blood near a - journal of epidemiology and community health | … · analyses were carried out at...

British Journal ofPreventive and Social Medicine, 1977, 31, 154-163

Lead in human blood and in the environment neara battery factory

W. J. ELWOOD1, B. E. CLAYTON2, R. A. COX3, H. T. DELVES2,E. KING4, D. MALCOLM5, J. M. RATCLIFFE6, AND J. F. TAYLORs

From lBolton Area Health Authority (formerly MOH Swinton & Pendlebury MB), 2Hospital for SickChildren, Great Ormond Street, 3Cremer and Warner, London (formerly Imperial College),4 NationalOccupational Hygiene Service Ltd, Manchester, 5Chloride Group Ltd, London, and 6Mathematics Department,Imperial College

SUMMARY Samples of blood, air, dust, soil, vegetation, and tap water were examined between1973 and 1975 to determine whether a large battery factory (with a smelter) was contributing tolead in the environment and to lead absorption by the local population. Mean blood lead levelsin the children of lead workers were about 6 pg/100 ml higher (P <0.001) than in otherwise com-parable children. Capillary blood samples in wives of lead workers were 1.7 ,tg/100 ml higher(P <0.05) than those of otherwise comparable wives, but venous blood samples from the same

subjects showed no significant difference. Lead in dust, soil and vegetation, although variable,decreased in concentration with distance from the factory. This relationship with distance fromthe factory was not however found in blood lead levels. No consistent effect of distance was foundwith lead in air, but significantly higher concentrations were recorded at downwind than upwindsites. The blood lead results have been analysed to assess the influence of domestic factors ofpossible relevance-such as, lead pipes, car ownership, age of house, etc. The presence of a lead-worker in the household appears to outweigh these other factors. The findings are consistentwith the work of Burrows (1976) who showed that lead workers take lead home. An interlaboratorycomparison on lead in blood was carried out. The two laboratories concerned were found to beequally consistent, but there were differences between them and the design of this comparisondid not make it possible to say that the results of either were 'absolute' values.

Anxiety about possible harm to health from leadhas been evident in recent years. Lead may enterthe body from sources such as food and drink,water, or air (Department of the Environment,1974; Ministry of Agriculture Fisheries and Food,1972, 1975). An industrial works using lead is anobvious source of emission of lead, which is liableto attract public attention (Godber, 1971; Depart-ment of the Environment and the Welsh Office,1973).The battery factory studied is situated at Clifton

Junction near Manchester and is in an industrialzone with residential areas abutting on the industrialzone but not adjacent to the factory itself. Theprocesses carried out include lead smelting andvarious other processes which give rise to leadoxide containing dusts which for reasons of industrialhygiene must be removed from workplaces by theventilation system and are discharged at low level

to the air (some after filtration). Dust may beblown about by the wind, carried on lorries orpersons, or on shoes or clothing of workers.

Design of survey

The survey was designed to show whether or notthe blood lead levels of people living near thefactory or of close relatives of those employed in itwere significantly higher than those of a suitablecontrol group. A further objective was to study thedistribution of lead from the factory up to a distanceof two miles. A similar survey was carried out atthe same time for another battery factory at Dagen-ham in the London borough of Barking, but as thiswas much smaller in scale than the Clifton Junctionstudy it is not described in this paper.

Six groups of test and control subjects weredefined as follows:

154

Protected by copyright.

on June 29, 2020 by guest.http://jech.bm

j.com/

Br J P

rev Soc M

ed: first published as 10.1136/jech.31.3.154 on 1 Septem

ber 1977. Dow

nloaded from

http://jech.bmj.com/

Lead in human blood and in the environment near a battery factory

1. Leadworkers families (LF) childrenChildren under five years of age of battery factoryleadworkers, wherever they lived, whose nameswere on the register maintained in the medicaldepartment of the factory in December 1972;

2. Births register (BR) children (near)A random sample of children under five years whowere not the children of a leadworker or batteryfactory employee and who had lived within one mile(1 '6 km) of the factory all their lives;

3. Births register (BR) children (far)A random sample of children under five years whowere not the children of a leadworker or batteryfactory employee and who had always lived morethan one mile (1 * 6 km) from the factory (butwithin the local authority areas indicated below);

4, 5, 6. The mothers of subjects in groups 1, 2,and 3 above; these were named LF mothers, BRmothers (near), and BR mothers (far), respectively.NoTE: children selected in groups 1, 2, and 3 arereferred to below as 'index' children.

Method

SAMPLING

Survey subjects and controlsLF children were identified by a census of employeeswhich asked for names and dates of birth of allchildren under five years of age. BR children weresampled from the registers of births in three con-tiguous former local authority areas-Swinton andPendlebury MB, Prestwich MB, and WhitefieldUD (total population 1971 census 95 327). Fromthese registers for each year of birth children whohad died, children whose families were known tohealth visitors to have changed address during theprevious five years, and children whose parents'employment involved the use of lead or its com-pounds were eliminated. The remaining childrenwere divided into those resident within, and thoseresident more than one mile (1 '6 km) from theworks: from each of these sets a random samplewas drawn.Each mother was invited to attend a local clinic

for blood sampling with all of her children underfive, both by introductory letter and by a personalvisit from a public health inspector or health visitor.Transport by taxi was provided for her. Those whofailed to attend were visited or telephoned and everyeffort made to secure attendance; in a few cases therequired blood was taken in the home.

Blood specimensFrom each family the following samples were

taken: 5 ml of venous blood and 0 *2 ml of capillaryblood by finger prick from the mother; 0 '2 ml ofcapillary blood by finger or heel prick from theindex child, and from each of his siblings underfive years of age.

In addition, microaliquots from a proportion ofthe venous samples were put in capillary containers.A proportion of all of the above were replicated inorder to check the reproducibility of the results.All blood specimens were taken with the stringentprecautions subsequently published by the Depart-ment of Health and Social Security (1976).The bloods from the test and control groups were

examined contemporaneously so that each batch ofspecimens sent to the laboratories contained a cross-section of specimens from the different groupswithout means of identification by which the labora-tories could tell to which group any individualsample belonged.For each blood sample a record sheet was pre-

pared giving details of the subject (age, sex, relation-ship to leadworker, and population sample group),details of his household (social class, geographicallocation, age and rateable value of house, use ofvacuum cleaners and cars, public health inspector'sassessment of such hazards as lead pipes, flakingpaint, and traffic fumes) and details about thesample itself (capillary, venous, first specimen,replicate, repeat, laboratory used, and result).The record sheets were cross-referenced so thatdifferent samples from the same subject could beidentified and individual subjects related to theirrespective families.

If the blood lead level was 'high' (which was takeninitially as over 60 pg/100 ml, but later revised toover 50 pg/100 ml) the survey organiser arrangedfor a repeat blood sample. If this was also high thematter was reported to one of us (WJE or JFIT) whodiscussed the finding with the family doctor andarranged referral to a paediatrician in the case of achild. If it was satisfactory, a report was sent to thefamily doctor indicating the range within which theresult fell and a letter of thanks and reassurancewas sent to the mother. In four cases (includingboth BR and LF children) the first test was over100 tig/100 ml but the subsequent test was acceptable(38, 33, 18, and 18 ,ug/100 ml), and these cases weretherefore excluded from the statistical analysis.In 17 other cases where a repeat was taken the firstresult alone was used for analysis whether or notthe second result confirmed the first. This was doneto avoid any bias.

AirSome 265 daily (24 hour average) air samples wereobtained from six fixed sites near the Clifton

155



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from


W. J. Elwood et al.

works during the period April to May 1974. Thesampling method was by analysis of the particulatefilters on standard UK National Survey of AirPollution Instruments. These samplers draw about2 cubic metres of air each day through a WhatmanNo. 1 filter paper. Background levels of lead inunexposed Whatman No. 1 paper yielded an averagevalue of 0 002 pg/m2, which corresponds to lessthan 1 % of the values found after exposure. Errorsowing to blank values have therefore been ignored.DustA soft new paintbrush was used to brush surfacedust on to clean paper from which it was transferredinto an already numbered container. Samples weretaken from the roadside gutter and from thepavement along the main transport route from thefactory gate at the distances shown in Fig. 1;similar samples were taken from other roadscarrying a comparable volume of traffic but notused by factory vehicles; samples were also takenfrom selected reference points near the factory.

SoilFour cores of soil 15 cm in depth x 2 5 cm indiameter were taken at each corner of a 1 metresquare from undisturbed grassed soil at each of theselected points (as for dust). Cylinders 2-5 cm longfrom top and bottom of cores constituted thespecimens of 'surface' and of 'deep soil' respectively.VegetationThe study of vegetation has been described byRatcliffe (1975).

Tap waterA random sample of 20 households was selectedfrom those households in which blood had beentaken and 47 specimens of tap water were obtainedfrom 18 of these homes; the first run-off in earlymorning from cold and hot taps were sampledseparately.

ANALYTICAL PROCEDURES

BloodAnalyses were carried out at two separate labora-tories which will be referred to here as Lab A andLab B. Lab A used anodic stripping voltametry,checking by dithizone and by atomic absorptionwith flameless atomisation of paper discs on aproportion of the samples. Lab B used the DelvesCup method with atomic absorption spectro-photometry (Delves, 1970). An interlaboratorycomparison was carried out by comparing resultsby Lab B for microaliquots taken from adultvenous blood submitted to Lab A.

Environmental specimensApart from the specimens of vegetation which wereanalysed by one of us (JMR), all the environmentalspecimens were examined at Lab A. For air andwater samples the monocolour dithizone methodwas used. For dust (which was sieved through a 200mesh sieve) and soil samples polarographic methodswere used coupled with confirmation of absence oftin by atomic absorption spectroscopy.

STATISTICAL METHODS

BloodWe found no significant difference between themean blood lead levels of the index children (whetherLF or BR) and their siblings, and moreover therewas no correlation between lead levels and size offamily so it was decided for the final presentationto merge these groups within each year of birthsince this would increase the number of subjectswithout introducing a bias. The results presentedare arithmetic means and standard deviations;log transformation of all the original data beforeanalysis did not alter any of our conclusions.The 'within laboratories' and 'between laboratories'variances were examined, and the linear regressionof one laboratory's results on the other's wascalculated, using adult venous blood and thecorresponding microaliquots.

Environmental specimensAll the values for soil and dust from ledges quotedin this paper are the means of four specimenstaken from the same place at the same time. Ateach point selected for dust sampling on transportroutes means of two specimens from the gutter andtwo from the pavement are given.

Associations between blood levels and environmentalfactorsSimple bivariate regression analyses were madetaking blood lead as 'dependent' variable andconsidering in turn the following factors: age,length of residence, age of house, rateable valueof house, distance and direction from the works,length of time during which there had been a lead-worker in the family. Children and adults wereconsidered separately in most cases, but in thecorrelation with age they were also taken together,using the capillary samples from the adults.

ResultsPOPULATION SAMPLE CHARACTERISTICS ANDRESPONSE RATESTable 1 sets out relevant information. The numberof 'index children included in results' was less than

156



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from



Table 1 Population samples and response rates

LF BR (Births register)Population (Leadworkers'

families) Near Far(< 1 mile) (> I mile)

Leadworkers on register in December 1972 1387 N/A N/AResponders to census 905 N/A N/AResponse rate (°.) 65 N/A N/A

Samples of index children drawn 158 165 195Index children who attended 152 88 129Response rate (Y.) 96 53 66

Included in results (children)Index children 147 71 122Siblings 45 32 48Total 192 103 170

Included in results (mothers)Social class I 0 10

II 0 25III 0 100IV 146 47V 0 2

Not specified 1 9

Table 2 Children's capillary blood-blood lead levels of leadworkers'families compared with children selected frombirths register (pug lead/100 ml blood-Lab B)

LF children BR childrenAge Difference Significance(years) No. Mean SD No. Mean SD of means kvel

< 1 10 251 11 0 18 21-8 11-8 3-3 NS1-2 20 33'S 11 1 38 28-9 7-6 4-6 NS2-3 73 33-4 91 107 28-0 8-1 54 P<00013-4 63 34-8 7'S 70 24-7 6-3 10.1 P < 0 0014-5 26 31-0 6-0 40 28-6 7-3 2-4 NSAll 192 33-1 8-6 273 27-0 7-8 6-1 P < 0'001

the 'index children who attended' for technicalreasons such as an unsatisfactory sample whichwas not repeated, or because (in the case of 12 BRchildren) a previously unknown occupationalassociation with lead did not come to light untilthe clinic attendance. The mis-match in socialclass was unavoidable because of the samplingmethods used. No significant difference was,however, found between the mean blood lead levelof BR children in each social class and that of allBR children, or between the corresponding meansfor BR adults.

BLOOD LEADTable 2 shows the findings for children's blood.At each year of age the blood lead in LF childrenis higher than that in BR children. For the totalgroup this difference is statistically significant andit is also significant for the 2 and 3-year-old groups.It appears that three years is the most critical age,for the LF children have maximum blood leadlevels then and at that age the difference betweenthem and the BR children is greatest. If childrenof all ages in social class IV are taken alone, themeans are 33'0 for LF children and 26-8 for BRchildren; the difference of 6'2 jg/100 ml is significant(P >0'01). The mean blood level for the 103 BRchildren living within one mile of the factory was

27*3 pg/lOO ml whereas the mean for the 170children living over one mile away was 26'7 pg/100ml, a difference which was not statistically significant(P >0 05)The results for the 28 children under 12 months of

age were examined. There was no child under threemonths old. No child between three and five monthsof age had a blood lead level over 30 gg/100 ml;three children between five and 12 months (1 LF and2 BR) had blood lead levels between 40 and 60Ag/100 ml.The mean lead level in capillary blood examined

by one laboratory from the 147 LF mothers was23-6 pg/100 ml whereas that from 193 BR motherswas 21'9 jig/100 ml, a difference which althoughsmall was statistically significant (P <0 05). Therewas however no statistical significance in thevenous blood of these two groups examined bythe other laboratory. Neither laboratory found asignificant difference between the BR (near) andBR (far) groups of mothers.The blood lead level for the BR children was

analysed by distance and direction of their homesfrom the factory (Table 3). No significant differencewas found between the mean blood lead of childrenliving in different sectors or at any distance andthe mean of the whole group-except that it wassignificantly higher (P <0'05) for the 53 children

157



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from


Table 3 Children's capillary blood-blood lead levels ofBR children living at various distances and in various directionsfrom the factory (4ug lead/100 ml blood-Lab B)

Direction SignificantDistance (miles) difference

N NE E SE S SW W NW All from mean

<* 0 0 0 0 1 1 2 0 434 0 27 5 27 5 29 1 NS

I-I 0 0 0 0 13 14 13 0 4033 4 24-6 27 9 28 6 NS

1-1 0 0 18 0 1 18 22 0 5927 1 20 0 27-3 25-2 26-3 NS

1-2 6 12 28 3 22 37 11 0 11925 3 26-3 29-0 27-7 23-7 24-9 28-0 26 1 NS

2-3 4 25 7 0 0 12 0 0 4824 0 26 2 36 1 30-2 28 5 NS

>3 0 3 0 0 0 0 0 0 321-7 21-7 NS

All 10 40 53 3 37 82 48 0 27324-8 25 9 29 3 27 7 27-3 26 2 26-7 26-9 -

Significant difference P <from mean NS NS 0 05 NS NS NS NS - -

Number of children.Mean blood level (in italics if 10 or more subjects in group).

Table 4 Correlation of various factors wiih group nwans

Difference inFactor mean blood lead Significance Comments on blood lead levels

(pg/100 ml) levelLab B

Leadworker 6 1 P<0*001 See Table 2. 192 LF children higher than 273 BR children.in family (Note: Significant difference also with adults' capillary blood)Lead pipes 4-2 P<0 005 40 BR children in houses with lead pipes higher than 85 BR childrenin house in houses on which Environmental Health Officer reported 'No lead

hazard'.(Note: No significant difference with adults or with LF children)

Age of 3-8 P<0-01 133 BR girls higher than 193 BR mothers.subject (Note: Significant difference also between LF girls and mothers;

confirmed for all groups by regression analysis, see text)Car 3*2 P<0*005 182 BR children with car higher than 90 BR children without.ownership (Note: No significant difference with LF children or adults or with

BR adults)Sex 2-4 P<0-02 140 BR boys higher than 133 BR girls.

(Note: No significant difference with LF children)Capillary or 1-6 P<0 001 Capillary higher. 222 womenvenous blood

living in the eastern sector. Similarly constructedtables for the BR adults, for the LF children andfor the LF adults showed no significant differencesof mean-with the only exception that 30 BR adultsliving in the north-eastern sector had blood leadlevels significantly lower (P <0 02) than the meanfor their group.

Regression analysis also gave anomalous resultswith the four groups of subjects, no correlationwith direction from the factory being stronger thanr = 0'25. Thus the evidence on direction from thefactory is confficting: no consistent effect wasdemonstrated. No consistent effect on blood leadwas found for: rateable value of house, length ofresidence, or length of time a leadworker had beenresident. Table 4 sets out, for a number of factorswhich did seem to have a significant influence, thedifference between the mean blood lead of thegroups specified. The difference of 3 '8 ig shown for'age of subject' is a straightforward difference of

means between that for 133 BR girls and thatfor 193 BR mothers. Linear regression gave abest-fit relationship:CaPillary level = 27 3 - 0.178 x (agein) r= 0 31blood lead I years j P <0-01For the leadworkers families, the effect of age wasmore pronounced:Capillary blood _34 0-0-439 (age) + r = 0'49lead level -0 0033 (age)2 P < 0O01With the exception of the LF children, the other

groups did show a significant weak correlationbetween blood lead and the age of the house theylived in. Quite clearly, this factor is linked withother environmental factors, such as lead pipesor type of paint, and does not operate independently.For the BR children linear regression gave a bestfit relationship:Capillary age of r= 0'18blood lead = 25*8 + 0'0622 x houselevel in years P <0 01

158 W. J. Elwood et aL



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from



LEAD IN AIRBecause sampling was limited to a period of twomonths, it was not possible to study concentrationsunder all kinds of weather conditions. The windspeed and direction however were monitored andthese records have been used to estimate the numberof hours for which each sampling site was within15° of the downwind direction from the factoryon each day. This number of hours is referred tohere as the 'downwind time', and is naturally only anotional figure, owing to the arbitrary nature of itsdefinition. Lead concentrations on days withdifferent downwind times are given in Table 5.

Table 5 Mean lead-in-air concentrationsMean

Downwind time No. of concentration SD(hours) samples (g/rlm3)

0 178 0-45 0-27> 0 265 0-53 0-36> 1 87 0-69 0-45>6 21 0-66 0-36

From this it can be seen that the general backgroundis about 0 45 pg/m3 (based on zero downwindtime). Although there is considerable variability, asindicated by the standard deviation results, it canalso be said that the downwind sites are on averagehigher in lead by some 0 2 jig/M3. The effect ofincreasing downwind time from one to six hours ishowever not clear, presumably because of the smallnumber of samples and the high variability.The results for the six different sites are sum-

marised in Table 6. The slight effect of downwind

Table 6 Mean lead-in-air concentrations for individualsitesFrom factory Average No. of Mean

downwind days concerDistance Direction time per sampler (Ug/nm(metres) day (hours) operated800 NE 0-71 59 0-60800 East 0-63 38 0-48800 SSW 2-63 61 0-65

2400 NE 0*82 52 0*462400 East 0-86 59 0-402400 SSW 3*43 24 0-81

ntration SDn3)

0-290-300-430-260-270-58

position relative to the factory is again indicatedby these results, but no consistent effect of distancecan be discerned. Results over 1 0 g.g/m3 have beenexamined individually in an attempt to identifyspecific causes of high concentrations. Out of 20such results 70% were associated with downwindtimes of at least one hour, but the converse relation-ship did not hold, since of some 17 samples withover eight hours' downwind only two gave resultsof over 1 0 pg/m3. It was thought that the weather

may have been a contributory factor in this variationand so a preliminary examination of the currentweather was made. This seemed to indicate thathigh concentrations were more likely to occur oncalm days, but owing to the small number of samplesthis could not be established with confidence.

OTHER ENVIRONMENTAL SAMPLINGRESULTSThe results for dust on transport routes, dust fromledges, and surface soil are shown diagrammaticallyin Figs 1 to 3, while data for indigenous mossesand moss bags and for grasses have been presentedin a previous paper (Ratcliffe, 1975). Results fordeep soil were lower than for surface soil by afactor of 5 within one-third of a mile, and byprogressively smaller factors at greater distances.The mean lead level found in 47 specimens of tapwater taken at 18 homes in the survey area was0048 parts per million (ppm). Although the meanfor all hot tap specimens (0 049 ppm) was similarto that for all cold tap specimens (0 047 ppm), inindividual houses a high level of lead was oftenfound in one tap or the other but less often inboth. Although the number of samples is too smallto be statistically significant, it was noted thatin four BR families living in houses where tap waterlead was high their mean blood lead levels werehigher than the general mean for BR subjects-theadults by 2 jsg/lOO ml and the children by 6 jig/lOOml.

3,

A-r

.0

0.b

-C

.0

E,

0

*+"+HO Transport routesKerbside edge of pavement

Transport routesRoadside gutter-- --- Mean value for other roads

= 1640 ppm

0 1 2 3 4 b 8 10 12 14 160-1 0-5 1.1 1-5

Distance from works -lOOs of yords (above) km (below)

Fig. 1 Mean values of lead in specimens of dust ontransport routesfrom factory and on other roads carryingcomparabk traffic (I unit = 105 ppm by weight)

159



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from


W. J. Elwood et al.

N

1300 04E

/ 850 70 \

\ 80 /\30170 /

2SW 230 SE/S

SN-- Factory boundary

33

/ \ ~ ~~~~9\N

1

s

Fig. 2 Mean values of lead in specimens ofdust from ledges (1 unit = 10ppm byweight).Fig. 3 Mean values of lead in specimens of surface soil (1 unit = 10 ppm by weight).-

160



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from



Discussion

ENVIRONMENTAL LEAD IN THE VICINITYOF A LEAD WORKSApart from air samples and some anomalousresults-such as the lead in dust in the northernsector (Fig. 2)-generally the environmental samp-ling showed higher concentrations of lead near thefactory with a sharp fall-off as distance increased.Thisiswellillustrated,forexample, by the patternof lead concentration in dust on the transportroutes (Fig. 1); after 400 yards (0 4 km) it haddeclined to a level of about one-tenth of thatfound in specimens taken at the factory gates;it should be noted, however, that this level of leadin the road dust was about eight times as great asthat found in samples of dust taken from comparableroads not carrying lead works traffic. It was alsofound that mean lead levels for soil, grass, and mosswere very much higher within one-third of a mile(0 - 5 km) than beyond (Fig. 3). Surveys in the vicinityof lead works in other parts of the country (Depart-ment of the Environment, 1974) have shown asimilar pattern of high values for lead in dustand soil close to the works. Our studies revealedslightly increased mean concentration of lead in airin the downwind position relative to the factory,but no consistent effect of distance.

Unlike the position at Tower Hamlets andRotherhithe (Department of the Environment,1974) where many children lived close to the works,and where higher blood lead levels were found inthem than in those who lived further away, only afew of our subjects lived within one-third of a mile(0-5 km) of the factory (Table 3), and we found noconsistent difference in mean blood lead levelsbetween those who lived near and those who livedfurther away which could confidently be attributedto the effect of the factory. At the other extreme,the survey area did not extend beyond the localauthority boundaries (which were roughly threemiles-4.8 km-from the works). It cannot beclaimed that the 'control' population sampleliving within those boundaries and more than onemile (1 6 km) from the works is drawn from anenvironment (if such could be found) free fromsources of lead; their area of residence formspart of a large conurbation-with all the variedsources of lead that an urban environment canproduce. This population sample is howeverprobably representative of many urban populationstoday.

FACTORS AFFECTING THE INTERPRETATIONOF RESULTSA survey of the type described can yield results

which may be interpreted qualitatively, but it isnecessary to express some reservations about precisenumerical values and comparisons with othersurveys. Cochrane (1954) has stressed the importanceof response rates in population sampling work.Notwithstanding the efforts made, our responserate was only 53% in the BR (near) group, althoughit was rather better in the others.Lead in dust values are particularly erratic and

differences of the order of tenfold can be foundbetween specimens taken at the same spot. For thisreason dust and soil values quoted in this paper arethe means of four specimens (or two in the caseof the transport routes, Fig. 1) taken from thesample place at the same time.

In the interlaboratory comparison of blood leadanalysis it was found that the two laboratories wereequally consistent within themselves, both having'within samples' variance of about 2-8 pg/100 ml.Although one laboratory was giving results some6-9 pg/100 ml higher than the other, there wassome evidence that the interlaboratory differencedeclined in magnitude as the survey proceeded.Further investigation indicated that some 2*7pg/100 ml of the total interlaboratory difference of6 9 jg/100 ml could be accounted for by the use ofdifferent containers for the venous samples and thecorresponding microaliquots. We conclude thatinterlaboratory difference is a factor which must beconsidered when comparing blood lead levelsascertained in different surveys.

Either capillary or venous blood may be used in asurvey: we found that the mean lead level in capillaryblood was 1 * 6 pg/100 ml higher than that in venousblood. This observation rests on the analysis by thesame laboratory of specimens of capillary bloodand microaliquots of venous blood taken on thesame occasion from 222 women.

SIGNIFICANCE OF BLOOD LEAD LEVELSThe outstanding finding in this survey was that thechildren of the leadworkers had higher mean bloodlead levels than comparable children of people whowere not leadworkers: this confirms the findings ofthe surveys at Welwyn Garden City and Darley Dale(Department of the Environment, 1974) which werebased on relatively small numbers of subjects. In ourstudy 17% of LF children had blood lead levels inexcess of 40 jig/100 ml compared with 5% of BRchildren. The difference between means-6- 1pig/100 ml averaged over the whole range of ourunder-fives-was particularly striking at ages twoand three. This confirms the impression that it isbecause of behaviour such as crawling and suckingthat children of this age are more liable than othersto absorb lead that may be present in their environ-

161



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from


W. J. Elwood et al.

ment. We also found a significant difference betweenthe means of LF and BR mothers when capillaryblood data were used. Since, however, this differencewas relatively small, and not confirmed by the datafor mothers' venous blood, we attach far less weightto it than we do to the children's result.The absence of significant association for non-

leadworkers between their blood lead levels whethertheir residence was near or far from the factory,or was in any particular direction from the factory,is consistent with the hypothesis that it is the lead-worker himself who brings the lead home. It seemsreasonable to suspect that the lead is carried homein the form of dust on his person or on his clothing.This hypothesis has been elegantly validated byBurrows (1976), who, in a survey in which Lab Awas used, demonstrated that of all the means bywhich a leadworker may bring lead home, hisshoes are by far the most important.The data were analysed to assess the effect of a

number of factors each of which was looked at as ifit were acting alone (Table 4). The magnitudes of theeffects indicated in this table should be regardedas upper limits, as in many cases the effects werenot equal in all population subgroups. In whateverway this exercise may be interpreted, there was nodoubt that the presence of a leadworker in the homewas the biggest single factor, and the one whichoperated most consistently, in bringing about asignificant raising of the blood lead level of thechildren in the household.

This survey was planned and conducted withoutany preconceptions about the range of blood leadlevels or what limits were 'normal', 'acceptable', or'dangerous'. Moncrieff et al. (1964) took 36 pg/100ml as 'the upper limit of normal', and since thennumerous studies (for example, Betts et al., 1973)have suggested upper limits of normal rangingfrom 35 to 40 pg/100 ml. Our work exemplifies thedifficulty of laying down and applying rigid numeri-cal standards in this field: although 95% of our BRchildren had blood lead levels below 40 jsg/100 ml,this proportion would have been higher, and themean lead levels in each group lower than themeans shown in Tables 2 and 3, if we had takenvenous blood from the children and used Lab Ainstead of Lab B. During the survey three childrenwho were found after a second blood test to have ahigh level (64/60, 60/59, and 61/54) were referredto a paediatrician. All were discharged from out-patients after investigations and a number ofattendances, and a check with both general prac-titioner and paediatrician 9-12 months later indicatedno known departure from normal health.

DAGENHAM SURVEYThe survey undertaken at a battery factory atDagenham, while basically similar in concept tothat at Clifton Junction, was on fewer samples ofboth human blood and environmental samplesand therefore the results showed generally lowerlevels of statistical significance than did this largersurvey. Despite this, the main results of the CliftonJunction survey were confirmed by the Dagenhamwork.

Conclusions

Lead is a poison and the aim must be 'minimalexposure to lead at work, as little contamination ofthe environment as possible, and no involvement ofthe surrounding populace and workers' families'(Department of the Environment, 1974). We havefound no consistent evidence-in spite of demon-strable dissemination of lead to the environment-that the factories studied are producing measurablelead absorption in people living in their vicinity.We have however shown that of the sources studiedthe presence of a leadworker in the home is thefactor which operates most strongly and mostconsistently in raising the blood lead level of thefamily. Our data suggest however that the influenceof other factors-such as lead pipes in the homeand car ownership-merits further evaluation.

This survey was sponsored and funded by ChlorideGroup Ltd, with the co-operation of the former andpresent local authorities and health authorities.Our thanks are due to: Dr E. G. Bellinger, Dr P. C.Elwood, Mrs A. H. George, Dr J. A. Gillett,Dr H. C. Gupta, Dr T. S. Jones, Dr A. J. Lane,Mrs P. Mahgoub, Mr C. Robinson, Dr A. S. StLeger, Dr M. H. P. Sayers, Mrs J. Shepherd,Professor R. S. Scorer, Mr F. E. Whitehead, Dr M.K. Williams and the public health inspectors andhealth visitors of the boroughs of Barking, Prest-wich, Swinton and Pendlebury, and WhitefieldUrban District. Financial support for communitysurvey work is given by the DHSS to the GreatOrmond Street Laboratory.

Reprints (or copies of a fuller report on the survey)from DM, Chloride Group Ltd, 52 GrosvenorGardens, London SW1.

References

Betts, P. R., Astley, R., and Raine, D. N. (1973). Leadintoxication in children in Birmingham. British MedicalJournal, 1, 402-406.

162



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from



Burrows, G. E. (1976). Lead: From the factory to thehome. MD thesis. University of Liverpool.

Cochrane, A. L. (1954). The detection of pulmonarytuberculosis in a community. British Medical Bulletin,10, 91-95.

Delves, H. T. (1970). A micro-sampling method for therapid determination of lead in blood by atomicabsorption spectrophotometry. Analyst, 95, 431-438.

Department of the Environment (1974). Lead in theEnvironment and its Significance to Man. HMSO:London.

Department of the Environment and the Welsh Office(1973). Lead and the Environment. Joint circulars 6/73and 53/73. HMSO: London.

Department of Health and Social Security (1976). Supra-regional Assay Service, pp. 84-85. DHSS: London.

Godber, Sir George (1971). Environmental HazardsfromLead. CMO letter 20/71 to Medical Officers of Health.DHSS: London.

Ministry of Agriculture, Fisheries, and Food (1972).Second Report on the Monitoring of Foodstuffs forHeavy Metals. HMSO: London.

Ministry of Agriculture, Fisheries, and Food (1975).Survey of Lead in Food: First Supplementary Report.HMSO: London.

Moncrieff, A. A., Koumides, 0. P., Clayton, Barbara E.,Patrick, A. D., Renwick, A. G. C., and Roberts, G. E.(1964). Lead poisoning in children. Archives ofDiseasein Childhood, 39, 1-13.

Ratcliffe, J. M. (1975). An evaluation of the use ofbiological indicators in an atmospheric lead survey.Atmospheric Environment, 9, 623-629.

163



j.com/

Br J P

rev Soc M


ber 1977. Dow

nloaded from


blood near a - journal of epidemiology and community health | … · analyses were carried out at...

Documents