Article Text

Developing a new formulation of sodium phenylbutyrate
  1. Nathalie Guffon1,
  2. Yves Kibleur2,
  3. William Copalu3,
  4. C Tissen4,
  5. Joerg Breitkreutz4
  1. 1Hôpital Femme-Mère-Enfant, Centre de Référence des maladies héréditaires du métabolisme, Lyon, France
  2. 2Lucane Pharma, Paris, France
  3. 3Parexel International, London, UK
  4. 4Institute of Pharmaceutics and Biopharmaceutics, Heinrich-Heine-University Düsseldorf, Düsseldorf, Germany
  1. Correspondence to Dr Yves Kibleur, Lucane Pharma, 172 rue de Charonne, Paris, France; ykibleur{at}


Background Sodium phenylbutyrate (NaPB) is used as a treatment for urea cycle disorders (UCD). However, the available, licensed granule form has an extremely bad taste, which can compromise compliance and metabolic control.

Objectives A new, taste-masked, coated-granule formulation (Luc 01) under development was characterised for its in vitro taste characteristics, dissolution profiles and bioequivalence compared with the commercial product. Taste, safety and tolerability were also compared in healthy adult volunteers.

Results The in vitro taste profile of NaPB indicated a highly salty and bitter tasting molecule, but Luc 01 released NaPB only after a lag time of ∼10 s followed by a slow release over a few minutes. In contrast, the licensed granules released NaPB immediately. The pharmacokinetic study demonstrated the bioequivalence of a single 5 g dose of the two products in 13 healthy adult volunteers. No statistical difference was seen either for maximal plasma concentration (Cmax) or for area under the plasma concentration–time curve (AUC). CI for Cmax and AUC0–inf of NaPB were included in the bioequivalence range of 0.80–1.25. One withdrawal for vomiting and five reports of loss of taste perception (ageusia) were related to the licensed product. Acceptability, bitterness and saltiness assessed immediately after administration indicated a significant preference for Luc 01 (p<0.01), confirming the results of the taste prediction derived from in vitro measurements.

Conclusions In vitro dissolution, in vitro and in vivo taste profiles support the view that the newly developed granules can be swallowed before release of the bitter active substance, thus avoiding stimulation of taste receptors. Moreover, Luc 01 was shown to be bioequivalent to the licensed product. The availability of a taste-masked form should improve compliance which is critical to the efficacy of NaPB treatment in patients with UCD.

  • Therapeutics
  • Paediatric Practice
  • Metabolic

Statistics from

What is already known on this topic

  • Sodium phenylbutyrate (NaPB) is notoriously bitter, leading to compliance difficulty and treatment rejection in chronic treatment.

  • NaPB is a life-saving drug in the management of urea cycle disorders, a rare disease.

What this study adds

  • A new coated formulation of NaPB has been developed with taste-masking properties.

  • This new formulation is bioequivalent to the licensed product.


Urea cycle disorders (UCD) are inherited deficiencies of one of the six enzymes involved in the metabolic pathway for amino acids. When the cycle is not fully functioning nitrogen accumulates and raised plasma levels of ammonia and glutamine may cause acute and/or chronic cerebral diseases with several debilitating and life-threatening clinical consequences.1

Alternative pathways for nitrogen excretion were described nearly a century ago2 ,3 and the successful use of phenylacetic acid (PA) in UCD was demonstrated by Brusilow et al in 1980.4 PA is acetylated to phenylacetylglutamine, incorporating two moles of nitrogen/mole, which is freely excreted through the kidneys.

PA has an extremely unpleasant odour and was eventually replaced by its pro-drug phenylbutyrate (PB), which undergoes rapid and complete hepatic conversion to PA in vivo. Two formulations of sodium phenylbutyrate (NaPB) (both uncoated 500 mg tablets and 940 mg/g granules) are available in Europe for the treatment of UCD. The granules are usually prescribed for infants and children and also in adults when a large dose is required. However, a notoriously bad taste and taste aversion remain among the most common adverse events (AEs) reported with NaPB.5

It has been widely demonstrated that the taste of medicines can influence compliance with treatment—in particular, in chronic paediatric diseases.6–9 Bitter compounds stimulate rejection reflexes such as nausea and vomiting, and, furthermore, children are more sensitive to bitterness than adults.10 In UCD, failure to comply with NaPB or learning to reject food associated with the bad taste of the co-administered drug are both counterproductive, could compromise adequate control of plasma ammonia and may result in life-threatening situations. Improving compliance will reduce the burden associated with chronic treatment of this rare disorder.11

The taste of NaPB has continuously been identified as a problem resulting in adequate compliance112 and several clinical studies and case narratives report a clear link between poor compliance and the taste problem using NaPB.10 ,12–16 Additionally, in the exploratory treatment of refractory solid tumour malignancies with oral NaPB, compliance was shown to be inversely proportional to dose up to 45 g NaPB/day.17 Furthermore, gastrointestinal AEs, potentially reflecting rejection of the drug, were also frequently reported with existing NaPB formulations.13 ,18 19 In an effort to improve compliance, extemporaneously prepared sweetened suspensions of NaPB granules have been used,20 although the taste cannot be easily disguised. A dietician's support may be required.12 In patients who cannot tolerate the taste, the drug has to be administered by a nasogastric tube or gastrostomy.1

The new formulation (Luc 01) was developed to improve adherence and compliance and hence effectiveness of treatment. The objectives were to mask the NaPB taste as completely as possible while maintaining the therapeutic equivalence to the reference product. Taste masking of solid drug formulations with immediate release properties such as the reference NaPB powder is difficult to accomplish without changing the pharmacokinetic profiles, which is mandatory for the development of a generic product. After experimentation with several different formulations, this was achieved by application of the active substance onto microgranular sugar cores followed by coating with a taste-masking solution. The final coating, although water soluble, prevents the immediate release of NaPB in the mouth but still enables rapid drug dissolution in the gastrointestinal tract. The preclinical development of Luc 01 was therefore driven and controlled by laboratory testing of the new formulation using a taste sensing system, often called an ‘electronic tongue’,21 and comparative dissolution studies using a fibre-optic setup for ultra-rapid dissolution detection.22 A pharmacokinetic study in adult healthy volunteers was carried out to determine if the new NaPB formulation is bioequivalent to the licensed product. Taste attributes of both products were characterised by the volunteers and the results should be compared with the predictions from the preclinical experiments.

Materials and methods

Dissolution testing

Comparative in vitro dissolution tests were performed using a fibre-optic setup for ultra-rapid dissolution detection22 to mimic the NaPB dissolution within the oral cavity. Real-time dissolution studies were carried out with a Sotax AT-6 dissolution tester (Hopkinton, MA, USA) in pharmacopoeial setup at a rotation speed of 100 rpm. All experiments were carried out in triplicate and arithmetic means determined. Direct spectroscopic UV measurements were made using a fibre-optic sensing system (Ocean Optics, Ostfildern, Germany) fully dipped into dissolution fluid consisting of purified water at 37±0.5°C. The spectrometer signal was sent to the computer and recorded every 0.2 s (with no time gap) by direct spectroscopic measurements of the light beam and residual after absorption by the sample.

In vitro taste studies (electronic tongue)

The taste profile of this new formulation was evaluated in vitro using a new taste sensing system, often referred to as an ‘electronic tongue’,21 and compared with the licensed granule formulation. The TS-5000Z taste sensing system (Insent, Atsugi-chi, Japan) was equipped with several sensors (TecLabS, Essen, Germany), each being associated with a unique prediction ability for specific taste characteristics. Five gustatory stimuli (some with multiple modalities) were tested: SB2CT0 for saltiness, SB2CA0 for sourness, SB2AAE for umami and aftertaste, SB2AE1 for astringency and aftertaste, SB2C00 for bitterness 3 (anionic) and aftertaste, SB2AN0 and SB2AC0 for bitterness 1 and 2 (cationic) and aftertaste. After dipping the probe for 90–120 s in three different cleaning solutions prepared according to the supplier's recommendations, a sensor check was performed in conditioning solution for 30 s. The sample was then measured over 30 s. After 3 s in two further cleaning solutions, aftertaste was measured over 30 s. Signals were obtained as electric potentials (in mV) from molecular adsorption to the sensor membranes. Longlasting adsorption to the sensors SB2C00, SB2AC0, SB2AN0, SB2AAE and SB2AE1 provided aftertaste signals.23 Sensor signals from the electronic tongue can be either positive or negative and the relationship between sensor response and drug concentration depends on the substance under investigation, therefore different concentrations of pure NaPB in water were measured. The observed log-linear relationship between drug concentration and sensor signals is in perfect agreement with the Nernst law which is the underlying principle for the electrochemically working sensors. Samples from dissolution testing were withdrawn after 2, 5 and 30 min for electronic tongue analysis. Each sample was measured in quadruplicate. The first result was excluded from calculations to avoid any unstable data and results provided represent arithmetic means of three consecutive measurements. The sensor data were analysed with Microsoft Excel 2007 and the multivariate statistics evaluations performed with SIMCA-P+ V.11.5.

Bioequivalence study


This was a randomised, open-label, two-treatment, two-period, two-sequence, two-way crossover bioequivalence study between the new formulation and the currently marketed formulation. A two-step sequential design was applied using adjusted statistics for interim analysis. The protocol was approved on 6 October 2010 by the Pharma-Ethics Independent Research Lyttelton Manor, Republic of South Africa, and all subjects provided their written informed consent. The study was conducted under the responsibility of Pretorius A, principal investigator (Kampuslaan Suid Campus of the University of the Free State, 9301 Bloemfontein, South Africa).

Subjects and study conduct

This study recruited healthy young male and female subjects. The study consisted of a screening visit up to 2 weeks before two treatment periods, separated by a 1-week washout and a follow-up visit 24 h after the last dose of the study drug. The treatment periods order was randomised with RANDPLAN (V.3.1), using the PROC PLAN procedure in SAS.

PB pharmacokinetics was assessed after study drug dosing in each period once fasted subjects had received a single 5 g dose as either a NaPB reference granules formulation or test granules formulation. Breakfast was supplied 2 h after drug administration. A standard lunch and dinner were served after the 4 h and 12 h blood sampling, respectively.

Subjects were resident in the study centre from the morning before study drug dosing (ie, day −1) until their discharge after the final pharmacokinetic sampling on day 1. Physical examinations, vital signs, 12-lead ECGs and laboratory safety tests were performed at intervals during each treatment period. AEs and concomitant drugs were recorded throughout the study.

The washout duration was regarded as sufficient as compared with the 0.8 h mean terminal half-life reported for NaPB in healthy adults.24 Sixteen 9 ml venous blood samples were collected before dosing (T0) and at the following times: 15 min, 30 min, 45 min, 1 h, 1.25 h, 1.5 h, 2 h, 2.5 h, 3 h, 3.5 h, 4 h, 6 h, 8 h, 12 h and 16 h after the drug intake. Samples were centrifuged at 2700 g at 0–8°C for 10 min immediately after collection. The plasma was stored in airtight stoppered polystyrene crystal tubes at −20°C until analytical determination.

Analytical determinations

PB plasma concentrations were determined using a sensitive, selective and fully validated liquid chromatography with tandem mass spectrometry (LC-MS/MS). The lower limit of quantitation for PB was 488 ng/ml.

Pharmacokinetic analysis

The pharmacokinetic parameters were calculated by non-compartmental analysis using WinNonlin Professional Software V.5.2 (2006). The terminal plasma half-life (t1/2) was calculated according to the following equation: Embedded Image

The area under the concentration–time curve from time zero (pre-dose) to time of last quantifiable concentration (AUC0–t) was calculated using a linear trapezoidal method. The AUC from time zero to infinite time (AUC0–∞) was calculated as follows:  Embedded Image where Ct is the last quantifiable concentration and λz is the terminal elimination rate constant.

Statistical analysis

The SAS package V.9.1 (2002–2003, SAS Institute Inc, Cary, North Carolina, USA) running on a personal computer was used to compare Cmax and AUCs by analysis of variance on the logarithmically transformed data with sequence, subject (sequence), product and period effects. A test for carryover effect was performed at the 5% level by a test of a sequence effect using the between-subject error term.

This study followed a two-stage sequential design and the results were tested as those from an interim analysis. Adjusted (adjustment of α level due to multiple testing) 94.12% CI for relative treatment differences were calculated.

Criteria for bioequivalence between formulations were based on Cmax, AUC0–t and AUC0–∞ exposure. The two formulations were considered bioequivalent if the ratios between geometric mean values and adjusted CIs for sodium 4-phenylbutyrate (NaPB) pharmacokinetic parameters all lay within the range 80–125%.25

A non-parametric Wilcoxon signed rank test was performed on the tmax.

For exploratory purposes, pharmacokinetic parameters were compared between male and female subjects for each product (data not shown).

Taste evaluation

The same healthy subjects were asked to complete a taste test questionnaire after NaPB dosing in order to provide subjective information on the acceptability, bitterness, saltiness and sweetness. Responses were compared between the two NaPB formulations at various times (immediately after, 0.5 h after and 2 h after ingestion) using an analysis of variance with sequence, product, period and product×period effects on untransformed data. A two-sided 5% significance level was used and point estimates and 95% CI for the ‘test minus reference’ mean differences on these variables at each time point were calculated and tabulated.


AEs and medical history were coded according to the Medical Dictionary for Regulatory Activities (MedDRA, V.13.0).


Drug dissolution

NaPB release properties of the two formulations showed very different drug dissolution profiles in a real-time experiment (figure 1). The licensed product released the drug completely within <60 s. With Luc 01, a lag time of 10 s was seen before release started followed by a slow progressive release of 60% of the drug over this 8 min experiment.

Figure 1

Sodium phenylbutyrate release from uncoated reference formulation granules (blue; reference) and coated test formulation granules (red).

In vitro taste assessment

For all sensors a log-linear relationship was seen between the intensity of the signal and the drug concentration in the relevant concentration range of 300–1000 mg/ml NaPB. Within this determined concentration window the electronic tongue sensors predicted that the free amount of NaPB substance in aqueous solution could be mainly characterised as bitter, sour, salty and astringent. The kinetic response of the taste sensors to Luc 01 indicated a modified drug release over time compared with the concentrations series of NaPB containing liquids (data not shown).

Clinical study results

Fourteen normal healthy Caucasian male (eight) and female (six, including one who dropped out) subjects, aged 19–50 years, weighing 51–82 kg, and 158–182 cm in height, were included after a prestudy medical examination, ECG, haematological and urine analysis in the 2 weeks before. The demographic characteristics of the study population are summarised in online supplementary table S1. Mean, SD and median (range) pharmacokinetic results are presented together with summarised statistical analysis in the following table 1.

Table 1

Pharmacokinetic parameters (geometric means, n=13) of sodium phenylbutyrate (NaPB) obtained for the test granules formulation Luc 01 and the reference granules formulation (reference sodium phenylbutyrate (NaPB))

Figure 2 represents the PB mean plasma concentrations over time for the two formulations. After administration of Luc 01, maximal plasma concentrations of PB ranged from 137 to 318 μg/ml, with tmax of 0.5–1.5 h. After ingestion of reference NaPB, maximal plasma concentrations of PB ranged from 146 to 326 μg/ml, with the same tmax range (p=0.5205). Plasma concentrations of PB declined with a mean terminal half-life of 0.39±0.07 h and 0.38±0.06 h after the test and reference formulation, respectively. Mean (SD) extrapolated AUC0–∞ were 466.920 (190.356) μg/ml.h and 448.220 (171.880) μg/ml.h after reference NaPB and Luc 01, respectively, with a point estimate (94.12% adjusted CI range) of 95.80% (90.80% to 101.08%). The point estimates (CI) of the ‘test/reference’ mean ratios of Cmax and AUC0–t for PB were 94.32% (86.95% to 102.31%) and 95.61% (90.34% to 101.19%), respectively. These CIs were within the bioequivalence acceptance range of 80–125%. In addition, no significant difference was seen for median tmax (Wilcoxon signed rank test, p value=0.5205). Therefore, the two products were bioequivalent with respect to the rate and extent of absorption of PB. Noteworthy, in accordance with previous data, gender differences in the pharmacokinetic parameters of PB with AUC and Cmax 30% greater in women were found in this study (data not shown).

Figure 2

Plasma concentration–time curve of phenylbutyrate concentrations after single oral dose of 5 g sodium phenylbutyrate (NaPB) as Luc 01 granules (test, triangles) or of the reference NaPB granule formulation (reference, circles) in 13 healthy male and female volunteers (plasma concentrations are expressed as geometric means).

Taste evaluation in healthy volunteers

Immediately after the dose the 95% CIs for the (test minus reference) difference in scores for acceptability, bitterness and saltiness were: (−56.79 to −11.71), (−62.20 to −33.65) and (−49.26 to −12.50), respectively, thus indicating statistical significant differences between the two products. The direction of the differences confirmed that the acceptability of the licensed product was significantly worse than that of Luc 01 (difference −34.25). The reference product was also perceived as significantly more bitter (difference −47.93) and salty (difference −30.88) than Luc 01.


Both formulations were well tolerated. Assessment of clinical biochemistry, vital signs, ECG, laboratory and physical examination did not show major changes from screening to the end of the study. The most commonly reported AE was a mild to moderate headache considered as related to treatment (15 occurrences out of 22 AEs). A loss of taste perception (ageusia) occurred in five cases, only after the reference drug and vomiting occurred in one subject also after the reference drug (and led to study drop out), one episode of body weakness occurred after the reference drug.


In the electronic tongue experiment, the in vitro taste profile of NaPB is that of a highly stimulant molecule, predicted to be bitter (cationic and anionic), salty, astringent, sour and umami tasting. These are all very uncomfortable attributes, especially for use in children. This pattern is rarely encountered as most other compounds show only signals on certain sensors, but not on all of them. Luc 01 granules provided similar results to the licensed granules, an expected result as they both contain the same active substance, but this occurs after a few minutes’ lag time as more clearly shown on the dissolution profiles. As the electronic tongue experiments need the withdrawal of a huge volume of dissolution fluid at each time point, they are difficult to conduct for multiple solid drug carriers with a short lag time such as in Luc 01. The dissolution experiments showed a lag time before release of the active substance NaPB with Luc 01. This delay should prevent interaction of NaPB with the taste receptors in the mouth, indicating typical taste-masking properties. This profile indicates that Luc 01 granules can be swallowed before releasing the bitter active substance. In contrast, the licensed product granules showed an immediate and complete release of the active substance.

In this study, pharmacokinetic parameters after administration of 5 g NaPB in healthy adults under fasting conditions with Luc 01 or the reference compound were similar to results reported after a single therapeutic dose of 5 g NaPB.24 Gender differences were also as already described and attributed to the lipophilicity of the drug and differences in volume of distribution or enzymatic reaction rates.

The taste evaluation conducted during the bioequivalence study immediately after drug intake confirmed in vitro findings of much less taste stimulation after Luc 01 than after the reference granules.

Among the subjects who had AEs, five healthy volunteers reported ageusia after the administration of the reference product as compared with none after Luc 01. Moreover one volunteer was withdrawn following severe vomiting soon after administration of the reference granules. While acknowledging the small number of subjects, the difference in reporting of ageusia is interesting as this is a recognised AE associated with NaPB, which may be a consequence of direct interaction with taste receptors which did not occur with Luc 01. Because of a real taste-masking effect the unpleasant taste of NaPB will not be experienced with Luc 01 granules, which is a major improvement over the existing formulation for the clinical management of children with UCD.

Existing marketed formulations of NaPB have an extremely unpleasant taste which interferes with compliance to the lifelong treatment of UCD. Children, in particular, might refuse to take the drug, which severely affects treatment. The bad taste was sometimes even associated with the intake of food, which can create problems as these patients require strict nutritional regimens, and this might lead to clinical decompensations. Avoidance of the taste of this essential drug by using intranasal tubing, gastrostomy or the addition of sweeteners, is onerous and adds to the familial burden of the disease.

These concerns stimulated the development of a bioequivalent taste-less granule formulation adapted to children. This new individual-granule coated formulation of NaPB is expected to improve compliance with treatment and its long-term benefit in preventing metabolic decompensations in UCD. Despite the different drug release pattern leading to an improved taste profile, these studies showed that the test granule formulation and the reference NaPB formulation are bioequivalent in rate and extent of absorption. This new formulation may lead to better acceptance of the treatment and to improved compliance and thus improved efficiency of the treatment.


Supplementary materials

  • Supplementary Data

    This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.

    Files in this Data Supplement:


  • Contributors NG is the first author, being directly interested in the clinical management and treatment of paediatric patients with UCD. JB is the last author and responsible for in vitro taste and dissolution studies and a major contributor to the development of pharmaceutical formulations adapted to the treatment of children in the EU. CT is a collaborator to JB and was reponsible for the conduct of in vitro taste and dissolution studies. WC is the head of Parexel International pharmacokinetic and pharmacodynamic evaluation. YK is the head of medical affairs at Lucane Pharma, the sponsor of the studies.

  • Funding The in vitro and in vivo studies were sponsored by Lucane Pharma, 172 rue de Charonne, 75011 Paris, France.

  • Competing interests YK is an employee of Lucane Pharma. There are no other competing interests.

  • Ethics approval Pharma-Ethics Independent Research Lyttelton Manor, Republic of South Africa.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.