We comment on the recent paper by Zhou et al. [ Appl. Opt. 45, 4433 (2006)], in which transmission losses of $0.2–0.3\mathrm{dB}/\mathrm{m}$ were claimed across the wavelength range $420–900\mathrm{nm}$ in a high-index ( $n_d=1.80518$ at $587.6\mathrm{nm}$ ) SF6 glass-based photonic crystal fiber fabricated by novel die-cast technique. If confirmed, these losses are at least 1 order of magnitude lower than previous reported losses of SF6 photonic crystal fibers from other fabrication approaches. Here we present a statistic survey on the relationship between the refractive index and the bulk material attenuation, based on a large number of commercial Schott optical glasses with the $n_d$ ranging between 1.40 and 2.05. It shows that the loss of a high-index ( $n_d=1.80$ ) glass optical fiber should be at the levels of $10–50\mathrm{dB}/\mathrm{m}$ at $420\mathrm{nm}$ and $1–10\mathrm{dB}/\mathrm{m}$ at $500\mathrm{nm}$ , respectively. Moreover, the material attenuation of such a high-index glass fiber should intrinsically show a large decay, from $10–50\mathrm{dB}/\mathrm{m}$ at $420\mathrm{nm}$ to the level of $1\mathrm{dB}/\mathrm{m}$ at $700\mathrm{nm}$ , which arises from the tail on the UV absorption edge of the high-index glass extending to the visible region. Therefore, we conclude that: (1) the low loss of $0.2–0.3\mathrm{dB}/\mathrm{m}$ reported in the cited paper is abnormally one or two magnitudes lower than the material attenuation that a high-index ( $n_d=1.80$ ) glass optical fiber should have in the range between 420 and $500\mathrm{nm}$ and that (2) the flat loss curve between 420 and $700\mathrm{nm}$ in the cited paper deviates greatly from the intrinsic behavior of a high-index ( $n_d=1.80$ ) glass fiber.