Background: The Deyo adaptation of the Charlson comorbidity index (DaCCI), which relies on 17 comorbid condition groupings, represents one of the most frequently used baseline comorbidity assessment tools in retrospective database studies. However, this index is not specific for patients with bladder cancer (BCa) treated with radical cystectomy (RC). The goal of this study was to develop a short-form of the original DaCCI (DaCCI-SF) that may specifically predict 90-day mortality after RC, with equal or better accuracy. Patients and Methods: Between 2000 and 2009, we identified 7,076 patients in the SEER-Medicare database with stage T1 through T4 nonmetastatic BCa treated with RC. We randomly divided the population into development (n=6,076) and validation (n=1,000) cohorts. Within the development cohort, logistic regression models tested the ability to predict 90-day mortality with various iterations of the DaCCI-SF, wherein <17 original comorbid condition groupings were included after adjusting for age, sex, race, T stage, and N stage. We relied on the Akaike information criterion to identify the most parsimonious and informative set of comorbid condition groupings. Accuracy of the DaCCI and the DaCCI-SF was tested in the external validation cohort. Results: Within the development cohort, the most parsimonious and informative model resulted in the inclusion of 3 of the 17 (17.6%) original comorbid condition groupings: congestive heart failure, cerebrovascular disease, and chronic pulmonary disease. Within the validation cohort, the accuracy was 68.4% for the DaCCI versus 69.7% for the DaCCI-SF. Higher accuracy of the DaCCI-SF was confirmed in subgroup analyses performed according to age (≤75 vs >75 years), stage (organ-confined vs non–organ-confined), type of diversion (ileal-conduit vs non–ileal-conduit), and treatment period. Conclusions: DaCCI-SF relies on 17.6% of the original comorbid condition groupings and provides higher accuracy for predicting 90-day mortality after RC compared with the original DaCCI, especially in most contemporary patients.
Paolo Dell'Oglio, Zhe Tian, Sami-Ramzi Leyh-Bannurah, Vincent Trudeau, Alessandro Larcher, Marco Moschini, Ettore Di Trapani, Umberto Capitanio, Alberto Briganti, Francesco Montorsi, Fred Saad and Pierre I. Karakiewicz
Carlotta Palumbo, Francesco A. Mistretta, Sophie Knipper, Angela Pecoraro, Zhe Tian, Shahrokh F. Shariat, Fred Saad, Claudio Simeone, Alberto Briganti, Alessandro Antonelli and Pierre I. Karakiewicz
Background: Conditional survival (CS) may reveal important differences in cancer-specific mortality (CSM) among patients with nonmetastatic renal cell carcinoma (nmRCC). This study assessed CS according to T and N stages in patients treated surgically for nmRCC. Patients and Methods: Within the SEER database (2001–2015), all patients with nmRCC treated with either partial or radical nephrectomy were identified. CSM-free estimates according to T and N stage and substage groupings (pT1aN0–pT4N0 and pTanyN1) and multivariable Cox regression models with adjustment for Fuhrman grade and histologic subtype were assessed. Results: According to T and N stage and substage groupings, the following patients were included in the study: 35,966 (46.2%) with pT1aN0 disease; 18,858 (24.2%) with pT1bN0; 5,977 (7.7%) with pT2aN0; 2,511 (3.2%) with pT2bN0; 11,839 (15.2%) with pT3aN0; 1,037 (1.3%) with pT3b–cN0; 402 (0.5%) with pT4N0; and 1,302 (1.7%) with pTanyN1. Conditional CSM-free survival estimates were 98.2% at 1 year versus 98.0% at 10 years of event-free follow-up for patients with pT1aN0 disease, relative to baseline. Conversely, pT4N0/pTanyN1 conditional CSM-free survival estimates were 55.8% at 1 year versus 77.9% at 8 years of event-free follow-up. Attrition due to mortality was highest in patients with pT4N0/pTanyN1 disease. In multivariable Cox regression analyses, T stage, tumor grade, and histologic subtype represented independent predictors, but no interactions were identified. Conclusions: Tumor stage and its substages represent extremely important determinants of prognosis after lengthy event-free follow-up. The recorded observations have critical importance for physicians regarding patient follow-up and counseling.
Elio Mazzone, Sophie Knipper, Francesco A. Mistretta, Carlotta Palumbo, Zhe Tian, Andrea Gallina, Derya Tilki, Shahrokh F. Shariat, Francesco Montorsi, Fred Saad, Alberto Briganti and Pierre I. Karakiewicz
Background: Use of inpatient palliative care (IPC) in the treatment of advanced cancer represents a well-established guideline recommendation. A recent analysis showed that patients with genitourinary cancer benefit from IPC at the second lowest rate among 4 examined primary cancers, namely lung, breast, colorectal, and genitourinary. Based on this observation, temporal trends and predictors of IPC use were examined in patients with metastatic urothelial carcinoma of the bladder (mUCB) receiving critical care therapies (CCTs). Patients and Methods: Patients with mUCB receiving CCTs were identified within the Nationwide Inpatient Sample database (2004–2015). IPC use rates were evaluated in estimated annual percentage change (EAPC) analyses. Multivariable logistic regression models with adjustment for clustering at the hospital level were used. Results: Of 1,944 patients with mUCB receiving CCTs, 191 (9.8%) received IPC. From 2004 through 2015, IPC use increased from 0.7% to 25.0%, respectively (EAPC, +23.9%; P<.001). In analyses stratified according to regions, the highest increase in IPC use was recorded in the Northeast (EAPC, +44.0%), followed by the West (EAPC, +26.8%), South (EAPC, +22.9%), and Midwest (EAPC, +15.5%). Moreover, the lowest rate of IPC adoption in 2015 was recorded in the Midwest (14.3%). In multivariable logistic regression models, teaching status (odds ratio [OR], 1.97; P<.001), more recent diagnosis (2010–2015; OR, 3.89; P<.001), and presence of liver metastases (OR, 1.77; P=.02) were associated with higher IPC rates. Conversely, Hispanic race (OR, 0.42; P=.03) and being hospitalized in the Northeast (OR, 0.36; P=.01) were associated with lower rate of IPC adoption. Finally, patients with a primary admission diagnosis that consisted of infection (OR, 2.05; P=.002), cardiovascular disorders (OR, 2.10; P=.03), or pulmonary disorders (OR, 2.81; P=.005) were more likely to receive IPC. Conclusions: The rate of IPC use in patients with mUCB receiving CCTs sharply increased between 2004 and 2015. The presence of liver metastases, infections, or cardiopulmonary disorders as admission diagnoses represented independent predictors of higher IPC use. Conversely, Hispanic race, nonteaching hospital status, and hospitalization in the Midwest were identified as independent predictors of lower IPC use and represent targets for efforts to improve IPC delivery in patients with mUCB receiving CCT.