EPA 560/6-81-005
                           EPIDEMIOLOGY STUDIES
                      MAGNETIC LUNG MEASUREMENTS IN
                    RELATION TO OCCUPATIONAL EXPOSURE
                 TN ASBESTOS MINERS AND MILLERS OF QUEBEC
                               JANUARY 1981
                               FINAL REPORT
                   U.S. ENVIRONMENTAL PROTECTION AGENCY
                         OFFICE OF PESTICIDES AND
                             TOXIC SUBSTANCES
                             WASHINGTON, D.C.

-------
                                     EPA 560/6-81-005
                                     January 1981
     MAGNETIC LUNG MEASUREMENTS IN
   RELATION TO OCCUPATIONAL EXPOSURE
IN ASBESTOS MINERS AND MILLERS OF QUEBEC
                   by
 David Cohen^,
Thomas Crowtherl

Maret Decklake -
    Massachusetts Institute of Technology
    Cambridge, Massachusetts
    M
            Project Officer
             Jane E. Keller

Health and Environmental Review Division
Office of Pesticides and Toxic Substances
          Washington, D.C.  20460
       U.S. ENVIRONMENTAL
               WASHINGTON,
PROTECTION AGENCY
 D.C.  20*60

-------
                        DISCLAIMER
This project has been funded with federal funds from the
Environmental Protection Agency under contract No. 68-01-3859,
The content of this publication does not necessarily reflect
the views or policies of the U.S. Environmental Protection
Agency, nor does mention of trade names, commercial products,
or organizations imply endorsement by the U.S. Government.

-------
                        ACKNOWLEDGEMENTS


      The authors wish  to  thank  the  Quebec asbestos Mining industry
and their employees  for their  cooperation, as well as the Institute
of Occupational and  Environmental  Health of the Quebec Mining
Association.

      The authors are grateful to  Mr.  Ed Givler for his help during
the early phases of  this project,  to Mrs.  Eleanor Pyle for her
encouragement and editorial  help,  and  to Mrs. Donna Neuberg and the
MIT Statistics Center for  their  statistical evaluation of the data.
M. IK Don1clHlf.n in n  n*v.^- Tnvrnl-i rj-iJ-nr  nf tihP M^rl LcaJ— Re-SgHTTSK
     i 1 of Canada*
      Last we would  also  like  to thank Emanuel Landau,  Ph.D.,  Lenora
Barnes, M.P.H.,  and  the entire staff at the American Public Health
Association; Joseph  Seifter, M.D.,  Charles Poole,  M.P.H.,  of the U.S.
Environmental Protection  Agency for their cooperation and  support.

      The experimental measures and initial data analysis  of this
project were supported by NIH  grant #HL17543 and NSF grant #APR-761019,

-------
 ABSTRACT
        Fe 0. particles  (magnetic)  are  often  attached to  asbestos  fibers
 (non-magnetic)  in  the primary  asbestos industries,-  therefore,  a measurement
 of Fe O could  help  determine  the  amount  of  asbestos in  the  lungs of
 workers in these industries.   As a first  assessment of this  method of
 determining retained dust,  magnetic measurements were made of  the amount
 of Fe,0. in the lungs of  115 miners and millers of  chrysotile  asbestos.
 The  performance of these  measurements  at  an  industrial site  was found to
 be feasible and practical.  A  relatively  large amount of Fe  O  was seen
 in the  lungs of those with  welding experience, which masked  the Fe O
 contributed by  asbestos,  therefore this group was considered separately.
 For  the remainder  (non-welders), the amount  of Fe O was plotted against
 a total dust exposure index which  was  available for each individual.  The
 correlation between  these quantities was  not high,  but was statistically
 significant at  the 0.01 level.  For the non-smokers within that group, the
 correlation was higher and  the amount  of  Fe  0  was  relatively  greater.
 These results suggest that  the magnetic measurement of a chrysotile miner
 and miller reflects, at least  to some  extent, the amount of  asbestos in
his lung; the scatter could be due  to  individual differences in deposition
and clearance,  to  which this measurement  should be  sensitive.  These
results  are also consistent with the possibility that less dust is deposited
or retained in  smokers than in non-smokers.

-------
INTRODUCTION
       During the past decade, a method has been developed to measure
ferrimagnetic dust in the human lung  (Cohen, 1973, 1975).  In this method,
the lungs of the subject are first magnetized by an external magnetic
field. Then, after the external field is removed, the remanent field
produced by the magnetized particles is measured over the subject's torso;
this yields the amount of this dust in the lungs. Because chrysotile
asbestos often occurs with attached ferrimagnetic Fe O  particles, and
preliminary measurements of several chrysotile miners and millers had
revealed a measurable amount of Fe 0  in their lungs, the question arose:
Can the magnetic method be useful in determining the amount of asbestos
in the lungs of miners and millers?  The study reported here begins to
answer that question.
       We concentrated on mining and milling because the Fe O  content of
asbestos is high at this primary stage (Gibbs, 1971).  In particular, we
concentrated on the miners and millers of chrysotile asbestos in
Quebec. For this well-studied population, previously the amount of asbestos
in a worker's lung had been inferred from his total dust index, defined
as (concentration of airborne dust) x (period of exposure), summed over
his various jobs (Gibbs and Lachance,  1972).  Epidemiological studies
have shown a relationship between respiratory abnormality and this index
(Becklake et al., 1972; McDonald et al. ,  1972;  Rossiter et al. , 1972).
However, in some of these studies the correlations, while statistically
significant, have been low.  This may be due to the indirect nature of
this index,  which does not take into account individual variations
in dust deposition and clearance in the lung.  The magnetic method, while
it has its own drawback of being completely dependent on the ratio of
Fe 0  to asbestos, may nevertheless be more direct.  Our main objective
was to examine the relationship between the amount of Fe^ in the lungs
and the total dust index.  If the amount of Fe 0  was indeed related to
the amount of asbestos in the lung, then we would expect at least some
correlation with this dust index.
       The remanent field produced by the lung particles is very weak

-------
 Cv> 10   gauss),  in comparison to the background earth's field in which it
 is measured (^ 1 gauss).   This field must therefore be measured with care
 and attention.   While it can readily be measured in a well-equipped
 laboratory, we  previously had little experience in measuring this field
 at a distant industrial  site,  such as at the mining town where this
 study was performed.   Hence our second objective was to see if such on-site
 measurements are  practical,  not only for possible  use  in the primary
 asbestos  industries,  but for other "magnetic"  industries as  well.
        The measurements  were made  in 1974,  and preliminary results had been
 reported  (Cohen,  1978).   Following this  study,  other magnetic studies
 of occupational  groups were performed.   These include  two studies of
 welders (Kalliomaki  et al.,  1978;  Freedman  et al.,  1979)  and a study of
 coal miners (Freedman et al.,  1980).   In addition,  a magnetic study of
 stone workers is  being completed (M.  Kotani of Tokyo Denki University,
 personal  communication).   However,  in none  of these studies  was the
 relationship examined between  magnetic reading and dust exposure,  nor  was
 the  practical aspect  evaluated.  In  addition  to  these  studies,  magnetic
 studies of only  small occupational groups have been described (Cohen,  1978).
 MATERIALS AND METHODS

 The Group Studied
       The target group had already been selected for certain  lung  function
 measurements (Peress et al., 1977), and were without radiologic  abnormality
 in that their most recent chest film was read as 0/0.  This group contained
both smokers and non-smokers, an age range from 26-50, and a wide range  of
total dust indices.  These indices applied only to exposures up  to  1967;
the period of 1967-74 was not included, but the omission is not  likely to
be serious because the dust level had become reduced by the late 1960's.
Unfortunately,  the general nature of the grouping changed after measurements

-------
began, when it was seen that workers with welding exposure had a relatively
large amount of Fe O  in their lungs, which could mask the Fe O  from
asbestos.  We therefore divided the original group of 115 into those with
and without welding exposure. The relation between magnetic readings and
total exposure index was investigated only for the non-welding group, now
reduced to 51.  For the remainder, our efforts were salvaged by estimating
their exposure to welding dust and examining the relationship between the
amount of Fe O  in their lungs and this exposure; if the Fe O  amount was
a reflection of occupational exposure, we would again expect to see a
correlation between the two.
       The divisions within the group are given in Table I.   Those called
 smokers  included  11 who were  ex-smokers  at the time of this study; they were
 included because  their smoking had taken place when most of their dust was
 inhaled,  some  years earlier,  hence it would have affected their response to
 dust.  Welders were divided into  those with only a small  (<0.1 year) and
 those with  a  greater  ( >0.1 year) exposure to welding dust, to allow various
 correlations.  However, it  is seen that some correlations would be limited
 by  very  small  sub-groups, such as 5  or 12 welders.

                Table I.   Divisions  Within the  Group  Studied
Smokers Non-smokers
Welders ( > 0.1 year)
Welders ( <0.1 year)
Non-welders
Total
33
5
33(33)
71
12
12
20(18)
44
Total
45
17
53(51)
115
*
 Numbers in parentheses are those non-welders for whom total dust exposure
indices were available.

Magnetic Measurements

       We measured not only the amount of Fe 0  in the lung, but also its
crude distribution within the lung; in addition, we measured two quantities
which are unique to magnetic particles.  During the application of the
external field, the particles are both magnetized and become rotationally

-------
 aligned with  this  field.  However,  their rotation  is  impeded by their viscous
 environment;  we measured  a quantity involving  this impedance,  called the
 apparent  viscosity.  After magnetization,  the  remanent  field produced by
 the  particles is not steady but  always  decreases in time, typically by  a
 factor of three or four in an hour.  This  decrease, called  relaxation,  is
 due  to random rotations imparted to the initially-aligned particles by
 local motions in the lung; because  the  remanent field is the vector sum of
 the  fields from all particles, it becomes  reduced  as  the particles become
 randomized. We measured the rate of relaxation.  Our  purpose in measuring
 these auxiliary quantities was to see if they  were useful in measurements
 of occupational groups.
        The magnetic measurements were made during  each  subject's visit,
 mainly for his lung function tests,  to  a clinic in the  town of Thetford
 Mines. Between tests  the subject made  an  initial  15-minute visit to the
 magnetic  station in the clinic,  during  which most  of  our measurements
 were made.  If he  showed enough Fe,O., he then  returned  about 20 minutes
 later for a 5-minute relaxation  measurement; in some  cases  the schedule
 allowed a third similar visit.
        During the  initial visit, the subject changed  from his  clothing
 into special  shorts, thereby removing all magnetic items from  his person
 (zipper,  shoes, etc.); dental plates, which are always  magnetic, were also
 removed.  In order  to be magnetized,  the subject was placed  between two  coils,
 as shown  in Fig. 1(A).  These coils  were powered by  two car batteries, and
 generated a magnetic field which was uniform over  the lung  to  ± 10%.  The
 subject was magnetized twice.  The  first was only  for viscosity purposes; a
 field of  400-gauss strength was  applied for only 0.35 seconds  (called the
 short pulse), which magnetized the particles but only partly rotated
 them.  After their remanent field was measured, the  subject  received the
 second,main magnetization; a 750-gauss  field was applied for 30 seconds
 (the  long pulse),  which was enough strength and time to produce complete
 alignment of particles. In Fig.  1(B), the remanent  field due to the
magnetized particles is seen to be  oriented almost horizontally, called
the z-direction,  at the chest and back;  the field  is called B  when
                                                             z
measured over those areas.

-------
 Fig.  1.   Sequence of magnetization and measurement.   (A)  Subject being
 magnetized by the external field (broken lines);  this  field both magnetizes
 and aligns the particles.   (B)  The remanent field around  the  torso produced
 by the magnetized particles.   Over the chest and  back  the field is
 approximately horizontal, and is called BZ.   (C) Subject performing a
 measurement of Bz by moving up to the  gradiometer, which  is fixed to  the
 table.   He is shown at the "near" position  of his far-near-far motion.  Not
 shown here is a plastic shield which is mounted up from the floor, and
 prevents him from touching the fluxgate and causing  an artifact.
                                                          RF   MF   LF
                                                               10 SEC
Fig. 2.  An example of the gradiometer output,  due  to measurements  of a
subject's Bz at the lung points.  The first  three measurements  are  at the
subject's right front  (RF) , middle front  (MF) ,  and  left  front  (LF) ,  recorded
about 30 sec after magnetization; the three  back points  (LB, MB,  RB)  were
recorded 15 sec later.  The B ' s at R and L  are larger than at  M because the
                             £i
detector views more Fe3O4 there.  The final  three signals  (back points
omitted) were recorded 27 min later; the typical  decrease  of  Bz  due  to
relaxation is seen, as well as some baseline  (background)  disturbance.

-------
       To measure B , we used the magnetic detector called the fluxgate.
This detector is compact, simple to use, and has a suitable sensitivity.
It can measure fields down to 5 x 10~  gauss in the 0-3 Hz bandpass in
which it was used, corresponding to the detection of about 0.05 mg of
Fe,O  in the lungs. In dealing with the problem of the magnetic background,
the steady and fluctuating backgrounds are considered separately. To deal
with the steady background, the detector was always rigidly fixed in
position, and the subject moved up to the detector for a measurement; in
this way the change in detector output was only due to the subject's B  .
                                                                      z
The problem of the fluctuating background, due for example to moving cars,
 (which act as large moving magnets), was minimized by using the fluxgate
in the gradiometer mode as follows. The output of the model we used
 (#MF-5000, Automation Industries, Ltd.) is the sum of outputs from its
two identical probes; they were mounted in-line about 10 cm apart,
 (horizontally, as shown in Fig. 1(C)), and oriented oppositely so that
the output was the difference in B  between them. Thus a fluctuating B
                                  z                                   z
produced by a distant source was largely cancelled, while the B  from a
                                                               z
subject's lung was not cancelled because it was much larger in the nearer
than the further probe. The fluctuating background in this mining town was
negligible when dealt with in this way.
       For a measurement the subject first stood out of range of the detector,
then moved inward and placed a point of his torso at the detector as shown
in Fig. 1(C), then stepped back again.  The detector has a bell-shaped
response curve in angle, with the maximum at the 0° or z-line and a
half-maximum at about ± 22°, corresponding to about ± 3 cm at the lung.
Measurements were made at three marked points on the chest and three
corresponding points on the back, called the six lung points; the chest points
were on a horizontal line 10 cm above the xiphoid, one at the midline and
two 10 cm on each side, at about the nipples.  In addition, measurements
were made at the chin and three abdomen points, 20 cm below the chest
points.  Measurements at these non-lung points were made in order to detect
any magnetic contamination occasionally present in the abdomen and the
head, which could mask the Fe30. in the lung.  When such contamination was
seen, it could usually be demagnetized with a hand-held magnetic tape
eraser; in five subjects the contamination was large enough to resist this
procedure,  and they were not included in the study.

-------
 Processing the Magnetic Data
      We here describe how the values of B  from the six lung points were
                                          Z
 converted into the amount of Fe,O. in the lung and the auxiliary quantities.
 The BZ values were first summed in these various ways: EF (three front B 's),
 EB (three back B 's), EM (two middle B 's), EL (two left B 's),  ER (two
                 z                     z                   z
 right BZ'S), and E6 (all 6 B •s).  By using the relaxation rates  (see below),
 these sums, from both short and long pulses, were extrapolated back to 0-time
 (end of the magnetizing pulse).   Next, the long-pulse sums were  combined in
 various ways to yield the crude distributions in the lung; EL/ER indicates
 the amount of Fe O  in the left lung compared to the right,  2EM/ (EL +ER)
 indicates the amount in the middle compared to the average side,  and EF/EB
 is the front/back ratio.
      To calculate the amount of Fe O , we first consider the simplest
 relationship between this quantity and B ; this is when all the  Fe O  in the
                                         Z                         J rr
 lung is imagined to be compressed into a point source, called the magnetic
 dipole; for which

              B  =2Imz~3         or     B / (2 I m)  = z~3          (1)
               z                            z
 where B  is in gauss, m is the Fe,0. mass in grams, and z is the distance
        Z                         J fi
 to the field point in cm.  I is a property of the magnetic dust  called the
 magnetization (in emu/gm); based on our measurements of laboratory samples,
 and on literature values, we chose I = 10, so that eqn.  (1)  becomes
 B /20m = z~ .  For large z, an actual lung behaves as eqn. (1),  with the z
  z
 falloff.   However, for z < 20 cm,  the B  dependence on z is the  curve shown
                                        Z
 in Fig. 3(A).  To use this curve,  if there were only one probe,  say the near
 probe of the gradiometer, we would simply use the z for that subject to
 determine the ordinate B /20 m. and knowing B ,  we would solve for m.  However,
                         z                  z
 because of the far probe of the gradiometer (z+8.6 cm),  we use the
 difference of the ordinates, called F , so that

                      Fg = (1/20 m)  
-------
 where m is now the amount of Fe 0  in mg.   The error involved in using this
 formula is due to two sources.   The first  is a probable error of ± 30% in I;
 since I is a constant for all subjects,  this error does not affect their
 relative m.   The second is a probable error of ± 30% in F ; this is due to
 uncertainty in individual lung spacing,  hence this is the relative error in
 m for the data presented here.
      For the viscosity quantity,  we defined and used this term:
 apparent viscosity = Z6    / Z6     ,  which should increase with the
  "^              J     long    short
 viscosity experienced by the particles;  however,  it would also be sensitive
 to the shape and size of particles.   In  laboratory measurements of this type
 of quantity, we found it to be reproducible for an individual,  but it does
 not correlate with obvious variables,  such as smoking or the residence time
 of Fe O. in the lung.
      Relaxation curves (B  vs time)  have been well studied; they have a
                          z
 characteristic mathematical shape (Cohen,  1974,  1978)  and a dropoff rate
 which depends on the residence time of Fe  O. in the lung and on smoking.  The
 shape of the curves is seen in Fig.  3(B);  recently inhaled Fe_O. results in
 a rapid dropoff such as the lowest  curve of Fig.  3(B),  while Fe_O.  inhaled
 years ago results in a slower dropoff, such as one of the upper curves.
 Smokers who have recently inhaled Fe O  show a much more rapid dropoff than
 non-smokers  (Cohen et al.,  1979).   It  would seem,  therefore, that the
 relaxation rate of these workers  might reflect both the residence time of
 Fe,O.  in the lung,  and the amount of smoking.   Because  of time restraints,
 which allowed only two or three B  measurements  following long-pulse
                                  2
 magnetization,  the following system was  chosen to  quantify the relaxation.
 Five  relaxation groups were  arbitrarily  defined, .as shown in Fig.  3(B).   Two
 measurements  >20  min  apart  were  usually enough  to place a subject  in a
 particular group;  three or more measurements yielded increased accuracy
 and allowed  divisions  within  the  group.  Once the  relaxation group  was
 determined,  the various sums  of B  could be extrapolated back to 0-time.
The Indices of Exposure
     Dust sampling had been carried out since 1948 at various locations
in the local mines and mills, using the midget impinger.  This method yields
the density of respirable dust in the air, where the dust here consists of
both asbestos fibers and other particles.  Using this density and the
                                    8

-------
 10
                  PURE DIPOLE (Z"3)
      LUNGS
      (EXTENDED
      SOURCE)
   10
  15       20     25
Z-DISTANCE  IN CM
                                                      10    20    30   40    50    60
                                                    MINUTES AFTER MAGNETIZATION
Fig. 3.  Curves used in processing  the magnetic measures.   (A)  Falloff  from a
dipole source (z~ )  and from an  extended lung. The extended lung falloff  is an
average of falloff measurements  from various subjects.  The distance z  is from
the lung center to the field point, estimated from the thickness of the
subject's chest.   (B)  The  five relaxation groups which are used here.   These
are bounded by four relaxation curves of characteristic shapes  derived  from
actual, measured curves; however, the four dropoff rates are arbitrarily  chosen.

-------
 month-by-month employment  record  of  each worker, a total dust exposure index
 was calculated,  defined as (density) x (years worked at this  location,
 corrected to a 40-hour working week),  summed over each job held by the
 worker,  up to the end of 1966.  The  index is given in units  of  (millions of
 particles/cu ft) x (years), or mpy/cf.  That the termination  of the index
 by the end of 1966   does  not seriously  affect its  use for  1974  supported  by
 data on the steady decrease in dust  density in the mills; for example, it had
 fallen from an average of  75 mp/cf in  1948 to 10 mp/cf in 1966, with
 reductions thereafter.
      The exposure index for welding  was derived more crudely, without dust
 sampling.  The index we used was  simply the period of welding exposure in
 years, corrected to a 40-hour working  week; e.g., for a 10% work-time
 exposure for 20 years,  the exposure  index was 2.0 years.   A welding history
 was obtained by careful questioning  of each worker during the clinic visit
 and, in cases of ambiguity,  by telephone follow-up.  In contrast to the
 total dust index,  the index used  here  applied up to the time of this study.
 If a worker had any welding exposure whatever, for example if only to nearby
 welding, he was  grouped by us with the welders.  Table I, therefore, shows
 17 "welders" with exposure of only <0.1 year.  We included torch cutting,
 often used in the mines, as  a welding  procedure.
 RESULTS
      The  amount of Fe,O  , its crude distribution within the lung, the
 apparent  viscosity^ and the relaxation rate were readily measured; their
 distributions within the group* are given in Figs. 4 and 5, and in Table II.
 Fig.  4 shows that the amount of Fe3O4 in the lungs of welders is indeed
 greater than in non-welders, as we had seen during the measurements.  The
 average amount of Fe 0. in the lungs is 1.3 mg for a non-welder, and
 7.8 mg for a welder; if the first interval is excluded, the amounts increase
to 1.7 and 8.3 mg.  There is therefore about fives times more Fe 0. in the
lungs of the welders of this group.
     The results of one type of distribution measurement of Fe 0  within the
lung, the ratio of left/right, is shown in Fig. 5.  The fact that less
*  As distinct from Fe.,0. distribution within the lung.
                     '34

                                    10

-------
       20
     CO

     o
     CO
     en
     UJ
     a.
       15
     feio
     oc.
     UJ
     CD
          0
                                                   — NON-WELDERS

                                                   	WELDERS
0.10    0.32     1.0     3.2      10      32

      Fe304 IN MILLIGRAMS (LOG SCALE)
                                                           100
320
         Distribution within the group of the amount  of  Fe  0   found in the lungs

                                            A -       -   '
Fig. 4.

of the 53 non-welders, and the 62 welders.  A log scale is~used, except for the

first interval which extends down to  and  includes zero. This interval is elevated

for the non-welders because it contains office workers, and for the welders

because it contains borderline cases  where it was not certain that there had beer,

exposure to welding dust.
        40 r
     CO
     •z.
     o
     CO
     CC
     UJ
     Q.
       30 -
-





0.4 0.






5 0.






8 1.






D 1.





	 1 1
2 1.4 1.6 1.8
     O  20 h

     cc
     UJ
     CD
        10 -
         0
             u.^     u.o      u.o


                               LEFT/RIGHT



Fig. 5.  Distribution within the  group of the ratio of  the  remanent  field on the

left side of the torso to that on the right (ZL/ER).  On  the  average,  the left

side produces less field, hence contains less Fe  O  than  the  right side. This

is because of the reduced volume  of lung on the left, due to  the heart.
                                    11

-------
 dust is seen on the left,  demonstrates that the magnetic method is
 capable, by measurements at an industrial site, of yielding at least some
 distribution information.   The large left/right spread is due to our crude
 B  measurements;  if the measurements would have been made with more
  z
 accurate positioning of each subject,  then the spread would have been far
 narrower.  The remaining magnetic data are shown in Table II, along with
 data on dust exposures.  For the magnetic quantities, the total number of
 subjects in each  row (All)  is limited to those who contain enough Fe 0  for
 this type of measurement.
      For the middle/side ratio,  the table shows a most probable value of
 about 0.5;  this indicates  the smaller  amount of Fe O  "seen" by the
 detector at the middle,  where there is almost no lung,  in comparison to
 the side.  Again,  the method reveals at least some distribution information.
 However, the meaning of the most probable value of about 0.7 for the ratio of
 front/back  is not  as obvious: although there is more posterior than anterior
 lung, there are also differences in lung-to-detector spacing between front
 and back, as suggested by  one of the correlations presented below.   The
 large spread about the 0.7 value is again due to coarseness of the
 z-spacing,  to which this ratio is sensitive.   In the next quantity, the
 relaxation  group,  the coarseness has been reduced by averaging (16) , and the
 spread about the most probable group II is actual Fe O  behavior.   This
 applies to  the apparent  viscosity as well,  where the most probable value
 is about 1.8.
      For the dust  exposure  data,  the distribution within the group is
 different for  the  years  welding  in  comparison to the total dust index.  While
 they both have a large number in the first interval,  due to "borderline"
 welders  for the former and  office workers  for the latter,  the distribution
 for  the  latter has  an obvious dip in the  middle interval.   This was a
 deliberate  choice  in the makeup  of  the group for the lung function
 measurements;  most workers were  selected  to have either low or high exposure,
 but  a small  group was selected at the  mid-level as  well.
     To visually examine the  relationship  between the amount  of  Fe_O  and
 the  total dust  exposure  index, these quantities are  plotted  against  each
other in Fig. 6 for all  the non-welders,  and  in Fig.  7  for these
non-welders who are non-smokers.  In both  figures  it  is  seen that there  is
considerable scatter, with no obvious  correlation between the two quantities;

                                    12

-------
         Table II.  Number of Persons vs. Magnetic and Dust Quantities
Quantity
Middle/Side

Front/Back

Relaxation
Rate

Apparent
Viscosity
Welder's
Exposure


Total Dust
Exp. Index

Definition
2EM/(EL+ER)
number
EF/EB
number
group

number
E6 /E6
X. S
number
years

number
*
mpy/cf

number

0.0-
0
0.4-
7



< 1.0
0
<0.1

17

0-3.2

21

0.2-
4
0.6-
27
I

6
1.0-
7
0.1-

8

3.2-

8
Ir
0.4-
39
0.8-
26
II

40
1.4-
20
0.32-

13

10-

27
iterval;
0.6-
32
1.0-
16
III

15
1.8-
20
1.0-

18

32-

10
3
0.8-
20
1.2-
9
IV

2
2.2-
11
3.2-

2

100-

26

1.0-
2
1.4-
1
V

0
2.6-
10
10-

4

320-

20

1.2-
0
1.6-1.8
1



3.0-3.4
3
32-

0

1000-

0
All

97

87


63

72


62



112
Intervals are in log scales except first interval
 this applies to both the points clumped near the origin, and the more visible
 points.  The same scatter and low visible correlation is seen in the welding
 plot, in Fig. 8.  In order to see if the correlations between these various
 quantities were indeed as low as appears visually,  a statistical analysis was
 performed; the analysis was extended to include correlations between any of
 the variables in this study where there might be some basis for correlation.
 The analysis involved three computations.  The first was least squares,  which
 provided the correlation coefficient r and a straight line; the second was of
 the Spearman's rank correlation coefficient r  (Snedecor and Cochran, 1967);
                                              S
 the third was a robust regression to provide a straight line, using an
 iteratively weighted   procedure in which the weights are inversely
 proportional to the relative size of the residuals  in the previous procedure
                                     13

-------
   10
    8
CO
o:
CD
 ro
                                                                  ROBUST.
     0
 200              400              600
DUST  EXPOSURE INDEX (MPY/CF)
800
   Fig 6.   Amount of Fe O  in the lung of each worker vs his total dust
   exposure index, for the 51 non-welders.  Six points near the origin have been
   omitted here.  Although visual inspection shows  no obvious correlation,
   Spearman's rank correlation gives r  =0.50 and  the least-squares regression
                                    s
   gives r = 0.45; both are significant at the 1% level.  The straight lines
   are due to the least squares  fitting of the data, and the robust regression
   procedure.
                                      14

-------
         10
         8
     CD
     S
     <  6
     cr
                          200             400              600
                   DUST EXPOSURE INDEX (MPY/CF)
Fig. 7.  Amount of Fe 0  vs.  total dust exposure  index, for the 18  non-
welders non-smokers.  Again visual inspection shows no obvious correlation,
but Spearman's rank correlation gives r  =0.62 and least-squares regression
                                     s
gives r = 0.80, again significant at the 1% level.  The line slopes are
greater than in Fig. 6,  therefore these non-smokers appear to have  more dust
in their lungs than the  smokers.
                                   15

-------
                      5                10
                          YEARS  OF  WELDING
15
20
Fig. 8.  Amount of Fe 0^ vs. years of exposure to welding dust,  for the
45 asbestos workers with >0.1 years of welding.  The arrow indicates that
one point is off the figure, at the upper right, with 29 years and  140 mg.
Again visual inspection shows no obvious correlation, but Spearman's rank
correlation gives rg = 0.37  (significant at about the 1% level),  and
least-squares regression gives r = 0.88 (high significance).  The least
squares line is  pulled up by the high point.
                                  16

-------
    Table III.
Correlation Coefficients r  due to Spearman's Rank Analysis
                          5
                  Group
 Vital Capacity    all
                   non-sm.
              Chest
            Thickness
Relaxation
   Rate
 Fe304
Density
                                          low(104)t
                                          low  (39)
Fe304
Amount
                              low(106)
                              low  (41)
 Front/Back
  all       0.38  (88)**     low  (76)
  non-sm.                   low  (28)
 Middle/Side
  all       0.21  (93)*      low  (76)
  non-sm.                   low  (28)
                             0.22  (93)*
                              low  (32)
 Appl. Viscosity  all
                   non-sm.
                            low  (64)
               0.25 (67)*
               0.49 (18)*
               0.23 (72)*
               0.47 (19)*
 Relaxation  Rate   all         low (76)
                   non-sm.     low (28)
                                        0.25  (72)*
                                        0.40  (26)*
                              low (77)
                              low (27)
 Cigarettes/Day    all
                            low  (77)
 Years Welding
  all
  non-sm.
  >0.1 yrs.
  >0.1 yrs.
   + non-sm.
   low (41)
                                            0.64 (12)*
 0.58(109)**
 0.53 (40)**
0.55(115)**
0.50 (42)**
0.37 (45)**
 low (12)
 Total Dust Index  all
 (only non-weld.)
                  non-sm.
                                        0.47  (50)**    0.50  (51)**
                                        0.67  (18)**    0.62  (18)**
 t  r  is simply called low  (low correlation)  if it is not significant at
       the 5% level; the number of  cases  is  given in parentheses.
 *  significant at the 5% level
**  significant at the 1% level or  less
                                         17

-------
 (Hosteller and Tukey,  1977)-   This method was  appropriate because it is not
 sensitive to the many  outlying points  of our data to which the least squares
 regression is quite  sensitive.
      The results of  the Spearman's rank  correlations are  given in Table III;
 the coefficient r is  simply  called "low" unless  the correlation is
 significant at the level of 5%  or  less,  (indicating a probability of 5% or
 less that this could have occurred at  random).  The chest thickness used
 in the table had been  coarsely  measured,  with  the subject's back at the
 wall.   The relaxative  rate had  been quantified so that the number increased
 linearly with the relaxation  group,  and  subdivisions within a  group were
 also used.   The density of Fe 0 was the amount of Fe O.  divided by the lung
                              J  4r                      .5 ft
 volume,  obtained from  their lung function measurements.   The vital capacity
 (% of  predicted value)  was similarly obtained;  it can be  an early indicator
 of pulmonary abnormality due  to exposure  to  asbestos dust,  however the  range
 is small in this group because  of  their  0/0  chest films.   For  the cases of
 direct interest (Figs.  6,  7 and 8),  we here  give  least-squares  results.  For
 the amount  of Fe 0  vs.  the total  dust index,  r = 0.45  for all  non-welders
 (51) and r  = 0.80  for  those that are non-smokers  (18);  these are both
 significant at lower  than the  1%  level,  in  agreement with the  rank
 correlation.  For the amount of  Fe  O. vs.  the years  of welding,  r = 0.88
 for those welders  with  >0.1 years  exposure (45);  this  is  certainly
 significant  at lower  than the  1%  level.   Further comments  on this table
 are made below.
DISCUSSION
     If the ratio of Fe 0 /asbestos was always a constant, both in respirable
dust and after deposition in the lung, then this study would not have been
necessary; a measurement of the amount of Fe_0. would directly give the  amount
of asbestos in the lung.  However, while this ratio is quite constant for the
airborne dust at most locations of the Quebec mines and mills, it is not
known if it changes in the lung.  The ratio in more than 100 samples from the
mines and mills was recently measured (O. Djamgouz of Laurentian University,
personal communication); the ratio of Fe_0./chrysotile was always found  to
be in the range of 2-5% by weight, where the lower the grade of asbestos, the
higher the percentage.  In the lung, however, the Fe3°4 particles probably

-------
 become detached from the fibers  to which  they were  stuck, and may clear out
 differently from the fibers;  this is  suggested  by the  fact  that chrysotile
 fibers in the lung are known  to   dissociate,  after  some  months  or sooner,  into
 fibrils (Suzuki and Churg,  1969).   Further,  there may  be some inhaled
 Fe3°4 dust which was not attached to  fibers; and which may  also clear out
 differently.   If the clearance rates  are  indeed different,  then the ratio
 of Fe3O4/asbestos will change in time;  the  greater  the difference in
 clearance rates and the variation in  this difference from person to person,
 then the greater the uncertainty in the magnetic method for  determining
 the amount of asbestos.
      Assuming that the ratio  is  not constant, the straightforward way to
 find out how well the magnetic method can determine the amount of asbestos
 would have been to magnetically  measure autopsied lungs of miners and
 millers, and correlate the  amount of  Fe_O with the actual amount of
 asbestos present.  An alternative approach would have  been to correlate the
 amount of Fe O. with lung abnormality in  miners and millers; here we would
 have found out how well the magnetic  method  determines both  the amount of
 asbestos and its biological effects,  which is another  issue.  Neither way
 was available to us at that time,  hence we chose a variation of the second
 approach.  This was to find the  correlation  between the amount of Fe O  and
 a dust exposure index which does correlate with lung abnormality;  namely
 the total dust exposure index.   If the amount of Fe O. would correlate
 highly  with the amount of  asbestos,  then at best we can expect a low but
 significant correlation between  the amount of Fe O. and this total index:
 this is because  the correlation  is  not  high between this index and lung
 abnormality.   On the other  hand,  if there was an intrinsically low
 correlation between the Fe  0.  and  the asbestos, then we would expect the
 correlation between the Fe  0.  and  the index to  vanish.
    Although  the  eye  shows  no  correlation in Figs. 6 and 7,  the statistical
analysis  does show a  correlation which is  low but statistically significant.
This is compatible with  a high correlation between the amount of Fe_O  and
asbestos, and incompatible with a  low correlation between them.   Stated
otherwise,  the magnetic method appears  to indicate the amount of asbestos in
the lung, at  least to some extent;  the  relative FeO /asbestos clearance
rates do not  appear to vary greatly from person-to-person.    It follows that
the large scatter in Figs. 6 and 7 must be due,  in large part, to individual

                                     19

-------
differences in deposition and clearance.  The points well below the robust
lines, for example, are due to less deposition or more clearance than
normal (per unit of exposure), while the points well above the least squares
line are due to more deposition or less clearance than normal.
   It is tempting to speculate that those points very high off the line,
say the two points near the upper left corner of Fig. 6, may represent
individuals at increased risk, in the sense that their dust clearance
mechanisms may be impaired. This would suggest that the magnetic method may
be useful in the primary industries for screening of workers with supposed
impaired clearance, hence at increased risk; although the dust levels are
now reduced, impaired clearance may nevertheless result in enough Fe3C>4 to
be measured.  In that regard, we note in Table III that the amount or
density of Fe 0  does not correlate with the vital capacity, the possible
early indicator of abnormality;  this strengthens the idea that the
measurement of amount of Fe.O. should be only for the purpose of determining
the amount of asbestos, not its biological effects; the increased risk
mentioned above would only be due to an increase of asbestos in the lung,
not its effects.
   The conclusion that Fe O. indicates in part the amount of asbestos is indirectly
supported by the data on welders.  Whether for all workers, as shown in
Table III, or for welders with >0.1 years exposure (Fig. 8), there is a
correlation between the Fe,O. and the welding exposure which is statistically
significant.  Again, therefore, the magnetic method indicates/ to some
extent, the amount of occupational (welding) dust in the lung.
   The large increase in the line slopes in Fig. 7 in comparison to Fig. 6
deserves comment.   If statistically valid, this increase implies that
smokers have less dust in their lungs than non-smokers, per unit of
exposure.   This would appear to be at variance with the,recent result (Cohen
etal., 1979)  that smokers show impaired long-term clearance from the lung,
hence retain more dust than non-smokers.   Because the dust-count (alveolar)
in that work did  not  begin  until  short-term clearance  (from airways) had  been
completed,  this discrepancy could be resolved if one assumes that smokers
have far less alveolar deposition than non-smokers, perhaps because of
different airway diameters, etc.   Stated otherwise, if this result is
valid,  then in smokers perhaps less dust is deposited in the alveoli, but
this alveolar dust is cleared away more slowly than in non-smokers.   But

                                    20

-------
 how believable is the slope increase?
      On the one hand,  Fig. 7 has only 18 points,  with the large slopes
 therefore depending on only a small number of special points.  On the other
 hand, the slope increase exists for both the least squares and the robust
 lines.  Perhaps we can only say that these results are consistent with
 smokers having less dust in their lungs than non-smokers,  for the same
 exposure.  Certainly we do not  see  here  the  large,  reverse effect  seen by
 Cohen et al;  in that case both slopes in Fig. 7 would have been smaller
 by a factor of ten, which is clearly ruled out.  It is unfortunate that the
 group of non-smoking welders was too small (12)  to see if their slopes would
 also rise, compared to Fig. 8; Table III shows  only a low correlation,  hence
 the line is meaningless.
      Because we see that the magnetic method can be useful in the primary
 asbestos industry for determining asbestos in the lung, we will comment
 about various other aspects of the magnetic technique which have shown up
 in this study, which may be of value in applying this method.
      Concerning the practical aspects of these measurements, the fact that the
 Fe_O  was readily measured in a mining town  does  not  insure that other
 groups can easily be  measured in  other settings.   It  is  a  matter of
 magnetic signal/noise;  more Fe O.  in the lungs produces  larger gradiometer
 signals which  can override  larger  background  noise.   In  this  regard, the
 average amount of Fe  0  in  the lungs of different  groups is important.  We
 saw  that  the average  in  this group was about  2 mg  and  10 mg for non-welders.
 We had measured some  foundry workers while in the  Thetford area and found
 the  average to be 140  mg(!)7 while measurements  at the MIT  lab of workers in the
 secondary asbestos  industries  (finished products)  yielded  the much lower
 average of about  0.10  mg.   Foundry workers can therefore be measured  in an
 urban  setting  with  high background;  while  workers  in  secondary asbestos
 industries could  only be measured  with a detector  more  sensitive than a
 fluxgate, in a magnetically  quieter  setting or with better
 gradiometer-cancelling techniques.
      We next  consider the correlations in Table  III of the auxiliary magnetic
quantities.  There  is a significant  correlation between the front/back ratio
and chest thickness, indicating the  front  B  's are sensitive to the
                                           z
thickness of the muscle and  fat in the  chest, therefore our methods of
                                     21

-------
positioning subjects should be improved, perhaps with a  standard  E-spacing
used  for  all  subjects.  The same applies to the middle/side  ratio, which  also
shows  some correlation; however, its correlation with the amount  of  Fe3C>4
is probably meaningless, because the correlation became  lower with the
more  uniform  group of non-smokers.  The applied viscosity shows correlation
with  both Fe  0  amount and density, which is believable  because there could
well  be some  sort of "clogging" viscosity with increased amounts  of  dust.
       The relaxation shows a similar but lower correlation,  again believable
for the same  reason.  However, the lack of correlation between relaxation
and amount of smoking is surprising, because this correlation in  laboratory
measurements  was high; one explanation could be that the latter result
was based on  relaxation of dust recently inhaled, and perhaps the long
residence time of Fe O  in our asbestos subjects had washed  out the
correlation.   Another surprising result with relaxation is not shown in
the table.  The correlation between relaxation rate and the  total dust
index  (74 subjects including welders)  showed r  = 0.37, which is  significant
                                              s
at the 1% level.  However,  for the smokers within that group the  correlation
vanished,  so that the first correlation is not believable, and could be only
a random  effect which is to be expected occasionally.
      The data from Table III suggests, therefore,  that the apparent viscosity
and relaxation rate are worth pursuing further as auxiliary quantities in
future measurements.   The data also show that density of Fe O. in the lung
generally yields a somewhat higher correlation in comparison to the amount
of Fe3°4'  this suggests that the density is a more  useful quantity than the
amount of  Fe,0  in the lung,  and that  this comparison should be pursued
further as well.
                                    22

-------
REFERENCES



Becklake, M.  R. , Fournier-Massey,  G.,  Rossiter,  C. E.,



    McDonald, J. C.  (1972).   Lung  Function  in  Chrysotile Asbestos



    Mine and  Mill Workers  of  Quebec.   ARCH,.i ENVIRON. HEALTH  24,



    401.



Cohen, D.  (1973).  Ferromagnetic Contamination in the Lungs  and



    Other Organs of  the  Human Body.  SCIENCE 180, 745.



Cohen, D.  (1975).  Measurements of the Magnetic  Fields Produced



    by the  Human Heart,  Brain and  Lungs.  IEEE TRANS. MAG.-11,



    #2,  694.



Cohen, D.  (1978).  "Report of the  Low-Fiend Group:  The Magnetic



    Field of  the Lung."  Publication MIT/FBNML-78-1.  National



    Technical Information  Service, Springfield,  Virginia.



Cohen, D.,  Arai, S.  F.,  and Brain, J.  D.  (1979).  Smoking Impairs



    Long-Term Dust Clearance  from  the  Lung.  SCIENCE 204, 514.



Freedman, A.  P., Robinson, S.  E.,  Goodman, L., et al (1979).




    Non-Invasive Magnetopneumographic  Determination of Luna  Dust



    Loads in  Steel Arc Welders.  CHEST 7_6, 352.



Freedman, A.  P., Robinson, S.  E.,  and  Johnston,  R. J.  (1980).



    Non-Invasive Magnetopneumographic  Estimation of Lung Dust



    Loads and Distribution in  Bituminous Coal  Workers.  J.




    OCCUP- MED.  22,  613.



Gibbs, G. W.  (1971).   Qualitative  aspects of dust exposure in



    the Quebec asbostos mining and milling industry.  In




    "Inhaled Particles III."   (W.  H. Walton, Ed.), Vol. II,



    pp. 783-798, Unwin Brothers Ltd.,  Surrey,  England.
                           23

-------
Gibbs, G. W., and Lachance, M.  (1972).  Dust Exposure  in  the




    Chrysotile Asboestos Mines and Mills of Quebec.  ARCH.




    ENVIRON. HEALTH 2_4_, 189.



Kalliomaki, P. L., Korhonen, 0., Vaaranen, V-, et al  (1976).




    Lung Retention and Clearance of Shipyard Arc Welders.   INT.




    ARCH. OCCUP. ENVIRON. HEALTH 42, 83.




McDonald, J. C., Becklake, M. R., Fournier-Massey, G., and



    Rossiter, C. E. (1972).  Respiratory Symptoms in Chrysotile




    Asbestos Mine and Mill Workers in Quebec.  ARCH. ENVIRON.




    HEALTH 24, 358



Mosteller, F. and Tukey, W. (1977).  "Data Analysis and Regression."




    Section 14H.  Addison Wesley, Reading, Massachusetts.




Peress. L., Hoag, H.,  White, F., and Becklake, M. R. (1975).  The




    Relationship between Closing Volume, Smoking, and Asbestos




    Dust Exposure.   CLIN. RES. 23.  647A.




Rossiter, C. E., Bristol, L. J., Cartier,  P. H., Gilson, J. G.,




    Grainger, T. R., Sluis-Cremer,  G.  K.,  and McDonald, J. C.




    (1972).   Radiographic Changes in Chrysotile Asbestos Mine




    and Mill Workers in Quebec.   ARCH.  ENVIRON. HEALTH 24, 388.




Snedecor, G. W., and Cochran,  W. G. (1967).  "Statistical Methods."




    6th Edition, p.  194.   Iowa State University Press, Ames, Iowa.




Suzuki, Y.,  and  Churg,  J.  (1969).   Structure and Development of the



    Asbestos Body.   AM.  J.  PATH. 55, 79.
                           24

-------
                           APPENDIX
The preceeding text, without this appendix, has been submitted




for journal publication.  The material in this appendix therefore




is available only in this report.  The material consists of




details of methods  (Appendix A) and of the results (Appendix B),




and is intended for those who wish more knowledge of this work.




However, Appendix A contains some details which are useful for




magnetic measurements of the lung generally.  The style of presen-




tation in the appendix is occasionally more casual than in the




main text.
                               25

-------
 APPENDIX A.   SOME DETAILS OF METHODS

 Performing the Magnetic Measurements
     We first give a  brief outline  of the  procedure  we  used during the
 first, 15-min  visit of the  subject  to  the magnetic station.   We then pre-
 sent some details involved in this procedure.

 Outline
 1.  The test  is explained to  the  subject;  the subject gives his I.D.  and
    welding (torch cutting) history.   Next the  subject  removes  all his
    clothing  except socks,  and changes into athletic shorts furnished by
    us.
 2.  The chest thickness  is  measured,  as  well as the  subject's chest height.
    Three  points are  marked on his  chest on a horizontal  line 10 cm above
    the ziphoid, and  10  cm  apart.   The height of the platform between coils
    is  adjusted  so  that  the subject's chest will be  centered between  the
    coils.
 3.  The subject  is  magnetized  with a short pulse of 0.3 sec, of  strength of
    about  400  gauss.  Exact pulse length is noted on  an oscilloscope.
 4.  Subject's  chin  and abdomen are  checked for magnetic interferences by
    measurement with  the  flux  gate  gradiometer.
 5.  Guided by  the operator, the subject  undergoes his far-near-far motions,
    touching the shield  for each  "near"  position.  First  the three chest
    points are measured, then  the three back points.
 6.  The  subject is magnetized  for 30 sec at 750 gauss.
 7.  Again the  gradiometer measurements are made for  each  of the chest and
    back points.
 8.  Step 7 is repeated approximately 20 min later, then again later if the
    subject's time permits.

Various Details
       The MF-5000 gradiometer was arranged so the outer tips of  the  probes
                                     26

-------
 were 4 11/16 inches apart (hot-spots were 3 3/8 inches apart).   The MF-5000
 was always used on the 1 milligauss range;  it fed a Varian recorder used
 on either a 10,  100,  or 1,000 millivolt range.   The chart speed was always
 4  inches  per minute.   A calibrated coil was always present on the sensor
 closest to the subject,  fed by the current  from a calibration unit.  It
 yielded either 1,  5,  or 10,  or 50 or 100 or 500 or 1,000 or 5,000 or
 10,000 times 10~  gauss, applied to one probe only.   The capacity across
 the output of the  MF-5000,  which filtered out the higher frequencies,  was
 850 mf; this yielded  a 3db  point somewhere  between 0.5 and  1.0 Hz, with a
 fall-off  of 6db per octave,  since filtering took place only at one point
 in the circuit;  the chart recorder also had its own,  slow time constant.
        The two large  magnetizing coils were set with an inside spacing of
 12.0 inches, and were connected in series and fed by two storage batteries
 in series (24 v),  yielding  750 gauss.   When the switch was first closed,  a
 current of about 400  amps flowed, which drooped during the 30-second appli-
 cation time to about  375 amps.  During the  30-second magnetization,  each
 subject was rocked from side to side slowly,  in order to get fairly uni-
 form magnetization across the entire chest; he  was not moved up and down,
 however,  or in and out,  but placed about half-way, fore and aft.   One  sub-
 ject was  too heavy to fit between the magnets,  and he was magnetized in-
 stead with a powerful Alnico magnet (about  250  gauss  at the lung),  slowly
 rubbed over his front and back.
        Various difficulties  were encountered  during the  measurements of
 the first few subjects;  these included inability to exactly position the
 measuring points at the  fluxgate,  much magnetic contamination  on the skin
 especially around  the chin and neck,  and the  phenomenon  we later called
 spiking.   To cope  with the first difficulty,  we experimented with the
 subject's  posture.  During the first  day or two,  we  asked the  subject  to
 "hang  loose"  while placing marks and  making measurements.   Subsequently,
we  found that this led to the trouble  of back measurements,  which were
too low on the back;  it  was  better to  have  the  subject stand erect at  all
times, so that the  back measurement lined up more  on  the  upper  part of  the
back,  closer  to the center of activity.   Concerning the  next problem,  we
began  to vacuum the chest and often the  back, certainly  the neck,  of each
                                    27

-------
 subject.   We  also  tacked Kleenex to the subject's  chest  and back,  in  order
 to prevent magnetite  in the air from clinging to his  skin  during magnetization.
 For spiking,  we eventually found, by trial and error, that if we covered
 the front  of  the lucite shield with ordinary brown masking tape, the  spiking
 disappeared.   The  spiking therefore is probably associated with some
 electrostatic discharge, at frequencies to which the  fluxgate is sensitive.
 That is, skin rubbing against the lucite appears to generate charges  in
 some fashion  or other; this does not appear to be  the case with skin  against
 masking tape.
       One phenomenon we noticed, as yet unexplained, was  with subjects
 who had very  high  readings, such as some welders.  It was  noticed, when
 they stood at the  "far" position, that we could see their  breathing on the
 chart recorder.  The  amplitude of this modulation  seemed to be far greater
 than one would expect at that distance.  We should also  note that, when a
 subject stood at the  far position, with perhaps 10 or 12 inches between his
 skin and the  lucite,  the chart reading was not a true zero B ; when a subject
 was asked  to  do a  180° turn at this far position,  and we could see the
 modulation of the baseline accordingly, all readings should really be corrected
 upward by  a certain percentage, because of the distance  effect; that  is,
 subjects were not adequately far away in their far positions.  It appeared,
 on  visual  inspection, to be a correction of perhaps 3 or 5%.  Another point,
 of  some interest, is  the modulation of the signal during breathing, by
 breathing, when the subject is at the lucite.  It has been noticed that when
 the  subject stands with his front pressing the lucite, there is almost no
 modulation by breathing;  however,  when his back is against the lucite, it
 is  almost  always seen that there is a large modulation of  the signal.
       For investigators using this technique for  determining magnetic
 particles  in  the lung, the following list of interferences (both correctable
 and  non-correctable) may be useful.
       It  is,   of course,  important that non-magnetic clothing be worn, and
 that  shoes, belts,  watches and false teeth be removed.   In this study white
 athletic shorts with elastic waist bands were provided, although hospital
 trousers with  drawstrings would have been preferable.   Magnetic checks at
neck height revealed steel pins in teeth;  they also revealed a magnetic
piece in a toupe!   Abdominal checks showed up steel sutures from surgery,
                                   28

-------
 and often steel chips ingested with food.   Steel items in the  mouth  and
 abdomen were often correctable by using a  large but hand-held  magnetic 60-Hz
 eraser.   Steel particles embedded in the skin were  occasionally erasable
 by  a short-range eraser.
        The  sensitivity of the fluxgate gradiometer was  limited by  its rms
noise  of  about ^5x10   gauss (rms //Hz").   In  addition  to  this intrinsic
noise  the field changes produced by  passing cars and trucks  would  occasionally
introduce a total  reading uncertainty of up to  10x10  gauss.  Assuming that
they were random,  this corresponds to an uncertainty in calculated total
Fe3O4  from  one reading of ±0.3  mg of Fe 0  or  ±0.1 mg for  the average of
12 readings, because
                               a= a//n
       where  a  is  the  standard error between the  sample mean  and population mean,
              a  is  the  standard deviation  of  a single measurement,
              n  is  the  number  of trials  (12 here).
       For  subjects having  a  magnetic lung burden so great  that it exceeded
 the background  noise by more  than two orders of magnitude,  it was found that
 difference  in any  two  consecutive readings of the same point  averaged about
 9%.  This variability  is  a  function  of  how accurately the subject aligned the
 point on his  body  with the  detector  and on how hard he pressed against the
 plastic shield.  These uncertainties are  negligible compared  to the  ± 20%
 uncertainty which  we believe  was introduced  by using an average chest thickness
 correction  curve  (Fig. 3 (A) ) ;  they are  also  negligible compared to the  ± 30%
 uncertainty in  the I   value ,  which  depends  on the particle size and shape of
                    RS
 the Fe O  in  the lung.  However,  as  noted in the  text, this latter error is
 an absolute error, similar  for all subjects,  and  does not affect the relative
 amount of Fe  O. between subjects.

 Processing the  Magnetic Data

 Obtaining the Amount of Fe  O   in the Lung from Measurements of the Remanent
                                  Field
       In order to convert  gradiometer  readings into amount of magnetite in
the lungs, two  factors must be used.  The first is the remanent magnetic
moment of the Fe_O. particles  in the lung following magnetization in the
 coils.  It will be shown  that  this value  depends  on the grain size, aspect

                                     29

-------
ratio, and orientation of the magnetic particles.  The  second  is  a  factor
to account for the distance between the lungs and the detector (which
varies with the individual subject's chest thickness, lung thickness and
shape) and for the magnetic particle extended distribution within the
lungs.
1.  The Magnetic Moment
       The remanent magnetization of dispersed magnetite powders  has been
measured by Parry  (Parry, 1965) who obtained good agreement compared to  the
theory  (Stacey and Banerjee, 1974).  The theory uses
                              IBS-VN                                     (1)
       where I   is the remanent magnetization in emu/cc,
              Ro
             H  is the coercive force in oersteds, and
              c
             N is the effective self-demagnetizing factor for  the grains
                   (For spherical grains N=4ir/3).
Experimentally Parry found that for a 1% dispersion
                           I_  = 3.0xlO~3H                                  (2)
                            JxO             C
which means  (using a factor of 100) that the value of N is 3.32,  i.e., 0.79
times that of a spherical grain.  This means that some of the  grains were
elongated and aligned in the direction of the applied field.
       Another useful relationship is that   H  «
-------
 5 mg  into  2  cc  of  an  epoxy  (EZ Mount plastic, Donsel Equipment Co., Westboro,
 MA) in a clear  plastic  cubical box  2 cm  on  a  side.  Pairs of samples differed
 by the magnetic field environment in which  they were placed while the epoxy
 hardened;  unoriented  samples were placed in a magnetic shield, while oriented
 samples were placed in  a field of  ^ 700  gauss.
       We  then  measured the remanent field  of the samples using a fluxgate
 probe (MF-500,  manufactured by Automation  Industries) in the MIT shielded
 room.  A single probe was used in these  experiments (the second probe of
 this gradiometer was placed in a separate shield).

       The procedure before measuring each  sample was to first demagnetize
 it in a 60-Hz demagnetizer, then magnetize  it in a 750-gauss field for  30
 seconds.   The remanent  magnetization was calculated from the formula for the
 field from a magnetic dipole
                            2 Ips V   2 I   m
                        B  =    rc    =    RS                                   (4)
                              z3      z3p
       where m  is the mass of magnetite  in  grams,
             I   is the remanent magnetization in emu/cc (and is therefore of
              Ro
                  of different dimensions than the I used in the text),
             z  is the distance between the  center of the dipole and the
                  detector in cm,
             p  is the density of Fe O  in grams/cc,
             B  is the  axial magnetic field in gauss.
              Z
Typical values  for an unoriented sample  were m=4.8 mg, z = 12. 7 cm, p =  5.2
     "3            —6
am/cm.   B  =35x10    gauss; this gives I  =38.9 emu/cc or 7.5 emu/gm.  Since
         z                               R*-3
the saturation  magnetization of magnetite, I , is 450 emu/cc, I   /I  =0.086.
                                            t>                  Ro   S
       An  independent calculation of I   / I  was obtained by measuring  the
                                      i\O   O
 coercive force  H  .for an unoriented sample  in a PAR Vibrating Sample Magnetometer
                c
Model CC-1.  Here an  H  of 150 Oe was measured.  Using eqn.  (1) with N=4ir/3,
this gives I   /I  =0.080, in good  agreement with the direct measurement.
            Ro   o
       For the  samples  which were oriented  in the 700-gauss field while the
plastic hardened, the measured value of  1^ was 64.5 emu/cc, or 12.4 emu/gm,
*No longer manufactured by them.  The basic design is by the Forster Co. of
West Germany, and their units are available.
                                     31

-------
which is 70% greater than for the unoriented particles,  or I   /I  =0.143.
                                                            I\O   O
The increase is due to the decreased demagnetizing  factor  N for  particles
with their longer axes aligned in the direction of  the applied field.   Since
it is assumed that at least some of the Fe,O  particles  in the lungs  rotate
during the time the field is applied, a value of I   of 10  emu/gm (or  52  emu/cc)
                                                 rd>
is used in this study, since it lies between the unoriented and completely
aligned values.
2.  The Spatial Factor: The Extended Lung vs. the Dipole
       For the actual extended lung, we can assume  that  its z-dependence can
be approximated as that magnetic dipole (z  ) for distances greater than
20 cm between the center of the lung and the detector.   For distances  less
than 20 cm it is found that the B  for the human lung falls off  less  rapidly
      _3                         z
than z  .  This falloff distance factor, F (z) , is  plotted as the lower  curve
                                          m
in Fig. 3 (A) ;  (the sub-m refers to a magnetometer or single-probe measurement,
as opposed to a gradiometer, the latter designated  as sub-g; we  also  note that
in the tables F (z) is used as F (d) , as they are meant to be interchangeable
               g                g
here).  This curve is an average of two experimentally determined curves,
one for a tall and thin subject, the other for a short and stocky subject.
F  (z) is defined as
 m
                  F(z) -
                   m            '                r
       where M  is the total remanent magnetic moment.
       The net reading for the gradiometer used in this study is the difference
between two probes whose sensing elements were 8.6 cm apart.  Thus the
gradiometer distance factor, F (z) ,  is defined in terms of the magnetometer
distance factors, F (z) as
                   m
                      F (z) = F (z)  - F (z + 8.6) , and                         (6)
                       g       mm
                      Bg(z) = 2MrFg(z) = 2IRsmFg(z) /p                     (7)
where B (z) is the average gradiometer field reading in units of 10   gauss, or
                           B (z)  =1/6  E 6                                  (8)
                            g          t=0
where  Z_ 6 is the sum of the six gradiometer readings taken at the six  lung
points; these are three on the chest, three on the back, each extrapolated
back to time zero, which is just at the end of the magnetization in the
coils.  Combining egns.  (7) and  (8) and solving for m gives
                                     32

-------
              m (mg of Fe  O.) = [83.3 p E  6j 10 3 / [F (z) I   1              (9)
                          •3 1           t_0            g    RS
and since I   = 52  emu/cc  and p = 5.2 gm/cc, then
              m (mg of Fe  O ) = [8.33   E  6] 10~3 /F (Z)                   (10)
                          34           t=Q           g
or since the average lung  volume,  V.,  is  equal to FRC+ TV/2,
the density    m/V =[8.33 E  6]lO~3 / {F (z) [FRC + TV/2] }.                (11)
                    *•        t=o            g
References
Dunlop, D. J.  (1972).  Magnetite:  Behavior near the Single Domain Threshold.
   Science 176, 41.
Parry, L. G.  (1965).  Magnetic  Properties  of Dispersed Magnetite Powders.
   Phil. Mag.  1.1,  303.
Stacey, F. D., and  Banerjee,  S.  K.  (1974).   "The Physical Principles of Rock
   Magnetism."  Elsevier Scientific  Pub. Co.,  New York.
                                     33

-------
APPENDIX B.  SOME DETAILS OF RESULTS

       In this part of the Appendix two types of details are presented of
the results of measurements.  The first type consists of data of each of
the 115 individuals.  The second type consists of bar graphs which illus-
trate the distribution of quantities within the group.  The individual
data is given in the following Table BI.  We here explain the column head-
ings of that table.

SUBJECT NO.:  This is the subject number, only for purposes of this table.
   We are following the rule of preserving the subject's anonymity.
AGE:  This was the subject's age at the time of the study, in 1974.  For
   purposes of the lung function tests for which this group was chosen,
   the ages were confined to the ranges 25-35 and 41-50.  In addition,
   three individuals were measured magnetically whose ages are not shown
   (#3, 5, 97), but who are  over  50;  there  were  no dust indices
   and other data available for them, hence they were not included in the
   main correlations.
HEIGHT:  Included only as information on the body build of the subject.
WEIGHT:  Included only as information on the body build of the subject.
CHEST TH:  This is the chest thickness, used to calculate the distance from
   the lung center to the nearer probe, for application to Fig.  3(A)  of
   the text.  The distance used (d)  was (half the chest thickness)  +1.5 cm.
   Thickness is here given in inches (not uncommon in 1974J)
LUNG VOL(£):  This is the lung volume (in liters), used for calculating the
   density of Fe 0  in the lung (further on).  We calculated this volume by
   multiplying the 'FRC (measured by the McGill Group during the clinic
   visit)  by 1.08, which is an approximate way to include 0.5 of the tidal
   volume.
V.CAP:   The vital capacity, also measured by the McGill group during the
   clinic visit, expressed as % of predicted value;  because the subjects
   were chosen to be without pulmonary abnormality,  the range of values is
   restricted accordingly.
                                    34

-------
 CIGS/DAY:  The number of cigarettes smoked per day,  up to the time of the
    study.  Ex-smokers are indicated by ( );  in our correlations,  we included
    these with the smokers.  The values in this column were determined by care-
    ful questioning by the McGill group.
 YRS WELD.:   The years of welding exposure,  as defined in the text (METHODS).
    The designation < 0.2 indicates that,  as  far as we can determine,  the
    subject  had some welding exposure,  but certainly  less than 0.2 years.
    A bar under a value indicates that  the welding exposure was recent,  as
    opposed  to "old";  recent was defined  as  greater than 0.10 of his peak
    yearly exposure was obtained during the past year (ending in Aug.  '74).
 DUST:   This is the total dust  exposure index in units of (millions of par-
    ticles/cu ft)  x yrs.   As explained  in  the text, it is based on midget
    impinger counts,  etc.,  and  only applied up to the end of 1966.   There
    was  no inclusion for  dust exposure  after  that;  however,  for the period
    not  included,  the  dust  levels were  generally much lower than previously.
 I6(10~'g):   This  is the  sum of the  BZ  values measured over the six lung
    locations  on the torso;  it  is 6  x (the average  remanent  field  from the
    lung).   It  is  used  in the formula (see METHODS) to yield the amount  of
    Fe 0  .   A  designation such  as 40->75 indicates  that the  "true  value"
    can be anywhere  in  that  range,  i.e., the  error  is greater than  normal.
 CONV.(F  (d)):  This is the  geometric conversion factor involving  both the
    extended lung  and the gradiometer (differences),  as explained  in METHODS
    (both in the text and in App. A).
AMT.Fe-jO-:  After the  data  was  processed  and prepared,  it was  seen that
 (REVISED)
            F  (d) should have been  revised downward  by about 13%,  with
   variations  among individuals; hence, the  amt. Fe  0   should  have been
    revised  upward by a corresponding amount.   This new,  revised value was
   not used as data in the  text.   For  future use,  it is  a better value than
   the unrevised  value.
AMT.Fe-aO  :  This  is the unrevised value, actually used in the  text  for the
 (UNREV.)    measured amount of  Fe  0  in the  lungs of  a worker.
DENS.:  This  is the density of  Fe  0  in the  lung, calculated by dividing
   the revised amount  of Fe 0   by  the  lung volume.
                                    35

-------
MIDDLE/SIDE:  This is information re distribution of Fe 0  in the lung. It
   is the middle B , summed front and back, divided by the sum  (front and
                  z
   back) of the left side, and the right side.  It is otherwise written as
   2ZM/ZL+ZR.
LEFT/RIGHT:  This is also the distribution within the lung and is the two
   left points divided by the two right points, otherwise written as ZL/ZR.
APP.VISC.:  This is the apparent viscosity, defined as Z6 from the long
   pulse, divided by Z6 from the short pulse.  Numbers with ( )  indicate
   larger errors than usual.
RELAXATION:  This is relaxation group into which this subject fits.   A
   minus after the group refers to the lower range within that group, a
   plus refers to the higher range within that group, and no sign refers
   to the middle of the group.  It is a finer designation than just the
   course grouping given in the text RESULTS.  The (  )  refers to a larger
   error than usual, perhaps by a half-group; e.g. (Ill)  could mean III-
   or III+.
                                      36

-------
                                             TABLE BI.   DATA OF EACH SUBJECT
co
SUBJECT NO.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
W
a
49
49
-
49
-
28
43
49
44
50
43
50
47
49
47
48
50
50
46
49
31
48
28
HEIGHT (cm)
160
170
173
155
167
174
173
162
172
168
167
174
175
166
176
162
167
172
167
161
165
167
178
WEIGHT (kg)
71
75
71
57
64
63
71
64
83
71
75
89
75
90
88
57
61
70
60
58
76
62
84
"c
•H
E^
EH
U)
W
s
10.4
10.6
9.8
9.2
9.5
9.4
9.6
9.6
10.9
9.8
9.6
11.2
9.8
12.0
10.0
9.0
8.8
9.8
8.9
9.2
10.0
9.0
9.8
<••*
=><
I
O
,J
1.8
4.1
-
2.1
-
5.0
2.7
2.4
4.1
2.9
-
2.6
3.8
1.7
2.9
3.2
2.7
2.9
2.9
3.7
2.2
3.5
2.9
'M
a
Of
1
>
87
94
-
78
-
116
-
88
130
114
104
111
112
87
93
120
96
92
114
-
112
97
119
CIGS/DAY
0
20
20
20
20
0
20
50
(20)
0
0
20
0
(8)
(25)
22
20
20
0
0
0
22
0
•
3
£
2
>H
<0.1
<0.1_
0
0
0
<0.1
1-2.
0
0
<0._1
<0.2^
0
0
0
0
0
4. £
0
0
2.2_
0
3-2.
<-o.i
DUST (mpy/cf)
26
107
—
589
—
1
116
5
10
54
45
111
588
665
407
618
525
106
29
30
2
692
40
"5»
t^
o
•H
10
W
0->€5
665
1140
590
380
360
5610
40+75
480
2050
1450
0->-70
2710
285
125
145
2800
265
550
1300
145
1210
335
*••*
3
*"&>
&4_
8
1.63
1.56
1.80
1.97
1.88
1.93
1.85
1.85
1.50
1.80
1.85
1.43
1.81
1.23
1.74
2.05
2.10
1.80
2.07
1.97
1.74
2.05
1.80
AMT.Fe304(ing
(REVISED)
<0.4
3.6
5.3
2.5
1.7
1.6
25.3
0.2
2.7
9.5
6.5
<0.5
12.5
1.9
0.6
0.6
11.1
1.2
2.2
5.5
0.7
4.0
1.6
£
^r
o -~
ro .

•
OH
PH
<
._
2.8
1.8
2.2
1.9
_
1.3
__
2.6
_
2.3
_
1.9
2.4
(2.9)
1.8
1.8
1.55
-
2.6
-
1.6
2.5
RELAXATION

II-
11+
II
11+
11+
II

II-
1+
II
	
III
II
(III)
(1+)
II
(ID
II-
III-
(IV)
III-
(I)

-------
                                          TABLE BI.  DATA OF EACH  SUBJECT (CON'T.)
00
SUBJECT NO.
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
W
0
50
42
47
31
43
47
45
34
49
28
31
48
45
44
26
47
43
27
31
48
46
48
41
HEIGHT (cm)
163
164
179
173
178
165
173
171
187
171
189
186
174
167
181
163
179
169
172
177
177
161
172
WEIGHT (kg)
86
68
96
82
66
60
80
70
80
77
82
92
74
67
75
66
74
88
80
84
73
87
82
'H
•H
EH
EH
CO
W
5
9.8
9.6
11.4
10.0
9.0
9.0
10.2
10.2
10.6
9.4
9.5
11.4
10.5
9.2
10.0
10.5
9.2
10.5
10.5
11.0
9.4
11.2
10.6
o
2.1
2.4
1.6
3.2
4.1
2.9
2.7
3.0
5.2
3.7
4.0
3.2
2.6
4.4
4.5
2.7
3.8
3.0
3.0
2.4
2.8
2.8
2.7
a
df
I
83
104
110
109
103
-
116
98
89
120
112
133
98
106
109
92
109
115
114
101
102
118
91
CIGS/DAY
0
(10)
0
15
15
18
12
15
18
0
0
35
25
0
(10)
18
9
0
15
0
0
40
30
a
s
co
K
0
0
< 0.1
0
3.0
1.2
2.6_
2-1
0.5_
1.0_
<0.1
0
0
0
<0.1_
0.8
18-2
0
0._3
0.4
18.0
0.2
<0.1
DUST(mpy/cf)
187
9
442
1
123
117
124
20
17
1
3
22
125
214
23
579
10
2
21
507
73
133
18
0
rH
tO
0+65
0+70
8430
0+70
630
675
800
2570
1310
2140
85+95
0+105
140
350
1190
1250
12300
0+70
425
240
7020
460
35+80
•a
tn
8
1.80
1.85
1.37
1.74
2.05
2.05
1.66
1.66
1.56
1.93
1.88
1.37
1.60
1.97
1.74
1.60
1.97
1.60
1.60
1.46
1.93
1.43
1.56
AMT.F6304 (mg
(REVISED)
<0.2
<0.2
51.3
<0.4
2.6
2.7
4.0
12.9
7.0
9.2
0.4
<0.7
0.7
1.5
5.7
6.5
52.0
<0.4
2.2
1.4
30.3
2.7
0.2
o —
n •
0) >
<0.2
<0.2
45.0
<0.3
2.0
2.1
3.3
11.0
5.9
7.3
0.3
<0.6
0.6
1.2
4.6
5.4
40.0
<0.3
1.8
1.2
24.0
2.3
0.3
DENS. (ygm/cc
MIDDLE/SIDE
0.0 -
0.0 -
2.0 0.4
0.0 -
0.6 0.3
0.9 0.2
1.5 0.2
4.3 0.4
1.3 0.3
2.5 0.4
0.1 0.1
0.0 -
0.3 0.4
0.3 0.3
1.3 '0.3
2.4 0.4
14.0 0.2
0.0 -
0.7 0.5
0.6 0.3
11.0 0.2
1.0 0.3
0.8 -
FRONT/BACK
-
-
0.7
-
1.3
1.3
0.8
0.7
1.0
0.9
-
-
0.8
0.9
0.8
0.7
0.8
-
1.2
1.1
1.2
0.7
-
LEFT/RIGHT
-
-
0.9
-
1.0
0.7
1.2
0.7
1.1
1.0
1.0
-
0.9
0.6
1.0
1.4
0.7
-
0.8
0.6
1.1
0.8
1.0
u
co
H
1
-
-
2.9
-
1.5
-
1.9
2.0
2.0
2.5
(1.5)
-
-
1.9
1.7
2.8
1.3
-
1.3
(2.2)
2.6
1.4
-
RELAXATION
-
-
Ill
-
II
II-
III-
I
II-
11+
-
-
-
II
11+
III-
III+
-
II-
(III)
nn-
iit
-

-------
                                        TABLE BI.   DATA OF  EACH  SUBJECT (CON'T.)
w
SUBJECT NO.
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
w
u
<
41
44
50
50
49
46
47
44
50
27
46
45
43
43
46
43
46
49
43
46
44
48
50
HEIGHT (cm)
172
175
173
174
155
173
166
175
175
180
178
167
172
171
168
172
170
176
178
162
159
163
177
WEIGHT (kg)
70
73
89
89
49
75
60
78
77
94
74
70
67
77
66
73
65
61
63
79
64
89
87
*-x
e
•rH
EH
EH
CO
W
5
9.6
10.2
10.5
11.4
8.0
9.5
9.5
10.5
11.2
10.8
9.0
11.2
9.6
10.2
9.0
9.5
9.0
9.5
8.2
10.0
10.0
12.0
11.2
^^
o*
I
CJ
^
3.8
4.8
1.9
-
2.2
2.8
3.3
3.5
3.2
3.0
3.2
2.5
3.6
2.6
2.6
2.4
2.9
4.1
3.9
2.5
2.6
3.0
2.7
V.CAP(%pr)
127
95
115
-
51
103
127
80
113
104
85
101
144
105
87
99
88
96
94
106
98
-
80
CIGS/DAY
15
25
0
(50)
(12)
20
15
0
0
0
20
0
0
(50)
38
20
0
25
55
0
(25)
12
25
•
Q
^
3
>•
15. £
<0.1
2-i
0
0
0
0
0
0
1.3
0
0
0
<0.1
0
1.2
0
0
0
1.5_
0
0.8
0.8_
DUST (mpy/cf
11
9
30
835
534
21
11
227
72
32
116
7
7
57
110
13
154
641
8
523
308
119
644
f~*.
Cn
r^
0
•H
10
W
9530
170
455
1160
525
560
0+70
570
450
630
130
35+65
0+70
150
285
150
230
1140
245
2900
1400
430
1200
^^*
s
r01
CM
•
O
U
1.85
1.66
1.60
1.37
2.35
1.88
1.88
1.60
1.43
1.51
2.05
1.43
1.85
1.66
2.05
1.88
2.05
1.88
2.30
1.74
1.74
1.23
1.43
AMT.Fe304(ir
(REVISED)
42.9
0.8
2.4
7.0
1.9
2.5
<0.2
3.0
2.6
3.5
0.5
0.2
<0.2
0.8
1.2
0.7
0.9
5.0
0.9
13.9
6.7
2.9
7.0
>:r
0
n ^
0) .
frj £>
s* B
34.0
0.7
2.0
6.2
1.3
2.0
<0.2
2.5
2.3
2.9
0.4
0.2
<0.2
0.6
0.9
0.5
0.7
4.0
0.6
11.0
5.4
2.6
6.1
o
1
•
w
z
3
11.0
0.2
1.3
-
0.9
0.9
0.0
0.9
0.8
1.2
0.2
0.1
0.0
0.3
0.5
0.3
0.3
1.2
0.2
5.6
2.6
1.0
2.6
MIDDLE/SIDE
0.2
0.3
0.5
0.3
0.3
0.5
-
0.4
0.4
0.3
0.2
-
-
0.5
0.4
0.4
0.4
0.5
0.2
0.3
0.4
0.3
0.3
FRONT/BACK
1.4
1.1
1.1
-
1.1
0.8
-
0.8
0.9
0.8
1.2
-
-
0.9
1.1
0.9
0.8
1.1
1.0
1.1
0.8
0.6
0.6
LEFT/RIGHT
1.0
0.7
1.1
1.1
0.6
1.1
-
0.8
0.8
0.9
0.9
-
-
1.2
1.1
1.0
0.9
1.0
1.4
1.6
0.8
0.9
0.7
•
O
C/3
H
>
•
ft
8!
2.8
(1.5)
-
1.6
2.1
2.4
-
-
2.0
2.3
-
-
-
(1.6)
1.1
-
-
2.7
(4.0)
2.1
-
(2.7)
2.6
RELAXATION
II
-
(Ill)
III
III
(II)
-
II
II
1+
-
-
-
-
Ill
-
-
II
1+
III
11+
11+
11+

-------
TABLE BI.  DATA OF EACH  SUBJECT (CON'T.)
SUBJECT NO.
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
w
34
48
43
28
48
45
42
31
46
49
41
49
29
35
43
27
45
26
32
27
41
31
46
HEIGHT (cm)
165
170
181
176
168
164
174
178
161
,161
175
175
178
173
172
164
166
172
169
176
164
176
175
WEIGHT (kg)
53
61
75
77
66
71
69
60
89
73
69
63
72
75
66
75
85
65
56
56
54
113
84
•H
CO
W
5
8.0
10.0
9.2
10.0
9.5
9.0
9.4
8.4
10.9
11.0
10.2
9.4
9.0
10.6
10.2
10.0
11.5
9.0
8.0
8.0
8.4
12.0
11.0
1
2.9
3.6
3.7
3.8
3.0
1.9
3.0
4.0
1.5
1.5
2.6
3.2
2.9
1.9
5.2
1.7
4.0
5.0
3.9
2.9
2.6
1.9
3.3
of
§
88
92
91
117
84
95
126
104
65
88
108
110
104
105
140
97
129
119
84
89
81
89
99
CIGS/DAY
18
30
25
10
25
18
0
18
40
0
0
25
20
20
0
15
50
0
0
20
0
25
28
9
1
0
0
5.3
0.4
0
1.9
0.4
0
0.6_
0
<0.1
0
1.2
0.2
0
0.5_
0
0.5
0
0
29. £
0.2
0
DUST(mpy/cf)
1
540
8
0
20
3
20
0
595
241
89
10
0
58
228
20
126
41
26
0
45
1
129
o
•H
325
20+75
1230
370
225
135+150
2390
45+70
1490
170
200
65+75
430
910
1910
785
155
225
0+70
820
51800
230
1490
2
Cn
8
2.35
1.74
1.97
1.74
1.88
2.05
1.90
2.22
1.50
1.46
1.66
1.90
2.05
1.56
1.66
1.74
1.35
2.05
2.35
2.35
2.22
1.23
1.46
O Q
row
01 CO
(J-, HI
1.2
0.1
5.2
1.8
1.0
0.5
10.5
0.2
8.3
1.0
1.0
0.3
1.7
4.8
9.6
3.8
1.0
o.e
<0.2
2.9
194.0
1.6
8.5
o — .
ro .
0) >
PM W
0.8
0.2
4.0
1.4
0.8
0.4
8.2
0.1
7.0
0.8
0.8
0.2
1.4
4.1
7.9
3.1
0.9
0.7
<0.2
2.1
140.0
1.4
7.3
DENS. (ygm/cc
0.4
0.0
1.4
0.5
0.3
0.3
3.5
0.1
5.5
0.7
0.4
0.1
0.6
2.5
1.8
2.2
0.3
0.2
0.0
1.0
75.0
0.8
2.6
MIDDLE/SIDE
0.3
-
0.3
0.4
0.3
0.3
0.3
-
0.3
0.4
0.4
-
0.2
0.4
0.2
0.3
0.3
0.4
-
-
0.5
0.4
0.2
FRONT/BACK
0.8
-
0.7
0.8
0.6
-
1.1
-
0.6
0.5
0.6
-
0.9
1.2
0.9
0.9
1.2
1.2
-
0.6
0.9
0.5
0.6
LEFT/RIGHT
0.9
-
0.9
0.9
0.9
-
1.2
1.0
1.7
0.7
0.7
0.8
0.9
1.3
1.1
1.1
0.7
0.8
1.0
0.6
1.2
0.7
0.9
o
CO
H
§
-
-
1.7
1.7
-
(2.2)
2.0
-
1.8
-
(2.1)
-
2.1
1.7
3.1
1.9
-
-
-
1.7
3.0
(1.8)
1.7
RELAXATION
(ID
-
11+
(1 + )
-
-
11+
-
11+
(III)
(11+)
-
II-
11+
11+
II-
-
11+
-
II
III+
-
II

-------
TABLE BI.   DATA OF EACH  SUBJECT (CON'T.)
•
0
z
E-i
U
W
5 w
5 u
in <
93 31
94 29
95 27
96 28
97 -
98 34
99 43
100 34
101 29
102 27
103 49
104 29
105 27
106 45
107 49
108 47
109 26
110 48
111 29
112 49
113 49
114 49
115 41
1
EH"
EC
U
H
3
i/y
178
174
173
175
180
171
178
172
180
169
171
175
165
166
162
172
183
173
174
175
173
167
^ *c
O> -H
* H
^^ r^
g H
u (/>
M W
g 5
75 9.5
68 9.0
107 12.0
96 11.0
87 11.5
84 10.6
69 10.2
100 11.6
60 10.0
65 8.9
78 10.2
91 10.5
92 11.2
69 10.2
85 11.0
67 10.2
61 9.6
83 10.2
83 10.0
83 11.0
74 9.5
175 14.2
68 8.9
5
.j
n
>
CJ
z
ED
^
2.9
3.7
2.8
2.2
-
3.3
4.4
2.8
3.0
4.5
2.7
2.3
2.1
2.6
1.9
2.9
3.8
3.1
3.9
2.7
3.9
-
2.6
*•"•*
k
ft ><
<#> <
* Q
ft \
< w
u e>
H
> u
125 0
111 25
104 23
95 0
28
99 255
144 10
106 4
117 0
103 10
90 0
77 (20)
91 0
20
100 0 •
90 20
119 0
97 20
117 0
97 (25)
92 25
75 0 <
87 18
•
p
iJ

10
«
><
<0.1
!._!
0
<0._1
1.0
<0.1
2.8
0.2
0
0.2_
0.8
<-o.i
0
0
<0.1
°-l
0
2.0
0
5.£
0
:°-I
0
^•^
IM
U
£
JE
f-t
u1
Cfl
3
O
2
0
0
23
-
30
19
1
22
1
267
32
2
137
378
113
2
822
1
125
596
202
7
•^*»
o^
i^
i
o
i— H
vO
W
125
760
0->80
0+150
325
0+70
1600
230
85+110
1120
1070
55+75
0+70
210
1050
1040
0+70
230
240
2500
190
0+245
150+160
V
•»_•
&
CL,
*
8
1.88
2.05
1.23
1.46
1.35
1.56
1.66
1.33
1.74
2.09
1.66
1.60
1.43
1.66
1.46
1.66
1.85
1.66
1.74
1.46
1.88
0.84
2.09
£
*r *-*
0 Q
ro W
0) W
t. H
EH fj
0.6
3.1
0
<0.9
2.0
<0.4
8.0
1.4
0.4
4.5
5.4
0.3
<0.5
1.0
6.0
5.2
<0.3
1.2
1.1
14.3
0.8
<2.8
0.6
I

h i
EH 1
0.4
2.4
<0.5
<0.8
1.8
<0.3
6.6
1.3
0.3
3.4
4.4
0.2
<0.4
0.9
5.1
4.3
<0.2
0.9
0.9
12.0
0.7
<2.4
0.5
u
CJ
^L
•
C/3
2
W
Q
0.2
0.8
0.0
0.0
-
0.0
1.8
0.5
0.1
1.0
2.0
0.1
0.0
0.4
3.2
1.8
0.0
0.4
0.3
5.3
0.2
-
0.2
w
Q
H
\
s
Q
Q
H
S
0.4
0.3
0.9
-
0.4
-
0.3
0.3
0.4
0.3
0.3
-
-
0.4
0.2
0.4
-
0.3
0.4
0.4
0.4
-
0.4
:*;
u
S
§
§
K
Cn
0.6
0.8
-
_
0.8
-
1.0
0.6
-
0.9
0.7
-
_
0.6
0.8
1.0
-
0.9
0.7
0.8
1.3
-
1.1
EH
X
u
H
«
E>
&
^
0.9
0.7
-
_
0.9
_
0.8
1.1
_
1.2
0.4
_
_
1.0
1.0
0.9
-
0.6
0.9
1.0
0.9
-
1.0
U
to
H
•>
*-*^
ft
ft
<=c
_
1.8
(1.3)

1.0

2.8
1.9
_
2.0
1.9
_

(3.9)
2.9
2.0
_
1.6
-
2.2
1.6
_
- 1
§
H
EH
3
r%
3
§

II-


IV

I
(ID

1+
III+


__
IV-
II-
_
_
(II-)
11+
11+
_
:IID

-------
The  following six  figures are  visual  displays of  the distribution
data presented in  Table  II in  the text.   Our purpose is  to clarify
these quantities.
          o
          C/)
          Ql
          UJ
          O_
          cn
          UJ
          m
            30
            20
             10
             0
                   0.2     0.4     0.6     0.8

                          MIDDLE/SIDE
 1.0
 1.2
 Fig.  Bl.  Distribution of the ratio of the remanent field at the middle of the
 torso  to  that on the average of both sides, for all workers who had enough  Fe3O4
 to yield  a ratio.  The sum of  front and back is  used for right, middle and left.
             30 r
z
o
0? 20
UJ
CL
u_
o
or in
LU IU
m
^
r>
z
0

-
















































1 	 1 	 1 	 1 	
                  0.4     0.6     0.8     1.0

                              FRONT/BACK
1.2
1.4
1.6
1.8
 Fig. B2.  Distribution of the ratio of the remanent  field at the front of the
 torso to that at the back.  The sum of the right,  middle, and left is used.
 The  spread is quite large here because of the coarse z-spacing used, front and
 back.  The spread would presumably narrow if a constant z-spacing was used for
 all  subjects, as suggested by the correlation in Table III in the text.
                                     42

-------
PERSONS
OJ j
o c
fe 20
o:
LJ
CD
^
i 10
n
















1
                       1       JI      _
                       RELAXATION GROUP
Fig.  B3.   Distribution of the relaxation  rate  for  all workers who had
enough Fe304 to yield a crude relaxation  curve.  The groups correspond to
those of  Fig. 3 in the text.
d(J
en
en
or
Q_
t in
f~j i \s
DC.
LU
CD
Z 5
O




_



-






































                  1.0     1.4     1.8    2.2     2.6     3.0

                       APPARENT VISCOSITY   ( L/S )
3.4
Fig. B4.  Distribution of the apparent viscosity for all workers where
there was enough Fe3O4 to yield the  ratio of remanent field from the  long
pulse to that of the short pulse.
                                     43

-------
NUMBER OF PERSONS
_ f\3 OJ
3 O O 0

/











           0     3.2      10     32      100     320   1000

                 DUST EXPOSURE INDEX (LOG SCALE)
 Fig. B5.  Distribution of total dust exposure index.   Welders  are  included
 here.   (Three workers were measured who had no indices.)   A log  scale is
 used, except at the first interval, which goes down to zero.   The  workers in
 this interval, mostly office workers, therefore have  relatively  low indices
 (-0).
         20 r
CO
0 15
CO
QC.
LU
CL
0 10
cr
UJ
CO
ID
^ 5

0

-




-





>0
/



































































"N -^















                 0.10    0.32    1.0     3.2     10

                 YEARS OF WELDING (LOG  SCALE)
32
Fig.  B6.   Distribution of years welding for all those with any welding ex-
posure whatever.   A log  scale is used, except for the first interval which
extends down to but does not include zero.   This interval contains many
workers who are borderline between welders  and non-welders.
                                     44

-------
                            Sample C. Technical Report Data Sheet, EPA Form 2220-1
                                           TECHNICAL REPORT DATA
                                   (Please read Instructions on the reverse before completing)
 . REPORT NO.
   560/6-81-005
                                                                       3. RECIPIENT'S ACCESSION NO.
4. TITLE AND SUBTITLE
                       MAGNETIC  LUNG
   MEASUREMENTS IN RELATION TO OCCUPATIONAL EXPOSURE TO
   ASBESTOS MINERS AND MILLERS OF QUEBEC
                                                                      5. REPORT DATE
                                                                          January, 1981
                                                                      6. PERFORMING ORGANIZATION CODE
 '. AUTHOFUS)
                David Cohen *
                                  Thomas S. Crowther **
(Salmi W. Gibbs *       Margaret R. Bccklaka,1
                                                                         8. PERFORMING ORGANIZATION REPORT NO
9. PERFORMING ORGANIZATION NAME ANO AOORESS
   American Public Health Association
   1015 Fifteenth Street, N.W.
   Washington, D.C.  20005
                                                                         10. PROGRAM ELEMENT NO.
                                                                       1. CONTRACT/GRANT NO.
                                                                         68-01-3859
12. SPONSORING AGENCY NAME ANO AOORESS
   Office of Pesticides and Toxic Substances
   U.S. Environmental Protection Agency
   401 M Street, S.W.
   Washington, DC  20460
                                                                       3. TYPE OF REPORT ANO PERIOD COVERED
                                                                         Final Report
                                                                       4. SPONSORING AGENCY CODE
 s. SUPPLEMENTARY NOTES      Massachusetts Institute of Technology *  ^Ctelanese
   U.S. Department of Labor --- OSHA **       OteGill University 1
 6. ABSTRACT
       Feo04 particles (magnetic) are often attached to asbestos fibers (non-magnetic) in the primary asbestos
   industries; therefore, a measurement of Fe.j04 could help determine the amount of asbestos in the lungs of
   workers  in these industries. As a first  assessment of this method of determining retained dust, magnetic
   measurements were made of the  amount of Fe-j04 in the lungs of  115 miners and millers of chrysotile
   asbestos.  The performance  of these measurements at  an  industrial site was found to be feasible and
   practical. A relatively large amount of Fe-j04 was seen in the lungs of those with welding experience, which
   masked the  Fe-^O^ contributed by asbestos, therefore this group was considered separately. For  the
   remainder (non-welders), the amount of Fe304  was plotted against a total dust exposure index which was
   available for  each individual. The correlation between these quantities was not high, but was  statistically
   significant at the 0.01 level. For the  non-smokers within that group, the correlation was higher  and  the
   amount  of  Fe304 was  relatively greater. These  results suggest that the magnetic measurement of a
   chrysotile miner and miller  reflects, at  least to some extent, the amount of asbestos in his lung; the scatter
   could be due to individual  differences in deposition and clearance, to which this measurement should be
   sensitive. These  results are  also consistent with the possibility that  less dust is deposited or retained in
   smokers than in non-smokers.
 7.
                                       KEY WORDS ANO DOCUMENT ANALYSIS
                      DESCRIPTORS
                                                          .IDENTIFIERS/OPEN ENDED TERMS
                                                                                           .  COSATI Field/Group
   Asbestos
   Occupational Exposure
   Magnetic Fe304
                                                       Amount of Fe3O4 in Lung
                                                       Magnetic Dust Study
                                                       Univariate Analysis
 IB. DISTRIBUTION STATEMEN1
    Unlimited
                                                       19. SECURITY CLASS (This Report/'
                                                        Unclassified
21. NO. OF PAGES
                                                         2O. SECURITY CLASS (This page I
                                                                                            AA.
                                                                                           22. PRICE
                                                                                                                r
EPA Form 2220-1 (9-73)
                                                       45

-------