Page 132 - An Evidence Review of Active Surveillance in Men With Localized Prostate Cancer

pulmonary embolism and deep-vein thrombosis were uncommon (cumulative one-year incidence
proportions 2.1 percent, 1.4 percent and 1 percent, respectively). Data on adverse events in the
WW group for the first year were not provided.
An ancillary investigation 214 from the Swedish component of the SPCG-4 RCT (including a
total of 326 patients) evaluated the quality of life outcomes at different followup timepoints (2 to
3 years, 4 to 5 years, and 6 to 8 years postrandomization). The study found that anxiety
(moderate or high) and depressed mood (moderate or high) were less common in the RP group
compared to the WW group, and that sense of well being and quality of life were not
significantly different between groups at 2-3 years post randomization. Depressed mood was
more common in the RP group compared to the WW group at 4-5 years post randomization;
significant differences were not observed for other quality of life outcomes. No significant
differences were observed at and 6-8 years post randomization. This ancillary investigation also
reported that about 80 percent and 42 percent of patients in the RP group reported erectile
dysfunction and urinary leakage, compared to 37 percent and 11 percent in the WW group
respectively 2 to 3 years post randomization. At 6-8 years post randomization the corresponding
frequencies were 83 percent and 56 percent in the RP group and 55 percent and 25 percent in the
WW group.

Observational studies—Cancer-specific mortality. Three cohort studies 207,209,223 compared
prostate cancer-specific mortality between patients on observational strategies and patients
treated with RP based on large databases (SEER-Medicare, 207 the National Cancer Register of
Sweden Follow-up Study, NCRSFS, 223 and the Connecticut Tumor Registry 209 ). All three studies
identified a statistically significant difference between treatments favoring RP (HR 0.63; 95
percent CI 0.55, 0.71 [using a propensity score 207 ], HR 0.49; 95 percent CI 0.34, 0.71, 223 and HR
209
0.29; 95 percent CI 0.17, 0.52, respectively). However, an instrumental variable analysis of the
SEER-Medicare study did not identify any significant difference between treatments with wide
confidence intervals around the estimated treatment effect (HR 1.37; 95 percent CI 0.15, 12.5). 207

Observational studies—All-cause mortality. Five studies compared all-cause mortality
between observational management strategies and RP. 207,208,223,230,232 One was based on the
CDC-NPCR Patterns of Care Study (POCS), 230 two used the SEER-Medicare database, 207,208 one
used NCRSFS data, 223 and one used the Center for Prostate Disease Research database to
identify men of 70 years of age or older who were diagnosed with localized prostate cancer. 232
From CDC-NPCR POCS, the 5-year all-cause mortality rate was significantly higher in patients
230
on WW than in patients treated with RP (adjusted HR 2.3; 95 percent CI 1.70, 3.12).
Similarly, the two reports that analyzed data from the SEER-Medicare database, 207,208 one report
analyzed data from the Center for Prostate Disease Research database, 232 and the NCRSFS
223
study showed favorable outcome in the RP group (HR 0.65; 95 percent CI 0.62 ,0.68[using a
propensity score 207 ]; adjusted HR 0.49; 95 percent CI 0.41, 0.57 207 ; adjusted HR 0.50; 95 percent
CI 0.47, 0.53; 208 and adjusted HR 0.52; 95 percent CI 0.32, 0.84, 232 respectively; the latter
estimate was derived from an analysis comparing WW to RP (reference group), where patients
initially managed with WW were stratified by whether they later received secondary therapy.
This analysis may be susceptible to bias because the decision to administer secondary treatment
is often made based on information not available at baseline (e.g., tumor progression status
during followup); and this information is likely to be associated with the clinical outcome of
interest. This study was exclusively limited to patients who were aged 70 years or older. 232 An

78