br charges Finally no differences

257 charges. Finally, no differences in mortality rates were recorded. In consequence, our findings
258 validate the feasibility and safety of robotically-assisted CRP relative to open CRP, within a large
259 population-based cohort of contemporary North American patients. However, due to missing
260 prospective studies demonstrating a survival advantage for CRP relative to no local treatment in
261 mPCa patients, CRP, regardless of the used approach, should only be performed inside of clinical
262 trials and does not represent the current standard of care for mPCa patients.
264 retrospective analysis with all of its inherent limitations. Even though retrospective analyses usually
265 meet with criticism, it should be noted that no prospective trial, comparing complications between
266 robotically-assisted and open CRP, is ongoing, designed or even planned. The conduct of such trial
267 is unlikely in the near future. In consequence, retrospective data from large population-based data
268 repositories will provide the most robust and generalizable evidence for robotically-assisted relative
269 to open CRP. Additionally, it should be noted that the definition of CRP is not standardized and the
270 definition of mPCa patients undergoing CRP through coding ICD might be inaccurate and
272 Beside its retrospective nature, our study also suffers of several other limitations that are
273 inherent to population-based analyses. For example, we were unable to adjust for tumor
274 characteristics. It could be argued that open CRP patients harbored more advanced prostate cancer
275 stages, which could adversely affect intraoperative and postoperative complication rates. In
276 consequence, adjustment for tumor stage, 3-Methyladenine node metastases, as well as lymph node dissection
277 extent, which might have been more extensive in open CRP patients, could have influenced the
278 complication rates. Moreover, we were also unable to adjust for patient characteristics, such as
279 performance status, American Society of Anesthesiologists status, differences in laboratory values
280 and opioid use, as well as presence or absence of exposure to neoadjuvant chemotherapy, androgen 13
ACCEPTED MANUSCRIPT
281 deprivation and radiotherapy. This said, despite multivariable adjustments are residual selection
282 bias may have cofounded our results. Moreover, ideally a grading system for complications, such as
283 the Clavien-Dindo system, should be used when surgical outcomes are observed 25. However, this
284 information is unavailable in large data repositories such as the NIS. Finally, only inpatients
285 information’s were available in the database we analyzed and no data were available regarding
286 readmissions and late complications 17. Last but not least, vas deferens should be underlined that our study is
287 not able to provide any information regarding the oncologic outcomes of open vs. robot assisted
288 CRP, as well as that no information on positive surgical margins and data on surgeons’ experience
289 is available within the NIS.
293 In conclusion, the intraoperative and postoperative complications associated with robotically-
294 assisted CRP are lower than those of open CRP. Similarly, robotically-assisted CRP is associated
295 with shorter stay. Conversely, an increase in total hospital charges is associated with robotically-
296 assisted CRP. Nonetheless, robotically-assisted CRP complication profile validates its safety and