RESUMO
OBJECTIVE: To evaluate predictive ability of a novel combined index, Charlson comorbidity index and C-reactive protein (CCI-CRP), for outcomes in renal cell carcinoma (RCC), and compare predictive outcomes with of CCI-CRP to its separate components and to the UCLA integrated staging system (UISS). PATIENTS AND METHODS: We retrospectively analyzed INMARC registry of RCC patients. Receiver Operator Characteristics (ROC) analysis was fitted to identify threshold defining low-CRP (LCRP) and high-CRP (HCRP). Patients were stratified according to CCI [low-CCI ≤ 3 (LCCI); intermediate-CCI 4-6 (ICCI); high-CCI > 6 (HCCI)] and CRP level. Kaplan-Meier analysis (KMA) was conducted for overall (OS) and cancer-specific survival (CSS). Based on survival analysis distribution we proposed a new stratification: CCI-CRP. Model performance was assessed with ROC/area under the curve (AUC) analysis and compared to CCI and CRP alone, and UISS. RESULTS: We analyzed 2,890 patients (median follow-up 30 months). ROC identified maximum product sensitivity and specificity for CRP at 3.5 mg/L. KMA revealed 5-year OS of 95.6% for LCRP/LCCI, 83% LCRP/ICCI, 73.3% LCRP/HCCI, 62.6% HCRP/LCCI, 51.6% HCRP/ICCI and 40.5% HCRP/HCCI (P < .001). From this distribution, new CCI-CRP is proposed: low CCI-CRP (LCRP/LCCI and LCRP/ICCI), intermediate CCI-CRP (LCRP/HCCI and HCRP/LCCI), and high CCI-CRP (HCRP/ICCI and HCRP/HCCI). AUC for CCI-CRP showed improved performance for predicting OS/CSS vs. CCI alone (0.73 vs. 0.63/0.77 vs. 0.60), CRP alone (0.73 vs. 0.71/0.77 vs. 0.74) and UISS (0.73 vs 0.67/0.77 vs 0.73). CONCLUSIONS: CCI-CRP, exhibits increased prognostic performance for survival outcomes in RCC compared to CCI and CRP alone, and UISS. Further investigation is requisite.
RESUMO
BACKGROUND: To evaluate relationship between histological subtypes of renal cell carcinoma (RCC) and preoperative c-reactive protein (CRP). PATIENTS AND METHODS: We queried the International Marker Consortium for Renal Cancer database for patients affected by RCC. Patients were classified according to their histology: benign tumors, clear cell (cc) RCC, chromophobe (ch) RCC, papillary (p) RCC, and variant histology (vh) RCC; and according to CRP (mg/L): low CRP ≤5 and high CRP >5. Primary outcome was all-cause mortality (ACM). Secondary outcomes were cancer-specific mortality (CSM), recurrence and association between CRP and histology. Multivariable analysis (MVA) via Cox regression and multivariable logistic regression were fitted to elucidate predictors of outcomes. RESULTS: Total 3902 patients (high CRP n = 1266) were analyzed; median follow up 51 (IQR 20-91) months. On MVA elevated CRP was an independent risk factor associated with increased risk of ACM in benign tumors (HR 5.98, P < .001), ccRCC (HR 2.69, P < .001), chRCC (HR 3.99, P < .001), pRCC (HR 1.76, P = .009) and vhRCC (HR 2.97, P =.007). MVA for CSM showed CRP as risk factor in ccRCC (HR 2.77, P < .001), chRCC (HR 6.16, P = .003) and pRCC (HR 2.29, P = .011), while in vhRCC was not (P = .27). MVA for recurrence reported CRP as risk factor for ccRCC (HR 1.30, P = .013), while in chRCC (P = .33), pRCC (P = .34) and vhRCC (P = .52) was not. On multivariable logistic regression CRP was a predictor of pRCC (OR 1.003, P = .002), while decreasing CRP was associated with benign tumors (OR 0.994, P = .048). CONCLUSION: Elevated CRP was a robust predictor of worsened ACM in all renal cortical neoplasms. While most frequently observed in pRCC patients, elevated CRP was independently associated with worsened CSM in non-vhRCC. Conversely, elevated CRP was least likely to be noted in benign tumors, and elevation in this subgroup of patients should prompt further consideration for surveillance given increased risk of ACM. Further investigation is requisite.
