Predicting success using response after lead implantation with sacral neuromodulation for urgency incontinence.

Importance: Many women report inadequate symptom control after sacral neuromodulation (SNM), despite 50% reduction in urgency incontinence episodes (UUIE) after test stimulation.

Objective: To determine the ideal percent UUIE reduction after test stimulation that predicts 24-month success.

Study design: Using data from a multicenter SNM trial, we constructed receiver operating characteristic curves to identify an ideal threshold of percent UUIE reduction after test stimulation. We defined 24-month success as Patient Global Impression of Improvement of "very much better" to "better." We compared predictive accuracy of two models predicting success: (1) percent UUIE reduction alone and (2) with baseline characteristics.

Results: Of 149 women (median [IQR] baseline daily UUIE 4.7 [3.7, 6.0]), the ideal threshold for 24-month success was 72% (95% confidence interval 64,76%) UUIE reduction with accuracy 0.54 (0.42, 0.66), sensitivity 0.71 (0.56, 0.86) and specificity 0.27 (0.05, 0.55). The accuracy of the 50% reduction threshold was 0.60 (0.49, 0.71), sensitivity 0.95 (0.88, 1.0) and specificity 0.04 (0.0, 0.12). Percent reduction in UUIE was not better than chance in predicting 24-month success (concordance index [c-index] 0.47 [0.46, 0.62]); adding age, body mass index, diabetes mellitus and visual or hearing impairment the c-index was 0.68 (0.61, 0.78).

Conclusions: Among women who received an internal pulse generator (IPG) due to ≥50% UUIE reduction after test stimulation, we found no ideal threshold that better predicted 24-month success. Percent reduction in UUIE after test stimulation poorly predicts 24-month success with or without clinical factors. Given this, re-evaluating how we determine who should receive an IPG is needed.