In our original paper (Section 4.4) we compared the ratios of hard X-ray to nuclear mid-infrared emission for Seyfert 1 and 2 (Sy1 and Sy2) nuclei, finding that both distributions appear significatively different according to the Kolmogorov-Smirnov test. We have detected a mistake in the computation of the X-ray luminosities (the lack of a 4π factor). Additionally, for two galaxies, the nuclear mid-infrared fluxes used within calculations are slightly different from the correct values which are reported in Table 3 of the original version. These mistakes affect the mentioned ratios of hard X-ray to nuclear mid-infrared emission, although they do not alter the result concerning the distributions, being statistically different. The correct values of these ratios are log(L X intr/L MIR) = –0.21 ± 0.33 for the Seyfert 1 galaxies, and log(L X intr/L MIR) = 0.17 ± 0.62 for the type 2 nuclei. Figure 8 has been updated with the newly computed values. We have also corrected the X-axis with respect to the original figure, which is now ν L ν. We have performed again Spearman's rank correlation test, finding that the correlation is significant when all galaxies and types 1 and 2 are considered separately. It is worthwhile to note that here we have excluded from the analysis two Seyfert 2 galaxies (NGC 1386 and NGC 7674) due to their extremely low hard X-ray luminosities, consistent with being optically thick sources. Recently, Horst et al. (2008) have estimated an intrinsic X-ray luminosity for NGC 7674 of log L X intr = 44.56 erg s–1, in contrast to the non-absorption-corrected value reported by Lutz et al. (2004) (log LX obs = 41.91 erg s–1). Such a difference of about 2.5 dex corroborates our decision of not including these galaxies. In a recent work, Horst et al. (2008) have argued that our luminosity ratios are ~8 larger than what they have found for their well-resolved objects, claiming that our nuclear data are heavily contaminated by nuclear star formation. Our new values are now of the same order of magnitude than theirs, even smaller for the case of Seyfert 2 galaxies. Therefore, we believe that despite the limited resolution of ISOCAM images, our nuclear fluxes are representative of the torus emission, and the contamination by circumnuclear star formation, although present, is not dominant. There is a typographical mistake in Table 3 for the point-spread function (PSF) flux in the LW7 filter corresponding to the galaxy ESO144-G55; the correct value is 84.2 ± 2.44 mJy.