A Multi-Institutional Propensity Score Matched Comparison of Transperitoneal and Retroperitoneal Partial Nephrectomy for cT1 Posterior Tumors

Abstract

Objective:

To compare the perioperative and renal functional outcome between transperitoneal and retroperitoneal robotic partial nephrectomy (TP-RPN and RP-RPN) in the largest cohort to date of RP-RPN for posterior tumors.

Methods:

We identified 519 patients who met eligibility criteria and underwent TP-RPN (n = 357, 68.8%) or RP-RPN (n = 162, 31.2%) for a posteriorly located cT1 tumor. Patients were propensity score (PS) matched on preoperative and tumor-specific characteristics. Perioperative outcome and renal function outcome at median follow-up 22 months were compared.

Results:

Between the PS matched TP-RPN (n = 157, 50%) and RP-RPN (n = 157, 50%) patients, operative time (OT) (185.0 versus 157.0, P < .001) was longer in TP-RPN versus RP-RPN patients. No significant differences in ischemia time (P = .618), blood loss (P = .178), positive surgical margins (P = .501), overall postoperative complications (P = .861), or progression of chronic kidney disease stage at median 22 months (P = .599) were identified. Length of stay (LOS) was reduced in RP-RPN patients (P = .017), but was not different once an institution used a postoperative day (POD)-1 discharge protocol (P = .579). Operative times were similar between groups in patients with obesity (P = .293) or a cT1b renal mass (P = 908).

Conclusion:

RP-RPN for posterior tumors resulted in reduced OT and a shorter LOS compared to TP-RPN. When surgeons aimed to routinely discharge patients on POD-1, the surgical approach did not influence LOS. Operative time was similar between RP and TP-RPN among patients with obesity or a cT1b renal mass. All other measures, including ischemia time, blood loss, margin rates, complications, and renal function, did not differ between the two approaches.

Introduction

Following the reports demonstrating comparable oncologic and functional outcome to radical nephrectomy, partial nephrectomy (PN) has become the recommended treatment for cT1a renal masses. Regarding cT1b masses, guidelines recommend PN when feasible.¹ Use of minimally invasive PN has increased since the adoption of the robotic platform. Transperitoneal (TP) approach is widely adopted due to increased working space. However, retroperitoneal (RP) approach is an alternative option to facilitate access of posteriorly located renal tumors and to avoid adhesions in patients with prior abdominal surgery.^2,3

While both approaches are safe and feasible approaches during PN, reports on comparative effectiveness between TP and RP are conflicting due to confounding by tumor location.^4–11 In these studies, TP approach is preferred for anterior tumors, as the RP approach is mostly used for posterior tumors. Only two studies have compared the two approaches exclusively for posterior tumors, with decreased length of stay (LOS) being the only replicated result.^12,13 In the largest series to date, we therefore sought to perform a propensity score (PS) matched comparison of TP and RP robotic partial nephrectomy (RPN) for posteriorly located clinical T1 tumors.

Patients and Methods

Data source

We utilized a prospectively maintained, multi-institutional database of patients who underwent PN as definitive treatment for a renal mass. Institutional Review Board approval and data sharing agreements were obtained by all institutions to participate in this multi-institutional consortium using REDCap.¹⁴

Patients

There were 1868 patients identified in the database who underwent PN as definitive treatment for a renal mass. We excluded patients with a tumor anterior (n = 547) or neither anterior nor posterior to the coronal plane of the kidney (n = 253) and patients with no data available on tumor location (n = 253). Patients were also not included if they had a clinical T2 or higher mass (n = 51) or missing data on the clinical stage (n = 7), multiple tumors resected (n = 42), prior nephrectomy (n = 47) and/or a solitary kidney (n = 23), a horseshoe kidney (n = 2), metastatic disease (n = 8), and a clamp technique other than main artery clamping, including off-clamp technique (n = 44), selective arterial clamping (n = 55), or clamp technique unknown (n = 17).

Following exclusion, there were 519 patients included in this study who underwent TP (n = 357, 68.8%) or RP (n = 162, 31.2%) RPN for a solitary clinical T1 posteriorly located renal tumor.

Variables and statistical analysis

Between patients who underwent TP and RP-RPN, baseline variables, including age at surgery, gender, body mass index (BMI), age adjusted Charlson comorbidity index (CCI), and comorbidities, including obesity (i.e., BMI >30), diabetes mellitus, chronic obstructive pulmonary disease (COPD), baseline estimated glomerular filtration rate (eGFR) (mL/min/1.73 m²) according to the modification of diet in renal disease (MDRD) formula,¹⁵ baseline chronic kidney disease (CKD) stage and tumor characteristics (stage, size [cm], R.E.N.A.L. Nephrometry score, endophytic component of the R.E.N.A.L. Nephrometry score, tumor laterality, and hilar location), and date of surgery, were compared.

We then PS matched TP and RP patients using a nearest neighbor approach with a 1:1 ratio¹⁶ using the MatchIt package in R.¹⁷ Groups were matched on age at surgery, baseline eGFR, gender, presence of obesity, diabetes mellitus (DM), COPD, clinical stage, endophytic component of the R.E.N.A.L. Nephrometry scoring system, and date of surgery to generate a balanced PS-matched cohort of TP (n = 157, 50.0%) and RP (n = 157, 50.0%).

Between the PS-matched TP and RP patients, perioperative outcomes, including operative time (OT; minutes), warm ischemia time (WIT; minutes), estimated blood loss (EBL; mL), collecting system injury, LOS (days), and positive surgical margins (PSM), were compared. Postoperative complications were classified according to the Clavien-Dindo classification system¹⁸ and were further classified as major (i.e., Clavien score ≥3), medical, and surgical as previously defined.¹⁹ Rates of overall, major, medical, and surgical complications, as well as rates of blood transfusions, were compared between groups. The progression of baseline CKD stage defined as an increase from CKD 1 to ≥2, 2 to ≥3, 3 to ≥4, or 4 to 5 was compared between groups at most recent follow-up (median 22 months; interquartile range 8.9–41.6 months; range 3.2–95.9 months).

Subanalyses of LOS between TP and RP were conducted for surgeons using a postoperative day (POD)-1 discharge protocol after RPN defined as >80% discharge on day 1 versus those without this protocol. In brief, the implementation of a POD-1 discharge protocol involves the institution of a pathway in which patients are counseled and given the expectation that they will be discharged the next day, and a checklist with timely goals (i.e., removal of catheter first thing in the morning, removal of drain if output is low, and so on) is implemented to ensure that the goal of POD-1 discharge is met. If all discharge criteria are met POD-1, the patient is discharged that day. These comparisons were done among 85 TP (n = 56, 65.9%) and RP-RPN patients (n = 29, 34.1%) who were under a surgeon's POD-1 discharge protocol and 239 TP (n = 101, 44.1%) and RP (n = 128, 55.9%) RPN patients who were not under a surgeon's POD-1 discharge protocol. Furthermore, multivariable analysis of LOS stratified by POD-1 discharge protocol was conducted.

Additional PS-matching and subanalyses were performed comparing all outcomes between TP and RP-RPN for patients with obesity (TP: n = 64; RP: n = 64) and separately for patients with cT1b tumors (TP: n = 44; RP: n = 44).

Baseline and outcome variables were compared using Mann–Whitney U tests for continuous variables and Chi-squared tests of independence or fisher's exact tests (when expected cell counts were less than five) for categorical variables. Outcome variables were only compared for the PS-matched cohorts. Poisson regression analysis was used to compare LOS between groups. A log-rank test was used to compare progression of CKD at most recent follow-up. The type I error rate in all analysis was fixed at α = 0.05 and statistical significance considered when the P value was below this level in all analyses.

Measures of association between the surgical approach and outcome were summarized with coefficients and 95% confidence intervals (CIs) generated from linear, binary logistic, Poisson, and Cox proportion hazards regression models.

Results

Baseline characteristics

Baseline characteristics of the pre and post PS-matched TP and RP-RPN patients are presented in Table 1.

Table 1.

Comparison of Demographic, Clinical, and Tumor-Specific Characteristics of Patients Undergoing Transperitoneal and Retroperitoneal Robotic Partial Nephrectomy

	Prepropensity score matched			Propensity score matched
	Transperitoneal	Retroperitoneal	P	Transperitoneal	Retroperitoneal	P
N	357 (68.8%)	162 (31.2%)		157 (50.0%)	157 (50.0%)
Clinical characteristics
Date of surgery	6/2013 (10/2011–11/2014; 11/2007–8/2016)	6/2012 (10/2010–1/2014; 9/2006–7/2016)	<.001	11/2012 (11/2010–1/2014; 11/2007–8/2016)	6/2012 (11/2010–1/2014; 9/2006–7/2016)	.855
Age	59 (49–66; 22–87)	61 (53–68; 17–88)	.042	61 (50–69; 24–87)	61 (53–68; 17–88)	.651
Gender			.112			.481
Male	205 (57.4%)	105 (64.8%)		97 (61.8%)	103 (65.6%)
Female	152 (42.6%)	57 (35.2%)		60 (38.2%)	54 (34.4%)
BMI	29.7 (25.9–34.0; 18–52)	28.6 (25.5–32.8; 21–57.5)	.306	29.1 (25.8–33.3; 18.8–50.4)	28.8 (25.6–33.1; 21.0–57.5)	.996
Obesity	169 (47.6%)	64 (39.5%)	.086	65 (41.4%)	64 (40.8%)	.909
Age adjusted CCI	3 (2–4; 0–8)	3 (2–5; 0–11)	.010	3 (2–4; 0–8)	3 (2–4; 0–11)	.053
Diabetes	79 (22.3%)	24 (14.8%)	.050	23 (14.6%)	23 (14.6%)	>.999
COPD	19 (5.4%)	17 (10.5%)	.033	18 (11.5%)	17 (10.8%)	.858
Baseline eGFR	82.4 (68.8–98.8; 15.1–214.3)	78.3 (64.2–93.3; 38.7–143.3)	.045	80.2 (65.2–94.8; 15.1–144.4)	77.9 (64.2–92.6; 38.7–143.3)	.502
Baseline CKD ≥3	52 (14.8%)	28 (17.4%)	.456	27 (17.2%)	28 (17.8%)	.882
Tumor characteristics
Clinical stage			.122			.610
T1a	282 (79.0%)	75 (21.0%)		117 (74.5%)	113 (72.0%)
T1b	118 (72.8%)	44 (27.2%)		40 (25.5%)	44 (28.0%)
Tumor size (cm)	3.0 (2.2–3.9; 0.8–7.0)	2.9 (2.1–4.1; 0.9–7.0)	.907	3.0 (2.2–4.1; 0.8–7.0)	2.9 (2.1–4.3; 0.9–7.0)	.741
R.E.N.A.L. nephrometry score	7 (6–9; 4–11)	7 (6–8; 4–11)	.674	7 (5–8; 4–11)	7 (6–8; 4–11)	.308
Tumor location
Laterality			.424			.954
Right kidney	185 (52.6%)	79 (48.8%)		76 (48.7%)	77 (49.0%)
Left kidney	167 (47.4%)	83 (51.2%)		80 (51.3%)	80 (51.0%)
% Endophytic			.139			.878
<50%	120 (34.9%)	66 (41.8%)		64 (40.8%)	66 (42.0%)
>50%	164 (47.7%)	74 (46.8%)		78 (49.7%)	74 (47.1%)
100%	60 (17.4%)	18 (11.4%)		17 (10.8%)	17 (10.8%)
Hilar	64 (18.6%)	12 (11.9%)	.117	23 (14.9%)	12 (12.0%)	.507

P values in bold indicate statistical significance at P < .05 level.

Continuous variables: medians presented with interquartiles; ranges in parenthesis; compared with Mann–Whitney U tests.

Categorical variables: frequencies presented with percentages in parenthesis; compared with chi-squared tests of independence.

ASA, American Society of Anesthesiologist Score; BMI, body mass index; CAD, coronary artery disease; CCI, Charlson comorbidity index; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; eGFR, estimated glomerular filtration rate.

In the PS-matched cohort of 157 TP and 157 RP patients, there were no significant differences noted, including age (P = .651), obesity (P = .909), DM (P > .999), COPD (P = .858), baseline eGFR (P = .502), date of surgery (P = .855), clinical stage (P = .610), tumor size (P = .741), R.E.N.A.L. Nephrometry score (P = .308), endophytic (P = .878), or hilar component (P = .507).

Outcomes

In the PS-matched cohort, there were no significant differences between RP and TP in WIT (P = .618), EBL (P = .178), collecting system entry (P = .761), PSM (P = .501), in addition to overall (P = .861), major (P = .295), surgical (P = .396), or medical (2.5% versus 7.0%, odds ratio = 0.49; 95% CI = 0.11–1.11; P = .064) postoperative complications. There were no differences in rates of blood transfusions (P > .999). No differences in progression of CKD stage (P = .599) at most recent follow-up were observed.

Operative time was significantly lower for RP compared to TP-RPN patients (157 versus 185 minutes; β = −27.5; 95% CI = −41.4 to −13.6; P < .001). In addition, LOS was significantly shorter for RP-RPN patients (median 1 versus 2 days; mean 2.42 versus 1.65; β = 0.77; 95% CI = 0.14–1.40; P = .017) (Table 2).

Table 2.

Comparison of Perioperative and Renal Function Outcome Between Patients Undergoing Transperitoneal and Retroperitoneal Robotic Partial Nephrectomy

	Prepropensity score matched patients
	Transperitoneal	Retroperitoneal	OR/β/HR (95% CI)	P
Operative time	185.0 (152.5–215.8; 111.0–400.0)	157.0 (129.8–180.3; 80.0–341.0)	β = −27.5 (−41.4 to −13.6)	<.001
Warm ischemia time	17.0 (13.3–21.6; 5.0–52.0)	17.0 (14.0–21.5; 7.0–51.0)	β = 0.11 (−1.5 to 1.7)	.618
Estimated blood loss	100.0 (50.0–150.0; 0.0–1025.0)	100 (50.0–150.0; 20.0–600.0)	β = −11.8 (−40.5 to 17.0)	.178
Collecting system entry	55 (61.1%)	99 (63.1%)	OR = 1.09 (0.64 to 1.85)	.761
Length of stay	2 (1–2; 1–47)	1 (1–2; 1–5)	β = 0.77 (0.14 to 1.40)	.017
True PSM	3 (2.4%)	5 (3.9%)	OR = 1.64 (0.38 to 7.01)	.501
Postoperative complication
Overall	19 (12.1%)	18 (11.5%)	OR = 0.94 (0.47 to 1.87)	.861
Major (Clavien ≥3)	5 (3.2%)	8 (5.1%)	OR = 1.63 (0.52, 5.10)	.295
Medical	11 (7.0%)	4 (2.5%)	OR = 0.35 (0.11 to 1.11)	.064
Surgical	10 (6.4%)	14 (8.9%)	OR = 1.45 (0.62 to 3.35)	.396
Blood transfusion	2 (1.3%)	1 (0.6%)	OR = 0.49 (0.05 to 5.54)	>.999
Follow-up
Months	20.3 (9.6–23.8; 3.5–95.9)	23.0 (6.3–40.3; 3.2–95.5)		.529
Progression of CKD stage (24 month presented)	21.6%	22.5%	HR = 0.88 (0.53 to 1.44)	.599

P values in bold indicate statistical significance at P < .05 level.

Continuous variables: medians presented with interquartiles; ranges in parenthesis, compared with Mann–Whitney U tests.

Categorical variables: frequencies presented with percentages in parenthesis; compared with chi-squared tests of independence or Fisher's exact tests.

CI, confidence interval; CKD, chronic kidney disease; HR, hazard ratio; OR, odds ratio; PSM, positive surgical margins.

Length of stay

Subanalyses showed that when the surgeon did not utilize a POD-1 discharge protocol LOS was significantly shorter for RP-RPN (median 2 versus 2 days; mean 3.16 versus 1.78; β = 1.34; 95% CI = 0.49–2.19; P = .002). However, when a surgeon utilized a POD-1 discharge protocol, LOS was not significantly different between PS-matched RP (n = 29, 34.1%) and TP-RPN (n = 56, 65.9%) patients (median 1 versus 1 day; mean 1.09 versus 1.07, β = 0.037; 95% CI = 0.09–0.17; P = .579).

Multivariable analysis showed that no factors influenced LOS when a surgeon utilized a POD-1 protocol, but when no POD-1 protocol was utilized, male (P = .013), DM (P = .049), RPNs performed earlier in the surgeon's experience (P < .001) were associated with longer LOS with a trend for older age (P = .068) and lower baseline eGFR to also increase LOS (P = .054).

TP versus RP-RPN for patients with obesity

Between 128 PS matched TP (n = 64) and RP (n = 64) RPN patients with obesity, there were no baseline differences. LOS was shorter in the RP group (mean 1.56 versus 2.73; median 1.0 versus 2.0. P < .001), with no differences in OT (187.0 versus 167.5, P = .293), WIT (P = .828), EBL (P = .355), collecting system entry (P = .328), overall, medical, surgical, major postoperative complications (P = .811, P = 273, P = .784, P = .679, respectively), PSM (P = .299), or progression of CKD stage at 12.8 months (P = .921).

TP versus RP-RPN for patients with cT1b tumors

Between 88 PS matched TP (n = 44) and RP (n = 44) RPN patients with a cT1b, there were no baseline differences noted after PS matching. There were no differences observed in OT (189.5 versus 183.5, P = .908), EBL (P = .632), collecting system entry (P = .263), overall, medical, surgical, major postoperative complications (P = .777, P = 360, P = .747, P > .999, respectively), PSM (P = .226), progression of CKD stage at 15.2 months (P = .410), LOS (P = .343), or WIT (17.7 minutes TP versus 20.0 minutes RP, P = .071).

Discussion

In the largest series to date of RP-RPN comparing 314 PS matched patients undergoing RP (n = 157) and TP (n = 157) RPN for cT1 posterior tumors, we found that a RP approach is associated with shorter OT and a shorter LOS with no differences in EBL or postoperative complications.

These findings validate a comparison of TP and RP-RPN for posterior tumors by Kim et al. who also found RP for posterior tumors to be associated with a shorter LOS with no differences in EBL and complication rates (overall and major).¹² With a larger sample size, our findings also validate the findings of Maurice et al., who also performed a PS-matched comparison of TP and RP-RPN for posterior tumors. They have also reported a shorter LOS for RP-RPN with no differences in positive margin rates, blood loss, functional outcome, recurrence, and death at median follow-up of 15 months. Finally, our findings also validate a comparative analysis of TP and RP not confounded by tumor location by Marszalek et al. who also found shorter OT and shorter LOS for RP with no differences in blood loss, complications, WIT, or immediate renal function outcome.²⁰ Our results however dispute those by Kim et al. and Maurice et al. and confirm those by Marszalek et al. since we have found that RP-RPN for posterior tumors results in a reduction of OT by 27.5 minutes. Adding to these three previous studies, results from the present study show no significant differences in blood transfusion rates or renal function outcome at a longer median follow-up (22 months) between the two approaches.

Outside of the present study, Kim et al., and Maurice et al.,^12,13 all previous studies comparing TP and RP-PN have been confounded by tumor location with the majority of anteriorly located renal tumors treated transperitoneally and the majority of posteriorly located renal tumors treated retroperitoneally.^4–11 For the most part, these studies have shown OT^4,6–9,11 and LOS^6–9,11 to be lower for RP-PN with only one study showing no difference in OT and LOS.⁵ Findings regarding blood loss are inconsistent across studies,^5–9,11 with studies consistently reporting no difference in complication rates between the two approaches.^4–6,8,11 Regardless, results from these studies are indeed confounded by tumor location; and as a result, studies showing improved OT or a shorter LOS for RP-PN are actually showing that RP for a posterior tumor is a shorter operation with a lower LOS than TP for an anterior tumor. Hence, these studies have begged the question, whether either approach yields better outcome irrespective of tumor location. We have addressed that this resulted in shorter OT and a shorter LOS and similar outcome with regards to WIT, EBL, positive margin rates, complications, and renal functional outcome.

The majority of surgeons opt for TP approach during PN because it is easier, it is the approach they have been trained in, and it has inherent technical benefits compared to RP, including a larger working area, as well as better angles for reconstruction of the kidney.^3,7 Indeed, the present study shows that most posterior tumors can be managed safely and effectively with a TP approach. However, RP approach is useful in the presence of significant adhesions and prior abdominal surgery. Furthermore, technical benefits of a RP approach during PN include superior access to posterior hilar structures and the posterolateral surface of the kidney, as well as not having to mobilize the bowel or completely mobilize and medially rotate the kidney, ultimately leading to better visualization of posteriorly located tumors.³ As evidenced in the present study, the technical benefits of RP translated to clinical benefit with patients undergoing RP-RPN for a cT1 posterior tumor having lower OT and a shorter LOS.

This decreased LOS may be attributable to the lower OT, as well as reported faster gastrointestinal recovery. Regardless, we found no difference in complication rates between groups. In addition, our study highlights that LOS depends on many other factors outside of the surgical approach. Additional factors associated with LOS included older age, DM, lower baseline eGFR, male gender, and RPNs performed earlier in the surgeon's experience. However, these factors were not associated with LOS when a surgeon implemented a POD-1 discharge protocol. Ultimately, LOS is a complex process driven by many factors, and the surgical approach is only one component of this dynamic process.

TP approach would be favorable over RP in management of obese patients with significant RP fat, and/or in the presence of a large tumor, which warrants a larger working space.^3,7 In these settings (i.e., in obese patients or cT1b tumors), our results show that RP-PN did not offer any difference in OT, EBL, complications, PSM, and renal functional outcome. We have seen a trend for WIT to be lower in patients undergoing TP-RPN for posterior cT1b renal masses; however, no benefit was seen regarding LOS. The LOS benefit of the RP approach was limited to obese patients.

The weakness of our study lies in the retrospective design and the inherent selection bias. We tried to overcome this bias by performing PS matched analysis. As prior abdominal surgery is a factor that both influences the approach to PN and can also influence TP-RPN outcome,²¹ the inability to control for prior abdominal surgery due to the lack of these data may confound these results and is therefore a limitation of this study. However, the majority of RP-RPN procedures were performed by surgeons who routinely utilized this approach for posterior tumors and, therefore, should minimize this potential bias. An additional limitation due to the retrospective nature of this study is our inability to evaluate whether there were differences in postoperative paresthesia or flank bulge which may be more likely during RP as skin incisions using this approach are closer to the spine; future studies are needed to evaluate this. A final limitation is that most of the RP cases were performed by 1 surgeon and that there are inherent biases driven by this that cannot be controlled for; however, it is likely that because all of the surgeons in this analysis are high volume, biases associated with differences in experience and proficiency are minimal.

Conclusion

Operative time and LOS were found to be shorter in patients undergoing RP-RPN. However, this effect disappeared when surgeons instituted a protocol to discharge patients on POD-1. Both approaches are safe and efficacious in the treatment of posterior renal tumors.

Footnotes

Disclosure Statement

Dr. Eun is a lecturer for Intuitive Surgical and also a consultant for Conmed and Metronic. Dr. Abaza is a lecturer for Intuitive Surgical and also had a scientific study being conducted with Conmed, Inc. Dr. Porter is a speaker/trainer for Intuitive Surgical, Inc. and receives research support from Conmed. All of these financial conflicts of interest are unrelated to and are outside of the submitted work. All other authors have nothing to disclose.

References

Campbell

, Uzzo

, Allaf

, et al. Renal mass and localized renal cancer: AUA Guideline. J Urol, 2017; 198:520–529.

Abdullah

, Rahbar

, Barod

, et al. Multicentre outcomes of robot-assisted partial nephrectomy after major open abdominal surgery. BJU Int, 2016; 118:298–301.

Patel

, Porter

. Robotic retroperitoneal partial nephrectomy. World J Urol, 2013; 31:1377–1382.

Choo

, Lee

, Sung

, et al. Transperitoneal versus retroperitoneal robotic partial nephrectomy: Matched-pair comparisons by nephrometry scores. World J Urol, 2014; 32:1523–1529.

Gin

, Maschino

, Spaliviero

, Vertosick

, Bernstein

, Coleman

. Comparison of perioperative outcomes of retroperitoneal and transperitoneal minimally invasive partial nephrectomy after adjusting for tumor complexity. Urology, 2014; 84:1355–1360.

Hughes-Hallett

, Patki

, Patel

, Barber

, Sullivan

, Thilagarajah

. Robot-assisted partial nephrectomy: A comparison of the transperitoneal and retroperitoneal approaches. J Endourol, 2013; 27:869–874.

, Gill

, Ramani

, et al. Transperitoneal versus retroperitoneal laparoscopic partial nephrectomy: Patient selection and perioperative outcomes. J Uurol, 2005; 174:846–849.

Ouzaid

, Xylinas

, Pignot

, et al. Laparoscopic partial nephrectomy: Is it worth still performing the retroperitoneal route?. Adv Urol, 2012; 2012:473457.

Wright

, Porter

. Laparoscopic partial nephrectomy: Comparison of transperitoneal and retroperitoneal approaches. J Urol, 2005; 174:841–845.

10.

Tanaka

, Shigemura

, Furukawa

, et al. Comparison of the transperitoneal and retroperitoneal approach in robot-assisted partial nephrectomy in an initial case series in Japan. J Endourol, 2013; 27:1384–1388.

11.

Kieran

, Montgomery

, Daignault

, Roberts

, Wolf

Jr . Comparison of intraoperative parameters and perioperative complications of retroperitoneal and transperitoneal approaches to laparoscopic partial nephrectomy: Support for a retroperitoneal approach in selected patients. J Endourology/Endourological Society, 2007; 21:754–759.

12.

Kim

, Larson

, Potretzke

, Hulsey

, Bhayani

, Figenshau

. Retroperitoneal robot-assisted partial nephrectomy for posterior renal masses is associated with earlier hospital discharge: A single-institution retrospective comparison. J Endourol, 2015; 29:1137–1142.

13.

Maurice

, Kaouk

, Ramirez

, et al. Robotic partial nephrectomy for posterior tumors through a retroperitoneal approach offers decreased length of stay compared with the transperitoneal approach: A propensity-matched analysis. J Endourol, 2017; 31:158–162.

14.

Harris

, Taylor

, Thielke

, Payne

, Gonzalez

, Conde

. Research electronic data capture (REDCap)—a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform, 2009; 42:377–381.

15.

Matsushita

, Mahmoodi

, Woodward

, et al. Comparison of risk prediction using the CKD-EPI equation and the MDRD study equation for estimated glomerular filtration rate. JAMA, 2012; 307:1941–1951.

16.

Rassen

, Shelat

, Myers

, Glynn

, Rothman

, Schneeweiss

. One-to-many propensity score matching in cohort studies. Pharmacoepidemiol Drug Saf, 2012; 21 Suppl 2:69–80.

17.

, Imai

, King

, Stuart

. MatchIt: Nonparametric preprocessing for parametric causal inference. J Stat Softw, 2011; 42:199–236.

18.

Clavien

, Barkun

, de Oliveira

, et al. The Clavien-Dindo classification of surgical complications: Five-year experience. Ann Surg, 2009; 250:187–196.

19.

Moskowitz

, Paulucci

, Reddy

, et al. Predictors of medical and surgical complications after robot-assisted partial nephrectomy: An analysis of 1139 patients in a multi-institutional kidney cancer database. J Endourol, 2017; 31:223–228.

20.

Marszalek

, Chromecki

, Al-Ali

, et al. Laparoscopic partial nephrectomy: A matched-pair comparison of the transperitoneal versus the retroperitoneal approach. Urology, 2011; 77:109–113.

21.

Desai

, Strzempkowski

, Matin

, et al. Prospective randomized comparison of transperitoneal versus retroperitoneal laparoscopic radical nephrectomy. J Urol, 2005; 173:38–41.