Open Partial Nephrectomy: Incidence

Classically, the critical technical challenge of partial nephrectomy has been balancing vascular control to reduce the risk of significant intraoperative hemorrhage against the risk of ischemic injury resulting from prolonged clamping. In most open series, hemorrhage has been the most common intraoperative complication. The reported incidences of other significant postoperative complications include those of urinary fistula (0–17.4%), acute renal failure (0–26%), and postoperative bleeding (0–4.5%).

In the early experience with open partial nephrectomy, reported complications were high, at ≤30.1% in one series. In that series, hemorrhage was the fourth most common complication, whereas acute renal failure was more common, occurring in 42.3% of patients. In preventing bleeding and renal dysfunction, the surgeon’s challenge is to balance efficient reconstruction with as minimal an ischemic injury as possible. Recognizing the deleterious effect of prolonged arterial clamping on renal function, surgeons have adopted a more permissive bleeding strategy and improved reconstructive techniques. With the increased experience of the surgeons, complications of open partial nephrectomy have declined; in a series comparing 1049 consecutive open radical or partial nephrectomies, complications of partial nephrectomy occurred in 9% of patients including urinary fistula (5.5%), acute renal failure (1.3%), perinephric abscess (1.1%), pneumothorax (0.8%), and retroperitoneal hemorrhage (0.8%). Thompson and colleagues, when comparing two groups of patients operated before and after 1995, reported an incidence of hemorrhage of 1.5% and 1.2%, respectively. The incidence of acute renal failure was also much lower (3.8% and 1%, respectively); dialysis was required in only 2% of those operated before 1995 and in 0.6% of those treated thereafter. Urinary fistula or leak has also been a historically frequent complication in some early series (57.6%), whereas in a retrospective review of 1118 partial nephrectomies at one institution, urinary leak was observed in 52 patients for a rate of 4.4%. Clearly, the learning curve of partial nephrectomy has reduced morbidity.

Van Poppel and colleagues prospectively compared the complications of nephron-sparing surgery and radical nephrectomy in a phase III multicenter trial. These investigators found in the nephron-sparing group a rate of severe hemorrhage (defined as blood loss >1 L) of 3.1% and urinary fistula of 4.4%, whereas in the radical nephrectomy group severe hemorrhage was seen in only 1.2%. Reoperation for complications was necessary in 4.4% of the nephron-sparing group and in 2.4% of the radical nephrectomy group. As such, partial nephrectomy has been considered safe and desirable, despite a higher rate of perioperative and early postoperative complications, given the potential benefit of preserving the renal function. We have experienced an extremely low rate of complications with contemporary partial nephrectomy by adhering to the specific technical concepts outlined later in this chapter. With careful attention to detail, partial nephrectomy can be very safely performed for most patients with early-stage renal tumors.

Positive Margin

Oncologic efficacy in renal cancer relies on a negative tumor margin during partial resection. Although investigators have demonstrated that the thickness of the margin does not appear to be important in predicting recurrence-free survival, a negative margin remains essential. Simple enucleation of the tumor is certainly technically the simplest approach, but in our opinion it increases the likelihood of microscopic residual disease.

Planning the line of renal incision is heavily dependent on the tumor position and the extent of subcortical tumor. Tumors that are predominantly exophytic can be incised in a shallow incision around the circumferential base of the tumor. Tumors that are predominantly endophytic require true resection of a segment of the kidney to ensure an adequate margin. We have found the technique of intraoperative ultrasound most useful for mapping the line of renal incision during both open and laparoscopic partial nephrectomy ( Fig. 38.4 ). The ultrasound probe is passed over the kidney from normal cortex toward the center of the tumor. Radial passes of the ultrasound probe allow mapping of the subcortical extent of the tumor. The line of renal incision is then planned on this basis.

In mapping a tumor by ultrasound scan, the probe is passed across the tumor in each direction (arrows point in the direction of the probe movement) to determine the extent of subcortical tumor and to ascertain the optimal line of incision. Tumor can often extend tangential or lateral to the visible exophytic component of the tumor. The probe can be passed across the tumor in radial or perpendicular lines (like the spokes of a wheel) to assess the extent of tumor fully.

During excision, arterial ischemia is useful in allowing direct visualization of the renal incision as it progresses. This feature is perhaps most important in laparoscopic partial nephrectomy because tactile feedback is less available to assess the surgeon’s position in the kidney relative to the tumor. Incision into the tumor, or too close to it, can be redirected by a repeat incision at a greater depth starting 1–2 cm proximal to the leading edge of the incision. We have used visual landmarks in the kidney to guide the adequacy of our renal incision during laparoscopic partial nephrectomy. For example, if a tumor is radiologically positioned within 10 mm of the renal sinus or collecting system ( Fig. 38.5A ), then visual entry of these structures confirms the adequacy of the margin (see Fig. 38.5B ).

On observing the position of a tumor to be within 10 mm of the renal sinus (A), exposure of the sinus during resection (B) is essential to ensure adequacy of the margin.

Frozen section analysis of shaved margins is widely used to confirm the adequacy of the incision and, although this analysis may reassure the surgeon and patient, it is likely inaccurate in predicting oncologic efficacy. Sampling error and artifactual expansion of the defect during incision render this technique of margin analysis of little value.

Bleeding

Because partial nephrectomy requires transection of multiple blood vessels during renal incision, it is perhaps intuitive that bleeding is a potential consequence. Bleeding can occur intraoperatively, immediately postoperatively, and at a delayed interval up to several days or weeks after the surgical procedure. Intraoperative bleeding usually occurs from the cut parenchyma, but it can occur from the hilum if the vessels are not carefully dissected during exposure. Preoperative knowledge of the individual renal vascular anatomy is essential in this regard. Bleeding has generally been feared in partial nephrectomy. For large and central tumors, selective embolization or extracorporeal resection of the tumor followed by autotransplantation has been performed, in an attempt to provide clean margins with limited blood loss. In the contemporary era, such techniques are generally not necessary if one adheres to basic tenets of vascular control and renal reconstruction.

The fundamental tenets of renal reconstruction are as follows:

• 1.
  

  Closure of the collecting system

  
  
• 2.
  

  Suture or coagulation control of small transected intraparenchymal vessels

  
  
• 3.
  

  Prevention of retraction within cut large central blood vessels

  
  
• 4.
  

  Compression of the defect to aid in venous tamponade.

A variety of tools and techniques can be employed for each tenet, but with attention to each, good outcomes can be achieved.

Parenchymal bleeding during resection is usually handled by clamping of the renal artery or vein. In most open partial nephrectomies, arterial clamping is adequate because venous tamponade through hilar stretch and manual compression is easily achieved. Brisk bleeding on renal incision despite clamping can indicate a missed renal artery, and if it is not easily identified, a hilar cross-clamp can be placed during both open and laparoscopic procedures. When performing this maneuver on the left side, the position of the aorta, adrenal, pancreas, and superior mesenteric artery should be known before the clamp is placed. On the right, avoidance of the duodenum and porta hepatis is essential. If brisk bleeding is noted with the venous clamp in place, then venous hypertension resulting from continued arterial inflow should be suspected. Removal of the venous clamp can reduce bleeding in these cases.

Numerous techniques exist for repair of cut renal vessels. Direct suture ligature with 4-0 absorbable sutures is generally preferred. We find it most effective to identify vessels during renal parenchymal incision and suture ligate them before division. The renal parenchyma is separated with a small Freer instrument ( Fig. 38.6A ), and on identification of the vessel ( Fig. 38.6B ), a direct suture ligature is placed and the vessel is divided with tenotomy scissors. In this fashion, the line of resection can be controlled and redirected ( Fig. 38.7 ).

Separation of the renal parenchyma with a Freer instrument (A) allows visualization of arterial branches as they enter the specimen (B). On recognition of the vessel, it is directly suture ligated, thereby avoiding placement of deep sutures that may inadvertently injure or ligate a more proximal branch point of the vessel or adjacent vessels.

(B, From Nieder A, Taneja SS. The role of partial nephrectomy for renal cell carcinoma in contemporary practice. Urol Clin North Am. 2003;30:536.)

By ligating and incising small vessels entering the partial nephrectomy specimen, while they are exposed and stretched from the parent vessel, one can avoid injury to deeper vessels (thereby minimizing radial infarct) and have directional control of the depth and direction of the incision.

If brisk bleeding occurs after removal of the arterial clamp, one should again consider the possibility of venous compression or twisting of the renal pedicle, thus impairing venous outflow. If the vein is not compromised, then suture repair of bleeding should be performed while one manually compresses the defect. Reclamping of the artery should be avoided because of the increased risk of ischemic injury following reperfusion.

In cases of immediate postoperative bleeding, initial management should be conservative with transfusion, bed rest, and serial monitoring of hematocrit. Expanding retroperitoneal hematoma can worsen bleeding through stretch of the renal defect. Therefore if multiple transfusions are required, hemodynamic instability is noted, or the hematocrit is not responsive to transfusion, then one should give early consideration to angiography and selective embolization. In these patients, unlike in renal trauma, attention must be focused on the resection bed in which any visible vessels can be embolized, even if they are not actively bleeding. Intermittent hemorrhage is common, and bleeding may not actually be seen on angiography.

In patients with delayed bleeding (hematuria or flank hematoma) following discharge from the hospital, early angiography should be performed given the high likelihood of arteriovenous fistula or pseudoaneurysm. If the index of suspicion is low, or if the bleeding is minimal and the patient is hemodynamically stable, then one could first perform a duplex ultrasound scan (see Chapter 34 , Fig. 34.6 ) or magnetic resonance angiography (MRA) because these techniques are less invasive. If an abnormality is not noted and bleeding persists, then angiography would be required anyway. Selective angioembolization is extremely effective in the management of these complications. We have on occasion seen delayed hematuria resulting from decompression of an extrarenal hematoma through the collecting system. In these cases, a small, fluid-filled defect (closed urinoma) is noted at the resection bed on an MRI urogram.

Infarct

In performing partial nephrectomy, preservation of the vascular supply of the retained segments of kidney is essential to avoid renal infarction. Although the functional significance of infarct is not known, certainly infarcted tissue likely decreases the preserved renal reserve over time. Large infarcted segments can result in renin-mediated hypertension, and if a large portion of the repaired collecting system loses blood supply, this situation can promote leakage.

Avoiding infarct again relies on knowledge of renal arterial anatomy. Given the end-arterial nature of renal blood flow, maximizing proximal vascular preservation should be the goal of resection. In approaching polar lesions, straight rather than tangential amputation is desirable, but if the lesion is primarily located within the anterior or posterior segmental vasculature, then a significant amount of parenchyma can be preserved through a tangential cut. Within the anterior and posterior segment, we find that starting the parenchymal incision away from the hilum and cutting toward the center of the kidney will allow identification of arterial branches as they enter the specimen, thereby avoiding radial infarct ( Fig. 38.8 ). We find that starting the incision from the hilum and working outward results in a higher likelihood of injury to main branches and thereby promotes additionally infarcted kidney ( Figs. 38.9A–D ).

Partial nephrectomy incisions carried outward from the hilum (double arrows) often injure larger blood vessels feeding renal parenchyma beyond the planned resection. As a result the risk of radial infarct, or adjacent segmental infarct, is relatively high. Cutting toward the hilum (single arrows) allows incision of only those vessels that are feeding the resected parenchyma as they are encountered, resulting in a lower risk of radial infarct.

The direction of incision can influence the likelihood of injuring arterial branches that feed the residual kidney. As arteries and more proximal branch points are present in the hilar region, we generally prefer to start resection on the side of the tumor opposite the hilum. This is particularly relevant when operating in the anterior or posterior segment. A, If the incision is carried out from the hilum, infarct radial to the defect can often be seen. B–D, When starting the resection away from the hilum, larger central vessels are preserved and the surrounding kidney remains well perfused.

On separation of renal parenchyma along the planned line of incision, individual vessels are identified as they enter the specimen, are suture ligated, and are transected sharply to avoid the necessity for deep suture placement for retracted blood vessels (see Figs. 38.6 and 38.7 ). This technique is more difficult to recapitulate in laparoscopic surgery, and as such, the likelihood of radial infarct is higher in laparoscopic partial nephrectomy owing to the suturing techniques typically used in the laparoscopic procedure. We have adapted a modified suturing technique (described in Chapter 34 ) to reduce the likelihood of major infarct or vascular complication.

The general management of infarct is observation. In patients with severe hypertension or persistent leak resulting from infarcted kidney, nephrectomy should be performed.

Urine Leak

Urine leak rates reported in the urologic literature depend on the operator’s definition. This definition has varied widely from leakage for >72 hours to leakage persisting for >3 months postoperatively. We generally define urinary fistula following partial nephrectomy as urine leakage persisting for >4 weeks after the surgical procedure. Nonetheless, any duration of postoperative urine leak must be appropriately managed by the surgeon. Simple principles of drainage, prevention of infection, and avoidance of ureteral obstruction generally allow resolution of the leak. Management of urine leak following partial nephrectomy is additionally discussed in Chapters 13 and 34 .

Careful closure of the collecting system generally prevents urine leak. In cases of large resection bed, or in tumors resected from the anterior or posterior segment with tangential renal incision, small calyceal injuries may not be identified, and these are often the site of persistent leak. In general, meticulous closure of the collecting system is performed with a layer of interrupted absorbable 4-0 braided sutures. This stage is followed by a second layer of imbricated sutures that attempt to pull the parenchyma together around the defect and thereby reduce tension on the primary closure ( Fig. 38.10 ).

Two-layer collecting system closure involves two interrupted layers of absorbable suture. The first layer (A) incorporates the edges of the collecting system, whereas the second layer (B) attempts to imbricate normal renal parenchyma to remove tension from the first line. The use of interrupted rather than running suture reduces the likelihood of tearing the collecting system with the suture.

Two modifications of technique have greatly reduced our leak rate with open partial nephrectomy: (1) retrograde instillation of methylene blue through the renal pelvis and (2) the use of a layer of tissue adhesive over the closure. The use of tissue adhesive is most important in laparoscopic partial nephrectomy because small caliceal injuries are often difficult to identify. We have adopted a standardized technique of infiltrating Gelfoam (Pfizer Inc., New York) with fibrin sealant (Tisseel, Baxter International Inc., Deerfield, Illinois), molding it to the resection defect, and then activating it through infiltration of thrombin. Additional hemostatic materials are packed into the defect, and then the kidney is folded over the materials using horizontal mattress sutures of 3-0 absorbable nonbraided suture anchored to the surrounding renal capsule.

Urine leak presents in one of two ways. Early leaks often become evident in the recovery room and persist, whereas delayed leaks manifest 5–14 days postoperatively and may become symptomatic. It is not clear whether most delayed leaks were simply unrecognized early because of inadequate drainage and lack of symptoms. The old tenet that urine leakage frequently occurs in the first 24–72 hours but then resolves is not entirely true. In fact, most renal reconstruction for elective partial nephrectomy should be watertight. Early leakage generally indicates poor collecting system closure or unrecognized collecting system injury, and in our experience this leakage rarely stops within the first few days.

Early leak is suspected if drainage of >30 to 40 mL per shift is noted >48 hours after the surgical procedure. Large-volume drain outputs within the first 24 hours may lead one to suspect leak, but intervention is not required unless infection is noted, serum creatinine becomes markedly elevated because of reabsorption, or the patient has no urine output (solitary kidney). In these cases, renal obstruction may be suspected and early imaging is advised. Urine leak beyond 48 hours is confirmed by measurement of drainage creatinine level relative to serum. If the drainage is pure urine, creatinine levels will generally be >30 mg/dL. When the drainage creatinine is even moderately higher than that of the serum, then at least part of the draining fluid is urine and a urine leak is likely present. Minimal creatinine elevation may suggest a resolved urine leak with some dilute urine still in the retroperitoneum. After leak confirmation, we generally perform an immediate ultrasound scan or noncontrast CT study for assurance that the drain is properly positioned, that no undrained urinoma is present, and that the kidney is well drained.

If all these tenets are met, the patient is discharged home with prophylactic antibiotics and a drain in place. Patients are monitored weekly with electrolytes and renal ultrasound studies to rule out worsening urinoma or drain migration. In cases of undrained fluid, the drain is repositioned to sit immediately adjacent to the leak point. This maneuver can be facilitated by contrast CT with delayed cuts. Ideal drain placement is immediately adjacent to the kidney to allow a drain tract to form. In cases of high-output leak, the drain can be converted from suction to gravity, but a follow-up ultrasound scan within a few days is warranted to rule out a secondary fluid collection.

Presentation of delayed leak is usually the result of flank pain, fever, or drainage from an incision. In a patient with fever, evaluation should include a complete blood count with manual differential, blood and urine cultures, and a chest radiograph to identify other potential sources of infection. CT with intravenous contrast should include delayed images to identify urinary extravasation, provided renal function is acceptable for receiving iodinated contrast. Ultrasound can alternatively be used to identify a perirenal fluid collection. Before percutaneous drainage, broad-spectrum antibiotics should be administered. The drain is best positioned immediately adjacent to kidney, but in cases of infection, maximal drainage is paramount to allow defervescence, and this may be best achieved with the patient in the most dependent position.

In the absence of infection, similar evaluation of renal function and imaging to identify fluid collections should be performed. On percutaneous drainage, the fistula output may be quite high initially because of the presence of a large potential space outside the kidney. Maximal kidney drainage is necessary to promote antegrade drainage over time. In general, a ureteral stent is not necessary unless evidence indicates ureteral obstruction, blood clots or debris within the collecting system, or a very large collecting system opening. In our experience, stents often worsen the leak initially, possibly as a result of reflux. It is essential to have a concomitant urethral catheter in place while the leak is perpetuated, particularly in a man with high-pressure voiding. We have generally avoided stent placement unless absolutely necessary because most leaks will close with prolonged drainage.

Resolution of urine leak generally requires scarring in of the potential space around the kidney. This occurs as scar tissue fills in around the drain forming a drain tract. We generally prefer to leave the drain on gravity drainage to promote antegrade urine flow within the collecting system. On converting to gravity drainage, a follow-up ultrasound scan within 36–48 hours will confirm that no urine collection has formed around the kidney or drain tract. If it has, then the drain should be returned to suction. Urine collections prevent the formation of a drain tract and delay leak closure. Over a period of 2–4 weeks, if properly drained, a drain tract should have formed, and on confirmation of the absence of urinoma, the drain can be advanced back 2–3 cm to allow the tract to close between the tip of the drain and the kidney. Every 48–72 hours, the drain can be pulled back another 2–3 cm until it is removed. During this time of drain advancement, one should monitor for urinoma formation and fever, a finding suggesting recurrence of the urinoma within the drain tract. In cases of high drain outputs (>300 mL/24 hours) more prolonged observation (≤12 weeks) may be necessary before attempting drain advancement. Low-output leaks require minimal observation before closure can be achieved. In this regard, closure of a leak is an active process initiated by the surgeon rather than a passive physiologic process.

Urine leak can also result from ureteral injury. Excessive dissection of the renal pelvis, the lower pole of the kidney, and the proximal ureter should be avoided because it can cause devascularization, ischemia, and necrosis with subsequent fistula formation in the dissected segment. On completion of the procedure, fat should be interposed between the lower pole of the kidney and the upper ureter to avoid adherence, scarring, and subsequent stricture of the ureter.

Renal Function

Measuring renal functional outcomes after elective partial nephrectomy is most difficult given the compensatory ability of the remaining kidney. Conventional beliefs have been that renal dysfunction following partial nephrectomy is related to the length of renal arterial clamping and the amount of parenchyma resected. In truth, many factors may influence the function of the operated kidney including intraoperative blood pressure variations, fluid resuscitation during the case, and aggressive handling of the kidney. Intraoperative blood loss may result in a higher risk of renal dysfunction because of the possibility of hypotension and reduced tissue oxygenation. When operating on an individual with two kidneys, the risk of postoperative azotemia and the necessity for dialysis depend on the functional reserve of the remaining kidney. Patients with diabetes, hypertension, and known vascular disease are more likely to have transient azotemia postoperatively. Individuals with a solitary kidney are also affected by these factors, but clearly, the risk of azotemia is more strongly affected by renal arterial clamping than in patients with two kidneys and normal renal function.

In patients with bilateral functioning kidneys, the reported rate of acute kidney injury after open surgery has been 4%, whereas chronic kidney disease has been reported to occur in 2–8% of patients. Similar patients in laparoscopic series were shown to have negligible rates of renal dysfunction. In patients with solitary kidneys, the rates of acute kidney injury after open procedures are more evident, ranging between 13% and 38%, with chronic kidney disease occurring in from 3% to 30% of these patients. In a series of 103 partial nephrectomies in a solitary kidney, Ghoneim et al. reported that eight patients developed acute kidney failure, but nonmodifiable factors such as age and preoperative eGFR predict long-term postoperative eGFR, rather than modifiable factors such as ischemia time. In laparoscopy series, these data have not been consistently reported. Analyzing an initial cohort of 430 patients who underwent laparoscopic partial nephrectomy at the Cleveland Clinic, Gill and colleagues observed that in 22 cases with solitary kidneys and two of which electively converted to an open procedure, a single patient (4.5%) required temporary dialysis and the chronic kidney disease incidence rate was 13%.

Several maneuvers to reduce renal ischemia have been proposed. The use of an osmotic diuretic, such as mannitol, immediately before and shortly after clamping of the renal artery is thought to reduce the accumulation of free oxygen radicals in the tissues. Similarly, cooling of the kidney during ischemia is believed to reduce the likelihood of tissue injury. During open partial nephrectomy, we have used both these maneuvers. The kidney is placed within an intestinal bag with the opening loosely cinched around the renal pedicle. On ice slush packing, the bag serves to hold the ice to the renal surface. After clamping, a period of 5–10 minutes is necessary to reach the nadir core temperature within the kidney. We generally begin to operate on the kidney during this period to reduce total ischemia time, but other surgeons advocate waiting a full 5 minutes before renal incision.

In general, we do not clamp the renal vein because stretch of the pedicle usually sufficiently reduces venous backflow. Investigators have suggested that retrograde circulation of venous blood may allow low-level tissue oxygenation in the setting of arterial clamping. This possibility has been suggested for laparoscopic partial nephrectomy as well.

An alternative to the use of vascular clamping is manual compression of the renal parenchyma. This technique is quite effective in open partial nephrectomy, particularly for polar lesions. In laparoscopic surgery, the technique can be applied through hand assistance, but this approach is more cumbersome because one-hand suturing is required. Even if the renal vessels are not clamped, they should be fully dissected in case of the need for urgent clamping. Surgery can then be performed by squeezing the renal parenchyma with the hands and using only enough pressure to control the bleeding.

Because of the increased risk of azotemia in the setting of a solitary kidney, we generally avoid arterial clamping in these patients whenever possible. In most cases, this situation requires open surgery, but in the case of small exophytic tumors, a laparoscopic approach with no arterial clamping can be considered.

Postoperatively, attention should be given to avoiding or minimizing nephrotoxic or harmful medications such as angiotensin-converting enzyme inhibitors, nonsteroidal antiinflammatory drugs, and aminoglycosides. Individuals at high risk, including those with solitary kidney, large (>50%) renal resections, prolonged ischemia times, preexisting renal insufficiency, and severe vascular disease, should be prepared mentally for the possibility of temporary or even permanent hemodialysis.

Reoperation

Urine leak rarely requires reoperation, but in cases of persistent ureteral obstruction perpetuating the leak, recurrent abscess, or severe intraperitoneal leak (laparoscopic), a repeat operation may be necessary. Obstruction may be addressed by temporary stenting maneuvers, or reconstruction if it is clearly the cause of recurrent or persistent leak. In cases of refractory leak, either due to persistent obstruction or renal infarct, nephrectomy is generally performed, particularly if the original procedure was elective in nature. Reoperation for bleeding is required when angiographic techniques fail to control postoperative bleeding. In most cases, reoperation for leak or bleeding results in nephrectomy.