It remains unclear how the detection of hepatitis B core antibody (anti-HBc) in the absence of hepatitis B surface antigen (HBsAg) and antibody (anti-HBs) should be interpreted and whether all patients with this pattern need to be tested for hepatitis B virus (HBV)-DNA. This study aimed at reassessing the significance of 'anti-HBc alone' in unselected sera referred to the clinical laboratory and determining whether significant HBV viraemia can be found in this setting. Of the 6431 patients tested for HBsAg, total anti-HBc and anti-HBs in a Paris hospital over a 1-year period, 362 (5.6%) had 'anti-HBc alone' (24.8% of anti-HBc-positive patients). Only 11 of the 362 sera (3.0%) were found to be false positive. One patient was in the resolving phase of acute hepatitis B. HBV-DNA was detected in 10 of 362 (2.8%) patients, using a commercial standardized assay (threshold: 350 IU/mL). Viral loads exceeded 10(4) copies/mL in 6 of 10 patients. Mutations in the HBsAg immunodominant region were identified in seven of the viraemic patients. HBsAg was detected in only two cases when retested by one of the latest, multivalent assays. Neither human immunodeficiency virus nor hepatitis C virus serostatus distinguished between patients with and without HBV-DNA. In conclusion, 'anti-HBc alone' should be considered a risk marker for a so-called 'false occult' HBV infection with significant viraemia. Indeed, results in this hospital population indicate that a small proportion of patients with 'anti-HBc alone' have high viral loads, revealing the occurrence of infection with HBV mutants that escape detection even by multivalent HBsAg assays.