P J Tschopp1, T Wesley-James, A Spekkens, L Lohfeld. 1. Department of Clinical Epidemiology and Biostatistics, Faculty of Health Sciences, McMaster University, Hamilton and St. Joseph's Hospital and Home, Guelph, Ontario, Canada.
Abstract
PURPOSE: We examined the durability of collagen injection response for female incontinence using Kaplan-Meier survival statistics and the log rank test. MATERIALS AND METHODS: A total of 99 women 26 to 84 years old (mean age 60) underwent collagen injections for urethral incompetence. In this retrospective chart audit treatment failures were analyzed using survival statistics. In addition, we conducted a sensitivity analysis of the data using a worst and best case model to account for variability of the data due to the high amount of censored observations. RESULTS: Mean incidence of success was 56% with a mean followup of 9.5 months (95% confidence interval 0.17 to 24.6). Median duration of success was 4.7 months (95% confidence interval 3.4 to 5.9). Of the 10 time to failure function comparisons only cystometric incontinence classification had a statistically significant predictive value for treatment failures (p = 0.003). The survival analysis curve provided a 13% success probability at 18 months and was close to the worst case time to failure function. CONCLUSIONS: Collagen injection has minimal morbidity and a low median success rate. Bladder instability, as determined by cystometrogram, is an important determinant of success. Survival statistics methods should be used more widely to describe urological outcomes because they provide dynamic and, thus, more meaningful information to urologists and patients than summary statistics.
PURPOSE: We examined the durability of collagen injection response for female incontinence using Kaplan-Meier survival statistics and the log rank test. MATERIALS AND METHODS: A total of 99 women 26 to 84 years old (mean age 60) underwent collagen injections for urethral incompetence. In this retrospective chart audit treatment failures were analyzed using survival statistics. In addition, we conducted a sensitivity analysis of the data using a worst and best case model to account for variability of the data due to the high amount of censored observations. RESULTS: Mean incidence of success was 56% with a mean followup of 9.5 months (95% confidence interval 0.17 to 24.6). Median duration of success was 4.7 months (95% confidence interval 3.4 to 5.9). Of the 10 time to failure function comparisons only cystometric incontinence classification had a statistically significant predictive value for treatment failures (p = 0.003). The survival analysis curve provided a 13% success probability at 18 months and was close to the worst case time to failure function. CONCLUSIONS: Collagen injection has minimal morbidity and a low median success rate. Bladder instability, as determined by cystometrogram, is an important determinant of success. Survival statistics methods should be used more widely to describe urological outcomes because they provide dynamic and, thus, more meaningful information to urologists and patients than summary statistics.
Authors: Sumana Koduri; Roger P Goldberg; Christina Kwon; Deborah G Dobrez; Peter K Sand Journal: Int Urogynecol J Pelvic Floor Dysfunct Date: 2005-10-14