Comparison of pyogenic liver abscesses caused by hypermucoviscous Klebsiella pneumoniae and non- Klebsiella pneumoniae pathogens in Beijing: A retrospective analysis

Abstract

Objective

To perform a retrospective comparison of the clinical and radiological features of Klebsiella pneumoniae (KP)-associated and non-KP-associated pyogenic liver abscesses (PLA) in Chinese patients.

Methods

Patients with confirmed diagnoses of bacterial liver abscess at three Beijing hospitals were enrolled. Clinical isolates from liver abscesses were used to determine serology and expression of hypermucoviscosity genes. Basic clinical, ultrasonographic (US) and computed tomography (CT) data were recorded and compared between patients with KP- and non-KP-associated PLA.

Results

A total of 101 (77.10%) and 30 (22.90%) cases were due to KP and non-KP pathogens, respectively. Compared with the non-KP cohort, the KP cohort demonstrated a significantly higher incidence of underlying diabetes mellitus, and more gas-forming abscesses, as demonstrated by US and CT examinations. Prior abdominal surgery or chemoradiation therapy was significantly associated with non-KP cases. The non-KP group had a higher chance of a clear edge, compared with the KP group, on pre-contrast CT images.

Conclusion

KP and non-KP-associated PLA have distinctive risk factors and unique US and CT features, in Chinese patients.

Keywords

Liver abscess hypermucoviscosity

Introduction

Liver abscesses can be caused by intra-abdominal infection due to bacterial, fungal or parasitic organisms.¹ It was initially believed that Escherichia coli was the single most important pathogen in the aetiology of pyogenic liver abscess (PLA).¹ Klebsiella pneumoniae is an enteric Gram-negative nosocomial and/or community-acquired human pathogen, commonly associated with pneumonia or urinary tract infections.² In the last two decades, it has also emerged as the major pathogen causing PLA. K. pneumoniae-mediated PLA was first detected in Taiwan^3–5 and has now assumed the status of an emerging global disease, with anecdotal accounts and case series reported worldwide.⁶

A hypermucoviscous phenotype is associated with serotypes K1 and K2 K. pneumoniae strains, and this is largely responsible for causing invasive liver abscesses.^2,7 Several virulence factors have been described for K. pneumoniae including presence of: the capsular serotype; mucoviscosity-associated gene A (magA); regulator of mucoid phenotype gene A (rmpA); aerobactin.^8–10 Approximately 50% of reported patients with PLA have diabetes mellitus (type 1 or type 2: the type is not relevant to the development of PLA), while the remainder of patients do not appear to exhibit any other underlying disease.¹¹ Worldwide, 43 (5.28%) of the 813 invasive PLA cases reported in the past decade have died.² The high mortality rate can be attributed to the high antibiotic resistance profile of K. pneumoniae, which is further compounded by its unique capacity to cause invasive and metastatic complications.^12–14

It is imperative to understand the differences observed between patients with K. pneumoniae- or non-K. pneumoniae-associated PLA. Such an understanding may help to guide the development of comprehensive recommendations regarding the early diagnosis and treatment of K. pneumoniae-associated PLA. The present study, therefore, investigated the distinctive clinical and radiological features found among patients with K. pneumoniae- or non-K. pneumoniae-associated PLA.

Patients and methods

Patient population

Consecutive patients with bacterial liver abscesses, admitted to Beijing Shijitan Hospital, Capital Medical University, Chinese PLA General Hospital, and Peking University First Hospital (all in Beijing, China) between January 2008 and December 2011 were recruited into this study. Patients were included if they had: (i) clinical symptoms of liver abscess including fever, chills and liver pain; (ii) one or more area(s) of echolucency in the liver on ultrasonography (US); (iii) undergone computed tomography (CT)-guided or US-guided percutaneous liver abscess pus puncture therapy; (iv) a culture of liver aspirates positive for bacteria (single pathogenic strain).

The study was approved by the Institutional Review Board at the Chinese PLA General Hospital and patients provided written informed consent.

Outcome measures

Basic clinical data including patient information, evidence of underlying disease, clinical manifestations, laboratory tests, and US and CT imaging characteristics were obtained from all enrolled patients by a blinded independent researcher (X-Y.X).

Culture and aetiology identification

Clinical isolates from liver aspirates (obtained from enrolled patients with a confirmed diagnosis of PLA) were identified in the clinical microbiology laboratories of each participating hospital, using the API® 20E test kit system (bioMérieux, Marcy I'Etoile, France).¹⁵ All cultures were collected using a culture collection and transport system (BBL™ CultureSwab™, Becton Dickinson and Co., Franklin Lakes, NJ, USA). Isolates were stored in brain–heart infusion broth at −80℃ for a mean of 2 years until use. Liver isolates were collected from the same patient only once during each episode of infection (nonduplicate isolates) and were serotyped using a countercurrent immunoelectrophoresis method, as previously described.¹⁶

String test

The hypermucoviscosity phenotype of K. pneumoniae can be identified by a positive string test.⁷ Briefly, MacConkey agar plates (Thermo Fisher Scientific, Rockford, IL, USA) were inoculated with the cultures and incubated at 35℃ without CO₂. The plates were examined the day after inoculation for the presence of hypermucoviscous, lactose-fermenting Gram-negative rods, using an automated bacterial identification system (VITEK®2, bioMérieux, Hazelwood, MI, USA). Identified isolates were subcultured on a 5% sheep blood agar plate, to ensure purity. Any isolate yielding a ≥ 90% probability of K. pneumoniae was further analysed using the string test: a bacterial loop was touched to a suspect colony on the agar plate and withdrawn slowly; bacteria forming a mucoid ‘string’ of ≥ 5 mm were considered to be positive.

Molecular analyses

DNA from isolates determined to be hypermucoviscous K. pneumoniae was extracted by a standard boiling lysis method and was frozen at −80℃ for subsequent molecular analyses. The primers used for amplification of target genes, serotype analysis and virulence genetic testing are presented in Table 1. PCR was undertaken in a final reaction volume of 25 µl, which contained 1 µl cDNA, 250 µM dNTPs, 0.1 µM of each primer, Taq polymerase buffer, and 2.5 U of LA Taq DNA polymerase (Takara,Shiga, Japan). The PCR reaction mixtures were kept at 95℃ for 5 min, followed by 40 cycles of 95℃ for 1 min, 50℃ for 1 min, 72℃ for 2 min, then 72℃ for 7 min using a xxxx thermal cycler (GeneAmp® PCR System 9700 (Applied Biosystems, Foster City, CA). Amplified products were run on a 1.2% agarose gel and visualized by ethidium bromide staining; glyceraldehyde 3-phosphate dehydrogenase was used as the internal control gene. The hypermucoviscosity serotypes of the clinical isolates were confirmed by two independent blind readings of the results (Y.Y. and K-F.W).

Table 1.

Sequences of primers used for amplification of target genes, serotype analysis and virulence genetic testing of hypermucoviscous isolates of Klebsiella pneumoniae, obtained from patients with pyogenic liver abscesses.

Target genes	Forward primer, 5′–3′	Reverse primer, 5′–3′
K1 (wzyKPK1)	GGTGCTCTTTACATCATTGC	GCAATGGCCATTTGCGTTAG
K2 (wzyKPK2)	GACCCGATATTCATACTTGACAGAG	CCTGAAGTAAAATCGTAAATAGATGGC
rmpA	ACTGGGCTACCTCTGCTTCA	CTTGCATGAGCCATCTTTCA
Aerobactin	GCATAGGCGGATACGAACAT	CACAGGGCAATTGCTTACCT
magA – outer	GGTGCTCTTTACATCATTGC	GCAATGGCCATTTGCGTTAG
magA – inner	CGCCGCAAATACGAGAAGTG	GCAATCGAAGTGAAGAGTGC

Statistical analyses

All statistical analyses were performed using SPSS® statistical software, version 16.0 (SPSS, Inc. Chicago, IL, USA) for Windows®. Data are expressed as mean ± SD. Student’s t-test was used to determine any significant differences in gene expression and outcome measures between patients with K. pneumoniae-associated PLA and those with non-K. pneumoniae-associated PLA. A P-value of < 0.05 was considered statistically significant.

Results

A total of 131 patients with a definitive diagnosis of PLA were included in the present study; clinical isolates from liver aspirates were positive for K. pneumoniae in 101 (77.10%) cases and negative for K. pneumoniae in 30 (22.90%) cases. Escherichia coli was the next most-common pathogen (accounting for 20 [15.27%] of cases), followed by Pseudomonas aeruginosa (3 [2.29%] cases), Enterococcus (2 [1.53%] cases), Citrobacter freundii (2 [1.53%] cases), Enterobacter cloacae (1 [0.76%] case), Staphylococcus aureus (1 [0.76%] case) and coagulase-negative Staphylococci (1 [0.76%] case).

Of the 101 K. pneumoniae cases, K1 and K2 serotypes were found in 43 (42.57%) and 37 (36.63%) cases, respectively; the remaining 21 (20.79%) cases were due to non-K1/K2 serotypes. A hypermucoviscosity phenotype was determined in 90 (89.11%) cases (Table 2), a finding that was confirmed by two independent blind readings of the 101 clinical isolates, yielding a consistency of 99%. A high proportion of the K. pneumoniae-positive isolates expressed rmpA and aerobactin, but not magA (Table 2).

Table 2.

Virulence gene expression profile associated with hypermucoviscosity (HMV) in 101 cases of Klebsiella pneumoniae, detected over 2 consecutive years (during the period between 2008 and 2011) in patients with pyogenic liver abscesses.

Profile	2008–2009 n = 50	2010–2011 n = 51	Total n = 101
Any HMV gene	44 (88.00)	46 (90.20)	90 (89.11)
magA	12 (24.00)	15 (29.41)	27 (26.73)
rmpA	42 (84.00)	40 (78.43)	82 (81.19)
Aerobactin	43 (86.00)	48 (94.12)	91 (90.10)

Data presented as n (%) of patients.

magA, mucoviscosity-associated gene A; rmpA, regulator of mucoid phenotype gene A.

There were significantly more male patients with K. pneumoniae-associated PLA compared with the non-K. pneumoniae cohort (P < 0.001; Table 3). Fatigue and abdominal pain were significantly more common in patients with non-K. pneumoniae-associated PLA compared with patients with K. pneumoniae-associated PLA (P < 0.001; Table 3). There were no significant differences in any of the physical examination findings between the two study groups (Table 3).

Table 3.

Clinical characteristics and laboratory test results in 131 patients with Klebsiella pneumoniae- or non-K. pneumoniae-associated pyogenic liver abscesses (PLA).

Characteristic/test result	K. pneumoniae- associated PLA n = 101	Non-K. pneumoniae- associated PLA n = 30	Statistical significance^a
Age, years	53.93 ± 1.45	56.33 ± 2.25	NS
Range	16–79	18–81
Sex
Male	77 (76.24)	6 (20.00)	P < 0.001
Female	24 (23.76)	24 (80.00)	P < 0.001
Maximum body temperature, ℃	39.12 ± 0.08	39.01 ± 0.11	NS
Clinical manifestations
Chills	65 (64.36)	21 (70.00)	NS
Fatigue	17 (16.83)	13 (43.33)	P < 0.001
Abdominal pain	41 (40.59)	19 (63.33)	P < 0.001
Anorexia	44 (43.56)	16 (53.33)	NS
Vomiting	17 (16.83)	5 (16.67)	NS
Physical examination
Liver area percussion pain	33 (32.67)	13 (43.33)	NS
Liver tenderness	27 (26.73)	10 (33.33)	NS
Hepatomegaly	7 (6.93)	3 (10.00)	NS
WBC, × 10⁹/l	12.67 ± 1.21	12.91 ± 1.11	NS
Neutrophil ratio, %	78.31 ± 2.23	76.75 ± 1.96	NS
Haemoglobin, g/l	123.78 ± 4.93	117.82 ± 3.85	NS
Albumin, g/l	37.72 ± 1.21	36.73 ± 2.36	NS
ALT, U/l	78.91 ± 5.38	69.35 ± 7.12	NS
TB, µmol/l	19.69 ± 4.41	20.14 ± 3.75	NS
FBG, mmol/l	8.32 ± 0.87	5.48 ± 0.53	P < 0.001

Data presented as mean ± SD or n (%) of patients.

Student’s t-test.

WBC, white blood cell count; ALT, alanine transaminase; TB, total bilirubin; FBG, fasting blood glucose; NS, not statistically significant (P ≥ 0.05).

Fasting blood glucose levels were significantly higher among the K. pneumoniae cohort compared with the non-K. pneumoniae cohort (P < 0.001; Table 3) and diabetes mellitus (type 1 or 2) was significantly more common in K. pneumoniae cases compared with non-K. pneumoniae cases (P < 0.001; Table 4). The incidence of comorbid diabetes plus hepatobiliary disease was not significantly different between the two study cohorts. Prior history of abdominal surgery or chemoradiation therapy was significantly more common in patients with non-K. pneumoniae-associated PLA compared with K. pneumoniae-associated cases (P < 0.001 and P < 0.05, respectively; Table 4). There was no statistically significant difference in the rates of comorbid biliary tract disease, cirrhosis or fatty liver complications between the two groups (Table 4).

Table 4.

Underlying diseases in 131 patients with Klebsiella pneumoniae- or non-K. pneumoniae-associated pyogenic liver abscesses (PLA).

Concomitant disease	K. pneumoniae- associated PLA n = 101	Non-K. pneumoniae- associated PLA n = 30	Statistical significance^a
Diabetes mellitus (type 1 or 2)	59 (58.42)	7 (23.33)	P < 0.001
Biliary tract disease	26 (25.74)	9 (30.00)	NS
History of abdominal surgery	17 (16.83)	13 (43.33)	P < 0.001
Cirrhosis	12 (11.88)	5 (16.67)	NS
Chemoradiotherapy	5 (4.95)	4 (13.33)	P < 0.05
Fatty liver complications	3 (2.97)	1 (3.33)	NS
Diabetes and biliary disease	10 (9.90)	1 (3.33)	NS

Data presented as n (%) of patients.

Student’s t-test.

NS, not statistically significant (P ≥ 0.05).

All enrolled patients had a history of B-mode US inspection. US examination revealed single abscesses in 84/101 (83.17%) and 26/30 (86.67%) cases of K. pneumoniae-associated and non-K. pneumoniae-associated PLA, respectively (Table 5). Single abscesses were found to be more prominently localized to the adipose tissue in the right lobe of the liver, in both study groups. Consistent with the observed association between diabetes and K. pneumoniae-associated PLA, US imaging indicated significantly more gas-forming abscesses among K. pneumoniae cases compared with non-K. pneumoniae cases (P < 0.001; Table 5).

Table 5.

B-mode ultrasonography (US) and computed tomography (CT) scanning examination results in patients with Klebsiella pneumoniae- or non-K. pneumoniae-associated pyogenic liver abscesses (PLA).

Examination/findings	K. pneumoniae- associated PLA^a	Non-K. pneumoniae- associated PLA^a	Statistical significance^b
US
Abscesses, n
1	84 (83.17)	26 (86.67)	NS
> 1	17 (16.83)	4 (13.33)	NS
Single abscess area
Left lobe	13 (12.87)	4 (13.33)	NS
Right lobe	68 (67.33)	20 (66.67)	NS
Junction of right and left lobes	3 (2.97)	2 (6.67)	NS
Abscess diameter, mm	63.43 ± 3.74	72.95 ± 2.67	NS
Observed abscesses
Gas	35 (34.65)	4 (13.33)	P < 0.001
Separated	26 (25.74)	7 (23.33)	NS
CT
Pre-contrast abscess edge
Clear	20 (25.00)	12 (66.67)	P < 0.001
Unclear	60 (75.00)	6 (33.33)	P < 0.001
Gas seen in abscess	31 (38.75)	4 (22.22)	P < 0.05
Pneumatosis in biliary system	5 (6.25)	1 (5.56)	NS
Pleural effusion	9 (11.25)	3 (16.67)	NS

Data presented as mean ± SD or n (%) of patients.

US examination performed in all enrolled patients (K. pneumoniae-associated PLA, n = 101; non-K. pneumoniae-associated PLA, n = 30); CT examination performed in 98/131 patients (K. pneumoniae-associated PLA, n = 80; non-K. pneumoniae-associated PLA, n = 18).

Student’s t-test.

NS, not statistically significant (P ≥ 0.05).

Computed tomography examination was performed in 98/131 patients (80 K. pneumoniae cases and 18 non-K. pneumoniae cases). Pre-contrast CT images demonstrated that there was a greater chance of a clear edge in the non-K. pneumoniae group compared with the K. pneumoniae group (P < 0.001; Table 5). Presence of gas in the abscess was significantly higher in the K. pneumoniae group (P < 0.05; Table 5). Rates of pneumatosis in the biliary system or pleural effusion were not significantly different between the two study groups.

Discussion

Klebsiella pneumoniae-mediated community-acquired PLA is an emerging infectious disease worldwide.^3–6 Results from the present study corroborate this observation, demonstrating that the main bacterial liver abscess pathogen was K. pneumoniae in 77.10% of the 131 cases examined. Furthermore, it was shown that ∼80% of the K. pneumoniae serotypes were K1 and K2, and that 89.11% of the strains exhibited the hypermucoviscosity phenotype, which is consistent with other published studies.²

Molecular analyses undertaken in the present study indicated that expression of rmpA and aerobactin was high in K. pneumoniae-positive isolates, whereas magA expression was low. These data suggest that rmpA and aerobactin expression may be closely related to K. pneumoniae-mediated invasion of the liver and the subsequent development of liver abscesses. Siderophores, including enterochelin and aerobactin, are intracellular iron chelators that allow bacteria to compete for iron sources in the host.^8–10 In combination with rmpA, aerobactin plays an important part in the virulence and hypermucoviscosity phenotype of K. pneumoniae isolates other than those of serotype K1 and K2.² In fact, it has been shown that the nonserotype K1 and K2 isolates (21 cases in the present investigation) that express aerobactin genes display comparative virulence to serotype K1 and K2 isolates from patients with liver abscesses.¹⁷ As such, the increased virulence and invasiveness of the K. pneumoniae isolates in the present study may be explained by the high levels of rmpA and aerobactin expression that were found in the majority of these isolates. magA has been described as the causative gene for K. pneumoniae liver abscess and septic metastatic complications.^18–21 K1 serotypes were detected more often than K2 serotypes in the present study, however, which may be indicative of the involvement of other genes in the regulation of K1. rmpA, which aids capsule synthesis, has also been proposed as a virulence factor, in addition to magA and capsular serotypes K1/K2.^17,22,23

A high blood-glucose level is conducive to bacterial growth and inhibition of phagocyte chemotaxis, phagocytosis and bactericidal function.^24,25 In addition, the negative impact of diabetes mellitus on haemodynamics hinders immune-cell mobilization. Half of reported patients with PLA also have diabetes mellitus.^9,11,12 Data from the present study are consistent with these findings: patients with K. pneumoniae-associated PLA had higher fasting blood-glucose levels and this was corroborated by the finding that diabetes was significantly more common in these patients compared with non-K. pneumoniae cases. These results suggest that diabetes mellitus may be one of the most important risk factors for K. pneumoniae-associated PLA.

Interestingly, the present study also established that prior history of abdominal surgery and chemoradiation therapy are the most important risk factors for non-K. pneumoniae-associated PLA. These data suggest that non-K. pneumoniae-associated PLA largely occurs as a nosocomial infection. In accordance with previous findings, the present study showed that prior hepatobiliary disease did not seem to be a risk factor in either study group.²

Research has indicated that PLA has no sex predilection but most commonly involves patients ≥ 65 years of age,^26,27 although the mean ages of the patients in our study were younger than this. In the present study, the K. pneumoniae cohort consisted of significantly higher numbers of male patients than females, compared with the non-K. pneumoniae cohort. It is of note, however, that there were a higher number of cases in general, in the K. pneumoniae cohort compared with the non-K. pneumoniae cohort, which may have biased the observed difference in the male : female ratio.

Others have reported that solitary abscesses were more frequently found in patients with K. pneumoniae-associated PLA compared with non-K. pneumoniae-associated PLA.^28,29 The present study did not, however, reveal any between-group differences in this regard. Our findings replicate previous observations that the size of abscess was similar in both groups, that most cases had single abscesses and that abscesses were often situated in the right lobe of the liver.^30,31 Other studies have also reported that gas is not commonly present in CT images obtained from patients with K. pneumoniae-associated PLA^32–34 and that there was no significant difference in the presence of gas between K. pneumoniae and non-K. pneumoniae cases.¹ We observed significantly more gas in liver abscesses from patients with K. pneumoniae compared with non-K. pneumoniae, using either US or CT imaging techniques. In addition, pre-contrast CT imaging showed that the non-K. pneumoniae group had a higher chance of a clear abscess edge than the K. pneumoniae group, in the present study.

Bacterial culture is affected by objective conditions, such as time and environment. The positive culture rate may also decrease significantly when using empirical antibiotics, which in turn can delay diagnosis and subsequent treatment. Incorporation and interpretation of improved clinical and imaging data, combined with detection at the molecular level, is required to make fast and accurate diagnosis of pathogens (particularly in acute-onset cases or severe infections). The present study showed that K. pneumoniae and non-K. pneumoniae-associated PLA have distinctive risk factors and unique US and CT features. Taken together, these data suggest that rapid and accurate diagnosis of K pneumoniae infection in invasive liver abscesses could be achieved by analysis of the presenting symptoms, culture reports and radiological investigations, which may ultimately prevent severe metastatic complications. Further elucidation of the genomic basis of the higher risk of K. pneumoniae infection and the pathological mechanism of PLA in an Asian population are warranted.

Footnotes

Declaration of conflicting interest

The authors declare that there are no conflicts of interest.

Funding

This study was supported by the Capital Health Research and Development of Special Capital fund (Grant No. 2011-2008-07).

Acknowledgements

The authors wish to thank everyone who helped with this study.

References

Kim

Lee

. Computed tomographic differences of pyogenic liver abscesses caused by Klebsiella pneumoniae and non-Klebsiella pneumoniae. J Comput Assist Tomogr 2007; 31: 59–65.

Siu

Yeh

Lin

. Klebsiella pneumoniae liver abscess: a new invasive syndrome. Lancet Infect Dis 2012; 12: 881–887.

Chang

Chou

Fan

. A clinical study of Klebsiella liver abscess. Taiwan Yi Xue Hui Za Zhi 1988; 87: 282–287.

Chang

Chou

. Comparison of pyogenic liver abscesses caused by Klebsiella pneumoniae and non-K pneumoniae pathogens. J Formos Med Assoc 1995; 94: 232–237.

Liu

Cheng

Lin

. Klebsiella pneumoniae liver abscess associated with septic endophthalmitis. Arch Intern Med 1986; 146: 1913–1916.

Shon

Bajwa

Russo

. Hypervirulent (hypermucoviscous) Klebsiella pneumoniae: a new and dangerous breed. Virulence 2013; 4: 107–118.

Pinsky

Baron

Janda

. Bartholin's abscess caused by hypermucoviscous Klebsiella pneumoniae. J Med Microbiol 2009; 58: 671–673.

Podschun

Ullmann

. Klebsiella spp. as nosocomial pathogens: epidemiology, taxonomy, typing methods, and pathogenicity factors. Clin Microbiol Rev 1998; 11: 589–603.

Keller

Tsai

Lin

. Risk of infections subsequent to pyogenic liver abscess: a nationwide population-based study. Clin Microbiol Infect 2012. doi: 10.1111/1469-0691.12027. epub ahead of print.

10.

Lin

Hsieh

. Isolation of genes involved in biofilm formation of a Klebsiella pneumoniae strain causing pyogenic liver abscess. PLoS One 2011; 6: e23500–e23500.

11.

Fang

Chuang

Shun

. A novel virulence gene in Klebsiella pneumoniae strains causing primary liver abscess and septic metastatic complications. J Exp Med 2004; 199: 697–705.

12.

Fung

Lin

. Klebsiella pneumoniae in gastrointestinal tract and pyogenic liver abscess. Emerg Infect Dis 2012; 18: 1322–1325.

13.

Lederman

Crum

. Pyogenic liver abscess with a focus on Klebsiella pneumoniae as a primary pathogen: an emerging disease with unique clinical characteristics. Am J Gastroenterol 2005; 100: 322–331.

14.

Rahimian

Wilson

Oram

. Pyogenic liver abscess: recent trends in etiology and mortality. Clin Infect Dis 2004; 39: 1654–1659.

15.

Abbott

Klebsiella, Enterobacter, Citrobacter and Serratia. In: Murray

Baron

Pfaller

Tenover

Yolken

(eds). Manual of Clinical Microbiology, 7th ed. Washington D.C: American Society for Microbiology Press, 1999, pp. 475–482.

16.

Palfreyman

. Klebsiella serotyping by counter-current immunoelectrophoresis. J Hyg (Lond) 1978; 81: 219–225.

17.

Cheng

. Comparison of prevalence of virulence factors for Klebsiella pneumoniae liver abscesses between isolates with capsular K1/K2 and non-K1/K2 serotypes. Diagn Microbiol Infect Dis 2008; 62: 1–6.

18.

Fang

Chuang

Shun

. A novel virulence gene in Klebsiella pneumoniae strains causing primary liver abscess and septic metastatic complications. J Exp Med 2004; 199: 697–705.

19.

Fung

Chang

Lin

. Immune response and pathophysiological features of Klebsiella pneumoniae liver abscesses in an animal model. Lab Invest 2011; 91: 1029–1039.

20.

Hunt

Wang

Callegan

. Contribution of mucoviscosity-associated gene A (magA) to virulence in experimental Klebsiella pneumoniae endophthalmitis. Invest Ophthalmol Vis Sci 2011; 52: 6860–6866.

21.

Lin

Yang

. Amino acid substitutions of MagA in Klebsiella pneumoniae affect the biosynthesis of the capsular polysaccharide. PLoS One 2012; 7: e46783–e46783.

22.

Cheng

. Association between rmpA and magA genes and clinical syndromes caused by Klebsiella pneumoniae in Taiwan. Clin Infect Dis 2006; 42: 1351–1358.

23.

Yeh

Kurup

Siu

. Capsular serotype K1 or K2, rather than magA and rmpA, is a major virulence determinant for Klebsiella pneumoniae liver abscess in Singapore and Taiwan. J Clin Microbiol 2007; 45: 466–471.

24.

Lin

Wang

. Klebsiella pneumoniae liver abscess in diabetic patients: association of glycemic control with the clinical characteristics. BMC Infect Dis 2013; 13: 56–56.

25.

Hagiya

Kuroe

Nojima

. Emphysematous liver abscesses complicated by septic pulmonary emboli in patients with diabetes: two cases. Intern Med 2013; 52: 141–145.

26.

Eckburg

Montoya

Hepatobiliary infections. In: Wilson

Sande

(eds). Diagnosis and treatment in infectious diseases: Lange current series, New York, NY: McGraw-Hill, 2001, pp. 269–286.

27.

Law

. Older age as a poor prognostic sign in patients with pyogenic liver abscess. Int J Infect Dis 2013; 17: e177–e184.

28.

Yang

Yen

. Comparison of pyogenic liver abscess caused by non-Klebsiella pneumoniae and Klebsiella pneumoniae. J Microbiol Immunol Infect 2004; 37: 176–184.

29.

Hui

Yang

Cho

. Pyogenic liver abscesses caused by Klebsiella pneumoniae: US appearance and aspiration findings. Radiology 2007; 242: 769–776.

30.

Alvarez

González

Baldonedo

. Pyogenic liver abscesses: a comparison of older and younger patients. HPB (Oxford) 2001; 3: 201–206.

31.

Târcoveanu

Vlad

Moldovanu

. Pyogenic liver abscesses. Chirurgia (Bucur) 2008; 103: 417–427. in Romanian, English abstract).

32.

Lee

Kim

Lee

. CT differentiation of pyogenic liver abscesses caused by Klebsiella pneumoniae vs non-Klebsiella pneumoniae. Br J Radiol 2011; 84: 518–525.

33.

Lin

Chen

Lai

. Gas forming liver abscess caused by Klebsiella pneumoniae. BMJ Case Rep 2009; 2009. bcr08.2008.0638.

34.

Alsaif

Venkatesh

Chan

. CT appearance of pyogenic liver abscesses caused by Klebsiella pneumoniae. Radiology 2011; 260: 129–138.