The Distant Metastasis of Pyogenic Liver Abscess Caused by Klebsiella pneumoniae Serotype K2 and the Underlying Disease of Diabetes Mellitus Should Be Carefully Interpreted

Tthe Editor—A recent article by Fang et al. [1] mentioned that diabetes mellitus (DM) is not a significant risk factor for septic ocular or CNS complications in Klebsiella pneumoniae liver abscess (KP-LA). In the study by Fang and colleagues, 13 (57%) of 23 patients with distant metastasis to eye or CNS were affected with DM. Previous studies showed that 67%–95.5% of the KP-LA-affected patients with distant metastasis were affected with DM [2–5]. In 2 studies, with overall enrollment of 354 patients with KP-LA, DM was verified as a significant factor for developing severe extrahepatic complications [4, 5] and poor visual outcome of endophthalmitis [5]. We suggest that subjects with DM should not be ignored in the future study of KP-LA and that the definition of severe metastatic complications of KP-LA should not be confined to eye and CNS only. Although the introduction in the article by Fang et al. [1] provided references to support the low prevalence of DM among subjects with liver abscess, all references have at least 64% of patients with KP-LA had DM (see the references [6–9], which Fang et al. [1] cited). The only reference from Korea [10], with ∼40% of patients affected with DM, actually mentioned that DM was at highest prevalence among all underlying diseases. Taken together, the work of Fang and colleagues seems to completely misrepresent the significance of DM in KP-LA. Fang et al. [1] disagree with our conclusions that K1/K2 is a major virulence factor, conclusions that are based on our findings from an in vitro neutrophil phagocytosis model and an animal lethality study [11, 12]; we used isolates from 2 geographical regions. Instead, Fang and colleagues concluded that K1 only (and not K2) was the major virulence factor, conclusions that were based on their results from a serum-resistance assay of K1 and K2 isolates from their single hospital [1]. Might the results of their in vitro assay of serum resistance be less pertinent than the findings of virulence from the animal model? In the study by Mizuta et al. [13], K1 strains had lower virulence than did K2 strains. Although 2 strains of K2 isolates from the study by Mizuta et al. [13] were avirulent, 7 of 9 K2 isolates did cause lethality with <10 colony-forming units. It is also interesting that 1 patient infected with a serum-sensitive K1 strain in the study by Fang et al. [1] developed endophthalmitis that led to loss of vision. Although we highlight the fact that serotypes K1 and K2 play an important role in KP-LA, as well as the extrahepatic complications [2, 11], the significance of interactions between the virulence of K. pneumoniae and DM should not be ignored [12]. We should not exclude the associated factors other than serotypes that were involved in the KP-LA, because non-K1/K2 KP-LA still has an unknown etiology that magA or serotypes cannot explain. Potential conflicts of interest. All authors: no conflicts.