Occupational and Entironmental Medicine 1994;51:347-353 Liver function tests and urinary albumin in house painters with previous heavy exposure to organic solvents Ingvar Lundberg, Gun Nise, Gunilla Hedenborg, Margareta Hogberg, Olof Vesterberg Abstract The serum activities or concentrations of aspartate aminotransferase (ASAT), ala- nine aminotransferase (ALAT), alkaline phosphatase (ALP), albumin, y-glutamyl transpeptidase (GGT), bilirubin (BIL), cholic acid (CHOL), chenodeoxycholic acid (CHENO), and transferrin with iso- electric point 5-7, and the urinary excre- tion of albumin were determined among male current or former house painters (n = 135) and house carpenters (n = 71) who had worked in their trades for at least 10 years before 1970. Workers who showed a value above the 90th percentile among the carpenters in at least one of the tests ASAT, ALAT, GGT, BIL, CHOL, or CHENO were regarded as showing "pos- sible signs of liver dysfunction". Each participant's lifetime solvent exposure was evaluated by interview. The painters were divided into categories with low, intermediate, and heavy cumulative exposure during life (LTSE) or during the most exposed year (MEYSE). All participants stated none or slight recent exposure. The prevalence of possible signs of liver dysfunction increased with solvent exposure category according to LTSE as well as MEYSE with a numeri- cally higher risk estimate in the heavy exposure category for MEYSE than for LTSE. ALP activity increased with expo- sure category according to both exposure estimates. This increase seemed to be due to an interaction between exposure to solvents and current or previous long term intake of medicines potentially toxic to the liver. None of these results was affected by whether or not the subjects had been exposed to solvents during the year before the investigation. The expo- sure to solvents was not significantly related to any other outcome variable. It is concluded that long term heavy expo- sure to solvents may elicit changes in conventional liver function tests indica- tive of a mild chronic effect on the liver. The findings also suggest that heavy sol- vent exposure during shorter time peri- ods is a more likely cause of the findings than lifetime cumulative solvent expo- sure and that an interaction between sol- vent exposure and medicines potentially harmful to the liver may be important in the causation of the effects. (Occup Environ Med 1994;51:347-353) Some halogenated solvents-for example, 1,1,2,2-tetrachloroethane, carbon tetra- chloride, and chloroform-are classical hepto- toxins. The liver toxicity of these solvents is due to metabolic activation with formation of reactive metabolites. Because hepatotoxicity is a striking feature of these solvents it has been suspected that other solvents commonly used today may also be hepatotoxic.' Most epidemiological studies to determine the hepatotoxicity of industrial solvents in widespread use-such as white spirit, toluene, and xylene-have shown no effect of ongoing exposure on routine liver enzymes in serum,24 but recent studies have found an increase in bile acids in serum among workers exposed to mixed solvents as well as among workers exposed to styrene."- Some studies have suggested that exposure to organic solvents may induce glomerulo- nephritis,9 l0 whereas other studies did not find an increased risk." 12 Increased albumin excretion, as a marker of glomerular damage, has been found in groups of workers with ongoing solvent exposure in some studies"-'6 and other studies have found increased excre- tion of enzymes or of retinol binding protein, indicating tubular dysfunction, but not of albumin. 17-19 The long term significance of the findings concerning markers of liver and kidney dysfunction in groups of active workers has, however, to our knowledge never been inves- tigated. This study examines liver function and albumin excretion in a group of current or former house painters who had been active in their trades for at least 10 years before 1970. They had all been exposed to high sol- vent concentrations for extended periods, but their current exposure was low or none. The decrease in solvent exposure was due primarily to an extensive shift from solvent based to water based paints during the past decades. Thus the study also provided a possibility to examine potential effects of exposure to water based paints. Materials and methods STUDY GROUP The original population considered for the study were the members of the painters union and the carpenters in the union of building workers in Stockholm who were born in 1925 or later, who were members of the Stockholm branches of their respective trade unions in 1965, and who had been members for at least 10 years before 1970. These eligibility criteria Department of Occupational Health, Karolinska Hospital, Stockholm, Sweden I Lundberg G Nise Department of Occupational Medicine, National Institute of Occupational Health, Solna, Sweden M Hogberg Department of Clinical Chemistry, Karolinska Hospital, Stockholm, Sweden G Hedenborg Department of Chemistry, National Institute of Occupational Health, Solna, Sweden o Vesterberg Requests for reprints to: Dr Ingvar Lundberg, Department of Occupational Health, Karolinska Hospital, S-104 01 Stockholm, Sweden. Accepted 25 October 1993 347 on April 28, 2021 by guest. Protected by copyright. http://oem.bmj.com/ Occup Environ Med: first published as 10.1136/oem.51.5.347 on 1 May 1994. Downloaded from
Transcript
Occupational and Entironmental Medicine 1994;51:347-353
Liver function tests and urinary albumin in housepainters with previous heavy exposure to organicsolvents
AbstractThe serum activities or concentrations ofaspartate aminotransferase (ASAT), ala-nine aminotransferase (ALAT), alkalinephosphatase (ALP), albumin, y-glutamyltranspeptidase (GGT), bilirubin (BIL),cholic acid (CHOL), chenodeoxycholicacid (CHENO), and transferrin with iso-electric point 5-7, and the urinary excre-tion of albumin were determined amongmale current or former house painters (n= 135) and house carpenters (n = 71) whohad worked in their trades for at least 10years before 1970. Workers who showed avalue above the 90th percentile amongthe carpenters in at least one of the testsASAT, ALAT, GGT, BIL, CHOL, orCHENO were regarded as showing "pos-sible signs of liver dysfunction". Eachparticipant's lifetime solvent exposurewas evaluated by interview. The painterswere divided into categories with low,intermediate, and heavy cumulativeexposure during life (LTSE) or duringthe most exposed year (MEYSE). Allparticipants stated none or slight recentexposure. The prevalence of possiblesigns of liver dysfunction increased withsolvent exposure category according toLTSE as well as MEYSE with a numeri-cally higher risk estimate in the heavyexposure category for MEYSE than forLTSE. ALP activity increased with expo-sure category according to both exposureestimates. This increase seemed to bedue to an interaction between exposure tosolvents and current or previous longterm intake ofmedicines potentially toxicto the liver. None of these results wasaffected by whether or not the subjectshad been exposed to solvents during theyear before the investigation. The expo-sure to solvents was not significantlyrelated to any other outcome variable. Itis concluded that long term heavy expo-sure to solvents may elicit changes inconventional liver function tests indica-tive of a mild chronic effect on the liver.The findings also suggest that heavy sol-vent exposure during shorter time peri-ods is a more likely cause of the findingsthan lifetime cumulative solvent expo-sure and that an interaction between sol-vent exposure and medicines potentiallyharmful to the liver may be important inthe causation ofthe effects.
(Occup Environ Med 1994;51:347-353)
Some halogenated solvents-for example,1,1,2,2-tetrachloroethane, carbon tetra-chloride, and chloroform-are classical hepto-toxins. The liver toxicity of these solvents isdue to metabolic activation with formation ofreactive metabolites. Because hepatotoxicity isa striking feature of these solvents it has beensuspected that other solvents commonly usedtoday may also be hepatotoxic.'Most epidemiological studies to determine
the hepatotoxicity of industrial solvents inwidespread use-such as white spirit, toluene,and xylene-have shown no effect of ongoingexposure on routine liver enzymes in serum,24but recent studies have found an increase inbile acids in serum among workers exposed tomixed solvents as well as among workersexposed to styrene."-Some studies have suggested that exposure
to organic solvents may induce glomerulo-nephritis,9 l0 whereas other studies did notfind an increased risk." 12 Increased albuminexcretion, as a marker of glomerular damage,has been found in groups of workers withongoing solvent exposure in some studies"-'6and other studies have found increased excre-tion of enzymes or of retinol binding protein,indicating tubular dysfunction, but not ofalbumin. 17-19The long term significance of the findings
concerning markers of liver and kidneydysfunction in groups of active workers has,however, to our knowledge never been inves-tigated. This study examines liver functionand albumin excretion in a group of currentor former house painters who had been activein their trades for at least 10 years before1970. They had all been exposed to high sol-vent concentrations for extended periods, buttheir current exposure was low or none. Thedecrease in solvent exposure was due primarilyto an extensive shift from solvent based towater based paints during the past decades.Thus the study also provided a possibility toexamine potential effects of exposure to waterbased paints.
Materials and methodsSTUDY GROUPThe original population considered for thestudy were the members of the painters unionand the carpenters in the union of buildingworkers in Stockholm who were born in 1925or later, who were members of the Stockholmbranches of their respective trade unions in1965, and who had been members for at least10 years before 1970. These eligibility criteria
Department ofOccupational Health,Karolinska Hospital,Stockholm, SwedenI LundbergG NiseDepartment ofOccupationalMedicine, NationalInstitute ofOccupational Health,Solna, SwedenM HogbergDepartment ofClinical Chemistry,Karolinska Hospital,Stockholm, SwedenG HedenborgDepartment ofChemistry, NationalInstitute ofOccupational Health,Solna, Swedeno VesterbergRequests for reprints to:Dr Ingvar Lundberg,Department of OccupationalHealth, Karolinska Hospital,S-104 01 Stockholm,Sweden.Accepted 25 October 1993
were chosen to locate painters who had beenworking during the 1950s and 1960s whenexposure to organic solvents was high amongpainters in Stockholm.20
Information on the eligibility criteria wascollected from the membership registries ofthe respective trade union. The inclusioncriteria were met by 767 painters and 1212carpenters.A sample of 149 painters and 85 carpenters
were invited, in a random order, to a clinicalexamination in 1988-9 as part of a largerongoing study focusing on the relationbetween organic solvent exposure and neuro-psychiatric effects21; 135 painters and 71carpenters came to the examination.
EXPOSUREInterviewAll participating workers were interviewed byan occupational hygienist (GN) regardingtheir entire exposure history. The exposure tosolvents during work and leisure time sincefirst exposure was obtained in detail. In theinterview the different jobs held by the subjectwere listed and the duration of the jobs wasestimated. The average work time per day indifferent jobs was noted. For the painters theaverage amount of time actually spent paint-ing, as compared with other tasks, wastotalled. The time spent painting indoors v
outdoors, the amount of time working withroller or brush v spraying, and the fraction oftime wearing a respirator was estimated as apercentage of total work time for every jobheld.
Outdoor exposure to all types of paints wasestimated to have provided only a small frac-tion of the cumulative exposure of each par-ticipant and was excluded from all exposurecalculations. Taking roller or brush paintingas unity it was assumed that spray paintingdoubled and adequate use of respiratorshalved the exposure. These relations werederived from the papers by Riala et al 22 andFidler et al,23 and from personal experience.
Painting or other exposure during leisuretime was included in the exposure calcula-tions in the same way as occupational expo-sure.
Finally the time actually spent painting,with appropriate corrections, was totalled in"painting-months"; one painting-month beingequivalent to 176 work hours.The participants were exposed to several
solvents although exposure to white spiritalways constituted a large share of the totalexposure for all painters. On certain occasionssuch as painting ships, tanks, industrial build-ings with special requirements, or paintingin hospitals and gluing laminated boards inelevators or kitchens toluene or xylene hadconstituted the major part of the solvent expo-sure. Such work was totalled in painting-months in the same way as work in whichwhite spirit was the major exposure, andincluded in the exposure calculations.Working with water based paints was
judged to involve a very low solvent exposure,which was not included in the calculations.
The direct solvent exposure of the carpen-ters was almost exclusively due to the use ofsolvent based glues. This use was calculatedas equivalent to painting with a roller or brushindoors. The solvents most often used by thecarpenters were xylene, toluene, butanols,and white spirit.
Exposure estimatesLifetime cumulative solvent exposure (LTSE)The exposure to solvent in air during a paint-ing-month is likely to have been differentduring different time periods. Compilations ofmeasurements from previous decades showedthat the average eight hour time weightedaverage exposure to white spirit during apainting-month was close to the hygienicstandard during that period (Nise et al,unpublished observations). A reduction of thesolvent content in the paints generally coin-cided with a lowered hygienic standard. Thusthe fraction of the 1988 Swedish white spirithygienic standard enforced during a particulartime period (white spirit standard factor) wasmultiplied by the number of painting-monthsduring that period to get a number of hygienicstandard-months (HSMs). Table 1 shows thehygienic standards for Swedish white spiritduring different periods and the white spiritstandard factor applied during each period.Finally the HSM accumulated during differ-ent periods was totalled. For each subject thegeneral form of our cumulative solvent expo-sure index is given by:
HSM = z E (T. x Fjklm)
where i = the ith job; T = painting months;F = the fraction of time; j = form of applica-tion (1 = roller or brush, 2 = spray); k = workwith solvent based paints ( = 1), with waterbased paints ( = 0); 1 = respirator (protectingfactor = 0-5); m = white spirit standardfactor.
Most exposedyear solvent exposure (MEYSE)-This consisted of HSMs accumulated duringthe year with the most intense solvent expo-sure during the career as a painter.
Last year solvent exposure (LYSE)-This con-sisted of HSMs accumulated during the yearbefore the investigation. Sixty six of thepainters had no current solvent exposure asthey worked only with water based paints, hadchanged to a non-exposed occupation, or hadretired early.
Lifetime water-based paint exposure (LTWBE)-This exposure was estimated in accordance
Table 1 The Swedish hygienic standd-sfor white spiritduring different periods and the coreponding white spiritstandard-factor
Liverfunction tests and urinary albumin in house painters with previous heavy exposure to organic solvents
Table 2 Distribution of different exposure estimates in exposure categories among the painters. The carpenters' LTSErangedfrom 0 to 46 hygienic standard months (HSMs) (n = number ofpainters in each category)
Exposure category
Exposure estimate Category 1 Category 2 Category 3(units) n (lowest) n (intermediate) n (highest)
with that of the solvent based paints but wasnot weighted according to any hygienic stan-dard. It was totalled in water-based-months.
Last year water based paint exposure(LYWBE)-This was the number of water-based-months during the year before theinvestigation.
Table 2 shows the variations in the differentexposure estimates. In the statistical analysesthe exposure estimates were categorised as inthis table. Most of the painters had their heav-iest solvent exposure during the 1960s and theuse of solvent based paints decreased consid-erably after that. During the five years beforethe investigation only 0-14% (median 2%) ofthe total HSMs were accumulated among theindividual painters (figure).
INFORMATION ON POTENTIAL CONFOUNDINGFACTORSThe participants were interviewed for poten-tial confounding factors. The confoundingfactors considered were:
(a) Age.(b) Alcohol consumption. The participants
were asked about their average consumptionof spirits, wine, and beer each week during thepast year. The consumption of the differentbeverages was totalled as g 100% alcohol fromdata on their average alcohol contents.
(c) The degree of obesity. This may affectseveral tests of liver function. The weight andthe length of the subjects were measured andBroca's index (100 x (weight in kg)/(heightin cm - 100)) was calculated.
1953 1958 1963 1968
(d) Present or previous long term intake ofmedicines potentially harmful to the liver(used in analyses of liver function tests) or thekidney (used in analyses of urinary albumin)according to the Swedish reference book ofdrug information (FASS), which contains alldrugs approved by the Medical ProductsAgency in Sweden. The drugs were mostlyantihypertensive agents, analgesics, sleepingpills and tranquilisers, and anti-inflammatorydrugs.
(e) Blood pressure. This may affect albu-min excretion in urine. Blood pressure wasmeasured after five minutes of rest with a con-ventional cuff. The systolic pressure showed astronger relation to urinary albumin excretionthan the diastolic pressure and was used in theanalyses.
(f) Past or present liver or kidney disease.The large majority of previous kidney diseaseswere kidney stones and previous liver diseasesincluded hepatitis and those of the gall blad-der. Few subjects stated such diseases andthey were not shown to affect any outcome.Thus these variables were not used as poten-tial confounders in the data analysis.
For the analyses the potential confounderswere categorised. Age, Broca's index, and sys-tolic blood pressure were dichotomised by themedians, 50 years, and 104, and 140 mm Hgrespectively. Alcohol consumption was dicho-tomised at a weekly consumption of 126 galcohol.
Table 3 gives the distribution of the poten-tial confounders according to LTSE.
OUTCOMESThe participants were examined on a
-- __~ Monday, Tuesday, Wednesday, or Thursday.They had been instructed not to eat or drinkfor eight hours before the investigation. When
* Most exposed painter coming to the investigation at around 8.00 amo Median exposed painter blood samples were obtained from the cubitalA Least exposed painter vein and the participants were asked toA Most exposed carpenter urinate and the time of urination was noted.
After some initial tests of neuropsychiatricC ~ ~function and the exposure interview the sub-
jects were told to urinate again at around11.00 am and the time of the urination wasnoted. The volume of this urine sample wasmeasured. Then the urine samples werefrozen and stored at - 700C until the analysis
1973 1978 1983 1988 of urinary albumin, as an index of glomerularr kidney function, and urinary creatinine. Thelaryears. creatinine analyses were performed at the hos-
500 _
400 _
300 _
en_,,+1
0E_0
a)
_0
CIC.)
CDamT
200
100
I/
YearCumulative solvent exposure accumulated by calent
()-rb ^ - -_nA
349
on April 28, 2021 by guest. P
rotected by copyright.http://oem
.bmj.com
/O
ccup Environ M
ed: first published as 10.1136/oem.51.5.347 on 1 M
Age >50years 31 7 40 25Average alcohol consumption per week during past year> 126 g 16 10 21 4
Broca's index (see text) >104 42 15 31 15Systolic blood pressure 140 mm Hg 43 17 46 24Past or present liver disease 5 0 6 2Use of medicines potentially harmful to the liver 1 1 5 1 1 10Past or present kidney disease 7 6 11 4Use of medicines potentially harmful to the kidney 4 1 2 2
pital laboratory according to Kodak EktachemT-700 test methodology and the albuminanalyses at the Division of MedicalChemistry, Swedish National Institute ofOccupational Health by the method ofVesterberg.24 The urinary albumin concentra-tion was corrected for urinary creatinine.The blood was centrifuged and most of the
serum was immediately transferred to the hos-pital laboratory.
Aspartate aminotransferase (ASAT), ala-nine aminotransferase (ALAT), alkaline phos-phatase (ALP), y-glutamyl transpeptidase(GGT), albumin, and bilirubin (BIL) weredetermined the same day (Kodak EktachemT-700 test methodology).
Serum for the determination of cholicacid (CHOL) and chenodeoxycholic acid(CHENO) was frozen at - 20'C and analysedwithin two months by radioimmunoassay withantisera produced at the laboratory.25Serum for the analysis of transferrin was
frozen at - 20'C and delivered to theDivision of Medical Chemistry at the SwedishNational Institute of Occupational Healthwhere it was stored at - 70'C until analysis.The determination of total transferrin and thefraction of transferrin with an isoelectric pointof 5'7 (TF5..) was performed according toPetren and Vesterberg.26
Subjects with values above the 90th per-centile of the distribution among the car-penters in any of ASAT, ALAT, ALP, BIL,GGT, or the bile acids were also defined asshowing possible signs of liver dysfunction.
STATISTICAL METHODSThe data were analysed with multiple linearregression and logistic regression. Logisticregression was used when analysing data con-cerning possible signs of liver dysfunction.For the logistic regression analyses of separatetests the results were dichotomised at the 90thpercentile of the distribution among the car-penters. Multiple linear regression was usedfor continuous outcome variables. All out-comes that were not normally distributedwere positively skewed and were convertedinto logarithms for the multiple linear regres-sion analyses.
Multiple linear regression analyses wereperformed with continuous or discrete expo-sure or potential confounder variables cate-gorised as stated earlier. Dummy variableswere created from exposure variables withmore than two categories.
Cross tabulations were performed between
all exposure and all potential confounder vari-ables. In general there were weak associationsbetween exposure variables and potentialconfounder variables and between differentpotential confounder variables. In instanceswhere obvious associations existed-as forexample between lifetime cumulative expo-sure estimates and age-there was still areasonable representation in all cells of thecross tables and there was never reason toexclude any variable due to collinearity. Theassociations between different exposure vari-ables were weak or moderate except for theassociation between LTSE and MEYSEwhere only one painter in the category withhigh LTSE belonged to the category with lowMEYSE and only two painters with highMEYSE belonged to those with low LTSE.
In the analyses the effect of the differentexposure estimates, one at a time, was investi-gated in the entire study group with no adjust-ment for potential confounders. Then theeffects of the exposure estimates, one at atime, were investigated adjusting for all poten-tial confounder variables relevant to the par-ticular analysis. The computer packages SASand EGRET were used for the analyses.
In all statistical analyses the regressioncoefficients for the exposure estimates weresimilar with and without adjustment forpotential confounders. The results presentedare adjusted for potential confounders.
ResultsPOSSIBLE SIGNS OF LIVER DYSFUNCTIONPossible signs of liver dysfunction were morecommon among painters than among car-penters (OR 1-5, 95% CI = 0-8-2-8 afteradjustment for potential confounders). Theproportion of subjects with this outcomeseemed to increase with increasing MEYSEand LTSE (table 4) but not with increasingLYSE, LTWBE, or LYWBE. The dose-response relations in table 4 were similaramong those who had been exposed or notexposed to solvent and water based paintsduring the past year.Among the painters analyses were per-
formed with the MEYSE and LTSE exposureestimates entered together. In these analysesthe ORs for MEYSE were unchanged whereasestimates for LTSE decreased. There was atendency towards an interaction betweenintake of medicines and MEYSE or LTSE sothat the increase in ORs with increasing expo-sure was more evident among those taking
350
on April 28, 2021 by guest. P
rotected by copyright.http://oem
.bmj.com
/O
ccup Environ M
ed: first published as 10.1136/oem.51.5.347 on 1 M
Liverfunction tests and urinaty albumin in house painters with previous heavy exposure to organic solvents
Table 4 ORs, with adjustmentfor potential confounders, ofshowing an abnormal value in at least one of the tests GGT,ALAT, ALP, ASAT, CHENO, CHOL, BIL among painters in different LTSE categories compared with carpenters(n = number affected, N = number in group)
Exposure category
Exposure Carpenters Painters low Painters inte e Painters highestimate n/N OR n/N OR (95% CI) n/N OR (95% CI) n/N OR (95% CI)LTSE 24/70 1-0 11/34 0-9 (0-4-2-3) 28/65 1-6 (0-8-3-3) 18/34 2-4 (1-0-5-7)MEYSE 24/70 1-0 10/32 1-0 (0-4-2-4) 27/68 1-3 (0-62-7) 20/33 3-1 (1-3-7-4)
medicines potentially harmful to the liver.None of the potential confounders was
significantly associated with the outcome.
ANALYSES OF SEPARATE TESTS OF LIVERFUNCTION AND URINARY ALBUMINThere were only small differences in themedian values of separate function testsbetween the carpenters' and the painters'groups formed on the basis ofLTSE (table 5).In the multiple linear regression analyses thehighest LTSE category showed higher ALPvalues than the carpenters (p = 0-02). Thecategory with the highest MEYSE alsoshowed significantly raised ALP activities(p = 0.05). These effects were, however, duein large extent to an interaction between thestated intake of medicines potentially harmfulto the liver and these exposure estimates(table 6). Medicine significantly increasedALP activity among those with intermediateas well as heavy solvent exposures accordingto the MEYSE and LTSE exposure estimates.There was no difference in ALP activitybetween those with a high LTSE or MEYSEwho had been exposed to solvents during thepast year and the corresponding groups with-out solvent exposure.
No exposure estimate was significantlyassociated with any other outcome variable.The logistic regression analyses of the sepa-
rate tests did not reveal any associations notapparent in the multiple linear regressionanalyses.Of the potential confounders studied a high
alcohol consumption and a high Broca's indexwere associated with significantly higherGGT, urinary albumin, and ALAT valuesthan low alcohol consumption and lowBroca's index. A high systolic blood pressure
was significantly associated with higher uri-nary albumin values than low systolic bloodpressure.
DiscussionThe results suggest that heavy exposure topaint solvents has effects on the liver, but noton the kidney glomeruli. Our estimates ofexposure to water based paints were notsignificantly associated with any outcome.We firstly consider whether biases could
explain these findings and which type of expo-sure caused the effects. Finally we discuss thefindings in the light of previous investigations.
Table S Median values and ranges of tests of liver and kidneyfunction among carpenters and among painters with low,intermediate, and high LTSE
LTSE category
Carpenters Painters low Painters intermediate Painter highTest (unit and reference (n = 71) (n = 34) (n = 67) (n = 34)interval or upperreference lmit in laboratory) Median Range Median Range Median Range Median Range
Statistical comparisons are based on log ALP values.*p = 0-02 v carpenters (with adjustment for potential confounders).
Table 6 Median and range ofALP values (Mukat/l), in solvent exposure categories, among those stating (sm), and notstating (nsm), intake of medicines potentially harmful to the liver
Exposure categoryCarpenters Painters low Painters intermediate Painters high
Exposureestimate N Median Range N Median Range N Median Range N Median RangeLTSEsm 11 2-5 1-0-43 5 3-1 1-8-3-9 10 3-3 2-1-4-9* 10 3-2 2-1-68*msm 60 2-8 1-4-4-9 29 2-6 1-1-4-1 56 2-7 1-9-3-8 24 3-0 1-9-4-0
Statistical comparisons are based on log ALP values.p = 0-01; tp = 0-03; 1p = 0-002 compared with carpenters stating intake of medicines (with adjustment for potential
confounders).
351
on April 28, 2021 by guest. P
rotected by copyright.http://oem
.bmj.com
/O
ccup Environ M
ed: first published as 10.1136/oem.51.5.347 on 1 M
POTENTIAL SOURCES OF BIASSensitivity and specificity of testsThe most reliable way of measuring possiblechronic changes in liver function is to employsome clearance test, such as galactose elimi-nation or indocyananine green clearance.2728Such methods are, however, time consuming,generally require hospital facilities, or involvesome risk for complications. Thus they arenot readily used for screening workers for sub-clinical effects.Many of the tests used by us, such as
ASAT, ALAT, GGT, and TF5.7, mostlyreflect current hepatocellular injury27 andwould be expected to be sensitive to effectsfrom ongoing exposure. Some of these testswere in fact found to be quite sensitive toalcohol consumption as well as the degree ofobesity.The compounds measured in the routine
liver function tests are usually also increasedin liver cirrhosis and other chronic liver dis-
27eases. For example, an increase of at leastone of ASAT, ALAT, GGT, ALP, BIL, orCHOL was found in all 51 cases of liver cir-rhosis among outpatients with liver disease.29Thus the combination of tests used in possiblesigns of liver dysfunction was the outcomemost likely to reflect chronic effects on theliver. We made calculations to determinewhether ALP or any other particular test had alarge influence on possible signs of liver dys-function but this was not found to be the case.
On the other hand the specificity of thepossible sign of liver dysfunction outcome israther weak. It has been shown that 30% ofoutpatients with various gastrointestinal dis-turbances showed abnormal levels in at leastone of the tests used for the outcome in ourstudy.29 Thus our findings seem to requirefurther investigation concerning the tissue oforigin of the increased occurrence of possiblesigns of liver dysfunction in heavily exposedworkers in our study.
Urinary albumin has been shown to beincreased among workers with ongoing sol-vent exposure in several investigations andalso in early stages of diabetes nephropathy.30Therefore it seems likely that clinically signifi-cant acute as well as chronic effects on kidneyglomeruli induced by solvents should havebeen detected in our study, had they beenpresent.
Misclassification and confoundingIn this study information on exposure wasobtained from the subjects themselves. Thusthere is a possibility that unhealthy subjectswith different symptoms and increased levelof some liver function test may have exagger-ated their exposure. We believe, however, thatthe careful exposure interview covering allactivities involving solvent exposure for allyears since the beginning of exposure willhave greatly diminished this possibility.
In particular ALP activity increased withincreasing solvent exposure among those who,currently or previously for a long period, tookmedicines with potential liver toxicity. Thismay have reflected the fact that subjects with
heavy exposure had taken larger quantities ofsuch medicines than those with low, or no,exposure. No difference was evident, how-ever, between those with low and heavy expo-sure either concerning types of medicinesused or the amount of medicine taken.Neither were the results changed if thosecurrently taking medicines were calculatedseparately.
WHICH TYPE OF EXPOSURE MAY HAVE CAUSEDTHE EFFECTS ON LIVER FUNCTION TESTS?Increasing solvent exposure according to theLTSE as well as MEYSE estimates were asso-ciated with increasing prevalence of possiblesigns of liver dysfunction and of increasingALP activity among those taking medicine.Exposure to solvents during the past yearseemed to lack importance for all outcomes.The LTSE and MEYSE exposure estimateswere collinear and it is not possible to deter-mine whether one of them were more influen-tial than the other. Moreover they may bothbe related to some other exposure estimatethat is really causing the effects. HeavyMEYSE exposure yielded the highest riskestimate for possible signs of liver dysfunc-tion, however, and when LTSE and MEYSEwere included in the same logistic regressionmodel the risk estimates for LTSE tended todecrease. This suggests that not only cumula-tive solvent exposure but also particularlyheavy exposures during shorter time periodsneed to be assessed in studies of liver effectsfrom solvents.
COMPARISON WITH PREVIOUS INVESTIGATIONSThe association and dose-response trendbetween two measures of solvent exposureand possible signs of liver dysfunction supportthe idea of a chronic effect from solvent expo-sure on the liver. We are not aware of previousinvestigations using this or similar outcomesto measure such effects.The only separate test of liver dysfunction
related to any exposure estimate was measure-ment of ALP activity. In a previous study itwas suggested that this test is one of the mostspecific for mild chronic hepatic injury due toenvironmental agents.28 Workers with currentexposure to mixed solvents and toluene haveshown specific increases in this enzyme inserum3' 32 but interactive effects between med-icines taken and solvent exposure were notreported. Such an interaction has been sug-gested, however, by previous case reports.33Our results concerning ALP suggest a mild,
chronic solvent effect on the liver that is man-ifested among those who take medicines withpotential effects on the liver.
Several studies have shown that exposureto mixed solvents does not affect commonliver enzymes such as ALAT, ASAT, andGGT in serum2-4 6 and our results are compat-ible with these findings.
Recent studies of workers with ongoingexposure to mixed solvents have shown anincrease in serum bile acids6 7 but anotherrecent study34 did not confirm this effectamong workers with probably heavy current
352
on April 28, 2021 by guest. P
rotected by copyright.http://oem
.bmj.com
/O
ccup Environ M
ed: first published as 10.1136/oem.51.5.347 on 1 M
Liverfunction tests and urinary albumin in house painters with previous heavy exposure to organic solvents
exposure to mixed solvents. Workers exposedto rather low concentrations of mixed solventshave shown an increase of TF5.735.Furthermore some other studies havereported effects on TF5.7 in workers exposedto various organic chemicals.3637The painters examined by us had a previ-
ously heavy solvent exposure with no or lowcurrent exposure. Thus it is possible thatthese painters had previously had increasedbile acids and TF57, which had returned tocontrol group concentrations with decreasingexposure. It has been shown that serum bileacid concentrations normalise within a weekafter ceasing heavy alcohol intake.38 TF, 7decreases noticeably in alcoholic patientswithin a few weeks of abstinence andafter moderate misuse it remains at normallevels.26No effect on urinary albumin from solvent
exposure was seen although increases havebeen found in several studies of workers with asubstantial current solvent exposure.13-'6 It ispossible that the workers studied by us haveexperienced reversible kidney effects.
It is concluded that heavy previous expo-sure to solvents from solvent based paints,with no or low current exposure, may elicitchanges in routine liver function tests indica-tive of mild chronic effects on the liver.Furthermore our findings suggest that aninteraction between exposure to solvents andmedicines potentially harmful to the liver maybe important in the causation of these effects.
Urinary albumin concentrations seemednot to be affected by solvent exposure.
Our estimates of exposure to water basedpaints seemed unrelated to all outcomesinvestigated.The results merit further investigations to
ascertain whether the liver is the true tissue oforigin of the indications of chronic liver effectsfrom solvent exposure obtained in this study.If so, it also needs to be clarified whetherthese effects have clinical significance.
We thank Annika Gustavsson who assisted in the data analy-sis, and Sven Petren, Birgit Akerlund, and Helene Saraniuswho performed the transferrin and urinary albumin analyses.Bengt Sjogren gave critical comments on the manuscript. Thestudy was supported by the Swedish Work EnvironmentFund.
1 Lundberg I, Hogstedt C, Liden C, Nise G. Organicsolvents. In: McCullen R, Rosenstock L, eds. Clinicaloccupational and environmental medicine. New York:Saunders Company, 1994 (in press).
2 Kurppa K, Husman K. Car painters exposure to a mixtureof organic solvents. Serum activities of liver enzymes.ScandJV Work Environ Health 1982;8:137-40.
3 Lundberg I, Hakansson M. Normal serum activities ofliver enzymes in Swedish paint industry workers withheavy exposure to organic solvents. Br J Ind Med 1985;42:596-600.
4 Waldron HA, Cerry N, Venables H. Solvent exposure andliver function. Lancet 1982;ii:1276.
5 Edling C, Tagesson C. Raised serum bile acid concentra-tions after occupational exposure to styrene: a possiblesign of hepatotoxicity. Br. _JInd Med 1984;41:257-9.
6 Franco G, Fonte R, Tempini G, Candura F. Serum bileacid concentrations as a liver function test in workersoccupationally exposed to organic solvents. Int ArchOccup Environ Health 1986;58:157-64.
7 Franco G, Santagostino G, Lorena M, Imbriani M.Conjugated serum bile acid concentrations in workersexposed to low doses of toluene and xylene. Br IndMed 1989;46:141-2.
8 Franco G. New perspectives in biomonitoring liver func-tion by means of serum bile acids: experimental andhypothetical biochemical basis. Br J (nd Med 1991;48:557-61.
9 Ravnskov U, Forsberg B, Skerfving S. Glomerulonephritis
and exposure to organic solvents. A case-control study.Acta Med Scand 1979;205:575-9.
10 Bell G, Gordon A, Lee P. Doig A, MacDonald M,Thomson D, et al. Proliferative glomerulonephritis andexposure to organic solvents. Nephron 1985;40:161-5.
11 van der Laan G. Chronic glomerulonephritis and organicsolvents. Int Arch Occup Environ Health 1980;47:1-8.
12 Harrington J, Whitby H, Gray C, Reid F, Aw T,Waterhouse J. Renal disease and occupational exposureto organic solvents: a case referent approach. Br J IndMed 1989;46:643-50.
13 Askergren A, Allgen L-G, Karlsson C, Lundberg I, NybergE. Studies on kidney function in subjects exposed toorganic solvents. I. Excretion of albumin and beta-2-microglobulin in the urine. Acta Med Scand 1981;209:479-83.
14 Lauwerys R, Bernard A, Viau C, Buchet J. Kidney disor-ders and hematotoxicity from organic solvent exposure.ScandJ7 Work Environ Health 1985;11(suppl 1):83-90.
15 Hotz P, Pilliod J, Berode M, Rey F, Boillat M.Glycosaminoglycans, albuminuria and hydrocarbonexposure. Nephron 1991;58:184-91.
16 Hotz P, Pilliod J, S6derstrom D, Rey F, Boillat M,Savolainen H. Relation between renal function tests anda retrospective organic solvent exposure score. Br J IndMed 1989;46:815-9.
17 Franchini I, Cavatorta A, Falzoi M, Lucertini S, Mutti A.Early indicators of renal damage in workers exposed toorganic solvents. IntArch Occup Environ Health 1983;52:1-9.
18 Ng T, Ong S, Lam W, Jones M, Cheung C, Ong C.Urinary levels of proteins and metabolites in workersexposed to toluene. A cross-sectional study. Int ArchOccup Environ Health 1990;62:43-6.
19 Vyskocil A, Popler A, Skutilova I, et al. Urinary excretionof proteins and enzymes in workers exposed to hydrocar-bons in a shoe factory. Int Arch Occup Environ Health1991;63:359-62.
20 Bobjer 0, Knave B. Work load and exposure to solventsand dust hazard factors in house painting. In:International symposium on the control of air pollution inthe working environment. Stockholm: The WorkEnvironment Fund International Labour Office, 1977:37-61.
21 Lundberg I, Gustavsson A, Hogberg M, Nise G.Diagnoses of alcohol abuse and other neuropsychiatricdisorders among house painters compared with housecarpenters. BrJ Ind Med 1992;49:409-15.
22 Riala R, Kalliokoski P, Pyy L, Wickstrom G. Solvent expo-sure in construction and maintenance painting. Scand JWork Environ Health 1984;10:263-6.
23 Fidler AT, Baker EL, Letz R. Estimation of long term
exposure to mixed solvents from questionnaire data: atool for epidemiological investigations. Br J Ind Med1987;44:133-41.
24 Vesterberg 0. Quantification of albumin in urine by a newmethod: zone immuno-electrophoresis assay (ZIA). ClinChimActa 1981;113:305-10.
25 Samuelsson K, Johansson C, Norman A. Serum bile acidsafter a test meal in Crohn's disease. Scand Jf Clin Invest1979;39:511-8.
26 Petren S, Vesterberg 0. Concentration differences in iso-forms of transferrin in blood from alcoholics duringabuse and abstinence. Clin ChimActa 1988;175:183-8.
27 Price C, Alberti K. Biochemical assessment of liver func-tion. In: Wright R, Alberti K, Karran S. Millward-SadlerG, eds. Liver and biliary diseases. London: WB Saunders,1979:381-416.
28 Tamburro C, Greenberg R. Effectiveness of federallyrequired medical laboratory screening in the detection ofchemical injury. Environ Health Perspect 1981;41:117-22.
29 Samuelsson K, Aly A, Johansson C, Norman A.Evaluation of fasting serum bile acid concentration inpatients with liver and gastrointestinal disorders. ScandJGastroent 1981;16:225-34.
30 Mogensen CE. Microalbuminuria as a predictor of clinicaldiabetic nephropathy. Kidney Int 1987;31:673-89.
31 Elofsson SA, Gamberale F, Hindmarsh T, et al. Exposure to
organic solvents. A cross-sectional epidemiologicalinvestigation on occupationally exposed car and indus-trial spray painters with special reference to the nervous
system. ScandJ Work Environ Health 1980;6:239-73.32 Svensson BG, Nise G, Erfurth EM, Olsson H. Neuro-
endocrine effects in printing workers exposed to toluene.BrJ Ind Med 1992;49:402-8.
33 Edling C. Interaction between drugs and solvents as a
cause of fatty change in the liver. Br J Ind Med 1982;39:198-9.
34 Hotz P, Pilliod J, Bourgeois R, Boillat M-A. Hydrocarbonexposure, pancreatitis, and bile acids. Br Y Ind Med1990;47:833-7.
35 Petren S, Vesterberg 0. Studies of transferrin in serum ofworkers exposed to organic solvents. Br J Ind Med1 987;44:566-8.
36 Beving H, Malmgren R, Petren S, Vesterberg 0.Hematological changes in house painters using epoxypaints. Y Soc Occup Med 1991;41:102-6.
37 Beving H, Petren S, Vesterberg 0. Increased isotransferrinratio and reduced erythrocyte and platelet volumes inblood from thermoplastic industry workers. Ann OccupHyg 1990;34:391-7.
38 Ahlgren A, Hedenborg G, Norman A, Wisen 0. Serumbilirubin subtractions in patients with alcohol abuseduring detoxication. Scand Y Clin Lab Invest 1988;48:319-26.
353
on April 28, 2021 by guest. P
rotected by copyright.http://oem
.bmj.com
/O
ccup Environ M
ed: first published as 10.1136/oem.51.5.347 on 1 M
