (84)33 (80)0.6
Right laterality, n (%)22 (38)20 (49)0.28
Mechanism, n (%):   
 Deceleration injury10 (20)20 (57)<0.001*
 Direct impact51 (94)38 (97)0.47
Presentation, n (%):   
 Gross haematuria54 (96)33 (85)0.09
 Haemodynamic instability4 (7)8 (22)0.043*
Median (IQR) haemoglobin, mg/dL127.5 (115–138)108 (86.2–123.7)<0.001*
Associated lesions, n (%)22 (38)29 (71)0.001*
Median (IQR) uro-haematoma diameter, cm8 (4–12)11 (9–15)0.07
Urinary extravasation, n (%)50 (86)22 (54)<0.001*
Management, n (%):   
 Angioembolisation6 (10)19 (46)<0.001*
 Surgical intervention6 (10)11 (27)0.033*
 JJ stent17 (29)10 (24)0.58
 Median (IQR) CT number2 (2–3)2 (2–3.7)0.97
 Median (IQR) length of stay, days12 (9–18.5)18 (11.5–26)0.008*
 Transfusion, n (%)30 (42)23 (57)0.09
 Median (IQR) relative renal function, %40 (35–45)10 (5–37.5)<0.001*

There was no de novo arterial hypertension at the last follow-up.


Managing high-grade blunt renal trauma remains a controversial issue. To standardise management, the AAST created the RISC [6]. It provided a standardised classification of renal injuries. A recent revision in 2011 was published by Buckley and McAninch [5] to establish a more rigorous definition of severe renal trauma. The main difference between the two classifications was to limit grade V injuries to main renal artery and/or vein injuries. Since then, grade IV injuries include all injuries involving the collecting system and segmental arteries and/or veins. In previous studies, the AAST classification was validated as a predictor of trauma outcomes, including renal function [7,8]. However, very few studies evaluated the functional outcomes of urinary extravasation [3].

This prospective study aimed to evaluate the management of collecting system injuries and to predict the outcome of blunt renal trauma. To our knowledge, this represents the largest study to have ever included patients with urinary leakage after blunt renal trauma initially managed with a non-operative approach.


The main issue in this situation is to decide whether or not a ureteric stent is required and when it should be placed. A ureteric stent was placed in the present series when needed, as suggested by Alsikafi et al. [3]. Overall, it was required in 37% of patients. It is unclear whether persistent urinary leakage is an appropriate indication of urinary stenting. Indeed, 48% of the patients denied having any symptoms at the time of stent placement. It can be speculated that in some cases, especially early in our experience, stenting could have been avoided, and an exclusively non-operative approach used.

Time from trauma to ureteric stenting raises issues. In the present series, CT revaluation was systematically performed after 3–5 days and a JJ stent was individually discussed based on the significance of urinary extravasation. On multivariate analysis, predictors of ureteric stent placement were the symptoms (clot obstruction and fever). Nonetheless, the rationale for an endoscopic procedure in case of suspected infection is questionable, as endoscopic manoeuvres may contaminate the urinoma [9]. In our experience, few patients were discharged with an active urinary extravasation without stenting without experiencing adverse effects. These observations are now encouraging us to continue monitoring these patients and extend follow-up unless a sepsis or a clot obstruction occurs. This assumption should be confirmed by a randomised study and the present study's design does not allow for conclusions to be drawn about the optimal timing of ureteric stent placement.

Surgery was required in only 15% of patients presenting with a urinary extravasation. When surgery was necessary, very few conservative procedures were performed. These findings are consistent with previous series showing that a conservative approach usually leads to a radical treatment in case of failure [10,11]. However, it is a good indication for renorrhaphy, which was performed in two patients in our series.


On multivariate analysis after adjustment for confounding, the independent predictors of renal function were renal surgery and the size of devitalised fragments. This difference was found with the low threshold of 25% of renal parenchyma devascularisation. Urinary leakage did not represent an independent predictor of poor functional outcome.

Renal surgery was an independent predictor of a worse late renal function, which is likely to be related to the non-conservative type of surgeries performed. Indeed, non-operative management failures include severe injuries leading to unavoidable total nephrectomy [12].

The present study assessed renal function based on DMSA renal scans that seem more reliable than calculated estimated GFR using the Modification of Diet in Renal Disease (MDRD) equation, especially in young patients with two healthy kidneys [13,14]. We confirmed in a previous study that the AAST classification was able to predict renal function according to this evaluation [8].

The overall good prognosis of collecting system lesions compared with distal vasculature lesions encourages us to be the most conservative in these cases.


When separating the population into two subgroups with a threshold of 25% of devascularised parenchyma, we found that the two cohorts were different in terms of trauma mechanism. The group with isolated vascular lesions had a higher rate of deceleration injury, similar to the rate found with main renal artery lesions (current grade V of the revised AAST classification). An important point to note is that surgical intervention was needed more frequently and relative function was significantly decreased.

Given these findings, we think that the RISC could be improved by separating grade IV injuries into two groups accordingly to the importance of devitalised segments [15]. A threshold set at 25% of the entire kidney was shown in the present study to be a predictor of long-term renal function.


In the present study, we showed that non-operative management was a suitable option for a more conservative approach to managing high-grade renal injuries. As long as patients were strictly monitored and angiography was available to perform angioembolisation when needed, it provided results that were comparable with the results in previous published series [16,17]. We think that angioembolisation is a suitable alternative to surgical exploration, allowing a conservative approach even in the case of active bleeding [18]. However, decreased renal function due to endovascular obliteration was not evaluated and may depend highly on the number and calibre of obliterated vessels, as well as the experience of the radiologist.

In the present study, we showed that both the need for surgery and renal function were not predicted by urinary extravasation. Consequently, we would recommend managing these traumas conservatively and placing a ureteric stent if needed in cases of a ureteric clot obstruction and sepsis not resolving with antibiotics. The percentage of parenchymal devascularisation can help to predict both functional and surgical outcomes.

The main limitation of the present study is the few patients who completed the functional 6-month evaluation (65%). This is explained by the fact that the population consisted primarily of young, active subjects from other regions and countries. Another significant limitation is that the classification was not reviewed by a radiologist ‘blinded’ to the patient outcome. However, the revised version of the AAST classification regroups all collecting system injuries as grade IV lesions. As this lesion is easy to detect, we can consider it as reproducible.

Another limitation concerns the comparison between urinary and non-urinary extravasation, as both lesions are usually associated. Not surprisingly, the few patients without urinary extravasation reflects the fact that most blunt traumas lead to combined vascular and urinary lesions. Isolated urinary injuries are uncommon. However, the multivariate analysis was conducted among all grade IV traumas, providing a reliable evaluation of outcome predictors.

Furthermore, the present series involved lower velocity injuries than other series, with a higher proportion of traffic accidents than sports injuries. These findings cannot be applied to penetrating injuries. Finally, the limited follow-up in the present series could not provide valid data about delayed arterial hypertension.

In conclusion, urinary extravasation after blunt renal trauma can be successfully managed conservatively and does not predict long-term renal function and surgery requirement. A devascularised parenchyma volume of >25% predicts a higher rate of surgery and poorer renal function.


The authors thank Laura Miller for her support in the translation.


None declared.