Stress Incontinence During the Learning Curve of Holmium Laser Enucleation of the Prostate

Introduction

O ver the last decade, urologists have witnessed an expansion of techniques used for the treatment of benign prostatic hyperplasia. Since 1998, when Gilling and associates ¹ first described the procedure, holmium laser enucleation of the prostate (HoLEP) combined with mechanical morcellation has gained increasing popularity. HoLEP has received wide acclaim as treatment for patients with bladder outlet obstruction with distinct advantages over other surgical approaches, including efficacy despite prostate size, low morbidity, and shorter hospital stay. ^{1 –5}

A limitation often stated by urologists, however, is the prolonged learning curve associated with HoLEP. This poses a major obstacle to its acceptance among urologists. Numerous studies have demonstrated a steep curve, suggesting a threshold of 20 cases. ² Others have reported that an inexperienced urologist gains acceptable efficiency after approximately 50 cases. ³ Published reports identify decreases in operative time, increases in operative and morcellation efficiency, and comfort with the procedure as the end points of the learning curve. ^{1 –5} Data on decreased complication rates with surgeon experience, however, are limited. In addition, no study to date has evaluated the frequency with which the procedure should be performed to decrease the complication rate.

We retrospectively assessed urinary incontinence after HoLEP performed by a single urologist. We sought to determine if surgeon experience and frequency at which the procedure was performed affected outcomes. If so, we hoped to determine identifiable predictors of postoperative voiding complications as a means to improve surgical outcome and better patient selection, particularly during the learning curve.

Patients and Methods

A retrospective chart review was performed of all patients undergoing HoLEP in a single New England hospital from May 2002 to September 2007, yielding a total of 77 patients. All cases were performed and/or supervised by a single surgeon from the completion of preceptor training to 5.4 years in practice within a teaching institution, meaning residents were involved in nearly all cases to varying degrees. This period includes multiple periodic cessations of up to 7 months between cases because of institutional employment arrangements and maternity leave. The teaching institution had no previous laser prostate program. All patients were referred by practicing urologists.

Preoperative evaluation included a focused history with International Prostate Symptom Score (IPSS) and symptom assessment, physical and rectal examinations, ± urodynamic studies, and serum prostate-specific antigen (PSA) levels. Prostate volume was determined by transrectal ultrasonography in selected patients. The instrumentation and surgical technique for HoLEP as developed by Dr. Peter Gilling and colleagues ^6,7 have been described previously in detail.

We reviewed data related to length of hospitalization, duration of catheterization, readmission, urinary incontinence (stress, urge, and/or mixed), urinary tract infection, dysuria, and length of operating room time, which includes setup and resident teaching. Data regarding enucleation and morcellation time are not available in our retrospective data. All measures were evaluated at 3 months and 1 year. Stress incontinence (SUI) was defined as any patient wearing at least one pad for protection against urinary leakage during stress maneuvers. Patients with mixed urge and stress leakage were included in the SUI group.

All statistical calculations were computed using Stata/IC® v 10.0 (College Station, TX) for Mac OS^® X. Categorical data were analyzed using the chi-square test. The Student t test with unequal variance was used to compare two groups of continuous numerical data. All continuous variables were tested for normality. In instances where the normality assumption was violated, the Mann-Whitney U test was applied. Odds ratios were determined using logistic regression analysis. Approval was obtained from the Committee for the Protection of Human Subjects, Dartmouth Medical School's Institutional Review Board.

Results

Of the 77 HoLEPs performed at our institution between May 2002 and Sept 2007, 75 had at least 3 months follow-up. Of these, 54 had urodynamic studies: 49 (90.7%) were obstructed; 5 (9.3%) were equivocal. Patients without urodynamics had appropriate responses to medications, but they were not sufficiently relieved of their symptoms and were thought to be good candidates for surgery. Patients referred with an indwelling Foley catheter were transitioned to clean intermittent catheterization before surgery.

Nine (12%) had pure urge symptoms at 3 months and were thus excluded. For the remaining 66 HoLEPs, 17 (26%) had SUI necessitating the use of at least one pad, while 49 (74%) reported being continent at 3 months. Only two patients with SUI at 3 months remained incontinent at 1 year. One was an 84-year-old retired physician with urinary retention poststroke. He was receiving anticoagulants and had an indwelling Foley catheter, recurring infections, and hematuria; he requested surgery over the Foley catheter. The second man was a 66-year-old dialysis-dependent patient with nonoliguric renal failure that was secondary to bladder outlet obstruction who was being considered for transplantation. The first patient was in three pads per day at 1 year; the second was only in a security pad. Furthermore, all 36 patients in the continent group for whom 1-year follow-up data were available remained continent at 1 year. Interestingly, there were no differences with respect to incontinence and when the procedure was performed. Of the 25 cases of incontinence at 3 months, 7 were from the first 25 cases and 5 were from the last 25 cases (P = 0.29).

There were no differences between those with incontinence and those without as regards age at time of procedure (P = 0.54), operating room time (P = 0.18), prostate size (P = 0.85), PSA levels (p = 0.96), and IPSS (P = 0.11). The number of days between cases, however, was significantly higher in the SUI group (18.1 vs 44.5, P = 0.02) (Table 1). Therefore, further investigation into the duration between cases was conducted. Using a simple logistic regression model, the number of days between operative cases as a continuous variable was found to be significantly associated with SUI (OR 1.25; 95% CI: 1.04, 1.51; P = 0.019) (Table 2). Furthermore, as the duration between cases increased, the risk for SUI increased. Whereas the number of days between cases as a dichotomous variable (>5 weeks) was strongly associated with SUI at 3 months (OR 4.69; 95% CI: 1.21, 18.26, P = 0.026), the analysis at 1 and 3 weeks revealed only trends toward associations with incontinence.

With surgeon experience, the younger age population showed improvement in both transient incontinence (66.7% vs 11.8%, P = 0.002) and acute retention (25% vs 0, P = 0.029). For patients <70, mean age was 62 and 62.9 for groups 1 and 2, respectively. No significant change was shown in patients >70 years.

Discussion

The learning curve that is associated with HoLEP has been a source of concern for many urologists and is often referenced as a reason why they are reticent to adopt the procedure. Urologists experienced with the approach often quote 20 cases as the threshold number to have comfort with the technique. This number is rather arbitrary, however. El-Hakim and associates ² determined that a senior resident felt adept with the procedure after 20 cases when supervised by an experienced senior urologist for all cases. Shah and coworkers ³ determined enucleation efficiency and percentage of tissue resected reached a plateau at 50 cases, suggesting that while most surgeons can perform the procedure earlier, their mastery of the technique takes longer.

These studies and others address operative time, efficiency, tissue removal, intraoperative complications, conversion rates, and patient outcomes. ^{2 –5,8,9} It is accepted in surgery that all new procedures have a higher incidence of complications during the learning phase. ^{10 –12} The establishment of a new program, even with an experienced surgeon, can lead to increased complications during the learning phase of the operating room personnel and the institution. ^5,13 Despite these accepted issues with new procedures, the threshold number for surgeons to have full competence must include the point at which complications decrease. No published reports to date have determined the number required to reduce SUI with HoLEP, however, nor has it been determined at what frequency the procedure should be performed to decrease both the learning curve and complications.

We believe there is no set number of cases a surgeon must complete to become proficient. There are too many variables that must be considered, including the individual skills of a surgeon, which can vary considerably. While our study was unable to identify a threshold number of cases, we show for the first time that days between cases during the initial experience may be just as important as, if not more than, a set number. This is evidenced by the strong statistical association between time interval between cases and persistence of SUI at 3 months follow-up. We can speculate that 25 to 30 cases is likely a reasonable number to decrease complications, provided that the surgeon performs HoLEP at regular intervals to truly maximize learning.

Similar findings regarding time intervals between cases during the learning curve have been reported in both orthopedics and vascular surgery. ^{13 –15} Barizlay and colleagues ¹³ thought that long time intervals between robot-assisted spine cases and limited total numbers of cases had a significant effect on performance, concluding that spine surgeons should have a concentrated exposure over a 2 to 4 week period when learning. Howells and coworkers ¹⁴ showed that when orthopedic surgeons were trained on a simulator, then had a forced 6-month hiatus immediately after this first learning experience, all performance was lost, and they had to relearn skills. Lobato and colleagues ¹⁵ found that a ≤10-day span between endovascular abdominal aortic aneurysm repair was associated with a higher success rate.

The surgeon who performed the procedures in this study had several long-term interruptions (up to 7 months) during which no HoLEPs were performed (institutional changes, contract issues, maternity leave). This clearly had an impact on both her learning curve and complications. Initially after fellowship, SUI was less. With subsequent long hiatuses during her learning curve, however, the transient incontinence rates increased. Once she was able to perform the procedure at no less than 1 to 2 week intervals, there was a statistically significant decrease in SUI.

Presumably, most surgeons who intend to adopt the procedure will not stop and start, which will impact the proficiency number and keep it lower. An urologist who does not have the patient load to maintain regular operative frequency will have a longer learning curve and experience more SUI. This last point raises another recommendation. If there are multiple surgeons in one institution and/or practice who perform surgery for benign prostatic hyperplasia and are interested in HoLEP, it is likely best for one surgeon at a time to perform all cases and master the technique rather than several surgeons with small patient numbers learning concurrently.

This study did not identify what specific factors change when the time interval is reduced. Operative time seems the most likely factor, but we found no difference. This warrants discussion, however, because analysis of operative time in our group is fraught with confounding issues. Residents who varied as regards their year of training and experience with HoLEP were involved in all cases. Much of the recorded operative time involved equipment setup, equipment delays, teaching/demonstrating, etc. We do not know how much time was actually spent resecting and morcellating and therefore no valuable information can be gleaned from the dataset. We strongly believe that resection and morcellation time decreased with surgeon experience, both in the performance of HoLEP by the operating surgeon as well as in her teaching abilities, but we cannot prove this.

Other possible reasons why SUI is more common during the learning curve include the technique. The operative planes are challenging to learn in HoLEP and during the learning phase a surgeon may violate those planes and/or enucleate too deeply, leaving a larger fossa that allows for urine trapping and/or decreased coaptation. In addition, overdissection at the level of the apex could occur, because this portion of the procedure can be somewhat tricky in the beginning. The length of time the resectoscope is across the sphincter may lead to transient hypotonicity of the external sphincter complex, as well. As the surgeon becomes more efficient, follows correct planes, and performs more specific moves, surgical trauma/inflammation will be reduced. We also believe that shorter operative times that occur with experience lead to less transient SUI for these same reasons, regardless of our inability to prove operative time differences in this study.

An important consideration with this study is that all patients were referred by urologists. This was a more complicated patient population, because likely most straightforward patients were treated by their primary urologist and not referred. Referral reasons were mostly for medical comorbidities, anticoagulation, complicated voiding problems that did not fit into clear obstructive voiding patterns, and large prostates where the urologist had been contemplating an open procedure. That being said, only patients we determined would benefit from HoLEP underwent the procedure (proven obstruction on urodynamic studies, symptoms consistent with obstruction, and/or appropriate responses to medications). Reduction in operative and voiding complications could have occurred at an earlier number if a more typical patient population was used during the learning period.

Conclusions

The proficiency number of any new technique can only be a reference, because the skills of individual surgeons vary. To our knowledge, this study is the first to show that transient postoperative SUI is reflective of the frequency the procedure is performed during the learning phase as opposed to a set number of cases. Our patients were treated in a teaching institution and residents were involved in all cases. For the surgeon without resident involvement, the number may be less. Regardless of the occurrence of SUI during any stage of training, in nearly all patients, this was only transient. This information should be useful to the urologist beginning HoLEP and suggests that during the learning period, patients should be scheduled at short intervals.