Quality and stability of extemporaneous pyridoxal phosphate preparations used in the treatment of paediatric epilepsy

Article abstract of DOI:10.1111/jphp.12701

Objectives: To assess the pyridoxal 5′-phosphate (PLP) content and stability of extemporaneous PLP liquids prepared from dietary supplements used for the treatment of vitamin B₆-dependent epilepsy. Methods: Pyridoxal 5′-phosphate liquids were prepared in accordance with the guidelines given to patients from marketed 50 mg PLP dietary capsules and tablets. The PLP content and its stability were evaluated under conditions resembling the clinical setting using reverse phase HPLC and mass spectrometry. Key findings: Pyridoxal 5′-phosphate content in most of the extemporaneously prepared liquids from dietary supplements was found to be different from the expected amount (~16–60 mg). Most of these PLP extemporaneous liquids were stable at room temperature (protected from light) after 24 h but unstable after 4 h when exposed to light. A key photodegradation product of PLP in water was confirmed as 4-pyridoxic acid 5′-phosphate (PAP). Conclusion: Pyridoxal 5′-phosphate tablets from Solgar were found to be the most reliable product for the preparation of extemporaneous PLP liquids. This work highlighted the difference between the marketed PLP dietary supplements quality and the importance of proper storage of aqueous PLP. There is a need to develop pharmaceutical forms of PLP that ensure dose accuracy and avoid potentially unsafe impurities with the aim of enhancing safety and compliance.

Full text of DOI:10.1111/jphp.12701

Research Paper
Quality and stability of extemporaneous pyridoxal
phosphate preparations used in the treatment of paediatric
epilepsy
Abeer H. A. Mohamed-Ahmeda,*, Matthew P. Wilsonb,*, Maedelyn Albueraa, Ting Chena,
Philippa B. Millsb, Emma J. Footittc, Peter T. Claytonb and Catherine Tuleua
aUCL School of Pharmacy, University College London, London, UK, bInstitute of Child Health, University College London, London, UK and
cMetabolic Medicine Unit, Great Ormond Street Hospital for Children NHS Foundation Trust, London, UK
Keywords
Abstract
Dietary supplements; epilepsy; pyridox(am)
ine 50-phosphate oxidase deficiency;
Objectives To assess the pyridoxal 50-phosphate (PLP) content and stability of
extemporaneous PLP liquids prepared from dietary supplements used for the
treatment of vitamin B6-dependent epilepsy.
Methods Pyridoxal 50-phosphate liquids were prepared in accordance with the
guidelines given to patients from marketed 50 mg PLP dietary capsules and
tablets. The PLP content and its stability were evaluated under conditions
resembling the clinical setting using reverse phase HPLC and mass
spectrometry.
pyridoxal-5-phosphate; stability
Correspondence
Abeer H. A. Mohamed-Ahmed, UCL School
of Pharmacy, University College London,
29-39 Brunswick Square, London WC1N
1AX, UK.
E-mail: abeer.ahm0065d@ucl.ac.uk
Key findings Pyridoxal 50-phosphate content in most of the extemporaneously
prepared liquids from dietary supplements was found to be different from the
expected amount (~16–60 mg). Most of these PLP extemporaneous liquids were
stable at room temperature (protected from light) after 24 h but unstable after
4 h when exposed to light. A key photodegradation product of PLP in water was
confirmed as 4-pyridoxic acid 50-phosphate (PAP).
Received June 17, 2016
Accepted January 12, 2017
doi: 10.1111/jphp.12701
Conclusion Pyridoxal 50-phosphate tablets from Solgarâ were found to be the
most reliable product for the preparation of extemporaneous PLP liquids. This
work highlighted the difference between the marketed PLP dietary supplements
quality and the importance of proper storage of aqueous PLP. There is a need to
develop pharmaceutical forms of PLP that ensure dose accuracy and avoid
potentially unsafe impurities with the aim of enhancing safety and compliance.
*These authors contributed equally to the
manuscript.
also essential for the recycling of PLP from degraded
enzymes in a ‘salvage pathway’.[2]
Introduction
Pyridoxal-5-phosphate (PLP) is the biologically active form
of vitamin B6 (Figure 1) and it is involved in the catalysis
of more than 140 enzymatic reactions in the body including
those involved in the synthesis and degradation of neuro-
transmitters. There are multiple disorders relating to a defi-
ciency of intracellular PLP. Several of these results in
neurological dysfunction and epilepsy, an unsurprising out-
come given the integral role PLP plays in neurotransmitter
metabolism.[1] Pyridox(am)ine 50-phosphate oxidase
(PNPO) is a flavin mononucleotide (FMN)-dependent
enzyme required for the synthesis of PLP from dietary pyri-
doxine (PN), pyridoxamine (PM) and their phosphates
(PNP and PMP), as well as the glucoside of pyridoxine. It is
PNPO deficiency is an autosomal recessive inborn
error of metabolism that leads to neonatal epileptic
encephalopathy. This disorder is characterised by drug-
resistant seizures that can be fatal. Classically, seizures are
responsive exclusively to PLP but patients show a good
neurodevelopmental outcome when treated promptly.[3–5]
Recently, it has become apparent that in some cases of
PNPO deficiency, pyridoxine is an effective treatment.[6–
9]
However, in more than half of the patients, PLP is still
the only effective form of treatment. PLP has also been
shown to aid seizure control in 11.7% of idiopathic
intractable epilepsies and is seen to produce a better out-
come than pyridoxine.[10]
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**
1

Pyridoxal-5-phosphate
Abeer H. A. Mohamed-Ahmed et al.
Figure 1 The human vitamin B6 pathway from dietary to intracellular pyridoxal 50-phosphate showing the activity of intestinal phosphatases,
pyridoxal kinase, pyridox(am)ine 5-phosphate oxidase, tissue non-specific alkaline phosphatase, aldehyde oxidase and aldehyde dehydrogenases
(adapted from Clayton 2006).[1]
There is currently no pharmaceutically licensed form of
PLP; it is only available in the form of dietary supplements.
There are several PLP dietary supplements in the form of
capsules, tablets and liquid (Appendix 1). Unlike medici-
nes, dietary supplements are not strictly regulated and their
manufacture is not standardised. Therefore, the content of
PLP in these dietary supplements should be determined to
ensure dose accuracy. Patients with PNPO deficiency are
treated by dissolving a crushed tablet or the content of cap-
sule (50 mg PLP) in 5–10 ml water before oral administra-
tion or via a nasogastric tube. Because of the difficulties
experienced in the preparation, this aqueous PLP liquid is
often prepared in advance and given to the patient some
hours later. Often, PLP tablets are enteric coated and hence
difficult to crush them. Additionally, large volumes of liq-
uid are required because of the poor solubility, and high
amounts of PLP are needed for seizure control. Along with
poor palatability, these large volumes have been known to
cause vomiting in patients. Overall, the current acceptabil-
ity of PLP when used as an anti-seizure medication is a
major concern, and it increases the burden of disease on
both patient and carer.
(LFTs) and, with time, hepatic cirrhosis.[11,12] Transient
LFT increases were shown to occur in 14/28 patients treated
with PLP for infantile spasms, these were resolved on cessa-
tion of PLP supplementation.[13] In addition, one reported
homocystinuria patient treated with 1000 mg/day of PLP
developed hepatitis and deranged LFTs within 4 days of his
dose being raised.[14]The cause of this liver toxicity is not
clear but is hypothesised that it may occur as a result of
either the high doses of PLP itself (30–100 mg/kg/day in 4–
6 doses) or ingestion of compounds arising as a result of
degradation of PLP when in solution. Importantly, hepatic
disease is not seen in PNPO patients supplemented with
high dose of pyridoxine or in other B6 responsive disorders
treated similarly.
Pyridoxal 50-phosphate is known to be unstable in aque-
ous form and undergoes photodegradation. Several photol-
ysis products of PLP were first postulated nearly sixty years
ago.[15] This photochemical reaction is irreversible and the
degradation compounds formed are oxygen concentration-
dependent and PLP concentration-dependent.[16] One of
the degradation products has been identified as 4-pyridoxic
acid 50- phosphate (PAP), formed by the oxidation of the
aldehyde at the 40 position of PLP (Figure 2). In the pres-
ence of O2, PAP is readily formed when PLP is present at
It has been reported that some patients receiving this
aqueous form of PLP develop deranged liver function tests
2
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**

Abeer H. A. Mohamed-Ahmed et al.
Pyridoxal-5-phosphate
heptafluorobutyric acid (HFBA) and acetic acid were pur-
chased from Sigma.
Weight uniformity of dosage form unit
[British Pharmacopoeia (BP) 2014]
Figure 2 The photooxidation of pyridoxal 50-phosphate to PAP.
Pharmaceutical tablets and capsules must comply with the
tests for content uniformity or the test for uniformity of
mass. The tablets were weighed directly on a precision bal-
ance whilst the content of the capsules was evaluated by
taking the difference between the mass of each of the
twenty filled capsules and the average mass of 20 empty
capsule shells. Data were expressed as an average of content
mass ꢀ standard deviation (mg). The uniformity of con-
tent was measured using high performance liquid chro-
matography (HPLC).
low concentrations, however, at high concentrations other
yet unidentified species, with a peak absorbance of 288 nm
are also formed, these are postulated to be dimers of
PLP.[16] These higher concentrations are likely more repre-
sentative of those seen on the use of PLP as an anti-seizure
medication. It is possible that one or more currently
unidentified degradation products are hepatotoxic and
responsible for the liver dysfunction recently identified in
patients on high-dose PLP supplementation.
The aims of our work are to understand what patients
receive when marketed supplement products are manipu-
lated to prepare a liquid dosage form to be administered to
children for the treatment of PLP-responsive seizures by:
(1) Assessing the available marketed PLP dietary supplements
when prepared in aqueous solution, (2) Investigating the sta-
bility of the liquid PLP derived from different dietary supple-
ments depending upon the storage conditions (temperature
and light), (3) Characterising the photostability of PLP in
aqueous solution. It is evident that this work is novel because;
(i) this is the first study in which the quality of the marketed
PLP products has been assessed and, (ii) the knowledge of
the quality of PLP products (PLP content and stability) will
help understand their clinical efficacy and safety.
Quantification of pyridoxal 50-phosphate in
dietary supplements using HPLC
Pyridoxal 50-phosphate was quantified using reverse phase
HPLC. Quantification and stability evaluations were per-
formed using an Agilent Technologies 1200 Series HPLC sys-
tem (Agilent Technologies, Palo Alto USA), equipped with
an autosampler (injection volume of 10 ll) and a UV-VIS
detector, set to perform measurements at a fixed absorption
maximum wavelength of 285 nm at 25°C. A Synergi Polar
(RP) HPLC column from Phenomenex (250 9 4.60 mm,
4 lm), was used with a gradient elution method at a flow
rate of 1.0 ml/min. The mobile phase was initially 95% water
(containing 0.2% trifluoroacetic acid) and 5% methanol.
The percentage of methanol increased slowly to 40% over
20 min then gradually reduced to 5% over 5 min. The run
time was 25 min. The retention time of PLP under the
HPLC conditions was 5.5 min. Area under the peak was
used to quantify PLP using a calibration curve (Appendix 2).
Seven marketed PLP dietary products were quantified.
The crushed PLP tablets (using a mortar and pestle) and
capsule contents were dissolved in two volumes of deionised
water (10 and 50 ml volumetric flasks) resulting in a yellow
cloudy solution at theoretical concentration of 50 mg/10 ml
(5 mg/ml) or 50 mg/50 ml (1 mg/ml). Pure PLP dissolved
in the same volume of water was used as control. These liq-
uids were protected from exposure to light with foil and left
to stir for 40 min at room temperature. The whole suspen-
sions were filtered using a micro-syringe filter (Millex PES
membrane 0.22 lm) before being analysed by HPLC.
Materials and Methods
Materials
Pyridoxal phosphate hydrate (>98%) was purchased from
Sigma Aldrich. The marketed PLP dietary products include:
Solgar 50 mg tablets (Batch No. 746087-03), Country Life
tablets (Lot No. 13L622A), Thorne capsules (LB11606),
Food Science capsules (21412000 1115), Biocare capsules
(BN 32711) and Metabolics PLP liquid (800 lg/drop) were
purchased from the Amazon website. PLP Bonusan tablets
(Batch No. 0822) were purchased from Bonusan company.
PLP Vitacost tablets (Batch No. 151027) were purchased
from Vitacost company, Their composition is described in
Appendix 1. HPLC grade methanol and trifluoroacetic acid
(TFA) were purchased from Fisher Scientific. A custom
synthesised PAP standard was purchased from Dr Herman
ten Brink, Vrije Universiteit Medisch Centrum, Amster-
dam, The Netherlands. D3-PLP was purchased from
Buchem BV, Apeldoorn, The Netherlands. HPLC grade
methanol was purchased from VWR, both LC-MS grade
The maximum solubility of PLP in distilled water was
investigated under the same conditions as the quantifica-
tion test. A supersaturated solution of PLP in water
(10 mg/ml) was prepared, protected from light exposure by
foil, and left to stir for 40 min at room temperature. This
suspension was filtered and diluted tenfold before analysis
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**
3

Pyridoxal-5-phosphate
Abeer H. A. Mohamed-Ahmed et al.
by HPLC. All experiments were repeated thrice. Data were
expressed as average concentration (mg/ml) or PLP
amount (mg) ꢀ standard deviation.
Pyridoxal 50-phosphate and PAP were quantified by mea-
suring the ratio of the peak area to that of D3-PLP. Calibra-
tion curves for PLP and PAP were constructed in MQ-H20
using a D3-PLP internal standard concentration of 25 nmol/
l. These calibration curves were shown to be linear between
1–200 nmol/l (PLP: R2 = 0.9982; PAP: R2 = 0.9990). (Fig-
ure S4) This allowed the accurate quantification of PLP and
PAP, which were distinguishable by both retention time and
m/z ratio (Table S1). Data acquisition and analysis were per-
formed using Masslynx software. Statistical analysis was per-
formed using Graphpad Prism version 5.00 for Windows.
Stability testing of marketed pyridoxal 50-
phosphate products using HPLC-UV/VIS
The content of PLP capsules or crushed tablets were sus-
pended in deionised water (10 ml volumetric flasks) to
achieve a 5 mg/ml strength. For liquid PLP (Metabolics),
4.5 ml (11.5 mg/ml calculated concentration) of the liquid
was withdrawn and diluted using water to a PLP concentra-
tion of 5 mg/ml. Pure PLP powder (50 mg) was dissolved
in distilled water (10 ml) and used as a control. These solu-
tions were protected from light exposure by foil and left to
stir for 40 min. The stability studies were conducted at
room temperature (25°C) over a 24 h period.
For light stability studies, the liquids were exposed to
light with 2 Philips TL 8W/35 fluorescent lamps enclosed
in a wooden box fitted with a cooling fan 240 V AC run-
ning at a speed of 50–60 Hz with a 23 W power to keep the
atmosphere at ambient temperature. The PLP liquids were
exposed for 4 h along with a solution of pure PLP as a con-
trol. PLP liquids were filtered and then analysed by HPLC.
All the stability tests were repeated thrice. Results from the
stability tests were expressed as the mean of the amount of
PLP remaining (%) ꢀ standard deviation.
Investigation of pyridoxal 50-phosphate
photodegradants using LC-MS/MS
Preliminary mass spectra were obtained upon the pho-
todegradation of 1 mM PLP solutions in deionised water
exposed to sunlight. Investigative MS1, MS2, daughter ion
and neutral loss scans (Figure S5) were performed by either
direct or combined infusion mode using the Xevo TQ-S
instrument described above. For specific identification of
photoproducts previously separated by HPLC, the peaks of
interest were collected and analysed in an identical fashion
by either direct or combined infusion.
Statistical analysis
Statistical analysis for the comparison of different PLP solu-
tions prepared from the marketed products and pure PLP
after stability testing was conducted using Kruskal-Wallis
(using statistic programme Originâ). Conover-Iman test
with P adjustment (bonferroni) was conducted using R
(version 3.0.1) to compare between individual products.
Assessment of pyridoxal 50-phosphate
Photodegradation by LC-MS/MS
The rate of PLP photodegradation was measured by the
incubation of freshly prepared solutions at a theoretical
concentration of 5 mg/ml, according to the stated PLP con-
tent in each product. The marketed formulations or analyt-
ical grade PLP were incubated for 1, 4 and 24 h under the
same conditions described in the above photostability stud-
ies (n = 3). Solutions were analysed using a modified ver-
sion of the LC-MS/MS method published by Footitt
et al.[17] Samples were diluted to 100 nM and an equal vol-
ume of 0.3 N TCA containing 50 nmol/l D3-PLP internal
standard was added.
Results
Weight uniformity and pyridoxal 50-
phosphate content
Weight variation is associated with non-uniformity of the
amount of drug in solid formulations and poses a risk of
dose variation; therefore, the weight uniformity of PLP diet-
ary supplements was examined. The BP requirements for
tablets weighing more than 250 mg is that no more than 2/
20 tablets can deviate from the mean weight by more than
5% and none should deviate by more than 10%. For capsules
weighing less than 300 mg the BP guidelines state that no
more than 2/20 tablets can deviate from the mean by more
than 10% and none should deviate by more than 20%.
LC-MS/MS analysis was performed using a Waters H-
Class FTN LC linked to a Waters Xevo TQ-S mass spec-
trometer in Multiple Reaction Monitoring (MRM) mode.
An HSS T3 column was used with an HSS T3 guard and
the mobile phases employed in the analysis consisted of A:
3.7% acetic acid, 0.01% HFBA and B: 100% methanol at a
flow rate of 0.4 ml/min. The mobile phase composition
graduated linearly from 97.5% A, 2.5% B at 0 min to
0.01% A, 99.9% B at 4.5 min before returning to initial
conditions for a re-equilibration step of 2 min. The
injected volume of the sample was 8 ll.
Among the seven marketed PLP dietary products anal-
ysed, Food Science, Biocare and Bonusan did not comply
with the BP requirements with weight variability being
most marked for the Biocareâ capsules (Table 1). PLP
4
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**

Abeer H. A. Mohamed-Ahmed et al.
Pyridoxal-5-phosphate
tablets (Solgar, Country Life and Vitacost) and Thorne cap-
sules met the BP requirements for weight uniformity
(Table 1).
11.01 ꢀ 5.78 mg/ml PLP as well as significant amounts of
degradation products. There is therefore a high risk of inac-
curate dosing if the PLP dietary supplements from Biocare,
Food Science, Vitacost, Thorne or Metabolics are used for
preparation of PLP extemporaneous liquids.
The content of PLP in the marketed products was
assessed after dissolving the crushed tablets/contents of the
capsules in 10 ml of deionised water as in the clinic or
50 ml to ensure complete dissolution of PLP (solubility of
PLP is 5.7 mg/ml in water http://www.drugbank.ca/drugs/
DB00114). The maximum experimentally calculated solu-
bility of PLP in water at room temperature was
8.37 ꢀ 0.94 mg/ml. According to the manufacturers, all
the dietary products contained 50 mg PLP. Products from
Solgar, Bonusan and Thorne contained the stated amount
~ 50 mg when dissolved in 10 ml water (Table 2) whilst
Country Life, Food Science and Biocare products contained
less than 50 mg (~2–30 mg) (Table 2). When the volume
of water used for dissolution was increased to 50 ml,
the same PLP content was obtained for all the products
except for Country Life and Thorne (Table 2). The content
of PLP in both Vitacost and Thorne was higher than
50 mg. The liquid form of PLP (Metabolicsâ) contained
Stability Studies
Photostability of pyridoxal 50-phosphate and
its marketed products
Quantification of 4-pyridoxic acid 50-phosphate as
a major degradation product of pyridoxal 50-phos-
phate
Analysis of a PLP solution by HPLC-UV/VIS after exposure
to light revealed the presence of degradation products (Fig-
ure S2). In particular, a major product was seen to elute at
3.5 min. There were also two more minor photoproducts
found to elute after PLP between 6.5 and 10 min HPLC-
UV/VIS analysis (Figure S2).
To identify these compounds and study their rate of for-
mation, mass spectral analysis was performed, followed by
multiple reactions monitoring (MRM)-based LC-MS/MS
for accurate quantification. Upon direct infusion of a pho-
todegraded analytical grade PLP solution, several potential
products were detected. Some of these were shown to con-
tain phosphoric acid, signified by a loss of m/z 98 upon
fragmentation with significant products seen at m/z 397.2,
381.1, 264.0, 262.9 and 219.9 (Figure S5). Because of its
fragmentation pattern and abundancy, the product at m/z
264.0 was hypothesised to be 4-pyridoxic acid 50-phosphate
(PAP), the only previously confirmed PLP pho-
todegradant.[16] Based on daughter scans of m/z 264, m/z
166.0 was identified as the most abundant product ion and
a theoretical transition for PAP pertaining to a loss of phos-
phoric acid (m/z 264.0 > 166.0) was included in all subse-
quent LC-MS/MS analyses. An ion corresponding to this
mass eluting at 0.78 min was evident after injection of the
degraded PLP solution onto the LC-MS/MS with pyridoxal
phosphate eluting at 0.94 min (Table S1).
Table 1 Weight uniformity of pyridoxal 50-phosphate dietary prod-
ucts. Twenty tablets or capsules content were weighed from each
product. Data are expressed as mean ꢀ standard deviation
BP criteria
No more
than 2/20
masses
None
deviate by
more than
Dosage
form
Product
Weight (mg)
deviate by 5% 10%
Solgar
Tablet
Tablet
Tablet
Tablet
350.7 ꢀ 5.4
330.4 ꢀ 4.9
337.8 ꢀ 4.1
250.2 ꢀ 10.7
√
√
√
√
√
x
√
√
Country Life
Vitacost
√
√
Bonusan
x
x
x
√
Food Science Capsule 393.8 ꢀ 10.0
Biocare
Thorne
Capsule 288.5 ꢀ 21.2
Capsule 197.2 ꢀ 6.3
BP, British Pharmacopoeia.
X did not meet B.P requirements for weight uniformity test.
Table 2 Pyridoxal 50-phosphate content of dietary products dis-
solved in 10 and 50 ml deionised water. Data are expressed as
mean ꢀ standard deviation (n = 3)
The fragmentation and elution pattern of PAP was con-
firmed using the synthesised standard. Light irradiation of
pure PLP over 24 h resulted in a dramatic decrease in PLP
concentration. This was concomitant with the appearance
of PAP, the concentration of which increased with time
(Figure 3). Interestingly, at the concentrations of PLP used,
the proportion of PAP formed after photolysis was found
to only constitute 22.7 (ꢀ4.6%), 20.2 (ꢀ1.8%) and 27.5
(ꢀ4.6%) of the ‘missing fraction’ of PLP after 1, 4 and 24 h
respectively. This confirms earlier work by Reiber (1972)
that showed a large quantity of PLP is photolysed into
compounds other than PAP.
Amount of dissolved drug (mg)
Products
Dosage form
10 ml water
50 ml water
Country Life
Solgar
Tablet
15.9 ꢀ 5.8
51.3 ꢀ 2.0
53.1 ꢀ 1.0
62.7 ꢀ 1.5
30.9 ꢀ 6.0
2.1 ꢀ 0.6
46.2 ꢀ 4.4
53.4 ꢀ 2.2
51.4 ꢀ 0.3
62.3 ꢀ 1.7
36.6 ꢀ 4.4
2.41 ꢀ 0.1
64.4 ꢀ 2.8
51.2 ꢀ 0.1
Tablet
Bonusan
Vitacost
Tablet
Tablet
Food Science
Biocare
Capsule
Capsule
Capsule
Powder
Thorne
54.7 ꢀ 4.0
49.3 ꢀ 1.1
Pure PLP
PLP, pyridoxal 50-phosphate.
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**
5

Pyridoxal-5-phosphate
Abeer H. A. Mohamed-Ahmed et al.
Figure 3 LC-MS/MS analysis of the stability of analytical grade aque-
ous pyridoxal 50-phosphate and the formation of 4-pyridoxic acid 50-
phosphate under light irradiation. Data are expressed as mean ꢀ stan-
dard deviation (n = 3).
Figure 4 The stability of marketed pyridoxal 50-phosphate products
dissolved in 10 ml water after 4 h of exposure to light. Data are
expressed as mean ꢀ standard deviation (n = 3).
Subsequent analysis of the pure PAP standard on the
HPLC/UV-VIS system revealed that PAP elutes at 7.2 min
and does not correspond to the major photodegradation
peak seen at 3.5 min (Figure S3). Upon collection of this
peak and direct infusion into a Xevo TQ-S mass spectrome-
ter, abundant ions were seen at m/z 492.9, 395.0, 297.1 and
247.2. These are hypothesised to correspond to [M+H]+,
[M+H-H3PO4]+, [M+H-2(H3PO4)]+ and [M+2H]2+ of a
diketone dimer of PLP, respectively, as predicted by Mor-
rison and Long in 1958[15] Spectrophotometric analysis of
this peak showed strong absorbance at 288 nm and is
therefore considered highly likely to be one of the afore-
mentioned ‘288 nm absorbing species’ identified by Reiber
in 1972.[16]
Several of the PLP dietary supplements identified as
being the most reliable by HPLC-UV/VIS were further
studied by LC-MS/MS and a similar loss of PLP was seen
over time. In addition, in the marketed products, PAP was
shown to form by photodegradation at a similar rate to that
seen in pure PLP (Table 3).
Stability of light protected marketed
pyridoxal 50-phosphate dietary supplements
in solution at room temperature
At room temperature whilst protected from light, the sta-
bility of PLP was improved: with ~ 90% of drug remaining
after 24 h (Figure 5) for all the products except for aqueous
liquids made from Food Science (82.7 ꢀ 1.8 drug) and
Country Life (40.2 ꢀ 6.9%) products. There was signifi-
cant difference between all marketed PLP and pure PLP
(P = 0.012). Country life was significantly unstable com-
pared with the pure PLP (P = 0.029), Solgar (P = 0.006),
Bonusan (P = 0.0002) and Vitacost (P = 0.008) (Figure 5).
Pure PLP solution was stable at room temperature
(93.1 ꢀ 4.8% drug remaining) but PLP Metabolics aque-
ous liquid (56.0 ꢀ 0.8% drug remaining) was not.
Photostability of marketed pyridoxal 50-phosphate
dietary supplements in solution
Extemporaneously prepared aqueous liquids of marketed
PLP dietary products were found to be unstable after 4 h
exposure to light (Figure 4). Solgar and Thorne were the
most stable (82.9 ꢀ 7.9 and 81.7 ꢀ 6.6% drug remaining
respectively) followed by Bonusan (72.5 ꢀ 18.9%), Food
Science (69.7 ꢀ 2.2%) and Vitacost (66.3 ꢀ 1.2%); Coun-
try Life was the least stable (34.6 ꢀ 7.2%). Pure PLP was
unstable and displayed 68.0 ꢀ 5.6% drug remaining. There
was significant difference between all marketed PLP and
pure PLP (P = 0.025) (Figure 4). Country life was signifi-
cantly unstable compared with Solgar and Thorne
(P = 0.003 and 0.011) (Figure 4).
The stability of the marketed liquid form of PLP
(Metabolicsâ) was also assessed and revealed that only
47.0 ꢀ 3.2% drug remained after 4 h exposure to light.
This product was impure to begin with and contained
degradation products irrespective of whether it had been
stored in the fridge or at room temperature.
Discussion
Pyridoxal 50-phosphate is used for the treatment of epilepsy
caused by a deficiency of PNPO which, if untreated, can be
fatal. Currently, there is no pharmaceutical form of PLP for
paediatric use, so liquid forms of PLP are prepared from
dietary products using crushed tablets or capsule contents
mixed in a specific volume of tap water. There are major
concerns about these extemporaneous PLP liquids regard-
ing dose accuracy, stability and safety.
There are several issues with the feasibility of preparation
of these extemporaneous liquids. At 25°C analytical grade
6
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**

Abeer H. A. Mohamed-Ahmed et al.
Pyridoxal-5-phosphate
Table 3 PLP remaining, taken as a percentage of initial pyridoxal
50-phosphate (PLP), because of light degradation over 24 h, as measured
by LC-MS/MS. Quantity of 4-pyridoxic acid 50-phosphate formed over
24 h because ofphotodegradation of PLP, calculated as a percentage of
PLP lost over the same period. Initial concentrations were nominally
5 mg/ml. Data are expressed as mean ꢀ standard deviation (n = 3)
formulation, carers usually only take the supernatant after
solubilising PLP and leave any insoluble excipients behind,
not knowing what is left, but in this context pragmatism
prevails. This study has replicated the clinical scenario by
only measuring PLP in the supernatant of the solutions
prepared for analysis and has not measured any PLP that
may be sequestered in the insoluble fraction. It is possible
(indeed likely given its reactivity) that additional PLP is
bound to other insoluble excipients (Appendix 1). If these
products are taken as a dietary supplement it is likely the
acidic conditions found in the stomach will release all PLP
found therein. Moreover, it is also thought that these experi-
ments being performed in the laboratory possibly improved
the preparation methods of these PLP liquids and are not
strictly what an inexperienced parent would do.
Initial PLP
(mg/ml)
PLP remaining
PAP Formed
Products
(% of initial PLP)
(% of PLP lost)
Thorne
Solgar
5.7 ꢀ 0.7
4.1 ꢀ 0.4
4.3 ꢀ 0.1
4.6 ꢀ 0.7
21.7 ꢀ 9.3
18.8 ꢀ 3.8
16.0 ꢀ 11.1
16. 5 ꢀ 2.7
26.6 ꢀ 11.5
23.1 ꢀ 5.3
18.7 ꢀ 10.7
27.5 ꢀ 0.7
Vitacost
Pure PLP
PAP, 4-pyridoxic acid 50-phosphate.
All of the extemporaneous liquids prepared from the
tested marketed PLP products showed degradation when
exposed to light even for short periods of time (4 h). This
was consistent with the studies of Morrison and Long[15]
and Ubbink,[21] which showed that pure PLP solutions
were unstable after 4 h incubation in simulated daylight
(66.5 ꢀ 4.3% drug remaining). Morrison and Long[15]
observed that pure PLP dissolved in air-free water com-
pletely photolysed after one hour in bright summer sun-
light. Our advice is that PLP should always be made up
immediately before administration. However, if this is
impossible, it is imperative to protect PLP aqueous liquids
from light when prepared in advance to avoid risk of dose
inaccuracy and potential toxicity from any degradation
products that may form. Unfortunately, this might be diffi-
cult to implement in the patient’s home.
Figure 5 The stability of marketed pyridoxal 50-phosphate products
in water (5 mg/ml PLP) after 24 h at room temperature (25°C). Data
are expressed as mean ꢀ standard deviation (n = 3).
PLP requires continuous stirring for 40 min to completely
dissolve at a concentration of 5 mg/ml, therefore, it is likely
that PLP prepared from dietary products by manual shak-
ing within a clinical setting or at home might not com-
pletely dissolve to give 5 mg/ml. This was evidenced by our
findings; PLP contents in liquids that were prepared from
dietary products according to common clinical usage were
not accurate. Some of these products contained less than
the 50 mg stated on the label (Country Life, Food Science,
Biocare and Metabolics) and others more (Vitacost and
Thorne). This variation in PLP content can result in dose
inaccuracy leading to either inefficacy or toxicity.[18–20] The
low amounts of PLP solubilised in some products are par-
ticularly dangerous, given that non-response to these sup-
plements can be taken to be diagnostically indicative of a
seizure disorder that does not respond to PLP.
Most of the extemporaneous liquids prepared from the
tested marketed PLP products and pure PLP were stable at
room temperature when protected from light (~90% drug
remaining). This was not consistent with the study of Shep-
hard and Labadarios,[22] who have reported that 95% of a
solution of pure PLP was degraded (mostly by hydrolysis to
PL) after storage at room temperature in the dark for 24 h.
The concentration of PLP is known to affect its degradation
pattern.[14] This might explain the different findings in
Shephard and Labadario’s studies in which PLP was
assessed at low concentrations (1 lg/ml), while in our
studies a higher concentration relevant to the clinical set-
ting was investigated (5 mg/ml).
The photodegradation products of PLP were further
characterised using LC-MS/MS and showed that after 24 h
exposure to light 27.4% of degraded PLP had been con-
verted to PAP with a further 72.6% degraded to other
products including a diketone dimer of PLP. This is crucial
as these degradation products could cause toxic side effects.
PAP, naturally not present in humans, has been shown to
inhibit PLP-dependent enzymes and thus theoretically
could be a cause of the hepatotoxicity seen in PNPO
Several dietary PLP products (Biocare, Bonusan and
Food Science) did not comply with the B.P guidelines with
regard to weight uniformity. This increases the risk of dose
inaccuracy. Furthermore, some of the tablets (Solgar and
Vitacost) are enteric coated making them difficult to crush
and cause the generation of residues that can impair
administration. When used clinically as
a
liquid
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**
7

Pyridoxal-5-phosphate
Abeer H. A. Mohamed-Ahmed et al.
patients on high-dose PLP supplementation.[23] However,
in vivo it is likely that PAP is hydrolysed and subsequently
excreted as pyridoxic acid in urine.[24,25] It is known that
the B6 vitamers are mostly absorbed in their non-phos-
phorylated form and do not cross the gastro-intestinal wall
in appreciable amounts before hydrolysis.[26] In addition,
upon measurement of the B6 vitamer blood levels of
patients on PLP supplementation, PAP is not detected in
significant amounts (Unpublished data Laboratory of Mills
P, 2016).
Conclusions
Pyridoxal 50-phosphate from Solgar, Thorne, Bonusan
and Vitacost are the most stable products at room tem-
perature, protected from light. Bonusan tablets are of
non-uniform weight. Throne and Vitacost contain higher
than stated amounts of PLP [~ 60 mg] which increases
the risk of dose inaccuracy and toxicity. Although Solgar
was of the best quality, none of the currently marketed
products are truly suitable for long-term use at high
doses because of possible liver toxicity. There is a need
for a more stable pharmaceutically licensed treatment
that is easily prepared and administered to PLP-respon-
sive patients.
While it seems unlikely that PAP is a major toxic degra-
dation product, other structural analogues of the B6 vita-
mers are already known to have toxic effects. Gingko toxin
(40-O-methylpyridoxine),
a naturally occurring plant
extract, causes seizures thought to be because of inhibition
of the pyridoxal kinase enzyme, responsible for phosphory-
lation of the B6 vitamers: pyridoxal, pyridoxamine and
pyridoxine.[27,28] If other structural analogues of PLP are
also capable of inhibition of B6-metabolising or PLP-
dependent enzymes, the effects could be wide ranging and
could indeed result in hepatic dysfunction.
The quantity of PAP formed upon the photodegradation
of PLP in aqueous solution was accurately quantified. This
accounted for 27% of the PLP lost/24 h. The implications
of this are significant for patients with seizure disorders on
high doses of PLP as are the presence of other degradation
products (73% of total degraded PLP) which could be
responsible for hepatic cirrhosis in these patients.
Coman et al. postulate that the hepatic cirrhosis seen on
high-dose PLP supplementation results because of the aber-
rant purinoceptor activation of hepatic stellate cells
(HSC).[11] P2 receptors of HSC are linked with a fibrogenic
response through increased collagen production. PLP itself
acts on P2 receptors in an antagonist role but other PLP
analogues show both agonistic and antagonist effects. It is
possible that degradation products or metabolites of PLP
may cause the hepatic fibrosis and deranged LFTs seen in
PNPO deficient patients on high doses of PLP by
purinoceptor activation.
We recommend that (1) high-dose PLP supplements be
prepared immediately before administration and protected
from light to avoid degradation in solution, (2) PLP dietary
supplements that do not meet the standards required for
pharmaceutical products should be avoided and (3) the
LFTs of all patients on this treatment should be closely
monitored.
Declarations
Acknowledgement
The identification of other photodegradants of PLP goes
beyond the scope of this study, but would be an important
focus for any future work investigating the hepatotoxicity
seen in certain patients on large doses of PLP.
We would like to thank Dr. Yucheng Sheng for his help
with the statistical analysis and Dr. Elin Haf Davies for her
support and mentoring.
4. Hatch J et al. Normal Neurodevelop-
mental Outcomes in PNPO Deficiency:
8. Pearl PL et al. Partial Pyridoxine
Responsiveness in PNPO Deficiency.
JIMD Rep 2013; 9: 139–142.
References
a Case Series and Literature Review.
JIMD Rep 2015; 16: 1392–1394.
5. Clayton PT et al. Neonatal epileptic
encephalopathy. Lancet 2003; 361:
1614.
1. Clayton PT. B6-responsive disorders:
a model of vitamin dependency. J
Inherit Metab Dis 2006; 29: 317–326.
2. Musayev FN et al. Structure and
properties of recombinant human
pyridoxine 50-phosphate oxidase. Pro-
tein Sci 2003; 12: 1455–1463.
9. Pearl PL, Gospe SM Jr. Pyridoxine or
pyridoxal-50-phosphate for neonatal
epilepsy: the distinction just got
murkier. Neurology 2014; 82: 1392–
1394.
6. Mills PB et al. Epilepsy due to PNPO
mutations: genotype, environment 10. Wang HS et al. Pyridoxal phosphate
and treatment affect presentation and
outcome. Brain 2014; 137: 1350–1360.
7. Plecko B et al. Pyridoxine responsive-
is better than pyridoxine for control-
ling idiopathic intractable epilepsy.
Arch Dis Child 2005; 90: 512–515.
3. Mills PB et al. Neonatal epileptic
encephalopathy caused by mutations
in the PNPO gene encoding pyridox
(am)ine 50-phosphate oxidase. Hum
Mol Genet 2005; 14: 1077–1086.
ness in novel mutations of the PNPO 11. Coman D et al. PNPO Deficiency and
gene. Neurology 2014; 82: 1425–1433. Cirrhosis: expanding the Clinical
8
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**

Abeer H. A. Mohamed-Ahmed et al.
Pyridoxal-5-phosphate
Phenotype? JIMD Rep 2015; 25: 71–
75.
method. J Inherit Metab Dis 2013; 36: 23. Churchich JE. Fluorescence changes
139–145.
12. Sudarsanam A et al. Cirrhosis associ- 18. Hartmann H et al. Status epilepticus
induced by binding of 4-pyridoxic
acid 5 -phosphate to proteins. J Biol
Chem 1972; 247: 6953–6959.
ated with pyridoxal 50-phosphate
treatment of pyridoxamine 50-phos-
phate oxidase deficiency. JIMD Rep
2014; 17: 67–70.
in a neonate treated with pyridoxine
because of a familial recurrence risk 24. Fonda ML. Purification and character-
for antiquitin deficiency: pyridoxine
toxicity? Dev Med Child Neurol 2011;
53: 1150–1153.
ization of vitamin B6-phosphate phos-
phatase from human erythrocytes. J
Biol Chem 1992; 267: 15978–15983.
13. Takuma Y, Seki T. Combination ther-
apy of infantile spasms with high-dose 19. Hammen A et al. A paradoxical rise 25. Gao GJ, Fonda ML. Kinetic analysis
pyridoxal phosphate and low-dose
corticotropin. J Child Neurol 1996; 11:
35–40.
of neonatal seizures after treatment
with vitamin B6. Eur J Paediatr Neurol
1998; 2: 319–322.
and chemical modification of vitamin
B6 phosphatase from human erythro-
cytes. J Biol Chem 1994; 269: 7163–
7168.
14. Yoshida I et al. Pyridoxal phosphate- 20. Salazar P, Tapia R. Seizures induced
induced liver injury in a patient with
homocystinuria. J Inherit Metab Dis
1985; 8: 91.
by intracerebral administration of 26. Albersen M et al. The intestine plays a
pyridoxal-50-phosphate:
effect of substantial role in human vitamin B6
metabolism: Caco-2 cell model.
GABAergic drugs and glutamate
receptor antagonists. Neuropharmacol-
ogy 2001; 41: 546–553.
a
15. Morrison AL, Long RF. The photoly-
sis of pyridoxal phosphate. Journal of
PLoS One 2013; 8: e54113.
27. Kastner U et al. The human pyridoxal
kinase, a plausible target for ginkgo-
toxin from Ginkgo biloba. FEBS J
2007; 274: 1036–1045.
the Chemical Society (Resumed) 1958; 21. Ubbink JB et al. Stability of pyri-
211–215.
doxal-5-phosphate
semicarbazone:
16. Reiber H. Photochemical reactions of
vitamin B 6 compounds, isolation and
properties of products. Biochim Bio-
phys Acta 1972; 279: 310–315.
applications in plasma vitamin B6
analysis and population surveys of 28. Gandhi AK et al. Crystal structures of
vitamin B6 nutritional status. J Chro-
matogr 1985; 342: 277–284.
human pyridoxal kinase in complex
with the neurotoxins, ginkgotoxin and
theophylline: insights into pyridoxal
kinase inhibition. PLoS One 2012; 7:
e40954.
17. Footitt EJ et al. Measurement of 22. Shephard GS, Labadarios D. Degrada-
plasma B6 vitamer profiles in children
with inborn errors of vitamin B6
metabolism using an LC-MS/MS
tion of vitamin B6 standard solutions.
Clin Chim Acta 1986; 160: 307–311.
Figure S2. Chromatogram of Pure
Figure S5. Neutral loss (98) mass
Supporting Information
PLP after 4 h exposure to light. Intact spectra of a PLP solution before (a)
PLP appears at retention time and after (b) incubation for 5 h in
5.8 min, while the main degradation sunlight. The major degradation
Additional Supporting Information
may be found in the online version of
this article:
product appears at 3.5 min products seen to lose a phosphoric
Figure S3. Chromatogram of Pure acid residue (m/z 98) can be seen at
Appendix 1. PLP dietary products
on the market (none have marketing
authorisation in the EU)
PAP appears at retention time m/z’s 397.21, 381.15, 264.01(pyridoxic
7.2 min
acid phosphate), 262.97 and 219.99.
Figure S4. Calibration curves for m/z’s 270.12 and 269.81 are likely
PLP and PAP, R2 values are 0.98824 PLP [M+Na]+
Appendix 2. Calibration curve
Figure S1. Standard Calibration
Curve of PLP in water. Equation of
the line is y = 10643x. Data is
expressed as mean ꢀ standard devia-
tion (n = 3)
and 0.9990 respectively. Response is
Table S1. The mass spectral specifi-
calculated by normalising the peak cations for detection of PLP, PAP and
areas of the MRMs for PLP and PAP D3-PLP. All are detected in positive
against that of D3-PLP
ion mode.
© 2017 Royal Pharmaceutical Society, Journal of Pharmacy and Pharmacology, ** (2017), pp. **–**
9