br Gastrointestinal br In treatment phase expected number of
In treatment phase, expected number of GI complica-tions was 20% higher for EBRT + ADT, and 31% higher
ARTICLE IN PRESS
R. Jayadevappa et al. / Urologic Oncology: Seminars and Original Investigations 00 (2019) 1−8
Phase-specific Poisson regression for health service use and complications.
Health resource utilization, OR (95% CI)
ER Inpatient Outpatient
ADT = androgen deprivation therapy; BT = brachytherapy; CI = confidence interval; EBRT = external bean Solasodine therapy; ED = erectile dysfunction; ER = emergency room; GI = gastrointestinal; OR = odds ratio.
for EBRT + BT § ADT, compared to RP. During follow-up phase, this number was 24% higher for EBRT + BT § ADT. In terminal phase, compared to RP, number of GU complications was 55% higher for EBRT + ADT.
3.6.3. Erectile dysfunction
During treatment phase, expected number of ED compli-cations was 71% lower for EBRT + ADT, and 58% lower for EBRT + BT § ADT, compared to RP. In follow-up phase, compared to RP, this number was 45% lower for EBRT + ADT. In terminal phase, number of ED complica-tions was comparable across groups.
This study has several pertinent findings. In the elderly Medicare population, 10-year overall mortality was compa-rable between RP and EBRT + ADT. On the other hand, EBRT + BT § ADT had overall survival advantage com-pared to RP that continued until year 8 of follow-up. RP
was superior to EBRT + ADT for PCa-specific mortality between years 2 and 10 of follow-up. However, 10-year PCa-specific mortality was comparable between EBRT + BT § ADT and RP. The 10-year cost analysis demonstrated that this clinical equivalence was associated with 30% to 70% lower overall cost for RP.
For phase-specific health services use, RP was associ-ated with lower health service use during follow-up and ter-minal phases, compared to both radiation groups. Additionally, RP was associated with early (treatment phase) GU and ED complications, which emerged later for the radiation groups. This indicates opportunities for treat-ment-specific quality improvement initiatives as part of cancer survivorship care.
For high-risk patients unwilling to undergo surgery or those with at the most 10-years of life expectancy, RT could be a suitable treatment option. However, in localized PCa patients, RP was associated with a significant decline in PCa-specific mortality relative to RT and ADT . A meta-analysis of published trials compared surgery and RT outcomes among PCa patients. Surgery had better overall
ARTICLE IN PRESS
(49%) and cancer-specific survival (44%), compared to RT . One cohort study observed higher incidence of sec-ondary malignancies, and complications requiring hospital admissions, and open surgical procedures among those receiving RT . Series analyzing effects of RP in high-risk PCa patients showed PCa-specific 10-year survival rates up to 92% (72%-92%) [28,29].
A population-based study found that after adjusting for disease aggressiveness, those receiving RP had a low rate of PCa mortality . In another study, compared to high-dose RT, RP was associated with 65% risk reduction for metastasis . The mortality outcomes that we observed are consistent with results from previous clinical trials and retrospective case studies [16−20,30]. Additionally, a recent study found that if appropriate radiation doses and 2 years of ADT were given, the apparent survival benefit of ADT + EBRT + BT compared to ADT + EBRT was removed .
This emphasizes the need for appropriate treatment strat-egies as the risk of morbidity, and complications are no worse for RP compared to EBRT + ADT [20,31].
We note following limitations. Our study lacks intermedi-ate outcomes like follow-up treatment, a measure of distant disease control that would support the association of RP with survival. Our study sample consists of Medicare beneficiaries aged ≥66 years, not enrolled in HMO, with continuous cov-erage and living in a SEER region. The SEER regions have a higher proportion of nonwhite persons, and SEER mortality rates may not be representative of national cancer mortality rates . Earlier research has observed that 15% to 20% of high-risk cases were downgraded in the final pathological examination. However, our study does not address this issue. Another limitation is that a bias may exist as healthier patients are selected for surgery, and those with higher comorbidities are referred to RT. In our study, the RP group was younger with lower comorbidity, compared to the RT group. To address this issue, we determined the propensity of receiving specific treatment based on age, race and ethnic-ity, geographic location, socioeconomic status, marital status and comorbidity, and weighted our analysis by the inverse probability of PS. Finally, our study did not account for radi-ation doses. Current radiation practice uses higher doses than those used during our study period of 1996 to 2003, and may be more effective.