Occupational and Environmental Medicine 1994;51:700-705

Evaluation -of phosphine genotoxicity atoccupational levels of exposure in New SouthWales, Australia

Anadergh Barbosa, Antonio M Bonin

AbstractPhosphine has been claimed to causechromosomal damage at exposures closeto the current time weighted averageexposure standard of 0'3 ppm (0.4mg/m3). The current study involved 31phosphine fumigators and 21 controlsduring the high fumigation season. Allwere volunteers and were evaluated forgenotoxicity variables, including micro-nuclei in peripheral blood lymphocytesand urine mutagenicity. In parallel, allfumigators and 17 controls were evalu-ated- -for full haematology, multiplebiochemical analysis, whole bloodorganochlorines, and whole blood andserum cholinesterase activity. The resultsfor- micronuclei showed no significantdifferences between fumigators and con-trols, but detected a strong associationbetween age and increased frequency ofmicronuclei. Measurement of urinemutagenicity did not show any significantdifference between fumigators and con-trols, but did show increased excretion ofmutagens in smokers. All haematologicaland biochemical variables were withinnormal ranges, except for some non-specific changes in biochemistry. Atmonitored occupational exposures of< 2-4 ppm/h our results show no associa-tion between phosphine exposure andgenotoxic or toxicological effects in fiuni-gators.

(Occup Environ Med 1994;51:700-705)

Keywords: phosphine, micronucleus induction, urinemutagenicity

Department ofOccupational Health,FHDF, Brasilia,BrazilA BarbosaToxicology Unit,National Institute ofOccupational Healthand Safety, WorksafeAustralia, Sydney,AustraliaA M BoninCorrespondence to:Dr AM Bonin, ToxicologyUnit, National Institute ofOccupational Health andSafety, Worksafe AustraliaGPO Box 58, Sydney,New South Wales 2001,Australia.

Accepted 6 June 1994

Phosphine has been an occupational healthissue in Australia since the early 1950s. The

establishment of codes of practice for fumiga-tion with phosphine and improvements in thetechnology of storage and fumigation haveresulted in lower occupational exposure tothis fumigant than in previous decades inAustralia' and internationally.2 Despite theseachievements, exposures can be furtherdecreased.

Effects of acute and sub-acute exposure inhumans to phosphine, mainly suicidal, havebeen described and symptoms includeheadache, diarrhoea, tightness of chest, pul-monary oedema, and death.35 Phosphine is a

recognised poison to mammals, birds, andmost insects and its toxicity and subsequent

lethality is dependent on the duration andconcentration of the exposure.6 Exposure byinhalation of phosphine seems to be the mosteffective route of absorption as it targets thewhole respiratory tract, especially the lungs.5-7

Within the state of New South Wales,Australia, Grain Corp is the governmentorganisation responsible for about 85% of thecommercial grain stores (about 270 sites, 500individual stores), with 31 fumigators beingresponsible for all phosphine fumigation.During a normal year, there are about170-200 fumigations with phosphine thatoccur mainly during summer months(December-March). There are four maintypes of stores used by this corporation:sealed storages (gas-tight), bunker stores(temporary), SIROFLO stores (sealed base+ open top), and unsealed non-SIROFLOstores.

In 1988, after the release of environmentalhealth criteria on phosphine and selectedmetal phosphides,6 it became obvious thatthere was a lack of genotoxicity data on thiscompound. At the same time, phosphine wasunder review in the United States by theEnvironmental Protection Agency for renewalof the licence for its use in that country.Because of this lack of genotoxicity data, theEnvironmental Protection Agency demandedfrom the manufacturers of phosphine andmetal phosphides a battery of tests for muta-genicity and carcinogenicity. These testsincluded an acute inhalation toxicity, aninhalation developmental toxicity, a 13 weekinhalation toxicity, and Ames/Salmonellaplate incorporation assay.7 In rats exposed to10 ppm for 3 days (6 h/day), there weredecreased erythrocytes, lung congestion, andincreased kidney weights with coagulativenecrosis of the tubular epithelium in the outercortex. Subchronically (6 h/day, 13 weeks),there was also a dose-related decrease in bodyweight gain at 1 and 3 ppm, with a 5%decrease in erythrocytes, haemoglobin, andpacked cell volume found in the 3 ppm group.Exposure of pregnant rats up to 4-9 ppmshowed no evidence of maternal or develop-mental toxicity.

In 1989, Garry et al published a study thatassociated occupational exposure to phos-phine with increases in chromosomal aberra-tions, which may have implications forcarcinogenicity." This generated the need toevaluate the occupational health and safety ofworkers handling this fumigant in Australia.Our study was therefore undertaken primarilyto evaluate two biomarkers of genotoxicity

700

on Novem

ber 3, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.10.700 on 1 October 1994. D

ownloaded from

Evaluation ofphosphine genotoxicity at occupational levels ofexposure in New South Wales, Australia

(induction of micronuclei in peripheral lym-phocytes and urine mutagenicity), as well asseveral other variables (full blood count,multiple biochemical analysis, whole bloodorganochlorines, serum and whole bloodcholinesterase) on long term fumigators whowere occupationally exposed to phosphine.Environmental monitoring was also per-formed to provide an indication of exposure.

Materials and methodsSUBJECT SELECTIONFumigators (n = 31) working with phosphine(mean (range) 11 6 (1-5-32) years) for theNew South Wales Grain Corp and controlsubjects (n = 21) working on the same sitesas grain handlers, mechanics, and clerks,volunteered for the study and were matchedby sex, age, and smoking habit. All subjectsgave signed consent for their participation inthis investigation and the project wasapproved by the Sydney University HumanEthics Committee. Samples of blood andurine were collected over the period fromFebruary-April 1992. All subjects were givenan approved personal health questionnaire byan experienced registered nurse, and detailswere obtained on chronic and acute illnesses,use of medication, diagnostic exposure tox rays in the past 12 months, and exposureto potential mutagens. Subjects with a historyof having medication or exposure to x rayswere subdivided into a separate group forevaluation of induction of micronuclei toensure that such exposure was not aconfounder.

BLOOD AND URINE COLLECTIONBlood and urine samples collected at thework sites were transported to the laboratoryand processed within 12 hours. Blood wascollected into six 10 ml vacuum tubes(Greiner, three tubes with sodium heparin,two tubes with EDTA and one tube withoutadditive). The tubes with sodium heparinwere used for lymphocyte culture anddetermination of organochlorines andcholinesterase activity. Tubes with EDTAwere used for full blood counts and tubeswithout preservatives for multiple biochemicalanalysis. Urine samples (100 ml) werecollected into 120 ml sterile plastic bottlesand transported in a container with ice. In thelaboratory, the urine samples were storedat -80'C until processing (up to four months)for urine mutagenicity.

INDUCTION OF MICRONUCLEIPeripheral blood lymphocytes were separatedfrom whole blood with a sterile gradient solu-tion of Lymphoprep (sodium metrizoate,9-6% w/v; polysaccharide, 5'6% w/v) with amean (SD) density of 1-077 (0.001) g/ml andosmolarity of 280 (15) mOsm (NycomedPharma AS, Oslo, Norway). The cytokinesis-block method as described by Fenech andMorley (1985) was used.9 Briefly, lympho-cytes were cultured in Roswell Park MemorialInstitute (RPMI)-1640 medium (Sigma

Chemicals, USA) supplemented with 15%foetal calf serum (FCS, CommonwealthSerum Laboratories, Australia), and stimu-lated to divide by addition of purifiedphytohaemagglutinin (2 ,g/ml, PHA HA-16,Wellcome Diagnostics, France), in a concen-tration of 0-75 x 106 cells/ml of medium, in 24well tissue culture plates (Linbro, FlowLaboratories). Two cultures for each subjectwere prepared, and cells were cultured at37°C in a humidified atmosphere containing5% CO2 for 72 hours. After 44 hours ofculture, Cytochalasin-B (Cyt-B, SigmaChemicals, USA) was added at a concentra-tion of 5 ug/ml. After 72 hours of culture, cellswere harvested and slides were prepared from120,1 of cell suspension (2/culture) with acytospin Shandon, England; (600 rpm, 5min). The slides were stained with Diff-QuikStain Set (modified Wright stain, Lab-Aids,cat 64851, Sydney, Australia), air dried for aminimum of two hours and mounted withcover slips, in Gurr XAM neutral medium(BDH Chemicals UK). A minimum of 1200binucleated cells/subject were scored underoil immersion at 1000 x magnification. Allslides were scored blind. Micronuclei werescored in binucleated cells by the criteria. (a)diameter between 1/16 to 1/3 that of the mainnuclei; (b) non-refractile; (c) not linked to themain nuclei; (d) morphologically identical tobut smaller than the main nuclei.

URINE MUTAGENICITYAmberlite XAD-2 resin (BDH Chemicals,UK, 20-50 mesh) was washed with acetoneonce, with methanol twice, and several timeswith distilled water. The resin was stored at4°C and used as required. The urine sampleswere centrifuged for 10 min at 800 rpm andthe supernatant was passed through the chro-matography column containing 2 ml ofXAD-2 resin, at a rate of 2 ml/min. Adsorbedmutagens were eluted with 6 ml of acetoneand collected in two 5 ml glass tubes/sample.Samples were freeze-dried and the urineresidue was reconstituted in 1-5 ml ofdimethyl sulphoxide. All media and solutionsused in this assay were prepared as describedby Maron and Ames.'0 Two strains ofSalmonella typhimurium (TA100 and TA98),with and without metabolic activation (S9mix, male Sprague-Dawley rats treated withAroclor 1254) were used with two differentvolumes of urine extract (50 and 100,pl). Atotal of 50 urine extracts were tested in fivedifferent batches (10/batch) with appropriatepositive (2-5 pg/plate 2-aminoanthracene and2 pg/plate sodium azide in the presenceand absence of S9) and negative (dimethylsulphoxide) controls.

HAEMATOLOGYThe following haematological variables wereevaluated: haemoglobin, red cell counts,packed cell volume, mean cell volume, meancell haemoglobin concentration, plateletcounts, whole blood counts, differential whitecell counts, and erythrocyte sedimentationrate.

701

on Novem

ber 3, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.10.700 on 1 October 1994. D

ownloaded from

Barbosa, Bonin

BIOCHEMISTRY-SPECIAL AND MULTIPLEANALYSESTwo special tests were performed: wholeblood organochlorines and serum andwhole blood cholinesterase activity. Wholeblood was screened for the presenceof 13 organochlorines: hexachlorobenzene,Heptachlor epoxide, y-chlordane, a-chlor-dane, Endrin, DDE, Lindane, Oxychlordane,Heptachlor, Aldrin, DDD, DDT, andDieldrin, with: a limit of detection of 1 ,ug/l.

Multiple biochemical tests, conductedaccording to routine pathology laboratoryanalyses, were performed on blood samplesfrom 31 fumigators and -17 controls (foursamples of the control group were determinedto be unsatisfactory for analysis) and included19 variables: sodium, potassium, chloride,bicarbonate, urea, creatinine, uric acid, glu-cose, protein, albumin, total bilirubin, alka-line phosphatase, y-glutamyl transpeptidase,serum aspartate aminotransferase (also knownas serum glutamate oxaloacetate transami-nase), serum alanine aminotransferase (alsoknown as serum glutamate pyruvate transami-nase), calcium, inorganic phosphate, choles-terol, and triglycerides.

STATISTICAL ANALYSISThe Wilcoxon rank sum test was used to com-pare the frequencies of micronuclei betweenfumigations and controls, smokers and non-smokers, and younger (< 35 y) and older(> 35 y) age groups. The X2 test was used tocompare the incidences of abnormal liverfunction tests between fumigators and con-trols. A one sided unpaired t test was used toevaluate the significance of the mutagenicresponse in the urine mutagenicity assay.

ENVIRONMENTAL MONITORINGPhosphine concentration in the breathingzone of fumigators was recorded during 8fumigations, with phosphine badges (Drager,Germany) with detection limits ranging from0-01 to 2 4 ppm/h, depending on the durationof the measurement (30 min-8 h). Just beforephosphine release, each fumigator involved in aspecific fumigation procedure placed a badge(in the appropriate housing) at the front of thecollar as close as possible to the breathingzone. On the back of each badge was recordedthe date, name of the user, finigation tech-nique and times of start and finish of the oper-ation. During each fumigation that wasmonitored with badges, we also used phos-phine tubes (Drager, Germany), attached to agas detector pump (Drager) with detection

Table 1 Mean (SD) incidence of micronuclei in phosphine fumigators and controlsexposed to medication and x rays

Fumigators Controls

M+ M- X+ X- F M+ M- X+ X- C(n = 6) (n = 25) (n = 2) (n = 29) (n = 31) (n = 4) (n = 17) (n = 5) (n = 16) (n = 21)

8-2# 6-6 6-7 7 0 6-9 6-3 7-3 5-3 7-6 7-1(4-7) (4 3) (30) (44) (44) (40) (40) (3 6) (40) (3-9)

M+ = medicated; M- = non-medicated; X+ = exposed to x rays; X- = not exposed to x rays;F = total fumigators; C = total controls.

limits of 0-1 to 40 ppm. These tubes served toovercome the limitations of the badges, whichdo not detect short term peak concentrations,and were used to obtain readings at a mini-mum of five and a maximum of 10 samplingpoints, based on the perceived relative risk ateach point, for each fumigation.When we used badges for monitoring, the

manufacturer's recommendations for calcu-lating the average phosphine concentration inppm were followed, where the reading inppm x h, obtained from a printed colourcomparison code on the badge, was dividedby the exposure time (h):

reading in ppm x hPH3 concentration in ppm = measuring period (h)

ResultsAll fumigators and controls were assessedfor history of illness, use of medication andexposure to x rays in the past 12 months, aswell as for diet and alcohol intake. No ill-nesses were recorded in fumigators or controlsexcept for one fumigator who reported anoccasional respiratory allergy. Of the 52 sub-jects, 10 received some form of medication,mainly antitetanus vaccine and vitamins andseven had been exposed to x rays in the past12 months. These two events were deter-mined to have no confounding effect on theincidences of micronuclei in the respectivegroups (table 1).The figure shows that the incidence of

micronuclei in peripheral blood lymphocytesof 52 subjects divided into three comparativegroups; (a) phosphine funigators (n = 31)and controls (n = 21); (b) smokers (n = 19)and non-smokers (n = 33), and (c) younger(<35 y, n = 16) and older (>,35 y, n = 36)subjects. In the first two groups, there wereno significant differences in the incidenceof micronuclei/1000 binucleated cellsbetween fumigators (6-9 (4 5) and controls(7-1 (4-0) Wilcoxon, P = 0-88), and smokers(7-2 (3 9) and non-smokers (6-8 (4 2)

CO 14Z P 000o1, 12 _ P=0-88 P=0.8

10 i0 10

4E8.0 2-

0

o LO LOo 0 0 0 CC) CDa) 0 0 V A\-.0 E E

L C0z

Frequency of micronuclei (mean (SD)) in fumigators vcontrols, smokers v non-smokers, and < 3 5 v > 35 ycompared by Wikoxon test.

702

on Novem

ber 3, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.10.700 on 1 October 1994. D

ownloaded from

Evaluation ofphosphine genotoxicity at occupational levels of exposure in New South Wales, Australia

Table 2 Urine mutagenicity in fumigators and controls

Fumigators (n = 27) Controls (n = 19)Degree ofmutagenicity* Smoker Non-smoker Smoker Non-smoker

0 - 12 1 8+ 4 3 - 4++ 1 2 - -+++ 5 5 1Total 10 17 6 13

*Mutagenicity was assessed as follows: 0 = no significantactivity in either TA100 or TA98 ±S9; + = 1 -5x background(P < 0 05; one-sided unpaired t-test); + + = 2x background(P < 0-01); + + + = 3x background (P < 0-001). Thespontaneous revertant ranges for TA98 were 17-33 (-S9) and31-50 (+ S9) and for TA100, 153-178 (-S9), and 148-169(+ S9). The positive control, 2-aminoanthracene, consistentlyinduced about 2500 revertants when applied at 2-5 jg/plate inthe presence of S9, in all experiments.

Wilcoxon, P = 0 8). In the third group, how-ever, there was an age related significant dif-ference, with the younger subjects having alower incidence of micronuclei/l000 binucle-ated cells (4 5 (3 4)) than the older subjects(8X1 (4 0), Wilcoxon, P < 0-001).

Table 2 summarises the urine mutagenicityin Salmonella typhimurium strains TA100 andTA98 (±S9). Of a total of 27 fumigators, 15(56%) had mutagenic activity compared with10 (53%) out of 19 control subjects. Thisactivity seemed to be unrelated to phosphinefumigation but there was a relation withsmoking: 100% of the fumigators and 83% ofthe controls who smoked excreted mutagenicurine.

Haematological tests on 31 fumigators andfour controls were consistent with normal val-ues, but there was a slightly raised packed cellvolume in one fumigator and one control sub-ject. Whole blood and serum cholinesterasewere within the normal range for all subjects(3-6-6-3 KU/i and 1-0-4-3 KU/l) and concen-trations of organochlorines were withinacceptable limits (Work Cover, New SouthWales).

Table 3 Percentage offumigators and controls with raised liverfunction variables

Liver Normal Fumigators Controlsvariables range n = 31 (%a) Range n = 17 (%) Range

Total bilirubin 3-18 prmol/l 2 (6-5) 19-21 0 -

Alkaline phosphatase 30-120 U/I 1 (3-2) 173 1 (5-9) 167y-Glutamyltranspeptidase 0-50U/R 11 (35-54) 53-163 3 (17-6) 52-117Alanine aminotransferase 0-45 U/I 8 (25 8) 47-156 2 (11-7) 48-71Aspartate aminotransferase 0-45 U/I 1 (3-2) 47 0> 1 Variable* 17 (54-8) - 6 (35-3)

*Subjects with one or more raised liver function variables.

Table 4 Environmental monitoring ofphosphine fumigation sites

Date Type Type Individual measurements*ofl of 01 Durationfumigation fumigation store (A) (B) (C) (h)

11/2/92 Gas Sealed 1-6 2-4 < 0-1 102/4/92 AIP pellets Bunker < 0-it < 01it < 0-it 0-202/4/92 SIROFLO Unsealed < 0-1 < 0-1 < 0-1 0-813/2/93 Blankets Sealed 0-4 0-1 -- 114/2/93 Blankets Bunker < 0-1 0-4 0-8 116/2/93 AIP tablets Unsealed 0-8 1-2 1-2 201/6/93 SIROFLO Unsealed < 0-1 < 0-1 < 0-1 101/6/93 SIROFLO Sealed < 0-1 < 0-1 < 0-1 0-7

*= Badges, ppm/h; t= fumigation time < 30 min was unsuitable for monitoring with phosphinebadges (measurements with phosphine tubes showed values < 0-4 ppm for no longer than 10min); t = monitoring with phosphine tubes showed values < 2-4 ppm during the last hour of theprocedure.

Table 3 shows biochemical results fromfumigators and controls. Overall, 54-8% offumigators and 35 3% of controls had one ormore increased liver function variables.Furthermore, 51 6% of fumigators had raisedcholesterol (range 5-7-7 1 mmol/l) v 47% ofcontrols (range 5-7-6-4 mmol/l); 45-2% offumigators had raised triglycerides (range2-2-4 3 mmol/l) v 52-9% of controls (range2-2-3-8 mmol/l); 19-3% of fumigators hadraised glucose (range 3 1-9-6 mmol/l) v 5 9%of controls (range 3-1-6-0 mmol/l); 12-9% offumigators had raised urea (range 2-5-3-2mmol/l) v 5 9% of controls (2-7 mmol/l);53-1% of fumigators had altered concentra-tions of some of the liver function variables(y-glutamyl transpeptidase, serum aspartateaminotransferase, serum alanine aminotrans-ferase, total bilirubin, alkaline phosphatase) v35*3% of controls. All other variables werewithin the normal range.The results of environmental monitoring

showed phosphine exposures that varied inconcentration and duration (table 4), depend-ing on the techniques used. In the threeSIROFLO fumigations monitored, no phos-phine was detected in the work environment,whereas fumigation with blankets, either insealed stores or bunkers, showed detectableconcentrations varying from 0 1 to 0 8 ppmover a maximum period of one hour. In theunsealed store where fumigation was with AlP(aluminium phosphide) tablets by the "probeand sheet" method, phosphine badges (detec-tion limits of 0-01 to 2-4 ppm) worn by threefumigators showed results of 0-8, 1 2, and 1 2ppm/h. This technique took two hours to beperformed. During the last hour of the proce-dure, concomitant monitoring with Dragertubes (range 0-1 to 40 ppm) showed a mean(SD) concentration of 2 (0 5) ppm.

DiscussionPhosphine has been implicated in the induc-tion of chromosomal aberrations in phosphinefumigators at occupational exposures of0.1-0.9 mg/M3 (007-0-63 ppm) in open airand 0-4-5-8 mg/M3 (0-28-4 11 ppm) inenclosed spaces.8 Analyses of chromosomalaberrations and micronuclei show that theyare strongly associated, especially when thedamage is due to breaks or aneuploidy.Formation of micronuclei is an indirect indi-cator of chromosomal damage and is compa-rable in sensitivity to metaphase analysis."The results of the incidences of micronuclei inphosphine fumigators in this study do notassociate occupational exposure to phosphinewith any clastogenic or aneugenic effect inlymphocytes. When data from fumigators andcontrols are pooled and analysed by age ( <35 and > 35), the increase in frequency ofmicronuclei seen in the older group is signifi-cant (Wilcoxon, P < 0x001) compared withthe younger group. These results are inaccordance with several other reports onincreased DNA damage with age.'2 13

Urine mutagenicity has been successfullyapplied as a screening technique to detect

703

on Novem

ber 3, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.10.700 on 1 October 1994. D

ownloaded from

Barbosa, Bonin

occupational exposure to mutagenic sub-stances'4 1' and dietary mutagens.'6 Ourresults showed no significant difference inexcretion of mutagens in fumigators and con-trols. It is important, however, to considerthat due to logistical limitations, some of theurine samples could not be collected in the 24hour period after fumigation-neither werethere pooled 24 hour specimens, as recom-mended by Falck et al.14 Nevertheless, whendata from all subjects (fumigators and con-trols) were analysed for the influence of smok-ing habits, it was clear that cigarette smokingincreased urine mutagenicity, where 100% offumigators who smoked had mutagenic urinecompared with 29% of fumigators who didnot smoke. It is also important to note thatnone of the fumigators who did not smokehad more than double the background muta-tion frequency whereas 50% of the fumigatorswho smoked trebled their mutation fre-quency, compared with the spontaneousbackground. Of the control subjects whosmoked, 83% had mutagenic urine comparedwith 38% of controls who did not smoke.Furthermore, 100% of the controls whosmoked and had mutagenic urine had a three-fold increase, whereas 80% of the controlswho did not smoke but had mutagenic urinehad only a 1*5-fold increase over background.This is consistent with the well documentedinfluence of cigarette smoking on urine muta-genicity.'4 17The results of multiple biochemical analy-

ses showed a variety of mild changes thatcould be associated with phosphine exposure.Serum cholinesterase activity was within thenormal range for all subjects investigated inour study. Inhibition of cholinesterase activityhas been found in acute phosphine poisoningwhere ingestion of aluminium phosphidetablets resulted in a high percentage of caseswith decreased serum cholinesterase activity." 18On the other hand, 31 crew members of agrain freighter accidentally exposed throughinhalation showed normal concentrations ofserum cholinesterase activity except for onesubject.'9 Our results do not show any associa-tion between inhibition of serumcholinesterase activity and phosphine expo-sure. Recently, Potter et al reported a signifi-cant decrease in anticholinesterase and(butyryl)cholinesterase activity in phosphinefumigators, and suggested a dose-relatedeffect.20 We think, however, that some of thepreviously reported levels of exposure' 2 8 wereprobably higher than most of the levels cur-rently found, due to the poorer work practicesand technology that existed at that time.Nevertheless, our data show that at the cur-rent occupational levels of exposure encoun-tered in the New South Wales stores, such aneffect is not present.

Increases in serum transaminases anddecreases in serum cholinesterase activityhave been found in oral phosphine poisoning.3Although such exposure would be expected tocause liver damage with a subsequent increasein serum transaminase concentrations, theeffect on transaminase concentrations at a low

level of chronic exposure still needs furtherinvestigation. In our study, 29% of fumigatorshad raised transaminases (serum aspartateaminotransferase, serum alanine aminotrans-ferase) compared with 11-7% of controls,which may be associated and consistent withliver damage due to phosphine toxicity.6 Anincrease in yglutamyl transpeptidase thatoccurred in 34-4% of fumigators comparedwith 17-6% of controls, may be partlyexplained by alcohol consumption, where77-4% of the fumigators and 61-9% of con-trols reported a medium (216-432 g/week) tohigh (450-648 g/week) alcohol intake.Cholesterol and triglycerides were raised infumigators (51 6% and 45-2%, respectively)as well as in controls (47% and 52-9%,respectively). This suggests that diet (data notshown) is a significant contributor to theseincreases. Considering liver function in gen-eral, 53 1% of fumigators had an increase ofone or more variables compared with 35.3%of controls. Although this difference was notsignificant (X2 = 1-68, P > 0-19), furtherinvestigations are needed to evaluate if- thereare any possible biological implications. Thehaematological variables we evaluated werewithin the normal range, consistent with otherfindings in humans,5 but in contrast to animalstudies.7

In spite of our small sample, the results ofenvironmental monitoring showed that theexposures experienced by the fumigators inthis study are far lower than those reportedpreviously. 28 Although phosphine is a potentpoison, our results show that it is possible tokeep exposure to this fumigant at levels thatdo not present detectable health risks as mea-sured by the genotoxicity biomarkers used inthis study. Furthermore, monitored concen-trations of phosphine in the work environ-ment showed that the SIROFLO fumigationtechnique resulted in the lowest occupationalexposures whereas the probe and sheetmethod resulted in the highest. The probeand sheet method is being phased out but theSIROFLO technique is increasing in use, cur-rently accounting for about 70% of all fumiga-tions conducted by Grain Corp. Undercurrent occupational exposures encounteredin grain stores in New South Wales, the con-centration of phosphine as measured by us ( <2-4 ppm over a period of one hour) has notbeen shown to elicit any detectable adverseeffects, as measured by induction of micronu-clei, urine mutagenicity, and standard haema-tological analysis.

We thank Dr Robert Baker for his significant input during theinitial phase of this project. We are also indebted to Mr CarlosCorvalan for assisting in the statistical analyses of the data, DrNeill Stacey for his participation during part of this work, andDr Grahame Budd for critical evaluation of the manuscript.This research was supported by Grain Research andDevelopment Corporation (GRDC), the University of Sydney,Conselho Nacional de Desenvolvimento Cientifico eTecnologico (CNPq-Brazil), and the New South Wales GrainCorporation.The views expressed in this article are those of the authors

and do not necessarily reflect those of the NationalOccupational Health and Safety Commission.

704

on Novem

ber 3, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.10.700 on 1 October 1994. D

ownloaded from

Evaluation ofphosphine genotoxicity at occupational levels ofexposure in New South Wales, Australia

1 Jones AT, Jones RC, Longley EO. Environmental andclinical aspects of bulk wheat fumigation with alu-minium phosphide. Am Ind HygAssocJI 1964;25:376-9.

2 Zaebst DD, Blade LM, Burroughs GE, Morrelli-SchrothP, Woodfin WJ. Phosphine exposure in grain elevatorsduring fumigation with aluminum phosphide. AppliedIndustrial Hygiene 1988;3:146-54.

3 Singh RB, Rastogi SS, Singh DS. Cardiovascular manifes-tations of aluminum phosphide intoxication. AssocPhysicians India 1989;37:590-2.

4 Chugh SN, Singhal HR, Girdhar NK, Arora BB, MalhotraKC. Aluminium phosphide poisoning: Analysis of 228cases. JAssoc Physicians India 1989;37:28-9.

5 Chungh SN, Ram S, Mehta LK, Arora BB, Malhotra KC.Adult respiratory distress syndrome followingaluminium phosphide ingestion. Report of 4 cases.Assoc Physicians India 1989;37:271-2.

6 World Health Organization. Phosphine and selected metalphosphides. International programme on chemical safety.Geneva: WHO, 1988. (Environmental health criteriaNo 73.)

7 Newton PE, Schroeder RE, Sullivan JB, Busey WM,Banas DA. Inhalation toxicity of phosphine in the rat:acute, subchronic, and developmental. InhalationalToxicology 1993;5:223-39.

8 Garry VF, Griffith J, Danzl TJ, Nelson RL, Whorton E,Krueger LA, Cervenka J. Human genotoxicity: pesticideapplicators and phosphine. Science 1989;246:251-5.

9 Fenech M, Morley AA. Measurement of micronuclei inlymphocytes. Mutat Res 1985;147:29-36.

10 Maron DM, Ames BN. Revised methods for theSalmonella mutagenicity test. Mutat Res 1983;113:173-215.

11 Klisech U, Danford N, Adler D. Micronucleus test andbone-marrow chromosome analysis. A comparison of 2

methods in vivo for evaluating chemically induced chro-mosomal alterations. Mutat Res 1981;80:21-32.

12 Fenech M, Morley AA. The effect of donor age on sponta-neous and induced micronuclei. Mutat Res 1985;148:99-105.

13 Richard F, Aurias A, Couturier J, Dutrillaux A, Fhiry-Herard A, Gerbault-Sereau M, et al. Aneuploidy inhuman lymphocytes: an extensive study of eight individ-uals of various ages. Mutat Res 1993;295:71-80.

14 Falck K, Grohn P, Sorsa M, Vainio H, Heinonen E, HolstiLR. Mutagenicity in urine of nurses handling cytosticdrugs. Lancet 1979;i:1250-1.

15 Mielzynska D, Snit M. Urine mutagenicity in workersdirectly employed in coke production. Polish Journal ofOccupational Medicine 1992;5:363-71.

16 Poyen D, De Meo MP, Botta A, Gouvernet J, Dumenil G.Handling of cytostatic drugs and urine mutagenesis. IntArch Occup Environ Health 1988;61:183-8.

17 Baker RSU, Darnton-Hill I, Bonin AM, Arlauskas A,Braithwaite C, Wootton M, Truswell AS. Urine muta-genicity as an indicator of exposure to dietary mutagensformed during cooking of foods. Environ Health Perspect1986;67: 147-52.

18 Rastori P, Raman R, Shanker A. Serum cholinesteraseactivity and brain tissue (acetyl cholinesterase activity) inaluminium phosphide poisoning. JAssoc Physicians India1989;37:28-9.

19 Wilson R, Lovejoy FH, Jaeger RJ, Landrigan PL. Acutephosphine poisoning abroad a grain freighter.Epidemiologic, clinical and pathological findings. JAMA1980;244: 148-50.

20 Potter WT, Garry VF, Kelly JT, Tarone R, Griffith J,Nelson RL. Radiometric assay of red blood cell andplasma cholinesterase in pesticide appliers fromMinnesota. ToxicolAppl Pharmacol 1993;119:150-5.

Rejected manuscriptsFrom February 1994, authors whose sub- be returned to them. The Journal will destroymutted articles are rejected will be advised of remaining copies of the article but corres-the decision and one copy of the article, pondence and reviewers' comments will betogether with any reviewers' comments, will kept.

705

on Novem

ber 3, 2020 by guest. Protected by copyright.

http://oem.bm

j.com/

Occup E

nviron Med: first published as 10.1136/oem

.51.10.700 on 1 October 1994. D

ownloaded from