Experience With Laparoscopic Donor Nephrectomy Among More Than 1000 Cases:  Low Complication Rates, Despite More Challenging Cases FREE

The overall safety of laparoscopic donor nephrectomy (LDNX) has been documented in several large case series^1- 4 and was as safe as open donor nephrectomy in a large-scale multi-institutional database study⁵ and in a recent meta-analysis.⁶ Moreover, this technique has been shown to result in reduced postoperative pain, earlier return to work, and improved cosmetic results compared with open donor nephrectomy.^7- 8 These substantial benefits may have contributed to increased use of living donors for kidney transplantation.⁹ Therefore, LDNX has become the procedure of choice at most institutions in the United States.

However, concern remains about the application of this technique to several patient groups, including those with right-sided nephrectomies, donors with complex kidney anatomy, and obese donors (body mass index [BMI, calculated as weight in kilograms divided by height in meters squared] >30). The safety of LDNX in each of these complex categories has been evaluated in smaller case series. Specifically, right-sided nephrectomy was shown to result in shorter operating times, with an equivocal complication profile.^10- 11 A series among obese donors showed slightly increased rates of minor complications but no increase in major complications.¹² Finally, series using donors with multiple arteries demonstrated longer operative times but no increase in rates of donor complications.^13- 17

The University of California, San Francisco, has experienced excellent results with donors in all 3 of these complex categories.^1,10,13 Since 2003, only 1 patient has undergone an elective open donor nephrectomy for kidney donation. This retrospective medical record review was designed to evaluate complication rates as we have offered LDNX to all medically acceptable donors. By comparing the complication rates of our first 250 patients (when LDNX was selectively offered) with those of our last 795 patients (when LDNX was offered to all medically acceptable donors), we sought to determine if rates changed as cases became progressively more challenging. Moreover, we independently analyzed the overall and major complication rates in 3 complex donor categories (right-sided nephrectomies, donors with complex kidney anatomy, and obese donors), hypothesizing that complication rates in the 3 complex categories would be equivalent to those among more straightforward cases. The overall objective of this study was to demonstrate that this LDNX could be offered to all medically acceptable donors with equal safety once the surgical team had achieved adequate experience.

One thousand forty-five LDNXs were performed at the University of California, San Francisco (a tertiary care teaching institution) between November 3, 1999, and August 28, 2009. A transplant surgeon (C.E.F.) performed the first 27 procedures alongside a laparoscopic general surgeon and a laparoscopic urologist. All subsequent procedures were performed by 1 of 3 transplant surgeons (A.M.P., S.-M.K. and C.E.F.) with the assistance of a transplant surgery fellow or a surgical resident (A.J.A. and colleagues).

Preoperative donor evaluations included a history and physical examination by a nephrologist, standard kidney donor laboratory measurements (including urine collection for creatinine clearance), and health care maintenance studies (eg, colonoscopy and mammograms). Early in the series, renal anatomy was evaluated using renal angiograms and intravenous pyelograms. More recently, donors were evaluated using high-resolution computed tomographic angiography, generally at the transplant center. A selection committee (A.J.A., A.M.P., S.-M.K., J.P.R., C.E.F., and others) met to determine the final acceptance of all donors. Criteria for donor acceptance changed slightly over the course of the study, with a more liberal weight criterion (BMI, ≤36). The left kidney was generally preferred because of its longer vein unless arterial anatomy was more complicated on the left or some other feature distinguished the left kidney as the better kidney to leave in the donor (ie, a small stone in the right kidney, a large cyst in the right kidney, or a larger left kidney).

The surgical procedure has been described previously.^1,10,13 In brief, the donor was placed in a modified lateral position. Four radially dilating ports were used. Liberal fluid was administered during the case, with mannitol (25 g) given as the dissection progressed. Once the vascular structures had been isolated, heparin sodium was administered (3000-5000 U). The procedure for control and division of renal vasculature evolved during the study. For the first 5 cases, an endoscopic stapler (GIA; US Surgical, Norwalk, Connecticut) was used for vascular control and ligation. For the subsequent 1040 cases, venous control was obtained using a vascular stapler (TA; Tyco-Healthcare, Pleasanton, California). The renal artery was controlled by self-locking plastic clips (Heme-o-lok; Teleflex Medical, Research Triangle Park, North Carolina) and was further secured using a metal clip. After occlusion of the vessels, they were transected using laparoscopic shears. Following documented failure of the self-locking plastic clips,¹⁸ the renal artery was also controlled with a vascular stapler beginning with patient 654. Extraction of the kidney was generally through a 7- to 8-cm Pfannenstiel incision, without the use of an extraction bag. Port sites were closed at the skin level only, and the extraction incision was closed in layers.

Patient-controlled analgesia was used for the first 24 hours after surgery. Donors were allowed to drink and eat ad libitum within 8 hours after surgery and were encouraged to walk within 18 hours. Bladder catheters were removed within 24 to 48 hours. Discharge goals consisted of good pain control on oral analgesics, adequate oral intake, and resolution of any complications.

Demographic data were gathered from the patient medical records, and clinic visit records were analyzed retrospectively per a protocol approved by the Committee on Human Research at the University of California, San Francisco. Data collected included the following: age, sex, BMI, nephrectomy side, vessel management technique, operative time, need for blood transfusion, intraoperative complications, postoperative complications, length of hospital stay, and readmissions, as well as the number of arteries, veins, and ureters in the donor kidney. Comparisons between study groups were performed using a 2-tailed t test, assuming unequal variance.

The donor characteristics of all cases, the first 250 cases, and the last 795 cases are summarized in Table 1. The mean age of donors was 41 years (age range, 18-77 years) and did not change over the course of the series. At our institution, most donors were female (61.1%), and this did not change over time. When comparing our first 250 patients with the last 795 patients, our policy of accepting all medically acceptable donors for LDNX led to more complex cases in the latter part of the series. The percentage of right-sided nephrectomies increased from 12.4% to 21.3% (P < .001). Although the mean donor BMI did not increase substantially (25.5 vs 26.9), the percentage of donors with BMI exceeding 30 significantly increased (8.9% vs 23.2%, P < .001). In this study, complex kidney anatomy was defined as any case in which the recovered kidney had multiple arteries, multiple or retroaortic veins, or multiple ureters. The percentage of cases with complex kidney anatomy increased during the study period (16.4% among the first 250 patients vs 20.5% among the last 795 patients, P < .07). Overall, we performed 200 right-sided nephrectomies, 204 cases with complex kidney anatomy, and 148 cases in obese donors.

Our overall mean operative time was 184 minutes (Table 2). Among the entire series, we had 60 overall complications (5.7%) and 21 major complications (2.0%). All complications are listed in Table 3. There were no deaths in the series. Three patients required conversion to open surgery. The first conversion occurred in 1 patient after an endoscopic stapler failed to control bleeding at the renal artery stump. The second conversion occurred in patient 219, with a history of multiple abdominal operations and in whom an enterotomy occurred on initial incision (laparoscopic instruments were never placed). The third conversion occurred in patient 829, in whom a lumbar vein tore during mobilization of a retroaortic renal vein and could not be safely controlled. This patient was the only patient who required a blood transfusion among the last 795 patients. Two patients required reoperation. The first reoperation was in a patient who was readmitted with an elevated creatinine level and abdominal pain. Diagnostic laparoscopy revealed no intra-abdominal problem, and the patient had an uneventful recovery with hydration and laxatives. The second reoperation was performed in a patient who had been noted to have an expanding hematoma in the Pfannenstiel incision when in the recovery room. Using general anesthesia, the wound was explored, and the bleeding source was controlled in the subcutaneous tissue.

Operative time was defined as the time from skin incision to skin closure. Operative times statistically significantly improved, averaging 212 minutes among the first 250 patients and 176 minutes among the last 795 patients (P < .001) (Table 2). No significant change was noted in the overall complication rate (6.4% vs 5.5%). A major complication was defined as an event that changed the patient's operative course or required significant medical intervention, reoperation, or readmission. There was a statistically significant improvement in our major complication rate (4.0% among the first 250 patients vs 1.4% among the last 795 patients, P < .05). This was primarily owing to an improvement in our blood transfusion rate (1.6% vs 0.1%). To evaluate operative trends that might be associated with subsequent complications, operating room events were assessed. Examples of these events included minor injuries, such as liver or splenic lacerations, injuries to the small-bowel serosa, and bleeding that required extra surgical clips. All of these injuries were repaired laparoscopically and resulted in no negative effects on the patients. Although they did not alter the patients' outcomes, we wanted to confirm that these operating room events were not more frequent among the more complex cases, as they have the potential to result in true complications. However, no differences were noted when comparing the early vs the late study groups.

The patients who underwent right-sided nephrectomy (n = 200) had shorter operative time compared with that in the total series (mean, 167 vs 184 minutes, P < .05) (Table 2). There was also a slight trend toward more minor complications (6.0% vs 3.3%) and a trend toward fewer major complications (2.0% vs 2.4%), but these did not reach statistical significance (Table 3). Right-sided nephrectomy did not lead to conversion to open surgery, except for the patient in whom an enterotomy occurred during the initial attempt to gain peritoneal access (no laparoscopic instruments were used). No patients required blood transfusion in this group. However, 2 patients experienced carbon dioxide embolism, which is specific to right-sided nephrectomies, caused by the introduction of carbon dioxide into the vascular system secondary to a liver laceration with the Veress needle. Following these experiences, we routinely use a special port system (Opti-view; Ethicon Endosurgery, Inc, Cincinnati, Ohio) to gain laparoscopic access in right-sided nephrectomies.

One hundred forty-eight LDNXs were performed in obese donors (BMI, >30). Operative times were only slightly longer (mean, 189 minutes) in this patient population (Table 2). This group had a low rate of complications, with only 3 minor complications (2 postoperative ileus and 1 pneumothorax that did not require intervention) and no major complications (Table 3).

Complex renal anatomy in the donor organ was present in 204 patients, including 169 with multiple arteries, 41 with multiple or retroaortic veins, and 3 with multiple ureters (some cases involved >1 complex feature). Among donor organs in which the accessory artery was small (<2 mm) and supplied the upper pole, the accessory artery was sacrificed during procurement. In all other cases, the aberrant anatomy was preserved for implantation. Overall, the cases with complex renal anatomy took slightly longer (mean, 192 minutes) (Table 2), but complication rates were not increased (3.4% for minor complications and 2.5% for major complications) (Table 3). The single patient among the last 795 cases who required conversion to open surgery and blood transfusion had a retroaortic vein that was torn and bled. However, the overall rate of adverse events was not increased among donors with complex kidney anatomy compared with that among the total series.

Meta-analysis⁶ and review of a large multicenter donor database⁵ in combination with a body of case reports^1- 4 have demonstrated that LDNX is as safe as open donor nephrectomy for procurement of live donor kidneys. With the improved donor experience associated with laparoscopic procurement,^7- 8 these findings continue to favor LDNX as the procedure of choice for living renal graft procurement. To continue to expand the donor pool, it makes sense to offer the benefits of this procedure to the largest possible cohort of donors who would otherwise be considered acceptable to donate. However, given that LDNX is a major operative procedure with no physical benefit to the healthy donor, it is imperative that safety can continue to be demonstrated as the procedure is applied to more challenging cases.

This series reinforces and expands on earlier findings.¹ Our rate of conversion to open surgery (0.3%) and blood transfusion rate (0.5%) continue to be lower than the rates reported in other large series.^2- 4 This finding supports our belief that a purely laparoscopic procedure without the use of a hand-assistance device can provide adequate patient safety and vascular control. These results also demonstrate the importance of continued effort to improve our operative technique. We attribute our few vascular complications to progress in our methods of managing renal vessels. In our early operative experience, we used an endoscopic stapler system for vascular ligation. However, a stapler misfire that fails to ligate but cuts the donor vessel can lead to a large-volume bleed and no alternative but to convert to an open operation. A previous study¹ discussed a modification of our procedure in which self-locking plastic clips and metallic clips are used to ligate the renal artery before transecting the artery with laparoscopic scissors. We had no cases of significant bleeding using this technique, but documented donor deaths after failure of these locking clips led to warnings against the use of these clips for arterial control in living kidney donors.¹⁸ Therefore, we subsequently developed a third technique using a vascular stapler to ligate the renal artery and vein. Because the vascular stapler does not transect the vessel as the stapler mechanism fires, the staple line can be inspected before transection of the artery or vein. This technique provides the security of a staple line and conserves artery length, with reduced risk of bleeding. Our rates of major and minor complications have not changed significantly since we began using this technique, and there has been no need for conversion to open surgery because of failure of vascular control at the final stages of kidney removal.

This study describes 200 right-sided donor nephrectomies and adds to the series reported in detail previously.¹⁰ As described in our methods, we preferentially perform left-sided nephrectomies but do not avoid right-sided nephrectomies if otherwise indicated. Our findings are consistent with previous studies^10- 11 that show shorter operating times in right-sided nephrectomies. This is because the right-sided procedure requires little or no colon mobilization to expose the kidney, and venous branches are rarely encountered. Despite the advantage of shorter operative time, many centers are reluctant to offer right-sided LDNX because of concern about vascular control and vessel length. In the present series, the overall complication rate among right-sided nephrectomies was not significantly different from that among the total series. However, a complication unique to right-sided nephrectomies is carbon dioxide embolism secondary to Veress needle puncture of the liver and subsequent insufflation. Two severe cases of carbon dioxide embolus led to abortion of the planned laparoscopic procedures. After the second occurrence, we changed our technique of establishing insufflation on the right from the Veress needle approach to the use of a special port system that allows entry into the abdomen under direct visualization. This is an important technical detail in performing right-sided nephrectomy but does not preclude safe procurement without this trocar.

Our low complication rate among obese donors is striking in that only 3 minor complications were noted among 148 patients. We believe that LDNX can be offered to patients with BMI exceeding 30 without additional risk to the patient compared with open donor nephrectomy. A more important question relates to long-term safety of kidney donation in obese donors, and we set a BMI limit of 36 for donors. This cutoff is somewhat arbitrary, and we have no long-term data to support the concept that donor nephrectomy in obese patients is safe over the life span of the patient or that a BMI of 36 is the correct cutoff. However, the operation seems to be safe in this cohort of patients. The question of long-term safety will need to be answered with more rigorous long-term follow-up studies, and the need for answers supports the concept of a donor registry.

We also report on more than 200 donors with complex kidney anatomy. There has been concern about using a laparoscopic technique to procure kidneys with multiple renal arteries, as these procedures are associated with longer warm ischemia times and more ureteral complications in the recipient population compared with cases involving kidneys having single arteries.¹³ However, there are similar risks for renal grafts with multiple arteries in the cadaveric setting.¹⁴ Although the risks are higher using a donor organ with multiple vessels, this risk is far outweighed by the benefit of avoiding time on hemodialysis if only a multiple-artery kidney is available for transplantation. This study shows that these complex kidney anatomy cases can be performed with no additional risk to the donor. However, because of concern about ureteral complications, we continue to prefer right-sided nephrectomies in preference to procuring left-sided kidneys with multiple arteries.¹³

In conclusion, the operative times and major complication rates associated with LDNX improved in our last 795 patients compared with our first 250 patients. This division roughly approximates our institution's decision to offer LDNX to all potential donors. Since 2003, we have performed only 1 scheduled open donor nephrectomy at the University of California, San Francisco, for a pelvic kidney with extremely complex vascular anatomy. It is clear from this experience that laparoscopic nephrectomy can be offered to all medically acceptable donors once adequate experience is obtained by a surgical team.