Efficacy and safety of i.v. sodium benzoate in urea cycle disorders: a multicentre retrospective study

The urea cycle is the single metabolic pathway that eliminates circulating ammonia, which is the nitrogen waste product from protein catabolism and a potent neurotoxin. It involves a series of biochemical steps in which nitrogen is removed from the blood and converted to urea, which is non-toxic and excreted in the urine. Urea cycle disorders (UCD) is a group of congenital disorders caused by enzyme or transporter deficiencies that impair the conversion of ammonia to urea. In UCDs, nitrogen removal is blocked, and nitrogen accumulates in the form of ammonia and other nitrogenous precursors of urea, causing acute episodes of hyperammonemia. The overall incidence of UCDs caused by enzyme deficiencies is approximately 1:8000 [1, 2] or newly 1:35,000 newborns [3]. UCDs include ornithine transcarbamylase deficiency (OTCD), the most common form; carbamoyl phosphate synthetase deficiency (CPSD); arginino-succinate synthetase deficiency (ASSD, citrullinemia type I); arginosuccinate lyase deficiency (ASLD, arginino-succinic aciduria); and more rarely, arginase deficiency (ARGD) and N-acetylglutamate synthetase deficiency (NAGSD). Hyperammonemia is also found in the hyperornithinemia-hyperammonemia-homocitrullinuria (HHH) syndrome caused by a mutation in the SLC25A15 gene, which encodes the mitochondrial ornithine and citrulline transporter [4], and in lysinuric protein intolerance (LPI) caused by inherited recessive mutations in the SLC7A7 gene, which encodes for the cationic amino-acids transporter subunit y + LAT1 located on the basolateral plasma membrane of epithelial cells and macrophages. Hyperammonemia can also be associated with other inherited metabolic diseases such as organic acidurias, mitochondrial fatty acid oxidation defects and Carbonic anhydrase VA (CAVA) deficiency, a new disease regulating carbamoyl phosphate synthetase 1, and with liver failure and certain cases of iatrogenic intoxication, e.g., those caused by sodium valproate [5].

If untreated, UCDs can cause increased levels of ammonia in the bloodstream (a normal level of plasma ammonium is < 100 μmol/L in neonates and < 50 μmol/L in children and adults), resulting in vomiting, lethargy, seizures, brain damage, coma and eventually, death. The first episode of hyperammoniemia may occur at any age, depending on the severity of the UC blockade. Neonatal coma is the most common presentation, with a risk of decompensation at any time throughout the lifespan, despite adequate medical and dietary treatment. Despite medical progress, UCDs remain at high risks of acute decompensation with neurological sequelae and vital risk [6], underlining the necessity of early diagnosis [7] and intensive consensual protocols to optimise neonatal but also long-term prognosis [2, 8‐10].

The current treatment of UCD consists of a low-protein diet associated with medications that provide alternative pathways for the removal of nitrogenous precursors of urea or ammonia from the bloodstream [11]. Depending on the enzyme deficiency, these treatments must be combined with daily L-arginine and/or citrulline supplementation [12].

In cases of acute decompensation, and particularly in emergency situations caused by infection, the major complication of hyperammonemia is cerebral oedema [2, 13]. First-line emergency treatment depends on the clinical status and whether the oral route is available for drug administration. Treatment aims to normalize the plasma ammonium level and restore the nitrogen balance (to ensure that input = loss), stop protein intake, administrate emergency regimen with high lipid and carbohydrate intake, re-establish the excretion of excess nitrogen via the administration of ammonium ion scavengers (orally or via i.v.), and provide appropriate symptomatic management (rehydration, correction of electrolyte imbalance). If necessary, extracorporeal toxin removal procedures should be considered [14] (see below). After recovery and for the long term, a low-protein diet and oral ammonium ion scavengers should be prescribed.

Two active substances have been identified as good candidates to promote the removal of nitrogenous residues (and mainly ammonium ions) from the blood by acting as an ammonium ion trap [15]: i) Sodium benzoate conjugates with glycine to form hippurate, which is eliminated by the kidneys (one mole of sodium benzoate eliminates one mole of ammonia); ii) Sodium phenylacetate conjugates with glutamine to form phenylacetylglutamine, which is eliminated by the kidneys (one mole of sodium phenylacetate eliminates 2 moles of ammonia) [16]. Sodium or glycerol (not yet marketed in France) phenylbutyrate (per os), precursors of phenylacetate, also eliminates 2 moles of ammonia per mole.

For the daily treatment and prevention of hyperammonemia, sodium benzoate per os can be used with or without sodium or glycerol phenylbutyrate.

In cases of acute hyperammonemia [4, 17], emergency treatment in France consists of i) Sodium benzoate at a loading dose by i.v. route [18]. This treatment, ‘sodium benzoate AP-HP 1 g/10 mL solution for infusion’, is available as a hospital preparation. It is produced by the Pharmaceutical Establishment of the Paris Hospitals (Assistance Publique Hopitaux de Paris). Sodium phenylacetate can be combined with sodium benzoate (an equimolar mixture of both products). The combination of these two active substances is available in France on a named-patient prescription basis (ATU) [19]. It must be administered via a central venous catheter, which carries an increased risk of complications (infection, thrombosis) and the need for anaesthesia [20]. In cases of severe hyperammonemia (>400 μmol/L), extracorporeal toxin removal procedures [21] should be considered.

Despite the availability in France of the above-mentioned ammonium scavengers and a relatively consensual protocol for the treatment of acute episodes of hyperammonemia [22], there are no data specifically evaluating the efficacy of sodium benzoate for the treatment of hyperammonemia. The aim of this study was to evaluate the risk/benefit of providing i.v. sodium benzoate treatment at the onset of the acute hyperammonemic episodes and during the follow-up of UCD. We retrospectively studied patient files from all of the French reference centres specializing in the treatment of metabolic diseases where this drug is the reference treatment for acute episodes of hyperammoniemia.

Sodium benzoate is an established and accepted adjunctive therapy for UCDs, as mentioned in the European guidelines [10], and for all causes of hyperammoniemia. Clinicians have used i.v.sodium benzoate routinely from the past 25 years to treat hyperammonemia in situations of acute decompensation (hyperammonemia) and during fasting, such as in cases of surgery or vomiting. Its efficacy and safety have been well known for years by metabolic physicians. Moreover, as previously reported (22), the most frequent cause of metabolic decompensation in our study was a catabolism episode, such as an infection. This underlines assertions that the availability of intravenous sodium benzoate is crucial for the management of UCD patients [23].

However, i.v. sodium benzoate has no marketing authorisation, and very little information on the use of this drug has been published. This retrospective study was established in 2011–2012 to consolidate the French physicians’ previous clinical experiences in the treatment of hyperammonemia in recent years. Data from 61 patients and 95 episodes of hyperammonemia were analysed during this study. These data showed that i.v. sodium benzoate decreased ammonia levels to ≤ 100 μmol/L by the end of treatment in most cases of late decompensation and that normal levels of NH3 were maintained in other situations where this treatment was prescribed to prevent decompensation. Also, this treatment is used for the prevention of decompensations and for the treatment of overt decompensations. Before treatment, the median ammonemia level was 246 μmol/L whatever the situations (overt decompensation or prevention of decompensation) (range: 20–2274 μmol/L, N = 82 episodes); it decreased to 40 μmol/L (range: 13–181 μmol/L, N = 77 episodes) among the patients who were still alive at the end of treatment with i.v. sodium benzoate. Deaths were caused by the underlying metabolic disease and occurred during the neonatal coma in 6 cases and during surgery/liver transplantation in 2 patients. Our study also confirmed that OTCD and ASSD are the most frequent UCDs (33 % and 30 %, respectively), while OTCD and CPSD are the most severe (representing all deaths in our study). It is well established that deaths occur mainly in severe neonatal coma, usually caused by OTCD and CPSD.

Because the first dose of i.v. sodium benzoate may be administered to patients both in the intensive care unit of the hospital than in other hospital departments, it is important for local hospitals to have i.v. sodium benzoate readily available before transferring a patient to a reference centre. Importantly, our work described the use of i.v. sodium benzoate in patients hospitalized in the French reference centers only. It is not representative of all episodes of hospitalizations and decompensations of the patients who firstly go to their nearby hospital, with telephone advices and emergency certificates from their reference center. Moreover initial comas are not managed only in reference centers as to date a lot of hospitals perform haemodialysis and acute medical treatments. All these considerations reinforce the necessity to have i.v. medications readily available in the whole national territory. The use of nitrogen scavengers is crucial, but of course associated with an intensive care management including rehydration, induction of anabolism with adequate amounts of calories, administration of i.v. arginine, carglumic acid for NAGS, CAVA and sometimes CPS1 deficiencies.

In our cohort, eighteen side effects were related to the infusion (extravasation of the venous catheter, local oedema, inflammation or abscess, haematoma). They were probably under-estimated in this retrospective study, but importantly no severe side effects were attributed specifically to i.v. sodium benzoate. Serious complications like comas, death and liver failure have been attributed to the underlying UCD and its decompensations.

A loading dose of i.v. sodium benzoate was given for 43 % of episodes, and the dose was variable. The fact that there were no guidelines available for most of the study period probably contributed to the lack of consistency in the administration and reporting of loading doses. Maintenance doses were given via peripheral i.v. infusion in most cases; central i.v. infusion was used for only one episode. The median maximum dose was 246.1 mg/kg/day, in accordance with the guidelines [10].

The total duration of i.v. sodium benzoate treatment per episode was a median of 2 days [0.1 –13], patients stayed in the intensive care unit for a median of 4 days, and the total length of stay in hospital was a median of 10 days. This point is important to note: not only the peak of plasma ammonium level, but also the duration of hyperammoniemia until normalization, have an impact on the outcome. More precisely, ammonemia was still > 100 μmol/L at the end of i.v. treatment for only 5 episodes. However, the 5 patients were successfully discharged from hospital (4 with ammonemia ≤ 100 μmol/L at discharge, the results were missing for one episode).

During 10 episodes, patients were treated with the combination of sodium phenylacetate and sodium benzoate (administered via central i.v. infusion) in association with i.v. sodium benzoate. This underlines the interest in i.v. sodium benzoate, which only requires peripheral vein access. Haemofiltration was required for 16 episodes, mostly in cases of neonatal coma for which i.v. sodium benzoate was not sufficient.

In our study, i.v. benzoate was switched to the oral form of sodium benzoate (sodium benzoate p.o., 78.3 %, 65/83 recorded episodes).

Whilst data were being collected for this French study, Häberle et al. published guidelines with the aim of harmonising the diagnosis and management of UCD in Europe [10]. These guidelines pool the results of and experience with all UCD treatments from several European countries, with most of the evidence based on expert opinion or case studies because of a lack of formal clinical trial data in Europe. The guidelines state that the ammonia scavengers sodium benzoate and sodium phenylacetate (one of the 2 active ingredients in the combination of sodium phenylacetate and sodium benzoate, along with sodium benzoate) are among the main treatments for acute hyperammonemia. The dose of i.v. sodium benzoate quoted in the guidelines is the same as that suggested in the product information sheet provided by AGEPS, AP-HP, which summarizes the doses of sodium benzoate AP-HP 1 g/10 mL to be used in clinical practice in France.

The experiences of the French clinicians indicated that i.v. sodium benzoate has advantages over the other treatments mentioned above. First of all, it can be administered during a coma or in cases of gastroenteritis, whereas the p.o. drugs containing sodium or glycerol phenylbutyrate cannot be used or are not well absorbed; and secondly, it is administered via a peripheral venous catheter, which is thought to be less risky [16, 20] and more practical than administration via a central venous catheter, as is the required for the combination of sodium phenylacetate and sodium benzoate. Another advantage concerns the availability of the products: i.v. sodium benzoate is available in France with a hospital preparation status, whereas the combination of sodium phenylacetate and sodium benzoate has to be imported on a named-patient basis from the USA (it has temporary use agreement —ATU - status in France). These advantages make i.v. sodium benzoate an essential tool in the treatment of hyperammoniemia, and all hospitals should own this treatment, particularly to treat the patient during the transfer to a specialized hospital. Importantly, secondary care centres and local hospitals may need to treat hyperammonemic decompensations related to UCD, as reference centres may have difficulty managing all UCD patients.

The 10-year retrospective study from 2000 until 2010 in the 6 French reference centrers for metabolic diseases shows that i.v. sodium benzoate treatment is effective for the treatment and/or the prevention of hyperammonemia in UCD. This study also confirms the safety of this treatment. I.v. sodium benzoate is the only peripheral treatment recommended and used in France in acute episodes of UCD, associated with an emergency regimen. Consequently, this precious i.v. treatment should be readily available in emergency in all hospitals.