INTRODUCTION
Pyogenic liver abscess (PLA) is a potentially life-threatening condition, complicating diverse abdominal pathologies. The global reported incidence for PLA is variable, ranging from 3 to 25 per 100,000 pediatric hospital admissions (1-4). Because of exposure to a high number of environmental microorganisms, PLA is a significant cause of morbidity in children in the developing world. It has been suggested that the overall incidence of PLA in adults is increasing because of a wider availability of diagnostic procedures. The mortality rate, however, has improved because of the use of draining techniques, earlier identification of immunocompromised patients, and more effective antibiotic treatment (4). Although PLAs are thought to be rare in the developed world, their incidence is uncertain. They have been described with disorders of innate immunity (5,6), malignancies (7), Crohn disease (7,8), diabetes mellitus (9), and abdominal sepsis (10-17). The aim of this study was to identify the prevalence of PLAs in children referred over the last decade to a tertiary pediatric hepatology center and to review their clinical features, predisposing factors and management.
METHODS
King's College Hospital is the largest referral center for pediatric liver disease in the United Kingdom, receiving approximately 300 new referrals per year. We have examined the database of all children (<16 years of age) referred between 1993 and 2003 and reviewed retrospectively medical records of those with a proven diagnosis of PLA. Clinical features, demographic details, laboratory and radiologic information, treatment, and outcome were analyzed. All children underwent baseline blood cultures, full blood count and C-reactive protein (CRP) measurements. Radiologic assessment of the intrahepatic lesions and of portal vein or other intra-abdominal organ involvement was performed in all children using ultrasound scan (USS) and computed tomography (CT). Double-phase contrast CT scans were used for better definition of the lesions and also to obtain pelvic images to try to visualize the appendix. Whenever possible, ultrasound-guided aspiration of the hepatic lesion under sedation was attempted. All children were screened for disorders of the innate immune system by performing absolute lymphocyte and neutrophil counts, serum immunoglobulins, and nitroblue-tetrazolium (NBT) test. Patients diagnosed with portal vein thrombosis (PVT) underwent procoagulant screen including the determination of anti-thrombin III, protein C and S serum levels, factor V Leiden phenotype, and presence of anticardiolipin antibodies.
RESULTS
Demographics
Of 3,024 patients referred for investigation of liver disease in the study period, 15 (0.5%, 7 boys) had PLA. Their median age at presentation was 10 years (range 2 months-15 years) (Table 1). Four of the children had families who originated from the developing world, three from Pakistan, one from Africa. Two of the three Pakistani children were born in the United Kingdom and had no prior history of travel abroad. The third (patient 1) had been in Pakistan for 3 months just before her presentation. The African child (patient 14) was born and had lived in the United Kingdom all her life. All other children in the series had no relevant travel history.
TABLE 1: Clinical features of patients with pyogenic liver abscesses
Clinical Features
All 15 children presented with fever. Seven (47%) had abdominal pain. The median length of duration of symptoms at the time of referral was 14 (range 4-49) days.
Radiologic Evaluation
In all children, PLAs were identified initially on USS. The right lobe of the liver was affected in all (Fig. 1), with additional involvement of the left lobe in four patients. Nine (60%) children had multiloculated abscesses (Fig. 2). Four children (patients 4, 5, 12, 13) had evidence of PVT on USS or CT. Three of them (patients 4, 5, 12) also had ileocecal lymphadenopathy or abnormal bowel within the right iliac fossa on CT images, suggesting possible intra-abdominal infection. One child (patient 4) who was initially deemed too unwell to undergo aspiration under intravenous sedation, which is our standard protocol, developed signs of clinical peritonitis, and subsequently underwent laparotomy. One further child (patient 15) without ascending pylephlebitis had an appendix abscess diagnosed on CT (and confirmed at surgery), whereas another (patient 7) had multiple splenic abscesses in addition to the hepatic lesions visualized on CT.
FIG. 1: CT scan of the abdomen showing a right lobe hepatic abscess (patient 6).
FIG. 2: CT scan of the abdomen showing a multilocuted hepatic abscess (patient 14).
Laboratory Investigations
At the time of referral to our unit, the median white cell count was 23.7 × 109/L (range 5.32-34.2 × 109/L), with 14 (93%) patients having leukocytosis (defined as >11 × 109/L). All children had an elevated CRP (median 200.5 g/dL, range 25-309 g/dL; normal <10 g/dL) that normalized with treatment after a median period of 3 (range 2-12) weeks.
Aspiration of the abscess under radiologic guidance was attempted in all cases. In two of the patients with single abscesses (patients 2 and 15), therapeutic aspiration of 205 and 750 mL of pus, respectively, led to a dramatic reduction of the maximum temperature within 48 hours and a subsequent shortened course of antibiotics. The results of the microbial isolates are shown in Table 1. Microorganisms were cultured in 10 (67%) patients, despite all children being already on antibiotic treatment at the time of aspiration.
There were five patients in whom an insufficient amount of fluid was aspirated for culture; Propionibacterium was grown in the blood culture of one of them (patient 7), which was probably a contaminant. No patient had positive bacterial stool cultures or microscopy for Ascaris spp.
One of the patients with PVT (patient 4) had low levels of antithrombin III (55 μg/dL, normal range 80-130 μg/dL) and protein C (42 μg/dL, normal range 80-130 μg/dL), whereas the other three patients with PVT (patients 5, 12, and 13) had a normal procoagulant screen.
Treatment
All patients initially received broad-spectrum intravenous antibiotics (ampicillin and cephalosporins) and antifungals (fluconazole, itraconazole, or liposomal amphotericin). These were subsequently modified according to antibiotic sensitivities of the aspirate and clinical and laboratory response (Table 1). Metronidazole (15 mg/kg/d) was given to 13 of 15 children. All 15 children were given intravenous antibiotics until CRP was normal for a median of 8 (range 2-32) weeks.
Presence of Intra-Abdominal Sepsis
Five children had additional intra-abdominal pathology (Table 1). Three children (patients 4, 5, 12) had radiologic evidence of ileocecal lymphadenopathy or abnormal bowel within the right lower quadrant, suggestive of concomitant abdominal infection. The other two children were the boys with the splenic abscesses (patient 7) and histologically confirmed appendicitis (patient 15), respectively. Of all these patients, only two (patients 4 and 15) underwent exploratory laparotomy. One of them (patient 4) became acutely more unwell necessitating an urgent surgical intervention. Her presurgery CT scan suggested bowel perforation and showed thickened intestinal loops in the right lower quadrant. At laparotomy, there was abnormal bowel in the right lower quadrant with hematoperitoneum, but no intestinal perforation or appendicitis. Another child (patient 15), who had insulin-dependent diabetes mellitus (IDDM), had an urgent laparotomy and appendectomy after a CT scan suggested the presence of an appendiceal abscess, which was confirmed on surgery. A further 400 mL of pus was drained from his liver abscess during surgery (having had 350 mL drained earlier under CT guidance). One girl (patient 14) had already had an exploratory laparotomy at her local hospital before her referral to us. At laparotomy, an incidental ovarian cyst was removed, and there was no other evidence on intra-abdominal sepsis. One child (patient 7) in addition to the hepatic had also splenic abscesses.
Three of the four patients with PVT had radiologic evidence of abdominal sepsis, and both lobes of the liver were affected. The child with splenic abscesses also had both lobes of the liver affected. In all other children, only the right lobe was involved.
Associated Disorders
Three (20%, 1 girl) children were diagnosed as having chronic granulomatous disease (CGD) on NBT test. This was a new diagnosis in all; only one boy had a history of recurrent and prolonged skin infections. One child had been diagnosed with IDDM 5 years before, and another was known to be homozygous hemoglobin S sickle cell disease. None of these patients has been subsequently diagnosed with inflammatory bowel disease.
Portal Vein Thrombosis and Portal Hypertension
Three (patients 4, 5, 12) of the four children with PVT, diagnosed on either USS or CT scan, also had evidence of intra-abdominal sepsis. Three children (patients 4, 5, 13) developed splenomegaly on subsequent USS after a median period of 3 (range 2-9) months. Elective upper gastrointestinal endoscopy confirmed presence of grade I to II esophageal varices in two of them.
DISCUSSION
This single center study suggests that PLAs are rare in the developed world but are frequently associated with disorders of granulocyte function and ascending pylephlebitis complicating previous abdominal infection. PLAs are much more common in the developing world (18), presumably because of an increased environmental exposure to microorganisms in the food and water, often leading to chronic gastrointestinal infection. It has also been suggested that PLA can occur as a consequence of chronic malnutrition affecting defense mechanisms (19,20) and leading to a higher incidence of chronic gastrointestinal infections (21,22). In a study in a predominantly Melanesian population in New Caledonia, 39% of the children affected were below 1 SD for their weight (20), whereas only 2 of 15 of our patients were below the 25th percentile (1 SD) for their weight (Table 1). Ascaris infestation has been implicated in the etiology of liver abscesses in some reports from the developing world (23,24); none of our patients had Ascaris spp. detected in their stool or aspirate. In malnourished children with chronic gastrointestinal infections, the protective mechanisms of macrophages within the hepatic reticuloendothelial system could potentially be overcome even in the absence of immunodeficiency. Finally, overcrowding, poor sanitation, and poor healthcare access may contribute to the different incidence of PLAs in the developing and developed world.
Similar to other series from both developing and developed world (7,25-27), there was no male predominance in this series. However, a pediatric meta-analysis (4) and a recent population-based study from Canada (1) noted an unexplained higher incidence of PLAs in males. In our series, the median age at presentation was 10 years, similar to previously reported pediatric series from Taiwan (26), where 13 of 15 children were older than 8 years, and Texas (4), where the mean age at the diagnosis in 11 cases was 7.5 years. A large series from Brazil including 65 children reported a median age at diagnosis of 7.5 years (18), suggesting a trend for children from the developing world to present earlier, perhaps in relation to the exposure to chronic infection.
Akin to other series (3,4,28-30), the clinical features we have observed were nonspecific, including malaise, fever, abdominal pain, and, rarely, hepatomegaly. An early ultrasound assessment of any child with abdominal symptoms and raised inflammatory markers is therefore indicated. The clinical response to antibiotic treatment may not be prompt because of the time required to achieve adequate liver tissue concentrations, but gradual improvement in the inflammatory markers should be observed within the first 2 to 3 weeks of treatment.
Because of few complications and useful microbiologic information, percutaneous aspiration and drainage has been advocated as the first line of treatment in the management of PLA (31-33). Recently, a large prospective randomized study has demonstrated a superiority of intermittent percutaneous drainage over continuous catheter drainage in adult patients (27). Moore et al. (34), looking at 124 children in South Africa with liver abscesses, demonstrated 37% healing after conservative management without percutaneous aspiration or drainage. They allow the possibility that liver abscess that have been reported to resolve with drainage in several studies could have improved without intervention, albeit much slower (35). The aspirations are often diagnostic, guiding appropriate antibiotic therapy. In our series, even the four children in whom aspiration was unsuccessful went on to improve with prolonged antibiotic therapy. All PLAs in our series involved the right lobe, making percutaneous aspiration feasible and safe. Therapeutic mechanical removal of the pus led to a swift improvement of the symptoms in two of the patients (patients 2 and 15) in our series and thus a shorter duration of antibiotic course. The relatively high incidence of failed aspirations we have observed may be related to our use of conscious sedation, which was preferred to general anesthesia in unwell children. However, in retrospect, given the good clinical response to successful therapeutic aspirations in this series, a formal surgical drainage under general anesthesia may be justified. In contrast with PLA aspiration, in which the number of positive cultures was high even after antibiotic therapy, the information obtained from blood cultures was much less helpful. We would, therefore, recommend a routine attempted radiologic aspiration of the abscess, if this is considered safe.
In adult patients with PLAs, the commonest isolates are Gram-negative bacteria, such as E. coli, Klebsiella and Pseudomonas, anaerobes, and streptococci (7). Catalase-positive Staphylococcus aureus is one of the most common etiologic agents of PLA in the pediatric population (3,4,18,28). In our series, there were three children with Staphylococcus aureus cultured from aspiration of the hepatic lesion, of whom two were diagnosed with CGD. Bacteroides fragilis and other anaerobes (found in 2 of our patients) colonize the gastrointestinal tract and may easily be translocated from their normal habitat. They are common isolates in intra-abdominal and rectal abscesses (36). Consequently, antibiotics with antistaphylococcal and anaerobic activity should be used as initial therapy for children with PLAs. Furthermore, in patients with PLA, when neutrophil disorders are suspected, antifungal treatment should also be begun because children with CGD are also at risk of life-threatening fungal infections such as Aspergillus (5,6,19). Our experience indicates that CGD should be considered in both girls and boys to exclude the less common autosomal recessive form of the disease.
PLA is a well-recognized complication of IDDM, as seen in one of our patients (37-39). Previous microorganisms reported in IDDM-related PLA include Klebsiella endophthalmitis, in which the liver abscess may be a complication of the ocular involvement (40). There are numerous reports of PLAs occurring in individuals with SCD (41-44), Crohn disease (8), malignancies (1), and after organ transplantation (1), in which a lesser degree of impaired host defense probably facilitates their development.
Mortality rates for PLA have been reported to be as high as 36%, although they have decreased in more recent publications (3,4,45). For example, Tsai et al. (26) report no fatalities in their series on Taiwanese children. The absence of mortality in our series probably reflects the early therapeutic and diagnostic interventions, identification of immunocompromised patients, and targeted and prolonged intravenous antibiotic therapy. Moreover, we have had only one patient requiring open drainage, an intervention associated with increased morbidity because of its invasive nature and the risk of secondary infections (27). This patient also had a good outcome (patient 15).
We confirm previous reports that ascending pylephlebitis after intra-abdominal infection, such as appendicitis, can be the cause of PVT and PLA in some immunocompetent patients (46,47). Interestingly, one of our four children who developed PVT had an abnormal procoagulant profile. In the present series, USS was a useful modality to diagnose PLA and PVT, assess the clinical response, and monitor development of portal hypertension, whereas abdominal CT, including the pelvic region, was important to assess the presence of appendicitis or other signs of abdominal sepsis, such as abdominal lymphadenopathy, that could have initiated the pathologic process.
In conclusion, PLA is a rare disorder in children in the developed world but should be suspected and diagnosed promptly using appropriate investigations to achieve effective treatment and good prognosis. Disorders of the innate immunity, particularly neutrophil disorders, diabetes mellitus, and SCD, should be excluded in all cases, even when there is no suggestive history. In patients who develop PVT, procoagulant disorders should be excluded. Children with PLA are potentially at risk of developing PVT and portal hypertension and should be monitored for the development of this complication.
REFERENCES
1. Kaplan GG, Gregson DB, Laupland KB. Population-based study of the epidemiology of and the risk factors for pyogenic liver abscess.
Clin Gastroenterol Hepatol 2004;2:1032-8.
2. Kong MS, Lin JN. Pyogenic liver abscess in children.
J Formos Med Assoc 1994;93:45-50.
3. Chusid MJ. Pyogenic hepatic abscess in infancy and childhood.
Pediatrics 1978;62:554-9.
4. Pineiro-Carrero VM, Andres JM. Morbidity and mortality in children with pyogenic liver abscess.
Am J Dis Child 1989;143:1424-7.
5. Finn A, Hadzic N, Morgan G, et al. Prognosis of chronic granulomatous disease.
Arch Dis Child 1990;65:942-5.
6. Lublin M, Bartlett DL, Danforth DN, et al. Hepatic abscess in patients with chronic granulomatous disease.
Ann Surg 2002;235:383-91.
7. Mohsen AH, Green ST, Read RC, et al. Liver abscess in adults: ten years experience in a UK centre.
QJM 2002;95:797-802.
8. Narayanan S, Madda JP, Johny M, et al. Crohn's disease presenting as pyogenic liver abscess with review of previous case reports.
Am J Gastroenterol 1998;93:2607-9.
9. Gaskin DA, Bodonaik NC, Williams NP. Hepatic abscesses at the University Hospital of the West Indies. A 24-year autopsy review.
West Indian Med J 2003; 52:37-40.
10. Chang TN, Tang L, Keller K, et al. Pylephlebitis, portal-mesenteric thrombosis, and multiple liver abscesses owing to perforated appendicitis.
J Pediatr Surg 2001;36:E19.
11. Dieulafoy. Le foie appendiculaire: abscès du foie consécutifs à l'appendicite.
Semaine Med 1898;18:449.
12. Kader HA, Baldassano RN, Harty MP, et al. Ruptured retrocecal appendicitis in an adolescent presenting as portal-mesenteric thrombosis and pylephlebitis.
J Pediatr Gastroenterol Nutr 1998;27:584-8.
13. Ochsner A, De Bakey M, Murray S. Pyogenic abscess of the liver.
Am J Surg 1938;40:292-319.
14. Slovis TL, Haller JO, Cohen HL, et al. Complicated appendiceal inflammatory disease in children: pylephlebitis and liver abscess.
Radiology 1989;171:823-5.
15. Soro Y. Pylephlebitis and liver abscesses due to appendicitis.
J Int Coll Surg 1948;11:464-8.
16. van Spronsen FJ, de Langen ZJ, van Elburg RM, et al. Appendicitis in an eleven-year-old boy complicated by thrombosis of the portal and superior mesenteric veins.
Pediatr Infect Dis J 1996;15:910-2.
17. Chubachi A, Miura I, Ohshima A, et al. Risk factors for hepatosplenic abscesses in patients with acute leukemia receiving empiric azole treatment.
Am J Med Sci 1994;308:309-12.
18. Ferreira MA, Pereira FE, Musso C, et al. Pyogenic liver abscess in children: some observations in the Espirito Santo State, Brazil.
Arq Gastroenterol 1997;34:49-54.
19. Kumar A, Srinivasan S, Sharma AK. Pyogenic liver abscess in children: South Indian experiences.
J Pediatr Surg 1998;33:417-21.
20. Guittet V, Menager C, Missotte I, et al. Hepatic abscesses in childhood: retrospective study about 33 cases observed in New-Caledonia between 1985 and 2003.
Arch Pediatr 2004;11:1046-53.
21. Moreira-Silfa SF, Pereira FE. Brief report. Intestinal nematodes, Toxocara infection, and pyogenic liver abscess in children: a possible association.
J Trop Pediatr 2000;46:167-72.
22. Pereira FE, Musso C, Castelo JS. Pathology of pyogenic liver abscess in children.
Pediatr Dev Pathol 1999;2:537-43.
23. Rossi MA, Bisson FW. Fatal case of multiple liver abscesses caused by adult Ascaris lumbricoides.
Am J Trop Med Hyg 1983;32:523-5.
24. Reay HA, Maunder C. Liver abscess caused by adult
Ascaris lumbricoides.
Br Med J 1964;5408:553-4.
25. Chou FF, Sheen-Chen SM, Chen YS, et al. Single and multiple pyogenic liver abscesses: clinical course, etiology, and results of treatment.
World J Surg 1997;21:384-8.
26. Tsai CC, Chung JH, Ko SF, et al. Liver abscess in children: a single institutional experience in southern Taiwan.
Acta Paediatr Taiwan 2003;44:282-6.
27. Yu SC, Ho SS, Lau WY, et al. Treatment of pyogenic liver abscess: prospective randomized comparison of catheter drainage and needle aspiration.
Hepatology 2004;39:932-8.
28. Arya LS, Ghani R, Abdali S, et al. Pyogenic liver abscesses in children.
Clin Pediatr (Phila) 1982;21:89-93.
29. Dehner LP, Kissane JM. Pyogenic hepatic abscesses in infancy and childhood.
J Pediatr 1969;74:763-73.
30. Kaplan SL, Feigin RD. Experience and reason: briefly recorded.
Pediatrics 1976;58:614-6.
31. Attar B, Levendoglu H, Cuasay NS. CT-guided percutaneous aspiration and catheter drainage of pyogenic liver abscesses.
Am J Gastroenterol 1986;81:550-5.
32. Berger LA, Osborne DR. Treatment of pyogenic liver abscesses by percutaneous needle aspiration.
Lancet 1982;1:132-4.
33. Ch YS, Hg LR, Kan PS, et al. Pyogenic liver abscess: treatment with needle aspiration.
Clin Radiol 1997;52:912-6.
34. Moore SW, Millar AJ, Cywes S. Conservative initial treatment for liver abscesses in children.
Br J Surg 1994;81:872-4.
35. Wong KP. Percutaneous drainage of pyogenic liver abscesses.
World J Surg 1990;14:492-7.
36. Brook I. Microbiology of polymicrobial abscesses and implications for therapy.
J Antimicrob Chemother 2002;50:805-10.
37. Wong WM, Wong BC, Hui CK, et al. Pyogenic liver abscess: retrospective analysis of 80 cases over a 10-year period.
J Gastroenterol Hepatol 2002;17:1001-7.
38. Cheng HP, Chang FY, Fung CP, et al. Klebsiella pneumoniae liver abscess in Taiwan is not caused by a clonal spread strain.
J Microbiol Immunol Infect 2002;35:85-8.
39. Holt JM, Spry CJ. Solitary pyogenic liver abscess in patients with diabetes mellitus.
Lancet 1966;2:198-200.
40. Tan YM, Chee SP, Soo KC, et al. Ocular manifestations and complications of pyogenic liver abscess.
World J Surg 2004;28:38-42.
41. Shulman ST, Beem MO. An unique presentation of sickle cell disease: pyogenic hepatic abscess.
Pediatrics 1971;47:1019-22.
42. Lama M. Hepatic abscess in sickle cell anaemia: a rare manifestation.
Arch Dis Child 1993;69:242-3.
43. Chong SK, Dick MC, Howard ER, et al. Liver abscess as an unusual complication in sickle cell anemia.
J Pediatr Gastroenterol Nutr 1993;16:221-2.
44. Brittain HP, De la TA, Willey EN. A case of sickle cell disease with an abscess arising in an infarct of the liver.
Ann Intern Med 1966;65:560-3.
45. Hashimoto L, Hermann R, Grundfest-Broniatowski S. Pyogenic hepatic abscess: results of current management.
Am Surg 1995;61:407-11.
46. Chang TN, Tang L, Keller K, et al. Pylephlebitis, portal-mesenteric thrombosis, and multiple liver abscesses owing to perforated appendicitis.
J Pediatr Surg 2001;36:E19.
47. Vanamo K, Kiekara O. Pylephlebitis after appendicitis in a child.
J Pediatr Surg 2001;36:1574-6.