9
Device Survival
Although it is not possible to predict exactly how long
an implanted prosthesis will function in a particular patient,
American Medical Systems, Inc. gathered data from two sources
on device removals and revisions to help gain insight into
product performance over time. Figure 1 presents device survival
results from the prospective clinical study and a Bayesian analysis
that uses data from the prospective clinical study and the PIF
Study to estimate device survival at five years.
Prospective Clinical Study
– A device survival curve was calculated
from data collected during a prospective clinical study (n=85)
with two-year follow-up. Using Kaplan-Meier analysis, the two-
year revision-free rate for the AMS 800 was 79.5% (95% CI
with 95% lower confidence bound 69.8%).
Bayesian Analysis
– A Bayesian hierarchical model was used
to evaluate device safety in the prospective clinical study. The
Bayesian model estimated device survival using historical data
(PIF Study n=12,713) on the AMS 800 and prospective clinical
study data (n=85) on the AMS 800. A log-normal distribution
fit the AMS 800 historical data. Based on the log-normal
hierarchical model, it was estimated that the five-year revision-
free rate for the AMS 800 is approximately 73.8% with 95% CI
ranging from 67.3% to 79.6%. The results met the primary safety
endpoint for the clinical study of a five-year revision free rate at
75% using a 10% delta with two-sided 95% lower bound greater
than 65%.
Estimated Survival Rates
Re
vision F
ree R
at
e
Time in Months
- - - -
Prospective Survival Rate
_____ Bayesian Survival Rate
Figure 1: Estimated Survival Rates for the AMS 800
Device Use
The study included male AMS 800 patients (n=390) implanted
between 1987-1990. Data from this study was used to estimate
the device use for patients (n=356) with available data through
1997 (range: 0.03-10.3 years). For the retrospective study,
“device use” means the span of time from implant to removal,
including revisions. Life table analysis indicated that the
probability for a 9-year span of device use was 83.9%. Thirty-
three (33) of the 356 patients had their device removed. The
remaining 323 patients had a functioning device at last contact.
d. Prospective PIF Post-market Registry Study – The AMS 800™
Urinary Control System, with InhibiZone™ Treatment, received
FDA approval in 2006 with the approval condition that AMS
conduct a five (5) year registry study. The actual average follow-
up period study was 5.8 years.
The Inhibizone™ (IZ) antibiotic treatment, a combination of
minocycline and rifampin, is impregnated into the surfaces
of AMS 800 Cuff and Control Pump components in order
9
