Pharmacokinetic data reduce uncertainty in the acceptable daily intake for benzoic acid and its salts

Pharmacokinetic data reduce uncertainty in the acceptable daily intake for benzoic acid and its salts

Accepted Manuscript Pharmacokinetic data reduce uncertainty in the acceptable daily intake for benzoic acid and its salts K. Zu, D.M. Pizzurro, T.A. L...

2MB Sizes 35 Downloads 83 Views

Accepted Manuscript Pharmacokinetic data reduce uncertainty in the acceptable daily intake for benzoic acid and its salts K. Zu, D.M. Pizzurro, T.A. Lewandowski, J.E. Goodman PII:

S0273-2300(17)30208-8

DOI:

10.1016/j.yrtph.2017.07.012

Reference:

YRTPH 3874

To appear in:

Regulatory Toxicology and Pharmacology

Received Date: 14 April 2017 Revised Date:

13 July 2017

Accepted Date: 14 July 2017

Please cite this article as: Zu, K., Pizzurro, D.M., Lewandowski, T.A., Goodman, J.E., Pharmacokinetic data reduce uncertainty in the acceptable daily intake for benzoic acid and its salts, Regulatory Toxicology and Pharmacology (2017), doi: 10.1016/j.yrtph.2017.07.012. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

2

Pharmacokinetic Data Reduce Uncertainty in the Acceptable Daily Intake for Benzoic Acid and Its Salts

3 4

Zu, K1; Pizzurro, DM1; Lewandowski, TA2; Goodman, JE1*

1

6

1

Gradient. 20 University Road, Cambridge, MA, 02138, USA

7

2

Gradient. 600 Stewart Street, Suite 1900, Seattle, WA, 98101, USA

9

SC

8

Corresponding Author:

M AN U

10

Julie E. Goodman, Ph.D., DABT, ATS, FACE

12

Gradient

13

20 University Road

14

Cambridge, MA 02138

15

Email: [email protected]

AC C

EP

TE D

11

16

RI PT

5

1

ACCEPTED MANUSCRIPT

5 6

RI PT

SC

4

Introduction ............................................................................................................ 4 Methods .................................................................................................................. 7 Interspecies Comparison of Pharmacokinetics of Benzoic Acid and Its Salts ......... 8 3.1 Qualitative Interspecies Comparison of Pharmacokinetics ........................ 9 3.2 CSAF Derivation for Interspecies Differences in Pharmacokinetics ......... 11 3.2.1 CSAF Derivation .......................................................................... 12 3.2.2 Alternative Approaches .............................................................. 15 3.2.3 Discussion ................................................................................... 18 Intraspecies Comparison of Pharmacokinetics of Benzoic Acid and Its Salts in Humans ............................................................................................................. 20 Evaluation of Human Clinical Data of Sodium Benzoate ...................................... 23 Proposed ADI for Benzoic Acid and Its Salts ......................................................... 26

M AN U

1 2 3

References .................................................................................................................................... 28

TE D

32 33

Page

EP

18 19 20 21 22 23 24 25 26 27 28 29 30 31

Table of Contents

AC C

17

2

ACCEPTED MANUSCRIPT

Abstract

35

The current acceptable daily intake (ADI) for benzoic acid and its salts as food additives is 0-5 mg/kg bw-

36

day. This accounts for a total uncertainty factor (UF) of 100, which includes a default factor of 10 for

37

interspecies differences. Based on pharmacokinetic data in rodents and humans, we derived a chemical-

38

specific adjustment factor (CSAF) of 2 for the pharmacokinetic component of the interspecies UF.

39

Additional analyses indicate that this CSAF is conservative and interspecies differences between rats and

40

humans are likely closer to unity. Human clinical studies indicate that the pharmacokinetics of benzoic

41

acid and its salts are similar in children and adults, and that there is a lack of adverse events in humans at

42

doses comparable to the no observed adverse effect level (NOAEL) in rodents; this suggests that the

43

pharmacokinetic UF for intraspecies variability, as well as the pharmacodynamic components of the UFs,

44

may also be reduced, although we did not calculate to what degree. In conclusion, the total UF can be

45

reduced to 50 (2 for interspecies differences in pharmacokinetics, 2.5 for interspecies differences in

46

pharmacodynamics, and 10 for intraspecies variability), which would increase the ADI to 0-10 mg/kg bw-

47

day.

48

50

Keywords:

EP

49

TE D

M AN U

SC

RI PT

34

Benzoic Acid; Sodium Benzoate; Potassium Benzoate; Pharmacokinetics; Chemical-specific Adjustment

52

Factor; Data-derived Exploration Factor; Uncertainty Factor; Acceptable Daily Intake

AC C

51

3

ACCEPTED MANUSCRIPT

1

Introduction

54

Benzoic acid and its salts, particularly sodium and potassium benzoates (Figure 1), are commonly used as

55

preservatives in acidic foods, beverages, pharmaceuticals, and cosmetic products because they inhibit

56

mold, yeast, and certain bacterial growth (Brul and Coote, 1999; Nair, 2001; JECFA, 2015, 2016).

57

Sodium benzoate, benzoic acid, and precursor compounds are evaluated for safety as a group because,

58

upon ingestion and rapid absorption, they all result in circulating benzoic acid.

SC

RI PT

53

59

Benzoic acid and its salts have a long history of safe use as food additives. A number of scientific and

61

regulatory bodies worldwide have conducted comprehensive assessments of available acute and chronic

62

toxicity, reproductive/developmental toxicity, carcinogenicity, and genotoxicity studies, and concluded

63

that these substances have low toxicity and are generally safe for human consumption (Adams et al.,

64

2005; IOMC, 2000; JECFA, 1997; OECD, 2001; Nair, 2001; SCCP, 2005; EFSA, 2016). For example,

65

benzoic acid and its salts are "generally recognized as safe" (GRAS) for current use in food by the United

66

States Food and Drug Administration (US FDA, 1973a,b, 2016), and Food Standards Australia New

67

Zealand (FSANZ) evaluated the safety of benzoates in foods and concluded that "there were no issues

68

with benzoates" (FSANZ, 2016).

TE D

EP

69

M AN U

60

In addition to its use as a preservative, sodium benzoate was used historically to assess liver function (by

71

measuring hippuric acid formation in urine) with oral doses on the order of 5,000 mg (Borgstrom, 1949).

72

Sodium benzoate is currently used clinically to treat urea cycle disorders because it can increase waste

73

nitrogen excretion and improve clinical outcomes (IOMC, 2000). Currently, there are three US FDA-

74

approved drugs that contain sodium benzoate as one of the main active ingredients (Ammonul®, a generic

75

version of Ammonul®, and Ucephan®) (US FDA, 2017). Therapeutic doses of sodium benzoate can be as

76

high as 250-500 milligrams per kilograms of body weight per day (mg/kg bw-day), equivalent to 211.9-

77

423.7 mg/kg bw-day benzoic acid (IOMC, 2000). Such treatment typically results in improved clinical

AC C

70

4

ACCEPTED MANUSCRIPT

outcomes with no serious side effects reported (IOMC, 2000). Notably, these human pharmaceutical uses

79

of sodium benzoate that result in systemic exposures up to nearly two orders of magnitude higher than

80

exposure levels from dietary intake (IOMC, 2000) have not been previously considered in any safety

81

evaluation of benzoic acid or its salts for use as a food preservative.

RI PT

78

82

The World Health Organization (WHO) Joint Expert Committee on Food Additives (JECFA) first

84

presented an ADI of 0-5 mg/kg bw1 for benzoic acid and sodium benzoate in 1962 based on a four-

85

generation rat study and a total UF of 100 (JECFA, 1962). This study was conducted in 1960, and no

86

adverse effects were observed at up to the highest dose tested (i.e., the no observed adverse effect level

87

[NOAEL] of 500 mg/kg bw-day) (Kieckebusch and Lang, 1960). The basis of the UF was not specified

88

in the 1962 report (JECFA, 1962). In 1997, JECFA maintained this ADI for benzoic acid and confirmed

89

its use for benzoic acid and its salts (calcium, potassium, and sodium), as well as for its precursors

90

(benzaldehyde, benzyl acetate, benzyl alcohol, and benzyl benzoate); the rationale for applying a UF of

91

100 was not discussed in this assessment (JECFA, 1997). In 2000, JECFA reaffirmed the group ADI of

92

0-5 mg/kg bw-day and specified that the total UF of 100 included 10 for database uncertainty and 10 for

93

interspecies differences (IOMC, 2000). It is unclear why JECFA did not consider a UF for intraspecies

94

variability at this time, as a default UF of 100 (based on a factor of 10 for interspecies differences and

95

another factor of 10 for intraspecies variability) has historically been used when JECFA derived an ADI

96

from a NOAEL based on animal toxicity data (IPCS, 2005, 2009).

AC C

EP

TE D

M AN U

SC

83

97 98

Recently, JECFA conducted a comprehensive dietary assessment on benzoate exposure in the general

99

population and determined that high-end dietary exposures to benzoates in children (1-17 years of age),

100

may exceed the upper bound of the current ADI (0-5 mg/kg bw-day) (JECFA, 2015, 2016).

101

Consequently, the 48th Codex Committee on Food Additives lowered the level of benzoates in beverages

102

by 50-75% in 2016 (Joint FAO/WHO Food Standards Programme, 2016). However, given the long 1

Traditionally, JECFA has developed ADIs as ranges, from 0 to an upper limit.

5

ACCEPTED MANUSCRIPT

103

history of safe use of benzoates as food additives and therapeutic agents, the current ADI may not

104

appropriately reflect the general safety of dietary exposures.

105 Current risk assessment guidelines, such as those from the WHO International Programme on Chemical

107

Safety (IPCS) and the US Environmental Protection Agency (US EPA), encourage the use of human and

108

animal data to develop adjustment factors (AFs) (e.g., chemical-specific adjustment factors [CSAFs] and

109

data-derived exploration factors [DDEFs]) to replace the default UF values, where possible (IPCS, 2005,

110

2009; US EPA, 2014a). Such factors allow for the incorporation of quantitative pharmacokinetic and

111

pharmacodynamic chemical-specific data on interspecies differences or human variability into the risk

112

assessment process. In doing so, these chemical-specific values enhance the assessment by more fully

113

utilizing the entirety of the dataset available for a particular substance and reducing the uncertainty

114

surrounding the ultimate safety criteria value. Development of data-derived AFs in place of default UFs

115

for health-based toxicity criteria derivation is becoming more of a standard practice, and such values are

116

increasingly incorporated into human health risk assessments by a variety of agencies. JECFA, for

117

example, has set a precedent for using such AFs in previous assessments and has subdivided the default

118

UFs in their evaluations of dioxins (JECFA, 2002) and methylmercury (JECFA, 2004).

119

assessments from US EPA incorporate pharmacokinetic data and modeling to adjust default uncertainty

120

values to establish chronic oral and inhalation reference concentrations for environmental exposures (e.g.,

121

for lead [US EPA, 2013], perfluorooctanoic acid [PFOA], and perfluorooctane sulfonate [PFOS] [US

122

EPA, 2014b,c]).

SC

M AN U

TE D

EP

Recent

AC C

123

RI PT

106

124

WHO/IPCS guidance divides UFs into specific pharmacokinetic and pharmacodynamic components. As

125

illustrated in Figure 2, based on empirical analyses, the interspecies UF default value of 10 can be

6

ACCEPTED MANUSCRIPT

126

subdivided into a factor of 4 for pharmacokinetics and 2.5 for pharmacodynamics, while the intraspecies

127

UF default value of 10 can be subdivided into two equal subfactors of 3.16 (IPCS, 2005).2

128 The database on the safety, toxicity, and pharmacokinetics of benzoic acid and its salts has expanded

130

considerably since the ADI was developed in the early 1960s. We evaluated the available animal and

131

human pharmacokinetic data, and human clinical data to determine whether the current default UFs

132

should be updated, and, if so, to develop CSAFs for deriving the ADI for benzoic acid and its salts.

SC

RI PT

129

133

2

135

We conducted literature searches for the available pharmacokinetic studies of benzoic acid and its salts in

136

humans, experimental animals, and in vitro, as well as other human clinical studies, using PubMed and

137

Scopus to identify clinical and laboratory studies published through October 2016. We also considered

138

studies of precursor compounds including benzyl alcohol, benzyl acetate, and benzaldehyde, which are

139

exclusively metabolized to benzoic acid in both humans and rodents. We also searched the Drugs@FDA

140

online database, which includes detailed chemistry, pharmacology, and toxicology profiles for drug

141

products approved by US FDA to identify any potentially unpublished studies of benzoates. In addition,

142

we searched Tox Planet, an online database that compiles all safety assessment documents for chemicals

143

from national and international regulatory and research agencies, to identify existing comprehensive

144

assessments for benzoates. We cross-referenced the bibliographies of review articles and assessment

145

documents to identify additional relevant studies.

TE D

EP

AC C

146

Methods

M AN U

134

2 Based on an empirical analysis of the variability between various animal species and humans for a variety of pharmacokinetic parameters, the default UF for interspecies differences of 10 is subdivided into a factor of 4 (100.6) for pharmacokinetic and 2.5 (100.4) for pharmacodynamic differences (Renwick, 1993; IPCS, 1994). However, the database for intraspecies differences is not considered sufficient to justify uneven subdivision of the default UF of 10. This UF is therefore divided evenly into factors of 100.5 (3.16), which is supported by a later analysis of various kinetic parameters for 60 chemicals in humans (Renwick and Lazarus, 1998).

7

ACCEPTED MANUSCRIPT

147

We evaluated and integrated the available pharmacokinetic literature on benzoic acid and its salts in

148

humans, animals, and in vitro to qualitatively compare interspecies (i.e., human and animal) and

149

intraspecies (i.e., age-related) pharmacokinetics for these substances.

150

evaluation, we selected key studies for the CSAF derivation analysis if they contained sufficient

151

quantitative information to allow for quantitative analysis of key pharmacokinetic parameters (e.g.,

152

plasma benzoate area under the curve [AUC], maximum biotransformation rate [Vmax], or clearance rate)

153

for humans and rats.

SC

RI PT

Based on this review and

154

In addition, we evaluated the human clinical studies of sodium benzoate with a specific focus on the

156

therapeutic doses and any associated observed/reported adverse effects. We determined whether the data

157

were sufficiently robust for identifying a NOAEL or deriving CSAFs for the pharmacodynamic

158

components of the UFs.

M AN U

155

159

3

Interspecies Comparison of Pharmacokinetics of Benzoic Acid and Its Salts

161

In its guidelines for risk assessment, WHO/IPCS specifies that the evaluation of qualitative and

162

quantitative differences in absorption, distribution, metabolism, and excretion (ADME) between

163

experimental animals and humans greatly informs hazard identification and risk characterization (IPCS,

164

2009). The uncertainty of extrapolating animal toxicity data to humans is reduced when the same ADME

165

processes occur in both species. In addition, identifying key ADME determinants and corresponding

166

differences in both species provides a mechanistic basis for estimating a CSAF. Thus, below are our

167

qualitative and quantitative evaluations of the robust pharmacokinetic data available in humans and

168

animals for benzoic acid, its salts, and its precursor compounds.

AC C

EP

TE D

160

169

8

ACCEPTED MANUSCRIPT

170

3.1

Qualitative Interspecies Comparison of Pharmacokinetics

171

Absorption.

172

completely absorbed from the gastrointestinal tract (Tremblay and Qureshi, 1993; US FDA, 1972). In

173

humans, peak plasma benzoic acid levels are reached 1-2 hours after oral ingestion of sodium benzoate

174

(Kubota et al., 1988; Kubota and Ishizaki, 1991) or after a 90-minute "bolus" intravenous (i.v.) infusion

175

of sodium benzoate (administered as Ammonul® (10% sodium benzoate/10% sodium phenylacetate) up to

176

5.5 g/m2 (US FDA, 2004a; MacArthur et al., 2004). In rats, peak benzoic acid plasma levels were

177

reached within 3 hours of administration after oral gavage administration of benzyl acetate (500 mg/kg

178

bw) (Yuan et al., 1995). In both humans and rats, administered doses of benzoic acid are recovered

179

almost entirely in urine as the metabolite hippuric acid, suggesting that essentially none of the

180

administered parent compound passes through the body unabsorbed (Kubota et al., 1988; Kubota and

181

Ishizaki, 1991; Bridges et al., 1970; Thabrew et al., 1980).

M AN U

SC

RI PT

The available information suggests that, after ingestion, benzoic acid is quickly and

TE D

182

Distribution. Following systemic exposure of benzoic acid, its salts, or precursor compounds, there is no

184

evidence of tissue accumulation or sequestration in humans (Bridges et al., 1970; Kubota et al., 1988;

185

Kubota and Ishizaki, 1991) or rats (reviewed by US FDA, 1972; FASEB, 1973; IOMC, 2000; OECD,

186

2001; JECFA, 1996; SCCP, 2005; EFSA, 2016). The observed volumes of distribution for benzoates in

187

human adults were 0.14-0.18 L/kg bw,3 which is consistent with a distribution primarily in the plasma

188

compartment in the absence of any sequestration in tissues (US FDA, 2004a). Similarly, the volume of

189

distribution measurements in rats administered sodium benzoate via i.v. injection (28-288 mg/kg bw)

190

ranged from 0.46 to 0.61 L/kg bw for different doses (with no apparent relationship with increasing

191

doses), which also supports a lack of tissue sequestration at all doses tested (Gregus et al., 1992).

AC C

EP

183

192

3

Converted from units of L/m2 based on a 60 kg adult and a body surface area of 1.62 m2 (Nair and Jacob, 2016).

9

ACCEPTED MANUSCRIPT

Metabolism. In humans and rats, benzoic acid and its salts are primarily metabolized in the liver by

194

conjugation with glycine, resulting in the formation of hippuric acid (Tremblay and Qureshi, 1993; US

195

FDA, 1972). The conversion of benzoate to hippurate also takes place in the kidneys, although to a lesser

196

extent than it does in the liver (Tremblay and Qureshi, 1993). Benzoic acid is rapidly converted to the

197

intermediate, benzoyl-CoA, which is subsequently conjugated with glycine in a reaction catalyzed by the

198

enzyme glycine N-acyltransferase (Tremblay and Qureshi, 1993; US FDA, 1972). Studies have shown

199

that the formation of hippuric acid is a saturable process following Michaelis-Menten kinetics and is

200

limited by glycine availability at high benzoate/benzoic acid doses (Tremblay and Qureshi, 1993; US

201

FDA, 1972; Adams et al., 2005; Kubota et al., 1988; Kubota and Ishizaki, 1991; MacArthur et al., 2004;

202

reviewed in IOMC, 2000 and JECFA, 1996). This is supported by studies showing that percent benzoic

203

acid clearance decreased as the dose increased (ranging from 40 to 160 mg/kg bw), while percent hippuric

204

acid clearance was stable with increasing dose. In these studies, peak plasma benzoic acid concentrations

205

increased with increasing dose while peak hippuric acid concentrations did not change. Multiple kinetic

206

studies conducted in rats demonstrate that there is a similar saturable process for benzoate metabolism at

207

high exposures (i.e., > 120 mg/kg bw) (Schwab et al., 2001; Gregus et al., 1992; Simkin and White, 1957;

208

Bray et al., 1951). Together, these data indicate that the saturable metabolism of benzoic acid to hippuric

209

acid due to saturation of the glycine conjugation pathway at high doses is the key pharmacokinetic driver

210

for potential benzoic acid toxicity.

EP

TE D

M AN U

SC

RI PT

193

AC C

211 212

After saturation of the glycine conjugation pathway and depletion of glycine after high doses of sodium

213

benzoate (e.g., > 500 mg/kg bw), the benzoyl glucuronide metabolite appears; however, this only occurs

214

at high doses that are orders of magnitude greater than typical levels in the diet (US FDA, 1972; IOMC,

215

2000; Kubota and Ishizaki, 1991; as reviewed in Tremblay and Qureshi, 1993 and Adams et al., 2005).

216

Thus, saturation of the glycine pathway does not appear to be significant in the context of typical human

217

exposures.

218

10

ACCEPTED MANUSCRIPT

Excretion. The amount of benzoate excreted as hippuric acid is similar in rats and humans: 75-100% of

220

benzoic acid is excreted as hippuric acid within 6-24 hours in both species (IOMC, 2000; Bridges et al.,

221

1970; Thabrew et al., 1980). Benzoic acid sometimes appears in its free form in urine (Bridges et al.,

222

1970).

RI PT

219

223

Overall, the general pharmacokinetic pathways of benzoic acid and its salts are very similar in humans

225

and rats (summarized in Table 1).

SC

224

226

3.2

228

Evaluation of interspecies pharmacokinetic differences requires careful consideration and selection of key

229

kinetic dose metrics. As recommended by WHO/IPCS guidance (IPCS, 2005, 2010), these considerations

230

include the toxicologically active substance (e.g., parent chemical or metabolite), internal dose levels,

231

exposure route, duration, intensity (e.g., peak or average dose), and biological matrix (e.g., blood, target

232

tissue) that are consistent with the mode of action of the substance of interest. In this case, the active

233

substance is benzoic acid, and the systemic exposures resulting from chronic, oral exposure to benzoic

234

acid and its salts are most relevant for their use as food additives. While the exact mode of action of

235

benzoic acid toxicity is not certain, the suspected mode of action is related to the saturation of the glycine

236

conjugation pathway at very high doses in both animals and humans.

TE D

EP

AC C

237

CSAF Derivation for Interspecies Differences in Pharmacokinetics

M AN U

227

238

It is important to note that while toxicity is not observed at levels below saturation, evidence of saturation,

239

alone, is not evidence of toxicity; studies in rats and humans demonstrate some level of saturation (as

240

indicated by AUC values that increase in a greater than dose-proportional manner) at doses greater or

241

equal to 120 and 160 mg/kg bw, respectively. However, no evidence of toxicity is observed in either

242

species at or significantly above these doses. Our review of the literature indicates lowest observed

243

adverse effect levels (LOAELs) of 800 mg/kg bw in humans and greater than or equal to 750 mg/kg bw in

11

ACCEPTED MANUSCRIPT

rats (discussed further in Section 5) (IOMC, 2000).

245

metabolism to hippuric acid can be viewed as the primary detoxification pathway, the pharmacokinetic

246

determinants that best characterize the potential toxicity of benzoic acid and its salts are the overall

247

systemic exposure to benzoic acid over time and the metabolic capacity for glycine conjugation.

248

3.2.1

249

For these reasons, and because benzoic acid

RI PT

244

CSAF Derivation

To derive a CSAF for interspecies differences in pharmacokinetics (i.e., PKinterCSAF), we considered

251

studies that reported sufficient quantitative data on several dose metrics, such as plasma AUC, clearance,

252

and maximum biotransformation rate (Vmax). We selected AUC as the key dose metric because it

253

provides a specific comparison of chronic and systemic benzoic acid exposure based on in vivo human

254

and animal data, is reflective of metabolic saturation with increasing dose, and is also a conservative

255

metric due to the greater human variability in AUC compared to other metrics (e.g., Cmax4) (IPCS, 2005).

M AN U

SC

250

TE D

256

We primarily relied on pharmacokinetic data obtained from studies of i.v. exposures to sodium benzoate

258

or benzoic acid in humans and rats to derive the PKinterCSAF because extensive pharmacokinetic data for

259

oral exposure are not available in rats. For the human study (US FDA, 2004a), we first converted i.v.

260

doses of sodium benzoate (in units of mg/m2) to oral doses (in units of mg/kg bw) by dividing by a

261

conversion factor of 37, which is based on the default body surface area for a 60 kg adult5 (US FDA,

262

2005). For consistency across studies, we converted the administered sodium benzoate dose in animals

263

and humans to an internal benzoic acid equivalent dose for each dose group in each study by multiplying

264

the administered dose (in units of mg/kg bw) by the ratio of the molecular weights of benzoic acid

265

(122.123 g/mol) to sodium benzoate (144.11 g/mol).

AC C

EP

257

266 4

Cmax is the maximum concentration of a drug achieved in a specified body compartment (usually plasma) after the drug's administration and before the administration of a second dose. 5 The WHO standard body weight of an adult human is 60 kg (IPCS, 2009).

12

ACCEPTED MANUSCRIPT

As discussed above, absorption after oral exposure is rapid and essentially 100% both in humans and in

268

rats. In addition, available data in healthy adult humans indicate that mean AUCs after oral exposure are

269

similar, if not lower, to those obtained after i.v. bolus infusions of comparable doses of benzoic acid (e.g.,

270

385.4 µg × hour/mL at 80 mg/kg bw and 564.6 µg × hour/mL at 86 mg/kg bw,6 respectively) (Kubota and

271

Ishizaki, 1991; US FDA, 2004a). Therefore, we determined that the pharmacokinetics for benzoic acid

272

and its salts are similar after exposure via the i.v. and oral routes and that the use of pharmacokinetic data

273

obtained from i.v. exposures in humans and animals is appropriate for PKinterCSAF derivation.

SC

RI PT

267

274

We identified two independent pharmacokinetic studies in healthy human volunteers conducted by US

276

FDA with the drug Ammonul® (10% sodium benzoate/10% sodium phenylacetate) (US FDA, 2004a;

277

MacArthur et al., 2004), from which an AUC for plasma benzoate could be derived.

278

administration, phenylacetate conjugates with glutamine in the liver and kidneys to form

279

phenylacetylglutamine, via acetylation, and is then excreted by the kidneys. The metabolic process of

280

phenylacetate is distinct from that of benzoate and the pharmacokinetics of benzoate and phenylacetate in

281

combination are similar to those for the chemicals when administered individually (e.g., Kubota and

282

Ishizaki, 1991; reviewed in MacArthur et al., 2004). Therefore, the pharmacokinetic data of these two

283

substances can be evaluated independently.

Upon

EP

TE D

M AN U

275

284

In one study, 20 healthy adult volunteers (9 males, 11 females) were administered Ammonul® via i.v.

286

infusion at doses of 5.5 g/m2 or 3.75 g/m2 (approximately 150 or 100 mg/kg bw sodium benzoate,

287

respectively) 7 and pharmacokinetic measurements were obtained after an initial 90-minute loading

288

infusion (US FDA, 2004a). Sodium benzoate demonstrated saturable elimination kinetics with decreased

289

clearance with increased dose. The authors also noted a greater than dose-proportional increase in both

290

AUC and Cmax for benzoic acid. The AUC for the primary metabolite, hippurate, increased proportionally

AC C

285

6

3.75 g/m2 × 1,000 × (1/37) = 101 mg/kg bw sodium benzoate × (122.123/144.11) = 86 mg/kg bw benzoic acid equivalent dose. A dose of 5.5 g/m2 caused severe emesis in the first three patients; the dose was then reduced to 3.75 g/m2 for the remaining 17 subjects.

7

13

ACCEPTED MANUSCRIPT

with dose, though the increase in Cmax was less than dose-proportional. The authors noted no marked sex

292

differences. In the second study, similar results were obtained for parent compounds and metabolites

293

with escalating dose (1, 2, and 4 g/m2) after a 90-minute administration using similar methods (US FDA,

294

2004a). Administered i.v. doses, benzoic acid equivalent doses, and dose-normalized AUCs for these

295

studies are presented in Table 2.

RI PT

291

296

The dose-normalized AUCs increased with increasing doses, indicating that the clearance of circulating

298

benzoic acid is nonlinear (equal dose-normalized AUCs would be expected for substances with linear

299

elimination kinetics). Standard errors (i.e., standard deviations divided by the square root of the subject

300

numbers) for all except the highest dose group are less than or equal to 10-11% of mean AUC values, and

301

thus, mean AUCs from these studies are considered adequate to provide an accurate measure of the

302

central tendency (IPCS, 2005). The largest inter-individual variability was observed in the highest dose

303

group, as indicated by the standard deviations in the AUC values and standard error of approximately

304

17% of the mean AUC. This is likely a result of the relatively small sample size (n = 3) and general

305

human variability in AUC measurements.

M AN U

TE D

306

SC

297

Gregus et al. (1992) conducted the most relevant animal study in adult male Wistar rats with sufficient

308

quantitative pharmacokinetic data for calculating the AUCs for plasma benzoate. As with the human

309

data, we converted administered i.v. doses from Gregus et al. (1992) from mmol/kg to oral doses (in units

310

of mg/kg bw) based on a molecular weight of sodium benzoate of 144.11 g/mol, and converted sodium

311

benzoate to benzoic acid equivalent doses based on the molecular masses of sodium benzoate and benzoic

312

acid (122.123 g/mol). Administered i.v. doses, benzoic acid equivalent doses, and dose-normalized

313

AUCs for each dose group in this study are presented in Table 3. Similar to humans, the clearance of

314

circulating benzoic acid in rats is nonlinear, as indicated by increasing dose-normalized AUCs with

315

increasing doses, and standard errors of mean AUC values indicate that sample size is acceptable as they

316

are all less than 5% of the mean.

AC C

EP

307

14

ACCEPTED MANUSCRIPT

317 We then compared dose-normalized AUCs at the most comparable doses of benzoic acid from the US

319

FDA human studies (US FDA, 2004a) and the rat study (Gregus et al., 1992) to calculate the

320

PKinterCSAF. Candidate values derived this way range from 2.1 at the lowest dose (i.e., 23 mg/kg bw and

321

24 mg/kg bw, respectively) to 6.7 at the highest dose (i.e., 126 mg/kg bw and 122 mg/kg bw,

322

respectively). Owing to the large inter-individual variability in the AUC results from the highest dose

323

group in humans from US FDA (2004a), we do not consider the high-dose comparison to be reliable for

324

CSAF derivation. Based on a comparison of AUCs at dose levels that are most comparable to the ADI

325

and levels experienced in the general population (i.e., the lowest dose groups, here), we ultimately

326

selected a CSAF of 2 for interspecies differences in pharmacokinetics.

M AN U

SC

RI PT

318

327 328

3.2.2

Alternative Approaches

Additional analyses further support a reduction of the pharmacokinetic component of the interspecies UF

330

and, together with the qualitative assessment presented in Section 3.1, indicate that the CSAF derived in

331

Section 3.2.1 is conservative.

332

TE D

329

First, to further inform potential species-related differences in glycine conjugation—a key step in the

334

metabolic pathway of benzoic acid—we evaluated in vitro metabolic studies of the maximum rate of

335

hippuric acid formation (i.e., Vmax) from benzoic acid in adult human and rat liver tissues. We compared

336

interspecies Vmax values because the saturation of this biotransformation pathway is presumably a vital

337

governing process in the clearance and potential toxicity of benzoic acid. Thus, understanding any

338

species-related differences in this pathway is essential to adequate evaluation of potential interspecies

339

pharmacokinetic differences for benzoic acid and its salts.

AC C

EP

333

340

15

ACCEPTED MANUSCRIPT

In human liver specimens, an average Vmax value of 204 nmol/min/mg liver was estimated in one study

342

(Pacifici et al., 1991), and the rate of hippuric acid formation at a high concentration (200 µM) was

343

reported to be 254 nmol/min/mg liver in another (Temellini et al., 1993). Because Temellini et al. (1993)

344

only used a single dose to determine this value, it is unclear whether the rate of hippuric acid formation

345

they calculated reflects a state where the enzyme is fully saturated (i.e., Vmax); however, plots of hippuric

346

acid versus benzoic acid concentration from Pacifici et al. (1991) suggest that the 200 µM dose by

347

Temellini et al. (1993) was in the saturated range. In rats, two studies examined glycine conjugation in

348

adult rat liver samples using single-pass perfused adult rat liver preparations and reported Vmax values of

349

101 and 78 nmol/min/g liver, respectively (Chiba et al., 1994; Schwab et al., 2001).

M AN U

SC

RI PT

341

350 351

Comparing the Vmax values from rat liver studies to those observed in the human liver studies yields a

352

mean ratio of 0.4 (Table 4). This indicates that humans have an approximately 2.5-fold greater metabolic

353

capacity to conjugate benzoic acid than rats at levels spanning metabolic saturation.

TE D

354 355

We next explored chemical-specific interspecies differences through the use of default allometric scaling

356

methods for a comparison to the CSAF.

357

pharmacokinetic UF default value is meant to address two different concepts: the known and expected

358

differences across species that are due to differences in basal rates of metabolism (i.e., rodents have a

359

faster metabolism and elimination of compounds relative to humans), and unusual and unanticipated

360

differences when humans are more sensitive than rodents even beyond the expected differences in

361

metabolism (e.g., due to differences in metabolic pathways or enzyme efficiencies). We compared AUCs

362

between humans and rats after allometric scaling of the rat-administered doses to explore whether there

363

were differences in benzoic acid metabolism across species beyond what would be expected based on

364

differences in basal metabolism. For this analysis, we converted the benzoic acid equivalent doses to

365

human equivalent doses (HEDs) via body weight scaling to the ¾ power (US EPA, 2011; Rhomberg and

366

Lewandowski, 2006) to place the rodent and human doses on a "basal metabolism equivalent" basis,

AC C

EP

When applied to derive a safety value, the interspecies

16

ACCEPTED MANUSCRIPT

367

which allows for a more direct and accessible evaluation of chemical-specific differences in metabolism

368

across species (Table 5).8

369 If metabolism between species were identical (i.e., same pathways, same saturation potential, same

371

enzyme affinities) and fully explained by the default allometric scaling approach, then the ratios of the

372

benzoic acid equivalent dose and HED presented in Table 6 should be equal or close to 1. Here, the ratios

373

range between 0.3 and 0.4, which indicates that there are metabolic differences between humans and rats

374

beyond the expected difference due to different body size or metabolic rate and that the default scaling is

375

too conservative.

M AN U

SC

RI PT

370

376

From these data, we can derive a DDEF for benzoic acid, consistent with US EPA (2014a) methodology,

378

by combining the allometric scaling value reflective of inherent slower metabolism in humans (i.e., by

379

dividing by 0.269 or multiplying by its inverse, 3.8) with the chemical-specific metabolic factor derived in

380

Table 6 that suggests more efficient benzoic acid metabolism in humans than rats (i.e., 0.3) to yield an

381

overall DDEF for benzoic acid and its salts of 1.1. This essentially means that humans' apparent greater

382

ability to metabolize benzoic acid almost completely offsets the rat's higher basal metabolism, and

383

therefore there is no significant difference between the overall pharmacokinetics of benzoic acid and its

384

salts in animals and humans.

EP

AC C

385

TE D

377

386

Lastly, WHO/IPCS guidance (2005) states that one can compare AUC values in animals at the dose used

387

for risk assessment (e.g., the NOAEL for the point of departure [POD]) to the AUC in humans at doses

388

expected after typical human exposures. While we do not have data in rats at the POD of 500 mg/kg bw,

389

metabolic saturation is observed at the higher dose groups in Gregus et al. (1992) of 122.1 and 244.2 8 As per US EPA (2011) guidance: HED = Dose (mg) × (BWa/BWh)(3/4), where Dose is the administered dose in benzoic acid equivalents (in units of mg); BWh is 60 kg, the standard adult human body weight (IPCS, 2009); and BWa is 0.3 kg, the average rat body weight based on data reported in Gregus et al. (1992). While this scaling approach was developed based on oral routes of exposure, its use is acceptable here because sodium benzoate is still extensively metabolized in both species and AUCs for sodium benzoate administration are comparable after oral and i.v. administration. 9 Scaling factor = (BWa/BWh)¼ = (0.3 kg/70 kg)¼ = 0.26 (US EPA, 2011).

17

ACCEPTED MANUSCRIPT

mg/kg bw benzoic acid, and thus reflects the nonlinear pharmacokinetics expected at higher doses closer

391

to the POD. A comparison of AUC values at these top doses in rats to the AUC in humans at the level

392

closest to the upper end of expected human exposures (i.e., the 22.9 mg/kg bw benzoic acid group) yields

393

human to rat ratios10 of 0.2-0.5. In other words, humans are capable of clearing the upper end of expected

394

human exposures of benzoic acid and its salts approximately 2 to 5 times faster than the rat clears the high

395

levels closer to the NOAEL used for basis of the ADI. This indicates that additional adjustment of the

396

POD for interspecies pharmacokinetic extrapolation is unnecessary for the purposes of the ADI and thus

397

should be reduced to 1.

SC

RI PT

390

3.2.3

399

Discussion

M AN U

398

We derived a PKinterCSAF of 2 in accordance with WHO/IPCS methodology from comparisons of AUC

401

values based on in vivo human data from US FDA (2004a) and rat data from Gregus et al. (1992). The

402

human data were obtained from a high-quality clinical study with a sufficient number of subjects in the

403

relevant dose groups. The key animal study had a sufficient number of rats, used appropriate controls,

404

and provided adequate documentation of the experimental procedures and results. While the dose ranges

405

used in these studies are lower than that of the POD for the ADI, these dose ranges are greater than

406

typical exposures in the general population and the current and proposed ADIs. Further, metabolic

407

saturation was observed in both species at the doses utilized in the available pharmacokinetic studies, and

408

these levels (including that of the POD) are not associated with toxicity in either species.

409 410

Several lines of evidence indicate that a PKinterCSAF of 2 is conservative, and that a value of 1 might be

411

more appropriate:

AC C

EP

TE D

400

412

10

At 122.1 mg/kg bw: Human AUCNorm/Rat AUCNorm = (Human AUC/Dose)/(Rat AUC/Dose) = (20.3/22.9)/(1,100/244.2) = 0.2. At 244.1 mg/kg bw: Human AUCNorm/Rat AUCNorm = (Human AUC/Dose)/(Rat AUC/Dose) = (20.3/22.9)/(239.5/122.1) = 0.5.

18

ACCEPTED MANUSCRIPT

413



414 415

Vmax comparisons from in vitro liver tissues indicate that humans more efficiently metabolize benzoic acid than rats;



AUC comparisons indicate that humans are capable of clearing the upper end of expected human exposures of benzoic acid faster than the rat clears the higher, saturated levels closer to the

417

NOAEL that is the basis of the ADI; and

418



RI PT

416

AUC ratios between humans and rats after allometric scaling of the rat-administered doses indicate that humans' apparent greater ability to metabolize benzoic acid almost completely

420

offsets the rat's higher basal metabolism, and that additional pharmacokinetic adjustment is

421

unnecessary.

M AN U

SC

419

422

We also note that Hoffman and Hanneman (2017) recently reported that physiologically based

424

pharmacokinetic (PBPK) modelling indicated human-to-rat margin-of-exposure (MOE) ratios of 0.3 to

425

0.4 based on predicted plasma concentrations of benzoic acid after simulated repeated daily oral

426

exposures. The authors ultimately concluded that benzoic acid is cleared more quickly in humans than

427

rats at a steady-state or AUC-based exposure level, and that the PKinterCSAF could be reduced to 1 for

428

ADI derivation.

EP

429

TE D

423

In conclusion, a PKinterCSAF of 2 is consistent with WHO/IPCS methodology, and several lines of

431

evidence indicate that it is conservative.

432

PKinterCSAF of 2 should be applied to the default interspecies UF for ADI derivation.

433

AC C

430

Given its scientific support and conservative nature, a

19

ACCEPTED MANUSCRIPT

434 435

4

Intraspecies Comparison of Pharmacokinetics of Benzoic Acid and Its Salts in Humans

Because of its history of therapeutic uses to treat urea cycle disorders (IOMC, 2000), clinical studies of

437

therapeutic drugs containing sodium benzoate involving infants, children, and adults are available and

438

provide robust pharmacokinetic information that can be used to evaluate potential sensitivities across life

439

stages. Comparable pharmacokinetic data for other potentially sensitive populations (e.g., the elderly or

440

those with specific polymorphisms that may affect metabolism) are not available.

SC

RI PT

436

M AN U

441

The available pharmacokinetic studies in humans indicate that ADME patterns of benzoic acid and its

443

salts are generally similar in children and adults. However, quantitative pharmacokinetic data in neonates

444

and infants are limited to those with severe genetic defects in urea synthesis (Green et al., 1983; Brusilow

445

et al., 1984; Simell et al., 1986). They therefore are not appropriate for a robust quantitative evaluation of

446

such differences for children in the general population without these conditions. For these reasons, and

447

because data on other potentially sensitive populations are not available, we did not derive an CSAF for

448

the intraspecies pharmacokinetic component of the UF. The available data support the use of a default

449

intraspecies pharmacokinetic UF value of 3.16 until more robust data are available. Below, we provide a

450

qualitative discussion of potential age-related pharmacokinetic differences for benzoic acid and its salts

451

based on available pharmacokinetic information in humans.

EP

AC C

452

TE D

442

453

Absorption. US FDA (2004b) noted that peak plasma benzoate levels from seven children (aged 3-26

454

months) following administration of 250-500 mg/kg bw of sodium benzoate (equivalent to 212-424

455

mg/kg bw of benzoic acid) for 1-2 hours were qualitatively similar to those in adults. In a clinical study

456

of children with inborn errors of urea synthesis, Brusilow et al. (1984) documented the time course of

457

plasma benzoate concentrations in two infants (aged 5 months and 12 months) after receiving a single

20

ACCEPTED MANUSCRIPT

dose of sodium benzoate (250 mg/kg bw) and sodium phenylacetate (250 mg/kg bw) via i.v.

459

administration. The plasma benzoate concentrations in these two infants rose quickly and peaked at 2.5

460

mmol/L and 3.2 mmol/L at 2.7 and 1 hours, respectively, following i.v. infusion. Similarly, Simell et al.

461

(1986) reported peak plasma benzoic acid concentrations of 5.1-6.2 mmol/L at approximately 2 hours

462

following i.v. administration of 2 mmol/kg bw (approximately 288 mg/kg bw) sodium benzoate plus L-

463

alanine (6.6 mmol/kg) to five Finnish children aged 2.8-12.6 years (mean of 8.8) with a urea cycle

464

disorder (lynsinuric protein intolerance).11

SC

RI PT

458

465

In a study of four neonates with hyperammonemia administered 3.5 mmol/kg bw-day (460 mg/kg bw-

467

day) sodium benzoate via i.v. injection, Green et al. (1983) found large inter-patient variability, observing

468

an eight-fold range in serum benzoate concentrations (mean ± SD: 7.38 ± 6.39 mmol/L; range: 2.14-16.0

469

mmol/L) after sodium benzoate administration for at least 24 hours. These levels are higher than those

470

observed in infants and children and may be a result of a small sample size and/or decreased metabolism

471

or excretion in these health-compromised neonates.

TE D

M AN U

466

472

Overall, the available information in infants and children demonstrates that peak plasma benzoic acid

474

levels are achieved between 1 and 3 hours after i.v. administration, which is similar to that of adults (1-2

475

hours; see Section 3.1). Data on neonates are limited and may not currently support assuming the same

476

degree of absorption as seen in adults.

AC C

EP

473

477 478

Distribution. There are limited data on the distribution of benzoic acid and its salts in children. The

479

mean volume of distribution reported for neonates was 0.14 L/kg bw with an SD of 0.07 L/kg bw (Green

480

et al., 1983), which is almost identical to those reported for adults (US FDA, 2004a; MacArthur, 2004)

481

and is consistent with a lack of tissue sequestration. No data are available for infants and older children.

482 11

Data obtained from figures provided in Brusilow et al. (1984) and Simell et al. (1986) with the use of GetData Software.

21

ACCEPTED MANUSCRIPT

Metabolism. Evidence regarding whether a particular age group may be more sensitive than others with

484

respect to metabolic capability remains equivocal. In neonates, some studies reported reduced glycine

485

conjugation capacity after i.v. exposure to benzoate (reviewed in Dorne et al., 2005), while others

486

reported capacities after oral exposure that are similar to those of adults (Gow et al., 2001). In two infants

487

aged 5 months and 12 months, sodium benzoate appeared to be completely metabolized within 20 and 10

488

hours of i.v. infusion, respectively (Brusilow et al., 1984), potentially supporting an age-dependent effect

489

on benzoate metabolism in infants less than 1 year of age. The wide variability of benzoate metabolism

490

in infants and neonates has been suggested to indicate the immaturity of the acylation system in the liver

491

and kidneys (Batshaw et al., 2001), as studies have shown that mitochondrial glycine N-acyltransferase

492

activity in the liver varied from 5% to 40% of peak activity between birth and 7 months of age, and peak

493

activity was observed by 18 months of age (Mawal et al., 1997, as reviewed in Batshaw et al., 2001).

M AN U

SC

RI PT

483

494

In older children (i.e., 1 year of age or older), some studies reported higher rates of metabolic clearance

496

compared to healthy adults for a number of metabolic pathways, including for glycine conjugation

497

(e.g., Dorne et al., 2005), but other studies reported no significant differences among the mean rate of

498

glycine conjugation in children, healthy adults, or the elderly12 (e.g., Dorne et al., 2004; reviewed in

499

Badenhorst et al., 2013). The latter finding is supported by evidence indicating that adult-level kidney

500

filtration rates (and thus typical clearance rates) are achieved by approximately 7 months of age in

501

humans (Besunder et al., 1988).

AC C

EP

TE D

495

502 503

Further, Dorne et al. (2005) analyzed metabolic pathway-related UFs for a number of compounds.

504

They determined that the UFs for the pharmacokinetic component of intraspecies variability for glycine

505

conjugation pathway were consistently less than the default value of 3.16 for children, younger adults,

12

Dorne et al. (2004) stratified the age groups as follows: neonates (< 1 month old), infants (> 1 month to < 1 year old), children (> 1 to < 16 years old), healthy adults (16 to < 70 years old), and elderly (healthy adults > 70 years old).

22

ACCEPTED MANUSCRIPT

506

and older adults. Specifically, they reported that the 99th percentile UFs for adults and children were 1.6

507

and 1.8, respectively.13

508 Excretion. As discussed in Section 3.1, i.v. and orally administered sodium benzoate or benzoic acid is

510

almost completely excreted as hippuric acid in adult volunteers. The same is observed in studies of

511

children, ranging from neonates (i.e., 3 days old) to 12.6 years of age (Simell et al., 1986; Green et al.,

512

1983).

SC

RI PT

509

513

Taken together, the evidence indicates that the overall ADME for adults and infants, and children are

515

similar. Metabolic capacity for benzoate may be limited in neonates because of their immature acylation

516

systems, but human variability in benzoate pharmacokinetics appears to be negligible after the age of 1

517

year. The clinical studies in infants and neonates are limited in that the study populations were of poor

518

health status and likely had impaired renal function and clearance as well as limited benzoate conjugation

519

capacity (reviewed by Batshaw et al., 2001; Brusilow et al., 1984; Simell et al., 1986). It is worth noting

520

that the doses given in these studies are quite large compared to potential dietary intake. Healthy

521

neonates and infants in the general population have little, if any, potential dietary exposure to benzoic

522

acid or its salts, so their lower metabolizing rates should not substantially contribute to uncertainties in

523

deriving the ADI for benzoic acid and its salts. Also, the JECFA dietary assessment for benzoate did not

524

include children younger than 1 year old (JECFA, 2015).

525

AC C

EP

TE D

M AN U

514

526

5

527

As discussed above, studies in humans and laboratory animals demonstrate that benzoate metabolism to

528

hippurate by glycine conjugation in the liver and excretion in urine diverts nitrogen from urea production,

529

and thus provides an alternative pathway for the elimination of waste nitrogen (Tremblay and Qureshi, 13

Evaluation of Human Clinical Data of Sodium Benzoate

This analysis was based on data from various substrates that undergo glycine conjugation and is not specific to benzoate.

23

ACCEPTED MANUSCRIPT

1993). Because of this, sodium benzoate has been used as a therapeutic agent to treat hyperammonemia

531

(Tremblay and Qureshi, 1993). More recently, sodium benzoate has also been tested as a treatment for

532

certain neuropsychological conditions (Lane et al., 2013) due to its capability in inhibiting D-amino acid

533

oxidase, a flavoenzyme of peroxisomes in the central nervous system, via benzoyl CoA, the common

534

intermediate to hippuric acid for benzoate.

RI PT

530

535

We evaluated available clinical studies of sodium benzoate, including case reports (Trijbels et al., 1974;

537

Brusilow et al., 1979; Kodama et al., 1981; Brubakk et al., 1982; Mizutani et al., 1983; Takeda et al.,

538

1983; Call et al., 1984; Qureshi et al., 1984; Van de Bor et al., 1984; Letarte et al., 1985; McCormick et

539

al., 1985; Sharp and Lang, 1987; Walter et al., 1992), case series (Brusilow et al., 1980, 1984; Batshaw

540

and Brusilow, 1980; Batshaw et al., 1981, 1982; Watson et al., 1985; Sakuma et al., 1992; Tuchman et

541

al., 1992; Feoli-Fonseca et al., 1996; Maestri et al., 1996; Enns et al., 2007), and clinical trials

542

(Mendenhall et al., 1986; Uribe et al., 1990; Sushma et al., 1992; Lane et al., 2013; Lin et al., 2014), to

543

determine treatment-related adverse effects in these studies associated with doses above the current ADI.

544

The clinical studies involved a number of individuals across life stages, including infants, small children,

545

adolescents, adults, and older adults. Therapeutic doses of sodium benzoate varied considerably, ranging

546

from 4.2 to 500 mg/kg bw-day (equivalent to 3.6 to 423.7 mg/kg bw-day benzoic acid), and were well

547

tolerated in the study subjects with no apparent adverse effects. Observed toxicity, such as vomiting and

548

irritation, occurred only after accidental exposure at doses around 800 mg/kg bw-day (equivalent to 677.9

549

mg/kg bw-day benzoic acid) and ceased once the exposure was discontinued. These studies are described

550

in detail in Table S1.

M AN U

TE D

EP

AC C

551

SC

536

552

Most of the clinical studies were observational in nature (vs. experimental), involved only a small number

553

of subjects, and were not designed to evaluate long-term adverse effects in a systematic manner.

554

Therefore, these studies are not informative regarding a POD for deriving an ADI for benzoates and are

555

not suitable for the development of pharmacodynamic CSAFs. Setting this aside, these studies support a

24

ACCEPTED MANUSCRIPT

NOAEL of 500 mg/kg bw-day (423.7 mg/kg bw-day benzoic acid equivalent) and a lowest observed

557

adverse effect level (LOAEL) of 800 mg/kg bw-day (677.9 mg/kg bw-day benzoic acid equivalent) in

558

humans, which are comparable to the rodent NOAEL of 500 mg/kg bw-day of benzoic acid. This

559

provides supporting evidence – at least qualitatively – that the UF for interspecies differences in

560

pharmacodynamics should be reduced or possibly eliminated.

RI PT

556

561

In addition, because a therapeutic use of sodium benzoate is to treat genetic disorders of urea synthesis,

563

the recipients of treatment are generally children – the population of interest in the most recent JECFA

564

assessment on benzoic acid and its salts. The therapeutic doses are generally higher and the treatment

565

durations are often longer in children than in adults. In terms of safety, the clinical reports do not show

566

that children are more sensitive to the potential toxicity of sodium benzoate. In fact, as discussed in

567

Section 4, there is evidence suggesting that there are no significant differences in benzoate metabolism,

568

the key driver of potential benzoate toxicity, between adults and children over 1 year of age.

M AN U

SC

562

TE D

569 570

Similarly, we investigated studies of individuals with impaired liver function as a possible sensitive

571

population to the potential toxic effects of sodium benzoate.

572

(Mendenhall et al., 1986; Uribe et al., 1990; Sushma et al., 1992) conducted in patients with liver

573

dysfunction showed that there were no treatment-related adverse effects from doses of 5,000-10,000

574

mg/day of sodium benzoate (equivalent to approximately 70-140 mg/kg bw-day of benzoic acid),

575

indicating that sodium benzoate at doses 14-28 times the current ADI is well tolerated in this potentially

576

sensitive population.

EP

AC C

577

These clinical intervention studies

578

Observed NOAELs in children with urea cycle defects and in individuals with impaired liver functions

579

are similar and are at least one order of magnitude higher than the current ADI of 5 mg/kg bw-day,

580

indicating that benzoate exposures at levels considerably higher than the current ADI do not pose any

581

health risk to potentially sensitive populations. Collectively, the clinical evidence indicates that the UF

25

ACCEPTED MANUSCRIPT

582

for human variability in pharmacodynamics should be reduced or possibly eliminated, but as discussed

583

above, the data are not robust enough to derive a specific CSAF.

584

6

586

In 2000, JECFA identified a group ADI of 0-5 mg/kg bw-day for benzoic acid, its salts, and precursor

587

compounds based on a four-generation rat study conducted in 1960 in which no adverse effects were

588

observed at up to the highest dose tested (i.e., the NOAEL of 500 mg/kg bw-day), and a total UF of 100:

589

10 for interspecies differences and 10 for database uncertainty (IOMC, 2000). Many studies on benzoic

590

acid and its salts have been conducted since the ADI was first derived, including toxicity,

591

pharmacokinetic, and clinical studies. For example, the animal toxicity database for oral exposures to

592

benzoic acid, its salts, and related precursor compounds is extensive, and includes acute and chronic

593

toxicity, reproductive/developmental toxicity, carcinogenicity, and genotoxicity studies. These studies

594

consistently report a lack of toxicity at doses up to 1,000 mg/kg bw-day (SCCP, 2005; EFSA, 2016).

595

Human clinical data on sodium benzoate as a therapeutic agent provide additional evidence supporting the

596

lack of toxicity at doses up to 500 mg/kg bw-day. This indicates that the UF for database uncertainty is

597

not warranted and should be changed to 1.

SC

M AN U

TE D

EP

598

Proposed ADI for Benzoic Acid and Its Salts

RI PT

585

The pharmacokinetics of sodium benzoate and benzoic acid are nearly identical in humans and rats.

600

Quantitative comparisons of AUCs support a CSAF of 2 for interspecies differences in pharmacokinetics

601

(in place of the default value of 4), and is supported by a comparison of the maximum rate of metabolic

602

transformation in rat and human liver tissues in vitro. In addition, while information from studies of

603

healthy adults and children with urea cycle defects support similar pharmacokinetic profiles across adults

604

and children, additional studies are needed to assess the generalizability of these findings before a CSAF

605

for the pharmacokinetic component of the intraspecies UF can be derived.

AC C

599

606 26

ACCEPTED MANUSCRIPT

Clinical studies of sodium benzoate in children and adults also support reducing the UFs for interspecies

608

and intraspecies variability in pharmacodynamics, but the data are not sufficient to derive specific CSAFs.

609

Thus, the pharmacodynamic components of the UFs for interspecies and human variability should remain

610

at their default values of 2.5 and 3.16, respectively, until more robust data are available.

RI PT

607

611

Taken as a whole, the current safety database of benzoic acid and its salts supports a total UF of 50;

613

individual UF components are illustrated in Figure 3. Applying this revised total UF of 50 to the NOAEL

614

of 500 mg/kg bw-day established by JECFA results in an ADI of 0-10 mg/kg bw-day for benzoic acid and

615

its salts.

M AN U

SC

612

616 617

In this analysis, we demonstrated that the incorporation of chemical-specific information enhances ADI

618

derivation by reducing uncertainty and developing health-protective values based on the best available

619

science.

TE D

620

Acknowledgments

622

We thank Dr. Lorenz R. Rhomberg, Dr. Isaac Mohar, and Dr. Chester Rodriguez for their technical

623

insights, Ms. Jessica Goldstein and Ms. Carla Walker for editorial support, and Mr. Aaron Bazzle for

624

graphics support.

AC C

625

EP

621

626

Funding Source

627

The authors are employed by Gradient, a private environmental consulting firm. The work reported in

628

this paper was conducted during the normal course of employment. This work was funded by the

629

American Beverage Association. The authors have the sole responsibility for the writing, content, and

630

conclusions of this article.

27

ACCEPTED MANUSCRIPT

RI PT

References

Adams, TB; Cohen, SM; Doull, J; Feron, VJ; Goodman, JI; Marnett, LJ; Munro, IC; Portoghese, PS; Smith, RL; Waddell, WJ; Wagner, BM. 2005. "The FEMA GRAS assessment of benzyl derivatives used

SC

as flavor ingredients." Food Chem. Toxicol. 43(8):1207-1240. doi: 10.1016/j.fct.2004.11.014.

Badenhorst, CP; van der Sluis, R; Erasmus, E; van Dijk, AA. 2013. "Glycine conjugation: Importance in

M AN U

metabolism, the role of glycine N-acyltransferase, and factors that influence interindividual variation." Expert Opin. Drug Metab. Toxicol. 9(9):1139-1153. doi: 10.1517/17425255.2013.796929. Batshaw, ML; Brusilow, S; Waber, L; Blom, W; Brubakk, AM; Burton, BK; Cann, HM; Kerr, D; Mamunes, P; Matalon, R; Myerberg, D; Schafer, IA. 1982. "Treatment of inborn errors of urea synthesis:

TE D

Activation of alternative pathways of waste nitrogen synthesis and excretion." N. Engl. J. Med. 306(23):1387-1392. doi: 10.1056/NEJM198206103062303. Batshaw, ML; Brusilow, SW. 1980. "Treatment of hyperammonemic coma caused by inborn errors of

EP

urea synthesis." J. Pediatr. 97(6):893-900. doi: 10.1016/S0022-3476(80)80416-1. Batshaw, ML; MacArthur, RB; Tuchman, M. 2001. "Alternative pathway therapy for urea cycle

AC C

631

disorders: Twenty years later." J. Pediatr. 138(1 (Suppl.)):S46-S55. doi: 10.1067/mpd.2001.111836. Batshaw, ML; Painter, MJ; Sproul, GT; Schafer, IA; Thomas, GH; Brusilow, S. 1981. "Therapy of urea cycle enzymopathies: Three case studies." Johns Hopkins Med. J. 148(1):34-40.

28

ACCEPTED MANUSCRIPT

Besunder, JB; Reed, MD; Blumer, JL. 1988. "Principles of drug biodisposition in the neonate. A critical evaluation of the pharmacokinetic-pharmacodynamic interface (Part I)." Clin. Pharmacokinet. 14(4):189-

RI PT

216. doi: 10.2165/00003088-198814040-00001. Borgstrom, B. 1949. "Detoxification of benzoic acid by glycuronic acid under normal conditions and in liver disease." Acta Med. Scand. 133(1):7-18. doi: 10.1111/j.0954-6820.1949.tb09332.x.

SC

Bray, HG; Thorpe, WV; White, K. 1951. "Kinetic studies of the metabolism of foreign organic compounds. 1. The formation of benzoic acid from benzamide toluene, benzyl alcohol and benzaldehyde

M AN U

and its conjugation with glycine and glucuronic acid in the rabbit." Biochem. J. 48(1):88-96. Bridges, JW; French, MR; Smith, RL; Williams, RT. 1970. "The fate of benzoic acid in various species." Biochem. J. 118(1):47-51. doi: 10.1042/bj1180047.

Brubakk, AM; Teijema, LL; Blom, W; Berger, R. 1982. "Successful treatment of severe OTC

TE D

deficiency." J. Pediatr. 100(6):929-931. doi: 10.1016/S0022-3476(82)80518-0. Brul, S; Coote, P. 1999. "Preservative agents in foods. Mode of action and microbial resistance mechanisms." Int. J. Food Microbiol. 50 (1-2) :1-17.

EP

Brusilow, S; Tinker, J; Batshaw, ML. 1980. "Amino acid acylation: A mechanism of nitrogen excretion

AC C

in inborn errors of urea synthesis." Science 207(4431):659-661. doi: 10.1126/science.6243418. Brusilow, SW; Danney, M; Waber, LJ; Batshaw, M; Burton, B; Levitsky, L; Roth, K; McKeethren, C; Ward, J. 1984. "Treatment of episodic hyperammonemia in children with inborn errors of urea synthesis." N. Engl. J. Med. 310(25):1630-1634. doi: 10.1056/NEJM198406213102503. Brusilow, SW; Valle, DL; Batshaw, M. 1979. "New pathways of nitrogen excretion in inborn errors of urea synthesis." Lancet 314(8140):452-454. doi: 10.1016/S0140-6736(79)91503-4.

29

ACCEPTED MANUSCRIPT

Call, G; Seay, AR; Sherry, R; Qureshi, IA. 1984. "Clinical features of carbamyl phosphate synthetase-I deficiency in an adult." Ann. Neurol. 16(1):90-93. doi: 10.1002/ana.410160118.

RI PT

Chiba, M; Poon, K; Hollands, J; Pang, KS. 1994. "Glycine conjugation activity of benzoic acid and its acinar localization in the perfused rat liver." J. Pharmacol. Exp. Ther. 268(1):409-416.

Dorne, JL; Walton, K; Renwick, AG. 2004. "Human variability for metabolic pathways with limited data

42(3):397-421. doi: 10.1016/j.fct.2003.10.003.

SC

(CYP2A6, CYP2C9, CYP2E1, ADH, esterases, glycine and sulphate conjugation)." Food Chem. Toxicol.

M AN U

Dorne, JL; Walton, K; Renwick, AG. 2005. "Human variability in xenobiotic metabolism and pathwayrelated uncertainty factors for chemical risk assessment: A review." Food Chem. Toxicol. 43(2):203-216. doi: 10.1016/j.fct.2004.05.011.

Enns, GM; Berry, SA; Berry, GT; Rhead, WJ; Brusilow, SW; Hamosh, A. 2007. "Survival after

TE D

treatment with phenylacetate and benzoate for urea-cycle disorders." N. Engl. J. Med. 356(22):22822292. doi: 10.1056/NEJMoa066596.

European Food Safety Authority (EFSA). 2016. "Scientific Opinion on the re-evaluation of benzoic acid

EP

(E 210), sodium benzoate (E 211), potassium benzoate (E 212) and calcium benzoate (E 213) as food additives." Panel on Food Additives and Nutrient Sources (ANS). EFSA J. 14(3):4433. doi: 10.2903/

AC C

j.efsa.2016.4433. 110p.

Federation of American Societies for Experimental Biology (FASEB). 1973. "Evaluation of the Health Aspects of Benzoic Acid and Sodium Benzoate as Food Ingredients." Life Sciences Research Office. Report to US Food and Drug Administration (US FDA). NTIS PB-223837; SCOGS-7. 24p., January.

30

ACCEPTED MANUSCRIPT

Feoli-Fonseca, JC; Lambert, M; Mitchell, G; Melancon, SB; Dallaire, L; Millington, DS; Qureshi, IA. 1996. "Chronic sodium benzoate therapy in children with inborn errors of urea synthesis: Effect on carnitine metabolism and ammonia nitrogen removal." Biochem. Mol. Med. 57(1):31-36. doi: 10.1006/

RI PT

bmme.1996.0006. Food Standards Australia New Zealand (FSANZ). 2016. "Sulphites." June. Accessed at

SC

http://www.foodstandards.gov.au/consumer/additives/sulphite/Pages/default.aspx.

Gow, PJ; Ghabrial, H; Smallwood, RA; Morgan, DJ; Ching, MS. 2001. "Neonatal hepatic drug

M AN U

elimination." Pharmacol. Toxicol. 88(1):3-15. doi: 10.1034/j.1600-0773.2001.088001003.x. Green, TP; Marchessault, RP; Freese, DK. 1983. "Disposition of sodium benzoate in newborn infants with hyperammonemia." J. Pediatr. 102(5):785-790. doi: 10.1016/S0022-3476(83)80259-5. Gregus, Z; Fekete, T; Varga, F; Klaassen, CD. 1992. "Availability of glycine and coenzyme A limits

TE D

glycine conjugation in vivo." Drug Metab. Dispos. 20(2):234-240.

Hoffman, TE; Hanneman, WH. 2017. "Physiologically-based pharmacokinetic analysis of benzoic acid in rats, guinea pigs and humans: Implications for dietary exposures and interspecies uncertainty." Comput.

EP

Toxicol. doi: 10.1016/j.comtox.2017.06.002.

AC C

International Programme on Chemical Safety (IPCS). 1994. "Environmental Health Criteria No. 170: Assessing human health risks of chemicals: Derivation of guidance values for health-based exposure limits." WHO EHC 170. 73p.

31

ACCEPTED MANUSCRIPT

International Programme on Chemical Safety (IPCS). 2005. "Chemical-Specific Adjustment Factors for Interspecies Differences and Human Variability: Guidance Document for Use of Data in Dose/Concentration-Response Assessment." World Health Organization (WHO); International Labour

RI PT

Organisation; United Nations Environmental Programme. IPCS Harmonization Project Document No. 2. 100p. Accessed at http://www.who.int/ipcs/methods/harmonization/areas/uncertainty/en.

International Programme on Chemical Safety (IPCS). 2009. "Environmental Health Criteria 240:

SC

Principles and Methods for the Risk Assessment of Chemicals in Food." United Nations Environmental Programme; International Labour Organisation; World Health Organization (WHO) Environmental

M AN U

Health Criteria 240. 731p. Accessed at http://www.who.int/foodsafety/publications/chemical-food/en. International Programme on Chemical Safety (IPCS). 2010. "Characterization and Application of Physiological Based Pharmacokinetic Models in Risk Assessment." World Health Organization (WHO) IPCS

Harmonization

Project

Document

No.

9.

97p.

Accessed

at

TE D

http://www.inchem.org/documents/harmproj/harmproj/harmproj9.pdf.

Inter-Organization Programme for the Sound Management of Chemicals (IOMC). 2000. "Benzoic Acid and Sodium Benzoate." World Health Organization (WHO) Concise International Chemical Assessment

AC C

pdf?ua=1.

EP

Document (CICAD) 26. 52p. Accessed at http://www.who.int/ipcs/publications/cicad/cicad26_rev_1.

Joint FAO/WHO Expert Committee on Food Additives (JECFA). 1962. "Evaluation of the Toxicity of a Number of Antimicrobials and Antioxidants: Sixth Report of the Joint FAO/WHO Expert Committee on Food Additives." World Health Organization (WHO), Geneva, Switzerland, WHO Technical Report Series No. 228. 104p. Accessed at http://apps.who.int/iris/bitstream/10665/40518/1/WHO_TRS_228.pdf.

32

ACCEPTED MANUSCRIPT

Joint FAO/WHO Expert Committee on Food Additives (JECFA). 1996. "Toxicological Evaluation of Certain Food and Additives: Report of the Forty-Sixth Meeting of the Joint FAO/WHO Expert Committee on Food Additives (JECFA)." World Health Organization (WHO), Geneva, Switzerland,

RI PT

WHO Food Additives Series 37. Accessed at http://www.inchem.org/documents/jecfa/jecmono/ v37je01.htm.

Joint FAO/WHO Expert Committee on Food Additives (JECFA). 1997. "Evaluation of Certain Food

SC

Additives and Contaminants. Forty-Sixth Report of the Joint FAO/WHO Expert Committee on Food Additives." World Health Organization (WHO), Geneva, Switzerland, WHO Technical Report Series

M AN U

868. 80p. Accessed at http://apps.who.int/iris/bitstream/10665/41962/1/WHO_TRS_868.pdf. Joint FAO/WHO Expert Committee on Food Additives (JECFA). 2002. "Polychlorinated dibenzodioxins, polychlorinated dibenzofurans and coplanar polychlorinated biphenyls." WHO Technical Report Series 909. In Evaluation of Certain Food Additives and Contaminants: Fifty-seventh report of the Joint

TE D

FAO/WHO Expert Committee on Food Additives (JECFA). World Health Organization (WHO), Geneva, Switzerland. p121-149. Accessed at http://apps.who.int/iris/bitstream/10665/42578/1/WHO_TRS_ 909.pdf.

EP

Joint FAO/WHO Expert Committee on Food Additives (JECFA). 2004. "Methylmercury." WHO Technical Report Series 922. In Evaluation of Certain Food Additives and Contaminants: Sixty-

AC C

first Report of the Joint FAO/WHO Expert Committee on Food Additives (JECFA). World Health Organization (WHO), Geneva, Switzerland. p132-143. Accessed at http://apps.who.int/iris/bitstream/ 10665/42849/1/WHO_TRS_922.pdf.

33

ACCEPTED MANUSCRIPT

Joint FAO/WHO Expert Committee on Food Additives (JECFA). 2015. "Safety Evaluation of Certain Food Additives and Contaminants: Prepared by the Eightieth Meeting of the Joint FAO/WHO Expert Committee on Food Additives (JECFA)." World Health Organization (WHO), Geneva, Switzerland.

RI PT

WHO Food Additives Series 71. 142p. Joint FAO/WHO Expert Committee on Food Additives (JECFA). 2016. "Evaluation of Certain Food Additives and Contaminants: Eightieth Report of the Joint FAO/WHO Expert Committee on Food

SC

Additives." World Health Organization (WHO), Geneva, Switzerland, WHO Technical Report Series

M AN U

995. 132p.

Joint FAO/WHO Food Standards Programme. 2016. "Report of the Forty-Eighth Session of The Codex Committee on Food Additives, Xi'an, China, 14-18 March, 2016." Codex Alimentarius Commission, REP16/FA,

101p.,

March.

Accessed

at

http://www.fao.org/fao-who-codexalimentarius/sh-

proxy/en/?lnk=1&url=https%253A%252F%252Fworkspace.fao.org%252Fsites%252Fcodex%252FMeet

TE D

ings%252FCX-711-48%252FReport%252FREP16_FAe.pdf.

Kieckebusch, W; Lang, K. 1960. "The tolerability of benzoic acid in chronic feeding experiments."

EP

Arzneimittelforschung 10:1001-1003.

Kodama, H; Nose, O; Tajiri, J; Maki, I; Sano, T; Harada, T; Kai, H; Yabuuchi, H. 1981. "Treatment with

AC C

sodium benzoate in a patient with ornithine transcarbamylase deficiency." Pediatr. Int. 23(3):399. doi: 10.1111/j.1442-200X.1981.tb01658.x. Kubota, K; Horai, Y; Kushida, K; Ishizaki, T. 1988. "Determination of benzoic acid and hippuric acid in human plasma and urine by high-performance liquid chromatography." J. Chromatogr. 425(1):67-75. doi: 10.1016/0378-4347(88)80007-0.

34

ACCEPTED MANUSCRIPT

Kubota, K; Ishizaki, T. 1991. "Dose-dependent pharmacokinetics of benzoic acid following oral administration of sodium benzoate to humans." Eur. J. Clin. Pharmacol. 41(4):363-368. doi: 10.1007/

RI PT

BF00314969. Lane, HY; Lin, CH; Green, MF; Hellemann, G; Huang, CC; Chen, PW; Tun, R; Chang, YC; Tsai, GE. 2013. "Add-on treatment of benzoate for schizophrenia: a randomized, double-blind, placebo-controlled trial of D-amino acid oxidase inhibitor." JAMA Psychiatry 70(12):1267-1275. doi: 10.1001/

SC

jamapsychiatry.2013.2159.

ornithine

transcarbamylase

deficiency."

3476(85)80359-0.

M AN U

Letarte, J; Ouellet, R; Qureshi, IA; Godard, M. 1985. "Chronic benzoate therapy in a boy with partial J.

Pediatr.

106(5):794-797.

doi:

10.1016/S0022-

Lin, CH; Chen, PK; Chang, YC; Chuo, LJ; Chen, YS; Tsai, GE; Lane, HY. 2014. "Benzoate, a D-amino acid oxidase inhibitor, for the treatment of early-phase Alzheimer disease: A randomized, double-blind,

TE D

placebo-controlled trial." Biol. Psychiatry 75(9):678-685. doi: 10.1016/j.biopsych.2013.08.010. MacArthur, RB; Altincatal, A; Tuchman, M. 2004. "Pharmacokinetics of sodium phenylacetate and

EP

sodium benzoate following intravenous administration as both a bolus and continuous infusion to healthy adult volunteers." Mol. Genet. Metab. 81(Suppl. 1):S67-S73. doi: 10.1016/j.ymgme.2003.12.011.

AC C

Maestri, NE; Brusilow, SW; Clissold, DB; Bassett, SS. 1996. "Long-term treatment of girls with ornithine transcarbamylase deficiency." N. Engl. J. Med. 335(12):855-859. McCormick, K; Viscardi, RM; Robinson, B; Heininger, J. 1985. "Partial pyruvate decarboxylase deficiency with profound lactic acidosis and hyperammonemia: Responses to dichloroacetate and benzoate." Am. J. Med. Genet. 22(2):291-299. doi: 10.1002/ajmg.1320220211.

35

ACCEPTED MANUSCRIPT

Mendenhall, CL; Rouster, S; Marshall, L; Weesner, R. 1986. "A new therapy for portal systemic encephalopathy." Am. J. Gastroenterol. 81(7):540-543.

RI PT

Mizutani, N; Maehara, M; Hayakawa, C; Kato, T; Watanabe, K; Suzuki, S. 1983. "Hyperargininemia: Clinical course and treatment with sodium benzoate and phenylacetic acid." Brain Dev. 5(6):555-563. doi: 10.1016/S0387-7604(83)80059-X.

Basic Clin. Pharm. 7(2):27-31. doi: 10.4103/0976-0105.177703.

SC

Nair, AB; Jacob, S. 2016. "A simple practice guide for dose conversion between animals and human." J.

M AN U

Nair, B. 2001. "Final report on the safety assessment of benzyl alcohol, benzoic acid, and sodium benzoate." Int. J. Toxicol. 20(Suppl. 3):23-50. doi: 10.1080/10915810152630729. Organisation for Economic Co-operation and Development (OECD). 2001. "SIDS Initial Assessment Report for SIAM 13 on Benzoates: Benzoic Acid (CAS No. 65-85-0), Sodium Benzoate (CAS No. 532-

UNEP Publications. 320p.

TE D

32-1), Potassium Benzoate (CAS No. 582-25-2), and Benzyl Alcohol (CAS No. 100-51-6)."

Pacifici, GM; Mogavero, S; Giuliani, L; Rane, A. 1991. "Conjugation of benzoic acid with glycine in the

EP

human fetal and adult liver and kidney." Dev. Pharmacol. Ther. 17(1-2):52-62.

AC C

Qureshi, IA; Letarte, J; Ouellet, R; Batshaw, ML; Brusilow, S. 1984. "Treatment of hyperargininemia with sodium benzoate and arginine-restricted diet." J. Pediatr. 104(3):473-476. doi: 10.1016/S00223476(84)81120-8.

Renwick, AG. 1993. "Data-derived safety factors for the evaluation of food additives and environmental contaminants." Food Addit. Contam. 10(3):275-305.

36

ACCEPTED MANUSCRIPT

Renwick, AG; Lazarus, NR. 1998. "Human variability and noncancer risk assessment - An analysis of the default uncertainty factor." Regul. Toxicol. Pharmacol. 27(1):3-20.

RI PT

Rhomberg, LR; Lewandowski, TA. 2006. "Methods for identifying a default cross-species scaling factor." Hum. Ecol. Risk Assess. 12:1094-1127.

Sakuma, T; Sugiyama, N; Wada, Y. 1992. "The urinary acylcarnitine profile in three cases of transient

SC

hyperammonemia of the newborn." Acta Paediatr. 81(5):436-438. doi: 10.1111/j.1651-2227.1992. tb12264.x.

M AN U

Schwab, AJ; Tao, L; Yoshimura, T; Simard, A; Barker, F; Pang, KS. 2001. "Hepatic uptake and metabolism of benzoate: A multiple indicator dilution, perfused rat liver study." Am. J. Physiol. Gastrointest. Liver Physiol. 280(6):G1124-G1136.

Scientific Committee on Consumer Products (SCCP). 2005. "Opinion on Benzoic Acid and Sodium

TE D

Benzoate." SCCP/0891/05. European Commission, Health & Consumer Protection Directorate-General. 30p., June 21. Accessed at http://ec.europa.eu/health/ph_risk/committees/04_sccp/docs/sccp_o_015.pdf. Sharp, RA; Lang, CC. 1987. "Hyperammonaemic encephalopathy in chronic myelomonocytic leukaemia

EP

(Letter)." Lancet 329(8536):805. doi: 10.1016/S0140-6736(87)92827-3.

AC C

Shen, DD. 2013. "Toxicokinetics." In Casarett and Doull's Toxicology: The Basic Science of Poisons (Eighth Edition). (Ed: Klaassen, CD), McGraw-Hill Education, New York, NY. p367-390. Simell, O; Sipila, I; Rajantie, J; Valle, DL; Brusilow, SW. 1986. "Waste nitrogen excretion via amino acid acylation: Benzoate and phenylacetate in lysinuric protein intolerance." Pediatr. Res. 20(11):11171121. doi: 10.1203/00006450-198611000-00011.

37

ACCEPTED MANUSCRIPT

Simkin, JL; White, K. 1957. "The formation of hippuric acid; The influence of benzoate administration on tissue glycine levels." Biochem. J. 65(3):574-582.

RI PT

Sushma, S; Dasarathy, S; Tandon, RK; Jain, S; Gupta, S; Bhist, MS. 1992. "Sodium benzoate in the treatment of acute hepatic encephalopathy: A double-blind randomized trial." Hepatology 16(1):138-144. doi: 10.1002/hep.1840160123.

SC

Takeda, E; Kuroda, Y; Toshima, K; Watanabe, T; Naito, E; Miyao, M. 1983. "Effect of long-term administration of sodium benzoate to a patient with partial ornithine carbamoyl transferase deficiency."

M AN U

Clin. Pediatr. (Phila) 22(3):206-208.

Temellini, A; Mogavero, S; Giulianotti, PC; Pietrabissa, A; Mosca, F; Pacifici, GM. 1993. "Conjugation of benzoic acid with glycine in human liver and kidney: A study on the interindividual variability." Xenobiotica 23(12):1427-1433. doi: 10.3109/00498259309059451.

TE D

Thabrew, MI; Bababunmi, EA; French, MR. 1980. "The metabolic fate of [14C] benzoic acid in proteinenergy deficient rats." Toxicol. Lett. 5(6):363-367. doi: 10.1016/0378-4274(80)90016-8. Tremblay, GC; Qureshi, IA. 1993. "The biochemistry and toxicology of benzoic acid metabolism and its

AC C

7258(93)90022-6.

EP

relationship to the elimination of waste nitrogen." Pharmacol. Ther. 60(1):63-90. doi: 10.1016/0163-

Trijbels, JM; Monnens, LA; van der Zee, SP; Vrenken, JA; Sengers, RC; Schretlen, ED. 1974. "A patient with nonketotic hyperglycinemia: Biochemical findings and therapeutic approaches." Pediatr. Res. 8(5):598-605. doi: 10.1203/00006450-197405000-00009.

38

ACCEPTED MANUSCRIPT

Tuchman, M; Mauer, SM; Holzknecht, RA; Summar, M; Vnencak-Jones, CL. 1992. "Prospective versus clinical diagnosis and therapy of acute neonatal hyperammonaemia in two sisters with carbamyl

RI PT

phosphate synthetase deficiency." J. Inherit. Metab. Dis. 15(2):269-277. doi: 10.1007/BF01799641. Uribe, M; Bosques, F; Marin, E; Cervera, E; Gil S; Luis Poo, J; Garcia Compean, D; Santoyo, R; Huerta, E; Garcia-Ramos, G; Cadena, M; Merikansky, A; Robledo, J; Cortes, M; Guevara, A; Lopez, N; Aguilar, A; Alejandra Ochoa, E. 1990. "Sodium benzoate in portal-systemic-encephalopathy-induced

SC

blood ammonia normalization and clinical improvement. Interim report of a double-blind multicenter

M AN U

trial." Rev. Invest. Clin. 42(Suppl.):149-154.

US EPA. 2011. "Recommended Use of Body Weight(3/4) as the Default Method in Derivation of the Oral Reference Dose (Final)." Risk Assessment Forum. EPA/100/R11/0001. 50p. US EPA. 2013. "Integrated Science Assessment for Lead (Final)." National Center for

TE D

Environmental Assessment (NCEA). EPA/600/R-10/075F. 1885p., June. US EPA. 2014a. "Guidance for Applying Quantitative Data to Develop Data-Derived Extrapolation Factors for Interspecies and Intraspecies Extrapolation." Risk Assessment Forum, Office of the Science

EP

Advisor, EPA/100/R-14/002F, 109p., September.

US EPA. 2014b. "Health Effects Document for Perfluorooctanoic Acid (PFOA) (Draft)." Office of

AC C

Water, Health and Ecological Criteria Division. EPA 822-R-14-001. 268p., February. US EPA. 2014c. "Health Effects Document for Perfluorooctane Sulfonate (PFOS) (Draft)." Office of Water, Health and Ecological Criteria Division. EPA 822-R-14-002. 208p., February.

39

ACCEPTED MANUSCRIPT

US Food and Drug Administration (US FDA). 1972. "GRAS (Generally Recognized as Safe) Food Ingredients - Benzoic Acid and Sodium Benzoate." Prepared by Informatics, Inc. , NTIS PB-221208,

RI PT

Rockville, MD, 135p., September. US Food and Drug Administration (US FDA). 1973a. "Select Committee on GRAS Substances (SCOGS) Opinion: Benzoic Acid (CAS No. 65-85-0)." Accessed at http://www.fda.gov/Food/IngredientsPackaging

SC

Labeling/GRAS/SCOGS/ucm260036.htm.

US Food and Drug Administration (US FDA). 1973b. "Select Committee on GRAS Substances (SCOGS)

M AN U

Opinion: Sodium Benzoate (CAS No. 532-32-1)." Accessed at http://www.fda.gov/Food/Ingredients PackagingLabeling/GRAS/SCOGS/ucm260074.htm.

US Food and Drug Administration (US FDA). 2004a. "Clinical Pharmacology and Biopharmaceutics Review(s) for Ammonul." Center for Drug Evaluation and Research. 16p.

TE D

US Food and Drug Administration (US FDA). 2004b. "Pharmacology/Toxicology Review and Evaluation for Ammonul." Center for Drug Evaluation and Research. 16p.. December 8. Accessed at http://www.accessdata.fda.gov/drugsatfda_docs/nda/2005/020645s000_PharmR.pdf.

EP

US Food and Drug Administration (US FDA). 2005. "Guidance for Industry: Estimating the Maximum Safe Starting Dose in Initial Clinical Trials for Therapeutics in Adult Healthy Volunteers." Center for

AC C

Drug Evaluation and research (CDER). 30p., July. US Food and Drug Administration (US FDA). 2016. "Listing of specific substances affirmed as GRAS: Benzoic acid." 21 CFR 184.1021. Accessed at https://www.gpo.gov/fdsys/pkg/CFR-2016-title21vol3/pdf/CFR-2016-title21-vol3-sec184-1021.pdf.

40

ACCEPTED MANUSCRIPT

US Food and Drug Administration (US FDA). 2017. "Drugs@FDA: FDA Approved Drug Products." Accessed on January 19, 2017 at http://www.accessdata.fda.gov/scripts/cder/daf/

RI PT

Van de Bor, M; Mooy, P; van Zoeren, D; Berger, R; van Gelderen, HH; Teijema, HL. 1984. "Successful treatment of severe carbamyl phosphate synthetase I deficiency." Arch. Dis. Child. 59(12):1183-1185. Walter, JH; Allen, JT; Holton, JB. 1992. "Arginosuccinate synthetase deficiency: Good outcome despite

SC

severe neonatal hyperammonaemia." J. Inherit. Metab. Dis. 15(2):282-283. doi: 10.1007/BF01799645. Watson, AJ; Chambers, T; Karp, JE; Risch, VR; Walker, WG; Brusilow, SW. 1985. "Transient idiopathic

M AN U

hyperammonaemia in adults." Lancet 326(8467):1271-1274. doi: 10.1016/S0140-6736(85)91554-5. Yuan, JH; Goehl, TJ; Abdo, K; Clark, J; Espinosa, O; Bugge, C; Garcia, D. 1995. "Effects of gavage versus dosed feed administration on the toxicokinetics of benzyl acetate in rats and mice." Food Chem.

EP AC C

632

TE D

Toxicol. 33(2):151-158. doi: 10.1016/0278-6915(94)00123-6.

41

ACCEPTED MANUSCRIPT

Figure Titles

RI PT

Figure 1 Chemical Structures of Benzoates Figure 2 Subdivision of Inter and Intraspecies Uncertainty Factors. Adapted from IPCS (2005).

AC C

EP

TE D

M AN U

SC

Figure 3 Proposed Uncertainty Factors for Acceptable Daily Intake for Benzoic Acid and Its Salts

1

ACCEPTED MANUSCRIPT

Table 1 Pharmacokinetics of Benzoic Acid and Its Salts in Humans and Rats •

Distribution



Rapid distribution with no evidence of tissue sequestration (JECFA, 1996; US FDA, 2004a)



Rapid distribution with no evidence of tissue sequestration (JECFA, 1996; Gregus et al., 1992)

Metabolism



Rapidly and completely metabolized to hippuric acid (US FDA, 1972; IOMC, 2000; Tremblay and Qureshi, 1993)



Rapidly and completely metabolized to hippuric acid (IOMC, 2000; Bridges et al., 1970; Thabrew et al., 1980)



Peak plasma benzoic acid levels at 1-2 hours after oral administration (Kubota et al., 1988; Kubota and Ishizaki, 1991)



Peak plasma benzoic acid levels 3 hours after oral gavage administration (Adams et al., 2005; JECFA, 1996)a



At high doses (> 500 mg/kg), benzoyl glucuronide is a secondary metabolite (Kubota and Ishizaki, 1991; JECFA, 1996)



At high doses (> 500 mg/kg), benzoyl glucuronide is a secondary metabolite (Adams et al., 2005; JECFA, 1996)



Metabolism driven by conjugation with glycine; saturable process at high doses (i.e., ≥ 160 mg/kg) (Kubota et al., 1988; Kubota and Ishizaki, 1991; MacArthur et al., 2004)



Metabolism driven by conjugation with glycine; saturable process at high doses (i.e., > 120 mg/kg) (Schwab et al., 2001; Gregus et al., 1992; Simkin and White, 1957; JECFA, 1996)



Almost entirely excreted as hippuric acid within 6-24 hours (Kubota et al., 1988; Kubota and Ishizaki, 1991)



Almost entirely excreted as hippuric acid within 24 hours (Bridges et al., 1970; Thabrew et al., 1980)

Rat Essentially complete absorption after oral ingestion (e.g., US FDA, 1972; IOMC, 2000)

SC



b

M AN U

Excretion

Human Essentially complete absorption after oral ingestion (e.g., US FDA, 1972; IOMC, 2000)

RI PT

Endpoint Absorption

AC C

EP

TE D

Notes: (a) Data from a study of F344 rats administered 500 mg/kg benzyl acetate via gavage (Yuan et al., 1995). (b) Data from studies of oral, subcutaneous, or intraperitoneal (i.p.) administration of benzyl acetate to rats (reviewed in Adams et al., 2005 and JECFA, 1996).

1

ACCEPTED MANUSCRIPT

Table 2 Pharmacokinetic Parameters for Humans After Intravenous Administration of Sodium Benzoate Administered i.v. Dose (g/m2) 1 2 4 3.75 5.5

a

BAED (mg/kg bw) 22.9 45.8 91.6 85.9 126.0

AUC (μg × h/mL) Mean (SD) 20.3 (3.6) 114.9 (31.3) 562.8 (142.3) 564.6 (103.9) 1,599.1 (463.1)

AUC/BAED Mean (SD) 0.9 (0.2) 2.5 (0.7) 6.1 (1.6) 6.6 (1.2) 12.6 (3.6)

RI PT

Number of Subjects 6b b 6 b 6 c 17 c 3

AC C

EP

TE D

M AN U

SC

Notes: AUC = Area Under the Curve; BAED = Benzoic Acid Equivalent Dose; i.v. = Intravenous; SD = Standard Deviation; US FDA = United States Food and Drug Administration. (a) Calculated internal sodium benzoate doses from reported i.v. doses with a 2 conversion factor of 37 to convert mg/m to mg/kg bw for adult humans (US FDA, 2005). Converted sodium benzoate to benzoic acid equivalent doses based on the molecular masses of sodium benzoate (144.11 g/mol) and benzoic acid (122.123 g/mol). (b) US FDA Study 973600. (c) US FDA Study 951603. Source: US FDA (2004a).

1

ACCEPTED MANUSCRIPT

Table 3 Pharmacokinetic Parameters for Rats After Intravenous Administration of Sodium Benzoate

5-8 5-8 5-8 5-8

Administered i.v. Dose (mmol/kg) 0.2 0.5 1 2

a

BAED (mg/kg bw) 24.4 61.1 122.1 244.2

b

AUC (μg × h/mL) Mean (SD) 10.4 (0.8) 58.2 (6.3) 239.5 (22.5) 1,100.2 (89.2)

c

AUC/BAED Mean (SD) 0.4 (0.03) 1.0 (0.1) 2.0 (0.2) 4.5 (0.4)

RI PT

Number of Rats

AC C

EP

TE D

M AN U

SC

Notes: AUC = Area Under the Curve; BAED = Benzoic Acid Equivalent Dose; bw/BW = Body Weight; i.v. = Intravenous; SD = Standard Deviation. (a) Converted sodium benzoate to benzoic acid equivalent doses based on the molecular masses of sodium benzoate (144.11 g/mol) and benzoic acid (122.123 g/mol). (b) Calculated based on the formula: AUC = F × Dose/CL, where F = Fraction absorbed (1 for i.v. administration) and CL = Clearance (in units of mL/kg × min) (Shen, 2013). (c) Units are μg × h/mL per mg/kg dose. Source: Gregus et al. (1992).

1

ACCEPTED MANUSCRIPT

RI PT

Table 4 Comparison of the Relative Maximum In Vitro Metabolism in Adult Humans and Rat Livers Vmax Rats (SD) Vmax Humans (SD) Rat Vmax/Human Vmax (nmol/min/g liver) (nmol/min/g liver) 101 (NR)a 204 (47.8)b 0.5 c 78 (6) 204 (47.8)b 0.38 101 (NR)a 254 (90.5)d 0.4 c d 78 (6) 254 (90.5) 0.31

AC C

EP

TE D

M AN U

SC

Notes: NR = Not Reported; SD = Standard Deviation; Vmax = Maximum in vitro biotransformation rate of benzoic acid to hippuric acid. (a) Chiba et al. (1994); Tested concentrations: 0.004 to 600 µM benzoic acid. (b) Pacifici et al. (1991); Tested concentrations: 6 to 200 µM benzoic acid. (c) Schwab et al. (2001); Tested concentrations: 5 to 1,000 µM benzoic acid. (d) Temellini et al. (1993); Tested concentration: 200 µM benzoic acid.

1

ACCEPTED MANUSCRIPT

Table 5 Pharmacokinetic Parameters for Rats After Intravenous Administration of Sodium Benzoate a

b

BAED (mg/kg bw)

HED (mg/kg bw)

24.4 61.1 122.1 244.2

6.5 16.2 32.5 64.9

c

AUC (μg × h/mL) Mean (SD) 10.4 (0.8) 58.2 (6.3) 239.5 (22.5) 1,100.2 (89.2)

d

AUC/HED Mean (SD) 1.6 (0.1) 3.6 (0.4) 7.4 (0.7) 16.9 (1.4)

RI PT

Administered i.v. Dose (mmol/kg) 0.2 0.5 1 2

AC C

EP

TE D

M AN U

SC

Notes: AUC = Area Under the Curve; BAED = Benzoic Acid Equivalent Dose; bw/BW = Body Weight; HED = Human Equivalent Dose; i.v. = Intravenous; SD = Standard Deviation. (a) Converted from mmol/kg bw to mg/kg bw based on the molecular mass of benzoic acid (122.123 g/mol). 3/4 (b) Converted to HEDs based on allometric scaling of BW (US EPA, 2011; Rhomberg and Lewandowski, 2006). (c) Calculated based on the formula: AUC = F × Dose/CL, where F = Fraction absorbed (1 for i.v. administration) and CL = Clearance (in units of mL/kg × min) (Shen, 2013). (d) Units are μg × h/mL per mg/kg dose. Source: Gregus et al. (1992).

1

ACCEPTED MANUSCRIPT

Table 6 Comparison of Dose-normalized AUCs in Humans and Rats After Allometric Scaling a

b

HED in Rats (mg/kg bw) 16.2 32.5 64.9

Mean AUCNorm c in Humans 0.9 2.5 6.6

Mean AUCNorm c in Rats 3.6 7.4 16.9

Human AUCNorm/Rat AUCNorm 0.3 0.3 0.4

RI PT

BAED in Humans (mg/kg bw) 22.9 45.8 85.9

AC C

EP

TE D

M AN U

SC

Notes: AUCNorm = Dose-normalized Area Under the Curve; BAED = Benzoic Acid Equivalent Dose; bw = Body Weight; HED = Human Equivalent Dose. (a) Converted to benzoic acid equivalent doses. 3/4 (b) Converted to HEDs by scaling the rat benzoic acid equivalent doses by BW (US EPA, 2011; Rhomberg and Lewandowski, 2006). (c) Units are μg × h/mL per mg/kg dose. Sources: US FDA (2004a); Gregus et al. (1992).

1

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

Benzoic Acid

Sodium Benzoate

Potassium Benzoate

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT

Benzoate pharmacokinetics are similar between humans and rodents.



A conservative CSAF of 2 is derived for interspecies variability in pharmacokinetics.



Use of chemical-specific data reduces the default UF of 100 for the ADI to 50.



An ADI of 10 mg/kg bw-day is proposed for benzoic acid and its salts.



Clinical data support the safety of benzoates at current dietary levels.

AC C

EP

TE D

M AN U

SC

RI PT



1