BPH Treatment Device Safety in the FDA’s MAUDE Database Using Total Procedural Context: PUL Exhibits the Lowest Complication Rates, WVTT Intermediate, and AQB the Highest

1. INTRODUCTION

Benign Prostatic Hyperplasia (BPH) is a progressive disease associated with Lower Urinary Tract Symptoms (LUTS) affecting more than half of men over age 50 in the United States (US) [1]. Minimally Invasive Surgical Therapies (MISTs) can offer lasting relief from LUTS with an improved patient experience, treatment in an outpatient setting, minimal or no post-procedure catheterization, and improved early recovery compared to more invasive therapies such as Transurethral Resection of the Prostate (TURP) [2]. While these therapies have been shown to be safe and effective in clinical trials, their real-world safety profiles continue to benefit from ongoing surveillance.

The Manufacturer and User Facility Device Experience (MAUDE) database is a publicly accessible, prospective registry of Medical Device Reports (MDRs) submitted to the US Food and Drug Administration (FDA) by mandatory reporters, such as manufacturers, importers, and device user facilities, and voluntary reporters, including healthcare professionals, patients, and consumers [3]. These reports are designed to capture events in which a device may have caused or contributed to a serious injury or death, or in which a device malfunction would likely cause harm if it were to occur [4]. The MDRs contained in MAUDE, although limited and not formally adjudicated [5], constitute a valuable source of patient experience and safety data on the real-world performance of BPH MISTs outside of controlled trial settings. In the context of procedural safety analysis, MAUDE is especially valuable in the early identification of Adverse Event (AE) trends associated with specific devices that may not have emerged in pre-market trials [6]. MAUDE is limited by its failure to capture denominator data, which precludes estimation of event incidence rates because the reports are not normalized to the “at-risk” population. While previous analyses of BPH treatment devices have used MAUDE to examine safety and patient experience patterns [7-9], these analyses failed to incorporate data on the full procedural context to assess the national incidence of MDRs. This study aimed to characterize adverse event reporting for BPH devices of interest through formal adjudication, while also estimating procedural denominators using sales and administrative claims data, thereby situating the MDRs of interest within the broader context of real-world clinical practice. This approach enabled a pragmatic assessment of risk by enhancing the interpretability of MAUDE-reported events relative to the total procedural context.

2. MATERIALS AND METHODS

3. RESULTS

4. DISCUSSION

New technologies and procedures continue to capture an increased share of the BPH surgical treatment market, driving TURP rates from 39% of all BPH surgeries in 2015 to 26.9% in 2021 [14]. While randomized controlled trials offer the highest level of evidence for the safety and efficacy of a procedure in the controlled population studied, once deployed into broader patient populations, real-world results may indeed vary [15-17]. Surveillance data, such as those collected in MAUDE, contribute to an expanded and nuanced understanding of BPH therapeutic performance in large, real-world populations beyond controlled settings. Real-world database data can uncover effects such as the physician learning curve and technique development, as well as adverse effects that cannot be detected in relatively small, controlled studies. Additionally, it can serve as a check on marketing claims that might be based on a small study yet extrapolated to the general population across unstudied cohorts. Only data generated by randomizing between procedures can yield high-confidence comparisons of these treatment options. As such, this study provides instead a view of how each treatment option is performing in the patient populations to which urologists applied them during the period studied.

Given the heterogeneity of MAUDE data sources, which range from voluntary submissions by clinicians and patients to mandatory reports from manufacturers, we approached the dataset with caution, adjudicating each MDR for quality, clinical relevance, and narrative completeness. Previous studies examining BPH devices have utilized MAUDE to identify post-market safety signals, though most reported absolute event counts without contextualizing these figures against real-world procedural volumes [7-9]. Our findings confirm that MDR frequency tends to rise in parallel with device utilization, a trend observed in other post-market surveillance studies [18]. As such, raw counts of adverse events, while valuable for signal detection, may be misleading when used in isolation. Volume-adjusted event rates provide a more meaningful framework for clinical interpretation and comparative safety assessment.

PUL adoption increased since its FDA clearance in 2013, accounting for approximately 30% of BPH surgical procedures between 2019 and 2022 [19, 12, 20]. AQB has seen a rapid increase in utilization since its introduction in 2019, with total sales increasing by 97% between 2021 and 2022 [21-23]. Following FDA approval in 2015, WVTT annual utilization decreased 48% between 2019 and 2021 [24, 25].

The use of MAUDE data in published analyses increased during the COVID-19 pandemic; in 2019, 45 publications used MAUDE data, while by 2022 the number had increased to 89 [26]. PUL, WVTT, and AQB all saw increased total MDRs submitted over the course of the study, perhaps signaling an increased awareness of the MAUDE database’s existence amongst voluntary reporters.

Intraoperative events for the BPH treatment devices in this analysis were less severe overall than post-operative events, which may highlight an inherent bias in the MAUDE sample – the database is designed to capture device malfunctions which, particularly in the intraoperative setting, are often as mild as a failure in the user interface and may be corrected with minor intervention, resulting in lower Gupta scale ratings. In the case of PUL, these intraoperative events are often failed device deployments remedied by obtaining a new device. The WVTT and AQB intraoperative event narratives were similar in describing device malfunctions; however, obtaining a new device in both cases is considerably more difficult. The reliance on an electronic interface to complete the WVTT and AQB procedures means that a malfunction often mandates procedural abandonment. The decrease in intraoperative event rates for AQB between 2019 and 2022 may be explained by the technology’s relative novelty in 2019, with physicians gaining increased proficiency in the procedure by 2022. In terms of overall intraoperative event rates, as well as serious or life-threatening events, AQB’s intraoperative safety profile differs from the MISTs studied, consistent with the procedure’s more invasive nature.

The decrease in post-operative events as a proportion of total MDRs between 2019 and 2022 across all devices may signal greater physician experience in performing the procedures, leading to fewer adverse events. AQB is performed using a robotically controlled high-velocity waterjet to resect obstructing prostatic tissue, employing a technique similar to TURP but without the use of electrocautery. AQB’s 15% severe post-operative complication rate observed in 2019 points to the difficulty in achieving hemostasis with water therapy alone, reminiscent of the 9.9% transfusion rate observed in the WATER II trial [27]. In January 2020, PROCEPT BioRobotics, the makers of the Aquabeam System, implemented global training to perform focal bladder neck cauterization to achieve hemostasis following the primary AQB procedure [28]. The subsequent decreases in overall MDRs and severe post-operative Clavien-Dindo events for AQB in MAUDE appear to reflect the effectiveness of this procedure update in reducing bleeding. However, overall and serious MDRs remain elevated for AQB compared with PUL and WVTT, suggesting that AQB has a safety profile distinct from minimally invasive BPH treatments and is not overcome by bladder neck cauterization.

Adverse event adjudication in a clinical trial setting includes a determination of the relatedness of the event to the device and remains a vital component of scientific analysis; while other publications have retroactively attributed relatedness to serious or life-threatening events (for example, bladder perforation during TURP attributed to prior PUL procedure) for which no such attribution existed in the MAUDE database itself, our analysis is the first to adjudicate MDRs in a more rigorous manner [29]. Because many MDRs in MAUDE are submitted by voluntary reporters, including patients with varying levels of medical fluency, the quality of entries is heterogenous, constituting the entire bell curve from social media posts to full case reports; to account for this heterogeneity, we have employed a novel completeness scale which rates the ability to determine if the patient received treatment for the stated event, if that determination could be incompletely made, or if no information about interventions was provided. These methods aim to facilitate rigorous analysis, and we encourage all scientists to account for data quality and event-relatedness when approaching this valuable yet highly heterogeneous data source.

Once these methods were applied, the most common AE classification in WVTT was infection, which may be consistent with the longer indwelling catheterization durations following that procedure. Bleeding complications following AQB and PUL manifest in the classified hematomas and transfusions. When relatedness and completeness are accounted for, PUL emerges with the fewest events. When similar standards are applied to AQB’s MDRs, particularly severe to life-threatening events, we see that most MDRs are clearly related to the procedure and were treated with a return to the operating room for hemostasis. Although the grading of related and complete MDRs produces event rates that are lower for all BPH treatments in this analysis, rates of 2 vs. 5 events per 10,000 cases lack clinical meaningfulness. As would be expected for a minimally invasive treatment for a quality-of-life disease, mortality following MIST procedures is rare and often unrelated to the procedure.

The MAUDE database offers insights into real-world device performance in the hands of practicing urologists beyond the confines of a clinical trial. However, MAUDE is a limited sample, with entries that, according to the FDA, are potentially biased, unverified, untimely, incomplete, or inaccurate [30]. Data collection is also limited by under-reporting of events by professionals, unclear submission responsibilities, and continued lack of awareness of the database itself [31, 32]. The extraction of MDRs without context, including total procedures per year, patient baseline demographic data (e.g., gland size, International Prostate Symptom Score [IPSS], and comorbidities), should be undertaken with caution. Despite these limitations, MAUDE provides an additional opportunity to prioritize patient experience in shared decision-making when approaching the myriad treatment options for BPH.

CONCLUSION

The MAUDE database constitutes a valuable source of up-to-date information, providing early insights into device safety and performance in the real world. In using the MAUDE database, previous publications have failed to account for event incidence, at-risk population sizes, device-relatedness, and data quality. When these features are accounted for, AQB emerges with significantly higher rates of mild, moderate, and severe events than the minimally invasive therapies analyzed. PUL has the lowest year-over-year rates of mild, moderate, and severe events in the MAUDE database.

REFERENCES