Wastewater analysis to monitor use of caffeine and nicotine and evaluation of their metabolites as biomarkers for population size assessment

Accepted Manuscript Wastewater analysis to monitor use of caffeine and nicotine and evaluation of their metabolites as biomarkers for population size assessment Ivan Senta, Emma Gracia-Lor, Andrea Borsotti, Ettore Zuccato, Dr. Sara Castiglioni, Head of the Environmental Biomarkers Unit PII:

S0043-1354(15)00075-5

DOI:

10.1016/j.watres.2015.02.002

Reference:

WR 11138

To appear in:

Water Research

Received Date: 10 September 2014 Revised Date:

19 December 2014

Accepted Date: 1 February 2015

Please cite this article as: Senta, I., Gracia-Lor, E., Borsotti, A., Zuccato, E., Castiglioni, S., Wastewater analysis to monitor use of caffeine and nicotine and evaluation of their metabolites as biomarkers for population size assessment, Water Research (2015), doi: 10.1016/j.watres.2015.02.002. 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.

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT 1

Wastewater analysis to monitor use of caffeine and nicotine and evaluation

2

of their metabolites as biomarkers for population size assessment

4

RI PT

3

Ivan Senta1, Emma Gracia-Lor2, Andrea Borsotti2, Ettore Zuccato2, Sara Castiglioni2*

5

6

1

7

10000 Zagreb, Croatia.

8

2

9

Health Sciences, Via La Masa 19, 20156, Milan, Italy.

M AN U

SC

Rudjer Boskovic Institute, Division for Marine and Environmental Research, Bijenicka c. 54,

IRCCS – Istituto di Ricerche Farmacologiche “Mario Negri”, Department of Environmental

TE D

10

* Corresponding author:

12

Dr. Sara Castiglioni,

13

Head of the Environmental Biomarkers Unit

14

Department of Environmental Health Sciences

15

IRCCS-Istituto di Ricerche Farmacologiche "Mario Negri"

16

Via La Masa 19, 20156 Milan, Italy

17

Tel: +39 02 39014776

18

Fax: +39 02 39014735

19

e-mail: [email protected]

AC C

EP

11

20

21

Keywords: wastewater analysis, caffeine, nicotine, urinary metabolites, population

22

biomarkers

23 1

ACCEPTED MANUSCRIPT

Abstract

25

The use of caffeine, nicotine and some major metabolites was investigated by wastewater

26

analysis in 13 sewage treatment plants (STPs) across Italy, and their suitability was tested as

27

qualitative and quantitative biomarkers for assessing population size and dynamics. A

28

specific analytical method based on mass spectrometry was developed and validated in raw

29

urban wastewater, and included two caffeine metabolites, 1-methylxanthine and 7-

30

methylxanthine, never reported in wastewater before. All these compounds were found

31

widely at the μg/L level. Mass loads, calculated by multiplying concentrations by the

32

wastewater daily flow rate and normalized to the population served by each plant, were

33

used to compare the profiles from different cities. Some regional differences were observed

34

in the mass loads, especially for nicotine metabolites, which were significantly higher in the

35

south than in the center and north of Italy, reflecting smoking prevalences from population

36

surveys. There were no significant weekly trends, although the mean mass loads of caffeine

37

and its metabolites were slightly lower during the weekend. Most caffeine and nicotine

38

metabolites fulfilled the requirements for an ideal biomarker for the assessment of

39

population size, i.e. being easily detectable in wastewater, stable in sewage and during

40

sampling, and reflecting human metabolism. Nicotine metabolites were tested as

41

quantitative biomarkers to estimate population size and the results agreed well with census

42

data. Caffeine and its metabolites were confirmed as good qualitative biomarkers, but

43

additional information is needed on the caffeine metabolism in relation to the multiple

44

sources of its main metabolites. This exploratory study opens the way to the routine use of

45

nicotine metabolites for estimating population size and dynamics.

AC C

EP

TE D

M AN U

SC

RI PT

24

46

2

ACCEPTED MANUSCRIPT

1. Introduction

48

Caffeine and nicotine are the most widely used legal stimulants in modern societies

49

(Garattini, 1993; WHO, 2013). Caffeine is the main stimulating ingredient in coffee, but is

50

also found in other widely-consumed products, such as tea, soft and “energy” drinks.

51

Nicotine is contained in cigarettes and other tobacco products and is the major addictive

52

component of tobacco. Once consumed, these substances are extensively metabolized in

53

the human body and excreted, mostly in the urine, as complex mixtures of parent

54

compounds and metabolites (Garattini, 1993; Debry, 1994; Hukkanen et al., 2005), which

55

end up in the sewage system. In fact, these substances are frequently detected in municipal

56

wastewater at μg/L concentrations ( Buerge et al., 2003; Buerge et al., 2008; Huerta-Fontela

57

et al., 2008; Santos et al., 2009; Rosal et al., 2010; Bueno et al., 2011) and are among the

58

most ubiquitous waste and surface water microcontaminants (Focazio et al., 2008).

59

However, comprehensive data on the occurrence of their metabolites are still scarce,

60

especially for caffeine. In fact, some of caffeine’s major metabolites, such as 1-

61

methylxanthine and 7-methylxanthine, have never been analyzed in environmental samples.

62

In view of the widespread use of caffeine and nicotine, it has been suggested that these

63

compounds could be used as anthropogenic markers to indicate the discharge of domestic

64

wastewater in rivers and lakes (Buerge et al., 2003; Buerge et al., 2008). Caffeine has also

65

been proposed as a human biomarker for assessing population size and the dynamics of

66

people served by a particular sewage treatment plant (STP) (Daughton, 2012).

67

Population size and dynamics are important parameters in many human activities, including

68

material-flow (chemicals-flow) analysis in different environmental matrices and the per

69

capita contributions of different pollutants (e.g. pharmaceuticals, personal care products,

70

household chemicals, pesticides, biocides, nanomaterials) to the environment ("pollutant

AC C

EP

TE D

M AN U

SC

RI PT

47

3

ACCEPTED MANUSCRIPT load per capita"; (Tsuzuki, 2006). This is particularly relevant for the "wastewater-based

72

epidemiology” approach, which involves analysis of urban wastewater for the combined

73

excretion products of different substances to track human habits and lifestyle (Thomas and

74

Reid, 2011; Castiglioni et al., 2014). This approach was originally developed for the

75

estimation of illicit drug consumption through wastewater analysis (Zuccato et al., 2005;

76

Zuccato et al., 2008; van Nuijs et al., 2011), but can be extended to other applications, such

77

as alcohol (Reid et al., 2011) and nicotine (Castiglioni et al., 2015). The rationale for this

78

approach is based on the fact that almost everything we consume is excreted unchanged

79

and/or as a mixture of metabolites in our urine and feces and ultimately ends up in the

80

sewage network. Thus, the concentrations of metabolic residues in raw municipal

81

wastewater can reflect the collective consumption of a substance in a community.

82

The accuracy of comparison of the profiles of consumption of illicit drugs and other

83

substances in different communities relies on the estimation of the population size, i. e. the

84

number of persons served by the STPs investigated, and the characterization of population

85

dynamics. Current methods for population size assessment are based mainly either on public

86

surveys, such as a census, or certain hydro-chemical parameters that are routinely

87

determined at the STPs, including chemical oxygen demand (COD), biological oxygen

88

demand (BOD) and total nitrogen and phosphorus (Andreottola et al., 1994; Daughton,

89

2012). However, since these parameters are greatly influenced by the wastewater

90

composition (i.e. industrial, domestic or mixed), the measurement of specific substances in

91

urban wastewater, which univocally indicates the persons served by a STP, has been

92

proposed as an alternative for estimating population size (Daughton, 2012).

93

An ideal biomarker should fulfill several requirements: 1) be unique to human metabolism;

94

2) have no or minimal exogenous sources; 3) have a stable daily per capita excretion with

AC C

EP

TE D

M AN U

SC

RI PT

71

4

ACCEPTED MANUSCRIPT minimal intra-individual and inter-individual variability; 4) be evenly distributed and stable in

96

sewage; 5) be only minimally formed by microbial activity in sewage; 6) be determined

97

easily, quickly and safely in environmental samples. Obviously, finding a suitable compound

98

is a great challenge. The viability of the compounds proposed as population biomarkers,

99

such as pharmaceuticals, coprostanol, caffeine, biocides and food additives has not been

100

experimentally verified, but a few studies tested some of these substances in the last few

101

years. Measuring pharmaceutical loads was first suggested to estimate the number of

102

persons contributing to the wastewater (Lai et al., 2011). More recently, several compounds,

103

including cotinine, have been screened using different criteria, such as quantification

104

methods, affinity to particulate, stability in wastewater, constancy of inter-day excretion,

105

and correlation with census population (Chen et al., 2014).

106

The present study tested caffeine and nicotine derivatives for the first time as quantitative

107

human biomarkers for the assessment of population size in raw wastewater. The aims of the

108

study were: a) to investigate the occurrence of caffeine, nicotine and some of their major

109

metabolites in raw wastewater in Italy; b) to assess their patterns of use in different

110

communities through wastewater analysis; c) to explore their potential as human

111

biomarkers for population size assessment. A specific analytical method, including for the

112

first time an almost complete set of metabolites of caffeine and nicotine, was developed and

113

validated in raw urban wastewater. The reliability of these compounds as human biomarkers

114

was investigated by stability tests during residence in sewage, during storage and

115

wastewater sampling and by comparing the figures for inhabitants obtained from

116

hydrochemical parameters and from nicotine derivatives.

AC C

EP

TE D

M AN U

SC

RI PT

95

117

5

ACCEPTED MANUSCRIPT

2. Material and Methods

119

2.1. Selection of analytes

120

At the beginning of the study, nine compounds were selected as possible human biomarkers

121

for population size assessment: caffeine and its major metabolites paraxanthine (1,7-

122

dimethylxanthine), 1-methylxanthine, 7-methylxanthine, 1,7-dimethyluric acid, 1-methyluric

123

acid (Garattini, 1993; Baselt, 2004), and nicotine and its metabolites cotinine and trans-3’-

124

hydroxycotinine (Byrd et al., 1992; Hukkanen et al., 2005). For caffeine, the first criterion for

125

selecting metabolites was the percentage of excretion of each metabolite which was

126

available for a relatively low number of subjects (range 7-68) (Table S1). Generally, those

127

accounting for more than 5% in the total human metabolism were taken into account as

128

possible biomarkers. However, the literature suggested that some were not good candidates

129

for biomarkers, because of instability in urine and/or in wastewater. For example, AFMU (5-

130

acetyl-amino-6-formylamino-3-methyluracil), one of the main metabolites of caffeine, was

131

unstable in urine (Krul and Hageman, 1998; Wong et al., 2002). For nicotine, we selected two

132

main metabolites excreted in urine: cotinine (30%) and trans-3’-hydroxycotinine (44%)

133

(Castiglioni et al., 2015). Glucuronide forms of nicotine and its major metabolites were

134

considered as deconjugated compounds, because they are completely converted to the free

135

form by β-glucuronidase enzymes from fecal bacteria in raw wastewater (D'Ascenzo et al.,

136

2003; Castiglioni et al., 2006).

AC C

EP

TE D

M AN U

SC

RI PT

118

137

138

139

6

ACCEPTED MANUSCRIPT 2.2. Chemicals and materials

141

Cotinine and cotinine-d3 were purchased from Cerilliant Corporation (Round Rock, Texas,

142

USA). Caffeine, caffeine-13C3 , paraxanthine, 1-methylxanthine, nicotine and nicotine-d3 were

143

purchased from Sigma Aldrich (St. Louis, MO, USA). Trans-3’-hydroxycotinine, 1-methyluric

144

acid, 1,7-dimethyluric acid, 1,7-dimethyluric acid-d3 and 7-methylxanthine were obtained

145

from Santa Cruz Biotechnology, Inc. (Santa Cruz, California, USA). All the acquired standards

146

were of analytical grade in liquid or powder form with purity higher than 98%. Individual

147

stock solutions at the concentration of 1 mg/mL were prepared in methanol, except for 1-

148

methylxanthine, 7-methylxanthine, paraxanthine, 1-methyluric acid and 1,7-dimethyluric

149

acid, for which stock solutions were prepared in methanol/water (50/50) at pH 8.5-10

150

(adjusted with 25%ammonia solution). Stock solutions were stored in the dark at -20°C.

151

Working standard solutions containing all target analytes at 1, 0.1 and 0.01 ng/μL were

152

prepared by diluting the individual stock solutions with methanol. Labeled standard

153

mixtures, used as internal standards (IS), were prepared separately following the same

154

procedure.

155

All the solvents used were of reagent grade or higher. Methanol for pesticide analysis and

156

ammonium acetate were from Carlo Erba Reagents (Italy). Ammonium hydroxide solution

157

(25%) was acquired from Fluka (Buchs, Switzerland). LC-MS grade acetonitrile and

158

hydrochloric acid (37%) were supplied by Riedel de Haen (Seelze, Germany). Water was

159

purified using Milli-RO Plus 90 apparatus (Millipore, Molsheim, France). Solid-phase

160

cartridges (3-mL Oasis HLB, 60 mg) and HPLC columns (XTerra C18, 100 x 1 mm; 3.5 µm)

161

were obtained from Waters Corp. (Milford, MA).

AC C

EP

TE D

M AN U

SC

RI PT

140

162

7

ACCEPTED MANUSCRIPT 2.3. Sampling

164

24-h volume or time-proportional composite raw wastewater samples were collected at the

165

entrance of 13 STPs serving large (>500 000 inhabitants) and medium (>50 000 inhabitants)

166

cities across Italy. Five cities (Milan, Como, Bologna, Turin, Verona) are in the north of Italy,

167

four (Pescara, Florence, Perugia, Rome) are in the center, and four (Bari, Palermo, Naples,

168

Potenza) in the south. The main characteristics of the STPs and the sampling periods are

169

listed in the Supplementary Information (SI) (Table S2). Samples were collected daily

170

considering the mean travel time for wastewater from use to plant which was 7 hours. In all

171

the analytical campaigns samples were collected for 7-18 consecutive days. Two seven-days

172

sampling campaigns were carried out during spring and autumn 2012 in Como, Rome and

173

Naples. In Milan, four sampling campaigns were carried out in March (7 days), April-May (15

174

days), September-October (18 days) and November-December (18 days) 2012. At all other

175

sampling locations, one seven-day sampling campaign was conducted in October 2012

176

(Table S2). All the samples were collected in polyethylene terephthalate (PET) 500-mL

177

bottles, immediately frozen and transported to the laboratory, where they were kept at -

178

20°C until extraction.

SC

M AN U

TE D

EP

AC C

179

RI PT

163

180

2.4. Extraction and extract work-up

181

The solid-phase extraction (SPE) method for the selected analytes has been modified from

182

previous publications that included caffeine and nicotine analyses (Huerta-Fontela et al.,

183

2007; Bueno et al., 2011). After thawing in a warm bath, samples were filtered to remove

184

the suspended matter. They were passed on 1.6 μm GF/A glass microfiber filters (Whatman,

185

Kent, UK) and then on 0.45 μm mixed cellulose membrane filters (Whatman, Kent, UK). 8

ACCEPTED MANUSCRIPT Filtered samples (3 mL) were spiked with an IS mixture (2 ng of cotinine-d3 and 20 ng of

187

nicotine-d3 and caffeine-13C3) and, if necessary, the pH was adjusted to between 6.0 and 7.5

188

using 12% HCl (v/v). Samples were loaded on Oasis HLB cartridges previously equilibrated

189

with 6 mL of methanol and 3 mL of ultrapure water. After percolation, cartridges were

190

vacuum-dried for 5 minutes, wrapped in aluminum foil and immediately stored at -20°C.

191

For analysis, cartridges were eluted with 2 mL of MeOH and the eluates were evaporated to

192

dryness under a nitrogen stream. Dry residues were redissolved in 100 μL of H2O/MeOH

193

mixture (80/20, v/v), centrifuged and transferred into glass vials for instrumental analysis.

SC

RI PT

186

M AN U

194

2.5. Liquid chromatography–tandem mass spectrometry (LC-MS/MS)

196

The analyses were done using an API 5500 QqQ equipped with a Turbo Ion Spray source

197

(Applied Biosystems - Sciex, Thornhill, Ontario, Canada) and a 1200 Series pumps system

198

(Agilent Technologies, Santa Clara, CA, USA).

199

The compounds were separated on 100 x 1 mm X-Terra C18 column using 10 mM

200

ammonium acetate (eluent A) and acetonitrile (eluent B), with the following gradient

201

program: initial condition 98% of eluent A, followed by a 10-min linear gradient to 100% of

202

eluent B, 4-min isocratic elution and 1-min linear gradient back to 98% of eluent A, which

203

was held for 13 minutes to equilibrate the column. The flow rate was maintained at 70

204

μL/min and the injection volume was 2 μL.

205

Samples were ionized using electrospray ionization in positive polarity. The Turbo Ion Spray

206

source settings were: ion spray voltage (IS) 5500V; source temperature 400°C; curtain gas

207

25; collision gas (CAD) 7; ion source gas 1 (GS1) 30; ion source gas 2 (GS2) 35. Mass

208

spectrometric analysis was done in the selected reaction monitoring (SRM) mode under

AC C

EP

TE D

195

9

ACCEPTED MANUSCRIPT time-scheduled conditions, with a time window of 240 s. Scheduled SRM enables one to

210

maximize dwell times during acquisition, optimizing cycle times in order to provide good

211

analytical precision even for multiple transitions. Two most abundant precursor/product ion

212

transitions were obtained for each compound. The choice of the transitions and optimization

213

of the corresponding collision energies (CE) was done in continuous-flow mode using

214

standard solutions at 50-100 pg/µL. The entrance potential (EP) and the collision cell exit

215

potential (CXP) were respectively 8.0 and 12.0 for all the target compounds. The precursor

216

and product ions selected for analytes and IS, with their collision energies, are listed in Table

217

S3.

M AN U

SC

RI PT

209

218

2.6. Quantification and validation of the method

220

Recoveries and repeatabilities were determined by analyzing raw wastewater samples,

221

spiked with the target analytes at two concentrations: 100 μg L-1 for caffeine and

222

metabolites and 20 μg L-1 for nicotine and metabolites. Non-spiked samples were analyzed

223

as well, in order to subtract the amount of target analytes already present in the original

224

sample. Spiked and non-spiked samples were analyzed in triplicate. Repeatability was

225

determined as the relative standard deviation (RSD) of the analysis of spiked wastewater

226

samples. Blank samples, prepared with mineral water and spiked only with the IS mixture,

227

were processed with each analytical batch to check for any contamination and correct for

228

biases.

229

Target analytes were quantified using the isotopic dilution method and isotopically labeled

230

analogs of nicotine, caffeine and cotinine were used. A six-point calibration curve was

231

constructed by injecting standard solutions containing different amounts of target analytes

AC C

EP

TE D

219

10

ACCEPTED MANUSCRIPT 232

(0-600 ng) and a fixed amount of IS (2 ng of cotinine-d3 and 20 ng of nicotine-d3 and caffeine-

233

13

234

Limits of detection and limits of quantification were calculated as the concentrations of

235

analytes giving signal-to-noise ratio of 3 and 10, respectively. Method detection limits

236

(MDLs) and method quantification limits (MQLs) were calculated using real wastewater

237

samples. Instrumental detection limits (IDLs) and instrumental quantification limits (IQLs)

238

were calculated by injecting standard solutions containing small amounts of each target

239

analyte, until signal-to-noise ratios of 10 were achieved.

M AN U

240

SC

RI PT

C3). The first point was spiked only with the IS and was also used as an instrumental blank.

2.7. Stability of analytes

242

The degradation of a substance in wastewater can occur due to the high microbial activity of

243

different species of bacteria. However, due to the impossibility to perform a degradation

244

study of the compounds on a real scale, stability tests were done in the laboratory mimicking

245

“real conditions” for temperature to check the stability of selected compounds in the sewer

246

system (residence time about 7 hours) and during the collection of 24-h composite samples.

247

The experiments were conducted at 4°C, which is the common temperature during the

248

sample collection, and 22°C as the expected highest temperature in a sewer system in Italy

249

(worst-case scenario). Additional aliquots were stored at -20°C to check stability during

250

storage before analysis. All the aliquots were prepared from raw wastewater, divided into

251

three sets, and stored as follows: one set was frozen at -20°C, one was kept at 4°C and one

252

at room temperature (22°C). Aliquots stored at 4°C and at room temperature were

253

extracted immediately (T0) and after 2, 4, 6, 8 and 24 h. Aliquots stored at -20°C were

AC C

EP

TE D

241

11

ACCEPTED MANUSCRIPT 254

analyzed after one and four weeks. High concentrations of all target analytes in wastewater

255

enabled us to do these experiments without any spiking.

256

2.8. Statistical analysis

258

GraphPad Prism (version 6) was used for the statistical analysis of results. One-way ANOVA

259

with p <0.05 was applied for the stability experiment at 4°C and 20°C, and the t-test with

260

p<0.05 for the stability experiment at -20°C.

261

The spatial and temporal variabilities of the mass loads were also analyzed but only for the

262

October 2012 sampling campaign, when were collected samples from all the cities. For

263

spatial variability, the cities were grouped according to their location (north, center, south)

264

and the mass loads were compared. Statistical analysis was done using a non-parametric

265

Kruskal-Wallis test followed by Dunn’s multiple comparison test on account of the non-

266

normal distribution of some data. In view of the small dataset and the consequent difficulty

267

in correctly establishing the normality of data distribution, statistical significance was also

268

verified using a parametric test (one-way ANOVA followed by Tukey’s multiple comparison

269

test) and the results were similar in all cases. For inter-day variability, the mass loads on

270

different days of the week were compared using one-way ANOVA with p <0.05.

SC

M AN U

TE D

EP

AC C

271

RI PT

257

272

3. Results and discussion

273

3.1. Method validation

274

Method validation parameters, including linearities, recoveries, repeatabilities and

275

quantification limits, are reported in Table 1. Analytical parameters were good for most of 12

ACCEPTED MANUSCRIPT the compounds, with recoveries between 64% and 89% and variability lower than 14%

277

(RSD). However, the method did not perform acceptably for 1-methyluric acid and 1,7-

278

dimethyluric acid and the results were omitted. 1-methyluric acid was excluded because the

279

recovery was low (14%), with high variability (RSD 50%), and the instrumental sensitivity was

280

poor. For 1,7-dimethyluric acid, only caffeine-13C3 could be used as IS at the beginning of the

281

study, but it was inappropriate for quantification of this compound, on account of the

282

different instrumental response. A deuterated analog of 1,7-dimethyluric acid was acquired

283

later and quantification was much more reliable, but only very few samples could be

284

processed under these conditions. Future analytical investigations will aim for more reliable

285

results for this compound too. Contaminations checked by running instrumental and

286

analytical blanks were generally below the LOQs, some traces of caffeine were detected, but

287

they were far below the 10% of the level in the samples.

288

Calibration curves were linear in the range analyzed (0-600 ng/mL) with r2 values above

289

0.999 for all compounds. Instrumental detection and quantification limits ranged,

290

respectively, from 0.11 to 2.97 and from 0.37 to 9.9 pg/injected, and MQLs ranged from 0.43

291

to 28.5 ng/L.

SC

M AN U

TE D

EP AC C

292

RI PT

276

293

3.2. Stability in wastewater

294

All compounds were stable during 24 hours storage at 4°C and 20°C and they fulfill one of

295

the most important requirements for reliable human biomarkers (Table S4). The stability test

296

was conducted in raw wastewater, but not on a real scale in the sewer pipe where a

297

complex mixture of biofilms occurs, thus the assessment of stability in urban wastewater

298

could be improved through additional modeling studies to investigate in-pipe

13

ACCEPTED MANUSCRIPT biotrasformation. All the compounds were also stable at -20°C for up to four weeks (Table

300

S4). Some degradation was observed only for 7-methylxanthine, but this was not statistically

301

significant. Therefore, freezing is the best method for sample preservation between

302

collection and extraction.

RI PT

299

303

3.3. Occurrence of caffeine, nicotine and their metabolites in wastewater

305

The investigated compounds in raw wastewater of 13 Italian cities are listed in Table 2.

306

Caffeine, nicotine, and their metabolites were detected in the µg/L range in all samples,

307

which is not surprising considering the high daily consumption of these compounds in

308

modern societies. There are numerous reports on caffeine in wastewater, but data on its

309

metabolites are still very scarce. The mean concentrations of caffeine in this study were

310

between 17.6 and 67.6 μg/L, generally comparable to the levels previously reported (Buerge

311

et al., 2003; Santos et al., 2009; Rosal et al., 2010; Bueno et al., 2011). Paraxanthine was

312

mostly analyzed in Spanish wastewater and the mean concentrations in our study (17.5-77.3

313

μg/L) were similar (Huerta-Fontela et al., 2008; Rosal et al., 2010; Bueno et al., 2011) or

314

higher (Teijon et al., 2010).

315

To the best of our knowledge, this is the first report of 1-methylxanthine and 7-

316

methylxanthine in urban wastewater (Table 2). The mean concentrations in Italian

317

wastewaters were 5.0-61.9 μg/L and 7.3-84.1 μg/L for 1-methylxanthine and 7-

318

methylxanthine, respectively.

319

The mean concentrations of nicotine (1.36-6.87 μg/L) were comparable to the

320

concentrations in Spanish wastewaters (Teijon et al., 2010). The mean cotinine and trans-3’-

321

hydroxycotinine levels (0.65-3.12 μg/L and 2.14-7.00 μg/L, respectively) were similar to the

AC C

EP

TE D

M AN U

SC

304

14

ACCEPTED MANUSCRIPT 322

concentrations found in Swiss wastewaters (Buerge et al., 2008). However, there are some

323

reports of higher concentrations of both nicotine and cotinine in Spanish wastewater

324

(Huerta-Fontela et al., 2008; Bueno et al., 2011).

RI PT

325

3.4. Mass loads profiles of caffeine, nicotine and their metabolites in wastewater

327

The mass loads of caffeine, nicotine and their metabolites were calculated by multiplying the

328

analyte concentrations (ng/L) by the daily flow rate in each STP (m3/day) (Table S2). To

329

compare loads from different cities, mass loads were normalized to the number of people

330

served by each STP, which was selected case by case from the most reliable estimates

331

following a recently suggested protocol (Castiglioni et al., 2013). Mass loads are reported in

332

Figures 1 A and B and can indicate the profile of excretion of each substance.

333

The mass loads of caffeine and its metabolites (Figure 1A) ranged between 10 and 15

334

g/day/1000 inhabitants in all the cities, except Pescara and Perugia where they were lower

335

than 10 g/day/1000 inhabitants for all the substances, and for Florence, Palermo and Naples

336

where caffeine and paraxanthine mass loads reached 20 g/day/1000 inhabitants. The mean

337

mass load of caffeine was 14 ± 5.2 g/day/1000 inhabitants, which is closely comparable with

338

the loads found in Swiss wastewater treatment plants a few years ago (15.8 ± 3.8

339

g/day/1000 inhabitants) (Buerge et al., 2003).

340

The quantitative profiles of the different metabolites were similar in all the cities (Figure 1A)

341

and was caffeine > paraxanthine > 7methylxanthine > 1 –methylxanthine, but the mass load

342

profiles do not seem to completely agree with the human excretion profile of caffeine (Table

343

S1). In particular, the loads of 1-methylxanthine, which is one of the caffeine metabolites

344

excreted in the highest percentages (9-18%), were lower than those of paraxanthine and 7-

AC C

EP

TE D

M AN U

SC

326

15

ACCEPTED MANUSCRIPT methylxanthine, whose excretion is reported to be lower (2-7%). Caffeine mass loads were

346

also higher than expected from human metabolism (1-3% of caffeine is excreted as parent

347

compound). This can be easily explained by the presence of this substance in several

348

different products, such as a wide range of beverages (coffee, tea, soft and “energy” drinks),

349

foods (chocolate) and dietary supplements (guaranà). Thus it can come from multiple

350

sources. Certain caffeine metabolites are also metabolites of other naturally occurring

351

alkaloids with similar structures, such as theobromine and theophylline. For example, 7-

352

methylxanthine is the major metabolite of theobromine (Rodopoulos et al., 1996), the main

353

alkaloid found in cocoa beans and therefore in many widely-consumed chocolate products

354

(Srdjenovic et al., 2008). Moreover, 1-methylxanthine is also a minor metabolite of

355

theophylline (Rodopoulos and Norman, 1997), an alkaloid naturally present in tea leaves, but

356

which can be also used for the treatment of asthma and other lung diseases (Barnes, 2003).

357

This is a quite complex situation that requires further investigation of the human

358

metabolism of these substances related to their sources. Moreover, also pharmacokinetic

359

data on caffeine metabolism can be a source of biases, since most of the studies are quite

360

old and include relatively few subjects (see SI). This was an exploratory study and it was

361

anyway able to show a constant profile of the different metabolites in almost all the cities

362

indicating a common profile of use of these substances. Another possible explanation of the

363

differences between the mass load profiles and the human excretion profiles of caffeine and

364

its metabolites might be related to their different in-sewer degradation rates which should

365

be verified with ad-hoc modeling studies.

366

The mass loads of nicotine, cotinine and trans-3’-hydroxicotinine varied in the different cities

367

and ranged between 0.5 and 3 g/day/1000 inhabitants (Figure 1B). Instead, the profiles of

368

cotinine and trans-3’-hydroxicotinine in wastewater were very similar and reflected their

AC C

EP

TE D

M AN U

SC

RI PT

345

16

ACCEPTED MANUSCRIPT human excretion profiles (Hukkanen et al., 2005; Castiglioni et al., 2015), trans-3’-

370

hydroxicotinine being the most abundant metabolite. Nicotine itself was found in larger

371

amounts than expected considering only urinary excretion. This is in accordance with the

372

hypothesis that other sources of nicotine, such as the improper disposal of ash and cigarette

373

butts or pharmacological use, may contribute to the total amount of nicotine in wastewater

374

(Castiglioni et al., 2015).

RI PT

369

SC

375

3.4.1. Weekly profiles of mass loads

377

Weekly patterns of mass loads of selected compounds were also investigated. To check

378

inter-day variabilities, the mass loads for all the samples collected in September-October

379

2012 were used to calculate the mean mass load for each day of the week (Fig. 2). The mean

380

mass loads of caffeine and its metabolites tended to be lower during the weekend than on

381

weekdays (12-19%). This difference was more pronounced for Sundays (19-27%) than

382

Saturdays (1-10%), but was not statistically relevant. These results are in line with the

383

findings of Brewer et al. (2012), who reported decreasing of non-normalized daily mass loads

384

of caffeine from Wednesday to Saturday. One possible explanation could be that coffee, as

385

the main caffeine source, is used less at weekends. However, Brewer et al. (2012) found that

386

caffeine mass loads normalized to creatinine did not follow the same trend and actually

387

rose. Although creatinine has been used as a biomarker in clinical chemistry for decades,

388

recent studies have indicated that this compound is not suitable for population

389

quantification purposes (Chen et al., 2014; Thai et al., 2014) because of its poor stability in

390

wastewater.

AC C

EP

TE D

M AN U

376

17

ACCEPTED MANUSCRIPT 391

The mean mass loads of nicotine showed no differences between weekdays and weekends.

392

However, the mean mass loads of both nicotine metabolites were lower over weekends than

393

on weekdays (8-10%), although this difference was not significant.

RI PT

394

3.5 Spatial variability in Italy

396

The results were then grouped in three main areas (north, center and south of Italy),

397

according to the position of each city, in order to explore potential differences in the mass

398

loads (Figure 3 A and B). The normalized mass loads of caffeine were similar in the north and

399

center of Italy (12.3±1.4 and 13.1±6.7 g/day/1000 inhabitants, respectively), but significantly

400

higher in the south (16.7±6.2 g/day/1000 inhabitants, p<0.05 north vs. south and center vs.

401

south by Dunn’s multiple comparison test). The mass loads of paraxanthine and 1-

402

methylxanthine were almost identical in the south and north of the country (15.2±1.5 and

403

15.1±3.8 g/day/1000 inhabitants), while in the cities in the center they were about 30%

404

lower (11.2±4.7 g/day/1000 inhabitants). The differences north vs. center and center vs.

405

south were significant for paraxanthine (p<0.001 by Dunn’s multiple comparison test) and 1-

406

methylxanthine (p<0.01 by Dunn’s multiple comparison test). The mass loads of 7-

407

methylxanthine decreased in the following order: north > south > center. The difference was

408

significant only between north (13.9±3.1 g/day/1000 inhabitants) and center (8.4±4.0

409

g/day/1000 inhabitants), with p <0.0001 by Dunn’s multiple comparison test. The mass load

410

profiles were lower in the center of Italy for all the metabolites, while the mass loads of

411

caffeine were higher in the south. These results may have been biased by the high variability

412

among the cities investigated in the center of Italy (Figure 1A), which include both the lowest

413

mass loads, Pescara and Perugia, and two of the highest, Florence and Rome. Moreover, the

AC C

EP

TE D

M AN U

SC

395

18

ACCEPTED MANUSCRIPT multiple sources of these substances could make it even harder to understanding these

415

profiles. Additional data are therefore required to verify the trends.

416

The normalized mass loads of nicotine and metabolites were higher in the south than in the

417

center and north of Italy (Figure 3B). The nicotine mean mass load in the south was 1.7±0.6

418

g/day/1000 inhabitants, higher than in the north (1.2±0.02 g/day/1000 inhabitants, p<0.01

419

by Dunn’s multiple comparison test) and center (1.1±0.6 g/day/1000 inhabitants, p<0.001 by

420

Dunn’s multiple comparison test). Cotinine mass loads ranged from 0.5±0.1 and 0.4±0.1

421

g/day/1000 inhabitants respectively in the north and center to 0.7±0.2 g/day/1000

422

inhabitants in the south (p<0.01 north vs. south and p<0.0001 center vs. south by Dunn’s

423

multiple comparison test). Trans-3’-hydroxycotinine mass loads ranged from 1.4±0.3 and

424

1.1±0.4 g/day/1000 inhabitants respectively in the north and center to 1.6±0.4 g/day/1000

425

inhabitants in the south (p<0.05 north vs. south and p<0.001 center vs. south by Dunn’s

426

multiple comparison test). These wastewater analysis findings reflect the recent prevalence

427

data on smoking habits in Italy (Gallus et al., 2013) which indicated a higher prevalence in

428

the south of Italy (23.3% smokers in the population > 15 years, versus 19.6% and 19.0% in

429

the north and center of Italy). This encouraging result indicates the possibility of tracking

430

smoking habits in a population by measuring nicotine metabolites in urban wastewater

431

(Castiglioni et al., 2015).

SC

M AN U

TE D

EP

AC C

432

RI PT

414

433

3.6. Caffeine, nicotine and their metabolites as population biomarkers

434

The potential of caffeine, nicotine and their metabolites as human biomarkers for the

435

assessment of population size was evaluated by taking Milan and Como as case studies.

19

ACCEPTED MANUSCRIPT Figure 4 shows the complete profiles of daily mass loads of caffeine, nicotine and

437

metabolites in Milan, where two 18-day monitoring campaigns were available, and the

438

corresponding BOD and COD mass loads, reported for comparison. The daily loads of

439

caffeine were very similar to those of its metabolites (paraxanthine, 1-methyxanthine, 7-

440

methylxanthine). The loads decreased slightly over the weekends, as previously observed

441

(Fig. 2). In contrast, nicotine loads showed higher variability than its metabolites (cotinine

442

and trans-3’-hydroxycotinine) and were higher than expected from nicotine metabolism

443

(Castiglioni et al., 2015). As already discussed, this might be due to the presence of

444

unspecific sources other than human excretion. Moreover, the peaks of nicotine loads

445

coincided with rain occurred several times during sampling (Fig. 4). The same pattern was

446

observed for COD which were higher on rainy days, and particularly each first rainy day (Fig.

447

4). This can be expected from the general wash-out when it rains, which can increase the

448

amount of organic material entering the sewage system, as well as the amount of nicotine

449

probably with cigarette butts and ash washed out from the streets.

450

The profiles of all the substances investigated were very similar indicating a similar source

451

which is likely to be human excretion (no peaks ascribable to additional sources were

452

observed, except for nicotine and COD). Moreover, the profiles moslty decreased during the

453

weekend and this can be due either to the decrease of consumption or to the decrease of

454

the population consuming these substances. In Milan, which is a big city with commuters

455

travelling every day in and out, the latter hypothesis is more probable and could also

456

support our initial thesis that these substances can be used as qualitative biomarkers to

457

check the population size and identify some population dynamics. Similar results were found

458

by Brewer et al. (2012) in USA as previously discussed (Paragraph 3.4.1).

AC C

EP

TE D

M AN U

SC

RI PT

436

20

ACCEPTED MANUSCRIPT The key issue is the possibility of using these substances also as quantitative biomarkers for

460

population size assessment, as already done for BOD and COD (Castiglioni et al., 2013). So

461

far, only cotinine has been tested as a population biomarker and appeared eligible,

462

especially for comparisons of areas with similar smoking prevalence and for temporal

463

comparisons of the same area (Chen et al., 2014).

464

Since we found that the back-calculation of tobacco use by measuring nicotine metabolites

465

matched the prevalence from population surveys and confirmed epidemiological reports

466

(Castiglioni et al., 2015), we used these metabolites to estimate the daily population in Milan

467

and Como, in the north of Italy. These cities were chosen because they have different

468

demographic characteristics and composition of urban sewage. Milan is a big city with a

469

large number of daily commuters and no industrial input in sewage (Table S2). Como is

470

smaller, with few commuters and an important contribution of the textile industry in sewage

471

(Table S2). The amounts of nicotine consumed (nicotine equivalents) were calculated by

472

multiplying the sum of cotinine and trans-3’-hydroxycotinine loads by a correction factor

473

(1.35) developed from analysis of the excretion rates of the two metabolites, as described

474

elsewhere (Castiglioni et al., 2015). The numbers of cigarettes and smokers were obtained

475

considering, respectively, the average amount of nicotine systemically absorbed from one

476

cigarette (1.25 mg) (Hukkanen et al., 2005), and the average number of cigarettes consumed

477

daily by a smoker in the north of Italy (12.8 ± 6.8) (Gallus et al., 2013). Finally, the population

478

size was calculated from the number of smokers, considering the prevalence of smokers in

479

the population aged >15 years available from population surveys (19.6% in the north of Italy)

480

and the population aged < 14 years in Milan and Como (ISTAT, 2011) (Table 3 and raw data

481

are in Table S5 and S6).

AC C

EP

TE D

M AN U

SC

RI PT

459

21

ACCEPTED MANUSCRIPT No census data of people served by this STP are available for Milan and to date the most

483

reliable population estimates have been calculated from BOD, even if these can

484

overestimate population, because they depend on the content of organic matter in sewage,

485

which can come from multiple sources. Mean population estimates from nicotine

486

metabolites (830,000 estimated inhabitants) are slightly lower than those obtained from

487

BOD and COD (1,090,000 estimated inhabitants), but generally there is good agreement.

488

Considering that Milan is a difficult case because of the numbers of commuters staying in

489

this area only during the day and/or coming for nightlife over the weekend, it is likely that

490

nicotine metabolites account for the real number of people in this area but further research,

491

including additional sampling campaigns over longer periods, will be necessary to confirm

492

this.

493

In Como, mean results from nicotine metabolites (88,271 estimated inhabitants) agree with

494

the known census data (91,344 residents), and are far lower than those from BOD and COD

495

(174,000 and 128,000 estimated inhabitants). This demonstrates that when wastewater

496

receives industrial wastes, as in Como, the hydrochemical parameters BOD and COD, are

497

greatly affected and cannot be used to estimate the population served by the plant. This

498

case study does indicate that nicotine metabolites can be suitable quantitative biomarkers to

499

estimate the real population served by a STP.

SC

M AN U

TE D

EP

AC C

500

RI PT

482

501

4. Conclusions

502

This study showed the widespread presence of caffeine and nicotine and their main

503

metabolites in Italian wastewater in μg/L concentrations. Some regional differences were

504

observed in the mass loads, especially for nicotine and its metabolites, whose normalized 22

ACCEPTED MANUSCRIPT mass loads were significantly higher in the south of Italy than in the center and north. No

506

significant weekly trend was observed, although the mean mass loads of caffeine and its

507

metabolites were lower during the weekend, probably due to less consumption of coffee.

508

To the best of our knowledge, this is the first study investigating the daily profiles of all these

509

potential human biomarkers in order to verify their suitability for estimating population size.

510

The daily profiles of caffeine and nicotine metabolites in Milan were similar, indicating that

511

they can be used as qualitative biomarkers to check the population size and identify

512

population dynamics. Nicotine metabolites were also tested as quantitative biomarkers to

513

estimate population size and good agreement was found with the resident population in

514

Como, which is extremely encouraging for extending research to a larger number of case

515

studies. Additional information is needed on caffeine metabolism in relation to the different

516

sources of its main metabolites to test these compounds as quantitative biomarkers of

517

population. This exploratory study can therefore open the way to the routine use of some of

518

these substances for estimating population size and dynamics.

TE D

M AN U

SC

RI PT

505

519

Acknowledgements

521

The study was supported by Fondazione Cariplo (Grant 2009-3468-2009-3513) and

522

Dipartimento Politiche Antidroga (Presidenza del Consiglio dei Ministri, Rome) – Project

523

Aqua Drugs. Part of this study has received funding from the European Union’s Sevenths

524

Framework Programme for research, technological development and demonstration under

525

grant agreement no [317205]. We acknowledge the personnel from sewage treatments

526

plants who helped us for samples collection and Judith Baggott for English revision.

AC C

EP

520

527 23

ACCEPTED MANUSCRIPT

References

529

Andreottola, G., Bonomo, L., Poggiali, L. and Zaffaroni, C. 1994. A methodology for the

530

estimation of unit nutrient and organic loads from domestic and non-domestic

531

sources. European Water Pollution Control 4(6), 13-19.

534 535 536

167(6), 813-818.

Baselt, R. (ed) 2004. Disposition of toxic drugs and chemicals in man., Biomedical

SC

533

Barnes, P.J. 2003. Theophylline: new perspectives for an old drug. Am J Respir Crit Care Med

Publications, California, USA.

Brewer, A.J., Ort, C., Banta-Green, C.J., Berset, J.D. and Field, J.A. 2012. Normalized diurnal

M AN U

532

RI PT

528

537

and between-day trends in illicit and legal drug loads that account for changes in

538

population. Environ Sci Technol 46(15), 8305-8314.

539

Bueno, M.J., Ucles, S., Hernando, M.D., Davoli, E. and Fernandez-Alba, A.R. 2011. Evaluation of selected ubiquitous contaminants in the aquatic environment and their

541

transformation products. A pilot study of their removal from a sewage treatment

542

plant. Water Res 45(6), 2331-2341.

545 546 547 548

EP

544

Buerge, II, Poiger, T., Muller, M.D. and Buser, H.R. 2003. Caffeine, an anthropogenic marker for wastewater comtamination of surface waters. Environ Sci Technol 37(4), 691-700.

AC C

543

TE D

540

Buerge, I.J., Kahle, M., Buser, H.R., Muller, M.D. and Poiger, T. 2008. Nicotine derivatives in wastewater and surface waters: application as chemical markers for domestic wastewater. Environ Sci Technol 42(17), 6354-6360.

Byrd, G.D., Chang, K.M., Greene, J.M. and deBethizy, J.D. 1992. Evidence for urinary

549

excretion of glucuronide conjugates of nicotine, cotinine, and trans-3'-

550

hydroxycotinine in smokers. Drug Metab Dispos 20(2), 192-197.

24

ACCEPTED MANUSCRIPT 551

Castiglioni, S., Zuccato, E., Crisci, E., Chiabrando, C., Fanelli, R. and Bagnati, R. 2006.

552

Identification and measurement of illicit drugs and their metabolites in urban

553

wastewater by liquid chromatography-tandem mass spectrometry. Anal Chem

554

78(24), 8421-8429. Castiglioni, S., Bijlsma, L., Covaci, A., Emke, E., Hernandez, F., Reid, M., Ort, C., Thomas, K.V.,

556

van Nuijs, A.L., de Voogt, P. and Zuccato, E. 2013. Evaluation of uncertainties

557

associated with the determination of community drug use through the measurement

558

of sewage drug biomarkers. Environ Sci Technol 47(3), 1452-1460.

SC

Castiglioni, S., Thomas, K.V., Kasprzyk-Hordern, B., Vandam, L. and Griffiths, P. 2014. Testing

M AN U

559

RI PT

555

560

wastewater to detect illicit drugs: state of the art, potential and research needs. Sci

561

Total Environ 487, 613-620.

562

Castiglioni, S., Senta, I., Borsotti, A., Davoli, E. and Zuccato, E. 2015. A novel approach for monitoring tobacco use in local communities by wastewater analysis. Tobacco

564

Control 24(1), 38-42.

565

TE D

563

Chen, C., Kostakis, C., Gerber, J.P., Tscharke, B.J., Irvine, R.J. and White, J.M. 2014. Towards finding a population biomarker for wastewater epidemiology studies. Sci Total

567

Environ 487, 621-628.

569 570 571 572 573

AC C

568

EP

566

D'Ascenzo, G., Di Corcia, A., Gentili, A., Mancini, R., Mastropasqua, R., Nazzari, M. and Samperi, R. 2003. Fate of natural estrogen conjugates in municipal sewage transport and treatment facilities. Sci Total Environ 302(1-3), 199-209.

Daughton, C.G. 2012. Real-time estimation of small-area populations with human biomarkers in sewage. Sci Total Environ 414, 6-21. Debry, G. 1994. Coffee and Health, John Libbey Eurotext, Paris.

25

ACCEPTED MANUSCRIPT 574

Focazio, M.J., Kolpin, D.W., Barnes, K.K., Furlong, E.T., Meyer, M.T., Zaugg, S.D., Barber, L.B.

575

and Thurman, M.E. 2008. A national reconnaissance for pharmaceuticals and other

576

organic wastewater contaminants in the United States — II) Untreated drinking water

577

sources. Sci Total Environ 402, 201-216

579

Gallus, S., Lugo, A., Colombo, P., Pacifici, R. and La Vecchia, C. 2013. Smoking prevalence in

RI PT

578

Italy 2011 and 2012, with a focus on hand-rolled cigarettes. Prev Med 56(5), 314-318. Garattini, S. 1993. Caffeine, Coffee and Health, Raven Press, New York.

581

Huerta-Fontela, M., Galceran, M.T. and Ventura, F. 2007. Ultraperformance liquid

SC

580

chromatography-tandem mass spectrometry analysis of stimulatory drugs of abuse in

583

wastewater and surface waters. Anal Chem 79(10), 3821-3829.

584

M AN U

582

Huerta-Fontela, M., Galceran, M.T., Martin-Alonso, J. and Ventura, F. 2008. Occurrence of psychoactive stimulatory drugs in wastewaters in north-eastern Spain. Sci Total

586

Environ 397(1-3), 31-40.

589 590 591 592 593 594

nicotine. Pharmacol Rev 57(1), 79-115.

Italian National Institute of Statistics (ISTAT) 2011. Censimento Popolazione 2011. Rome,

EP

588

Hukkanen, J., Jacob, P., 3rd and Benowitz, N.L. 2005. Metabolism and disposition kinetics of

Italy. http://www.istat.it/it/ (accessed 16 December 2014).

AC C

587

TE D

585

Krul, C. and Hageman, G. 1998. Analysis of urinary caffeine metabolites to assess biotransformation enzyme activities by reversed-phase high-performance liquid chromatography. J Chromatogr B Biomed Sci Appl 709(1), 27-34.

Lai, F.Y., Ort, C., Gartner, C., Carter, S., Prichard, J., Kirkbride, P., Bruno, R., Hall, W.,

595

Eaglesham, G. and Mueller, J.F. 2011. Refining the estimation of illicit drug

596

consumptions from wastewater analysis: co-analysis of prescription pharmaceuticals

597

and uncertainty assessment. Water Res 45(15), 4437-4448. 26

ACCEPTED MANUSCRIPT 598

Reid, M.J., Langford, K.H., Morland, J. and Thomas, K.V. 2011. Analysis and interpretation of

599

specific ethanol metabolites, ethyl sulfate, and ethyl glucuronide in sewage effluent

600

for the quantitative measurement of regional alcohol consumption. Alcohol Clin Exp

601

Res 35(9), 1593-1599.

605 606

RI PT

604

healthy adults. Scand J Clin Lab Invest 56(4), 373-383.

Rodopoulos, N. and Norman, A. 1997. Elimination of theophylline metabolites in healthy adults. Scand J Clin Lab Invest 57(3), 233-240.

SC

603

Rodopoulos, N., Hojvall, L. and Norman, A. 1996. Elimination of theobromine metabolites in

Rosal, R., Rodriguez, A., Perdigon-Melon, J.A., Petre, A., Garcia-Calvo, E., Gomez, M.J.,

M AN U

602

607

Aguera, A. and Fernandez-Alba, A.R. 2010. Occurrence of emerging pollutants in

608

urban wastewater and their removal through biological treatment followed by

609

ozonation. Water Res 44(2), 578-588.

Santos, J.L., Aparicio, I., Callejon, M. and Alonso, E. 2009. Occurrence of pharmaceutically

TE D

610 611

active compounds during 1-year period in wastewaters from four wastewater

612

treatment plants in Seville (Spain). J Hazard Mater 164(2-3), 1509-1516. Srdjenovic, B., Djordjevic-Milic, V., Grujic, N., Injac, R. and Lepojevic, Z. 2008. Simultaneous

614

HPLC determination of caffeine, theobromine, and theophylline in food, drinks, and

616 617

AC C

615

EP

613

herbal products. J Chromatogr Sci 46(2), 144-149.

Teijon, G., Candela, L., Tamoh, K., Molina-Diaz, A. and Fernandez-Alba, A.R. 2010. Occurrence of emerging contaminants, priority substances (2008/105/CE) and heavy

618

metals in treated wastewater and groundwater at Depurbaix facility (Barcelona,

619

Spain). Sci Total Environ 408, 3584-3595.

27

ACCEPTED MANUSCRIPT 620

Thai, P.K., O'Brien, J., Jiang, G., Gernjak, W., Yuan, Z., Eaglesham, G. and Mueller, J.F. 2014.

621

Degradability of creatinine under sewer conditions affects its potential to be used as

622

biomarker in sewage epidemiology. Water Res 55C, 272-279.

623

Thomas, K.V. and Reid, M.J. 2011. What else can the analysis of sewage for urinary biomarkers reveal about communities? Environ Sci Technol 45(18), 7611-7612.

625

Tsuzuki, Y. 2006. An index directly indicates land-based pollutant load contributions of

RI PT

624

domestic wastewater to the water pollution and its application. Sci Total Environ

627

370(2-3), 425-440.

van Nuijs, A.L., Castiglioni, S., Tarcomnicu, I., Postigo, C., Lopez de Alda, M., Neels, H.,

M AN U

628

SC

626

629

Zuccato, E., Barcelo, D. and Covaci, A. 2011. Illicit drug consumption estimations

630

derived from wastewater analysis: a critical review. Sci Total Environ 409(19), 3564-

631

3577.

633 634

World Health Organization (WHO) 2013. WHO Report on global tobacco epidemic, 2013.

TE D

632

World Health Organization, Geneve, Switzerland. Wong, P., Villeneuve, G., Tessier, V., Banerjee, K., Nedev, H., Jean-Claude, B.J. and LeylandJones, B. 2002. Stability of 5-acetamido-6-formylamino-3-methyluracil in buffers and

636

urine. J Pharm Biomed Anal 28(3-4), 693-700.

638 639 640

AC C

637

EP

635

Zuccato, E., Chiabrando, C., Castiglioni, S., Calamari, D., Bagnati, R., Schiarea, S. and Fanelli, R. 2005. Cocaine in surface waters: a new evidence-based tool to monitor community drug abuse. Environ Health 4, 14.

Zuccato, E., Chiabrando, C., Castiglioni, S., Bagnati, R. and Fanelli, R. 2008. Estimating

641

community drug abuse by wastewater analysis. Environ Health Perspect 116(8),

642

1027-1032.

643 28

ACCEPTED MANUSCRIPT Figures Captions

645

Fig. 1. Normalized mass loads (g/day/1000 inhabitants) of caffeine, nicotine and their

646

metabolites in 13 Italian cities. Results from the September-October 2012 sampling

647

campaign.

648

Fig. 2. Weekly mass loads profiles of caffeine, nicotine and their metabolites.

649

Fig. 3. Normalized mass loads (g/day/1000 inhabitants) of caffeine, nicotine and their

650

metabolites in north, center and south of Italy. *= p<0.05; **= p<0.01; ***p<0.001;

651

****=p<0.0001.

652

Fig. 4. Profiles of the population biomarkers investigated in Milan for two 18-day monitoring

653

campaigns.* = rain.

AC C

EP

TE D

M AN U

SC

RI PT

644

29

ACCEPTED MANUSCRIPT

Linearity IS used for

Coefficient of

Recovery

Repeatability

IQL

MQL

correlation (r2)

(%)

RSD (%)

(pg/injected)

(ng/L)

89

6

0.37

0.43

0.9998

70

6

4.2

11.5

0.9998

87

8

1.0

1.9

range quantification (ng/mL) Cotinine-d3

0-600

Nicotine

Nicotine-d3

0-600

Cotinine-d3

0-600

trans-3’hydroxycotinine

0.9993

M AN U

Cotinine

SC

Compound

RI PT

Table 1. Linearities, recoveries, repeatabilities and quantification limits (IQL and MQL).

13

0-600

0.9989

88

12

1.4

3.6

1,7-dimethylxanthine

13

0-600

0.9996

76

5

2.4

6.6

1-methylxanthine

13

0-600

0.9996

72

14

5.9

6.1

7-methylxanthine

13

0-600

0.9999

64

10

9.9

28.5

C3-Caffeine C3-Caffeine

AC C

C3-Caffeine

EP

C3-Caffeine

TE D

Caffeine

30

ACCEPTED MANUSCRIPT

1Methylxanthine

7Methylxanthine

Caffeine

Paraxanthine

North

Milan

25.3 ± 8.8

24.3 ± 10.2

14.8 ± 6.2

27.0 ± 7.0

Como

27.3 ± 6.1

30.4 ± 8.2

23.2 ± 7.4

31.6 ± 14.8

Bologna

54.8 ± 7.5

72.1 ± 8.0

57.4 ± 7.8

Turin

36.4 ± 2.6

41.9 ± 5.1

25.5 ± 5.0

Verona

41.7 ± 10.7

58.3 ± 15.6

33.8 ± 10.3

Pescara

26.1 ± 2.6

26.7 ± 2.9

Florence

33.6 ± 3.7

17.5 ± 3.4

Perugia

38.0 ± 10.0

Rome

Bari

South

2.40 ± 0.92

M AN U

2.97 ± 0.72

Cotinine

trans-3’hydroxycotinine

1.05 ± 0.45

2.58 ± 0.50

1.09 ± 0.29

2.93 ± 0.78

6.08 ± 1.00

2.45 ± 0.43

6.71 ± 1.02

35.1 ± 3.7

3.19 ± 0.43

1.81 ± 0.11

4.71 ± 0.39

52.2 ± 13.0

4.48 ± 1.05

1.63 ± 0.47

4.34 ± 1.27

18.6 ± 2.5

25.3 ± 5.6

2.31 ± 0.20

1.07 ± 0.14

2.55 ± 0.41

5.00 ± 1.08

7.29 ± 0.96

3.21 ± 0.55

2.56 ± 0.40

4.73 ± 0.63

41.2 ± 12.1

24.0 ± 9.7

39.2 ± 12.6

3.77 ± 1.03

2.29 ± 0.54

5.19 ± 1.51

17.6 ± 12.3

17.7 ± 10.9

11.7 ± 7.0

18.8 ± 9.9

1.36 ± 0.71

0.65 ± 0.33

2.14 ± 1.14

52.7 ±

77.3 ± 8.4

61.9 ± 4.7

64.2 ± 10.2 31

6.87 ± 0.48

3.12 ±

7.00 ± 0.34

EP

TE D

84.1 ± 7.9

AC C

Center

Nicotine

SC

Cities and region

RI PT

Table 2. Mean concentrations (µg/L) ± standard deviation (SD) of the selected compounds in wastewater in Italy.

ACCEPTED MANUSCRIPT

11.9

2.37 ± 0.38

5.38 ± 0.86

6.43 ± 0.71

2.59 ± 0.56

6.01 ± 1.15

2.20 ± 1.00

0.99 ± 0.39

2.56 ± 0.95

52.0 ± 8.0

31.6 ± 4.9

35.5 ± 6.2

6.00 ± 1.39

Naples

51.4 ± 13.2

31.7 ± 14.7

13.2 ± 5.3

24.9 ± 3.8

Potenza

27.9 ± 12.6

25.2 ± 10.1

12.0 ± 8.1

19.6 ± 9.6

RI PT

67.6 ± 9.0

SC

Palermo

0.15

AC C

EP

TE D

M AN U

Number of samples: Milan: 58 for caffeine, paraxanthine, 1-methylxanthine, nicotine and cotinine, 36 for 7-methylxanthine and trans-3'-hydroxycotinine; Como, Rome and Naples: 14 for caffeine, paraxanthine, 1-methylxanthine, nicotine and cotinine, 7 for 7-methylxanthine and trans-3'-hydroxycotinine; other cities: 7 for all analytes.

32

ACCEPTED MANUSCRIPT

Table 3. Estimated population size in Milan and Como using cotinine and trans-3’-hydroxycotinine loads measured in raw wastewater and comparison with the figures from BOD and COD. Como (resident population 91,344)

RI PT

Milan (resident population 1,100,000 from BOD) Inhabitants from nicotine a

Inhabitants from

Inhabitants from

Inhabitants from

BOD

COD

893,97

255,127

174,423

745,367

81,816

210,948

131,380

842,921

90,677

196,793

145,838

694,525

80,858

170,250

137,903

1,135,196

119,979

143,325

141,619

1,331,054

88,504

130,935

95,084

1,035,542

66,667

107,572

72,186

973,188 ± 22,6458

88,271 ± 16,207

173,564 ± 51,133

128,347 ± 34074

COD

774,419

1,183,800

1,027,708

Tuesday

769,038

835,733

Wednesday

900,767

970,683

Thursday

825,719

874,100

Friday

949,411

1,204,072

Saturday

813,522

1,411,813

Sunday

785,971

1,171,950

Mean

831,264 ± 68,666

TE D

EP

Monday

AC C

metabolites

metabolites

M AN U

from BOD

week

nicotine

SC

Day of the

Inhabitants from Inhabitants

1,093,165 ± 207,172

a

Calculated from the means of a 18 day monitoring campaign in September-October 2012.

b

Calculated from a 7-day monitoring campaign in October 2012.

33

b

30

Caffeine

Paraxanthine

RI PT

20 15 10

SC

5 0

2.5 2.0 1.5 1.0

trans-3'-hydroxycotinine

AC C

0.5

Cotinine

TE D

3.0

Nicotine

EP

3.5

Loads g/day/1000 inhabitants

7-Methylxanthine

25

A

B

1-Methylxanthine

M AN U

Loads g/day/1000 inhabitants

ACCEPTED MANUSCRIPT

0.0

Fig. 1. Normalized mass loads (g/day/1000 inhabitants) of caffeine, nicotine and their metabolites in 13 Italian cities. Results from the September-October 2012 sampling campaign.

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

A

B Fig. 2. Weekly mass loads profiles of caffeine, nicotine and their metabolites.

North ITALY

25

20

****

15

RI PT

*** **

10

SC

5

0 Paraxanthine

A North ITALY

2.0

1.0

AC C

0.5

**

EP

1.5

***

1-methylxanthine

Centre ITALY

TE D

2.5

Loads (g/day/1000 inhabitants)

South ITALY

*

Caffeine

7-methylxanthine

**

South ITALY

* ***

****

0.0

Nicotine

B

Centre ITALY

M AN U

Loads (g/day/1000 inhabitants)

ACCEPTED MANUSCRIPT

Cotinine

trans-3'-hydroxycotinine

ACCEPTED MANUSCRIPT

Fig. 3. Normalized mass loads (g/day/1000 inhabitants) of caffeine, nicotine and their metabolites in north, center and south of Italy. *= p<0.05; **= p<0.01; ***p<0.001;

AC C

EP

TE D

M AN U

SC

RI PT

****=p<0.0001.

*

*

* *

*

TE D

M AN U

SC

*

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

Fig. 4. Profiles of the population biomarkers investigated in Milan for two 18-day monitoring campaigns.* = rain.

ACCEPTED MANUSCRIPT

Highlights

Wastewater analysis was applied to monitor the use of caffeine and nicotine in Italy Caffeine and nicotine metabolites were tested as biomarkers for population size assessment

RI PT

The selected compounds showed a widespread occurrence at the μg/L concentration level Mass loads of nicotine metabolites indicate the highest use in the south of Italy

AC C

EP

TE D

M AN U

SC

Nicotine metabolites resulted good quantitative biomarkers for estimating population size

ACCEPTED MANUSCRIPT

Supplementary Information

Wastewater analysis to monitor use of caffeine and nicotine and evaluation of

RI PT

their metabolites as biomarkers for population size assessment

Rudjer Boskovic Institute, Division for Marine and Environmental Research, Bijenicka c. 54, 10000

M AN U

1

SC

Ivan Senta1, Emma Gracia-Lor2, Andrea Borsotti2, Ettore Zuccato2, Sara Castiglioni2*

Zagreb, Croatia. 2

IRCCS – Istituto di Ricerche Farmacologiche “Mario Negri”, Department of Environmental Health

* Corresponding author: Dr. Sara Castiglioni,

TE D

Sciences, Via La Masa 19, 20156, Milan, Italy.

EP

Head of the Environmental Biomarkers Unit

Department of Environmental Health Sciences

AC C

IRCCS-Istituto di Ricerche Farmacologiche "Mario Negri" Via La Masa 19, 20156 Milan, Italy Tel: +39 02 39014776

Fax: +39 02 39014735

e-mail: [email protected]

1

ACCEPTED MANUSCRIPT Human metabolism of caffeine Caffeine, 1,3,7-trimethylxanthine, is a weekly basic alkaloid found in seeds, leaves and fruit of more than 60 plants (Heckman et al., 2010). It occurs in many products extensively consumed on daily-bases, such as coffee, tea, energy and soft drinks. Caffeine is a mild central nervous system

RI PT

and metabolic stimulant which reduce physical fatigue, restore mental alertness and increase metabolism rate (Smit and Rogers, 2002). It also produces diuresis, myocardial and respiratory stimulation and coronary vessel dilation (Baselt, 2004). Caffeine is found in many over-the-counter and prescription drugs, as well as dietary supplements. It is administered orally in analgesics

SC

mixtures, migraine remedies and antisoporific preparations. Occasionally is also used

intravenously for treatment of some other conditions, like apnea in premature infants.

M AN U

Caffeine is rapidly metabolized in the liver by the cytochrome P450 oxidase enzyme system (Krul and Hageman, 1998) and excreted mostly via urine (Callahan et al., 1982). First, it undergoes three oxidative N-demethylations, mostly 3-demethylation, to form 1,7-dimethylxanthine (paraxanthine). This metabolite then undergoes three different reactions: 8-hydroxylation to form 1,7-dimethyluric acid, 7-demethylation to form 1-methylxanthine and formation of the open ring

TE D

product 5-acetylamino-6-formylamino-3-methyluracil (AFMU), unstable product which can be deformylated nonenzymatically to 5-acetylamino-6-amino-3-methyluracil (AAMU). Part of 1methylxanthine is further metabolized to 1-methyluric acid. Two other demethylation products of

EP

caffeine, teobromine (3,7-dimethylxanthine) and teophylline (1,3-dimethylxanthine) are also naturally-present alkaloids. 7-methylxanthine, another caffeine metabolite, can be formed both by demethylation of theobromine and paraxanthine (Blanchard, 1985). Overall, metabolite of

AC C

caffeine is quite complex and variable (Crews et al, 2001; Grant et al, 1983), with several other minor metabolites (Garattini, 1993). The major metabolic pathways are presented at Fig. S1.

2

ACCEPTED MANUSCRIPT O

CH3 N

N

O

N

N

Caffeine

CH3

H O

N N

O

CH3 N

H3C

N

O

H3C

N

N

N

N H

H

O

CH3

O

O

1,7-dimethyluric acid

O

H O N

N

N

N

CH3

N

N CHO

H

H

H3C

H N

M AN U

H3C

OH

H

7-methylxanthine

O

N

N

Paraxanthine

CH3

SC

O

RI PT

H3C

N

N

O

H

5-acetylamino-6-formylamino-3-methyluracil

TE D

O

H3C

O

N

N H

N

1-methylxanthine

H

N OH N 1-methyluric acid

AC C

highlighted).

EP

Fig. S1. Major pathways of caffeine metabolism (substances analyzed in this study are

The main metabolites of caffeine, which were selected within this study according to their pattern of urinary excretion, are shown in Table S1. The different values found in the literature and used for selection are also reported separately.

3

ACCEPTED MANUSCRIPT

Table S1. Profile of human excretion of several caffeine metabolites selected among those excreted in higher amounts.

Human metabolism of nicotine

Excretion profile from Grant et al., 1983 (n=68) 4.8 10.1 2.5* 6.0 11.8 4.3

RI PT

Excretion profile from Baselt, 2004 (n=7) 2.4 3.5 9.5 2.4 7.8 22 16

SC

Caffeine Paraxanthine 1-methylxanthine 7-methylxanthine 1,7 dimethyluric acid 1 methyluric acid AFMU *n=59

Excretion profile from Garattini, 1993 1.2 6 18 7 6 25 15

M AN U

Compound

Nicotine is an alkaloid found in tobacco and other nightshade plants. It constitutes approximately 0.5-8.0% of dry weight of tobacco. Nicotine is highly toxic and cause stimulation of autonomic ganglia and the central nervous system (Baselt, 2004). It acts as a nicotinic acetylcholine receptor

active or passive smoking.

TE D

agonist. Virtually every member of the tobacco-smoking societies is exposed to this drug, either by

The major pathways of the nicotine metabolism are presented at Fig. S2. The principal metabolite,

EP

cotinine, is formed in the liver by C-oxidation, but nicotine also undergoes N-oxidation, Ndemethylation and N-glucuronidation. Cotinine is further metabolized by several reactions:

AC C

hydroxylation to trans-3'-hydroxycotinine and, to a lesser extent, to 5'-hydroxycotinine, Noxidation to cotinine-N-1-oxide and N-glucuronidation. Trans-3'-hydroxycotinine is further metabolized by O-glucuronidation. The major urine metabolites are cotinine and trans-3'hydroxycotinine, which are excreted either in free form (13% and 35%, respectively) or as conjugates (17% and 9%, respectively) (Byrd, et al., 1992). Other minor metabolites have been identified as well, but account for less than 10% of the total nicotine metabolism (Tricker, 2003).

4

ACCEPTED MANUSCRIPT H

H

H

N

N

N O

CH3

CH3

N

N

Gluc

Nicotine

Nicotine-1'-N-oxide

CH3

N

Nicotine-N-glucuronide OH

N

O

N

O

CH3

N

H

H

H

N

CH3

N

N

O

CH3

Gluc Cotinine

H

Cotinine-N-glucuronide

SC

trans-3'-hydroxycotinine

OGluc

N

O

CH3

trans-3'-hydroxycotinine-O-glucuronide

M AN U

H

N

RI PT

H

AC C

EP

TE D

Fig. S2. Major pathways of nicotine metabolism (substances analyzed in this study are highlighted)

5

ACCEPTED MANUSCRIPT

Table S2. Main characteristics of the selected sewage treatment plants (STPs). WWTPs investigated

Daily flow rate (m3/d)

Population served by the plant

Wastewater Composition

Periods of Sampling

RI PT

26/03/12 - 01/04/12

Milan (Nosedo)

17/04/12 - 01/05/12

370,000

1,100,000

100% domestic

20/09/12 - 07/10/12 22/11/12 - 09/12/12

44,000

80% domestic; 20% industrial

SC

Como

91,344 937,000

mostly domestic

1,279,000 Naples

195,000

mostly domestic

650,000 544,000

1,370,000

Bologna

109,000

500,000

Verona

78,000

Florence

146,000

Palermo

85,000

Bari

80,000

Pescara Perugia

01/10/12 - 07/10/12 26/03/12 - 01/04/12 22/10/12 - 28/10/12 26/03/12 - 01/04/12 05/11/12 - 11/11/12

70% domestic; 30% industrial

17/10/12 - 23/10/12

76% domestic; 24% industrial

15/10/12 - 21/10/12

TE D

Turin

mostly domestic

05/11/12 - 11/11/12

200,000

93% domestic; 7% industrial

16/10/12 - 22/10/12

200,000

100% domestic

22/10/12 - 28/10/12

340,000

90% domestic; 10% industrial

15/10/12 - 21/10/12

53,000

154,000

100% domestic

15/10/12 - 21/10/12

15,000

47,800

85% domestic; 15% industrial

05/11/12 - 11/11/12

80,000

90% domestic; 10% industrial

22/10/12 - 28/10/12

EP

300,000

AC C

Potenza

M AN U

Rome

26/03/12 - 01/04/12

32,000

6

ACCEPTED MANUSCRIPT Table S3. Precursor and products ions of the analyzed compounds with the associated collision energies Precursor ion

Product ion 1 (m/z) and

Product ion 2 (m/z) and

(m/z)

collision energy (eV)

collision energy (eV)

Compound

Nicotine-d3

166.1

130 (26)

Cotinine

177.1

80 (30)

Cotinine-d3

180.1

80 (30)

trans-3’-hydroxycotinine

193.1

80 (33)

Caffeine

195.1

138 (25)

Caffeine-3C13

198.1

140 (25)

1,7-dimethylxanthine

181.1

1-methylxanthine

167.1

7-methylxanthine

167.1

117.1 (34)

RI PT

130.1 (26)

-

98 (27) -

SC

163.1

AC C

EP

TE D

M AN U

Nicotine

7

134 (24)

110 (30) -

124 (26)

96 (32)

110 (25)

82 (33)

124 (24)

150 (24)

ACCEPTED MANUSCRIPT Table S4. Stability of selected compounds in wastewater expressed as a percentage of the concentrations measured at t0 (immediately after the spiking).

-20 °C

7methylxant hine

Nicoti ne

Cotini ne

trans-3’hydroxycoti nine

97 ± 2

100 ± 3

104 ± 3

101 ± 4

102 ± 2

102 ± 3

98 ± 2

99 ± 4

98 ± 4

104 ± 2

103 ± 3

104 ± 2

101 ± 2

101 ± 5

105 ± 9

102 ± 2

4h

100 ± 6

101 ± 5

100 ± 10

106 ± 1

6h

98 ± 6

99 ± 5

104 ± 5

96 ± 4

8h

101 ± 4

100 ± 1

111 ± 7

99 ± 3

24 h

100 ± 6

99 ± 1

103 ± 6

98 ± 6

103 ± 6

106 ± 2

105 ± 2

2h

101 ± 2

100 ± 3

104 ± 4

100 ± 5

101 ± 3

100 ± 2

100 ± 1

4h

102 ± 2

103 ± 1

106 ± 4

102 ± 2

102 ± 3

102 ± 1

100 ± 3

6h

99 ± 0

103 ± 3

98 ± 8

104 ± 3

98 ± 1

104 ± 4

101 ± 4

8h

99 ± 3

101 ± 3

107 ± 8

102 ± 2

101 ± 1

103 ± 4

105 ± 3

24 h

98 ± 2

114 ± 3

106 ± 3

103 ± 3

105 ± 2

107 ± 1

7 day s

99 ± 3

97 ± 3

100 ± 4

83 ± 6

96 ± 4

101 ± 6

103 ± 5

28 day s

104 ± 1

106 ± 1

107 ± 6

85 ± 8

102 ± 5

107 ± 4

106 ± 3

107 ± 4

M AN U

SC

RI PT

2h

EP

20 °C

1methylxant hine

AC C

4 °C

Paraxant hine

TE D

Temperat Tim Caffei ure e ne

8

ACCEPTED MANUSCRIPT Table S5. Daily loads of cotinine and trans-3’-hydroxycotinine measured in raw wastewater in Milan and further procedure used to estimate the number of inhabitants. Milan

Saturday

1483

Sunday

1424

Monday

1544

Tuesday

1410

Wednesday

1821

Thursday

1449

Friday

1688 1513

Sunday

1467

Monday

1321

Tuesday

1430

Wednesday

1631

Thursday

1577

Friday

2041

Saturday

1574

Sunday

1487

1759320 1931040 1601640 1537920 1667520 1522800 1966680 1564920 1823040 1634040 1584360 1426680 1544400 1761480 1703160 2204280 1699920 1605960

TE D

Saturday

2199 2414 2002 1922 2084 1904 2458 1956 2279 2043 1980 1783 1931 2202 2129 2755 2125 2007

Number of smokers c

Number of inhabitants estimated d

RI PT

1788

Number of cigarettes b

137447 150863 125128 120150 130275 118969 153647 122259 142425 127659 123778 111459 120656 137616 133059 172209 132806 125466

SC

Friday

Nicotine equivalents (g/day)a

M AN U

Thursday

Sum of the daily loads of cotinine and trans-3’hydroxycotinine (g/day) 1629

18 day sampling (20-09-12/07-1012)

859010 927457 796159 770760 822418 764733 941663 781522 884408 809073 789271 726420 773343 859871 836624 1036369 835333 797881

AC C

EP

1571 2121 1696620 132548 834018 Mean Calculated using a correction factor (1.35) that consider human metabolism of these substances as developed in Castiglioni et al., 2014. b Calculated using the mean content of nicotine absorbed from one cigarette (1.25 mg) (Hukkanen et al., 2005). c Calculated considering the average number of cigarettes smoked per day (12.8±6.8 in the North of Italy) (Gallus et al., 2013). d Estimated considering the prevalence of smokers in the population aged > 15 years (19.6% in the North of Italy) (Gallus et al., 2013) and the population aged < 14 years in Milan (Italian National Institute of Statistics, 2011). a

9

ACCEPTED MANUSCRIPT Table S6. Daily loads of cotinine and trans-3’-hydroxycotinine measured in raw wastewater in Como and further procedure used to estimate the number of inhabitants. Como

Tuesday

164

Wednesday

184

Thursday

161

Friday

252

Saturday

179

Sunday

128

245 221 249 218 340 242 173 241

195634 176614 198845 174212 272359 193393 138608 192809

b

Number of smokers c

Number of inhabitants estimated d

15284 13798 15535 13610 21278 15109 10829 15063

89397 81816 90677 80858 119979 88504 66667 88271

RI PT

181

Number of cigarettes

SC

Monday

Nicotine equivalents (g/day).

M AN U

7 day sampling (01-10-12/07-1012)

Sum of the daily loads of cotinine and trans-3’hydroxycotinin e (g/day)

179 Mean Calculated using a correction factor (1.35) that consider human metabolism of these substances as developed in Castiglioni et al., 2014. b Calculated using the mean content of nicotine absorbed from one cigarette (1.25 mg) (Hukkanen et al., 2005). c Calculated considering the average number of cigarettes smoked per day (12.8±6.8 in the North of Italy) (Gallus et al., 2013). d Estimated considering the prevalence of smokers in the population aged > 15 years (19.6% in the North of Italy) (Gallus et al., 2013) and the population aged < 14 years in Como (Italian National Institute of Statistics, 2011).

AC C

EP

TE D

a

10

ACCEPTED MANUSCRIPT References Baselt RC. Disposition of Toxic Drugs and Chemicals in Man. 3rd ed. Foster City: Biomedical Publications; 2004.

RI PT

Blanchard J, Sawers JA, Jonkman JHG, Tang-Liu DDS. Comparison of the urinary metabolite profile of caffeine in young and elderly males. Brit J Clin Pharmaco 1985;19:225-32.

Byrd GD, Chang KM, Greene JM, Debethizy JD. Evidence for urinary excretion of glucuronide conjugates of nicotine, cotinine, and trans-3'-hydroxycotinine in smokers. Drug Metab Dispos

SC

1992;20:192-7.

Callahan MM, Robertson RS, Arnaud MJ, Branfman AR, McComish MF, Yesair DW. Human

1982;10:417-23.

M AN U

metabolism of [1-methyl-14C]- and [2-14C]caffeine after oral administration. Drug Metab Dispos

Castiglioni S, Senta I, Borsotti A, Davoli E, Zuccato E. A novel approach for monitoring tobacco use in local communities by wastewater analysis. Tob Control 2014 (in press; DOI:

TE D

10.1136/tobaccocontrol-2014-051553)

Crews HM, Olivier L, Wilson A. Urinary biomarkers for assessing dietary exposure to caffeine. Food Addit Contam 2001;18:1075-87.

EP

Gallus S, Lugo A, Colombo P, Pacifici R, La Vecchia C. Smoking prevalence in Italy 2011 and 2012, with a focus on hand-rolled cigarettes. Prev Med 2013;56:314-8.

AC C

Garattini S. Ed. Caffeine, Coffee and Health. 1993. In: Monograph of the Mario Negri Institute for Pharmacological Research, Milan. Raven Press, New York Grant DM, Tang BK, Kalow W. Variability in caffeine metabolism. Clin Pharmacol Ther 1983;33:591601.

Heckman MA, Weil J, Gonzalez de Mejia E. Caffeine (1, 3, 7-trimethylxanthine) in Foods: A Comprehensive Review on Consumption, Functionality, Safety, and Regulatory Matters. J Food Sci 2010;75:77-87.

11

ACCEPTED MANUSCRIPT Hukkanen J, Jacob P, Benowitz NL. Metabolism and disposition kinetics of nicotine. Pharmacol Rev 2005;57:79-115. Krul C, Hageman G. Analysis of urinary caffeine metabolites to assess biotransformation enzyme activities by reversed-phase high-performance liquid chromatography. J Chromatogr B

RI PT

1998;709:27-34. Smit HJ, Rogers PJ. Effects of ‘energy’ drinks on mood and mental performance: critical methodology. Food Qual Prefer 2002;13:317-26.

SC

Tricker AR. Nicotine metabolism, human drug metabolism polymorphisms, and smoking behaviour. Toxicology 2003;183:151-73.

M AN U

Italian National Institute of Statistics (ISTAT). Censimento Popolazione 2011. Rome, Italy 2011.

AC C

EP

TE D

http://www.istat.it/it/

12