Which women with stress incontinence require urodynamic evaluation?

OBJECTIVE: This study was undertaken to determine the predictive value of the symptom of stress urinary incontinence and to evaluate the ability of other factors suggested by a published Agency for Health Care Policy and Research guideline for the discrimination of patients unlikely to require urodynamic testing before surgical management. STUDY
DESIGN: We evaluated 950 consecutive women without advanced (stage III or IV) pelvic organ prolapse who were referred with symptoms of incontinence. Incontinence was recorded by means of standard forms and was characterized as "any stress loss" (76.4%), "primarily stress loss" (58.9%), "stress loss only" (29.8%), "stress and urge loss" (52.2%), "urge loss only" (13.8%), "constant and stress loss" (1.9%), or "constant loss" (2.3%). Other variables were assessed by means of a standardized history, physical examination (including urethral axis determination and stress test), 1-week urinary diary, and postvoid residual volume measurement. A urodynamic diagnosis of pure genuine stress incontinence was used as the criterion standard. Sensitivity, specificity, and positive and negative predictive values were calculated. Logistic regression models incorporating various combinations of stress loss only, previous prolapse or incontinence surgery, nocturia, voiding frequency, urethral hypermobility, and postvoid residual volume <100 mL (the factors recommended by the Agency for Health Care Policy and Research guidelines), along with age and race as predictors of genuine stress incontinence, were constructed to evaluate the predictive ability of the guideline in a subset of 447 patients for whom data on all variables were available.
RESULTS: Of the entire population 480 (50.5%) had pure genuine stress incontinence, 134 (14.1%) had both genuine stress incontinence and detrusor instability, 180 (18.9%) had pure detrusor instability, and 40 (4.2%) had intrinsic sphincter deficiency. Fifty-four (5.7%) had normal study results, and 62 (6.5%) had other nonincontinence diagnoses. Among the subjects with symptoms of stress loss only, 10.8% did not have genuine stress incontinence confirmed on urodynamic examination. Agency for Health Care Policy and Research guideline criteria had excellent discrimination (C statistic of 0.807) compared with the sole criterion of stress urinary incontinence only (C statistic of 0.574), with a positive predictive value of 85.7%. Only 7.8% of subjects met all the criteria, however, and 5.7% of these ultimately had a urodynamic diagnosis of either detrusor instability or normal study result.
CONCLUSION: The predictive value of stress symptoms alone was not high enough to serve as the basis for surgical management. Agency for Health Care Policy and Research guidelines improved the predictive value but were applicable to only a small subset of patients referred with urinary incontinence.