T:1119
INDEPENDENT SCIENTIFIC COMMITTEE ON SMOKING Si HEALTH
CONEOLIRSIND NMI N MOMWIffil 04 IUMMY QE MUM& MM Luz=
1. Yields of "tar" (particulate matter % water & nicotine from) and
nicotine since 1972, and of carbon monoxide (CO) since 1977, have been
regularly determined by the Laboratory of the Government Chemist , on
behalf of the DHSS, for up to 150 major selling brands of cigarettes in
the United Kingdom. At the request of the ISCSH the LGC has conducted a
special survey o4 additional noxa, using 75 key brands representative of
all types and designs together accounting for just over 90% of the
current UK market.
2. LGC has determined the mean yields of four ci Iiatoxic substances,viz.
hydrogen cyanid*(HCN), acroloin, formaldehyde and total aldebydes (RCHO)
expressed as Pacetaldshydo'3 and also determined nitrogen monox
These analytes may be grouped according as to whether the smoke analysis
was performed an the QZCJ1&UJ&:t1t (condensed) phase (tar & nicotine)
whgle MOM ( RCHO, HCN & formaldehyde) or y122Ur phase only CO
,--4crelain and nitric oxide ) .
'1@. For each analyte existing METHODS, described by the TAC or published
elsewhere, were considered and appropriate, sometimes extensive,
adjustments mad* to facilitate semi-automated analysis. The individual
methods are summarised in Annex 1.
The Committee then sought studies on the components of tar fractions
a survey is in hand with monchydric and dihydric phenol and subsequent
attention will be given to polycyclic aromatic hydrocarbons and to
non-volatile nitrosamineig.
4. T&bJ1_J 0 gives the mean yields for individual brands for 3 regular
and i addi onal noxa. The 75 brands have been ranked in order of
increasinqltarl yield. Only one hY22C_-X2[1tJJjLt2SJ brand (Players KS
Ultra-Mild) was used and Is the source of Jq11ltL yield of all eight
AM41 Yt *S. A number of brands, as indicated by boxes in Table I and
footnotes in Table I IA, variously contribute the max yield for each
analyte.The Tarle Nh&Cl of each brand in 1982 is included for purpose
of comparison
summarises for each analyto the lVA11t and qE:SAtRAJ brand
iolds. and the 1AJS2=gSLghtSjJ 22M yields, per cigarette with an
indication of the typical 95% %2n:EJ#2QTLS JJMJt2 about the moans and the
number of determinations that contributed to each brand mean.
7. TAbJZ_JJR shows the mean yields for subsets of the 75 brands
partitioned into six discrete but partially overlapping categories, vi=.
imported brands (11), ai r-curod tobacco (4), plain cigarettes (6) and
three filter groups i Mini&Rogular size (14), King Si=& (51) and all.
Ventilated (22).
c::)
'IO
00
4-4
CO
BATCo document for Province of British Columbia 19 April 1999

'reemasaaaaa Zmat
C a Socimical-wim AMMOD
V^" 3MVIUL.00 OGW OZENION441two,
. a wwwwoeso W-ftwomms m@ Om OWN sense Wm so arms 0
Csa
CD Asset to am
Nomads Ovem @- Nov "6. to is to to is to %IF as
gram nam -
J Flavor X.16 v. "040 1.6 *.&6 464 4611 410 426 %10 CIO 0.0
Kit, cut 9.014 ffthe 6.3 4649 a.& 0.66 311 so 26 96 0.0
saw SmatatsIrre IL I.: 0.76 6. 0. 74 " 66 42 37 O..
COMM4,0419 K. 0. li.= R.: 0.01 46 OF -.v 9.4
J.pt~ Van~@ 6.7 *.M 12.3 L. 29 19 136 43 463,
lowalls U.S. $hid ::93 10,3 "" " " '* *-d
U11, tot 40 9., 1.23 .0 139 so .1 0.1
Ofts, 9.9 O." 9.4 9.00 78 011 91 43 1.&
101wUlliew 9.0 0.06 0.6 1.44 72 6 %7 44 39 0.7
tsb@"V M41.9 9. Rate 9.2 ..= 80.3 S." 47 INS 73 6& LA
sill Cot King St&* 9.2 0.92 W. &.06 73 13* 40 311 2.3
2wassew Extra Mae 9.3 0.76 9.9 J.as a In " so 0.16
be -ftwow K S. 9.4 O." 9.9 9.30 39 9113 M 49 t.3
I afte a 211109M 1%Ld 9.9 0.87 LS.& 9.30, M 930 66 0.6
4.16 Cot
.PIA"W S. C. mat ;:-' 9:0"I "J 11:330 V I.- 7". to
7 a
=3 Sems&i 41110 S, sli. 1. 64 tv ST 42 0.3
at."watt "" ..23 ..4,41 70 on n 63 0.3
awmm trAwl"asse"GA 12.3 11.011 17.2 t.31 434 143 32 39 ..O
::%are% *Ittqw IM.3 S." 94.3 6.71$ 309 236 " as 0.0
SW swessOL -6.0. ou :.$ S.82 12. 2 1.31, 67 too 64 so 0.0
rarstaft %M" lits 12.0 9.23 .10.3 9.44 69 an " 0.4
POLO 3 an 0.
ft.ewo Its. S:'. 1, 31, 4S 03 '10
2 364 30 V.:
@saw stw".L.9- 13.3 1.22 11.2 1.66 63 in 69 76 Z.:
"Or 13,: 1:30 111.9 1:211 VEI 00 " 42 ::,
In lism@ MASS 3. , to .3.0 LIf 64 976 at 49
-10 M: 9: 19 L3.1 6.34 29 133 66 63 ..9
both a I uwwrv 9 13. a 29 12.7 I.MS 414 160 is 60 9.7
OMAN:&& U.N. 14.0 1.20 14.4, S." 72 L92 ON, 69 3.0
Ca"V910 lai6i K.$
MMISIMM 0141 Ps
':7 99
14-a"LIr 11 U_. Is.; 1.320 I'S t.62 40 jn63 73 72 I.:
J." 6@ S., &.. 16.2 9.12 " M " 44 G.
coma -11= 91. LI.%? 14.7 9.60 992 =6 as
Past WI I top ,6, :,U la,', 1:114 3" 21. W .1 C:
, a . V st, 67
aw~qmwv 16 16 Us. 80 M 219
%Wft&% tatem"40-41 &S.2 1.30, is. a 1.74, 72 972 76 E.
L. am 0 "I" 1"aw 1163 6:4L V111 3.00 79 SM M ': Is
M. Wm .,. 18.4 3 22 14 1.61 64 2" ON 35
IAD= W161 U.S. 15.11 1.311 as. 6 9.87 To gas 3 94 Vt 3.3
9"40" Fist alt 14.
".4ar '21 ':31 a I ':" as a S a I es 1% I
ai-nover 1. 7 1 11.3 11.90 67 " 11,
0 spar's "S.; V.S. ., ... 92., 2" 76 an ." go 3.3
.014 so woo. la.? I. 34 So,2 9.74 2" 222 73 64
sewhod"s "v&&* A. 60 L 0 3 . 96
Is.: ..Za 6"A Iss
go"Reed" 6"6 .9. 2 96 4. E.
4"61 . S. 96.0 I.410 12.6 3. 69 72 906 410 so 2.2
I 400 - short" 16.0 1:33 M: 1:6* W 973, a 36 40.0
,a,1sq,e la.I 1 32 111. & " 211 ass a 44 1:3
lot.. ?%I,. 16.2 1.40 96.7 I.93 k" 2" 0 n I
lo"W.ben Filter 96.3 6. 1: 96.2 1. Pe 4 I =0 9 I I.. ..:
I ^a LAnew LwboM 10.4 1.112 16.3 6.79 96 sea 06 37 I..
1LOV40,11 40.6 Pitter lb," 1:311 111,1 1:91, 74 203 V.9 964 ::
;.W. .6-11 S , ,3-1, ,I" `17 11., Pe d,
gn Potter P.Aft" .6. I..= . 6-11 ..W 77 22:11 . 116 2.2
olower's M.M Filter 96.7 S. as 94.9 IF 21016 9 & 102 0.2
*%~Is said U00, 16", I:r 1. M 246 I" E
'IMot" we. -..*. .6.0 . . 31 "', " as 60 ::,
-Nowde-e, Posts, 96.0 1.37 13.1 S." 62 161 NO, 93 0.0
MethasOM K. 2. 16.9 9.62 96.4 B." 97 2" Vs 117 3.3
6oboom ftle
FE a. It "::: a: T,
- a 7 ad 7a 'J" 74
stover me. Is. C. 40of 92.3L.36 &a.& 1.87 144 113 11:1
1.014vop 1110111CL&I 114. 17.4I.. 49.4 S." To 196 S"
"Glar OPtre ILLaQ, 17.3 t.,.v 16.7 2. 38 04 ass 893 1031 -.0
4"1"0 Club k. S. 17.6 t. PI to. I I.90 sag 244 &OR 1.3
uwwm K.& 97.2 I.VV " =5 9.10
Part, Onve4fil 7:11
I :7 ::" 61 NOL 63 IN 1:.
"Memang, 01 a 46 9 83 as M 00 Ed I 5
*",or so two#01 LS. t&.at 19.1 t.79 VP 142 72 62 0.5
ftes 0166.0% 18.3 &.= IS.3 1.77 33 in 60 To 9.0
9 ww " some 44 U.s. 917 2fAZ " 74 2
02 2.91 360 67 a
52*0 &O., 22:*'
caestan AA ILIP$ 24.6 2.40 93.4 1.93 " 2" 73 SW
Mg?wqv= pion 14. X4. If 93.6 2.66 SW its 76 66
ftL-90@ SMLJM-Wlg Is. 61.32 is. a 9.20 a 994, 89 79
LOU TM 1014.910 win -, E ad .40 Ps
L-" "A *%= VTD 11. 3 Ed
"B" la* SALIM W" ty., S.1 as 6' Ws a
TWon. "AN&Cr 94W4 W O" 78 SPAPOM 10 "C TOOLE M6Z
.910oMf SM09 ; o%A. 600 TOM MONM IN "C Th" L*. 619
-rafts "ft %.L j.-O TAMI MMMM In TM TO" 36.42
@aftv pooft OF ALL Hts" Mo "Om IN "O ?11" 0 St.=
TAN "$a co MCIVI 00 "M Seems Mcps
TOR " *v co
polcavsma -0 :!! -t.00
to --J
R"
-*. 26 -0.111 -2.00 -Cb-
-0.37 -0. M -O." -O-Z4 -9-04
QZ .6 .4 .0 do -*.*@ -4.6d, -1.00 Q1
-0:35 -0:0: -0.26 -0.64 -1.00 co
#a" is owed 0 so 3 hOwl 118
BATCo document for Province of British Columbia 19 April 1999

8 sets out the Rank Order (without sale% weighting) in terms
o; increasing arithmetic mean.
For ItKI, K_N and T9tjj ALIt2tXdga (RCHO) the order is roughly the same,
le Plain Air-cured MIM&Regularv KS 9 and Ventilated .
UIS21ing is similar &part from the expected such lower yields for
air-curod brand%.
For raACUM 2902iiLdit (CO) & tfUr-22ta moa2xLs1n(NO) the order is Air-cured,
MiniLRogular, KS , Plain, and Ventilated.
The figures for E2C_%Ajjj2hX!ft and err_212io do not seem to fit in either
with tar, MCN, nicotine & RCHO or with COMO . The decreasing rank order
is Mini&Regular, Plain, KS Filter, Air-Cured and Ventilated.
The position of IMUCtRA kCAM02 ( North American blends and French dark
air-curod tobacco ) is low -for formaldehyde and nicotine,
intermediate for acrolein and total aldehydes, and high for tar, MCN, CO
and NO. It is particularly high for NO, being over 190 , whereas all
(except one) UK brand is below 120 micrograms NO per cigarette.
TASUS 11A 119 and 11C
Mean yields of noxa per cigarette (milligrams)
MU I I A s IUMM GE ZZ BRAN22
- Pvticolate - - %dip Soots V"W
OW licotime NM Total WM NM knisin M
Lowt mm (brad a) 1.0 1.16 1.10 0.17 (0.81 (8.11 M 6.91
$dn-voijktW we W 1.32 4.114 1.0 1.075 IMR W LM
Hilhoot walbrad 6) 24JIM 2.46(k) 4.310(s) 2.nidi 1.1121j) 0.1144d ITMO 1.434M
Aq.M CoMisit (41 31 n U n n n U it
ft.dotov.por brand 31 38 A. L 6 36 A
a0layws KS LVA I buCaWn FS Maial ; c*AmM No.1 KS dsftjw utra sit I
o4addsn M&W f4itan lotwntiondo I rtlayws ML14 Filtor
do
%-0
CO
4
CO
BATCo document for Province of British Columbia 19 April 1999

TAMA III
Arithmetic means, and salou-stelghtwd means, for PARTITIONED DATA
-"icslato- -wolewl- @@Vapw@@ Nutot
%.brands OW, licotin 10 Total Co Cm Acrolvis CB No dwe
Ventilated F. e22 16.619.3 LUIS." .11U.1221.2411.19.1471.141 MI1.W 1.9111.1 .0611.663 14.1
0 FlItw 951 ILUILS 1.1711.32 .46U.475 .6741.98V IL21IL4 M.w TU
NiniNnolar F.*14 14.71ILl 1.1911.26 .011W.192 IM113.9 .1691.165 MI
Aircwtd tdwcow4 IL4/- CM- ZW- 1.01- .%$/- .4W- U.71- .4191-
Plain cilwattan4 19.11ILI LWI.47 .1731.0 I.IVI.7V .101.110 .%V.US IIXII.4 M/.M 4.9
M.M1 JW.085 16.111M .3091.215 LI
aportod broadoull ILMILI 1.1411.31 .2311.231 l.W1.7? V
IBILE
P&Mmlato - Ubdo godo Vapow
ft.bruds OTar" Iticatin KI Total ACE N= Acrolsis 0 No skwv
hatilated Filtv all 1 2 I I I I I 1 14.1
King Ilin Filtw aS1 2 31, 2 2 3 3 I 3 73.k
MaMplar Filtwxl@ 3 4 3 I 5 5 4 4 14.1
Air-Cw2d 4. 4 I 3 3 2 2 5 5 -
plain 5 5 4 3 4 4 2 2 4.9
lowt 415 213 85 314 112 U4 4/3 (5 2.3
Co
BATCo document for Province of BritiSh Columbia 19 April 1999

W
0
0
CL
sr
L LLI
P.r
L p v p p low,
St
p
cr
LL w ft ir
F ? -, le,
t
p d
co
lp
.04 0 We
co
co
co
HgkZ8631
i A-K

ANNEX I
6MMOX OE MIMI MU=
ALDEHYDES AND HCN
The analytes are trapped by passing the %make stream through a tube
containing silica gel. The material absorbed onto the silica g*1 is
recovered by shaking with water and both IgI&L &jd2tXdS2 and HQN are
determined on the same extract.
The W:N determination is based an the reaction of cyanide with
chicramine-T to form cyanogen chloride which in turn reacts with
pyridine and pyrazolone to form a colored product. The sample must be
made alkaline before the reaction to break down cyanide complexes to
simple cyanide.
The total al dehyde determination Is based on the formation of
2,4-dinitrophanylhydrazone complexes. Before the addition of the reagent
the sample is mixed with an aqueous solution of methanol and sodium
hydroxide to liberate al dehydes from complexes such as cyanchydrins.
Determination of the total aldthyde content Is mad* by reference to, and
is expressed as,scataldshyde. Only very rough weighting factors are
-ailable for subtracting separately determined quantities of HCHO and
icro lei if ane, wishes to derive a crude estimate of acetald*hyde (and
higher homoloquos).
ACROLIEIN
Acrolein is known to be present only In the vapour phase of the smak
and that it can react quickly with substance% in the the smoke
ph&%&. It is therefore determined on the vapour phase only . Filter
pads are changed after smoking each cigarette to prevent build up of
particulate matter which could degrade the acral ei as it passes
through the pad. eSE2121a is trapped by passing the smoke stream
through a bubbler containing ethanol cooled to 'dry ice' temperature.
It is determined by forming an addition compound with 1-butanothiol
which is measured quantitatively by gas chromatography.
F13RMALDEHYDE
carmaldshyde is determined calorimetrically using its reaction with
p-rosaniline in the presence a+ sulphur dioxide. The trapping technique
11--@ s Identical to that described for RCHO & HCN and for survey purposes
.he three analytes are conveniently determined an the same sample.
NITROGEN MONOXIDE
The vapour phase o4 the fresh smoke -From each puff is analysed by
measurement a+ the chemiluminescence produced by the roacti an of
nitrogen monoxide with ozone. Immediate analysis of the fresh smoke
from each puff is necosary because of the rapid reaction of nitrogen
monoyide with other vapour phase constituents.
Cigarettes are smoked In a Bor-gwal dt RM"-O automatic rotary smoking
machine and the chemiluminescence measured in a Cheml ab nitric oxide
an a I yq or.
co
--@j
-Ch.
4 Ln
.co
BATCo document for Province of British Columbia 19 April 1999

ANNEX I I
QgaaffLengb WeLMI
QgCCRjA:tLgQ between pairs of variables can be considered in a number
04 ways.
Correlation coefficients can be derived from calculation of the
, 1 east squares fits" for any two noxa . Over all the 75 brands 9 is
wall correlated with 0"g&Lat and ALfttXA2jv and GQ with HGN ; the
latter pair might be implicated In vascular diseases. ftEglain and
+crmaLd2b,X92 each correlate moderately well only with other aldshydes.
For Nitric oxide (NO) the marked difference between imported and home
produced brands leads to very low correlation coefficients between NO
and the other anal yt*s .
Correlation coefficients, both for the full 75 brand set and for the
data RjLr_tJjL9pSd as in table 119 (summarized in Para 7) , are set out in
a revised TAU& 111.
Correlation can also be Illustrated by scatter-plots, taking pairs
of parameters as rectangular co-ordinates. Ton ELa!Ar_vs_Z0-_N were
previously supplied to the Committee * these Indicated the typo of
correlation diagram that can be presented. A complete set is being
-spared for all combination%.
With Par_&LUga" _"& care should be taken in interpreting the very
small groups of partitioned data (air-cured and plain brands); however,
removal orf these data improves the correlation of the remaining data
base and certain further conclusions can be drawn
(I) 'Tar', when broken down into groups, not only correlates well with
n1r.21 "n and GO as recognised for all brands , but is now fairly good
for all anal ytes except Ng from filter brands.The correlation is still
poor with ALS1Rbyd2j for plain and air-cured brands.
(II) For Nicotine all comparisons with air-cured brands are distorted by
their inherently lower nicotine content. Nicotine generally correlates
well with tAC and is fairly good with QQ and UCL4 when data are
partitioned; this is also true for A"Mbydgj with all filter groups but
not for plain and air-cured brands.
(iii) Formaldehyde correlates well with tic and other
-. S12hy, (la reverse,
@U,2 for filter brands but consistently negatively
slope) with all analytes for plain brands.
1'0@'(iv) Acroloin correlates m dwAtRLY wall with all analytes except Ng
for filter cigarettes and (not unnaturally) gy
its wall with other
algshxd2s. It is poor for plain brands and (except with HCN and RGHQ)
for air-cured.
Cv) Total aldohydes correlate quite well with all analytes except
for filter groups (particularly the Ventilated brands) and fairly well
for air-cured brands (except with Ng and nic2&LM) but poorly with
ell anal ytoo for plain brands.
(vi) HCN also correlates moderately well with all anal ytou (even NC)
except for ajC=SUC2Sj tCandg) - There is a-captional I poor correlation
with all the gL!fth_y. 2j for plain brands.
(Vii) CO correlates quite well with ALL analytes except UQ for *ilter
brands but only with tac, Di%2=2 and H( for plain and air-cured cx:)
brands
BATCo document for Province of British Columbia 19 April 1999

(vtii) NO , with knocud brands segregated gives somewhat improved
correlations and is best with ventilated brands. For imports only QQ
shows moderately good correlation;. the aircured imports are strongly
negatively correlated.
Health & Forensic Services Division,
Laboratory of the Governm*nt Chemist
la May 1985
co
(-n
BATCo document for Province of British Columbia 19 April 1999