In a previous paper [1], we presented a three-flavour oscillation analysis of the solar neutrino measurements and of the first data from the KamLAND experiment, in terms of the relevant mass-mixing parameters (delta m^2, theta_12, theta_13). The analysis, performed by including the terrestrial neutrino constraints coming from the CHOOZ (reactor), KEK-to-Kamioka (K2K, accelerator) and Super-Kamiokande (SK, atmospheric) experiments, provided a stringent upper limit on theta_13, namely, sin^2(theta_13)<0.05 at 3 sigma. We reexamine such upper bound in the light of a recent (although preliminary) reanalysis of atmospheric neutrino data performed by the SK collaboration, which seems to shift the preferred value of the largest neutrino square mass difference Delta m^2 downwards. By taking the results of the SK official reanalysis at face value, and by repeating the analysis in [1] with such a new input, we find that the upper bound on theta_{13} is somewhat relaxed: sin^2(theta_13)<0.067 at 3 sigma. Related phenomenological issues are briefly discussed.