Acute liver failure (ALF) in children, although a rare condition, is associated with high mortality (1). Patients with ALF are prone to infections because they have impaired polymorphonuclear leukocyte function (2), impaired cell-mediated and humoral immunity (3), and diminished opsonic and complement activity (4). Some of the factors that predispose to infection are presence of indwelling catheters, H2-receptor blockers (5,6), steroid therapy (5,6), and broad-spectrum antibiotics. Infectious complications (IC) in adults are considered to be an important factor that increases mortality and morbidity; however, there are few data in the pediatric population. To our knowledge, this is the first study to describe IC in pediatric ALF and its effect on morbidity and mortality in a pediatric liver transplant center.
MATERIALS AND METHODS
The medical records of children with ALF admitted to King's College Hospital between 1995 and 2004 were reviewed. Data beyond 2005 were not included because of the center's participation in a multimember ALF trial. ALF was defined as hepatic international normalized ratio (INR) > 2 with evidence of significant hepatic dysfunction, in the absence of chronic liver disease (1). All of these patients had weekly surveillance cultures from significant sites (eg, blood, urine, bronchoalveolar lavages, drained abdominal collections) and more often if clinically indicated. Infectious complications (IC) were defined as the presence of clinical infection with proven organisms in cultures, positive pathogenic organisms by polymerase chain reaction (PCR) in sterile body fluids, or positive serology for bacteria and viruses requiring treatment. Standard definitions were used to describe types of infection (7–10). Bloodstream infection is defined as presence of pathogenic organisms in at least 1 positive blood culture obtained from a peripheral vein or catheter plus features of sepsis (eg, fever or hypothermia, and/or hypotension, raised white cell count, response to treatment) (8). If a different organism grew from a patient at any time after the initial positive blood culture had cleared, then it was considered to be a separate episode of bacteremia or fungemia. Catheter-associated bloodstream infection was referred to as positive organisms from central venous catheter without any additional clinically and microbiologically confirmed focus of infection, along with either positive culture of the same organism in the peripheral blood or presence of >15 colony forming units of the organism from the catheter tip in cases in which the catheter was removed (9). Lower respiratory tract infection (LRTI) was defined on the basis of increased production of purulent sputum, fever, an increased respiratory rate, raised white cell count, and radiographic changes (10). Urinary tract infection (UTI) was defined as presence of symptoms with or without pyuria and >105 colonies per milliliter of urine of a single type of organism in pure culture (7). All of the patients with ALF were commenced on antimicrobials when admitted to our unit, or in some patients as soon as diagnosis was confirmed at referring centers, where appropriate advice was given at first contact with our unit. Each patient received amoxicillin 20 to 30 mg/kg body weight intravenously, the frequency of doses depending on the age of the child, and cefuroxime 20 mg/kg body weight every 8 hours intravenously (choice of antibiotics was to cover endogenous flora and also was based on our local epidemiology of organisms in this group of patients). Amoxicillin added additional enterococcal cover, and an antifungal (fluconazole 3–6 mg/kg body weight) also was commenced for 14 days. Fluconazole was given at 3 mg/kg body weight earlier in the study; however, the practice was changed to 6 mg/kg body weight in the year 2000, after increasing candiduria was noticed in patients. High-dose acyclovir (20 mg/kg body weight every 8 hours intravenously) for herpesviruses was given to all neonates with ALF. Acyclovir was stopped once herpes simplex was ruled out as a cause of ALF.
Patients were broadly divided into those with infectious complications (IC group) and those without infectious complications (non-IC group). Comparison between non-IC and IC groups was done looking at age, hematological parameters (hemoglobin, platelets, and white cell count), INR and biochemical parameters (albumin, bilirubin, transaminases, gamma glutamate, ammonia, creatinine), mortality, duration of pediatric intensive care, ventilation, and overall hospital stay. Because mortality and surgical intervention directly affect the duration of hospital stay, patients in IC and non-IC groups were further divided into 4 subgroups: patients who recovered on medical treatment without liver transplant (subgroup 1), patients with liver transplant and recovered (subgroup 2), mortality without liver transplant (subgroup 3), and mortality after liver transplant (subgroup 4), and their duration of hospital stay was analyzed. For statistical analysis, GraphPad Prism version 5.00 for Windows (GraphPad Software, San Diego, CA) was used. Measures of central tendency for continuous variables were compared using the Mann-Whitney U test (nonparametric variables) or the unpaired t test (parametric variables) following normality testing. The log-rank (Mantel-Cox) test was used to compare survival curves between the IC and the non-IC group.
During the 10 years, 145 children (78 boys), median (range) age 4.22 years (1 day–16 years), were admitted with ALF. The median age of IC group was 25 (1–209) months and that of non-IC was 73 (1–205) months, which was statistically significant (P = 0.05). The underlying etiology of ALF was outlined in Table 1. During the study period, 37 (25%) patients had proven IC with a total of 52 infectious episodes. IC occurred in patients after a median (range) duration of 16 days (0–54 days) of admission.
The most common infection was bacteremia with 14 episodes observed in 13 patients (Table 2). Ten episodes of LRTI were seen in our patients. These LRTIs included pneumonia, pleural effusion, bronchitis, and bronchiolitis. Infections were predominantly bacterial (Table 2), with Pseudomonas species as the most common organism isolated. All of these patients were on assisted ventilation.
There were a total of 8 episodes of UTI; 4 were bacterial and other 4 were caused by Candida albicans. Other infections included gastroenteritis (4), intraabdominal collections (2), wound infections (5), line site infections (4), and miscellaneous (5). Gastroenteritis was seen in 4 patients: 3 had rotavirus, and 1 had colitis associated with Clostridium difficile toxin. Both intraabdominal infections were caused by Enterococcus faecium. Three of 5 wound infections were caused by Staphylococcus aureus; Pseudomonas aeruginosa and Klebsiella aerogenes accounted for the rest. Line site infections were caused due to S aureus, K aerogenes, Serratia species, and P aeruginosa. Conjunctivitis was caused by Moraxella catarrhalis.
Comparison between IC and non-IC groups for age, underlying diagnosis, hematological parameters (hemoglobin, platelets, and white blood cell count), INR, and biochemical parameters (albumin, bilirubin, transaminases, gamma glutamate, ammonia, creatinine) showed no significant difference, apart from bilirubin and aspartate aminotransferase (Table 3). Patients with IC had higher values of bilirubin in comparison with the non-IC group of ALF, whereas aspartate aminotransferase in the IC group was higher statistically than in the non-IC group. Median (range) duration of hospital stay in patients with IC was 38 (1–201) days compared with patients without IC (10.5 [1–74] days, P < 0.0001; Mann-Whitney U test; Fig. 1). The median duration of hospital stay in all 4 subcategories of patients with IC was prolonged significantly when compared with the non-IC group (Table 4). More than 88% (113/145) of patients with ALF, including 36 of 37 patients with IC, were admitted to the intensive care unit. The median duration of ventilation was significantly higher in the group with IC (11 days) than in the non-IC group (5 days, P < 0.008; Mann-Whitney U test).
Twenty-five patients received liver transplantation (LT) in the non-IC group, and 21 received LT in the IC group. There was no significant difference in age or sex between the 2 groups; however, the patients with LT and IC had a significantly longer duration of hospital stay (P = 0.03) and a significantly longer duration of ventilation (P = 0.01) as compared with the children without IC with LT (Table 3). In patients with LT, 3 and 5 patients died in the non-IC and IC groups, respectively. A total of 31 (21%) patients died, of which 11 (7%) were from the IC group and the rest from the non-IC group. The 31 patients who died were distributed as follows: 23 before LT and 8 after LT, 20 without IC (17 before LT, 3 after) and 11 with IC (6 before LT, 5 after) (Table 4). Some patients who died were not listed for LT because of underlying etiology of ALF: 9 patients with herpes simplex virus infections, 1 hypoxic ischemic damage, 3 with metabolic or mitochondrial disorders, and 1 with hemophagocytosis. Few other patients were listed for LT but did not undergo transplantation: 9 in the non-IC group (6 had spontaneous recovery, so they were removed from transplant list, and 3 died before LT) and 3 in the IC group (all spontaneously recovered).
Patients in the non-IC group died after a median of 4.5 (1–24) days of hospital admission, whereas death in the IC group was after 42 (1–75) days of admission, which was statistically significant P < 0.0001. The cause of death was attributed to infection and sepsis in 4 patients in the IC group. In the remaining patients with IC, causes were multifactorial, including multiorgan failure and graft failure, and they were culture negative at the time of death. The log-rank (Mantel-Cox) test failed to show any statistically significant difference in mortality between groups with and without IC (Fig. 2). Subgroup analysis with or without LT failed to show any significant difference in mortality between the 2 groups (Fig. 3).
IC are a major contribution to mortality in adults with ALF; however, etiology of liver failure is also a major determinant of outcome. Etiology of ALF in pediatric patients differs from that seen in adult patients (11), with certain diseases specific to each population. In adults, common causes of ALF were acetaminophen overdose (17%), idiosyncratic drug reactions (13%), and viral hepatitis A and B combined (12%) and indeterminate (17%) (11). In this study, in one-quarter of the cases, the etiology was indeterminate (26%) (Table 1). Unlike adults, there were no cases of hepatitis B or C causing ALF in children; all of the viral infections belonged to herpes simplex or enteroviruses. Disseminated herpes simplex viruses 1 and 2 caused ALF in 9 neonates and all of them died even before the option of LT could be considered (12).
Thirty-seven (25%) patients had culture-proven sepsis in the present study with a total of 52 episodes, in comparison with the incidence of 80% in previous adult studies (13). Younger patients with ALF had more IC. The most common infection in our series was bacteremia. Ten percent (14/145) of patients had bloodstream infection compared with a 35% (72/206) incidence of bacteremia in adult patients with ALF (14). Incidence of overall infections and bloodstream infection in pediatric patients was lower in comparison with adult patients with ALF despite both groups having had antimicrobial prophylaxis (13). We speculate that the lower incidence may have been caused by the use of prophylactic antibiotics for all of the patients with ALF in our unit; however, we can only compare our findings to historical data because no recent pediatric studies are available.
There were only 2 bacteremias, both with coagulase-negative staphylococci, which fulfilled the criteria of catheter-related bloodstream infection, which may be an underestimate because not always were paired peripheral and catheter blood cultures taken; also, line tips were only sent in cases in which the catheter was removed. Poor venous line accessibility in pediatric patients meant that first-line treatment in catheter infections was to rescue the central line and treat bloodstream infections by using antimicrobials intravenously and lock therapy.
The sites of infection in pediatric ALF were similar to those found in pediatric intensive care studies from the United States (15), with bloodstream infections being the most common site followed by UTI and LRTI. In pediatric ALF, however, the most common organisms were nonlactose fermenting organisms (eg, pseudomonads), Staphylococcus aureus, and Enterococci, whereas those in pediatric intensive care studies were coagulase-negative staphylococci (skin flora causing catheter-related bloodstream infections) (15). Enterococci and coliforms were reported to be the most common cause of bacteremia in adult patients with ALF (14).
LRTI was seen in 10 patients, Pseudomonas species being the most common organism isolated in 5 of them; all of these were ventilation-associated pneumonias. In critically ill patients, there is an increased risk of infections with environmental organisms such as Pseudomonas species, especially after prolonged hospitalization or multiorgan failure (16).
Eight patients had UTI and half of these were caused by C albicans. These findings mirror the data from other studies on patients from pediatric intensive care (11). A large number of samples were from urinary catheters, and although catheter colonization is commonly seen, patients were treated for UTI if they had features of sepsis (raised white cell response, fever despite broad-spectrum antibacterials). Appropriate doses of antifungals have been shown to reduce Candida UTI. Fluconazole was used at a dose of 3 mg/kg body weight in our unit until the year 2000, during which most (3 cases) Candida UTI cases were seen. The dose was subsequently (after the year 2000) increased to the highest prophylactic value of 6 mg/kg and thereafter only 1 case occurred. There were no cases of disseminated fungal infection; this is in direct contrast to adult studies with ALF, in which up to 32% of patients have been described as having fungal infection (13).
The present study highlights the timing of breakthrough infections and the common organisms causing them in pediatric ALF. The median duration of appearance of IC was 16 days, suggesting that 14-day prophylaxis is helpful in preventing early infections. Antifungal prophylaxis and the use of appropriate doses and choice of antifungals on clinical suspicion of infection are associated with a low incidence of invasive fungal infections.
IC in pediatric ALF is directly associated with prolonged hospital stay and increased duration of ventilation, but it is not associated with increased mortality. Recent adult studies show that bacteremias, which are a surrogate marker for sepsis, do not influence mortality in adult ALF (14). This is in contrast to previous adult ALF studies, in which infection is associated with increased mortality (12). There are no contemporary pediatric studies to compare our data.
Supportive management including prophylactic antimicrobials and availability of various modalities of LT has improved the survival of patients with ALF. Breakthrough infections in these patients pose an important problem because they could lead to prolonged hospital stay. IC is more common in younger age groups, and was associated with increased morbidity but with no significant effect on mortality. Standard antimicrobial prophylaxis for 14 days may have a protective effect in reducing infectious complications; however, a longer duration of antibiotics, chosen depending on local epidemiology of organisms, may be helpful in patients with more severe ALF, prolonged ventilation, or those who undergo LT. Appropriate doses of antifungal prophylaxis seem effective in preventing invasive fungal infections. Further randomized controlled trials will be helpful to better understand the risk factors for IC and the efficacy of antimicrobial prophylaxis.
1. Squires RH Jr, Schneider BL, Bucuvalas J, et al. Acute liver failure in children: the first 348 patients in the pediatric acute liver failure study group. J Pediatr 2006; 148:652–658.
2. Bailey RJ, Woolf IL, Cullens H, et al. Metabolic inhibition of polymorphonuclear leucocytes in fulminant hepatic failure. Lancet 1976; 1:1162–1163.
3. Fagan EA, Williams R. Fulminant viral hepatitis. Br Med Bull 1990; 46:462–480.
4. Larcher VF, Wyke RJ, Mowat AP, et al. Bacterial and fungal infection in children with fulminant hepatic failure: possible role of opsonisation and complement deficiency. Gut 1982; 23:1037–1043.
5. Schubert ML, Peura DA. Control of gastric acid secretion in health and disease. Gastroenterology 2008; 134:1842–1860.
6. Stoll BJ, Temprosa M, Tyson JE, et al. Dexamethasone therapy increases infection in very low birth weight infants. Pediatrics 1999; 104:e63.
7. Welsh A. Urinary tract infection in children. Diagnosis, treatment and long-term management. National Institute of Health and Clinical Excellence guidelines 54.2007. http://guidance.nice.org.uk/CG54
. Accessed August 15, 2010.
8. Weinstein MP, Towns ML, Quartey SM, et al. The clinical significance of positive blood cultures in the 1990s: a prospective comprehensive evaluation of the microbiology, epidemiology, and outcome of bacteremia and fungemia in adults. Clin Infect Dis 1997; 24:584–602.
9. Mermel LA, Farr BM, Sheretz RJ, et al. Guidelines for the management of intravascular catheter-related infections. Clin Infect Dis 2001; 32:1249–1272.
10. Korppi M. Physical signs in childhood pneumonia. Pediatr Infect Dis J 1995; 14:405–406.
11. Ostapowicxz G, Fontana R, Frank V, et al. Results of a prospective study of acute liver failure at 17 tertiary care centers in the United States. Ann Intern Med 2002; 137:947–954.
12. Verma A, Dhawan A, Zuckerman M, et al. Neonatal herpes simplex virus infection presenting as acute liver failure: prevalent role of herpes simplex virus type I. J Pediatr Gastroenterol Nutr 2006; 42:282–286.
13. Rolando N, Philpott-Howard J, Williams R. Bacterial and fungal infection in acute liver failure. Semin Liver Dis 1996; 16:389–402.
14. Karvellas CJ, Pink F, McPhail M, et al. Predictors of bacteraemia and mortality in patients with acute liver failure. Intensive Care Med 2009; 35:1390–1396.
15. Richards MJ, Edwards J, Culver DH, et al. Nosocomial infections in pediatric intensive care units in the United States. National Nosocomial Infections Surveillance System. Pediatrics 1999; 103:e39.
16. Park DR. Microbiology of ventilator associated pneumonia. Respiratory Care 2005;50:742–65.