Page 133 - An Evidence Review of Active Surveillance in Men With Localized Prostate Cancer

instrumental variable analysis performed in one of these studies however, did not find a
significant difference between treatments (HR 0.92; 95 percent CI 0.39, 2.17); however
confidence intervals were wide indicating substantial uncertainty around the HR estimate. 207

Observational studies—Morbidity of primary treatment. One report analyzed the data from
the CaPSURE database and found that men treated with RP had a higher rate of receiving
treatments for urethral stricture than men on WW over a median followup of 2.7 years (adjusted
218
HR 10.4; 95 percent CI 3.28, 33.3). One study analyzed the risk of additional surgical
procedures following primary treatment with WW, RP, RT or ADT. 231 The reported
multivariable-adjusted estimates compared WW only with RP and demonstrated that bladder
irrigation/cystostomy procedures and TURP/bladder neck incision were more common in the
WW group compared to the RP group (HR 1.71; 95 percent CI 1.33-2.20 and HR 2.63; 95
percent CI 2.08-3.33, respectively). In contrast urethra dilation procedures were less common in
the WW group (HR 0.71; 95 percent CI, 0.61-0.84) and cystoscopy procedures were not
significantly different (HR 1.00; 95 percent CI, 0.88-1.13), compared to the RP group.

Observational studies—Quality of life. Four studies reported QoL outcomes. 226-229 Two studies
analyzed data from the CaPSURE registry. One of them reported that patients on observational
management strategies (WW or AS, reported in aggregate) had significantly lower mean SF-36
scores in the social function domain compared to patients treated with RP (89 vs. 100; P < 0.05,
229
respectively), but no difference was found in three other domains of SF-36. The other study
using CaPSURE data 227 presented trends in QoL scores from immediately after treatment
through 2 years of followup, and reported that scores for most general and disease-specific QoL
domains improved statistically significantly at one and two years post treatment compared with
scores immediately after treatment among patients who received RP. Patients on observational
management strategies (WW or AS, reported in aggregate) did not experience statistically
significant changes in their scores over time with the exception of a statistically significant
improvement in “health change” at one year of followup (compared to immediately post
treatment assignment), however the sample size of the group was small (n=87). Patients
receiving radiotherapy experienced improvements in “health change,” bowel function, and
urinary bother scores between the immediate post-treatment period and one year of followup
(however, “health change” significantly worsened between year 1 and year 2 in this group).
The third study originated from four academic centers in Wisconsin and suggested that
disease-specific QoL declined more among patients receiving RP compared to WW (for the
domains of urinary and sexual function) but did not identify a difference between treatment
groups for urinary bother, sexual bother, bowel function, bowel bother, or general QoL. 226
The fourth study included a subgroup of patients with available longitudinal data from the
Health Professionals Followup Study followed for up to 3 years to assess the changes in general
QoL and UCLA Prostate Cancer Index. The study reported that WW was not associated with
statistically significant changes on any of the SF-36 scales compared to RP (data were extracted
only for the longitudinal component of this study). 228
Studies of QoL often report multiple outcomes and perform multiple comparisons. None of
the studies we reviewed reported the use of statistical procedures that account for multiple
testing. However, we note that such procedures are not universally recommended. 233,234

79