- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Nephron-sparing surgery has been accepted as the standard treatment for small (≤4 cm) renal tumours in appropriately selected patients and is increasingly being performed at many centers worldwide. The increased experience with laparoscopic renal surgery has meant that its use has now been extended to the treatment of larger and more complex lesions [1–4]. Oncological outcomes for laparoscopic partial nephrectomy (LPN) are comparable with radical nephrectomy as well as with open partial nephrectomy (OPN) [5–7].
Complications of LPN have been well described in the literature and are similar to those reported for OPN. Bleeding, one of the most serious complications after LPN, ranges from 4.2 to 6% including both immediate postoperative and delayed haemorrhage. The data suggests patients undergoing LPN for centrally located tumours are at greater risk of bleeding [4,8,9]. A small number of studies have focused on defining delayed haemorrhage after LPN or OPN. The management of delayed bleeding has been described and ranges from observation to selective angioembolization (SAE) and in rare cases, completion nephrectomy. The present study was undertaken to define the frequency, clinical presentation and angiographic findings of delayed haemorrhage after LPN. A secondary analysis was performed to determine the impact of this complication on renal function. The indications and outcome for SAE are also discussed.
PATIENTS AND METHODS
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Data were prospectively collected from 640 LPNs performed between August 1993 and May 2009. The procedures were performed by three experienced laparoscopic surgeons using a standard transperitoneal technique. Selected LPNs were performed off-clamp and are included in our analysis. The indication for surgery was a single renal tumour deemed amenable to resection via LPN as identified on axial imaging, either CT or MRI.
Delayed haemorrhage was defined as ‘gross haematuria occurring ≥7 days after LPN and requiring medical assessment and intervention’. Of the 13 patients that were identified, characteristics from the perioperative and postoperative period recorded in our database were reviewed. These included age, sex, medical comorbidities, tumour size, tumour location, warm ischaemia time (WIT), use of adjunct haemostatic agents, method of renal parenchymal repair, tumour pathology, serum creatinine level, haemoglobin and haematocrit levels, estimated blood loss (EBL), number of units transfused, length of hospitalization, and time to clinical presentation with delayed haemorrhage. Other data collected included radiographic findings from patients with gross haematuria at time of presentation and at 6-month follow-up. The angiographic findings were obtained and reviewed with an interventional radiologist. Vascular lesions were identified according to type and location of the segmental artery involved.
The transperitoneal approach used has been previously described [10,11]. Selected cases were performed off-clamp. Indications for off-clamp LPN are patients with exophytic lesions and favorable (i.e. anterior, lower pole) location. In cases where hilar control was obtained, the renal vessels were clamped individually using bulldog clamps. In all cases, the tumour was excised using cold scissors. Renal parenchymal repair was similar for all cases. However, the use of adjunct haemostatic agents varied according to surgeon preference. Any actively bleeding vessels seen in the tumour resection bed were clipped or oversewn. In half of the identified cases, the capsule was approximated over Surgicel bolsters (Johnson & Johnson, New Brunswick, NJ, USA) using 2-0 polyglactin 910 sutures. In the other half of the cases, the argon beam coagulation (ABC) as well as FloSeal (Baxter, Deerfield, IL, USA) were used as part of the haemostasis regimen. The renal parenchyma was re-approximated with running 2-0 polyglactin 910 suture in all cases.
Selective angioembolization was performed by an experienced interventional radiologist using a transfemoral approach. The feeding arteries of the bleeding site were identified and embolized with endovascular platinum microcoils. Follow-up included postembolization assessment of haemoglobin and creatinine trends before discharge from the hospital. Long-term follow-up was part of the oncological protocol, with initial assessments at 1, 6 and 12 months and imaging obtained at 6 and 12 months. Laboratory values including complete blood count and metabolic panel were obtained at each follow-up visit. Successful treatment was defined as ‘clinical and radiographic absence of bleeding and stable renal function’.
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Delayed haemorrhage after LPN requiring SAE occurred in 13 (2%) patients. Angiographic findings showed renal artery pseudoaneurysm (RAP) in 10 patients, and contrast extravasation in three patients, two of whom had an arteriovenous fistula (AVF). Of the patients with delayed haemorrhage, three were female and 10 were male, with an overall median (range) age of 64.7 ± 8.5 (49–78) years. The mean (range) pathological tumour size was 2.7 ± 0.79 (1.7–4) cm. The tumour location was central and endophytic in five patients, whereas the remaining eight patients had peripheral or exophytic lesions. The mean (range) WIT was 28.2 ± 12.1 (0–50) min. In comparison, the mean (range) tumour size and mean (range) WIT in patients without haemorrhage were 3.3 ± 2.0 (0.4–13.5) cm and 14.3 ± 13.9 (0–58) min, respectively. There was no significant difference in mean tumour size: P= 0.31. However, WIT was found to be significantly higher in the group of patients who experienced delayed haemorrhage: P < 0.001. Of the 640 LPNs, 68 (10.6%) were performed off-clamp, of which one (1.5%) had delayed bleeding (Table 1).
Table 1. Comparison of patients who underwent LPN with and without delayed postoperative haemorrhage
| ||No haemorrhage||Delayed haemorrhage||P|
|Patients, n||627|| 13||–|
|Age, years, mean ±sd (range)|| 59 ± 11.7 (21.4–84)|| 64.7 ± 8.52 (49–78||0.084|
|EBL, mL, mean ±sd (range)|| 308 ± 290.2 (25–2000)||403.8 ± 321.1 (50–1000)||0.26|
|WIT, min, mean ±sd (range)||14.3 ± 13.9 (0–58)|| 28.2 ± 12.1 (0–50)||<0.001|
|Tumour size, cm, mean ±sd (range)|| 3.3 ± 2.0 (0.4–13.5)|| 2.7 ± 0.79 (1.7–4)||0.31|
For the patients without delayed haemorrhage, mean (range) EBL at LPN was 308.2 ± 290.2 (25–2000) mL. The mean (range) EBL for patients with delayed haemorrhage was 403.8 ± 321.1 (50–1000) mL (P= 0.26) with the highest EBL occurring in a LPN that was performed without hilar control (Table 2). None of the patients with delayed haemorrhage required perioperative blood transfusion and serial postoperative haemoglobin, haematocrit and electrolyte levels were stable before discharge. The mean (range) length of hospital stay after the LPN was 3.3 (2–6) days. Final pathological diagnosis showed RCC in 12 patients (nine clear-cell, two papillary and one chromophobe) and mixed epithelial stromal tumour in the remaining patient. The surgical margins were negative in all the patients later presenting with delayed haemorrhage.
Table 2. Clinical and angiographic characteristics of patients with presentation of delayed haemorrhage after LPN
|Patient no.||Sex/Age||Tumour size, cm)||Location (pole)||EBL, mL||WIT, min||Angiographic finding|
| 3||F/56||4||Mid/Hilar|| 50||21||RAP|
The mean (range) time to readmission for delayed haemorrhage was 16.8 (9–30) days. All patients presented with gross haematuria accompanied by other symptoms such as ipsilateral flank pain, weakness and diffuse abdominal pain. Initial CT showed a perirenal haematoma without contrast extravasation (Fig. 1). Five patients (38.5%) received a mean (range) of 2.4 (2–4) units of packed red blood cells upon admission for low (mean 23.6%) haematocrit level associated with signs of haemodynamic instability. Angiography and SAE were performed after conservative measures failed or when patients who received a transfusion did not respond appropriately. At our institution conservative measures include intravenous fluid hydration, blood transfusion, continuous bladder irrigation, and appropriate laboratory studies to monitor patient response to treatment. Failure to respond after more than 24 h or progression of symptoms determined the need for intervention with SAE. Angiography showed a definitive aetiology for the delayed postoperative haemorrhage in all 13 patients. Specifically, RAP was observed in 10 patients and contrast extravasation was observed in the three remaining patients. An AVF was identified in two of the patients who had contrast extravasation. SAE of the segmental artery involved was performed with endovascular platinum microcoils, resulting in complete cessation of bleeding in all patients (Figs 2,3). The median (range) length of stay after SAE was 3.2 (1–5) days. At a median (range) follow-up of 13 (11–22) months no patient had recurrent haemorrhagic episodes, creatinine levels remained stable over the period of follow-up, and repeat imaging at 6 months showed no evidence of the perirenal haematoma. There were no major complications associated with SAE. Post-procedural minor complications included fever and ureteric obstruction secondary to blood clot in one patient. Cystoscopy and stent placement resulted in defervescence and resolution of the ureteric obstruction. Another patient was readmitted for fever and found to have a small subsegmental pulmonary embolus necessitating anticoagulation therapy. Figure 4 summarizes a proposed treatment algorithm for patients presenting with delayed haemorrhage after LPN.
Figure 3. A, Selective angiography showing contrast extravasation and AVF. B, Closure of fistula and resolution of contrast extravasation after selective coil embolization.
Download figure to PowerPoint
- Top of page
- PATIENTS AND METHODS
- CONFLICT OF INTEREST
Although still considered a major challenge, LPN has gained significant acceptance worldwide and its use has been extended for the resection of larger and more centrally located tumours [1–4]. Several large series have been published showing short- and long-term oncological outcomes for LPN to be similar to those for OPN with the added benefit of a shorter hospital stay, faster convalescence and less pain in the postoperative period [7,9]. The number and complexity of LPNs performed has significantly increased at large academic centers as well as smaller institutions and the focus has been on researching complications of LPN and how these compare with those of OPN. One of the largest series comparing a total of 1800 LPNs and OPNs found that postoperative urological complications, particularly renal haemorrhage were slightly more common in the LPN group (4.2% vs 1.6%). Most patients were successfully managed via conservative measures. However, haemorrhage was higher after LPN compared with OPN, more commonly necessitating a secondary procedure including embolization, re-exploration, or completion nephrectomy . Haemorrhage associated with LPN can occur intraoperatively or postoperatively. While intraoperative bleeding from the tumour resection bed is usually recognized and can be addressed by oversewing the parenchyma or the bleeding vessel, recognizing postoperative haemorrhage can be more challenging, especially when patients present days or weeks after surgery. Overall, postoperative haemorrhage after LPN has been reported to be 4.2–6%[8,9]. Delayed postoperative haemorrhage has not been well defined separately but appears to occur in 1.2% of small tumour cases and in up to 7.5% of more complex cases involving centrally located tumours [4,6,11,12].
These cases of delayed bleeding most commonly occur from injury of a segmental artery, which can be shown with angiography, providing a definitive diagnosis. When high-pressure arterial blood leaks freely, contrast extravasation is observed. Drainage of high-pressure flow into a low-pressure system can result in an AVF (leakage into an adjacent vein) or RAP (leak into renal parenchyma or hilar areolar tissue). Review of the literature showed that delayed bleeding after LPN was most often associated with RAP. To the best of our knowledge, there are eight case reports describing a total of 11 cases as well as four case series describing an additional 22 cases of delayed haemorrhage after LPN secondary to RAP (Table 3[4,8,13–23]). Additionally, one series by Ramani et al.  found delayed haemorrhage in eight out of 66 patients (4%) undergoing LPN. Half of these patients were treated conservatively, only two underwent SAE and one patient underwent completion nephrectomy. However, in five out of the eight patients a possible precipitating cause of delayed haemorrhage could be identified. Of the four patients who responded to conservative measures (bedrest and blood transfusions), three had identifiable causes (splenic tear, fall, strenuous exercise) which leaves only one patient responding to conservative measures for delayed bleeding after LPN. Combining results from previous studies and data from the present series, RAP appears to be the most common aetiology in patients presenting with delayed haemorrhage after LPN. In the absence of identifiable risk factors for bleeding, conservative measures have not been shown to be successful for patients with haemorrhage after LPN.
Table 3. Summary of LPN series and case reports of delayed postoperative haemorrhage
|Reference||No. of cases with haemorrhage/Total||% delayed haemorrhage||Intervention (% success)||Mean POD of presentation (range)|
|Present series||13/640||2||SAE (100)||16.8 (9–30)|
|Nadu et al. ||4/53 (central tumours only)||7.5||SAE (100)||20.5 (7–45)|
|Ramani et al. ||8/66||4||2 of 8 SAE (100)||15.6 (6–30)|
|Singh and Gill ||6/485||1.2||SAE (100)|| 11.7 (8–15)|
|Shapiro et al. ||6/259||2.3||SAE (100)||12.6 (5–23)|
|Inci et al. ||2/25||n/a||SAE (100)||17 (12–22)|
|Cohenpour et al. ||1/Case report||n/a||SAE (100)||14|
|Negoro et al. ||1/Case report||n/a||SAE (100)|| |
|Wright et al. ||2/Case report||n/a||SAE (100)|| 9|
|Uberoi et al. ||2/Case report||n/a||SAE (100)||21|
|Zorn et al. ||1/Case report||n/a||SAE (100)||24|
|Hayn et al. ||1/Case report||n/a||SAE (100)|| |
|Shigeta et al. ||1/Case report||n/a||SAE (100)|| |
|Moore et al. ||1/Case report||n/a||SAE (100)|| |
Renal artery pseudoaneurysm and AVF causing renal haemorrhage are well documented and are often reported after renal biopsy, blunt and penetrating renal trauma, renal transplantation, and are found in 1.2% of cases after percutaneous renal surgery [12,24–27]. After OPN, a large series by Albani and Novick  showed that three out of 698 patients (0.4%) were found to have a RAP. The incidence of RAP formation appears to be higher after LPN and ranges from 1.2 to 2.3 %[13,14]. The incidence is even higher for central tumours treated with LPN. Nadu et al.  reviewed a series of 212 LPN cases of which 53 patients had central tumours including 12 hilar lesions. All LPNs were performed with hilar control and there was no difference in EBLor other intraoperative parameters analysed. Postoperative bleeding occurred in 7.5% (4 of 53) of patients, all presented in a delayed fashion (>7 days after surgery) and all were found to have RAP. The reported cases of RAP were identified by angiography and successfully treated with SAE. In the present series we included all LPNs regardless of tumour location, including central lesions as well as all LPNs done without hilar control. The incidence of bleeding after LPN in our series falls within previously reported ranges.
To identify the incidence, presentation and angiographic findings of delayed haemorrhage this study only included cases that occurred on or after postoperative day 7. All patients presented with gross haematuria and at least some degree of flank or abdominal pain. Initial management included fluid resuscitation, assessment of haemoglobin level and blood transfusion as clinically warranted. Routine CT imaging was obtained in most patients presenting with haematuria after LPN early in our experience. However, more recently the imaging of choice is CT angiography with embolization if needed.
In all cases CT imaging showed perinephric haematoma. However, the decision to further intervene with angiography and SAE was based on haemoglobin level and clinical signs/symptoms of persistent haemorrhage (persistent severe gross haematuria, haemodynamic instability, or inappropriate haematocrit response to blood transfusions). In our experience, initial CT imaging failed to contribute significantly to the diagnosis or treatment in this clinical context. In rare cases where patients present with concomitant obstruction secondary to ureteric clots, imaging would help make this diagnosis. In the absence of renal colic or pain, imaging may be obviated. Reviewing the published data and data from the present series, it is evident that conservative management for treatment of delayed haemorrhage after LPN plays a role in select patients only, perhaps those with identifiable risk factors for bleeding or those with only minimal signs and symptoms. The majority of patients presenting with delayed and persistent gross haematuria after LPN will require intervention as shown by the data accumulated to date. All patients in this series eventually underwent SAE. Identifying and treating select patients who present with delayed and persistent haemorrhage after LPN early with SAE could potentially reduce the number of transfusions needed and decrease the length of stay.
The most common aetiology identified in our series was RAP (77%). Three cases showed contrast extravasation on angiography and in two of the patients a definite AVF was found. In all cases, SAE was successful and no recurrence was reported after a single treatment with endovascular embolization coils. Superselective angiography with embolization was able to identify the focal point of the bleed and microcoils were delivered with great specificity, allowing for minimal devascularization and maximal preservation of renal parenchyma and, in turn, renal function.
Mechanisms for formation of RAP have been proposed and previously described in detail [16,27]. Prevention of RAP is important and the importance of technical caveats such as meticulous oversewing or clipping of transected vessels, haemostatic renorrhaphy and careful reinspection of the surgical site after unclamping and deflation of pneumoperitoneum cannot be emphasized enough. The use of adjunct haemostatic agents such the ABC and gelatin matrix thrombin sealant (FloSeal) are useful but cannot be used instead of meticulous parenchymal repair. The ABC is an excellent device allowing for control of minor parenchymal oozing and for superficial coagulation of the tumour bed. However, this device does char the tissue and may cause thermal damage to the collecting system, theoretically contributing to a prolonged urinary leakage. It may also weaken or burn any polyglactin 910 suture material used for underlying pelvicalyceal repair. FloSeal has been shown to reduce the number of haemorrhagic complications when used together with parenchymal repair and renorrhaphy. Gill et al.  compared two small groups of patients undergoing LPN and the impact of FloSeal use on haemorrhagic complications. However, RAP formation was still observed in 2 out of 52 patients in the FloSeal group. In the current series adjunct haemostatic agents were used at the discretion of the surgeon. RAP was observed irrespective of parenchymal repair alone or combined with ABC or haemostatic agents.
Comparing EBL and tumour size among patients with and without delayed haemorrhage no significant difference was found. The mean WIT in the group of patients without haemorrhage was found to be significantly shorter (14.4 vs 28.2 min, P < 0.001) than in patients with delayed haemorrhage. However, this difference is attributable to the large number (68 cases) of off-clamp LPNs (zero WIT) in the group of patients that did not bleed. Off-clamp LPN is a relatively new experience and is usually applied to highly selected patients with small and mostly exophytic lesions. This selection bias probably explains the lower rate of delayed haemorrhage after off-clamp LPNs identified in this series. Statistical subgroup analysis for differences with regard to delayed haemorrhage was not intended in this series. Future studies are needed to compare complications including delayed haemorrhage between well-matched patients undergoing standard hilar control vs off-clamp LPN.
The present study is limited by its retrospective nature and the relatively small number of delayed haemorrhage cases encountered. However this series represents the largest to date and supports previously published findings. The true incidence of RAP is probably higher than reported in this series or others, given that only symptomatic patients are evaluated and angiography is reserved for patients with severe symptoms.
In conclusion, delayed haemorrhage is an uncommon but serious event, which must be considered in all patients presenting with postoperative haematuria or other signs of bleeding after LPN. Angiographic imaging and SAE could be part of an early intervention protocol in well-selected patients presenting with delayed and persistent bleeding. The minimally invasive endovascular approach can accurately identify and treat delayed bleeding after LPN regardless of aetiology (AVM vs RAP) and SAE allows for preservation of renal tissue. Early identification and treatment can reduce morbidities associated with vascular complications and eliminate the potential risk of nephrectomy.