Urinary incontinence after 125I prostate brachytherapy was evaluated using a patient self-assessment questionnaire based on the NCI Common Toxicity Criteria (version 2). Grade 0 is defined as no incontinence; Grade 1 incontinence occurs with coughing, sneezing, or laughing; Grade 2 is spontaneous incontinence with some control; and Grade 3 is no control. One hundred fifty-three patients received monotherapy (145 Gy) 125I implants between October 1996 and December 2001, and 112 (75%) responded to our survey. Median follow-up was 47 months (range, 14–74 months). Patient characteristics included a preimplant prostate-specific antigen ?10, Gleason score ?6, and stage ?T2b. CT-based postimplant dosimetry was analyzed approximately 30 days after the procedure, and dose–volume histograms of the prostate and the prostatic urethra were generated based on contoured volumes. Dosimetric parameters evaluated as predictive factors for incontinence included the prostate volume; total activity implanted; number of needles; number of seeds; seed activity; urethral D5, D10, D25, D50, D75, and D90 doses; prostate D90 doses; and prostate V100, V200, and V300. Clinical parameters evaluated included age, Gleason score, prostate-specific antigen, preimplant International Prostate Symptom Score (I-PSS), and length of follow-up.

Results

Urethral D10 dose and preimplant I-PSS predicted for urinary incontinence on multivariate analysis (p = 0.002 and p = 0.003, respectively). Twenty-eight patients reported Grade 1 incontinence (26%), and 5 patients reported Grade 2 (5%). Patients with Grade 1 and 2 incontinence were analyzed together, because of the small number of patients who experienced Grade 2. No patients reported Grade 3 incontinence. Mean urethral D10 was 314 ± 78 Gy in patients with Grade 0 compared with 394 ± 147 Gy in patients with Grades 1, 2 incontinence (p = 0.002). The incidence of incontinence doubled as the urethral D10 dose increased above 450 Gy. Patients with Grade 0 had a mean preimplant I-PSS score of 6.6 ± 4.5 compared with 10.0 ± 6.4 for Grades 1, 2 (p = 0.003). A significant increase in the incidence of incontinence was noted when the preimplant I-PSS was greater than 15. No relationship was noted between incontinence and prostate volume, total activity implanted, or the number of needles used (p = 0.83, p = 0.89, p = 0.36, respectively).

Conclusion

Urethral D10 dose and preimplant I-PSS are predictive for patients at higher risk of urinary incontinence. To decrease the risk of this complication, an effort should be made to keep the urethral D10 dose as close to the prescribed dose as possible, and the preimplant I-PSS should be thoroughly evaluated in an attempt to select patients with scores less than 15.

DISCUSSION: Prostate brachytherapy has evolved considerably since its inception and has demonstrated excellent clinical outcomes comparable with radical prostatectomy and external beam radiation therapy. Despite its success, a considerable amount of controversy continues to exist, and long-term data are now being recorded regarding morbidity, dosimetry, and quality of life.

We generated a self-assessment questionnaire based on the National Cancer Institute Common Toxicity Criteria, version 2 to evaluate urinary incontinence in our patient population and found a 31% incidence of incontinence. We evaluated patient and dosimetric parameters to determine predictive factors for this complication and found that both the magnitude of the urethral D10 dose and the magnitude of the preimplant I-PSS correlate directly with the incidence of urinary incontinence.

To our knowledge, only the study by Merrick et al. evaluated patients treated with brachytherapy alone and correlated urinary incontinence to dosimetric and clinical parameters . They found no significant difference in the EPIC scores (Expanded Prostate cancer Index Composite) of implanted patients compared with newly diagnosed controls. They reported a mean maximal urethral dose of only 130 ± 18 Gy on Day 0. Our results indicate that the incidence of incontinence would be low at this dose level. It is probable that their lack of an observed dose–response relationship is due to the low urethral doses delivered to their patients.

Several investigators have studied the relationship between the urethral dose and urethral morbidity, in which incontinence was not a specific end point. Wallner et al. found that urinary morbidity was related to the maximum central urethral dose and to the length of the urethra that received a dose greater than 450 Gy. With intraoperative planning, Zelefsky et al. were able to limit the median urethral dose to 201 Gy compared to 378 Gy with conventional planning, and they noted that acute urinary symptoms resolved more quickly with the lower doses. Desai et al. reported increased urinary toxicity scores in patients receiving higher urethral doses. Merrick et al. found that the mean membranous urethral dose was a predictive factor for urethral strictures. These reports suggest a dose response for the urethra, but to date no reported series has demonstrated an association between urethral dose and urinary incontinence. Our data clearly demonstrate a dose response for the urethra and indicate that the urethral D10 dose is a predictive factor for urinary incontinence.

In addition to determining predictive factors for incontinence, we sought to evaluate the incidence of this complication in our patient population. The risk of urinary incontinence after prostate brachytherapy has not been clearly defined with reported incontinence rates ranging from 0% to 40%. Available reports are limited and difficult to interpret because of differing implant techniques, the variety of tools used to assess this complication, the various definitions of incontinence, and the differences in the mode of data collection, such as physician grading or patient self-assessment.

The techniques of prostate brachytherapy have evolved considerably over the last decade, and a greater emphasis has been placed on urethral sparing. Talcott et al. reported a 40% incidence of incontinence in patients with implants performed using a uniform seed distribution. Postimplant dosimetry was not obtained for these patients; however, this type of seed distribution produces high central doses within the prostate and likely resulted in high urethral doses similar to those delivered in our earlier implants, when our incidence of incontinence ranged from 30% to 50%. At the other extreme, Merrick et al. reported no difference in EPIC scores of patients after brachytherapy compared with newly diagnosed controls with implants carried out using a technique that limited the mean maximal urethral dose to 130 Gy. Our results reflect this evolution of technique in that our incidence of incontinence decreased significantly between 1997 and 2001 as we reduced the urethral doses delivered. Differences in implant techniques that have an impact on the urethral dose likely contributed to the wide range of incontinence reported in the literature.

The variety of tools used to assess incontinence also contributes to the confusing results reported in the literature and to the wide range of reported rates. Several investigators used the Radiation Therapy Oncology Group toxicity scale or a modification to grade morbidity and reported 0%–5% incidence of incontinence.. This scale, however, does not specifically include incontinence as an end point or detail its severity. It is not surprising, then, that the incidence of incontinence is generally low in studies that use this scale. The I-PSS has been shown to be useful for monitoring urinary morbidity, but it also does not include incontinence. The EPIC is a frequently used self-assessment quality of life tool developed by expanding the University of California Los Angeles Prostate Cancer Index and asks 4 questions regarding urinary incontinence. Both the EPIC and the University of California Los Angeles Prostate Cancer Index are scaled to standardized values from 0 to 100, with 100 being the most favorable outcome. This tool provides an evaluation of the impact of incontinence on the patient's quality of life, but it does not describe the crude rates of incontinence in the patient population, making comparisons with studies using this scale difficult.

Another reason for the controversy about the incidence of incontinence after prostate brachytherapy is the definition of incontinence. Incontinence is considered any involuntary leakage of urine, but this general term encompasses occasional leakage of a few drops to complete loss of the capacity to contain urine. Ragde et al., for example, reported the incidence of incontinence to be 0% in patients without a history of transurethral resection of the prostate. However, patients were classified as incontinent only if they required the use of a protective pad. Many patients who reported incontinence in our study would have been regarded as continent using these criteria, which would have resulted in a significantly lower incidence of incontinence. How investigators define incontinence has a significant effect on the incidence reported.

Differences in the mode of data collection, be it physician grading or patient self-assessment, also have been shown to significantly affect results. Some reported studies are based on physician assessment of patient symptoms, whereas others are based on self-assessment. In the CaPSURE database of more than 3,900 patients, physicians rated patients in several areas, including urinary incontinence. These data were compared with patient-completed self-assessment questionnaires, and significant differences were noted in all quality of life and clinical domains. The incidence of incontinence, then, varies with different modes of data collection and is likely to be an underreported complication in studies based solely on physician assessment.

When evaluating incontinence in brachytherapy patients, it is important to know the incidence of this problem in the general population. In studies of men without prostate cancer or patients who opted for observation of their cancer, the incidence of incontinence has not been as low as could be expected. In a population of normal older men, Litwin found that 33% had some degree of urinary leakage. Similarly, in a review of the literature including over 21 studies and 12,000 men, Thom found the prevalence of incontinence among community-dwelling older men to be 11%–34% with a median of 17% . Reports in the literature indicate that incontinence is not uncommon in the general population and highlight the fact that a portion of brachytherapy patients would be expected to experience some level of urinary leakage independent of their diagnosis or its treatment. Series that report extremely low incontinence rates may be underestimating this symptom for any of the reasons noted in this discussion.

It is understandable, considering all of the above, that a wide variation exists in the incidence of incontinence after prostate brachytherapy as reported in the literature. More standardization is needed in the definition of incontinence, in the tools that are used to assess incontinence, and in the manner in which the urethral dose is calculated and reported. Investigators also need to report the urethral dose and preimplant I-PSS when reporting the incidence of incontinence.

Similar to sexual dysfunction, urinary incontinence is a feared complication of prostate cancer treatment and is, therefore, an important end point to evaluate when investigating treatment outcomes and quality of life. There are several established types of incontinence, including urge incontinence (detrusor overactivity), neurogenic incontinence (detrusor underactivity), overflow incontinence (urethral obstruction), and stress incontinence (urethral incompetence). The mechanism of incontinence in patients that have received prostate brachytherapy is not well defined. Our study does not attempt to speculate on the mechanisms of incontinence in our population; nor does it indicate how incontinence may change with time, age, medication, or surgical intervention. An attempt to collect this information was made via telephone interviews; however, these details were difficult for patients to recall, precluding reliable conclusions.

The data presented identify two risk factors for urinary incontinence, urethral D10 and the preimplant I-PSS. These factors should be used to guide physicians' recommendations but do not represent contraindications to implantation. In selecting patients for this procedure, a global picture, including stage of disease, Gleason score, PSA, medical comorbidities, prior urologic procedures, medications, urinary symptoms, and preimplant dosimetry, should be evaluated. Limiting the urethral D10 dose and selecting patients with preimplant I-PSS less than 15 will minimize the incidence of urinary incontinence in patients receiving prostate brachytherapy.