World J Surg DOI 10.1007/s00268-015-3126-1
ORIGINAL SCIENTIFIC REPORT
Pyogenic Liver Abscess: Does Escherichia Coli Cause more Adverse Outcomes than Klebsiella Pneumoniae? Vishal G. Shelat1 • Clement L. K. Chia1,3 • Charleen S. W. Yeo2 • Wang Qiao3 Winston Woon1 • Sameer P. Junnarkar1
•
Ó Socie´te´ Internationale de Chirurgie 2015
Abstract Background This paper aims to demonstrate if Escherichia coli pyogenic liver abscess (ECPLA) results in adverse outcomes compared to Klebsiella pneumoniae PLA (KPPLA). Methods A retrospective review of all patients admitted at a tertiary hospital in Singapore from 2003 to 2011 was performed. Patients with age \18 years, amoebic liver abscess, infected liver cyst, culture negative abscess or ruptured liver abscess requiring urgent surgical intervention were excluded. Only patients with blood or pus culture confirmation of ECPLA (n = 24) or KPPLA (n = 264) were included. Median length of hospital stay, failure of nonoperative therapy and 30-day mortality are the reported outcomes. Results ECPLA affects older patients (68 vs. 62 years, p = 0.049). Ischemic heart disease was more common in ECPLA (29 vs. 14 %, p = 0.048) and there was no difference in diabetic state (42 vs. 38 %, p = 0.743). ECPLA is more commonly associated with hyperbilirubinemia (60 vs. 34 lmol/L, p = 0.003), increased gamma-glutamyl transpeptidase (236 vs. 16 IU/L, p = 0.038) and gallstones (58 vs. 30 %, p = 0.004). KPPLA are larger in size (6 vs. 4 cm, p = 0.006) and had percutaneous drainage (PD) more frequently (64 vs. 42 %, p = 0.034). There was no difference in median hospital stay (14 vs. 14 days, p = 0.110) or 30-day mortality (17 vs. 10 %, p = 0.307) between ECPLA and KPPLA. Among patients with ECPLA, antibiotic treatment with PD appeared to have higher mortality compared to antibiotic treatment alone (30 vs. 7 %) but this was not significant (p = 0.272). Conclusion In the setting of multimodal care, outcomes of ECPLA are comparable to KPPLA.
Dr Vishal G Shelat and Dr Clement LK Chia contributed equally to the work. & Vishal G. Shelat [email protected] 1
Department of General Surgery, Tan Tock Seng Hospital, Singapore 308433, Singapore
2
Yong Loo Lin School of Medicine, National University of Singapore, Singapore 119228, Singapore
3
Ministry of Health Holdings, 1 Maritime Square, Singapore 099253, Singapore
Introduction Pyogenic liver abscess (PLA) is a major hepato-biliary infection with incidence ranging from 1.1 to 3.6 per 100,000 reported in previous population based studies [1– 4]. Despite advances in critical care medicine, anti-microbial therapy and interventional radiology; morbidity and mortality of PLA remains high [5–8]. Mortality rates of 1–31 % have been reported globally [2–4, 6, 7, 9–11]. Escherichia coli were the predominant pathogen worldwide until the 1980 s [6, 9, 12]. However, studies in the
123
World J Surg
past two decades report that Klebsiella pneumoniae has emerged as the leading pathogen in many countries [5, 10, 13, 14]. As antibiotic therapy coupled with percutaneous drainage (PD) is the mainstay of management, knowledge of causative pathogen is essential for treatment. Although E. coli PLA (ECPLA) and K. pneumoniae PLA (KPPLA) have been described separately in several studies [11, 14–20], there are few studies that directly compare these two common causative pathogens [21]. It is generally believed that outcomes of ECPLA are inferior compared to KPPLA and this may be due to inherent demographic and clinical differences. Locally, all the PLA patients are managed by a multidisciplinary team comprising of gastroenterologist (physician or surgeon), infectious disease specialist, nursing team and interventional radiologist. Our study aims to demonstrate if length of hospital stay, failure of non-operative therapy and 30 day mortality differ between ECPLA and KPPLA when a multimodal care is provided.
Materials and methods A retrospective medical record review of all PLA patients admitted to university affiliated tertiary hospital in Singapore was performed from Jan 2003 to Dec 2011. The list of patients is generated by accessing the ICD-10 (International Classification of Diseases, 10th Revision) codes of the hospital discharge database. This study was approved by hospital institutional review board. Patient demography, co-morbidity, clinical presentation, laboratory, radiographic, microbiology and clinical outcome data were collected and analysed.
Inclusion criteria PLA is suspected when a patient has pyrexia, abdominal pain or deranged liver enzymes and is confirmed by imaging. Patients with age \18 years, amoebic liver abscess, infected liver cyst, abscess due to other microorganisms, culture negative abscess or ruptured liver abscess requiring an urgent surgical intervention are excluded. We excluded patients with ruptured liver abscess as their management and outcomes are understandably different. Only patients with blood or pus culture confirmation of either K. pneumoniae or E. coli formed the study cohort. A total of 635 patients were treated for PLA during the study period and 288 patients [KPPLA, n = 264 (91.7 %) and ECPLA, n = 24 (8.3 %)] fulfilled the inclusion criteria. One patient grew both E. coli (blood culture) and K. pneumoniae (pus culture) and was included as ECPLA based on blood culture results.
123
Treatment protocol All patients with PLA had at least one set of blood culture taken prior to the initiation of empirical broad spectrum intravenous (IV) antibiotics, usually a combination of a third generation cephalosporin, such as ceftriaxone, along with metronidazole. Patients would generally undergo PD of PLA at the discretion of treating gastroenterologist (physician or surgeon) and clinical condition of patient. We use 10–12 Fr size drains for PD. PD would be performed when any of the following criteria was met: (1) Size of PLA [4 cm, solitary or dominant, (2) Presence of hemodynamic instability or need for ionotropic support on admission, (3) Gas within the abscess cavity regardless of size and (4) Failure of antibiotic therapy for PLA \4 cm. The size of 4 cm is arbitrary and is based on the mathematical principles of sphere volumes. It becomes obvious that with increment in diameter from 4 to 5 cm size, the volume of sphere doubles from 33 to 65 cc. The 4 cm size also allows for a satisfactory placement of pigtail drains and hence seems a logical cut off for PD. This is supported by reports on PD for colonic diverticular abscess [22, 23]. The specimens obtained were processed for gram stain, bacterial cultures (aerobic and anaerobic) and tests for antibiotic susceptibility (organism identification is performed using a mix of MicrobactTM 12A (Oxoid, UK), VITEK 2 system (bioMe´rieux), and the Bruker Biotyper (version 2.0) matrix-assisted laser desorption/ionisation time-of- flight (MALDI-TOF) mass spectrometry system; susceptibility testing was performed using the VITEK 2 system (bioMe´rieux) and Kirby–Bauer method). Antibiotic therapy was subsequently tailored to culture and sensitivity results of blood or pus. All the patients with positive blood culture were treated with 10–14 days of IV antibiotics either as an inpatient or as an outpatient. Total duration of antibiotic therapy varied according to the clinical and radiological response. Drains were flushed with 10 mL of saline to prevent blockages and accurate logs were maintained for the quality and quantity of effluents. A contrast dye study via the drain was routinely obtained when the drain output was \10 mL/24 h for at least 2 days. Depending on the radiologic finding of size of residual cavity; the drainage tube was either repositioned, upsized or withdrawn. The drains were removed upon resolution of sepsis as evidenced by stable vital parameters, total white cell count down trending, C reactive protein (CRP) down trending and \10 mLs/24 h drain aspirates for at least 2 days. Treatment outcomes were measured by length of in hospital stay, failure of non-operative therapy and 30 day mortality. For the purpose of this study, failure of nonoperative therapy was defined as patients having persistent signs of sepsis after 5 days of intravenous antibiotics and PD, hence requiring escalation of therapy (i.e. surgical
World J Surg
intervention) or death. All the patients underwent a repeat radiologic imaging at 2–3 week intervals to evaluate treatment response. Antibiotic therapy was discontinued when there was a clinical resolution and near complete to complete radiological resolution of abscess, after a minimum duration of 6 weeks. In instances where a patient has completed at least 6 weeks of antibiotics (intravenous and oral) and sepsis resolves, antibiotics were discontinued despite a small residual persistent abscess on imaging. Figure 1 shows the radiological diagnosis, drainage and resolution of a patient treated for KPPLA at our institution. All the patients who had an index computerised tomography (CT) scan for diagnosis of PLA and no evidence of biliary lithiasis, an ultrasound of abdomen was performed at next available opportunity to establish diagnosis of gallstones. All the patients with gallstones, regardless of imaging features suggestive of cholecystitis were offered a laparoscopic cholecystectomy after completion of antibiotic therapy. Patients without gallstones were evaluated for colonic source. Colonoscopy was performed for patients with colonic symptoms (change in bowel habits, history of bleeding per rectum) and/or imaging abnormality of colon (e.g. thickened wall). Figure 2 illustrates the flowchart of management. Statistical Analysis Data were analysed using commercial statistical software (SPSS for Windows, version 14.0, Chicago, Illinois, USA). All continuous data were expressed as mean ± standard deviation (SD) and analysed by Independent-Samples-T test. All categorical variables were described as percentage and compared by either v2 or Fisher exact test. p value \0.05 was considered statistically significant. Due to small number of patients in ECPLA group, multivariate adjustment was not performed.
Results Demographic data, symptoms and signs The demographic data, co-morbidity and presenting symptoms are summarised in Table 1. ECPLA patients were older (68 vs. 62 years, p = 0.049). Ischemic heart disease was more common in ECPLA (29 vs. 14 %, p = 0.048) but there was no difference in diabetic status (p = 0.743). Fever was the most common presenting symptom in both groups. Biochemistry and imaging ECPLA patients had higher levels of total bilirubin (60 vs. 34 lmol/L, p = 0.003) and gamma-glutamyltranspeptidase
Fig. 1 A 70-year-old man was admitted for abdominal pain and fever of 2 days duration. Axial computerized tomography (CT) scan image showing a 9 cm multi-loculated segment VI liver abscess (a). This was treated with antibiotics and percutaneous drainage. Tube review demonstrated persistent cavity and hence tube was left in situ (b). Axial CT scan image showing resolution at 3 months follow-up (c)
(GGT) (236 vs. 16 IU/L, p = 0.038). Gallstones were more common in ECPLA patients (58 vs. 30 %, p = 0.004). KPPLA were larger in size (6 vs. 4 cm, p = 0.006). Majority of the KPPLA (70 %) and ECPLA (71 %) PLA were solitary.
123
World J Surg
Jan 2003 – Dec 2011 n = 635 (100)
Excluded: - 83 (13) other microorganisms - 264 (41.6) Culture negative PLA
Klebsiella pneumonia and Escherichia coli n = 288 (100)
Klebsiella pneumonia n = 264 (92) Antibiotics only 95 (36) Resolved 83 (87)
Death 12 (13)
Escherichia coli n = 24 (8)
Antibiotics with PD 169 (64) Resolved 154 (91)
Cause of death: 5 MOF 2 Pneumonia 2 Hepatocellular Cancer 1 Lung cancer 1 Pancreas Cancer 1 Urosepsis
a
ESRF – end stage renal failure
b
MOF – multiorgan failure
c
PD – percutaneous drainage
Death 15 (9) Cause of death: 6 Pneumonia 4 Cardiac event 3 MOF 1 Psoas abscess 1 ESRF
Antibiotics only 14 (58) Resolved 13 (93)
Death 1 (7) Cause of death: 1 MOF
Antibiotics with PD 10 (42) Resolved 7 (70)
Death 3 (30) Cause of death: 1 MOF 1 Hepatocellular Cancer 1 Cardiac event
Fig. 2 Outcomes of patients with pyogenic liver abscess (PLA) secondary to Klebsiella pneumoniae and Escherichia coli—n (%)
Treatment and outcome
Discussion
More KPPLA patients were treated with PD (64 vs. 42 %, p = 0.034). The outcomes measured in this study are summarised in Table 2. There was no difference in median hospital stay (14 vs. 14 days, p = 0.110). Overall 30 day mortality in our series was 31/288 (10.7 %). There was no difference in 30 day mortality among ECPLA and KPPLA patients (17 vs. 10 %, p = 0.307). One-third of the deaths (10 out of 31) in the study cohort were attributable to multiorgan failure. The distribution and characteristics of patients with multi-organ failure is summarised in Table 3. Surgical drainage was not performed in this series. Among the patients with ECPLA, antibiotic treatment along with PD appeared to have higher mortality compared to patients treated with antibiotics alone (30 vs. 7 %) but this was not significant (p = 0.272).
The past two decades have seen a gradual shift from E. coli to K. pneumoniae as the leading pathogen isolated from PLA; especially in Asian countries, including Singapore [5, 10, 13, 14, 24, 25]. This trend is reflected in our study, which shows a predominance of KPPLA. This is the largest study comparing outcomes between the two most common PLA pathogens. Our study demonstrates that ECPLA occurs more commonly in older patients with ischemic heart disease and is more likely to be associated with underlying biliary disease. Hence ECPLA management does pose distinct challenges. Figure 3 demonstrates adversaries associated with ECPLA. We have shown that despite demographic and clinical differences, outcomes of ECPLA are comparable to KPPLA when multimodal care is provided.
123
World J Surg Table 1 Demography, co-morbidity, symptoms and imaging of patients with Klebsiella pneumoniae and Escherichia coli pyogenic liver abscess Klebsiella pneumoniae (n = 264)
Escherichia coli (n = 24)
OR (95 % C.I.)
p value
Male:Female
160:104
12:12
1.54 (0.67–3.55)
0.310
Mean Age in years
62 (25–94)
68 (21–89)
NA
0.049
Hypertension
120 (46)
15 (63)
0.50 (0.21–1.18)
0.109
Hyperlipidaemia
80 (30)
5 (21)
1.65 (0.60–4.58)
0.483
Diabetes
101 (38)
10 (42)
0.87 (0.37–2.03)
0.743
Renal failure
14 (5)
2 (8)
0.62 (0.13–2.89)
0.632
Ischaemic heart disease
37 (14)
7 (29)
0.40 (0.15–1.02)
0.048
I
109 (41)
7 (29)
1.68 (0.69–4.26)
0.246
II CIII
135 (51) 20 (8)
12 (50) 5 (21)
1.05 (0.45–2.41) 0.41 (0.13–1.32)
0.915 0.130
Fever
225 (85)
21 (88)
0.82 (0.24–2.90)
1.0
Abdominal pain
114 (43)
8 (33)
1.52 (0.63–3.68)
0.350
Jaundice
21 (8)
4 (17)
0.43 (0.14–1.38)
0.142
Constitutional symptomsa
219 (83)
21 (88)
0.70 (0.20–2.43)
0.777
Co-morbidity [n (%)]
ASA [n (%)]
Symptoms [n (%)]
Computer Tomographic imaging [n (%)] Solitary lesion
185 (70)
17 (71)
0.96 (0.39–2.42)
0.938
Mean size of abscess (cm)
6 (0.8–15.4)
4 (1–11.2)
NA
0.006
NA not applicable, ASA American Society of Anaesthesiologists classification system a
Constitutional symptoms—Non-specific symptoms such as anorexia, malaise and nausea/vomiting
Table 2 Treatment outcomes of patients with Klebsiella pneumoniae and Escherichia coli pyogenic liver abscess Outcomes
Klebsiella pneumonia (n = 264)
Escherichia coli (n = 24)
p value
Median length of hospital stay (days)
14 (1–80)
14 (4–99)
0.110
Failure of non-operative therapy [n (%)] 30 day mortality [n (%)]
15 (9)
3 (13)
0.180
27 (10)
4 (17)
0.307
Our study found that ECPLA patients were older than KPPLA patients and this was consistent with the findings by Chen et al. [21]. Chen et al. also found a significant association between ECPLA and female gender which was consistent with our study findings although this did not reach statistical significance, possibly due to smaller sample size. The incidence of diabetes in our study is comparable to other studies 25–70 % [11, 16, 17, 20, 21, 25, 26]. Ours is the first study in the literature to report that ECPLA is more likely to occur in patients with ischemic heart disease. In a local study on KPPLA [25], the prevalence of ischemic heart disease was even lower at 6 %. This low incidence could probably be due to the fact that
only patients receiving outpatient antibiotic treatment were included. Patients with ECPLA are older with higher prevalence of ischemic heart disease and are more likely to receive in hospital treatment. There was no difference in presenting symptoms between the two pathogens. This is similar to previous report of Chen et al. [21]. However, we found fever and constitutional symptoms as the most common presenting symptoms as compared to Chen et al. [18] who reported fever and right upper quadrant pain. This difference in the presentation could be explained by larger size of abscess in their series ([50 % of patients with [5 cm PLA). Our study found an association between ECPLA and underlying biliary pathology. This is consistent with other reports [18, 21, 27, 28]. Biliary disease is more common in females and jaundice is a common clinical presenting symptom, but our study did not show this difference. This may be due to small number of ECPLA patients. In our institution, all PLA patients are managed by a multidisciplinary team of gastroenterologist (physician and/or surgeon), infectious disease specialist, nursing team and interventional radiologist. The current mainstay of PLA treatment in our institution involves a combination of systemic antibiotics and PD. Size and presence of loculation within the abscess are important determinants of the
123
World J Surg Table 3 Characteristics of patients who died from multi-organ failure Patient number
1
2
3
4
5
6
7
8
9
10
Demographics Age (years)
66
67
57
53
61
53
70
79
69
79
Gender
F
F
F
M
F
M
F
M
M
M
Comorbidities Hypertension
?
–
–
?
–
–
–
–
?
–
Dyslipidemia
–
–
?
–
–
–
?
–
?
–
Diabetes
?
?
?
?
–
–
–
–
–
–
Renal failure Ischaemic heart disease
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
ASA score
II
I
I
II
I
I
II
III
IV
III
Fever
?
?
?
?
?
–
–
?
?
?
Abdominal pain
?
–
–
–
?
–
?
–
–
–
Jaundice
?
–
?
–
–
–
–
–
–
–
Constitutional symptoms
?
–
?
?
?
?
?
?
?
?
Symptoms
Investigation findings Bilirubin (lmol/L)
80
21
74
10
60
35
79
94
26
94
GGT (IU/L)
381
31
335
14
24
90
980
224
245
224
Gallstones
–
–
–
–
–
–
–
–
?
–
Abscess characteristics Solitary
–
?
?
?
–
?
?
?
?
?
Size (cm)
6.5
10
8.4
3
7.7
9.7
5.8
5.2
1.4
5.2
KP
KP
KP
KP
KP
KP
KP
KP
EC
EC
Organism Treatment and outcome Percutaneous drainage
–
?
–
–
–
?
–
?
–
?
Length of stay (days)
9
28
2
3
2
10
15
98
14
98
ASA American Society of Anaesthesiology, GGT Gamma-glutamyl transpeptidase, KP Klebsiella pneumoniae, EC Escherichia coli
Fig. 3 Adversaries attributed to Escherichia coli pyogenic liver abscess
123
World J Surg
need for PD and surgical drainage [24]. KPPLA has a tendency to form larger abscess compared to other bacteria [29]. KPPLA size of more than 5 cm is reported in 71 % (104/147) by Chen et al. [21]. The larger mean abscess size of KPPLA patients could be a possible contributory factor for the higher proportion of PD observed in our study. Multiple liver abscesses could suggest increased virulence and more severe nature of illness with a potential for higher mortality [10]. Our study showed that majority of PLA was solitary regardless of the causative pathogen and majority of patients who died of multi-organ failure also had a solitary PLA. We reserve surgical drainage for patients who have ruptured abscesses or who demonstrate failure of non-operative therapy. Although there have been local studies that advocate the use of surgical drainage for PLA [5 cm [24] and laparoscopic drainage in selected patients [30] with reportedly better clinical outcomes than PD, none of our patients required surgical drainage. Claims favouring routine surgical drainage based on large size needs to be verified by prospective randomised trials. Average hospital stay of patients with PLA tends to be high due to need for intravenous antibiotics and PD. In the endeavour to reduce overall length of stay, outpatient parenteral antibiotic therapy (OPAT) is preferred. This is only feasible in selected clinically stable patients with appropriate antibiotic sensitivity and in healthcare systems where such a facility is available. We do have an active and integrated multidisciplinary management protocol and framework in place for OPAT. Despite the presence of such framework, median length of stay of our patients is 2 weeks which indicates that PLA is a drain to healthcare resources. Mortality from PLA is globally reported as 1–31 % [2– 4, 6, 7, 9–11]. It could be speculated that as ECPLA tend to be small in size, they are less likely to undergo PD and this could lead to poor outcomes. However, ECPLA patients treated with antibiotics and PD paradoxically showed a trend towards increased mortality suggesting that more sick patients had a higher mortality despite PD. Hence it appears that systemic antibiotic treatment is more important in ECPLA compared to local therapy with PD. Chen et al. [21] reasoned that the higher mortality of ECPLA patients in his series could be contributed by higher Acute Physiology and Chronic Health Evaluation (APACHE II) score at admission. In our study, we examined baseline American Society of Anaesthesiology (ASA) score and there was no statistical difference. It is possible that proactive upfront PD could reduce the progression or worsening of systemic sepsis and multi-organ failure while a salvage PD for failure of ‘antibiotic only’ therapy has limited role in already septic patient. Due to retrospective nature of our study, we are unable to determine to timing of PD with regards to clinical condition and hence did not define a failure of ‘antibiotic only’ therapy.
Figure 2 shows causes of mortality in patients with PLA and it is apparent that one-third of all the mortality is attributed to multi-organ failure. Majority of patients who succumbed to multi-organ failure had PLA [5 cm (Table 3). It remains to be determined if early elective surgical drainage in selected patients could reduce this mortality. Treatment of PLA requires a quartet of early diagnosis, prompt resuscitation, aggressive proactive policy of PD and integrated multidisciplinary management protocols. In a local study on outpatient management of KPPLA involving our institution, Chan et al. [25] reported zero mortality in his series, suggesting that the prognosis of this group of patients remain favourable if they recover from acute hepatobiliary sepsis in the hospital and continue with OPAT. The strength of our study lies in large sample which is possible due to higher prevalence of PLA in local population. The main limitation of our study is its retrospective nature with potential recall bias. It is also possible that some patients with polymicrobial infection which have been categorised as ECPLA or KPPLA based on their presence in the culture, the outcomes were influenced by presence of other bacteria. The increased mortality of ECPLA patients with PD compared to antibiotics alone appears paradoxical and calls for an international collaboration to standardise terminology with regards to ‘proactive’ PD and ‘salvage’ PD. Only when such distinction in terminology is made and timing of PD is known, it could be proven if systemic antibiotic therapy plays a dominant role in ECPLA management and PD plays a dominant role in KPPLA management. In conclusion, there are differences in patient demography and clinical profile of PLA patients with these two different pathogens. Despite such differences, nonoperative management of PLA appears to be highly successful regardless of aetiology. The higher mortality of ECPLA reported by others is not observed in our series. While, it is likely that such good outcomes are secondary to multimodal care, such conclusions can rightfully be derived only by future multicentre collaborative studies. Conflict of interest The authors do not report any conflicts of interest.
References 1. Kaplan GG, Gregson DB, Laupland KB (2004) Population-based study of the epidemiology of and the risk factors for pyogenic liver abscess. Clin Gastroenterol Hepatol 2:1032–1038 2. Jepsen P, Vilstrup H, Schonheyder HC et al (2005) A nationwide study of the incidence and 30-day mortality rate of pyogenic liver abscess in Denmark, 1977–2002. Aliment Pharmacol Ther 21:1185–1188 3. Hansen PS, Schonheyder HC (1998) Pyogenic hepatic abscess. A 10-year population based retrospective study. APMIS 106: 396–402
123
World J Surg 4. Meddings L, Myers RP, Hubbard J et al (2010) A populationbased study of pyogenic liver abscesses in the United States: incidence, mortality and temporal trends. Am J Gastroenterol 105:117–124 5. Branum GD, Tyson GS, Branun MA et al (1990) Hepatic abscess: changes in etiology, diagnosis, and management. Ann Surg 212:655–662 6. Seeto RK, Rockey DC (1996) Pyogenic liver abscess: changes in etiology, management, and outcome. Medicine 75:99–113 7. Chu KM, Fan ST, Lai ECS et al (1996) Pyogenic liver abscess: an audit of experience over the past decade. Arch Surg 131:148–152 8. Yu SC, Lo RH, Kan PS et al (1997) Pyogenic liver abscess: treatment with needle aspiration. Clin Radiol 52:912–916 9. Gyorffy EJ, Frey CF, Silva J Jr et al (1987) Pyogenic liver abscess: diagnostic and therapeutic strategies. Ann Surg 206:699–705 10. Huang CJ, Pitt HA, Lipsett P et al (1996) Pyogenic hepatic abscess: changing trends over 42 years. Ann Surg 223:600–609 11. Li J, Fu Y, Wang JY et al (2010) Early diagnosis and therapeutic choice of Klebsiella pneumoniae liver abscess. Front Med China 4(3):308–316 12. Cerwenka H (2010) Pyogenic liver abscess: differences in aetiology and treatment in Southeast Asia and Central Europe. World J Gastroenterol 16(20):2458–2462 13. Rahimian J, Wilson T, Oram V et al (2004) Pyogenic liver abscess: recent trends in etiology and mortality. Clin Infect Dis 39:1654–1659 14. Lederman ER, Crum NF (2005) Pyogenic liver abscess with a focus on Klebsiella pneumoniae as a primary pathogen: an emerging disease with unique clinical characteristics. Am J Gastroenterol 100:322–331 15. Liu Y, Wang JY, Jiang W (2013) An increasing prominent disease of Klebsiella pneumonia liver abscess: etiology, diagnosis and treatment. Gastroenterol Research and Practice 1–12 16. Chung DR, Lee SS, Lee HR et al (2007) Emerging invasive liver abscess caused by K1 serotype Klebsiella pneumonia in Korea. J Inf 54:578–583 17. Chan KS, Yu WL, Tsai CL et al (2007) Pyogenic liver abscess caused by Klebsiella pneumonia: analysis of the clinical characteristics and outcomes of 84 patients. Chin Med J 120(2):136–139
123
18. Chen SC, Yen CH, Lai KC et al (2005) Pyogenic liver abscesses with Escherichia coli: etiology, clinical course, outcome, and prognostic factors. Wien Klin Wochenschr 117(23–24):809–815 19. Pastagia M, Arumugam V (2008) Klebsiella pneumoniae liver abscesses in a public hospital in Queens, New York. Travel Med Infect Dis 6:228–233 20. Wang JH, Liu YC, Lee SSJ et al (1998) Primary liver abscess due to Klebsiella pneumoniae in Taiwan. Clin Infect Dis 26:1434–1438 21. Chen SC, Wu WY, Yeh CH et al (2007) Comparison of Escherichia coli and Klebsiella pneumoniae liver abscesses. Am J Med Sci 334(2):97–105 22. Durmishi Y, Gervaz P, Brandt D et al (2006) Results from percutaneous drainage of Hinchey stage II diverticulitis guided by computed tomography scan. Surg Endosc 20:1129–1133 23. Siewert B, Tye G, Kruskal J, Sosna J, Opelka F (2006) Impact of CT-guided drainage in the treatment of diverticular abscess: size matters. Am J Roentgenol 186:680–686 24. Tan YM, Chung AYF, Chow PKH et al (2005) An appraisal of surgical and percutaneous drainage for pyogenic liver abscesses larger than 5 cm. Ann Surg 241:485–490 25. Chan DS, Archuleta S, Llorin RM et al (2013) Standardized outpatient management of Klebsiella pneumoniae liver abscesses. Int J Infect Dis 17(3):185–188 26. Yang CC, Yen CH, Ho MW et al (2004) Comparison of pyogenic liver abscess caused by non-Klebsiella pneumoniae and Klebsiella pneumoniae. J Microbiol Immunol Infect 37:176–184 27. Chen SC, Yen CH, Tsao SM et al (2005) Comparison of pyogenic liver abscesses of biliary and cryptogenic origin: an eight-year analysis in a University Hospital. Swiss Med Wkly 135:344–351 28. Cohen JL, Martin FM, Rossi RL et al (1989) Liver abscess. The need for complete gastrointestinal evaluation. Arch Surg 124:561–564 29. Kim SB, Je BK, Lee KY et al (2007) Computed tomographic differences of pyogenic liver abscesses caused by Klebsiella pneumoniae and non-Klebsiella pneumoniae. J Comput Assist Tomogr 31:59–65 30. Tan L, Zhou HJ, Hartman M et al (2013) Laparoscopic drainage of cryptogenic liver abscess. Surg Endosc 27(9):3308–3314
ORIGINAL SCIENTIFIC REPORT
Pyogenic Liver Abscess: Does Escherichia Coli Cause more Adverse Outcomes than Klebsiella Pneumoniae? Vishal G. Shelat1 • Clement L. K. Chia1,3 • Charleen S. W. Yeo2 • Wang Qiao3 Winston Woon1 • Sameer P. Junnarkar1
•
Ó Socie´te´ Internationale de Chirurgie 2015
Abstract Background This paper aims to demonstrate if Escherichia coli pyogenic liver abscess (ECPLA) results in adverse outcomes compared to Klebsiella pneumoniae PLA (KPPLA). Methods A retrospective review of all patients admitted at a tertiary hospital in Singapore from 2003 to 2011 was performed. Patients with age \18 years, amoebic liver abscess, infected liver cyst, culture negative abscess or ruptured liver abscess requiring urgent surgical intervention were excluded. Only patients with blood or pus culture confirmation of ECPLA (n = 24) or KPPLA (n = 264) were included. Median length of hospital stay, failure of nonoperative therapy and 30-day mortality are the reported outcomes. Results ECPLA affects older patients (68 vs. 62 years, p = 0.049). Ischemic heart disease was more common in ECPLA (29 vs. 14 %, p = 0.048) and there was no difference in diabetic state (42 vs. 38 %, p = 0.743). ECPLA is more commonly associated with hyperbilirubinemia (60 vs. 34 lmol/L, p = 0.003), increased gamma-glutamyl transpeptidase (236 vs. 16 IU/L, p = 0.038) and gallstones (58 vs. 30 %, p = 0.004). KPPLA are larger in size (6 vs. 4 cm, p = 0.006) and had percutaneous drainage (PD) more frequently (64 vs. 42 %, p = 0.034). There was no difference in median hospital stay (14 vs. 14 days, p = 0.110) or 30-day mortality (17 vs. 10 %, p = 0.307) between ECPLA and KPPLA. Among patients with ECPLA, antibiotic treatment with PD appeared to have higher mortality compared to antibiotic treatment alone (30 vs. 7 %) but this was not significant (p = 0.272). Conclusion In the setting of multimodal care, outcomes of ECPLA are comparable to KPPLA.
Dr Vishal G Shelat and Dr Clement LK Chia contributed equally to the work. & Vishal G. Shelat [email protected] 1
Department of General Surgery, Tan Tock Seng Hospital, Singapore 308433, Singapore
2
Yong Loo Lin School of Medicine, National University of Singapore, Singapore 119228, Singapore
3
Ministry of Health Holdings, 1 Maritime Square, Singapore 099253, Singapore
Introduction Pyogenic liver abscess (PLA) is a major hepato-biliary infection with incidence ranging from 1.1 to 3.6 per 100,000 reported in previous population based studies [1– 4]. Despite advances in critical care medicine, anti-microbial therapy and interventional radiology; morbidity and mortality of PLA remains high [5–8]. Mortality rates of 1–31 % have been reported globally [2–4, 6, 7, 9–11]. Escherichia coli were the predominant pathogen worldwide until the 1980 s [6, 9, 12]. However, studies in the
123
World J Surg
past two decades report that Klebsiella pneumoniae has emerged as the leading pathogen in many countries [5, 10, 13, 14]. As antibiotic therapy coupled with percutaneous drainage (PD) is the mainstay of management, knowledge of causative pathogen is essential for treatment. Although E. coli PLA (ECPLA) and K. pneumoniae PLA (KPPLA) have been described separately in several studies [11, 14–20], there are few studies that directly compare these two common causative pathogens [21]. It is generally believed that outcomes of ECPLA are inferior compared to KPPLA and this may be due to inherent demographic and clinical differences. Locally, all the PLA patients are managed by a multidisciplinary team comprising of gastroenterologist (physician or surgeon), infectious disease specialist, nursing team and interventional radiologist. Our study aims to demonstrate if length of hospital stay, failure of non-operative therapy and 30 day mortality differ between ECPLA and KPPLA when a multimodal care is provided.
Materials and methods A retrospective medical record review of all PLA patients admitted to university affiliated tertiary hospital in Singapore was performed from Jan 2003 to Dec 2011. The list of patients is generated by accessing the ICD-10 (International Classification of Diseases, 10th Revision) codes of the hospital discharge database. This study was approved by hospital institutional review board. Patient demography, co-morbidity, clinical presentation, laboratory, radiographic, microbiology and clinical outcome data were collected and analysed.
Inclusion criteria PLA is suspected when a patient has pyrexia, abdominal pain or deranged liver enzymes and is confirmed by imaging. Patients with age \18 years, amoebic liver abscess, infected liver cyst, abscess due to other microorganisms, culture negative abscess or ruptured liver abscess requiring an urgent surgical intervention are excluded. We excluded patients with ruptured liver abscess as their management and outcomes are understandably different. Only patients with blood or pus culture confirmation of either K. pneumoniae or E. coli formed the study cohort. A total of 635 patients were treated for PLA during the study period and 288 patients [KPPLA, n = 264 (91.7 %) and ECPLA, n = 24 (8.3 %)] fulfilled the inclusion criteria. One patient grew both E. coli (blood culture) and K. pneumoniae (pus culture) and was included as ECPLA based on blood culture results.
123
Treatment protocol All patients with PLA had at least one set of blood culture taken prior to the initiation of empirical broad spectrum intravenous (IV) antibiotics, usually a combination of a third generation cephalosporin, such as ceftriaxone, along with metronidazole. Patients would generally undergo PD of PLA at the discretion of treating gastroenterologist (physician or surgeon) and clinical condition of patient. We use 10–12 Fr size drains for PD. PD would be performed when any of the following criteria was met: (1) Size of PLA [4 cm, solitary or dominant, (2) Presence of hemodynamic instability or need for ionotropic support on admission, (3) Gas within the abscess cavity regardless of size and (4) Failure of antibiotic therapy for PLA \4 cm. The size of 4 cm is arbitrary and is based on the mathematical principles of sphere volumes. It becomes obvious that with increment in diameter from 4 to 5 cm size, the volume of sphere doubles from 33 to 65 cc. The 4 cm size also allows for a satisfactory placement of pigtail drains and hence seems a logical cut off for PD. This is supported by reports on PD for colonic diverticular abscess [22, 23]. The specimens obtained were processed for gram stain, bacterial cultures (aerobic and anaerobic) and tests for antibiotic susceptibility (organism identification is performed using a mix of MicrobactTM 12A (Oxoid, UK), VITEK 2 system (bioMe´rieux), and the Bruker Biotyper (version 2.0) matrix-assisted laser desorption/ionisation time-of- flight (MALDI-TOF) mass spectrometry system; susceptibility testing was performed using the VITEK 2 system (bioMe´rieux) and Kirby–Bauer method). Antibiotic therapy was subsequently tailored to culture and sensitivity results of blood or pus. All the patients with positive blood culture were treated with 10–14 days of IV antibiotics either as an inpatient or as an outpatient. Total duration of antibiotic therapy varied according to the clinical and radiological response. Drains were flushed with 10 mL of saline to prevent blockages and accurate logs were maintained for the quality and quantity of effluents. A contrast dye study via the drain was routinely obtained when the drain output was \10 mL/24 h for at least 2 days. Depending on the radiologic finding of size of residual cavity; the drainage tube was either repositioned, upsized or withdrawn. The drains were removed upon resolution of sepsis as evidenced by stable vital parameters, total white cell count down trending, C reactive protein (CRP) down trending and \10 mLs/24 h drain aspirates for at least 2 days. Treatment outcomes were measured by length of in hospital stay, failure of non-operative therapy and 30 day mortality. For the purpose of this study, failure of nonoperative therapy was defined as patients having persistent signs of sepsis after 5 days of intravenous antibiotics and PD, hence requiring escalation of therapy (i.e. surgical
World J Surg
intervention) or death. All the patients underwent a repeat radiologic imaging at 2–3 week intervals to evaluate treatment response. Antibiotic therapy was discontinued when there was a clinical resolution and near complete to complete radiological resolution of abscess, after a minimum duration of 6 weeks. In instances where a patient has completed at least 6 weeks of antibiotics (intravenous and oral) and sepsis resolves, antibiotics were discontinued despite a small residual persistent abscess on imaging. Figure 1 shows the radiological diagnosis, drainage and resolution of a patient treated for KPPLA at our institution. All the patients who had an index computerised tomography (CT) scan for diagnosis of PLA and no evidence of biliary lithiasis, an ultrasound of abdomen was performed at next available opportunity to establish diagnosis of gallstones. All the patients with gallstones, regardless of imaging features suggestive of cholecystitis were offered a laparoscopic cholecystectomy after completion of antibiotic therapy. Patients without gallstones were evaluated for colonic source. Colonoscopy was performed for patients with colonic symptoms (change in bowel habits, history of bleeding per rectum) and/or imaging abnormality of colon (e.g. thickened wall). Figure 2 illustrates the flowchart of management. Statistical Analysis Data were analysed using commercial statistical software (SPSS for Windows, version 14.0, Chicago, Illinois, USA). All continuous data were expressed as mean ± standard deviation (SD) and analysed by Independent-Samples-T test. All categorical variables were described as percentage and compared by either v2 or Fisher exact test. p value \0.05 was considered statistically significant. Due to small number of patients in ECPLA group, multivariate adjustment was not performed.
Results Demographic data, symptoms and signs The demographic data, co-morbidity and presenting symptoms are summarised in Table 1. ECPLA patients were older (68 vs. 62 years, p = 0.049). Ischemic heart disease was more common in ECPLA (29 vs. 14 %, p = 0.048) but there was no difference in diabetic status (p = 0.743). Fever was the most common presenting symptom in both groups. Biochemistry and imaging ECPLA patients had higher levels of total bilirubin (60 vs. 34 lmol/L, p = 0.003) and gamma-glutamyltranspeptidase
Fig. 1 A 70-year-old man was admitted for abdominal pain and fever of 2 days duration. Axial computerized tomography (CT) scan image showing a 9 cm multi-loculated segment VI liver abscess (a). This was treated with antibiotics and percutaneous drainage. Tube review demonstrated persistent cavity and hence tube was left in situ (b). Axial CT scan image showing resolution at 3 months follow-up (c)
(GGT) (236 vs. 16 IU/L, p = 0.038). Gallstones were more common in ECPLA patients (58 vs. 30 %, p = 0.004). KPPLA were larger in size (6 vs. 4 cm, p = 0.006). Majority of the KPPLA (70 %) and ECPLA (71 %) PLA were solitary.
123
World J Surg
Jan 2003 – Dec 2011 n = 635 (100)
Excluded: - 83 (13) other microorganisms - 264 (41.6) Culture negative PLA
Klebsiella pneumonia and Escherichia coli n = 288 (100)
Klebsiella pneumonia n = 264 (92) Antibiotics only 95 (36) Resolved 83 (87)
Death 12 (13)
Escherichia coli n = 24 (8)
Antibiotics with PD 169 (64) Resolved 154 (91)
Cause of death: 5 MOF 2 Pneumonia 2 Hepatocellular Cancer 1 Lung cancer 1 Pancreas Cancer 1 Urosepsis
a
ESRF – end stage renal failure
b
MOF – multiorgan failure
c
PD – percutaneous drainage
Death 15 (9) Cause of death: 6 Pneumonia 4 Cardiac event 3 MOF 1 Psoas abscess 1 ESRF
Antibiotics only 14 (58) Resolved 13 (93)
Death 1 (7) Cause of death: 1 MOF
Antibiotics with PD 10 (42) Resolved 7 (70)
Death 3 (30) Cause of death: 1 MOF 1 Hepatocellular Cancer 1 Cardiac event
Fig. 2 Outcomes of patients with pyogenic liver abscess (PLA) secondary to Klebsiella pneumoniae and Escherichia coli—n (%)
Treatment and outcome
Discussion
More KPPLA patients were treated with PD (64 vs. 42 %, p = 0.034). The outcomes measured in this study are summarised in Table 2. There was no difference in median hospital stay (14 vs. 14 days, p = 0.110). Overall 30 day mortality in our series was 31/288 (10.7 %). There was no difference in 30 day mortality among ECPLA and KPPLA patients (17 vs. 10 %, p = 0.307). One-third of the deaths (10 out of 31) in the study cohort were attributable to multiorgan failure. The distribution and characteristics of patients with multi-organ failure is summarised in Table 3. Surgical drainage was not performed in this series. Among the patients with ECPLA, antibiotic treatment along with PD appeared to have higher mortality compared to patients treated with antibiotics alone (30 vs. 7 %) but this was not significant (p = 0.272).
The past two decades have seen a gradual shift from E. coli to K. pneumoniae as the leading pathogen isolated from PLA; especially in Asian countries, including Singapore [5, 10, 13, 14, 24, 25]. This trend is reflected in our study, which shows a predominance of KPPLA. This is the largest study comparing outcomes between the two most common PLA pathogens. Our study demonstrates that ECPLA occurs more commonly in older patients with ischemic heart disease and is more likely to be associated with underlying biliary disease. Hence ECPLA management does pose distinct challenges. Figure 3 demonstrates adversaries associated with ECPLA. We have shown that despite demographic and clinical differences, outcomes of ECPLA are comparable to KPPLA when multimodal care is provided.
123
World J Surg Table 1 Demography, co-morbidity, symptoms and imaging of patients with Klebsiella pneumoniae and Escherichia coli pyogenic liver abscess Klebsiella pneumoniae (n = 264)
Escherichia coli (n = 24)
OR (95 % C.I.)
p value
Male:Female
160:104
12:12
1.54 (0.67–3.55)
0.310
Mean Age in years
62 (25–94)
68 (21–89)
NA
0.049
Hypertension
120 (46)
15 (63)
0.50 (0.21–1.18)
0.109
Hyperlipidaemia
80 (30)
5 (21)
1.65 (0.60–4.58)
0.483
Diabetes
101 (38)
10 (42)
0.87 (0.37–2.03)
0.743
Renal failure
14 (5)
2 (8)
0.62 (0.13–2.89)
0.632
Ischaemic heart disease
37 (14)
7 (29)
0.40 (0.15–1.02)
0.048
I
109 (41)
7 (29)
1.68 (0.69–4.26)
0.246
II CIII
135 (51) 20 (8)
12 (50) 5 (21)
1.05 (0.45–2.41) 0.41 (0.13–1.32)
0.915 0.130
Fever
225 (85)
21 (88)
0.82 (0.24–2.90)
1.0
Abdominal pain
114 (43)
8 (33)
1.52 (0.63–3.68)
0.350
Jaundice
21 (8)
4 (17)
0.43 (0.14–1.38)
0.142
Constitutional symptomsa
219 (83)
21 (88)
0.70 (0.20–2.43)
0.777
Co-morbidity [n (%)]
ASA [n (%)]
Symptoms [n (%)]
Computer Tomographic imaging [n (%)] Solitary lesion
185 (70)
17 (71)
0.96 (0.39–2.42)
0.938
Mean size of abscess (cm)
6 (0.8–15.4)
4 (1–11.2)
NA
0.006
NA not applicable, ASA American Society of Anaesthesiologists classification system a
Constitutional symptoms—Non-specific symptoms such as anorexia, malaise and nausea/vomiting
Table 2 Treatment outcomes of patients with Klebsiella pneumoniae and Escherichia coli pyogenic liver abscess Outcomes
Klebsiella pneumonia (n = 264)
Escherichia coli (n = 24)
p value
Median length of hospital stay (days)
14 (1–80)
14 (4–99)
0.110
Failure of non-operative therapy [n (%)] 30 day mortality [n (%)]
15 (9)
3 (13)
0.180
27 (10)
4 (17)
0.307
Our study found that ECPLA patients were older than KPPLA patients and this was consistent with the findings by Chen et al. [21]. Chen et al. also found a significant association between ECPLA and female gender which was consistent with our study findings although this did not reach statistical significance, possibly due to smaller sample size. The incidence of diabetes in our study is comparable to other studies 25–70 % [11, 16, 17, 20, 21, 25, 26]. Ours is the first study in the literature to report that ECPLA is more likely to occur in patients with ischemic heart disease. In a local study on KPPLA [25], the prevalence of ischemic heart disease was even lower at 6 %. This low incidence could probably be due to the fact that
only patients receiving outpatient antibiotic treatment were included. Patients with ECPLA are older with higher prevalence of ischemic heart disease and are more likely to receive in hospital treatment. There was no difference in presenting symptoms between the two pathogens. This is similar to previous report of Chen et al. [21]. However, we found fever and constitutional symptoms as the most common presenting symptoms as compared to Chen et al. [18] who reported fever and right upper quadrant pain. This difference in the presentation could be explained by larger size of abscess in their series ([50 % of patients with [5 cm PLA). Our study found an association between ECPLA and underlying biliary pathology. This is consistent with other reports [18, 21, 27, 28]. Biliary disease is more common in females and jaundice is a common clinical presenting symptom, but our study did not show this difference. This may be due to small number of ECPLA patients. In our institution, all PLA patients are managed by a multidisciplinary team of gastroenterologist (physician and/or surgeon), infectious disease specialist, nursing team and interventional radiologist. The current mainstay of PLA treatment in our institution involves a combination of systemic antibiotics and PD. Size and presence of loculation within the abscess are important determinants of the
123
World J Surg Table 3 Characteristics of patients who died from multi-organ failure Patient number
1
2
3
4
5
6
7
8
9
10
Demographics Age (years)
66
67
57
53
61
53
70
79
69
79
Gender
F
F
F
M
F
M
F
M
M
M
Comorbidities Hypertension
?
–
–
?
–
–
–
–
?
–
Dyslipidemia
–
–
?
–
–
–
?
–
?
–
Diabetes
?
?
?
?
–
–
–
–
–
–
Renal failure Ischaemic heart disease
– –
– –
– –
– –
– –
– –
– –
– –
– –
– –
ASA score
II
I
I
II
I
I
II
III
IV
III
Fever
?
?
?
?
?
–
–
?
?
?
Abdominal pain
?
–
–
–
?
–
?
–
–
–
Jaundice
?
–
?
–
–
–
–
–
–
–
Constitutional symptoms
?
–
?
?
?
?
?
?
?
?
Symptoms
Investigation findings Bilirubin (lmol/L)
80
21
74
10
60
35
79
94
26
94
GGT (IU/L)
381
31
335
14
24
90
980
224
245
224
Gallstones
–
–
–
–
–
–
–
–
?
–
Abscess characteristics Solitary
–
?
?
?
–
?
?
?
?
?
Size (cm)
6.5
10
8.4
3
7.7
9.7
5.8
5.2
1.4
5.2
KP
KP
KP
KP
KP
KP
KP
KP
EC
EC
Organism Treatment and outcome Percutaneous drainage
–
?
–
–
–
?
–
?
–
?
Length of stay (days)
9
28
2
3
2
10
15
98
14
98
ASA American Society of Anaesthesiology, GGT Gamma-glutamyl transpeptidase, KP Klebsiella pneumoniae, EC Escherichia coli
Fig. 3 Adversaries attributed to Escherichia coli pyogenic liver abscess
123
World J Surg
need for PD and surgical drainage [24]. KPPLA has a tendency to form larger abscess compared to other bacteria [29]. KPPLA size of more than 5 cm is reported in 71 % (104/147) by Chen et al. [21]. The larger mean abscess size of KPPLA patients could be a possible contributory factor for the higher proportion of PD observed in our study. Multiple liver abscesses could suggest increased virulence and more severe nature of illness with a potential for higher mortality [10]. Our study showed that majority of PLA was solitary regardless of the causative pathogen and majority of patients who died of multi-organ failure also had a solitary PLA. We reserve surgical drainage for patients who have ruptured abscesses or who demonstrate failure of non-operative therapy. Although there have been local studies that advocate the use of surgical drainage for PLA [5 cm [24] and laparoscopic drainage in selected patients [30] with reportedly better clinical outcomes than PD, none of our patients required surgical drainage. Claims favouring routine surgical drainage based on large size needs to be verified by prospective randomised trials. Average hospital stay of patients with PLA tends to be high due to need for intravenous antibiotics and PD. In the endeavour to reduce overall length of stay, outpatient parenteral antibiotic therapy (OPAT) is preferred. This is only feasible in selected clinically stable patients with appropriate antibiotic sensitivity and in healthcare systems where such a facility is available. We do have an active and integrated multidisciplinary management protocol and framework in place for OPAT. Despite the presence of such framework, median length of stay of our patients is 2 weeks which indicates that PLA is a drain to healthcare resources. Mortality from PLA is globally reported as 1–31 % [2– 4, 6, 7, 9–11]. It could be speculated that as ECPLA tend to be small in size, they are less likely to undergo PD and this could lead to poor outcomes. However, ECPLA patients treated with antibiotics and PD paradoxically showed a trend towards increased mortality suggesting that more sick patients had a higher mortality despite PD. Hence it appears that systemic antibiotic treatment is more important in ECPLA compared to local therapy with PD. Chen et al. [21] reasoned that the higher mortality of ECPLA patients in his series could be contributed by higher Acute Physiology and Chronic Health Evaluation (APACHE II) score at admission. In our study, we examined baseline American Society of Anaesthesiology (ASA) score and there was no statistical difference. It is possible that proactive upfront PD could reduce the progression or worsening of systemic sepsis and multi-organ failure while a salvage PD for failure of ‘antibiotic only’ therapy has limited role in already septic patient. Due to retrospective nature of our study, we are unable to determine to timing of PD with regards to clinical condition and hence did not define a failure of ‘antibiotic only’ therapy.
Figure 2 shows causes of mortality in patients with PLA and it is apparent that one-third of all the mortality is attributed to multi-organ failure. Majority of patients who succumbed to multi-organ failure had PLA [5 cm (Table 3). It remains to be determined if early elective surgical drainage in selected patients could reduce this mortality. Treatment of PLA requires a quartet of early diagnosis, prompt resuscitation, aggressive proactive policy of PD and integrated multidisciplinary management protocols. In a local study on outpatient management of KPPLA involving our institution, Chan et al. [25] reported zero mortality in his series, suggesting that the prognosis of this group of patients remain favourable if they recover from acute hepatobiliary sepsis in the hospital and continue with OPAT. The strength of our study lies in large sample which is possible due to higher prevalence of PLA in local population. The main limitation of our study is its retrospective nature with potential recall bias. It is also possible that some patients with polymicrobial infection which have been categorised as ECPLA or KPPLA based on their presence in the culture, the outcomes were influenced by presence of other bacteria. The increased mortality of ECPLA patients with PD compared to antibiotics alone appears paradoxical and calls for an international collaboration to standardise terminology with regards to ‘proactive’ PD and ‘salvage’ PD. Only when such distinction in terminology is made and timing of PD is known, it could be proven if systemic antibiotic therapy plays a dominant role in ECPLA management and PD plays a dominant role in KPPLA management. In conclusion, there are differences in patient demography and clinical profile of PLA patients with these two different pathogens. Despite such differences, nonoperative management of PLA appears to be highly successful regardless of aetiology. The higher mortality of ECPLA reported by others is not observed in our series. While, it is likely that such good outcomes are secondary to multimodal care, such conclusions can rightfully be derived only by future multicentre collaborative studies. Conflict of interest The authors do not report any conflicts of interest.
References 1. Kaplan GG, Gregson DB, Laupland KB (2004) Population-based study of the epidemiology of and the risk factors for pyogenic liver abscess. Clin Gastroenterol Hepatol 2:1032–1038 2. Jepsen P, Vilstrup H, Schonheyder HC et al (2005) A nationwide study of the incidence and 30-day mortality rate of pyogenic liver abscess in Denmark, 1977–2002. Aliment Pharmacol Ther 21:1185–1188 3. Hansen PS, Schonheyder HC (1998) Pyogenic hepatic abscess. A 10-year population based retrospective study. APMIS 106: 396–402
123
World J Surg 4. Meddings L, Myers RP, Hubbard J et al (2010) A populationbased study of pyogenic liver abscesses in the United States: incidence, mortality and temporal trends. Am J Gastroenterol 105:117–124 5. Branum GD, Tyson GS, Branun MA et al (1990) Hepatic abscess: changes in etiology, diagnosis, and management. Ann Surg 212:655–662 6. Seeto RK, Rockey DC (1996) Pyogenic liver abscess: changes in etiology, management, and outcome. Medicine 75:99–113 7. Chu KM, Fan ST, Lai ECS et al (1996) Pyogenic liver abscess: an audit of experience over the past decade. Arch Surg 131:148–152 8. Yu SC, Lo RH, Kan PS et al (1997) Pyogenic liver abscess: treatment with needle aspiration. Clin Radiol 52:912–916 9. Gyorffy EJ, Frey CF, Silva J Jr et al (1987) Pyogenic liver abscess: diagnostic and therapeutic strategies. Ann Surg 206:699–705 10. Huang CJ, Pitt HA, Lipsett P et al (1996) Pyogenic hepatic abscess: changing trends over 42 years. Ann Surg 223:600–609 11. Li J, Fu Y, Wang JY et al (2010) Early diagnosis and therapeutic choice of Klebsiella pneumoniae liver abscess. Front Med China 4(3):308–316 12. Cerwenka H (2010) Pyogenic liver abscess: differences in aetiology and treatment in Southeast Asia and Central Europe. World J Gastroenterol 16(20):2458–2462 13. Rahimian J, Wilson T, Oram V et al (2004) Pyogenic liver abscess: recent trends in etiology and mortality. Clin Infect Dis 39:1654–1659 14. Lederman ER, Crum NF (2005) Pyogenic liver abscess with a focus on Klebsiella pneumoniae as a primary pathogen: an emerging disease with unique clinical characteristics. Am J Gastroenterol 100:322–331 15. Liu Y, Wang JY, Jiang W (2013) An increasing prominent disease of Klebsiella pneumonia liver abscess: etiology, diagnosis and treatment. Gastroenterol Research and Practice 1–12 16. Chung DR, Lee SS, Lee HR et al (2007) Emerging invasive liver abscess caused by K1 serotype Klebsiella pneumonia in Korea. J Inf 54:578–583 17. Chan KS, Yu WL, Tsai CL et al (2007) Pyogenic liver abscess caused by Klebsiella pneumonia: analysis of the clinical characteristics and outcomes of 84 patients. Chin Med J 120(2):136–139
123
18. Chen SC, Yen CH, Lai KC et al (2005) Pyogenic liver abscesses with Escherichia coli: etiology, clinical course, outcome, and prognostic factors. Wien Klin Wochenschr 117(23–24):809–815 19. Pastagia M, Arumugam V (2008) Klebsiella pneumoniae liver abscesses in a public hospital in Queens, New York. Travel Med Infect Dis 6:228–233 20. Wang JH, Liu YC, Lee SSJ et al (1998) Primary liver abscess due to Klebsiella pneumoniae in Taiwan. Clin Infect Dis 26:1434–1438 21. Chen SC, Wu WY, Yeh CH et al (2007) Comparison of Escherichia coli and Klebsiella pneumoniae liver abscesses. Am J Med Sci 334(2):97–105 22. Durmishi Y, Gervaz P, Brandt D et al (2006) Results from percutaneous drainage of Hinchey stage II diverticulitis guided by computed tomography scan. Surg Endosc 20:1129–1133 23. Siewert B, Tye G, Kruskal J, Sosna J, Opelka F (2006) Impact of CT-guided drainage in the treatment of diverticular abscess: size matters. Am J Roentgenol 186:680–686 24. Tan YM, Chung AYF, Chow PKH et al (2005) An appraisal of surgical and percutaneous drainage for pyogenic liver abscesses larger than 5 cm. Ann Surg 241:485–490 25. Chan DS, Archuleta S, Llorin RM et al (2013) Standardized outpatient management of Klebsiella pneumoniae liver abscesses. Int J Infect Dis 17(3):185–188 26. Yang CC, Yen CH, Ho MW et al (2004) Comparison of pyogenic liver abscess caused by non-Klebsiella pneumoniae and Klebsiella pneumoniae. J Microbiol Immunol Infect 37:176–184 27. Chen SC, Yen CH, Tsao SM et al (2005) Comparison of pyogenic liver abscesses of biliary and cryptogenic origin: an eight-year analysis in a University Hospital. Swiss Med Wkly 135:344–351 28. Cohen JL, Martin FM, Rossi RL et al (1989) Liver abscess. The need for complete gastrointestinal evaluation. Arch Surg 124:561–564 29. Kim SB, Je BK, Lee KY et al (2007) Computed tomographic differences of pyogenic liver abscesses caused by Klebsiella pneumoniae and non-Klebsiella pneumoniae. J Comput Assist Tomogr 31:59–65 30. Tan L, Zhou HJ, Hartman M et al (2013) Laparoscopic drainage of cryptogenic liver abscess. Surg Endosc 27(9):3308–3314
