The diffusion of robotic-assisted laparoscopic prostatectomy

One of the most remarkable developments in urology over the last decade has been the shift from open to robotic-assisted laparoscopic prostatectomy (RALP) for localized prostate cancer. This change has occurred particularly in the United States, but is also increasing in other wealthy countries of the world. At the time when Binder described the first RALP in the year 2000, the vast majority of radical prostatectomies were performed using a retropubic approach involving an infraumbilical midline incision. Pioneering work by Walsh and coworkers on the location and function of the neurovascular bundles had led to the introduction of the nerve-sparing retropubic approach in appropriately selected patients. Similar techniques were demonstrated to be effective with perineal radical prostatectomy but had never reached the widespread dissemination seen with the retropubic approach. Laparoscopic prostatectomy, the third surgical treatment option prior to the advent of RALP, was developed and technically perfected in centers of excellence in France, but for a variety of reasons, including surgical learning curve, was never widely used in the United States outside of select institutions.

Approval of the robotic-assist device, more accurately described as a master-slave system, by the US Food and Drug Administration (FDA) in 2001 and the events that followed dramatically changed the playing field for radical prostate surgery. Initially developed by the United States military, then purchased by a small start-up company called Intuitive, this technology offered the initial appeal of performing minimally invasive surgery from a remote location. It also promised to reduce blood loss and length of hospital stay and, most importantly, improve on the critical, patient-important outcomes of radical surgery, namely oncological control, urinary control, and erectile function. From the surgeon’s perspective the technique offered the advantages of 3-dimensional visualization, magnification, improved ergonomics, and (when compared with pure laparoscopic surgery) a shortened surgical learning curve. It is noteworthy that FDA approval was granted through the mechanism of a 501(k) premarket notification process. This approval process is open to technology that is perceived as sufficiently similar to other surgical devices that are already in the marketplace. The evidentiary standards for FDA approval in this framework are relatively low, in particular when compared with the hurdles faced by new drugs, which are ultimately subjected to head-to-head clinical trials to demonstrate superiority, equivalence, or at least noninferiority to an existing standard drug. As a result, very limited evidence about the actual outcomes was required to receive FDA approval.

Since that time, the rate of uptake for RALP has been remarkable. Factors that have contributed to the rapid uptake have been the minimally invasive nature of the procedure and plausible advantages over open approaches in terms of blood loss and length of stay. The rapid dissemination, however, is also a product of aggressive marketing by the manufacturer, with claims of superior oncological and functional outcomes. It has undoubtedly benefited from a broad-based fascination for new technology, which has been labeled “gizmo idolatry.”

Although nerve-sparing radical prostatectomy was a major advance in preserving erectile function, it remains an imperfect technique because of various factors. The hope of a better method, which might “guarantee” preservation of erections, has been a major driver for the adoption of this new technology. The following sections review the guiding principle of evidence-based decision making and apply it to the assessment of erectile function outcomes after RALP, with special emphasis on the implications for future research in this field.

Evidence-based clinical decision making

As it is understood today, clinical and health policy decision making should ideally be based on sound scientific research that provides high-quality evidence. For questions of therapy the highest quality of evidence is potentially provided by randomized controlled trials. These studies stand out by using established methodological techniques such as randomization and allocation concealment to create groups that are comparable at baseline. Blinding, intention to treat analysis, and completeness of follow-up then provide assurance that the groups are treated similarly (with the exception of the intervention of interest) and the results remain balanced. As a result, high-quality randomized controlled trials stand at the top of the hierarchy of evidence for questions of therapy. Observational studies are unable to offer the same safeguards against bias. Common issues of observational studies include the lack of comparability at baseline, and differential treatment and outcomes assessment of the groups being compared. Only under select circumstances are observational studies considered to provide high-quality evidence, the most common of which relates to a very large effect size that would be unlikely to be invalidated by potential biases. This particular scenario is exemplified by the example of the parachute, which has never been prospectively evaluated as an intervention to prevent death from falling from an airplane, but unquestionably works. However, interventions with such a large magnitude of effectiveness are rare in medicine, which is why randomized controlled trials are needed in most settings.

The reality is that the highest level of evidence is not often feasible for surgical interventions. Acknowledging the specific challenges of randomizing patients to different surgical approaches, prospectively designed, hypothesis-driven, nonrandomized observational studies provide the next highest category of evidence quality.

Evidence-based clinical decision making

As it is understood today, clinical and health policy decision making should ideally be based on sound scientific research that provides high-quality evidence. For questions of therapy the highest quality of evidence is potentially provided by randomized controlled trials. These studies stand out by using established methodological techniques such as randomization and allocation concealment to create groups that are comparable at baseline. Blinding, intention to treat analysis, and completeness of follow-up then provide assurance that the groups are treated similarly (with the exception of the intervention of interest) and the results remain balanced. As a result, high-quality randomized controlled trials stand at the top of the hierarchy of evidence for questions of therapy. Observational studies are unable to offer the same safeguards against bias. Common issues of observational studies include the lack of comparability at baseline, and differential treatment and outcomes assessment of the groups being compared. Only under select circumstances are observational studies considered to provide high-quality evidence, the most common of which relates to a very large effect size that would be unlikely to be invalidated by potential biases. This particular scenario is exemplified by the example of the parachute, which has never been prospectively evaluated as an intervention to prevent death from falling from an airplane, but unquestionably works. However, interventions with such a large magnitude of effectiveness are rare in medicine, which is why randomized controlled trials are needed in most settings.

The reality is that the highest level of evidence is not often feasible for surgical interventions. Acknowledging the specific challenges of randomizing patients to different surgical approaches, prospectively designed, hypothesis-driven, nonrandomized observational studies provide the next highest category of evidence quality.

The role of systematic review to guide evidence-based practice

Systematic reviews have a preeminent role in guiding evidence-based clinical practice by providing summaries of the totality of evidence. Decisions ideally should not be made on the basis of individual studies but on the totality of evidence on a certain clinical question summarizing both the benefits and harms associated with an intervention. For questions of therapy, such as the comparative effectiveness of RALP to treat clinically localized prostate cancer, such data should be derived from several randomized controlled trials performed comparing RALP to more established approaches such as laparoscopic prostatectomy, radical retropubic prostatectomy, or radical perineal prostatectomy. The methodology of how to perform systematic reviews has been pioneered by the Cochrane Collaboration and may be considered well established. Defining quality characteristics include a focused clinical question that is being addressed, defined inclusion and exclusion criteria, a systematic literature search of the published and unpublished literature, as well as disclosure of any conflict of interest of both the individual trials that are incorporated in the systematic review and the authors of the systematic review themselves. Systematic reviews should furthermore consider the heterogeneity of study results across studies that can potentially be explained by differences in the clinical (patients, interventions, outcomes) and methodological (use of methodological safeguards against bias) study characteristics.

Assessing male sexual dysfunction after radical prostatectomy

For a large proportion of men undergoing radical prostatectomy for clinically localized prostate cancer, erectile function is a critically important outcome that is only superseded by oncological control (overall survival, disease-specific survival, recurrence-free survival) and urinary control (absence of stress incontinence, urinary quality of life) in importance. In contrast to localized radiation therapy that is characterized by a slow but progressive loss of erectile function over the course of a year, patients undergoing radical surgery experience an immediate loss of their ability to achieve erections, followed by a potential recovery of function that may take up to 2 years. The recovery phase is characterized by the steepness of the recovery curve as well as the degree of sexual function that is ultimately regained. Both outcomes are influenced by several variables that should find consideration in well-designed studies that assess sexual function outcomes. These factors include the patients’ baseline erectile function and level of sexual activity, age, and the extent to which—based on the surgeon’s subjective assessment—the nerve-sparing was performed. To measure the recovery of function over time, frequent assessments are necessary to establish a time profile. Because sexual function may also be affected by urinary outcomes and pathologic features (such as stage and margin status) that may prompt postoperative radiation, erectile function outcomes should not be measured in isolation but in the context of other outcome variables.

In the absence of more objective end points such as tumescence studies, assessment of erectile function requires the use of a validated patient self-assessment questionnaire that is administered by an independent third party who is not involved in the analysis nor vested in the outcomes in any way other than the comprehensive and accurate collection of data. Examples of validated instruments that assess erectile function are the International Index of Erectile Function (IIEF) and the Sexual Health Inventory for Men (SHIM, or IIEF-5). Aside from accounting for adjuvant and salvage treatment, studies further need to account for the use of erection-enhancing interventions such as phosphodiesterase-5 inhibitors, injection therapies, and vacuum erection devices (VEDs).

Whereas much attention is given to a sample size as a (misleading) surrogate of study quality, completeness of follow-up is a much underappreciated methodological safeguard against bias that is highly relevant to observational studies of surgical interventions. It is well established that patients lost to follow-up frequently have a different prognosis to patients with complete follow-up. This scenario can thereby bias the results favorably if those lost to follow-up had worse outcomes. Well-designed studies should therefore aim to achieve a high proportion of patients with complete follow-up data and account for all patients who are lost in a flow diagram. In the commonly used Kaplan-Meier-type analysis, this requires not only the transparent reporting of the patients at risk but the clear distinction of patients who were censored because their follow-up ended or they were lost to follow-up.

Lastly, there is the issue of selective reporting and publication bias. Given the importance of erectile function to most patients undergoing radical prostatectomy as a critical outcome, one would expect most if not all studies assessing the outcomes of robotic-assisted prostatectomy to provide some assessment of the observed outcomes. Failure to do so raises the concern of selective reporting of only those sets of outcomes with favorable results, thereby biasing the body of literature toward better outcomes.

Evidence-based outcomes of robotic-assisted laparoscopic prostatectomy

In 2010 the authors reported a systematic review of the published literature on RALP. In brief, the published literature through MEDLINE and EMBASE was systematically searched, focusing on original research publications. Two reviewers independently performed the data abstraction using a standardized form derived from the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) criteria. This systematic review identified 75 original research studies. The following analysis represents a secondary analysis of a subset of studies that addressed erectile function outcomes.

Randomized controlled trials on robotic-assisted laparoscopic prostatectomy

The systematic review by Kang and colleagues identified two randomized controlled trials of RALP. Neither compared robotic surgery to alternative surgical approaches. Both studies focused on technical modifications of the robotic-assisted approach in terms of extraperitoneal versus intraperitoneal access and bladder neck reconstruction. Neither study reported sexual function outcomes; therefore no level I evidence on this topic exists.

Prospective cohort studies on robotic-assisted laparoscopic prostatectomy

Well-designed prospective studies are protocol-driven investigations with predefined inclusion and exclusion criteria, active patient follow-up, blinded outcomes assessment, and an analysis plan that is triggered by a certain length of follow-up or event rate not dissimilar a randomized controlled trial. These types of prospective cohort studies need to be carefully distinguished from retrospective analyses of prospectively collected data, which have become the predominant study type reported in the urological literature yet represents retrospective case series nevertheless.

The systematic literature review identified 4 comparative studies that addressed erectile function outcome following RALP. Although some of these were labeled as prospective, none of them met the criteria of prospective study design as defined here. There are therefore currently no prospective and dedicated studies of erectile function outcomes for robotic-assisted laparoscopic surgery to help define these outcomes or establish the comparative effectiveness of RALP to alternative surgical approaches.

Health services research on robotic-assisted laparoscopic prostatectomy

In the absence of evidence stemming from either randomized controlled trials or prospective cohort studies, the best quality of evidence on the outcomes of RALP can be drawn from a single much criticized study, based on an administrative database, by Hu and colleagues. In brief, these investigators used US Surveillance, Epidemiology, and End Results (SEER)/Medicare-linked data from 2003 to 2007 to compare the short-term complication rates and intermediate-term health-related quality of life outcomes for minimal-invasive (laparoscopic or robotic) prostatectomy and radical retropubic prostatectomy. In a propensity-score–adjusted analysis, the investigators found a significantly higher risk of being diagnosed with erectile dysfunction in the minimal-invasive than in the open surgery group (26.8% vs 19.2% per 100 person-years), thereby contradicting the widespread claims of superior outcomes of RALP compared with alternative approaches.

Several methodological issues should be recognized when critically appraising this study. First, the investigators were unable to distinguish between patients undergoing RALP versus laparoscopic prostatectomy, thereby leading to potential contamination of (superior) RALP outcomes by (inferior) laparoscopic results. However, given the low prevalence of laparoscopic prostatectomy in the United States, this contamination is unlikely to have exceeded 5% to 10%. Second, the study period likely represented the early learning curve for robotic surgery, which may not be reflective of contemporary outcomes. This is an important and valid argument but also raises serious ethical questions about how surgical procedures and devices are adopted in the urological community. Third, the study results were informed by data coded for billing purposes, rather than via patient query using validated patient self-assessment tools. In the absence of more objective data on the patients’ actual sexual function, it is difficult to accurately characterize the effects of the different surgical approaches. The higher incidence of diagnoses codes of impotence may reflect unrealistic patient expectations rather than inferior outcomes, which relates back to how RALP has been marketed not only by the manufacturer but also by the hospitals and surgeons. Recognizing all these issues, one might argue that the importance of this study stems less from its validity in assessing the erectile function outcomes, but from how it highlights the major limitations of the remaining body of evidence, which only provides very low-quality evidence for the outcome of erectile function recovery, as is now described.