Smokeless Tobacco Sub-Group - CORESTA · Smokeless Tobacco Sub-Group ... December 2015 Author and...

Smokeless Tobacco Sub-Group

Technical Report

CORESTA Reference Products

2015 Analysis

December 2015

Author and Sub-Group Coordinator:

Karl Wagner, Ph.D.

Altria Client Services

Study Coordinator:

John Bunch, Ph.D.

American Snuff Company

Co-Author and Statistical Analysis:

Michael Morton, Ph.D.

Altria Client Services

Table of Contents

1. Summary ........................................................................................................................... 3

2. Introduction ....................................................................................................................... 3

2.1 Objective ................................................................................................................. 4

3. Organisation ...................................................................................................................... 4

3.1 Participants .............................................................................................................. 4

3.2 Protocol ................................................................................................................... 5

3.2.1 Sample Shipment ......................................................................................... 5

3.2.2 Within Laboratory Sample Preparation ....................................................... 5

3.2.3 Sample Analysis and Data Reporting .......................................................... 6

4. Data – Raw ........................................................................................................................ 7

5. Data – Statistical Analysis .............................................................................................. 12

5.1 Exclusion of Outliers ............................................................................................. 12

5.2 Calculation of Repeatability (r) and Reproducibility (R) ...................................... 13

5.3 Calculation of Z-Scores ......................................................................................... 14

5.4 CRP Stability Assessment ..................................................................................... 16

6. Data Interpretations ......................................................................................................... 19

7. Recommendations ........................................................................................................... 20

APPENDIX A: Raw Data Plots for 2015 WG4 Study ............................................................ 21

APPENDIX B: CRP Stability Assessment, Mean Data Plots ................................................. 35

STS-CTR 2015 Analysis of CRPs – December 2015 3/48

1. Summary

In May 2015, the CORESTA Smokeless Tobacco Sub-Group (STS) conducted the annual

Working Group 4 (WG4) inter-laboratory study designed to assess the stability of the four

CORESTA Reference Products (CRPs). Sixteen laboratories participated in the 2015 study.

The participating laboratories reported the levels of nicotine, pH, moisture (oven volatiles),

and tobacco specific nitrosamines (TSNAs) using CORESTA Recommended Methods

(CRMs). In an initial study conducted in 2010, the STS recommended monitoring these

analytes as appropriate markers for monitoring product stability. Tabulated data are presented

along with repeatability (r), reproducibility (R) and z-scores.

The results from the stability analysis over all study years showed only a few trends.

Specifically:

In the 2014 report1, increased variability between labs was noted across the reporting

period (2010 – 2014) for pH for CRP3 and moisture for CRP1.

In the 2012/2013 report2, a statistically significant increase in moisture was noted for

CRP3 over the study years (2010-2013). This upward trend did not continue in 2014,

but was evident again with the inclusion of 2015 data.

In the 2012/2013 report, an increase in inter-laboratory variability for pH was noted

for CRP3 over the study years (2010-2013). There is still an indication that this is the

case; however, the trend did not reach the level of statistical significance in 2014 or

2015.

In 2012, there was a notable increase in inter-laboratory variability for moisture for

CRP4. Since that time the lab-to-lab differences have been much smaller, possibly

because of improved sample handling, which included grinding and homogenising an

entire pouch prior to analysis.

The results of this study demonstrate that the CRPs are suitable for continued use as reference

products; however, manufacture of new CRPs should be considered due to the dwindling

supply of the current products.

2. Introduction

In November 2008, the Smokeless Tobacco Sub-Group (STS) was established by

recommendation of the CORESTA Scientific Commission. In 2009, STS Working Group

Three (WG3) cooperated to design and manufacture four CORESTA Reference Products

(CRPs) referred to as CRP1, CRP2, CRP3, and CRP4. These products were intended as

replacements for the Smokeless Tobacco Research Products: 2S3 (Moist Snuff), 1S2 (Dry

Snuff) and 2S1 (Loose-leaf Chewing Tobacco), which were more than ten years old.

At the Amelia Island, Florida, STS meeting (October 2009), Working Group 4 (WG4) was

organized to proceed with the chemical characterization of the four CRPs. The protocol for

the first WG4 study3 was distributed in December 2009 and the study was conducted in 2010.

This study included 43 analytes and did not specify methods of analysis. Many of the results

from this study showed a wide range in analyte levels, which is not unexpected since the

methods were not harmonized. The recommendations from the 2010 study were to monitor

1 STS Technical Report: CORESTA Reference Products 2014 Analysis, January 2015.

2 STS Technical Report: CORESTA Reference Products 2012 and 2013 Analysis, July 2014.

3 STS Technical Report: CORESTA Reference Products 2010 Analysis, February 2014.


the stability of the CRPs on an annual basis, by determining the levels of nicotine, pH,

moisture (oven volatiles) and TSNAs using CORESTA Recommended Methods (CRMs) or

draft CRMs. This annual analysis would allow the STS to determine when the CRPs should

be remanufactured.

At the STS Meeting in Hamburg, Germany (May 2011), it was decided to adopt the

recommendations from the 2010 WG4 study and initiate the first annual analysis of the four

CRPs using CRMs or draft CRMs. The first WG4 study using CRMs or draft CRMs was

conducted in 2011, and a similar study has been carried annually since that time. The focus of

this report is the 2015 study. Repeatability and reproducibility (r & R) and z-scores were

calculated for the 2015 study while the stability analysis includes data from 2010-2015. As

mentioned earlier, methods of analysis were not specified for the 2010 WG4; however, the

data were considered to be similar enough for inclusion in the analysis.

The CRPs continue to be stored at -20 °C and distributed by the North Carolina State

University (NCSU) Tobacco Analytical Services Lab under the direction of Dr. Ramsey

Lewis4 and Karen Andres

5.

2.1 Objective

The 2015 participating laboratories were to provide analytical results for pH, moisture (oven

volatiles), nicotine, and the four TSNAs (N-nitrosonornicotine (NNN), N-nitrosoanatabine

(NAT), N-nitrosoanabasine (NAB) and 4-(N-nitrosomethylamino)-1-(3-pyridyl)-1-butanone

(NNK)). This work was conducted using the applicable CRMs referenced in Section 3.

These studies were conducted to support the assessment of stability of the CRPs and to

provide an assessment of inter-laboratory variability. Data were collected from the

participating laboratories and statistically evaluated in basic conformance with the

recommendations of ISO 5725-2:1994 and ISO/TR 22971:2005. The stability analysis was

conducted using the study data from 2010 through 2015. Additionally, z-scores were

calculated for the 2015 data as an additional measure of each laboratory’s performance as

compared to the results of other laboratories.

3. Organisation

3.1 Participants

A list of the participating laboratories is provided in Table 1. The laboratories are listed in

alphabetical order. Letter codes were assigned to each laboratory and do not correspond to

the order in the table below. Since the statistical analysis covers several years, the laboratory

letter codes were maintained between the tables for all years. Furthermore, not all

laboratories participated in all time points or submitted data for all analyses.

Table 1: List of Participating Laboratories in the 2015 WG4 Study

2015 WG4 Participants

Altria Client Services, United States

American Snuff Company, United States

Arista Laboratories, United States

4 [email protected]

5 [email protected]


2015 WG4 Participants

Essentra Scientific Services, United Kingdom

Global Laboratory Services, United States

Imperial Tobacco Group, Reemtsma, Germany

Imperial Tobacco Group, Seita, France

ITC Limited, India

Japan Tobacco, Inc., Japan

Labstat International, Canada

Montepaz S.A., Uruguay

R.J. Reynolds Tobacco Company, United States

Swedish Match Northern Europe, Sweden

Swedish Match, Owensboro, United States

Swisher International, United States

University of Kentucky, United States

3.2 Protocol

Specific details from the protocol are described below:

3.2.1 Sample Shipment

Laboratories were responsible for procuring approximately 150g of each of the CRPs from

NCSU immediately before starting the 2015 study. Laboratories were requested to store the

samples at approximately 4 °C upon receipt if the analyses would be conducted within one

week or to store the samples at approximately -20 °C if the analyses would be delayed. The

study was to be conducted in May through early June of 2015. Laboratories were requested to

submit data by June 9, 2015. The samples are identified in Table 2.

Table 2: Sample Identification

Product Type

CRP1 - Swedish style snus pouch

CRP2 - American-style loose moist snuff

CRP3 - American-style loose dry snuff powder

CRP4 - American-style loose-leaf chewing tobacco

3.2.2 Within Laboratory Sample Preparation

The laboratories were directed to remove the samples from cold storage the day before testing

and to not open the samples until equilibrated to ambient temperature. Once samples were

opened, they could be stored at approximately 4 °C for up to one week. Additionally:

The snus pouch (CRP1) was to be cut into two halves and placed directly into the

extraction vessel. Both the tobacco and pouch material were included in the analysis.

The moist snuff reference product (CRP2) and the dry snuff reference product (CRP3)

did not require sample grinding and were analyzed as received.


An entire pouch of the loose leaf reference product (CRP4) was to be cryoground

before removing aliquots for analysis.

3.2.3 Sample Analysis and Data Reporting

The WG4 participating laboratories were instructed to conduct triplicate analyses for the

following: pH, moisture (oven volatiles), nicotine, and TSNAs. The laboratories were

requested to use the current versions of the following CRMs:

pH: CRM N° 69, Determination of pH in Smokeless Tobacco Products

Moisture (oven volatiles): CRM N° 76, Determination of Moisture Content (Oven

Volatiles) of Smokeless Tobacco Products

Nicotine: CRM N° 62, Determination of Nicotine in Tobacco and Tobacco Products

by Gas Chromatographic Analysis

TSNAs: CRM N° 72, Determination of Tobacco Specific Nitrosamines in Smokeless

Tobacco Products by Liquid Chromatography - Tandem Mass Spectrometry

Participating laboratories were requested to document any deviations from the protocol and

the CRMs and submit the deviations with their results. The deviations are identified below as

well as if the data were withdrawn from the study.

Lab A: CRP4 was stored in a freezer at approximately -20 °C. Two frozen pouches

were ground at 2000rpm for 15 seconds with a knife mill (Grindomax GM300

Retsch).

Lab D: One pouch of CRP4 was ground at room temperature to approx. 4mm and

then mixed thoroughly prior to sampling for analysis.

Lab G: The grinder and tobacco was placed in refrigerator overnight prior to grinding

at low speed to avoid excess heat.

Lab K: CRP4 was manually cut into very small pieces and mixed thoroughly.

Lab K: CRM N° 72 was not followed for the determination of TSNAs. The

laboratory reported the use of a different extraction solvent. The data were identified

as outliers and removed from the study.

Lab K: Moisture content is measured at 110 °C for 3.5 hours. Moisture for CRP4 was

determined without particle size reduction. These data were not included in the

analysis.

Lab M: One pouch of CRP4 was ground at room temperature.

Lab N: CRP4 was not ground, but cut into smaller pieces, without sufficient

homogenization. The TSNA data were identified as extreme outliers. The participant

requested that CRP4 data for all analytes be removed.

Lab N: Moisture determination was performed by applying CRM 56, Karl Fischer.

These data were removed from the study since the participant determined water

instead of moisture.

Lab P: CRP4 was ground at room temperature in a Retsch Grindomix for 8 seconds.

Lab T: CRP4 was manually cut into smaller pieces.

Lab U: Two frozen pouches of CRP4 (0 °C) were ground in a Robo Coupe RSI 6V at

3000 rpm for 20 seconds.

Lab U: TSNAs were determined using GC-TEA instead of using CRM N° 72.

All test results were to be reported on an as-is basis with no correction for moisture content.

The results were not to be rounded and ideally reported to at least one more digit than

typically required.


The study results and the comments were to be sent by e-mail to John E. Bunch and Karl

Wagner.

4. Data – Raw

The full data set for the 2015 study is listed in Table 3. The results are presented on an as-is

basis, without correction for moisture. Each analysis includes three replicates. Raw data

plots that include all replicates, without removal of outliers, are given in Appendix A.

As mentioned above, several data sets were removed from the study due to significant

deviations from the study protocol and the fact that the data were preliminarily identified as

outlying data. Those data are included in Table 3, but were eliminated prior to analysis.

Table 3: Full Data Set for 2015 WG4 Study (results are presented on an as-is basis)

Lab Code

Product Nicotine pH Moisture NNN NNK NAT NAB

% % µg/g µg/g µg/g µg/g

A CRP1 0.96 7.87 49.98 0.72 0.22 0.57 0.033

A CRP1 0.98 7.88 49.94 0.70 0.21 0.55 0.038

A CRP1 0.97 7.89 50.00 0.72 0.21 0.57 0.038

A CRP2 1.26 7.64 54.57 1.94 0.43 1.81 0.163

A CRP2 1.27 7.64 54.68 1.85 0.45 1.67 0.157

A CRP2 1.26 7.63 54.53 1.90 0.44 1.86 0.152

A CRP3 2.13 6.79 10.87 8.64 4.39 5.84 0.444

A CRP3 2.14 6.74 10.79 8.42 4.22 5.80 0.423

A CRP3 2.13 6.72 10.63 8.46 4.45 6.04 0.427

A CRP4 1.08 5.85 22.11 2.13 0.48 1.43 0.071

A CRP4 1.11 5.82 22.13 2.09 0.46 1.38 0.068

A CRP4 1.10 5.82 22.15 2.00 0.43 1.29 0.071

D CRP1 1.07 7.88 52.11 0.61 0.21 0.41 0.032

D CRP1 0.811

7.86 50.31 0.70 0.23 0.44 0.033

D CRP1 1.02 7.88 52.27 0.65 0.22 0.42 0.029

D CRP2 1.21 7.68 54.38 1.76 0.47 1.49 0.147

D CRP2 1.21 7.67 54.36 1.89 0.49 1.53 0.161

D CRP2 1.21 7.68 54.21 1.92 0.47 1.56 0.153

D CRP3 2.07 6.76 9.50 8.20 4.41 5.23 0.401

D CRP3 2.08 6.78 9.60 8.29 4.19 4.91 0.383

D CRP3 2.03 6.77 9.40 8.47 4.12 5.09 0.384

D CRP4 1.02 5.88 21.29 1.98 0.47 1.05 0.057

D CRP4 1.03 5.91 21.19 1.94 0.46 1.03 0.061

D CRP4 1.03 5.90 21.10 2.04 0.47 1.05 0.060

F CRP1 1.03 7.82 51.82 0.86 0.22 0.46 0.034

F CRP1 0.98 7.86 51.57 0.85 0.22 0.47 0.043


Lab Code



F CRP1 0.93 7.90 51.72 0.80 0.20 0.40 0.034

F CRP2 1.29 7.63 54.60 1.94 0.43 1.47 0.146

F CRP2 1.31 7.64 54.54 1.99 0.44 1.45 0.128

F CRP2 1.28 7.57 54.29 1.99 0.41 1.44 0.128

F CRP3 2.17 6.75 10.14 8.12 4.11 4.26 0.373

F CRP3 2.22 6.71 10.12 8.16 4.09 4.33 0.360

F CRP3 2.20 6.74 10.16 8.15 4.04 4.43 0.376

F CRP4 0.93 5.82 22.09 2.64 0.41 0.98 0.053

F CRP4 1.04 5.84 21.93 3.16 0.45 1.02 0.052

F CRP4 1.07 5.88 21.91 2.68 0.45 0.97 0.058

G CRP1 0.87 7.79 52.06 0.72 0.20 0.54 0.035

G CRP1 0.91 7.81 51.99 0.71 0.22 0.59 0.039

G CRP1 0.91 7.82 52.42 0.69 0.21 0.55 0.039

G CRP2 1.24 7.68 54.48 1.81 0.44 1.64 0.161

G CRP2 1.22 7.66 54.36 1.70 0.44 1.65 0.157

G CRP2 1.23 7.65 54.42 1.74 0.43 1.63 0.160

G CRP3 2.09 6.72 10.18 8.91 4.27 5.93 0.454

G CRP3 2.08 6.72 10.16 8.61 4.28 5.83 0.458

G CRP3 2.07 6.72 10.09 8.89 4.24 5.75 0.425

G CRP4 1.15 5.93 21.98 1.98 0.50 1.31 0.070

G CRP4 1.12 5.91 21.92 2.11 0.55 1.29 0.073

G CRP4 1.13 5.92 21.85 2.01 0.51 1.24 0.070

H CRP1 1.07 7.79 49.22 0.70 0.23 0.54 0.038

H CRP1 1.08 7.80 49.07 0.72 0.24 0.54 0.039

H CRP1 1.14 7.79 49.58 0.73 0.23 0.54 0.042

H CRP2 1.35 7.61 54.28 2.03 0.49 1.78 0.160

H CRP2 1.35 7.63 54.25 1.96 0.48 1.71 0.161

H CRP2 1.33 7.62 54.08 1.98 0.48 1.76 0.160

H CRP3 2.23 6.70 10.49 9.31 4.92 5.48 0.411

H CRP3 2.27 6.71 10.63 9.13 5.18 5.65 0.421

H CRP3 2.26 6.73 10.54 9.18 5.19 5.33 0.412

H CRP4 1.12 5.87 22.78 2.03 0.45 1.23 0.063

H CRP4 1.12 5.89 22.81 1.94 0.46 1.22 0.062

H CRP4 1.12 5.92 22.88 2.04 0.47 1.19 0.061

I CRP1 0.95 8.04 51.04 0.63 0.20 0.50 0.028

I CRP1 0.98 8.00 51.67 0.62 0.20 0.52 0.029

I CRP1 0.99 7.96 50.98 0.61 0.18 0.51 0.031


Lab Code



I CRP2 1.23 7.78 54.47 1.77 0.43 1.76 0.135

I CRP2 1.25 7.75 54.67 1.76 0.44 1.72 0.142

I CRP2 1.27 7.75 54.63 1.78 0.41 1.74 0.136

I CRP3 2.17 6.80 9.90 7.99 3.83 5.46 0.330

I CRP3 2.16 6.80 9.91 8.54 3.91 5.38 0.346

I CRP3 2.17 6.80 9.89 7.98 3.75 5.50 0.333

I CRP4 1.11 5.94 22.87 1.92 0.41 1.33 0.054

I CRP4 1.10 5.95 22.88 1.98 0.43 1.28 0.051

I CRP4 1.11 5.94 22.68 1.97 0.46 1.28 0.057

K CRP1 0.98 7.67 54.60 0.49 0.13 0.34 0.025

K CRP1 1.01 7.63 51.74 0.50 0.12 0.33 0.030

K CRP1 0.99 7.62 53.10 0.47 0.12 0.32 0.025

K CRP2 1.36 7.38 54.90 1.40 0.30 1.23 0.112

K CRP2 1.36 7.44 54.90 1.35 0.29 1.18 0.114

K CRP2 1.36 7.42 54.70 1.38 0.29 1.18 0.113

K CRP3 2.30 6.68 11.90 6.70 3.54 4.19 0.339

K CRP3 2.27 6.71 12.10 6.20 3.51 4.09 0.332

K CRP3 2.26 6.70 12.20 6.71 3.39 3.99 0.326

K CRP4 1.22 5.87 22.80 1.37 0.27 0.78 0.041

K CRP4 1.24 5.93 22.80 1.35 0.26 0.80 0.039

K CRP4 1.20 5.89 22.80 1.34 0.26 0.78 0.036

L CRP1 1.02 7.96 51.35 0.65 0.22 0.54 <LOQ

L CRP1 1.03 7.98 51.50 0.66 0.21 0.52 <LOQ

L CRP1 1.03 7.98 51.59 0.68 0.21 0.54 <LOQ

L CRP2 1.29 7.72 54.36 1.80 0.44 1.88 0.154

L CRP2 1.29 7.73 54.58 1.79 0.45 1.87 0.151

L CRP2 1.29 7.72 54.42 1.80 0.46 1.88 0.152

L CRP3 2.19 6.72 9.50 8.09 3.94 5.96 0.385

L CRP3 2.19 6.71 9.34 8.25 3.76 5.87 0.383

L CRP3 2.22 6.73 9.15 8.18 3.82 5.91 0.380

L CRP4 1.11 5.96 22.38 1.86 0.41 1.26 0.056

L CRP4 1.11 5.94 22.32 1.91 0.43 1.27 0.056

L CRP4 1.14 5.94 22.31 1.91 0.41 1.27 0.057

M CRP1 1.09 8.00 46.46 0.64 0.21 0.53 0.030

M CRP1 1.08 7.97 48.81 0.68 0.22 0.56 0.036

M CRP1 0.93 7.95 45.93 0.57 0.18 0.48 0.028

M CRP2 1.30 7.75 54.45 1.63 0.44 1.79 0.146


Lab Code



M CRP2 1.30 7.79 53.95 1.73 0.44 1.78 0.130

M CRP2 1.29 7.78 54.02 1.65 0.46 1.77 0.140

M CRP3 2.18 6.88 9.44 8.48 4.66 6.12 0.428

M CRP3 2.20 6.88 9.21 7.59 3.79 5.57 0.353

M CRP3 2.19 6.93 9.56 7.42 3.73 5.38 0.370

M CRP4 1.05 6.10 23.10 1.81 0.43 1.18 0.049

M CRP4 1.03 6.07 22.92 1.79 0.39 1.17 0.054

M CRP4 1.08 6.08 22.51 1.72 0.40 1.18 0.055

N CRP1 0.92 7.81 NA 0.65 0.16 0.46 0.030

N CRP1 0.93 7.81 NA 0.66 0.17 0.44 0.030

N CRP1 0.93 7.81 NA 0.67 0.17 0.47 0.030

N CRP2 1.29 7.74 NA 1.72 0.40 1.72 0.140

N CRP2 1.28 7.74 NA 1.73 0.42 1.76 0.150

N CRP2 1.28 7.74 NA 1.78 0.39 1.74 0.150

N CRP3 2.16 6.95 NA 7.79 4.06 5.98 0.430

N CRP3 2.16 6.93 NA 7.67 4.15 6.03 0.420

N CRP3 2.15 6.93 NA 7.82 4.18 6.21 0.450

N CRP4 NA NA NA NA NA NA NA



P CRP1 0.97 7.75 49.79 0.67 0.21 0.51 0.040

P CRP1 0.99 7.72 50.61 0.69 0.20 0.51 0.040

P CRP1 0.97 7.78 50.27 0.63 0.20 0.51 0.040

P CRP2 1.34 7.60 54.08 1.77 0.41 1.62 0.150

P CRP2 1.32 7.60 54.11 1.80 0.41 1.63 0.150

P CRP2 1.34 7.60 54.14 1.79 0.41 1.62 0.150

P CRP3 2.26 7.26 10.19 8.10 3.72 5.05 0.350

P CRP3 2.22 7.36 10.20 8.38 3.73 5.09 0.350

P CRP3 2.24 7.23 10.16 8.25 3.72 5.04 0.360

P CRP4 1.18 5.93 22.24 2.02 0.39 1.19 0.060

P CRP4 1.18 5.92 22.12 2.02 0.39 1.19 0.060

P CRP4 1.16 5.98 22.29 2.02 0.40 1.18 0.060

Q CRP1 0.96 7.87 51.61 0.66 0.20 0.59 0.033

Q CRP1 0.97 7.86 51.52 0.68 0.21 0.60 0.032

Q CRP1 0.98 7.87 51.37 0.69 0.21 0.57 0.032

Q CRP2 1.39 7.63 54.25 1.82 0.46 1.94 0.166

Q CRP2 1.26 7.63 54.55 1.77 0.42 1.92 0.169


Lab Code



Q CRP2 1.27 7.62 54.55 1.78 0.43 1.88 0.174

Q CRP3 2.16 6.66 10.67 8.07 4.28 6.32 0.414

Q CRP3 2.18 6.65 10.66 8.30 4.23 6.55 0.424

Q CRP3 2.22 6.64 10.60 8.03 4.56 6.42 0.443

Q CRP4 1.07 5.81 23.49 1.87 0.46 1.39 0.058

Q CRP4 1.10 5.82 23.50 1.99 0.45 1.44 0.065

Q CRP4 1.08 5.82 23.53 1.95 0.44 1.48 0.059

R CRP1 0.93 7.81 55.27 0.68 0.19 0.57 0.035

R CRP1 0.91 7.80 53.57 0.67 0.20 0.57 0.036

R CRP1 1.08 7.80 47.14 0.68 0.19 0.61 0.035

R CRP2 0.93 7.59 54.30 1.77 0.40 1.97 0.164

R CRP2 1.09 7.60 54.61 1.76 0.36 1.93 0.157

R CRP2 1.08 7.60 54.48 1.77 0.39 1.90 0.164

R CRP3 1.65 6.65 6.23 7.80 3.75 5.90 0.417

R CRP3 1.85 6.63 6.24 7.70 3.81 6.00 0.418

R CRP3 1.81 6.64 6.35 7.75 3.67 5.65 0.419

R CRP4 0.94 5.85 22.08 1.89 0.41 1.35 0.064

R CRP4 0.92 5.83 21.88 1.86 0.39 1.26 0.066

R CRP4 0.94 5.83 20.99 1.85 0.41 1.32 0.062

S CRP1 1.04 7.94 51.60 0.66 0.21 0.50 0.031

S CRP1 1.14 7.95 55.24 0.69 0.22 0.48 0.033

S CRP1 1.15 7.95 51.55 0.67 0.20 0.45 0.030

S CRP2 1.42 7.68 55.01 1.83 0.47 1.64 0.151

S CRP2 1.42 7.67 53.54 1.89 0.46 1.66 0.147

S CRP2 1.42 7.66 53.88 2.01 0.46 1.71 0.150

S CRP3 2.30 6.78 10.15 7.95 3.93 5.16 0.346

S CRP3 2.34 6.80 9.83 8.48 4.17 5.24 0.349

S CRP3 2.29 6.80 9.80 8.17 4.18 5.36 0.376

S CRP4 1.23 5.96 23.09 1.85 0.45 1.08 0.052

S CRP4 1.13 5.88 22.99 1.85 0.40 1.13 0.054

S CRP4 1.28 5.90 23.01 1.98 0.49 1.14 0.054

T CRP1 0.82 7.78 52.09 NA NA NA NA

T CRP1 0.87 7.78 52.52 NA NA NA NA

T CRP1 0.89 7.80 51.59 NA NA NA NA

T CRP2 1.26 7.71 52.73 NA NA NA NA

T CRP2 1.25 7.68 52.39 NA NA NA NA

T CRP2 1.28 7.71 52.71 NA NA NA NA


Lab Code



T CRP3 2.08 6.79 10.04 NA NA NA NA

T CRP3 2.09 6.78 10.00 NA NA NA NA

T CRP3 2.09 6.79 10.16 NA NA NA NA

T CRP4 0.98 6.00 22.14 NA NA NA NA

T CRP4 0.92 5.95 22.48 NA NA NA NA

T CRP4 0.95 6.00 22.68 NA NA NA NA

U CRP1 0.94 7.88 49.38 0.58 0.22 0.46 0.035

U CRP1 0.95 7.90 53.44 0.60 0.25 0.51 0.032

U CRP1 0.89 7.86 52.59 0.65 0.27 0.53 0.055

U CRP2 1.22 7.58 54.09 1.68 0.51 1.65 0.155

U CRP2 1.23 7.55 54.12 1.67 0.52 1.66 0.174

U CRP2 1.23 7.55 54.10 1.70 0.52 1.65 0.155

U CRP3 2.06 6.67 10.12 7.86 4.48 5.36 0.432

U CRP3 2.06 6.71 10.12 7.35 4.22 5.11 0.428

U CRP3 2.06 6.71 10.20 7.63 4.32 5.27 0.438

U CRP4 1.04 5.86 22.85 1.57 0.45 0.90 0.055

U CRP4 1.02 5.86 22.56 1.68 0.43 0.94 0.050

U CRP4 1.04 5.85 22.48 1.57 0.39 0.87 0.049

1 Eliminated as an outlier prior to conducting Levene’s Test and Grubbs’ Test. Reps 1 and 3 were retained.

NA means not available. Lab N moistures were dropped because the lab analysed for water by the Karl Fischer

method, rather than oven volatiles by the prescribed method. Lab N analyses of CRP4 were dropped because the

samples were not homogenized prior to testing. Lab T did not conduct the TSNA analyses.

5. Data – Statistical Analysis

A statistical analysis was conducted in basic conformance with ISO 5725-2:1994 and ISO/TR

22971:2005. A summary of the results from outlier detection and the calculated results for

repeatability (r) and reproducibility (R) are given below in sections 5.1 and 5.2, respectively.

Even though ISO 5725-2:1994 does not suggest calculation of z-scores, z-scores are presented

in section 5.3 so that the participating laboratories would have an additional measure of their

performance compared to their peers. Section 5.4 describes the stability analysis conducted

using the data from the 2010-2015 studies. Raw data plots that include all replicates shown in

Table 3, without removal of outliers, are given in Appendix A.

5.1 Exclusion of Outliers

Procedures outlined in ISO 5725-2:1994 and ISO/TR 22971:2005 were generally used for the

exclusion of outliers. An adaptation of Levene’s Test was used for eliminating laboratories

with overly large repeatability standard deviations and Grubbs’ Test was used to eliminate

laboratories with outlying mean values.

ISO 5725(2) also recommends the use of Mandel’s h and k plots. Mandel’s h statistic is the

same as the statistic used in Grubbs’ Test. Similarly Mandel’s k statistic, associated with


within lab standard deviation, is statistically equivalent to the c-value calculated in Cochran’s

Test ( cnk labs ). However, the critical values associated with Mandel’s h and k statistics do

not make allowance for multiple testing and can therefore, give a false impression of

statistical significance. Thus, Mandel’s h and k statistics do not add fundamentally new

information and may lead to incorrect conclusions. For those reasons, we do not include

Mandel’s h and k plots.

The intent of ISO 5725-2:1994 is to eliminate outliers that exceed a 1% critical value. This

was accomplished by an adaptation of Levene’s Test. Levene’s Test is preferable to

Cochran’s Test, which is recommended in ISO 5725-2:1994, because of Cochran’s Test’s

extreme sensitivity to deviations from normality. Grubbs’ Test and an adaptation of Levene’s

Test were applied at the standard nominal 1% significance level to determine outliers and the

results are shown in Table 4. Levene’s Test is mentioned in ISO/TR 22971:2005 as an

alternative to Cochran’s Test. However, Levene’s Test does not directly apply without

adaptation. For more details, see the footnote below.6

Table 4: Outliers among 2015 Data

Analyte Product Levene’s Outliers Lab Grubbs’ Outliers Lab

Nicotine CRP2 − R

Nicotine CRP3 − R

Moisture CRP2 − T

Moisture CRP3 − R

pH CRP3 − P

NNN CRP1 − F

NNN CRP4 − F

The (-) symbol indicates an outlier was not detected.

5.2 Calculation of Repeatability (r) and Reproducibility (R)

After removal of outlying data based on numerical data consistency methods (Grubbs’ Test

and Levene’s Test), the final repeatability and reproducibility (r & R) results were calculated.

The r & R results are shown in Table 5 for the 2015 WG4 study. Since this test was

conducted to evaluate the stability of the CRPs, it should be understood that the r & R results

reflect both laboratory variability and product consistency. This is always true to some

degree, but is particularly true in studies such as this in which the product stability is being

evaluated and may play a non-negligible role in the results.

6 Levene’s Test is commonly used to determine if each of several subpopulations have the same variance. Since

it was designed to test for overall differences, not to determine if the largest variance is significantly greater than

the others, some adaptation is necessary to use the approach to eliminate laboratories whose within lab variation

is too large. Levene’s test was adapted to this purpose by Morton, who presented the approach utilized in this

report at the 2014 CORESTA Congress (Quebec, Canada, presentation ST28, October 14, 2014). Specifically,

the approach taken here is a two-step process with a lab being eliminated as an outlier if both steps are

statistically significant. First, Levene’s Test was run at a nominal -level of 0.02. Second a comparison of the

largest variance to the remaining variances is carried out at a one-sided nominal level of =0.01/number of labs.

Dividing by the number of labs is to account for multiple testing, since it is not known a priori which lab will

have the largest variance. Simulation studies were carried out by Morton and presented at the 2014 CORESTA

Congress and these results demonstrated that this process has an overall -level near 0.01 and is robust to

deviations from normality.


Table 5: Repeatability (r) and Reproducibility (R) for 2015 WG4 Study

Parameter Product N° of

Labs* Mean

Repeatability Reproducibility

r % r R % R

Nicotine (%) CRP1 16 0.98 0.121 12.3% 0.211 21.5%

Nicotine (%) CRP2 15 1.29 0.059 4.5% 0.163 12.6%

Nicotine (%) CRP3 15 2.17 0.050 2.3% 0.223 10.3%

Nicotine (%) CRP4 15 1.09 0.087 8.1% 0.248 22.9%

pH CRP1 16 7.85 0.057 NA 0.267 NA

pH CRP2 16 7.65 0.046 NA 0.250 NA

pH CRP3 15 6.75 0.047 NA 0.232 NA

pH CRP4 15 5.91 0.062 NA 0.200 NA

Moisture (%) CRP1 14 51.06 4.212 8.2% 5.397 10.6%

Moisture (%) CRP2 13 54.34 0.695 1.3% 0.763 1.4%

Moisture (%) CRP3 13 10.05 0.294 2.9% 1.284 12.8%

Moisture (%) CRP4 14 22.39 0.576 2.6% 1.758 7.9%

NNN (µg/g) CRP1 13 0.67 0.075 11.2% 0.112 16.8%

NNN (µg/g) CRP2 14 1.81 0.123 6.8% 0.294 16.2%

NNN (µg/g) CRP3 14 8.20 0.620 7.6% 1.299 15.8%

NNN (µg/g) CRP4 12 1.92 0.146 7.6% 0.381 19.8%

NNK (µg/g) CRP1 14 0.21 0.030 14.2% 0.056 27.1%

NNK (µg/g) CRP2 14 0.44 0.032 7.2% 0.100 22.5%

NNK (µg/g) CRP3 14 4.15 0.490 11.8% 1.059 25.5%

NNK (µg/g) CRP4 13 0.44 0.063 14.2% 0.107 24.4%

NAT (µg/g) CRP1 14 0.51 0.065 12.6% 0.153 29.7%

NAT (µg/g) CRP2 14 1.72 0.098 5.7% 0.393 22.9%

NAT (µg/g) CRP3 14 5.54 0.424 7.7% 1.517 27.4%

NAT (µg/g) CRP4 13 1.20 0.094 7.8% 0.435 36.3%

NAB (µg/g) CRP1 13 0.035 0.012 33.6% 0.015 42.5%

NAB (µg/g) CRP2 14 0.152 0.016 10.6% 0.032 20.9%

NAB (µg/g) CRP3 14 0.398 0.041 10.4% 0.104 26.2%

NAB (µg/g) CRP4 13 0.059 0.007 11.1% 0.019 31.5%

*This is the number of laboratory data sets included after removal of outliers.

NA = Since pH is not a proportional scale, it is not appropriate to calculate % r or % R.

5.3 Calculation of Z-Scores

Although calculation of z-scores is not suggested in ISO 5725-2:1994, z-scores were

calculated so that the participating laboratories could compare their results to those of their

peers. The z-scores were calculated using methods suggested in ISO 13528: 2005. The

calculations were based on Algorithm A given in Annex C of ISO 13528: 2005. It is expected

that most of the data should fall within the range of ±2, and that laboratories having values

with |z|>3 should be treated as an “action signal” to investigate laboratory performance. Since


this study is also intended to evaluate the suitability of the CRPs for continued use, large z-

values could also be an indication of inconsistencies (instability) in the CRPs. A final

summary table of z-scores is presented in Table 6 for the 2015 WG4 study.

Table 6: Z-Scores for 2015 WG4 Study

Product Lab Nicotine pH Moisture NAB NAT NNK NNN

CRP1 A -0.21 0.27 -1.97 0.48 0.98 0.71 1.13

CRP1 D 1.00 0.19 0.07 -0.65 -1.49 0.95 -0.28

CRP1 F -0.04 0.03 0.25 0.63 -1.11 0.65 3.94

CRP1 G -1.39 -0.61 0.83 0.78 0.91 0.13 0.95

CRP1 H 1.81 -0.77 -2.85 1.23 0.63 2.38 1.19

CRP1 I -0.15 1.71 -0.36 -1.10 0.03 -1.28 -1.01

CRP1 K 0.17 -2.61 2.11 -1.70 -3.08 -7.63 -4.23

CRP1 L 0.71 1.38 -0.04 NA 0.47 0.34 -0.13

CRP1 M 0.81 1.39 -5.71 -0.65 0.26 -0.56 -0.86

CRP1 N -0.90 -0.57 NA -0.95 -0.87 -3.60 -0.13

CRP1 P -0.10 -1.29 -1.65 1.30 0.05 -0.29 -0.05

CRP1 Q -0.21 0.11 -0.01 -0.42 1.37 0.10 0.26

CRP1 R -0.15 -0.65 0.62 0.25 1.29 -1.13 0.19

CRP1 S 2.05 1.07 1.66 -0.65 -0.51 -0.05 0.19

CRP1 T -1.98 -0.85 0.72 NA NA NA NA

CRP1 U -0.94 0.27 0.38 1.45 -0.13 3.38 -1.35

CRP2 A -0.37 -0.27 1.07 0.57 0.51 0.10 1.07

CRP2 D -1.36 0.31 -0.16 0.23 -1.36 1.19 0.63

CRP2 F 0.20 -0.62 0.55 -1.57 -1.92 -0.37 1.88

CRP2 G -0.99 0.12 0.30 0.75 -0.52 0.03 -0.46

CRP2 H 1.11 -0.52 -0.67 0.84 0.29 1.38 2.04

CRP2 I -0.62 1.53 1.05 -1.23 0.22 -0.27 -0.30

CRP2 K 1.44 -3.54 2.14 -3.49 -3.82 -4.41 -4.38

CRP2 L 0.13 0.99 0.44 0.11 1.23 0.32 0.01

CRP2 M 0.26 1.73 -0.95 -1.14 0.50 0.22 -1.31

CRP2 N 0.01 1.24 NA -0.41 0.21 -1.03 -0.54

CRP2 P 0.94 -0.81 -1.08 -0.10 -0.65 -0.83 -0.09

CRP2 Q 0.44 -0.42 0.43 1.69 1.52 0.02 -0.04

CRP2 R -4.67 -0.86 0.49 0.96 1.63 -1.61 -0.31

CRP2 S 2.57 0.21 -0.94 -0.16 -0.30 0.77 1.19

CRP2 T -0.37 0.65 -7.76 NA NA NA NA

CRP2 U -0.98 -1.40 -1.11 0.93 -0.45 2.50 -1.17

CRP3 A -0.36 0.00 1.28 0.99 0.76 0.90 0.90


Product Lab Nicotine pH Moisture NAB NAT NNK NNN

CRP3 D -1.22 0.31 -1.04 -0.12 -0.79 0.52 0.45

CRP3 F 0.39 -0.26 0.14 -0.64 -2.19 -0.01 0.03

CRP3 G -0.99 -0.46 0.14 1.37 0.66 0.61 1.61

CRP3 H 1.05 -0.57 0.90 0.55 -0.01 3.35 2.59

CRP3 I 0.04 0.78 -0.30 -1.51 -0.08 -0.83 0.09

CRP3 K 1.34 -0.82 3.68 -1.62 -2.65 -2.00 -3.84

CRP3 L 0.43 -0.46 -1.35 -0.29 0.80 -0.80 0.10

CRP3 M 0.31 2.28 -1.22 -0.27 0.37 -0.08 -0.73

CRP3 N -0.08 2.90 NA 1.04 1.10 0.16 -0.89

CRP3 P 0.90 8.28 0.22 -1.07 -0.82 -1.19 0.27

CRP3 Q 0.27 -1.55 1.06 0.87 1.78 0.91 0.00

CRP3 R -4.65 -1.70 -6.97 0.64 0.68 -1.13 -0.92

CRP3 S 1.73 0.68 -0.25 -0.97 -0.45 0.03 0.16

CRP3 T -0.91 0.57 0.00 NA NA NA NA

CRP3 U -1.25 -0.82 0.15 1.02 -0.47 0.86 -1.26

CRP4 A 0.12 -1.27 -0.58 2.00 1.20 0.75 1.25

CRP4 D -0.74 -0.09 -2.40 0.23 -0.97 1.13 0.45

CRP4 F -0.90 -0.98 -0.88 -0.60 -1.31 0.12 8.07

CRP4 G 0.57 0.33 -1.00 2.17 0.61 2.78 0.86

CRP4 H 0.38 -0.15 0.76 0.68 0.18 0.95 0.61

CRP4 I 0.24 0.74 0.73 -0.65 0.73 0.02 0.17

CRP4 K 1.63 -0.09 0.71 -3.20 -2.66 -5.40 -5.24

CRP4 L 0.41 0.80 -0.18 -0.27 0.53 -0.54 -0.36

CRP4 M -0.41 3.24 0.80 -0.88 -0.09 -0.83 -1.47

CRP4 N NA NA NA NA NA NA NA

CRP4 P 1.06 0.74 -0.41 0.34 0.01 -1.27 0.76

CRP4 Q -0.04 -1.51 2.08 0.45 1.65 0.51 -0.01

CRP4 R -1.89 -1.16 -1.51 1.00 0.83 -1.05 -0.61

CRP4 S 1.55 0.21 1.16 -0.77 -0.47 0.43 -0.39

CRP4 T -1.68 1.46 0.00 NA NA NA NA

CRP4 U -0.65 -0.80 0.39 -1.10 -1.86 -0.35 -2.98

NA means not applicable. Lab N moistures were not considered because they analysed for moisture instead of

oven volatiles. Lab N analyses of CRP4 were not considered because the samples were not homogenized prior

to analysis. Lab T did not conduct TSNA analyses and the Lab L NAB analyses for CRP1 were only reported as

<LOQ.

5.4 CRP Stability Assessment

The data were evaluated for stability by examining the results from the 2010-2015 WG4

studies for trends over time, for all analytes. Even though the stability trend analysis,


described below, was the primary evaluation of stability, the data were also graphically

evaluated for changes in between-lab variability over the study years. As mentioned in the

introduction, the 2010 study did not specify methods of analysis and the 2011 study was the

first study year initiated to assess stability of the CRPs where methods were specified.

However, after evaluating the data for all five years, it was determined that the 2010 data

were consistent with the data from later years and were included in the stability analysis.

The stability trend analysis included two factors: year of analysis and laboratory. To account

for potential year-to-year correlation in a laboratory’s results, the laboratory variable was

treated as a random factor.

There is also a potential need to account for multiplicity of testing. That is, if four statistical

tests are each carried out with a 5% chance to falsely declare a time trend, then there is just

under a 20% chance that at least one of them will give a false indication of a time trend even

in the ideal case of absolute stability. In this instance there are four reference products and

seven separate analytes, thus twenty-eight comparisons in all. The approach taken was to

correct for testing multiplicity across products for each analyte and not across all analytes.

The correction employed was to allow for four tests being conducted not all twenty-eight.

That meant that each test was judged statistically significant if the associated p-value was less

than 5% divided by 4, or 1.25%. Therefore, there is roughly a 5% chance that each analyte

will be declared to show a time trend on one or more of the CRPs by chance alone.

Graphs of the means for all years and all analytes are shown in Appendix B. These graphs

include the complete data set, without removal of outliers, to give the reader a perspective

both on the retained points and those that were excluded as outliers. Table 7 summarizes the

trend analysis of each analyte and each product. Since the testing was done on the mean

values, mean value outliers were removed prior to analysis, but no outliers were removed for

excessive intra-laboratory variation. The data points dropped as outliers are given in Table 8.

The only significant trend in the data was an increase in moisture for CRP3. It should be

noted that in most years, for most analytes, there is significant lab-to-lab variation. This

limits the ability to detect statistically significant trends. Also points that are eliminated as

analytical outliers, in some instances, could represent evidence of product changes.

There has been a tendency for CRP3 moisture to increase slightly and that pattern continued

this year.

Table 7: Summary of Stability Analysis

Product Variable Estimated Slope p-value1

CRP1 Nicotine 0.0067 9.4%

CRP2 Nicotine 0.0026 59.8%

CRP3 Nicotine -0.0018 78.0%

CRP4 Nicotine -0.0077 24.5%

CRP1 pH -0.0133 5.9%

CRP2 pH -0.0112 3.1%

CRP3 pH -0.0093 6.3%

CRP4 pH -0.0022 73.9%

CRP1 Moisture

-0.0257 75.6%

CRP2 Moisture -0.0235 45.2%

CRP3 Moisture 0.2908 <.0001


Product Variable Estimated Slope p-value1

CRP4 Moisture 0.1646 15.8%

CRP1 NNN 0.0023 58.9%

CRP2 NNN -0.0019 85.5%

CRP3 NNN -0.0070 89.7%

CRP4 NNN 0.0010 94.2%

CRP1 NNK

-0.0014 24.9%

CRP2 NNK -0.0024 37.0%

CRP3 NNK

0.0128 72.5%

CRP4 NNK 0.0024 54.3%

CRP1 NAT -0.0020 59.0%

CRP2 NAT -0.0193 9.9%

CRP3 NAT -0.0357 39.3%

CRP4 NAT 0.0071 48.8%

CRP1 NAB

0.0008 15.9%

CRP2 NAB 0.0001 89.2%

CRP3 NAB -0.0007 82.9%

CRP4 NAB

0.0003 62.6%

1. P-values are declared statistically significant if p < 1.25%. 1.25% is derived by dividing the nominal 5%

level by 4 to account for testing multiplicity per analyte.

Table 8: Summary of Points Dropped in Stability Analysis Calculations

Product Variable Laboratory Year

All CRPs All TSNAs K All years

All CRPs Moisture K 2015

CRP1 Nicotine F 2012 & 2014

CRP2 Nicotine S 2014

CRP2 Nicotine R ;2015

CRP3 Nicotine C 2010

CRP3 Nicotine S & R 2014

CRP3 Nicotine R 2015

CRP4 Nicotine F 2014

CRP1 pH P 2014

CRP2 pH M 2010

CRP2 pH K 2015

CRP3 pH P 2015

CRP1 Moisture M 2014 & 2015

CRP1 Moisture E 2010

CRP1 Moisture N 2013

CRP2 Moisture P 2014

CRP3 Moisture R 2015

CRP4 Moisture F 2012


Product Variable Laboratory Year

CRP1 NNN F 2012 & 2015

CRP2 NNN P 2013

CRP4 NNN F 2015

CRP1 NNK P 2013

CRP4 NNK R 2014

CRP2 NAT P 2013

CRP3 NAT N 2013

CRP4 NAT R 2014

CRP2 NAB O 2014

6. Data Interpretations

The trend of increased average moisture content noted at times in CRP3 was again

statistically significant. The increase from the time of manufacture is approximately 2%.

CRP3 has the lowest moisture content so this product would be most likely to absorb water

over time as compared to the other CRPs. As mentioned earlier, results are presented on an

as-is basis; however, this negligible shift in moisture for CRP3 would not be evident in the

other chemical analyses considering the negligible change in moisture in comparison with

lab-to-lab method variability.

In examining the mean data plots for all years (Appendix B), a number of potential outliers

are evident; therefore, the stability analyses were conducted both with and without the outliers

included, but only shown in the table above with the outliers excluded. There were some

indications of changes in between-lab variability over the study years. Two notable examples

of increased variability between labs are pH for CRP3 and moisture for CRP1. Both have

increased in variability across the reporting period (2010 Lab E was dropped as an outlier for

CRP1).

As reported last year, the inter-laboratory variability for CRP4 moisture has been very high in

some years, such as 2012, but was improved in 2014 and 2015. In 2014, one laboratory

reported that a package of CRP4 that was recently received from NCSU appeared to be

uncharacteristically dry. The laboratory confirmed that the moisture content was

approximately half of what it should have been and retested another package (all analyses).

Although the original package of CRP4 was clearly defective a smaller loss of moisture

would not have been readily apparent and any differences in the moisture content determined

would have likely been attributed to method variability.

Due to the product type, CRP4 is the least homogenous of the four CRPs. This suggests that

sample preparation (particle size reduction) and how tobacco aliquots are removed for testing

are important components of the variability seen within labs for CRP4. Labs reported a range

of sample preparation techniques ranging from cryogrinding an entire package to simply

removing individual aliquots for testing. Variability is reduced by generating homogeneous

composite sample and removing aliquots for the replicates from that homogeneous composite.

Creation of a homogeneous composite can be accomplished by either cryogrinding an entire

package or using a razor blade to cut and mix the entire contents of a package. We

recommend either of these approaches in the future as simply removing aliquots from an

unhomogenised pouch of CRP4 for analysis has proven not to be suitable.


Overall, both nicotine and TSNAs are shown to be stable over the six study years, for all

CRPs. This is not unexpected as both nicotine and TSNAs should be stable in these products

when the products are maintained at -20 °C.

Based on the results taken in their entirety, the conclusion of this study is that all four CRPs

are suitable for continued use as reference products. Lack of any clear and significant trends

in the stability results indicate that storage at -20 °C is an appropriate storage condition for the

four reference products. Furthermore, since the changes in moisture, if any, are small relative

to the inherent method reproducibilities, it is unlikely that they will have an easily detectable

effect on other results.

7. Recommendations

The results of this study demonstrate that the CRPs ae suitable for continued use as reference

products; however, manufacture of new CRPs should be considered due to the dwindling

supply of the 2009 products.


APPENDIX A: Raw Data Plots for 2015 WG4 Study


APPENDIX B: CRP Stability Assessment, Mean Data Plots

Smokeless Tobacco Sub-Group - CORESTA · Smokeless Tobacco Sub-Group ... December 2015 Author and...

Documents

Transcript of Smokeless Tobacco Sub-Group - CORESTA · Smokeless Tobacco Sub-Group ... December 2015 Author and...