Background. Serum CRP levels are elevated in the majority of patients with HL at diagnosis; reflecting tumor burden and aggressive biologic behavior. Despite being an easily and frequently measured marker, data on its potential prognostic significance are extremely limited. Aim. To analyze the correlation between baseline CRP levels and clinical-labora¬tory findings and outcome of patients with HL treated with anthracy-cline-based chemotherapy with or without radiotherapy (RT). Patients and Methods. Baseline CRP levels were recorded in 496 patients with HL7 who were treated with anthracycline-based chemotherapy with or without radiotherapy (RT) in 2 Centers. Baseline CRP levels were cor¬related with other baseline clinical-laboratory features (Spearman corre¬lation coefficient or Mann-Whitney test), Failure Free Survival (FFS) and Overall Survival (OS) (Kaplan-Meier curves, log-rank comparison). Mul-tivariate FFS analysis was based on Cox's proportional hazards model. CRP levels s>5 mg/L were considered elevated. Results. The median value of CRP levels in the 496 evaluable patients was 21.10 mg/L; 27% and 73% had normal and elevated CRP levels respectively. Baseline CRP levels correlated with virtually all other parameters reflecting tumor burden or disease aggressiveness: There were strong correlations with advanced stages and B-symptoms (p<0.001) as well as specific extran-odal sites (bone marrow, lung, liver). Baseline CRP presented a very strong correlation with ESR (Spearman's rho 0.74, p<0.001), strong cor¬relations with haemoglobin (inverse), platelet counts and albumin (in¬verse) (S-rho 0.40-0.60, p<0.001) and looser correlations with white blood cell counts, serum LDH; and absolute lymphocyte counts (in¬verse) (S-rho <0.35, p<0.001). A trend towards a dose-response effect with FFS was observed: 5-year FFS for patients with CRP levels <5, 5-21.09, 21.1-69.99 and s=70 mg/L (roughly the CRP quartiles) was 81%, 78%, 68% and 66% (p=0.02). The difference was more pronounced for patients with normal versus elevated CRP levels (81% versus 71%, p=0.006). Differences in OS were not significant (p>0.10). The prog¬nostic impact of CRP on FFS was borderline in early stage patients (IA,IIA: 5-year FFS 83% versus 75% for normal versus elevated levels, p=0.09), but no difference was detected in advanced disease (68% versus 66%, p=0.48). CRP was an independent adverse prognostic factor for FFS, when adjusted for stage (III/IV versus I/II), B-symptoms and IPS factors (except of lymphocytopenia). However, baseline CRP was not analyzed along with ESR in the same model, because of their very strong correlation, raising issues of collinearity. Conclusions. CRP levels are elevated in -70% of HL patients at diagnosis and correlated with FFS independently from other established prognostic factors in¬cluded in the IPS. The very strong correlation between baseline CRP and ESR probably suggests that only one of them should be included in a given prognostic model. Even larger patient series should be analyzed in order to draw definite conclusions, but the final selection might be re¬lied on specific biologic and technical advantages of either marker.