OBJECTIVE To study the laboratory predicted anticaries efficacy of five commercially available fluoride varnishes (FV) by determining their ability to reharden and to deliver fluoride to an early caries lesion when applied directly or in close vicinity to the lesion (halo effect). METHODS Early caries lesions were created in 80 polished bovine enamel specimens. Specimens were allocated to five FV groups (n = 16) based on Knoop surface microhardness (KHN) after lesion creation. All tested FV claimed to contain 5% sodium fluoride and were: CavityShield, Enamel Pro, MI Varnish, Prevident and Vanish. FV were applied (10 +/- 2 mg per lesion) to eight specimens per FV group (direct application); the remaining eight specimens received no FV but were later exposed to fluoride released from specimens which received a FV treatment (indirect application). Specimens were paired again and placed into containers (one per FV). Artificial saliva was added and containers placed into an incubator (27 hours at 37 degrees C). Subsequently, FV was carefully removed using chloroform. Specimens were exposed to fresh artificial saliva again (67 hours at 37 degrees C). KHN was measured and differences to baseline values calculated. Enamel fluoride uptake (EFU) was determined using the acid etch technique. Data were analyzed using two-way ANOVA. RESULTS The two-way ANOVA highlighted significant interactions between FV vs. application mode, for both deltaKHN and EFU (P < 0.001). All FV were able to reharden and deliver fluoride to caries lesions, but to different degrees. Furthermore, considerable differences were found for both variables between FV when applied either directly or in close vicinity to the lesion: MI Varnish and Enamel Pro exhibited greater fluoride efficacy when applied in vicinity rather than directly to the lesion, whereas CavityShield and Vanish did not differ. Prevident exhibited a higher EFU when applied directly, but little difference in rehardening.