Adoption of Technology and Its Impact on Nephrectomy Outcomes,a U.S. Population-Based Analysis (2008–2012)

Introduction

Radical nephrectomy (RN) remains the standard of care for surgical management of large renal tumors or tumors not amenable to nephron-sparing surgery. For decades, open radical nephrectomy was the preferred and recommended approach for all renal tumors. In the early 1990's the first documented pure laparoscopic radical nephrectomy (LRN) was performed. ¹ Since then, LRN has had a well-documented increase in utilization ^2,3 and by many is currently considered the gold standard for RN. ⁴

The first robotic radical nephrectomy (RRN) for renal-cell carcinoma (RCC) was described by Klingler at al. in 2005. ⁵ Since its introduction, multiple case series have demonstrated that RRN is both a safe and effective approach to renal extirpative surgery. ⁶ Exclusion criteria for RRN and LRN are no different, and both are considered suitable options for surgical management of renal tumors. ⁷

LRN and RRN are often combined to report on generalizations about minimally invasive surgery (MIS). Nevertheless, from a technical standpoint there are distinct differences between the approaches, with clinical, oncologic, and cost implications. Pure laparoscopic surgery is thought to have a steeper learning curve than a robotic platform, as laparoscopic surgery has limited maneuverability. ⁸ Robotic surgery offers potential advantages in surgical dexterity and intraoperative imaging, which may facilitate hilar dissection and control of the kidney vessels. ⁷

Although the robotic approach to RN is increasing, ⁹ little data to support this increased adaptation exist, especially comparing the two surgical approaches in regard to perioperative and oncologic outcomes. We sought to compare LRN vs RRN in regard to cost, complications, and survival.

Methods

Results

Intermediate-term survival

During a mean follow-up of 3.2 years, 101 patients undergoing LRN and 31 patients undergoing RRN died from any cause. More specifically, 24 patients in the LRN and <10 patients in the RRN group died from kidney cancer during follow-up. Before propensity score matching, the estimated overall survival at 3 years was 88.0% (95% CI 82.5%, 91.8%) in RRN and 83.7% (95% CI 80.3%, 86.6%) in LRN groups. The estimated RCC-specific survival at 3 years was 97.8% (95% CI 94.8%, 99.1%) and 95.7% (95% CI 93.6%, 97.1%) in the RRN and LRN groups, respectively. After propensity score matching, the estimated overall survival was 88.0% (82.5%–91.9%) for RRN compared with 87.9% (82.5%–91.7%) in the LRN group. The estimated RCC-specific survival was 98.1% (95% CI 95.1%, 99.3%) vs 96.4% (95% CI 92.9%, 98.2%) in RRN vs LRN patients, respectively (Table 3). Generally, overall survival (Prentice–Wilcoxon test: p = 0.90) and RCC-specific survival (Prentice–Wilcoxon test: p = 0.25) were similar between the two groups in the matched cohort (Fig 1).

OPEN IN VIEWER

FIG. 1.

Kaplan–Meier estimates with number of subjects at risk and 95% Hall–Wellner Bands for overall and cancer-specific survival among Medicare beneficiaries undergoing robotic vs laparoscopic radical nephrectomy.

OPEN IN VIEWER

Table 3.

Estimated Overall and Cancer-Specific Survival at 3 Years Following Laparoscopic and Robotic Radical Nephrectomy

	Full cohort			Matched cohort
	Laparoscopic (n = 574)	Robotic (n = 241)	p	Laparoscopic (n = 230)	Robotic (n = 230)	p
Overall survival			0.27			0.90
At 3 years	83.7% (80.3%–86.6%)	88.0% (82.5%–91.8%)		87.9% (82.5%–91.7%)	88.0% (82.5%–91.9%)
Cancer-specific survival			0.15			0.25
At 3 years	95.7% (93.6%–97.1%)	97.8% (94.8%–99.1%)		96.4% (92.9%–98.2%)	98.1% (95.1%–99.3%)

Discussion

RRN is considered a safe procedure and a suitable option for patients who require surgical removal of the entire renal unit. ⁵ However, the utility of the robotic platform in the setting of RN has been questioned largely due to concerns such as cost and increased time and effort associated with robotic docking. ¹⁶ Nevertheless, it has been difficult to conclude if further advantages offered from robotic technology exist, as data on direct comparison of RRN to LRN are few and often consist of small cohorts with limited follow-up. ¹⁷ In the current population-based contemporary study, we found no difference in survival or adverse event rates between patients undergoing RRN or LRN, although cost estimates were significantly higher with RRN. These results contribute to the limited but growing data that fail to demonstrate a surgical or economic benefit to use of a robotic platform in the RN setting.

We found no difference in cancer-specific or overall survival between RRN and LRN with intermediate-term mean follow-up of 3.2 years. These results contribute to prior studies that have demonstrated short-term oncologic efficacy of minimally invasive RN. A recent systematic review by Asimakopoulos and colleagues ⁷ included 10 studies evaluating the oncologic efficacy of RRN for clinically localized and/or advanced cases of RCC. Of those, only one study in 2009 by Hemal and Kumar ¹⁷ directly compared the efficacy of RRN to LRN. In comparing 15 patients who underwent RRN with 15 patients who underwent LRN for cT1-T2 RCC, authors found no difference in local, port-site, or distal recurrence rates after a mean follow-up of 8.3 and 9.1 months, respectively. Proper use of robotic arms can reduce the complexity of laparoscopic tasks being performed by the bedside assistant and, therefore, maximize console surgeon independence. ¹⁸ While it stands to reason that this could lead to different outcomes between the two approaches, our results do not support this theory.

We found no difference in major in-hospital events, procedural complications, or LOS between the two approaches. Potential advantages of the robotic approach over laparoscopic have been suggested, which may provide better perioperative outcomes for RRN. For instance, the use of an articulated clip applier allows for the surgeon to approach the kidney vessels at angles not achievable with a conventional laparoscopic clip applier, and EndoWrist technology may allow for easier ligation of the kidney vessels. However, we saw no difference in major events between groups, including death, ICU use, and blood transfusion. Our results at the population-level complement single-center or single-surgeon studies reporting no significant difference in measured perioperative outcomes. ^19,20

Total inpatient charges were significantly higher when surgery was performed by a robotic approach. Our results are consistent with multiple reports that have shown robotic surgery to be equal to or more expensive than laparoscopic surgery. ²¹ Using the Nationwide Inpatient Sample database, Yang and colleagues ²² found median total charges were higher for RRN vs LRN ($47,036 vs $38,068, p < 0.001), finding total charges and hospital costs of RRN were around $9000 and $3500 more than LRN, respectively. A single institution study by Boger and colleagues ²⁰ concluded that robotic assistance adds about $1300 in direct costs to surgery compared with a laparoscopic approach and nearly $400 more in disposable costs per patient. In a prospective comparison, Hemal and Kumar ¹⁷ reported RRN added a mean of 46 minutes of operative time compared with LRN. Although they did not directly analyze for cost, others have suggested time-based operating room costs are around $54/minute for the first 30 minutes and $11/minute thereafter. ²⁰ In the setting of partial nephrectomy, Alemozaffar and colleagues ²³ reported no difference in variable costs between robotic-assisted and laparoscopic approaches, and Yu and colleagues ²⁴ reported no difference in costs among open, laparoscopic, and robotic partial nephrectomy. This suggests that the need for reconstruction (i.e., partial nephrectomy) may be the difference in a procedure being cost neutral or cost negative. ²² However, when removing the entire renal unit, our estimates are consistent with previous reports of significantly higher costs incurred when utilizing a robotic platform.

Finally, we identified several sociodemographic and tumor characteristics associated with the surgical approach. Higher utilization of RRN over LRN was seen in those who live in metropolitan areas, consistent with reports that patients undergoing total nephrectomy at hospitals in moderately or highly competitive markers are more likely to undergo robotic-assisted surgery. ²⁵ In addition, we noted some baseline differences in patient characteristics between surgical approaches, including race and education level. Racial, gender, and socioeconomic status have been previously investigated in the setting of RCC. In small renal masses, black patients are more likely to undergo nonsurgical management and less likely to undergo partial nephrectomy than RN if operated on. ²⁶ Similarly, men are more likely than women to be nonsurgically managed, but more likely to receive a partial nephrectomy. Risk aversion, access to care, and acceptability of novel cancer treatments have all been implicated in treatment disparities. Furthermore, marketing for robotic surgery is increasingly web based, often promoting improved outcomes without describing a comparison group or making reference to evidence-based data to support claims. ²⁷ Thus, it is possible that susceptibility to marketing may factor into patient decision-making, as inequalities in communication, leading to differences among social groups to access, seek, process, and use information in direct to consumer advertising, are well documented. ²⁸ However, it is also possible that sociodemographic differences in our study are reflective of unidentified confounders, as by utilizing an administrative data set designed for billing purposes, we fail to capture individual patient and physician preferences that impact operative decision-making. Regardless, as we found no difference in perioperative or survival outcomes between surgical approaches, it is unlikely these differences led to any significant treatment disparities.

This study possesses several strengths. Through utilization of two large population databases (SEER-Medicare), we provide a comprehensive representation of patient outcomes across the United States. In doing so, we provide population-based evidence that with intermediate-term follow-up, no difference in overall or cancer-specific survival exists between minimally invasive surgical approaches to RN. These results have not been previously reported on this scale, and we believe they have important significance to the patient, surgeon, and medical community. That said, our study does have limitations, as we were unable to capture patients who underwent minimally invasive RN before 2008, when robotic codes became effective. While this limits our sample size, it frames our study in a contemporary, relevant setting.

Second, as Medicare provides healthcare benefits to Americans ≥65 years old, our results may have limited generalizability to younger patients and those treated internationally or in non-SEER regions. Third, while propensity score methods allowed us to adjust for known confounders, neither could we adjust for unobserved differences such as tumor complexity nor were we able to distinguish pure laparoscopic from hand-assisted laparoscopic surgical technique. While some have suggested that RRN may minimize the technical difficulties associated with laparoscopy and permit surgeons to undertake more challenging cases such as complex IVC thrombi, these are currently limited to case reports and modest case series. ^29,30 In addition, we were unable to compare measures of postoperative convalescence, such as quality of life metrics or analgesic requirements, as SEER does not capture this information. However, we would not anticipate a difference in these measures between the different minimally invasive approaches. Finally, we do not account for the capital and maintenance costs of robotic equipment, which therefore limits our cost analysis. This study has adhered to all appropriate ethical standards. This study is IRB exempt.

Conclusion

The robotic platform demonstrated no benefit over standard laparoscopy for RN, as we found no difference in oncologic efficacy or adverse event rates. Given escalating medical costs and focusing on reducing low value care, robotic nephrectomy offers no advantage over standard laparoscopy for patient outcomes.