Introduction Telomere restriction fragments (TRF) length is a well-known prognostic indicator in CLL. Long telomeres have been associated with VH-mutated status, while VH-unmutated patients are regarded as having uniformly short telomeres. However, little is known on cases in which discordance exists between these two parameters. Patients and methods 138 B-CLL patients were analyzed for TRF length and VH mutational status. All samples were taken before the start of anti-CLL treatment. M/F ratio was 93/45. Median age was 60 years. According to Binet staging, 82 patients were A, 30 B and 26 C. CD38 and ZAP70 expression and FISH for Ch 11, 12, 13, 17, were available in 72% of patients. Median follow-up was 33 months (range 6–290). TRF length was evaluated by Southern blot as previously described (Ladetto M et al, Blood 2004). VH mutational status was evaluated by direct sequencing according to standard methods. Results VH sequencing was successful in 127 patients: 82 were VH-mutated and 45 VH-unmutated. Median TRF length was 6097bp (1465–14837). As expected VH-unmutated patients had shorter telomeres (median 4000bp) compared to VH-mutated patients (median 7399bp) (p<0.0001). A mathematical model has been employed to analyze the distribution of TRF length in VH-mutated and unmutated patients. While mutated patients had a homogeneous distribution, unmutated patients showed a bimodal distribution with no patients falling between 4137 and 4710bp (fig1a). This allowed to discriminate a larger population (26 patients) with very short TRF length (concordant patients), and a smaller (19 pts) with a TRF length superimposable to that of VH-unmutated patients (p=0.21) (discordant patients). Concordant and discordant patients had similar patterns of VH usage and similar levels of homology (H) to the germline IgH sequence (i.e. H=100% vs H<100% and >99% vs H<99% and >98%). In addition they showed no difference for all the available (previously mentioned) biological and clinical parameters. However the two population had significant differences in terms of clinical outcome: 5-year PFS was 22% in concordant patients and 55% in discordant (p<0.05) (fig1b). Moreover, only 1 patient has already died among discordant patients as opposed to 9 among concordant (fig1c). Of note, the outcome of discordant patients was not significantly different from that of VH-mutated patients. The role of telomere length as survival predictor was further supported by performing univariate and multivariate analysis indicating that TRF length is the most powerful prognostic indicator among those considered (age, sex, Binet, CD38, ZAP70, FISH and mutational status). Conclusion VH-unmutated patients are highly heterogeneous and this heterogeneity can be revealed through evaluation of TRF length.