Virulence of Klebsiella pneumoniae Serotype K2 Should Not Be Underestimated in K. pneumoniae Liver Abscess

To the Editor—Fang et al.[1] have observed that serotype K2 Klebsiella pneumoniae isolates are not a significant risk factor for septic ocular or CNS complications. According to our previous study [2] and studies from outside Taiwan [3, 4], serotype K2 is the second largest serotype (after K1) to cause pyogenic liver abscess (PLA) and complications. Actually, the study by Fang et al. [1] made the same observation. One of the problems that has always been confusing to us is the number of cases in the series of studies by Fang et al. [1, 5] and Chuang et al. [6], which we tried to understand previously [7]. In their series of studies in the same hospital, 53 cases of PLA occurred during 1996–2001 [5] in which only 1 case was not caused by K1, then 42 PLA cases occurred consecutively in 1997–2003, of which 35 cases were caused by serotype K1 [6]. Although Fang and colleagues have explained that the errors occurred in sampling between 2 studies, the latter study was more accurately collected [7]. In the recent study [1], strains were collected in 1997–2005, and numbers of cases were increasing remarkably in 2004–2005, with 135 cases of PLA within 2 years, ∼11.25 times the average increase per year after 2003. We believe that the increasing number of PLA cases was possibly due to selection criteria for PLA. Previous studies of K. pneumoniae liver abscess focused on the cryptogenic PLA, whereas there were 3 different categories of PLA in the study by Fang et al. [1]. We actually do not understand how to combine the so-called secondary PLA and undefined (noncryptogenic and nonsecondary) PLA to compare with other studies. According to Fang et al. [1, 5] and Chuang et al. [6], K2 isolates first occurred after 2002, and the proportion then increased dramatically, to 20% in total, for PLA in 1997–2005. Thus, there was a sharp increase in K2 PLA. In fact, K2 should be considered a significant factor that we should not separate from K1 liver abscess. Whether the cases with complications in the study of Fang et al. [1] were all from cryptogenic PLA and without cases from secondary or undefined PLA, another question will be raised as to whether the cases of complication in the study of were totally from cryptogenic PLA. With the assumption that the complications in the study by Fang et al. [1] were all from cryptogenic PLA, 43 (81%) of 53 and 6 (11%) of 53 isolates had K1 and K2 serotype, respectively. Among them, 19 and 2 of the K1 and K2 isolates, respectively, had endophthalmitis and meningitis. Thus, 19 (43%) of 43 and 2 (33%) of 6 patients with K1 and K2 liver abscess, respectively, had further complications. Therefore, K2 strains play an important role and not a minor or nonsignificant role. The authors should clarify the origin of those complicated cases. If not, the analysis will become inconclusive and uncertain. However, for either of the situations above, K2 is still an important factor for complications or liver abscess. Finally, if serotyping for the K antigen is used only to differentiate K1 from non-K1 K. pneumoniae, detection could be done within 10 min by a simple antiserum Quellung test. PCR detection should not be the rapid test used to differentiate serotype K1 and non-K1 K. pneumoniae. Antiserum could also be purchased easily, as described in the previous study by Chuang et al. [6]. Potential conflicts of interest. All authors: no conflicts.