Randomized control trials (RCTs) are very rare in pediatric hepatology. This specialty, which developed in earnest only in the last few decades, can be seen as “pharmacological orphans” because it frequently lacks age-specific trials, whereas in adult hepatology RCTs are achievable and commonplace. Moreover, children with liver problems are by default occasionally still managed by pediatric gastroenterologists or by adult hepatologists. The consequence is that the clinical treatment of many pediatric conditions is heavily influenced by experience from adult patients. In some instances, such as the treatment of chronic hepatitis C or autoimmune hepatitis, this was proven correct, whereas in other conditions, such as acute liver failure (ALF) or sclerosing cholangitis, the debate continues as to whether these disorders have different pathophysiology, immunopathology, or response to treatment in adults and children.
Squires et al.1 have made an important contribution to the discussion of one of these contentious topics‒the management of ALF in children. They have set an ambitious goal to test the widespread assumption, based on previously reported experience in children and adults2-4 that N-acetyl-cysteine (NAC) improves the outcome in ALF. NAC has a number of different pharmacological effects,5 predominantly based on an increased mitochondrial and cytosolic production of glutathione, which is one of the principal free radical-scavenging agents in the human body. The reported clinical applications are diverse and include the treatment of pulmonary fibrosis and chronic obstructive pulmonary disease, minimizing contrast-induced renal injury during imaging, and infertility management in women with clomiphene-resistant polycystic ovary syndrome.6 Oral NAC was reportedly effective as a mucolytic7 and in treatment of cannabis addiction.8 In acetaminophen toxicity glutathione repletion reduces the level of covalent binding of the toxic metabolite, making NAC an effective antidote.9
The rationale for the study by Squires et al. was based on the reported improved outcome of non-acetaminophen-related ALF in a retrospective noncontrolled pediatric study2 and in two prospective studies in adults,3, 4 where the beneficial effect of NAC was particularly evident in patients with less severe hepatic encephalopathy (HE).4 Based on these data, the tacit consensus among pediatric hepatologists over the years has been to use this inexpensive medication in ALF under the “primum non nocere” umbrella. When a pediatrician faces the dilemma of giving or not giving a medication with a neutral evidence of benefits, his/her decision is almost always to use it!!
It took almost a decade for the study by Squires et al. to recruit the required 184 children from 20 pediatric units in North America and Europe. Patients were randomized to receiving NAC (150 mg/kg/d) or placebo for up to 7 days, or until liver transplantation (LT) or death.1 To illustrate the difficulties in recruitment, it is worth emphasizing that 32% of the eligible 271 families did not consent to the randomization, and that five children died within 24 hours from admission, before the allocated treatment could be started.1 Additional treatment, including intensive care, anti-infectious measures, and consideration of LT remained under the direction of local pediatricians who used their own management protocols. Survival at 1 year was the primary outcome measure.
The study by Squires et al. concluded that NAC treatment does not influence the 1-year survival (73% in treated patients versus 82% in the placebo group). The authors were also unable to demonstrate differences between the treatment arms with regard to progression of HE, severity of multiorgan failure, and overall intensive care and hospital stay. Of note, however, at 1-year follow-up, only 35% of NAC-treated, but 53% of placebo-treated children did not require LT (P = 0.03). This trend was borderline significant in children younger than 2 years of age with grade 0-1 HE (NAC 25%, placebo 60%; P = 0.0493). The number of post-LT deaths was identical: five in both arms.1
This type of study is inevitably open to criticism. The definition of ALF was straightforward, including an international normalized prothrombin ratio (INR) >1.5 in the presence of encephalopathy, or an INR >2 in the absence of encephalopathy without preexisting liver disease. The multicenter nature of the study, however, is likely to have resulted in inconsistencies in the management and perhaps even in the LT criteria. The aggressive monitoring and supportive measures that some pediatric centers use, such as direct monitoring of carotid flow, cerebral perfusion and intracranial pressure, continuous hemofiltration and exchange transfusions are not universally applied and depend on the local facilities and expertise. The fact that more than 20 different diagnoses with diverse natural histories were enrolled could be easily targeted by the purists. In particular, the inclusion of conditions as diverse as biliary atresia, galactosemia, Niemann-Pick type C, or infantile hemangioendothelioma in a study of ALF is debatable. However, their numbers in this study were minimal and probably unlikely to have affected the overall results.
A critical aspect of this study was obtaining an accurate history of acetaminophen exposure. Acetaminophen is often given to children in the prodromal phase of ALF and it is difficult to ascertain its role, if any, in the progression of the disease. To complicate matters further, some genetic susceptibility and different acetaminophen pharmacokinetics in the presence of liver injury in children have been suggested.10 In the study by Squires et al. testing for acetaminophen-cysteine adducts (A-CA), which is often used as a marker of acetaminophen exposure and of hepatic injury (at concentration >1.1 nmol/mL),11 was performed only retrospectively. A-CA were reported positive in 9 out of 84 patients tested, six from the placebo and three from the treatment arm.1 Among these, cryptogenic ALF was diagnosed in six, and ALF related to viral hepatitis and sepsis in one each. The ninth patient eventually diagnosed as having acetaminophen-related ALF, after accurate review of his medication history, was removed from the study. Notably, there were no statistical differences in the demographic and clinical characteristics of children who were and were not tested for A-CA.1
Given the critical condition of children in ALF, it is reassuring that the side and adverse effects were few and randomly distributed in the two study arms. Arguably, some of the reported side effects, such as dilated pupils, aplastic anemia, and posttransplant lymphoproliferative disease, were the consequence of the original disease or immunosuppression, but were not in numbers significant enough to modify the overall safety record.
To explain the outcome differences between the present study1 and a comparable randomized adult study4 is not easy; one possibility is the higher prevalence of cryptogenic ALF in children as compared to adults (54% in the study by Squires et al. versus 24% in the RCT in Ref.4). This subgroup of ALF could represent a mixture of undiagnosed viral and metabolic disorders, possibly less responsive to NAC than other forms of ALF. The authors themselves speculate that some unrecognized immune dysregulation could play a more prominent role in the cryptogenic than in the other forms of ALF.1 Interestingly, NAC has been shown to interfere with the production of anti-tumor necrosis factor,12 one of the key proinflammatory cytokines produced in ALF.13 On the other hand, in mitochondrial cytopathies, which represent an emerging subgroup of ALF in children, often challenging to diagnose and sometimes still labeled cryptogenic,14 NAC theoretically could play a more prominent role. These difficult questions will need to be addressed in future studies.
The failure of NAC to modify the outcome of ALF in children may not be surprising for everyone. Many pediatric hepatologists will be aware of the limitations of the medical treatment in ALF and the desperate anticipation for the graft availability and LT while anxiously observing the fragile neurological condition of a child. The reported tie in this RCT might still indicate someone's victory; in these cost-conscious times the hospital managers and insurance companies could have the last laugh.